ADSORBENTS AND ADSORPTION PROCESSES FOR POLLUTION

CONTROL
A. Buekens and N. N. Zyaykina
Department of Chemical Engineering - CHIS 2, Vrije Universiteit Brussel, Belgium
Keywords: Adsorption, Breakthrough, Desiccant, Desorption, Fixed bed adsorber, Fluidized
bed adsorber, Mass transfer zone, Moving bed adsorber, Pressure-swing, Regeneration,
Thermal-swing.
Contents
1. Survey
2. Properties of adsorbents
3. Fixed bed adsorbers
4. Continuous counter-current flow systems
5. Application of adsorbents for desiccation
6. Desorption and adsorbent regeneration
7. Design of adsorbers
8. Conclusions
Related Chapters
Glossary
Bibliography
Biographical Sketches
Summary
This chapter briefly covers the major technical issues related to industrial adsorbents and their
application in gas cleaning and desiccation. Fixed, moving, and fluidized bed types of
adsorbers are schematically shown and their operation principles are discussed. General
guidelines for an adsorber design are presented: however, they are mostly limited to the most
usual case of the fixed-bed adsorber. Several approaches, based on theoretical models for
calculation of breakthrough curves, are outlined. Because of the simplifications used, it is
advisable to treat these models with caution when applying them to real cases. Formulas for
calculation of the essential parameters such as pressure drop and bed length are explained.
Finally, the attention is drawn to some safety aspects, which should be considered while
operating an adsorber.
1. Survey
Technical adsorption units are so conceived that they provide and optimize the operating
conditions required for bringing the gas to be treated in close contact with an adsorbent during
a sufficient time period for contact and ensure that adsorption has taken place to the required
extent.
This implies that the molecules to be adsorbed may migrate to the surface, continue their path
by internal pore diffusion, and adsorb. It should be noted that each of these three consecutive
steps has distinct requirements, namely:

A high relative velocity between gas and particle reduces the external mass transfer
resistance and helps enhancing the migration of adsorbate towards the surface, through

the laminar layer surrounding individual particles. On the other hand, high gas
velocities also increase the pressure drop and reduce the residence time available in
the adsorption vessel;
Selecting adsorbent particles of a smaller size strongly increases pressure drop in a
fixed adsorbent bed, but also abridges the inner pathways required for internal
diffusion and increases the amount of external surface available. Disregarding the
effect of pressure drop, small adsorbent particles are highly desirable. However, in
practice particle size is fixed mainly by pressure drop considerations.
A rise in temperature slightly accelerates internal diffusion. Conversely, adsorption
capacity increases and residual pressure at equilibrium reduces when the operating
temperature is lowered. Moreover, adsorption remains an exothermic process, which
may somewhat increase the operating temperature.
When activated carbon is used for adsorbing pollutants from air it is necessary to
monitor exit temperature and/or carbon monoxide, in order to detect a possible
temperature run-away and eventual ignition of the carbon.
2. Properties of Adsorbents
The most important technical adsorbents are activated carbon, which is used for almost all
duties, except drying, and molecular sieves, silicagel, and alumina. Typical requirements for
commercial adsorbents are:

High porosity, high internal surface.

High adsorption efficiency in a wide range of adsorbate concentrations.
Good balance between macro-pores (for fast internal transport) and micro-pores (for
large internal surfaces).
Hydrophobic chemical structure (for treatment of moist gases) unless the adsorbent is
to be used as a desiccant.
Thermal stability unaffected by a cyclic regeneration.
Mechanical integrity during handling.
Low pressure drop over the adsorber bed.
Low cost for acquisition (and
eventually
disposal) of
adsorbents.
The most important properties of technical adsorbents are shown in the Table 1:

EOLSS - ADSORBENTS AND ADSORPTION PROCESSES FOR

POLLUTION CONTROL
Average pore
Sorptive
Particle porosity, Surface area, m2
diameter,
capacity,
%
g-1
nm
kg kg-1 (dry)

Adsorbent

Nature

Activated
alumina

Hydrophilic,
amorphous

4 14

50

320

0.1 0.33

Activated carbon

Hydrophobic,
amorphous

1-4

40 - 85

200 - 1200

0.3 0.7

0.3 0.6

35 - 50

400

0.2 0.5

Molecular-sieve Hydrophobic,

carbon

structured

Molecular-sieve
zeolites

Hydrophilic,
crystalline

0.3 -1

20 - 50

600 - 700

0.12 0.42

Polymeric
adsorbents

Hydrophobic,
amorphous

4 - 25

40 - 60

80 - 700

0.45 0.55

Silica
gel

Hydrophilic,
amorphous

2-5

47 - 71

300 - 850

0.35 0.5

Table 1: Main properties of technical adsorbents

[Values are taken from two sources: 1. Perry, R. H.; Green, D. W. eds. (1997). Perrys
Chemical Engineers' Handbook, 7th ed., New York, NY, USA: McGraw-Hill and 2. Seader, J.
D.; Henley, E. J. (1998). Separation process principles. 886 pp., New York, NY, USA: Wiley
and combined into one table especially for this article]

Table 1: Main properties of technical adsorbents

Apolar, low volatile organics, such as PAHs, adsorb best on activated carbon. Some inorganic
vapors (e.g. mercury) can be removed efficiently by adsorption, using specially prepared,
impregnated adsorbents (e.g. as a sulfide, arsenic- or antimony-compound). Activated carbon
and many polymers are hydrophobic; inorganic adsorbents are hydrophilic.
2.1. Activated Carbon
Activated carbon is quite easy to manufacture, but much more difficult to adapt to specific
applications. Traditional starting materials are charred wood, peat, lignite, or coal, but almost
any organic material, including sewage sludge, animal bone, blood or hair, may be used with
fair success. Production often proceeds in two either successive or overlapping steps:
1) Charring the original material, and
2) Activating it thermally or chemically in order to create the desired chemical pore
structure.
Charred materials still retain the external aspects of the original feed. Charcoal, for instance
still recognizably shows the cellular structure specific to the wood used. Coconut shells are a
premium starting material, because of the superior crushing strength of the resulting
carbonized coconut scale.
Adsorption on activated carbon is capable of a deep elimination of minor amounts of almost
all low to medium volatile pollutants from a gas stream. Optimal operating conditions are:

a relatively cold gas (optimal < 30 C),

dry (relative humidity < 60 %), and
dust free (particulate concentration < 1 mg m-3) gas feed.

