186
KINETICS OF THE VAPOUR-PHASE CATALYTIC DISPROPORTIONATION OF TOLUENE L.E. ANEKE f /.I? ^^^y

ke^etics of the vapour-phase catalytic disproportionation of

Embed Size (px)

Citation preview

KINETICS OF THE VAPOUR-PHASE CATALYTIC DISPROPORTIONATION

OF TOLUENE

L.E. ANEKE

f /.I? ^^^y

KE^ETICS OF THE VAPOUR-PHASE CATALYTIC DISPROPORTIONATION

OF TOLUENE

PROEFSCHRIFT

TER VERKRIJGING VAN DE GRAAD VAN DOCTOR IN DE TECHNISCHE WETENSCHAPPEN AAN DE Ti;CH-NISCHE HOGESCHOOL DELET, OP GEZAG VAN DE RECTOR MAGNIFICUS PROF. DR. JR. H. VAN BEKKUM, VOOR EEN COMMISSIE AANGEWEZEN DOOR HIT

COLLEGE VAN DEKANEN TE VERDEDIGEN OP WOENSDAG 4 FEBRUARI 1976 TE 14.00 UUR

DOOR

P1138 5104

C10026 42385

LINUS ENEMMOR ANEKE

MASTER OF SCIENCE

GEBOREN TE ABIA, NIGERIA / / 3 (P •*>""/ 0 y

1976

DrukkcriJ J.H. Pasmans, 's-Gravcnhagc

BIBLIOTHEEK TU Delft

P 1138 5104

264238

r

2

Dit proefschrift is goedgekeurd door de promotoren

Prof.Drs.P.J.van den Berg

Prof.Ir.W.A.de Jong

k

ACKNOWLEDGEMENT

I gratefully acknowledge my indebtedness to all the people who

contributed in various ways to the realization of the work reported

here. In particular, I would like to express my deep appreciation

to Prof.Dr.J.J.F.Scholten, Lector Ir.A.C.Montfoort, Dr.Ir.T.van

Herwijnen and Mrs.L.A.de Wit for their interest and stimulating

discussions; to Mr.R.Th.Nijvenheim and his colleagues for their

efforts in the construction of the equipment used. Sincere thanks

are also due to R.A.Betckt, J.Eilers, L.A.Gerritsen,

R.E.van Iddekinge and R.Trion for their assistance with the

experiments and in the development of new ideas. Finally, I would like

to thank my family and friends for their faith and understanding which

were a constant source of encouragement.

F

4

Like a dawn unheralded at midnight

it opened abruptly before me - a rough

circular clearing, high cliffs of deep

forest guarding it in amber-tinted spell

A long journey^ end it was but how

long and from where seemed unclear,

unimportant.

Chinua Achebe, The Explorer

DEDICATION

To my parents

Our ancestors, soul brother, were wiser

than is often made out.

Chinua Achebe, Beware,Soul Brother

6

CONTENTS

SUMMARY 9

LIST OF SYMBOLS 11

CHAPTER 1 Introduction 15

1.1 Economic Aspects of the Production of Aromatics 15

1.2 The Disproportionation of Toluene 17

1.3 Objectives of the present work 20

1.4 Synopsis of thesis 21

CHAPTER 2 Thermodynamics of the Disproportionation of Toluene 23

2.1 Introduction 23

2.2 Stability diagram 25

2.3 Calculation of Equilibrium Compositions 27

2.3.1 The disproportionation of toluene 28

2.3.2 The disproportionation and transalkylation of

toluene 32

2.4 Heat of reaction 37

2.5 Conversion, Selectivity and Yield of the disproportionation

of toluene 39

2.5.1 Feed consisting of toluene and hydrogen 39

2.5.2 Feed consisting of benzene, toluene, xylenes and

hydrogen 40

2.6 Conclusions 43

CHAPTER 3 Catalyst preparation and test of catalytic performance 45

3.1 Introduction 45

3.2 Experimental 48

3.3 Results and discussion 53

3.4 Conclusions 62

7

CHAPTER 4 Characterisation of the Physico-Chemical Properties

of the Catalysts 64

4.1 Introduction 64

4.2 Texture of catalysts 65

4.2.1 Introduction 65

4.2.2 Experimental 67

4.2.3 Results 68

4.3 Acidity of catalysts 81

4.3.1 Introduction 81

4.3.2 Acidity measurement by base titration method 83

4.3.3 Acidity measurement by ammonia adsorption method 85

4.4 Conclusion 98

CHAPTER 5 Kinetics of toluene disproportionation on an HY/6-A1F.-/

/Cu catalyst 100

5.1 Introduction 100

5.2 Measurement of the kinetics of heterogeneous catalytic

reactions 102

5.2.1 Differential reactor method 102

5.2.2 Integral reactor method 103

• 5.2.3 The Initial rate method 103

5.3 Influence of physical transport processes on the kinetics

of heterogeneous catalytic reactions 105

5.3.1 Non-ideality of the reactor 106

5.3.2 External and Internal Mass Transport Resistance 106

5.3.3 Pressure drop in the reactor 107

5.4 Reaction rate models 107

5.4.1 Power function rate models 108

5.4.2 Hougen-Watson models 108

5.5 Experimental 109

5.5.1 Materials 109

5.5.2 Equipment 110

5.5.3 Procedure 111

8

5.6 Results

5.6.1 Preliminary experiments

5.6.2 Kinetic measurements

5.7 Discussion and Conclusion

112

112

119

132

CHAPTER 6 Design Considerations 134

APPENDIX 1 Standard Free Energy and Enthalpy of formation of

components. 137

APPENDIX 2 Derivation of conversion, selectivity and yield. 139

APPENDIX 3 Adsorption Isotherms. 151

APPENDIX 4 Test of the Influence of transport processes. 154

APPENDIX 5 Internal normalization. 160

APPENDIX 6 Correction for catalyst deactivation. 161

APPENDIX 7 Derivation of a Langmuir-type rate model. 163

APPENDIX 8 F-test on the variances of the models. 165

APPENDIX 9 Formation of trimethylbenzene. 167

REFERENCES 169

SAMENVATTING 181

9

SUMMARY

Disproportionation is a potential alternative to methods of using

surplus toluene from the manufacture of aromatics. Although the

reaction may be carried out both in the liquid and vapour phases, the

latter is commercially the more important process and takes place in

the presence of solid acidic catalysts. Most of the work reported here

centres on the preparation and characterization of a catalyst with the

level of activity, selectivity and stability required for a commercial

process and for a study of the reaction kinetics.

First, the preparation of such a catalyst, designated ABl and with

the composition 72% HY/18% 6-AlF ./10% Cu, and the effect of some

process variables on its performance for the reaction are described.

The results reported show that the catalyst has good toluene

disproportionation performance and reveal that 500 C is its optimum

activation temperature, the activity all but disappearing when a

higher temperature is employed.

Subsequently, the physical and chemical properties of the

catalyst determined by analysis of nitrogen adsorption isotherms,

mercury penetration porosimetry and acidity measurements, are

discussed. The results of the texture determinations confirm that

most conventional methods for porous substances are inapplicable to

zeolites and zeolite-containing catalysts. Accordingly, a new method

is proposed and used to obtain values of the surface area in zeolitic

micropores which are in agreement with values computed from

crystallographic data. The mercury porosimetry experiments provide the

sizes of the voids between the catalyst particles as well as those of

the pores due to the 6-a.luminium fluoride and the copper present in

the catalyst. The results of the acidity measurements, determined by

n-butylamine titration and ammonia adsorption, are combined with the

texture of the catalysts to show that only about 10% of the total

surface area consists of acidic sites. Combination of the results of

texture and activity measurements also suggests that toluene

10

disproportionation activity of catalyst ABl is localized in its

transitional pores and that the micropores only serve to collect heavy

reaction products which would otherwise lead to deactivation. The

results of ammonia adsorption combined with the effect of activation

temperature on the catalytic activity suggest thatBrcinsted acid sites

are responsible for toluene disproportionation activity.

Reaction rate models derived from a number of postulated mechanisms

consisting of simple adsorption, surface reaction and desorption steps

are used to correlate the kinetic data. Non-linear weighted regression

analysis is applied to isolate a group of models with the smallest

variances. An F-test on these variances demonstrates that the difference

between those of the two models with the smallest values is not

significant at the 95% confidence level;thus, the two models are

equivalent from a statistical point of view. Experiments with reaction

products show that benzene has no measurable influence on the rate of

toluene disproportionation whereas xylenes have a definite retarding

effect.

The thesis concludes with some design considerations which are

regarded as relevant for the realization of an industrial toluene

disproportionation process.

11

LIST OF SYMBOLS

a reaction order 2

A area occupied by a molecule; m

response factor in GLC detection

ABl 72 w%HY/18 w% B-ALF /10w% Cu catalyst

b reaction order

B benzene

BET Brunauer, Emmett and Teller 3

C concentration kg/m

d reactor diameter m

2, d catalyst pa r t i c l e diameter ra D dif fusivi ty m /s

2

D. longitudinal dispersion coefficient m /s

E activation energy cal/mol

E apparent activation energy cal/mol a

E, porosity of catalyst bed

F F-value from F-test

g ammonia adsorption; ml/g

moles of a component in GLC detection moles

G Free Enthalpy (Gibb's Free Energy) cal/mole

HEXA hexamethylbenzene

HY Hydrogen Y zeolite

k reaction rate constant g moles/(g.s.atm)

K equilibrium constant;

moles xylenes per mole toluene in the feed;

response factor in GLC detection; 2

mass transfer coefficient g moles/(m .s.atm)

L total number of active sites on catalyst

surface;

reactor length m

m mole fraction in GLC detection

M molecular weight; kg/kg mole methane

12

MCH

n

N

N s

% P.P

p/p°

PENTA

''st

%.a

Q r

R

s

2 s

S

\ SSQR

T

TMB

TETRA

V

V

methylcyclohexane

number of data points in statistical

analysis

number of active sites

number of acid sites

Avogadro's number

pressure;partial pressure

relative pressure

Pentamethylbenzene

isosteric heat of adsorption

constants in Temkin isotherm

objective function

pore radius;

reaction rate

reaction rate;

gas constant;

hydrogen/aromatics ratio

standard deviation;

active site

variance

selectivity;

specific surface area;

adsorption entropy

bulk density of catalyst bed

sum of squares

toluene;

temperature

t r imethy1ben z ene s

tetramethylbenzenes

superficial fluid velocity

superficial fluid velocity;

pore volume;

methyl groups per phenyl group in the

feed

--

~"

--

--

--

atm

--

--

cal/mol

--

--

m

g mole/g.hr

g mole/g.hr

mol/raol

--

--

--2,

m /g

cal/mole.!'

kg/m

--

--

°C, K

--

--

m/s

m/s

ml/g

--

13

W/F space time g.hr/mol

X,x xylenes

Y mole fraction

Y_ yield of disproportionation products

3P LAL3P, a low-alumina silica alumina

5P LAL5P, steamed LAL3P

GREEK SYMBOLS

experimental response factor - —

intergranular porosity of catalyst bed --

volumetric flow rate ml/s 2

n ,y fluid viscosity N.s/m 3

p catalyst particle density kg/m

V

•J tortuosity factor

8 surface coverage

C conversion

DIMENSIONLESS GROUPS

P = axial Peclet number for mass transport u.d

R = Reynolds p a r t i c l e number

Sc = Schmidt number

B = Bodenstein number o

^L

u . p

y

V p . D

u.L

%

-

Pe-L

15

C H A P T E R 1

INTRODUCTION

1.1 Economic Aspects of the Production of Aromatics

Benzene, toluene and xylenes are the most commercially important

aromatic hydrocarbons. There are two traditional sources of these

aromatics:1) Fractional distillation of the light oil obtained as a

by-product of the pyrolysis of coal.

2) Catalytic reforming of naphtha fractions, where the aro­

matics are separated from the rest of the reformate by selective ex­

traction and the raffinate is refined into benzene, toluene and xylenes

by fractional distillation.

A third source, which is gaining in importance, is the fractional dis­

tillation of the benzene-rich by-product of the thermal cracking

(pyrolysis) of naphtha for the manufacture of ethylene.

Until shortly after World War II, industrial demand could be met

by the production from coal pyrolysis. After the war, however, the

demand rose so rapidly that increasing amounts were produced from

naphtha. Table 1-1 shows the typical composition of the aromatic

product obtained by the three sources mentioned above, while the pro­

duction of these aromatics in some of the major industrial countries

is shown in Table 1-2. Table 1-1 shows that catalytic reforming, which

by 1970 accounted for almost 100% of the total production of the aro­

matics, yields more toluene than either benzene or xylenes. However,

because of the use of large amounts of benzene for the manufacture of

nylon and polystyrene, and of p-xylene for the production of polyester

fibres, the demand for benzene and xylenes increased faster than that

for toluene. At the same time, the quantity of toluene commercially

available exceeded its demand as a chemical raw material, with the

result that the price of toluene is lower than those of benzene and

xylenes. This situation has led to attempts to develop processes based

on toluene as a raw material. The most important channels for the

16

Composition of aromatics produced, vol %

Component

Benzene

Toluene

Xylenes

Coal Pyrolysis

(184)

80

12

8

Catalytic Reforming

(168)

16

43

41

Naphtha Cracking

(184)

57

29

14

Table 1-1 Typical composition of the aromatic products from

various sources.

disposal of this surplus toluene are:addition to gasoline to produce

high octane fuel, hydrodealkylation to benzene and methane, and dis­

proportionation to benzene and xylenes,

The choice of a particular method of disposal is strongly in­

fluenced by the prevailing prices of benzene, toluene and xylenes.

When the prices of both benzene and xylenes are low, it may be more

economical to dispose of the toluene by blending it into gasoline.

If the price of benzene is high in comparison with that of toluene

and xylenes, hydrodealkylation may be the most economical process. If,

on the other hand, the availability of petroleum reformate is limited,

demand for xylenes is high, with the consequent high price of xylenes,

while the demand and prices of both benzene and toluene are low, dis­

proportionation may be the most attractive process for the conversion

of toluene.

Disproportionation has important advantages over hydrodealkylation.

In disproportionation, benzene and xylenes, both of which are commer­

cially more useful than toluene, are produced. Hydrodealkylation, on

the other hand, produces only benzene and methane. Since methane is

almost worthless in comparison with either toluene or benzene, hydro­

dealkylation is more sensitive to changes in feedstock and product

17

prices than disproportionation. Furthermore, hydrodealkylation uses up

hydrogen. If hydrogen is used in disproportionation in order to pre­

vent catalyst deactivation, it is not consumed and can, therefore, be

recycled. When hydrogen is not readily available, its price may ad­

versely affect the profitability of hydrodealkylation.

1.2 The disproportionation of toluene.

The disproportionation of toluene is the reaction in which toluene

is converted into a mixture of benzene and the isomers of xylene:

2(0/ TIZ;(Q} -C^CHJ

AH =0.8 kJ/mole toluene (800K)

U.S.A.

Japan

W.Germany

Benzene

Toluene

Xylenes

Benzene

Toluene

Xylenes

Benzene

Toluene

Xylenes

1969

3937

2478

1242

1221

590

542

538

141

124

1970

3861

2429

1679

1585

775

760

830

189

128

1971

3861

2429

1679

1698

791

856

867

181

183

1972

3665

2710

1742

1852

833

929

855

204

342

Table 1-2 Production (10 metric tons) of BTX in

some major industrial countries (174)

18

The reaction was first reported in 1884 in the pioneering work of

Anschiitz (70,71,72), who refluxed toluene at atmospheric pressure with

aluminium chloride as a catalyst, identified the products and postu­

lated a mechanism for the reaction. Besides benzene and xylenes from

the main reaction, hydrocarbons with higher boiling points and some

tarry products were formed by side reactions. Since that time, the

liquid-phase reaction has invariably been carried out over metal ha-

lides acting as classical Friedel-Crafts catalysts (73-76) .

The reaction also proceeds in the vapour phase. Solid catalysts,

such as silica-alumina and natural or synthetic zeolites, are used for

the vapour phase reaction. The first reported use of a solid catalyst,

silica-alumina, was in 1943 (80). After this time, most investigators

used silica-alumina (82) or boria-alumina catalysts (83). It is now

clear, however, that zeolites such as mordenite and faujasite are

superior in many respects to silica-alumina (87). Mordenite (88,90),

rare-earth-exchanged X-zeolite (87,89) and cation-exchanged Y-zeolite

(91,118) have been shown to possess the highest activity for this

reaction. Unfortunately, many of these catalysts show a low selectivity

as a result of hydrodealkylation and cracking reactions. Furthermore,

they have a low stability, with the result that the rate at which their

activity declines as a result of coke formation is usually so fast that

the activity decreases to a low value in a short time. In order to

improve their properties, the trend in the development of toluene

disproportionation catalysts has shifted towards the use of composite

catalysts (121). Satisfactory results have been obtained with alumina-

aluminium fluoride (92), mordenite-aluminium fluoride (94),

clinoptilolite-aluminium fluoride-copper (100), and mordenite-

aluminium fluoride-copper (93), combinations.

In spite of the attention which the process received after World

War II, especially in the U.S.A. and Japan, and the large number of

patents which have appeared on the subject (84), the disproportionation

of toluene as an industrial process is a recent development. In 1969,

Toray Industries (169,170,171) started up the first commercial plant

which is based on the reaction. In this process, shown schematically

19

in Fig.1-1, the reaction takes place in the vapour phase. Hydrogen is

used as a diluent gas in order to avoid coke deposition and, conse­

quently, catalyst deactivation. The reaction takes place at a pressure

of 30 atm, a temperature of 440°C, and with the molar ratio H./

aromatics in the reactor feed kept at 10. The concentration of hydrogen

in the make-up hydrogen stream is at least 70% by volume. The process

is claimed to be capable of an overall selectivity higher than 97%.

The details of the catalyst used have not been revealed. However, it

is probably a zeolite or a mixture of zeolites, alone or promoted with

other components.

A liquid phase process, the Mobil Low Temperature Disproportion­

ation (LTD) process, has been developed by Mobil Oil Corporation

(172,173). The process, shown in Fig.i_2, uses a zeolite catalyst at a

pressure of 45 atm and a temperature of 260°C. The process uses no

diluent gas, but the reactor temperature is raised from the initial

value of 260°C during the process cycle in order to maintain the

conversion at a satisfactory level. After the temperature reaches about

Toluene recycle ca-Aromatks C 0 - Arontotlcs

hydrogen rich gas recycle

^

A \ j

r~\

R«oclor Separator Froctlonotors

V Hlgn boiling fraction

C9-Aromotics r«cyci«

Figure 1-1 Toray toluene disproportionation process

20

Toluene • • Toluene recycle

Non-Aromat ics Benzene

• ^

r~\

r\

feed heater Reactor Cooler

Distillation columns 08-Aromatics

Figure 1-2 Mobil LTD (Low Temperature Disproportionation) process

300°C, the catalyst is regenerated by controlled burning of the coke

deposited on it. The life of the catalyst is claimed to be 1 years.

There is as yet no industrial plant based on this process.

1.3 Objectives of the present work.

The goal of the investigation reported in this thesis is to prepare

an active, selective and stable catalyst for the disproportionation of

toluene and, subsequently, to use the catalyst to arrive at a descrip­

tion of the kinetics of the reaction in the vapour phase. Although the

reaction has been known for almost a century, only a few kinetic

studies on it have been reported (83,84,85,90,175). Because of the poor

activity, selectivity and stability of some of the catalysts which have

been used in such kinetic studies (85,90), the results are of doubtful

accuracy. Inasmuch as a reaction rate equation is useful for an accu­

rate design of a commercial reactor for the process, the aim of the

kinetic experiments is to obtain such a rate equation.

Linde molecular sieve Y-zeolite, SK-40 was selected as the basic

material for the preparation of the catalyst because the hydrogen form

21

of this zeolite has a very high initial activity for the disproportion­

ation of toluene (88). Furthermore, preliminary experiments showed

that H-Y zeolite was much more active in this application than silica-

alumina and more stable than hydrogen mordenite. The results obtained

by other investigators with the type of composite catalysts mentioned

above suggest that promotion of hydrogen Y-zeolite with g-aluminium

fluoride and a metal such as copper may have a beneficial effect on its

selectivity and stability as a toluene disproportionation catalyst.

As previously mentioned, catalysts based on combinations of alumina,

mordenite or clinoptilolite with 6-aluminium fluoride and copper are

known. However, a catalyst in which hydrogen Y-zeolite is substituted

for these zeolites has not yet been mentioned in any of the large

number of patents and articles on this subject. Furthermore it was

considered potentially useful to try and prepare the B-aluminium

fluoride in a less laborious manner than has hitherto been the case

(93,108,109).

1.4 Synopsis of thesis.

After the introductory material dealt with in the present chapter,

the thermodynamics of the disproportionation of toluene is taken up

in chapter 2. This topic has not yet been sufficiently studied in the

literature, even though it is one of the prerequisites for an under­

standing of the relative importance of the many possible reactions of

toluene as a function of temperature, pressure and concentration.

The method of preparation and the toluene disproportionation ac­

tivity of the catalysts used in this study are described in chapter 3.

The results are presented in graphs of conversion and selectivity, at

"standard"reaction conditions, against time on stream. The graphs

facilitate comparison of the catalysts and show the effect of reaction

conditions, catalyst composition, and activation procedure on the dis­

proportionation activity of the catalysts.

Experiments aimed at the characterisation of some physico-chemical

properties of the catalysts are described in chapter 4. The texture,

that is the pore structure properties, of the catalysts is determined.

22

Texture plays an important part in the transport of reactants and

products in a catalyst and, therefore, exerts an influence on the ef­

fectiveness of a catalyst. For this reason, characterisation of the

texture of catalysts is essential for an understanding of their proper­

ties. Also, acidic properties are among the determining factors of the

activity of a toluene disproportionation catalyst. Hence the amount,

strength, type and distribution of acid sites on the catalysts are

determined as well. The limitations of the methods used and the sig­

nificance of the results obtained when applied to zeolites and zeolite-

based catalysts are discussed.

In chapter 5 the kinetic experiments are taken up. After a brief

introduction of the methodology for the study of the kinetics of

heterogeneous catalytic reactions, these methods are applied to the

kinetics of the disproportionation of toluene. The catalyst used is

selected on the basis of the work reported in chapter 3. The

derivation of the Langmuir-type reaction rate models which are used

to describe the kinetic data, and the significance of these models,

are discussed.

Finally, in chapter 6, the results of the experiments described in

previous chapters are combined and some design considerations of an

industrial toluene disproportionation process are examined.

23

C H A P T E R 2

THERMODYNAMICS OF THE DISPROPORTIONATION OF TOLUENE

2.1 Introduction

Pitzer and Scott (75) first reported on the thermodynamics of the

disproportionation of toluene. In their experimental investigation

liquid toluene was maintained in contact for five days at 50°C with a

catalyst consisting of an anhydrous AlBr /HBr mixture. Analysis of the

reaction products showed an abnormally low concentration of xylenes,

which they attributed to the consumption of xylenes by such side re­

actions as the formation of trimethylbenzenes They determined that the

equilibrium constant for the disproportionation of toluene in the liquid

phase lies between 0.15 and 0.22.

Later, Hastings and Nicholson (176) and Egan (177) independently made

a theoretical study of the gas-phase equilibrium.

Their calculations were based on the assumption that methyl group trans­

fer is the most probable reaction of methylbenzenes. The transfer re­

actions are assumed to proceed in a step-wise manner, the products of

previous transfer steps undergoing subsequent transfer reactions

(isomerisation and either disproportionation or transalkylation with

toluene, see Table 2-5). At equilibrium, a mixture of products consist­

ing of benzene and the twelve methylbenzenes is thermodynamically

possible.

As was pointed out in chapter 1, the vapour phase disproportionation

of toluene is carried out industrially with hydrogen as a diluent

gas in order to minimize coke-formation by cracking reactions and thus

avoid deactivation of the catalyst. In such a system many reactions

other than disproportionation are possible. The most probable among

these reactions are shown in Table 2-1. In these reactions, it is assum­

ed that toluene is the only methylbenzene present. Actually, similar

reactions are possible with the benzene, xylenes and trimethylbenzenes

formed as well as with the higher methylbenzenes produced in subsequent

24

1 Disproport ionation

C H T t.M3 1 -

'.(of ^ZliCo) 'C^CH3

2 Hydrodealkylation

CH3

( g / • H2 t = i (5 ) . CH4

3 Transalkylation

(of . ( ^ c H a t Z ; (o) . CH3 9^^3 CH3

CH3

CH-,

4 Cracking to Methane

CH3

(of • 10 Hg *• 7CH4

5 Cracking to Carbon and Hydrogen

CH3

- - - -'12 to/ 7C . 4H'3

6 Ring Hydrogenotion

CH3 9^3

(3/ . 3H2 • O

Table 2-1 Probable reactions of toluene in the presence of hydrogen

reactions. However, the reactions involving toluene are considered

the most probable, since the concentration of toluene is higher than

those of the other aromatic species present. Below, a stability dia­

gram (178) is used to explore the relative importance of the reactions

listed m Table 2-1. After calculation of the equilibrium conversion

of the disproportionation reaction as a function of temperature, the

heat effect of the reaction is considered. Subsequently, expressions

for toluene conversion, the selectivity of the disproportionation

reaction and the yield of disproportionation products are derived for

use m the analysis of the experimental results described in chapters

3 and 5.

25

2.2 Stability diagram

In order to bring the reaction equations into a form suitable for

making easy comparisons, the equations shown in Table 2-1 are re­

arranged so that each of them contains but a single molecule of

toluene:

1.

2.

3.

4.

5.

6.

T

T + H

T + >

T + lOH

T

T + 3H

-> 1/2B + 1/2 m-X

B + M

T + X J 2/3B + 1/3 1,3,5-TMB

•+ 7M

J 7C + 4H2

-> MCH

The sum of the free enthalpies of formation, lAG^/ of the products of

each of these reactions is calculated from the standard free enthalpy

of formation, AG^ , of the various components (Appendix 1). This sum

of free enthalpies is shown in Table 2-2 for each reaction and plotted

as a stability diagram in Figure 2-1. The distance between each line

and the line for toluene gives the free enthalpy change, AG°, of the

reaction concerned.

Temperature, K

reaction

1

2

3

4

5

6

300

29,80

18,95

30,17

-84,77

0,0

6,79

400

36,01

24,94

36,51

-70,49

0,0

21,84

500

42,58

31,39

43,21

-54,95

0,0

37,51

600

49,38

38,15

50,15

-38,57

0,0

53,55

700

56,36

45,15

57,26

-21,42

0,0

69,80

800

63,43

52,28

64,48

- 3,92

0,0

86,16

900

70,50

59,52

71,77

13,93

0,0

102,59

1000

77,80

66,85

79,13

32,06

0,0

119,03

Table 2-2:Sum of free enthalpies, £AG^, of the products of the

reactions of Table 1-1(in kcal/raol).

26

o E o u

C

O <

II

300 500 750 1000

^ — T in K

Figure 2-1 Stability diagram

The stability diagram shows that both the disproportionation and

transalkylation reactions have a slightly positive change in free en­

thalpy. This means that the conversion for each of these reactions is

incomplete, that is, high conversions are improbable. Below about 550K,

ring hydrogenation is thermodynamically favoured, while above this

temperature the reaction becomes unfavourable. Hydrodealkylation, on

the other hand, has a negative free enthalpy change throughout the

range of temperatures covered, which indicates that this reaction

probably goes to completion. The large negative value of the free en­

thalpy change of the cracking reactions means that they are the most

thermodynamically favoured of the reactions considered. Below 820 K,

cracking to methane is favoured over cracking to carbon and hydrogen.

Above 820 K the reverse is true.

The results of the thermodynamic analysis show that in a situation

150

100

-100

27

where all the reactions considered proceed at the same rate, the se­

lectivity of the disproportionation reaction will be vanishingly small.

In order to increase the selectivity for disproportionation, a catalyst

is needed which will selectively accelerate the rate of the reaction

at the expense of the hydrodealkylation and cracking reactions.

In the experiments described in chapters 3 and 5, analysis of the

gaseous product obtained after condensation and separation of the aro­

matics shows that only methane and hydrogen are present. This means

that if cracking to low-molecular weight hydrocarbons occurred, only

methane was formed (reaction 4 in Table 2-1). However, a methyl-group

balance shows that all the methane in the product can be accounted for

on the basis of the hydrodealkylation reaction. Thus, it can be con­

cluded that no cracking to methane or other low molecular weight hydro­

carbons occurs.

Cracking to carbon and hydrogen (reaction 5 in Table 2-1) results in

coke deposition on the catalyst. Catalyst deactivation is usually

attributed to this coke formation. From the slow deactivation of the

catalyst prepared in chapter 3 (30% in 3000 hours) and the fact that

the maximum carbon content determined for a sample of this catalyst

after deactivation with toluene at 500°C was 4.5%, this reaction can

be neglected under the conditions used in the experiments described in

chapters 3 and 5.

The product of ring hydrogenation, methylcyclohexane, was not detec­

ted in the reaction products. Therefore, this reaction (reaction 6 in

Table 2-1) can also be left out of consideration. This means that

reactions 1, 2, and 3 of Table 2-1 are the only reactions of toluene

which are relevant for further consideration.

2.3 Calculation of Equilibrium Compositions

Reactions 1 and 2 as well as 2 and 3 are parallel reactions.

However, while reactions 1 and 3 are equilibrium reactions, reaction

2 essentially goes to completion. On the other hand, reactions 1 and

3 are consecutive types. As previously mentioned, the methylbenzenes

formed in these consecutive reactions undergo transalkylation reactions

28

with toluene to yield other methylbenzenes. Furthermore, these

methylbenzenes undergo isomerisation reactions.

