ULSD Brochure

Ultra Low Sulfur Diesel
2010
Advanced Refining Technologies Hydroprocessing Catalysts from

Chevron and Grace Davison
Coker Naphtha
Naphtha
Kerosene
& Jet
ATTM Series
ATTM Series
DXTM Series
Diesel
Hydrocracker
Pretreat
ATTM Series
DXTM Series
ICR Series
FCC Pretreat
StARTTM
Catalyst
System
AT Series
SmART
Catalyst
System
DX Series
ApARTTM
Catalyst
System
AT, DX Series
Vacuum Tower
ART brings new standards of product innovation and customer service to the global
refining industry by harnessing the technical expertise of both Chevron and Grace
Davison in catalyst development and their
broad experience in supporting refiners.
With the combination of Grace Davisons
material science, manufacturing, marketing
and sales strength, and Chevrons extensive
experience from operating its own refineries
and leadership in design and process licensing, ART offers refiners one-stop access to hydroprocessing knowledge and
technical service unparalleled in the industry.
Product Overview Chart
Crude Tower
Advanced Refining Technologies (ART), is

the joint venture of Chevron Products Company and W. R. Grace & Co.s Grace Davison catalysts business unit, created to
develop, market and sell a comprehensive
line of state-of-the-art hydroprocessing catalysts. Since our formation in March 2001,
we have firmly established ourselves as a
leading supplier of premium hydroprocessing catalysts and technical service to the
petroleum refining industry worldwide.
ICR Series
Fixed Bed Resid
HOP Series
Fixed Bed Resid
ART offers refiners the unique synergies deEbullating Bed

GR Series
Resid for LC-Fining
LSTM Series
rived from our parent joint venture partners
HSLSTM Series and H-Oil processes
being leaders in hydroprocessing catalyst
technologies. Chevron, and now Chevron
Lummus Global (CLG) represent over 35
years of experience designing and licensing RDS and VRDS units, as well as operating resid hydroprocessing and distillate hydrotreating units at Chevron refineries. Grace Davison has over 50 years of experience in extruding hydroprocessing catalysts. Leveraging the refining
and hydroprocessing experience of Chevron with Grace Davisons expertise in specialty materials and catalyst technologies makes ART uniquely equipped to quickly commercialize and deliver innovative products to the dynamic hydroprocessing marketplace.
To keep pace with the high demand for ART catalysts driven by heavy resid hydroprocessing applications and more
stringent clean fuels standards, ART has expanded manufacturing capacity at its existing North American plants. ART
acquired Orient Catalyst Companys (OCC) hydroprocessing catalyst technologies, and the HOP catalyst product line
in 2002; and obtained an ownership position in Kuwait Catalyst Company (KCC) which has manufactured HOP resid
hydroprocessing catalysts under license since 2001.
ARTs comprehensive line of hydroprocessing catalysts deliver maximum sulfur removal and upgrading for a wide
range of feedstocks. ART offers hydroprocessing catalysts for the resid segment of the hydroprocessing market, including the fixed-bed Onstream Catalyst Replacement (OCR) resid process and ebullating bed applications, as well
as a complete product line of premium catalysts for distillate hydrotreating of heavy VGO/DAO, diesel, kerosene and
light naphtha applications.
Contents
Understanding Ultra Low Sulfur Diesel
By Gerianne DAngelo and Charles Olsen
This article discusses the importance of feedstock properties and operating conditions to successfully producing ultra low sulfur diesel.
The SmART Catalyst System: Meeting the Challenges of Ultra Low Sulfur Diesel
By Charles Olsen
A disucssion of the various components of the SmART catalyst system and the factors that go
into designing customized systems to meet a variety of ULSD processing needs.
No Need to Trade ULSD Catalyst Performance for Hydrogen Limits: SmART Approaches
By Charles Olsen and Geri DAngelo
The SmART Catalyst System provides the optimum balance between activity and hydrogen consumption. This article discusses the chemistry and kinetics involved in designing a system, and
reviews the commercial performance observed in several ULSD applications.
15
ART Excels In ULSD Service: Update on Sulfur minimization By ART

By Greg Rosinski, Dave Krenzke, and Charles Olsen
ULSD production with the first SmART Catalyst System Series began early in 2004 at a North American
refinery processing a feed containing 40% of a high endpoint LCO. Since that time DX Platform Catalysts
have been selected for over 70 ULSD applications as either stand-alone catalysts or as components in
SmART System. The technology has been a great success since its introduction with millions of pounds
installed in commercial units around the world.
23
Distillate Pool Maximization by Exploiting the use of Opportunity

Feedstocks Such as LCO and Syncrude
By Brian Watkins and Charles Olsen
The use of opportunity feeds such as FCC LCGO, diesel streams from other hydroprocessing
units and feeds from synthetic crude sources has helped refiners to maximize their diesel pool.
In this article we highlight differences in feed reactivity for various feed components and explore
the impact of various Hydrotreating Catalysts and operating conditions on diesel production.
30
Cetane Improvement In Diesel Hydrotreating

by Greg Rosinski and Charles Olsen
This article discusses the importance of cetane in ULSD. The SmART Catalyst System, which utilizes both the CoMo and NiMo catalyst, results in a cetane uplift which is nearly two numbers
higher than an all-CoMo system with only a small increase in hydrogen consumption. For H2 constrained refiners this is an ideal solution for improving the product cetane.
38
Factors Influencing ULSD Product Color

By Greg Rosinski, Charles Olsen and Brian Watkins
Product Color of petroleum products such as kerosene, jet fuel, diesel fuel and lube oils is a concern. Unit cycle length can be shortened due to product color degradation. In this paper we identify components that contribute to color degradation and report on the effects of feedstocks and
operating conditions on ULSD color.
43
Understanding
Charles Olsen
New Product Development
Manager
Gerianne DAngelo
Senior Technical Services Engineer
ADVANCED REFINING
TECHNOLOGIES
Chicago, IL USA
Very tight specifications on sulfur content in diesel fuel

have recently been announced by the European Union,
the U.S. EPA and others. Refiners will be required to
produce transportation diesel fuel with sulfur levels
below ten ppm (Ultra Low Sulfur Diesel or ULSD).
These new specifications will place a severe burden on
diesel hydrotreaters as refiners struggle to keep up
with diesel fuel demand and quality.
As the diesel sulfur specification drops to the low ppm
range, it is important to look at the types of sulfur compounds that need to be removed in order to meet the
lower sulfur levels. The sulfur compounds can be classified into different groups commonly termed easy
and hard sulfur, based on the ease of sulfur removal.
The easy sulfur compounds are represented by dibenzothiophenes (DBT) and lighter. The hard sulfur
species include multi-substituted dibenzothiophenes
commonly represented by 4,6 dimethyl-dibenzothiophene (DMDBT), and these compounds are extremely
difficult to desulfurize. The reason for the reactivity differences can be explained by the different reaction
pathways for sulfur removal from DBTs and substituted
DBTs.
DBTs are more effectively desulfurized via a direct sulfur abstraction route, and it is the generally accepted
reaction pathway for HDS of diesel to 500 ppm sulfur.
Substituted DBTs, on the other hand, are not effectively desulfurized via direct abstraction because the
methyl group(s) adjacent to the sulfur atom effectively
shield the sulfur from the catalyst active sites. However,
hydrogenation of an aromatic ring allows the molecule
to flex enabling access to the sulfur atom, and C-S
bond scission readily follows. Direct abstraction is
catalyzed more effectively by CoMo catalysts, while the
hydrogenation-abstraction route is typically more facile
over NiMo catalysts. This forms the basis for ARTs
www.artcatalysts.com
Figure 22
Fraction Hard Sulfur Correlates with Feed End Point
0.7
0.6
Fraction Hard Sulfur
0.5
0.4
0.3
0.2
0.1
0.0
232
260
288
316
343
371
399
Feed D86 90% Point, C
SR
SmART Catalyst System which is tailored to optimize

both reaction pathways.
The relative amounts of easy and hard sulfur in a feed
are critical parameters to consider when choosing a
catalyst and operating conditions for production of
ULSD. Their concentration can vary significantly from
feed to feed depending on crude source, the boiling
range, and the treatment the feed has been subjected
to whether it be thermal (coker) or catalytic treating
(cycle oil).
Analysis of a wide array of diesel feeds shows that the
fraction of hard sulfur correlates quite well with the
end point of the feed. Figure 22 shows the fraction of
hard sulfur as a function of D86 90% point for a variety
of feeds. The figure clearly shows a strong correlation
with the 90% point of the feed. The fraction of hard
sulfur increases rapidly for feed T90 between about
316C and 343C. Interestingly, the fraction of hard
sulfur does not correlate strongly with feed type;
whether it is straight run, light cycle oil, or light coker
gas oil the fraction of hard sulfur is largely determined
by the feed endpoint.
The amount of nitrogen in the feedstock is another critical parameter that must be considered. Figure 23
demonstrates the detrimental effects nitrogen has on
the removal of hard sulfur. The relative rate constant for
hard sulfur removal decreases significantly in going
LCGO
LCO
from a feedstock containing 50 ppm nitrogen to a feedstock containing 150 ppm nitrogen. The relative rate
constant continues to decrease as nitrogen content
increases beyond 150 ppm, but the decrease is at a
somewhat slower rate. This inhibiting effect of nitrogen
is a result of poisoning of the acid sites needed for aromatic ring saturation, and as discussed above, the saturation-abstraction route is the preferred reaction pathway for hard sulfur removal.
Thermodynamic equilibrium is another issue related to
poly nuclear aromatic (PNA) saturation which must be
taken into account. The saturation of PNAs does reach
a thermodynamic constraint on conversion at high
enough temperatures. Figure 24 shows the expected
PNA conversion as a function of temperature for a variety of LHSVs. The equilibrium constraint is readily
apparent for the higher temperatures in the chart.
Decreasing the LHSV serves to increase conversion in
the kinetically controlled region (i.e. low temperatures),
but has no effect on conversion in the thermodynamically controlled regime (i.e. high temperatures). This
suggests that adding catalyst volume will help in the
removal of hard sulfur, but a point is reached where
decreases in sulfur do not occur due to unfavorable
thermodynamic equilibrium at those conditions. This
raises the potential for situations where significant
increases in temperature will not result in appreciable
decreases in product sulfur at very low levels.
reprinted from Catalagram 95 2004
Figure 23
Nitrogen Inhibits Hard Sulfur Removal
A number of operating parameters take on greater

importance in ultra low diesel HDS due to the issues
raised above. ART has conducted extensive pilot plant
work to define appropriate operating conditions for
ULSD and to define the influence of cracked stocks.
One study utilized a straight run feed and a 20% LCO
blend with the same straight run component, and
selected properties of the two feeds are shown in
Table X.
For the high pressure operation, it is apparent that

increasing H2/oil is beneficial for both feeds up to a
certain point. The reactivity of the straight run feed
increases from about 60 at low H2/oil (i.e. 60% of the
highest activity achievable for that feed) to nearly 100
at a H2/oil ratio / H2 consumption ratio of about five.
Further increases in H2 beyond that provide little additional benefit.
The operating conditions used in the work are representative of the pressure and H2/oil ranges typically
found in diesel hydrotreaters. Figure 25 summarizes
the influence of H2/oil ratio on ULSD production. It is
a plot of the relative HDS rate constant as a function of
the relative excess hydrogen (H2/oil ratio divided by the
hydrogen consumption). Data for both high and low
pressure operation are shown for each feed.
Similarly, the 20% LCO feed shows a reactivity of about

20 at low hydrogen rate, and steadily increases
towards 100 near a H2/oil ratio / H2 consumption ratio
of about six. Further increases in hydrogen rate again
do not add much additional benefit. These data suggest that for high pressure more excess hydrogen is
required for ULSD compared to conventional diesel
hydrotreating (i.e. 500 ppm sulfur) where H2/oil ratio
consumption ratios of between three and four are generally recommended.
Table X
Feed Properties
Specific Gravity
Sulfur, wt.%
Nitrogen, ppm
D86 Dist., C, IP/50/FP
Straight Run
20% LCO/Straight Run
0.847
1.21
95
212/288/354
0.868
1.56
200
208/286/353
Figure 24
Thermodynamic Equilibrium Can Limit Conversion
The effect of H2/oil ratio at lower pressure is quite different. In this case, the relative HDS rate constant for
both feeds shows a steady increase with increasing
hydrogen rate. Note that the activity is relative to the
straight run feed at high pressure and high H2 rate.
The benefit of increasing H2/oil never reaches a

plateau as observed in the high pressure case, suggesting that more hydrogen is always better at low
pressure. The effect of hydrogen pressure is also
readily apparent in Figure 25. At relative excess hydro-
Figure 25
Excess Hydrogen Improves Reactivity
Figure 26
H2S is Detrimental to Hard Sulfur Removal
gen values around three to four, the straight run feed at

low pressure has a relative rate constant of about 20
compared to the high pressure case, and the difference for the LCO feed is even greater.
Figure 26 summarizes these data in another way. This
chart shows the relative activity plotted as a function of
H2S partial pressure at reactor outlet conditions.
Again, quite different behavior is observed for the low
and high pressure cases. At high pressure, the data
suggest some tolerance for low H2S pressure and a
steady decline in activity as H2S pressure increases
beyond that. The straight run feed also appears more
forgiving, as it has a slower activity decline compared
to the LCO containing feed. Looking at the low pressure data, it is apparent that even a small amount of
H2S is detrimental to activity. There is a rapid initial
decrease in activity with the first increments of H2S,
and thereafter a slow steady drop-off in activity as H2S
level increases further. These data are from the pilot
plant where there is no H2S in the treat gas to the unit.
A hydrogen stream containing H2S will exacerbate the
activity loss due to H2S especially at low pressure.
This study also investigated the effects of adding catalyst volume and the potential tradeoffs between cata-
lyst volume, operating pressure and H2/oil ratio. Figure