High temperatures favors desorption and thus lower adsorption capacity. The presence of
water in the activated carbon or of water vapor in the adsorbate can affect the adsorption of
organic gases. Molecules with a high affinity to the carbon surface, whether soluble or not, are
less influenced by the presence of water. Solubility is not a dominant factor that affects

adsorption on a wet carbon bed. Water vapor competes for adsorption sites and dust may clog
the bed of activated carbon, if adsorption takes place in fixed bed layer.
Activated carbon can be supplied under the form of fibrous mats or even shaped filter
elements. Given the supplemental cost for such conditioning these address only rather specific
applications.
2.2. Specialty Impregnated Carbons
Physical forces are not always sufficient for total adsorption of a particular component.
Specialty impregnated carbon refers to an activated carbon that is chemically coated or treated
to enhance its adsorptivity towards specific compounds. Impregnated carbons are developed
for pollutants that are difficult to control with standard activated carbons. Such adsorbents
retain specific contaminants long enough for their conversion by the chemical impregnating
agent, forming a stable compound within the carbon. They are used in a wide range of air, gas
and water purification and industrial process applications. Some examples of contaminants
removed with specially impregnated carbons are listed in Table 2.

EOLSS - ADSORBENTS AND ADSORPTION PROCESSES FOR

POLLUTION CONTROL
Impregnating
agent

Quantity, wt.
%

Sulfuric acid

2 - 25

Application
Gas purification
mercury

from

ammonia,

amine,

Potassium
carbonate

10 - 20

Gas purification from acid gases: HCl, HF,

SO2, NO2

Manganese IV oxide

no data

Gas purification from aldehyde

Triethylene diamine

25

Gas protection
iodide

from

radioactive

methyl

Zinc oxide

10

In gas masks against hydrogen cyanide

Chromium copper
silver salts

10 - 20

Military and civil protection: phosgene,

chlorine, arsine, sarin and other nerve gases

Silver

0.1 - 3

Domestic drinking water filters

Table 2: Applications of impregnated carbons

[Source:
Source:
adapted
from
http://chemistry.about.com/cs/adsorption1/index.htm.
Henning
K.-D.;
Schfer S. Impregnated activated carbon for environmental protection.]

Table 2: Applications of impregnated carbons

A typical application of such adsorbents is the reduction of mercury emissions. Even in small
amounts, mercury and its compounds have an extremely harmful effect on human health.
Mercury also represents a very complex problem in hydrocarbon exploration and production.
Moreover, mercury corrosion attack endangers dramatically process plants and facilities.
Therefore, prevention of this contaminant entering into either the biosphere or a technologybased system is absolutely critical. At least two techniques are feasible:

Injection of powdered activated carbon into combustion flue gas, with subsequent
collection in a particulate control device
Granular activated carbon fixed-bed adsorption.
Several studies indicate that sulfur-impregnated carbons (e.g., Mersorb , with 13 % elemental
sulfur content) exhibit significantly greater mercury uptake capacities when compared to
virgin carbons. The reaction results in a stable and insoluble compound, the mercury sulfide.
2.3. Zeolites
Zeolites, also called molecular sieves, are crystalline silicates with a general chemical formula
Me2/nO Al2O3 x SiO2 y H2O (n = valence) containing oxides of alkali- or alkali-earth metals
(Na, K, Ca) and characterized by regular structure of the pores with their dimensions in the
range of the size of a molecule. By changing the Si to Al ratio (which is never less than 1.0,
but there is no upper limit), it is possible to vary the negative charge on the framework of the
zeolite. Since each Al atom introduces one negative charge, this must be balanced by
exchangeable cations. Molecular sieves with specific adsorptive properties can be produced
by ion exchange or chemical treatment. An example of such a structure is given in Figure 1.
Activation typically requires 300 C, applied under full vacuum or an inert purge gas.

Figure 1: Structure of a molecular sieve adsorbent

Molecular sieve separations may be based on molecular size and shape because of the small
pore sizes in the sieves, typically less than one nanometer, and their narrow pore size
distribution.
High thermal and chemical stability make zeolites ideal materials for high temperature
applications such as catalytic membrane reactors.
2.4. Silicagel
Silicagel has an amorphous structure and is described by the general formula SiO 2 nH2O. The
adsorption ability and chemical properties of silicagels largely depend on the presence of
structural Si-OH groups at their surface. The main properties are outlined in Table 1.
Silicagels are subdivided into three classes according to their pore size: large pores (
5 nm) and medium pores (5 - 1.5 nm) adsorbents are used for sorption of organic vapors;
small pores (1.0 - 1.5 nm) serve for removal of easy condensing vapors and gases. Generally,
silicagel is used for sorption of polar compounds, especially for drying of different gaseous
media. The advantages of silicagel are:
A low temperature of regeneration: 100-200 C
Non-flammable
A high mechanical strength

2.5. Activated Alumina

It is produced from hydrated alumina, Al 2O3 nH2O (n = 1-3) by dehydration under carefully
controlled conditions to get n = 0.5. Common forms are balls 1 to 8 mm diameter, granules,
pellets 2 to 4 mm diameter, and powder. The two widest applications of activated alumina are
as a catalyst and a desiccant. It is also used as an adsorbent are for removal of oxygenates and
mercaptans from hydrocarbon feed streams, fluoride ions from water, etc. It has superior
mechanical strength and resistance to attrition than molecular sieves and silicagel, which is
important to consider in moving bed applications. Regeneration requires heating to about
205 C.
2.6. Selection of a Suitable Adsorbent
The selection of adsorbents follows a few simple rules:

Chemisorption is much more selective than van der Waals forces.

