Vous êtes sur la page 1sur 26

ADSORPTION OF GASEOUS POLLUTANTS

A. Buekens and N. N. Zyaykina


Department of Chemical Engineering CHIS 2, Vrije Universiteit Brussel, Belgium
Li Xianwei
Institute of Multidisciplinary Research of Advanced Materials (IMRAM), Tohoku University,
Japan
Keywords: Adsorption isotherm, Adsorbent, Adsorbate, Diffusion, Internal mass transfer,
Knudsen diffusion, Surface diffusion.
Contents
1. Scope
2. Scientific Principles of Adsorption
3. Adsorption kinetics
4. Conclusions
Acknowledgement
Related Chapters
Glossary
Bibliography
Biographical Sketches

Summary
This chapter introduces the fundamentals of the adsorption of gaseous pollutants, including
laws of its equilibrium and kinetics. Various forms of internal diffusion are methodically
discussed. The chapter Adsorbents and Adsorption Processes for Pollution Control is devoted
to the main characteristics of adsorbents and associated processes with special emphasis on
technical aspects and general design guidelines in environmental applications, particularly in
gas cleaning and separation.

1. Scope
1.1. Definitions
Adsorption is an important unit operation in chemical engineering, even though it is much less
used than most of its counterparts. Its major handicap is in requiring an auxiliary solid, the
adsorbent, whereas most other operations involve only fluids that are much simpler to
circulate. This explains the batch operating mode and cyclic character of numerous adsorption
operations.
In a wider sense adsorption can take place on any interface between any set of two phases, on
either a fluid-fluid interface (liquid-gas, liquid-liquid), or a solid-fluid interface (solid-gas,
solid-liquid). Adsorption usually refers to the increase in concentration of the target
compound on the solid surface of a porous material, selected as adsorbent, since the latter
found most practical applications.
Adsorption is often confounded with absorption. Adsorption involves a concentration of a
chemical at the surface; absorption addresses a concentration in the bulk of the material.
Organics dissolve into plastics, hence absorb, rather than adsorb on the surface.
Adsorption is often reserved for cases in which all simpler methods have failed. Since it
achieves a high degree of removal of the pollutant (up to 99 % and sometimes even higher), it
is often used as a final polishing step, after removing all easier separable pollutants.
Generally a gas flow is treated by adsorption with the particular purpose of removing a trace
impurity or only a minor amount of pollutants: since the adsorption capacity of any adsorbent
is limited, it is most rational to apply this method only when the concentration of a target
compound is low. Adsorption is effective mainly when a contaminant ranges from 1 ppb to
1000 ppm. Such an approach allows increasing considerably the duration of operation of a gas
cleaning facility in its adsorption stage, before switching to the stages of regeneration of the
adsorbent (desorption, drying and cooling). When large amounts are to be cleaned, other
methods are mostly preferred, for the temporal ratio of loading the adsorbent versus its
regeneration becomes too unfavorable.
Adsorption is both a phenomenon of partition between phases and a mass transfer process, in
which molecules from the gas phase or from a solution are bound onto a solid surface in a
condensed layer (generally).
1.2. Applications
Adsorption is used for separation, cleaning or drying of both liquid and gaseous effluents in
such different fields as industry, medicine, environmental protection, etc. Some examples of
technical applications are:
Drying of gases, e.g. to obtain a low moisture content as required for
instrumentation or in cryogenic applications (see cryo-condensation
Condensation, Control of Carbon Monoxide and Volatile Organic
Compounds);
Catalytic processes, where adsorption is one first step of a complex
phenomenon leading to chemical conversion. Chemical reaction is
facilitated while molecules are adsorbed or more precisely chemisorbed
onto a surface. After a molecular rearrangement, the converted molecule
should be desorbed and enabled to leave the catalyst;
Gas/Solid chromatography, an analytical technique, in which compounds
are separated by their specific rate of migration through a partition
column, filled with suitable adsorbent (see Effluent Gas Monitoring);
Separation of linear and branched or cyclic hydrocarbons, present in oil
mixtures;
Food discoloring, e.g. in sugar production;
Purification of pharmaceutical products.
There are also many other ways, in which adsorption can be used. Purification of liquid
streams is based on the similar principles as that of gases: both proceed in a fixed bed
arrangement, but liquids can also be treated by stirring/mixing. In what follows, emphasis is
placed strongly on environmental applications, particularly the treatment of off-gases, and
much less - on industrial applications, such as:
Pressure Swing Adsorption, a cyclic process using molecular sieves for
separating gases, e.g. air into oxygen and nitrogen;
Drying of gas streams, on mineral adsorbents, such as alumina, molecular
sieves or silica gel;
Heterogeneous catalysis.
The above processes may also be applied in an environmental context. Some examples of
applications in a specific pollution control include:
Removal of odorous substances from landfill gas;
Recovery of volatile organic compounds (VOCs) and solvent vapors from
air;
Treatment of industrial off-gases containing products of incomplete
combustion (PAHs, dioxins and other halogenated aromatic compounds)
as well as CO, SO2 and CO2;
Decrease of NOx emissions from exhaust of vehicles;
Cleaning of wastewater from persistent or non-biodegradable pollutants,
etc.
Regardless of the specific application, the selection of the most suitable adsorbent and
designing the equipment are the main practical issues. A brief overview of technical
adsorption units and guidelines for their design are treated more fully in a second article.
1.3. Main Features
For each particular task an adsorbent is selected on the basis of its main features:
Capacity,
Selectivity,
Kinetics,
Regeneration possibilities,
Compatibility and
Cost.
Usually, these characteristics are to be established experimentally, which generally requires
extensive testing, since each adsorbent and adsorbate behaves differently. Experimental
results are also not to be extrapolated beyond the investigated domains.
Pressure drop over a packed bed is well predictable, from correlations based on the Darcys or
Kozeny-Carmans Law (see Adsorbents and Adsorption Processes for Pollution Control).
Adsorption as a rule is stronger, when the adsorbate has a lower vapor pressure and a higher
molecular weight. The presence of water vapor tends to reduce adsorptive capacity, water
molecules competing with the adsorbate for adsorption sites or surface. Polarity and
concentration levels also play an important role. In case different compounds are present in
the gas phase, adsorption may become a competitive process, so that stronger adsorbing
chemicals may replace previously adsorbed molecules. Conversely, weaker adsorbing
molecules may also displace strongly adsorbed ones, if the former are very numerous and the
residence time is sufficiently long for effecting such replacement.
Selectivity is strongly affected by polarity, vapor pressure, and the nature and structure of the
adsorbent. Physical and possibly chemical kinetics is another very important issue since it
controls the cycle time of an adsorption bed, markedly influences the shape of the
breakthrough curve and hence the effective time available to conduct the operation.
1.4. Transport Phenomena
The target compound travels from the bulk gas to the external surface (external diffusion),
then continuing its route inside the particle by a combination of molecular, surface, Knudsen
and other types of internal diffusion. Smaller pore diameter leads to higher internal surface,
but may retard or impede internal diffusion.
1.5. Regeneration
Regeneration and reuse of adsorbent are technically possible and desired in almost all cases,
unless the cost of adsorbent is negligible, as is the case with waste materials, put to use as an
adsorbent, or when the operation is close to a quarry of mineral adsorbents. The recovery of
adsorbate is also desirable, e.g. of solvents, unless there is no use for the compounds
separated, e.g. in case these are smells. The decision whether to employ such recovery and
reuse is based on several economic, environmental and technical considerations. However,
currently it is becoming uncommon to design a system in which a high-value adsorbent is
discarded after one use.
The regeneration can be performed either in-place, or off-site. Much depends on the duration
of the adsorption run: this is typically of the order of one day or some fraction thereof when
recovering solvents, of several months when using the adsorbent bed as a police filter.