27 summarizes the relationship between 1/LHSV or
residence time (i.e. catalyst volume), H2/oil ratio and
hydrogen pressure at 10 ppm product sulfur for the
straight run feed. Not surprisingly, there is a significant difference in required catalyst volume between
high and low pressure operation at constant H2/oil
ratio. The catalyst volume at low pressure is nearly
three times higher than required for the high pressure
operation at typical H2 rates.
Figure 28 is a similar plot, but in this case it is comparing the performance of the straight run feed with the
20% LCO blend at an intermediate pressure. The intermediate pressure is the practical lower limit for ULSD
production for the 20% LCO feed. Lower pressures
required unreasonably high gas rates and or long residence times (very large reactors). This chart clearly
shows the difficulty in treating cracked stocks relative
to a straight run feed. At high H2 rates, 1.5 times higher catalyst volume is required for the LCO, and this
increases to nearly two times higher at the low H2 rate.
Of course, increasing the pressure helps somewhat,
but even at high pressure the LCO case requires 1.31.7 times higher catalyst volume depending upon the
H2 rate.
Figure 27
Trade-offs Between Pressure, H2/Oil and LHSV
In summary, it is possible in some cases to trade off

pressure, H2/oil and catalyst volume, when designing
for ULSD production. However, there is a limit which
depends strongly upon the feedstock. For example,
cracked stocks require higher pressure and higher

H2/oil ratio compared to a straight run feed. To learn
more about producing ULSD at your refinery, contact
your ART technical or sales representative.
Figure 28
Cracked Stocks Change The Rules
The SmART Catalyst SystemTM:

Meeting the Challenges of
Charles Olsen
Manager
ADVANCED REFINING
TECHNOLOGIES
Chicago, IL USA
In 2001, ART introduced the SmART Catalyst SystemTM

to help refiners deal with the severe demands of ultra
low sulfur diesel (ULSD). The SmART Catalyst System
utilizes state-of-the-art catalyst technology which is
staged in the proper proportions to provide the best
performance while at the same time meeting individual
refiner requirements. The catalyst staging is designed
to take advantage of the different reaction mechanisms for sulfur removal; ART CDX, a high activity
CoMo catalyst, efficiently removes the unhindered,
easy sulfur via the direct abstraction route and ART
CDY, a high activity NiMo catalyst, then attacks the
remaining sterically hindered, hard sulfur. Pilot plant
work has proven that the properly configured SmART
Catalyst System provides higher activity than either the
CoMo or NiMo catalyst alone.
ART CDX and ART CDY, individually or as part of a
SmART Catalyst System, were selected for 14 diesel
units in 2004, and most of these applications aim to
evaluate ULSD capability and/or produce ultra low sulfur fuels in advance of the regulations for economic
benefit.
Optimizing the SmART Catalyst System
An important aspect of the SmART Catalyst System is
determination of the optimum proportions of the CoMo
and NiMo catalysts that will deliver the best performance. This is dependent upon a number of factors,
including the refiners requirements, and selected feed
properties and operating conditions as discussed in

detail previously in Catalagram No. 95 (March 2004).
One clearly important parameter which must be considered is the boiling range of the feedstock. Sulfur
speciation on a wide variety of feedstocks has shown
that there is a strong correlation between the fraction of
multi-substituted dibenzothiophenes (hard sulfur) and
the feed endpoint. Once the D86 endpoint increases
beyond about 329C there is a rapid increase in the
fraction of hard sulfur contained in the feed. This has
a large impact on catalyst activity as shown in Figure
21. The figure shows pilot plant data comparing
results from treating two feeds with different endpoints
over the same catalyst under identical conditions. At
ultra low sulfur levels there is about 17C difference in
reactivity of the two feeds with the lower endpoint feed
more reactive. Clearly, feed endpoint and the amount
of hard sulfur are critical parameters that influence the
optimum SmART configuration.
Another critical feed property that must be accounted
for is the nitrogen content. It is generally accepted that
nitrogen inhibits aromatic saturation reactions through
poisoning of acidic sites on the catalyst. Recall that
the primary reaction pathway for removal of hard sulfur
is via hydrogenation of an aromatic ring, and it is not
surprising that feed nitrogen content has a serious,
negative impact on HDS activity. The magnitude of the
impact can be seen in Figure 22 which summarizes
data for NiMo and CoMo catalyst activity on an SR feed
Figure 21
Feed Endpoint Impacts HDS Activity
357C D86 EP
390C D86 EP
Temperature, C
17C
31F
10C
1000
100
10
Product Sulfur, ppm
before and after selectively removing the nitrogen via

an adsorption process. The difference in activity on
the two feeds is quite large. Increasing the nitrogen
content from 25 to 160 ppm results in a loss in HDS
activity of 22-28C for both catalysts. Comparing the
catalysts on the low nitrogen feed shows that the NiMo
catalyst has about 8C higher activity relative to the
CoMo, and that decreases to an advantage of about
3C or less on the higher nitrogen feed. This suggests
the impact of nitrogen is different for NiMo and CoMo
catalysts with the CoMo catalyst more tolerant of nitro-
gen. This is another important consideration when

designing the optimum SmART Catalyst System.
Hydrogen availability, in terms of hydrogen pressure
and hydrogen circulation, also takes on greater importance in ULSD. Figure 23 is a chart showing how the
relative HDS rate constant changes as a function of the
excess hydrogen (H2/Oil ratio divided by the hydrogen
consumption) for both high and low pressure operation. Note the range in operating pressure from low to
high represents that typically encountered in diesel
Figure 22
Nitrogen Impacts Catalysts Differently
Increase in required Temperature, C
28
22
17
11
6
Base
CoMo
0
160 ppm Nitrogen
NiMo
25 ppm Nitrogen
Figure 23
Hydrogen Availability is Critical
hydrotreating. At high pressure, increasing the H2/Oil

is beneficial for both SR and 20% LCO feeds up to a
point, after which further increases in H2 rate provide
little additional benefit. At low pressure, the effect of
H2/Oil ratio is quite different. In that case, the relative
rate constant for both feeds shows a steady increase
with increasing hydrogen rate. The benefit of increasing H2/Oil never reaches a plateau as observed in the
high pressure case indicating that more hydrogen is
always better at low pressure. The effect of pressure
is also readily apparent in the figure. Comparing the
relative rate constant for high and low pressure at a
typical H2/Oil ratio reveals that the activity at low pressure is only 10-20% of that at high pressure for 20%
LCO and SR feed respectively.
Controlling Hydrogen Consumption
One of the key advantages of the SmART Catalyst
System is the efficient use of Hydrogen. Figure 24
illustrates how the system can be tailored to provide
the best balance of high HDS activity while minimizing
H2 consumption. The figure shows that as NiMo catalyst is added to the system there is a large increase in
HDS activity relative to the all CoMo reference, and
eventually, a minimum in the product sulfur curve is
reached (i.e. maximum HDS activity). The position and
magnitude of this minimum varies with feed and operating conditions such as those discussed above. The
figure also shows the relative H2 consumption, and
again, as the percentage of the NiMo component
10 www.artcatalysts.com
increases, the H2 consumption relative to the base

CoMo system increases. In the region where the system shows the best activity, the hydrogen consumption
is only slightly greater than that for the all CoMo system, and well below that for the all NiMo catalyst. This
is a result of the different kinetics for sulfur and aromatics removal and is a critical consideration when
customizing a SmART Catalyst System.
To help understand the differences in kinetics it is useful
to compare the performance of CoMo and NiMo catalysts alone. Figure 25 shows a comparison of the hydrogen consumption over a NiMo and CoMo catalyst for a
straight run feed at ULSD conditions. The amount of
hydrogen consumed by sulfur, nitrogen and olefins
removal is essentially the same for each catalyst and is
not shown. What separates the two catalysts is the
amount of aromatics saturation which occurs, and in
particular, the amount of mono ringed aromatics which
are hydrogenated. In this case, an additional 12 Nm3/m3
of hydrogen is consumed with the NiMo catalyst due to
mono ringed aromatics saturation. This represents
excess hydrogen consumption above that required for
the removal of sulfur.
Figure 26 is a similar chart for LCGO feed. In this
example, the NiMo catalyst hydrogenates more PNA (2
rings and greater) and mono ringed aromatics compared to the CoMo catalyst accounting for an additional 19 Nm3/m3 of hydrogen consumption above that
required for sulfur, nitrogen and olefins removal.
Clearly, in cases where hydrogen consumption needs
Figure 24
Optimizing HDS and H2 Consumption
Figure 25
Excess H2 Consumption with SR Feed in ULSD
NiMo
CoMo
HDS
HDS
HDN
Olefins
28 Nm3/m3
Poly aromatics
32 Nm3/m3
HDN
Olefins
20 Nm3/m3
3
28 Nm /m
Mono aromatics
Mono aromatics
Poly aromatics
to be minimized, a NiMo catalyst for ULSD is the wrong

choice. Unfortunately, in many of these cases a CoMo
catalyst does not provide the best activity for ULSD,
and it is precisely these units where the SmART
Catalyst System is ideal as it offers the highest activity
and is more efficient with hydrogen compared to an all
NiMo system.
It perhaps appears contradictory that a NiMo catalyst
is included in the SmART Catalyst System due to its
high hydrogenation activity making it preferred for hard
sulfur removal, and yet, under some conditions the
hydrogenation activity is too high and excess aromatics saturation occurs. This highlights one of the keys
to designing the proper system. The design involves
increasing the hydrogenation selectivity of the system
to provide highest HDS activity, while at the same time
minimizing hydrogen consumption (i.e. minimizing

excess aromatics saturation).
Figure 27 shows a schematic of the reaction pathway
for poly aromatics saturation. Naphthalene, a two
ringed aromatic molecule, is featured at the top of the
figure. The reaction begins with the hydrogenation of
one of the aromatic rings to form tetralin, a mono ring
aromatic. The next reaction in the sequence is hydrogenation of the remaining aromatic ring to produce
decalin, a fully saturated molecule. Each step is
reversible and subject to equilibrium constraints. The
slowest reaction in the sequence is the saturation of
the mono ring aromatic. The reaction sequence for a
substituted biphenyl is also shown at the bottom of the
figure. The sequence is essentially the same. In both
examples the hydrogenation of the mono aromatic is
11
Figure 26
Excess H2 Consumption with LCGO Feed in ULSD
CoMo
NiMo
HDS
HDS
HDN
HDN
Olefins
Olefins
3
46 Nm /m
22 Nm3/m3
Mono aromatics
Poly aromatics
the slowest step and also consumes more hydrogen

compared to the other hydrogenation reactions shown;
three moles of hydrogen per mole of aromatic are consumed by the saturation of the mono ring aromatic
compared with two moles of hydrogen/mole for the
multi ringed aromatics.
The reaction sequences shown in Figure 27 can be
treated as a series of first order reversible reactions
where the intermediate (the mono ring aromatic) is the
desired product. For these sorts of reaction systems
the intermediate often tends to be favored at shorter
contact times (see Chemical Reaction Engineering by
Octave Levenspiel [Wiley, 1998], for example), and it
this tendency which is exploited when customizing a
SmART Catalyst System. The effective residence time
in both the CoMo and NiMo beds of a system is short
compared to the overall reactor residence time, and
this helps minimize the hydrogenation of the mono ring
aromatic (i.e. minimize hydrogen consumption).
42 Nm3/m3
Poly aromatics
8.0 Nm3/m3
Mono aromatics
Figure 28 shows the effect of residence time, as indicated by 1/LHSV, on aromatics saturation. For PNA
saturation, the two ringed aromatic going to the mono
ring aromatic, there is a fairly steep curve for conversion as a function of residence time below about 1 hr.
Above that point, which represents space velocities of
1 hr-1 or less there is very little change due to equilibrium constraints. For mono ring aromatic saturation
there is a steady increase in conversion as the residence time is increased indicating more and more saturation as the residence time is increased (i.e., LHSV is
decreased). Both sets of curves suggest aromatic saturation can be limited by appropriate choice of LHSV,
or in the case of a SmART Catalyst System, by adjusting the relative quantities of CoMo and NiMo catalyst.
SmART Catalyst System Experience
The SmART Catalyst System is the culmination of an
extensive effort put towards understanding the chemistry and process conditions required for ultra low sulfur fuels. In addition, properly designed high activity
Figure 27
Reaction Sequence for Poly Aromatics Hydrogenation
Figure 28
Limiting Aromatics Saturation
Table VII
Commercial Experience with ART CDX, CDY and the SmART Catalyst System
Refiner
Pilot Testing
Yes
Yes
Yes
Yes
Yes
Startup Date
4Q04
4Q04
CDX
CDY
CDX
CDY
CDX
CDY
CDX
CDY
CDX
CDY
40
10-20
30
Catalysts
Feedstock:
% cracked stock
Sp Gravity
8
Yes
Yes
Yes
1Q05
2006
2006
CDX
CDX
CDY
CDX
CDY
10
100
4Q04 2Q05 2Q04 3Q04
Yes
Yes
Yes
2006
2006 2006
2006
CDX
CDY
CDX
CDX
CDY
CDX
CDY
CDX
36
60
0.8628 0.8514 0.8468 0.8665 0.8735 0.8534 0.8448 0.8956 0.8649 0.8198 0.8676 0.8933 0.8328
Sulfur, wt.%
1.66
0.99
0.07
1.26
0.34
1.26
1.00
0.90
0.86
0.82
0.87
0.89
0.83
EP D86 Dist, C
371
377
335
396
354
380
360
360
285
545
346
363
349
Product:
Sulfur, ppm
10
10
35
50
Conditions
LHSV
0.7
1.5
2.8
0.7
1.1
1.6
1.9
1.0
1.1
2.6
1.3
1.0
1.6
Inlet P, BARG
54
64
48
56
56
51
44
88
56
41
73
88
50
13
catalysts must be used in order to take full advantage