The higher the molecular weight, the stronger the retention.
Kinetic and steric effects may enhance selectivity.
Inorganic sorbents (silicagel, alumina, molecular sieves) are particularly suitable for
dehydration and for retaining hydrophilic substances.
Organic compounds are well retained by activated carbon and less well by lignite
coke.
3. Fixed bed Adsorbers
A fixed bed adsorber represents a cyclic batch system, in which the adsorbent bed is saturated
and also regenerated in a periodic manner. This type of adsorption units is a standard choice
for small and most medium sized installations, as well as for police filters. In the German
literature this term denotes a bed of adsorbent, which is inactive under normal circumstances,
but completely arrests a brief surge of a pollutant under emergency conditions.
A typical scheme is shown in Figure 2. The fixed bed adsorber consists of a steel vessel, with
a distributor plate at the bottom carrying the granular adsorbent medium, and ducting for the
gas feed, the cleaned gas, and the inlet and outlet for the regeneration medium (often steam).
To avoid entraining adsorbent particles or even fluidizing the bed during adsorption, the flow
is usually directed downwards. Regeneration proceeds in the reverse sense, so that the
capacity is fully restored at the exit side of the adsorbing mode.

Figure 2: Scheme of fixed bed adsorber: vertical and horizontal units

Depending on space and process requirements two types of adsorber vessel can be
distinguished:

a wide version, with large cross-section. Such type is destined to cope with a large gas
flow to be cleaned while maintaining a still reasonable linear gas velocity.
a radial or annular configuration, in which an expansion of capacity takes place in the
vertical rather than the horizontal direction. The latter option is much more
economical in case of pressurized vessels, since the wall thickness required by
mechanical strength considerations is proportional to the diameter of the vessel.
A single fixed bed adsorber can never operate on a continuous basis, since some provisions
for regeneration or replacement of the adsorbent must always be made. Of course, a police
filter may operate for many months before the available adsorption capacity is exhausted.
Therefore, adsorption plants usually consist of several fixed bed units. This can only be
avoided if the pollutant load is so low that periodic replacement and external regeneration or
even disposal of the adsorbent is acceptable.

Figure 3: Operation of a two-column adsorption system

The

availability

of

fixed

bed

adsorbers

follows

from

the

ratio:

The numerator is substantial, because the bed must be not only regenerated thoroughly, but
also dried and cooled before returning to activity. Before or during breakthrough the active
bed unit is taken out of service for regeneration and replaced by another one, with its
adsorbent already treated.

3.1. Breakthrough Curves

A breakthrough curve is a graphical representation showing the outlet adsorbate concentration
as a function either of operating time or of total volume of gas treated, while emphasizing the
time period in which the bed is confronted with saturation and a rise in concentration is taking
place. Such breakthrough curves are an important operational and design feature of fixed bed
adsorbers. Plant design requires data related to timing and shape of the breakthrough curve. It
should be noted that the removal of a series of chemical compounds A, B, C... by adsorption
gives rise to a separate breakthrough curve for each of these compounds. Two distinct cases
should then be considered:

Only the toxic target compound A must be retained, at all cost. The breakthrough
curve of A determines the length of the cycle and schedule of operation and the other
curves (for B, C, etc.) have no real importance in establishing the
adsorption/regeneration cycle;
All compounds must be retained equally well, or almost so. The breakthrough curve
of the compound breaking through first then determines the length of the cycle and the
schedule of the cyclic operation.
Generally, the most volatile compound breaks through faster than others do, unless it is
retained because of particular features, such as polarity or specific retention agents (e.g.,
sulfur for mercury vapors).
It is important to analyze the various phenomena determining the transition from an adsorbent
bed that retains the target compounds almost completely into one that allows dangerous
pollutants to seep through. In the ideal case the adsorption system reaches almost
instantaneous equilibrium between the target compound in the gas phase and its adsorbate
bound to the particles. If there is no axial dispersion in the bed, the breakthrough curve is
represented by a step function moving abruptly from zero exit concentration for the fully
treated effluent to the original and also final concentration, corresponding to the passage of
the gas to be treated through an entirely depleted bed. Since there is always some axial
dispersion in the bed, the breakthrough curve deviates considerably from the idealized step
model, which is only used for simple calculations, e.g., of maximum sorption capacity.
Another factor to be accounted for is competitive adsorption. It is not so simple to predict the
features of this phenomenon: on the one hand, there is a gradation in strength of adsorption,
so that some compounds displace others, with a higher vapor pressure, on the other hand,
there is the Law of mass action, implying that a high concentration of weakly adsorbed
compound can supplant a low concentration of stronger adsorbing ones. In practice, the
strongly adsorbing compounds move through the bed more slowly than poorly adsorbing
ones. Since adsorption remains complex, empirical study prevails over fundamental data and
theoretical considerations. Breakthrough curves are a standard feature of any study on
adsorption processes. Still, various models of a highly distinctive complexity can simulate this
transition. The simplest possible transition is a linear one, as depicted in Figure 4.

Figure 4: Ideal (left) and real (right) breakthrough curves

The real case curve assumes a S-shape, incorporating various effects due to diffusion, sorption
kinetics and mass transport resistances. If it is steeply ascending, it indicates a high capacity
and selectivity of adsorption. A slowly rising curve is undesirable, since it corresponds with
early breakthrough.
Figure 5 shows the gradual development of breakthrough through the fixed bed filled with
adsorbent. In this graph, initial retention is complete, without any leak stream, as could
develop following inadequate or incomplete contact between the fluid and adsorbent. Fluid
may find preferential pathways that more or less allow avoiding contact with adsorbent: a
packed bed is most porous at the wall, since this wall cannot be penetrated by the particles.
This effect is larger for lower values of the ratio of column diameter to particle diameter.

Figure 5: Breakthrough curve from a fixed bed adsorber

As time proceeds, an increasing fraction of the bed becomes saturated and hence inactive, so
that the lower part gradually becomes more and more actively engaged. At a given moment
most spare capacity is almost depleted and the effluent concentration first becomes detectable,
then at first increases exponentially, later linearly, finally rising to the original inlet value. At a

certain time it attains the value fixed in the emission standards, so that from that time on the
effluent quality will no longer be satisfactory. At that moment adsorption process may be
either switched to another fixed-bed unit, which halts the pollutants breaking through, or
continued until exhaustion. The latter may be permissible in the following cases:

The operation is followed by other treatment methods.

There is adequate buffering capacity, e.g. by mixing the particular effluent with other
streams either in a waiting basin or directing them towards the same chimney.
The average concentrations over certain periods are still in line with the average
values as specified in Regulation Codes.
At the breakthrough point the bed still has considerable adsorption capacity, which can be
expressed through the equivalent length of bed: the Length of Unused Bed (LUB).