2. Scientific Principles of Adsorption

2.1. Adsorption Forces


A simple picture regarding the nature of physical adsorption relates to the reduction in vapor
pressure, exerted by attraction forces between a surface and a condensable vapor.
Physical adsorption is the result of van der Waals interactions between adsorbate(s) and the
adsorbent. These forces are weak and characterized by a small change of enthalpy, which is
insufficient to lead to bond breaking. Physical adsorption does not involve a sharing or
transfer of electrons, so that all interacting species maintain their identity. The interactions
between the partial electric charges of polar molecules are responsible for creating attractive
contributions to these forces. Repulsive components prevent the complete collapse of matter
to nuclear densities.
Chemical adsorption or chemisorption is characterized by the superior strength of the
adsorbent adsorbate bond even approaching that of an ordinary covalent bond, so
that the chemical nature of gases or vapors may be significantly different in the adsorbed state
(cf. catalysis). There is no sharp distinction, however, between physical or chemical
adsorption, rather a somewhat gradual transition scaled by the strength of adsorption.
Moreover, there is no material distinction between physical and chemical adsorption in actual
applications or in their technical aspects.
The main features of the both types are as follows:

The spectroscopic, electron spin resonance and magnetic susceptibility


measurements are the most reliable way to identify the type or strength of
adsorption. They can confirm whether a transfer of electrons occurs and
chemical bonds are formed.
The simplest test criterion for distinguishing chemisorption from
physisorption is based on a comparison of the enthalpies of adsorption.
The heat of physical adsorption is normally below 20 kcal/mol (84
kJ/mol), but this value may be exceeded for adsorbents with very narrow
pores. The heat of chemical adsorption is approaching that of chemical
bonds and varies in a very wide range, starting at less than 40 kcal/mol
(168 kJ/mol), to more than 100 kcal/mol (418 kJ/mol). Thus, activation
energy is required for chemisorption.
Physical adsorption is not site-specific; adsorbed molecules are free to
cover the entire surface, so that the surface of solid adsorbents can be
measured by determining the adsorbent amount (B.E.T.-method: see under
Section 2.6). Chemisorption is restricted to specific surface sites so, that
chemisorbed molecules are no longer free to migrate about the surface.
These physical adsorption interactions are fully reversible, whereas
chemisorption is irreversible under specific operating conditions.
Desorption at the temperature levels required is accompanied by
decomposition of adsorbate.
Therefore, chemical sorption has certain advantages over its physical alternative in gas
cleaning. However, regeneration of adsorbent becomes more difficult and energy demanding
at the same time, and recovery of adsorbate is sometimes impossible, since destructive
desorption takes place at the temperature and operating conditions required.

