European Symposium on Computer Aided Process Engineering - 11

R. Gani and S.B. Jorgensen (Editors)

9 2001 Elsevier Science B.V. All rights reserved. 357

Reactor Selection and Design for Heterogeneous Reaction Systems

Christian Btihner, Gerhard Schembecker

Process Design Center B.V., Joseph-von-Fraunhofer-Str. 20, D-44227 Dortmund, Germany

This paper presents a heuristic-numeric strategy for the selection and the design of technical
reactors for multiphase reaction systems in particular. On basis of fundamental data of the
reaction system and its structure essential decisions are made with regard to relevant reactor
selection criteria and the reactor selection itself.

1. INTRODUCTION
Selection and design of a reactor is a decisive step within the framework of process syn-
thesis. Chemical processes consist of the reaction section and the separation section. The
reaction section and its outlet streams have strong influence on the structure of the separation
section whereas recycle streams which origin from the separation section affect the conditions
of the reaction section. Apart from this, the selection and design of a reactor strongly in-
fluence the conversion of reactants to desired products and the selectivity respectively.
This short description shows that the reactor type, its characteristics and the reaction con-
ditions within this reactor are important questions which should be answered at an early stage
of process synthesis. The fact that many parameters to be considered during the selection pro-
cedure depend on one another makes the reactor choice a difficult task.
In addition only fundamental data from initial experiments will be present at an early stage
of process synthesis. This means that selection methods which are based on mathematical
models and algorithms [1] are not suitable for the selection process since the necessary data
for the utilization of such tools are usually not available.
This paper presents a strategy for the selection of suitable reactors for multiphase reaction
systems which uses a heuristic-numeric approach for the selection procedure. On the basis of
fundamental data of the reaction system (e.g. kind of side reactions, qualitative data of
kinetics) the strategy applies selection levels which are based on one another (see Fig. 1). Each
selection level determines relevant reactor selection criteria to be used in the subsequent
selection level and for the final selection of suitable technical reactors respectively. The stra-
tegy especially takes into consideration that many of the criteria investigated are based on one
another or depend on each other. Therefore in contrast to publications of other authors (e.g.
[2-3]) the strategy presented is not a simple straight forward procedure to determine selection
criteria. The strategy has been implemented in the heuristic-numeric consulting system
READPERT (_Reactor E_valuation and Design Expert System). The following chapters
describe the major steps of the decision making process.
358

2. INVESTIGATION OF THE REACTOR PHASE

Beside homogeneous reaction systems the reactor selection and design strategy especially
focuses on the investigation of multiphase reaction systems. The following phase systems can
be handled:

Fig. 1: Structure of the strategy

9 homogeneous: gas, liquid

9 heterogeneous: gas-liquid, liquid-liquid, gas-catalyst, gas-liquid-catalyst
The first step to investigate multiphase reaction systems is to analyze the reaction phase.
E.g. for a gas-liquid reaction system with the reaction taking place in the liquid phase the
liquid phase will be investigated. The gas phase is necessary for the supply of additional
reactants. This strategy step proposes basic reaction conditions to be realized in the reaction
phase by using fundamental data of the reaction system. These are mainly:
9 structure o f the reaction system: parallel reaction, consecutive reactions etc.
9 qualitative data o f kinetics: irreversible, reversible, autocatalytic, inhibited etc.
9 ratio o f reaction orders
9 temperature dependency of reactions (ratio of activation energies)
Based on this information READPERT proposes qualitatively favorable conditions with
regard to the concentration levels of the reactants (by comparison of reaction orders), the resi-
dence time distribution and the temperature profile (by comparison of activation energies,
structure of the reaction system). The goal for the investigations is the maximization of the
reactor performance (i.e. conversion, selectivity or yield). A multitude of heuristic rules is
applied which for homogeneous reactions systems have been created by Westhaus [4], now
extended for the reaction phase of multiphase reaction systems.

3. DETERMINATION OF PHASE AND MIXING CONDITIONS

This step of the strategy takes into account the effects and characteristics typical for a mul-
tiphase reaction system. The following important criteria are part of this selection level:
9 backmixing characteristic of the phases present (total, partial, no backmixing)
9 contacting of the phases (co-, counter-current)
9 interracial area (area between the fluid phases, catalyst structure)
9 holdup of reaction phase
 359