In the calculations that follow, two separate situations will be

considered. Firstly, it is assumed that only toluene disproportionation

and subsequent isomerisation of the xylenes occur. Hydrodealkylation,

as well as the transalkylations is neglected. Secondly, toluene

disproportionation accompanied by isomerisations and transalkylations

of higher methylbenzenes is considered. Again hydrodealkylation is

left out. The first situation would be applicable in a case where a

highly selective catalyst is available so that only toluene dispropor­

tionation and isomerisation of xylenes take place. The second is more

realistic because of the difficulty of finding such a selective

catalyst. In practical situations, satisfactory disproportionation

catalysts are also active for the isomerisations and, to a lesser

extent, for transalkylationsof methylbenzenes.

2.3.1 The disproportionation of toluene

Since the exact mechanism of the disproportionation of toluene is

unknown, it may be supposed that there are three disproportionation

reactions, one for the formation of each isomer of xylene:

2T J B + o-X

2T i B + m-X

2T J B + p-X

Since the experiments described in chapters 3 and 5 demonstrate that

the xylene isomers are in thermodynamic equilibrium, the system can

be represented by the following reactions:

2T J B + m-X 1

m-X J o-X 2

m-X I p-X 3

The following equilibrium relationships hold for these reactions:

Y, . Y 4 V _ b m-X

1 Y2 T

29

• 2

h-p-x

where K , K_ and K are equilibrium constants and Y is the mole frac­

tion. From the stoichiometry of the reactions and by using the

equilibrium relationships, the following expressions can be derived

for the mole fractions of the components:

Y^ = 1 - C 7

Y3 = (1-K2^K3)Y^_^ = 0.5 C

TEMP.

K

300

400

500

600

700

800

900

1000

AG°(1)

kcal/mol

1.07

1.42

1.75

2.12

2.49

2.90

3.31

3.77

AG°(2)

kcal/mol

0.78

0.88

0.95

1.00

1.06

1.10

1.14

1.18

AG°(3)

kcal/mol

0.55

0.68

0.82

0.96

1.11

1.27

1.42

1.58

h

0.166

0.168

0.172

0.169

0.167

0.161

0.157

0.150

h

0.270

0.330

0.384

0.432

0.467

0.501

0.529

0.552

h

0.397

0.425

0.438

0.447

0.450

0.450

0.452

0.452

Table 2-3: AG and K-values for the disproportionation of toluene

to m-xylene (1) and isomerisation of the m-xylene to

o-xylene (2) and p-xylene (3)

30

TEMP.

K

300

400

500

600

700

800

900

1000

Toluene

conversion

51.2

52.1

52.8

53.0

53.1

52.8

52.7

52.3

Benzene

25.6

26.1

26 4

26.5

26.6

26.4

26.4

26.2

Toluene

48.8

47.9

47.2

47.0

46.9

47.2

47.3

47.7

o-xylene

4.1

4.9

5.6

6.1

6.5

6.8

7.0

7.2

m-xylene

15.4

14.8

14.5

14.1

13.8

13.5

13.3

13.0

p-xylene

6.1

6.3

6.3

6.3

6.2

6.1

6.0

5.9

Table 2-4:Toluene conversion and equilibrium product concentrations

for the disproportionation of toluene (mole %)

Y = K„Y o-X 2 m-X

0.5C

9

10 (I+K2+K3)

Y = K^Y p-X 3 m-X

11

where E, is the fractional conversion of toluene. By substituting 7, 8

and 10 in 4 and rearranging, one obtains an expression for the conver­

sion in terms of the equilibrium constants:

1/2

S = 2(K^+K^K2+KjK3)

1*2(K^+K^K2+K^K3)

12

1/2

The values of K., K_, and K are calculated from the free enthalpy

changes, AG , of reactions 1, 2, and 3, which are in turn obtained

from the free enthalpy data of the components involved (see Appendix 1).

Table 2-3 shows these AG and K-values for temperatures between 300

and 1000 K. By assuming ideal gas and ideal mixture behaviour, the

31

toluene conversion and the concentrations of the species present can

be calculated from equations 7-12 for any particular temperature,

as is shown in Table 2-4 for temperatures between 300 and lOOOK. This

table shows that the toluene conversion and the product distribution

do not vary much over the temperature range covered.

No.

1

2

3

4

5

6

7

8

9

10

11

Reaction

2(Toluene) j' Benzene + m-Xylene

m-Xylene j" o-Xylene

K, m-Xylene ^^ p-Xylene

Toluene+m-Xylene j'* Benzene+1,2,4-Trimethylbenzene

1,2,4-Trimethylbenzeit -*- 1,2,3-Trimethylbenzene

1,2,4-Trimethylbenzene j^ 1,3,5-Trimethylbenzene

1 Toluene+1,2,4-Trimethylbenzene j^ B+1,2,3,5-Tetramethylbenzene

1,2,3,5-Tetramethylbenzene j^ 1,2,3,4-Tetramethylbenzene

1,2,3,5-Tetramethylbenzene ^^ 1,2,4,5-TetramethyIbenzene

Toluene+l,2,3,5-Tetramethylbenzene j Benzene+Pentamethylbenzene

Toluene+Pentamethylbenzene i Benzene+Hexamethylbenzene

Table 2-5:Reactions in the disproportionation and transalkylation of

toluene.

32

2.3.2 The disproportionation and transalkylation of toluene:

In this case, toluene undergoes disproportionation as described

above as well as transalkylation with the methylbenzenes formed. The

possibility exists t'nat benzene and the twelve methylbenzenes will be

present at equilibrium. The reactions which can occur are summarized

in Table 2-5. The following equations can be written for the equili­

brium constants of these reactions:

K. = Benzene.m-Xylene 1 2

(Toluene)

K_ = o-Xylene 2 m-Xylene

K_ = p-Xylene 3 m-Xylene

K. = Benzene.124-Trimethylbenzene 4

Toluene.m-Xylene

K_ = 123-Trimethylbenzene 5 124-Trimethylbenzene

K, = 135- Trimethylbenzene 6 124- Trimethylbenzene

K_ = Benzene.1235-Tetramethylbenzene 7 Toluene.124-Trimethylbenzene

K. = 1234-Tetramethylbenzene 8 1235-Tetramethylbenzene

Kg = 1245-Tetramethylbenzene 9 1235-Tetramethylbenzene

K. _= Benzene. Pentamethylbenzene 10 Toluene.1235-Tetramethylbenzene

K.,= Benzene.Hexamethylbenzene 11 Toluene.Pentamethylbenzene

In the equations above, the names of the components designate their

mole fractions. It is assumed that the reaction mixture is ideal so

that the fugacity and activity coefficients are both unity. Further­

more, since each reaction consists of the same number of moles of

33

products as reactants, the total pressure does not enter into the

expressions for the equilibrium constants.

The eleven equations above contain thirteen variables. Two more

equations are required to define the system completely. These are the

methyl group and the phenyl group balances:

T+2(o-X+m-X+p-X)+3(123TMB+124TMB+135TMB)+4(1235TETRA

+1234TETRA+1245TETRA)+5(PENTA)+6(HEXA) = 1

B+T+o-X+'m-X+p-X+123TMB+124TMB+135TMB+

1235TETRA+1234TETRA+1245TETRA+PENTA+HEXA = 1

It is not possible to solve the thirteen equations above explicitly

for the mole fractions. They were solved numerically on an IBM 360/65.

To this end, equations 1-11 were rearranged to derive the mole frac­

tions in terms of those of two independent components (180): m-xylene

and 1,2,4,-trimetlylbenzene. The expressions for the mole fractions of

benzene and toluene were obtained by combining equations 1 and 4, and

were substituted in the other equations above where necessary to derive

those of the other componenets. The resulting expressions are:

2 3 K^m-X 12

' - - 2 Kj.I24TMB

K,.m-X^ 13 J = _4

K^.124TMB

o-X = K2.m-X 14

p-X = K^.m-X 15

123TMB = Kg.l24TMB 16

13STMB = K^.124TMB 17 D K.,.124TMB 18

1235TETRA = _£ K^.m-X

34

K.,.Ko.l24TMB iq 1234TETRA = 7 S i»

K..m-X 4

K .K 124TMB^ 20 123i5TETRA =—^ ^

> K .m-X

K.,.K,„.124TMB^ 21 PENTA = '^' ^0

2 2 K^.m-X 4

K,„.K,,.K^.124TMB^ 22 HEXA = 1° ^1 '

3 3 K,.m-X- 4

Equations 12-22 were solved by assuming successive values for m-X and

124TMB, until the mass balance criteria (the conservation of the number

of methyl and phenyl groups) were fulfilled, the sum of the mole frac­

tions then being unity. The standard free enthalpies of formation used

in the calculations are listed in Appendix 1. The results of the cal­

culations are shown in Tables 2-6 and 2-7 and in Figure 2-2. The change

in free enthalpies, AG , and the equilibrium constants, of the reac­

tions are given in Table 2-6. The results show that, as in disproportionation

alone, the equilibrium conversion and product distribution are not

strongly dependent on temperature.

The results obtained here are in statisfactory agreement, but not

identical, with those of Egan (177) and Hastings and Nicholson (176)

for the same problem. The difference in results is attributable to

differences in the standard free enthalpies, AG„, used. Hasting and

Nicholson used a method suggested by Kandiner and Brinkley (180) to

derive their equations. The discussion presented above demonptrates

that the same relationships, equations 12-22, can be derived analyti­

cally and not simply by visual inspection as was done by Hastings and

Nicholson. Egan did not specify the equations used in his calculations.

The results obtained here may be considered an improvement over those

of Egan and Hastings and Nicholson in view of the more up-to-date free

enthalpy values used.

35

Temperature, K

Reaction

1

2

3

4

5

6

7

8

9

10

11

AG,

K

AG,

K

AG,

K

AG,

K

AG,

K

AG,

K

AG,

K

AG,

K

AG,

K

AG,

K

AG,

K

300

1.07

0.17

0.78

0.27

0.55

0.40

1.38

0.10

1.83

0.05

0.24

0.67

2.27

0.02

1.12

0.15

0.17

0.75

2.94

0.01

3.49

0.003

400

1.42

0.17

0.88

0.33

0.68

0.43

1.70

0.12

2.08

0.07

0.50

0.53

2.81

0.03

1.26

0.20

0.28

0.70

3.63

0.01

4.49

0.004

500

1.75

0.17

0.95

0.38

0.82

0.44

1.99

0.13

2.34

0.09

0.78

0.46

3.31

0.04

1.35

0.26

0.37

0.69

4.24

0.01

5.4

0.004

600

2.12

0.17

1.00

0.43

0.96

0.45

2.30

0.15

2.61

0.11

1.06

0.41

3.79

0.04

1.44

0.30

0.48

0.67

4.83

0.02

6.29

0.005

700

2.49

0.17

1.06

0.47

1.11

0.45

2.62

0.15

2.88

0.13-

1.35

0.38

4.24

0.05

1.50

0.34

0.60

0.65

5.37

0.02

7.14

0.005

800

2.90

0.16

1.10

0.50

1.27

0.45

2.95

0.16

3.16

0.14

1.64

0.36

4.69

0.05

1.56

0.37

0.72

0.64

5.90

0.02

8.00

0.007

900

3.31

0.16

1.14

0.53

1.42

0.45

3.29

0.16

3.43

0.15

1.93

0.34

5.12

0.06

1.61

0.41

0.84

0.63

6.41

0.03

8.82

0.007

1000

3.77

^.15

1.18

0.55

1.58

0.45

3.65

0.16

3.71

0.15

2.23

0.33

5.57

0.06

1.66

0.43

0.98

0.61

6.93

0.03

9.67

0.008

Table 2-6: AG O*-cal/mol) and K-values for the disproportionation

and transalkylation of toluene.

36

1 TEMPERATURE, K |

Toluene Conversion

Benzene

Toluene

1,2-Dimithyl benzen

1,3-Dimethyl benzene

1,4-Dimethyl benzene

1,2,3-Trimethyl benzene

1,2,4-Trimethyl benzene

1,3,5-Trimethyl benzene

1,2,3,4-Tetra methylbenzene

1,2,3,5-Tetra methylbenzene

1,2,4,5-Tetra methylbenzene

Pentamethyl benzene

Hexamethyl benzene

300

53.4

27.0

46.6

3.6

13.4

5.3

0.1

2.3

1.5

0,2

0.0

0.0

0.0

0.0

400

54.3

27.3

45.7

4.3

12.9

5.5

0.2

2.6

1.4

0.1

0.0

0.0

0.0

0.0

500

55.0

27.7

45.0

4.8

12.5

5.5

0.3

2.7

1.2

0.2

0.0

0.1

0.0

0.0

600

55.4

27.5

44.6

5.3

12.2

S.5

0.3

2.9

1.2

0.3

0.0

0.2

0.0

0.0

700

55.5

27.5

44.5

5.6

12.0

5.4

0.4

3.0

1.1

0.2

0.1

0.2

0.0

0.0

800

55.4

26.9

44.6

6.0

11.9

5.4

0.4

3.1

1.1

0.3

0.1

0.2

0.0

0.0

900

55.2

27.3

44.8

6.1

11.6

5.2

0.4

3.0

1.0

0.3

0.1

0.2

0.0

0.0

1000

55.5

27.5

44.5

6.3

11.4

5.1

0.5

3.1

1.0

0.3

0.1

0.2

0.0

0.0

Table 2-7:Toluene conversion and equilibrium product concentrations

for the disproportionation and transalkylation of toluene

(mole %)

37

400 600 800 1000 Temperature . K

Figure 2-2 Toluene conversion and equilibrium composition

o Toluene conversion

• Benzene;x Toluene;V o-Xylene;* m-Xylene

• p-Xylene;A 1, 2, 4-Trimethylbenzene;

• 1, 3 , 5-Trimethylbenzene

2.4. Heat of reaction

The heat effects of the disproportionation, isomerisation and trans­

alkylation reactions(l-ll) listed in Table 2-5 are shown in Table 2-8.

Reactions 12-15 in Table 2-8 are, respectively, the hydrodealkylation

of toluene, the hydro-cracking of toluene to methane, the thermal

cracking of toluene to carbon and hydrogen, and the hydrogenation of

toluene to methylcyclohexane. The standard enthalpies of formation of

the components are shown in Appendix 1. Table 2-8 shows that the heat

effects of the disproportionation,isomerisation and transalkylation

38

Temperature, K

Reaction

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

300

0.02

0.4

0.2

0.4

1.0

-0.5

0.5

0.7

-0.1

0.8

0.4

-10.0

-137.2

-11.9

-49.0

400

0.03

0.5

0.1

0.5

1.0

-0.6

0.7

0.8

-0.1

1.0

0.7

-10.3

-140.8

-10.3

-50.1

500

0.01

0.6

0.1

0.5

1.0

-0.6

0.9

0.9

-0.2

1.3

0.9

-10.8

-144.2

-9.1

-51.0

600

-0.04

0.7

0.1

0.4

1.0

-0.7

1.0

1.0

-0.2

1.5

1.1

-11.2

-147.3

-8.0

-51.5

700

-0.1

0.7

0.04

0.4

1.0

-0.7

1.1

1.1

-0.2

1.6

1.2

-11.6

-150.0

-7.2

-51.7

800

-0.2

0.8

0.01

0.3

1.0

-0.7

1.2

1.1

-0.3

1.7

1.3

-12.0

-152.4

-6.7

-51.7

900

-0.3

0.8

-o.o;

0.1

0.9

-0.7

1.2

1.2

-0.3

1.8

1.3

-12.3

-154.3

-6.2

-51.6

1000

-0.4

0.8

I -0.03

0.02

0.9

-0.7

1.2

1.3

-0.3

1.8

1.4

-12.6

-155.8

-6.0

-51.2

Table 2-8:Heats of the reactions of toluene, AH J: cal/mol), in the

presence of hydrogen.

reactions (reactions 1-11) are small. This is in line with the above

(section 2-3) conclusion that the equilibrium conversion and product

distribution are not strongfy dependent on temperature. The other reac­

tions, hydrodealkylation, hydro-cracking, thermal cracking and hydro­

dealkylation are strongly exothermic.

39

2.5. Conversion, Selectivity and Yield of the disproportionation of

toluene

The conversion of toluene, the selectivity of the disproportionation

reaction and the yield of required products are important parameters

for judging the performance of disproportionation catalysts and in

studying the kinetics of the reaction. Furthermore, they determine the

amount of heat evolved or absorbed during the reaction.

2.6.1 Feed consisting of toluene and hydrogen

As was demonstrated at the end of section 2.2, the reaction prod­

ucts obtained from a feed consisting of toluene and hydrogen in the

experiments described in chapters 3 and 5 can be completely accounted

for by solely considering disproportionation, hydrodealkylation and

transalkylation reactions, viz. the disproportionation of toluene, the

hydrodealkylation of toluene and the transalkylati'on of toluene and

xylenes (reactions 1, 2 and 3 in Table 2-1) as well as the dispropor­

tionation of xylenes and the hydrodealkylation of the xylenes and

trimethylbenzenes formed in the afore-mentioned reactions. Thus, under

the conditions of the experiments, the following reactions are probable

-*-1

2

3

4

5

2X t T+TMB 6

where T,B,X,TMB and M designate toluene, benzene, xylenes, trimethyl­

benzenes and methane. When considering the thermodynamics of the dis­

proportionation of toluene, reactionsI,2 and 3 are sufficient to ac­

count for the six species present in the products of the experiments

described in chapters 3 and 5, since reactions 4,5 and 6 can be derived

2T ^ B+X

T+H^ ^ B+M

T+X J

X+H^

TMB+H^

B+TMB

^ T+M

^ X+M

40

from 1,2 and 3. When dealing with rate problems, however, such as the

calculation of the selectivity for one of the above reactions, all six

reactions need to be considered simultaneously, at least in principle,

since, because of their individual rates, all six reactions may occur

at the same time. Nevertheless, under the conditions of the experiments

described in chapters 3 and 5 and where only toluene and hydrogen were

present in the feed, the yield of xylenes and trimethylbenzenes were

small in comparison with the concentration of toluene. Therefore,

reactions 4,5 and 6 were neglected.

If Y stands for the mole fraction of a particular component in the

reaction product, it can be derived (see Appendix 2) that the conver­

sion of toluene is given by:

(Y,+Y +Y^ , )100 ^ b X tmb- „

" (YK+Y^+Y +Y, , ) "'' b t X tmb

the disproportionation selectivity by:

2(Y +Y^ u)100 _ X tmb^ o ^ ' (Y.+Y +Y^ , °

b X tmb

and the yield of disproportionation products by:

2(Y +Y^ u)100 _ X tmb^ ^ D (Y,+Y^+Y +Y^ , ) °

b t X tmb-

2.5.2 Feed consisting of benzene, toluene, xylenes and hydrogen

As was previously stated, under certain conditions some or all of

the reactions which were neglected in the above derivations of conver­

sion, selectivity and yield must be considered. One such condition is

when benzene and/or xylenes are present along with toluene and hydrogen

in the feed (see chapter 5). Analysis of the reaction products of the

experiments described in chapter 5 with feeds consisting of hydrogen

and mixtures of the aromatics showed that only a negligible amount of

methane was formed and that this amount could be completely accounted

for by solely considering the hydrodealkylation of toluene. The reactions

41

which must be considered are, therefore, reactions 1,2,3 and 6 above.

It is , however, impossible to derive the selectivity for the dispro­

portionation reaction and the yield of the disproportionation products

in terms of the mole fractions of the components present in the products

by considering these four reactions simultaneously, because it is im­

possible to distinguish by analytical techniques between trimethylben­

zenes formed by reactions 3 and 6. Data on the rates of these reactions

over the catalyst used in the experiments of chapter 5 are needed in

order to decide whether reaction 3 or 6 is the most probable under the

conditions of the experiments described in that chapter.

Since no such rate data are available, an experiment was performed

using the apparatus described in chapter 5. The conditions and results

of the experiment are summarized in Appendix 9. The choice of flow rates

in the two experiments shown in Table 5-11 ensured that the two aromat­

ics mixtures had the same number of methyl groups per phenyl group and,

therefore, the same equilibrium compositions. The choice of reaction

pressure gave the same xylene partial pressure (0.17 atm) in both feeds.

The fact that roughly as much trimethylbenzenes were formed with mix­

tures 1 and 2 indicates that under the conditions of the experiments

reaction 6 is favoured over 3. On the strength of these results it was

concluded that reaction 6 was more likely than reaction 3. Using reac­

tions 1,2 and 6 and similar notations as in section 2.5.1 and using V

to represent the average number of methyl groups per phenyl group in

the feed and K to represent the mole ratio of xylenes and toluene in

the feed, one can derive (see Appendix 2) that when the feed contains

benzene, toluene, xylenes and hydrogen, the conversion of toluene is

given by:

((Y.+Y^+Y +Y^ ,)V-Y^(1+2K))100 ^ b t X tmb' t^ ^ 5,

^ " (Y,+Y^+Y +Y^ , )V '° ' ^ b t X tmb-

the disproportionation selectivity by:

((2Y +4Y^ , )(1 + 2K)-2VK(Y,+Y^+Y +Y^ ,))100 q _ X tmb-^ ' *• b t X tmb- - ,.

(Yv+Y^+Y +Y^ u)V-Y^ (1 + 2K) ^ b t X tmb" X.

42

and the yield of disproportionation products by:

° " f b ^ V mb ^

when the feed contains only hydrogen, toluene and benzene, the corres­

ponding quantities are:

((Y,+Y^+Y +Y^ , )V-Y^)100 ^ b t X tmb^ t- „

C = (Y,+Y^+Y +Y^ , )V b t X tmb'

(2Y +4Y^ u)100 ^ ^ X tmb

((Y,+Y,+Y +Y, v,)V-Y ) b t X tmb t-

(2Y^MY^^,)100 ^

D - ((Yb^Y^.Y^.Y^^^)V) '»

When the feed contains only hydrogen, toluene and xylenes , the corres­

ponding quantities are:

(CY,^Y^.Y^.Y^^^)(2-V)-Y^)100 ^

(Y,+Y^+Y Y^ , )(2-V) b t X tmb" -

(2Y +4Y^ ,-2(Y,+Y,+Y +Y, ,)(V-1))100 X tmb ^ b t X tmb '- •"' „

S = (^b^\^V\mbn2-V)-Y,)

Y (^V^^mb-^V^-V\mb^^^-^»^°V D- (Y^.Y^.^.Y^^j^)(2-V)

The expressions for the conversion, selectivity and yield for the

reaction of a feed consisting of toluene and hydrogen only is obtained

by substituting K = 0 and V = 1 in the appropriate relationships

above: (Y.+Y +Y^ ,)100 b X tmb J,

" (Y,+Y^+Y +Y^ , ) "" b t X tmb^

(2Y +4Y^ u)100 ^ X tmb (Y,+Y +Y^ , ) b X tmb

(2Y +4Y^ u)100 ^ X tmb

43

2.6 Conclusions

The stability diagram, Figure 2-1, shows that disproportionation is

equilibrium limited;hydrodealkylation, hydrocra:king and thermal cacking

are not. Ringhydrogenation is also equilibrium limited. At high hydro­

gen/toluene ratios and low temperatures this reaction is likely to

become a competing reaction. A selective catalyst is therefore needed

in order to accelerate disproportionation at the expense of the other

reactions.

Even in the presence of such a highly selective catalyst, isomeri­

sations and transalkylations are likely to occur along with the dis­

proportionation of toluene. If only the disproportionation reaction

and the isomerisation of xylenes are considered, it is seen that the

equilibrium conversion is not strongly dependent on temperature and

has an average value of 52.5%. When the formation of the twelve methyl­

benzenes by transalkylation reactions coupled with isomerisations is

considered, the equilibrium conversion increases to an average value

of 55% but remains still weakly dependent on temperature.

Disproportionation, isomerisations and transalkylations have small

heats of reaction. On the other hand, hydrodealkylation, cracking and

hydrogenation are highly exothermic.

Finally, relationships for toluene conversion, selectivity for the

disproportionation reaction and the yield of disproportionation pro­

ducts have been derived in terms of the mole fractions of the components

present in theproducts by setting up chemical reactions which account

for the components present in these reaction products and then making

a mass balance. These relationships will be used in the ahalysis and

interpretation of the experimental results of chaptffs 3 and 5. The fact

that the expressions derived for the conversion, selectivity and yield

using reactions 1,2 and 3 are different from those obtained with 1,2

and 6 demonstrates the importance of the selectivity of the catalyst.

The set of reactions which take place determines the mathematical form

of the relationships derived for the conversion, selectivity and yield.

However, for the two sets of reactions considered above, the differen­

ces in the relationships derived (see Appendix 2) are significant only

44

if trimethylbenzenes are formed in appreciable quantities.

In the experiments described in chapters 3 and 5, the relationships

derived above using reactions 1,2 and 3 are applied to analyse the

results in which no benzene or xylenes are present in the feed, the

assumption being that the high concentration of toluene would favour

reaction 3. The results of the experiments in which either benzene or

xylenes are present in the feed (chapter 5) are analysed with the rela­

tionships derived above using reactions 1,2 and 6. It is assumed that,

under the conditions of the latter experiments, reaction 6 is more

probable than reaction 3.

45

C H A P T E R 3

CATALYST PREPARATION AND TEST OF CATALYTIC PERFORMANCE

3.1 Introduction

Among the catalysts which have been claimed for the disproportion­

ation of toluene are halides (70-76,78) such as aluminium chloride and

borontrifluoride-hydrogen fluoride, oxides (80,82,83,85,86), like

silica-alumina, silica-magnesia and silica-boria, and zeolites (87-94,

100, 118,121), such as faujasite and mordenite. While the halide

catalysts are normally used for the liquid phase reaction, the vapour

phase reaction is carried out over the metal oxide or zeolite catalysts.

The liquid phase reaction catalysed by halides has very low selec­

tivity towards toluene disproportionation and is, therefore, unsatis­

factory for this process. Various authors ascribe this low selectivity

to the following factors: the higher stability and basicity of

m-xylene relative to o-xylene and p-xylene, the faster rate of iso­

merisation of both o-xylene and p-xylene to m-xylene than that of

m-xylene to the other isomers, and to the fact that a-complexes between

m-xylene and highly acidic halide catalysts are more stable than those

of the other isomers (75,76,79). One of the effects of these factors

is the presence at equilibrium of more m-xylene in the catalyst phase

at low temperatures than corresponds to thermodynamic equilibrium be­

tween the three xylenes. At higher temperatures, enough energy is

available to make the o-complex less stable and to cause the m-xylene

to undergo not only isomerisation but also disproportionation and

transalkylation (with toluene). The two last-mentioned reactions de­

crease the selectivity of the disproportionation of toluene.

The results of several studies (88,89,90) and the claims in patents

(84, Table 3-1) show that zeolites make better disproportionation cata­

lysts than silica-aluminas. Among the zeolites mordenite possesses the

highest activity for this reaction (88,90). However, its activity de­

teriorates rapidly as the reaction progresses. Faujasite, which has a

good activity, also suffers from too rapid a loss of activity (91).

No. Catalyst

1 Al^Oj/P^

2 SiO^/Al^Oj

3 S1O2/AI2O3

4 H-mordenite

5 H-mordenite/Al 0

6 H-faujasite/Al^Oj

7 H-mordenite/NiS

8 H-mordenite/e-AlF /Cu

9 H-mordenite

10 SiO^/Al^O^

11 SiO /NiSe

12 AlCl^

13 HF-BFj

14 H-mordenite

15 Al^Oj/B-AlFj

Temp.

°C

425

460

538

300

420

440

302

450

400

400

340

100

23

410

SCO

Pressure

kg/cm

1

28

27

35.2

34

34

35

30

1

1

78

33

33

30

30

mol H^/

mol toluene

2.5

1

1

10

10

10

5

20

5

5

2.5

-

-

8

20

LHSV

mol/mol/hr

0.2

2.3

1.5

0.83

5

2

2.3

1.8

0.27

0.27

4.0

-

-

-

-

Conversion

%

39

15

30

17

33

27

33

41

40

0.7

35

-

0

44

44

Selectivity 0 0

100

100

90

100

92

85

100

-

95

100

94

-

0

98

95

Reference

(185)

(186)

(187)

(188)

(189)

(190)

(191)

(192)

( 90)

( 90)

(193)

(194)

(195)

(196)

(197)

16 H-faujasite/P^

17 H-mordenite/Ag

18 Si02/Al203/Cr202

19 ^^0^/kl^Q^/Cr^Q^

20 Y-zeolite/Cr20

21 B^O^/Al^Oj/Pt

22 B^Oj/Al^Oj/Pd

23 B^O^/Al^O^/Ni

24 B^Oj/Al^O^/SnO/Pd

25 B^Oj/Al^Oj/SnO/Ni

26 Y-zeolite/Cr'^"

27 Mordenite/MnO /

SnO/Cr ""

28 Mordenite/MnO / 4 +

SnO/V^

29 NiY/SiO^/Al^Oj/

CoO/MoO

30 H-mordenite/CoS

Table :

480

400

566

538

538

538

538

538

538

538

538

483

483

480

280

33

67

28

56

56

28

28

28

56

56

56

35

35

21

53

5.1. Catalysts used

10

23

2

2

1.2

3.0

3.0

3.0

3.0

3.0

1.2

2.0

2.0

3.8

4

for the dis

8

-

1

1

1

1

1

1

0.5

0.5

1

1

1

1

1.7

30

21

14

20

29

36

27

21

42

48

29

57

56

43

50

100

100

64

72

90

88

93

96

46

50

90

100

100

100

100

jroportionation of toluene.

(190)

(198)

(203)

(203)

(203)

(200)

(200)

(200)

(200)

(200)

(201)

(201)

(201)

(199)

(202)

48

Studies (92,93,94,100) have shown that combining mordenite with alu­

minium fluoride and copper results in an improvement of its selectivity

and stability for the reaction. It may therefore be expected that

aluminium fluoride and copper would have a similar effect on the per­

formance of faujasite for the same reaction. The method of preparation

of a toluene disproportionation catalyst, based on a faujasite-type

zeolite, NaY, combined with g-aluminium fluoride and copper, is des­

cribed in the next section. Afterwards, the activity of the catalyst,

measured in terms of the conversion of toluene, and the selectivity

of the reaction for the disproportionation of toluene, are presented

as a function of the time the catalyst is in use and compared with the

same quantities determined for other catalysts. Data on the effects

of catalyst composition and activation procedure on the performance

of the catalyst are also presented. The influence of the reaction con­

ditions such as pressure, temperature and space velocity will be dis­

cussed in chapter 5 along with the kinetics of the reaction.