(1)

where: q0 is the amount of substance adsorbed in equilibrium with fluid having initial
concentration of the target compound C0, qb - the amount of substance adsorbed at breakpoint,
and L - length of the bed.
Monitoring of the effluent concentration of an adsorption column requires starting effluent
analysis from the very low levels. Since the break point represents the limit of capacity, it is
common design practice to provide a factor of safety and end the operation before the
breakthrough occurs.
3.2. Mass Transfer Zone (MTZ)
Conceptually a fixed bed adsorber can be subdivided into three zones (see Figure 5):

An inactive saturated zone, in which the adsorption potential has already been fully
depleted and the adsorbent is in equilibrium with the adsorbate in the inlet fluid.
The active mass transfer zone (MTZ), in which the adsorbate concentration drops
from close to the feed value to the final value near zero. The concentration does not
reach nil because of non-idealities in the bed.
The lower bed zone, which is almost untouched.
It is important that the active mass transfer zone is short, and correspond to a fast decline of
the adsorbate concentration. Obviously, a slow descent and a long MTZ would encourage a
much faster breakthrough.
To establish the height of the MTZ a system consisting of a total depth with one, two, three,
and so on pilot beds is considered. The time it takes each system to reach breakthrough and
exhaustion is plotted against the total depth of the system. The results provide two lines
parallel to each other. The distance between those two lines, in a direction parallel to the axis

of cumulative depth is equal to the length of the MTZ. The same graph provides also the
relationship between bed depth and service time.
MTZ is partly depends on the size of the adsorbent particles: the fine particles tend to shorten
it.
4. Continuous Counter-current Flow Systems
For large plants it is disadvantageous to operate in a cyclic fashion, since the entire
installation including adsorption, desorption, drying and cooling stages is rather space
consuming and is not easily amenable to automatic control. Therefore, other types of
adsorbers are developed, which operate in a continuous mode. This is possible, when the
adsorbent moves as a countercurrent against the fluid at such a rate that the adsorption zone is
stationary. Such arrangements maximize the driving force of mass transfer, however require
much more complex design since it is obviously more difficult to move solids in a continuous
manner than it would be the case with moving fluids.
The technical units used for continuous operation are mainly:
1) Moving bed adsorbers, in which the bed of adsorbent moves downwards and the feed
gas enters near the center of the adsorption column, then rises through the charge, first
in contact with used adsorbent, later with freshly fed material (Figure 6a);
2) Fluidized bubbling bed types, in which the adsorbent is very well mixed by the rising
gas to be treated and moves up and down with respect to the containing vessel, pushed
around by rising gas bubbles (Figure 6b);
3) Circulating fluidized bed types, in which the adsorbent is pneumatically conveyed by
the gas to be treated. Pneumatic systems are used for removing dioxins from off-gas.
The bed is quite lean, typically with an adsorbent load of 50
500 mg Nm-3;
4) Rotary bed adsorbers, in which the drum moves together with the solid phase
relatively to a fixed inlet for the feed and fixed outlet for the clean gas.

Figure 6: Moving bed (a) and fluidized bed (b) adsorbers

Adsorbent is easily added to and removed from fluidized beds and pneumatic systems. The
pressure drop through a fluidized bed is effectively constant over a wide range of fluid flow
rates, and this makes it possible to treat large flow rates in relatively compact equipment. On
the other hand, attrition losses of an adsorbent are high. Moreover, in a bubbling bed
operating at a high gas velocity, an important fraction of the gas traverses the bed as bubbles,
with limited internal dust content and hence contact with adsorbent.
Different particle sizes are used in fluidized bed and in pneumatic systems. For smooth
fluidization a particle size of 0.2 to 1 mm is desirable, whereas pneumatic systems operate
best with very small particles, typically 10 to 50 m.
5. Application of Adsorbents for Desiccation
When an essentially complete water removal is desired, as in cryogenic operations
(distillation of liquefied air; VOCs removal; see also Condensation, Control of Carbon
Monoxide and Volatile Organic Compounds) or when (the usually small) drying installations

require utmost operating simplicity, the application of a granular adsorbent for the drying
purposes is usually the best technical and economic choice.
Adsorbents suitable for gas dehydration, due to their high affinity to water, are called
desiccants. The adsorbents most commonly included in this group are:

Silicagel and silica-base beads ;

Activated alumina and alumina-gel balls;
Activated bauxite;
Molecular sieves.

Synthetic molecular sieves are relatively more expensive. However, in return zeolites have a
number of advantages over other desiccants: they may adsorb water selectively or can be used
to remove other impurities as well together with water; their capacity of low relative humidity
is superior; and they are not damaged by condensation of water vapor to liquid water.
In many cases the drying agents are interchangeable (with exception of molecular sieves) and
equipment designed for one desiccant can be efficiently operated with another. A scheme of
principle showing a drying installation is illustrated in Figure 7.

Figure 7: Scheme of an adsorption set-up for the drying of gases

Humid gas where possible is first cooled and conducted to the moisture separator (1) for
removal of water droplets or fog, and then it goes for desiccation to the adsorbers (2). Once
saturated with moisture the adsorbent is regenerated in situ in the adsorber (4) by blowing
through the bed a hot gas, previously heated in the apparatus (5). After regeneration of the
drying agent the hot gas is cooled in the unit (7); then it proceeds to the second moisture
separator (8) and mixes with the main stream. The adsorbent is cooled down to 30-40 C in
the adsorber (3), which is returned afterwards to the drying stage of the cycle. It should be
noted that the typical regenerating gas temperature required (of about 175
205 C)
is identical for all desiccants, molecular sieves excepted. These materials require a
regeneration temperature of 260 - 375 C, if they have to recover their maximum capacity.
The capacity of desiccants to adsorb water vapor depends on several parameters:

Nature of the adsorbent

Time: as a general rule, the adsorption capacity decreases with time of adsorbent
service, which can reach several years. The rate of ageing is specific and depends on
the nature of desiccant, impurities present in the gas and operating conditions.
However, the most dramatic decrease of capacity usually occurs within the first 12

months of exploitation (down to 50 % of the original value) and after a certain time
the decrease practically stops and the capacity stabilizes.
Relative humidity (RH): for silicagel, activated alumina and activated bauxite the
equilibrium capacity rapidly increases with rising RH. Only molecular sieves provide
sufficient capacity when this parameter is initially already low. For instance, if the RH
is only 10 %, a zeolite Type A adsorbs about 20 % of water on the dry basis when
activated alumina can adsorb only less than 5 %. On the contrary, when the RH is as
high as 60-80 % molecular sieves are comparatively less efficient. A staged approach
(first alumina, then zeolites) is hence conceivable.
Temperature: The adsorption of water results in the evolution of heat set free in the
active adsorption zone. This heat raises the temperature of the bed and the gas,
decreasing the operating capacity as a result of the effect of temperature on
equilibrium. For example, for air dehydration with activated alumina after 2.5 hours of
operation and at the distance of about 0.3 m from the inlet, the increase of temperature
may reach 75 C comparing to the starting inlet conditions.
6. Desorption and Adsorbent Regeneration
Regeneration is conducted under conditions favorable for a shift of the adsorption equilibrium
towards desorption. Physical adsorption can be reversed rather easily. In practice,
regeneration is accomplished by one of the four following methods:
1. Vaporizing the adsorbate by heating the adsorbent with a hot purge gas or steam
(thermal-swing adsorption);
1. Reducing the pressure to low values and vaporize the adsorbate (pressure-swing
adsorption);
2. Inert purge gas stripping without effecting changes in temperature or pressure;
3. Displacement desorption by a fluid containing a more strongly adsorbing species.
The thermal-swing regeneration mode (Figure 8) is based on increasing the vapor pressure
exerted by the adsorbed phase with temperature. This makes the molecules to desorb until a
new equilibrium with the fluid phase is established. It is the simplest and the most common
way to regenerate an adsorbent. Its disadvantage, however, is in the considerable time needed
for a bed to acquire the regeneration temperature then cooling down again to the operating
temperature.

Figure 8: Schematic representation of pressure-swing and thermal-swing adsorption

For such treatment of activated carbon temperature should attain at least 150-200 C.
Manufacturers usually recommend the best regenerating temperature for their particular
adsorbent. Exceeding that temperature may accelerate the ageing processes, which cause
pores to coalesce and capacity to be reduced. Too low regeneration temperatures may result in
incomplete regeneration so that the effluent concentration in subsequent adsorption stages will
be higher than its design value.
The purge stream during thermal-swing may consist of either a hot gas or steam. Hot gas is
the preferred choice in case the adsorbate is a generic group of high boiling compounds, or
that no steam is available, or if the desorbed adsorbate is too difficult to separate from
condensed steam, because of foaming, emulsification, mutual dissolution, or other problems.
Saturated or slightly superheated steam provides the required heat of desorption and the low
adsorbate partial pressure required to reverse the equilibrium. Since it is often available at
industrial sites, steam is very convenient for regenerating saturated adsorbent. The steam
displaces the adsorbate, and is condensed jointly with it. A disadvantage of purging with
steam is the necessity to cool and also dry the regenerated load, at the end of the cycle and
prior to its reuse. The possibility to use steam for regeneration depends on the boiling point of
the adsorbate, the ease of separating the condensate (e.g. a mixture of toluene and water,

which separate in two phases), and the difficulty in treating the resulting aqueous effluent,
which always contains a trace of organics, etc.
Change of pressure is an alternative way to influence the equilibrium conditions (Figure 8).
Adsorption takes place at high pressure, and desorption of the charged adsorbent follows at
low pressure, therefore this cyclic process is called Pressure Swing Adsorption (PSA).
Because changes in pressure can be realized more rapidly than changes in temperature this
type of regeneration can be used with shorter cycle times, which allows smaller beds to be
used.
PSA is often used for other applications, such as for separating gases, e.g. air into oxygen and
nitrogen. Two different types of processes are extensively used:
1. When separation takes place on a zeolite adsorbent (Type 5A or 13X), the process is
characterized by the preferential adsorption of nitrogen under equilibrium conditions
and the maximum purity of the oxygen product is limited to 96-97 % due to the
unavoidable presence of argon. The process operates at two intermediate pressures
between the feed pressure and the exhaust pressure (usually atmospheric).
2. When a carbon molecular sieve (e.g. prepared from anthracite or hard coal by
controlled oxidation and subsequent thermal treatment) is utilized, the process depends
on a kinetically controlled separation. There is little difference in equilibrium values,
but a large difference in diffusivity, with oxygen being the most rapidly adsorbed
species. The high-pressure product in the carbon sieve process is now nitrogen with
purities as high as 99.99 %. Oxygen is set free either at low pressure during the purge
and blowdown, or during vacuum desorption. In the latter case, the cycle is commonly
operates between about 4 and 0.1 atmosphere.
The carbon sieve process is preferred to the zeolite process when a pure nitrogen product is
the primary requirement. On the contrary, it is inferior as a process to be selected for oxygen
production.
Another way of sorbent regeneration, is purging with a non-adsorbing inert gas (as in
chromatography) while operating at an essentially constant temperature and pressure. It is
applicable only when adsorbed species are weakly held, otherwise too high a volume of purge
gas is required. Since the desorbate concentrations are low, this method is used when recovery
is not a target.
Extracting adsorbents with a solvent, e.g. toluene or chlorinated solvents, is also a possibility
for regeneration, but its use seems to be strictly limited to the laboratory practice. Concerning
non-thermal regeneration, the attainable degrees of desorption, as well as the dosage of
regenerating agent required may need to be investigated. Several alternative adsorbents and
ways of regeneration may have to be tested.
Regeneration is normally economically desirable, because:
1. The majority of adsorbents are relatively expensive. Only low cost natural adsorbents
(e.g. clay minerals) are proposed for one-way use.
2. Desorption often allows recovery of a valuable product. It should be noticed, however,
that most industrial solvents are too cheap for making their recovery worthwhile.
Moreover, the reclaimed solvent has to be filtered and dried and may have assimilated
impurities.