of the SmART System concept. ART has devoted significant resources to designing the best ULSD catalysts for use in the SmART Catalyst System, and this
effort has lead to the commercialization in 2004 of a
new CoMo catalyst CDX, and a new NiMo catalyst CDY.
These new technologies benefit from the optimization
of the alumina chemistry to give the right surface area
and pore size distribution, as well as providing the right
surface chemistry (i.e., acidity).
ULSD production with the CDX/CDY SmART System
began early in 2004, in a North American refinery processing a feed containing 20-40% of a high endpoint
LCO. Since that time, ART CDX and ART CDY have
been selected for 14 different diesel applications either
as components in SmART Systems or as a complete
charge. A list of commercial applications of CDX, CDY
and SmART Catalyst Systems is shown in Table VII. It
is apparent from the table that the SmART catalyst
technology has been successfully applied to a wide
range of feeds and conditions. In most cases it has
been selected based on the high activity demonstrated in pilot plant testing.
As these and other recent successes demonstrate,
Advanced Refining Technologies has developed stateof-the-art technologies aimed at helping refiners meet
the challenges of clean fuels. These successes are
the result of harnessing the unique heritage of ART
which includes a collective expertise in material sci-
ence, catalyst formulation, and manufacturing knowhow. The science of designing specific catalyst components to operate in an optimum system is a fundamental part of ART's catalyst technology. With the
advent of clean fuels, ART seized the opportunity to
extrapolate their catalyst system expertise from resid to
lighter feedstocks. As a result, ART has been able to
deliver high performance technologies ranging from
StART Catalyst Systems for Si tolerance in coker naphtha applications, SmART Catalyst Systems for ULSD,
and ApART Catalyst Systems for cat feed hydrotreating.
ART strives to continuously improve the performance of
its catalysts, and the current focus of that effort is on
maximizing the effectiveness of the catalytic metals.
As described in Catalagram No. 96 (October 2004),
ART's new CoMo catalyst, CDX, benefits from
improved metals utilization through the application of
novel metals chemistry and a unique impregnation
technique. These same techniques have been successfully applied to a new NiMo catalyst called ART
NDXi. As shown in Figure 26, this catalyst has significantly higher activity for ULSD than ART CDY and the
conventional reference NiMo catalyst.
ART NDXi will be commercialized in early 2005 and is
slated to become the latest NiMo catalyst component
of the SmART Catalyst System.
Figure 29
ART NDXi has Step Out Activity
No Need to Trade
ULSD Catalyst Performance
for Hydrogen Limits:
SmART Approaches
Charles Olsen
Manager
Gerianne DAngelo
Senior Technical Services Engineer
ADVANCED REFINING
TECHNOLOGIES
Chicago, IL USA
t has been widely discussed that

desulfurization of dibenzothiophene and substituted dibenzothiophenes occurs through two
reaction pathways, the direct sulfur
abstraction route and the hydrogenation-abstraction route.
The former
involves adsorption of the molecule on
the catalyst surface via the sulfur atom
followed by C-S bond scission. This is
the primary pathway over cobaltmolybdenum
(CoMo)
based
hydrotreating catalysts. The second
pathway involves saturation of one

aromatic ring of the dibenzothiophene species followed by the
extraction of the sulfur atom.
Nickel-molybdenum (NiMo) catalysts have a higher activity for desulfurization via this route.
It has become a fairly common
practice to model ULSD applications by lumping the various sulfur
species into easy sulfur and hard
sulfur. The so-called easy sulfur
15
Figure 9
Optimizing HDS and Hydrogen Consumption
1.15
is this ability to balance HDS activity and H2 consumption to meet individual refiner requirements that sets
SmART apart.
Product Sulfur
H2 Consumption
Product Sulfur, ppm
Relative H2 Consumption
1.00
All CoMo
Reference
SmART Systems
is made up of compounds which

are readily desulfurized via direct
abstraction and boil below about
363CF whereas hard sulfur is
made up of compounds which are
more readily removed via hydrogenation followed by abstraction.
These
compounds
include4,6
dimethyl-dibenzothiophene
and
other di- and tri-substituted dibenzothiophenes. The relative amounts
of easy and hard sulfur in a feed are
a critical property to consider since
the concentration of each can vary
significantly from feed to feed
depending on crude source, boiling
range and the prior thermal or catalytic treatment of the feedstock.
ART first introduced the SmART
Catalyst System in 2001 (1,2) in
anticipation of the stringent new
demands required for ULSD. The
SmART system concept is based on
a staged catalyst approach
designed to exploit the fact that
there are two reaction pathways for
desulfurization. The system utilizes
a high activity CoMo catalyst like
ART CDXi for efficient removal of
sulfur via the direct abstraction
route and a high activity NiMo catalyst like ART NDXi which effectively
removes the multisubstituted dibenzothiophenes via the hydrogenation
route.
All NiMo
Reference
A number of factors need to be considered when designing the optimum

SmART system for a given application.
These have been reviewed in detail
previously (3, 4), and include feed endpoint (amount of hard sulfur), feed
nitrogen and H2 availability.
One of the big advantages of the
SmART System is illustrated in Figure
9. The figure shows that as NiMo catalyst is added to the SmART system
there is a large increase in HDS activity relative to the all CoMo reference,
and eventually, a maximum in HDS
activity is reached. The position and
magnitude of this maximum varies
with feed and operating conditions,
especially H2 partial pressure. The
figure also includes the relative H2
consumption, and again, as the percentage of the NiMo component
increases, the H2 consumption relative
to the base CoMo system increases.
Notice, however, that in this case the
relationship between H2 consumption
and the fraction of NiMo catalyst is
nonlinear. In the region where the system shows the highest activity the
hydrogen consumption is only slightly
greater than that for the all CoMo system, and well below that for the all
NiMo catalyst. Again, the nature of
this relationship varies with feed and
operating conditions, with a strong
correlation to hydrogen availability. It
The balancing of hydrogen consumption and HDS activity is possible because of the different kinetics
and reaction pathways for sulfur
removal and aromatic hydrogenation. The reaction pathways for sulfur removal were discussed above
with the main point being the hydrogenation pathway is critical for
ULSD production. The multisubstituted dibenzothiophenes (hard sulfur) are polynuclear aromatic
species containing two aromatic
rings, one of which must be hydrogenated for efficient removal of the
sulfur atom. Thus, the catalyst system needs good hydrogenation
activity and selectivity in order to
minimize the use of hydrogen.
To more fully understand how this
works, it is useful to review polynuclear aromatic hydrogenation in
general. To begin with, hydrogenation of aromatics is reversible, and
equilibrium conversion is less than
100% under practical hydrotreating
conditions. The equilibrium conversion decreases with increasing temperature, and therefore, increasing
temperature to get higher hydrogenation rates may ultimately result
in lower conversion. These reactions are also exothermic which can
have an impact on conversion in
adiabatic systems.
The hydrogenation of a number of
poly aromatic species such as
naphthalene and biphenyl have
been studied by a number of investigators (5) and the work has lead to
the reaction networks presented in
Figure 10. In the case of naphthalene, the reaction begins with the
hydrogenation of one of the aromatic rings to form tetralin, a mono ring
aromatic. The next reaction is hydrogenation of the remaining aromatic
ring to produce decalin, the fully
saturated species. As indicated in
the figure, the reactions proceed
sequentially with the rate of hydro-
genation of the final aromatic ring

(tetralin) at least an order of magnitude lower than saturation of the
first aromatic ring (naphthalene).
Interestingly, the rate of tetralin
hydrogenation is about the same as
that observed for benzene hydrogenation. A variety of substituted
naphthalenes have also been
shown to follow a similar reaction
network with the rate of hydrogenation of the first aromatic ring roughly the same as that observed for
naphthalene.
The hydrogenation of biphenyl proceeds similarly. The hydrogenation
is a stepwise reaction with the rate
of hydrogenation of the first aromatic ring about an order of magnitude
faster than that of the mono ring
compound. An important difference is that the rate of the first
hydrogenation reaction in naphthalene is about and an order of magnitude faster than the rate of hydrogenation of the first ring in biphenyl.
Thus, there is a significant difference between the hydrogenation of
an aromatic with two fused rings
compared to a two ring aromatic
where the rings are not fused. This
is an important difference when
considering hard sulfur removal
since these species do not contain
two fused aromatic rings and as
such, can be expected to behave
more like a biphenyl species.
Figure 10
Reaction Pathways for Poly Aromatics Hydrogenation
naphthalene
k1
k2
k3
bicyclohexyl
k4
k1>k2, k3>k4, and k1>k3
The challenge when designing a

SmART system is to provide enough
hydrogenation activity to efficiently
saturate the first ring on the two ring
aromatic (biphenyl type, sulfur containing) molecule, but not so much as
to catalyze the final hydrogenation
step in the reaction pathway discussed above.
The poly aromatic hydrogenation
reaction networks shown in Figure 10
can be modeled as first order
reversible reactions in series. Figure
11 shows the species concentration
profiles as a function of residence
time for a hydrogenation reaction
sequence such as that for naphthalene. In the example shown, the rate
1.0
Concentration
decalin
cyclohexylbenzene
biphenyl
Figure 11
First Order Reversible Reactions
in Series: Concentration Profile
0.8
tetralin
mono
di
Satd
k1=10*k2
of the first hydrogenation reaction in

the series is an order of magnitude
faster then the rate of the second
hydrogenation reaction. There is a
rapid decrease in the concentration
of the two ringed aromatic at short
times, and a corresponding
increase in the mono ringed
species. As time increases however, the mono ring aromatic concentration begins to decrease and the
fully saturated species begin to
build up. This type of concentration
profile suggests that there is a window of residence times corresponding to a maximum in the mono
ringed aromatic concentration. The
way to minimize hydrogen consumption is by avoiding saturation
of the mono ringed compounds.
This can be accomplished by
adjusting the residence time in the
CoMo and NiMo beds of a SmART
system through varying the
amounts of each catalyst. This
effectively tunes the hydrogenation
activity of the overall catalyst bed to
maximize hydrogenation of the poly
aromatics and minimize hydrogenation of the mono ringed species.
This is explored further on a diesel
feed comparing a CoMo and NiMo
catalyst in a pilot plant test. Figure
12 compares the HDS activity of the
two catalysts as a function of residence time at constant reactor temperature. It is apparent that the
NiMo catalyst is more active than
the CoMo catalyst at the indicated
0.6
0.4
0.2
0.0
Contact time
17
Figure 12
HDS Activity Comparison at High Pressure
50
CoMo
Product Sulfur, ppm
40
NiMo
30
20
10
0
0.00
0.20
0.80
0.60
0.40
1.00
1.20
1/LHSV, hrs
Figure 13
Aromatics Concentration Profile
for CoMo Catalyst at High Pressure
Concentration, %
40
30
Mono's
Sat'd
Poly's
20
10
0.00
0.20
0.40
0.60
0.80
1.00
1.20
1/LHSV, hrs
Figure 14
Aromatics Concentration Profile
for NiMo Catalyst at High Pressure
Concentration, %
40
Mono's
Sat'd
Poly's
30
20
10
0.00
0.20
0.40
0.60
0.80
1/LHSV, hrs
1.00
1.20
Figure 15
Cetane Index Increase with H2 Consumption
6.5
Cetane Index Increase
6.0
5.5
CoMo
5.0
NiMo
4.5
4.0
3.5
3.0
2.5
2.0
0.5
1.5
1.0
2.0
2.0
3.5
3.0
Relative H2 Consumption
Product Sulfur, ppm
Figure 16
HDS Activity Comparison at Moderate Pressure
100
90
80
70
60
50
40
30
20
10
0
CoMo
At the longest residence time (lowest LHSV), the NiMo catalyst has
about 15 numbers (absolute) higher
saturate concentration (and about
15 numbers lower mono ringed aromatic concentration) than the CoMo
catalyst. That translates to about 53
Nm3/m3 higher hydrogen consumption compared to the CoMo catalyst.
This difference decreases rapidly
as the residence time decreases.
NiMo
0.00
0.20
0.40
0.80
0.60
1.00
1.20
1/LHSV, hrs
Figure 17
Aromatics Concentration Profile for CoMo Catalyst
at Moderate Pressure
Concentration, %
40
30
Mono's
Poly's
Sat'd
20
10
0.00
0.20
0.40
0.60
0.80
1/LHSV, hrs
1.00
conditions, although both catalysts

are able to achieve <10 ppm product sulfur if the residence time is
long enough. Figures 13 and 14
compare the concentration profiles
for poly aromatics, mono ringed aromatics and the saturates as a function of residence time for the CoMo
and NiMo catalyst respectively.
Both charts show the same concentration trends as shown in Figure 11.
There is a rapid decrease in poly
aromatics concentration and a corresponding increase in mono
ringed aromatics for both catalysts.
Clearly, however, the NiMo catalyst
is much more efficient at hydrogenating the final aromatic ring as
evidenced by a lower maximum
mono ringed aromatic concentration and a faster increase in saturates with increasing residence time
compared to the CoMo catalyst.
1.20
Of course, if there is sufficient H2

available, the incremental increase
in aromatics saturation and the correspondingly higher hydrogen consumption can offer some benefits
such as cetane improvement.
Figure 15 summarizes the cetane
index increase observed for this
example. The cetane index increases linearly with increasing H2 consumption, and ultimately, a cetane
index improvement of six numbers
was achieved by the high activity
NiMo catalyst.
Not surprisingly, pressure has a dramatic effect on the performance of
NiMo and CoMo catalysts in ULSD.
Figure 16 compares the HDS activity of the NiMo and CoMo catalyst
for the same feed and conditions as
above, but at lower pressure. At
19
Figure 18
Aromatics Concentration Profile for
NiMo Catalyst at Moderate Pressure
these conditions the CoMo catalyst

is more active at the higher LHSVs
(shorter residence times), while at
the lower LHSVs the catalysts have
essentially the same activity. A
comparison of the aromatic species
concentration profiles is shown in
Figures 17 and 18. In this case, the
difference between the catalysts is
much less than observed for the
higher pressure case, and the
degree of saturation achieved by
either catalyst is lower.
For the
CoMo catalyst in Figure 17, the concentration of Mono ringed aromatics never begins to decrease, and
the concentration of saturates
remains very low. The NiMo catalyst
data shown in Figure 18 looks similar, although at long residence time
there is a decrease in mono ringed
aromatics, the poly aromatic level is
lower and the saturates concentration is higher than for the CoMo catalyst case.
SmART Performance
Since introducing the SmART system for ULSD, ART has conducted a
significant amount pilot work to
demonstrate both the activity and
stability of these systems. Figure
19 summarizes some of that work
comparing ART CDXi and a SmART
system consisting of CDXi and
NDXi.
This particular study
Concentration, %
30
Mono's
Poly's
Sat'd
20
10
0.00
0.20
0.40
0.60
0.80
1.00
1.20
1/LHSV, hrs
388
Figure 19
Pilot Plant Aging of a SmART System
Normalized to 1.25 LHSV and 55 BARG
1.26 wt.% feed sulfur and 10 ppm product sulfur
Temperature, C
382
377
371
366
360
354
CDXi
349
SmART
343
0
250
500
750
1000 1250 1500 1750 2000
Hours on Stream
Figure 20
Pilot Plant Aging of a SmART System
Aromatics Hydrogenation
Product Aromatics, vol.%
These data are a good example of

