High Temp Corrosion

Name ///sr-nova/Dclabs_wip/High Temp/5208_1.
pdf/Chap_01/
26/10/2007 1:51PM Plate # 0
High-Temperature Corrosion And Materials Applications

George Y. Lai, editor, p1
DOI: 10.1361/hcma2007p001
Copyright 2007 ASM International

All rights reserved.
www.asminternational.org
CHAPTER 1
Introduction
METALS AND ALLOYS will react during
high-temperature service with the surrounding
environment, resulting in high-temperature
corrosion. In gaseous environments, high-temperature corrosion is dened as the corrosion that
takes place above the maximum temperature
at which acids condense and dew-point corrosion takes place. Although a majority of hightemperature corrosion reactions take place at
temperatures above 500 C (930 F), severe
high-temperature corrosion has been encountered in many cases at temperatures below
500 C (930 F). In waste-to-energy boilers, for
example, carbon and low-alloy steels have
experienced severe reside corrosion problems
in the waterwalls of the boilers at the tube metal
temperatures of approximately 260 to 315 C
(500 to 600 F).
This book is intended primarily for engineers
and metallurgists who are concerned with hightemperature materials problems in the following
industries: aerospace/gas turbine, chemical processing, rening and petrochemical, fossil-red
power generation, coal gasication, waste-toenergy industry, pulp and paper, heat treating,
mineral and metallurgical processing, and
others. The technical data presented in this book
are pertinent to real materials problems related
to the aforementioned industries. The book will
also be useful for both undergraduate and graduate students who are interested in studying or
pursuing research on the subject of hightemperature corrosion.
The book covers eight basic modes of hightemperature corrosion. A brief description of
thermodynamics is included for most chapters to
help readers to understand the corrosion reactions. The external stresses (or strains) can cause
alloys to suffer preferential corrosion penetration
attack in a certain corrosive environment, such as
suldizing environments. In addition, external
stresses or residual stresses can cause the alloy
to suffer brittle, intergranular cracking when
exposed to the lower end of the intermediate
temperatures for certain alloys. This type of
cracking is frequently referred to as reheat
cracking, stress-relaxation cracking, or

strain-age cracking (for nickel-base alloys).
Both of these subjects are covered in Chapter 14,
Stress-Assisted Corrosion and Cracking. The
subject of erosion and erosion/corrosion is also
reviewed with an attempt to offer readers general
guidance on materials selection and application.
Discussion also includes hydrogen attack of
carbon steels in boilers and renery equipment.
Finally, extensive discussion on the materials
problems in coal-red boilers, oil-red boilers,
waste-to-energy boilers, and black liquor recovery boilers is included. In summary, the subjects covered extensively in this book include:

Oxidation
Nitridation
Carburization and metal dusting
Corrosion by halogen and hydrogen halides
Suldation
Hot corrosion
Molten salt corrosion
Liquid metal corrosion and embrittlement
Erosion and erosion/corrosion
Stress-assisted corrosion and cracking
Hydrogen attack
Coal-red boilers
Oil-red boilers and furnaces
Waste-to-energy boilers and waste incinerators
Black-liquor recovery boilers
The focus of this book is on commercial

alloys, including both generic and proprietary
alloys. Most data are presented to reveal alloy
ranking and thus serve as a general guide to
materials selection and application. Engineers
can thus use the data and information to compare alloys that are commercially available.
The effects of alloying elements, temperature,
and environmental conditions on the corrosion
behavior of alloys are also discussed, providing
information about the capability of an alloy in
terms of useful temperature limitation. Trademarks for alloys and alloy manufacturers are
listed in Appendix 1. The compositions of alloys
are tabulated in Appendix 2.
pg 1
Name ///sr-nova/Dclabs_wip/High Temp/5208_3-4.pdf/Chap_02/
26/10/2007 12:10PM Plate # 0

George Y. Lai, editor, p3-4
DOI: 10.1361/hcma2007p003

CHAPTER 2
Challenges in Materials Applications

for High-Temperature Service
IN MANY INDUSTRIAL SYSTEMS, plant
operating conditions can be quite complex; it is
rather difcult to use laboratory tests to simulate
plant conditions. However, laboratory tests can
provide good general guidance for making preliminary alloy selection. In situ eld testing or
eld trials of candidate alloys in the operating
plant provides the best way of obtaining the
corrosion information that can be reliably used
for nal materials selection.
During the preliminary alloy selection process, it is important to evaluate not only the
high-temperature corrosion resistance of the
alloy, but also its mechanical properties such as
tensile and creep-rupture strengths. The microstructural changes at the application temperatures such as thermal stability of the alloy, should
also be considered. For example, duplex stainless
steels are known to suffer 475 C (885 F)
embrittlement caused by the formation of alpha
prime (0 ) coherent precipitates. Accordingly,
these stainless steels should be avoided for
use as a structural component at temperatures
approximately, above 340 C (650 F). ASME
Codes may have lower maximum service temperature limits for these alloys. Consideration
should also be given to fabrication issues, such
as weldability and welding procedures, annealing heat treatments, postweld heat treatment
(PWHT) and stress relieving, and codes and
standards requirements.
The availability of the alloy can also be an
issue. It is not uncommon to nd that some
alloys are no longer commercially available
in stock due to a number of reasons, which may
include poor market demands in the past, difculty in manufacturing, and so forth. In some
cases, the alloy may only be available on order
for a whole production heat, which can be tens
of thousands of pounds of material. Another
important factor is the alloy price. A cost
analysis needs to be conducted to balance the

material cost with the expected life for the
component to ensure the alloy is cost effective.
Often the life-cycle cost is a better criterion than
the initial material cost in making an alloy selection.
Selection of an appropriate ller metal for
welding is important for component fabrication
involving welding. Normally, it is a simple process when the candidate alloy has a ller metal
with matching chemical composition. However,
many high-temperature alloys do not have ller
metals with matching chemistries. The widely
used Fe-Ni-Cr alloy 800H is a good example.
Many heat-resistant cast alloys also do not have
matching ller metals. Thus, when no matching
chemistry ller metal is available for welding, it
is critical to select a ller metal that not only
possesses excellent weldability but, also exhibits
comparable or better high-temperature corrosion
resistance along with comparable strengths,
thermal stability, and other relevant properties.
Some fabricators sometimes use weldability to
select a ller metal without considering the
resistance of the weld metal to the specic hightemperature corrosive environment in the end
application. This can lead to premature failures.
For example, because of their good weldability,
high nickel ller metals, such as ller metal alloy
82 (ERNiCr-3), are sometimes used for welding
the alloys that are to be in service in suldizing
environments. This can cause preferential suldation attack at the weld joint because of the
relatively poor suldation resistance of high
nickel alloys.
Welding can still be an issue for some hightemperature alloys even with matching ller
metals. This is because some high-temperature
alloys contain many alloying elements for various metallurgical reasons, such as improving the
resistance to a certain mode of high-temperature
pg 3
26/10/2007 12:10PM Plate # 0
4 / High-Temperature Corrosion and Materials Applications
corrosion, increasing tensile and creep-rupture

strengths, or increasing wear resistance. Increasing the levels of some of these alloying
elements can increase the difculty in the weldability of the alloy. For example, an alloy containing high silicon, high aluminum, high
carbon, or very high chromium can be difcult to
weld even though a matching ller metal is
available.
For construction of a component, engineers
have the option to consider whether a wrought
alloy or a cast alloy will be more suitable metallurgically and/or economically for the intended
high-temperature application. Engineers may
also consider a totally different approach to
address the high-temperature corrosion issue
for some existing plant equipment that has
suffered corrosion. In reneries, many reactor
vessels, such as crude towers, hydrocrackers,
and hydrodesulfurizers, are made of clad plates
with a corrosion-resistant cladding in original
installations. Cladding can be corroded after
years of operation. One common approach is to
refurbish the corroded vessels by applying a
corrosion-resistant weld overlay instead of replacing it with a new construction. This
approach has been adopted in the boiler industry
in recent years to address the severe corrosion
problems with the waterwalls of boilers in
waste-to-energy boilers, coal-red boilers, basic
oxygen furnace hoods in steel mills, and so forth.
With automatic controls for gas metal arc welding machines, a large scale of weld overlay can
be applied in vessels or boilers with engineering
quality. Laser cladding can also be applied in the
shop on large equipment such as waterwall
panels. Coextruded composite tubes with a
corrosion-resistant alloy cladding on the outer
diameter have long been available for construction of waterwalls as well as superheaters in
boilers. Composite tubes manufactured by a
spiral weld overlaying process have been made
available in recent years. These composite tubes
use the outer diameter cladding for providing
corrosion protection and the substrate base tube
for the load-bearing structural part. Most of these
composite tubes are used in superheaters and
reheaters in boilers with metal temperatures
being likely less than about 650 C (1200 F).
Many furnace tubes used in petrochemical processing, such as ethylene cracking furnace tubes,
are exposed to temperatures higher than 980 C
(1800 F) and carburizing gas streams on the
internal diameter (ID) of the tube, application
of composite tubes with a carburization- and
coking-resistant alloy cladding on the tube ID
can potentially increase the operating temperature and/or prolong the tube life.
Aluminide coatings reportedly have been used
in ethylene cracking furnace tubes. At the writing
of this book, it appears no commercial companies
in the United States provide aluminizing coating
services for ethylene furnace tubes or pipes.
Another diffusion coating, chromized coating,
has also reportedly been used in boilers. Both of
these diffusion coatings are very thin. Coatings
have been highly successful in providing protection against oxidation and hot corrosion for
the high-temperature components, such as airfoils, in gas turbines. The coatings used involve
aluminide coatings, overlay MCrAlY coatings
by vapor deposition processes (e.g., electron
beam physical vapor deposition), and ceramic
thermal barrier coatings (e.g., stabilized ZrO2).
Coatings are considered sacricial and are to be
replaced periodically.
pg 4

DOI: 10.1361/hcma2007p005
31/10/2007 12:43PM Plate # 0

CHAPTER 3
Oxidation
3.1 Introduction
Oxidation is the most important hightemperature corrosion reaction. Metals or alloys
are oxidized when heated to elevated temperatures in air or in highly oxidizing environments,
such as combustion atmospheres with excess air
or oxygen. Many metallic components, such as
retorts in heat treat furnaces, furnace heater tubes
and coils in chemical and petrochemical plants,
waterwalls and superheater/reheater tubes in
boilers, and combustors and transition ducts in
gas turbines, are subject to oxidation.
For many industrial processes, combustion
involves relatively clean fuels such as natural
gas or No. 1 or No. 2 fuel oil. These fuels generally have low concentrations of contaminants,
such as sulfur, chlorine, alkali metals, and
vanadium. In many cases, excess air is used to
ensure complete combustion of the fuel. The
combustion products thus consist primarily of
O2, N2, CO2, and H2O. Although alloys in these
environments are oxidized by oxygen, other
combustion products, such as H2O, may play an
important role in affecting the oxidation behavior
of the alloy. The presence of N2 in the combustion gas stream can cause signicant internal
nitridation attack under certain conditions, which
is discussed in Chapter 4 Nitridation.
Oxidation can also take place in a reducing
environment (i.e., the environment with a low
oxygen potential created by the combustion
under a substoichiometric condition). When
combustion takes place under stoichiometric or
substoichiometric conditions, the resultant
environment becomes reducing. This type of
environment is generally characterized by low
oxygen potentials. Under this condition, the
oxygen potential of the environment is typically
controlled by pH2 =pH2 O or pCO =pCO2 ratio, and
the oxidation kinetic is generally slow. The
development of a protective oxide scale can be
sluggish for most alloys. As a result, the effects of

corrosive contaminants can become more pronounced, resulting in other modes of hightemperature corrosion. For example, if the sulfur
level in the environment is high, suldation then
becomes the predominant mode of corrosion,
even though oxidation also takes part in the
corrosion reaction. Thus, a majority of hightemperature corrosion problems in reducing
environments are caused by modes of corrosion
attack other than oxidation. Most industrial
environments have sufcient oxygen activities
(or potentials) to allow oxidation to participate in
the high-temperature corrosion reaction regardless of the predominant mode of corrosion. In
fact, the alloy often relies on the oxidation reaction to develop a protective oxide scale to resist
corrosion attack, such as suldation, carburization, hot corrosion, and so forth. The oxidation
behavior under these conditions is discussed in
other chapters dealing with different modes of
corrosion attack.
There is a large spectrum of engineering alloys
available for applications in different temperature ranges. This chapter presents a large oxidation database for a wide spectrum of engineering
alloys, ranging from carbon and Cr-Mo steels,
which serve the low end of the temperature
spectrum, to superalloys serving the highest
temperature regime. The data are organized by
alloy groups to help readers to compare alloys
within the same alloy group and also to compare
alloys between different alloy groups. The focus
is to present comparison data, thus allowing
readers to consider candidate alloys for applications in the temperature regime of interest. Also
included are some important metallurgical and
environmental factors that can affect the oxidation behavior of the alloy.
It should be noted that a majority of the database has been generated in laboratory tests that
were conducted at temperatures higher than those
at which the tested alloys would normally be
pg 5
31/10/2007 12:43PM Plate # 0
pg 6
used. The intent for this approach was to determine the oxidation behavior of alloys within
relatively short test durations. For example,
many tests have been conducted at 980 to
1200 C (1800 to 2200 F) for stainless steels,
Fe-Ni-Cr alloys, and Ni-Cr alloys. This may
result in the alloy performance ranking based
more on scaling (metal loss) than on internal
oxidation attack which the alloys would most
likely have encountered at lower application
temperatures. An example is given in Fig. 3.1,
which illustrates an actual eld experience with a
furnace heater coil made of a Ni-Cr alloy that had
been in service for about 4 to 5 years at temperatures less than 900 C (1650 F), suffering
extensive internal oxidation attack with very little scaling (metal loss) (Ref 1). Few long-term
tests have been conducted at temperatures of
650 to 980 C (1200 to 1800 F) where most
stainless steels, Fe-Ni-Cr alloys, and Ni-Cr
alloys are used in high-temperature applications.
Furthermore, many test results were presented
as weight changes instead of actual measurements of the damage to the metal, such as the
total depth of oxidation attack including both
metal loss (thickness reduction) and internal
penetration.
3.2 Thermodynamic Considerations

3.2.1 Formation of Oxides
Thermodynamically, an oxide is likely to form
on a metal surface when the oxygen potential
(pO2 ) in the environment is greater than the
oxygen partial pressure in equilibrium with the
oxide. The oxygen partial pressure in equilibrium
with the oxide can be determined from the standard free energy of formation of the oxide.
Consider the reaction:
M+O2 MO2
3:1

aMO2
DG =7RT ln
aM pO2
3:2
Assuming the activities of the metal and the

oxide are unity, Eq 3.2 becomes:
DG =RT ln pO2
3:3
Then
pO2 =eDG
=RT
3:4
Equation 3.4 permits the determination of the

oxygen partial pressure in equilibrium with the
0.5 mm
Fig. 3.1
A Ni-Cr alloy furnace heater coil suffering extensive internal oxidation attack with little surface scaling after service for 4 to 5
years at temperatures below 900 C (1650 F). Source: Ref 1
31/10/2007 12:43PM Plate # 0
pg 7
Chapter 3:
partial pressure of oxygen in equilibrium with

Cr2O3 at 1000 C (1830 F) is about 1021 atm
from Fig. 3.2. This implies that the formation of
Cr2O3 is favored thermodynamically at 1000 C
in environments with oxygen potentials higher
than 1021 atm.
When the environment is reducing (e.g., the
environment generated by stoichiometric or
substoichiometric combustion), the oxygen
potential is controlled by either pH2 =pH2 O or
pCO =pCO2 ratio. The oxygen potential can be
determined by the reaction:
oxide from the standard free energy of formation.

The standard free energies of formation of
selected oxides as a function of temperature are
shown in Fig. 3.2. The gure also allows quick
determination of the oxygen partial pressure
(pO2 ) in equilibrium with the oxide. This oxygen
partial pressure can be read by drawing a straight
line from the point marked O on the left vertical axis of Fig. 3.2 through the free-energy line
of the oxide at the intersecting point with the
temperature of interest. This line continues to
extend until it intersects with the pO2 scale located at the right-hand side and bottom of the
Fig. 3.2. The intersecting point shows the oxygen
partial pressure in equilibrium with the oxide
of interest. If the oxygen partial pressure in the
environment is greater than the oxygen partial
pressure in equilibrium with the oxide, the oxide
is likely to form on the metal surface. Conversely,
the oxide is not likely to form. For example, the
2H2 +O2 2H2 O
3:5
The standard free energy of formation is related

to the partial pressures of hydrogen, oxygen, and
water by:
p2H2 O
2
pH2 pO2
DG =7RT ln
H2/H2O ratio
108
CO/CO2 ratio
108
106
3:6
pO
104
104
106
102
102
100
O4
Fe 3
200
+O
O3
e2
6F
oO
NiO
+
Ni
O2
O2
o+
1
102
= 2C
2C
=2
104
300
102
102
G =RT In pO2 (kJ/mole O2)
400
500
H
C
600
4- Cr +
3
700
Si +
2 Cr2O 3
=O2 3
O2
I 2O 3
O
I+
104
4- A
3
B
O
O
Ca
Mg
=2
2
B
108
=
2
O
+
2
+ O 2Ca
g
2M
Change of state Element Oxide
M
M
Melting point
M
M
B
Boiling point
B
1010
900
1014
106
1016
1000
1100
1200
200 400 600 800 1000 1200 1400 1600 1800 2000 2200 2400
10
Temperature, C
CO/CO2 ratio
1014
1012 10
1010
1012
106
2- A
= 3
106
108
104
iO 2
=S
800
1018
108
1020
OK
1022
H2/H2O ratio
1020010100 1070 1060 1050 1042 1038 1034 1030 1028 1026 1024
pO
Fig. 3.2
Oxidation / 7
Standard free energies of formation of selected oxides as a function of temperature. Source: Ref 2
31/10/2007 12:43PM Plate # 0
pg 8
Rearranging the equation results in:

pO2 =eDG
=RT
1
(pH2 =pH2 O )2
3:7
Thus, the oxygen partial pressures at various

temperatures can be determined as a function of
pH2 =pH2 O values. The pH2 =pH2 O value in equilibrium with the oxide can be read from Fig. 3.2,
using the method discussed previously, except
that the starting point for the straight line is H
and the pH2 =pH2 O value is determined from the
H2/H2O scale. For example, the oxygen potential, in terms of pH2 =pH2 O , in equilibrium with
Cr2O3 at 1000 C, is about 5 103 from Fig. 3.2.
Thus, Cr2O3 is likely to form at 1000 C
when the pH2 =pH2 O ratio in the environment
is less than 5103. The equilibrium reaction
for an environment whose oxygen potential is
controlled by pCO =pCO2 is:
2CO+O2 2CO2
3:8
The corresponding oxygen potential is:

pO2 =eDG
=RT
1
(pCO =pCO2 )2
3:9
The pCO =pCO2 value can be read from Fig. 3.2

using the method discussed previously, with the
exception that a straight line is drawn from point
C to the CO/CO2 scale.
Thus, it is possible to obtain the oxygen
potential of the environment in terms of
pO2 , pH2 =pH2 O , pCO =pCO2 , and the oxygen partial
pressure in equilibrium with the oxide of interest
from Fig. 3.2, to determine whether or not the
oxide is likely to form thermodynamically.
Figure 3.2 also illustrates the relative stability of
various oxides. The most stable oxides have the
largest negative values of G, or the lowest
value of pO2 , or the highest values of pH2 =pH2 O
and pCO =pCO2 .
It is clear from Fig. 3.2 that oxides of iron,
nickel, and cobalt, which are the alloy bases for
the majority of engineering alloys, are signicantly less stable than the oxides of some
solutes (e.g., chromium, aluminum, silicon, etc.)
in engineering alloys. When one of these solute
elements is added to iron, nickel, or cobalt,
internal oxidation of the solute is expected to
occur if the concentration of the solute is relatively low. As the solute concentration increases
to a sufciently high level, oxidation of the solute
will be changed from internal oxidation to
external oxidation, resulting in an oxide scale
that protects the alloy from rapid oxidation. This

process is known as selective oxidation. The
majority of iron-, nickel-, and cobalt-base alloys
rely on selective oxidation of chromium to form a
Cr2O3 scale for oxidation resistance. Some hightemperature alloys use aluminum to form an
Al2O3 scale for oxidation resistance.
Most oxides exhibit high melting points and
remain in a solid state for the temperature range
in which the alloys are used. If the oxide is present as a liquid state, catastrophic oxidation can
occur. Since many engineering alloys contain
many alloying elements for various metallurgical
reasons, formation of oxides that become liquid
at the service temperature should be prevented.
Table 3.1 shows the melting points of selected
oxides of alloying elements commonly found in
high-temperature alloys. Most oxides remain
solid until they reach extremely high temperatures. Oxides of molybdenum (MoO3) and
vanadium (V2O5), however, exhibit very low
melting points. Vanadium (V), which is a strong
carbide former, is often used in alloy steels for
increasing the strength of the material. However,
the amount used typically is quite small and is not
likely to form V2O5. Molybdenum (Mo) is also a
strong carbide former and is used in a small
amount to strengthen low-alloy steels (e.g., CrMo steels). It is unlikely these steels will be
affected by MoO3-related oxidation problems.
However, molybdenum is an effective alloying
element for improving the resistance of the alloy
to aqueous corrosion. Some stainless steel grades
contain molybdenum, with superaustenitic
stainless steels containing much higher levels of
molybdenum. Some nickel-base alloys contain
very high levels of molybdenum for either
aqueous corrosion resistance or solid-solution
Table 3.1 Melting points of selected oxides

for alloying elements commonly found in
high-temperature alloys
Qxide
Al2O3
CoO
Cr2O3
FeO
Mn3O4
MoO3
Nb2O5
NiO
SiO2
TiO2
V2O5
WO3
Source: Ref 3
Melting point, C (F)
2015 (3659)
1935 (3515)
2435 (4415)
1420 (2588)
1705 (3101)
795 (1463)
1460 (2660)
1990 (3614)
1713 (3115)
1830 (3326)
690 (1274)
1473 (2683)
31/10/2007 12:43PM Plate # 0
pg 9
Chapter 3:
Oxidation / 9
volatile CrO3 (Ref 58). Table 3.2 shows the

weight-loss data for Cr2O3 pellets after heating to
1000 to 1200 C (1830 to 2190 F) in dry O2 due
to formation of gaseous CrO3 by oxidation of
Cr2O3 (Ref 5). Caplan and Cohen (Ref 5) also
observed that moisture promoted volatilization
of Cr2O3. Asteman et al. (Ref 9) indicated that
high vapor pressure of CrO2(OH)2 can form by
reacting Cr2O3 with H2O in O2-containing
environments. The theoretical calculated partial
pressure of CrO2(OH)2 as a function of temperature for the O2-containing environment with
pO2 =0:9 atm and pH2 O =0:1 atm is shown in
Fig. 3.4 (Ref 9).
strengthening. The formation of MoO3 and its

effect on oxidation are discussed in section 3.2.2
Volatility of Oxides and section 3.4.17 Catastrophic Oxidation.
3.2.2 Volatility of Oxides
Some oxides exhibit high vapor pressures at
very high temperatures (e.g., above 1000 C, or
1830 F). Oxide scales become less protective
when their vapor pressures are high. Figure 3.3
shows vapor pressures of several refractory metal
oxides exhibiting high vapor pressures at temperatures above 1000 C (1830 F) (Ref 4).
Vanadium is typically used in small quantities as
a carbide former in alloy steels. Thus, the volatility of VO2 is generally of no concern in oxidation of alloys. Molybdenum (Mo) and tungsten
(W) are often used as alloying elements in signicant amounts in Ni- or Co-base alloys as
solution-strengthening elements. Formation of
WO3 or MoO3 may occur under certain conditions in some alloy systems, particularly in alloy
systems containing insufcient chromium for
forming a protective Cr2O3 scale.
A majority of engineering alloys rely on the
Cr2O3 scale to provide resistance to oxidation.
When heated to very high temperatures (i.e.,
above 1000 C), Cr2O3 can react with O2 to form
Table 3.2 Weight loss of Cr2O3 on heating in

dry O2 and Ar environments
Run
Temperature,
C (F)
Time, h
Gas
Gas ow,
mL/min
Weight
loss, mg
1
2
3
4
5
6
7
8
9
1100 (2010)
1200 (2190)
1200 (2190)
1200 (2190)
1200 (2190)
1200 (2190)
1200 (2190)
1200 (2190)
1200 (2190)
20
20
20
20
20
20
42
66
115
Dry O2
Dry O2
Dry O2
Dry O2
Dry O2
Dry O2
Dry O2
Dry Ar
Dry Ar
200
10
10
20
200
200
200
200
192
0.6
2.1
1.3
1.8
2.3
2.6
8.0
0
0
Source: Ref 5
Temperature, C
1
1500
1000
600
102
WO3
CrO3
104
Vapor pressure, atm
MoO3
106
10
VO2
1010
1012
1014
1016
1018
0.5
0.6
0.7
0.8
0.9
1.1
1.2
Inverse temperature, 103/T (K1)
Fig. 3.3
Vapor pressures of several refractory metal oxides exhibiting high vapor pressures at temperatures above 1000 C (1830 F).
Source: Ref 4
31/10/2007 12:43PM Plate # 0
pg 10
4
CrO2(OH)2, g
Log p, X, atm
10
12
CrO3, g
14
16
600
700
800
900
1000
1100
1200
1300
Temperature, K
Theoretical partial pressures of CrO2(OH)2 and CrO3 as a function of temperature for the environment with pO2 =0:9 atm and
pH2 O =0:1 atm: Source: Asteman et al. (Ref 9)
For service temperatures above 1200 C

(2190 F), the increasing volatility of oxides can
progressively cause the oxide to lose its protective capability. SiO2 and Al2O3 are the only two
oxides that are capable of forming a very protective barrier against oxidation at temperatures
above 1200 C (Ref 9). However, the SiO2 scale
may lose some protective capability by forming
gaseous SiO at low oxygen partial pressures
(Ref 10).
Inverse log
Parabolic
Oxide mass, m
Fig. 3.4
Log
Linear
Time, t
3.3 Kinetic Considerations

The kinetics of oxidation of metals and alloys
generally follow several reaction rates. Most
reactions follow a parabolic rate. Some reactions
follow a linear rate. Some other reaction kinetics
may include logarithmic and inverse logarithmic
rates. These reaction kinetics are illustrated
schematically in Fig. 3.5 (Ref 11), and a brief
summary, based on the article by Danielewski
(Ref 11), is presented in sections 3.3.1 to 3.3.3.
Fig. 3.5
Different oxidation kinetics. Source: Ref 11
decreases with increasing time due to the

increasing diffusion distance for ions. The oxidation rate is, thus, inversely proportional to the
thickness of the oxide scale:
X 2 =kt
3:10
where X is the oxide scale thickness, t is the

exposure time, and k is the parabolic constant;
when t = 0, X = 0.
3.3.1 Parabolic Kinetics

When the oxide scale forms on the metal surface, the oxidation reaction is controlled by the
diffusion of ions through the oxide scale, which
is in turn controlled by the chemical potential
gradient as a driving force. As the thickness of
the oxide scale increases, the rate of oxidation
3.3.2 Linear Kinetics

When the oxide scale forming on the metal
surface provides no protection barrier due to
oxide cracking and spalling, volatile oxides, and
molten oxidation products, the oxidation rate
generally remains constant with increasing
31/10/2007 12:43PM Plate # 0
pg 11
Chapter 3: Oxidation / 11
time. The linear oxidation kinetic rate can be

expressed by:
The inverse logarithmic rate can be expressed

by the following equation:
X =kl t
1=X =b7ki log t
3:11
where X is the mass (or thickness) of the oxide, t

is the exposure time, and kl is the linear rate
constant; when t = 0, X = 0.
3.3.3 Logarithmic and Inverse
Logarithmic Kinetics
where b and ki are constants.
3.4 Oxidation in Air, O2, and Clean

Combustion Atmospheres
3.4.1 Carbon and Cr-Mo Steels
At very low temperatures when the oxide lm

forms on the metal surface, the oxidation rate
usually follows either a logarithmic or inverse
logarithmic rate. The driving force for the oxidation is the electric eld across the oxide lm.
The logarithmic rate can be expressed by:
X =ke log (at+1)
3:13
3:12
where ke and a are constants.
Carbon and Cr-Mo steels are the most widely

used engineering materials and are used extensively for high-temperature applications in
power generation, chemical and petrochemical
processing, petroleum rening, pulp and paper
industry, industrial heating, and metallurgical
processing.
At temperatures below 570 C (1060 F), iron
(Fe) oxidizes to form Fe3O4 and Fe2O3. Above
120
110
100
1400 F
90
Weight-loss, mg/cm2
80
70
1200 F
60
50
40
30
20
1000 F
10
800 F
100
200
300
400
500
600
700
Time, h
Fig. 3.6
Oxidation behavior of plain low-carbon steel in air at 430, 540, 650, and 760 C (800, 1000, 1200, and 1400 F). Source:
Ref 12
31/10/2007 12:43PM Plate # 0
570 C (1060 F), it oxidizes to form FeO,

Fe3O4, and Fe2O3. The oxidation behavior of
carbon steel in air at 430, 540, 650, and 760 C
(800, 1000, 1200, and 1400 F) is summarized in
Fig. 3.6 (Ref 12). At 430 and 540 C (800 and
Average penetration/side, mils
10.5
Calculated continuing
penetration rate
53 mpy*
9.0
Carbon steel - 1200 F
7.5
6.0
4.5
5.3 mpy*
A242 type 1 HSLA steel - 1200 F
3.0
Carbon steel - 1000 F
1.5
2.3 mpy*
A242 type 1 HSLA steel - 1000 F 1.0 mpy*
200
400
600
800 1000
Exposure time, h
Fig. 3.7
Oxidation of carbon steel and high-strength low-alloy

(HSLA) steel in air. Source: Ref 13, reproduced
from Ref 14
1000 F), carbon steel showed very little weight

gain after exposure for 500 h. As the temperature
was increased to 650 C (1200 F), the oxidation
rate was signicantly increased. Carbon steel
suffered rapid oxidation at 760 C (1400 F),
exhibiting essentially a linear rate of oxidation
attack. Vrable et al. (Ref 13) reported oxidation
data for carbon steel and high-strength low-alloy
(HSLA) steel, as shown in Fig. 3.7 (Ref 14). At
650 C (1200 F), carbon steel suffered an oxidation rate of about 1.3 mm/yr, or 53 mpy (mils
per year). The oxidation rate is expected to be
much higher when exposed to temperatures
higher than 650 C (1200 F). Recent test results
by John (Ref 15) showed that carbon steel
exhibited about 0.25 mm/yr (10 mpy) of oxidation at 604 C (1120 F). Figure 3.7 also illustrates that HSLA steel is more oxidation resistant
than carbon steel, presumably due to minor
alloying elements such as manganese, silicon,
chromium, and nickel.
Cr-Mo steels are used at higher temperatures
than carbon steel because of higher tensile and
creep-rupture strengths as well as better microstructural stability. Molybdenum and chromium
provide not only solid-solution strengthening but
also carbide strengthening. Low-alloy steels with
chromium and silicon additions exhibit better
oxidation resistance than carbon steel. The benecial effects of chromium and silicon additions
to carbon steel are summarized in Fig. 3.8 (Ref
16). Silicon is very effective in improving the
oxidation resistance of Cr-Mo steels. Addition of
1.5% Si to 5Cr-0.5Mo steel signicantly
improved its oxidation resistance. The most
important alloying element for improving oxidation resistance is chromium. As shown in
Fig. 3.8, for 0.5% Mo steels, increasing chromium from 1 to 9% signicantly increases oxidation resistance. The 7Cr-0.5Mo and 9Cr-1Mo
steels showed negligible oxidation rates at temperatures up to 680 C (1250 F) and 700 C
(1300 F), respectively. Further increases in
chromium improve oxidation resistance even
more. Alloys become martensitic or ferritic
grades of stainless steels (400 series) when
chromium content is increased to 12% or higher.
3.4.2 Martensitic, Ferritic, and
Austenitic Stainless Steels
Fig. 3.8
Effects of chromium and/or silicon on the oxidation

resistance of steels in air. Source: Ref 16
The superior oxidation resistance of martensitic and ferritic stainless steels to that of carbon
and Cr-Mo steels is illustrated in Fig. 3.9
(Ref 17). As chromium content in the straight
pg 12
31/10/2007 12:44PM Plate # 0
Fig. 3.9
Oxidation resistance of carbon, low-alloy, and stainless steels in air after 1000 h at temperatures from 590 to 930 C (1100 to
1700 F). Source: Ref 17
chromium steels increases from 9 to 25%, resistance to oxidation improves signicantly. The
25Cr steel (Type 446) is the most oxidation
resistant among the 400 series stainless steels,
due to the development of a continuous Cr2O3
scale on the metal surface. In Fe-Cr alloys, it
appears that a minimum of approximately 18wt
% Cr is needed to develop a continuous Cr2O3
scale against further oxidation attack (Fig. 3.10)
(Ref 18). Cyclic oxidation studies conducted by
Grodner (Ref 19) also revealed that Type 446
was the best performer in the 400 series stainless
steels, followed by Types 430 (1418Cr), 416
(1214Cr), and 410 (11.513.5Cr) (Fig. 3.11).
Figure 3.11 shows that Fe-12Cr steels, such as
Types 410 and 416, showed increased rates of
cyclic oxidation above 760 C (1400 F). At
650 C (1200 F) in air, cycling from 650 to
300 C, 12Cr-1Mo steel (X20 CrMoV 12 1) steel
exhibited a thin, adherent (Fe,Cr)2O3 scale, as

observed by Walter et al. (Ref 20). The growth of
the (Fe,Cr)2O3 scale as a function of the accumulated isothermal hold time up to 1000 h is
shown in Fig. 3.12 (Ref 20). John (Ref 15)
reported that Fe-12Cr steel (Type 410) exhibited
an air oxidation rate of 0.25 mm/yr (10 mpy) at
832 C (1530 F).
Another ferritic stainless steel, 18SR (about
18% Cr), was found to be as good as, and
sometimes better than, Type 446 (25% Cr), as
illustrated in Tables 3.3 and 3.4 (Ref 21). This
was attributed to the addition of 2% Al and 1% Si
to the alloy. Furthermore, both of these ferritic
stainless steels, Type 446 and 18SR, showed
better cyclic oxidation resistance than some
austenitic stainless steels, such as Type 309 and
310, when cycled to 980 to 1040 C (1800 to
1900 F), as shown in Table 3.4.
pg 13
31/10/2007 12:44PM Plate # 0
pg 14
Fe2O3
Fe3O4
FeO
Fe
106
Fe2O3
Fe3O4
Parabolic rate constant, g2 cm4 s1
FeO
107
Fe 2Cr
Fe/Cr oxide
Fe2O3
(Fe,Cr)2O3
108
Fe 9Cr
Fe2O3
Fe3O4
109
Cr2O3
Fe 16Cr
FeFe(2 ... x)CrxO4
1010
Cr2O3
Fe 28Cr
1011
0
10
20
30
40
50
60
70
80
90
100
Alloy chromium content, wt%
Fig. 3.10
Effect of chromium content on oxidation of Fe-Cr alloys at 1000 C (1830 F) in 0.13 atm O2. Source: Ref 18
When the service temperature is above

650 C (1200 F), ferritic stainless steels, which
have a body-centered cubic (bcc) crystal structure, drastically lose their elevated-temperature
strength (both tensile and creep-rupture
strength). As a result, the application of ferritic
stainless steels becomes limited at higher temperatures. At these temperatures, alloys with a
face-centered cubic (fcc) crystal structure are
preferred because of their higher creep-rupture
strength. Nickel is added to Fe-Cr steels to stabilize the austenitic structure. The austenitic
structure is inherently stronger and more creep
resistant than the ferritic structure (Ref 22).
The 300 series austenitic stainless steels have
been widely used for high-temperature components in various industries because of their
strength and high-temperature corrosion resistance, including oxidation resistance. These
alloys exhibit higher elevated-temperature
strength than do ferritic stainless steels. Furthermore, they do not suffer 475 C (885 F)
embrittlement or ductility-loss problems in thick
sections and in heat-affected zones as do ferritic
stainless steels. Nevertheless, some austenitic
stainless steels can suffer some ductility loss
upon long-term exposure to intermediate temperatures (e.g., 540 to 800 C, or 1000 to
1470 F) due to sigma-phase formation.
The oxidation resistance of two austenitic
stainless steels, Types 309 and 310, is compared
with that of a number of ferritic stainless steels in
Fig. 3.11. Several austenitic stainless steels are
compared in Fig. 3.13 (Ref 23). Nickel improves
the resistance of alloys to cyclic oxidation.
Moccari and Ali (Ref 24) also observed the
similar benecial effects of nickel in improving
the oxidation resistance of alloys. Brasunas et al.
(Ref 25) studied the oxidation behavior of
about 80 experimental Fe-Cr-Ni alloys exposed
31/10/2007 12:44PM Plate # 0
pg 15
Oxidation resistance of several stainless steels as a function of temperature. Source: Ref 19
Thickness of the oxide layer (d ), m
Fig. 3.11
40
35
30
25
20
15
10
5
0
0
100
200
300
400
500
600
700
800
900
1000
Testing time (t ), h
Fig. 3.12
Oxidation behavior of 12Cr-1Mo steel (X20 CrMoV 12 1) at 650 C (1200 F) in air with every 8 h of exposure specimens
being cycled from 650 to 300 C. Source: Ref 20
to air-H2O mixture at 870 to 1200 C (1600 to

2190 F) for 100 and 1000 h. They observed that
increases in nickel in excess of 10% in alloys
containing 11 to 36% Cr improved the oxidation
resistance of the alloys. John (Ref 15) reported
that Type 304 and 310 exhibited 0.25 mm/yr
(10 mpy) of oxidation attack (both metal loss and
internal oxidation penetration) at 893 and 982 C

(1640 and 1800 F), respectively, in air. In
Fig. 3.13, several high-nickel alloys were found
to be more resistant to oxidation than austenitic
stainless steels. The oxidation behavior of highnickel Fe-Ni-Cr alloys and Ni-base alloys is
discussed in later sections of this chapter.
31/10/2007 12:44PM Plate # 0
Table 3.3 Cyclic oxidation resistance of several

stainless steels in air cycling to 870 to 930 C
(1600 to 1700 F) temperature range
Specimen weight changes after indicated cycles, mg/cm2
Alloy
288 cycles
409 + Al Destroyed
430
9.9
22-13-5
0.5
442
0.7
446
0.3
309
0.3
18SR
0.3
480 cycles
750 cycles
958 cycles
Destroyed
3.0
1.2
0.4
4.6
0.4
18.8
1.5
0.2
23.7
0.5
35.7
1.5
0.1
32.6
0.6
Note: 15 min in furnace and 15 min out of furnace. Source: Ref 21
Table 3.4 Cyclic oxidation resistance of several

stainless steels in air cycling to 980 to 1040 C
(1800 to 1900 F) temperature range
Specimen weight changes after indicated cycles, mg/cm2
Alloy
130 cycles
368 cycles
561 cycles
753 cycles
1029 cycles
446
18SR
309
310
0.4
0.7
24.2
1.5
0.5
1.1
77.5
11.3
0.2
1.5
178.3
29.3
7.0
2.2
242
62.8
19.4
3.0
358
107
In evaluating materials for automobile emission-control devices, such as thermal reactors

and catalytic converters, Kado et al. (Ref 26) and
Michels (Ref 27) have carried out cyclic oxidation tests on various stainless steels. In cyclic
oxidation tests performed by Kado et al. (Ref 26)
in still air at 1000 C (1830 F) for 400 cycles
(30 min in the furnace and 30 min out of the
furnace), Types 409 (12Cr), 420 (13Cr), and 304
(18Cr-8Ni) suffered severe attack. Type 420
(13Cr) was completely oxidized after only 100
cycles, although the sample did not show any
weight changes. Alloys that performed well
under these conditions were Types 405 (14Cr),
430 (17Cr), 446 (25Cr), 310 (25Cr-20Ni), and
DIN 4828 (19Cr-12Ni-2Si), as illustrated in
Fig. 3.14. When cycled to 1200 C (2190 F) for
400 cycles (30 min in the furnace and 30 min out
of the furnace), all the alloys tested except F-1
alloy (Fe-15Cr-4Al) suffered severe oxidation
attack (Fig. 3.15). This illustrates the superior
oxidation resistance of alumina formers (i.e.,
alloys that form Al2O3 scales when oxidized at
elevated temperatures). The data also illustrate
that for temperatures as high as 1200 C
(2190 F), Cr2O3 oxide scales can no longer
provide adequate oxidation resistance. Oxidation
of Fe-Cr-Al alloys is discussed in Section 3.4.7.
Kado et al. (Ref 26) also investigated oxidation behavior in a combustion environment that
simulated the gasoline engine. Their test
involved air-to-fuel ratios of 9 to 1 and 14.5 to 1
and regular gasoline that contained 0.01wt% S.
Exhaust gas taken from the exhaust manifold was
mixed with air before being piped into a furnace
retort where tests were performed. Test specimens were exposed to the mixture of exhaust gas
and air. The gas mixture consisted of 72.4% N2,
9.7% H2O, 9.93% O2, 8% CO2, and 507 ppm
NOx, when an air-to-fuel ratio of 14.5 to 1 was
used for combustion, while that coming from the
combustion using an air-to-fuel ratio of 9 to 1
consisted of 70.6% N2, 13.7% H2O, 3.21% O2,
12.5% CO2, and 34 ppm NOx. The total accumulated test duration was 200 h (400 cycles
with 30 min in the hot zone and 30 min out of
the hot zone). Test results along with air oxidation data are summarized in Fig. 3.16. There were
no signicant differences between air and
exhaust gas test environments when tested at
800 C (1470 F). All the alloys tested showed
negligible attack except Type 304, which
exhibited much more severe attack in the exhaust
environment. When the test temperature was
increased to 1000 C (1830 F), all the 400
series stainless steels with less than 17% Cr (i.e.,
Types 409, 405, 410, and 430) and Type 304
exhibited signicantly more oxidation attack in
the exhaust environment. Type 310, Type 446,
DIN 4828 (Fe-19Cr-12Ni-2Si), F-1 alloy (Fe15Cr-4Al), A-1 alloy (Fe-16Cr-13Ni-3.5Si), and
A-2 alloy (Fe-20Cr-13Ni-3.5Si) performed well.
At 1200 C (2190 F), all the alloys tested suffered severe oxidation attack with the exhaust
environment being more aggressive than air.
These authors attributed the enhanced attack to
the presence of sulfur in the exhaust gas environment, although low-sulfur (0.01%) gasoline
was used for testing. Sulfur segregation to the
scale/metal interface was detected.
In a study by Michels (Ref 27), the engine
combustion atmosphere was also found to be
signicantly more corrosive than the air-10%
H2O environment. The engine combustion
exhaust gas contained about 10% H2O along
with 2% CO, 0.33 to 0.55% O2, 0.05 to 0.24%
hydrocarbon, and 0.085% NOx. The balance was
presumably N2 (not reported in the paper). The
engine exhaust gas was piped into a furnace
retort where the tests were performed. The
results, which were generated in the air-10%H2O
and the engine exhaust environment, are shown
in Fig. 3.17. After exposure to the air-10%H2O
pg 16
31/10/2007 12:44PM Plate # 0
pg 17
10
625
600
80Ni-20Cr
60Ni-15Cr
800
22Cr-32Ni
10
Type 310
25Cr-20Ni
20
Change in weight, %
19Cr-14Ni
20Cr-25Ni
Type 309
23Cr-13Ni
30
40
Type 347
18Cr-8Ni(Cb)
50
Type 304
18Cr-8Ni
60
70
0
200
400
600
800
1000
Hours of cyclic exposure

(15 min heating 5 min cooling)
Fig. 3.13
Cyclic oxidation resistance of several stainless steels and nickel-base alloys in air at 980 C (1800 F). Source: Ref 23
environment at 980 C (1800 F) for 102 h, Type

309, Type 310, 18SR, alloy OR-1 (Fe-13Cr-3Al),
alloy 800, and alloy 601 were all relatively
unaffected. On the other hand, only 18SR and
alloy 601 were relatively unaffected by the
engine exhaust gas environment, with alloy
OR-1, Type 309, Type 310, and alloy 800 suffering severe oxidation attack. The sulfur content
in the gasoline used in this test was not reported.

The relatively high gas velocity, about 6.1 to 9.2
m/s (20 to 30 ft/s) was considered by the author to
be one of the possible factors responsible for
much higher oxidation attack in the engine
exhaust gas test.
Oxidation data generated in combustion
atmospheres is relatively limited. No systematic
31/10/2007 12:44PM Plate # 0
pg 18
Fig. 3.14
Cyclic oxidation resistance of several ferritic and austenitic stainless steels in still air at 1000 C (1830 F) for up to 400 cycles
(30 min in furnace and 30 min out of furnace). Source: Ref 26
409
0
F1
(Fe-15Cr-4AI)
Weight change, mg/cm2
430
100
DIN 4828
(19Cr-12Ni-2Si)
420
200
310
304
446
300
100
200
300
400
Number of cycles
Fig. 3.15
Cyclic oxidation resistance of several ferritic and austenitic stainless steels in still air at 1200 C (2190 F) for up to 400 cycles
(30 min in furnace and 30 min out of furnace). Source: Ref 26
31/10/2007 12:44PM Plate # 0
pg 19
1200 C (2190 F)/400 cycles

100
80
60
40
Thickness loss, %
20
1000 C (1830 F)/400 cycles
100
80
60
40
20
0
800 C (1470 F)/400 cycles
20
)
.5
Si
)
3N
(F
e-
20
r-1
3N
(F
e-
i-3
.5
i-3
31
A
2
16
r-1
A
1
In air
In exhaust gas (R = 9)
Si
0S
IN
48
31
28
4
30
44
r-4
(F
e-
15
43
)
AI
0
41
5
40
40
Fig. 3.16
Comparison of cyclic oxidation resistance between air and gasoline engine exhaust gas environments at 800, 1000, and
1200 C (1470, 1830, 2190 F) for 400 cycles (30 min in hot zone and 30 min out of hot zone). Alloy F-1 suffered localized
attack at 1200 C in engine exhaust gas. Source: Ref 26
studies have been reported that varied combustion conditions, such as air-to-fuel ratios. In
combustion atmospheres, the oxidation of metals
or alloys is not controlled by oxygen only. Other
combustion products, such as H2O, CO, CO2,
N2, hydrocarbon, and others, are expected to
inuence oxidation behavior. When air is used
for combustion, nitridation in conjunction with
oxidation can occur in combustion atmospheres
under certain conditions. This nitridation/
oxidation behavior of alloys is discussed in
Chapter 4 Nitridation. The presence of water

vapor can be an important factor in affecting
oxidation behavior of alloys. The effect of water
vapor on the oxidation resistance of alloys is
covered in Section 3.4.15.
3.4.3 Surface Chemistry versus
Bulk Chemistry
It is important to point out that the surface
chemistry may not be the same as the reported
31/10/2007 12:44PM Plate # 0
Fig. 3.17
Cyclic oxidation resistance of several ferritic and austenitic stainless steels in (a) air-10H2O at 980 C (1800 F) cycled every
2 h, and (b) gasoline engine exhaust gas at 980 C (1800 F) cycled every 6 h. Source: Ref 27
bulk chemistry. Some of the manufacturing

processes involved in the production of an alloy
product, such as plate, sheet, or tubular products,
may result in lower chromium contents at or
near the surface of the product. This is particularly important for austenitic stainless steels,
since the chromium specication range for austenitic stainless steels can be at or near the borderline for forming a continuous Cr2O3 scale. For
example, ASTM A 213/A 213M (or ASME SA213/SA-213M) specication for the chromium
range is 18.0 to 20.0% for Type 304, 16.0 to
18.0% for Type 316, 17.0 to 20.0% for Type 321,
and 17.0 to 20.0% for Type 347. The lower end of
the chromium content in these alloys is essentially at the minimum level that is considered to
be required for forming a continuous Cr2O3 scale
when exposed to elevated temperatures. A slight
surface depletion in chromium due to some
manufacturing processes may result in the
condition that a continuous, protective Cr2O3
scale cannot be formed, thus resulting in accelerated oxidation due to formation of nonprotective iron oxides.
Some manufacturing processes are prone to
producing the nal nished product with surface
depletion of chromium. This surface depletion
becomes more critical for sheet products or thinwall tubular products because of much higher
percentage of the surface-depletion zone in the
overall thickness of the component. In alloy
manufacturing, annealing is required after each
cold-rolling step for sheet product manufacturing
(or cold pilgering for reduction in thickness and
sizes for tubular product manufacturing) to
soften the metal for further cold-reduction steps
until a nal product is produced. Stainless steels
are typically annealed in the temperature range of
1010 to 1120 C (1850 to 2050 F) (Ref 28).
When annealing is performed in air, heavy
chromium oxide scales form on the metal surface. As a result, the matrix immediately underneath the oxide scales can be depleted in
chromium. Figure 3.18 shows the concentration
prole of chromium near the surface of the plate
of alloy AL-6XN (Fe-21Cr-24Ni-6.5Mo-0.2N)
after annealing in air at 1120 and 1175 C (2050
and 2150 F), respectively (Ref 29). Oxidation
during annealing at either temperature resulted
in a signicant chromium depletion near the
surface of the plate. It is quite common to perform annealing in air during manufacturing of
stainless steels. This is commonly referred to
as black annealing, as opposed to bright
pg 20
31/10/2007 12:44PM Plate # 0
pg 21
annealing, which is performed in a protective

atmosphere, such as hydrogen atmosphere. The
oxide scales on the alloy plate, sheet, or tubing
are generally removed by acid pickling. This
manufacturing process can often produce at
products (plate or sheet) as well as tubular products with surface depletion of chromium. When
bright annealing is performed, the alloy surface is
protected from oxide scale formation. As a result,
chromium depletion at and near the product
surface is minimized.
Surface depletion in chromium becomes
increasingly critical as the thickness of the component, such as sheet or tubular product,
decreases. Figure 3.19 shows an example of

surface depletion of chromium in a thin-gage
commercial heat-exchanger tube made from
Type 321 in the as-fabricated condition (Ref 30).
The manufacturing process involved was not
known. The Type 321 tube is shown to exhibit
depletion in chromium near the tube surface
when analyzed by energy-dispersive x-ray
spectroscopy (EDX). The analysis was terminated at approximately about 0.5 m from the
tube surface. It is expected that the chromium
content would decrease further if the analysis
was performed at locations closer to the surface.
After service for 6 months as a recuperator tube at
22
As hot rolled
Chromium content, wt%
20
18
Annealed 1175 C
Annealed 1120 C
16
AL-6XN
14
2
10
12
14
16
18
20
Distance, m
Fig. 3.18
Surface deletion of chromium near the surface of the plate of alloy AL-6XN (Fe-21Cr-24Ni-6.5Mo-0.2N) after annealing in
air at 1120 and 1175 C (2050 and 2150 F), respectively. Source: Ref 29
18.4
Cr concentration, wt%
18.2
18
17.8
17.6
17.4
17.2
17
16.8
0
10
Distance from tube OD surface, m
Fig. 3.19
Surface depletion of chromium observed in a thin-gage commercial heat-exchanger tube in the as-fabricated condition
made from Type 321. Source: Ref 30
31/10/2007 12:44PM Plate # 0
metal temperatures of approximately 620 to

670 C (1150 to 1240 F) for preheating air,
signicant oxide spalling and scaling was
observed on the air side of the heat-exchanger
tube. Figure 3.20 shows heavy oxide scales
formed on the side of the tube exposed to the
incoming air after 6 months of service (Ref 30).
The oxide scales were analyzed by scanning
electron microscopy/energy-dispersive x-ray
spectroscopy (SEM/EDX) analysis; the results
are shown in Fig. 3.21. The analysis showed that
the outermost oxide layer was essentially iron
oxides with very little chromium. Thus, the
stainless steel recuperator no longer exhibited
adequate resistance to oxidation because of its
failure to form and maintain a protective Cr2O3
scale. Once a protective Cr2O3 scale is no longer
present on the stainless steel surface, iron oxides
then take over. This results in scaling and accelerating oxidation. This is often referred to as
breakaway oxidation (or corrosion). Breakaway oxidation is discussed in Section 3.4.13.
Stainless steels manufactured by different
producers can exhibit different chemical compositions and different surface characteristics.
This is illustrated in Fig. 3.22 (Ref 30). Two Type
321 heat-exchanger tubes, manufactured by
two different suppliers (supplier A and B), were
tested in the eld in the same recuperator as
described previously for preheating air at
approximate metal temperature of 620 to 670 C
(1150 to 1240 F) for about 1008 h. Figure 3.22
(a) shows the formation of mushroom-type oxide
nodules on the surface of the Type 321 tube,
produced by Supplier A. This is considered the
initiation of breakaway oxidation. As shown in
Fig. 3.22(a), the mushroom-type oxide nodule
consisted of two layers of oxides, light grayish
outer oxide scale and darkish inner oxide layer.
The rest of the metal surface was still protected
by a thin, adherent oxide scale, which was not

clearly revealed in the gure at about 400
magnication. Once the formation of the light,
grayish outer oxide scale formed on the metal
surface, the inner oxidation took place with
accelerated growth. The outer oxide scale was
found to be iron-rich oxides with little chromium,
as shown in Fig. 3.23 (Ref 30). With the formation of the iron-rich oxide scale at the outer layer,
which failed to provide protection, the inner
Fe-Cr oxides were found to penetrate into the
metal, causing a relatively massive internal oxidation penetration (Fig. 3.23). In the same test,
the other Type 321 heat-exchanger tube (produced by supplier B), which was subjected to the
same test condition and duration, was found to
exhibit a thin, adherent oxide scale with no evidence of mushroom-type oxide nodules, as
shown in Fig. 3.22(b). The oxide scale was
analyzed by SEM/EDX, showing an Fe-Cr oxide
scale formed on the metal surface (Fig. 3.24).
Bulk chemical compositions of two tubes were
analyzed; results are shown in Table 3.5. The
chemical analysis results showed that the chromium in the Type 321 tube from supplier A was
essentially at the lower end of the specication
range. Any depletion in chromium near the surface could result in a chromium level that is
below the specication limit. On the other hand,
the tube from supplier B contained much higher
chromium and was found to be much more
resistant to oxidation under the same test conditions. It appears that some stainless steel manufacturers produce their products with leaner
chemistry in terms of major alloying elements,
such as chromium and nickel. For hightemperature oxidation and other modes of corrosion, stainless steel with the chromium in the
lower end of the specication range can be
potentially more prone to breakaway oxidation
25 m
Fig. 3.20
Heavy oxide scales formed on the side of Type 321 recuperator tube that was exposed to the incoming air after 6 months of
service with the metal temperatures approximately 620 to 670 C (1150 to 1240 F). This tube was from the same batch of
tubes that shows surface chromium depletion (Fig. 3.19). Source: Ref 30
pg 22
31/10/2007 12:44PM Plate # 0
pg 23
B
C
D
E
25 m
(a)
30 m
Fig. 3.21
Scanning electron micrograph (backscattered

electron image) showing the oxide scales formed
on the outside diameter of the heat-exchanger tube (from the same
batch of tubes that showed surface chromium depletion) exposed
to air for 6 months. Energy-dispersive x-ray spectroscopy (EDX)
analysis was performed to determine the chemical compositions
at different locations, marked A to E. The results (wt%) of the
EDX analysis are summarized below (minor elements not included). Source: Ref 30
A: 1.9% Cr, 97.6% Fe.
B: 44.2% Cr, 44.4% Fe, 6.8% Ni.
C: 48.8% Cr, 40.4% Fe, 4.4% Ni.
D: 28.3% Cr, 46.4% Fe, 21.0% Ni.
E: 38.7% Cr, 45.3% Fe, 10.2% Ni.
or corrosion. A brief discussion of this important

issue is presented in Section 3.4.5.
3.4.4 Surface Conditions
As discussed in Section 3.4.3 Surface
Chemistry versus Bulk Chemistry, the concentration of chromium at and near the surface of
the alloy product plays a signicant role in the
oxidation of stainless steels such as Types 304,
316, 321, 347, and so forth. This is because the
chromium concentration of these stainless steels
is at the lower end of the chromium range that is
generally required to form a continuous Cr2O3
scale when heated to elevated temperatures.
25 m
(b)
Fig. 3.22
Type 321 heat-exchanger tubes, which were

manufactured by two different alloy suppliers,
were tested in the same facility as described previously for preheating air at approximate metal temperature of 620 to 670 C
(1150 to 1240 F) for about 1008 h. (a) Supplier A. (b) Supplier B.
Note the tube from supplier A showed the initiation of accelerated
oxidation attack (a) as opposed to the tube from supplier B
showing no sign of accelerated oxidation attack (b).
Manufacturing processes can greatly inuence

the surface chemistry of an alloy product.
Stainless steels can be nished into the nal
product by bright annealing (i.e., annealing is
performed in a protective atmosphere, such as
hydrogen environment or dissociated ammonia
environment). This process generally produces
a product with minimal depletion of chromium
at or near the surface. On the other hand, when
the alloy product is nished by black annealing
(i.e., annealing is performed in air or combustion atmosphere in the furnace) and followed
by acid pickling, there is a good chance that
the alloy product may exhibit a surface depletion
of chromium. This is particularly important
for thin-gage sheet products or thin tubular products.
31/10/2007 12:44PM Plate # 0
pg 24
In most laboratory oxidation tests, the test

specimens are typically prepared by grinding and
polishing with different grits of emery papers
prior to testing. The objective of grinding and
polishing the test specimens to a certain surface
nish condition is to keep the surface condition
of all test specimens constant in order to compare
the oxidation behavior of different alloys. However, the mechanical forces of grinding and
polishing can produce a thin cold-worked layer
on the specimen surface. This cold-worked
structure at the surface layer can signicantly
enhance the diffusion of chromium from the
interior to the surface of the metal to form
chromium oxide scales when heated to elevated

temperatures, thus increasing the oxidation
resistance of the alloy. For some critical applications involving a thin-gage sheet (or foil)
product or a thin tubular product, testing should
be carried out on the specimen that retains the
surface condition of the product without prior
surface grinding or mechanical polishing.
Electropolishing, which is not commonly used to
improve the surface nish of the alloy product for
high-temperature services, may cause surface
depletion of chromium for the product. Nevertheless, some investigators may use electropolishing to prepare the surface condition of the
test specimens. Table 3.6 shows some comparison oxidation data generated in wet O2 between
the wet ground and electropolished surface
conditions for several stainless steels (Ref 31).
3.4.5 Todays Stainless Steels
Some stainless steel producers may manufacture stainless steels at the bottom of the
1
3
2
4
13 m
13 m
Fig. 3.23

electron image) showing the oxide scales formed
on the outside diameter of Type 321 tube (from supplier A)
exposed to air at approximately 620 to 670 C (1150 to 1240 F)
for 1008 h. Energy-dispersive x-ray spectroscopy (EDX) analysis
was performed to determine the chemical compositions at different locations as marked 1 to 4 in the oxides. The results (wt%) of
the EDX analysis are (minor elements not included):
1: 12% Cr, 86% Fe.
2: 52% Cr, 24% Fe, 18% Ni.
3: 32% Cr, 25% Fe, 38% Ni.
4: 37% Cr, 52% Fe, 7% Ni.
Table 3.5
Fig. 3.24

image) showing the oxide scales formed on the
outside diameter of Type 321 tube (from supplier B) exposed to air
at approximately 620 to 670 C (1150 to 1240 F) for 1008 hours.
EDX analysis was performed to determine the chemical compositions at different locations, marked as No. 1 and 2, in the oxides.
The results (wt%) of the EDX analysis are (minor elements not
included):
1: 35% Cr, 50% Fe, 6% Ni.
2: 21% Cr, 59% Fe, 4% Ni.
Chemical compositions of Type 321 tubes from Suppliers A and B

Composition, wt%
Supplier
A
B
Cr
Ni
Ti
Mn
Si
Mo
Cu
Fe
0.043
0.072
17.01
17.77
8.85
8.81
0.32
0.38
1.04
0.94
0.54
0.76
0.35
0.21
0.26
0.09
0.032
0.035
<0.005
0.018
bal
bal
31/10/2007 12:44PM Plate # 0
pg 25
specication range for key alloying elements,

such as chromium, to reduce materials cost.
Kelly (Ref 32) discusses this point in his 1992
paper, as shown in Tables 3.7 and 3.8. Accordingly, the chromium content can be insufcient
to maintain a continuous chromium oxide scale
during prolonged service, or when subjected to
thermal cycling, or overheating conditions, thus
promoting breakaway oxidation. These lean
stainless steels can be further aggravated by
the surface depletion of chromium resulting
from manufacturing processes that may involve
excessive pickling after black annealing
(annealing in air or combustion atmosphere),
during successive reductions in cold rolling in
at product manufacturing, or pilgering in tubular manufacturing. This can result in a product,
particularly a thin-gage sheet or tube, in which
the chromium concentration at the surface is too
low to form or maintain a continuous chromium
Table 3.6 Metal losses of several stainless steels
tested at 648 C (1200 F) for 168 h in wet O2
Alloy
Surface nish
Metal loss after

descaling, mg/cm2
304
Wet ground
Electropolished
Wet ground
Electropolished
Wet ground
Electropolished
Wet ground
Electropolished
3.6
5.9
1.2
6.5
1.8
6.6
2.3
6.4
321
316
347
oxide scale during service. As a result, iron oxides and isolated nonprotective Fe-Cr oxide
nodules can develop on the metal surface, thus
resulting in breakaway oxidation.
3.4.6 Special, Proprietary Austenitic
Stainless Steels
There are several commercial proprietary heatresistant alloys that belong to the austenitic
stainless steel group in terms of chromium
and nickel contents, but with the addition of
silicon for increasing the resistance of the alloy
to oxidation and other high-temperature corrosion attack. These alloys include 253MA (Fe21Cr-11Ni-1.7Si-0.17N-0.04Ce) and 85H (Fe18.5Cr-14.5Ni-3.5Si-0.2C-1.0Al). Figure 3.25
shows the cyclic oxidation resistance of 253MA
in air at 1090 C (2000 F) compared with that
of Type 309 and some higher-alloyed Fe-NiCr alloys (e.g., 800H and RA330) (Ref 33).
Figure 3.26 shows more cyclic oxidation tests
comparing 253MA with 800H and 353MA at
1093 and 1150 C (2000 and 2100 F) (Ref 34).
Figure 3.27 shows the oxidation resistance of
253MA compared with austenitic stainless
steels, such as 18-10Ti (Type 321), 20-12Si (DIN
1.4828, 20Cr-12Ni-2Si), and 25-20 (Type 310),
and 353MA and Ni-Cr alloy 601 as a function of
temperature (Ref 35). Figure 3.28 shows that
12
Source: Ref 31
Composition, wt%
ASTM A 240
Pre-1965 heats
Current production
Cr
Ni
18.0020.00
18.7
18.3
8.0010.50
9.9
9.0
Source: Ref 32
10
RA309
800H
RA310
Weight gain, mg/cm2
Table 3.7 Chromium and nickel contents for

Type 304 between pre-1965 heats and current
production heats
1090 C
(2000 F)
20 h cycles in air
8
RA330
RA333
6
RA601
4
RA 253 MA
Table 3.8 Chromium, nickel, and molybdenum

contents for Type 316 between pre-1965 heats
and current production heats
Composition, wt%
ASTM A 240
Pre-1965 heats
Current production
Source: Ref 32
Cr
Ni
Mo
16.0018.00
17.9
16.3
10.0014.00
12.4
10.2
2.003.00
2.4
2.1
Fig. 3.25
100
300
200
Exposure time, h
400
500
Cyclic oxidation resistance of 253MA in air at

1090 C (2000 F) with 20 h cycles compared
with that of Type 309 and some higher-alloyed Fe-Ni-Cr alloys
(e.g., 800H and RA330) up to 500 h testing. Source: Ref 33
31/10/2007 12:44PM Plate # 0
specimen was cathodically descaled, instead of

weight gain. The oxidation resistance of RA330
along with other alloys in the 2025Cr/3040Ni
group of Fe-Ni-Cr alloys is discussed in Section
3.4.8.
160
870 C
(1600 F)
1090 C
(2000 F)
120
1150 C
(2100 F)
100
80
60
Weight gain, mg/cm2
140
980 C
(1800 F)
40
20
0
RA 253 MA
RA 353 MA RA800H/AT
Fig. 3.26
Cyclic oxidation tests performed on 253MA,

353MA, and 800H at 871, 982, 1093, and 1150 C
(1600, 1800, 2000, and 2100 F) for 1640 h with specimens
cycling to room temperature every 160 h. Source: Ref 34
20 - 12Si
Weight increase, g/m2 h
18 - 10Ti
253 MA
2
153 MA
25 - 20
Alloy 601
353 MA
0
1000
1100
1200
Temperature, C
Isothermal oxidation rates (g/m2 h) obtained by
testing for 45 h at various temperatures for 253MA
and 353MA compared with some austenitic stainless steels, such
as 18-10Ti (Type 321), 20-12Si (DIN 1,4828, 20Cr-12Ni-2Si), 2520 (Type 310), and nickel alloy 601. 1.0 g/m2 = 0.1 mg/cm2.
Source: Ref 35
Fig. 3.27
RA85H was better than alloy 800H, but not as

good as some higher-alloyed Fe-Ni-Cr alloys
(e.g., RA330, 353MA) at 1150 C (2100 F) in
air with 164 h cycles (Ref 33). In a long-term
cyclic oxidation test in air at 1090 C (2000 F)
for times up to 1 year with specimens cycling to
room temperature after every 20 h of exposure,
both 253MA and 85H were not as good as
RA330, as shown in Table 3.9 (Ref 36). The
oxidation resistance in this test was measured
by metal loss, which was determined after the
3.4.7 Fe-Cr-Al Alloys

Aluminum is a very effective alloying element
in improving the resistance of the alloy to oxidation and other high-temperature corrosion
attack. The alloy normally requires about 4% Al
to form a continuous Al2O3 scale. The Al2O3
scale provides exceptional protection against
oxidation attack. Figure 3.29 shows parabolic
rate constants of some important oxides, such as
Al2O3, Cr2O3, SiO2, FeO, NiO, and CoO (Ref 2).
Al2O3 exhibits the lowest parabolic rate constants. When the metal is heated to 1200 C
(2200 F) and higher, the Cr2O3 scale, which
exhibits high growth rates as well as forming
volatile CrO3, becomes essentially nonprotective. Under these very high temperature
conditions, Al2O3 scale provides excellent protection against oxidation. Because of extremely
slow growth rates at low and intermediate temperatures, Al2O3 scale provides less protection at
these temperatures. As a result, high-temperature
alloys that are designed to form Al2O3 scale for
very high temperatures also contain adequate
chromium to form Cr2O3 scale for intermediate
temperatures.
Some commercial electrical resistance heating
elements are made of Fe-Cr-Al alloys, such as
Kanthal alloys, which rely on the formation of
the Al2O3 scale for applications up to 1400 C
(2550 F) (Ref 37). For example, some of the
Kanthal alloys that are available in wire, strip,
and ribbon product forms are Kanthal A-1
(Fe-22Cr-5.8Al), AF (Fe-22Cr-5.3Al), and D
(Fe-22Cr-4.8Al). Since these wrought alloy
products are essentially ferritic alloys, they
exhibit low creep-rupture strengths when the
temperature exceeds 650 C (1200 F) and cannot be used for high-temperature structural
components. Thus, the electrical resistance
heating elements made of these alloys have to be
properly supported to avoid creep deformation,
such as sagging. These Kanthal wires can be used
in arc or ame spraying to produce an oxidationresistant coating or in weld overlay cladding by
using gas metal arc welding (GMAW) process.
A powder metallurgy (P/M) process was used
to produce a new alloy product, Kanthal APM.
This P/M-produced Kanthal APM alloy was
pg 26
31/10/2007 12:44PM Plate # 0
pg 27
120
1150 C (2100 F)
164 h cycles in air
164
100
80
80
74
Weight gain, mg/cm2
72
60
40
36
23
20
20
RA85H
RA310
RA330
RA333
RA 353MA
RA600
800H
Total exposure time, 1640 h
Fig. 3.28
Cyclic oxidation tests at 1150 C (2100 F) for 1640 h in air, with specimens cycling to room temperature every 164 h for
Type 310, RA85H, 800H, 353MA, RA330, and nickel-base alloys 600 and RA333. Source: Ref 33
Table 3.9 Isothermal and cyclic oxidation tests

in air at 1090 C (2000 F)
Metal loss, m (mils)
Alloy
Exposure time
Cyclic(a)
Isothermal
85H
3000 h
1 year
3000 h
1 year
3000 h
1 year
3000 h
883 (34.8)
2665 (105)
1033 (40.7)
3730 (147)
683 (26.9)
2005 (79)
41 (1.63)
22 (0.88)
44 (1.74)
34 (1.34)
253MA
RA330
310
(a) Specimens were cycled to room temperature after every 20 h of exposure.

Source: Ref 36
reported to exhibit higher creep-rupture strengths

(Ref 38).
Other commercial Fe-Cr-Al alloys include
ALFA-I (Fe-13Cr-3Al), ALFA-II (Fe-13Cr4Al), and ALFA-IV (Fe-20Cr-5Al-Ce) developed by Allegheny Ludlum (Ref 39), and
Fecralloy (Fe-16Cr-4Al-0.3Y) developed by

U.K. Atomic Energy Authority (Ref 40). Allegheny Ludlum (Ref 39) conducted cyclic oxidation tests in air by repeatedly resistance heating
a thin ALFA IV foil (0.050 mm, or 0.02 in.,
thick) to various temperatures for 2 min until
it failed. The cycles to failure of the foil are
summarized in Table 3.10. No denition of the
cycle to failure was given in the report. It is
believed that the failure of the foil was dened as
when oxidation penetrated through the thin foil.
ALFA IV uses a rare earth element, cerium, for
enhancement of the Al2O3 scale. Very few studies have been conducted on the effectiveness of
cerium on the adhesion of the Al2O3 scale. More
studies have been conducted on the effects of
yttrium, zirconium, and other reactive elements.
In the Fecralloy alloy, the rare earth element
yttrium is added to increase the adhesion of the
31/10/2007 12:45PM Plate # 0
pg 28
Al2O3 scale. Yttrium and reactive elements such

as zirconium have been found to increase the
adhesion of the Al2O3 scale formed on Fe-Cr-Al
alloys, thus improving the oxidation resistance
of the alloy (Ref 4144). An Fe-Cr-Al-Y alloy
can be strengthened by the oxide-dispersion
strengthening (ODS) mechanism to signicantly
increase its elevated-temperature strengths by the
mechanical alloying process (Ref 45). One such
commercial ODS alloy is MA956 (Fe-20Cr4.5Al-0.5Y2O3). Oxidation resistance of some
ODS alloys is discussed in Section 3.4.10 on
ODS alloys as well as in Section 3.4.12 on oxidation in high-velocity combustion gas streams.
Temperature, K
1350
1400
1300
1250
FeO
Log k, g2 cm4s1
Table 3.10 Cycles to failure of a thin foil

(0.050 mm, or 0.002 in., thick) of ALFA IV
repeatedly resistance heated to the indicated
temperatures for 2 min in still air from room
temperature
NiO
10
Cr2O3
12
Fig. 3.29
As the nickel content in the Fe-Ni-Cr system

increases from austenitic stainless steels to a
group of iron-base alloys with 2025Cr and
3040Ni, the alloys become more stable in terms
of metallurgical structure and more resistant to
creep deformation (i.e., higher creep-rupture
strengths). In general, this group of alloys also
exhibits better oxidation resistance. Some of
the wrought alloys in this group are 800H/800HT
(Fe-21Cr-32Ni-Al-Ti), RA330 (Fe-19Cr-35Ni1.2Si), HR120 (Fe-25Cr-37Ni-0.7Nb-N), AC66
(Fe-27Cr-32Ni-0.8Nb-Ce), 353MA (Fe-25Cr35Ni-1.5Si-Ce), and 803 (Fe-26Cr-35Ni-Al-Ti).
The oxidation resistance of 353MA compared
with RA330, 800H, several stainless steels, and
nickel-base alloys is shown in Fig. 3.26 to 3.28.
Figure 3.30 shows the comparison of 353MA
CoO
14
3.4.8 Fe-Ni-Cr Alloys (2025Cr/3040Ni

Alloys)
SiO2
AI2 O
3
Temperature, C (F)
7.0
7.8
7.6
7.4
1 104, K1
T
7.2
8.0
8.2
Cycles to failure
1090 (2000)
1150 (2100)
1200 (2200)
1260 (2300)
Parabolic rate constants of several oxides. Source:

Ref 2
2600
1000
460
200
Source: Ref 39
4
253MA
2
0
353MA
HK 4M
2
4
6
8
Inco 803
10
HP45-Nb
500
1500
2500
3500
Time, h
Fig. 3.30
Isothermal oxidation tests in air at 1000 C (1830 F) for 353MA, alloy 803, HK4M (Fe-25Cr-25Ni-0.3Al-0.4Ti), and
HP45Nb (Fe-24Cr-37Ni-1.4Si-1.2Nb). Source: Ref 46
31/10/2007 12:45PM Plate # 0
pg 29
and 803 in comparison with some casting alloys

with similar chromium and nickel contents when
tested in air at 1000 C (1830 F) (Ref 46).
Additional oxidation data on 353MA, 803,
HR120, 800H, and RA330 are summarized in
Fig. 3.31 to 3.35 (Ref 4749). Particular
emphasis should be placed on long-term
oxidation tests (Fig. 3.33 and 3.35), where some
alloys are shown to suffer breakaway oxidation
after some incubation time of relatively little or
mild weight losses. More discussion on breakaway oxidation is presented in Section 3.4.13.
3.4.9 Ni-Cr/Co-Cr Superalloys
In many Ni-Cr alloys, many alloying
elements, such as those for solid-solution
strengthening (e.g., Mo, W) and precipitation
strengthening (e.g., Al, Ti, Nb), are added into
the alloys to provide strengthening of the alloy
at elevated temperatures. Many of these alloys
are commonly referred to as superalloys.
The superalloys also include oxide dispersion
strengthened (ODS) alloys, which are briey
discussed in Section 3.4.10.
Similar to Fe-Cr-Al alloys, aluminum is also
used as an alloying element in Ni-Cr alloys to
improve the oxidation resistance. Although it
generally requires a minimum of 4% Al in order
for a Ni-Cr alloy to form a protective Al2O3

scale, the addition of less than 4% Al can also
signicantly improve the oxidation resistance of
the alloy. Alloy 601 with only about 1.3% Al
shows excellent oxidation resistance, as shown in
Fig. 3.32 to 3.34 and Fig. 3.36 to 3.38. In both
Fig. 3.37 and 3.38, alloy 601GC was included in
the testing to compare with alloy 601. Alloy
601GC is essentially alloy 601 with small additions of titanium, zirconium, and nitrogen for
grain-size control involving nitrides and carbonitrides. The development of alloy 601GC was
intended to prevent excessive grain coarsening
when exposed to 1100 C (2010 F) and higher
with the formation of nitrides and carbonitrides
of zirconium, titanium, and Zr + Ti (Ref 51). In
long-term oxidation testing (for up to close to
500 days) in air at 1000 C (1830 F), alloy
601GC was found to be better than alloy 601, as
shown in Fig. 3.38. The improvement of oxidation resistance in alloy 601GC over alloy 601
may be the result of ner grain sizes as well as the
presence of zirconium, a reactive element that
has been found to be benecial in improving
oxidation resistance for Fe-Cr-Al alloys by
several investigators cited previously.
Even though the presence of about 1.4% Al in
alloy 601 is benecial in improving oxidation
resistance, the oxide scales formed on alloy 601
50
N08811
R30566
40
Weight gain, mg/cm2
N08330
30
S35315
20
N06333
HR-120
10
500
1000
1500
2000
2500
3000
Time, h
Fig. 3.31
Cyclic oxidation tests for 353MA (S35315), HR120, RA330 (N08330), 800HT (N08811), RA333 (N06333), and 556
(R30556) in air at 1090 C (2000 F) with specimens cycling to room temperature once a week. Source: Ref 47
31/10/2007 12:45PM Plate # 0
pg 30
350
300
N08810
Weight gain, mg/cm2
250
HR-120
200
150
100
N06601
50
500
1000
1500
2000
N06045
2500
S35315
3000
Time, h
Fig. 3.32
Cyclic oxidation tests for 353MA (S35315), HR120, 800H (N08810), 601 (N06601), and 45TM (N06045) in air at 1150 C
(2100 F) with specimens cycling to room temperature once a week. Source: Ref 47
100
100
Incoloy alloy DS
UNS N06601
UNS N06601
100
Incoloy alloy 803
200
300
UNS N06600
400
500
UNS N08330
600
UNS N08810
700
Mass change, mg/cm2
Mass change, mg/cm2
Incoloy alloy DS
100
UNS N06600
200
UNS N08330
Incoloy alloy 803
300
400
500
UNS N08810
800
0
5000
10,000
15,000
20,000
600
0
Exposure time, h
Fig. 3.33
Oxidation tests for 803, RA330 (N08330), 800H

(N08810), alloy DS, and 600 (N06600) and 601
(N06601) in air + 5% H2O at 1000 C (1830 F). Source: Ref 48
are primarily chromium-rich oxides. Chromiumrich oxide scales are prone to scaling, cracking,
and spalling, particularly under cyclic conditions
at very high temperatures, such as 1100 C
(2010 F) and higher. This is illustrated in
Fig. 3.39 for alloy 601 tested in air for 1056 h at
850 C (1560 F), 1000 C (1830 F), 1100 C
(2010 F), and 1200 C (2190 F) (Ref. 52). For
testing at 850, 1000, and 1100 C, specimens
were cycled to room temperature from the test
1000
2000
3000
4000
5000
6000
Exposure time, h
Fig. 3.34
Oxidation tests for 803, RA330 (N08330), 800H

(N08810), and alloy DS in compared with Ni-Cr
alloys 600 (N06600) and 601 (N06601) in air + 5% H2O at
1100 C (2010 F). Source: Ref 48
temperature every 16 h, with 2 h of heating and

6 h of cooling. For 1200 C testing, specimens
were removed from the hot zone every 16 h. In
addition to scaling and spalling of external oxide
scales when exposed to 1100 and 1200 C (2010
and 2190 F), signicant internal oxide penetration was also observed. However, with slightly
higher aluminum content (i.e., about 2%), the
31/10/2007 12:45PM Plate # 0
pg 31
HR120
Weight change, mg/cm
50.0
100.0
RA85H
150.0
800HT
200.0
0
90
180
270
360
450
540
630
720
Time, days
Fig. 3.35
Oxidation tests for HR120, 800H, and RA85H in still air at 980 C (1800 F) for times up to 720 days with specimens being
cooled to room temperature every 30 days for weight measurement. Source: Ref 49
alloy can signicantly reduce its scaling and

spalling of external oxide scales as well as
internal oxide penetration. This is illustrated in
Fig. 3.40, which shows the oxide scales formed
on alloy 602CA (Ni-25Cr-9Fe-2.2Al-0.2C0.06Zr-0.08Y) after the same test conditions as
alloy 601 as shown in Fig. 3.39. The improved
oxidation resistance of alloy 602CA may also be
attributed to the presence of yttrium and zirconium. The benecial effects of yttrium and zirconium along with other reactive elements in
aluminum-containing alloys are discussed later
in the chapter. The cyclic oxidation resistance
of alloy 602CA compared with alloy 800H
and alloy 601 as well as several nickel- and
cobalt-base superalloys at temperatures from 750
to 1200 C (1380 to 2190 F) are summarized in
Table 3.11 (Ref 53). With further increase in
aluminum to 2.8% in the same alloy composition
series, alloy 603GT (Ni-25Cr-9Fe-2.8Al-0.22C0.1Zr-0.1Y) exhibits more compact, adherent
oxide scales (Fig. 3.41) under the same test
conditions as Fig. 3.40 (Ref 52).
Ni-Cr alloys containing about 4% Al or higher
form a very protective Al2O3 scale when heated
to very high temperatures. This is illustrated in
Fig. 3.42, where alloy 214 (Ni-16Cr-4.5Al-Y)

was compared with alloy 601 and alloy 800H in
cyclic oxidation tests performed in still air at
1150 C (2100 F) with specimens cycling to
room temperature once a day except weekends
(Ref 54). Alloy 214 showed essentially no
weight loss after 42 days of testing, while alloy
601 suffered a linear weight loss. Figure 3.43
shows the cross section of the specimen between
alloy 214 and alloy 601 oxidation tested in air at
1090 C (2000 F) for 1008 h. The Al2O3 oxide
scale formed on alloy 214 remains very protective even when the temperature is increased to
close to its incipient melting point, as shown in
Fig. 3.44 and 3.45 (Ref 55). Figure 3.44 shows
the alloy 214 coupon after exposure to owing
air at 1320 C (2400 F) for 200 h with the
specimen being cycled to room temperature
every 24 h. The incipient melting temperature
for alloy 214 is believed to be slightly higher
than 1345 C (2450 F) (Ref 56). The specimen
showed no evidence of oxidation attack
(Fig. 3.44). SEM/EDX examination of the cross
section of the tested specimen showed a thin
aluminum-rich oxide scale formed on the metal
surface (Fig. 3.45).
31/10/2007 12:45PM Plate # 0
pg 32
As shown in Fig. 3.29, in addition to Cr2O3

and Al2O3, SiO2 also exhibits very low parabolic
rate constants and can provide an effective barrier to oxidation attack at very high temperatures.
Fig. 3.36
Cyclic oxidation resistance of alloy 601 (Ni-23Cr14Fe-1.4Al) compared with alloy 600 (Ni-16Cr8Fe) and alloy 800 (Fe-20Cr-32Ni-0.4Al-0.4Ti) in air at 1090 C
(2000 F) with specimens being in hot zone for 15 min and out of
hot zone for 5 min. Source: Ref 50
Birks and Pettit (Ref 57) suggested that both

Al2O3 and SiO2 are capable of providing adequate oxidation resistance above 1200 C
(2200 F). There are a number of commercial
high-temperature alloys that contain relatively
high levels of silicon for improving hightemperature corrosion resistance. However, the
amount of silicon addition to an alloy that can be
manufactured into a wrought product or casting
product is quite limited. Silicon can make an
alloy difcult to cast and also very difcult to
weld. For some commercial, wrought alloys,
silicon levels are up to 3.5%. The formation of a
continuous, external SiO2 scale was not observed
for these commercial Fe-Ni-Cr alloys or Ni-Cr
alloys in a way that a continuous, external Al2O3
scale forms on Fe-Cr-Al alloys or Ni-Cr-Al
alloys. A Ni-Cr-Co-Si alloy, HR160, was
developed in late 1980s for applications in severe
suldizing environments (Ref 58). The oxidation
resistance of HR160 alloy (Ni-28Cr-30Co2.75Si-0.5Ti-0.5Nb) was found to be quite
comparable to alloy 601 and signicantly better
than alloy 800HT when tested for time up to
about 1 year in air at 1090 C (2000 F), as
shown in Fig. 3.46 (Ref 59). However, in contrast
to Ni-Cr-Al alloy 214 that forms an external
Al2O3 with essentially no internal oxide or void
formation when exposed to very high temperatures, such as 1200 C (2200 F), HR160 forms
external Cr2O3/SiO2 oxides scales with internal
oxide/void formation (Ref 49). Figure 3.47
shows oxidation data for alloys 214 and HR160
along with several other nickel- and iron-base
alloys. The oxidation tests were conducted at
100
Mass change, mg/cm2
0
601
100
601GC
200
330
300
RA85H
400
50
100
150
200
Exposure time, days
Fig. 3.37
Oxidation tests in air + 5% H2O at 1100 C (2010 F) for up to 200 days for alloys 601 (Ni-23Cr-14Fe-1.4Al), 601GC
(Ni-23Cr-14Fe-1.3Al-0.2Ti-0.2Zr-0.05N), RA85H (Fe-18Cr-14Ni-3.5Si-1Al), and RA330 (Fe-19Cr-35Ni-1.2Si). Specimens
were cooled to room temperature and weighed at the indicated data points. Source: Ref 51
31/10/2007 12:45PM Plate # 0
pg 33
10
Mass change, mg/cm2
10
601GC
20
30
330
601
40
50
0
100
300
200
400
500
600
Exposure time, days
Fig. 3.38
Oxidation tests in air + 5% H2O at 1000 C (1830 F) for times up to close to 500 days for alloys 601 (Ni-23Cr-14Fe-1.4Al),
601GC (Ni-23Cr-14Fe-1.3Al-0.2Ti-0.2Zr-0.05N), and RA330 (Fe-19Cr-35Ni-1.2Si). Specimens were cooled to room
temperature and weighed at the indicated data points. Source: Ref 51
(a)
(b)
Fig. 3.39
20 m
(c)
(d)
Alloy 601 tested in air for 1056 h at (a) 850 C (1560 F), (b) 1000 C (1830 F), (c) 1100 C (2010 F), and (d) 1200 C
(2190 F). For testing at 850, 1000, and 1100 C, specimens were cycled to room temperature from the test temperature
every 16 h, with 2 h of heating and 6 h of cooling. For 1200 C testing, specimens were removed from the hot zone every 16 h.
Magnification bar represents 20 m for all micrographs. Source: Ref 52. Courtesy of ThyssenKrupp VDM
31/10/2007 12:45PM Plate # 0
(a)
20 m
(c)
(d)
(b)
Fig. 3.40
Alloy 602CA (Ni-25Cr-9Fe-2.2Al-0.2C-0.06Zr-0.08Y) tested in air for 1056 h at (a) 850 C (1560 F), (b) 1000 C (1830 F),
(c) 1100 C (2010 F), and (d) 1200 C (2190 F). For testing at 850, 1000, and 1100 C, specimens were cycled to room
temperature from the test temperature every 16 h, with 2 h of heating and 6 h of cooling. For 1200 C testing, specimens were removed
from the hot zone every 16 h. Magnification bar represents 20 m for all micrographs. Source: Ref. 52. Courtesy of ThyssenKrupp VDM
Table 3.11 Weight change data (mg/m2h)

from cyclic oxidation tests in air for 1200 h at
indicated temperatures
Test temperature
Alloy
602CA
X
800H
625
601
617
188
750 C
(1380 F)
850 C
(1560 F)
1000 C
(1830 F)
1100 C
(2010 F)
1200 C
(2190 F)
+0.4
+1
+7
+1
+1
+4
+1
+3
+8
+8
+6
+10
+12
+4
+12
+5
24
100
+7
+19
+7
+7
5
162
1410
24
19
302
310
820
Note: Specimens were held at the test temperature for 16 h followed by cooling to
room temperature. For testing at 750 to 1100 C, specimens were cycled by
furnace cooling and furnace heat up (about 1.5 hours heat up). Cycling for
1200 C testing involved air cooling and inserting the specimen directly to the
furnace hot zone. Source: Ref 53
1200 C (2200 F) in air for almost 1 year.

HR160 showed extensive internal attack with
relatively little metal loss due to external oxidation, compared with the alumina-former alloy
214. The internal attack observed in alloy HR160

was caused by formation of internal oxides and
voids. Alloy 214, on the other hand, showed
essentially no metal loss due to the formation of a
compact external Al2O3 scale.
Also shown in Fig. 3.47 is a high-silicon ironbase alloy, RA85H (Fe-19Cr-15Ni-3.5Si-1Al),
suffering much more extensive internal attack
under the same test conditions. The oxidation
resistance data for alloy RA85H is also shown in
Table 3.9 and Fig. 3.28 and 3.37. Figure 3.48
shows the oxide morphology of a high-silicon
Ni-Cr-Fe alloy, 45TM (Ni-27Cr-23Fe-2.7Si),
after oxidation testing in air for 1154 h at
850 to 1200 C (1560 to 2190 F) (Ref 52).
Extensive internal void formation was observed
in the specimen tested at 1100 C (2010 F).
The oxidation data in comparing 45TM with
other high-temperature alloys is shown in
Fig. 3.32.
pg 34
31/10/2007 12:45PM Plate # 0
(a)
20 m
(b)
(c)
(d)
Fig. 3.41
Alloy 603GT (Ni-25Cr-9Fe-2.8Al-0.22C-0.13Zr-0.01Y) tested in air for 1056 h at (a) 850 C (1560 F), (b) 1000 C (1830 F),
(c) 1100 C (2010 F), and (d) 1200 C (2190 F). For testing at 850, 1000, and 1100 C, specimens were cycled to room
temperature from the test temperature every 16 h, with 2 h of heating and 6 h of cooling. For 1200 C testing, specimens were removed
from the hot zone every 16 h. Magnification bar represents 20 m for all micrographs. Source: Ref. 52. Courtesy of ThyssenKrupp VDM
In addition to maximizing the oxidation

resistance by adjusting the level of chromium
and/or aluminum, or silicon, a majority of
high-temperature alloys have been developed to
attain elevated-temperature strengths by alloying
with many other elements. A large number of
superalloys have been developed in response to
the demands of gas turbine engines for critical
operating conditions involving high stresses
and high temperatures. In response to the
demands for high stresses at intermediate temperatures, one group of wrought superalloys is
strengthened by precipitation strengthening with
Ni3X precipitates (X represents aluminum, titanium, niobium, etc.) along with solid-solution
strengthening using molybdenum or tungsten.
These alloys include 718, R-41, X-750, Waspalloy, and Nimonic 80A. Some of the applications for these alloys in gas turbines include
compressors, diffusers, turbine disks and cases,
heat shields, exhaust system, thrust reversers,
and turbine shroud rings. Most alloys in this
group are used in the heat treated conditions to
take advantage of the precipitation strengthening. Most heat treating procedures are performed
in the intermediate temperature range. As a
result, the applications for these alloys are
in the intermediate temperature range to prevent
overaging of the strengthened precipitates.
Oxidation of these alloys at intermediate temperatures generally does not present a signicant
issue in terms of the performance.
Another important group of wrought
superalloys is generally classied as solidsolution-strengthened alloys. Solid-solutionstrengthening alloying elements typically are
molybdenum and tungsten. These alloys are also
strengthened with carbides. This group of alloys
is typically used in stationary components with
lower mechanical stresses than rotating parts
such as disks and blades. However, operating
temperatures for these alloys are generally
higher. Typical applications in gas turbines are
combustors, transition ducts, and afterburners.
The alloys require good combined properties
of creep and fatigue strengths, fabricability,
pg 35
31/10/2007 12:45PM Plate # 0
pg 36
weldability, thermal stability, and oxidation

resistance. Some of these alloys are being used
increasingly in nongas turbine industries.
+20
214
0
20
40
60
601
80
100
120
140
800H
160
180
200
10
20
30
40
50
Exposure time, days
Fig. 3.42
Cyclic oxidation resistance of alloy 214 compared

to alloys 601 and 800H in still air at 1150 C
(2100 F) cycled once a day every day except weekends.
Source: Ref 54
sample
surface
Fig. 3.43
For cast nickel-base superalloys, substantial

amounts of aluminum and titanium are used to
produce a large volume fraction of gamma prime
() precipitates to further increase strengthening
of the alloy. They also contain large amounts of
refractory elements, such as molybdenum and
tungsten for solid-solution strengthening, along
with boron, zirconium, carbon, and hafnium for
grain-boundary strengthening. Since these alloys
do not require hot and cold working during
manufacturing, they can be designed to contain
maximum amounts of these alloying elements to
attain the maximum strength requirements.
These cast nickel-base superalloys include
IN713C, IN713LC, IN738X, IN100, B-1900,
Rene 80, IN792, MAR-M246, and MAR-M247.
Lacking precipitation-strengthening phases in
cobalt-base alloys, cast cobalt-base superalloys
derive their strengths from solid-solution
strengthening as well as carbide strengthening.
These alloys typically contain high levels of
carbon. Some of these alloys are X-40 (or alloy
31), MAR-M509, WI-52, and MAR-M302.
They are widely used for high-pressure vanes.
Another group of superalloys are referred to as
the ODS alloys. These alloys are produced by the
mechanical alloying process and are strengthened by oxide dispersoids (Ref 45). MA956 and
MA754 were originally developed for gas turbine combustors and stator vanes, respectively
(Ref 60), and MA6000 was developed for
rotor blades (Ref 61). ODS alloys are much
more difcult to fabricate than conventional
wrought alloys. Joining can present a particularly
signicant challenge to this group of alloys. ODS
sample
surface
Cross sections of the specimens for alloy 601 (a) and alloy 214 (b) after oxidation testing in flowing air at 1090 C (2000 F) for
1008 h (samples were cycled to room temperature once every week). Samples were cathodically descaled to remove oxide
scale prior to metallographic mounting. Top edge of the micrograph represents the original specimen surface. Source: Ref 55
31/10/2007 12:45PM Plate # 0
pg 37
HR160
0
100
601
200
300
400
500
600
700
800
800HT
900
100
200
Time, days
300
400
Fig. 3.46
Fig. 3.44
Alloy 214 specimen tested in air at 1320 C

(2400 F), which was about 25 C below the
incipient melting point of the alloy, for 200 h with the specimen
being cycled to room temperature every 24 h. Source: Ref 55
Fig. 3.45
Scanning electron micrograph showing the adherent aluminum-rich oxide scale formed on alloy 214
after exposure in flowing air at 1320 C (2400 F) for 200 h with
the specimen being cycled to room temperature every 24 h. EDX
analysis was performed at three different locations, marked 1, 2,
and 3, in the aluminum oxide scale. The results of the analysis
(wt%) are shown in: Area. 1: 69.9% Al, 17.0% Cr, 10.2% Ni, 1.1%
Fe, and 1.8% Zr. Area 2: 97.8% Al, 1.3% Cr, 0.7% Ni, and 0.2%
Zr. Area 3: 98.3% Al, 0.2% Cr, and 1.5% Ni. Source: Ref 55
alloys are briey described in Section 3.4.10.

More information about superalloys for gas
turbines is available in Ref 62 to 66.
Long-term oxidation data in air at an intermediate temperature were generated by Barrett
(Ref 67) for 33 alloys, from ferritic stainless
steels to superalloys. Tests were performed at
815 C (1500 F) for 10,000 h with 1000 h
cycles (a total of 10 cycles). The results are
shown in Fig. 3.49. Alloys, which contain no or
Oxidation resistance of HR160 alloy (Ni-28Cr30Co-2.75Si-0.5Ti-0.5Nb) compared with alloys

601 and 800HT when tested for up to about 1 year in air at
1090 C (2000 F). Source: Ref 59
low chromium content, such as nickel-base alloy

B (Ni-28Mo) and alloy N (Ni-7Cr-16.5Mo) and
Type 409 (Fe-11Cr), can suffer severe oxidation.
It was surprising to nd that Type 321 exhibited
poor oxidation resistance compared with Types
304, 316, and 347. The author offered no
explanation in the paper. As is discussed in other
sections, austenitic stainless steels, such as Types
304, 347, and 321, containing a borderline level
of chromium, are very sensitive to the chromium
content for the bulk chemistry as well as the
surface chemistry in the alloy. When the surface
depletion of chromium occurs for a stainless steel
product whose bulk chromium content is at
the low end of the specication, breakaway
oxidation is likely to take place, thus resulting in
severe oxidation attack.
The majority of the oxidation data reported so
far has been presented in terms of weight changes
(mg/cm2) as a function of times or temperatures.
However, it is impossible to use the weight
change data (either weight gain or weight loss) to
estimate the life of the component due to oxidation attack. The oxidation data that is of engineering importance is the depth of oxidation
attack, which includes the depth of metal loss and
the depth of internal oxidation attack. The total
depth of the oxidation attack (or the depth of the
metal affected) is responsible for the reduction
of the load-bearing capability for the component.
A large oxidation database in terms of the depth
of oxidation attack for commercial alloys from
stainless steels to high-alloy Fe-Ni-Cr alloys
(2025Cr/3040Ni alloys), and nickel- and
31/10/2007 12:45PM Plate # 0
Average metal affected, mm
8
7
>6.35
>6.35
>6.35
>6.35
RA85H
617
HR120
800HT
Average internal attack

Metal loss
6
5
4
3
2
1
0
214
HR160
230
601
Fig. 3.47 Oxidation data in terms of metal loss, resulting from external oxide scales, and internal attack, resulting from internal oxide
and/or void formation, for alumina-former alloy 214 and chromia/silica-former alloy HR160 along with several other nickeland iron-base alloys, generated at 1200 C (2200 F) in air for 360 days. Source: Ref 49
(a)
20 m
(b)
(c)
(d)
Fig. 3.48
A high-silicon Ni-Cr-Fe alloy, 45TM (Ni-27Cr-23Fe-2.7Si), after oxidation testing in air for 1056 h at (a) 850 C (1560 F),
(b) 1000 C (1830 F), (c) 1100 C (2010 F), and (d) 1200 C (2190 F). For testing at 850, 1000, and 1100 C, specimens
were cycled to room temperature from the test temperature every 16 h, with 2 h of heating and 6 h of cooling. For 1200 C testing,
specimens were removed from the hot zone every 16 h. Magnification bar represents 20 m for all micrographs. Source: Ref. 52.
Courtesy of ThyssenKrupp VDM
cobalt-base superalloys, is summarized in

Table 3.12 (Ref 68). Tests were conducted in
owing air (30 cm/min) at 980, 1095, 1150, and
1200 C (1800, 2000, 2100, and 2200 F) for
1008 h. The specimens were cooled to room

temperature for visual examination once a week
(168 h). Specimens from sheet products were
surface ground to maintain the uniform surface
pg 38
31/10/2007 12:45PM Plate # 0
Fig. 3.49
Long-term oxidation tests (10,000 h) in air at 815 C (1500 F) with 1000 h cycles (a total of 10 cycles to room temperature for
the entire test) for iron-, nickel-, and cobalt-base alloys. Also included is the upper limit of the metal loss for isothermal tests
(i.e., 10,000 h without cycling to room temperature) for same alloys. Source: Ref 67
condition for all test specimens. Some of the

observations of the data are summarized:
Types 304 and 316 specimens were consumed completely at 1095, 1150, and
1200 C (2000, 2100, and 2200 F).
Type 304 was found to be much more resistant than Type 316 at 980 C (1800 F).
Type 446 was not as good as Type 310 at all
test temperatures.
Type 446 specimens were consumed at 1150
and 1205 C (2100 and 2200 F).
Type 310 appeared to be slightly better than
RA330 and 800H.
For applications at high temperatures, many
superalloys contain numerous alloying elements
for increasing the elevated-temperature strength
of the alloy. Molybdenum and tungsten are
common alloying elements for providing solidsolution strengthening for increasing the creeprupture strength of the alloy. Two iron-base
superalloys, Multimet alloy (Fe-20Ni-20Co21Cr-3Mo-2.5W-1.0Nb+Ta) and alloy 556
(Fe-20Ni-18Co-22Cr-3Mo-2.5W-0.6Ta-0.02La0.02Zr), are good examples. However, the oxides

of both molybdenum and tungsten (MO3 and
WO3) exhibit high vapor pressures at very high
temperatures, as shown in Fig. 3.3. Multimet
alloy suffered rapid oxidation attack at 1150 and
1200 C (2100 and 2200 F), with specimens
completely consumed at both temperatures.
However, formation of the volatile oxides of
MO3 and WO3 can be minimized by modication of some key alloying elements in Multimet
alloy. The development of alloy 556 was aimed at
improving the oxidation resistance of Multimet
alloy without losing the elevated-temperature
strength by making some modication of alloying elements in Multimet alloy. The modication
involved a slight increase in chromium, a
decrease in cobalt, replacement of niobium with
tantalum, and addition of a rare-earth element,
lanthanum, and a reactive element, zirconium,
but the same amounts of molybdenum and
tungsten were kept. The result was a much
more oxidation-resistant alloy, alloy 556, at 1095
pg 39
31/10/2007 12:45PM Plate # 0
pg 40
and 1150 C (2000 and 2100 F), although the

alloy still suffered rapid oxidation at 1200 C
(2200 F).
As shown in the Table 3.12 nickel- and cobaltbase alloys containing molybdenum and/or
tungsten were also found to suffer rapid oxidation at very high temperatures (i.e., specimens
were consumed during the tests). Specimens of
nickel-base alloys that were consumed at
1200 C (2200 F) were alloy S (14% Mo), alloy
X (9% Mo, 0.6% W), and alloy 625 (9% Mo,
3.5% Nb). Some of those nickel-base alloys
containing molybdenum and/or tungsten that
were not consumed at 1200 C (2200 F) were
alloy 230 (14% W), alloy 617 (9% Mo), and
RA333 (3% Mo, 3% W). From these two different sets of oxidation behavior at very high
temperatures, one can design nickel-base alloys
(relying on chromium oxide scales) containing molybdenum and tungsten for elevatedtemperature strengthening to resist oxidation
resistance at very high temperatures by adjusting
other alloying elements.
For nickel-base alloys containing high levels

of molybdenum and/or tungsten, it is believed
that increasing chromium is probably the most
important factor in suppressing rapid oxidation
involving molybdenum and/or tungsten. Some
nickel-base precipitation-strengthened alloys
containing high titanium as well as molybdenum
that were consumed at 1200 C (2200 F) were
Waspaloy (4.3% Mo, 3.0 Ti), Ren 41 (10% Mo,
3.0 Ti), and alloy 263 (6% Mo, 2.2Ti). Titanium
was found to be very active in oxide-scale
formation. Figure 3.50 illustrates the oxide
scale formed on alloy 263 after exposure to air
for 1 h at 1200 C (2200 F), showing mainly
titanium-rich oxides and Cr-Ti oxides. The
formation of titanium-rich oxides apparently
disrupts the Cr2O3 scale. Nagai et al. (Ref 69)
found that titanium was detrimental to the oxidation resistance of Ni-20Cr alloy. In the Fe-CrAl system, however, the addition of 1% Ti to Fe18Cr-6Al was found to improve resistance in
cyclic oxidation in air at 950 C (1740 F)
(Ref 70). It is not clear whether the benecial
Table 3.12 Results of oxidation tests for various alloys at indicated temperatures in flowing air
(30 cm/min) for 1008 h
980 C (1800 F)
Alloy
Metal loss,
mm (mils)
1095 C (2000 F)
Average
metal
affected,
mm (mils)
Metal loss,
mm (mils)
1150 C (2100 F)
Average
metal
affected,
mm (mils)
Metal loss,
mm (mils)
1205 C (2200 F)
Average
metal
affected,
mm (mils)
Metal loss,
mm (mils)
Average
metal
affected,
mm (mils)
214
0.0025
(0.1) 0.005
(0.2)
0.0025
(0.1)
0.0025
(0.1)
0.005
(0.2)
0.0075
(0.3)
0.005
(0.2)
0.018
(0.7)
601
600
0.013
0.0075
(0.5) 0.033
(0.3) 0.023
(1.3)
(0.9)
0.03
0.028
(1.2)
(1.1)
0.067
0.041
(2.6)
(1.6)
0.061
0.043
(2.4)
(1.7)
0.135
0.074
(5.3)
(2.9)
0.11
0.13
(4.4)
(5.1)
0.19
0.21
(7.5)
(8.9)
230
S
617
333
X
671
0.0075
0.005
0.0075
0.0075
0.0075
0.0229
(0.3)
(0.2)
(0.3)
(0.3)
(0.3)
(0.9)
0.018
0.013
0.033
0.025
0.023
0.043
(0.7)
(0.5)
(1.3)
(1.0)
(0.9)
(1.7)
0.013
0.01
0.015
0.025
0.038
0.038
(0.5)
(0.4)
(0.6)
(1.0)
(1.5)
(1.5)
0.033
0.033
0.046
0.058
0.069
0.061
(1.3)
(1.3)
(1.8)
(2.3)
(2.7)
(2.4)
0.058
0.025
0.028
0.05
0.11
0.066
(2.3)
(1.0)
(1.1)
(2.0)
(4.5)
(2.6)
0.086
0.043
0.086
0.1
0.147
0.099
(3.4) 0.11
(1.7) >0.8I
(3.4) 0.27
(4.0) 0.18
(5.8) >0.9
(3.9) 0.086
(4.5) 0.20
(31.7) >0.8I
(10.6) 0.32
(7.1) 0.45
(35.4) >0.9
(3.4) 0.42
(7.9)
(31.7)
(12.5)
(17.7)
(35.4)
(16.4)
625
Waspaloy
R-4I
263
0.0075
0.0152
(0.3) 0.018
(0.6) 0.079
(0.7)
(3.1)
0.084
0.036
(3.3)
(1.4)
0.12
0.14
(4.8)
(5.4)
0.41
0.079
(16.0)
(3.1)
0.46
0.33
(18.2) >1.2
(13.0) >0.40
(47.6) >1.2
(15.9) >0.40
(47.6)
(15.9)
0.0178
0.0178
(0.7) 0.122
(0.7) 0.145
(4.8)
(5.7)
0.086
0.089
(3.4)
(3.5)
0.30
0.36
(11.6)
(14.2)
0.21
0.18
(8.2)
(6.9)
0.44
0.41
(17.4) >0.73
(16.1) >0.91
(28.6) >0.73
(35.7) >0.91
(28.6)
(35.7)
188
25
150
6B
0.005
0.01
0.01
0.01
(0.2)
(0.4)
(0.4)
(0.4)
0.015
0.018
0.025
0.025
(0.6)
(0.7)
(1.0)
(1.0)
0.01
0.23
0.058
0.35
(0.4)
(9.2)
(2.3)
(13.7)
0.033
0.26
0.097
0.39
(1.3) 0.18
(10.2) 0.43
(3.8) >0.68
(15.2) >0.94
(7.2) 0.2
(16.8) 0.49
(26.8) >0.68
(36.9) >0.94
(8.0)
(19.2)
(26.8)
(36.9)
(21.7)
(37.9)
(46.1)
(36.8)
(21.7)
(37.9)
(46.1)
(36.8)
556
Multimet
0.01
0.01
(0.4) 0.028
(0.4) 0.033
(1.1)
(1.3)
0.025
0.226
(1.0)
(8.9)
0.067
0.29
(2.6) 0.24
(11.6) >1.2
(9.3) 0.29
(47.2) >1.2
(11.6) >3.8
(47.2) >3.7
800H
RA330
310
316
304
446
0.023
0.01
0.01
0.315
0.14
0.033
(0.9)
(0.4)
(0.4)
(12.4)
(5.5)
(1.3)
0.046 (1.8) 0.14

0.11
(4.3) 0.02
0.028 (1.1) 0.025
0.36 (14.3) >1.7
0.21
(8.1) >0.69
0.058 (2.3) 0.33
(5.4) 0.19
(0.8) 0.17
(1.0) 0.058
(68.4) >1.7
(27.1) >0.69
(13.1) 0.37
(7.4) 0.19
(7.5) 0.23
(6.7) 0.041
(1.6) 0.22
(2.3) 0.075
(3.0) 0.11
(68.4) >2.7
(105.0) >2.7
(27.1) >0.6
(23.6) >0.6
(14.5) >0.55
(21.7) >0.55
>0.55
>0.96
>1.I7
>0.94
>0.55
>0.96
>1.I7
>0.94
(150.0) >3.8
(146.4) >3.7
(8.9) 0.29
(11.3) 0.35
(8.7) 0.096
(3.8) 0.21
(4.4) 0.2
(8.0) 0.26
(105.0) >3.57 (140.4) >3.57
(23.6) >1.7
(68.0) >1.73
(21.7) >0.59
(23.3) >0.59
(150.0)
(146.4)
(13.6)
(8.3)
(10.3)
(140.4)
(68.0)
(23.3)
Note: 3304 cm3/min of ow rate in a 1.75 in. diam furnace tube. The moisture was removed from the air by a lter prior to entering into the furnace tube. Specimens were
cathodically descaled for measurement of the metal loss. The average metal affected is the sum of the metal loss and the depth of internal attack. The depth of internal attack
was measured by metallography. Source: Ref 68
31/10/2007 12:45PM Plate # 0
pg 41
effect of titanium for oxidation resistance is only

for alumina formers such as in this case (Fe18Cr-6Al), but not for chromia formers in Ni20Cr alloy. Niobium is another alloying element
that may be detrimental to alloy oxidation resistance at very high temperatures. The relatively
poor oxidation resistance of alloy 625 at 1095
and 1150 C (2000 and 2100 F) can be attributed to niobium.
Cobalt-base alloys with tungsten, such as
alloy 188 (Co-22Cr-22Ni-14W-0.04La), alloy 25
(Co-20Cr-10Ni-15W), and alloy 6B (Co-30Cr4.5W-1.2C), suffered rapid oxidation at 1205 C
(2200 F). A cobalt-base alloy, alloy 150 (Co27Cr-18Fe), containing no tungsten also suffered
rapid oxidation attack at 1205 C (2200 F).
Again, the oxidation of a cobalt-base alloy can be
signicantly improved with some modication
of alloying elements. Alloy 25 with 15% W
exhibits excellent creep-rupture strengths at high
temperatures. However, because of the high level
of tungsten, the alloy suffers high oxidation rates
at very high temperatures, such as 1095 and
1150 C (2000 and 2100 F). With slight
increase in chromium and nickel along with the
addition of lanthanum, the result of the modication was alloy 188. As shown in Table 3.12,
alloy 188 exhibits signicantly better oxidation
resistance than alloy 25 at 1095 and 1150 C
(2000 and 2100 F).
The best alloy among those investigated was
an alumina former, alloy 214 (Ni-16Cr-4.5Al-Y).
The depth of oxidation attack (average metal
affected) was found to be less than 0.025 mm
(1.0 mils) after 1008 h at temperatures up to
Fig. 3.50 Scanning electron micrograph showing the early

stage of oxidation in air at 1200 C (2200 F) for
1 hour for alloy 263, revealing titanium-rich and Cr-Ti oxides on
the outermost oxide scale. Area 1: 28.1% Cr, 70.9% Ti, 0.8% Co,
0.2% Ni. Area 2: 57.0% Cr, 36.8% Ti, 2.8% Co, 25% Ni, 1.0% Fe.
1205 C (2200 F). The oxidation resistance of

alloy 214 is also presented in Fig. 3.42 to 3.45
and 3.47.
The oxidation data presented in Table 3.12
were generated with a weekly cycle (168 h).
When the cyclic frequency was increased to 25 h
cycles, oxidation rates were increased for all the
alloys tested. However, some alloys are more
sensitive to cyclic oxidation than others. The
effect of thermal cycling on the oxidation resistance of various alloys in air at 1095 C
(2000 F) is illustrated in Table 3.13, which
compares once-a-week (168 h) cycle data (Ref
68) with 25 h cycle data (Ref 71). All the data are
presented in terms of the average depth of metal
affected, which represents the metal loss plus the
depth of internal oxidation attack. Both sets of
the data were generated under the same test
conditions using the same test furnaces and test
procedures except the differences in cyclic frequencies. Some chromia formers, such as alloys
230, S, 188, 556, and 310, showed good resistance to thermal cycling.
A long-term oxidation test program was
undertaken to test alloys up to 2 years at 980,
1095, and 1150 C (1800, 2000, and 2100 F)
Table 3.13 Comparative oxidation resistance of

various alloys in flowing air between 168 h and
25 h cycles at 1095 C (2000 F)
Total depth of attack,
mm (mils)
Extrapolated oxidation
rate, mm/yr (mpy)
Alloy
1008 h/168 h
1050 h/25 h
168 h cycles
25 h cycles
214
601
600
671
230
S
G-30
617
RA333
625
Waspaloy
263
188
25
150
6B
556
Multimet
800H
RA330
310
446
0.003 (0.1)
0.066 (2.6)
0.041 (1.6)
0.061 (2.4)
0.003 (1.3)
0.003 (1.3)
0.122 (4.8)
0.046 (1.8)
0.058 (2.3)
0.122 (4.8)
0.137 (5.4)
0.361 (14.2)
0.003 (1.3)
0.259 (10.2)
0.097 (3.8)
0.394(15.5)
0.066 (2.6)
0.295 (11.6)
0.188 (7.4)
0.170 (6.7)
0.058 (2.3)
0.368 (14.5)
0.025 (1.0)
0.297 (11.7)
0.185 (7.3)
0.584 (23.0)
0.086 (3.4)
0.061 (2.4)
0.203 (8.0)
0.267 (10.5)
0.130 (5.1)
0.414 (16.3)
0.414 (16.3)
0.478 (18.8)
0.058 (2.3)
0.490 (19.3)
0.353 (13.9)
>0.800 (31.5)
0.117 (4.6)
0.381 (15.0)
0.406 (16.0)
0.442 (17.4)
0.112 (4.4)
0.655 (25.8)
0.025 (1)
0.58 (23)
0.36 (14)
0.53 (21)
0.28 (11)
0.28 (11)
1.07 (42)
0.41 (16)
0.51 (20)
1.07 (42)
1.19 (47)
3.12 (123)
0.28 (11)
2.26 (89)
0.84 (33)
3.43 (135)
0.58 (23)
2.57 (101)
1.63 (64)
1.47 (58)
0.51 (20)
3.20 (126)
0.20 (8)
2.49 (98)
1.55 (61)
4.88 (192)
0.71 (28)
0.51 (20)
1.70 (67)
2.24 (88)
1.09 (43)
3.45 (136)
3.45 (136)
3.99 (157)
0.48 (19)
4.09 (161)
2.95 (116)
>6.68 (263)
0.97 (38)
3.18 (125)
3.40 (134)
3.68 (145)
0.94 (37)
5.46 (215)
Note: 3304 cm3/min of ow rate in a 1.75 in. diam furnace tube. The moisture was
removed from the air by a lter prior to entering into the furnace tube. Specimens
were cathodically descaled for measurement of the metal loss. The total depth of
attack is the sum of the metal loss and the depth of internal attack. Metal loss was
measured by cathodically descaling the oxide scale prior to measurement of the
specimen thickness. Source: Ref 71
31/10/2007 12:45PM Plate # 0
pg 42
(Ref 49). Test specimens, which were cut from

plate products, had dimensions of 1.27 cm
(thick) by 2.54 cm (width) by 2.54 cm (length).
Tests were conducted in a box furnace with still
air; specimens were removed from the furnace
every 30 days to allow air cooling. Weights were
then measured to determine the time to initiation
of breakaway oxidation. For block specimens
(e.g., 1.25 cm, or 0.5 in., thick specimens), oxide
scales were cracking, breaking, and spalling with
bing noises as soon as the specimens were
removed from the hot zone in the furnace.
Cracking noises would not stop until almost all
the oxide scales were broken and spalled off.
Thus, no cathodic descaling was necessary to
remove the oxide scales for measurement of the
specimen thickness at the end of the test. When
similar oxidation testing was performed with
thin test coupons (about 3.2 mm, (0.125 in.) or
200
Weight gain, mg/cm2
0
200
400
600
800
1000
0 30 60 90 120 150 180 210 240 270 300 330 360

Days
Fig. 3.51
The oxidation behavior of alloy 800H tested in still

air at 1095 C (2000 F) involving a thick, blocky
specimen (1.25 cm, or 0.5 in., thick) cycling to room temperature
every 30 days for weight measurement, showing the alloy was
under protective scales initially for about 30 days and then suffered breakaway oxidation. Courtesy of Haynes International, Inc.
thinner), no cracking noises were heard during

specimen cooling from the hot zone, and oxide
scales remained on the specimen surface for
these thin test coupons. Figure 3.51 shows
weight change data for alloy 800H in this longterm testing at 1095 C (2000 F) using thick,
blocky specimens, showing breakaway oxidation with linear weight loss after 60 days of
exposure. Some alloys showed no breakaway
oxidation even after 2 years of testing. Figure
3.35 shows no breakaway oxidation for HR120
after 2 years of testing at 980 C (1800 F),
while 800H and RA85H suffered breakaway
oxidation. HR160 and 601 showed no breakaway oxidation after one year of testing at
1095 C (2000 F), while alloy 800H suffered
breakaway oxidation (Fig. 3.46). At the end of
testing for 2 years (720 days) at 980 C
(1800 F) and 1 year (360 days) at 1093, 1150,
and 1200 C (2000, 2100, and 2200 F), specimens were cut, mounted, and polished for
metallographic determination of the depth of
internal attack. Metal loss was determined by
subtracting the original specimen thickness from
the thickness after testing. The data generated
from blocky specimens are summarized in
Tables 3.14 to 3.16. The annual oxidation rates in
terms of the total depth of oxidation attack are
included in Tables 3.14 to 3.16 at 980, 1090, and
1150 F (1800, 2000, and 2100 F), respectively. These oxidation rate values would be
considered to be quite reasonable, since the test
duration was almost 2 years for 980 C
(1800 F) and about 1 year for 1090 and 1150 C
testing. At 980 C (1800 F), alloys 230, 617,
HR120, 556, and HR160 exhibited oxidation
rates of less than 10 mpy. At 1090 C (2000 F),
only alloy 230 exhibited about 10 mpy of oxidation rate, while other alloys tested exhibited
more than 20 mpy. At 1150 C (2100 F), all
alloys tested exhibited more than 30 mpy of
Table 3.14 Oxidation of several high temperature alloys in still air at 980 C (1800 F) for 720 days
with specimens cycling to room temperature every 30 days
Alloy
230
617
HR120
556
HR160
601
RA85H
800HT
Weight change,
mg/cm2
Metal loss,
mm (mils)
Total depth of
attack, mm (mils)
Oxidation rate,
mm (mpy)
1.4
1.0
33.7
19.8
51.2
9.9
122.2
417.8
0.00254 (0.1)
0
0.04060 (1.6)
0.02286 (0.9)
0.0635 (2.5)
0.0127 (0.5)
0.16002 (6.3)
0.52578 (20.7)
0.14732 (5.8)
0.23876 (9.4)
0.30988 (12.2)
0.38608 (15.2)
0.42418 (16.7)
0.56896 (22.4)
1.36398 (53.7)
2.02692 (79.8)
0.0508 (2)
0.127 (5)
0.1524 (6)
0.2032 (8)
0.2286 (9)
0.2794 (11)
0.6858 (27)
1.0414 (41)
Note: Total depth of attack = metal loss + internal attack. Source: Ref 49
31/10/2007 12:45PM Plate # 0
pg 43
Table 3.15 Oxidation of several high-temperature alloys in still air at 1095 C (2000 F) for
360 days with specimens cycling to room temperature every 30 days
Alloy
230
556
RA330
HR160
HR120
601
800HT
RA85H
Metal loss, mm (mils)
Total depth of attack, mm (mils)
Oxidation rate, mm/yr (mpy)
42.1
298.1
405.0
73.1
665.7
110.5
893.9
348.1
0.04826 (1.9)
0.36322 (14.3)
0.50800 (20.0)
0.09144 (3.6)
0.82804 (32.6)
0.13716 (5.4)
1.12522 (44.3)
0.45466 (17.9)
0.27178 (10.7)
0.53848 (21.2)
0.60706 (23.9)
0.73660 (29.0)
0.96520 (38.0)
1.14554 (45.1)
1.29540 (51.0)
2.03962 (80.3)
0.27940 (11)
0.55880 (22)
0.60960 (24)
0.73660 (29)
0.99060 (39)
1.16840 (46)
1.32080 (52)
2.05740 (81)
Table 3.16 Oxidation of several high-temperature alloys in still air at 1150 C (2100 F) for
360 days with specimens cycling to room temperature every 30 days
Alloy
230
617
HR120
HR160
800H
601
RA85H
Total depth of attack, mm (mils)
Oxidation rate, mm/yr (mpy)
249.7
452.7
894.2
155.5
1315.6
258.7
389.2
0.28194 (11.1)
0.54102 (21.3)
1.10998 (43.7)
0.19304 (7.6)
1.65608 (65.2)
0.32004 (12.6)
0.50800 (20.0)
0.83680 (34.0)
0.94488 (37.2)
1.34620 (53.0)
1.49098 (58.7)
1.78562 (70.3)
1.84912 (72.8)
2.40792 (94.8)
0.88900 (35)
0.96520 (38)
1.37160 (54)
1.52400 (60)
1.80340 (71)
1.87960 (74)
2.4384 (96)
oxidation rate. At 1200 C (2200 F), aluminaformer alloy 214 showed little or no oxidation
attack (Fig. 3.47).
These data are valuable in providing readers
with the oxidation data in terms of the total depth
of oxidation attack in air based on the actual
measurements of the specimens after 1 to 2 years
of testing. Since the data were generated from
thick, blocky specimens, caution should be used
when the data are being considered for application in thin-gage sheets or foils. This is related to
the reservoir effect of a solute alloying element
for the formation of a protective oxide scale.
The discussion of the reservoir issue and the
oxidation in thin foils is presented later.
In this test program (Ref 49), in addition to the
metal loss caused by formation of external oxide
scales and internal attack caused by formation of
internal oxides and/or voids, the weight-loss
values were also determined. The weight loss of
the specimens is related mostly to the metal loss
resulting from the removal of the external oxide
scales and is not signicantly affected by the
formation of internal oxides and/or voids. The
weight-loss values of the alloys tested are plotted
against their corresponding metal-loss values
at 980, 1090, and 1150 C (1800, 2000, and
2100 F), revealing a nice straight line correlation (Fig. 3.52). Alloys tested were 230, 617,
601, 556, HR160, HR120 RA330, 800HT, and
RA85H. These alloys are primarily chromia

formers. This correlation may be useful in making rough estimates of the metal loss for an alloy
that showed only weight-loss data.
The depth of oxidation attack was also investigated by John (Ref 15) for a wide variety of
commercial alloys in isothermal air oxidation
testing. Table 3.17 summarizes his data in terms
of the temperature at which the oxidation rate
reaches 10 mpy. Figure 3.53 illustrates some
oxidation data in terms of oxide penetration as a
function of test temperature in air after 1 year for
some alloys (Ref 15). Table 3.18 shows the depth
of oxidation attack of various heat-resistant
alloys after cyclic oxidation tests at 1100 C
(2010 F) in air + 5% H2O for 504 h with specimens cycling out of the furnace every 15 min
(Ref 72).
Lai et al. (Ref 73) reported the oxidation data
generated from a eld test inside a radiant tube
red with natural gas with an average temperature of 1010 C (1850 F) (Table 3.19). The test
rack containing coupons of various alloys was
exposed for about 3000 h. Many chromia formers, such as alloys 601, 230, 556, 310, 600, and
RA330, were found to perform well, with
extrapolated oxidation rates of less than 0.5 mm/
yr (20 mpy). Type 304, however, suffered severe
oxidation attack with an extrapolated oxidation
rate of more than 4.4 mm/yr (>175 mpy). The
31/10/2007 12:45PM Plate # 0
pg 44
70
60
1.50
1.00
40
30
20
Metal loss, mm
Metal loss, mils
50
0.50
10
0
0
200
400
600
800
1000
1200
1400
Weight loss, mg/cm2

Correlation between weight loss (mg/cm2) and the depth of metal loss (mils) for commercial alloys that are primarily chromia
formers tested in air at 980 C (1800 F)/720 days, 1095 C (2000 F)/360 days, and 1150 C (2100 F)/360 days. Alloys
tested were 230, 617, 601, 556, HR160, HR120, RA330, 800HT, and RA85H. 1.0 mil = 0.0254 mm
Fig. 3.52
Table 3.17 The oxidation rate of total depth of

attack was reached after 1 year in air
Maximum temperature for
0.25 mm/yr (10 mpy), C (F)
Carbon steel
Copper
Nickel
9Cr-1Mo
410
304
617
803
625
800H
601GC
DS
230
310
RA330
446
556
HR120
253MA
602CA
MA956
214
C11000
N02270
S50400
S41000
S30400
N06617
N06625
N08810
N06230
S31000
S33000
S44600
R30556
S30815
S67956
N07214
604 (1120)
677 (1250)
782 (1440)
799 (1470)
832 (l530)
893 (1640)
938 (1720)
954 (1750)
960 (1760)
966 (1770)
977 (1790)
977 (1790)
982 (1800)
982 (l800)
999 (1830)
1010 (1850)
1010 (1850)
1010 (1850)
1082 (1980)
1121 (2050)
>1150 (>2100)
>1150 (>2100)
AISI 410
AISI 304
10
0.3
Alloy 617
Nickel
AISI 310
Alloy 800 H
Penetration, mils
UNS No.
2.5
9Cr -1Mo
Carbon steel
Penetration, mils
Alloy
102
1
0.03
0.1
1000
1200
1400
1600
1800
2000
Temperature, F
Source: Ref 15
Fig. 3.53
alumina former, alloy 214, showed little or no

oxidation attack with an extrapolated oxidation
rate of about 0.076 mm/yr (3 mpy). In another
eld test (Ref 74), a test rack containing coupons
of various alloys was placed in a natural-gas-red
furnace used for reheating ingots and slabs of
nickel- and cobalt-base alloys. The test was
conducted for about 113 days at temperatures
varying from 1090 to 1230 C (2000 to
2250 F), with frequent cycles to 540 C
Oxidation penetration (metal loss + internal attack)

as a function of test temperature for 1 year in air for
a variety of commercial alloys. Source: Ref 15
(1000 F) during furnace idling. The results are

summarized in Table 3.20. All the chromia formers tested suffered severe oxidation attack. The
alumina former (alloy 214), however, exhibited
little attack. Examination of the oxide scale
formed on alloy 214 was found to consist of
essentially aluminum-rich oxides.
31/10/2007 12:45PM Plate # 0
pg 45
Table 3.18 Cyclic oxidation resistance of various heat-resistant alloys at 1100 C (2010 F) for 504 h
in air-5H2O
Specic weight change (descaled), mg/cm2
Alloy
ACI grade HK
310SS
800
601
617
X
RA333
IN-814
188
MA-956
Mean
Range
Metal loss,
mm (mils)
Maximum attack,
mm (mils)
105.8
149.0
168.6
11.0
13.5
20.0
30.5
3.5
25.0
1.0
98 to 124
92 to 235
83 to 223
6.3 to 17.2
6.5 to 17.5
10.0 to 29.5
2.5 to 4.9
13.0 to 40.5
0.3 to 1.5
0.25 (9.9)
0.31 (12.2)
0.39 (15.4)
<0.02 (0.8)
0.05 (2.0)
0
0.03 (1.2)
0.01 (0.4)
0.35 (13.8)
0.38 (15.0)
0.59 (23.2)
0.12 (4.7)
0.25 (9.9)
<0.01 (0.4)
0.15 (5.9)
0.02 (0.8)
Table 3.19 Results of field test in a

natural-gas-fired radiant tube at 1010 C
(1850 F) for 3000 h
Alloy
214
601
230
556
310
600
RA330
800H
309
304
Metal loss,
mm (mils)
Maximum metal
affected(a), mm (mils)
Oxidation rate,
mm/yr (mpy)
0.003 (0.1)
0.023 (0.9)
0.028 (1.1)
0.018 (0.7)
0.041 (1.6)
0.018 (0.7)
0.048 (1.9)
0.12 (4.7)
0.50 (19.7)
>1.5 (60)(b)
0.025 (1)
0.076 (3)
0.10 (4)
0.10 (4)
0.10 (4)
0.15 (6)
0.15 (6)
0.30 (12)
0.50 (20)
>1.5 (60)(b)
0.076 (3)
0.23 (9)
0.30 (12)
0.30 (12)
0.30 (12)
0.46 (18)
0.46 (18)
0.89 (35)
1.5 (58)
>4.4 (175)
(a) Metal loss + maximum internal penetration. (b) Sample was consumed. Source:
Ref 73
Table 3.20 Results of field test in a

natural-gas-fired furnace for reheating nickeland cobalt-base alloy ingots and slabs for
113 days at 1090 to 1230 C (2000 to 2250 F)
with frequent cycles to 540 C (1000 F)
Alloy
214
RA330
601
600
800H
310SS
304SS
316SS
446SS
Metal loss,
mm (mils)
Maximum metal
affected(a),
mm (mils)
0.013 (0.5)
0.39 (15.5)
0.18 (7.2)
0.64 (25.0)
>0.79 (31.0)(b)
>1.0 (41.0)(b)
>1.5 (60.0)(b)
>1.6 (63.0)(b)
>0.61 (24.0)(b)
0.11 (4.5)
0.65 (25.5)
0.95 (37.2)
1.1 (45.0)
>0.79 (31.0)(b)
>1.0 (41.0)(b)
>1.5 (60.0)(b)
>1.6 (63.0)(b)
>0.61 (24.0)(b)
(a) Metal loss + internal penetration. (b) Samples were consumed. Source: Ref 74
3.4.10 Oxide-Dispersion-Strengthened
(ODS) Alloys
Oxide-Dispersion Strengthened alloys use
very ne oxide particles that are uniformly distributed throughout the matrix to provide excessive strengthening at very high temperatures.
These oxide particles, typically yttrium oxide, do
not react with the alloy matrix so no coarsening
or dissolution occurs during the exposure to
very high temperatures, thus maintaining the
strengthening of the alloy. This group of superalloys is produced using specialty powders that
are manufactured by the mechanical alloying
process. These powders are essentially composite powders with each particle containing a
uniform distribution of submicron oxide particles
in an alloy matrix. The process of producing
these ODS powders involves repeated fracturing
and rewelding of a mixture of powder particles in
vertical attritors or horizontal ball mills (Ref 75).
Alloy powders are then canned, degassed, and
hot extruded, followed by hot working and
annealing to produce a textured microstructure

(Ref 75). Alloys are available in mill products
such as bar, plate, sheet, and so forth, or custom
forgings. Some ODS alloys are shown in
Table 3.21 (Ref 75). The oxidation behavior
of some of these ODS alloys tested in air containing 5% H2O at 1200 C (2190 F) is shown
in Fig. 3.54 (Ref 75). The oxidation behavior
of MA956 compared with those of several ironand nickel-base alloys at 1100 C (2010 F)
is shown in Fig. 3.55 (Ref 76). Additional oxidation data for some ODS alloys is presented in
Section 3.4.12.
3.4.11 Effect of Oxygen Concentration
on Oxidation
Air atmosphere consists primarily of oxygen
and nitrogen with some water vapor and small
amounts of inert gases, such as argon, neon, and
helium. Dry air consists of essentially 21% O2
31/10/2007 12:45PM Plate # 0
pg 46
Table 3.21
Nominal chemical compositions of several ODS alloys
Alloy
Ni
Fe
Cr
Al
Ti
Mo
Ta
Y2O3
Zr
MA754
MA758
MA760
MA6000
MA956
bal
bal
bal
bal
bal
20
30
20
15
20
0.3
0.3
6.0
4.5
4.5
0.5
0.5
2.5
0.5
3.5
4.0
2.0
2.0
2.0
0.6
0.6
0.95
1.1
0.5
0.01
0.01
0.15
0.15
Composition, wt%
Note: All alloys contain 0.05% C. Source: Ref 75
50
Mass change, mg/cm2
MA 760
MA 754
50
100
0
0.5
1.0
1.5
2.0
150
MA 6000
200
250
0
10
20
30
40
50
2.5
3.0
3.5
60
Mass change, Ib/in.2 103
0.5
MA 956
0
Exposure time, days
Fig. 3.54
Oxidation behavior of several ODS alloys in

air containing 5% H2O at 1200 C (2190 F).
Source: Ref 75
and 78% N2 with about 1% inert gases. In combustion, the concentration of oxygen may vary.
Also, in some processes, oxygen may be the only
gaseous component to which the equipment is
exposed. Thus, the effect of oxygen concentration on the oxidation behavior of alloys may need
to be evaluated for some applications. John (Ref
15, 77) investigated the oxidation behavior of a
wide variety of commercial alloys in N2-O2
mixtures with the concentration of oxygen
varying from 1 to 100%. The data generated at
871 C (1600 F) are shown in Fig. 3.56 (Ref
77), and data generated at 927 C (1700 F) are
shown in Fig. 3.57 (Ref 15). The data presented
in Fig. 3.56 were based on the 1152 h testing,
while the data in Fig. 3.57 were based on tests
after 1 year. At 871 C (1600 F), Type 304 was
the only alloy that showed signicant increase in
oxidation attack from about 0.25 mm/yr
(10 mpy) in N2-21%O2 to close to 2.5 mm/yr
(100 mpy) in 100% O2. All other alloys in the
gure showed about 10 mpy or less of oxidation
attack at three different levels of O2 concentrations (1%, 21%, and 100%). Figure 3.57 shows
the oxidation behavior of a number of alloys at
927 C (1700 F). The alloys that were found to
increase oxidation attack with increasing oxygen
Fig. 3.55
Cyclic oxidation resistance of ODS alloy MA956

compared with alloy 601, HK alloy, alloy 800, and
Type 310. Source: Ref 76
concentration included 9Cr-1Mo steel, 410, 304,

and 617, while carbon steel, nickel, 800H, and
310 were relatively unaffected by oxygen concentrations. In Fig. 3.57, Type 304 was found to
exhibit approximately 0.25 mm/yr (10 mils) of
attack at 927 C (1700 F) after 1 year in 100%
O2, while close to 2.5 mm/yr (100 mpy) of
attack was extrapolated based on 1152 h exposure at 871 C (1600 F) in 100% O2, as shown
in Fig. 3.56. Extrapolation from short-term tests
here showing a higher oxidation rate at lower
temperature could be an issue here. More longterm tests are needed. It is also of practical
interest to perform long-term tests to evaluate the
oxidation behavior in 100% O2 environments for
some alloys that are to be used in chemical processes involving 100% O2.
3.4.12 High-Velocity Combustion
Gas Streams
Oxidation of alloys can signicantly increase
under high-velocity gas streams. Combustors
and transition ducts in the gas turbine are subject
to such conditions. These components are also
31/10/2007 12:45PM Plate # 0
103
AISI 304
AISI 310
102
956MA
Penetration rate, mpy
10
253MA
HR120
800HT
AISI 446
101
617
102
230
214
103
601GC
104
102
101
Partial pressure of O2, atm
Fig. 3.56
602CA
556
Effect of oxygen concentration in the N2-O2 mixture on the oxidation penetration (metal loss + internal attack) at 871 C
(1600 F) for 1152 h. 1.0 mil = 0.025 mm. Source: Ref 77
subject to severe thermal cycling, particularly for

gas turbines in airplane engines. Laboratory
burner rigs have been developed to evaluate this
type of oxidation, often referred to as dynamic
oxidation, under the condition of very high gas
velocities. Some of these dynamic oxidation
burner rigs are described elsewhere (Ref 7883).
Lai (Ref 82) investigated a wide range of
alloysfrom stainless steels to superalloysin a
burner rig that generated a combustion gas
stream with 0.3 Mach (100 m/s) velocity. The
specimens were held in a carousel-type holder
rotating at 30 rpm with respect to the combustion
gas stream. Every 30 min the carousel was
withdrawn from the hot zone, and quenched to
less than 260 C (500 F) by a blast of cold air,
and then automatically reinserted back into the
hot zone. The specimens were subject to severe
thermal cycling. Combustion was generated
using No. 2 fuel oil with an air-to-fuel ratio of 50
to 1, producing a high-velocity (0.3 Mach, or
100 m/s) test gas. The tests were conducted at
1090 C (2000 F) with 30 min cycles, and the
results are tabulated in Table 3.22.
At 1090 C (2000 F) with a high-velocity
gas stream plus severe thermal cycling, most
alloys suffered signicant metal loss, which
constituted a large portion of the total depth of
oxidation attack. With protection by an aluminum oxide scale, alloy 214 suffered very little
attack. The alloy showed no sign of breakaway

oxidation after 500 h. Figure 3.58 shows the
oxide scale formed after testing for 500 h
(1000 cycles) (Ref 82). The scale consisted of
aluminum-rich oxides. After 1000 h (2000
cycles) of testing, the scale remained aluminumrich. The maximum metal affected (metal loss
+maximum internal penetration) remained about
the same after 1000 h compared to after 500 h
(Ref 82).
The test results generated at 980 C (1800 F)
for 1000 h (2000 cycles) are shown in Table 3.23
(Ref 82). Unlike 1090 C (2000 F) testing
(Table 3.22), testing at 980 C (1800 F) resulted
in internal oxidation and nitridation in addition
to metal loss. Internal nitridation penetrated
deeper into metal interior than internal oxidation
penetration. Table 3.23 included only internal
oxidation penetration data (i.e., maximum
metal affected = metal loss + internal oxidation
penetration). Limited tests were conducted to
determine the effect of thermal cycling by testing
at 980 C (1800 F) for 1000 h with 30 min
cycling and without thermal cycling in dynamic
oxidation testing (Ref 82). As expected, thermal
cycling primarily contributed metal loss portion
of the oxidation attack. The results are summarized in Table 3.24 (Ref 82). Total attack presented in Table 3.24 was based on metal loss and
internal oxidation penetration. Since internal
pg 47
31/10/2007 12:46PM Plate # 0
10
Carbon steel
10
9Cr-1Mo
Penetration, mils
10
Penetration, mm
AISI 410
1
Nickel
Alloy 800 H
10
AISI 310
10
AISI 304
Alloy 617
1
10
10
10
10
pO , atm
2
Fig. 3.57
Effect of oxygen concentration in the N2-O2 mixture on the oxidation penetration (metal loss + internal attack) after 1 year at
927 C (1700 F) for various commercial alloys. 1.0 mils = 0.025 mm. Source: Ref 15
nitridation attack was found to penetrate deeper

into the alloy than internal oxidation attack does
for many alloys, the total depth of attack for
many alloys under dynamic oxidation test conditions was more than that reported in Tables
3.23 and 3.24. The oxidation/nitridation behavior of various alloys under dynamic oxidation
test conditions is discussed in Chapter 4 Nitridation.
Hicks (Ref 83) performed dynamic oxidation
tests with 170 m/s gas velocity at 1100 C
(2010 F) with 30 min cycles for several
wrought chromia-former superalloys and an
ODS alumina-former (MA956). Alumina former
MA956 was found to be considerably better than
chromium formers, such as alloys 230, 86, 617,
188, and 263. His results are shown in Fig. 3.59.
MA956 along with some ODS alloys was tested

by Lowell et al. (Ref 78) with 0.3 Mach gas
velocity at 1100 C (2010 F) with 60 min
cycles. ODS alloys tested included MA956 (Fe19Cr-4.4Al-0.6Y2O3), HDA8077 (Ni-16Cr4.2Al-1.6Y2O3), TD-NiCr (Ni-20Cr-2.2ThO2)
and STCA264 (Ni-16Cr-4.5Al-1Co-1.5Y2O3).
Also included in the test was physical vapor
deposition (PVD) coating of Ni-15Cr-17Al-0.2Y
on MAR-M-200 alloy (Ni-9Cr-10Co-12W-1Nb5Al-2Ti). Their results are shown in Fig. 3.60.
MA956 and HDA8077 as well as PVD Ni-Cr-AlY coating were found to perform well. No
explanation was offered in the paper for
STCA264, which did not perform as well as
HDA8077 although both alloys had similar
chemical compositions.
pg 48
31/10/2007 12:46PM Plate # 0
pg 49
Table 3.22 Dynamic oxidation resistance

of iron-, nickel-, and cobalt-base alloys in
high-velocity combustion gas stream (0.3 mach
velocity) with 30 min cycles at 1090 C (2000 F)
for 500 h
Alloy
214
230
RA333
188
556
X
RA330
S
600
310
601
617
800H
625
Multimet
Metal loss,
mm (mils)
Maximum metal
0.013 (0.5)
0.056 (2.2)
0.10 (4.0)
0.19 (7.5)
0.22 (8.7)
0.23 (9.0)
0.28 (10.9)
0.30 (11.8)
0.44 (17.2)
0.54 (21.2)
0.27 (10.7)
0.32 (12.4)
0.77 (30.5)(b)
>0.79 (31.0)(c)
1.25 (49.1)(d)
0.046 (1.8)
0.15 (5.7)
0.22 (8.7)
0.27 (10.7)
0.30 (11.7)
0.34 (13.5)
0.35 (13.6)
0.39 (15.2)
0.53 (20.7)
0.61 (24.1)
0.61 (24.0)
0.61 (24.0)
0.86 (34.0)(b)
>0.79 (31.0)(c)
1.42 (55.8)(d)
Note: Gas velocity was 0.3 mach (100 m/s, or 225 mph); samples were cycled to
less than 260 C (500 F) once every 30 min; 50 to 1 air-to-fuel ratio; two parts
No. 1 fuel oil and one part No. 2 fuel oil. Internal nitridation occurred in some
alloys, but is not included in the current data. See section 4.3.3 in Chapter 4.
(a) Metal loss + maximum internal, penetration. (b) Extrapolated from 400 h;
sample was about to be consumed after 400 h. (c) Sample was consumed in
500 h. (d) Extrapolated from 225 h; sample was about to be consumed after
225 h. Source: Ref 82
Table 3.23 Dynamic oxidation resistance

of iron-, nickel-, and cobalt-base alloys in
high-velocity combustion gas stream at 980 C
(1800 F) for 1000 h
Alloy
214
230
188
556
X
S
RA333
625
617
RA330
Multimet
800H
310
600
601
304
316
Metal loss,
mm (mils)
Maximum metal
0.010 (0.4)
0.020 (0.8)
0.028 (1.1)
0.043 (1.7)
0.069 (2.7)
0.079 (3.1)
0.064 (2.5)
0.12 (4.9)
0.069 (2.7)
0.20 (7.8)
0.30 (11.8)
0.31 (12.3)
0.35 (13.7)
0.31 (12.3)(b)
0.076 (3.0)
>9.0 (354)(c)
>9.0 (354)(c)
0.031 (1.2)
0.089 (3.5)
0.107 (4.2)
0.158 (6.2)
0.163 (6.4)
0.17 (6.6)
0.18 (7.0)
0.19 (7.6)
0.27 (10.7)
0.30 (11.8)
0.38 (14.8)
0.39 (15.3)
0.42 (16.5)
0.45 (17.8)(b)
0.51 (20.0)
>9.0 (354)(c)
>9.0 (354)(c)
Note: Gas velocity was 0.3 mach (100 m/s, or 225 mph); samples were cycled to
less than 260 C (500 F) once every 30 min; 50 to 1 air-to-fuel ratio; two parts
No. 1 fuel oil and one part No. 2 fuel oil. Internal nitridation occurred in some
alloys, but is not included in the current data. See section 4.3.3 in Chapter 4.
(a) Metal loss + maximum internal penetration. (b) Extrapolated from 917 h;
sample was about to be consumed after 917 h. (c) Extrapolated from 65 h; sample
was consumed in 65 h. Source: Ref 82
3.4.13 Breakaway Oxidation

In Fe-Cr, Fe-Ni-Cr, Ni-Cr, and Co-Cr alloy
systems, the formation of an external Cr2O3
oxide scale provides the oxidation resistance for
Fig. 3.58
Oxide scales formed on alloy 214 in a high-velocity

gas stream (0.3 Mach velocity) with 30 min cycles
at 1090 C (2000 F) for 500 h. Area 1: 96.5% Al, 1.5% Cr, 0.1%
Fe, 1.9% Ni. Area 2: 75.2% Al, 6.2% Cr, 2.6% Fe, 16.0% Ni. Area
3: 95.8% Al, 1.0% Cr, 0.1% Fe, 3.1% Ni. Area 4: 53.0% Al, 2.8%
Cr, 9.2% Fe, 35.0% Ni. Source: Ref 82
Table 3.24 Effect of thermal cycling in

dynamic oxidation behavior of several
nickel-base alloys at 980 C (1800 F) for 1000 h
No thermal cycling
Thermal cycling
Alloy
Metal loss,
mm
Total attack(a),
mm
Metal loss,
mm
Total attack(a)
mm
230
617
X
263
0.04
0.03
0.03
0.07
0.11
0.16
0.12
0.21
0.07
0.17
0.16
0.32
0.16
0.24
0.23
0.42
(a) Metal loss + internal oxidation. Source: Ref 82
the alloy. The growth of the Cr2O3 oxide scale

follows a parabolic rate law as the exposure time
increases. As the temperature increases, the
oxide scale grows faster. The growth of the
Cr2O3 oxide scale requires a continuous supply
31/10/2007 12:46PM Plate # 0
Total depth of oxidation, mm/side
MA 956
230
0.1
617
0.2
86
188
(1971 Cast)
0.3
263
188
(1980 Cast)
0.4
(C)
0
200
400
600
800
Number of cycles
Fig. 3.59
Dynamic oxidation resistance of several wrought

superalloys including MA956 alloy in highvelocity combustion gas stream (170 m/s) at 1100 C (2010 F)
with 30 min cycles. Source: Ref 83
Specific weight change, mg/cm2
10
Coated MAR-M-200
0
MA 956
10
HDA8077
20
30
40
STCA-264
50
0
400 800 1200 1600 2000 2400 2800 3200

Cycles
Fig. 3.60
Dynamic oxidation tests at 1100 C (2010 F) in a

Mach 0.3 gas stream with each cycle consisting of
1 h at temperature followed by quenching to ambient temperature for 3 min. Source: Ref 78
of chromium from the alloy interior diffusing to

the oxide/metal interface. Continued oxidation
can eventually deplete chromium in the alloy
matrix immediately underneath the oxide scale.
When the chromium concentration in the alloy
matrix immediately beneath the oxide scale is
reduced to below a critical concentration, the
alloy matrix no longer has adequate chromium to
reform a protective Cr2O3 oxide scale when the
scale cracks or spalls due to oxide growth stresses
or thermal cycling. Once this occurs, fastgrowing, nonprotective iron oxides, or nickel
oxides or cobalt oxides (i.e., oxides of base
metal) form and grow on the alloy surface.
Breakaway oxidation, thus, initiates, and the
alloy begins to undergo oxidation at a rapid rate.
This is illustrated in Fig. 3.51. The alloy thus
requires the level of chromium immediately
underneath the chromium oxide scale to have a
critical level to allow the chromium oxide scale
to reheal.
Gleeson (Ref 84) presented air cyclic oxidation data for three chromia formers tested at
982 C (1800 F) for up to 360 days. Also presented were the corresponding chromium concentration analyzed by EDX on the surface of the
metal when the oxide scale was spalled off from
the test specimen. The data are presented in
Fig. 3.61. In this test program, thick, blocky
specimens (13 mm thick, 25 mm wide, and
25 mm long) instead of typically thin coupons
were used. The specimens were cycled to room
temperature once every 30 days for weight
measurement. Oxide scales were found to completely spall off while the specimens were
removed from the furnace for cooling to room
temperature with popping noises being heard
during cooling (Ref 85). Figure 3.61 shows that
alloy 230 exhibited very little weight loss with no
evidence of breakaway oxidation after 360 days
at 982 C (1800 F). The chromium concentration of the alloy immediately underneath the
spalled oxide scale was found to remain at about
16% with no sign of decreasing with increasing
exposure time. For HR120 alloy, the weight loss
data also revealed no evidence of breakaway
oxidation up to 360 days of exposure. The corresponding chromium concentration of the alloy
on the surface underneath the spalled oxide scale
remained approximately about 18 to 20% up to
240 days, and then dropped to about 13% after
360 days. Alloy 800HT, on the other hand,
showed breakaway oxidation after 180 days. The
corresponding chromium concentration of the
alloy on the surface immediately underneath of
the spalled oxide scale after 180 days was found
to be about 10%.
With continuing oxidation, alloy 800HT suffered linear weight loss and continued the
decrease in chromium concentration to 8% when
the exposure reached to 360 days. However,
when alloy 800HT was oxidized after 90 days
showing no sign of weight loss, the chromium
concentration of the alloy underneath the oxide
scale was about 11%. The data appear to suggest
pg 50
31/10/2007 12:46PM Plate # 0
pg 51
that the critical chromium concentration of the

alloy underneath the oxide scale is approximately 11% for alloy 800HT. When the chromium concentration underneath the oxide scale
is below 11%, the reformation of a protective
chromium-rich oxide scale is not possible, thus
resulting in breakaway oxidation.
In the same oxidation test program at 1150 C
(2100 F), similar analysis on the chromium
concentration prole underneath the oxide
scale was performed for alloys 230, HR160
and HR120 with the data presented in
Fig. 3.62 and 3.63 (Ref 86). Figure 3.62 shows
the weight-loss data for three alloys up to
360 days. The chromium concentration prole
from the surface immediately underneath the
230
0.0
HR120
100.0
200.0
800HT
300.0
0
100
(a)
200
Time, days
300
400
spalled oxide scale to the alloy interior after 360

days of testing was determined using EDX analysis with the results shown in Fig. 3.63. The
chromium concentration immediately underneath the spalled oxide scale for both alloys 230
and HR120 was well below 10%, while that of
alloy HR160 was about 10%. Alloy HR120
showed sign of breakaway oxidation after
90 days of exposure. Continuing oxidation testing resulted in a linear weight-loss rate. Alloy
230 showed signs of breakaway oxidation after
240 days. HR160, however, showed a linear
weight loss up to 360 days, although suffering
the least weight-loss rate, with no clear sign of
breakaway oxidation. Oxidation of alloy HR160
is involved the formation of Cr2O3 and SiO2.
This may explain that HR160, although continuing to lose weight, still showed no sign of
breakaway oxidation after 360 days at 1150 C
(2100 F). In a long-term oxidation study of
Fe-20Cr-25Ni alloy in CO2 containing 1% CO,
300 ppm H2O, and 300 ppm CH4 at 1023 to
1173 K (750 to 900 C), Evans et al. (Ref 87)
found the critical chromium level for rehealing of
chromium oxide scales to be about 16%. The
minimum level of chromium needed to maintain
a protective chromium oxide scale to prevent
breakaway oxidation may vary from environment to environment and from alloy to alloy.
To prolong the time for the initiation of
breakaway oxidation, it is necessary to have an
adequate reservoir for chromium immediately
below the oxide scale to provide adequate chromium to maintain a protective chromium oxide
28.0
Surface Cr concentration, wt%
HR160
24.0
HR120
20.0
230
16.0
12.0
250.0
230
500.0
HR120
750.0
800HT
8.0
(b)
100
200
300
400
1000.0
0
Time, days
60
120
180
240
300
360
Time, days
Fig. 3.61
Weight changes as a function of exposure time in

long-term cyclic oxidation tests in air at 982 C
(1800 F) for alloys 800HT, HR120, and 230 (a), and the corresponding changes in the surface chromium concentration (measured after the scale was spalled off) as a function of exposure time
(b). Source: Ref 84
Fig. 3.62
Weight changes as a function of exposure time for

alloys 230, HR160, and HR120 in air oxidation
tests at 1150 C (2100 F) with thermal cycling to room temperature for weight measurement once every 30 days. Source:
Ref 86
31/10/2007 12:46PM Plate # 0
80.0
70.0
Ni
Concentration, wt%
60.0
50.0
40.0
30.0
Cr
20.0
W
10.0
0
0
200
400
600
800
1000
1200
Distance from surface, m
(a)
50.0
Ni
Concentration, wt%
40.0
Co
30.0
20.0
Cr
10.0
0
0
400
800
1200
1600
2000
(b)
60.0
Concentration, wt%
50.0
Ni
40.0
30.0
Fe
20.0
Cr
10.0
0
0
(c)
Fig. 3.63
200
400
600
800
1000
Concentration profiles for (a) alloy 230, (b) HR160,

and (c) HR120 after air oxidation tests for 360 days
at 1150 C (2100 F), as shown in Fig. 3.62. Source: Ref 86
scale or to reheal the oxide scale that suffered

local cracking or failure. Brady et al. (Ref 88)
proposed that chromium carbides might provide
a reservoir of chromium for maintaining the
growth and rehealing of chromium oxide scales.
These authors observed that an as-cast Fe-15Cr0.5C formed a nonprotective iron-rich oxide
scale when exposed in O2 at 850 C (1562 F).
When the alloy was forged at 1150 C (2100 F)
to produce a uniformly distributed ne carbide
phase, the forged alloy showed a thin protective
Cr2O3 scale under the same test condition (Ref
88). During oxidation testing of this forged
sample, the ne chromium-rich carbides are
dissolved into the underlying alloy substrate to
continue supplying chromium to maintain and
reheal the chromium oxide scales (Ref 88).
For alumina formers, such as Fe-Cr-Al alloys,
and Fe-Cr-Al-base and Ni-Cr-Al-base ODS
alloys, breakaway oxidation occurs when aluminum concentration underneath the Al2O3 scale
has reduced to a critical level such that healing of
the Al2O3 is no longer possible, thus resulting in
the formation of nonprotective, fast-growing
oxides of base metals (e.g., iron oxides or nickel
oxides). The breakaway oxidation due to rapid
growth of iron oxides or nickel oxides becomes
essentially a life-limiting factor. This critical
aluminum concentration was found to be about
1.0 to 1.3% for Fe-Cr-Al-base ODS alloys (e.g.,
MA956, ODM751) at 1100 to 1200 C (2012
to 2192 F) (Ref 89, 90). These values were
obtained from foil specimens (0.2 to 2 mm thick)
tested in still air at 1100 to 1200 C. For the nonODS Fe-20Cr-5Al alloy, this critical aluminum
concentration was found to be higher (about
2.5%) at 1200 C (Ref 89). Since the breakaway
oxidation is related to aluminum reservoir in the
alloy, and the aluminum reservoir becomes a
critical issue when the component is made of thin
sheet or foil. Because of excellent oxidation
resistance at very high temperatures, there is
increasing interest in looking at alumina formers
for products that require thin foils, such as honeycomb seals in gas turbines, metallic substrates
for automobile catalyst converters, and recuperators in microturbines. The oxidation behavior
of several commercial alumina formers in thin
foils is summarized in Section 3.4.14.
For alumina formers to improve their resistance to breakaway oxidation, yttrium is frequently used to increase the adhesion of the
aluminum oxide scale. Other alloying elements
that are known to increase the adhesion of the
aluminum oxide scale include zirconium and
pg 52
31/10/2007 12:46PM Plate # 0
pg 53
hafnium. Quadakkers (Ref 91) shows that both

MA956 (Fe-20Cr-4.5Al-0.5Y2O3) and Aluchrom (Fe-20Cr-5Al-0.01Y) exhibited much
more cyclic oxidation resistance than Fe-20Cr5Al when tested at 1100 C in synthetic air with a
hourly cycle to room temperature (Fig. 3.64).
Addition of Y2O3 to an alumina former has a
similar benecial effect as yttrium alloying element. Klower and Li (Ref 92) studied the oxidation resistance of Fe-20Cr-5Al alloys in 10
different compositions containing various
amounts of yttrium ranging from 0.045 to 0.28%.
All 10 compositions contained 0.002% S, and
eight compositions contained 0.04 to 0.06% Zr
with two compositions containing no zirconium.
Cyclic oxidation tests were performed at 1100
and 1200 C (2012 and 2192 F), respectively,
with each cycle consisting of 96 h at temperature
and rapid air cooling to room temperature. These
authors concluded that the yttrium addition
of about 0.045% was sufcient to prevent the
oxide scales from spalling and when the yttrium
concentration was increased to more than
0.08%, substantial internal oxidation could
occur, resulting in rapid metal wastage, as shown
in Fig. 3.65 (Ref 92).
Sulfur in the alloy is known to play a very
signicant role in the adhesion of the aluminum
oxide scale to the alloy substrate for alumina
formers. The role of yttrium is believed to prevent the preferential segregation of sulfur in the
alloy to the scale/metal interface to weaken the
adhesion of the oxide scale (Ref 9395). Reducing the concentration of sulfur in a Ni-Cr-Al
alloy can signicantly improve the oxidation
resistance of the alloy. Smeggil (Ref 96) compared cyclic oxidation resistance between the
normal purity Ni-Cr-Al alloys (approximately 30

to 40 ppm S) with the high-purity Ni-Cr-Al
alloys (approximately 1 to 2 ppm S), showing a
signicant improvement in cyclic oxidation
resistance when sulfur in the alloy was signicantly reduced. This is illustrated in Fig. 3.66
(Ref 96). Also demonstrated in the gure is the
benecial effect of yttrium addition to the normal
purity Ni-20Cr-12Al alloy, showing signicant
improvement in the cyclic oxidation resistance of
the alloy without reducing the sulfur content in
the alloy. Sulfur has been found to segregate to
the oxide/alloy interface during oxidation in FeCr-Al alloys (Ref 97, 98). The role of yttrium is
believed to tie up sulfur at the oxide/metal
interface, thus improving the oxide-scale adhesion (Ref 96).
3.4.14 Thin Foils
There are some industrial applications that
require thin-gage sheet materials or thin foils
for construction of some critical components.
As the component thickness decreases, oxidation
becomes a major limiting factor for its service
life. When the component is made of thin foil,
prolonging the incubation time before the
initiation of breakaway oxidation is the controlling factor for extending the service life of the
component. Thus, as applications are being
pushed toward higher and higher temperatures,
alloys that form aluminum oxide scales can offer
tremendous advantages in performance over
those alloys that form chromium oxide scales.
In gas turbine applications, one important
component made of a thin foil is a turbine seal
ring assembly that controls the turbine tip
Depth of internal oxidation, m
4
Fe-20Cr-AI
MA956
Aluchrom
3
2
1
0
1
0
200
400
600
800
1000
1200
800
1100 C
1200 C
600
400
200
0
0
Fig. 3.64
Cyclic oxidation resistance of MA956, Aluchrom,

and Fe-20Cr-5Al tested in synthetic air at 1100 C
(2012 F) with an hourly cycle (each cycle consisted of 56 min
heating and 4 min cooling). Source: Ref 91
0.1
0.2
0.3
Yttrium, wt%
Time, h
Fig. 3.65
Maximum internal oxidation depth as a function of

yttrium content in the alloys after 3000 h of cyclic
oxidation tests at 1100 and 1200 C (2012 and 2192 F) in air.
Source: Ref 92
31/10/2007 12:46PM Plate # 0
pg 54
Mass change/unit area, mg/cm2
High-purity NiCrAI
8
Normal-purity
NiCrAIY
16
24
Normal-purity
NiCrAI
32
0
20
40
60
80
100
Number of 1 h cycles
Fig. 3.66
Cyclic oxidation resistance of the normal purity Ni-20Cr-12Al (30 to 40 ppm S), the high-purity Ni-20Cr-12Al (1 to 2 ppm S)
and the normal purity Ni-20Cr-12Al-Y at 1180 C. Source: Ref 96
clearance for improving thermal efciency.