3. Environmental considerations. These militate against a throwaway operating mode.

The main exception to this rule is that of activated carbon that was used to adsorb
dioxins, even though these can be in situ destroyed during high-temperature
regeneration. Recycling is also problematic when adsorbing
Sometimes the adsorption forces are too strong, and the adsorbed product undesirable (smelly
compounds, in odor control), or the vapor pressure too low to reverse the operation. In that
case custom regeneration is still possible using a high temperature treatment, leading to the
destructive removal of the adsorbate, which in the process is thermally decomposed or
oxidized.
The choice of regeneration method for a particular system should be based on the balance
between economic factors and technical possibilities.
7. Design of Adsorbers
7.1. General Principles
Certain physical and chemical properties of adsorbents are involved in all adsorber design
calculations. The usual starting constraint in such a design exercise is that even basic
properties governing the equilibrium and kinetics of the system adsorbent/adsorbate are not
widely available, e.g., in handbooks, property databases or vendor-supplied charts. The more
features a model takes into account, the more adjustable parameters are required, and the
more experimental data are necessary to evaluate them. The complexity of data necessary for
an accurate adsorber design is illustrated in Table 3:

EOLSS - ADSORBENTS AND ADSORPTION PROCESSES FOR

POLLUTION CONTROL
No

Themes

Groups
1.
data

II

Basic
adsorbent
properties

Application
considerations

Issues

Isotherm Hysteresis; pretreatment conditions;

multicomponent effects

2.
transfer
behaviour

Interface
character;
intraparticle
Mass diffusion; film diffusion; dispersion

Porosity;
pore
size
distribution;
specific
surface
area;
density;
3. Particle
particle size distribution; particle
characteristics shape; abrasion resistance; crush
strength; composition / stability;
hydrophobicity
1. Operating
conditions
2.
Regeneration
technique

Flow
rate;
feed
and
product
concentrations;
pressure;
temperature; desired recovery; cycle
time; contaminants
Thermal: steam or hot fluid; pressure
shift; chemical: acid / base / solvent;

3. Energy
requirements
4.
life

regenerant / adsorbate recovery or

disposal

Adsorbent
Attrition / swelling; aging

1.
Contactor Axial / radial flow; pulsed / fluidized
type
bed
2. Geometry

III

Number of beds; bed dimensions;

flow distribution; dead volumes

Equipment /
support
/
ballast;
flowsheet 3.
Column Bed
distribution; insulation
internals

flow

Instrumentation;
safety
/
4.
maintenance;
construction
materials;
Miscellaneous
operation, start-up, shut-down
Table 3: Adsorber design considerations
[Source:
Source:
adapted
from
http://www.adsorption.com/publications.htm. Knaebel, K. S. A "How to"
guide for adsorber design.]
It may be useful to start with a simple model using bounding assumptions, which are known
to be reasonably valid, so that the outcome will be reasonably correct. It is crucial to
recognize the particular characteristics of a situation and to apply a model, which is neither
unnecessarily complicated nor one that is oversimplified.
To account for complex phenomena that may be encountered in adsorbers, it is generally
desirable to have a documented and user-friendly software package (e.g., AdDesignS,
EnviroPro Designer), a comprehensive database, and a fast PC or workstation. Even with
such tools, it may require great effort and considerable time to account for the mass transfer
rates, deviations from ideal plug flow, heat effects and other aspects.
It should be also remembered that cost is involved in every design decision.
7.2. Fixed-bed Adsorber Design
7.2.1. Approaches and Methods
Since the design of a continuous countercurrent adsorption system is even more complicated,
the following recommendations are limited to the preliminary design of fixed-bed adsorbers.
There are two very different approaches:
1. The first one is based on theory. The breakthrough curve for a selected bed length is
calculated using an appropriate theoretical model solving differential mass balance
equations, for example:

1. Local equilibrium model (a very simplified case, neglecting all forms of diffusion,
applicable to favorable isotherm and restricted to isothermal plug flow) evaluates
the ideal breakthrough time as follows:

(2)
where: is the void space of the bed, L - length of the bed, - bulk density, Acs cross-sectional area of the empty bed, n - averaged adsorbent loading, Q - volumetric
flow rate, and C
solute concentration in the fluid.
2.

The Hougen-Marshall model (1947): employs a Henrys Law isotherm,

assuming that film diffusion resistance predominates and neglecting axial
dispersion. It results in the following set of equations used to predict the
breakthrough curve:

(3)

,
(4)

(5)

where: J is an integral function, called J-function, C the current concentration,

the number of transfer units, C0 the initial concentration, a
dimensionless time, s the density of the solid phase, keff an effective mass transfer
coefficient, A the slope of the isotherm, v the interstitial velocity, t time, and z an axial
distance coordinate taken as positive in the flow direction
3.

The Thomas model (1944): is similar to the previous one, except that it
incorporates the Langmuir isotherm.

, (6)
where: , are defined as above, r is separation factor, which equals [1/(1+bC0] for
a Langmuir isotherm.
4. The Length of Unused Bed (LUB) approach: assumes that the transition of
concentration from the initial value to that of the feed is only due to mass transfer
resistance. The concept of LUB is introduced already earlier in this Article.
As follows from these examples, the models often require knowledge of mass transfer
coefficient and the size of the mass transfer zone, diffusivity, adsorption isotherm, etc. If
breakthrough occurs earlier than desired, then the bed length should be taken larger by the
amount necessary to produce the acceptable breakthrough time.
2. The second approach is almost completely empirical. Tests on laboratory- or pilot-scale
processes are conducted in order to determine the breakthrough curves. Usually, for such
experimental studies of adsorption, equilibrium and rate gravimetric, columnar, or
diffusion cell reactors are used. For example, a set of adsorption columns of different
length is a suitable option. All important parameters such as the type of adsorbent, the
mesh size range, the initial concentration of the compound(s) to be adsorbed, the fluid
flow rate, etc. must be the same in the pilot runs as one would have in the full-scale
adsorber.
The following assumptions can be made in the fixed-bed analysis:
1. Isothermal adsorption of a single component
2. Negligible concentration gradient in the radial direction
3. Constant cross section and uniform properties of the adsorbent bed throughout the
column.
4. At a steady state the rate of mass transfer of adsorbate from the bulk fluid to the
external pellet surface is equal to the net rate of solid diffusion. This assumption is
made to eliminate the surface concentration from any equation involving the rate of
mass transfer, because it cannot be measured by standard techniques.
5.