3.2 Experimental

Materials. The toluene, analytical grade, was used without further

purification. However, precautions were taken to prevent it from coming

into contact with the atmosphere in order to avoid contamination with

oxygen and water vapour. The hydrogen, of chemically pure quality, was

purchased commercially, dried over molecular sieve 3A and passed over

reduced copper oxide (BASF R3-11 BTS) catalyst to remove traces of

oxygen.

The sodium Y zeolite, SK-40, was purchased from Union Carbide Cor­

poration in the form of pellets with binder and as a powder without

binder. Other chemicals (copper nitrate, aluminium chloride, ammonium

chloride and ammonium fluoride) were reagent grade products purchased

commercially. The two silica-aluminas, which were studied along with

the zeolite-based catalysts, were obtained from AKZO Chemie (Ketjen,

Amsterdam). The first, LAL5P, is a low-alumina (15% alumina) sample

which had been subjected to steam treatment by the manufacturer. The

second, LAL3P, is similar to the first type, except that it was not

49

treated with steam. The hydrogen mordenite, Zeolon 100, was obtained

from Norton Company, U.S.A.

Preparation of catalysts. The silica-aluminas and the hydrogen mor­

denite were used without further treatment. The ammonium form of the

above Y zeolite was prepared by exchanging the Na ions in SK-40 for

NH ions by three successive immersions of the SK-40 (25 g) in a

2.23N NH CI solution (250 ml). During each exchange, the resulting

slurry was heated under reflux for 2 hours, with thorough stirring.

After the final exchange the zeolite was filtered off, washed free of

chloride and sodium ions with deionized water and then dried overnight

in an oven at 110 C. Analysis of the NH.Y zeolite by atomic absorption

spectrophotometry showed that the unit cell composition had changed to

Na. ,(NH ) (AlO )^,(SiO )j^^.nH 0. This corresponds to a level of

ammonium exchange of 91.8%. Figure 3-0 shows the sodium content and the

level of ammonium ion exchange as a function of the frequency of treat­

ment. A higher level of exchange was not attempted in order not to

damage the crystalline structure of the zeolite.

The 6-AlF was prepared by a double decomposition reaction between

solutions, in deionized water, of stoichiometric quantities of aluminium

chloride (120.7 AlCl .6H 0 in 250 ml) and ammonium fluoride

(55.6 NH.F in 150 ml). The resulting solution, clear presumably as a

result of the preponderance of the soluble a-form of aluminium fluoride,

was partially evaporated and allowed to stand for two days. During this

time a white precipitate was formed. This precipitate was filtered off,

washed with cMonized water and dried in an oven at 110 C. X-ray dif­

fraction analysis showed that it was ammonium aluminium fluoride,

NH AlF , formed according to the equations 4NH F + AlCl -*•

NH.AIF. + 3NH.C1. Chemical anylysis also showed that it was contamin-4 4 4 ^ •'

ated with excess ammonia and about 0.5 w% chloride. After calcination

at 500 C for 24 hours chemical analysis showed it to be free of

ammonium ions. X-ray diffraction analysis of the residue confirmed

that it was pure g-AlF , formed according to:

NH.AIF^ -> B-AIF, + NH. + HF 4 4 3 3

50

100

J 20

1 2 No of Exchanges

Figure 3-0 Sodium content and ammonium ion exchange of SK40

The average crystallite size of the 6-AlF , also determined by X-ray

diffraction, was 18.5 nm (185 A°).

Catalysts containing NH Y and 6-AlF were prepared by adding the

required quantity of NH AlF to a suspension in deionized water of the

ammonium Y zeolite. The mixture was heated and stirred in order to

mix the two components intimately. It was then evaporated almost to

dryness and further dried in an oven at 110 C.

Catalysts consisting of NH Y and copper were prepared in a similar

manner by mixing the zeolite with copper nitrate dissolved in de­

ionized water. Catalysts containing the three components were also

prepared in the same way. The dried catalysts were crushed and sieved

to obtain the particle size, 0.210-0.420 mm, used in the experiments.

Equipment. The catalytic performance measurements were carried out

51

© ®

Figure 3-1 Flow scheme of equipment for catalytic performance

measurements.

in a continuous flow apparatus designed to operate at 1 atm total

pressure. A flow scheme of the equipment is shown in Figure 3-1.

The gases used, hydrogen and helium, supplied from gas cylinders, were

led through reduced copper oxide (BASF R3-11 BTS) catalyst (T) to

remove traces of oxygen and other impurities, and dried over molecular

sieve 3A (7). Their flow rates and pressures were then measured by

means of rotameters fs) and pressure gauges.

The helium flowed directly to the analysis section, where it was

used as a carrier gas for the gas chromatograph. The hydrogen passed

through a saturator consisting of a glass apparatus which was packed

with cylindrical glass particles, partially filled with toluene and

immersed in thermostatically controlled water bath Q4_). The vapour

pressure over the toluene and therefore also the concentration of

toluene in the constant flow hydrogen streampassing through it is

solely determined by the temperature of the water bath. The saturator

52

was calibrated in order to verify that its efficiency was constant

under the conditions of the experiments. To this end, the saturator

effluent was channelled through a trap immersed in liquid nitrogen and

the amount of toluene collected in a known period of time was weighed.

The results showed that the efficiency of the saturator was constant

for the duration of the experiments.

The hydrogen-toluene reactant stream then flowed through a heated

copper tube to the stainless steel reactor (20 cm long, 0.8 cm internal

diameter), which was placed in the middle of cylindrical electrical

oven (¥). The catalyst was kept in place by a sintered steel plate on

either end of the reactor. The temperature of the reactor was measured

by means of chromel-alumel thermocouples placed at three different

positions along the axis of the catalyst bed and was registered on a

Philips twelve-point recorder. It was constant within +_ 2°C.

The reactor effluent flowed to sampling valve (6), where a sample

was taken for analysis, and then to a water-cooled condenser where the

condensable components were separated. The non-condensables, mainly

hydrogen and some methane, were vented through a soap-film meter, with

which the flow rate of the gas leaving the system was measured.

Analysis. The sample taken above was carried in the helium gas stream,

flowing at 66 ml/min., into gas chromatographic column KjJ• The column

was made of a copper tubing (3 m long, 4 mm I.D. and 6 mm O.D.) and'

packed with chromosorb W impregnated with bentone and diisodecyl-

phthalate. The components in the column effluent were detected in a

catharometer connected to wheatstone bridge(8^. The signal from the

catharometer was electronically integrated by means of analog

integrator(9) before being registered on Hitachi-Perkin Elmer model 159

recorder (lO) . The sampling valve and the column with the catharometer

were maintained at 90°C by means of a thermostatically controlled air

bath.

The mole fractions of the components were calculated from the elec­

tronically integrated peak areas by the method of internal normalisa­

tion in which toluene was used as the internal standard. With these

mole fractions the conversion and selectivity were calculated using the

relationships derived in chapter 2.(see also Appendix 5).

53

Proaedure. The reactor was filled with catalyst and the system was

then tested under 2 atm hydrogen pressure to ensure that it was leak-

proof. After the pressure was reset to 1 atm, the hydrogen flow rate

was adjusted to 60 ml/min. The catalyst was activated by heating the

reactor from room temperature to 230°C at the rate of 1°C per minute

and holding it at this temperature for 2 hours. The temperature was

then increased to 500°C at the rate of 2''C per minute and held at this

temperature until a total of 24 hours had elapsed from the beginning

of the activation period.

At the end of the catalyst activation, the system was brought to

the standard conditions used to compare the activities of the cata­

lysts. These standard conditions are shown in Table 3.2.

Reactor temperature

Reactor pressure

Reactor volume

Hydrogen flow-rate

Toluene flow-rate

Hydrogen/Toluene mole ratio

Saturator temperature

Toluene vapour pressure

Apparent residence time in the reactor

Particle size of catalyst

Liquid hourly space velocity

W/F

500 *_ 2°C

1 atm. absolute

10.4 cm^(L=20cm,I

0.590 ml/s (STP)

1.58 X 10'^ mol/s

16.7

33.9°C

0.056 atm

6 sees

0.210-0.420 mm

0.06 hr"^

859 g.cat.hr/mol.

.D.=0.8cm)

toluene

Table 3.2 Standard conditions for catalytic activity

measurements.

3.3 Results and Discussion

The toluene disproportionation activity and selectivity of the

HY zeolite obtained by activation of ammonium Y zeolite is shown in

Figure 3.2. The activity goes through a maximum at about 3 hours

54

n l p I I I I I 1 L_ 0 20 40 60 80 100 120 140

^ — Streom time ,hrs

Figure 3-2 Activity and Selectivity of HY zeolite

stream time and then decreases steadily, becoming almost constant at

around 12% after about 130 hours. The selectivity,on the other hand,

goes through a minimum at the same stream time where the conversion

reaches a maximum and rises as the conversion decreases, eventually

levelling off at about 85%. These results suggest that more and more

of the active sites are eliminated as the reaction proceeds (progres­

sively lower conversion) and that the remaining sites are the ones that

are active for toluene disproportionation (progressively higher se­

lectivity) . Apparently the most active sites are eliminated first;

these sites are responsible for the most important side reactions,

that is, the hydrodealkylation and cracking of toluene (see chapter 2).

The elimination of the active sites and the resulting progressive de­

crease in activity can most probably be attributed to the coverage of

these sites by the coke formed in the cracking reactions.

The activity and selectivity of a composite catalyst, designated

AB1(72%HY + 18%e-AlF + 10%Cu), are compared with those of HY zeolite,

H-mordenite and the silica-aluminas 5P and 3P in Figures 3-3 and 3-4.

The results show that the activity of H-mordenite is high initially,

55

4 8 12 — stream time , hrs

16 20 24 28..

Figure 3-3 Comparison of the activity of various catalysts.

5 10

— Stream time , hrs

Figure 3-4 Comparison of the selectivity of various catalysts.

56

but decreases rapidly in a very short time. The selectivity of this

catalyst for toluene disproportionation is very poor. The two silica-

aluminas have very low activity and selectivity for this reaction. The

steamed sample is a particularly poor catalyst, with very little ac­

tivity and zero selectivity. The composite catalyst ABl is less active

than HY zeolite but it is nevertheless the best toluene disproportion­

ation catalyst among the fj.ve because of its reasonable activity

coupled with a high selectivity and good stability.

Figures 3-5 and 3-6 show the activity and selectivity of catalysts

of widely different aluminium fluoride contents. The results indicate

that, with the copper content constant, a higher proportion of B-AIF

results m a higher selectivity, but at the expense of the activity.

Figure 3-5 also shows that, within experimental error, there is no

difference m performance between a catalyst prepared from SK-40 with

binder and one prepared from bmderless SK-40.

Effect of &-AlF^. The effect of B-AIF was further studied by prepar-o 3

ing and testing catalysts containing HY zeolite and varying percentages

of g-AlF . Figure 3-7 shows the results for a catalyst containing 17%

B-AIF . Comparison of the data in Figure 3-7 with those for HY zeolite,

Figure 3-2, shows that the selectivity of the catalyst with g-AlF

IS higher especially at the beginning than that of HY zeolite and in­

creases rapidly to a final value of 88%. However, the initial activity

of this catalyst is rather lower than that of HY zeolite, its activity

decreasing less rapidly with time than that of HY zeolite.

After only 50 hours the selectivity over this catalyst has already

attained a steady value of 88% at a constant conversion of 6%.

Table 3-3 summarizes the results of the effect of aluminium fluoride

on the activity and selectivity of HY zeolite. From this table it is

clear that the catalyst with the higher 6-AlF content (17%) has the

highest selectivity and the most stable activity and selectivity,

although Its activity is low when compared with the other catalysts.

Effect of copper. Figure 3-8 contains data on the activity and selec­

tivity of a catalyst containing 12% copper on HY zeolite. The results

for other catalysts containing varying amounts of copper are summarized

57

20 40 60 — Stream time .hrs

140

Figure 3-5 Activity and Selectivity of catalyst (55% HY/34% B-ALF^/

11% Cu) with a high B-ALF, content

X Catalyst prepared from SK40 with binder

o Catalyst prepared from SK40 without binder

80,

60 k 40

20

Selectivity

Conversion

o

20 40 60 — Stream time , hrs

80 100 120 140

Figure 3-6 Activity and Selectivity of catalyst (84% HY/5% B-ALF^/

/11% CuJ with a low B-ALF, content.

58

Catalyst

A

B

C

D

% B-AIF^

0

5

9

17

At 2 hours

S%

84

74

54

20

S%

2

4

46

84

At 50 hours

5%

27

28

20

6

S%

70

77

84

88

At 100 hours

C%

13

16

13

6

S%

84

89

86

88

At 140 hours

C%

12

14

11

6

S%

85

89

86

88

Table 3-3 Effect of 6-AlF on the activity and selectivity of

HY zeolite

in Table 3-4. These results show that the catalysts containing 8%

and 12% copper have a reasonable activity and selectivity even at 50

hours stream time. Catalysts containing different copper percentages

Catalyst

1

2

3

4

5

6

7

% Cu

0

2

3

6

8

12

25

At 0

e%

55

56

53

68

73

40

50

lours

S%

41

36

52

0

5

46

15

At 2 hours

5%

84

83

82

85

82

66

82

S%

2

0

0

0

4

6

0

At 25

5%

46

10

17

12

33

27

11

hours

S%

48

56

68

78

80

70

55

At 50

5%

27

2

6

6

13

17

3

hours

S%

70

0

62

85

71

81

30

Table 3-4 Effect of copper on the activity and selectivity of

HY zeolite

59

60 r 6 0

4 0

20 Ax

o o - 0 0 - Selectlvlty

%

SB-

—^"^ J L

-XMt^. Conversion

20 40 60 Stream time , hrs

80 100 120 140

Figure 3-7 Activity and Selectivity of a catalyst containing 17%

6-ALF, on HY zeolite

80

20 40 60 — Stream time ,hrs

80 100 120 140

Figure 3-8 Activity and Selectivity of a catalyst containing

12% Cu on HY zeolite.

60

than these are much less selective and stable, the main reaction of

toluene over them being hydrodealkylation rather than disproportion­

ation. The data of Table 3-4 indicate that the improvement by copper

alone of the toluene disproportionation performance of HY zeolite is

slight. However, comparison of this table with Table 3-3 and both

Tables 3-3 and 3-4 with Figures 3-3 and 3-4 reveals the obvious im­

provement by B-AIF and copper in combination of the performance of

HY zeolite as a toluene disproportionation catalyst.

The study of the influence of 3-AlF and copper as reported above

is a preliminary attempt to help in selecting a suitable catalyst for

the reaction. Of course, if the aim is to investigate the effect of

the two components exhaustively, then tests of the performance of such

catalysts under other conditions of temperature and pressure need to

be carried out. On the strength of the performance of HY with 17%

0-AlF (Figure 3-7) and HY with 8% and 12% copper (Figure 3-8,

Table 3-4), a composite catalyst, designated ABl was prepared with the

following composition:72%HY, 18% S-AIF and 10% Cu. The composition of

this catalyst, which is used for the kinetic studies described in

chapter 5, is in agreement with those of other toluene disproportion­

ation catalysts reported in the literature (92,93,100).

Effect of activation temperature on the performance of catalyst ABl.

The catalyst was activated according to the procedure previously de­

scribed but at four different temperatures (400°, 450°, 500°, 540°C).

After activation, each catalyst was used for the disproportionation of

toluene. The reaction conditions were those shown in Table 3-2, except

for the reaction temperature which was 450°C. The results are shown in

Figure 3-9. The catalyst activated at 540°C had very little activity.

The catalyst activated at 400°C had the highest initial activity but

also the highest rate of deactivation. The catalyst activated at 500°C

had a lower initial activity than the one activated at 450°C but also

a lower deactivation rate. The influence of activation temperature

appears to be that, as it increases up to 500°C, initial activity

decreases, rate of deactivation decreases and catalyst stability

61

100

80

60

J 40 • c _o m > ° 20

O

400 450 500 550 ^"—Activation temperature, "C

Figure 3-9 Effect of activation temperature on Conversion and

Selectivity over catalyst ABl.

A Performance at t=0 hrs.;• Performance at t=20 hrs.;

X Performance at t=28 hrs.

increases. The final activity level increases up to 500°C and then

decreases. A similar trend is observed for the selectivity which in­

creases with the activation temperature but above 500°C decreases

again. The best activation temperature appears therefore to be 500°C.

Above this temperature the performance of the catalyst clearly dete­

riorates. This may be a result of the destruction of the crystalline

structure of the catalyst by such a high temperature or by a change in

the nature of the active sites such as by transfoinnation of the sites

from Br0nsted acid types, which are considered responsible for this

type of reaction (88,91) into Lewis acid types which are believed to

be inactive. The discovery that rehydrating the catalyst after activa-

62

- 600

- 400

200

0 0

Figure 3-10 Weight change of catalyst ABl during activation.

tion at 550°C restored both its toluene disproportionation activity

and ammonia adsorption capacity (see chapter 4) and that the specific

surface area of the catalyst did not change appreciably when activated

above 500°C appear to support the hypothesis that the loss of catalytic

activity above 500°C is a result of the elimination of Br^nsted

sites by dehydration rather than the destruction of the structure of

the zeolite.

The weight change of catalyst ABl during activation was studied in

the same thermobalance used for the ammonia adsorption measurements

described m chapter 4. The result is shown in figure 3-10. The total

weight loss is approximately 32% and is a result of the evolution of

adsorbed water, the deammoniation of NH Y, the decomposition of

NH.AIF , the decomposition of copper nitrate, the reduction of copper

oxide to metallic copper and the loss of water of constitution of the

zeolite.

3.4 Conclusions.

The activity, selectivity and stability of a number of catalysts

63

for the disproportionation of toluene have been compared at a set of

standard conditions. The results show that a catalyst consisting of

72 w% HY zeolite, 18 w% B-AIF and 10 w% Cu has the most satisfactory

overall activity, selectivity and stability of the catalysts tested.

While this catalyst is less active than HY zeolite, its selectivity

and stability are much higher. Although neither B-AIF nor copper is

in itself active for the disproportionation of toluene, the combination

of these materials with HY zeolite results in a catalyst with an im­

proved performance. It has also been demonstrated that B-AIF can be

prepared in a very simple manner. More laborious methods have often

been used to prepare this compound for incorporation in toluene

disproportionation catalysts (93,108,109). It was also shown that the

best activation temperature for the catalyst is 500°C. When activated

above this temperature its activity all but disappears. Activation

below 500°C, on the other hand, leads to a faster deactivation of the

catalyst with use.

64

C H A P T E R 4

CHARACTERISATION OF THE PHYSICO-CHEMICAL PROPERTIES OF THE CATALYSTS.

4.1 Introduction

The performance of a catalyst is stronglyinfluenced by its physico-

chemical properties. For the disproportionation of toluene, texture and

acidity are the most important determinants of catalytic behaviour.

Texture influences the transport of reactants and products through the

catalyst, while acidity is responsible for activity and selectivity .

The texture of zeolites and zeolite-based catalysts is characteri­

zed by their highly porous structure combined with their very regular

crystalline framework. The adsorption selectivity shown by these mole­

cular sieves depends on the differences in size and shape between the

molecules and the apertures in the zeolite crystals. The catalytic

properties, on the other hand, depend on the number and strength of

acidic surface sites. Thus, characterization of texture and acidity is

essential for a complete understanding of the properties of zeolite-

based catalysts.

Extensive crystallographic and structural information on zeolites,

mainly from X-ray crystal structure analysis, is available. However,

when zeolites are, for example, impregnated or combined with other

chemical components, as is often the case with industrial catalysts,

data on the zeolite carrier alone do not suffice to characterize the

composite catalyst and it becomes necessary to determine the properties

of such catalysts separately.

In the following sections, the texture and acidity of the catalysts

used in the experiments described in chapters 3 and 5 are determined.

The validity of the methods used when applied to zeolites and the

significance of the results are also discussed.

65

4.2. Texture of catalysts

4.2.1 Introduction

The most useful parameters for the characterization of the texture

of porous solids, namely the specific surface area, the total pore vol­

ume , and the pore-size distributrion, are often determined by adsorp­

tion measurements such as the adsorption of nitrogen at low temperatures.

With zeolites, thisposes the problem of selecting the correct adsorption

isotherm to be used for interpreting the data. A cursory inspection of

the low-temperature nitrogen adsorption isotherms of microporous adsor­

bents such as zeolites gives the impression that they are of the Lang-

muir type. However, the adsorption isotherm cannot be a true Langmuir

type because zoelites contain not only micrpores but also meso- and

macropores , although they are usually classified simply as microporous

solids, that is solids to which a maximum pore radius of 2nm(20A) is

arbitrarily assigned. The meso- and macropores contribute to the total

adsorption by zeolites. Firstly there is, at low relative pressures,

the pore-filling adsorption in micropores, which occurs with a high

heat of adsorption. Then subsequently, at higher partial pressures,

adsorption in the meso- and macropores and on the external surfaces of

the zeolite crystals sets in, which is accompanied by a lower heat of

adsorption.

Similarly, the BET equation, which has often been used to estimate

the specific surface of porous and non-porous solids does not apply here :

since the isotherms of zeolites are obviously not of the BET type it

would be quite erroneous to apply the BET theory to such microporous

solids. Moreover, determination of the pore-size distribution by the

classical method, which utilizes Kelvin's theory of capillary conden­

sation is of doubtful utility in the case of zeolite-based catalysts for

two reasons. Firstly, their nitrogen adsorption isotherms show no hys­

teresis loop in the range of relative pressures at which the pores are

filled, which is characteristic of adsorption in pores with diameters of

less than 2 nm(20A). Secondly, the validity of Kelvin's theory becomes

66

questionable below approximately 2nm pore radius. Even the corrected

Kelvin equation derived by Broekhoff and de Boer (143,156) breaks down

at radii less than 2nm. The theory of capillary condensation in such

narrow pores is, therefore, still undeveloped. This is caused mainly by

the lack of knowledge of the density of the adsorbed phase and the ab­

sence of concrete data on the surface tension of strongly curved nitro­

gen/solid interfaces.

The use of mercury penetration porosimetry to determine the pore

size distribution of solids containing micropores is difficult because

very high pressures are needed to force mercury into the narrow pores.

Furtliermore, application of the Washburn equation, r(in A )=7500/p(in atm)

to analyse mercury penetration data is of questionable validity even in

transitional pores, since the values of surface tension and contact an­

gle for mercury in strongly curved pores are not accurately known.

Since there is as yet no universally accepted method of determining

the relevant parameters, several methods are usually applied simulta­

neously to analyse the texture of zeolites (44). For example. X-ray data

have been used to estimate the specific surface of CaA and NaX zeolites

(147), while the same quantity has been evaluated for H-Y zeolite by

low temperature nitrogen adsorption (23,89). However, the correctness of

the various methods which have been advanced for estimating the texture

of microporous solids remains a matter of active discussion (147,148,

149).

In the following sections, the texture of a number of catalysts, most

of them predominantly microporous in structure, is determined by some

well-known methods. The results obtained by the various methods serve

not only to characterize the catalysts but also to compare these methods,

in the hope that such a comparison may shed some light on their relative

utility. Also, an alternative method is suggested for the determination

of the specific surface area of zeolites. Finally, an attempt is made to

relate the texture of the catalysts to their activities for the dispro­

portionation of toluene.

67

4.2.2. Experimental

Catalysts investigated

The catalysts investigated can be summarized as follows:

1. Linde SK-40 molecular sieve Y zeolite with binder, which was used as

purchased from the manufacturer.

2. The hydrogen form of the above Y zeolite, obtained by in situ eva­

cuation at 350 C of ammonium Y zedite, which was prepared as described

in chapter 3.

3. A catalyst prepared by impregnating NH -Y with 18% 6-ALF, and 10%

copper, the preparation of which has been described in chapter 3.

4. The catalyst described in 3, above, activated for 24 hours at 500°C

under a constant flow of hydrogen (60ml/min).

5. The catalyst described in 4, above, after it had been used for 2

hours for the vapour-phase disproportionation of toluene at 500 C.

6. The same catalyst as described in 4., above, which had been used for

Ij months for the disproportionation of toluene at 500°C.

7. A steam-treated low-alumina silica-alumina catalyst (LAL5P).

8. An untreated low-alunina silica-alunina catalyst (LAL3P).

Procedure

Nitrogen adsorption measurements were carried out in a micro-BET

apparatus at -196 C according to the method described by Lippens and

co-workers (156,158). The system containing the catalyst sample was

evacuated for 16 hours at 350 C before measurement of each adsorption

isotherm.

A mercury penetration porosimeter, model 905-1, operating pressure

range 0-3500 atm., manufactured by Micromeritics Instruments Corp.,

was used for the mercury penetration measurements. The appropriate

corrections for mercury compressibility were applied (154). Further

details on the instrument and on the theory of its applications are

available in the literature (157).

68

4.2.3. Results

nitrogen adsorption measurements '

The nitrogen adsorption isotherms of the catalysts are shown in

Figure 4-1. The shapes of the isotherms of catalysts 1-7 show that they

are of type I according to the BET classification (44). This is charac­

teristic of solids which are predominantly microporous in structure.

The isotherms of these catalysts do not show hysteresis in the range of

relative pressures where their pores are filled. Catalyst 8, the un-

steamed silica-alumina sample (LAL3P), on the otherhand, shows a type

IV BET isotherm. This suggests that this catalyst contains few micro­

pores or none at all and that capillary condensation would be expected

Figure 4-1 Nitrogen adsorption isotherms of the catalysts at -

196 °C. Points are measured;lines calculated with the

Langmuir isotherm.

• Cat.No. 1 o Cat.No.2

» Cat.No.3 V Cat.No.4

D Cat.No.5 • Cat.No.6

X Cat.No.7 A Cat.No.8

69

to occur in its pores. The corrected Kelvin equation (143,156) can be

used for the calculation of the pore-size distribution of this catalyst.

The isotherms of catalysts 1-7 are Langmuir-type in shape in the

relative pressure range of 0.02-0.3. Between relative pressures 0.1 and

0.3, only a limited amount of adsorption is observed. This suggests

that the surface of these catalysts is mostly made up of micropores and

that only a small proportion of the total surface is in the meso- and

macropores. On the other hand, the substantial adsorption of catalyst 8

above a relative pressure of 0.1 points to a preponderance of pores of

the transitional type.

The five methods listed below were used to study the texture of the

catalysts. In all cases,surface areas were calculated by assuming a val­

ue of 0.162nm (16.2A'' ) for the cross-sectional area of an adsorbed

nitrogen molecule at -196 C, while pore volumes were determined by as­

suming that nitrogen was adsorbed as a liquid having a density of

0.8081 ml/g.

1) The t-method of De Boer (143,146,159), Figure 4-2, was used to esti­

mate the volume of the micropores, V , the area in transitional pores

S..., and the area in micropores, S . The "common t-curve " used in t' ^ ' m

this method is the curve determined by De Boer and co-workers (143,

158) for oxides, hydroxides and graphite.

2. The BET method was used to calculate the BET surface area, S_„_. bbl

However, as outlined above, the result has physical significance

for catalyst No.8 only, since multilayer adsorption, which is one of

the basic assumptions of the BET theory, can hardly occur in micro­

pores. The surface area calculated by this method is, however, use­

ful for comparison with values determined by other methods.

3. The nitrogen adsorption data were fitted to a Langmuir isotherm,

Figure 4-3, and used to calculate the specific surface, S,, of the

catalysts.

4. The methods of Dubinin and of Kaganer (44), Figure 4-4, were used to

evaluate the pore volumes, Vp,, and the specific surface areas,S re-U K

spectively of catalysts 1 through 7, which are microporous. 5. The Gurvitsch rule (44) was applied to calculate the pore volumes,

70

140-

120=-

100

0. I-z

CM

z E o >

80

60

40

20

2 0 3 0 - t A

4 O 5 0

F i g u r e 4-2 t - p l o t of t h e c a t a l y s t s .

X C a t . N o . 1 o C a t . N o . 2

• Ca t .No .3 V C a t . N o . 4

A Ca t .No .5 • C a t . N o . 6

D Ca t .No .7

V , of catalysts 1 through 7. For this purpose, the volume of

nitrogen adsorbed at saturation was estimated by extrapolating the

adsorption data to a relative pressure P/P =1 by means of the

Langmuir isotherm. The fit of the data to the Langmuir isotherm was

excellent, which justifies the use of the isotherm for extrapolation.

The straight lines of methods 1-4 above were determined by a least-

squares computer programme.

Table 4-1 summarized the texture parameters determined by the various

methods. The t-plots obtained for catalysts 1 through 7 correspond to

71

o 1 0 2 0 3

-8.0

- 7 0

- 6 0

5 O

P/Po

Figure 4-3 Langmuir plot for the catalysts

D Cat.No.l • Cat.No.2

• Cat.No.3 A Cat.No.4

o Cat.No.5 V Cat.No.6

X Cat.No.7

type2 in De Boer's classification (144,145,155,159). This is a further

indication that these catalysts contain micropores. The t-plot for cata­

lyst Sis of type 3. Such a plot results when capillary condensation

occurs in transitional pores of a certain shape and size at a certain

relative pressure. Owing to this capillary condensation, the adsorbent

takes up more adsorbate than corresponds to multilayer adsorption on a

non-curved, non-porous solid surface at the same relative pressure.