2.2. Pore and Surface Characteristics


Each adsorbent is characterized by its structural properties, including its

Surface area (external, internal, total and specific),


Porosity, tortuosity, and pore size distribution.

Surface area plays a fundamental role in the physical adsorption of a specific adsorbent. For
understanding the various kinetic issues (see Section 3) it is important to distinguish between
external and internal surfaces. The former constitutes only the minor part of the total surface
area of an adsorbent and consists from an ideal geometric surface, i.e. that of an equivalent
sphere or cylinder and augmented with the surface represented by all those prominences and
cracks, which width is greater than their depth. The internal surface is associated with the
walls of pores that have narrow openings extending inward from the granule surface to the
interior of this granule. Another important parameter for the theory of adsorption is specific
surface area (as, m2 g-1), which is defined as the total surface area divided by the mass of the
adsorbent.
According to their size, pores are classified into three groups; each of these has its special
functions during adsorption:

Macropores (> 0.2 m) are relatively large channels allowing internal


transport to the sites of adsorption proper;
Mesopores (1.5 nm - 0.2 m) are responsible for further distribution of the
adsorbate throughout the porous structure and feature mono- and
multilayer adsorption;
Micropores (< 1.5 nm) are with dimensions in the range of the size of a
molecule. The total volume of micropores in industrial adsorbents usually
does not exceed 0.5 cm3 g-1 and constitute the most of the internal surface.
Adsorption takes place according to capillary condensation mechanism.
For the sake of comparison, a dioxin molecule, consisting of three condensed cycles, more
precisely two aromatic nuclei internally linked by two ether bonds, has dimensions of the
order of 0.7 to 1.4 nm. The surface area values for nil- to octachlorinated dioxins vary from
186 to 315 2. The molecular volume of a series of polyaromatic hydrocarbons (PAHs) lies
between ca. 128 to 261 3; their surface area ranges from ca. 156 to 282 2.
Porosity (, %) is the percentage of the total volume of the adsorbent that consists of pore
space. In other words, it reflects the portion of the free volume. Porosity can be related either
to a separate granule or to a layer of the adsorbent. The latter depends on the shape of the
granules and the density of packing of the layer.
Tortuosity, , (in German: Labyrinthfaktor, Zick-Zack-Faktor) reflects the shape of the pores
and affects the rate of molecular diffusion in the particle. It can be defined as a ratio of actual
pore length to the superficial diffusion path. Its value has to be determined for each kind of
adsorbent to estimate the pore diffusion coefficient. Available methods and models are (see
later in this article):

The pore diffusion model with chromatography method


The pore diffusion model with the Wicke-Kallenbach method
The combined diffusion model with the transient response method.

Since tortuosity is a structural parameter, it is independent of temperature and of the nature of


the diffusing species.

2.3. Equilibrium Data


The free, non-adsorbed gas and the adsorbed gas (adsorbate) are in a dynamic equilibrium that
can be moved by acting upon temperature, total pressure, and composition of the gas phase, in
particular the concentration of adsorbate and the presence or absence of vapors competing for
adsorption. The amount of adsorbate can be expressed using either surface concentration (in
mol m-2), or as surface coverage (, in %) defined as ratio of number of adsorption sites
occupied to the number of adsorption sites available.
The equilibrium partition between the free and adsorbed state depends on temperature and
pressure and can be described in terms of an empirical function q = f (P, T), which can vary
as:
q = f(P) Isotherm T = const
q = f(T) Isobar P = const where q is the
amount adsorbed.
q = f(T) Isostere Fixed surface coverage
An adsorption isotherm is the most widely used form for representing equilibrium data,
because it is the most convenient form both to obtain and to plot experimental data. Moreover,
this is the form in which theoretical treatments are most easily developed. The most common
types of isotherms and related theories are described below.
2.4. Langmuir Theory

2.4.1. Basic Assumptions


Langmuirs theory (1918) is based on the following assumptions:

The adsorbing surface area consists of a series of sorption sites that are
equivalent to each other
The ability of a particle to bind to a sorption site is independent of
whether or not nearby sites are occupied
Adsorbed molecules form at most a monolayer on the sorption surface.
The latter corresponds to a complete coverage.
Each of these assumptions is somewhat open to criticism. The Langmuir theory is most
popular in studies of catalysis and widely applied in a number of variations.

Figure 1: Monolayer and multilayer


2.4.2. Derivation of Equilibrium Conditions
In this derivation the dynamic equilibrium between the free and the adsorbed condition is
described as a chemical equilibrium between the two states.
Starting from equilibrium between on the one hand the adsorbate molecule G, on the other the
adsorbent site S, one finds:

. (1)
This equilibrium parameter K depends on temperature, not on concentration or pressure.
The adsorption sites (total concentration CT) are either occupied by G (CGS), or free from
adsorbent (CS). Hence:
CT = CS + CGS = CS (1 + K CG) , (2)
CS = CT / (1 + K CG). (3)
The adsorbate concentration is given by:

. (4)
There are two limiting cases of interest:

CG is small, approaching zero. The denominator thus tends towards one,


so that there is a linear relation between CGS and CG: CGS = K CT CG

CG is very large. The denominator thus tends towards K CG, so that there is
a limit value of CT for CGS: CGS = CT
The temperature-dependent equilibrium constant, K, follows the van t Hoff equation:

(5)

, (6)
where:
H - heat of adsorption,
- area of a site,
K0 - frequency factor (in principle, the frequency of collisions),
t0 - constant, about 10-12 to 10-13 second, and
Nav - Avogadros number.
Finally, the Langmuir adsorption isotherm based on this derivation is described by the
formula
, (7)
where fraction of surface coverage; bL = ka kd-1 Langmuir
constant; ka - a rate constant for adsorption, which is function of the number of available
unoccupied sites and a number of collisions of gas molecules with the surface; kd - a rate
constant for desorption, which is a function of the heat of desorption from the surface and
temperature.
The Langmuir isotherm predicts that a straight line should be obtained when is
plotted against pressure for relatively low values of P. In general this model is too simple for
really detailed agreement with experimental adsorption systems. However, a variety of
experimental adsorption data and catalytic processes has been found to fit the Langmuir
equation fairly well. Moreover, it provides a point of departure for many refinements
proposed later.

2.5. Freundlich Adsorption Isotherm


The Freundlich adsorption isotherm is a semi-empirical correlation; the validity of the
isotherm is strictly limited to the range of concentrations investigated:
q = K C 1/n (8)
where q is the amount adsorbed by a unit mass of the adsorbent; C is the equilibrium
concentration; K and n are the empirical constants dependent on the nature of solid and
adsorbate, and on the temperature. The dimensions of the experimental data must be
monitored and followed carefully.
In actual practice the Freundlich adsorption isotherm is more often used than Langmuir
theory, since it tends to yield better results, in particular for correlating adsorption data in the
aqueous phase. For many systems a good correlation results over a range of two or three
decades in concentration, with deviations at both the high side and the low side, which is
difficult to approach experimentally.

2.6. BET Adsorption Isotherm


Brunauer, Emmett and Teller in 1938 proposed another physical adsorption model, creating
the BET multilayer adsorption theory, which extends Langmuirs approach to the multilayer
coverage of the adsorbent:

, (9)
where
- Q0 is the temperature-independent constant, supposed to represent either a fixed number of
surface sites or some specific surface area,
- P is the saturation vapor pressure of the sorbate at the relevant temperature.

The equation works well in the range of the pressure ratios P/P of 0.05 0.35 and
its major advantages are:

It is easy to apply;
It serves as a base method for obtaining surface areas from adsorption
data;
It allows deriving a multi-component isotherm as an extension of the one-
component one.
BET-measurement is a standard test in characterizing adsorbents and catalysts. The most
traditional shortcoming of the BET equation is that it predicts too little adsorption at low
pressures and too much at high pressures.

2.7. Classification of Adsorption Isotherms


Adsorption isotherms can be grouped into five principal classes, as illustrated in Figure 2.

Figure 2: The five types of adsorption isotherms according to the classification of Brunauer
Type I is a characteristic Langmuir-type adsorption, which shows a linear increase for low
values of the partial pressure, followed by a monotonic asymptotic approach to a limiting
value that corresponds theoretically to the completion of a surface monolayer.
Type II is the most common found for physical adsorption on relatively open surfaces, where
adsorption proceeds from a first sub-monolayer to the final multilayer. The isotherm exhibits
a distinct concave-downward curvature at low relative pressure and a sharply rising curve at
high P/P. The point B at the knee of the curve signifies completion of an adsorbed
monolayer.
A type III isotherm represents a relatively weak adsorbate-adsorbent interaction, as
exemplified by the adsorption of water and alkanes on nonporous low-polarity solids such as
polytetrafluorethylene PTFE, well known under the DuPont de Nemours trade name Teflon. In
this case, the adsorbate does not effectively spread on the solid surface. It should be noted that
adsorption on surfaces is a potentially serious source of errors in environmental analysis of
semi-volatiles It was found that 85% of dioxins in water phase report to the glass surface of
the condensate container. Adding methanol to the water, thus enhancing dioxins solubility,
can significantly reduce this elevated value.
Type IV and V isotherms are characteristic of adsorption by capillary condensation into tiny
adsorbent pores, in which the adsorption reaches an asymptotic value as saturation pressure is
approached.
Adsorption of organic vapors on activated carbon is typically type IV, whereas adsorption of
water vapor on activated carbon is type V.
Figure 3 shows several adsorption isotherms for the same adsorbent-adsorbate system, but
measured at different temperatures. Such a graph, reflecting the wide range of operating
conditions, is necessary for engineering calculations.

Figure 3: Influence of temperature on sorption capacity


Isosteric heats of adsorption may be measured simultaneously with the loading and
composition of gaseous mixtures. The combined measurements at a single reference
temperature provide complete thermodynamic information on the equilibrium behavior of a
binary mixture. The components of the mixture are dosed alternately so that the individual
isosteric heats can be determined from two successive measurements.