The seal ring assembly is typically constructed
out of a honeycomb seal brazed onto a superalloy casing. The traditional alloys used for
honeycomb seals are chromia formers, such as
nickel-base alloy X. Lai (Ref 99) evaluated
honeycomb samples made of wrought alloy 214
(Ni-16Cr-3Fe-4.5Al-Y) and alloy X (Ni-22Cr18.5Fe-9Mo) in dynamic burner rig testing that
simulated gas turbine hot gas conditions. Tests
were conducted at 954 C (1750 F) in a highvelocity combustion gas stream (0.3 Mach or
100 m/s) with cycling every 30 min. Both alloy
X and alloy 214 honeycomb samples were
made of 0.076 mm (3 mils) foils. The alloy X
honeycomb sample was completely oxidized
(destroyed) after 154 h of testing, while the
alloy 214 honeycomb sample was unaffected
after 317 h when the test was terminated.
Figure 3.67 shows the condition of both samples
after tests.
Simms et al. (Ref 100) conducted extensive
oxidation studies on several commercial foil
materials with different thicknesses (from 0.05 to
0.127 mm) in a simulated combustion environment (nominally N2-14%O2-3%CO2-7.8%H2O)
at 950 to 1250 C (1742 to 2282 F). Test specimens were cycled to room temperature every
100 h when tested at 950 and 1050 C (1742 and
1922 F), every 40 h at 1150 C (2102 F), every
20 h at 1250 C (2280 F). Each specimen was
contained in an individual alumina crucible so

the spalled oxides could be included in the
weight measurement. All the specimens were
preoxidized in air at 1050 C (1922 F) for 1 h
prior to oxidation testing. The alloys tested
included Kanthal AF (Fe-20Cr-5Al-0.05Y0.08Zr wrought alloy), Aluchrom YHf (Fe-20Cr5.8Al-0.04Y-0.05Zr-<0.1Hf wrought alloy),
PM2000 (Fe-20Cr-5.5Al-0.5Y2O3 ODS alloy),
and Haynes 214 alloy (Ni-3Fe-16Cr-4.5Al0.01Y wrought alloy). The 0.127 mm thick
Haynes 214 foil specimens were the only foils
exhibiting no breakaway oxidation over the test
duration of 300 h (15 cycles) at 1250 C
(2282 F). Kanthal AF, Aluchrom YHf and
PM2000 foil specimens with 0.127 mm thick
suffered breakaway oxidation well before completing 300 h (15 cycles) test duration. Thinner
foils suffered breakaway oxidation in shorter
times than thicker foils. As the test temperature
was lowered, the life of the foil was progressively
longer. At 950 C, tests were conducted only on
alloys 214 and PM2000. Both alloys exhibited
no breakaway oxidation for up to 1500 h (15
cycles) for all the thicknesses (0.05 to 0.127 mm)
tested.
Klower (Ref 101) studied the effect of the
foil thickness varying from 0.049 to 0.25 mm
(2 to 10 mils) on breakaway oxidation of Fe20Cr-5Al alloys, which contained rare-earth
metals (mischmetal, such as Ce), in air at
31/10/2007 12:46PM Plate # 0
(a)
(b)
Fig. 3.67 Honeycomb samples after testing at 950 C (1750 F) for 154 h for the alloy X honeycomb sample (a) and 317 h for the 214
honeycomb sample (b) in a high-velocity combustion gas stream (0.3 Mach or 100 m/s) generated by a dynamic burner rig.
The samples were also subjected to rapid quenching from the test temperature to less than 260 C (500 F) for 2 min every 30 min. Both
honeycomb samples were made of 0.076 mm (3 mil) foils. Courtesy of Haynes International, Inc.
1100 C (2012 F). The specimens were cooled
to room temperature for weighing every 96 h.
The time to breakaway oxidation increased
with increasing foil thickness, as illustrated in
Fig. 3.68 (Ref 101).
The incubation time for breakaway oxidation
is dependent on the amount of aluminum in the
reservoir for alumina formers. The total amount
of aluminum in the reservoir of the alloy
increases with increasing foil thickness. Thus, it
takes much longer for a thicker foil to reduce the
concentration of aluminum below the critical
level such that rehealing of the protective oxide
scale is not possible, resulting in breakaway
oxidation.
Thin foils of oxidation-resistant alloys

including stainless steels, Fe-Ni-Cr alloys and
nickel-base alloys have been extensively evaluated for high-temperature recuperators in
microturbines (Ref 102104). The use of a
recuperator for preheating the incoming air for
combustion can signicantly increase the efciency of the microturbine. The oxidation tests
were carried out on foils with about 100 m
(0.1 mm, or 4 mils) thick in air containing 10%
H2O, which was to simulate the level of H2O that
would be present in the exhaust gas stream in
microturbines. It was found that air containing
10% H2O was signicantly more aggressive than
dry air (i.e., laboratory air) (Ref 102104). The
pg 55
31/10/2007 12:46PM Plate # 0
60
0.072
50
Mass change, g/m2
Foil thickness, mm
0.25
0.165
0.125
0.091
0.072
0.058
0.05
0.049
0.058
40
0.05
0.091
30
0.049
20
10
0
0
500
1000
1500
2000
2500
3000
3500
Exposure time, h
Fig. 3.68
Weight gain as a function of exposure time in air at 1100 C (2012 F) for Fe-20Cr-5Al foils with various thicknesses. The
alloy also contained 0.015% Mischmetal. Source: Ref 101
effect of water vapor on oxidation behavior of

alloys is discussed in next section.
Pint (Ref 104) reported that Type 347 foil
(100 m, or 4 mils, thick) suffered accelerated
oxidation (or breakaway oxidation) after less
than 2000 h of exposure at 650 C (1200 F) in
air containing 10% H2O. The foil was also subjected to thermal cycling every 100 h. Under the
same test condition, alloy HR120, and alloy 625
showed no accelerated oxidation (or breakaway
oxidation) after more than 8000 h. Both alloys
HR120 and 625 exhibited a slight mass loss
of about less than 0.2 mg/cm2 after 8000 h of
exposure. The author attributed this small mass
loss to volatilization of CrO2(OH)2 that formed
during exposure of the water vapor in the environment (Ref 104). Both alloys HR120 and 625
showed that the oxidation behavior at 700 C
(1292 F) was similar to that at 650 C
(1200 F), exhibiting very little mass loss after
more than 8000 h (Ref 104). At 800 C
(1472 F), alloy HR120 showed no sign of
accelerated oxidation after 7500 h. Alloy 625
again showed excellent oxidation resistance with
little mass loss at 800 C (1472 F) for times up
to 6000 h when the test was terminated (Ref
104). Both HR120 and 625 exhibited excellent
oxidation resistance in air containing 10% H2O
at 650 to 800 C (1200 to 1472 F). However,
Type 347 performed poorly at 650 C (1200 F)
in air containing 10% H2O. Nevertheless, when
Type 347 foil was tested at 650 C (1200 F) in
laboratory air (i.e. dry air) for up to 40,000 h, the
alloy showed a thin oxide scale with no sign of
accelerated oxidation (Ref 105). The effect of

water vapor on oxidation of alloys is discussed in
the next section.
3.4.15 Effect of H2O on Oxidation
In high-temperature combustion atmospheres,
water vapor is invariably present in the environment. In some cases, the level of water vapor
in the environment can be signicant. The effect
of water vapor on the oxidation of alloys is thus
an important factor in the alloy selection process.
Most oxidation data are generated in laboratory
air, which generally contain low levels of water
vapor.
Tuck et al. (Ref 106) investigated the effect of
water vapor on oxidation of iron in air and air
with 5%, 10%, 15%, and 20% H2O at 800 and
1000 C (1472 and 1832 F) for times up to
200 min. Figure 3.69 shows the weight changes
as a function of time at 800 C (1472 F), indicating essentially no effects on oxidation by the
presence of water vapor up to 20% (Ref 106).
Also, no water vapor effects were observed when
tested at 1000 C (1832 F). The tests, however,
were conducted at such high temperatures that
carbon steels would never be used. In boilers
burning fuels containing large amount of moisture (e.g., 15 to 40% in coal) (Ref 107), a large
amount of H2O is expected to be present in the
combustion atmosphere in a coal-red boiler.
The ue gas analysis from a utility pulverized
coal-red boiler showed approximately 10%
H2O (Ref 108). Carbon and low-alloy steels,
pg 56
31/10/2007 12:46PM Plate # 0
pg 57
40
Weight gain, mg/cm2
30
20
10
Air 20% steam
Air 15% steam
Air 10% steam
Air 5% steam
Dry air
0
50
100
150
200
250
Time, min
Fig. 3.69
Effect of water vapor on oxidation of iron in air and air with 5%, 10%, 15%, and 20% H2O at 800 C (1472 F) for times up to
200 min. Source: Ref 106
which are used for the construction of the boiler

waterwalls in the combustion zone, have
performed well under normal oxidizing conditions. More detailed discussion on oxidation
and high-temperature corrosion of carbon and
low-alloy steels in coal-red boilers is presented
in Chapter 10.
Segerdahl et al. (Ref 109, 110) studied the
oxidation behavior of 11Cr steel (X20
11Cr1MoV) in O2 and O2 with 10% and 40%
H2O at 600 C (1112 F). In tests at 600 C
(1112 F) for 700 h at a gas ow rate of 0.5 cm/s,
the steel showed no weight changes for 100% O2
and O2-10H2O, while it suffered breakaway
oxidation after 336 h when tested in O2-40%
H2O (Ref 109). The steel exposed to dry O2
exhibited a protective (Cr,Fe)2O3 scale after
testing, while the steel showed hematite in the
outer layer and FeCr spinel in the inner layer
when exposed to O2-40H2O mixtures (Ref 109).
When tests were conducted at 600 C (1112 F)
for 168 h with test gas velocity at 0.5, 1.0, 2.5, 5,
and 10 cm/s, the steel showed no weight changes
in 100% O2 and O2-10% H2O at all gas velocities, while it suffered accelerated weight gain
at 5 and 10 cm/s when tested in O2-40% H2O
(Ref 110).
Asteman et al. (Ref 9) studied oxidation
behavior of Type 304L at 873 K (600 C) in O2
and O2 with 10% H2O. In the O2 environment,

the weight of the alloy increased initially with
increasing time, which was then followed by a
decrease in the rate of increase with further
exposure time, indicating the formation of a
protective oxide scale. As for the environment
consisting of O2 with 10% H2O, the weight
of the alloy initially increased and then decreased
with increasing exposure time. The authors
observed a brownish layer of deposits in the
exhaust end of the furnace tube downstream from
the test specimen. The analysis of the deposits
revealed a high concentration of chromium.
They proposed the weight loss in the O2-10%
H2O environment was due to the evaporation
of CrO2(OH)2 that formed on the alloy as a
result of water vapor. The author showed that
the theoretical partial pressure of CrO2(OH)2
in the O2-10%H2O environment was quite
high at low temperatures, as shown in Fig. 3.4
(Ref 9).
The effect of water vapor on the oxidation
behavior of Type 321 at 800 C (1472 F) was
illustrated in Fig. 3.70, showing Type 321 foil
suffered severe oxidation attack in air containing
10% H2O compared with the sample tested in dry
air (laboratory air) (Ref 111). The gure also
shows a severe oxidation attack for Type 347
foil tested at 800 C (1472 F) in air with 10%
31/10/2007 12:46PM Plate # 0
10
347, 800 C 10% H2O
9
8
321, 800 C 10% H2O
Mass change, mg/cm2
7
6
321, 700 C 10% H2O
5
4
3
2
347, 700 C 10% H2O
321, 800 C dry air
0
0
100
200
300
400
500
600
700
800
900
1000
Time, h
Fig. 3.70
Oxidation behavior of Type 321 and 347 foils (100 m thick) at 700 and 800 C (1292 and 1472 F) in dry air and air with
10% H2O. Specimens were cycled to room temperature every 100 h. Source: Ref 111
H2O. The detrimental effect of water vapor in

air on the oxidation resistance of Type 347 is
clearly illustrated in Fig. 3.71 (Ref 112). Water
vapor has been found to be detrimental to
the oxidation resistance of Type 310 (Fig. 3.72)
(Ref 114), alloy X (Fig. 3.73) (Ref 114), and
CMSX-4 (Fig. 3.74) (Ref 115). Both Type
310 and alloy X are chromia formers, while
CMSX-4 (Ni-9.5Co-6.4Cr-5.7Al-6.3W-6.5Ta2.9Re) is an alumina former. Onal et al. (Ref 116)
investigated the cyclic oxidation for a number of
alumina-formers cyclic oxidation in both dry air
and air containing 30% H2O at temperatures
from 700 to 1000 C (1292 F to 1832 F). The
authors proposed that water vapor reduced the
adhesion of aluminum oxide scale, thus causing
oxide spallation (Ref 116).
3.4.16 Intermetallic Compounds
Tremendous interest has been generated
among academic researchers in intermetallic
materials and their structures, properties, and
high-temperature corrosion behaviors. However,
industrial applications of these materials are still
limited. Limited data are presented here to briey
compare some promising nickel aluminides

(Ni3Al) or nickel aluminide-base materials with
some commercial nickel-base alloys. Readers are
referred to Ref 117 for more information about
the oxidation and corrosion of various types of
intermetallic alloys.
Comparison oxidation tests were performed
in air at 1150 and 1200 C (2100 and 2200 F)
between two promising Ni3Al-base nickel
aluminides, IC-50 and IC-218 (developed by
Oak Ridge National Laboratory), and nickelbase alloys, alloy 214 (alumina former) and
alloys X and 230 (chromia formers) (Ref
55). Test results for IC-50 (Ni-11.3Al-0.6Zr0.02B), IC-218 (Ni-8.5Al-7.8Cr-0.8Zr-0.02B),
alloy 214 (Ni-4.5Al-16Cr-3Fe-0.01Y), alloy X
(Ni-22Cr-18Fe-9Mo-0.6W), and alloy 230
(Ni-22Cr-14W-2Mo-0.02La) are summarized
in Table 3.25. Test results on other alloys
under the same test conditions are shown in
Table 3.12.
Samples of IC-218 were completely turned
into oxides after 1008 h at 1150 and 1200 C
(2100 and 2200 F), although they still maintained the sample shape. Their specimen weight
gain data did not reveal the complete oxidation of
the specimen. IC-50 suffered internal oxidation
pg 58
31/10/2007 12:46PM Plate # 0
pg 59
347 foil, 40,000 h, air
50 m
Cu-plating
(a)
Cu-plating
347 foil, 40,000 h, air
10 m
(b)
347 foil, 10,000 h, air+10% H2O
50 m
(c)
Fig. 3.71
Optical micrographs showing a thin oxide scale formed on Type 347 foil when tested in laboratory air after 40,000 h at
650 C (1200 F) (a) and (b) and thick unprotective oxide scales formed on Type 347 foil when tested in air containing 10%
H2O after only 10,000 h at the same temperature (c). Source: Ref 112. Courtesy of Oak Ridge National Laboratory
10
50
100
0.01% H2O
5% H2O
10% H2O
150
200
400
600
Time, h
Fig. 3.72
800
1000
Descaled
Effect of water vapor in air on the oxidation resistance of Type 310 at 1100 C (2010 F) cycled
every 100 h. Source: Ref 113
that penetrated through the specimen thickness,

as illustrated in Fig. 3.75. Again, the weight
change data failed to reveal the severity of the
oxidation attack for IC-50. This example
emphasizes the fact that specimen weight gain
(or weight change) data can be misleading.
Metallographic examination of the test specimens is also required to determine the total depth
of attack, which included both metal loss and
internal penetration.
Comparing IC-50 and IC-218 with alloy 214,
it is interesting to note that both nickel aluminides, with signicantly higher aluminum contents (twice of that in alloy 214), failed to
produce a protective aluminum oxide scale.
SEM/EDX analyses of the scales formed on both
nickel aluminide samples after testing at 1150 C
(2100 F) revealed both nickel- and aluminumrich oxides for IC-50 and mostly nickel-rich
oxides for IC-218 (Fig. 3.76) (Ref 55). It is
believed that chromium is needed to increase
aluminum activities in the alloy for forming a
continuous aluminum oxide scale. In his studies
on oxidation of IC-50 in air at 1000 C
(1832 F), Natesan (Ref 118) found that NiO was
31/10/2007 12:46PM Plate # 0
pg 60
10
+1
CMSX4 in dry air

1
m/A, mg/cm2
W/A, mg/cm2
0.01% H20
5% H20
50
10% H20
CMSX4
in wet air
4
5
6
7
8
100
0
200
400
600
800
Time, h
1000
Descaled
Fig. 3.73
Oxidation of alloy X at 1100 C (2010 F) in air

containing various amounts of H2O. Cycling to
room temperature every 100 h for weighing. Source: Ref 114
100
200
300
400
Time, h
Fig. 3.74
Cyclic oxidation behavior of CMSX4 (Ni-9.5Co6.4Cr-5.7Al-6.3W-6.5Ta-2.9Re) in dry air and air

containing 0.1 atm H2O (10% H2O in 1.0 atm test gas), cycling
once an hour with 45 min at 1100 C (2012 F) and 15 min
cooling. Source: Ref 115
Table 3.25 Results of oxidation tests on nickel aluminides and nickel-base alloys in air for 1008 h at
indicated temperatures
1150 C (2100 F)
1200 C (2200 F)
Material
Weight gain,
mg/cm2
Metal loss,
mm (mils)
Average metal
affected(a),
mm (mils)
IC-50
IC-218
214
230
X
2.7
16.7
1.2
33.5
68.5
0.008 (0.3)
>0.36 (14)(c)
0.005 (0.2)
0.058 (2.3)
0.11 (4.5)
>0.38 (15)(b)
>0.36 (14)(c)
0.0075 (0.3)
0.086 (3.4)
0.147 (5.8)
Weight gain,
mg/cm2
Metal loss,
mm (mils)
Average metal
affected,
mm (mils)
8.8
61.5
1.3
81.2
169(d)
0.02 (0.8)
>0.36 (14)(c)
0.005 (0.2)
0.11 (4.5)
>0.9 (35.4)(e)
>0.38 (15)(b)
>0.36 (14)(c)
0.018 (0.7)
0.20 (7.9)
>0.9 (35.4)(e)
Note: Flowing air 30 cm/min (472 cm3 min in a 1.75 in. diam tube); cycled to room temperature once a week (168 h cycles). (a) Metal loss + average internal penetration.
(b) Internal penetration through thickness. (c) Sample was completely oxidized. (d) Specimen weight gain after 504 h. (e) Extrapolated from 504 h; specimen was
completely oxidized (consumed) in 504 h. Source: Ref 55
the only oxide formed on the sample; no Al2O3

was detected.
Burner rig dynamic oxidation tests under a
high-velocity combustion gas stream (0.3 Mach
or 100 m/s) were also performed on IC-50
compared with commercial alloys (Ref 55).
The nickel aluminide IC-50 suffered signicantly more severe oxidation attack than
nickel-base alloys 214 (alumina former) and 230
(chromia former) after testing at 1090 C
(2000 F) for 500 h with 30 min cycles. IC-50
suffered more than 0.38 mm (15 mils) of oxidation attack, compared with 0.05 mm (1.8 mils)
for alloy 214 and 0.14 (5.7 mils) for alloy 230.
The data for other commercial alloys tested
under the same conditions are shown in Table
3.22. The nickel aluminide was very susceptible
to internal oxidation. Signicant internal oxidation was observed after only 50 h of testing.
SEM/EDX analysis of the scale formed on the
50 h tested sample revealed mainly nickel-rich

oxides.
3.4.17 Catastrophic Oxidation
As temperature increases, metals and alloys
generally suffer increasingly higher rates of
oxidation. When the temperature is excessively
high, metals and alloys can suffer rapid oxidation. There is, however, another mode of rapid
oxidation that takes place at relatively lower
temperatures. This mode of rapid oxidation,
which is often referred to as catastrophic oxidation, is associated with the formation of a
liquid oxide. The liquid oxide disrupts and dissolves the protective oxide scale, causing the
alloy to suffer rapid oxidation at relatively low
temperatures. Leslie and Fontana (Ref 119)
observed an unusually rapid oxidation for Fe25Ni-16Cr alloy containing 6% Mo when heated
31/10/2007 12:46PM Plate # 0
pg 61
Original
Sample Thickness
33.3 m
Fig. 3.76
Nickel-rich oxides formed on nickel aluminide IC218 after 1008 h at 1150 C (2100 F) in air with
168 h cycles. Area 1: 11.7% Al, 80.0% Ni, 8.3% Cr. Area 2:
18.8% Al, 49.0% Ni, 32.2% Cr. Area 3: 22.7% Al, 57.1% Ni,
20.2% Cr. Source: Ref 55
200 m
Fig. 3.75
Specimen cross section of IC-50 after oxidation

testing in air at 1150 C (2100 F) for 1008 h with
168 h cycles. Samples were descaled prior to metallographic
mounting. Source: Ref 55
in static air to 900 C (1650 F). The same

alloy exhibited good oxidation resistance when
heated to the same temperature in owing
air. They postulated that the rapid oxidation
was due to the accumulation of gaseous MoO3
on the metal surface. The oxidation is accelerated
by the thermal dissociation of MoO3 into MoO2
and O. Meijering and Rathenau (Ref 120),
Brasunas and Grant (Ref 121), and Brennor (Ref
122), however, attributed this to the presence
of a liquid oxide phase. The MoO3 oxide
melts at about 795 C (1463 F). The 19Cr-9Ni
steel suffered catastrophic oxidation attack in
the presence of MoO3 at 770 C (1420 F).
This temperature was very close to the eutectic
temperature of MoO2-MoO3-Cr2O3, which was
reported to be 772 C (1420 F) (Ref 123).
Other mixed oxides involving MoO3 that exhibit

low melting points are MoO2-MoO3 (778 C,
or 1435 F), MoO2-MoO3-NiO (764 C, or
1400 F), Fe2O3-MoO3 (730 C, or 1345 F),
Fe2O3-Fe-MoO3 (725 C, or 1335 F), V2O5MoO3 (610 to 718 C, or 1130 to 1330 F),
Na2O-MoO3 (499 C, or 930 F), and Cu2OMoO2-MoO3 (470 C, or 880 F) (Ref 123).
Other oxides, such as PbO and V2O5, can also
cause metals or alloys to suffer catastrophic
oxidation in air at intermediate temperatures
of 640 to 930 C (1200 to 1700 F) (Ref 124).
PbO and V2O5 melt at 888 and 690 C (1630
and 1270 F), respectively. The deleterious
effect of lead oxide was believed to be related to
exhaust-valve failures in gasoline engines.
Gasoline additives were a primary source for lead
compounds. Vanadium is an important contaminant in residual or heavy fuel oils. Therefore,
V2O5 plays a signicant role in oil ash
corrosion in oil-red boilers, which is discussed
in Chapter 11.
Sawyer (Ref 124) indicated that accelerated
oxidation of Type 446 stainless steel in the
presence of lead oxide can proceed at temperatures where the liquid phase does not exist.
Experiments carried out by Brasunas and Grant
(Ref 125) showed that 16Cr-25Ni-6Mo alloy
specimens placed adjacent to, but not in contact
with, 0.5 g samples of WO3 oxides suffered
accelerated oxidation attack when tested in
air at 868 C (1585 F), which is well below
the melting point of WO3 (i.e., 1473 C, or
2683 F).
31/10/2007 12:46PM Plate # 0
Because molybdenum and tungsten are very

important solid-solution-strengthening alloying
elements, many superalloys containing either
or both have been developed since Leslie and
Fontana rst observed catastrophic oxidation
in 1948. Some of these alloys, including alloys
X, R-41, 625, 617, S, 230, 25, and 188, have
been used successfully in service for the hot
section of gas turbine engines. Many of them
have also been used successfully in heat treating,
chemical processing, and related industries. The
most effective way to alleviate the potential
catastrophic oxidation problem is to avoid a
stagnant condition for the gaseous atmosphere.
3.5 Oxidation/Nitridation in Air and

Nitridation can take place in conjunction with
oxidation in oxidizing atmospheres including air
and combustion environments. This can result in
signicant internal nitridation penetration and
affect the mechanical properties of the alloy. This
topic is covered in detail in Chapter 4 Nitridation.
3.6 Summary
Oxidation data for carbon and low-alloy steels,
ferritic stainless steels, austenitic stainless steels,
Fe-Cr-Ni alloys with 20-25Cr/30-40Ni, Fe-Cr-Al
alloys, and superalloys including ODS alloys are
presented. The data are presented in such a way
for readers to make comparisons between alloys
within the same alloy group or between alloys
from different alloy groups. The chapter focuses
on long-term oxidation behavior. Oxidation
resistance of some nickel aluminides is also
compared with that of several commercial
nickel-base superalloys.
A large portion of commercial hightemperature alloys relies on chromium for
forming a protective chromium oxide scale to
resist oxidation attack. The temperature range in
which the chromium oxide scale is effective in
providing adequate oxidation resistance varies
from 540 to 1090 C (1000 to 2000 F). This
group of alloys is frequently referred to as
chromia formers. An adequate supply of chromium from the alloy interior to the metal surface
to form and maintain a continuous, protective
chromium oxide scale is necessary for an alloy to
maintain its oxidation-resistant capability under
the service condition. Once the chromium supply

drops below the critical level to maintain and
reheal the chromium oxide scale, breakaway
oxidation is initiated followed by rapid growth of
oxides of base metals, such as iron, nickel, or
cobalt oxides. Detailed discussion on the effect
of chromium on the oxidation behavior and
breakaway oxidation of various alloys is presented. Effects of other minor elements on the
oxidation behavior of chromia formers are also
discussed.
A relatively small number of commercial
high-temperature alloys rely on aluminum for
forming a protective aluminum oxide scale to
resist oxidation attack at very high temperatures.
This group of alloys that form aluminum oxide
scales are typically referred to as alumina
formers. The oxidation behavior of alumina
formers that include Fe-Cr-Al, Ni-Cr-Al, and
ODS alloys is presented. The effect of aluminum
as well as yttrium and sulfur on the oxidation
resistance of alumina formers is discussed.
Discussion also includes oxidation under
high-velocity gas streams, oxidation of thin foils,
effect of surface depletion of chromium in austenitic stainless steels, effect of water vapor, and
catastrophic oxidation (oxidation under molten
oxides). Most oxidation data were generated at
980 to 1200 C (1800 to 2200 F). However,
many industrial applications are in the temperature range of 650 to 980 C (1200 to 1800 F),
which are below the test temperatures at which
most data were generated. More long-term oxidation data need to be generated at 650 to 980 C
(1200 to 1800 F) for stainless steels, Fe-Ni-Cr
and some simple Ni-Cr alloys to provide a more
reliable database at intended application temperatures.
REFERENCES
1. G.Y. Lai, unpublished data, 2004

2. N. Birks and G.H. Meier, Introduction to
High Temperature Oxidation of Metals,
Edward Arnold Ltd., London, 1983
3. R.C. Weast, Ed., Physical Constants of
Inorganic Compounds, Handbook of
Chemistry and Physics, 65th ed., The
Chemical Rubber Company, 1984, p B68
B161
4. C.E. Ramberg, P. Beatrice, K. Kurokawa,
and W.L. Worrell, Mater. Res. Soc. Proc.,
Vol 322, C.L. Briant, J.J. Petrovic, B.P.
Bewlay, A.K. Vasudevan, and H.H. Lipsitt,
pg 62
31/10/2007 12:46PM Plate # 0
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
Ed., Materials Research Society, 1994,

p 243
D. Caplan and M. Cohen, The Volatilization of Chromium Oxide, J. Electrochem.
Soc., Vol 108 (No. 5), 1961, p 438
E.A. Gulbranson and S.A. Jansson, in
Heterogenous Kinetics at Elevated Temperatures, G.R. Belton and W.L. Warrell,
Ed., Plenum Press, 1970, p 181
H.C. Graham and H.H. Davis, Oxidation/
Vaporization Kinetics of Cr2O3, J. Am.
Ceram. Soc., Vol 54 (No. 2), 1971, p 89
C.A. Stearns, F.J. Kohl, and G.C.
Fryburg, Oxidative Vaporization Kinetics
of Cr2O3 in Oxygen from 1000 to 1300 C,
J. Electrochem. Soc.: Solid-State Sci.
Technol., Vol 121 (No. 7), 1974, p 945
H. Asteman, J.E. Svensson, L.G. Johansson, and M. Norell, Indication of Chromium Oxide Hydroxide Evaporation
During Oxidation of 304L at 873 K in the
Presence of 10% Water Vapor, Oxid. Met.,
Vol 52 (No. 1/2), 1999, p 95
N. Birks and F.S. Pettit, Environmental
Effects During Application of Materials at
Temperatures Above 1200 C, Mater. Sci.
Eng., A143, 1991, p 187
M. Danielewski, Kinetics of Gaseous
Corrosion Processes, in Corrosion:
Fundamentals, Testing and Protection,
Vol 13A, ASM Handbook, ASM International, 2003, p 97
W.R. Patterson, in Designing for Automotive Corrosion Prevention, Proceedings
P-78 (Troy, MI), Nov 810, 1978, Society
of Automotive Engineers, p 71
J.B. Vrable, R.T. Jones, and E.H. Phelps,
The Application of Corrosion-Resistant
High-Strength Low-Alloy Steels in the
Chemical Industry, presented at Fall
Meeting, American Society for Metals,
Chicago, 1977
R.T. Jones, in Process Industry Corrosion,
NACE, 1986, p 373
R.C. John, Compilation and Use of
Corrosion Data for Alloys in Various HighTemperature Gases, Paper No. 73, Corrosion/99, NACE International, 1999
A.W. Zeuthen, Heating, Piping and Air
Conditioning, Vol 42 (No. 1), 1970, p 152
H.E. McGarrow, Ed., The Making, Shaping
and Treating of Steel, United States Steel
Corp., 1971, p 1136
I.G. Wright, in Corrosion, Vol 13, Metals
Handbook, ASM International, 1987, p 97
19. A. Grodner, Weld. Res. Counc. Bull.,

No. 31, 1956
20. M. Walter, M. Schutze, and A. Rahmel,
Behavior of Oxide Scales on 12Cr-1Mo
Steel During Thermal Cycling, Oxid. Met.,
Vol 39 (No. 5/6), 1993, p 389
21. S.B. Lasday, Ind. Heat., March 1979, p 12
22. O.D. Sherby, Acta Metall., Vol 10, 1962,
p 135
23. H.E. Eiselstein and E.N. Skinner, STP 165,
ASTM, 1954, p 162
24. A. Moccari and S.I. Ali, Br. Corros. J.,
Vol 14 (No. 2), 1979, p 91
25. A.S. Brasunas, J.T. Gow, and O.E. Harder,
Proc. ASTM, Vol 46, 1946, p 870
26. S. Kado, T. Yamazaki, M. Yamazaki,
K. Yoshida, K. Yabe, and H. Kobayashi,
Trans. Iron Steel Inst. Jpn., Vol 18 (No. 7),
1978, p 387
27. H.T. Michels, Met. Eng. Q., Aug 1974, p 23
28. J. Douthett, Heat Treating of Stainless
Steels, in Heat Treating, Vol 4, ASM
Handbook, ASM International, 1991,
p 769
29. J.F. Grubb, in Proc. International Conf. on
Stainless Steels, Iron and Steel Institute of
Japan, Tokyo, 1991, p 944
30. G.Y. Lai, Unpublished data, 2005
31. W.E. Ruther and S. Greenberg, J. Electrochem. Soc., Vol 111, 1964, p 1116
32. J.C. Kelly, Todays Versus Yesterdays
Stainless Steels, Mater. Perform., March
1992
33. Heat Resistant Alloy Specications and
Operating Data, Bulletin 150, Rolled
Alloys, Temperance, MI, 1997
34. RA 353MA alloy, Bulletin No. 1069,
Rolled Alloys, Temperance, MI
35. Avesta 353MA, Information 9270,
Avesta AB, S-77480 Avesta, Sweden
36. G.R. Rundell, Evaluation of A High Silicon-Aluminum Alloy for Waste Incineration and Fluidized Bed Combustion
Systems, Performance of High Temperature Materials in Fluidized Bed Combustion Systems and Process Industries, P.
Ganesan and R.A. Bradley, Ed., ASM
International, 1987, p 193
37. Kanthal Electrical Resistance and High
Temperature Alloys, Kanthal Alloys Data
Sheet, Sandvik, Sweden
38. R. Berglund and B. Jonsson, New Powder
Metallurgical Fe-Cr-Al Alloy for High
Temperature and Corrosion Resistance in
Furnace Uses, Ind. Heat., Oct 1989, p 21
pg 63
31/10/2007 12:46PM Plate # 0
39. Allegheny Ludlum Data Sheet, Allegheny Ludlum, Pittsburgh, PA, 1990
40. P.T. Moseley, K.R. Hyde, B.A. Bellamy,
and G. Tappin, The Microstructure of the
Scale Formed During the High Temperature Oxidation of a Fecralloy Steel,
Corros. Sci., Vol 24 (No. 6), 1984, p 547
41. E. Tsuzi, The Role of Yttrium on the Oxide
Adherence of Fe-24Cr Base Alloys, Metall.
Trans. A, Vol 11A, Dec 1980, p 1965
42. T.A. Ramanarayanan, M. Raghavan,
and R. Petkovic-Luton, Metallic Yttrium
Additions to High Temperature Alloys:
Inuence on Al2O3 Scale Properties, Oxid.
Met., Vol 22 (No. 3/4), 1984, p 83
43. J.L. Pandey, S. Prakash, and M.L. Mehta,
Effect of Zirconium Concentration on High
Temperature Cyclic Oxidation Behavior of
Fe-15Cr-4Al at 1150 C, J. Electrochem.
Soc.: Solid-State Sci. Technol., Vol 135
(No. 1), Jan 1988, p 209
44. J.L. Pandey, S. Prakash, and M.L. Mehta,
Effects of Varying the Zirconium Concentration and 1 wt.% Y on High Temperature Oxidation of Fe-15wt.%Cr-4wt.%
Al Alloy under Isothermal and Cyclic
Conditions, J. Less-Common Met., Vol
159, 1990, p 23
45. J.S. Benjamin, Metall. Trans., Vol 1, 1970,
p 2943
46. M. Lundberg, L.-P. Bergmark, and M.
Ramberg, Mechanical and Chemical
Properties of 353MAA Seamless Tube
for High-Temperature Petrochemical
Applications, 1999 Stainless Steel World
Conf. Proc., Book 2, KCI Publishing BV,
Zutphen, The Netherlands, 1999, p 563
47. J.C. Kelly and J.D. Wilson, Oxidation
Rates of Some Heat Resistant Alloys,
Heat-Resistant Materials II: Conf. Proc.
Second International Conference on HeatResistant Materials, K. Natesan, P. Ganesan, and G. Lai, Ed., ASM International,
1995, p 53
48. P. Ganesan, G.D. Smith and C.S. Tassen,
Mechanical Properties and Corrosion
Resistance of Incoloy Alloy 803, Applications and Materials Performance: Proc.
Nickel-Cobalt 97 International Symposium,
F.N. Smith, J.F. McGurn, G.Y. Lai, and
V.S. Sastri, Ed., The Metallurgical Society
of CIM, Montreal, Canada, 1997, p 97
49. M.A. Harper, J.E. Barnes, and G.Y. Lai,
Paper No. 132, Corrosion/97, NACE
International, 1997
50. Inconel Alloy 601 Brochure, Huntington

Alloys, Inc., 1969
51. P. Ganesan, G.D. Smith, and C.S. Tassen,
Performance of A New Alloy in High
Temperature Service, Paper No. 234, Corrosion/93, NACE, 1993
52. D.C. Agarwal, ThyssenKrupp VDM
unpublished data
53. D.C. Agarwal and U. Brill, Performance of
Alloy 602CA (UNS N06025) in High
Temperature Environments up to 1200 C,
International, 2000
54. G.Y. Lai, J. Met., Vol 37 (No. 7), July 1985,
p 14
55. G.Y. Lai, unpublished results, Haynes
International, Inc., 1988
56. Haynes Alloy No. 214, H-3008B, Haynes
International, Inc. Kokomo, IN
57. N. Birks and F.S. Pettit, Environmental
Effects During Application of Materials at
Temperatures above 1200 C, Mater. Sci.
Eng., Vol A143, 1991, p 187
58. G.Y. Lai, Suldation-Resistant Co-Cr-Ni
Alloys with Critical Contents of Silicon and
Cobalt, U.S. Patent No. 4711763, Dec 1987
59. G.Y. Lai, Meeting the Challenge of Materials Development for Coal Combustion
Plants, Mater. High Temp., Vol 11
(No. 14), 1993, p 143
60. W. Crawford, in Proc. Conf. Frontiers
of High Temperature Materials II, London,
Inco Alloys International, May 1983, p 272
61. R.F. Singer, in Proc. Conf. Frontiers of
High Temperature Materials II, London,
Inco Alloys International, May 1983, p 336
62. Superalloys Source Book, M.J. Donachie,
Jr., Ed., American Society for Metals, 1984
63. C.T. Sims, in Proc. Fifth International
Symposium on Superalloys (Seven
Springs, Champion, PA), Metallurgical
Society of AIME, 1984, p 399
64. W. Betteridge and W.W.K. Shaw, Mater.
Sci. Technol., Vol 3, 1987, p 682
65. B.H. Kear and E.R. Thompson, Science,
Vol 208, May 23, 1980, p 847
66. M.J. Donachie and S.J. Donachie, Superalloys: A Technical Guide, 2nd ed., ASM
International, 2002
67. C.A. Barrett, in Proc. Conf. Environmental
Degradation of Engineering Materials,
M.R. Louthan, Jr. and R.P. McNitt, Ed.,
Virginia Polytechnic Institute, 1977, p 319
68. M.F. Rothman, Cabot Corporation internal
report, 1985
pg 64
31/10/2007 12:46PM Plate # 0
69. H. Nagai, M. Okaboyashi, and H. Mitani,

Trans. Jpn. Inst. Met., Vol 21, 1980, p 341
70. G.M. Kim, E.A. Gulbranson, and G.H.
Meier, in Proc. Conf. Fossil Energy
Materials Program, ORNL/FMP 87/4,
May 1921, 1987, R.R. Judkins, Ed., Oak
Ridge National Laboratory, 1987, p 343
72. R.H. Kane, J.W. Schultz, H.T. Michels,
R.L. McCarron, and F.R. Mazzotta, Ref 30
of the paper by R.H. Kane in Process
Industries Corrosion, B.J. Moniz and W.I.
Pollock, Ed., NACE, 1986, p 45
73. G.Y. Lai, M.F. Rothman, and D.E. Fluck,
Paper No. 14, Corrosion/85, NACE, 1985
74. J.J. Barnes and S.K. Srivastava, Paper No.
527, Corrosion/89, NACE, 1989
75. J.J. deBarbadillo and J.J. Fischer, Dispersion-Strengthened Nickel-Base and IronBase Alloys, in Properties and Selection:
Nonferrous Alloys and Special-Purpose
Materials, Vol 2, Metals Handbook, ASM
76. R.H. Kane, G.M. McColvin, T.J. Kelly, and
J.M. Davidson, Paper No. 12, Corrosion/
84, NACE, 1984
77. R.C. John, Oxidation Studies of Commercial Alloys at 8711093 C (1600
2000 F), in Heat-Resistant Materials
IIConf. Proc. Second International
Conference on Heat-Resistant Materials,
K. Natesan, P. Ganesan, and G. Lai, Ed.,
ASM International, 1995, p 41.
78. C.E. Lowell, D.L. Deadmore, and J.D.
Whittenberger, Long-Term High-Velocity
Oxidation and Hot Corrosion Testing of
Several NiCrAl and FeCrAl Base Oxide
Dispersion Strengthened Alloys, Oxid.
Met., Vol 17 (No. 3/4), 1982, p 205
79. M.F. Rothman, Oxidation Resistance of
Gas Turbine Combustion Materials,
Paper No. 85-GT-10, presented at the Gas
Turbine Conference (Houston, TX), March
1821, 1985, ASME, 1985
80. J.V. Wright, The Effects of Gas Velocity
and of Temperature on the Oxidative
Response of Selected Sheet Superalloys,
Paper No. 88-GT-281, presented at the
Gas Turbine and Aeroengine Congress
(Amsterdam, The Netherlands), June 69,
1988, ASME, 1988
81. U. Brill and T.I. Haubold, Corrosion
Behaviour of Some Gas Turbine Alloys
under High Velocity Burnt Fuels, Paper
82.
83.
84.
85.
86.
87.
88.
89.
90.
91.
92.
93.
94.
95.
96.
No. 522, Corrosion/2000, NACE International, 2000

G.Y. Lai, unpublished results, Haynes
B. Hicks, Mater. Sci. Technol., Vol 3,
Sept 1987, p 772
B. Gleeson, High-Temperature Corrosion
of Metallic Alloys and Coatings, in Corrosion and Environmental Degradation,
Vol II, Materials Science and Technology,
M. Schutze, Ed., Wiley-VCH, Weinheim,
Germany, 2000, p 173
B. Gleeson and M.A. Harper, The LongTerm, Cyclic-Oxidation Behavior of
Selected Chromia-Forming Alloys, Oxid.
Met., Vol 49 (No. 3/4), 1998, p 373
H.E. Evans, D.A. Hilton, R.A. Holm, and
S.J. Webster, Oxid. Met., Vol 14, 1980,
p 235
M.P. Brady, B. Gleeson, and I.G. Wright,
Alloy Design Strategies for Promoting
Protective Oxide-Scale Formation, JOM,
Jan 2000, p 16
W.J. Quadakkers and K. Bongartz, The
Prediction of Breakaway Oxidation for
Alumina forming ODS Alloys Using Oxidation Diagrams, Werkst. Korros., Vol 45,
1994, p 232
I. Gurrappa, S. Weinruch, D. Naumenko,
and W.J. Quadakkers, Factors Governing
Breakaway Oxidation of FeCrAl-Based
Alloys, Mater. Corros., Vol 51, 2000, p 224
W.J. Quadakkers, Growth Mechanisms
of Oxide Scales on ODS Alloys in
the Temperature Range 10001100 C,
Werkst. Korros., Vol 41, 1990, p 659
J. Klower and J.G. Li, Effects of Yttrium on
the Oxidation Behavior of Iron-ChromiumAluminum Alloys, Mater. Corros., Vol 47,
1996, p 545
J.G. Smeggil, A.W. Funkenbusch, and
N.S. Bornstein, High Temp. Sci., Vol 20,
1985, p 163
A.W. Funkenbusch, J.G. Smeggil, and
N.S. Bornstein, Met. Trans. A, Vol 16,
1985, p 1164
J.G. Smeggil, A.W. Funkenbusch, and
N.S. Bornstein, Met. Trans. A, Vol 17,
1986, p 923
J.G. Smeggil, Some Comments on the
Role of Yttrium in Protective Oxide Scale
Adherence, Mater. Sci. Eng., Vol 87, 1987,
p 261
pg 65
31/10/2007 12:46PM Plate # 0
97. P.Y. Hou and J. Stringer, Oxide Scale

Adhesion and Impurity Segregation at the
Scale/Metal Interface, Oxid. Met., Vol 38
(No. 5/6), 1992, p 323
98. P.Y. Hou, Compositions at Al2O3/
FeCrAl Interfaces after High Temperature
Oxidation, Mater. Corros., Vol 51, 2000,
p 329
99. G.Y. Lai, Several Modern Wrought
Superalloys for Gas Turbine Applications,
Paper 96-TA-030, presented at ASME
Turbo Asia 96 (Jakarta, Indonesia),
Nov 57, 1996
100. N.J. Simms, R. Newton, J.F. Norton, A.
Encinas-Oropesa, J.E. Oakey, J.R.
Nicholls, and J. Wilber, Mater. High Temp.,
Vol 20 (No. 3), 2003, p 439
101. J. Klower, Factors Affecting the Oxidation
Behaviour of Thin Fe-Cr-Al Foils, Mater.
Corros., Vol 49, 1998, p 758
102. P.J. Maziasz, B.A. Pint, R.W. Swindeman,
K.L. More, and E. Lara-Curzio, Advanced
Stainless Steels and Alloys for High
Temperature Recuperators, DOE/CETC/
CANDRA Workshop on Microturbine
Applications (Calgary, Alberta, Canada),
Jan 2123, 2003
103. B.A. Pint, The Effect of Water Vapor on
Cr Depletion in Advanced Recuperator
Alloys, GT2005-68495, ASME Turbo
Expo 2005 (Reno-Tahoe, Nevada),
June 69, 2005
104. B.A. Pint, Stainless Steels with Improved
Oxidation Resistance for Recuperators,
J. Eng. Gas Turbines Power, Vol 128,
2006, p 1
105. B.A. Pint, The Effect of Water Vapor on
Cr Depletion in Advanced Recuperator
Alloys, GT2005-68495, ASME Turbo
Expo 2005, (Reno-Tahoe, Nevada), June
69, 2005
106. C.W. Tuck, M. Odgers, and K. Sachs,
Scaling Rates of Pure Iron and Mild Steel in
Oxygen, Steam, Carbon Dioxide in the
Range 8501000 C, Anti-Corrosion,
June 1966, p 14
107. S.C. Stultz and J.B. Kitto, Ed., Steam and
Its Generation and Use, 40th ed., Babcock
& Wilcox, 1992, p 13-2
108. S.C. Stultz and J.B. Kitto, Ed., Steam and
Its Generation and Use, 40th ed., Babcock
& Wilcox, 1992, p T-17
109. K. Segerdahl, J.E. Svensson, and
L.G. Johansson, The High Temperature
Oxidation of 11% Chromium Steel: Part
110.
111.
112.
113.
114.
115.
116.
117.
118.
119.
120.
121.
122.
123.
124.
125.
IInuence of pH2 O , Mater. Corros., Vol

53, 2002 p 247
K. Segerdahl, J.E. Svensson, and L.G.
Johansson, The High Temperature Oxidation of 11% Chromium Steel: Part II
Inuence of Flow Rate, Mater. Corros.,
Vol 53, 2002, p 479
B.A. Pint, R. Peraldi, and P.F. Tortorelli,
The Effect of Alloy Composition on the
Performance of Stainless Steels in Exhaust
Gas Environments, Paper No. 03499,
Corrosion/2003, NACE International,
2003
B.A. Pint, private communication, 2007
R.L. McCarron and J.W. Schultz, in Proc.
Symp. High Temperature Gas-Metal
Reactions in Mixed Environments, AIME,
1973, p 360
C.C. Clark and W.R. Hulsizer, Superalloys
Development for Gas Turbines Operating
in the Marine Environment, Conf. Proc.,
Gas Turbine Materials Conference, Naval
Ship Engineering Center, 1972, p 35
C. Sarioglu et al., The Adhesion of
Alumina Films to Metallic Alloys and
Coatings, Mater. Corros., Vol 51, 2000,
p 358
K. Onal, M.C. Maris-Sida, G.H. Meier, and
F.S. Pettit, Water Vapor Effects on the
Cyclic Oxidation Resistance of Alumina
Forming Alloys, Mater. High Temp., Vol 20
(No. 3), 2003, p 327
G. Welsch and P.D. Desai, Ed., Oxidation
and Corrosion of Intermetallic Alloys,
Purdue University, 1996
K. Natesan, Oxid. Met., Vol 30 (No. 1/2),
1988, p 53
W.C. Leslie and M.C. Fontana, Paper
No. 26, 30th Annual Convention of
ASM (Philadelphia, PA), Oct 2529,
1948
J.K. Meijering and G.W. Rathenau, Nature,
Vol 165, Feb 11, 1950, p 240
A.D. Brasunas and N.J. Grant, Iron Age,
Aug 17, 1950, p 85
S.S. Brennor, J. Electrochem. Soc., Vol 102
(No. 1), Jan 1955, p 16
J.H. DeVan, Catastrophic Oxidation of
High Temperature Alloys, ORNL-TM-51,
Oak Ridge National Laboratory, Oak
Ridge, TNn, Nov 10, 1961
J.W. Sawyer, Trans. TMS-AIME, Vol 221,
1961, p 63
A. de S. Brasunas and N.J. Grant, Trans.
ASM, Vol 44, 1950, p 1133
pg 66

DOI: 10.1361/hcma2007p067
30/10/2007 3:55PM Plate # 0
pg 67

CHAPTER 4
Nitridation
4.1 Introduction
In air or combustion atmospheres containing
nitrogen, nitridation can take place under certain
exposure conditions. In most cases, oxidation
dominates the high-temperature corrosion reaction. However, nitridation can take place for
some alloys when oxide scales no longer provide
protection. The alloys that are particularly susceptible to oxidation/nitridation attack are those
containing strong nitride formers, such as titanium and aluminum. Many high-temperature
nickel-base alloys containing both aluminum and
titanium are strengthened by phase, Ni3(Al,Ti).
For these alloys, nitridation by forming internal
nitrides of aluminum and titanium can deplete the
surface layer with aluminum and titanium, thus
weakening the alloy. Under a high-velocity combustion gas stream with severe thermal cycling,
similar to the conditions in ying gas turbines
(aircraft engines), nitridation can be particularly
severe in oxidation/nitridation attack.
In nitrogen-base atmospheres, such as N2 or
N2-H2, metals and alloys can also suffer nitridation attack. This type of atmosphere is often
used as a protective atmosphere in heat treating
and sintering operations. Molecular nitrogen can
be severely nitriding for many metals and alloys,
particularly when temperatures are sufciently
high.
Ammonia (NH3) is a commonly used nitriding
gas for case hardening at temperatures from
500 to 590 C (925 to 1100 F) (Ref 1). Furnace
equipment and components repeatedly subjected
to these service conditions frequently suffer
brittle failures as a result of nitridation attack.
Carbonitriding is another important method of
case hardening that produces a surface layer of
both nitrides and carbides. The process is typically carried out at 700 to 900 C (1300 to 1650
F) in ammonia, with additions of carbonaceous
gases, such as CH4 (Ref 2). Thus, the heat treat
retort, xtures, and other furnace equipment are
subject to both nitridation and carburization.
Cracked ammonia (i.e., ammonia that is completely dissociated into H2 and N2) provides an
economical protective atmosphere for processing
metals and alloys. Many bright annealing operations for stainless steels use a protective atmosphere consisting of N2 and H2, generated by
dissociation of ammonia. With three parts H2
and one part N2 produced in cracked ammonia,
nitridation is less critical for the heat treating
equipment.
In the chemical processing industry, nitriding
environments are generated by processes employed for production of ammonia, nitric acid,
melamine, and nylon 6-6 (Ref 3, 4). Ammonia is
produced by reacting nitrogen with hydrogen
over a catalyst at temperatures of typically 500
to 550 C (930 to 1020 F) and pressures of 200
and 400 atm. Commercial processes for ammonia synthesis are discussed in detail in Ref 5. The
converter, where the ammonia synthesis reaction
takes place, may suffer nitridation attack. Brittle
failure of the welds for the waste heat boiler of
an ammonia plant has been reported by Van der
Horst (Ref 6). These tube-to-tube sheet welds
were high-nickel alloy 182, which suffered
nitridation attack.
Production of nitric acid involves the oxidation of ammonia over a platinum gauze catalyst
at temperatures of about 900 C (1650 F)
(Ref 5). The catalyst grid support structure and
other processing components in contact with
ammonia may also be susceptible to nitridation
attack. Figure 4.1 shows the nitrided structure of
a nickel-base alloy catalyst grid support after
two years of service in a nitric acid plant.

When metal is exposed to nitrogen gas at
elevated temperatures, nitridation proceeds
according to:
1=2N2 (gas)=[N] (dissolved in metal)
4:1
30/10/2007 3:56PM Plate # 0
pg 68
[%N] = k(pN2 )1=2
4:2
where k is the equilibrium constant and pN2 is the

partial pressure of N2 in the atmosphere.
In nitrogen-base atmospheres, the nitriding
potential is proportional to (pN2 )1=2 . Increasing
the nitrogen partial pressure (or nitrogen concentration) increases the thermodynamic potential for nitridation. Molecular nitrogen is less
aggressive than ammonia in terms of nitridation
of metals. However, when metal is heated to
excessively high temperatures (e.g., 1000 C
(1830 F) or higher), nitridation by molecular
nitrogen can become a serious material issue.
Under oxidation/nitridation conditions, nitrogen molecules permeate through cracks and
pores and reach the metal underneath the oxide
scales when the oxide scale is no longer protective. Nitridation then proceeds by dissociation of
100 m
nitrogen molecules and absorption of nitrogen

atoms by the metal following Reaction 4.1.
Nitridation often takes place in the metal at the
vicinity of cracks developed under creep conditions in air or N2-containing combustion atmospheres. In this case, oxides are often associated
with the crack under creep deformation in air.
Oxidation consumes the oxygen molecules from
air, which penetrates into the crack, thus depleting oxygen and increasing nitrogen potential
(or concentration) around the crack. As a result,
nitridation takes place in the vicinity of the creep
crack.
When the environment is ammonia (NH3) or
contains NH3, metals or alloys may undergo rapid
nitridation reactions. It is precisely for this reason
that NH3 is frequently used for case hardening.
Ammonia is metastable and dissociates into
molecular N2 and molecular H2 when heated to
elevated temperatures. Once NH3 is completely
dissociated into N2 and H2, the nitriding potential
is dened by Reaction 4.3. To increase nitrogen
absorption by steel, molecular NH3 should be
allowed to dissociate on the steel surface, thus
allowing dissociated atomic nitrogen to be dissolved at the metal surface (Ref 7, 8). Thus, to
increase nitridation reactions, it is necessary to
bring as much fresh, uncracked NH3 as possible
in contact with the surface of the metal to minimize the production of molecular nitrogen.
At temperatures below 600 C (1110 F) and at
high gas ow rates, the production of nitrogen
is minimized and the nitrogen solubility at the
surface of iron is determined by (Ref 8):
(a)
100 m
(b)
Fig. 4.1
Alloy X (Ni-22Cr-18.5Fe-9Mo-0.6W) catalyst grid

support structure bar after 2 years of service in a nitric
acid plant. (a) Internal nitride precipitates (about 20 mils in depth)
containing mainly chromium-rich nitrides along with some carbides formed during thermal aging. (b) Microstructure in the
unaffected interior containing mainly carbides due to thermal
aging at the service temperature
NH3 $ 3=2 H2 +[N] (dissolved in Fe)
4:3
[%N]=k(pNH3 =(pH2 ))3=2
4:4
where k is the equilibrium constant, and pNH3 ,

and pH2 are partial pressures of NH3 and H2,
respectively.
The nitriding potential is proportional to
pNH3 =( pH2 )3=2 . Increasing ammonia partial
pressure (or concentration) in the atmosphere
increases the thermodynamic potential for nitridation. When nitrogen in the metal exceeds its
solubility limit, nitrides will then precipitate
out. The nitrides of important alloying elements
for engineering alloys are tabulated in Table 4.1
(Ref 9). For iron, nickel, and cobalt, three important alloy bases for high-temperature alloys,
only iron forms stable nitrides. No nitrides of
nickel and cobalt have been reported. The
relative stabilities among various nitrides can
30/10/2007 3:56PM Plate # 0
pg 69
Chapter 4: Nitridation / 69
alloys with nickel ranging from 0 to 35%

(Ref 19). In Fe-18Cr-Ni-N system, increasing
nickel reduces the solubility of nitrogen, as
shown in Fig. 4.3 (Ref 11). It is also shown in
Table 4.2 that the solubility of nitrogen in stainless steels increased with increasing temperature.
Alloys with higher nitrogen solubilities generally
exhibit less resistance to nitridation attack.
e 4N
+10
2F
N
H
N
o2
10
2M
2
2C
2
1/
40
Iron
Chromium
Molybdenum
Tungsten
Aluminum
Titanium
Niobium
Tantalum
Zirconium
Hafnium
Silicon
Vanadium
Boron
Manganese
Magnesium
Fe4N
CrN
MoN
WN
AlN
TiN
NbN
TaN
ZrN
HfN
Si3N4
VN
BN
Mn4N
Mg3N
bN
60
2
3N
2
70
VN
N2
AI
N
2
Ta
100
N
120
Ti
130
Zr
2M + N2 = 2MN
140
Nb4N3
Ta3N5
500
1000
Hf4N3
Fig. 4.2
Mn3N
102
N2 partial pressure, bar
Nickel, %
20
+ Cr2N
10
+
0.1
0.2
0.3
1000 C
(1830 F)
+ CrN
1
+ Cr2N
0.1
10 2
10 3
10 4
0
+ + Cr2N
2000
Standard free energy of formation for selected

nitrides. Source: Ref 10
10
1500
Temperature, C
Source: Ref 9
BN
g3
90
V 2N
Mn2N
110
Hf3N2
Fe2N
Cr2N
Mo2N
W 2N
Ti2N
Nb2N
Ta2N
N4
Si 3
50
Table 4.1 Nitrides of important alloying

elements for engineering alloys
Nitrides
Cr
r 2N
30
80
Element
20
G, Kcal
be compared in terms of their free energies of

formation, as illustrated in Fig. 4.2 (Ref 10).
Physical-chemical properties of some of these
nitrides can be found in Ref 9. The types of
nitrides that are likely to form in the alloy can
be predicted by examining the phase-stability
diagram. Figure 4.3 shows a phase-stability
diagram of Fe-18Cr-Ni-N system at 900 C
(1650 F), indicating phase regions of Cr2N
in (or +) phase as a function of nickel
and nitrogen contents (Ref 11). Nitride phases
formed in alloys are also dependent on nitrogen
partial pressure ( pN2 ), as shown in Fig. 4.4 for
Ni-Cr-N system at 1000 C (1830 F) (Ref 12).
Nitrogen solubility in the alloy is important in
affecting the nitridation resistance of the alloy.
Table 4.2 summarizes some nitrogen solubility
data for iron, stainless steels, and nickel alloy
(Ref 1318). Iron and stainless steels exhibit
signicantly higher nitrogen solubility than a
nickel-base alloy (Ni-20Fe). Nickel has been
found to decrease the nitrogen solubility in Fe-Ni
0.4
+
10
20
30
40
50
60
Cr concentration, wt %
Nitrogen, %
Fig. 4.3
Phase stability diagram for Fe-18Cr-Ni-N system at

900 C (1650 F). Source: Ref 11
Fig. 4.4
Phase stability diagram for the Ni-Cr-N system as a

function of N2 partial pressure at 1000 C (1830 F).
Source: Ref 12
30/10/2007 3:56PM Plate # 0
Table 4.2
Nitrogen solubility in metals and alloys
Nitrogen, wt%
0.06
0.26 max
0.02 (at pN2 : 1 atm)
0.125
0.177
0.190
0.258
0.281
0.18
0.18
0.21
0.26
0.26
0.0001 (at pN2 : 1 atm)
Metal or alloy
-Fe
Fe
Fe-10%Ni
Type 304
Type 304
Type 304
Type 304
Type 304
Fe-18Cr-12Ni-2Ti
Fe-18Cr-12Ni-2Ti
Fe-18Cr-12Ni-2Ti
Fe-18Cr-12Ni-2Ti
Fe-18Cr-12Ni-2Ti
Ni-20%Fe
Temperature,
C (F)
502 (936)
region of Fe-C
1000 (1832)
538 (1000)
593 (1100)
927 (1700)
954 (1749)
981 (1800)
985 (1805)
1040 (1905)
1093 (2000)
1150 (2100)
1210 (2210)
1000 (1832)
Ref
13
14
15
16
17
17
17
17
18
18
18
18
18
15
4.3 Internal Nitridation in Oxidizing

Environments
In air or oxidizing combustion environments,
oxidation usually dominates high-temperature
corrosion reactions. However, under certain
conditions, alloys can suffer internal nitridation
attack along with oxidation. Internal nitridation
attack, when it occurs, can penetrate farther
into the metal interior than oxidation, thus signicantly affecting the creep-rupture behavior
of the alloy by accelerating the creep crack
growth. Discussion of internal nitridation under
no external stresses and under creep conditions
is presented in sections 4.3.1 and 4.3.2.
4.3.1 Internal Nitridation in Air under
No External Stress
Some high-temperature alloys that contain
strong nitride formers such as aluminum and
titanium can suffer internal nitridation even in air
environments. In an air oxidation study for two
nickel-base alloys, IN939 (Ni-22Cr-20Co-3.8Ti1.4Al-2W-1Nb-1.3Ta) and IN738LC (Ni-16Cr9Co-3.5Ti-3.3Al-1.8Mo-1Nb-1.8Ta) at 700,
900, and 1100 C (1290, 1650, and 2010 F).
Litz et al. (Ref 20) observed internal titanium
nitrides (needle shape) formed in front of internal
aluminum oxides that formed underneath the
external oxide scales. In an oxidation study of
alloy 800HT (Fe-21Cr-32Ni-0.5Al-0.5Ti) in air
at 980 C (1800 F) for about 2 years (720 days),
Harper et al. (Ref 21) observed Widmansttten
acicular chromium-rich nitrides along with
aluminum nitrides that formed below chromiumrich oxides. Lai (Ref 22) observed internal
aluminum nitrides (needle shape) that formed
underneath the external oxide scale and internal

oxides in alloy 601 exposed to a furnace oxidizing atmosphere for about 4 to 5 years at
temperatures probably between 760 and 870 C
(1400 and 1600 F), as shown in Fig. 4.5.
Severe thermal cycling that causes cracking
and spalling of oxide scales can also result in
severe internal nitridation. Han and Young (Ref
23) conducted cyclic oxidation tests by heating
the specimens to 1100 C (2010 F) in still air
for 1 h followed by cooling to room temperature
for 15 min then repeating the cycle again for 260
cycles. The alloys investigated were Ni-24 to
38%Cr-14 to 25%Al. The specimens suffered
severe oxide scale spallation. The internal nitridation attack was found to be extensive, and the
nitridation zone consisted of AlN beneath Cr2O3
and Al2O3, then AlN+Cr2N, and then AlN in the
deepest region (Ref 23).
Douglas (Ref 24) indicated that the diffusivity
of nitrogen appears to be two orders of magnitude greater than that of oxygen in nickel or
nickel alloys. Table 4.3 summarizes the diffusion
coefcients of nitrogen in nickel and iron alloys
compared with those of oxygen and carbon in
nickel, based on the diffusivity data from Rubly
and Douglas (Ref 25, 26), Grabke and Peterson
(Ref 27), Park and Alstetter (Ref 28), and Gruzin
et al. (Ref 29). The diffusivity of nitrogen is also
on the same order of magnitude as that of carbon
as shown in Table 4.3. It is thus not surprising to
nd internal nitrides were advancing in front of
internal oxides.
4.3.2 Internal Nitridation at Creep Cracks
in Air Environment
During creep testing in air, extensive internal
nitridation can develop in the vicinity of cracks.
Brickner et al. (Ref 30) found that types 302,
304, and 310 stainless steels showed signicant
nitridation after creep-rupture testing in air at
870 C (1600 F) in less than 1000 h. Acicular
nitrides (believed to be chromium nitrides) in
a Widmansttten pattern were found to form
extensively in the vicinity of microcracks, as
shown in Fig. 4.6 (Ref 30). Extensive nitridation
was conrmed by the chemical analysis of the
tested specimens for nitrogen, which showed
the nitrogen content was increased from about
0.058% before testing to 0.30 to 0.53% after
creep-rupture testing (Ref 30). Extensive internal
nitrides were also observed in the vicinity of
creep cracks in alloy 253MA (Fe-21Cr-11Ni)
pg 70
30/10/2007 3:56PM Plate # 0
pg 71
(a)
0.1 mm
1
2
(c)
10 m
(b)
0.0010 in
Formation of internal aluminum nitrides beneath external oxide scales and internal oxides in alloy 601 after exposing to a
furnace oxidizing atmosphere for approximately 4 to 5 years in a temperature range of 760 to 870 C (1400 to 1600 F). (a)
Optical micrograph showing the external oxide scales and the internal oxides, and then the chromium denuded zone immediately below,
followed by internal nitrides underneath the denuded zone. (b) Optical micrograph at higher magnification showing internal nitrides. (c)
SEM (backscattered electron image) showing internal aluminum nitrides and the EDX analysis of nitrides. Results of the semiquantitative
EDX analysis (at.%) on internal aluminum nitrides are summarized as:
Fig. 4.5
Phase 1
Phase 2
Table 4.3
41.5% Al, 24.7% Ni, 10.6% Cr, 6.8% Fe, 5.5% Ti, and 10.0% N
58.0% Al, 13.1% Ni, 6.1% Cr, 3.6% Fe, 0.8% Ti, and 17.3% N
Diffusion coefficients of nitrogen, oxygen, and carbon.

Diffusion coefcient, cm2/s
Temperature, C (F)
700 (1290)
800 (1470)
900 (1650)
N in Ni-Alloys(a)
N in Fe-20Ni(b)
O in Ni(c)
C in Ni(d)
9.5 109 to 2.3 108

3.2 108 to 8.5 108
1.4 107 to 4.0 107
1.17 108
3.86 108
1.47 107
7.4 1011
5.05 1010
2.3 109
3.19 109
1.47 108
0.55 107
(a) Ref 25, 26. (b) Ref 27. (c) Ref 28. (d) Ref 29
after creep-rupture testing at 900 C (1652 F)

for 11,800 h in air (Ref 31), in 800H after
creep-rupture testing at 900 to 1000 C (1650 to
1830 F) in air (Ref 32), and in alloy 800H
during the creep crack growth testing at 1000 C
(1830 F) (Ref 33). The nitrides identied were
Cr2N in 253MA (Ref 31), Cr2N (major) and CrN
(minor) in 800H creep-ruptured specimens
(Ref 32), and Cr2N and AlN in 800H creep-crack

growth specimens (Ref 33).
Hoffman and Lai (Ref 34) investigated an
alloy 800HT pigtail that suffered cracking after
about 7.5 years of service in a hydrogen reformer.
The pigtail section, which was exposed to air at
approximately 850 C (1565 F) at the outside
of the reformer furnace, was found to show
30/10/2007 3:56PM Plate # 0
pg 72
extensive blocky precipitates along grain

boundaries and acicular precipitates in the matrix
in the vicinity of cracks at the tube outer-diameter
side (exposed to air). Samples were cut from
this pigtail section and solution-annealed in a
furnace for 1 h at 1093, 1149, and 1204 C
(2000, 2100, and 2200 F), respectively. Microstructural examination of these samples indicated
that both blocky, grain-boundary phases and
acicular phases in the matrix remained in the
microstructure and were not put back into solution, suggesting those phases were nitrides
instead of carbides. Also, chemical analysis of
the samples from the pigtail indicated that carbon
content remained about the same as that of the
material before service (about 0.07%), while
nitrogen content was about 0.27 wt% with a
nominal nitrogen content of about 0.02% prior to
service. The process gas in the tube contained
essentially no nitrogen (typically about 0.01%).
Thus, the nitrogen ingress into the tube was primarily from air from the outside diameter side
of the pigtail. Using scanning electron microscopy with energy-dispersive x-ray spectroscopy
(SEM/EDX) analysis, acicular phases were
found to be enriched in aluminum, while blocky
phases were enriched in chromium; the former
was believed to be aluminum nitride and the
latter chromium nitride. Figure 4.7 shows the
acicular aluminum nitrides and blocky chromium
nitrides that remained in the microstructure after
solution annealing at 1150 C (2100 F) for 1 h
for the sample from the straight section of the
pigtail (Ref 34). Figure 4.8 shows extensive
nitride formation in the vicinity of creep cracks in
the bend section of the pigtail from another

hydrogen reformer (Ref 34).
When creep cracks initially develop at the
metal surface during creep testing in air, oxidation occurs at the crack surface including
the crack tip. The oxide scales formed on the
crack surface become nonprotective due to creep
deformation, thus causing the oxygen potential
to decrease signicantly with concurrent increase
in nitrogen potentials at the oxide/metal interface. As a result, nitrogen is absorbed by the
metal and is diffused into the metal in the vicinity
of cracks to form internal nitrides.
100 m
Fig. 4.7
Acicular aluminum nitrides and blocky chromium

nitrides, which formed in the vicinity of the creep
cracks in alloy 800HT pigtail in a hydrogen reformer, were not
dissolved into solution after the sample was resolution annealed at
1150 C (2100 F) for 1 h. Source: Ref 34
100 m
Fig. 4.6 Acicular nitrides (believed to be chromium nitrides)

in a Widmansttten pattern formed in the vicinity of
creep cracks in Type 302SS after creep-rupture testing at 870 C
(1600 F) in less than 1000 h. Original magnification, 500.
Source: Ref 30
Fig. 4.8
Extensive aluminum and chromium nitrides formed

in the vicinity of creep cracks in the bend section
of an alloy 800H pigtail in another hydrogen reformer. Source:
Ref 34
30/10/2007 3:56PM Plate # 0
pg 73
4.3.3 Oxidation and Nitridation in

High-temperature alloys that are exposed to a
high-velocity, oxidizing combustion gas stream
at high temperatures are susceptible to internal
nitridation attack. In investigating transition duct
component failures in a land-based gas turbine,
Swaminathan and Lukezich (Ref 35) observed
that alloy 617 (Ni-22Cr-12.5Co-9Mo-1.2Al) had
suffered severe oxidation and nitridation attack
from both the air side (outside diameter side
of the transition duct) and the combustion side
(inside diameter side of the duct) after service for
slightly less than 2 years (14,000 h). Extensive
internal nitridation from both air and combustion
gas sides of alloy 617 transition duct is shown in
Fig. 4.9 (Ref 35). Alloy 230 (Ni-22Cr-14W2Mo-0.3Al-La) was also tested for 16,000 h as a
transition duct, suffering similar oxidation/nitridation attack (Ref 35). However, no aluminum
nitrides were observed in alloy 230. Signicant
nitrogen pickup was observed from both transition ducts. Results of the chemical analyses of
nitrogen from samples at the exit end of the
transition duct and at the location far away from
the exit for both alloy 617 and 230 transition
ducts are shown in Table 4.4 (Ref 35).
Lai (Ref 34) used a high-velocity dynamic
burner rig test to simulate a gas turbine combustion environment. The simulated combustion
gas stream was generated by burning fuel oil
(a mixture of two parts No. 1 fuel and one
part No. 2 fuel) with an air-to-fuel ratio of approximately 50 to 1 in a laboratory burner rig.
Most of the air for combustion was from a
compressor. When combusted with fuel oil, a
OD
ID
high-velocity combustion gas stream with about

0.3 Mach (100 m/s) was generated. Specimens
were loaded in a carousel specimen holder
that rotated at 30 rpm during testing to ensure all
the specimens were subjected to the same test
conditions. Furthermore, the specimens were
subjected to severe thermal cycling once every
30 min by lowering the carousel from the test
chamber followed by rapid fan-air cooling to
below 260 C (500 F) for 2 min before returning the carousel back to the test chamber. A
schematic of this dynamic burner rig is shown
in Fig. 4.10. The combustion gas was determined
to consist of 76% N2, 13% O2, 6% CO2, and 5%
H2O.
The test on alloy 617 produced severe internal
nitridation, with the microstructure very similar
to that observed by Swaminathan and Lukezich
(Ref 35) from the transition duct in a land-based
gas turbine power plant. Figure 4.11 shows the
microstructure of an alloy 617 specimen after
testing at 980 C (1800 F) for 1000 h with
30 min thermal cycling. Extensive needle-shape
aluminum nitrides were observed. Some blocky
chromium nitrides were observed to form right
below the external oxide scales. Aluminum
nitrides were found to penetrate farther into the
metal interior than chromium nitrides. The oxide
scales were found to be porous and nonprotective.
Alloy 230 was included in the test and found
to show less nitridation attack under the same
test condition. Nitridation in alloy 230 involved
only the formation of internal chromium nitrides
below internal chromium oxides and chromium
denuded zone with no aluminum nitrides. Two
other common combustor alloys, alloys X and
263, were also included in the test. Figures 4.12
and 4.13 show the microstructures of alloys
X and 263, respectively, after testing at 980 C
(1800 F) for 1000 h with 30 min thermal
cycling. Alloy X showed mainly internal chromium nitrides, while alloy 263 showed mainly
tiny needle-shaped nitrides, presumably titanium
Table 4.4 Nitrogen contents at different

locations of alloy 617 and alloy 230 transition
ducts
Fig. 4.9
Cross section (2.5 mm, or 0.1 in.) of an alloy 617 (Ni22Cr-12.5Co-9Mo-1.2Al) transition duct after service
for less than 2 years (about 14,000 h) in a land-based gas turbine,
showing extensive formation of both aluminum and chromium
nitrides from both air side (outside diameter of the transition duct)
and the combustion gas side (inside diameter of the duct). Source:
Ref 35
Transition duct/service
Location
Alloy 617/14,000 h
Exit
Far away from exit
Exit
Far away from exit
Alloy 230/16,000 h
Source: Ref 35
Nitrogen, wt%
0.24
0.004
0.22
0.05
30/10/2007 3:56PM Plate # 0
pg 74
Thermocouple for
steady-state control
Thermocouple for recording

specimen temperature history
50 mm square insulated
flame tunnel
Specimen temperature measured
by pyrometer
Compressed
inlet air 425 C (800 F)
Fuel
Combustor
Rotating shaft
Blower
(thermal shock)
Thermal cycle
Fig. 4.10
The dynamic burner rig used by Lai (Ref 36) for simulating a gas turbine combustion environment in evaluating the
oxidation/nitridation behavior of gas turbine combustor alloys. Courtesy of Haynes International, Inc.
50 m
50 m
Fig. 4.11
Some blocky chromium nitrides and extensive

acicular aluminum nitrides formed in alloy 617
after testing in the dynamic burner rig at 980 C (1800 F) for
1000 h with 30 min thermal cycling. The combustion gas stream
with about mach 0.3 (100 meter/s) consisted of 76% N2, 13% O2,
6% CO2, and 5% H2O. Source: Ref 36
nitrides. Similar to both alloy 617 and 230, alloys

X and 263 also exhibited nonprotective, porous
oxide scales.
Fig. 4.12
Extensive internal chromium nitrides formed in

alloy X after testing in the dynamic burner rig at
980 C (1800 F) for 1000 h with 30 min thermal cycling. The
combustion gas stream with about mach 0.3 (100 m/s) consisted
of 76% N2, 13% O2, 6% CO2, and 5% H2O. Source: Ref 36
X-ray diffraction analysis of the oxide scales

formed on alloys 230, 617, and X was performed
with the results summarized in Table 4.5. The
results showed that NiO and NiCr2O4 along with
30/10/2007 3:56PM Plate # 0
pg 75
Cr2O3 oxides made up the external oxide scales

for alloy 230, and NiO and Cr2O3 made up the
oxide scales for alloys 617 and X. With the formation of NiO oxides, the alloys were no longer
protected by Cr2O3 oxide scales. An electrolytic
extraction technique was used to extract precipitate phases in the tested specimens for analysis,
and the results are summarized in Table 4.5,
revealing essentially nitride phases in all three
alloys. Alloy 230 also showed M6C carbides,
which were the carbides in the alloy in the
as-solution-annealed condition. Bulk nitrogen
contents for the original samples (before testing) and those after testing are summarized in
Table 4.6, showing signicant nitrogen adsorption for alloys 263, 617, and X. Alloy 230
showed little nitrogen adsorption. Bulk carbon
contents for the samples before and after testing
were also determined, and the results clearly
showed that carburization was not involved
(Table 4.7). The overall test results in terms of
weight loss (due to oxidation), metal loss (due
to oxidation), internal oxidation, internal nitridation, and total depth of attack are summarized
in Table 4.8.
In continuing his testing program for the same
simulated gas turbine environment involving the
same four combustor alloys (i.e., 230, 617, 263,
and X) at the same test temperature and duration
Table 4.6 Results of bulk nitrogen analysis before
and after testing at 980 C (1800 F) for 1000 h
with 30 min thermal cycling
Alloy
Original nitrogen, wt%
Nitrogen after testing, wt%
230
263
617
X
0.05
0.004
0.03
0.04
0.06
0.42
0.52
0.57
Source: Ref 36
50 m
Fig. 4.13
Extensive internal nitrides (believed to be titanium

nitrides) formed in alloy 263 after testing in the
dynamic burner rig at 980 C (1800 F) for 1000 h with 30 min
thermal cycling. The combustion gas stream with about mach 0.3
(100 m/s) consisted of 76% N2, 13% O2, 6% CO2, and 5% H2O.
Source: Ref 36
Table 4.5 Results of x-ray diffraction analysis

on oxide scales and extracted precipitate
phases for alloys 230, 617, and X after testing
thermal cycling
Alloy
Surface oxide scales
230
NiCr2O4 (strong)
Cr2O3 (medium)
NiO (medium)
NiO (strong)
Cr2O3 (medium)
NiO (strong)
Cr2O3 (weak)
617
X
Source: Ref 36
Table 4.7 Results of bulk carbon analysis before

and after testing in the dynamic burner rig at
980 C (1800 F) for 1000 h with 30 minute
thermal cycling
Alloy
Original carbon, wt%
Carbon after testing, wt%
230
263
617
X
0.09
0.06
0.05
0.08
0.09
0.03
0.04
0.01
Source: Ref 36
Table 4.8 Test results in terms of weight loss,

depth of oxidation penetration, depth of
nitridation, and total depth of attack after
testing at 980 C (1800 F) for 1000 h with
30 min thermal cycling
Alloy
Weight
change,
mg/cm2
Metal loss,
mm
Internal
oxidation,
mm
230
617
263
X
6.8
80.1
219.4
107.1
0.07
0.17
0.32
0.16
0.09
0.07
0.10
0.07
Extraction residues
M6C (strong)
Cr2N (medium)
AlN (strong)
TiN (medium weak)
CrN (medium strong)
Cr2N (medium strong)
Internal
nitridation,
mm
0.17
>0.41(b)
0.29
>0.40(b)
Total
attack(a),
mm
0.24
>0.58(b)
0.61
>0.56(b)
Note: 1.0 mm=39.4 mils. (a) Metal loss + internal oxidation or internal
nitridation (whichever is greater). (b) Internal nitridation through thickness.
Source: Ref 36
30/10/2007 3:56PM Plate # 0
pg 76
(i.e., 980 C for 1000 h), Lai (Ref 37) examined

the effect of the thermal cycling on the internal
nitridation. In this test, no thermal cycling was
involved. The results of this test (Ref 37) were
then compared with those in the earlier test
(Ref 36). The results of the high-velocity
dynamic burner rig test at 980 C (1800 F) for
1000 h without thermal cycling are summarized
in Table 4.9. The results of chemical analysis
showing nitrogen content before and after testing
are summarized in Table 4.10. In comparing
the test results with thermal cycling and those
without thermal cycling, thermal cycling signicantly accelerated oxidation attack by causing
oxide spallation, as shown in Fig. 4.14. Thermal
cycling was also found to accelerate nitridation
attack, as shown in Fig. 4.15. Among the four
combustor alloys, alloy 230, however, was least
affected by thermal cycling.
The test program was extended to include
some iron-base alloys under the same test
conditions using the same dynamic burner rig
(Ref 38). The results showed that iron-base
alloys suffered signicantly more nitridation
attack than nickel-base alloys. Figure 4.16(a)
shows the microstructure of alloy 556 (Fe-22Cr20Ni-18Co-3Mo-2.5W-0.6Ta-0.2N-La), revealing signicant internal nitridation attack with
formation blocky chromium nitrides after testing

thermal cycling. The nitrogen content was found
to increase to 1.27% after testing from the
original 0.13% before testing. Figure 4.16(b)
shows the microstructure of Type 310 stainless
steel (SS) after testing at 980 C (1800 F)
for 1000 h with 30 min thermal cycling, revealing signicant internal nitridation attack with
formation of blocky chromium nitrides. The
Table 4.9 Test results in terms of weight loss,

depth of oxidation penetration, depth of
nitridation, and total depth of attack after testing
at 980 C (1800 F) for 1000 h without thermal
cycling
Fig. 4.14
Metal
loss,
mm
Internal
oxidation,
mm
Internal
nitridation,
mm
Total
attack(a),
mm
230
617
263
X
3.7
12.1
19.3
5.0
0.039
0.031
0.072
0.030
0.065
0.126
0.135
0.086
0.18
>0.55(b)
0.16
0.10
0.22
>0.58(b)
0.23
0.13
Note: 1.0 mm = 39.4 mils. (a) Metal loss +internal oxidation or internal nitridation
(whichever is greater). (b) Internal nitridation through thickness. Source: Ref 37
Table 4.10 Results of bulk nitrogen analysis

before and after testing at 980 C (1800 F) for
1000 h without thermal cycling
Alloy
Original
nitrogen, wt%
Nitrogen after
testing, wt%
230
263
617
X
0.05
0.004
0.03
0.04
0.065
0.097
0.179
0.075
Source: Ref 37
263
40
60
80
100
219
Non cyclic test
Cyclic test
Comparison weight change data between the
thermal cycling test (30 min cycles) and no thermal cycle test during the dynamic burner rig testing at 980 C
(1800 F) for 1000 h for alloys 230, X, 617, and 263. Source:
Ref 36, 37
Nitrogen, wt%
Weight
change,
mg/cm2
617
20
Cyclic test
0.6
0.4
0.2
0
Nitrogen, wt%
Alloy
230
230
617
263
Non cyclic test
0.4
0.2
0
230
617
263
Original
After testing
Fig. 4.15
Comparison nitrogen gain data between the thermal

cycling test (30 min cycles) and no thermal cycle
test during the dynamic burner rig testing at 980 C (1800 F)
for 1000 h for alloys 230, X, 617, and 263. Source: Ref 36, 37
30/10/2007 3:56PM Plate # 0
pg 77
nitrogen content increased from the original

0.03% before testing to 1.69% after testing.
Alloy 800H was also shown to suffer severe
nitridation attack with chromium nitrides and
needle-shaped aluminum nitrides (Fig. 4.16c).
The oxide scales formed on these iron-base
alloys after testing were porous and nonprotective.
At 870 C (1600 F), internal nitridation
attack was found to be signicantly reduced.
Figure 4.17 shows the microstructures of alloys

230, 617, and X after testing at 870 C (1600 F)
for 2000 h with 30 min cycles. Some aluminum
nitrides and chromium nitrides were observed
in alloy 617, and only some chromium nitrides
were observed in alloy X, while no nitrides were
observed in alloy 230, as shown in Fig. 4.17
(a)
(a)
20 m
50 m
(b)
(b)
(c)
(c)
Fig. 4.16
Extensive internal blocky chromium nitrides

formed in alloy 556 (a), Type 310 (b), and
alloy 800H (c) after the dynamic burner rig testing at 980 C
(1800 F) for 1000 h with 30 min cycles. Courtesy of Haynes
International, Inc.
Fig. 4.17
Alloys 230 (a), 617 (b), and X (c) after the dynamic
burner rig testing at 870 C (1600 F) for 2000 h
with 30 min cycles. Alloy 230 revealed no nitrides, alloy 617
showed both chromium nitrides (blocky phases) and aluminum
nitrides (needle phases), and alloy X showed only blocky chromium nitrides. Internal oxides were observed for all three alloys,
and all three alloys showed porous and nonprotective external
oxide scales. Courtesy of Haynes International, Inc.
30/10/2007 3:56PM Plate # 0
pg 78
(Ref 38). Oxide scales for three alloys were

found to be porous and nonprotective. Iron-base
alloys, such as Type 310SS, on the other hand,
were found to suffer severe internal nitridation
attack, as shown in Fig. 4.18 (Ref 38). The difference in the internal nitridation attack between
nickel-base and iron-base alloys is likely to be
caused by the differences in nitrogen solubilities
in two different alloy systems with much lower
nitrogen solubilities in nickel-base alloys.
The above test data were generated from Ni-Cr
and Fe-Ni-Cr alloys. These alloys are chromia
formers (i.e., alloys forming Cr2O3 oxide scale).
These chromia formers are susceptible to oxidation/nitridation attack in varying degrees under
gas turbine combustion conditions.
For very high temperatures and harsh oxidizing combustion conditions, alumina formers
(i.e., alloys forming Al2O3 oxide scale) are
better performers. Lai (Ref 38) investigated two
alumina formers; one was wrought alloy 214
(Ni-16Cr-3Fe-4.5Al-Y) and the other oxidedispersion-strengthened alloy (produced by
powder metallurgy) MA956 (Fe-20Cr-4.5Al0.5Y2O3). Due to much more tenacious aluminum oxide scales, these two alloys were tested at
1150 C (2100 F) for 200 h with 30 min cycles.
The test results showed no nitridation in either
alloy. Figure 4.19 shows the cross section of
alloy 214 tested specimen, and Fig. 4.20 shows

the cross section of alloy MA956 tested specimen. Both alloys showed ngerlike preferential
oxidation penetration. The preferential oxidation
penetration was the result of thermal stresses
developed from severe thermal cycling from
1150 to less than 260 C (2100 to <500 F) every
30 min. No preferential oxidation penetration
was observed under the same test condition
without thermal cycling. Scanning electron
microscopy with energy-dispersive x-ray spectroscopy (SEM/EDX) analysis of the oxide
2
3
10 m
Fig. 4.19
Scanning electron micrograph showing the oxide

scale of alloy 214 after testing in the dynamic
burner rig at 1150 C (2100 F) with 30 min cycle. The results of
the energy-dispersive x-ray spectroscopy (EDX) analysis of the
oxide scale are summarized: 1, aluminum oxide; 2, aluminum
oxide; and 3, Al-rich (Ni,Cr) oxide
2
1
3
4
10 m
20 m
Fig. 4.18
Extensive blocky chromium nitrides formed in Type

310SS after testing in the dynamic burner rig testing
at 870 C (1600 F) for 2000 h with 30 min cycles. Courtesy of
Haynes International, Inc.
Fig. 4.20
Scanning electron micrograph showing the oxide

scale of alloy MA956 after testing in the dynamic
burner rig at 1150 C (2100 F) with 30 min cycle. The results of
the energy-dispersive x-ray spectroscopy (EDX) analysis of the
oxide scale are summarized: 1, Fe-Al-rich oxide; 24, aluminum
oxide
30/10/2007 3:56PM Plate # 0
pg 79
scales showed that both alloys exhibited an aluminum oxide scale. For MA956 some Fe-Al-rich
oxide phases were observed to form on the top of
the aluminum oxide scale.
A model for oxidation/nitridation reactions
for chromia formers (Ni-Cr and Fe-Ni-Cr alloys)
in oxidizing combustion atmospheres was proposed by Lai (Ref 36). This model, as schematically illustrated in Fig. 4.21, involves the
following reaction steps:
1. Chromium oxides form a protective scale
initially on the alloy surface in oxidizing
combustion atmospheres.
2. Cracks, pores, and other defects develop in
the chromium oxide scales after thermal
cycling and/or long-term exposure.
3. Chromium oxide scales become porous and
nonprotective with nickel oxides forming in
Ni-Cr alloys and iron oxides forming in ironbase alloys.
4. Both O2 and N2 molecules from the combustion gas stream permeate through the
oxide scales and reach the metal underneath.
5. Oxidation of the metal surface results in
lower oxygen potential with a concurrent
increase in nitrogen potential.
6. Nitridation then occurs following the reaction: 1=2N2 (gas) $ N (solution); the concentration of nitrogen absorbed in the metal is
then proportional to the nitrogen potential
( pN2 ) by:
[%N]=k( pN2 )1=2
where k is an equilibrium constant.

7. Nitrogen dissolves into the alloy and diffuses into the metal interior. Increasing
nitrogen concentration in the alloy with
increasing exposure times eventually leads

to the formation of nitrides in the alloy once
the solubility limits for CrN, Cr2N, AlN,
and/or TiN are exceeded.
4.4 Nitridation in NH3-H2O

Environments
A NH3-H2O mixture has been considered in
the Kalina cycle for power generation (Ref 39).
Very little published data are available for the
corrosion behavior of engineering alloys in this
type of environment. Grabke et al. (Ref 40)
investigated the corrosion behavior of a number
of commercial alloys in the NH3-30%H2O gas
mixture at 500 C (930 F). Several interesting
results were obtained from this investigation.
One of the most interesting observations was
that signicant nitridation attack and severe
intergranular cracking were observed in alloy
600 (Ni-16Cr-8Fe) after only 200 h of exposure,
as shown in Fig. 4.22. Alloy 600 has been
known to be one of the most nitridation-resistant
alloys in ammonia environments, and the
alloy has been widely used in ammonia plants
(data are presented in Section 4.5). Alloy 800
(Fe-22Cr-32Ni-Al-Ti) was also found to suffer
severe nitridation attack and intergranular
cracking (Fig. 4.23). For the ferritic stainless
steel, Fe-18Cr (Sicromal), a combination of
severe intergranular nitridation, cracking, and
oxidation attack caused rapid disintegration of
the alloy in 200 h (Fig. 4.24). An austenitic
stainless steel (Fe-18Cr-9Ni) was also found to
suffer severe nitridation attack. However, no
2
CrN
Cr2N
Log pN , atm
Test
environment
6
AIN
Cr2O3
10
Cr
14
18
AI
22
55
45
AI2O3
50 m
35
25
15
Log pO , atm
2
Fig. 4.21
Schematic showing a model for internal nitridation

attack in high-temperature alloys in a simulated
combustion environment. Source: Ref 36
Fig. 4.22
Optical micrograph showing the cross section

of alloy 600 (Ni-16Cr-8Fe) after exposure to the
NH3-30%H2O gas mixture at 500 C (930 F) for only 200 h,
revealing severe nitridation attack and intergranular cracking.
Source: Ref 40
30/10/2007 3:56PM Plate # 0
pg 80
cracking was observed in the austenitic stainless

steel. They found formation of very ne CrN
precipitates in these alloys. The authors (Ref 40)
pointed out that in these tests (low pressure) NH3
should decompose largely to N2 and H2, while
in the Kalina process (under high pressures)
decomposition of NH3 is considered to be
negligible. It is believed that at the test temperature of 500 C, a signicant amount of
NH3 is believed to be retained without decomposition. Robo (Ref 4) found that about 60%
NH3 was decomposed, leaving about 40%
NH3 at the exhaust end at 525 C (980 F) in
his laboratory testing using 100% NH3 in the
inlet gas. Barnes and Lai (Ref 41) found a larger
decomposition of NH3 (100% NH3 in the inlet
test gas and 30% NH3 in the exhaust) when tested
at a slightly higher temperature, 650 C. The
corrosion behavior of alloys in the NH3-H2O
mixture as observed by Grabke et al. (Ref 40)
appears to be different from that of the NH3-H2
mixture, as is discussed in the next section.
(a)
4.5 Nitridation in NH3 and H2-N2-NH3

Environments
Ammonia (NH3) is thermally unstable. It can
readily dissociate into N2 and H2 at elevated
temperatures. There appears to be no published
data available on the dissociation rate as functions of temperature and pressure. However,
several laboratory measurements have indicated
that dissociation rates were extremely fast at
high temperatures. Table 4.11 shows some dissociation data generated in laboratory test furnaces by measuring the amount of NH3 at the
exhaust end of the tube furnace when 100% NH3
was entered into the furnace tube. Both sets of
data were measured in laboratory setups with
pressures being close to 1 atm. In order to avoid
catalyst reactions by metals, the measurement of
ammonia dissociation at 650, 980, and 1090 C
was made with no metallic samples in the furnace tube, which was composed of high-purity
alumina (Ref 38). No published high-pressure
dissociation values are available. Thus, for
laboratory test data even with 100% NH3 as the
200 m
30 m
Fig. 4.24
Optical micrograph showing the cross section of

an Fe-18Cr (Sicromal alloy) tested specimen that
suffered combination of severe nitridation, oxidation, and intergranular cracking after the exposure to the NH3-30%H2O gas
mixture at 500 C (930 F) for 200 h. Source: Ref 40
50 m
(b)
Fig. 4.23 Severe nitridation attack and intergranular cracking

in alloy 800 (Fe-22Cr-32Ni-Al-Ti) after exposure
to the NH3-30%H2O gas mixture at 500 C (930 F) for 200 h. (a)
Surface appearance of the tested specimen showing intergranular
cracks. (b) Cross section of the tested specimen. Source: Ref 40
Table 4.11 NH3 content measured at

the exhaust end of the tube furnace at different
temperatures when 100% NH3 was injected
NH3 at inlet, vol%
Temperature, C (F)
100
525
650
980
1090
(a) Ref 4. (b) Ref 41
C (980 F)
C (1200 F)
C (1800 F)
C (2000 F)
NH3 at exhaust, vol%
40(a)
30(b)
<5(b)
<5(b)
30/10/2007 3:56PM Plate # 0
pg 81
inlet gas, the corrosion reactions generally

involve both N2 and NH3. At very high temperatures, such as 980 and 1090 C (1800 and
2000 F), nitridation is most likely involved in
the reaction with N2 because of rapid dissociation of NH3.
Verma et al. (Ref 42) reported that an ammonia
cracker unit, used to produce nitrogen and
hydrogen, failed after 1000 h of operation. The
preheater tubes (operating at 350 to 400 C,
or 660 to 750 F) were made of Type 304SS,
while the furnace tubes (operating at about
600 C, or 1110 F) were made of Type 310SS.
Both suffered severe nitridation attack. To select
an alternate alloy, nitriding tests were performed
on various alloy samples at 600 C (1110 F) in
an environment consisting of 6 to 8% NH3, 75.77
to 77.5 wt% N2, and 16.25 to 16.5 wt% H2. Test
results are summarized in Table 4.12. The alloys
that performed well include Types 347, 316,
321, SLX-254, and HV-9A. Type 347 was the
best performer, having a linearly extrapolated
penetration rate of about 0.13 mm/yr (5 mpy).
Alloy 800, which contains more nickel than any
of the above stainless steels, did not perform
as well. Furthermore, Type 304 was found to
suffer attack two orders of magnitudes higher
than that of Type 316L. The results also showed
that titanium suffered severe nitridation attack,
which resulted in severe sample cracking. Both
carbon steel and 1Cr-0.5Mo steel suffered decarburization after only 50 h.
Ammonia (NH3) is produced by synthesis
from hydrogen and nitrogen at high pressures
and elevated temperatures. The heart of the
process is the ammonia converter, where
hydrogen and nitrogen combine. Signicant
corrosion issues are associated with the converter and the internal components inside the
converter. The converters operate at high pressures (130350 atm or 8001000 atm) and
temperatures up to 650 C (1200 F) (Ref 43).
Cihal (Ref 43) discussed the major corrosion
problemshydrogen attack and nitridationfor
the ammonia converter. The converter usually
consists of a vessel with a catalyst basket and
an interchanger inside the vessel. Because of
high-pressure, high-temperature hydrogen in the
converter, early converters were constructed out
of a thick-wall steel vessel with an inner carbon
steel lining and vent holes through the vessel
wall. Thus, the inner carbon steel lining was
the only part suffering hydrogen attack, while the
main thick-wall vessel was unaffected by highpressure, high-temperature hydrogen (Ref 43).
Hydrogen attack is the damage of steel by the
reaction of hydrogen with cementite (Fe3C) in
steel to form methane gas (CH4), resulting in
formation of microcracks and ssures as well
as decarburization in steel. (Hydrogen attack is
reviewed and discussed in Chapter 17.) Later
designs of the converter allowed the cold inlet
gas owing along the vessel wall to keep the
vessel cold, thus eliminating the potential hydrogen attack problem for the vessel (Ref 43).
Cihal (Ref 43) indicated that the internal
components made of carbon steels exhibited a
short life due to hydrogen attack. Alloy steels
containing chromium were more resistant to
hydrogen attack, but had suffered severe embrittlement problems due to nitridation attack.
Figure 4.25 shows intergranular cracking in
the nitrided layer of an alloy steel (0.12C5.6Cr-0.42Mo) after exposure to the synthesis
gas inside the converter at 325 atm and 450 to
500 C (840 to 930 F) for 4380 h (Ref 43).
An alloy steel containing a strong nitride former
such as titanium, such as alloy steel with
Table 4.12 Nitridation attack of various alloys in an ammonia-bearing environment at 600 C

(1110 F) for indicated exposure times
Penetration depth of nitrtdiltion attack, mm (mils)
Alloy
Carbon steel
1Cr-0.5Mo steel
Titanium
304
316L
329
310
321
347
SLX-254(b)
HV-9A(c)
800
50 h
100 h
300 h
600 h
1000 h
1500 h
Decarb.
Decarb.
0.0066 (0.3)
0.02 (0.8)
0.013 (0.5)
0.013 (0.5)
0.01 (0.4)
0.02 (0.8)
Decarb.
Decarb.
0.0133 (0.5)
0.013 (0.5)
0.02 (0.8)
0.066 (2.6)
0.03 (1.2)
0.013 (0.5)
0.013 (0.5)
0.013 (0.5)
0.10 (3.9)
0.10 (3.9)
Decarb.
0.033 (1.3)
0.233 (9.2)
0.013 (0.5)
0.02 (0.8)
0.10 (3.9)
0.13 (5.1)
0.013 (0.5)
0.013 (0.5)
0.026 (1.0)
0.10 (3.9)
0.20 (7.9)
Decarb.
0.033 (1.3)
0.266 (10.5)
0.03 (1.2)
0.03 (1.2)
0.10 (3.9)
0.16 (6.3)
0.016 (0.6)
0.013 (0.5)
0.026 (1.0)
0.10 (3.9)
Decarb.
0.033 (1.3)
Cracked(a)
0.06 (2.4)
0.04 (1.6)
0.20 (7.9)
0.33 (13.0)
0.06 (2.4)
0.02 (0.8)
0.06 (2,4)
0.10 (3.9)
0.20 (7.9)
Decarb.
0.3 (11.8)
Cracked(a)
4.2 (165)
0.04 (1.6)
0.40 (15.7)
0.40 (15.7)
0.06 (2.4)
0.02 (0.8)
0.06 (2.4)
0.10 (3.9)
0.20 (7.9)
Decarb.: decarburized. (a) Nitridation through thickness. (b) SLX-254: Fe-19.7Cr-24.5Ni-4.35Mo-1.43Cu. (c) HV-9A: Fe-21.2Cr-24.6Ni-3.8Mo-1.5Cu. Source: Ref 42
30/10/2007 3:56PM Plate # 0
pg 82
0.05% C, 2.9% W, and 0.54% Ti, exposed to the

same converter environment under the same
test conditions as the 0.12C-5.6Cr-0.42Mo steel
(as shown in Fig. 4.25) was found to show no
cracking.
Nitridation resistance of various alloys was
studied by Moran et al. (Ref 44) in an ammonia
converter and preheater line. The results are
summarized in Table 4.13. Corrosion rates were
found to depend strongly on the concentration
of ammonia. Type 304, for example, suffered
corrosion rates that increased from 0.02 to
2.5 mm/yr (0.6 to 99 mpy) as the concentration
of NH3 was increased from 5 to 6% (in the
ammonia converter) to 99% (in the ammonia
preheater line) at about 500 C (930 F). In an
ammonia converter with about 5 to 6% NH3 and
490 to 550 C (910 to 1020 F), all stainless
steels tested (i.e., 430, 446, 302B, 304, 316,
321, 309, 314, 310, and 330) showed negligible
nitridation attack, with corrosion rates of
about 0.03 mm/yr (1 mpy) or less. For the plant
ammonia line (preheater exit), which was
exposed to 99% NH3, stainless steels, such as
446, 304, 316, and 309, suffered severe nitridation attack, with corrosion rates of about
2.54 mm/yr (100 mpy) or more. Moran et al.
(Ref 44) found that Type 316 suffered signicantly more attack than Type 304, contrary
to the observations of Verma et al. (Ref 42).
Robo (Ref 4) reported the performance of
several alloys in a Topsoe-type ammonia converter. Most of the components made of
Type 304, exposed to temperatures up to 500 C
(930 F) with ammonia concentration up to 20%,
exhibited negligible nitridation rates (0.4 to
4 mpy). One Type 304 sample showed a slightly

higher corrosion rate (10 mpy), presumably due
to a higher temperature. Alloy 600 (Ni-Cr-Fe
alloy) had signicantly better nitridation resistance than stainless steels, with corrosion rates
1 or 2 orders of magnitude lower. The results are
summarized in Table 4.14 (Ref 4).
McDowell (Ref 45) reported eld test results
performed in a Casale converter (540 C, or
1000 F, and 11 ksi) for 1 and 3 years. These
results are summarized in Table 4.15. AISI 502
(5Cr steel) was extremely susceptible to nitridation attack, with more than 2.54 mm (0.1 in., or
100 mils) of nitridation depth in a year. Results
showed a general trend of increased resistance
to nitridation as nickel content in the alloy
increased. One striking observation was that after
3 years of exposure, the alloys showed essentially similar depths of nitridation attack as they
did after 1 year.
Robo (Ref 4) reported kinetic data for
Type 304 and alloys 600 and 625 in laboratory
tests performed with pure ammonia as the inlet
test gas. The nitridation for Type 304 was found
to follow a linear rate law at 525 C (980 F) for
up to 1000 h. The maximum thickness of the
nitride layer (in the form of scale) measured
metallographically is shown in Fig. 4.26 as a
function of time. A growth rate of about 0.37
m/h was observed. This corresponds to about
3240 m/yr (128 mpy). This rate is signicantly higher than those observed in ammonia
converters. The ammonia concentration in this
test (reportedly, 40% ammonia was dissociated
Table 4.13 Corrosion behavior of various
alloys in an ammonia converter and plant
ammonia line
Corrosion rate. mm/yr (mpy)
Alloy
430
446
302B
304
316
321
309
314
310
330 (0.47Si)
330 (1.00Si)
600
80Ni-20Cr
Ni
Fig. 4.25
Intergranular cracking in the nitrided layer of an

alloy steel (0.12C-5.6Cr-0.42Mo) after exposure
to the synthesis gas inside the converter at 325 atm and 450 to
500 C (840 to 930 F) for 4380 h. Source: Ref 43
Ammonia
(converter)(a)
Plant ammonia
line(b)
0.022 (0.90)
0.028 (1.12)
0.019 (0.73)
0.015 (0.59)
0.012 (0.47)
0.012 (0.47)
0.006 (0.23)
0.003 (0.10)
0.004 (0.14)
0.002 (0.06)
0.001 (0.02)
4.18 (164.5)
2.53 (99.5)
>13.21 (520)
2.41 (95)
0.43 (17.1)
0.16 (6.3)
0.19 (7.4)
2.01 (79.0)
(a) 5 to 6% NH3, 29164 h at 490 to 550 C (910 to 1020 F), and 354 atm
(5200 psi) ( Haber-Bosch converter). (b) 99. 1% NH3, 1540 h at 500 C (930 F).
Source: Ref 44
30/10/2007 3:56PM Plate # 0
pg 83
Table 4.14
Converter No.
1
2
3
Corrosion behavior of various alloys in an ammonia converter

Component
Material
Gas
temperature,
C
Lining
Plate, 2nd bed
Bolt
Wire mesh, 2nd bed
Perforated plate. 1st bed
Inner shell, 2nd bed
Perforated plate. 2nd bed
Center tube, 2nd bed
Nut, bottom
Bolt, bottom
Wire mesh
ThermoweII
304
304
302
Alloy 600
304
304
304
304
304
403
Alloy 600
304
525
475
520
500
440
440
485
480
480
c.500
c.500
NH3
1520
1520
13
810
810
16
16
16
3.5
Time of
operation, yr
Thickness
of nitride,
m (mils)
4
7
7
7
5
5
5
5
5
5
4
8
1000 (39.4)
100 (3.9)
375 (14.8)
8 (0.3)
270 (10.6)
45 (1.8)
60 (2.4)
440 (17.3)
260 (10.2)
540 (21.3)
6 (0.2)
200 (7.9)
Average nitriding,
m/yr (mpy)
250 (9.8)
14 (0.6)
54 (2.1)
1 (0.04)
54 (2.1)
9 (0.4)
12 (0.5)
88 (3.5)
52 (2.0)
108 (4.3)
1.5 (0.06)
25 (1.0)
Note: Topsoe-type ammonia converter operated at 22 MPa (3.2 ksi). Source: Ref 4
Table 4.15 Depth of nitridation for various alloys

after 1 and 3 years in a Casale ammonia converter
500
Nitridation depth, mm (mils)
502 (5Cr steel)

446
304
316
321
347
309
310
800
804 (30Cr-42Ni)
600
Nickel 200
1 yr
3 yr
2. 88 (113.2)
1.06 (41.7)
1.08 (42.7)
0.46 (18. 2)
0.46 (18.3)
0.49 (19.2)
0.24 (9.5)
0.22 (8.8)
0.14 (5.4)
0.03 (1.2)
0.16 (6.4)
None
Completely nitrided
1.15 (45.3)
1.12 (44.0)
0.48 (18.7)
0.60 (23.6)
0.45 (17.6)
0.24 (9.6)
0.23 (9.2)
0.13 (5.3)
0.03 (1.2)
0.16 (6.4)
None
Note: Operated at 538 C (1000 F) and 76 MPa (11 ksi). Source: Ref 45
in the test furnace) was signicantly higher than

in the ammonia converters. The rate of about
3.25 mm/yr (128 mpy) was, however, of the
same order of magnitude as that observed by
Moran et al. (Ref 44) in the plant ammonia line
(about 100 mpy for Type 304).
At 700 C (1290 F), Robo (Ref 4) found that
the reaction rates for Type 304, alloy 600, and
alloy 625 can be described by:
X =kt n
2[N]
Dt
c[m]
300
200
100
500
1000
Exposure time, h
Fig. 4.26
Nitriding depth of Type 304SS in ammonia (100%

in the inlet gas and 60% in the exhaust) at 525 C
(980 F) as a function of exposure time. Source: Ref 4
4:5
where X is thickness of the nitrided layer in m,

k is reaction constant, t is time in hours, n is 0.66
for Type 304 and 0.26 for alloys 600 and 625.
Jack (Ref 46) developed a kinetic model based
on the models for internal oxidation, to describe
the growth of internal penetration (X) in the
absence of iron nitride formation:
X 2=
400
Thickness of nitride layer, m
Alloy
4:6
where [N] is the surface nitrogen concentration

(at.%), [m] is the alloy element concentration
(at.%), is the ratio of nitrogen to alloy element

in the nitride phase, D is diffusivity of nitrogen,
and t is time.
This model predicts that fast nitriding rates
can be achieved by increasing the ammonia
content in the gas mixture and thus the surface
nitrogen concentration. Another important factor
in nitriding kinetics is the concentration of the
alloy element that forms internal nitrides. The
model predicts that nitriding depth is inversely
proportional to the concentration of the nitrideforming alloy element.
30/10/2007 3:56PM Plate # 0
pg 84
The nitridation behavior of a wide variety of

commercial alloys in ammonia was extensively
investigated by Barnes and Lai (Ref 41). The
alloys tested included stainless steels, Fe-Ni-Cr
alloys, and nickel- and cobalt-base superalloys.
Many alloys tested contained alloying elements
(e.g., Al, Ti, Zr, Nb, Cr, Mo, W, Fe, etc.) that
form nitrides. The test results generated at 650,
980, and 1090 C (1200, 1800, and 2000 F)
are summarized in Tables 4.16 to 4.18. It was
found that nickel-base alloys are generally more
nitridation resistant than iron-base alloys. Increasing nickel content generally improves
the resistance of the alloy to nitridation attack.
Increasing cobalt content appears to have the
same effect. When nitrogen absorption was
plotted against Ni + Co content in the alloy for
the 650 C (1200 F) test data, resistance to
nitridation improves with increasing Ni + Co
content up to about 50 wt%, as shown in
Fig. 4.27. Further increases up to about 75% did
not seem to affect the nitridation resistance of the
alloy. For maximum resistance to nitridation
attack at 650 C (1200 F), it appears that alloys
with at least 50% Ni or Ni + Co are most suitable.
This is in general agreement with the results
reported by Moran et al. (Ref 44). Their results
suggested that improvement in nitridation resistance began to level off at about 40% Ni, with no
improvement resulting from further increases in
nickel up to about 80%. Pure nickel, however,
showed signicantly lower nitridation resistance
(Ref 43). At 980 C (1800 F), a slightly different relationship was observed, as shown in
Table 4.17 Nitridation resistance of various

alloys in ammonia at 980 C (1800 F) for
168 h
Alloy
214
600
S
601
230
617
HR-160
188
625
6B
253MA
25
X
RA333
RA330
800H
825
150
MULTIMET
316
556
304
310
446
Alloy
C-276
230
HR-160
600
625
RA333
601
188
S
617
214
X
825
800H
556
316
310
304
Alloy base
Nitrogen
absorption,
mg/cm2
Depth of nitride
penetration,
mm (mils)
Nickel
Nickel
Nickel
Nickel
Nickel
Nickel
Nickel
Cobalt
Nickel
Nickel
Nickel
Nickel
Nickel
Iron
Iron
Iron
Iron
Iron
0.7
0.7
0.8
0.8
0.9
1.0
1.1
1.2
1.3
1.3
1.5
1.7
2.5
4.3
4.9
6.9
7.4
9.8
0.02 (0.6)
0.03 (1.2)
0.01 (0.5)
0.03(1.3)
0.01 (0.5)
0.03 (1.0)
0.03 (1.0)
0.02 (0.6)
0.03 (1.1)
0.03 (1.0)
0.04 (1.5)
0.04 (1.5)
0.06 (2.2)
0.10 (4.1)
0.09 (3.5)
0.19 (7.3)
0.15 (6.0)
0.21 (8.4)
Note: 100% NH3 in the inlet gas and 30% NH3 in the exhaust gas. Source: Ref 41
Depth of
nitride penetration,
mm (mils)
Nickel
Nickel
Nickel
Nickel
Nickel
Nickel
Nickel
Cobalt
Nickel
Cobalt
Iron
Cobalt
Nickel
Nickel
Iron
Iron
Nickel
Cobalt
Iron
Iron
Iron
Iron
Iron
Iron
0.3
0.9
0.9
1.2
1.4
1.5
1.7
2.3
2.5
3.1
3.3
3.6
3.2
3.7
3.9
4.0
4.3
5.3
5.6
6.0
6.7
7.3
7.7
12.9
0.04 (1.4)
0.12 (4.8)
0.18 (7.2)
0.17 (6.6)
0.12(4.9)
0.38 (15.0)
0.18 (7.2)
0.19(7.4)
0.17 (6.9)
0.15 (5.8)
0.48 (19.0)
0.26(10.4)
0.19 (7.4)
0.42(16.4)
0.52 (20.6)
0.28 (11.1)
0.58 (23.0)
0.38 (15.1)
0.35 (13.6)
0.52 (20.3)
0.37 (14.7)
>0.58 (23.0)
0.38 (15.1)
>0.58 (23.0)
Note: 100% NH3 in the inlet gas and less than 5% NH3 (detection limit) in the
exhaust gas. Source: Ref 41

alloys in ammonia at 1090 C (2000 F) for
168 h
Alloy

alloys in ammonia at 650 C (1200 F) for 168 h
Alloy base
Nitrogen
absorption
mg/cm2
600
214
S
230
25
617
188
HR-160
601
RA330
625
316
304
X
150
556
446
6B
MULTIMET
825
RA333
800H
253MA
310
Alloy base
Nitrogen
absorption,
mg/cm2
Depth of nitride
penetration,
mm (mils)
Nickel
Nickel
Nickel
Nickel
Cobalt
Nickel
Cobalt
Nickel
Nickel
Iron
Nickel
Iron
Iron
Nickel
Cobalt
Iron
Iron
Cobalt
Iron
Nickel
Nickel
Iron
Iron
Iron
0.2
0.2
1.0
1.5
1.7
1.9
2.0
2.5
2.6
3.1
3.3
3.3
3.5
3.8
4.1
4.2
4.5
4.7
5.0
5.2
5.2
5.5
6.3
9.5
0
0.02 (0.7)
0.34 (13.4)
0.39 (15.3)
>0.65 (25.5)
>0.56 (22)
>0.53 (21)
0.46 (18)
>0.58 (23)
>0.56 (22)
>0.56 (22)
>0.91 (36)
>0.58 (23)
>0.58 (23)
0.51 (20)
>0.51 (20)
>0.58 (23)
>0.64 (25)
>0.64 (25)
0.58 (23)
>0.71 (28)
>0.76 (30)
>1.5 (60)
>0.79 (31)
Note: 100% NH3 in the inlet gas and less than 5% (detection limit) in the exhaust
gas. Source: Ref 41
30/10/2007 3:56PM Plate # 0
pg 85
15
Nitrogen absorption, mg/cm2
10
20
40
60
10
80
20
40
60
80
Ni + Co, wt%
Ni + Co, wt%
Fig. 4.28
Fig. 4.27
Effect of the Ni + Co content in iron-, nickel-,

and cobalt-base alloys on nitridation resistance at
650 C (1200 F) for 168 h in ammonia (100% NH3 in the inlet
gas and 30% NH3 in the exhaust). Source: Ref 41
Effect of the Ni + Co content in iron-, nickel-,

and cobalt-base alloys on nitridation resistance at
980 C (1800 F) for 168 h in ammonia (100% NH3 in the inlet
gas and <5% NH3 in the exhaust). Source: Ref 41
Fig. 4.28. Nitrogen absorption was reduced

drastically with an initial 15% Ni (or Ni + Co). As
Ni + Co content increased from 15 to 50%, no
drastic improvement in nitridation resistance was
noted. Further increases in Ni + Co content in
excess of about 50% caused a sharp improvement. Alloys with Ni (or Ni + Co) in excess of
about 60% showed the most resistance to nitridation. In this test program (Ref 41), no alloys
with 80% Ni (or Ni + Co) and higher were tested.
It is generally believed that the benecial
effect of nickel or cobalt in increasing nitridation
resistance is caused by the reduced solubility of
nitrogen in the alloy. Nickel and cobalt were
found to reduce the solubility of nitrogen in iron
(Ref 15, 47).
Morphology of nitrides formed in alloys as a
result of exposure to NH3 is widely different
between low- and high-temperature exposures.
Nitridation at low temperatures (e.g., 650 C, or
1200 F) generally results in a surface nitride
layer. For iron-base alloys, the surface nitride
layer consists of mostly iron nitrides (Fe2N or
Fe4N), while for nickel- and cobalt-base alloys,
the nitride layer consists of mainly CrN. Hightemperature exposures, on the other hand, result
in formation of internal nitrides, which are
mostly CrN, Cr2N, (Fe,Cr)2N, AlN, and TiN.
Table 4.19 summarizes the results of the x-ray
diffraction analysis performed on the surfaces of
the selected test specimens tested at different
temperatures. Figure 4.29 illustrates the morphology of the surface nitride layer formed at
Table 4.19 Phases detected from the x-ray

diffraction analysis performed on the surfaces
of test specimens after exposure to NH3 at
temperatures indicated for 168 h
Phases
Alloy
304SS
800H
556
230
188
650 C (1200 F)
980 C (1800 F)
1090 C (2000 F)
Fe2N
(Fe3Ni)N
(Fe3Ni)N
CrN
CrN
CrN
CrN
CrN
CrN
CrN
(Cr,Fe)2N1-x
(Cr,Fe)2N1-x
(Cr,Fe)2N1-x
(Cr,Mo)12(Fe, Ni)8-xN4-z
Cr2N, CrN
Note: At 650 C (1200 F), 100% NH3 in the inlet gas and 30% NH3 in the
exhaust. At 980 and 1090 C (1800 and 2000 F), 100% NH3 in the inlet gas and
<5% NH3 in the exhaust. Source: Ref 41
650 C (1200 F) for iron-base alloys, Type

446SS (Fe-25Cr) and Type 304SS (Fe-18Cr8Ni), and nickel-base alloys, alloys 600 (Ni16Cr-8Fe), 625 (Ni-22Cr-9Mo-3.5Nb-3Fe), X
(Ni-22Cr-18.5Fe-9Mo-0.6W), and C-276 (Ni16Cr-5Fe-16Mo-4W). Type 446SS containing
about 25% Cr with no nickel exhibited a thick
nitride layer (about 0.72 mm, or 28 mils, thick)
formed on the alloy surface, as shown in
Fig. 4.29(a). With about 8% Ni in the alloy, Type
304SS showed a signicantly thinner nitride
layer (about 0.2 mm, or 0.008 in., 8 mils thick)
formed on the alloy surface. For four nickelbase alloys, alloys 600, 625, X, and C-276, an
extremely thin nitride layer (about 1.2 to 2.2 m
thick) was found to form on the alloy surface.
The iron content in these four nickel-base alloys
varies from about 3 to 19%.
30/10/2007 3:56PM Plate # 0
pg 86
When exposed to high temperatures, the alloy

forms internal nitrides due to increased diffusivity of nitrogen. Figure 4.30 shows internal
nitrides formed in Type 446SS, Type 304SS,
alloy 800H (Fe-21Cr-32Ni-0.4Al-0.4Ti), and
alloy 188 (Co-22Cr-22Ni-14W-La). For Type
304SS, alloy 800H, and alloy 188, nitrides
appeared to be etched differently with the nitrides
formed near the surface etched darker than those
in the interior. The nitrides (etched darker) that
formed in the surface zone are believed to be
CrN. The x-ray diffraction analysis of the test
specimen surface showed CrN for these three
alloys (Table 4.19). As nitrogen diffuses farther

into metal interior, nitrogen activities become
lower, thus forming Cr2N. Thus, the nitrides
(etched lighter) formed in the metal interior are
believed to be Cr2N.
Ni-Cr alloys containing aluminum or titanium
(or both) form internal nitrides of not only
chromium but also aluminum or titanium (or
both) at high temperatures. Both aluminum and
titanium are stronger nitride formers, and AlN
and TiN can form and penetrate farther into the
metal interior than CrN and Cr2N. When alloys
contain relatively low aluminum (e.g., about
(a)
200 m
(d)
(b)
200 m
(e)
(c)
Fig. 4.29
200 m
(f)
200 m
200 m
200 m
Optical micrographs showing typical nitride morphology of a surface nitride layer that formed on the alloy surface when
exposed to NH3 (100% NH3 in the inlet gas and 30% NH3 in the exhaust) for 168 h at 650 C (1200 F) for (a) Type 446, (b)
Type 304, (c) alloy 600, (d) alloy 625, (e) alloy X, and (f) alloy C-276. Courtesy of Haynes International, Inc.
30/10/2007 3:56PM Plate # 0
pg 87
1%), such as alloys 601 and 617, a signicant

amount of internal aluminum nitrides formed in
the alloy, as shown in Fig. 4.31(a). Also observed
in alloy 601 (Fig. 4.31a) are chromium nitrides
(blocky-type phases) that formed near the alloy
surface. For alloys containing high concentrations of aluminum (e.g., 4.5% Al in alloy 214),
aluminum nitrides formed on the alloy surface,
as shown in Fig. 4.31(b). Al2O3 oxide is also
believed to form on the alloy 214 surface.
4.6 Nitridation in N2 Atmosphere

Metals and alloys are also susceptible to
nitridation attack in N2 or N2-H2 environments,
particularly at high temperatures. The N2 or
N2-H2 atmosphere is commonly used as a protective atmosphere in heat treating and sintering operations. Figure 4.32 shows extensive
nitridation attack of Type 314 wire mesh belt
in a sintering furnace after 2 to 3 months of
(a)
50 m
(c)
(b)
50 m
(d)
Fig. 4.30
50 m
50 m
Optical micrographs showing typical nitride morphology in form of internal nitrides penetrating into the metal interior when
exposed to NH3 for 168 h at 980 C (1800 F) for (a) Type 446, (b) Type 304, (c) alloy 800H, and (d) alloy 188. Courtesy of
Haynes International, Inc.
30/10/2007 3:56PM Plate # 0
service at 1120 C (2050 F) in the N2-10%H2

atmosphere. Type 314SS (Fe-26Cr-20Ni-2Si) is
commonly used for wire mesh furnace belts.
Smith and Bucklin (Ref 48) investigated nitridation reactions in 100% N2 for several iron- and
nickel-base alloys. Their results, generated
at 980, 1090, and 1200 C (1800, 2000, and
2200 F), are tabulated in Table 4.20. As shown
in the table, the nitridation kinetics in 100% N2 is
extremely rapid. Even nickel-base alloys were
found to suffer severe nitridation attack even
when the temperature was reduced to 980 C
(1800 F). Both AlN and Cr2N were found in
alloys 600 and 800 after exposure to 100% N2 at
1200 C (2200 F) for 100 h. For RA330, only
Cr2N was detected after exposure to the same test
conditions. Ganesan and Smith (Ref 49) identied the nitride phases formed near the surface of
the test specimens after exposure at 980 C
(1800 F) for 1008 h in pure nitrogen atmosphere using x-ray diffraction. The major phases
identied are summarized in Table 4.21 (Ref 49).
200 m
(a)
200 m
(b)
Fig. 4.31
(a) Extensive internal aluminum nitride (long

needle phase) formation in alloy 601 and (b)
insignificant AlN formation in alloy 214 after exposure to NH3
at 1090 C (2000 F) for 168 h
For alloys containing 20% or more chromium,

Cr2N-type nitrides were found in the region near
the surface, except alloy 600, which contains
only about 16% Cr. This is in agreement with the
phase stability diagram at 1000 C in terms of
pN2 versus Cr content in Ni-Cr alloys as shown in
Fig. 4.4, which shows CrN is the most likely
nitride in Ni-16Cr alloy (Ref 12).
Barnes and Lai (Ref 50) conducted an extensive nitridation study in pure nitrogen atmosphere for iron-, nickel-, and cobalt-base alloys
at 1090 C (2000 F) for 168 h. Test results
in terms of nitrogen absorption (mg/cm2) and
the depth of nitridation are summarized in
Table 4.22. As a result of rapid nitridation
kinetics under the test condition, nitridation
attack penetrated through the thickness of the
test specimen for many alloys. Due to different
thicknesses for different alloys, the ranking of
alloy performance in terms of nitridation depths
became difcult for most alloys tested. (The
thickness of the test specimen varied from alloy
to alloy, because of the use of whatever sheet
products were available for preparation of test
specimens.) Iron-base alloys, the last group from
RA330 to Type 310SS, suffered the worst nitridation attack. Two cobalt-base alloys, alloys 188
(Co-22Cr-22Ni-14W-La) and 150 (Co-27Cr18Fe), exhibited poor resistance, with alloy 150
(high Cr and no Ni) showing extremely poor
nitridation resistance similar to iron-base alloys.
The nitride phases formed in alloys were
analyzed using x-ray diffraction performed on
the chemical extraction residues obtained from
the test specimens. Selected alloys (six nickelbase alloys, two cobalt-base alloys, and one
iron-base alloy) were analyzed, and the x-ray
diffraction analysis results are summarized in
Table 4.23. All the alloys except alloy 214
exhibited Cr2N nitrides. No internal nitrides
were observed in alloy 214, which contains 4.5%
Al. The alloy 214 specimen showed only surface
Al2O3 and AlN phases, as analyzed by x-ray
diffraction analysis performed on the surface
scales of the specimen. For nickel-base alloys
containing low levels of aluminum, such as
alloys 601 and 617 (both contain about 1.3% Al),
extensive AlN nitrides formed in metal interior.
Figure 4.33 shows a through-thickness nitrided alloy 617 specimen, exhibiting extensive
needle-shaped internal AlN nitrides along with
Cr2N nitrides. Figure 4.34 shows needle-shaped
internal AlN nitrides as well as Cr2N nitrides at
high magnication in alloy 601. The addition
of 4.5% Al to a nickel-base alloy can provide
pg 88
30/10/2007 3:56PM Plate # 0
pg 89
a very effective protection against nitridation

attack at high temperatures. The nitride formed
on alloy 214 surface after 168 h was too thin to
be identied. However, after 500 h of exposure,
the surface scale was found to be composed
of AlN and Al2O3, with AlN predominating
(Ref 50). The nitrogen absorbed after 168 h
was 0.3 mg/cm2, and no further increase in

nitrogen absorption was observed after 504 h.
Comparing alloy 214 with another nickel-base
alloy containing little aluminum, such as alloy
230, in 100% N2 at 1090 C (2000 F) for up to
500 h of exposure clearly showed the superior
resistance of alloy 214 against nitridation attack
1.3 mm (50 mils)

100 m
Fig. 4.32
Optical micrograph showing extensive internal chromium nitrides that formed in the entire cross section of a wire sample
obtained from a Type 314 wire mesh belt in a sintering furnace after service for 2 to 3 months at 1120 C (2050 F) in N2-10%
H2. Courtesy of Haynes International, Inc.
30/10/2007 3:56PM Plate # 0
pg 90
Table 4.20
Alloy
600
601
800
520
330
DS
314SS
Nitridation resistance of iron- and nickel-base alloys in pure nitrogen

982 C (1800 F)/1008 h(a)
nitrided depth, mm (mils)
1093 C (2000 F)/900 h(b)

1204 C (2200 F)/100 h(c)

1.30 (51)
1.55 (61)
1.85 (73)
2.57 (101)
>3.81 (150)
1.85 (73)
2.79 (110)
>3.81 (150)
>3.81 (150)
>3.81 (150)
>3.81 (150)
>3.81 (150)
2.16 (85)
>3.81 (150)
>3.81 (150)
>3.81 (150)
(a) Specimens were cycled to room temperature once every 24 h for the rst 3 days and then weekly for the remainder of the test. (b) Specimens were cycled to room
temperature once every 96 h (4 days). (c) Isothermal exposure. Source: Ref 48
Table 4.21 Major phases formed in the nearsurface region of the test specimens after exposure
to 100% N2 at 980 C (1800 F) for 1008 h, as
determined by x-ray diffraction
Table 4.23 Results of x-ray diffraction analysis of

extraction residues obtained from specimens after
exposure to 100% N2 at 1090 C (2000 F) for
168 h
Alloy
Alloy
Major phases
Type 314SS
Type 330SS
Alloy 800
Alloy 601
Alloy 600
(Cr,Fe)2N
(Cr,Fe)2N
(Cr,Fe)2N, AlN
(Cr,Fe)2N, AlN
CrN
Source: Ref 49
Table 4.22 Nitrogen absorbed (mg/cm2) and the

average depth of internal nitridation for iron-,
nickel-, and cobalt-base alloys after exposure in
100% N2 at 1090 C (2000 F) for 168 h
Alloy
214
600
230
HR160
X
617
601
188
150
RA330
RA85H
556
HR120
253MA
800H
800HT
Type 310 SS
Nitrogen absorbed,
mg/cm2
Depth of internal
nitridation, mm
0.2
1.1
2.7
3.9
6.0
5.1
7.2
3.7
9.0
6.6
8.5
9.0
9.6
10.0
10.3
11.4
12.3
0.0
0.41
0.46
1.19
0.63
>0.58
>0.59
>0.51
>0.80
>1.52
>1.44
>1.52
>0.86
>1.50
>1.50
>1.46
>0.79
214 (a)
230
600
601
617
HR160
188
150
RA85H
Phases detected
AlN, Al2O3
Cr2N, (Cr,Mo)12(Fe, Ni)8-xN4-z, M6C
Cr2N, TiN
Cr2N, AlN
Cr2N, AlN
CrN, Cr2N
Cr2N
Cr2N
Cr2N, AlN
(a) Surface analysis. Source: Ref 50
Source: Ref 50
Fig. 4.33
with AlN/Al2O3 surface scales, as illustrated in

Fig. 4.35.
Alloy 150 (Co-27Cr-18Fe) suffered nitridation attack as severe as that experienced by
some iron-base alloys. Figure 4.36 shows a
through-thickness nitrided alloy 150 compared
with Type 310SS. Extensive Cr2N nitrides were
Optical micrographs showing a through-thickness

nitridation attack for alloy 617, a nickel-base alloy
containing about 1.3%Al, after exposure to 100% N2 at 1090 C
(2000 F) for 168 h. Note extensive blocky chromium nitrides and
needle-shaped aluminum nitrides. Magnification bar represents
200 m. Courtesy of Haynes International, Inc.
observed throughout the alloy 150 specimen

cross section.
30/10/2007 3:56PM Plate # 0
pg 91
Nickel-base alloys are in general more resistant to nitridation attack than iron-base alloys.
This is illustrated in Fig. 4.37 comparing alloy X
with 253MA after 168 h in 100% N2 at 1090 C
(2000 F). Similar ndings were observed in
nitridation studies in nitrogen atmospheres by
Smith and Bucklin (Ref 48) and Tjokro and
Young (Ref 51). Tjokro and Young (Ref 51)
investigated a number of commercial alloys
in N2-5%H2 at 1100 and 1200 C. Their results
10 m
showed the nitridation rate constants decreased

with increasing nickel concentration, as illustrated in Fig. 4.38.
(a)
200 m
(b)
200 m
Fig. 4.34

image) showing internal chromium nitrides
(blocky phases) and aluminum nitrides (long needle-shaped
phases) formed in alloy 601, a nickel-base alloy containing about
1.3% Al, after exposure to 100% N2 at 1090 C (2000 F) for
4.50
230 alloy
N absorbed, mg/cm2
4.00
3.50
3.00
2.50
2.00
1.50
1.00
0.50
0.00
0.00
214 alloy
200.00
400.00
600.00
Time, h
Fig. 4.35
Nitridation kinetic data for alloy 214 (nickel-base

alloy containing 4.5% Al) and alloy 230 (nickelbase alloy containing little aluminum) after exposure to 100% N2
at 1090 C (2000 F) for 168 h. Source: Ref 50
Fig. 4.36
Optical micrographs showing through-thickness

nitridation attack for (a) Type 310SS (Fe-25Cr20Ni) and (b) alloy 150 (Co-27Cr-18Fe) after exposure to 100%
N2 at 1090 C (2000 F) for 168 h. Courtesy of Haynes International, Inc.
30/10/2007 3:57PM Plate # 0
4.7 Nitridation Kinetics between NH3

and N2 Atmospheres
Nitridation data generated in 100% NH3
(Ref 41) and those generated in 100% N2
(Ref 50) were compared. Since both test programs were generated using the same test
apparatus and procedures, and both tests were
carried out by same technicians, the laboratoryto-laboratory variation was signicantly minimized. Thus, the comparison between these two
sets of test results could yield a more meaningful
comparison in terms of the difference in environments. The test results generated at 1090 C
(a)
200 m
(2000 F) for 168 h in 100% NH3 (Ref 41) and

100% N2 (Ref 50) are tabulated in Table 4.24.
The data are also presented in terms of nitrogen
absorption as a function of Ni + Co content in the
alloy (Fig. 4.39). The results clearly indicated
that the nitrogen atmosphere was a more severe
nitriding environment than the ammonia environment at 1090 C (2000 F). The amount of
nitrogen absorbed in N2 environment was more
than double that in NH3 environment for many
alloys. Figure 4.40 shows two nitrided alloy
601 specimens, one exposed to NH3 environment and the other to N2 environment. The
N2 environment caused signicantly more
internal nitride formation than for the NH3
environment. More chromium nitrides (blocky
shaped) and aluminum nitrides (needle shaped)
formed in the N2 environment than in the NH3
environment.
Ammonia readily dissociates to one part N2
and three parts H2 at 1090 C (2000 F). With
the test system used in the study by Barnes and
Lai (Ref 41), 100% NH3 was fed into the alumina
test tube with no test specimens inside, and the
exhaust gas was measured to contain less than
5% NH3, which was the detection limit of the the
apparatus used for measuring NH3 (Table 4.11).
It is believed most, if not all, of the ammonia
had been dissociated into H2 and N2 before the
test gas was in contact with the test specimens.
The NH3 test environment was essentially a
cracked ammonia, which was dissociated into
H2 and N2. Thus, the nitridation potential ( pN2 )
in the NH3 test environment (0.25 atm) was
much lower than that in the N2 test environment
(1.0 atm). As a result, the N2 test environment
was found to produce more severe nitridation
attack for most of the alloys tested (Table 4.24
and Fig. 4.39).
4.8 Summary
(b)
Fig. 4.37
200 m
Optical micrographs showing a through-thickness

nitrided Fe-20Cr-10Ni-1.7Si-Ce alloy 253MA (a)
and a better nitridation resistant Ni-22Cr-9Mo-18Fe-0.6W alloy X
after exposure to 100% N2 at 1090 C (2000 F) for 168 h.
Courtesy of Haynes International, Inc.
Nitridation behavior of metals and alloys in

(a) air, (b) gas-turbine combustion gas, (c)
NH3-H2O, (d) NH3, and (e) N2 environments
is reviewed. Nitridation attack can occur in air
and oxidizing, combustion environments. Under
certain conditions, alloys can suffer oxidation/
nitridation attack. Internal nitridation attack is
much more prevalent in a high-velocity combustion gas stream with thermal cycling. In NH3H2O environments, alloys appear to behave
differently under nitridation attack. Extensive
review is carried out on the behavior of metals
pg 92
30/10/2007 3:57PM Plate # 0
pg 93
6000
Intragranular
1000 C (1830 F)
1100 C (2010 F)
309S
4000
kp, m2/h
310S
153MA
2000
RA330
AC66
800
253MA
353MA
IN601
0
0.200
0.400
0.600
0.800
1.000
XNi'
(a)
12
Intergranular
1000 C (1830 F)
1100 C (2010 F)
153MA
kp, 103 m2/h
253MA
800
309S
310S
RA330
AC66
353MA
IN601
0.200
0.400
0.600
0.800
1.000
XNi'
(b)
Fig. 4.38
Nitridation rate constants as a function of the alloys nickel concentration when tested in N2-5%H2 at 1000 and 1100 C
(1830 and 2010 F). Source: Ref 51
and alloys in NH3 and N2 environments. Comparative resistance to nitridation attack for a wide
variety of alloys is presented.
REFERENCES
1. Metals Handbook, Vol 2, 8th ed., American

Society For Metals, 1964, p 149
2. Metals Handbook, Vol 2, 8th ed., American
Society For Metals, 1964, p 119
3. G.L. Swales, Behavior of High Temperature
Alloys in Aggressive Environments, Proc.
1979 Petten International Conference,

I. Kirman et al., Ed., The Metals Society,
London, 1980, p 45
4. K. Rorbo, Environmental Degradation
of High Temperature Materials, Series 3,
No. 13, Vol 2, The Institution of Metallurgists, London, 1980, p 147
5. R.N. Shreve, The Chemical Process Industries, McGraw-Hill, 1956
6. J.M.A. Van der Horst, Corrosion Problems
in Energy Conversion and Generation,
C.S. Tedmon, Jr., Ed., The Electrochemical
Society, 1974
30/10/2007 3:57PM Plate # 0
pg 94
Table 4.24 Nitrogen absorption in NH3 and N2

environments at 1090 C (2000 F) for 168 h
Alloy
214
600
230
617
601
X
188
150
556
RA330
800H
253MA
NH3
N2
0.2
0.3
1.5
1.9(a)
2.6(a)
3.8(a)
2.0(a)
4.1
4.2(a)
3.1(a)
5.5
6.3(a)
0.0
1.1
2.7
5.1(a)
7.2(a)
6.0(a)
3.7(a)
9.0(a)
9.0(a)
6.6(a)
10.3(a)
10.0(a)
(a) Nitrided all the way through specimens. Source: Ref 41, 50
13
1090 C (2000 F) / 168 h
N2
12
(a)
200 m
(b)
200 m
NH3
10
20
40
60
80
Ni or Ni + Co, wt%
Fig. 4.39
Nitrogen absorption as a function of Ni + Co content in the alloy for 100% NH3 and 100% N2
environments at 1090 C (2000 F) for 168 h. Source: Ref 41, 50
7. M.B. Bever and C.F. Floe, Source Book on

Nitriding, American Society For Metals,
1977, p 125
8. B.J. Lightfoot and D.H. Jack, Source Book
on Nitriding, American Society For Metals,
1977, p 248
9. K.N. Strafford, Corros. Sci., Vol 19, 1979,
p 49
10. T. Rosenquist, Principles of Extractive
Metallurgy, McGraw-Hill, 1974
Fig. 4.40
Optical micrographs showing both chromium

nitride and aluminum nitride (needle-shaped
phase) formed in alloy 601 after exposure to (a) 100% NH3 and (b)
100% N2 at 1090 C (2000 F) for 168 h. Courtesy of Haynes
International, Inc.
11. T. Masumoto and Y. Imai, J. Jpn. Inst. Met.,

Vol 33, 1969, p 1364
12. H.J. Christ, S.Y. Chang, and U. Krupp,
Thermodynamic Characteristics and Numerical Modeling of Internal Nitridation of
Nickel Base Alloys, Mater. Corros., Vol 54
(No. 11), 2003, p 887
30/10/2007 3:57PM Plate # 0
13. N.S. Corney and E.T. Turkdogan, The Effect

of Alloying Elements on the Solubility
of Nitrogen in Iron, J. Iron Steel Inst., Aug
1955, p 344
14. D.H. Jack and K.H. Jack, Carbides and
Nitrides in Steel, Mater. Sci. Eng., Vol 11,
1973, p 1
15. H.A. Wriedt and O.D. Gonzalez, Trans.
AIME, Vol 221, 1961, p 532
16. J.F. Eckel, T.P. Floridis, and B.N. Ferry,
Nitrides in Type 304 Stainless Steel, Virginia
J. Sci., Vol 17, 1966, p 325
17. J.F. Eckel and T.B. Cox, J. Mater., Vol 3,
1968, p 605
18. L.E. Kindlimann and G.S. Ansell, Kinetics
of the Internal Nitridation of Austenitic
Fe-Cr-Ni-Ti Alloys, Metall. Trans., Vol 1,
1970, p 163
19. A.J. Heckler and J.A. Peterson, The Effect
of Nickel on the Activity of Nitrogen in
Fe-Ni-N Austenite, Trans. Metall. Soc.
AIME, Vol 245, 1969, p 2537
20. J. Litz, A. Rahmel, M. Schorr, and J. Weiss,
Scale Formation on the Ni-Base Superalloys
IN 939 and IN 738LC, Oxid. Met., Vol 32,
1989, p 167
21. M.A. Harper, J.E. Barnes, and G.Y. Lai,
Long-Term Oxidation Behavior of Selected
High Temperature Alloys, Paper No. 132,
Corrosion/97, NACE International, 1997
22. G.Y. Lai, unpublished results, 2003
23. S. Han and D.J. Young, Simultaneous
Internal Oxidation and Nitridation of NiCr-Al Alloys, Oxid. Met., Vol 55, 2001,
p 223
24. D.L. Douglass, Anomalous Behavior
During Internal Oxidation and Nitridation,
JOM, Nov 1991, p 74
25. R.P. Rubly and D.L. Douglass, Oxid. Met.,
Vol 35, 1991, p 269
26. R.P. Rubly and D.L. Douglass, Internal
Nitridation of Ni-Cr-Al Alloys, Proc. Int.
Symp. On Solid-State Chemistry of Advanced Materials: High-Temperature Corrosion Workshop, 1992
27. H.J. Grabke and E.M. Peterson, Scr. Met.,
Vol 12, 1978, p 1111
28. J.-W. Park and C. J. Alstetter, Metall. Trans.
A, Vol 18A, 1987, p 43
29. P.L. Gruzin, Y.A. Polikarpov, and G.B.
Federov, Fiz. Metal. I Metalloved., Vol 4
(No. 1), 1957, p 94
30. K.G. Brickner, G.A. Ratz, and R.F.
Domagala, Creep-Rupture Properties of
Stainless Steels at 1600, 1800, and 2000 F,
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
Advances in the Technology of Stainless

Steels and Related Alloys, STP 369, ASTM,
1965, p 99
M. Yu, R. Sandstrom, B. Lehtinen, and
C. Westman, Scand. J. Metall., Vol 16, 1987,
p 154
V. Guttmann and R. Burgel, CreepStructural Relationship in Steel Alloy 800H
at 9001000 C, Met. Sci., Vol 17, 1983,
p 549
M. Welker, A. Rahmel, M. Schutze, Oxidation and Nitridation of Alloy 800H at a
Growing Creep Crack and for Unstressed
Samples, Metall. Trans. A, Vol 20A, 1989,
p 1541
J.J. Hoffman and G.Y. Lai, Paper No. 5402,
Corrosion 2005, NACE International,
2005
V.P. Swaminathan and S.J. Lukezich,
Degradation of Transition Duct Alloys in
Gas Turbines, Advanced Materials and
Coatings for Combustion Turbines, Proc.
ASM 1993 Materials Congress Materials
Week (Pittsburgh, PA), Oct 1721, 1993,
V.P. Swaminathan and N.S. Cheruvu, Ed.,
ASM International, 1994, p 99
G.Y. Lai, Nitridation of Several Combustor
Alloys in a Simulated Gas Turbine Combustion Environment, Advanced Materials
and Coatings for Combustion Turbines,
Proc. ASM 1993 Materials Congress
Materials Week (Pittsburgh, PA), Oct
1721, 1993, V.P. Swaminathan and
N.S. Cheruvu, Eds., ASM International,
1994, p 113
G.Y. Lai, Nitridation Attack in a Simulated
Gas Turbine Combustion Environment,
Materials for Advanced Power Engineering,
Part II, D. Coutsouradis et al., Ed., Kluwer
Academic Publishers, The Netherlands,
1994, p 1263
Y.M. Park and R.E. Sonntag, Int. J. Energy
Res., Vol 14, 1990, p 153
H.J. Grabke, S. Strauss, and D. Vogel,
Nitridation in NH3-H2O Mixtures, Mater.
Corros., Vol 54 (No. 11), 2003, p 895
J.J. Barnes and G.Y. Lai, High Temperature
Nitridation of Fe-, Ni-, and Co-base Alloys,
Corrosion & Particle Erosion at High
Temperatures, Proc. TMS-ASM Symposium, V. Srinivasan and K. Vedula, Ed., The
Minerals, Metals & Materials Society, 1989,
p 617
pg 95
30/10/2007 3:57PM Plate # 0
42. K.M. Verma, H. Ghosh, and J.S. Rai, Brit.

Corros. J., Vol 13 (No. 4), 1978, p 173
43. V. Cihal, Corrosion Mechanisms in
Ammonia Synthesis Equipment, Conf.
Proc., First International Congress on
Metallic Corrosion (London, U.K.), April
1015, 1961, L. Kenworthy, Ed., Butterworths, London, 1962, p 591
44. J.J. Moran, J.R. Mihalisin, and E.N. Skinner,
Corrosion, Vol 17 (No. 4), 1961, p 191t
45. D.W. McDowell, Jr., Mater. Protect., Vol 1
(No. 7), 1962, p 18
46. K.H. Jack, High Temperature Gas-Metal
Reactions
in
Mixed
Environments,
S.A. Jansson and Z.A. Foroulis, Ed., The
Metallurgical Society of AIME, 1973,
p 182
47. H. Schenck, M.G. Frohberg, and F.
Reinders, Stahl Eisen, Vol 83, 1963, p 93
48. G.D. Smith and P.J. Bucklin, Some

Observation on the Performance of
Nickel-Containing Commercial Alloys in
Nitrogen-Based Atmospheres, Paper No.
49. P. Ganesan and G.D. Smith, Performance of
Selected Commercial Alloys in Nitrogen
Based Sintering Atmospheres, Paper
No. 278, Corrosion/90, NACE, 1990
50. J.J. Barnes and G.Y. Lai, Factors Affecting
the Nitridation Behavior of Fe-Base, NiBase and Co-Base Alloys in Pure Nitrogen,
J. Physique IV, Colloque C9, supplemental
au Journal de Physique III, Vol 3, 1993,
p 167
51. K. Tjokro and D.J. Young, Comparison of
Internal Nitridation Reactions in Ammonia
and in Nitrogen, Oxid. Met., Vol 44, 1995,
p 453
pg 96

DOI: 10.1361/hcma2007p097
26/10/2007 12:21PM Plate # 0

CHAPTER 5
Carburization and Metal Dusting

5.1 Introduction
Metals and alloys are susceptible to carburization when exposed to an environment containing CO, or CH4 or other hydrocarbon gases,
such as ethane (C2H6), propane (C3H8), and so
forth, at elevated temperatures. Carburization
attack generally results in formation of internal
carbides, which often cause the alloy to suffer
embrittlement as well as other mechanical
property degradation.
Carburization problems are quite common to
heat treating equipment, particularly furnace
retorts, baskets, fans, and other components used
for case hardening of steels by gas carburizing.
A common commercial practice for control of
gas carburizing is to use an endothermic gas as a
carrier enriched with one of the hydrocarbon
gases, such as CH4, C3H8, and so forth (Ref 1).
An endothermic gas enriched with about 10%
natural gas (CH4) is a commonly used atmosphere (Ref 2). The typical endothermic gas
consists of 39.8% N2, 20.7% CO, 38.7% H2, and
0.8% CH4, with a dew point of 20 to 4 C
(5 to +25 F) (Ref 1). Gas carburizing occurs
typically at 840 to 930 C (1550 to 1700 F).
Furnace equipment and components repeatedly
subjected to these service conditions frequently
suffer brittle failures as a result of carburization
attack.
In the petrochemical industry, carburization is
one of the major modes of high-temperature
corrosion for processing equipment. The pyrolysis furnace tubes for production of ethylene and
olens are a good example (Ref 35). Ethylene is
formed by cracking petroleum feedstock, such as
ethane and naphtha, at temperatures up to 1150
C (2100 F). This generates a strong carburizing
gas stream inside the tubes. As a result, carburization was found to be a major mode of tube
failure in a survey of ethylene and olen pyrolysis furnaces conducted by Moller and Warren
(Ref 3).
Production of carbon bers also generates

carburizing atmospheres in a furnace. As a result,
the furnaces retorts, xtures, and other components require frequent replacement because of
carburization attack.
Metal dusting, a form of catastrophic carburization, can occur at intermediate temperatures
when a process gas stream consists of primarily
H2, CO, and CO2 along with some hydrocarbons
with high carbon potentials (ac > 1). Metals or
alloys can suffer rapid metal wastage in a form of
pitting or general thinning of the cross-sectional
thickness of a metallic component. Metal dusting
typically occurs at temperatures between 430
and 900 C (800 and 1650 F). Materials failures
associated with metal dusting have been encountered in rening and petrochemical processing, such as production of syngas in hydrogen,
ammonia, and methanol plants, heat treating, and
other industrial processes (Ref 510).
5.2 Carburization
5.2.1 CarburizationThermodynamic
Considerations
Whether an alloy is likely to be carburized or
decarburized depends on the carbon activity (ac)
in the environment and that of the alloy. The
thermodynamic condition that dictates either
carburization or decarburization can be described
simply.
The alloy is likely to be carburized when:
(ac )environment 4 (ac )metal
The alloy is likely to be decarburized when:

(ac )environment 5 (ac )metal
Thus, in order to predict whether an alloy

will be carburized, one needs to know the
carbon activities of both the environment and
the alloy.
pg 97
26/10/2007 12:21PM Plate # 0
pg 98
Carburization can proceed by one of the following reactions when the environment contains
CH4, CO, or H2 and CO:
CO+H2 =C+H2 O
Eq 5:1
2CO=C+CO2
Eq 5:2
CH4 =C+2H2
Eq 5:3
Assuming that carburization follows Reaction

5.1, the carbon activity in the environment can be
calculated by:
DG =7RT ln
ac pH2 O
pCO pH2
Eq 5:4
ac =e7DG
=RT
pCO pH2
pH2 O
DG =7RT ln
ac =e7DG
=RT
Eq 5:5
p2CO
pCO2
Eq 5:6

Eq 5:7
pCH4
p2H2
Eq 5:8
Carbon activities as a function of ( pCH4 =p2H2 ) are

plotted in Fig. 5.3.
Reactions 5.1 and 5.2 have a similar characteristic, showing lower carbon activities with
increasing temperature (Fig. 5.1 and 5.2).
H2 = C + H2O
Carbon activity (ac) as a function of gaseous composition in terms of (pCO pH2 =pH2 O ) ratios based on
Eq 5.1 for various temperatures. Also plotted are carbon activities
for carbon steel (in equilibrium with Fe3C), and for 2.25Cr-1Mo
and austenitic stainless steels (both measured ac).

ac pCO2
p2CO
Plots of carbon activities as a function of gas

compositions in terms of (p2CO =pCO2 ) for various
temperatures are shown in Fig. 5.2.
When carburization follows Reaction 5.3, the
carbon activity in the environment is:
ac =e
From Eq 5.5, one can construct graphs of carbon

activity as a function of gaseous composition
in terms of ( pCO pH2 =pH2 O ) ratios for various
temperatures, as shown in Fig. 5.1.
Fig. 5.1
7DG =RT
Rearranging the equation changes it to:

Similarly, if carburization follows Reaction

5.2, the carbon activity of the environment can
also be calculated:
2CO = CO2 + C
Fig. 5.2
Carbon activity (ac) as a function of gaseous com2

position in terms of (pCO
=pCO2 ) based on Eq 5.2 for
various temperatures. Also plotted are carbon activities for carbon
steel (in equilibrium with Fe3C), and for 2.25Cr-1Mo and austenitic stainless steels (both measured ac).
26/10/2007 12:21PM Plate # 0
Chapter 5:
Reaction 5.3, on the other hand, shows increased

carbon activities with increasing temperature
(Fig. 5.3). If the environment contains CH4,
the carbon activity of the environment at higher
temperatures is likely to be dominated by Reaction 5.3. When no CH4 is present in the
environment, Reaction 5.1 and/or 5.2 will dictate
the carbon activity.
The carbon activity maps shown in Fig. 5.1 to
5.3 were previously described by Mazandarany
and Lai (Ref 11) in assessing the carburizationdecarburization behavior of alloys in hightemperature gas-cooled helium environments
containing H2, CO, CO2, CH4, and H2O. These
activity maps provide a simple means of estimating an environments carbon activity for
predicting whether or not the environment is
thermodynamically capable of carburizing an
alloy.
When the gas stream contains many gaseous
components, such as H2, CO, CO2, CH4, and
pg 99
Carburization and Metal Dusting / 99
H2O, and under very dynamic conditions with a

high gas velocity such that the gaseous components do not have time to react to reach a
thermodynamic equilibrium (i.e., nonequilibrium
conditions), the gas-metal reaction can be reasonably assumed to follow the dominating
reaction from one of those shown in Reaction
5.1, 5.2, or 5.3, and thus, one of the activity maps
shown in Fig. 5.1, 5.2, or 5.3.
Data for carbon activities of commercial alloys
at temperatures below 1200 C (2190 F) is very
limited. Natesan (Ref 12) reported that ac for
2.25Cr-1Mo steel is in the range of 1101 to
102 from 550 to 750 C (1020 to 1380 F).
Natesan and Kassner (Ref 13) reported the
carbon activities of Fe-18Cr-8Ni alloys. These
values are superimposed in Fig. 5.1 to 5.3.
For carbon steels, carbon activity can be estimated by assuming that it is in equilibrium with
cementite (Fe3C):
Eq 5:9
3Fe+C=Fe3 C
DG =7RT ln
DG =7RT ln
aFe3 C
(ac ) (aFe )3

1
ac

Eq 5:10
Eq 5:11
where aFe3 C and aFe are assumed to be unity.

ac =eDG
CH4 = 2H2 + C
Fig. 5.3
Carbon activity (ac) as a function of gaseous com2

position in terms of (pCH4 =pH
) based on Eq 5.3 for
2
various temperatures. Also plotted are carbon activities for carbon
steel (in equilibrium with Fe3C), and for 2.25Cr-1Mo (measured
ac). Carbon activities of austenitic stainless steels are below 102 at
800 to 1000 C (1470 to 1830 F).
=RT
Eq 5:12
The ac values for carbon steel based on Eq 5.12

are plotted in Fig. 5.1 to 5.3. Using Fig. 5.1, 5.2,
or 5.3 one can make a quick determination as to
whether an environment has a carbon potential
(or activity) high enough to carburize the alloy of
interest.
Even though in cases where the gas mixture
may not reach an equilibrium condition, it will be
of great benet to better understand the gas-metal
reaction in terms of the thermodynamic equilibrium condition in multicomponent gases
environments. The thermodynamic equilibrium
gaseous composition along with its thermodynamic potentials, such as carbon activity (ac),
oxygen potential ( pO2 ), and other potentials, can
be determined using a commercial software
program such as HSC Chemistry for Windows
(Ref 14) and ChemSage (Ref 15).
The environment can also be characterized in
terms of ac and pO2 to determine the relative
severity of its carburization potential. Both carbon activity and oxygen potential can be calculated by a computer program. The environment
26/10/2007 12:21PM Plate # 0
Log PCO /PCO
During carburization, the relative stabilities of

these carbides can be best described by a stability
diagram, such as the one shown in Fig. 5.5. If the
carbon and oxygen activities of the environment
are in the Cr3C2 region, conditions will favor
formation of Cr3C2 on the surface and/or in the
underlying metal. As carbon diffuses farther into
the alloys interior, carbon activities will be
lowered, thus favoring Cr7C3. Moving even
farther into the interior, carbon activities will
be further reduced, favoring the formation of
Cr23C6. These chromium carbides can incorporate other alloying elements depending on alloy
system. For example, in Fe-Ni-Cr system, iron
with very little nickel can be incorporated into
these chromium carbides. The combined metal
elements in the carbide are then represented by
M, as M3C2, M7C3, and M23C6. An example
of the metallic compositions of M7C3 and M23C6
formed in carburization of Type 304L is
illustrated in Fig. 5.6 (Ref 20). Both M7C3 and
M23C6 contain essentially Cr and Fe with
negligible amount of Ni.
For many high-temperature alloys, particularly superalloys, there are other alloying
elements, such as Ti, Ta, Nb (or Cb), Mo, and W,
that can form carbides. The carbides of these
alloying elements are important to the physical
metallurgy of high-temperature alloys in
that they provide an important strengthening
mechanism. A general review of binary metallic
Log PCH /PH 2
4
2
can then be presented in a stability diagram of

a metal-carbon-oxygen system. The stability
diagrams of Fe-C-O and Cr-C-O systems are
shown in Fig. 5.4 and 5.5 (Ref 16). From the
stability diagram, the possible phases that the
alloy may form at the gas/metal interface can be
predicted. As the activities of both carbon and
oxygen are decreasing from the gas/metal interface to the metal interior, the possible phases
that the alloy may form beneath the gas/metal
interface can also be predicted.
For carbon and alloy steels with low concentrations of chromium, ingress of carbon into
the metal or alloy may result in the formation of
iron carbides. Several forms of iron carbides
have been reported (Ref 17), with compositions
ranging from Fe4C to Fe2C. They are phase
(Fe4C), phase (Fe3C), phase (Fe2.2C), and
phase (Fe2-3C). Fe3C (cementite) is the most
stable iron carbide. Other iron carbides are less
stable. Browning et al. (Ref 18) found that
phase, which formed by carburizing Fe with
butane at 275 C (530 F), was converted to
Fe3C when heated to 500 C (930 F). The
phase (Fe2-3C) is a transition phase that forms in
martensite during tempering of steel (Ref 19).
In ferritic and austenitic stainless steels and
nickel- and cobalt-base alloys, ingress of carbon
into the alloy results in the formation of mainly
chromium carbides. There are three forms of
chromium carbides: Cr23C6, Cr7C3, and Cr3C2.
Log PCO /PCO 18 16 14 12 10 8 6 4 2

2
Log P
/P 18 16 14 12 10 8 6 4 2
H2O
H2
10 12
10 12
0
2
C(s)
2
0
Fe3C(s)
Log aC
6
4
6
8
10
8
Fe(s)
Fe3C4(s)
6
10
Fe0.95O(s)
8
12
10
12
14
14
16
16
12
14
50 45 40 35 30 25 20 15 10
Log PO , atm
2
Fig. 5.4
Fe2O3(s)
Stability diagram of Fe-C-O system at 870 C (1600 F). Source: Ref 16
pg 100
26/10/2007 12:21PM Plate # 0
pg 101
H2
Cr3C2(s)
0
2
10 12 14
10 12 14
C(s)
Cr7C3(s)
4
Cr23C6(s)
Log aC
6
8
Cr2O3(s)
10
H2O
Log PCO /PCO 14 12 10 8 6 4 2 0 2 4

2
14 12 10 8 6 4 2 0 2 4
Log P
/P
Log PCO /PCO
Log PCH /PH2
Chapter 5: Carburization and Metal Dusting / 101
10 10
Cr(s)
12
12 12
14
14 14
16 16
16
50 45 40 35 30 25 20 15 10
Log PO , atm
(a)
H2O
H2
0
2
8 10 12
8 10 12
0
2
Cr3C2(s)
C(s)
2
0
Cr7C3(s)
4
2
Log aC
Cr23C6(s)
8
6
8
10
Log PCH /PH 2
Log PCO /PCO 18 16 14 12 10 8 6 4 2

2
18 16 14 12 10 8 6 4 2
Log P
/P
4
2
2
Log PCO /PCO
2
10
Cr2O3(s)
Cr(s)
8
12
10
12
14
14
16
12
14
16
50 45 40 35 30 25 20 15 10
(b)
Fig. 5.5
Log PO , atm
2
Stability diagrams of Cr-C-O system at (a) 620 C (1150 F), (b) 870 C (1600 F), and (c) 1090 C (2000 F). Source: Ref 16
carbides can be found elsewhere (Ref 21, 22).