Hence, the amount of adsorbate increases steeply as the relative pres­

sure is increased, with the result that if the t-plot were extrapolated

72

1 8

1 6

1 5

1 3

o

» 0

• •

O Q o

1

i^~~i-

1

-—o _ ^

i •

-

• i •

1

O 2 0 4

(log Po/P)

0 6 0 8

- 2 2

2 1

1 9

1 8

Figure 4-4 Dubinin and Kaganer plots for the catalysts

• Cat.No.l V Cat.No.2

D Cat.No.3 A Cat.No.4

o Cat.No.5 X Cat.No.6

* Cat.No.7

Catalyst

No. Description

1 Na-Y

2 H-Y

3 H-Y/AIF^/ Cu-oxide

4 H-Y/AlFj/Cu

H-Y/AIF /Cu used 2hrs.

g H Y/AlFj/Cu used l^months

._ Silica-alumina (steamed)

g Silica-alumina (not steamed)

Method 1

1

24

25

65

98

77

70

71

:-method

1

2

S m

2, m /g

454

618

243

308

154

24

20

t-method

1

3

s BET 2,

m /g

342

422

237

301

173

87

85

400

BET-method

4

2, m /g

546

648

370

491

270

144

146

5

corrected m /g

1079

1280

731

970

533

284

288

1 1 ^angmuirLangmuir

1 1

6

2, m /g

544

641

354

480

262

134

134

Kaganer

7

S

2, m /g

855

980

447

557

255

48

This work

8

^D

ml/g

0.192

0.226

0.125

0.169

0.093

0.047

0.047

9

ml/g

0.178

0.204

0.093

0.116

0.053

0.010

0.010

1 1 Dubinin t-method

1 1

10

ml/g

0.193

0.229

0.129

0.171

0.094

0.048

0.049

Gurvitsch

Table 4-1:Texture of the various catalysts.

74

to lower t values, it would not pass through the origin, but would

intercept the t-axis at a positive value.

It has been assumed that the "common t-curve"of De Boer and co-workers

(143,158) applies to the solids analysed in this study. The appropriate

'l;ommon t-curve"is one obtained by measuring the adsorption isotherm on

a non-porous reference substance, with a surface as nearly as possible

identical with the porous solid under investigation, as regards adsorp-

tive properties. When this condition is satisfied, the two solids have

identical adsorption potentials and,therefore, heats of adsorption. At

the same relative pressure, the thickness of the adsorbate multilayer

then is the same on the two solids. The difficulty of finding a truly

comparable reference substance for the determination of the t-curve of

a particular adsorbent is well-known (44, 160).

The results of the t-method (Table 4-1) show that when sodium Y zeolite

(catalyst No.l) is converted into the H-form (catalyst No.2), the

transitional pore area, S , does not change. This area consists of the

external surface of the zeolite crystals, which normally does not exceed 2

10m /g (147), and the area of the binding material used in the manufac-2

ture of zeolite pellets, which is normally 20-25m /g.

Comparison of No.2 with No.3 shows that the micropore volume has

decreased from 0.204ml/g to 0.093 ml/g, while the transitional pore 2 2

area, has increased from 25m /g to 65m /g. This decrease in micropore

volume indicates that the aluminium fluoride and copper oxide resulting

from the decomposition of copper nitrate have partially blocked the 2

micropores. The increase in the transitional pore area of 40m /g suggests

that the presence of aluminium fluoride and copper oxide particles in­

troduces extra transitional pores.

When catalyst No. 3 is activated to form No.4, volatile components

are lost, notably by the decomposition of copper nitrate to copper oxide

and the reduction of the latter to metallic copper. The corresponding

weight loss causes an apparent increase in the micropore volume, V ,

and in the total pore volume, V . Comparison of S. found for catalysts VJ t

No.2 and No.4 indicates that the addition of copper and aluminium fluo-

de produces more surface area in the transitional pores.

75

Texture of zeolite-containing catalysts.

The specific surfaces of the microporous catalysts examined, as cal­

culated from the Langmuir isotherm and by Kaganer's method, are in 2

satisfactory agreement. Benson et al. (23) determined a value of 850m /g

for H-Y zeolite by applying the Langmuir isotherm to nitrogen adsorption

2

data. This value is significantly higher than the value of 648m /g ob­

tained in this study for the same catalyst, using the same method. The

discrepancy between the two results may well be significant since the

Y zeolite used here was ammonia-exchanged to a higher level (91.8%)

than the catalyst of Benson and co-workers (69%) and would, therefore, be expected to possess a higher surface area. Venuto et al. (89) ob-

2 tained a value of 676m /g for H-Y zeolite using the BET method, a value

which is higher than the BET surface obtained for a similar catalyst in

this work (Table 4-1). Since the deammoniation procedure employed here

(evacuation at 350 C for 16 hours) is considered adequate, the differ­

ence between the two sets of results may be due to the presence of a

binder in our zeolite sample or to poorer crystallization of our sample

than those of the other investigators. It is known that incomplete

crystallization of zeolite preparations sometimes occurs during manufac­

ture, resulting in inaccessibility of some pores and low values of

experimental surface areas and pore volumes.

All the values of specific surface referred to above were calculated

by assuming that the cross-sectional area of a nitrogen molecule at

-196°C is 0.162nm (16.2A°^). While this value may be valid for a flat

adsorbent surface, it is less justifiable for zeolites, owing to the

strong curvature of the surface of the zeolite cavity. From the molecu­

lar diameter of the nitrogen molecule (0.315nm at 20 C) and the radius

of the large cage of Y zeolite (0.625nm), it can be calculated geometri-2

cally that the area occupied by a spherical nitrogen molecule is 0.32nm

(32A ). This value has been used to correct the surface areas calcula­

ted from the Langmuir isotherm (Table 4-1). Nevertheless, this improve­

ment notwithstanding , the basic assumptions underlying the Langmuir

theory, that is, monolayer adsorption, absence of mutual interaction

between adsorbed molecules, and constancy of the heat of adsorption as

76

a function of surface coverage, are not at all fulfilled. Therefore,

application of the Langmuir equation remains unsatisfactory.

Since the structure of zeolites is regular and well-known, it is

possible to calculate their texture analytically. By assuming the large

cages of Y zeolite to be perfectly spherical and by making use of the

fact that this zeolite has a cage radius of 0.625nm and a pore volume

of 0.296ml/g (181), its surface area, based on the area of the inner 2 -20 2

walls of its cages, is calculated to be 1419m /g or 491x10 m /cage. 2

A similar value, 1400m /g, was calculated by Dubinin for NaX zeolite

(147). The large discrepancy between such calculated results and expe­

rimentally determined areas (Table 4-1) was also noted by other inves­

tigators (147).

In the next section, a method which elimates this discrepancy is pre­

sented.

The pore volumes of the microporous catalysts as calculated by the

method of Dubinin, Vp, and by the Gurvitsch rule,V , are in good agree-

ment (Table 4-1). These values and similar results by others (19,20)

are slightly higher than the estimates obtained in this study by the

t-method. This is easily explained by the fact that the t-method esti­

mates the volume of the micropores only, whereas the other methods

determine the total pore volume. The t-method enables a distinction to

be made between the volume of nitrogen taken up in the cages of the

zeolite from the volume adsorbed on the external surface of the zeolite

crystals, on the binder and on the other materials brought on the zeolite.

An alternative method of determination of the specific surface area of

zeolites.

As has been pointed out above, the t-method provides, separately,

the amount of material adsorbed in the micropores and the amount in

transitional and larger pores. The slope of the t-plot gives an estimate

of the surface area, S , of the transitional and larger pores (Table 4-1).

If the cross-sectional area of an adsorbed nitrogen molecule and the

configuration of the adsorbed molecules in the micropores were known

with sufficient precision, it would be possible to estimate the surface

77

area of the micropores. Since neither is accurately known, the following

relationship, which combines calculated and experimental parameters, is

proposed for the determination of the surface area of zeolite catalysts:

S = V^.S^/V^ . 4-1

2 where S is the surface area in m /g, V is the volume of the micropores

3 determined by the t-method, cm /g, S is the calculated surface area of

2 '' one cage in m /cage, and V is the calculated volume of one cage in 3

cm /cage. S and V can be calculated from X-ray data (181). For NaY

zeolite,S , calculated by assuming a perfectly spherical cage (see

above) is 491x10 m /cage while V is 1023x10 cm /age.

This method can be used for other zeolites than NaY. The cross-sectional

area of the nitrogen molecule is not required. Furthermore, although

nitrogen is used as adsorbate, as is usual in texture studies, the re­

sult should in theory be the same irrespective of the adsorbate used,

as long as the value of V is known. Nevertheless, since results are in

practice somewhat dependent on the adsorbate (44), it is better to use

the standard adsorbate, nitrogen. There are zeolites, however, such as

molecular sieves 3A with pores that are too narrow to admit nitrogen

molecules. In such cases another adsorbate than nitrogen may be used,

for example hydrogen or krypton. All the same, the t-method on which the

present method depends is as yet developed onlv for nitrogen adsorption.

The surface areas of the micropores of the catalysts as calculated by

this method using V are shown in Table 4-1 under S. The value for NaY 2 2

zeolite, 855m /g, is much lower than the expected value of 1419m /g

calculated above. The reason must be that the value obtained for

V ,0.178ml/g, is lower than it should be. If it is accepted that the

t-method determines the correct volume of the micropores, then the

explanation for the low value of V must lie either in the poor crystal­

lization of the NaY sample or in the presence of a binder which has

blocked some of the micropores. It is also possible that both factors

come into play. The value of V calculated for binderless NaY is

0.334ml/g. With this value and using the method described above, a value 2

for the micropore surface area of S=1603m /g was calculated. This result

78

suggests that the low value obtained for NaY with binder can at least in

part be attributed to the blockage of the micropores by the binder.

This conclusion was further substantiated by employing the method

proposed above to recalculate the data of Benson and co-workers (23), 2

who used the Langmuir equation to obtain 850m /g as the surface area of 2

binderless HY zeolite. A value of 1451 m /g was found, which is consid­

ered to be consistent with the value expected for the surface area on

the strength of X-ray data.

Mercury penetration measurements.

The mercury porosimeter used in this study has a maximum operating

pressure of 3500 atm. Substitution Of this value in the Washburn equa­

tion sho'.vs that at this pressure pores narrower than approximately

2.2nm(22°A) will not be filled. Clearly, this method provides the pore

size distribution of the transitional pores and the macropores, but not

of the micropores.

The pore size distribution of the H-Y zeolite sample is shown in

Figure 4-5. Figure 4-6 shows the same type of distribution for catalyst

No.4. In both graphs, the cumulative volume and dV/d(logr), the deriva­

tive of the pore volume with respect to the logarithm of the pore radius,

^— log P

10 0 0 10 2 0 3 0 1 96

1 68

1 40

Z 1 12 o - 0 84 "O

? 0 56

0 28

O 00 -0 13

6 0 5 0 4 0 3 0 2 0 10 log r •

Figure 4-5 Pore size distribution in HY zeolite with binder

79

» — log P

10 0 0 10 2 0 3 0

o> o

x>

>

6 0 5 0 4 0 3 0 2 0 1 0 log r —^

Figure 4-6 Pore size distribution of HY zeolite impregnated with

g-ALF and Cu (catalyst A B l ) .

are plotted against the logarithm of the pore radius and the penetration

pressure. Figure 4-5 shows that the distribution curve of H-Y zeolite

consists of two distinct parts. The volume of mercury penetrating the

sample at pressures below about 10 atm presumably went into void spaces

among the individual particles, while the volume above 10 atm penetrated

the pores of the binding material of the catalyst. This would imply that

the volume of the pores of the binder is only slightly more than that

between the particles. From the graph, it can also be inferred that the 4

space between the particles is mainly of the order of 3x10 nm in radius,

while the average radius of the pores of the binding material is about 2

3x10 nm. The two types of void space thus differ greatly in size. Figure

4-6 shows that, besides the pores identified in Figure 4-5, catalyst

No.4 has pores predominantly of about 4nm (40 A) radius. These pores can

only be due to the aluminium fluoride component of the catalyst. The 2

average pore radius of the binder has shifted to around 1.5x10 nm.

The mercury penetration data were also used to estimate the specific

surface area of the penetrated pores by application of the Rootare 2 2

equation (161):12m /g was calculated for catalyst no.2 and 56m /g for No.4. Comparison of these values with the areas of the transitional

2-0 r z o

1-6-

1-2-

0 8 •

0 4-

0 0 .

.xwKxW*""'*'

1 -6

1 2

•0 8

0 4

0 0

80

pores, S , of the same catalysts estimated by the t-method and listed

in Table 4-I shows that the t-method gives consistently higher values.

This is due to the fact that S includes the areas of pores with radii

as small as 0.5nm(5A) while mercury penetrates pores with radii no smal

ler than 2.2nm at the maximum pressure at which the porosimeter was

operated.

Texture and toluene disproportionation activity.

The activities of catalysts No.5-No.8 for the disproportionation of

toluene are shown in Table 4-2. The conversion of toluene was used to

characterize the catalytic activity. Total toluene conversion and the

disproportionation selectivity were calculated with the expressions

derived in chapter 2.

Table 4-1 shows that the micropore volume and the micropore surface

area decrease drastically to about half their original values, that is, 2

from 0.116ml/g to 0.053ml/g for the pore volume, and from 557m /g to 2

255m /g for the micropore surface area, when catalyst No.4 is used for

2 hoin-s for the disproportionation of toluene (sample No.5). At the same

Catalyst

No. Description

5 HY/A1F,/Cu used 2hrs.

6 HY/AIF /Cu used if months

7 Silica-alumina (steamed)

8 Silica-alumina (not steamed)

Toluene disproportionation performance

Total conversion, %

42

25

2

15

Selectivity, %

85

92

0

70

Reaction conditions:P=10 ata, T=500°C, H2/toluene=16.7mol/mol,

W/F=176 g.hr/mol

Table 4-2:Toluene disproportionation performance of the catalysts.

81

time, the transitional pore area undergoes a comparatively less pronoun-2 2

ced decrease from 98m /g to 77m /g. The same trend is observed in the

catalyst which was used for 1 1/2 months for the same reaction (No.6).

Although it is still catalytically active (Table 4-2), its micropore

volume (Table 4-1) has all but disappeared (0.Olml/g);the decrease in

2 2

the transitional pore area, form 98m /g to 70m /g, is small in compari­

son. These results suggest that the seat of the catalytic activity of

catalysts No.4 to No.6 for the disproportionation of toluene is in their

transitional pores and that the contribution of the micropores to cata­

lytic activity is marginal if not minimal. The micropores appear only to

collect heavy reaction products which would otherwise lead to total

deactivation of the catalyst. These heavy products gradually fill up or

close off these micropores.

It can also be seen from the results, as would be expected, that the

texture of the catalysts is not the sole determinant of their activity

for the disproportionation of toluene. For example, catalyst No.4 has

a smaller specific surface area than No.8, but is more active and se­

lective for the reaction. The texture of No. 7 is comparable to that of

No.6. However, No.7 has little or no activity for the reaction, whereas

No.6 is reasonably active and selective. As was previously pointed out,

the amount, strength, type, and distribution, of acid sites present

on the surface of the catalysts, and for catalysts No.5 and No. 6 the

free-copper and aluminium fluoride surface area as well, are the co-

determining variables of their performance for this reaction. Therefore,

the acidity of the catalysts is considered next.

4.3. Acidity of catalysts

4.3.1 Introduction

The total acidity of a solid is measured by determining the quantity

of base which it can adsorb chemically. Acid strength, on the other

hand, is expressed either by the fraction of the base retained at a

particular temperature or by the Hammett acidity function, Ho(l). The

type of acid sites is usually described in terms of the Bronsted and

Lewis definitions of an acid (2).

82

Several methods have been used to measure the total acidity (2-30),

the acid strength (2,7,16,24,31-35) and the Bronsted and Lewis acidity

(2,15,36,37) of solids. Two methods which are widely used are the

Benesi titration method (9,10) and the base adsorption method (2,7),

both of which can be used to measure acid amount as well as acid

strength. Since the titration method depends on the detection of changes

in the colour of adsorbed indicators, it is best suited for whitish or

light coloured solids. The base adsorption method is not affected by the

colour of the solid. It enables simultaneous measurement of the amount,

strength and type of acid sites present even at the elevated temperatures

at which disproportionation catalysts are used. Whereas n-butylamine is

generally used as base for the titration method, different bases are

suitable for the adsorption method, such as pyridine (11,15), trimeth-

ylamine (11,12,13,14), ammonia (13,16-25,29,30), n-butylamine (26),

quinoline (27,28), pyrrole (13) and piperidine (12). The possible tech­

niques for detecting the amount of base adsorbed include infra-red

spectrometry, gravimetry, calorimetry, volumetry, differential thermal

analysis, thermogravimetry and thermal conductivity measurement.

Among the catalysts whose acid properties have been most widely in­

vestigated are alumina (15,16,24), silica (15,24) and silica-alumina

(15,17,22,24,36,37). A few studies (23,24,120,182) on the acidity of

zeolites have been reported. Stone and Walley (24) found that CaX

zeolite has stronger acid sites than NaX.Bcison et al. (23) measured

the acid strength and ammonia adsorption entropy of H-Y zeolite and

found, by comparison with the results of Clark and co-workers (17) for

silica-alumina gel, that H-Y, has many more acid sites than the

silica-alumina sample but that the strength of the acid sites is mode­

rate.

In the next sections, the amounts of acid present in some of the

catalysts described in chapter 3 is determined by the n-butylamine

titration method. Owing to the colour of the HY/6-A1F /Cu cata­

lyst, that is, deep red changing to black when coke is deposited, the

indicator method could not be used for it and ammonia adsorption at the

tenperatures used in the toluene disproportionation experiments was

applied.

83

4.3.2. Acidity measurement by base titration method.

Experimental

Materials

The following catalysts were investigated:

1. Linde SK-40 molecular sieve, a Y zeolite with binder.

2. The toluene disproportionation catalyst prepared as outlined in

chapter 3 and consisting of ammonium-exchanged Y zeolite, 18%B-A1F„

and 10% copper as copper nitrate

3. An untreated low-alumina silica-alumina catalyst (LAL3P).

4. A steam-treated low-alumina silica-alumina catalyst (LAL5P).

Reagent grade benzene, n-butylamine and crystal violet were purchased

commercially.

Procedure.

The acidity of each catalyst was measured before and after activa­

tion, which was effected by heating the catalyst sample in a stream of

hydrogen (60ml/min) in a fixed bed reactor from room temperature to

230 C at the rate of 1 C/min., holding for 2 hours, then raising the

temperature to 500 C at the rate of 2 C/min and holding at this temper­

ature until a total of 24 hours had elapsed from the beginning of the

activation period.

About Igof catalyst, particle size less than 0.21 mm, was placed in

each of three previously weighed 20ml-weighing bottles. Precautions

were taken to avoid contamination of the catalysts by moisture from

the atmosphere. Ten ml benzene, dried over molecular sieves 3A, was

added to each bottle. The approximate quantity of n-butylamine required

to neutralize the acid was calculated from the rule of thumb that -4 2

8x10 m mcles of base are needed per m surface area (10). After this,

three different amounts of a O.IM -.:tyiamine sclution in benzene

were added to the three catalyst suspensions prepared above so as to

bracket the approximate quantity previously calculated. After shaking

for about 4 hours, a small quantity (1ml)of suspension was withdrawn

from each bottle and tested with one drop of crystal violet solution

84

(0.1% in benzene) in order to identify the two bottles containing

amounts of n-butylamine which bracket the quantity needed for neutral­

ization. The procedure was repeated with three new suspensions of the

same catalyst, but with a different quantity of base, in order to

determine the acidity more accurately.

Results

Table 4-3 shows the total acidity of the catalysts determined by

titration with n-butylamine and using crystal violet as Hammett indica­

tor. Since the pKa of the indicator is +0.8, the quantities shown in

Table 4-3 are the amounts of acid sites (in mmol n-butylamine per gram

catalyst) whose acid strengths are greater than +0.8, that is, whose

H < 0.8(2). o — ^

Table 4-3 shows that NaY appears to have a higher acidity per gram

than the other catalysts. On the same basis the steamed silica-alumina

is much less acidic than the unsteamed sample. After activation, the

acidity of the catalysts per gram is either unchanged or the change is

within experimental error. The acidity of the catalysts are also shown 2

in Table 4-3 as the number of acid sites per m and as the percentage

of the total surface sites. For these calculations the surface areas,

S, calculated by the alternative method proposed in section 4.2.3 and

No.

1.

2.

3.

4.

Catalyst Before activa-After activa-Number of acid tion mmol/g tion mmol/g sites per m

NaY

NH Y+18%B-A1F + +10%Cu

Silica-alumina 3P (untreated)

Silica-alumina (steamed)

0.41

0.31

0.30

0.10

0.45

-

0.30

0.10

3.0x10^^

4.0x10 ''

4.0x10^^

7.0x10 ''

Percent total sites

5.0

7.0

7.0

12.0

Table 4-3:Total acidity of the catalysts at H £ 0.8

85

listed in Table 4-1, were used for the zeolite-based catalysts (No. 1

and No.2 in Table 4-3 are No.l and No.3 in Table 4-1) whereas the BET

surface areas were used for the silica-aluminas (No.3 and No.4 in Table

4-3 are No.8 and No.7 in Table 4-1). Table 4-3 shows that No.2 and

No. 3 have the same level of acidity per unit area. NaY has the lowest

acidity per unit area whereas steamed silica alumina has the highest.

4.3.3. Acidity measurement by ammonia adsorption method.

Experimental

Materials

Two different catalysts were used in this study:

1. The catalyst prepared by impregnating NH.-Y with 18%3-A1F_ and 10%

copper as copper nitrate. The preparation of this catalyst has been

described in chapter 3.

2. The untreated low-alumina (15% alumina), silica-alumina catalyst

(LAL3P), which has been used in many other experiments in this study.

Both catalysts were activated in situ at the beginning of the

experiment.

Ammonia, nitrogen and hydrogen, purchased commercially, were of

chemically pure quality. The ammonia, whose purity was verified by gas

chromatography to be 99.92%, was used without further purification. The

nitrogen and hydrogen were purified over reduced copper (BASF BTS R3-11)

catalyst to remove traces of oxygen and molecular sieve 3A to remove

moisture.

Equipment

The adsorption measurents were carried out in a Perkin-Elmer model

TGS-1 thermobalance equipped with a temperature programming unit. The

thermobalance was connected to a gas-dosing panel which made it possible

to work with a specified gas mixture. A baffle plate was installed just

above the furnace in order to minimize the effects of convection currents

on the temperature of the furnace.

86

Procedure

At the beginning of an adsorption series, the thermobalance was cali­

brated by placing five ferromagnetic standards (monel, alumel, nickel,

mumetal and nicoseal deep draw) m the sample pan, followed by tempera­

ture-programmed heating in a strong magnetic field. When a standard

reached its Curie Point the resulting loss of magnetic properties was

registered as an apparent loss of weight. From temperature readings on

the programmer at the known Curie Points a calibration curve for the

temperature of the furnace m terms of the programmer dial reading was

plotted.

After calibration of the Cahn electrobalance itself with standard

weights, about lOmg catalyst was introduced into the sample pan and ac­

tivated for 24 hours with a mixture of hydrogen (60ml/min) and nitrogen

(108ml/min). The furnace was heated from room temperature to 230 C at

the rate of 1 C/min., held at this temperature for 2 hours, then to

500 C at 2 C/mm and maintained at this temperature until the end of

activation.

The temperature of the furnace was then adjusted to the desired

value and a mixture of ammonia and nitrogen passed over the catalyst.

Care was taken to use the same total volumetric flow each time.

Adsorption isotherms were measured at 400, 430, 450, 470, and 500 C

for the zeolite-based catalyst and at 300, 350, 400, 450, 500°C for

the silica-alumina. The adsorption of ammonia on both catalysts was

fast and completely reversible. Adsorption equilibrium, which was

presumed when there was no further increase m the weight of the sample,

was reached m about a quarter of an hour. After adsorption, the ammo­

nia was desorbed by passing nitrogen over the sample. Although desorp­

tion was somewhat slower than adsorption, it, too, was complete withm

half an hour.

The measurements were duplicated, approadiir^ the adsorption equili­

brium from the high and the low temperature sides of the isotherms.

Isotherms obtained m duplicate runs with different batches of catalysts

agreed to within 5%.

87

Results

Ammonia adsorption on EY/^-AlF^Cu

Isotherms. The shapes of the ammonia adsorption isotherms (Figure 4-7)

are in agreement with the results of other investigators for similar so­

lids (17, 152). They rise steeply at low coverages and gradually at

higher surface coverages. The coverage at which this transition takes

place varies with the temperature. This is typical of surfaces having

strong as well as weak adsorption sites (17).

In order to characterize the type of adsorption, the experimental

data were analysed according to the Langmuir, Freundlich and Temkin

theories of adsorption. These theories are discussed in detail in the

literature (44). The linearized forms of the Langmuir, Freundlich and

Temkin equations are shown in Figures 4-8, 4-9 and 4-10. The straight

lines in these figures were determined by a least-squares computer

programme. Visual inspection indicates that the Freundlich isotherm ap­

pears to fit the experimental data best (see also Appendix 3).

Isosteric Heats of Adsorption. The isosteric heats of adsorption, q

were calculated from the Clausius-Clapeyron equation (39):

1 ^ = r i 1 "1 P R *• T^ T^ -'

where P and P_ are the equilibrium pressures corresponding to the same

level of adsorption on two isotherms at absolute temperatures T. and T_.

The surface coverage 0 was calculated from: 0 = g/g^, where g is milli­

grams ammonia adsorbed per gram catalyst and g^ is the maximum ammonia

adsorption, milligrams per gram catalyst. The lines for 400 and 500 C

(Figure 4-9) do not converge to a point like the rest. This may be due

to experimental error or may mean that the adsorption isotherms at these

temperatures do not conform to the Freundlich theory. The value of g

was estimated from the point of convergence of the lines for 430, 450

and 470 C to be 10 milligrams ammonia per gram catalyst.

The isosteric heats are plotted against the surface coverage in Fig­

ure 4-11. The average isosteric heats are also plotted in Figure 4-11.

8 0 -

6 0 -

00 I 2

4 0 -

100 200 300

Figure 4-7 Ammonia adsorption isotherms on ca ta lys t ABl.

0 400 °C;A 430 °C;x 450 °C; • 470 °C;V 500 °C.

1 0 -

0 8

^ 0 6 -

0 4

0 2

0 010 0 020 0 030 - 1 ' P N H , f"^"" Hg-'')

Figure 4-8 Langmuir plot for ammonia adsorption on catalyst ABl.

• 400 °C; A 430 °C;x 450 °C;o 470 °C; 0 500 °C.

89

2 0

1 5 -

1 0

c 0 5

3 O 4 0 5 O 6 0

NH3

Figure 4-9 Freundlich plot for ammonia adsorption on ca ta lys t ABl.

A 400 °C;o 430 °C;a450 °C,x 470°C;«500 °C.

10 0

8 O

- , 6 0

4 0 -

2 O

Figure 4-10 Temkin plot for ammonia adsorption on catalyst ABl

X 400 °C;o 430 °C;A 450 °C;D470 °C; • 500 °C.

90

160

120

8 0

4 0

0

-

-

~

y

\ ^

^V * " f V *

' f^

1

S . s ^ ^ \ ^

, /

A X

/

^

/

1

/ ^ ^

^ ^

-

-

:

-

-

0 2 - 9

70

60

- 5 0

- 4 0

0 4 0 6 0 8 1 0

Figure 4-11 I sos te r i c hea ts , q and adsorption ent ropies , S , of

ammonia on cat ABl.

• 430-450 °C;o 450-470 °C; D Average.

The gradual decrease of q with 0 indicates that the adsorption of

ammonia on the catalyst follows the Freundlich equation, 0 = g/g = DT /n

= (a p) ™ (Table 4-4). If the adsorption would have obeyed the

Langmuir theory, q would be invariant with temperature, whereas the

decrease of q would follow a straight line if the Temkin theory was

applicable.

Differential Entropies. The experimental differential entropies of am­

monia in the adsorbed state, S , were calculated with the following a

relationship (23,39,42): ^a = -'^s^'^ * "^^P" '^ ^ ^g

where q is the isosteric heat of adsorption, p is the equilibrium pres­

sure at constant coverage for the isotherm of absolute temperature T,

R is the gas constant, and p is the standard state pressure at which

the entropy of gaseous ammonia, S , at the temperature T, is evaluated. o

Values of S for ammonia at different temperatures were obtained from

the literature (41). S is plotted against coverage in Figure 4-11, a.

91

T,°C

430 - 470

g^,. mg/g

10

a ,(mmH ) 0 ' ' g'

8.8x10'"*

q ,kcal/mole

4 .3

Table 4-4. Constants of the Freundlich equation for ammonia

adsorption on HY/B-AIF^/Cu

which shows that the entropies are low at low coverage and increase to

rather high values as 0 increases. This may mean that the adsorbed

molecules are mobile on the catalyst surface.

The question of the mobility of the adsorbed molecules was further

investigated by comparing experimental and theoretical entropies (39,

43, 260). The three-dimensional entropy of translation of ammonia at

1 atm., assuming ideal gas behaviour, was calculated from the relation­

ship:

3/2 5/2 S, = RlnM T -2.30. The two-dimensional entropy of trans­

lation of the molecule in the adsorbed state was calculated with the relationship:

S2 = RlnJ TA + 65.80.

In both equations, M is the molecular weight, T is the absolute temper-2

ature, and A is the area (cm ) occupied by each molecule. If single site

adsorption is assumed, that is without association or dissociation, the

configurational entropy is given by

S = R[xlnx-(x-l)ln(x-l)]

where x = 1/0. Various values of the surface coverage and a cross-

sectional area of an ammonia molecule of 0.129nm (12.9A ) were used in

the calculations. The experimental entropy loss AS = S -S and the

theoretically calculated values, (S^-S,) and (S -S,), are shown in Table

4-5. (S_-S,) is the minimum entropy loss fin* mobile layers whereas

(S -S-) is the minimum loss for immobile layers. The results show that

92

9

0 . 5

0 . 6

0 . 7

0 . 8

0 . 9

S a

e .u .

4

17

28

37

44

S g

e .u .

54.8

54.8

54.8

54.8

54.8

h e .u .

38.8

38.8

38.8

38.8

38.8

^2

e .u .

