Modelling of Reactive Separation Processes

See
discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/223525930
Modelling of Reactive Separation Processes:

Reactive Absorption and Reactive Distillation
Article in Chemical Engineering and Processing March 2003
DOI: 10.1016/S0255-2701(02)00086-7
CITATIONS
READS
144
898
3 authors, including:
Eugeny Y. Kenig
Andrzej Grak
Universitt Paderborn
Technische Universitt Dortmund
211 PUBLICATIONS 2,369 CITATIONS
255 PUBLICATIONS 2,488 CITATIONS
SEE PROFILE
All in-text references underlined in blue are linked to publications on ResearchGate,

letting you access and read them immediately.
SEE PROFILE
Available from: Eugeny Y. Kenig

Retrieved on: 27 August 2016
Chemical Engineering and Processing 42 (2003) 157 /178

www.elsevier.com/locate/cep
Modelling of reactive separation processes: reactive absorption and

reactive distillation
C. Noeres, E.Y. Kenig *, A. Gorak
Department of Chemical Engineering, Chair of Fluid Separation Processes, Dortmund University, 44221 Dortmund, Germany
Received 6 July 2001; received in revised form 28 October 2001; accepted 30 October 2001
Abstract
In the last years chemical process industries have shown permanently increasing interest in the development of reactive separation
processes (RSP) combining reaction and separation mechanisms into a single, integrated unit. Such processes bring several
important advantages among which are increase of reaction yield and selectivity, overcoming thermodynamic restrictions, e.g.
azeotropes, and considerable reduction in energy, water and solvent consumption. Important examples of reactive separations are
reactive distillation (RD) and reactive absorption (RA). Due to strong interactions of chemical reaction and heat and mass transfer,
the process behaviour of RSP tends to be quite complex. This paper gives an overview of up-to-date reactive separation modelling
and design approaches and covers both steady-state and dynamic issues. These approaches have been applied to several different
RA and RD processes including the absorption of NOx , coke gas purification, methyl acetate synthesis and methyl tertiary butyl
ether (MTBE) synthesis.
# 2003 Elsevier Science B.V. All rights reserved.
Keywords: Reactive separation; Reactive distillation; Reactive absorption; Rate-based approach; Film reaction; Catalytic internals; Maxwell /Stefan
equations
1. Introduction
Chemical manufacturing companies produce materials based on chemical reactions between selected feed
stocks. In many cases the completion of the chemical
reactions is limited by the equilibrium between feed and
product. The process must then include the separation
of this equilibrium mixture and recycling of the reactants. Usually reaction and separation stages are carried
out in discrete equipment units, and thus equipment and
energy costs are added up from these major steps.
In recent decades, a combination of separation and
reaction inside a single unit has become more and more
popular. This combination has been recognised by the
chemical process industries for having favourable economics of carrying out reaction simultaneously with
separation for certain classes of reacting systems, and
many new processes (called reactive separations) have
been invented based on this technology [1 /6].
* Corresponding author. Tel.: /49-231-755-2357; fax: /49-231755-3035
E-mail address: [email protected] (E.Y. Kenig).
The most important examples of reactive separation

processes (RSP) are reactive distillation (RD) and
reactive absorption (RA). In RD, reaction and distillation take place within the same zone of a distillation
column. Reactants are converted to products with
simultaneous separation of the products and recycle of
unused reactants. The RD process can be efficient both
in size and cost of capital equipment as well as in energy
used to achieve complete conversion of reactants. Since
reactor costs are often less than 10% of the capital
investment the combination of a relatively cheap reactor
with a distillation column offers great potential for
overall savings. Among suitable RD processes are
etherifications, nitrations, esterifications, transesterifications, condensations and alcylations [1].
Similarly, in RA reactions occur simultaneously with
the component transport and absorptive separation.
These processes are predominantly used for the production of basic chemicals, e.g. sulphuric or nitric acids and
the removal of components from gas and liquid streams.
This can be either the clean up of process gas streams or
the removal of toxic or harmful substances in flue gases
0255-2701/03/$ - see front matter # 2003 Elsevier Science B.V. All rights reserved.
PII: S 0 2 5 5 - 2 7 0 1 ( 0 2 ) 0 0 0 8 6 - 7
158
C. Noeres et al. / Chemical Engineering and Processing 42 (2003) 157 /178
Table 1
Applications of RA processes
Aim of the process
Example
Application area
References
Removal of harmful substances

Retrieval/regeneration of valuable substances or nonreacted reactants
Production/preparation of particular products
Coke oven gas purification, amine washing

Solvent regeneration
Gas purification
Gas separation
[15 /17]
[18]
Sulphuric acid manufacturing, formaldehyde

preparation
Water removal from natural gas, air drying
Synthesis gas conditioning
Chemical synthesis
[19,20]
Gas drying
Gas separation/gas purification
[21,22]
[18]
Water removal
Conditioning of gas streams
and liquids. Absorbers or scrubbers where RA is

performed are often considered as gas /liquid reactors
[7]. If more attention is paid to the mass transport, these
apparatuses are rather treated as absorption units.
Optimal functioning of RSP depends largely on
relevant process design, properly selected column internals, feed locations, place of catalyst as well as on
sufficient understanding of the process behaviour. All
this unavoidably necessitates application of well working, reliable and adequate process models [8].
As both RD and RA occur in multiphase fluid
systems and involve gas/vapour and liquid phases,
with an important role of the interfacial transport
phenomena, they have much in common. For this
reason, the modelling methods for these both operations
are largely based on the same framework, especially
regarding the models of trays or packing sections. On
the other hand, RD and RA have a number of specific
features, which should be considered with care and
modelled by different approaches. Examples are provided by the presence of solid catalyst in several RD
processes, by different thermodynamics (frequently
electrolytes in RA), reaction kinetics as well as different
process configurations.
The objective of this paper is to give a comprehensive
overview of both basics and peculiarities of RA and RD
modelling and design. A detailed description covers
mass and heat transfer, reaction kinetics including
reaction-mass transfer coupling, as well as steady-state
and dynamic modelling issues. The achievements in the
theoretical description are illustrated with a number of
our own case studies which have been supported by the
results of laboratory-, pilot- and industrial-scale experimental investigations.
2. RSP characteristics
2.1. Reactive absorption
RA represents a process in which a selective solution
of gaseous species by a liquid solvent phase is combined
with chemical reactions. As compared with purely
physical absorption, RA does not necessarily require
elevated pressure and high solubility of absorbed

components: because of the chemical reaction, the
equilibrium state can be shifted favourably resulting in
enhanced solution capacity [9]. Most of the RA processes involve reactions in the liquid phase only, in some
of them both liquid and gas reactions occur [10,11].
Usually the effect of chemical reactions in RA
processes is advantageous only in the region of low
gas-phase concentrations due to limitations by the
reaction stoichiometry or equilibrium [12]. Further
difficulties of RA applications may be caused by the
reaction heat through exothermic reactions and by
relatively difficult solvent regeneration [13,14]. Most of
the RA processes are steady-state operations, either
homogeneously catalysed or auto-catalysed. Some important industrial applications of RA are given in Table
1.
2.2. Reactive distillation

RD, the combination of chemical reaction and
distillation in a single column, is one of the most
important industrial applications of the multifunctional
reactor concept. Recently, it has drawn considerable
attention because of its advantages, especially for
equilibrium-limited reactions. Several reviews have
been published in the last decade which give an excellent
introduction to and overview of RD processes [1,23 /26].
RD processes offer several advantages, e.g.:
. increased yield due to overcoming of chemical and
thermodynamic equilibrium limitations;
. increased selectivity through suppression of undesired
consecutive reactions;
. reduced energy consumption via direct heat integration in case of exothermic reactions;
. avoidance of hot spots by simultaneous liquid
evaporation;
. separation of close boiling components.
These advantages result in reduced capital investment
and operating costs. However, the application of RD is
somewhat limited by constraints, like, e.g.
159
Table 2
Applications of RD processes (hom., homogeneously catalysed; het., heterogeneously catalysed)
Reaction type
Synthesis
Catalyst
References
Esterification
Methyl acetate from methanol and acetic acid

Methyl acetate from methanol and acetic acid
Ethyl acetate from ethanol and acetic acid
Butyl acetate from butanol and acetic acid
Ethyl acetate from ethanol and butyl acetate
Diethyl carbonate from ethanol and dimethyl carbonate
Acetic acid and methanol from methyl acetate and water
MTBE from isobutene and methanol
ETBE from isobutene and ethanol
TAME from isoamylene and methanol
Cumene from propylene and benzene
Diacetone alcohol from acetone
Bisphenol-A from phenol and acetone
Monosilane from trichlorsilane
Mono ethylene glycol from ethylene oxide and water
4-Nitrochlorobenzene from chlorobenzene and nitric acid
hom.
het.
no data
hom.
hom.
het.
het.
het.
het.
het.
het.
het.
no data
het.
hom.
hom.
[27]
[28,29]
[30]
[31]
[32]
[33]
[34]
[35,36]
[37]
[38]
[39]
[40]
[41]
[42]
[43]
[44]
Transesterification
Hydrolysis
Etherification
Alcylation
Condensation
Dismutation
Hydration
Nitration
. common operation range (temperature and pressure)

for distillation and reaction;
. proper boiling point sequence: the key component
should be a top or a bottom product, undesired side
or consecutive products should be medium boiling
components;
. difficulties in providing proper residence time characteristics.
RD processes can be divided into homogeneous ones,
either auto-catalysed or homogeneously catalysed, and
heterogeneous processes, in which the reaction is
catalysed by a solid catalyst. The latter, often referred
as catalytic distillation (CD), permits an optimum
configuration of the reaction and separation zones in a
RD column whereas expensive recovery of liquid
catalysts may be avoided. However, it should be
mentioned that CD shows general drawbacks of heterogeneous catalysis like, e.g. limited catalyst life time due
to deactivation usually caused by undesired reaction
products and increased operating temperature.
Up to now, there is a rather limited number of RD
processes realised on industrial scale. Table 2 gives a
short overview of possible RD applications.
2.3. Internals for RSP

