This article was downloaded by: [University North Carolina - Chapel Hill]

On: 15 July 2013, At: 08:05

Publisher: Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number:
1072954 Registered office: Mortimer House, 37-41 Mortimer Street,
London W1T 3JH, UK

Catalysis Reviews: Science

and Engineering
Publication details, including instructions for
authors and subscription information:
http://www.tandfonline.com/loi/lctr20

LABORATORY CATALYTIC
REACTORS
a b
L. K. Doraiswamy & D. G. Tajbl
a
National Chemical Laboratory, Poona, 8, India
b
Mobil Research and Development Corporation,
Paukboro, New Jersey, 08066
Published online: 19 Dec 2006.

To cite this article: L. K. Doraiswamy & D. G. Tajbl (1974) LABORATORY CATALYTIC

REACTORS, Catalysis Reviews: Science and Engineering, 10:1, 177-219

To link to this article: http://dx.doi.org/10.1080/01614947408079629

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all
the information (the “Content”) contained in the publications on our
platform. However, Taylor & Francis, our agents, and our licensors
make no representations or warranties whatsoever as to the accuracy,
completeness, or suitability for any purpose of the Content. Any opinions
and views expressed in this publication are the opinions and views of
the authors, and are not the views of or endorsed by Taylor & Francis.
The accuracy of the Content should not be relied upon and should be
independently verified with primary sources of information. Taylor and
Francis shall not be liable for any losses, actions, claims, proceedings,
demands, costs, expenses, damages, and other liabilities whatsoever
or howsoever caused arising directly or indirectly in connection with, in
relation to or arising out of the use of the Content.

This article may be used for research, teaching, and private study
purposes. Any substantial or systematic reproduction, redistribution,
 reselling, loan, sub-licensing, systematic supply, or distribution in any
form to anyone is expressly forbidden. Terms & Conditions of access
and use can be found at http://www.tandfonline.com/page/terms-and-
conditions
Downloaded by [University North Carolina - Chapel Hill] at 08:05 15 July 2013
 Laboratory Catalytic Reactors
Downloaded by [University North Carolina - Chapel Hill] at 08:05 15 July 2013

L. K. DORAISWAMY
National Chemical Laboratory
Poona.8, India

D .G. TAJBL
Mo bil Research and Development Corporation
Paulsboro. New Jersey 08066

I. INTRODUCTION .................................... 178

I1. PACKED BED INTEGRAL AND DIFFERENTIAL REACTORS . 179
.
A Basic Principles .................................... 179
B. IntegralReactors ................................... 180
C. Differential Reactors ................................ 182
I11. INTEGRAL VS DIFFERENTIAL REACTORS:
DATA ANALYSIS .................................... 183
IV . EXTERNAL RECYCLE REACTORS ...................... 184
A. Steady-State Reactors ............................... 184
.
B Batch and Semibatch Reactors ........................ 186
V. AGITATED REACTORS ............................... 187
A. Continuous Stirred Gas-Solid Reactors .................. 187
B. Rotating Basket Reactors ............................ 187
C. Reactors with Catalyst Impregnated on Reactor Walls
or Placed in an Annular Basket ........................ 189
D. Reactor with Catalyst Placed in a Stationary
Cylindrical Basket .................................. 190
E . Internal Recirculation Reactor ........................ 192
F. General Considerations in Stirred Reactors ............... 193
G. Fluidized Bed Reactor with Agitation ................... 195
H. BallMillReactor .................................. 197
VI . TUBULARREACTORS ................................ 197

177
Copyright 0 1 9 7 5 by Marcel Dekker. Inc . All Rights Reserved . Neither this work nor any part
may be reproduced or transmitted in any form or by any means. electronic or mechanical. including
photocopying. microfilming. and recording. or by any information storage and retrieval system .
without permission in writing from the publisher.
 178 L. K. DORAISWAMY and D. G. TAJBL

VII. ADIABATIC REACTORS ............................... 199

VIII. MICROREACTORS ................................... 201
A. Typical Microreactors ............................... 205
B. Chromatographic Column Reactors ..................... 208
IX. ISOTHERMAL HEAT TRANSFER MEDIA ................. 209
X. CRITERIA FOR THE ABSENCE OF HEAT AND MASS
TRANSFER EFFECTS ................................. 209
Downloaded by [University North Carolina - Chapel Hill] at 08:05 15 July 2013

SYMBOLS ........................................... 214

REFERENCES ....................................... 216

I. INTRODUCTION

A basic requirement in the analysis of catalytic reactors is a rate

expression for the reaction concerned. In the design of any equipment
the equations to be used contain certain basic physical-chemical co-
efficients whose values must be known. In the case of physical opera-
tions, these coefficients can usually be predicted from available corre-
lations and the design of equipment is often carried out from first
principles. But there is no method available by which the kinetic
rate parameters of a chemical reaction can be predicted; thus ex-
perimental determination of these parameters is unavoidable.
Hence the choice of a suitable reactor for carrying out experi-
ments under conditions where meaningful kinetic rate expressions
can be obtained is of very great importance. This is particularly
true of catalytic reactions in which considerations such as external
and pore diffusional resistances come into the picture. In addition,
radial and axial transport effects must also be accounted for, and
therefore, the choice of a laboratory catalytic reactor has to be
made on the basis of several considerations.
In this review some of the important types of laboratory reactors
a r e discussed, with particular reference to the stirred reactor which
offers several advantages over other laboratory catalytic reactors.
The references cited here are not only those that describe new lab-
oratory reactor types per se, but also references in which a kinetic
study was the main purpose and the reactor used had some interest-
ing or unique aspect which would be important to the reader who has
a specific laboratory reactor design problem at hand. An effort has
been made to include a s many references to the various classes of
laboratory reactors as is possible so that the reader can judge the
merits of each relative to h i s own application. A t the end of this re-
view a brief summary is also presented of the various transport
 LABORATORYCATALYTIC REACTORS 179

effects that must be accounted for, and criteria are listed in a tabu-
lar form for neglecting these effects. Also, the various “diagnostic
tests” a r e briefly discussed and some warnings a r e given to pitfalls
among them.
The emphasis here will be placed on isothermal laboratory cata-
lytic reactors. It is also possible to obtain rate data from noniso-
thermal reactors, but due to the availability of isothermal reactors
which yield precise rate data, nonisothermal reactors a r e not dis-
Downloaded by [University North Carolina - Chapel Hill] at 08:05 15 July 2013

cussed in detail in this paper. Nonisothermal reactors a r e useful

for quick testing of catalysts and scoping of possible operating con-
ditions for maximum selectivity benefit. As will be elaborated later,
isothermality is of utmost importance in kinetic work. The reader
interested in gaining kinetic information from nonisothermal reac-
tors, as for example semiworks scale reactors, should consult the
articles of Beek and Hassell [l], Pexidr et al. [2], and Hawthorn et
al. [3], where this subject is handled in detail.
Although it is a valuable research tool, the “single pellet reactor”
will not be discussed in this review. This novel reactor is useful
mainly for measuring catalyst pore diffusion phenomena (with and
without reaction) and poisoning effects. The single pellet reactor
is expected to be reviewed in this journal by E. E. Petersen who has
done most of the development work on this reactor [4-61.

11. PACKED BED INTEGRAL AND DIFFERENTIAL REACTORS

A. Basic Principles

The most important feature of all experimental reactors is con-

stancy of temperature in the catalyst bed. A large catalyst bed
usually does not satisfy t h i s primary condition, so that experimental
catalyst beds a r e seldom more than 10 to 20 cm in length, the diam-
eter of the containing tube being of the order of 1 to 2 cm. (Much
smaller packed beds a r e possible and these “microreactors” will
be discussed later.) If plug flow is assumed in such a reactor, a
simple material balance over a differential element of catalyst, dW,
gives

r,dW = F dx (1)

where dx represents the differential conversion of A, rAis the reac-

tion rate (moles A converted per unit weight of catalyst per unit
time), and F is the feed rate (moles per unit time). Integration of
this equation between the limits
 180 L. K. DORAISWAMY and D. G. TAJBL

w=o, x=x,

w=w, x=x*

leads to the well-known result

Downloaded by [University North Carolina - Chapel Hill] at 08:05 15 July 2013

If, then, the reaction rate is known at various conversions, W/F

can be determined by graphical integration. But our concern here is
with the reverse problem of determining r Aat various conversions.
For this purpose a plot of x vs. W/F can be differentiated to provide
the desired values of rA. This is the principle of integral reactor
analysis, so-called because it involves analysis of experimentally
determined integral conversion data (with attendant concentration
profiles in the reactor).
Equation (1) shows that if a differential quantity of catalyst is
used, then the conversion obtained (dx) would be so small that the
compositions of the inlet and outlet streams would be almost iden-
tical, Under these conditions the experimental conversion gives the
rate directly, and t h i s rate corresponds to the partial pressures
(average of inlet and outlet values) of the reactants and products in
the differential bed. Thus the rate determined in this so-called dif-
ferential reactor would be equivalent to t h e rate in an integral bed at
a position where the partial pressures of the reaction components
are equal to those in the differential reactor.
Both integral and differential reactors a r e commonly used in rate
measurements, with several variations in design. These a r e briefly
discussed below.

B. Integral Reactors

The integral reactor has the inherent advantage that large conver-
sions are involved so that routine analytical methods are adequate.
Kinetic rate equation modeling using integral reactor data centers
around obtaining reaction rates from the x-(W/F) plot by some
numerical means. Several methods can be employed, and these
have been discussed by Corrigan [7]. The two most useful methods
a r e (1) graphical differentiation and (2) analytical differentiation of
an equation fitted to the data. An equation of the form

xR = a($) + b(;)' + c@y + d():4 (3)

 LABORATORY CATALYTIC REACTORS 181

usually fits the x-(W/F) data satisfactorily in the least squares sense.
Here xR is the conversion to product R, and a , b, c, and d are con-
stants that can be readily determined by a least squares analysis of
the experimental conversion data a t different values of W/F.
It can be readily seen that Eq. (3) would represent a smooth curve
through the data points. Since usually there is a considerable scatter
of the data points, such a curve cannot be regarded a s indicative of
the true picture. Let it be assumed that the curve drawn through
Downloaded by [University North Carolina - Chapel Hill] at 08:05 15 July 2013

t h e points is represented by Eq. (3) and that a rate expression can

be developed by differentiating this equation to obtain the experi-
mental rates. It is instructive now to examine to what extent such a
rate expression would be meaningful. Let us take data points 1 and 2.
The average rate calculated from these two points would be

where the subscripts 1 and 2 correspond to the two data points. When
average rates a r e calculated in this manner for successive pairs of
experimental points, it will be noticed that one would be entirely
justified in assuming zero-order, first-order, or Hougen-Watson
kinetics with a successive increase in accuracy. The question that
emerges, however, is: Is it really necessary to use a complex
kinetic expression when the e r r o r s involved in the experimental
data a r e as serious a s suggested by t h e above analysis? A s White
and Churchill [8] have pointed out, such a ‘differential treatment”
of the rate data must be regarded as a necessary adjunct to the
analysis of integral rate data, since it provides a clear indication
of the scope and accuracy of the data.
The main reason for the e r r o r s in the integral data is the lack of
isothermicity, which is particularly serious in highly exothermic
reactions. One remedy for this problem is to dilute the catalyst bed
with inert solids. Rihani et al. [9] have successfully employed a
diluted fixed bed in studying the kinetics of a highly exothermic r e -
action: catalytic vapor phase hydrogenation of nitrobenzene to ani-
line. They found that merely mixing the catalyst and diluent was not
adequate, and that different degrees of dilution were necessary in
different parts of the bed to ensure isothermal operation. By a trial-
and-error procedure a bed was formed which was sufficiently iso-
thermal to permit rate measurements. Calderbank et al. [lo] em-
ployed catalyst dilution a s a moderating influence in the design of
fixed bed reactors for exothermic reactions. They developed equa-
tions to be used for calculating the dilution required to maintain iso-
 182 L. K. DORAISWAMY and D. G. TAJBL

thermal conditions. This dilution is a function of distance along the

packed bed.
A closer examination of a diluted bed will show that several dis-
tributions of the diluent a r e possible, and that the exit reactant con-
centration can be influenced by the type of distribution. Although, in
practice, this effect of dilution is not significant in relation to the
isothermicity ensured by it, nevertheless the necessity to account
for t h e dilution effect (even in a randomly packed bed of catalyst
Downloaded by [University North Carolina - Chapel Hill] at 08:05 15 July 2013

and inert particles) is brought out by the circumstance that, under

conditions of high experimental precision, even a small effect due
to dilution (say of the order of 5%) can cause avoidable errors. Van
den Bleek et al. [ll]have recently attempted a theoretical analysis
of the problem of catalyst dilution in fixed bed reactors and have
formulated the following general criterion for neglecting the effect
of catalyst dilution:

B=
(1 - b)L6 < 4
bdp x 10-3 (5)

where 6 is the experimental e r r o r percent, b is the inert fraction,

and d, is the particle size.
Musick, Thomas, and Johnson [12] used an integral reactor with
several axial vapor sampling points to measure the kinetics of oxi-
dation of low- concentration multicomponent contamination in air.
Small samples of reacting mixture were withdrawn at points along
the bed and sent to a gas chromatograph. Such equipment allows
much data to be taken in a relatively short time.

C. Differential Reactors

Differential reactors conforming strictly to their definition are

difficult to operate. This is so in view of the differential conversion
involved, which calls for a high degree of analytical precision. In
addition to the high degree of analytical precision required, a prac-
tical difficulty in using t h e conventional differential reactor is the
necessity to prepare synthetic feeds conforming to various composi-
tions to facilitate the measurement of the reaction rate a s a func-
tion of composition. While this is possible in simple reactions with
known reaction paths, in complex reactions some compounds may be
formed which may be difficult to introduce in the feed o r whose
presence might not be known a priori. A simple way of surmounting
this difficulty is to use an integral reactor to provide the feed for
the differential reactor. Lunde and Kester [13] used this technique
in studying CO, methanation over ruthenium catalyst. By analyzing
t h e inlet and outlet streams of the differential reactor, the average
 LABORATORY CATALYTIC REACTORS 183

composition in the reactor can be determined so that the measured

rate would correspond to this composition. In any differential reac-
tor the rate can be directly computed from the relation
Downloaded by [University North Carolina - Chapel Hill] at 08:05 15 July 2013

where yAoand yAI represent the inlet and outlet mole fractions of A ,
respectively, and rA is the rate of disappearance of reactant A . It
has been assumed that the feed rate, F, is unaffected by the differen-
tial reaction. The very sensitive detectors now available for gas
chromatography have made operation of true differential reactors
possible in some cases. Pansing and Malloy [14] show that a reac-
tor heated by fluidizing solids can be used under conditions of differ-
ential conversion provided the reaction products a r e prevented from
condensing by injecting a stream of nitrogen, and the analysis is
carried out in a sensitive gas chromatograph equipped with a tem-
perature-programmed column and an electronic integrator.
The shortcomings of the differential reactor mentioned above can
be overcome by two methods to be discussed later: (1) recycling a
large part of the effluent stream or (2) using a stirred reactor.
These reactors may be regarded as pseudodifferential reactors in
that they give the rates directly (like the true differential reactor)
but at integral conversion levels.

111. INTEGRAL VS DIFFERENTIAL REACTORS: DATA ANALYSIS

There are several reasons why in some cases integral reactors,

provided they a r e isothermal, are preferred over differential reac-
tors. These reasons involve the statistical analysis of the data ob-
tained and its use in rate equation model discrimination. The papers
of Barnard and Mitchell [15], Mezaki and Happel [16], Peterson and
Lapidus [17], and Mears [18] a r e excellent references for the reader
interested in the relative merits of handling data from integral and
differential reactors. Included here is a short summary of the major
points of these papers.
Integral reactor data provide the following advantages over differ-
ential reactor data:

When using a differential reactor in complex reaction systems,

synthetic feeds must be prepared which match the composition at
various stages of conversion.
Integral reactor data a r e intrinsically more accurate than differ-
ential reactor data.
 184 L. K. DORAISWAMY and D. G. TAJBL

Data a t high conversion are usually preferred in discriminating

statistically between possible rate models.
Barnard and Mitchell [15] contend that measuring integral data
over the complete conversion-reciprocal flow rate graph gives more
accurate results than when only a portion of the graph is used.
Integral reactor data must (1) fit the conversion-reciprocal flow
rate graph and (2) vary in a particular way with reactant concentra-
tion; differential data have only to fit the measured variation of re-
Downloaded by [University North Carolina - Chapel Hill] at 08:05 15 July 2013

action rate with initial concentration.

However, the main disadvantage of integral reactors may be over-
whelming. The effect of temperature on the rate of a chemical reac-
tion is exponential, as described by the Arrhenius law. This means
that relatively small deviations from isothermality can lead to dis-
astrous e r r o r s in measured reaction rates for reactions with large
heat effects. Thus the choice between integral and differential labor-
atory reactors might be difficult. Fortunately, the reactor types to
be discussed subsequently a r e useful alternatives to t h i s dilemma.

IV. EXTERNAL RECYCLE REACTORS

A . Steady-State Reactors

If a differential reactor is operated such that most of the effluent

stream is recycled, a small amount of feed is continuously added,
and a small net-product stream continuously removed, a reactor
system is obtained which has several advantages over the conven-
tional differential reactor. Since recirculation rates 10 to 1 5 times
the feed rate are usually employed, the system would tend to be
more nearly isothermal. Also, the resultant high velocities past
the catalyst particles eliminate almost completely any external
mass transfer influence. By varying the circulation rate, the rate
beyond which the mass transfer is negligible can be established.
A l l kinetic experiments a r e then carried out at circulation rates
higher than this value. Steady-state recycle reactors of this type
have been described by Langer and Walker 11191, Perkins and Rase
[20], Temkin [21], and Satterfield and Roberts [22]. Sidorov et al. [23]
used a thermosyphon circulation system to study high-pressure am-
monia synthesis at 450°C.
A sketch of the recycle reactor system is shown in Fig. 1. In the
absence of recycle the rate is given by Eq. (6). If the final mole frac-
tion, yAt, is not very different from the initial value, yAO,then this
equation would give an average rate corresponding to some average
conversion. But these cease to have any meaning if ynf is consider-
 LABORATORY CATALYTIC REACTORS 185

PRODUCT OUT

t
Downloaded by [University North Carolina - Chapel Hill] at 08:05 15 July 2013

RECYCLE ( FR)
' 'At

1 A
',

FEED ( F )

FIG. 1. External recycle reactor.

ably less than ypi that is, a fair amount of reaction has taken place.
Thus in a conven ional differential reactor the conversion should
necessarily be very low. If now part of the product stream is r e -
cycled, a material balance over the entrance section (see Fig. 1)
leads to

Equation (7) clearly shows that yAI approaches yA, a s the recycle
 186 L. K. DORAISWAMY and D. G. TAJBL

rate, F,, is increased. Thus, a t high rates of recirculation, the reac-

-conditions and is given by
tion rate corresponds to the exit

The use of exit conditions is to be preferred since they conform to

conditions within the reactor (when there is full mixing). Knowing the
Downloaded by [University North Carolina - Chapel Hill] at 08:05 15 July 2013

exit composition and the reaction rate from Eq. (8), the kinetics of
the reaction can be elucidated. The use of exit conditions obviates
the need for employing synthetic feeds of different compositions.
A critical hardware problem associated with recycle reactors is
the recirculation pump. For careful kinetic or mechanistic studies,
the pump should not contaminate the reaction mixture. The pump
should have low holdup so that steady state is reached rapidly. Re-
circulation pump development h a s been done by Boudart e t al. [24]
and Bernard and Teichner [25]. Hanson and Benson [26] have de-
veloped a noncontaminating gas recirculation pump for use in gradi-
entless reactors. Their pump requires no cooling air or water, is
almost vibrationless, and a vacuum of about 2 X T o r r can be ob-
tained. A t 1 atm and a pressure drop of 20 T o r r s in the reactor loop,
the possible dry air flow rates ranged from 33 to 100 cc/sec. The
pumping rate varied nearly linearly with shaft rpm.

B. Batch and Semibatch Reactors

A batch recycle reactor can also be used instead of the steady-

state type described above. In this reactor a batch of the reactant
is continuously recirculated over a bed of catalyst. If the conver-
sion per pass is kept a t a very low value, then this would in effect
constitute a batch differential reactor system. By incorporating
suitable sampling points fitted with serum stoppers for withdrawing
samples through a hypodermic syringe, the concentration vs time
curve can be obtained for any given run. This can be repeated a t
several temperatures and pressures to provide the necessary data
for kinetic analysis. Reactor systems of this type have been used
by Butt et al. [27] and Cassano et al. [28].
A semi-batch recycle reactor has also been used for obtaining
kinetic data (Leinroth and Sherwood [29]), but t h i s is not a particu-
larly useful reactor to employ.
An important aspect of the recycle reactor is that the degree of
mixing can be calculated from the recycle rate, F,, and depending
on the value of F, the reactor can be operated to the extremes of
zero and complete mixing. Gillespie and Carberry [30, 311 have
given procedures for estimating the effect of recycle ratio on reac-
tor perform.ance.
 LABORATORY CATALYTIC REACTORS 187

V. AGITATED REACTORS

A. Continuous Stirred Gas-Solid Reactors

The stirred reactor is fundamentally similar to the external re-

cycle reactor described earlier in that both give a direct measure
of the rate from the measured difference between the inlet and out-
let concentrations (see Eq. 8). The principle features of these reac-
Downloaded by [University North Carolina - Chapel Hill] at 08:05 15 July 2013

t o r s a r e that gas-solid mass transfer resistance is almost eliminated

and perfect mixing can be achieved, thus making it possible to oper-
ate them under practically gradientless conditions.
Though the basic principles and the general features of all the
stirred reactors a r e the same, they can in general be divided into
groups and subgroups according to whether the catalyst is stationary
or nonstationary and the arrangement of the catalyst and stirring
assembly .
B. Rotating Basket Reactors