the flexibility of the SmART system.
In applications where there is sufficient H2 partial pressure, a NiMo
catalyst is likely the most active system for HDS. However, it will consume significantly more hydrogen
due to its efficiency at catalyzing
hydrogenation reactions. If the
incremental hydrogen consumption
cannot be tolerated, a SmART system can be designed which will
deliver high HDS activity and minimize hydrogen consumption. In
cases where the hydrogen pressure
is lower, the SmART system is often
more active than either component
alone without increasing the H2 consumption significantly over the all
CoMo system.
40
30
28
26
24
22
1.25 LHSV and 55 BARG
20
18
16
14
12
10
CDXi
SmART
250
500
750
1000 1250 1500 1750 2000
Hours on Stream
Figure 21
Commercial SmART System Performance
spanned about 2000 hours at the

conditions shown in the chart processing a straight run feed from the
West Coast. The H2 partial pressure
is quite good in this example which
explains the large activity advantage for the SmART system relative
to the all CoMo system. Notice that
the slopes of the lines through each
data set are essentially the same
indicating the ART CDXi and a
SmART system of CDXi/NDXi have
the same relative stability when producing ULSD.
Figure 20 shows how the aromatic
saturation activity changed over the
course of the 2000 hour run. The
total aromatics saturation stayed
essentially constant for the SmART
system throughout the run, while for
CDXi there is a suggestion of some
slight loss in saturation activity at
the end of the run. Although not
shown in the chart, the PNA content
in the product was also analyzed,
and the data tell the same story.
Another thing to note is the similarity in aromatics (and PNA) conversion between the SmART system
and CDXi. This is an indication that
the hydrogen consumption is similar
for both systems, even though the
SmART system has a substantial
activity advantage.
The SmART catalyst system or its
components have been selected for
over 20 applications since being
introduced to the market. The technology has been successfully
applied to a wide range of feeds
427
Temperature, C
30-50% LCO blend with 360-393C end point
399
371
343
316
20
40
and conditions, and in most cases has

been chosen based on the high activity demonstrated in pilot plant testing,
both in ART labs and refiner testing
facilities. A few of these applications
are discussed in the paragraphs that
follow.
ULSD production with the SmART
System began early in 2004 in a North
American refinery processing a feed
containing 30-50% of a refractory,
high endpoint LCO. The LCO end
point can gat as high as 391C, and
sulfur speciation indicates that the
concentration of hard sulfur often
exceeds 4000 ppm. The performance
of the SmART system is shown in
Figure 12. Despite processing a feed
which was more difficult than anticipated, the SmART system provided
excellent performance. The typical
deactivation period following start up
was exacerbated by the fact that the
1.5
15
LHSV
0.75
0.75
0.75
H2/Oil, Nm3/m3
166
166
166
Inlet pressure, BARG
65
65
65
WABT, C
334
336
346
Product sulfur
80
100
120 140 160
180
Days on Stream
Table V
Commercial ULSD Performance for Refiner A
Months on stream
60
refiner had to run off a tank of LCO

which had built up during the unit
turnaround. Even when processing
a blend of 50-60% LCO at SOR, the
SmART system had no problems
making the <10 ppm sulfur target.
Once the feed and operating conditions lined out the operation was
very stable with essentially no deactivation apparent over the last several months.
In another example, Refiner A conducted in-house testing for their
ULSD catalyst selection. ART proposed a SmART catalyst system
which ultimately tested substantially
more active than other catalysts in
the program including one from the
perceived market leader. As a
result, this refiner selected a
CDX/CDY SmART system for their
ULSD unit. The commercial unit
began operation in the fourth quarter of 2004, and a snapshot of the
performance thus far is shown in
Table V. The feed is a blend of up to
40% LCO and light coker gas oil
with a sulfur content of about 1.3
wt.%. The unit has decent pressure, but the H2/Oil ratio is lower
than optimum especially given that
the H2 consumption is typically
around 98 Nm3/m3. The performance of the SmART system has met
the high expectations set by the
pilot plant testing.
21
Table VI
SmART System Performance for Refiner C
Reactor Train
Months on stream
LHSV
chart shows, the SmART system has

experienced essentially no deactivation since start up.
1.7
1.7
1.7
1.7
H2/Oil, Nm3/m3
247
247
247
247
Inlet pressure, BARG
43
43
43
43
WABT, C
346
347
352
345
Product sulfur, ppm
45
45
45
45
ART and Refiner B worked together

to generate data for a planned
ULSD unit revamp. The feedstock
had a very high endpoint of 396C
which corresponded to over 6000
ppm of hard sulfur. ART proposed
a SmART system consisting of its
new high activity catalysts ART CDX
and CDY.
This system was
designed to provide high HDS
activity and minimize hydrogen consumption in order to stay within the
refiners hydrogen availability limits.
The refiner conducted pilot plant
testing on a variety of proposed catalysts, and found that the SmART
system easily met the targets and
outperformed the other vendors in
the program.
As a result of this work, Refiner B

selected the SmART system for the
first cycle in the ULSD unit. The unit
started up in the second quarter of
2005, and has been online for about
nine months. The performance of the
SmART system is shown in Figure 22.
This refiner conducted a ULSD test
run early in the cycle, and the performance met all expectations and in
fact, closely matched the pilot plant
results. The unit was then operated to
produce 200 ppm sulfur for several
months before pushing again to make
ULSD. The hydrogen consumption
increased from about 55 Nm3/m3 to
62 Nm3/m3 when they went from 200
ppm sulfur to 10 ppm sulfur. As the
Temperature, C
0.72 LHSV and 43 BARG H2 Pressure, 1.24 wt.% feed sulfur
371
10 ppm sulfur
357
343
329
200 ppm sulfur
316
302
25
50
75
100
125
Days on Stream
150
175
References
1. Olsen, C., Krenzke, L.D., Watkins, B.,
AICHE Spring National Meeting, New
Orleans, March 2002.
2. Krenzke, D., Armstrong, M., 2001 ERTC
Meeting, Madrid, Spain.
3. DAngelo, G., Olsen, C., Davison
Catalagram 95, March, 2004, pp. 38-44
4. Olsen, C., Krenzke, L.D. 2005 NPRA
Annual Meeting, Paper AM-05-17
5. Girgis, M.J., Gates, B.C., Ind. Eng.
Chem. Res., 30, 1991, p 2021.
Figure 22
SmART System Performance for Refiner B
385
In the final example, Refiner C wanted to produce <50 ppm sulfur

diesel in an existing reactor. ART
optimized a SmART loading based
on the product sulfur requirement
while at the same time staying within the constraints of the refinerys
hydrogen system. The unit started
up with the SmART system in the
second quarter of 2005, and has
been producing <50 ppm sulfur
diesel for about 9 months. A summary of the performance is shown
in Table VI. The performance of the
SmART system has been excellent
even with the high LHSV and low
hydrogen partial pressure. The
required temperature is within a few
degrees of the predicted performance, and the catalyst deactivation
observed to date is minimal.
Hydrogen consumption has also
been within expectations at 45-48
Nm3/m3. Refiner C is extremely
pleased, and is now working with
ART on a unit revamp which will
allow production of <10 ppm sulfur
diesel.
200
ART Excels In ULSD Service: Update

on Sulfur minimization by ART
Greg Rosinski
Technical Service Engineer
Dave Krenzke
Regional Technical Services
Manager
Charles Olsen
Worldwide Technical Services
Manager
ADVANCED REFINING
TECHNOLOGIES
Chicago, IL USA
RT first introduced the SmART

Catalyst System Series for
ultra-low sulfur diesel (ULSD)
in 2001. Since that time the technology has been widely accepted by the
refining industry as top tier for ULSD.
As detailed previously (Catalagram
99, 2006), the SmART System is a
staged catalyst system customized to
meet individual refiners objectives
with the performance of the system
driven by ARTs DXTM Catalyst
Platform.
ULSD production with the first SmART

System began early in 2004 at a North
American refinery processing a feed
containing 40% of a high endpoint
Light Cycle Oil (LCO). Since that time
DX Catalyst Platform has been selected for over 35 ULSD applications as
either stand-alone catalysts or as
components in SmART System. The

technology has been a great success since its introduction with millions of pounds installed in commercial units around the world.
Several of the first refiners to utilize
a SmART System are still enjoying
benefits today, or have reloaded
another SmART System based on
the exceptional first cycle they
received from the technology.
This article contains several case
studies highlighting the performance of ART catalysts in a variety of
ULSD units around the world.
These are summarized in Table I.
Refiner A is an initial user of a
SmART System in Asia Pacific. This
refiner conducted in-house testing
for their ULSD catalyst selection.
reprinted from Catalagram 104 SE 2008
23
Table I
Summary of ULSD Case Studies
A
B
C
D
E
F
G
H
Region
AP
AP
AP
AP
NA
NA
NA
NA
Feed
40% cracked stock
Straight Run
Straight Run
Straight Run
40% coker/LCO
50% LCO
70% LCO
45% LCO/LCGO
Inlet Pressure BARG

64
53
59
78
74
N/A
86
90
LHSV
0.7
0.7
1.2
1.1
1.3
1.1
1.0
0.8
% CoMo/ % NiMo
70/30
90/10
55/45
30/70
35/65
70/30 + Dewax
35/65
25/75
Figure 1
Asia Pacific Refiner A
382
5 ppm product sulfur
WABT, C
360
338
316
Actual
Normalized
293
271
0
200
400
600
800
1000
Days on Stream
Figure 2
Asia Pacific Refiner B
427
413
High severity mode

Product Sulfur: 85% <8ppm
WABT, C
399
385
Low severity mode

Product Sulfur: 15% <8ppm
371
357
343
Return to low severity mode:

WABT dropped about 8C
329
316
50
100 150 200 250 300 350 400 450 500 550 600
Days on Stream
Actual
Normalized to 10 ppm
They requested catalyst samples

from ART and a market leading
domestic supplier.
The DXTM
Catalyst Platform tested as a substantially more active catalyst than
the others in the program. As a
result, ART was selected as the catalyst supplier for their unit, which
started up in the last half of 2004.
This refiner was completing a
revamp, which involved the addition
of a new reactor in front of an existing reactor. They decided to use
fresh catalyst in the new lead reactor while keeping used ART catalyst
in the lag reactor. The unit conditions are characterized by an inlet
pressure of 64 BARG and a LHSV of
0.7 hr-1. The feed contains up to
40% cracked stocks, and the product sulfur has averaged 5 ppm. The
unit ran for three years before the
refiner decided to change out the
used catalyst in the lag reactor
with fresh catalyst from ART. The
performance of the unit is summarized in Figure 1 and, as the figure
depicts, stability has been exceptional.
Figure 2 shows data for Refiner B.
This was another major Asia Pacific
refiner using a SmART System. This
refiner needed to produce 10 ppm
sulfur diesel for two years in a newly
revamped unit. The operating conditions for this unit are 53 BARG inlet
pressure with an LHSV around 0.7
hr-1. The feed was a high end point,
straight run diesel with typical sulfur
content of 1.25 wt.%. This refiner
was concerned with minimizing
hydrogen consumption as hydrogen
availability was limited at the refinery. This was a situation well suited
for the flexibility of the SmART
System to tune the hydrogen consumption/activity relationship. The
optimum loading for this unit was
90% cobalt-molybdenum (CoMo)
and
10%
nickel-molybdenum
(NiMo) catalyst. This design was
expected to be significantly more
active than an all CoMo loading and
would consume the same amount of
hydrogen as an all CoMo loading.
The final selection of the SmART
System was based on competitive
pilot plant testing by the refiner.
Figure 3
Asia Pacific Refiner C
382
371
WABT, C
About eight months into the ULSD

portion of the cycle, the refiner significantly increased the operating
severity and began producing very
low sulfur diesel. Typically, 8 ppm is
the recommended product sulfur
target when producing <10 ppm
ULSD. In the higher severity mode,
the product sulfur was well below 8
ppm 85% of the time, whereas in
the lower severity mode the product
sulfur was below 8 ppm only 15% of
the time. The higher severity led to
a much higher deactivation rate and
potentially jeopardized the two year
cycle length. After discussions with
the refiner, the severity was reduced
and the two year cycle length was
achieved.
360
349
338
327
316
50
150
100
Actual
ART was chosen for the second

cycle which again was based on
competitive testing and the outstanding performance demonstrated in the first cycle. The current system is performing very well.
demanding operation than those discussed previously, and a required system designed for maximum activity.
The catalyst loading in this case was
55% CoMo and 45% NiMo. The performance is summarized in Figure 3.
Refiner C, another Asia Pacific refiner, selected ART catalysts for their
ULSD unit based on the strong reference from the refiners mentioned
above. This refiner needed to produce 8 ppm sulfur diesel for two
years and hydrogen availability was
not a constraint. The operating conditions included an inlet pressure of
59 BARG and LHSV of 1.2 hr-1. The
feed was also a high endpoint,
straight run diesel with a sulfur content of 1.8 wt.%. This was a more
This unit met all performance targets

during the two year cycle and has just
been reloaded with a new SmART
System.
Operating data for Refiner D is shown
in Figure 4. This is a grass roots ULSD
unit in Asia Pacific. ART was chosen
to participate in this project because
of the excellent performance of the
SmART System from previous references in the region. ART worked
closely with the refiner and engineer-
Figure 4
Asia Pacific Refiner D
382
WABT, C
371
360
349
338
327
316
0
25
50
75 100 125 150 175 200 225 250 275 300 325 350 375
Days on Stream
Actual
200
250
300
Days on Stream
Normalized to 8 ppm
Normalized to 8 ppm
ing construction firm on the design

basis for this project. A 30% CoMo
and 70% NiMo SmART System was
designed for this unit in order to
deliver maximum activity. The operating conditions for this unit included an inlet pressure of 78 BARG
and LHSV of 1.1 hr-1. This unit also
processes a straight run feed which
has a relatively high level of nitrogen
and an endpoint of 388C (by D86).
The unit is running well and meeting
all performance expectations with a
somewhat lower than estimated
deactivation rate, and is well on
track to meet the targeted threeyear cycle length.
Refiner E is another grassroots
ULSD unit, which started up in the
second quarter of 2006 in North
America. ART worked with the
licensor on the unit design and proposed a SmART System loading
consisting of about 65% NiMo. This
loading was designed for maximum
activity as the design feed contained a high percentage of
cracked stocks. The conditions of
the unit included an inlet pressure of
74 BARG with an LHSV of 1.25 hr-1.
This refinery processes mostly
sweet crudes, and the feed to the
unit typically contains about 40%
FCC LCO and coker LGO. The
activity of the system has been
extremely high, allowing this refiner
to operate at 20% over design
charge rate. Figure 5 shows the
25
Figure 6 summarizes data from