7.2.2. Estimation of Pressure Drop

A cardinal feature in any fixed bed of particles is the head loss, experienced by the fluid in its
flow. Generally speaking, such a problem is tackled on the basis of a similarity approach and
dimensional analysis. The flow properties can be characterized by an appropriate form of the
Reynolds number Re as a dimensionless parameter:

, (7)

where Rp is the adsorbent pellets radius, f and are the density and viscosity of the fluid,
respectively, and Vs is the superficial velocity. The latter is defined as volumetric flow per unit
cross-section of the void bed: Vs = Q/S.
Low values of Re correspond to laminar flow and high values to turbulent or eddy flow. The
transition between the two regimes may be either abrupt as e.g., flow in an empty tube, or
gradual, as for the sedimentation of a single particle.
For the sake of simplicity, flow through the bed of an adsorbent can be roughly regarded as a
flow passing a large number of submerged uniform spherical particles of diameter d. When
there is no flow, the net gravitational force in the packed bed acts downward, but this force is
small when considering gases. When the flow enters at the bottom of the adsorber, the friction
forces exerted on the particles act upward and counterbalance the net gravitational force. This
causes a slight expansion of the bed, which gradually extends with rising upward flow rate.
When the fluid velocity is high enough to balance gravity, the friction force becomes
sufficient for lifting the particles. Theoretically, this point of balance represents the onset of
fluidization. The corresponding velocity is described as the minimum velocity for fluidization
and the frictional force can be expressed in terms of a friction factor fp, which is defined as:

(8)
where: p - pressure drop across the bed,
- the void space of the bed, d - particle
diameter (use for non-spherical particles an equivalent diameter), - fluid density v - fluid
velocity, g - gravitational acceleration = 9.81 m s-2, and L - length of the bed.
Ergun (1952) developed the following relationship for flow through packed beds of spherical
particles:

(9)
where K1 and K2 are constants, generally accepted to be 150 and 1.75, respectively, fluid viscosity.
Combining expressions (8) and (9), the following form of Ergun equation is obtained:

(10)

This relationship reveals the following information necessary for fluidized bed adsorber
design:
1. Value of pressure drop along the length of the packed bed given some fluid velocity
2. Factors influencing the pressure drop: packing size, length of bed, fluid viscosity and fluid
density
The Ergun equation (10) can be simplified depending on the turbulence of the fluid:
1. For low values of Reynolds number (Re < 1) it is called Kozeny-Carman equation and
it contains only the first component of the sum, linear in terms of velocity, v:

.
(11)
2.
The high Reynolds number asymptote (Re > 1000) of the same equation is called
after Burke-Plummer: only the second component of the sum is important, and the
expression is of the second order in terms of v:

(12)

The practical requirement to maintain an acceptable pressure drop at relatively high flow
rates means that in a large-scale adsorption process it is generally necessary to use relatively
large adsorbent particles: typically 1-3 mm in diameter.
In case of an adsorption treatment of liquid fluids, the Darcy's law (1856) is useful for
description of the rate of flow of the solvent through porous media:

,
(13)

where Q is the volume of a liquid passed in unit time, H - the height of the fluid on top of
the bed, S - the cross-section area of the bed, k - a coefficient depending on the nature of the
adsorbent and L - the thickness of the bed.
The conclusion from Darcys law is that the rate of viscous flow of homogeneous fluids
through isotropic porous media is proportional to, and in the direction of, the driving force.
7.3. Safety Measures
Adsorption units are prone to fire and explosion. This hazard is related to:
1. the usually flammable nature of the organic vapors being adsorbed
2. the use of activated carbon as an adsorbent.
Activated carbon beds have a tendency to develop hot spots, as a consequence of catalytic
oxidation of organic compounds or other exothermic processes, such as polymerization or
adsorption itself. Another cause is in clogging of the bed or other problems of misdistribution.
Conversely, hot-spot development may lead to desorption of organics, with local
concentrations above the lower explosion limit. While regenerating activated carbon,
formation of explosive mixtures of desorbate with flue gas is also possible, if the latter still
has appreciable oxygen content.
Activated carbon is often injected pneumatically, and after adsorbing the pollutants it is
separated in an electrostatic precipitator or a baghouse filter: the latter provides a
supplemental contact layer, which achieves a step increase in collection efficiency. Depending
on the system and the operating conditions (temperature, oxygen content of the gas to be
treated) the following precautions are required:
1. The activated carbon is diluted with an inert material, generally slaked lime, Ca(OH) 2.
Depending on the results from preliminary testing, only 5
12 % of
activated carbon is used.
2.
The operating temperature is carefully controlled.
3. Storage bins of spent material are emptied carefully, to avoid the onset of slow,
glowing combustion.
In fixed bed units an incipient run-away is best monitored via the content of carbon monoxide
in the off-gas. A steep rise should promptly trigger emergency measures to stop an incipient
run-away and fire.
Some classes of organic compounds have a problematic record for adsorption on activated
carbon; this is the case e.g. with cyclohexanone, and higher molecular weight - aldehydes,
ketones, aromatic amines and fatty acids. For example, ketones tend to polymerize upon
adsorption into certain activated carbons. Adsorption of acetone, methylethylketone,
methylisobutylketone, formaldehyde and acetic acid demands special precautions.

8. Conclusions
In this chapter some properties of conventional commercial adsorbents are analyzed and
attention is further devoted to the various technical means for conducting an adsorption
process, whether in the most usual configuration of a fixed bed adsorber, or in more
continuous arrangements. The design and operating problems of fixed bed adsorbers are
also briefly described. The use and development of adsorbents gains more momentum, as
air pollution codes become stringent. Currently, they are already State of the Art in
removing trace impurities, such as semi-volatile and volatile organic contaminants, as
well as volatile metals, mercury mainly.
Related Chapters
Click Here To View The Related Chapters
Glossary
Activated
carbon
Adsorbate
Adsorbent
Adsorption