The relative stabilities of some binary carbides
are shown in Fig. 5.7 (Ref 22).
When the environment contains oxygen and
carbon activities, temperature is an important
factor in determining whether the oxide or carbide will be thermodynamically stable. Considering a chemical reaction, such as 3Cr2O3 +
4C = 2Cr3C2 + 9/2 O2, Cr3C2 will remain stable
when the reaction goes from left to right. In order
to keep the reaction going from left to right (i.e.,

keeping Cr3C2 stable), the pO2 of the environment shall be lower than the equilibrium pO2
associated with the above reaction. On the other
hand, if the pO2 of the environment is higher than
the equilibrium pO2 associated with the above
reaction, Cr2O3 will become stable. Temperature
can be a signicant factor in determining whether
chromium oxide or chromium carbide is stable,
and thus signicantly affects the carburization
26/10/2007 12:21PM Plate # 0
H2
Cr3C2(s)
10
10
4
2
C(s)
Cr7C3(s)
4
0
6
Cr23C-6(s)
4
Log aC
2
2
2
8
10
Cr2O3(s)
Cr(s)
6
12
10
8
14
12
10
16
14
12
18
16
50 45 40 35 30 25 20 15 10
Concentration, wt.%
1123 k
150 h
80
M23C6
Cr
40
Fe
Ni
M7C3
0
0.2
(c) 1090 C (2000 F). Source: Ref 16
100
Fig. 5.5 (continued)
20
Log PO , atm
(c)
60
H2O
0
0
Log PCO /PCO 20 18 16 14 12 10 8 6 4 2

2
Log P
/P 20 18 16 14 12 10 8 6 4 2
Log PCO /PCO
Log PCH /PH 2
0.4
0.6
0.8
Distance to surface, mm
Fig. 5.6
Compositions of the metallic components of M7C3

and M23C6 formed in Type 304L after carburizing at
1123 K (850 C) in H2-2.6CH4 (ac = 0.9) for 150 h. Source: Ref 20
behavior of an alloy. This can be nicely illustrated in a plot that contains the oxygen potentials of the environment (Boudouard reaction
[2CO = C + CO2] is used for this example) and
those in equilibrium with Cr2O3/Cr3C2 as a
function of temperature. This is shown in Fig. 5.8
(Ref 23). The gure shows the equilibrium pO2
line of Cr2O3/Cr3C2 intersecting with the environments pO2 lines (pCO = 1.0 bar, pCO =
0.5 bar, and pCO = 0.25 bar). The intersections
are between 1000 and 1200 C. On the right side
of the intersections (i.e., lower temperatures), pO2
(environment) is greater than pO2 (Cr2O3/Cr3C2).
Thus, Cr2O3 is stable. On the left side of the

intersections (i.e., higher temperatures), pO2
(environment) is lower, thus favoring the formation of carbides, but not oxides. Accordingly,
higher temperatures favor carburization thermodynamically and lower temperatures favor
formation of oxides thus retarding carburization.
Nishiyama et al. (Ref 24) examined the effect
of the temperature on the stability of chromium
oxide and carbides based on the ethylene
pyrolysis environment, which is generated by the
reaction of naphtha with steam for the production
of ethylene (C2H4) and propylene (C3H6) at
temperatures of approximately 900 to 1100 C
(1650 to 2012 F). In their calculation of the
oxygen potential for the reaction of naphtha with
steam, three naphtha feedstocks were used with
the steam/naphtha weight ratios of 0.4 and 0.5.
The calculated pO2 for the pyrolysis environment
as a function of temperature is plotted in Fig. 5.9.
Also plotted in Fig. 5.9 are pO2 values in equilibrium with Cr3C2/Cr2O3 and those in equilibrium with Cr7C3/Cr2O3. The results in Fig. 5.9
are very similar to those shown in Fig. 5.8, where
the environment was calculated from CO-CO2
reaction (Boudouard reaction). The calculation
by Nishiyama et al. showed that the chromium
oxide was stable up to 1030 to 1040 C (1886
to 1904 F) and became unstable above those
temperatures. At temperatures above 1030 to
pg 102
26/10/2007 12:21PM Plate # 0
Fig. 5.7
Standard free energies of formation for carbides. Source: Ref 22
1040 C (1886 to 1904 F), carbides such as

Cr3C2 and Cr7C3 became stable. This oxidecarbide transition temperature can vary depending on the steam/naphtha ratio, as illustrated in
Fig. 5.10 (Ref 24). As shown in the gure, when
the steam/naphtha ratio is decreased to 0.1 from
the common operating ratios of 0.35 to 0.5, the

oxide-carbide transition temperature is decreased
to about 970 C (1778 F).
In Reactions 5.1 to 5.3, carbon deposition
(coking) can occur when the carbon activity (ac)
in the environment is greater than 1.0 (ac = 1
pg 103
26/10/2007 12:22PM Plate # 0
in equilibrium with graphite). Many laboratory

carburization tests have been conducted in
H2-CH4 mixtures. Figure 5.11 shows the carbon
activities of H2-1%CH4 and H2-2%CH4 as
a function of temperature (Ref 25). Grabke
600
800
1000
Cr3C2/Cr2O3
10
PCO1 bar
Cr3C2
15
PCO0.5 bar
PCO0.25 bar
Log PO , bar
1200
1800
1600
1400
T, C
(Ref 25) recommended that the H2-1%CH4

mixture with ac <1.0 be used for laboratory
testing to correspond to the industrial process that
does not form coking. He also recommended the
use of the H2-2%CH4 mixture with ac > 1 for
testing to correspond to the processes where
coking is taking place. The test temperature is
thus recommended to be higher than 1000 C
(1832 F) (see Fig. 5.11). Ethylene pyrolysis
environment is known to develop coking on
the internal surface of the pyrolysis furnace
tubes. The deposition of carbon is the result of
decomposition of ethylene (C2H4) in the reaction
as described in Eq 5.13 (Ref 26). Signicant
coking in the internal surface of the ethylene
pyrolysis furnace tube and the repeated decoking
20
C/CO-CO2
Cr2O3
25
5.104
75.104
10.104
12.5.104
1/T, K4
Fig. 5.8
The oxygen potentials of the environment based on

the Boudouard reaction (2CO = C + CO2) and those
in equilibrium with Cr2O3/Cr3C2 as a function of temperature.
Source: Ref 23
T, C
0,150 1,100 1,050 1,000
17
900
850
Equilibrium pO2 of the environment based on

ethylene pyrolysis of naphtha I with various steam/
naphtha weight (S/O) ratios as a function of temperature, and pO2
in equilibrium with Cr3C2/Cr2O3 and Cr7C3/Cr2O3 as a function of
temperature. Source: Ref 24
Fig. 5.10
Cr-carbides + metastable Cr2O3
18
Cr2O3+ metastable Cr-carbides

19
ac
Log PO .atm
950
20
)
=1
(a c
3
)
r 2O
=1
/C
C2
(a c
3
Cr 3
r 2O
/C
C3
Cr 7
II (0
Na
.5)
pht
21
Cr-carbides
22
7.5
ha
and
III (
.4)
8.5
2% CH4
0.5
I (0
Carbon activity in H2-CH4

1 bar
1/T 104, 1/K

Equilibrium pO2 of the environment based on ethylene pyrolysis of naphtha with steam/naphtha ratios
of 0.4 and 0.5 (in parentheses) as a function of temperature,
and pO2 in equilibrium with Cr3C2/Cr2O3 and Cr7C3/Cr2O3 as a
function of temperature. Naphtha I, II, and III represent different
naphtha feed stocks. Source: Ref 24
1% CH4
2
1
Fig. 5.9
800
Fig. 5.11
Ref 25
850
900
950 1000 1050 1100 1150 C
Carbon activities (ac) of H2-1%CH4 and H2-2%CH4

are plotted as a function of temperature. Source:
pg 104
26/10/2007 12:22PM Plate # 0
operation to remove this coke deposit can

signicantly degrade the tube life (Ref 27).
C2 H4 =2C+2H2
Eq 5:13
5.2.2 Resistance to Carburization

Carburization attack generally results in the
formation of internal carbides in the alloy matrix
as well as at grain boundaries. The gravimetric
method has been widely used for studying carburization kinetics. This method can sometimes
produce a misleading result when the environment exhibits an oxygen potential high enough to
form oxides of some active alloying elements.
The weight gain, in this case, is the result of both
carbon ingress and oxide formation. Measurements of carburization depth have also been used
by some investigators. Different alloy systems
can produce signicant differences in the concentration prole for the carburized layer. Thus,
one alloy may exhibit a large carburization depth
with only a slight concentration gradient, while
another alloy may show a narrow carburization
depth with a steep concentration prole. Furthermore, measurements of carburization depth
by the metallographic method can be difcult
when separating the carbides formed by carburization from those formed by thermal aging.
Some investigators measured the total amount of
carbon in the alloy after the exposure. The measurement of the carbon concentration prole as
a function of distance from the metal surface
may be an excellent method for characterizing
Carbon content, %
4
B
2
C
1
0
ID
Distance from bore of tube, mm
OD
Wall thickness
Fig. 5.12
Three possible carbon concentration profiles of

carburized alloys. Source: Ref 28
the carburized alloy. Each evaluation method

has its merit. It certainly will be benecial to
use as many evaluation methods as possible for
characterizing the carburized alloy.
With respect to the impact of carburization on
an alloys performance, Krikke et al. (Ref 28)
believed that not only the total amount of carbon
absorbed but also the maximum carbon level and
the maximum carbon concentration gradient are
the most important factors. Figure 5.12 illustrates
three possible carbon concentration proles,
as suggested by Krikke et al. (Ref 28). They
considered prole A with a steep concentration
prole to be the most damaging. Heubner (Ref
29) tested various commercial alloys in H2-CH4
gas mixture (ac = 0.8) at 1000 C (1832 F) and
observed a steep carbon concentration prole in
Fe-Ni-Cr alloys and a low at concentration
prole in Ni-Cr alloys, as shown in Fig. 5.13.
One Ni-Cr alloy (alloy 45TM), which contained
relatively high Fe and high Si, was an exception
showing a steep carbon concentration prole
similar to Fe-Ni-Cr alloys such as 800H, AC66,
and DS. The Fe-Ni-Cr alloys were found to have
suffered more room-temperature impact toughness drop in general than Ni-Cr alloys (Ref 29).
However, the relative room-temperature impact
toughness loss (%) was found to increase with
increasing total carbon pickup (Ref 29). For
carburization, the real issue is the effect of carburization on the alloys mechanical properties,
such as creep-rupture properties and toughness
or ductility. This type of data, however, is quite
limited and is inadequate as a basis for making an
informed materials selection. Thus, this chapter
reviews mainly the carburization data in terms of
mass gain, mass of carbon absorption, carburization depth, and concentration prole of the
carburized layer.
When the environment is such that no protective oxide scale (e.g., Cr2O3 scale) is formed
on the metal surface, carburization is controlled
by diffusivity and solubility (Ref 30). The ingress
of carbon will be greatly reduced when a chromium oxide scale is developed. Carburization
kinetics in this case are then controlled by the
diffusion of carbon through the oxide scale. Wolf
and Grabke (Ref 31) demonstrated that there was
no detectable solubility of carbon in Cr2O3 oxides by equilibrating the oxides with CO2-CO
mixtures tagged with radioactive 14C at 1000 C
(1832 F). Thus, carbon permeation is not possible through the perfectly dense Cr2O3 oxide
layer, unless the oxide layer contains pores and
ssures (Ref 31).
pg 105
26/10/2007 12:22PM Plate # 0
alloy compared with the as-cast surface nish

(Fig. 5.17).
Norton and his colleagues (Ref 3537) at
Petten Laboratories have conducted a series of
studies on the effects of silicon, niobium, chromium, and iron in Fe-Ni-Cr alloys, including
four commercial alloys (HK-40, HP-40Nb, Type
314 SS, and alloy 800H) and three experimental
alloys. Their test environments had xed carbon
activities (ac) of 0.3 and 0.8, with various oxygen
potentials (pO2 ) at temperatures from 825 to
1050 C (1520 to 1920 F), as illustrated in
Fig. 5.18. The oxygen potentials of the test
environments were below that in equilibrium
with Cr2O3 (Fig. 5.18). That means that no
chromium oxide scales should have formed on
the metal surface. However, a SiO2 scale was
likely to form at 825 C (1520 F), but not at
1000 C (1830 F). The test results at 825 C
(1520 F) showed that Type 314 stainless steel
(2.04% Si) was signicantly more carburization
Resistance to carburization is an important

factor in the performance of pyrolysis furnace
tubes as well as pigtails for ethylene and olen
plants. Furnace tubes are typically constructed of
Fe-Ni-Cr cast alloys, such as HK (Fe-25Cr-20Ni
or 25/20), HP (Fe-25Cr-35Ni or 25/35) and their
variants. Some of these variants involved additions of niobium (or columbium), tungsten,
molybdenum, silicon, and titanium. Some also
involved increases in nickel and/or chromium.
Some of the modied alloys are referred to as
microalloyed castings. It has been found that
these additions and increases improve carburization resistance as well as creep-rupture
strengths. Figures 5.14 to 5.17 illustrate the carburization resistance of some of these modied
alloys compared with HK alloy (Ref 3234).
Also shown in Fig. 5.17 is the effect of the
surface nish on carburization resistance of the
alloy. The machined-nished surface signicantly reduced the carbon ingress into the
3.5
AC 66
Carbon concentration, %
3
2.5
alloy 800 H
alloy DS
1.5
1
0.5
0
0
(a)
2.5
45 TM
Carbon concentration, %
1.5
alloy 600H
0.5
alloy 617
alloy 602 CA
0
0
(b)
Fig. 5.13
Carbon concentration profile for alloys after testing at 1000 C (1832 F) for 1008 h in a H2-CH4 mixture (ac = 0.8) for (a)
Fe-Ni-Cr and (b) Ni-Cr alloys. Source: Ref 29
pg 106
26/10/2007 12:22PM Plate # 0
resistant than HK-40 (1.35% Si), HP-40Nb

(1.29% Si), and alloy 800H (0.4% Si), as shown
in Fig. 5.19. Also revealed by the test results was
that the model alloys were signicantly less
resistant to carburization than the commercial
alloys with the same chromium, nickel, and iron.
The model alloys had much lower silicon levels
(0.130.28%) as well as manganese, aluminum,
titanium, and so forth. The benecial effect of
silicon on carburization resistance was clearly
demonstrated when the data were replotted
(Fig. 5.20). The presence of a SiO2 scale was
conrmed by Van der Biest et al. (Ref 36). When
the test temperature was increased to 1000 C
(1832 F), where the pO2 was below that in
equilibrium of SiO2 (Fig. 5.18), Type 314 SS was
found to be similar to HK-40, alloy 800H, and
HP-40Nb (Fig. 5.21). Under these conditions,
SiO2 was no longer thermodynamically stable.
Thus, the silicon effect was diminished. The
model alloy 50/50 (50Ni-50Cr with very low
silicon level) was among the best performers in
the alloys tested.
The ratio of Ni to Cr + Fe is an important factor
(Ref 35) in governing carburization resistance, as
shown in Fig. 5.22. Decreases in Cr+ Fe in
Fe-Ni-Cr alloys improved carburization resistance. Both chromium and iron are carbide
formers, constituting the major elements in M7C3

and M23C6 carbides resulting from carburization.
Harrison et al. (Ref 37) found that the surface
carbides removed from HK-40 sample after
testing at 1000 C (1830 F) contained 55 to
58% Cr and 41 to 43% Fe, with very little nickel
(approximately Cr4Fe3C3). Nickel reduces the
diffusivity of carbon in Fe-Ni-Cr alloys, as
demonstrated by Demel et al. (Ref 38) in
Fig. 5.23. Nickel also decreases the solubility of
carbon in Fe-Ni alloy system as shown in
Fig. 5.24 (Ref 39). Decreases in carbon diffusivity and solubility can result in increases in
carburization resistance. Grabke et al. (Ref 40)
observed that increasing nickel improved carburization resistance in Fe-Ni-Cr alloys, with the
maximum resistance achieved when the ratio of
Ni to Fe was 4 to 1. This is in general agreement
with the product of carbon solubility and diffusivity (Ref 41). High-nickel alloys are generally
more resistant to carburization than Fe-Ni-Cr
alloys. This is illustrated by the test results of
Klower and Heubner (Ref 29), as shown in
Fig. 5.25. The benecial effect of nickel on carburization resistance can also be clearly revealed
in Fig. 5.26 when the data of Fe-Ni-Cr alloys
Fig. 5.14
Fig. 5.15
Carburization resistance of HK (25Cr-20Ni) and

several HP alloys (Cr/Ni) as a function of temperature in pack carburization tests. Source: Ref 32
Carbon concentration profiles for HK (25Cr-20Ni)

and several HP alloys (Cr/Ni) carburized at 1100 C
(2010 F) for 520 h in pack carburization tests. Source: Ref 32
pg 107
26/10/2007 12:22PM Plate # 0
and Ni-Cr alloys generated at 1000 C for

1008 h were plotted as a function of iron content
(i.e., decreasing nickel content) (Ref 29). The
oxygen partial pressure (pO2 ) of the test environmentwhich was 1027.92, 1025.25,
and 1023.88 for 850, 1000, and 1100 C,
respectivelywas below the pO2 in equilibrium
of Cr2O3 (Ref 29). Accordingly, no chromium
oxide scale would be present to provide protection for the alloys in these tests.
In an extensive study undertaken by Steel and
Engel (Ref 42) on the relative inuence of nickel
and chromium on the carburization resistance of
Fe-Ni-Cr alloys, standard heats of ASTM grades
varying from HC to HX cast alloys were
investigated, along with many experimental cast
alloys. They found nickel to be benecial, as
illustrated in Fig. 5.27. The role of chromium,
however, appeared to be different for different
levels of nickel, as shown in Fig. 5.28. For ironbase alloys with 25% or less nickel, increasing
chromium signicantly reduced carbon pickup.
A slight decrease in carbon pickup with
increasing chromium was noted for alloys
Fig. 5.16
containing 26 to 45% Ni. For alloys containing

46 to 70% Ni, increasing chromium resulted in an
increase in carbon pickup. In this study, although
there is no mention of the oxygen potential of the
test environment, it is believed that the chromium
oxide scale was not involved in the carburization
reaction. Wolfe (Ref 43) also found that a higher
Ni-containing Fe-Cr-Ni cast alloy, alloy HU
(Fe-19Cr-39Ni-0.5C), was signicantly better
than Type 304 (19Cr-9Ni) and Type 321 (18Cr12Ni) with similar amounts of chromium but
lower nickel. Even a low Cr-containing Fe-Cr-Ni
cast alloy, alloy HT (Fe-15Cr-35Ni-0.5C) was
signicantly better than both Types 304 and 321
(Ref 43). These data are shown in Fig. 5.29.
Small additions of some minor elements,
such as titanium, niobium, tungsten, and rare
earth elements, may also improve an alloys
resistance to carburization in test environments
of H2-8.6%CH4-7%H2O and H2-12%CH4-10%
H2O. This is illustrated in Table 5.1 (Ref 44).
The superior carburization resistance of TMA
4750 alloy to HK-40 (2% Si) can be attributed to
small additions of titanium, niobium, tungsten,
Carbon concentration profiles of HK and HP alloys tested at 1050 C (1920 F) for 1200 h in 37%N2-40%H2-20%CO-3%
CH4 (ac = 1.0, pO2 = 3:4 10720 atm). Source: Ref 33
pg 108
26/10/2007 12:22PM Plate # 0
and rare earth elements. In both test environments, the oxygen potentials, although not discussed, are believed to be high enough to form
oxide scales. Thus, improvements in carburization resistance may be partly due to the improved
oxide scale.
Silicon has also been found to be very effective in improving carburization resistance. The
benecial effect of silicon on carburization
resistance has been reported by Norton and
Barnes (Ref 35), Steinkusch (Ref 45), Wolfe
(Ref 46), and Van den Bruck and Schillmoller
(Ref 47), and the data are illustrated in Fig. 5.19,
and Fig. 5.30 to 5.32. Kane (Ref 48) investigated
a large number of centrifugally cast tubes of HK
(Fe-25Cr-20Ni) and HP (Fe-25Cr-35Ni) alloys
from four producers. The tubes contained silicon
varying from about 1% to more than 2%. Tests
were conducted at 980 and 1090 C (1800 and
2000 F). A unit carbon activity (ac = 1.0) was
maintained for all the test environments. Oxygen
potentials of the test environments were varied
by injecting different levels of H2O. With no
H2O injection, the environments pO2 was below
that in equilibrium with SiO2 (i.e., SiO2 could not

form). Thus, silicon played no role in carburization resistance. At high oxygen potentials
(i.e., 1% and 10% H2O injections) where SiO2
was stable, silicon improved carburization
resistance.
The focus thus far has been primarily on alloys
used in ethylene cracking and steam hydrocarbon
reforming operations. Most of the alloys are
cast alloys used for furnace tubes. A variety of
wrought alloys of stainless steels, Fe-Ni-Cr
Fig. 5.18
Oxygen potentials of the test environments used by

Norton and his colleagues in carburization studies
at Petten Laboratories. Source: Ref 37
Fig. 5.17
Carbon concentration profiles of several centrifugally cast alloys in (a) the as-cast surface
condition and (b) the machined surface condition after 1 year of
field testing in an ethylene cracking furnace. Source: Ref 34
Fig. 5.19
Carburization rate constants of several Fe-Ni-Cr

alloys at 825 C (1520 F) in the test environment
with a carbon activity of 0.8 and an oxygen potential such that
SiO2 is stable (but not Cr2O3), as shown in Fig. 5.18. Source: Ref 35
pg 109
26/10/2007 12:22PM Plate # 0
Fig. 5.20
Carburization rate constants as a function of silicon content in the alloy for several Fe-Ni-Cr alloys tested at 825 C (1520 F)
in the test environment with a carbon activity of 0.8 and an oxygen potential such that SiO2 is stable (but not Cr2O3), as
shown in Fig. 5.18. Source: Ref 35
Fig. 5.22
Carburization rate constants of several Fe-Ni-Cr

alloys at 1000 C (1830 F) in the test environment
with a carbon activity of 0.8 and an oxygen potential such that
SiO2 is not stable as shown in Fig. 5.18. Source: Ref 35
Carburization rate constants as a function of Ni to

Cr + Fe ratio [Ni/(Cr + Fe)] for several Fe-Ni-Cr
alloys tested at 1000 C (1830 F) in the test environment with a
carbon activity of 0.8 and an oxygen potential such that SiO2 is not
stable as shown in Fig. 5.18. Source: Ref 35
alloys, and Ni-Cr alloys have been widely used in

various industries, including heat treating and
chemical processing. Mason et al. (Ref 49)
investigated various stainless steels by performing pack carburization tests. Their results are
summarized in Table 5.2. Silicon was again
noted for its benecial effect, as illustrated by

Type 330 (0.47% Si) versus Type 330 (1.0% Si)
and Type 304 (0.39% Si) versus Type 302B
(2.54% Si). Chromium was found to be benecial in Fe-Cr alloys, as shown by Type 446
(27% Cr) versus 430 (16% Cr). Small additions
Fig. 5.21
pg 110
26/10/2007 12:22PM Plate # 0
Fig. 5.23
Effect of nickel content on the diffusion coefficient of carbon in Fe-15Cr-Ni alloys. Source: Ref 38
of titanium or niobium appeared to be benecial

when comparing Type 321 and Type 347 to
Type 304.
Several nickel-base alloys along with HK-40
were investigated by Kane and Hosier (Ref 50).
Tests were conducted in environments with unit
carbon activity and various oxygen potentials.
Different rankings were obtained at different
oxygen potentials. Test results for two environments are summarized in Table 5.3. In the
test environment of H2-12%CH4-10%H2O with
1.3 1020 atm of pO2 , where SiO2 was thermodynamically stable, HK-40 (1.19% Si) performed the best. When pO2 was reduced to 1.9 1024
atm in H2-0.1%C7H13OH, where SiO2 was
not thermodynamically stable, and the carbon
activity was maintained at unity, alloys containing aluminum, such as alloys 601 and 617,
were much more resistant to carburization than
HK-40. The authors (Ref 50) attributed this to the
formation of an Al2O3 scale, although alloys
601 and 617 contain only about 1.3% Al. Other
investigators (Ref 24, 51) also showed benecial
effect of silicon in 25Cr-30/35Ni type alloys.
Nishiyama et al. (Ref 24) performed their
study of carburization in a simulated ethylene
pyrolysis environment (H2-15%CH4-3%CO2)
for Fe-Ni-Cr alloys with high Si (1.7%) and low
Si (0.30.5%). When tested at 1000 C, where
Cr2O3 was stable (see Fig. 5.9), chromium,
not silicon, was more effective in improving
carburization resistance as shown in Fig. 5.33
pg 111
26/10/2007 12:22PM Plate # 0
pg 112
of alloy 214 compared with those of alloys X,

601, and 150 after exposure to the test environment at 980 C (1800 F) for 55 h (Ref 52). The
Mass change, g/m2
alloy 602CA
100
50
200
alloy 617
alloy 601
45TM
400
(a)
Mass change, g/m2
alloy 800H
AC66
150
600
800
1000
1200
Time, h
350
300
250
200
150
100
50
0
AC66
alloy 800H
alloy DS

alloy 600H

alloy 601
45TM
alloy 602CA
alloy 617
(b)
Mass change, g/m2
(Ref 24). Alloy B (26Cr-35Ni-0.5Si) was signicantly better than alloy D (21Cr-31Ni-0.3Si),
while alloy B (26Cr-35Ni-0.5Si) was similar to
alloy A (25Cr-37Ni-1.8Si). At 1150 C, where
Cr2O3 oxide was not stable, high-Si alloys
(alloy A: 25Cr-37Ni-1.8Si, alloy C: 32Cr-43Ni1.7Si) exhibited signicantly better carburization
resistant than low-Si alloys (alloy B: 26Cr-35Ni0.5Si, alloy D: 21Cr-31Ni-0.3Si), as shown in
Fig. 5.34.
Aluminum is the most effective alloying
element in improving an alloys carburization
resistance at high temperatures. Lai (Ref 52)
showed that when tested at 870, 930, and 980 C
(1600, 1700, and 1800 F) alloy 214 (Ni-16Cr3Fe-4.5Al-Y) was the most carburizationresistant alloy among more than 20 commercial
wrought alloys, ranging from stainless steels and
Fe-Ni-Cr alloys to nickel- and cobalt-base
superalloys in the test environments, which were
characterized by a unit carbon activity and oxygen potentials such that Cr2O3 was not expected
to form on the metal surface. Oxides of silicon,
titanium, and aluminum were expected to be
stable under the test conditions. The excellent
carburization resistance of this alloy was attributed to the Al2O3 oxide scale formed on the metal
surface. Figure 5.35 illustrates the microstructure
200
400
600
800 1000 1200 1400
Time, h
320
240
160
80
200
400
(c)
45TM
alloy DS
600
AC66
alloy 601
alloy 617
alloy 602CA
alloy 600H
800
1000
1200
Time, h
Fig. 5.25
Carbon pick-up after 1000 h

of exposure at 1000 C g/cm2
Results of carburization tests in H2-CH4 mixtures

(ac = 0.8) at (a) 850 C, (b) 1000 C, and (c) 1100 C
for Fe-Ni-Cr alloys (800H, AC66, and DS) and nickel-base alloys
(alloys 600H, 601, 602CA, 617, and 45TM). Source: Ref 29
400
AC66
300
200
alloy 600H
45TM
alloy 800H
alloy DS
HPM
alloy 625 alloy 601
alloy 602CA
alloy 617
100
0
10
20
30
40
50
Fe, %
Fig. 5.24
Carbon solubility as a function of nickel content at

different carbon activities in Fe-Ni alloy system at
1000 C (1830 F). Source: Ref 39
Fig. 5.26
Weight gain as a function of iron content (i.e., a

function of nickel) in Fe-Ni-Cr and Ni-Cr alloys
tested at 1000 C for 1008 h in a H2-CH4 mixture (ac = 0.8 and
725:25
pO2 =10
). Source: Ref 29
26/10/2007 12:22PM Plate # 0
Fig. 5.27
Effect of nickel content on the carburization resistance of Fe-Ni-Cr alloys. Source: Ref 42
Fig. 5.29
Fig. 5.28
Carburization resistance of several wrought and

cast Fe-Cr-Ni alloys (Type 304, 321, HT, HU, and
HK) after testing in dry ethane (C2H6) for 24 h at temperatures from
880 to 1000 C. Source: Ref 43
existence of this oxide scale was conrmed by

Auger analysis (Ref 52). Similar test results were
also observed by Lai et al. (Ref 53) when tested in
H2-2%CH4 at 982 C (1800 F) for 96 h. The

carbon activity for the test environment at the test
temperature was greater than 1.0 (see Fig. 5.11).
Their results are summarized in Fig. 5.36. Alloy
Effect of chromium on the carburization resistance

of Fe-Ni-Cr alloys after testing in H2-2.5%CH4 at
pg 113
26/10/2007 12:22PM Plate # 0
pg 114
214 showed no evidence of carburization,

while Type 310 and alloys 800H, 25-35NbMA,
25-35Nb, 803, and 602CA showed different
degrees of carburization attack. Alloy 602CA
with about 2% Al (Ni-25Cr-10Fe-2Al-Y-Zr),
although not as good as alloy 214, was signicantly better than other alloys. A similar observation was also made by Kane et al. (Ref 54),
showing that the alumina-forming MA956
Table 5.1 Weight gain (mg/cm2) for several
cast alloys after 100 h at 1090 C (2000 F) in
H2-CH4-H2O mixtures
Weight gain, mg/cm2
Alloy(a)
HK-40 (1% Si)

HK-40 (2% Si)
TMA-4750
HP-45
TMA-6350
H2-8.6CH4-7H2O
H2-12CH4-10H2O
25.0
16.8
2.0
19.0
3.8
21.8
10.2
1.0
4.3
2.3
Table 5.2 Results of pack carburization tests

at 980 C (1800 F)(a) for various stainless steels
Nominal
composition
Si content,
%
Increase in C
content(b), %
Fe-21Cr-34Ni
Fe-15Cr-35Ni
Fe-15Cr-35Ni-Si
Fe-25Cr-20Ni
Fe-25Cr-20Ni-Si
Fe-25Cr-12Ni
Fe-18Cr-8Ni-Nb
Fe-18Cr-8Ni-Ti
Fe-18Cr-8Ni
Fe-18Cr-8Ni-Si
Fe-28Cr
Fe-16Cr
0.34
0.47
1.00
0.38
2.25
0.25
0.74
0.49
0.39
2.54
0.34
0.36
0.04
0.23
0.08
0.02
0.03
0.12
0.57
0.59
1.40
0.22
0.07
1.03
Alloy
800
330
330
310
314
309
347
321
304
302B
446
430
(a) 40 cycles of 25 h each cycle at 980 C (1800 F). Carburizer was renewed after
each cycle. (b) Bulk analysis. Source: Ref 49
(a) HK-40 (1% Si): 0.43C-0.60Mn-0.96Si-25.4Cr-20.7Ni. HK-40 (2% Si): 0.41C0.60Mn-1.98Si-25.0Cr-20.7Ni. TMA-4750: 0.44C-0.69Mn-1.99Si-24.9Cr20.8Ni-0.11Ti-0.29Nb-0.30W-REM. HP-45: 0.51C-0.54Mn-1.65Si-25.5Cr36.1Ni. TMA-6350: 0.50C-0.70Mn-1.84Si-25.1Cr-38.4Ni-0.13Ti-0.28Nb-0.27WREM. REM denotes rare earth metals. Source: Ref 44
Fig. 5.31
Effect of silicon on the carburization resistance of

cast Fe-20Ni-Cr alloys tested at 1090 C (2000 F)
for 24 h in wet ethane (C2H6). Source: Ref 46
Fig. 5.30
Ref 45
Effect of silicon on the carburization resistance

of 25Cr-20Ni and 35Cr-25Ni-Nb alloys. Source:
Fig. 5.32
Effect of silicon on the carburization resistance

of HK-40 with different silicon levels tested at
1100 C (2010 F) for 520 h in carbon granulate (pack carburization test). Source: Ref 47
26/10/2007 12:22PM Plate # 0
alloy (Fe-20Cr-4.5Al-0.5Y2O3) performed signicantly better than alloys 601, 800H, Type
310, and HK alloy (Table 5.4).
Aluminum, which has been used for alloy
addition to form external Al2O3 scale in ironand nickel-base wrought alloys, has been used
in centrifugally cast alloys for oxidation or
carburization resistance. Recently a commercial,
centrifugally cast nickel-base alumina-forming
alloy, alloy 60HT, containing approximately
25% Cr, 11% Fe, 0.4% C, and Al, was developed
(Ref 55). In the paper published by Kirchheiner
et al. (Ref 55), no carburization data were
reported. However, the resistance to coking was
studied on alloy 60HT containing three levels of
aluminum (i.e., 2.35, 3.55, and 4.81%). They
found signicant reduction in coking rates for the
samples containing 3.55 and 4.81% Al, with the
sample containing 2.35% Al exhibiting only
slight reduction of coking rates compared with
the conventional HP-40 alloy (4852). Coking,
which is an important phenomenon in ethylene
cracking, develops on the internal surface of the
pyrolysis tube and reduces the heat transfer. The
ethylene cracking operation has to be interrupted
Table 5.3 Weight gain (mg/cm2) for several
Fe-Ni-Cr and Ni-base alloys after ten 24 h cycles
at 1100 C (2010 F) in H2-12%CH4-10%H2O
and H2-0.1%C7H13OH environments
Weight gain, mg/cm2
Alloy
HK-40 (1.19% Si)

800H
600
617
601
690
H2-12CH4-10H2O(a)
H2-0.1C7H13OH(b)
6.97
23.46
12.00
16.84
22.74
25.54
54.38
46.60
4.85
0.50
1.81
54.38
(a) Inlet gas mixture: ac = 1.0, pO2 1:3 10720 atm at 1100 C. (b) Inlet gas
mixture: ac = 1.0, pO2 =1:9 10724 atm at 1100 C. Source: Ref 50
Fig. 5.33
Carbon profiles of high-Si (HSi) and low-Si (LSi)

Fe-Ni-Cr alloys after testing at 1000 C (1832 F)
for 96 h in H2-15%CH4-3%CO. Source: Ref 24
to allow decoking, typically with steam and air to

remove the coked layer. This decoking operation
can have detrimental effects on the tube properties, thus reducing the tubes service life.
Other factors such as surface nish have been
found to be very important in affecting carburization reactions. Machining the metal surface to
improve the surface nish can signicantly
increase an alloys carburization resistance. It is
common practice to bore or hone the internal
diameter of a centrifugally cast tube to remove
surface shrinkage pores. Figures 5.17, 5.37, and
5.38 illustrate the signicant improvement in
carburization resistance as a result of surface
machining (Ref 34, 56). A cast metal surface with
shrinkage pores can generate stagnant conditions
in crevices, which are very conducive to carburization attack. In addition, a machined surface
exhibits a cold-worked layer, which tends to
accelerate the diffusion process and results in
rapid formation of oxide scale or lm, thus
slowing subsequent carbon ingress.
Compared to the machined or ground surface,
the electropolished surface exhibited accelerated
carburization (Fig. 5.39). Norton and Barnes
(Ref 35) considered the electropolished surface a
work-free surface that failed to develop a surface
oxide lm as readily as the cold-worked ground
surface. Not mentioned by Norton and Barnes
in their paper (Ref 35), electropolishing can
produce a surface layer that may be depleted
in chromium, and the surface depletion of
chromium may result in the observed accelerated
carburization. In their study of oxidation/
carburization of a high-temperature gas-cooled
reactor (HTGR) helium environment containing
parts per million (ppm) levels of H2, CO, CO2,
CH4, and H2O, Mazandarany and Lai (Ref 57)
observed that Type 316SS specimen (obtained
Fig. 5.34
Carbon profiles of high-Si (HSi) and low-Si (LSi)

Fe-Ni-Cr alloys after testing at 1150 C (2100 F)
for 48 h in H2-15%CH4-3%CO. Source: Ref 24
pg 115
26/10/2007 12:22PM Plate # 0
from sheet material) with the as-received surface

condition suffered severe carburization after
testing at 649 C (1200 F) for 5000 h in
He-1500 atm H2-450 atm CO-50 atm CH4-50
atm H2O. The 316 specimen with the as-ground
specimen, on the other hand, showed no evidence of carburization tested in the same
retort under the same test condition. This observation was supported by both microstructure
and microhardness prole results, as shown in
Fig. 5.40 (Ref 57). The oxygen potentials of
the test environment were reported to be high
enough to form chromium oxides and too low
to form iron and nickel oxides. The unusual
Fig. 5.35
observation for the 316 specimen with the asreceived surface condition after testing at 649 C
(1200 F) in He-1500 atm H2-450 atm CO-50
atm CH4-50 atm H2O was the formation of an
Fe-Ni metallic scale on the specimen surface and
internal oxides underneath the metallic scale.
This is shown in Fig. 5.41 (Ref 57). The 316SS
specimen with the as-ground surface condition
after testing under the same condition showed
surface oxide scales with no outer metallic scale
and internal oxides. The authors thus believed
that the initial surface of the 316SS sheet was
depleted in chromium at a signicant degree,
such that not enough chromium was available
Optical micrographs showing the microstructures of (a) alumina-former alloy 214 (Ni-16Cr-3Fe-4.5Al-Y), and several
chromia-former alloys (b) 601 (Ni-23Cr-14Fe-1.4Al), (c) X (Ni-22Cr-18Fe-9Mo), and (d) 150 (Co-27Cr-18Fe) after testing at
980 C (1800 F) for 55 h in Ar-5%H2-5%CO-5%CH4 (ac = 1.0,pO2 =9 10722 atm). Source: Ref 52
pg 116
26/10/2007 12:23PM Plate # 0
pg 117
Fig. 5.36
Weight gain as a function of cycles (24 h cycle) tested at 982 C (1800 F) for 96 h with 24 h cycles in H2-2%CH4 for Type
310SS, 800H (N08810), 25-35Nb (25Cr-35Ni-1.3Nb-0.4C-2.0max Si), 25-35NbMA (microalloyed, 25Cr-35Ni-1.1Nb0.4C-1.4Si, Ti, REM), 803 (S35045: Fe-25Cr-35Ni), 602CA (N06025: Ni-25Cr-10Fe-2.1Al-Y-Zr), and 214 (N07214: Ni-16Cr-3Fe-4.5AlY). Note: All four alloy 800H (N08810) specimens were control samples. Source: Ref 53
Table 5.4 Carburization resistance of Fe-Ni-Cr,

Ni-base, and ODS alloys in H2-2%CH4 at
1000 C (1830 F) for 100 h
Alloy
MA956
601
800
310SS
HK (1.07% Si)
HK (2.54% Si)
Weight gain, mg/cm2
<0.3
10
19
36
34
29
Source: Ref 54
Fig. 5.37
Comparative carburization resistance of as-cast and

machined tubes of alloys 30Cr-30Ni, 36X (Fe-0.4C25Cr-34Ni-1.2Nb), and 36XS (Fe-0.4C-25Cr-34Ni-1.5Nb-1.5W)
after 3 years of field testing in an ethylene pyrolysis furnace.
Source: Ref 34
to form a continuous external chromium oxide

scale. Then, the nucleation of internal chromiumrich oxides resulted in a chromium-depleted
metal on the metal surface. When the surface

of the test specimen from the same sheet
material was ground in specimen preparation, the
chromium-depleted surface layer was removed.
When tested under the same condition, this
ground specimen formed an external chromium
oxide scale on the metal surface without the
formation of an Fe-Ni metallic surface layer. The
Fe-Ni metallic layer formed on the as-received
surface of the 316SS specimen provided rapid
absorption of carbon and a rapid diffusion path
of carbon in the metal, resulting in signicant
carburization. When tested at 870 C (1600 F)
in the same environment, an external chromium
oxide scale was observed to form on the
26/10/2007 12:23PM Plate # 0
as-received surface with no evidence of carburization attack. No explanation was offered by

the authors on this. It is believed that higher
temperatures caused a rapid homogenization of
the chromium concentration at the initially
depleted surface layer due to faster diffusion rate.
Since this type of environment contains only
ppm levels of corrosive gaseous components, the
reaction kinetics are believed to be relatively
slow. This slow reaction kinetics allowed the
homogenization of the chromium concentration
at the metal surface when tested at 870 C
(1600 F).
Injecting sulfur compounds into the carburization gas stream can signicantly reduce the
carburization kinetics. Sulfur compounds, such
as H2S, can retard carburization kinetics. This is
illustrated by the test results shown in Fig. 5.42,
obtained by Ramanarayanan et al. (Ref 58).
Benecial effect of sulfur in reducing carburization attack was demonstrated by Norton
and Barnes (Ref 35) in their laboratory tests.
Figure 5.43 shows that carburization of HK-40
alloy was signicantly reduced when 100 ppm
H2S was injected into the test environment. The
effect of different levels of H2S injection on the
carburization of HK-40 is illustrated in Fig. 5.44.
Figure 5.45 shows the carburization of alloy

800H in H2-CH4 (ac=1.0) with injection of different levels of H2S (H2S/H2) (Ref 59). When
this approach is used for controlling carburization, it is important to determine the optimum
level of sulfur compounds needed to achieve
maximum effectiveness. Too much can lead to
accelerated corrosion due to suldation, which
may be worse than carburization. Figure 5.46
shows that a low H2S injection failed to reduce
carburization and a high H2S injection caused
suldation attack for alloy 800 (Ref 59). It
has been proposed by Grabke et al. (Ref 60)
and Ramanarayanan and Srolovitz (Ref 61) that
sulfur absorbed on the metal surface blocks
the potential sites for carbon absorption from the
carburizing gas, thus signicantly reducing the
carbon concentration in the metal surface layer
and retarding the overall carbon transfer into the
metal.
5.2.3 Effect of Carburization on Mechanical
Properties
Carburization results in formation of internal
carbides in the alloy. The room-temperature
ductility or toughness can be severely degraded
Fig. 5.39
Fig. 5.38
Effect of machining on the carburization resistance

of cast HK-40 alloy. Source: Ref 56
Detrimental effect of the electropolished surface

condition on the carburization resistance of HK-40
alloy tested at 825 C (1520 F) in an environment with 0.8 ac and
the oxygen potential shown in Fig. 5.18. Source: Ref 35
pg 118
26/10/2007 12:23PM Plate # 0
pg 119
Oxalic etch
Exposed at 649 C (1200 F) for 5000 h, as-received surface
(649 C)1200 F
As-received surface
500
Microhardness, DPH
400
300
649 C(1200 F)
and
760 C(1400 F)
Ground surface
Oxalic etch
Exposed at 649 C (1200 F) for 5000 h, ground surface
200
871 C(1600 F)
100
0
0
100
200
300
400
500
600
700
Distance from exposed surface, m
Fig. 5.40
Microhardness profile and optical micrograph showing severe carburization attack on the 316 specimen with the original
as-received surface (solid circle data point) after testing at 649 C (1200 F) for 5000 h in He-1500 atm H2-450 atm
CO-50 atm CH4-50 atm H2O. Also shown are microhardness profile and optical micrograph of the 316 specimen with the as-ground
surface condition (solid square data point) after testing under the same condition showing no sign of carburization attack. The specimens
with the ground surface condition showed no carburization attack after testing at a higher temperature (i.e., 760 C) in the same
environment for the same test duration. The specimen with the as-received surface condition showed no evidence of carburization attack
when tested at 871 C (1600 F). Source: Ref 57
26/10/2007 12:23PM Plate # 0
pg 120
when the alloy is heavily carburized (Ref 29). For

high-temperature components that are subjected
to severe carburizing environments at very high
temperatures, such as ethylene pyrolysis furnace
tubes, the component can be carburized through
the thickness of the component with signicant
amount of carbon pickup. Grabke and Jakobi
(Ref 62) performed a metallurgical analysis of a

failed ethylene pyrolysis furnace tube made of
HP40Nb and found the tube was carburized
4
T = 1000 C
3.5
3
H2S Off
2.5
H2S On
2
Carburization
1.5
1
Cr2O3 Growth
Cr2O3 to Cr7C3 Conversion
0.5
0
20
40
60
80
100
120
Fig. 5.42
Fig. 5.41
As-received surface 316 specimen after testing at

649 C (1200 F) for 2000 h in He-1500 atm H2450 atm CO-50 atm CH4-50 atm H2O. Area 1: Fe-Ni metallic
phase, Area 2: Cr-Mn-Si oxides, and Area 3: Fe-Ni metallic phase.
Unetched condition. Source: Ref 57
Effect of sulfur on the carburization kinetics of Ni30Cr alloy. Vertical axis is mass gain per unit area,
and horizontal axis is exposure time in hour. The test began with
oxidation at 1000 C (1832 F) in a CO-CO2 environment for
about 40 h, and then switched to carburization in a H2-CH4
mixture with a carbon activity of 0.9, thus prompting the conversion of oxide scale to carbides and a rapid carburization. This
was then followed by injecting 100 ppm of H2S into the carburizing environment, thus immediately stopping carburization
reaction until the injection of H2S was turned off, which set off a
rapid carburization again. Source: Ref 58
Fig. 5.43
Carbon concentration profiles for HK-40 alloy after

testing at 1000 C (1830 F) for 100 h in a carburizing environment (ac = 0.8, pO2 is shown in Fig. 5.18) with and
without injection of 100 ppm H2S. The specimen was a flat
coupon with both sides being exposed to the carburizing gas.
Source: Ref 35
Fig. 5.44
Effect of H2S on the carburization behavior of

HK-40 alloy. Source: Ref 35
26/10/2007 12:23PM Plate # 0
throughout the tube wall with 3.14 to 3.3% C, as

opposed to the original carbon content of about
0.5%. Furthermore, carburization can increase
the volume of the carburized zone, thus developing internal stresses. When the alloy is
severely carburized, these internal stresses can be
Fig. 5.45
Weight gain of alloy 800 after 100 h in H2-CH4-H2S environment (ac = 1.0) at 900, 1000, and 1100 C (1652, 1832, and
2012 F) for different H2S/H2 ratios. Source: Ref 59
Fig. 5.46
Effect of H2S on the carburization behavior of alloy 800 in H2-CH4-H2S environments (ac = 1.0) at 1000 C (1832 F) for
different H2S/H2 ratios. Also shown are corresponding microstructures of the tested specimens. Source: Ref 59
pg 121
26/10/2007 12:23PM Plate # 0
signicant enough to develop cracking when the

carburized alloy is cooled to room temperature.
This is illustrated in Fig. 5.47 (Ref 25).
When the alloy is in the high-temperature
creep range, a thoroughly carburized alloy can
retain reasonable ductility at elevated temperatures. In fact, a thoroughly carburized Fe-Ni-Cr
alloy can exhibit a better creep ductility than the
same alloy tested in air. Guttmann and Schonherr
(Ref 63) studied the creep-rupture behavior of
HK alloys in a heavily precarburized condition
and found that the creep ductility was signicantly improved by carburization when creep
tested at 1000 C (1832 F). In their study, they
precarburized both HK-40 and HK-30 specimens
at 1000 C (1832 F) in H2-CH4 to a complete
saturation of M7C3 carbides (about 55 vol%) in
the material with about 4% C. The creep-rupture
testing was conducted at 1000 C (1832 F) in
H2-1%CH4 (ac = 0.8) to avoid decarburization
during testing. The rupture ductility of precarburized specimens was compared with that of
HK-40 and HK-30 in the as-cast condition
tested in air at the same temperature, as shown
in Fig. 5.48. The authors also made a similar
observation in comparing the rupture ductility
when tested in H2-1%CH4 (ac = 0.8) at 1000 C
(1832 F) (i.e., specimens were not precarburized prior to testing) with that tested in air, as
shown in Fig. 5.49. Carburization-accelerated
creep deformation at high temperatures is clearly
illustrated by the two comparison creep curves
tested at 1000 C (1832 F) in H2-1%CH4 (ac =
0.8) and air reported by Guttmann and Schonherr
(Ref 63), as shown in Fig. 5.50. Because carburization accelerates creep deformation, creep
strength is signicantly reduced. Carburization
as the result of either testing in a carburization

environment or precarburization can signicantly reduce the 1% creep strengths of both
HK-40 and HK-30. The authors attributed this
to both the internal stresses developed due to
Fig. 5.47
Fig. 5.49
Cracking developed in alloy 800 due to internal

stresses resulting from heavy carburization in the
test coupon with no external loading during the carburization test.
Source: Ref 25
Fig. 5.48
Elongation to fracture as a function of rupture life

of HK-40 and HK-30 comparing the as-cast
specimens tested in air and the precarburized (thoroughly carburized) specimens tested in H2-1%CH4 (ac = 0.8) (to avoid
decarburization) after creep-rupture testing at 1000 C (1832 F).
Source: Ref 63
Elongation to fracture as a function of rupture life of

HK-40 and HK-30 comparing the data from tests in
the carburizing environment (i.e., H2-1%CH4 [ac = 0.8]) and that
from air tests at 1000 C (1832 F). Source: Ref 63
pg 122
26/10/2007 12:23PM Plate # 0
volume increase in carburization and carbide

coarsening during creep testing at high temperatures. This is illustrated in Fig. 5.51.
Ethylene pyrolysis furnace tubes that are
subjected to severe carburization during service
at high temperatures typically experience signicant creep deformation as a result of carburization. Accordingly, the tube suffers stretching
due to creep elongation, and frequent shortening
of tube coils is required in order to avoid the
tube from reaching to the furnace oor to cause
tube damages (Ref 64). Jakobi and Gommans
(Ref 64) also indicated brittle fracture was one of
the failure modes for centrifugally cast pyrolysis
furnace tubes. The brittle fracture was primarily
due to the furnace trip that caused a large
temperature drop of 500 to 1000 C (932 to
1832 F). Such a furnace trip can produce a strain
of 0.75 to 1.5%, which is equivalent to about the
rupture ductility of aged, carburized, and nitrided
cast furnace tubes at room temperature to about
600 C (1112 F) (Ref 64). Thus, the furnace
tubes suffer brittle fracture at low temperatures
due to the large temperature drop.
Taylor et al. (Ref 65) studied the effect of
carburization on the creep-rupture behavior of
alloy 800H at 800 C (1472 F) and found that
carburization increased rupture strength and
life. They compared the creep-rupture data of
precarburized specimens (partially carburized
specimens and fully carburized specimens) with
that of as-received specimens as well as pre-aged
specimens. The data for precarburized specimens
were tested in a H2-CH4 environment (ac > 0.5)
to avoid decarburization during testing, while asreceived and pre-aged specimens were tested in
air. Precarburization was carried out at 1000 C

(1832 F) for 1200 h in a H2-CH4 mixture with
ac being 0.8, which resulted in a fully carburized
condition containing only M7C3 (about 30 vol
%). Partially carburized specimens were carburized at 1000 C (1832 F) for 100 h in a H2-CH4
mixture with ac being 0.8, which produced a
carburized depth of about 1 mm (equivalent to
56% of cross-sectional area being carburized).
Pre-aging treatment consisted of heating the
specimen at 1000 C (1832 F) for 100 h in air to
produce a comparable structure to the uncarburized core of the partially carburized specimens.
Their creep-rupture data are summarized in
Fig. 5.52. Fully carburized specimens were
found to exhibit signicantly higher creeprupture strengths than the as-received specimens
tested in air. Partially carburized or pre-aged
specimens showed strengths comparable to those
of the as-received specimens. As for rupture
ductility, the fully carburized specimens showed
lower rupture ductility for shorter rupture times
(at high stresses), but comparable to those of
as-received specimens and pre-aged as well as
partially carburized specimens for longer rupture
times (i.e., at lower stresses). This is illustrated in
Fig. 5.53. The authors also investigated the effect
of the temperature on creep-rupture ductility of
alloy 800H that was fully carburized. The alloys
rupture ductility (when fully carburized) was
found to decrease signicantly with decreasing
temperature. The carburized specimen failed in
a completely brittle manner on loading when
tested at 600 C (1112 F). This is illustrated in
Fig. 5.54.
Fig. 5.50
Fig. 5.51
Comparative creep curves of HK-40 tested at

1000 C (1832 F) and 15 MPa in air and H2-1%
CH4 (ac = 0.8). Source: Ref 63
1% creep strengths of HK-40 and HK-30 tested at

1000 C (1832 F) in air, H2-1%CH4 (ac = 0.8), and
for precarburized specimens tested in H2-1%CH4. Source: Ref 63
pg 123
26/10/2007 12:23PM Plate # 0
5.3 Metal Dusting

5.3.1 Metal DustingThermodynamic
Considerations
Metal dusting occurs in carburizing atmospheres at intermediate temperatures (approximately between 430 and 900 C (800 and
1650 F) with the maximum rate of attack
occurring at about 600 to 700 C (1112 to
1292 F), depending on alloys and environments. It is now well understood that metal
dusting occurs in environments that exhibit highcarbon activities (i.e., ac > 1) (Ref 6675).
As discussed in section 5.2.1, three chemical
Fig. 5.52
Stress rupture data of precarburized (fully carburized and partially carburized) specimens tested at
800 C (1472 F) in a carburizing environment (to prevent decarburization) is compared with that of as-received specimens and
pre-aged specimens tested in air at the same temperature. Source:
Ref 65
Fig. 5.53
Comparison rupture ductility of alloy 800H tested

at 800 C (1472 F) for fully and partially carburized specimens in comparison with as-received and pre-aged
specimens. Source: Ref 65
reactions (Eq 5.15.3) that can cause carburization can also cause metal dusting attack. Metal
dusting most often occurs in syngas environments, which consist of primarily H2 and CO.
In the earlier discussion of Eq 5.1: H2 + CO =
C + H2O, where graphical presentation of the
chemical reaction is presented as a function of
carbon activity (ac), temperature, and the gas
composition in terms of (pCO pH2 = pH2 O ), as
shown in Fig. 5.1. Figure 5.1 illustrates that the
carbon activity of the environment decreases
with increasing temperature. Thus, at high temperatures, many gaseous compositions are not
thermodynamically favored for causing metal
dusting. Conversely, lower temperatures with
increasing carbon activity are thermodynamically favored for causing metal dusting. This is
why metal dusting attack diminishes as the
temperature increases after passing the peak
reaction temperature due to decreasing carbon
activity. For example, for the gas mixture with
(pCO pH2 =pH2 O ) ratio being 1.0, its carbon
activity varies from more than 102 to approximately 101 as the temperature increases from
400 to 800 C (752 to 1472 F). Metal dusting
attack diminishes as the temperature decreases
after dropping below the peak temperature. This
is due to the reactions kinetics that decreases
with decreasing temperature.
The reaction Eq 5.2, as graphically presented
in Fig. 5.2, shows a similar fashion. When the
environment consists of CO and CO2, metal
dusting attack follows that involves H2 and CO.
However, when the environment consists of H2,
CO, and CO2, the Boudouard reaction (CO-CO2
Fig. 5.54
Effect of temperature on rupture ductility of fully

carburized alloy 800H tested at different temperatures with stresses to cause rupture in 200 h. Source: Ref 65
pg 124
26/10/2007 12:23PM Plate # 0
reaction) is believed to be slower in reaction

kinetics than that of Eq 5.1, thus making Eq 5.1 to
be the dominant reaction for the metal dusting
(Ref 76, 77). The reaction 5.3 (Eq 5.3), on the
other hand, increases the carbon activity with
increasing temperature. The graphic presentation
of this reaction is shown in Fig. 5.3. The carbon
activity is generally too low for many environments containing CH4 where metal dusting
occurs at 600 or 700 C (1112 or 1292 F). The
presence of CH4 in the gas mixture containing
CO and CO2 may help bring up the carbon
activity at higher temperatures, thus increasing
the upper temperature limit at which metal
dusting can take place.
ac (gas)
CO + H2
H2O + C (diss.)
ac (Fe3C1Fe)
ac=1
(a)
Fe3C
C (diss.) + Fe
CO + H2
H2O + C (in Fe3C)
(b)
CO + H2
H2O + C (graphite)
5.3.2 Mechanisms of Metal Dusting Attack

Signicant advancement in understanding the
mechanisms of metal dusting has been achieved
in the past decade, with most of the contribution
being made by Professor Grabke and his group
(Ref 6875). Two models have been proposed
with one model explaining the metal dusting
behavior of iron and low-alloy steels and
the other model explaining the behavior of Crcontaining high-temperature alloys (i.e., Fe-Cr,
Fe-Ni-Cr, and Ni-Cr alloys). These two models,
which are summarized by Grabke (Ref 73), are
briey described here. The model for iron and
low-alloy steels is summarized in Fig. 5.55.
Based on the model, the prerequisite for metal
dusting to occur is a high carbon activity environment (ac > 1) that causes the oversaturated
metal with the carbon activity being more
than 1. This condition promotes the formation of
cementite (Fe3C) on the metal surface. As shown
in Fig. 5.1, the calculated carbon activities in
equilibrium with cementite are higher than 1.0
at 700 C (1292 F) and lower temperatures.
Figure 5.56 shows a metastable Fe-C-O phase
diagram at 600 C indicating that metastable
cementite forms at ac > 1 (Ref 78). Once graphite
deposits on the cementite with the carbon activity
of the graphite at 1, cementite in contact with
graphite becomes unstable and then decomposes
into iron and carbon with iron migrating outward
into the graphite layer. As a result, metal wastage
takes place as a repeated reaction of formation
of metastable cementite and decomposition of
cementite into iron and carbon. For iron and
low-alloy steels with no protective oxide scales,
metal wastage attack in general follows uniform
(c)
CO + H2
H2O + C
(graphite)
(d)
3 Fe + C
Fe3C
Fig. 5.55
Schematic showing mechanism of metal dusting for

iron and low alloy steels with the following steps:
(a) The metal is oversaturated with carbon (ac > 1) due to carbon
transfer from a high carbon activity environment (ac > 1) to the
metal, (b) thus resulting in the formation of cementite (Fe3C) at the
surface, (c) and later the formation of graphite on top of cementite
and then decreases in ac to one (ac = 1 for graphite) at the
cementite in contact with graphite, (d) thus, cementite becomes
unstable and decomposes into iron and carbon, with iron
migrating into graphite layer to form iron particles (embedded in
graphite), which act as catalysts for more carbon deposition and
coking (d). Source: Ref 73
Fig. 5.56
Fe-C-O metastable phase stability diagram showing

that metastable cementite (Fe3C) forms under high
carbon activities (ac > 1) at 700 C and lower. Source: Ref 78
pg 125
26/10/2007 12:23PM Plate # 0
pg 126
metal thinning (Ref 73). A layer of cementite

(Fe3C) was found to occur on 1Cr-0.5Mo steel
after exposure in metal dusting conditions with a
layer of coke on top of the cementite (Ref 71).
For chromia formers in Fe-Cr, Fe-Ni-Cr, and
Ni-Cr alloys, a different mechanism was proposed, which was summarized in Fig. 5.57
(Ref 73). These high-chromium alloys form a
protective chromium oxide scale. Metal dusting
attack is initiated when the oxide scale develops
local defects, allowing carbon transfer from a
high carbon activity environment (ac > 1) to
metal causing oversaturation of carbon, and
subsequent formation of graphite after formation
of metastable M3C carbides in low nickel alloys
Cr2O3 alloy
5.3.3 Alloy Resistance to Metal Dusting

Graphite
(a)
(d)
Internal
carbides
3M+C
and C
M3C
'Coke'
(b)
(e)
C
(or direct graphite formation in high nickel

alloys), thus resulting in lower carbon activity
(ac = 1) and decomposition of metastable carbides into metal particles and carbon. As a result,
the attack starts locally and often leads to the
formation of hemispherical pits (Ref 73). This
type of morphology is shown in Fig. 5.58.
Pippel et al. (Ref 79) found graphite on
cementite that formed on Fe-5Ni alloy, when
exposed to metal dusting conditions. For
Fe-10Ni and Fe-25Ni to Fe-30Ni alloys, graphite
was found to grow into metal with no metastable
carbide formed on the alloy (Ref 79). Graphite
was also observed to grow into metal directly
in high-nickel Fe-Ni alloys, such as Fe-40Ni,
Fe-50Ni, and Fe-80Ni (Ref 72).
Process equipment failures due to metal dusting were reported in the renery industry during
the 1950s. Eberle and Wylie (Ref 80) reported
metal wastage of uncooled components, such as
soot blower elements, made of Types 347SS and
310SS in the waste heat boiler of a synthesis
gas reactor. The synthesis gas, predominantly
CO and H2 with some water vapor and carbon
particles, was produced by combustion of
methane with oxygen. The wastage took place
at temperatures between 480 and 900 C (900
and 1650 F). Both 347SS and 310SS suffered
severe metal wastage after only 3 weeks of
service.
Prange (Ref 8) reported that tubes containing
a chromia-alumina catalyst at a temperature of
about 590 C (1100 F) in a butane dehydrogenation system, where butane was converted to
butene, suffered metal loss problems. The oxide
Metastable
carbide
(c)
Fig. 5.57
Mechanism of metal dusting for chromia formers

with the following steps: (a) Development of local
defects in the oxide scale, allowing carbon transfer from the
environment to the metal, (b) carbon ingress into the metal
resulting in the formation of stable carbides, M23C6 and M7C3, (c)
continued carbon ingress increases the carbon activity to more
than one (ac > 1), resulting in the formation of metastable M3C
carbides in low-nickel alloys, or resulting in direct growth of
graphite in high-nickel alloys, (d) graphite deposition occurs
decreasing the carbon activity to one, thus, causing the decomposition of metastable M3C into metal particles and carbon with
metal particles catalyzing more coking (e). Source: Ref 73
1cm
Fig. 5.58
Metal dusting attack in alloy 800 in synthesis gas

environment at 550 C (1022 F). Source: Ref 75
26/10/2007 12:23PM Plate # 0
pg 127
dusts resulting from metal dusting contaminated

the catalyst and caused undesirable side effects
for the operation. The alloys reported to perform
poorly were 12Cr steel, 18Cr-11Ni, and alloy
600. The alloys that performed well included
Fe-27Cr, 25Cr-20Ni, and Type 302B (18Cr-8Ni
with 2.4% Si). Severe metal loss in the form of
pitting was also observed by Hoyt and Caughey
(Ref 81) for Type 310SS equipment exposed to a
gas mixture rich in H2 and CO at temperatures of
650 to 700 C (1200 to 1300 F) in a plant that
converted coal to gasoline and other products.
Metal dusting problems were also encountered
in reforming plants (Ref 5), where a synthesis
gas (i.e., H2 + CO) for methanol manufacturing
was produced. Type 304SS and 310SS reformer
outlet tubes were perforated by severe pitting
attack at 650 to 725 C (1200 to 1340 F).
Hydrodealkylation units (Ref 5), acetic acid
cracking furnaces (Ref 5), and coal gasication
plants (Ref 6) were reported to suffer metal
dusting problems. Perkins et al. (Ref 6) reported
pitting attack of alloy 800 tubes in a preheater for
gasier recycle gas rich in H2 and CO with some
H2O. The tubes, with a wall thickness of 0.38 cm
(0.15 in.), were perforated in a few thousand
hours. The attack occurred at 540 to 870 C
(1000 to 1600 F).
Dunmore (Ref 82) reported a failure of the
waste heat boiler in an ammonia plant. The alloy
800 exit ferrules suffered severe metal wastage,
with large uniform circular pits penetrating
through the tubes and black sooty deposits on the
pitted surface. The attack occurred at a location
where the metal temperature was about 600 C
(1110 F). The environment was highly enriched
in H2 and CO, along with large quantities of
steam. This was unusual in that steam was considered an effective additive to mitigate the metal
dusting problem (Ref 7). Grabke and Spiegel
(Ref 83) discussed several industrial failure
cases that were caused by metal dusting. These
cases involved an alloy 800 heat exchanger for
synthesis gas production, HK-40 gas heaters in a
direct iron reduction plant, 5Cr and 9Cr steels in
catalyst regeneration units in a renery, and an
alloy 601 heat exchanger in an ammonia plant.
Metal dusting has also been encountered
in the heat treating industry (Ref 84). Refractory
anchors, fan housing assemblies, and other
components in carburizing furnaces frequently
suffer metal dusting problems. Alloys typically
used include stainless steels, such as Type 310,
nickel-base alloys, such as alloys X and 333,
and iron-base alloys, such as Multimet alloy
(or N-155). Metal dusting typically occurred at

temperatures between 540 and 820 C (1000 and
1500 F). Severe attack frequently took place
after 1 year of service and at locations where
the gaseous environment became stagnant.
Favorite locations included the interface with
refractories, where small gaps or dead spaces
were created. Figure 5.59 (Ref 84) shows a
sample of Multimet alloy obtained from a
furnace fan housing in a carburizing furnace. The
fan housing suffered metal dusting at the metal
Cm
Fig. 5.59
Multimet alloy fan box suffering metal dusting

attack in a carburizing furnace. (a) General view of
failed the sample. Note the perforated edge of the fan box. (b)
Cross section of the sample showing pitting attack and severe
metal thinning. (c) Severe carburization attack beneath the pitted
and wasted area. Attack was initiated in the dead spaces created
by the refractory and the fan box. Source: Ref 84
26/10/2007 12:23PM Plate # 0
surface that was in contact with the furnace

refractory lining. The 1.9 cm (0.75 in.) thick fan
housing was perforated in about a year. The
nearby metal surface that was exposed to the
circulating carburizing gas, on the other hand,
showed no sign of metal dusting (Fig. 5.60). A
chromium-rich oxide scale was observed on the
metal surface. One common feature associated
with metal dusting is carburization beneath the
pitted area (Fig. 5.59).
Manufacturing of carbon bers for carbon
composite materials can also generate high
temperature carburizing environments for causing metal dusting problems to some furnace
components. Carbon bers are manufactured
from polyacrylonitrile (PAN) (Ref 85). The
last step of manufacturing, carbonization, is
carried out at about 900 C (1650 F) in an
environment enriched with CO, CO2, CH4, N2,
HCN, H2O, and so forth. Furnace components
made of nickel- and iron-base alloys were found
to suffer general metal thinning and pitting
attack. An example is shown in Fig. 5.61.
The general characteristics of metal dusting
observed in the industrial environments were
reproduced in many laboratory tests involving
primarily gas mixtures of H2-CO-H2O. Accordingly, laboratory testing has become an important
tool for generating relevant data to allow
performance ranking among engineering alloys.
Laboratory data generated by various laboratories are summarized below.
Maier and Norton (Ref 86) investigated
9Cr-1Mo steel (P91), 12Cr steel (Type 410),
Fig. 5.60
Oxide scale and no evidence of metal dusting on

the surface of a Multimet alloy fan box exposed to
flowing carburizing gas (same fan box as that shown in Fig. 5.59).
The sample was plated with nickel before mounting for metallographic examination.
Model alloy (Type 410 + 2.75Si), and Type

310 (Fe-25Cr-20Ni) in H2-24.4CO-2.4H2O at
560 C (ac > 1). The specimens were annealed in
H2 at 1000 C (1832 F) for 1 h to eliminate
surface deformation structure and produced grain
coarsening. The test specimens were ground.
Their results are summarized in Fig. 5.62. In this
short-term test, alloys with low chromium contents (9Cr for P91 and 12Cr for 410SS) showed
metal dusting attack. Type 310SS (25Cr-20Ni)
showed no sign of metal dusting attack up to
200 h. Silicon was found to be a very effective
alloying element in improving the resistance to
metal dusting. This is illustrated by the data
comparing model alloy (410SS + 2.75Si) with
410SS.
Grabke et al. (Ref 69) investigated commercial
alloys, which included Fe-Cr, Fe-Ni-Cr, Ni-base
alloys, and silicon-containing alloys, at 650 C
(1200 F) for 7 days in H2-24.7CO-1.9H2O (ac =
15). All specimens were annealed in dry hydrogen at 1000 C (1832 F) for 1 h to achieve large
grains, followed by grinding and polishing. All
test specimens were in as-ground surface conditions. The data are summarized in Fig. 5.63. The
Fig. 5.61
Type 310SS furnace component suffering metal

dusting in a furnace used for manufacturing carbon
fibers. (a) General view of the failed component. (b) Cross section
of the sample showing pitting and thinning
pg 128
26/10/2007 12:23PM Plate # 0
pg 129
hatched data show mass gain of coke deposits on

the test specimen, while the solid data represent
the either weight loss due to metal dusting or
weight gain due to oxidation and/or carbon gain
(no metal dusting attack). Sicromal (Fe-18Cr1Al-1Si) contained not very high chromium
level, but showed no metal dusting. This was due
to the presence of some aluminum and silicon

in the alloy. Both aluminum and silicon are
effective alloying additions in improving the
metal dusting resistance. X18 CrN28 (Fe-28Cr)
showed signicantly better performance than
Sanicro 28 (Fe-27Cr-30Ni). Both alloys contained about the same level of chromium, but one
250
560 C, 1.5 bar
Mean metal loss, m
200
150
100
50
0
0
50
100
150
200
Exposure time, h
Fig. 5.62
Average metal loss as a function of exposure time (hours) for P91, 410, 310, and model alloy (410 + 2.75Si). Source: Ref 86
1000
100
10
Mass gain, mg/cm
1
0.1
0.01
0.001
0.01
0.1
1
10
Removable deposits
Fig. 5.63
Inconel 600
AC 66
Sanicro 28
HK 40
X15 CrNiSi 2012
253 MA
12 CrMoV
Alloy 410
X18 CrN28
Sicromal
100
Cleaned specimen
Mass gain (hatched data) due to coke deposits and metal loss of specimen (solid data) due to metal dusting as well as
mass gain of specimen (solid data) due to possible oxidation/carburization after exposure at 650 C (1200 F) in
H2-24.7CO-1.9H2O (ac = 15). X18 CrN28 (Fe-28Cr); Sanicro 28 (Fe-27Cr-30Ni); 253MA (Fe-21Cr-11Ni-2Si-0.05Ce); X15 CrNiSi 2012
(Fe-20Cr-12Ni-2Si). Source: Ref 69
26/10/2007 12:23PM Plate # 0
Fe25Cr2.5Ni
0
Fe60Cr
Mass change, mg/mm2
0.02
Fe25Cr
0.04
0.06
Fe25Cr5Ni
0.08
0.1
0.12
0.14
Fe25Cr25Ni
0.15
Fe25Cr10Ni
0.18
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210
Exposure cycles, h
Fig. 5.64
Weight loss of specimens after removing carbon deposits as a function of thermal cycles for Fe-Cr and Fe-Cr-Ni alloys in
metal dusting tests at 680 C (1256 F) in H2-68%CO-6%H2O (ac = 2.9 and pO2 =2 10723 atm) cycling specimens to room
temperature every 60 min in the test environment. Source: Ref 90
was a ferritic alloy (Fe-Cr) while the other was an

austenitic alloy (Fe-Ni-Cr). The diffusivity of
chromium in ferritic alloys is generally about two
orders of magnitude higher than that in austenitic
Fe-Ni-Cr alloys (Ref 8789). As a result, ferritic
stainless steels form chromium oxide scales
much more readily than austenitic stainless
steels, as was observed in Fig. 5.63. Both 12Cr
steels (410SS and 12CrMoV) showed some
metal dusting attack. Again, silicon was found
to greatly improve metal dusting resistance.
Both alloy 253MA (Fe-21Cr-11Ni-2Si-0.05Ce)
and X15 CrNiSi 2012 (Fe-20Cr-12Ni-2Si)
showed no metal dusting attack. Another siliconcontaining alloy, HK-40 (Fe-25Cr-20N-2Si),
showed some metal dusting attack. The alloys
that showed the worst attack in this test were
Fe-Ni-Cr alloys (Sanicro 28 and AC66), and a
low-chromium nickel alloy (alloy 600).
The results of Toh et al. (Ref 90) further conrmed that Fe-Cr alloys exhibit better metal
dusting resistance than Fe-Ni-Cr alloys. They
investigated Fe-25Cr alloy and Fe-25Cr with 5,
10, and 25% Ni. All alloys were annealed in
Ar-10%H2 at 1050 C (1922 F) for 100 h.
Test specimens were then cut, ground, and
polished to 3 m before being electrolytically
polished. Testing was conducted at 680 C
(1256 F) in H2-68%CO-6%H2O (ac = 2.9 and
pO2 =2 10723 atm). Specimens were cycled by
heating them to the test temperature for 60 min
for each cycle. Their test results are summarized
in Fig. 5.64. Fe-25Cr alloy was found to be signicantly more resistant to metal dusting than
Fe-25Cr alloys with 5, 10, and 25% Ni. Fe-25Cr
with 2.5% Ni appeared to perform well. It was

likely that the alloy with only 2.5% Ni remained a
ferritic structure. The test also included Fe-60Cr,
which showed good metal dusting resistance.
Ferritic Fe-Cr alloys are more resistant to
metal dusting than austenitic Fe-Ni-Cr alloys
presumably due to much faster chromium diffusion rates, thus resulting in faster formation of
chromium oxide scales, and then better metal
dusting resistance. Rapid diffusion of chromium
to the metal surface to form a protective chromium oxide scale is important in retarding metal
dusting attack. Accordingly, for the same alloy, a
ne-grained material can form a surface oxide
scale more readily than a coarse-grained material, thus better metal dusting resistance. This is
illustrated in Fig. 5.65, where the ne-grained
Type 304SS (as-received from the supplier)
with a grain size of ASTM No. 10 (average size
of 10 m) showed no metal dusting attack while
the coarse-grained material with a grain size
of about ASTM No. 3 to 7 (about 30100 m)
suffered severe attack (Ref 91). The coarsegrained material was obtained by annealing
the sample to 1000 C (1832 F) for 1 h.
Metal dusting testing was conducted at 600 C
(1112 F) in H2-24CO-2H2O. More discussion
on the effects of surface conditions on the metal
dusting behavior is presented later.
Fe-Ni-Cr alloys, which showed less resistance
to metal dusting than Fe-Cr alloys due to lower
chromium diffusivity, are also much less resistance to metal dusting as compared with Ni-base
alloys. This is illustrated in Fig. 5.66, which were
generated on the commercial alloys without prior
pg 130
26/10/2007 12:23PM Plate # 0
annealing treatments. Specimens were in the asreceived condition from the supplier and were
ground to a 120-grit nish prior to testing in
H2-90%CO for 672 days at 482, 566, 649, and
732 C (900, 1050, 1200, and 1350 F) (Ref 92).
10
18Cr-8Ni-Steel
Mass gain, mg/cm2
6
Coarse grain
4
2
Fine grain
0
50
100
150
200
Time, h
Fig. 5.65
Effect of grain size on the metal dusting behavior

of Type 304SS tested at 600 C (1112 F) in H224CO-2H2O. Source: Ref 91
Fig. 5.66
Type 304, 310, and 800Hwhich are Fe-Ni-Cr

alloyswere found to suffer metal dusting attack
(i.e., the reaction rates in weight loss) while two
nickel-base alloys (alloy 601 and RA333)
showed no metal dusting attack (i.e., the reaction
rates in weight gain). The gure also shows that
alloy 85H (Fe-18.5Cr-14.5Ni-3.5Si-1Al) showed
no metal dusting attack. This was due to benecial effect of silicon along with aluminum in the
alloy. Benecial effects of silicon and aluminum
are demonstrated by additions of these two elements to alloy 800 (Fe-20Cr-32Ni), as shown in
Fig. 5.67 (Ref 93). Also in this study by Strauss
and Grabke (Ref 93), other alloying elements,
such as niobium, molybdenum, and tungsten,
were found to have some benecial effects as
well.
Klower et al. (Ref 94) investigated a large
number of commercial alloys including several
Fe-Ni-Cr alloys, such as alloys 800H, HK-40,
HP-40, and DS, and many nickel-base alloys.
Test results are summarized in Table 5.5.
Fe-Ni-Cr alloys were much less resistant to metal
dusting than nickel-base alloys. High-siliconcontaining alloy DS (Fe-18Cr-35Ni-2.2Si) was
much better than other Fe-Ni-Cr alloys. This
is nicely illustrated in Fig. 5.68. In Ni-Cr-Fe
Metal dusting resistance of several Fe-Ni-Cr alloys and Ni-base alloys tested in H2-90%CO for 672 days at 482, 566, 649,
and 732 C (900, 1050, 1200, and 1350 F). The alloys tested were Type 304 (S30403), Type 310 (S31000), alloy 85H
(S30615), alloy 800H (N08810), alloy 601 (N06601), and RA333 (N06333). Source: Ref 92
pg 131
26/10/2007 12:23PM Plate # 0
pg 132
system, increases in chromium along with

aluminum additions can signicantly improve
the alloys metal dusting resistance, as illustrated
in Fig. 5.69. Alloy 601 was more resistant than
alloy 600 because of higher chromium and aluminum addition in alloy 601. Alloy 602CA was
better than alloy 601 because of higher aluminum
content. There were two different surface nishes
involved in this test, with alloys 600 and 601 in as
ground surface, and alloy 602CA in black
surface nish. There was no discussion about
the procedure for preparing the black surface
nish in the paper. If the 602CA specimens were
black annealed (heat treated in air), with about
20
+4.5%AI
+2.5%Si
+
+
Mass loss, mg/cm2
+3%Mo
+3%Nb
20
+3%W
Fe-32%Ni-20%Cr
60
+0.05%Ce
100
500
Fig. 5.67
1000
Time, h
1500
Effects of Al, Si, Nb, Mo, and W additions to a model

alloy (alloy 800, Fe-20Cr-32Ni) on the metal dusting behavior of modified experimental alloys. Source: Ref 93
2.3% Al in alloy 602CA, alloy 602CA specimens

might have preformed an Cr/Al-rich oxide scales
prior to metal dusting testing. The effects of
surface nishes are discussed further.
In Table 5.5, it is surprising to nd that alloy
214 with about 4.5% Al did not perform well
compared with 602CA (2.3% Al), 617 (1.3% Al),
and 690 (30% Cr, no Al). It is believed that with
only about 16% Cr, alloy 214 did not contain
enough chromium to rapidly develop a protective
chromium oxide scale at such a low temperature.
And, at this low temperature, development of an
exclusive Al2O3 is difcult in short time. Thus, a
Table 5.5 Metal wastage rates after exposure
to H2-CO-H2O gas mixtures at 650 C (1200 F)
for various commercial alloys
Alloy
Surface
condition
Total time, h
Wastage rates,
mg/cm2h
800H
HK-40
HP-40
DS
600H
601
601
601
C-4
214
HR160
45TM
602CA
617
690
Ground
Ground
Ground
Black
Polished
Ground
Ground
Ground
Ground
Black
Black
Ground
Ground
95
190
190
1,988
5,000
6,697
1,988
10,000
10,000
9,665
9,665
10,000
10,000
7,000
10,000
0.21
0.04
0.038
4.3 103
3.3 102
7.3 103
4.9 103
5.8 104
1.1 103
1.2 103
6.3 104
1.0 105
1.1 105
3.7 106
2.0 106
Note: Gas mixtures H2-24CO-2H2O (ac = 14 at temperature) for the rst 5000 h of
testing and H2-49CO-2H2O for the subsequent 5000 h. Source: Ref 94
102
Metal wastage rate, mg/cm2h
10
1
101
102
103
104
105
Fig. 5.68
Metal dusting resistance of Fe-Ni-Cr alloys (800H and HP40) in comparison with Ni-base alloy 600H tested at 650 C
(1200 F) in H2-24CO-2H2O (ac = 14). Source: Ref 94
26/10/2007 12:23PM Plate # 0
pg 133
high-chromium nickel-base alloy, alloy 690

(30% Cr), performed the best in this test. It is also
surprising to nd that HR-160 alloy with both
high Cr (28%) and high Si (2.75%) did not perform as well as some other nickel-base alloys.
Alloy 45TM, also with high Cr (27%) and high Si
(2.7%) performed better than HR-160 alloy. The
alloy 45TM specimen was in a black surface

condition. The black surface condition might
be resulted from a preoxidation treatment that
might have produced chromium- and silicon-rich
oxide scales prior to metal dusting testing.
In conducting metal dusting testing that
involved only the specimens with as-ground
101
Metal wastage rate, mg/cm2h
Alloy 600
102
Alloy 601
103
104
Alloy 602CA
105
106
0
2,000
4,000
6,000
8,000
10,000
Exposure time, h
Fig. 5.69
Increasing Cr along with addition of aluminum improves the metal dusting resistance of the alloy in Ni-Cr-Fe alloys. Source:
Ref 94
Fig. 5.70
Metal dusting behavior of various Ni-base alloys tested at 593 C (1100 F) in H2-18.4CO-5.7CO2-22.5H2O at 14.3 atm of
pressure. Source: Ref 95
26/10/2007 12:23PM Plate # 0
pg 134
surface nish, Natesan and Zeng (Ref 95) performed tests at 593 C (1100 F) in H2-18.4CO5.7CO2-22.5H2O at 14.3 atm of pressure. Alloy
45TM suffered rapid metal dusting attack with
about 59.1 mg/cm2 of weight loss after only
3300 h while HR-160 alloy exhibited only
7.3 mg/cm2 of weight loss after 9700 h. Their
results are summarized in Fig. 5.70. The bestperforming alloy was alloy 693 (30% Cr, 3.3Al),
followed by alloy 602CA (25% Cr, 2.3% Al),
and HR-160 (28% Cr, 2.75% Si) and alloy 690
(30% Cr). The worst performer was alloy 45TM,
followed by alloys 617, 214, and 601. These tests
conducted in Argonne National Laboratory were
under a high pressure (14.3 atm, or 210 psi), and
all other data generated in other laboratories were
under gas pressure close to 1 atm (atmosphere
pressure). Natesan and Zeng (Ref 95) observed
that high pressure could signicantly reduce
the time to initiate metal dusting attack for
some alloys. Tests were performed at 593 C
(1100 F) for 246 h in 1, 14.3, and 40.8 atm
pressures. Specimens were then examined for
signs of metal dusting attack. Their results are
summarized in Table 5.6, showing alloys 601,
690, 617, and 214 suffering metal dusting attack
at high pressures but not at 1 atm pressure after
246 h. On the other hand, alloys 45TM, 602CA,
and HR160 showed no metal dusting attack at
both low and high pressures after 246 h. They
also measured the maximum pit size and average
pit depth in addition to weight loss. This is
illustrated in Table 5.7. In terms of the pitting
depth, HR160 was found to perform the best,
while alloy 690 with little weight loss showed
signicant pit depth. The data suggested that the
total weight loss was not correlated well with
pitting depth. Alloy 214 was found to show
uniform metal wastage. This was apparently due
to insufcient chromium content (16%) at such a
low temperature to form a continuous Cr2O3
scale. The Raman spectra of the alloy 214 tested

specimen showed low intensity for the Cr2O3
band (Ref 95).
In manufacturing of sheet products (or thin
gage tubular products) in nickel-base alloys, the
manufacturing process in the later stage typically
consists of cold working (cold pilgering) and
bright annealing (i.e., annealing in H2 atmosphere). The hydrogen atmosphere in bright
annealing typically contains some moisture, thus
it is typically characterized with a certain dew
point (i.e., a certain oxygen potential, pO2 ). In
general, annealing furnaces exhibit low enough
dew points such that chromium oxide scales do
not form on the metal surface when processing
stainless steels or nickel-base alloys. Sheet products of these alloys typically look shiny. However, the dew points or oxygen potentials in
hydrogen-annealing furnaces generally are high
enough that Al2O3 or SiO2 will form on the metal
surface of a bright-annealed sheet of a nickel
alloy containing a sufcient level of aluminum or
silicon. Klarstrom and Grabke (Ref 96) tested
bright-annealed sheet specimens of alloy 214
(alumina former) and alloy HR160 (silica
former) along with alloy 230 and HR120 in the
bright-annealed condition. Also included in
testing were black annealed and pickled sheet
specimens of alloys 800H and 601 (no brightannealed sheet samples were available for these
two alloys at the time of testing). No surface
grinding for bright-annealed sheet specimens
(alloys 214, HR160, 230, and HR120). Surface
grinding to a 120-grit surface nish was performed for black-annealed and pickled sheet
specimens of alloys 800H and 601. Testing
was performed at 650 C (1200 F) for up to
10,000 hours in H2-49CO-2H2O (ac =18.9).
Test results are summarized in Table 5.8. HR160
alloy was found to perform very well, showing
no evidence of metal dusting after 10,000 h
Table 5.6 Surface conditions of alloys

after testing at 593 C (1100 F) for 246 h in
H2-18.4CO-5.7CO2-22.5H2O at 1, 14.3, and
40.8 atm pressures
Table 5.7 Pit size, pit depth, and weight loss

for alloys after testing at 593 C (1100 F) for
9700 h in H2-18.4CO-5.7CO2-22.5H2O
at 14.3 atm pressure
Condition at pressure
Alloy
601
690
617
602CA
214
45TM
HR160
Source: Ref 95
Alloy
1 atm
14.3 atm
40.8 atm
Clean surface
Clean surface
Clean surface
Clean surface
Clean surface
Clean surface
Clean surface
Pits
Pits
Pits
Clean surface
Pits
Clean surface
Clean surface
Pits
Pits
Pits
Clean surface
Pits
Clean surface
Clean surface
601
690
617
602CA
214
45TM(b)
HR160
693
Weight loss, mg/cm2
Pit depth, m
Pit diameter, m
19.5
6.5
35.1
2.1
25.6
59.1
7.3
0.1
110
147
201
96
(a)
141
13
37
450
440
887
374
(a)
600
210
99
(a) Specimen uniformly corroded. (b) Exposed for only 3300 h. Source: Ref 95
26/10/2007 12:23PM Plate # 0
of exposure. The alloy, however, showed slight

carburization underneath the metal surface.
Alloy 230 showed slight metal dusting attack
with a few spots 1 to 2 mm in diameter on each
side of the specimen after 10,000 h of exposure.
Alloy 601 suffered signicantly more metal
dusting attack than alloy 230. Typical surface
conditions for these three alloys after 10,000 h
of testing are shown in Fig. 5.71. Alloy 214,
however, did not perform well, suffering metal
Table 5.8 Final metal wastage rates for various
alloys tested at 650 C (1200 F) for up to
10,000 h in H2-49CO-2H2O (ac = 18.9)
Alloy
HR120
800H
214
601
230
HR160
Total exposure time, h
Wastage rate, mg/cm2h
190
925
5,707
10,000
10,000
10,000
4.1 102
2.7 103
1.0 103
2.5 103
3.2 104
0.0(a)
dusting attack with fairly high wastage rates.

Both Fe-Ni-Cr alloys, alloys 800H, and HR120,
performed the worst among the alloys tested.
Figure 5.72 shows typical surface conditions of
alloys 214, HR120, and 800H after 5707 h,
190 h, and 925 h, respectively.
Baker and Smith (Ref 97, 98) and Baker et al.
(Ref 99) investigated about 20 commercial alloys
in H2-80CO at 621 C (1150 F) for times up to
16,000 h. All materials were cold rolled and
annealed sheets (i.e., bright annealed products)
except alloys K-500 and 617, which were hotrolled, annealed plates (i.e., black annealed and
pickled products), and alloy DS, which was
extruded and annealed tubing. Nevertheless, test
coupons were all ground to a 120-grit nish.
For exposure times less than 10,000 h, weight
changes as a function of exposure time for
Fe-Ni-Cr alloys as well as 9Cr steel and Monel in
comparison with some nickel-base alloys are
(a) Attack too small for analysis. Source: Ref 96
Fig. 5.71
Optical micrographs showing typical surface conditions for (a) HR160 alloy, (b) alloy 230, and (c)
alloy 601 after testing at 650 C (1200 F) for 10,000 h in H249CO-2H2O (ac = 18.9). Magnification bar represents 20 m for
(a), 200 m for (b) and (c). Source: Ref 96
Fig. 5.72
Typical surface conditions for (a) alloy 214

after 5707 h, (b) alloy HR120 after 190 h, and
(c) alloy 800H after 925 h at 650 C (1200 F) in H2-49CO-2H2O
(ac = 18.9). Source: Ref 96
pg 135
26/10/2007 12:23PM Plate # 0
pg 136
summarized in Fig. 5.73. Nickel-base alloys

except alloy 600 performed much better than
Fe-Ni-Cr alloys. Nickel-base alloy 600 performed poorly because of its low chromium
content (1516%). MA956 (Fe-20Cr-4.5AlY2O3 ODS alloy) performed very well. The test
results are also presented in Fig. 5.74 in terms of
maximum pitting depth as a function of exposure

time. Alloys 690, 617, and MA956 showed the
minimum pitting attack. Figure 5.75 shows
the test results for some additional alloys
and some better performing alloys tested up
to 16,000 hours (Ref 99). Alloys 690, 263, and
MA956 were observed to show pits with
20
263
617
0
825
864
MA 956
690
601
600
40
9Cr-1Mo
60
K-500
Mass change, mg/cm2
20
DS
803
800
80
100
330
120
0
1000
2000
3000
4000
5000
6000
7000
8000
9000
10000
Exposure time, h
Fig. 5.73
Mass change as a function of exposure time for various alloys including 9Cr steel, Ni-Cr alloy (Monel K-500), Fe-Ni-Cr
alloys, and Ni-base alloys tested at 621 C (1150 F) in H2-80CO. All surfaces were ground to a 120-grit finish prior to
testing. Source: Ref 98
762
30
800
803
601
o
-1M
20
508
330
15
381
864
10
254
82
5
690
617
600
127
Maximum pitting depth, m
635
DS
9Cr
Maximum pitting depth, mils
25
MA956
0
0
K500
5
0
1000
2000
3000
4000
5000
6000
7000
8000
127
9000
Exposure time, h
Fig. 5.74
Maximum pit depths as a function of exposure time for various alloys including 9Cr steel,Fe-Ni-Cr alloys and Ni-base alloys
tested at 621 C (1150 F) in H2-80CO. All surfaces were ground to a 120-grit finish prior to testing. Source: Ref 98
26/10/2007 12:24PM Plate # 0
pg 137
25
24
23
22
21
20
19
18
17
16
15
14
13
12
11
10
9
8
7
6
5
4
3
2
1
0
1
800
than that of alloy 693 (Fig. 5.76). It should be

cautioned in interpreting pit depth data, when the
specimen surface has shown general wastage
without the original specimen surface surrounding the pit, pitting depth data thus measured then
becomes questionable. This would particularly
be the case after very long exposure when pits
have spread throughout the specimen surface.
In reviewing the data summarized in Fig. 5.75
and 5.76, one alloy of interest was alloy 671
(Ni-46Cr), which was found to exhibit excellent
metal dusting resistance. Unfortunately, the test
on this alloy was terminated at close to 10,000 h.
This alloy exhibited low pit depths and very low
metal loss rates (lower than alloy 693). Grabke et
al. (Ref 100) also found that Ni-50Cr alloy was
very resistant to metal dusting, showing a thin
chromium oxide scale (about 3 m thick) after
testing in H2-49CO-2H2O at 650 C (1200 F)
(ac = 18.9) for 10,000 h with no evidence of
metal dusting or carburization. Only minute
amounts of coke were observed on the specimen
surface. Under this test condition, alloy 601
had already suffered metal dusting attack after
only 1993 h of exposure. Similarly, a very
thin chromium oxide scale was observed in
DS
610
803
559
601
508
33
956
263
457
690
356
305
254
86
5
82
203
TD
758
LCE
625
152
C276
671
102
754
617
693
600
0
Pit depth, m
406
602CA
400
Pit depth, mils
signicant pit depths after 14,00016,000 h of

exposure. Alloys 617 and 693, on the other hand,
still exhibited pits with insignicant pit depths.
Alloy 693 was a recently developed commercial
alloy targeting for metal dusting environments by
adding about 3% Al to alloy 690 with same
amount of Cr (about 30%) but slightly lowered
Fe. The alloy was found to perform much better
than alloy 690. It is understandable that high Cr
and high Al in nickel-base alloys should perform
well in metal dusting environments. However, it
was surprising to nd that alloy 617 with only
22% Cr (quite normal level chromium for hightemperature alloys) and fairly low aluminum
content (about 1.3%) exhibited a pitting depth
similar to that of alloy 693. In metal dusting
testing by Natesan and Zeng (Ref 95) involving
several nickel-base alloys including alloy 617,
the test results, which were presented as weight
loss, showed alloy 617 was worse than alloy 601,
602CA, and several other nickel-base alloys
(Fig. 5.70). When Baker and Smith (Ref 98)
presented their test results in terms of weight loss
rate as a function of exposure time, alloy 617 was
found to exhibit metal loss rate (in the range of
alloys 263 and MA956) signicantly higher
51
0
2000
4000
6000
8000
10000
12000
14000
16000
18000
Exposure time, h
Fig. 5.75
Maximum pit depths as a function of exposure time for various alloys including 9Cr steel, Fe-Ni-Cr alloys, and Ni-base alloys
tested at 621 C (1150 F) in H2-80CO for exposure time up to 16,000 h. All surfaces were ground to a 120-grit finish prior to
testing. Source: Ref 99
26/10/2007 12:24PM Plate # 0
pg 138
0
80
6 0
82 00
5
101
33
40
100
80
4
75
864
DS
60
Mass loss rate, mg/cm2h
102
625
103
690
C 276
58
263
602CA
MA
TD
617
956
104
105
693
671
0
2000
4000
6000
8000
10000
12000
14000
16000
18000
Exposure time, h
Fig. 5.76
Metal loss rate as a function of the exposure time for various alloys tested at 621 C (1150 F) in H2-80CO for exposure time
up to 16,000 h. All surfaces were ground to a 120-grit finish prior to testing. Source: Ref 99
Cr-5Fe-1Y2O3 after 10,000 h of exposure in the

same test, showing no evidence of metal dusting
attack.
5.3.4 Effects of Surface Conditions and
Finish
From the discussion so far on metal dusting,
formation of a good, protective chromium oxide
scale is a very effective way to provide protection
against metal dusting attack. Since the temperature range for metal dusting attack is quite low
and chromium diffusivities are relatively low at
these temperatures, rapid formation of a good,
protective chromium oxide scale is critical.
Accordingly, the concentration of chromium at
the surface of the metal becomes important.
Higher chromium concentration at the metal
surface provides faster formation of a protective
chromium oxide scale. Thus, alloys with higher
bulk chromium concentration (higher surface
chromium concentration) are more resistant to
metal dusting as was discussed in the previous
section. Also, for the same alloy, the ne-grained
structure exhibits better metal dusting resistance
than the coarse-grained structure, as illustrated in
Fig. 5.65. This is because grain boundaries provide fast diffusion paths for chromium to reach
to the metal surface. A ne-grain-sized material
will have more grain boundaries and, thus, more
chromium reaching to the metal surface to form a
better chromium oxide scale faster than a coarsegrained material, thus, resisting metal dusting
attack much better.
The surface of the metal can also be prepared
by grinding or machining to produce a thin coldworked layer with high density of dislocations,
which also provide fast diffusion paths for
chromium to reach to the metal surface to form a
protective chromium oxide scale. As a result,
metal dusting is greatly improved when the surface is ground or machined. The benecial effect
of the ground surface condition is illustrated
Fig. 5.77 and 5.78 in a thermogravimetric study
by Grabke et al. (Ref 101). For both alloy
800 and Type 310SS, the as-ground specimen
was most resistant to metal dusting compared
with the as-received surface (cold rolled) and
electropolished surface. Electropolished surface
condition was the worst because of possible
surface depletion in chromium during electropolishing. Both alloy 800 and Type 310SS were
26/10/2007 12:24PM Plate # 0
tested under the same condition; Type 310

was much more resistant to metal dusting than
alloy 800. This is primarily because Type
310SS contains more chromium. Specimens
with ground surface nish, alloy 601, showed
better metal dusting resistance than those with
electropolished surface nish, as shown in
Fig. 5.79 (Ref 94). In this study, the black
sample (believed to be a black-annealed sample)
performed as poorly as the electropolished
10
Alloy 800
8
Mass gain, mg/cm2
Electropolished
Cold rolled
6
Ground
10
15
20
25
Time, h
Fig. 5.77
Thermogravimetric testing in H2-24CO-2H2O at

600 C (1112 F) for alloy 800 (Fe-21Cr-32Ni) in
three different surface conditions: as-ground surface (to 600-grit),
cold-rolled (as-received surface), and electropolished surface.
Source: Ref 101
sample. The black-annealed sample, as in

electropolished specimen, exhibited surface
chromium depletion due to annealing in air (i.e.,
black annealing). The formation of chromium
oxide scales during annealing in air results in
some chromium depletion in the surface layer.
Similar results were obtained by Baker
and Smith (Ref 98) in their study on alloy 601
with different surface conditions, as shown in
Fig. 5.80 and 5.81. In this study, samples from
as-received sheet manufactured from blackannealing and pickling (i.e., annealing in air
followed by acid pickling to remove oxide
scales), which were identied as as-produced
(annealed+ pickled), were slightly better than
those black-annealed samples and electropolished samples. Black annealing and acid
pickling can often result in surface depletion in
chromium. These processes are often involved
in producing sheet products by hot rolling.
When sheet products are produced via cold
rolling, the sheet is typically bright annealed.
Under this condition, no chromium oxide scales
formed on the sheet metal; thus no chromium
depletion occurred at the metal surface. For highaluminum or high-silicon alloys, bright-annealing atmospheres typically exhibit high enough
oxygen potentials (or dew points) that a thin
aluminum or silicon oxide lm can form, which
can be exceedingly benecial to the subsequent
service in metal dusting environments. This
was discussed earlier for HR160. Alloy 214,
unfortunately with too low a chromium level,
showed no benecial effect from bright annealing in its resistance to metal dusting attack.
Mass gain, mg/cm2
25Cr-20Ni - Steel
5.3.5 Metal Dusting Behavior of Weldments

Selection of an appropriate ller metal for
metal dusting environments is also critical, since
Electropolished
2
Cold rolled
Ground
0
0
Fig. 5.78
10
15
Time, h
20
25
Thermogravimetric testing in H2-24CO-2H2O at

600 C (1112 F) for Type 310SS (Fe-25Cr-20Ni)
in three different surface conditions: as-ground surface (to 600grit), cold-rolled (as-received surface), and electropolished surface. Source: Ref 101
Fig. 5.79
Metal dusting behavior at 650 C (1200 F) in

H2-49CO-2H2O for alloy 601 in three different
surface conditions: as-ground surface, elecropolished surface,
and black-annealed surface. Source: Ref 94
pg 139
26/10/2007 12:24PM Plate # 0
pg 140
not every wrought alloy has a matching ller

metal. Thus, it is necessary to select a ller metal
that is at least as good as, but preferably better
than, the wrought alloy selected for the application. The suitable ller metal requires not only
good resistance to metal dusting but also good
weldability. Also, in many cases, machining or

grinding the weld joint may not be possible,
particularly in tubular butt welds where inside
diameter (ID) grinding or machining is not
feasible after joining. Furthermore, alloys that
resist metal dusting contain high aluminum or
Ground
Fig. 5.80
Metal dusting behavior in terms of weight change tested at 621 C (1150 F) in H2-80CO for alloy 601 in various surface
conditions. Source: Ref 98
Fig. 5.81
Metal dusting behavior in terms of weight loss rate tested at 621 C (1150 F) in H2-80CO for alloy 601 in various surface
conditions. Source: Ref 98
26/10/2007 12:24PM Plate # 0
silicon along with high chromium, as discussed

in previous sections. If matching ller metals are
available for these alloys, the weldability of these
high-aluminum or high-silicon ller metals can
be an issue.
Grabke et al. (Ref 101) investigated the metal
dusting behavior of different weldments involving base metals of alloys 800H, 600H, 601, and
602CA using different ller metals, such as alloy
82 (Nicrofer S 7020 or UTP 068 HH) and alloy
602CA (Nicrofer 6025 or UTP 6225 Al). Weldment specimens included alloy 800H to alloy
800H with alloy 82 ller metal, alloy 600H to
alloy 600H with alloy 82 ller metal, alloy 601H
to alloy 601H with alloy 602CA, and alloy 800H
to alloy 602CA with alloy 602CA as a cap layer
and alloy 82 ller metal for root pass and ller.
Welding processes involved were listed as TIG
welding (i.e., gas-tungsten arc welding) and
MMAW (probably shielded metal arc welding).
The surface nish for most weldment specimens
1.0
no H2S
0.1 ppm H2S

0.3 ppm H2S 0.5 ppm H2S 0.75 ppm H2S
m/A, mg/cm2
0.8
0.6
0.4
1 ppm H2S
0.2
0.0
50
100
150
200
250
were brushed, with one weldment being

ground, one being sandblasted, and one being
pickled. Tests were conducted at 600 and 650 C
(1112 and 1200 F) in H2-24CO-2H2O. In
almost all cases, metal dusting was initiated at the
interface between the weld metal and the base
metal (i.e., the heat-affected zone surface). Alloy
82 weld metal was found to be more resistant to
metal dusting than alloy 800H and alloy 600H.
The authors (Ref 101) also concluded that TIG
welding led to a better resistance to metal dusting
than hand-welding, and grinding led to a
modest delay in initiation of metal dusting attack.
5.3.6 Effects of Sulfur on Metal Dusting
Sulfur, which has a strong tendency to segregate to the surface and grain boundaries, can
act as an inhibitor to metal dusting. Hochman
(Ref 67) rst reported benecial effect of sulfur
against metal dusting attack. Extensive investigations on the effect of sulfur on metal dusting
behavior of iron were carried out by Grabke
and Muller-Lorenz (Ref 102), Schneider et al.
(Ref 103), Schneider et al. (Ref 104), and
Schneider and Grabke (Ref 105). The mechanism for sulfur to inhibit metal dusting attack,
as proposed by Grabke and Muller-Lorenz
(Ref 102), is the absorption of sulfur on the
surface of cementite suppresses the nucleation of
graphite, thus inhibiting metal dusting attack.
Figure 5.82 shows the benecial effect of H2S on
the kinetics of metal dusting of iron at 500 C
300
Time, h
Fig. 5.82
Effect of H2S on metal dusting behavior of iron

at 500 C (932 F) in H2-CO-H2O-H2S gas mixture
(ac = 100). Source: Ref 103
1.50
no H2S
0.7 ppm H2S
0.1 ppm
H2S
1 ppm H2S
0.5 ppm
H2S
m/A, mg/cm2
1.25
1.00
0.75
5 ppm H2S
0.50
15 ppm H2S
0.25
0.00
T=700 C
ac = 100
Fig. 5.83
50
100
150 200
Time, h
250
300
350
Effect of H2S on metal dusting behavior of iron at

700 C (1292 F) in H2-CH4-H2S gas mixture (ac =
100). Source: Ref 104
Fig. 5.84
Effect of H2S on metal dusting behavior of iron in

terms of pH2 S =pH2 versus 1/T. Open data points
represent that the onset of metal dusting was retarded for more
than 48 h, while the solid data points represent metal dusting
without retardation. The hatched region represents a transition to
an iron surface saturated with sulfur. Source: Ref 74
pg 141
26/10/2007 12:24PM Plate # 0
(932 F) in H2-CO-H2O-H2S gas mixture (ac =

100) (Ref 103), and Fig. 5.83 shows the similar
effect for iron at 700 C (1292 F) in H2-CH4H2S gas mixture (ac = 100) (Ref 103). Grabke
(Ref 74) provides metal dusting behavior of iron
in terms of pH2 S =pH2 ratio as a function of temperature (1/T) in Fig. 5.84. The gure shows the
region when metal dusting was avoided by sulfur
injection under certain pH2 S =pH2 ratios at different temperatures from about 500 to 1000 C (932
to 1832 F). Injecting a right amount of sulfur
into the environment to retard metal dusting
without causing accelerated suldation attack is a
balancing act.
5.4 Summary
Metals and alloys are generally susceptible to
carburization when exposed to environments
containing CO, CH4, or other hydrocarbon gases
at elevated temperatures. Carburization typically
results in the formation of internal carbides in the
matrix as well as boundaries, causing the alloy
to lose its room-temperature ductility and/or
creep-rupture strengths.
Fe-Ni-Cr alloys are widely used for processing
equipment to resist carburization in the petrochemical industry. The cast 25Cr-20Ni, HK40,
was once the workhorse of pyrolysis furnace
tubes in ethylene cracking operations. Many
modications based on HK40 have been developed and used now with improved carburization
resistance as well as increased creep-rupture
strengths. These alloy modications involve the
use of alloying elements such as, titanium,
niobium, tungsten, molybdenum, and silicon, as
well as increases in nickel and/or chromium.
Increasing nickel in Fe-Ni-Cr alloys improves
carburization resistance. Nickel reduces the diffusivity of carbon in Fe-Ni-Cr alloys. Nickel also
reduces the solubility of carbon in Fe-Ni alloys.
Among these alloying elements, silicon is the
most effective in improving carburization resistance. This is attributed to the formation of
SiO2 scale, which is more impervious to carbon
ingress than Cr2O3 scale. However, when the
silicon content in the alloy is too high, the
weldability of the alloy can become a serious
issue.
Aluminum is another alloying element that
can signicantly improve carburization resistance. However, effectiveness generally requires
about 4% or higher aluminum, the amount
needed to form a continuous Al2O3 scale.
Alumina formers (i.e., alloys forming Al2O3

scale, such as alloy 214 and MA956) are signicantly better than chromia formers (i.e., alloys
forming Cr2O3 scale).
Surface nish plays an important role in
carburization resistance. For cast products,
machining the metal surface can signicantly
reduce carburization attack. Injecting sulfur
compounds (e.g., 50 to 100 ppm) into the processing gas stream is also effective in reducing
carburization attack.
Metal dusting is another form of carburization
attack; it typically causes an alloy to suffer pitting
attack and/or thinning. The metal beneath the
pitted area generally shows carburization. The
corrosion products typically consist of carbon
soots, metal particles, carbides, and oxides. The
environment in which metal dusting occurs
generally contains H2, CO, CO2, and H2O with
high carbon activities (i.e., aC > 1). Stagnant gas
conditions can be conducive in initiating metal
dusting attack. The metal temperatures at which
metal dusting occurs are between 430 and 900 C
(800 and 1650 F). Metal dusting data for various
commercial alloys are presented. Nickel-base
alloys containing high chromium and high
aluminum (e.g., alloys 602CA and 693) or containing high chromium and high silicon (e.g.,
alloy HR160) showed excellent resistance
to metal dusting attack. Surface conditions
also play an important role in metal dusting
resistance. Sulfur may also retard metal dusting
attack.
REFERENCES
1. Metals Handbook, 8th ed., Vol 2, American

Society for Metals, 1964, p 93
2. L.J. Haga, Heat Treat., Dec 1986, p 6
3. G.E. Moller and C.W. Warren, Paper No. 237,
Corrosion/81, NACE, 1981
4. A.J. McNab, Hydrocarbon Process., Dec
1987, p 43
5. G.L. Swales, in Behavior of High Temperature Alloys in Aggressive Environments, Proc.
1979 Petten International Conference, I.
Kirman et al., Ed., The Metals Society,
London, 1980, p 45
6. R.A. Perkins, W.C. Coons, and F.J. Radd, in
Properties of High Temperature Alloys, Proc.
1976 Fall Meeting of the Electrochemical
Society, The Electrochemical Society
7. R.C. Schueler, Hydrocarbon Process., Aug
1972, p 73
pg 142
26/10/2007 12:24PM Plate # 0
8. F.A. Prange, Corrosion, Vol 15 (No. 12),

Dec 1959, p 619t
9. G.Y. Lai, M.F. Rothman, and D.E.
Fluck, Paper No. 14, Corrosion/85, NACE,
1985
10. O.J. Dunmore, Proc. UK Corrosion Conf.
1982, Institution of Corrosion Science and
Technology, Birmingham, UK, 1982, p 101
11. F.N. Mazandarany and G.Y. Lai, Nucl.
Technol., Vol 43, 1979, p 349
12. K. Natesan, Nucl. Technol., Vol 28, 1976,
p 441
13. K. Natesan and T.F. Kassner, Metall. Trans.,
Vol 4, 1973, p 2557
14. HSC, Chemistry for Windows, Version 6.0,
A. Roine, Outokumpu Technology, Finland,
www.outokumputechnology.com, accessed
Dec 2006
15. ChemSage, Version 4.16, GTT-Technologies, Aachen, 1998
16. P.L. Hemmings and R.A. Perkins, Thermodynamic Phase Stability Diagrams for
the Analysis of Corrosion Reactions in
Coal Gasication/Combustion Atmospheres,
Report FP-539, EPRI, Palo Alto, CA, 1977
17. R. Hultgren, P. Desai, D. Hawkins, M.
Gleiser, and K.K. Kelley, Selected Values of
the Thermodynamic Properties of Elements
and Binary Alloys, ASM, 1973
18. L.C. Browning, T.W. Dewitt, and P.H.
Emmett, J. Am. Chem. Soc., Vol 72, 1950,
p 4211
19. K.H. Jack, J. Iron Steel Inst., Vol 169, 1951,
p 26
20. H.J. Christ, Experimental Characterization
and Computer-Based Description of the
Carburization Behaviour of the Austenitic
Stainless Steel AISI 304L, Mater. Corros.,
Vol 49, 1998, p 258
21. R.G. Coltters, Mater. Sci. Eng., Vol 76,
1985, p 1
22. S.R. Shatynski, Oxid. Met., Vol 13 (No. 2),
1979, p 105
23. W.F. Chu and A. Rahmel, Oxid. Met., Vol
15, 1981, p 331
24. Y. Nishiyama, N. Otsuka, and T. Nishizawa,
Carburization Resistance of Austenitic
Alloys in CH4 -CO2 -H2 Gas Mixtures at
Elevated Temperatures, Corrosion, Vol 59
(No. 8), 2003, p 688
25. H.J. Grabke, Carburization: A High
Temperature Corrosion Phenomenon, MTI
Publications No. 52, Materials Technology
Institute of the Chemical Process Industries,
St. Louis, Missouri, 1998
26. S. Forseth and P. Kofstad, Carburization

of Fe-Ni-Cr Steels in CH4-H2 Mixtures at
8501000 C, Mater. Corros., Vol 49, 1998,
p 266
27. D. Jakobi and R. Gommans, Typical Failures
in Pyrolysis Coils for Ethylene Cracking,
Mater. Corros., Vol 54 (No. 11), 2003, p 881
28. R.H. Krikke, J. Horing, and K. Smit, Mater.
Perform., Aug 1976, p 9
29. J. Klower and U. Heubner, Carburization of
Nickel-Base Alloys and its Effects on the
Mechanical Properties, Mater. Corros., Vol
49, 1998, p 237
30. H.J. Grabke, Metallic Corrosion, Proc.
Eighth Int. Cong. Metallic Corrosion, Vol
III, Mainz, Germany, Sept 611, 1981
31. I. Wolf and H.J. Grabke, Solid State Commun., Vol 54, 1985, p 5
32. C.M. Schillmoller, Chem. Eng., Jan 6, 1986,
p 87
33. D.J. Hall, M.K. Hossain, and J.J. Jones,
Mater. Perform., Jan 1985, p 25
34. J.A. Thuillier, Mater. Perform., Nov 1976,
p9
35. J.F. Norton and J. Barnes, in Corrosion in
Fossil Fuel Systems, I.G. Wright, Ed., The
Electrochemical Society, 1983, p 277
36. O. Van der Biest, J.M. Harrison, and J.F.
Norton, in Behavior of High Temperature
Alloys in Aggressive Environments, Proc.
International Conference (Patten, The
Netherlands), Oct 1518, 1979, The Metal
Society, London, 1980, p 681
37. J.M. Harrison, J.F. Norton, R.T. Derricott,
and J.B. Marriott, Werkst. Korros., Vol 30,
1979, p 785
38. O. Demel, E. Keil, and P. Kostecki, SGAW
Report No. 2538, Osterreichische, Studiengesellschaft fur Atomenergie, Lenaugasse
10, A-1082 Wien, Forschungszentrum Seibersdorf, Institut fur Metallurgie
39. R.P. Smith, Trans. Met. AIME, Vol 224,
1962, p 10
40. H.J. Grabke, U. Gravenhorst, and W. Steinkusch, Werkst. Korros., Vol 27, 1976, p 291
41. S.K. Bose and H.J. Grabke, Z. Metallkde.,
Vol 69, 1978, p 8
42. C. Steel and W. Engel, AFS Int. Cast Metals
J., Sept 1981, p 28
43. L.H. Wolfe, Laboratory Investigation of
High Temperature Alloy Failure Mechanisms, Paper No. 12, Corrosion/77, NACE,
1977
44. I.Y. Khandros, R.G. Bayer, and C.A. Smith,
pg 143
26/10/2007 12:24PM Plate # 0
45. W. Steinkrsch, Werkst. Korros., Vol 30,

1979, p 837
46. L.H. Wolfe, Mater. Perform., April 1978,
p 38
47. U. Van den Bruck and C.M. Schillmoller,
48. R.H. Kane, Paper No. 266, Corrosion/83,
NACE, 1983
49. J.F. Mason, J.J. Moran, and E.N. Skinner,
Corrosion, Vol 16, 1960, p 593t
50. R.H. Kane and J.C. Hosier, Carburization
Resistance of Some Wrought NickelContaining Alloys in Simulated Industrial
Environments, Inco Alloys International
Technical Report, Inco Alloys International,
Huntington, WV, 1985
51. D.R.G. Mitchell, D.J. Young, and W.
Kleemann, Carburization of Heat-Resistant
Steels, Mater. Corros., Vol 49, 1998, p 231
52. G.Y. Lai, in High Temperature Corrosion in
Energy Systems, Proc. TMS-AIME Symposium, M.F. Rothman, Ed., The Metallurgical
53. G.Y. Lai, B. Li, B. Gleeson, and H.L.
Craig, Proposed Standard Carburization
Test Method, Paper No. 3473, Corrosion/
2003, NACE International, 2003
54. R.H. Kane, G.M. McColvin, T.J. Kelly, and
J.M. Davison, Paper No. 12, Corrosion/84,
NACE, 1984
55. R. Kirchheiner, D.J. Young, P. Becker, and
R.N. Durham, Improved Oxidation and
Coking Resistance of a New Alumina
Forming Alloy 60HT for the Petrochemical
Industry, Paper No. 5428, Corrosion/2005,
2005
56. F. Pons and M. Hugo, Paper No. 272,
57. F.N. Mazandarany and G.Y. Lai,
Corrosion Behavior of Selected Structural
Materials in a Simulated Steam-Cycle
HTGR Helium Environment, GA-A14446,
General Atomic Company, San Diego, CA,
Oct 1977
58. T.A. Ramanarayanan, R.A. Petkovic, J.D.
Mumford, and A. Ozekcin, Carburization
of High Chromium Alloys, Mater. Corros.,
Vol 49, 1998, p 226
59. H.J. Grabke, R. Moller, and A. Schnaas,
Werkst. Korr., Vol 30, 1979, p 794
60. H.J. Grabke, E.M. Peterson, and S.R.
Srinivasan, Surf. Sci., Vol 67, 1977, p 501
61. T.A. Ramanarayanan and D.J. Srolovitz,
Carburization Mechanisms of High Chromium Alloys, J. Electrochem. Soc.: Solid-
62.
63.
64.
65.
66.
67.
68.
69.
70.
71.
72.
73.
74.
75.
76.
State Sci. Technol., Vol 132 (No. 9), 1985,

p 2268
H.J. Grabke and D. Jakobi, High Temperature Corrosion of Cracking Tubes, Mater.
Corros., Vol 53, 2002, p 494
V. Guttmann and H. Schonherr, Creep
Properties of Two 25Cr-20Ni Cast Alloys
in Air and under Carburizing Conditions,
High Temp. Technol., Vol 3 (No. 2), 1985,
p 79
D. Jakobi and R. Gommans, Typical Failures
Mater. Corros., Vol 54, 2003, p 881
N.G. Taylor, V. Guttmann, and R.C. Hurst,
The Creep Ductility and Fracture of Carburized Alloy 800H at High Temperatures,
in High Temperature Alloys: Their Exploitable Potential, J.B. Marriott, M. Merz,
J. Nihoul, and J. Ward, Ed., Elsevier Applied
Science, 1987, p 475
R.F. Hochman, Proc. Fourth International
Congress on Metal Corrosion, NACE, 1972,
p 258
R.F. Hochman, Proc. Symposium on
Properties of High Temperature Alloys with
Emphasis on Environmental Effects, Z.A.
Foroulis and F.S. Pettit, Ed., The Electrochemical Society, 1977, p 715
J.C. Nava Paz and H.J. Grabke, Metal
Dusting, Oxid. Met., Vol 39, 1993, p 437
H.J. Grabke, R. Krajak, E.M. MullerLorenz, Metal Dusting of High Temperature
Alloys, Werkst. Korros., Vol 44, 1993, p 89
H.J. Grabke, C.B. Bracho-Troconis, and
E.M. Muller-Lorenz, Metal Dusting of Low
Alloy Steels, Werkst. Korros., Vol 45, 1994,
p 215
H.J. Grabke, Metal Dusting of Low- and
High-Alloy Steels, Corrosion, Vol 51, 1995,
p 711
H.J. Grabke, R. Krajak, E.M. MullerLorenz, and S. Straub, Metal Dusting of
Nickel-Base Alloys, Mater. Corros., Vol 47,
1996, p 495
H.J. Grabke, Thermodynamics, Mechanisms and Kinetics of Metal Dusting, Mater.
Corros., Vol 49, 1998, p 303
H.J. Grabke, Corrosion by Carbonaceous
Gases, Carburization and Metal Dusting,
and Methods of Prevention, Mater. High
Temp., Vol 17 (No. 4), 2000, p 483
H.J. Grabke, Metal Dusting, Mater. Corros.,
Vol 54, 2003, p 736
H.J. Grabke and G. Tauber, Arch. Eisenhttenwes., Vol 46, 1975, p 215
pg 144
26/10/2007 12:24PM Plate # 0
77. S.R. Shatynski and H.J. Grabke, Arch.

Eisenhttenwes., Vol 49, 1978, p 129
78. F. Bonnet, F. Ropital, Y. Berthier, and
P. Marcus, Filamentous Carbon Formation
Caused by Catalytic Metal Particles from
Iron Oxide, Mater. Corros., Vol 54, 2003,
p 870
79. E. Pippel, J. Woltersdorf, and H.J. Grabke,
Microprocesses of Metal Dusting on IronNickel Alloys and Their Dependence on the
Alloy Composition, Mater. Corros., Vol 54,
2003, p 747
80. F. Eberle and R.D. Wylie, Corrosion, Vol 15
(No. 12), 1959, p 622t
81. W.B. Hoyt and R.H. Caughey, Corrosion,
Vol 15 (No. 12), 1959, p 627t
82. O.J. Dunmore, A Case History of a Metal
Dusting Problem Which Led to a Boiler
Failure, presented at UK Corrosion/82
(London), Nov 1618, 1982
83. H.J. Grabke and M. Spiegel, Mater. Corros.,
Vol 54, 2003, p 799
84. G.Y. Lai, J. Met., Vol 37 (No. 7), 1985,
p 14
85. E. Fitzer, W. Frohs, and M. Heine, Carbon,
Vol 24 (No. 4), 1986, p 387
86. M. Maier and J.F. Norton, Studies Concerned with the Metal Dusting of Fe-Cr-Ni
Materials, Paper No. 75, Corrosion/99,
NACE International, 1999
87. R.A. Perkins, R.A. Padgett, and N.K. Tunali,
Met. Trans. AIME, Vol 4, 1973, p 2535
88. P.J. Albery and C.W. Haworth, Met. Sci.,
Vol 8, 1974, p 407
89. A.F. Smith, Met. Sci., Vol 9, 1975, p 375,
425
90. C.H. Toh, P.R. Munroe, and D.J. Young,
Metal Dusting of Fe-Cr and Fe-Ni-Cr Alloys
under Cyclic Conditions, Oxid. Met., Vol 58
(No. 1/2), 2002, p 1
91. H.J. Grabke, E.M. Muller-Lorenz, S.
Strauss, E. Pippel, and J. Woltersdorf,
Effects of Grain Size, Cold Working, and
Surface Finish on the Metal-Dusting Resistance of Steels, Oxid. Met., Vol 50 (No 3/4),
1998, p 241
92. A.S. Fabiszewski, W.R. Warkins, J.J.
Hoffman, and S.W. Dean, The Effect of
Temperature and Gas Composition on the
Metal Dusting Susceptibility of Various
Alloys, Paper No. 532, Corrosion/2000,

93. S. Strauss and H.J. Grabke, Mater. Corros.,
Vol 49, 1998, p 321
94. J. Klower, H.J. Grabke, E.M. MullerLorenz, and D.C. Agarwal, Metal Dusting
and Carburization Resistance of NickelBase Alloys, Paper No. 139, Corrosion/97,
95. K. Natesan and Z. Zeng, Metal Dusting
Performance of Structural Alloys, Paper
96. D.L. Klarstrom and H.J. Grabke, The Metal
Dusting Behavior of Several High Temperature Alloys, Paper No. 1379, Corrosion/
97. B.A. Baker and G.D. Smith, Metal Dusting
Behavior of High-Temperature Alloys,
Paper No. 54, Corrosion/99, NACE International, 1999
98. B.A. Baker and G.D. Smith, Alloy Selection
for Environments Which Promote Metal
Dusting, Paper No. 257, Corrosion/2000,
99. B.A. Baker, G.D. Smith, V.W. Hartmann, L.
E. Shoemaker, and S.A. McCoy, NickelBase Material Solutions to Metal Dusting
Problems, Paper No. 2394, Corrosion/2002,
100. H.J. Grabke, H.P. Martinz, and E.M.
Muller-Lorenz, Metal Dusting Resistance
of High Chromium Alloys, Mater. Corros.,
Vol 54, 2003, p 860
101. H.J. Grabke, E.M. Muller-Lorenz, and M.
Zinke, Metal Dusting Behaviour of Welded
Ni-Base Alloys with Different Surface
Finish, Mater. Corros., Vol 34, 2003, p 785
102. H.J. Grabke and E.M. Muller-Lorenz, Steel
Res., Vol 66, 1995, p 252
103. A. Schneider, H. Viefhaus, G. Inden, H.J.
Grabke, and E.M. Muller-Lorenz, Mater.
Corros., Vol 49, 1998, p 330
104. A. Schneider, H. Viefhaus, and G. Inden,
Surface Analytical Studies of Metal
Dusting of Iron in CH4-H2-H2S Mixtures,
105. A. Schneider and H.J. Grabke, Effect of
H2S on Metal Dusting of Iron, Mater.
Corros., Vol 54, 2003, p 793
pg 145
26/10/2007 12:25PM Plate # 0

DOI: 10.1361/hcma2007p147

CHAPTER 6
Corrosion by Halogen and

Hydrogen Halides
6.1 Introduction
Many metals react readily with halogen gases
at elevated temperatures to form volatile metal
halides. Some metal halides also exhibit low
melting points, and some even sublime at relatively low temperatures. As a result, alloys
containing elements that form volatile or lowmelting-point halides may suffer severe hightemperature corrosion.
Industrial environments often contain halogen
gases. Because of high vapor pressures of many
metal chlorides, the chlorination process is an
important step in processing metallurgical ores
for production of titanium, zirconium, tantalum,
niobium, and tungsten (Ref 13). Chlorination is
also used for extraction of nickel from iron
laterites (Ref 4) and for detinning of tin plate
(Ref 5). Production of TiO2 and SiO2 involves
processing environments containing Cl2 and/or
HCl, along with O2 and other combustion products. Calcining operations in the production
of (a) lanthanum, cerium, and neodymium for
electronic and magnetic materials and (b) ceramic ferrites for permanent magnets, frequently
generate environments contaminated with
chlorine. In the chemical process industry, many
processing streams also contain chlorine. Manufacturing of ethylene dichloride (EDC), which is
an intermediate for the production of vinyl
chloride monomer, generates chlorine-bearing
environments. The reactor vessels, calciners,
and other processing equipment for the above
operations require alloys that are resistant to
high-temperature chloridation attack.
In the manufacture of uorine-containing
compounds, such as uorocarbon plastics, refrigerants, and re-extinguishing agents, the
processing equipment requires alloys with good
resistance to corrosion by uorine and hydrogen
uorides at elevated temperatures. During the
rening operation in the production of uranium,

UO2 is uorinated at elevated temperatures
(e.g., 500 to 600 C) with HF to produce UF4 or
UF6 for separation of U235 (Ref 6). The materials of construction for this processing equipment must resist corrosion by HF at both high
and low temperatures.
This chapter reviews data on primarily gaseous corrosion by halogen and hydrogen halides.
In many high-temperature industrial processes
where fuels and/or feedstocks are often contaminated with impurities, such as alkaline
metals, halogen may react readily with these
metals to form halide salts. Corrosion reactions
under these conditions are to be discussed in later
chapters dealing with high-temperature corrosion in gas turbines, coal-red boilers, oil-red
boilers, waste-to-energy boilers, black liquor
recovery boilers, and so forth.

The relative stabilities of various chlorides,
uorides, bromides, and iodides are presented in
Fig. 6.1 to 6.3, in terms of standard free energies
of formation (G) versus temperature (Ref 7).
The gures also include the information about
the melting points and boiling points of halides.
Some of the halides exhibit low melting and
boiling points. Tables 6.1 to 6.4 list melting
and boiling points of some important halides and
oxyhalides (Ref 8, 9). Since industrial environments typically have nite oxygen activities, the
thermodynamic phase stability diagram in terms
of the M-O-Cl (M represents metal, O represents
oxygen, and Cl represents chlorine) system is
quite useful in describing the possible corrosion
products that may form on the metal. These
diagrams can be constructed for major alloying
pg 147
26/10/2007 12:25PM Plate # 0
elements of high-temperature alloys (e.g., Fe,

Ni, Co, Cr, Mo, W, Al, Si, etc.). Commercial
computer programs, such as HSC (Ref 10) and
Chemsage (Ref 11), are available for construction of these phase-stability diagrams. Figure 6.4
shows a Ni-O-Cl stability diagram in terms of
pO2 and pCl2 at 723 C (1333 F) (Ref 12). The
Fig. 6.1
diagram denes the boundary ( pO2 10715 atm)

between Ni and NiO, the boundary (pCl2 1078 )
between Ni and NiCl2, and the boundary between NiO and NiCl2. This means that if the
equilibrium gas mixture of the environment
exhibits an oxygen potential ( pO2 ) higher than
about 1015 atm, NiO may form. Similarly,
Standard free energies of formation for chlorides. Source: Ref 7
pg 148
26/10/2007 12:25PM Plate # 0
Chapter 6: Corrosion by Halogen and Hydrogen Halides / 149
NiCl2 may form if the environment exhibits

a chlorine potential ( pCl2 ) higher than about
108 atm when the oxygen potential ( pO2 ) is
below about 1015 atm.
With phase-stability diagrams, one can predict
the possible phases that are likely to form on
Fig. 6.2
a metal. Lets consider an example where nickel

is exposed to an environment consisting of air
with 0.1% Cl2 at 723 C (1333 F). The environment would be at the location that identies
pO2 being very close to 100 and pCl2 being 103 in
the 723 C Ni-O-Cl stability diagram (Fig. 6.4).
Standard free energies of formation for fluorides. Source: Ref 7
pg 149
26/10/2007 12:25PM Plate # 0
Based on the stability diagram shown in Fig. 6.4,

the environment is in the NiO regime. NiO
oxide is expected to form on the surface of
nickel when exposed to this environment. If NiO
oxide scale formed on nickel is defect free, the
pCl2 at the interface between NiO and Ni would
Fig. 6.3
be below the value needed to form NiCl2, thus

preventing the formation of NiCl2. However,
when defects and cracks develop in the NiO
scale, Cl2 can permeate through the oxide scale
and reach the nickel with high enough pCl2 to
form NiCl2, initiating chloridation attack. The
Standard free energies of formation for bromides and iodides. Source: Ref 7
pg 150
26/10/2007 12:25PM Plate # 0
pg 151
NiCl2 formed at 723 C (1333 F) would be

a solid phase. As shown in Table 6.1, NiCl2
melts at relatively high temperature (1030 C)
compared with other metal chlorides, such as
CoCl2 (740 C), FeCl2 (676 C), and FeCl3
(303 C).
The phase-stability diagram for Co-O-Cl at
723 C (1333 F) is shown in Fig. 6.5 (Ref 12),
and those for Cr-O-Cl and Fe-O-Cl at 600 C
(1112 F) are shown in Fig. 6.6 and 6.7,
respectively (Ref 13). In addition to low melting
points, some chlorides have low boiling points
also. In Fig. 6.7, FeCl3 is in a gaseous state at
600 C (1112 F). Figure 6.8 shows WCl4 and
WO2Cl2 in a gaseous state at 900 C (1650 F).
Some uorides, bromides, and iodides also have
either low melting points or low boiling points
(see Tables 6.26.4). Corrosion reactions can be
signicantly increased when the corrosion product is in either a liquid state or a gaseous state.
Furthermore, many halides, although in a solid
state, may exhibit high vapor pressures. When
the corrosion product in a solid state exhibits a
Table 6.2 Melting points, temperatures at which

fluoride vapor pressure reaches 104 atm, and
boiling points of various fluorides
Fluorides
Melting
point, C (F)
Temperature
at 104 atm, C (F)
Boiling
point, C (F)
FeF2
FeF3
NiF2
CoF2
CrF2
CrF3
CuF
MoF5
MoF6
WF6
TiF3
TiF4
AlF3
SiF4
MnF2
ZrF4
NbF5
HfF4
TaF5
NaF
KF
LiF
MgF2
CaF2
BaF2
ZnF2
PbF2
1020 (1868)
1027 (1881)
1450 (2642)
1250 (2282)
894 (1641)
1404 (2559)
908 (1666)
64 (147)
17 (63)
2 (36)
1200 (2192)
90 (130)
920 (1688)
932 (1710)
79 (174)
97 (207)
992 (1818)
857 (1575)
848 (1558)
1263 (2305)
1418 (2584)
1290 (2354)
875 (1607)
822 (1512)
906 (1663)
673 (1243)
939 (1722)
962 (1764)
928 (1702)
855 (1571)
24 (75)
82 (116)
91 (132)
108 (226)
825 (1517)
160 (256)
992 (1818)
583 (1081)
615 (1139)
37 (99)
928 (1702)
788 (1450)
908 (1666)
1257 (2295)
1429 (2604)
1581 (2878)
806 (1483)
664 (1227)
1800 (3272)
36 (97)
17 (63)
283 (541)
1270 (2318)
95 (139)
233 (451)
1704 (3099)
1502 (2736)
1681 (3058)
2230 (4046)
2500 (4532)
2215 (4019)
1500 (2732)
1293 (2359)
Source: Ref 8, Ref 9
Table 6.1 Melting points, temperatures at

which chloride vapor pressure reaches 104 atm,
and boiling points of various chlorides
Chlorides
Melting
point, C (F)
Temperature
at 10 4 atm, C (F)
Boiling
point, C (F)
FeCl2
FeCl3
NiCl2
CoCl2
CrCl2
CrCl3
CrO2Cl2
CuCl
MoCl5
WCl5
WCl6
TiCl2
TiCl3
TiCl4
AlCl3
SiCl4
MnCl2
ZrCl4
NbCl5
NbCl4
TaCl5
HfCl4
CCl4
NaCl
KCl
LiCl
MgCl2
CaCl2
BaCl2
ZnCl2
PbCl2
676 (1249)
303 (577)
1030 (1886)
740 (1364)
820 (1508)
1150 (2102)
95 (139)
430 (806)
194 (381)
240 (464)
280 (536)
1025 (1877)
730 (1346)
23 (9.4)
193 (379)
70 (94)
652 (1206)
483 (901)
205 (401)
216 (421)
434 (813)
24 (11)
801 (1474)
772 (1422)
610 (1130)
714 (1317)
772 (1422)
962 (1764)
318 (604)
498 (928)
536 (997)
167 (333)
607 (1125)
587 (1089)
741 (1366)
611 (1132)
387 (729)
58 (136)
72 (162)
11 (52)
921 (1690)
454 (849)
38 (36)
76 (169)
87 (125)
607 (1125)
146 (295)
239 (462)
80 (176)
132 (297)
80 (112)
742 (1368)
706 (1303)
665 (1229)
663 (1225)
1039 (1902)
349 (660)
484 (903)
1026 (1879)
319 (606)
987 (1809)
1025 (1877)
1300 (2372)
945 (1733)
117 (243)
1690 (3074)
268 (514)
337 (639)
750 (1382)
137 (279)
58 (136)
1190 (2174)
250 (482)
455 (851)
240 (464)
77 (171)
1465 (2669)
1407 (2565)
1382 (2520)
1418 (2584)
2000 (3632)
1830 (3326)
732 (1350)
954 (1749)
Table 6.3 Melting points, temperatures at which

bromide vapor pressure reaches 104 atm, and
boiling points of various bromides
Bromides
Melting
point, C (F)
Temperature
at 104 atm, C (F)
Boiling
point, C (F)
FeBr2
FeBr3
NiBr2
CoBr2
CrBr2
CrBr3
CrBr4
CuBr
WBr5
WBr6
AlBr3
SiBr4
MnBr2
ZrBr4
NbBr5
HfBr4
TiBr4
TaBr5
NaBr
KBr
LiBr
MgBr2
CaBr2
BaBr2
ZnBr2
PbBr2
689 (1272)
965 (1769)
678 (1252)
842 (1548)
>800 (1472)
488 (910)
276 (529)
309 (588)
97 (207)
5 (41)
695 (1283)
450 (842)
267 (513)
424 (795)
267 (513)
750 (1382)
740 (1364)
550 (1022)
710 (1310)
742 (1368)
854 (1569)
398 (748)
373 (703)
509 (948)
156 (313)
580 (1076)
716 (1321)
615 (1139)
516 (961)
435 (815)
53 (127)
169 (336)
137 (279)
145 (293)
690 (1274)
671 (1240)
630 (1166)
626 (1159)
320 (608)
432 (810)
974 (1785)
919 (1686)
1318 (2404)
255 (491)
153 (307)
361 (682)
232 (450)
347 (657)
1393 (2539)
1383 (2521)
1310 (2390)
1230 (2246)
1800 (3272)
650 (1202)
914 (1677)
26/10/2007 12:25PM Plate # 0
high vapor pressure, the corrosion reaction can

increase as well. It is generally considered that
when the vapor pressure of a corrosion product
reaches to 104 atm, the corrosion reaction can
become signicant. Tables 6.1 to 6.4 also include
the temperature at which a halides vapor pressure reaches to 104 atm.
The chlorine partial pressure (pCl2 ) needed to
form a chloride corrosion product with 104 atm
will be lower than that needed to form either
solid or liquid chloride. This is illustrated in
Fig. 6.9 (Ref 14). Bender and Schutze (Ref 15)
Table 6.4 Melting points, temperatures at

which iodide vapor pressure reaches 104 atm,
and boiling points of various iodides
Iodides
Melting
point, C (F)
Temperature
at 104 atm, C (F)
Boiling
point, C (F)
FeI2
NiI2
CoI2
CrI2
CrI3
CuI
AlI3
SiI4
MnI2
ZrI4
NbI4
HfI4
TaI5
NaI
KI
LiI
MgI2
CaI2
BaI2
ZnI2
PbI2
594 (1101)
780 (1436)
515 (959)
869 (1596)
>600 (1112)
588 (1090)
191 (376)
122 (252)
613 (1135)
499 (930)
503 (937)
449 (840)
496 (925)
660 (1220)
685 (1265)
469 (876)
650 (1202)
740 (1364)
712 (1314)
446 (835)
412 (774)
476 (889)
702 (1296)
529 (984)
144 (291)
55 (131)
227 (441)
244 (471)
208 (406)
651 (1204)
629 (1164)
621 (1150)
425 (797)
316 (601)
397 (747)
935 (1715)
1207 (2205)
385 (725)
301 (574)
545 (1013)
1304 (2379)
1330 (2426)
1170 (2138)
730 (1346)
872 (1602)
Fig. 6.4
Phase stability diagram for Ni-O-Cl system at 723 C

(1333 F). Both corrosion products (NiO and NiCl2)
are solid phases at this temperature. Source: Ref 12
Fig. 6.5
Phase stability diagram for Co-O-Cl system at 723 C

(1333 F). All the corrosion products (i.e., CoO,
Co3O4, and CoCl2) are solid phases at this temperature. Source:
Ref 12
Fig. 6.6
Phase stability diagram for Cr-O-Cl system at 600 C

(1112 F). All the corrosion products (i.e., Cr2O3,
CrCl2, and CrCl3) are solid phases at this temperature. Source:
Ref 13
Fig. 6.7
Phase stability diagram for Fe-O-Cl system at 600 C

(1112 F). All the corrosion products are solid phases
except FeCl3 at this temperature. Source: Ref 13
pg 152
26/10/2007 12:26PM Plate # 0
Fig. 6.8
Phase stability diagram for the W-O-Cl system at 900 C (1650 F), showing tungsten chloride (WCl4) and tungsten
oxychloride (WO2Cl2) in a gaseous state.
Fig. 6.9
Phase stability diagram for Fe-O-Cl system at 700 C

(1292 F). The solid line represents the boundary for
forming solid FeCl2, while the dotted line represents the boundary
for forming FeCl2 with 104 atm pressure. Source: Ref 14
extensively examined the phase-stability diagrams involving the vapor pressures of chlorides
reaching 104 atm for various alloying elements.
The authors termed these phase-stability diagrams quasi-stability diagrams. Figure 6.10
shows the pCl2 values that are needed to form
NiCl2 with 104 atm at various temperatures
(Ref 15). Also included in the gure are two
environments (air + 0.1% Cl2, and air + 2% Cl2)
for illustration purpose. For both of these
environments, NiO, not NiCl2 with vapor
pressures of 104 atm or higher, is likely to form
on nickel at 500 and 650 C. However, at 800 C
and higher, NiCl2 with vapor pressures of 104

atm and higher (not NiO) is to form on nickel.
Molybdenum oxychlorides also exhibit high
vapor pressures. Figure 6.11 shows the quasistability diagram for Mo-O-Cl system at 500 C
(Ref 15). The gure also shows that MoO2Cl2
with vapor pressures of 104 atm and higher is to
form in the environments of air + 0.1% Cl2, and
air + 2% Cl2.
The metal-halogen reaction differs from
other reactions, such as oxidation, in that most
reaction products are characteristic of high vapor
pressures and, in some cases, low melting points.
The volatile halides (reaction products) formed
on the metal surface can no longer provide
protection against further corrosion. This is in
contrast to most oxides, which generally exhibit
very low vapor pressures and high melting
points. Furthermore, many halides exhibit low
melting points. Once the reaction products
become molten, the alloy loses all protection against further corrosion, leading to rapid
attack.
6.3 Corrosion in Cl2- and HCl-Bearing

Environments
6.3.1 Corrosion in Cl2 Environments
(No O2)
This section focuses on the corrosion behavior
of alloys in environments containing essentially
Cl2 with no oxygen (O2) present. An excellent
review of halogen corrosion data up to the
mid-1970s was presented by Daniel and Rapp
pg 153
26/10/2007 12:26PM Plate # 0
(Ref 8). They summarized the test results

obtained by various authors (Ref 1619) on
chloridation of iron (Table 6.5). It should be
noted that the data in Table 6.5 were obtained
from short-duration tests, ranging from a few
minutes to hours. Use of these test results for
extrapolation to 1 year could result in signicant
errors. They should not be used to estimate the
service life of equipment, but should instead be
used for comparison purposes. As illustrated in
Table 6.5, iron exhibited little corrosion attack
in Cl2 at temperatures up to 250 C (480 F).
Above 250 C, corrosion rates abruptly increased. Iron forms two types of chlorides:
FeCl2 and FeCl3. The melting and boiling
points of FeCl2 are 676 and 1026 C (1249 and
1879 F), respectively. FeCl3, on the other hand,
is extremely unstable. Its melting and boiling
points are 303 and 319 C (577 and 606 F),
respectively. Bohlken et al. (Ref 20, 21) suggested that the abrupt increase in the corrosion rate of iron in Cl2 at temperatures above
250 C (482 F) was related to the formation
of FeCl3.
p(NiCl2) 104 bar
500 C
2
650 C
Log p(Cl2), bar
4
NiCl2
850 C
1000 C
8
10
NiO
12
Ni
14
30
25
20
15
10
Log p(O2), bar
Fig. 6.10
Quasi-stability diagram for Ni-O-Cl system for NiCl2 with vapor pressures of 104 atm (bar) and higher at temperatures from
500 to 1000 C (932 to 1832 F). Source: Ref 15
p(MoOxCly) 104 bar

MoOCl4
Log p(Cl2), bar
2
4
MoOCl3
6
MoCl4
8
MoO2Cl2
10
Mo
12
40
35
30
25
MoO2
20
15
MoO3
10
Log p(O2), bar
Fig. 6.11
Quasi-stability diagram for Mo-O-Cl system for vapor pressures of chlorides and oxychlorides being 104 atm (bar) and
higher at 800 C (1472 F). Source: Ref 15
pg 154
26/10/2007 12:26PM Plate # 0
pg 155
Adding chromium and/or nickel to iron

improves the alloys resistance to chloridation
attack. Ferritic and austenitic stainless steels can
resist chloridation attack at higher temperatures
than cast iron and carbon steels. Brown et al.
(Ref 22) reported a corrosion rate of about
600 mpy at 232 C (450 F) for both carbon
steel and cast iron. A ferritic stainless steel
(Fe-17Cr) showed a corrosion rate of about
79 mpy at 360 C (680 F), and a titaniumstabilized austenitic stainless steel showed a rate
of 24 mpy at 418 C (784 F) (Ref 8). The
results of studies on several stainless steels by
Tseitlin and Strunkin (Ref 16) and Brown et al.
(Ref 22) were summarized by Daniel and Rapp
(Ref 8) and are presented in Table 6.6.
Adequate aluminum when added to iron to
form aluminum oxide scales is also benecial in
improving the chloridation resistance. Han and
Cho (Ref 23) studied corrosion behavior of
Fe3Al (Fe-12.11%Al) in Ar-1%Cl2 at 750, 800,
and 900 C using a thermogravimetric method.
The alloy behaved similarly at three different
temperatures, showing an initial stage of an
incubation time before the breakaway corrosion showing a drastic weight loss, as shown in
Fig. 6.12. The authors observed a thin protective
Al2O3 scale during the initial incubation stage,
and nonprotective oxide scales (Al2O3, Fe2O3)
Table 6.5
Temperature,
C (F)
77 (170)
166 (330)
198 (388)
200 (392)
230 (446)
240 (464)
240 (464)
247 (477)
251 (484)
255 (491)
260 (500)
268 (514)
279 (534)
285 (545)
285 (545)
302 (576)
304 (579)
310 (590)
323 (613)
327 (621)
381 (718)
381 (718)
540 (1004)
540 (1004)
595 (1103)
599 (1110)
and small amounts of FeCl3 and FeCl2 formed at

the breakaway stage. Also, at the breakaway
stage, the specimen showed aluminum depletion
at the metal/scale interface. The thin aluminum
oxide scale was observed to form on the specimen during heating to the test temperature with
argon gas owing through the test chamber
(approximately 2 h) prior to switching to the test
gas. The test gas (i.e., Ar-1%Cl2) was found to
contain 1 ppm O2. Thus, under the test condition,
Al2O3 could form on the metal as well as FeCl2,
as shown in Fig. 6.13.
Nickel and nickel-base alloys are widely used
in chlorine-bearing environments. The corrosion
behavior of nickel in chlorine at various temperatures was analyzed by Daniel and Rapp (Ref
8), using the test results of Downey et al. (Ref
24), Tseitlin and Strunkin (Ref 16), and McKinley and Shuler (Ref 25) (see Table 6.7). At temperatures up to 500 C (930 F), nickel showed
relatively low corrosion rates. Corrosion rates
became suddenly and signicantly higher at
temperatures over 500 C (930 F).
Nickel reacts with chlorine to form NiCl2,
which exhibits relatively high melting point
(1030 C) compared to FeCl2 and FeCl3 (676
and 303 C, respectively). This may be an important factor, making nickel much more resistant to chloridation attack than iron. Brown et al.
Corrosion of iron in chlorine

Flow rate,
(L/min)
100
100
100
15
100
100
15
100
100
120
15
120
120
15
15
120
120
15
120
120
120
120
15
15
120
120
pCl2,
atm
1
1
1
1
1
1
1
I
1
(c)
1
(c)
(c)
1
1
(c)
(c)
1
(c)
(c)
(c)
(c)
(c)
(c)
(c)
(c)
Diluent
gas
None
None
None
None
None
None
None
None
None
Ar
None
He
Ar
None
None
He
Ar
None
Ar
He
Ar
He
N2
N2
Ar
He
Test duration,
min
480
015, 15480
015, 15480
360
015, 15480
015, 15480
360
015, 15480
480
360
360
60
60
Linear rate constant(a),

m/min
4
3 10
3.3 103, 3.8 104
5.2 103, 3.7 104
2 104
8.5 103, 2.2 104
9.5 103, 2.4 104
2 10-4
1 102, 1.9 104
1.55
0.94
4
2 10
1.78
2.45
4
4 10
20.4
4.04
4.17
3.9
8.94
11.4
40.4
65.3
6.6
1.5
187
624
Corrosion rate(b),
mm/yr (mpy)
0.16 (6.3)
0.20 (7.9)
0.19 (7.5)
0.11 (4.3)
0.12 (4.7)
0.13 (5.1)
0.11 (4.3)
0.10 (3.9)
820 (32 in.)
490 (19 in.)
0.11 (4.3 mils)
940 (37 in.)
1,300 (51 in.)
0.21 (8 mils)
11,000 (433 in.)
2,100 (83 in.)
2,200 (87 in.)
1,700 (67 in.)
4,700 (185 in.)
6,000 (236 in.)
21,000 (827 in.)
34,000 (1,338 in.)
3,500 (138 in.)
830 (33 in.)
98,000 (3,858 in.)
328,000 (13,000 in.)
(a) Rate constants are given as metal loss rates, m/min. (b) Estimated metal loss after 1 yr of exposure. (c) Tests were conducted with chlorine partial pressures up to 0.3 atm
and total pressures of 1.0 atm. However, the metal loss rates were extrapolated to 1.0 atm chlorine pressure. Source: Ref 8
26/10/2007 12:26PM Plate # 0
pg 156
Table 6.6
Corrosion of stainless steels in chlorine(a)
Alloy
Ferritic stainless
(Fe-17Cr)
Austenitic stainless
(Fe-18Cr-9Ni-Ti)
(Fe-18Cr-8Ni-Mo)
(Fe-18Cr-8Ni)
Temperature,
C (F)
Flow rate,
L/min
Linear rate
constant(b),
m/min
Corrosion rate(c),
mm/yr (mpy)
300 (572)
360 (680)
440 (824)
540 (1,004)
418 (784)
450 (842)
480 (896)
535 (995)
640 (1,184)
315 (599)
340 (644)
400 (752)
450 (842)
480 (896)
290 (554)
315 (599)
340 (644)
400 (752)
450 (842)
15
15
15
15
15
15
15
15
15
28
28
28
28
28
28
28
28
28
28
4 104
3.8 103
6.7 102
1.35
1.1 103
4.3 102
0.13
0.47
46
1.4 103
2.9 103
5.9 103
2.9 102
5.9 102
1.5 103
2.9 103
5.9 103
2.9 102
5.9 102
0.2 (7.9)
2 (79)
40 (1.6 in.)
700 (28 in.)
0.6 (24)
20 (787)
70 (2.8 in.)
200 (7.9 in.)
20,000 (787 in.)
0.8 (31)
1.5 (59)
3 (118)
15 (590)
30 (1.2 in.)
0.8 (31)
1.5 (59)
3 (118)
15 (590)
30 (1.2 in.)
(a) Chlorine pressure was approximately 1.0 atm. (b) Duration of these tests was 60360 min for the rst two alloys and 1201200 min for the last two alloys. (c) Estimated
metal loss after one year of exposure. Source: Rcf 8
10
1% Cl2/Ar
FeCl3
0
5
Log p Cl , g
2
750 C
10
FeCl2
10
AlCl3
15
15
900 C
800 C
750 C
Al2O3
Fe3O4
Fe
20
Fe2O3
Al
20
25
25
0
10
20
30
40
50
60
70
30
60
Exposure time, h
Fig. 6.12
Thermogravimetric results for Fe3Al (Fe-12.11%Al)

tested in Ar-1%Cl2 at 750, 800, and 900 C.
Source: Ref 23
(Ref 22) conducted short-term laboratory tests

in chlorine on various commercial alloys. The
results (see Table 6.8) suggested that, in an
environment of 100% Cl2, carbon steel and
cast iron are useful at temperatures up to 150 to
200 C (300 to 400 F) only. The 18-8 stainless
steels can be used at higher temperaturesup to
320 to 430 C (600 to 800 F). Nickel and
nickel-base alloys (e.g., Ni-Cr-Fe, Ni-Mo, and
Ni-Cr-Mo alloys) were most resistant. The
benecial effect of nickel on the resistance
of chloridation attack in Cl2 environments is
FeO
50
40
30
20
10
10
Log p O , g
2
Fig. 6.13
Phase stability for Al-O-Cl system at 750 C. The

solid circle indicates the test environment. Source:
Ref 23
illustrated in Fig. 6.14 (Ref 26). This trend is

also reected in long-term tests (Table 6.9).
Alloy 600 is the most commonly used alloy for
high-temperature services in Cl2 environments.
Figure 6.15 shows the corrosion rates of alloy
600 in dry chlorine gas as a function of temperature (Ref 22). MTI Publication MS-3
(Ref 27) suggests corrosion guidelines for Ni200,
alloy 600, alloy 400, Type 304, and steel in dry
chlorine gas applications, as shown in Fig. 6.16.
Tu et al. (Ref 28) performed phase analysis
using x-ray diffraction on the external corrosion
26/10/2007 12:26PM Plate # 0
pg 157
Table 6.7
Corrosion of nickel in chlorine

pCl2,
atm
Temperature,
C (F)
350 (662)
375 (707)
400 (752)
407 (765)
433 (811)
465 (869)
485 (905)
500 (932)
500 (932)
525 (977)
540 (1004)
550 (1022)
550 (1022)
660 (1220)
740 (1364)
770 (1418)
~1
~1
~1
~1
~1
Test duration,
min
0.13
0.13
0.13
0.13
0.13
0.13
0.13
0.13
atm at 16 cm3/min
0.13
atm at 16 cm3/min
0.13
0.13
atm at 16 cm3/min
atm at 16 cm3/min
atm at 16 cm3/min
30005000
3600
3600
3600
60
60
Corrosion rate(a),
mm/yr (mpy)
0.0012 (0.05)(b)
0.0014 (0.06)(b)
0.0036 (0.14)(b)
0.005 (0.2)(b)
0.0058 (0.2)(b)
0.107 (4.2)
0.142 (5.6)
0.319 (12.6)
0.492 (19.4)
2.41 (95)
2.07 (82)
6.02 (237)
5.85 (230)
40.3 (1,587)
120 (4,724)
2,150 (84,646)
(a) Estimated metal loss after 1 year of exposure to chlorine. (b) These estimates are probably low. Source: Ref 8
Table 6.8 Corrosion of selected alloys

in chlorine
Approximate temperature, C (F),
at which given corrosion rate is exceeded
Alloy
0.8 mm/yr
(30 mpy)
1.5 mm/yr
(60 mpy)
3.0 mm/yr
(120 mpy)
15 mm/yr
(600 mpy)
Nickel
Alloy 600
Alloy B
Alloy C
Chromel A
Alloy 400
18-8 Mo
18-8
Carbon steel
Cast iron
510 (950)
510 (950)
510 (950)
480 (900)
425 (800)
400 (750)
315 (600)
288 (550)
120 (250)
93 (200)
538 (1000)
538 (1000)
538 (1000)
538 (1000)
480 (900)
455 (850)
345 (650)
315 (600)
175 (350)
120 (250)
593 (1100)
565 (1050)
593 (1100)
565 (1050)
538 (1000)
480 (900)
400 (750)
345 (650)
205 (400)
175 (350)
650 (1200)
650 (1200)
650 (1200)
650 (1200)
620 (1150)
538 (1000)
455 (850)
400 (750)
230 (450)
230 (450)
Source: Ref 22
products formed on Ni-4Cr alloy when exposed

in 105 Pa (1 atm) Cl2 at 575 and 700 C. The
authors found that the scales consisted of mainly
NiCl2, CrCl3, and CrCl2. The deposits on the
quartz test assembly during testing of Ni-4Cr
alloy were also analyzed. These deposits were
mainly NiCl2 and CrCl3 with very little CrCl2,
indicating both NiCl2 and CrCl3 are major vapor
phases during testing Ni-4Cr alloy.
6.3.2 Corrosion in O2-Cl2 Environments
Many industrial environments may contain
both chlorine and oxygen. Metals generally follow a parabolic rate law by forming condensed
phases of oxides, if the environment is free of
chlorine. With the presence of both oxygen and
chlorine, corrosion of metals then involves a
combination of condensed oxides and volatile
chlorides. Depending on the relative amounts of
oxides and chlorides formed, corrosion can
follow either a paralinear rate law (a combination

of weight gain due to oxidation and weight loss
due to chlorination) or a linear rate law due to
chlorination. This is illustrated by the results of
Maloney and McNallan (Ref 29) on corrosion
of cobalt in Ar-50O2-Cl2 mixtures (Fig. 6.17).
As shown in the gure, at high Cl2 levels, the
corrosion products are primarily cobalt chloride
vapor, causing the weight loss to follow a linear
rate law. Because of volatile corrosion products,
the reaction rate can be highly dependent on the
gas ow rate. McNallan and Liang (Ref 30)
showed that CoO specimens exhibited increased
linear weight loss rates with increasing gas
velocity when exposed in the Ar-O2-1Cl2 mixtures with pO2 being 0.01 and 0.15 atm pressures
at 723 C (1000 K). Furthermore, the oxygen
partial pressure (pO2 ) of 0.01 atm resulted in
higher weight loss rates than that of 0.15 atm.
This is also illustrated in Fig. 6.18 (Ref 31),
showing the corrosion of cobalt in Ar-O2-1Cl2
mixtures with three different concentrations
(1, 10, and 50% O2) of oxygen at 650 C
(1200 F). When the environment contained 10
and 50% O2, the corrosion reaction involved
mainly the formation of cobalt oxide, thus following an approximate parabolic rate. When the
oxygen concentration reduced to 1%, the corrosion reaction involved mainly volatile CoCl2,
thus following an approximate linear weight
loss with time. The gure also shows that
the linear weight loss agreed very well with the
volatilization of CoCl2. It should be noted that in
the above test environments containing 10 and
50% O2, the corrosion reaction, which followed a
parabolic rate due to formation of condensed
cobalt oxides (Fig. 6.18), involved only a very
26/10/2007 12:26PM Plate # 0
mg/cm2
690
Fig. 6.14
Effect of nickel on the corrosion resistance of alloys in Ar-30Cl2 at 704 C (1300 F) for 24 h. Source: Ref 26
Table 6.9 Corrosion of several alloys

in Ar-30Cl2 after 500 h at 400 to 704 C
(750 to 1300 F)(a)
Descaled weight loss, mg/cm2
Alloy
400 C
(750 F)
500 C
(930 F)
600 C
(1110 F)
704 C
(1300 F)(a)
Ni-201
600
601
625
617
800
310
304
347
0.2
0.02
0.3
0.7
0.6
6
28
108
215
0.3
5
3
7
7
13
370
1100
Total
47101
127180
85200
200270
97
160
215
180
190
890
820
>1000
Total
(a) 24 h test period. Source: Ref 26
short-term test (2 h). It is extremely likely that

upon longer exposure times the cobalt oxide
scales may eventually crack (or spall), thus
allowing chlorine gas to reach the underlying
metal and causing formation of volatile CoCl2
and resulting in a linear corrosion rate.
Corrosion of pure nickel in O2-Cl2 environments was found to be dominated by the formation of volatile NiCl2 corrosion product, and
the weight loss essentially followed a linear rate
law (Ref 32, 33). Figure 6.19 shows thermogravimetric results at 927 C in Ar-50O2 containing various amounts of Cl2 from 0 to 5%
(Ref 32). The gure also shows that the experimental data were in good agreement with the
pg 158
26/10/2007 12:26PM Plate # 0
/month
theoretical calculation based on vapor pressures

of NiCl2. In their testing of nickel in Ar-50O2-Cl2
containing 0, 0.5, 1.0, 1.75, 3, and 5% Cl2 at 627
to 927 C, Lee and McNallan (Ref 32) observed
a very rapid reaction referred to as ignition at
727 and 827 C in the environments containing
higher concentrations of Cl2. McNallan (Ref 33)
suggested that this rapid corrosion reaction was
caused by the reaction of chlorine with nickel
at the metal/oxide scale interface to form nickel
chloride vapor, which then diffuses out and is
converted into powdery nickel oxide. This rapid
corrosion reaction causes metal temperature
to increase. He further indicated that this rapid
reaction (ignition) can be prevented at temperatures below 727 C due to low vapor pressures
of nickel chloride and can also be prevented at
temperatures higher than 827 C due to formation of a protective oxide scale by rapid oxidation of nickel to reduce the ingress of chlorine
through the oxide scale and thus the formation of
nickel chloride vapor at the metal/oxide scale
interface.
In an O2-Cl2 containing environment, oxidation and chloridation can take place. As discussed in section 6.2, a protective oxide scale can
lower the pCl2 below the value for forming metal
chlorides at the metal/oxide scale interface.
However, cracking and spalling of this oxide
scale resulting from, for example, thermal
Fig. 6.15
Corrosion rate of alloy 600 in Cl2 as a function of

Fig. 6.16
MTI corrosion guidelines for Ni200, alloy 600, alloy 400, Type 304SS and steel in dry chlorine (Cl2) as a function of
temperature. Source: Ref 27. Courtesy of Materials Technology Institute
pg 159
26/10/2007 12:26PM Plate # 0
Fig. 6.17
Corrosion of cobalt in Ar-50O2-Cl2 at 927 C (1700 F). Source: Ref 29
Fig. 6.18
Thermogravimetric results of corrosion of cobalt in

Ar-O2-1Cl2 with 1, 10, and 50% O2 at 650 C
(1200 F). Source: Ref 31
cycling can allow permeation of chlorine to reach

the metal underneath to initiate chloridation
attack. Figure 6.20 illustrates the effect of thermal cycling on the initiation of accelerated
chloridation attack for Fe-20Cr at 927 C in
Ar-20O2-0.5Cl2 (Ref 34). The gure shows the
thermogravimetric results for three separate test
Fig. 6.19 Thermogravimetric results of corrosion of nickel at

927 C in Ar-50O2-Cl2 (0, 0.5, 1.0, 1.75, 3, and 5%
Cl2). Source: Ref 32
runs, showing similar behavior in initiating an
accelerated chloridation attack right after a thermal cycle, which involved cooling the specimen
for 30 min by lowering it from the test temperature (927 C) to 100 C in 5 min. As soon
pg 160
26/10/2007 12:27PM Plate # 0
pg 161
as the test was resumed after the thermal cycle,

specimens (three separate test runs) showed
accelerated chloridation attack.
A signicant amount of data have been generated for commercial alloys in environments
containing both oxygen and chlorine. Table 6.10
summarizes short-term test results generated
in Ar-20O2-2Cl2 at 900 C (1650 F) for 8 h
(Ref 35). The results revealed several interesting trends. The best-performing alloy was an
aluminum-containing alloy 214 (Fe-16Cr-3Fe4.5Al) with a small amount iron. Two worstperforming alloys were cobalt-base alloys (alloys
188 and 6B) containing large amounts of tungsten. Molybdenum-containing nickel-base alloys
also did not perform well. Oh et al. (Ref 36)
attributed this to the formation of oxychlorides
of molybdenum and tungsten, which have very
high vapor pressures. The partial pressures of
WO2Cl2 and MoO2Cl2 in equilibrium with the
oxides (WO3 and MoO3, respectively) and the
test environment (Ar-20O2-2Cl2) at 900 C
(1650 F) were 7.52 102 and 2.1 100 atm,
respectively. Accordingly, alloy 188 (14% W),
alloy C-276 (16% Mo, 4% W), alloy 6B (4.5%
W, 1.5% Mo), alloy X (9% Mo), and alloy S
(14.5% Mo) suffered relatively high rates of
corrosion attack. Simple Fe-Ni-Cr (Type 310 SS)
and Ni-Cr-Fe (alloy 600) performed better than
molybdenum- or tungsten-containing alloys. The
100
Mass change, mg/cm2
80
36 h
60
18 h
40
gravimetric results for representative alloys

are summarized in Fig. 6.21 (Ref 36). Thermodynamic phase stability diagrams showing high
vapor pressures of oxychlorides of tungsten and
molybdenum were presented earlier in Fig. 6.8
and 6.11, respectively.
In order to determine whether molybdenum
oxychlorides would contribute to high corrosion
rates in high-molybdenum-containing nickel
alloys, such as alloy S (Ni-16Cr-14.5Mo) in
O2-Cl2 environments, Jacobson et al. (Ref 37)
used a high-pressure sampling mass spectrometer to measure volatile species produced from
the preoxidized specimen of alloy S with 14.5%
Mo in comparison with alloy 600 (Ni-16Cr-9Fe)
with no Mo during the exposure of Ar-50O21Cl2. Thermogravimetric data for these two
preoxidized alloys under the test condition are
shown in Fig. 6.22, showing a signicantly
higher weight loss rate for alloy S (14.5Mo)
than alloy 600 (no Mo) during the exposure of
the preoxidized specimens to Ar-50O2-1Cl2 at
900 C (1650 F). The mass spectrometer results indicated that MoO2Cl2 along with NiCl2
and CrO2Cl2 were major vapor phases in the
case of alloy S. For alloy 600, NiCl2 and CrO2Cl2
were detected.
Alloy R-41 (nickel-base alloy with 1.5% Al,
3% Ti and 10% Mo) suffered less chloridation
attack than other nickel-base alloys containing
molybdenum despite high molybdenum content
in a short-term test presented in Table 6.10.
However, the results of long-term tests in
Ar-20O2-0.25Cl2 by Rhee et al. (Ref 38) and
McNallan et al. (Ref 39) showed that these
nickel-base alloys with molybdenum, such
as R-41 (Ni-19Cr-11Co-10Mo-1.5Al-3Ti) and
alloy 263 (Ni-20Cr-20Co-5.8Mo-0.5Al-2.2Ti),
eventually suffered severe attack despite the
presence of aluminum and titanium. Figure 6.23
24 h
Table 6.10 Corrosion of selected alloys in

Ar-20O2-2Cl2 at 900 C (1650 F) for 8 h
20
Pure oxygen
0
Alloy
12
16
20
24
28
32
36
Time, h
Fig. 6.20
Thermogravimetric results of three test runs for Fe20Cr alloy tested in Ar-20O2-0.5Cl2 at 927 C isothermally for the first 12 h, followed by a thermal cycle by cooling
the specimen to 100 C for 30 min and raising the specimen to the
test temperature to resume testing. Note that the thermal cycle
resulted in the initiation of an accelerated chloridation attack.
Source: Ref 34
214
R-41
600
310SS
S
X
C-276
6B
188
Metal loss,
mm (mils)
Average metal
affected(a),
mm (mils)
0
0.004 (0.16)
0.012 (0.48)
0.012 (0.48)
0.053 (2.08)
0.020 (0.80)
0.079 (3.12)
0.014 (0.56)
0.014 (0.56)
0.012 (0.48)
0.028 (1.12)
0.035 (1.36)
0.041 (1.60)
0.063 (2.48)
0.071 (2.80)
0.079 (3.12)
0.098 (3.84)
0.116 (4.56)
(a) Metal loss + average internal penetration. Source: Ref 35
26/10/2007 12:27PM Plate # 0
pg 162
310SS
Fig. 6.21
Gravimetric results for selected Fe-, Ni-, and Co-base alloys in Ar-20O2-2Cl2 at 900 C (1650 F). Source: Ref 36
Fig. 6.22
Thermogravimetric data showing weight loss of

two preoxidized specimens during the exposure
in Ar-50O2-1Cl2 at 900 C (1650 F). Source: Ref 37
shows corrosion test results for aluminumcontaining nickel-base alloys with and without
molybdenum, such as 214, 601, R-41, and 263,
tested at 900 C (1650 F) (Ref 39). Test results

for all the alloys tested at 900 and 1000 C (1650
and 1830 F) are summarized in Tables 6.11 and
6.12 (Ref 39).
The benecial effect of aluminum, as well as
the detrimental effect of molybdenum and tungsten, on the resistance to chloridation attack in
oxidizing environments is further substantiated
by the results of long-term tests in another
environment with a higher concentration of Cl2,
as shown in Fig. 6.24 (Ref 35). Similar results
were obtained by Elliott et al. (Ref 40) from tests
conducted in air-2Cl2 at 900 C (1650 F) for
50 h (Fig. 6.25).
Chloridation attack in O2-Cl2 environments at
these high temperatures primarily consisted of
metal wastage and internal penetration for most
alloys, with the exception of Ni-Cr-Mo alloys
containing high levels of molybdenum, such as
alloys S and C-276, which showed mainly metal
wastage with little or no internal penetration. In
general, the scales formed on the alloy surface
were loose when the test specimens were cooled
to room temperature after the exposure test, as
illustrated in Fig. 6.26. Scales were basically
26/10/2007 12:27PM Plate # 0
pg 163
oxides, as shown in Fig. 6.27 and 6.28. Fe-Cr

oxides were found to form on Type 310 SS, while
aluminum-rich oxides were the major oxide
phase along with little nickel-rich oxides that
formed on alloy 214. Internal attack consisted of
voids for some alloys and of oxides for other
alloys. Some alloys appeared to contain both
internal voids and oxides. Figure 6.29 shows
internal void formation in alloy R-41 (Ni-Cr-CoMo-Al-Ti) and alloy 25 (Co-Cr-Ni-W) after
exposure for 50 h at 900 C (1650 F) in air2Cl2. It was suggested that the internal penetration may involve halogen-carbide reactions
and simultaneous void formation (Ref 25).
Figure 6.29 shows some evidence of carbides
being converted into voids during chloridation
attack. Some alloys, however, showed internal
oxides instead of voids, as illustrated in Fig. 6.30.
Elliott et al. (Ref 40) have identied volatile
species of the condensed products removed from
the exit end of the test apparatus during their
investigation in air-2Cl2. Their results are shown
in Table 6.13. No oxychlorides were detected.
Since the analysis of the volatile species was
performed on the condensed phases collected at
the exit end of the test apparatus, oxychlorides
were likely to be in a gaseous state at the exit end,
and thus were not collected. As discussed earlier,

Jacobson et al. (Ref 37) used a high-pressure
sampling mass spectrometer to measure volatile
species from alloy S with 14.5% Mo and alloy
600 (Ni-16Cr-9Fe) with no Mo during the
exposure of Ar-50O2-1Cl2 at 900 C (1650 F)
and found that MoO2Cl2 along with NiCl2 and
CrO2Cl2 were major vapor phases for alloy S
and NiCl2 and CrO2Cl2 for alloy 600.
McNallan et al. (Ref 39) reported corrosion behavior in Ar-20O2-0.25Cl2 at 900 and
1000 C (1650 and 1830 F). This was followed
by a study (Ref 41) using the same environment
to investigate the same alloys at lower temperatures (i.e., 700, 800, and 850 C). The results of
the tests at 700, 800, and 850 C (1290, 1470,
and 1560 F) are summarized in Table 6.14
Table 6.11 Corrosion of various alloys in
Ar-20O2-0.25Cl2 for 400 h at 900 and 1000 C
(1650 and 1830 F)
Weight loss, mg/cm2
Alloy
214
601
600
800H
310SS
556
X
625
R-41
263
188
S
C-276
900 C (1650 F)
1000 C (1830 F)
4.28
20.67
72.08
26.91
47.15
40.29
54.41
99.07
63.83
82.57
139.77
228.21
132.05
9.05
124.99
254.96
87.05
97.40
82.74
153.49
220.09
207.32
229.53
156.30
248.98
298.85
Source: Ref 39
Table 6.12 Depth of attack after 400 h

at 900 and 1000 C (1650 and 1830 F) in
Ar-20O2-0.25Cl2
900 C (1650 F)
Alloy
Fig. 6.23
Corrosion of several aluminum-containing nickelbase alloys with and without molybdenum in

Ar-20O2-0.25Cl2 at 900 C (1650 F). Source: Ref 39
214
601
600
800H
310SS
556
X
625
R-41
263
188
S
C-276
Metal loss,
mm (mils)
Total depth(a),
mm (mils)
0.023 (0.9)
0.150 (5.9)
0.061 (2.4)
0.264 (10.4)
0.127 (5.0)
0.252 (9.9)
0.043 (1.7)
0.191 (7.5)
0.086 (3.4)
0.152 (6.0)
0.046 (1.8)
0.152 (6.0)
0.099 (3.9)
0.218 (8.6)
0.208 (8.2)
0.272 (10.7)
0.114 (4.5)
0.244 (9.6)
0.130 (5.1)
0.193 (7.6)
0.216 (8.5) >0.356 (14.0)
0.315 (12.4) 0.353 (13.9)
0.300 (11.8) 0.320 (12.6)
1000 C (1830 F)
Metal loss,
mm (mils)
Total depth(a),
mm (mils)
0.013 (0.5)
0.203 (8.0)
0.330 (13.0)
0.203 (8.0)
0.191 (7.5)
0.152 (6.0)
0.318 (12.5)
0.356 (14.0)
0.381 (15.0)
0.368 (14.5)
0.254 (10.0)
0.419 (16.5)
0.419 (16.5)
0.051 (2.0)
0.295 (11.6)
0.386 (15.2)
0.424 (16.7)
0.246 (9.7)
0.300 (11.8)
0.434 (17.1)
0.437 (17.2)
0.457 (18.0)
0.424 (16.7)
0.417 (16.4)
0.472 (18.6)
0.450 (17.7)
(a) Metal loss + internal penetration. Source: Ref 39
26/10/2007 12:27PM Plate # 0
Fig. 6.24
Corrosion of several nickel- and cobalt-base alloys in Ar-20O2-1Cl2 at 900 C (1650 F). Source: Ref 35
Fig. 6.25
Corrosion of several iron- and nickel-base alloys

in air-2Cl2 at 900 and 1000 C (1650 and 1830 F)
for 50 h. Source: Ref 40
(Ref 41). The corrosion behavior of alloys as a

function of temperature from 700 to 1000 C
(1290 to 1830 F) can best be summarized in
Fig. 6.31 using three different alloy systems
(Ni-Cr-Mo alloys S, Fe-Ni-Cr alloy 800H, and
Ni-Cr-Al alloy 214). As discussed earlier, refractory metals, such as molybdenum and tungsten, are detrimental to chloridation resistance
in oxidizing environments at high temperatures.

Alloy S was found to be less corrosion resistant
than alloy 800H. However, both alloys S and
800H suffered increasing corrosion attack with
increasing temperatures. This represents a typical
trend for most alloys in oxidizing environments.
One exception is the Ni-Cr-Al system. As illustrated in Fig. 6.31, alloy 214 showed a sudden
decrease in corrosion attack as the test temperature was increased from 900 to 1000 C (1650
to 1830 F). This sharp reduction in corrosion
attack at 1000 C (1830 F) was attributed to the
formation of a protective Al2O3 scale. At lower
temperatures, such as 900 C or less, the kinetics
of Al2O3 formation was not fast enough to form a
protective oxide scale in O2-Cl2 environments.
Formation of a protective Al2O3 scale is favored
at higher temperatures (e.g., 1000 C or higher).
Most of the data presented so far were generated at fairly high temperatures (i.e., 900 C and
higher). Chloridation was quite aggressive at
those high temperatures. Schwalm and Schutze
(Ref 4244) investigated a large number of
commercial alloys at signicantly lower temperatures, varying from 300 to 800 C (572 to
1472 F) for times up to 300 h in air-2Cl2. The
corrosion behavior of various alloys were generated in terms of the decrease in specimen crosssection thickness (i.e., metal loss) and the depth
of internal attack. Extensive characterization of
the corrosion products formed on the alloys was
pg 164
26/10/2007 12:27PM Plate # 0
investigated using SEM/EDX analyses, which

showed distribution of elements including oxygen and chlorine in the corrosion products. Their
results are briey summarized below.
Figure 6.32(a) shows the decrease in metal
cross-section thickness for 2.25Cr-1Mo steel
(10CrMo9 10), alloy 800H, alloy AC66, alloy
45TM, and alloy 690 after testing for 300 h as
Fig. 6.26
a function of temperature. The data on the depth

of internal corrosion attack are presented in
Fig. 6.32(b). The corrosion attack at 300 C
(572 F) was quite negligible after 300 h for
the alloys tested including 2.25Cr-1Mo steel.
Nevertheless, the Cr-Mo steel was found to
exhibit a fragile oxide scale, which contained
Fe, O, and Cl. All other alloys including two
Loose scales on samples of several nickel-base alloys after testing at 900 C (1650 F) in Ar-20O2-1Cl2 for 100 h
Semiquantitative EDX analysis, wt%

Area
1
2
Fig. 6.27
Fe
Cr
Si
85.1
69.5
14.9
29.9
0.6
Scanning electron micrograph showing oxide scale formed on Type 310SS sample exposed at 900 C (1650 F) for 400 h
in Ar-20O2-0.25Cl2. The results of the EDX analysis of the corrosion products on the areas, as marked No. 1 and No. 2,
are tabulated. Magnification bar represents 33.3 m
pg 165
26/10/2007 12:27PM Plate # 0
nickel alloys (alloys 45TM and 690) showed

pitting type of attack. The phase in the pit was
heavily enriched in Cl and O with some Cr, Fe,
and Ni. Figure 6.33 shows the morphology of
the pit along with an x-ray map for Cl. At 500 C
(932 F), 2.25Cr-1Mo steel suffered both signicant thickness reduction and internal attack,
while alloys AC66, 800H, 45TM, and 690
showed little attack. At 650 C, alloy AC66
suffered signicantly more thickness loss than
alloys 800H, 45TM, and 690. The elemental
distribution in the corrosion products for alloy
AC66 (worst alloy in this group) is shown in
Fig. 6.34. The chromium oxides with a layer of
iron-rich oxide that formed on AC66 became
convoluted. Chlorine was detected at the metal/
oxide scale interface and in the metal underneath
the metal/oxide scale interface, where internal
attack was observed. Alloy 690, on the other
hand, showed a continuous chromium oxide
scale. No chlorine was detected. Some internal
attack was observed underneath the oxide scale,

and these internal particles are believed to be
aluminum oxides. The morphology of the corrosion products formed in alloy 690 is shown in
Fig. 6.35.
Corrosion behavior for alloys 59, C-2000, and
HR160 after testing for 300 h as a function of
temperature is shown in Fig. 6.36. All three
alloys exhibited little corrosion attack at 300 and
500 C. At 650 C, both alloys 59 and HR160
continued to exhibit little corrosion attack, while
alloy C-2000 suffered much more corrosion
attack. Both C-2000 (Ni-23Cr-16Mo-1.6Cu) and
59 (Ni-23Cr-16Mo-0.3Al) exhibit similar chemical compositions except C-2000 contains additional 1.6% Cu and alloy 59 contains additional
0.3Al. It is not clear whether Cu in alloy C-2000
was responsible. At 650 C, both alloys 59 and
HR160 were found to perform better than alloys
690 and 45TM (Fig. 6.32). Alloy 59 was found to
exhibit a thin, continuous chromium-rich oxide
Fig. 6.28
Area
Ni
Cr
Al
Fe
1
2
72
6
20
5
4
88
4
1
Scanning electron micrograph showing oxide scale formed on alloy 214 sample tested at 900 C (1650 F) for 400 h in
Ar-20O2-0.25Cl2. The results of the EDX analysis of the corrosion products on the areas, as marked No. 1 and No. 2,
are tabulated. Magnification bar represents 33.3 m
pg 166
26/10/2007 12:27PM Plate # 0
scale after exposure at 800 C for 300 h in air2Cl2. as shown in Fig. 6.37. Also observed was
an outer Fe-, Ni-, and Cr-rich oxide scale. Also
observed were a small amount of chlorine and
slight molybdenum enrichment at the metal/
chromium-oxide scale interface. At 800 C, all
three alloys showed higher corrosion attack.
Alloy 59 continued to perform the best. The
Fig. 6.29
high concentration of molybdenum (16%)

showed no detrimental effect on the alloys
corrosion resistance in this oxidizing environment containing 2% Cl2. Thermodynamically,
this environment at 800 C, molybdenum oxychloride (MoO2Cl2) would be stable, as shown
in Fig. 6.11. Molybdenum oxychloride was believed to contribute to high corrosion rates for
Scanning electron micrographs showing internal void formation in (a) alloy R-41 and (b) alloy 25 after exposure for 50 h at
900 C (1650 F) in air-2Cl2
pg 167
26/10/2007 12:27PM Plate # 0
pg 168
high-molybdenum-containing alloys, such as

alloys C-276, S, and so forth when exposed to
O2-Cl2 environments at 900 C (1650 F) and
higher (Ref 35, 39, 40). The chloride phases
exhibiting 104 atm pressure that are predicted
thermodynamically under this test condition are
NiCl2, CoCl2, CrO2Cl2, FeCl3, and MoO2Cl2
(Table 6.1), while the oxides are Al2O3, SiO2,
Cr2O3, and Fe2O3. HR160 alloy showed an inner
layer of silicon-rich oxide scale underneath the

chromium-rich oxide scale with slight internal
silicon oxides after exposure at 800 C for 300 h
in air-2Cl2, as shown in Fig. 6.38. The gure also
shows that silicon oxides contained what
appeared to be tiny internal voids.
Schwalm and Schutze (Ref 44) also tested
alumina former alloy 214 (Ni-16Cr-3Fe-4.5AlY-Zr), intermetallic Fe3Al (Fe-5.5Cr-15.9Al0.2Zr), and intermetallic TiAl (Ti-36Al). The
corrosion data are shown in terms of crosssection thickness loss (Fig. 6.39a) and internal
attack (Fig. 6.39b) after exposure for 300 h at
temperatures from 300 to 800 C (572 to
1472 F) in air-2Cl2. Alumina former alloy 214
performed very well, comparable to alloy 59
(Fig. 6.36). Fe3Al, while showing little corrosion
attack at 300 and 500 C, suffered severe corrosion at 650 C. At 800 C, Fe3Al showed very
little corrosion again. The oxide scales formed
on Fe3Al at 650 C consisted of nonprotective,
multilayers of Cr-containing Fe2O3, Al2O3,
and Fe2O3. At 800 C, a thin, continuous
aluminum-rich oxide scale was found to form
on Fe3Al when tested at 800 C. It is clear that
Table 6.13 Volatile species of condensed

products(a) after testing at 900 C (1650 F)
for 50 h in air-2Cl2
Area
Ni
Cr
Fe
Al
Ti
Alloy
1
2
3
4
5
1
19
25
29
56
62
66
60
45
26
1
7
12
13
18
27
6
2
10
9
2
1
3
Alloy 214
Alloy 601(b)
Type 310SS
Alloy 800H(b)
Alloy 25
Alloy 625
Alloy 617
Alloy 263
Alloy C-276
Fig. 6.30
Scanning electron micrograph showing oxide

scales and internal oxides for alloy 601 exposed at
900 C (1650 F) for 400 h in Ar-20O2-0.25Cl2. The results of the
EDX analysis of the corrosion products on the areas, as marked
No. 1, No. 2, No. 3, No. 4, and No. 5, are listed.
Major constituents
NiCl2, AlCl3
FeCl32H2O, NiCl26H2O
FeCl32H2O, NiCl26H2O
CoCl2, NiCl26H2O, WCl6
NiCl26H2O, FeCl32H2O, MoCl5
NiCl26H2O, FeCl32H2O, CoCl2
NiCl26H2O, CoCl2, MoCl5, FeCl32H2O
NiCl26H2O, MoCl5
(a) Collected at the downstream, cooler section of the test apparatus. (b) Tested at
Table 6.14 Depth of attack for various alloys after 400 h at 700, 800, and 850 C (1290, 1470,
and 1560 F) in Ar-20O2-0.25Cl2
700 C (1290 F)
Alloy
214
600
800H
310SS
556
S
C-276
188
Source: Ref 41
800 C (1470 F)
850 C (1560 F)
Metal loss,
mm (mils)
Total depth,
mm (mils)
Metal loss,
mm (mils)
Total depth,
mm (mils)
Metal loss.
mm (mils)
Total depth,
mm (mils)
0.010 (0.4)
0.025 (1.0)
0.079 (3.1)
0.033 (1.3)
0.010 (0.4)
0.033 (1.3)
0.081 (3.2)
0.046 (1.8)
0.018 (0.7)
0.020 (0.8)
0.023 (0.9)
0.036 (1.4)
0.020 (0.8)
0.145 (5.7)
0.066 (2.6)
0.058 (2.3)
0.061 (2.4)
0.086 (3.4)
0.046 (1.8)
0.053 (2.1)
0.051 (2.0)
0.150 (5.9)
0.071 (2.8)
0.074 (2.9)
0.018 (0.7)
0.038 (1.5)
0.031 (1.2)
0.031 (1.2)
0.020 (0.8)
0.224 (8.8)
0.163 (6.4)
0.025 (1.0)
0.066 (2.6)
0.132 (5.2)
0.097 (3.8)
0.061 (2.4)
0.079 (3.1)
0.257 (10.1)
0.175 (6.9)
0.264 (10.4)
26/10/2007 12:28PM Plate # 0
higher temperatures are favored for forming

an aluminum oxide scale. Unlike Fe3Al, TiAl
suffered severe corrosion attack at 500, 650,
and 800 C.
In many combustion processes involving
burning fuels or feedstock containing hydrocarbon, combustion environments under substoichiometric combustion conditions most often
would contain CO2. McNallan et al. (Ref 45)
examined the effect of CO2 on the chloridation
resistance of Fe-Cr and Fe-Ni-Cr alloys and
found that CO2 signicantly increased the
alloys metal wastage and internal attack. The
authors compared the environments between
Ar-20O2-2500 ppm (0.25%) Cl2 and Ar-20CO22500 ppm Cl2 for Fe-20Cr and 800H at 927 C
(1700 F). Figure 6.40 shows metal loss data
for both environments as a function of exposure
time. For both alloys, the CO2-Cl2 environment
caused more metal wastage than the O2-Cl2
environment with Fe-20Cr being more severely
affected than alloy 800H. The authors did not
offer an explanation in the paper. It was likely
that when the environment was switched from
Ar-20O2-Cl2 to Ar-20CO2-Cl2 the environment
changed from oxidizing to reducing with its
oxygen potential being reduced to closer to, or in,
the CrCl3 and FeCl2 regimes in the Cr(Fe)-O-C
diagrams, resulting in formation of more volatile
chlorides, thus more metal loss. In addition to

metal loss, the internal attack for alloy 800H was
more severe in the CO2-Cl2 environment than
the O2-Cl2 environment (Fig. 6.41). The authors
hypothesized the mechanism of the internal
corrosion by forming internal chromium carbides, which are then converted to chromium
chlorides with carbon reacting with chlorine to
form more chromium carbides. Thus, internal
corrosion is the result of formation of internal
voids and pores in the metal. Fe-20Cr alloy, on
the other hand, suffered no internal corrosion.
The authors attributed this lack of internal corrosion to the alloys low carbon content and ferritic
structure.
In another paper by McNallan et al. (Ref 46),
the CO2-containing environments were further
examined. The authors observed internal carburization of Type 310SS in Ar-20CO2 at 800 C
for 24 h. When the test was conducted in
Ar-20CO2-Cl2 at 800 C for 24 h, Type 310SS
suffered internal corrosion attack in forms of
voids and carbides. Similar internal attack was
observed for alloy 800. The corrosion, suggested
by the authors, proceeded in two stages with
carburization preceding the chlorine-accelerated
oxidation.
In combustion environments, H2O is invariably present among the combustion products
produced. The oxygen potential is dictated by
partial pressures of oxygen and hydrogen.
Hydrogen reacts with Cl2 to form more stable
HCl molecule. The chloridation behavior in HClcontaining environments is discussed in the next
two sections.
6.3.3 Corrosion in O2-HCl Environments
In some combustion environments, chlorine is
present as HCl instead of Cl2. This section covers
the corrosion data in oxidizing environments
containing HCl. In an oxidizing environment
containing HCl, chlorine partial pressure, pCl2 ,
can be calculated from the equilibrium condition
of the reaction below. The oxide and chloride
phases that are likely to form thermodynamically
can then be determined from the thermodynamic
phase stability diagrams discussed earlier in
section 6.2.
4HCl(g)+O2 (g)=2Cl2 (g)+2H2 O(g)
Fig. 6.31
Corrosion behavior of alloy 214 (Ni-Cr-Al-Y), alloy

S (Ni-Cr-Mo) and alloy 800H (Fe-Ni-Cr) in Ar20O2-0.25Cl2 for 400 h at 7001000 C (12901830 F). Source:
Ref 39 and Ref 41
The effects of adding O2 to HCl environment

were extensively studied by Ihara et al. (Ref 47)
on iron, by Ihara et al. (Ref 48) on nickel, and
by Ihara et al. (Ref 49) on chromium. Adding O2
pg 169
26/10/2007 12:28PM Plate # 0
to HCl signicantly accelerates the corrosion

of iron (Ref 47), as shown in Fig. 6.42. The
increased corrosion was attributed to the formation of highly volatile FeCl3 (Ref 47). For
nickel, adding O2 to HCl does not signicantly
affect the metals corrosion rate, as shown in
Fig. 6.43 (Ref 48). This is due to the formation
of primarily NiCl2 whether it is 100% HCl or
O2-HCl mixtures (Ref 48). Adding O2 to HCl
suppresses the corrosion rate of chromium at
lower temperatures (400600 C) by forming
Cr2O3 and accelerates the corrosion rate at
higher temperatures (700 and 800 C) by forming highly volatile CrCl3 (Ref 49). The corrosion
rate of chromium as a function of O2-HCl
mixtures at different temperatures is shown in

Fig. 6.44 (Ref 49).
Devisme et al. (Ref 50) investigated the effect
of O2 and CO2 in HCl-bearing environments on
the corrosion behavior of four nickel-base alloys,
alloys C-276, 600, 601, and 214. Corrosion data
in terms of metal loss, which was determined
by weight change and metallographic examination of the test specimen, are summarized
in Tables 6.15 and 6.16. Adding 2% O2 to Ar20HCl signicantly increased corrosion attack
for alloy C-276 (Ni-Cr-Mo alloy), but signicantly decreased corrosion attack for alloy 214
(Ni-Cr-Al alloy). Both alloys 600 and 601 were
not signicantly affected. Nevertheless, adding
1600
1575
Alloy AC66
Decrease, m
500
400
Alloy 800H
10CrMo 9 10*
300
Alloy 45TM
200
Alloy 690
100
0
300
350
400
450
500
550
600
650
700
750
800
Temperature, C
(a)
225
Alloy AC66
200
10CrMo 9 10*
175
Depth, m
150
125
Alloy 800H
100
75
50
Alloy 45TM
Alloy 690
25
0
300
(b)
350
400
450
500
550
600
650
700
750
800
Temperature, C
Fig. 6.32 Corrosion behavior of 2.25Cr-1Mo steel (10CrMo9 10), alloy 800H, alloy AC66, alloy 45TM and alloy 690 tested for 300 h
at temperatures from 300 to 800 C (572 to 1472 F) in air-2Cl2; (a) decrease in thicknesses as a function of temperature, and
(b) depth of internal corrosion attack as a function of temperature. Source: Ref 42
pg 170
26/10/2007 12:28PM Plate # 0
10O2 to Ar-5HCl signicantly increased corrosion attack for alloys C-276, 600, and 601. For
Ar-5HCl, adding 0.5% H2O to the environment
caused increased corrosion for all four alloys,
particularly severe for alloys 600 and 601. The
test results obtained by Devisme et al. (Ref 50)
showed that overall, alloy C-276 (Ni-Cr-Mo)
performed better than alumina former alloy 214
in reducing environments, such Ar-HCl and ArHCl-H2 (Table 6.15). Addition of 10% CO2 was
found have less effect on corrosion attack for
four alloys in Ar-20HCl environment at 600 and
700 C, as shown in Table 6.16.
Ganesan et al. (Ref 51) investigated the
corrosion behavior of nickel-base alloys (625,
825, and 600) and iron-base alloys (800HT,
316SS, and 347SS) in a combustion environment
consisting of N2, 4 and 9O2, 12CO2, 500 ppm
SO2 with two levels of HCl (1% HCl and 4%
HCl). For both levels of HCl, three nickelbase alloys were signicantly more resistant to
chloridation attack than iron-base alloys. For
three nickel-base alloys, alloys 625 and 600
Fig. 6.33
Scanning electron backscattered image (a) and

an x-ray map for Cl (b), showing a typical pit on
alloy 800H tested for 300 h at 300 C (572 F) in air-2Cl2. Source:
Ref 42
were better than alloy 825. Their test results are

shown in Fig. 6.45 to 6.47.
Smith and Ganesan (Ref 52) conducted further
extensive studies on the corrosion behavior of
iron-base alloys (Type 316SS, Type 347SS, and
alloy 800HT) and nickel-base alloys (alloys 825,
600, and 625) in simulated combustion environments consisting of N2, O2, SO2 and various
amounts of HCl at 426, 593, and 704 C (800,
1100, and 1300 F). Also included in their
studies was the effect of H2O in the environment
on the alloys corrosion behavior. Table 6.17
summarizes the test results generated from tests
conducted in N2-10O2-50 ppm SO2-500 ppm
HCl at 426, 482, and 593 C (800, 900, and
1100 F) for 1008 h. HCl is known to be more
corrosive than SO2 in high-temperature corrosion. Thus, in this environment, corrosion attack
is primarily from HCl. The test results show that
the environment that contained about 500 ppm
HCl was not corrosive at all for Type 316SS,
Type 347SS, 800HT, 825, 600, and 625 at
temperatures up to 593 C (1100 F). When the
HCl content was increased to 4%, both Types
316 and 347 showed much higher corrosion rates
at 593 C (1100 F), while alloys 800HT, 825,
600, and 625 showed little corrosion attack at
593 C (1100 F) after 1008 h, as shown in
Table 6.18. However, when the temperature was
increased to 704 C (1300 F), only high-nickel
alloys, such as alloys 600 and 625 exhibited
good corrosion resistance in the environment
containing 4% HCl (Table 6.19). The environment containing 10% HCl became very corrosive
to nickel-base alloys 825, 600, and 625 even at
593 C (1100 F) (Table 6.18).
At high temperatures, nickel- and cobalt-base
alloys, while exhibiting low metal loss, suffered
more internal attack in O2-HCl environments.
Elliott et al. (Ref 53) examined the corrosion
attack in terms of the metal loss and internal
penetration for nickel-base alloys (alloys 214,
600, and 601) and cobalt-base alloys (alloys 25
and 188) along with Fe-Ni-Co-Cr alloy (alloy
556), Fe-Ni-Cr alloy (alloy 800H), and Type
310SS after testing in Ar-5.5O2-1HCl-1SO2 at
900 C (1650 F) for 800 h with thermal cycling
to 200 C (390 F) every 100 h. Although the
environment contained SO2, hydrogen chloride
(HCl), which is known to be more corrosive
than SO2, was the primary corrodent causing
the corrosion attack. All alloys, while exhibiting
low metal losses except Type 310SS, suffered
signicant internal penetration attack. These test
results are shown in Fig. 6.48.
pg 171
26/10/2007 12:28PM Plate # 0
6.3.4 Corrosion in HCl and HCl-Bearing

Reducing Environments
Brown et al. (Ref 22) reported corrosion rates
of various commercial alloys in a dry HCl
environment (Table 6.20). Test duration varied
from 2 to 20 h. Thus, extrapolation to a year

would yield an unreliable corrosion rate. Hossain
et al. (Ref 54) performed long-term tests in HCl
on several nickel alloys and one stainless steel
(Type 310SS). Their results are summarized in
Table 6.21 and Fig. 6.49. Type 310SS was found
(a)
(d)
(b)
(e)
(c)
(f)
Fig. 6.34
(a) Scanning electron backscattered image of the corrosion products formed on alloy AC66 tested at 800 C for 300 h in air2Cl2 and the x-ray maps showing elemental distribution for (b) chlorine, (c) chromium, (d) oxygen, (e) iron, and (f) nickel.
Source: Ref 42
pg 172
26/10/2007 12:28PM Plate # 0
to be the worst among the alloys tested. All nickel

alloys (Ni-201, 601, 625, C-4, B-2, and 600)
were much better than Type 310SS. In this
HCl test environment containing no O2, the
molybdenum-containing nickel-base alloys,
such as alloys 625 (Ni-22Cr-9Mo-3.5Nb) and
C-4 (Ni-16Cr-15.5Mo), were found to be the
best performers. In a study by Hossain et al.