The design of internals for RSP is more severe than
for conventional non-reactive counter-current gas/
vapour/liquid processes [24 /26]. Feasibility and efficiency of a RSP under study strongly depend on the
appropriate choice of internal characteristics like liquid
residence time distribution (RTD), liquid hold-up,
separation efficiency and pressure drop. Even if column
internals for homogeneous RSP are identical to noncatalytic column internals, modifications are often
necessary to meet optimum reactor design demands. If

the use of solid catalysts is necessary, specific bifunctional internal structures capable of combining the
separation and catalytic functions, the so-called catalytic
internals, are needed.
The design of catalytic internals for RSP is presently
based on two general concepts, namely immobilisation
of commercial catalyst pellets and catalytic activation of
conventional internals for gas/vapour/liquid contactors
(see Fig. 1).
When using small solid catalyst particles under
counter-current flow conditions, the following additional requirements should be met:
. uniform liquid flow in the catalyst bed without
stagnant zones and liquid bypassing;
. wide gas/vapour and liquid loading ranges without
flooding;
. limited catalyst abrasion;
. possibility of variable catalyst amount;
. simple catalyst exchange.
One of possible ways to immobilise the catalyst pellets
is to use wire gauze mesh from which objects of different
shape can be manufactured. These objects together with
the catalyst itself form certain structures, which should
meet the demands of optimum hydrodynamic behaviour
under gas/vapour /liquid counter-current flow conditions. Examples of such structures are different random
packings made of cylindrical baskets [45], wire gauze
boxes [46], wire mesh bales [47], etc.
Additionally to this type of random packings, a
significant effort has been made in the last decade
towards the development of structured sandwich-type
packings. Here, catalyst pellets are immobilised between
two sheets of corrugated wire gauze forming a sandwich.
A parallel arrangement of these layers results into open
160
Fig. 1. Internals for RSP.
and closed channels for the gas/vapour and liquid flow,

respectively, which show hydrodynamic characteristics
similar to those of traditional structured packings, e.g.
reduced pressure drop and optimum flow conditions
within a wide operating range (KATAMAX [48],
KATAPAK-S [49]). Furthermore, it is possible to
combine catalyst sandwiches with conventional corrugated wire gauze layers. This hybrid sandwich structure
meets the demands of flexible catalyst amount and
separation efficiency (MULTIPAK [50], KATAPAKSP [51]).
Wire gauze envelopes filled with catalyst pellets can
also be placed on trays, either across the tray [52] or in
the downcomer section [53]. Catalyst beads can be
immobilised like in a fixed-bed reactor between two
non-reactive distillation trays [54,55] as well as in an
external side stream reactor [56,57].
The second concept for catalytic column internals is
the use of catalytically active structures instead of those
filled with catalyst. Such structures are either carrier
supported catalysts or solid catalytic structures. Carrier
supports can be coated with any kind of catalyst (e.g.
GPP-rings and some specific structured packing [58],
KATAPAK-M [59]). Moreover, it is possible to
develop solid catalytic structures without any carrier.
The so-called BP-rings, for example, are produced by
polymerisation in an annular gap [58], whereas the
monolithic structures are made by extrusion of catalytic
material [60]. Up to now, catalytic internals are primarily used for RD processes. However, there is also some
potential for their application in RA processes.
3. RSP modelling
3.1. General aspects
In general, RSP are of a multicomponent nature. This
means that they are qualitatively more complex than
similar binary processes. Thermodynamic and diffusional coupling in the phases and at the interface which
are of multicomponent character are accompanied by
complex chemical reactions [8,61,62]. As a consequence,
to describe such processes adequately, specially developed mathematical models are required capable of
taking into consideration column hydrodynamics,
mass transfer resistances and reaction kinetics.
The feasibility of a sophisticated RSP model application is questionable if the problems of plant design,
model-based control and on-line process optimisation
have to be solved. In this case a reasonable model
reduction should be applied [63].
An overview of possible modelling approaches for RA
and RD is shown in Fig. 2. A process model consists of
sub-models for mass transfer, reaction and hydrodynamics whose complexity and rigour vary within a
broad range. For example, mass transfer between the
gas/vapour and the liquid phase can be described on the
grounds of the most rigorous rate-based approach, with
the Maxwell /Stefan diffusion equations, or it can be
accounted for by the simple equilibrium stage model
assuming thermodynamic equilibrium between the both
phases.
Homogeneously catalysed RD, with a liquid catalyst
acting as a mixture component, and auto-catalysed RD
161
Fig. 2. Modelling approaches for RA and RD.
present essentially a combination of transport phenomena and reactions taking place in a two-phase system
with an interface. In this respect they are very similar to
RA, and, generally, reaction has to be considered both
in the bulk and in the film region. For slow reactions, a
reaction account in the bulk phase only is usually
sufficient.
For heterogeneous systems, it is generally necessary to
additionally consider the phenomena in the solid
catalyst phase. In this case, very detailed models using
intrinsic kinetics and covering mass transport inside the
porous catalyst arise (see, e.g. [64 /66]). However, it is
often assumed that all internal (inside the porous
medium) and external mass transfer resistances can be
lumped together [29,67]. The catalyst surface is then
totally exposed to the liquid bulk conditions and can be
completely described by the bulk variables. This results
in the so-called pseudo-homogeneous models. By this
name a similarity to a simpler homogeneous bulk-phase
reaction is reflected (cf. Fig. 2). If the reaction (either
homogeneous or heterogeneous) is fast, it can be
described using the data on chemical equilibrium only.
Modelling of hydrodynamics in multiphase gas/
vapour/liquid contactors includes an appropriate description of axial dispersion, liquid hold-up and pressure
drop. The correlations giving such a description have
been published in numerous papers and are collected in
several reviews and textbooks (e.g. [68,69]). Nevertheless, there is still a need for a better description of
hydrodynamics in catalytic column internals, this is
being reflected by research activities in progress [70,71].
The description of thermodynamics and chemical
reaction kinetics of the RSP is very process-specific,
and hence its general detailed discussion would constitute a separate issue. Therefore, in this paper we will
only give a brief discussion of these topics in the context

of the following case studies. Further related details can
be found in [72,73].
In order to model large industrial reactive separation
units, a proper sub-division of a column apparatus into
smaller elements is usually necessary. These elements
(the so-called stages) are identified with real trays or
segments of a packed column. They can be described
using different theoretical concepts, with a wide range of
physicochemical assumptions and accuracy.
3.2. Equilibrium stage model
In the last decades, modelling and design of RSP has
usually been based on the equilibrium stage model.
Since 1893, as the first equilibrium stage model was
published by Sorel [74], numerous publications discussing various aspects of model development, application
and solution have appeared in the literature [75]. The
equilibrium stage model assumes that each gas /vapour
stream leaving a tray or a packing segment is in
thermodynamic equilibrium with the correspondent
liquid stream leaving the same tray or segment. In case
of RSP the chemical reaction has to be additionally
taken into account, either via reaction equilibrium
equations, or via rate expressions integrated into the
mass and energy balances.
Different model assumptions reflect the relation
between the mass transfer and reaction rates. The
definition of the Hatta-number representing the reaction
rate in reference to that of the mass transfer helps to
discriminate between very fast, fast, average and slow
chemical reactions [76,77].
If a fast reaction system is considered, the RSP can be
satisfactorily described assuming reaction equilibrium.
162
Here, a proper modelling approach is based on the nonreactive equilibrium stage model, extended by simultaneously using the chemical equilibrium relationship. An
alternative approach proposed by Davies and Jeffreys
[32] includes two separate steps. First, the concentrations and flow rates of the leaving streams are calculated
with the simple non-reactive equilibrium stage model.
Afterwards, the leaving concentrations are adapted by
using an additional equilibrium reactor concept. However, the latter approach does not consider direct
interactions between the chemical and thermodynamic
equilibrium.
Such descriptions can be appropriate for instantaneous and very fast reactions. Contrary, if the chemical
reaction is slow, the reaction rate dominates the whole
process, and, therefore, a reaction kinetics expression
has to be integrated into the mass and energy balances.
This concept has been used in a number of studies, both
for RD (e.g. [78,79]) and RA (e.g. [80,81]) process
simulation.
In practice, RSP rarely operate close to thermodynamic equilibrium. Therefore, some correlation parameters like tray efficiencies or HETS-values have been
introduced to adjust the equilibrium-based theoretical
description to the reality. For multicomponent mixtures,
however, this concept often fails, since diffusion interactions of several components result into unusual
phenomena like osmotic or reverse diffusion and mass
transfer barrier [82,83]. These effects cause a strange
behaviour of the efficiency factors, which are different
for each component, vary along the column height and
show a strong dependency on the component concentration [61,83,84].
The acceleration of mass transfer due to chemical
reactions in the interfacial region is often accounted for
via the so-called enhancement factors [19,77,85]. They
are either obtained by fitting experimental results or
derived theoretically on the grounds of simplified model
assumptions. It is not possible to derive the enhancement factors properly from binary experiments, and
significant problems arise if reversible, parallel or
consecutive reactions take place.
geous since there is a broad spectrum of correlations

available in the literature, for all types of internals and
systems. For the penetration/surface renewal model,
such a choice is limited, and therefore, in this work the
two-film model is used.
In the two-film model (Fig. 3), it is assumed that all of
the resistance to mass transfer is concentrated in thin
films adjacent to the phase interface and that transfer
occurs within these films by steady-state molecular
diffusion alone. Outside the films, in the bulk fluid
phases, the level of mixing is so high that there is no
composition gradient at all. This means that in the film
region we have one-dimensional diffusion transport
normal to the interface.
Multicomponent diffusion in the films is described by
the Maxwell /Stefan equations, which can be derived
from the kinetic theory of gases [90]. The Maxwell /
Stefan equations connect diffusion fluxes of the components with the gradients of their chemical potential.
With some modification these equations take a generalised form in which they can be used for the description
of real gases and liquids [61]:
di
n
X
xi NLj xj NLi
j1
cLt ij
i 1; . . . ; n
where di is the generalised driving force:

di
xi dmi
RT dz
i 1; . . . ; n
(2)
Similar equations can be also written for the gas/

vapour phase.
Thus the gas/vapour /liquid mass transfer is modelled
as a combination of the two-film model presentation
and Maxwell /Stefan diffusion description. In this stage
model, the equilibrium state exists only at the interface.
The film thickness represents a model parameter,
which can be estimated using mass transfer coefficient
correlations. These correlations govern the mass transport dependence on physical properties and process
hydrodynamics and are available from the literature
(see, e.g. [61,77,91]).
3.3. Rate-based approach

A more physically consistent way to describe a
column stage is known as the rate-based approach
[61,86,87]. This approach implies that actual rates of
multicomponent mass and heat transfer and chemical
reactions are taken into account directly.
Mass transfer at the gas/vapour/liquid interface can
be described using different theoretical concepts [8,61].
Most often the two-film model [88] or the penetration/
surface renewal model [19,89] are used, whereas the
model parameters are estimated via experimental correlations. In this respect the two-film model is advanta-
(1)
Fig. 3. Two-film model for a differential packing segment.
3.3.1. Balance equations