Refiner F in North America. This
refiner had completed a project to
revamp an existing hydrocracker to
ULSD service. The objectives were
to increase feed capacity from 4770
to 7155 MTD, while ensuring the
capability to process 50% or more
LCO, as well as provide for cold flow
improvement during the winter
months. This refiner also wanted to
minimize the hydrogen consumption
so that feed rate could be maximized within make-up hydrogen
constraints.
Along with Sd
Chemie, ART designed a dewaxing/ULSD catalyst system meeting
the unit objectives. The unit started
up successfully in the 2nd quarter
of 2006 and processes both sweet
and sour crude derived feeds in
block operation. The feed is also
comprised of 40-50 vol.% of a high
endpoint LCO. The performance of
the unit is summarized in Figure 6.
The unit came on stream with higher than expected activity, and is well
on its way to exceeding the target
three year cycle length. Additional
details on this unit can be found in
Catalagram 103, Spring 2008.
Refiner G is another grassroots unit
which started up in the fourth quar-
50
45
40
35
WABT, C
371
360
30
25
20
15
349
338
327
10
5
0
316
304
0
200
100
400
300
500
600
800
Days on Stream
WABT
Kerosene
Figure 6
North American Refiner F
382
371
360
349
338
327
316
304
0
100
200
300
400
500
600
Days on Stream
Figure 7
North American Refiner G
393
5 ppm average product sulfur
382
371
360
349
338
327
316
304
0
100
200
300
400
Days on Stream
700
Kerosene, vol.%
382
Normalized WABT, C
This unit also typically processes

10% kerosene with occasional
increases to as much as 30%. As
can be seen in Figure 5, the addition of kerosene to the feed has no
negative impact on the activity of
the system, and may even improve
the performance. Notice how the
WABT decreases when processing
higher amounts of kerosene. This
suggests that the increase in feed
vaporization (decrease in H2 pressure) is offset by the decrease in
hard to treat, substituted dibenzothiophene sulfur species.
Figure 5
North American Refiner E
Normalized WABT, C
performance for the cycle thus far,

and it is evident from the figure that
the catalyst stability has been excellent. In fact, the unit was designed
for a 24-month cycle, but 48 months
appears possible even at higher
than design feed rates.
500
600
97
100
90
95
90
83
85
Recycle H2 Purity, vol.%
80
76
75
70
69
65
60
62
0
100
200
400
300
500
600
Days on Stream
ter of 2006. The performance of

this unit is summarized in Figure 7.
ART again worked closely with the
licensor on the new unit design and
proposed a SmART System loading
using 65% NiMo catalyst for maximum activity. The unit typically
processes >70% LCO, including
both LCO produced within the refinery and purchased externally. The
initial deactivation rate of the unit
was significantly higher than
expected, and it was determined
that this was due to lower H2 partial
pressure than design combined
with lower H2/Oil ratio. This can be
seen in Figure 8 which shows the
trend of reactor inlet pressure and
the recycle hydrogen purity. Early in
the cycle there were control issues
and the inlet pressure showed a
steady decline. The recycle hydrogen purity also decreased steadily
and had periods where it fell dramatically. Once the H2 partial pressure concerns were addressed,
combined with better bed temperature management, the deactivation
rate decreased significantly. The
operation became much more consistent, and since that time the
deactivation rate indicates that the
unit will easily meet the expected
24-month cycle length.
Finally for Refiner H, Figure 9 summarizes the performance of a
Recycle H2
SmART System at a U.S. Gulf Coast

refinery. This was another grassroots
unit that started up in the Fall of 2006.
SOR activity met expectations, and
since that time the unit operating
severity has steadily increased. The
current feedrate is 22% over design
and the feedstock endpoint has
increased by 17-22C. Typical unit
conditions include 0.8 hr-1 LHSV and
90 BARG inlet pressure, and, on average, the unit processes a feed containing 15% FCC LCO, 30% LCGO
and 5% coker naphtha. At times the
unit has processed as much as 80%
cracked stocks in the feed, and still
the product sulfur has averaged less
than 6 ppm for the cycle. The target
cycle length was 24 months, and the
unit is currently on track for a 36-
As this sampling of case studies

demonstrates, the SmART System
has been employed in a wide variety of ULSD applications around
the world. The technology has been
successfully operating over a broad
range of operating conditions from
low to high pressure with feeds
ranging from straight run to 80%
cracked stocks. In each application, all expectations have been
met or exceeded, and in a number
of cases ART catalyst has been
selected for the second cycle
based on the excellent performance.
ADVANCED
REFINING
TECHNOLOGIES continues to
develop higher performance products for ULSD as evidenced by the
recent introduction of the newest
DXTM Catalyst Platform, 420DX. The
addition of this catalyst to the
ULSD portfolio is an example of
the commitment that ART will continue to deliver state-of-the-art
technology for ULSD.
Figure 9
North American Refiner H
20
18
16
14
12
10
8
6
4
2
0
371
360
349
WABT, C
Pressure
month cycle. Also shown in Figure

9 is the API uplift from the unit. From
the chart, the API typically increases 6-10 numbers depending on the
feedstock.
338
327
316
304
293
0
50
100
150
200
250
300
350
400
450
API Gravity Increase
Inlet Pressure, BARG
Figure 8
North American Refiner G Hydrogen Availability
500 550
Days on Stream
WABT
API Uplift
27
Editors Note:
A lot has changed since ART first introduced the SmART Catalyst System series for ULSD in 2001. Many ULSD units
have been built or retrofitted and are now into their second or third cycles. Most countries in the world are either mandating ULSD from their refineries, or in the process of doing so.
ART has made many improvements in its offerings for making ULSD as a result of extensive investment in the
research and development of new and improved catalysts for expanding performance and flexibility. Since the introduction of CDY and CDX, ART is already on its third generation of catalysts. The premier catalysts used in todays
SmART system are 420DX and NDXi.
The table below shows the commercial experience of the SmART System. There are many repeat users among the
more than 70 unit start-ups that have occurred using the SmART catalyst system. A wide variety of feedstocks and
operating conditions are represented from units all over the world.
SmART Catalyst System Users List

Start-up
Date
% Cracked
Stock
Product Sulfur
(ppm)
Pressure
BARG
% NiMo
Catalyst
Mid Continent, USA
2004
48
35%
Japan
2004
10-20
65
55%
Japan
2004
40
55
25%
Japan
2004
35
52
0%
Mid Continent, USA
2004
20
10
57
25%
Mid Continent, USA
2005
15
52
25%
Japan
2005
Japan
2005
35
88
100%
Korea
2005
10
57
10%
Russia
2005
10
50
45
30%
South Africa
2006
15
400
53
15%
West Coast
2006
Mid Continent, USA
2006
51
0%
Eastern Canada
2006
35-50
145
30%
South America
2006
25
87
100%
Korea
2006
59
50%
Russia
2006
50
48
25%
Mid Continent, USA
2006
36
74
0%
Mid Continent, USA
2006
60
89
67%
Gulf Coast, USA
2006
100
89
67%
Gulf Coast, USA
2006
57
67%
Gulf Coast, USA
2006
41
0%
Western Canada
2007
15
61
30%
Country/Region
100%
100%
Start-up
Date
% Cracked
Stock
Product Sulfur
(ppm)
Pressure
BARG
% NiMo
Catalyst
Gulf Coast
2007
50
350
69
50%
United Kingdom
2007
32
54
50v
Mid Continent, USA
2007
60-100
110
70%
Taiwan
2007
450
42
0%
Korea
2007
10
79
70%
Japan
2007
35
88
100%
South America
2007
25
54
100%
South America
2007
25
54
100%
Korea
2007
57
70%
Korea
2007
53
10%
Korea
2007
65
10%
Singapore
2007
66
0%
Russia
2007
50
25%
Russia
2007
15
50
32
0%
Russia
2007
10
50
45
45%
Thailand
2008
45
60
70%
Thailand
2008
45
60
70%
Russia
2008
10
50
45
45%
South Africa
2008
15
400
53
15%
Gulf Coast, USA
2008
60
15
69
50%
India
2008
10
50
51
65%
India
2008
50
51
0%
Thailand
2008
50
42
0%
Korea
2008
59
10%
Korea
2008
52
50%
Gulf Coast, USA
2008
57
0%
Russia
2009
50
39
0%
Australia
2009
40
50
33%
United Kingdom
2009
30-40
40
0%
Gulf Coast, USA
2009
74
100%
Chile
2009
40
10
97
100%
Gulf Coast, USA
2009
60
66
20%
Japan
2009
Poland
2009
Singapore
2009
10
3000
23
0%
Singapore
2009
30
66
70%
East Coast, USA
2009
20
27
0%
Russia
2009
10
50
45
30%
Taiwan
2009
40
450
41
0%
Taiwan
2009
35
52
100%
Mid Continent, USA
2010
52
50%
Country/Region
0%
10
60%
29
Distillate Pool Maximization by

Exploiting the Use of Opportunity
Feedstocks Such as LCO and Syncrude
Brian Watkins
Hydrotreating Technical Services
Engineer
Charles Olsen
Manager
ADVANCED REFINING
TECHNOLOGIES
Chicago, IL USA
s the distillate market demand

has increased over the last few
years, the production of ultra
low sulfur diesel (ULSD) has prompted refiners to look for ways to maximize their diesel pool. One way to
accomplish this has been to increase
the use of opportunity feedstocks
such as additional LCO, diesel
streams from other hydroprocessing
units, and feeds from various synthetic crudes. Some of these opportunity
feedstocks, having already been
processed through conventional refinery processes, may pose unexpected
challenges to refiners wishing to incorporate them into the distillate pool.
Some of these streams have proven to
be significantly more difficult to
process, underscoring the fact that it
is important to understand the poten-
tial impact of processing new feed

streams in order to avoid unpleasant surprises. This paper highlights
a few examples demonstrating significant differences in feed reactivity
for a variety of different feed components which are not necessarily
anticipated from the usual bulk feed
analyses.
FCC LCO and coker diesels have
long been used as feed components combined with a straight run
(SR) feed source to produce ULSD
products. The quality of the LCO
varies with distillation range, and
depends on the severity of the pretreatment of the FCC feed as well as
on the conditions in the FCC and
the FCC catalyst employed. A common element in LCO is a very high
Table I
Diesel Feedstock Analysis
Light SR
Gas Oil
1.11
138
0.8635
LCO
0.17
203
0.9094
EB Diesel
0.017
135
0.8695
Synthetic
Diesel
0.07
261
0.8657
FB Diesel
0.006
71
0.8685
24.18
14.47
9.71
64.61
31.97
32.64
41.26
36.55
4.71
36.58
32.52
4.06
44.83
41.95
2.88
Thiophenes
141
256
316
388
426
8
104
186
243
303
340
0
151
208
309
375
404
0
97
176
269
350
392
0
141
193
268
344
413
0
Benzothiophenes (BT)
Substituted BTs
2
2793
69
1083
0
0
0
26
0
0
Di Benzothiophene (DBT)
C2-DBTs
4,6 DiMethyl DBT
C3-DBTs
222
3453
199
4410
110
436
0
0
0
72
78
17
7
266
29
370
0
15
43
24
Sulfur, wt.%
Nitrogen, wppm
Specific Gravity
Aromatics, vol.%
Total
Mono
Poly
Distillation, D2887, C
0.5
10
50
90
99.5
concentration of polynuclear aromatic compounds relative to other

feeds.
Synthetic diesel material is often initially processed by either a coker or
ebullating bed residue hydrocracking unit, and then processed
through
a
hydrotreater
or
hydrotreater/hydrocracker combination. These hydrocracking units
tend to operate at severe conditions
in conjunction with high hydrogen
partial pressures. At these conditions, the removal of all the easy,
less refractory sulfur is readily
achieved, and the majority of the
multi-ring aromatics are saturated.
This leaves a product which is relatively low in sulfur and PNAs and,
when added to the feed to a ULSD
unit, gives rise to a surprisingly difficult feedstock to process.
Likewise, the use of diesel range
products from an H-Oil, LC-FININGTM unit or fixed bed resid desulfurizer can also have a significant
impact on downstream diesel catalyst

activity for similar reasons. The general properties of these types of diesel
feeds often indicate that they may be
relatively easy to hydrotreat due to
their low sulfur content and API gravity which is often similar to SR materials. Table I lists the properties for several diesel feeds including the diesel
product fractions from an ebullating
bed resid (EB) unit, a fixed bed resid
(FB) unit, and a diesel fraction from a
Canadian synthetic crude.
ART conducted pilot plant testing to
investigate the impact of various
diesel feed components on catalyst
activity. The pilot work utilized the SR
diesel shown in Table I as the base
case feed. The other components
shown in Table I were blended into the
base feed at 20% by volume to show
the effects on catalyst performance.
The pilot plant work involved several
tailored catalyst systems as well as
changes to operating pressure and
hydrogen rate in order to cover a
broad range of operation.
The base case testing was done at

a hydrogen pressure of about 48
BAR, a LHSV of 0.7 and 232
Nm3/m3 hydrogen/oil ratio. The catalyst system was a stacked system
of high activity CoMo and NiMo catalysts containing >80% CoMo catalyst. This system was chosen due
to limited hydrogen availability and
a desire to minimize hydrogen consumption Additional information on
the theory, design and use of this
type of staged catalyst loading can
be found in references 1-5.
Table II shows the analysis of the
different feed blends. The 20% LCO
has 1600 ppm lower sulfur, a one
number lower API, and 20 ppm
higher nitrogen content compared
to the SR feed. The total aromatic
content in the blend is also higher
by 10 volume percent absolute.
Compare this to the feed blends
containing the EB diesel, FB diesel,
or the synthetic diesel where all
three have even lower sulfur com-
31
Table II
Blended Diesel Feedstock Analysis
Light SR
Gas Oil 20% LCO
20%
EB Diesel
20% Syncrude
20%
FB Diesel
Sulfur, wt.%
1.11
0.95
0.88
0.92
0.92
Nitrogen, wppm
138
158
144
179
131
Specific Gravity
0.8635
0.8668
0.8593
0.8589
0.8592
Total
24.18
34.58
30.1
29.09
30.12
Mono
14.47
19.7
21.2
20.33
21.94
Poly
9.71
14.9
8.8
8.76
8.18
0.5
141
125
142
129
150
10
256
220
243
228
238
50
316
301
314
308
311
90
388
379
383
381
384
99.5
426
424
423
424
427
12
10
Aromatics, vol.%
Distillation, D2887, C
Thiophenes
Benzothiophenes (BT)
Substituted BTs
13
2793
2279
2161
2105
1861
Di Benzothiophene (DBT)
222
206
35
148
133
C2-DBTs
3453
2497
2324
2809
2464
4,6 DM-DBT
199
119
116
148
130
C3-DBTs
4410
3598
3508
4061
3583
pared to the SR, and slightly higher