17.3

17.3

17.3

17.3

17.3

S c

e .u .

2 . 8

2 . 3

1.7

1.3

0 . 7

S -S° a g

e .u .

-50 .4

-37 .8

-26 .8

-17 .8

-10 .8

S2-S3

e .u .

-21 .5

-21 .5

-21 .5

-21 .5

-21 .5

S -S^ c 3

e . u .

-36 .0

-36 .5

-37 .1

-37 .5

-38 .1

Table 4-5: Experimental and theoretical entropy losses for ammonia

adsorption on HY/&-AlF,/Cu.

Temperature range: 430-470 C

the experimental entropy loss is less than the quantities (S„-S,) and

(S -S_) at surface coverages higher than 0.8. This indicates that under

these conditions the adsorbed molecules are mobile. This conclusion is

in agreement with the results of other investigators on the adsorption

of ammonia on related catalysts (17).

The number of acid sites per gram of HY/B-AIF /Cu catalyst,N , was cal­

culated from the maximum ammonia adsorption, g (=10mg/g) by the follow­

ing relationship:

N = g. N /M, s ^ o '

where g is in grams ammonia per gram catalyst, M is the molecular weight

of ammonia and N is Avogadro's number. The surface area of the catalyst,

S, as computed by the alternative method proposed in section 4.2.3. and

listed in Table 4.1 was used to calculate the number of acid sites per

unit area and the percent of the total sites that are acidic. The re­

sulting value for the number of acid sites on the activated catalyst

93

Acid sites/g

20 3.54x10

Acid sites/m

6.46x10^^

% acid sites

11

Table 4-6:Acid sites on HY/6-A1F /Cu at 450°C

(Table 4.6) is higher than the result obtained before activation using

the n-butylamine titration method at room temperature (Table 4-3).

The type of acid sites present on the HY/B-A1F„/Cu catalyst was inves­

tigated by determining the ammonia adsorption capacity of samples ac­

tivated at different temperatures. Table 4-7 shows the adsorption capa­

city, at 500 C, of samples activated at 500°C and 550°C, as well as that

of a sample activated at 550 C to which a small quantity of water had

been added by injecting lyl into the nitrogen stream flowing over it.

The toluene disproportionation performance of the various samples is

also included in Table 4-7. The results show that the adsorption capa­

city of the catalyst drops by 70% after activation at 550 C instead of

500 C. Injection of a small amount of water virtually restores its

ammonia adsorption capacity to its orginal level. An analogous result

is obtained for the toluene disproportionation performance of the cat­

alyst: the activity of the catalyst activated at 550 C has all but dis­

appeared, but addition of a small amount of water partly restores the

activity. The results can be explained on the basis of the formation of

Br(6nsted and Lewis sites on the catalyst surface during activation of

zeolites (88,112). At a certain water content, dissociation of hydrate

water occurs, leading to the formation of protons which combine with

lattice oxygen atoms in the catalyst to form hydroxyl groups which act

as Brjinsted acid sites. Still further water loss, as the temperature of

activation increases to 550 C, results in dehydroxylation of the zeolite

and the formation of inactive and possibly also Lewis acidic sites.

Upon injection of water, these sites revert to Bronsted sites by becom­

ing rehydroxylated.

94

Ammonia adsorption (mg/g)

Ammonia partial pressure:

2U5 mm Hg

Nitrogen flowrate:

193 ml/min.

Total pressure :

1 ata

Activated at 500 °C

5.0

Toluene disproportionation

performance:

Total conversion, %

Selectivity , %

Reaction pressure: 1 ata

^/^toluene = 859 g.h/mol

H /toluene : I6.7 mol/mol

51

57

Activated at 550 °C

1.5

2

0

Activated at 550 °C then water -treated

5.2'

11 XX

100

KAfter injection of 1.0 yl water, xx After injection of 0.3 ml water.

Table 1t-7 Effect of activation temperature on ammonia adsorption

capacity and toluene disproportionation performance of

HY/e-AlF /Cu

Ammonia adsorption on a low-alumina catalyst.

The ammonia adsorption isotherms of this catalyst are shown in Figure

4-12. The isotherms have the same general shape as those of the zeolite-

based catalyst shown in Figure 4-7. The corresponding Langmuir,

Freundlich and Temkin plots are shown in Figure 4-13to4-15. The straight

lines were determined by a least-squares computer programme. It is clear

that the Langmuir plot does not yield a straight line and, therefore,

that the adsorption of ammonia on this catalyst does not follow the

Langmuir theory.

The Freundlich plot gives a straight line. However, because these

95

4 0 -

I z

Ol E

- P N H 3 ' " " " ^ '

Figure 4-12 Ammonia adsorption isotherms on LAL3P

o 300 °C;A 350 °C; • 400 °C;V 450 °C;x 500 °C,

10 -

0 8

:;: oe

I Z 0.4

O 2

0 01 - 1/P NH3

0 0 2 ( mm Hg " ' )

O 03

Figure 4-13 Langmuir plot for ammonia adsorption on LAL3P

D300 °C;o 350 °C; x 400 °C;A 450 °C; • 500 °C.

96

1 5 r

-o 5

0 0

-0 5 -

Figure 4-14 Freundlich plot for ammonia adsorption on LAL3P.

A 300 °C;o 350 °C;a400 °C; V 450 °C;x 500 °C

10K

8 -

X 4 z

4/f

" NH-.

Figure 4-15 Temkin plot for ammonia adsorption on LAL3P

V 300 °C;o 350 °C;A 400 °C;^450 °C;x 500 °C.

97

30

20

10

1 0 2 0 g (mgNH3/gcat)

3 0

Figure 4-16 Isosteric heats of ammonia adsorption on LAL3P.

A 450-500 °C;o 400-450 °C;D350-400 °C;V 300-350 °C.

Straight lines do not converge, it is not possible to estimate the

maximum adsorption, g . Consequently, the isosteric heats of adsorption

q . ,are plotted in Figure 4-16 as a function of the quantity of ammonia

adsorbed, g, instead of the more usual surface coverage, 0. The non-

convergence of the straight lines of the Freundlich plot may be due to

the narrow range of ammonia adsorption, 1.0-4.0mg/g catalyst, used or

to errors in the measurements of such low adsorption values. Nevertheless,

the plot of q , Figure 4-16, shows a decrease of q as a function of

g as postulated by the Freundlich and Temkin theories of adsorption.

The Temkin plot also gives a straight line (Figure 4-15). On closer

inspection, however, the Temkin theory appears to be inapplicable in

interpreting the results. Since the slopes of the lines shown in Figure

4-15 are positive, the quantity, g /q_-oi (Appendix 3) , must also be

positive. The parameters g and a are positive, therefore q must be

positive. However, the values of q calculated from the experimental

results. Table 4-8, are negative, which indicates that the data are

inconsistent with the theoretical basis of the Temkin isotherm. Thus

only the Freundlich isotherm is applicable.

98

T, °C

500

450

400

350

a , dimensionless

403 X 10^

103 X 10^

143

1.38

q , kcal/mol

-14.6

-16.2

- 6.1

- 0.39

Table 4-8: Constants of the Temkin equation for a low-

alumina silica-alumina. (LAL3P)

Figures 4-7 and 4-12 show that the silica-alumina is less acidic than

the zeolite-based catalyst. Although the range of the adsorption mea­

surements for the silica-alumina sample is limited. Figures 4-11 and

4-16 show that, under comparable adsorption values, the zeolite-based

catalyst shows higher heats of adsorption than the silica-alumina sample

and, therefore, possesses sites with higher acid strengths.

Acidity and toluene disproportionation activity.

The toluene disproportionation activities of the zeolite-based and

silica-alumina catalysts. Table 4-2, show that the former catalyst is

superior to the latter for this reaction. Figures 4-7 and 4-12 and the

discussion above show that, under toluene disproportionation conditions^

the zeolite-based catalyst is more acidic than the silica-alumina. The

superior activity of the zeolite-based catalyst is probably attributable

to its higher acidity.

4.4. Conclusion

Several methods have been used to characterize the texture and acid­

ity of zeolite-based and silica-alumina catalysts. The influence of

these properties on the toluene disproportionation activity of these

catalysts have been discussed.

The usual methods of estimating the surface areas of catalysts

proied inadequate when applied to zeolite-based catalysts. However, these

99

methods yielded results which were useful in comparing the texture of the

catalysts used in the disproportionation studies reported here.

An alternative method of estimating the surface area of zeolites and

zeolite-based catalysts has been proposed. The surface area obtained by

this method for Y zeolite containing binder suggests that this binder

blocks some of the micropores. The micropore volumes required for this

calculation have been obtained by applying the t-method.

Mercury penetration was used to determine the pore-size distribution

in the transitional pores, but mercury was incapable of penetrating the

micropores.

The ammonia adsorption isotherms of the zeolite-based and silica-

alumina catalysts were consistent with the Freundlich theory of adsorp­

tion.

The activity of the HY/B-AIF /Cu catalyst for toluene disproportion­

ation appears to be localized mainly in its transitional pores. The

micropores, on the other hand, gradually fill up, presumably with coke

or heavy products, as the reaction progresses.

The acidic properties of the zeolite-based catalyst are consistent

with the hypothesis that its high performance for the disproportionation

of toluene is due to its acidity. The lower acidity and acid strength

of silica-alumina are thought to be responsible for its lower dis­

proportionation performance

100

C H A P T E R 5

KINETICS OF TOLUENE DISPROPORTIONATION ON AN HY/B-ALF^/Cu CATALYST

5.1 INTRODUCTION

Very little of the large body of published information on the

disproportionation of toluene deals with the kinetics of the reaction.

The results of the kinetic studies reported in the literature are

summarized in Table 5-1. Izumi and Shiba (83), as well as Ogawa (175)

and Iwamura (84) identified the surface reaction as the rate-determining

step. The value of -1.0 kcal/mol for the heat of adsorption of toluene

in Ogawa and Iwamura's rate equation seems rather low in view of the

fact that chemisorption is very likely under their experimental

conditions. None of the above-mentioned authors corrected their data

for the effect of catalyst deactivation on the rate of reaction.

Yashima's (90) studies on H-mordenite took the rather pronounced

catalyst deactivation into account by using the reaction rates extra­

polated to zero time on stream. Since the initial catalyst activity

was invariably high, the rates obtained in this manner were also high

and it is, therefore, not surprising that above 350 C pore diffusion

was the rate-determining step: above this temperature, the apparent

activation energy decreased to 11.8 kcal/mol from 88.8 kcal/mol as was

determined below 350 C. Yashima found a zero order dependency of the

rate on the partial pressure of toluene and ascribed this to the strong

adsorption of toluene on the catalyst.

In this chapter, the kinetics is studied with the HY/fs-ALF /Cu

catalyst (18% B-ALF. and 10% Cu) prepared in chapter 3, since

experimental results showed that it possessed a satisfactory activity,

selectivity and stability. After a discussion of the methodology of

measuring and interpreting the kinetics of heterogeneous catalytic

reactions, some preliminary experiments are performed in order to study

the influence of some important process variables on the reaction, viz.

space time, temperature, toluene partial pressure, hydrogen partial

No.

1

2

3

Rate equation

T > 623 K: pore diffusion rate-determining

T < 623° K: r s exp (-18800/RT)

34.9 exp(-14700/RT)(P^ - P P /K^ )

(1 + 4.12 x 10'^ n exp(18000/RT))^

exp(-21000/RT) P^

'° (1 + exp(-1000/RT + M i ) P j 2

Catalyst

H-Mordenite

Alumina/Boria

unknown

T/K

548-696

723-783

653-713

P/Nm'^

1.0x10^

1.0x10^

27.4x10^ -70.9x10

Reference

(90)

(83)

(84, 175)

r^ = initial rate; P., P and P = partial pressures of benzene, toluene and xylenes (atm.).

n = total pressure (atm.).

Table 5-1 Literature results on the kinetics of the disproportionation of toluene.

102

pressure, hydrogen/toluene ratio and total pressure. Subsequently,

initial reaction rates are determined at 400, 430, 450, 470 and 500 C

and at 2.0, 3.0, 4.5, 6.0 and 9.0 atm total pressure. The hydrogen/

toluene ratio is kept constant at 16.7 mol/mol. A correction was applied

to account for catalyst deactivation. Other experiments were performed

to investigate the effects of the disproportionation products, benzene

and xylenes and the influence of the hydrogen/toluene ratio.

The results were used to establish a kinetic model for toluene

disproportionation.

5.2 Measurement of the kinetics of heterogeneous catalytic reactions

The term kinetics as applied to chemical processsrefers to the effects

of variables such as temperature and concentration on the rates of

chemical reactions, their determination and the subsequent interpretation

of the observed rates in terms of structures, interaction of reactants

and other relevant physico-chemical properties of the system. An

important goal of any kinetic investigation is to arrive at a reaction

rate equation, that is a mathematical model which describes the rate

in terms of such variables as temperature, pressure and concentrations

of reactants and products. When dealing with catalytic gas/solid

reactions such a model is developed by postulating likely mechanisms

and deriving theoretical rate models, which are then compared with

experimentally determined rates. The model which agrees most closely

with the observed data is selected as the "best" description of the

kinetics.

The experiments described in this study were performed in a

continuous fixed-bed micro reactor (227,228), which was operated

integrally for the preliminary runs and differentially for the kinetic

measurements.

5.2.1 Differential reactor method

In a differential reactor the reaction rate changes so little that

it can be considered constant at some average value throughout the

103

reactor. For a tubular reactor, the rate will be approximately constant

if changes in composition in the reactor are small. For this reason

such reactors behave differentially when small conversions take place.

The rate is calculated as follows:

r = L_ W/F

where r is the rate, E, is the fractional conversion and W/F is the

reciprocal space velocity of the reacting species.

5.2.2 Integral reactor method

In an integral reactor variations in the reaction rate occur which

are large enough to cause an appreciable conversion and changes in

composition which need to be considered in evaluating the rate.

Integral reactor data may be obtained under isobaric, isothermal, or

constant reactants ratio,conditions.In the isobaric method, the total

pressure and reactants ratio are held constant and for a series of

temperatures the conversion is measured as a function of reciprocal

space velocity. Isothermal data are obtained by keeping the temperature

and reactants ratio constant and determining curves of conversion as a

function of time at varying total pressures. Either the isobaric or the

isothermal method may be repeated at a different reactants ratio.

Figures 5-1 and 5-2 illustrate the different kinds of sych data.

5.2.3 The initial rate method

The initial rate is the, reaction rate at zero time or, alternatively,

for an infinitely small change in the composition of the reactant. The

initial rate may be determined from either integral or differential

reactor data. If a differential reactor is operated with reaction

products absent from the feed stream, the rate calculated with the

differential reactor formula given above is a close approximation to

the initial rate.

If the curve of conversion versus reciprocal space velocity obtained

104

430°C

400 °C

W/F (ghr/mol)

Figure 5-1 Hypothetical example of isobaric plot of integral reactor

data.

— W/F ( g h r / m o l )

Figure 5-2 Hypothetical example of isothermal plot of integral reactor data.

by the integral method as described above is extrapolated to zero

reciprocal space velocity and different ia ted graphically or analyt ical ly

at that point , the slope evaluated gives the i n i t i a l r a t e .

105

5,3 Influence of physical transport processes on the kinetics of

heterogeneous catalytic reactions

A heterogeneous catalytic process is made up of the following

sequence of steps:

(i) Transport of reactants from the bulk fluid to the adjoining fluid-

catalyst interface by film diffusion of reactants through the

Laminar boundary layer surrounding the catalyst particle.

(ii) Pore diffusion of reactants to the internal surface of the

catalyst, if porous.

(iii) Adsorption (chemisorption) of reactants on the active sites on

the catalyst.

(iv) Surface reaction of the adsorbed reactants to form adsorbed

products.

(v) Desorption of the adsorbed products from the active sites,

(vi) Pore diffusion of the products from the internal surface to the

outer surface of the particle.

(vii) Film diffusion of products through the laminar boundary layer

around the catalyst pellet to the bulk fluid stream.

If any of the physical steps, (i), (ii), (vi) and (vii) is slow in

comparison with the chemical reaction step, the overall rate as

determined by measurements in the bulk fluid phase will not accurately

reflect the intrinsic reaction rate and even the selectivity or product

distribution of the reaction may be affected (210). These physical

factors which can falsify the measured kinetics of a heterogeneous

catalytic reaction are (211-214): non-ideal flow in the reactor, film

diffusion, pore diffusion, adsorption, desorption and heat effects.

Heat effects in kinetic measurements are obviated by operating the

reactor isothermally. The effects of non-ideality and diffusion are

eliminated by choosing the reaction conditions in such a way that they

either do not occur or have been reduced to a level where their effects

can be neglected. With most of these steps eliminated as described above

the observed rate includes the adsorption-desorption and chemical

reaction steps, (iii)-(v), described above. The criteria which ensure

106

the absence or the elimination of the complicating effects of these

physical factors are described individually below whereas the

calculations concerning the kinetic studies discussed in this thesis

are presented in Appendix 4.

5.3.1 Non-ideality of the reactor

For the fixed-bed reactor employed here to approach plug-flow

behaviour, the radial velocity profile must be as flat as possible.

The following criteria (208, 210, 235, 236), which ensure that these

conditions are met, were applied to test the ideality of the

experimental reactor (see Appendix 4):

d/d ^ 20 P

L/d 5- 100 P

Re > 10 P

where d is the diameter of the fixed-bed reactor, d is the diameter of P

the catalyst particles, I, is the length of the fixed-bed reactor and

Re is the Reynolds number based on the particle diameter.

5.3.2 External and Internal Mass Transport Resistance

The absence of film and pore diffusion (209-211, 237, 238) was

checked by calculation. The significance of film diffusion was estimated

by calculating the drop in the partial pressure of the reactant over the

laminar boundary layer surrounding a catalyst particle using an

experimental reaction rate (see Appendix 4). The criterion applied was

that if ' p/p < 0.01, film diffusion could be considered negligible.

The criterion proposed by Weisz (240) was used to decide whether

pore diffusion could be disregarded:

d rP ^ = 4Vc <

e s

107

where $ is the Thiele modulus for a spherical catalyst pellet, d is

the diameter of the pellet,]' is the density of the catalyst pellet and

D is the effective diffusivity within the catalyst.

5.3.3 Pressure drop in the reactor

An excessive pressure drop over the reactor is undesirable not only

because it invalidates the plug-flow behaviour assumed for kinetic

analysis but also because it may adversely affect the selectivity of the

reaction under investigation. Methods for calculating the pressure

drop when compressible fluids flow over particles of different shapes

are described in the literature (242-247). Erguns equation is perhaps

the most widely used (245, 246) and is valid for spherical particles

and for a ratio of catalyst particle diameter to reactor inside diameter

of less than 0.05:

^ = ifL (l-j) C150_y ^ L d e- f Vd '• ^' • -•

2 where Ap = pressure drop, N/m

L = reactor length, m

J = fluid density, kg/m

V = superficial fluid velocity, based on empty reactor, m/s

d = catalyst particle diameter, m

E = intergranular porosity of catalyst bed 2

y = fluid viscosity, N.s/m

The pressure drop over the reactor under the conditions of the

experiments described in this chapter is calculated in Appendix 4 using

the relationship given above and assuming the catalyst particles to be

spherical.

5.4 Reaction rate models

Various kinds of rate models are used for correlating reaction rate

data (212). A model which is derived from some basic physical and

chemical phenomena taking place is described as mechanistic whereas one

108

which is not based on any such phenomena is termed empirical. The two

most widely used types are: power function and Hougen-Watson models

(213).

5.4.1 Power function rate models

In terms of partial pressures these are of the general form:

n ^j m , r = k, n R - ko n P.J

li=l' j=l -l

where r is the reaction rate, k and k are the rate constants for the

forward and reverse reactions respectively, P. and P. are the partial

pressures of the reactants and products respectively and a. and b. are

the reaction orders. The problem of using such a model to correlate

experimental data boils down to that of determining the reaction orders

and the rate constant (212, 219-226).

5.4.2 Hougen-Watson models

These models, which include Langmuir-Hinshelwood and Eley-Rideal

types, have the general form:

koexp(-E/RT) f(p) (1-a)

(1 + g(T, p) )"

In this equation the numerator is the product of a pure kinetic term,

f(p), and a potential term, (I-a), which corrects for the deviation

from thermodynamic equilibrium. The denominator contains adsorption

terms, g(T,P), which originate from the coverage of the active sites

by species present in the reacting system and n is the number of sites

involved in the rate-determining step. The Hougen-Watson (247, 250,

251 ) approach to reaction rate modelling can be summarized as follows:

The initial stage.

A family of plausible models is formulated by postulating a chemical

dissociation mechanism, assuming one of the elementary steps in the

mechanism to be rate-controlling while all the others are at

109

equilibrium, and deriving a reaction rate equation for each case. If

two adjacent active sites are involved in the rate-determining step,

the mechanism is the classical Langmuir-Hinshelwood dual-site type,

whereas if only one site is concerned it is an Eley-Rideal single-site

mechanism.

The intermediate stage.

The models developed in the previsous stage are fitted to the data by

non-linear regression analysis. Physical and chemical criteria are

used to reject those models which may be considered implausible. The

F-test is applied to the variances obtained for the unrejected models

in order to select those that have the lowest variances which are

indistinguishable from each other in a statistical sense.

The final stage.

If at this stage morethan one model is left, such quantities as surface

coverage and apparent activatian energy are calculated from these models and

compared with the corresponding values determined 'from the experimental

measurements (233). The trends in these quantities may provide further

arguments for a final choice among the models.

The Hougen-Watson method has been critized on theoretical (253) as

well as statistical grounds (205, 255). Weller (253) has challenged

the theoretical validity commonly attributed to Hougen-Watson models

and argued that power function models usually correlate kinetic data

just as well. Boudart (254), on the other hand, has defended the

rational application of Hougen-Watson models, emphasizing their value

in extrapolation and in gaining some insight into the mechanism of

reactions.

5.5 Experimental

5.5.1 Materials

The toluene, analytical grade, was used without further purification.

Chemically pure hydrogen was dried over molecular sieves 3A and passed

over reduced copper oxide (BASF R3-11 BTS) catalyst to remove traces of

oxygen and other impurities.

110

1

i ®

loMta

1 © ©

® _0]

_K_

^

£?1

®

n ®

®-<D HSh

©

CarriarQasCH*}

®

-® ^ (D

®

J ®

s

TTAP

@

Hj

@

® • - ® @

Figure 5-3 Flow sheet of the equipment.

The catalyst used is catalyst ABl (72% HY zeolite 18% 6-ALF +10% Cu),

the preparation of which has been described in chapter 3 and the physico-

chemical properties of which have been studied in chapter 4.

5.5.2 Equipment

The kinetic measurements were carried out in the continuous flow

apparatus of Figure 5-3. Dry and deaerated aromatics were fed from

reservoirs (Y) by Hughes micrometering pumps (2j to evaporator (3J,

the temperature of which was maintained at 230 C. The flow rates were

determined with the aid of microburettes (s) by closing solenoid valves

Hydrogen, purified over reduced copper oxide on silica (BASF R3-11

BTS) catalyst and molecular sieves 3A was metered with Brooks ELF

precision flow controllers (6). The gas flow rate was determined from

Ill

the pressure drop across a calibrated stainless steel capillary tubing

immersed in a thermostatically controlled water bath maintained at

40 °C.

The stainless steel reactor (30 cm long, 1 cm internal diameter)was

placed in a fluidized bed of carburundum (V) acting as a thermostatic

bath. The reactor temperature, which was measured at three points along

the axis of the reactor with chromel-alumel thermocouples, could be

kept constant to within + 1 C.

For off-line analysis, the aromatics were condensed in high pressure

condenser (sj cooled to -78 C. For on-line analysis, valve (V) reduced

the pressure to 1 atm., which was required for sampling valve H2) with

which samples of the product stream could be injected into gas

chromatograph Q_^. Except during sampling periods, the product stream

was freed of condensable components in low-pressure condenser \}v) ; the

non-condensable gases were passed through Brooks volumeter Q_l) to

measure the flow rate, and vented.

Analysis was accomplished by separating the components at 100 C

over GLC column (jj) (3 m long, 4 mm I.D. and 6 mm O.D.) packed with

chromosorb W impregnated with bentone and diisodecylphthlate, using

helium as the carrier gas. Flame Ionization Detection Q4) was used.

The mole fractions of the aromatics were calculated from the peak area

counts, obtained with digital integrator n6) , by the method of inter­

nal normalization, using toluene as the internal standard (see Appendix

5).

5.5.3 Procedure

For the kinetic experiments, a catalyst bed of 30 cm length,

containing 15-20 g catalyst, particle size 0.21-0.42 mm, was used.

These dimensions are such that a good approach to plug flow was assured.

Calculations also showed that, under the conditions of the experiments,

neither pore diffusion nor film diffusion limited the rate of toluene

disproportionation (see Appendix 4).

The catalyst was activated at 1 atm. total pressure and a hydrogen

flow rate of 60 ml/min. To this end, the reactor was heated from room

112

temperature to 230 °C at the rate of 1 C per minute and held at this

temperature for 2 hours. The temperature was then increased to 500 C

at the rate of 2 °C per minute and held at this temperature for about

18 hours.

5.6 Results

5.6.1 Preliminary experiments

After activation, toluene and hydrogen were passed over the catalyst.

Initially, catalyst activity increased to a maximum value, then

decreased rapidly shortly thereafter and more gradually as the stream

time increased. At the same time the catalyst selectivity goes through

a minimum, eventually reaching more than 90% (see chapter 3). The kinetic

measurements were performed during the period of slow deactivation and

high selectivity. In order to correct for the loss in activity, the

conversion, 5, was measured as a function of time on stream, t, at one

standard condition (P = 6 ata, T = 450 °C, H /Toluene = 16.7,and W/F =

= 176.6 g.hr/mol) at the end of each experimental run. The results

(Figure 5-'4) were fitted to a straight line by a computer program

utilizing the least-squares criterion:

C = 1 1 . 6 - O.OOlSt

1 5 -

o u

6 0 0 1000

" - Stream time,hrs

2000 3000

Figure 5-4 Conversion at standard conditions as a function of stream

time.

113

This relationship was used to correct for deactivation by normalizing

the measured conversions to the catalyst activity observed at 1400

hours stream time. The theory behind this method of correction is

presented in Appendix 6; the normalization equation is:

5 = 5 (11.6 - O.OOlSt) m.

where £ is the corrected conversion, £ is the measured conversion ' m

which is to be corrected and £ is the conversion measured at standard conditions at about the same stream time as £ . The value of 1400 hours

m was substituted for t in the above equation.

Influence of H^toluene ratio and total pressure on catalyst deactivation.

Figure 5.5 shows the results of two experiments, one with hydrogen

and toluene as the feed, the other with argon and toluene. The gas/

toluene ratio (16.6 mol/mol and the total pressure (1.1 ata) were the

same in both cases. The results show that deactivation is faster in the

absence of hydrogen and suggest that, at constant total pressure and

o u

50 100

Stream t ime , hrs 150

Figure 5-5 Influence of Hj/Toluene ratio on catalyst deactivatio

at 500 °C and 1.1 ata.

o H^/Toluene = 16.7 mol/mol;• Hj/Toluene = 0

114

Ot I I l_ 0 50 100 150

^m— Stream time ,hrs

Figure 5-6 Influence of t o t a l pressure on ca ta lys t deac t iva t ion .

o P^ ^ , = 1 . 1 a ta ;x P.,. .„ , = 3 a t a ; •P^ ^ , = 6 a t a . t o t a l t o t a l ' t o t a l

constant toluene partial pressure, the rate of deactivation increases

as the hydrogen/toluene ratio decreases.

Figure 5-6 shows the results of experiments designed to demonstrate

the effect of total pressure on catalyst deactivation, with temperature

and hydrogen/toluene ratio constant. The results indicate that, at a

constant hydrogen/toluene ratio, the rate of deactivation decreases as

the total pressure increases and becomes all but negligible at 6 ata

total pressure and 16.7 mol/mol hydrogen/toluene ratio.

Gonversion as a function of space time.

The variation of toluene conversion with space time is given in

Figures 5-7 and 5-8 for different temperatures and pressures and a

hydrogen/toluene ratio of 16.7 mol/mol. Both figures show a definite

curvature in the graphs at higher conversions, especially when the

temperatures and pressures are high. It appears that, for conversions

below 10%, the reactor may be considered differential. Since at such

low conversions the reaction products, benzene and xylenes, do not

influence the rate of toluene disproportionation (see section 5.6.2),

115

3 0 -

j 20 -

100 200 W / F g h r / m o l

300

Figure 5-7 Variation of conversion as a function of space time.

o 500 °C;x 450 °C.

30-

10O 200

W/Fg hr /mol

Figure 5-8 Variation of conversion as a function of space time at

500 °C and H2/Toluene = 1 6 . 7 mol/mol.

o P^ ^ = 10.5 ata;V P^ ^ = 6 a t a ; DP^ ^ ,= 3 a ta t o t a l ' t o t a l t o t a l

the i n i t i a l reaction r a t e , r , can be calculated by:

^o W/ F. D

toluene

where Yn is the yield of disproportionation products defined in

chapter 2 and W/F^ , is the space time. toluene

116

Influence of temperature.

The temperature dependency of the reaction is expressed in terms of

an apparent activation energy as determined by an Arrhenius plot,

Iigure 5-9, m which the initial rates calculated from Figure 5-7 are

used. The results show that the apparent activation energy is 19.8

kcal/mol.

-5 5

-6 0

-6 5

S -7 0 c

-7 5

-8 0

12 13 14 15 ^ — lOOO/T.K"''

Figure 5-9 Arrhenius plot for toluene disproportionation

''total ' "" ata.H^/Toluene = 16.7 mol/mol.

Influence of process variables on reaction rate.

The effect of hydrogen and toluene partial pressures, Pu and P„,

respectively, which are the basic independent variables of the process,

on the reaction rate was investigated qualitatively in the preliminary

experiments. The results are shown m Figure 5-10. Most but not all of

the measurements on which the figures are based were differential.