The mass balance equations of the traditional multicomponent rate-based model (see, e.g. [61,62]) are
written separately for each phase. As chemical reactions
take place in the fluid phases, the steady-state balance
equations should include the reaction source terms:
d
B I
0 (LxBi )(NLi
a RBLi fL )Ac ;
dl
0
d
dl
B I
(GyBi )(NGi
a RBGi fG )Ac ;
i 1; . . . ; n
i 1; . . . ; n
(3)
(4)
Eqs. (3) and (4) represent the mass balances for

continuous systems (packed columns). For discrete
systems (tray columns), the differential terms transform
to finite differences, and the balances are reduced to
algebraic equations.
If chemical reactions take place in the liquid phase
only (this is valid for most of RD processes), the gas/
vapour-phase balances simplify to:
0
d
dl
B I
(GyBi )NGi
a Ac ;
i 1; . . . ; n
(5)
The bulk-phase balances are completed by the summation equation for the liquid and gas/vapour bulk
mole fractions:
n
X
xBi 1
(6)
yBi 1
(7)
i1
n
X
i1
The volumetric liquid hold-up fL depends on the gas/

vapour and liquid flows and is calculated via empirical
correlations (e.g. [69]). For the determination of axial
temperature profiles, differential energy balances are
formulated including the product of the liquid molar
hold-up and the specific enthalpy as energy capacity.
The energy balances written for continuous systems are
as follows:
0
0
d
0
(LhBL )(QBL aI RBL fL DHRL
)Ac
dl
d
0
(GhBG )(QBG aI RBG fG DHRG
)Ac
dl
(8)
(9)
If the dynamic process behaviour has to be considered, Eqs. (3) /(5), (8) and (9) become partial differential
equations including derivatives of the hold-up in respect
to time (see more details in Sections 4.2 and 4.3).
3.3.2. Mass transfer and reaction coupling in the fluid

film
The component fluxes NBi entering into Eqs. (3) /(5)
are determined based on the mass transport in the film
163
region. As the key assumptions of the film model result

in the one-dimensional mass transport normal to the
interface, the differential component balance equations
including simultaneous mass transfer and reaction in the
film are as follows:
dNLi
dz
RLi 0;
i 1; . . . ; n
(10)
Eq. (10) which is generally valid for both liquid and

gas phases represent nothing but a differential mass
balance for the film region with the account of the
source term due to the reaction. To link this balance to
the process variables like component concentrations,
some additional relationship often called constitutive
relations (see [61]) are necessary. For the component
fluxes NI , these constitutive relations result from the
multicomponent diffusion description (Eqs. (1) and (2))
and, for the source terms, from the reaction kinetics
description.
The latter strongly depends on the specific reaction
mechanism, stoichiometry and presence or absence of
parallel reaction schemes [85]. The rate expressions for
Ri usually represent non-linear dependences on the
mixture composition and temperature. Specifically for
the coupled reaction-mass transfer problems like Eq.
(10), it is always essential whether the reaction rate is
comparable with that of diffusion or not [76,77]. Eq.
(10) should be completed by the boundary conditions
relevant to the film model. These conditions specify the
values of the mixture composition at both film boundaries. For example, for the liquid phase:
xi (z 0)xIi ;
xi (z dL )xBi ;
i 1; . . . ; n
(11)
Combining Eq. (10) with the boundary conditions,

Eq. (11), written in a vector form, and using the
constitutive relations like Eqs. (1) and (2), we obtain a
vector-type boundary value problem which permits the
component concentration profiles to be obtained as
functions of the film co-ordinate. These concentration
profiles, in turn, allow to determine the component
fluxes. Thus the boundary value problem describing the
film phenomena has to be solved in conjunction with all
other model equations.
The composition boundary values entering into Eq.
(11) represent external values for Eq. (10). With some
further assumptions concerning the diffusion and reaction terms, this allows an analytical solution of the
boundary value problem, Eqs. (10) and (11), in a closed
matrix form (see [62,92]). On the other hand, the
boundary values need to be determined from the total
system of equations describing the process. The bulk
values in both phases are found from the balance
relations, Eqs. (3) and (4). The interfacial liquid-phase
concentrations xIi are related to the relevant gas-phase
concentrations yIi by the thermodynamic equilibrium
164
relationships and by the continuity condition for the

molar fluxes at the interface [61,92].
Due to the chemical conversion in the liquid film, the
molar fluxes at the interface and at the boundary
between the film and the bulk of the phase differ. The
system of equations is completed by the conservation
equations for the mass and energy fluxes at the phase
interface and the necessary linking conditions between
the bulk and film phases (see [8,61]).
Generally, all these considerations are also valid for
the gas/vapour film phase provided that reactions occur
there [92]. Both analytical and numerical solution of the
coupled diffusion-reaction film problem are analysed at
full length in [93], their particular applications are
considered in the following Sections 4.1 and 4.2.
3.3.3. Non-ideal flow behaviour in catalytic column
internals
The mass balances (Eqs. (3) and (4)) assume plug flow
behaviour for both the gas/vapour and the liquid phase.
However, real flow behaviour is much more complex
and constitutes a fundamental issue in multiphase
reactor design. It has a strong influence on the reactor
performance, for example via backmixing of both
phases, which is responsible for significant effects on
the reaction rates and product selectivity. Possible
development of stagnant zones results in secondary
undesired reactions. To ensure an optimum model
development for CD processes, we performed experimental studies on the non-ideal flow behaviour in the
catalytic packing MULTIPAK [94].
The experimental results confirm that the fluid flow in
the packing deviates from the plug flow behaviour (Fig.
4). Calculated axial dispersion coefficients are about
104 /102 m2/s, this is several orders of magnitude
larger than that of molecular diffusion (Fig. 5). Therefore, in the investigated operating range non-ideal
Fig. 5. Axial dispersion coefficients of the catalytic packing MULTIPAK (dC /0.1 m) calculated with the ADM model [94].
mixing effects are rather caused by hydrodynamic than

by molecular diffusion effects. Calculated Bodenstein
numbers for the packing are one order of magnitude
smaller than for fixed-bed reactors, which may be
caused by two effects: the occurrence of stagnant zones
in the catalyst layer and liquid bypassing due to the
hybrid structure of the catalytic packing [94].
The rate-based models suggested up to now do not
take liquid backmixing into consideration. The only
exception is the non-equilibrium cell model for multicomponent RD in tray columns presented in [95]. In this
work a single distillation tray is represented by a series
of cells along the vapour and liquid flow paths, whereas
each cell is described by the two-film model. Using
different number of cells in both flow paths allows to
describe various flow patterns. However, a consistent
experimental determination of necessary model parameters (e.g. cell film thickness) appears difficult, whereas
the complex iterative character of the calculation
procedure in the dynamic case limits the applicability
of the non-equilibrium cell model.
A substantially more promising approach is represented by the so-called differential models like the axial
dispersion model (ADM) [96] as well as the piston flow
model with axial dispersion and mass exchange (PDE)
[97]. Experimental studies show that the ADM gives an
appropriate description of the non-ideal flow behaviour
of the liquid phase in catalytic packings (see Fig. 4) [94].
Applying the ADM to cover this non-ideality, the mass
balances (Eq. (3)) transform to the following equations:
0
Dax @ 2
@
B I
(LxBi ) (LxBi )(NLi
a RBLi fL )Ac ;
2
@l
uL @l
(12)
i 1; . . . ; n
Fig. 4. Comparison between the experimental RTD curve of the
catalytic packing MULTIPAK (dC /0.1 m), the ADM and the PDE
model [94].
A thorough investigation of the influence of flow nonideality in catalytic packings on the process behaviour of
specific CD processes is an objective of our current

studies whose results will be published in the near future.
4. Case studies
4.1. Absorption of NOx
4.1.1. Chemical system
The reactive system considered is a basic one in the
production of nitric acid as well as in some other
industrial processes [11]. It consists of ten components
including air (N2, O2), water (H2O), oxyacids of nitrogen (HNO2, HNO3), and nitrogen oxides (NO, NO2,
N2O, N2O3, N2O4). The components are involved in
simultaneous, parallel, and consecutive reactions occurring in both phases. The reactions are of high orders and
most of them are exothermic.
Reaction kinetics is described by the scheme suggested
in [98] and modified in [99]. This scheme involves eight
reactions and can be regarded as the most extensive
reaction system so far. The gas-phase reactions are
governed by the following equations:
2NOO2 0 2NO2
DHR0 114 kJ=mol
NONO2 l N2 O3
DHR0 39:9 kJ=mol
0
2NO2 l N2 O4
DHR 57:2 kJ=mol
3NO2 H2 O l 2HNO3 NO
DHR0 35:4 kJ=mol
(R1)
(R2)
(R3)
(R4)
whereas the corresponding equations for the liquid

phase are:
2NO2 H2 O 0 HNO2 HNO 3
DHR0 10:72 kJ=mol
N2 O3 H2 O 0 2HNO2
DHR0 3:99 kJ=mol
N2 O4 H2 O 0 HNO2 HNO3
DHR0 5:03 kJ=mol
3HNO2 l HNO3 H2 O2NO
DHR0 7:17 kJ=mol
(R5)
(R6)
(R7)
(R8)
The liquid-phase reactions are valid for nitric acid

concentrations below 34 wt.%. In the case of higher
nitric acid concentrations reactions (R5), (R6) and (R7)
become reversible. The oxydation of NO (R1) is the
slowest reaction in this system. Therefore, the total gasphase hold-up in absorbers can be determined using the
kinetic data of this reaction [100]. The other gas-phase
reactions are instantaneous reversible reactions.
4.1.2. Process set-up
Measurements of an industrial NOx absorption
process schematically shown in Fig. 6 were described
in [101]. The absorption plant represents a sequence of
165
four units used for the removal of nitrogen oxides from

the waste gas of an adipin acid factory. Each unit is
separated by a metal plate into two sections. In fact
there are eight columns joined together as a countercurrent absorption plant. This plant is operated at
atmospheric pressure. The columns 1 /7 have a pump
around for cooling of the liquid. The diameter of each
column is 2.2 m, its height is 7 m. The packing height is
3.2 m. The packing consists of 35 mm INTALOX
ceramic saddles.
The liquid feeds entering columns seven and eight are
low concentrated nitric acids. The liquid product has a
concentration of HNO3 of about 35 wt.%. The gas feed
has a concentration of NOx of about 60 000 vppm. A
quarter of NOx is NO, the rest is NO2.
4.1.3. Modelling peculiarities
In terms of the concentration vector, Eq. (10) is a nonlinear differential equation of the second order. The
boundary value problem, Eqs. (10) and (11), is usually
solved numerically. However, it is also possible to use
another approach employing an analytical treatment.
The analytical solution of the linearised boundary value
problem in the film region is obtained in [62]:
x sinhf[C](dL z)gsinh1 f[C]dL gxI
sinhf[C]zgsinh1 f[C]dL gxB
(13)
where:
[C]([D]1 [K])0:5
(14)
Matrix [D ] results from the transformation of the

Maxwell /Stefan Eq. (1) to the form of the generalised
Ficks law [83]. This matrix is generally a function of the
mixture composition and is assumed constant along the
diffusion path [83]. The direct expressions for the
elements of the diffusion matrix [D ] can be found, for
example, in [61].
Matrix [K ] follows from the linear approximation
suggested by Wei and Prater [102]:
R $[K]x
(15)
Eq. (15) represents one of the best approaches to the

modelling of complex reaction systems providing a
satisfactory representation for many rate processes
over the entire range of reaction and linear approximations for most systems in a sufficiently small range (see,
e.g. [77,103 /105]). Eq. (15) has gained widespread
acceptance in various chemical and reactor engineering
areas [106] and is recommended for use in the modelling
of reactive separation operations [8,105]. The analytical
solution, Eqs. (13) and (14), can be of a particular value
when large industrial reactive separation units are
considered and designed.
Differentiating Eq. (13) give simple analytical expressions for the component fluxes with regard to the
166
Fig. 6. Absorption plant consisting of four units (eight columns).
homogeneous reaction in the fluid films (see [92]). The

methods of determination of the reaction matrix [K ] are
considered in [93,102,104,105,107].
4.1.4. Model parameters
In order to calculate the multicomponent diffusion
matrices [D ], the binary diffusivities in both phases
should be known. The film thickness, which represents
an important model parameter, is estimated via the mass
transfer coefficients [61,83]. The binary diffusivities and
mass transfer coefficients are calculated from the
correlations summarised in Table 3.
The correlations of Onda et al. [109] and Billet [68] are
valid for various mixtures and packings and cover both
absorption and distillation processes. The correlation of
Kolev [112] is obtained for RA and certain random
packings. In general, the mass transfer coefficient
correlations need to be compared with one another
and validated using experimental data. This shows, in
particular, the way the mass transfer correlations
influence the concentration profiles of the components
and other relevant process characteristics.
Nitric acid is a strong electrolyte. Therefore, the
solubilities of nitrogen oxides in water given in [114]
and based on Henrys law are utilised and further
Table 3
Binary diffusion coefficients and mass transfer coefficients
Phase
Binary diffusion
coefficient
Mass transfer coefficient correlation
Gas
Fuller et al. [108]
Onda et al. [109]; Billet [68]; Wehmeier