3. Adsorption Kinetics
Adsorption takes place in a sequence of four consecutive steps describing the progression of a
molecule of adsorbate from the bulk of the fluid towards the adsorbate state. These steps are:

Movement in and with the bulk of the fluid


External mass transfer
Internal diffusion
Adsorption

The movement with the bulk of the fluid is very fast compared to the further external and
internal diffusion steps. The mass balance and flow characteristics in the adsorber establish
the bulk fluid concentrations (see Adsorbents and Adsorption Processes for Pollution
Control).
Adsorption on a solid adsorbent takes place in several stages as:
(1) External diffusion refers to the mass transfer by diffusion of adsorbate molecules from
the bulk fluid phase through an almost stagnant boundary layer, surrounding each
adsorbent particle, to the external surface of the solid.
(2) Internal diffusion is the transfer of adsorbate to the interior of the particle by migration
of adsorbate molecules from the relative limited external surface to the inner surface
of the pores within each particle of the adsorbent. This rate is controlled by the
diffusion of adsorbate molecules through the pores of the particle.
(3) The actual adsorption process, in which adsorbate molecules moving in the pores are
adsorbed from the vapor phase or from a solution onto the solid phase is considered to
be relatively fast, compared to the first two steps, taking only 10-8 10-9 s.
An instantaneous reversible reaction is assumed for the adsorption step, an empirical
equilibrium law locally relating the concentration of adsorbate and the adsorbate load at the
carbons surface.
Hence, local equilibrium is usually assumed between these two phases and migration is rate
controlling. However, it should be noted that it is still a matter of debate whether semi-
volatiles can still be desorbed from adsorbents. In many adsorbate/sludge systems it has been
noted that an ageing adsorbate becomes more difficult to desorb.

3.1. External Diffusion


Diffusion is an important transport phenomenon (see Control of Gaseous Emissions) arising
as a result of random molecular motion in a fluid, which tends to reduce differences in
concentration, since there will be more molecules moving from a region of high concentration
into an area of low concentration than vice versa.
The concentration in the bulk flow is lowered close to the particle, due to depletion of the
adsorbate.
A linear driving force model p or c, describes the rate of mass transfer through a diffusion
boundary layer, as in Ficks Law. The rate of transfer, NA, for the first stage can be expressed
by the following equation:

, (10)
where kf is the external film mass transfer coefficient, ap is the external surface area of the
particle, is the void fraction, b is the bulk density of the particle, and C, Cs are the
concentrations in the bulk of gas or solution and at the surface, respectively. As an alternative,
these concentrations may be represented as pressures p, ps.
Empirical equations are available to characterize the coefficient kf, both in adsorption and
catalysis (see formulas 17-23). They give kf as a function of the hydraulic behavior in the
vessel or in the surrounding of the particle, some physical properties of the solvent and the
molecular diffusion of the solute. The kf value is affected by particle size, but not by the
nature or the porosity of the adsorbent. Furthermore the molecular diffusion coefficient, Dmol,
is also needed for calculations.
These empirical equations include a series of dimensionless numbers, which allow
characterizing:

Flow properties:

- Reynolds number:

(11)

- the particle-related Reynolds number:

(12)
Transfer properties:

- the Nusselt number or Nu for heat transfer:

(13)
- the Sherwood number or Sh for mass transfer:

(14)
Physical properties:

- the Prandtl number Pr:

(15)
- the Schmidt number Sc uniting diffusion-related properties:

(16)
where:
Vs - the superficial velocity
f - the density of fluid
Rp - the adsorbent pellets radius
Dm - the molecular diffusivity
- the viscosity of the fluid
h - the heat transfer coefficient
Cp - the specific heat of the fluid at constant pressure
k - the thermal conductivity of the fluid
These parameters are essential in the design of fixed bed adsorbers.
According to different authors, the external mass transfer coefficient can be estimated using
the following equations:
Wakao and Funazkri:

. (17)
Petrovic and Thodos:

. (18)
Lightfoot et al.:

(19)
Chu et al.:

. (20)
Ranz and Marshall:

. (21)
Wilke and Hougen:

. (22)
Williamson et al.:
(23)

3.2. Internal Diffusion


The rate of adsorption in a bed of porous adsorbents is generally controlled by the second step
in the sequence, i.e. transport within the pore network. Ficks first law of diffusion for a
spherical particle is expressed as:

, (24)
where NA is the diffusion flux, D is the intraparticle diffusion coefficient, Cp is the
concentration in the pore space and r the particle radius.
Intraparticle diffusion in porous solids is about two orders of magnitude slower than in open
space, because:

only a fraction of the total volume, i.e. the pore volume, is available for gas diffusion;
pores are not necessarily oriented in the direction of the concentration gradient; rather,
they constitute tortuous channels;
part of the pores lead to a dead end, not supporting diffusion any further.