:highly porous hydrophobic adsorbent, produced by activation with steam or a

chemical reagent of e.g. charred wood, peat, lignite, bituminous coal, etc.
:substance in the adsorbed state.
:solid phase adsorbing the compounds to be removed.
:partition process, fixing molecules from fluids onto the surface of a solid or
liquid.
Breakpoint :moment in the operation of fixed bed adsorbers, when the concentration of the
compound being adsorbed abruptly starts increasing at the outlet.
Breakthrough :conventionally, the minimum detectable or maximum allowable solute
concentration concentration in the effluent fluid.
Breakthrough :effluent concentration curve registered versus time in fixed bed adsorbers.
curve
Chemical
:adsorbent regeneration method using an extracting agent, such as a solvent,
regeneration acid or base, or another chemical for which the adsorbate has a high affinity.
Desiccants
:group of adsorbents suitable for gas dehydration due to their hydrophilic
properties.
Hydraulic
:volumetric flow rate expressed per unit of empty-bed cross-sectional area, for
loading
example m3 h-1 m-2 bed area.
Mass transfer :part of an adsorption bed, located downstream the saturated zone, in which
zone
the adsorption process actually takes place.
Molecular
:are crystalline silicates containing oxides of alkali- or alkali-earth metals with
sieves
chemical formula Me2/nOAl2O3xSiO2yH2O (n = valence) and characterized
(zeolites)
by a regular structure of the pores with their dimensions in the range of the size
of a molecule.
Police filters :an adsorbent bed, which is inactive under normal circumstances, but only
removes a brief surge of pollutants under emergency conditions (German
literature).
Pressure drop :difference of fluid pressure between inlet and outlet of a packed bed caused
by the resistance of adsorbent particles to a flow.
Pressure:method to regenerate an adsorbent by a periodic lowering of the pressure,
Swing
when adsorption takes place at high pressure and desorption at low pressure.
Adsorption

(PSA)
Regeneration
of an
adsorbent
ThermalSwing
Adsorption
(TSA)

:procedure to reverse the conditions required for adsorption, generally by using

a hot gas, or applying steam or a chemical agent to dislodge the adsorbate(s).
:commonly used method for adsorbent regeneration using desorption
temperatures that are higher than those used during the adsorption step of the
cycle.

Bibliography
1.

Cooney, D. O. (1999). Adsorption design for wastewater treatment. 190 pp., Boca Raton, USA: Lewis
Publishers. [This reference briefly describes several approaches to the design of granular carbon fixed-bed
adsorbers]

2.

Coulson, J. M.; Richardson, J.F. (1991). Chemical Engineering: particle technology and separation
processes. Vol. 2, 4th ed., 968 pp. London, UK:Pergamon Press. [In this textbook the basic description of
various types of adsorbers is given]

3.

http://www.adsorption.com/publications.htm. Knaebel, K. S. A How to guide for adsorber design.

[Gives a broad list of parameters, which must be considered while designing an adsorber]

4.

Kohl, A.L.; Riesenfeld, F. C. (1985). Gas purification, 4th ed., 900 pp. Houston, TX, USA: Gulf
Publishing Company. [This book describes dehydration and gas purification processes most commonly used
in industry; it covers not only treatment with adsorption, but also applying absorption and chemical
conversion]

5.

Perry, R. H.; Green, D. W. eds. (1997). Perrys Chemical Engineers' Handbook, 7th ed., New York, NY,
USA: McGraw-Hill.

6.

Ruthven, D. M. (1984). Principles of adsorption and adsorption processes. 433 pp., New York, NY,
USA: John Wiley & Sons. [This is the one of most fundamental overviews of all adsorption related issues;
contains several hundred references to the original publications; is one of the most cited references in
adsorption domain].

7.

Seader, J. D.; Henley, E. J. (1998). Separation process principles. 886 pp., New York, NY, USA: Wiley.
[This textbook for undergraduate students contains a very compact and well structured chapter on adsorption
theory and applications]

Biographical Sketches
Alfons Buekens was born in Aalst, Belgium; he obtained his M.Sc. (1964) and his Ph.D (1967) at Ghent
University (RUG) and received the K.V.I.V.-Award (1965), the Robert De Keyser Award (Belgian Shell Co.,
1968), the Krber Foundation Award (1988) and the Coca Cola Foundation Award (1989). Dr. Buekens was full
professor at the Vrije Universiteit Brussel (VUB), since 2002 emeritus. He lectured in Ankara, Cochabamba,
Delft, Essen, Sofia, Surabaya, and was in 2002 and 2003 Invited Professor at the Tohoku University of Sendai.
Since 1976 he acted as an Environmental Consultant for the European Union, for UNIDO and WHO and as an
Advisor to Forschungszentrum Karlsruhe, T.N.O. and VITO. For 25 years, he advised the major industrial
Belgian Bank and conducted more than 600 audits of enterprise.
Main activities are in thermal and catalytic processes, waste management, and flue gas cleaning, with emphasis
on heavy metals, dioxins, and other semi-volatiles. He coordinated diverse national and international research
projects (Acronyms Cycleplast, Upcycle, and Minidip). Dr. Buekens is author of one book, edited several books
and a Technical Encyclopedia and authored more than 90 scientific publications in refereed journals and more
than 150 presentations at international congresses. He is a member of Editorial Boards for different journals and
book series.
He played a role in the foundation of the Flemish Waste Management Authority O.V.A.M., of a hazardous waste
enterprise INDAVER, and the Environmental Protection Agency B.I.M./I.B.G.E. He was principal ministerial
advisor in Brussels for matters regarding Environment, Housing, and Classified Enterprise (1989). Since 1970 he

has been a Member of the Board of the Belgian Consumer Association and of Conseur, grouping more than a
million members in Belgium, Italy, Portugal, and Spain.
He is licensed expert for conducting Environmental Impact Assessments (Air, Water, Soil) and Safety Studies
regarding large accidents (Seveso Directive).
Nadezhda Zyaykina obtained her degree of an Environmental Engineer at The Mendeleev University of
Chemical Technology of Russia, Moscow (1999) with specialization in recuperation technologies of secondary
industrial materials. Then she followed a Master program in Human Ecology at Vrije Universiteit Brussel
(VUB), 1999
2001. This interdisciplinary study of environmental problems and environmental
management includes scientific, economic and regulatory approaches. Since 2001 she is doing research in the
frame of a PhD program in Applied Sciences, also at VUB. The research is dedicated to analysis of
adsorption/desorption properties of filter dust from various metallurgical processes in respect to dioxins and
related chlorinated aromatic compounds. The investigation is also related to formation of dioxins and their
surrogates by carbon gasification reactions.

To cite this chapter

A. Buekens and N. N. Zyaykina, (2005), ADSORBENTS AND ADSORPTION PROCESSES FOR
POLLUTION CONTROL, in Pollution Control Technologies, [Eds. Bhaskar Nath, and Georgi St. Cholakov], in
Encyclopedia of Life Support Systems (EOLSS), Developed under the Auspices of the UNESCO, Eolss
Publishers, Oxford ,UK, [http://www.eolss.net] [Retrieved July 5, 2007]