The investigation of the backmixing characteristic of the phases utilizes the results of the
preceding selection level with regard to concentration levels and residence time distribution.
In case two fluid phases exist (e.g. gas-liquid), the backmixing characteristic of the non-reac-
tion phase (gaseous) only depends on the proposed concentration levels of the reactants trans-
ferred from the gas to the liquid phase whereas the backmixing characteristic of the reaction-
phase also depends on the residence time distribution. An example for the determination of
the reaction phase backmixing characteristic is given for the following reaction system:
2
(1) Ag + B!--~ CI (CI: desired product) r1 - k 1.c A .c a
2
(2) Ag + Bl-~ D! rE = k2 "CA "Ca
(3) Cl --~ El r3 = k3 "Cc
On account of the reaction orders, the presence of a consecutive reaction and the ratio of the
reaction rates of the side reactions the following heuristic is applied for the reaction phase
(liquid phase):
Rule: The backmixing characteristic of the reaction phase is, total backmixing ', if
1) the concentration level of the reactant supplied by the reaction phase(Bt) has to be
kept ,as low as possible'
2) a narrow residence time distribution is necessary,
3) the reaction rate o f the parallel reaction is faster than the reaction rate o f the
consecutive reaction.
Due to the ratio of the reaction rates of the side reactions the residence time distribution is
not as important as the desired concentration level. As a result, total backmixing' of the reac-
tion phase is proposed.
The investigation of criteria such as e.g. interracial area requires knowledge about mass
transfer limitations between the phases existing. It is sufficient to have knowledge about the
ratio of the intrinsic reaction rate and the mass transport, called ,,kinetic regime":
9 mainly controlled by intrinsic kinetics
9 controlled by kinetics and mass transfer
9 mainly controlled by mass transfer
In order to decide which kinetic regime is present READPERT needs information about
possible conversion changes checked by simple laboratory experiments such as conversion
change by variing the agitator speed of a laboratory reactor for gas-liquid or liquid-liquid
reactions. Concerning gas-catalyst reaction systems the gas velocity and catalyst size are pos-
sible features for the determination of the kinetic regimes.
The interracial area between the phases is an important aspect as it is possible to influence
the conversion and selectivity. For reaction systems with two fluids (gas-liquid, liquid-liquid)
one can affect the concentration level of the transferred components by changing the inter-
facial area [5-8]. The following rule is a simple example for the approach:
Rule: The interracial area for a liquid-liquid system has to be as low as possible with regard to
goal,selectivity'if
1) the reaction is controlled by kinetics and mass transfer,
2) the concentration level of the transferred components has to be as low as possible.
In case a heterogeneous catalyst is present (i.e. gas-catalyst, gas-liquid-catalyst) the deter-
mination of the interracial area refers to favorable catalyst sizes and pore structures [2]. Not
only the concentration level is influenced by the catalyst size and pore structure but also the
residence time in the catalyst pores as the heuristic given below shows:
360

Rule: The catalyst size should be ,as small as possible' and the mean pore size should be a 'wide pore
structure' with regard to the goal selectivity, if
1) the reaction is controlled by kinetics and mass transfer (pore diffusion resistance),
2) parallel reactions and consecutive reactions occur as side reactions,
3) the reaction rate of the consecutive reaction is faster than the rate of the parallel reaction.
The consecutive reaction has more influence than the parallel reaction and therefore the
reduction of the residence time is more important than the consideration of the concentration
levels of the reactants. In addition to the catalyst investigation concerning concentration
levels, the strategy also offers the option to determine the catalyst with respect to heat transfer
from and to the catalyst as this also may have strong effects on conversion and selectivity.
Another criterion dependent on the kinetic regime is the holdup o f the reaction phase. First
it is checked whether explosive or poisonous phases are present and for which the holdup has
to be kept low. For two fluid systems the holdup of the reaction phase should be as high as
possible if the reaction is ,kinetically' or ,kinetically and mass transfer' controlled. In this
case the reaction takes place in the bulk of the phase. If only mass transfer limitation occurs
the holdup is not a decisive aspect, the interfacial area is much more important. Proposals for
the catalyst holdup are made on the consideration of the proposed catalyst size and the kinetic
regime as both give hints how much reaction zone is available.
Within this selection level also hints are given for counter- or co-current phase contacting
of two fluid phase systems. First of all it is revised whether a runaway of the reactor is pos-
sible (---~ no co-current operation) or if some reactants have to be supplied in excess (---~ phase
contacting is without influence). In a second step the structure of the reaction system is
regarded. If conversion is the goal for reactor selection a counter-current phase contacting is
obviously preferable, especially for equilibrium and inhibited reactions. For the goal selec-
tivity sensitivity studies have been carried out. The results show that the phase contacting
does not have significant effect on the selectivity for parallel reactions whereas in the
presence of consecutive reactions a counter-current operation improves selectivity.

4. OPERATING CONDITIONS
This third step of the strategy provides recommendations for significant operating
conditions of the reactor which are listed in figure 2. The figure shows that many of these
operating conditions are coupled and cannot be analyzed separately. Dependent on the
complexity of the reaction system it is possible that some of these criteria may have already
been derived in the two preceding selection levels before which indicates that the strategy has
to consider the dependencies between the criteria. This clarifies that the strategy is not a
simple straight forward procedure of determining selection criteria. The proposals for the
different operating conditions such as e.g. component removal or ratio of reactants are based
on the investigation of the structure of the given reaction system. A product removal for
example is recommended if consecutive reactions occur. A removal of side products may be
advantageous in case the side products react with one reactant or product. The heuristics
applied in this strategy level are meant to give hints and to generate ideas. The question if it is
possible to transfer the recommendations to technical practice is not the aim here.
 361