In this reactor the catalyst is held in a wire mesh container (bas-

ket) which is attached to a stirrer rotating inside a reaction pot.
Variations in design pertain chiefly to basket design and stirring
arrangement. When such a basket is rotated at high speeds, the
gas flowing into the pot would be instantaneously mixed with the re-
circulating gas, so that the concentration of the outflowing gas would
be identical with the bulk gas concentration in the pot, and the reac-
tion occurring on the catalyst pellets placed in the basket would be
dependent on this concentration. This reactor is superior to the r e -
cycle reactor in that very high speeds of rotation can be employed
(say up to 5000 rev/min) to ensure the absence of any mass transfer
effect. A reactor of this type was first proposed and used by Car-
berry [32] and Tajbl e t al. [33-361.
Sketches of three basket designs a r e shown in Fig. 2: (a) a basket
with four rectangular paddles at right angles to accommodate a single
layer of catalyst, (b) a basket with four cylindrical paddles at right
angle for mounting catalyst of any form, and ( c ) cylindrical and cru-
ciform baskets.
A disadvantage of the basket reactor is that the actual temperature
of the surface of the catalyst, or even in the field immediately exter-
nal to it, cannot be directly measured, since the basket (with its
catalyst-fluid system) constitutes the moving part, so that the tem-
perature measured is that of the fluid in the pot and hence one is
forced to accept predicted catalyst temperatures. This point is dis-
cussed further in a subsequent section.
There has also been some question whether all of the catalyst in
a paddle-type basket [Fig. 2(a)] “sees” the gas uniformly. For this
 188 L. K. DORAISWAMY and D. G. TAJBL
Downloaded by [University North Carolina - Chapel Hill] at 08:05 15 July 2013

( a ) FOUR RECTANGULAR PADDLE (b) FOUR CYLINDRICAL BASKETS MADE

BASKETS FOR SINGLE LAYER OF WIRE MESH FOR CATALYST OF
OF CATALYST PELLETS ANY FORM
(Carbcrry,[32] i Tajbl Gal,
[33j) (Choudhary and Doraiswarny, [45] 1

( c ) WIRE MESH CIRCULAR BASKET

( Brisk fiO1.[120])
FIG. 2. Basket designs.

reason it seemed that use of a very small catalyst particle size was
not recommended in such a reactor. However, studies by Tajbl [35]
in which small particles were confined in a specially-designed thin
basket show that the stirred basket reactor can match published re-
sults in which other reactor types were used. Furthermore, the ques-
tion of whether all of the catalyst is uniformly involved in reaction
was answered by W. D. Smith [37]. Professor Smith constructed a
stirred basket reactor like the design of Tajbl et al. Several mass
transfer tests were made to measure the efficiency of contacting of
gas and catalyst. In the early tests the baffles were inadvertently
absent from the reactor. In these tests indeed there was a distribu-
 LABORATORY CATALYTIC REACTORS 189

tion of contacting within the baskets. However, when the baffles were
installed, the contacting was in fact quite uniform. It is well known
that in stirred liquid-phase reactors baffles greatly improve the
contacting pattern. Baffles are also essential in the stirred basket
reactor.
Halladay and Mrazek [38] used a unique glass backmix reactor to
study heterogeneous and homogeneous decompositions of nitrous
oxide occurring simultaneously. The gold ribbon catalyst was
Downloaded by [University North Carolina - Chapel Hill] at 08:05 15 July 2013

mounted on support racks which were rotated at r a t e s sufficient to

insure perfect mixing. The paper also demonstrates that the r a t e
of chemical reaction occurring by simultaneous homogeneous and
heterogeneous paths can be separated into two contributing rates
for r a t e equation modeling purposes.
Trifiro et al. [39]studied the oxidation of butenes to maleic anhy-
dride in a s t i r r e d gas-solid reactor in which the catalyst was placed
in a cylindrical basket mounted on the rotating shaft between two
propellers. They checked the mixing characteristics by operating
at a constant temperature and getting a s e r i e s of conversions at
different shaft speeds and at the lowest flow rate. Their data indi-
cated perfect mixing prevailed a t rotation speeds greater than about
2200 rpm.
The catalyst in the stirred reactors described above is not sta-
tionary. Stirred reactors with stationary catalyst a r e described
below.

C. Reactors with Catalyst Impregnated on Reactor Walls o r

Placed in an Annular Basket

In some cases a thin layer of the desired catalyst can be directly

coated on the inner walls of the reactor (Ford and Perlmutter [40])
or on a removable liner fitted snugly inside the reactor (Lakshmanan
and Rouleau [41]), or the reactor wall itself can have catalytic prop-
e r t i e s (Wentzheimer [42]). The reactor is then provided with an effi-
cient stirrer to ensure thorough mixing [Fig. 3(a)]. Under conditions
of full mixing and because of the absence of any catalyst pore diffu-
sional intrusion (as the catalyst layer is very thin), the true kinetics
can be studied in any desired temperature range. Such a reactor is
almost a single phase s t i r r e d gas reactor with a large ratio of f r e e
volume to catalyst volume which limits its use in kinetics if non-
catalytic homogeneous reactions or homogeneous side reactions a r e
occurring simultaneously with the catalytic reaction. The tempera-
ture of the catalyst (that is, the reactor wall) can be directly mea-
sured.
In another design the catalyst is in the form of pellets or granules
and is placed in an annular basket made of wiremesh fitted close to
 190 L. K. DORAISWAMY and D. G. TAJBL

I
FEED
r--bPRODUCT

FEED
T P R O D U C T d
Downloaded by [University North Carolina - Chapel Hill] at 08:05 15 July 2013

CATALYST LINER OR
IMPREGNATED CATALYST

(0) REACTOR WITH CATALYST LINED t b ) REACTOR WITH CATALYST PLACED

OR COATED ON REACTOR WALL IN AN ANNULAR BASKET

FIG. 3. Reactors with catalyst impregnated o n reactor walls or placed in an annular

basket.

the reactor wall [Fig. 3(b)]. The use of such an annular basket-type
reactor has been reported by Tajbl et al. [34], Relyea and Perlmutter
[43], and Lakshmanan and Rouleau [44]. In t h i s reactor, direct mea-
surement of catalyst temperature is possible, but the large ratio of
fre e volume to catalyst volume limits its use if homogeneous reac-
tions occur. The reactor used by Tajbl et al. was easily converted
to the rotating paddle-basket type. It was designed for high-pressure,
high-temperature catalytic methanation of CO/H, o r CO,/H, mixtures.

D. Reactor with Catalyst Placed in a Stationary Cylindrical Basket

In this reactor, catalyst of any form and size or even a single pel-
let is placed in a cylindrical basket (provided with a fine wiremesh)
fitted at the center of the bottom of the reactor pot, and stirred tank
performance is obtained by rotating a special impeller [Fig. 4(a)],
 LABORATORY CATALYTIC REACTORS 191

o r the reactor pot with baffles [Fig. 4(b)] itself rotates around the
basket with some clearance (Choudhary and Doraiswamy [45]). The
ratio of the free volume to catalyst volume can be minimized by
proper design of the reactor pot, impeller, and basket. The im-
portant feature of this type of reactor is that the actual temperature
of the catalyst, which is one of the most important requirements in
kinetic studies, can be measured directly; the free volume to cata-
lyst volume ratio is low; and the reactor can be used a s a single
Downloaded by [University North Carolina - Chapel Hill] at 08:05 15 July 2013

pellet reactor and can also be conveniently employed for rapid

catalyst evaluations. This reactor can also be used for impregnated
catalyst by replacing the cylindrical catalyst basket by a cylinder
(closed a t both ends) with catalyst impregnated (or lined) on its cy-
lindrical surface, thus minimizing free gas volume as compared to
the reactors with catalyst impregnated on the reactor walls. They
can be operated under perfect isothermal conditions.

F E E D 1
n

,CYLINDRICAL
CATALYST BAiSKET

,IMPELLER
?=- THERMOWELL

INTERNAL
BAFFLES

EXTERNAL
2THERMOWELI BAFFLES

1 1 'CATALYST BASKET

FEED I LPRODUCT

( a ) STATIONARY CYLINDRICAL B A S K E T ( b ) STATIONARY CYLINDRICAL BASKET

REACTOR WITH ROTATING IMPELLER REACTOR W I T H ROTATING REACTOR POT

FIG. 4. Stationary cylindrical basket reactors.

A similar type of reactor has been employed by Costa and Smith

[46]for studying a gas-solid noncatalytic reaction: hydrofluorina-
tion of uranium dioxide using a spherical single pellet with a special
impeller designed to rotate around the spherical pellet with a small
clearance in order to approach stirred tank performance.
 192 L. K. DORAISWAMY and D. G. TAJBL

E. Internal Recirculation Reactor

In this reactor mixing is achieved by recirculating the reaction

mixture internally through a stationary catalyst bed using a spe-
cially designed impeller (turbine type). Variations in the design of
this reactor have been reported by Garanin et al. [47], Weychert
and Trela [48], Gelain [49],Berty et al. [50], Brown and Bennett [51],
Livbjerg and Villadsen [52], and Bennett et al. [53]. The recircula-
Downloaded by [University North Carolina - Chapel Hill] at 08:05 15 July 2013

tion rate may be determined by the rotational speed of a centrifugal

blower mounted at the top or bottom of the catalyst chamber. Two
typical designs a r e shown in Fig. 5.

r
FEED
I
I

FEED

CATALYST
/

CENTRIFUGAL 1
IMPELLER
I I

PRODUCT u IP

( a ) REACTOR WITH CATALYST ( b ) REACTOR WITH CATALYST

AT THE WALL AT THE CENTER
( Brown and Bennett, [Sl]) (Berty e,[50])
FIG. 5. Internal recirculation reactors.

The flow rate of the reacting mixture and the degree of recircula-
tion through the catalyst bed are well-defined quantities for these re-
actors and the actual rate of internal recycle can be estimated from
a few introductory experiments. Hence the mass and heat transfer
correlations developed for fixed bed reactors can be successfully
applied for estimating the mass and heat transfer rates from gas to
 LABORATORY CATALYTIC REACTORS 193

solid particles. Bertyl [54] has shown that the Union Carbide Auto-
clave Reactor (U.C.A.R.) can be used for process studies conducted
in the flow and mass transfer regime of the commercial scale, Thus
one can easily look in detail experimentally at the phenomena occur-
ring at a specific point in a commercial fixed bed reactor by operat-
ing the U.C.A.R. with the commercial catalyst formulation and a t
commercial Reynolds numbers. Mahoney [55] has employed a ver-
sion of the U.C.A.R. modified for operation at high temperatures and
Downloaded by [University North Carolina - Chapel Hill] at 08:05 15 July 2013

high pressures (1100°F max up to 1000 psig, 1000°F max up to 2000

psig) which a r e prevalent in petroleum processing.
Recently, Bennett et al. [53] have presented a review describing
internally recycled reactors incorporating various designs and their
advantages, and have also given a flow scheme to generate sharp
concentration signals for carrying out transient experiments in
kinetic studies. They have studied the kinetics of nitrous oxide de-
composition and carbon monoxide oxidation by transient experiments
in an internally recycled reactor.