(Ref 54), nickel performed reasonably well in
HCl until the temperature reached 700 C
(1290 F). At 700 C, nickel was inferior to
many nickel-base alloys, such as alloys 600,
625, and C-4 (Table 6.21). Alloy 400 was
found to be very susceptible to chloridation
(a)
(d)
(b)
(e)
(c)
(f)
Fig. 6.35
(a) Scanning electron backscattered image of the corrosion products formed on alloy 690 tested at 800 C for 300 h in air2Cl2 and the x-ray maps showing elemental distribution for (b) chlorine, (c) chromium, (d) iron, (e) oxygen, and (f ) aluminum.
Source: Ref 42
pg 173
26/10/2007 12:28PM Plate # 0
attack in HCl (Ref 54). Alloy 400 specimens

(6 mm diam 12 mm length) were completely
destroyed after exposure for 100 h at 400 C
(750 F) in HCl. Alloy 600 is commonly used
for high-temperature service in HCl environments. Figure 6.50 shows the corrosion rates
of alloy 600 in HCl gas as a function of
temperature (Ref 22). MTI Publication MS-3
(Ref 27) suggests corrosion guidelines for
Ni200, alloy 600, alloy 400, Type 304, and steel
in HCl environments, as shown in Fig. 6.51.
In reducing environments, such as Ar-4H24HCl, investigated by Baranow et al. (Ref 35),
Ni-Cr-Mo alloys, such as alloys C-276 and S,
were signicantly better than alloys 600, 625,

188, and X. Their test results generated from the
8 h tests at 900 C (1650 F) in Ar-4H2-4HCl
are shown in Fig. 6.52. In general, alloys suffered
very little metal losses, but suffered signicant
internal penetration attack. Figure 6.53 shows
the cross sections of the specimens in various
iron- and nickel-base alloys after testing in Ar4H2-4HCl at 900 C (1650 F) for 8 h (Ref 55).
A study was performed by Brill et al. (Ref 56)
investigating the chlorination resistance of nickel
and nickel-base alloys in H2-10HCl. Pure nickel,
Ni-Mo, and Ni-Cr-Mo alloys containing little
or no iron were found to be more resistant
Alloy C-2000
500
Decrease, m
400
Alloy HR160
300
40
30
20
Alloy 59
10
0
300
350
400
450
500
550
600
650
700
750
800
Temperature, C
(a)
22
20
Alloy C-2000
Alloy 59
18
16
Depth, m
14
Alloy HR160
12
10
8
6
4
2
0
300
(b)
Fig. 6.36
350
400
450
500
550
600
650
700
750
800
Temperature, C
Corrosion behavior of alloys 59, C-2000, and HR160; tested for 300 h at temperatures from 300 to 800 C (572 to 1472 F)
in air-2Cl2; (a) decrease in thicknesses as a function of temperature, and (b) depth of internal corrosion attack as a function of
pg 174
Fig. 6.37
26/10/2007 12:28PM Plate # 0
(a) Scanning electron backscattered image of the corrosion products formed on alloy 59 tested at 650 C for 300 h in air-2Cl2
and the x-ray maps showing elemental distribution for (b) chromium, (c) chlorine, (d) molybdenum, (e) iron, (f ) nickel, and
(g) oxygen. Source: Ref 43
pg 175
26/10/2007 12:28PM Plate # 0
than some nickel-base alloys with higher iron

contents. This is illustrated in Fig. 6.54, showing alloy 205 (pure nickel), B-2 (Ni-28Mo), C-4
(Ni-16Cr-16Mo), and 59 (Ni-23Cr-16Mo) being
much more resistant than alloys 625 (Ni-22Cr9Mo-3Fe) and 600H (Ni-16Cr-9Fe) (Ref 56).
Fig. 6.38
In another study by Devisme et al. (Ref 50)

on chlorination of Ni-Cr-Mo alloy C-276, NiCr-Fe alloy 600, Ni-Cr-Fe-Al alloy 601 (1.4Al),
and Ni-Cr-Al-Fe alloy 214 in Ar-HCl environments. Their test results conducted at 600 C
in Ar-5HCl, Ar-10HCl, and Ar-20HCl are
(a) Scanning electron backscattered image of the corrosion products formed on alloy HR160 tested at 800 C for 300 h
in air-2Cl2, and the x-ray maps showing elemental distribution for (b) chromium, (c) chlorine, (d) silicon, (e) oxygen, and
(f ) nickel. Source: Ref 43
pg 176
26/10/2007 12:28PM Plate # 0
shown in Fig. 6.55. Barnes (Ref 57) investigated

nickel-base alloys in Ar-HCl mixtures with
concentrations of HCl varying from 13 to 100%
and found that Ni201 (pure nickel) was slightly
more resistant than alloy 600 (8% Fe) for all the
concentrations tested except 100% HCl, as
shown in Table 6.22. Other nickel-base alloys
were also included in his test program, with the
test results summarized in Table 6.23. Except for
pure nickel (Ni201), nickel-base alloys suffered
internal attack by formation of internal voids.
This surface zone with extensive internal voids
was found to be highly depleted in chromium in
the matrix near voids. For example, alloy 600
showed about 1.7% Cr and 4.7% Fe in the
matrix in the porous zone after testing at 735 C
(1355 F) for 100 h in Ar-33HCl (Ref 57).
Barness test environments were strictly HCl
(inlet gas), thus making pCl2 in the environment
signicantly higher than that if the initial inlet
gas contain both H2 and HCl. This is shown in
Fig. 6.56, where pCl2 was plotted as a function
of temperature for 100% HCl, Ar-33HCl, and
H2-30HCl along with several metal chlorides
(Ref 57). The gure shows that there is no signicant difference in pCl2 between the 100HCl
and Ar-33HCl environments. However, with the
Decrease, m
400
350
300
250
200
150
Fe3Al
50
TiAl
40
30
20
10
Alloy 214
0
300 350 400 450 500 550 600 650 700 750 800
Temperature, C
(a)
105
100
Depth, m
95
30
25
20
15
10
5
0
Fe3Al
Alloy 214
TiAl
300 350 400 450 500 550 600 650 700 750 800
(b)
Temperature, C
Fig. 6.39
Corrosion behavior of alumina former alloy 214

and intermetallics Fe3Al and TiAl; tested for 300 h
at temperatures from 300 to 800 C (572 to 1472 F) in air-2Cl2;
(a) decrease in thicknesses as a function of temperature, and
(b) depth of internal corrosion attack as a function of temperature.
Source: Ref 44
Fig. 6.40
Metal loss as a function of time for alloys Fe-20Cr

and 800H tested at 927 C (1700 F) in (a) Ar20O2-2500 ppm (0.25%) Cl2 and (b) Ar-20CO2-2500 ppm Cl2.
Source: Ref 45
pg 177
26/10/2007 12:28PM Plate # 0
Fig. 6.41
Depth of internal penetration as a function of time

(h0.5) for alloy 800H tested at 927 C (1700 F) in
(a) argon 20% O2, 0.25% Cl2 and (b) argon 20% CO2, 0.25% Cl2.
Source: Ref 45
introduction of H2 in the H2-HCl gas mixture,

pCl2 becomes signicantly lower, as shown in
H2-30HCl. Even though chlorine partial pressure
(pCl2 ) stays relatively unchanged (Fig. 6.56)
thermodynamically, the kinetic of the HCl-metal
reaction (corrosion rate) in Ar-HCl mixtures
was found to increase with increasing HCl concentration (HCl partial pressure). This is shown
in Fig. 6.57 for both Ni201 and alloy 600 tested
at 735 C (1355 F). It should also be noted
that Barness test environments involving
mainly HCl were extremely corrosive for nickel
alloys tested. In addition to high pCl2 values, the
environments were so reducing that no oxides
were thermodynamically stable. Thus, in those
test environments, the gas-metal reaction mainly
involved formation of chlorides (highly volatile
corrosion products) and no oxides (nonvolatile
solid phases). Figure 6.58 shows nickel chlorides
that formed on a Ni201 coupon after testing at
735 C (1355 F) in Ar-33HCl (Ref 57).
In a simulated waste incineration environment
consisting of N2, 12% CO2, 500 ppm SO2, and
Fig. 6.42
Effect of oxygen in O2-HCl mixtures on the corrosion rate of iron at 300700 C (5701290 F).
Source: Ref 47
Fig. 6.43
Effect of oxygen in O2-HCl mixtures on the corrosion rate of nickel at 400700 C (7501290 F).
Source: Ref 48
pg 178
26/10/2007 12:28PM Plate # 0
pg 179
1% HCl, Ganesan et al. (Ref 51) observed that

nickel-base alloys, particularly alloy 625, performed signicantly better than iron-base alloys,
such as alloy 800HT, Type 316SS, and Type
347SS, as shown in Fig. 6.59 and 6.60. Although
not discussed in the paper, it is believed that
the oxygen potential (pO2 ) was high enough to
form some oxides, such as Cr2O3. Thus, little
corrosion attack was detected for nickel-base
alloys even when the test temperature was increased to 704 C (1300 F) (Ref 51). Formation
of chromium oxides can signicantly reduce
the corrosion attack in reducing, HCl-bearing
800HT
316SS
347SS
Fig. 6.44
Effect of oxygen in O2-HCl mixtures on

the corrosion rate of chromium at 400800 C
(7501470 F). Source: Ref 49
Table 6.15 Corrosion of nickel-base alloys in

terms of metal loss m (mils) after 500 h at 600 C
(1112 F) in the indicated test environments
Fig. 6.45
Weight change as a function of exposure time

for nickel-base alloys (alloys 625, 600, and 825)
and iron-base alloys (alloy 800HT, 316SS, and 347SS) in N24O2-12CO2-1HCl-500 ppm SO2. Testing was initially performed
at 649 C, then increased to 704 C, and finally to 760 C as
indicated. Source: Ref 51

Environment
Ar-20HCl
Ar-20HCl-2O2
Ar-5HCl
Ar-5HCl-10O2
Ar-5HCl-0.5H2O
Ar-5HCl-3H2
C-276
600
601
214
60 (2.4)
330 (13.0)
35 (1.4)
120 (4.7)
90 (3.5)
5 (0.2)
150 (5.9)
185 (7.3)
50 (2.0)
140 (5.5)
240 (9.5)
15 (0.6)
150 (5.9)
120 (4.7)
90 (3.5)
160 (6.3)
255 (10.0)
15 (0.6)
260 (10.2)
65 (2.6)
30 (1.2)
55 (2.2)
80 (3.2)
20 (0.8)
800HT
Source: Ref 50
347SS
Table 6.16 Metal loss after 500 h at 600 and

700 C in the indicated test environments
316SS
Metal loss, m
Environment
Ar-20HCl
Ar-20HCl-10CO2
Ar-20HCl
Ar-20HCl-10CO2
Source: Ref 50
Temperature, C
C-276
600
601
214
600
600
700
700
60
110
200
280
150
100
280
195
150
120
225
250
260
230
360
190
Fig. 6.46

and iron-base alloys (alloy 800HT, 316SS, and 347SS) in N29O2-12CO2-1HCl-500 ppm SO2. Testing was initially performed
at 649 C, then increased to 704 C, and finally to 760 C as
indicated. Source: Ref 51
26/10/2007 12:29PM Plate # 0
pg 180
800HT
316SS
347SS
704
Fig. 6.48
Fig. 6.47

and iron-base alloys (alloys 800HT, 316SS, and 347SS) in N29O2-12CO2-4HCl-100 ppm SO2. Testing was initially performed
at 593 C, then increased to 704 C, and to 816 C, and finally
to 927 C as indicated. Source: Ref 51
Table 6.17 Corrosion rates in terms of metal loss

for iron- and nickel-base alloys at 426, 482 and
593, 900 C (800, 900, and 1300 F) for 1008 h
in N2-10O2-50 ppm SO2-500 ppm HCl
The metal loss and internal penetration for nickelbase alloys (alloys 214, 600, and 601) and cobaltbase alloys (alloys 25 and 188) along with Fe-Ni-Co-Cr alloy
(alloy 556), Fe-Ni-Cr alloy (alloy 800H), and Type 310SS tested in
Ar-5.5O2-1HCl-1SO2 at 900 C (1650 F) for 800 h with thermal
cycling to 200 C (390 F) every 100 h. Source: Ref 53
Table 6.19 Corrosion rates in terms of metal

loss for iron- and nickel-base alloys at 593
and 704 C (1100 and 1300 F) in N2-10O250 ppm SO2-4HCl
Metal loss
593 C(a)
Metal loss
426 C
482 C
593 C
Alloy
m/yr
mpy
m/yr
mpy
m/yr
mpy
Type 316
Type 347
Alloy 800HT
Alloy 825
Alloy 600
Alloy 625
0.07
0.09
0.07
0.02
0.04
0.02
0.003
0.004
0.003
0.0008
0.002
0.0008
2.54
2.29
1.02
1.27
2.03
1.78
0.1
0.09
0.04
0.05
0.08
0.07
5.08
7.11
5.33
2.54
3.05
2.79
0.2
0.28
0.21
0.1
0.12
0.11
704 C(b)
Alloy
m/yr
mpy
m/yr
mpy
Type 316
Type 347
Alloy 800HT
Alloy 825
Alloy 600
Alloy 625
914
1245
74
20
25
16
36
49
2.9
0.8
1.0
0.6
3,810
12,039
2,083
41
152
149
470
81
1.6
5.9
(a) Test duration: 72 h. (b) Test duration: 192 h. 1.0 m = 0.001 mm = 0.0394 mil.
Source: Ref 52
1.0 m = 0.001 mm = 0.0394 mil. Source: Ref 52
Table 6.20 Corrosion of alloys in dry HCl(a)

Approximate temperature, C (F),
at which given corrosion rate is exceeded
Table 6.18 Corrosion rates in terms of

metal loss for iron- and nickel-base alloys at
593 C (1100 F) for 72 h in N2-9O2-12CO2100 ppm SO2-4 and 10HCl
Metal loss
4% HCl
10% HCl
Alloy
m/yr
mpy
m/yr
mpy
Type 316
Type 347
Alloy 800HT
Alloy 825
Alloy 600
Alloy 625
914
1245
74
20
25
16
36
49
2.9
0.8
1.0
0.6
1066
1219
1549
42
48
60
Alloy
0.8 mm/yr
(30 mpy)
1.5 mm/yr
(60 mpy)
3.0 mm/yr
(120 mpy)
15 mm/yr
(600 mpy)
Nickel
600
B
C
D
l8-8Mo
25-12Cb
18-8
Carbon steel
Ni-resist
400
Cast iron
Copper
455 (850)
425 (800)
370 (700)
370 (700)
288 (550)
370 (700)
345 (650)
345 (650)
260 (500)
260 (500)
230 (450)
205 (400)
93 (200)
510 (950)
480 (900)
425 (800)
425 (800)
370 (700)
370 (700)
400 (750)
400 (750)
315 (600)
315 (600)
260 (500)
260 (500)
148 (300)
565 (1050)
538 (1000)
480 (900)
480 (900)
455 (850)
480 (900)
455 (850)
455 (850)
400 (750)
370 (700)
345 (650)
315 (600)
205 (400)
675 (1250)
675 (1250)
650 (1200)
620 (1150)
650 (1200)
593 (1100)
565 (1050)
593 (1100)
565 (1050)
538 (1000)
480 (900)
455 (850)
315 (600)
(a) Based on short-term laboratory tests. Source: Ref 22
environments. This is illustrated by the test

results generated by Strafford et al. (Ref 58).
The authors tested Fe-Cr alloys containing 2, 5,
9, 14, and 25% Cr at 1000 C (1832 F) in a
H2-H2O-HCl mixture giving pO2 value of

1.5 1016 atm and pCl2 value of 108 atm at the
test temperature. Based on the M-O-Cl stability
26/10/2007 12:29PM Plate # 0
pg 181
diagram at the test temperature, the environment

was in the location where Cr2O3, but not FeO,
was to form, and FeCl2 was to form. Test results
show that Fe-25Cr alloy exhibited no weight loss
for the exposure time up to 100 h. This suggests
that once the alloy contains sufcient chromium
to form a continuous Cr2O3 scale, as in the case
of Fe-25Cr, corrosion can be signicantly
reduced or prevented. The results also indicate

that Fe-Cr alloys with low chromium contents,
such as 2%, 5%, and 9%, suffered signicant
corrosion attack (Fig. 6.61).
In an investigation into possible candidate
alloys for a process developed by the Bureau
of Mines for extracting alumina from Kaolinitic clay, Carter et al. (Ref 59) tested various
Table 6.21 Corrosion of selected alloys in HCl at 400, 500, 600, and 700 C (750, 930, 1110,
and 1290 F)
Metal loss mg/cm2
400 C (750 F)
500 C (930 F)
600 C (1110 F)
700 C (1290 F)
Alloy
300 h
1000 h
100 h
300 h
1000 h
100 h
300 h
96 h
Ni-201
601
310
625
C-4
B-2
600
1.19
1.58
3.26
0.74
0.55
0.75
0.93
0.91
1.47
5.16
1.1
1.12
0.76
0.81
1.60
2.57
6.74
2.42
2.09
2.10
1.69
2.89
4.14
13.65
3.78
3.36
2.65
3.31
4.86
9.38
46.60
8.64
7.24
5.87
7.81
11.46
9.01
15.65
6.79
7.31
12.93
7.67
37.7
19.46
32.6
14.6
19.14
62.3
17.3
377
102.5
1025
26.5
34.9
126.4
49.6
Source: Ref 54
Fig. 6.49
Corrosion rates of several iron- and nickel-base alloys in HCl at 400 to 700 C (750 to 1290 F). Source: Ref 54
26/10/2007 12:29PM Plate # 0
alloy. Long-term tests (1584 h) at 260 to 380 C

(500 to 720 F) showed corrosion rates of 0.3
to 8 m/yr for alloy 625 and 4 to 10 m/yr for
alloy R-41. The results were in agreement with
those obtained from short-term tests. Titanium
and Ti-0.2Pd were also tested in the same
environment for 15 days. The corrosion rates
for titanium were found to be 246, 28, and 13
m/yr at 400, 300, and 200 C, respectively.
Ti-0.2Pd was found to corrode at 1900, 108, 0,
and 2 m/yr at 500, 400, 300, and 200 C,
respectively.
In a study by Reeve (Ref 60) involving 80%
HCl and 20% H2O, corrosion rates were obtained
for steel and stainless steels (Fig. 6.62). Mild
steel was found to suffer corrosion rates of
less than 0.76 mm/yr (30 mpy) up to 400 C
(752 F), and the 18-8 stainless steel, less than
1.0 mm/yr (40 mpy) up to 500 C (930 F).
These values were signicantly lower than those
predicted by Brown et al. (Ref 22) in dry HCl
(Table 6.20). It appears that addition of signicant amount of H2O into HCl environment
makes HCl less corrosive.
month
commercial alloys in an environment of 40%

HCl and 60% H2O at various temperatures up to
500 C (930 F). The results of iron- and nickelbase alloys are summarized in Tables 6.24
and 6.25. Corrosion data at temperatures below
200 C (390 F) are not included because of dew
point corrosion. Stainless steels and nickel alloys
tested showed low corrosion rates at all test
temperatures in this environment. A cobalt-base
alloy (alloy 188) and an Fe-Ni-Co-Cr alloy
(Multimet alloy) were also tested in the temperature range of 315 to 375 C (600 to 710 F),
showing no measurable attack for alloy 188 and
a corrosion rate of only 0 to 3 m/yr for Multimet
6.4 Corrosion in F2- and HF-Bearing

Environments
6.4.1 Corrosion in F2 Environments
Fig. 6.50
Early studies of corrosion in uorine gas for

various metals and alloys were carried out by
Corrosion rate of alloy 600 in HCl as a function of

Corrosion rate, mm/yr

0.025
700
0.05
0.10
0.25
Tubes / Internals
0.50
1.00
2.5
1290
Vessels / Pipes
1110
600
Nickel 200 and Alloy 600
400
1020
930
Type 304
300
750
Alloy 400
572
Carbon steel
200
Temperature, F
Temperature, C
500
100
390
0
1
10
20
40
60
100
Corrosion rate, mpy
Fig. 6.51
MTI corrosion guidelines for Ni200, alloy 600, alloy 400, Type 304SS and steel in HCl as a function of temperature. Source:
Ref 27. Courtesy of Materials Technology Institute
pg 182
26/10/2007 12:29PM Plate # 0
Metal loss
214
Continuous
internal penetration
C-276
601
R-41
600
625
188
X
0
40
80
120 160 200 240 280 320
360
Corrosion rate, mils/month
Fig. 6.52
Corrosion rates in terms of metal loss and internal

penetration for nickel- and cobalt-base alloys at
900 C (1650 F) in Ar-4HCl-4H2. Data was based on 8 h tests.
Source: Ref 35
Myers and DeLong (Ref 61). Their test duration

was short, normally 4 h. The longest test run
was 15 h. The corrosion rates they obtained are
summarized in Table 6.26. These rates, extrapolated from short-term tests, are not recommended for estimating the service life of process
equipment. Rather, the data are useful for
making performance comparisons among various alloys.
The results obtained by Myers and DeLong
(Ref 61) indicated that nickel had good corrosion
resistance in uorine gas at temperatures up to
500 C (930 F). Even at temperatures higher
than 500 C (930 F), corrosion rates for nickel
were signicantly lower than those of other
alloys. Nickel is commonly used for plant equipment handling uorine at temperatures up to
500 C (930 F) (Ref 62). The resistance of
nickel to uorine gas was attributed to the formation of an adherent nickel uoride scale
(Ref 62). The reaction of nickel and uorine
was found to follow a parabolic rate law (Ref 63).
Hauffe (Ref 64) summarized the results of

corrosion tests by Myers and DeLong (Ref 61),
Jarry et al. (Ref 62), and Lukyanchev et al.
(Ref 65) in Table 6.27. The tests were very short
in duration, from 30 min to 32 h. The corrosion
rates obtained by Myers and DeLong were two
orders of magnitude higher than those observed
by Jarry et al. and Lukyanchev et al. Based on
the corrosion rates observed by Jarry et al. and
Lukyanchev et al., nickel exhibited good resistance in uorine at temperatures up to 810 C
(1490 F) (about 12 mpy).
Hale et al. (Ref 66) tested six different types
of nickel with various degrees of purity at 590 C
(1100 F) for 95 h in uorine and found no
measurable corrosion for all the samples, except
one with a slightly higher silicon content (about
8 mils of general corrosion). The results by
Hale et al. (Ref 66) are shown in Table 6.28.
They also found that the materials purity
played an important role when tested at 700 C
(1290 F) for 210 h (Ref 66). Both high-purity
vacuum-melted nickel and electrolytic nickel
showed negligible attack. Low-carbon nickel
I (about 0.015% Si), carbonyl nickel, and A
nickel (commercial grade pure nickel) began to
show intergranular void formation at 700 C
(1290 F). At the same temperature, low-carbon
nickel II (about 0.05% Si) suffered signicant
general corrosion and intergranular attack (void
formation). Silicon appears to be a detrimental
impurity in pure nickel in resisting corrosion
attack by uorine. Results obtained by Steindler
and Vogel (Ref 67) showed that A nickel
(commercial pure nickel) suffered signicant
corrosion at 750 C (1380 F), as shown in
Table 6.29. A general review on the kinetic
aspects of nickel-uorine reactions can be found
in Ref 8.
Additions of alloying elements to nickel
generally are detrimental to uorine corrosion
resistance. Many nickel-base alloys were found
to be signicantly more susceptible than
nickel to uorine corrosion (Ref 61, 66, 67).
Alloy 600 (Ni-15.5Cr-8Fe), Inco 61 weld wire
(Ni-1.5Al-3Ti), Ni-O-Nel (Ni-21Cr-31Fe-3
Mo-1.75Cu), INOR-1 (Ni-20Mo), INOR-2 (Ni5Cr-16Mo), INOR-3 (Ni-16Mo-1Al-1.6Ti),
INOR-4 (Ni-17Mo-1.7Fe-2Al-1.7Ti), INOR-5
(Ni-13Mo-2.7W-2.2Cb+Ta-1Mn), Hastelloy W
(Ni-25Mo-5.5Fe-2.5Co), Hastelloy B (Ni25Mo-6Fe), HyMa 80 (Ni-16Fe-4Mo), and
Monel (Ni-30Cu) suffered signicantly more
corrosion than nickel in uorine at 590 C
(1100 F), as shown in Table 6.28 (Ref 66).
pg 183
26/10/2007 12:29PM Plate # 0
pg 184
Fig. 6.53
Internal penetration in terms of voids for various iron- and nickel-base alloys after testing in 900 C (1650 F) for 8 h in
Ar-4H2-4HCl. Source: Ref 55
Metal loss, m
50
100
150
200
Fig. 6.54
Corrosion of nickel-base alloys in H2-10HCl

at 850 C with 24 h cycles. Source: Ref 56
250
5% HCl
10% HCl
20% HCl
300
C276
Dura-Nickel (Ni-3Ti-1.5Al), 70Cu-30Ni, and

Ni-10Co suffered slightly more corrosion than
nickel at 590 C (1095 F) (Ref 66). Fluorides of
molybdenum, tungsten, titanium, and other elements have low melting points and/or high vapor
pressures (Table 6.2).
Iron is signicantly less resistant to uorine
attack than nickel. Myers and DeLong (Ref 61)
observed that Armco Iron is resistant to uorine
600
601
214
Alloy
Fig. 6.55
Corrosion in terms of loss of sound metal (m) of

alloys C-276, 600, 601, and 214 in Ar-5HCl, Ar10HCl, and Ar-20HCl at 600 C for 500 h. Source: Ref 50
at temperatures up to 250 C (480 F). Steels

were found to be less resistant than Armco Iron
(Ref 61). The concentration of silicon appears
26/10/2007 12:29PM Plate # 0
Table 6.22 Metal loss rate(a) in Ar-HCl

and 100HCl environments at 735 C (1355 F)(b)
Metal loss
Alloy
Ni201
Alloy 600
Ni201
Alloy 600
Ni201
Alloy 600
Ni201
Alloy 600
Environment
mm/yr
mpy
Ar-13HCl
Ar-13HCl
Ar-33HCl
Ar-33HCl
Ar-52HCl
Ar-52HCl
100HCl
100HCl
2.8
3.5
5.3
7.9
9.6
11.4
18.4
14.2
110
138
209
311
378
449
724
559
(a) Metal loss rate did not include internal void penetration. Internal void penetration was a small portion of the total metal loss. (b) Test durations from 15100 h.
Source: Ref 57
to be important in affecting the steels uoridation resistance, as shown in Table 6.26

(Ref 61). SiF4 has a very low melting point
and high vapor pressure (Table 6.2). Ferritic
and austenitic stainless steels, except Type
347SS, showed negligible corrosion at 200 and
250 C (390 and 480 F) (Table 6.26). At
higher temperatures, corrosion of these alloys
became signicant. Jackson (Ref 68) also
reported signicant corrosion rates for several
austenitic stainless steels at 370 C (700 F), as
Table 6.23 Estimated metal loss rates for

pure nickel and nickel-base alloys tested at 685,
735, and 785 C (1265, 1355, and 1445 F)
in 100HCl
Metal loss
Alloy
Ni201
Alloy 600
Alloy HR160
Alloy 214
Alloy 602CA
Alloy HR160
Ni201
Alloy 600
Alloy HR160
Source: Ref 57
Temperature, C (F)
mm/yr
mpy
685 (1265)
685 (1265)
685 (1265)
686 (1265)
686 (1265)
735 (1355)
785 (1445)
785 (1445)
785 (1445)
4.2
7.5
7.0
5.7
28.6
45.6
16.7
26.4
96.0
165
295
276
224
1130
1800
657
1040
3780
Fig. 6.57
Metal loss rate (mm/yr) as a function of HCl concentrations (pHCl) in Ar-HCl mixtures at 735 C
H2-30% HCI
Fig. 6.56
Thermodynamic equilibrium chlorine partial pressure (pCl2 ) as a function of temperature for several environments (100HCl,
Ar-33HCl, and H2-30HCl) and several chlorides. Source: Ref 57
pg 185
26/10/2007 12:29PM Plate # 0
pg 186
Fig. 6.58
Nickel chlorides formed on Ni201 after testing at 735 C (1355 F) for 15 h in Ar-33HCl. Source: Ref 57
347SS
316SS
316SS
347SS
Fig. 6.59
Mass change as a function of time for nickeland iron-base alloys tested initially at 593 C
(1100 F), then increased to 649 C (1200 F), and finally to
704 C (1300 F) in N2-12%CO2-500ppm SO2-1%HCl. Source:
Ref 51
Mass change as a function of time for nickel- and

iron-base alloys tested at 593 C (1100 F) in N212%CO2-500ppmSO2-1%HCl. Source: Ref 51
shown in Table 6.30. Cobalt and cobalt-base

alloys are not as resistant to uorine as nickel
(Ref 66) (Table 6.28). Limited data for other
metals, such as copper, aluminum, and magnesium, are shown in Tables 6.26, 6.29, and
6.31. Aluminum was resistant to uorine at
temperatures up to approximately 500 C
(930 F) (Table 6.26 and 6.31). AlF3 has a
relatively high melting point (1197 C or
2187 F) (Ref 64).
Corrosion of nickel, alloy 400, and alloy 600
by various volatile metallic uorides was investigated by Vogel et al. (Ref 69) in their studies on
volatile ssion product uorides, which were

associated with the development of a process to
recover uranium and plutonium from partially
spent nuclear reactor fuels. Most of these ssion
product uorides were much less corrosive than
uorine gas. However, whenever uorine gas
was present along with the uoride, the corrosion
rate was generally more aggressive. Nickel and
alloy 400 exhibited relatively low corrosion rates
for all the volatile uorides at 500 C (930 F),
except TeF6, as shown in Table 6.32. Alloy 600,
on the other hand, suffered extremely high corrosion rates.
Fig. 6.60
26/10/2007 12:29PM Plate # 0
6.4.2 Corrosion in HF Environments

Hydrogen uoride is generally less corrosive
than uorine for most metals and alloys. However, it is still very corrosive at elevated temperatures. Corrosion rates of various metals and
Fig. 6.61
Weight change as a function of exposure time at

1000 C (1832 F) in a H2-H2O-HCl mixture giving
pO2 value of 1.5 1016 atm and pCl2 value of 108 atm for Fe-Cr
alloys containing various amounts of chromium. Source: Ref 58
Table 6.24 Corrosion of iron-base alloys in

40HCl-60H2O
Corrosion rate(a) m/yr
Temperature,
C (F)
500 (930)
400 (750)
300 (570)
210 (410)
200 (390)
316LSS
29-4SS
430SS
E-Brite
26-1
1020
steel
4130
steel
483
15
3
10
3
363
5
2
2100
326
700
46
1700
406
870
51
(a) Linearly extrapolated from 15 d (360 h) laboratory tests. Note: mpy = (m)
0.0394. Source: Ref 59

40HCl-60H2O
alloys in HF based on short-term tests were

reported by Myers and DeLong (Ref 61), as
shown in Table 6.33. Nickel was the most resistant among the alloys tested. Copper, alloy 400,
and alloy 600 were slightly worse than nickel.
Carbon steels and stainless steels showed poor
resistance. Tyreman and Elliott (Ref 70) found
that nickel, cobalt, copper, and molybdenum
were more resistant to HF corrosion than iron,
chromium, niobium (columbium), and tantalum
(Fig. 6.63). Chromium was found to be detrimental to resistance to corrosion by HF for Ni-Cr
alloys (Fig. 6.64) (Ref 71). Chromium uorides
were the major corrosion products for Ni-Cr
alloys tested in HF (Ref 71).
Marsh (Ref 72) did an extensive investigation
for his Ph.D. thesis on the resistance of several
nickel-base alloys to corrosion attack in HF
environments at elevated temperatures. In NiCr alloys tested in HF at 650 C (1200 F),
chromium uorides were found to form on the
metal surface as well as internally as internal
phases. This is illustrated in Fig. 6.65, showing
elemental distribution for Ni-40Cr alloy tested at
650 C (1200 F) in anhydrous HF environment
(Ref 72). Surface corrosion products were found
to be enriched in chromium and uorine, but
depleted in nickel, and similar elemental
distribution was observed for the internal phases.
Several commercial nickel-base alloys were also
tested. Figure 6.66 shows corrosion attack of
alloy 600 in HF after 92 h at 650 C (1200 F).
In this case, both the corrosion products formed
on the metal surface and the internal phases
were found to consist of both CrF3 and FeF2,
as determined by x-ray diffraction analysis
(Ref 72). In Ni-Cr alloys, it appears that NiCr-Mo alloy 625 was more resistant to HF
corrosion attack than Ni-Cr alloys X750 and
Corrosion rate(a), m/yr

Temperature,
C (F)
500 (930)
380 (720)
380 (720)
380 (720)
375 (710)
375 (710)
365 (690)
350 (660)
350 (660)
315 (600)
310 (590)
290 (550)
260 (500)
Ni-200
600
625
R-41
B-2
13
15
15
25
13
20
3
18
13
30
58
46
58
48
48
46
30
33
15
13
132
13
23
10
15
13
8
15
8
0
3
3
3
5
0
3
3
5
0
0
15
18
20
13
28
33
36
28
(a) Linearly extrapolated from 15 d (360 h) laboratory tests. Note: mpy = (m)
0.0394. Source: Ref 59
Fig. 6.62
Corrosion rates of mild steel, Type 304SS, and alloy

800 in 80HCl-20H2O at 300 to 600 C (570 to
pg 187
26/10/2007 12:29PM Plate # 0
pg 188
Table 6.26
Corrosion of various metals and alloys in fluorine

Corrosion rates(a), mpy
Materials
200 C
250 C
300 C
350 C
400 C
450 C
500 C
600 C
650 C
700 C
Nickel
400
600
Copper
Aluminum
Magnesium
430SS
347SS
309CbSS
310SS
Armco Fe
Steel (0.007% Si)
1020 (0.22% Si)
1030 (trace Si)
1030 (0.18% Si)
1015 (0.07% Si)
Music wire (0.13% Si)
Nil
8.4
Nil
Nil
Nil
Nil
Nil
456
24
Nil
Nil
1,740
Nil
Nil
24
192
5,760
96
Nil
3,060
2,556
900
372
108
48
7,920
108
9,000
9,960
4,800
936
6,204
5,544
4,248
96
2.4
1,764
Nil
8
6
456
1,920
Nil
936
9,540
7,980
6,732
288
144
6,480
180
23
18
1,152
Nil
3600
18 in.
6,480
61
24
744
1,440
156
139 in.
238 in.
348
720
2,040
11,880
216
192
960
1,560
408
1,800
6,120
(a) Note: mpy 0.0254 = mm/yr. Source: Ref 61
Table 6.27 Corrosion of nickel in fluorine

at various temperatures
Temperature,
C (F)
300 (570)
300 (570)
400 (750)
400 (750)
400 (750)
500 (930)
500 (930)
500 (930)
550 (1020)
600 (1110)
660 (1220)
720 (1330)
810 (1490)
Table 6.28 Results of corrosion tests in

fluorine at 590 C (1100 F) for 95 h
Pressure,
atm
Test
duration, h
Corrosion rate,
mm/yr (mpy)
Ref
0.9
0.9
0.9
0.9
0.99
0.9
0.9
0.99
0.99
0.9
0.99
0.99
0.99
8
32
0.5
28
0.25
30
2
32
1.5
2
1.7
0.005 (0.2)
0.00073 (0.03)
0.015 (0.59)
0.0012 (0.05)
0.21 (8.3)
0.018 (0.7)
0.003 (0.12)
1.5 (59.1)
0.036 (1.4)
0.017 (0.7)
0.150 (5.9)
0.240 (9.4)
0.310 (12.2)
62
62
62
62
61
62
62
61
65
62
65
65
65
Source: Ref 64
601 containing no signicant amounts of

molybdenum, as illustrated in Fig. 6.67 (Ref 71).
Corrosion attack in HF for alloys 601, 625, and
N is shown in Fig. 6.68 to 6.71. Alloy N (Ni-5Cr16Mo) was found to be signicantly more
resistant to HF corrosion than nickel-base alloys
600, 601, and 625. This suggests that nickel-base
alloys containing low chromium and high
molybdenum would be more resistant to HF
corrosion attack than Ni-Cr alloys.
Molybdenum showed little corrosion attack
in HF after 10 h at 850 C (1560 F), as shown
in Fig. 6.63 (Ref 70). The results of Zotikov and
Semenyuk (Ref 73) indicated that both molybdenum and tungsten were quite resistant to
corrosion attack by HF at temperatures from
300 to 800 C (up to 700 C for tungsten), as
illustrated in Table 6.34. Molybdenum appears
Alloy
Depth of corrosion
attack(a),
mm (mils)
High-purity nickel sheet

High-purity nickel rod
Low-carbon nickel (II)
Low-carbon nickel (I)
Carbonyl nickel
Electrolytic nickel
A nickel
70-30 Cupronickel
Inconel (low carbon)
Inco 61 weld wire
Duranickel
Ni-O-NEL
NM
NM
0.20 (8)
NM
NM
NM
NM
0.06 (2.5)
0.64 (25)
0.64 (25)
0.10 (4)
>0.42 (16.5)
INOR-1
>0.83 (32.5)
INOR-2
>0.83 (32.5)
INOR-3
>0.83 (32.5)
INOR-4
>0.81 (32)
INOR-5
Hastelloy alloy B
0.76 (30)
>0.48 (19)
HyMa 80
90Ni-10Co
80Ni-20Co
Monel
0.37 (14.5)
0.06 (2.5)
>0.13 (5)
>0.80 (31.5)
Hastelloy alloy W
>0.79 (31)
310SS
>0.60 (23.5)
Carpenter 20
>0.66 (26)
Haynes alloy No. 25
>0.34 (13.5)
Cobalt
>1.52 (60)
(a) NM, not measurable. Source: Ref 66
Comments
General attack
General attack
General attack
General attack
General attack
Completely converted
to uorides
to uorides
to uorides
to uorides
to uorides
General attack
to uorides
General attack
General attack
General attack
to uorides
to uorides
to uorides
to uorides
to uorides
to uorides
26/10/2007 12:29PM Plate # 0
pg 189
to improve the corrosion resistance of nickelbase alloys in HF. The results published by
International Nickel Company (Ref 74) showed
that alloy N (Ni-16Mo-5Cr) and alloy B (Ni25Mo) were slightly better than nickel in HF
(Table 6.35). Field testing in an HF-bearing
environment showed that alloy N and alloy S
(Ni-16Cr-15Mo) were better than alloys 625 and
C-22 at 900 C (1650 F) (Ref 75).
Barnes (Ref 76) investigated the corrosion
behavior of a wide range of materials in HF
environments at 1000 C (1832 F). Tests were
Table 6.29 Corrosion of several metals and
alloys in fluorine(a)
Metal
Test temperature,
C (F)
Time, h
Weight gain,
mg/cm2
Corrosion rate,
mm/y (mils/y)
Ni(b)
600
Cu
Ni(b)
400
600
Cu
Ni(b)
400
600
600
Cu
550 (1,020)
550 (1,020)
550 (1,020)
650 (1,200)
650 (1,200)
650 (1,200)
650 (1,200)
750 (1,380)
750 (1,380)
750 (1,380)
750 (1,380)
750 (1,380)
6.2
5.0
2.42
5.28
6.0
5.6
4.8
4.1
6.1
4.7
4.7
5.8
1.9
706
17.4
21.5
16.0
1,743
134
100
1, 831
4, 907
12, 220
278
0.11 (4.4)
81 (3,200)
2.8 (110)
1.5 (59)
1.0 (41)
180 (7,100)
10.9 (430)
9.0 (353)
74 (2,900)
610 (24,000)
660 (26,000)
19 (750)
(a) Tests were conducted in owing gas (30 to 130 cc/min). (b) Nickel A
(commercial grade pure nickel). Source: Ref 67
Table 6.30 Corrosion of several alloys

in fluorine(a)
Corrosion rate, mm/yr (mpy)
Alloy
Exposure
time, h
400
Ni-200
304
304L
347
Illium R
600
5
24
120
5
24
120
5
24
120
5
5
5
200 C
(400 F)
370 C
(700 F)
540 C
(1000 F)
0.013 (0.5)
0.048 (1.9)
0.76 (29.8)
0.013 (0.5)
0.043 (1.7)
0.29 (11.3)
0.003 (0.1)
0.031 (1.2)
0.18 (7.2)
0.084 (3.3)
0.043 (1.7)
0.62 (24.5)
0.013 (0.5)
0.031 (1.2)
0.41 (16.1)
0.003 (0.1)
0.010 (0.4)
0.35 (13.8)
0.155 (6.1) 40 (1565)
0.191 (7.5) 153 (6018)
0.65 (25.4)
0.102 (4.0) 108 (4248)
0.152 (6.0)
0.32 (12.7) 103 (4038)
0.015 (0.6)
2.0 (78.0)
88 (3451)
(a) Tests were conducted in owing uorine. Source: Ref 68
Table 6.31
Corrosion of aluminum in fluorine
Test temperature, C (F)
26 (79)
201 (394)
356 (673)
543 (1009)
Source: Ref 68
0.00087 (0.03)
0.0003 (0.01)
0.57 (22)
2.2 (87)
conducted in Ar-5HF, Ar-15HF, and Ar-35HF

mixtures. In nickel-base alloys, alloys B-3 (Ni28Mo-1.5Cr) and 242 (Ni-25Cr-8Mo) were
found to be signicantly better than alloys 600,
617, and 602CA. His test results are summarized
in Table 6.36 and Fig. 6.72. Both Ni-Mo alloys
containing low Cr formed thin surface scales
with very little internal uoride penetration.
Surprisingly, alloy 188, a cobalt-base alloy with
high chromium (22%) and high tungsten (14%),
also exhibited good corrosion resistance with
a thin surface scale and little internal uoride
penetration. Nickel aluminide intermetallic
(Ni-8Al-8Cr-1.4Mo-1.7Zr) was found to exhibit
good corrosion resistance in Ar-5HF. However,
when tested in Ar-35HF, the nickel-aluminide
Table 6.32 Corrosion of nickel and nickel-base
alloys by various volatile fluorides at 500 C
(930 F)
Corrosive
environment
Test
duration,
h
GeF4 + F2
AsF5
AsF5 + F2
SeF6
SeF6
SeF6 + F2
MoF6
MoF6 + F2
MoF6 + F2
TeF6
TeF6
TeF6 + F2
SF6
UF6
F2
F2
8.4
7.0
7.2
7.0
29.5
6.6
9.2
6.0
7.0
18.9
5.7
7.8
28.7
28.8
7.0
8.6
Corrosion rate(a), mm/yr (mpy)

Ni-200
Alloy 400
Nil(b)
0.22 (8.8)
0.22 (8.8)
Nil
0.44 (17.5) 0.67 (26.3)
0.22 (8.8)
0.22 (8.8)
(c)
(c)
0.22 (8.8)
0.22 (8.8)
0.22 (8.8)
Nil
0.22 (8.8)
Nil
(c)
(c)
8.9 (350)
1.8 (70)
135 (5326) 56 (2190)
0.22 (8.8)
0.22 (8.8)
Nil
Nil
0.22 (8.8)
0.66 (26)
0.22 (8.8)
Alloy 600
29 (1139)
Nil
45 (1770)
0.44 (17.5)
22 (850)
0.22 (8.8)
44 (1726)
63 (2488)
2.4 (96)
5.6 (219)
40 (1568)
Nil
9.8 (385)
31 (1209)
(a) Calculated from weight loss after descaling. (b) Rates reported as nil are
less than 0.001 mils/h. (c) Scale was not completely removed by descaling.
Source: Ref 69
Table 6.33 Corrosion of various metals and

alloys in anhydrous HF
Material
Nickel
400
600
Copper
Aluminum
Magnesium
Carbon steel (1020)
304
347
309Cb
310
430
Source: Ref 61
500 C
(930 F)
550 C
(1020 F)
600 C
(1110 F)
0.9 (36)
1.2 (48)
1.5 (60)
1.5 (60)
4.9 (192)
12.8 (504)
15.5 (612)
183 (7,200)
5.8 (228)
12. 2 (480)
1.5 (60)
1.2 (48)
14.6 (576)
457 (18,000)
43 (1,680)
100 (3,960)
9.1 (360)
0.9 (36)
1.8 (72)
1.5 (60)
1.2 (48)
14.6 (576)
7.6 (300)
13.4 (528)
177 (6,960)
168 (6,600)
305 (12,000)
11.6 (456)
26/10/2007 12:29PM Plate # 0
suffered extensive corrosion attack, while alloy

242 remained resistant under the same test condition. The results indicated that Ni-Al system
was not as good as Ni-Mo system in resisting HF
corrosion attack.
Barnes (Ref 76) also tested various pure
metals, which included gold (Au), platinum
(Pt), palladium (Pd), chromium, Ni200, Ni201,
Fig. 6.63
Corrosion of various metals in HF and HF-50H2O

at 850 C (1562 F) for 10 h. Source: Ref 70
Fig. 6.64
Effect of chromium on resistance to HF at 650 C

(1200 F) for Ni-Cr alloys. Source: Ref 71
Ni270, and copper. The test results on pure

metals are summarized in Table 6.37 (Ref 76).
Precious metals (Au, Pt, and Pd) showed no
weight changes and little changes in specimen
surface appearance after testing. However, both
gold and platinum specimens were found to
exhibit surface etching on the specimen surface.
When tested in a higher concentration of HF
(Ar-35HF), palladium was heavily corroded with
some molten corrosion product formed on the
specimen surface. There was no discussion in
the paper about the performance of gold and
platinum in Ar-35HF mixture. Copper was also
found to exhibit few changes in specimen
appearance after testing in Ar-5HF. In fact, the
copper specimen retained its original metallic
luster after testing in Ar-5HF. Chromium was
found to be quite susceptible to uoridation
attack when tested in Ar-5HF. The metal suffered
extensive internal chromium uoride penetration
and extensive weight loss due to vaporization
of chromium uorides.
Nickel was found to be extremely resistant
to HF corrosion attack. Three different types of
nickel were tested: Ni200 and Ni201 (99% pure
nickel), and Ni270 (99.9% pure nickel) (Ref 76).
Ni201 is a low carbon nickel (0.02% C), while
Ni200 is a high carbon nickel (0.15% C). The
impurities and minor elements in these three
nickel specimens in the test program are: 0.15C,
0.4Fe, 0.35Si, 0.35Mn, and 0.25Cu for Ni200;
0.02C, 0.4Fe, 0.35Si, 0.35Mn, and 0.25Cu for
Ni201; and 0.02C, 0.001Cu, and 0.05 max Fe
for Ni270. The test results after 15 h in Ar-5HF
are summarized in Table 6.37 (Ref 76). All nickel
specimens retained metallic luster appearance
after testing. A nickel wire that had been used
for holding test specimens during testing of
specimens in the Ar-5HF test gas at 1000 C for
1600 h was sectioned for metallographic examination, showing only a penetration of about
20 m (0.8 mils) of internal nickel uoride
precipitates. These nickel uorides, which were
randomly distributed, were found to be less than
2 m in diameter. Some nickel uorides were
also detected on the metal surface.
Ceramic materials and graphite were also
tested by Barnes (Ref 76) under the same test
conditions. The results are summarized in
Table 6.38. Graphite was found to be extremely
resistant to HF corrosion. Ceramic materials,
such as alumina (polycrystalline or single
crystal), sintered silicon carbide, chemical vapor
deposited (CVD) silicon carbide, and silicon
nitride, suffered high corrosion rates.
pg 190
26/10/2007 12:29PM Plate # 0
Fig. 6.65
Scanning electron backscattered image (a) and x-ray maps for Cr (b), Ni (c), and F (d) for Ni-40Cr alloy tested at 650 C
(1200 F) for 22 h in anhydrous HF environment. Source: Ref 72. Courtesy of Glyn Marsh
Fig. 6.66
Optical micrograph showing corrosion attack of alloy 600 after testing in HF for 92 h at 650 C (1200 F). Source: Ref 72.
Courtesy of Glyn Marsh
pg 191
26/10/2007 12:29PM Plate # 0
Adding H2O to HF appears to make HF less

corrosive. Figure 6.63 shows the corrosion data
generated by Tyreman and Elliott (Ref 70) by
Fig. 6.69
Optical micrograph showing corrosion attack

on alloy 601 after testing in HF for 16 h at 650 C
(1200 F). Source: Ref 72. Courtesy of Glyn Marsh
Fig. 6.67
Corrosion kinetics of alloys 625, X750, and 601 as a

function of time in HF at 650 C (1200 F). Source:
Ref 71
Fig. 6.70
Optical micrograph showing very little corrosion

attack (about 1 to 2 m in depth) on alloy N
(Ni-5Cr-16Mo) after testing in HF for 46 h at 650 C (1200 F).
Source: Ref 72. Courtesy of Glyn Marsh
Fig. 6.71
Fig. 6.68
Optical micrograph showing corrosion attack on

alloy 625 after testing in HF for 142 h at 650 C
(1200 F). Source: Ref 72. Courtesy of Glyn Marsh
Scanning electron micrograph showing the

corrosion scale (mainly NiF2 with some CrF3, as
determined by x-ray diffraction analysis) formed on alloy N after
testing in HF for 115 h at 650 C (1200 F). Source: Ref 72.
Courtesy of Glyn Marsh
pg 192
26/10/2007 12:29PM Plate # 0
pg 193
comparing the 100% HF and HF-50%H2O

environments for nickel, copper, cobalt, iron,
chromium, molybdenum, niobium, and tantalum. All metals except molybdenum showed
more corrosion attack in HF than in HF-50H2O.
Corrosion data generated in HF-10H2O and HF50%H2O are summarized in Tables 6.39 and
6.40, respectively.
6.4.3 Corrosion in O2-HF Environments
Oxygen appears to make HF more corrosive.
During nuclear fuel reprocessing, stainless steel
fuel cladding was being chemically removed by
reacting it with an O2-HF mixture (4060% HF)
at 377 to 627 C (710 to 1160 F) (Ref 77). The
Table 6.34 Corrosion of molybdenum and
tungsten in HF containing 0.6% H2O
Test temperature,
C (F)
Molybdenum
Tungsten
300 (570)
400 (750)
500 (930)
600 (1110)
700 (1290)
800 (1470)
0.003 (0.1)
0.011 (0.4)
0.014 (0.6)
0.023 (0.9)
0.144 (5.7)
1.3 (51)
0.003 (0.1)
0.009 (0.4)
0.017 (0.7)
0.022 (0.9)
0.91 (36)
removal rates (or corrosion rates) were more than

1 mm/h (40 mils per h (Ref 77). Macheteau et al.
(Ref 78) also found that oxygen contamination
accelerated uoridation attack of iron. Marsh and
Elliott (Ref 79) observed that cobalt exhibited a
protective CoF2 lm with a very low corrosion
rate when exposed to HF at 650 C (1200 F).
However, once air was introduced to mix with
the test gas of HF (i.e., HF-O2-N2 mixture), the
corrosion rate increased signicantly. This is
illustrated in Fig. 6.73 (Ref 79). The rapid
corrosion attack of cobalt when air was mixed
with HF was associated with the formation of
numerous Co3O4 oxide protrusions, as shown in
Fig. 6.74. At the Co3O4 oxide and the cobalt
metal interface, no protective uoride lms were
observed except remnant segments of CoF2 lms
were still present as indicated the lower line in
Fig. 6.74. It is believed that as soon as the air
was introduced to mix with HF gas stream, a
protective CoF2 lm was broken, prompting
the formation of Co3O4 oxide protrusions.
Schutze and Simon (Ref 80) tested a number of
alloys in N2-5O2-3HF at 1100 C for 75 h in
Source: Ref 73
Table 6.35 Corrosion of various nickel-base

alloys in HF at 500 to 600 C (930 to 1110 F)
Alloy
Alloy C
Alloy 600
Alloy B
Ni-200
Ni-201
Alloy 400
Alloy K-500
70Cu-30Ni
0.008 (0.3)
0.018 (0.7)
0.051 (2)
0.229 (9)
0.356 (14)
0.330 (13)
0.406 (16)
0.406 (16)
Tests were performed in HF (7 lb/h) for 36 h. Source: Ref 74
Fig. 6.72
Depth of internal fluoride penetration for alloys

tested in Ar-5HF at 1000 C (1832 F) for 15 h.
Source: Ref 76
Table 6.36 Corrosion behavior of nickel-base alloys along with one cobalt-base alloy (alloy 188)
and nickel aluminide Intermetallic (IC221M) in Ar-5HF at 1000 C (1832 F) for 15 h
Alloy
Alloy 242
Alloy B-3
IC221M
Alloy 188
Alloy 617
Alloy 600
Alloy 602CA
Maximum attack(a),
mm/side (mils/side)
Specimen appearance
Comments
0.015 (0.6)
0.02 (0.8)
0.025 (1.0)
0.051 (2.0)
0.13 (5.3)
0.15 (5.9)
0.18 (7.1)
Surface scale
Surface scale
Surface scale
Surface scale
Nodules on surface
Nodules on surface
Molten corrosion product
Good resistance at 35HF
Extensive corrosion at 35HF(b)
Extensive internal uorides

(a) Surface metal loss + internal uoride penetration. (b) IC221M exhibited good resistance in Ar-5HF, but poor resistance in Ar-35HF. The specimen was almost completely
converted to uorides. Source: Ref 76
26/10/2007 12:29PM Plate # 0
pg 194
Table 6.37
Materials
Corrosion behavior of pure metals in Ar-5HF at 1000 C (1832 F) for 15 h

Corrosion rate, mpy
Specimen appearance
0
0.2
0.2
35.25
0.1
0.1
0.0
0.0
1.1
1454(b)
1.4
1.4
0.0
Unchanged
Matte grey
Unchanged
Green
Unchanged
Unchanged
Unchanged
Gold
Platinum
Palladium
Chromium(a)
Ni200
Ni201
Ni270
Comments
Heavily corroded in 35HF

Total depth of attack = 8.3 mils
No internal attack
No internal attack
No internal attack
(a)Tested for 50 h at 1000 C in Ar-5HF. (b) Based on total depth of attack (maximum metal affected) of 8.3 mils/side in a 50 h test. Source: Ref 76
Table 6.38 Corrosion behavior of ceramic

materials and graphite in Ar-5HF at 1000 C
(1832 F) for 15 h
Materials
Graphite (CZR-1)
Graphite (DFP-2)
Alumina (polycrystalline)
Alumina (sapphire)
CVD SiC
Sintered SiC
Si3N4
Weight
change,
mg/cm2
Corrosion
rate, mpy
Comments
0
0
2.4
0.9
5.7
6.4
2.2
0.0
0.0
141
52
411
458
155
Unchanged
Unchanged
General wastage
Frosted appearance
General wastage
General wastage
General wastage
Source: Ref 76
Table 6.39 Corrosion of several alloys in

HF-10H2O at 850 C (1560 F)
Alloy
304LSS
310SS
800H
600
625
100130 (40005000)
100130 (40005000)
100130 (40005000)
23 (900)
6.7 (265)
Source: Ref 70
Table 6.40 Corrosion of nickel and Alloy 400 in

50HF-50H2O at various temperatures
their search for a candidate alloy for air nozzles

in a circulating uidized-bed combustor for reprocessing spent pot lining material in aluminum
production. Among the metallic materials tested,
alloy 242 was found to exhibit the smallest
metal loss (50 m). Although the alloy suffered
extensive nitridation attack, the authors considered that nitridation attack would not affect
the alloys performance as air nozzles. In a lowtemperature test (450 C) in N2-8O2-10CO215H2O-5HF, Crum et al. (Ref 81) observed no
measurable corrosion attack for many nickelbase alloys and a superaustenitic stainless steel
after 155 h of exposure. Their test results are
summarized in Table 6.41. Stress-corrosion
cracking (SCC) resistance of these alloys was
also included in the test program using the same
test environment with U-bend test specimens.
All alloys showed no cracking after 100 and
155 h except alloy 600. Authors did not explain
why alloy 600 suffered SCC while other alloys
including many nickel-base alloys and one
superaustenitic stainless steel (alloy 25-6MO)
showed no cracking.
6.5 Corrosion in Bromine and Iodine

Environments
Very little data have been reported on the
performance of metals and alloys in bromine and
550 C
(1020 F)
600 C
(1110 F)
650 C
(1200 F)
700 C
(1290 F)
750 C
(1380 F)
Nickel
0.79 (31) 1.83 (72) 2.74 (108) 3.66 (144) 3.05 (120)
Alloy 400
0.61( 24) 1.52 (60) 3.96 (156) 5.18 (204)

Source: Ref 61
iodine environments at elevated temperatures.

Miller et al. (Ref 82) investigated the corrosion
behavior of copper, nickel, and nickel-base
alloys in bromine at 300 and 500 C (470 and
930 F) (Table 6.42). Copper suffered rapid
attack by bromine at 300 C (570 F). The CuBr
compound exhibits very high vapor pressures.
The vapor pressure of CuBr reaches 1 104 atm
(a value considered high enough to cause hightemperature corrosion) when the temperature is
as low as 435 C (815 F) (Table 6.3). Nickel
exhibited good corrosion resistance in Br2 at
300 and 500 C (470 and 930 F). Bromine
reacts with nickel to form NiBr2, which melts at
965 C (1769 F), signicantly higher than the
melting point of CuBr. Duranickel 301 (Ni-5Al)
had a corrosion resistance similar to nickel.
Alloy 400 was less resistant than nickel and
Duranickel 301.
Smith and Ganesan (Ref 83) investigated the
corrosion behavior of various commercial alloys
in a simulated combustion environment containing a very high level of HBr (about 4%) at
593 and 927 C (1100 and 1700 F). The test
26/10/2007 12:30PM Plate # 0
pg 195
gas (the inlet gas) consisted of 9% O2, 12% CO2,

4% HBr, 100 ppm SO2, and balance N2. The
equilibrium partial pressures of the test environment at 593 and 927 C (1100 and 1700 F) are
listed in Table 6.43. The test results, which are
summarized in Table 6.44, showed surprisingly
very little corrosion attack at 593 C (1100 F)
for all the alloys tested, which included
austenitic stainless steels (Type 309, 316, and
347), alloy 800, and nickel-base alloys. At

927 C (1700 F), nickel-base alloys showed
little weight loss, while austenitic stainless steels,
Fe-Ni-Cr alloy 800, and high-iron-containing
nickel alloy 825 suffered high weight losses.
The authors (Ref 83) analyzed the corrosion
products spalled from test specimens as well as
Table 6.41 Corrosion rates generated from
tests in N2-8O2-10CO2-15H2O-5HF at 450 C
(842 F) for 100 and 155 h
Alloy
Alloy 600
Alloy 400
Alloy 825
Alloy C-276
Alloy 686
Alloy 622
Alloy 25-6MO
Corrosion rate,
mm/yr (mpy)
from 100 h tests
Corrosion rate,
mm/yr (mpy)
from 155 h tests
0.01 (0.4)
0
0.01 (0.4)
0
0
0
0
0
0.002 (0.1)
0
0
0
0
0
Source: Ref 81
Table 6.42 Corrosion of several alloys in

bromine at 300 and 500 C (570 and 930 F)
Alloy
Ni-201
Fig. 6.73
Weight change of cobalt at 650 C (1200 F) when

exposed to HF during the first 12 h, showing a very
low corrosion rate. At time X, air was introduced into the test gas
to mix with HF, rapid corrosion attack was observed. Source:
Ref 79
Temperature,
C (F)
300 (570)
300 (570)
300 (570)
Duranickel 301 300 (570)
300 (570)
400
300 (570)
300 (570)
Copper
300 (570)(a)
300 (570)(a)
Ni-201
500 (930)
500 (930)
Duranickel 301 500 (930)
500 (930)
Time,
days
Weight loss,
mg/cm2
Corrosion rate,
mm/yr (mpy)
11
11
11
10
10
2
2
<1
<1
10
10
10
10
0.28
0.16
0.14
0.17
0.07
7.07
6.65
Rapid attack
Rapid attack
3.83
1.40
2.14
3.64
<0.03 (1.2)
<0.03 (1.2)
<0.03 (1.2)
<0.03 (1.2)
<0.03 (1.2)
1.5 (60)
1.5 (60)
0.15 (6.0)
0.06 (2.4)
0.09 (3.6)
0.15 (6.0)
(a) Tests were terminated before 300 C was reached because of rapid corrosion
attack. Source: Ref 82
Table 6.43 Thermodynamic equilibria in terms

of partial pressures (atm) for the inlet gas of
N2-9O2-12CO2-4HBr-100 ppm SO2 at 593 and
927 C (1100 and 1700 F)
Gas component
Fig. 6.74
A Co3O4 oxide protrusion formed on cobalt

immediately after test gas was switched from HF
to air-HF mixture at 650 C. The upper line indicates the Co3O4
oxide protrusion, and the lower line indicates the CoF2 phase
formed at the interface between the Co3O4 oxide and the substrate
cobalt. Source: Ref 72. Courtesy of Glyn Marsh
N2
CO2
O2
Br2
H 2O
HBr
SO3
SO2
H2
Source: Ref 83
Partial pressure
at 593 C, atm
Partial pressure
at 927 C, atm
0.78
0.12
0.81 101
0.20 101
0.20 101
0.11 103
0.90 104
0.63 106
0.39 1014
0.78
0.12
0.81 101
0.19 101
0.19 101
0.22 102
0.82 105
0.82 104
0.82 109
26/10/2007 12:30PM Plate # 0
the extraction residue obtained from the test

specimens using x-ray diffraction technique to
determine the phases present. The results of the
x-ray diffraction analysis are summarized in
Table 6.45 for the corrosion scales spalled from
tested specimens and in Table 6.46 for the
extraction residue obtained from tested specimens by electrolytic extraction of phases. In both
cases, the phases found were oxides. No bromide
Table 6.44 Mass change data for alloys tested

in N2-9O2-12CO2-4HBr-100 ppm SO2 at 593
and 927 C (1100 and 1700 F) for 300 h
Mass change, mg/cm2
Alloy
593 C (1100 F)
927 C (1700 F)
309
316
347
800
825
625
601
600
617
690
0.19
0.75
0.59
0.55
0.11
0.10
0.56
0.55
0.15
0.50
218.70
83.35
67.57
53.25
28.57
8.77
7.78
6.04
5.57
3.88
phases were detected. The results by Smith and

Ganesan (Ref 83) suggest that nickel-base alloys
appear to be suitable for applications in HBrcontaining environments up to about 4%. It is
also suggested that HBr is not as corrosive as HCl
at elevated temperatures. More tests, particularly
long-term tests, are needed to conrm the current
ndings.
Corrosion of various alloys in I2 was investigated by Shapiro (Ref 84) when determining the
most suitable alloy for construction of vessels
for iodine zirconium processing in the production of zirconium. Tests were conducted at
300 and 450 C (570 and 840 F) in iodine vapor
with 400 mm Hg (0.53 atm) pressure. Results
are shown in Table 6.47 and 6.48. Platinum, gold,
tungsten, and molybdenum were very resistant
to corrosion attack in I2. Alloy B (Ni-Mo), alloy
C (Ni-Cr-Mo), alloy 600 (Ni-Cr-Fe), and nickel
were quite resistant to I2. Stainless steels and
alloy 400 were less resistant. In the production
of zirconium using the iodide process, closure
gaskets made of gold are used for handling dry
iodine vapors at 500 C (939 F) (Ref 85).
Source: Ref 83
Table 6.45 Results of x-ray diffraction analysis

on the spalled corrosion products for the test
specimens after testing in N2-9O2-12CO2-4HBr100 ppm SO2 at 927 C (1700 F) for 300 h
Alloy
Major phases
Minor phases
316
800
825
600
601
617
625
690
Fe2O3
NiFe2O4, Cr2O4
Cr2O3, NiCr2O4
Cr2O3, NiFe2O4
Cr2O3
Cr2O3
Cr2O3
Cr2O3, Fe2O3
NiCr2O4, M3O4
Fe2O3
M 3O 4
M 3O 4
M 3O 4
M 3O 4
Table 6.47 Corrosion rates of several metals

and alloys in iodine at 300 C (570 F)(a)
Alloy
Platinum
Gold
Tungsten
Molybdenum
Tantalum
Alloy B
Alloy C
Alloy 600
Nickel
Alloy 400
0
0
0
0.003 (0.11)
0.004 (0.16)
0.044 (1.7)
0.057 (2.2)
0.107 (4.2)
0.27 (10.6)
0.56 (22.0)
(a)400 mm Hg (0.53 atm) iodine pressure. Source: Ref 84
Source: Ref 83
Table 6.48 Corrosion rates of several metals

and alloys in iodine at 450 C (840 F)(a)
Table 6.46 Results of x-ray diffraction
analysis on the extracted residue from the test
specimens after testing in N2-9O2-12CO2-4HBr100 ppm SO2 at 927 C (1700 F) for 300 h
Alloy
Major phases
Minor phases
316
800
825
600
601
617
625
690
Cr2O3, M3O4
NiFe2O4, Cr2O3
Cr2O3
NiCrO3
Al2O3, Cr2O3
Cr2O3
Cr2O3, NiCr2O4
Cr2O3
NiO
NiFe2O4
NiO
M3O4
NiFe2O4
Source: Ref 83
Alloy
Platinum
Tungsten
Gold
Molybdenum
Alloy B
Alloy 600
Tantalum
Nickel
Type 310SS
Type 316SS
Type 347SS
Alloy 400
Type 304SS
0.0055 (0.2)
0.008 (0.3)
0.024 (0.9)
0.033 (1.3)
0.46 (18)
0.54 (21)
0.88 (35)
1.2 (47)
1.8 (71)
2.1 (83)
2.1 (83)
2.6 (102)
3.2 (126)
(a)400 mm Hg (0.53 atm) iodine pressure. Source: Ref 84
pg 196
26/10/2007 12:30PM Plate # 0
6.6 Summary
High-temperature corrosion behavior of
metals and alloys in gaseous environments containing (a) chlorine and hydrogen chloride, (b)
uorine and hydrogen uoride, (c) bromine and
hydrogen bromide, and (d) iodine and hydrogen
iodide is reviewed. Metals or alloys in these
environments form metal halides as corrosion products. Many metal halides exhibit high
vapor pressures; the corrosion products can
thus be in a gaseous phase. Some metal halides
also exhibit low melting points; the corrosion
products can thus be in a liquid phase. As a result,
the metal-halogen or metal-hydrogen-halide
reactions can proceed at a rapid rate at elevated
temperatures.
The corrosion behavior of alloys can be
signicantly different between oxidizing
environments that contain O2 and reducing
environments that contain no O2. Also, corrosion
by halogen or by hydrogen halide can proceed at
different rates. Accordingly, the corrosion data
are categorized in different environment types.
The arrangement of the corrosion data in this way
may also help readers to determine the type of
data that may be more pertinent to their intended
applications or conditions. For the corrosion
behavior of alloys in environments containing
Cl2 and HCl, the data are presented in (a) Cl2
environments containing no O2, (b) O2-Cl2
mixtures, (c) O2-HCl mixtures, and (d) HCl and
HCl-bearing, reducing environments. For corrosion in F2- and HF-bearing environments, the
data are presented in (a) F2 environments, (b)
HF environments, and (c) O2-HF mixtures. For
bromine- and iodine-containing environments
the corrosion data are very limited, and a brief
review is presented.
REFERENCES
1. W.J. Kroll, Method of Manufacturing Ti

and Alloys Thereof, U.S. Patent 2205854,
Jan 1940
2. W.A. Henderson, J. Met., Vol 16, 1964, p 155
3. S.M. Shelton, The Metallurgy of Zirconium,
B. Lustman and F. Kerze, Jr., Ed., McGrawHill, 1955, p 59
4. I. Iwasaki, Y. Takahasi, and H. Kahata,
Trans. SME AIME, Vol 243, 1966, p 308
5. C.L. Mantell, Tin, Reinhold, 1949
6. W.E. Berry, Corrosion in Nuclear Applications, John Wiley & Sons, 1971
7. T. Rosenquist, Principles of Extractive

Metallurgy, McGraw-Hill, 1974
8. P.L. Daniel and R.A. Rapp, Advances in
Corrosion Science and Technology, Vol 5,
M.G. Fontana and R.W. Staehle, Ed., Plenum Press, 1970, p 55
9. O. Kubaschewski and E. Evans, Metallurgical Thermochemistry, Pergamon Press,
1958
www.outokumpotechnology.com, accessed
Dec 2006
11. ChemSage, Version 4.16, GTT-Technologies, Aachen, 1998
12. M.J. McNallan, W.W. Liang, J.M. Oh,
and C.T. Kang, Morphology of Corrosion
Products Formed on Cobalt and Nickel
in Argon-Oxygen-Chlorine Mixtures at
1000 K, Oxid. Met., Vol 17, 1982, p 371
13. A. Zahs, M. Spiegel, and H.J. Grabke, The
Inuence of Alloying Elements on the
Chlorine-Induced High Temperature Corrosion of Fe-Cr Alloys in Oxidizing Atmospheres, Mater. Corros., Vol 50, 1999, p 561
14. F.H. Stott and C.Y. Shih, The Inuence of
HCl on the Oxidation of Iron at Elevated
Temperatures, Mater. Corros., Vol 51, 2000,
p 277
15. R. Bender and M. Schutze, The Role of
Alloying Elements in Commercial Alloys
for Corrosion Resistance in OxidizingChloridizing AtmospheresPart I: Literature Evaluation and Thermodynamic
Calculations on Phase Stabilities, Mater.
Corros., Vol 54, 2003, p 567
16. K.L. Tseitlin and V.A. Strunkin, J. Appl.
Chem. USSR, Vol 31, 1958, p 1832
17. K.L. Tseitlin, J. Appl. Chem. USSR, Vol 28
(No. 5), 1955, p 467
18. G. Heinemann, F.G. Garrison, and P.A.
Haber, Ind. Eng. Chem., Vol 38, 1946, p 497
19. R.J. Fruehan, Metall. Trans., Vol 3 (No. 10),
1972, p 2585
20. S.F. Bohlken, A. Klinkenberg, and
H.W. Nicolai, Ind. Chim. Belge., Vol 20,
1955, p 579
21. S.F. Bohlken, A. Klinkenberg, and H.W.
Nicolai, C. R. Cong. Int. Chim. Ind., Vol 27,
1954, p 2
22. M.H. Brown, W.B. DeLong, and J.R. Auld,
Ind. Eng. Chem., Vol 39 (No. 7), 1947, p 839
23. G. Han and W.D. Cho, High-Temperature
Corrosion of Fe3Al in 1% Cl2/Ar, Oxid.
Met., Vol 58, 2002, p 390
pg 197
26/10/2007 12:30PM Plate # 0
24. B.J. Downey, J.C. Bermel, and P.J.

Zimmer, Corrosion, Vol 25 (No. 12), 1969,
p 502
25. J.D. McKinley, Jr. and K.E. Shuler, J. Chem.
Phys., Vol 28, 1958, p 1207
26. R.H. Kane, Process Industries Corrosion,
B.J. Moritz and W.I. Pollock, Ed., NACE,
1986, p 45
27. MTI Publication MS-3, Materials Selector
for Hazardous Chemicals, Vol 3, Hydrochloric Acid, Hydrogen Chloride and
Chlorine, Materials Technology Institute
of the Chemical Process Industries, Inc.,
C.P. Dillon and W.I. Pollock, Ed., 1999
28. J.P. Tu, Z.Z. Li, and Z.Y. Mao, Internal
Chlorination of Ni-Based Alloys and Its
Relation to Volatilization Corrosion, Mater.
Corros., Vol 48, 1997, p 441
29. M.J. Maloney and M.J. McNallan, Met.
Trans. B, Vol 16, 1985, p 751
30. M.J. McNallan and W.W. Liang, GaseousTransport-Controlled Chlorination of CoO
in Flowing Oxygen-Chlorine-Argon Mixtures at 1000 K, J. Am. Ceram. Soc., Vol 64
(No. 5), 1981, p 302
31. N.S. Jacobson, M.J. McNallan, and
Y.Y. Lee, The Formation of Volatile Corrosion Products During the Mixed OxidationChlorination of Cobalt at 650 C, Metall.
Trans., Vol 17A, 1986, p 1223
32. Y.Y. Lee and M.J. McNallan, Ignition of
Nickel in Environments Containing Oxygen
and Chlorine, Metall. Trans., Vol 18A, 1987,
p 1099
33. M.J. McNallan, High-Temperature Corrosion in Halogen Environments, MP, Sept.
1994, p 54
34. J.C. Liu and M.J. McNallan, Effects of
Temperature Variations on Oxidation of
Iron-20% Chromium Alloys at 1200 K
in Ar-20% O2-Cl2 Gas Mixtures, Mater.
Corros., Vol 50, 1999, p 253
35. S. Baranow, G.Y. Lai, M.F. Rothman, J.M.
Oh, M.J. McNallan, and M.H. Rhee, Paper
36. J.M. Oh, M.J. McNallan, G.Y. Lai, and M.F.
Rothman, Metall. Trans. A, Vol 17A, 1986,
p 1087
37. N.S. Jacobson, M.J. McNallan, and Y.Y.
Lee, Mass Spectrometric Observations of
Metal Oxychlorides Produced by OxidationChlorination Reactions, Metall. Trans. A,
Vol 20A, 1989, p 1566
38. M.H. Rhee, M.J. McNallan, and M.F.
Rothman, High Temperature Corrosion in
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
Energy Systems, M.F. Rothman, Ed., The

Metallurgical Society of AIME, 1985,
p 483
M.J. McNallan, M.H. Rhee, S. Thongtem,
and T. Hensler, Paper No. 11, Corrosion/85,
NACE, 1985
P. Elliott, A.A. Ansari, R. Prescott, and
M.F. Rothman, Paper No. 13, Corrosion/85,
NACE, 1985
S. Thongtem, M.J. McNallan, and G.Y. Lai,
C. Schwalm and M. Schutze, The Corrosion Behavior of Several Heat Resistant
Materials in Air + 2 vol-% Cl2 at 300 to
800 C: Part 1Fe-Base and Fe-Containing
Alloys, Mater. Corros., Vol 51, 2000, p 34
C. Schwalm and M. Schutze, The Corrosion
Behavior of Several Heat Resistant Materials in Air +2% Cl2 at 300 to 800 C: Part
2Nickel Base Alloys, Mater. Corros.,
Vol 51, 2000, p 73
C. Schwalm and M. Schutze, The Corrosion
Behavior of Several Heat Resistant Materials in Air + 2% Cl2 at 300 to 800 C: Part
3Alumina Formers and Intermetallics,
M.J. McNallan, S. Thongtem, J.C. Liu,
Y.S. Park, and P. Shyu, Corrosion of Chromium Containing Alloys in Non-steady
State Environments Containing Oxygen,
Carbon, and Chlorine, J. Phys. IV, Coll. C9,
Suppl. J. Phys. III, Vol 3, Dec 1993, p 143
M.J. McNallan, Z. Niemczura, H.H. Lu,
and Y.S. Park, Carbide Formation and Oxidation of Austenitic Alloys in Oxidizing/
Carburizing Environments Contaminated
with Chlorine, Paper No. 252, Corrosion/93,
NACE, 1993
Y. Ihara, H. Ohgame, K. Sakiyama, and
K. Hashimoto, Corros. Sci., Vol 21 (No. 12),
1981, p 805
Y. Ihara, H. Ohgame, and K. Sakiyama,
Y. Ihara, H. Ohgame, and K. Sakiyama,
F. Devisme, P. Falgoux, F. Lefebvre, and
T. Flament, High Temperature Corrosion
in Atmospheres Containing Hydrogen
Chloride, 11th International Incineration Conference (Albuquerque, NM), May
1115, 1992
P. Ganesan, G.D. Smith, and L.E.
Shoemaker, The Effects of Excursions of
Oxygen and Water Vapor Contents on
Nickel-Containing Alloy Performance in
pg 198
26/10/2007 12:30PM Plate # 0
52.
53.
54.
55.
56.
57.
58.
59.
60.
61.
62.
63.
A Waste Incineration Environment, Paper

G.D. Smith and P. Ganesan, Metallic
Corrosion in Waste Incineration: A Look at
Selected Environmental and Alloy Fundamentals, Heat-Resistant Materials II, Conf.
Proc. Second International Conference
on Heat-Resistant Materials, K. Natesan,
P. Ganesan, G. Lai, Ed., ASM International,
1995, p 631
P. Elliott, R. Prescott, and C.J. Tyreman,
Aspects of Halogen Corrosion at High
Temperatures, Paper No. 373, Corrosion/86,
NACE, 1986
M.K. Hossain, J.E. Rhoades-Brown, S.R.J.
Saunders, and K. Ball, Proc. UK Corrosion/
83, 1983, p 61
G.Y. Lai, M.F. Rothman, S. Baranow, and
R.B. Herchenroeder, Recuperator Alloys
for High-Temperature Waste Heat Recovery,
J. Met., Vol 35 (No. 7), 1983, p 24
U. Brill, J. Klower, and D.C. Agarwal,
Inuence of Alloying Elements on the
Chlorination Behavior of Nickel- and
Iron-Based Alloys, Paper No. 439, Corrosion/96, NACE International, 1996
J.J. Barnes, Materials Limitations in High
Temperature Chlorination Environments,
K.N. Strafford, P.K. Datta, and G. Forster,
High-Temperature Chloridation of Binary
Fe-Cr Alloys at 1000 C, High Temperature
Corrosion 2: Advanced Materials and Coatings, Proc. Second International Symposium
on High Temperature Corrosion of Advanced Materials and Coatings (Les Embiez,
France), May 2226, 1989, R. Streiff,
J. Stringer, R.C. Krutenat, and M. Caillet,
Ed., Elsevier Science Publishers, London,
1989, p 61
J.P. Carter, B.S. Covino, Jr., T.J. Driscoll,
W.D. Riley, and M. Rosen, Corrosion,
Vol 40 (No. 5), 1984, p 205
L. Reeve, J. Iron Steel Inst., Sept 1955, p 26
W.R. Myers and W.B. DeLong, Chem. Eng.
Prog., Vol 44 (No. 5), 1948, p 359
R.L. Jarry, J. Fischer, and W.H. Gunther,
J. Electrochem. Soc., Vol 110 (No. 4), 1963,
p 346
R.L. Jarry, W.H. Gunther, and J. Fischer,
The Mechanism and Kinetics of the
Reaction Between Nickel and Fluorine,
ANL-6684, Argonne National Laboratory,
Argonne, IL, Aug 1963
64. K. Hauffe, Z. Werkstofftech., Vol 15, 1984,

p 427
65. Y.A. Lukyanchev, I.I. Astakhov, and N.S.
Nikolaev, Mechanism of the Formation of
Fluoride Films on Nickel and Their Properties, Bull. Acad. SSSR. Div. Chem. Sci.,
1965, p 577
66. C.F. Hale, E.J. Barber, H.A. Bernhardt, and
K.E. Rapp, High Temperature Corrosion of
Some Metals and Ceramics in Fluorinating
Atmospheres, Report K-1459, Union Carbide Nuclear Co., Sept 1960
67. M.J. Steindler and R.C. Vogel, Corrosion
of Materials in the Presence of Fluorine
at Elevated Temperatures, ANL-5662,
Argonne, IL, Jan 1957
68. R.B. Jackson, Corrosion of Metals and
Alloys by Fluorine, NP-8845, Allied
Chemical Corp., Morristown, PA, 1960
69. R.C. Vogel, J.H. Schraidt, and J. Royal,
Chemical Engineering Division SemiAnnual Report, ANL-7125, Argonne
National Laboratory, Argonne, IL, May
1966
70. C.J. Tyreman and P. Elliott, Paper No. 135,
71. G. Marsh and P. Elliott, High Temperature Corrosion in Energy Systems, M.F.
Rothman, Ed., The Metallurgical Society of
AIME, 1985, p 467
72. G. Marsh, Ph.D. thesis, UMIST, Manchester, U.K., Dec 1982
73. V.S. Zotikov and E.Ya, Semenyuk, Metallschutz Moscow, Vol 6, 1970, p 218
74. Corrosion Resistance of Nickel-Containing
Alloys in Hydrouoric Acid, Hydrogen
Fluoride and Fluorine, CEB-5, International Nickel Co., 1968
75. Haynes International, Inc., unpublished
results, 1988
76. J.J. Barnes, Materials Behavior in High
Temperature HF-Containing Environments,
International, 2000
77. G. Dumont and W. Goossens, Chemical
Decladding of Fuel Elements with HF/O2
Mixture, Chem. Eng. Technol., Vol 43, 1971,
p 800
78. Y. Macheteau, J. Gillardeau, P. Plurien, and
J. Oudar, Oxid. Met., Vol 4 (No. 3), 1972,
p 141
79. G. Marsh and P. Elliott, Aspects of High
Temperature Hydrouorination of Cobalt,
p 215
pg 199
26/10/2007 12:30PM Plate # 0
80. M. Schutze and P. Simon, Materials for Use

in Atmospheres Containing HF in a Pyrohydrolysis Plant for Re-processing Spent Pot
Lining from Aluminum Production, Werkst.
Korros., Vol 45, 1994, p 435
81. J.R. Crum, G.D. Smith, M.J. McNallan, and
S. Himyj, Characterization of Corrosion
Resistant Materials in Low and High Temperature HF Environments, Paper No. 382,
82. P.D. Miller, E.F. Stephan, W.E. Berry, and
W.K. Boyd, Corrosion Resistance of
Nickel and Two Nickel Alloys to Gaseous
Bromine, BMI-X-489, Battelle Memorial

Institute, Columbus, Ohio, Jan 1968
83. G.D. Smith and P. Ganesan, Performance
and Scale Formation of Selected High
Temperature Alloys in Simulated Waste
Incineration Environments Containing
Gaseous Bromides and Chlorides, Paper No.
84. Z.M. Shapiro, Metallurgy of Zirconium, B.
Lustman and F. Kerze, Jr., Ed., McGrawHill, 1955, p 135
85. Handbook of Corrosion Data, B.D. Craig,
Ed., ASM International, 1989
pg 200

DOI: 10.1361/hcma2007p201
26/10/2007 11:13AM Plate # 0

CHAPTER 7
Sulfidation
7.1 Introduction
Sulfur is one of the most common corrosive
contaminants in high-temperature industrial
environments. Sulfur is generally present as an
impurity in fuels or feedstocks. Typically, fuel
oils are contaminated with sulfur varying from
fractions of 1% (No. 1 or No. 2 fuel oil) to about
3% (No. 6 fuel oil). United States coal may
contain from 0.5 to 5% S, depending on where it
is mined. Feedstocks for calcining operations in
mineral and chemical processing are frequently
contaminated with various amounts of sulfur.
Sulfur is one of the corrosive species in petroleum-rening operations.
When combustion takes place with excess air
to ensure complete combustion of fuel for generating heat in many industrial processes, sulfur
in the fuel reacts with oxygen to form SO2 and
SO3. A combustion atmosphere of this type is
oxidizing in nature. Oxidizing environments
are usually much less corrosive than reducing
environments, where sulfur is in form H2S.
However, suldation in oxidizing environments
(as well as in reducing environments) is frequently accelerated by other fuel impurities, such
as sodium, potassium, and chlorine, which may
react among themselves and/or with sulfur during combustion to form salt vapors. These salt
vapors may then deposit at lower temperatures
on metal surfaces, resulting in accelerated suldation attack. This accelerated suldation attack
due to salt deposits is frequently referred to
as hot corrosion in gas turbines (Chapter 9),
coal ash corrosion in coal-red boilers
(Chapter 10), and oil ash corrosion in oil-red
boilers (Chapter 11). They are discussed in these
chapters.
Oxidizing environments that contain high
concentrations of SO2 can be produced in the
chemical process used to manufacture sulfuric
acid. Sulfur, in this case, is used as a feedstock.
Combustion of sulfur with excess air takes place
in a sulfur furnace at about 1150 to 1200 C
(2100 to 2200 F). The product gas typically

contains about 10 to 15% SO2 along with 5 to
10% O2 (balance N2), which is then converted to
SO3 for sulfuric acid.
In many industrial processes, combustion is
carried out under stoichiometric or substoichiometric conditions to convert feedstock to process
gases consisting of H2, CO, CH4, and other
hydrocarbons. Sulfur is converted to H2S. The
environment, in this case, is reducing and is
characterized by low oxygen potentials. Coal
gasication, which converts coal to substitute
natural gas or medium- and low-BTU fuel gases,
is an example of a process that generates this type
of atmosphere. Reducing conditions may also
prevail in localized areas, in some cases, even
when combustion is taking place with excess air.
A coal-red boiler equipped with low NOx
combustion burners is an example. More discussion on this issue is presented in Chapter 10
on high-temperature corrosion and other
materials issues in coal-red boilers. Furthermore, ash deposits on the metal surface can
sometimes turn an oxidizing condition in the
gaseous environment into a reducing condition
beneath the deposits.
In most cases, alloys rely on oxide scales to
resist suldation attack; most high-temperature
alloys rely on chromium oxide scales. In oxidizing environments, oxide scales form readily
because of high oxygen activities. Thus, oxidation dominates the corrosion reaction. When the
environment is reducing (i.e., low oxygen
potentials), the corrosion reaction becomes a
competition between oxidation and suldation.
Thus, lowering the oxygen activity tends to make
the environment more suldizing, resulting in
increased domination by sudation. Conversely,
increasing the oxygen activity generally results
in a less-suldizing environment with increased
domination by oxidation. Suldation reaction is
thus controlled by both sulfur and oxygen
activities (also referred to as potentials). When
corrosion involves more than one mode,
pg 201
26/10/2007 11:13AM Plate # 0
including suldation, suldation generally dictates materials selection.

Suldation attack, in many cases, can be quite
localized. Figure 7.1 shows a high-nickel alloy
that suffered suldation attack at about 930 C
(1700 F) in a furnace ring ceramic tiles. The
cross section at the corroded area showed suldes through the cross section of the component.
The breakdown of a protective oxide scale (i.e.,
Cr2O3 scale for most high-temperature alloys)
usually signies the initiation of breakaway
corrosion, which is generally followed by rapid
corrosion attack. Another example of component
failure due to suldation attack is illustrated in
Fig. 7.2. Alloy 800H suffered breakaway corrosion. The chromium oxide scale broke down and
was replaced by iron-rich oxides, nickel-rich
suldes, and chromium-rich suldes, as illustrated in Fig. 7.2(b). The alloy matrix beneath
the corrosion products was found to be severely
depleted of chromium, as shown in Fig. 7.2(b).
Materials problems related to suldation
attack have been encountered in various industries. Some of these problems have been reported
in calcining of mineral and chemical feedstock
(Ref 1), petrochemical processing (Ref 2), fossilred boilers (Ref 2), petroleum rening (Ref 3),
coal gasication (Ref 4, 5), waste incineration
(Ref 6, 7), uidized-bed coal combustion (Ref
810), and oil-red boilers.
This chapter reviews corrosion data involving
mainly gaseous environments. The data are
grouped into three different types typical of
industrial environments: (1) H2/H2S mixtures
and sulfur vapor with extremely low oxygen
activities such that a Cr2O3 scale is not thermodynamically stable, (2) reducing, mixed-gas
environments with sufciently high oxygen
activities such that Cr2O3 is thermodynamically
stable, and (3) SO2-bearing environments.
Suldation accelerated by salt deposits (hot
corrosion in gas turbines, coal ash corrosion, and
oil ash corrosion) are covered in Chapters 9 to 11,

respectively.

Iron, nickel, and cobalt are the alloy bases for a
majority of high-temperature alloys. Most hightemperature alloys rely on chromium to form a
protective chromium oxide scale to resist oxidation and other high-temperature corrosion
attack. When the sulfur potential is high enough,
suldes of iron, nickel, cobalt, and chromium are
likely to form under suldation attack. An
Ellingham diagram such as the one shown in
Fig. 7.2
Fig. 7.1
Alloy 601 tube suffering localized sulfidation attack.

The tube was in service at about 930 C (1700 F) in
a natural gas-fired furnace making ceramic tiles. Sulfur was
believed to come from the ceramic feedstock.
Alloy 800H recuperator suffering severe sulfidation

attack in a nonferrous metal scrap melting furnace.
The 9.5 mm (0.4 in) thick recuperator was perforated in less than
2 years at metal temperatures of about 650 to 760 C (1200 to
1400 F). (a) General view of a corroded sample. (b) Cross section
of the sample showing corrosion products. 1, Fe-rich oxide (57Fe18Ni-16Cr-5S-4K); 2, Fe-rich oxide (80Fe-8Ni-12Cr); 3, Ni sulfide
(light gray phase) (72Ni-2Cr-27S); 4, Cr depleted zone (72Ni28Fe); 5, Cr-rich sulfide (internal phase) (40Cr-10 Fe-50S).
pg 202
26/10/2007 11:13AM Plate # 0
Chapter 7:
Fig. 7.3 (Ref 11) can help determine whether an

environment has a sulfur potential high enough
to form suldes. The sulfur potential is represented by either pS2 or pH2 S =pH2 . The sulfur
partial pressure (pS2 ) in equilibrium with a sulde
Sulfidation / 203
can be read from Fig. 7.3 by drawing a straight

line from point S through the free-energy line
of the sulde phase through the temperature of
interest, and intersecting with the pS2 scale. The
intersection at the pS2 scale gives the sulfur
H2S/H2 ratio
Fig. 7.3
pg 203
Standard free energies of formation of selected sulfides. Source: Ref 11
26/10/2007 11:13AM Plate # 0
pg 204
Metal-sulfur-oxygen stability diagrams at

870 C (1600 F) for iron, nickel, cobalt, and
chromium are shown in Fig. 7.4 to 7.7 (Ref 15).
The diagrams map out the regions within which
different phases will be in equilibrium at the gas/
metal interface. Gordon and Nagarajan (Ref 16)
constructed stability diagrams for the Fe-Cr-Ni
system at 870 C (1600 F) as shown in Fig. 7.8.
The spinel phases were deleted from the original
gure for better clarity. In an environment
marked X, for example, phases of Cr2O3, FeS,
and NiS are likely to form at the gas/metal
interface for the Fe-Cr-Ni system.
Perkins (Ref 14) proposed that in the environment within the upper right region of the CrS-O diagram (Fig. 7.9) both CrS and Cr2O3 will
form initially on the metal surface of chromium
and high-chromium alloys. The Cr2O3 will grow
and overtake the sulde because it is a stable
phase in this phase stability region. The alloying
elements that form stable suldes can diffuse
through the oxide scale and eventually form
suldes on the surface of the oxide scale, leading
to breakaway corrosion. It is particularly serious
when the suldes become liquid. Some metalmetal sulde eutectics have low melting
partial pressure in equilibrium with the sulde

phase. The pH2 S =pH2 value in equilibrium with a
sulde can be obtained similarly by using the
starting point H for the straight line; the intersection with H2S/H2 scale gives the equilibrium
H2S/H2 value. It is clear from Fig. 7.3 that the
sulde of chromium is more stable than those of
iron, nickel, and cobalt.
Most suldizing environments exhibit both
sulfur and oxygen activities. It is thus better to
characterize the environment in terms of pS2 and
pO2 . Both sulfur and oxygen potentials of an
environment containing various gaseous components in an equilibrium condition can be calculated using commercial software, such as HSC
(Ref 12) and Chemsage (Ref 13). The environment in terms of pS2 and pO2 can be presented in a
stability diagram of a metal-sulfur-oxygen system. Perkins (Ref 14) and Hemmings and Perkins
(Ref 15) discuss briey how stability diagrams
can be used to aid understanding corrosion
behavior. Various metal-sulfur-oxygen stability
diagrams at different temperatures can be found
in Ref 15.
Stability diagrams allow one to predict the
phases that are likely to form on pure metals.
Log pSO /pSO

3
2
Log pCO /pCO
26 24
22 20 18
18 16 14
16 14
12 10 8
12 10
10
12
16 14 12 10
10
12
Log pH O /pH
2
2
18
S(l )
0
Fe2(SO4)3(s)
FeS1+x (s)
0
2
FeSO4 (s)
15
20
Fe3O4 (s)
10
Fe(s)
25
Fe0.950(s)
12
30
14
Fe2O3(s)
16
35
18
50
45
40
35
30
25
20
15
Log pO , atm
2
Fig. 7.4
Stability diagram of the Fe-S-O system at 870 C (1600 F). Source: Ref 15
10
Log pH S/pH
2
2
Log pS , atm
10
26/10/2007 11:13AM Plate # 0
pg 205
Chapter 7:
Log pCO /pCO

2
26 24
22 20 18
18 16 14
Log pH O /pH
2
2
18
16 14
12 10 8
12 10
10
12
16 14 12 10
10
12
S(l )
NiSy (l)
NiSO4 (s)
4
6
15
Log pS , atm
10
20
10
Ni(s)
25
Log pH S/pH
2
2
Log pSO /pSO

3
2
Sulfidation / 205
12
NiO(s)
30
14
35
16
18
50
45
40
35
30
25
20
15
10
Log pO , atm
2
Fig. 7.5
Stability diagram of the Ni-S-O system at 870 C (1600 F). Source: Ref 15
26 24
Log pSO /pSO

3
22 20 18
16 14
12 10 8
Log pCO /pCO

2
18 16 14
Log pH O /pH
2
2
18
12 10
10
12
16 14 12 10
10
12
S(l )
CoS1+x (s)
5
10
15
20
10
CoO(s)
Co(s)
25
12
30
14
Co3O4 (s)
35
16
18
50
45
40
35
30
25
20
15
10
Log pO , atm
2
Fig. 7.6
Stability diagram of the Co-S-O system at 870 C (1600 F). Source: Ref 15
Log pH S/pH
2
2
Log pS , atm
CoSO4 (s)
Co4S3+x (s)
26/10/2007 11:13AM Plate # 0
pg 206
chromium oxide scale for protection against

suldation. Most industrial environments have
sufcient oxygen activities to form a chromium
oxide scale. The scale may eventually break
points: 635 C (1175 F) for Ni-Ni3S2, 880 C

(1616 F) for Co-Co4S3, and 985 C (1805 F)
for Fe-FeS (Ref 17). Most high-temperature
alloys are chromia formers, which rely on
26 24
Log pSO /pSO

3
22 20 18
16 14
12 10 8
Log pCO /pCO

2
18 16 14
Log pH O /pH
2
2
18
12 10
10
12
16 14 12 10
10
12
S(l )
2
0
CrS (s)
2
15
Cr2O3 (s)
20
Log pH S/pH
2
2
Log pS , atm
10
10
25
Cr (s)
12
30
14
35
16
18
50
45
40
35
30
25
20
15
10
Log pO , atm
2
Fig. 7.7
Stability diagram of the Cr-S-O system at 870 C (1600 F). Source: Ref 15
0
Cr2 S3
Cr2 O3
FeS
NiS
FeO
Log pS
NiO
8
Fe
Ni
NiO
CrS
FeO
12
Fe
Cr
Ni
Cr2 O3
16
36
32
28
24
16
20
12
Log pO
Fig. 7.8
Stability diagram for the M-S-O systems at 870 C (1600 F), where M stands for metal. Source: Ref 16
26/10/2007 11:13AM Plate # 0
Chapter 7:
down, leading to breakaway corrosion. Perkins

(Ref 14) referred to this region as limited
life; see the upper right region of the Cr-S-O
diagram shown schematically in Fig. 7.9. The
environments within this region typically contain
H2, H2O, CO, CO2, H2S, and so forth, referred
to in this chapter as reducing, mixed-gas
environments, which are characterized by low
oxygen and high sulfur potentials. Another type
of suldizing environment with common characteristics for the suldation of various alloys is
the upper left region in Fig. 7.9. In this region,
suldes are stable phases. Alloys in these environments form sulde scales. The environments
include sulfur vapor and H2/H2S mixtures with
extremely low oxygen activities such that Cr2O3
is not thermodynamically stable.
7.3 Sulfidation in Sulfur Vapor,

Hydrocarbon Streams (No H2),
H2S, and H2-H2S Mixtures
Sulfur vapor, hydrocarbon streams with no

hydrogen (H2), H2S, and H2-H2S mixtures have
a common feature in suldation reactions. This
common feature is that the corrosion products in
these environments are suldes. The data presented in this section are not applicable to other
environments where oxides and suldes coexist.
Suldation behavior of metals and alloys in
sulfur vapor environments has been studied generally at sulfur pressures higher than 103 atm.
Studies of H2-H2S mixtures have been typically
carried out at sulfur partial pressures less than
102 atm. Mrowec and Przybylski (Ref 18) and
Young (Ref 19) gave excellent reviews on suldation of metals and alloys in sulfur vapor and
Fig. 7.9
Stability diagram for the Cr-S-O system. Source:

Ref 14
pg 207
Sulfidation / 207
H2-H2S environments. Suldation of pure

metals, such as iron, nickel, and cobalt, and iron-,
nickel-, and cobalt-base alloys was found to
proceed with the formation of suldes (Ref 18,
19) in these environments. High sulfur potentials
combined with extremely low oxygen potentials
have resulted in formation of suldes. Thus, data
generated in these types of environments are
applicable only to industrial environments where
suldes, but not oxides, are thermodynamically
stable.
Various investigations of suldation of
iron (247977 C, or 4751790 F), nickel
(400640 C, or 7501185 F), and cobalt (497
1000 C, 9251830 F) in sulfur vapor and
H2-H2S mixtures were summarized by Mrowec
and Przybylski (Ref 18). Suldation of these
metals was found to follow a parabolic rate law,
with sulde scales forming on metal surfaces in
all cases. Chromium also suldizes at a parabolic
rate in sulfur vapor and H2-H2S mixtures with
sulfur potentials of 107 to 1.0 atm at 700 to
1000 C (1290 to 1830 F), with only sulde
scales forming on metal surfaces (Ref 18).
Data for Fe-Cr, Ni-Cr, and Co-Cr alloys were
generated by several studies (Ref 2026).
Mrowec and Przybylski (Ref 18) summarized
these data in Fig. 7.10, which shows that at sulfur
potentials greater than 102 atm, all three alloy
systems follow a similar trend of decreasing
suldation rate with increasing chromium content. Suldation rates were expressed in terms
of parabolic rate constants since all three alloy
systems were found to follow a parabolic rate
law. It is surprising to nd that there is no signicant difference in suldation resistance
among the three alloy systems. It appears that the
binary alloys with less than 15% Cr showed more
scattering. Improved suldation resistance for
alloys with less than 40% Cr (at.%) was attributed to the formation of an inner sulde layer:
Fe(Fe2xCrx)S4 for Fe-Cr alloys, chromium suldes with nickel for Ni-Cr alloys, and chromium
suldes with cobalt for Co-Cr alloys (Ref 18). In
alloys with higher chromium (40% at.), a single
layer of chromium sulde (Cr2S3) was observed
(Ref 18). Davin and Coutsouradis (Ref 20)
investigated the relative suldation resistance
of Fe-Cr, Ni-Cr, and Co-Cr alloys with 20 and
35% Cr in H2S at 800 C (1470 F). For 20% Cr
alloys, the Co-Cr alloy was most resistant, followed by the Fe-Cr alloy. The Ni-Cr alloy was
least resistant. When chromium was increased to
35%, however, all three alloys showed similar
resistance.
26/10/2007 11:13AM Plate # 0
Aluminum is a benecial alloying element

in Fe-Cr and Co-Cr systems in both sulfur vapor
and H2-H2S environments. The effect of aluminum on the parabolic rate constant of the
Fe-Cr-Al system was summarized in Fig. 7.11 by

Mrowec and Przybylski (Ref 18) using data
generated by various studies (Ref 2730). The
improved suldation resistance of Fe-Cr alloys
p
p
Fig. 7.10
Parabolic rate constants for Fe-Cr, Ni-Cr, and Co-Cr alloys as a function of chromium content. Source: Ref 18
106
800 C (1470 F)
--Fe-19 Cr-AI, MROWEC,WEDRYCHOWSKA

(1979)
--Fe-17 Cr-AI, JALLOULI et al. (1979)
--Co-25 Cr-AI, BIEGUN, BRCKMAN
(1980)
--Fe-25 Cr-AI, SMELTZER et al. (1982)
pS = 1 atm
2
" g2cm4 s1
kp,
107
108
pS = 1 atm
2
pS = 8.105atm
2
109
pS = 107atm
Fe-Cr-AI
Co-Cr-AI
10
20
30
% at AI
Fig. 7.11
Parabolic rate constants for Fe-Cr-Al and Co-Cr-Al alloys as a function of aluminum content. Source: Ref 18
pg 208
26/10/2007 11:13AM Plate # 0
Chapter 7:
due to aluminum additions was attributed to the

formation of an inner sulde layer of presumably
Fe(FexAlyCrz-x-y)S4 (Ref 27). This sulde
decomposed at room temperature (Ref 27). The
outer layer consisted of Fe1yS (Ref 27). Biegun
and Bruckman (Ref 29) reported a similar benecial effect in Co-25Cr alloys when aluminum
content was varied 1 to 22% (at.%) at temperatures of 700 to 1100 C (1290 to 2010 F) in
sulfur vapor at 1 atm. However, the inuence of
aluminum in the Co-Cr-Al system was not as
effective as in the Fe-Cr-Al system (Fig. 7.11).
The Ni-Cr-Al system behaved differently from
both the Fe-Cr-Al and Co-Cr-Al systems
(Ref 31). Biegun and Bruckman (Ref 31) investigated Ni-25Cr alloys containing 1 to 10 at.%
aluminum in sulfur vapor (1 atm) at 580 to
950 C (1080 to 1740 F) with erratic results.
The kinetics rates observed ranged from parabolic to linear.
Corrosion rates of several austenitic stainless
steels in sulfur vapor at 571 C (1060 F) are
shown in Table 7.1 (Ref 32). Higher-chromium
alloys, such as 314SS (24Cr, 2Si), 310SS (25Cr),
and 309SS (23Cr), were found to perform the
best. Silicon may not be very effective at this
relatively low temperature. In addition to Type
314SS, which contains about 2% Si, there was
another silicon-containing stainless steel, Type
302B (18Cr, 2.5Si). Type 302B performed quite
similarly to Type 304SS (19Cr, no Si). For
comparison, the corrosion rates of austenitic
Table 7.1 Corrosion rates of several

austenitic stainless steels in sulfur vapor at
571 C (1060 F) for 1295 h
Alloy
Corrosion rates,
mm/yr (mpy)
314
310
309
304
302B
316
321
0.43 (16.9)
0.48 (18.9)
0.57 (22.3)
0.69 (27.0)
0.76 (29.8)
0.79 (31.1)
1.39 (54.8)
Source: Ref 32
Table 7.2
Alloy
Type 310
Type 304
Type 316
Source: Ref 32
pg 209
Sulfidation / 209
stainless steels in H2-50%H2S at 500 to 700 C

(932 to 1292 F) is shown in Table 7.2 (Ref 32).
Corrosion becomes much more aggressive in
100% H2S. Malinowski et al. (Ref 33) reported
corrosion rates of iron and iron-aluminum alloys
in 100 H2S at 400, 600, and 700 C (752, 1112,
and 1292 F) (Fig. 7.12). Although aluminum
was found to be benecial in corrosion rates of
Fe-Al system, the content of aluminum that was
needed to signicantly reduce the alloys corrosion rates was found to be about 10% at 600 and
700 C (1112 and 1292 F), and slightly lower at
400 C (752 F). Bruns (Ref 34) tested many
commercial alloys in 100% H2S at 593 C
(1100 F). Also included in the tests were
experimental alloys based on Fe-32Ni-20Cr with
various levels of aluminum (Ref 34). His test
results are summarized in Fig. 7.13 in terms of
the percent of either chromium or aluminum
content in the alloy (Ref 34). The data generated
in 100% H2S by Bruns indicate that increasing
chromium signicantly improves the alloys
suldation resistance. It also indicated that adding 8.5% Mn to Fe-18Cr alloy failed to improve
the alloys corrosion resistance. For Fe-32Ni20Cr alloy (the base composition is similar to
alloy 800/800H) with various aluminum contents, Fig. 7.13 shows strong benecial effect for
aluminum in increasing the alloys suldation
resistance. Furthermore, the aluminum content
that signicantly increased the suldation resistance of Fe-32Ni-20Cr alloy was found to be in a
range of only 2 to 4%.
In reneries, suldation, which is commonly
referred to as suldic corrosion in renery
industry, is a common materials problem. The
temperatures of the renery equipment having
suldation problems are typically between about
260 and 540 C (500 and 1000 F). Sulfur
compounds originating from crude oils include
polysuldes, hydrogen sulde, mercaptans, aliphatic suldes, disuldes, and so forth (Ref 35).
Suldation occurs in some processing units
where no hydrogen is present in the system, such
as crude distillation units. The crude distillation
units that process mostly sweet crude oils (less
Corrosion rates of Types 310, 304, and 316 in H2-50%H2S at 500700 C (9321292 F)
Corrosion rate 500 C
(932 F), mm/yr (mpy)

(1022 F), mm/yr (mpy)

(1112 F), mm/yr (mpy)

(1292 F), mm/yr (mpy)
0.91 (36)
1.12 (44)
1.50 (59)
1.45 (57)
1.57 (62)
2.44 (96)
2.79 (110)
2.95 (116)
4.45 (175)
8.94 (352)
10.2 (400)
10.8 (424)
26/10/2007 11:13AM Plate # 0
than 0.6% total sulfur, with essentially no

hydrogen sulde) experience relatively less suldation problems. More suldation problems are
encountered in the distillation units that process
mostly sour crudes (Ref 36). The main fractionation tower is usually made of carbon steel
with the lower part lined with a stainless steel
containing about 12% Cr, such as Type 405
(Ref 37). Type 309L weld overlay cladding is
most commonly used for replacing the ferritic
stainless steel lining or cladding that suffered
suldation attack. The corrosion rates of the
crudes and their liquid fractions can be predicted
by the so-called modied McConomy curves.
Figure 7.14 shows the modied McConomy
curves for liquid hydrocarbon streams containing
0.6 wt.% S (Ref 38). For hydrocarbon streams
containing more than 0.6% sulfur, the corrosion
rate multiplier as shown in Fig. 7.15 can be used
in conjunction with the data shown in Fig. 7.14
for making the prediction of the corrosion rate
for various steels at different temperatures
(Ref 38).
Hydrogen in hydrotreating, hydrocracking,
and hydrodesulfurizing processes is used to
remove sulfur (to convert it to hydrogen sulde)
and nitrogen (to ammonia) for separation from
Fig. 7.12
Corrosion rates of iron and iron-aluminum alloys in

100% H2S at 400, 600, and 700 C (752, 1112, and
the hydrocarbon streams (Ref 39). Hydrocracking processes combine desulfurization and
cracking operation that can convert hydrocarbon
feedstocks into various products (Ref 39). Suldation in these processing units is dictated by
the H2S concentration in the H2-H2S environment. Suldation in H2-H2S environments is
severe since the corrosion products are suldes
with no protection by oxide scales in these
environments. Austenitic stainless steels are
generally used as a cladding or weld overlay for
suldation resistance. Because of high hydrogen
temperatures and pressures in the system, the
vessels generally are made of low-alloy steels for
resisting hydrogen attack, and the selection of
materials to avoid hydrogen attack problem
should follow the recommendations of the API
Publication 941 (Ref 40). Materials issues related
to hydrogen attack are discussed in Chapter 17:
Hydrogen Attack. The reactors in hydrotreating
and hydrocracking units are normally made of
low-alloy steels with Type 347 cladding or weld
overlay (Ref 41). The use of Type 347 (a stabilized grade of austenitic stainless steel) for
cladding or weld overlay is to avoid intergranular
stress-corrosion cracking by polythionic acid
during downtime (Ref 39, 41).
Corrosion rates of metals and alloys in H2-H2S
environments are dependent on the H2S concentration. Corrosion of austenitic stainless steels
was very aggressive in H2-50%H2S and 100%
H2S environments, as shown in Table 7.2 and
Fig. 7.13. However, the concentrations
of hydrogen sulde (H2S) in H2-H2S environments in hydrotreating, hydrocracking, hydrodesulfurizing, and catalytic reforming processes
are signicantly lower, thus making carbon
steels, low-alloy steels and stainless steels capable of resisting suldation attack in different
temperature ranges. For example, Neumair and
Schillmoller (Ref 42) reported the corrosion rates
of steels and austenitic stainless steels as a
function of hydrogen sulde concentration for
a hydrodesulfurization equipment at 415 C
(780 F) and 27.2 atm (400 psig) (Fig. 7.16) and
for a catalytic-reformer equipment at 510 C
(950 F) and 27.2 atm (400 psig) (Fig. 7.17).
The corrosion data in catalytic reforming units
generated by different renery companies was
summarized by Sorell (Ref 43) and are shown in
Table 7.3 to illustrate the suldation of various
steels in H2-H2S mixtures. Catalytic reforming is
used in petroleum reneries to upgrade the
octane number of gasoline (Ref 44). Severe
corrosion attack on processing equipment by
pg 210
26/10/2007 11:13AM Plate # 0
Chapter 7:
hydrogen sulde has been encountered in several

catalytic reforming and desulfurizing units
(Ref 43, 45, 46). In Table 7.3, Sorell (Ref 43)
reported hydrogen sulde corrosion data generated by various renery operators in laboratory
tests, pilot plant testing, eld testing in commercial units, and inspection of commercial
operating equipment. Austenitic stainless steels
(18Cr-8Ni) were most resistant, followed by
straight chromium stainless steels (1216% Cr),
with low-chromium steels (09% Cr) being
worst.
Fig. 7.13
pg 211
Sulfidation / 211
Sorell and Hoyt (Ref 45) reported the suldation resistance of Fe-Cr and Fe-Cr-Ni alloys
in H2-16H2S at 315 to 480 C (600 to 900 F)
(Fig. 7.18). Slight to moderate addition of nickel
(e.g., 8 or 40%) to Fe-20Cr improved the alloy
suldation resistance. Greater nickel addition
(e.g., 65% or more) reduced suldation resistance. Figure 7.18 also surprisingly revealed that
Fe-20Cr-65Ni exhibited a suldation resistance
similar to that of Fe-20Cr (Ref 45). Suldation of
alloys in H2-H2S mixtures has been described by
isocorrosion rate curves, which show corrosion
Corrosion rates of commercial alloys as well as experimental Fe-30Ni-20Cr with various aluminum contents in 100% H2S at
593 C (1100 F). Note the upper curve shows the corrosion rates as a function of chromium content in the alloy, while the
lower curve show the corrosion rates as function of Al content in the alloy. Note: alloy 202 (Fe-18Cr-8.5Mn), CL (Fe-32Ni-20Cr-0.25C), CI
(Fe-32Ni-21Cr-0.025C), 804 (Fe-41Ni-29Cr-0.97Ti), 2.24% Si alloy (Fe-32Ni-21Cr-2.24Si). Source: Ref 34
26/10/2007 11:13AM Plate # 0
rate as a function of H2S concentration and

temperature. Backensto and Sjoberg (Ref 48)
determined isocorrosion rate curves for chromium steels (0 to 5% Cr) and austenitic stainless
steels, shown in Fig. 7.19 and 7.20, respectively.
Couper-Gorman curves are generally
believed to provide the most practical correlation
among corrosion rates, temperature and
concentration of H2S for various steels (Ref 38).

These curves were based on a survey conducted
by NACE Committee T-8 on Rening Industry
Corrosion (Ref 49). The modied Couper-Gorman curves (i.e., the original curves were
extended to higher concentrations of H2S with
dashed lines) for 5Cr-0.5Mo steels (also applicable to carbon steels) and 18Cr-8Ni steel are
shown in Fig. 7.21 and 7.22, respectively. It was
found that total pressure between 1 and 18 MPa
(150 and 2650 psig) was not a signicant variable. The authors also observed that no suldation occurred at very low H2S concentrations and
at temperatures above 315 C (600 F) because
formation of iron suldes in that regime is not
feasible thermodynamically.
7.4 Sulfidation in Environments with

Low-Oxygen and High-Sulfur Potentials
Fig. 7.14
Modified McConomy curves providing predicted

corrosion rates as a function of temperatures for
various steels in a liquid hydrocarbon stream containing 0.6% S.
Source: Ref 38
Fig. 7.15
Effect of sulfur content in hydrocarbon streams

on corrosion rates predicted by the modified
McConomy curves in the 290400 C (550750 F) range.
Source: Ref 38
The environments covered in this section

typically contain H2, CO, CO2, H2O, H2S, and
other gaseous components with oxygen and
sulfur potentials high enough to form oxides and
suldes for most high-temperature alloys. These
types of reducing environments are generated by
combustion under substoichiometric conditions
with insufcient air or oxygen. The corrosion
reactions in these environments generally involve oxidation and suldation. Suldation is
usually the determining factor in the alloys
performance capability. In most cases, the alloy
remains protected during the oxidation period
until the initiation of breakaway corrosion, which
is then followed by rapid suldation attack by
forming primarily suldes.
Suldation behavior of alloys in this type of
environments, where oxidation and suldation
reactions are involved, has been extensively
studied during the 1970s, 1980s, and early 1990s
when coal gasication programs had been
actively pursued. Coal gasication atmospheres
typically have sulfur potentials (pS2 ) of 105 to
1010 atm and oxygen potentials (pO2 ) of 1015 to
1020 atm (Ref 50). Signicant understanding of
corrosion reactions in these environments has
been achieved (Ref 5167). In addition, a large
engineering database on commercial alloys has
been generated. Major contributions to this
database came from Metal Properties Council
(MPC) programs from 1972 to 1985 that evaluated more than 80 commercial alloys and
coatings for coal gasication. The tests were
pg 212
26/10/2007 11:14AM Plate # 0
Chapter 7:
carried out mostly by IIT Research Institute, and

results were documented in MPC annual reports
(Ref 5159). All these results were summarized
pg 213
Sulfidation / 213
by Howes (Ref 60) and Verma (Ref 63). The

test gas mixtures used are tabulated in Table 7.4,
and their thermodynamic potentials (pS2 and
Fig. 7.16
Maximum anticipated corrosion rates for alloys in hydrodesulfurization equipment at 415 C (780 F) and 27.2 atm
(400 psig). Source: Ref 42
Fig. 7.17
Maximum anticipated corrosion rates for alloys in catalytic reforming equipment at 510 C (950 F) and 27.2 (400 psig).
Source: Ref 42
26/10/2007 11:14AM Plate # 0
pg 214
Table 7.3 Summary of corrosion data in H2-H2S mixtures typical of catalytic reforming generated
by various petroleum refining companies
Type of
corrosion
data(a)
Source
Temperature,
C (F)
Test
duration,
days
Hydrogen
pressure,
atm (psi)
H2S
concentration,
vol%
Materials
Corrosion rate,
mm/yr (mpy)
1.5 (59)
1.5 (59)
0.76 (30)
0.33 (13)
0.12.5 (5100)
0.031.5 (160)
0.030.2 (18)
1.0 (38)
0.18 (7)
1.0 (40)
0.4 (16)
0.2 (8)
0.06 (2.5)
1.910 (73398)
0.11.1 (542)
0.0310 (1400)
0.011.5 (0.460)
2.3 (90)
5.1 (200)
0.31.0 (13.541.5)
American Oil
I
C
510 (950)
510550 (9501025)
3689
89
1827 (265400)
1827 (265400)
0.030.07
0.030.07
Atlantic
L, P, C, I
480 (900)
1365
34 (500)
480 (900)
180
34 (500)
D-X Sunray
Canadian Petrona
I
C
465 (870)
480490 (890920)
180
90
27 (400)
2122 (310330)
0.0130.074
0.0110.13
0.0150.27
0.036
0.016
0.04
0.008
Humble
480510 (900956)
14
20 (300)
0.0130.14
540 (1000)
20 (300)
0.0070.15
450520 (850975)
480550 (9001025)
470520 (875960)
127
20 (300)
0.035
139577
3436 (500535)
0.0350.09
1Cr, 2Cr
09Cr
12Cr
18Cr8Ni
05Cr
1113Cr
18Cr8Ni
2Cr
12Cr
5Cr
09Cr
12Cr
18Cr8Ni
012Cr
18Cr8Ni
012Cr
18Cr8Ni
1Cr
CS, 2Cr
2Cr, 5Cr
0.015
12Cr
18Cr8Ni
05Cr
0.280.7 (1127)
0.1 (3.7)
0.5 (18.9)
12Cr
18Cr8Ni
2Cr
09Cr
12Cr
18Cr8Ni
09Cr
12Cr
5Cr
9Cr
09Cr
12Cr
18Cr8Ni
07Cr
13Cr
18Cr8Ni
07Cr
13Cr
18Cr8Ni
5Cr
12Cr
05Cr
716Cr
CrNi
5Cr
09Cr
1216Cr
CrNi
09Cr
12Cr
Cr-Ni
09Cr
12Cr
18Cr8Ni
CS
0.2 (8.3)
0.1 (4.5)
0.8 (33)
0.47.9 (15310)
0.36.1 (10240)
0.10.5 (3.520)
0.5 (18)
0.2 (8)
1.43.4 (56135)
1.7 (68)
0.91.2 (3547)
0.40.6 (1522)
0.10.2 (4.37.2)
2.011 (80450)
0.55.3 (20210)
0.11.2 (348)
0.10.3 (2.410)
0.050.2 (1.96.5)
0.010.02 (0.550.6)
0.2 (90).
0.2 (6.5)
0.037.2 (1285)
0.032.6 (1102)
0.030.2 (18.7)
1.1 (44)
2.7 (108)
0.5 (21)
0.1 (5)
0.51.6 (1862)
0.30.6 (1323)
0.020.1 (0.74.3)
0.313 (10500)
0.15 (4200)
0.031.6 (1.362.5)
1.8 (70)
1.1 (45)
1.0 (37.5)
0.1 (3.3)
0.20.8 (6.129.7)
0.10.3 (5.811.7)
0.020.05 (0.71.9)
0.07 (2.9)
0.05 (1.9)
0.01 (0.3)
Major U.S.
Ref. Comp.
Major West Coast
Ref. Comp.
480490 (900920)
72
3738 (540565)
Pure Oil
Richeld
I
L, P, C, I
480540 (9001000)
510 (950)
240
16.7
34 (500)
27 (400)
Shell Oil
510 (950)
71
43 (625)
0.026
0.0111.0
0.0141.0
0.0251.0
0.015
I
C
510 (950)
510610 (9551135)
500 (925)
148758
758
148251
46 (680)
46 (670)
46 (680)
0.050.08
0.08
0.050.08
510 (950)
4.2
34 (500)
490500 (910940)
212382
34 (500)
0.081.0
0.11.0
0.11.0
0.0010.0089
120270
34 (500)
0.005
480550 (9001025)
480510 (900950)
530 (985)
6.320.8
33 (485)
L
C
490 (905)
470500 (885935)
21
55
31 (460)
12 (175)
0.0090.2
0.0170.2
0.060.2
0.085
0.5
470490 (885920)
50.5213
3739 (550575)
0.0180.04
480 (900)
0.711.7
19 (285)
0.0130.44
16.546
20 (300)
510 (950)
510540 (9501000)
550 (1030)
47
23 (340)
0.023
0.032
0.029
Sun Oil
500520 (925960)
188395
2041 (300600)
0.0170.04
Texas
500 (940)
133668
41 (600)
0.0016
Sinclair
Socony Mobil
Standard Oil (Ind.)
09Cr
18Cr8Ni
05Cr
12 Cr
18Cr8Ni
05Cr
12Cr
18Cr8Ni
(a) L, laboratory corrosion tests; P, pilot plant corrosion tests; C, commercial unit corrosion tests; I, inspection of commercial operating equipment. Source: Ref 43
26/10/2007 11:14AM Plate # 0
Chapter 7:
pO2 ) at various test temperatures and pressures

are shown in Table 7.5. The test environments
in terms of pO2 and pS2 in the phase-stability
diagrams for two test temperatures650 and
980 C (1200 and 1800 F)are shown in
Fig. 7.23 and 7.24 (Ref 63).
Most alloys suffered breakaway corrosion in
several thousands of hours or less. Breakaway
corrosion is illustrated in Fig. 7.25 (Ref 60). Both
Type 310SS and alloy 800H followed a parabolic
reaction rate prior to rapid corrosion attack.
However, a few alloys did not exhibit breakaway
corrosion within the test duration of up to
10,000 h, including the Co-Cr-W alloy 6B as
shown in Fig. 7.25. Verma (Ref 63) discussed in
detail the performance of various alloys in the
MPC environments. Chromium was the most
important alloying element in resisting suldation attack.
Most alloys were protected by chromium
oxide scales. The scale may eventually break
down, leading to breakaway corrosion
(Fig. 7.26) (Ref 63). Breakdown of the protective
chromium oxide scale results from the suldes
formed over it (Ref 14, 61). Formation of liquid
Fig. 7.18
Corrosion rates of Fe-Cr and Fe-Cr-Ni alloys in H216H2S at 315 to 480 C (600 to 900 F) and 75 atm
(1100 psig) pressure. Source: Ref 45, summarizing results of
Ref 47
pg 215
Sulfidation / 215
sulde slag over the chromium oxide scale is

most damaging. Alloying elements, such as
manganese, iron, cobalt, nickel, and so forth,
diffuse through the chromium oxide scale and
react with the environment on top of the oxide
scale to form external suldes. This was proposed by Perkins (Ref 14) as a possible mechanism for breakaway corrosion, schematically
illustrated in Fig. 7.27. Manganese is the fastest
diffusing element, followed by iron, cobalt,
nickel, and chromium (Ref 14). In addition to
outward diffusion of alloying elements to form
external suldes, the corrosion reaction also
involves sulfur penetrating through the oxide
scale to form discrete particles of suldes in the
matrix (Ref 14). Natesan (Ref 62) showed that
for a given sulfur potential, there exists a
threshold value for oxygen potential beyond
which a continuous protective oxide scale is
developed. This threshold oxygen partial pressure (kinetic boundary) is about 103 times the
oxygen partial pressure for chromium oxide and
Fig. 7.19
Corrosion rates of chromium steels (05% Cr)

generated from laboratory tests in H2-H2S at
hydrogen pressures of 12 to 34 atm (175 to 500 psig) as a function
of H2S concentration and temperature. IPY, inch per year. Source:
Ref 48
26/10/2007 11:14AM Plate # 0
chromium sulde equilibrium (thermodynamic

boundary) (Ref 62).
The suldation resistance of various commercial alloys generated in the MPC programs
is summarized in Fig. 7.28 to 7.31 (Ref 60) and
Fig. 7.32 and 7.33 (Ref 67). For nickel-base
alloys, increasing nickel generally increases
susceptibility to suldation attack (Fig. 7.28).
High-nickel alloys, such as alloys 600 and 601,
are particularly susceptible to suldation attack,
as illustrated in Fig. 7.32 (Ref 67). The Ni-Ni3S2
eutectic melts at 635 C (1175 F). Molten sulde slag can easily destroy the chromium oxide
scale and cause catastrophic suldation attack.
Increasing chromium generally improves suldation resistance in Fe-Cr-Ni alloys (Fig. 7.29)
(Ref 60). This also holds true for iron- and nickelbase alloys. Type 446SS (Fe-27Cr) and 50Ni50Cr alloys (alloys 671 and 657) are good
examples (Fig. 7.30) (Ref 60). Despite its high
nickel content, the 50Ni-50Cr alloy exhibited
good suldation resistance due to its extremely
high chromium content.
In the MPC data summarized in Figs. 7.28 to

7.31, many stainless steels and nickel alloys,
while showing very high wastage rates at high
temperatures (e.g., 816 C and higher), showed
low corrosion rates at 482 and 650 C (900 and
1200 F). Review of the test data generated
at 650 C (1200 F) for 10,000 h in the MPC test
programs summarized in Howess report
(Ref 60) showed that alloys containing at least
18% Cr, which included 304, 309, 310, 312, 329,
22-13-5, 29-4-2, Ebrite 26-1, 446, 20Cb-3,
800, 253MA, Crump 25, 825, RA333, 617,
and Multimet, exhibited corrosion rates of
0.02 mm/yr or less (1 mpy or less) when tested in
a 500 psig gas pressure and 1% H2S in the MPC
coal gasication environment. Type 410 (12Cr)
showed the corrosion rate of 0.05 mm/yr
(2 mpy), and 9Cr-1Mo steel suffered the corrosion rate of 1.01 mm/yr (40 mpy) (Ref 60).
John et al. (Ref 68) reported the corrosion data
generated at 700 C (1292 F) and lower (1 atm
pressure) for various commercial alloys in syngas environments that were rich in CO and H2
with low water. The corrosion data for carbon
Fig. 7.20
Fig. 7.21
Corrosion rates of Cr-Ni austenitic stainless steels

generated from laboratory tests in H2-H2S at
hydrogen pressures of 12 to 34 atm (175 to 500 psig) as a function
of H2S concentration and temperature. IPY, inch per year. Source:
Ref 48
Modified Couper-Gorman curves showing corrosion rates as a function of H2S concentration (mol.
%) and temperatures for 5Cr-0.5Mo steel. The data also is
applicable to carbon steels and alloy steels with less than 5% Cr
(naphtha desulfurizers). Source: Ref 38
pg 216
26/10/2007 11:14AM Plate # 0
pg 217
Chapter 7:
steel and 1Cr-0.5Mo steel tested in CO-51.2H20.8CO2-1.7H2S-0.2HCl at 330, 400, and 500 C
(626, 752, and 932 F) are summarized in
Fig. 7.34 (Ref 68). The corrosion behavior of
alloy 625 in the same test environment at 330
to 700 C (626 to 1292 F) is summarized in
Fig. 7.35 (Ref 68). The data for various Fe-, Ni-,
and Co-base alloys at 700 C (1292 F) and
lower in CO-26H2-1.0CO2-0.9 H2S-2H2O-10N2
is summarized in Fig. 7.36 (Ref 68). It was surprising that Co-base alloys, such as alloys 25
Sulfidation / 217
Table 7.5 Oxygen and sulfur potentials ( pO2 and

pS2 ) in the MPC-CGA test gas mixtures at 68 atm
(1000 psig)
Temperature,
C (F)
H2S in test gas,

vol%
pO2 (atm)
pS2 (atm)
0.1
0.1
0.1
0.1
0.1
0.5
0.5
0.5
0.5
0.5
1.0
1.0
1.0
1.0
1.0
1.2 1023
7.0 1021
3.0 1018
6.4 1017
1.3 1015
1.2 1023
7.0 1021
3.0 1018
6.4 1017
1.3 1015
1.2 1023
7.0 1021
3.0 1018
6.4 1017
1.3 1015
1.5 1010
8.8 1010
5.4 109
1.3 108
3.1 108
3.5 109
2.2 108
1.4 107
3.1 107
7.6 107
1.5 108
8.8 109
5.5 107
1.3 106
3.1 106
480 (900)
650 (1200)
816 (1500)
900 (1650)
980 (1800)
480 (900)
650 (1200)
816 (1500)
900 (1650)
980 (1800)
480 (900)
650 (1200)
816 (1500)
900 (1650)
980 (1800)
Source: Ref 60
NiO
Co/CoO
Fig. 7.23
Fig. 7.22 Modified Couper-Gorman curves showing corrosion rates as a function of H2S concentration (mol.
%) and temperatures for 18Cr-8Ni steel. Source: Ref 38
Table 7.4
mixtures
The phase-stability diagram and the test environment (star) in terms of equilibrium pO2 and pS2 in
the MPC coal gasification test programs at 650 C (1200 F).
Source: Ref 63
Inlet and equilibrium gas compositions at different temperatures for the MPC-CGA test gas
Equilibrium gas composition(a), vol%
Gaseous component
H2
CO
CO2
CH4
NH3(b)
H2S
H2O
Inlet gas, vol%
480 C (900 F)
650 C (1200 F)
816 C (1500 F)
900 C (1650 F)
980 C (1800 F)
24
18
12
5
1
01.5
bal
4
5
25
19
1
0.5
bal
11
8
22
14
1
1.5
bal
23
11
19
9
1
0.11.5
bal
27
14
17
6
1
0.51.0
bal
31
17
15
3
1
01.0
bal
(a) Computer calculations at 6.9 MPa (1000 psig). (b) NH3 will decompose to N2 and H2 on increasing temperature. Source: Ref 60
26/10/2007 11:14AM Plate # 0
(Co-20Cr-10Ni-115W) and 150 (Co-27Cr18Fe), were not signicantly better than Type
310, and but not as good as Type 446. The suldation behavior of Co-base alloys at higher
temperatures is discussed later.
Bakker (Ref 69) and Bakker and Bonvallet
(Ref 70) investigated the effect of H2O concentration on the corrosion behavior of Type 310
and alloy 800 in coal gasication environments
at 540 C (1004 F) for 600 h. The test gas
environments with different levels of H2O from
0 to 15% along with oxygen partial pressures
(pO2 ) at 540 C (1004 F) are tabulated in
Table 7.6. The test results are summarized in
Fig. 7.37. Both oxides and suldes were observed to have formed on both alloys in all test
environments. Type 310 was found to suffer
signicantly less metal loss than alloy 800, and
its corrosion was relatively unaffected by H2O
concentrations. Alloy 800 suffered much higher
metal loss, with highest metal loss surprisingly at
about 3% H2O. Long-term tests are needed to
conrm this abnormal observation.
Cobalt-base alloys as well as cobaltcontaining alloys (e.g., Fe-Ni-Co-Cr) generally
have better suldation resistance than nickelbase alloys and Fe-Cr-Ni alloys (Fig. 7.33)
(Ref 67). Alloys 6B and 188 (cobalt-base alloys)
and alloy N155 (Fe-Ni-Co-Cr alloy) are better
than alloys RA333 and X (nickel-base alloys)
and alloys 800H, HK-40, and Type 310SS
(Fe-Ni-Cr alloys). In general, cobalt-base alloys

and cobalt-containing alloys have higher temperature capabilities and are more resistant to
breakaway corrosion. Results generated by Lai
(Ref 71) at 760, 870, and 980 C (1400, 1600,
and 1800 F) also revealed that cobalt-base
Fig. 7.25
Corrosion behavior of Type 310 stainless steel,

alloy 800, and alloy 6B at 980 C (1800 F) in the
MPC coal gasification atmosphere, showing breakaway corrosion
for Type 310SS and alloy 800. Inlet gas: 24H2-18CO-12CO25CH4-1NH3-0.5H2S (balance H2O) at 6.9 MPa (100 psig). pO2 =
1.3 1015 atm. pS2 = 7.6 107 atm. Source: Ref 60
Fe/FeS
5
Ni/NiS
Cr/CrS
40
35
30
25
20
FeO/Fe
Al2O3/Al
25
30
Cr2O3/Cr
20
15
Ni/NiO
Al/AlS
15
FeO/Fe3O4
Fe3O4/Fe2O3
Log pS , atm
10
10
Log pO , atm
2
Fig. 7.24
The phase stability diagram and the test environment (star) in terms of equilibrium pO2 and pS2 in the MPC coal gasification
test programs at 980 C (1800 F). Source: Ref 63
pg 218
26/10/2007 11:14AM Plate # 0
Chapter 7:
Fig. 7.26
pg 219
Sulfidation / 219
Corrosion behavior of alloy 800 in the MPC coal gasification atmosphere (0.5% H2S at 900 C or 1650 F, and 6.9 MPa
or 1000 psig) showing oxide scales during the protective stage and oxides/sulfides after breakaway corrosion. Source:
Ref 63
alloys were more suldation resistant than

nickel-base and Fe-Ni-Cr alloys with similar
chromium contents (Fig. 7.38).
Higher chromium (preferably more than 25%)
is needed for an alloy to form a protective
oxide scale in coal gasication environments.
Verma (Ref 63) suggested that chromium oxide
scale would be more protective if doped with
cobalt, as was the case in alloy 6B (Co-Cr-W)
that formed a (Cr,Co)2O3 scale. Bradshaw et al.
(Ref 65) believed that titanium modied the
Cr2O3 scale, thus improving the alloys suldation resistance. They compared 3% Ti-modied
Type 310SS with Type 310SS after exposure in
an MPC gas mixture with 1% H2S at about
1.0 atm pressure for 100 h at 980 C (1800 F).
Fig. 7.27
Formation of external sulfides on top of the chromium oxide scale and the formation of internal
sulfides. Source: Ref 14
26/10/2007 11:14AM Plate # 0
Fig. 7.28
Corrosion rates of high-nickel alloys in the MPC coal gasification atmosphere with 1.0 and 1.5% H2S (see Table 7.4 and 7.5
for gas composition). Source: Ref 60
330
Fig. 7.29
Corrosion of Fe-Cr-Ni alloys in the MPC coal gasification atmosphere with 1.0 and 1.5% H2S (see Tables 7.4 and 7.5 for gas
composition). Source: Ref 60
pg 220
26/10/2007 11:14AM Plate # 0
Chapter 7:
Type 310 sample showed some sulde nodules

as well as some spalled oxides, while Ti-modied Type 310 sample showed an adherent oxide
scale with no sulde nodules. Their analysis
indicated a signicant amount of titanium in the
Cr2O3 scale (Ref 65). They also tested alloys
800 and 801 in the same environment. Alloy
800 sample was totally corroded after 100 h at
980 C (1800 F), while the alloy 801 sample
showed an adherent oxide scale with only about
1.4 mg/cm2 weight gain. The only apparent difference in the chemical composition between the
two alloys is titanium (about 0.4% for alloy 800
and about 1.1% for alloy 801). Additional data
indicating the benecial effect of titanium are
shown in Table 7.7 (Ref 64). Results obtained by
Matl
446
302
304
309
310
314
316
800
793
600
601
671
310
(Al)
Sound metal loss, mils/yr
482
Temperature, C
650
816
900
800
(Al)
982
310
(Cr)
Completely
corroded
800
(Cr)
0.1
0.5
1.0
0.1
0.5
0.0
0.1
0.5
1.0
0.1
0.5
1.0
0.1
0.5
1.0
0.1
0.5
1.0
0.1
0.5
1.0
0.1
0.5
1.0
0.1
0.5
1.0
0.1
0.5
1.0
0.1
0.5
1.0
0.1
0.5
1.0
0.1
0.5
1.0
0.1
0.5
1.0
0.1
0.5
1.0
0.1
0.5
1.0
NT
Scaling loss
NT
0.5% H2S
NT
671
657
<2
446
671
657
900
1800
1650
1500
1200
Temperature, F
Internal penetration
Not tested
Welded specimens
Completely corroded
125
NT
125
75.1
Completely corroded
43.8
NT
NT
0
50
100
150
2050
220
Sulfidation / 221
% H2S
446
>50
pg 221
12
8
10
Total corrosion, mil
14
16
18
20
200 250 300 350

Total corrosion, m
400
450
500
Fig. 7.32
Corrosion of stainless steels and nickel-base alloys

at 816 C (1500 F) for 100 h in the MPC coal
gasification atmosphere with 0.1, 0.5, and 1.0% H2S. Also
included were aluminized Type 310 and alloy 800 [310 (Al) and
800 (Al], and chromized Type 310 and alloy 800 [310 (Cr) and
800 (Cr)] (see Tables 7.4 and 7.5 for gas composition). Source:
Ref 67
Fig. 7.30
Corrosion of Fe-27Cr (Type 446) and 50Ni-50Cr

alloys (alloy 671 and 657) in the MPC coal gasification stmosphere with 1.0 and 1.5% H2S (see Tables 7.4 and 7.5
for gas composition). Source: Ref 60
Alloy
310
HK-40
Cru-25
Sound metal loss, mils/yr
482
Temperature, C
650
816
900
N155
982
T63wC
Completely
corroded
188-Breakway occurs
in 200010,000 h
>50
Alloy X
2050
556
250
N155
6B
188
6B
6B
556 188
<2 N155
RV-18 & 19
900
Fig. 7.31
RA333
800
RV-18 & 19
1200
1500
1650
Temperature, F
671
1800
Corrosion of cobalt-base alloys (alloys 188 and 6B)

and cobalt-containing alloys (alloys 556, N155,
RV-18, and RV-19) in the MPC coal gasification atmosphere with
1.0 and 1.5% H2S (see Tables 7.4 and 7.5 for gas composition).
Source: Ref 60
Time,
h
1000
2000
1000
2000
5000
10,000
1000
2000
5000
10,000
1000
2000
5000
10,000
1000
2000
5000
10,000
1000
2000
5000
7278
1000
2000
5000
10,000
1000
2000
5000
10,000
1000
2000
5000
10,000
1000
2000
5000
7000
8000
1000
2000
5000
10,000
2.0
Legend
Scaling
Penetration
0.67
50
Fig. 7.33
1.5
0.9
8
10
12
14
Total corrosion, mil
16
18
20
100 150 200 250 300 350 400 450 500

Total corrosion, m
Corrosion of Fe-Cr-Ni, nickel-base, and cobalt-base

alloys at 980 C (1800 F) in the MPC coal gasification atmosphere with 0.5% H2S (see Tables 7.4 and 7.5 for gas
composition). Source: Ref 67
26/10/2007 11:14AM Plate # 0
pg 222
Lai (Ref 71) also revealed the benecial effect of

titanium in resisting suldation attack. As shown
in Fig. 7.38, alloy R-41 and Waspaloy alloy
(both contain about 3% Ti) were the best of the
nickel-base alloys tested. In fact, they approached some cobalt-base alloys. Alloy 263 (2.5%
Ti), while performing well at 760 C (1400 F),
suffered severe suldation attack at 870 and
980 C (1600 and 1800 F) (Ref 71).
Bradshaw et al. (Ref 64) also examined the
effects of molybdenum and aluminum. Addition
of 6% Mo to high-purity Type 310SS
Temperature, C
600
500
Alloy 600
Alloy 625
Alloy 25
10
Alloy DS
AISI 310
0.1
AISI 446
Corrosion after 1 year, mm
Corrosion after 1 year, mils
100
700
1
1.00 1.04 1.08 1.12 1.16 1.20 1.24 1.28
1000/(Temperature), 1/K
Fig. 7.34
Total surface recession including internal attack for

carbon steel and 1Cr-0.5Mo steel at 330, 400, and
500 C (626, 752, and 932 F) in CO-51.2H2-0.8CO2-1.7H2S0.2HCl. Source: Ref 68
Corrosion after 1 year, mils
100
Temperature, C
600
700
500
HK4M
Cr30A
1
Alloy 556
10
Alloy 188
Alloy 150
0.1
Cr35A
Corrosion after 1 year, mm
5000
1
1.00 1.04 1.08 1.12 1.16 1.20 1.24 1.28
1000/(Temperature), 1/K
Fig. 7.36
Total surface recession including internal attack for

Fe-, Ni-, and Co-base alloys at 700 C (1292 F)
and lower in CO-26H2-1.0CO2-0.9H2S-2H2O-10N2. Source:
Ref 68
Table 7.6 Test gas environments (nonequilibrium) used for investigating the effect of
water content on corrosion behavior of Type 310
and alloy 800 at 540 C (1004 F) for 600 h
Gas composition, vol %
Fig. 7.35
Total surface recession including internal attack

for alloy 625 at 330 to 700 C (626 to 1292 F) in
CO-51.2H2-0.8CO2-1.7H2S-0.2HCl. Source: Ref 68
Component
No. 1
H 2O
H2
CO
CO2
H 2S
HCl
pO2 at
540 C
0.0
1.0
2.0
3.0
5.0
10.0
15.0
32.0
30.2
29.2
28.8
28.2
28.2
30.0
64.0
64.0
64.0
64.0
60.0
52.0
45.0
3.2
4.0
4.0
4.0
6.0
9.0
9.2
0.8
0.8
0.8
0.8
0.8
0.8
0.8
0.04
0.04
0.04
0.04
0.04
0.04
0.04
1029 1028.4 1027.8 1027.4 107 1026.4 1026
Source: Ref 70
No. 2
No. 3
No. 4
No. 5
No. 6
No. 7
26/10/2007 11:15AM Plate # 0
Chapter 7:
signicantly improved suldation resistance

(Table 7.7). However, a layer of (Fe, Ni} suldes
was observed on top of the chromium oxide
scale. Thus, formation of molten sulde slag is
expected upon further exposure. No explanation
was given in the paper by Bradshaw et al.
(Ref 64) on the benecial effect of molybdenum.
The test temperature used by Bradshaw et al.
(Ref 64) was excessively high (i.e., 1000 C).
It would be of interest if the tests were conducted
at much lower temperatures. Natesan (Ref 72)
found that molybdenum and TZM (molybdenum
with 0.5Ti and 0.04Zr) formed very thin adherent
sulde scales when tested at 871 C (1600 F) in
a suldizing environment with low oxygen and
high sulfur potentials. Test results are shown in
Fig. 7.39 (Ref 72). He et al. (Ref 73) showed that
Ni-10Mo was more resistant to suldation than
nickel (Fig. 7.40) in short-term tests. The benecial effect of molybdenum in improving the
alloy suldation resistance in the environments
with low oxygen, high sulfur potentials would
be of interest, since molybdenum has been used
in many high-temperature nickel-base alloys to
provide solid solution strengthening for the
alloy. Some of these wrought alloys include
alloys X, S, 617, 625, and R-41. Furthermore,
molybdenum is also a major alloying element
in Ni-Cr-Mo corrosion-resistant alloys, such as
C-276, C-22, (622), 59, C-2000, and 686. These
Ni-Cr-Mo alloys contain molybdenum in a range
of 13 to 16%. High levels of molybdenum
in these Ni-Cr-Mo alloys can cause a thermal
pg 223
Sulfidation / 223
stability issue when used at high temperatures

because of formation of intermetallic phases
during the long-term, high temperature exposures. However, application of these alloys at
temperatures below 650 C (1200 F) may not
raise this thermal stability issue.
Addition of 3% Al to high-purity Type 310SS
signicantly improved suldation resistance, as
shown in Table 7.7. However, this composition
was not considered worthwhile for further
development because of expected fabrication
problems (Ref 64).
Manganese has a deleterious effect on an
alloys suldation resistance. In their tests in
Ar-40H2-15H2O-1.4H2S at 1000 C (1830 F)
for 24 h, Bradshaw et al. (Ref 64) found
that addition of 2% Mn adversely affected suldation resistance of Ni-30Cr and Type 310SS,
particularly a high-purity 310SS (Table 7.8).
Manganese diffuses the quickest through chromium oxide scale to form external suldes on
top of it, thus causing accelerating corrosion
(Ref 14).
The effect of silicon was investigated by
Nagarajan et al. (Ref 74) in Fe-18Cr alloys in
24H2-39H2O-18CO-12CO2-5CH4-1H2S-1NH3.
The Fe-18Cr-2Si alloy exhibited suldation
resistance signicantly better than that of
Fe-18Cr-0.5Si alloy (i.e., 20 mg/cm2 versus
slightly more than 200 mg/cm2) after 120 h at
980 C (1800 F) (pO2 =9:9 10716 atm, pS2 =
2:4 1076 atm, and ac = 0.3 at the test temperature). A nickel-base wrought alloy (HR160 alloy)
Metal loss (m)

200
Alloy 800
100
SS 310
20-35-3.25
0
0
10
15
% H2O
Fig. 7.37
Effect of H2O on the metal loss of Type 310, alloy 800, and experimental alloy Fe-20Cr-35Ni-3.25Si (20-35-3.25) tested at
540 C (1004 F) for 600 h. The test gas environments are shown in Table 7.6. Source: Ref 70
26/10/2007 11:15AM Plate # 0
was developed by Lai (Ref 75) using a combination of silicon, chromium, and cobalt to maximize the alloys suldation resistance as well as
its metallurgical stability, creep-rupture properties, and weldability. With high chromium (28%)
and high silicon (2.75%), a protective oxide scale
consisting of a chromium oxide scale with
underlying silicon enrichment formed on the
alloy surface (Ref 76, 77). In addition, cobalt
provided further improvement in the alloys
suldation resistance. Figure 7.41 shows the
suldation resistance of alloy HR160 compared
to some familiar alloys, such as alloys 556,
800H, and 600 in Ar-5H2-5CO-1CO2-0.15H2S
Fig. 7.38
at 870 C (1600 F) (Ref 77). The suldation

resistance of alloy HR160 was quite comparable
to that of cobalt-base alloy 6B and signicantly
better than other cobalt-base alloys, such as
alloys 188, 25, and 150 (Table 7.9). Norton et al.
(Ref 78) conducted corrosion tests on several
Fe-Ni-Cr and Ni-base alloys in comparison
with alloy HR160 at 700 C (1290 F) for
up to 1000 h in H2-7CO-1.5H2O-0.6H2S
(pO2 =10723 atm, pS2 =1079 atm, and ac = 0.3
to 0.4). Their test environment, as located in the
phase stability diagram at the test temperature, is
shown in Fig. 7.42 (Ref 78), and their test results
are summarized in Fig. 7.43 (Ref 78).
Corrosion of iron-, nickel-, and cobalt-base alloys after 215 h at (a) 760 C (1400 F), (b) 870 C (1600 F), and (c) 980 C
(1800 F) in Ar-5H2-5CO-1CO2-0.15H2S. Source: Ref 71
pg 224
26/10/2007 11:15AM Plate # 0
Chapter 7:
Norton et al. (Ref 79) also conducted corrosion

tests on an alumina-forming alloy MA956
(Fe-20Cr-4.5Al-0.5Y2O3),
which
is
an
oxide-dispersion-strengthened alloy. Also included in the tests were HR160, 45TM (Ni-Cr-FeSi), HR 3C (Fe-25Cr-20Ni-0.5Nb-0.2N), and an
experimental alloy (Fe-26Cr-39Ni-3.2V). Tests
were conducted in CO-32H2-3.8CO2-0.2H2S at
600 C (1112 F) and the test environment in
terms of pO2 and pS2 potentials in the phasestability diagram as shown in Fig. 7.44. Their
test results are summarized in Fig. 7.45 (Ref 80).
MA956 was found to be slightly less resistant
compared with HR160 and 45TM. Norton and
Levi (Ref 80) reported that a mixed Cr-Al oxide
scale with Cr-rich oxide on the outer layer and
Al-rich oxide on the inner layer formed on
MA956, and Fe-rich suldes grew on the top of
pg 225
Sulfidation / 225
the oxide scale. On the other hand, a thin, continuous and adherent Si-rich oxide scale enriched
in titanium in the outer region of the oxide scale
formed on HR160, and Cr-rich suldes grew on
the top of the oxide scale (Ref 80). The corrosion
(a)
Table 7.7 Results of corrosion tests at 1000 C

(1832 F) for 100 h in Ar-30H2-30H2O-1H2S(a)
Total affected
depth, m (mils)
Alloy
310SS
310HP(b)
310HP + 2% Ti
310HP + 3% Ti
310HP + 6% Mo
92 (3.6)
810 (31.9)
605 (23.8)
>1500 (59.1)
650 (25.6)
330 (13.0)
62 (2.4)
38 (1.5)
38 (1.5)
34 (1.3)
12 (0.5)
310HP + 3% Al
Comments
Internal suldation
Liquid suldes
Liquid suldes
Liquid suldes
Sulde penetration
Sulde penetration
Adherent oxide
Adherent oxide
Adherent oxide
Adherent oxide
Spalling oxides and (Fe.Ni)
suldes over Cr2O3
Spalling oxides
Spalling oxides
5 (0.2)
5 (0.2)
(a) Test gas was at 1 atm; pO2 =3 10715 atm and pS2 =3 1076 atm: (b) HP
indicates high-purity material. Source: Ref 64
(b)
Fig. 7.40
Weight change data comparing nickel (a) with Ni10Mo alloy (b) when tested at 550 and 600 C
(1022 and 1112 F) in Ar-13.56H2-0.6H2O-1.89H2S (pO2 = 2.7
27
10 , pS2 = 4.1 108 at 600 C; pO2 = 4.5 1029, pS2 = 1 108 at
550 C). Source: Ref 73
Weight change, mg/mm2
2.0
1.6
Alloy 800
Test temperature 871 C

p = 4.1 x 1018 atm
O
p 2 = 9.4 x 107 atm
Table 7.8 Results of corrosion tests at 1000 C

(1832 F) for 24 h in Ar-40H2-15H2O-1.4H2S(a)
S2
1.2
Alloy
0.6
Total affected
depth, m (mils)
310SS
56 (2.2)
310SS + 2% Mn
72 (2.8)
310 SS
0.4
Ta
TZM
Nb
0
30
60
90
120
Mo
v
150
Exposure time, h
Fig. 7.39
Thermogravimetric data for refractory metals, such

as, Mo, TZM, Nb, Ta, and V in comparison with
conventional alloys in a sulfidizing environment with high sulfur
and low oxygen potentials tested at 871 C (1600 F). Source:
Ref 72
310HP(b) + 2% Mn
Ni-30Cr
Ni-30Cr-2Mn
420 (16.5)
38 (1.5)
1220 (48.0)
Comments
Sulde layer
Oxide layer
Intergranular suldes
Internal suldes
Sulde layer
Oxide layer
Internal suldes
Liquid suldes
Oxides
Internal suldes
Liquid suldes
(a) Test gas was gas at 1 atm; pO2 =3 10716 and pS2 =3 1076 atm: (b) HP
indicates high-purity material. Source: Ref 64
26/10/2007 11:15AM Plate # 0
pg 226
morphologies for both alloys after exposure for

2000 h are schematically illustrated in Fig. 7.46
(Ref 80). Additional low temperature suldation
data generated in reducing environments can be
found in Tables 10.5 and 10.6 and Figures 10.17
and 10.18.
7.5 Corrosion in SO2-Bearing

Environments
binary alloys in SO2 and SO2-O2 environments.