(see [110])
Onda et al. [109]; Billet [68]; Kolev
[112]; Mika [113]
Liquid Siddiqi and Lucas

[111]
corrected by using the method of van Krevelen and

Hoftijzer [76] for electrolyte solutions. The chemical
equilibrium is calculated in terms of liquid-phase
activities. The local composition model of Engels [115]
based on the UNIQUAC model is used for the calculation of vapour pressures and activity coefficients of
water and nitric acid. Multicomponent diffusion coefficients in the liquid phase are corrected for the nonideality as suggested in [61].
4.1.5. Results and discussion
The sensitivity analysis performed in [116] proves that
the suggested model provides qualitatively correct
behaviour of the concentration profiles. For the simulation of the industrial absorption process shown in Fig. 6
the following correlations ensuring the most reliable
results are selected: the rate constant of reaction (R1)
(the slowest and hence the most important reaction in
the system) according to [117]:
kp1 217347:91041:156T 1:3605T 2
(16)
the liquid-side mass transfer coefficient according to

[112], the gas-side mass transfer coefficient according to
Wehmeier (see [110]).
Figs. 7 and 8 illustrate how the suggested model work.
Fig. 7 shows a comparison of the simulated and
measured gas-phase concentrations of NO and NO2
throughout the whole absorption plant. The zigzag form
of the simulated concentration profiles results from
switching different sections of each single column (see
[92]). Good agreement between experimental and simulation results can be readily observed here.
In Fig. 8, experimental and simulated liquid-phase
concentrations of HNO3 and HNO2 throughout the
absorption plant are demonstrated. They also match
167
The reactions including CO2 obey first and second

order kinetics, whereas the other reversible reactions are
based on simple proton transfers and are, therefore,
regarded as instantaneous by the corresponding mass
action law equations. The formation of bicarbonate ions
(HCO3) takes place via two different mechanisms. The
rate of the direct reaction between carbon dioxide and
hydroxyl ions (the most important step) is given by:
r1 k1 (T)cOH (cCO2 ceq
CO2 )
Fig. 7. Experimental and simulated gas-phase concentrations of NO

and NO2 throughout the absorption plant.
Fig. 8. Experimental and simulated liquid-phase concentrations of

HNO3 and HNO2 throughout the absorption plant.
each other. Only in the first two columns we have larger

deviations between experiments and simulated results.
This can be attributed to the fact that at high concentration of HNO3 reactions (R5), (R6) and (R7) assumed
as irreversible reactions convert to reversible ones, the
data on their rate constants is lacking.
4.2. Coke gas purification
The selective absorption of coke plant gas contaminations results from a complex system of parallel liquidphase reactions.
Instantaneous reversible reactions:

NH3 H2 O l NH
4 OH
H2 SH2 O l HS H3 O
HCNH2 O l CN H3 O
2

HCO
3 H2 O l CO3 H3 O

H3 O OH l 2H2 O
(R9)
(R10)
(R11)
(R12)
(R13)
Finite-rate reversible reactions:

CO2 OH l HCO
3

CO2 2H2 O l HCO
3 H3 O
CO2 NH3 H2 O l H2 NCOO H3 O
(R14)
(R15)
(R16)
(17)
where cCO2eq represents the molar concentration at

chemical equilibrium [20]. Usually the reaction between
CO2 and water is very slow and hardly contributes to the
total rate of reaction of carbon dioxide. Nevertheless, in
this work it has been considered to be of the first order
with respect to CO2, since the reaction kinetics depends
on the carbonation ratio [118].
The absorption rate of carbon dioxide increases in the
presence of amines or ammonia. Therefore, the reaction
kinetics of NH3 and CO2 has been considered in the
model equations, too:

1
r3 k3 (T) cCO2 cNH3 eq cH2 NCOO cH3 O
(18)
K3
The rate constant as function of temperature has been
determined according to [118]. The coefficients for the
calculation of the chemical equilibrium constants in this
system of volatile weak electrolytes are taken from [119].
The CO2 absorption is hindered by a slow chemical
reaction by which the dissolved carbon dioxide molecules are converted into the more reactive ionic species.
Therefore, when gases containing H2S, NH3 and CO2
are contacted with water, the H2S and ammonia are
absorbed much more rapidly than CO2 and this
selectivity can be accentuated by optimising the operating conditions [15]. Nevertheless, all chemical reactions
are coupled by hydronium ions and additional CO2
absorption leads to the desorption of hydrogen sulfide
and decreases the scrubber efficiency.
Pilot plant experiments have been carried out at real
process conditions in the coke plant August Thyssen
(Duisburg, Germany). The DN 100 pilot column (Fig.
9) was made of stainless steel and equipped with about 4
m of structured packing (Sulzer MELLAPAK 350Y),
three liquid distributors and a digital control system.
Several steady-state experiments have been compared
with the simulation results and supported the design
optimisation of the coke gas purification process [16].
Coke oven gas mainly consists of a mixture of carbon
monoxide, hydrogen, methane and carbon dioxide. It is
contaminated with a variety of organic and inorganic
compounds which have to be separated in absorption
columns before its further use as a synthesis gas. The
168
namely the electroneutrality, which has to be met in

each point of the liquid phase:
n
X
xi zi 0
(21)
i1

Thermodynamic non-idealities are considered both in
the transport equation, (Eq. (10)) and in the equilibrium
relationships at the phase interface. If electrolytes are
present, the liquid-phase diffusion coefficients should be
corrected to account for the specific transport properties
of electrolyte systems.
The thermodynamic equilibrium at the gas /liquid
interface is usually described as follows:
Fig. 9. H2S absorber used in the ammonia hydrogen sulfide circulation
scrubbing process for the coke oven gas purification.
objective of these scrubber units is a selective removal of

H2S, NH3 and HCN, whereas competing reactions of
the major impurity CO2 have to be suppressed. An
example of innovative purification processes is the
ammonia hydrogen sulfide circulation scrubbing
(ASCS), in which the ammonia available from the raw
gas is used instead of an external solvent [15].
In [120] a purely numerical approach to the solution
of this complex RA problem has been suggested. The
liquid film is considered as an additional balance region,
in which reaction and mass transfer occur simultaneously. Therefore, the reactions are considered both
in the liquid bulk-phase mass balances, Eq. (3), and in
the differential balances for the liquid film, Eq. (10).
To be able to describe the presence of electrolytes in
the system, the electrical driving force needs to be
additionally taken into account [61]. Therefore, the
gradient of the electrical potential 98 is introduced
into the generalised driving force di :
di
xi 1 @mi
F 1 d8
xi zi
;
RT dL dh
RT dL @h
i 1; . . . ; n
(19)
In dilute electrolyte systems, the diffusional interactions can usually be neglected, and the generalised
Maxwell /Stefan equations are reduced to the Nernst/
Planck equations (Eq. (20)):

c D
dxi
F d8
NLi Li Li:eff
xi zi
xi NLn ;
(20)
RT dh
dL
dh
i 1; . . . ; n
where n is the solvent index. The consideration of the
electrical potential requires an additional condition,
yIi Ki xIi ;
i 1; . . . ; n
(22)
where the distribution coefficient Ki comprises fugacities

in both phases and activity coefficients in the liquid
phase. For the considered system, the values of Ki are
determined from the three-parametric ElectrolyteNRTL method. This model is based on the local
composition concept and satisfactorily represents physical interactions of true species in aqueous multicomponent electrolyte systems.
The liquid-phase diffusion coefficients are found with
the Nernst/Hartley equation [121], which describes the
transport properties in weak electrolyte systems. The
gas-phase diffusion coefficients are estimated according
to the Chapman/Enskog/Wilke /Lee model [72]. The
correlations for the mass transfer coefficients are taken
from [122].
Several steady-state simulations have been performed
with the aim to analyse the influence of numerical and
physicochemical parameters, beginning with a single
stage and ending with a column simulation. Different
film and packing section discretisations, several mass
transfer and hydrodynamic correlations, and different
driving forces and diffusion models have been thoroughly tested. The most sensitive components appeared
to be those involved in finite-rate reactions, especially
CO2. Furthermore, the influence of the electrical potential gradient on the axial concentration profiles has been
investigated, which is found to be significant for ionic
components [120,123].
The coupling of reaction and mass transfer in the
liquid film region has also been studied and significant
changes in the species concentration profiles due to the
film reaction have been established. The molar fluxes
calculated with the effective diffusion approach differ
only slightly from those obtained by the Maxwell /
Stefan equations without the consideration of general-
ised driving forces. This result is as expected for dilute

solutions and allows to reduce the model complexity for
the process studied [124].
The model optimised in respect to the numerical
parameters and physicochemical properties has been
validated against experimental data, whereas the axial
concentration and temperature profiles for both phases
demonstrated a good agreement (Fig. 10). It has been
also found that the simulations of the scrubber based on
the equilibrium stage model extended by the chemical
reaction kinetics yield results completely inconsistent
with the experimental studies, namely, the selectivity
towards H2S and HCN absorption cannot be reflected
(Fig. 10). In this case, the film reaction constitutes an
essential element of the rate-based approach, which has
to be considered in the model. As a result, the only
feasible simplification is represented by a linearisation of
the film concentration profiles including the implementation of the average reaction kinetics in the liquid film
region [124].
4.2.6. Dynamic modelling