API gravities, despite having a higher total aromatic content. Other
changes to note are the fact the
feed nitrogen content stays fairly
constant, and the mono aromatic
content is higher and PNA content
lower for these blends compared to
the SR feed.
Required Temperature Increase, C
Figure 1 summarizes some of the

pilot plant data comparing the SR
and LCO feed blends. It shows that
the SR diesel requires a 23C
increase in temperature to go from
about 100 ppm sulfur down to 10
ppm sulfur at base LHSV and pressure. The LCO blend requires
almost 11C higher temperature to
achieve the same product sulfur relative to the SR feed. The product
from the LCO blend has a two to
three number lower API compared
to the SR product, and hydrogen
consumption increases significantly
for the LCO blend due to saturation
of additional polyaromatic compounds found in the LCO. These lat-
Figure 1
Activity Comparison on SR and Blended SR/LCO
39
33
LCO
28
SR
22
17
11
6
0
0
20
40
60
80
Product Sulfur, ppm
100
120
Figure 2
Comparison of Product Gravity for the SR and LCO Blend
0.832
0.835
LCO
Results indicate that both feeds

have similar API number upgrades
(i.e. product feed) as the reactor
increases in temperature; however,
the actual product API is different at
equal product sulfur. Even though
the LCO blended feedstock requires
a higher temperature to achieve the
same product sulfur, the product
API is still about a full number lower
as shown in Figure 2.
The LCO also has the additional
issue of increasing hydrogen consumption when added to the ULSD
operation. Figure 3 compares the
aromatic saturation achieved on the
blended LCO feedstock as compared to the SR. The majority of the
aromatic saturation occurs with the
poly aromatic compounds and, as
shown in Table II, the LCO blend
contains significantly more PNAs
compared to the SR feed. The
hydrogen consumption is estimated
to be 22-27 Nm3/m3 higher for the
LCO blend at 10 ppm product sulfur.
The figure above shows that processing LCO is significantly more
difficult than processing the SR
feed. One option to gain back some
of the lost activity is to change the
SR
0.837
0.840
0.842
0.845
0.847
0.850
0.852
0.855
11
22
33
end point of the LCO in the feed. ART

completed pilot plant testing on an
LCO stock as received and the same
LCO with a 22C end point reduction
to simulate how this can affect catalyst
performance. Table III lists the major
component analysis of the two LCO
feeds. The decrease in endpoint lowers the total sulfur by almost 1000
ppm and total nitrogen decreases by
129 ppm.
The impact this degree of LCO endpoint reduction has on ULSD performance is over 17C difference in activity
which corresponds to additional life in
Increasing WABT, C
11
22
33
44
55
66
Increasing WABT, C
Figure 3
Comparison of Aromatic Saturation
44
55
66
the hydrotreater. A comparison of

the two LCO feeds blended at 30%
into SR feed is shown in Figure 4.
The addition of LCO has a major
impact on activity for both the low
and high endpoint LCO materials.
The required temperature increase
for ULSD in going from 0 to 30%
LCO for the lower endpoint material
is about 0.7C per percent LCO.
Processing the higher endpoint
LCO increases the required temperature to about 0.8C per percent
LCO. Figure 5 demonstrates this
more clearly in the form of a plot of
the required temperature increase
as a function of LCO content.
Notice from the chart that the activity effects are not exactly linear with
increasing LCO content. The first
15% LCO has a larger impact on
activity than the next 15%.
1
0
-1
LCO
-2
SR
-3
-4
-5
-6
-7
-8
-9
Change in Total Aromatics, vol.%
Specific Gravity
ter consequences set limits on the

amount of LCO which can be
processed and still meet product
cetane specifications and hydrogen
availability constraints.
The diesel products from an EB unit,

a FB unit and the synthetic crude
diesel provide very different sulfur
distribution patterns compared to
the SR feed and LCO shown in
Table I. Almost all of the sulfur
species in those feeds are multisubstituted dibenzothiophenes, the
so-called hard sulfur species. The
species groupings from sulfur speciation using a GC-AED technique,
however, indicate little about what
the actual molecular structure is
33
Table III
Impact of End Point Reduction on FCC LCO
Type
LCO
(Low FBP)
LCO
(High FBP)
0.945
0.948
708
66.86
22.65
44.21
0.964
1.041
837
68.81
18.44
50.37
121
218
277
316
358
411
124
222
288
327
371
433
Specific Gravity
Sulfur, wt.%
Nitrogen, ppm
Aromatics, lv.%
Mono-, lv.%
Poly-, lv.%
Dist., D2887, C
IBP
10%
50%
70%
90%
FBP
Figure 4
Impact of LCO Endpoint Reduction
on Hydrotreating Performance
89
SR
78
30% Hi FBP LCO
66
30% Lo FBP LCO
55
44
33
22
11
0
0
100
200
300
400
500
600
Product Sulfur, ppm
Figure 5
Activity Comparisons at Different LCO
FBP and Concentration
since the basic technique separates out the sulfur based on boiling
point distribution. The sulfur molecules left in these previously treated
feeds have already been processed
once in a high temperature, high
pressure hydrotreating application.
Those conditions easily remove the
majority of sulfur molecules and
leave only those sulfur species that
are multi-ring, sterically hindered
molecules and other aromatic nitrogen compounds. It is these species
that require a greater level of saturation or ring opening before the
nitrogen or sulfur can be removed.
It is likely for there to be very low
concentrations of multiple-ring, partially saturated compounds that
need to be more fully saturated in
order to remove the sulfur. This is
enough to make it more difficult to
produce 10 ppm sulfur product
from such feeds.
28
25
Hi EP LCO
22
Lo EP LCO
19
17
14
11
8
6
3
0
0
10
15
% LCO
20
25
30
35
An understanding of the upstream

processing is important when considering the use of synthetic
crudes. Production of synthetic
fuels involves a combination of several processes in order to accommodate downstream processing.
These upstream processes include
coking or an ebullating bed resid
operation, followed by a hydrotreating or hydrocracking operation in
order to produce a lighter grade
material. These products are then
blended in with other heavier materials as a diluting or cutting stock
and sent downstream as synthetic
crude. The synthetic diesel used in
this work is taken from a product
diesel cut from a synthetic VGO
hydrocracker. Figure 6 shows the
activity difference between the SR
and the blended SR/synthetic
diesel. Note that at higher product
sulfur, the two feedstocks respond
fairly similarly to each other. As the
application becomes more demanding, the required reactor temperature increases dramatically for
the synthetic diesel feed as compared to the SR feed. The blended
feed requires more than 14C higher temperature relative to the SR to
achieve ULSD sulfur levels.
Figure 6
Activity Comparison of the SR and Synthetic Diesel Blend
50
44
SR
Synthetic
39
33
28
22
17
11
6
0
20
40
60
80
100
120
Product Sulfur, ppm
Figure 7
Activity Comparison of Previously Hydrotreated Streams
It is reasonable to expect that the

upstream hydroprocessing of the
synthetic diesel material results in a
feed which behaves similarly to
other previously hydrotreated feedstocks like those from the EB and
FB resid applications. Two feedstocks from these sources are
shown in Figure 7. These two feedstocks have a remarkably similar
response as that observed for the
synthetic diesel feedstock. The
fixed bed diesel fraction, which has
significantly lower sulfur and nitrogen than the other two feedstocks,
shows over 22C higher SOR than
either the EB or synthetic diesels at
10 ppm sulfur. These data show
that upstream processing prior to
treating in a ULSD unit can have a
dramatic effect on the activity of the
unit and consequently decrease
cycle length.
66
Figure 8 examines how the product

API is changed during processing
for the synthetic diesel blend. As
can be seen, there is only a one
number increase in product API
over an almost 56C WABT change
compared to greater than two number increase for the SR feed over a
similar temperature span.
55
FB
EB
44
33
22
11
0
0
20
40
60
80
100
120
Product Sulfur, ppm
Figure 8
Comparison in Specific Gravity
0.832
Syn Diesel
0.835
SR
0.837
0.840
0.842
0.845
0.847
0
11
22
33
44
56
Increasing WABT, C
67
78
89
Specific Gravity
Synthetic
Aromatic saturation in the ULSD unit

is also a concern in order to meet
required cetane and aromatic targets. The higher temperature
required to process these previously processed streams may make it
difficult to achieve much aromatics
saturation because of the approach
to the thermodynamic equilibrium
limit for aromatic saturation. Figure
9 compares the aromatic saturation
achieved for the SR diesel and the
synthetic diesel blend. The synthetic diesel has a low level of poly aromatic compounds, and the blend
actually has a slightly lower concentration of PNAs compared to the SR
feed. Less saturation is achieved
on the synthetic blend, probably a
reflection of the fact that mono aromatic molecules are the predominant species, and these are quite
difficult to saturate. The equilibrium
limit on conversion is readily apparent in the figure. The synthetic
35
Making the switch to using a NiMo

catalyst in this application has much
more significant effect on unit performance. In Figure 11 the base
case conditions of low pressure
and low hydrogen/oil ratio actually
result in activity which is similar to
what was observed at these conditions for the CoMo catalyst. After
increasing the hydrogen rate at low
pressure, the all NiMo system gains
over 14C relative to the base condi36
33.0
29.0
27.0
25.0
23.0
21.0
Syn Diesel
19.0
SR
17.0
15.0
11
22
33
44
56
67
78
89
Increasing WABT, C
Figure 10
CoMo Catalyst Activity on Synthetic Diesel
66
48 BAR & 232 H2/Oil
55
48 BAR & 463 H2/Oil

90 BAR & 463 H2/Oil
44
90 BAR & 712 H2/Oil
33
22
11
0
0
20
40
60
80
100
120
Product Sulfur, ppm
Figure 11
NiMo Catalyst Activity on Synthetic Diesel
66
49 BAR & 232 H2/Oil
55
48 BAR & 463 H2/Oil

90 BAR & 463 H2/Oil
44
90 BAR & 712 H2/Oil
33
22
11
0
0
20
40
60
80
Product Sulfur, ppm
Total Aromatics, vol.%
31.0
When evaluating opportunities to

hydrotreat
previously
treated
streams for ULSD, the need to
examine the catalytic effects of
these feeds is important. ART completed a series of pilot tests using
the synthetic diesel feed blend covering a wide range in operating conditions.
Figure 10 summarizes
some of the results for one of ARTs
high activity CoMo catalysts. The
base case condition was again 48
BAR, 232 Nm3/m3 hydrogen/oil ratio
and 0.7 LHSV. As you can see from
the figure, the base case conditions
result in the highest start of run temperature for ULSD. Increasing the
H2/Oil ratio to 463 Nm3/m3 results in
a decrease in the SOR WABT of
over 8C. The third data set shows
results for higher hydrogen partial
pressure, 90 BAR, and 463
Nm3/m3hydrogen to oil ratio. This
results in a gain of 14C lower SOR
compared to the base case conditions or an incremental 6C lower
temperature due to the increase in
hydrogen pressure. Finally, increasing the hydrogen rate to 712
Nm3/m3 at high pressure provides
over 17C lower SOR as compared
to the base case system, but only
an additional 3C relative to the high
pressure lower gas rate case.
These data clearly demonstrate the
significant benefits of increasing
hydrogen partial pressure when
treating these types of difficult
feeds.
Figure 9
Change in Total Aromatics on SR and Synthetic Diesel
diesel provides a two number

decrease in total aromatics while
the SR diesel has an almost 10 number decrease. This can be problematic if trying to meet an aromatic
target in the diesel pool.
100
120
tions, a larger activity gain than

observed for the CoMo catalyst.
The activity benefit increases to
over 22C lower required temperature at 1300 psia, and increasing
the hydrogen rate to 712 Nm3/m3
results in over 28C lower temperature compared to the base conditions. The NiMo catalyst is at least
11C more active than the CoMo
catalyst at this last set of conditions.
Clearly there is a need to maximize
saturation when treating these preprocessed types of feeds. Any
increase in hydrogen partial pressure helps this and, as the data just
discussed indicates, the catalyst
selection has a significant impact.
High activity NiMo catalysts are better saturation catalysts compared to
high activity CoMo catalysts, and this

appears critical to removing the harder sulfur species present in these preprocessed feeds. In units that are
constrained by limited hydrogen or
lower hydrogen pressures, the use of
even a small amount of NiMo catalyst
will prove to be beneficial in order to
remove the remaining difficult sulfurs.
Advanced Refining Technologies can
work closely with refining technical
staff to help plan for processing
opportunity feeds such as those discussed above. One of the keys is
being aware of the potential impact
processing certain feeds will have on
unit performance. Feeds which have
been previously processed present
unique challenges and ART is well
positioned with its experience at pro-
viding customized catalyst systems

for ULSD applications. Opportunity
feeds provide yet another objective
to consider when designing the
appropriate catalyst system to maximize unit performance.
References
1. Olsen, C., Krenzke, L.D., Watkins, B.,
AICHE Spring National Meeting, New
Orleans, March 2002.
2. Krenzke, D., Armstrong, M., 2001 ERTC
Meeting, Madrid, Spain.
3. Olsen, C., Krenzke, L.D., 2005 NPRA
Annual Meeting, Paper AM-05-17.
4. Olsen, C., DAngelo, G., 2006 NPRA
Annual Meeting, Paper AM-06-06.
5. Olsen, C., Watkins, B., Shiflett, W., ERTC
Meeting, Barcelona, Spain, November 2007
37
Cetane Improvement In Diesel