Metals and alloys exposed to these environments
generally form oxides and/or suldes as corrosion products. Corrosion products and corrosion rates depend strongly on temperature. The
Table 7.9 Sulfidation resistance of Alloy HR160

compared to cobalt-base alloys at 870 C (1600 F)
for 500 h(a)
Alloy
Most corrosion studies of SO2-bearing environments have been conducted in either SO2 or
O2-SO2 mixtures. Materials investigated include
nickel (Ref 8184), iron (Ref 8587), cobalt
(Ref 8889), chromium (Ref 90), Co-Cr alloys
(Ref 91), and Ni-Cr alloys (Ref 9297). Kofstad
(Ref 98) reviews the corrosion of pure metals and
Fig. 7.41
HR160
556
188
25
150
6B
Weight change,
mg/cm2
Metal loss,
mm (mils)
Maximum metal
affected(b), mm (mils)
0.5
183.6
40.6
33.0
131.7
10.7
0.01 (0.2)
0.52 (20.6)
0.19 (7.6)
0.10 (4.1)
0.26 (10.3)
0.01 (0.3)
0.13 (5.2)
0.90 (35.6)
0.60 (23.6)
0.37 (14.6)
0.72 (28.3)
0.08 (3.3)
(a) Ar-5H2-5CO-lCO2-0.15H2S; pO2 =3 10719 atm, pS2 =0:9 1076 atm:

(b) Metal loss + maximum internal penetration. Source: Ref 76
Sulfidation resistance of alloy HR160 compared to those of alloys 556, 800H, and 600 after 215 h at 870 C (1600 F) in
Ar-5H2-5CO-1CO2-0.15H2S (pO2 = 3 1019 atm, pS2 = 0.9 106 atm). Samples were cathodically descaled before
being mounted for metallographic examination. Source: Ref 77
26/10/2007 11:15AM Plate # 0
pg 227
Chapter 7:
0
FeS2
reaction between SO2 and O2 leads to formation

of SO3. The kinetics for this reaction are relatively slow without catalysts such as platinum.
Most tests were conducted in conjunction with
platinum catalysts to obtain an equilibrium SO3
partial pressure. SO3 is an important reactant in
the corrosion reaction involving SO2 (Ref 83,
93). Corrosion rate was found to be dependent on
the ratio of SO2:O2. When that ratio is 2:1, which
gives the highest partial pressure of SO3, the
corrosion rate is generally fastest, as illustrated in
Fig. 7.47 (Ref 83). The corrosion reaction is also
temperature dependent. Corrosion rate increases
with increasing temperature until a maximum is
reached. Further increases in temperature result
in a reduced corrosion rate (Ref 83). With an
SO2:O2 ratio of 2:1, nickel corroded fastest at
700 to 800 C (1290 to 1470 F) (Ref 83).
250
W, mg/cm2
150
Type 347 SS
100
Alloy 800H
50
0
0
200
(a)
600
400
Exposure, h
800
1000
100
80
Alloy 625
60
Alloy X
40
Alloy 617
20
NiS
0
Cr2S3
Test gas
Co9S8
200
(b)
oO
400
600
Exposure, h
800
1000
Log pS , bar
Type 321 SS
200
W, mg/cm2
reaction rate generally peaks at a certain temperature, then decreases with increasing temperature. The highest corrosion rate is normally
related to the formation of suldes. Suldes
provide paths for rapid outward diffusion of
metals, such as nickel, iron, and chromium, and
so forth, resulting in rapid corrosion attack
(Ref 98). The rate is particularly rapid when
liquid sulde eutectics are formed. The Ni-Ni3S2
eutectic melts at 635 C (1175 F), the Fe-FeS
eutectic at 985 C (1805 F), and the Co-Co4S3
eutectic at 880 C (1616 F) (Ref 17).
The temperature at which the corrosion rate is
highest varies from metal to metal. For nickel, it
is around 600 C (1110 F). At temperatures
above 800 C (1470 F), the rate decreases with
increasing temperature (Ref 98). Suldes
become less and less stable as temperature
increases. Eventually, NiO becomes the only
corrosion product (Ref 98). For cobalt, the
corrosion rate is highest at 920 C (1688 F),
and decreases above that temperature (Ref 98).
The corrosion rate for iron is high above
940 C (1724 F) (Ref 98), which corresponds
approximately to the iron sulde eutectic temperature. Chromium, on the other hand, does not
form suldes in SO2 at temperatures from 700 to
1000 C (1290 to 1830 F), but instead forms a
Cr2O3 scale (Ref 90). Thus, chromium as an
alloying element improves suldation resistance
in Co-Cr (Ref 91) and Ni-Cr alloys (Ref 92).
Corrosion tests in environments containing
both O2 and SO2 were conducted primarily
in SO2-O2 mixtures with various ratios. The
Sulfidation / 227
Ni3S2
10
FeS
Oo
10
Ni
Type 347 SS
Fe
W, mg/cm2
CrS
15
FeO
Fe3O4
Cr Cr2O3
NiO
Alloy 800H
Alloy HR-120
Alloy 556
4
2
20
35
30
25
20
15
Log pO , bar
Alloy HR-160
0
0
(c)
Fig. 7.42
Ref 78
Test environment in terms of pO2 and pS2 is plotted

in the stability diagram at 700 C (1202 F) Source:
Fig. 7.43
200
400
600
Exposure, h
800
1000
Resuts of corrosion tests in H2-7CO-1.5H2O0.6H2S at 700 C (1292 F). Source: Ref 78
26/10/2007 11:15AM Plate # 0
p
The test environment in terms of pO2 and pS2 potentials at 600 C in CO-32H2-3.8CO2-0.2H2S is plotted as an equilibrium
condition (identified as E) and a nonequilibrium (NE) is plotted in the phase-stability diagram. Source: Ref 79
weight gain, mg/cm2
Fig. 7.44
8
7
6
5
4
3
2
1
0
26/37/3
HR3C
MA960 (ave)
500
1000
1500
45 TM
MA956 (low)
2000 HR160
time, h
Fig. 7.45
Corrosion behavior of MA956 in comparison with HR160, 45TM, HR3C, and an experimental alloy 26Cr/37Ni/3V (26/37/3)
in CO-32H2-3.8CO2-0.2H2S at 600 C (1112 F). Source: Ref 79
Fig. 7.46
Corrosion scales formed on MA956 (left picture)

and HR160 (right picture) after exposure at 600 C
(1112 F) for 2000 h in CO-32H2-3.8CO2-0.2H2S. For MA956,
the oxide scale (in black) was a mixed Cr-Al-rich oxide scale with
the outer layer being chromium-rich oxide scale and the inner
layer being aluminum-rich oxide scale. For HR160, the oxide
scale (in black) was silicon-rich oxide enriched with titanium on
the outer layer of the oxide scale. Source: Ref 80
Nickel forms nickel oxides, nickel suldes, and

NiSO4 when the corrosion rate is highest
(Ref 98). The suldes provide paths for outward
diffusion of nickel. At sufciently high temperatures, NiO becomes the only stable corrosion
product (Ref 98). Then the corrosion rate is
lowered. Ni-20Cr alloy in SO2-O2 mixture
behaved similarly to nickel (Ref 98). Vasantasree
and Hocking (Ref 93) investigated Ni-Cr alloys
with various chromium contents. Their results
are shown in Fig. 7.48. Rates were highest at
700 C (1290 F) for Ni-Cr alloys. Formation of
oxides was favored at high temperatures, particularly 900 and 1000 C (1650 and 1830 F).
Therefore, corrosion rates became much lower at
these temperatures.
Very few investigators have tested commercial
alloys in SO2-bearing environments. Barnes and
Lai (Ref 99) examined the oxidation behavior of
two iron-base commercial alloys, Type 304SS
and alloy 556, in Ar-5O2-5CO2 with and without
10% SO2 at 980 C (1800 F). The equilibrium
partial pressures for the Ar-5O2-5CO2-10SO2 at
pg 228
26/10/2007 11:15AM Plate # 0
Chapter 7:
Fig. 7.47
Initial linear rate constants of nickel in different

SO2-O2 mixtures at 1 atm and 800 C (1470 F).
Also shown are equilibrium SO3 pressures at different SO2-O2
mixtures. Source: Ref 83
Parabolic rate constant, mg2/cm4/h
103
700
SO2:O2::2:1
Temp. C
600
700
800
900
1000
L=Linear
800
102
600
101
900
100
1000
101
102
100
Cr
Fig. 7.48
50
20 10
at.%, Cr
0.1 0
Ni
Parabolic rate constants of Ni-Cr alloys as a function of chromium concentration in SO2:O2 mixture
(2:1 ratio) at various temperatures. Source: Ref 93
pg 229
Sulfidation / 229
the test temperature were 4.9 102 atm for pO2 ,

9.7 102 atm for pSO2 , 3.6 103 atm for pSO3 ,
and 1 1022 atm for pS2 . In the SO2-bearing test
environment, Type 304 sample suffered severe
oxidation/suldation attack, with some areas
being completed consumed by corrosion attack.
On the other hand, alloy 556 formed a thin,
compact chromium-rich oxide scale with some
internal chromium-rich suldes underneath the
oxide scale. Table 7.10 summarizes the corrosion
attack for both alloys in both environments. The
results suggest that an alloy that forms a protective chromium oxide scale in a purely oxidizing
environment is likely to form a similar scale in
SO2-O2 environment. Many industrial processes
that generate SO2-bearing environments are
generally at much lower temperatures. Yates
et al. (Ref 100) tested alloys X (identied
as HX Ni-22Cr-9Mo-18Fe), 617 (Ni-22Cr12Co-9Mo-1.2Al), 230 (Ni-22Cr-14W-La), 188
(Co-22Cr-20Ni-14W-La), and 214 (Ni-16Cr3Fe-4.5Al-Y) at 704 C (1300 F) in O2- 4%
SO2 for more than 40 days. Their test results are
summarized in Fig. 7.49. Based on these test
results, Ni-Cr and Co-Cr alloys containing
about 22% Cr are considered to have adequate
corrosion resistance in O2-4%SO2 at 704 C
(1300 F). All four 22Cr alloys (X, 617, 188, and
230) exhibited a parabolic reaction kinetics with
low mass changes over more than 40 days of
exposure, indicating formation of protective
chromium-rich oxide scales. On the other hand,
Ni-Cr-Al alloy 214 with only about 16% Cr
showed some indication of breakaway corrosion.
The test temperature of 704 C (1300 F) was
likely too low for rapid formation of Al2O3 in
O2-SO2 mixtures. The amount of chromium
(about 16%) is considered to be inadequate in
forming protective Cr2O3 scales under the test
condition. Field test data, which was generated
in a chemical plant with the environment containing about 18% SO2 in the temperature range
of 260 to 371 C (500 to 700 F), showed
minimal corrosion rates (about 0.3 mpy and less)
for Type 316, 317, alloy 20, and alloy 825
(Ref 101).
The levels of SO2 used for most corrosion
studies in either SO2 environments or SO2-O2
mixtures are signicantly higher than the
amounts expected in the combustion of sulfurbearing fuels such as coal or oil. Typical combustion ue gas produced in a coal-red boiler,
for example, contains approximately 0.25% SO2
(Ref 102). Very few investigators have studied
SO2 corrosion at this low level (i.e., less than
26/10/2007 11:15AM Plate # 0
pg 230
Table 7.10 Corrosion of Type 304SS and alloy 556 at 980 C (1800 F) for 550 h in an oxidizing
environment with and without SO2
Type 304SS
Test environment
Ar-5O2-5CO2
Ar-5O2-5CO2-10SO2
Metal loss,
mm (mils)
0.31 (12.2)
> 0.61 (24)
Alloy 556
Maximum depth of
attack, mm (mils)(a)
0.44 (17.2)
> 0.61 (24)
Metal loss,
mm (mils)
Maximum depth of
attack, mm (mils)(a)
0.005 (0.2)
0.06 (2.5)
0.056 (2.2)
0.10 (4.0)
(a) Metal loss + maximum internal penetration. Source: Ref 99
Table 7.11 Corrosion of Ni-15Cr alloy in

SO2-bearing environments at 870 C (1600 F)
for 30 h
Test environment
Weight loss,
mg/cm2
Depth of attack,
mm (mils)
Pure SO2
N2-1.35SO2
N2-1SO2
N2-0.2SO2
N2-0.05SO2
N2-0.2SO2-0.01O2
N2-0.2SO2-0.1O2
N2-0.2SO2-2O2
N2-0.2SO2-3O2
0.6
1.8
2.0
86.0
13.7
6.5
0.3
0.3
0.008 (0.3)
0.52 (20.5)
0.008 (0.3)
Source: Ref 103
Fig. 7.49
Mass change as a function of exposure time

for alloys X (identified as HX) 617, 230, 188 and
214 at 704 C (1300 F) in O2-4% SO2. Source: Ref 100
1%). However, studies have been conducted of

corrosion in environments containing low concentrations of SO2 in conjunction with ash/salt
deposits in order to simulate fuel ash corrosion in
combustion systems, particularly fossil-red
boilers. These data are discussed in later chapters
covering specically on coal-red boilers and
oil-red boilers. Viswanathan and Spengler
(Ref 103) observed that Ni-15Cr alloy suffered
signicantly more corrosion attack in 0.2% SO2
(balance N2) than in pure SO2 at 870 C
(1600 F). The sample exposed to N2-0.2SO2 at
870 C (1600 F) for 30 h showed about
0.5 mm (20 mils) of attack with internal globular
sulde phases, while the sample exposed to
100% SO2 for same temperature and time
exhibited only about 0.008 mm (0.3 mil) of
attack. Addition of oxygen signicantly reduced
the alloys corrosion rates. The weight loss for
the sample after 30 h fell from 86 mg/cm2 in N20.2SO2 to about 0.3 mg/cm2 in N2-0.2SO20.1O2. The results of their tests (Ref 103) are
summarized in Table 7.11. Low concentrations
of SO2 in the environment with no free oxygen

can cause Ni-Cr alloys to suffer severe suldation attack even without ash/salt deposits.
7.6 Summary
The suldation behavior of a wide variety of
alloys was reviewed. The data presented are
primarily related to gaseous environments;
sulfate-accelerated suldation (hot corrosion) is
covered in Chapter 9. The corrosion data are
grouped into three different types of environments: (1) sulfur vapor, hydrocarbon streams
containing no hydrogen gas, H2S, and H2-H2S,
(2) reduced, mixed gas environments with low
oxygen and high sulfur potentials, and (3) SO2bearing environments.
Sulfur vapor, hydrocarbon streams (no H2),
H2S, and H2-H2S environments have a common
feature in that the corrosion products formed in
these environments are essentially suldes. The
reduced, mixed gas environments with low
oxygen and high sulfur potentials typically contain H2, CO, CO2, H2O, H2S, and other gaseous
components where oxides and suldes can form
on most high temperature alloys. In these environments, alloys are protected by a protective
26/10/2007 11:15AM Plate # 0
Chapter 7:
oxide scale before the initiation of breakaway

corrosion due to formation of fast-growing,
nonprotective suldes. The SO2-bearing environments are generated by combustion of a sulfurcontaining fuel or feedstock with excess air or
oxygen. These environments are oxidizing, and
are generally less corrosive than reducing
environments.
REFERENCES
1. G.Y. Lai, J. Met., July 1985, p 14

2. G.L. Swales, in Behavior of High Temperature Alloys in Aggressive Environments, I. Kirman et al., Ed., Proc. Petten
International Conference, Oct 1518,
1979, The Metals Society, London, 1980,
p 45
3. G. Sorell, M.J. Humphries, E. Bullock, and
M. Van de Voorde, Int. Met. Rev., Vol 31
(No. 5), 1986, p 216
4. J.F. Norton, Ed., High Temperature Materials Corrosion in Coal Gasication
Atmospheres, Elsevier, Amsterdam, 1984
5. K.J. Barton, V.L. Hill, and R. Yurkewycz,
in The Properties and Performance of
Materials in the Coal Gasication Environments, V.L. Hill and H.L. Black, Ed.,
American Society For Metals, 1981, p 65
6. S.K. Srivastave, G.Y. Lai, and D.E. Fluck,
7. J.A. Harris, W.G. Lipscomb, and G.D.
Smith, Paper No. 402, Corrosion/87,
NACE, 1987
8. J. Stringer, in High Temperature Corrosion, R.A. Rapp, Ed., Conference Proceedings (San Diego, CA) March 26,
1981, NACE, 1981, p 389
9. A.J. Minchener, D.M. Lloyd, and P.T.
Sutcliffe, Materials Evaluation for Fluidized Bed Combustion Systems, CS3511, Final Report to EPRI on Research
Project RP979-11, Electric Power
Research Institute, Palo Alto, CA, 1984
10. J. Stringer, Paper No. 90, Corrosion/86,
NACE, 1986
11. S.R. Shatynski, Oxid. Met., Vol 11 (No. 6),
1977, p 307
www.outokumputechnology.com,
accessed Dec 2006
13. ChemSage, Version 4.16, GTT-Technologies, Aachen (1998)
pg 231
Sulfidation / 231
14. R.A. Perkins, in Environmental Degradation of High Temperature Materials, Series

3, Vol 2 (No. 13), 1980, p 5/1
15. P.L. Hemmings and R.A. Perkins,
Thermodynamic Phase Stability Diagrams for the Analysis of Corrosion
Reactions in Coal Gasication/Combustion Atmospheres, EPRI Report FP-539,
Lockheed Palo Alto Research Laboratories, Palo Alto, CA, 1977
16. B.A. Gordon and V. Nagarajan, Oxid. Met.,
Vol 13 (No. 2), 1979, p 197
17. M. Hansen and K. Anderko, Constitution
of Binary Alloys, McGraw-Hill, 1958
18. S. Mrowec and K. Przybylski, High
Temp. Mater. Proc., Vol 6 (No. 1 and 2),
1984, p 1
19. D.J. Young, Rev. High Temp. Mater., Vol 4
(No. 4), 1980, p 299
20. A. Davin and D. Coutsouradis, Cobalt, Vol
17, 1962, p 23
21. S. Mrowec, T. Walec, and T. Werber, Oxid.
Met., Vol 1, 1969, p 93
22. T. Narita, W.W. Smeltzer, and K. Nihida,
Oxid. Met., Vol 17, 1982, p 299
23. T. Narita and K. Nihida, Oxid. Met., Vol 6,
1973, p 157 and 181
24. S.K. Mrowec, T. Werber, and M. Zastawnik, Corros. Sci., Vol 6, 1966, p 47
25. D.P. Whittle, S.K. Verma, and J. Stringer,
Corros. Sci., Vol 13, 1973, p 247
26. T. Biegun, A. Bruckman, and S. Mrowec,
Oxid. Met., Vol 12, 1978, p 157
27. S. Mrowec and M. Wedrychowska, Oxid.
Met., Vol 13, 1979, p 481
28. E.M. Jallouli, J.P. Larpin, M. Lambertin,
and J.C. Colson, J. Electrochem. Soc., Vol
126, 1979, p 2254
29. T. Biegun and A. Bruckman, Bull. Acad.
Polon. Ser. Sci. Chim., Vol 28, 1980, p 377;
Vol 29, 1981, p 69
30. W.W. Smeltzer, T. Narita, and K. Przybylski, in Proc. Corrosion-Erosion, Wear
of Materials in Emerging Fossil Energy
Systems, A.V. Levy, Ed., NACE, 1982,
p 860
31. T. Biegun and A. Bruckman, Suldation
of Ni-Cr-Al Alloys, Report No. 2.34264,
Institute of Physical Chemistry, Polish
Academy of Science, Warsaw, 1980
32. L.A. Morris, Chapter 17: Resistance to
Corrosion in Gaseous Atmospheres, in
Handbook of Stainless Steels, D. Peckner
and I.M. Bernstein, McGraw-Hill, 1977,
p 17.1
26/10/2007 11:15AM Plate # 0
33. E. Malinowski, R. Bigot, and E. Herzog,

Le Metallurgic, Vol 94 (No. 4), 1962
34. F.J. Bruns, Corrosion of Ni-Cr-Al-Fe
Alloys by Hydrogen Sulde at 1100 to
1800 F, Corrosion, Vol 25 (No. 3), 1969,
p 119
35. Z.A. Foroulis, High Temperature Degradation of Structural Materials in Environments Encountered in the Petroleum and
Petrochemical Industries: Some Mechanistic Observations, Anti-Corrosion, Vol 32
(No. 11), 1985, p 49
36. J. Gutzeit, R.D. Merrick, and L.R. Scharfstein, Corrosion in Petroleum Rening and
Petrochemical Operations, in Corrosion,
Vol 13, 9th ed., Metals Handbook, ASM
37. F.A. Hendershot and H.L. Valentine,
Materials for Catalytic Cracking Equipment (Survey), Mater. Prot., Vol 6 (No. 10),
1967, p 43
38. J. Gutzeit, High Temperature Suldic
Corrosion of Steels, in Process Industries
CorrosionThe Theory and Practice,
NACE, 1986
39. The Role of Stainless Steels in Petroleum
Rening, Originally published by the
Committee of Stainless Steel Producers,
AISI (1977), Nickel Development Institute, Toronto, Ontario, Canada, April 1996
40. Steels for Hydrogen Service at Elevated
Temperatures and Pressures in Petroleum
Reneries and Petrochemical Plants,
Publication 941, 3rd ed., American Petroleum Institute, 1983
41. R.A. White and E.F. Ehmke, Materials for
Reneries and Associated Facilities,
NACE, 1991, p 51
42. B.W. Neumaier and C.M. Schillmoller,
How Richeld Plans to Combat HighTemperature Sulde Corrosion in Its
New Catalytic Reformer, presented at the
21st Midyear Meeting of the American
Petroleum Institutes Division of Rening
(Montreal, Canada), May 14, 1956
43. G. Sorell, Compilation and Correlation of
High Temperature Catalytic Reformer
Corrosion Data, Technical Committee
Report, Publication 58-2, NACE, Houston,
TX, 1957
44. E.B. Backensto, R.E. Drew, J.E. Prior, and
J.W. Sjoberg, High-Temperature Hydrogen Sulde Corrosion of Stainless Steels,
Technical Committee Report, Publication
58-3, NACE, Houston, TX, 1957
45. G. Sorell and W.B. Hoyt, Collection and

Correlation of High Temperature Hydrogen Sulde Corrosion Data, NACE
Technical Committee Report, Publication
56-7, NACE, Houston, TX, 1956
46. E.B. Backensto, Corrosion in Catalytic
Reforming and Associated Processes,
Summary Report of the Panel on Reformer
Corrosion to the Subcommittee on Corrosion, presented at the 22nd Midyear
Meeting of APIs Division of Rening
(Philadelphia, PA), May 13, 1957
47. E. Dittrich, Chem. Fab., Vol 10 (No. 13/
14), 1947, p 145
48. E.B. Backensto and J.W. Sjoberg,
Iso-Corrosion Rate Curves for High
Temperature Hydrogen-Hydrogen Sulde, Technical Committee Report,
Publication 59-10, NACE, Houston, TX,
1958
49. A.S. Couper and J.W. Gorman, Computer
Correlations to Estimate High Temperature
H2S Corrosion in Renery Streams, Mater.
Prot. Perform., Vol 10 (No. 4), 1971, p 31
50. V.L. Hill and H.S. Meyer, in High Temperature Corrosion in Energy Systems,
M.F. Rothman, Ed., The Metallurgical
51. A.O. Schaefer, C.H. Samans, M.A. Howes,
S. Bhattacharyya, E.R. Bangs, V.L. Hill,
and F.C. Chang, A Program to Discover
Materials Suitable for Service under
Hostile Conditions Obtaining in Equipment for the Gasication of Coal and
Other Solid Fuels, 1975 Annual Report,
The Metal Properties Council, New York,
1976
52. A.O. Schaefer, A Program to Discover
Materials Suitable for Service Under
Hostile Conditions Obtaining in Equipment for the Gasication of Coal and Other
Solid Fuels, 1976 Annual Report, The
Metal Properties Council, New York, 1977
Materials Suitable for Service Under Hostile Conditions Obtaining in Equipment for
the Gasication of Coal and Other Solid
Fuels, 1977 Annual Report, The Metal
Properties Council, New York, 1978
pg 232
26/10/2007 11:15AM Plate # 0
Chapter 7:

Fuels, Supplement to 1978 Annual
Report, The Metal Properties Council, New
York, 1979
Materials Suitable for Service Under Hostile Conditions Obtaining in Equipment
for the Gasication of Coal and Other Solid
Materials Suitable for Service Under
Hostile Conditions Obtaining in Equipment for the Gasication of Coal and Other
Solid Fuels, 1981 Annual Report, The
Metal Properties Council, New York, 1982
59. A. Humphreys and A.O. Schaefer, A
Program to Discover Materials Suitable for
Service under Hostile Conditions Obtaining in Equipment for the Gasication of
Coal and Other Solid Fuels, 1982 Annual
Report, The Metal Properties Council, New
York, 1983
60. M.A.H. Howes, High Temperature Corrosion in Coal Gasication Systems, Final
Report GRI-8710152, Gas Research Institute, Chicago, Aug 1987
61. R.A. Perkins and S.J. Vonk, EPRI Report
FP-1280, Electric Power Research Institute, Palo Alto, CA, Dec 1979
62. K. Natesan, in High Temperature Corrosion, R.A. Rapp, Ed., NACE, 1983, p 336
63. S.K. Verma, Corrosion of Commercial
Alloys in A Laboratory-Simulated Medium-BTU Coal Gasication Environment,
64. R.W. Bradshaw, R.E. Stoltz, and D.R.
Adolphson, Report SAND 77-8277,
Sandia Laboratories, Livermore, CA 1977
65. R.W. Bradshaw, A.N. Nagelberg, R.E.
Stoltz, and D.R. Adolphson, Report SAND
78-8260, Sandia Laboratories, Livermore,
CA, 1978
66. J.A. Kneeshaw, I.A. Menzies, and J.F.
Norton, Werkst. Korros., Vol 38, 1987,
p 473
pg 233
Sulfidation / 233
67. J.L. Blough, V.L. Hill, and B.A. Humphreys, in The Properties and Performance
of Materials in the Coal Gasication
Environment, V.L. Hill and H.L. Black,
Ed., American Society For Metals, 1981,
p 225
68. R.C. John, W.C. Fort III, and R.A. Tait,
Prediction of Alloy Corrosion in the Shell
Coal Gasication Process, Mater. High
Temp., Vol 11 (No. 14), 1993, p 124
69. W.T. Bakker, Effect of Gasier Environment on Materials Performance, Mater.
High Temp., Vol 11 (No. 14), 1993, p 81
70. W.T. Bakker and J.A. Bonvallet, Corrosion
of Stainless Steels on the Wrong Side of
the Kinetic Boundary, in Heat-Resistant
Materials II, Conf. Proc. Second International Conference on Heat-Resistant
Materials, K. Natesan, P. Ganesan, G. Lai,
Ed., ASM International, 1995, p 121
71. G.Y. Lai, in High Temperature Corrosion
in Energy Systems, M.F. Rothman, Ed.,
The Metallurgical Society of AIME, 1985,
p 227
72. K. Natesan, Surface Modication for
Corrosion Resistance, Mater. High Temp.,
Vol 11 (No. 14), 1993, p 36
73. Y.-R. He, D.L. Douglass, and F. Gesmundo, The Corrosion Behavior of
Ni-Mo Alloys in a H2/H2O/H2S Gas
Mixture, Oxid. Met., Vol 37 (No. 5/6),
1992, p 413
74. V. Nagarajan, R.G. Miner, and A.V. Levy,
J. Electrochem. Soc., Vol 129 (No. 4),
1982, p 782
75. G.Y. Lai, Suldation-Resistant Co-Cr-Ni
Alloy With Critical Contents of Silicon and
Cobalt, U.S. Patent No. 4711763, Dec
1987
76. G.Y. Lai, Paper No. 209, Corrosion/89,
NACE, 1989
77. G.Y. Lai, J. Met., Vol 41 (No. 7), 1989,
p 21
78. J.F. Norton, F.G. Hodge, and G.Y. Lai,
A Study of the Corrosion Behavior of
Some Fe-Cr-Ni and Advanced Ni-Based
Alloys Exposed to a Sulphidising/Oxidising/Carburising Atmosphere at 700 C, in
High Temperature Materials for Power
Engineering (Part I), Conf. Proc., E.
Bachelet et al., Ed., Kluwer Academic
Publishers, Dordrecht, The Netherlands,
1990, p 167
79. J.F. Norton, T.P. Levi, and W.T. Bakker,
High Temperature Corrosion of Candidate
26/10/2007 11:15AM Plate # 0
80.
81.
82.
83.
84.
85.
86.
87.
88.
89.
90.
91.
Heat Exchanger Alloys in a Dry-Feed

Entrained Slagging Gasier Atmosphere,
in High Temperature Materials for Power
Engineering (Part II), Conf. Proc., E.
Bachelet et al., Ed., Kluwer Academic
Publishers, Dordrecht, The Netherlands,
1994, p 1617
J.F. Norton and T.P. Levi, A Laboratory
Study of the Corrosion Behaviour of
Alloys Exposed in a Non-Equilibrated
Coal-Gasication Atmosphere at 600 C,
M. Seiersten and P. Kofstad, Corros. Sci.,
Vol 22 (No. 5), 1982, p 487
M.R. Wootton and N. Birks, Corros. Sci.,
Vol 12, p 829
B. Haan and P. Kofstad, Corros. Sci.,
Vol 23 (No. 12), 1983, p 1333
A. Andersen, B. Haan, P. Kofstad, and
P.K. Lillerud, Mater. Sci. Eng., Vol 87,
1987, p 45
J. Gilewicz-Wolter, Oxid. Met., Vol 11,
1977, p 81
A. Rahmel, Oxid. Met., Vol 9, 1975, p 491
A. Rahmel, Corros. Sci., Vol 13, 1975,
p 125
P. Singh and N. Birks, Oxid. Met., Vol 12,
1978, p 1
P. Singh and N. Birks, Oxid. Met., Vol 12,
1978, p 22
C.D. Asmundis, F. Gesmundo, and C.
Bottino, Oxid. Met., Vol 14 (No. 4), 1980,
p 351
P. Singh and N. Birks, Oxid. Met., Vol 13
(No. 5), 1979, p 457
92. H. Lewis and J.E. Whittle, in Proc. Fourth

Int. Cong. Met. Corros., NACE, TX, 1972
93. V. Vasantasree and M.G. Hocking, Corros.
Sci., Vol 16, 1976, p 261
94. M.G. Hocking and V. Vasantasree, Corros.
Sci., Vol 16, 1976, p 279
95. K.N. Strafford, P.K. Datta, A.F. Hampton,
and P. Mistry, Corros. Sci., Vol 29 (No. 6),
1989, p 673
96. P.S. Sidky and M.G. Hocking, Corros. Sci.,
Vol 27 (No. 2), 1987, p 183
97. M.G. Hocking and P.S. Sidky, Corros. Sci.,
Vol 27 (No. 2), 1987, p 205
98. P. Kofstad, High Temperature Corrosion,
Elsevier Applied Science, 1988
99. J.J. Barnes and G.Y. Lai, Paper No. 90276,
100. D.H. Yates, P. Ganesan, and G.D. Smith,
Recent Advances in the Enhancement of
Inconel Alloy 617 Properties to Meet the
Needs of the Land Based Gas Turbine Industry, in Advanced Materials and Coatings for Combustion Turbines Conference
Proceedings, V.P. Swaminathan and N.S.
Cheruvu, ASM International, 1994, p 89
101. A Guide to Corrosion Resistance, Climax
Molybdenum Company, Greenwich, CT,
1981
102. R.W. Borio, A.L. Plumley, and W.R. Sylvester, in Ash Deposits and Corrosion Due
to Impurities in Combustion Gases, R.W.
Bryers, Ed., Hemisphere Publishing, 1978,
p 163
103. R. Viswanathan and C.J. Spengler, Corrosion, Vol 26 (No. 1), 1970, p 29
pg 234

DOI: 10.1361/hcma2007p235
26/10/2007 12:33PM Plate # 0

CHAPTER 8
Erosion and Erosion-Corrosion

8.1 Introduction
Industrial plants are often involved in processes where gas streams are laden with particles.
Metallic components can suffer severe metal
wastage when subjected to constant impingement by this particle-laden gas stream under
high velocity and high particle loading. The
continual removal of the material from a component by this process is referred to as erosion.
When the component is exposed to elevated temperatures where oxidation and/or other modes of
high-temperature corrosion are involved, it is
frequently referred to as erosion-corrosion.
Materials problems that are caused by erosion or
erosion-corrosion in industries are numerous.
In the petrochemical industry, the pyrolysis
furnace tubes for the production of ethylene
are a good example. Ethylene is produced by
cracking petroleum feedstocks, such as ethane
and naphtha, at temperatures up to 1150 C
(2100 F), thus making the process gas stream
inside the tube highly carburizing in nature. The
furnace tubes suffer both carburization and
coking on the internal surface of the tube. In
order to maintain the process efciency, the coke
deposits have to be regularly removed from the
tube inner diameter (ID) surface by a process
referred to as decoking, which involves injecting a mixture of steam and air into the furnace
tube. Thus, during the decoking operation, the
return bends of the pyrolysis furnace tubes can
suffer erosion or erosion-corrosion due to the
coke particle-laden gas stream (Ref 1, 2), with
average tube skin temperatures varying from 800
to 1120 C (1475 to 2050 F) and gas velocities
greater than 200 m/s (656 ft/s).
In coal-red boilers, it is well known that yash erosion can be a serious problem for boiler
tubes. When coal (which contains ash) is combusted in the lower furnace of the boiler, some of
the ash drops out of the furnace from the bottom
with remaining ash being carried by the combustion ue gas stream to the top of the furnace
and through the convection pass. The ash carried

by the ue gas stream is referred to as y-ash.
With pulverized coal ring, about 70 to 90% of
the ash in the coal is carried by the ue gas
stream, while only about 40% of the ash is carried
by the ue gas in a stoker-red furnace (Ref 3).
The cyclone-red boiler generates about 15 to
30% of the ash from the coal in the ue gas.
Major constitutes of y-ash are SiO2, Al2O3, and
Fe2O3. (More information about ash constituents
in coal is available in Chapter 10 Coal-Fired
Boilers.) This y-ash-laden ue gas stream can
pose y-ash erosion problems to convection pass
tubes, such as superheater, reheater, boiler bank,
and economizer tubes. To reduce y-ash erosion
problems for the convection pass tubes, boiler
designers typically have set maximum ue gas
velocities in these areas. For example, Babcock
& Wilcox typically limits ue gas velocity to
19.8 m/s (65 ft/s) or less for relatively nonabrasive low ash coal, and 13.7 m/s (45 ft/s) or
less for coals with high ash quantities and/or
abrasive ash (Ref 4). Combustion Engineering
(now Alstom Power) typically set the design
velocity in the range of 12 to 18 m/s (40 to 60 ft/
s) (Ref 5). Lower velocities would be used for a
boiler burning coals that yield heavy loading of
erosive ash, which is usually indicated by high
silica content (Ref 5). For uidized-bed coalred boilers, ue gas streams are laden with not
only y-ash but also sands from the bed. The
convection pass tubes in these boilers are subject
to erosion attack.
Waste-to-energy (WTE) boilers burning municipal and industrial waste can also experience yash erosion problems for their convection pass
tubes (e.g., superheater and economizer). Because of much more corrosive environments in
WTE boilers than coal-red boilers, the boiler
designers have used lower maximum design
velocities in WTE boilers. Combustion Engineering, for example, generally limits the ue gas
velocities entering into the superheater or economizer to 6 to 7.5 m/s (20 to 25 ft/s) (Ref 6).
pg 235
26/10/2007 12:33PM Plate # 0
Babcock & Wilcox sets the maximum design

velocity for superheaters and generating bank
at 9.1 m/s (30 ft/s) (Ref 7). In practice, lower
velocities (3 to 4.6 m/s, or 10 to 15 ft/s) are used
(Ref 7).
Local ow disturbances can sometimes create
conditions that are more conducive to erosion
attack. This is illustrated in a case of the y-ash
erosion in the backpass area of the economizer
when the ue gas stream makes a 90 directional
turn and creates nonuniform ow of gas stream
and uneven distribution of y-ash caused by
centrifugal force (Ref 8).
Other systems where erosion can present a
serious material issue include coal gasication,
combined cycles, and gas turbine. Gas turbines
may involve extremely high gas velocities, up to
250 m/s (820 ft/s), but with small particles
(typically 5 m) (Ref 9).
The erosion problems that have been
encountered in the aforementioned systems all
occur at elevated temperatures where the metal
surface not only experiences mechanical damage
caused by erosion attack but also experiences
oxidation or other high-temperature corrosion
reactions at the same time. Although it is often
referred to as erosion in industry, the damage
reactions, in many cases, may involve a combination of erosion and corrosion, and are more
appropriately referred to as erosion-corrosion.
Since industry continues to use a trial-and-error
approach to solve this practical erosion-related
materials issue, the question is whether the
Fig. 8.1
research data generated can provide industry

with some practical guidance for materials
selection based on either the mechanical properties of the alloy (to select a more erosionresistant alloy) or based on the resistance of the
alloy to oxidation or high-temperature corrosion
(to select a more erosion-corrosion resistant
alloy). The objective of this chapter is to answer
this question by reviewing the relevant erosion
and erosion-corrosion data that are mainly related to a particle-laden gas stream jet impacting
on the metal surface. The erosion or erosioncorrosion that is related to (a) in-bed components
of bubbling uidized-bed boilers and (b) waterwalls of circulating uidized-bed boilers is not
covered in this chapter. Discussion of erosion in
uidized-bed coal-red boilers is presented in
Chapter 10 Coal-Fired Boilers.
8.2 Erosion and Erosion-Corrosion

Finne (Ref 10) described the response of a
material to erosion at room temperature in two
distinctively different modes in terms of the
particle incident angles (or particle impingement
angle). For a ductile material, such as aluminum,
the maximum erosion attack occurs at low incident angles (less than 30) with respect to the
particle impingement direction and the minimum
attack at about 90, as shown in Fig. 8.1 (Ref 10).
On the other hand, the maximum erosion attack
on a brittle material, such as glass, occurs at close
Effect of the particle incidence angle on the room-temperature erosion wear for a ductile material (aluminum) and a brittle
material (glass). Source: Ref 10
pg 236
26/10/2007 12:33PM Plate # 0
Chapter 8: Erosion and Erosion-Corrosion / 237
to 90 (Ref 10). In both cases, no erosion occurs

when the incident angle is 0 indicating the surface of the object is in parallel with the particle
impingement direction.
The erosion model described by Finne (Ref
10) is related to the room-temperature response
of materials. In industrial environments, erosion
problems are most often related to elevated service temperature. Levy (Ref 11) examined the
erosion response of Type 310 as a function of
temperature at incident angles of 30 and 90.
The test environment consisted of nitrogen gas
(N2), thus eliminating the effect of oxidation. The
data indicated that there was signicantly less
erosion attack at a 90 impingement angle than
30 particularly at high temperatures, as shown in
Fig. 8.2 (Ref 11). The erosion behavior of Type
310 in N2 with no oxidation in the erosion
reaction is quite similar to the room-temperature
erosion model of a ductile material proposed by
Finne. Figure 8.2 also shows that erosion attack
increased with advancing temperature at 30
impingement angle. Similar temperature behavior was also observed for Type 304 tested in N2
environment, as shown in Fig. 8.3 (Ref 11).
Carbon steel was also observed to behave

similarly in air, showing increasing erosion (or
erosion-corrosion) with advancing temperature,
as shown in Fig. 8.4 (Ref 12). Shida and Fujikawa (Ref 13) examined the effect of the
impingement angle on the erosion rate for carbon
steel, 1.25Cr-1Mo-0.3V steel, and Type 304 at
300 C (570 F) in an argon environment (i.e.,
elimination of oxidation participation). Their test
results are shown in Fig. 8.5 (Ref 13). The
maximum erosion attack was found to occur at
about a 30 impingement angle for all three
alloys with very little erosion attack at a 90
impingement angle. The test was carried out in an
inert environmentargon. Type 304 suffered
much more erosion attack than both carbon steel
and Cr-Mo-V steel in an inert environment. This
suggests that austenitic stainless steels may not
Fig. 8.2
Fig. 8.3
Erosion rate as a function of temperature in N2 for

Type 310 steel for 30 and 90 impingement angles
(). Source: Ref 11
N2
Ref 11
Erosion rate as a function of temperature in N2 for

Type 304 steel for 30 impingement angle. Source:
pg 237
26/10/2007 12:33PM Plate # 0
pg 238
be as resistant to erosion attack as ferritic steels

when oxidation is not involved. This is further
substantiated by the data that included carbon
steel, 2.25Cr-1Mo steel, and 12Cr steel compared with Type 304 and alloy 800 (Fig. 8.6).
Alloy 800 suffered the worst erosion attack at all
temperatures from room temperature to 650 C
(1200 F) when oxidation was not involved.
Nagarajan and Wright (Ref 14) investigated
iron-, nickel-, and cobalt-base alloys and the
effect of an oxidizing environment on the erosion
resistance of the alloys. Their test environments
included argon (an inert atmosphere) and a
Test conditions:
1018 steel, erodent: 180 m angular Al2O3
v = 10 m/s, = 30,
t = 8 h, loading: 600 g, air
11
Thickness loss, m
10
simulated uidized-bed combustion atmosphere

(N2-3%O2-15%CO2-0.026%SO2), which was
referred to as FBC environment in the paper.
For all three alloy systems, erosion rates were
found to be signicantly higher in the oxidizing
environment than in the inert environment.
Figure 8.7 shows the behavior of Stellite No. 1
(HS 1) and Stellite 6B (HS 6B) at 760 C
(1400 F) at both 30 and 90 impingement
angles. Both alloys No. 1 and No. 6B are wearresistant alloys. With much more carbon and
tungsten, alloy No. 1 is generally more resistant
to wear than alloy 6B. The data, however, show
that both alloys exhibited similar resistance to
erosion attack under the test condition in an inert
environment (Ar) at both 30 and 90 impingement angles. When tested in an oxidizing environment (FBC), both alloys also exhibited similar
erosion rates, although alloy 6B showed only
slightly higher erosion rates at the 90 impingement angle. However, both alloys showed higher
erosion rates in an oxidizing environment (FBC)
than in an inert environment (argon) for both
impingement angles. As for the effect of the
impingement angle, a 90 impingement angle
produced slightly higher erosion rates than did
30 in the oxidizing environment (FBC) while
25
100
200
300
400
500
600
Test temperature, C
Fig. 8.4
200
304
150
Max thickness loss, m/h
Max thickness loss, m/h
Effect of temperature on erosion (or erosioncorrosion) of carbon steel in air at 30 impingement

angle under the particle velocity of 10 m/s (32.8 ft/s) with 180 m
alumina particles. Source: Ref 12
Alloy 800
100
C-steel
304
2.25Cr-1Mo
50
C-steel
12Cr-1Mo-V
100
1.25Cr-1Mo-0.3V
0
Fig. 8.5
10
20
30 40 50 60 70 80
Angle of impingement, degree
0
RT
90
Effect of impingement angle on the erosion of Type

304, carbon steel and Cr-Mo-V steel at 300 C
(570 F) in argon with 120 m/s (394 ft/s) particle velocity,
3
120 g/m particle concentration, and silica particles of 120 m
average particle size. Source: Ref 13
300
500
650
Temperature, C
Fig. 8.6
Effect of temperature on the erosion rates of various

alloys in argon at the impingement angle of 20 with
120 m/s (394 ft/s) particle velocity, 120 g/m3 particle concentration, and silica particles of 120 m average particle size. Source:
Ref 13
26/10/2007 12:33PM Plate # 0
both 30 and 90 angles showed similar erosion

rates in an inert environment (argon).
In the same study by Nagarajan and Wright
(Ref 14), ferritic FeCrAlY (Fe-25Cr-4Al-1Y)
and high-strength nickel-base superalloy IN-100
(Ni-15Co-10Cr-5.5Al-4.7Ti-3Mo) were tested.
For the ferritic FeCrAlY alloy, no signicant
difference in erosion rates was observed between
argon and the FBC oxidizing environment. The
alloy also exhibited similar erosion rates at
both 30 and 90 impingement angles in both
HS 1, argon
HS 6B, argon
HS 1, FBC
(a)
HS 6B, argon
HS 1, argon
(b)
Fig. 8.7
Effect of the test environments (argon and fluidizedbed combustion) on the erosion behavior of HS 1
(Co-30Cr-12W-2.5C) and HS 6B (Co-30Cr-4W-1C) at 760 C
(1400 F), 42.7 m/s (140 ft/s), with 12 m alumina (Al2O3) as
erodent (a) 30 impingement angle and (b) 90 impingement
angle. The argon test environment contained 1% H2, where H2
was used to remove O2, while fluidized-bed combustion (FBC)
environment was a simulated test gas consisting of N2, 3% O2,
15% CO2, and 0.026% SO2. Source: Ref 14
environments (Fig. 8.8). For the nickel-base

superalloy IN-100, signicantly higher erosion
rates were observed in an oxidizing environment
(FBC) than in an inert environment (argon) at
both impingement angles. The most interesting
nding from the study by Nagarajan and Wright
(Ref 14) (which included high particle velocity
of 42.7 m/s, or, 140 ft/s test conditions) was
that the low-strength ferritic FeCrAlY alloy was
signicantly more resistant to erosion than the
high-strength IN-100 superalloy in an oxidizing
environment (Fig. 8.8). FeCrAlY alloy was also
found to be much more erosion resistant than
wear-resistant cobalt-base alloys No. 1 and 6B in
an oxidizing environment.
There was no direct comparison of tensile
strength data between FeCrAlY and IN-100.
However, FeCrAlY (Fe-25Cr-4Al-1Y) is believed to exhibit tensile strengths similar to Type
446 (Fe-25Cr). For example, the ultimate tensile
strength of Kanthal D (Fe-22Cr-4.8Al) at 900 C
(1650 F) is 34.5 MPa (5 ksi) (Ref 15), while
that of Type 446 is 29 MPa (4.2 ksi) (Ref 16). At
700 C (1300 F), Type 446 exhibits ultimate
tensile strength of about 9.0 MPa (1.3 ksi)
(Ref 16) as opposed to 1270 MPa (184 ksi) for
IN-100 (Ref 17). Thus, IN-100 would be expected to exhibit signicantly higher tensile
strength than FeCrAlY at 700 C (1300 F) and
would likely be the same at the test temperature
of 760 C (1400 F) in the erosion study conducted by Nagarajan and Wright. It is thus surprising to see that IN-100 alloy with its tensile
strength approximately more than 30 times
higher than that of FeCrAlY alloy suffered signicantly higher erosion rates than FeCrAlY
under a particle velocity of about 42.7 m/s
(140 ft/s) in an oxidizing environment. It is also
surprising to see that the ferritic FeCrAlY alloy
was much more resistant to erosion than Stellite
alloys No. 1 and No. 6B, which are considered to
be excellent wear-resistant alloys. Both alloys
No. 1 and No. 6B also exhibit much higher tensile strengths compared with FeCrAlY. For
example, at 675 C (1250 F), the ultimate tensile strength of Stellite 6B is about 793 MPa
(115 ksi) (Ref 18). Stellite No. 1 and No. 6B also
show much higher hardness than ferritic Type
446, as shown in Fig. 8.9 (Ref 19).
Pettit and Birks along with their research
group have investigated erosion behavior of
nickel and cobalt (Ref 20, 21), and Cr2O3- and
Al2O3-forming alloys (Ref 22, 23) at elevated
temperatures under very high particle velocities
with 20 m alumina particles. Kang et al.
pg 239
26/10/2007 12:33PM Plate # 0
(Ref 20) indicated that air caused much higher

erosion rates than N2 for nickel under their test
conditions involving particle velocities of 90 and
1
FeCrAlY, FBC
FeCrAlY (2541), argon
Specimen weight change, mg/cm2
1
2
IN-100, argon
3
4
5
6
7
IN-100, FBC
8
9
10
0
10
15
20
25
30
35
40
Erodent impacted, g/cm2

(a)
2
140 m/s (295 and 459 ft/s). They observed that

at 800 C (1470 F) there was essentially no
erosion in N2 at both 90 and 140 m/s (295 and
459 ft/s) (Ref 20). Erosion rates for nickel were
signicantly increased when the test environment was switched from N2 (an inert environment) to air (oxidizing environment). This is
illustrated in Fig. 8.10 (Ref 20). Similar results
were also observed for cobalt (Ref 20). The effect
of the impingement angle on erosion under the
same test condition for both nickel and cobalt
was reported Chang et al. (Ref 21). The data for
nickel are summarized in Fig. 8.11 and for cobalt
in Fig. 8.12 (Ref 21). Erosion tests under the
same conditions also included Cr2O3- and
Al2O3-forming alloys (Ref 22, 23). Figure 8.13
shows the effect of the impingement angle on
erosion rates of MA754 (Ni-20Cr-0.6Y2O3, an
oxide-dispersion-strengthened alloy) at 780 C
(1435 F) under 140 m/s (459 ft/s) particle
velocity (Ref 22). Erosion test results for nickel,
cobalt, MA754, Ni30Cr, CoCrAlY (Co-22Cr11Al-0.17Y), and Ni20Al (Ni-20Al) are summarized in Fig. 8.14 (Ref 22). Both nickel and
cobalt were found to suffer high erosion rates due
to formation of nickel and cobalt oxides,
respectively. Both MA754 and Ni30Cr alloys
that formed Cr2O3 scales showed better resistance to erosion attack, and Al2O3-forming
0
Specimen weight change, mg/cm2
IN-100, argon
2
4
FeCrAlY, argon
FeCrAlY, FBC
6
8
10
12
14
IN-100, FBC
16
0
10
20
30
40
50
60
Erodent impacted, g/cm2
70
80
(b)
Fig. 8.8
Effect of the test environments (argon and fluidizedbed combustion) on the erosion behavior of FeCrAlY
(Fe-25Cr-4Al-1Y) and IN-100 (Ni-15Co-10Cr-5.5Al-4.7Ti-3Mo)
at 760 C (1400 F), 42.7 m/s (140 ft/s), with 12 m alumina
(Al2O3) as erodent (a) 30 impingement angle and (b) 90 impingement angle. The argon test environment contained 1% H2,
where H2 was used to remove residual O2, while fluidized-bed
combustion (FBC) environment was a simulated test gas consisting of N2, 3% O2, 15% CO2, and 0.026% SO2. Source: Ref 14
Temperature, C (F)
Fig. 8.9
Hot hardness data for various alloys as a function of

pg 240
26/10/2007 12:33PM Plate # 0
pg 241
alloys (CoCrAlY and Ni20Al) were much more

resistant to erosion attack presumably due to
formation of aluminum oxide scales. Ives
(Ref 24) conducted erosion tests in a combustion

gas stream using a propane-red abrasive jet rig
at 975 C (1790 F) with 55 m/s (123 mph)
particle velocity of about 135 m (5 mil) SiC
2.5
780 C, 140 m/s
Erosion rate, 105 g/cm2 s
1.5
600 C, 140 m/s
0.5
780 C, 70 m/s
0
0
30
60
Impact angle, degrees
90
Fig. 8.10
Effect of the test environments (N2 and air) on the

erosion behavior of nickel with particle velocities
of 90 and 140 m/s (295 and 459 ft/s) at 90 impingement angle.
Source: Ref 20
Fig. 8.11
Effect of the impingement angle on erosion rate of

nickel in air at 780 C (1435 F) and 140 m/s
(459 ft/s) particle velocity. Source: Ref 21
Fig. 8.12
Fig. 8.13

cobalt in air. Source: Ref 21

MA754 (Ni-20Cr-0.6Y2O3, an oxide-dispersionstrengthened alloy) at 780 C (1435 F) in air under 140 m/s
(459 ft/s) particle velocity. Source: Ref 22
26/10/2007 12:33PM Plate # 0
particles. The erosion test results in terms of the

specimen thickness loss (m/h) are summarized
in Table 8.1. The data indicate that nickel-base
alloys were better than Fe-Ni-Cr alloys, which
were better than Fe-Cr alloys.
Tabakoff et al. (Ref 25) found that aluminized
coatings were capable of providing a signicant
reduction in erosion rates under very high particle velocities and high temperatures in the conditions that were similar to gas turbines. This is
illustrated in Fig. 8.15 and 8.16 (Ref 25). In the
erosion tests conducted by Tabakoff et al., coalash particles consisting of primarily SiO2, Al2O3,
and Fe2O3 with a mean particle diameter of
about 15 m were used. MAR-M246 (Ni-10Co9Cr-10W-2.4Mo-1.5Ta-1.5Ti-5.5Al) and X40
(Co-25Cr-10Ni-7.5W) were tested.
Aluminized coatings tested were C coating
on X40 and N coating on M246. Also included
were platinum-modied aluminized coating,
RT22, and rhodium/platinum-modied aluminide coating, RT22B. Both RT22 and RT22B
(a)
Fig. 8.14
Erosion behavior of nickel, cobalt, Cr2O3-forming

alloys (MA754 and Ni30Cr alloys), and Al2O3forming alloys (CoCrAlY and Ni20Al alloys) in air at 600 and
780 C (1110 and 1435 F) under 140 m/s (459 ft/s) particle
velocity and 30 impingement angle. Source: Ref 22
(b)
Table 8.1 Metal loss rates for alloys tested in

a combustion atmosphere using a propane-fired
abrasive jet rig at 975 C (1790 F), 55 m/s
(123 mph) particle velocity of about 135 m
(5 mils) SiC abrasive particles
Metal loss rate, m/h
Alloy
45 impinging angle
90 impinging angle
671
601
800
310
446
316
304
31
62
80
80
130
160
190
19
29
70
73
90
150
195
Source: Ref 24
(c)
Fig. 8.15
Erosion behavior of M246 (a), X40 (b), and RT22

coating (c) under hot, oxidizing combustion gas
stream at 815 C (1500 F) and 366 m/s (1200 ft/s) with fly-ash as
erodent. Source: Ref 25
pg 242
26/10/2007 12:33PM Plate # 0
were deposited on M246. The thickness was

0.076 mm (3 mils) for aluminized coating C,
N, and RT22B, and 0.127 mm (5 mils) for
RT22. Test results showed that aluminized
coatings signicantly reduced the erosion rates of
the alloy with both platinum- and rhodium/platinum-modied aluminide coatings providing
the most resistance.
The erosion and erosion-corrosion (oxidation)
data presented so far involve very high particle
velocities. Erosion or erosion-corrosion rates of
materials can be a strong function of particle
velocities. Many of the industrial applications
involve particle velocities that are much lower
than the data presented earlier. Wright et al. (Ref
26) investigated the erosion-corrosion behavior
of 10 commercial alloys as a function of particle
velocities (27 to 52 m/s, or 88 to 170 ft/s) at
760 C (1400 F) in an oxidizing environment
(N2-15CO2-3O2-0.03SO2), which was referred
to as FBC gas. Alumina particles (15 m size)
with a loading in the gas stream of about 15,000
ppm (by wt) were used in the tests with each test
being about 4 to 6 h. Test results showed a sharp
transition from low metal-loss rates to a regime
of rapid metal-loss rates at particle velocities in
the range of about 27 to 34 m/s (90 to 110 ft/s)
(Ref 26). The authors termed this regime of high
metal-loss rates as erosion dominated. The test
results on some of the alloys are summarized in
Fig. 8.17 (Ref 26).
Under the particle velocities of 90 and 140 m/s
(295 and 459 ft/s), Kang et al. (Ref 20) and
Chang et al. (Ref 21) observed that signicant
4
X40
Erosion rate, mgm/gm
M246
3
2
C
1
N
RT22B
RT22
Fig. 8.16
Erosion behavior of alloys and coatings under

hot, oxidizing combustion gas stream at 815 C
(1500 F) and 366 m/s (1200 ft/s) with fly-ash as erodent. Source:
Ref 25
plastic deformation developed on the metal

surface. This was caused by the impinging particles for nickel and cobalt tested in air at 800 C
(1470 F). In addition, thin, discontinuous oxide
scales were observed to form on the eroded surface (Ref 20, 21). The impinging particles not
only removed the oxide scales, but also caused
signicant plastic deformation on the metal surface, developing a rippled surface with mounds
and valleys (Ref 20 to 23). However, the repeated
removal and reformation of oxide scales was
attributed to the accelerated erosion-corrosion
rates in oxidizing environments (Ref 20 to 23).
Since signicant plastic deformation was
observed under high particle velocities, the
mechanical properties, hardness, and the microstructure of the alloy may play an important
role in affecting the erosion-dominated erosioncorrosion behavior of the alloy. Various mechanisms for erosion have been proposed in the
literature, such as, cutting by Finne (Ref 10),
cutting and ploughing by Hutchings and Winter
(Ref 27), ake formation by Brown et al.(Ref 28),
platelet formation by Bellman and Levy (Ref 29),
and deformation wear by Bitter (Ref 30).
Hardness has often been used in industry as a
key material property for making materials
selection for resisting wear. Hardness increases
can arise from different hardening mechanisms,
such as cold working, formation of ne-coherent
precipitates, martensitic transformation, and
second-phase hard particles (e.g., eutectic carbides, tungsten carbides, etc.). In industrial trialand-error tests, hardening that resulted from cold
working, coherent precipitation, and martensitic
transformation is somewhat benecial in resisting sliding wear but not erosion or erosioncorrosion. Hardness increases by second-phase,
hard particles appear to be benecial in increasing the erosion resistance of the alloy. Under
erosion-corrosion conditions at high particle
velocities and elevated temperatures, alloys
that exhibit good oxidation resistance (or hightemperature corrosion resistance) and are hardened by second-phase hard particles are likely
to be promising candidates for the application.
Increasing hardness can adversely affect alloy
toughness. Thus, application of a suitable coating
or weld overlay hardfacing material appears to be
a more viable practical approach to the problem.
At lower particle velocities, the erosioncorrosion behavior can be strongly dependent on
the type of oxide scales formed on the alloy.
Steels and low-alloy steels that form iron oxides
can suffer much more erosion-corrosion attack
pg 243
26/10/2007 12:33PM Plate # 0
than stainless steels with adequate chromium for

forming chromium oxide scales. This is illustrated by the test results generated by Levy and
Man (Ref 31) in Fe-Cr alloys with various
chromium contents at 850 C (1560 F) in air
under the particle velocity of 35 m/s (115 ft/s)
with alumina particles (130 m particle size).
The results are summarized in Fig. 8.18 (Ref 31).
Levy and Man (Ref 31) also performed static
oxidation tests on the same alloys in air at the
same temperature (850 C, or 1560 F). Their
oxidation tests showed decreasing weight gain
with increasing chromium content in the alloy, as
shown in Fig. 8.19. This behavior is quite similar
to that generated under erosion-corrosion conditions (Fig. 8.18).
Erosion-corrosion behavior of alloys in a
simulated coal gasication environment was
studied by Agarwal and Howes (Ref 32). The
alloy that was less resistant to suldation attack
in the coal gasication environment was also

found to be less resistant to erosion-corrosion
attack in the same environment. This is illustrated in Fig. 8.20 (Ref 32), showing Type 310
suffering signicantly more erosion-corrosion
attack than alloy 6B in a MPC coal gasication
environment (24H2-18CO-12CO2-39H2O-5CH41NH3-1H2S) tested at 815 C (1500 F) with a
particle velocity of 15.3 m/s (50 ft/s) and using
metallurgical coke as the erodent (300 to 600 m
particle size). [MPC, Materials Property Council
coal-gasication test program (See Chapter 7
Suldation).] Hard particles, such as alumina,
can cause more erosion-corrosion attack than soft
particles, such as coke. This is illustrated in
Fig. 8.21 (Ref 33). In general, the alloys that
are more resistant to suldation are better at
resisting erosion-corrosion in suldizing environments. Figure 8.22 shows the results of
erosion-corrosion tests at 980 C (1800 F) in
(a)
(b)
(c)
(d)
Fig. 8.17
Effect of particle velocity on the erosion-corrosion rate for (a) Type 446, (b) alloy 671, (c) alloy 188, and (d) alloy 6B tested at
760 C (1400 F) in a simulated combustion gas stream (designated as FBC gas, N2-15CO2-3O2-0.03SO2) containing 15 m
alumina particles with a particle loading of 15,000 ppm (by wt). 1 ft = 0.305 m. Source: Ref 26
pg 244
26/10/2007 12:33PM Plate # 0
pg 245
a coal-gasication environment eroded with

metallurgical coke at 30.5 m/s (100 ft/s) particle
velocity as a function of exposure time up to
200 h (Ref 33). Some less-resistant alloys were
found to suffer breakaway erosion-corrosion.
This breakaway erosion-corrosion is quite
similar to breakaway corrosion in reducing
suldizing environments. After 200 h, cobaltbase alloy No. 1 (Co-30Cr-12W-2.5C) was found
to be most resistant among alloys tested, showing
no breakaway erosion-corrosion. Alloy 671
(Ni-48Cr-0.5Ti) also showed no breakaway
erosion-corrosion. Both of these were highchromium alloys, which are known to be highly
120
Corrosion-erosion
110
resistant to suldation attack. (For more information about the suldation behavior of alloys in
coal-gasication environments, readers are
referred to Chapter 7.)
In the regime where particle velocities are
lower, the erosion-corrosion behavior is likely to
be dominated by corrosion. The alloys that are
more resistant to corrosion (i.e., high-temperature corrosion) are generally better at resisting
erosion-corrosion in the same environment. This
may provide a practical guide to materials
selection for applications that are under erosioncorrosion conditions. In this regime, it is believed
that impinging particles cause damage only
to oxides or corrosion products without signicantly affecting the underlying metal. The
most likely scenario in this regime involves the
impinging particles removing oxide scales or
corrosion products and allowing the fresh metal
to be exposed to the corrosive environment.
Oxide scales or corrosion products form and are
then removed by subsequent impinging particles.
This process involves repeated removal of the
corrosion products by eroding particles and
reformation of the fresh corrosion products, thus
resulting in erosion-induced accelerated corrosion. This is best described graphically by a
schematic showing a corrosion mode where
60
Alloy 310
Fig. 8.18
Effect of chromium in Fe-Cr alloys on the erosioncorrosion resistance of the alloys at 850 C
(1560 F) in air with 35 m/s (115 ft/s) particle velocity (130 m
alumina particles). Source: Ref 31
Maximum thickness loss, mils
50
40
30
20
Alloy 6B
10
0
20
40
60
80
100
Impingement angle, degrees
Fig. 8.20
Fig. 8.19
Effect of chromium in Fe-Cr alloys on the oxidation

resistance of the alloys at 850 C (1560 F) in air.
Source: Ref 31
Erosion-corrosion behavior of Type 310 and alloy

6B in MPC coal-gasification environment (24H218CO-12CO2-39H2O-5CH4-1NH3-1H2S) tested for 100 h at
250 psig, 1500 F (815 C), and 50 ft/s, with metallurgical coke
as erodent (300 to 600 m). Source: Ref 32
26/10/2007 12:33PM Plate # 0
Fig. 8.21
Erosion-corrosion behavior of various alloys tested at 980 C (1800 F) for 50 h in MPC coal-gasification environment at
1000 psig with 30.5 m/s (100 ft/s) at 45 impingement angle, using coke and alumina as erodents. S-1: Stellite No. 1: Co30Cr-12W-2.5C; Cru 25: Fe-25Cr-25Ni; LM 1866: Fe-18Cr-6Al-0.6Hf; Alloy 671: Ni-48Cr-0.5Ti; RA333: Ni-25Cr-18Fe-3Mo-3W; Alloy
188: Co-22Cr-22Ni-14W-0.04La; 800AL (aluminized alloy 800); and 310AL (aluminized 310). Source: Ref 33
the corrosion scale reaches a steady state and an

erosion-corrosion mode where the corrosion
scale is continuously reformed after it is repeatedly removed by eroding particles, as shown in
Fig. 8.23 (Ref 34).
8.3 Summary
Erosion and erosion-corrosion behavior of
alloys are reviewed. The data examined are generated mainly in laboratory tests using a jettype apparatus where the metal is impacted by a
particle-laden gas stream. Under conditions
involving very high particle velocities, such
as 100 m/s (328 ft/s) or higher, at elevated
temperatures, oxidizing environments (e.g., air)

signicantly accelerate the erosion-corrosion
rates compared with an inert environment, such
as N2 or argon. Under such high particle velocities, thin, disconnected oxide scales (instead
of a continuous oxide scale) were observed to
form in addition to severe plastic deformation
on the underlying metal that caused the formation of a rippled surface with mounds and valleys. Some limited data suggest alloys that form
Cr2O3 or Al2O3 showing less scaling are better
than those having high scaling rates for resisting
erosion-corrosion attack. The data also suggest
that aluminized coatings are capable of reducing
erosion-corrosion rates in oxidizing environments. Hardness has often been used as an
important material property in resisting wear.
pg 246
26/10/2007 12:33PM Plate # 0
Fig. 8.22
Erosion-corrosion behavior of various alloys tested

at 980 C (1800 F) in MPC coal-gasification
environment, 500 psig, 100 ft/s, coke as erodent. S-1: Stellite No.
1: Co-30Cr-12W-2.5C; Cru 25: Fe-25Cr-25Ni; LM 1866: Fe-18Cr6Al-0.6Hf; Alloy 671: Ni-48Cr-0.5Ti; RA333: Ni-25Cr-18Fe3Mo-3W; Type 310: Fe-25Cr-20Ni. Source: Ref 33
Fig. 8.23
A corrosion mode where the corrosion rate is

diminishing in time and an erosion-corrosion (E-C)
mode where the corrosion products are repeatedly removed by
impinging particles and reformed subsequently. Source: Ref 34
REFERENCES
Limited laboratory data generated so far have

failed to provide practical guidance on the
hardness of the metal in resisting erosioncorrosion attack at elevated temperatures. In
industrial trial-and-error tests, hardening of the
alloy by second-phase, hard particles appears to
be benecial in improving resistance to erosioncorrosion attack at elevated temperatures.
Erosion-corrosion studies conducted by
Wright et al. (Ref 26) suggest that an erosiondominated erosion-corrosion regime occurs
when the particle velocities are in excess of the
range of about 27 to 34 m/s (90 to 110 ft/s), and
below that particle velocity range is a corrosiondominated erosion-corrosion regime. At lower
particle velocities, that is, in a corrosiondominated erosion-corrosion regime, in both
oxidizing and suldizing environments, the data
suggest that the alloys that provide better oxidation or suldation resistance are likely to provide better erosion-corrosion resistance in the
same environment. This may provide a general
practical guide to materials selection for application in the regime where the corrosion dominates the erosion-corrosion reactions.
1. G.E. Moller and C.W. Warren, Survey of

Tube Experience in Ethylene and Olens
Pyrolysis Furnaces: T-5B-6 Task Group
Report, Corrosion/81, NACE
2. D. Jakobi and R. Gommans, Typical Failures
Mater. Corros., Vol 54 (No. 11), 2003, p 881
3. S.C. Stultz and J.B. Kitto, Ed., Steam: Its
Generation and Use, Babcock & Wilcox,
1992, p 33-1
1992, p 20-16
5. J.G. Singer, Ed., Combustion Fossil Power,
Combustion Engineering, Inc. (now Alstom
Power), 1991, p 78
Power), 1991, p 814
1992, p 27-1
Power), 1991, p 2321
pg 247
26/10/2007 12:34PM Plate # 0
9. M.M. Stack, F.H. Stott, and G.C. Wood,

Mater. High Temp., Vol 9 (No. 3), 1991, p 153
10. I. Finne, Wear, 1960, p 87
11. A.V. Levy, Chapter 5: Erosion and ErosionCorrosion
of
Steels
at
Elevated
Temperatures, Solid Particle Erosion and
Erosion-Corrosion of Materials, ASM
International, 1995
12. A.V. Levy, Erosion-Corrosion of Tubing
Steels in Combustion Boiler Environments,
Paper No. 236, Corrosion 93, NACE, 1993
13. Y. Shida and H. Fujikawa, Particle Erosion
Behaviour of Boiler Tube Materials at
Elevated Temperature, Wear, Vol 103, 1985,
p 281
14. V. Nagarajan and I.G. Wright, Inuence of
Oxide Scales on High Temperature Corrosion Erosion Behavior of Alloys, High
Temperature Corrosion, R.A. Rapp, Ed.,
Conf. Proc. (San Diego, CA), March 26,
1981, NACE, 1981, p 398
15. Kanthal Handbook: Resistance Heating
Alloys and Elements for Industrial Furnaces, Kanthal Heating Systems, Hallstahammar, Sweden, p 7
16. AISI Type 446, Alloy Digest, Engineering
Alloys Digest, Inc., May 1982
17. M.J. Donachie and S.J. Donachie, Superalloys: A Technical Guide, 2nd ed., ASM
18. Haynes Stellite Alloy No. 6B, Alloy Digest,
Engineering Alloys Digest, Inc., Sept 1960
19. I.G. Wright, V. Nagarajan, and R.B. Herchenroeder, Some Factors Affecting Solid
Particle Erosion-Corrosion of Metals and
Alloys, Corrosion-Erosion Behavior of
Materials, K. Natesan, Ed., Conf. Proc., Fall
Meeting of The Metallurgical Society of
AIME (St. Louis, MO), October 1718, 1978
20. C.T. Kang, F.S. Pettit, and N. Birks,
Mechanisms in the Simultaneous ErosionOxidation Attack of Nickel and Cobalt
at High Temperatures, Metall. Trans. A,
Vol 18, 1987, p 1785
21. S.L. Chang, F.S. Pettit, and N. Birks, Effect
of Angle of Incidence on the Combined
Erosion-Oxidation Attack of Nickel and
Cobalt, Oxid. Met., Vol 34 (No. 1 & 2), 1990,
p 47
22. S.L. Chang, F.S. Pettit, and N. Birks, Some

Interactions in the Erosion-Oxidation of
Alloys, Oxid. Met., Vol 34 (No. 1 & 2), 1990,
p 71
23. D.M. Rishel, F.S. Pettit, and N. Birks, Some
Principle Mechanisms in the Simultaneous
Erosion and Corrosion Attack of Metals at
High Temperatures, Mater. Sci. Eng., Vol
A143, 1991, p 197
24. L.K. Ives, Erosion of 310 Stainless Steel
at 975 C in Combustion Gas Atmospheres,
Trans. ASME, April 1977, p 126
25. W. Tabakoff, A. Hamed, M. Metwally, and
M. Pasin, High-Temperature Erosion
Resistance of Coatings for Gas Turbine,
Trans. ASME, Vol 114, 1992, p 242
26. I.G. Wright, V. Nagarajan, W.E. Merz, and
J. Stringer, The Kinetics of HighTemperature Erosion-Corrosion of Oxidation-Resistant Alloys, Corrosion/81, NACE,
1981
27. I.M. Hutchings and R.E. Winter, Wear,
Vol 27, 1974, p 121
28. R. Brown, E.J. Jun, and J.W. Edington,
Wear, Vol 70, 1981, p 347
29. R. Bellman and A.V. Levy, Wear, Vol 70,
1981, p 1
30. J.G.A. Bitter, Wear, Vol 6, 1963, p 5
31. A.V. Levy and Y.F. Man, Erosion-Corrosion
Mechanisms and Rates in Fe-Cr Steels,
Wear, Vol 131 (No. 1), 1989, p 39
32. S.C. Agarwal and M.A.H. Howes,
Erosion-Corrosion Materials in HighTemperature Environments: Impingement
Angle Effects in Alloys 310 and 6B under
Simulated Coal Gasication Atmosphere,
J. Mater. Energy Syst., Vol 7 (No. 4), 1986,
p 370
33. M.A.H. Howes, Elevated Temperature
Erosion-Corrosion of Alloys in Suldizing
Gas/Solid Streams: Mechanistic Studies,
Proc. Conf., Corrosion-Erosion-Wear of
Materials at Elevated Temperatures, NACE,
1986, p 230
34. V.K. Sethi and I.G. Wright, Observations
on the Erosion-Oxidation Behavior of
Alloys, in Proc. TMS Conf. on Corrosion
and Particle Erosion, V. Srinivasan and
K. Vedula, Ed., 1986, p 245
pg 248
26/10/2007 2:02PM Plate # 0

DOI: 10.1361/hcma2007p249
pg 249

CHAPTER 9
Hot Corrosion in Gas Turbines

9.1 Introduction
During combustion in the gas turbine, sulfur
from the fuel reacts with sodium chloride from
ingested air at elevated temperatures to form
sodium sulfate. The sodium sulfate then deposits
on the hot-section components, such as nozzle
guide vanes and rotor blades, resulting in accelerated oxidation (or suldation) attack. This
is commonly referred to as hot corrosion.
Sulfur in the fuel is generally limited to 0.3%
for commercial jet engines and to 1.0% for
marine gas turbines (Ref 1). Sodium chloride
comes from seawater (see Table 9.1) (Ref 2).
Seawater is also a source of sulfur. For aircraft
engines, Tschinkel (Ref 1) suggested that runway
dust may be a source of salts.
Gas turbines generally use large amounts
of excess air for combustion (a large fraction of
air, in fact, is also used to cool the combustor),
with a typical air-to-fuel ratio from about 40 to 1
(during takeoff) to 100 to 1 (at cruising speed)
for aircraft gas turbine engines (Ref 2). These
air-to-fuel ratios correspond to about 0.12 to 0.18
mole fractions of oxygen in the combustion zone
(Ref 2). Thus, the combustion gas atmosphere
is highly oxidizing. The sulfur partial pressure
in the atmosphere can be extremely low, varying from 1040 to 1026 atm over the range from
330 to 1230 C (620 to 2240 F) (Ref 2). These
sulfur partial pressures are well below those
necessary to form chromium suldes, which
are frequently observed in alloys suffering hot
corrosion attack.
High-temperature alloys that suffered hot
corrosion attack were generally found to exhibit
both oxidation and suldation. The hot corrosion
morphology is typically characterized by a thick,
porous layer of oxides with the underlying
alloy matrix depleted in chromium, followed
by internal chromium-rich suldes. It is generally believed that the molten sodium sulfate
deposit is required to initiate hot corrosion attack.
The temperature range for hot corrosion attack,
although dependent on alloy composition, is

generally 800 to 950 C (1470 to 1740 F). The
lower threshold temperature is believed to be
the melting temperature of the salt deposit, and
the upper temperature is the salt dew point
(Ref 3). This type of corrosion process is sometimes referred to as Type I hot corrosion to differentiate it from Type II hot corrosion, which
occurs at lower temperatures (typically 670 to
750 C, or 1240 to 1380 F) (Ref 4). Type II
hot corrosion is characterized by pitting attack
with little or no internal attack underneath the
pit (Ref 4). Type II hot corrosion is rarely
observed in aeroengines because the blades
are generally operated at higher temperatures
(Ref 5). However, marine and industrial gas
turbines, which operate at lower temperatures,
can experience low-temperature Type II hot
corrosion.
Type I hot corrosion generally proceeds in two
stages: an incubation period exhibiting a low
corrosion rate, followed by accelerated corrosion
attack. The incubation period is related to the
formation of a protective oxide scale. Initiation
of accelerated corrosion attack is believed to be
related to the breakdown of the protective oxide
scale. Many mechanisms have been proposed
to explain accelerated corrosion attack; the salt
uxing model is probably the most widely
accepted. Oxides can dissolve in Na2SO4 as
anionic species (basic uxing) or cationic species
(acid uxing), depending on the salt composition
(Ref 6). Salt is acidic when it is high in SO3,
and basic when low in SO3. The hot corrosion
Table 9.1
Element
Chlorine
Sodium
Magnesium
Sulfur
Calcium
Potassium
Beryllium
Source: Ref 2
Composition of seawater
Composition, ppm
18,980
10,561
1,272
884
400
380
65
26/10/2007 2:02PM Plate # 0
pg 250
mechanism by salt uxing has been discussed

in detail in Ref 7, 8, and 9. The topic of hot
corrosion has been extensively covered in
several reports and conference proceedings
(Ref 2, 1013).
9.2 Alloys Resistant to Hot Corrosion

Various test methods have been used to study
hot corrosion. Immersion testing (or crucible testing), which was the rst laboratory test method, is
not considered reliable for simulating the gas
turbine environment (Ref 14, 15). The salt-coated
method is quite popular in academia for studying
corrosion mechanisms. Engine manufacturers,
however, use the burner rig test system to determine relative alloy performance ranking. The rig
burns fuel with excess air to produce combustion
gases with continuous injection of a synthetic
sea-salt solution. This type of test system represents the best laboratory apparatus for simulating the gas turbine environment. A special
issue of High Temperature Technology published
in 1989 contained a number of papers discussing
burner rig test procedures (Ref 16). The data
reviewed here are limited to those generated by
burner rig test systems.
9.2.1 High Temperature or Type I
Hot Corrosion
Bergman et al. (Ref 17) studied hot corrosion
resistance of various nickel- and cobalt-base
alloys at temperatures from 870 to 1040 C
(1600 to 1900 F) with 5 ppm sea-salt injection.
Their results are tabulated in Table 9.2. The
Table 9.2
data show a good correlation between alloy

performance and chromium content. Increasing
chromium in the alloy signicantly improves
resistance to hot corrosion. Alloys with 15% Cr
or less are very susceptible to hot corrosion
attack. Cobalt-base alloys are generally better
than nickel-base alloys. This may simply be due
to higher chromium contents in cobalt-base
alloys. One nickel-base alloy (Hastelloy X) with
a chromium level similar to those of cobalt-base
alloys was found to behave similarly to cobaltbase alloys.
Among the alloys tested (Ref 17), alloy X-40
(Co-25Cr-10Ni-7.5W) performed best. This is in
good agreement with the operating experience
obtained by Royal Navy Ship (U.K.), which has
demonstrated the superior hot corrosion resistance of alloy X-40 in a marine environment
(Ref 18). Alloy X-40 was also found to be signicantly better than nickel-base alloys (Ref 19),
such as B-1900, U-700, U-500, and IN738
(Table 9.3). After 240 h, alloy X-40 showed
hardly any corrosion attack, while alloy B-1900
(Ni-10Co-8Cr-6Mo-4.3Ta-6Al-1Ti)
suffered
severe attack. Alloy U-500 (Ni-18Co-19Cr4Mo-2.9Al-2.9Ti) and IN738 (Ni-8.5Co-16Cr1.7Mo-2.6W-1.7Ta-0.9Nb-3.4Al-3.4Ti)
were
similar, suffering only mild attack. Surprisingly,
alloy U-700 (15% Cr) was found to be slightly
worse than alloy B-1900 (8% Cr). Alloy B-1900
along with IN100 (10% Cr) and Nimonic
100 (11% Cr) were considered to be poor in
hot corrosion and suggested that they not be
considered for use without coatings, even in
mildly corrosive environments (Ref 20).
Burner rig tests were conducted (Ref 21)
using residual oil, containing 3% S and 325 ppm
Results of burner rig hot corrosion tests on nickel- and cobalt-base alloys
Loss in sample diameter, mm (mils)
Alloy
SM-200
IN100
SEL-15
IN713
U-700
SEL
U-500
Rene 41
Hastelloy alloy X
L-605 (alloy 25)
WI-52
MM-509
SM-302
X-40
Chromium content
in alloy, %
870 C (1600 F)
500 h
950 C (1750 F)
1000 h
980 C (1800 F)
1000 h
1040 C (1900 F)
1000 h
9.0
10.0
11.0
13.0
14.8
15.0
18.5
19.0
22.0
20.0
21.0
21.5
21.5
25.0
1.6 (64.4)
3.3+ (130+)
3.3+ (130+)
3.3+ (130+)
1.7+ (66+)
1.2 (45.8)
0.2 (7.6)
0.3 (10.3)
0.5 (21.4)
0.14 (5.4)
0.11 (4.2)
3.3+ (130+)
3.3+ (130+)
3.3+ (130+)
2.0+ (77+)
1.6 (63.9)
1.3 (51.8)
0.8 (31.7)
0.3 (12.0)
0.4 (15.3)
0.5 (18.2)
0.3 (10.9)
0.3 (10.0)
0.3 (11.6)
0.3 (11.4)
0.7 (29.3)
0.8 (30.8)
0.4 (15.2)
0.3 (11.3)
Note: 5 ppm sea salt injection. Source: Ref 17
1.1 (41.9)
1.9 (73.9)
0.8 (31.8)
0.6 (23.1)
0.5 (18.5)
26/10/2007 2:02PM Plate # 0
Chapter 9:
NaCl (equivalent to 5 ppm NaCl in air), at

870 C (1600 F) for 600 h on several cobaltbase alloys, which were X-45 (Co-25Cr-10Ni7.5W), MAR-M302 (Co-21.5Cr-10W-9Ta-0.2Zr),
MAR-M509 (Co-21.5Cr-10Ni-7W-3.5Ta-0.2T0.5Zr), S-816 (Co-20Cr-20Ni-4W-4Nb-4Mo),
FSX-418 (Co-30Cr-10Ni-7W-0.15Y), and FSX414 (Co-30Cr-10Ni-7W). The test results are
shown in Fig. 9.1. All six cobalt-base alloys with
chromium varying from 20 to 30% suffered little
corrosion attack (about 0.04 to 0.12 mm, or
0.002 to 0.005 in., or 2 to 5 mils). Under the
same test condition, Udimet 700 (Ni-15Cr18.5Co-5.2Mo-5.3Al-3.5T) suffered about 0.76
mm (0.03 in., or 30 mils) of attack (Ref 21).
Figure 9.2 summarizes the data for a group of
nickel- and cobalt-base alloys at 870 to 1040 C
(1600 to 1900 F) (Ref 22). Alloy U-700 was
found to be inferior to IN713 at 950 C
(1750 F). This was contrary to eld experience.
Alloy U-700 has served most reliably in aircraft
jet engines, whereas IN713 has suffered severe
hot corrosion in many applications (Ref 22).
The authors attributed the high corrosion rate
of alloy U-700 to the low chromium content in
this heat (i.e., 13.6% versus 15.0% for regular
heats).
A systematic study was conducted (Ref 22)
to determine the effects of alloying elements on
hot corrosion resistance. In the Ni-10Co-15Cr4Al-2Ti system, decreasing chromium from 25
to 10% resulted in increases in hot corrosion
attack (Fig. 9.3). The data also suggest that
decreasing aluminum while increasing titanium
improves hot corrosion resistance. Furthermore,
addition of 8% W to Ni-10Co-15Cr-4Al-2Ti
alloy resulted in no apparent change in hot
corrosion resistance. In binary and ternary alloy
systems of nickel- and cobalt-base alloys, these
authors further observed the effectiveness of
chromium in improving hot corrosion resistance
at 910, 950, and 1040 C (1675, 1750, and
1900 F) (Ref 22). Results are shown in Fig. 9.4.
pg 251
Hot Corrosion in Gas Turbines / 251
In examining the effect of the third alloying

element in Ni-Cr alloys, they found that (Ref 22):
Tungsten (8%) showed little effect at 910

and 950 C (1675 and 1750 F), but a
slightly detrimental effect at 1040 C
(1900 F).
Cobalt was only slightly benecial.
Molybdenum was detrimental at 1040 C
(1900 F), but had little effect at lower temperatures.
Titanium (5%) showed signicant improvement at 1040 C (1900 F), but little effect at
lower temperatures.
Aluminum (8%) was detrimental, causing
severe hot corrosion attack at 1040 C
(1900 F).
For Co-25Cr alloys, the effect of the third
alloying element was summarized as (Ref 22):
Tungsten (8%) was detrimental at 1040 C

(1900 F), with little effect at lower temperatures.
ATTACK, mm per side

0
0.04
0.08
0.12
FSX 414
FSX418
S816
MARM 509
MARM 302
X45
Table 9.3 Results of burner rig tests at 874 C

(1605 F) for nickel- and cobalt-base alloys
ATTACK, mils per side
Penetration depth, mm/1000 h (mils/1000 h)

Exposure
time, h
100
170
240
B-1900
U-700
U-500
IN738
X-40
2.8 (111)
2.5 (97)
2.1 (83)
3.3 (129)
0.7 (29)
0.5 (20)
0.9 (35)
Slight
Note: Diesel fuel containing 1.0% S, 125 ppm Na, 15 ppm Mg, 4.8 ppm Ca,
4.1 ppm K, and 225 ppm Cl, air-to-fuel ratio was 50 : 1, and 100 h cycle. Source:
Ref 19
Surface loss
Maximum penetration
Fig. 9.1
Relative hot corrosion resistance of cobalt-base alloys

obtained from burner rig tests using 3% S residual oil
and 325 ppm NaCl in fuel (equivalent to 5 ppm NaCl in air) at
870 C (1600 F) for 600 h. Source: Beltran (Ref 21)
26/10/2007 2:02PM Plate # 0
pg 252
Molybdenum (6%) was detrimental at 950

and 1040 C (1750 and 1900 F), with little
effect at 910 C (1675 F).
Tantalum (7%) and nickel (10%) showed
little effect.
Burner rig tests were conducted (Ref 23) at
900 C (1650 F) on several wrought superalloys and nickel aluminides. The combustion
gas stream was generated by using No. 2 fuel oil
containing about 0.4 wt% S with an air-to-fuel
ratio of 35 to 1 and injection of either 5 or 50 ppm
sea salt into the combustion gas stream. The
specimens were loaded in a carousel, that rotated
at 30 rpm during testing to ensure that all the

specimens were subjected to the same test condition. The specimens were cycled out of the
combustion gas stream once every hour for
2 min, during which time the specimens were
cooled by forced air (fan cool) to less than
205 C (400 F). Superalloys tested were
alloy X (Ni-22Cr-18.5Fe-9Mo-0.5W), alloy S
(Ni-15.5Cr-14.5Mo-0.05La), alloy 230 (Ni-22Cr14W-2Mo-0.02La), alloy 625 (Ni-21.5Cr-9Mo3.6Nb), alloy 188 (Co-22Cr-22Ni-14W-0.04La),
alloy 25 (Co-20Cr-10Ni-15W), and alloy 150
(Co-27Cr-18Fe). Two nickel aluminides,
IC-50 (Ni-11.3Al-0.6Zr-0.02B) and IC-218
Surface loss
Maximum
penetration
1600 F (870 C)
1750 F (950 C)
1900 F (1040 C)
SL- Surface loss
MP- Maximum penetration
Alloy
U700
75.6 SL
92.4 MP
80.2
SM200
SEL
IN713
U500
X
WI52
109.2
SM302
X45
10
(250)
20
(500)
30
(750)
40
(1000)
50
(1250)
60
(1500)
70
(1750)
Loss in diameter, mils (m)
Fig. 9.2
Relative hot corrosion resistance of nickel- and cobalt-base alloys obtained from burner rig tests at 870, 950, and 1040 C
(1600, 1750, and 1900 F) for 100 h, using 1% S diesel fuel, 30:1 air-to-fuel ratio, and 200 ppm sea-salt injection. Source:
Bergman et al. (Ref 22)
26/10/2007 2:02PM Plate # 0
Chapter 9:
(Ni-7.8Cr-8.5Al-0.8Zr-0.02B), which were

developed by Oak Ridge National Laboratory,
were included in the test program.
The results of tests at 900 C (1650 F) for
200 h with 50 ppm sea salt are summarized in
Table 9.4 (Ref 23). Both IC-50 and IC-218 nickel
aluminides suffered severe hot corrosion attack
after 200 h at 900 C (1650 F) with 50 ppm sea
salt being injected into the combustion gas
stream. Scanning electron microscopy with
energy-dispersive x-ray spectroscopy (SEM/
EDX) analysis showed that both nickel aluminides exhibited porous nickel or nickel-rich
oxides with nickel sulde penetrating through
the remaining metal (Ref 23). Figure 9.5 shows
the cross section of a corroded IC-218 specimen
after hot corrosion burner rig testing at 900 C
pg 253
(1650 F) for 200 h with 50 ppm sea salt, revealing the formation of nickel oxides and nickel
suldes. SEM/EDX analysis showed that a thin,
protective chromium-rich oxide scale formed on
alloys X, 230, and 188.
Alloy 25, although exhibiting little weight
change (Table 9.4), showed evidence of initial
breakdown of the chromium-rich oxide scale.
SEM/EDX analysis revealed the formation of
cobalt-rich oxide nodules on the outer oxide
scale on alloy 25. This indicated the initiation of
the breakaway corrosion for alloy 25 after 200 h
at 900 C (1650 F) with 50 ppm sea salt. Longterm test results under the same test condition
clearly showed that alloy 25 suffered severe hot
corrosion in excess of 200 h of testing, as shown
in Fig. 9.6 (Ref 23).
Loss in diameters, m
0
Alloy
250
500
750
1000
1250
1500
1750
2000
Ni-10Co-0.5Nb-4Al-2Ti-25Cr TEL-1
140
Ni-10Co-0.5Nb-8W-4Ai-2Ti-15Cr TEL-8
Ni-10Co-0.5Nb-4Mo-15Cr TEL-9
Ni-0Co-0.5Nb-2Mo-4W-15Cr TEL-10
Ni-25Co-0.5Nb-15Cr TEL-11
Ni-15Co-4Mo-3.7Al-1.8Ti-19 Cr MELINI-1
Ni-15Co-3.5Al-1.8Ti-4.5 Re-19Cr MELINI-2
Ni-16Co-3.6Al-2.1Ti-21Cr MELINI-3
Surface
loss
Ni-15Co-4Mo-2.9Al-1.8Ti-0.2Y-19Cr MELNI-4
Maximum
penetration
950 C (1750 F)
1040 C (1900 F)
10
20
30
40
50
60
70
80
Loss in diameters, mils
Fig. 9.3
(Ref 22)
Relative hot corrosion resistance of experimental alloys obtained from burner rig tests at 950 and 1040 C (1750 and
1900 F) for 100 h, using 1% S diesel fuel, 30:1 air-to-fuel ratio, and 200 ppm sea-salt injection. Source: Bergman et al.
26/10/2007 2:02PM Plate # 0
Figure 9.6 also shows alloys 230 and 188

exhibiting very little weight change for exposure
time up to 1000 h. The results of the 1000 h tests
Maximum penetration
measurements
1675 F (910 C)
1750 F (950 C)
1900 F (1040 C)
CP-Complete penetration
through the specimen
Alloy
RL-1
RL-2
RL-3
RL-4
RL-5
RL-6
Ni-15Cr
Ni-25Cr
Ni-15Cr-8W
Ni-15Cr-24Co
130CP
Ni-15Cr-6Mo
9.2.2 Low-Temperature or Type II

Hot Corrosion
Ni-15Cr-5Ti
RL-9
RL-10
130CP
Ni-10Cr
RL-7
RL-8
UNALLOYED NICKEL
90
Ni-15Cr-8Al
UNALLOYED COBALT
RL-11
130CP
Co-25Cr
RL-12
Co-25Cr-8W
RL-13
130CP
Co-25Cr-6Mo
RL-14
Co-25Cr-7Ta
RL-15
Co-25Cr-10Ni
20
30
40
50
60
70
10
(250) (500) (750) (1000) (1250) (1500) (1750)
Loss in diameter, mils (m)
Fig. 9.4
are summarized in Table 9.5 (Ref 23). Three

nickel-base alloys (alloys S, X, and 625) and one
cobalt-base alloy (alloy 25) were completely
corroded before the test reached 1000 h, while
nickel-base alloy 230 and cobalt-base alloys 150
and 188 exhibited little corrosion attack after
1000 h. Under the same test conditions in the
same burner rig, nickel-base alloy HR-160 with
29% Co, 28% Cr and 2.75% Si was found to
perform as well as alloys 230, 150, and 188.
Figure 9.7 shows the conditions of the test specimens comparing HR-160 with other wrought
alloys (Ref 24). Another burner rig test was
conducted (Ref 23) with 5 ppm sea salt at 900 C
(1650 F) under the same combustion conditions
(i.e., No. 2 fuel oil with 0.4% S and 35-to-1 airto-fuel ratio). The results are summarized in
Table 9.6. Even at this low level of sea salt
(5 ppm) in the combustion gas stream, cobaltbase alloy 25 continued to exhibit very poor hot
corrosion resistance compared with some nickelbase alloys.
Relative hot corrosion resistance of experimental

alloys obtained from burner rig tests at 910, 950, and
1040 C (1675, 1750, and 1900 F) for 100 h, using 1% S diesel
fuel, 30:1 air-to-fuel ratio, and 200 ppm sea salt injection. Source:
Bergman et al. (Ref 22)
Low-temperature or Type II hot corrosion

has been observed at temperatures lower than
the temperature range where Type I hot corrosion
has been encountered. Severe hot corrosion
of alloy S590 and Nimonic 80A after several
thousand hours of operation in a gas turbine that
burned blast-furnace gas with the 700 to 730 C
(1290 to 1345 F) turbine entry temperature was
reported in Ref 25. In 1976, a new form of hot
corrosion attack of gas turbine airfoil materials
in a marine gas turbine was reported (Ref 26).
The rst-stage turbine blades coated with a
CoCrAlY coating, which had exhibited satisfactory hot corrosion resistance for metal temperatures in the range 800 to 1000 C (1470 to
1830 F), were found to suffer corrosion attack
for metal temperatures at about 600 to 730 C
(1110 to 1345 F) (Ref 26). The corrosion
Table 9.4 Results of burner rig hot corrosion tests at 900 C (1650 F) for 200 h with 50 ppm
sea salt with specimens being cycled once every hour
Alloy
X
230
25
188
IC-50
IC-218
Total depth of attack(a), mm (mils)
0.76
1.35
1.62
0.93
72
83
0.02 (0.6)
0.02 (0.8)
0.02 (0.9)
0.02 (0.7)
>0.72 (28.3), completely corroded
0.07 (2.8)
0.06 (2.4)
0.07 (2.8)
0.04 (1.6)
pg 254
26/10/2007 2:02PM Plate # 0
Chapter 9:
3
2
1
100 m
Fig. 9.5
Scanning electron backscattered image showing the cross section of a corroded IC-218 nickel aluminide specimen after hot
corrosion burner rig testing at 900 C (1650 F) for 200 h with 50 ppm sea salt using No. 2 fuel oil (0.4% S) for combustion at
35:1 air-to-fuel ratio. The results (wt%) of EDX analysis are: 1: 100% Ni; 2: 74% Ni, 26% S; 3: 100% Ni; 4: 88% Ni, 8% Cr, 4% Al; and 5:
98% Ni, 2% Al. Areas 1, 4, and 5 were essentially nickel oxides, area 2 was nickel sulfide, and area 3 was pure nickel. Courtesy of Haynes
International, Inc.
products formed on the CoCrAlY coating were

found to contain CoSO4 and NiSO4 (Ref 26). It
was proposed (Ref 27) that the mechanism of
low-temperature hot corrosion attack of a
CoCrAlY coating involved the formation of the
low melting Na2SO4-CoSO4 eutectic (melting
point of 565 C, or 1045 F). Nickel-base alloys
were, in general, more resistant to Type II hot
corrosion than cobalt-base alloys as found in Ref
28. It was also found (Ref 29) that NiCrAlY and
NiCoCrAlY coatings were, in general, more
resistant than CoCrAlY coatings.
Increases in chromium content in superalloys
and coatings provided signicant increases in
low-temperature hot corrosion resistance
(750 C, or 1380 F) for these materials
(Ref 30). Both IN939 (23% Cr) and NiCrAlY
coating (39% Cr) were found to exhibit good
low-temperature hot corrosion resistance.
NiCrAlY coatings with 26, 34, and 42% Cr were
tested and found signicant resistance to lowtemperature hot corrosion for coatings with only
34 and 42% Cr (Ref 31). MCrAlY coatings
(M = Co and/or Ni) with 30 to 35% Cr were
tested (Ref 32). All of the coatings showed
improved resistance to low-temperature hot
corrosion (705 C, or 1300 F) compared with

CoCrAlY coating with about 20% Cr. A critical
chromium content of no less than 37% was
required for cobalt-base coatings to provide
resistance to both low- and high-temperature hot
corrosion (Ref 33).
9.3 Summary
High-temperature or Type I hot corrosion
generally occurs in the temperature range of
800 to 950 C (1470 to 1740 F). It is believed
that the molten sodium sulfate deposit is required to initiate hot corrosion attack. The Type I
hot corrosion morphology is typically characterized by a thick, porous layer of oxides with the
underlying alloy matrix depleted in chromium,
followed by internal chromium-rich suldes.
Low-temperature or Type II hot corrosion generally occurs in the temperature range of 670
to 750 C (1238 to 1382 F). Type II hot corrosion is characterized by pitting attack with
little or no internal attack underneath the pit.
Cobalt-base alloys are more susceptible to Type
II hot corrosion, which generally involves
pg 255
26/10/2007 2:02PM Plate # 0
20
Haynes alloy 230
10
Haynes alloy 188
Table 9.5 Results of burner rig hot corrosion

tests at 900 C (1650 F) for 1000 h with 50 ppm
sea salt with specimens being cycled once every
hour
10
Alloy
20
Weight change,
mg/cm2
625
230
25
11.5
80
150
188
10.0
9.9
90
30
40
50
Total depth of attack(a), mm (mils)
60
70
Completely corroded in 350 h

(1.3 mm, or 49.2 mil thick specimen)
0.11 (4.4)
0.13 (5.2)
0.08 (3.2)
100
Haynes alloy 25
110
120
130
140
150
160
170
180
0 100 200 300 400 500 600 700 800 9001000

Exposure time, h
Fig. 9.6
Results of burner rig tests at 900 C (1650 F) with

50 ppm sea salt using No. 2 fuel oil (0.4% S) for
combustion at 35:1 air-to-fuel ratio for alloys 230, 188, and 25.
Source: Lai et al. (Ref 23)
Na2SO4 and CoSO4. Increasing chromium in

alloys or coatings will improve the resistance
of the material to both Type I and Type II hot
corrosion attack.
REFERENCES
1. J.G. Tschinkel, Corrosion, Vol 28 (No. 5),

1972, p 161
2. J. Stringer, Hot Corrosion in Gas Turbines, Report MCIC-72-08, Battelle
Columbus Laboratories, Columbus, OH,
1972
3. J. Stringer, High Temperature Corrosion in
Energy Systems, M.F. Rothman, Ed., The
Metallurgical Society of AIME, 1985, p 3
4. J. Stringer, Coatings in the Electricity Supply Industry: Past, Present, and Opportunities for the Future, Surf. Coat. Technol.,
Vol 108/109, 1998, p 1
5. A.K. Koul, J.P. Immarigeon, R.V. Dainty,

and P.C. Patnaik, Degradation of High Performance Aero-Engine Turbine Blades,
Advanced Materials and Coatings for
Combustion Turbines, V.P. Swaminathan
and N.S. Cheruvu, Ed., ASM International,
1994, p 69
6. G.H. Meier, High Temperature Corrosion 2
Advanced Materials and Coatings (Les
Embiez, France), May 2226, 1989, Elsevier Science, 1989, p 1
7. J.A. Goebel, F.S. Pettit, and G.W. Goward,
Metall. Trans., Vol 4, 1973, p 261
8. J. Stringer, Am. Rev. Mater. Sci., Vol 7, 1977,
p 477
9. R.A. Rapp, Corrosion, Vol 42 (No. 10),
1986, p 568
10. J. Stringer, R.I. Jaffee, and T.F. Kearns, Ed.,
High Temperature Corrosion of Aerospace
Alloys, Advisory Group for Aerospace
Research and Development, North Atlantic
Treaty Organizations, AGARD-CP-120,
Harford House, London, 1973
11. J.W. Fairbanks and I. Machlin, Ed., Proc.
1974 Gas Turbine Materials in The Marine
Environment Conf., MCIC-75-27, Battelle
Columbus Laboratories, 1974
12. Hot Corrosion Problems Associated with
Gas Turbines, STP 421, ASTM, 1967
13. A.B. Hart and A.J.B. Cutler, Ed., Deposition
and Corrosion in Gas Turbines, Applied
Science, London, 1973
14. J.F.G. Conde and G.C. Booth, Deposition
and Corrosion in Gas Turbines, A.B. Hart
and A.J.B. Cutler, Ed., John Wiley & Sons,
1973, p 278
pg 256
26/10/2007 2:02PM Plate # 0
Chapter 9:
pg 257
310SS
RA330 alloy
Alloy 800H
Alloy 625
HR160 alloy
1000 h
476 h
1000 h
940 h
1000 h
Fig. 9.7
Test specimens alloy HR-160, 625, 800H, RA330, and Type 310 at 900 C (1650 F) in the combustion gas stream generated
by a burner rig using No. 2 fuel oil (0.4% S) for combustion at 35:1 air-to-fuel ratio and with injection of 50 ppm sea salt into the
combustion gas stream. During testing, specimens were cycled once every hour. Source: Ref 24
Table 9.6 Results of burner rig hot corrosion

tests at 900 C (1650 F) for 1000 h with 5 ppm
sea salt with specimens being cycled once
every hour
Alloy
Weight change,
mg/cm2
Metal loss,
mm (mils)
Total depth of attack(a),

mm (mils)
X
230
625
25
0.24
0.79
5.87
0.04 (1.6)
0.03 (1.2)
0.05 (1.9)
188
1.09
0.02 (0.8)
0.14 (5.5)
0.13 (5.1)
0.14 (5.3)
Completely corroded
(1.09 mm, or 43 mils)
0.07 (2.8)
15. M.J. Donachie, R.A. Sprague, R.N. Russell,

K.G. Boll, and E.F. Bradley, Hot Corrosion
Problems Associated with Gas Turbines,
STP 421, ASTM, 1967, p 85
16. High Temperature Technology, Special Issue
on Hot-Salt Corrosion Standards Test Procedures and Performance, Vol 7 (No. 4),
Nov 1989
17. P.A. Bergman, A.M. Beltran, and C.T. Sims,
Development of Hot Corrosion-Resistant
Alloys for Marine Gas Turbine Service,
Final Summary Report to Marine Engineering Lab., Contract N600 (61533) 65661,
Navy Ship R&D Center, Annapolis, MD,
Oct 1, 1967
18. J. Clelland, A.F. Taylor, and L. Wortley,
Proc. 1974 Gas Turbine Materials in the
Marine Environment Conf., MCIC-75-27,
J.W. Fairbanks and I. Machlin, Ed., Battelle
Columbus Laboratories, 1974, p 397
19. M.J. Zetlmeisl, D.F. Laurence, and K.J.

McCarthy, Mater. Perform., June 1984, p 41
20. J. Stringer, Proc. Symp. Properties of High
Temperature Alloys with Emphasis on
Environmental Effects, Z.A. Foroulis and
F.S. Pettit, Ed., The Electrochemical Society,
1976, p 513
21. A.M. Beltran, Cobalt, Vol 46, 1970, p 3
22. P.A. Bergman, C.T. Sims, and A.M. Beltran,
Hot Corrosion Problems Associated with
Gas Turbines, STP 421, ASTM, 1967,
p 38
23. G.Y. Lai, J.J. Barnes, and J.E. Barnes, A
Burner Rig Investigation of the Hot Corrosion Behavior of Several Wrought Superalloys and Intermetallics, Paper 91-GT-21,
International Gas Turbine and Aeroengine
Congress and Exposition (Orlando, FL),
June 36, 1991
24. Haynes HR-160 Alloy, H-3129A, Haynes
International, Inc., Kokomo, IN
25. W. Moller, Deposition and Corrosion in Gas
Turbines, A.B. Hart and A.J.B. Cutler, Ed.,
John Wiley and Sons, 1973, p 1
26. D.J. Wortman, R.E. Fryxell, and I.I. Bessen,
A Theory for Accelerated Turbine Corrosion
at Intermediate Temperatures, Proc. Third
Conference on Gas Turbine Materials in a
Marine Environment, Session V, Paper 11
(Bath, England), 1976
27. K.L. Luthra, Metall. Trans. A, Vol 13, 1982,
p 1853
28. D.J. Wortman, R.E. Fryxell, K.L. Luthra,
and P.A. Bergman, Proc. Fourth US/UK
Conf. on Gas Turbine Materials in a Marine
26/10/2007 2:02PM Plate # 0
Environment (Annapolis, MD), U.S. Naval

Academy, 1979, p 317
29. L.F. Aprigliano, Proc. Fourth US/UK Conf.
on Gas Turbine Materials in a Marine
Environment (Annapolis, MD), U.S. Naval
Academy, 1979, p 151
30. A.R. Taylor, B.A. Wareham, G.C. Booth,
and J.F. Conde, Low and High Pressure Rig
Evaluation of Materials and Coatings, Proc.
Third Conference on Gas Turbine Materials
in a Marine Environment, Session III, Paper
No. 3 (Bath, England), 1976
31. J.F.G. Conde, G.C. Booth, A.F. Taylor,
and C.G. McGreath, Hot Corrosion in
Marine Gas Turbines, Proc. Conf. High Temperature Alloys for Gas Turbines, COST 50,
1982, p 237
32. J.A. Goebel, Advanced Coating Development for Industrial/Utility Gas Turbine
Engines, Proc. First Conf. on Advanced
Materials for Alternative Fuel Capable
Directly Fired Heat Engines (Castine, ME),
Department of EnergyElectric Power
Research Institute, 1979, p 473
33. K.L. Luthra and J.H. Wood, High Chromium Cobalt-Base Coatings for Low Temperature Hot Corrosion, Thin Solid Films,
Vol 119, 1984, p 271
pg 258

DOI: 10.1361/hcma2007p259
30/10/2007 4:01PM Plate # 0

CHAPTER 10
Coal-Fired Boilers
10.1 Introduction
A coal-red electric power generating station
is a highly complex plant. A major piece of the
equipment in the power plant is the boiler, which
generates steam that is then delivered to the turbine for generation of electricity. This chapter
reviews high-temperature corrosion problems
associated with utility boilers. A brief description
of boilers is included to help readers understand
some of the basics in the boiler. A brief discussion of coal is also presented since different types
of coal may present different issues in combustion, thus resulting in different types of materials
issues and corrosion. More detailed discussion
and coverage of boilers and the fuel can be found
in SteamIts Generation and Use (Babcock &
Wilcox, Ref 1) and Combustion Fossil Power
(Combustion Engineering, Ref 2).
10.2 Boiler Description

The primary function of the boiler is to generate steam for the turbine to convert to electricity. The steam is converted from water in a
furnace where heat is generated by combustion
of coal. The modern furnace in a square or
rectangular horizontal cross section is enclosed
with water-cooled tubes with a narrow plate
(commonly referred to as a membrane) connected between two adjacent tubes. When the
furnace is heated, the water under high pressure
inside these water-cooled tubes (also referred
to as waterwalls) is converted to steam. Some
boilers use a tangent tube design for waterwalls
where no membranes are used to connect adjacent tubes.
In terms of the operating pressure of the water/
steam in the waterwalls, there are two classications of boilers: subcritical and supercritical
units. The subcritical unit is the boiler that
operates at pressures below the 218.2 atm
(3208 psig) critical point, while the supercritical

unit operates at pressures higher than 218.2 atm
(3208 psig).
When heated at atmosphere pressure, water
boils at 100 C (212 F). As the pressure of the
water increases, the boiling (saturation) temperature also increases. The saturation temperature becomes constant once the water is boiling at
a given pressure. Accordingly, the saturation
temperature increases as the pressure increases.
Thus, for subcritical units, the saturation temperature is established by a given operating
waterwall water/steam pressure. For example,
for the pressure of 170 atm (2500 psig), the
saturated temperature is 353.4 C (668.11 F)
(Ref 3). In a subcritical unit, the water (slightly
undercooled) in the waterwall tubes rises from
the bottom of the waterwall. As the water absorbs
heat, bubbles form. More bubbles form as the
water continues to rise in the waterwall tubes as
shown in Figs. 10.1 and 10.2 (Ref 1). At the top
of the waterwall tubes, a mixture of steam and
water leaves the waterwall tubes and enters the
steam drum where steam is separated to be heated
further in the convection section for higher
temperatures prior to delivery to the turbine for
electricity generation. Water at the bottom of the
steam drum is then mixed with the replacement
water (feedwater) and is returned to the bottom of
the boiler through an unheated downcomer to
reenter the waterwall tubes.
In a supercritical unit, water at pressures above
218.2 atm (3208 psig) does not boil. The temperature rises steadily, and the water changes
completely to steam. There is no saturation temperature and no need for a steam drum. The
supercritical unit operates on a once-through
ow system with water owing in the furnace
waterwall tubes from the bottom of the furnace
and changing to 100% steam with a continuous
temperature increase. The steam then leaves the
furnace waterwall tubes and is further heated
in superheaters before entering into the steam
turbines.
pg 259
30/10/2007 4:01PM Plate # 0
There are several methods for burning coal.

Firing with pulverized coal has been the most
dominant method for utility boilers. Coal is
burned as ne powers suspended in the furnace,
and almost all types of coal from anthracite
to lignite can be burned by pulverized ring
(Ref 1). Pulverized coal in particle sizes of
50 m diameter or smaller can be completely
combusted in a matter of 1 to 2 s (Ref 1).
Babcock & Wilcox developed the Cyclone furnace for ring coal grades that have a low fusion
temperature and are not suitable for pulverized
coal ring because of potentially forming a
molten slag, thus developing a severe slagging
problem in superheaters (Ref 1). In Cyclone
boilers, the Cyclone barrels burn coal in such a
way that most of the coal ash is captured to form a
molten slag that coats the inside surface of the
Cyclone barrels (Ref 1). The combustion ue gas
from the Cyclone barrels then enters the main
furnace to generate steam.
Stoker-ring boilers are very versatile for
burning a wide range of solid fuels including
various types of coals, municipal waste, wood
waste, and other types of biomass fuels. In a
stoker-ring boiler, a mechanical stoker at the
Fig. 10.1
Water enters the waterwall tubes at the furnace

bottom and turns into a mixture of water and steam
that leaves the waterwall tubes at the top (C) and enters the steam
drum where steam and water is separated. Water mixed with the
replacement water (A) is returned to the waterwall tubes at the
furnace bottom (B). Source: Ref 1. Courtesy of Babcock & Wilcox
bottom of the furnace is used to feed fuel, such as

coal, onto a grate where coal is burnt. Fluidizedbed combustion (FBC) technology is considered
to be an emerging technology for power generation. In a FBC boiler, combustion of coal (or
other type of fuel) takes place in a uidized bed.
There are two types of bed: bubbling bed and
circulating bed. The bed temperature is typically
maintained in a range of 816 to 900 C (1500 to
1650 F) for maximum efciency of both combustion and sulfur capture. The bed consists of
essentially sand and limestone or dolomite. Air
enters the bed from the air distributor plate at the
bottom and renders the bed into a uidization
condition. Limestone or dolomite is added to the
bed as a sorbent to capture SO2 from the ue gas
by forming calcium sulfate (CaSO4). In a bubbling bed, an evaporator tube bundle is typically
immersed in the bed to help maintain the
designed bed temperature. Figure 10.3 shows a
schematic of a bubbling uidized-bed boiler
(Ref 1).
10.3 Coal and Coal Ash

Coal is generally classied by rank, which
indicates the geological formation history of the
coal. There are four different types of coal. The
youngest, or lowest-rank coal, is lignite, which is
then followed by an older, or higher-rank subbituminous coal, then bituminous coal, and then
anthracitic coal. Heating values, moisture content, volatile matter content, ash content, and
sulfur content can be different among these different types of coals. Bituminous coal is the most
commonly used coal for utility boilers in the
United States (Ref 1). Subbituminous coal in the
United States generally contains very low sulfur,
with many deposits containing less than 1%.
Table 10.1 shows the properties of some U.S.
coal (Ref 1). There is a big variation in moisture,
ash content, ash softening temperatures, and
sulfur content from various grades of coals. The
material factors associated with coal can directly
or indirectly affect the boiler tube material performance at different locations in the boiler.
Sulfur is the most important impurity in coal
for causing high-temperature corrosion in the
boiler. Sulfur is present in coal in forms of
organic sulfur, pyritic sulfur (i.e., pyrite), and
iron sulfate. High-sulfur coals also cause SOx
emission problems and require expensive air
pollution control equipment. As a result of the
pg 260
30/10/2007 4:01PM Plate # 0
Chapter 10:
U.S. Federal Clean Air Act emissions issues, the

low-sulfur, Powder River Basin (PRB) coal (a
subbituminous coal) has become extremely
popular for the past few years (Ref 4). The PRB
coal, which is from mines in southern Montana
through northern Wyoming, contains less than
1.2 lb of sulfur per million btu, making it compliant with Clean Air Act emissions limits
without air pollution control equipment (Ref 4).
Ash from coal after combustion can be
entrained in ue gas to cause y-ash erosion on
heat-absorbing surfaces in the lower furnace,
such as waterwalls, and in the convection pass,
such as superheater, reheater, generating bank,
and economizer. Ash can also deposit on the
furnace wall, causing a slagging problem, and on
superheater and reheater bundles, causing a
fouling problem. Ash deposits can cause heat
transfer problems, and they are required to be
removed regularly using devices such as soot
Fig. 10.2
Coal-Fired Boilers / 261
blowers (with steam), waterlances, or water

cannons (with room-temperature water),
depending on the nature of the ash deposits.
One of the important characteristics of ash is
its fusion temperature. Table 10.1 lists the ash
fusion temperatures of several coals under
reducing conditions (Ref 1). These fusion temperatures can affect the nature of the ash deposits,
whether in the form of dust or a tenacious slag.
If ash reaches the heat-absorbing surface at a
temperature near its softening temperature, the
resulting deposits are likely to be porous and can
be easily removed by sootblowing. Also, if such
a deposit is subjected to high gas temperature, the
ash deposit can reach its melting point (due to the
thermal insulating properties of the ash) and run
down the furnace wall surface (Ref 1). This
solidied slag is tightly bonded and is difcult to
remove. This slag may require water lances or
water cannons to create thermal shock for the
Same water and water/steam circulation in the furnace waterwall tubes as in Fig. 10.1 with illustration of the furnace
waterwall tubes. Source: Ref 1. Courtesy of Babcock & Wilcox
pg 261
30/10/2007 4:01PM Plate # 0
removal of this slag deposit. The furnace walls

that are subject to radiant heat are likely locations
for developing this slagging problem.
The analyses of ash for several types of U.S.
coal are shown in Table 10.1. The compositions
shown in Table 10.1 are presented as oxides.
However, most ash constituents in coal are minerals. Typical minerals found in coal are shown
in Table 10.2 (Ref 2). When these minerals are
exposed to oxidizing environments at appropriate high temperatures, oxides are likely to be
the stable phases. Most of these oxides typically
exhibit very high melting points. However, some
minerals may themselves react at combustion
temperatures to form reaction products, most
likely complex salts, that may exhibit much
lower melting temperatures. This is illustrated
in Table 10.3 (Ref 2). If the ash constituents
are present as oxides, the ratio of basic oxides
Distributor
Plate
Fig. 10.3
(e.g., Fe2O3, CaO, MgO, Na2O, and K2O) to

acidic oxides (e.g., SiO2, Al2O3, and TiO2) may
determine the fusion (fusibility) temperature of
the reaction product. It was reported that the ash
may exhibit low fusibility temperature with
higher slagging potential when its base/acid ratio
is in a range of 0.4 to 0.7 (Ref 2). Many other
parameters, such as SiO2/Al2O3 ratio, Fe2O3/
CaO ratio, Fe2O3/(CaO+MgO) ratio, (Na2O +
K2O), and so forth are also used for predicting the
fusibility temperature of the coal ash. It has been
suggested that SiO2 is more likely than Al2O3 to
form lower melting species (Ref 2). The fusibility
temperature of coal ash will be lowered when the
Fe2O3/CaO ratios are in a range of 0.2 to 10.
Alkalies are important in affecting the fusibility
of coal ash and the furnace slagging potential
(Ref 2). Many sodium compounds melt at
temperatures below 900 C (1650 F) (Ref 2).
Windbox
Bubbling fluidized-bed boiler. Source: Ref 1. Courtesy of Babcock & Wilcox
pg 262
30/10/2007 4:02PM Plate # 0
Chapter 10:
Table 10.1
pg 263
Properties of several types of U.S. coal

Coal type
Property
Anthracite
Illinois #6
Bituminous
Illinois
Spring Creek
Subbituminous
Wyoming
Bryan
Lignite
Texas
Proximate composition, %
Moisture
Volatile matter, dry
Fixed carbon, dry
Ash, dry
7.7
6.4
83.1
10.5
17.6
44.2
45.0
10.8
24.1
43.1
51.2
5.7
34.1
31.5
18.1
50.4
Composition, %
Carbon
Hydrogen
Nitrogen
Sulfur
Oxygen
Ash
83.7
1.9
0.9
0.7
2.3
10.5
69.0
4.9
1.0
4.3
10.0
10.8
70.3
5.0
1.0
0.4
17.7
5.7
33.8
3.3
0.4
1.0
11.1
50.4
Thermal properties
Heating value (as received), Btu/lb
11,890
10,300
9,190
3,930
1930
2040
2700
2100
2160
2700
2370
2580
2900+
51.0
34.0
3.5
2.4
0.6
0.3
0.7
2.6
1.4
41.7
20.0
19.0
0.8
8.0
0.8
1.6
1.6
4.4
32.6
13.4
7.5
1.6
15.1
4.3
7.4
0.9
0.4
14.6
62.4
21.5
3.0
0.5
3.0
1.2
0.6
0.9
3.5
Ash fusion temperatures (reducing atmosphere), F

Initial deformation
Softening
Fluid
Ash analysis(a), %
SiO2
Al2O3
Fe2O3
TiO2
CaO
MgO
Na2O
K 2O
P 2O 5
SO3
(a) Elements present in the ash are determined and reported as oxides. Source: Ref 1
More detailed discussion about these parameters

can be found in Ref 1 and 2. Chlorine content
is also an important indication for fouling
potential. When chlorine in coal is greater than
0.3%, fouling potential becomes high (Ref 2).
Alkali metals and chlorine can also play a signicant role in high-temperature corrosion in
boilers.
10.4 Combustion Environments

Stoichiometric combustion is a complete
combustion of a fuel with a theoretically calculated amount of oxygen in the combustion air.
Complete combustion of all combustible constituents in coal requires excess air beyond the
theoretically calculated value needed in the
combustion air. Excess air is typically expressed
in terms of percentage above the theoretically
calculated value. For pulverized coal, typically
15 to 30% excess air is required to enssure adequate combustion (Ref 2). This will make the
combustion environment an oxidizing atmosphere.
Table 10.2
coal
Mineral species
Kaolinite
Illite
Biotite
Orthoclase
Albite
Calcite
Dolomite
Siderite
Pyrite
Gypsum
Quartz
Hematite
Magnetite
Rutile
Halite
Sylvite
Typical mineral species found in

Formula
Al2O32SiO2H2O
K2O3Al2O36SiO22H2O
K2OMgOAl2O33SiO2H2O
K2OAl2O36SiO2
Na2OAl2O36SiO2
CaCO3
CaCO3MgCO3
FeCO3
FeS2
CaSO42H2O
SiO2
Fe2O3
Fe3O4
TiO2
NaCl
KCl
Source: Ref 2
Nitrogen oxides (NOx), which form during

combustion under oxidizing conditions, are an
undesirable pollutant from the boiler and contribute to acid rain and ozone formation. In order
to reduce this NOx (i.e., NO and NO2) emission
from the coal-red boiler, low NOx burner
30/10/2007 4:02PM Plate # 0
pg 264
systems and staged combustion techniques have

been developed and employed with great success
in signicantly reducting NOx emission. The
majority of NOx is formed by high-temperature
oxidation of N2 in the combustion air under
oxidizing conditions. Nitrogen in coal can also
lead to formation of NOx during combustion.
Reducing the availability of air during initial
combustion can signicantly reduce the formation of NOx. This results in reducing environments. For staged combustion, the lower furnace
is under reducing conditions, and additional air is
then introduced at a higher elevation in the furnace to complete the combustion process. More
detailed discussion of NOx control in combustion
is available in Ref 1 and 2. As a result of low NOx
combustion techniques, the furnace environment
changes from an oxidizing to a reducing condition. The location for this change is variable and
will uctuate as the load changes for the boiler
operation especially under non-base-load operation. Sulfur in the coal forms H2S in reducing
conditions instead of SO2 in oxidizing conditions. High wastage rates for waterwall tubes
under reducing conditions and oxidizing/reducing alternating conditions can become a very
serious materials issue.
Ash released from coal during combustion can
cause signicant problems to the behavior of the
boiler components. During combustion, signicant amounts of ash are carried by the ue gas
stream and cause y-ash erosion problems particularly for superheaters and reheaters in the
convection path. In addition, it can produce
slagging on the furnace walls and fouling for
superheaters and reheaters. The amount of ash
being carried by the ue gas stream depends
on the type of ring, for example, approximately
70 to 90% for pulverized coal units, approximately 40% for stoker-ring units, and all the
ash along with some uidized-bed material for
circulating uidized-bed boilers (Ref 1).
During combustion, the gas temperature
can reach a range of 1370 to 1650 C (2500 to
Table 10.3
3000 F), some of the mineral constituents and

compounds can be in a molten or plastic state
(see Tables 10.1 and 10.3). These ash constituents can then deposit on the heat-absorbing
surfaces in the boiler. Two general types of ash
deposition, slagging and fouling, can take place.
Slagging is the deposition of molten, partially
fused deposits on the furnace walls and the upper
furnace radiant superheaters exposed to radiant
heat (Ref 1). Fouling is the deposition of more
loosely bonded deposits on the heat-absorbing
surfaces in the convection pass, such as superheater and reheater, that are not exposed to
radiant heat (Ref 1). These ash deposits, which
are impeding the heat transfer, require their
regular removal from the tube surfaces. Soot
blowers using steam are generally adequate for
removing the ash deposits on the fouled tube
surfaces, while waterlances or water cannons
using thermal shock by water may be needed
to clean the slagged tube surfaces. Material
behavior under the inuence of ash deposits is
discussed in later sections.
10.5 Fireside Corrosion and Other

Materials Problems in Boilers
This section discusses the materials problems
related to the heat-absorbing surfaces in the furnace combustion area (i.e., waterwalls) and also
in the convection pass, such as superheaters
and reheaters. A schematic of a coal-red boiler
showing these two areas is shown in Fig. 10.4.
Accelerated tube wall wastage attack resulting
from the change of the combustion environment
from oxidizing to reducing due to installation of
low-NOx combustion technology intended to reduce NOx emissions is discussed. Also discussed
in detail are circumferential cracking of the furnace waterwalls, soot blower erosion/corrosion,
thermal fatigue cracking induced by waterlances
and water cannons for deslagging the furnace
Melting points of coal ash constituents
Element
Oxide
Compound
Si
Al
Ti
Fe
Ca
Mg
Na
K
SiO2
Al2O3
TiO2
Fe2O3
CaO
MgO
Na2O
K2O
1715 (3120)
2043 (3710)
1838 (3340)
1565 (2850)
2521 (4570)
2799 (5070)
Sublimes at 1277 (2330)
Decomposes at 349 (660)
Na2SiO3
K2SiO3
Al2O3Na2O6SiO2
Al2O3K2O6SiO2
FeSiO3
CaOFe2O3
CaOMgO2SiO2
CaSiO3
877 (1610)
977 (1790)
1099 (2010)
1149 (2100)
1143 (2090)
1249 (2280)
1391 (2535)
1540 (2804)
Source: Ref 2
30/10/2007 4:02PM Plate # 0
Chapter 10:
waterwalls, and superheater/reheater corrosion.

Limited discussion includes erosion issues in
uidized-bed boilers. Hydrogen attack encountered in subcritical units as a result of water
corrosion in the internal diameter of the waterwall tubes is discussed in Hydrogen Attack in
Chapter 17.
10.5.1 Fireside Corrosion of Furnace
Waterwalls
The furnace is enclosed by four walls (or
waterwalls) with front, rear, and two side walls in
a square or rectangular form for the horizontal
cross section of the furnace. The waterwall consists of tubes with membranes (i.e., steel plates)
connecting adjacent tubes (a tube-membranetube construction), or of tubes without membranes connecting adjacent tubes (a tangent tube
construction). These waterwalls, which enclose
the furnace combustion zone, are heat-absorbing
surfaces that heat the rising water in the tubes
(from bottom) to become steam. The gas temperature in the combustion inside the furnace
can reach a range of 1370 to 1650 C (2500 to
3000 F). As a result, these waterwalls, particularly the crown location of the tube on the reside
Fig. 10.4
Source: Ref 5
(i.e., the side facing the combustion), are subject

to extreme radiant heat from the combustion. The
tube wall exhibits temperature gradient across
the tube wall thickness. The outer metal temperature depends on the temperature of the water
and/or steam inside the tube along with many
other factors. For subcritical units of drum-type
boilers, the temperature of water/steam depends
on the pressure. For example, for the pressure of
170 atm (2500 psig), the saturated temperature
is 353.4 C (668.11 F) (Ref 3). In a supercritical
unit with pressures above 218.2 atm (3208 psig),
the water inside the tube changes to 100% steam
with a continuous temperature increase. The tube
metal temperature for the supercritical unit is
generally higher than that of a subcritical unit.
Other factors that can affect tube metal temperature include oxide scales formed on the tube
inside diameter (ID) surface, slag deposits on the
tube outside diameter (OD) surface, ame
impingement, and so forth. Typical temperature
range for the waterwall tube on the reside has
been given as 380 to 450 C (716842 F) by
Lees and Whitehead (Ref 6), below about 450 C
(842 F) by Hay (Ref 7), 427 to 482 C (800 to
900 F) by Cunningham and Webster (Ref 8),
and 400 to 500 C (752 to 932 F) by Stringer
Coal-fired boiler, showing two main areas for discussion of materials problems in a boiler: the furnace combustion
area (i.e., furnace walls) and the heat-absorbing surfaces in the convection path, such as superheaters and reheaters.
pg 265
30/10/2007 4:02PM Plate # 0
(Ref 9). Eurich and Kramer (Ref 10) indicated

that the design waterwall tube surface metal
temperature was between 482 and 510 C (900
and 950 F) for each of the ve 676 MW(e)
opposed-red supercritical boilers. Yeager (Ref
11) indicated that the design waterwall surface
temperature for the two supercritical, tangentially red boilers (each with 750 MWe) at
Montour Station was between 482 and 538 C
(900 and 1000 F). Figure 10.5 (Ref 12) shows a
schematic of the temperature gradient across the
tube wall and different oxide layers and ash/slag
deposits.
Wastage. Excess air is normally used for
combustion to ensure adequate combustion of
coal. For example, typically 15 to 30% excess
air is required for ring pulverized coal (Ref 2).
This produces an oxidizing condition with
major combustion gaseous products being N2,
O2, CO2, and H2O. Under this normal oxidizing
condition, waterwalls made of carbon steels
and low-alloy steels have been found to show
metal wastage rates of approximately 40 nm/h
(13 mpy) (Ref 12). A typical tube wastage prole
for the corroded waterwall tube is illustrated in
Fig. 10.6 (Ref 13). The maximum wastage typically occurred at the crown location (12 oclock
position). Figure 10.7 shows a cross section of a
carbon steel waterwall tube removed from a
subcritical unit after 21 years of service with an
observed wastage rate of about 0.19 mm/y
(7.5 mpy) (Ref 14). The front face of the tube at
the crown location facing the combustion radiant
heat suffers the worst metal wastage. One boiler

designer suggests the maximum tube metal
temperature limits based on oxidation for typical
ferritic steels used for waterwalls, as shown in
Table 10.4 (Ref 2). For example, carbon steel is
limited to 454 C (850 F), but without mentioning the design service life. However, due to
Fig. 10.6
Fig. 10.7
Fig. 10.5
Temperature gradients through the inner oxide

scale, tube wall, outer oxide scale, and ash/slag
deposits. Source: Ref 12
Typical wastage profile of a corroded waterwall

tube. Source: Ref 13
Cross section of a carbon steel waterwall tube from

a subcritical unit in the United States after 21 years
of service, showing the maximum wastage at the crown location
with about 1.9 mm/yr (7.5 mpy) of wastage rate. Source: Ref 14.
Courtesy of Welding Services Inc.
pg 266
30/10/2007 4:02PM Plate # 0
Chapter 10:
the complexity of the combustion in a large coalred boiler it is almost impossible to establish a
normal oxidizing condition uniformly in the ue
gas throughout the entire furnace cross section in
a dynamic condition. As a result, reducing conditions are developed in some localized areas
near the waterwall. Chemical analysis of ash
deposits formed on tube surfaces revealed an
appreciable amount of free carbon (Ref 15). The
presence of free carbon indicates the establishment of reducing conditions at the furnace wall
surface (Ref 16). Localized reducing conditions
with up to 10% CO in the furnace atmosphere in
the vicinity of the furnace walls have been
observed (Ref 17, 18).
Suldation. One of the most corrosive impurities in coals is sulfur, which is present in coals
as pyrite (FeS) and organic sulfur. Under an
Table 10.4 Maximum outer tube metal
temperature limits based on oxidation for
common ferritic steels used in waterwalls, as
suggested by a boiler designer
Steel
designation
Type
SA210 A1
T1
T11
T22
Carbon Steel
Carbon-0.5Mo
1.25Cr-0.5Mo
2.25Cr-1Mo
Source: Ref 2
Fig. 10.8
Oxidation limit,
C (F)
454 (850)
482 (900)
552 (1025)
593 (1100)
oxidizing condition, sulfur oxidizes to form SO2

during combustion. The presence of SO2 in an
oxidizing combustion atmosphere does not generally cause accelerated corrosion. However,
under reducing conditions, sulfur reacts to form
H2S. This can signicantly increase corrosion
rates by suldation. Lees and Whitehead (Ref 6)
examined several corroded carbon steel waterwall tube samples as well as Type 310 coextruded tube samples removed from several
power plants in the United Kingdom that suffered severe wastage problems. They had
observed suldation in some severely corroded
carbon steel tubes in addition to a chlorine-rich
phase at the boundary between the oxide/sulde
phases and the metal. (Chlorine issues are discussed next in this section). In one carbon steel
sample that had suffered a wastage rate of about
530 nm/h (170 mpy), they found the corrosion
product to be predominantly suldes. They even
observed a substantial layer of oxides/suldes on
the corroded Type 310 cladding in a Type 310/CS
composite tube that suffered more than 50 nm/h
(16 mpy) (Ref 6). Figure 10.8 shows a waterwall
carbon steel tube after only 1 year of service in a
subcritical unit in the United States, showing
pitting attack. The boiler was in the United
States, burning a coal containing about 3.0 to
3.5% S and about 300 to 400 ppm chlorine
(Ref 14). Metallurgical examination of the
Close-up view of a waterwall carbon steel tube showing pitting attack after 1 year of service in a subcritical unit in the United
States, burning coal containing about 3.0 to 3.5% S and about 300 to 400 ppm chlorine. Source: Ref 14. Courtesy of
Welding Services Inc.
pg 267
30/10/2007 4:02PM Plate # 0
pg 268
sample indicated that pitting was associated with

suldation attack, as shown in Fig. 10.9, which
shows initiation of internal suldation attack
(Ref 14). The tube was in the initiation stage of
the accelerated corrosion by suldation. There
were also areas in the same tube that showed only
oxidation attack, as illustrated in Fig. 10.10
(Ref 14).
Suldation of the waterwall tubes in the lower
furnace causes a serious wastage issue when the
furnace is red intentionally under reducing
conditions (i.e., substoichiometric conditions)
when retrotted with low-NOx burners to reduce
the formation of NOx. This subject is discussed in
section 10.5.2.
Chlorine. Another important impurity in coal
that can signicantly affect the corrosion of
metals in boilers is chlorine. Most coals from the
U.S. contain low chlorine (<0.1% Cl) (Ref 19).
Thus, there is very little experience and few
studies have been conducted on the effects of
high-chlorine coals on corrosion of metals in
U.S. boilers. British coals contain much higher
chlorine. The average chlorine content in coals
used in British power plants is about 0.25% with
some as much as 0.8% (Ref 20). Chlorine in coal
converts mostly to HCl during combustion. It is

believed that every 0.1% Cl in coal produces
approximately 80 ppm HCl in the ue gas (Ref
21). Clarke and Morris (Ref 18) performed gas
analysis on the gas samples obtained from the
vicinity of the furnace wall at various locations in
a 120 MWe boiler that historically suffered high
wastage rates (excess of 150 nm/h, or 48 mpy, or
1.2 mm/yr). This boiler burned high-chlorine
coal containing about 0.4 to 0.6% Cl. The authors
observed that reducing conditions occurred in the
areas of high wastage, while oxidizing conditions
were observed in the areas that were outside of
these high-wastage areas. In addition, the concentration of HCl was found to be about
400 ppm in the high-wastage areas. The level of
H2S was found to be about 300 to 400 ppm
only when CO concentration exceeded 3%, but
in some areas no H2S was observed under
those reducing conditions. In laboratory tests
in a simulated combustion gas environment
19 m
Fig. 10.9
Fig. 10.10
1: 1.1% S, 0.7% Al, 0.8% Si, 0.6% Mn, 95% Fe, and trace
elements
2: 0.8% S, 0.3% Al, 0.6% Si, 0.5% Cl, 0.5% Ca, 1.9% Zn, 94% Fe,
and trace elements
3: 13.4% S, 0.8% Al, 0.4% Si, 0.8% Mn, 84.1% Fe, and trace
elements
4: 9.2% S, 0.3% Al, 0.4% Si, 0.6% Mn, 89.3% Fe, and trace
elements
5: 1.7% S, 0.8% Al, 1.0% Si, 1.2% Cl, 0.2% Ca, 0.7% Mn, 91.8%
Fe, and trace elements
1: 8.9% Si, 1.5% Al, 86.9% Fe, and trace elements

2: 44.3% Si, 22.9% Al, 2.2% Mg, 6.7% Ca, 5.6% K, 14.7% Fe, and
trace elements
3: 1.7% Si, 1.0% Al, 1.0% S, 1.5% Zn, 93.7% Fe, and trace
elements
4: 1.3% Si, 1.0% Cu, 1.3% Zn, 1.2% S, 91.7% Fe, and trace
elements
5: 2.9% Cu, 93.8% Fe, and trace elementsa
6: 1.0% Si, 1.3% Zn, 93.1% Fe, and trace elements
Scanning electron microscopy backscattered

electrons image of the corrosion products showing initiation of sulfidation attack on the tube, where pitting attack
was observed on the tube surface as shown in Fig. 10.8 Chemical
compositions of the phases of the corrosion products at different
locations were analyzed by energy dispersive x-ray spectroscopy
(EDX) with the results summarized as:
Scanning electron microscopy backscattered

electrons image of the corrosion products
showing ash deposits and iron oxides with no evidence of sulfidation attack on other area of the tube that did not suffer pitting
attack (Fig. 10.8). Chemical compositions of the phases of
the corrosion products at different locations were analyzed
by energy dispersive x-ray spectroscopy (EDX) with the results
summarized as:
30/10/2007 4:02PM Plate # 0
Chapter 10:
However, after 27,000 h, both co-extruded tube

alloys showed a maximum corrosion rate of about
60 nm/h (19 mpy). Both Type 304 and 310 alloys
exhibited signicant performance improvement
compared with carbon steel. Examination of
corroded 304 and 310 samples revealed substantial amounts of suldes along with some
chlorine.
The effect of chlorine in coals on the furnace
wall corrosion was summarized by Daniel et al.
(Ref 19) based on CEGB data generated in boilers in the United Kingdom. The effect is illustrated in Fig. 10.13 and 10.14. Coals with more
than 0.2% Cl may result in accelerated reside
Fig. 10.12
Metal loss data for carbon steel in a simulated

combustion atmosphere consisting of N2-10CO10H2O-0.5SO2 containing different levels of HCl (0, 400, and
2000 ppm) at 500 C. Source: Ref 22
Corrosion rate, mpy
Metal loss, m
(N2-10%CO-10%H2O-0.5%SO2), Brooks and

Meadowcroft (Ref 22) showed that carbon
steel corroded at linear rates of about 130 and
550 nm/h (42 and 176 mpy) at 400 and 500 C
(752 and 932 F), respectively, with 400 ppm
HCl added to the test environment. Their test data
are shown in Fig. 10.11 and 10.12. The laboratory 400 C (752 F) test data of Brooks and
Meadowcroft (Ref 22) and the eld data of Clarke
and Morris obtained from a boiler (Ref 18) were
in the same order of magnitude. Figures 10.10
and 10.11 also show that the corrosion of carbon
steel followed a parabolic reaction rate when the
environment contained no chlorine. In the
examination of waterwall tube samples from
several plants by Lees and Whitehead (Ref 6),
corrosion rates were found to vary from about
200 to 530 nm/h (64 to 170 mpy) in different
plants. The corrosion products the authors
observed were mainly suldes with a chlorinerich phase along the metal/scale interface for
some samples. Other samples contained mainly
suldes with no chlorine-rich phases. If iron
chlorides formed at the metal/scale interface, the
phases that were highly volatile with high vapor
pressures could easily escape into the environment (see Chapter 6). In a eld testing of coextruded tubes with Type 304 and Type 310
stainless steel claddings in a Central Electricity
Generating Board (CEGB) boiler (500 MW) that
had suffered aggressive furnace wall corrosion
problems, Flatley et al. (Ref 23) found that the
corrosion rates were about 77 nm/h (25 mpy)
for Type 304, 55 nm/h (17 mpy) for Type 310,
and 180 to 250 nm/h (58 to 80 mpy) for
carbon steel after 15,000 h of exposure.
10
Fig. 10.11
Metal loss data for carbon steel in a simulated

combustion atmosphere consisting of N2-10CO10H2O-0.5SO2 containing different levels of HCl (0, 400, and
2000 ppm) at 400 C. Source: Ref 22
Fig. 10.13
Effect of chlorine in coals on the corrosion rate of

carbon steel and low-alloy steels under reducing
conditions, based on CEGB laboratory data (Ref 16). Source:
Ref 19
pg 269
30/10/2007 4:02PM Plate # 0
corrosion for furnace wall tubes under reducing

conditions (Ref 19). In 1997, James and Pinder
(Ref 24) provided a different perspective on the
effect of coal chlorine on the furnace wall corrosion. They indicated that one source of data
that contributed to the correlation of furnace wall
tube corrosion and the chlorine concentrations
in coals was obtained from a power station with
ve front-red boilers (each with 120 MW).
Between 1967 and 1977, the chlorine content in
coals for this station in United Kingdom was
increased from 0.35 to 0.65%. The maximum
corrosion rates over this period showed an
increase from 170 to 550 nm/h (58.5 to
189.2 mpy) when chlorine was increased from
0.35 to 0.65%, independent of sulfur content and
load factor (Ref 24). James and Pinder summarize the data in Fig. 10.15 (Ref 24). Also included
are data from the other plant. The data show a
strong dependence of furnace wall corrosion on
the chlorine content in coal. Figure 10.16 shows
the data from many plants plotted together
(Ref 25), showing a signicant scattering.
This may not be unexpected, since there were so
many plants that were probably of different
Fig. 10.14
boiler designs and were under various operating

parameters, along with many other variables that
were involved. The correlation would be best
established when the data were obtained from the
same boiler under constant operating parameters
with the chlorine content in the coal as the only
variable.
10.5.2 Fireside Corrosion of Furnace Walls
under Low NOx Combustion Conditions
In order to comply with the Clean Air Act
Amendments of 1990, power plants are required
to reduce NOx emissions from boilers. Nitrogen
oxides (NOx) are produced during combustion
under oxidizing conditions, and they are undesirable pollutants from the boiler because they
contribute to acid rain and ozone formation. In
order to reduce NOx emissions, staged ring is
used to produce a substoichiometric combustion
(i.e., combustion with insufcient oxygen or a
reducing condition) in the lower furnace, which
is followed by introduction of adequate air at the
overre air ports at a higher elevation to complete
the combustion process. As a result, the lower
Corrosion rate as a function of temperature for carbon steel and austenitic stainless steels obtained from laboratory testing
as well as field testing at Eggborough Power Station. Source: Ref 19
pg 270
30/10/2007 4:02PM Plate # 0
Chapter 10:
content in the alloy produces a more protective

chromium oxide scale with less probability of
forming suldes.
Similar results (Table 10.6) were reported by
Gilroy (Ref 27) using a simulated test environment with much lower H2S level, showing that
Type 310, Type 446, alloy 671 and chromized
coatings were signicantly more resistant to
suldation than carbon steel. The simulated
environment used by Gilroy contained about
600 ppm HCl. Where chloridation attack was
involved in this testing, chromium was also
effective in increasing the alloys resistance to
chloridation attack. Kung and Eckhart (Ref 28)
conducted tests in H2S-containing environments
for carbon steel, low-alloy steel, and several
stainless steels. Figure 10.17 shows the corrosion
rates for carbon steel and 2.25Cr-1Mo steel tested at 370 C (700 F) in N2-5.1CO-16.7CO24.6H2O-0.55H2 containing several levels of H2S
(0.05, 0.25, 0.5, 4.8% H2S) for 1000 h. Similar
tests were conducted for Types 304L and 310
(Fig. 10.18). Both 304L and 310 showed signicant reduction in corrosion rates compared
with carbon and 2.25Cr-1Mo steels. At 480 C
(900 F) in the gas mixture with 0.05% H2S, the
corrosion rate was about 0.05 mm/yr (2 mpy) for
Corrosion rate (max), nm/h
furnace is under reducing conditions and/or

shifting alternatively from oxidizing to reducing
conditions. Under reducing conditions, sulfur
from coal forms H2S instead of SO2. H2S can
cause severe suldation attack on carbon steel
furnace walls.
Chou and Daniel (Ref 26) conducted a simulated test with a relatively high level of H2S for a
variety of materials. Their results are summarized in Table 10.5 (Ref 26). Carbon and lowalloy steels were found to suffer corrosion rates
in a 1 to 1.3 mm/yr (40 to 50 mpy) range. Type
304, which is signicantly more resistant than
carbon and low-alloy steels, showed corrosion
rates of 0.2 to 0.3 mm/yr (8 to 12 mpy). This is
because the corrosion products were changed
from iron oxides/suldes on steels to chromium
oxides (or possibly with chromium suldes) on
Type 304. Both chromium oxides and/or chromium suldes exhibit much slower growth rates
than iron oxides/suldes; thus much less metal is
consumed and accordingly metal wastage rates
are much lower. Type 309 showed a very low
corrosion rate. Alloy 671 and chromized coatings
showed extremely low corrosion rates. Increases
in corrosion resistance are mainly related to the
chromium content in the alloy. Higher chromium
Weighted mean chlorine content, %
Fig. 10.15
Corrosion rates of furnace walls (carbon steel) as a function of chlorine content in coal from boilers in two plants. The
equation of the line is Rc =1380 (%Cl)-208, where Rc is corrosion rate (nm/h). Source: Ref 24 from original data (Ref 25)
pg 271
30/10/2007 4:02PM Plate # 0
pg 272
Corrosion rate, nm/h
Weighted mean coal chlorine content
Fig. 10.16
Corrosion rates of furnace walls (carbon steel) as a function of chlorine content in coal from boilers in many plants. Source:
Ref 24 from original data (Ref 25)
Table 10.5 Corrosion of various materials in

N2-5CO-16CO2-10H2O-0.5H2-2H2S at 482 C
(900 F) for 4000 h
Alloy
Carbon steel
2.25Cr-1Mo
Type 304
Type 304L
Type 309
Alloy 800
Alloy 671
Chromized carbon steel
Chromized 2.25Cr-1Mo
Corrosion rate,
mm/yr (mpy)
1.04 (41)
1.32 (52)
0.2 (8.2)
0.3 (12.0)
0.04 (1.6)
0.33 (13.0)
0.005 (0.18)
0.006 (0.25)
0.008 (0.32)
0.007 (0.28)
Source: Ref 26
Type 310 and about 0.2 mm/yr (8 mpy) for

Type 304L (Ref 28).
Dooley et al. (Ref 29) indicated that most
boilers experienced waterwall wastage rates of
29 nm/h (10 mpy) or less prior to installation of
low NOx burners. The wastage rates have been
dramatically increased to a range of 145 to 350
nm/h (50 to 120 mpy) for some boilers after
installing low NOx burners (Ref 29). In some
boilers, waterwall wastage rates, which were in a
range of 49 to 87 nm/h (17 to 30 mpy), were
found to increase to a range of 96 to 226 nm/h
(33 to 78 mpy) after retrotting with low-NOx
burners (Ref 30). It was also reported that the
corrosion scales were found to consist of iron
suldes and oxides. Furthermore, the ue gas in
Table 10.6 Weight gain (mg/cm2) of several

alloys after testing in N2-10CO-5CO2-10H2O0.1H2S-600 ppm HCl at 400 and 500 C
(752 and 932 F) for 3000 h
Alloy
Carbon steel
50Ni/50Cr, sprayed coating
Aluminized (A)
Aluminized (B)
Fe-27Cr-6Al-2Mo
FAL (Fe-13Cr-4Al)
Ferralium (Fe-25Cr-5Ni-4Mo)
44-LN (Fe-26Cr-5Ni-1.5Mo)
Monit (Fe-25Cr-5Ni-4Mo)
29-4-C (Fe-28Cr-4Mo)
29-4-2 (Fe-29Cr-4Mo-2Ni)
E Brite (Fe-26Cr-1Mo)
Fecralloy (Fe-16Cr-5Al-0.35Y)
Fecralloy A (Fe-16Cr-4.5Al-0.26Al)
GE2541 (Fe-26Cr-5Al-0.45Y)
Type 310
Type 310Nb (0.8Nb)
Alloy 800H
Type 446
Alloy 671
Chromized 2.25Cr-1Mo
400 C
(752 F)
25.0
8.0
5.5
3.1
0.7
0.8
0.3
0.2
0.2
0.7
0.6
0.7
0.7
0.2
0.6
0.7
0.4
0.3 (1500 h
exposure)
0.5
0.6
0.3
0.2
500 C
(932 F)
90.0
9.5
8.8
3.8 (2000 h
exposure)
0.1
2.1
0.5
0.7
1.1
0.9
0.8
0.6
0.7
0.4
0.2
0.4
0.6
0.3
0.3 (1500 h
exposure)
0.3
0.5
0.2
0.3
Source: Ref 27
the vicinity of the furnace walls that suffered

accelerated wastage rates was observed to have
levels of CO greater than 5% and of H2S up to
30/10/2007 4:02PM Plate # 0
Chapter 10:
500 ppm (Ref 30). Also observed in the high

wastage areas were carbon-rich deposits and
pyrites. Urich and Kramer (Ref 10) reported that
four supercritical boilers (each with 676 MWe,
opposed-red) at Gibson Generating Station
were retrotted with low NOx burners and
overre air. Prior to the installation of low NOx
burners, tube wastage problem was only limited
to a small area on the side walls at locations
where H2S concentrations were measured to be
as high as 1000 ppm. Since the retrot, a marked
increase in waterwall tube wastage was observed
for all four boilers. Severe corrosion was found to
occur on the side walls, with the most severe
corrosion occurring in the top burner elevation
up to the level of the overre air ports (Ref 10). In
some cases, wastage rates were on the order of
285 nm/h (100 mpy). Furthermore, the corrosion scales in these areas were mainly iron suldes. Gas sampling from the side wall in one of
the boilers showed the levels of CO as high as 11
or 12% and those of H2S as high as 1000 ppm
(Ref 10). It was reported that the service life of
the waterwall panels has been reduced from 12 to
15 years to 4 years after installation of the low
NOx burners in these boilers at Gibson Station
(Ref 31). Two supercritical units with tangentially ring (582 MWe each) were also reported
to have a more serious waterwall wastage issue
after installation of low NOx burners with overre air (Ref 31).
Yeager (Ref 11) reported that a supercritical
boiler, Unit No. 2, at Montour Station was
retrotted with low NOx burners in 1994. In
1996, an inspection revealed approximately

370 m2 (4000 ft2) of the waterwall had suffered
severe wastage with maximum wastage rates
being in excess of 2.2 mm/yr (85 mpy). The
boiler, a tangentially red unit (750 MWe),
burned Eastern bituminous coal with 1.9 to 2.2%
S, had a waterwall wastage problem prior to
installation of low NOx burners with the average
wastage rates in the range of 0.25 to 0.38 mm/yr
(10 to 15 mpy). Yeager indicated that the reball
had tendency to be closer to certain burner corners, thus causing higher wastage rates. The
furnace wall gas sampling tests revealed high
total reduced sulfurs (TRS) and CO levels at the
burner corners that suffered high wastage rates
and an oxidizing atmosphere at the burner corners that showed little wastage (Ref 11). Clark
and Morris (Ref 18) reported that a level of H2S
in a range of 300 to 400 ppm and of CO in a
range of greater than 3.5% were observed in the
vicinity of the furnace walls with accelerated
wastage rates in U.K. boilers. The level of H2S
observed in U.K. boilers was similar to what was
observed in U.S. boilers.
The areas that suffer severe waterwall wastage
can be highly dependent on boiler design and
ring conguration, among other factors. James
and Pinder (Ref 32) indicated that the side and/or
rear walls were most vulnerable in a front-red
boiler, while in a tangentially red boiler, the
wastage was invariably highest on the front wall.
Bakker and Kung (Ref 33) observed that FeS
deposits can signicantly increase the corrosion
rates of carbon and low-alloy steels. They found
that FeS deposits showed no effects in reducing
environments, but accelerated the corrosion in
alternating oxidizing and reducing environments
or oxidizing environments. This is shown in
Fig. 10.17
Fig. 10.18
Corrosion rates of carbon steel and 2.25Cr-1Mo

steel (T-22) as a function of H2S in the N2-5.1CO16.7CO2-4.6H2O-0.55H2 gas mixture at 370 C (700 F) for
Corrosion rates of Types 304L and 310 as a

function of H2S in the N2-5.1CO-16.7CO24.6H2O-0.55H2 gas mixture at 370 C (700 F) for 1000 h.
Source: Ref 28
pg 273
30/10/2007 4:02PM Plate # 0
Tables 10.7 and 10.8. The authors believe that

FeS under oxidizing conditions can produce
volatile reduced sulfur species including sulfur
vapor, H2S, COS, and so forth, thus making the
environment more aggressive and causing accelerated wastage rates. Since FeS is stable only
under reducing conditions and causes accelerated
corrosion to carbon and low-alloy steels only
under oxidizing conditions, the waterwall area
that is likely to suffer accelerated wastage rates
appears to be subject to alternating reducing and
oxidizing conditions. These alternating reducing
and oxidizing conditions may exist in staged
ring boilers where the overre air mixes with
substoichiometric ue gas from the burner zone
(Ref 33). In one boiler where two locations at the
waterwall made of T2 steel were monitored for
their wastage rates as well as the ash deposits, the
maximum wastage rate of 1.5 mm/yr (60 mpy)
at one location and of 0.37 mm/yr (15 mpy) at
the other location were observed after 12,633 h
of service (Ref 34). The ue gas measurements
near the waterwall at those two locations showed
highly reducing conditions with about 7% CO
and 500 ppm H2S during full-load operation and
slightly oxidizing conditions during low-load
conditions. However, there are signicant differences in the ash deposits between the two
locations; the area suffering high wastage rates
had an ash deposit consisting of 90% FeS and
10% y ash, while the low wastage rate area had
an ash deposit of mainly Fe3O4 and y ash
(Ref 34). This eld experience was consistent
with the laboratory test results observed by
Bakker and Kung (Ref 33). In the same boiler as
discussed previously in this paragraph, another
panel testing was conducted after 17,155 h of
exposure, the FeS content in the ash deposits was
in the 40 to 90% range, but with chlorine content
in the range of 500 to 1800 ppm. The maximum
wastage rate in the area was about 3.5 mm/yr
(140 mpy), a signicantly higher wastage rate
than the earlier test panel with about 90% FeS
Table 10.7 Corrosion rates of T2 in

N2-7CO-12CO2-6H2O-0.13H2S-0.008HCl with
various ash deposits at 427 C (800 F) for 600 h
Ash deposit
No deposit
30% FeS, 70% y ash
60% FeS, 40% y ash
60% FeS, 20% carbon, 20% y ash
90% FeS, 10% y ash
Source: Ref 33
0.420.57 (1723)
0.10.3 (512)
0.10.17 (57)
0.20.6 (922)
0.10.27 (511)
and only 90 ppm chlorine in the ash deposits

(maximum wastage rate of 1.5 mm/yr or
60 mpy) (Ref 34). Higher chlorine content in the
ash deposits might play a role in causing higher
wastage rates (Ref 34).
Bakker (Ref 34) reported preliminary results
of the laboratory test involving the effect of the
KCl-NaCl (equal amounts) in the ash deposits of
char, FeS, and y ash. The test involved ue gas
temperature of 1000 C (1830 F), maximum
ash deposit temperature of 650 C (1200 F),
metal tube sample temperature of 450 to 470 C
(850 to 900 F), and the temperature of the
cooling air (inside the tube sample to produce the
temperature gradient) of 320 C (610 F) at the
exit end. The metal loss was found to be about 20
to 22 m in 100 h when the ash deposits contained no chloride and about 44 to 50 m in
100 h when 0.2% alkali chloride (KCl-NaCl in
equal amounts) was added to the ash deposits.
Lees and Whitehead (Ref 6) observed suldes
as well as a chlorine-containing phase at the
sulde/metal interface when they examined the
corroded carbon steel waterwall samples using
SEM/EDX analyses. Similar observation was
also found by Lai (Ref 14). Figure 10.19 shows
a corroded carbon steel waterwall sample from
a boiler red with low NOx burners (Ref 14).
The boiler, a subcritical unit, had been burning
Western Kentucky bituminous coal containing
about 3.5% S. The waterwalls started to experience accelerated wastage rates after the boiler
began stage ring with low NOx burners and
overre air. SEM/EDX analysis of the corrosion
products formed on the sample revealed iron
sulde phases and a chlorine-containing phase
at the sulde/metal interface, as shown in
Fig. 10.20 (Ref 14).
Coating and Weld Overlay. As more boilers
experienced accelerated wastage rates for carbon
and low-alloy steels at waterwalls in boilers that
had been retrotted with low-NOx burners,
Table 10.8 Corrosion rates of T2 in alternating
reducing(a) and oxidizing environments(b) with
various ash deposits at 427 C (800 F) for 400 h
Ash deposit
No deposit
30% FeS, 70% y ash
60% FeS, 40% y ash
60% FeS, 20% carbon, 20% y ash
90% FeS, 10% y ash
50% carbon, 50% y ash
0.370.5 (1520)
1.0 (41)
0.82 (33)
1.2 (49)
1.2 (49)
0.5 (20)
(a) 16 h in N2-7CO-12CO2-6H2O-0.13H2S-0.008HCl. (b) 8 h in N2-1O2-17CO20.13SO2-0.008HCl. Source: Ref 33
pg 274
30/10/2007 4:02PM Plate # 0
Chapter 10:
industry looked for a viable coating or weld

overlay to extend the waterwall life. Bakker et al.
(Ref 35) reported the results of in-plant tests of
thermal sprayed coatings (arc sprayed coatings
and HVOF coatings), diffusion coatings (chromized coating and Cr/Si diffused coating), and
weld overlays (stainless steels 410, 309, and
312, and alloy 625). Waterwall test panels, which
consisted of uncoated T2 tubes, coated and weld
overlay tubes, were tested in No. 2 boiler at
Hatelds Ferry Station. The boiler, a supercritical unit, was retrotted with a low NOx cell
burner system. Test panels were installed at
North side wall at the burner elevation. Side
walls typically suffered the worst corrosion
attack for an opposed wall ring boiler such as
this one. During testing, the boiler burned Eastern bituminous coals with about 2.2% S with no
measurements or control of chlorine content.
However, the Pennsylvania coals typically contained about 0.05 to 0.15% Cl. The test results
generated by Bakker et al. (Ref 35) are summarized:
After 12,683 h of exposure, T2 waterwall

tubes were found to suffer a wastage rate of
about 1.6 mm/yr (62 mpy) at the burner
elevation. The arc sprayed coating of Ni44Cr alloy wire suffered minor spallation.
The high-velocity oxyfuel (HVOF) coating
of Ni-42Cr-2Si alloy also suffered minor

spallation. The arc sprayed coating of Ni45Cr-2Al was completely spalled off from
the tube. Metallurgical evaluation of all the
coated tubes showed some sulfur penetration
and corrosion attack at the coating/metal
interface. Furthermore, a thickness loss of 20
to 50% of the original thickness (typically
about 0.5 mm, or 20 mils) was observed at
the areas where no spallation occurred after
12,683 h of exposure. After additional
10,072 h of exposure, most coatings had lost
35 to 75% of the original coating thickness.
The authors concluded that thermal sprayed
coatings, which might require replacement
after 21,000 h exposure, were not adequate
in providing corrosion protection for waterwalls under low NOx combustion conditions.
The thickness of the coating in the as-coated
condition for both diffusion coatings was
about 0.25 mm (10 mils). The chromized
coating exhibited a surface layer containing
about 20 to 30% Cr in the as-coated condition. The Cr/Si diffusion coating contained
about 10% Cr, 14% Si, and balance Fe. In a
Fig. 10.20
Fig. 10.19
A corroded carbon steel tube sample from the

waterwall of a boiler (subcritical unit) retrofitted
with a low NOx burner system with overfire air ports. The waterwall tube suffered accelerated wastage after the furnace was retrofitted with NOx burner system. Courtesy of Welding Services
Inc.

electron image) showing the corrosion scales
formed on the fireside of the tube sample (shown in Fig. 10.19).
The chemical compositions of the corrosion scales at different
locations were analyzed semiquantitatively by energy dispersive
x-ray spectroscopy (EDX) analysis. The results are summarized as:
1: 66.9% Fe, 26.6% S, 1.9% Al, 2.6% Si, and minor elements
2: 69.5% Fe, 30.0% S, and minor elements
5: 91.0% Fe, 3.2% S, 4.4% Cl, and minor elements
pg 275
30/10/2007 4:02PM Plate # 0
second test panel in the same boiler after

exposure for about 17,155 h, both chromized
coating and Cr/Si diffusion coating had
completely failed, with corrosion penetrating
into the substrate steel.
Four weld overlays were also included in the
second test panel; they were Type 410 (12%
Cr), Type 309 (23.5% Cr), Type 312 (31%
Cr), and alloy 625 (21% Cr). The chromium
content in the weld wire, as reported in the
paper, is indicated in the parentheses. The
overlays were reported to exhibit about 15%
dilution. Thus, the chromium content in the
overlay was, then, about 10% for Type 410
overlay, 20.0% for Type 309 overlay, 26.4%
for Type 312 overlay, and 17.9% for alloy
625 overlay. After exposure for about
17,155 h, Type 410 overlay with only about
10% Cr suffered general wastage and
suldation at a wastage rate equivalent to half
of that of carbon steel, thus providing no
protection. Alloy 625 with about 18% Cr
showed wastage rates of about 0.125 to
0.25 mm/yr (5 to 10 mpy). Both 309 and 312
overlays with 20% or higher Cr showed no
measurable losses in the overlay thickness.
Fig. 10.21
Weld overlay using the composition that

matches the boiler tube chemistry had often been
used in the past to make temporary repair of the
boiler tube in the eld by manual welding techniques. These manually applied weld overlays
were often referred to as pad welds. In mid1980s, an automated weld overlay technology
using gas metal arc welding (GMAW) process
was developed to apply a corrosion-resistant
overlay alloy onto a large area of waterwall in a
waste-to-energy boiler that burned municipal
solid waste (Ref 36). This modern, automated
weld overlay technology has been expanded
from relatively small waste-to-energy boilers to
large coal-red boilers (Ref 3740). A schematic
showing an example of overlay weld beads
covering the waterwall by a GMAW overlay
welding process is shown in Fig. 10.21.
Figure 10.22 shows an example of the cross
section of a corrosion-resistant overlay on a
waterwall sample that was cut from an overlaid
waterwall. The weld overlay was applied in the
eld using automated overlay welding process.
The weld overlay was characterized with consistency in weld bead sequence and overlay
thickness of each weld bead. This automated
Overlay weld beads applied to cover the waterwall during GMAW overlay welding process. The indicated bead sequence
number is for illustration purposes. The actual overlay welding of the waterwall may not follow the bead sequence number
indicated here. The weld bead typically progresses in a vertical down mode from top (left side of the schematic) to the bottom (right side of
the schematic). Courtesy of Welding Services Inc.
pg 276
30/10/2007 4:02PM Plate # 0
Chapter 10:
GMAW overlay welding process can be carried

out on-site in the boiler or in-shop for overlaid
panels (Ref 3740). Figure 10.23 shows a partial
view of a waterwall that was overlaid in the eld
using a nickel-base alloy. The overlay cladding
can also be accomplished using a laser technology, but the laser cladding can only be performed
in-shop (Ref 4143).
For the weld overlay to provide adequate
resistance to reside corrosion and erosion/corrosion, dilution in chemistry must be low when
overlay welding is applied onto the waterwall. In
overlay welding, the surface layer of the substrate steel must be melted to establish a metallurgical bond between the overlay and the
substrate steel. If too much substrate steel is
melted and mixed with the molten overlay weld
wire, the overlay chemistry will thus be signicantly diluted. This can result in lowering the
concentration of chromium in the overlay too
low to form a protective chromium oxide scale,
thus losing the protective nature of a weld overlay. For example, one utility company selected
Type 430 stainless steel (18% Cr) for eld
application of the overlay on the waterwall of a
boiler using a submerged arc welding process
(Ref 14). The 430 overlay was found to
experience severe wastage problems due to
signicant dilution in overlay chemistry
(Fig. 10.24). The overlay was found to contain
only about 11% Cr, thus making the overlay
vulnerable to oxidation/suldation attack.
This overlay was, thus, not capable of forming
protective chromium oxide scales, instead
forming heavy scales in Fe-Cr oxides and sulfurrich Fe-Cr phases at the scale/metal interface.
Figure 10.25 shows the heavy Fe-Cr oxide scales

formed on this overlay during service.
The most common overlay alloys that have
been applied in both subcritical and supercritical
units to protect the waterwalls from severe reside corrosion attack under the low NOx combustion conditions are austentic stainless steel
Type 309 and two nickel-base alloys 625 and 622
(Ref 3941). Chromium is the most important
alloying element in the alloy (or overlay) to allow
Fig. 10.23
Partial view of a waterwall that was overlaid with

a nickel-base alloy using automated GMAW
process on-site. Courtesy of Welding Services Inc.
Fig. 10.22
An example of an overlaid waterwall sample that

was cut from the overlaid waterwall applied by
automatic overlay welding in the field, showing consistency in
weld bead sequence and overlay thickness of each weld bead.
The waterwall shown in the figure is of a tangent tube design (no
membranes). Courtesy of Welding Services Inc.
Fig. 10.24
A Type 430 stainless steel overlaid waterwall

sample, showing severe overlay wastage and
cracking. Courtesy of Welding Services Inc.
pg 277
30/10/2007 4:02PM Plate # 0
the overlay to form chromium oxide scales to

provide adequate protection by preventing the
formation of suldes and/or iron-rich oxides. The
specication range of chromium is 23.0 to 25.0%
for Type 309 weld wire (ER309), 20 to 23% for
alloy 625 weld wire (ERNiCrMo-3), and 20.0 to
22.5% for alloy 622 (ERNiCrMo-10). The automated GMAW process typically aims to achieve
a dilution in the overlay of approximately 10% or
less. Figure 10.26 shows the chronology of the
total square feet of waterwall overlays from early
1990s to 2001 by a leading applicator of weld
overlay cladding in terms of the year of application for three major overlay alloys (Ref 14).
Beginning in 2000 and 2001, the industry
began switching from alloy 625 to alloy 622 in
waterwall overlay cladding applications in coalred boilers, particularly in supercritical units.
This switch was related to circumferential
grooving of the alloy 625 overlay on the waterwalls of several supercritical units. (The issue of
circumferential grooving and cracking is discussed in section 10.5.3.)
In reside corrosion and erosion/corrosion, it
is desirable to select an overlay alloy with adequate chromium to form protective chromium
oxide scales. Another factor that is viewed by
some boiler operators to be important is the
thermal expansion coefcient mismatch between
the overlay and the substrate ferritic steel.
Nickel-base alloys, such as alloys 625 and 622,
generally have thermal expansion coefcients
that match quite well with carbon and Cr-Mo
steels. Austenitic stainless steels, such as Type
309, exhibit thermal expansion coefcients that
are higher than those of carbon and Cr-Mo steels.
Mean coefcients of thermal expansion for
carbon and Cr-Mo steels (Ref 44), Type 309 (Ref

14), alloy 625 (Ref 45), and alloy 22 (or 622)
(Ref 46) at temperatures of interest for waterwalls are shown in Table 10.9. There exists some
thermal expansion coefcient mismatch between
the 309 overlay and T-11 substrate steel. Nevertheless, no cracks have been found to occur at the
interface (or the fusion boundary) between the
309 overlay and the substrate steel waterwall
tube due to this thermal expansion mismatch.
Figure 10.27 shows typical interface structure
between the 309 overlay and the T11 substrate
steel waterwall tube in a supercritical unit after 10
years of service. The stresses developed due to
this thermal expansion coefcient mismatch
appear to be too small to initiate cracking. Furthermore, both the 309 overlay and the substrate
carbon and Cr-Mo steels exhibit excellent ductility and toughness to prevent any development
of cracks at the interface.
The concern of high stresses that can lead
to cracking at the overlay/substrate interface
during operation in the 309 overlay because of its
higher thermal expansion coefcients prompted
Coleman and Gandy (Ref 47) to examine Type
312 (a duplex stainless steel) as an alternative
stainless steel overlay alloy. Type 312 exhibits
thermal expansion coefcients that are much
lower than those of Type 309, and the alloy
contains higher chromium (about 30% Cr in
the weld wire), making it a potentially good
candidate overlay alloy. However, this alloy is a
duplex stainless steel, which consists of austenite
and ferrite phases, and can suffer severe embrittlement when exposed to a temperature range of
Total overlay area, ft2
35,000
Type 309
Alloy 625
Alloy 622
30,000
25,000
20,000
15,000
10,000
5,000
0
1993 1994 1995 1996 1997 1998 1999 2000 2001

Year overlay performed
Fig. 10.26
Fig. 10.25
Optical micrograph showing nonprotective Fe-Cr

oxide scales formed on the 430SS weld overlay.
Chronology of the total area of the weld overlay

applied to the waterwalls of coal-fired boilers
(both subcritical and critical units) from early 1990s to 2001 as a
function of the year of application for three major overlay alloys
Type 309, alloy 625, and alloy 622by a weld overlay application company. The data covers only field application overlays.
1.0 ft2 = 0.093 m2. Source: Ref 40
pg 278
30/10/2007 4:03PM Plate # 0
Chapter 10:
Table 10.9
pg 279
Mean coefficient of thermal expansion

Coefcient of thermal expansion, 106 in/in/F
Temperature
Carbon steel(a)
1.25Cr-0.5Mo(a)
309 SS(b)
Alloy 625(c)
Alloy 22(d)
70600 F
70800 F
701000 F
7.2
7.6
8.0
7.7
7.8
9.3
9.5
9.7
7.4
7.6
7.8
7.0
7.4
7.7
(a) EPRI data Ref 44. (b) Weld overlay data Ref 14. (c) Wrought alloy data Ref 45. (d) Wrought alloy data Ref 46. Alloy 22 and alloy 622 meet UNS N06022 specication.
Unit conversions: 1.8 106 in/in/F = 106/K = 106 m/m/C; 70 F = 21 C; 600 F =316 C, 800 F = 427 C, 1000 F = 538 C
Overlay
Substrate
0.010 in.
Fig. 10.27
Optical micrograph showing typical interface

structure in longitudinal orientation between the
309 overlay and T11 substrate steel (waterwall tube) after 10 years
of service in a supercritical boiler, revealing no cracking at the
interface. Substrate steel was etched with nital to show the interface. Courtesy of Welding Services Inc.
approximately 371 to 593 C (700 to 1100 F)

due to formation of ordered phases. This
embrittlement phenomenon is commonly referred to as 475 C (885 F). This embrittling
temperature range is within the waterwall metal
temperature range. Furthermore, Type 312 weld
overlay is susceptible to weld solidication
cracking when a weld overlay having a low
dilution is attempted (Ref 48). There have been
several boilers where Type 312 overlay was
applied to protect the waterwalls against reside
corrosion.
In one supercritical unit where Type 312
waterwall overlay was inspected recently after
about 6.5 years of service, and an overlay sample
was obtained from the waterwall for metallurgical examination. Surprisingly, the overlay was
found to exhibit excellent condition, showing
no evidence of cracking or embrittlement, or
circumferential grooving/cracking during service (Ref 49). Typical macro cross sections of
the overlay waterwall tubes after about 6.5
years of service in a supercritical unit are shown
in Fig. 10.28. Figure 10.29 shows typical
Fig. 10.28
Macro cross sections cut from a Type 312 overlay

waterwall panel sample obtained from a supercritical unit showing the overlay after 6.5 years of service in
(a) the transverse cross section and (b) the longitudinal cross
section. The macro samples were etched in nital to reveal the
overlay. The scale is in inches. 1.0 in. = 25.4 mm. Source: Ref 49
metallographic cross section of the 312 overlay

after about 6.5 years of service, showing essentially no corrosion and no cracking.
Bonnington and Brennan (Ref 43) reported
installation of laser-clad 312 overlay panels in
supercritical boilers at the Mirant Mid-Atlantic
station in 2000. Mirant also tested a laser-clad
30/10/2007 4:03PM Plate # 0
Type 446SS overlay in a high-wastage area of a

supercritical unit in 1991. The test results showed
excellent performance of Type 446 overlay in the
boiler (Ref 43). However, the cost of producing a
laser-clad panel with Type 446 was considered to
be prohibitive at that time (Ref 43). Type 446 is a
Fe-25Cr ferritic stainless steel, which is also
susceptible to the 475 C (885 F) embrittlement.
Ferritic steels, such as carbon and low-alloy
steels, typically exhibit higher thermal conductivities than austenitic stainless steels and
nickel-base alloys. As a result, the outer skin
metal temperature of the waterwall tube will be
increased somewhat during service when a weld
overlay cladding of an austenitic stainless steel
(e.g., Type 309) or a nickel-base alloy (e.g.,
alloys 625 and 622) is applied onto a carbon steel
or low-alloy steel. Blough (Ref 50) performed
heat transfer calculation using the thermal conductivities of wrought alloys. His results are
summarized in Fig. 10.30, which compares outer
skin overlay temperatures for Type 309, alloy
625, and alloy 622 overlays with T2 (0.5 Mo
steel) weld metal buildup to the same thickness of

about 0.090 in. (2.3 mm) on the bare steel tube
(Ref 50). For example, for a bare tube with an
outer metal skin temperature of 482 C (900 F),
the application of an overlay of T2 of about
90 mils (2.3 mm) thick to the bare tube increases
the skin temperature by approximately 33 C
(60 F). In comparison with T2 overlay, Type
309 overlay would increase the skin temperature
by about 22 C (40 F) higher than that of T2
overlay. Similarly, alloy 622 would increase the
skin temperature by about 28 C (50 F) higher
than that of T2 overlay, and alloy 625 by about
36 C (65 F) higher than that of T2 overlay.
Type 312 overlay is expected to show lower outer
skin metal temperature than Type 309 (or alloys
625 and 622) because of its duplex microstructure containing a mixture of austenite and
ferrite. Since the application of a weld overlay
using Type 309, Type 312, or nickel-base alloy
has essentially eliminated the general wastage
problem, it becomes desirable to apply a thinner
overlay to lower the outer skin overlay metal
temperature.
The overlays of Type 309, alloy 625, alloy
622, and Type 312 have so far accumulated
various service times with 309 and 625 overlays
staying in service the longest (about 9 to 10
years) in some boilers. These overlays have
essentially eliminated the reside wastage problems caused by staged combustion with low
NOx burners and overre air (Ref 35, 3743).
However, several supercritical boilers have been
found to experience circumferential grooving
and cracking problem. The circumferential
grooving/cracking problem is discussed in section 10.5.3.
09 in.
Fig. 10.29
Optical micrograph showing the transverse cross

section of Type 312 overlay on the waterwall in a
supercritical unit after about 6.5 years of service, revealing
essentially no corrosion and no cracking
Fig. 10.30
Calculated overlay metal temperatures for T2

bare tube with 0.090 in weld metal build-up of
T2 in comparison with 0.090 in overlay of Type 309, alloy 625,
and alloy 622. Source: Ref 50
pg 280
30/10/2007 4:03PM Plate # 0
Chapter 10:
10.5.3 Circumferential Grooving of

Furnace Waterwalls
Fireside corrosion of the furnace waterwalls
particularly in supercritical units has caused
carbon and low-alloy steel waterwall tubes to
suffer circumferential (transverse) grooves that
resemble an alligator hide or an elephant hide. An
example of a waterwall tube exhibiting this type
of circumferential grooving appearance is shown
in Fig. 10.31. This type of waterwall corrosion
has been discussed by Wright (Ref 51) and
French (Ref 15). Wright (Ref 51) indicated that
circumferential grooving has been found to occur
more frequently in supercritical boilers than
subcritical boilers. He also indicated that the
waterwall areas that suffered this type of attack
received the highest heat ux with superimposed
thermal stresses. French (Ref 15) suggested that
the circumferential grooving was caused by a
corrosion-fatigue mechanism. He indicated that
the corrosion products in the circumferential
groove contained sulfur and the ash deposits
on the waterwall tube surface contained carbon
(Ref 15). He thus suggested that the reducing
atmospheres were an essential part of the
grooving-corrosion mechanism. The circumferential grooves in the later stage exhibit a thermalfatigue crack appearance. Some supercritical
boilers historically had circumferential grooving/
cracking problems even prior to the installation
of weld overlay claddings, such as the ones in
pg 281
Conemaugh Station (Ref 52), Morgantown, and