The steady-state modelling is not sufficient if one
faces various disturbances in RA operations (e.g. feed
variation), or tries to optimise the start-up and shutdown phases of the process. In this case, a knowledge on
dynamic process behaviour is necessary. Further areas
where the dynamic information is crucial are the process
control as well as safety issues and training. The
dynamic modelling can also be considered as the next
169
step toward the deep process analysis, which follows the

steady-state modelling and is based on its results.
In the dynamic rate-based stage model molar hold-up
terms have to be considered in the mass balance
equations, whereas the change of both, the specific
molar component hold-up and the total molar hold-up,
are taken into account. For the liquid phase, these
equations are as follows:
@
@t
ULi
@
@l
B I
(LxBi )(NLi
a RBLi fL )Ac ;
(23)
i 1; . . . ; n
ULi xBi ULt ;
i 1; . . . ; n
(24)
The gas hold-up can often be neglected due to the low

gas-phase density, and the component balance equation
reduces to Eq. (4) (see also [120]).
The dynamic formulation of the model equations
requires a careful analysis of the whole system in order
to prevent high index problems during the numerical
solution [125]. As a consequence, a consistent set of
initial conditions for the dynamic simulations and a
suitable description of the hydrodynamics have to be
introduced. For instance, pressure drop and liquid holdup must be correlated with the gas and liquid flows.
The dynamic behaviour of the coke gas purification
process has been investigated systematically in
[120,123,126,127]. For instance, local perturbations of
the gas load and its composition have been observed.
Fig. 11 demonstrates the response to a sudden increase
of the gas flow by 20% and its H2S load by 100%. As
Fig. 10. Axial liquid bulk concentration profiles for the H2S scrubber: comparison between experimental and simulation results based on different
model approaches.
170
expected, the H2S load increases along the column

height in the gas-phase. In the lower part of the absorber
the change is more significant than at the top. The new
steady-state is achieved after about 30 min which
confirms the small time constants of the simulated pilot
plant absorber.
4.3. Methyl acetate synthesis
The synthesis of methyl acetate from methanol and
acetic acid is a slightly exothermic equilibrium-limited
liquid-phase reaction:
CH3 OH(CH3 )COOH l (CH3 )COO(CH3 )H2 O
(R17)
DHR0 4:2 kJ=mol
The low equilibrium constant and the strongly nonideal behaviour which gives rise to the binary azeotropes
methyl acetate/methanol and methyl acetate /water
make this reaction system interesting as a possible RD
application [27]. Therefore, the methyl acetate synthesis
has been chosen as a test system and investigated in a
semi-batch RD column. Since the process is carried out
under atmospheric pressure, no side reactions in the
liquid phase occur [128].
4.3.2. Process set-up and operation
A batch distillation column with a column diameter of
100 mm and a reactive packing height of 2 m (MULTIPAK I ) in the bottom section and an additional meter
of conventional packing (ROMBOPAK 6M ) in the
top section has been used. The flow sheet of the column
is shown in Fig. 12.
At first, the distillation still is charged with
methanol */the low boiling educt */and heated up under
total reflux until steady-state conditions are achieved. At
Fig. 11. Dynamic change of the H2S gas-phase concentration along

the column after a sudden increase of the gas flow and its H2S load.
Fig. 12. Flow sheet of the RD column for methyl acetate synthesis.
this moment, acetic acid */the high boiling educt */is fed
above the reaction zone to the second distributor. After
30 min the reflux ratio is turned from infinity to two and
the product withdraw at the top of the column begins.
During the column operations, the liquid-phase concentration profiles along the column and the temperature profiles are measured. For the determination of the
liquid-phase composition two methods are applied
simultaneously. On the one hand, samples are taken
and analysed by gas chromatography. On the other
hand, an online-NIR-spectrometer is used to determine
the concentration without taking any samples [129].

The model is based on the two-film theory and
comprises the material and energy balances of a
differential element of the two-phase volume in the
packing. The classical two-film model shown in Fig. 3 is
extended here to consider the catalyst phase (Fig. 13). A
pseudo-homogeneous approach is chosen for the catalysed reaction (see also [130,131]), and the correspondent overall reaction kinetics is determined by fixed bed
experiments [28]. This macroscopic kinetics includes the
influence of the liquid distribution and mass transfer
resistances at the liquid /solid interface as well as
diffusional transport phenomena inside the porous
catalyst.
The reaction kinetics is integrated into the mass
balances, and the liquid hold-up, as accumulation
term, is accounted for simultaneously, like in Eqs. (23)
and (24). Similar as in the dynamic simulation of the
coke gas purification process, the vapour hold-up is
Fig. 13. Extended two-film model for a differential packing segment

with heterogeneous catalyst.
neglected, and the vapour-phase component balance

equation reduces to Eq. (4).
Ranzi et al. [132] found that the full energy balances
including the accumulation term have to be considered
in order to predict correct dynamic process behaviour.
Therefore, the differential dynamic energy balance for
the liquid phase is applied as follows:
@EL
@t
@
@l
0
(LhBL )(QBL aI RBL fL DHRL
)Ac ;
n
X
l
QBL L (TLB T I )
NLi hLi ;
dL
i1
(25)
i 1; . . . ; n
Similar to the mass balance equation, the vapourphase energy balance simplifies to Eq. (9).
Experimental studies were carried out to derive
correlations for mass transfer coefficients, reaction
kinetics, liquid hold-up and pressure drop for the new
catalytic packing MULTIPAK (see [28,29]). Suitable
correlations for ROMBOPAK 6M are taken from [91]
and [133]. The vapour /liquid equilibrium is calculated
using the modification of Wilson method [28]. For the
vapour phase, the dimerisation of acetic acid is taken
into account using the chemical theory to correct
vapour-phase fugacity coefficients [134]. Binary diffusion coefficients for the vapour phase and for the liquid
phase are estimated via the method proposed by Fuller
et al. and Tyn and Calus, respectively (see [72]). Physical
properties like densities, viscosities and thermal conductivities are calculated from the methods given in [72].
Heat losses through the column wall are measured at
pilot scale.
Figs. 14 and 15 show the liquid-phase compositions
for the reboiler and condenser as functions of time.
After column start-up, the concentration of methanol
decreases continuously, whereas the distillate mole
fraction of methyl acetate reaches about 90%. A
171
comparison of the rate-based simulation (with the

Maxwell /Stefan diffusion equations; in the following
denoted as reference model) and experimental results for
the liquid-phase composition at the column top and in
the reboiler demonstrates their satisfactory agreement.
Fig. 16 shows the simulation results obtained with
different modelling approaches, including the reference
model, the effective diffusion model and an equilibrium
stage model, after an operation time of 10 000 s. Both
the reference model and effective diffusion model show
similar results. The equilibrium stage model is able to
describe the process behaviour qualitatively. This is in
contrast to the RA processes (see above, Sections 4.1
and 4.2) and can be explained by the low reaction rate
dominating the whole process kinetics.
4.4. Methyl tertiary butyl ether synthesis
The synthesis of methyl tertiary butyl ether (MTBE) is
one of the most important applications of RD. MTBE is
produced via an acid catalysed reaction between methanol and isobutylene as shown by (R18):
CH3 OHC(CH3 )2 CH2 l C(CH3 )3 OCH3
DHR0 37:7 kJ=mol
(R18)
This reaction has been extensively investigated by

several authors, e.g. [135 /137].
The synthesis of MTBE via RD was one of the key
processes studied in the context of a large research
project supported by the European Commission (BE951335). Five major industrial companies and five universities from Finland, Germany, Great Britain and
Italy built a consortium with the aim to perform a
comprehensive theoretical and experimental programme
and to develop special CAPE tools for feasibility and
design studies of the process.
MTBE synthesis was investigated both theoretically
and experimentally. In this paper we present some
results for a pilot scale RD column at Neste Oy
Engineering (Finland). The column (used as an example
here) has a catalytic section in the middle part. This
catalytic section consists either of a packed bed of
catalytically active rings (see [138]) or of catalytic
packing MULTIPAK [50]. The rectifying and stripping sections are filled with Intalox Metal Tower
Packing. The methanol feed is introduced just above
and the hydrocarbon feed just below the catalyst section
of the column.
The mathematical description considered in Section
3.3 was used in the project as a modelling basis for the
172
Fig. 14. Liquid bulk mole fractions in the column reboiler: lines, simulations; dots, experiments.
Fig. 15. Liquid bulk mole fractions in the column condenser: lines,
simulations; dots, experiments.
specially developed completely rate-based simulator

DESIGNER [139]. This tool consists of several blocks
including model libraries for physical properties, mass
and heat transfer, reaction kinetics and equilibrium as
well as specific hybrid solver and thermodynamic
package.
DESIGNER also contains different hydrodynamic
models (e.g. completely mixed liquid-completely mixed
vapour, completely mixed liquid /vapour plug flow,
mixed pool model, eddy diffusion model) and a model
library of hydrodynamic correlations for the mass
transfer coefficients, interfacial area, pressure drop,
hold-up, weeping and entrainment, which cover a
number of different column internals and flow conditions.
In DESIGNER, different ways of taking account of
heterogeneous reaction kinetics are available, depending
Fig. 16. Axial liquid bulk concentration profiles for the semi-batch
column (t /10 000 s).
on the reaction rate and character. One possibility is to

use a detailed model for the heterogeneous catalyst mass
transfer efficiency, which is based on the approach of
[138]. When applying this type of kinetic models, the
intrinsic kinetics data are needed (see Section 3.1).
Another way is the pseudo-homogeneous approach
with effective kinetic expressions, by which the kinetics
description is introduced as source terms into the
balance equations (cf. Eqs. (3) and (4)).
In the system considered here, the reaction is slow as
compared with the mass transfer rate. For this reason,
the pseudo-homogeneous approach is used, the reaction

being considered in the liquid bulk only.

Basically, DESIGNER can use different physical
property packages, which are easy to interchange. One
of these packages is the thermodynamic interface IKCAPE developed in a co-operative project involving all
large German chemical companies (see [140]). For the
example given here, all necessary physicochemical
properties are computed with the help of the IK-CAPE
tool, except for the diffusion coefficients. Thermodynamic equilibrium description is based on the UNIQUAC model. The liquid-phase binary diffusivities are
determined using the method of Tyn and Calus (see [72])
for the diluted mixtures corrected by the Vignes
equation [61] to account for finite concentrations. The
vapour-phase diffusion coefficients are assumed constant. The reaction kinetics parameters taken from [141]
are implemented directly into the DESIGNER code.

Fig. 17 demonstrates the simulated and measured
concentration profiles for the pilot test made in the
column with the reactive section filled with catalytically
active rings. In the simulations, four components,
namely, methanol, iso butene, MTBE and 1-butene,
were chosen to describe the system under consideration.
Here, segment 1 corresponds to the reboiler. A satisfactory agreement between calculated and measured values
can be clearly observed. In Fig. 18, the simulation results
for the column packed with MULTIPAK are shown.
Here, 16 components are considered. Again, the liquid
bulk composition profiles agree well with the experimental data.
Fig. 17. Calculated and experimental liquid compositions for experiments with catalytically active rings.
173
5. Conclusions and outlook

The paper reviews most important RSP - RA and
RD. Both these operations combining the separation
and reaction steps inside a single column are advantageous as compared with traditional unit operations. The
paper suggests a detailed discussion of the process basics
and peculiarities, considers the up-to-date applications
and thoroughly discusses the RA and RD modelling and
design issues. The theoretical description is illustrated by
several case studies and supported by the results of
laboratory-, pilot- and industrial-scale experimental
investigations. Both steady-state and dynamic issues
are treated, and in addition, the design of column
internals is addressed.
RA and RD occur in multicomponent multiphase
fluid systems, and thus a single modelling framework for
these processes is desirable. The paper discusses different possible ways to build such a framework and
advocates the rate-based approach as the most rigorous
and appropriate way. This approach provides a direct
consideration of the diffusional and reaction kinetics.
In terms of the rate-based approach, the peculiarities
of the specific process applications and the different
solution strategies are treated while discussing the case
studies. A detailed general description covers mass and
heat transfer, reaction kinetics including the reactionmass transfer coupling, as well as steady-state and
dynamic modelling issues.
The modelling of RA is illustrated by the absorption
of NOx and by the coke gas purification process. The
first case is modelled by using an analytical treatment of
the film phenomena, whereas the second one is solved by
purely numerical technique. The simulation results are
compared with the experimental data obtained at an
industrial absorption plant consisting of eight units with
pump around (NOx ) and at a pilot column for the
ASCS process (coke gas purification). For the later case,
both steady-sate and dynamic conditions are considered.
The comparison results, on the one hand, demonstrate a
good agreement between the rate-based simulations and
experimental data, and, on the other hand, warn of
using equilibrium approach which appears completely
inappropriate in the case of complex finite-rate reactions.
The modelling of RD processes is illustrated with the
heterogeneously catalysed syntheses of methyl acetate
and MTBE using sandwich-type packings and catalytically active rings. Both processes are described based on
the pseudo-homogeneous approach for the reaction
kinetics.
For the methyl acetate synthesis, both steady-state
and dynamic modelling issues are investigated. The
comparison between the simulation and experimental
data made for all RD case studies proves that the rate-
174
Fig. 18. Calculated and experimental liquid compositions for experiments with catalytic structured packing.
based approach is capable of predicting correct process

behaviour, both steady-state and dynamic.
The key RSP topics to be addressed in the near future
are a proper hydrodynamic modelling for catalytic
internals including RTD account and scale-up methodology. Further studies on the hydrodynamics of
catalytic internals are essential for a better understanding of RSP behaviour and the availability of optimally
designed catalytic column internals for them. In this
regard, the methods of computational fluid dynamics
appear very helpful.
The development of new methodologies enabling
creation of intelligent, tailor-made column internals
and consequent RSP optimisation constitutes one of
the burning present-day challenges. Such a development
is already in progress in some European research
projects (see [70,71]).
Despite a fast recent development of computer
technology and numerical methods, the rate-based
approach in its current realisation still requires a
significant computational effort, with related numerical
difficulties. This is one of the reasons why the application of rate-based models to industrial tasks is rather
limited. Therefore, further work is required in order to
bridge this gap and provide chemical engineers with
reliable, consistent, robust and comfortable simulation
tools for RSP.
Acknowledgements
We would like to thank our colleagues at the Chair of
Thermal Process Engineering and all project partners
who have been involved in the research activities. We are
also grateful to the German Research Foundation