F. General Considerations in Stirred Reactors

The various types of stirred gas-solid reactors mentioned above

have their own specific advantages and no a priori decision on the
choice of stirred reactor is possible. Choice is dependent on the
catalyst form, prior knowledge of the homogeneous reactions occur-
ring concurrently with the heterogeneous reaction (which dictates
the maximum allowable free volume to catalyst volume), and on the
thermodynamics of the reaction. The rotating basket-type reactor
should be used for studying catalytic reactions involving low heats
of reaction only, because for highly exothermic or endothermic re-
actions its use is limited due to the uncertainty in the predicted
catalyst temperature. The use of a reactor with catalyst lined on
the wall or placed in an annular basket within is recommended only
when homogeneous reactions do not occur to any appreciable extent.
On the other hand, reactors with stationary cylindrical basket and
internal recycle a r e versatile and may be used for studying any gas-
solid reaction.
The most important and basic requirement of the stirred gas-
solid reactor is that it should be operated under gradientless condi-
tions. It is, thus, necessary to minimize the resistance to heat and
mass transfer between the bulk gas phase and the catalyst surface
and also ensure complete mixing within the gas phase. Physical

The Union Carbide Autoclave Reactor has been thoroughly described recently in
the article by J. M. Berty, Chem. Eng. Progr., 70, 78, (1974).
 194 L. K. DORAISWAMY and D. G. TAJBL

measurements should, therefore, be carried out in order to evaluate

the reactor behavior both with respect to the heat-mass transfer and
mixing characteristics under different conditions of flow rate and
stirring speed before undertaking any kinetic study.
In order to determine the mass transfer coefficient experiment-
ally, the basket may be randomly packed with naphthalene pellets
and the loss in weight measured in a stream of nitrogen as a func-
tion of stirring speed, flow rate, and particle size. The mass trans-
Downloaded by [University North Carolina - Chapel Hill] at 08:05 15 July 2013

fer rate can thus be easily computed from the driving force obtained
by calculating the outlet concentration of naphthalene from the flow
rate and measured loss in weight. Based on this procedure, Choud-
hary and Doraiswamy [45]obtained the following j, correlations for
the three reactor types used by them:

Rotating cylindrical basket reactor [Fig. 3(b)]:

j, = 2.0 X 10-3Re-0-70 (9)
Stationary basket reactor with rotating impeller [Fig. 4(a)]:

Stationary basket reactor with rotating pot [Fig. 4(b)]:

j, = 2.0 X 10-3Re-0-55 (11)

where

Re = nd,Npd,/p

While these correlations can be used for reactors of design simi-

lar to those for which they were developed, separate mass transfer
studies should be carried out if the designs are even slightly differ-
ent. The surface temperature of the catalyst may be calculated by
estimating ,j through the following relationship between j, and j,:

j, = 1.37jD (13)
Mixing tests may be carried out by studying the kinetics of a sim-
ple (first-order) reaction, the true kinetics of which have already
been well established, o r (more directly) by a tracer technique.
Three tracer techniques are possible: (1) inject a small amount of
tracer into a steady stream of an inert gas over a short time (pulse
test); (2) switch over from a steady flow of inert gas to a steady flow
of tracer gas (step test) o r vice versa (purge test); and (3) introduce
 LABORATORY CATALYTIC REACTORS 195

a steady flow of tracer gas into a steady flow of inert gas (step test)
o r vice versa (purge test). If the mixing is perfect, the tracer re-
sponses to the above tests a r e a s follows:

Pulse Test:
Downloaded by [University North Carolina - Chapel Hill] at 08:05 15 July 2013

Step Test:

1 - (C/C,! = exp(-tQ,/V)

Purge Test:

C, is the gas concentration in the reactor a t the instant the pulse

enters, and in an actual test is the maximum gas concentration mea-
sured in the effluent; C, is the steady-state tracer concentration;
Q, is t h e volumetric flow rate; and V is the reactor volume.
Any of these tests can be carried out, the easiest perhaps being
the step test in which a steady stream of a tracer (carbon dioxide,
ethylene, etc.) is introduced into a steady stream of an inert (usually
nitrogen). A plot of ln[l - (C/C,)] vs tQ,/V must give a straight
line with a negative slope of unity if the reactor is perfectly mixed.
Deviations from such a linearity and the regions over which the de-
viations occur provide a clear indication of whether the reactor is
perfectly mixed and, if not, the region in which it may be regarded
as well mixed.
Three problems a r e faced in the design of stirred reactors: to
protect the shaft bearing from overheating when high temperature
reaction systems a r e studied, to seal effectively against leakage
along the shaft, and to minimize contamination of catalyst due to
possible poisons produced by the mixer hardware. The shaft bear-
ings may be protected from overheating by cooling them o r by using
a long shaft with the bearings at one end safely removed from the
high-temperature zone. The problem of leakage along the shaft may
be solved by using a hydraulic seal a t atmospheric pressure or a
magnetically coupled drive system.

G. Fluidized Bed Reactor with Agitation

The fluidized bed can be made fully mixed by providing suitable

baffles or by using a stirrer as shown in Fig. 6(a). A stirred fluid-
ized bed reactor of this design has been used by Trotter [56]. The
 196 L. K. DORAISWAMY and D. G. TAJBL

4
PRODUCT PRODUCT
Downloaded by [University North Carolina - Chapel Hill] at 08:05 15 July 2013

FEED
=--- PULSATOR
FEED

( a ) FLUlDlSED BED REACTOR ( b ) FLUlDlSED BED REACTOR

WITH MECHANICAL AGITATOR WITH PULSATOR

FIG. 6. Fluidized bed reactor with agitator.

basic disadvantage of this reactor is that the catalyst is subject to

disintegration due to the mechanical agitation involved. This disad-
vantage has been overcome by Ramaswamy and Doraiswamy 1571
who employed external agitation through a pulsator operating inde-
pendently of the feed device [Fig. 6(b)].
The main advantages of a pulsed fluidized bed in relation to a sim-
ple fluidized bed a r e that better mixing is obtained, a wider range of
particle sizes can be used, and better solid-gas contacting is possible.
Instead of using an independent pulsator, the feed itself can be intro-
duced through a pulse pump.
One should, however, exercise the greatest caution in choosing a
pulsed fluidized bed for kinetic studies by ensuring that suitable baf-
fles are provided and that the proper range of pulsation parameters
 LABORATORY CATALYTIC REACTORS 197

is employed. In view of the availability of other (more reliable) types

of reactors and the inherent uncertainty regarding complete mixing
in a pulsed fluidized bed, this reactor is not normally recommended
for securing kinetic data.

H. Ball Mill Reactor

A rotating mixer reactor (Fig. 7) which is based on the concept

Downloaded by [University North Carolina - Chapel Hill] at 08:05 15 July 2013

of the ball mill has been recently described by Barrett [58]. It offers
an alternative to fluidized bed systems for providing good solids
mixing. Mixing of solid and gas is obtained by rotating the reactor
which causes “cascading” of the solid particles. The temperature
difference between the particles and the gas is also reported to be
low (less than 2°C).

‘ ROTATING REACTOR

FIG. 7. Ball mill reactor.

VI. TUBULAR REACTORS

Packed bed reactors a r e usually characterized by short circuit-

ing of the catalyst particles by the fluid, resulting in irregular flow
patterns which affect the accuracy of rate measurements; also the
fluid may not “see” all the catalyst, so that the quantity of catalyst
actually involved in a reaction cannot always be unequivocally fixed.
Problems of inter- and intraphase transport of mass and heat a r e
involved, and except in very well designed reactors (say the basket
reactor) these effects cannot be altogether eliminated; one can a t
best account for them from generalized correlations and then ex-
tract the true chemical kinetic component of a rate measurement.
A simple way to overcome some of these problems is to deposit
the catalyst on the wall of an inert tube or to f i x a catalyst liner
within the tube. Such a reactor, usually called the tubular reactor,
has the following advantages: (1)the flow pattern is simple; (2) cata-
lyst surface temperature can be accurately controlled, leading to iso-
 198 L. K. DORAISWAMY and D. G. TAJBL

thermal operation; (3) all parts of the catalyst are equally accessible
to the reactants; and (4) samples may be obtained from any point in
the reactor. The principal experimental difficulty in such reactors
is that the tube diameter should be as small as possible in order to
minimize the free volume, which makes it impossible to obtain
velocity or concentration gradients experimentally.
The chief feature of this reactor is that its hydrodynamics a r e
well known, so that the true rate constant can be obtained easily
Downloaded by [University North Carolina - Chapel Hill] at 08:05 15 July 2013

provided the velocity profile is known. A s will be pointed out later

in this review, several criteria must be fulfilled for ensuring the
absence of diffusional effects in the integral and stirred reactors
already described. This is avoided in the tubular reactor since
the diffusional effect can, in principle, be precisely calculated from
hydrodynamic considerations.
These reactors, due to the very nature of their design and con-
struction, have not found widespread use in reaction engineering.
However, in view of some of their advantages in extracting kinetic
data from first-order systems in which the diffusional effect cannot
be ignored, they have been employed in investigating some systems.
It must be noted that, in theory, the first-order restriction is one of
convenience only, for conceivably solutions to the diffusion equation
can be found for a kinetic boundary condition defined by a reaction of
any order or mechanism (including reactions that follow Hougen-
Watson kinetics). However, one must remember that accurate kinetic
data in heterogeneous catalytic systems a r e difficult to secure and
that the introduction of a complicated mathematical problem which
can probably only be solved numerically will only make the job more
difficult .
The reactor used is usually 0.5 to 1 cm in diameter and about a
meter long. The catalyst is deposited by soaking or a liner is in-
troduced. The reactor is preceded by a calming section in order to
allow the velocity profile to develop fully before entering it. When
coated reactors of this type are operated under fully mixed condi-
tions, they can be considered a s pseudodifferential reactors, which
have already been described.
The first application of the tubular reactor in heterogeneous cata-
lytic studies was the work of Baron et al. [59]on the kinetics of SO,
oxidation catalyzed by vanadium oxide. Using the mathematical analy-
sis developed by Damk6hler [60] and with knowledge of the diffusivity
of the limiting reactant, they were able to calculate reaction rate
constants from their experimental data.
Katz [61] showed how steady-state data obtained when laminar
flow prevailed in an isothermal, round reactor could be used to de-
velop a kinetic model. The key to the applicability of the Katz analy-
sis is that the reaction rate and diffusion rate should be comparable.
 LABORATORY CATALYTIC REACTORS 199

In liquid systems t h i s balance is difficult to obtain, but in gas-solid

systems it might be obtainable. Lupa and Dranoff [62] numerically
determined the appropriate eigenvalues and eigenfunctions for the
Sturm-Liouville problem which results from the analysis of a cir-
cular annular reactor with a single catalytic surface. The catalytic
surface was assumed to be the inner convex surface, which would
be easy to prepare in practice and would permit better control of
temperature .
Downloaded by [University North Carolina - Chapel Hill] at 08:05 15 July 2013