Chalk Point Generating Stations (Ref 53).
In order to determine the root cause of circumferential grooving, it is best to examine the
initial stage of the groove formation. This is
illustrated by several examples for carbon and
low-alloy steel tubes. Figure 10.32 shows the
initiation of several circumferential grooves on a
T22 tube in a supercritical boiler. Scanning
electron microscopy with energy dispersive
x-ray spectroscopy (SEM/EDX) was used to
determine semiquantitatively the chemical compositions of the corrosion products inside the
groove. One of the grooves shown in Fig. 10.32
was analyzed by SEM/EDX with the results
shown in Fig. 10.33. The corrosion products
were mainly enriched in iron and sulfur particularly in light grayish stringers or channels
in the interior region containing higher sulfur.
Areas 2 and 3 are located in the light grayish
channels (or stringers), as shown in Fig. 10.33.
Similar observation was made for a T2 tube
in another supercritical boiler, as shown in
Fig. 10.34, revealing largely iron suldes in the
corrosion products. Again, there were light
grayish phases in a form of stringers or channels containing higher sulfur (Area 2, 3, and 4 in
Fig. 10.34). These stringers or channels, which
tended to form in the interior region, extended
from the surface of the corrosion products and
penetrated along with the overall corrosion
penetration. The one in the center tends to be
much wider and longer. These stringers, which
were found to be more enriched in sulfur than the
surrounding phases, appear to be gas paths
0.0010 in.
Fig. 10.31
An example of the circumferential grooving

encountered in the T2 steel waterwall. Courtesty
of Welding Services Inc.
Fig. 10.32
Optical micrograph showing the initial development of several circumferential grooves on a T22
waterwall tube in a supercritical unit. Courtesy of Welding Services Inc.
30/10/2007 4:03PM Plate # 0
pg 282
formed during the corrosion process. In some

cases, these stringers were quite complex and do
not appear to be the cracks which are subsequently lled with another corrosion product or
ash deposit. Figure 10.35 shows the case of a
very dense group of stringers or channels formed
in the circumferential groove on T11 tube in a
supercritical boiler.
Several supercritical boilers have been found
to experience circumferential grooving for Type
309 overlay (Ref 39) and alloy 625 overlays (Ref
39, 54) on waterwalls. Figure 10.36 shows a
longitudinal cross section of an alloy 625 overlay
waterwall tube at the crown location after 1 year
of service in a supercritical boiler. Two tiny
circumferential grooves are visible. One of the
grooves was analyzed by SEM/EDX, showing
that the alloy 625 overlay had suffered essentially
suldation penetration. Figure 10.37 shows the
corrosion products to be essentially chromium
suldes. There was a centerline channel which
was depleted in chromium (Fig. 10.37b), but

enriched in sulfur, as shown in Fig. 10.37(c), and
also in Ni (the Ni dot map was not shown). Thus,
(a)
140 m
60 m
Fig. 10.33

electron image) showing one of the circumferential grooves formed on a T22 waterwall tube (2.25Cr-1Mo) as
shown in Fig. 10.32. Semiquantative energy dispersive x-ray
spectroscopy (EDX) analysis in terms of weight percent at different
locations of the corrosion products inside the groove is summarized below. Courtesy of Welding Services Inc.
1: 83.9% Fe, 7.1% S, 4.6% Cr, 2.1% Mo, 0.5% Al, 0.7% Si, 0.9%
Mn, and trace elements
2: 72.2% Fe, 21.2% S, 3.6% Cr, 1.3% Mo, 0.5% Al, 0.9% Mn, and
trace elements
3: 73.3% Fe, 22.8% S, 2.2% Cr, 0.6% Mn, and trace elements
4: 80.6% Fe, 9.8% S, 5.1% Cr, 2.5% Mo, 0.9% Mn, 0.5% Al, and
trace elements
5: 87.6% Fe, 1.3% S, 6.7% Cr, 2.6% Mo, 1.2% Al, 0.8% Si, and
trace elements
6: 87.0% Fe, 0.8% S, 7.5% Cr, 2.8% Mo, 0.8% Si, 0.7% Mn, and
trace element
(b)
60 m
Fig. 10.34
(a) Scanning electron micrograph (backscattered

electron image) showing typical morphology of
the circumferential groove formed on a T2 tube (0.5Cr-0.5Mo) in a
supercritical boiler. The results (in wt%) obtained from the semiquantitative analysis using EDX on the corrosion products are
summarized as:
1: 86.8% Fe, 10.5% S, 1.4% Cr, and minor elements
2: 79.7% Fe, 18.2% S, 0.64% Cr, and minor elements
5: 93.7% Fe, 0.5% S, 2.5% Cr, 1.9% Mo, 1.2% Si, and minor
elements
6: 91.8% Fe, 3.7% S, 1.3% Cr, 1.2% Mo, and minor elements
(b) Detail of regions 1 to 4
30/10/2007 4:03PM Plate # 0
Chapter 10:
this centerline channel appeared to be nickel

sulde. Similar observation of the corrosion
products in the circumferential groove formed in
alloy 625 overlay cladding was also reported by
Luer et al. (Ref 54). Figure 10.38 shows a circumferential groove that formed on alloy 625
overlay on a waterwall tube after 2 years of service in another supercritical boiler. The corrosion
products were found to consist of essentially
chromium suldes. The alloy 625 was etched to
show the dendritic microstructure of the weld
overlay. Figure 10.39 shows the formation of
the initial preferential suldation penetration at
a high magnication, revealing in detail the
corrosion phases, which consisted of essentially
chromium suldes.
Type 309 overlay on the waterwall of a supercritical unit was found to suffer circumferential
grooving after 10 years of service. Figure 10.40
shows circumferential grooves at the early stage.
The circumferential groove consisted of several
branches of preferential penetrations. These
several branches of preferential penetrations
tended to converge into a groovelike attack. The
morphology of the circumferential groove (at the
early stage of development) formed in Type 309
overlay (an austenitic stainless steel) is somewhat different from that formed in nickelbase alloy 625. The corrosion products in one of
the branches of preferential penetrations, as
shown in Fig. 10.40, were analyzed by SEM/
EDX with the results summarized in Fig. 10.41.
The SEM/EDX results indicate the corrosion
pg 283
products were enriched in chromium and iron

with some sulfur. Sulfur was particularly high
within the phases inside the light grayish stringers that were clustered around the inner portion
of the corrosion penetration attack. However, the
levels of sulfur were signicantly lower than
those observed in alloy 625 overlay. It is thus
believed that corrosion penetration was essentially combined oxidation and suldation in the
309 overlay.
From the previous discussion of the initial
development of the circumferential grooving in
both alloy 625 and 309 overlays, it is concluded
that the circumferential grooves were initially
developed by way of preferential suldation
penetration. For nickel-base alloy 625 overlay,
suldation was the major mode of preferential
attack, while for Type 309 overlay (an austenitic
stainless steel), suldation along with oxidation
was involved in the preferential attack. As the
preferential attack continued, the circumferential
groove, which penetrated farther into the overlay
with increasing stresses at the penetration tip,
eventually developed into a crack at a later stage.
This is illustrated in Fig. 10.42.
In another power station, Type 309 overlay has
provided protection for the waterwalls of two
supercritical units for 8 years in one unit and
10 years for the other unit. The 309 overlay in
one unit was inspected using nondestructive
testing including dye-penetrant testing in 2006
after 8 years of service. The inspection revealed
0.5 mm
0.0010 in.
Fig. 10.35
Optical micrograph showing a circumferential

groove containing a complex network of
stringers formed on T11 tubes (1.25Cr-0.5Mo) in a supercritical
boiler. Courtesy of Welding Services Inc.
Fig. 10.36
Optical micrograph showing the initial development of two tiny circumferential grooves formed
on the alloy 625 overlay at the crown bead of a waterwall tube in a
supercritical unit after 1 year of service. The metallographic
mount was in the longitudinal cross section. One of the grooves
formed on the overlay (left side of the micrograph) was analyzed
by SEM/EDX with the results shown in Fig. 10.37. Source: Ref 40
30/10/2007 4:03PM Plate # 0
pg 284
no sign of tube wastage and circumferential

cracking (Ref 14). Typical weld overlay is shown
in Fig. 10.43. The unit was equipped with low
NOx burners and overre air, burning bituminous
coal containing 3.0 to 3.5% sulfur. Type 309
waterwall overlay, which has accumulated so far
for 10 years of service, was to be inspected during the next maintenance shutdown.
In laboratory testing of various alloys in
suldizing environments, several authors (Ref
5558) have found that applied tensile stresses
during exposure to a suldizing environment at
elevated temperatures can produce preferential

suldation penetration on wrought alloys. Guttmann et al. (Ref 56) showed that several alloys
were suffering preferential suldation penetration under tensile strains in a suldizing environment. Their tests were conducted at 600 C
(1112 F) in CO-32H2-4CO2-0.2H2S (inlet
gas mixture) with equilibrium sulfur and
oxygen potentials at the test temperature being
about 1011 and 1028 bar (or atm), respectively.
Figure 10.44 shows alloy 45TM (Ni-27Cr-23Fe2.8Si) after exposure to the environment at
(a)
(b)
Fig. 10.37
0.0010 in.
50 m
(c)
50 m
(a) Optical micrograph of a circumferential groove in Fig. 10.36. (b) Chromium x-ray dot map for the corrosion products
inside the groove. (c) Sulfur x-ray dot map for the corrosion products inside the groove. Source: Ref 40
30/10/2007 4:03PM Plate # 0
Chapter 10:
temperature for 2000 h under (a) no external

strain and (b) a 2% strain. The specimen under no
external strain during exposure showed essentially uniform corrosion attack, while the one
under a 2% strain showed preferential penetration attack. The morphology of the preferential
penetration attack is similar to the circumferential groove observed in nickel-base alloy 625
overlay on the waterwall. The preferential
penetration was found to be essentially oxidation/suldation. Also observed in this preferential penetration attack were channels
that were similar to those observed in circumferential groove. The channel and stringers
contain phases with much lighter color, similar to
what was observed in circumferential grooves
formed in Cr-Mo steel waterwall tubes or alloy
625 overlay in few supercritical boilers. Coze
et al. (Ref 59) observed similar phenomena in
testing Fe-12Cr-3Al-3Ti experimental alloy in
H2-34.3H2O-18.5CO2-3.8CH4-7.9CO-1.3H2S
at 600 C (1112 F) 615 h under no external
stresses and external stresses, as shown in
Fig. 10.45. The internal penetration attack under
stresses (upper micrograph of Fig. 10.45) is very
similar to circumferential grooves formed in
Type 430 overlay on the waterwall tube in a
boiler, as shown in Fig. 10.46.
Fig. 10.38
For Fe-Cr-Ni alloys, such as austenitic stainless steels, the stress-enhanced suldation penetration attack takes a different morphology.
Guttmann et al. (Ref 56) found that alloy HR3C
(Fe-25Cr-20Ni-0.5Nb-0.2N), which is a wrought
alloy, suffered stress-enhanced preferential
attack along grain boundaries instead of a ngerlike protrusion, as was observed in alloy
45TM (a nickel-base alloy). Figure 10.47 shows
alloy HR3C after testing in CO-32H2-4CO20.2H2S (inlet gas mixture) at 600 C (1112 F)
under (a) no external stresses showing general
uniform corrosion attack after 2100 h and (b)
a 1.3% strain after 250 h of exposure showing
grain-boundary attack. For Type 309 overlay
(approximately Fe-21Cr-12Ni) on the waterwall, preferential suldation penetration attack
followed individual branches of continuous
penetrations during the initial stage of development in the circumferential grooving, as shown
in Fig. 10.40. These individual branches of suldation penetrations were most likely to be
interdendritic boundaries.
From the discussions in this section, it can be
concluded that circumferential grooving is a
preferential suldation penetration attack under
tensile stresses. Tensile stresses can be varying.
However, thermal cycling is not a prerequisite for
A circumferential goove formed on alloy 625 overlay applied to the waterwall of a supercritical boiler after 2 years
of service. The overlay was etched to show the dendritic microstructure of the alloy 625 overlay. Courtesy of
Welding Services
pg 285
30/10/2007 4:03PM Plate # 0
initiating circumferential grooving, even though

the circumferential groove at a later stage can
develop into a crack.
Wright (Ref 51) indicated that circumferential
grooving typically occurs in the areas of the
waterwall receiving the highest heat ux and is
apparently a result of superimposed thermal
stress. It is believed that the preferential suldation penetration that initiates the circumferential
grooving at certain locations of the waterwall is
due to overheating under suldizing or alternating suldizing/oxidizing conditions. Although a
temperature range for waterwall tubes has been
mentioned to be 400 to 500 C (752 to 932 F)
(Ref 9) and the design temperature of some
(b)
Fig. 10.39
supercritical units was mentioned to be 482

to 510 C (900 to 950 F) (Ref 10) and 482 to
538 C (900 to 1000 F) (Ref 11), it is most
likely that the overheating at certain locations
of the waterwall is due to ame impingement.
Since alloy 625 overlay has been found to suffer
circumferential grooving in several supercritical
units, the outer overlay metal temperature of the
overlay can be estimated based on the aging
characteristics of alloy 625.
Alloy 625 is known to exhibit age hardening
at intermediate temperatures of 538 to 760 C
(1000 to 1400 F). The alloy age hardens by
forming (Ni3Nb) precipitates in this temperature range. At 538 to 593 C (1000 to 1100 F)
(c)
Scanning electron micrograph (a) showing a very early stage of the circumferential grooving formed in alloy 625 overlay
on a waterwall tube in the same supercritical boiler, as shown in Fig. 10.38. EDX spectra (b) and (c) show the alloying
elements associated with the light phases (b) and grayish phases (c), indicating the corrosion phases were essentially chromium sulfides.
The sample surface was plated prior to mounting to protect the corrosion products for SEM examination. Courtesy of Welding Services Inc.
pg 286
30/10/2007 4:03PM Plate # 0
Chapter 10:
0.0010 in.
Fig. 10.40 Optical micrograph showing circumferential

grooves that formed in Type 309 overlay applied
to the waterwall of a supercritical boiler after 10 years of service.
for 1 to 2 years of aging, the alloy forms ne,

coherent precipitates causing signicant hardness increases particularly at 593 C (1100 F).
With aging at 650 C (1200 F), initial precipitates formed in the alloy are ne, coherent ,
but later become incoherent long, platelet Ni3Nb
precipitates, as shown in Fig. 10.48 (Ref 61).
The microhardness prole was determined on
the alloy 625 overlay that was found to show
initiation of circumferential grooving after about
1 year in service (shown in Fig. 10.36). The
results, which are shown in Fig. 10.49, indicate
age hardening to 35 to 45 HRC at the outermost
surface layer of the overlay (Ref 40). The interior
part of the overlay exhibited a hardness level (23
to 25 HRC) that essentially corresponded to the
hardness of the as-overlay alloy 625. Based on
the hardness data, the metal temperature of the
overlay surface layer is believed to have increased to above 538 C (1000 F), most likely
to about 593 C (1100 F), the temperature that
causes signicant hardening for alloy 625.
In another supercritical boiler where alloy 625
overlay on the waterwall suffered circumferential
50 m
Fig. 10.41

electron image) showing the corrosion products
in a circumferential groove formed in Type 309 overlay on the
waterwall of a supercritical boiler after 10 years of service. The
results of the semiquantative EDX analysis of the corrosion products at different locations are as summarized below. Courtesy of
1: 62.2% Cr, 29.2% Fe, 4.6% Mn, 1.7% Ni, 1.4% S, and trace
elements
2: 40.7% Cr, 39.5% Fe, 4.2% Mn, 4.1% Ni, 8.4% S, 2.4% Si, and
trace elements
3: 28.0% Cr, 59.8% Fe, 2.6% Mn, 4.5% Ni, 2.9% S, 1.5% Si, and
trace elements
4: 18.1% Cr, 62.4% Fe, 1.5% Mn, 4.7% Ni, 11.6% S, 1.0% Si, and
trace elements
5: 46.8% Cr, 40.2% Fe, 2.3% Mn, 2.2% Ni, 5.8% S, 2.2% Si, and
trace elements
6: 45.8% Cr, 43.2% Fe, 3.2% Mn, 2.3% Ni, 3.5% S, 1.5% Si, and
trace elements
Fig. 10.42
Some well-developed circumferential grooves,

which developed into cracks at a later stage,
were observed in the 309 overlay on the waterwall after 10 years
of service. There was no evidence of cracking developed at the
interface between the 309 overlay and the T11 substrate even
with some thermal expansion coefficient mismatch. The substrate
steel was etched with nital to reveal the fusion boundary. Original
magnification: 25
pg 287
30/10/2007 4:03PM Plate # 0
grooving after 2 years of operation, the overlay

surface layer was also found to exhibit age
hardening to about 40 to 42 HRC.
Unlike alloy 625, Type 309 does not exhibit
age-hardening characteristics. As a result, the
possible overheated overlay surface temperature
could not be estimated using microhardness data.
Chromium carbides have been found to precipitate along grain boundaries in a Type 309
waterwall overlay after 10 years of service in a
supercritical unit, as shown in Fig. 10.50. Precipitation of these chromium carbides may be an
indication that the overlay was overheated to
above 538 C (1000 F). In performing metallurgical evaluation of a 309 overlay sample that
was removed from the same boiler in a waterwall
area (probably from a different location as the
area discussed in Fig. 10.50) after 7 years of
service, the microstructure of the overlay showed
no chromium carbide precipitates formed along
grain boundaries (Fig. 10.51) (Ref 37). The
temperature of the steam in this area was estimated by a plant engineer to be 371 to 427 C
(700 to 800 F) (Ref 14). Without the evidence of
chromium carbides along grain boundaries, the

overlay metal temperature was likely to be below
538 C (1000 F). Microhardness measurements
made across the overlay that corresponds to
the microstructure containing carbides, as shown
in Fig. 10.50, indicated about 250 HV (about 100
HRB) as opposed to 200 HV (about 90 HRB)
for the overlay that corresponds to the
microstructure with no carbides, as shown in
Fig. 10.51. The 309 overlay is typically applied
using a low-carbon Type 309 weld wire
(ER309LSi). Thus, age-hardening response is
generally minimal. Accordingly, the development of preferential grooving and later cracking
was not related to the aging behavior of the 309
weld overlay, which is expected to exhibit good
ductility, after exposure to 538 to 593 C (1000
to 1100 F) for 10 years.
Fig. 10.43
Fig. 10.44
Type 309 overlay on the waterwall in another

supercritical unit equipped with low NOx burners and overfire air after service for 8 years, showing no sign of
metal wastage or circumferential cracking. The boiler reportedly
burned bituminous coal containing 3 to 3.5% S. Courtesy of
Alloy 45TM (Ni-27Cr-23Fe-2.8Si) after exposure

to the test gas of CO-32H2-4CO2-0.2H2S (inlet
gas mixture) at 600 C (1112 F) for 2000 h under (a) no external
strain and (b) 2% strain. The equilibrium sulfur and oxygen
potentials for the test environment at the test temperature were
about 1011 and 1028 bar (or atm), respectively. Source: Ref 56
pg 288
30/10/2007 4:03PM Plate # 0
Chapter 10:
Tensile ductility after long-term aging may

not be an important factor in the development
of circumferential cracking. An example is illustrated by Type 312 waterwall overlay that
has been in service in a supercritical unit for
about 6.5 years with no evidence of circumferential cracking being revealed by dye-penetrant
testing during a recent inspection (Ref 14).
Samples obtained from the overlaid waterwall
for metallurgical examination also showed
similar results. Figure 10.28(b) shows a longitudinal macro cross section of the overlay sample, revealing no circumferential grooving or
cracking. Figure 10.52 shows typical overlay
surface condition in a longitudinal metallographic cross section, again, revealing no
circumferential grooving or cracking. The

microhardness prole across the 312 overlay at
the crown bead from two samples is shown in
Fig. 10.53. Both samples showed some hardening at the region close to the overlay surface
with one sample showing signicantly more
hardening (close to 50 HRC). Type 312 overlay
is a duplex stainless steel containing a mixture
of austenite and ferrite phases. When exposed to
a temperature range of approximately 343 to
593 C (650 to 1100 F), the ferrite phase is
subject to ordering reaction forming ordered
precipitates, resulting in signicant hardening
and causing the alloy to become brittle. This
phenomenon is commonly referred to as the
475 C (885 F) embrittlement. Even under this
brittle condition, the 312 overlay showed no
evidence of circumferential grooving or cracking. The overlay was found to contain about 29%
Cr. The corrosion products observed on the
Fig. 10.45
Fe-12Cr-3Al-3Ti experimental alloy after testing

in H2-34.3H2O-18.5CO2-3.8CH4-7.9CO-1.3H2S
at 600 C (1112 F) for 615 h under no external stresses (lower
micrograph) and external stresses (upper micrograph). The
external strain applied during test was unknown. Source: Ref 59
Fig. 10.47
Fig. 10.46
Circumferential grooves formed in Type 430 weld

overlay on the waterwall tube in a coal-fired
boiler. Source: Ref 60
Alloy HR3C after testing in CO-32H2-4CO20.2H2S (inlet gas mixture) at 600 C (1112 F)
under (a) no external stresses showing general uniform corrosion
attack after 2100 h and (b) a 1.3% strain after 250 h of exposure
showing grain-boundary attack. Source: Ref 56
pg 289
30/10/2007 4:03PM Plate # 0
overlay were found to be chromium-rich oxides,

thus providing adequate protection against
preferential suldation penetrations, a precursor
of circumferential grooving and cracking.
In the laboratory tests discussed above,
applications of about 1 to 2% tensile strains on
the test specimens were capable of causing preferential suldation penetration at about 593 C
(1100 F) for less than a couple of thousands of
hours of test duration for various alloys including
nickel-base alloys and Fe-Cr-Ni alloys. It would
appear that under lower tensile strains (e.g., 0.5
Fig. 10.48
to 1%) the alloys could develop preferential

suldation penetration after longer exposure
times, such as a year or longer). For circumferential grooving to develop, the axial stresses
would have to be large enough to develop tensile
strains of approximately 0.5 to 1%. Axial stresses
generated on the waterwall tubes may come from
the dead load of the waterwall panels, thermal
stresses, pressures of water/steam inside the tube,
and bending stresses due to localized ame
impingement (overheating) due to constraints
from the surrounding cooler waterwall areas as
Aging behavior of alloy 625 (wrought alloy samples) at 650, 760, and 870 C (1200, 1400, and 1600 F) for 16,000 h in
terms of microstructure, losses in impact toughness and elongation, and increases in strength. Source: Ref 61
pg 290
30/10/2007 4:03PM Plate # 0
Chapter 10:
scales in a suldizing environment, thus allowing

preferential suldation attack to take place. Three
conditions are required to be met for developing
circumferential grooving for the weld overlaid waterwall: (a) suldizing environment,
(b) adequate tensile strains (e.g., 0.5 to 1%), and
(c) overheating of tube outer skin metal to
approximately 593 C (1100 F).
For widely used weld overlays of Type 309,
alloy 625 and alloy 622, the chromium content
of the weld overlay at the crown location (i.e.,
the location subjected to the highest heat ux)
was typically 20%. This level of chromium in the
Rockwell C, HRC
well as buckstays. (Buckstays are structural

shapes or trusses that encircle and restrain the
movement of the furnace waterwalls caused by
uctuation in furnace pressure, or transient
internal or external loads.) The combination of
these stresses may be adequate to lead to the
development of preferential suldation penetrations and circumferential grooving when the
overlay outer layer is heated to approximately
593 C (1100 F) under localized suldizing
conditions. The development of preferential
suldation penetrations are believed to result
from local breakdown of the protective oxide
Fig. 10.49
Microhardness profile measured using Vickers hardness tester with a 500 g load as a function of the distance from the
overlay surface for alloy 625 weld overlay on the waterwall of a supercritical boiler after 1 year of operation when
circumferential grooves, as shown in Fig. 10.36, were initiated. Vickers hardness values (HV) were converted to Rockwell C (HRC) values.
Hardening of the overlay surface layer (within 0.5 mm, or 20 mils) is believed to result from age hardening of alloy 625 due to formation of
fine, coherent (Ni3Nb) precipitates when heated to probably 593 C (1100 F). Source: Ref 40
Fig. 10.50
Optical micrograph showing typical microstructure of the 309 overlay on the waterwall that suffered circumferential
grooving (Fig. 10.40) in a supercritical boiler after 10 years of service. Original magnification: 200
pg 291
30/10/2007 4:03PM Plate # 0
weld overlay may not be adequate when the overlaid waterwall is subject to the aforementioned
three conditions. To avoid the local breakdown
of the chromium oxide scales under these conditions, an overlay containing a higher level of
chromium is required to ensure the adequate
chromium content to maintain the protective
chromium oxide scales over a prolong service
duration. Alloy 52 (AWS ERNiCrFe-7 or UNS
N06052) with 30Cr, 9Fe and balance Ni provides
a good candidate overlay alloy with potentially
improved resistance to circumferential grooving
and cracking. Alloy 52 overlay, which was applied to the waterwall of a supercritical unit, has
now been in service for 3 years with no reported
degradation problems (Ref 14). Paul et al.
(Ref 62) indicated some eld testing of alloy 33
(UNS R20033: Fe-33Cr-32Ni) overlay test
panels in two supercritical boilers, showing
no cracking after slightly less than 2 years of
testing. Both boilers were tangential ring with
low NOx and overre air. Alloy 33 test panels
were located in the overre air region in both
boilers. In both cases, alloy 622 overlay panels
were installed at the same time in the same area.
After slightly less than 2 years of exposure, both
alloy 33 and alloy 622 overlays tube samples
with similar exposure times were removed for
metallurgical evaluation, showing no evidence of
circumferential cracking for both overlay alloys
(Ref 62).
10.5.4 Sootblowing
0.10 mm
Fig. 10.51
Microstructure of Type 309 overlay on the

waterwall after 7 years of service from the same
boiler, but likely from different area in the boiler. No chromium
carbides along grain boundaries were detected in this area. The
overlay showed no circumferential grooving or cracking. Source:
Ref 37
One important by-product of the combustion

of coal is ash. During combustion, some of the
mineral constituents and compounds can be in a
molten or plastic state. These ash constituents
can then deposit on the heat-absorbing surfaces
causing slagging and fouling. If ash reaches the
heat-absorbing surface at a temperature near its
softening temperature, the resulting deposits are
likely to be porous and can be removed by
sootblowing. Slagging is the deposition of molten, partially fused deposits on the furnace walls
and the upper furnace radiant superheaters
exposed to radiant heat (Ref 1). Fouling is the
deposition of more loosely bonded deposits on
the heat-absorbing surfaces in the convection
path, such as superheater and reheater, that
0.0010 in.
Fig. 10.52
Optical micrograph showing typical longitudinal

cross section at the crown bead of the 312
waterwall overlay after about 6.5 years of service in a supercritical
unit. The corrosion scales were chromium-rich oxides as identified by SEM/EDX analysis. Courtesy of Welding Services Inc.
Fig. 10.53
Microhardness profile across the overlay from the

overlay surface measured using Vickers hardness
tester with a 500 g load. Vickers hardness values (HV) are converted to Rockwell C (HRC) values. Data were obtained from two
different overlay samples (Series 1 and 2). 1 in. = 25.4 mm.
Source: Ref 49
pg 292
30/10/2007 4:04PM Plate # 0
Chapter 10:
are not exposed to radiant heat (Ref 1). Soot

blowers using steam are generally adequate
for removing the ash deposits on the fouled
tube surfaces, while waterlances or water cannons using thermal shock by room-temperature
water may be needed to clean the slagged tube
surfaces.
The damage that causes the tube surface due to
steam sootblowers is generally referred to as
sootblower erosion. Under normal operations,
steam sootblowing is in the erosion/corrosion
(i.e., erosion/oxidation) regime. An example is
given here to illustrate this phenomenon. A
905 MW(e) supercritical unit had experienced
severe waterwall tube wastage problem with its
SA213 T11 tubes (1.25 in. outside diameter
0.220 in. minimum wall thickness) subjected to
steam sootblowing by wall blowers. These
waterwall tubes had to be replaced every 18
months because of severe tube wastage. The
boiler, which was not equipped with low NOx
burners, burned the Illinois basin coal with about
2.5% S. The corrosion products were most likely
iron oxides with possibly some iron suldes
(if suldation also took place). No analysis was
done on the corrosion products formed on T11
waterwall tube. However, when the soot blower
affected area was weld overlaid with Type 309
using automatic GMAW process, the wastage
problem was essentially eliminated. A Type 309
overlay tube sample was removed after 7 year of
service for metallurgical evaluation. During this
7 year period, the boiler had burned the Illinois
basin coal (2.5% S) for about 4 years and an
Eastern Appalachian low sulfur coal for 3 years
with low NOx burners and separated overred air
during these 3 years. The overlay showed no
evidence of metal wastage and erosion/corrosion
attack by steam sootblowing (Fig. 10.54 and
10.55).
The reason for signicant life improvement
offered by the 309 overlay is the erosion/corrosion resistance of the 309 overlay. The 309
overlay forms chromium oxide scales, which are
much thinner and grow signicantly more slowly
than iron oxides that form on T11 (1.25Cr-0.5Mo
steel). Steam impingement on the steel removes
iron oxide scales, which are thick, nonprotective,
and fast growing. The iron oxide scales reformed
repeatedly after they were removed by steam
impingement, thus resulting in accelerated
wastage. Once the tube was protected by the 309
overlay, metal wastage due to erosion/corrosion
was signicantly reduced due to the formation
of thin, protective chromium oxide scales.
Nickel-base alloy overlays, such as alloy 622

overlay, that form protective chromium oxide
scales will have similar resistance to sootblower
erosion-corrosion. More discussion on erosion/
corrosion is covered in Chapter 8: Erosion and
Erosion-Corrosion.
Fig. 10.54
Field-applied Type 309 overlay on the waterwall

tube after 7 years of service in a 905 MW(e)
supercritical boiler: (a) close-up view of the crown and side beads
of the overlay, and (b) cross section of the crown bead of the
overlay showing both the overlay surface and the overlay/substrate interface (as-polished condition, original magnification:
25). Courtesy of Welding Services Inc.
pg 293
30/10/2007 4:04PM Plate # 0
pg 294
10.5.5 Deslagging by Waterlances and

Water Cannons
One of the important characteristics of ash is
its ash fusion temperatures. If the ash fusion
temperatures are low enough, the ash deposit
reaches its melting point (due to the ashs thermal
insulating properties) and the molten slag runs
down the furnace wall surface (Ref 1). When
solidied, this slag is tightly bonded and is difcult to remove. This slag may require waterlances or water cannons to create thermal shock
for the removal of this slag deposit. The furnace
walls are likely locations for developing this
slagging problem when coal with low ash fusion
temperatures is combusted.
When room-temperature water is used for
deslagging the waterwall through either waterlances or water cannons, the waterwall tube is
subjected to thermal shock. Repeated use of
waterlances or water cannons for deslagging can
cause thermal fatigue cracking of the waterwall
tubes. Thermal fatigue cracking of waterwall
tubes due to deslagging by waterlances or water
cannons has been reported by Kessler (Ref 63),
Carlisle et al. (Ref 64), Ray et al. (Ref 65), and
Blinka (Ref 66). Kessler (Ref 63) reported eld
measurement data that indicated a rapid temperature drop from approximately 379 to 317 C
(715 to 603 F) for the waterwall tube right after
water spraying from water cannons. He indicated
in Ref 63, that this temperature drop measured
from the eld was signicantly lower than the

EPRIs 288 C (550 F) drop in temperature
measured by a waterlance simulator used to test
the waterwall panels when sprayed without a slag
(no thermal insulating by a slag). The boiler at
Plant Miller (Alabama Power), which burned
Powder River Basin (PRB) coal, used waterlances for continuous waterwall cleaning with a
3 hour cycle time. This resulted in reduced
waterwall tube life due to quench cracking (Ref
64). Ray et al. (Ref 65) reported that a boiler
suffered waterwall tube damage (a waterwall
blowout) after 2 years of unrestricted
waterlance usage.
The morphology of thermal fatigue cracking
on carbon steel waterwall tube as a result of
water spraying from water cannons is illustrated
in Fig. 10.56. Surface appearance of thermal
fatigue cracks on a waterwall tube due to
water spraying from waterlances is shown in
Fig. 10.57. Figure 10.58 shows typical morphology of the circumferential cracks in a longitudinal cross section from the sample shown in
Fig. 10.57. Figure 10.59 shows an SEM of a
circumferential crack along with the energy dispersive x-ray spectroscopy (EDX) analysis on
the corrosion product inside the crack. The EDX
analysis showed the corrosion product inside the
crack to be essentially iron oxides.
One boiler operator decided to repair the
cracked waterwall tubes, which were made of
SA210 A1 steel, due to water spraying from
waterlances by applying alloy 622 weld overlay
cladding in the eld after the cracks were ground
0.1 mm
Fig. 10.55
Optical micrograph showing a very thin oxide

scale on the surface of the 309 overlay on a
waterwall tube (T11) after 7 years of service subjected to steam
sootblowing in a supercritical boiler, revealing no evidence of
erosion/corrosion damage or cracking. The 309 overlay was not
etched showing white portion of the micrograph with a magnification marker at the bottom of the micrograph. Source: Ref 40
0.5 mm
Fig. 10.56
Optical micrograph showing typical morphology

of circumferential thermal fatige cracking on a
carbon steel waterwall tube due to water spraying from water
cannons. The tube OD (fireside) is on the left side of the micrograph and the ID (water/steam side) is on the right side. Courtesy of
30/10/2007 4:04PM Plate # 0
Chapter 10:
off from the tubes. The repaired waterwall area

was more than 420 m2 (4500 ft2). After 4 years
of service, dye penetrant testing (PT) of the
whole overlaid waterwall area showed no evidence of cracking. Figure 10.60 shows typical
overlay area after PT testing.
10.5.6 Superheater and Reheater Corrosion
Steam from either the steam drum in a subcritical unit or from the furnace wall in a supercritical unit rst passes through a primary
Fig. 10.57
Appearance of thermal fatigue cracks occurred

on a carbon steel waterwall tube (viewed from 12
oclock crown position) due to water spraying from waterlances.
Source: Ref 40
0.5 mm
Fig. 10.58
Optical micrograph showing circumferential

thermal fatigue cracks that developed on a carbon steel waterwall tube (shown in Fig. 10.57) due to water
spraying from waterlances. Source: Ref 40
superheater, typically a horizontal heat exchanger located above the economizer, and then
through a secondary or nishing platen superheater (i.e., tubes are in a at arrangement and
hung in the furnace roof) with a number of platens in parallel. In some boilers, this nishing
superheater is heated by radiant heat from the
furnace combustion. The steam leaving the secondary or nishing superheater passes through a
high-pressure steam turbine. After expanding
through the high-pressure steam turbine, the
steam is returned to the boiler to be reheated in a
reheater. Steam from the reheater then passes
through an intermediate-pressure turbine, followed by passing through a low-pressure turbine.
Typical superheated steam temperature in most
utility boilers in the United States is about
538 C (1000 F). The metal temperatures of
superheaters and reheaters may be up to 650 C
(1200 F).
Superheater and reheater tubes suffer oxidation attack at lower temperatures. Oxidation
attack generally results in lower wastage rates.
When ue gas stream is entrained with y
ash, the tubes can also suffer erosion/corrosion
attack. The wastage rates under erosion/corrosion conditions can be much higher. When
the metal temperature is approaching 650 C
(1200 F) or higher, superheater and reheater
tubes can suffer accelerated wastage rates due
to coal ash corrosion. The accelerated wastage
rates due to coal ash corrosion are the result
of molten salt (sulfate) corrosion. French
(Ref 15) suggests that the corrosion follows a
hockey stick type behavior with two distinctive corrosion regimes: low metal wastage
rate by oxidation and accelerated wastage
rates by molten salt corrosion, as shown in
Fig. 10.61. The molten salt corrosion behavior
exhibits a bell-shaped curve with respect to
temperature for austenitic stainless steels. The
rate increases with temperature to a maximum,
then decreases with increasing temperature.
The accelerated corrosion associated with this
bell-shaped curve is related to the formation
of molten alkali metal-iron-trisulfate [(Na,K)3Fe
(SO4)3] (Ref 67, 68). Variation of the sodium-topotassium ratio greatly affects the melting point
of the complex sulfate in an ash deposit (Ref 68).
Figure 10.62 illustrates a wide variation of
melting points of mixtures of sodium iron trisulfate and potassium iron trisulfate (Ref 68).
Corrosion rate was also affected by the sodiumto-potassium ratio in the complex sulfate (Ref
68). Nelson and Cain (Ref 69) conducted a series
pg 295
30/10/2007 4:04PM Plate # 0
Fe
10,000
9,000
8,000
Counts
7,000
6,000
EDS-1 corrosion products
5,000
4,000
3,000
2,000
1,000
(a)
120 m
O
Fe
C Mn
Fe
Si S
Cr Mn
0
0.000 1.000 2.000 3.000 4.000 5.000 6.000 7.000 8.000 9.000 10.000 11.000 12.000 13.000 14.000 15.000
keV
(b)
Fig. 10.59
(a) Scanning electron micrograph (backscattered electron image) showing a circumferential thermal fatigue crack (from
the sample shown in Fig. 10.57) along with (b) an EDX spectrum showing the corrosion product inside the crack to be
essentially iron oxides. Source: Ref 40
Fig. 10.61
Fig. 10.60
Tube wastage as a function of tube metal temperature after 125,000 h, with low wastage rate
by oxidation and accelerated wastage rates by molten salt corrosion. Source: Ref 15
of laboratory tests with owing synthetic ue gas

(N2-15CO2-3.6O2-0.25SO2) over a mixture of
potassium sulfate, sodium sulfate, and iron oxide
(molecular ratio 1.5:1.5:1.0) that covered the test
coupons. Samples were exposed at different temperatures for 5 days. Results are summarized in
Fig. 10.63 (Ref 69). The corrosion rate was
greatly enhanced when the sulfate was molten.
Corrosion rate also depends on SO2 concentration in the ue gas and Na2SO4 + K2SO4 concentration in the ash deposit, as illustrated in
Fig. 10.64 and 10.65 (Ref 70). For alkali sulfates,
only the acid-soluble ones are of concern in coalash corrosion (Ref 71).
The typical corroded superheater or reheater

tube is characterized by two wastage ats at
about the 2 oclock and 10 oclock positions, as
shown schematically in Fig. 10.66 (Ref 13). At
these two locations, where the ash was thin and
the heat ux was high, a molten sulfate layer
formed, resulting in severe corrosion attack
(Ref 13). On the other hand, the front face of the
tube, with a heavy ash deposit providing sufcient insulation to keep the metal surface temperature below the melting point of the sulfate,
suffered signicantly less corrosion attack
(Ref 13).
Figure 10.67 shows the cross section of a
Type 304H reheater tube after service for about
4 years with the reheated steam at about 538 C
(1000 F) suffering coal-ash corrosion attack in a
subcritical boiler. The locations that suffered the
Alloy 622 overlay (dye penetrant tested) after 4

years of service involving the use of waterlances
for deslagging. The overlay was applied onto the carbon steel
waterwall after thermal fatigue cracks caused by waterlances
were ground off. The dye penetrant testing showed no cracking.
pg 296
30/10/2007 4:04PM Plate # 0
Chapter 10:
worst coal-ash corrosion attack were at

approximately the 2 and 10 oclock positions
(when the ue gas impinges at 12 oclock position). The boiler used staged combustion and
overre air for NOx reduction control. The surface appearance of the maximum wastage area is
shown in Fig. 10.68. SEM/EDX analysis was
performed on the corrosion scale formed on the
maximum wastage area with the results being
summarized in Fig. 10.69 (Ref 14). The analysis
shows suldation attack along with oxidation.
Fig. 10.64
Effect of SO2 content in flue gas on the corrosion

of several superheater/reheater materials exposed to synthetic ash containing 5 wt% (Na2SO4 + K2SO4) at
650 C (1200 F). Source: Ref 70
Fig. 10.62
Melting point curve for the K3Fe(SO4)3-Na3Fe

(SO4)3 system. Source: Ref 68
Fig. 10.65
Effect of Na2SO4 + K2SO4 content in synthetic ash

on the corrosion of several superheater/reheater
materials at 650 C (1200 F) in flue gas containing 0.25% SO2.
Source: Ref 70
Fig. 10.63
Results of laboratory tests with flowing synthetic

flue gas (N2-15CO2-3.6O2-0.25SO2) over a
synthetic coal ash (K2SO4, Na2SO4, and Fe2O3 with a molecular ratio of 1.5:1.5:1.0) that covered the test coupons. Source:
Ref 69
Fig. 10.66
Ref 13
Typical wastage feature of a corroded superheater

and reheater tube from a coal-fired boiler. Source:
pg 297
30/10/2007 4:04PM Plate # 0
It was further observed that carburization

occurred at the locations (i.e., about 2 and 10
oclock positions) where coal-ash corrosion took
place, as shown in Fig. 10.70. However, the
Fig. 10.67
Cross section of a Type 304H reheater tube

showing two wastage flats with the maximum
wastage. Note the wastage flats on both sides of the tube surface
where the flue gas impinging at the 90 location (i.e., facing the
ruler in the photo). Courtesy of Welding Services Inc.
Fig. 10.68
surface that did not suffer coal-ash corrosion

attack, such as at the 3 or 9 oclock position,
showed no carburization (Fig. 10.71). It is
believed that carburization was not the cause that
resulted in coal ash corrosion because of the
formation of chromium carbides that reduced the
chromium level in the metal matrix. On the
contrary, coal-ash corrosion (or suldation)
caused the metal surface to form suldes along
with unprotective chromium oxide scales, thus
resulting in carburization when the localized
environment was under reducing conditions with
the presence of CO and/or unburned carbon soot.
A superheater tube made of Type 304 suffering coal-ash corrosion attack is shown in
Fig. 10.72. Similar to the reheater tube discussed
earlier (Fig. 10.67), the worst attack occurred at
the 2 and 10 oclock positions (when the ue gas
was impinging at the 12 oclock position). The
morphology of corrosion attack at the worst
corrosion attack area is shown in Fig. 10.73,
showing general wastage and some internal
attack in the matrix as well as intergranular attack. Figure 10.74 shows a scanning electron
micrograph (backscattered electron image),
showing both general corrosion scales and
internal attack. The EDX analysis of the corrosion products indicated suldes (marked as 1)
and internal oxide phases (marked as 2). Area 1
area was found to contain primarily 45.5%
Surface appearance of one of the wastage flats with the maximum wastage of the 304H reheater tube shown in Fig. 10.67.
pg 298
30/10/2007 4:04PM Plate # 0
Chapter 10:
Cr, 33.5% Fe, 9.2% Ni, and 8.6% S (in wt%)

and area 2 primarily 56.8% Fe, 27.5% Cr, and
11.7% Ni.
In the same boiler, Type 304H reheater tubes
suffered maximum wastage rates at about 30
locations on either side of the direct ue gas
impingement position, as shown in Fig. 10.75.
The wastage pattern in this case is believed to
be caused by y-ash induced erosion/corrosion
instead of coal-ash corrosion. Figure 10.76
shows the cross section of the maximum wastage
area where y-ash deposits were in contact with
the metal surface. The EDX analysis of the yash deposits, which are marked as 1, showed
primarily 46.4% Si, 21.6% Al, and 20.7% Fe.
Impinging y-ash particles removed oxide scales
from the metal surface that prompted the fresh
metal surface to form oxides again, which were
21 m
Fig. 10.69

electron image) showing the corrosion products
formed on the maximum wastage area of Type 304H reheater
shown in Fig. 10.67. Semiquantative EDX analysis shows the
compositions (wt%) at different locations as indicated below.
1: 34.4% Fe, 49.1% Cr, 5.6% Ni, 4.9% S, 3.7% Mn, 1.3% Si, and
trace elements
2: 34.7% Fe, 17.6% Cr, 22.8% Ni, 22.3% S, 1.6% Mn, 0.6% Si,
and trace elements
3: 32.1% Fe, 45.5% Cr, 6.8% Ni, 7.6% S, 4.5% Mn, 0.9% Si, 0.7%
Na, and trace elements
4: 28.9% Fe, 49.0% Cr, 6.3% Ni, 6.6% S, 5.0% Mn, and trace
elements
5: 33.7% Fe, 17.6% Cr, 22.4% Ni, 23.1% S, and trace elements
elements
7: 37.7% Fe, 48.3% Cr, 1.8% Ni, 6.8% S, and trace elements
elements
again removed by subsequent impinging y-ash

particles. This repeated oxide removal and
reformation process caused the wastage rate to be
increased due to erosion-enhanced corrosion
reaction. (Detailed discussion of erosion/corrosion is covered in Chapter 8) Some internal attack
was observed underneath the yash deposits.
Internal attack was essentially in form of oxides
along grain boundaries (marked 2, 3, and 4).
Trace of chlorine was detected in the phase
marked by 4. In adjacent area, which was still on
the at wasted face (but near the area shown in
Fig. 10.76) relatively thin oxide scales were
observed. These oxide scales appeared to be
fragmented and fractured possibly by impinging
yash (Fig. 10.77). The EDX analysis showed
presence of sulfur in these oxides. These oxides
were likely the subsequently reformed oxides
that formed after preceding oxides were removed
by impinging y-ash particles.
At the position where the ue gas directly
impinging on the tube in the same 304H reheater
tube sample, signicant amount of ash deposits
was observed, as shown in Fig. 10.78. At this
location, the tube wall wastage was signicantly
less. The EDX analysis was performed in the
areas in (a) the top portion of the deposits
(Fig. 10.79), (b) midsection of the deposits/
corrosion products (Fig. 10.80), and (c) the
deposits/corrosion products near and at the
interface (Fig. 10.81). The top layer was found to
be essentially ash deposits enriched in aluminum,
silicon, iron, sulfur and arsenic, and calcium.
The midsection also consisted of ash deposits
enriched primarily in silicon, arsenic and iron.
Corrosion products formed on the tube were
found to consist of essentially Cr-Fe-rich suldes
and Fe-Ni-Cr-rich suldes. Some internal suldes were also observed.
Many investigators have been using laboratory simulation tests involving synthetic ash
(typically Na2SO4, K2SO4, and Fe2O3) to cover
the test specimens with a owing synthetic ue
gas (typically N2-O2-CO2-H2O-SO2) in a test
retort to rank alloy performance for resistance to
coal-ash corrosion. Figure 10.82 shows the
results of laboratory coal-ash corrosion tests
involving various austenitic stainless steels,
Fe-Ni-Cr alloy 800H, and 50Ni-50Cr alloy
671 (Ref 72). The best performer was the alloy
with the highest chromium content (alloy 671).
Resistance to coal-ash corrosion has been found
to increase with increasing chromium content in
the alloy. Castello et al. (Ref 73) summarized the
coal-ash corrosion data generated by various
pg 299
30/10/2007 4:04PM Plate # 0
investigators to show the benecial effect of

chromium in resisting coal-ash corrosion in
Fig. 10.83.
Kihara et al. (Ref 78) also presented their
laboratory simulated coal-ash corrosion tests as a
function of chromium contents in the alloys in
Fig. 10.84. Synthetic ash used in laboratory
testing consisted of 2.5, 5, and 20% Na2SO4 +
K2SO4 (equal amounts) with Fe2O3, Al2O3 and
SiO2 (equal parts). Tests were performed at 600,
650, and 700 C (1112, 1200, and 1292 F) for
times up to 300 h in a synthetic ue gas containing 0.05, 0.1, 0.25, and 1.0% SO2. Plant
exposure data were also included in the gure.
Plant exposure data were obtained using corrosion probes inserted into the operating boiler at
Tennessee Valley Authoritys Gallatin Station
Unit No. 2. Alloys tested included 17-14 CuMo
(17Cr-14Ni austenitic stainless steel), Type
347, 25Cr-20Ni-Nb, 21Cr-11Ni-Si-Ce austenitic
stainless steel, 20Cr-18Ni-Si-Al austenitic stainless steel, alloy 800, 30Cr-45Ni-2Mo, and
Fig. 10.70
Services Inc.
chromized T91. During the 16,000 h exposure,

the boiler burned coal containing 3.08 to 3.26%
S, 8 to 11% ash. Field data showed signicant
scattering with alloys containing about 20% Cr.
Alloys containing about 25% Cr and higher were
found to perform signicantly better with less
data scattering.
At TVAs Gallatin Station, previous 304 and
321 stainless steel reheaters in Unit No. 2 (a
subcritical boiler) typically required replacement
after 7 years of service (Ref 79). The reheater
made of higher chromium-containing alloy
HR3C (25Cr-20Ni-0.5Nb-0.25N) had been in
service for about 45,004 h (approximately 5
years) without causing any forced outage due to
corrosion-induced tube wastage problem
(Ref 79). In order to develop coal-ash corrosion
data for superheaters and reheaters at higher
steam temperatures in ultra-supercritical boilers,
a eld exposure test program was initiated that
involved using tube sleeves to increase metal
temperatures. Samples were exposed at 521 to
Optical micrograph showing Type 304H reheater (the same one shown in Fig. 10.67) that suffered carburization
on the surface where severe coal-ash corrosion took place (about 2 and 10 oclock positions). Courtesy of Welding
pg 300
30/10/2007 4:04PM Plate # 0
Chapter 10:
Fig. 10.71
Type 304H reheater (the same one shown in Fig 10.67) at 3 or 9 oclock positions, where no severe coal-ash corrosion
attack occurred, showing no carburization attack. Courtesy of Welding Services Inc.
685 C (970 to 1265 F) for about 45,004 h.

Blough et al. (Ref 79) summarized the test
results on the tube shields (higher metal temperatures) in Fig. 10.85, showing metal wastage
rates as a function of metal temperature. Two
highest chromium-containing alloys, HR3C
(25Cr), and CR30A (30 Cr), performed the
best. The data in Fig. 10.85 show what appeared
to be low wastage rates for chromized T22 at
585 C (1085 F). In fact, the chromizing
layer with about 0.33 mm (0.013 in.) in thickness was found to be completely corroded at
certain locations during the test duration, and
the corrosion had penetrated into the substrate
T22 steel (Ref 79). Thus, the authors concluded
that chromizing layer can only offer temporary
protection to the substrate steel (Ref 79).
A long-term plant exposure program to
determine suitable superheater and reheater
alloys against coal-ash corrosion for applications
in ultra-supercritical boilers was conducted by
Babcock & Wilcox (B&W), the U.S. Department
of Energy (DOE), and the Ohio Coal Development Ofce (OCDO) under the Coal Ash
Fig. 10.72
Cross section of a superheater tube made of Type

304SS suffering coal-ash corrosion attack. Note
the wastage flats on both sides of the tube surface where the flue
gas impinged at the 90 location (i.e., facing the ruler in the
photo). Courtesy of Welding Services Inc.
pg 301
30/10/2007 4:04PM Plate # 0
Corrosion Resistant Materials Testing Program.

The results of the test program were published
by McDonald (Ref 80) and McDonald and
Robitz (Ref 81). Plant exposure tests were conducted in a B&W 120 MW(e) cyclone-red
boiler (a subcritical unit burning a 3 to 3.5% S
Ohio coal) at Reliant Energys plant in
Niles, Ohio. Test sections, which were cooled by
600 F/315 psi reheat steam, were located within
the superheater bank. Figure 10.86 shows cross
sections of SAVE 25 (Sumitomos 25Cr-20Ni
steel) tube specimens tested at three different
temperatures, showing increased wastage rates
with increasing exposure temperature. The
results summarizing the performance of various
alloys are presented in Fig. 10.87. The alloys that
exhibit acceptable wastage rates were found to
be alloy 671 cladding and alloy 72 weld overlay.
Both alloys 671 and 72 were nickel alloys with
very high chromium contents (47% Cr for former
and 44% Cr for the latter). They were both used as
a cladding. Alloy 671 clad tubing was produced
by coextrusion method, while alloy 72 weld
Fig. 10.73
overlay tube was produced by spiral overlay

welding method using gas metal arc welding
(GMAW) process.
The Central Electricity Generating Board
(CEGB) in the United Kingdom tested and tried
co-extruded tubing with a high chromium alloy
as a cladding for superheaters and reheaters in
1970s and 1980s. Latham et al. (Ref 82) reported
test trials of Type 310/Esshete 1250 co-extruded
tubes and 671/800H co-extruded tubes in CEGB
boilers. One test trial was conducted in a
550 MW(e) boiler burning coal with 0.45% Cl.
The test tubes were welded to the hottest section
of the reheater with maximum tube metal temperature of approximately 650 C (1200 F), and
ue gas temperature of about 1150 C (2100 F).
After 15,400 h of exposure, 671/800H coextruded tubes along with Type 316 and
347 tubes were removed for evaluation. In
addition, Type 310/E1250 and 671/800H coextruded tubes were removed after 18,000 h of
exposure, and 671/800H co-extruded tubes were
removed after 34,000 h of exposure. The results
Optical micrograph showing the morphology of the corrosion attack on the 304H superheater tube at the severely wasted
area. Courtesy of Welding Services Inc.
pg 302
30/10/2007 4:04PM Plate # 0
Chapter 10:
in terms of wastage rates (nm/h or mpy) are

summarized in Table 10.10.
In another test conducted in a 275 MW(e)
boiler burning coal with less than 0.15% Cl,
a complete nal reheater was retubed with
pg 303
Type 310/E1250 co-extruded tubes. After 17,500

and 29,000 h of exposure, Type 310/E1250
co-extruded tubes showed a wastage rate of only
6 nm/h (1.9 mpy) versus a wastage rate of about
30 m
11 m
Fig. 10.76
Fig. 10.74
Scanning electron micrograph showing the corrosion products (general corrosion scales and
internal corrosion phases) on Type 304H superheater at the
severely wasted area. The EDX analysis showed that the corrosion
product (area 1) was Cr-Fe-rich sulfide (45.5Cr-33.5Fe-9.2Ni8.6S), and the internal corrosion phase (area 2) was Fe-Cr-rich
oxide (56.8Fe-27.5Cr-11.7Ni). Area 3 is the alloy matrix. Courtesy
of Welding Services Inc.

electron image) showing fly-ash deposits
(46.4 Si-21.6Al-20.7Fe) (marked 1) on the surface of Type 304H
reheater (Fig. 10.75) that suffered the maximum wastage
at location 30 away from the direct flue gas impingement point.
The 304H also suffered intergranular corrosion attack (areas 24).
50 m
Fig. 10.75
Cross section of a Type 304H reheater tube

suffering fly-ash erosion/corrosion damage.
Fig. 10.77
Scanning electron micrograph showing the

oxide scales formed on the nearby location of
the one shown in Fig. 10.76 on the severely wasted area for Type
304H reheater tube. Courtesy of Welding Services Inc.
30/10/2007 4:04PM Plate # 0
pg 304
50 nm/h (16 mpy) for previous Type 347 tubes.

A nal superheater was retubed with Type 310/
E1250 co-extruded tubes for a 500 MW(e) boiler
burning a 0.3% Cl coal, Type 310/E1250 coextruded tubes showed maximum corrosion rate
of about 50 to 60 nm/h (16 to 19 mpy) versus
maximum corrosion rates of about 100 nm/h
(32 mpy) for the original Type 316 tubes.
Experiences in the United Kingdom with coextruded tubes for superheaters and reheaters
were summarized in a 1987 paper by Flatley et al.
(Ref 83). Flatley et al. (Ref 83) indicated that
boilers in several major power stations retubed
the hottest sections of superheaters and reheaters
with Type 310/E1250 co-extruded tubes with
excellent results. In one installation, however,
some of the hotter leading tubes were less satisfactory in performance advantage of the 310/
E1250 co-extruded tubes over the original Type
316 tubes with wastage rates being about 50 nm/
h (16 mpy) for Type 310 cladding versus about
70 to 100 nm/h (22 to 32 mpy) for the original
Type 316 tubes. For alloy 671/800H co-extruded
tubes installed in a complete reheater in a
500 MW(e) boiler burning high chlorine (0.5%)
coal, the alloy 671 cladding showed a wastage
rate of less than 0.1 mm/yr (4 mpy) after
60,000 h of operation as opposed to a wastage
11 m
Fig. 10.79
Scanning electron micrograph showing only the

top portion of the ash deposits/corrosion scale
shown in Fig. 10.78 at higher magnification. The gray and white
phases were analyzed by EDX with the results (in wt%) summarized below. Courtesy of Welding Services Inc.
1: 30.2% Fe, 21.0% As, 14.1% Si, 11.0% Al, 8.7% Ca, 9.7% S,
2.3% P, and residual elements
2: 80.7% Fe, 11.1% Si, 2.1% Al, 1.5% As, and residual elements
11 m
30 m
Fig. 10.78
Scanning electron micrograph showing ash

deposits that formed at the 12 oclock position
where the flue gas made a direct impingement on the Type 304H
reheater tube (Fig. 10.75). This location showed significantly less
tube wall wastage. Courtesy of Welding Services Inc.
Fig. 10.80
Scanning electron micrograph showing the middle section of the ash deposits/corrosion scale
shown in Fig. 10.78 at higher magnification. The dark gray (3) and
light gray (4) phases were analyzed by EDX with the results (in wt
%) summarized below. Courtesy of Welding Services Inc.
3: 32.8% Si, 26.1% As, 22.6% Fe, 9.6% Al, 2.4% P, 1.5% S, and
residual elements
4: 93.8% Fe, 2.4% Cr, and residual elements
30/10/2007 4:04PM Plate # 0
Chapter 10:
rate of about 2 mm/yr (79 mpy) for the original

Type 347 tubes.
Some of the performance experiences for this
671/800H co-extruded tubing have been reported
by Fahrmann and Smith (Ref 84) and Kiser and
Orsini (Ref 85). Fahrmann and Smith (Ref 84)
reported the results of the metallurgical evaluation of two superheater tubes and one reheater
tube after 18 years of service in a utility coal-red
boiler. The alloy 671 cladding in the reheater tube
was found to exhibit very little corrosion attack,
with the cladding thickness very close to that of
the original as-fabricated tube. For two superheater tubes, the alloy 671 cladding generally
showed good conditions except at few locations
where pitting attack was observed. In one
case, the pitting attack was almost penetrated
to the entire cladding thickness. Even at this
location, the corrosion rate would still be very
low (less than 0.1 mm/yr, or 4 mpy, assuming
the original cladding thickness being 2 mm, or
80 mils).
In late 1990s, a spiral overlay welding technology involving a tandem gas metal arc and
pg 305
gas tungsten arc welding method was developed

for manufacture of weld overlay bimetallic tubing for superheaters and reheaters (Ref 86).
Two spiral weld overlay tubes tested in the Coal
Fig. 10.82
Results of laboratory tests conducted in synthetic

flue gas (80N2-15CO2-4O2-1SO2, saturated with
H2O) with synthetic ash (37.5 mol% Na2SO4, 37.5 mol% K2SO4,
and 25 mol% Fe2O3) covering the samples. Exposure time was
11 m
Fig. 10.81
Scanning electron micrograph showing the bottom portion of the corrosion products shown in
Fig. 10.78 at higher magnification. The phases (58) were analyzed by EDX with the results (in wt%) summarized below.
5: 47.2% Cr, 35.9% Fe, 13.0% S, 2.5% Mn, and residual elements
6: 54.2% Cr, 31.3% Fe, 5.7% S, 3.7% Mn, 2.8% Ni, and residual
elements
7: 45.9% Fe, 27.5% Ni, 15.9% Cr, 6.4% S, 3.7% Si, and residual
elements
8: 50.9% Fe, 31.9% Ni, 10.3% Cr, 5.4% S, and residual elements
Fig. 10.83
Metal loss as a function of chromium contents in

the alloys generated by various investigators in
laboratory coal-ash corrosion tests as well as plant exposure.
Source: Ref 73. Note: This work by Castello et al. in Lab tests:
10% alkali, 1% SO2, 700 C (Ref 73); Plumley et al. (Ref 74): Plant
exposure, 677727 C, 1.7% S in coal, 6% alkali in ash; Kihara
et al. (Ref 75): Lab test, 7.3% alkali, 0.3% SO2, 700 C; Van Weele
et al. (Ref 76): Lab test, 10% alkali, 1% SO2, 10% H2O, 700 C;
Blough (Ref 77): Plant exposure, 538871 C, 0.63.6% S, 0.4
0.9% Na in coal.
30/10/2007 4:04PM Plate # 0
Ash Corrosion Resistant Materials Testing Program (Ref 80, 81) were identied as IN72WO
and IN52WO in Fig. 10.87. For the sample tube
Fig. 10.84
Metal loss as a function of chromium contents in

the alloys in laboratory coal-ash corrosion tests
(solid line) and plant exposure using corrosion probes inserted
into the operating boiler at Tennessee Valley Authoritys Gallatin
Station Unit No. 2. The vertical axis on the right-hand side is in
mils. Source: Ref 78
Fig. 10.85
identied as IN72WO, the weld overlay was produced using alloy 72 weld wire (AWS ERNiCr4: Ni-44Cr), while the alloy 52 weld overlay for
IN52WO was produced using alloy 52 weld
wire (AWS ERNiCrFe-7: Ni-30Cr-9Fe). In
Fig. 10.87, alloy 72 overlay was found to perform similarly to alloy 671 cladding. Alloy 52
overlay, although not performing as well as alloy
72 overlay and alloy 671 cladding, performed
signicantly better than many austenitic alloys
particularly at higher temperatures.
In one subcritical boiler (255 MWe) burning
high-chlorine coal (about 0.3% Cl), Type 304H
reheater tubes typically lasted for about 4 years.
Test trials were performed for alloy 72/304H and
alloy 52/304H weld overlay tubes as part of the
reheater (538 C/1000 F outlet steam) for 3
years. Typical tube cross sections are shown in
Fig. 10.88 for Alloy 72/304H tube and Fig. 10.89
for alloy 52/304H tube. Very little tube thinning
was observed for alloy 72 overlay. Slight corrosion was observed for alloy 52 overlay. The
maximum wastage rate was estimated to be less
than 0.05 mm/yr (2 mpy) for alloy 72 and about
0.2 mm/yr (8 mpy) for alloy 52. Kiser et al. (Ref
87) indicated that, in one boiler with low NOx
burners, alloy 72 overlay reheater tubes were
found to show negligible wastage of about
0.076 mm (3 mils) after service for about 6 years
at the metal temperature of about 649 to 677 C
(1200 to 1250 F). The authors (Ref 87) also
Metal wastage rates as a function of metal temperature for 304HSS, 347SS, alloy 800H, alloy NF709 (20Cr-25Ni1.5Mo-0.2Nb-Fe), alloy HR3C (25Cr-20Ni-0.5Nb-0.25N-Fe), alloy CR30A (30Cr-48Ni-2Mo-0.25Al-0.25TiFe), and chromized T22 (T22Cr). The data were generated from plant exposure tests at TVAs Gallatin Station Unit No. 2 boiler.
Source: Ref 79
pg 306
30/10/2007 4:04PM Plate # 0
Chapter 10:
indicated that, in another boiler, alloy 72 overlay

superheater tubes showed essentially no sign of
corrosion or cracking after 6 years of service in
sootblower lanes (Fig. 10.90).
10.6 Erosion in Fluidized-Bed Boilers

In xed-bed, uidized-bed boilers, the major
high-temperature materials issues are essentially
corrosion and erosion of in-bed heat-exchanger
components. Corrosion is mainly suldation/
oxidation. In general, the 300 series austenitic
pg 307
stainless steels have adequate resistance, since

the metal temperatures for these heat exchangers
normally do not exceed 650 C (1200 F)
(Ref 8890). Materials wastage due to erosion is
a major issue for the in-bed heat exchangers.
Stringer (Ref 88, 89, 91), Stallings and Stringer
(Ref 92), and Rademakers et al. (Ref 90) provided excellent reviews on erosion related to inbed heat exchangers.
The wastage problem for the in-bed heat
exchanger appears to be a low-temperature phenomenon. Rademakers et al. (Ref 90) showed
that the tube wastage rates were high at low
temperatures and decreased with increasing
temperature, as shown in Fig. 10.91. Stringer
Table 10.10 Results of field tests on austenitic

stainless steel tubes and co-extruded tubes in the
reheater of a 500 MW (e) boiler
Wastage rate
Material
Fig. 10.86
Tube cross-section wastage profile for alloy Save

25 (Sumitomos 25Cr-20Ni steel, similar to ASME
SA213TP310H or SA213TP310HCbN) after exposure to the
temperature for about 15.5 months. Wastage rates in each case
are also indicated. Source: Ref 81
Fig. 10.87
316
347
310/E 1250
671/800H
Exposure, h
15,400
15,400
18,000
34,000
nm/h
mpy
160
200
60
51
64
19
Negligable
Source: Ref 82
Tube metal wastage rates as a function of surface metal temperature for various alloys tested at Reliant Energys Niles plant
under the Coal Ash Corrosion Resistant Materials Testing Program conducted jointly by B&W, DOE, and the Ohio Coal
Development Office. Source: Ref 80
30/10/2007 4:04PM Plate # 0
(Ref 89) indicated that this tube wastage does not

appear to be a problem above approximately
500 C (932 F). Wastage has taken place on
the in-bed evaporator tubes, but seldom occurred
on the in-bed superheater tubes (Ref 92). This

may suggest a temperature effect.
The wastage damage is commonly referred to
be that of erosion. Stringer and Wright (Ref 88),
Stringer (Ref 89), and Stallings and Stringer (Ref
92) considered that mechanical processes, such
as erosion and abrasion wear, are involved in the
damage mechanisms. Stringer (Ref 91) denes
erosion and abrasion: erosion is material removal
by the impact of particles moving freely before
and after the impact; abrasion wear is material
removal by particles that are loaded onto the
surface and remain in contact for a period of time.
Possible wastage mechanisms are discussed in
Fig. 10.88
The cross section of alloy 72/304H overlay tube

after testing as part of the reheater (1000 F
steam) for 3 years in a 255 MW(e) subcritical boiler,
which burned high chlorine coal (about 0.3%). The 304H
reheater tubes exhibited a typical 4 year life. Courtesy of Welding
Services Inc.
Fig. 10.89
The cross section of alloy 52/304H overlay tube

after testing as part of the reheater (538 C, or
1000 F steam) for 3 years in a 255 MW(e) subcritical boiler,
which burned high chlorine coal (about 0.3%), showing slight
pitting attack. The 304H reheater tubes exhibited a typical 4 year
life. Courtesy of Welding Services Inc.
Fig. 10.90
Alloy 72 overlay superheater tubes in a sootblower lane after 6 years of service. Original
weld bead ripples are still clearly visible. Source: Ref 87
pg 308
30/10/2007 4:05PM Plate # 0
Chapter 10:
detail by these authors in Ref 88, 89, 92, and 93.

Stringer and Wright (Ref 88) briey summarize
possible wastage mechanisms: (a) erosion by
in-bed moving particles, (b) abrasion wear by
loaded particles, (c) erosion by particles in the
wake of bubbles, (d) bubbles track along vertical
or sloping tubes, (e) erosion by particles thrown
into the metal surface by bubble collapse, (f)
erosion in the splash zone by bubble collapse at
the bed surface, (g) erosion induced by in-bed
jets, associated with coal or acceptor injection
ports, particle recirculation ports, and so forth,
and (h) erosion by gulf stream (long-range
ow patterns) in the bed.
Materials wastage issues in uidized-bed
boilers have been extensively reviewed in Ref 88
to 93. This section discusses the major erosion
and/or abrasion problem (a) for in-bed evaporator tubes in bubbling uidized-bed boilers
and (b) at the refractory/waterwall interface
region in circulating uidized-bed boilers.
10.6.1 In-Bed Evaporator Tubes in
Bubbling Fluidized-Bed Boilers
The evaporator tubes in a bubbling uidizedbed boiler (TVA 20 MWe boiler) experienced
severe wastage problems with carbon steel tubes
during operation in 1986. The average wastage
rates for the carbon steel (SA210 A1) tubes were
about 13 mils/1000 h with some tubes as high as
35 mils/1000 h (Ref 94). A test program was
subsequently conducted to evaluate several
cladding and coating materials that included
Fig. 10.91
Type 304/SA210 A1 co-extruded tubes, two

proprietary thermal sprayed coatings, sprayed
and fused WC-based coated (Extendalloy coating) tubes, and chromized T11 tubes. The
thickness of Extendalloy coating was about 25
mils. The in-bed evaporator tubes were operated
at approximately 343 to 371 C (650 to 700 F).
Test tubes were removed for metallurgical evaluation after exposure of about 6500 h. The
ndings, as reported by Lewis et al. (Ref 94),
indicated that the maximum wastage rate was
8 mils/1000 h (70 mpy) for carbon steel tubes
(SA210 A1 control samples) and 95 mils/1000 h
(83 mpy) for Type 304 cladding (co-extruded
tubes). Extendalloy, chromized, and thermal
sprayed coatings, however, showed no measurable wastage. Nevertheless, the authors observed
that oxide phases had formed at the coating/
substrate interface for the thermal sprayed coatings and concluded that this could lead to coating
spallation (Ref 94). As for a chromized coating, it
is questionable whether such a thin coating
(about 400 m thick) could sustain a long-term
performance under uidized-bed erosive conditions.
In examining evaporator tube wastage
experienced by the TVA 20-MW(e) pilot plant
(atmospheric bubbling uidized-bed boiler)
ring Pyro coal during operation in 1986,
Stallings and Stringer (Ref 92) concluded that the
tube wastage occurred mainly at the bottom of
the tube. This was because the larger, harder,
more angular bed materials tend to migrate
toward the bottom region of the bed where they
Tube wall wastage rates as a function of tube wall temperature from the in-bed tube tests in a 4 MW atmospheric fluidizedbed combustor (AFBC). St. 35.8 is a carbon steel. Source: Ref 90
pg 309
30/10/2007 4:05PM Plate # 0
were picked up in the wakes of bubbles and rose

upward. The tube wastage was primarily caused
by the impact and motion of particles entrained in
bubble wakes on the underside surface of the
tubes.
In a 130 MW(e) atmospheric bubbling uidized-bed boiler, the in-bed evaporator was
constructed out of sprayed and fused WC-based
hardfacing coated tubes. These coated tubes were
in general satisfactory in performance. Nevertheless, occasional local failures did occur
according to the plant operator. Repair of the
localized failures could be difcult for these
sprayed and fused hardfacing coating. The plant
operator was interested in weld overlay
hardfacing materials that were weld repairable
when local wear took place. A test tube consisting of test sections of different weld overlay
alloys was tested as part of an evaporator tube.
The steam in the evaporator tube element was
about 336 C (637 F). The bed temperature was
typically 788 to 870 C (1450 to 1600 F). Test
duration was 14,021 operating hours. The overlay alloys tested included Type 309, Type 312,
alloy 82, alloy 625, a proprietary WC-based
hardfacing alloy (HF60), and a thermal sprayed
coating on alloy 625 overlay (butter layer).
Evaluation of test specimens after 14,021 operating hours showed that thermal sprayed coating
was gone (worn away). Figure 10.92 shows
spallation of the coating. Type 309, 312, alloy
0.1 mm
Fig. 10.92
Spallation of the thermal sprayed coating on the

bottom surface of the test tube sample after
exposure for 14,021 h as part of the in-bed evaporator tubes in a
130 MW(e) bubbling fluidized-bed boiler. The coating was
applied to alloy 625 overlay, which acted as a butter layer
between the coating and the substrate steel tube. Courtesy of
82, and alloy 625 overlays were completely worn

away at the underside of the tube samples. The
WC-base hardfacing overlay (HF60) showed no
sign of erosion or abrasive wear. The cross sections of all the test specimens are shown in
Fig. 10.93 (Ref 14). The wastage was found to be
the greatest at the bottom of the tube surface
(6 oclock position). Traditional stainless steels
and nickel alloys are not adequate in providing
erosion or abrasive wear protection. A proprietary WC-based hardfacing weld overlay showed
no evidence of wastage. The overlay surface
showed a polished appearance. The microstructure of this hardfacing weld overlay and
the analyses of the phases are shown in
Fig. 10.94. The hardness of this weld overlay
was found to be in the range of 50 to 60 HRC
(converted from Vickers hardness values) across
the overlay.
A proprietary hardfacing weld overlay
(HF35), based on chromium eutectic carbides,
was developed and tested in the same bubbling
uidized-bed boiler. Testing was conducted on
tubes in the loop section, which was historically a
high wear area. Since it was a loop section of
the tube, the tube was weld overlaid using a
manual technique. Figure 10.95 shows one of the
overlaid tubes in the loop section after exposure
for close to 3 years as part of the in-bed evaporator tube bundle. The manually applied
longitudinal weld beads are still visible. Two
tubes in the loop section were later removed from
the boiler for metallurgical evaluation. The weld
overlay overall remained protective under the
erosive conditions for about 3 years. There was,
however, a localized wear area near the pin
studded carbon steel tube end. Both tubes
showed the similar locations that suffered localized wear.
An in-bed superheater tube bundle located
above this in-bed evaporator bundle reportedly
showed no erosion or abrasion problems. The
superheater was made of an austenitic stainless
steel. This appears to conrm earlier observation
by Rademakers et al. (Ref 90) that the erosion
or abrasion wear of the in-bed tubes were signicantly reduced at higher temperatures and
by Stringer (Ref 89) that the in-bed tube
wastage problem does not appear to be a problem above approximately 500 C (932 F).
The plant engineer indicated that the in-bed
superheater tube bundle was more tightly arranged than the in-bed evaporator tubes
(Ref 14). It is not clear whether this could be a
factor also.
pg 310
30/10/2007 4:05PM Plate # 0
Chapter 10:
10.6.2 Waterwalls in Circulating

Fluidized-Bed Boilers
One serious materials issue for circulating
uidized-bed boilers is the erosion or abrasion
wear that occurs on the waterwall at the refractory lining-waterwall transition region (Ref 90,
9597). It is generally believed that in a circulating uidized-bed boiler the bed particles ow
upward in the center of the boiler and migrate
downward along the waterwall (Ref 98, 99). This
downward ow of particles is believed to be
responsible for the erosion wear above the
refractory lining. The authors suggested that the
(a)
(b)
(c)
(d)
Fig. 10.93
Cross sections of test tube samples showing wear profiles for (a) Type 309 overlay tube, (b) alloy 625 overlay tube, (c) Type
312 overlay tube, and (d) HF60 hardfacing overlay tube after exposure for 14,021 h as part of the in-bed evaporator tubes
in a 130 MWe bubbling fluidized-bed boiler. Courtesy of Welding Services Inc.
pg 311
30/10/2007 4:05PM Plate # 0
particles owing downward along the waterwall

were forced to change ow direction at a tapered
(or sloped) refractory lining forcing the particles
to ow from the membrane (or web) outward
against the tube walls on the tapered refractory
19 m
Fig. 10.94

electron image) showing various hardface particles in the proprietary tungsten carbide based hardfacing weld
overlay, HF60. The results (wt%) of semiquantative EDX analyses
of various phases are summarized as: Light color phases (A, B, C,
E, and F) are tungsten carbides (6070W, 1416Ni, 8Cr, 5Si),
grayish phase (H) is also tungsten carbide ( 36W, 39Ni, 18Cr,
2.5Si), and dark phases (I and G) are matrix (7080Ni, 10Cr,
510W). Courtesy of Welding Services Inc.
Fig. 10.95
surface. Slusser et al. (Ref 95) reported the

waterwall erosion wear problem at the transition
region in a 50 MW(e) boiler. Figure 10.96
illustrates schematically the relationship between
the waterwall and the refractory lining at the
transition region (Ref 95). The gure also indicates that the potentially high wear area was
protected by a weld overlay. No technical information about the weld overlay alloy used was
discussed in the paper. However, the weld overlay was reported to have suffered severe metal
loss (Ref 95).
Some of the protective methods suggested by
the authors (Ref 9597), although somewhat
helpful, may not be long-term solutions. For
example, installing a shelf above the refractory
lining by interrupting the downward particle ow
was found to be helpful in reducing wastage rates
at the waterwall-refractory transition region, but
induced wear at the shelf location (Ref 95).
Because of the interruption of the particle ow
direction by the tapered refractory surface, particles are then forced to ow against the tube
walls on the tapered refractory surface, thus
resulting in high waterwall tube wear at the
transition region. This is illustrated schematically
in Fig. 10.97. Nevertheless, Slusser et al. (Ref 96)
reported in their 1992 paper indicating that
shelves (7 in. wide sheets) installed perpendicular to the waterwalls around the entire
perimeter of the boiler at 2.2 m (7 ft) and 5.2 m
(17 ft) above the refractory transition area were
capable of reducing wastage rates from >25 mm/
yr (>1000 mpy) to a manageable rate, thus
allowing reasonable operating periods between
One tube with a proprietary HF35 hardfacing weld overlay showing manually applied weld overlay beads along the tube
that are still visible after exposure for approximately 3 years as part of an in-bed evaporator tube bundle in a 130 MWe
bubbling fluidized-bed boiler. Courtesy of Welding Services Inc.
pg 312
30/10/2007 4:05PM Plate # 0
Chapter 10:
inspections. Wastage rates at the shelves were

found to be accelerated. For example, the
waterwalls at the 17 ft shelf location suffered
wastage rates of about 0.76 to 1.78 mm/yr (30 to
70 mpy) (Ref 96). Some other protection methods including cast erosion blocks made of Type
310 and thermal sprayed coatings of various
alloys were mentioned under trials in a 1997
paper by Solomon (Ref 97). Some coatings
such as thermal sprayed coatings of 50Ni-50Cr
alloy and Cr2O3-base coating were found to be
unacceptable in performance (Ref 97). Some
success was attained in an approach involving
bending the waterwall tubes out of the vertical
plane and into the wall at the interface with the
refractory lining making a smooth, vertical
transition from the waterwall tubes to the
refractory (Ref 93).
10.6.3 Erosion in the Convection Pass
Tubes in Fluidized-Bed Boilers
In uidized-bed boilers, ue gas stream
leaving the combustor and enters a cyclone,
where particles are removed, before exiting to
the convection pass. Typically, a superheater
and an economizer are in the convection pass
in the downstream of the cyclone. The ue
gas stream is entrained with ne particles that
are not removed by the cyclone. High tube
metal wastages could occur at superheater
tube banks and economizer tube banks
(Ref 93). Erosion by y ash can be serious
when the uneven distribution of y ash in the
Fig. 10.96
Waterwall in the transition region that suffers

severe erosion wear immediately above the
tapered refractory surface. Source: Ref 95
ue gas stream occurs under certain conditions.

For example, excessive y-ash erosion of the
economizer can occur in the backpass area
when the ue gas stream making the direction change and creating uneven distribution
of y-ash particles due to centrifugal force
(Ref 1).
An example is given below on erosion or
erosion/corrosion of the hanger tubes in the
backpass area for the economizer of a circulating
uidized-bed boiler (90 MWe) burning Eastern
bituminous coal (Ref 14). The current hanger
tubes are protected with tube shields, which
typically last for a year. A Type 309 overlay tube
and an HF35 hardfacing overlay tube were tested
as part of the hanger tubes at top of the backpass.
The temperature of the steam inside the tubes was
reportedly about 315 C (600 F). The tubes
were removed for metallurgical evaluation after
2 years of exposure. For both tubes, the leeward
side typically exhibited some thin scales,
while the windward side showed no scales.
Figure 10.98 shows typical cross sections of
these tube samples after 2 years of exposure. The
Type 309 overlay tube showed more thinning
on the overlay on the windward side of the tube
(top side of the cross section in the photograph).
This can be seen in tube cross sections as
shown in Fig. 10.98. The metallographic cross
sections showing the overlay in comparing
the windward side with the leeward side of the
tube for both 309 and HF35 overlay tubes, as
shown in Fig. 10.99 and 10.100. The surface
condition as well as the microstructure of the
weld overlay at the windward side for both
overlays is shown in Figs. 10.101 and 10.102.
The 309 overlay exhibits austenitic structure with
some delta ferrite phases, with a hardness range
of about 91 to 92 HRB (converted from Vickers
hardness values). The HF35 overlay exhibits
Fig. 10.97
Particles flowing outward against the waterwall tube wall from the membrane due to interruption of the downward flow of particles on a shelf. Source:
Ref 96
pg 313
30/10/2007 4:05PM Plate # 0
pg 314
chromium eutectic carbides formed along interdendritic boundaries, with a hardness range of
about 35 to 40 HRC (converted from Vickers
hardness values). The wastage rate was calculated to be about 0.38 mm/yr (15 mpy) for
Type 309 overlay and 0.2 mm/yr (8 mpy) for
HF35 hardfacing overlay. Both overlays contained similar chromium contents, about 21% Cr
for the 309 overlay and about 20% Cr for the

HF35 overlay, and both should form chromium
oxide scales. The reduced wastage rate exhibited
by the HF35 overlay over the 309 overlay is
believed to be due to its erosion resistance
resulting from hardfacing particles of chromium
eutectic carbides.
10.7 Summary
A general description of coal-red boilers and
their combustion conditions is presented. Fireside corrosion of the furnace waterwalls is discussed with particular emphasis on the materials
problems associated with a staged ring by substoichiometric combustion (i.e., combustion with
insufcient oxygen) in the lower furnace for
reducing NOx emissions. As a result of the staged
Fig. 10.98
Cross sections of Type 309 overlay tube (a) and

HF35 overlay tube (b) showing the wastage
profile of the weld overlay after exposure of 2 years as part of the
hanger tube for the economizer in a circulating fluidized-bed
boiler. The top of the tube cross section was the windward side
while the bottom (i.e., the ruler side) was the leeward side.
(a)
0.010 in.
(b)
0.010 in.
Fig. 10.99
Type 309 overlay tube in the unaffected leeward

side (a) and the windward side (b) after testing as
part of hanger tubes for the economizer in a circulating fluidizedbed boiler for 2 years. The substrate steel was etched with nital to
reveal the fusion boundary. Courtesy of Welding Services Inc.
30/10/2007 4:05PM Plate # 0
Chapter 10:
ring, suldation dominates the corrosion process for waterwall materials. The wastage rates
for carbon and low alloy steels have been found
to increase to an unacceptable rate for many
boilers, particularly those of supercritical units.
The current most widely used waterwall protection method against tube wall wastage problems involves weld overlay cladding in the
boiler on the affected waterwall area with a corrosion-resistant alloy using automatic gas-metalarc welding (GMAW) process. Cladding of
waterwall panels can also be applied in shop
using either GMAW or laser cladding techniques,
and the cladded panels are then installed in
the boiler. The use of a corrosion-resistant
cladding has essentially eliminated the tube
wastage problems. For some supercritical
units, circumferential grooving and cracking has
been encountered on overlaid waterwall tubes
pg 315
presumably due to overheating and preferential

suldation penetration. The possible causes to
circumferential grooving and cracking are discussed.
Also, included in the discussion are erosioncorrosion caused by steam sootblowing and
thermal fatigue cracking due to the use of water
lances or water canons for deslagging. Superheater/reheater wastage problems as well as
erosion-corrosion issues in convection pass tubes
are discussed. Erosion, erosion-corrosion, or abrasion wear issues encountered in uidized-bed
boilers including both bubbling beds and circulating beds are also discussed.
0.0010 in.
Fig. 10.101
(a)
0.010 in.
(b)
0.010 in.
Fig. 10.100
HF35 overlay tube in the unaffected leeward

side (a) and the windward side (b) after testing
as part of the hanger tube for the economizer in a circulating
fluidized-bed boiler for 2 years. The substrate steel was etched
with nital to reveal the fusion boundary. Courtesy of Welding
Services Inc.
The surface of Type 309 overlay tube at the

windward side after exposure for 2 years as part
of the hanger tube for the economizer in a circulating fluidizedbed boiler. Microstructure of Type 309 overlay consists of delta
ferrite (dark phases) in austenite. Courtesy of Welding Services
Inc.
0.0010 in.
Fig. 10.102
The surface of HF35 weld overlay tube at the

windward side after exposure for 2 years as part
of the hanger tube for the economizer in a circulating fluidizedbed boiler. The overlay contained hardfacing chromium eutectic
carbides (dark phases) formed along interdendritic boundaries.
30/10/2007 4:05PM Plate # 0
REFERENCES
1. S.C. Stultz and J.B. Kitto, Ed., Steam and Its

Generation and Use, 40th ed., Babcock &
Wilcox, Barberton, OH, 1992
4th ed., Combustion Engineering, Inc.,
Windsor, CT, 1991
3. C.A. Meyer, et al., ASME Steam Tables,
5th ed., The American Society of Mechanical Engineers, 1983
4. R. Peltier and K. Wicker, PRB Coal Makes
the Grade, Power, Oct 2003, p 28
5. J. Stringer, High Temperature Corrosion
Problems in the Electric Power Industry
and Their Solution, in High Temperature
Corrosion, R.A. Rapp, Ed., Proc. Conf.
(San Diego, CA, March 26, 1981, NACE,
1981
6. D.J. Lees and M.E. Whitehead, Microanalysis of Scales and Deposits Formed on
Corroding Furnace Tubes in Coal-Fired
Boilers, in Corrosion Resistant Materials for
Coal Conversion Systems, D.B. Meadowcroft and M.I. Manning, Ed., Applied Science, London, 1982, p 63
7. K.A. Hay, High Temperature Environmental
Degradation of Materials and Future
Requirements in The Power Generation
Industry, in Environmental Degradation of
High Temperature Materials, Vol 2 (No. 3),
Series 3, 1980, p 4/1
8. G.T. Cunningham and T. Webster, Design,
Manufacturing, and Testing of a Boiler
Metal Surface Temperature Probe, Paper
No. IJPCG 2000-15075, presented at the
2000 International Joint Power Generation
Conference (Miami Beach, FL), July 2326,
2000
9. J. Stringer, Materials for Fossil Energy
SystemsPast, Present, and Future, Paper
10. J.A. Urich and E. Kramer, Designing Solutions for Low NOx-Related Water Wall
Corrosion, Conf. Proc., 1996 ASME
Joint Power Generation Conference,
S.M. Smouse and A. Gupta, Ed., ASME,
1996, p 25
11. M.D. Yeager, The Effect of Low NOx Staging on Waterwall Wastage, presented at
the EPRI International Conference on Boiler
Tube Failures in Fossil Plants (Nashville,
TN), Nov 1113, 1997
12. A.J.B. Cutler and E. Raask, External Corrosion in Coal-Fired Boilers: Assessment
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
from Laboratory Data, Corros. Sci., Vol 21

(No. 12), 1981, p 789
A.J.B. Cutler, T. Flatley, and K.A. Hay,
Metall. Mater. Technol., Feb 1981, p 69
Welding Services Inc., Norcross, GA
D.N. French, Metallurgical Failures in
Fossil Fired Boilers, John Wiley & Sons,
1993
D.N. French, Paper No. 133, Corrosion/88,
NACE, 1988
T. Flatley, E.P. Latham, and C.W. Morris, in
Inst. Metallurgists Spring Residential Conference: Environmental Degradation of
High Temperature Materials, Vol 2 (No. 13),
Series 3, 1980
F. Clarke and C.W. Morris, Combustion
Aspects of Furnace Wall Corrosion, in
Corrosion Resistant Materials for Coal
Conversion Systems, D.B. Meadowcroft and
M.I. Manning, Ed., Applied Science Publishers, London, 1982, p 47
P.L. Daniel, L.D. Paul, J.M. Tanzosh, and
R. Hubinger, Estimating Effects of Chlorine
on Fireside Corrosion of Furnace Walls in
Coal-Fired Boilers, Paper No. 138, Corrosion/88, NACE, 1988
E. Latham, D.B. Meadowcroft, and L. Pinder, The Effects of Coal Chlorine on Fireside
Corrosion, Chlorine in Coal, J. Stringer and
D.D. Barnerjee, Ed., Elsevier Science Publishers B.V., Amsterdam, 1991, p 225
S. Brooks and D.B. Meadowcroft, The Role
of Chlorine in the Corrosion of Mild and
Low Alloy Steels in Low pO2 and High pS2
Environments, in High Temperature Corrosion in Energy Systems, M.F. Rothman, Ed.,
p 515
S. Brooks and D.B. Meadowcroft, The
Inuence of Chlorine on the Corrosion of
Mild and Low Alloy Steels in Sub-stoichiometric Combustion Gases, in Corrosion
Resistant Materials for Coal Conversion
Systems, D.B. Meadowcroft and M.I. Manning, Ed., Applied Science, London, 1982,
p 105
T. Flatley, E.P. Latham, and C.W. Morris,
Co-Extruded Tubes Improve Resistance to
Fuel-Ash Corrosion in U.K. Utility Boilers,
P.J. James and L.W. Pinder, The Impact of
Coal Chlorine on the Fireside Corrosion
Behavior of Boiler Tubing: A UK Perspective, Corrosion/97, NACE International,
1997
pg 316
30/10/2007 4:05PM Plate # 0
Chapter 10:
25. D.J. Lees, A Summary of Observations

Relating Furnace Wall Fireside Corrosion to
Chlorine Content of Coal, SSD/MID/M26/
79, 1979
26. S.F. Chou and P.L. Daniel, in High Temperature Corrosion in Energy Systems, M.F.
Rothman, Ed., The Metallurgical Society of
AIME, 1985
27. K.S. Gilroy, Laboratory Evaluation of Candidate Materials for Furnace Wall Applications, in High Temperature Corrosion in
Energy Systems, M.F. Rothman, Ed.,
p 345
28. S.C. Kung and C.F. Eckhart, Corrosion of
Iron-Base Alloys in Reducing Combustion
Gases, Paper No. 242, Corrosion/93,
NACE, 1993
29. B. Dooley, R. Tilley, T.P. Sherlock, and
C.H. Wells, State of Knowledge Assessment for Waterwall Wastage, presented at
EPRI International Conference on Boiler
Tube Failures in Fossil Plants (Nashville,
TN), Nov 1113, 1997
30. Workshop on Materials Issues Associated
with Low-NOx Combustion in Fossil-Fired
Boilers, Summary of Workshop held during
Advanced Research and Technology
Developments Tenth Annual Conference
on Fossil Energy Materials (Knoxville, TN),
May 1416, 1996, Materials & Components
in Fossil Energy Applications, Department
of Energy and Electric Power Research
Institute, No. 123, August 1, 1996
31. C. Jones, Maladies of Low-NOx Firing
Come Home to Roost, Power, Jan/Feb,
1997, p 54
32. P.J. James and L.W. Pinder, Furnace Wall
Fireside Corrosion: Taming the Beast
Within, presented at EPRI International
Conference on Boiler Tube Failures in
Fossil Plants (Nashville, TN), Nov 1113,
1997
33. W.T. Bakker and S.C. Kung, Waterwall
Corrosion in Coal-Fired BoilersA New
Culprit: FeS, Paper No. 246, Corrosion/
34. W. Bakker, Root Causes of Accelerated
Waterwall Wastage in Coal-Fired Boilers,
presented at EPRI International Conference
on Materials and Corrosion Experience for
Fossil Power Plants (Isle of Palms, SC), Nov
1821, 2003
35. W.T. Bakker, J.L. Blough, S.C. Kung,
T.L. Baneld, and P. Cunningham, Long
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
Term Testing of Protective Coatings and

Weld Overlays in a Supercritical Boiler,
Retrotted with Low NOx Burners, Paper
P. Hulsizer, Paper No. 246, Corrosion/91,
NACE, 1991
G. Lai, P. Hulsizer, and R. Lee, Waterwall
Wastage Mitigation for Coal-Fired Boilers
Using Automatic Pulse Spray GMAW
Overlay Technology, presented at 1999
EPRI Fossil Plant Maintenance Conference
(Atlanta, GA), 1999
G. Lai and P. Hulsizer, Corrosion & Erosion/
Corrosion Protection by Modern Weld
Overlays in Low NOx Coal-Fired Boilers,
International, 2000
G.Y. Lai, Performance of Automatic
GMAW Overlays for Waterwall Protection
in Coal-Fired Boilers, presented at EPRI
fth International Conference on Welding
and Repair Technology for Power Plants
(Point Clear, AL), June 2628, 2002
G.Y. Lai, Fireside Corrosion and Erosion/
Corrosion Protection in Coal-Fired Boilers,
International, 2004
M.S. Brennan and R.C. Gassmann, Laser
Cladding of Nickel- and Iron-Base Alloys
on Boiler Pressure Panels and Tubes, presented at the EPRI Third International
Conference on Welding and Repair Technology for Power Plants (Scottsdale, AZ),
June 912, 1998
M.S. Brennan and R.C. Gassmann, Laser
Cladding of Nickel and Iron Base Alloys on
Boiler Waterwall Panels and Tubes, Paper
A.J. Bonnington and M.S. Brennan, Type
312 Stainless Steel Laser Cladding for
Waterwalls in Supercritical Units, presented at the EPRI/DOE Conference on
Advances in Life Assessment and Optimization of Fossil Power Plants (Orlando, FL)
March 1113, 2002
Boiler Tube Failure Metallurgical Guide,
Volume 2: Appendices, EPRI TR-102433V2, Electric Power Research Institute, Oct
1993
INCONEL alloy 625, Inco Alloys International Huntington, WV
HASTELLOY C-22 Alloy, H-2019D,
Haynes International Inc., Kokomo, IN
pg 317
30/10/2007 4:05PM Plate # 0
47. K. Coleman and D. Gandy, Corrosion

Resistant Waterwall Overlays Selection of
Alternative Materials, presented at International Conference on Boiler Tube Failures
and HRSG Tube Failures, and Inspections
(Phoenix, AZ), Nov 68, 2001
48. G. Lai, The Microstructure, Properties and
Corrosion Resistance of Type 312 Overlay
in Batch Digesters, Conf. Proc., 2002
TAPPI Fall Technical Conference (San
Diego, CA), Sept 811, 2002
49. G.Y. Lai, presented at WSI Boiler Tube
Overlay Forum (Atlanta, GA) July 1012,
2006
50. J.L. Blough, private communication,
First Energy, Mayeld Village, OH, Sept 29,
2006
51. I.G. Wright, Hot Corrosion in Coal- and OilFired Boilers, Metals Handbook, Vol 13, 9th
ed., Corrosion, ASM International, 1987, p
995
52. S. French, K. Rumbaugh, and P.N. Hulsizer,
Fireside Corrosion-Erosion Mitigation via
the Application of Weld Metal Overlay,
Proceedings: Welding and Repair Technology for Power Plants, EPRI, Charlotte, NC
1997, p 477
53. A.J. Bonnington and T.M. Cullen, Performance of Chromized Waterwall Panels in
Supercritical Units, presented at EPRI
International Conference on Boiler Tube
Failures in Fossil Plants, (Nashville, TN),
Nov 1113, 1997
54. K. Luer, J. DuPont, A. Marder, and C.
Skelonis, Corrosion Fatigue of Alloy 625
Weld Claddings in Combustion Environments, Mater. High Temp., Vol 18 (No. 1),
2001, p 11
55. K. Stein, V. Guttmann, and W.T. Bakker,
The Inuence of Deformation on High
Temperature Corrosion of CRONIFER
45TM, in Heat-Resistant Materials II, Conf.
Proc. of the Second International Conference on Heat-Resistant Materials,
K. Natesan, P. Ganesan, G. Lai, Ed., ASM
56. V. Guttmann, K. Stein, and W.T. Bakker,
Deformation-Corrosion Interactions in
Selected Advanced High Temperature
Alloys, Mater. High Temp., Vol 14 (No. 2/3),
1997, p 61
57. P. Castello, V. Guttmann, N. Farr, and
G. Smith, Simulated Coal Ash Corrosion of
Ni-Based Alloys, Mater. High Temp., Vol 19
(No. 1), 2002, p 29
58. M.F. Stroosnijder, V. Guttmann, and

R.J.N. Gommans, Inuence of Creep
Deformation on the Corrosion Behavior of
a CeO2 Surface-Modied Alloy 800H in a
Suldizing-Oxidizing-Carburizing Environment, Mater. Sci. Eng., Vol A121, 1989,
p 581
59. J.L. Coze, et al., The Development of HighTemperature Corrosion-Resistant Aluminum-Containing Ferritic Steels, Mater. Sci.
Eng., Vol A120, 1989, p 293
60. E.C. Lewis and A.L. Plumley, Chromizing
for Combating Fireside Corrosion, in
Advances in Materials Technology for Fossil
Power Plants, R. Viswanathan and
R.I. Jaffee, Ed., ASM International, 1987,
p 291
61. H.M. Tawancy, Structure & Properties of
High Temperature Alloys: Applications of
Analytical Electron Microscopy, King Fahd
University of Petroleum & Minerals,
Dhahran, Saudi Arabia, 1993
62. L. Paul, G. Clark, and A. Ossenberg-Engels,
Protection of Waterwall Tubes from Corrosion in Low NOx Coal Fired Boilers,
presented at Coal-Gen Conference (Cincinnati, OH), Aug 1618, 2006
63. R.E. Kessler, Thermal Fatigue Cracking of
Waterwall Tubes from Waterlances and
Water Cannons, presented at EPRI Intelligent Sootblowing Workshop (Houston, TX),
March 1921, 2002
64. M. Carlisle, J. Sorge, and B. Mead, Water
Cannon Application at Alabama Powers
Plant Miller, presented at EPRI Intelligent
Sootblowing Workshop (Houston, TX),
March 1921, 2002
65. B. Ray, R. Hemperley, and R. Courtney,
W.A. Parish Units 7 & 8 ISB Project,
presented at EPRI Intelligent Sootblowing
Workshop (Houston, TX), March 1921,
2002
66. H.S. Blinka, W. A. Parish Unit 5 ISB Project, presented at EPRI Intelligent Sootblowing Workshop (Houston, TX), March
1921, 2002
67. R.W. Borio, A.L. Plumley, and W.R. Sylvester, in Ash Deposits and Corrosion Due
to Impurities in Combustion Gases,
R.W. Bryers, Ed., Hemisphere Publishing/
McGraw-Hill, 1978, p 163
68. C. Cain, Jr. and W. Nelson, J. Eng. Power,
Trans. ASME, Oct 1961, p 468
69. W. Nelson and C. Cain, Jr., J. Eng. Power,
Trans. ASME, July 1960, p 194
pg 318
30/10/2007 4:05PM Plate # 0
Chapter 10:
70. J.L. Blough and S. Kihara, Paper No. 129,

71. R.W. Borio and R.P. Hensel, J. Eng. Power,
Trans. ASME, Vol 94, 1972, p 142
72. J. Stringer, in, Corrosion, Vol 13, 9th ed.,
73. P. Castello, V. Guttmann, N. Farr and
G. Smith, Laboratory-Simulated Fuel-Ash
Corrosion of Superheater Tubes in CoalFired Ultra-Supercrtical-Boilers, Mater.
Corros., Vol 51, 2000, p 786
74. A.L. Plumley, J.I. Accort, and W.R. Roczniak, NACE-DOE Conference, Berkeley,
CA, 1979
75. S. Kihara, K. Nakagawa, A. Ohtomo, H.
Aoki, and S. Ando, Simulated Test
Results for Fireside Corrosion of Superheater and Reheater Tubes Operated at
Advanced Steam Condition in CoalFired Boilers, in High Temperature Corrosion in Energy Systems, M.F. Rothman, Ed.,
p 361
76. S. Van Weele, J.L. Blough, and J.H. DeVan,
77. J.L.
Blough,
ORNL/Sub/93-SM401/
01, Foster Wheeler Corp., Livingston, NJ,
1996
78. S. Kihara, J.L. Blough, W. Wolowodiuk, and
W.T. Bakker, Prediction of Corrosion Rate
of Superheater Tube in Boilers Burning
Various Kinds of Coals, presented at the
NACE International Conference on Life
Prediction of Corrodible Structures (Kauai,
HA), 1991
79. J.L. Blough, G.J. Stanko, W.T. Bakker, and
J.B. Brooks, Superheater Corrosion in
Ultra-Supercritical Power Plants, Paper No.
250, Corrosion/2000, NACE International,
2000
80. D.K. McDonald, Coal Ash Corrosion
Resistant Materials Testing Program: Evaluation of the First Section Removed in
November 2001, Babcock & Wilcox
Report, Barberton, OH
81. D.K. McDonald and E.S. Robitz, Coal
Ash Corrosion Resistant Materials Testing
Program: Evaluation of the Second Section
Removed in August 2003, Babcock &
Wilcox Report, Barberton, OH
82. E.P. Latham, T. Flatley, and C.W. Morris,
Comparative Performance of Superheated
Steam Tube Materials in Pulverized Fuel
Fired Plant Environments, in Corrosion
83.
84.
85.
86.
87.
88.
89.
90.
91.
92.
Resistant Materials for Coal Conversion

Systems, D.B. Meadowcroft and M.I. Manning, Ed., Applied Science Publishers,
London, U.K., 1982, p 137
T. Flatley, E.P. Latham, and C.W. Morris,
CEGB Experience with Co-Extruded Tubes
for Superheated and Evaporative Sections of
PF Fired Boilers, in Advances in Materials
Technology for Fossil Power Plants, Conf.
Proc., R. Viswanathan and R.I. Jaffee, Ed.,
M.G. Fahrmann and G.D. Smith, Evaluation
of Clad Tubing after 18 Years of Service
in a Coal-Fired Utility Boiler, Paper No.
2000
S.D. Kiser and T. Orsini, Benets of Chromium in Nickel for Corrosion Protection of
Superheater and Reheater Tubes in Coal
Fired Boilers, Paper No. 5455, Corrosion/
P.N. Hulsizer, Dual Pass Weld Overlay
Method and Apparatus, U.S. Patent No.
6,013,890, Jan 11, 2000
S.D. Kiser, E.B. Hinshaw, and T. Orsini,
Extending the Life of Fossil Fired Boiler
Tubing with Cladding of Nickel Based
Alloy Materials, Paper No. 06474, Corrosion/2006, NACE International, 2006
J. Stringer and I.G. Wright, Materials Issues
in Fluidized Bed Combustion, J. Mater.
Energy Systems, Vol 8 (No. 3), 1986, p 319
J. Stringer, Alloys for Advanced Power
Systems, Heat-Resistant Materials, Proc. of
the First International Conference, (Fontana,
WI), Sept 2326, 1991, K. Natesan and
D.J. Tillack, Ed., ASM International, 1991
P.L.F. Rademakers, D.M. Lloyd, and
V. Regis, AFBCs: Bubbling, Circulating
and Shallow Beds, in High Temperature
Materials for Power Engineering 1990,
Part I, Conf. Proc. (Liege, Belgium), Sept
2427, 1990, E. Bachelet, et al., Ed., Kluwer
Academic Publishers, Dordrecht, Netherlands, 1990, p 43
J. Stringer, Erosion/Corrosion in Fluidized
Bed Combustion Boilers, in Advances in
Materials Technology for Fossil Power
Plants, Conf. Proc., R. Viswanathan and
R.I. Jaffee, Ed., ASM International, 1987,
p 319
J.W. Stallings and J. Stringer, Mechanisms
of In-Bed Tube Wastage in Fluidized-Bed
Combustors, in Corrosion-Erosion-Wear of
Materials at Elevated Temperatures, Conf.
pg 319
30/10/2007 4:05PM Plate # 0
93.
94.
95.
96.
Proc., A.V. Levy, Ed., NACE International,

1991, p 19-1
I.G. Wright, A Review of Experience
of Wastage in Fluidized-Bed Boilers,
Mater. High Temp., Vol 14 (No. 2/3), 1997,
p 207
E.C. Lewis, D.A. Canonico, and R.Q. Vincent, Metal Wastage Experiences in BFBC
Environments, in Corrosion-Erosion-Wear
of Materials at Elevated Temperatures,
Conf. Proc., A.V. Levy, Ed., NACE International, 1991, p 20-1
J.W. Slusser, A.D. Bixler, and S.P. Bartlett,
Materials Experience from A Circulating
Fluidized Bed Coal Combustor, Paper No.
J.W. Slusser, A.D. Bixler, and D.E.
Thompson, Four Years of Field Performance
of A Circulating Fluidized Bed Combustor,

97. N.G. Solomon, Erosion-Resistant Coatings
for Fluidized-Bed Boilers, Paper No. 135,
98. M.D. Mirolli and W.P. Bauer, A Summary
of Combustion Engineerings Programs to
Control CFB Material Wastage, presented
at the EPRI-Argonne Workshop on Materials Issues in Circulating Fluidized Bed
Combustors, Argonne, IL, 1989
99. J. Zhao, R. Wu, R. Senior, R. Legros,
C. Brereton, J. Grace, and C. Lim, Spatial
Variation Inside a Pilot Scale Circulating
Fluidized Bed Combustion Unit, presented
at the EPRI-Argonne Workshop on Materials Issues in Circulating uidized Bed
Combustors, Argonne, IL, 1989
pg 320
26/10/2007 12:44PM Plate # 0

DOI: 10.1361/hcma2007p321
pg 321

CHAPTER 11
Oil-Fired Boilers and Furnaces

11.1 Introduction
Fuel oils primarily consist of residues from
distillation of crude oil in petroleum rening.
Sulfur in fuel oils can vary from a fraction of a
percent for lighter oils to 3.5% for some residual
oils (Ref 1). Two other corrosive constituents in
fuel oils are vanadium and sodium. Vanadium
exists in certain crude oils as an oil-soluble porphyrin complex (Ref 2). Fuel oils from some
crudes contain low levels of vanadium, while
certain Middle Eastern crudes and Venezuelan
crudes contain high levels of vanadium with up
to several hundred parts per million (ppm)
(Ref 2). Sodium can originate from crudes, the
neutralizer used in crude distillation during the
rening process, and contamination with seawater in transportation and storage (Ref 2).
During combustion, compounds formed by these
three constituents make up a major part of the ash
that deposits on the metallic components. For
example, superheater and reheater tubes in the
boiler, and tube hangers and other uncooled parts
in boilers or renery and petrochemical furnaces.
When the metal temperature of the component
reaches 540 C (1000 F) or higher, compounds
in the ash may become molten and corrosion
attack can become severe. This mode of corrosion is frequently referred to as oil-ash corrosion for boilers or furnaces red with fuel oils.
11.2 Oil-Ash Corrosion

During combustion of the fuel oils that contain
vanadium, sulfur, and sodium, low-melting
compounds can form. These compounds are
mixtures of vanadium pentoxide (V2O5), sodium
oxide (Na2O), or sodium sulfate (Na2SO4) (Ref
3). Signicant amounts of vanadium, sodium,
and sulfur were observed in the deposits collected from superheater/reheater tubes, or from
the uncooled tube hangers for the superheater/
reheater tubes in boilers. The analysis of the
deposits collected from corrosion probes at metal

temperatures of 475 to 675 C (890 to 1250 F)
from a utility boiler at Marchwood Power Station
(United Kingdom) is shown in Table 11.1 (Ref
4). Table 11.2 shows the results of the analysis
of the deposits from another utility boiler burning Buncker C fuel oil (Ref 5). Ash deposits
formed on metallic components whether cooled
(e.g., superheaters/reheaters) or uncooled (e.g.,
tube hangers) can exhibit low melting points.
Table 11.3 lists a number of oil-ash constituents
and their melting points (Ref 6). It is clear from
the table that the melting point of the ash salt
Table 11.1 Analysis of deposits in terms

of percent water soluble collected from the
corrosion probe at metal temperatures of
475675 C (8901250 F) at 0.4% boiler
oxygen in a utility boiler fired with fuel oil
containing 2.65% S, 49 ppm V, and 44 ppm Na
Content, %
Constituent
S as water soluble SO3

Na as water soluble Na2O
V as water soluble V2O5
V as water insoluble V2O5
Fe as water insoluble Fe2O3
Ni as water insoluble NiO
Other metal oxides
1-h exposure
5-h exposure
43.4
31.8
10.9
1.0
2.6
1.3
9.0
44.1
32.6
6.8
4.1
4.5
2.0
5.9
Source: Ref 4
Table 11.2 Analysis of deposits/scales

removed from the test specimens exposed in the
superheater area at temperatures from 565
to 850 C (1050 to 1560 F) in a utility boiler
Constituent
Si as silicon oxide
Al as aluminum oxide
V as vanadium pentoxide
S as sulfuric anhydride
Ni as nickel oxide
Na as sodium oxide
Mg as magnesium oxide
P as phosphate anhydride
Source: Ref 5
Content, wt %
7.8
2.0
36.37
32.80
2.16
5.12
2.35
0.5
26/10/2007 12:44PM Plate # 0
deposit can vary widely, depending on composition. The phase diagram for V2O5-Na2O system showing a series of low-melting eutectic
compounds is shown in Fig. 11.1 (Ref 7). Lowmelting compounds in the V2O5-Na2SO4 system
are shown in Fig. 11.2 (Ref 7). Some of the
eutectics become molten at 538 C (1000 F) or
even lower.
Table 11.3 Melting points of some oil-ash
constituents
Compound
Ferric oxide, Fe2O3

Ferric sulfate, Fe2(SO4)3
1565 (2850)
Decomposes at 480
(895) to Fe2O3
2500 (4530)
Decomposes at 1125
(2060) to MgO
2090 (3795)
Decomposes at 840
(1545) to NiO
1720 (3130)
880 (1615)
250 (480)
400 (750)
1970 (3580)
1970 (3570)
675 (1250)
630 (1165)
640 (1185)
850 (1560)
>900 (>1650)
>900 (>1650)
860 (1580)
855 (1570)
625 (1160)
Magnesium oxide, MgO

Magnesium sulfate, MgSO4
Nickel oxide, NiO
Nickel sulfate, NiSO4
Silicon oxide, SiO2
Sodium sulfate, Na2SO4
Sodium bisulfate, NaHSO4
Sodium pyrosulfate, Na2S2O7
Vanadium trioxide, V2O3
Vanadium tetraoxide, V2O4
Vanadium pentoxide, V2O5
Sodium metavanadate, Na2O.V2O5
Sodium pyrovanadate, 2Na2O.V2O5
Sodium orthovanadate, 3Na2O.V2O5
Nickel pyrovanadate, 2NiO.V2O5
Nickel orthovanadate, 3NiO.V2O5
Ferric metavanadate, Fe2O3.V2O5
Ferric vanadate, Fe2O3.2V2O5
Sodium vanadic vanadate,
Na2O.V2O4.V2O5
Sodium vanadic vanadate,
5Na2O.V2O4.11V2O5
535 (995)
Source: Ref 6
Fig. 11.1
Ref 7
Phase diagram for V2O5-Na2O system showing a

series of low melting eutectic compounds. Source:
Kawamura and Harada (Ref 8) observed that

when the value of the (Na +S)/V in at.% ratio
was 20 or higher, the melting point of the
deposits on the superheater tubes was more than
800 C (1470 F). However, this melting point
could be reduced to about 500 C (930 F) when
the ratio became less than 3 to 4. This is
illustrated in Fig. 11.3 (Ref 8). The analysis of the
chemical compositions (wt%) of the deposits that
formed on the superheater tubes when the boiler
was red with fuel oils with various ratios of
(Na + S)/V showed that the amounts of S as SO3,
V as V2O5, and Na as Na2O varied as a function
of the (Na +S)/V ratios in the fuel, as shown in
Table 11.4 (Ref 8).
It has been well accepted that the oil-ash corrosion is the result of the formation of molten
vanadate compounds involving V2O5-Na2O and/
or V2O5-Na2SO4 systems. Molten vanadate
compounds ux away the oxide scales formed on
the metal, thus causing rapid corrosion attack.
The corrosion mechanism is essentially hot corrosion involving a fused salt in an oxidizing
environment. The hot corrosion mechanism
involving fused Na2SO4 has been extensively
studied in high-temperature corrosion of gas
turbine components (see Chapter 9). The solubility of an oxide in a fused salt can vary as the
melt chemistry changes. A salt can dissociate
into a base (Na2O) and acid (SO3). The melt is
generally characterized by Na2O concentration
(melt basicity). When the melt exhibits high
solubility of an oxide formed on an alloy, a high
corrosion rate will occur. Rapp (Ref 9) provides
Fig. 11.2
Source: Ref 7
Melting points of low melting compounds in the

V2O5-Na2SO4 system during heating and cooling.
pg 322
26/10/2007 12:44PM Plate # 0
Chapter 11:
Fig. 11.3
pg 323
Oil-Fired Boilers and Furnaces / 323
Melting of the deposits formed on the superheater tubes as a function of the value of (Na + S)/ V ratio (in atomic percent) in the
fuel oils used in firing a boiler (375 MW) producing superheated steam of 570 C (1060 F). Source: Ref 8
Table 11.4 Chemical compositions (wt %) of the deposits formed on the superheater tubes in a boiler
when fired with fuel oils with different concentrations of vanadium
Fuel oil(a)
S as SO3
V as V2O5
Na as Na2O
Fe as Fe2O3
Ni as NiO
Ca as CaO
Mg as MgO
SO3+V2O5+Na2O
0.20.3% S
13 ppm V
2.72.8% S
4565 ppm V
1.61.8% S
130150 ppm V
2.42.5% S
200250 ppm V
51.8
0.85
34.4
4.70
3.38
2.06
1.92
87.1
24.4
30.0
17.6
13.0
6.42
2.25
1.41
72.0
21.6
49.7
17.8
11.2
2.24
1.17
0.88
72.0
0.89
83.0
2.69
6.48
7.45
0.22
0.20
86.6
(a) Sodium in fuel oils was in a range of 815 ppm. Source: Ref 8
a general review on the subject of the hot corrosion mechanism in fused salt. Figure 11.4
shows the solubilities of various oxides in fused
Na2SO4 at 927 C (1700 F) and 1.0 atm O2 as a
function of the activity of Na2O (aNa2O) (Ref 9).
Oil-ash corrosion may involve fused salt of
sodium vanadate and/or sodium sulfate. The
partial pressure of SO3 plays an important role in
the stability of the fused salt, such as sodium
vanadate, or sodium vanadate/sodium sulfate
mixture, or sulfate, formed on the metal. Seiersten and Kofstad (Ref 10) discuss the effect of
SO3 on vanadate-induced hot corrosion.
Fig. 11.4
11.3 Corrosion Problems in Oil-Fired

Boilers and Furnaces
Oil-ash corrosion problems in boilers are
typically associated with superheaters and reheaters because the tube metal temperature may
reach the melting point of the ash deposit. The
The solubilities of various oxides in fused Na2SO4 at

927 C (1700 F) and 1.0 atm O2 (The solubility of
SiO2 is at 900 C, or 1650 F). Source: Ref 9
uncooled components, such as tube hangers, in

oil-red boilers are subject to much higher
temperatures and thus can suffer more serious
corrosion problems. Many furnaces in renery
and petrochemical plants often use residual fuel
oils for heating high-temperature processing
26/10/2007 12:45PM Plate # 0
equipment. The uncooled components in these

furnaces can suffer severe oil-ash corrosion
attacks. This section discusses oil-ash corrosion
problems in both power boilers and renery
furnaces.
11.3.1 Waterwall Corrosion of Oil-Fired
Boilers
Because of much lower metal temperatures,
furnace waterwalls generally do not suffer serious corrosion problems in oil-red boilers. The
temperature of the reside metal surface for
the furnace waterwalls for fossil-red boilers including oil-red boilers in the United Kingdom
is generally limited to about 450 C (840 F)
(Ref 11). Reichel (Ref 12) indicates the metal
temperatures for the furnace waterwalls were in
the range of 300 to 460 C (570 to 860 F) for
fossil-red boilers including oil-red boilers in
Germany. However, in case of ame impingement, the furnace waterwall may be subjected
to higher-temperature exposure. Formation of
oxide scales or corrosion products on the internal
diameter (ID) of the waterwall tubes can also
signicantly increase the outer tube metal temperature. French (Ref 13) indicated that a thin
internal deposit can raise the tube metal temperature into the ash-corrosion range, into the
creep-failure range, or into the rapid-oxidation
range, leading to serious furnace-tube problems.
In some cases, oil-red boilers are used as a
peaking unit.* As a result, the load in these
units is often cycled daily or weekly, thus subjecting the boiler tubes to thermal cycling. The
thermal cycling can lead to circumferential
grooving and cracking of the waterwall tubes
under corrosive conditions such as under the oilash deposits and higher tube metal temperatures.
The development of circumferential grooving/
cracking on the waterwall tubes in coal-red
boilers is discussed in Chapter 10 Coal-Fired
Boilers. The example that is given below shows
the circumferential grooving/cracking of the
waterwall tubes that is believed to be the result of
thermal cycling, higher metal temperatures, and
oil-ash corrosive conditions (Ref 14).
A supercritical oil-red boiler (590 MWe)
required replacement of waterwall panels, which
were made of 2.25Cr-1Mo steel (T22), every 2
to 3 years because of severe corrosion. The
* The peaking unit is a boiler that is used mainly during highpower demand seasons, such as summer.
waterwall tube had the dimensions of 11/8 in.

outside diameter (OD) by 0.220 in. MWT.
Several panels that were aluminized by a commercial company were tried in the boiler. A tube
leakage developed after about 2.5 years of
service. Thus, the aluminized coating failed to
extend the life of the waterwall tube. A close-up
view of the waterwall tube showing circumferential grooves at the location near the localized
rupture area is shown in Fig. 11.5. The aluminized coating was completely destroyed. Oxidation attack penetrated through the aluminized
coating and into the steel tube wall (Fig. 11.6).
Fig. 11.5
Close-up view of the tube surface near the rupture

area showing numerous circumferential grooves
and cracks. Courtesy of Welding Services Inc.
Fig. 11.6
Optical micrograph showing the oxidation penetrated through the aluminized coating and into the
tube wall. Note the tube inside diameter surface is visible at
the right side bottom of the micrograph and the fire side is on the
leftside of the micrograph. Courtesy of Welding Services Inc.
pg 324
26/10/2007 12:45PM Plate # 0
Chapter 11:
The remnant material of the aluminized coating

can be seen in Fig. 11.7 (area No. 5 in Fig. 11.7).
The deposits and corrosion phases were analyzed
by scanning electron microscopy with energy
dispersive x-ray spectroscopy (SEM/EDX) with
the results summarized in Fig. 11.7. A signicant
amount of vanadium along with magnesium,
phosphorus, sodium, zinc, and other elements
was observed in the ash deposit and corrosion
products. The preferential corrosion penetration
that forms a groove into the metal was essentially
iron with little sulfur and is believed to consist of
iron oxides. Thermal cycling of the waterwall
caused by load changes and oil-ash corrosion
with possible higher waterwall tube temperatures
caused by ame impingement are believed to
have caused preferential oxidation penetrations,
which developed into circumferential grooves
and cracks. Kawamura and Harada (Ref 8) also
observed circumferential grooving or cracking
on furnace waterwall tubes made of 2.25Cr-1Mo
steel in an oil-red supercritical unit.
Fig. 11.7
Scanning backscattered electron image showing

the oxidation attack of the diffusion coating and the
substrate steel. The chemical analysis at different locations was
performed using EDX. The results of the analysis (wt %) are summarized as:
No. 1: 23% V, 20% Ni, 17% Si, 10% Fe, 8% Mg, 9% Al, 5% P, 3%
Na, 2% Zn, and trace elements.
No. 2: 53% Fe, 19% Ni, 12% Zn, 11% Al, 2% Mg, 2% V, and trace
elements.
No. 3: 34% Si, 24% Al, 12% V, 8% Fe, 7% Mg, 5% Na, 4% Ca, 2%
Ni, and trace elements.
No. 4: 54% Si, 20% Al, 8% Na, 6% Fe, 3% P, 4% Ca, and trace
elements.
No. 5: 70% Fe, 28% Al, 1% Cr, and trace elements.
No. 6: 74% Fe, 14% Al, 10% V, and trace elements.
No. 7: 90% Fe, 4% Cr, 2% Mo, 1% S, and trace elements.
No. 8: 90% Fe, 4% Cr, 2% Mo, 2% S, and trace elements.
No. 9: 97% Fe and trace elements.
11.3.2 Superheater and Reheater

Corrosion in Oil-Fired Boilers
To avoid the formation of molten vanadate
compounds in the ash deposits formed on
superheaters or reheaters in boilers, European
practice in general has limited the steam temperatures to 540 C (1000 F) (Ref 15). This
steam-outlet temperature limit implies that the
outer tube metal temperature is not greater than
580 C (1080 F) (Ref 11). In the United States,
oil-red boilers ordered after 1965 have design
steam-exit temperatures limited to 540 C
(1005 F) to prevent oil-ash corrosion problems
(Ref 16). The changes in steam-exit temperatures
for oil-red boilers constructed during the rst
eight decades of the 20th century are summarized
in Fig. 11.8 by Paul and Seeley (Ref 16). In
general, oil-ash corrosion problems tend to occur
at the secondary superheaters and reheaters
where the metal temperatures are the highest.
For austenitic stainless steels, corrosion rates
were found to be strongly dependent on the ue
gas temperature. Figure 11.9 shows the corrosion
rates of austenitic stainless steels as a function
of the tube surface temperature with ue gas
temperatures of 800 and 1150 C (1470 and
2100 F) (Ref 11). Austenitic stainless steels
suffered a rapid increase in wastage rates with
increasing tube surface temperature when the
ue gas was 1150 C (2100 F) compared with
the 800 C (1470 F) ue gas. Ferritic steels, on
the other hand, were less sensitive to the ue gas
temperature.
Holland et al. (Ref 15) conducted corrosion
probe tests in an oil-red boiler at Marchwood
Power Station for three austenitic stainless steels
(Type 316, 321, and 347) and a 12Cr ferritic steel
(12Cr, 0.5Mo, 0.25V). The boiler was operated
under base-load conditions with 0.5% O2 in ue
gas. The analysis of the fuel oil used during
testing showed 3.54% S, 91 ppm V, and 71 ppm
Na. The test results are summarized in Fig. 11.10
and 11.11 (Ref 15). The results conrmed that the
12Cr ferritic steel performed much better than all
three austenitic stainless steels. Among three
austenitic stainless steels, Type 347 was found
to be much better than Type 316 and 321. The
authors did not offer any explanation on the
performance ranking of the three austenitic
stainless steels. However, the results of the
chemical analyses of the three alloys showed
18.2% Cr for Type 347, 17.1% Cr for Type 321,
and 16.7% Cr for Type 316. It is, thus, believed
that the performance of three austenitic stainless
pg 325
26/10/2007 12:45PM Plate # 0
Fig. 11.8
Steam exit temperature, C
Steam exit temperature, F
Summary of the design steam-exit temperatures for oil-fired boilers built during the 20th century. Source: Ref 16
Fig. 11.9
Corrosion rates of austenitic stainless steels and

ferritic steels as a function of metal temperature
and flue gas temperatures. Source: Ref 11
steels followed the chromium concentration

levels in the alloys. The metal-loss data appear
to follow a linear kinetic when plotted as a
function of exposure time, as shown in Fig. 11.12
(Ref 15). This gure clearly shows the superior
performance of the 12Cr ferritic steel compared
with austenitic stainless steels containing 16 to
18% Cr.
Alexander et al. (Ref 17) found the similar test
results showing ferritic steels with less chromium
being signicantly more resistant to oil-ash corrosion than austenitic stainless steels with more
chromium. The authors compared several ferritic

steels with several austenitic stainless steels by
welding the tube samples together to form a test
tube. This test tube was then installed in the
secondary superheater tube bank in a boiler
red with fuel oil containing 4.0% S, 0.007% Na,
0.007% V, and 0.017% Cl at Bromborough
Power Station (United Kingdom). During the
exposure test, the metal temperatures ranged 315
to 680 C (600 to 1225 F) for about 20,026 h.
The tube was exposed to 590 to 620 C (1100
to 1150 F) for about 8000 h. Test tube samples
were measured for maximum metal loss caused
by the loss in thickness by corrosion (measured
after the corrosion scales were removed by descaling) and for the total metal wastage that
included maximum metal loss and intergranular
penetration (measured by metallography). Ferritic steels showed very little internal penetration,
while austenitic stainless steels suffered some
internal penetration. The data were plotted in
terms of the chromium concentration in alloys
for both ferritic steels and austenitic stainless
steels and are summarized in Fig. 11.13 (Ref 17).
More data comparing ferritic steels with austenitic stainless steels were generated by Parker
et al. (Ref 4) using corrosion probes tested in a
boiler red with fuel oil containing 2.65% S, 49
ppm V, and 44 ppm Na at Marchwood Power
Station (United Kingdom). Tube metal loss data
for ferritic steels and austenitic stainless steels are
summarized in Fig. 11.14 and 11.15 (Ref 4). The
test results showed both 9Cr and 12Cr ferritic
steels were signicantly more corrosion resistant
pg 326
26/10/2007 12:45PM Plate # 0
Chapter 11:
than austenitic stainless steels including Type

310 (25Cr-20Ni). It is somewhat surprising to
nd that 2.25Cr-1Mo steel was much more
resistant than austenitic stainless steels (Fig.
11.14). It is more surprising to nd that 2.25Cr1Mo steel was comparable to 9Cr and 12Cr steels
at approximately 550 to 625 C (1020 to
1155 F). This was in contrast with test results
that were generated by Alexander et al. (Ref 17)
in a boiler at Bromborough Power Station,

showing that increasing chromium concentration
signicantly increased the corrosion resistance
for ferritic test steels 2.25Cr-1Mo, 9Cr, and 12Cr
steels (Fig. 11.13). Among the four austenitic
stainless steels, Type 310 (25Cr-20Ni) was more
resistant than Type 316, 321, and 347 due to
higher chromium content. The test results generated at the upper furnace are summarized
Fig. 11.10
Results of corrosion probe tests for 1000 h as a function of the tube metal temperature in a boiler fired with fuel oil
containing 3.54% S, 91 ppm V, and 71 ppm Na. Source: Ref 15
Fig. 11.11
Results of corrosion probe tests for 2000 h as a function of the tube metal temperature in a boiler fired with fuel oil
containing 3.54% S, 91 ppm V, and 71 ppm Na. Source: Ref 15
pg 327
26/10/2007 12:45PM Plate # 0
in Fig. 11.15 (Ref 4), again showing both ferritic steels were more resistant than austenitic
stainless steels. In this upper furnace test, 2.25Cr1Mo steel was not included.
The plant exposure tests have shown that 9Cr

and 12Cr ferritic steels are much more resistant to
oil-ash corrosion than austenitic stainless steels,
such as Type 316, 321, and 347, typically used
Exposure time, h
Fig. 11.12
Fig. 11.13
Metal loss as a function of exposure time for 12Cr ferritic steel and three austenitic stainless steels. Source: Ref 15
Tube wastage data, which included both metal loss (tube thickness loss) and total wastage (metal loss+intergranular
penetration), for ferritic steels and austenitic stainless steels in terms of chromium concentration in alloys. The data were
generated in a boiler at Bromborough Power Station (United Kingdom) fired with fuel oil containing 4.0% S, 0.007% Na, 0.007% V, and
0.017% Cl. Open data points: total wastage (maximum metal loss+intergranular penetration). Solid data points: Maximum metal loss.
Source: Ref 17
pg 328
26/10/2007 12:45PM Plate # 0
Chapter 11:
for superheaters and reheaters in utility boilers.

Nevertheless, the ferritic steels may not have
adequate creep strengths at higher temperatures.
Composite tubes with austenitic stainless steels
for providing materials strength and outer tube
cladding for providing corrosion resistance offer

better alternatives for higher-temperature applications. Bolt (Ref 18) evaluated some of the
composite tubes in an oil-red experimental
boiler heated with fuel oil containing 2.2% S, 200
ppm V, and 50 ppm Na at metal temperatures
from 500 to 700 C (930 to 1290 F) and two
ue gas temperatures 1000 and 1125 C (1830
and 2060 F). The tests were conducted in
corrosion probes for 2000 h of exposure with a
continuous load (i.e., constant metal temperature) and a discontinuous load (varying temperatures8 h at higher and 16 h at lower levels).
Four coextruded composite tubes tested were
Esshete 1250/Type 310, 800H/Type 446, 1714
CuMo/35CrA, and 800H/671; Type 310, Type
446, 35CrA, and 671 were cladding alloys. The
35CrA cladding was 35Cr-45Ni-Fe alloy, and
the 671 cladding was 46Cr-Ni alloy. The test
results are summarized in Fig. 11.16 and 11.17.
A bell-shaped curve with the maximum attack at
about 650 C was observed for Type 347 for both
continuous and discontinuous loads. Type 347
and 310 showed unacceptable corrosion rates (>1
mm/yr, or >39 mpy) at 630 to 675 C (1165
to 1250 F). Type 446 cladding appeared to be
Fig. 11.14
Maximum metal loss as a function of metal temperature tested for 10,000 h in the superheater
inlet zone in a boiler fired with fuel oil containing 2.65% S, 49
ppm V, and 44 ppm Na. Source: Ref 4
Type 347H
Fig. 11.15
Maximum metal loss as a function of metal temperature tested for 10,000 h in the combustion
zone in a boiler fired with fuel oil containing 2.65% S, 49 ppm V,
and 44 ppm Na. Source: Ref 4
Fig. 11.16
Corrosion rates of Type 347 in comparison

with 310, 446, 35CrA, and 671 claddings as a
function of temperature (C) with flue gas temperatures of 1000
and 1125 C (1830 and 2060 F) in a constant thermal load.
Source: Ref 18
pg 329
26/10/2007 12:45PM Plate # 0
sufciently corrosion resistant. Both 35CrA and

671 claddings performed extremely well even
though alloy 35CrA contained about 45% Ni and
alloy 671 contained about 54% Ni with 35% Cr
in the former and 46% Cr in the latter.
Carburization has been observed at times in
superheaters and reheaters. Lopez-Lopez et al.
(Ref 19) found severe carburization in a heavily
corroded Type 321H superheater tube after
107,000 h of service in a boiler that had been
used as a peaking unit. The tube had suffered
severe metal wastage. However, underneath the
corroded surface was a shallow, severely carburized layer about 80 m deep and a peak
hardness of about 400 HV (Vickers hardness
number). In this study, the authors failed to report
(or analyze) the chemistry of the ash deposits
including carbon collected from the superheater
tube. In a separate study, Wong-Moreno et al.
(Ref 20) analyzed nine different ash deposits
collected from superheaters in different oil-red
boilers ranging from 84 to 350 MW. Most deposits contained about 0.03 to 0.06% C, with
only two deposits showing about 0.14% C. This
level of carbon in the deposits is unlikely to cause
severe carburization in austenitic stainless steels
at the superheater tube temperatures. Therefore,
Discontinuous load,
Discontinuous load,
Type 347H
Fe
it is believed that the observed carburization

was not caused by oil-ash deposits. Internal
carburization can occur in CO2-containing
environments. McNallan et al. (Ref 21) observed
internal carburization of Type 310 in Ar-20CO2
at 800 C (1470 F) for 24 h. Furthermore, it is
believed that the observed carburization was not
the major factor causing the severe superheater
wastage problem. Carburization was often observed to take place after the alloy suffered severe
wastage with no protective oxide scales on the
corroded surface.
In addition to selecting a better-performing
alloy to resist oil-ash corrosion, methods that
could reduce superheater and reheater corrosion
include (a) decreasing excess air for combustion
and (b) the use of additives to raise the oil-ash
melting point. Excess air promotes the formation
of SO3 and of V2O5, which melts at 675 C
(1250 F), instead of V2O3 or V2O4, which melts
at 1970 C (3580 F) (Ref 6). However, keeping
very low excess air for combustion can probably
be difcult to maintain in practice. The Central
Electricity Generating Board (CEGB) in the
United Kingdom claimed to have made signicant achievements in combustion in oil-red
boilers with low excess air (Ref 22). Boilers at
Merksem Power Station were reported to have
operated at 1% excess air (0.2% O2) as a regular
operating practice (Ref 23).
The use of additives to increase the melting
point of the vanadate or sulfate is likely to be
a more practical approach. Useful additives include (Ref 8):
Magnesium compounds: MgO, Mg(OH)2,

MgCO3, MgSO4, MgCO3CaCO3, organic
magnesium compounds
Calcium compounds: CaO, Ca(OH)2, CaCO3
Barium compounds: BaO, Ba(OH)2
Aluminum and silicon compounds: Al2O3,
SiO2, 3Al2O32SiO2
The additive reacts with vanadium compounds to
form reaction products with higher melting
points. When magnesium compounds are used,
some of the reaction products and their melting
points are:
Fig. 11.17
Corrosion rates of Type 347 in comparison with

310, 446, 35CrA, and 671 claddings as a function
of temperature (C) with flue gas temperatures of 1000 and
1125 C (1830 and 2060 F) in varying loads. Source: Ref 18
MgOV2O5: 671 C (1240 F)

2MgOV2O5: 835 C (1535 F)
3MgOV2O5: 1191 C (2175 F)
With injection of magnesium compounds into
the fuel, the higher Mg/V ratios increase the
pg 330
26/10/2007 12:45PM Plate # 0
Chapter 11:
melting point of the oil-ash deposits (Fig. 11.18)

(Ref 6). Increasing the melting point of oil-ash
deposits would result in lowering the corrosion
rates. Figure 11.19 shows the effectiveness of the
Mg(OH)2 additive injection in reducing the corrosion rate of Type 321 superheater tubes (Ref 8).
Disadvantages of the additive injection approach
include additional operating costs and a
substantial increase in ash volume that may
require additional furnace downtime for tube
cleaning (Ref 24).
11.3.3 Tube Supports or Hangers in Boilers
and Refinery/Petrochemical Furnaces
Many renery and petrochemical furnaces are
red with residual fuel oils containing signicant
amounts of sulfur, vanadium, and sodium.
Structural supports and tube hangers for the
Fig. 11.18
Effect of MgO addition on the melting point of oilash deposit on superheater tubing of an oil-fired
boiler. Source: Ref 6
Fig. 11.19
Source: Ref 8
Corrosion of Type 321 superheater tubes with and

without Mg(OH)2 injection in an oil-fired boiler.
heater tubes used to process hydrocarbon uid

in the furnace are uncooled and thus can be
subjected to temperatures as high as 900 C
(1650 F) or higher. Severe materials problems
due to oil-ash corrosion were illustrated by
numerous case histories presented in a 1958
NACE Technical Committee Report (Ref 25).
For example, Type 309 tube supports suffered a
corrosion rate of about 12.7 mm/yr (500 mpy) in
an oil-red heater in a renery plant (Ref 25).
This heater increased the temperature of the oil
charge to 350 to 470 C (670 to 880 F) in a
furnace operating at 940 to 980 C (1700 to
1800 F) red with residual fuel containing
about 4.1% S and 0.05% ash composed of about
16.2% V2O5. Another example involving
uncooled superheater spacers made of cast HH
stainless steel (25Cr-12Ni) failed after only
7 months of service in a boiler red with a
Buncker C oil, producing 565 C (1050 F)
steam (Ref 25). The corrosion of the HH tube
spacers was found to proceed by suldation/
oxidation attack (Ref 25).
McDowell et al. (Ref 5) conducted eld test
racks in a utility boiler red with Buncker C
fuel oil. The test racks were exposed to the
ue gas in the second bank of the superheater.
During testing, the boiler operated at full load
during the day for 5 days a week and at half
and low loads during early morning hours and
weekends. The temperature of the ue gas was
about 850 C (1560 F) during full load of
power, 660 C (1220 F) during half load, and
565 to 590 C (1050 to 1100 F) during minimum loads. No fuel analysis was reported in the
paper. Analysis of the deposits/corrosion scales
showed a signicant amount of V2O5 (Table
11.2). The test results, which were extrapolated
from three test racks exposed for 504, 648, and
707 h, respectively, are summarized in Table
11.5. None of the alloys tested showed acceptable wastage rates.
More rack tests in boilers red with Buncker
C oils containing high concentrations of
vanadium (150 to 450 ppm) were reported by
McDowell and Mihalisin (Ref 26). Test racks
were exposed to the ue gas in the superheater
section. Alloys ranging from low-alloy steels
to iron- and nickel-base alloys suffered severe
corrosion attack. The results of one test rack are
shown in Table 11.6 (Ref 26). Specimens
included 5Cr steel, stainless steels (both 400 and
300 series), Fe-Ni-Cr alloys, Ni-Cr-Fe alloy, and
50Ni-50Cr alloy, along with two cast stainless
steels (HE and HH alloys). All the alloys tested
pg 331
26/10/2007 12:45PM Plate # 0
exhibited unacceptable corrosion rates. Even the

best performer (50Ni-50Cr alloy) suffered a
corrosion rate of 3.1 mm/yr (121 mpy) (Ref 26).
Some investigators have reported excellent
performance of 50Cr-50Ni alloy as tube supports in furnaces or boilers red with fuel oils
containing sulfur and vanadium. Spafford (Ref
27) reported good performance of the 50Ni-50Cr
alloy in renery heaters for coking and catalytic
reformer units. The heaters were red with heavy
fuel oil containing 2.5 to 4% S and 50 to 70 ppm
V (occasionally up to 150 ppm). The original
Table 11.5 Results of field rack tests conducted
in the second bank of the high-temperature
superheater at the location about 3 ft from the
brick side wall in a utility boiler fired with
Buncker C fuel oil with flue gas stream
fluctuating from 565 to 850 C (1050 to 1560 F)
while the boiler changed from half to full loads
during the exposure test
Alloy
Carbon steel
2.25Cr-1Mo
502 (5Cr steel)
410 (12Cr steel)
406 (12Cr-3Al steel)
430
446
302
321
309
800
HW (12Cr-60Ni)
HF (21Cr-9Ni)
HE (28Cr-10Ni)
Inconel 600 (Ni-15Cr-7Fe)
Wastage rate, mm/yr (mpy)
17.7 (695)
17.019.1 (670750)
6.214.2 (244557)
14.1 (555)
2.7 (108)
9.526.9 (3741060)
4.8 (189)
10.3 (406)
17.5 (690)
6.27.7 (244305)
5.5 (217)
5.815.3 (230601)
10.7 (422)
3.1 (124)
2.9 (113)
hangers and tube supports made of cast HH alloy

(25Cr-12Ni steel) suffered severe corrosion
attack. Metal temperatures were in the range of
730 to 890 C (1350 to 1630 F). The highest
corrosion rates of cast HH alloy were 6.4 to 9.5
mm/yr (250 to 375 mpy). Replacements with
alloy 657 (a cast 50Ni-50Cr alloy) were reported
to perform very well, with minimal maintenance
and repair (Ref 27).
Swales and Ward (Ref 2) reported the results
of a eld test that showed alloy 657 performed 10
times better than HH and HK alloys, as illustrated
in Fig. 11.20. There are additional cases where
alloy 657 tube supports provided excellent performance; two such examples (Ref 2) are given
below. The authors also concluded that at temperatures higher than 900 C (1650 F), alloy
657 often suffered severe corrosion attack, but
provided good protection up to that temperature
(Ref 2).
Catalytic Reformer Heaters in a Refinery.
The furnace was red with a mixture of fuel oil
(about 50%) and gas. The ash analysis showed
48% V2O5 and 2% Na2O. Tube supports were
exposed to temperatures up to 900 C (1650 F).
The tube supports made of cast HH stainless steel
(25Cr-12Ni) suffered severe corrosion after
2 years of service. Tube supports were replaced
with cast alloy 657 (Ni-48Cr) showing no noteable corrosion after 7 years of service.
Source: Ref 5
Table 11.6 Results of a field test (uncooled

specimens) exposed in the superheater section
at 815 C (1500 F) in a boiler fired with high
vanadium (150450 ppm) Buncker C fuel
Alloy
5Cr steel
406
431
446
302B
309
321
347
310
Incoloy(a)
Incoloy 804
Inconel(a)
50Cr-50Ni
HE
HH
Wastage rate, mm/yr (mpy)
3245 (12701775)
5.7 (224)
1723 (655925)
3.8 (149)
1416 (533644)
5.0 (196)
913 (346505)
6.2 (243)
4.7 (187)
912 (364458)
1318 (495710)
5.0 (196)
3.1 (121)
4.8 (187)
1216 (467645)
(a) Incoloy may refer to Incoloy alloy 800, while Inconel may refer to Inconel alloy
600. Source: Ref 26
Fig. 11.20
Results of a field test at 700 C (1290 F) for

6 months in a refinery (crude oil) heater comparing alloy 657 to HH and HK alloys. Source: Ref 2
pg 332
26/10/2007 12:45PM Plate # 0
Chapter 11:
Catalytic Reformer Heater in a Refinery.

The furnace was red with fuel oil containing
3 to 4% S, 40 to 50 ppm V (90 to 150 ppm
occasionally for short periods). Tube hangers
were exposed to 700 to 900 C (1290 to
1650 F). The original HH cast roof tube hangers
suffered severe wastage problems after 1 to
2 years of service. All tube hangers were then
replaced with alloy 657. No failures occurred
after more than 4 years of service for alloy 657
hangers.
11.4 Summary
Fireside corrosion can present a serious problem in oil-red boilers or renery/petrochemical
furnaces red with low-grade fuels with high
concentrations of vanadium, sulfur, and sodium. This corrosion is frequently referred to as
oil-ash corrosion. Accelerated attack by oil-ash
corrosion is related to the formation of lowmelting point molten vanadium pentoxide and
sodium sulfate eutectics, which ux the protective oxide scale from the metal surface. In
boilers, superheater and reheater tubes are susceptible to oil-ash corrosion attack. Uncooled
components in the boilers, such as tube supports
and spacers, can suffer severe corrosion attack
because of higher temperatures. Oil-ash corrosion can also occur in renery and petrochemical
furnaces burning low-grade fuels. The resistance
of oil-ash corrosion for various alloys in both
boilers and renery/petrochemical furnaces is
reviewed.
REFERENCE

Wilcox, Barberton, OH, 1992
2. G.L. Swales and D.M. Ward, Strengthened
50% Chromium, 50% Nickel Alloy (IN657)
Renery Heater Tube Supports to Combat
Fuel Ash CorrosionA Review of Service
Caswe Histories, Paper No. 126, Corrosion/
79, NACE, 1979
3. D.N. French, Corrosion of Superheaters
and Reheaters in Fossil Fired Boilers,
Environmental Degradation of Engineering
Materials in Aggressive Environments,
Conf. Proc., Virginia Polytechnic Institute,
Blacksburg, VA, 1981, p 407
4. J.C. Parker, D.F. Rosborough, and M.J. Virr,

High Temperature Corrosion Trials at
Marchwood Power Station10,000 Hour
Corrosion Probe Trials, J. Inst. Fuel, Feb
1972, p 95
5. D.W. McDowell, R.J. Raudebaugh, and
W.E. Somers, High-Temperature Corrosion
of Alloys Exposed in the Superheater of
an Oil-Fired Boiler, Trans. ASME, Feb 1957,
p 319
6. M. Fichera, R. Leonardi, and C.A. Farina,
Fuel Ash Corrosion and Its Prevention with
MgO Addition, Electrochim. Acta, Vol 32
(No. 6), 1987, p 955
7. W.T. Reid, External Corrosion and Deposits
Boiler and Gas Turbines, Elsevier Publishing, 1971
8. T. Kawamura and Y. Harada, Control of
Gas Side Corrosion in Oil Fired Boilers,
Mitsubishi Technical Bulletin No. 139,
Mitsubishi Heavy Industries, Ltd., May
1980
9. R.A. Rapp, Hot Corrosion of Materials, Pure
Appl. Chem., Vol 62 (No. 1), 1990, p 113
10. M. Seiersten and P. Kofstad, The Effect of
SO3 on Vanadate-Induced Hot Corrosion,
p 115
11. A.J.B. Cutler, T. Flatley, and K.A. Hay, FireSide Corrosion in Power Station Boilers,
Metall. Mater. Technol., Feb 1981, p 69
12. H.H. Reichel, Fireside Corrosion in German
Fossil-Fuel Fired Power Plants: Appearance,
Mechanism and Causes, Werkst. Korros.,
Vol 39, 1988, p 54
13. D.N. French, Metallurgical Failures in
Fossil Fired Boilers, 2nd ed., John Wiley &
Sons, 1993, p 370
14. Welding Services Inc. unpublished data
15. N.H. Holland, D.F. ODwyer, D.F. Rosborough, and W. Wright, High-Temperature
Corrosion Investigations on an Oil-Fired
Boiler at Marchwood Power Station, J. Inst.
Fuel, May 1968, p 206
16. L.D. Paul and R.R. Seeley, Oil Ash
CorrosionA Review of Utility Boiler
Experience, Paper No. 267, Corrosion/90,
17. P.A. Alexander, R.A. Marsden, J.M. NelsonAllen, and W.A. Stewart, Operational Trials
of Superheater Steels in a C.E.G.B. OilFired Boiler at Bromborough Power Station,
J. Inst. Fuel, Feb 1964, p 59
18. N. Bolt, Fireside Corrosion Phenomena in
Superheaters of Coextruded Materials with
pg 333
26/10/2007 12:45PM Plate # 0
19.
20.
21.
22.
18, 25, 35 and 50% Cr at Metal Temperatures of up to 700 C, International Congress on Metallic Corrosion, Conf. Proc.,
Vol IV, Oxford & IBH Publishing Co. Pvt.
Ltd, 1987, p 3593
D. Lopez-Lopez, A. Wong-Noreno, and
L. Martinez, Unusual Superheater Tube
Wastage Associated with Carburization,
Mater. Perform., Dec 1994, p 45
A. Wong-Moreno, Y.M. Martinez, and L.
Martinez, High Temperature Corrosion
Enhanced by Residual Fuel Oil Ash
Deposits, Paper No. 185, Corrosion/94,
M.J. McNallan, S. Thongtem, J.C. Liu, Y.S.
Park, and P. Shyu, Corrosion of Chromium
Containing Alloys in Non-steady State
Environments Containing Oxygen, Carbon,
and Chlorine, J. Phys. IV, Coll. C9, Suppl. J.
Phys. III, Vol 3, Dec 1993, p 143
E.M. Hamilton and G.R. Stern, Operation
of Oil Fired Boilers with Very Low Excess
23.
24.
25.
26.
27.
Air, CEGB Report RD/H/MI, Central

Electricity Generating Board
J. Remeysen, Operation of Large Boilers at
Very Low Excess Air Levels, Proceedings
of Brussels Conference of Institute of Fuel,
Paper 1, Sept 1964
J.R. Wilson, Understanding and Preventing
Fuel Ash Corrosion, Paper No. 12, Corrosion/76, NACE, 1976
The Present Status of the Oil Ash Corrosion
Problem, NACE Technical Committee
Report, Pub. 58-11, Corrosion, 1958,
p 369t
D.W. McDowell, Jr. and J.R. Mihalisin,
Paper No. 60-WA-260, presented at ASME
Winter Annual Meeting, Nov 27Dec 2,
1960
B.F. Spafford, in UK Corrosion 82, Conf.
Proc. (Birmingham, U.K.), Nov 1517
1982, Institution of Corrosion Science &
Technology, Birmingham, U.K., 1982,
p 67
pg 334

DOI: 10.1361/hcma2007p335
30/10/2007 4:09PM Plate # 0

CHAPTER 12
Waste-to-Energy Boilers and

Waste Incinerators
12.1 Introduction
Municipal solid waste (MSW) is a combination of residential, commercial, and industrial
refuse. Historically, the typical means of disposal of this refuse has been landlling. A novel
approach in recovering the energy from this
resource in boilers for electricity generation
and minimizing the landll requirements began
in the 1960s in Europe. It became apparent
from these early European boilers that corrosion of the boiler tube materials was a major
obstacle for operating these so-called wasteto-energy (WTE) boilers (Ref 1, 2). Managing
the corrosion problems continues to be a big
challenge for operators of modern boilers
worldwide. This chapter focuses on the corrosion
issues related to boilers burning municipal solid
waste.
12.2 Fuels, Combustion Environments,

and Boilers
Municipal solid waste typically contains paper
and paperboard, plastics, rubber, textile, leather,
batteries, food waste, yard waste, metal, glass,
and other miscellaneous materials. This type
of fuel is very heterogeneous and varies greatly
by geographic location, country, and culture.
In North America, typical constituents of
the MSW in 1990 were paper and paperboard
(34.240.0%), plastics (7.29.2%), food waste
(7.38.5%), yard waste (17.619.9%), metal
(8.19.6%), glass (7.79.7%), and other (10.3
13.2%) (Ref 3). Polyvinyl chloride (PVC) plastic
is a dominant source of chlorine, which is a major
corrosive constituent making the combustion
environment corrosive to boiler tube materials.
Polyvinyl chloride contains about 36 wt%
chlorine (Ref 4). Rubbers such as Hypalon,

chloroprene, and neoprene also contain high
concentrations of chlorine (Ref 4). Other waste
constituents that can contribute to corrosive
environments under combustion are batteries
(automobile batteries, button-type batteries in
watches and calculators, alkaline, zinc, nickel,
and cadmium batteries) and consumer electronics. Batteries contribute lead and cadmium;
household batteries contribute cadmium; and
consumer electronics contribute both lead and
cadmium (Ref 5). Rademakers et al. (Ref 6)
reported their calculation of the heavy metal
composition of Dutch solid waste based on 1994
data in g/kg waste: 0.01 Cd, 0.0004 to 0.0006 Hg,
0.22 to 0.56 Pb, 0.02 to 0.04 Sb, 0.63 to 1.04 Zn,
0.01 As, 3.96 to 6.89 Cl, 0.11 to 0.19 Br, 0.1 to
0.2 F, and 1.48 to 2.96 S. During combustion,
these heavy metals can react with chlorine to
form some low-melting-point chlorides, thus
causing severe boiler tube material corrosion
problems. Corrosion mechanisms are discussed
in the next section.
There are primarily two types of combustion
technologies, namely, mass-burning and refusederived fuel (RDF) burning. For mass-burning
units, the fuel, which is not segregated except for
appliances, furniture, and other large articles,
burns in the as-received condition. In RDF units,
on the other hand, the fuel is segregated, classied, and shredded to size. Metals and glass are
typically removed from RDF fuel prior to combustion. Thus, RDF fuel has a higher heating
value than that of a mass-burning unit. A small
number of boilers in use today are based on the
uidized-bed technology. The mass-burning unit
uses a mechanical stoker to feed the fuel onto a
grate where the fuel is combusted. For most RDF
units, the fuel is blown into the furnace sidewalls
through an RDF burner with a fuel-distribution
impeller (Ref 3). Most RDF units burn the fuel in
pg 335
30/10/2007 4:09PM Plate # 0
pg 336
boilers developed in the 1960s were designed

to operate at higher steam pressures and temperatures, particularly those in Germany, with
one in Mannheim operating at 12.5 MPa/530 C
(1800 psig/980 F) steam and another one in
Munich at 18 MPa/540 C (2650 psig/1000 F)
steam (Ref 1). Higher steam temperatures result
in higher tube metal temperatures and thus more
serious corrosion problems.
Combustion of the MSW fuel is signicantly
different from that of coal or oil because the solid
waste is a heterogeneous fuel. In addition, the
fuel contains numerous impurities that include
chlorine, sulfur, sodium, potassium, cadmium,
zinc, lead, and other heavy metals. Because of
these impurities, combustion of this fuel generates a very corrosive environment that causes
serious corrosion problems for boiler tube
materials. Gaseous combustion products include
N2, O2, CO2, H2O, SO2, HCl, HF, and other
gaseous impurities such as CO and HBr. These
gaseous constituents are often measured by
plant operators. Examples of these ue gas
compositions measured in mass-burning units at
different plants are shown in Table 12.1 (Ref 7,
1113). However, vapors of metal chlorides and
sulfates are also produced during combustion.
These compounds are normally not quantied by
plant personnel. Many of these metal chlorides
exhibit high vapor pressures and/or low melting
points. Some physical properties of many metal
chlorides can be found in Chapter 6 Corrosion
by Halogen and Halides. As is discussed in this
chapter, some of these metal chlorides are primarily responsible for the corrosion of the boiler
tube materials.
The furnace is typically enclosed by four walls
(rear, front, and two side walls) using a tubemembrane construction (i.e., individual tubes
are connected by narrow plates). These tubemembrane walls, which are commonly referred
to as waterwalls, provide heat-absorbing
suspension (Ref 3). RDF units also employ

moving grates.
Combustion gas temperatures are generally
at or below approximately 1090 C (2000 F) in
mass-burning units and approximately 1315 to
1370 C (2400 to 2500 F) in RDF units (Ref 7).
The guidelines for solid waste combustion
introduced by the European authorities require
that the temperature should be above 850 C
(1560 F) for a 2 s residence time at the level
1 m above the secondary or tertiary air (Ref 6).
The MSW boilers operate at much lower steam
pressures and temperatures compared with coalred utility boilers, as discussed in Chapter 10.
Kubin (Ref 7) reported that Ogden Martin boilers
in the United States typically operated at steam
pressures of 4 to 6 MPa (615 to 880 psig) and
temperatures of 370 to 440 C (700 to 830 F)
for mass-burning units, and steam pressures of
4.5 to 6 MPa (665 to 900 psig) and temperatures
of 370 to 440 C (700 to 830 F) for RDF units.
These pressures and temperatures appear to be
representative of other WTE boilers in the United
States. For example, two boilers at the Wheelabrator Concord facility operate at a steam pressure of 650 psig (4.5 MPa) and a temperature of
400 C (750 F), and two at the Wheelabrator
Spokane facility operate at 6.2 MPa (900 psig)
and 440 C (830 F) steam (Ref 8), with some
other Wheelabrator boilers operating at higher
pressures and temperatures (Ref 9). Some of the
European boilers operate at higher steam pressures and temperatures. For example, the boilers
at Ivry Paris (France) operate at 7.5 MPa
(1103 psig, or 75 bar) and 480 C (900 F)
steam, at HKW Mannheim (Germany) at 8 MPa
(1176 psig, or 80 bar) and 500 C (930 F)
steam, at BSR Berlin (Germany) at 7.5 MPa
(1088 psig, or 74 bar) and 420 C (790 F)
steam, and at EVO Oberhausen (Germany)
at 7 MPa (1029 psig, or 70 bar) and 480 C
(896 F) steam (Ref 10). Early European WTE
Table 12.1
Examples of flue gas compositions generated by different WTE mass-burning boilers

Composition, vol%
Gas
Ogden-Martin (US)(a)
Saint Ouen (France)(b)
Richtlinien (Germany)(c)
Kuririn (Japan)(d)
O2
CO2
SO2
HCl
HF
CO
H 2O
N2
8.59.5
89
100200 ppm
400600 ppm
520 ppm
1315
bal
910
912
90130 ppm
6001200 ppm
2040 ppm
<20 ppm
NR
bal
NR
NR
1002000 ppm
5602240 ppm
64640 ppm
NR
NR
9.1
11.1
40 ppm
810 ppm
42 ppm
21.4
bal
NR, not reported. (a) Source: Ref 7. (b) Source: Ref 11. (c) Source: Ref 12. (d) Source: Ref 13
30/10/2007 4:09PM Plate # 0
Chapter 12: Waste-to-Energy Boilers and Waste Incinerators / 337
surfaces for converting the water inside the tubes

to steam as the water rises from the bottom of the
waterwall. At the top of the waterwall tubes, a
mixture of steam and water leaves the waterwall
tubes and enters the steam drum where steam is
separated from water. The furnace waterwall
typically operates with water at a saturation
temperature. Thus, the temperature of the water/
steam in the waterwall tubes, which depends on
the pressure of the water/steam, is generally at or
below 277 C (530 F) (Ref 14). The ranges of
the waterwall tube metal temperatures were cited
to be 260 to 293 C (500 to 560 F) (Ref 7), 260
to 290 C (500 to 550 F) (Ref 15), and 260 to
315 C (500 to 600 F) (Ref 16). For higherpressure boilers, the temperature of the steam
will be higher. The steam from the steam drum
after separating from water is then further heated
in superheaters (typically two superheaters
primary and secondary or nal superheaters) to
higher temperature in the convection path before
it is delivered to turbines for electricity generation. The combustion ue gas exits from the
furnace at the top and then enters into the convection path, owing typically through the
superheater, then the boiler bank, and the economizer. The ue gas temperature entering into
the superheater may vary from approximately
650 to 900 C (1200 to 1650 F).
Figure 12.1 shows a schematic of a massburning unit. For some mass-burning units,
screen tubes are installed in front of the superheater to lower the temperature of the ue gas
entering into the superheater section. In some
other mass-burning units, there are additional gas
passes (one or more additional walls to allow ue
gas to pass through in the convection path) to
allow the ue gas stream to lower its temperature
before entering into the superheater section.
This type of design is shown schematically in
Fig. 12.2 (Ref 17). Licata et al. (Ref 17) indicated
that passing ue gas through two 180 turns and
one 90 turn prior to entering the superheaters
can not only lower the ue gas temperature but
also remove particles from the ue gas stream. In
addition, the ue gas stream can achieve better
mixing to minimize local reducing conditions
(Ref 17). Figure 12.3 shows a schematic of an
RDF unit. Carbon steels and sometimes lowalloy steels are typical construction materials
for waterwalls, screen tubes, the boiler bank,
and economizer. For superheaters, carbon and
low-alloy steels are typically the materials of
construction.
This chapter focuses mainly on waterwalls
and superheaters that have experienced severe
corrosion problems. Also discussed are the
latest protection methods against corrosion and
Refuse
feed
hopper
Boiler
Refuse
receiving area
Refuse
pit
Plunger ash
extractor
Ash removal
Fig. 12.1
Mass-burning unit, with a grate where the fuel burns and the superheater right above the arch (or bull nose) in the upper
furnace, followed by a boiler bank and an economizer in the convection pass downstream of the superheater. Source: Ref 3
pg 337
30/10/2007 4:10PM Plate # 0
pg 338
13
Drum
Radiant
pass
2ND
pass
SH.
2
3RD
pass
Evap.
SH.
1
Econ.
14
6
10
2
4
12
3
11
B
C
15
7
9
8
Fig. 12.2
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Refuse feed hopper

Refuse chute
Refuse incineration grate
Secondary air supply
Furnace
Auxiliary burner
Primary air hopper
Ash extractor
Scraper conveyor
Fly-ash hopper
Fly-ash conveyor
Boiler ash
Superheater steam outlet
Boiler feed water inlet
Primary air supply
Mass-burning unit of different design involving multiple passes for the flue gas stream before entering the superheaters. Also
shown is the grate where the fuel is combusted. Source: Ref 17
erosion/corrosion at these two critical areas in the

boiler.
12.3 Boiler Tube Fireside Corrosion

Problems
Prepared
fuel
storage
Boiler
Ash removal
system
Fig. 12.3
Refuse-derived fuel unit, with superheaters (right

above the nose arch or bull nose) and a boiler
(generating) bank in the convection pass. The bottom of the furnace is a traveling grate. Source: Ref 3
Materials problems and high wastage rates

experienced by furnace waterwalls and superheaters in numerous waste-to-energy plants
worldwide have been extensively reviewed by
Krause (Ref 1), Sorell (Ref 2), and Krause and
Wright (Ref 18). Many of the early boilers, particularly European boilers, suffered severe boiler
tube failure problems. A boiler (410 C, or
770 F, and 5 MPa, or 725 psig steam) at the Issy
plant in Paris suffered tube failures in the side
wall after only about 5000 h of operation, and the
same boiler suffered serious superheater corrosion problems from the start of the operation
(Ref 1). In a boiler (525 C, or 980 F, and
12 MPa, or 1800 psig steam) in Mannheim,
Germany, the furnace wall tubes failed only after
30/10/2007 4:10PM Plate # 0
3000 h of operation, and superheater tube failed

after only 2000 h of operation (Ref 1). The rst
U.S. WTE boiler with relatively low steam
temperature (320 C, or 610 F) in Nashville,
TN, suffered both waterwall and superheater
failure within the rst year of operation (Ref 18).
Many other tube failures have been cited in Ref 1,
2, 1821). All these initial boiler tubes materials
were carbon and low-alloy steels with no corrosion protection.
Modern WTE boilers have increasingly used
various corrosion protection methods to protect waterwalls, superheaters, and boiler banks
against corrosion and erosion/corrosion. Many
boilers with no corrosion protection for their
waterwalls and superheaters continue to suffer
premature failures. Figure 12.4 shows an example of a tube failure at the waterwall in one boiler
after only 8 months of operation (Ref 22). One
effective corrosion protection method for the
furnace waterwalls was developed in the 1985 to
1986 period by using alloy 625 overlay cladding,
which was applied on-site in an RDF boiler in
Lawrence, MA using an automatic gas metal arc
welding (GMAW) process (Ref 23). A total of
about 21,000 lb of alloy 625 weld overlay metal
was applied to this boiler. The performance of the
waterwall overlay of alloy 625 in this rst overlaid waterwall proved to be very successful
during the next 3 to 4 years of boiler operation.
Between 1989 and 1990, a total of about
260,000 lb of alloy 625 weld overlay metal was
applied to the waterwalls of 29 boilers (Ref 23).
Fig. 12.4
Alloy 625 overlay has been found to provide

dramatic reduction in metal loss in areas where
carbon or low-alloy steels have suffered unacceptable wastage rates (Ref 7, 16, 2225). For
mass-burning units, the waterwalls are typically
protected by alloy 625 weld overlay above the
refractory lining at the lower, high-radiant section of the boiler (Ref 9). The waterwalls in an
RDF boiler are generally fully protected with
alloy 625 weld overlay (Ref 9). Kubin (Ref 7)
indicated that all RDF boiler waterwalls for
Ogden plants were virtually fully covered with
alloy 625 weld overlay.
In many plants, alloy 625 as an outer tube
cladding for corrosion protection of superheaters
(with higher metal temperatures) has also proved
to provide signicant reduction in tube-wall
wastage rates compared to carbon and low-alloy
steels (Ref 10, 22, 24). Excellent performance
was observed for alloy 625 overlay superheater
tubes in a boiler in Europe, as shown in Fig. 12.5.
The cladding can either be in a form of weld
overlay (Ref 24) or coextruded tube cladding
(Ref 10). In some plants, however, alloy 625
cladding in superheaters was found to experience
high wastage rates (Ref 10, 12, 22, 24). This is
illustrated in Fig. 12.6 (Ref 10, 22). The wastage
rates ranged from less than 10 mpy to hundreds
of mpy. Superheater corrosion appears to be
highly dependent on the individual boiler.
Possible corrosion mechanisms involved in
WTE environments are discussed in the next
section.
Carbon steel waterwall suffering blown tubes due to high wastage rates resulting in significant tube-wall thinning after
8 months of service in a WTE boiler. Courtesy of Welding Services Inc.
pg 339
30/10/2007 4:10PM Plate # 0
12.4 Corrosion Mechanisms

As described in Section 12.3, the furnace
waterwalls and superheaters made of carbon or
low-alloy steels suffered severe wastage problems in numerous boilers, experiencing tube
failures after only thousands of hours of operation. Corrosion was the result of exposure to the
combustion products generated during the combustion of MSW. Typical gaseous components
generated during combustion are O2, CO2, H2O,
SO2, and HCl with HF in some units. In some
cases, local areas at the waterwall may be in
reducing conditions. In those local furnace wall
areas, there could be some CO present. Among
these gaseous components, HCl is considered to
be the most corrosive. In general, the level of HF,
Fig. 12.5
if present, is signicantly lower than that of HCl

(see Table 12.1). Thus, HF plays no signicant
role in corrosion of boiler tubes.
Carbon and low-alloy steels are very susceptible to corrosion by HCl at elevated temperatures. The level of HCl generated in combustion
of MSW fuel is typically in the range of 100 to
2000 ppm. The tube metal temperatures of the
waterwall, although depending on the operating
pressure of the boiler along with other factors, are
generally in the range of 260 to 315 C (500 to
600 F). The superheater tube metal temperatures are generally in the range of 370 to 540 C
(700 to 1000 F), depending on the steam temperature, along with other factors. A brief review
of the corrosion data in terms of the direct HCl
corrosion is discussed below to see whether HCl
Alloy 625 overlay superheater tubes (on 15Mo3 steel substrate) after 4.5 years of service in a superheater producing 405 C
(760 F)/42 bar (609 psi) superheated steam, showing no evidence of corrosion or erosion/corrosion. Source: Ref 24
Steam temperature, C
18
95
205
315
425
535
350
Series 1
Series 2
Wastage rate, mpy
300
250
200
150
100
50
0
0
200
400
600
800
1000
Steam temperature, F
Fig. 12.6
Wastage rates as a function of steam temperature for alloy 625 cladding in weld overlay tubes and coextruded tubes tested as
part of superheater tube bundles at various WTE boilers. Source: Ref 10, 22
pg 340
30/10/2007 4:10PM Plate # 0
pg 341
Table 12.2 Corrosion rates in terms of metal loss for iron- and nickel-base alloys at indicated
temperatures for 1008 h in N2-10O2-50ppmSO2-500ppm HCl
Metal loss
at 425 C (800 F)
Alloy
m/yr
Type 316
Type 347
Alloy 800HT
Alloy 825
Alloy 600
Alloy 625
0.07
0.09
0.07
0.02
0.04
0.02
mpy
0.003
0.004
0.003
0.0008
0.002
0.0008
at 480 C (900 F)
at 590 C (1100 F)
m/yr
mpy
m/yr
mpy
2.54
2.29
1.02
1.27
2.03
1.78
0.1
0.09
0.04
0.05
0.08
0.07
5.08
7.11
5.33
2.54
3.05
2.79
0.2
0.28
0.21
0.1
0.12
0.11
corrosion can be responsible for severe waterwall

and superheater corrosion.
Paul and Daniel (Ref 26) conducted laboratory
tests at 315 C (600 F) for 720 h in simulated
ue gas environments with one being an oxidizing atmosphere (N2-2.6O2-15.0CO2-11.9H2O1000ppmSO2-1000ppmHCl) and the other
a reducing atmosphere (N2-11.4CO2-7.0CO4.4H2-11.9H2O-1000ppmHCl). Specimens were
not coated with salts nor covered with ash deposits. Both environments, containing about
1000 ppm HCl, showed that the corrosion rates
for both carbon steel (SA178C) and Type 304
were less than 0.25 mm/yr (<10 mpy). Smith
and Ganesan (Ref 27) conducted extensive tests
in simulated ue gas environments with HCl concentrations varying from 500 ppm to 10% and
temperatures from 425 C (800 F) to 590 C
(1100 F). Corrosion rates were found to be quite
low for austenitic stainless steels and nickel-base
alloys at temperatures up to 590 C (1100 F)
in ue gas environment containing 500 ppm
HCl (Table 12.2). Even for the environment
containing 4% HCl, the nickel-base alloys (alloys
825, 600, and 625) continued to show low corrosion rates. However, austenitic stainless steels
exhibited high corrosion rates. This is illustrated
in Table 12.3. A level of 4% HCl was more than
an order of magnitude higher than what can be
expected in a refuse-red combustion environment, and the temperature of 590 C (1100 F)
was much higher than the superheater metal
temperature of current WTE boilers. Nevertheless, alloy 625 as either coextruded cladding
or weld overlay was observed to suffer wastage
rates of more than 1.3 mm/yr (>50 mpy) in some
aggressive refuse-red boilers (see Fig. 12.6).
Table 12.3 shows that alloys 625 suffered a
corrosion rate of about 1.53 mm/yr (60.4 mpy)
when HCl was increased to 10% in the environment. Additional data for nickel-base alloys
are shown in Table 12.4 (Ref 27). Devisme

loss for iron- and nickel-base alloys at 590 C
(1100 F) for 72 h in N2-9O2-12CO2100ppmSO2-4 and 10HCl
Metal loss
4% HCl
Alloy
Type 316
Type 347
Alloy 800HT
Alloy 825
Alloy 600
Alloy 625
10% HCl
m/yr
mpy
m/yr
mpy
914.4
1244.6
73.66
20.32
25.4
15.75
35.7
48.5
2.9
0.8
1.0
0.6
1066
1219
1549
42.0
47.5
60.4
et al. (Ref 28) conducted tests in Ar-5HCl

and Ar-5HCl-10O2 environments at 600 C
(1110 F), showing much higher corrosion rates
in an oxidizing environment. However, the test
tempered was much higher than the current
superheater metal temperature. Their test results
are summarized in Table 12.5. More corrosion
data in halogen and halide environments are
available in Chapter 6.
It is generally agreed that chlorine in the fuel is
an important cause of corrosion problems of
boiler tubes. However, from the previous discussion, the direct corrosion of HCl, a gaseous
component that formed when the chlorine in the
fuel is combusted in the boiler, is not likely to
be primarily responsible for the boiler tube corrosion. In Section 12.2, it was explained that
there are a number of heavy metals, such as lead
(Pb), zinc (Zn), cadmium (Cd), arsenic (As), and
tin (Sb), along with alkali metals (i.e., Na, K) that
are present in MSW. These metallic elements
can react with chlorine to form various chlorides that have very low melting points. Many
investigators (Ref 12, 14, 16, 20, 24, 29, 30) have
considered that corrosion attack by low melting
point chlorides is a principal mode of corrosion
for boiler tubes in WTE plants.
30/10/2007 4:10PM Plate # 0
pg 342

loss for iron- and nickel-base alloys at 590 C
(1100 F) for 72 h in N2-9O2-12CO2100ppmSO2-4HCl with 0% and 10% H2O
Metal loss
0% H2O
10% H2O
Alloy
m/yr
mpy
Type 316
Alloy 825
Alloy 600
Alloy 625
914.4
20.32
25.4
15.75
35.7
0.8
1.0
0.6
m/yr
mpy
1.85
0.07
3.81
0.15
8.85
0.35
No measurable corrosion

terms of metal loss after 500 h at 600 C (1110 F)
in the indicated test environments
Ar-5HCl
Ar-5HCl-10O2
Source: Ref 28
: Field test data

400 h, gas temp.: 735 C (1355 F)
0.30
: Laboratory test data

40-fold 100 h thickness loss
0.25
Melting point of deposits
on field test tube
301 C (570 F)
0.20
0.15
0.10
0.05
0
0

Environment
exhibited melting points from approximately 330

to 650 C (625 to 1200 F), as shown in Fig. 12.8
(Ref 32). The gure also indicates that larger
amounts of salt deposits melt at around 380 and
500 C (715 and 930 F). In the deposits collected from the 3000 h exposure tests, sulfur
(as SO3) as high as about 40% and chlorine as
high as 9%, along with Na, K, Pb, and Zn were
detected. The authors observed that the corrosion
attack increased with higher fused salt content
in the deposits. This is illustrated in Fig. 12.9
(Ref 32) for Type 347 and alloy 625 in the
corrosion probe tests. The gure shows the corrosion attack increases as the sum of the heats of
fusion (or the amount of fused salt) of the deposit
increases.
The salts in the deposits are mainly chlorides
and sulfates. Chlorides exhibit much lower
melting points than sulfates. Under the operating
conditions of WTE boilers, many metal chlorides
can be in a molten state in the deposits on
waterwall/screen tubes (evaporator tubes) and
superheaters. This is illustrated in Fig. 12.10,
showing many low melting point chlorides.
Approximate steam temperatures at the waterwall and superheater are shown for general
comparison (Ref 30).
In Fig. 12.9, it is shown that increasing the
amount of salts in the deposit could increase
Average corrosion thickness loss, mm/400 h
Kawahara and Kira (Ref 31) observed that

the corrosion rate of carbon steel changed to a
much higher rate above approximately 300 C
(150 F), which corresponded to the melting
point of the deposits collected from the tube
samples in eld testing. This is illustrated in
Fig. 12.7. They also performed laboratory
tests, which involved actual deposits collected
from the boiler to which ZnCl2 was added as 8%
ZnO. The laboratory data conformed well with
the eld data. The temperature at which carbon
steel suffered accelerated corrosion attack was
found to coincide with the melting point of the
boiler tube deposits. The data strongly suggest
that accelerated corrosion attack in the WTE
boiler combustion environment was caused by
the formation of molten phases in the tube
deposits.
Otsuka et al. (Ref 32) performed an extensive
analysis on the salt deposits collected from three
commercial WTE boilers (stoker type furnaces).
A total of 23 salt deposits were collected from
three boilers using corrosion probes at 550 C
(1020 F) (probe metal temperature) for exposure of 700 and 3000 h, respectively. Each collected deposit was analyzed for the heat of fusion
and the melting point by differential scanning
calorimeter measurements. The heat of fusion is
related to the amount of the fused salt in the
deposit, with higher heat of fusion indicating
a larger amount of fused salt. The deposits
C-276
600
601
214
35 (1.4)
120 (4.7)
50 (2.0)
140 (5.5)
90 (3.5)
160 (6.3)
30 (1.2)
55 (2.2)
200
250
300
350
400
Metal temperature, C
Fig. 12.7
Corrosion in terms of thickness loss for carbon steel

(SA178) as a function of the metal temperature in
field exposure tests. Also superimposed are data generated from
laboratory tests. Source: Ref 31
30/10/2007 4:10PM Plate # 0
pg 343
the corrosion attack on the boiler tube,

and Fig. 12.10 shows that there are many lowmelting salts, particularly metal chlorides, that
can become molten at the operating temperatures
of furnace walls and superheaters. The type of
35
30
Heat of fusion, J/g
25
20
15
10
chloride that forms on the deposit, as well as the

type of the corrosion products formed between
the chloride deposit and the underlying tube
metal, can signicantly affect the corrosion
attack. For example, some of the eutectic chloride salts involving FeCl3 can become molten
at 150 to 205 C (300 to 400 F) as shown in
Fig. 12.10. At higher metal temperatures, some
chloride salts exhibit high vapor pressures.
Figure 12.11 shows vapor pressures of some
alkali metal salts (KCl, NaCl, KOH, and NaOH)
and heavy metal chlorides (ZnCl2 and PbCl2)
(Ref 26). K, Na, Zn, and Pb are among the most
frequently detected elements in the tube deposits.
It is generally believed that vapor pressures of
about 104 atm are adequate to cause signicant
high-temperature corrosion attack. The temperature at which the vapor pressure of ZnCl2
is at 104 atm is about 350 C (660 F) (see
Chapter 6). At temperatures of 350 C (660 F)
5
0
300
400
500
600
PbCl2
700
Melting point temperature, C
900
(480)
Fig. 12.8
Melting point temperatures versus the heat of fusion

for the deposits collected at the corrosion probes
(with the probe metal temperature of about 550 C, or 1020 F) in
three commercial WTE boilers. The open data points are from the
outer portion of the deposit, and the solid data points are from the
inner portion of the deposit. Source: Ref 32
ZnSO4
Na2S2O7
800
(425)
PbCl2/FeCl2
KCl/PbCl2
NaCl/PbCl2
700
K2SO4Na2SO4ZnSO4 (370)
(3050%:1030%:4060%)
NaCl/FeCl2
KCl/FeCl2
600
(315)
1.5
K2S2O7
Max. waterwall
fluid temperature
Temperature F (C)
Maximum corrosion thickness loss, mm/3000 h
Max. superheater
steam temperature
NaCl/CaCl2
PbCl2/CaCl2
PbCl2/MgCl2
0.5
0
0
10
12
ZnCl2
500
(260)
400
(205)
NaCl/ZnCl2
SnCl2
ZnCl2/KCl
ZnCl2/FeCl3
PbCl2/FeCl3
300
(150)
SnCl2/ZnCl2
NaCl/SnCl2
KCl/SnCl2
NaCl/FeCl3
Total sum of heat of fusion up to 550 C, J/S
Fig. 12.9
Maximum thickness loss of corrosion probes

exposed at metal temperature 550 C (1020 F)
for 3000 h in WTE boilers as a function of the sum of heat of fusion
(corresponding to the amount of fused salt) of the deposits. The
open data points are Type 347H, and the solid data points are
alloy 625. Source: Ref 32
Fig. 12.10
Melting temperatures of various salts that are

likely to form in WTE boilers. The temperatures
of waterwall saturated fluid and superheater steam are also indicated for comparison. Source: Ref 30
30/10/2007 4:10PM Plate # 0
and higher, vapor-phase corrosion by ZnCl2 can

be signicant. Below this temperature, ZnCl2
condenses, thus causing its vapor pressure below
104 atm and making vapor-phase corrosion less
likely. Accordingly, waterwall corrosion is less
likely to be caused by vapor-phase corrosion
involving ZnCl2. The temperature at which the
vapor pressure of PbCl2 is at 104 atm is about
485 C (900 F) (see Chapter 6). If the vapor
phase of PbCl2 is involved in the corrosion
attack, it is likely to be for superheaters. Both
KCl and NaCl will be at much higher temperatures when these chlorides reach the vapor
pressure of 104 atm. Thus, both KCl and NaCl
are most likely to be involved in molten salt
corrosion not vapor corrosion.
Gleeson et al. (Ref 33) examined the effect of
ZnCl2 on the corrosion of several iron- and
nickel-base alloys using a modied Dean test
with a three-zone furnace arrangement. The test
involved a owing ue gas (N2-3.6O2-14CO20.25SO2) passing rst through a crucible
containing molten ZnCl2 salt at 425 to 540 C
(800 to 1000 F) to pick up ZnCl2 vapor
(1.3103 atm at 425 C, or 800 F, and
2.8102 at 540 C, or 1000 F), followed by the
second zone heated to about 815 C (1500 F)
with a catalyst to equilibrate the gas mixture, and
to the third zone where specimens were exposed
to a 540 C (1000 F) gas mixture containing
ZnCl2 vapor. Because the gas mixture drops in
temperature from 815 C (1500 F) in the
second zone to 540 and 510 C (1000 and

950 F) in the third zone where specimens were
tested, the salt deposition rate onto the test specimen was found to be about 0.07 mg/cm2 h of
molten ZnCl2. The test specimens were thus
coated with molten ZnCl2 during the testing.
Their test results are summarized in Table 12.6
(Ref 33). The test temperature of 510 C (950 F)
is very close to some superheaters with
steam temperatures of about 455 to 480 C (850
to 900 F). The corrosion rates of alloy 625 were
on the same order of magnitude as those of alloy
625 (as a weld overlay or coextruded cladding)
Table 12.6 Corrosion rates after testing for

190 h in N2-3.6O2-14CO2-0.25SO2 containing
ZnCl2 vapor, which condensed onto the test
specimens as molten ZnCl2 at a rate of
0.07 mg/cm2 h
Material
540 C (1000 F) front row
C-2000
625
C-22
214
230
G-30
825
HR-160
HR-120
556
Type 310
510 C (950 F) back row
2.97 (117)
2.11 (83)
2.18 (86)
2.87 (113)
4.17 (164), 4.42 (174)
3.15 (124)
3.45 (136)
5.26 (207)
3.63 (143)
1.83 (72)
3.07 (121)
2.9 (114)
1.42 (56)
4.24 (167)
4.72 (186)
2.18 (86)
8.94 (352)
7.11 (280)
7.77 (306)
Source: Ref 33
Temperature, C
200
400
600
800
1000
1200
1
ZnCI2
101
KOH
Condensation
occurs
Vapor pressure, atm
PbCI2
KCI
102
103
104
Species remains
as a vapor
NaOH
105
NaCI
106
600
800
1000
1200
1400
1600
Temperature, K
Fig. 12.11
Vapor pressures of some alkali and heavy metal salts as a function of temperatures. Source: Ref 26
pg 344
30/10/2007 4:10PM Plate # 0
and corrosion products. Wright et al. (Ref 16)

reported that the deposits (on a severely corroded
waterwall tube) were found to contain 10 to 30%
chloride, 33 to 47% Pb, 3 to 13% Zn, 3 to 14%
Na, and 2 to 11% K. Spiegel (Ref 12) analyzed
two overlay superheater tube samples that were
severely corroded in German boilers. One tube
sample was overlaid with Ni-16.7Cr-8Fe-4.6Si
alloy, while the other overlaid with alloy 625.
In both cases, the overlays were severely corroded. Analysis of the deposits in both cases
showed salt melts were sulfates in contact with
the ue gas and chlorides at the deposit/scale
phase boundary (Ref 12). The sulfate smelt
was a CaSO4-K2SO4-Na2SO4 system including
ZnSO4 and PbSO4. The chloride melt was KClZnCl2 and also NaCl.
Heavy metals, such as Zn and Pb, along with
Cl, have been detected in the corrosion front
in some very aggressive boilers, as shown in

Fig. 12.6. It is thus believed that the severe corrosion of the alloy 625 overlay (or cladding)
superheater tubes was most likely due to molten
ZnCl2 and/or molten PbCl2. Because of considerable scattering in the data, the test results
were not adequate for making an alloy ranking
among nickel-base alloys. X-ray diffraction
analysis of the scale and corrosion products was
performed on alloys C-22, 625, G-30, 825, HR120, and Type 310, showing mainly Cr2O3 and
NiCr2O4 for all alloys, with NiSO4 being detected for C-22, 625, 825, and Type 310 (Ref 33).
From analysis of severely corroded tube
samples obtained from operating boilers, it was
learned that Cl, Zn, Pb, Cd, Na, K, and S, along
with major alloying elements from the tube (e.g.,
Fe if the tube was steel, or Cr and Ni if alloy 625
overlay) were often detected in the deposit
60 m
Fig. 12.12
Scanning electron micrograph (backscattered electron image) showing the deposits and corrosion scales formed on a
carbon steel (SA178A) superheater tube suffering severe tube-wall wastage. Chemical compositions at different locations
were analyzed by energy-dispersive x-ray spectroscopy (EDX) analysis (trace elements not reported here):
1: 31% Ca, 29% Si, 14% Mg, 15% Fe, 9% S, and 2% Zn
5: 72% Fe, 6% Cl, 7% Zn, 4% S, 4% Na, and 2% K
2: 63% Fe, 16% Cl, 9% Zn, 4% Pb, and 2% S
6: 88% Fe, 2% Cl, 2% Zn, 2% Cd, and 1% Na
3: 20% Fe, 13% Cl, 3% Zn, 41% Pb, 11% S, 4% Na, 3% K, and 2% Ca
7: 76% Fe, 8% Cl, 5% Zn, 2% Cd, 2% Na, and 2% S
4: 67% Fe, 12% Cl, 7% Zn, 4% S, and 6% Na
pg 345
30/10/2007 4:10PM Plate # 0
of superheater tubes from operating boilers