(DFG, Grant No. Schm 808/5-1), the VolkswagenFoundation (Project No. I/70 875, 876, 877) and the
European Commission (BRITE-EURAM program,
CEC Project No. BE95-1335) for financial support.
Appendix A: Nomenclature
Ac
aI
B
c
d
dC
[D ]
//
Dax
Deff
E
EA
Eu
F
FC
G
DH0R
h
k
kp1
column cross section (m2)

specific gas /liquid interfacial area (m2/
m 3)
liquid load (m3/m2 per h)
molar concentration (mol/m3)
generalised driving force (l/m)
column diameter (m)
diffusion matrix (m2/s)
Maxwell /Stefan diffusion coefficient
(m2/s)
axial dispersion coefficient (m2/s)
effective diffusion coefficient (m2/s)
specific energy hold-up (J/mol per s)
activation energy (J/mol)
dimensionless residence time distribution
Faradays constant (9.65/104 C/mol)
gas capacity factor (Pa0.5)
gas molar flow rate (mol/s)
reaction enthalpy (kJ/mol)
molar enthalpy (J/mol)
second order reaction rate constant (m3/
mol per s)
rate constant of reaction 1 related to the
component partial pressures (1/bar2 per
s)
Ki
Keq
i
[K ]
l
L
n
N
Q
r
R
R
ReBL
t
T
u
U
x
y
z
zi
Greek letters
d
f
8
h
l
m/
u
Subscripts
G
i, j
L
n
t
Superscripts
B
I
Abbreviations
ADM
CD
PDE
RA
RD
RSP
RTD
distribution coefficient
equilibrium constant of reaction i
reaction matrix (Eqs. (14) and (15)) (1/s)
axial coordinate (m)
liquid molar flow rate (mol/s)
number of components of mixture
molar flux (mol/m2 per s)
heat flux (W/m2)
equivalent reaction rate (mol/m3 per s)
total component reaction rate (mol/m3
per s)
gas constant (8.3144 J/mol per K)
liquid Reynolds number
time (s)
temperature (K)
liquid velocity (m/s)
length specific molar hold-up (mol/m)
liquid mole fraction (mol/mol)
gas mole fraction (mol/mol)
film coordinate (m)
ionic charge of component i
film thickness (m)
volumetric hold-up (m3/m3)
electrical potential (V)
dimensionless film coordinate
molecular thermal conductivity (W/m
per K)
chemical potential (J/mol)
dimensionless time
gas phase
component/reaction indices
liquid phase
solvent index
total
bulk phase
phase interface
axial dispersion model
catalytic distillation
piston flow with axial dispersion and
mass exchange
reactive absorption
reactive distillation
reactive separation process
residence time distribution
References
[1] M.F. Doherty, G. Buzad, Reactive distillation by design, Trans.
Inst. Chem. Eng. 70 (1992) 448 /458.
175
[2] R. Zarzycki, A. Chacuk, Absorption: Fundamentals and Applications, Pergamon Press, Oxford, 1993.
[3] D.W. Agar, Multifunctional reactors: old preconceptions and
new dimensions, Chem. Eng. Sci. 54 (1999) 1299 /1305.
[4] V.V. Kelkar, K.M. Ng, Design of reactive crystallization systems
incorporating kinetics and mass-transfer effects, Am. Inst.
Chem. Eng. J. 45 (1999) 69 /81.
[5] J. Fricke, H. Schmidt-Traub, M. Kawase, Chromatographic
reactors, in: Ullmanns Encyclopedia of Industrial Chemistry,
Wiley-VCH, Weinheim, 2001.
[6] H.-J. Bart, Reactive Extraction, Springer, Berlin, 2001.
[7] W.J. Hatcher, Reaction and mass transport in two-phase
reactors, in: N.P. Cheremisinoff (Ed.), Handbook of Heat and
Mass Transfer, vol. 2, Gulf Publ Comp Book Division, Houston,
1986, pp. 837 /868.
[8] E.Y. Kenig, Modeling of Multicomponent Mass Transfer in
Separation of Fluid Mixtures, VDI-Verlag, Dusseldorf, 2000.
[9] T.K. Sherwood, R.L. Pigford, C.R. Wilke, Mass Transfer,
McGraw-Hill, New York, 1975.
[10] T.K. Sherwood, R.L. Pigford, Absorption and Extraction,
McGraw-Hill, New York, 1952.
[11] R.M. Counce, J.J. Perona, Designing packed-tower wet scrubbers: emphasis on nitrogen oxides, in: N.P. Cheremisinoff (Ed.),
Handbook of Heat and Mass Transfer, vol. 2, Gulf Publ Comp
Book Division, Houston, 1986, pp. 953 /966.
[12] K. Sattler, Thermische Trennverfahren, VCH, Weinheim, 1988.
[13] S. Strelzoff, Choosing the optimum CO2-removal system, Chem.
Eng. 82 (1975) 115 /120.
[14] J. Falbe, Chemierohstoffe aus Kohle, Georg-Thieme-Verlag,
Stuttgart, 1977.
[15] A.L. Kohl, F.C. Riesenfeld, Gas Purification, Gulf Publ Comp
Book Division, Houston, 1985.
[16] H. Thielert, Simulation und Optimierung der Kokereigaswasche,
Ph.D. thesis, Technical University of Berlin, Germany, 1997.
[17] W.-C. Yu, G. Astarita, Selective absorption of hydrogen
sulphide in tertiary amine solutions, Chem. Eng. Sci. 42 (1987)
419 /424.
[18] A. Kobus, Ein heuristisch-numerischer Ansatz zum systematischen Entwurf und Design von Absorptionsverfahren, VDIVerlag, Dusseldorf, 1999.
[19] P.V. Danckwerts, Gas /Liquid Reactions, McGraw-Hill, New
York, 1970.
[20] G. Astarita, D.W. Savage, A. Bisio, Gas Treating with Chemical
Solvents, Wiley, NewYork, 1983.
[21] E.Y. Kenig, L.P. Kholpanov, L.I. Katysheva, I.H. Markish,
V.A. Malyusov, Calculation of two-phase non-isothermal absorption in a liquid film in downward co-current flow, Theor.
Found. Chem. Eng. 19 (1985) 97 /102.
[22] P. Gandhidasan, Heat and mass transfer in solar regenerators,
in: N.P. Cheremisinoff (Ed.), Handbook of Heat and Mass
Transfer, vol. 2, Gulf Publ Comp Book Division, Houston, 1986,
pp. 1475 /1499.
[23] O. Worz, H.-H. Mayer, Reaction columns, in: Ullmanns
Encyclopedia of Industrial Chemistry, Wiley-VCH, Weinheim,
2001.
[24] M. Sakuth, D. Reusch, R. Janowsky, Reactive distillation, in:
Ulmanns Encyclopedia of Industrial Chemistry, Wiley-VCH,
Weinheim, 2001.
[25] R. Taylor, R. Krishna, Modelling reactive distillation, Chem.
Eng. Sci. 55 (2000) 5183 /5229.
[26] G.P. Towler, S.J. Frey, Reactive distillation, in: S. Kulprathipanja (Ed.), Reactive Separation Processes, Taylor & Francis,
Philadelphia, 2001.
[27] V.H. Agreda, L.R. Partin, W.H. Heise, High-purity methyl
acetate via reactive distillation, Chem. Eng. Prog. 86 (2) (1990)
40 /46.
176
[28] L.-U. Kreul, Discontinuous and reactive distillation, Ph.D.

thesis, University of Dortmund, Germany, 1998.
[29] A. Gorak, A. Hoffmann, Catalytic distillation in structured
packings: methyl acetate synthesis, Am. Inst. Chem. Eng. J. 47
(2001) 1067 /1076.
[30] H. Komatsu, C.D. Holland, A new method of convergence for
solving reacting distillation problems, J. Chem. Eng. Jpn. 4
(1977) 292 /297.
[31] H. Hartig, H. Regner, Verfahrenstechnische Auslegung einer
Veresterungskolonne, Chem. Ing. Tech. 43 (1971) 1001 /1007.
[32] B. Davies, G.V. Jeffreys, The continuous trans-esterification of
ethyl alcohol and butyl acetate in a sieve plate column. Part III:
Trans-esterification in a six plate sieve tray column, Trans. Inst.
Chem. Eng. 51 (1973) 275 /280.
[33] H.-P. Luo, W.-D. Xiao, A reactive distillation process for a
cascade and azeotropic reacting system: carbonylation of ethanol
with dimethyl carbonate, Chem. Eng. Sci. 56 (2001) 403 /410.
[34] Y. Fuchigami, Hydrolysis of methyl acetate in distillation
column packed with reactive packing of ion exchange resin, J.
Chem. Eng. Jpn. 23 (1990) 354 /358.
[35] J.L. DeGarmo, V.N. Parulekar, V. Pinjala, Consider reactive
distillation, Chem. Eng. Prog. 88 (3) (1992) 42 /50.
[36] K. Sundmacher, U. Hoffmann, Activity evaluation of a catalytic
distillation packing for MTBE production, Chem. Eng. Technol.
16 (1993) 279 /289.
[37] C. Thiel, Modellbildung, Simulation, Design und experimentelle
Validierung von heterogen katalysierten Reaktivdestillationsprozessen zur Synthese der Kraftstoffether MTBE, ETBE und
TAME, Ph.D. thesis, Technical University of Clausthal, Germany, 1997.
[38] C. Oost, Ein Beitrag zum Umweltschutz: Reaktionstechnische
Untersuchungen zur Flussigphasesynthese des Antiklopfmittels
TAME, Ph.D. thesis, Technical University of Clausthal, Germany, 1995.
[39] J.D. Shoemaker, E.M. Jones, Cumene by catalytic distillation,
Hydroc. Proc. 66 (1987) 57 /58.
[40] G.G. Podrebarac, F.T.T. Ng, G.L. Rempel, The production of
diacetone alcohol with catalytic distillation. Part I: catalytic
distillation experiments, Chem. Eng. Sci. 53 (1998) 1067 /1075.
[41] S. Ung, M.F. Doherty, Vapor /liquid equilibrium in systems
with multiple chemical reactions, Chem. Eng. Sci. 50 (1995) 23 /
48.
[42] D. Muller, J.-P. Schafer, H.-J. Leimkuhler, The economic
potential of reactive distillation processes exemplified by silane
production, Proceedings of the GVC, DECHEMA and EFCE
Meeting on Distillation, Absorption and Extraction, Bamberg,
Germany, 2001.
[43] A.R. Ciric, P. Miao, Steady-state multiplicities in an ethylene
glycol reactive distillation column, Ind. Eng. Chem. Res. 33
(1994) 2738 /2748.
[44] D.J. Belson, A distillation method of aromatic nitration using
azeotropic nitric acid, Ind. Eng. Chem. Res. 29 (1990) 1562 /
1565.
[45] K.H. Johnson, Catalytic distillation structure, US Patent
5189001.
[46] B.W. van Hasselt, H.P.A. Calis, S.T. Sie, C.M. van den Bleek,
Liquid holdup in the three-levels-of-porosity reactor, Chem.
Eng. Sci. 54 (1999) 1405 /1411.
[47] L.A.J. Smith, Catalytic distillation structure, US Patent 4232177.
[48] A.P. Gelbein, M. Buchholz, Process and structure for effecting
catalytic reactions in distillation structure, US Patent 5073236.
[49] R. Shelden, J.-P. Stringaro, Vorrichtung zur Durchfuhrung
katalysierter Reaktionen, European patent 0396650 B1.
[50] A. Gorak, L.U. Kreul, M. Skowronski, Strukturierte Mehrzweckpackung, German Patent 19701045 A1.
[51] L. Gotze, O. Bailer, Katalysatorsandwich-Reaktivdestillation