Intraparticle diffusion may occur by several different mechanisms:

Ordinary, molecular diffusion as based on the kinetic theory of gases and the
statistical movement of molecules, leading to the uniform spreading by a haphazard
progression halted and redirected by molecular collisions. Molecular diffusion is
important only in large pores, the diameter of which is much larger than the free path
length.
Knudsen diffusion, an atypical variant of the former, in which molecules move either
in a high vacuum or inside a channel so narrow that the distance between successive
molecular collisions, i.e. the free path length, is comparable with the channel diameter.
If the free path length exceeds the pore diameter the collision frequency with the wall
becomes larger than that with the same or different molecules. Knudsen diffusion rates
are pressure independent and vary with the square root of temperature.
Surface diffusion is characterized by molecular movement under such influence from
the channel walls that the desorbed condition becomes more plausible than the free
condition. Under such circumstances, it can be visualized that movement proceeds
along the surface and is directed either by the freedom of movement in the adsorbed
state, or by periodic desorption immediately followed by re-adsorption.
The diffusion regime can be derived by plotting the logarithm of the diffusion rate parameter
as a function of either temperature or pressure. The formula of Bosanquet was empirically
developed as a transition between those two domains.
The existence of surface diffusion was shown by Wicke and Kallenbach in 1941, using a
diffusion cell bearing their name. This device consists of a cell divided into two halves by
e.g., a zeolite partition wall. These two parts are respectively fed with a constant gas flow, of
nitrogen and carbon dioxide. As a consequence of diffusion some nitrogen enters into the
carbon dioxide, and vice versa. The diffusion rate is derived from a gas chromatographic
analysis of both gas flows. The rate of this third diffusion mechanism no longer depends on
the pore diameter, but only on the nature of gases and materials. Such diffusion can also be
observed using smooth surfaces.
It is required that the molecules adsorb on the surface and that their movement to a nearby,
void site requires less energy than adsorption. Under such conditions molecules become
capable of free movement along the surface, in a two-dimensional space. The direction of
motion is independent of concentration gradients, and only depends on geometry. The rate of
surface diffusion varies with the partial pressure of the adsorbate, but is independent of total
pressure. The temperature dependence is determined by the energy involved in switching
positions and by a statistical parameter related to probability to finding another location. The
contribution of surface diffusion is sizeable only for microporous substances, such as
activated carbon.
The relative importance of these three types of motion depends on the pore size distribution,
tortuosity, the sorbate concentration, and other conditions, such as temperature and pressure.
Molecular diffusion is the dominant diffusion process if pores are large and the fluid relatively
dense. If the fluid is a gas phase where density is low, or if the pores are quite small, or both,
the molecules collide with the pore wall much more frequently than with each other. This
process is known as Knudsen diffusion. Molecules adsorbed on a solid surface may evidence
considerable mobility. Transport by movement of molecules over a surface is considered as
surface diffusion.
The permeability is defined as a product of porosity derived by comparing the real and
apparent density and tortuosity.
For a gas-solid system, the effective molecular diffusivity, Dm,eff, can be calculated using a
Chapman-Enskog equation correlated with tortuosity, , and particle porosity, p, as follows:

, (25)
, (26)

, (27)
where:
Dm - molecular diffusivity,
- a function of /kbT,
M1, M2 - molecular weights,
= 12 - the Lennard-Jones force constant,
P - total pressure,
kb - Boltzmann constant, and
12 = (1 + 2) collision diameter from the Lennard-Jones potential, .
The pore diffusion coefficient can be obtained from correlations that use the molecular
diffusion coefficient, the porosity, and the tortuosity factor of the porous materials.

3.3. Diffusion Models


Mass transfer inside the particle involves two different phenomena:

Porous diffusion, where the adsorbate first diffuses in pores and is eventually
adsorbed, and
Surface diffusion, where the adsorbate first adsorbs, then diffuses from site to site.

Some authors proposed that internal mass transfer is only due to surface diffusion. The
required equations and assumptions are therefore:
1) Mass balance;
2) External mass transfer, governed by a linear driving force;
3) Internal mass transfer is only governed by surface diffusion.
It is assumed that:

At the interface, there is continuity in rate between external and internal diffusion.
At the interface, there is equilibrium between adsorbate concentration in the fluid and
adsorbate load on the surface.
The particle of adsorbent is spherical and homogeneous.

On the internal pore surface, an adsorbed molecule may hop along the surface when it attains
sufficient activation energy and when an adjacent adsorption site is available. Although the
mobility of the adsorbed phase will generally be much less than that in the fluid, the
concentration is also much higher. So a significant contribution to the flux is possible. It has
been shown that surface diffusion is very important when appreciable adsorption occurs. In
order to study surface diffusion, it is therefore necessary to eliminate the contribution from the
fluid phase. The procedure involves making measurements over a wide range of temperatures.
The flux through the gas phase is determined from the high temperature measurements, since
under these conditions the surface flux can be neglected. The flux at the lower temperature is
then found by extrapolation.
3.3.1. External Diffusion Model
This model pictures the diffusion of an adsorbate through an external film while its molecules
are moving from the bulk fluid flow towards the outer surface of a particle. Once arrived at
this surface, adsorption occurs instantaneously and equilibrium between adsorbate in the fluid
and that on the external surface is assumed to be established. The non-adsorbed material then
diffuses further into the pores in the adsorbed state. A material balance for spherical particle
results in the following mass transfer equations:

, (28)

,
where Ds is the surface diffusion coefficient, r is the radial coordinate, t is the time, and p
is the pellet density.
3.3.2. Pore or Internal Diffusion Model
In the pore diffusion model, it is assumed that the adsorbate diffuses into the pores in the fluid
phase and at some moment it ends being taken up by adsorption on the walls of the pores. In
most models it is assumed that there are uniformly distributed sinks for adsorption within the
pores. The molecule cannot diffuse along the walls of the pores in the adsorbed state but can
migrate only by desorbing first. If one assumes (1) a spherical particle, (2) dilute fluids, (3)
local equilibrium, then the equations, which describe the adsorbate mass transfer, are as
follows:

, (29)

,
where Dp is the pore diffusion coefficient, Cp0 is the initial solution concentration, and q0 is
the initial adsorbed amount on solid phase.
3.3.3. Combined Diffusion Model
Adsorbent particles are heterogeneous systems formed by a porous solid phase and fluid
phase filling the void fraction of the solid. The internal diffusion can be expressed by the three
possible mechanisms of diffusion simultaneously, i.e. molecular, Knudsen diffusion, and
surface diffusion. If one considers this combined parallel resistance within the adsorbent
pellet, a material balance for spherical particle results in the following partial differential
equations:
, (30)

, (31)

.
In the dilute solution range of an adsorption system having a favorable isotherm, Equation
(30) can be reduced to:

. (32)

3.4. Rate Controlling Step


The over-all rate of adsorption in porous adsorbents is generally controlled by the transport
steps within the pore network. This intra-particle transport in narrow pores is a very particular
diffusion process, since it simultaneously involves several mechanisms, such as

ordinary diffusion or molecular diffusion,


Knudsen diffusion, and
surface diffusion.

Their relative contribution depends on pore size, the sorbate concentration, its interactions
with the surface, and other conditions. In macropores there is still ordinary diffusion, with the
diffusion coefficient proportional to the length of free path and the Brownian molecular
velocity V:
D = 1/3 V. (33)
It has a typical order of magnitude 10-5 10-4 m2 s-1.
For pores with a diameter smaller than the length of free path Knudsen diffusion takes
over.

3.5. The Biot Number


Biot Number is a dimensionless ratio, which compares:

the initial external mass transfer for a non-loaded particle, and


the initial surface diffusion for a particle without boundary layer

Thus, for a given situation the largest possible external mass transfer rate is compared to the
largest possible internal diffusion rate.

(34)
where kg is the mass transfer coefficient, ri is the outlet radius of the pores inside a
spherical pellet, and De is effective diffusivity.
Traegner and Suidan (1989) considered 3 typical cases:

when Biot >> 100, surface and thus intra-particle diffusion is rate limiting, and the
rate is insensitive to kf variations
when Biot < 1, external mass transfer is rate limiting; rates are very sensitive to kf
and insensitive to Ds variations. The transport is controlled by film diffusion
There is also a transition zone, with 1 < Biot < 100, in which both mechanisms are
important.
Surface diffusion plays an important role in the transport of gases through beds of porous and
non-porous solid. Surface diffusivity in a spherical particle has been correlated with surface
coverage and heat of adsorption using the following correlations:

, (35)

where x = surface coverage;


D0 = surface diffusivity for zero surface coverage;
E = activation energy and
R = gas constant.

, (36)

, (37)

where a = constant and H = heat of adsorption, which is expressed as:

, (38)

where P is the vapor pressure. A plot of ln P versus T-1 can be analyzed graphically for the
slope at constant loading amount to obtain H.

3.6. Temperature Programmed Physisorption


Temperature Programmed Physisorption (TPP) is used to determine gas adsorption isotherms
based on one simple temperature-programmed experiment. The exit composition of an
adsorption column, through which a gas mixture is flowing, is monitored, whereby the
temperature of the column is varied in time. All adsorption isotherms can directly be
calculated from the exit gas composition response of one single run, with the help of an
axially dispersed plug flow model, assuming uniform temperatures in the column.
This method allows determining isotherms in a wide range of temperatures and compositions
in one simple experiment. Since TPP requires operation sufficiently close to local equilibrium,
the experimental conditions must be chosen so that mass and heat transfer limitations are
negligible.
A non-isothermal, non-isobaric Maxwell-Stefan model for gas adsorption kinetics in single
porous particles is used to describe responses to pressure and concentrations steps. This model
is valid in a broad range of operating conditions due to the contribution of three different mass
transport processes leading to pressure and concentration equalization: bulk and Knudsen
diffusion, viscous flow and surface diffusion. The unknown surface diffusion transport
parameter is determined by fitting the model to response curves measured in a microbalance.
This model is also used for the analysis of mass and heat transfer limitations in TPP systems
leading to practical criteria for local equilibrium in TPP experiments.

4. Conclusions

A survey is presented of the principles of physical and chemical adsorption, even though there
is no sharp distinction between both modes of adsorption and the technical application of both
are even harder to distinguish. Generally, rates of adsorption are controlled by mass transfer;
hence attention is paid to both external diffusion through an outer film in laminar flow and to
inner diffusion, proceeding by the usual molecular diffusion in large pores, by the less
tractable Knudsen type and surface-type diffusion in small pores, responsible for most of the
inner surface.

In Adsorbents and Adsorption Processes for Pollution Control the properties of adsorbents
and the ways of using them in actual adsorption systems will be considered. Adsorption is
also an essential step in heterogeneous catalysis, a process with numerous industrial
applications, e.g. in oil refining and petrochemical processes.

Acknowledgement

This chapter was prepared while one of the authors (Alfons Buekens) resided as a visiting
professor at Tohoku University (IMRAM). Special thanks are due to Prof. T. Nakamura for
his support and advice, and for creating pleasant working surroundings.