Hydrotreating
Greg Rosinski
Technical Services Engineer
Charles Olsen
Manager
ADVANCED REFINING
TECHNOLOGIES
Chicago, IL USA
emand for higher performance

in diesel engines has resulted
in an increase in minimum
cetane numbers for diesel fuel. It is
expected that the desire for higher
cetane will continue as indicated by
the recommendations of the World
Wide Fuels Charter. Thus, it is important for refiners to understand the
effects of both feedstock and processing parameters on the cetane of
diesel fuel to enable them to more
effectively manage their distillate
hydrotreating units to meet ever more
stringent fuels specifications.
While some diesel is produced in

hydrocrackers, the vast majority of
diesel is processed in diesel
hydrotreaters (DHTs) which usually
co-process streams such as FCC
LCO, LCGO, SR diesel, and

Kerosene. The units processing
significant amounts of cracked
stocks need special attention in
order to meet product cetane
requirements. To understand why
this is so it is necessary to know
what cetane is, and how the different molecular species influence it.
The cetane number is a measure of
the ignition quality of diesel fuel and
is based upon the compound
cetane or hexadecane which is
assigned a cetane number of 100.
It is analogous to the octane number in gasoline. Gasoline octane
increases with olefin, aromatic, and
iso-paraffin contents, whereas
cetane number increases with
paraffin and naphthene contents.
Table I
Cetane Number of Pure Compunds
Paraffins
Isoparaffins
Naphthenes
Aromatics
Compound
Formula
n-Decane
n-Pentadecane
C10H22
C15H32
Cetane
Number
76
95
n-Eicosane
3-Ethyldecane
4,5-Diethyloctane
Heptamethylnonane
8-Propylpentadecane
7,8-Diethyltetradecane
9,10-Dimethyloctadecane
Decalin
3-Cyclohexylhexane
2-Methyl-3-cyclohexylnonane
2-Cyclohexyltetradecane
1-Methylnaphthalene
n-Pentylbenzene
Biphenyl
1-Butylnaphthalene
n-Nonylbenzene
2-Octylnaphthalene
n-Tetradecylbenzene
C20H42
C12H26
C12H26
C16H34
C18H38
C18H38
C20H42
C10H18
C12H24
C16H32
C20H40
C11H10
C11H16
C12H10
C14H16
C15H24
C18H24
C20H34
110
48
20
15
48
67
59
48
36
70
57
0
8
21
6
50
18
72
Thus, a fuel with a high cetane value

has low octane and visa versa.
Table 1 lists some pure compounds
and their corresponding cetane
number. As can be seen, paraffins,
particularly normal paraffins, have
very high cetane numbers while
aromatics, especially naphthalene
type aromatics, have very low
cetane numbers. Certain distillate
range materials like FCC LCO are
high in naphthalenes which
explains the low cetane number of
LCO feedstocks.
The actual cetane number is rarely
analyzed in refineries since it
requires a specialized motor for its
determination. Most refiners use
cetane index, typically, ASTM D-976
and ASTM D-4737. D976 uses the
API gravity and the 50% distillation
point, whereas D4737 uses the
gravity with the 10%, 50% and 90%
distillation points. The two equations are shown below.
ASTM D-976
cetane index = -420.34+0.016*API 2
+0.192*API*(log(T50))2+65.01*log
(T50)0.0001809 *T502
Where T50 is the D86 50% point in
degrees F
ASTM D-4737
Cetane index= 45.2+0.0892*(T10215)+[0.131+0.901* B]*[T50-260]
+[0.0523(0.420)*(B)][T90310]+[0.00049]*[(T10-215)2
(T90-310)2] + (107)*(B) + (60)*B2
Where: B = Exp[-3.5*(sp. gr. 0.85)]
1 and the D86 temperatures are in C
Figure 1 compares the cetane index
(D976) for a number of different distillate feed sources. It is readily apparent that FCC LCOs have the lowest
cetane while straight run (SR) materials have the highest cetane. Distillate
feeds derived from coking operations
tend to have a cetane similar to SR
material, while kerosene tends to have
somewhat lower cetane owing to the
lower boiling point. For the diesel
range materials, the feeds with

lower API gravity (LCO) have lower
cetane index demonstrating that
within a given boiling range the API
is a reasonable tool for estimating
the cetane index.
Figure 2 shows the cetane index as
a function of poly aromatics content
for a variety of distillate feeds. The
LCOs clearly have the highest concentration of poly aromatics and
correspondingly lower cetane
index. The SR, LCGO and vacuum
bottom gas oil (VBGO) all have considerably lower PNA content with
higher cetane index values compared to the LCOs. Kerosenes
have very low polynuclear aromatics
(PNA) content, but because of the
lower molecular weight (kerosene
generally has compounds containing less than 16 carbon atoms) the
cetane index is slightly lower than
for SR diesel material. The figures
make it clear that when it comes to
cetane, LCO is a problem due to the
high PNA content.
39
Figures 3 shows the level of cetane

increase (measured by delta cetane
index) that has been achieved commercially as a function of unit LHSV.
Generally speaking, as LHSV
decreases the potential cetane
improvement increases. At a LHSV
around 1 hr-1 or less, cetane index
increases of 10 or more numbers
are achievable (provided the H2
pressure is high enough), while at a
LHSV greater than about 1.7 hr-1
the cetane improvement is about 4
numbers or less.
Figure 4 summarizes the cetane
increase as a function of unit pressure. Not surprisingly, higher pressure units tend to achieve much
larger cetane increases. In these
examples, the cetane uplift is typically less than 6 numbers when the
unit pressure is less than 69 BARG.
Cetane uplift increases to 8-10 numbers as pressure increases beyond
69 BARG. The data in Figures 3 and
4 also suggest there might be a
60
55
Cetane Index (D976)
50
45
40
35
30
25
20
15
10
1.037
1.000
0.966
0.904
0.934
0.876
0.850
0.825
0.802
0.780
Feed Specific Gravity

LCO
Kerosene
Straight Run
LCGO/VBGO
Figure 2
Cetane Index and Polynuclear Aromatics (PNAs)
60
55
Feed Cetane Index (D976)
There are a number of parameters

which influence cetane improvement in the diesel hydrotreater.
Hydrogen partial pressure and
LHSV are key operating conditions
which effect the product cetane.
Catalyst selection also plays an
important role since at higher pressures NiMo catalysts have a higher
PNA saturation activity compared to
CoMo catalysts.
Figure 1
Cetane Index of Various Distillate Feeds
50
45
40
35
30
25
20
15
10
0
10
15
20
25
30
35
40
45
50
55
60
65
70
75
Feed PNAs, vol.%

LCO
Kerosene
Straight Run
LCGO/VBGO
Figure 3
Effect of LHSV on Cetane in a Variety of Commercial Units
14.0
13.0
12.0
11.0
Delta Cetane Index
Treating feedstocks that contain

LCO will become more challenging
as LCO yields increase and cetane
requirements become more stringent. With this in mind, it is useful to
survey the level of performance currently being achieved in commercial
diesel hydrotreating units today.
This will help to define what the reasonable level of cetane uplift that
can be expected is, and if there is a
practical maximum uplift that can
be achieved via hydrotreating. It is
also useful to ascertain whether
current operations indicate if
cetane changes (decreases) during
the cycle. All of these are important
questions for refiners interested in
increasing diesel yields from lower
quality feedstocks.
10.0
9.0
8.0
7.0
6.0
5.0
4.0
3.0
2.0
1.0
0.0
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
LHSV
Refiner A
Refiner B
Refiner C
Refiner D
Refiner E
Refiner F
Figure 4
Effect of Unit Pressure on Product Cetane
practical limit to cetane improvement achieved from typical

hydrotreating. Comparing the
cetane uplift achieve by Refiners A
and C shows about 8-10 number
improvement for both units despite
the difference in operating pressure
at similar LHSV.
14.0
13.0
12.0
11.0
Delta Cetane Index
10.0
9.0
8.0
7.0
6.0
5.0
4.0
3.0
2.0
1.0
0.0
28
41
55
69
83
97
110
124
138
152
Reactor Inlet Pressure, BARG

Refiner A
Refiner B
Refiner C
Refiner D
Refiner E
Refiner F
Figure 5
Feed Gravity Has a Significant Impact on Cetane Uplift
14
13
12
Delta Cetane Index
11
10
9
8
7
6
5
4
3
2
1
0
0.959
0.934
0.910
0.887
0.865
0.845
0.825
0.806
Feed Specific Gravity

Refiner A
Refiner B
Refiner D
Refiner E
Refiner F
Figure 6
API Uplift and Cetane Increase
for Several Commercial Units
The hydrogen consumption is

another important consideration
when discussing cetane improvement. There is a general rule of
thumb that says the hydrogen consumption is roughly equal to (100*
API Uplift) or (100* Cetane uplift).
Averaging the H2 consumption
required for the observed cetane
increase with the units discussed
here indicates that the H2 consumption varies from about (80*Cetane
uplift) at low pressure (Refiner D) to
(150-175* Cetane uplift) for the high
pressure units (Refiners A & C). The
data suggests the rule of thumb is a
reasonable estimate for H2 consumption for units operating below
about 69 BARG.
These data demonstrate that as the
unit conditions (LHSV and pressure)
get more favorable for PNA saturation, the cetane uplift increases.
However, is cetane uplift constant
during an entire cycle?
11
10
9
Delta Cetane Index
Figure 5 summarizes the commercial data in another way. It shows

how the cetane increase correlates
with the API gravity of the feed. In
general, the cetane uplift increases
as the feed API decreases. Figure 6
shows how the observed API
increase correlates with the cetane
index increase. This data shows
that the API uplift is a reasonable
predictor for the cetane increase.
8
7
6
5
4
3
2
1
0
0
Delta, API
Refiner A
Refiner B
Refiner D
Refiner E
Refiner F
10
Figure 7 summarizes the observed

cetane from three ULSD units currently using SmART systems.
Refiner A is a high pressure unit with
a LHSV around 1. The feed to this
unit contains 40-50 vol.% LCO. This
unit has not experienced a significant decline in cetane uplift during
the cycle. Refiner B is a higher
LHSV unit, with a lower pressure
than Refiner A, but the feed is rela-
41
As mentioned previously, the catalyst system will also have an impact

on the degree of cetane uplift
achieved in a hydrotreater. It is
common knowledge that NiMo catalysts have a higher saturation activity than CoMo catalysts, and therefore a NiMo catalyst is expected to
deliver greater cetane uplift. Figure
8 summarizes pilot plant data which
demonstrates this. These data were
generated using a 50% LCO containing feed, and shows that the
NiMo catalyst results in almost twice
the cetane uplift compared to the
CoMo catalyst. The SmART Catalyst
System, which utilizes both the
CoMo and NiMo catalyst, results in
a cetane uplift which is nearly 2
numbers higher than the all CoMo
system with only a small increase in
hydrogen consumption. For H2 constrained refiners this is an ideal
solution for improving the product
cetane.
Figure 7
Variation in Cetane Uplift During a Cycle
14
13
12
Delta Cetane Index
11
10
9
8
7
6
5
4
3
2
1
0
0
100
200
300
400
500
600
700
800
900
1000
1100
Days on Stream
Refiner A
Refiner B
Refiner C
Figure 8
The Catalyst System Has a Large Impact on Cetane Uplift
15
14
13
Cetane Index Increase
tively light with about 35-40 vol.%

LCO and LCGO. Despite the lower
pressure and higher LHSV, this
refiner also did not see an appreciable decline in cetane uplift during the cycle. Refiner C is a higher
pressure unit with a lower LHSV
compared to Refiner B. The feed is
high in sulfur with large (>80 vol.%)
amounts of cracked stock, especially LCO. This unit does show a
slow, steady decline in cetane uplift;
the cetane uplift is 2-3 numbers
lower after more than two years on
stream. This suggests that units
with difficult feeds containing high
fractions of LCO and other cracked
stocks, or units without sufficient
hydrogen, will experience decreasing cetane uplift during the cycle.
12
11
10
9
8
7
6
5
4
327
329
332
335
338
341
343
346
Temperature, C
NiMo
CoMo
SmART
349
352
354
Factors Influencing ULSD Product Color
Greg Rosinski
Brian Watkins
Charles Olsen
Manager
ADVANCED REFINING
TECHNOLOGIES
Chicago, IL USA
roduct color is a common concern for refiners with a number

of petroleum products including kerosene, jet fuel, diesel fuel and
lube base oils. With the introduction of
ultra low sulfur diesel (ULSD) the issue
of diesel product color has become
more of an issue as the typical ULSD
unit cycle length may now be limited
by color degradation of the product.
Refiners have been uncertain about
end of run (EOR) reactor outlet temperatures with expectations in the
range of 377-404C. The typical
ULSD unit has a deactivation rate in
the range of 1.1-1.7C/month so an
increase in EOR temperature of 611C has a significant impact on a
refiners planning and economics.
Figure 1 summarizes data from a commercial ULSD unit using ART catalysts. The data shows that in this case
the product color exceeded 2.5