mit einer neuen strukturierten Packung, Chemie Technik 29 (2)
(2000) 42 /45.
[52] E.M.J. Jones, Contact structure for use in catalytic distillation,
US Patent 4536373.
[53] N. Yeoman, R. Pinaire, M.A. Ulowetz, T.P. Nace, D.A. Furse,
Method and apparatus for concurrent reaction with distillation,
World Patent 94/08679.
[54] T. Adrian, B. Bessling, H. Hallmann, J. Niekerken, V. Spindler,
A. Ohligschlager, M. Rumpf, Vorrichtung zur Durchfuhrung
von Destillationen und heterogen katalysierten Reaktionen,
German Patent 19869598 A1.
[55] J.L. Nocca, J.A. Chodorge, Catacol, a Low Cost Reactive
Distillation Technology for Etherification, IFP Industrial Division 2001.
[56] H. Schoenmakers, Reactive and catalytic distillation from an
industrial perspective, Chem. Eng. Process. 42 (2003) 145 /155.
[57] K. Jakobsson, A. Pyhalathi, S. Pakkanen, K. Keskinen, J.
Aittamaa, Modelling of a configuration combining distillation
and reaction in a side-reactor, Proceedings of the Second
International Symposium on Multifunctional Reactors (ISMR2), Nuremberg, Germany, 2001.
[58] U. Kunz, Entwicklung neuartiger Polymer-/Trager-Ionenaustauscher als Katalysatoren fur chemische Reaktionen in Fullkorperkolonnen, Papierflieger, Clausthal-Zellerfeld, 1998.
[59] Sulzer, KATAPAK (product information), Sulzer Chemtech
AG, 2000.
[60] F. Kapteijn, J.J. Heiszwolf, T.A. Nijhuis, J.A. Moulijn, Monoliths in multiphase catalytic processes, CATTECH 3 (1) (1999)
24 /41.
[61] R. Taylor, R. Krishna, Multicomponent Mass Transfer, Wiley,
New York, 1993.
[62] E. Kenig, A. Gorak, A film model based approach for
simulation of multicomponent reactive separation, Chem. Eng.
Process. 34 (1995) 97 /103.
[63] L.U. Kreul, A. Gorak, P.I. Barton, Modeling of homogeneous
reactive separation processes in packed columns, Chem. Eng.
Sci. 54 (1999) 19 /34.
[64] K. Sundmacher, U. Hoffmann, Multicomponent mass and
energy transport on different length scales in a packed reactive
distillation column for heterogeneously catalysed fuel ether
production, Chem. Eng. Sci. 49 (1994) 3077 /3089.
[65] R. Krishna, J.A. Wesselingh, The Maxwell /Stefan approach to
mass transfer, Chem. Eng. Sci. 52 (1997) 861 /911.
[66] A. Higler, R. Krishna, R. Taylor, Nonequilibrium modeling of
reactive distillation: a dusty fluid model for heterogeneously
catalyzed processes, Ind. Eng. Chem. Res. 39 (2000) 1596 /1607.
[67] Z. Yuxiang, X. Xien, Study on catalytic distillation processes.
Part II. Simulation of catalytic distillation processes /quasihomogenous and rate-based model, Trans. Inst. Chem. Eng. 70
(1992) 465 /470.
[68] R. Billet, Packed Towers, VCH, Weinheim, 1995.
[69] J. Mackowiak, Fluiddynamik in Kolonnen mit modernen Fullkorperpackungen und Packungen fur Gas/Flussigkeitssysteme,
Sauerlander, Frankfurt am Main, 1991.
[70] E.Y. Kenig, M. Kloker, Y. Egorov, F. Menter, A. Gorak,
Towards improvement of reactive separation performance using
computational fluid dynamics, Proceedings of the Second
International Symposium on Multifunctional Reactors (ISMR2), Nuremberg, Germany, 2001.
[71] http://www.cpi.umist.ac.uk/intint/.
[72] R. Reid, J.M. Prausnitz, B.E. Poling, The Properties of Gases
and Liquids, McGraw-Hill, New York, 1987.
[73] G.F. Froment, K.B. Bischoff, Chemical Reactor Analysis and
Design, Wiley, New York, 1990.
[74] E. Sorel, La Rectification de lAlcool, Gauthiers et Fils, 1893.