Fig. 2" Operating conditions and their dependencies

5. TECHNICAL R E A C T O R AND HEAT TRANSFER EQUIPMENT

The proposals of the preceding steps for different reactor selection criteria are used on the
final selection level for the selection of suitable technical reactors. Additional constraints and
further selection criteria will enter the selection process of suitable technical reactors.
First all reactors are collected from a database in READPERT which can handle the reac-
tion phase system considered (e.g. liquid-liquid). In the succeeding step so-called knock out
criteria as for example deposit forming or foaming caused in the reactor by the reaction mix-
ture are applied. In case a reactor cannot handle such reaction mixtures it is excluded from the
selection process. Having checked the knock out criteria the subsequent selection steps utilize
so-called ranking criteria to find favorable reactors. The list of ranking criteria consists of the
reactor selection criteria determined in the preceding selection levels (e.g. backmixing and
contacting of phases, interfacial area, kinetic regime etc.). Additional criteria such as phase
dispersion or corrosive phase are also part of the list of ranking criteria. Applying the ranking
criteria- for each step the user of READPERT can select one ranking criteria out of the list -
the number of favorable technical reactors will be reduced step by step, so that at the end of
the selection a list of most favorable reactors will be left. Moreover, READPERT offers the
option to change the input of a selection step in order to create alternative solution branches.
Having finished the reactor selection procedure it is possible to evaluate heat transfer
options. At first the reactors are analyzed by shortcut methods concerning a feasible adiabatic
operation. In case an adiabatic operation is not possible different heat transfer options will be
regarded for each reactor such as boiling cooling, external heat exchange equipment, internals
for heat exchange and heat exchange using the reactor wall. First general constraints are
checked which prevent the usage of a heat transfer option for the reactor (e.g. boiling cooling
cannot be applied in case strong foaming occurs). In a second step short cut methods for heat
transfer options are applied in order to roughly estimate whether the heat production can be
controlled by the heat transfer option. As a results one for example gets the necessary heat
transfer area for external heat exchange equipment.
An extra advantage of the presented heuristic-numeric strategy is the fact that it can pro-
vide useful hints with regard to additional laboratory experiments which may help to improve
the investigation of the selection criteria and the reactor selection itself.
362

6. EXAMPLE: ALKYLATION OF 1-BUTENE WITH ISOBUTANE PRODUCING Cs-

ALKYLATE
Data of reaction system
Reaction system 9 irreversible, catalyst: H2SO4, exothermic
9 phase system: liquid-liquid
1. Isobutane + 1-Butene --->C8-Alkylate 9 exothermic
2. Cs-Alkylate + 1-Butene --->C12-Alkylate 9 reaction orders: first order
9 ratio of activation energies: EA~< Em
Investigation of reaction phase Phase and mixing conditions
9 concentration level of isobutane: 9 no backmixing characteristic of both
without influence phases favorable
9 concentration level of 1-butene: 9 interfacial area: as high as possible
without influence 9 dispersion of reaction phase
9 residence time distribution: narrow 9 contacting of phases: counter-current
9 temperature profile: constant (as 9 holdup of reaction phase: without
low as possible) influence
Operating conditions
no special reactants supply 9 C8-Alkylate removal recommended
ratio of reactants" excess of isobutane 9 incomplete conversion of isobutane
recycle streams: recycle of isobutane 9 safety" no hint concerning necessity of
adding inerts

Technical reactor
9 checked criteria: deposit forming, stable emulsion, backmixing characteristic,
interracial area, kinetic regime, contacting, volume flow ratio
9 favorable: Rotating Disc Contactor, Kuehni or Scheibel extraction column
9 heat transfer: adiabatic operation not possible, employment of external heat exchangers

7. CONCLUSION
The presented selection procedure has been implemented in the heuristic-numeric con-
sulting system READPERT which is part of PROSYN | a tool for computer-aided process
synthesis. The application of READPERT in numerous industrial process synthesis projects
proofed that READPERT is a useful tool for improving the process structure and the reactor
performance in special.
REFERENCES
[1] V. Metha and A. Kokossis, Computers chem. Engng., Vol. 22 (1998), S 119
[2] R. Krishna and S. Sie, Chem. Eng. Sci., Vol. 49 (1994), No. 24A, 4029
[31 R. Jacobs and W. Jansweijer, Comp. & Chem. Eng., Vol. 24 (2000), 1781
[4] U. Westhaus, Beitrag zur Auswahl chemischer Reaktoren mittels heuristisch-nume-
rischer Verfahren, Ph.D. Thesis, Universit~it Dortmund, 1995, Shaker-Verlag
[5] O. Nagel, B. Hegner and H. Ktirten, Chem.-Ing.-Tech., Vol. 50 (1978), 934
[6] K. Schiigerl and R. H~insel, Chem.-Ing.-Tech., Vol. 58 (1986), 308
[7] R. Doraiswami and K. Yesuda, Chem. Eng. Comm., Vol. 147 (1996), 119
[8] K. Samant, AIChE J., Vol. 44 (1998), No. 10, 2212