(Ref 22). This is illustrated in Fig. 12.12 (Ref 22)
for a carbon steel superheater tube that suffered
severe corrosion attack. The steam temperature
and pressure of the superheater were reported
to be 400 C (750 F) and 4.5 MPa (625 psig),
respectively. The tube wall was reportedly reduced from the original 5.6 mm (0.220 in.) to
about 1.02 mm (0.040 in.) after 11 months of
service, with a wastage rate of approximately
5 mm/yr (196 mpy). Figure 12.12 shows the
deposit and corrosion products. The corrosion
front (location No. 5, 6, and 7 in Fig. 12.12) was
found to contain Cl, Zn, Cd, Na, S, and Fe. The
compounds are believed to contain iron chlorides
with Zn, Cd, and Na.
The composition of the deposit and corrosion

products can vary signicantly from plant to
plant. This is shown in another boiler described
in Fig. 12.13 and 12.14 (Ref 22), which involved
an alloy 625 overlay superheater tube (410 C, or
770 F, steam) after service for about 6.5 months.
Figure 12.13(a) shows the full cross section of
the weld overlay with slight surface pitting attack, and Fig. 12.13(b) shows one of the corrosion pits at a higher magnication. The chemical
compositions of the deposit and the corrosion
products, which were analyzed by SEM/EDX,
are shown in Fig. 12.14. Signicant amounts
of lead were found throughout the deposit
and corrosion products. It is also signicant
that Pb, Zn, and Cl, along with Cr, Ni, and
(a)
0.025 mm
(b)
Fig. 12.13
(a) Optical micrograph showing slight pitting corrosion attack on alloy 625 overlay in an alloy 625 spiral overlay
superheater tube after about 6.5 months of service. Micrograph (a) also shows the fusion boundary and substrate carbon
steel. (b) Higher-magnification view of one of the corrosion pits. SEM/EDX analysis on the corrosion products in one of the surface pits is
summarized in Fig. 12.14.
pg 346
30/10/2007 4:10PM Plate # 0
Mo (major alloying elements from the weld

overlay) were detected in the corrosion front,
as indicated in locations No. 10 and 12 in
Fig. 12.14.
Both the morphology and the corrosion front
that contained heavy metals, such as Zn and/or
Pb, as were observed in the above two cases,
suggest that the corrosion mechanism is uxing
by molten chloride salts.
In some cases, the corrosion morphology for a
weld overlay of a Ni-Cr-Mo alloy involves general metal wastage followed by internal corrosion
penetration along dendrites of the weld overlay. This internal dendrite corrosion attack (or
penetration) is in a way similar to internal oxidation attack or intergranular oxidation attack

following general oxidation (or general metal
wastage by oxidation) for wrought alloys. Similarly, internal corrosion attack (or intergranular
corrosion attack) also takes place in wrought
alloys under gaseous chloridation attack. Figure 12.15, shows general material wastage
followed by internal dendrite corrosion penetration in the overlay for an alloy 622 overlay
superheater tube after 225 days of exposure.
This weld overlay tube was still in excellent
condition, as shown in Fig. 12.16, showing the
entire cross section of this alloy 622 overlay (on
30 m
Fig. 12.14
Scanning electron micrograph (backscattered electron image) showing a localized corrosion pit on alloy 625 overlay of a
superheater tube. SEM examination using energy-dispersive x-ray (EDX) spectroscopy showed the chemical compositions
(wt%) at different phases, marked as No. 1 through No. 12. The chemical compositions (wt%) at different phases are:
1: 68% Pb, 11% Mo, 6% Cr, 3% Fe, 3% Ni, 4% S, 3% Cl, and trace elements
2: 63% Pb, 9% S, 6% Cl, 7% Cr, 5% Mo, 2% Fe, 3% Ni, 2% Na, and trace elements
3: 31% Cr, 24% Ni, 2% Fe, 27% Pb, 6% Zn, 5% S, 1% Cl, and trace elements
4: 32% Pb, 14% Cr, 8% Ni, 5% Fe, 10% Mo, 7% Zn, 9% Cl, 5% K, 3% Na, 3% S, and trace elements
5: 52% Pb, 11% K, 13% S, 7% Cl, 6% Mo, 4% Na, 4% Cr, 2% Ni, and trace elements
6: 34% Ni, 23% Cr, 3% Fe, 23% Pb, 7% Zn, 6% Mo, 2% Cl, and trace elements
7: 49% Pb, 12% K, 3% Na, 13% S, 7% Cl, 5% Cr, 6% Mo, 3% Ni, and trace elements
8: 30% Cr, 28% Pb, 13% Ni, 11% Zn, 7% Mo, 6% Cl, 2% K, and trace elements
9: 26% Pb, 22% Ni, 16% Cr, 11% Zn, 7% Cl, 5% Mo, 7% Na, 3% Fe, and trace elements
10: 27% Pb, 19% Cr, 14% Ni, 13% Zn, 9% Mo, 5% Cl, 4% Na, 2% K, and trace elements
11: 54% Ni, 24% Cr, 6% Pb, 5% S, 6% Fe, and trace elements
12: 27% Cr, 19% Ni, 19% Pb, 15% Zn, 9% Mo, 5% Cl, 2% K, 2% Fe, and trace elements
pg 347
30/10/2007 4:10PM Plate # 0
a carbon steel tube). The morphology of this

internal dendritic corrosion penetration is
believed to be the result of gaseous corrosion
reactions, but not by the molten salt uxing
mechanism. In the same exposure test, an alloy
625 overlay tube was installed at the next platen
on the same row as the alloy 622 overlay
tube discussed in Fig. 12.15 and 12.16. After
152 days of exposure, the tube was removed
for metallurgical examination. The alloy 625
overlay showed slight surface corrosion pitting
attack (similar to alloy 622 overlay), but
no internal dendritic corrosion attack, as illustrated in Fig. 12.17 (Ref 22). In another boiler
where alloy 625 overlay superheater tube
suffered severe corrosion attack, the overlay
showed only general metal wastage with no
internal dendritic corrosion attack (Ref 22).
Montgomery and Larsen (Ref 34) also found
that alloy 622 weld overlay, which was applied
to the waterwall (the rear wall of the boiler in
the rst pass) in a WTE boiler (Haderslev plant)
in Denmark, showed internal dendrite corrosion
penetration after 8000 h of exposure, while
alloy 625 weld overlay at a similar location
tested during the same time period showed
no internal dendrite corrosion penetration after
8000 h of exposure. Both overlays showed only
narrow pitting corrosion attack. Nevertheless,
Spiegel (Ref 12) observed similar internal dendrite corrosion attack on an alloy 625 overlay
superheater tube where alloy 625 overlay suffered severe corrosion attack. He found that
those internal dendrite corrosion phases were
chlorides of mainly iron, chromium, and nickel,
along with the corresponding oxides. He suggested that the molten chlorides reacted with
oxides (formed on the metal) to produce chlorine,
which then reacted with the metal to form metal
chlorides.
The internal dendrite corrosion penetration is
similar to internal corrosion attack within the
alloy matrix or internal grain-boundary corrosion
attack in wrought alloys under high-temperature
gaseous corrosion, such as oxidation and
chloridation. In high-temperature gaseous corrosion, the extent of internal corrosion attack
can vary from alloy to alloy. Some alloys may
exhibit more internal attack than others. It is
important to note, however, that the alloys with
less internal oxidation attack are not necessarily
more corrosion resistant. There is no clear
indication that the weld overlay with internal
dendrite corrosion penetration would be less
resistant in terms of the alloys overall corrosion
attack. It is believed that the corrosion rate due
to general metal wastage by molten salt uxing
is the rate-controlling factor.
0.025 mm
Fig. 12.15
Internal dendrite corrosion attack that followed general wastage was observed in the overlay of an alloy 622 weld overlay
in an alloy 622 overlay superheater tube. Lower-magnification micrographs showing the through-thickness overlay is
shown in Fig. 12.16(a) and general surface corrosion morphology is shown in Fig. 12.16(b).
pg 348
30/10/2007 4:10PM Plate # 0
12.5 Corrosion Protection for

Furnace Waterwalls
In mass-burning units, the lower part of the
waterwall is generally protected by refractories
(e.g., SiC) (Ref 3, 7, 8, 16). The refractories
provide protection for the waterwall against
mechanical damage from abrasion of sliding
large articles on the moving grate and also
against corrosion from the combustion products.
Relatively lengthy installation and upkeep time
as well as reducing the heat-absorbing surface of
the waterwalls can be an issue for refractory
linings (Ref 20). Nevertheless, Licata et al.
(Ref 17) indicated that refractory linings are
needed to provide thermal insulation for ensuring that the designed combustion ue gas
temperature is reached. For corrosion protection

of the waterwall above the refractory, the current
prevailing method is the application of alloy
625 weld overlay using automatic gas metal
arc welding (GMAW) on site for existing boilers.
The waterwall can also be constructed with shopfabricated weld overlay panels with alloy 625
overlay or with coextruded composite tubes with
alloy 625 cladding.
In RDF units, the same refractory design used
for mass-burning units was tried initially by a
major boiler designer resulting in slagging problems for the refractory wall surface (Ref 3).
This was caused by the insulation of the lower
furnace waterwalls by the refractory that resulted
in higher ame temperatures and caused signicant refractory wall slagging (Ref 3).
(a)
0.010 mm
(b)
Fig. 12.16
Alloy 622 overlay superheater tube after 225 days of exposure in a boiler. (a) Cross section of the overlay showing slight
pitting attack. Micrograph (a) also shows the fusion boundary and substrate carbon steel. (b) Higher-magnification
micrograph showing surface corrosion morphology with general metal wastage and internal dendrite corrosion penetration
pg 349
30/10/2007 4:10PM Plate # 0
Following rapid corrosion attack of the bare

carbon steel waterwalls in an RDF unit in Lawrence, MA, the lower furnace waterwall was weld
overlaid with Inconel material in 1986, and this
overlay protection proved to be very effective
(Ref 3). The overlay alloy applied in the Lawrence unit was alloy 625 (Ref 23). The furnace
waterwalls of modern RDF units are protected by
alloy 625 cladding either as weld overlay applied
on-site, or shop-applied panels, or as coextruded
composite tubes with no refractory. Kubin (Ref
7) indicated in 1990 that the waterwalls of all
Ogden RDF boilers were virtually fully covered
with alloy 625 weld overlay.
Figure 12.18 shows a general view of an
automatic GMAW-applied alloy 625 weld overlay on the waterwall in a boiler. Alloy 625 weld
overlay on the waterwalls of the boiler was

found to perform well in both mass-burning and
RDF units. Generally, the wastage rates of alloy
625 overlay on waterwalls in U.S. boilers were
quite low, typically approximately 125 m/yr
(5 mpy) (Ref 22, 25). The boiler tube metal
temperatures are in a range of 260 to 315 C
(500 to 600 F) for most U.S. boilers. For some
European boilers running higher water/steam
pressures, the waterwall tube metal temperatures
could be higher.
Figure 12.19 shows the cross section of an
overlaid waterwall tube sample obtained from a
RDF boiler in Lawrence, MA after 16 years of
service. The metallographic cross section of the
alloy 625 weld overlay of that overlay waterwall
tube is shown in Fig. 12.20, revealing the full
(a)
0.010 mm
(b)
Fig. 12.17
Alloy 625 overlay superheater tube after 152 days of exposure in the superheater platen next to the alloy 622 superheater
tube, which is shown in Fig. 12.16. (a) Cross section of the overlay showing slight pitting attack. Micrograph (a) also shows
the fusion boundary and substrate steel. (b) Higher-magnification micrograph showing surface corrosion morphology with general metal
wastage and no internal dendrite corrosion penetration.
pg 350
30/10/2007 4:10PM Plate # 0
cross section of the overlay with little evidence of

corrosion attack. Figure 12.21 shows a closeup view of alloy 625 overlaid waterwall after
about 10 years of service in another RDF boiler.
The common mode of corrosion attack on alloy
625 weld overlay on the waterwall has been
found to be pitting attack. This is illustrated in
Fig. 12.22. When corrosion or pitting attack
becomes extensive, the corroded waterwall area
can be repaired by rst grinding off the corroded
metal prior to performing overlay welding, as
shown in Fig. 12.23.
It has been reported by Vrchota (Ref 25) that
the bare carbon steel waterwall above the weld
overlaid waterwall had experienced higher
wastage rates as the boiler continued to increase
its service duration. As a result, the weld overlay
area has been creeping upward gradually,
leading to application of weld overlay at
increasingly higher elevation in the boiler. This
situation has also been experienced in other
plants (Ref 22). However, there is no consensus
on a technical explanation about this phenomenon.
Thermal sprayed coatings have not yet been
used on a large scale in WTE boilers. One of
the major issues is related to the intrinsic characteristics of the sprayed coating in terms of
interconnecting pores that allow the corrosive to
permeate through the coating and cause corrosion attack at the coating/substrate interface, thus
leading to spallation. Another issue is lack
of automatic application system that can cover a
large waterwall area in achieving a consistent
quality over the large coating area. In the 1990s,
some tests on sprayed coatings were performed
in boilers without satisfactory results (Ref 7, 16,
35). DeVincentis et al. (Ref 35) conducted corrosion probe tests on sprayed coatings of three
nickel-base alloys (Ni-Cr-Co-Si alloy HF-160,
Fig. 12.19
Cross section of an overlaid waterwall tube

showing alloy 625 overlay after 16 years of
service in a RDF unit in Lawrence, MA. Courtesy of Welding
Services Inc.
Fig. 12.20
Fig. 12.18
Services Inc.
General view of an overlaid waterwall with alloy

625 overlay in a WTE boiler. Courtesy of Welding

the alloy 625 overlay of an overlaid waterwall
sample (Fig. 12.19) obtained from a RDF boiler (Lawrence, MA)
after 16 years of service. Micrograph also shows the fusion
boundary and substrate carbon steel. Courtesy of Welding Services Inc.
pg 351
30/10/2007 4:10PM Plate # 0
Ni-Cr-Fe-Al-Y alloy 214, and Ni-Cr-Mo alloy C22). The corrosion probe tests were conducted in
the rst pass of the front wall in a boiler at
Hempstead plant for 76 days of exposure with the
metal temperature maintained between 180 and
230 C (350 and 450 F). Coatings were applied
by the inert gas electric arc spraying method.
Disbonding of the coating during exposure was
found to be a major problem. Alloy C-22 and 214
coatings were found to suffer disbonding after
exposure. Kubin (Ref 7) indicated that a sprayed
coating of alloy 50Ni-50Cr was tested in the
upper furnace of a mass-burning unit and failed
after 2.2 years of service. Furthermore, the
removal of all the coating material in preparation
for overlay welding with alloy 625 later was
Fig. 12.21
Close-up view of alloy 625 overlay on the

waterwall of another RDF boiler after 10 years of
service. Shown in the photograph are two tubes and a membrane.
found to be an arduous and time-consuming

operation (Ref 7).
12.6 Corrosion Protection for

Superheaters
Metal temperatures for superheater tubes can
be 370 to 480 C (700 to 900 F) or higher. In
this temperature range, there will be more
chloride salts that become molten, as shown in
Fig. 12.10. This may also result in a higher
concentration of molten chloride salts in the ash
deposits forming on the tube surface. Presence of
more chloride salts will result in more corrosion
attack (Fig. 12.9). At higher temperatures, some
of the heavy metal chlorides, such as ZnCl2 and
PbCl2, exhibit higher vapor pressures, as shown
in Fig. 12.11. This can make the environment
more corrosive for superheater tube materials. As
discussed in Section 12.3, premature failures of
superheater tubes made of carbon or low-alloy
steels were quite common. Figure 12.24 shows
the cross section of a carbon steel (SA178A)
superheater tube removed after 11 months of
service in a mass-burning unit (Ref 22). The
superheated steam temperature and pressure for
this boiler were 400 C (750 F) and 4.5 MPa
(625 psig), respectively.
Blough et al. (Ref 36) conducted corrosion
probe tests in a mass-burning unit at Charleston
Resource Recovery. A wide variety of commercial alloys were tested, including carbon and
low-alloy steels, austenitic stainless steels, FeNi-Cr alloys, and nickel-base alloys. Tests were
Fig. 12.23
Fig. 12.22
Pitting attack is the common mode of corrosion of

alloy 625 weld overlay on the waterwall. Shown
in the photograph are three tubes and three membranes.
Repair welding can be performed on the corroded

waterwall overlay by grinding followed by overlay welding. An alloy 625 weld overlay on the waterwall after
10 years of service in a RDF unit. Shown in the photograph are two
tubes and three membranes.
pg 352
30/10/2007 4:10PM Plate # 0
pg 353
Fig. 12.24
Carbon steel (SA178A) superheater tube after 11

months of service in a mass-burning unit. The
superheated steam temperature and pressure were 400 C
(750 F) and 4.5 MPa (625 psig), respectively. Courtesy of
conducted for 4492 h. The results are summarized in Fig. 12.25 (Ref 36). Alloy 625 was found
to be the best at all test temperatures. Austenitic
stainless steels and Fe-Ni-Cr alloys and some
nickel-base alloys except alloy 625 form a big
scattering band between alloy 625 and ferritic
steels. In examining the wastage proles of the
probe sample cross sections, some alloys showed
the maximum wastage at the 90 position, where
the ue gas stream impinged upon, while others
showed the maximum wastage at about 45 from
either side from the 90 position. In another
eld test conducted in an RDF unit at Elk River
Station, Blough et al. (Ref 37) reported that
alloy 625 was found to perform signicantly
better than T-22, chromized T-22, Type 304H,
HR3C, and 825 at both 470 and 500 C (880
and 935 F). Tests were conducted with tube
samples of different alloys welded together as
part of the lead superheater tubes after about
1180 h of exposure. The results are summarized
in Fig. 12.26 and 12.27. The chromized layer was
Temperature, C
205
260
315
370
425
480
535
0.160
Ferrite band
Ferritic steels
0.140
Austenitic steels
625
0.120
825
0.100
600
Max. wastage, in.
800H
Austenite
band
HR160
0.080
0.060
0.040
625
0.020
0.000
400
450
500
550
600
650
700
750
800
850
900
950
1000
Temperature, F
Fig. 12.25
Results of corrosion probe tests for various alloys in a boiler at Charleston Resource Recovery. The corrosion probes were
installed in the convection path with the exposure time of 4492 h. Ferritic steels included carbon steel, T-22, and T-91.
Austenitic steels included Type 304, 347, 310, and 27Cr-31Ni-3.5Mo (N08028). Source: Ref 36
30/10/2007 4:11PM Plate # 0
pg 354
cladding in coextruded composite tubes has also

been widely used in WTE boilers (Ref 10).
For superheaters, the corrosiveness of the
environment varies greatly from boiler to boiler.
In some boilers, alloy 625 has performed well
as an overlay or cladding in superheaters. One
example is shown in Fig. 12.5. In some other
boilers, the environment was so corrosive that
alloy 625 cladding lasted for only about 1 year.
For example, in boilers with superheated steam
temperatures between 400 and 455 C (750 and
850 F), the alloy 625 cladding was found to
exhibit a wide range of wastage rates from a
negligible rate to 7.5 mm/yr (300 mpy), as
shown in Fig. 12.6.
The factors that affect the corrosiveness of the
environment can be very complex. In his corrosion probe testing in an operating boiler, Krause
(Ref 29) found that the ue gas temperature
could be important in affecting the corrosion rate
of carbon steel. His data are shown in Fig. 12.28.
The gure shows that when the ue gas
2
80
315 wastage
60
1.5
40
20
0.5
Corrosion, mm
45 wastage
Corrosion, mils
found to be consumed in large areas and thus

provided no protection. Kubin (Ref 7) had tested
chromized coatings for superheater applications
with mixed results, showing some success at one
facility, but not at another facility. Furthermore,
chromizing was not found to be cost effective (Ref 7). Other diffusion coatings, such as
aluminizing and aluminum-silicon codiffusion
coatings, were tested and found to be inadequate
in performance (Ref 7).
Alloy 625 as a weld overlay in spiral overlay
tubing or a cladding in coextruded tubing offers a
viable solution to the superheater corrosion problems. Table 12.7 summarizes the comparative
performance between a bare carbon steel tube
and an alloy 625 overlay tube in a nishing
superheater in a side-by-side eld test in a boiler.
The wastage rate was found to be about 2.8 mm/
yr (110 mpy) for carbon steel and 0.46 mm/yr
(18.3 mpy) for alloy 625 overlay. Based on these
data, the expected replacement interval for carbon steel tubes and alloy 625 overlay tubes are
1.4 and 5.8 years, respectively (Table 12.7).
Figure 12.5 shows alloy 625 overlay superheater
tubes (405 C, or 760 F, and 42 bar, or 609 psi,
superheated steam) exhibiting excellent overlay
condition with no sign of corrosion or erosion/
corrosion after 4.5 years of service in a boiler in
the Netherlands (Ref 24). Alloy 625 overlay
tubing has been widely used for superheater
applications in WTE boilers. Alloy 625 as a
10
400
45 wastage
315 wastage
8
300
0
304
HR-3C T-22Cr 825
T-22
625
200
4
100
Corrosion, mm
Corrosion, mils
Alloys
6
Fig. 12.27
Tube metal loss of various alloys tested at 470 C

(880 F) for 1180 h in an RDF unit at Elk River
Station. T-22CR represents the chromized T-22 tube sample. The
chromized layer (on T-22) was found to be consumed in a large
area. Source: Ref 37
304
HR-3C T-22CR
825
T-22
625
Table 12.7 Wastage rates for the carbon

steel tube and the alloy 625 overlay tube in a
side-by-side test as part of a finishing superheater
in a boiler in New York
Alloys
Fig. 12.26
Tube metal loss of various alloys tested at 500 C

(935 F) for 1180 h in an RDF unit at Elk River
Station. T-22CR represents the chromized T-22 tube sample. The
chromized layer (on T-22) was found to be consumed in a large
area. Source: Ref 37
Wastage rate
Tube construction
Carbon steel
625 overlay
mm/yr
mpy
Expected tube life, years
2.8
0.46
110
18.3
1.4
5.8(a)
(a) Overlay (0.080 thick) + steel tube. Source: Ref 24
30/10/2007 4:11PM Plate # 0
temperature increased from 760 to 845 C (1400

to 1550 F), the corrosion rate of carbon steel
was found to accelerate as the metal temperature
exceeded 430 C (800 F). The author (Ref 29)
believed that the increased corrosion rate was the
result of the formation of volatile FeCl3, since
below 430 C (800 F) the corrosion product
was primarily FeCl2, which would not volatilize
at these temperatures. This ue gas temperature
increase is not likely to signicantly affect the
corrosion rate of alloy 625, which is a nickel-base
alloy and not likely to form either FeCl2 or FeCl3.
Metal temperature can signicantly affect the
corrosion rate of the alloy. The corrosion reaction
is a thermally activated process, thus increasing
metal temperature can result in a higher corrosion
rate. Furthermore, increasing metal temperature
can increase the range of various chloride salts
that become molten (Fig. 12.10). Increased
amounts of molten chloride salts can also increase corrosion rates. Furthermore, increasing
temperature can increase vapor pressures of
chloride salts, thus resulting in increased corrosion by chloride vapors.
The concentration of the chloride deposits and
the type of chlorides can vary from boiler to
boiler. Both the concentration of chloride salts
and the type of chlorides are not normally monitored in plants. Furthermore, ue gas velocity
can be an important factor in the superheater
wastage. In some cases, the superheater wastage
can be the result of erosion/corrosion. The only
way to judge the severity of the superheater

wastage issue in a particular boiler is to examine
the historical data of that particular boiler. For
aggressive boilers, efforts have been underway
in the industry to identify an alloy that can outperform alloy 625. The results, however, have
not been encouraging thus far. The ndings of
some of the comparison tests are summarized in
the next paragraph.
In a side-by-side eld test on alloys 625 and
622 overlay tubes (ve-tube platen panel each)
(a)
Metal temperature, C
200
0.6
600
760 C (1400 F) gas temperature

855 C (1550 F) gas temperature
0.5
13
11
0.4
9
0.3
0.2
Corrosion rate, m/h
Corrosion rate, mils/h
400
0.1
(b)
1
0
200
400
600
800
1000
1200
Metal temperature, F
Fig. 12.28
Effect of flue gas temperature on the corrosion rate

of carbon steel in short-time corrosion probe tests
(10 h exposure) in an operating boiler. Source: Ref 29
Fig. 12.29
Alloys 72 overlay superheater tube (a) and alloy

C276 overlay superheater tube (b) after 7200
operating hours (10 months) in a RDF unit. The windward side of
the tube, where the flue gas impinged upon the tube surface was
the top side of the tube cross section as shown in the figure. The
tube cross section was polished and etched with nital to reveal
alloy 625 overlay (white portion of the metal) which was not
etched by nital. Courtesy of Welding Services Inc.
pg 355
30/10/2007 4:11PM Plate # 0
with the steam temperature of about 340 C

(640 F) inside the tube and about 815 C
(1500 F) ue gas temperature, alloy 625 overlay tubes were exposed for 152 days and alloy
622 tubes for 225 days. The wastage rate was
estimated to be about 175 m/yr (7 mpy) for
alloy 625 and about 355 m/yr (14 mpy) for
alloy 622. Alloy 622 containing 13% Mo and 3%
W was found to be not as good as alloy 625 with
about 9% Mo and no W but about 3.5% Nb. Both
alloys contain about 21 to 22% Cr. Overlay tubes
made of C-276 (Ni-16Cr-16Mo) and alloy 72
(Ni-44Cr) were also tested in an RDF unit, where
alloy 625 overlay lasted for about 11 months.
Both overlays were found to be not as good as
alloy 625 overlay. The overlays of both alloys
were consumed after about 10 months as shown
in Fig. 12.29 (Ref 22). Alloy 686 (Ni-21Cr16Mo-4W) was compared with alloy 625 in a
side-by-side test as overlays applied to the
waterwall (72 bar and 290 C, or 555 F) of a
boiler in Denmark (Ref 34). Evaluation of the
overlay waterwall samples removed from the
boiler after 8000 h of exposure showed both 625
and 686 overlays exhibited shallow pitting with
pitting depth of about 50 m.
An alternate protection method for superheaters is the use of tube shields, which are
typically made of Type 309, 310, and 253MA.
Figure 12.30 shows carbon steel superheater
tubes protected by metallic tube shields awaiting
installation at a WTE plant. The shields are
typically attached to the tube by mechanical
straps and clamps with llet welds. The tube
shields generally do not receive adequate heat
transfer through the air gap between the shield
Fig. 12.30
and the tube. As a result, the shields can experience temperatures as high as those of ue gas
streams. The shields can thus suffer warping,
distortion, and creep damage in addition to hightemperature corrosion. Accumulation of ash/salt
deposits in the crevice behind the shields can
further accelerate the corrosion attack. Loosened
or fallen shields can cause problems by impeding
the gas ow. Tube shields are generally considered to be a sacricial part and are replaced
regularly during the plant maintenance shutdown. Tube shields are sometimes made of cast
stainless steels. A cast tube shield can be made
much thicker than a wrought alloy sheet, thus
enabling the tube shield to last until the next
annual maintenance shutdown. Vrchota (Ref 25)
reported that a 7.5 mm (0.3 in.) thick cast tube
shield made of HD stainless steel (Fe-27Cr-5Ni)
lasted for 12 months, which coincided with the
maintenance shutdown cycle in an RDF boiler.
To improve the heat transfer of the tube shield,
silicon carbide cement was used to ll the gap
between the shield and the superheater tube. This
allows for the reinstallation of new tube shields
every 12 months.
12.7 Summary
Materials issues related to waste-to-energy
boilers for burning municipal solid waste (MSW)
for electricity generation are presented. Combustion of MSW generates a very hostile environment for waterwall tubes and superheater
tubes. The wastage rates of waterwall tubes made
Carbon steel superheater tubes protected by metallic tube shields awaiting installation at one WTE plant.
pg 356
30/10/2007 4:11PM Plate # 0
of carbon or low-alloy steels have been found to

be unacceptably high if no protection method is
used. The corrosion is believed to result from
molten chloride salts. The current, widely used
method for protecting the waterwalls is the use of
alloy 625 overlay cladding applied by automatic
gas metal arc welding process. The waterwalls
can also be constructed out of alloy 625/carbon
steel coextruded tubes. Superheater tubes also
require some methods of protection against corrosion attack. Alloy 625 overlay tubes and
coextruded tubes have been used for superheaters successfully in many boilers. However,
for some boilers with higher steam temperatures
and/or more corrosive environments, alloy 625
overlay or cladding was found to be inadequate.
An alternate corrosion protection method is the
use of metallic tube shields or refractories. Tube
shields are generally considered to be a sacricial part and are replaced regularly during
the plant maintenance shutdown. Refractories
require regular repair or replacement.
9.
10.
11.
12.
13.
14.
REFERENCES
1. H.H. Krause, Historical Perspective of

Fireside Corrosion Problems in RefuseFired Boilers, Paper No. 200, Corrosion/93,
NACE, 1993
2. G. Sorell, The Role of Chlorine in High
Temperature Corrosion in Waste-To-Energy
Plants, Mater. High Temp., Vol 14 (No. 2/3),
1997, p 137
Wilcox, 1992
4. C.F. Knights, I.W. Cavell, and B.A. Phillips,
Corrosion During Incineration of a Sulfur
and Chlorine Bearing Mixture of Rubbers
and Plastics, Werkst. Korros., Vol 40, 1989,
p 163
5. E.A. Bretz, Energy from Wastes, Power,
March 1990, p S-1
6. P. Rademakers, W. Hesseling, L.A. Tange,
and R. Montaigne, Review on Corrosion in
Waste Incinerators, and Possible Effect of
Bromine, CEF-12, Laan van Westenenk,
The Netherlands, 2002
7. P.Z. Kubin, Materials Performance and
Corrosion Control in Modern Waste-ToEnergy Boilers Applications and Experience, Paper No. 90, Corrosion/99, NACE
International, 1999
8. L. Strach and D.T. Wasyluk, Experience
with Silicon-Carbide Tiles in Mass-Fired
15.
16.
17.
18.
19.
20.
21.
Refuse Boilers, Paper No. 219, Corrosion/

93, NACE, 1993
R.L. Anderson, Wheelabrator Technologies
Inc., private communication, 2006
A. Wilson, U. Forsberg, M. Lundberg, and
L. Nylof, Composite Tubes in Waste Incineration Boilers, Stainless Steel World 99
Conference on Corrosion-Resistant Alloys
(Conf. Proc.), Book 2, KCl Publishing BV,
The Netherlands, 1999, p 669
F. Soutrel, C. Rapin, P. Steinmetz, and G.
Pierotti, Corrosion of Fe, Ni, Cr and Their
Alloys in Simulated Municipal Waste
Incineration Conditions, Paper No. 428,
M. Spiegel, Salt Melt Induced Corrosion of
Metallic Materials in Waste Incineration
Plants, Mater. Corros., Vol 50, 1999, p 373
M. Noguchi et al., Experience of Superheater Tubes in Municipal Waste Incineration Plant, Mater. Corros., Vol 51, 2000,
p 774
P.L. Daniel, L.D. Paul, and J. Barna, FireSide Corrosion in Refuse-Fired Boilers,
Mater. Perform., May 1988, p 20
J.G. Singer, Ed., Combustion Fossil Power,
4th ed., Combustion Engineering, Inc.,
Windsor, CT, 1991
I.G. Wright, H.H. Krause, and R.B. Dooley,
A Review of Materials Problems and Solutions in U.S. Waste-Fired Steam Boilers,
A.J. Licata, L.A. Terracciano, R.W. Herbert,
and U. Kaiser, Design Features for Superheater Corrosion Control in Municipal
Waste Combustors, Materials Performance
in Waste Incineration Systems, G.Y. Lai and
G. Sorell, Ed., NACE, 1992, p 5-1
H.H. Krause and I.G. Wright, Boiler Tube
Failures in Municipal Waste-To-Energy
Plants: Case Histories, Paper No. 561, Corrosion/95, 1995
A. Pourbaix, Corrosion of a Waste Incinerator: Effects of Design and Operating
Conditions, Werkst. Korros., Vol 40, 1989,
p 157
A.L. Plumley, W.R. Roczniak, and E.C.
Lewis, Materials Performance of Heat
Transfer Surfaces in A MSW-Fired Incinerator, Materials Performance in Waste
Incineration Systems, G.Y. Lai and G.
Sorell, Ed., NACE, 1992, p 7-1
W.G. Schuetzenduebel, I.E. Johnson, and
C.W. Clemons, Accelerated Tube Metal
pg 357
30/10/2007 4:11PM Plate # 0
22.
23.
24.
25.
26.
27.
28.
29.
Wastage in Municipal Solid Waste Fired

Furnaces, Materials Performance in Waste
Incineration Systems, G.Y. Lai and G.
Sorell, Ed., NACE, 1992, p 10-1
Welding Services Inc., unpublished data,
Norcross, Georgia
P.N. Hulsizer, Problems and Solutions
in Applying Weld Overlay to Waste Boiler
Incinerators, Materials Performance in
Waste Incineration Systems, G.Y. Lai and
G. Sorell, Ed., NACE, 1992, p 11-1
G.Y. Lai, Corrosion Mechanisms and Alloy
Performance in Waste-To-Energy Boiler
Combustion Environments, presented at
the 12th North American Waste To Energy
Conference (NAWTEC 12) (Savannah, GA),
May 1719, 2004
S. Vrchota, Fireside Corrosion Management
in RDF Waste-To-Energy Boilers, Paper No.
2005
L.D. Paul and P.L. Daniel, Corrosion
Mechanisms in Oxidizing, Reducing, and
Alternating Combustion Gases in RefuseFired Boiler Environments, Paper No. 216,
G.D. Smith and P. Ganesan, Metallic
Corrosion in Waste Incineration: A Look
at Selected Environmental and Alloy
Fundamentals, Heat-Resistant Materials
II (Conf. Proc.), Second International Conference on Heat-Resistant Materials, K.
Natesan, P. Ganesan, and G. Lai, Ed., ASM
F. Devisme, P. Falgoux, F. Lefebvre, and
T. Flament, High Temperature Corrosion
in Atmospheres Containing Hydrogen
Chloride, presented at the 11th International Incineration Conference (Albuquerque, NM), May 1115, 1992
H.H. Krause, Chlorine Corrosion in Waste
Incineration, Materials Performance in
30.
31.
32.
33.
34.
35.
36.
37.
Waste Incineration Systems, G.Y. Lai and

G. Sorell, Ed., NACE, 1992, p 11
I.G. Wright, V. Nagarajan, and H.H. Krause,
Y. Kawahara and M. Kira, Corrosion Prevention of Waterwall Tube by Field Metal
Spraying in Municipal Waste Incineration
Plants, Corrosion, Vol 53 (No. 3), 1997,
p 241
N. Otsuka, Y. Tsukaue, K. Nakagawa, Y.
Kawahara, and K. Yukawa, A Corrosion
Mechanism for the Fireside Wastage of
Superheater Materials in Waste Incinerators,
B. Gleeson, J.E. Barnes, and M.A. Harper,
Corrosion Behavior of Various Commercial
Alloys in a Simulated Combustion Environment Containing ZnCl2, Paper No. 196,
M. Montgomery and O.H. Larsen, Field
Investigation of Various Weld Overlays
in a Waste Incineration Plant, Paper No.
2005
D.M. DeVincentis, S.P. Goff, J.W. Slusser,
Z. Zurecki, and J.T. Rooney, Solving Fireside Corrosion in MSW Incinerators with
Thermal Spray Coatings, Paper No. 198,
Corrosion/93, 1993
J.L. Blough, G.J. Stanko, and M.T. Krawchuk, In Situ Materials Testing in a WasteTo-Energy Power Plant, Mater. High Temp.,
Vol 14 (No. 2/3), 1997, p 181
J.L. Blough, G.J. Stanko, W.T. Bakker, and
T. Steinbeck, Superheater Corrosion in a
Boiler Fired with Refuse-Derived Fuel,
Heat-Resistant Materials II (Conf. Proc.),
Second International Conference on HeatResistant Materials, K. Natesan, P. Ganesan,
and G. Lai, Ed., ASM International, 1995,
p 645
pg 358

DOI: 10.1361/hcma2007p359
31/10/2007 12:47PM Plate # 0

CHAPTER 13
Black Liquor Recovery Boilers

in the Pulp and Paper Industry
13.1 Introduction
The black liquor recovery boiler is an
integral part of the kraft pulping process in
the pulp and paper industry. In the kraft process,
wood chips are cooked in an aqueous solution,
known as white liquor, containing primarily
sodium hydroxide (NaOH) and sodium sulde
(Na2S) under pressure in a digester for manufacturing pulp. After cooking, the pulp is
separated from residual liquor for paper manufacturing. The residual liquor, which contains
wood lignins, organic materials, and inorganic
compounds, then undergoes a series of evaporation processes to increase the concentration of
solids from approximately 15 to 70 wt% or
higher. This liquor with high concentrations of
solid is known as black liquor. The black liquor
is burned in a boiler to produce molten inorganic
material, known as molten smelt, which is rich in
sodium carbonate (Na2CO3) and sodium sulde
(Na2S). The molten smelt forms in a bed on the
furnace oor and is discharged from the furnace
oor into a water tank outside the furnace to form
green liquor containing primarily sodium carbonate (Na2CO3) and sodium sulde (Na2S).
This green liquor undergoes the causticizing
reaction to convert it to white liquor consisting of
sodium hydroxide (NaOH) and sodium sulde
(Na2S) for use in cooking wood chips to manufacture pulp again. This chemical recovery process in the kraft pulping is described in detail in
Ref 1 and 2. Because of its function of recovering
cooking chemicals, the boiler is often referred to
as a chemical recovery boiler or simply recovery boiler. In addition to recovering cooking
chemicals for the pulping process, the boiler also
recovers energy by producing process steam.
Grant (Ref 3) reported in 1997 that about 216 recovery boilers which were operated by 130 mills
in the United States produced approximately
41% of the total energy used by these mills.
The boiler construction is similar to that of

coal-red boilers (Chapter 10), oil-red boilers
(Chapter 11), and waste-to-energy (WTE) boilers
(Chapter 12) using the tube-membrane waterwalls for the furnace enclosure, with the exception that the furnace oor is also enclosed with a
tube-membrane construction. However, the
combustion characteristics and problems are
uniquely different from those of fossil-red boilers and WTE boilers. The rst and most signicant issue is the formation of molten smelt (or
molten salts) on the furnace oor, and thus a
boiler tube leak (under certain conditions) may
lead to a smelt-water explosion (Ref 4). Recovery boilers have the potential for devastating
explosions causing fatalities and injuries (Ref 4).
Combustion of the black liquor in the boiler
generates a sulfur-containing, reducing furnace
environment in the lower furnace and a pile of
char bed with molten smelt (or molten salts) on
the oor tube bed. The major materials problems
involved tube-wall thinning due to suldation
when carbon steels were used and subsequent
severe cracking problems for Type 304L/carbon
steel coextruded tubes when the furnace oor
tubes and waterwall tubes were upgraded with
these coextruded tubes. A review of the current
understanding on the tube cracking issues and
improved alloys is presented in this chapter.
13.2 Fuel, Combustion, and Boilers

Black liquor contains dissolved wood components, inorganic constituents (NaOH, Na2S,
Na2CO3, Na2SO4, Na2S2O3, and NaCl), and
water. An example of the elemental compositions
of black liquor from North American wood
species is shown in Table 13.1 (Ref 5). Black
liquor is injected by spraying into the boiler
through liquor guns at about 4.6 m (15 ft) above
pg 359
31/10/2007 12:47PM Plate # 0
the boiler at the location above the secondary

air ports. The liquor droplets injected into the
boiler are generally 0.5 to 5 mm in size (Ref 7).
The droplets should be small enough so they
become dry or nearly dry before reaching the
char bed to aviod smelt-water contact. However,
they should not be too small, otherwise the droplets can be entrained in the ue gas stream
resulting in fouling and plugging in the convection path. (Ref 7). Large droplets may travel
directly to the walls (Ref 7). In general, black
liquor burns in approximately four stages,
which are drying, devolatilization (pyrolysis),
char burning, and smelt coalescence and reactions (Ref 7). The furnace can typically be divided into three zones: the reducing zone at the
bottom, the drying zone where liquor is red, and
the oxidizing zone in the upper furnace, as
schematically illustrated in Fig. 13.2 (Ref 8). The
gure also shows the char bed that forms on the
furnace hearth during combustion of black
liquor.
the hearth for some boilers (Ref 6). Air for

combustion is injected into the furnace separately
from primary and secondary air ports for substoichiometric combustion and also from tertiary
air ports to complete the combustion. Figure 13.1
shows a schematic diagram of a recovery boiler
with primary, secondary, and tertiary air ports
(Ref 6). The liquor is shown to be sprayed into
Table 13.1 Elemental compositions of black
liquor from North American wood species
Element
Concentration, wt%
Carbon
Hydrogen
Oxygen
Sodium
Sulfur
Potassium
Chlorine
Nitrogen
Others
3439
35
3338
1725
37
0.12
0.22
0.040.2
0.10.3
Source: Ref 5
Steam flow
to mill
Stack gas
Feedwater
Sootblowing steam
Induced
draft
fan
Boiler
bank
Superheater
Economizer
Waterwalls
Bullnose
Electrostatic
precipitator
Dust
Recycle
Blowdown
Forced draft
fan
Furnace
Steam coil
air heater
Liquor
guns
Black
liquor
mix
tank
Tertiary air
Liquor
heater
Secondary air
Primary air
Smelt spouts
Strong black
liquor from
concentrator
Fig. 13.1
Makeup
salt cake
Recovery boiler. Source: Ref 6
Direct
heating
steam
Smelt to
dissolving tank
pg 360
31/10/2007 12:48PM Plate # 0
Chapter 13: Black Liquor Recovery Boilers in the Pulp and Paper Industry / 361
Flue
gas
Oxidizing
Secondary
air
Drying
Black liquor
Primary
air
Reducing
Smelt
Fig. 13.2
Recovery boiler illustrating three zones of reactions

in the furnace. Source: Ref 8
Formation of a char bed on the furnace hearth

during combustion of black liquor makes the
recovery boiler uniquely different from fossilred boilers and WTE boilers discussed in previous chapters. The char bed generally consists
of carbon, partially pyrolyzed black liquor solids,
and molten and frozen smelt (Ref 9). The characteristics of char beds are different for different
furnace oor designs. The sloped (or inclined)
furnace oor, such as for B&W boilers, and the
decanting oor, such as for Alstom CE boilers,
exhibit different characteristics of char beds,
and their characteristics are schematically illustrated in Fig. 13.3 and 13.4 for the sloped oor
and decanted oor, respectively (Ref 9). For the
sloped oor design, molten smelt ows from the
active char bed and is collected in troughs around
the perimeter of the bed and then discharged
through smelt spouts (Ref 9). For the decanting
oor design, molten smelt is collected and contained by the decanting bottom of the boiler
(Ref 9).
Since the oor tubes and lower furnace
waterwall tubes are cooled with water, a layer
of the frozen (or solidied) smelt is formed on
the tube surface, thus preventing rapid corrosion
of tube metal by the molten salts. With studded
carbon steel tubes, a thicker frozen layer is
formed (Ref 10). The frozen smelt is not corrosive as long as it remains solid on the tube
surface. However, once this frozen smelt layer
suffers localized disturbances or instability,
Active layer
Pyrolysis, combustion, reduction
1025 cm (410 in.) thick
8001200 C (15002200 F)
Secondary
air
Inactive core
Solidified smelt
Carbon
<760 C (1400 F)
Primary
air
Hearth 230290 C (450550 F)
Solid smelt
Smelt spout
Fig. 13.3
Liquid smelt
Char beds formed in the sloped floor boiler. Source: Ref 9
pg 361
31/10/2007 12:48PM Plate # 0
pg 362
Secondary
air
Chemically active layer

Pyrolysis, combustion, reduction
8001200 C (15002200 F)
Physically
active
layer
Primary
air
Liquid smelt
Solid smelt
Smelt spout
Fig. 13.4
Hearth 260340 C (500650 F)
Char beds formed in the decanting hearth boiler. Source: Ref 9
molten smelt may come within close proximity

to, and possibly in direct contact with, the oor
tubes, prompting severe tube overheating and
thus premature tube failures (Ref 11). Hogan
(Ref 11) reported several oor tube overheating
incidents; in one case the tubes were overheated
in excess of 675 C (1250 F) and in another
case in excess of 720 C (1330 F). The author
(Ref 11) attributed the overheating to localized
disturbances or instability of the frozen melt on
the tube surface. The furnace combustion conditions can be greatly affected by many boiler
operating factors. For example, plugging of air
ports by unburned material or unreacted smelt
as well as plugging of liquor gun nozzles can
greatly inuence the furnace combustion conditions (Ref 12).
Above the char bed up to the tertiary air ports,
the gaseous environment remains under reducing
conditions. Singh et al. (Ref 13) analyzed gas
samples collected from the areas near the waterwall between the secondary and tertiary air ports
in an operating boiler. The samples collected
were analyzed using the gas chromatograph
technique. The samples were collected from the
area of the carbon steel waterwall that had
experienced high corrosion. Their results are
summarized in Table 13.2 (Ref 13). The amount
of water vapor (% H2O) in the environment was
Table 13.2 The compositions of major gas

species detected in gas samples collected from the
different locations of the lower furnace waterwall
between the secondary and tertiary air ports
Composition, %
Gas species
N2
CO2
CO
H2
O2
CH4
H 2S
SO2/COS
Methyl mercaptan
At waterwall
surface
2.5 cm (1 in.)
away
24.8
29.4
7.5
2.4
2.9
4.0
18
0.2
0.17
32.9
23.9
9.1
2.3
1.0
2.0
24.2
0.27
0.51
30 cm (12 in.)
away
47.6
18.3
16.8
1.5
1.1
0.9
0.6
0.04
0.002
Source: Ref 13
excluded from the data in the table. In addition,

the authors indicated that during the present
study, the black liquor was intentionally sprayed
on the sidewalls for drying, thus leading to higher
concentrations of H2S. It should be noted that the
samples collected were not in equilibrium; thus,
oxygen was detected. The results show that the
concentration of H2S was signicantly reduced
as the probe moved away from the waterwall.
Hydrogen sulde (H2S) is the major sulfur
species of the gaseous sulfur compounds generated in the lower furnace. Other gaseous sulfur
31/10/2007 12:48PM Plate # 0
compounds include methyl mercaptan, dimethyl

sulde, and dimethyl disulde. These compounds are frequently referred to as total reduced sulfur (TRS) gases. In the lower furnace,
the concentration of TRS gases is typically in a
range of few hundred to a thousand ppm (Ref 14).
As the ue gas stream rises to the upper furnace and enters into the superheater bundles, the
oxidizing atmosphere converts H2S to SO2. The
ue gas stream is also entrained with y-ash
particles, which include carryover of black liquor
droplets and fume that is formed by condensation
of volatilized inorganic salts (Ref 15). These yash particles deposit on the cooler surfaces of
the heat-exchanger tubes, such as superheaters
and generating banks, in the convection path.
The deposits consist mainly of sodium sulfate
(Na2SO4) and sodium carbonate (Na2CO3) with
small amounts of sodium sulde (Na2S), sodium
chloride (NaCl), and potassium salts. The rst
melting temperature (FMT) of the deposits typically varies between 520 and 580 C (970 and
1075 F), depending on composition (Ref 16).
The superheater tube can suffer accelerated
wastage when the deposit in contact with the tube
surface exceeds its FMT (Ref 17).
13.3 Materials Problems in Lower

Furnace and Superheaters
The rst Tomlinson recovery boiler built in
1929 had refractory furnace walls that were too
costly to maintain (Ref 1). The furnace design
was later changed to a completely water-cooled
furnace enclosure including a furnace oor made
of carbon steel, with the rst such unit built in
1934 (Ref 1). Lower furnace waterwalls and oor
tubes were protected by pin studs that held frozen
smelt providing protection against molten smelt
(Ref 1). There have been several incidents of pin
stud oor tube wastage and failure involving
both sloped oor and decanting oor boilers (Ref
18). Clement and Blue (Ref 18) characterized the
failure of the pin stud carbon steel oor tubes to
be stud wastage, or burn-back, and tube-wall
thinning. Since the 1970s, an increasing number
of mills, rst in Scandinavia and then in North
America, gradually upgraded carbon steel
waterwalls and oors in the lower furnace to
coextruded tubes with Type 304L outer cladding
to provide corrosion protection for the inner
carbon steel tube (Ref 19). Singbeil et al. (Ref 19)
reported in 1997 that most of the 56 recovery
boilers in Scandinavia (34 in Sweden and

Norway and 22 in Finland) had coextruded tube
walls and more than half had coextruded tube
oors, and in North America about 65 recovery
boilers out of approximately 340 kraft recovery
boilers had coextruded tube oors and many
others had coextruded tube walls. One major
U.S. boiler manufacturer (Babcock & Wilcox)
installed coextruded tubes in the smelt ow areas
adjacent to the sidewalls with carbon steel studded tubes in the center of the oor in 22 boilers
and in the complete oor for 26 boilers (Ref 18).
In the 1990s, cracking of the 304L cladding of
the composite tubes had become a major materials issue in recovery boilers, although some
cracking was reported in the 1980s (Ref 19, 20).
In 1995, a United States Department of Energy
(DOE) program was established to determine
the cause of the tube cracking and to identify
alternative materials or process changes to prevent this type of cracking. This project, coordinated by J.R. Keiser of the Oak Ridge National
Laboratory (ORNL), was carried out by researchers at ORNL, the Pulp and Paper Research
Institute of Canada (PAPRICAN), the Institute of
Paper Science and Technology (IPST) with
support of more than a dozen paper companies,
and boiler and tubing manufacturers (Ref 20).
The results of the research projects under this
program, which were presented at regular ORNL
review meetings and published in various
technical journals, provided a signicant understanding on the root causes of the 304L composite tube cracking problems. This chapter
reviews (a) the cracking problems of the 304L
cladding in composite tubes used as oor tubes,
smelt openings, and primary air port openings
and (b) alternative materials. Superheater corrosion is also included in the discussion.
13.3.1 Furnace Floor Tubes
Cracking of the 304L cladding has been found
to occur in (a) both high-pressure (8.3 to
10.3 MPa, or 1200 to 1500 psi) and low-pressure (4 to 6.2 MPa, or 600 to 900 psi) boilers, (b)
both sloped oor and decanting oor designs,
and (c) boilers by different boiler manufacturers
(Ref 19). Cracks typically initiated from the
outer diameter of the clad tube and propagated
inward to the substrate steel; in most cases, the
cracks terminated at the cladding/steel interface
with some cracks changing the direction of
propagation along the cladding/steel interface
(Ref 1921). This is illustrated in Fig. 13.5 for the
pg 363
31/10/2007 12:48PM Plate # 0
pg 364
1250 m
500 m
Fig. 13.5
Fig. 13.6
clad tube and Fig. 13.6 for the clad membrane.

These cracks were found to be essentially transgranular in nature (Ref 19, 20). In order to
determine the cracking mechanisms, the DOE
program included the following studies: (a)
measurement and modeling of residual stresses
and the stress states of the composite tubes, (b)
tube metal temperature measurements during
operation, (c) thermal fatigue testing of tube
materials, and (d) the effect of waterwashing
during shutdowns (Ref 20).
To determine whether thermal fatigue was a
possible cause of oor tube cracking, oor tube
metal temperature was measured in a boiler at the
Weyerhaeuser mill in Prince Albert, Saskatchewan, Canada, during September 1, 1998
through February 28, 1999 (Ref 22). This boiler
with a sloped oor had experienced cracking of
Type 304L cladding in coextruded oor tubes. In
many boilers, cracking had occurred within 2 m
of the rear (spout) wall. Keiser et al. (Ref 22)
installed 25 thermocouples on the oor and
recorded the temperature spikes over a 6-month
period, showing an average of about one thermal
spike per thermocouple per day. A thermal spike
was dened as a temperature spike of more
than 50 C (90 F) above approximately 275 C
(530 F). The authors observed that the thermocouples located near the spout wall that suffered
tube cracking showed the fewest thermal spikes,
while the areas farther from the spout wall that
exhibited little or no cracking experienced the
greatest numbers of thermal spikes (Ref 22). The
authors concluded that the frequency of thermal
spikes was far too low for thermal fatigue to be
the cause for the cracking (Ref 22).
Thermal fatigue data for Type 304H and 304L

cycling to 500, 550, and 600 C (930, 1020, and
1110 F), respectively, were generated and compared with the ASME Sec. III Subsec. NH design curve for Type 304H at 430 C (800 F)
in isothermal fatigue. All the thermal fatigue
data fell near or above the ASME design curve
(Ref 20, 23). Assuming thermal cycling to
450 C (840 F) is considered, the cyclic strains
would be about 0.25% and the fatigue life for the
stainless steel cladding is expected to be in excess
of 100,000 cycles (Ref 23).
There is some thermal expansion mismatch
between the 304L cladding and the carbon steel
substrate in the composite tube due to differences
in mean coefcients of thermal expansion
between these two alloys (e.g., 7.6 106 and
10.0 106 in./in. F from 70 to 800 F for
carbon steel and Type 304, respectively). The
residual stresses formed in the cladding in both
the as-fabricated condition and after service were
studied under the DOE program. Also investigated in the program was nite element modeling
of stresses in the cladding. Keiser et al. (Ref 20)
summarized the results of both residual stress
measurements and nite element modeling of the
304L/SA210 coextruded composite tubes. Residual stress measurements using x-ray and neutron
diffraction showed compressive axial and tangential stresses on the outer surface of the cladding of the as-fabricated tube, and tensile axial
and hoop stresses (up to a maximum of 300 MPa)
on the crown location of the coextruded tube after
service in the boiler oor (Ref 20). Finite element
modeling studies showed that the 304L cladding
developed tensile stresses when cooled down
Cracks initiated on the outer diameter of the 304L

clad tube, propagated inward to the substrate steel
and terminated at the cladding-steel interface. Courtesy of Oak
Ridge National Laboratory.
Cracks initiated on the outer surface of the 304L

cladding, propagated inward to the substrate steel
of the membrane and terminated at the cladding-steel interface.
Courtesy of Oak Ridge National Laboratory.
31/10/2007 12:48PM Plate # 0
from the operating temperature (e.g., 300 C,

or 570 F) to room temperature (Ref 20). Furthermore, the 304 cladding developed tensile
stresses when cooled down from the thermal
spike to the operating temperature (e.g., cooling down from 550 to 300 C, or 1020 to 570 F)
(Ref 20). As a result, tensile stresses can be present in the 304 cladding at operating temperature after a thermal spike as well as when oor
tubes are cooled to room temperature.
The presence of tensile stresses in the 304L
cladding when the composite oor tubes are at
low temperatures is signicant in that it can make
stress-corrosion cracking (SCC) a potential
cracking mechanism since the morphology of
cracking resembles that of SCC. Prescott and
Singbeil (Ref 24) reported that Type 304L suffered stress-corrosion cracking under tensile
stresses when exposed to concentrated solutions
of alkali compounds containing Na2S at temperatures from 160 to 200 C (320 to 390 F). In
C-ring tests of Type 304L/SA210A1 coextruded
tubes in a salt consisting of 90% Na2S9H2O and
10% NaOH at about 170 C (340 F) for 48 h,
cracks developed at the outer diameter of the
cladding and propagated inward and terminated
at the cladding/substrate interface (Ref 20).
Stress-corrosion cracking of Type 304L takes
place in this type of environment within a certain
temperature range. For example, in constant load
tests with 275 MPa (40 ksi) in Na2S-10%NaOH,
stress-corrosion cracking of Type 304L readily
occurred at 150 to 250 C (300 to 480 F)
(Ref 23). It is generally believed that waterwashing of the boiler when there are signicant
remnants of smelt on the oor results in aqueous
solutions that contain Na2S along with NaOH or
Na2CO3. These aqueous solutions are believed
(Ref 20, 23, 25, 26) to be an essential requirement
for stress-corrosion cracking of the 304L cladding either during the boiler shutdown or startup
when the oor tubes are covered with these salt
solutions and subject to this SCC temperature
range. The critical temperature range is believed
to be 150 to 200 C (300 to 400 F) (Ref 23).
C-ring tests were conducted in a salt consisting of 90% Na2S9H2O and 10% NaOH
at about 170 C (340 F) for 48 h for Type
304L/SA210A1 coextruded tube, Sanicro 38*
* Sanicro 38 meets UNS N08825 (alloy 825) according to

Sandvik Steel's technical bulletin on Sanicro 38/4L7 coextruded tubes (Ref 27), thus the cladding should be considered
alloy 825, not modied alloy 825.
(modied alloy 825) coextruded tube (on carbon

steel), Type 309 overlay tube, alloy 625 overlay
tube, Type 310 coextruded tube, and carbon
steel (Ref 20). The test results of these tubes are
summarized: (a) Type 309 overlay tube was
found to be as bad as the 304L coextruded tube,
(b) Type 310 coextruded tube also suffered
SCC cracking with the longest crack being
about 325 m, (c) Sanicro 38 cladding (modied
alloy 825) exhibited cracks, but less than
50 m deep, (d) alloy 625 overlay showed no
cracking, and (e) carbon steel showed no cracking (Ref 20).
One of the conditions required for developing
stress-corrosion cracking is the presence of
tensile stress. In this regard, alloys 825 and 625
have advantages over Type 304L in that both
exhibit coefcients of thermal expansion much
closer to those of carbon steels. Instead of
developing axial tensile stresses in Type 304L
when cooled from the operating temperature
to room temperature, both alloys 825 and
625 develop axial compressive stresses (Ref
23). This is important in that cracking predominantly occurred in the transverse direction
(Ref 23). The alloy 625 overlay in the as-overlaid
tube typically exhibits residual tensile stresses
that can be eliminated by annealing the tube
at 900 C (1650 F) for 20 min (Ref 28). The
overlay tubing is produced by a spiral overlay
welding mode using gas metal arc and gas
tungsten arc welding processes (Ref 29), and its
structure and properties are reported elsewhere
(Ref 30, 31).
Keiser et al. (Ref 26) reported the results of the
tests comparing austenitic stainless steel claddings of Type 304L coextruded tubes and Type
309L overlay tubes with nickel-base alloy claddings of alloy 825 coextruded tubes, alloy 625
coextruded tubes, and alloy 625 overlay tubes in
a recovery boiler in North America. Both 304L
coextruded tubes and 309L overlay tubes
showed cracking during the second year of service, while alloy 825 coextruded, 625 coextruded, and 625 overlay tubes showed no
cracking after 5 years of service. Barna and
Rivers (Ref 21) reported in their 1999 paper
that several oors have been constructed with
alloy 825 coextruded tubes since 1996. Wilson
et al. (Ref 32) reported good performance
results of Sanicro 38 coextruded tubes as oor
tubes in recovery boilers. The authors (Ref 32)
indicated that the outer cladding alloy in Sanicro
38 coextruded tube was a modied alloy 825.
Nevertheless, the chemical composition of
pg 365
31/10/2007 12:48PM Plate # 0
the outer cladding published in the paper

was within the specication of UNS N08825
(alloy 825). Sandvik Steels technical bulletin
(Ref 27) on Sanicro 38/4L7 composite tubes
also indicates that the cladding meets UNS
N08825.
Lai and Wensley (Ref 31) reported the performance experience of alloy 625 overlay oor
tubes in several boilers where good results were
observed. One boiler (a Babcock & Wilcox
design) operated reportedly at 1420 psig and
produced 883,400 lb/h steam. After severe
cracking of Type 304L coextruded tubes in 1995,
alloy 625 overlay tubes were installed in the
smelt runs in 1997 replacing the 304L coextruded tubes. Figure 13.7 shows a general view of
the smelt run oor tubes in the boiler (Ref 31).
The alloy 625 overlay tubes had been in service
for about 7 years with no reported material problems when the paper was published in 2005.
Barna and Rivers (Ref 21) also reported the
installation of alloy 625 overlay tubes for boiler
oors in several boilers. One installation
involved covering the entire width of about onehalf the boiler oor in 1996, and three other
boilers were tted with partial oor installations
of alloy 625 overlay tubes in 1995 and 1997. No
cracks or other damage were reported when the
paper was presented in 1999 (Ref 21).
Fig. 13.7
13.3.2 Smelt Spout Openings

Cracking of Type 304L coextruded tubes
was initially observed at smelt openings in the
early 1980s before the observation of cracking of
Type 304L coextruded oor tubes (Ref 21).
Cracks can occur on adjacent tubes that form the
spout wall (Ref 21). The circumferential cracks
that occur in the 304L coextruded tubes at the
smelt openings have been found to propagate
from the outer stainless steel cladding through
the interface and into the carbon steel substrate
(Ref 19). Smelt openings are subject to thermal
uctuations due to intermittent exposure of the
owing smelt through the spout (Ref 21, 33). The
wall tubes including the tubes adjacent to spout
openings are also subject to thermal uctuations
due to rising and falling of the surface of the
smelt pool (Ref 18). It is believed that thermal
fatigue cracking may be a contributing factor for
cracking of smelt opening tubes made of Type
304L coextruded tubes (Ref 18). Dykstra et al.
(Ref 34) attributed the cracking of Type 304L
coextruded smelt spout opening tubes to thermal
fatigue. The authors (Ref 34) indicated that
fractographic examination of the crack surfaces
showed evidence of cyclic crack progression.
In addition to the thermal fatigue type cracks,
craze cracks were also observed in Type 304L
General view of the alloy 625 overlay smelt run floor tubes in the boiler. Source: Ref 31
pg 366
31/10/2007 12:48PM Plate # 0
coextruded spout opening tubes (Ref 34). These

craze cracks (typically branched) were found to
terminate at the cladding/steel interface (Ref 34).
In order to investigate alternate cladding
alloys, Clement and Blue (Ref 18) conducted
thermal fatigue testing of alloy 825 coextruded
tube and alloy 625 overlay tube compared with
Type 304L coextruded tube and Type 304L
monolithic tube. The substrate tubes for the
above three composite tubes were all SA210 A1
steel. The test involved a water-cooled test tube
that was rotated to make the tube-wall temperature uctuate from the cladding temperature at
the point of ame impingement to a low of about
40 C (105 F). The temperature of the cladding
at the point of ame impingement was measured
inside the cladding at the point about 0.25 mm
(10 mils) from the cladding surface. The cladding metal temperature cycled from about
650 C (1200 F) (high) to 40 C (105 F)
(low). Solid 304L tubes failed in less than 5000
cycles, and the 304L clad tubes failed in slightly
over 10,000 cycles to less than 10,000 cycles.
Alloy 825 coextruded tubes cycled for 25,000
cycles showing no cracks, and alloy 625 overlay
tubes cycled for 30,000 cycles showing no
cracks. Both 825 coextruded tubes and 625
overlay tubes were then cycled from 815 C
(1500 F) (high) to 40 C (105 F) (low) for
Fig. 13.8
30,000 cycles and showed no cracking. The test

results showed that both alloy 825 coextruded
tubes and alloy 625 overlay tubes were very
resistant to thermal fatigue cracking.
Barna and Rivers (Ref 21) reported in 1999
that there were about 60 smelt openings that were
made of alloy 825 coextruded tubes. However,
it has been reported that cracking of alloy 825
cladding was encountered in some boilers
(Ref 21). Furthermore, the cracks were found to
terminate at the fusion boundary or turn at 90
and proceed along the fusion boundary (Ref 21).
The authors (Ref 21) also reported that alloy
625 overlay tubing was rst installed in a smelt
opening in 1987 in a boiler where Type 304L
coextruded tube smelt openings required replacement approximately every 6 months as well as
several other installations including one boiler
where alloy 625 overlay smelt opening tubes
were installed in 1994. Lai and Wensley (Ref 31)
reported that a smelt spout wall (including
opening tubes) made of alloy 625 overlay tubes
was installed in a B&W 1050 psi boiler in 2000
replacing Type 304L coextruded tubes that had
experienced cracking problems. At the time the
authors (Ref 31) presented their paper in
2005, no cracking problems had been reported.
Figure 13.8 shows that the alloy 625 smelt spout
wall including opening tubes revealed no
Alloy 625 overlay smelt spout wall including opening tubes after 2 years of service showing no indication of cracking by
liquid dye penetrant testing. Source: Ref 31
pg 367
31/10/2007 12:48PM Plate # 0
cracking in dye penetrant testing after 2 years of

service (Ref 31).
13.3.3 Air Port Openings

Cracking in both craze (or mosaic) and circumferential patterns has been observed in Type
304L coextruded tubes at the air port openings
(Ref 3538). These authors observed that some
of the circumferential cracks penetrated into the
carbon steel tube. Most cracks, however, did not
progress into the carbon steel tube (Ref 36).
Almost all the cracking was found to occur at the
bottom half of the air port openings (Ref 36, 38).
Temperature measurements of the air port
opening tubes in several mills have indicated that
those air port opening tubes that had historical
cracking problems showed the greatest temperature uctuations in frequency and amplitudes (Ref 36). The air port opening tubes with
no historical cracking problems exhibited no
signicant thermal uctuations (Ref 36). In one
example, based on the 5-day temperature measurements, the air port opening tubes with
no history of cracking problems in one mill
showed temperature uctuations between 300
and 400 C (570 and 750 F), while the other
mill with a history of cracking problems showed
extensive thermal uctuations from about
350 C (660 F) to temperatures exceeding
500 C (930 F) (Ref 36).
In making mill visits to inspect the air port
openings, the authors (Ref 36) observed that
the air port opening tubes with a shorter, wider
design appeared to suffer more frequent cracking
problems than the ones with a longer, narrow
opening design. This certainly suggests that
the cold-worked conditions of the tube bends
and/or the increased stresses due to increased
constraint might be a factor. The results of the
residual stress measurements using x-ray and
neutron diffraction methods showed that the
304L coextruded tube typically exhibits compressive axial stresses on the surface of the
cladding at the unbent section and tensile axial
stresses in some areas of the bent section (Ref
37). One interesting observation was made with a
video camera used to record the air port opening
activities in operating boilers. What appeared to
be molten smelt, which was found on the opening
tube surfaces, owed down the tube surfaces
(Ref 37). This suggests that molten smelt could
likely be on air port opening tube surfaces during
operation, thus resulting in possible local
temperature excursions (spikes) and corrosion

reactions with the tube surface.
Shenassa et al. (Ref 39) indicated that the air
port design and fabrication may be a factor in
causing primary air port cracking problems. The
authors (Ref 39) discussed two air port opening
designs (the casting design and the nonwelded
insert design) provided by a major European
boiler designer. The opening is formed by
bending one or two composite tubes out of the
plane of the waterwall toward the cold side of the
boiler. It was observed that the casting design,
which involved a cast iron casting being bolted to
the bent tubes from the cold side, appeared to
show better resistance to cracking because of a
low bending angle and the optimum bending
radius of the opening tubes (Ref 39). This
observation was consistent with what was
observed by Keiser et al. (Ref 36), who observed
that opening tubes with a shorter, wider design
appeared to suffer more frequent cracking problems than the ones with a longer, narrow
opening design. In a search for alternate alloys in
coextruded tubes to replace Type 304L coextruded tubes in primary air port openings with the
same air port design, alloy 625 coextruded tubes
were installed. After 1 year of operation, cracking was observed on many of the primary air port
opening tubes (Ref 39). Alloy 825 coextruded
tubes have been installed in primary air port
openings in several boilers including one unit
with alloy 825 coextruded primary air port
openings having been in operation for 3 years; no
cracks had been observed on these air port
opening tubes (Ref 39).
Keiser et al. (Ref 36) reported that the
inspection of a mill revealed extensive thinning
at the reside of alloy 825 coextruded air port
opening tubes with no evidence of cracking. Lai
and Wensley (Ref 31) reported severe tube
thinning for primary air port openings fabricated
from spiral overlay Fe-20Cr-1Nb composite
tubes after 1 year of service in a mill in South
America. A close-up view of the ferritic alloy
overlay tubes is shown in Fig. 13.9. Keiser et al.
(Ref 37) reported a case where the air port
opening made of alloy 625 overlay tubes was
found to suffer overlay thinning of about 1 mm
(0.04 in.) after the rst 6 months of service.
However, no further additional wastage was
observed during subsequent exposure (Ref 37).
In fact, Wensley (Ref 40) reported that there was
no further overlay thinning after an additional 5
years of operation for this boiler. It is believed
that the initial corrosion was likely to result from
pg 368
31/10/2007 12:48PM Plate # 0
changes in liquor ring practices (Ref 40). Furthermore, no cracking has been encountered for
these air port opening tubes (Ref 40). Lai and
Wensley (Ref 31) summarized the performance
experience of alloy 625 overlays in primary air
port openings. In a B&W boiler operating at
1420 psig and producing 883,400 lb/h steam,
about 29 air port openings were replaced with
alloy 625 overlay tubes in 1999. These tubes
have been inspected annually using dye penetrant testing (PT). No cracking was reported
at the time the authors presented their paper
(Ref 31). Figure 13.10 shows some of these alloy
625 overlay air port opening tubes. In another
boiler, alloy 625 overlay air port opening tubes
were installed in 2000, and the inspection of
these tubes in 2002 did not reveal any cracks or
corrosion attack (Ref 31). However, inspection a
few years later showed cracks were developed in
few of these overlay tubes, and these cracks had
advanced at least to the carbon steel (Ref 41).
The performance experience of alloy 625
coextruded tubes used in air port openings
has been reported by several authors (Ref 36,
37, 39). Shenassa et al. (Ref 39) reported that air
port openings made of alloy 625 coextruded
tubes (Sanicro 63) in a mill suffered cracking

after only 1 year of operation. Cracking was
observed in many of the primary air port openings (Ref 39). Keiser et al. (Ref 37) observed
that the circumferential cracks that occurred in
alloy 625 coextruded tubes were intergranular
in nature (i.e., cracking along grain boundaries)
(Fig. 13.11). Intergranular cracking was also
Fig. 13.10
Primary air port openings fabricated from alloy

625 overlay tubes have been in service
since 1999. No cracking or corrosion has been observed.
The photo was taken after liquid dye penetrant testing. Source:
Ref 31
Fig. 13.9
Primary air port opening made of Fe-20Cr-1Nb

spiral overlay tubes suffered severe fireside
wastage after 1 year of service in a mill in South America. Source:
Ref 31
Fig. 13.11
Intergranular cracking that occurred in the alloy

625 coextruded air port opening tube. Courtesy
of Oak Ridge National Lab.
pg 369
31/10/2007 12:48PM Plate # 0
pg 370
observed in alloy 625 coextruded oor tubes and

smelt opening tubes (Ref 42).
Primary air port openings made of alloy 625
coextruded tubes in a boiler in the Southeast
(U.S.) were also found to suffer primarily circumferential cracking after 1 year of service (Ref 43).
The cracking was also found to be intergranular
(Fig. 13.12). However, these circumferential
cracks were found to terminate at the cladding/
steel interface and to change the direction of
propagation to follow the cladding/steel interface, as shown in Fig. 13.13 (Ref 43). Corrosion
products formed inside the crack at the top portion
(Fig. 13.14) and at the crack tip (Fig. 13.13b)
were analyzed using energy-dispersive x-ray
spectroscopy (EDX). In the top portion of the
crack, the corrosion products appeared to be primarily Ni-rich oxides (Fig. 13.14). The oxides
formed at the crack tip (Fig. 13.13b) were Ni-Cr
rich (area No.3) before the crack reached the
cladding/steel interface and Fe-Ni rich (area No.1
and 2) when the crack propagated along the
cladding/steel interface. It is more important to
note that both Na and K were detected in these
corrosion products, although it is not clear what
role these alkali metals played in cracking.
Very little data have been reported on the coldwork condition of the alloy 625 coextruded tube
that was formed for the air port opening. A primary air port opening tube made of alloy 625
coextruded tube that suffered circumferential
cracking after 1 year of service in a boiler in
Southeast (U.S.) was removed from the boiler

and examined for the cold-work condition at
different locations of the tube. The examination
was carried out by conducting microhardness
600 m
(a)
3
2
1
(b)
Fig. 13.13
Fig. 13.12
Intergranular cracking of the alloy 625 coextruded air port opening tube in a mill in Southeast
(U.S.) after 1 year of service. The microstructure was not etched,
but cracking could be traced to follow austenitic grain boundaries. Courtesy of Welding Services Inc.
11 m
Scanning electron micrographs (backscattered

electron image) showing a crack penetrated
through the 625 cladding (of a coextruded tube) and then terminated at the cladding/steel interface at low magnification (original
micrograph 35) (a), and at high magnification showing the crack
changed direction and followed the cladding-steel interface (b).
Semiquantitative EDX analysis (wt%) of the corrosion products on
the crack tip is summarized below. Courtesy of Welding Services
Inc.
1: 52.0% Fe, 29.4% Ni, 9.5% Cr, 3.4% K, 1.4% Na, 0.8% S, and
trace elements
2: 61.3% Fe, 25.8% Ni, 7.6% Cr, 2.1% K, and trace elements
3: 14.0% Fe, 50.3% Ni, 24.4% Cr, 4.0% K, 2.2% Na, and trace
elements
31/10/2007 12:48PM Plate # 0
pg 371
measurements across the cross section of the

cladding at 0.25, 0.50, 0.75, 1.00, and 1.25 mm
(0.010, 0.020, 0.030, 0.040, and 0.050 in.) from
the cladding surface (Ref 43). Vickers hardness
tester with a 500 g load was used for microhardness measurements. Vickers hardness values
(HV) were then converted to Rockwell C scale
(HRC). The schematic of the air port opening
tube sample along with measured hardness at
different locations is shown in Fig. 13.15.
According to the plant personnel, the air port
opening tubes were reportedly not stress relieved
or annealed after cold forming and prior to
installation (Ref 43). Circumferential cracks
were found to develop at the crown location (i.e.,
area with the highest heat ux) where the tube
section was bent in the bottom portion of the air
port opening. This area showed hardness of
about 36 to 39 HRC near the cladding surface
(about 0.25 mm, or 0.01 in., from the cladding
surface), which were slightly lower than either
the extrados (the exterior bend section) (40 HRC)
or the intrados (the inner bend section) (42 HRC)
of the bend. The straight section was found to be
about 30 HRC (about 0.25 mm, or 0.01 in., from
the cladding surface). The microhardness measurements of this air port opening sample showed
that the area that suffered cracking was in a coldworked condition.
From the above discussion, it appears that
cracking of the primary air port openings tends to
occur at the locations where the tube is bent. The
cladding in these locations is under a coldworked condition. The material at these locations
is characterized by high residual stresses. Some
alloys with high residual stresses (e.g., in a highly
cold-worked condition) can suffer brittle, intergranular cracking when heated to or in service at
temperatures of 425 to 650 C (800 to 1200 F),
depending on the alloy. The morphology of this
type of intergranular cracking, often referred to
as stress-relaxation cracking, is similar to the
morphology of cracking observed in Type 304L
and alloy 625 claddings of the coextruded primary air port tubes. This cracking mechanism,
which has not been discussed by any of the
authors, may play a role in the air port opening
cracking. The process of stress-relaxation
cracking can be described in a simple way: when
a heavily cold-worked alloy with high residual
stresses (i.e., the residual stress in the alloy in the
cold-worked condition is essentially its roomtemperature yield strength) is heated to 540 C
(1000 F), for example, the yield strength of the
alloy at that temperature will be lower than the
residual stress of the alloy. This condition

prompts the alloy to undergo local plastic
deformation to relieve the stresses. However,
when the grain matrix is strengthened by ne
precipitates, it cannot allow plastic deformation
to occur to relieve those stresses, thus causing the
alloy to develop cracking at grain boundaries,
which are often the weakest locations. Austenitic
stainless steels and nickel-base alloys that form
homogeneous, coherent precipitates, such as 0
(Ni3Al) or 00 (Ni3Nb), are more susceptible to
this type of brittle cracking. Detailed discussion
on this subject is covered in Chapter 14 StressAssisted Corrosion and Cracking.
13.3.4 Waterwalls above Cut Line
The waterwalls above the butt weld joints with
composite tubes typically are bare carbon steel.
Corrosion of these waterwall carbon steel tubes is
primarily caused by suldation by reduced sulfur
gases, primarily H2S. The corrosion rate is typically 0.2 mm/yr (8 mpy), but can be as high as
1
2
21 m
Fig. 13.14
Scanning electron micrographs (backscattered

electron image) showing the top portion of the
crack (shown in Fig. 13.13a) that penetrated through the 625
cladding (of a coextruded tube) and then terminated at the cladding/steel interface and then changed direction and followed the
cladding/steel interface. Semiquantitative EDX analysis (wt%) of
the corrosion products on the top portion of the crack is summarized below. Courtesy of Welding Services Inc.
1: 55.6% Ni, 20.9% Cr, 4.7% Fe, 4.6% S, 2.6% Na, 7.1% Si,
1.9% Ca, 1.5% Al, and trace elements
2: 84.6% Ni, 5.2% Cr, 3.1% Fe, 2.8% Na, 1.6% Si, 1.0% S, and
trace elements
3: 85.6% Ni, 5.7% Cr, 3.1% Fe, 2.2% Na, 1.0% Si, 0.7% S, and
trace elements
4: 71.1% Ni, 18.3% Cr, 3.5% Fe, 2.3% Na, 1.2% Si, 1.6% S, and
trace elements
31/10/2007 12:48PM Plate # 0
0.8 mm/yr (32 mpy) (Ref 10). When suldation

becomes excessive, one effective approach is to
apply Type 309 overlay using automatic overlay
welding process. Instead of forming FeS on
unprotected carbon steel, protective Cr2O3 scales
form on the 309 overlay containing typically
20% Cr. Numerous papers have been published
on the benecial effects of chromium on suldation resistance of alloys (see Chapter 7 Suldation). Moberg et al. (Ref 44) illustrates the
benecial effect of chromium on the suldation
resistance of steels containing chromium in a
suldizing environment at 400 C (750 F) (Fig.
13.16). Lai and Hulsizer (Ref 45) reported that
Type 309 overlay was applied on the waterwalls
of a recovery boiler in a Midwest mill (U.S.) in
1987, 1989, and 1992. Figure 13.17 shows the
cross section of a Type 309 overlay sample
obtained from the overlaid waterwall after 13
years of service, showing essentially the original
overlay thickness with only tiny surface corrosion pits (Ref 45). A close-up view of the 309
overlay on the front wall after 8 years of boiler
operation in the same boiler is shown in Fig.
13.18.
13.3.5 Superheater Tube Corrosion
As the combustion ue gas rises to the upper
furnace, black liquor droplets entrained in the
ue gas stream as carryover particles can
33 HRC at 0.01 in.

(3133 HRC)
potentially contribute to the deposits on superheater platens. The deposits form primarily
on the windward side of the tube. When the
tube metal temperature reaches the melting
temperature of the deposits, corrosion of the
superheater can become serious. Corrosion is
more serious on the windward side than the leeward side of the superheater tube. This is illustrated Fig. 13.19 (Ref 46). Suldation/oxidation
is the major corrosion mechanism for superheaters (Ref 10). Common superheater alloys are
T-11 or T-22. When T-11 or T-22 superheater
tubes suffer high wastage rates, one cost-effective solution will be to switch to austenitic
stainless steels that are capable of forming protective chromium oxide scales instead of formation of iron oxides and/or suldes on
unprotected T-11 or T-22 tubes.
An example was given by Lai and Wensley
(Ref 31) for a boiler in South America, where
T-11 superheater tubes suffered severe wastage at
the bend sections and required annual replacement. A T-11 superheater tube bend sample was
removed for metallurgical evaluation after 6
months of service. The sample is shown in
Fig. 13.20. The wastage rate was estimated to be
approximately 3.9 mm/yr (154 mpy). Superheater steam temperature and pressure were
reportedly 435 C (815 F) and 6.4 MPa (63.5
bars, or 930 psig), respectively. Examination of
the sample revealed that the corrosion products
39 HRC at 0.01 in.

(3639 HRC)
36 HRC at 0.01 in.
(3537 HRC)
37 HRC at 0.01 in.

(3739 HRC)
42 HRC at 0.01 in.

(4042 HRC)
Cracks
40 HRC at 0.01 in.
(4042 HRC)
40 HRC at 0.01 in.
(3840 HRC)
Fig. 13.15
30 HRC at 0.01 in.

(2830 HRC)
Air port opening made of alloy 625 coextruded tube that suffered cracking after 1 year of service in a boiler in Southeast
(U.S.) with hardness data at different locations of the tube. Cracking occurred at the bottom half of the opening (right-hand
side of the tube). Rockwell C hardness values (converted from Vickers microhardness values) were reported in a range at different locations
of the tube as well as the hardness at 0.25 mm (0.01 in.) from the cladding surface. Courtesy of Welding Services Inc.
pg 372
31/10/2007 12:48PM Plate # 0
pg 373
14
400 C (750 F)
Rate of weight gain, g/m2/h
12
10
10
15
20
25
30
Chromium, %
Fig. 13.16
Effect of chromium on sulfidation resistance of steels containing various amounts of chromium tested at 400 C (750 F) in
N2-15H2O-10CO2-10H2-0.1O2-0.1H2S. Source: Ref 44
were essentially iron oxides but with the presence

of chlorine. Figure 13.21 and 13.22 show the
corrosion products and the energy-dispersive
x-ray spectroscopy (EDX) analysis of the corrosion products. The results of the EDX analysis
indicated that the corrosion products were
essentially iron oxides with no sulfur. However, a
relatively high chlorine (Cl) peak was detected in
the corrosion scales at the scale/metal interface. It
is, thus, believed the corrosion was caused by
chlorine-accelerated oxidation. Type 310 overlay
tubes were thus selected to replace numerous
T-11 tube bends in this boiler in 2000, 2001, and
2003. So far, no corrosion or other degradation
has been encountered with these overlay tubes.
Figure 13.23 shows the 310 overlay superheater
tubes after 2 years of operation in the boiler,
revealing no evidence of corrosion attack.
13.4 Summary
0.5 mm
Fig. 13.17

Type 309 overlay at the crown location of the rear
waterwall tube after 13 years of boiler operation in a boiler in a
Midwest mill (U.S.). The metallographic mount was etched with
nital to reveal the substrate steel including the fusion boundary
and the heat-affected zone (HAZ). Source: Ref 45
A brief description of a black liquor recovery

boiler along with its fuel and combustion conditions is presented. Materials problems with
oor tubes, smelt spout openings, and air port
openings in the lower furnace are discussed.
The materials that have been tested and tried
with good results are reported. The corrosion
issue for the waterwalls above the cut line (i.e.,
the butt weld joints where carbon steel waterwall
31/10/2007 12:49PM Plate # 0
tubes join with the composite tubes at the lower

part of the furnace) is discussed. In addition, the
corrosion of superheater tubes is also discussed.
REFERENCES

Wilcox, (1992), p 261
2. G.A. Smook, Handbook for Pulp & Paper
Technologists, 2nd ed., Angus Wilde Publications, Vancouver, Canada, (1992), p 74
3. T.J. Grant, Update of the American Forest
& Paper Associations Recovery Boiler
Program, TAPPI Engineering & Papermakers Conference (Conf. Proc.), Book 2,
TAPPI, (1997), p 589
Fig. 13.18
A general view of the 309 overlay on the front

wall after about 8 years of boiler operation in
Midwest mill (U.S.). Source: Ref 45
4. J. Gommi, Root Causes of Recovery Boiler

Leaks, TAPPI Engineering & Papermakers
Conference (Conf. Proc.), Book 2, TAPPI,
(1997), p 509
5. W.J. Frederick, Chapter 3, Black Liquor
Properties, Kraft Recovery Boilers, T.N.
Adams, Ed., TAPPI Press, (1997), p 61
6. T.M. Grace, Chapter 5, Chemical Recovery
Process Chemistry, Chemical Recovery in
the Alkaline Pulping Processes, 3rd ed., R.P.
Green and G. Hough, Ed., TAPPI Press,
(1992), p 57
7. W.J. Frederick and M. Hupa, Chapter 5,
Black Liquor Droplet Burning Processes,
Kraft Recovery Boilers, T.N. Adams, Ed.,
TAPPI Press, (1997), p 131
8. G.A. Smook, Handbook for Pulp &
Paper Technologists, 2nd ed., Angus Wilde
Publications, Vancouver, Canada, (1992),
p 133
9. T. Grace and W.J. Frederick, Chapter 6, Char
Bed Processes, Kraft Recovery Boilers,
T.N. Adams, Ed., TAPPI Press, (1997), p 163
10. H. Tran, Chapter 10, Recovery Boiler Corrosion, Kraft Recovery Boilers, T.N. Adams,
Ed., TAPPI Press, (1997), p 285
11. E.F. Hogan, Investigation of Chemical
Recovery Unit Floor Tube Overheating
Failures, TAPPI Engineering & Papermakers Conference (Conf. Proc.), Book 2,
TAPPI, (1997), p 567
12. C.M. Wells, Chapter VIII, Chemical
Recovery Area, Pulp and Paper Manufacturing, Vol 10, Mill-Wide Process Control & Information Systems, D.B. Brewster
and M.J. Kocurek, Ed., Joint Textbook
5
T-22 probe
Metal loss, mm
Windward
2
Leeward
1
0
440
460
480
500
520
540
560
580
600
620
640
660
Temperature, C
Fig. 13.19
Results of corrosion probe tests for T-22 in the lower superheater region for 840 h in a boiler. Source: Ref 46
pg 374
31/10/2007 12:49PM Plate # 0
pg 375
C
B
Fig. 13.20
T-11 superheater tube sample showing severe

corrosion attack after 6 months of operation.
Source: Ref 31
13.
14.
15.
16.
17.
18.
19.
20.
Committee of the Paper Industry, Canada,

(1993), p 124
P.M. Singh, S.J. Al-Hassan, S. Stalder, and
G. Fonder, Corrosion in Kraft Recovery
BoilersIn-Situ Characterization of Corrosive
Environments,
1999
TAPPI
Engineering/Process and Product Quality
Conference (Conf. Proc.), Vol 3, TAPPI,
(1999), p 1047
A. Borg, A. Teder, and B. Warnqvist, TAPPI,
Vol 57 (No. 1), (1974), p 126
H. Tran, Chapter 9, Upper Furnace Deposition and Plugging, Kraft Recovery Boilers,
T.N. Adams, Ed., TAPPI Press, (1997),
p 247
H. Tran, M. Gonsko, and X. Mao, Effect
of Composition on the First Melting Temperature of Fireside Deposits in Recovery
Boilers, TAPPI J., Vol 82 (No. 9), (1999),
p 93
H. Tran, Recovery Boiler Plugging and
Prevention, TAPPI Kraft Recovery Operations Short Course Notes, TAPPI Press,
1992, p 209218
J.L. Clement and J.D. Blue, Recovery Furnace Floor Design and Alternative Materials, presented at Tenth Latin American
Recovery Congress (Concepcion, Chile),
Aug 2630, 1996
D. Singbeil, R. Prescott, J. Keiser, and
R. Swindeman, Composite Tube Cracking
in Kraft Recovery BoilersA State-OfThe-Art Review, 1997 Engineering &
Papermakers Conference (TAPPI Conf.
Proc.), Book 3, TAPPI Press, (1997), p 1001
J.R. Keiser et al., Analysis of Cracking
of Co-Extruded Recovery Boiler Floor
(a)
(b)
160 m
24 m
Fig. 13.21
Scanning electron micrographs showing (a) the

corrosion products formed on the T-11 superheater tube sample at the tube bend (Fig. 13.20) and (b)
the corrosion products on area C at high magnification.
EDX analysis showed essentially iron with tiny chromium
peak. Area C showed presence of chlorine (Cl) in addition to
iron. The x-ray spectra from area C are shown in Fig. 13.22.
Source: Ref 31
Tubes, 1997 Engineering & Papermakers

Conference (TAPPI Conf. Proc.), Book 3,
21. J.L. Barna and K.B. Rivers, Improving
Recovery Boiler Furnace Reliability with
Advanced Materials and Application Methods, presented at Canadian Pulp and Paper
Association Meeting, (Montreal, Quebec,
Canada), Jan 2529, 1999
31/10/2007 12:49PM Plate # 0
8000
Fe
7000
6000
Counts
5000
4000
3000
2000
1000
C
0
Fe
Cr
C1
Fe
Si
Cr
Cr Mn
0
0
10
11
12
13
14
15
keV
Fig. 13.22
X-ray spectra from area C, as shown in Fig. 13.21, showing a relatively high chlorine (Cl) peak. Source: Ref 31
(a)
Fig. 13.23
(b)
Type 310 overlay superheater tubes (a) and in close-up (b) after 2 years of operation in the boiler at a mill in South America.
Source: Ref 31
22. J.R. Keiser, L.M. Hall, K.A. Choudhury,

G.B. Sarma, J.P. Gorog, and R.E. Baker,
Thermal Behavior of Floor Tubes in a Kraft
Recovery Boiler, 1999 TAPPI Engineering/
Process and Product Quality Conference
(TAPPI Conf. Proc.), TAPPI Press, (1999),
p 1109
23. J.R. Keiser et al., Status Report on Studies of

Recovery Boiler Composite Floor Tube
Cracking, 1999 TAPPI Engineering/Process
and Product Quality Conference (TAPPI
Conf. Proc.), TAPPI Press, 1999, p 1099
24. R. Prescott and D.L. Singbeil, Stress Corrosion Cracking of Type 304L Stainless
pg 376
31/10/2007 12:49PM Plate # 0
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
Steel in Kraft Recovery Boiler Environments, Ninth International Symposium on

Corrosion in the Pulp and Paper Industry
(Conf. Proc.), CPPA (Montreal, Quebec,
Canada), (1998), p 185
H. Tran, B. Habibi, and C. Jia, Drying
Behavior of Waterwash Solution and the
Effect on Composite Floor Tube Cracking in
Recovery Boilers, 1999 TAPPI Engineering/
Process and Product Quality Conference
(TAPPI Conf. Proc.), TAPPI Press, (1999),
p 1061
J.R. Keiser et al., Why Do Recovery Boiler
Composite Floor Tubes Crack? 2000 TAPPI
Engineering Conference (TAPPI Conf.
Proc.), TAPPI Press, 2000
Sandvik Sanicro 38/4L7, S-12126-Eng,
Sandvik Steel, Sweden, June 1996
X.L. Wang, E.A. Payzant, B. Taljat,
C.R. Hubbard, J.R. Keiser, and M.J. Jirinec,
Experimental Determination of the Residual Stresses in a Spiral Weld Overlay
Tube, Mater. Sci. Eng. Vol A232, (1997),
p 31
P.N. Hulsizer, Dual Pass Weld Overlay
Method and Apparatus, U.S. Patent No.
6013890, Jan 2000
G. Lai, M. Jirinec, and P. Hulsizer, The
Properties and Characteristics of Unifuse
625 Overlay Tubing for Recovery Boiler
Applications, 1998 TAPPI Engineering
Conference (Conf. Proc.), Book 2, TAPPI
Press, 1998, p 417
G.Y. Lai and A. Wensley, Metallurgical
Characteristics and Performance Experience
of Spiral Overlay Tubes in Black Liquor
Recovery Boilers, 2005 TAPPI Engineering,
Pulping and Environmental Conference
(Conf. Proc.), TAPPI Press, 2005
A. Wilson, M. Lundberg, and U. Forsberg,
Alloy 825 Mod/SA210-A1 Composite Tube
for Black Liquor Recovery Boiler Floors,
1997 Engineering & Papermakers Conference (TAPPI Conf. Proc.), Book 3, TAPPI
Press, (1997), p 1043
D. Singbeil, Inspection for Cracking of
Composite Tubes in Black Liquor Recovery
Boilers, 2002 TAPPI Fall Conference (Conf.
H. Dykstra, N. Risebrough, and A. Wensley,
Corrosion and Cracking of Lower Furnace
Wall Tubes in Recovery Boilers, 1999
TAPPI Engineering/Process and Product
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
Quality Conference (TAPPI Conf. Proc.),

A. Wensley, Alternative Materials for Floor
and Lower Waterwall Tubes in Black Liquor
Recovery Boilers, 2000 TAPPI Engineering
Conference (Conf. Proc.), TAPPI Press,
2000
J.R. Keiser et al., Recent Observations
of Recovery Boiler Primary Air Port
Cracking and Characterization of Environmental Conditions, 2001 TAPPI Engineering/Finishing & Converting Conference
J.R. Keiser et al., Relationship of Recovery
Boiler Parameters and Primary Air Port
Cracking, 2002 TAPPI Fall Conference
A. Wensley and B. Woit, Inspection of
Primary Air Port Opening Tubes in Recovery Boilers, 2001 TAPPI Engineering/
Finishing & Converting Conference (Conf.
R. Shenassa, K. Haaga, and J. Tuiremo,
Primary Air Port Tube IntegrityA Critical
Review of Primary Air Port Design and
the Effect of Boiler Design Parameters,
2002 TAPPI Fall Conference (Conf. Proc.),
TAPPI Press, 2002
A. Wensley, presented at the International
Symposium on Corrosion in the Pulp and
Paper Industry, Charleston, SC, 2004
J.R. Keiser, private communication, 2006
J.R. Keiser et al., Causes and Solutions for
Cracking of Co-extruded and Weld Overlay
Floor Tubes in Black Liquor Recovery
Boilers, Ninth International Symposium on
Corrosion in the Pulp and Paper Industry
Welding Services Inc., unpublished data,
Norcross, GA
O. Moberg, P.E. Ahlers, and L. Dahl,
Recovery Boiler Corrosion, Pulp and Paper
Industry Corrosion Problems, NACE,
(1974), p 125
G. Lai and P. Hulsizer, Performance of Type
309 SS Overlay in the Lower Furnace of A
Black Liquor Recovery Boiler, 2001 TAPPI
Engineering Conference (Conf. Proc.),
TAPPI Press, 2001
H.N. Tran, D.C. Pryke, and D. Barham,
Local Reducing AtmosphereA Cause of
Superheater Corrosion in Kraft Recovery
Units, TAPPI, Vol 68 (No. 6), 1985, p 102
pg 377

DOI: 10.1361/hcma2007p379
31/10/2007 4:06PM Plate # 0

CHAPTER 14
Stress-Assisted Corrosion and Cracking

14.1 Introduction
Most high-temperature components are under
stress during service. The stress can be residual,
resulting from welding or forming operations
prior to service. When the temperature is not
high enough, these stresses can remain in the
component during service. The stress can also
come from externally imposed thermal stresses
or mechanical loading to which the component is
subjected during service. The stresses imposed
upon the component during service can affect the
behavior of high-temperature corrosion attack
or cause intergranular cracking.
One important subject area is the effect of
tensile stresses (or strains) on high-temperature
corrosion attack. Tensile stresses that are imposed on the component during service can cause
preferential corrosion penetration. As this preferential corrosion penetration grows into the
metal interior, it can eventually develop into a
crack. One important industrial example involving this phenomenon is the circumferential
cracking that occurs on the waterwall tubes of
some supercritical coal-red boilers, which are
red under low NOx combustion conditions.
The low NOx combustion produces suldizing
environments at or near the waterwall of the
boiler. The waterwall tubes, which are made of
a carbon or low-alloy steel, or even with a stainless steel or nickel-base alloy weld overlay that
protects the waterwall steel tubes, can suffer
circumferential cracking in some boilers. This
particular phenomenon (i.e., circumferential
cracking of the waterwall) is covered in Chapter
10 Coal-Fired Boilers. In that chapter, the
correlation between the circumferential cracking
and the stress-assisted preferential corrosion
penetration is discussed. The current chapter
discusses the phenomenon of the preferential
high-temperature corrosion penetration that
forms under the combination of stresses and corrosive conditions, primarily in suldizing environments.
The other subject area discussed in this chapter

is stress-induced cracking. This is a brittle
fracture with cracks propagating along grain
boundaries. This is often referred to as stressrelaxation cracking, reheat cracking, or
strain-age cracking in the literature. Typically,
a highly constrained component, such as a
heavy wall construction, or a welded component,
or a cold-worked structure, can be susceptible to
this type of intergranular, brittle cracking for
some alloys when exposed to lower end of the
intermediate temperature range, such as 480 to
700 C (900 to 1290 F). Alloys that are susceptible to this type of intergranular fracture
include ferritic steels, austenitic stainless steels,
Fe-Ni-Cr alloys, and nickel-base alloys. Some
[Ni3Al] strengthened nickel-base alloys can
be susceptible to this type of embrittlement at
temperatures higher than 700 C (1290 F). This
type of cracking can also occur during reheating
of the component in a postweld heat treatment
or annealing/stress-relieving. Weldments or weld
joints are particularly susceptible to this type of
cracking.
Some components can be subject to thermal
cycling during service. The components that are
subject to thermal cycling are generally more
easily understood from the operating standpoint.
When both stress magnitudes (or strain magnitudes) and frequencies are high enough, the metal
can eventually fail by thermal fatigue cracking.
The morphology of thermal fatigue cracking is
generally very well dened and easily identied.
The subject of thermal fatigue cracking has been
extensively discussed in the literature and is not
covered in this chapter.
14.2 Stress-Assisted Preferential

Corrosion Penetration
In most industrial environments, alloys rely
on the formation of a continuous oxide scale
pg 379
31/10/2007 4:07PM Plate # 0
for protection against accelerated corrosion

attack during service at elevated temperatures.
Most commercial alloys rely on chromium oxide
(Cr2O3) scales. The majority of the corrosion
data have been generated from test specimens
that were not subjected to external stresses
during exposure. However, as discussed earlier,
most high-temperature components are subject
to external stresses during plant operation.
When a metal with a protective oxide scale
is under tensile stress, the oxide scale may crack
as the tensile stresses or strains become high
enough. Figure 14.1 shows cracking of the
oxide scale on alloy 800 after testing at 800 C
(1470 F) in air under the strain rate of 106 s1
(Ref 1). Under certain conditions, these cracks
can be rehealed. This is illustrated in Fig. 14.2,
showing a healed oxide crack on alloy 800 at
800 C (1470 F) in air under the strain rate
of 108 s1 (Ref 2). Schtze (Ref 2) observed
that cracking of the oxide scale on a deforming
metal can cause signicant internal corrosion
penetration, as shown in Fig. 14.3 for Fe-18Cr
steel (with 0.8Al and 1.5Si) deforming at 800 C
(1470 F) in air under a strain rate of 108 s1.
The gure shows that once the oxide scale
reached the initial threshold strains for developing cracking, further deforming of the steel
caused signicant internal oxidation penetration.
On the other hand, the nondeformed specimen
(i.e., the specimen that was not under creep
deformation during the exposure) showed no
changes in the depth of the internal oxidation
penetration during the exposure for times up

to 1000 h. The internal oxidation penetration
in the deforming Fe-18Cr-0.8Al alloy specimen
was found to be essentially in the form of nitridation attack involving aluminum nitrides
(Ref 1). (Nitridation attack in air or combustion
environments is discussed in Chapter 4 Nitridation.)
Oxidation in air or oxidation in general is
considered to be the least corrosive among
various high-temperature corrosion modes. As
the corrosivity of the environment increases, the
effect of the external tensile stresses on preferential corrosion penetration becomes increasingly more signicant. This section focuses on
the stress-induced preferential corrosion penetration in suldizing environments.
Smolik and Flinn (Ref 3) examined the effects
of stresses on suldation of alloy 800H in coal
gasication environments, which were characterized with high sulfur and low oxygen potentials ( pS2 and pO2 ). Tests involved exposing the
outer diameter of the alloy 800H tube to three
test environments, which were Ar-23%H2O10%H2 with 0.1, 0.2, and 0.4% H2S, respectively. The corresponding partial pressures of
oxygen ( pO2 ) and sulfur ( pS2 ) at the test temperature of 870 C (1600 F) were 2 1018 and
1 108 atm for the 0.1 H2S test environment, 2 1018 and 5 108 atm for the 0.2 H2S
environment, and 2 1018 and 2 107 atm for
the 0.4 H2S environment. The circumferential
strain was created by pressurizing the test
tube internally with argon. Preferential corrosion
penetration was observed to take place along
Fig. 14.1
Fig. 14.2
Cracking of the oxide scale on alloy 800 during

deformation at the strain rate of 106 s1 at 800 C
(1470 F) in air. Source: Ref 1
Rehealed oxide-scale crack on alloy 800H during

deformation at the strain rate of 108 s1 at 800 C
(1470 F) in air. Source: Ref 2
pg 380
31/10/2007 4:07PM Plate # 0
Chapter 14:
grain boundaries. Both oxides and suldes were

found to form along grain boundaries. This intergranular oxidation/suldation penetration eventually affected the circumferential strains to
failure for the test tubes. The higher H2S concentration produced deeper penetrations, thus
resulting in smaller strains to failure. The circumferential strains to failure were found to be
about 8 to 10% for the 0.1% H2S, about 6 to 7%
for the 0.2% H2S, and 1 to 4% for the 0.4% H2S.
Researchers at the Joint Research Centre,
Petten Establishment, The Netherlands, have
conducted a series of extensive studies on the
effects of deformation on preferential corrosion in reducing, suldizing environments (i.e.,
simulated coal gasication environments) (Ref
48). Some of their results are summarized in this
section.
In their study of the effect of the creep deformation on the corrosion of alloy 800H in
pg 381
Stress-Assisted Corrosion and Cracking / 381
coal gasication environments, Guttmann and

Timm (Ref 4) observed that creep deformation
markedly accelerated preferential corrosion
penetration. Creep tests were conducted at
800 C (1470 F) in H2-1.2H2O-7CO-0.4H2S
(5 1022 atm pO2 , 5109 atm pS2 , and 0.3 ac)
and H2-1.2H2O-0.4H2S (51022 atm pO2 ,
4 109 atm pS2 ). Exposure tests were also conducted in the same test environments with
no applied stresses for comparison of the corrosion behavior. Under the stressed condition preferential corrosion penetration was found to be
deepest at grain boundaries that were normal to
the stress direction. Samples in the stress-free
condition showed shallow corrosion depths
following grain or twin boundaries without preferential penetration. Preferential intergranular
corrosion penetration was found to consist of
essentially suldation/oxidation. It was found
that the suldation/oxidation penetration in the
= 108s1
Without deformation
150
Scale
cracking
x i, m
= 0.7%
100
50
500
1000
Time, h
Fig. 14.3
Depth of internal corrosion penetration for Fe-18Cr-0.8Al-1.5Si between the undeformed specimens (open data points) and
the deforming specimens (solid data points) under the strain rate of 108 s1 at 800 C (1470 F) in air. Source: Ref 2
31/10/2007 4:07PM Plate # 0
Depth of corrosion, m
stressed specimens was about 2 to 3 times more

than that of the unstressed specimens. This
is illustrated in Fig. 14.4 (Ref 4). For the COcontaining environment, specimens in both
stressed and unstressed conditions suffered
carburization as well as suldation/oxidation.
Under these conditions, carburization dominated
the corrosion penetration depth; the authors
found no signicant effect on the corrosion depth
when the corrosion attack involved carburization
(Fig. 14.4). In examination of the surface cracks
of the specimens tested in air and those tested
in the suldizing environments, Guttmann and
Timm (Ref 4) found that the surface cracks,
which became blunted in air, were deeply penetrating and long and sharp with severe corrosion
at the crack tip and in the crack vicinity. Stroosnijder et al. (Ref 5) made similar observations
when comparing the creep deformation behavior
of alloy 800 at 700 C (1290 F) between air
and H2-1.2H2O-7CO-0.2H2S (2 1025 atm pO2 ,
2 1010 atm pS2 ). These authors observed
blunted surface cracks when tested in air and
deeply penetrating sharp cracks along grain
boundaries with deep corrosion paths in front of
cracks when tested in the suldizing environment. The corrosion products formed at the grain
boundary in the suldizing environment were
found to consist of oxides and suldes.
The above discussion focuses on the effects
of relatively large deformation (i.e., large strain
or high strain rate) during creep testing. For many
103
102
No load Load
Sulfidation
Oxidation
Carburization
10
10
102
103
104
Exposure time, h
Fig. 14.4
Corrosion depth for unstressed and stressed specimens of alloy 800H tested at 800 C (1470 F)
in H2-1.2H2O-0.4H2S (5 1022 atm pO2 , 4 109 atm pS2). Also
included was the data generated from H2-1.2H2O-7CO-0.4H2S
(5 1022 atm pO2 , 5 109 atm pS2 , and 0.3 ac). The latter
environment caused carburization in addition to sulfidation/
oxidation. Source: Ref 4
structural components during service at high

temperatures, the strains imposed on the component are likely to be much smaller. The design
strain rate (or creep rate) of ASME Boiler and
Pressure Vessel Codes is 105% h1 (2.8
1011 s1) (Ref 9). It is therefore important to
examine the effects of low tensile stress (or
strain) on the preferential corrosion penetration.
Stroosnijder et al. (Ref 6) investigated the
effect of strain on preferential corrosion of
alloy 800H when tested at 800 C (1470 F)
in H2-1.2H2O-7CO-0.4H2S (5 1022 atm pO2 ,
5 109 atm pS2 , and 0.3 carbon activity, ac)
with (a) no external stress (or strain) during
testing and (b) strains up to 4% and strain
rates down to 109 s1. In this study, alloy 800H
specimens were surface treated with CeO2 using
cerium sol-gel techniques. The authors observed
that the unstressed specimens showed some
external and internal corrosion with no evidence
of preferential corrosion penetration after 663 h
of exposure (Fig. 14.5a). For stressed specimens
after 663 h and subjected to about 2.3% strain
under a strain rate of 108 s1, preferential corrosion penetration along the grain boundary
began to take place (Fig. 14.5b). After testing
for 1247 h with about 5% strain (108 s1 strain
rate), intergranular corrosion attack along grain
boundaries was found to penetrate farther into
the metal interior (Fig. 14.5c).
Guttmann et al. (Ref 7) examined the effect
of a much lower strain range (typically 1 to 2%)
on the preferential corrosion penetration in a
suldizing environment. The strains the authors
examined were much closer to the strains that
might be experienced by operating hightemperature components. Tests were conducted
at 600 C (1110 F) in CO-32H2-4CO2-0.2H2S.
The strain rates during creep were 1010 s1 to
108 s1. Similar to what was observed earlier
in alloy 800H (Fe-33Ni-2Cr-0.4Al-0.4Ti)
(Ref 46), Guttmann et al. (Ref 7) observed
that alloy HR3C (Fe-20Ni-25Cr-0.5Nb-0.2N)
showed stress-assisted intergranular corrosion
attack under the test conditions. This is illustrated
in Fig. 14.6 (Ref 7). The corrosion products
formed at grain boundaries were enriched in
chromium, oxygen, and sulfur, but depleted in
nickel and iron. The intergranular corrosion
attack in alloy HR3C tested after 1810 h with
2.2% strain is shown in Fig. 14.7 for the scanning
electron microscopy (SEM) backscattered electron image of the intergranular corrosion attack
and the x-ray maps for chromium, oxygen, and
sulfur. The intergranular corrosion attack was
pg 382
31/10/2007 4:07PM Plate # 0
Chapter 14:
(a)
(b)
pg 383
(c)
Corrosion morphology of alloy 800H tested at 800 C (1470 F) in H2-1.2H2O-7CO-0.4H2S (5 1022 atm pO2 , 5 109 atm
pS2 , and 0.3 ac) (a) after 663 h under no external stresses, (b) after 663 h at about 2.3% strain under the strain rate of 108 s1,
and (c) after 1247 h under strain rate of 108 s1. Note: the specimen surface was plated prior to the mounting of the sample to retain the
corrosion products. Source: Ref 6
Fig. 14.5
(a)
25 m
(b)
25 m
Fig. 14.6
Strain-assisted intergranular corrosion attack in alloy HR3C after testing at 600 C (1110 F) for 250 h in CO-32H2-4CO20.2H2S with (a) 1.3% strain and (b) 2% strain. Corrosion products formed on the metal surface were also observed. Note:
the tested specimen surface was plated prior to the mounting of the metallographic sample to retain the surface corrosion products. Source:
Ref 7
oxidation/suldation, similar to the earlier

observations in alloy 800H. For nonstressed
specimens, the alloy suffered external suldation
with Fe- or Ni-rich suldes ([Ni,Fe]9S8) and
Ni3S2. Underneath the surface sulde scales
were ne internal corrosion phases randomly
distributed within the grain matrix with no deep
penetrating intergranular corrosion attack, as
shown in Fig. 14.8.
Both alloys 800H and HR3C are Fe-Ni-Cr
alloys with austenitic microstructure. Both alloys
were found to suffer preferential corrosion
penetration along grain boundaries. It is interesting to note that nickel-base alloys with the
same austenitic microstructure would follow a
completely different morphology in preferential

corrosion penetration under stress. In fact, the
morphology of stress-assisted preferential corrosion attack for nickel-base alloys tends to be
similar to that for iron-base alloys having a
ferritic structure. Guttmann et al. (Ref 7) tested
nickel-base alloy 45TM (Ni-27Cr-23Fe-2.76Si)
under the same test condition as for alloy HR3C.
Alloy 45TM exhibited a ngerlike preferential
corrosion penetration, as shown in Fig. 14.9
(Ref 7). The corrosion products formed in the
preferential corrosion penetration were found to
be enriched in chromium, oxygen, sulfur, and
silicon, as shown in Fig. 14.10. By contrast, the
unstressed specimens were found to exhibit a
31/10/2007 4:07PM Plate # 0
Fig. 14.7
Scanning electron backscattered images showing intergranular corrosion penetration along with the x-ray maps for Cr, O,
and S for alloy HR3C after testing at 600 C (1110 F) for 1810 h in CO-32H2-4CO2-0.2H2S with 2.2% strain. Source: Ref 7
Fig. 14.8
Corrosion products formed on the metal surface as

well as internal corrosion products formed in alloy
HR3C with no external strain imposed on the test specimen when
tested at 600 C (1110 F) for 2100 h in CO-32H2-4CO2-0.2H2S.
Note: the tested specimen surface was plated prior to the
mounting of the metallographic sample to retain the surface corrosion products. Source: Ref 7
fairly uniform surface corrosion attack with no

preferential corrosion penetration, as shown in
Fig. 14.11.
In the same study, the authors tested an
iron-base ferritic oxide-dispersion-strengthened
Fig. 14.9
Stress-assisted preferential corrosion penetration

for alloy 45TM tested to 2% strain at 600 C
(1110 F) for 1820 h in CO-32H2-4CO2-0.2H2S. Note: the tested
specimen surface was plated prior to the mounting of the metallographic sample to retain the surface corrosion products. Source:
Ref 7
(ODS) alloy, MA956 (Fe-20Cr-4.6Al-0.5Y2O3),

under the same condition and found that this
iron-base alloy with ferritic structure exhibited
pg 384
31/10/2007 4:07PM Plate # 0
Chapter 14:
a ngerlike preferential corrosion penetration

under stresses, as shown in Fig. 14.12. The result
of x-ray elemental mapping (Fig. 14.12) reveals
that (a) the external surface layer was ironrich suldes, (b) the inner surface layer was
chromium- and aluminum-rich oxides along
with some suldes, and (c) the inner surface
layer (Cr-Al rich oxides) also extended to the
preferential corrosion penetration as an outer
layer with a center core of iron-rich suldes that
were extended from the external surface layer
discussed in (a). When tested under a similar
Fig. 14.10
condition but with no external stresses, alloy

MA956 showed only uniform corrosion attack,
as shown in Fig. 14.13. For the unstressed
specimen, the external corrosion layer was found
to be iron-rich suldes and inner layer consisted
of (Fe,Cr)1xS, Cr2O3, and Al2O3. Ferritic
stainless steels appeared to follow the similar
ngerlike preferential corrosion penetration
when the specimen was under creep deformation,
as shown in Fig. 14.14 (Ref 8). This gure also
shows that the specimen suffered essentially
uniform attack when under no external stress.
Scanning electron micrograph for the stress-assisted preferential corrosion penetration in alloy 45TM tested to 4%
strain at 600 C (1110 F) for 2000 h in CO-32H2-4CO2-0.2H2S and elemental x-ray maps for the corrosion products
in chromium, oxygen, sulfur, silicon, and iron. Source: Ref 7
pg 385
31/10/2007 4:07PM Plate # 0
The authors did not report the percent of strain

when the test was terminated and the specimen
was examined; however, the test duration was
only 615 h, and the strain was likely to be high.
Upon closer examination of Fig. 14.14, it appears
that a surface crack formed in the center of each
corrosion penetration, further indicating that the
strain might be large.
Preferential corrosion penetration under
strain (or stress) was also observed in other
aggressive environments. The stress-assisted
preferential corrosion attack is quite sensitive to
the environment. Le Calvar et al. (Ref 10)
performed low strain rate tests in air + 4CO2 +
8H2O at 610 C (1130 F). Figure 14.15 shows
Type 304H suffering preferential oxidation
penetration under this test condition at a strain
rate of 3108/s with total strain of about 2%.
The authors referred to this preferential oxidation attack as cracking in their paper (Ref 10).
Similar tests were also performed in a vacuum
environment, and the results showed
no cracking. The gure suggests that the
phenomenon Le Calvar et al. observed was
more like stress-assisted preferential corrosion
penetration than stress-assisted cracking followed by formation of the corrosion products
inside the crack.
Rorbo (Ref 11) reported that the external
stresses in Type 304 caused cracks to develop
in the nitrided layer, and the cracks resulted in
increased nitridation attack in front of the crack.
This is illustrated in Fig. 14.16. The nitrided
layer is believed to be very brittle, and very little
strain is required to develop cracking in the
nitrided layer. In this case, cracking could
have developed in the nitrided layer and caused
preferential nitridation penetration in front of
the crack.
Fig. 14.11
Corrosion morphology for alloy 45TM tested

under no external strains at 600 C (1110 F)
for 2000 h in CO-32H2-4CO2-0.2H2S. Source: Ref 7
The waterwalls in some coal-red boilers,

particularly some supercritical units, have been
found to encounter preferential corrosion penetration. These supercritical boilers are typically
equipped with low NOx burners and overre
air, thus creating localized reducing, suldizing
environments at or near the waterwall in the
lower furnace. Furthermore, the waterwalls can
also be subjected to high heat ux and possibly
ame impingement, thus resulting in higher
metal temperatures and higher thermal stresses.
The combination of these conditions leads to
the development of stress-assisted preferential
suldation penetration. As the preferential suldation penetration continues to grow into
the metal interior, it eventually develops into a
crack. At later stages, these cracks resemble
thermal fatigue cracks. These cracks, which
are in the transverse direction with respect to
the tube axis, are commonly referred to as
circumferential cracks. Typical alloys of
construction have experienced circumferential
cracking. The commonly used weld overlay
cladding alloys including Type 309 and alloy 625
have also been observed to suffer circumferential
cracking. This phenomenon is covered in detail
in Chapter 10 Coal-Fired Boilers (Section
10.5.3).
The morphologies of the preferential suldation penetration are strikingly similar between
what has been observed in the waterwall of coalred boilers and that observed in test specimens
under tensile stresses (or low strain rate creep
tests) in laboratory suldizing environments.
Figure 10.32 in Chapter 10 shows an example
of preferential suldation penetrations observed
on T-22 (2.25Cr-1Mo) in a supercritical coalred boiler equipped with low NOx burners
and overre air. These preferential suldation
penetrations are considered to be precursors to
circumferential cracks observed from tubes that
failed or were about to fail. The preferential
suldation penetration often exhibits channels
in its core, as shown in Fig. 14.17. These
channels, which generally exhibited lighter color
(Fig. 14.17), were found to consist of suldes
(see Fig. 10.33 and 10.34 in Chapter 10). In some
cases, these channels were much more pronounced and numerous, such as the one shown
in Fig. 14.18.
The preferential suldation penetration
formed in steels was found to consist of essentially iron oxides in the outer region of the
penetration with the inner region (core) including channels being essentially iron suldes
pg 386
31/10/2007 4:07PM Plate # 0
Chapter 14:
(see the SEM/EDX analysis in Fig. 10.33 and

10.34 in Chapter 10). This is illustrated schematically in Fig. 14.19.
Proposed reaction steps involved in developing preferential suldation penetration in carbon
or low-alloy steel waterwall tubes are:
1. Iron oxide scales initially form on the steel
surface.
Fig. 14.12
2. Tensile stresses cause the breakdown of the

iron oxide scales.
3. The surface oxide scale breakdown initiates
the development of a preferential oxidation
penetration.
4. As the preferential oxidation penetration
continues to grow inward into the metal
interior, the inner region of the oxide penetration (i.e., penetration core) becomes more
Scanning electron micrograph in a backscattered electron image along with x-ray maps for Cr, O, S, Al and Fe showing
the stress-assisted preferential corrosion penetration (fingerlike penetration attack) on alloy MA956 tested to 5% strain at
600 C for 1830 h in CO-32H2-4CO2-0.2H2S. Source: Ref 7
pg 387
31/10/2007 4:07PM Plate # 0
pg 388
decient in oxygen, thus lowering oxygen

potentials in the core region.
5. Iron suldes begin to form in the core region
due to insufcient oxygen.
6. As the penetration continues to grow farther
into the metal interior, oxidation/suldation
processes continue to penetrate inward into
the metal interior with oxidation forming in
the outer region and suldation in the inner
region of the corrosion penetration.
7. Channels are likely created during the growth
of sulde phases in the core region under
the tensile stresses. These channels may
also serve as gas passages as the oxidation/
suldation penetration grows deeper into the

metal interior.
For a Ni-Cr-Fe alloy, the preferential suldation penetration under large strains could develop
a large core of iron suldes in the inner region
with chromium oxides formed in the region
next to the unaffected metal, such as in the case
for alloy 45TM tested to 4% strain at 600 C
(1110 F) for 2000 h in CO-32H2-4CO2-0.2H2S,
as shown in Fig. 14.10 (Ref 7). Because of a
large tensile strain (4%), the entire center core
became a one big channel. However, when the
strain was reduced to about 2%, alloy 45TM
showed one big channel through the external
surface scale followed by a number of ne
channels in the preferential suldation penetration, as shown in Fig. 14.9. For MA956
Fig. 14.13
Alloy MA956 tested under no external strain

(or stress) at 600 C (1110 F) for 2000 h in
CO-32H2-4CO2-0.2H2S, showing uniform corrosion with no
preferential corrosion penetration. Source: Ref 7
10 m
Fig. 14.14
Fe-12Cr-3Al-3Ti showing preferential corrosion

penetration for the specimen under creep deformation at 600 C (1110 F) in H2-34.3H2O-18.5CO2-3.8CH47.9CO-1.3H2S for 615 h (upper figure) and uniform corrosion
attack for the specimen under no external stress (or strain)
after exposure to the same test environment and duration (lower
figure). Source: Ref 8
Fig. 14.15
Type 304H tested in air + 4CO2 + 8H2O at 610 C

(1130 F) under the strain rate of 3 108/s with
about 2% strain, showing preferential oxidation penetration
(the authors referred to as cracking). SEM/EDX analysis showed
the oxide in area A was Fe-3Cr-4Si-0.5Mo, in area B was
Fe-29Cr-5Ni-0.8Mo-0.8Si, and in area C was Fe-37Cr-21Ni1.5Si. Source: Ref 10
31/10/2007 4:07PM Plate # 0
Chapter 14:
pg 389
(a Fe-Cr-Al alloy), the entire core became a one

big channel consisting of iron suldes with the
outer region consisting of a mixture of chromium
oxides and aluminum oxides at 5% strain, as
shown in Fig. 14.12 (Ref 7). For operating
components, such as waterwalls in boilers, the
strains imposed onto the component during
operation are likely to be much lower. As shown
in Fig. 14.20, alloy 625 weld overlay on the
waterwall of a supercritical unit suffered preferential suldation penetration, which exhibited
ne channels. The gure also shows no evidence
of the formation of a crack near or at the tip
Fig. 14.16
Cracks developed in the nitrided layer formed in

Type 304 due to external stresses and accelerated
nitridation attack in front of the crack. Source: Ref 11
Iron sulfides (higher sulfur)

Iron oxides
Iron sulfides
Iron oxides
Carbon or low-alloy steel
Fig. 14.19
Fig. 14.17 Preferential sulfidation penetration (a precursor of
the circumferential cracking) observed on a T-22
(2.25Cr-1Mo) wingwall tube of a supercritical boiler, showing
channels (light color stringers) in the core of the penetration.
Optical micrograph showing numerous channels in a circumferential groove formed on a

T-11 (1.25Cr-0.5Mo) tube in a supercritical coal-fired boiler.
Fig. 14.18
Fig. 14.20
Preferential sulfidation penetration formed in

waterwall steel tubes
Scanning electron (backscattered electron) image

showing a circumferential groove formed in alloy
625 weld overlay on a waterwall tube of a supercritical coal-fired
boiler. Courtesy of Welding Services Inc.
31/10/2007 4:07PM Plate # 0
of the preferential suldation penetration. The

corrosion products were essentially chromium
suldes with nickel-rich suldes forming the
channels. The results of the SEM/EDX analysis
of the corrosion products are shown in Fig. 10.37
in Chapter 10. Figure 14.21 shows another example of a circumferential groove that formed
in the alloy 625 weld overlay on the waterwall
tube after 6 years of service in another supercritical coal-red boiler. The circumferential
groove was found to contain essentially nickeland chromium-rich suldes with numerous light
grayish channels. Some light grayish channels
were nickel-rich suldes. Suldes in the groove
were also found to contain lead (Pb), arsenic
(As), and zinc (Zn), which are common impurities in coal ash. Fine channels were also
observed in the preferential suldation penetrations formed in the 309SS weld overlay on the
waterwall, as shown in Fig. 10.40 and 10.41 in
Chapter 10. In the waterwall, the 309SS overlay
(Fe-Cr-Ni alloy) suffered preferential suldation
80 m
Fig. 14.21
Scanning electron (backscattered electron) image

showing a circumferential groove formed in alloy
625 weld overlay on a waterwall tube after 6 years of service in a
supercritical coal-fired boiler under low NOx combustion. The
compositions (wt%) of the phases at different locations were
analyzed by EDX, showing either nickel-rich or chromium-rich
sulfides. Some sulfides were also found to contain Pb, As, and Zn
that are commonly found in coal ash. Note numerous light grayish
channels. Some of these channels (No. 3 and 4) were nickel-rich
sulfides. Courtesy of Welding Services Inc.
1: 56% Ni, 13.9% Cr, 6.2% Fe, 18.7% S, 2.5% Zn, 1.4% As
2: 53.8% Cr, 10.2% Ni, 7.0% Fe, 14.6% S, 9.4% Pb, 2.6% Zn
3: 58.9% Ni, 5.8% Cr, 4.6% Fe, 26.1% S, 1.9% As, 1.0% Pb
4: 56.4% Ni, 5.1% Fe, 3.3% Cr, 6.6% Mo, 24.4% S, 1.9% Pb
5: 52.5% Cr, 18.4% Mo, 7.5% Fe, 3.8% Ni, 9.5% Mo, 5.5% S
6: 18.7% Ni, 17.8% Cr, 13.7% Nb, 13.7% Mo, 21.3% Pb,
22.1% S, 4.6% Fe
penetration attack along dendritic boundaries,

as shown in Fig. 10.40 in Chapter 10. This is
quite similar to the wrought alloys, such as
HR3C (Fe-Ni-Cr alloy), that suffered preferential suldation penetrations along grain boundaries when tested in laboratory environments
(Fig. 14.6).
14.3 Stress-Assisted Intergranular

Cracking
When external stress is applied to a metallic
component, the metal rst undergoes elastic
deformation. With increasing applied stress, the
metal then deforms plastically and eventually
reaches the limit of its plastic deformation. At
this point, with further increase of the applied
stress, the metal can no longer deform to relieve
the stresses built up in the metal, thus resulting
in the initiation of cracks at the weakest locations
to relieve those stresses. An engineering alloy
can normally undergo a large plastic deformation
before developing cracks. When the metallic
component is subject to external stresses at
elevated temperatures, the metal undergoes creep
deformation, which is a time-dependent deformation governed by diffusion. An engineering
alloy typically exhibits a large creep deformation
prior to developing cracks and nal rupture.
However, as the temperature decreases, the creep
strain at rupture decreases.
Under certain conditions, a metal can only
sustain very little deformation (plastic and/or
creep deformation) before developing brittle
fracture. The fracture typically takes the form
of intergranular cracking (i.e., cracking along
grain boundaries). This brittle, intergranular
fracture has occurred during heat treatments or
service for some engineering alloys, including
ferritic steels, austenitic stainless steels, Fe-Ni-Cr
alloys, and nickel-base alloys. This type of
brittle, intergranular cracking phenomenon has
been described by different names, such as
reheat cracking, stress-relaxation cracking,
strain-age cracking, and gamma-prime embrittlement. This phenomenon typically occurs
at the lower end of the intermediate temperature
range, such as 480 to 700 C (900 to 1290 F).
14.3.1 General Conditions That Cause
Stress-Assisted Cracking Embrittlement
Several material conditions can increase the
susceptibility of the stress-assisted cracking
pg 390
31/10/2007 4:07PM Plate # 0
Chapter 14:
embrittlement for an alloy. An alloy in a coldworked condition, which exhibits higher yield
strength and higher hardness, and lower ductility,
is much more susceptible to this type of brittle
intergranular cracking. Figure 14.22 shows an
example of the effect of cold work on hardness
and yield strength of several nickel-base alloys
and the corresponding effect on its tensile
elongation (Ref 12). The hardness and yield
strength of an alloy can be signicantly increased by increasing the amount of cold work.
The yield strength of a cold-worked metal also
represents the residual stress in the cold-worked
structure. An alloy with a larger amount of cold
work exhibits a higher residual stress. Higher
residual stresses in a structural component can
increase the susceptibility to stress-relaxation
cracking. After the alloy is strengthened by
cold working, the capability of the grain matrix
to plastically deform to relieve additional external stresses imposed on the metal during service
at intermediate temperatures is signicantly
reduced. When these external stresses cannot
be relieved through deformation, cracking then
develops at grain boundaries to relieve those
stresses.
The matrix of an alloy can also be strengthened by precipitates, particularly ne, coherent
precipitates (e.g., [Ni3(Al,Ti)], [Ni3Nb],
Ni2Mo ordered phases). Figure 14.23 shows
ne, coherent [Ni3Nb] precipitates formed
in the alloy matrix (i.e., grain interior) of alloy
625 (Ni-22Cr-9Mo-3.5Nb) at 650 C (1200 F)
for 24 h (Ref 13). Examples of ne precipitates
formed in other alloy systems are shown in
Fig. 14.24 for Ni2(Cr,Mo) ordered phases formed
in Ni-16Cr-15Mo-3Fe alloy, Fig. 14.25 for
(Ni3Al) precipitates formed in alloy 214, and
Fig. 14.26 for (Ni3Al) precipitates in alloy 601.
Precipitation of these ne, coherent precipitates
not only strengthens the grain matrix but also
causes volume contraction that causes additional
internal stresses in the alloy, making it more
susceptible to stress-relaxation cracking. Carbides, such as TiC, NbC, and Cr23C6, also
produce some alloy strengthening and can
increase the susceptibility to stress-relaxation
cracking. The combination of a cold-worked
structure and ne precipitates in the matrix can
cause the material to be extremely susceptible to
stress-relaxation cracking. Formation of precipitates, such as Cr23C6, at grain boundaries can
further promote grain-boundary cracking, thus
increasing the susceptibility of stress-relaxation
cracking.
Another factor that is important in increasing the susceptibility of stress-assisted cracking

embrittlement is the surface geometry of the
component, which can lead to stress risers such
as an abrupt thickness change in a component. A
component that is under a highly constrained
condition, such as a heavy section part, is most
susceptible to this type of cracking embrittlement
since the additional stresses imposed on the
component render it incapable of accomodating
any dimensional change or deformation, and it
develops cracks to relieve those stresses.
A majority of cracking embrittlement cases
occur under two different conditions. The cracking can occur during heat treatment, such as
a postweld heat treatment (PWHT) for a
weldment, which involves a short duration. The
cracking can also occur during service, which
involves a relatively longer duration, but still
relatively short in terms of the design life of the
component, with most failures occurring after
1 to 2 years of service or less. In both cases, the
metal (or the metallic component) has undergone
very little deformation prior to intergranular
cracking fracture.
An example of heat-treatment-induced cracking embrittlement is described as follows. When
welding is performed on a metal, residual
stresses are developed in the heat-affected zone
(HAZ) of the base metal due to the volume
shrinkage of adjacent weld metal changing from
the molten state to the solid state. The maximum
residual stress is approximately the roomtemperature yield strength of the metal. When
this weldment receives a PWHT, the yield
strength of the metal at this particular PWHT
temperature would be lower than the residual
stresses (i.e., the room-temperature yield strength
of the metal), thus causing the metal to undergo deformation to accommodate this additional
stress. However, when the metal cannot deform
to relax or relieve this additional stress
through deformation, cracks develop at the
weakest locations, typically grain boundaries,
to relieve this additional stress, thus resulting
in this intergranular cracking embrittlement.
The service-related cracking embrittlement
typically occurs when the service temperatures
are relatively low such that creep deformation
is too low to relieve the applied stresses, and the
alloy matrix is signicantly strengthened such
that these stresses are not able to be relieved
through deformation, thus resulting in cracking
for the relief of this applied stress. Detailed discussion on stress-relaxation cracking is presented
pg 391
31/10/2007 4:07PM Plate # 0
Hardness, HRC
50
10
20
30
40
50
60
50
40
40
30
30
20
20
Alloy 25
Alloy 188
Alloy 625
230 Alloy
Alloy X
10
10
20
30
10
40
0
60
50
Cold work, %
(a)
0
10
20
30
40
50
60
200
1350
1100
150
125
850
350
10
20
30
40
75
50
60
50
Cold work, %
(b)
70
Elongation, %
100
Alloy 25
Alloy 188
Alloy 625
2230 Alloy
Alloy X
600
10
20
30
40
50
60
50
50
Alloy 25
Alloy 188
Alloy 625
230 Alloy
Alloy X
40
30
40
30
20
20
10
10
Fig. 14.22
60
70
60
0
(c)
0.2 % yield strength, ksi
0.2 % yield strength, MPa
175
10
20
30
40
50
0
60
Cold work, %
Effects of cold work on hardness (a) room-temperature yield strength (b) and the corresponding tensile elongation (c) for
several nickel-base alloys. Source: Ref 12
pg 392
31/10/2007 4:07PM Plate # 0
Chapter 14:
in the following sections based on the alloy

systems.
14.3.2 Cracking in Ferritic Steels
Cracking of CrMo and CrMoV steel weldments in steam pipework and valve assemblies
during stress-relief heat treatment (e.g., postweld
heat treatment) and high-temperature service
have been encountered in power-generating
equipment (Ref 17). Similar reheat cracking has
also been reported in the heat-affected zone
(HAZ) of low-alloy steels for pressure vessels
during hydrostatic pressure tests (Ref 17). The
phenomenon of reheat cracking also received
much attention when small cracks in the HAZ
under the austenitic alloy cladding, which were
referred to as underclad cracking, were encountered in nuclear pressure vessels in the

1970s (Ref 17). Other publications discussing
reheat and underclad cracking of ferritic steels
in the 1970s can be found in Ref 18 to 21.
Dhooge and Vinckier (Ref 17) provided a
detailed review of the factors that might be
responsible for causing reheat cracking of
ferritic steels. These factors included intragranular precipitation hardening that strengthens
the grain matrix and segregation of impurities to
grain boundaries that weakens grain boundaries.
Segregation of impurities to grain boundaries
may play a signicant role in reheat cracking
Fig. 14.25
Fig. 14.23
Transmission electron micrograph showing

fine, coherent c 00 (Ni3Nb) precipitates formed in
the grain matrix of alloy 625 at 650 C (1200 F) for 24 h. Source:
Ref 13
Fig. 14.24
Transmission electron micrograph showing a

dark-field image of fine, Ni2(Cr,Mo) ordered
phases formed in the grain matrix of Ni-16Cr-15Mo-3Fe alloy at
540 C (1005 F) for 16,000 h. Source: Ref 14

dark-field image of fine, coherent c 0 (Ni3Al)
precipitates formed in the grain matrix of alloy 214 (Ni-16Cr4.5Al-3Fe-Y) at 800 C (1470 F) for 8 h. Source: Ref 15
Fig. 14.26

dark-field image of fine, coherent c 0 (Ni3Al)
precipitates formed in the grain matrix of alloy 601 at about
590 C (1100 F) for 2.5 years. Original magnification: 97,000.
Source: Ref 16
pg 393
31/10/2007 4:07PM Plate # 0
pg 394
50
2.25Cr-1Mo, 2 kJ/mm
2.25Cr-1Mo, 3 kJ/mm
40
2.25Cr-1Mo, 4 kJ/mm
Reduction in area, %
HCM2S, 2 kJ/mm
HCM2S, 3 kJ/mm
30
HCM2S, 4 kJ/mm
20
10
0
550
600
650
700
750
PWHT, C
Fig. 14.27
Reduction in area for 2.25Cr-1Mo and HCM2S at various postweld heat treatment (PWHT) temperatures with an initial
applied tensile stress of 325 MPa (47 ksi). Source: Ref 22
in ferritic steels. However, reheat cracking (or

stress-relief cracking or strain-age cracking)
also occurs in austenitic stainless steels, alloy
800H, and gamma prime () strengthened alloys;
however, the impurity segregation is not likely
to be as critical a factor in these austenitic alloys.
Reheat cracking (or stress-relief cracking) of
ferritic steels can be described using the test
results generated by Nawrocki et al. (Ref 22, 23)
when comparing 2.25Cr-1Mo steel with a modied 2.25Cr steel (HCM2S). HCM2S contains
similar chromium amounts, but lower carbon
contents (to reduce hardness of HAZ) along
with the substitution of molybdenum (Mo) with
tungsten (W), vanadium (V), and niobium (Nb)
to increase creep strength, compared with
2.25Cr-1Mo steel (Ref 22). Comparing these
two alloy steels in terms of their respective
resistance to stress-relief cracking will allow
a better understanding of the important metallurgical factors that contribute to reheat cracking.
Nawrocki et al. (Ref 22, 23) used Gleeble
simulation techniques to evaluate stress-relief
cracking. Specimens were subjected to singlepass weld thermal simulation cycles with a
peak temperature of 1315 C (2400 F) followed
by cooling to room temperature. A tensile stress
was imposed on the specimen during cooling
and held for the duration of the test to simulate
the residual stresses present in an actual weldment. After cooling to room temperature, the
20 m
Fig. 14.28
Scanning electron micrograph of fracture surface

of HCM2S ruptured air at 675 C (1250 F) under
an initial stress of 325 MPa (47 ksi). Source: Ref 22
specimen was then subjected to a simulated,

programmed postweld heat treatment temperature and held at constant temperature and load
until failure (Ref 22). The rupture ductility of
HCM2S was found to be signicantly lower
than that of 2.25Cr-1Mo, as shown in Fig. 14.27
(Ref 22). HCM2S was found to be very brittle,
as shown in Fig. 14.28 which reveals essentially
complete intergranular fracture (Ref 22). The
authors (Ref 22) concluded that HCM2S was
more susceptible to stress-relief cracking than
2.25Cr-1Mo steel. Nawrocki et al. (Ref 23)
attributed the brittle, intergranular fracture of
31/10/2007 4:07PM Plate # 0
Chapter 14:
HCM2S to the formation of ne (5 to 40 m)

carbides (W/Fe-rich carbides) in the grain matrix
that were resistant to plastic deformation and
the formation of grain-boundary Fe-rich M3C
carbides that nucleated microcavities. With
more carbide-forming alloying elements, such as
tungsten, vanadium, and niobium, HCM2S is
expected to form more matrix-strengthening
carbides, thus resulting in brittle, intergranular
fracture due to signicantly reduced deformation
capability of the grain matrix.
14.3.3 Cracking in Austenitic Stainless
Steels and Fe-Ni-Cr Alloys
Among austenitic stainless steels, Type 321
and 347 have been frequently found to suffer
stress-relaxation cracking (or reheat cracking).
Titanium is used in Type 321 to stabilize the
alloy by forming titanium carbides (TiCs) in
the grain interior to prevent sensitization (i.e.,
formation of chromium carbides along grain
boundaries, thus creating a chromium-denuded
zone along grain boundaries) during heat treatment or cooling following welding. Type 347, on
the other hand, uses niobium (referred to as
columbium until 1968) to form NbC in the grain
interior to avoid sensitization during heat treatment or cooling following welding. Stressrelaxation cracking has been reported for these
two grades of austenitic stainless steels during
service at temperatures between approximately
500 and 750 C (930 and 1380 F) (Ref 24). De
Santis et al. (Ref 24) summarized a number of
failure cases involving these two grades of
stainless steels. These authors indicated that the
failures were the result of intergranular cracking,
and the components that failed were either in a
cold-worked structure or were associated with
welding. Also, all failures were found to have
occurred after relatively short service times.
Cases of service failures related to Type 321 and
347 reported by De Santis et al. (Ref 24) are
summarized:
A straight Type 347 tube, which was thermally insulated externally and carried a gaseous hydrocarbon stream with traces of
hydrogen sulde at 650 to 680 C (1200 to
1255 F), failed after only 96 h of service.
A similar failure also occurred in the same
plant after only 15 days of service involving
a heavier wall tube made of Type 347 at the
bend near the weld.
Failures of Type 321 tubes occurred after
service of about 1 month at the bend with the
tube carrying a synthesis gas stream at about

600 C (1110 F) with the outer diameter
being thermally insulated. The failure occurred at the intrados of the tube bend.
Type 321 tubes imbedded in a catalytic bed of
aluminum silicate of a cracking reactor and
carrying saturated steam failed at the bend
section after 3 months of operation. The tube
metal temperatures were between 650 and
750 C (1200 and 1380 F).
Type 321 serpentine heat-exchanger tubes,
which were exposed to synthesis gases at 700
to 750 C (1290 to 1380 F) with internal
temperature of about 300 C (570 F), failed
at the bend section after 3 months of service.
The authors (Ref 24) observed that these failures

exhibited a common characteristic in which
intergranular cracks were oxidized, leaving a
metallic lm in the center of the crack. The
metallic lm was found to be essentially iron
with nickel varying from 0 to about 15% and
little chromium. In addition to oxides observed
in the crack, the authors also observed in most
cases iron and nickel suldes.
Both Type 321 and 347 are known to be susceptible to intergranular cracking in the heataffected zone (HAZ) during services at a certain
temperature range (Ref 18). It is generally believed (Ref 2530) that titanium carbides (in
Type 321) and niobium carbides (in Type 347) in
the region next to the fusion are put back into
solution during welding and subsequent reprecipitation of these ne carbides (during cooling)
that strengthens the grain matrix in the HAZ
(particularly the coarse-grain HAZ) during service or reheating during PWHT. Stress relaxation
in the matrix-strengthened HAZ during service
or PWHT causes cracking to develop at grain
boundaries, thus resulting in brittle, intergranular
fracture at the HAZ.
Alloy 800H (Fe-32Ni-21Cr-0.4Al-0.4Ti) is
the most widely used Fe-Ni-Cr alloy for hightemperature applications in the petrochemical
industry. Stress-relaxation cracking has also been
observed in this alloy under certain conditions.
Kohut (Ref 31) reported intergranular cracking
failures in a transfer line made of alloy 800H after
several months of service in a temperature range
of 540 to 705 C (1000 to 1300 F). The transfer
line ran between a pressure vessel and a feedefuent exchanger in a petrochemical unit. The
pipe was made from solution-annealed alloy
800H plate. Fabrication of pipe involved press
breaking the plate into two pipe halves followed
pg 395
31/10/2007 4:07PM Plate # 0
pg 396
by seam welding using shielded metal arc

welding. Press breaking produced cold-worked
structure in the pipe halves. The 30 cm (12 in.)
(diameter) transfer line failed at about 1 m (3 ft)
from the ange of the feed-efuent exchanger.
Four months later, a second failure occurred at
about 10 m (30 ft) from the rst failure. Both
failures were found to occur in the areas with
some deformation, which was believed to be the
result of the local repair. A third failure occurred
in the heat-affected zone of a weld initiating from
the inside diameter of tube surface after 1 year
of service. Electron microprobe analysis showed
no chlorine, sulfur, or other contaminants that
would cause embrittlement of the alloy. The
material near the fracture still exhibited good
room-temperature tensile ductility but much
higher yield and tensile strengths. Table 14.1
shows the room-temperature tensile test results
of the alloy 800H material obtained from the
30 cm (12 in.) transfer line near the fracture area
compared with the tensile test results of the
material that was obtained from the same area but
was re-solution annealed. The stress-rupture tests
at 595 C (1100 F), however, revealed that the
material near the fracture was very brittle, as
shown in Table 14.2 (Ref 31). The sample that
was obtained from the material near the fracture
exhibited extremely low elongation under the
test condition, while the sample obtained from
the same area, but was given a re-solution annealing treatment, was found to exhibit excellent
ductility under the same test condition.
Korkhaus (Ref 32) observed similar brittle,
intergranular cracking in an alloy 800H pipe
that was part of an outlet header of a process
air preheater coil operated at about 600 C
(1110 F). The leakage of the header was
detected after 2 years of service. Cracks were
observed to occur primarily near the pipe-to-tube
weld and in the vicinity of the longitudinal
Table 14.1 Results of room-temperature
tensile tests on alloy 800H samples obtained
from the area near the fracture as well as the
material from the same area but was re-solution
annealed
Material
Material near
fracture
Re-solution
annealed
Source: Ref 31
Yield
strength,
MPa (ksi)
Ultimate tensile
strength,
MPa (ksi)
Elongation,
%
Reduction
in area, %
460 (68)
765 (111)
23.5
46.5
205 (30)
506 (74)
55.0
72.2
pipe seam weld. The alloy 800H header was

508 mm in diameter, had 34.5 mm wall thickness, and was 4720 mm long. The header was
fabricated by cold bending the plate material into
pipe halves that were then seam welded using a
nickel-base alloy ller metal. No heat treatment
of this cold-formed and seam-welded pipe was
performed prior to installation as a header.
A joint industry sponsored project addressing stress-relaxation cracking failures in austenitic welded joints was undertaken by Van
Wortel (Ref 33) at TNO Institute of Industrial
Technology in Apeldoorn, The Netherlands.
Van Wortel reported major observations and
conclusions derived from the test program. Since
the brittle, intergranular cracking failures were
mainly associated with cold working and/or
welding, the program mainly focused on the
effect of cold working and welding on the
susceptibility of stress-relaxation cracking. A
specially developed three-point bending test
rig, which was capable of producing a cracking
failure with a similar crack morphology to the
service-induced failure, was used to determine
the susceptibility of the alloy to stress-relaxation
cracking. Testing procedures were not described
in the paper. The test results generated by Van
Wortel are summarized in Table 14.3 (Ref 33).
Type 321 is generally considered to be
susceptible to stress-relaxation cracking in the
industry. In Van Wortels test program, however,
Type 321 was found to suffer no cracking under
the test condition. It is believed that the 321H
material used in Van Wortels test program was
a ne-grained material with ASTM grain size
No. 7. Fine-grained materials are less susceptible
to stress-relaxation cracking. All other materials
tested were coarse-grained materials with ASTM
grain sizes of No. 2 and coarser. Van Wortel
(Ref 33) reported that several heat treatments
were benecial in improving the resistance of
the alloy to stress-relaxation cracking based on
Table 14.2 Results of stress-rupture tests

performed at 595 C (1100 F) and 276 MPa
(40 ksi) on alloy 800H samples obtained from
the area near the fracture as well as the
material from the same area but was re-solution
annealed
Material
Material near fracture

Re-solution annealed
Source: Ref 31
Rupture
life, h
Elongation,
%
Reduction
in area, %
385
261
1.2
15.8
12.5
18.6
31/10/2007 4:07PM Plate # 0
Chapter 14:
Table 14.3
pg 397
Results of relaxation tests on cold-worked plates and welded joints

304H
Condition
at 575 C
Weldment
PWHT(a)
Aged(b)
CW(c)
HT(1)(d)
HT(2)(e)
HT(3)(f)
Cracks
No cracks
No cracks
Cracks
No cracks
No cracks
321H
AC66
at 650 C
at 575 C
at 650 C
at 575 C
800H
at 650 C
at 650 C
617
No cracks
No cracks
No cracks
No cracks
No cracks
No cracks
No cracks
No cracks
No cracks
Cracks
No cracks
Cracks
Cracks
No cracks
No cracks
No cracks
Cracks
No cracks
Cracks
Cracks
Note: Grain sizes for the alloys under testing were ASTM No. 2 for 304H, No. 7 for 321H, No. 2 for AC66, No. 0 for 800H, and No. 1 and No. 0 for 617. (a) PWHT at 875 C/
3 h for 304H and 800H; 980 C/3 h for 617 after welding and before testing. (b) Material aged at 650 C for 16,000 h before testing. (c) Cold worked (CW) the metal from
0 to 15% before testing. (d) Heat treated the metal at 875 C/3 h for 304H and 980 C/3 h for 800H before cold working and testing. (e) Heat treated the metal after cold
working and before testing at 875 C/3 h for 304H and 980 C/3 h for 800H. (f ) Heat treated the metal from in-service cracked material at 980 C/3 h before testing. Source:
Ref 33
Testing temperature, C
100
50
200
300
400
500
600
700
Solution annealed
Aged 4000 h at 540 C (1000 F)
Aged 4000 h at 595 C (1100 F)
Aged 4000 h at 650 C (1200 F)
45
300
250
35
30
200
25
150
15
10
5
Room temperature
20
100
0.2% offset yield strength, MPa
40
0.2% offset yield strength, ksi
his three-point bending testing method. However, no examples were given to illustrate the
effectiveness of these heat treatments in actual
applications.
Van Wortel (Ref 33) indicated that the alloy
is most susceptible to stress-relaxation cracking
when the alloy matrix is in an age-hardened
condition with a high density of ne precipitates
within the grain. This condition can signicantly
reduce the deformation capability of the alloy.
Alloys under this condition are expected to suffer
cracking after only 0.1 to 0.2% relaxation strain
(Ref 33). Cold working can accelerate the
precipitation process within the grain, making
the alloy more susceptible to stress-relaxation
cracking. Van Wortel (Ref 33) proposed that the
microstructure in alloy 800H susceptible to
stress-relaxation cracking consisted of very ne
matrix carbides, which strengthen the grain
matrix, and grain-boundary carbides along with a
denuded carbide zone along the grain boundaries
that weaken the grain boundaries. Van Wortel,
however, did not mention the formation of [Ni3
(Al,Ti)] precipitates in the grain matrix in the
age-hardened condition for Alloy 800 or 800H.
The major strengthening mechanism for
alloy 800H at temperatures at which the alloy is
susceptible to stress-relaxation cracking is, in
fact, the precipitation of ne, coherent [Ni3(Al,
Ti)] precipitates (Ref 3436). Lai and Kimball
(Ref 37) observed that the maximum age hardening for alloy 800H was at 595 and 650 C
(1100 and 1200 F). These authors further indicated that the strengthening was more pronounced at the elevated temperatures than
at room temperature. This is illustrated in
Fig. 14.29 (Ref 37). The retained tensile elongation at both room-temperature and aging temperatures after 4000 h of aging was found to be
quite good (approximately 35 to 45% tested at
50
0
200
0
400 600 800 1000 1200 1400
Testing temperature, F
Fig. 14.29
Strengthening at room temperature compared

with strengthening at the aging temperatures
after aging at 540, 595, and 650 C (1000, 1100, and 1200 F) for
4000 h for alloy 800H containing 0.39% Al and 0.44% Ti.
Source: Ref 37
room temperature and approximately 25 to 35%

tested at respective aging temperatures) (Ref 37).
It is well known that age hardening in alloy 800
or 800H is a strong function of combined Al+Ti
content in the alloy. Alloys with higher content of
Al + Ti exhibit higher degree of age hardening
due to formation of more ne [Ni3(Al,Ti)]
precipitates. An example is given in Fig. 14.30,
which shows the aging behavior of two heats
of alloy 800H with Heat A containing higher
combined Al + Ti content (0.83%) than Heat E
31/10/2007 4:07PM Plate # 0
pg 398
100
95
650 C
(1200 F)
Heat A
Heat E
90
595 C
(1100 F)
540 C (1000 F)
Hardness, HRB
85
540 C (1000 F)
595 C (1100 F)
80
650 C (1200 F)
75
760 C (1400 F)
70
760 C (1400 F)
815 C (1500 F)
65
60
1
Unaged
103
104
105
Aging time, h
Aging behavior of two heats of alloy 800H with Heat A containing higher combined Al + Ti (0.83%) showing significantly
more age hardening than Heat E with much lower combined Al + Ti (0.58%). Source: Ref 37
(0.58% of combined Al + Ti) (Ref 37). Heat A

showed signicantly more age hardening than
Heat E with both materials having similar grain
sizes. Increasing the combined Al + Ti content in
alloy 800H causes signicant strengthening
in creep-rupture strength. This is illustrated in
Fig. 14.31 (Ref 38). Increasing the combined
Al +Ti content in alloy 800H increases the 1%
creep strength at 650 C (1200 F), which was
the maximum age-hardening temperature. At
850 C, [Ni3(Al,Ti)] precipitates coarsen in
alloy 800H; thus the effect of the combined
Al +Ti content signicantly diminishes, as
shown in Fig. 14.31 (Ref 38).
Accompanying the strengthening of the
alloy produced by [Ni3(Al,Ti)] precipitates,
alloy 800H exhibits creep embrittlement at the
maximum age-hardening temperature, particularly for the heats containing higher levels of
combined Al +Ti content. This is illustrated in
Fig. 14.32 (Ref 38). Figure 14.33 also shows
the similar effect of drastic reduction of stressrupture ductility for alloys containing higher
levels of Al+Ti in alloy 800H when tested
at 650 C (1200 F)the maximum agehardening temperature (Ref 39). The data suggest that the low rupture ductility in alloy 800
at 600 and 650 C (1110 and 1200 F) is the
result of [Ni3(Al,Ti)] precipitates with higher
combined Al +Ti content producing a larger
150
RP1/10,000
650 C (1200 F)
Stress (N/mm2)
Fig. 14.30
102
10
815 C
(1500 F)
RP1/30,000
100
50
850 C (1560 F)
RP1/10,000
RP1/30,000
0
0.4
0.6
0.8
1.0
1.2
1.4
Ti + AI, %
Fig. 14.31
Effect of combined Al + Ti content in alloy 800H

on the creep strength (1% in 10,000 h and
30,000 h) at 650 and 850 C (1200 and 1560 F). Source: Ref 38
volume of these ne, coherent precipitates, thus

resulting in signicant reduction in stress-rupture
ductility at these temperatures. The rupture ductility of alloy 800 can be further reduced to close
to nil with prior cold work when tested at 600 or
650 C (1110 or 1200 F). Stone et al. (Ref 39)
showed the effect of cold work on stress-rupture
ductility when tested at 600 C (1110 F) for
alloy 800. Their test results showed that the
31/10/2007 4:07PM Plate # 0
Chapter 14:
rupture elongation was reduced from about 10%

in the solution-annealed condition to almost
nil with 10 to 20% prior cold work, as shown
in Fig. 14.34 (Ref 39). Similar results were
observed by Smith (Ref 40) when he conducted
stress-rupture tests at temperatures from 540 to
650 C (1000 to 1200 F). Smiths test results
are summarized in Table 14.4 (Ref 40). All the
prestrained specimens were quite brittle under
the test conditions, exhibiting rupture elongation of less than 1%. The elongation for the
60
50
Elongation, %
0.1 h
40
30
1000 h
20
10,000 h
10
30,000 h
0.4
0.6
0.8
1.0
1.2
1.4
Ti + AI, %
Fig. 14.32
Creep rupture ductility of alloy 800H as a function

of combined Al + Ti content in the alloy tested at
650 C (1200 F). Source: Ref 38
300
Stress, MPa
0.4% AI 0.6% Ti
0.2% AI 0.4% Ti
100
50
Elongation, %
30
50
40
30
20
10
0
10
Fig. 14.33
0.2% AI 0.4% Ti
0.4% AI 0.6% Ti
102
103
Time, h
104
105
Effects of combined Al + Ti content in alloy 800

on stress rupture strength and rupture ductility at
650 C (1200 F). Source: Ref 39
prestrained specimens, which was so low it could

not be obtained by the traditional method of
tting the ends of the ruptured specimen together
for determination of the percent of elongation,
was taken from the creep curves.
14.3.4 Cracking in Nickel-Base Alloys
Some wrought nickel-base chromia-forming
alloys (i.e., nickel-base alloys forming chromium
oxide scales for high-temperature corrosion resistance) that are designed for high-temperature
applications contain a relatively low level of aluminum, typically 1 to 2%, to further improve
the oxidation resistance of the alloy. Some of
the more familiar alloys in this group includes
alloys 617 (1.2% Al) and 601 (1.4% Al). As
discussed earlier, Van Wortel (Ref 33) reported
that alloy 617 was susceptible to stress-relaxation
cracking at 650 C (1200 F), as shown in
Table 14.3. The author, however, provided little
information about the microstructure and aging
characteristics of alloy 617. Alloy 617 also is
known to age harden at 600 and 650 C (1110
and 1200 F), as shown in Fig. 14.35 (Ref 41).
Alloy 617, similar to alloy 800H, exhibits
excellent room-temperature ductility after longterm aging, as shown in Fig. 14.36 (Ref 41).
Nevertheless, alloy 617, similar to alloy 800H,
was susceptible to stress-relaxation cracking
at 650 C (Table 14.3). Bassford and Schill
(Ref 42) believed that precipitates are a
major contributor to age hardening in alloy 617
at 650 C as well as 600 C. Both alloys 800H
and 617 are strengthened primarily by precipitates at temperatures when stress-relaxation
cracking has been observed. For nickel-base
alloys, this type of cracking is generally referred
to as strain-age cracking.
Stahl et al. (Ref 43) reported a reformer tube
failure in a steam reformer for hydrogen production. The reformer tube was made of alloy
601. The product gas cooled from 900 to 600 C
(1650 to 1110 F) in the tube and then owed
through the outlet manifold system. Cracking
was observed to occur near the outlet header
in the heat-affected zone parallel to the weld
fusion line after about 1 year of operation. Their
repair method involved cutting out all the problem connections between the reformer tube and
the outlet header, which were replaced with
new outlet manifold welded connections. These
welded connections were then heat treated at
950 C (1740 F) for 1 h before they were connected to the reformer tubes. The postfabrication
pg 399
31/10/2007 4:07PM Plate # 0
pg 400
Stress, MPa
500
5% CW
10% CW
20% CW
25% CW
300
200
ST. 1120 C (2050 F)
100
Elongation, %
40
30
20
ST. 1120 C (2050 F)
10
0
10
102
103
104
105
Time, h
Fig. 14.34
Effects of cold work (cw) on the rupture ductility of alloy 800 at 600 C (1110 F). Solid lines represent the data generated
from the specimens in the as-solution heat treated condition. The specimens were solution heat treated (ST) at 1120 C
540 (1000)
565 (1050)
595 (1100)
650 (1200)
Applied
stress,
MPa (ksi)
345 (50)
310 (45)
310 (45)
275 (40)
241 (35)
207 (30)
207(a) (30(a))
172 (25)
155 (22.5)
138 (20)
Rupture
life, h
SA
1214
3350
615
1953
772
2670
546
1152
2352
PS
129
319
92
546
247
721
7342(a)
861
1565
2589
Rupture
elongation, %
SA
PS
19
13
15
11
13
14
16
20
22
0.26
0.20
0.2
0.3
0.2
0.28
0.3(a)
0.45
0.4
0.33
Exposure temperature, C
650
705
760
90
600
Exposure time, h
48,000 h
80
12,000 h
70
500
4,000 h
1,000 h
60
400
100 h
50
Not exposed
300
Yield strength, MPa
Test temperature,
C (F)
595
Yield strength, ksi
Table 14.4 Effect of 20% prestrain in

compression on stress rupture ductility of alloy
800H tested at 540, 565, 595, and 650 C (1000,
1050, 1100, and 1200 F) using bar specimens
in both solution-annealed (SA) and prestrained
(PS) conditions
40
30
1100
1200
1300
Exposure temperature, F
1400
(a) Prestrained 20% in tension; broke at extensomer weld. Source: Ref 40
Fig. 14.35
heat treatment, according to the authors (Ref 43),

was (a) to eliminate the effect of retained coldwork and welding stresses and (b) to coarsen
both the grain-boundary precipitates and grainmatrix precipitates to reduce age-hardening
effects. These authors did not mention in their
paper about (Ni3Al) precipitates being a
possible strengthening mechanism in alloy 601,
which contains about 1.4% Al and is known to

form (Ni3Al) precipitates. The heat treatment
at 950 C is believed to mainly coarsen both
grain-boundary and intragranular carbides in
addition to relieving residual stresses resulting
from cold working and welding. It should provide some improvement to the resistance of
the alloy to strain-age cracking. Nevertheless, the
Room-temperature yield strengths of alloy 617

after aging at different temperatures for various
times up to 48,000 h. Source: Ref 41
31/10/2007 4:07PM Plate # 0
Chapter 14:
temperature of 950 C (1740 F) in the heat

treatment would be above the solvus temperature
of the (Ni3Al) precipitates, and (Ni3Al)
precipitates can form again during the service at
strain-age cracking temperatures of 600 and
650 C (1110 and 1200 F).
Lai (Ref 16) reported a case of strain-age
cracking involving a recuperator shell made
of alloy 601 that was used for preheating air
for combustion. The recuperator shell suffered
through-thickness cracking after 2.5 years of
service. The failure occurred at the location
where the metal temperature was approximately
590 C (1100 F). No cracking was observed at
other locations with temperatures either higher or
lower than 590 C (1100 F). Both the air side
and ue gas side of the recuperator shell showed
little oxidation or corrosion attack. The recuperator shell (7.9 mm, or 5/16 in. thick) failed by
intergranular fracture, as shown in Fig. 14.37.
Figure 14.38 shows a crack propagating along
grain boundaries. The gure shows some evidence of cold work that is likely to result from the
original fabrication of the recuperator shell.
The tensile property of alloy 601 from the
recuperator shell near the main fracture was
evaluated. Tensile blanks were obtained and
machined into round tensile specimens with
adequate material being machined off from both
internal and external surfaces (ID and OD) of the
shell. Table 14.5 shows the tensile properties of
the material from the area that was near the main
fracture (Ref 16). The material from the recuperator shell exhibited good tensile ductility at
room temperature. At 590 C (1100 F), however, the ductility was found to be much lower
and the yield strength was very high, which was
close to that at room temperature. Transmission
electron microscopy examination of the recuperator shell showed the presence of ne, precipitates, as shown in Fig. 14.26. Tensile blanks
were obtained from the recuperator shell and
re-solution annealed at 1150 C (2100 F) for
30 min followed by water quenching. Tensile
properties of these re-solution annealed specimens at both room temperature and 590 C
(1100 F) are included in Table 14.5 for comparison. The tensile properties of re-solution
annealed material were similar to those of solution-annealed material reported in the literature
(Ref 44). Assuming that the re-solution annealed
material has the properties of the original plate
material prior to service, the data indicated that
the alloy 601 recuperator shell increased its yield
strength at 590 C (1100 F) by almost three
times (but only two times at room temperature)
as a result of the service exposure at about
590 C (1100 F). In order to determine that
(a)
600
Exposure temperature, C
650
700
750
Not exposed
70
Elongation, %
60
50 Exposure
40
100 h
1000 h
time, h
4,000 h
12,000 h
30
48,000 h
8,000 h
20
1100
Fig. 14.36
1200
1300
Exposure temperature, F
1400
Room-temperature tensile elongation of alloy 617

after aging at different temperatures for various
times up to 48,000 h. Source: Ref 41
(b)
Fig. 14.37
Scanning electron micrographs, (a) low magnification and (b) high magnification, showing
intergranular fracture surface of the alloy 601 recuperator shell.
Original magnification: (a) 13 and (b) 100. Source: Ref 16
pg 401
31/10/2007 4:07PM Plate # 0
pg 402
the recuperator shell material was in a brittle

condition (extremely low creep ductility) under
creep condition, stress rupture testing was performed at 590 C (1100 F) involving smoothnotch specimens obtained from the recuperator
Fig. 14.38
Optical micrograph showing intergranular

cracking near the main fracture. Original magnification: 100. Source: Ref 16
shell material and the shell material after resolution annealing at 1150 C (2100 F). The
test results are shown in Table 14.6 (Ref 16). The
brittle nature of the alloy under the creep test
condition at 590 C (1100 F) was substantiated
from these tests. Furthermore, the test results
indicated that a full recovery of the original
ductile material condition can be obtained by
re-solution annealing the material.
Lai (Ref 16) presented additional test data to
further substantiate that alloy 601 creep embrittlement was related to the age-hardened condition produced during service at 590 C
(1100 F). An annealed alloy 601 plate of different heat number was obtained from an alloy
supplier for evaluation. Specimen blanks were
aged at 590 C (1100 F) for 10,000 h. Stressrupture tests were performed at 590 C (1100 F)
and 310 MPa (45 ksi) on both annealed and
aged specimens using smooth-notch specimens. Results of these tests are summarized in
Table 14.7. The aged specimen, although fractured at the smooth section, clearly showed an
embrittled condition with very low creep ductility and intergranular fracture. The as-received
specimen showed much better creep ductility
with dimple rupture, which is characteristic of
a ductile rupture.
Many high-temperature, nickel-base alloys
are solid-solution strengthened using primarily
molybdenum and/or tungsten to increase their
tensile and creep rupture strengths. These alloys
Table 14.5 Tensile properties of alloy 601 specimens obtained from the area near the main fracture
in the alloy 601 recuperator shell that suffered intergranular through-wall cracking after 2.5 years
of service with the metal temperature at about 590 C (1100 F) during service. For each test condition,
two tests were run and both values are reported.
Material
From recuperator
RSA(a)
Test temperature,
C (F)
0.2% yield strength,

MPa (ksi)
Ultimate tensile strength,

MPa (ksi)
Elongation,
%
Reduction in
area, %
Room temperature
590 (1100)
Room temperature
590 (1100)
426, 414 (61.8, 60.0)

372, 386 (54.0, 56.0)
232, 224 (33.7, 32.5)
154, 142 (22.4, 20.6)
883, 865 (128.0, 125.5)

606, 600 (87.9, 87.0)
616, 614 (89.3, 89.1)
481, 478 (69.7, 69.3)
36, 34
18, 13
56, 53
63, 62
32, 34
15, 15
66, 53
55, 54
(a) Test blanks from the recuperator shell near the main fracture were re-solution annealed (RSA) at 1150 C (2100 F) for 30 min followed by water quenching. Source:
Ref 16
Table 14.6 Results of stress rupture tests performed at 590 C (1100 F) and 310 MPa (45 ksi)
on smooth-notch round specimens obtained from the 601 recuperator shell and from the
re-solution-annealed recuperator shell material
Specimen condition
From recuperator shell

From recuperator shell + 1150 C (2100 F)
for 30 min and water quench
Source: Ref 16
Failure location
Fracture mode
Notch, rupture life: 3110 h

Smooth section, rupture life: 175 h,
22% elongation, 14% reduction in area
Intergranular fracture
Dimple rupture
31/10/2007 4:07PM Plate # 0
Chapter 14:
Table 14.7 Results of stress rupture tests at

590 C (1100 F) and 310 MPa (45 ksi) on
smooth-notch specimens of as-received and
aged alloy 601 specimens from alloy 601 plate
of different heat number obtained from an alloy
supplier
Specimen
Failure Rupture Elongation, Reduction

location
life, h
%
in area, %
As-received Smooth
Aged(a)
Smooth
963
5348
19
4
21
4
Fracture
mode
Dimple
Intergranular
(a) Aged at 590 C (1100 F) for 10,000 h. Source: Ref 16
Table 14.8 Results of stress rupture tests at

590 C (1100 F) and 310 MPa (45 ksi) on
smooth-notch specimens of as-received and aged
alloy X specimens
Specimen
Failure
location
Rupture
life, h
Elongation,
%
Reduction
in area, %
Fracture
mode
As-received
Aged(a)
Smooth
Smooth
1574
1919
42
55
44
63
Dimple
Dimple
is actually strengthened by ne, coherent

(Ni3Nb) precipitates, and alloy X-750 containing
Al, Ti, and Nb is strengthened by both and
precipitates. Many of the applications for these
alloys are in sheet products. As a result, fabricated products are under much less constraint
than are heavy section components. Furthermore, the grain sizes for sheet products are
typically ASTM No. 4 or ner. Accordingly,
components made of sheet products are less
susceptible to strain-age cracking. Nevertheless,
in some cases components made of sheet products for these alloys can be heavily formed
with severe cold-worked conditions, thus making
them susceptible to strain-age cracking even
during postfabrication heat treatments.
Rowe (Ref 45) studied strain-age cracking of
wrought - or -strengthened sheet alloys by
using a controlled heating rate testing (CHRT)
method that was developed in 1960s (Ref 46).
Testing involves heating a tensile specimen at a
controlled heating rate to a test temperature in
the or precipitation temperature range, and
then pulling the specimen in tension to failure.
The heating rate of the postweld annealing heat
treatment of a fabricated component could be
used for the CHRT in the test. The minimum
tensile elongation from the test results is then
used as a guide for the susceptibility to strain-age
cracking for the alloy. Figure 14.39 summarizes
the test results generated by Rowe (Ref 45). The
results indicated that increasing the combined
Al + Ti + Nb content in the alloy decreases the
minimum elongation, thus increasing the susceptibility to strain-age cracking.
20
Minimum CHRT elongation, %
are also strengthened by chromium-rich carbides.