[75] E.J. Henley, J.D. Seader, Equilibrium Stage Separation Operations in Chemical Engineering, Wiley, New York, 1981.
[76] D.W. van Krevelen, P.J. Hoftijzer, Kinetics of gas /liquid
reactions */Part I: General Theory, Recl. Trav. Chim. PaysBas 67 (1948) 563 /586.
[77] L.L. Doraiswamy, M.M. Sharma, Heterogeneous Reactions:
Analysis, Examples and Reactor Design, Wiley, New York,
1984.
[78] I. Suzuki, H. Yagi, H. Komatsu, M. Hirata, Calculation of
multicomponent distillation accompanied by a chemical reaction, J. Chem. Eng. Jpn. 4 (1971) 26 /33.
[79] Y.A. Chang, J.D. Seader, Simulation of continuous reactive
distillation by homotopy continuation method, Comp. Chem.
Eng. 12 (1988) 1243 /1255.
[80] H. Holma, J. Sohlo, A mathematical model of an absorption
tower of nitrogen oxides in nitric acid production, Comp. Chem.
Eng. 3 (1979) 135 /141.
[81] G. Carta, Scrubbing of nitrogen oxides with nitric acid solutions,
Chem. Eng. Comm. 42 (1986) 157 /170.
[82] W.E. Stewart, R. Prober, Matrix calculation of multicomponent
mass transfer in isothermal systems, Ind. Eng. Chem. Fundam. 3
(1964) 224 /235.
[83] H.L. Toor, Solution of the linearized equations of multicomponent mass transfer, Am. Inst. Chem. Eng. J. 10 (1964) 448 /455,
460 /465.
[84] A. Gorak, Simulation thermischer Trennverfahren fluider Vielkomponentengemische, in: G. Schuler (Ed.), Prozesimulation,
VCH-Verlag, Weinheim, 1995, pp. 349 /408.
[85] K.R. Westerterp, W.P.M. van Swaaij, A.A.C.M. Beenackers,
Chemical Reactor Design and Operation, Wiley, Chichester,
1984.
[86] J.D. Seader, The rate-based approach for modelling staged
separations, Chem. Eng. Prog. 85 (10) (1989) 41 /49.
[87] S.S. Katti, Gas /liquid /solid systems: an industrial perspective,
Trans. Inst. Chem. Eng. 73 (Part A) (1995) 595 /607.
[88] W.K. Lewis, W.G. Whitman, Principles of gas absorption, Ind.
Eng. Chem. 16 (1924) 1215 /1220.
[89] R. Higbie, The rate of absorption of a pure gas into a still liquid
during short periods of exposure, Trans. Am. Inst. Chem. Eng.
31 (1935) 365 /383.
[90] J.O. Hirschfelder, C.F. Curtiss, R.B. Bird, Molecular Theory of
Gases and Liquids, Wiley, New York, 1964.
[91] S. Pelkonen, Multicomponent mass transfer in packed distillation columns, Ph.D. thesis, University of Dortmund, Germany,
1997.
[92] E.Y. Kenig, U. Wiesner, A. Gorak, Modeling of reactive
absorption using the Maxwell /Stefan equations, Ind. Eng.
Chem. Res. 36 (1997) 4325 /4334.
[93] E.Y. Kenig, F. Butzmann, L. Kucka, A. Gorak, Comparison of
numerical and analytical solutions of a multicomponent reaction-mass transfer problem in terms of the film model, Chem.
Eng. Sci. 55 (2000) 1483 /1496.
[94] C. Noeres, C. Benvenuti, A. Hoffmann, A. Gorak, Reactive
distillation: non-ideal flow behaviour of the liquid phase in
structured catalytic packings, Proceedings of the Second International Symposium on Multifunctional Reactors (ISMR-2),
Nuremberg, Germany, 2001.
[95] A. Higler, R. Krishna, R. Taylor, Nonequilibrium cell model for
multicomponent (reactive) separation processes, Am. Inst.
Chem. Eng. J. 45 (1999) 2357 /2370.
[96] P.V. Danckwerts, Continuous flow systems */distribution of
residence times, Chem. Eng. Sci. 2 (1953) 1 /13.
[97] W.P.M. van Swaaij, J.C. Charpentier, J. Villermaux, Residence
time distribution in the liquid phase of trickle flow in packed
columns, Chem. Eng. Sci. 24 (1969) 1083 /1095.
177
[98] N.J. Suchak, K.R. Jethani, J.B. Joshi, Modeling and simulation
of NOx absorption in pilot-scale packed columns, Am. Inst.
Chem. Eng. J. 37 (1991) 323 /339.
[99] U. Wiesner, Mathematische Modellierung der Absorption nitroser Gase in Fullkorperkolonnen, Ph.D. thesis, University of
Dortmund, Germany, 1997.
[100] K. Ulrichs, W. Laue, H.-D. Renovanz, Salpetersaure und
salpetrige Saure, Stickstoffoxide, Nitrate, Nitrite, in Ullmanns
Encyklopadie der technischen Chemie, vol. 20, VCH, Weinheim,
1981, pp. 305 /362.
[101] U. Wiesner, A. Gorak, E.Y. Kenig, H. Steude, G.-G. Borger,
Modelling of absorption of nitrous waste gases, Inst. Chem. Eng.
Symp. Ser. 142 (1) (1997) 323 /333.
[102] J. Wei, C.D. Prater, The structure and analysis of complex
reaction systems, Adv. Catal. 13 (1962) 203 /392.
[103] H. Hikita, S. Asai, Gas absorption with (m , n )-th order
irreversible chemical reactions, Int. Chem. Eng. 4 (1964) 332 /
340.
[104] H.L. Toor, Dual diffusion-reaction coupling in first order
multicomponent systems, Chem. Eng. Sci. 20 (1965) 941 /951.
[105] G.B. DeLancey, Multicomponent film-penetration theory with
linearized kinetics-I Linearization theory and flux expressions,
Chem. Eng. Sci. 29 (1974) 2315 /2323.
[106] G. Astarita, S.I. Sandler, Kinetic and Thermodynamic Lumping
of Multicomponent Mixtures, Elsevier, Amsterdam, 1991.
[107] E.Y. Kenig, L.P. Kholpanov, Analysis of formulation and
solution of multicomponent reaction-diffusion problems, Theor.
Found. Chem. Eng. 26 (1992) 510 /521.
[108] E.N. Fuller, P.D. Schettler, J.C. Giddings, A new method for
prediction of binary gas-phase diffusion coefficients, Ind. Eng.
Chem. 58 (1966) 18 /27.
[109] K. Onda, H. Takeuchi, Y. Okumoto, Mass transfer coefficients
between gas and liquid phases in packed columns, J. Chem. Eng.
Jpn. 1 (1968) 56 /62.
[110] H. Brauer, Stoffaustausch einschlielich chemischer Reaktionen,
Sauerlander, Aarau, 1971.
[111] M.A. Siddiqi, K. Lucas, Correlations for prediction of diffusion
in liquids, Can. J. Chem. Eng. 64 (1986) 839 /843.
[112] N. Kolev, Wirkungsweise von Fullkorperschuttungen, Chem.
Ing. Tech. 48 (1976) 1105 /1112.
[113] V. Mika, Model of packed absorption column. I. Physical
absorption, Coll. Czech. Chem. Commun. 32 (1967) 2933 /2943.
[114] S.E. Schwartz, W.H. White, Solubility equilibria of the nitrogen
oxides and oxyacids in dilute aqueous solutions, in: J.R. Pfafflin,
E.N. Zeigler (Eds.), Advances in Environmental Science and
Engineering, vol. 4, Gordon and Breach, New York, 1981.
[115] H. Engels, Anwendung des Modells der lokalen Zusammensetzung auf Elektrolytlosungen, Ph.D. thesis, RWTH Aachen,
Germany, 1985.
[116] U. Wiesner, M. Wittig, A. Gorak, Design and optimization of a
nitric acid recovery plant from nitrous waste gases, Comp.
Chem. Eng. 20 (1996) S1425 /S1430.
[117] V.V. Kafarov, E.A. Shestopalov, E.A. Novikov, V.P. Belkov,
Mathematical model of absorption of NOx in production of
week nitric acid, Sov. Chem. Ind. 7 (1975) 1224 /1226.
[118] P.V. Danckwerts, M.M. Sharma, The absorption of carbon
dioxide into solutions of alkalis and amines (with some notes on
hydrogen sulfide and carbonyl sulfide), The Chemical Engineer,
July /August (1966) 244 /280.
[119] G. Maurer, On the solubility of volatile weak electrolytes in
aqueous solutions, ACS Symp. Ser. 133 (1980) 139 /172.
[120] R. Schneider, E.Y. Kenig, A. Gorak, Dynamic modeling of
reactive absorption with the Maxwell /Stefan approach, Chem.
Eng. Res. Des., Trans. Inst. Chem. Eng. Part A 77 (1999) 633 /
638.
[121] A.L. Horvath, Handbook of Aqueous Electrolyte Solutions,
Ellis Horwood, Chichester, 1985.
178
[122] J.A. Rocha, J.L. Bravo, J.R. Fair, Distillation columns containing structured packings-2. Mass transfer model, Ind. Eng. Chem.
Res. 35 (1996) 1660 /1667.
[123] E.Y. Kenig, R. Schneider, A. Gorak, Rigorous dynamic modelling of complex reactive absorption processes, Chem. Eng. Sci.
54 (1999) 5195 /5203.
[124] E.Y. Kenig, R. Schneider, A. Gorak, Reactive absorption:
optimal process design via optimal modelling, Chem. Eng. Sci.
56 (2001) 343 /350.
[125] C.C. Pantelides, The consistent initialization of differential /
algebraic systems, SIAMJ Sci. Stat. Comp. 9 (1988) 213 /231.
[126] R. Schneider, E.Y. Kenig, A. Gorak, Dynamische Simulation
reaktiver Absorptionsprozesse am Beispiel einer Sauergaswasche: Modellentwicklung, -analyse und -optimierung, Chem.
Ing. Tech. 72 (2000) 1224 /1229.
[127] R. Schneider, Modelloptimierung fur die dynamische Simulation
der reaktiven Absoprtion und Rektifikation, Ph.D. thesis,
University of Dortmund, Germany, 2001.
[128] W. Song, G. Venimadhavan, J.M. Manning, M.F. Malone, M.F.
Doherty, Measurement of residue curve maps and heterogeneous
kinetics in methyl acetate synthesis, Ind. Eng. Chem. Res. 37
(1998) 1917 /1928.
[129] R. Schneider, C. Noeres, L.-U. Kreul, A. Gorak, Dynamic
modelling and simulation of reactive batch distillation, Comp.
Chem. Eng. 25 (2001) 169 /176.
[130] F.G. Helfferich, Y.-L. Hwang, Ion exchange as catalysts, in: M.
Streat (Ed.), Ion Exchange for Industry, Ellis Horwood,
Chichester, 1988, pp. 585 /596.
[131] Z.P. Xu, K.T. Chuang, Kinetics of acetic acid esterification over
ion exchange catalysts, Can. J. Chem. Eng. 74 (1996) 493 /500.
[132] E. Ranzi, M. Rovaglio, T. Faravelli, G. Biardi, Role of energy

balances in dynamic simulation of multicomponent distillation
columns, Comp. Chem. Eng. 12 (1988) 783 /786.
berprufung unterschiedlicher Modelle fur den
[133] G. Ronge, U
Stoffaustausch bei der Rektifikation in Packungskolonnen,
VDI-Verlag, Dusseldorf, 1995.
[134] J. Gmehling, B. Kolbe, Thermodynamik, Verlag, Weinheim,
1992.
[135] A. Gicquel, B. Torck, Synthesis of methyl tertiary butyl ether
catalyzed by ion-exchange resin. Influence of methanol concentration and temperature, J. Catal. 83 (1983) 9 /18.
[136] A. Rehfinger, U. Hoffmann, Kinetics of methyl tertiary butyl
ether liquid phase synthesis catalyzed by ion exchange resin. I.
Intrinsic rate expression in liquid phase activities, Chem. Eng.
Sci. 45 (1990) 1605 /1617.
[137] D. Parra, J. Tejero, F. Cunill, M. Iborra, J.F. Izquierdo, Kinetic
study of MTBE liquid-phase synthesis using C4 olefin cut,
Chem. Eng. Sci. 49 (1994) 4563 /4578.
[138] K. Sundmacher, U. Hoffmann, Development of a new catalytic
distillation process for fuel ethers via a detailed nonequilibrium
model, Chem. Eng. Sci. 51 (1996) 2359 /2368.
[139] E. Kenig, K. Jakobsson, P. Banik, J. Aittamaa, A. Gorak, M.
Koskinen, P. Wettmann, An integrated tool for synthesis and
design of reactive distillation, Chem. Eng. Sci. 54 (1999) 1347 /
1352.
[140] G. Fieg, W. Gutermuth, W. Kothe, H.-H. Mayer, S. Nagel, H.
Wendeler, G. Wozny, A standard interface for use of thermodynamics in process simulation, Comp. Chem. Eng. 19 (1995)
S317 /S320.
[141] K. Sundmacher, U. Hoffmann, Macrokinetic analysis of MTBEsynthesis in chemical potentials, Chem. Eng. Sci. 49 (1994)
3077 /3089.

Modelling of Reactive Separation Processes

Uploaded by

Copyright:

Available Formats

Modelling of Reactive Separation Processes

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Modelling of Reactive Separation Processes

Uploaded by

Copyright:

Available Formats

See

Modelling of Reactive Separation Processes:

Technische Universitt Dortmund

211 PUBLICATIONS 2,369 CITATIONS

255 PUBLICATIONS 2,488 CITATIONS

All in-text references underlined in blue are linked to publications on ResearchGate,

Available from: Eugeny Y. Kenig

Chemical Engineering and Processing 42 (2003) 157 /178

Modelling of reactive separation processes: reactive absorption and

The most important examples of reactive separation

C. Noeres et al. / Chemical Engineering and Processing 42 (2003) 157 /178

Removal of harmful substances

Coke oven gas purification, amine washing

Sulphuric acid manufacturing, formaldehyde

and liquids. Absorbers or scrubbers where RA is

elevated pressure and high solubility of absorbed

2.2. Reactive distillation

C. Noeres et al. / Chemical Engineering and Processing 42 (2003) 157 /178

Methyl acetate from methanol and acetic acid

. common operation range (temperature and pressure)

2.3. Internals for RSP

necessary to meet optimum reactor design demands. If

C. Noeres et al. / Chemical Engineering and Processing 42 (2003) 157 /178

Fig. 1. Internals for RSP.

and closed channels for the gas/vapour and liquid flow,

C. Noeres et al. / Chemical Engineering and Processing 42 (2003) 157 /178

Fig. 2. Modelling approaches for RA and RD.

only give a brief discussion of these topics in the context

C. Noeres et al. / Chemical Engineering and Processing 42 (2003) 157 /178

geous since there is a broad spectrum of correlations

where di is the generalised driving force:

Similar equations can be also written for the gas/

3.3. Rate-based approach

Fig. 3. Two-film model for a differential packing segment.

C. Noeres et al. / Chemical Engineering and Processing 42 (2003) 157 /178

3.3.1. Balance equations

Eqs. (3) and (4) represent the mass balances for

The volumetric liquid hold-up fL depends on the gas/

3.3.2. Mass transfer and reaction coupling in the fluid

region. As the key assumptions of the film model result

Eq. (10) which is generally valid for both liquid and

Combining Eq. (10) with the boundary conditions,

C. Noeres et al. / Chemical Engineering and Processing 42 (2003) 157 /178

relationships and by the continuity condition for the

mixing effects are rather caused by hydrodynamic than

C. Noeres et al. / Chemical Engineering and Processing 42 (2003) 157 /178

specific CD processes is an objective of our current

whereas the corresponding equations for the liquid

The liquid-phase reactions are valid for nitric acid

four units used for the removal of nitrogen oxides from

Matrix [D ] results from the transformation of the

Eq. (15) represents one of the best approaches to the

C. Noeres et al. / Chemical Engineering and Processing 42 (2003) 157 /178

Fig. 6. Absorption plant consisting of four units (eight columns).

homogeneous reaction in the fluid films (see [92]). The

Mass transfer coefficient correlation

Fuller et al. [108]

Onda et al. [109]; Billet [68]; Wehmeier

Liquid Siddiqi and Lucas