VII. ADIABATIC REACTORS

The conventional method of obtaining rate data from isothermal

beds involves several analytical estimations, use of a recycle pump
in certain cases, and accurate temperature control. These can be
eliminated i f an adiabatic reactor system is used. In such a system,
preheated feed is passed into a thermally insulated reactor contain-
ing the catalyst bed and the temperature profile allowed to build up
in accordance with the kinetics and thermodynamics of the reaction
involved. If a one-dimensional model is assumed (which is usually
a valid assumption in an adiabatic reactor), and if interphase gra-
dients a r e minimized, then the experimentally measured longitudinal
temperature profile can give sufficient information to establish the
kinetics of the reaction.
The starting point in this procedure is the conservation equation
expressed in t e r m s of catalyst weight instead of reactor length:

dT
G A c A d~ - k,ApB d2T
hw" + r,[-AH,] =0

From the experimental axial profile of T v s W, numerical values of

dT/dW and d2T/dW2 can be obtained by graphical differentiation or
by analytical differentiation of an equation fitted to the experimental
profile. Then, if the effective thermal conductivity is known (either
from the literature or from a separate experiment), Eq. (17) can be
solved to provide values of rA at different temperatures.
The next step is to relate the measured temperature at any point
to the mole fractions of the various components of the reaction. This
can be easily done through an energy balance across a catalyst bed
which the reactants enter at a temperature To and leave at a temper-
ature T. Thus, if axial diffusion of heat and mass is neglected, and
Tois taken as the base temperature, the following equation can be
written (under adiabatic conditions) for Reactant A:
 200 L. K. DORAISWAMY and D. G. TAJBL

where % is the initial molecular weight of the feed and 6, is the in-
crease in number of moles per mole of A converted. Therefore,
Downloaded by [University North Carolina - Chapel Hill] at 08:05 15 July 2013

from a knowledge of the temperature profile, the mole fraction of A

at any point in the reactor can be computed. Similar equations can
be written for yB, yR, and ys, with [-AH] being expressed per mole
of the concerned component (B, R , o r S). More rigorous forms of
Eq. (17) should be used if the assumptions of constant C, and [-AH]
a r e to be relaxed, and an iterative technique for solving the equation
would then be necessary.
Equations (17) and (18) provide the basic data for testing different
r a t e expressions and selecting the most plausible one. For reactions
that follow simple power law kinetics, the testing procedure is
straightforward. If the reaction order is designated n, the rate equa-
tion can be written a s

or

Thus a plot of In (r,/[p,y,l") vs 1/T for an assumed value of n should

give a straight line of slope E/R, and intercept nln A, if the assumed
value of n i s correct. Several trials may be necessary before the
final kinetic expression is established. A single rate value at a given
temperature is adequate for this procedure; that is, one experimental
profile is sufficient in theory. In practice, however, more runs (at
least two) should be carried out so that reproducible values of A,
and E can be obtained; further, different values of n may give accept-
able straight lines, so that it would be necessary to test the linearity
for several runs with the selected value of n.
This method h a s been tested by Schmidt et al. [63] in a special
multibed adiabatic reactor. The catalyst bed is divided into several
cylindrical sections to enable gas mixing in the intermediate baffled
spaces and easy measurement of gas temperatures in these regions.
 LABORATORY CATALYTIC REACTORS 201

Further, cumulative build-up of radial gradients is avoided, so that

the one-dimensional assumption would be more truly valid. The
number and spacing of these beds can be altered at will to provide
temperature values at any desired number of axial positions. Mose-
ley e t al. [64] used a 0.44 c m i.d. reactor with three heating zones
that simulated closely the virtually adiabatic operation of large-
scale steam-hydrocarbon gasification units. The unit was used to
characterize loss of catalyst activity, but such a unit could be used
Downloaded by [University North Carolina - Chapel Hill] at 08:05 15 July 2013

for adiabatic kinetic experiments.

It must be ensured that the reactor is completely adiabatic. This
can be done by insulating it heavily and supplying controlled heat to
the insulation by heating coils wound round it. If the heating is so
adjusted (by manipulating the pitch and heat load of one of the coils)
that the temperature profile in the insulation is essentially the same
a s in the reactor (for the same run), then there would be no radial
transfer of heat through the reactor wall and the operation would be
adiabatic. A s a test of adiabatic behavior, the product issuing from
the reactor may be analyzed and [T - To] calculated from Eq. (18);
t h i s value should agree with the experimentally determined [T - To]
for the catalyst bed a s a whole. Note that if t h i s requirement is sat-
isfied, there would be no further need for any chemical analysis.
Using an adiabatic reactor similar in design to that outlined above,
Chakrabarti and Doraiswamy [65] studied the kinetics of the reduction
of nitrobenzene and obtained good agreement with rate data from
conventional reactors.
The use of an adiabatic packed bed reactor is not recommended
if interphase gradients a r e appreciable. A s Minhas and Carberry
[66] found in simulating an adiabatic, staged SO, converter, the
catalyst temperature can be significantly different from the bulk
gas phase temperature. The catalyst temperature is, of course, re-
quired in order for meaningful kinetics to be derived. Chambers
and Boudart [67]show how erroneous conclusions can be drawn
from experiments in which interphase gradients a r e not properly
accounted. They reanalyze some published experimental results in
which the experimenters postulated an unusual mechanism to account
for the fact that chemical equilibrium was violated. When, as Cham-
b e r s and Boudart show, the proper temperature, i.e., the catalyst
temperature, is used for correlation, the results a r e consistent with
thermodynamics.

VIII. MICROREACTORS
A significant recent development in laboratory reactor technology
is the microcatalytic reactor. A s the name implies, this reactor is
characterized by the use of a very small quantity of catalyst (usually
 202 L. K. DORAISWAMY and D. G. TAJBL

in the range 0.01 to 1.0 g), with correspondingly small quantities of

reactants. The advantages of such a reactor can be easily visualized.
In view of the microquantities involved, the heat of reaction would be
too small to cause any significant variation in temperature in the
bed; and as the length of the bed is usually less than 1.0 cm, the re-
action would proceed under practically isothermal conditions in the
external field. This means the apparatus can be relatively simple.
The nonisothermal character of the internal (or intraparticle) field
Downloaded by [University North Carolina - Chapel Hill] at 08:05 15 July 2013

in the case of exothermic o r highly endothermic reactions would,

however, remain unaltered. Since such a reactor is usually con-
nected to a gas chromatograph to enable the determination of product
composition, and since t h i s involves the use of a carrier gas in rela-
tively large quantities, it is clear that the resultant high velocities
would tend to eliminate any external mass transfer resistance to a
large extent. Another, rather obvious advantage of the microreactor
is that the raw material requirements a r e kept to a minimum, which
is of importance in the case of deuterium compounds and other such
r a r e materials. Also, any corrosion due to reactants or products
is likely to be negligible.
The microreactor can be operated either as a straightforward
steady-state reactor, like any of the differential (or integral) reac-
tors described earlier, or a pulse of reactant can be introduced into
the reactor and all the basic kinetic data extracted from a theoretical
analysis of the concentration pulse of the reactant through the reac-
tor. The latter is obviously an unsteady-state operation. Both these
types of operation a r e described in this section, with particular em-
phasis on the pulsed o r periodic operation since this has been the
subject of much recent experimental and theoretical work.
A s pointed out above, the microreactor can be operated in two
ways: a s a steady-state flow-type reactor and a s a pulsed (or
periodic flow) reactor. These a r e usually referred to a s “tail gas”
and ”slug” techniques, respectively, after Ettre and Brenner [68].
In the tail gas technique, the reactant is allowed to flow continuously
through the reactor, usually a small tube packed with catalyst, and
the products a r e continuously withdrawn. A special arrangement is
provided which enables periodic samples to be taken from the “tail
gas” and conducted into a gas chromatograph for analysis. The ex-
perimental scheme is illustrated in Fig. 8(a). A continuous stream
of carrier gas (usually hydrogen, helium, or nitrogen) is passed into
the sampling device S, from where it goes to vent, or is passed into
the chromatographic column C and then into the detector cells D.
Obviously the tail gas can be continuously taken through the chromat-
ograph if a continuous recording of the product composition is de-
sired. Note that the carrier gas does not pass through the reactor
and is used merely to pick up the tail gas for analysis. In principle,
 LABORATORY CATALYTIC REACTORS 203
Downloaded by [University North Carolina - Chapel Hill] at 08:05 15 July 2013

( a ) TAIL GAS TECHNIQUE

RG
I 4

I I

( b ) SLUG ( O R PULSE) TECHNIOUE

RG- REACTANTS R - MICROREACTOR

C G - CARRIER GAS C -CHROMATOGRAPHIC COLUMN
s - SAMPLING e DOSING DEVICE o -DETECTOR-CELL
FIG. 8. Tail gas and slug (or pulse) techniques [ 681.

this technique is not different from that of other steady-state experi-

mental reactors.
The slug technique, on the other hand, involves an unsteady-state
operation and is based on the concentration-time history of a slug
or pulse of reactant injected into the reactor through which a con-
tinuous stream of c a r r i e r gas is passing, The carrier gas flows
first into the reactor and then into the chromatographic column,
which is usually placed immediately following the reactor. The re-
actant pulse is injected into the c a r r i e r gas stream before the r e -
actor. The experimental arrangement is outlined in Fig. 8(b). Mathe-
 204 L. K. DORAISWAMY and D. G. TAJBL

matically t h i s is a far more complicated situation, as the governing

equation depends not only on the reaction order but also on the nature
of the pulse. Both techniques have been employed in obtaining kinetic
data (Yushchenko e t al. [691, Barbul et al. [70], Celler [71], and Luck-
n e r and Wills [?2]).
Consider a first-order rate equation:

rVA
= d x A / d t = -k,[l - x,]
Downloaded by [University North Carolina - Chapel Hill] at 08:05 15 July 2013

from which it is clear that the conversion in the reactor is inde-

pendent of the inlet concentration (2% and, therefore, of the nature
of the input pulse also. On the other hand, for a reaction of order n,
the rate equation is
A

rVA
= dCA/dt = kVChn-I[l - X*In (22)

so that the conversion in the reactor becomes a function of the inlet

concentration and, therefore, of the input pulse. Thus, unlike the
first-order case, the input pulse function forms a necessary boundary
condition in solving the nonfirst-order reactions.
Bassett and Habgood [73, 741 have shown that, if VR is the chromat-
ographic retention volume corresponding to W grams catalyst, Q, is
the volumetric feed rate of the reactant, and KA is the adsorption
equilibrium constant, then the following equation may be derived for
the conversion xA if we neglect any axial dispersion effects:

It will be seen that this equation is identical with the equation for
conversion in a first-order reaction occurring under steady-state
conditions. One has only to plot V, /QVl which comprises quantities
which a r e easily measurable, against In 1/(1 - xA) to obtain the
composite constant KAk,. There is seldom any necessity to deter-
mine k, and K A separately.
For reactions other than first order and where dispersion cannot
be neglected, the treatment gets more involved and will only be
briefly summarized here. Singh, Klinzing, and Coull [75] have done
a parametric study of pulse catalytic reactors in which axial dis-
persion, pulse shape, input amount, and reaction order were the
main variables. Some of their results a r e summarized here:

In nonlinear systems, dispersion effects depend strongly on pulse

shape, reaction order, and axial Peclet number.
 LABORATORY CATALYTIC REACTORS 205

For reactions of order n > 1, the conversion increases with de-

creasing pulse width and increasing input amount.
A s the reaction order increases, axial dispersion effects be-
come more serious. Also, larger dispersion e r r o r s will result
from thinner pulse shapes.
Reversible reactions are more seriously affected by dispersion.