Consequently, only qualitative conclusions may be drawn from the rates,

r, shown m them. During the experiments of Figure 5-10, hydrogen

partial pressure was varied while the total pressure and the toluene

partial pressure were constant. This was realised by replacing hydrogen

as a diluent gas with argon while maintaining the total gas/toluene

ratio constant. The results indicate a definite influence of hydrogen

partial pressure on the rate. At low partial pressures, the dispro-

117

100 -

o E

in o

40

20

O o

2 4 pH2 ,atm

Figure 5-10 Influence of toluene and hydrogen partial pressures on

reaction rate.

o P^ ^ , = 6 ata;V P ^ , = 3 ata;«P^ ^ , = 1.1 ata total total ' total

portionation rate is seen to decrease rapidly at first with increasing

hydrogen partial pressure and to become independent of this variable,

within experimental accuracy, at high values of I . This effect is

further confirmed by the kinetic measurements shown in Figure 5-14.

The influence of toluene partial pressure is also evident in Figure

5-10. The rate increases with increasing toluene partial pressure. This

is further substantiated in the kinetic measurements (see section

5.6.2) when the hydrogen/toluene ratio is constant (Figure 5-12) and

when it is varying (Figure 5-14).

In the actual kinetic measurements (section 5.6.2) it was

experimentally inconvenient to work with a third (inert) component such

118

60 100 W/F , g h r /mo i

150 200

Figure 5-11 Yield as a function of space time

o 500 °C;A 450 °C; D430 °C; •400 °C.

160

•? 120 -

o E

in O

0 2 0 4

Pj (atm)

Figure 5-12 Initial rates of toluene disproportionation as a function

of toluene partial pressure.

o 430 C

V 450 °C

X 470 °C

• 500 °C

H-/Toluene = 16.7 mol/mol

Points :measured data.

Lines :calculated values.

Model -27, Table 5-7.

119

- 6 V

-10 1 2 1 3

1 0 0 0 / T , K" 1 4 1 5

Figure 5-13 Arrhenius plot at d i f fe ren t react ion pressures .

° P»- i-oi = 9 a ta ;x P._^ , = 6 ata;A P^ , , = 4.5 a ta t o t a l t o t a l ' t o t a l O P^ . , = 3 a ta ;» P, . , = 2 a ta . t o t a l t o t a l

as argon in order to vary the toluene partial pressure whilst keeping

the hydrogen partial pressure and the total pressure constant. There­

fore, the experiments were carried out by varying the partial pressure

of the two components, hydrogen and toluene, simultaneously whilst

keeping the hydrogen/toluene ratio and the total pressure constant

(Figure 5-12). Subsequently, experiments were performed at varying

hydrogen/toluene ratios (Figure 5-14) in order to investigate the

effect of hydrogen partial pressure on the rate.

5.6.2 Kinetic measurements

The isobaric method described in section 5.2.2 was employed in the

kinetic measurements, with the hydrogen/toluene ratio fixed at 16.7

mol/mol. Each measured product yield was corrected for catalyst deacti­

vation as described in secti?)n 5.6.1 and initial reaction rates were

120

720 -

640 -

560 -

4 6 0

400 -

o E

in o

320 -

240 -

160 -

Figure 5-14 I n i t i a l r a tes of toluene dispropor t ionat ion as a function

of toluene p a r t i a l p ressure .

» 400 °C, 11 ata

o 430 °C, 11 a ta

t 450 °C, 11 a ta Points

+ 450 °C, 9 a ta Lines

Model

H,/Toluene-3-80 mol/mol

X 450 C, 6 ata

7 470 °C, 11 ata

• 500 °C, 11 ata

:measured data .

rcalculated values.

•27, Table 5-7.

121

calculated as previously explained in the same section. The results of

yield as a function of space time for 6 ata total pressure are shown

in Figure 5-11. In Figure 5-12 the initial rates are given as a function

of the partial pressure of toluene.

The apparent activation energy of the reaction at different pressures

was estimated by making an Arrhenius plot with the data of Figure 5-12.

The plots are shown in Figure 5-13 while the results are summarised in

Table 5-2. The apparent activation energy appears to vary with total

pressure. The pressure dependency of the apparent activation energy

normally points to a temperature dependency of at least one of the

adsorption terms in the reaction rate equation.

Figure 5-14 shows the initial reaction rates measured at varying

hydrogen/toluene ratios, total pressures, and temperatures.

Modelling of the kinetic data.

In this section, theoretical reaction rate models are derived and

applied to the experimental data. In the derivation of the models, the

possibility of the dissociative adsorption of toluene was neglected

in view of the absence of toluene dissociation components in the

product.

The following toluene disproportionation mechanisms can be

postulated:

1. Molecular adsorption of both toluene and hydrogen, without the

formation of a surface complex. The elementary steps involved in

P ,, atm total

2.0

3.0

4.5

6.0

9.0

P.p, atm

0.113

0.169

0.254

0.339

0.509

Ea kcal/mol

22.3

18.5

19.4

21.2

21.3

Table 5-2 Apparent activation energy, Ea, as a function

of pressure.

122

this mechanism can be represented as follows:

H2 + s J H2,s

T + s J Ts

2Ts J Bs + Xs

Bs J B + s

Xs J X + s

By assuming one of these steps to be rate-determining, one obtains

the Hougen-Watson models shown in Table 5-3. The table includes

models derived from a single-site mechanism in which the reaction

step between two adsorbed toluene molecules is replaced by one

between an adsorbed toluene molecule and one in the gas phase. An

example of the derivation of models is given in Appendix 7.

Model No.

1

2

3

i+

No.of sites

2

2

1

1

Rate determining step

Adsorption of

toluene

Surface reaction

adsorption of

toluene

Surface reaction

Initial rate

k.p,

^'^W-^ k.p^2

fl^\-PH/'^T-PT^'

k.p,

^'^W^ k.p^2

^^^^•PH/'S-PT^

Figure 5-3 Reaction rate models from mechanism No.l

123

2. Molecular adsorption of toluene, dissociative adsorption of hydrogen

without the formation of a surface complex:

"2 * ^^

T + s

2Ts

Bs

Xs

<-

-*-

->-

2Hs

Ts

Bs + Xs

B + s

X + s

The models resulting from this mechanism are similar to those given

in Table 5-3 except that the hydrogen partial pressure p is 1 '^2

replaced by p* "2

3. Molecular adsorption of toluene, molecular adsorption of hydrogen,

with formation of a reactive surface complex.

H2 . s J H2,s T .H2.S : T^2'^ 2TH2,s t BH2,s +XH2,s

BH2,s i ^ "" ^2'^

XH2,s J X + ^2'^

The models obtained from this mechanism are given in Tabel 5-4.

4. Molecular adsorption of toluene, dissociative adsorption of

hydrogen, with formation of a reactive surface complex.

H2 + 2s J 2Hs

T + Hs t THs

2THs J BHs +XHs

BHs J B + Hs

XHs ? X + Hs

The models corresponding to this mechanism are obtained from those 1

given in Table 5-4 by substituting ]l„ for p„ . "2 2

5. Molecular adsorption of toluene, molecular adsorption of hydrogen

and formation of a non-reactive surface complex.

H2 + s -f- ^^2'

T + H2, s J TH2,S

T + s -> Ts

124

2Ts

Bs

Xs

Bs + Xs

B + s

X + s

The corresponding models are shown in Table 5-5.

6. Molecular adsorption of tojuene, dissociative adsorption of hydrogen

and formation of a non-reactive surface complex.

H2 + 2 s

T + Hs

T + s

2Ts

Bs

Xs

-<-

-<-

- * •

2Hs

THs

Ts

Bs + Xs

B + s

X + s

The models for this mechanism are obtained by replacing

p„ in the models in Table 5-5 by p„ "2 2

Model No.

9

10

11

12

No. of sites

2

2

1

1

Rate-determining step

Adsorption of

Surface reaction

Adsorption of

Surface reaction

Initial rate

k.PT.PH2

iUK^^^K^^H^P^)

•^•PH^-PT^

(1+Kj pH2 + K^.pH2.p ICj-P )

•^•PT-PH,

(UKj^ ?H K PH PJ)

2 2 ' 2

k.p.j,2.p ^

f 'Sl2PH2 ' cPTPH2 'SPT)

Table 5-4 Reaction rate models from mechanism N3.3

125

Model NO.

17

18

19

20

NO.of sites

2

2

1

1

Rate-determining step

Adsorption of toluene

Surface reaction

Adsorption of

Surface reaction

Initial rate

Kp,j, 1

(l Kj PH *\PH PT^ 2 2 ' 2

Kp^2

(UK^ PH ^^VT^\VJP^ )' 2 2 2

• PT 1 (UK^^p^^.K^p^^p,j,)

KPT^ 1 (UKj^^p^^.lCj.p^.K^p^p^^) 1

Table 5-5 Reaction rate models from mechanism NO.5

The models corresponding to the desorption of reaction products as

rate-determining steps are omitted in Tables 5-3 to 5-5. They yield

rates which are independent of toluene partial pressure. Inspection

of these models and their comparison with figure 5-13 show that they

are implausible. They are therefore not considered further (261, 262).

A total of 24 models are obtained, of which for the sake of brevity

only the models (1 lD4,9tol2,17 to 20)incorporating molecular adsorption

of hydrogen (mechanisms 1, 3 and 5 above) are listed in Tables 5-3 to

5-5.

The kinetic data plotted in Figures 5-12 and 5-14 were fitted to the

remaining models using a non-linear regression computer program based

on the modified steepest descent optimization method described by

Powell (256-258). The objective function minimized is Q, the sum of

squares of the deviations between measured rates,r , and the

126

corresponding calculated rates,r_ , from each model, weighted with ''i

the reciprocal of the squared observed rate:

" 2 2 Q =iil f-o. - ^c.^ 1-0.

i l l

Weighting the deviations in the manner just described gives each

measured point an equal weight in the regression analysis. The

computer program also calculates the sum of squares of the weighted

residuals between measured and calculated rates at convergence, SSQR.

Using SSQR and the number of degrees of freedom, v, the variance about 2

regression, s", and the standard deviation, s, can be calculated:

SSQR = 0^. =.2:, (r - r )^/r^ ^ Tnin 1=1 o. c. o.

I l l

V = n - p, where n is the number of data points and p is the number of

parameters in a particular model; v = SSQR/v.

The variance and the standard deviation are a measure of the goodness

of fit of a model.

The starting values of the parameters of the various models were

estimated by linear regression. The results of the subsequent non­

linear regression analysis are shown in Table 5-6 for the models with

the smallest variances.

As was previously pointed out, the apparent activation energy, Ea,

seems to be a function of the total pressure at a constant hydrogen/

/toluene ratio (Figure 5-13, Table 5-2), which would mean that at

least one of the adsorption terms in the denominators of the reaction

rate equations is probably temperature dependent. Consequently, a

temperature dependency was introduced in the adsorption constants of

models 1, 5, 17 and 21 which have the smallest variances among models

1-24 by replacing each adsorption constant, K, in each of these models

by K exp (-AH/RT). The new models and their results are given in

Table 5-7 as models 25-28. These results show that of all the models

tested 27 and 28 have the smallest variances. An F-test (see Appendix

8) on these variances reveals that the difference between the two

models is not significant at the 95% confidence level. Hence it must

be concluded that from a statistical point of view it is impossible

Model No.

-

1

S

17

21

Initial rate, r ' 0

mol g'.hr.

ko exp(-E^/RT)p^

K exp(-E^/RT)p.j,

'''H2PH2

k exp(-E^/RT)p^

^^'^H2PH2"'^CPTPH2

k^ exp(-E /RT)p.j,

^"VH2 "' CPTPH2

k 0

mol g.hr.atm

5.99x10'*

1.29x10^

4.09x10'*

3.80x10^

E a

kcal mol

24.5

24.6

24.1

24.2

s atm"

0.109

0.924

0.035

0.200

K c

atm

0.038

0.150

sum of squares

-

3.708

3.935

2.869

2.965

variance for lack of fit

0.046

0.049

0.036

0.037

Table 5-6:Results of the regression on the initial toluene disproportionation rate.

128

to say that one of these model^ describes the data best: 1^ exp(-Ea/RT)(p -p| f /K^)

1. r = -1 1 B X e _ _ ^ ^

(1+K„ exp(-AHH /RT)p„ +Kc,oexp(- AH /RT)p p ) H, ,0 2 "2 y.

k^exp(-Ea/RT)(p.j,-Pg^ p 5 /K^) 2. r

1 + K ^exp(-AH^ /RT)if„ +K .,o exp(-6H /RT)p.j.if )

The final parameter values of these models are contained in Table 5-7.

The rates calculated with model 27 are plotted in Figures 5-12 and

5-14.

Influence of the reaction products.

The effect of the products on the rate was studied by adding either

benzene or m-xylene to the toluene-hydrogen feed. At temperatures

between 400 and 500 C, total pressures between 1 and 10 ata, and a

constant hydrogen/aromatics ratio of 16.7 mol/mol, either benzene or

m-xylene was added in a ratio to toluene of 1:4.

The values (Tables 5-9 and 5-10) of the reaction rates, r, obtained

under the above experimental conditions are compared with the

corresponding rates in the absence of reaction products, r , extracted

from Figure 5-12.

Table 5-9 shows that benzene appears to have no measurable effect

on the rate. Since deactivation in the presence of m-xylene was faster

than with toluene only, reference measurements were made with a pure

toluene feed at the same standard conditions and in the same manner

as described in section 5.6.1. Each reference measurement was used to

correct each experimental point for the effect of deactivation in the

usual manner. The disproportionation rate calculated after correction

for deactivation. Table 5-10, shows that addition of m-xylene to the

feed retards the rate of reaction. The experiments also revealed that

equilibrium among the xylene isomers is rapidly established. Furthermore,

on the strength of the results (see Table 5-11) of another set of

experiments, already mentioned in chapter 2, in which two aromatics

mixtures (see Appendix 9 for details of the experiments) were passed

Model No.

-

25

26

27«

28' '

I n i t i a l r a t e , r

mol g .h r .

k^ exp(-Ea/RT)P^

1+Kjj^^Qexp(-AH^^/RT)P^

k^ exp(-Ea/RT)P^

K

2

l+Kjj^^Qexp(-AHjj^/RT)Pjj^^

k^ exp(-Ea/RT)P^

k^ exp(-Ea/RT)P^

^ ^ H ^ ' " ^ C V H /

mol g . h r . a t m .

1. Ixio' '

O.ll+U

2.93x10^

0.31+8x10^

Ea

kcal mol

13.9

17.0

13.6

\ , o

atm

2.69x10"^

1.98x10"^

6.10x10"^

2.36x10"'''

' \

kcal mol

-25 .0

-28 .6

- 2 2 . 5

- 1 9 . 8

^ c , a

^ -2 atm

1.77x10"^

1.5itx10"^

AH c

k c a l mol

- 2 3 . 9

-29 .5

- Lof

2.05

2.55

1.28

1.3U

2 s

0.025

0.021+

0.015

0.016

a K = K exp(-AH„ /ET); K = K exp(-AH /RT) n- ii ,U li„ c 0,0 C

Table 5-T: Results of the regression on the initial toluene disproportionation rate.

130

Model No.

1

5

17

18

20

21

25

26

27

28

2 ^Lof

0.046

0.049

0.036

0.053

0.052

0.037

0.025

0.024

0.015

0.016

' Lof

76

76

75

74

74

75

75

75

73

73

F e

7.98

8.47

6.21

9.19

8.95

6.41

4.30

4.16

2.59

2.73

F

1.70

1.70

1.70

1.70

1.70

1.70

1.70

1.70

1.70

1.70

F s

3.10

3.30

2.40

3.58

3.48

2.48

1.66

1.60

1.00

1.05

F se

1.50

1.50

1.50

1.50

1.50

1.50

1.50

1.50

1.50

1.50

Table 5-8: Results of F-test on variances of the models.

P^ ^ ,,atm total'

3

6

3

6

3

6

T°C

400

400

450

450

500

500

P.j,,atm

0.135

0.270

0.135

0.270

0.135

0.270

P ,atm

0.034

0.068

0.034

0.068

0.034

0.068

P„ ,atm "2

2.83

5..66

2.83

5.66

2.83

5.66

H2 aromatics

16.7

16.7

16.7

16.7

16.7

16.7

rate.r, benzene added mol/g.hr

7x10'^

14x10"^

21x10"^

45x10'^

53x10'^

103x10'^

rate r no benzene added* mol/g.hr

7x10"^

15x10'^

23x10"^

44x10"^

57x10'^

112x10"^

K Extracted from Figure 5-12

Table 5-9:Effect of benzene on rate of toluene disproportionation.

131

P^ ^ ,,atm total'

3

6

3

6

3

6

T°C

400

400

450

450

500

500

P.j,,atm

0.135

0.269

0.135

0.269

0.135

0.269

P ,atm X '

0.034

0.070

0.034

0.070

0.034

0.070

P^ ,atm

2.83

5.66

2.83

5.66

2.83

5.66

"2 aromatic

16.7

16.7

16.7

16.7

16.7

16.7

s

rate,r, benzene added

rate r no benzene added ^

r,mol/g.hr

6x10"^

12x10"^

23x10"^

40x10"^

49x10"^

88x10"^

r ,mol/g.hr

7x10"^

15x10'^

23x10"^

44x10"^

57x10"^

112x10'^

X Extracted from Figure 5-12

Table 5-10:Effect of m-xylene on rate of toluene disproportionation.

over the catalyst under the same conditions, the trimethylbenzenes

formed when xylenes are present in the disproportionation feed can be

ascribed entirely to the disproportionation of xylenes. This means that

under the conditions of the experiments this reaction is faster than

the transalkylation between xylenes and toluene. Accordingly, the

relationships derived in chapter 2 for the case when xylene

disproportionation occurs to the exclusion of transalkylation were

used to calculate the conversion, yield and selectivity of toluene

disproportionation with m-xylene in the feed.

The retarding effect of m-xylene can be explained by supposing that

xylenes are adsorbed competitively on disproportionation sites, thus

decreasing the total surface available for toluene disproportionation.

That m-xylene is adsorbed more strongly than toluene is consistent

with the higher basicity of the former compound, and the acidic nature

of the catalyst as discussed in chapter 4.

In order to account for this effect of xylenes, the models given in

Table 5-7 need to be extended with an adsorption term for xylenes, K P

132

in the denominator. The results shown in Table 5-10 suggest that the

^lene adsorption equilibrium constant, K , is temperature-independent.

Mixture

1 BX

2 BTX

1 BX

2 BTX

Pfotal'^^"'

4

6

6

9

P ,atm x'

0.113

0.113

0.170

0.170

Y, %

11

10

14

15

r,mol/g .hr

32x10"^

30x10"^

40x10'^

42x10"^

Table 5-11:Results of experiments for comparison of rates of xylene

disproportionation and transalkylation.

Non-linear regression of the extended form of model 27 on all data

points shown in Figures 5-12 and 5-14 and in Table 5-10, again using

the sum of squares of the relative deviations as the objective function,

results in the following rate equation for the disproportionation of

k(P^-Pg^/^P^^/2/Kg)

= 1.54x10^ exp (-19504/RT)

= 1.81x10'•^°exp(26495/RT)

= 7.51xlO"^exp(18609/RT)

= 6.83

= thermodynamic equilibrium constant

The sum of squares SSQR yielded by the regression analysis is 2.94.

In view of the paucity of the data obtained with xylene in the feed

(Table 5-10), care must be taken in interpreting the above model.

5.7 DISCUSSION AND CONCLUSION

The results of the preliminary experiments established that a high

hydrogen/toluene ratio coupled with a high total pressure and therefore

a high hydrogen partial pressure are necessary in order to reduce

toluene:

where k

K "2

c K X

K

133

catalyst deactivation. The effect of hydrogen partial pressure as

determined in the same preliminary experiments is consistent with the

formation of a surface complex as postulated in the models used to

interprete the kinetic data. Nevertheless, definite conclusions cannot

be drawn from the preliminary experiments alone concerning the form

of the reaction rate equation.

A comparison between the rate of reaction for toluene disproportio­

nation found in this study with those of other investigators (Table

5-1) shows that the results are quite different. The results of our

experiments demonstrate that the adsorption of toluene is the rate-

determining step. Other authors ignored the influence of the partial

pressure of hydrogen which, as this study demonstrates, affects the

rate of reaction.

The present work shows that two models give a statistically

significantly better fit than the others examined. These models, shown

in Table 5-7, assume the adsorption of toluene to be rate-limiting and

contain a mixed adsorption term in the denominator. The presence of

such a term in a rate equation is usually indicative of the formation

of a complex on the surface of the catalyst. However, a definite

answer as to whether this complex is indeed present cannot be given

from kinetic data only. Among the two models, the one assuming

molecular adsorption of hydrogen yields the smaller variance, but the

difference between the two models is not significant at the 95%

confidence level. Hence it must be concluded, from a purely statistical

point of view, that it is impossible to choose between the models.

Experiments with reaction products added to the feed have

demonstrated that benzene has no measurable influence on the rate of

toluene disproportionation. Xylene, on the other hand, has a retarding

effect on the rate. This phenomenon has been explained by partial

coverage of the catalyst surface with this compound. Accordingly, the

reaction rate model should include an adsorption term for xylenes in

the denominator.

134

CHAPTER 6

DESIGN CONSIDERATIONS

In previous chapters, information has been obtained on the process

conditions at which the disproportionation of toluene should be carried

out. The following need to be specified when designing the reactor:

reactor type, reactor inlet temperature and pressure, toluene flow rate

and hydrogen/toluene ratio (T,P,F and R, respectively) as well as r,

the reaction rate equation, the required conversion, catalyst particle

dimensions and the mode of operation of the reactor, that is, whether

adiabatic, isothermal or otherwise.

In addition to these specifications, it is necessary to put

constraints on the temperature and pressure in the reactor. An

excessively high temperature (higher than 500 C) is detrimental to the

life of the catalyst and to the conversion and selectivity of the

reaction (cf. chapter 3). The reaction pressure and the hydrogen/toluene

ratio together exert a great influence on catalyst activity, which

increases with total pressure at constant T, R and W/F(see Figure 5-6).

Similarly, at constant T, P and W/F, catalyst activity is more stable

as R increases. Evidently, at a given value of P, there is a

corresponding value of R which yields a stable catalyst activity and

vice versa. The higher P is the lower R needs to be;the reverse is

also true. The lower R is the higher the stable activity level of the

catalyst at constant T, P and W/F, since the negative influence of

hydrogen (cf.chapter 5) is then smaller.

The above discussion demonstrates that the positive effect on catalyst

stability of high values of P and R has to be weighed against the

deleterious effect of a high hydrogen partial pressure on the rate of

reaction. Figure 6-1 shows the toluene disproportionation conversion

and selectivity of a fresh sample of catalyst ABl (72% HY/18% B-AlF,/

/10% Cu), described in chapteis 3 and 5, as a function of stream time

and indicates that the life and performance of the catalyst may be

regarded as satisfactory. Moreover, experiments with used samples have

135

TJBT

D . ^

selectivity j< „

conversion - 5 9 a c ^

25 50 Stream t ime hrs

75 100 125 150

Figure 6-1 Performance of catalyst ABl at 500 C and 10 ata.

(see Table 6-1 for other process conditions).

shown that a deactivated catalyst is completely regenerable. Regenera­

tion is accomplished in situ by passing air (60 ni/min) over the

catalyst while raising the reactor temperature to 230 °C at the rate

of 2 C/min, holding for 2 hours, heating again to 500 C at 1 °C/min

and holding for 18 hours. After this, the catalyst is cooled to room

temperature and subsequently subjected to the same heating procedure

just described, with hydrogen (60 ml/min) instead of air passing over

it, in order to activate it (265).

Industrial toluene disproportionation processes employ total

pressures of 30 ata and higher (cf. chapter 1). Presumably this may be

partly due to the fact that for these processes a high hydrogen partial

pressure is needed to maintain the catalyst at a reasonably stable

level of activity. The performance of the catalyst used to obtain the

data of Figure 6-1 is compared with that of a probable industrial

catalyst in Table 6-1. From the information shown in this table it can

be estimated that about 75,000 kg of catalyst ABl are needed to

136

Reaction variables

W/F, g. cat.hr/mol.toluene

Stream time, hrs

Temperature, °C

Pressure, ata

Hydrogen/toluene, mol/mol

Conversion, %

Selectivity, %

Catalyst ABl*''

177

100

500

10

16.7

32

91

Industrial

252

100

420

35

15

42

> 90

catalyst

K Catalyst = natural mordenite treated with Hci (259).

KK Catalyst AB1= 72% HY/18% 6-AlF2/10% Cu

Table 6-1 Comparison of the performance of catalyst ABl with that

of a probable industrial catalyst.

Q

disproportionate 1x10 kg of toluene per year under the conditions

specified, whereas 80,000 kg of the industrial catalyst are required.

However, in view of the differences in the experimental conditions

employed to obtain the data of Table 6-1, it is difficult to critically

compare the two catalysts.

Finally, it is expected that the performance of catalyst ABl can be

further improved. One way is to operate at a higher reaction pressure

than shown in Table 6-1. This would enable a lower hydrogen/toluene

ratio than 16.7 to be used without lowering catalyst stability. This

would result in a higher conversion than is given in Table 6-1 (268,

269).

137

APPENDIX I

Temperature, K

component

Methane

Benzene

Toluene

o-Xylene

m-Xylene

p-Xylene

1,2,3-tri-methylbenzene

1,2,4,-tri­methylbenzene

1,3,5-tri-methylbenzene

1,2,3,4-tetra-methylbenzene

1,2,3,4-tetra-methylbenzenc

1,2,4,5-tetra-methylbenzene

Pentamethyl benzene

Hexamethyl benzene

Methylcyclo­hexane

300

-12.11

31.06

29.27

29.33

28.55

29.10

29.97

28.14

28.38

29.74

28.62

28.79

29.77

31.47

6.79

400

-10.07

35.0

35.30

37.89

37.0

37.69

41.08

39.00

39.50

43.36

42.10

42.38

46.02

50.80

21.84

500

-7.85

39.24

41.70

46.86

45.9

46.73

52.70

50.36

51.14

57.48

56.13

56.50

62.83

70.69

37.50

600

-5.51

43.66

48.32

56.10

55.10

56.06

64.67

62.06

63.12

71.95

70.51

70.99

80.00

90.95

53.53

700

-3.06

48.2

55.1

65.56

64.50

65.61

76.90

74.02

75.37

86.66

85.16

85.76

97.43

111.47

69.79

800

-0.56

52.84

61.98

75.I2I

74.02

75.29

89.27

86.1

87.75

101.50

99.94

100.66

114.98

132.12

86.15

900

1.99

57.53

68.93

84.78

83.64

85.06

101.76

98.33

100.26

116.46

114.85

115.69

132.66

152.88

102.58

1000

4.58

62.27

75.91

94.50

93.32

94.90

114.32

110.61

112.84

131.48

129.82

130.80

150.39

173.70

ill9.03

Table 1: Standard

kcal/mol

f ree enthalpy of format ion, AG_, of the components,

(179)

138

Temperature, K

component

Methane

Benzene

Toluene

o-Xylene

m-Xylene

p-Xylene

1,2,3-Tr-mcthylbenzene

1,2,4-Tri­methylbenzene

1,3,5-Tri­methylbenzene

1,2,3,5-tetra-methylbenzene

1,2,3,5-tetra-methylbenzene

1,2,4,5-Tetra-methylbenzene

Pentamethyl benzene

Hexamethyl benzene

Methylcyclo­hexane

300

-17.90

19.79

11.92

4.50

4.08

4.25

-2.34

-3.38

-3.89

-10.07

-10.76

-10.87

-17.85

-25.31

-37.05

400

-18.63

18.56

10.34

2.72

2.18

2.32

-4.52

-5.56

-6.14

-12.27

-13.06

-13.18

-20.26

-27.81

-39.78

500

-19.30

17.54

9.05

1.19

0.57

0.68

-6.42

-7.44

-8.07

-14.14

-15.03

-15.18

-22.26

-29.86

-41.91

600

-19.90

16.71

8.02

-0.07

-0.75

-0.67

-8.00

-9.01

-9.66

-15.67

-16.66

-16.85

-23.88

-31.50

-43.46

700

-20.40

16.04

7.24

-1.07

-1.79

-1.75

-9.27

-10.25

-10.92

-16.87

-17.94

-18.17

-25.12

-32.72

-44.49

800

-20.82

15.51

6.65

-1.85

-2.60

-2.59

-10.26

-11.21

-11.90

-17.78

-18.91

-19.19

-26.04

-33.61

-45.10

900

-21.15

15.10

6.24

-2.43

-3.19

-3.21

-10.99

-11.92

-12.61

-18.41

-19.60

-19.91

-26.66

-34.18

-45.33

1000

-21.40

14.82

6.01

-2.79

-3.58

-3.61

-11.46

-12.37

-13.06

-18.76

-20.01

-20.35

-26.98

-34.42

-45.23

Table 2: Standard Enthalpy of

/mol (179) .

formation, AH^, of the components kcal /

139

APPENDIX 2

In order to derive the conversion of toluene by disproportionation,

the disproportionation selectivity and the yield of disproportionation

products, all in terms of the mole fractions of the components detected

in the reaction products, equations 1,2 and 3 (see section 2.5) are

rewritten as follows, assuming that the feed consists only of hydrogen

and toluene:

2T i B^+X A

T.H2 - B2^M2 B

T+X^ J B2+Tr4B2 C

The following relationships can be established between the mole frac­

tions of the components:

Y = Y bl xl

y = Y b2 m2 Y = Y = Y b3 tmb3 x3

\V = 'bl^^b2^\3

Y = Y ,-Y . xp xl x3

Y = Y tmbp tmb3

In these relationships, Y, ,Y ,Y and Y , stand for the mole fractions '^ b' x' m tmb

of benzene, xylenes, methane and trimethylbenzenes. The subscript p is

used to denote the product, while 1,2,3 denote the reaction in which a

particular component reacts or is formed. The total conversion is de­

fined as follows:

J- _ mo^es toluene converted by all reactions moles toluene in the feed

Alternatively, a definition based on a phenyl or methyl balance, or on

140

mole fractions can be used instead of the absolute number of moles of

the components. From equations A,B and C above and using mole fractions

it is clear that:

(Y. ,+Y ,+Y,-+Y, JlOO ^ bl xl b2 b3 o (Y, .+Y ,+Y. .+Y, .+Y^ , ' bl xl b2 b3 tp)

where Y is the mole fraction of toluene in the product. By making use

of the relationships established above, the following final expression

is obtained:

(Y, +Y +Y^ K )100 _ _ hp xp tmbp^ 0^ (Y +Y +Y +Y , ) '* ^ bp tp xp tmbp

The selectivity to the disproportionation reaction is defined as

follows:

„ _ moles toluene converted by disproportionation moles of toluene converted by all reactions

" f^l^\l^^b2^^b3 '

2(Y +Y^ , )100 xp tmbp 0

~ XYTTf +Y1 , ) "" ^ bp xp tmbp'

Since we are mainly interested in the disproportionation of toluene,

the yield of disproportionation products will also be derived. The

definition is as follows:

Y moles toluene converted by disproportionation D moles of toluene in the feed

(Ybl^^xl^l°°

(Y, ,+Y 1+Y, _+Y, .,+Y^ ) ^ bl xl b2 b3 tp

141

2(Y +Y , )100 xp tmbp ^

(Yu +Y^ +Y +Y^ , , * bp tp xp tmbp)

When, on the other hand, the most probable reactions are 1,2 and 6

(see section 2.5) and the feed contains benzene and xylenes as well as

toluene and hydrogen, the conversion, selectivity and yield can be

derived as follows:

2T B+X

2X t T+TMB

T+H2 +- B+M

Reaction

1

2

3

Converted

T = tl

X = t2

T = t3

"2 " ^•^

Formed

B = tl/2

X = tl/2

T = t2/2

TMB = t2/2

B = t3

M = t3

V = methyl groups in the feed phenyl groups in the feed

= Y +2Y to xo Y, +Y^ +Y bo to xo

where Y, , Y^ and Y are the mole fractions of benzene, toluene and bo to xo

xylenes in the feed. From the phenyl balance:

Y. +Y^ +Y = Y,+Y^+Y +Y^ , , bo to xo b t X tmb'

and from the methyl balance:

142

Y, +2Y = Y, + 2Y +3Y^ ,+M bo XO t X tmb

From the stoichiometry of the reactions and the notations of the table

above, the following equations can be derived for the mole fractions:

Yu = Y, +t-, , +t., b bo 1/2 3

Y.. = Y^ -t,-t,+t-,T t to 1 3 2/2

Y = Y +t. ,^-t-X XO 1/2 2

^tmb " ''2/2

By solving the above equations simultaneously, one obtains:

h = 2(Y^„-Y,)+2(Y^^^-Y^)+2TMB

t = 2TMB

*3 = -^\o-\^-'^\o-V-™^ The phenyl balance can be used to simplify t, and t,:

t, = 2Y -2Y +4Y^ , 1 X xo tmb

^7 = Y,-Y -2Y^ .+Y -Y, 3 b x tmb xo bo

The conversion is defined as follows:

moles toluene converted by all reactions

moles toluene in the feed

^ = h-^2/2^^3

\'h-^2/2'H

= (Y,+Y +Y^ ,-Y -Y, )100 b X tmb xo bo

(Y, +Y^+Y +Y^ ,-Y, -Y ) b t X tmb bo xo -"

y- -« to

The slectivity is given by:

143

S = moles of toluene converted by disproportionation moles of toluene converted by all reactions

h ^r^2/2"^^3

(2Y -2Y +4Y^ . )100 x xo tmb

(Y.+Y +Y^ ,-Y. -Y ) b X tmb bo xo

(2Y -2Y +4Y^ u)100 X xo tmb

^^o-^t^

The yield is given by:

Y„ = moles of toluene converted by disproportionation moles of toluene in the feed

h \'h-^2/2^H

(2Y -2Y +4Y^ ,)100 X xo tmb „

(Y,+Y^+Y +Y .-Y. -Y ) '° b t X tmb bo xo

(2Y -2Y +4Y^ u)100 x xo tmb' „

- Y -6 ^ 6 to

The unknowns, Y and Y , can be eliminated from the above expressions to xo ^

by defining an extra quantity, K, which is the mole ratio of xylenes

and toluene in the feed:

K = Y /Y^ xo to

Combining K with the phenyl balance and the definition of V, one obtains

V = Y^ +2Y to xo Y. +Y, +Y to bo xo

Y, +Y +Y = Y +Y,+Y +Y , bo to xo t b X tmb

144

Y^ +2Y = Y^+Y,+Y +Y^ , to xo t b X tmb

V

Y^ = (Y^+Y,+Y +Y^ , )V to ^ t b X tmb

(I+2K)

2Y = 2VK(Y,+Y^+Y +Y^ , ) xo b t X tmb

1 + 2K

Substituting for Y^ and Y in 4, 5 ani 6 above, ^ to xo ' D '

((Y^+Y.+Y +Y^ ,)V-Y^(1+2K))100 ^ t b X tmb t '' o

(Y^+Y.+Y +Y, , )V ^ t b X tmb

((2Y +4Y^ ,)(1 + 2K)-2VK(Y^+Y,+Y +Y^ ,))100 q - X tmb'^ ' ^ t b X tmb ' „

f^^W^mb)V-^fl^2K)

((2Y^^4Y^^^)(1 + 2K)-2VK(Y^+Y^+Y^+Y^^^))100 ^

D - (YK+Y^+Y +Y^ , ) V

' b t X tmb'

When the feed contains only hydrogen, toluene and benzene, Y =0, from

the definition of V,

Y^ = V(Y^ +Y, ) to "• to bo'

From the phenyl balance,

Y^ +Y, = Y,+Y^+Y +Y^ , to bo b t X tmb

Therefore, Y = (Y,+Y^+Y +Y , )V ' t o b t X tmb

By substituting for Y and Y in equations 4,5 and 6 above:

(Y,+Y^+Y +Y^ ,)V-Y^)100 ^ ^ b t X tmb' t' „,

(Y,+Y^+Y +Y^ , )V b t X tmb

(2Y +4Y^ ,)100 X tmb „ ,

and (Y,+Y^+Y +Y^ jy-y^ b t X tmb' t

145

(2Y +4Y, ,)100 Y = X tmb' „ D (Y.+Y.+Y +Y^ , )V

b t X tmb

When the feed contains only hydrogen, toluene and xylenes, Y, =0.

From the definition of V,

Y^^ = r2-V>Y

From the phenyl ba l ance ,

Y^ +Y = Y,+Y^+Y +Y^ , t o xo b t x tmb

Therefore ,

Y . = Y,+Y^+Y +Y^ ,-Y = r2-ViY t o b t X tmb xo lTr~rJ ^°

Therefore ,

c-2-4,Y +Y = Y,+Y^+Y +Y . [rrrrj xo xo b t X tmb

Therefore,

Y = (Y,+Y^+Y +Y^ ,)CV-1) and xo ^ b t X tmb'^ '

Y, = (Y,+Y^+Y +Y^ K)(2-V) to ^ b t X tmb ^ '

By substituting for Y and Y in equations 4,5 and 6 above, one

obtains;

((Y^+Y^+Y^+Y^^^)(2-V)-Yjl00 ^

^ ^\^^^ V \ m b ^ ^2-V) '°

, K-%mb-^V\-V\mbH^-^^)^"" „ (Y,+Y^+Y +Y, ,)(2-V)-Y ^ b t X tmb' t •

Y K-^\mb-^(VV\mb^(^-^^^^°° D fV^^V^.mb^f2-V)

When the feed contains only toluene and hydrogen, V=l and K=0.

Substituting for V and K in any of the relationships above,

146

^^b-\-\mb)^°°

f^^^t^^x^^mb^

S = ^^^x-^^mb^l""

(2Y +4Y ,)100 Y ^ X tmb 5, D (Y,+Y^+Y +Y^ , ) °

b t X tmb'

When the feed contains benzene and xylenes as well as toluene and

hydrogen, the conversion, selectivity and yield, using reactions 1,2

and 3 instead of 1,2 and 6 are derived as follows:

2T

T+X

T+H,

^ B+X

t B+TMB

t B+M

1

2

3

Reaction

1

2

3

Converted

T=t^

T=t2

X=t2

T=t3

"2=^3

Formed

B=t^/2

X=t^/2

B=t2

TMB=t2

B=t3

M=t3

methyl groups in the feed phenyl groups in the feed

Y^ +2Y to xo

Yu "Y, + Y ~ bo to xo

where Y^ Y and Y are the mole fractions of benzene, toluene and to xo

xylenes in the feed. From the phenyl balances:

147

Y, +Y^ +Y = Y,+Y^+Y +Y^ , , and from the methyl balance: bo to xo b t x tmb' '

Y^ +2Y = Y^ + 2Y +3Y^ ,+M. to xo t X tmb

From the stoichiometry of the reactions and using the notations of the

table above, one drives the following equations for the mole fract ions:

Y, = Y, +t,/2+t-+t_ b bo 1 2 3

Y4- = Y^ - t , - t - - t , t to 1 2 3

Y = Y + t , /2 - t „ X xo r 2

Y = t tmb 2

By solving the above equations simultaneously, one obtains:

t, = 2(T -T)+2(B -B) 1 0 0

t- = (T -T)+(B -B)+(X -X) = Y^ , 2 ^ o ' ^ o 0 ' tmb

t, = -2(T -T)-3(B -B)-(X -X) 3 o 0 0

The phenyl balance can be used to simplify t. and t»:

t, = 2Y -2Y +2Y^ , 1 X xo tmb

t, = Y,-Y -2Y^ , Y, +Y 3 b X tmb bo xo

By combining the definitions of V and K with the phenyl balance, one

obtains:

Y^ = (Y,+Y^+Y +Y^ , )V to ^ b t X tmb

(1+2K)

2Y = 2/K(Y,+Y^+Y +Y^ ,) xo ^ b t X tmb'

(1+2K)

The conversion is defined as follows:

_ = moles toluene converted by all reactions moles toluene in the feed

148

= ^1^^2^^3

\'h'h'^3

- \ o - \ \o

((Y,+Y^+Y +Y^ .)V-y^(l+2K))l00 '' b t X tmb t -' 5,

(Y,+Y^+Y +Y^ , )V b t X tmb

The selectivity for disproportionation is defined as:

S = moles of toluene converted by disproportionation moles of toluene converted by all reactions

h ^l^h^^3

2Y -2Y +2Y^ , X xo tmb

Y,+Y +Y ,-Y, -b X tmb bo

2Y -2Y +2Y , X xo tmb Y. -Y^ to t

Y xo

= 2Y +2Y^ ,-2VK(Y,+Y^+Y +Y^ i/(l + 2K) X tmb ^ b t X tmb' '

(Y,+Y^+Y +Y^ , )V ^ b t X tmb'

(1+2K) " t

(^^V^^tmb»^^^'^^-^^'^^^b-\-V^mb^^^°" „ (Y^.Y^.Y^.Y^^^)V-Y^(1+2K)

The yield of disproportionation products is given by:

Y_ = moles of toluene converted by disproportionation moles of toluene in the feed

't^^l^^2^^3

= 2Y -2Y +2Y^ , X xo tmb

\o

149

= V^\mb-^^'^V^t-\mb^/^^-^'^^

f^b^^t^^x^^tmb^V/fl^^'^^

(Y,+Y^+Y^+Y^^,)V/(1+2K)

When the feed contains only hydrogen, toluene and benzene, K=0.

By substituting for K in the relationships above,

((Y.+Y^+Y +Y^ ,)V-Y^]lOO '• b t X tmb' t-' „

" CY.+Y^+Y +Y^ , )V ^ b t X tmb

(2Y +2Y^ K)100 g X tmb' o.

(Y,+Y^+Y +Y^ K)V-Y^ ^ b t X tmb' t

(2Y +2Y^ , )100 Y X tmb' D (Y,+Y^+Y +Y^ , )V

^ b t X tmb'

When the feed contains only toluene, xylenes and hydrogen, Y, =0 and

it can be shown from the definitions of V and K that,

K = V-J^ 2-V

Substituting for K in the relations previously derived,

(fV^t^V\mb^f2-V)-YjlOO ^ 5 = (Yu+Y^+Y +Y . )(2-V)

^ b t X tmb'^ '

Y,

(2Y +2Y^ ,-2(Y.+Y^+Y +Y^ ,)(V-l)]lOO S = X tmb ^ b t X tmb'^ ''

^\^\^^x^^mbn2-V)-Y,

(V^^tmb-^^V^^\mb^^^-l^)^°° „ D - (Yb-Y^-Y^-Y^^b)f2-V)

When the feed contains only toluene and hydrogen, V=l and K=0.

Substituting for V and K in any of the relationships above,

150

e = (Y,+Y +Y^ , ) 1 0 0 ^ b X t m b ' „ (Y^,+Y^+Y +Y, . ) "

b t X t m b '

S = ^^V^^tmb^^°° f b^^x^^mb^

D - TY^^Y^+Y^+Y^^^)

151

A P P E N D I X 3

ADSORPTION ISOTHERMS

INTRODUCTION

The isotherms which result from physical adsorption of vapours have

been classified into five types by Brunauer, Emmett and Teller (45).

Physical adsorption (and chemisorption) of gases are found to follow

Type I isotherms, while physically adsorbed vapours are more in

conformity with the other isotherms. From these isotherms, important

properties of the adsorbent can be determined, such as specific surface

area (Types II and IV), and pore size distribution (Type IV). The best-

known of the theories which give a mathematical expression of the

isotherms are those due to Langmuir; Freundlich; Temkin; and Brunauer,

Emmett and Teller.

Langmuir Isotherm

This isotherm may be expressed as follows (46 - 49):

e = g/g = ^P ^'^m 1+Kp

where 6 is the fraction of the surface covered; g is the amount of

adsorbate taken up by the adsorbent at the partial pressure, p; g

is the amount adsorbed when the surface is covered with a monolayer;

and K is the adsorption equilibrium coefficient. The dependence of K

with temperature is exponential (50):

K = Ko exp (- AH/RT)

where AH is the heat of adsorption and R is the gas constant. The

Langmuir isotherm can be linearised into the following form in order to

test its validity:

1 ^ 1__ i + i .

S 81 " ^m

152

Adsorption data which follow the Langmuir isotherm should yield a

straight line in the whole range of coverages from 9 = 0 to 6 = 1 when

—• is plotted against —. From such a ]

and the parameter K can be estimated.

—• is plotted against —. From such a plot the monolayer adsorption g

Freundlich Isotherm

This isotherm is an empirical relationship between the amount of an

adsorbate on an adsorbent surface and the pressure of the gas or vapour

in equilibrium with it (48, 51-53):

e = g/g, = c pi/"

where c and n are parameters dependent on temperature and on the

particular system under study. It has, however, been demonstrated that

the isotherm may be derived theoretically, either from a thermodynamic

or a statistical approach (39). The final form of the Freundlich

isotherm is:

6 = -8- = (a p)' 'r/% g o ^ **m

where a and q are constants which do not vary with temperature.

Numerically, a is the reciprocal of the pressure, p, necessary to

cover the surface with a monolayer while q (= - AH ) is the heat of ' ^m ^ m'

adsorption at monolayer coverage. For purposes of data analysis, the

isotherm may be rearranged into the following form and Ing plotted

against Inp: Ing = (Ing + Ina ) + — Inp

^m ^m

Temkin Isotherm

This isotherm, which is a modification of the Langmuir isotherm, may

be expressed as follows (54-56):

RT = = — In A g q a o °m ^o

153

where a is a positive constant, q (= - AH ) is the heat of adsorption

at zero surface coverage, A = a exp (q /RT), and a is a constant.

For correlation purposes, the isotherm is expressed as follows and g

plotted versus Inp: n q ?

a = -^ RT (Ina + -£) t- JH RT Inp • q a o RT q a ^

^o ^o

The BET Isotherm

This isotherm is obtained by extending the Langmuir theory of

adsorption, which focuses attention mainly on monolayer adsorption,

to multilayer physical adsorption. The formal form of the isotherm is

^^'^^- C(P/P )

g„ ' (l-P/P^)(l+P/P^(C-l))

where P is the vapour pressure of the adsorbate at the temperature

under consideration and C is a constant which depends on the nature

of the adsorbate-adsorbent system. For use in correlation of data the

BET isotherm is linearised into the following form, P/V(P -P) being

plotted against the relative pressure, P/P :

P V(P^ - P)

I

m + -^ - 1.

m

P • P

o

where V is the volimie of gas (NTP) adsorbed at a pressure P, V is the

volume necessary to form a monolayer and P and C are constants.

154

APPENDIX 4

TEST OF THE INFLUENCE OF TRANSPORT PROCESSES

1. Film diffusion . At steady state, the rate of reaction is equal to

the rate of transport of reactants through the film:

r = K Sv(P -P ) g g s'

^^fli .100^ = K (P -P ) 6(1-E,) 3600 g g s

3 where r = reaction rate, g moles/m .s

2 3 S = specific surface of catalyst, m /m

r = reaction rate, g moles/g.hr 3

S, = bulk density of catalyst bed, kg/m

d = effective diameter of catalyst particle = 6/Sv,m

K = mass transfer coefficient from fluid to particle, g 2

g moles/(m .s.atm)

P = partial pressure in the bulk gas, atm

P = partial pressure at outer surface of catalyst particle

atm

E, = porosity of catalyst bed

The mass transfer coefficient can be calculated from the relation­

ship of Chu et al.(266): D

8 m ^ _ , p ,-0.78 „ -2/3 ^ , e „. — 5 — =5.7 (rj-%- ) . Sc for 1<_ p - 30

where P = total pressure, atm

M = molecular weight of gas mixture, kyk mole

V = superficial velocity (i.e. based on empty reaior),m/s 3

S = density of gas mixture, kg/m

R = Reynolds number based on catalyst particle.

155

Therefore,

P -P r S. d M , ' e „ .,0 ,._ Ap g s _ o b p m _1 , p ,0.78 2/3 P " P " 21.6(1-E,) • V • 5.7 4-E,-' •

^ b' s b m

The variables which affect film diffusion are r ,P,T,R and ^v,., , O n.,

where T= Temperature, R is the hydrogen/toluene ratio and iiiv,„ is 2

the volumetric flowrate of hydrogen. It can be shown from the

equations to be used below that film diffusion is more probable at:

1) higher reaction rates, r

2) higher total pressures, P

3) lower temperatures, T

4) lower hydrogen/toluene ratios , R

5) lower hydrogen flow rates, CIJV,^ "2

By considering the data points at 400 and 500 C with the highest

and lowest reaction rates (Figures 5-12 and 5-14), it can be

verified that the reaction rate is the more critical variable.

Consequently, the pressure drop over the film is calculated for

the highest reaction rate:

r = 0.006225 gnol/.g.hr

P =11 atm

T = 773 K • •

R =5.9 mol/mol

c)>v,u = 33.6 ml/s "2

P =1.60 atm

d = reactor diameter = 0.01 m

d = 420xl0'^m P

E, =0.5 ' 3

S^ = 550 kg /m The fol lowing phys i ca l cons t an t s can be computed:

_3 M = 15.08x10 kg/kg mole

" 3 S = 2.62 kg/m-^ m ^'

V = 0.12 m/s

cjj, „-, the constant in the Lennard-Jones potential function and n,

156

the collision integral are calculated (238):

" T , H 2 =4-38^ n^ = 0.79

-5 2 From these, the effective diffusivity D„ ^ = 1.72.10 m /s

1 ."2

«D,H2 = °-" .„ CL, = 5.79.10" m

-10 a„ =2.97.10 m "2

The viscosities are calculated:

n,p = 1.79.10'^ kg/m.s

n„ = 1.81.10"^ k /m.s Ho g *^; =4.59 ^ ^ = 0.099 "2'^ -5 r\ = 1.80.10 k /m.s m g'

where n = viscosity of gas mixture. m

The dimensionless numbers are calculated: S V d

R = ^ £ = 14.3

p h \ n

Sc = -^-^ = 0.409 m T,H2

With the values computed above, the relative pressure drop is

obtained:

| £ = 1.3.10-^

On the basis of this it is concluded that the effect of film

diffusion is negligible.

. Pore diffusion. Since the heat effect of the disproportionation

reaction is small, the porous sphere model (240) for an isothermal

catalyst particle is applied to analyse the effect of pore

diffusion on the kinetics. Since the reaction may be considered

irreversible (all measurements were in the differential range) and

first order in toluene (the kinetic results indicate this), Weisz'

157

criterion may be .pplied: r r

exp D ,-C^ $ = " P < 1

eff S

where cji is the experimental Thiele modulus, r is the reaction exp , * ' c rate in gmole/m .s,r- is the radius of the catalyst particle, in

P 2 m,D ^. is the effective diffusivity, m /s,and C is the concentration ' eff J > I s

of the reactant on the outside surface of the catalyst,gmol/m .

It can be shown from the equations to be used below that pore

diffusion is more serious at 1) higher reaction rates and 2)

higher temperatures, and that for the present case, the other

process variables (P, R,etc.) have no influence on pore diffusion.

On the basis of these considerations, the effect of pore diffusion

is calculated for the highest reaction rate and highest tempera­

ture as was also chosen above for film diffusion:

r = 0.006225 gmol/g.hr

P =1.6 atm

T = 773 K

E, =0.5 b 3 S = 550 kg/m d = 420.10"^m P The pore radius, r , tortuousity factor,"^ and porosity of the catalyst, E , and the diffusivity of toluene in hydrogen already calculated above, D ,„_, have the following values:

-10 r = 4.5.10 m po t - 2 E = 0.5

" -5 2 D„ „ = 1.72.10 m' /s '."2

Knudsen diffusion which plays an important role in pore diffusion is calculated as follows:

F T 1/2 c -7 2

\ = 9 ^ - V ^& - f = 0-32.10 ^m2/s The effective diffusivity is given by:

°eff = F^— * 4 - " = 0.32.10-V/s "^ T,H2 K

158

p T 3

Cg = Rj = 25.2 gmol/m

0.006225 S^ ^Qoo , o n c ^ i / 3 ^o = (1 - E^) • 3600 = 1-9° ^"l/™ • =

With these values, ci = 0.11, which indicates that the effect ' exp

of pore diffusion is negligible.

Ideality of the reactor.

Reactor diameter, , „ „, ' d = 0.01 m

Catalyst diameter, d = 420.10" m ' P

Reactor length, L = 0.30 m d/d = 2 4

P L/d = 715

P For the experiments, R varied between 1 and 24, so that P varied

®p between 1 and 2. Consequently B is large enough in all cases. It is, therefore, concluded that the reactor was ideal under the

conditions of the experiments.

Pressure drop over the reactor

The pressure drop is given by:

Ap = L S V^ OzEi .150 p(l-E) ^ P d • ^3 *• S V d * i--'^)

p E p 2

where Ap = pressure drop, N/m

L = reactor length, m 3

S = gas density, kg/m

V = superficial fluid velocity (based on empty reactor),

m/s

d = catalyst particle diameter, m

E = intergranular porosity of catalyst bed 2

y = fluid viscosity, N.s/m

Using the same reaction conditions as in 1. and 2. above, and the

quantities already calculated:

L = 0.30 m

S =2.62 kg/m'

159

V = 0.12 m/s

d = 420xl0"^m P E = 0.5

y = 1.80 X 10"^Ns/m2

0.30x2.62x0.122x0.5 ,150x1.8xlO"^x0.5 , _^,

P 420 X io-bxo.5^— ^7T7TT7T:7T7-6 " " ^ 2.62x0.12x420x10'

1290 N/m^ = 0.0128 atm

On the basis of the above result, it is concluded that the effect

of pressure drop is negligible.

160

APPENDIX 5

INTERNAL NORMALIZATION

The response of an arbitrary component can be represented as

follows:

A. = K. g. 1 1 " 1

where A; is the response, for example peak area of component i

K. is the response factor of component i

g. is moles of component i

Therefore, g. = = B. A. 'l KJ 1 1

A similar relationship may be written for the standard component, s: g = B A 's s s

By combining both expressions, one obtains: g. k A. 6. A. A.

g k. A 6 A i,s A 's 1 s s s s

where g. is the experimentally determined response of i relative 1, s

to s.

If m. is the mole fraction of i, one obtains: 1

8. g A. ,A 6. A. m = gi = i,s.^ s . 1/ s ^ i,s 1 i n n n

j=l J j=l J.s S J S ^^^ ],S J

161

APPENDIX 6

CORRECTION FOR CATALYST DEACTIVATION

In order to develop a method of correction for catalyst deactivation,

the following assumptions are made:

1. Catalyst deactivation is due to the elimination of active sites with

time of reaction. Thus the number of active sites is a function of

time only,if other reaction variables are constant:N. = f(t).

2. The rate of reaction is a function of the number of active sites

and the reaction conditions. The form of this deactivation function

can be of any type (linear, exponential, polynomial, etc.) but must

not change during the course of the kinetic measurements.

3. Tte rate of reaction is a power function of the conversion. This is

because, as will be shown below, it is necessary to solve explicit^

for the conversion, which is the quantity usually measured in

kinetic studies.

Assuming an n power dependency of the rate on the conversion and

an m power for the rate on the number of active sites, one can write:

• s,t - '^^^")s,t = f(P' T' C < 1

' s,ts = '^f?"^s,ts = ffP'T.ON-^^ 2

• m,t = '^o(5")m,t = g^P.T.ON"^ 3

•^Cts - V^"^c,ts = gtP.T.C^N^^s '

where k and k are constants, R is the reaction rate at standard o s

conditions, R is the measured rate, R is the corrected rate, ts is m ' c ' the standard time, t is an arbitrary time and C is the conversion.

These equations can be transformed as follows:

^''s,t^'^" = ' l s,t = f'''"(P.T,C)N™/" 5

^'^s,ts)''" = '^l^s,ts= f^/"CP.T.C)<^ 6

162

f'^m,t^''" = 2^m,t = .^'"(P>T,C)N"'/"

^^,ts^''" = '^2^c.ts = g^'"(P.T.C)N™/^

6/5 =

8/7 =

There]

( R s , t s ) ^ / " (R^^^) i /n

( R e t s ) ! / "

m,t

fore , c , t s

m,t

^ s . t s

^ s , t

_ ^ c , t s

S , t

s , t s

j^m/n

j^m/n

Note:The correction applied must not be too great.

163

APPENDIX 7

DERIVATION OF A LANGMUIR-TYPE RATE MODEL

The mechanism considered is the following:

H2 + s

T + s

2 T s

B s

Xs

kl H2 s

T s

Bs+XS

B+ s

X+s

:K = o

•1 =

h -h -K, =

B * T

PB-VS P .C /C X s X

1

2

3

4

where s = active site

Ts = adsorbed toluene

Bs = adsorbed benzene

Xs = adsorbed xylene

H„,s = adsorbed hydrogen

C = concentration on the catalyst surface

Assuming that both film and pore diffusion are negligible and that the

adsorption of toluene is the rate-limiting step, one writes:

- = kiPTC3 ^2^1

By combining 2, 3 and 4, one obtains:

2 V x ^ 4 - K2K3K4

If L is the total number of active sites then:

L = Cs + Cj H2 ' S ' S^ x

PRP (C +K P„ C + (-JTV „ ^ s o H s K-K K.

, ,T PQC P C ,l/2_ Bs x s , ) Cs + - j ^ — + - j ^ — )

s = '»-oS • '^,'"' B

P X ^

^3 ^ ^ ^

Substituting for C_ and C in 5.

164

k^P.^L-k2L(P3P^)^/^/(K2K3K^)^

P P P P , , ^ „ / B X , 1 / 2 B X,

k^L(P^ - C P B P ^ ) ' ^ V I K 2 K 3 K ^ ) ^

(1 + K„P +(!B!2i ) 1 / 2 , ! B , ! X 3 H2 '^2S'^4 S 4

k(PT,- ( P B P , ) ' ^ H ^

^'^W^^h'f^'/^^VH^h".^

165

APPENDIX 8

F-TEST ON THE VARIANCES OF THE MODELS

The residual sum of squares, SSQR, of the models are given in

Tables5-6 and 5-7. The'j3ure-error"sum of squares, which is a measure

of the experimental error, was calculated from those measurements

which were replicated at a certain condition as follows (263):

k m r . . - r. „ SSQR = E l [-^^ ) ^ p.e. . , . ., r. . ^ 1=1 3=1 i,j

where m=number of replications at one setting, r. . is one replication,

r. is the average of the replications (r. .) at one setting and k is 1 i> J

the number of settings. The pure error degrees of freedom was calculated from:

k n = T. (m.-l) p.e. i=i 1

The following values were computed:

SSQR = 0.1899 ^ p.e.

n =33 p.e.

s^ = 0.1899/33 = 0.0058 p.e. s =7.6% p.e.

The experimental error is, therefore, 7.6%. The sum of squares lack of

fit and the corresponding degrees of freedom were calculated as follows:

SSQR = SSQR - SSQR

n, J. = n - p - n Lof ^ p.e.

where n = number of experimental points used in the regression

analysis = 112 in this study.

p = number of parameters in each model.

166

The variance for lack of fit and the pure error variance were

calculated:

s = SSQR, Jn, ^ i of ^ Lof Lof

s^ = SSQR /n = 0.0058 p.e. ^ p.e. p.e.

The experimental F value was calculated:

F - 2 2 e ' Lof'^p.e.

A critical F value was calculated from an F-distribution table (263,

264), using n, ^ and n . The F (n, , n , 0.95) for all models '' ^ Lof p.e. ^ Lof p.e.' '

was 1.70. Table 5-8 contains the results of the F-test for the models

already shown in Tables 5-6 and 5-7. Since Fe>F for all the models, it

must be concluded that none of them adequately fite the data at the

95% confidence level.