Among the most widely used solid-solutionstrengthened alloys for high-temperature applications is alloy X (Ni-22Cr-18Fe-9Mo), which
among other applications is widely used for
gas turbine combustion chambers in both landbased gas turbine power generation and aircraft
engines. In smooth-notch stress rupture tests
to investigate the susceptibility of strain-age
cracking of alloy 601, Lai (Ref 16) included alloy X to examine this solid-solutionstrengthened alloy, which is also strengthened
by carbides (primarily M6C carbides) but with
no precipitates. The results of his tests on
alloy X are summarized in Table 14.8 (Ref 16).
Alloy X aged for 10,000 h at 590 C (1100 F),
under the same test conditions as were used for
alloy 601, was found to exhibit excellent creep
ductility (about 55% rupture elongation) with
ductile dimple rupture.
Some nickel-base alloys use [Ni3(Al,Ti)]
precipitates for providing strengthening in addition to solid-solution strengthening using
elements such as molybdenum and/or tungsten.
Some of the widely known wrought alloys in
this group include alloys 263 (Ni-20Cr-20Co6Mo-0.5Al-2.2Ti), 718 (Ni-18Cr-19Fe-3Mo5Nb), X-750 (Ni-16Cr-7Fe-0.7Al-2.5Ti-1Nb),
Waspaloy (Ni-19Cr-14Co-4Mo-1.5Al-3Ti), and
R-41 (Ni-19Cr-11Co-10Mo-1.5Al-3Ti). Alloy
718, which contains Nb instead of Al and Ti,
pg 403
263
18
16
14
12
X-750
10
718
8
6
Waspaloy
PK33
R-41
0
2
Al + Ti + Nb, at.%
Fig. 14.39
Results of controlled heating rate tests (CHRT)

in terms of minimum elongation (%) as a function of combined Al + Ti + Nb content in atomic percent for
0
selected c - and/or c 00 -strengthened wrought sheet alloys. Source:
Ref 45
31/10/2007 4:07PM Plate # 0
pg 404
Nickel-base alloys containing high levels of

aluminum or aluminum plus titanium can be
highly susceptible to strain-age cracking, particularly in plate or thick-wall tubing products.
Some of these alloys include alloy 214 (4.5%
Al), which forms a tenacious Al2O3 oxide scale
to resist oxidation at very high temperatures,
and alloy 693 (3% Al), which forms Cr2O3/
Al2O3 oxide scales to resist metal dusting in the
temperature range of 480 to 700 C (900 to
1300 F).
Ni-Cr-Mo and Ni-Mo alloys have the potential
for developing a long-range ordered Ni2(Cr,Mo)
phase after long aging times at a certain temperature range. Tawancy (Ref 47) observed the
formation of [Ni2(Cr,Mo)] phases in alloy S
(Ni-15.5Cr-14.5Mo) after aging for 8000 h
at 540 C (1000 F), as shown in Fig. 14.40
(Ref 47). The ordered phase caused the alloy
to signicantly increase its room-temperature
tensile strengths, particularly yield strength, as
shown in Table 14.9 (Ref 47). Tawancy and
Asphahani (Ref 48) reported that C-276 alloy
(Ni-16Cr-16Mo-4W) exhibited signicant roomtemperature strengthening when the alloy was
aged to develop a fully ordered structure, as
shown in Table 14.10. The Ni2(Cr,Mo) ordered
phase exhibits a morphology similar to and
precipitates and produces similar strengthening
effects as - and -strengthened alloys.
It is quite likely that a long-range ordered NiCr-Mo alloy may also be susceptible to strain-age
cracking. Alloy C-22 (Ni-22Cr-13Mo-3W) is
selected to be a candidate nuclear waste container
material for the Yucca Mountain site in Nevada
(Ref 49). Studies on the changes in microstructure, mechanical properties, and corrosion
resistance of C-22 alloy after long-term aging
have been carried out in recent years (Ref 50
52). Rebak et al. (Ref 52) reported the observation of long-range ordered phases, [Ni2(Cr,
Mo)], in C-22 alloy after aging at 595 C
(1100 F) for 1000 and 16,000 h, at 540 C
(1000 F) for 1000 h, and at 430 C (800 F) for
30,000 and 40,000 h. Room-temperature tensile
reduction in area of C-22 was reported to have
been about 80% in annealed condition to 75%
after aging for 40,000 h at 430 C (800 F),
which was apparently in the early stage of longrange ordered condition (Ref 52). In a similar
study on aging behavior of C-22 alloy, the alloy
after aging at 595 C (1100 F) for 16,000 h
showed no strengthening for one heat but slight
strengthening for the other heat, as shown in
Table 14.11 (Ref 51). The authors (Ref 51)
observed continuous grain-boundary carbides
in the aged samples and mentioned about the
formation of Ni2(Cr,Mo) ordered phases, but
without presenting positive identication of
the ordered phases using transmission electron
microscopy. With either no strengthening or
slight strengthening as presented in Table 14.11,
Table 14.9 Room-temperature tensile properties

of alloy S comparing annealed material with the
material aged for 8000 h at 540 C (1000 F)
Condition
Annealed specimen
Aged specimen(a)
0.2% yield strength, Ultimate tensile Elongation,

MPa (ksi)
strength, MPa (ksi)
%
436 (63)
824 (119)
889 (129)
1284 (186)
58
42
(a) Aged at 540 C (1000 F) for 8000 h. Source: Ref 47
Table 14.10 Room-temperature tensile

properties of C-276 alloy comparing annealed
material (disordered structure) with aged material
with a long-range ordered structure
Condition
Fig. 14.40
Transmission electron micrograph showing longrange ordered phases [Ni2(Cr,Mo)] in a dark field
image using a h220i reflection in alloy S after 8000 h at 540 C
Annealed
Ordered structure(a)
0.2% yield
strength,
MPa (ksi)
Ultimate
tensile strength,
MPa (ksi)
Elongation,
%
356 (51.6)
770 (111.7)
792 (114.8)
1240 (179.8)
61
28
(a) Aging condition was not reported in the paper. Source: Ref 48
31/10/2007 4:07PM Plate # 0
Chapter 14:
it is unlikely the ordered phases were fully

developed. The slight reduction in ductility as
shown in Table 14.11 is most likely caused by
the formation of continuous grain-boundary
carbides. The mechanical properties that were
under investigation in these studies (Ref 5052)
included tensile tests and impact toughness
testing. When long-range ordered phases are
fully developed, it may be imperative to perform
stress-relaxation tests or smooth-notch stress
rupture tests at the ordering temperatures to
determine the susceptibility of the alloy to stressrelaxation cracking (or strain-age cracking).
A relatively new age-hardenable Ni-Cr-Mo
alloy (C-22HS) with 21Cr and 17Mo in nickel
was developed using long-range ordered phases
[Ni2(Cr,Mo)] to produce strengthening through
a heat treatment (Ref 53). The heat treatment
produces a fully developed ordered structure that
produces signicant strengthening, particularly
at the ordering temperature (i.e., 595 C, or
1100 F). Tensile properties of Alloy C-22HS in
both as-annealed and as-heat-treated conditions
are summarized in Table 14.12 (Ref 53). This
behavior is quite similar to the alloys strengthened by precipitates, such as Waspaloy, R-41,
and so forth. Thus, the alloys that are strengthened by the ordered Ni2(Cr,Mo) phase may be
equally susceptible to strain-age cracking as
alloys.
Table 14.11 Room temperature tensile

properties of C-22 alloy comparing annealed
material with the material aged for 16,000 h at
595 C (1100 F)
Heat
Condition
No. 1 Annealed
Aged(a)
No. 2 Annealed
Aged(a)
0.2% yield
strength,
MPa (ksi)
Ultimate
tensile
strength,
MPa (ksi)
340 (49.3)
381 (55.2)
342 (49.6)
485 (70.3)
768 (111.3)
806 (116.9)
757 (109.7)
931 (135.0)
Elongation, Reduction
%
in area, %
66.2
55.4
70.0
39.0
76.1
50.7
77.0
31.1
pg 405
14.3.5 Mitigation of Reheat Cracking

and Stress-Relaxation Cracking or
Strain-Age Cracking
As was discussed in previous sections, the
reheat cracking due to heat treatment or the
stress-relaxation cracking (or strain-age cracking) that occurs during service is primarily related to strengthening of the grain matrix by ne
precipitates that formed at the intermediate
temperatures. Thus, to minimize the chances of
developing reheat cracking during postweld heat
treatment of a welded component or annealing/
stress-relieving of a cold-formed component, it
is recommended that the component be heated
through that temperature range quickly to minimize the formation of those detrimental precipitates. As for the stress-relaxation cracking
(or strain-age cracking) that occurs during
service, one effective method for mitigating
the cracking is to remove the residual stresses in
the component by performing a postweld heat
treatment for a welded component or an annealing/stress-relieving heat treatment for a coldformed component (e.g., pipe U-bend section)
prior to service. Other factors that may be benecial in reducing the cracking tendency during
service include (a) minimizing stress risers, (b)
the use of a ne-grained material, and (c) a heat
treatment involving heating the component to
a temperature high enough to coarsen precipitates (e.g., carbides) to reduce the matrix
strengthening.
14.4 Summary
Most high-temperature components are under
stress during service. The tensile stresses (or
strains) that are imposed on the component
can cause the alloy to suffer preferential corrosion penetration in high-temperature, corrosive
environments, particularly low-oxygen and highsulfur partial pressure conditions. The effects
Table 14.12 Tensile properties of C-22HS (a heat treatable alloy) comparing the annealed material
and heat treated material tested at both room temperature and 595 C (1100 F)
Condition
Test temperature, C (F)
0.2% yield strength,

MPa (ksi)
Ultimate tensile
strength, MPa (ksi)
Elongation, %
Reduction
in area, %
Annealed
HT(a)
Annealed
HT(a)
Room temperature
Room temperature
595 (1100)
595 (1100)
406 (59)
742 (108)
222 (32)
542 (79)
822 (119)
1232 (179)
640 (93)
934 (135)
61
40
72
48
75
50
67
66
(a) Heat treated at 705 C (1300 F)/16 h/furnace cool to 605 C (1120 F)/32 h followed by air cool. Source: Ref 53
31/10/2007 4:07PM Plate # 0
of the tensile stresses and strains on the preferential suldation penetration attack on various
alloys in suldizing environments are reviewed.
The preferential suldation penetration is considered to be a precursor to the circumferential
cracking that has been observed in the waterwall
tubes for some supercritical coal-red boilers
red under low NOx combustion conditions.
A detailed discussion on the preferential suldation penetration and the circumferential cracking of the waterwall tubes observed in some
supercritical coal-red boilers is presented.
Additional discussion on preferential suldation
penetration and circumferential cracking observed in waterwall tubes is also presented in
Chapter 10 Coal-Fired Boilers.
Stresses that are either residual stresses in the
component or externally applied stresses can
cause brittle, intergranular cracking when the
component is under service at the low end of the
intermediate temperature range. This cracking
phenomenon is frequently referred to as reheat
cracking, or stress-relaxation cracking, or
strain-age cracking. Reheat cracking occurs
when a welded component is subjected to a
postweld heat treatment (PWHT) or when a
cold-formed component is subjected to an
annealing or stress-relieving heat treatment.
Stress-relaxation cracking (or strain-age cracking) occurs when the component is under service.
This stress-induced cracking behavior of ferritic
steels, stainless steels, Fe-Ni-Cr alloys, and
nickel-base alloys is reviewed.
REFERENCES
1. M. Schtze, Deformation and Cracking

Behavior of Protective Oxide Scales on
Heat-Resistant Steels under Tensile Strain,
Oxid. Met., Vol 24 (No. 3/4), 1985, p 199
2. M. Schtze, The Healing Behavior of Protective Oxide Scales on Heat-Resistant
Steels After Cracking under Tensile Strain,
Oxid. Met., Vol 25 (No. 5/6), 1986, p 409
3. G.R. Smolik and J.E. Flinn, Stress and
Environmental Interactions for INCOLOY
800H in Coal Gasication Environments,
J. Mater. Energy Systems, Vol 8 (No. 3),
1986, p 297
4. V. Guttmann and J. Timm, Corrosion and
Creep of Alloy 800H under Simulated Coal
Gasication Conditions, Werkst. Korros.,
Vol 39, 1988, p 322
5. M.F. Stroosnijder, V. Guttmann, and J.H.W.

de Wit, Corrosion and Creep Behaviour
of Alloy 800H in Sulphidizing/Oxidizing/
Carburizing Environments at 700 CPart
II: Creep Behaviour, Werkst. Korros., Vol
41, 1990, p 508
6. M.F. Stroosnijder, V. Guttmann, and R.J.N.
Gommans, Inuence of Creep Deformation
on the Corrosion Behaviour of a CeO2
Surface-Modied Alloy 800H in a Sulphidising-Oxidixing-Carburising Environment,
Mater. Sci. Eng., Vol A121, 1989, p 581
7. V. Guttmann, K. Stein, and W.T. Bakker,
Deformation-Corrosion Interactions in Selected Advanced High Temperature Alloys,
Mater. High Temp., Vol 14 (No. 2/3), 1997,
p 61
8. J. Le Coze, U. Franzoni, O. Cayla, F.
Devisme, and A. Lefort, The Development
of High-Temperature Corrosion-Resistant
Aluminium-Containing Ferritic Steels,
Mater. Sci. Eng., Vol A120, 1989, p 293
9. Materials, Part DProperties 2005
Addenda, ASME Boiler and Pressure Vessel
Code, ASME, New York, July 1, 2005
10. M. Le Calvar, P.M. Scott, T. Magnin, and
P. Rieus, Strain Oxidation Cracking of
Austenitic Stainless Steels at 610 C, Corrosion, Vol 54 (No. 2), 1998, p 101
11. K. Rorbo, Experience with Nitriding of
Austenitic Stainless Steel and Inconel in
Ammonia Environments, Environmental
Degradation of High Temperature Materials, Series 3, No. 13, Vol 2, The Institution
of Metallurgists, London, 1980, p 147
12. D.L. Klarstrom, Metallurgical Factors that
Promote Cracking During the Heat Treatment of High Performance Alloys, HeatResistant Materials II (Conf. Proc.) Second
International
Conference
on
HeatResistant Materials, K. Natesan, P. Ganesan,
and G. Lai, Ed., ASM International, 1995,
p 487
Dhahran, Saudi Arabia, 1993, p 144
Dhahran, Saudi Arabia, 1993, p 176
15. R.B. Herchenroeder, G.Y. Lai, and K.V.
Rao, A New, Wrought, Heat-Resistant
pg 406
31/10/2007 4:07PM Plate # 0
Chapter 14:
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
Ni-Cr-Al-Fe-Y Alloy, J. Met., Vol 35

(No. 11), 1983, p 16
G.Y. Lai, presented at ASM Materials Week
(Cincinnati, OH), Oct 2024, 1991
A. Dhooge and A. Vinckier, Reheating
CrackingA Review of Recent Studies,
Int. J. Pressure Vessels Piping, Vol 27, 1987,
p 239
C.F. Meitzner, Stress-Relief Cracking in
Steel Weldments, WRC Bull, Vol 211, Nov
1975
A. Dhooge, R.E. Dolby, J. Sebille, R.
Steinmetz, and A.G. Vinckier, A Review
of Work Related to Reheat Cracking in
Nuclear Reactor Pressure Vessel Steels, Int.
J. Pressure Vessels Piping, Vol 6, 1978,
p 329
A.G. Vinckier and A.W. Pense, A Review
of Underclad Cracking in Pressure Vessel
Components, WRC Bulletin No. 197,
Welding Research Council, Aug 1974
R.E. Dolby and G.G. Saunders, A Review
of the Problem of Reheat Cracking in
Nuclear Vessel Steels, Report 3453/1/75,
The Welding Institute, April 1975
J.G. Nawrocki, J.N. DuPont, C.V. Robino,
and A.R. Marder, The Stress-Relief Cracking Susceptibility of A New Ferritic Steel
Part 1: Single-Pass Heat-Affected Zone
Simulations, Weld. J., Dec 2000, p 355s
J.G. Nawrocki, J.N. DuPont, C.V. Robins,
J.D. Puskar, and A.R. Marder, The Mechanism of Stress-Relief Cracking in a Ferritic
Alloy Steel, Weld. J., Feb 2003, p 25s
R. De Santis, W. Dumini, and G. Casarini,
Service Failure of Stressed Stabilized
Stainless Steels at Elevated Temperatures,
WRC Bulletin No. 421, Welding Research
Council, 1997, p 17
M. Chabaud, L. Allais, P. Dubuisson,
C. Hogrel, and A. Pineau, Ageing and
Cracking of Simulated Heat Affected Zones
in 321 Stainless Steel, Mater. High Temp.,
Vol 15 (No. 3/4), 1998, p 395
L. Li and R.W. Messler, Stress Relaxation
Study of HAZ Reheat Cracking in Type 347
Stainless Steel, Weld. J., 2000, p 137s
R.J. Christoffel, Notch-Rupture Strength
of Type 347 HAZ, Weld. J., July 1960,
p 315
R.J. Truman and H.W. Kirkby, Some Ductility Aspects of 18-12-1Nb Steel, JISI,
1960, p 180
N.E. Moore and J.A. Grifths, Microstructural Causes of HAZ Cracking in
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
Heavy Section18-12-Nb Austenitic Stainless Steel Welded Joints, JISI, 1961, p 29

R.N. Younger and R.G. Baker, Heat Affected Zone Cracking in Welded Austenitic
Steels During Treatment, Brit. Weld. J.,
1961, p 579
G.B. Kohut, A Case History of Cracking
Failures in Deformed Austenitic Alloy Pipe
Operating in the Temperature Range of
10001300 Deg F (538705 Deg C), Trans.
ASME, 1975, p 316
J. Korkhaus, Application of CorrosionResistant Steels in the Chemical Industry,
Stainless Steels World 1999 Conference,
Conf. Proc., Book 1, KCI Publishing BV,
The Netherlands, 1999, p 27
J.C. Van Wortel, Relaxation Cracking in the
Process Industry, an Underestimated Problem, BALTICA IV: Plant Maintenance
for Managing Life & Performance, Vol 2,
S. Hietanen and P. Auerkari, Ed., VTT
Technical Research Centre of Finland,
Espoo, Finland, 1998, p 637
P.G. Stone, J. Orr, and J.C. Guest, Status
Review of Alloy 800, Proc. British Nuclear
Energy Society Conference, S.F. Pugh, Ed.,
AERE Harwell, 1975, p 15
J. Orr, A Review of the Structural Characteristics of Alloy 800, Paper No. 2-1,
Proc. Petten Conference, North-Holland
Publishing, 1978
N. Persson, Mechanical Properties of Alloy
800 Above 600 C, Paper No. 3.2-1, Proc.
Petten Conference, North-Holland Publishing, 1978
G.Y. Lai and O.F. Kimball, Aging Behavior
of Alloy 800H and Associated Mechanical
Property Changes, Report GA-A15194,
General Atomic Company, San Diego, CA,
Nov 1978
V. Coppolecchia, J. Bryant, F. Hofmann, and
K. Drefahl, Loss of Creep Ductility in Alloy
800H with High Levels of Titanium and
Aluminum, Performance of High Temperature Materials in Fluidized Bed Combustion
Systems and Process Industries, Conf. Proc.,
P. Ganesan and R.A. Bradley, Ed., ASM
P.G. Stone, J. Orr, and J.C. Guest, A Status
Review of Alloy 800, Proc. British Nuclear
Energy Society Conference, University of
Reading, Sept 2526, 1974, S.F. Pugh, Ed.,
AERE Harwell, UK, 1975, p 15
A.B. Smith, Characterization of Materials
for Service at Elevated Temperatures,
pg 407
31/10/2007 4:07PM Plate # 0
41.
42.
43.
44.
45.
46.
47.
48.
MPC-7, G.V. Smith, Ed., ASME, 1978,

p 159
W.G. Lipscomb, J.R. Crum, and P. Ganesan,
Mechanical Properties and Corrosion Resistance of INCONEL Alloy 617 for Renery Services, Paper No. 259, Corrosion/89,
NACE, 1989
T.H. Bassford and T.V. Schill, A Review of
INCONEL Alloy 617 and Its Properties
after Long-Time Exposure to Intermediate
Temperatures, Special Metals, Inc., Huntington, WV
H. Stahl, G. Smith, and S. Wastiaux, StrainAge Cracking of Alloy 601 Tubes at 600 C,
Practical Failure Analysis, Vol 1 (No. 1),
Feb 2001
INCONEL Alloy 601, Special Metals
Product Literature, Special Metals Company, Huntington, WV
M.D. Rowe, Ranking the Resistance of
Wrought Superalloys to Strain-Age Cracking, Weld. J., Feb 2006, p 27s
R.W. Fawley, M. Prager, J.B. Carlton, and
G. Sines, Recent Studies of Cracking
During Postwelding Heat Treatment of
Nickel-Base Alloys, WRC Bulletin No.
150, Welding Research Council, 1970
H.M. Tawancy, Order-Strengthening in a
Nickel-Base Superalloy (Hastelloy Alloy S),
Metall. Trans., Vol 11A, 1980, p 1764
H.M. Tawancy and A.I. Asphahani, Ordering Behavior and Corrosion Properties of
Ni-Mo and Ni-Mo-Cr Alloys, HighTemperature Ordered Intermetallic Alloys
(Symposium Proc.), Vol 39, C.C. Koch,
49.
50.
51.
52.
53.
C.T. Liu, and N.S. Stoloff, Ed., Materials

Research Society, 1985, p 485
1997 Findings and Recommendations,
Report to The U.S. Congress and The
Secretary of Energy, U.S. Nuclear Waste
Technical Review Board, Arlington, VA,
April 1998
R.B. Rebak and N.E. Koon, Localized
Corrosion Resistance of High Nickel Alloys
as Candidate Materials for Nuclear Waste
RepositoryEffect of Alloy and Weldment
Aging at 427 C for up to 40,000 H, Paper
T.S.E. Summers, M.A. Wall, M. Kumar,
S.J. Matthews, and R.B. Rebak, Phase
Stability and Mechanical Properties of C-22
Alloy Aged in the Temperature Range 590 to
760 C for 16,000 Hours, Scientic Basis for
Nuclear Waste Management XXII, (Symposium Proc.), Vol 556, D.J. Wronkiewicz
and J.H. Lee, Ed., Materials Research
Society, 1999, p 919
R.B. Bebak, T.S.E. Summers, and R.M.
Carranza, Mechanical Properties, Microstructure and Corrosion Performance of
C-22 Alloy Aged at 260 to 800 C, Scientic
Basis for Nuclear Waste Management XXI
(Symposium Proc.), Vol 608, R.W. Smith
and D.W. Shoesmith, Ed., Materials Research Society, 2000, p 109
L.M. Pike and D.L. Klarstrom, A New
Corrosion-Resistant Ni-Cr-Mo Alloy with
High Strength, Paper No. 04239, Corrosion/
pg 408
26/10/2007 12:42PM Plate # 0

DOI: 10.1361/hcma2007p409

CHAPTER 15
Molten Salt Corrosion

15.1 Introduction
Molten salt technology plays an important role
in various industries. In the heat treating industry,
molten salts are commonly used as a medium for
heat treatment of metals and alloys (e.g.,
annealing, tempering, hardening, quenching, and
cleaning) as well as for surface treatment (e.g.,
case hardening). In nuclear and solar energy
systems, they have been used as a medium for
heat transfer and energy storage. Other applications include extraction of aluminum, magnesium, sodium, and other reactive metals; rening
of refractory metals; and high-temperature batteries and fuel cells. Table 15.1 summarizes the
general applications of molten salt technology in
several industries (Ref 1).
The containment material, which is in contact
with the molten salt, is subject to molten salt corrosion. This chapter reviews the data relevant to
the corrosion of containment materials. Although
the literature related to studies of corrosion in
molten salts is extensive, as can be seen in an annotated bibliography prepared by Janz and Tomkins (Ref 2), corrosion data useful in selecting
materials are rather limited and fragmented.
15.2 Corrosion Process

Molten salts generally are a good uxing
agent, effectively removing oxide scales from a
Table 15.1
metal surface. The corrosion reaction proceeds

primarily by oxidation, which is then followed
by dissolution of metal oxides in the melt. Oxygen and water vapor in the molten salt thus often
accelerate molten salt corrosion.
Corrosion can also take place through mass
transfer due to thermal gradient in the melt. This
mode of corrosion involves dissolution of an
alloying element at hot spots and deposition
of that alloying element at cooler spots. This
can result in severe fouling and plugging in a
circulating system. Corrosion is also strongly
dependent on temperature and velocity of the
salt.
Corrosion can take the form of uniform thinning, pitting, or internal or intergranular attack.
In general, molten salt corrosion is quite similar
to aqueous corrosion. More complete discussion
on the mechanisms of molten salt corrosion can
be found in Ref 3 to 5.
15.3 Corrosion in Molten Chlorides

Chloride salts are widely used in the heat
treating industry for annealing and normalizing
of steels. These salts are commonly referred to as
neutral salt baths. The most common neutral salt
baths are barium, sodium, and potassium chlorides, used separately or in combination in the
temperature range of 760 to 980 C (1400 to
General applications of molten salt technology in several industries
Power
Metals/materials
Chemicals
Solar/thermal: collection,
storage, transfer
Nuclear: homogeneous
reactors, reprocessing
Batteries
Extraction: refractory metals, actinides, lanthanides,

transition, and light metals
Processing: heat treatment, annealing, quenching, cleaning,
cementation, electroforming
Surface nishing: anodizing, plating
Fuels: cracking, catalysts
Fuel cells
Joining: uxes and slags for welding, brazing, soldering, and

electroslag rening
Composites: glasses, ceramics, slags
Recycling
Source: Ref 1
Plastics: curing, etching, vulcanizing

Pyrolysis: recycling, scrap treatment,
hazardous materials disposal
Synthesis: organics, gases
Special applications: liquid crystals,
single-crystal growing, matrix
pg 409
26/10/2007 12:42PM Plate # 0
1800 F). Compositions of some common neutral salt baths are (Ref 6):

50NaCl-50KCl
50KCl-50Na2CO3
20NaCl-25KCl-55BaCl2
25NaCl-75BaCl2
21NaCl-31BaCl2-48CaCl2
Jackson and LaChance (Ref 6) performed an

extensive study on the corrosion of cast Fe-Ni-Cr
alloys in the NaCl-KCl-BaCl2 salt bath. They
found that alloys suffered intergranular attack
more than metal loss. Corrosion data in terms of
metal loss and intergranular attack are shown in
Fig. 15.1 and 15.2, respectively. The gures also
indicate that resistance to the molten salt (NaClKCl-BaCl2) increases with decreasing chromium
and increasing nickel in Fe-Ni-Cr alloys. HW
alloy (Fe-12Cr-60Ni) was consistently the best
performer among the four commercial cast alloys
(HW, HT, HK, and HH alloys) studied. These
authors further noted that intergranular attack
generally followed grain-boundary carbides.
Fig. 15.1
Thus, lowering carbon from 0.4% to about

0.07% resulted in a threefold improvement.
Decreasing grain size also improved alloy resistance to intergranular attack. Five different neutral salt baths were compared for HW, HT, and
three Fe-Cr alloys, as shown in Fig. 15.3. In
general, the four chloride salt baths were quite
similar. The KCl-Na2CO3 salt bath was signicantly less aggressive than pure chloride
baths. It is also interesting to note that Fe-17Cr
alloy was better than HW (Fe-12Cr-60Ni) and
HT (Fe-15Cr-35Ni) alloys in NaCl-KCl, NaClKCl-BaCl2, and NaCl-BaCl2-CaCl2 salt baths.
Lai et al. (Ref 7) evaluated various wrought
iron-, nickel-, and cobalt-base alloys in a NaClKCl-BaCl2 salt bath at 840 C (1550 F) for
1 month (Fig. 15.4). Surprisingly, two highnickel alloys (alloys 600 and 601) suffered more
corrosion attack than stainless steels such as
Types 304 and 310. Co-Ni-Cr-W, Fe-Ni-Co-Cr,
and Ni-Cr-Fe-Mo alloys performed best.
Laboratory testing in a simple salt bath failed to
reveal the correlation between alloying elements
Corrosion rates in terms of metal loss for four commercial cast Fe-Ni-Cr alloys in a 20NaCl-25KCl-55BaCl2 salt bath under
different conditions of rectification at 870 C (1600 F) for 60 h. Source: Ref 6
pg 410
26/10/2007 12:43PM Plate # 0
Chapter 15:
Molten Salt Corrosion / 411
Fig. 15.2
Corrosion rates in terms of intergranular attack for four commercial cast Fe-Ni-Cr alloys in a 20NaCl-25KCl-55BaCl2 salt bath
under different conditions of rectification at 870 C (1600 F) for 60 h. Source: Ref 6
Fig. 15.3
Comparison of different neutral salt baths for HW, HT, and Fe-Cr alloys at 870 C (1600 F) for 60 h. Source: Ref 6
pg 411
26/10/2007 12:43PM Plate # 0
pg 412
and performance. Tests were conducted at

840 C (1550 F) for 100 h in a NaCl salt bath
with fresh salt for each test run. Results are
tabulated in Table 15.2 (Ref 7, 8). Similar to the
eld test results, Co-Ni-Cr-W and Fe-Ni-Co-Cr
alloys performed best.
Evaluating a eutectic sodium-potassium-magnesium chloride (33NaCl-21.5KCl-45.5MgCl2,
mol%) as a possible heat-transfer and energystorage medium in solar thermal energy systems
for power generation, Coyle et al. (Ref 9) conducted corrosion tests on various commercial
alloys at 900 C (1650 F) for 144 and 456 h
(Table 15.3). Fifteen alloys, including iron-,

nickel-, and cobalt-base alloys, were evaluated.
After 144 h of exposure, specimens of
eight alloys were consumed. The remaining
seven alloys disintegrated after 456 h of exposure. The authors concluded that the chloride salt
was too aggressive to be used at 900 C
(1650 F).
At lower temperatures, molten salts generally
become less aggressive. Susskind et al. (Ref 10)
conducted corrosion tests at 450 to 500 C (840
to 930 F) in molten NaCl-KCl-MgCl2 eutectic
and found many alloys resistant to molten salt
corrosion (Table 15.4). Low corrosion rates
may also be attributed to the vacuum environment used in these tests. Investigating the corrosion behavior of alloys at 400 and 500 C
Table 15.3 Results of corrosion tests in molten
eutectic NaCl-KCl-MgCl2 salt at 900 C (1650 F)
Alloy
Fig. 15.4
Results of a field rack test in a NaCl-KCl-BaCl2 salt

bath at 840 C (1550 F) for 1 month. Source: Ref 7
Table 15.2 Results of laboratory tests in a NaCl

salt bath at 840 C (1550 F) for 100 h
Alloy
188
25
556
601
Multimet
150
214
304
446
316
X
310
800H
625
RA330
617
230
S
RA330
600
304
316
800
800H
556
Nickel
600
214
X
N
S
230
X-750
R-41
188
144 h
456 h
Disintegrated
Disintegrated
Disintegrated
310
250
Disintegrated
280
120
Disintegrated
Disintegrated
400
300
Disintegrated
150
Disintegrated
Disintegrated
Disintegrated
Disintegrated
Disintegrated
Disintegrated
Disintegrated
Disintegrated
N2-(0.11H2O)-(110O2) was used for the cover gas. Source: Ref 9

mm (mils)
0.051 (2.0)
0.064 (2.5)
0.066 (2.6)
0.066 (2.6)
0.069 (2.7)
0.076 (3-0)
0.079 (3.1)
0.081 (3.2)
0.081 (3.2)
0.081 (3.2)
0.097 (3.8)
0.107 (4.2)
0.109 (4.3)
0.112 (4.4)
0.117 (4.6)
0.122 (4.8)
0.140 (5.5)
0.168 (6.6)
0.191 (7.5)
0.196 (7.7)
A fresh salt bath was used for each test run; air was used for the cover gas.
(a) Mainly intergranular attack; no metal wastage. Source: Ref 7 and 8
Table 15.4 Corrosion of alloys in molten

eutectic NaCl-KCl-MgCl2 salt at 450 to 500 C
(840 to 930 F) with 50 C temperature
differential under vacuum for 1000 h
Alloy
1020
2.25Cr-lMo
304
310
316
347
410
430
446
600
N
Molybdenum
Tantalum
Source: Ref 10
Maximum penetration,
mm/yr (mpy)
0
0.08 (3)
<0.01 (0.4)
0
<0.01 (0.4)
0.12 (4.7)
0.03 (1.0)
0.05 (2)
<0.01 (0.4)
0.05 (1.8)
0.05 (2)
0
0.07 (2.9)
26/10/2007 12:43PM Plate # 0
Chapter 15:
1100 F). It was not clear what type of cover

gas was involved in these tests. The results
from both salt mixtures are summarized in
Fig. 15.5 and 15.6. Steels and Fe-Cr alloys suffered severe corrosion in both types of salts.
Chromium in Fe-Cr alloys and nickel in Fe-Ni
alloys improved performance. Fe-Cr-Ni alloys
performed signicantly better than steels and
Fe-Cr alloys.
Intergranular corrosion is the major corrosion
morphology by molten chloride salts. Figures 15.7 and 15.8 show typical intergranular
corrosion by molten chloride salt. Figure 15.7
shows the intergranular attack of a Ni-Cr-Fe
alloy (alloy 600) coupon welded to a heat treat
basket that underwent heat treat cycles involving
(750 and 930 F) in the molten LiCl-KCl

eutectic, which was being considered as an
electrolyte for lithium-sulfur fuel cells, Battles
et al. (Ref 11) also found many alloys resistant
to molten salt (Table 15.5). Tests were conducted in closed quartz crucibles. All the alloys
tested showed negligible corrosion rates. Aluminum in the aluminum-clad Type 434 SS
sample corroded at a higher rate due to the
galvanic couple between aluminum and stainless steel (Ref 11).
Takehara and Ueshiba (Ref 12) investigated
the corrosion behavior of steel, Fe-Cr, Fe-Ni, and
Fe-Cr-Ni alloys in molten 20NaCl-30BaCl250CaCl2 and molten 25LiCl-25ZnCl2-16BaCl224CaCl2-10NaCl at 500 and 600 C (930 and
Table 15.5
pg 413
Corrosion rates of several metals and alloys in molten LiCl-KCl eutectic

Corrosion rate, m/yr (mpy)
Annealed samples
Material
304
316
347
430
E-Brite
Al-Clad Type 434
Iron (99.999% Fe)
Armco electromagnet iron
Sensitized samples(a)
400 C (750 F)
500 C (930 F)
400 C (750 F)
500 C (930 F)
2 (0.08)
2 (0.08)
2 (0.08)
8 (0.32)
130 (5.1)
41 (1.6)
12 (0.47)
6 (0.24)
2 (0.08)
6 (0.24)
3 (0.12)
1 (0.04)
4 (0.16)
5 (0.2)
1 (0.04)
5 (0.2)
Tests were conducted in closed quartz crucibles. (a) Samples were sensitized at 650 C (1200 F) for 120 h. Source: Ref 11
Fig. 15.5
Corrosion rates of steel, Fe-Cr, Fe-Ni, and Fe-Cr-Ni alloys in molten 20NaCl-30BaCl2-50CaCl2 at 500 and 600 C (930 and
26/10/2007 12:43PM Plate # 0
a molten KCl salt bath at 870 C (1600 F) and a

quenching salt bath of molten sodium nitrate and
sodium nitrite at 430 C (800 F) for 1 month
(Ref 13). Figure 15.8 shows the intergranular
attack of a heat treat basket made of the same
alloy after service for 6 months in the same heat
treat cycling operation (Ref 13).
Another frequently observed corrosion morphology is internal attack by void formation
(Fig. 15.9) (Ref 13). Voids tend to form at grain
boundaries as well as in the grain interior (Ref 4).
The continuing formation and growth of chromium compounds at the metal surface causes
outward migration of chromium and inward
migration of vacancies, thus leading to internal
void formation (Ref 4).
Cold work may signicantly affect corrosion

of alloys in molten chloride salts. Lai (Ref 14)
discovered severe intergranular corrosion at the
sheared edge of a Ni-Cr-Fe-Mo alloy (alloy X)
coupon after exposure in a molten CaCl2-NaCl
salt bath at 570 C (1050 F). The sample edge
Fig. 15.6
Corrosion rates of steel, Fe-Cr, Fe-Ni, and Fe-CrNi alloys in molten 25LiCl-25ZnCl2-16BaCl224CaCl2-10NaCl at 500 and 600 C (930 and 1110 F). Source:
Ref 12
Fig. 15.7
Intergranular attack of a Ni-Cr-Fe alloy coupon

welded to a heat treat basket after service for
1 month in a heat treat operation cycling between a molten KCl
bath at 870 C (1600 F) and a quenching salt bath of molten
sodium nitrate-nitrite at 430 C (800 F). Source: Ref 13
Fig. 15.8
Intergranular attack of a Ni-Cr-Fe alloy heat treat

basket after service for 6 months in the same heat
treat cycling operation described in Fig. 15.7. Source: Ref 13
pg 414
26/10/2007 12:43PM Plate # 0
Chapter 15:
was cold worked during sample shearing prior to

exposure. There was no evidence of intergranular
corrosion in the area away from the sheared edge
(Fig. 15.10). The prior cold work also resulted in
a thicker oxide scale during exposure in the
molten salt. A Co-Ni-Cr-W alloy (alloy 188)
exhibited similar behavior.
15.4 Corrosion in Molten Nitrates/

Nitrides
Molten nitrates or nitrate-nitride mixtures are
widely used for heat treat salt baths, typically
operating from 160 to 590 C (325 to 1100 F).
They are also used as a medium for heat transfer
or energy storage. Molten drawsalt (NaNO3KNO3) is being considered as a heat-transfer
and energy-storage medium for a solar central
receiver for power generation from solar energy.
Numerous studies (Ref 1521) have been
carried out to determine potential candidate
containment materials for handling molten
drawsalt. Bradshaw and Carling (Ref 22)
recently summarized these studies as well as the
results of their study (Table 15.6) (Ref 22). The
data suggest that, for temperatures up to 630 C
(1170 F), many alloys are adequate for handling
molten NaNO3-KNO3 salt. Carbon steel and

2.25Cr-1Mo steel exhibited low corrosion rates
(<0.13 mm/yr, or <5 mpy) at 460 C (860 F).
At 500 C (932 F), 2.25Cr-1Mo steel exhibited
a corrosion rate of about 0.026 mm/yr (1 mpy).
Aluminized Cr-Mo steel showed higher resistance, with a corrosion rate of less than
0.004 mm/yr (<0.2 mpy) at 600 C (1110 F).
Austenitic stainless steels, alloy 800, and alloy
600 were more resistant than carbon steel and
Cr-Mo steels. Nickel, however, suffered high
corrosion rates.
Slusser et al. (Ref 23) evaluated the corrosion behavior of a variety of alloys in molten
NaNO3-KNO3 (equimolar volume) salt with an
equilibrium nitrite concentration (about 6 to
12 wt%) at 675 C (1250 F) for 336 h. A
Fig. 15.10
Fig. 15.9
Corrosion attack consisting of voids in a nickel-base

alloy after 2 months at 870 C (1600 F) in a molten
BaCl2 salt bath. Source: Ref 13
Corrosion of Ni-Cr-Fe-Mo alloy in a molten

CaCl2-NaCl salt bath at 570 C (1050 F) for
6 months. (a) Little corrosion at the area not cold worked (away
from the sheared edge). (b) Intergranular attack at the cold worked
area (specimen sheared edge). Source: Ref 14
pg 415
26/10/2007 12:43PM Plate # 0
pg 416
constant purge of air in the melt was maintained during testing. Nickel-base alloys were
generally much more resistant than iron-base
alloys. Increasing nickel content improved
alloy corrosion resistance to molten nitratenitrite salt. However, pure nickel suffered rapid
corrosion attack. Figure 15.11 shows the corrosion rates of various alloys as a function of
nickel content (Ref 23). Silicon-containing
alloys, such as RA330 and Nicrofer 3718,
performed poorly. A long-term test (1920 h) at
675 C (1250 F) was performed on selected
alloys, showing corrosion rates similar to
those obtained from 336 h exposure tests
(Table 15.7). Alloy 800, however, exhibited a
higher corrosion rate in the 1920 h test than in
the 336 h test. As the temperature was
increased to 700 C (1300 F), corrosion rates
became much higher, particularly for iron-base
alloy 800, which suffered an unacceptably high
corrosion rates (Table 15.7). Boehme and
Bradshaw (Ref 24) attributed the increased
corrosion rate with increasing temperature to
higher alkali oxide concentration. Slusser et al.
(Ref 23) found that adding sodium peroxide
(Na2O2) to the salt increased the salt corrosivity.
resistant to molten NaOH (Ref 2629), particularly low-carbon nickel such as Ni 201 (Ref 30).
Gregory et al. (Ref 29) reported corrosion rates of
several nickel-base alloys obtained from
static tests at 400 to 680 C (750 to 1256 F)
(Table 15.8). Molybdenum and silicon appear to
be detrimental alloying elements in molten
NaOH salt. Iron may also be detrimental.
Molybdenum and iron were found to be selectively removed from nickel-base alloys with less
than 90% nickel, leading to the formation of
internal voids (Ref 31).
Molten sodium hydroxide becomes increasingly aggressive with increasing temperature.
Coyle et al. (Ref 9) evaluated a variety of
alloys for a possible containment material for
molten sodium hydroxide operating at 900 C
(1650 F) for a solar power generation system.
Test results are tabulated in Table 15.9. Many
15.5 Corrosion in Molten Sodium

Hydroxide (Caustic Soda)
The reaction of metals with molten sodium
hydroxide (NaOH) leads to metal oxide, sodium
oxide, and hydrogen (Ref 25). Nickel is most
Table 15.6 Corrosion rates of selected
metals and alloys in molten NaNO3-KNO3
Alloy
Carbon steel
2.25Cr-lMo
9Cr-lMo
Aluminized Cr-Mo steel
12Cr steel
304SS
316SS
800
600
Nickel
Titanium
Aluminum
Source: Ref 22
Temperature
C (F)
Corrosion rate,
mm/yr (mpy)
460 (860)
460 (860)
500 (932)
550 (1020)
600 (1110)
600 (1110)
600 (1110)
600 (1110)
600 (1110)
630 (1170)
565 (1050)
600 (1110)
630 (1170)
600 (1110)
630 (1170)
565 (1050)
565 (1050)
565 (1050)
0.120 (4.7)
0.101 (4.0)
0.026 (1.0)
0.006 (0.2)
0.023 (0.9)
<0.004 (0.2)
0.022 (0.9)
0.012 (0.5)
0.0070.010 (0.30.4)
0.106 (4.2)
0.005 (0.2)
0.0060.01 (0.20.4)
0.075 (3.0)
0.0070.01 (0.30.4)
0.106 (4.2)
>0.5 (20)
0.04 (1.6)
<0.004 (0.2)
Fig. 15.11
Corrosion rates of various alloys as a function of

nickel content in the alloy tested in molten
NaNO3-KNO3 salt at 675 C (1250 F). Source: Ref 23
Table 15.7 Corrosion rates of selected alloys

at 675 and 700 C (1250 and 1300 F) in
sodium-potassium nitrate-nitrite salt
Alloy
214
600
N
601
800
Source: Ref 23
675 C (1250 F)
1920 h
700 C (1300 F)
720 h
0.41 (16)
0.25 (10)
0.23 (9.1)
0.48 (19)
1.85 (73)
0.53 (21)
0.99 (39)
1.22 (48)
1.25 (49)
6.6 (259)
26/10/2007 12:43PM Plate # 0
Chapter 15:
pg 417
conditions to about 8 mm/yr (320 mpy) at a

rotational speed of 600 rpm (Ref 32).
Metals or alloys in molten sodium hydroxide
are susceptible to mass transfer due to thermal
gradients in the melt. This causes corrosion in the
hot zone, and potential tube plugging in the
cold zone, of a circulating system. For example,
6.35 mm (0.25 in.) nickel tubing was plugged
after 5000 h at 440 to 480 C (830 to 900 F),
and after 50 h at 690 to 730 C (1280 to 1350 F)
(Ref 27).
iron-, nickel-, and cobalt-base alloys disintegrated in 84 h. Samples of the alloys that survived the 84 h exposure test were severely
corroded. Scales that formed on these samples
were reportedly cracked and spalled. The weightgain or weight-loss data of surviving samples
were no longer indicative of alloy performance
ranking. No metallographic examination was
performed on these samples. The authors concluded that no further studies on molten sodium
hydroxide were necessary, because the salt was
too aggressive to metallic materials operating at
900 C (1650 F). The marked inuence of
temperature on the corrosiveness of molten
sodium hydroxide is also demonstrated by the
results shown in Table 15.10 (Ref 26).
Corrosion of metals and alloys in molten
NaOH depends strongly on the velocity of the
salt. Gregory et al. (Ref 32) showed that corrosion of nickel under dynamic conditions was
enhanced by as much as several times at 540 C
(1000 F) and higher. The corrosion rate for
nickel at 680 C (1250 F), for example, varied
from about 1 mm/yr (40 mpy) under static
15.6 Corrosion in Molten Fluorides

Corrosion of alloys in molten uoride salts has
been extensively studied for nuclear reactor
applications. The molten salt nuclear reactor uses
a LiF-BeF2 base salt as a fuel salt, containing
various amounts of UF4, ThF4, and ZrF4
(Ref 33). The reactor coolant salt is a NaBF4-NaF
mixture (Ref 33). A nickel-base alloy, Hastelloy
alloy N, has proved to be the most corrosion
resistant in molten uoride salts (Ref 34). The
Table 15.8 Corrosion rates of selected nickel-base alloys obtained from static tests in molten
sodium hydroxide
Alloy
Ni-201
C
D
400
600
301SS
75
400 C (750 F)
500 C (930 F)
580 C (1080 F)
680 C (1260 F)
0.023 (0.9)
0.018 (0.7)
0.046 (1.8)
0.028 (1.1)
0.043 (1.7)
0.028 (1.1)
0.033 (1.3)
2.54 (100)
0.056 (2.2)
0.13 (5.1)
0.06 (2.4)
0.08 (3.2)
0.36 (14.3)
0.06 (2.5)
(a)
0.25 (9.9)
0.45 (17.6)
0.13 (5.1)
0.26 (10.4)
0.53 (20.8)
0.96 (37.8)
(a)
1.69 (66.4)
1.03 (40.7)
1.21 (47.6)
(a) Severe corrosion. Source: Ref 29
Table 15.10 Corrosion of various metals

and alloys in molten sodium hydroxide
sodium hydroxide at 900 C (1650 F) for 84 h
Alloy
304
316
800
800H
556
Nickel
600
214
X
N
S
230
X-750
R-41
188
Disintegrated
Disintegrated
+60
+65
Disintegrated
50
27
+160
+22
Fractured
4
Disintegrated
+35
Disintegrated
Disintegrated
N2-(0.11H2O)-(110O2) was used for the cover gas. Source: Ref 9
Weight change rate(a), mg/cm2-day

Material
Ni-201
Copper
Chromium
Aluminum
Silver
Gold
Platinum
70Au-30Pt
Palladium
Colmonoy No. 5
Colmonoy No. 6
Colmonoy No. 9
Chromel P
Zirconium
538 C
(1000 F)
816 C
(1500 F)
+0.12
+1.54
0.12
0.34
+0.02
0.33
0.32
+0.36
+6.59
+0.63
0.48
0.23
+0.09
+0.56
+1.7
2.2
40
+2.9
Broke
Broke
Broke
+175
+70
72
Dissolved in 24 h
Dissolved in 24 h
(a) Based on 24 h exposure tests. Source: Ref 26
26/10/2007 12:43PM Plate # 0
pg 418
alloy was the primary containment material

for a molten salt test reactor successfully operated from 1965 to 1969 (Ref 35). Koger
(Ref 33) reported a corrosion rate of less than
0.0025 mm/yr (0.1 mpy) at 704 C (1300 F) in
the LiF-BeF2 base salt (fuel salt) and about
0.015 mm/yr (0.6 mpy) at 607 C (1125 F) in
the NaBF4-NaF coolant salt for alloy N.
Iwamoto et al. (Ref 36) performed corrosion
tests in eutectic LiF-NaF-KF salt in a test loop
with a 750 C (1380 F) hot leg and a 685 C
(1265 F) cold leg for 500 h. Alloy N exhibited
only 2.06 mg/cm2 of maximum weight loss
at the hot leg.
At lower temperatures, austenitic stainless
steels showed good performance. Type 316
exhibited 0.015 mm/yr (0.59 mpy) at 650 C
(1200 F) in 66LiF-34BeF2 (mol%), and about
0.002 mm/yr (0.08 mpy) at 530 C (986 F) in
22LiF-31LiCl-47LiBr (mol%) (Ref 37). In a LiFBeF2 fuel salt containing UF4, ThF4, and ZrF4,
Type 304 suffered a corrosion rate of only about
0.028 mm/yr (1.1 mpy) at 690 C (1270 F)
(Ref 33).
Corrosion can become more aggressive as
temperature increases. It is particularly severe
for stainless steels because of tube-plugging
problems due to mass transfer. Adamson et al.
(Ref 38) conducted corrosion tests in a thermal
convection loop involving 43.5KF-10.9NaF44.5LiF-1.1UF4 (mol%) with an 815 C
(1500 F) hot leg and a 704 C (1300 F) cold
leg. Types 410, 430, 316, 310, and 347 suffered
severe tube-plugging problems at the cold leg
within short test durations. Nickel and nickelbase alloys, on the other hand, showed no plugging even after 500 h of testing. However, these
alloys suffered corrosion at the hot leg after
500 h of exposure. Alloy 600 suffered internal
attack consisting of voids about 0.30 to 0.38 mm
(12 to 15 mils) deep. Nimonic alloy 75 suffered
intergranular pitting about 0.20 to 0.33 mm (8 to
13 mils) deep, and nickel suffered even metal
removal of about 0.23 mm (9 mils).
Misra and Whittenberger (Ref 39) reported
corrosion data for a variety of commercial alloys
in molten LiF-19.5CaF2, which was being considered for a heat-storage medium in an advanced
solar space power system, at 797 C (1467 F)
for 500 h. The tests were conducted in alumina
crucibles with argon as a cover gas. Results are
tabulated in Table 15.11. For nickel-base alloys,
chromium was detrimental. No inuence of
chromium, however, was noted on iron-base
alloys.
Molten uorides are generally used in a closed

system under vacuum or an inert atmosphere.
However, hydrogen uoride may be present
in the system, resulting in increased corrosion
rates. Moisture, a common impurity in uoride
salts, can react with uorides to produce gaseous
HF (Ref 39), some of which may dissolve in the
melt. Corrosion by HF will also be involved,
leading to production of hydrogen (Ref 39).
Therefore, it is important to reduce the moisture
level in the salt to reduce corrosion attack
(Ref 39).
15.7 Corrosion in Molten Carbonates

Molten carbonates are generally less corrosive
than molten chlorides or hydroxides. Coyle et al.
(Ref 9) evaluated three different molten salts for
a possible heat-transfer and energy-storage
medium capable of operating 900 C (1650 F)
for a solar power generation system. Both the
eutectic sodium-potassium-magnesium chloride
(33NaCl-21.5KCl-45.5MgCl2, mol%) and the
sodium hydroxide were found to be too corrosive for many commercial alloys. The eutectic
sodium-potassium
carbonate
(58Na2CO342K2CO3, mol%), on the other hand, showed
promise because of the much lower corrosion
rates exhibited by many commercial alloys.
Corrosion data generated in molten carbonate
salt at 900 C (1650 F) for 504 h are
summarized in Table 15.12. The best performer
was Ni-Cr-Fe-Mo alloy (alloy X), followed
by Ni-Cr-Fe-Al alloy (alloy 214) and CoNi-Cr-W alloy (alloy 188). The Ni-Cr-Mo
Table 15.11 Results of corrosion tests in
LiF-19.5CaF2 at 797 C (1467 F) for 500 h
Depth of attack, m (mils)
Alloy
Mild steel
304
310
316
RA330
B
N
S
X
600
718
75
25
188
General(a)
Grain boundary(b)
30 (1.2)
15 (0.6)
90 (3.5)
90 (3.5)
45 (1.8)
30 (1.2)
155 (6.1)
185 (7.3)
130 (5.1)
165 (6.5)
270 (10.6)
15 (0.6)
140 (5.5)
30 (1.2)
120 (4.7)
135 (5.3)
95 (3.7)
105 (4.1)
Tests were conducted in alumina crucibles under argon. (a) Intragranular voids
near surface. (b) Intergranular voids. Source: Ref 39
26/10/2007 12:43PM Plate # 0
Chapter 15:
alloy (alloy S) was severely corroded. No

evidence of a systematic trend for the correlation
between alloying elements and performance is
noted. A signicant difference in corrosion was
observed between two samples of alloy 800
obtained from different suppliers. However, two
samples of alloy 600 showed good agreement.
At lower temperatures, molten carbonate corrosion generally becomes less severe. In molten
eutectic alkali metal carbonate, Grantham et al.
(Ref 40) found that many commercial alloys,
including Types 304L, 310, and 347, and alloys
600, C, N, X, and 25, exhibited low corrosion
rates (about 0.01mg/cm2/h or less) at 600 C
(1110 F). Nonoxidizing N2 was used for the
cover gas in their corrosion tests, which may
also have contributed to low corrosion rates.
The temperature dependence of corrosion rate
can also be seen in the results generated by
Grantham and Ferry (Ref 41) in the eutectic
Li2CO3-Na2CO3-K2CO3 (about equal weight) at
500, 600, and 700 C (930, 1110, and 1290 F).
The cover gas in this case was not reported.
Corrosion rates were found to be less than 0.025,
0.025, and 2.54 mm/yr (1, 1, and 100 mpy) at
500, 600, and 700 C (930, 1110, and 1290 F),
respectively.
15.8 Summary
The corrosion behavior of alloys in molten
chlorides, nitrates/nitrites, sodium hydroxide,
uorides, and carbonates was reviewed. In
eutectic sodium-potassium carbonate at 900 C
(1650 F) for 504 h
Alloy
X
214
188
556
X-750
600(b)
600(b)
R-41
N
304SS
316SS
230
Nickel
800(b)
800(b)
S

mm (mils)
0.12 (4.7)
0.19 (7.5)
0.22 (8.7)
0.26 (10.2)
0.27 (10.6)
0.34 (13.4)
0.44 (17.3)
0.42 (16.5)
0.51 (20.1)
0.54 (21.3)
0.63 (24.8)
0.77 (30.3)
>0.30 (11.8)
0.25 (9.8)
>0.8 (31.5)
>1.43 (56.3)
N2-0.1CO2-(110O2) was used for the cover gas. (a) All alloys showed metal loss
only except nickel, which suffered 0.2 mm (7.9 mils) metal loss and more than
0.11 mm (4.3 mils) intergranular attack. (b) Two samples from two different
suppliers. Source: Ref 9
general, corrosion rates of metals and alloys are

strongly dependent on temperature and can
generally be reduced by decreasing the temperature. Reducing oxidizing impurities, such as
oxygen and water vapor, in the melt can also
signicantly reduce the corrosiveness of the
molten salt. Thermal gradients in the melt, in the
case of circulating systems, may cause dissolution of an alloying element at the hot leg and
deposition of that element at the cold leg, leading
to potential tube-plugging problems.
Corrosion data generated from different
chloride salts at temperatures ranging from 400
to 900 C (750 to 1650 F) for various cast and
wrought alloys were presented. The data are
rather limited and fail to show a denitive correlation between alloying elements and performance. Intergranular attack is the major
corrosion morphology in molten chlorides. Cold
work may signicantly accelerate intergranular
attack. If the cold work effect is conrmed to be a
general phenomenon in molten chloride salts,
annealing of fabricated components prior to
service is recommended.
The corrosion behavior of commercial alloys
in molten NaNO3-KNO3 is reasonably well
understood. For applications at temperatures up
to 630 C (1170 F), many commercial alloys,
including austenitic stainless steels, are capable
of handling the molten salt. At higher temperatures, nickel-base alloys are preferred because of
the increased salt corrosivity. The resistance of
alloys to molten NaNO3-KNO3 improves with
increasing nickel content. Nickel, however, performed poorly.
Corrosion data related to molten sodium
hydroxide (caustic soda) are rather limited. At
900 C (1650 F), the salt is probably too corrosive for metallic materials to handle. Metals
and alloys may have adequate resistance to
molten sodium hydroxide at 680 C (1260 F)
and lower temperatures. Nickel, particularly
low-carbon nickel such as Ni201, is most resistant to the molten salt. Corrosion rates can be
signicantly increased under dynamic conditions.
Corrosion of alloys in molten uorides has
been extensively studied for nuclear reactor applications. A nickel-base alloy, Hastelloy alloy N,
has been found to be most suitable in molten
uoride salts. Since molten uorides are used in
a circulating system with temperature differentials, mass transfer induced corrosion due to
thermal gradients in the loop becomes an important issue. It creates the potential for tube fouling
pg 419
26/10/2007 12:43PM Plate # 0
and plugging at the cold leg of the loop.

Stainless steels are generally worse than nickel
and nickel-base alloys in terms of tube-plugging
problems.
Molten carbonates are generally less corrosive
than molten chlorides or hydroxides. Some of the
alloys that perform best in molten eutectic
Na2CO3-K2CO3 at high temperatures such as
900 C (1650 F) include Ni-Cr-Fe-Mo alloy
(alloy X), Ni-Cr-Fe-Al alloy (alloy 214), and
Co-Ni-Cr-W alloy (alloy 188).
16.
17.
18.
REFERENCES
1. D.G. Lovering, in Molten Salt Technology,

D.G. Lovering, Ed., Plenum Press, 1982, p 1
2. G.J. Janz and R.P.T. Tomkins, Corrosion,
Vol 35 (No. 11), 1979, p 485
3. J.W. Koger, in Corrosion, Vol 13, 9th ed.,
Metals Handbook, ASM International,
1987, p 51
4. J.W. Koger and S.L. Pohlman, in Corrosion,
5. A. Rahmel, in Molten Salt Technology, D.G.
Lovering, Ed., Plenum Press, 1982, p 265
6. J.H. Jackson and M.H. LaChance, Trans.
ASM, Vol 46, 1954, p 157
7. G.Y. Lai, M.F. Rothman, and D.E. Fluck,
9. R.T. Coyle, T.M. Thomas, and G.Y. Lai, in
High Temperature Corrosion in Energy
Systems, M.F. Rothman, Ed., The Metallurgical Society of AIME, 1985, p 627
10. H. Susskind, F.B. Hill, L. Green, S. Kalish,
L. Kukacka, W.E. McNulty, and E. Wirsing,
Chem. Eng. Prog., Vol 56 (No. 3), 1960, p 57
11. J.E. Battles, F.C. Mrazek, W.D. Tuohig, and
K.M. Myles, in Corrosion Problems in
Energy Conversion and Generation, C.S.
Tedmon, Jr., Ed., The Electrochemical
Society, 1974, p 20
12. K. Takehara and T. Ueshiba, J. Soc. Mater.
Sci. Jpn., Vol 179, 1968, p 755
13. S.K. Srivastava, unpublished results,
Haynes International, Inc., 1989
15. R.W. Bradshaw, Thermal Convection Loop
Corrosion Tests of Type 316SS and Alloy
800 in Molten Nitrate Salts, SAND
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
81-8210, Sandia Laboratory, Livermore,

CA, Feb 1982
P.F. Tortorelli and J.E. DeVan, Thermal
Convection Loop Study of the Corrosion of
Fe-Ni-Cr Alloys by Molten NaNO3-KNO3,
ORNL TM-8298, Oak Ridge National
Laboratory, Oak Ridge, TN, Dec 1982
R.W. Bradshaw, A Thermal Convection
Loop Study of Corrosion of Alloy 800 in
Molten NaNO3-KNO3, SAND 82-8911,
Sandia Laboratory, Livermore, CA, Jan
1983
R.W. Bradshaw, Kinetic Oxidation and
Elemental Depletion of Austenitic and Ferritic Steels in Molten Nitrate Salt, SAND
87-8011, Sandia Laboratory, Livermore,
CA, 1987
R.W. Bradshaw, Oxidation of ChromiumMolybdenum Steels by Molten Sodium
Nitrate-Potassium Nitrate, SAND 87-8012,
Sandia Laboratory, Livermore, CA 1987
R.W. Carling, R.W. Bradshaw, and R.W.
Mar, J. Mater. Energy Sys., Vol 4 (No. 4),
1983, p 229
R.W. Bradshaw, Oxidation and Chromium
Depletion of Alloy 800 and Type 316SS in
Molten NaNO3-KNO3 at Temperatures
above 600 C, SAND 86-9009, Sandia
Laboratory, Livermore, CA, Jan 1987
R.W. Bradshaw and R.W. Carling, A
Review of the Chemical and Physical
Properties of Molten Alkali Nitrate Salts
and Their Effect on Materials Used for Solar
Central Receivers, SAND 87-8005, Sandia
Laboratory, Livermore, CA, April 1987
J.W. Slusser, J.B. Titcomb, M.T. Heffelnger, and B.R. Dunbobbin, J. Met., July 1985,
p 24
D.R. Boehme and R.W. Bradshaw, High
Temp. Sci., Vol 18, 1984, p 39
G.P. Smith, Corrosion of Materials in Fused
Hydroxides, USAEC Report ORNL-2048,
March 1956
C.M. Craighead, L.A. Smith, and R.I. Jaffee,
Screening Tests on Metals and Alloys in
Contact with Sodium Hydroxide at 1000 and
1500 F, USAEC Report BMI-705, Nov
1951
E.M. Simmons, N.E. Miller, J.H. Stang, and
C. Weaver, Corrosion and Components
Studies on Systems Containing Fused
NaOH, USAEC Report BMI-1118, July
1956
R.A. Lad and S.L. Simon, A Study of Corrosion and Mass Transfer of Nickel by
pg 420
26/10/2007 12:43PM Plate # 0
Chapter 15:
29.
30.
31.
32.
33.
34.
35.
Molten Sodium Hydroxide, Corrosion, Vol

10 (No. 12), 1954, p 435
J.N. Gregory, N. Hodge, and J.V.G. Iredale,
The Static Corrosion of Nickel and Other
Materials in Molten Caustic Soda, AEREC/M-272, Atomic Energy Research Establishment, Harwell, U.K., March 1956
R.R. Miller, Thermal Properties of Sodium
Hydroxide and Lithium Metal, Quarterly
Progress Report May 1Aug 1, 1952, NRL3230-201/52, 1952
G.P. Smith and E.E. Hoffman, Corrosion,
Vol 13, 1957, p 627t
J.N. Gregory, N. Hodge, and J.V.G. Iredale,
The Corrosion and Erosion of Nickel by
Molten Caustic Soda and Sodium Uranate
Suspensions under Dynamic Conditions,
AERE Report C/M 273, Atomic Energy
Research Establishment, Harwell, U.K.,
March 1956
J.W. Koger, Corrosion, Vol 29 (No. 3),
1973, p 115
J.W. Koger, Corrosion, Vol 30 (No. 4),
1974, p 125
J.R. Keiser, D.L. Manning, and R.E.
Clausing, in Proc. Int. Symp. Molten Salts,
36.
37.
38.
39.
40.
41.
J.P. Pemsler et al., Ed., The Electrochemical

N. Iwamoto, Y. Makino, K. Furukawa,
Y. Katoh, and H. Katsuta, Trans. JWRI, Vol 9
(No. 2), 1980, p 117
J.R. Keiser, J.H. DeVan, and E.J. Lawrence,
J. Nucl. Mater., Vol 85/86, 1979, p 295
G.M. Adamson, R.S. Crouse, and W.D.
Manly, Interim Report on Corrosion by
Alkali-Metal Fluorides, ORNL-2337, Oak
Ridge National Laboratory, Oak Ridge, TN,
1959
A.K. Misra and J.D. Whittenberger,
Fluoride Salts and Container materials for
Thermal Energy Storage Applications in
Temperature Range 973 to 1400 K, NASA
Tech. Memo. 89913, NASA Lewis Research
Center, Cleveland, OH, 1987
L.F. Grantham, P.H. Shaw, and R.D.
Oldenkamp, in High Temperature Metallic
Corrosion of Sulfur and Its Compounds,
Z.A. Foroulis, Ed., The Electrochemical
L.F. Grantham and P.B. Ferry, in Proc. Int.
Symp. Molten Salts, J.P. Pemsler et al. Eds.,
The Electrochemical Society, 1976, p 270
pg 421

DOI: 10.1361/hcma2007p423
31/10/2007 12:52PM Plate # 0

CHAPTER 16
Liquid Metal Corrosion and

Embrittlement
Liquid metals are frequently used as a heattransfer medium because of their excellent
heat-transfer properties, such as high thermal
conductivities, high heat capacities, and low
vapor pressures. Various metals have been
investigated for use as coolants in nuclear reactors. Sodium, for example, has been used as a
coolant in fast breeder nuclear reactors. Most
corrosion studies of molten metals have been
carried out in conjunction with nuclear reactor
applications (Ref 14). Other applications of
liquid metals include heat treat baths (e.g.,
molten lead) and power generation (Ref 5).
The containment material, which is in contact
with the molten metal, is subject to molten metal
corrosion. This chapter reviews data relevant
to the corrosion of containment materials. Discussion of liquid metal embrittlement is also
included.
major alloying elements) with the molten metal

to form an intermetallic compound. This requires
some solubility of the liquid metal in the containment metal. When a molten metal is used as
a heat-transfer medium in a loop, mass transfer
due to thermal gradients in the melt can be a
critical issue (Ref 6). It involves dissolution of
an alloying element in hot zones and deposition
of that alloying element in cold zones, resulting
in tube plugging. The corrosion can also involve
impurity and interstitial interactions, which are
generally associated with light elements such
as carbon, nitrogen, and oxygen. This type of
reaction dictates the corrosion process when
the solubilities of the major alloying elements
in the molten metal are low. An example is carburization-decarburization behavior of alloys in
a sodium loop. More complete discussions on the
mechanisms of molten metal corrosion can be
found in articles by numerous authors (Ref 1, 2,
710).
16.2 Corrosion Process
16.3 Corrosion in Molten Aluminum
Molten metal corrosion of a containment metal

is most often related to its solubility in the molten
metal. This type of corrosion is simply a dissolution-type attack. A containment metal with
higher solubility in the molten metal generally
exhibits a higher corrosion rate. In the case of
an alloy, the solubilities of the major alloying
elements could dictate the corrosion rate. The
solubility of an alloying element in a molten
metal typically increases with increasing temperature. As the temperature increases, the
diffusion rate also increases. Thus, the alloy
corrosion rate increases with increasing temperature.
Corrosion by molten metal can also proceed
by alloying of the containment metal (or its
Aluminum melts at 660 C (1220 F). Iron,

nickel, and cobalt, along with their alloys, are
readily attacked by molten aluminum. Extremely
high corrosion rates of iron-, nickel-, and cobaltbase alloys in molten aluminum are illustrated
by the laboratory test results shown in Table 16.1
(Ref 11). Samples of carbon steel and ironand nickel-base alloys were consumed in 4 h
at 760 C (1400 F). Cobalt-base alloys, which
appeared to be better than iron- and nickel-base
alloys, were corroding at rates too high to be considered for containment materials. In addition,
titanium, although exhibiting a corrosion rate
lower than iron-, nickel-, and cobalt-base alloys,
should not be considered for use as a containment
material because of its rapid corrosion rate.
16.1 Introduction
pg 423
31/10/2007 12:52PM Plate # 0
pg 424
16.4 Corrosion in Molten Zinc

Zinc melts at 420 C (790 F). Molten zinc is
widely used in the hot dip galvanizing process to
coat steel for corrosion protection. Galvanizing
tanks, along with baskets, xtures, and other
accessories, require materials resistant to molten
zinc corrosion.
Nickel and high-nickel alloys react readily
with molten zinc by direct alloying. This is illustrated in Fig. 16.1, which shows a nickel-base
alloy coupon after immersion in molten zinc at
455 C (850 F). Iron- and cobalt-base alloys
are generally corroded by dissolution, even those
containing up to 33% Ni, such as alloy 800H

(Fig. 16.2). The results of static tests in molten
zinc for selected iron-, nickel-, and cobalt-base
alloys are summarized in Table 16.2. Nickel-base
alloys suffered the worst attack, followed by
austenitic stainless steels, Fe-Ni-Cr alloys, and
Fe-Cr alloys. Cobalt-base alloys generally performed better. However, an Fe-Ni-Co-Cr alloy
Table 16.1 Results of static immersion tests in

molten aluminum at 760 C (1400 F) for 4 h
Maximum depth of corrosion
attack, mm (mils)
Alloy
Titanium
6B
188
150
556
X
671
Carbon steel
0.22 (8.5)
0.43 (16.8)
0.51 (20.2)
0.73 (28.9)
>0.5 (20.6)(a)
>0.6 (23.8)(a)
>0.7 (26.3)(a)
>1.6 (63.1)(a)
(a)
200 m
(b)
200 m
(c)
200 m
(a) Sample was consumed. Source: Ref 11
Fig. 16.2
Fig. 16.1
Ref 11
Nickel-base alloy coupon after immersion testing

in molten zinc at 455 C (850 F) for 50 h. Source:
Sample cross sections for (a) alloy 556, (b) alloy

800H, and (c) Type 446 after immersion testing in
molten zinc at 455 C (850 F) for 50 h. The top edge of each
photograph represents the original surface of the sample prior to
immersion testing. Uniform dissolution of metal from the sample
surface is noted for all three alloys. Source: Ref 12
Chapter 16:
(556 alloy) performed as well as cobalt-base

alloys.
Intergranular cracking was observed for some
alloys in contact with molten zinc. Field testing
in a galvanizing tank, with test coupons immersed in molten zinc at approximately 455 C
(850 F) for 19 runs (8 h immersion per run) for
a total of 152 h, revealed intergranular cracking
for coupons of alloy 25 (Co-Ni-Cr-W alloy) and
Type 316, as shown in Fig. 16.3 (Ref 12). The
other two alloys in the same eld test (cobaltbase alloy 188 and Fe-Ni-Co-Cr alloy 556) did
not suffer cracking. The metallurgical factors
responsible for the cracking of alloy 25 and Type
316 are unclear. Liquid metal embrittlement
(LME) is discussed in Section 16.9.
Molten zinc corrosion is expected to become
more severe at higher temperatures. However,
very little data, particularly at 500 C (930 F)
and higher, have been reported in the literature.
31/10/2007 12:52PM Plate # 0
pg 425
Liquid Metal Corrosion and Embrittlement / 425
(Ref 15). For example, cast iron centrifugal

pumps are used to pump liquid lead (Ref 15).
Ali-Khan (Ref 16) performed extensive corrosion tests in molten lead. The results of his
static corrosion tests at temperatures from 575
to 750 C (1070 to 1380 F) for chromium steels
(7 to 17Cr) and austenitic stainless steels (18Cr-8
to 16Ni) are shown in Tables 16.3 and 16.4,
respectively. Both exhibited similar performance, with corrosion rates of less than 0.15 mm
(6 mils) after 3250 h of exposure.
Wilkinson et al. (Ref 17) investigated molten
lead corrosion at a much higher temperature.
These authors performed static tests at 1000 C
(1830 F) for up to 408 h. Their test results
(Table 16.5) showed that the 17Cr steel (Type
430) was the most resistant. A sample of cast
iron, on the other hand, completely dissolved in
408 h.
16.5 Corrosion in Molten Lead

Lead melts at 327 C (620 F). Nickel and
nickel-base alloys generally have poor resistance
to molten lead corrosion (Ref 13, 14). The solubility of nickel in molten lead is higher than that
of iron. Cast iron, steels, and stainless steels
are commonly used for handling molten lead
Table 16.2 Results of static immersion tests
in molten zinc at 455 C (850 F) for 50 h
Alloy
Depth of corrosion attack, mm (mils)
556
25
188
446SS
800H
304SS
625
X
0.04 (1.6)
0.06 (2.3)
0.06 (2.5)
0.24 (9.3)
0.28 (11.0)
0.36 (14.1)
>0.61 (24.0)(a)
>0.61 (24.0)(a)
(a)
100 m
(b)
100 m
(a) Complete alloying. Source: Ref 11
Table 16.3 Results of static corrosion tests

at various temperatures for 3250 h in molten
lead for chromium steels
Alloy
7Cr-0.8Al
13Cr
13Cr-1.0Al
13Cr-1.0Mo
17Cr
575 C (1067 F)
650 C (1200 F)
750 C (1382 F)
25 (1.0)
10 (0.4)
10 (0.4)
10 (0.4)
(a)
(a)
15 (0.6)
15 (0.6)
30 (1.2)
(a)
30 (1.2)
100 (3.9)
150 (5.9)
35 (1.4)
(a) Little or no corrosion. Source: Ref 16
Fig. 16.3
Intergranular cracking for (a) alloy 25 and (b) Type

316 coupons after field testing in a hot dip galvanizing tank at 455 C (850 F) for a total of 152 h (19 runs).
Source: Ref 12
31/10/2007 12:52PM Plate # 0
pg 426
The corrosion behavior of both chromium

steels and austenitic stainless steels in a thermal
convection loop was studied by Ali-Khan
(Ref 16). The results are shown in Tables 16.6
and 16.7. Compared with tests under static conditions (Tables 16.3 and 16.4), alloys under
thermal convection conditions corroded at faster
rates. Tolson and Taboada (Ref 18) also reported
corrosion data generated from a thermal convection loop (Table 16.8). The results of the
Brookhaven loop, which used magnesium and
zirconium as inhibitors in the melt, are included
in the table for comparison. It appears that the
inhibitors signicantly reduced corrosion by
Table 16.6 Results of corrosion tests in a

molten lead thermal convection loop for 1002 h
at 600 C (1112 F, hot leg) with a temperature
differential of 232 C (450 F) for chromium steels
Table 16.4 Results of static corrosion tests

at various temperatures for 3250 h in molten
lead for austenitic stainless steels
Alloy

Alloy
18Cr-9Ni
17Cr-8Ni
18Cr-8Ni
18Cr-10Ni
18Cr-11Ni-2Mo
18Cr-12Ni-2.5Mo
18Cr-12Ni-2Mo
18Cr-16Ni-3.5Mo
18Cr-10Ni
575 C
(1067 F)
650 C
(1200 F)
750 C
(1382 F)
40 (1.6)
6090 (2.43.5)
10 (0.4)
40 (1.6)
55 (2.2)
25 (1.0)
40 (1.6)
80 (3.2)
110 (4.3)
10 (0.4)
15 (0.6)
20 (0.8)
25 (1.0)
15 (0.6)
(a)
10 (0.4)
35 (1.4)
35 (1.4)
25 (1.0)
10 (0.4)
(a)
150 (5.9)
10 (0.4)
70 (2.8)
15 (0.6)
20 (0.8)
20 (0.8)
(a) Little or no corrosion. Source: Ref 16
Table 16.5 Corrosion of several alloys in molten

lead at 1000 C (1830 F) for up to 408 h
Alloy
Cast iron
ARMCO iron
1020 steel
430SS
302SS
347SS
Complete dissolution in 408 h

4.0 (157)
2.5 (100)
0.7 (27)
10.0 (393)
9.8 (387)
Source: Ref 17
Table 16.8
molten lead. Croloy 2.25Cr steel revealed no

detectable corrosion attack in the Brookhaven
loop after 27,765 h of exposure, while the same
alloy showed 0.2 mm (8 mils) of attack in the
ORNL loop containing no inhibitors, although
the ORNL loop was operated 26 C (47 F)
higher. The benecial effect of inhibitors has
been conrmed by Asher et al. (Ref 19). It is
believed that inhibitors are instrumental in the
formation of a protective lm, which reduces the
solubility of the containment metal in molten
lead (Ref 20).
7Cr-0.8Al
13Cr steel
13Cr-1.0Al
13Cr-1.0Mo
17Cr
130180 (5.17.1)
273300 (1112)
187200 (7.47.9)
143240 (5.69.5)
130140 (5.15.5)
Source: Ref 16
Table 16.7 Results of corrosion tests in a

molten lead thermal convection loop for 1008 h
at 600 C (1112 F, hot leg) with a temperature
differential of 151 C (304 F) for austenitic
stainless steels
Alloy
18Cr-9Ni
17Cr-8Ni
18Cr-8Ni
18Cr-10Ni
18Cr-11Ni-2Mo
18Cr-12Ni-2.5Mo
18Cr-12Ni-2Mo
18Cr-16Ni-3.5Mo
18Cr-10Ni
80140 (3.25.5)
5380 (2.13.2)
5380 (2.13.2)
62100 (2.43.9)
80140 (3.25.5)
80 (3.2)
130200 (5.17.9)
120140 (4.75.5)
5570 (2.22.8)
Source: Ref 16
Results of thermal-convection loop tests in molten lead
Loop material
Maximum
temperature,
C (F)
DT, C (F)
Operated
time, (h)
Maximum
depth of attack,
m (mils)
Inhibitors
Croloy 2.25Cr steel

410
Croloy 2.25Cr steel
Carbon steel (A-106)
Nb-1Zr (cladded on type 446)
Croloy 2.25Cr steel(a)
654 (1210)
654 (1210)
593 (1100)
593 (1100)
760 (1400)
550 (1022)
149 (300)
149 (300)
93 (200)
93 (200)
204 (400)
105 (221)
266
1,346
5,156
5,064
5,280
27,765
25 (1.0)
50 (2.0)
203 (8.0)
254 (10)
None
None
None
None
None
None
None
Mg+Zr
(a) Brookhaven loop. Source: Ref 18
31/10/2007 12:52PM Plate # 0
Chapter 16:
16.6 Corrosion in Molten Lithium

Liquid lithium is an attractive candidate
blanket material for fusion reactors. Numerous
studies (Ref 2127) have evaluated the compatibility of various containment materials with
liquid lithium. This section gives a brief summary of the comparative performance of alloys
in molten lithium.
The solubilities of some important alloying
elements in liquid lithium are shown in Fig. 16.4
(Ref 25). Nickel has the highest solubility
among the metals investigated by Cleary et al.
(Ref 25). Thus, nickel and nickel-base alloys
are generally not considered good candidates
for handling liquid lithium. Cobalt-base alloys
are only slightly more resistant than nickel-base
alloys (Ref 21, 22, 28).
pg 427
Figure 16.5 illustrates the performance of

several iron-, nickel-, and cobalt-base alloys in
molten lithium at 890 C (1635 F) (Ref 28).
Two nickel-base alloys (RA333 and Hastelloy
alloy X) suffered the worst corrosion attack in
terms of weight loss, followed by a cobalt-base
alloy (Airesist 213) with two iron-base alloys
(E-Brite alloy and Type 304L) being the best
performers. However, in terms of penetration
depths, Airesist 213, E-Brite, and Type 304L
behaved similarly (Fig. 16.6). TZM (Mo-0.5Ti0.1Zr) was also included in the investigation,
revealing no corrosion attack at all.
Freed and Kelly (Ref 21) performed corrosion
tests in circulating molten lithium at 705 to
815 C (1300 to 1500 F) for iron and stainless
steels (Table 16.9). Both suffered considerable
metal loss due to dissolution. For austenitic
Temperature, F
101
1700
1600
1500
1400
Nickel
1300
1200
Nickel
Chromium
Iron
102
Titanium
Solubility, atomic percent solute
Titanium results reported as being <10 ppm

Molybdenum
Columbium
103
Chromium
Iron
104
Titanium
Columb
ium
Molybdenum
105
0.82
0.86
0.90
0.94
0.98
1.02
1.03
1.10
1000 / T , K
Fig. 16.4
Solubilities of some metals in molten lithium. Note: Columbium is the former (pre-1968) name of niobium. Source: Ref 25
31/10/2007 12:52PM Plate # 0
pg 428
stainless steels, intergranular attack was also

observed. The stabilized stainless steels, such as
Types 321 and 347, were more resistant to intergranular attack. Also observed were signicant
amounts of deposits due to mass transfer. This
can result in ow restrictions, or eventual tube
plugging, in a circulating system. Chopra et al.
(Ref 24) summarized the data generated from
different circulating loops (forced circulation
loop, FCL, and thermal convection loop, TCL)
at different laboratories. The data, illustrated in
Fig. 16.7, are presented in terms of an Arrhenius

plot of dissolution rates for ferritic steels (HT-9
and 9Cr-1Mo) and austenitic stainless steels
(Type 316, cold-worked Type 316, and Type
304). HT-9 exhibits excellent compatibility with
molten lithium.
16.7 Corrosion in Sodium

Liquid sodium is used as a coolant for fast
breeder reactors. Extensive studies (Ref 2935)
on the corrosion of various alloys in liquid
1
10
RA-333
Hastelloy x
1
Airesist-213
102
E-Brite 26-1
Penetration, mm
Weight loss, gm/cm2
101
RA-333
Hastelloy x
Airesist-213
101
304L s.s
E-Brite 26-1
304L stainless steel

103
TZM (no attack)
104
105
TZM (no attack)

106
102
104
Weight loss as a function of time for selected alloys

in molten lithium (Ti-gettered) at 890 C (1635 F).
Source: Ref 28
106
Time, s
Time, s
Fig. 16.5
105
Fig. 16.6
Total penetration as a function of time for selected

alloys in molten lithium (Ti-gettered) at 890 C
Table 16.9 Results of corrosion tests in molten lithium at 705 to 815 C (1300 to 1500 F) in a
forced-convection loop
Alloy
Exposure
time, h
Maximum depth of
attack(a), mm (mils)
Maximum thickness
of deposits, mm (mils)
Iron(b)
108138
0.320.38 (12.515.0)
Iron
138187
304
105138
310
6496
321
69200
347
82160
IG 0.0
ML 0.050.11 (2.04.5)
IG 0.00.02 (0.00.6)
ML 0.130.17 (5.06.6)
IG 0.030.13 (1.05.0)
ML 0.100.15 (3.86.0)
IG 0.080.10 (3.04.0)
ML 0.060.12 (2.24.7)
IG 0.00.05 (0.02.0)
ML 0.150.16 (6.26.4)
IG 0.010.02 (0.50.6)
ML 0.110.12 (4.34.9)
0.360.46 (14.018.0)
0.470.51 (18.520.0)
0.490.61 (19.524.0)
0.640.81 (25.032.0)
0.841.02 (33.040.0)
(a) ML, metal loss due to apparent solution attack, decrease in wall thickness; IG, intergranular attack. (b) Titanium getter in lithium ow stream. Source: Ref 21
31/10/2007 12:52PM Plate # 0
Chapter 16:
pg 429
corrosion database for a wide variety of alloys

has been published in Nuclear Systems Materials
Handbook (Ref 36) to provide guidance to
corrosion allowance for design calculation. This
sodium have been carried out in support of

sodium breeder reactor programs. In addition to
numerous papers and reports discussing the
behavior of alloys in liquid sodium, a large
Temperature, C
650
103
600
550
500
Type 316 stainless steel TCL

and
HT-9 TCL and FCL
450
400
350
Weight loss in lithium
ORNL
ANL
WARD
UW
102
SU
Dissolution rate, mg/m2.h
JAERI
10
Type 316 stainless

steel FCL
Scatter band
type 316 stainless
steel FCL
TCL
Type 316 stainless steel
PCA
HT-9
FCL
Type 316 CW stainless steel
PCA
HT-9
9 Cr-1Mo
101
1.0
1.1
1.2
1.3
1.4
1.5
1.6
1.7
1000 /T, K
Fig. 16.7
Corrosion rates of Types 304 and 316, PCA (primary candidate alloy), and HT-9 and 9Cr-1Mo steels in flowing lithium. CW,
cold worked; FCL, forced-convection loop; TCL, thermal-convection loop. Source: Ref 24
31/10/2007 12:52PM Plate # 0
pg 430
section gives a very brief summary of the comparative performance of alloys in liquid sodium.
Berry (Ref 1) summarized the data generated by numerous authors (Ref 2934). The data
were generated in both static and owing systems at temperatures from 550 to 595 C (1025
to 1100 F), with some data generated at
1000 C (1830 F). Carbon steel, Cr-Mo steels,
and ferritic and austenitic stainless steels exhibited low corrosion rates at temperatures up to
595 C (1100 F) (Table 16.10). The major problem for low-chromium alloy steels in sodium is
decarburization and resultant loss of strength
(Ref 35). For austenitic stainless steels and
nickel-base alloys, the reaction between the alloy
and the sodium leads to carburization (Ref 35).
Corrosion of alloys in liquid sodium can be
severe at higher temperatures. Ferritic and austenitic stainless steels suffered rapid corrosion
attack at 1000 C (1830 F) (Table 16.10). A
nickel-base alloy (alloy 600) also exhibited a
rapid corrosion rate (Table 16.10). Borgstedt
et al. (Ref 37) investigated the corrosion behavior
of several Fe-Ni-Cr and nickel-base alloys in
liquid sodium at 1000 C (1830 F). Their
results are summarized in Table 16.11. Nickelbase alloys suffered more attack than Fe-Ni-Cr
alloys.
16.8 Corrosion in Other Molten Metals

Corrosion data for other liquid metals, such
as magnesium (melting point of 651 C, or
1205 F); cadmium (melting point of 321 C,
or 610 F); mercury (melting point of 39 C,
or 38 F); tin (melting point of 232 C, or
450 F); antimony (melting point of 631 C, or
1165 F); and bismuth (melting point of 271 C,
Table 16.10 Corrosion of carbon steel,

chromium-molybdenum steels, and stainless steels
in liquid sodium under isothermal conditions
Materials
1010 steel
2.25Cr-1Mo
5Cr-0.5Mo
9Cr-1Mo
410
420
304
310
316
347
410
430
446
304
316
310
347
600
Test
temperature,
C (F)
Exposure
time, h
Test
system
Weight
change rate,
mg/cm2
per month
593 (1100)
593 (1100)
552 (1026)
556 (1033)
593 (1100)
593 (1100)
552 (1026)
566 (1033)
593 (1100)
593 (1100)
552 (1026)
566 (1033)
593 (1100)
593 (1100)
593 (1100)
593 (1100)
593 (1100)
593 (1100)
593 (1100)
593 (1100)
593 (1100)
593 (1100)
593 (1100)
593 (1100)
593 (1100)
593 (1100)
1000 (1830)
1000 (1830)
1000 (1830)
1000 (1830)
1000 (1830)
1000 (1830)
1000 (1830)
1000 (1830)
1000
1000
943
902
1000
1000
943
1913
500
500
943
902
500
500
1000
1000
1000
1000
1000
1000
500
500
1000
1000
500
500
400
400
400
400
400
400
400
400
Flowing
Static
Flowing
Static
Flowing
Static
Flowing
Static
Flowing
Static
Flowing
Static
Flowing
Static
Flowing
Static
Flowing
Static
Flowing
Static
Flowing
Static
Flowing
Static
Flowing
Static
Static
Static
Static
Static
Static
Static
Static
Static
0.49
0.37
0.12
0.12
0.14
0.09
+0.22
-0.06
+0.23
0.08
+0.35
0.05
+0.70
+0.29
+0.38
+0.35
+0.33
+0.31
+0.17
+0.15
+0.75
+0.27
+0.10
+0.13
+1.46
+0.22
+29.8
+46.8
+28.2
+25.5
+29.6
+28.2
+44.2
+18.7
Note: Sodium contained a maximum of 100 ppm oxygen. Source: Ref 1, based on
Ref 2934
Table 16.11 Corrosion of several iron-nickel-chromium and nickel-base alloys in liquid sodium
at 1000 C (1830 F)
Test
Run No. 1, 1000 h
Run No. 2, 1100 h
Source: Ref 37
Alloy
Cr
Ni
Fe
Thermon
617
X
AC-66
ASL71
Pyrotherm
800
625
625
617
X
Pyrotherm
253-MA
22
21
21
27
20
20
20
22
22
21
21
20
21
bal
bal
bal
32
20
33
33
bal
bal
bal
bal
33
11
30
1.5
18
bal
bal
bal
bal
2.5
2.5
1.5
18
bal
bal
Others
W, Nb
Co, Mo, Al
Mo
Nb, Ce
Co, W
Nb
Al, Ti
Mo, Nb
Mo, Nb
Co, Mo, Al
Mo
Nb
Si
13.0
13.0
4.5
2.35
1.94
2.0
0.7
36.01
35.35
25.11
16.32
+1.43
+9.95
Chapter 16:
or 520 F); are rather limited. Frequently, it is

possible to use phase diagrams to determine
which metals may react readily with the liquid
metal. It is clear from the Ni-Mg phase diagram
that magnesium reacts readily with nickel to
form a low-melting-point eutectic, with a large
liquid phase eld extending to low temperatures.
More than 50% Ni, for example, can be in solution with magnesium in liquid at 800 C
(1470 F). Nickel and nickel-base alloys are
therefore not suitable for use as a containment
material for molten magnesium. Cobalt also
forms a low-melting-point eutectic, but with a
much smaller liquid-phase eld. The solubility of
iron in liquid magnesium is very low. It is thus
reasonable to assume that iron-base alloys with
low nickel are more suitable than cobalt- and
nickel-base alloys to handle molten magnesium.
The solubility of iron increases with increasing
temperature. Thus, the corrosion rate is expected
to increase with temperature as well.
The solubility of iron in molten cadmium
is low. Daniels (Ref 38) indicated the inactivity
of steel in molten cadmium. Tammann and
Oelsen (Ref 39) reported a solubility of 2 to
3 104 wt% Fe in molten cadmium at 400 and
700 C (750 and 1290 F). Jackson and Adams
Temperature, C
780
1000
727
679
636
596
560
527
496
Ti
Zr
Ni
Ni
Zr
100
Ti
10
ppm in Hg
Cr
31/10/2007 12:52PM Plate # 0
pg 431
(Ref 15) indicated that cast iron has frequently

been used for handling molten cadmium.
Nickel and nickel-base alloys are not suitable
for handling molten tin because of relatively high
solubilities of nickel in molten tin. Cast iron is
often used in the laboratory for handling molten
tin (Ref 15). Cobalt, nickel, and iron have high
solubilities in molten antimony. These metals, as
well as their alloys, are expected to have poor
corrosion resistance in molten antimony.
The solubilities of a number of metals in liquid
mercury are illustrated in Fig. 16.8 (Ref 40).
Both iron and cobalt have low solubilities in
liquid mercury. Table 16.12 illustrates corrosion
rates of carbon steels and stainless steels in liquid
mercury (Ref 8).
16.9 Liquid Metal Embrittlement

Metallic components may suffer embrittlement when in contact with liquid metal. Components are more susceptible to liquid metal
embrittlement (LME) when under stresses.
McDonalt (Ref 41) indicated that stainless steels,
nickel alloys and Cu-Ni alloys may suffer LME
by the molten brazing alloy while the component
is stressed under brazing. Since the brazing
operation involves contact of the metallic component with a molten brazing alloy, LME can
occur during brazing operation. Heiple et al.
(Ref 42) reported that austenitic stainless steels
can be severely embrittled by copper and high
copper braze alloys. However, austenitic stainless steels are not affected by silver-base braze
alloys (Ref 42). It was also found that Type
430 (a ferritic stainless steel) was not embrittled
by copper-base braze alloys, silver-base braze
alloys, aluminum, or gold (Ref 42). Small
Co
Fe
1.0
0.10
Table 16.12 Corrosion of metals and alloys in

flowing liquid mercury
Nb
Ta < 0.002
0.01
0.95
1.0
1.05
1.1
Maximum
temperature,
C (F)
Test
duration, h
Corrosion rate,
mm/yr (mpy)
Mild steel
482 (900)
538 (1000)
593 (1100)
649 (1200)
482 (900)
538 (1000)
593 (1100)
649 (1200)
652 (1205)
650 (1200)
(a)
(a)
(a)
(a)
(a)
(a)
(a)
(a)
460
400500
0.10 (4)
0.23 (9)
0.56 (22)
1.35 (53)
0.05 (2)
0.10 (4)
0.25 (10)
0.64 (25)
0.51 (20)
1.19 (47)
5Cr steel
1.15
1.2
1.25
1.3
1000/ T, K 1
Fig. 16.8
Material
Solubilities of some metals in liquid mercury.

Note: Columbium is the former (pre-1968) name
of niobium. Source: Ref 40
304
310
(a) Average of a large number of laboratory tests as well as samples from largescale boiler operations; exposures up to 10,000 h. Source: Ref 8
31/10/2007 12:52PM Plate # 0
amounts of copper contaminant on the surface

of the alloy that is under welding can cause LME
in the heat-affected zone of cobalt-base alloys
(Ref 43, 44). Savage and Mushala (Ref 44)
believed that copper and cobalt are essentially
insoluble in each other and thus form a classic
LME couple.
Korb (Ref 45) summarized some of the LME
issues of various alloy combinations involved in
the construction of manned spacecraft as follows.
Many structural alloys are embrittled by lowmelting-point metals. Aluminum is embrittled
by mercury, indium, tin, and zinc; steel by tin,
cadmium, zinc, lead, copper, and lithium; stainless steels by cadmium, aluminum, lead, and
copper; titanium by cadmium and mercury; and
nickel by zinc, cadmium, and mercury. Many of
these embrittling elements are used in low-temperature solders for electrical and avionics
applications, for example, lead-tin for electrical
soldered connections, indium-base solders for
(a)
(b)
Fig. 16.9
(a) Fractured studs due to cadmium-induced

LME during service at 315 C (600 F), and
(b) optical micrograph showing intergranular cracks at the cross
section of a fractured stud. Source: Ref 46
glass-metal seals. Most silver solders (for higherstrength applications) melt in a temperature
range of 610 to 800 C (1125 to 1475 F) and
may contain copper, cadmium, zinc, lithium, and
tin. Alloys are more susceptible to LME under
stress. Many brazing alloys that are used for
many stainless steel plumbing systems contain
copper. For example, brazed manifold tube joints
using Nicoro 80 braze alloy (81.5Au-16.5Cu2Ni) to join 21-6-9 stainless steels for the shuttle
orbiter were found to suffer cracking. The
cracking also was observed for brazed joints for
304L and alloy 718.
Ebert (Ref 46) indicated that cadmium-plated
Cr-Mo steel (ASTM A 193 grade B) studs from a
steam line connector associated with a power turbine fractured during service at 315 C (600 F)
by cadmium-induced LME. Figure 16.9(a) shows
the fractured studs, and Fig. 16.9(b) shows intergranular cracks in the cross section of a fractured
stud.
Zinc has been widely used in the industry as
a corrosion-resistant coating for carbon steels
(e.g., hot dip galvanizing, electroplating, and
spray painting) (Ref 47). Korbrin (Ref 47) indicated (a) carbon steels are susceptible to LME by
zinc particularly under stresses or cold-worked
conditions, (b) austenitic stainless steels and
nickel alloys can suffer LME when in contact
with molten zinc, or when welded to galvanized
steels or parts contaminated with zinc. Intergranular cracking was observed in Type 316 and
alloy 25 (Co-20Cr-10Ni-15W) coupons after
exposure to molten zinc in a hot dip galvanizing
tank at 455 C (850 F) for 50 h, as shown in
Fig. 16.3 (Ref 12).
Dillon (Ref 48) found that a Type 321 nozzle
suffered LME due to molten zinc contamination
of welds from zinc-pigmented painting overspray during initial fabrication. Dillon (Ref 49)
observed the failure of ASTM A193 2H nuts
during service at 370 C (700 F) and attributed
the failure to LME by zinc due to cadmiumplated/zinc phosphate coated nuts. The Cd-Zn
eutectic melts at 270 C (515 F). Gutzeit et al.
(Ref 50) indicated that austenitic stainless
steels are susceptible to LME by zinc when
welding or during heat treatment of stainless steel
components contaminated with zinc-rich paint.
Figure 16.10 shows the formation of intergranular cracking at the heat-affected zone of a
Type 304 pipe weld joint when the area was
contaminated with zinc-rich paint during welding. Zinc-rich paints containing only metallic
zinc powders as a principal component can cause
pg 432
Chapter 16:
100 m
Fig. 16.10
Intergranular cracking in the heat-affected zone

of a weld joint for Type 304 pipe that was contaminated with zinc-rich paint when welding was performed.
Source: Ref 50
LME, while paints containing zinc oxide or zinc

chromates do not cause LME (Ref 50).
31/10/2007 12:52PM Plate # 0
Cast iron, steels, and stainless steels are commonly used for handling molten lead. Nickel and
nickel-base alloys have poor resistance to molten
lead corrosion because of the high solubility of
nickel in molten lead.
Nickel also has relatively higher solubility
than iron in molten lithium, as well as in molten
sodium. Thus, nickel-base alloys are not good
candidates for handling either metal. Iron-base
alloys are more suitable.
Nickel reacts more readily with molten magnesium. The solubility of iron in molten magnesium is low. Thus, iron-base alloys with low
nickel contents are preferred as containment
materials. Due to low solubilities in the molten
metal, iron and iron-base alloys may be suitable
for molten cadmium, while iron, cobalt, and
iron- and cobalt-base alloys may be suitable for
molten mercury. Iron, nickel, and cobalt exhibit
high solubilities in molten antimony. These
metals and their alloys are thus not recommended
for use as containment materials for molten
antimony.
Liquid metal embrittlement of various alloys
by low-melting-point metals, such as copper,
zinc, cadmium, mercury, lead, tin, and so forth, is
reviewed. LME cases are often associated with
welding, brazing, and soldering. Most brazing
and soldering alloys contain alloying elements
that can induce LME. The components are more
susceptible to LME when they are under stress
during welding or soldering.
16.10 Summary
The corrosion behavior of alloys in molten
aluminum, zinc, lead, lithium, sodium, magnesium, mercury, and other molten metals is
reviewed. Corrosion data useful in assisting
selection of materials are presented. Also presented is the information about the LME caused
by a wide variety of low-melting-point metals
during welding or heat treatment.
Molten aluminum is extremely aggressive.
Iron-, nickel-, and cobalt-base alloys are readily
attacked by molten aluminum. Molten zinc is
less aggressive. Nickel and nickel-base alloys,
however, react readily with molten zinc and
are not recommended for use. Cast iron, steels,
and iron- and cobalt-base alloys are generally
suitable for containment applications. Cobaltbase alloys are generally more resistant than ironbase alloys. However, under some conditions,
various metals or alloys may be susceptible to
LME by molten zinc.
REFERENCES
1. W.E. Berry, Corrosion in Nuclear Applications, John Wiley & Sons, 1971
2. H.U. Borgstedt, Ed., Materials Behavior
and Physical Chemistry in Liquid Metal
Systems, Plenum Press, 1982
3. J.E. Draley and J.R. Weeks, Ed., Corrosion
by Liquid Metals, Plenum Press, 1970
4. C. Bagnall and W.F. Brehm, Corrosion,
5. D.L. Katz, Liquid-Metals Handbook,
R.N. Lyon, Ed., NAVEXOS, P-733 (Rev.),
U.S. Government Printing Ofce, Washington, D.C., 1952, p 1
6. J.V. Cathcart and W.D. Manly, Corrosion,
Vol 12, 1956, p 87t
7. P.F. Tortorelli, Corrosion, Vol 13, 9th ed.,
Metals Handbook, ASM International,
1987, p 56
pg 433
31/10/2007 12:52PM Plate # 0
8. E.C. Miller, Liquid-Metals Handbook,

R.N. Lyon, Ed., NAVEXOS, P-733 (Rev.),
U.S.
Government
Printing
Ofce,
Washington, D.C., 1952
9. W.D. Manly, Corrosion, Vol 12 (No. 7),
1956, p 336t
10. A. Brasunas, Corrosion, Vol 9, 1953, p 78
12. S.K. Srivastava, Proc. Conf. Performance of
High Temperature Materials in Fluidized
Bed Combustion Systems and Process
Industries, P. Ganesan and R.A. Bradley,
13. F.R. Morrall, Wire Wire Prod., Vol 23, 1948,
p 484, 571
14. L.R. Kelman, W.D. Wilkinson, and F.L.
Yaggee, Resistance of Materials to Attack
by Liquid Metals, Report ANL-4417,
Argonne National Laboratory, 1950
15. C.B. Jackson and R.M. Adams, LiquidMetals Handbook, R.N. Lyon, Ed.,
NAVEXOS, P-733 (Rev.), U.S. Government
Printing Ofce, Washington, D.C., 1952
16. I. Ali-Khan, Materials Behavior and Physical Chemistry in Liquid Metal Systems,
H.U. Borgstedt, Ed., Plenum Press, 1982,
p 243
17. W.D. Wilkinson, E.W. Hoyt, and H.V.
Rhude, Attack on Materials by Lead at
1000 C, USAEC Report ANL-5449,
Argonne National Laboratory, 1955
18. G.M. Tolson and A. Taboada, A Study of
Lead and Lead-Salt Corrosion in Thermal
Convection Loops, ORNL-TM-1437, Oak
Ridge National Laboratory, 1966
19. R.C. Asher, D. Davies, and S.A. Beetham,
Corros. Sci., Vol 17, 1977, p 545
20. O.F. Kammerer et al., Trans. AIME, Vol 212,
1958, p 20
21. M.S. Freed and K.J. Kelly, Corrosion
of Columbium Base and Other Structural
Alloys in High Temperature Lithium,
Report No. PWAC-355, Pratt and Whitney
AircraftCANEL, Division of United Aircraft Corp., June 1961 (declassied in June
1965)
22. E.E. Hoffman, Corrosion of Materials by
Lithium at Elevated Temperatures, USAEC
Report ORNL-2924, Oak Ridge National
Laboratory, Oak Ridge, TN, 1960
23. J.R. DiStefano, Corrosion of Refractory
Metals by Lithium, USAEC Report ORNL3551, Oak Ridge National Laboratory, Oak
Ridge, TN, 1964
24. O.K. Chopra, D.L. Smith, P.F. Tortorelli,

J.H. DeVan, and D.K. Sze, Fusion Technol.,
Vol 8, 1985, p 1956
25. R.E. Cleary, S.S. Blecherman, and J.E.
Corliss, Solubility of Refractory Metals
in Lithium and Potassium, USAEC Report
TIM-850, Nov 1965
26. D.A. Bates, G.R. Edwards, and D.L.
Olson, An Evaluation of Engineering
Alloys for High Temperature Lithium Containment, Mater. Perform., March 1980,
p 41
27. V. Coen, H. Kolbe, L. Orecchia, and
T. Sasaki, Materials Behavior and Physical Chemistry in Liquid Metal Systems,
H.U. Borgstedt, Ed., Plenum Press, 1982,
p 121
28. P.A. Steinmeyer, D.L. Olson, G.R. Edwards,
and D.K. Matlock, Rev. Coatings Corros.,
Vol 4 (No. 4), 1981, p 349
29. R.F. Koening and E.G. Brush, Mater.
Methods, Vol 42, 1955, p 112
30. A. Brasunas, Interim Report on StaticLiquid Metal Corrosion, USAEC Report
ORNL-1647, Oak Ridge National Laboratory, Oak Ridge, TN, 1954
31. R.F. Dudek and K.M. Ferguson, The
Corrosion Testing of Various Materials
in Sodium: Part I and II, USAEC Report BW-7020, Babcock & Wilcox, April
1957
32. V.W. Eldred, Interactions Between Solid
and Liquid Metals and Alloys, British
Report AERE-X/R-1806, Nov 1955
33. W. Markert, Jr., The Corrosion Testing
of Various Materials in Sodium, USAEC
Report BW-3792, Babcock & Wilcox, Aug
1954
34. W.C. Hayes and O.C. Shepard, Corrosion and Decarburization of the Ferritic
Chromium-Molybdenum Steels in Sodium
Coolant Systems, USAEC Report NAASR-2973, North American Aviation, Dec
1958
35. J.H. Stang, E.M. Simons, J.A. DeMastry,
and J.M. Genco, Compatibility of Liquid
and Vapor Alkali Metals with Construction
Materials, DMIC Report 227, Defense
Metals Information Center, Battelle Memorial Institute, April 1966
36. Nuclear Systems Materials Handbook,
Hanford Engineering Laboratory, Richland,
WA, 1976
37. H.U. Borgstedt, G. Frees, and H. Jesper,
Werkst. Korros., Vol 40, 1989, p 525
pg 434
Chapter 16:
38. E.J. Daniels, J. Inst. Met., Vol 46, 1931,

p 87
39. G. Tammann and W. Oelsen, Z. Anorg.
Chem., Vol 186, 1930, p 277
40. J.R. Weeks, Corrosion, April 1967, p 98
41. M.M. McDonalt, Corrosion of Brazed
Joints, Corrosion, Vol 13, 9th ed., Metals
Handbook, ASM International, 1987, p 876
42. C. Heiple, W. Bennett, and T. Rising,
Embrittlement of Several Stainless Steels by
Liquid Copper and Liquid Braze Alloys,
Mater. Sci. Eng., Vol 52, 1982, p 177
43. S.J. Matthews, M.O. Maddock, and
W.F. Savage, How Copper Surface Contamination Affects Weldability of Cobalt
Superalloys, Weld. J., May 1972
44. W.F. Savage and M. Mushala, Copper
Contamination Cracking in the Weld Heat
Affected Zone, Weld. J., May 1978, p 145
45. L.J. Korb, Corrosion of Manned Spacecraft,
Corrosion, Vol 13, 9th ed., Metals Handbook, ASM International, Metals Park,
Ohio, 1987, p 1059
31/10/2007 12:52PM Plate # 0
46. H.E. Ebert, Liquid Metal Embrittlement

of Flange Connector Studs in Contact with
Cadmium, Handbook of Case Histories
in Failure Analysis, Vol 1, R.C. Uhl, et al.,
47. G. Korbrin, Materials Selection, Corrosion,
48. C.P. Dillon, Unusual Corrosion Problems
in the Chemical Industry, MTI Publication
No. 54, Materials Technology Institute of
the Chemical Process Industries, Inc., 2000,
p 161
49. C.P. Dillon, Unusual Corrosion Problems
in the Chemical Industry, MTI Publication
No. 54, Materials Technology Institute of
the Chemical Process Industries, Inc., 2000,
p 204
50. J. Gutzeit, R.D. Merrick, and L.R. Scharfstein, Corrosion in Petroleum Rening
and Petrochemical Operations, Corrosion,
pg 435

DOI: 10.1361/hcma2007p437
31/10/2007 4:06PM Plate # 0
pg 437

CHAPTER 17
Hydrogen Attack
17.1 Introduction
Hydrogen attack can result in brittle fracture
of a steel component during high-temperature
service. It has also been referred to as hydrogen
damage. Carbon steels or low-alloy steels
can suffer severe hydrogen attack during hightemperature service, resulting in brittle failure
due to severe loss in tensile and rupture strengths
as well as ductility. Hydrogen attack should
not be confused with hydrogen embrittlement.
Hydrogen attack occurs only when steel is
in service at elevated temperatures, while
hydrogen embrittlement occurs near or at room
temperature.
Carbon and Cr-Mo steels derive strength primarily from carbon in the metal. In carbon steels,
carbon in the metal combines with iron to form
iron carbides, primarily cementite (Fe3C), which
along with ferrite lamellae, constitutes to pearlite
phases in the matrix of ferrite, as shown in
Fig. 17.1. For low-alloy steels containing chromium and molybdenum, these two alloying elements can form carbides, such as M3C, M2C, and
M23C6 where M can represent Fe, Cr, and Mo.
When steel is in contact with hydrogen

molecules (H2) at elevated temperatures, hydrogen atoms can be absorbed at the steel surface
and then diffuse into the metal. Hydrogen atoms
in the metal then react with iron carbide (Fe3C) to
form methane (CH4):
Fe3 C+4H ! 3Fe+CH4
(17:1)
As Reaction 17.1 continues with ongoing

ingress of hydrogen atoms into the steel, methane
gas is increasingly generated in the steel. With its
low diffusivity in steel, methane gas is accumulating at grain boundaries and other interfaces,
causing the methane gas pressure to build up in
the steel. This eventually results in development
of microcracks and microssures. At the same
time, carbon is being removed from iron carbide
(Fe3C) as a result of this chemical reaction,
leading to decarburization in the affected region.
The formation of microcracks and microssures
as well as decarburization causes the steel to lose
its tensile and rupture strengths as well as ductility, thus resulting in rupture of the steel component during service.
There are two sources of hydrogen that contribute to hydrogen attack of steels in industries.
One source of hydrogen is from the corrosion
of steel by boiler water in the waterwall tube
for coal-red boilers, while the other source
of hydrogen is from high-pressure, hightemperature hydrogen-containing atmospheres
used in petroleum rening. Hydrogen attack
problems related to these two separate areas are
discussed in the following sections.
17.2 Hydrogen Attack in Coal-Fired

Boilers
Fig. 17.1
Typical microstructure of carbon steel, consisting of

ferrite (white grains) and pearlite (dark regions). The
pearlite consists of cementite (Fe3C) and ferrite lamellae.
Hydrogen attack occurs more often in subcritical drum boilers that use a recirculating
steam-generating system. Adequate control of
31/10/2007 4:06PM Plate # 0
water chemistry is critical to avoid internal

deposition and corrosion of boiler tubes. The
water chemistry is typically controlled by (a)
water purication to remove impurities using
makeup water, polishing of returned condensate,
deaeration, and blowdown, and (b) chemical
treatments to control pH, electrochemical
potential, and dissolved oxygen concentration
(Ref 1). In the waterwall tubes, as steam forms,
dissolved solids concentrate in the boiler water;
once the solubility limit of an impurity is
exceeded, deposits can precipitate out (Ref 1).
Typical boiler deposits are hardness deposits of
calcium and magnesium salts and metal oxides
(Ref 1). These deposits formed on the surface of
the waterwall tube can provide a location for
accumulation of corrosive boiler water contaminants and/or chemical additives and their
reaction compounds (Ref 2). The deposits may
lead to low pH water chemistry in the local area,
resulting in acidic corrosion attack (Ref 3).
With adequate water chemistry control, the
waterwall steel tube forms a protective magnetite
(Fe3O4) scale when steel is corroded by water
under normal operating conditions:
3Fe+4H2 O ! Fe3 O4 +4H2
17.1, the affected area is decarburized. The

microstructure of the steel suffering hydrogen
attack is shown in Fig. 17.2. Hydrogen attack
typically is associated with heavy scales
(Fig. 17.2). Figure 17.3 shows microcracks and
microssures that formed along grain boundaries
for the waterwall steel tube (ASME SA192)
suffering hydrogen attack in a subcritical coalred boiler. Also shown in the gure is the
complete decarburization of the steel in the
affected area. Figure 17.4 shows the typical
pearlite-ferrite microstructure in the unaffected
waterwall tube in the same boiler as discussed in
Fig. 17.2 and 17.3. Large cracks developed in the
waterwall steel tube through the linking of
numerous microcracks and microssures. This is
illustrated in Fig. 17.5. Large cracks were often
found to show oxides formed on the surface of
these cracks (Fig. 17.5). The tube eventually
ruptured from the internal uid pressure. The
locations that are often susceptible to hydrogen
attack in the boiler are the burner zone and
the bull nose area, as illustrated in Fig. 17.6
(Ref 4). Monitoring and controlling boiler water
(17:2)
The by-product of the reaction is hydrogen.

The growth of the magnetite (Fe3O4) scale that
forms on the steel surface follows a parabolic
rate law with the scale growth rate diminishing
with time. Accordingly, the generation of
hydrogen likewise diminishes with time once
a steady state is reached. Atomic hydrogen
produced combines with other hydrogen atoms
to form molecular hydrogen that is then dispersed in the boiler water as a gas or in solution
(Ref 2).
However, when water chemistry is poorly
controlled, heavy internal tube deposits may
develop. This condition may lead to acidic corrosion attack on the tube under the deposits,
resulting in large amounts of hydrogen atoms
being absorbed by the steel. Hydrogen atoms
then react with iron carbide (Fe3C) in the steel to
form methane gas (CH4) according to Reaction
17.1. Methane gas then accumulates at grain
boundaries and other interfaces due to its low
diffusivity. Microcracks and microssures are
eventually developed by increasing local gas
pressure created by the increasing amount of
methane gas produced by the iron-carbide/
hydrogen reaction. Furthermore, because of
iron carbides being reduced to iron by Reaction
Fig. 17.2
Low-magnification optical micrograph showing

numerous microcracks and microfissures formed
in the carbon steel (ASME SA192) of the waterwall tube that suffered hydrogen attack in a subcritical coal-fired boiler. Note
numerous microcracks and microfissures. Also shown are heavy
corrosion scales formed on the tube internal surface. Unetched
pg 438
31/10/2007 4:06PM Plate # 0
Chapter 17:
chemistry is critical in preventing internal tube

deposits and hydrogen attack (Ref 3).
17.3 Hydrogen Attack in Petroleum

Refining
In petroleum rening, some renery equipment, such as reactors in hydrotreating, reforming, and hydrocracking units, is exposed to
a high-temperature, high-pressure hydrogen
atmosphere. Hydrogen attack is potentially a
very serious materials issue in the design and
operation of this type of equipment in hydrogen
service (Ref 5, 6). Steels are susceptible to
hydrogen attack as functions of temperature and
Fig. 17.3
Optical micrograph showing the formation of

microcracks and microfissures along grain
boundaries and decarburization of carbon steel (ASME SA192) in
a waterwall tube that suffered hydrogen attack in a subcritical
coal-fired boiler
Fig. 17.4
Optical micrograph showing typical microstructure

(pearlite and ferrite) in the unaffected carbon steel
waterwall tube in the same boiler as discussed in Fig. 17.2 and
17.3
Hydrogen Attack / 439
partial pressure of hydrogen in the system. For

hydrotreating, reforming, and hydrocracking
units, steel can suffer hydrogen attack at temperatures above approximately 260 C (500 F)
and hydrogen partial pressures above 0.689 MPa
(100 psig) (Ref 7).
The mechanism for hydrogen attack of steels
in petroleum rening is essentially the same as
that described in Section 17.2. Atomic hydrogen
is rst absorbed by the steel at the surface of the
renery equipment that is in contact with the
high-temperature, high-pressure hydrogen gas
stream. Hydrogen atoms then diffuse into the
steel and react with iron carbide (Fe3C) in the
steel to form methane gas (CH4) according to
Reaction 17.1. Methane gas then accumulates at
grain boundaries and other interfaces due to its
low diffusivity. Continuous ingress of hydrogen
atoms into the steel results in an increasing
amount of methane gas being generated by the
iron-carbide/hydrogen reaction, eventually causing microcracks and microssures in the steel.
Cracks then develop by linking microcracks and
microssures. Furthermore, because iron carbides are being reduced to iron by Reaction 17.1,
the affected area is decarburized. As a result,
the steel eventually ruptures from the internal
pressure of the reactor or vessel. The microstructure of the steel that suffers hydrogen attack
is characterized by numerous microcracks and
microssures (or cracks) and decarburized
microstructure, which is essentially the same as
that described in the waterwall tube that suffers
hydrogen attack in the boiler. Typical optical
microstructure of the hydrogen-attacked steel is
similar to that shown in Fig. 17.3.
Fig. 17.5
Optical micrograph showing a large crack that was

developed by linking microcracks and microfissures in the steel (ASME SA192) of the waterwall tube that suffered hydrogen attack. Oxides were often observed to form
around the surface of the crack.
pg 439
31/10/2007 4:06PM Plate # 0
Adding chromium and/or molybdenum to the

steel to increase the stability of iron carbides can
increase the resistance of the steel to hydrogen
attack. Some concerns have been raised for the
long-term performance of C-0.5Mo steel in hightemperature hydrogen environments (Ref 8).
Thus, Cr-Mo steels are much more resistant to
hydrogen attack than carbon and C-0.5Mo steels.
The conditions under which carbon and Cr-Mo
steels can be used in high-temperature hydrogen
service are described in detail in API 941 (Ref 9).
The behavior of carbon and Cr-Mo steels with
respect to their resistance to hydrogen attack
is summarized in Nelson curves in Fig. 17.7
(Ref 9).
17.4 Summary
During service at elevated temperatures,
carbon steel can react with atomic hydrogen
and result in brittle fracture. This phenomenon
is often referred to as hydrogen attack or
hydrogen damage. The chemical reaction

involves atomic hydrogen reacting with iron
carbide in the steel to form methane gas. Continuing ingress of atomic hydrogen into the metal
causes an increasing amount of methane gas to be
generated and accumulated at grain boundaries
and other interfaces, resulting in the formation of
microcracks and microssures in the steel. In
addition, the steel is decarburized. Continuing
growth of microcracks and microssures along
with decarburization of steel eventually result in
rupture of the steel component during service at
elevated temperatures.
The source of atomic hydrogen can be from
the rapid waterside corrosion at the internal
diameter of the waterwall tubes in a coal-red
boiler when water chemistry is not properly
controlled, thus resulting in hydrogen attack.
Hydrogen attack can also occur in some renery
equipment, such as reactors in hydrotreating,
reforming, and hydrocracking units, which is
exposed to a high-temperature, high-pressure
hydrogen atmosphere. Hydrogen attack in both
of these systems is discussed.
Steam
Risers
1
3
2
Downcomer
Heat flux
profile
1
1
Furnace
tubes
Steamwater
mixture
1
2
Burners
1
1
Heat flux
Steam-free subcooled water in
Fig. 17.6
Wilcox
Locations, marked as 1, in the boiler that are susceptible to hydrogen attack. The area marked as 2 shows other modes of
waterside corrosion that are outside of the current discussion topic. Source: Stultz and Kitto (Ref 4) Courtesy of Babcock and
pg 440
31/10/2007 4:06PM Plate # 0
Chapter 17:
Hydrogen partial pressure, MPa absolute

1500
3.45
6.90
10.34
13.79
17.24
pg 441
Hydrogen Attack / 441
34.5 62.1
20.7 48.3 75.8
800
1400
1300
700
1200
600
1000
6.0Cr-0.5Mo steel
1.25Cr-0.5Mo steel
500
3.0Cr-0.5Mo steel
900
2.25Cr-1.0Mo steel
1.0Cr-0.5Mo steel
800
2.0Cr-0.5Mo steel
700
600
1.25Cr-0.5Mo or 1.0Cr-0.5Mo steel
Temperature, C
Temperature, F
1100
400
300
500
Carbon Steel
400
300
Legend:
Surface decarburization
Internal decarburization
(Hydrogen attack)
Fig. 17.7
200
500
1000
1500
2000
Hydrogen partial pressure, 1b/psia
2500
3000 7000 11,000

5000 9000 13,000
Scale Change
Nelson curves showing the temperature and hydrogen partial pressure conditions under which carbon and Cr-Mo steels are
susceptible to hydrogen attack. Source: API 421 (Ref 9). Courtesy of American Petroleum Institute.
REFERENCES
1. P. Daniel, Corrosion of Steam- and WaterSide of Boilers, Corrosion: Environments and

Industries, Vol 13C, ASM Handbook, ASM
2. H.A. Grabowski, Chapter 17, Management of
Cycle Chemistry, ASME Handbook on Water
Technology for Thermal Power Systems,
P. Cohen, Ed., ASME, 1989, p 1379
3. R.B. Dooley, Corrosion of Steam/Waterside Boilers, Corrosion, Vol 13, 9th ed.,
Metals Handbook, ASM International, 1987,
p 990
Wilcox, 1992, p 42-22
5. G. Sorell and M.J. Humphries, High Temperature Hydrogen Damage in Petroleum
6.
7.
8.
9.
Renery Equipment, Mater. Perform., Vol 17

(No. 8), 1978, p 33
A.R. Ciuffreda and W.R. Rowland, Hydrogen
Attack of Steel in Reformer Service, Proc.
API, Vol 37 (No. III), 1957, p 116
R.D. Merrick and A.R. Ciuffreda, Hydrogen Attack of Carbon-0.5 Molybdenum
Steels, Proc. API, Vol 61 (No. III), 1982,
p 101
R. Chiba, K. Ohnishi, K. Ishii, and K. Maeda,
Effect of Heat Treatment on Hydrogen Attack
Resistance of C-0.5Mo Steels for Pressure
Vessels, Heat Exchangers, and Piping, Corrosion, Vol 41 (No. 7), 1985, p 415
Steels for Hydrogen Service at Elevated
Temperatures and Pressures in Petroleum
Reneries and Petrochemical Plants, Publication 941, 3rd ed., American Petroleum
Institute, 1983
Name ///sr-nova/Dclabs_wip/High Temp/5208_445-453.pdf/Appendix_2/
26/10/2007 1:55PM Plate # 0

DOI: 10.1361/hcma2007p445
pg 445

APPENDIX 2
Chemical Compositions of Alloys and

Filler Metals
Table 1 Chemical compositions (wt.%) of wrought ferritic steels
Grade*
Cr
Mo
Mn
Si
Others
T2
T11
T12
T17
T21
T22
T23
0.100.20
0.050.15
0.050.15
0.150.25
0.050.15
0.050.15
0.040.10
0.500.81
1.001.50
0.801.25
0.801.25
2.653.35
1.902.60
1.902.60
0.440.65
0.440.65
0.440.65
0.801.06
0.871.13
0.050.30
0.300.61
0.300.60
0.300.61
0.300.61
0.300.60
0.300.60
0.100.60
0.100.30
0.501.00
0.50(a)
0.150.35
0.50(a)
0.50(a)
0.50(a)
T5
T9
T91
T92
0.15(a)
0.15(a)
0.080.12
0.070.13
4.006.00
8.0010.00
8.009.50
8.509.50
0.450.65
0.901.10
0.851.05
0.300.60
0.300.60
0.300.60
0.300.60
0.300.60
0.50(a)
0.251.00
0.200.50
0.50(a)
T122
0.070.14
10.0012.50
0.250.60
0.70(a)
0.50(a)
V: 0.15 min
W: 1.451.75, V: 0.200.30, Nb: 0.020.08,

B: 0.00050.006
V: 0.180.25, Nb: 0.060.10, N: 0.030.07

V: 0.150.25, W: 1.52.0, Nb: 0.040.09,
B: 0.0010.006, N: 0.030.07
V: 0.150.30, W: 1.502.50, Cu: 0.301.70,
Nb: 0.040.10, B: 0.00050.005,
N: 0.040.10
(a) Maximum, Nb = Cb. * For tubes
Table 2 Chemical compositions (wt.%) of wrought stainless steels

Alloy
UNS No.
Cr
Ni
Fe
403
410
414
416
416 (Se)
420
431
440A
440B
440C
GREEK ASCOLOY
154CM
405
430
430F
430F (Se)
446
409
439
444
S40300
S41000
S41400
S41600
S41623
S42000
S43100
S44002
S44003
S44004
S41880
S40500
S43000
S43020
S43023
S44600
S40900
S43035
S44400
0.15(a)
0.15(a)
0.15(a)
0.15(a)
0.15(a)
0.15(a)
0.20(a)
0.600.75
0.750.95
0.951.20
0.12
1.05
0.08(a)
0.12(a)
0.12(a)
0.12(a)
0.20(a)
0.08(a)
0.07(a)
0.025(a)
11.513.0
11.513.5
11.513.5
12.014.0
12.014.0
12.014.0
15.017.0
16.018.0
16.018.0
16.018.0
12.6
14.0
11.514.5
14.018.0
14.018.0
14.018.0
23.027.0
10.511.75
17.019.0
17.519.5
1.252.50
1.252.50
2.0
0.5(a)
0.5(a)
1.0(a)
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
E-BRITE
MONIT
S44627
S44635
0.01(a)
0.025(a)
25.027.5
24.526.0
0.5(a)
3.54.5
Bal
Bal
SEA-CURE
S44660
0.03(a)
25.028.0
1.03.5
Bal
AL 29-4C
S44735
0.03(a)
28.030.0
1.0
Bal
AL 29-4-2
18SR
12SR
S44800
0.01(a)
0.03
0.02
28.030.0
18.0
12.0
2.02.5
(continued)
Bal
Bal
Bal
(a) Maximum. Cb = Nb
Others
S: 0.15 min
Se: 0.15 min
Mo: 0.75(a)
Mo: 0.75(a)
Mo: 0.75(a)
W: 3.0
Mo: 4.0
At: 0.100.30
S: 0.15 min
Se: 0.15 min
N: 0.25(a)
Ti: (6 C) min, 0.75(a)
Ti: 0.2 + 4(C + N) min, 1.10(a)
Mo: 1.752.50, N: 0.035(a), Ti + Cb:
0.2 + 4(C + N) min, 0.8(a)
Mo: 0.751.50, N: 0.015(a), Cb: 0.050.20
Mo: 3.44.5, N: 0.035(a), Ti + Cb:
0.2 + 4(C + N) min, 0.80(a)
Mo: 3.03.5, N: 0.04(a), Ti + Cb:
0.21.0, 6(C + N) min
Mo: 3.64.2, N: 0.045(a), Ti + Cb:
0.21.0, 6(C + N) min
Mo: 3.54.2, N: 0.02(a), C + N: 0.025(a)
Al: 1.8, Ti: 0.4
Al: 1.2, Cb: 0.6, Ti: 0.3
26/10/2007 1:55PM Plate # 0
Table 2
(Continued)
Alloy
UNS No.
Cr
Ni
Fe
Others
ALFA-I
ALFA-II
329
URANUS 50
CD-4MCu
44LN
DP-3
3RE60
2205
FERRALIUM 255
7-Mo: PLUS
SUPEER-FERRIT
201
202
301
302
302B
303
303(Se)
304
304L
304H
305
308
309
309S
310
310S
314
316
316L
316H
317
321
321H
347
347H
348
253MA
RA85H
17-4PH
17-7PH
PH15-7Mo
A286
AM 350
AM 355
16-18
17-14CuMo
20-29CuMo
17-10P
HMN
TENELON
254SMO
19-9 DL
904L
16-25-6
15-5PH
CUSTOM 450
CUSTOM 455
AL-6XN
MVMA
22-4-9
NITRONIC 60
NITRONIC 50
NITRONIC 40
(21-6-9)
SNR-4
317LM
17-14-4LM
JS700
CRUTEMP: 25
S32900
S32404
J93370
S31200
S31260
S31500
S31803
S32550
S32950
S20100
S20200
S30100
S30200
S30215
S30300
S30400
S30403
S30409
S30500
S30800
S30900
S30908
S31000
S31008
S31400
S31600
S31603
S31609
S31700
S32100
S32109
S34700
S34709
S34800
S30815
S30615
S17400
S17700
S15700
K66286
S35000
S35000
S21400
S31254
K63198
N08904
S15500
S45000
S45500
N08367
S30415
S21800
S20910
S21900
0.025
0.025
0.08(a)
0.04(a)
0.04(a)
0.03(a)
0.03(a)
0.03(a)
0.03(a)
0.04(a)
0.03(a)
0.15(a)
0.15(a)
0.15(a)
0.15(a)
0.15(a)
0.15(a)
0.15(a)
0.08(a)
0.03(a)
0.040.10
0.12(a)
0.08(a)
0.2(a)
0.08(a)
0.25(a)
0.08(a)
0.25(a)
0.08(a)
0.03(a)
0.040.10
0.08(a)
0.08(a)
0.040.10
0.08(a)
0.040.10
0.08(a)
0.08
0.2
0.04
0.07
0.07
0.08
0.1
0.13
0.05
0.12
0.05
0.12
0.30
0.08
0.02
0.3
0.02
0.06
0.07
0.05(a)
0.03
0.03
0.05
0.5
0.05
0.03
0.05
13.0
13.0
23.028.0
20.522.5
24.526.5
24.026.0
24.026.0
18.019.0
21.023.0
24.027.0
26.029.0
28.0
16.018.0
17.019.0
16.018.0
17.019.0
17.019.0
17.019.0
17.019.0
18.020.0
18.020.0
18.020.0
17.019.0
19.021.0
22.024.0
22.024.0
24.026.0
24.026.0
23.026.0
16.018.0
16.018.0
16.018.0
18.020.0
17.019.0
17.019.0
17.019.0
17.019.0
17.019.0
21.0
18.5
16.5
17.0
15.0
13.516.0
16.5
15.5
16.0
16.0
20.0
17.0
18.5
17.0
19.520.5
19.0
19.023.0
16.0
15.0
15.5
11.75
20.022.0
18.5
21.5
17.0
22.0
21.0
2.55.0
5.88.5
4.756.0
5.56.5
5.57.5
4.255.25
4.56.5
4.56.5
3.55.2
3.2
3.55.5
4.06.0
6.08.0
8.010.0
8.010.0
8.010.0
8.010.0
8.010.5
8.012.0
8.010.5
10.013.0
10.012.0
12.015.0
12.015.0
19.022.0
19.022.0
19.022.0
10.014.0
10.014.0
10.014.0
11.015.0
9.012.0
9.012.0
9.013.0
9.013.0
9.013.0
11.0
14.5
4.25
7.0
7.0
24.027.0
4.25
4.25
19.0
14.0
29.0
10.5
9.5
17.518.5
9.0
23.028.0
25.0
4.5
6.0
8.5
23.525.5
9.5
4.0
8.5
12.5
6.0
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Al: 3.0, Ti: 0.4

Al: 4.0, Ti: 0.4
Mo: 1.02.0
Mo: 2.03.0, Cu: 1.5, N: 0.2(a)
Mo: 1.752.25, Cu: 3.0
Mo: 1.22.0, N: 0.140.20
Mo: 2.53.5, Cu: 0.20.8, N: 0.10.3, W: 0.10.4
Mo: 2.53.0, N: 0.080.15, Si: 1.42.0
Mo: 2.53.5, N: 0.080.2
Mo: 2.04.0, Cu: 1.42.5, N: 0.10.25
Mo: 1.02.5, N: 0.150.35
Mo: 2.1
Mn: 5.57.5, N: 0.25(a)
Mn: 7.510.0, N: 0.25(a)
Si: 2.03.0
S: 0.15 min
Se: 0.15 min
Si: 1.53.0
Mo: 2.03.0
Mo: 2.03.0
Mo: 2.03.0
Mo: 3.04.0
Ti: 5 C min
Ti: 5 C min
Cb + Ta: 10 C min
Cb + Ta: 10 C min
Cb + Ta: 10 C min, Ta: 0.1(a)
Si: 1.7, N: 0.17, Ce: 0.04
Si: 3.5, Al: 1.0
Cb: 0.25, Cu: 3.6
Al: 1.15
Mo: 2.25, A1: 1.15
Mo: 1.01.5, Ti: 1.92.35, V: 0.10.5
Mo: 2.75, N: 0.1
Mo: 2.75, N: 0.12
Mo: 2.5, Cb: 0.4, Ti: 0.3, Cu: 3.0

Mo: 2.20, Cu: 3.20
P: 0.28
Mn: 3.5, P: 0.25
Mn: 14.5, N: 0.4
Mo: 6.06.5, Cu: 0.51.0, N: 0.180.22
Mo: 1.25, W: 1.25, Cb: 0.4
Mo: 4.05.0, Cu: 1.02.0, V: 0.10.5
Mo: 6.0
Cb: 0.3, Cu: 3.5
Mo: 0.75, Cb: 8 C min, Cu: 1.5
Cb: 0.3, Ti: 1.2, Cu: 2.25
Mo: 6.07.0, N: 0.180.25
Si: 1.3, N: 0.15, Ce: 0.04
N: 0.4, S: 0.1
Si: 4.0, Mn: 8.0, N: 0.13
Mo: 2.0, Mn: 5.0, N: 0.3, Cb: 0.2, V: 0.2
Mn: 9.0
S31753
S31725
S31726
N08700
0.03
0.03
0.03
0.03
0.05
18.5
18.0
17.0
21.0
25.0
13.5
15.0
13.0
25.0
25.0
Bal
Bal
Bal
Bal
Bal
Mo: 3.6
Mo: 4.1
Mo: 4.2, N: 0.15
Mo: 4.5, Mn: 1.7
pg 446
26/10/2007 1:55PM Plate # 0
pg 447
Appendix 2: Chemical Compositions of Alloys and Filler Metals / 447
Table 3 Chemical compositions (wt.%) of cast corrosion- and heat-resistant alloys

Alloy
UNS No.
Cr
Ni
Fe
Others
CA-15
CA-40
CB-30
CC-50
CE-30
CF-3
CF-3M
CF-8
CF-20
CF-8M
CF-12M
CF-8C
CF-16F
CG-8M
CH-20
CK-20
CN-7M
HA
HB
HC
HD
HE
HF
HH
HI
HK
HL
HN
HT
HU
HW
HX
HP
THERMAX 40B
THERMAX 47
THERMAX 63
THERMAX 63WC
THERMALLOY 63W
THERMALLOY 63WC
(SUPERTHERM)
MANAURITE XU
MANAURITE 36X
MANAURITE 36XS
MANAURITE 900
MANAURITE XT
MO-RE 1
HOM-3
22-H (NA 22H or HOM-5)
SUPER 22-H
IN-657
J92605
J93005
J93403
J92603
J93503
J94003
J94224
J94604
J94213
J94605
J95405
N08001
N06006
J95705
0.15(a)
0.40(a)
0.30(a)
0.50(a)
0.30(a)
0.03(a)
0.03(a)
0.08(a)
0.20(a)
0.08(a)
0.12(a)
0.08(a)
0.16(a)
0.08(a)
0.20(a)
0.20(a)
0.07
0.2(a)
0.3(a)
0.5(a)
0.5(a)
0.20.5
0.20.4
0.20.5
0.20.5
0.20.6
0.20.6
0.20.5
0.350.75
0.350.75
0.350.75
0.350.75
0.4
0.4
0.45
0.45
0.45
0.45
0.5
11.514.0
11.514.0
1822
2630
2630
1821
1821
1821
1821
1821
1821
1821
1821
1821
2226
2327
1922
810
1822
2630
2630
2630
1923
2428
2630
2428
2832
1923
1317
1721
1014
1519
25
25
25
23
25
25
28
1.0(a)
1.0(a)
2.0(a)
4.0(a)
811
811
912
811
811
912
912
912
912
811
1215
1922
2831
2(a)
4(a)
47
811
912
1114
1418
1822
1822
2327
3337
3741
5862
6468
35
13
20
35
35
36
36
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Mo: 0.5(a)
Mo: 0.5(a)
Mo: 2.03.0
Mo: 2.03.0
Mo: 2.03.0
Cb: 8 C min, 1.0(a)
Mo: 1.5(a), Se: 0.20.35
Mo: 3.0 min
Mo: 2.03.0, Cu: 3.04.0
W: 0.5
W: 0.5
W: 0.5
Co: 15.0, W: 5.0
W: 5.0
Co: 15.0, W: 5.0
0.350.45
0.4
0.4
0.13
0.4
0.45
0.45
0.5
0.5
0.06
2427
25
25
21
35
26
26
28
28
48
3236
34
34
33
44
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
33
16
16
13
Nb: 0.5, W: 0.5

Cb: 1.2
Cb: 1.5, W: 1.5
Cb: 1.2
Cb: 1.5, W: 1.5
W: 1.6
Mo: 3.0, W: 3.0, Co: 3.0
W: 5.0
Co: 3.0, W: 5.0
Cb: 1.5
26/10/2007 1:55PM Plate # 0
Table 4
Chemical compositions (wt.%) of wrought iron-, nickel-, and cobalt-base alloys
Alloy
UNS No.
INCOLOY 800
INCOLOY 800H
INCOLOY 800HT
INCOLOY 802
INCOLOY 903
INCOLOY 904
INCOLOY 907
INCOLOY 909
INCOLOY DS
RA 330
RA 330HC
AC66
SANICRO 28
20Cb-3
20Mo-4
20Mo-6
Nicrofer 3033 (alloy 33)
HAYNES HR-120
HAYNES 556
MULTIMENT alloy
(N-155)
V-57
W-545
DISCALOY
PYROMET CTX-1
CHROMEL D
KANTHAL Al
KANTHAL AF
FECRALLOY A
Aluchrom S
Aluchrom ISE
Aluchrom Y
Aluchrom O
Aluchrom PSI
Ni 200
Ni 201
Ni 270
MONEL 400
MONEL 401
MONEL R-405
MONEL K-500
MONEL 450
FERRY alloy
CUPRO 107
INCONEL 600
INCONEL 601
Nicrofer 6025HT
(602CA)
INCONEL 617
INCONEL 625
INCONEL 690
INCONEL 693
INCOLOY 825
INCOLOY 890
INCOLOY 925
INCONEL 706
INCONEL 718
INCONEL X-750
INCONEL 751
INCONEL 671
INCONEL 686
Nicrofer 5923 (59)
HAYNES 214
HAYNES 230
HAYNES 242
HAYNES HR-160
Nicrofer 45TM
HASTELLOY X
HASTELLOY W
HASTELLOY S
HASTELLOY N
HASTELLOY C-22
N08800
N08810
N08811
N08802
N19903
N19907
N19909
N08330
N33228
N08028
N08020
N08024
N08026
R20033
R30556
R30155
Cr
Ni
0.06
0.05
0.4
0.05
0.01
0.02
0.02
0.02
0.015(a)
0.05
0.1
0.1
21
21
21
21
17
19
19
28
27
20
23.5
24
33
25
22
21
32.5
32.5
32.5
32.5
38
32.5
38
38
35
35
35
32
31
33
37
36
32
37
20
20
K66545
K66220
K92500
N02200
N02201
N02270
N04400
N04401
N04405
N05500
C71500
N06600
N06601
N06025
0.8(a)
0.08
0.06
0.03
0.03
0.08(a)
0.10(a)
0.010.1
0.08(a)
0.0150.03
0.08
0.02(a)
0.01
0.08(a)
0.10(a)
0.2
14.8
13.5
14.0
18.5
22.0
22.0
15.8
20
20
21
22
22.5
15.5
23.0
25
N06617
N06625
N06690
N06693
N08825
N08890
N09925
N09706
N07718
N07750
N07751
N06686
N06059
N06230
N12160
N06045
N06002
N10004
N06635
N10003
N06022
0.07
0.10(a)
0.02
0.2
0.03
0.1
0.01
0.03
0.04
0.04
0.05
0.05
0.01(a)
0.01(a)
0.04
0.1
0.03(a)
0.05
0.1
0.1
0.12(a)
0.02(a)
0.06
0.01(a)
22.0 Bal 12.5

1.5
21.5 Bal
2.5
29.0 Bal
9.0
29.0 Bal
4.0
21.5 Bal
30.0
25
42.5
Bal
21.0 Bal
28.0
16.0 Bal
37.0
18.0 Bal
18.5
15.5 Bal
7.0
15.0 Bal
7.0
48.0 Bal
21.0 Bal
5.0(a)
23
Bal
1.5(a)
16.0 Bal
3.0
22.0 Bal
3.0(a)
8.0 Bal
28.0 Bal 29.0

1.5
27
Bal
23
22.0 Bal
1.5
18.5
5.0 Bal
2.5
6.0
15.5 Bal
3.0(a)
7.0 Bal
5.0(a)
22.0 Bal
2.5(a) 3.0
(continued)
0.05
0.08
0.08
0.4
Co
Fe
Mo
15
14.5
13
13
18
20
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
3.5
2.2
3.8
5.6
1.25
3.0
3.0
2.5
2.5
Al: 0.3, Ti: 0.3

Al: 0.4, Ti: 0.4
Al + Ti: 1.0
Ti: 1.4, Al: 0.9, Cb: 3.0

Ti: 1.6
Ti: 1.5, Cb: 4.7, Si: 0.15
Ti: 1.5, Cb: 4.7, Si: 0.4
Si: 2.3
Si: 1.2
Si: 1.2
Nb: 0.8, Ce: 0.07
Cu: 1.0
Cu: 3.3, Cb: 0.5
Cu: 1.0, Cb: 0.25
Cu: 3.0
Cu: 0.75, N: 0.5
Cb: 0.7, N: 0.2
Ta: 0.6, La: 0.02, N: 0.2, Zr: 0.02
Cb + Ta: 1.0, N: 0.15
27.0
26.0
26.0
37.7 16.0
36.0
99.6
99.6
99.98
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
1.2
0.3
1.2
1.0
0.7
0.8
8.0
14.4
10
1.25
1.5
3.0
Al: 0.25, Ti: 3.0, V: 0.5(a), B: 0.01

Al: 0.2, Ti: 2.85, B: 0.05
Al: 0.25, Ti: 1.7
Cb: 3.0, Al: 1.0, Ti: 1.7
Si: 1.5
Al: 5.8
Al: 5.3, Y
Al: 4.8, Y: 0.3
Al: 4.5, Zr: 0.3(a)
Al: 5.0
Al: 5.5, Zr: 0.010.1, Y: 0.050.15
Al: 5.5, Zr: 0.3(a)
Al: 5.6, Hf: 0.3
Cu: 31.5, Mn: 1.1

Cu: 55.5, Mn: 1.63
Cu: 31.5, Mn: 1.1
Cu: 29.5, Ti: 0.6, Al: 2.7
Cu: 68.0, Mn: 0.7
Cu: 55.0
Cu: 68.0, Mn: 1.1
Al: 1.4
Al: 2.1, Ti: 0.15, Y: 0.050.12,
Zr: 0.010.1
A1: 1.2
Cb: 3.6
Al: 2.54.0, Nb: 0.52.5

Cu: 2.2
Ta: 0.2
Cu: 1.8, Ti: 2.1, Al: 0.3
Ti: 1.8, Al: 0.2, Cb: 2.9
Cb: 5.1
Ti: 2.5, Al: 0.7, Cb: 1.0
Ti: 2.5, Al: 1.1, Cb: 1.0
Ti: 0.35
Ti: 0.020.25
Al: 4.5, Y
La: 0.02, B: 0.015(a)
Si: 2.75, Ti: 0.5, Nb: 1.0(a)

Si: 2.53.0, Ce: 0.030.09
La: 0.05, B: 0.015(a)
9.0
9.0
3.0
1.5
3.0
3.0
16.0
15.5
2.0
25.0
9.0
24.0
14.5
16.5
13.0
4.0
14.0
0.6
3.0
Others
pg 448
26/10/2007 1:55PM Plate # 0
pg 449
Table 4 (Continued)
Alloy
UNS No.
Cr
Ni
Co
Fe
Mo
HASTELLOY C-276
HASTELLOY C-2000
HASTELLOY C-4
HASTELLOY C
HASTELLOY B
HASTELLOY B-2
HASTELLOY B-3
NICROFER 6224 (B-10)
HASTELLOY G
HASTELLOY G-3
HASTELLOY G-30
HASTELLOY G-35
HASTELLOY G-50
HASTELLOY H-9M
RA 333
CHROMEL A (or
NICHROME 80)
NA 224
NIMONIC 70
NIMONIC 75
NIMONIC 80A
NIMONIC 81
NIMONIC 86
NIMONIC 90
NIMONIC 91
NIMONIC 105
NIMONIC 115
NIMONIC 901
NIMONIC AP 1
NIMONIC PE 11
NIMONIC PE 16
NIMONIC PK 31
NIMONIC PK 33
NIMONIC PK 50
NIMONIC PK 37
WASPALOY alloy
263
HAYNES 282
REN 41
REN 95
REN 100
UDIMET 400
UDIMET 500
UDIMET 520
UDIMET 630
UDIMET 700
UDIMET 710
UDIMET 720
UNITEMP AF2-IDA
UNITEMP AF2-ID6
ASTROLOY
D-979
IN 100
IN 102
IN 587
IN 597
M 252
PYROMET 31
PYROMET 860
REFRACTORY 26
625 PLUS
IN 100 GATORIZE
HAYNES 188
HAYNES 25 (L-605)
HAYNES 150 (UMCo-50)
HAYNES 6B
S-816
MAR-M 918
MP 35N
MP 159
AR 213
ULTIMET alloy
N10276
N06200
N06455
N10002
N10001
N10665
N10675
N06007
N06985
N06030
N06035
N06950
N06333
0.01(a)
0.01(a)
0.01(a)
0.08(a)
0.05(a)
0.01(a)
0.01(a)
0.01(a)
0.05(a)
0.015(a)
0.03(a)
0.05(a)
0.02(a)
0.03(a)
0.05
15.5
23
16.0
15.5
2
8
22.0
22.0
29.5
33
20
22.0
25.0
20.0
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
2.5(a)
2.0(a)
2.5(a)
2.5(a)
5.0(a)
5.0(a)
2.5(a)
5.0(a)
3.0
5.5
3.0(a)
3.0(a)
6.0
5.0
2.0(a)
2
7
19.5
19.5
15.0
2.0(a)
17
19.0
18.0
16.0
16
15.5
17.0
28.0
28.0
2732
23
6.5
7.0
5.0
8
9
9.0
3.0
4.0
4.0
1.5(a)
2.5
2.0
3.0
N07080
N07090
N07001
N07041
N09979
N13100
N06102
N07252
N07031
N07716
R30188
R30605
R30816
R30035
R31233
0.5
27.0
20.0
19.5
19.5
30.0
25.0
19.5
28.5
15.0
15.0
12.5
15.0
18.0
16.5
20.0
18.0
19.5
19.5
19.0
20.0
19.5
19.0
14.0
9.5
17.5
18.0
19.0
18.0
15.0
18.0
17.9
12.0
12.0
15.0
15.0
10.0
15.0
28.5
24.5
20.0
22.5
13.0
18.0
20.0
12.4
22.0
20.0
27.0
30.0
20.0
20.0
20.0
19.0
19.0
26.0
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
22.0
10.0
20.0
20.0
35.0
25.5
9.0
18.5
25.0
36.0
34.0
34.0
1.5(a)
5.0(a)
18.0
27.0
7.0
15.0
28.9
16.0
5.0
3.0(a)
3.0(a)
18.0
4.0
9.0
3.0
6.0
10.0
5.0
4.0
5.8
5.0
5.2
3.3
4.5
7.0
4.25
4.3
5.8
8.5
10.0
3.5
3.5
3.0
4.0
4.0
6.0
1.0
3.0
3.0
5.2
3.0
1.5
3.0
1.3
3.0
6.0
2.7
6.5
5.3
4.0
4.0
3.0
3.0
3.0
1.5
10.0
2.0
6.0
3.2
9.0
3.2
14.0
15.0
1.5(a) 4.5
4.0
4.0
10.0
7.0
4.5
5.0
2.0
0.10
0.06
0.03
0.07
0.08
0.15
0.05
0.05
0.04
0.08
0.06
0.06
0.09
0.15
0.16
0.06
0.08
0.05
0.03
0.03
0.07
0.03
0.35
0.04
0.06
0.05
0.15
0.06
0.05
0.05
0.15
0.04
0.05
0.03
0.02
0.07
0.10
0.10
0.06
1.20
0.38
0.05
0.17
0.06
16.5
20.0
20.0
15.0
17.0
14.0
14.0
13.5
16.5
14.0
20.0
10
11.0
8.0
15.0
14.0
18.5
12.0
18.5
15.0
14.7
10.0
10.0
17.0
15.0
20.0
20.0
10.0
4.0
20.0
18.5
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Others
Cu: 1.6
V: 0.03
Cb + Ta: 2.0, Cu: 2.0

Cb + Ta: 0.3, Cu: 2.0
Cu: 2.0
Si: 1.0
Al: 1.0, Ti: 1.25, Cb: 1.5
Al: 1.4, Ti: 2.4

Al: 0.9, Ti: 1.8
Ce: 0.03
Al: 1.5, Ti: 2.5
Al: 1.2, Ti: 2.3
Al: 4.7, Ti: 1.3, B: 0.005
Al: 5.0, Ti: 4.0
Ti: 2.9
Al: 4.0, Ti: 3.5
Al: 0.8, Ti: 2.3
Al: 1.2, Ti: 1.2
Al: 0.4, Ti: 2.35, Cb: 5.0
Al: 2.1, Ti: 2.4
Al: 1.4, Ti: 3.0
Al: 1.5, Ti: 2.5
Al: 1.5, Ti: 3.0, Zr: 0.05, B: 0.006
Al: 0.5, Ti: 2.2
Al: 1.5, Ti: 2.1, B: 0.005
Al: 1.5, Ti: 3.0, B: 0.006
Cb: 3.5, Al: 3.5, Ti: 2.5, Zr: 0.05
Al: 5.5, Ti: 4.2, Zr: 0.06, B: 0.015
Cb: 0.5, Al: 1.5, Ti: 2.5, Zr: 0.06, B: 0.008
Al: 2.9, Ti: 2.9, Zr: 0.05, B: 0.006
Al: 2.0, Ti: 3.0, B: 0.005
Cb: 6.5, Al: 0.5, Ti: 1.0
Al: 5.3, Ti: 3.5, B: 0.03
Al: 2.5, Ti: 5.0
Al: 2.5, Ti: 5.0, Zr: 0.03, B: 0.033
Ta: 1.5, Al: 4.6, Ti: 3.5, Zr: 0.1
Ta: 1.5, Al: 4.0, Ti: 2.8, Zr: 0.1, B: 0.015
Al: 4.0, Ti: 3.5, B: 0.03
Al: 1.0, Ti: 3.0
Al: 5.5, Ti: 4.7, Zr: 0.06, V: 1.0, B: 0.015
Cb: 3.0, Al: 0.4, Ti: 0.6, Mg: 0.02, Zr: 0.03
Cb: 0.7, Al: 1.2, Ti: 2.3, Zr: 0.5
Cb: 1.0, Al: 1.5, Ti: 3.0, Zr: 0.5
Al: 1.0, Ti: 2.6, B: 0.005
Al: 1.4, Ti: 2.3, Cu: 0.9, B: 0.005
Al: 1.0, Ti: 3.0, B: 0.01
Al: 0.2, Ti: 2.6, B: 0.015
Cb: 3.1, Al: 0.2, Ti: 1.3
Al: 5.0, Ti: 4.3, Zr: 0.06, B: 0.02, V: 0.8
La: 0.04
Cb: 4.0
Ta: 7.5, Zr: 0.10
Cb: 0.6, Al: 0.2, Ti: 3.0

Al: 3.5, Ta: 6.5, Zr: 0.15, Y: 0.1
N: 0.08
26/10/2007 1:55PM Plate # 0
Table 5
Chemical compositions (wt.%) of cast nickel- and cobalt-base alloys
Alloy
Cr
Ni
B-1900
IN 100
IN 162
IN 731
IN 738
IN 792
IN 713C
IN 713LC
M 21
M 22
M 252
MAR-M 200
MAR-M 246
MAR-M 247
MAR-M 421
MAR-M 432
MC 102
REN 77
REN 80
TAZ-8A
TAZ-8B: (DS)
TRW-NASA-VIA
0.10
0.18
0.12
0.18
0.17
0.12
0.12
0.05
0.13
0.13
0.15
0.15
0.15
0.15
0.15
0.15
0.04
0.07
0.17
0.12
0.12
0.13
8.0
10.0
10.0
9.5
16.0
12.4
12.5
12.0
5.7
5.7
20.0
9.0
9.0
8.25
15.8
15.5
20.0
15.0
14.0
6.0
6.0
6.0
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
WAZ-20 (DS)
DELORO 40
DELORO 50
DELORO 60
AIRESIST 13
AIRESIST 213
AIRESIST 215
FSX 414
J-1650
MAR-M 302
MAR-M 322
MAR-M 509
MAR-M 918
NASA Co-W-Re
S-816
V-36
WI-52
X-40 (HAYNES 31)
Stellite 1
Stellite 6
Stellite 12
Stellite 21
Stellite 6K
Stellite 704
Stellite 706
Stellite 712
Stellite 706K
Tribaloy T-400C
Tribaloy T-401
0.20
0.2
0.45
0.7
0.45
0.20
0.35
0.25
0.20
0.85
1.00
0.60
0.05
0.40
0.40
0.27
0.45
0.50
2.45
1.2
1.6
0.25
1.6
1
1.2
1.6
1.6
0.1(a)
0.2
7.5
10.5
14.5
21.0
20.0
19.0
29.0
19.0
21.5
21.5
23.5
20.0
3.0
20.0
25.0
21.0
22.0
31
28
29.5
27
31
30
29
29
31
14
17
Bal
Bal
Bal
Bal
0.5
0.5
10.0
27.0
10.0
20.0
20.0
20.0
10.0
3(a)
3(a)
3(a)
2.5
3.0(a)
2(a)
3(a)
3(a)
3(a)
1(a)
0.8(a)
Co
Fe
Mo
Ta
Zr
Others
10.0
15.0
10.0
8.5
9.0
10.0
10.0
10.0
10.0
9.5
20.0
15.0
9.5
5.0
7.5
1.0
0.5
6.0
3.0
4.0
2.5
1.7
1.9
4.2
4.5
2.0
2.0
10.0
2.5
0.7
2.0
6.0
4.2
4.0
4.0
4.0
2.0
2.0
2.6
3.8
11.0
11.0
12.5
10.0
10.0
3.8
3.0
2.5
4.0
4.0
4.0
6.0
4.0
2.0
1.7
3.9
3.0
1.5
3.0
2.0
0.6
8.0
8.0
9.0
0.10
0.06
0.10
0.06
0.10
0.10
0.10
0.10
0.12
0.60
0.05
0.05
0.05
0.05
0.05
0.04
0.03
1.0
1.0
0.13
2.5
3.5
4.0
0.5
0.5
1.0
0.5
0.5
4.0
3.0
2.0
1.5
2.5(a)
3(a)
2.5(a)
3(a)
3(a)
2(a)
3(a)
3(a)
3(a)
1(a)
0.8(a)
4.0
4.0
5.5
14
4.5
8.5
4.5
27
22
20.0
11.0
4.5
4.5
7.5
12.0
10.0
9.0
7.0
25.0
4.0
2.0
11.0
7.5
13
4.5
8.5
4.5
2.0
6.5
7.5
2.0
9.0
4.5
3.5
7.5
1.5
0.1
0.1
0.2
2.0
0.5
0.1
1.0
Al: 6.0, Ti: 1.0, B: 0.015

Al: 5.5, Ti: 5.0, B: 0.01, V: 1.0
Al: 6.5, Ti: 1.0, Cb: 1.0, B: 0.02
Al: 5.5, Ti: 4.6, B: 0.015, V: 1.0
Al: 3.4, Ti: 3.4, Cb: 0.9, B: 0.01
Al: 3.1, Ti: 4.5, B: 0.02
Al: 6.0, Ti: 0.8, Cb: 2.0, B: 0.012
Al: 6.0, Ti: 0.6, Cb: 2.0, B: 0.01
Al: 6.0, Cb: 1.5, B: 0.02
Al: 6.3
Al: 1.0, Ti: 2.6, B: 0.005
Al: 5.0, Ti: 2.0, B: 0.015, Cb: 1.0
Al: 5.5, Ti: 1.5, B: 0.015
Al: 5.5, Ti: 1.0, Hf: 1.5, B: 0.015
Al: 4.3, Ti: 1.8, Cb: 2.0, B: 0.015
Al: 2.8, Ti: 4.3, Cb: 2.0, B: 0.015
Cb: 6.0
Al: 4.3, Ti: 3.3, B: 0.015
Al: 3.0, Ti: 5.0, B: 0.015
Al: 6.0, Cb: 2.5, B: 0.004
Al: 6.0, Cb: 1.5, B: 0.004
Al: 5.5, Ti: 1.0, Cb: 0.5, Hf: 0.4,
Re: 0.5, B: 0.02
Al: 6.5
Si: 3.5
Si: 4.0, B: 2.0
Si: 2.04.5, B: 3.3
Al: 3.4, Y: 0.1
Al: 3.5, Y: 0.1
Al: 4.3, Y: 0.1
B: 0.01
Ti: 3.8, B: 0.02
B: 0.005
Ti: 0.75
Ti: 0.2
Re: 2.0, Ti: 1.0

Cb: 4.0
Cb: 2.0
Cb + Ta: 2.0
Si: 2.6
Si: 1.3
1.5(a)
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Table 6
Chemical compositions (wt.%) of oxide-dispersion-strengthened (ODS) alloys
Alloy
INCOLOY MA 956
INCONEL MA 754
INCONEL MA 758
INCONEL MA 6000
Ni
Cr
Fe
Ti
Al
Others
0.05
0.05
0.05
Bal
Bal
Bal
20.0
20.0
30.0
15.0
Bal
1.0
0.5
0.5
0.5
2.5
4.5
0.3
0.3
4.5
Y2O3: 0.5
Y2O3: 0.6
Y2O3: 0.6
Y2O3: 1.1, Mo: 2.0, W: 4.0, Ta: 2.0
pg 450
26/10/2007 1:55PM Plate # 0
pg 451
Table 7 Chemical compositions of iron-, nickel- and cobalt-base filler metals

Filler Metal
UNS
Cr
Ni
Fe
Co
409
409Nb
410
410NiMo
420
430
26-1
630
ER409
ER409Nb
ER410
ER410NiMo
ER420
ER430
ER26-1
ER630
AWS
S40900
S40940
S41080
S41086
0.08(a)
0.08(a)
0.12(a)
0.06(a)
0.250.40
0.1(a)
0.01(a)
0.05(a)
10.513.5
10.513.5
11.513.5
11.012.5
12.014.0
15.517.0
25.027.5
16.016.75
0.6(a)
0.6 (a)
0.6 (a)
4.05.0
0.6(a)
0.6(a)
4.55.0
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
16-8-2
Nitronic 50W
ER16-8-2
ER209
S20980
0.1(a)
0.05(a)
14.516.5
20.524.0
7.59.5
9.512.0
Bal
Bal
Nitronic 60W
ER218
S21880
0.1(a)
16.018.0
8.09.0
Bal
219
240
307
308
308H
308L
308Mo
308LMo
308Si
308LSi
309
309L
309Mo
309LMo
309Si
309LSi
310
312
316
316H
316L
316Si
316LSi
317
317L
321
347
347Si
ER219
ER240
ER307
ER308
ER308H
ER308L
ER308Mo
ER308LMo
ER308Si
ER308LSi
ER309
ER309L
ER309Mo
ER309LMo
ER309Si
ER309LSi
ER310
ER312
ER316
ER316H
ER316L
ER316Si
ER316LSi
ER317
ER317L
ER321
ER347
ER347Si
S21980
S24080
S30780
S30880
S30880
S30883
S30882
S30886
S30881
S30888
S30980
S30983
S30982
S30986
S30981
S30988
S31080
S31280
S31680
S31680
S31683
S31681
S31688
S31780
S31783
S32180
S34780
S34788
0.05(a)
0.05(a)
0.040.14
0.08(a)
0.040.08
0.03(a)
0.08(a)
0.04(a)
0.08(a)
0.03(a)
0.12(a)
0.03(a)
0.12(a)
0.03(a)
0.12(a)
0.03(a)
0.080.15
0.15(a)
0.08(a)
0.040.08
0.03(a)
0.08(a)
0.03(a)
0.08(a)
0.03(a)
0.08(a)
0.08(a)
0.08(a)
19.021.5
17.019.0
19.522.0
19.522.0
19.522.0
19.522.0
18.021.0
18.021.0
19.522.0
19.522.0
23.025.0
23.025.0
23.025.0
23.025.0
23.025.0
23.025.0
25.028.0
28.032.0
18.020.0
18.020.0
18.020.0
18.020.0
18.020.0
18.520.5
18.520.5
18.520.5
19.021.5
19.021.5
5.57.0
4.06.0
8.010.7
9.011.0
9.011.0
9.011.0
9.012.0
9.012.0
9.011.0
9.011.0
12.014.0
12.014.0
12.014.0
12.014.0
12.014.0
12.014.0
20.022.5
8.010.5
11.014.0
11.014.0
11.014.0
11.014.0
11.014.0
13.015.0
13.015.0
9.010.5
9.011.0
9.011.0
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
318
ER318
S31880
0.08(a)
18.020.0
11.014.0
Bal
320
ER320
N08021 0.07(a)
19.021.0
32.036.0
Bal
320LR
ER320LR
N08022 0.025(a)
19.021.0
32.036.0
Bal
383
904L
2209
2507
253MA
RA85H
RA330
Marathon
21/33
353MA
33
ER383
ER385
ER2209
N08028
N08904
S39209
S32750
S30815
S30615
N08334
0.025(a)
0.025(a)
0.02(a)
0.02(a)
0.07
0.2
0.25
0.12
26.528.5
19.521.5
22.5
25
21
18.5
19
21.5
30.033.0
24.026.0
8
10
10
14.5
35
32
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
0.03
R20033 0.015(a)
28
31.035.0
34
30.033.0
Bal
Bal
556
ER3556
R30556 0.1
22
20
Bal
18
R30155 0.12
21
20
Bal
20
MULTIMET
(N-155)
NOREM 02A
NOREM 03A
61
67
ERNi-1
ERCuNi
(a) Maximum. Nb = Cb
1.101.35 23.026.0
0.91.2
20.023.5
N02061 0.15(a)
C71581 0.04(a)
3.75.0
4.05.0
93.0 min
29.032.0
(continued)
Bal
Bal
1(a)
0.400.70
Others
Ti: 10 C min/1.5 max

Nb: 10 C min/0.75 max
N: 0.015
Cu: 3.254.0,
Nb + Ta: 0.150.30,
Mo: 1.02.0
Mo: 1.53.0, Mn: 4.07.0,
N: 0.2
Mn: 8.0, Si: 3.54.5,
N: 0.080.18
Mn: 9.0, N: 0.2
Mn: 10.513.5
Mo: 0.51.5, Mn: 3.34.75
Mo: 2.03.0
Mo: 2.03.0
Si: 0.651.0
Si: 0.651.0
Mo: 2.03.0
Mo: 2.03.0
Si: 0.651.0
Si: 0.651.0
Mo: 2.03.0
Mo: 2.03.0
Mo: 2.03.0
Mo: 2.03.0
Mo: 2.03.0
Mo: 3.04.0
Mo: 3.04.0
Ti: 9 C min/1.0 max
Nb: 10 C min/1.0 max
Nb: 10 C min/1.0 max,
Si: 0.651.0
Mo: 2.03.0,
Nb: 8 C min/1.0 max
Mo: 2.03.0, Cu: 3.04.0,
Nb: 8 C min/1.0 max
Mo: 2.03.0, Cu: 3.04.0,
Nb: 8 C min/0.4 max
Mo: 3.24.2, Cu: 0.71.5
Mo: 4.25.2, Cu: 1.22.0
Mo: 3, N: 0.14
Mo: 4, N: 0.25
Si: 1.6, N: 0.16, Ce: 0.05
Si: 3.7, Al: 1.0
Mn: 5.25
Nb: 1.7
N: 0.15, Ce: 0.03
Mo: 0.52.0, Cu: 0.31.2,
N: 0.350.6
Mo: 3, W: 2.5, Ta: 0.6,
N: 0.2, Zr: 0.02, La: 0.02
Mo: 3, W: 2.5, Nb + Ta: 1,
N: 0.15
Mn: 4.5, Si: 3.3, Mo: 2.0
Mn: 4.0, Si: 2.8,
Mo: 1.72.2
Ti: 2.03.5, Al: 1.5(a)
Cu: Bal, Ti: 0.200.50
26/10/2007 1:55PM Plate # 0
Table 7
(Continued)
Filler Metal
Cr
Ni
Fe
Co
60
ERNiCu-7
AWS
N04060 0.15(a)
62.069.0
2.5(a)
ERNiCu-8
N05504 0.25(a)
63.070.0
2(a)
82
72
76
62
92
52
52M
ERNiCr-3
ERNiCr-4
ERNiCr-6
ERNiCrFe-5
ERNiCrFe-6
ERNiCrFe-7
ERNiCrFe-7A
N06082
N06072
N06076
N06062
N07092
N06052
N06054
18.022.0
42.046.0
19.021.0
14.017.0
14.017.0
28.031.5
28.031.5
67.0 min
Bal
75.0 min
70.0 min
67.0 min
Bal
Bal
3(a)
0.5(a)
2(a)
6.010.0
8(a)
7.011.0
7.011.0
ERNiCrFe-8
N07069 0.08(a)
14.017.0
70.0 min
5.09.0
53MD
ERNiCrFeAl-1
N06693 0.15(a)
27.031.0
Bal
2.56.0
601
602CA
ERNiCrFe-11
N06601 0.1(a)
N06025 0.2
21.025.0
25
58.063.0 14
Bal
10
214
65
ERNiFeCr-1
0.05(a)
N08065 0.05(a)
16
19.523.5
Bal
3
38.046.0 22.0 min
718
ERNiFeCr-2
N07718 0.08(a)
17.021.0
B
N
W
242
B-2
B-3
G
ERNiMo-1
ERNiMo-2
ERNiMo-3
ERNiMo-7
ERNiMo-8
ERNiMo-9
ERNiMo-10
ERNiCrMo-1
N10001
N10003
N10004
N10665
N10008
N10009
N10675
N06007
0.08(a)
1(a)
0.040.08 6.08.0
0.12(a)
4.06.0
0.03(a)
8
0.02(a)
1(a)
0.1(a)
0.53.5
0.1(a)
0.01(a)
1.03.0
0.05(a)
21.023.5
Bal
Bal
Bal
Bal
Bal
60.0 min
65.0 min
65.0 min
Bal
4.07.0
5(a)
4.07.0
2(a)
2(a)
10(a)
5(a)
1.03.0
18.021.0
X
625
C-276
C-4
S
ERNiCrMo-2
ERNiCrMo-3
ERNiCrMo-4
ERNiCrMo-7
N06002
N06625
N10276
N06455
0.050.15
0.1(a)
0.02(a)
0.015(a)
0.02(a)
20.523.0
20.023.0
14.516.5
14.018.0
14.517.0
Bal
58.0 min
Bal
Bal
Bal
17.020.0
5(a)
4.07.0
3.0(a)
3.0(a)
ERNiCrMo-8
N06975 0.03(a)
23.026.0
47.052.0 16
G-3
G-30
ERNiCrMo-9
ERNiCrMo-11
N06985 0.015(a)
N06030 0.03(a)
21.023.5
28.031.5
G-35
RA333
C-22/622
C-22HS
59
686CPT
C-2000
725
ERNiCrMo-10
ERNiCrMo-13
ERNiCrMo-14
ERNiCrMo-17
ERNiCrMo-15
N06035
N06333
N06022
N06059
N06686
N06200
N07725
33
25
20.022.5
21
22.024.0
19.023.0
23
19.022.5
617
230-W
ERNiCrCoMo-1 N06617 0.050.15 20.024.0

ERNiCrWMo-1 N06231 0.050.15 20.024.0
Bal
Bal
3.0(a)
3.0(a)
HR-160
R-41
WASPALOY
ERNiCoCrSi-1
Bal
Bal
Bal
2.0(a)
5.0(a)
2.0(a)
DELORO 40
DELORO 50
DELORO 60
Colmonoy 88
188
25 (L-605)
Bal
Bal
Bal
Bal
2.5
3.5
4
3
3.0(a)
3.0(a)
UNS
0.1(a)
0.010.10
0.080.15
0.08(a)
0.08(a)
0.04(a)
0.04(a)
0.05(a)
0.05
0.015(a)
0.01(a)
0.01(a)
0.01(a)
0.01(a)
0.03(a)
N12160 0.05
28
0.050.12 19
0.08
19
0.2
0.45
0.7
0.8
0.050.15
0.1
7.5
10.5
14.5
15
22
20
Bal
Bal
Bal
18.5
18.021.0
15
Bal
2.0(a)
Bal
17
Bal
2.06.0
Bal
2.0(a)
Bal
1.5(a)
Bal
5.0(a)
Bal
3.0(a)
55.059.0 7
22
10
(continued)
Others
Cu: Bal, Mn: 4.0(a),

Ti: 1.53.0, Al: 1.25(a)
Cu: Bal, Ti: 0.251.00,

Al: 2.04.0
Nb + Ta: 2.5, Mn: 3.0
Ti: 0.31.0
Ti: 0.150.50
Nb + Ta: 1.53.0
Ti: 3.0, Mn: 2.02.7
Ti: 1.0(a), Al: 1.1(a)
Ti: 1.0(a), Al: 1.1(a),

Nb + Ta: 0.501.0
Ti: 2.02.75, Al: 0.41.0,

Nb + Ta: 0.71.2
Al: 2.54.0, Nb + Ta:

0.52.5, Ti: 1.0(a)
Al: 1.01.7
Al: 2.1, Ti: 0.15,

Y: 0.050.12, Zr: 0.010.1
Al: 4.5, Y: 0.01
Mo: 3, Cu: 1.53.0,

Ti: 0.61.2
Ti: 0.651.15,
Al: 0.200.80,
Nb + Ta: 4.755.50
2.5(a)
Mo: 26.030.0, V: 0.3
Mo: 15.018.0
Mo: 23.026.0
Mo: 25.0
Mo: 26.030.0
Mo: 18.021.0, W: 3.0
Mo: 19.022.0, W: 3.0
Mo: 27.032.0, Mn: 3.0(a)

2.5(a)
Mo: 6.5, Cu: 2.0,
Nb + Ta: 1.752.5,
Mn: 1.5
0.52.5 Mo: 9.0, W: 0.21.0
Mo: 9.0, Nb + Ta: 3.154.15

2.5(a)
Mo: 16.0, W: 3.04.5
2.0(a)
Mo: 16.0
2.0(a)
Mo: 14.016.5, Al: 0.10.5,
La: 0.010.1
Mo: 6.0, Ti: 0.71.5,

Cu: 1.0
5.0(a)
Mo: 7.0, W: 1.5(a), Cu: 2.0
5.0(a)
Mo: 5.0, W: 1.54.0,
Nb + Ta: 0.31.5,
Cu:1.02.4
Mo: 8.0
3
Mo: 3.0, W: 3.0, Mn: 2.5
2.5(a)
Mo: 12.514.5, W: 2.53.5
Mo: 17.0
Mo: 16.0, Al: 0.10.4
Mo: 16.0, W: 3.04.4
Mo: 16.0, Cu: 1.6
Mo: 7.09.5, Nb + Ta:

2.754.0, Ti: 1.01.7
10.015.0 Mo: 9.0, Al: 0.81.5
5.0(a)
W: 14.0, Mo: 2.0,
Al: 0.20.5
29
Si: 2.75, Ti: 0.5, Nb: 1.0(a)
11
Mo: 10.0, Ti: 3.1, Al: 1.5
13.5
Mo: 4.3, Al: 1.5, Ti: 3.0,
Zr: 0.05
1.5(a)
Si: 3.5
Si: 4.0, B: 2.0
Si: 2.04.5, B: 3.3
W: 17.0, Si: 4.0, B: 3.0

Bal
W: 14.0, La: 0.020.12
Bal
W: 15.0
pg 452
26/10/2007 1:55PM Plate # 0
pg 453
Table 7 (Continued)
Filler Metal
ULTIMET
Stellite 1
Stellite 6
Stellite 12
Stellite 21
Stellite 6K
Stellite 704
Stellite 706
Stellite 712
Stellite 706K
Tribaloy T-400C
Tribaloy T-401
AWS
UNS
R31233
0.06
2.45
1.2
1.6
0.25
1.6
1
1.2
1.6
1.6
0.1(a)
0.2
Cr
26
31
28
29.5
27
31
30
29
29
31
14
17
Ni
9
3.0(a)
3.0(a)
3.0(a)
2.5(a)
3.0(a)
2.0(a)
3.0(a)
3.0(a)
3.0(a)
1.0(a)
0.8(a)
Fe
Co
Others
3
2.5(a)
3.0(a)
2.5(a)
3.0(a)
3.0(a)
2.0(a)
3.0(a)
3.0(a)
3.0(a)
1.0(a)
0.8(a)
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Bal
Mo: 5.0, W: 2.0

W: 13.0
W: 4.5
W: 8.5
W: 5.5
W: 4.5
Mo: 14.0
Mo: 4.5
Mo: 8.5
Mo: 4.5
Mo: 27.0, Si: 2.6
Mo: 22.0, Si: 1.3

High Temp Corrosion

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

High Temp Corrosion

Uploaded by

Copyright:

Available Formats

Name ///sr-nova/Dclabs_wip/High Temp/5208_1.

26/10/2007 1:51PM Plate # 0

High-Temperature Corrosion And Materials Applications

Copyright 2007 ASM International

cracking, stress-relaxation cracking, or

The focus of this book is on commercial

Name ///sr-nova/Dclabs_wip/High Temp/5208_3-4.pdf/Chap_02/

26/10/2007 12:10PM Plate # 0

High-Temperature Corrosion And Materials Applications

Copyright 2007 ASM International

Challenges in Materials Applications

analysis needs to be conducted to balance the

Name ///sr-nova/Dclabs_wip/High Temp/5208_3-4.pdf/Chap_02/

26/10/2007 12:10PM Plate # 0

4 / High-Temperature Corrosion and Materials Applications

corrosion, increasing tensile and creep-rupture

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

High-Temperature Corrosion And Materials Applications

31/10/2007 12:43PM Plate # 0

Copyright 2007 ASM International

sluggish for most alloys. As a result, the effects of

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:43PM Plate # 0

6 / High-Temperature Corrosion and Materials Applications

3.2 Thermodynamic Considerations

Assuming the activities of the metal and the

Equation 3.4 permits the determination of the

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:43PM Plate # 0

partial pressure of oxygen in equilibrium with

oxide from the standard free energy of formation.

2H2 +O2 2H2 O

The standard free energy of formation is related

G =RT In pO2 (kJ/mole O2)

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:43PM Plate # 0

8 / High-Temperature Corrosion and Materials Applications

Rearranging the equation results in:

Thus, the oxygen partial pressures at various

The corresponding oxygen potential is:

The pCO =pCO2 value can be read from Fig. 3.2

that protects the alloy from rapid oxidation. This

Table 3.1 Melting points of selected oxides

Melting point, C (F)

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:43PM Plate # 0

volatile CrO3 (Ref 58). Table 3.2 shows the

strengthening. The formation of MoO3 and its

Table 3.2 Weight loss of Cr2O3 on heating in

Vapor pressure, atm

Inverse temperature, 103/T (K1)

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:43PM Plate # 0

10 / High-Temperature Corrosion and Materials Applications

For service temperatures above 1200 C

3.3 Kinetic Considerations

Different oxidation kinetics. Source: Ref 11

decreases with increasing time due to the

where X is the oxide scale thickness, t is the

3.3.1 Parabolic Kinetics

3.3.2 Linear Kinetics

Name ///sr-nova/Dclabs_wip/High Temp/5208_5-66.pdf/Chap_03/

31/10/2007 12:43PM Plate # 0

time. The linear oxidation kinetic rate can be