They conclude that in any nonlinear system one should check for dis-
Downloaded by [University North Carolina - Chapel Hill] at 08:05 15 July 2013

persion effects by measuring the exit conversion a s a function of

pulse widths.
Merrill e t al. [76, 771 introduced a dispersion column ahead of
the catalyst bed to broaden the Gaussian-shaped pulses in their ex-
periments. This pulse-broadening technique allows one to neglect
the effect of axial dispersion on the peak width. Verma and Kalia-
guine [78] employed this technique and certain assumptions to sim-
plify the continuity equation in a pulse reactor study of ethylene oxi-
dation over silver. Their kinetic results compared well with steady-
state flow results, provided conversions were differential.
Galeski and Hightower [79], in a brief review of microcatalytic
reactor technology prior to 1970, indicate the utility of the micro-
catalytic technique to surface chemistry and poisoning studies. Small
amounts of reactants a r e in contact with the catalytic surface for
relatively short times in the slug technique. One can then easily
gather data on uncontaminated catalytic surfaces or on catalysts
contaminated to various extents.
Merrill et al. [77] extended the usefulness of the pulse-flow tech-
nique to the more general cases of nonfirst-order kinetics. Their
method involves matching the observed pulses with pulses computed
assuming certain kinetics. In principle, one should be able to dis-
tinguish between power-law and Langmuir-Hinshelwood-type kinetics
with their technique.

A. Typical Microreactors

In addition to the references cited above, several other important

microreactor studies a r e discussed briefly here. Kokes et al. [80]
were among the first to use a microcatalytic reactor for vapor-
phase catalytic reactions. Modifications in the original microreac-
to r have been suggested by Hall and Emmett [81], Keulemans and
Voge [82], Ettre and Brenner [68], Hartwig [83], Dutton and Mounts
[84], and Danforth and Roberts [85]. Semiautomatic (Hall et al. [86])
and completely automatic (Harrison et al. [87]) precision micro-
reactors have also been reported for obtaining data on reaction
kinetics and catalyst behavior. Recently, Binter and co-workers [88]
 206 L. K. DORAISWAMY and D. G. TAJBL

have described an integrated system comprising a microcatalytic

reactor combined with a gas chromatograph.
Sketches of a few typical microreactors a r e shown in Fig. 9.

1. This is the simplest microreactor and is similar to any steady-

state integral or differential reactor.
Downloaded by [University North Carolina - Chapel Hill] at 08:05 15 July 2013

CARRIER GAS
rTo GLC

GLASS WOOL GLASS WOOL

P R EHE ATER
loool V
J
0
GLASS W O O L 0
+ ( b ) VARIABLE CONTACT MICRO-REACTOR
( D u t t o n and Mounts,[84])

( a ) A SIMPLE MICRO-REACTOR
(Kokes =,[go] J

INJEC TOR PORTS

-CARRIER GAS

BY PASS TUBE

THERMOWELL

TO GLC COLUMN

( c ) T W I N T U B E INJECTOR MICRO-REACTOR
( D a n f o r r h and Raberts,[85] I

FIG. 9. Microreactors.
 LABORATORY CATALYTIC REACTORS 207

2. In this reactor there a r e separate provisions for introducing

the reactant and the carrier gas. The contact time can be varied a t
constant pressure and flow rate by altering the depth at which the
reactant is introduced into the catalyst bed through a long-needled
hypodermic syringe.
3. This is a more sophisticated design in which a special valve
permits t h e injection of feed at two points, Initially the valve posi-
tion is adjusted such that the c a r r i e r gas passes through the inert
Downloaded by [University North Carolina - Chapel Hill] at 08:05 15 July 2013

bed into which a pulse of the reactant is introduced. The effluent

is analyzed in the gas chromatograph attached. Then the valve posi-
tion is altered such that the c a r r i e r gas now passes through the
catalyst bed into which a reactant pulse (identical with the earlier
pulse) is introduced and the effluent analyzed in the gas chromato-
graph. Comparison of the reactant peaks in the effluent in the two
cases gives the conversion due to reaction directly. By introducing
the products of reaction into the first injection port (instead of the
single reactant as in the previous case), t h i s reactor can also be
used for reaction systems involving more than one reactant.

With the exception of Reactor (l), all the other reactors shown
in Fig. 9 can be used either a s pulsed o r steady-state microreac-
tors depending on whether the reactant is introduced in a single
short pulse or continuously. The ability to convert the reactor
easily from a slug reactor to a steady-state flow reactor is a
further advantage of the microcatalytic reactor.
A large number of catalysts can be readily evaluated for their
activity and selectivity by using any of the microreactors described
in the previous section at different reaction conditions (Hall et al.
1861, Norton [89], Steen e t al. [go], Norton and Moss [91], Topchieva
et al. 1921, and Yushchenko and Antipina [93]). The use of a set of
parallel microreactors, each containing a different catalyst fed by
a common pulse of reactants, permits the determination of relative
activity and selectivity data in a short period. It should be noted,
however, that the results obtained by this technique cannot describe
the process a s a whole; and if catalyst activity is time dependent,
as is often the case, the unsteady-state nature of the reactant pulse
exposure to a catalyst can yield deceptive data; hence, care must be
taken in selecting a catalyst on the basis of such data alone. This is,
in fact, the main disadvantage of the slug or pulsed microreactor,
Galeski and Hightower [79] point out: the partial pressures of t h e
components change continuously, the contact time may not be known
with certainty because of dispersion, and the residence time of the
reacting species may not all be the same because of adsorption-de-
sorption differences. These effects make securing meaningful ki-
netic data a difficult problem. It is the partitioning effect noted
 208 L. K. DORAISWAMY and D. G. TAJBL

above, i.e., the differences in adsorption-desorption behavior be-

tween reacting species, that can be used to advantage in a longer
chromatographic column reactor, which is the subject of the follow-
ing section.

B. Chromatographic Column Reactors

The chromatographic column itself can be used a s a reactor by

Downloaded by [University North Carolina - Chapel Hill] at 08:05 15 July 2013

packing it with catalyst for carrying out a chemical reaction under

chromatographic conditions. The special feature of this reactor is
that, under chromatographic conditions, not only a r e the products
of the reaction separated from the reactants, but they a r e also
separated from each other. Thus the reverse reaction does not take
place to any significant extent, leading to a sharp increase in the
conversion. Clearly, then, the use of a chromatographic column re-
actor is advantageous only for fast reactions with low equilibrium
constants which form at least two reaction products which can be
chromatographically separated from each other. Obviously, this
reactor will not be of any special advantage in cases where two re-
actants are present and which will also be separated in the chromato
graphic column, since in that case no reaction can take place at all.
The chromatographic column reactor would, therefore, be useful
only for reactions of the type

and

A+B----LC+D

with a very low equilibrium constant. Experimental and theoretical

studies of such reactions have been carried out by Masten e t al. [94],
Zimin et al. [95], and Langer et al. [96]. From their studies it has
been established that, provided the conditions stated earlier a r e
satisfied, the kinetics of the forward reaction can be easily studied
in a chromatographic reactor.
The use of these reactors h a s also been reported for reactions
of the type A === B (Klinkenberg [97]) as well a s for simple irrevers-
ible reactions (Roginskii et al. [981),but the results a r e not of any
practical value since the chromatographic reactor can truly be of
use only for reactions of the type mentioned earlier. Particularly
important uses of the chromatographic column reactor a r e in the
measurement of the physicochemical properties of solid catalysts
under reaction conditions and during catalysis, and in the determina-
tion of the effective diffusivity of gases in solid catalysts. The latter
is of great significance in catalytic reactor design.
 LABORATORY CATALYTIC REACTORS 209

Suzuki and Smith [99] obtained a solution for the chromatographic

curve resulting from a square injection pulse into an isothermal,
fixed bed reactor. The solution includes the effects of axial disper-
sion, external diffusion, intraparticle diffusion, reversible adsorp-
tion, and first-order reaction. The solution is then related to t h e
zeroth, first, and second moments of the chromatographic curve.
The general solution is rather complicated, but if experiments a r e
done in the regime where axial dispersion and fluid-to-particle
Downloaded by [University North Carolina - Chapel Hill] at 08:05 15 July 2013

diffusion a r e unimportant, then one can simplify the equations so

that the moments of the chromatographic peak a r e related to the
surface reaction rate constants. One can also evaluate the relative
importance of adsorption and surface rate processes using this
analysis.

IX. ISOTHERMAL HEAT TRANSFER MEDIA

Molten metal, solid metallic block, Dowtherm, fused salt, or in-

duction heating can be employed for heating the laboratory reactors.
Molten metal and fused salt systems can be hazardous although their
heat transfer properties a r e excellent. The most convenient and
efficient heating by far is provided by a fluidized solids bath. A
U-tube reactor is immersed into a fluidizing bed, one limb of which
serves to preheat the feed while the other holds the catalyst bed. If
such a reactor system is used, higher temperatures (say up to
1000°C) can be easily employed-which is not possible with other
common heat transfer systems. Also, temperature control is easier
and more accurate. Such a system was used by Noda, Hudgins, and
Silveston [loo]; Harrison, Hall, and Rase 1871; Livbjerg and Vallad-
sen [52]; Pansing and Malloy [14]; and Dalla Lana et al. [ l o l l . The
fluidized sand bath is easily constructed. It can also be obtained
commercially in several stock sizes, or custom manufactured to
given specifications. Verma and Kaliaguine [78] used a circulating
heating oil; Schriber and Parravano [lo21 used a molten salt bath
which operated between 246 and 345°C.

X. CRITERIA FOR THE ABSENCE OF HEAT AND MASS

TRANSFER EFFECTS

Rate data obtained under conditions of diffusional intrusion cannot

be used a s such for formulating meaningful kinetic expressions. A s
has been mentioned earlier, the chief characteristic of the various
laboratory reactors centers around the need for either eliminating
heat and mass transfer influences or accounting for them in a precise
 210 L. K. DORAISWAMY and D. G. TAJBL

manner. The only reactor for which the diffusional effect can be
precisely predicted is the tubular reactor. However, in view of the
limitations of this reactor, other reactor types, chiefly the stirred
reactor, a r e recommended for use in most kinetic studies. In these
reactors prior “calibration” of the reactor is necessary for a quan-
titative assessment of heat and mass transfer effects.
The purpose of this section is to outline briefly the various diffu-
sional effects which may be realized in laboratory catalytic reac-
Downloaded by [University North Carolina - Chapel Hill] at 08:05 15 July 2013

tors. These diffusional effects have been well-reviewed by others

(see Petersen [103], Satterfield 11041, Satterfield and Sherwood [105],
Mears [18], and Carberry [32, 1061); consequently only a summary
of the important conclusions is given here. General guidelines a r e
given for minimizing diffusional effects and equations from the lit-
erature a r e included in Table 1 for more accurate assessment. In
operating any of the reactors mentioned earlier it should be ensured
that the criteria as calculated from these equations a r e satisfied if
the kinetic data obtained a r e to reflect the intrinsic chemical be-
havior of the reaction involved. In using the equations, no more in-
formation is required than the diffusion coefficient, heat of reaction,
and the physical properties of the catalyst, together with an experi-
mental value of the observed, global reaction rate. Using these
values, the applicability of the various criteria can be easily tested.
Heat and mass transfer effects lead to temperature and concen-
tration gradients in the following regimes:

Intrapartic le -within the catalyst particles.

Interphase-between the external surface of the catalyst and the
adjacent bulk fluid phase.
Interparticle-between local bulk fluid regimes and from catalyst
particle to catalyst particle.