Related Chapters

Click Here To View The Related Chapters

Glossary

Activated carbon :highly porous hydrophobic adsorbent, produced by activation with steam
or a chemical reagent of e.g. charred wood, peat, lignite, bituminous coal,
etc.
Adsorbate :Substance in the adsorbed state.
Adsorbent :Solid phase adsorbing the compounds to be removed.
Adsorption :partition process, holding molecules from fluids onto the surface of a
solid or liquid.
Catalysis :chemical conversion facilitated by an external material that takes part in
the reaction, but is not consumed. Catalysis is generally a cyclic process,
featuring both catalyst consumption and regeneration.
Diffusion :spontaneous mass transfer process due to a random heat movement of
(molecular) atoms and molecules in a single- or multi-component medium.
Heterogeneous :catalysis involving a solid adsorbent as a catalyst.
catalysis
Hysteresis :deviation of adsorption and desorption values, as caused by e.g. capillary
condensation.
Internal surface :the (major) part of the surface of an adsorbent associated with the walls
of pores.
Regeneration of :procedure to reverse the adsorption conditions, generally by using a hot
adsorbent gas, or applying steam or a chemical agent.
Total surface :external and internal surface area of an adsorbent.
area
Tortuosity :structural property of an adsorbent reflecting the shape of its pores and
influencing upon adsorption kinetics. It can be defined as a ratio of actual
pore length to the superficial diffusion path.
van der Waals :forces responsible for physical adsorption, based on the attractive
forces interactions between the partial electric charges of polar molecules and
repulsive interactions that arise from the exclusion of electrons from
regions of space where the orbitals of species overlap.

Bibliography

Adamson, A. W. (1990). Physical chemistry of surfaces. 5th edition. 777 pp. London, UK: John Wiley & Sons.
[This is a synthesis of knowledge on surface structures and processes].

Atkins, P. W. (1990). Physical Chemistry. 995 pp. Oxford, UK: Oxford University Press. [This textbook provides
basic introduction to the adsorption related concepts].

Chiou, C. T. (2002). Partitioning and adsorption of organic contaminants in environmental systems. 257 pp.
Hoboken, NJ, USA: John Wiley & Sons. [This monograph gives overview of the processes by which nonionic
organic contaminants are sorbed to natural biotic and abiotic substances].

Coulson, J. M.; Richardson, J.F. (1991). Chemical Engineering, 4th ed., Vol. 2, 968 pp., London, UK: Pergamon
Press [Contains appropriate illustrative reference material: figures, graphs and photos related to adsorbent
structure, adsorption equilibrium, etc.].

Frost, A. C. (1981). Measurement of effective diffusivity from effluent concentration of a flow through diffusion
cell. AIChE Journal 27, 813-818. [Paper describing an experimental set-up for diffusivity measurements].

http://www.adsorption.com/publications/AdsorbentSel1B.pdf. Knaebel, K. S. Adsorbent selection. [a brief but


excellent guide for the procedure of an adsorbent selection]

Pankow, J. F. (1987). Review and comparative analysis of the theories on partitioning between gas and aerosol
particulate phases in the atmosphere. Atmospheric environment 21(11), 2275-2283. [Comprehensive discussion
of theories, applied to explain the partition of environmental contaminants between the gas and solid phases in
the atmosphere].

Ruthven, D. M. (1984). Principles of adsorption and adsorption processes. 433 pp. 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].

Siperstein, F.; Gorte, R. J.; Myers, A. L. (1999). A new calorimeter for simultaneous measurements of loading
and heats of adsorption from gaseous mixtures, Langmuir 15, 1570-1576[Paper describing an experimental set-
up for adsorption measurements].

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.

Xianwei Li was born in September, 1966 in The Jiangxi Province, China. In 1984 he entered the Undergraduate
Course of The Central South University of Technology, China. He was awarded the title of Good Postgraduate
Student in Three Aspects, and won the title of Excellent Graduate at graduation. His graduation thesis for a
Masters Degree was "Research of a New-type Efficient Composite Burden of Medium-sized & Small-sized
Blast Furnaces".
In 1991, after he received his Masters Degree, he entered the biggest iron and steel company of China, Baosteel.
From 1992 to 1997, he put forward some research projects which proved to be profitable.
From 1998 to 1999, he mainly studied the pisolitic ores, such as Yandi and Robe River. The aim of the study was
to increase the blending ratio of pisolite, and thus to reduce the cost of ore blending.
In 1998, he took part in the joint study projects between Baosteel and CVRD and BHP, and visited CVRD
Research Center and BHP Newcastle Laboratories.
From 2000 to 2003, he studied in the field of environment as a PhD student at Tohoku University, Japan. The
thesis title is Measurement and Calculation of Physicochemical Properties of Dioxin Congeners and Simulation
of Their Equilibrium in Thermal Processes. He obtained his PhD in 2003, and then returned to Baosteel as a
principal researcher in resources and environment.

To cite this chapter


A. Buekens, N. N. Zyaykina and Li Xianwei, (2005), ADSORPTION OF GASEOUS POLLUTANTS, 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 6, 2007]

Vous aimerez peut-être aussi