ASTM, the pipeline color specification for diesel, at a reactor outlet
temperature above 388C. The feed
to this unit contained 30% LCO and
it was operated at 1.0 LHSV and 59
BARG inlet pressure.
It is well known that the color of distillate products is affected by the reaction conditions in the hydrotreater,
especially temperature and hydrogen
partial pressure. As (outlet) temperature increases and/or hydrogen partial pressure decreases, the product
color degrades. It is also generally
accepted that the species responsible for color formation in distillates are
polynuclear aromatic (PNA) molecules. Some of these PNAs are
green/blue and fluorescent in color
which is apparent even at very low
concentrations of these species.
Certain nitrogen (and other polar)
43
Figure 1
ULSD Product Color Using a SmART Catalyst SystemTM
3.5
Product Color (ASTM)
3.0
2.5
2.0
1.5
1.0
0.5
0.0
304
316
327
338
349
360
371
382
393
404
416
Reactor Outlet Temperature C
compounds have also been implicated as problems for distillate

product color and product instability. These species can polymerize to
form condensed aromatic structures which tend to be green to yellow/brown in color and can also
form sediment via oxidation and
free radical reactions.(1) Work conducted by Ma et.al.(2) concluded that
the specific species responsible for
color degradation are anthracene,
fluoranthene and their alkylated
derivatives. These are both three
ringed aromatic structures and are
shown in Figure 2.
PNAs such as these are readily saturated to one and two ringed aromatics under typical diesel
hydrotreating conditions at start of
run (SOR), but as the temperature
of the reactor increases towards
EOR, an equilibrium constraint may
be reached whereby the reverse
dehydrogenation reaction becomes
more favorable. At some combination of low hydrogen partial pressure and high temperature the
dehydrogenation reaction predominates and PNAs begin to form
resulting in a degradation of the
color of the diesel product. Other
work completed by Takatsuka
et.al.(3,4) showed that the color bod44
ies responsible for diesel product

color degradation were concentrated
in the higher boiling points in the
diesel (>250C). This suggests that
color can be improved by adjusting
the diesel endpoint. They also suggest that the color bodies responsible
for color formation in desulfurized
diesel are newly formed PNA structures from desulfurized aromatic compounds.
To learn more about color degradation
in ULSD, ART completed a pilot plant
study which investigated diesel prod-
uct color over a wide range of operating conditions. The study utilized
spent ART CDXi, a premium high
activity CoMo catalyst for ULSD.
The sample of spent catalyst had
been in commercial diesel service
for well over a year and had a carbon content of 10.9 wt.%. The testing program included straight run
(SR) diesels, a 30 vol.% LCO blend
and a 30 vol.% light coker gas oil
(LCGO) blend. The properties of
all the feeds are listed in Table I.
The test was designed to examine
the effects of H2 partial pressure,
H2/Oil ratio and temperature on
ULSD product color. H2 partial
pressure varied from 20-80 BAR
and the H2/Oil ratio covered the
range of 125-375 Nm3/m3.
Figure 3 shows how the diesel product color changes with temperature
and pressure for the straight run
feed (SR #1). Not surprisingly, pressure clearly has a significant
impact. At the lowest operating
pressure, which corresponds to 2124 BAR H2 pressure, the product
color exceeds 2.5 ASTM at a temperature greater than 393-399C.
Doubling the unit pressure to 55
BARG allows the temperature to
increase to 416C before the product color reaches 2.5 ASTM, and at
even higher pressures the product
color is well below 2.5 ASTM for all
practical temperatures encountered
in ULSD processing. At these con-
Figure 2
Primary Fluorescence Species in
Hydrotreated Diesel (Ref 2)
anthracene
fluoranthene
Table I
Feedstock Properties
Specific Gravity
Sulfur, wt.%
Nitrogen, wppm
Total Aromatics, vol.%
PNAs (2-ring+), vol.%
ASTM Color
Distillation (D2887), C
IBP
10%
50%
90%
FBP
SR #1
0.8588
0.68
120
31.5
10.7
L3.5
SR #1/LCO
0.8887
0.93
293
40.0
21.9
L6.5
SR #2
0.8527
1.12
127
27.3
9.5
L2.0
SR #2/LCGO
0.8569
1.34
249
30.1
10.8
L5.5
135
234
304
316
447
129
226
296
373
449
106
247
323
361
393
114
225
304
353
401
Figure 3
Product Color Improves with Pressure
ditions the H2 partial pressure

increases by a factor of about 3.4
going from 28 BARG up to 83 BARG
total pressure for temperatures
around 400C.
3.5
Straight Run
3.0
ASTM Color
2.5
28 BARG
55 BARG
2.0
83 BARG
1.5
1.0
0.5
0.0
360
371
382
393
404
416
427
Temperature, C
Figure 4
Product Color Improvement with Increased H2/Oil Ratio
3.5
Straight Run
3.0
125 Nm3/m3 H2
375 Nm3/m3 H2
ASTM Color
2.5
2.0
1.5
1.0
0.5
0.0
360
366
371
377
382
388
393
Temperature, C
399
404
410
The data in Figure 3 were generated

at the low end of H2/Oil ratios investigated. Figure 4 shows the effect of
increasing the H2/Oil ratio at 28
BARG on the SR #1 feed. At this low
pressure, the H2/Oil ratio has a significant impact on product color.
The data show that the temperature
can be increased from 396C to
well above 404C before product
color exceeds 2.5 ASTM.
AT 28
BARG total pressure and 404C,
changing the H2/Oil ratio from 125375 Nm3/m3 results in a 10%
increase in H2 partial pressure
which appears to be enough to
keep the reaction environment on
the favorable side of the hydrogenation-dehydrogenation equilibrium curve. At higher operating pressures the impact of increasing the
H2/Oil ratio is reduced when processing the SR feed, but still has a
positive effect on suppressing product color.
As might be expected, adding LCO
to the ULSD unit feed makes the
product color situation worse.
Figure 5 compares the product
color for the SR feed and the 30%
LCO feed at 375 Nm3/m3 H2/Oil
ratio and two pressures. The SR
feed results in acceptable color
45
The pilot plant program also investigated the effects of a coker derived
material on ULSD product color.
Figure 7 compares the product
color for the second SR feed and a
30% LCGO/70% SR #2 blend at 55
BARG. The data indicates that the
feed containing LCGO behaves similarly to the SR feed. In both cases
the outlet temperature can exceed
416C before product color
approaches the ASTM 2.5 level.
This is not surprising when comparing the properties of the two feeds.
The aromatics level, and in particular the PNA concentrations, are
essentially the same for the SR and
the coker blend. Compare this with
the LCO blend shown in Table I
where the PNAs are twice that of
the SR or LCGO feeds.
As mentioned previously, it is generally accepted that product color
is related to PNAs, and earlier work
has concluded that specific threeringed aromatics are responsible
for color degradation in diesel.
Figure 8 shows a comparison of
the product PNAs (three-ring aromatics) and diesel product color
4.0
55 BARG
3.5
ASTM Color
3.0
LCO
83 BARG
55 BARG
2.0
1.5
83 BARG
1.0
0.5
0.0
338
349
360
371
382
393
404
416
427
Temperature, C
Figure 6
Effects of H2/Oil Ratio on Product Color for 30% LCO
416
410
83 BARG
55 BARG
404
399
393
388
382
377
371
89
178
267
356
445
H2/Oil Ratio, Nm3/m3
Figure 7
Product Color Comparison for LCGO and SR
2.0
SR #2
LCGO
1.5
1.0
0.5
0.0
338
SR
2.5
Temperature for 2.5 ASTM Color, C
Figure 6 demonstrates the effects of

the H2/Oil ratio on product color
when processing the 30% LCO
feed. It shows the temperature at
which the color reaches 2.5 ASTM
as a function of H2/Oil ratio for both
55 and 83 BARG total pressure.
The temperature increases by about
14C when the H2/Oil ratio is
increased from 125 to 375 Nm3/m3.
That range of H2 rates corresponds
to an increase in hydrogen partial
pressure of 5-10%
Figure 5
Comparison of Product Color for SR and 30% LCO
ASTM Color
over the wide range of temperatures for both pressures shown.

This compares with the 30% LCO
feed which goes off color at about
388-393C at 55 BARG; and at 83
BARG the temperature can exceed
404C before reaching 2.5 ASTM
color. This data demonstrates the
significant impact that pressure has
on diesel product color when processing feeds that contain LCO.
349
360
371
382
393
Temperature, C
404
416
427
From these data it is clear that

hydrogenation of PNAs is key to
maintaining acceptable product
color in ULSD. This suggests a couple of approaches that allow an
increase in EOR outlet temperatures
and thereby increase the ULSD unit
cycle length.
One approach which has been put
to commercial practice is to
increase quench to the bottom bed
of the hydrotreater. This accomplishes two things which are important to maintaining a good environment for hydrogenation of PNAs. It
reduces the outlet temperature and
helps to increase the outlet hydrogen partial pressure relative to no or
lower amounts of quench.
This, of course, requires that the
upper beds of the hydrotreater be
run at higher WABTs in order to
maintain the required HDS conversion. This means that the furnace
must have sufficient capacity to
achieve the higher inlet temperatures. Operating in this manner
offers the potential to add an additional 6-11C on to the cycle length
depending on the unit capabilities
(furnace, quench capacity).
Figure 8
3+ Ring Aromatics Correlate with Diesel Product Color
4.0
SR
3.5
LCO
3.0
ASTM Color
for all the feeds and conditions of

the study. It is readily apparent that
the PNAs correlate reasonably well
with product color..
LCGO
2.5
2.0
1.5
1.0
0.5
0.0
0.0
1.0
2.0
3.0
4.0
5.0
Product PNAs (3 rings+), vol.%
Another approach, which may be

implemented with the one just discussed, involves adjusting the feed to
the unit. The data from this work
shows the significant impact LCO has
on diesel product color. Reducing (or
eliminating) the amount of LCO in the
feed will help to suppress product
color degradation as the unit
approaches EOR. There is also data
showing that the color bodies that
cause problems for ULSD tend to be
concentrated at the higher boiling
points of the distillation on the
feed/product. Reducing the endpoint
of the LCO reduces the concentration
of these species which will help

maintain acceptable product color
as the unit moves towards EOR.
References
1. J. Pedley et.al, ACS Division of Fuel
Chemistry, 35 (4), 1100-1107 (1990).
2. X. Ma et. al., Energy and Fuels, 10, pp
91-96 (1996).
3. T. Takatsuka et.al., 1991 NPRA Annual
Meeting, Paper AM-91-39.
4. T. Takatsuka et.al., Journal of the Japan
Petroleum Institute, Vol. 23, No. 2, pp 179184, 1992.
47
Hydroprocessing Catalysts from Chevron and Grace

ART delivers innovative and economic hydroprocessing catalyst solutions worldwide, helping
refiners on their never-ending quest to upgrade difficult petroleum feedstocks into clean fuels
and other valuable products.
Supported by cutting edge research and development and world class technical service, the
joint venture between Chevron and Grace combines over 60 years of material science expertise
and catalyst formulation experience with over 3,300 unit-years operating experience. In addition,
our partner affiliate licenses exceed 260 units across the globe.
Let ART be part of your solution. For more information on our products, contact us.
Advanced Refining Technologies 7500 Grace Drive Columbia, MD 21044 USA

V +1.410.531.4000 F +1.410.531.8246 E artcatalysts.com
Advanced Refining Technologies

Columbia, Maryland USA 1.410.531.8282 Fax 1.410.531.4540
Houston, Texas USA 1.281.449.9949 Fax 1.281.442.0447
Singapore 65.6831.4132 Fax 65.6737.5826
Richmond, California USA 1.510.242.1314 Fax 1.510.242.1412
Worms, Germany 49.6241.4030 Fax 49.6241.403455
Toda, Japan 81.48.431.1952 Fax 81.48.431.1949
www.e-catalysts.com
Advanced Refining Technologies 2010
The information presented herein is derived from our testing and experience. It is offered, free of charge, for your consideration, investigation and verification. Since operating conditions vary significantly, and since they are not under our control, we disclaim any and all warranties on the results which might
be obtained from the use of our products. You should make no assumption that all safety or environmental protection measures are indicated or that other
measures may not be required.
GRACE, GRACE DAVISON, GR, and SmART Catalyst System are trademarks, registered in the United States and/or other countries, of W. R. Grace
& Co.-Conn.
H-OIL is a registered trademark of Axens North America, Inc. CHEVRON, ICR, LC-FINING is a trademark of Chevron Lummus Global. OCR is a
registered trademark of Chevron Intellectual Property LLC. HOP is a registered trademark of Japan Energy Corporation licensed to Advanced Refining
Technologies LLC.
This presentation is an independent publication and is not affiliated with, nor has it been authorized, sponsored, or otherwise approved by any of the aforesaid companies.
ART and Advanced Refining Technologies are trademarks, registered in the United States and/or other countries, of Advanced Refining Technologies
LLC. ApART , AT, DX, LS, HSLS and StART are trademarks of Advanced Refining Technologies LLC.
This trademark list has been compiled using available published information as of the publication date of this brochure and may not accurately reflect current trademark ownership.

ULSD Brochure

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

ULSD Brochure

Uploaded by

Copyright:

Available Formats

Ultra Low Sulfur Diesel

Advanced Refining Technologies Hydroprocessing Catalysts from

Product Overview Chart

Advanced Refining Technologies (ART), is

Fixed Bed Resid

Fixed Bed Resid

ART offers refiners the unique synergies deEbullating Bed

ART Excels In ULSD Service: Update on Sulfur minimization By ART

Distillate Pool Maximization by Exploiting the use of Opportunity

Cetane Improvement In Diesel Hydrotreating

Factors Influencing ULSD Product Color

Very tight specifications on sulfur content in diesel fuel

Fraction Hard Sulfur

Feed D86 90% Point, C

SmART Catalyst System which is tailored to optimize

reprinted from Catalagram 95 2004

A number of operating parameters take on greater

For the high pressure operation, it is apparent that

Similarly, the 20% LCO feed shows a reactivity of about

20% LCO/Straight Run

The benefit of increasing H2/oil never reaches a

reprinted from Catalagram 95 2004

gen values around three to four, the straight run feed at

lyst volume, operating pressure and H2/oil ratio. Figure

In summary, it is possible in some cases to trade off

cracked stocks require higher pressure and higher

reprinted from Catalagram 95 2004

The SmART Catalyst SystemTM:

In 2001, ART introduced the SmART Catalyst SystemTM

properties and operating conditions as discussed in

Product Sulfur, ppm

before and after selectively removing the nitrogen via

gen. This is another important consideration when

Increase in required Temperature, C

reprinted from Catalagram 97 2005

hydrotreating. At high pressure, increasing the H2/Oil

increases, the H2 consumption relative to the base

to be minimized, a NiMo catalyst for ULSD is the wrong

minimizing hydrogen consumption (i.e. minimizing

the slowest step and also consumes more hydrogen

4Q04 2Q05 2Q04 3Q04

reprinted from Catalagram 97 2005

catalysts must be used in order to take full advantage

t has been widely discussed that

pathway involves saturation of one

reprinted from Catalagram 99 2006

Product Sulfur, ppm

is made up of compounds which

A number of factors need to be considered when designing the optimum

genation of the final aromatic ring

k1>k2, k3>k4, and k1>k3

The challenge when designing a

of the first hydrogenation reaction in

reprinted from Catalagram 99 2006

Product Sulfur, ppm

Cetane Index Increase

Product Sulfur, ppm

conditions, although both catalysts

Of course, if there is sufficient H2

these conditions the CoMo catalyst

1000 1250 1500 1750 2000

These data are a good example of

1.25 LHSV and 55 BARG

1000 1250 1500 1750 2000