The magnitudes of the heat transfer effects in packed bed laboratory

reactors can be ordered:

Interparticle > Interphase > Intraparticle

The intraparticle transport effect, both isothermal and noniso-
thermal, has been analyzed for a multitude of kinetic rate equations
and catalyst particle geometries. It has been shown that the concen-
tration gradient within a catalyst particle is usually much more
serious than the temperature gradient. However, a s Hudgins [lo71
points out, intraparticle heat effects may not always be negligible
in H,-rich reaction systems. The classic experimental test to de-
termine the existence of intraparticle effects is to measure the de-
pendence of reaction rate on particle size, the lack of any dependence
 LABORATORY CATALYTIC REACTORS 211

indicating intraparticle effects a r e minimal. Koros and Nowak [lo81

have devised an alternate test which avoids any distortions in the
flow field which varying particle size might introduce.
Heat transfer effects become more serious, however, in the inter-
phase regime than in the intraparticle regime for vapor-phase sys-
tems. This is because the effective thermal conductivity of the gas
phase is much less than the effective thermal conductivity of the
catalyst. Indeed, bulk gas-catalyst temperature gradients can be-
Downloaded by [University North Carolina - Chapel Hill] at 08:05 15 July 2013

come quite large for reactions with large heat effects. Skinner and
Rase [ l o g ] used IR radiometry and a specially designed differential
reactor to measure steady-state and transient surface temperatures
on commercial nickel-kieselguhr catalysts during ethylene hydro-
genation. They noticed very large temperature gradients. Also, in
vapor phase systems, interphase heat transfer effects a r e more
serious than interphase mass transfer effects. Generally one can
expect that interphase transport effects will be noticed if reaction
r a t e s a r e high, heats of reaction a r e high, and flow r a t e s are low.
The usual experimental test for detection of interphase effects is to
check the effect of flow rate on the conversion while maintaining con-
stant space velocity. However, Chambers and Boudart [67]warn that
this test may be insensitive at the low Reynolds numbers usually en-
countered in laboratory reactors. The experimenter should plan on
running this diagnostic test over a wide range of flow rates and mea-
suring conversion very precisely.
Interparticle heat and mass transfer effects are manifest by
radial and axial temperature and concentration gradients in the r e -
actor a s a whole. Heat conduction problems due to a low effective
thermal conductivity of the bed can lead to severe radial tempera-
ture gradients in cases where the reaction rate and heat release a r e
large. Other factors such a s bypassing, axial dispersion, and heat
transfer resistance a t the wall further complicate the interparticle
problem. Generally, axial dispersion and axial heat conduction can
be neglected if the axial aspect ratio, L/d,, i s high (greater than
about 30), ensuring that plug flow is closely approached. Zahner
[110] has experimentally studied the axial dispersion problem in
laboratory reactors of various sizes. Guidelines a r e given in his
article for proper design to avoid axial dispersion effects. The
radial conduction problem is more serious. Radial mass transfer
effects a r e usually small compared to radial heat transfer effects.
The catalyst can be diluted (see Caldwell and Calderbank [lo]) to
spread the heat release more uniformly and thus avoid 'hot spots."
But dilution in large quantities introduces the hazard of bypassing.
Reduction of the reactor diameter, and thus the radial aspect ratio
R, /R, is helpful in cutting down radial temperature gradients, but
a s R,/R is reduced, the contribution of heat transfer a t the inside
 Downloaded by [University North Carolina - Chapel Hill] at 08:05 15 July 2013

TABLE 1
Criteria for the Absence of the Mass and Heat Transfer Effects

Interparticle transport
Interphase Intraparticle General (interphase and Axial mass
transport mass transfer intraparticle mass transfer) Radial heat transfer transfer

Mass Transfer Isothermal Isothermal

(Ruthven [ 1131) (Weisz and Prater (Guba and Narsimhan [ 1191) < 0.4 co
[1151) R . X In -
[1+8-@1)~,1 aw Cf
.. rvA (z) 0.15 Rt
11. Noniso thermal
<-
' CAkg n
1
(Mears [ 18J )
< ,;,,nfO
(Stewart and
Villadsen [ 116J ;
Mears [ 1141

(Mears [ 18I )
Downloaded by [University North Carolina - Chapel Hill] at 08:05 15 July 2013

LABORATORY CATALYTIC REACTORS

213
 214 L. K. DORAISWAMY and D. G. TAJBL

tube wall becomes increasingly more serious. The practical lower

limit of R,/R appears to be about 4.
The interparticle effects can be reduced by use of one of the agi-
tated reactors mentioned earlier. Interphase gradients can be re-
duced a s well since in most agitated reactors the velocity of the
vapor phase relative to the catalyst surface can be much higher
than in the packed bed with no recycle. The intraparticle effect is,
of course, not affected by reactor type, but rather by the kinetics
Downloaded by [University North Carolina - Chapel Hill] at 08:05 15 July 2013

and the catalyst physical structure.

Finally, it is important to mention that the packed bed reactors
with no recirculation and the recycled o r stirred reactors have dif-
ferent residence time distributions, In studying complex systems
of reactions, the residence time distribution of the reactor can
affect the selectivity of the system. A s Wei [ill] has shown, in a
reactor with a distribution of residence times (stirred tank or a re-
cycle reactor), all reactions are slowed down in comparison with
those in a plug flow reactor,but the faster reactions a r e slowed down
much more than are the slower reactions. Wei also shows that the
reaction paths will be more nearly straight lines. Also the classical
Wei-Prater analysis [112] for monomolecular systems must be
modified; that is, in stirred reactors or recycle reactors the decay
of the uncoupled characteristic species with time is not a simple ex-
ponential function a s it is in a packed bed which closely follows plug
flow behavior. When plug flow data a r e analyzed using the Wei-
Prater method, one can eliminate time in the characteristic species
equations and study the selectivity of the system easily. Time of re-
action, a s Wei and P r a t e r point out, is difficult to measure accurately.
When a stirred reactor is used, the selectivity equations in the space
of the characteristic species a r e rational functions of time and so
one is forced to use reaction time in the analysis. Therefore, if a
Wei-Prater analysis is being contemplated, one would probably be
advised to choose a reactor which approximates plug flow behavior
as closely a s possible.

SYMBOLS

A reactor cross section

A, Arrhenius frequency factor
b inert fraction
B dilution criterion defined by Eq. ( 5 )
[BijHw thermal Biot number at the wall
,CA concentrations
CA normalized concentration
c.4, surface concentration
 LABORATORY CA TA LY TIC REACTORS 215

exit concentration
initial concentration
maximum concentration
heat capacity
steady-state concentration
diameter of basket o r impeller
particle or pellet diameter
reactor diameter
Downloaded by [University North Carolina - Chapel Hill] at 08:05 15 July 2013

coefficient of molecular diffusion

axial eddy diffusivity
effective pore diffusivity
activation energy
void fraction
molar feed (or flow) rate
recycle rate
mass velocity
particle-fluid heat transfer coefficient
heat transfer coefficient at tube wall
heat of reaction
Colburn mass transfer factor
Colburn heat transfer factor
reaction rate constant
axial thermal conductivity
thermal conductivity of fluid
mass transfer coefficient
surface reaction rate constant
reaction rate constant
adsorption equilibrium constant
reactor length
average (or mean) molecular weight
reaction order
s t i r r e r speed
volumetric flow rate
total pressure
Peclet number (equal to ud,/D1)
reaction rate
radius of pellet o r particle
modified Reynolds number
gas constant
radius of tube
external surface area of pellet
total surface area
time
temperature
bulk gas temperature
 216 L. K. DOFWSWAMY and D. G. TAJBL

base temperature
surface temperature
wall temperature
average velocity
reactor volume
volume of pellet
retention volume
catalyst weight
Downloaded by [University North Carolina - Chapel Hill] at 08:05 15 July 2013

fractional conversion
mole fraction of Component A

Subscripts
0 initial o r inlet value
f final or outlet value
R product

Greek Letters
Arrhenius t e rm at the bulk temperature
Arrhenius t e rm at the wall temperature
dimensionless maximum temperature r ise
experimental e r r o r percent
increase in the number of moles per mole of A converted
external effectiveness factor
density of fluid
bulk density of particle or pellet
viscosity of fluid

REFERENCES

[ 11 J. Beek and H. L. Hassell, Proceedings of the Fourth European Symposium on

Chemical Reaction Engineering, Pergamon, Brussels, 1968,p. 229.
[ 21V. Pexidr, J. Cerny, and J. Pasek, Proceedings of the Fourth European Sympo-
sium on Chemical Reaction Engineering, Pergamon, Brussels, 1968,p. 239.
131 R. D. Hawthorn, G . H. Ackerman, and A. C. Nixon,Amer. Inst. Chem. Eng. J.,
14,69 (1968).
[4] J. R. Balder and E. E. Petersen, Chem. Eng. Sci., 23,1287(1968).
[ 51 J. R. Balder and E. E. Petersen, J. Catal., 11,195(1968).
[61 J. R. Balder and E. E. Petersen,lbid., 11,202 (1968).
[ 71 T.E. Corrigan, Chem. Eng., 62,199(April 1955).
[ 81 R. R. White and S. W. Churchill, Amer. Inst. Chem. Eng. J., 5,354 (1959).
[ 91 D.N. Rihani, T. K. Narayanan, and L. K. Doraiswamy. Znd. Eng. Chem., Process
Des. Deoelop., 4,403(1965).
[ 101 P. H. Calderbank, A. D. Caldwell, and G. L. Ross, Chim. Ind., Genie Chim.,
101(2), 215 (1969).
 LABORATORY CATALYTIC REACTORS 217

[ 11 ] C. M. Van Den Bleek, K. Van Der Wiele, and P. J. Van Den Berg, Chem. Eng.
Sci., 24,681 (1969).
[ 121 J. K. Musick, F. S. Thomas, and J. E. Johnson, Znd. Eng. Chem., Process Des.
Deuelop., 11,350 (1972).
[ 131 P. J. Lunde and F. L. Kester, Ibid., 13,27 (1974).
[14] W. F. Pansing and J. B. Malloy, Chem. Eng. Progr., 58(12), 53 (1962).
[ 151 J. A. Barnard and D. S. Mitchel1,J. Catal., 12,376 (1968).
[ 161 R. Mezaki and J. Happel, Catal. Reu., 3,241 (1969).
[ 171 T. I. Peterson and L. Lapidus, Chem. Eng. Sci., 21,655 (1966).
Downloaded by [University North Carolina - Chapel Hill] at 08:05 15 July 2013

[18] D. E. Mears, Ind. Eng. Chem., Process Des. Deuelop., 10,541 (1971).
[ 191 R. M. Langer and C. A. Walker, Ind. Eng. Chem., 46,1299 (1954).
[ZO] T. K. Perkins and H. F. Rase,Amer. Inst. Chem. Eng. J., 4,351 (1958).
[21] M. I. Temkin, Kinet. Kafal., 3,509 (1962).
[ 221 C. N. Satterfield and G. W. Roberts,Amer. Inst. Chem. Eng. J., 14,159 (1968).
[ 231 I. P. Sidorov, D. B. Kazarnovskaya, and P. P. Andreichev, Kinet. Katal., 3,523
(1962).
[24] R. P. Chambers, N. A. Dougharty, and M. Boudart, J. Catal., 4,625 (1965).
[ 251 J. R. Bernard and S. J. Teichner, Bull. SOC.Chim. Fr., 1969,3798.
[26] F. V. Hanson and J. E. Benson, J. Catal., 31,471 (1973).
[27] J. B. Butt, C. A. Walker, and H. Bliss, Amer. Znsf. Chem. Eng. J., 8,42 (1962).
[ 281 A. E. Cassano, T. Matsuura, and J. M. Smith, Znd. Eng. Chem. Fundam., 7,655
(1968).
[29] J. P. Leinroth and T.K. Sherwood, Amer. Inst. Chem. Eng. J., 10,524 (1964).
[ 301 B. M. Gillespie and J. J. Carberry, Ind. Eng. Chem., Fundam., 5,164 (1966).
[ 311 B. M. Gillespie and J. J. Carberry, Chem. Eng. Sci., 21,472 (1966).
[32] J. J. Carberry, Ind. Eng. Chem., 56(11), 39 (1964).
[ 331 D. G. Tajbl, J. B. Simons, and J. J. Carberry, Ibid., Fundam., 5,171 (1966).
[ 341 D. G. Tajbl, H. L. Feldkirchner, and A. L. Lee, Adu. Chem. Ser., 69,166 (1967).
[ 351 D. G. Tajbl, Can. J. Chem. Eng., 47,154 (1969).
[ 361 D. G. Tajbl, Ind. Eng. Chem., Process Des. Deuelop., 8,364 (1969).
[ 371 W. D. Smith, Paper Presented at Symposium on Progress Towards Gradientless
Reactors in Catalysis, 63rd Annual American Institute of Chemical Engineers
Meeting, Chicago, 1970.
[ 381 J. B. Halladay and R. V. Mrazek, J. Catal., 28,221 (1973).
[ 391 F. Trifiro, C. Banfi, G. Caputo, P. Forzatti, and I. Pasquon, Ibid., 30,393 (1973).
[40] F. E. Ford and D. D. Perlmutter, Chem. Eng. Sci., 19,371 (1964).
[41] R. Lakshmanan and D. Rouleau, Can. J. Chem. Eng., 47,45 (1969).
[ 421 W. W. Wentzheimer, Ph.D. Thesis, University of Pennsylvania, 1969.
[ 431 D. L. Relyea and D. D. Perlmutter, Ind. Eng. Chem., Process Des. Develop.,
7,261 (1968).
[44] R. Lakshmanan and D. Rouleau, J. Appl. Chem., 20,213 (1970).
[45] V. R. Choudhary and L. K. Doraiswamy, Ind. Eng. Chem., Process Des. Develop.,
11,420 (1972).
[46] E. C. Costa and J. M. Smith, Amer. Inst. Chem. Eng. J., 17,947 (1971).
[47] V. I. Garanin, V. M. Kirkchi, and K. M. Minachev, Kinet. Katal., 8,605 (1967).
[ 481 S. Weychert and M. Trela, Int. Chem. Eng., 8,658 (1968).
[49] C. Gelain, Chim. Ind., Genie Chim., 102,984 (1969).
[ 501 J. M. Berty, J. 0. Hambrick, T. R. Malone, and D. S. Ullock, American Institute
of Chemical Engineers Meeting, New Orleans, 1969.
 218 L. K. DORAISWAMY and D. G. T A J B L

[ 511 C. E. Brown and C. 0. Bennett,Amer. Inst. Chem. Eng. J., 16,817 (1970).
[ 521 H. Livbjerg and J. Valladsen, Chem. Eng. Sci.,26,1495 (1971).
[ 531 C. 0. Bennett, M. B. Cutlip, and C. C. Yang, Ibid., 27,2255 (1972).
[ 541 J. M. Berty, Paper Presented at Symposium on Laboratory Reactors for Process
Research, 66th Annual American Institute Chemical Engineers Meeting, Phila-
delphia, 1973.
[ 551 J. A. Mahoney, J. Catal., 32,247 (1974).
[ 561 I. Trotter, Ph.D. Thesis, Princeton University, 1960.
[ 571 V. Ramaswamy and L. K. Doraiswamy, Unpublished Material, National Chemical
Downloaded by [University North Carolina - Chapel Hill] at 08:05 15 July 2013

Laboratory, Poona, 1973.

[58] D. Barrett, Trans. Inst. Chem. Eng., 49,80 (1971).
[59] T. Baron, W. R. Manning, and H. F. Johnstone, Chem. Eng. Progr., 48,125
(1952).
[ 601 G. Damkohler, 2. Electrochem., 42,846 (1936).
[ 611 S. Katz, Chem. Eng. Sci., 10,202 (1959).
[62] A. J. Lupa and J. S. Dranoff,Ibid., 21,861 (1966).
[63] J. P. Schmidt, H. S. Mickley, and S. L. Grotch, Amer. Inst. Chem. Eng. J., 10,
149 (1964).
[64] F. Moseley, R. W. Stephens, K. D. Stewart, and J. Wood, J. Catal., 24,18 (1972).
[ 651 A. Chakrabarti and L. K. Doraiswamy, Unpublished Material, National Chemical
Laboratory, Poona, 1973.
[ 66[ S. Minhas and J. J. Carberry, Brit. Chem. Eng., 14(6),799 (1969).
[ 671 R. P. Chambers and M. Boudart,J. Catal., 6,141 (1966).
[68] L. S. Ettre and N. Brenner, J. Chromatogr., 3,524 (1960).
[69] T. M. Yushchenko, G. P. Korneichuk, V. P. Usha-Kova-Stasevich, and Yu. V.
Semenyuk, Katal. Katal., 4,154 (1968).
[70] M. Barbul, Gh. Serhan, I. Ghejan, and T. Filotti,Petrol Gaze, 19(3),181 (1968).
[71] W. Celler, Przem. Chem., 49(2), 68 (1970).
[72] R. C. Luckner and G. B. Wills, J. Catal., 28,83 (1973).
1731 D. W. Bassett and H. W. Habgood,J. Phys. Chem., 64,769 (1960).
[ 741 D. W. Bassett and H. W. Habgood, Chemical Institute of Canada, 44th Annual
Meeting, Montreal, 1961; Chem. Can., 13(5),50 (1961).
[ 751 H. B. Singh, G. E. Klinzing, and J. M. Coull, Paper Presented at Symposium on
Recent Advances in Reaction Kinetics and Catalysis, 66th Annual American
Institute of Chemical Engineers Meeting, Philadelphia, 1973.
[76] K. Makar and R. P. Merrill,J. Catal., 24,546 (1972).
[ 771 W. A. Blanton, C. H. Byers, and R. P. Merrill, Ind. Eng. Chem. Fundam., 7, 611
(1968).
[ 781 A. Verma and S. Kaliaguine, J. Catal., 30,430 (1973).
[79] J. B. Galeski and J. W. Hightower, Can. J. Chem. Eng., 48,151 (1970).
[SO] R. J. Kokes, H. Tobin, and P. H. Emmett, J. Amer. Chem. SOC., 77,5860 (1955).
[81] W. K. Hall and P. H. Emmett, Ibid., 79,2091 (1957).
[82] A. J. M. Keulemans and H. Voge, J. Phys. Chem., 63,476 (1959).
[83] M. Hartwig, Brennst.-Chem., 45,234 (1964).
[ 841 H. J. Dutton and T. L. Mounts, J. Catal., 3, 363 (1964).
[85] J. D. Danforth and J. H. Roberts,Zbid., 10, 252 (1968).
[86] W. K. Hall, D. S. MacIver, and H. P. Weber,Ind. Eng. Chem., 52,421 (1960).
[87] D. P. Harrison, W. K. Hall, and H. F. Rase,lbid., 57(1), 20 (1965).
 LABORATORY CATALYTIC REACTORS 219

E. D. Binter, V. L. Davison, and H. J. Dutton, J. Amer. Oil Chem. SOC.,46,

113 (1969).
C. J. Norton, Chem. Ind. (London), 6,258 (1962).
K. C. Steen, J. J. Feemann, L. Hofer, and R. B. Anderson, U S . Bur. MinesBull.,
608,19 (1962).
C. J. Norton and T. E. Moss, Ind. Eng, Chem., Process Des. Develop., 3,23
(1964).
K. V. Topchieva, E. Rosolovskaya, and 0. L. Shakhnoskaya, Vestn. Mosk.
Uniu., SerII, 23(1), 39 (1968).
Downloaded by [University North Carolina - Chapel Hill] at 08:05 15 July 2013

V. V. Yushchenko and T. V. Antipina, Zh. Fiz. Khim., 43,540,2982 (1969).

J. M. Masten, J. W. Harding, and E. M. Magee,J. Phys. Chern., 69,522 (1965).
R. A. Zimin, S. Z. Doginskii, and M. I. Yanovskii, Metody Issled. Katal. Katal.
Reakt., Red. l i e . Otd. Sihir. Otd. Akad. Nauk, SSSR, Novosibirsk, 3,279 (1965).
S. Langer, J. Yurchak, and E. Patton,Innd. Eng. Chem., 61(4), 11 (1969).
A. Klinkmberg, Chem. Eng. Sci., 15,255 (1961).
S. Z. Roginskii, M. I. Ianovskii, and G. A. Gaziev, Kinet. Katal., 3,529 (1962).
M. Suzuki and J. M. Smith, Chem. Eng. Sci., 26,221 (1971).
S. Noda, R. R. Hudgins, and P. L. Silveston, Can. J. Chem. Eng., 45,294 (1967).
I. G. Dalla Lana, D. E. McGregor, C. L. Liu, and A. E. Cormode, Proceedings of
the Fifth European/Second International Symposium on Chemical Reaction
Engineering, Elsevier, New York, 1972.
T. J. Schriber and G. Parravano, Chem. Eng. Sci.,22,1067 (1967).
E. E. Petersen, Chemical Reaction Analysis, Prentice-Hall, Englewood Cliffs,
New Jersey, 1965.
C. N. Satterfield, Mass Transfer in Heterogeneous Catalysis, M.I.T. Press, Cam-
bridge, Massachusetts, 1969.
C. N. Satterfield and T. K. Sherwood, The Role o f Diffusion in Catalysis,
Addison-Wesley, Reading, Massachusetts, 1963.
J. J. Carberry, Ind. Eng. Chem., 58(10), 40 (1966).
R. R. Hudgins, Chem. Eng. Sci., 23,93 (1968).
R. M. Koros and E. J. Nowak, Ibid., 22,470 (1967).
M. L. Skinner and H. F. Rase, J. Catal., 27, 335 (1972).
J. C. Zahner, Aduan. Chem., 109,210 (1972).
J. Wei, Can. J. Chem. Eng., 44, 31 (1966).
J. Wei and C. D. Prater, Advan. Catal., 13,203(1962).
D. M. Ruthven, Chem. Eng. Sci., 23,759 (1968).
D. E. Mears, J. Catal., 20, 127 (1971).
P. B. Weisz and C. D. Prater, Aduan. Catal.. 6,143 (1954).
1 W. E. Stewart and J. E. Villadsen, Amer. Inst. Chem. Eng. J., 15,28 (1969).
I E. E. Petersen, Chem. Eng. Sci., 20,587 (1965).

1 J. B. Anderson, Ibid., 18,147 (1963).

[119] B. K. Guha and G. Narsimhan, Ibid., 27, 703 (1972).
[ 1201 M. L. Brisk, R. L. Day, M. Jones, and J. B. Warren, Trans. Inst. Chem. Eng.,
46(l), T-3 (1968).