In order to investigate whether any of the models describes the

data better than the others, F was calculated as follows: s

2 2 F = sf ^ ./sf ^ s Lof,i Lof,m

where s, - . = variance for lack of fit for the i model, with Lof,i '

n-p.-n degrees of freedom. ^1 p.e. ^

n= number of experimental points = 112

p. = number of parameters in the i model

n is as given above p.e. ^ 2 2 s, ^ = variance for lack of fit for the model with the smallest s. ^ Lof,m Lor

the model has n-p -n degrees of freedom *m p.e. ^

From tables, F = F(n. _ ., n, ^ , 0.95) was computed. Model 27 was se Lot,i Lot,m ^

chosen as the model with the smallest s, - and F was 1.50 for all Lof se

models. The results are also shown in Table 5-8. Since F < F for s se

models 27 and 28, it is concluded that these models fit the data

significantly better than the others.

167

APPENDIX 9

FORMATION OF TRIMETHYLBENZENE

The r e a c t i o n s by which t r imethylbenzene can be formed from a feed

c o n s i s t i n g of t o luene and xylenes a r e :

2X ? T + TMB 1

T+X t B + TMB 2

The y ie ld of t r ime thy lbenzenes by the two r e a c t i o n s i s given by:

_ 2TMB 1 X

o

TMB ' 2 X

o

where TMB is the moles trimethylbenzene in the product and X is the

moles xylene in the feed.

Let SUM=total moles of aromatics in the reaction mixture. Then SUM =

B+T+X+TMB, where B,T and X denote the moles of benzene, toluene and

xylenes in the reaction product.

Let n = the mole fraction of xylenes in the feed.

Then X = n.SUM. o When the feed consists of a 1:1 mixture of benzene and xylenes, SUM =

= B +X = B+T+X+TMB, n= 1/2 and 0 0 ' '

B+T+X+TMB o 2

Since only reaction 1 above can occur with this mixture,

4TMB B+T+X+TMB

, where Y is the yield.

When the feed consists of a 1:1:1 mixture of benzene, toluene and

xylenes, SUM = B +T +X = B+T+X+TMB, n = 1/3 and X = B-'T+X+TMB ' ' o o o ' o 3

If reaction 1 above occurs to the exclusion of reaction 2 with this

mixture, then Y _ 6TMB

B+T+X+TMB

168

T4r .-I, • .. ^-u ^ 3TMB

If the reverse is true, then Y = _, _ „ .„•-D+1 + A+ IMrS

Now, assume that xylene disproportionation (reaction I) is the only

reaction which occurs in both mixtures. In other words, that toluene

has no effect on the reaction of xylene in the second mixture. This

hypothesis was tested by passing the first mixture (2.6 yl/s consisting

of 1:1 mole ratio of benzene and m-xylene) over the same catalyst

used in the kinetic experiments (chapter 5) at 4 and 6 ata total

pressure, and then the second mixture (3.9IJ1/S consisting of 1:1:1

mole ratio of benzene, toluene and m-xylene) at 6 and 9 ata. The

difference in pressure ensured the same m-xylene partial pressure, P ,

in both experiments. The remaining process conditions were the same in

both cases:

Temperature :450 C

H_/aromatics ratio :16.7 mol/mol

W/F , :352 g.hr/mol

m-xylene ^

If the lypothesis postulated above is valid, since the partial pressure

and space time of m-xylene is the same in both experiments, the yield

and the reaction rate determined in both cases should be the same.

The results shown in Table 5-11 support the hypothesis that xylene

disproportionation occurs to the exclusion of transalkylation between

toluene and m-xylene. The yields and rates shown in Table 5-11 were

corrected for catalyst deactivation.

169

REFERENCES

1. Hammett, L.P.,A.J.Deyrup, J.Am.Chem.Soc.,54,2721(1932)

Hammett, L.P. Chem.Rev.,16,67(1935)

Hammett, L.P.,Physical Organic Chemistry,Ch.9, Mc Graw-Hill, New

York, 1940.

2. Tanabe, K., Solid Acids and Bases, Academic Press, London, 1970.

3. Thomas, C.L.,Ind.Eng.Chem.Anal.,41,2564(1949).

4. Greenall, A., Ind.Eng.Chem.Anal.,41, 1485(1949).

5. Webb, A.N.,Ind.Eng.Chem., 49, 261(1957).

6. Tamele, M.W.,Disc.Faraday Soc.8,270(1950) .

7. Fiorni,L., Catalysis Reviews, 8,65(1973).

8. Johnson,0.,J.Phys.Chem.,59, 827(1955).

9. Benesi, H.A.,J.Am.Chem.Soc.,78, 5490(1956).

10. Benesi, H.A.,J.Phys.Chem.,61, 970(1957).

11. Richardson,R.L.,S.W.Benson, J.Physical Chemistry, 61, 405(1957).

12. Stone, R.L.,H.F.Rase, Anal.Chem., 29, 1273(1957).

13. Tezuka, Y.T.Takeuchi , Bull .Chem.Soc.Japan, 38, 485(1965) .

14. P ines , H.,W.O.Haag,J.Am.Chem.Soc.,82,2471(1960)

15. Parry, E.P.,J.Catalysis, 2, 371(1963).

16. Webb, A.N.,Ind.Eng.Chem.Anal.,49,261(1957).

17. Clark,A., V.C.F.Holm,and D.M.Balckburn, J.Catalysis, 1, 244(1962).

18. Clark,A., V.C.F.Holm, ibid, 2, 16, 21(1963).

19. Murakama,Y.,T.Shiba,Actes Congr.Intern.Catalyse, 2e,Paris III,No.

129(1960).

20. Nicholson, Nature, 186,630(1960).

21. Fripiat, J.J.,A.Leonard and J.B.Uytterhoeven, J.Phys.Chem.69,3274

(1965).

22. Hsieh, P.Y., J.Catalysis, 2, 211(1963).

23. Benson,J.E.,K.Ushiba and M.Boudart,J.Catalysis, 9, 91(1967).

24. Stone,F.S., L.Whalley, J.Catalysis, 8, 173(1967).

25. Peri, J.B., J.Phys.Chem.69, 231(1965).

26. Zettlemoyer,A.C., J.J.Chessick, J.Phys.Chem., 64, 1131(1960).

27. Mills, G.A., E.R.Boedecker,and A.G.Oblad,J.Am.Chem.Soc.72,1554(1950).

170

28. Milliken, T.H., G.A.Mills and A.G.Oblad, Discussions Faraday Soc.

8, 279(1950).

29. Schwab,G.M., H.Kral, Proc.Third Int.Congr.Catalysis,p.433, North-

Holland, Amsterdam (1965).

30. McIver,D.S., H.H.Tobin and R.T.Barth,J.Catalysis, 2, 485(1963).

31. Weil-Malherbe, H., Weiss,J., J.Chem.Soc.2164(1948).

32. Walling,C, J.Am.Chem.Soc.72, 1164(1950).

33. Hirschler, A.E., J.Catalysis, 2, 428(1963).

34. Leftin, H.P.,Hobson, M.C., Advances in Catalysis, Vol.14, p.115,

Academic Press London (1963).

35. Terenia, A.N., ibid, vol.15,p.227 (1964).

36. Mapes, J.E., Eischens, R.R.,J.Phys.Chem.,58, 809(1954).

37. Basila, M.R., T.R.Kantnes, ibid 71, 467(1967).

38. Stanislaus,A., M.J.B.Evans and R.F.Mann, Can.J.Chem.Eng.,51,725

(1973).

39. Hayward D.O., B.M.W.Trapnell, "Chemisorption", 2nded., Butterworths,

London (1964).

40. Weast, R.C., "Handbook of Chemistry and Physics, 52nded., The

Chemical Rubber Co., Cleveland (1971).

41. Kobe, K.A.,R.H.Harrison, Petroleum Refiner, 33, 11, 161(1954).

42. Gregg, S.J., "The Surface Chemistry of Solids", Chapman and Hall,

London (1961).

43. Kemball, C , "Advances in Catalysis", 2, 233(1950).

44. Gregg, S.J., K.S.W.Sing, "Adsorption, Surface Area and Porosity,

Academic Press, London (1967).

45. Brunauer, S., P.H.Emmett and E.Teller, J.Amer.Chem.Soc.,60, 309

(1938).

46. Langmuir, I., J.Amer.Chem.Soc.,38,2219(1916).

47. Langmuir, I., J.Amer.Chem.Soc., 37, 1139(1915).

48. Langmuir, I., J.Amer.Chem.Soc., 40, 1361(1918).

49. Langmuir, I Trans.Faraday Soc., 17, 621(1921).

50. Thomson, S.j., G.Webb, "Heterogeneous Catalysis", Oliver and Boyd

London (1968).

51. Freundlich,H.,"Kapillarchemie",Akad.Verlag M-B.H,Leipzig (1909).

171

52. Freundlich, H.,"Colloid and Capillary Chemistry", Methuen, London

(1926).

53. Davies, J.T., E.k.Rideal, "Interfacial Phenomena", Academic Press,

London (1963).

54. Temkin, J. J.Phys.Chem.Moscow, 15, 296(1941).

55. Fumkin, A., A.Slygin, Acta Physico chim.U.S.S.R., 3, 791(1935).

56. Brunauer, S., K.S.Love and R.G.Keenan, J.Amer.Chem.Soc.,64,751

(1942).

57. Brunauer, S., P.H.Emmett and E.Teller, J.Amer.Chem.Soc.,60,309

(1938).

58. Ries, H.E., "Catalysis", Vol.Eng., P.H.Emmett ed., Reinhold,

New York (1954) .

59. Emmett, P.H., "Catalysis"Vol.I, Reinhold, New York (1954).

60. Laidler, K.J.,"Catalysis" Vol.1.

61. Ciapetta, E.G., C.J.Plank, in "Catalysis" Vol.-I, Reinhold,

New York (1954).

62. Innes, W.B., in "Catalysis" Vol.1, Reinhold, New York (1954).

63. Beeck, 0., A.E.Smith and A.Wheeler, Proc.Roy.Soc-, 178A, 62(1940).

64. Rhodin, T.N., J.Phys.Chem., 57, 143(1953).

65. Kington, G.L., Trans.Faraday Soc, 49, 417(1953).

66. Balandin, A.A., "Scientific Selection of Catalysts", Monson,

Jerusalem (1968).

67. Balandin, A.A., E.I.Klabunovskii, and A.A.Tolstopyatova, Proc.IV

International Congress on Catalysis, Paper 41, 555(Moscow 1968).

68. Hall, H.J., Ind.Eng.Chem., 62, 3, 33(1970).

69. Roginskii, S.Z., Proc.IV Int.Congr.Catalysis, Lecture 2, 16

(Moscow 1968).

70. Anschiitz R. and H. Immendorff, Ber.der Deut.Chem. Gesell.l7, 2,

2816(1884).

71. Anschiitz R. and H. Immendorff, Ber.der Deut.Chem. Gesell, 18, 1,

657(1885).

72. Anschiitz R., Annalen der Chemie, 235, 150(1886).

73. Nightingale, D.V., Chem.Revs., 25, 329(1939).

172

74. Brown, H.C., C.R.Smoot, J.Am.Chem. Soc, 78, 2176(1956).

75. Pitzer, K.S., D.W.Scott, J.Am.Chem.Soc.,65, 803(1943).

76. Schriesheim, A., J.Org.Chem., 26, 3530(1961).

77. Hastings, S.H., D.E.Nicholson, J.Chem.Eng.Data, 6, 1(1961).

78. Lien, A.P., D.A.McCaulay, J.Am.Chem.Soc. 75, 2407(1953).

79. McCaulay, D.A., A.P.Lien, J.Am.Chem.Soc. 74, 6246(1952).

80. Natanson, and Kagan, J.Phys.Chem.U.S.S.R., 17, 381(1943).

81. Greensfelder, B.S., H.H.Voge, and G.M.Good, Ind.Eng.Chem., 37, 12,

1168(1945).

82. Given, P.H. and D.L.Hammick, J.Chem.Soc, 1779(1949).

83. Izumi, Y., and T.Shiba, Bull.Chem.SocJapan, 37(12), 1797(1964).

84. Iwamura, T., S.Otani, and M.Sato, Bull. Japan Petr.Inst.13(1),

116(1971).

85. Ai, M., E.Echigoya, and A.Ozaki, Bull.Japan Petr.Inst., 7, 46(1965).

86. Echigoya, E., Masai M., S.Saito, K.Murata, T.Nakamura and

K.Morikawa, KagaKu Kogaku, 31(4), 386(1967) (Abstract in:J.Chem.

Eng.Japan, 1(1), 96(1968).

87. H.Matsumoto, and Y.Morita, Bull.Japan Petr.Inst.10,8(1968).

88. Benesi, H.A., J.Catalysis, 8, 368(1967).

89. Venuto, P.B., L.A.Hamilton, P.S.Landis and J.J.Wise, J.Catalysis,

5, 81(1966).

90. Yashima, T., H.Moslehi and N.Hara, Bull.Japan Petr.Inst., 12, 106

(1970).

91. Jacobs, P.A., H.E.Leeman and J.B.Uytterhoeven, J.Catalysis, 33,31

(1974).

92. Toyo Rayon Kabushiki Kaisha, Tokyo, Dutch Patent No.6706218(1967).

93. Toyo Rayon Kabushiki Kaisha, Tokyo, Dutch Patent no.6817615(1968).

94. Otani, S., T.Iwamura, S.Hayashi, D.Ogawa, and M.Kanaoka, U.S.

Patent No.3, 597, 492 (1971).

95. E.Baud, Ann.Chem.Phys.,(8), 1, 60(1904). In French.

96. Roginskii, S.Z., Actes du 2e Congres Intem.de Cat., Vol.2, 1527

(Paris, 1960).

97. Ehret, W.F.and F.J.Frere, J.Am.Chem.Soc.,67, 64(1945).

98. Cowley, J.M., and T.R.Scott, J.Am.Chem.Soc.,70, 105^948).

173

99. Johnson, R.L. and B.Siegel, Nature, 210, 1256(1966).

100. Toyo Rayon Kabushiki Kaisha, Tokyo, Dutch Patent No.6817616(1968).

101. Mills, W., U.S.Patent No.582,938(1896).

102. Mills, W., British Patent No.20377(1895).

103. Mills, W., German Patent No.94849(1896).

104. Gmelin, Handbuch der Anorganischen Chemie, Teil B, 35, 157,

Verlag Chemie, Berlin (1934).

105. Kochloefl, K., M.Kraus and V.Bazant, Proc.IV Int. Congr.

Catalysis, Paper 85, 490(Moscow, 1968).

106. Kraus, M., Adv.Catalysis, 17, 76(1967).

107. Boreskov, G.K., Actes du 2e Congres Intern, de Cat., Vol.1, 163

(Paris, 1960).

108. Moerkerken, A., B.Behr, M.A.Noordeloos-Maas, and C.Boelhouwer,

J.Catalysis, 24, 177(1972).

109. Reitsma, H.J. and C.Boelhouwer, J.Catalysis, 33,39(1974).

110. Uytterhoeven, J.B., L.G.Christner, and W.K.Hall, J.Phys.Chem,69,

2117(1965).

111. Jacobs,P.A., H.E.Leeman, J.B.Uytterhoeven, J.Catalysis, 33, 17

(1974).

112. Moscou, L. and M.Lakeman, J.Catalysis, 16,173(1970).

113. Benesi, H.A., J.Catalysis, 173(1973).

114. Csicsery, S.M., J.Catalysis, 19, 394(1970).

115. Papadatos, K. and K.A.Shelstad, J.Catalysis, 28, 116(1973).

116. Moscow, L. and R.Mone, J.Catalysis, 30, 417(1973).

117. Ward, J.W. and R.C.Hansford, J.Catalysis, 13, 154(1969).

118. Wang, K.M. and J.H.Lansford, J.Catalysis, 24, 262(1972).

119. Lunsford, J.H., J.Phys.Chem., 72, 4163(1968).

120. Ward, J.W., J.Catalysis, 17,355(1970).

121. Gutberlet, L.C. and R.J.Bertolacini, U.S.Patent 3, 548,020(1970).

122. Voge, H.H., in "Catalysis", Vol.6, P.H.Emmett, ed., p.407,

Reinhold, New York (1958).

123. Loewenstein, W., Amer.Mineral., 39, 92(1954).

124. Breck, D.W., "Zeolite Molecular Sieves", Wiley, New York (1974).

174

126. Hersh, C.K., "Molecular Sieves", Reinhold, New York (1961).

127. Barrer, R.M. in "Molecular Sieves", Soc.Chem.Ind., London (1967).

128. Flanigan, E.M. and L.B.Sand, "Molecular Sieve Zeolites 1 and II",

Adv.Chem.Ser.Vols.101 and 102, Amer.Chem.Soc,Washington (1971).

129. Meier, W.M. and J.B.Uytterhoeven, eds. "Molecular Sieves", Adv.

Chem.Ser.Vol.121, Amer.Chem.Soc., Washington (197

130. Barrer, R.M., F.W.Bultitude and J.W.Sutherland, Trans.Faraday Soc.

53, 1111(1957).

131. Barrer, R.M., Endeavour, 23, No.90, 122(1964).

132. Collins, J.J., Chem.Eng.Progr.,64, No.8,66(1968).

134. Grubner, 0., P.Jiru and M.Ralek, "Molecular siebe", VEB Deutscher

Verlag der Wissenschaften, Berlin (1968).

135. Barrer, R.M., in "Non-stoicheiometric Compounds", L.Mandelcorn

(ed.). Academic Press, London (1964).

136. Eitel, W., "Silicate Science", Vols.l and IV, Academic Press,

London (1966).

137. Papp, J., D.Kallo and G.Schay, J.Catalysis, 23, 168(1971).

138. Karnaukhov, A.P., in '.'Pore structure and Properties of Materials",

P.A-3,S.Modry ed., Academia, Prague (1973).

139. Dubinin, M.M., Chem,Reviews, 60, 235(1960).

140. De Boer, J.H., in "The Structure and Properties of Porous Materials"

p.68, D.H.Everett, and F.S.Stone, eds., Butterworths, London

(1958).

141. Overbeek, J.Th.G., in "Surface Area Determination, P.3, D.H.

Everett and R.H.Ottewill, eds., Butterworths, London (1970).

142. Barrer, R.M., in "The Structure and Properties of Porous Materials"

p.6, D.H.Everett and F.S.Stone, eds. Butterworths, London (1958).

143. Broekhoff, J.C,P. and B.G.Linsen in "Physical and Chemical Aspects

of Adsorbents and Catalysts',' P.l, B.G.Linsen ed.. Academic Press,

London (1970).

144. Lippens, B.C. and J.H.de Boer, J.Catalysis, 4, 319(1965).

145. Sing, K.S.W. in "Surface Area Determination", P.25, D.H.Everett and

R.H.Ottewill, eds., Butterworths, London (1970).

175

146. Broekhoff, J.C.P.de Boer, J.H., in "Surface Area Determination",

P.97, D.H.Everett and R.H.Ottewill, eds, Butterworths, London

(1970).

147. Dubinin, M.M., Adv.Colloid Interface Sci, 2, 217, (1968).

148. Dubinin, M.M., in "Pore Structure and Properties of Materials",

p.C-27, S.Modry ed., Academia, Prague (1973).

149. Brunauer, S., Skalny, J., and I.Odler, in "Pore Structure and

Properties of Materials", p.c-3, S.Modry ed., Academia, Prague

(1973).

150. Gurvitsch, L., J.Phys.Chem.SocRuss., 47, 805(1915).

151. Steele, W.A., Advan.Colloid. Interface Sci.,1,3(1967).

152. Barrer, R.M. and R.M.Gibbons, Trans.Faraday Soc.,59,2569(1963).

153. Linsen, B.C. and A.van den Heuvel, in "The Solidgas Interface",

Vol.12, P., E.A. Flood ed.,Marcel Dekker, New York (1967).

154. Scholten, J.J.F., in "Porous Carbon Solids", P.225, R.L.Bond

ed. Academic Press, London (1967).

155. Broekhoff, J.C.P., Thesis, Delft (1969).

156. De Wit, L.A. and J.J.F.Scholten, J.Catalysis, 36,30(1975).

157. De Wit , L.A. and J.J.F.Scholten, J.Catalysis, 36, 36(1975).

158. Lippens, B.C., B.G.Linsen and J.H.de Boer, J. Catalysis, 3, 32

(1964).

159. De Boer, J.H., B.C.Lippens, B.G.Linsen, J.C.P., Broekhoff, A.v.d.

Heuvel and Th.J.Osinga, J.Colloid Interface Sci, 21,405(1966).

160. Linsen, B.G., Thesis, Delft (1964).

161. Rootare, H.M., and C.F.Prenzlow, J.Phys.Chem. 71, 2733(1967).

162. Zwietering, P., in "The Structure and Properties of Porous Solids"

D.H.Everett and F.S.Stone eds, P.287, Butterworths, London (1958).

163. Kaganer, M.G., Zh.fiz.Khim., 33, 2202(1959).

164. Mikhail, R.S., S.Brunauer and E.E.Bodor, J.Colloid Interface Sci.,

26, 45 (1968).

165. Brunauer, S., in "Surface Area Determination, P.63, D.H.Everett

and R.H.Ottewill, eds., Butterworths, London (1970).

166. Marsh, H., and B.Rand, J.Colloid Interface Sci.33, 478(1970).

167. Marsh, H., and B.Rand, J.Colloid Interface Sci,40,121(1972).

176

168. Kirk-Othmer, "Encyclopedia of Chemical Technology", Vol.3,

Inter-Science, New York (1964).

169. Anon., Oil and Gas Intern., Vol.9, no.10, 60(1969).

170. Anon., Hydrocarbon Processing, Novo, 155(1969).

171. Otani, S., Chem.Eng.,July 27, 118(1970).

172. Grandio, P., F.H.Schneider, A.B.Schwartz, and J.J.Wise, Oil and

Gas Journ. Nov,29, 62(1971).

173. Grandio, P., F.H.Schneider, A.B.Schwartz, and J.J.Wise,

Hydrocarb.Proc.,Vol.5I, No. 8, 85(1972).

174. O.E.C.D. The Chemical Industry, 1970-1973.

175. Ogawa, D., S.Hayashi, K.Marsumura, T.Iwamura, M.Sato, and

S.Otani, Kogyo Kagaku Zasshi, 72, 2165(1969).

176. Hastings, S.H. and D.H.Nicholson, J.Chem.Eng.Data, 6(1), 1(1961).

177. Egan, C.J., J. Chem.Eng.Data, 5, 298(1960).

178. Hedden, K., Chem.Ing.Technik, 34, 140(1962).

179. Stull, D.R., E.F.Westrum and G.C.Sinke, "The Chemical

Thermodynamics of Organic Compounds", Wiley, N.Y.(1968).

180. Kandiner, H.J. and S.R.Brinkley, Ind.Eng.Chem.42,5, 850(1950).

181. Breck, D.W., "Zeolite Molecular Sieves", Wiley, New York (1974).

182. Otouma, H., Y.Arai, and H.Ukihashi, Bull.Chem.SocJapan, 42,

2449(1969).

183. Kirk-Othmer, "Encyclopedia of Chemical Technology", Vol.15 p.188,

Inter-Science, New York (1956).

184. Kirk-Othmer, Encyclopedia of Chemical Technology", Vol.8, p.509,

Wiley, New York (1965).

185. British Petroleum Co., Belgian Patent 626, 446.

186. ESSO Oil Corp., U.S.Patent 3, 126, 422.

187. Houdry Process Corp., U.S.Patent 2, 795, 629.

188. Universal Oil Products, German Patent 1,925,102

189. Universal Oil Products, German Patent 1,946,187

190. Universal Oil Products U.S.Patent 3,417,157

191. Texaco Inc.Dutch Patent 6,903,210

192. Toyo Rayon Kabushiki Kaishi, Dutch Patent 6,817,615

193. California Research Corp.U.S.Patent 3,182,095

177

194. Sun Oil Co., U.S.Patent 2,739,993.

195. Standard Oil Indiana, U.S.Patent 3,009,004.

196. Toyo Rayon Kabushiki Kaishi, Dutch Patent 6,807,775.

197. Toyo Rayon Kabushiki Kaishi, Dutch Patent 6,706,218.

198. Shell Oil Co., U.S.Patent 3,281,483.

199. Standard Oil Indians, U.S.Patent 3,548,020.

200. Ashland Oil and Refining Co., U.S.Patent 3,607,961.

201. Ashland Oil and Refining Co., U.S.Patent 3,825,613.

202. Texaco Inc.,U.S.Patent 3,812,197.

203. Ashland Oil and Refining Co.,U.S.Patent 3,597,491.

204. Thomas, J.M. and W.J.Thomas, "Introduction to the Principles of

Heterogeneous Catalysis", Academic Press, London (1967).

205. Peterson, E.E., "Chemical Reaction Analysis", Prentice-Hall,

Engelwood Cliffs, N.J.(1965).

206. Kramers, H.and K.R.Westerterp, "Elements of Chemical Reactor

Design and Operation", Netherlands University Press, Amsterdam

(1963).

207. Aris, R., "Elementary Chemical Reactor Analysis", Prentice-Hall,

Engelwoods Cliffs, N.J.(1969).

208. Denbigh , K.G., J.C.R.Turner, "Chemical Reactor Theory", Cambridge

(1971), p.30.

209. Smith, J.M., "Chemical Engineering Kinetics", McGraw-Hill,

New York (1970) p.274.

210. Levenspiel, 0., "Chemical Reaction Engineering", Wiley, New York'

(1972).

211. Wei, J., Ind.Eng.Chem., 58,9,38(1966).

212. KittreU J.R., Advances in Chemistry Serie^ 97(1970).

213. Mizaki, R., Kittrell, J.R., lEC, Vol.59, No.5, 63(1967).

214. Paynter, J.D., Haskins, D.E., CES, Vol.35, 1415(1970)

215. Hicks, R.E., Ind.Eng.Chem.Fundam.Vol.9, no.3, 500(1970).

216. Carberry, J.J, lEC, Vol.56, No.11, 39(1964).

217. Smith, J.M., Chem.Eng.Progress, Vol.64, no.8, 78(1968)

178

218. Perkins, T.K., H.F.Rass, A.I.Ch.E.J., Vol.4, no.3, 351(1958).

219. Kittrell, J.R., R.Mezaki, C.C.Watson, lEC, Vol.58, No.5 51(1966).

220. Kittrell, J.R., R.Mezaki, lEC, Vol.59, No.2, 28(1967).

221. Laidler, K.J., "Chemical Kinetics", 2nded., McGraw-Hill,

New York, 1965.

222. Frost, A.A., R.G.Pearson, "Kinetics and Mechanism", Wiley,

New York, 1961.

223. Kropheller, H.W., Spikins, D.J., and Wildman, T. Brit.Chem.Eng.

10(2), 109(1965).

224. Hurt, D.M., Ind.Eng.Chem. 35, 522(1943).

225. Caddell, J.R. and D.M.Hurt, Chem.Eng.Progr., 47,333(1951).

226. Black, J.H., J.H.Wright and J.Coull, A.I.Ch.E.J., Vol.2, No.4,

572(1956).

227. Hall, J.W. and H.F.Rase, Ind.Eng.Chem.Fund., 3, 158(1964).

228. Harrison, D.P., J.W.Hall and H.F.Rase, Ind.Eng.Chem., 57, 18(1965).

229. Kokes, R.J., H.Tobin Jr. and P.H.Emmett, J.Am.Chem.Soc.77,5860(1955).

230. Hall,.W.K. and P.H.Emmett, J.Am.Chem.Soc, 79, 2091(1957).

231. Magee, E.M., Ind.Eng.Chem.Fund.,2, 32(1963).

232. Massaldi, and Maymo, J.Catalysis, 14, 61(1969).

233. Herwijnen, T.van, H.van Doesburg and W.A.de Jong, J.Catalysis,28,

391(1973).

234. Herwijnen, T.van, "On the Kinetics and mechanism of the CO-shift

coiversion on a copper/zincoxide catalyst". Dissertation, Delft

(1973).

235. Schwartz, C.E.and J.M.Smith, Ind.Eng.Chem.,45, 1209(1953).

236. Edwards, M.F.and J.F.Richardson, Chem.Eng.Sci., 23, 109(1968).

237. Satterfield, C.N.and T.K.Sherwood, "The Role of Diffusion in

Catalysis", Addison-Wesley, Reading, Mass.(1963).

238. Satterfield, C.N., "Mass Transfer in Heterogeneous Catalysis",

M.I.T. Press, Cambridge, Mass.(1970).

239. Chambert, R.P.and M.Boudart, J.Catalysis, 6, 141(1966).

240. Weisz, P.B., Z.Phys.Chem.,11, 1(1957).

179

241. Hudgins, R.R., Chem.Eng.Sci., 23, 94(1968).

242. Perry, J.H., "Chemical Engineers Handbook", 4. th ed., McGraw-Hill,

New York (1963).

243. Coulson,J.M. and J.F.Richardson, "Chemical Engineering", Vol.2,

Pergamon Press, London (1968).

244. Ergun, S., Chem.Eng.Progr., 48, 89(1952).

245. Ergun, S. Chem.Eng.Prog., 48, 227(1952).

246. Beek, J., in "Advances in Chemical Engineering", vol.Ill, Academic

Press, 203 (1962).

247. Hougen, O.A. and K.M.Watson, "Chemical Process Principles", John

Wiley, New York (1962).

248. Downes, H.R., "The Chemistry of Living Cells", 2nded., Longmans,

London (1962).

249. IVhite, A., P.Handler, E.L.Smith and D.Stetten, "Principles of

Biochemistry", 2nded., McGraw-Hill, New York (1959).

250. Hougen, O.A.and K.M.Watson, Ind.Eng.Chem. 35(5), 529(1943).

251. Yang, K.H. and O.A.Hougen, Chem.Eng.Prog.46, 146(1950).

252. Hinshelwood, C.N., "Kinetics of Chemical Change", Oxford

University Press, London (1940).

253. Weller, S., A.I.Ch.E.J., 2, 59(1956).

254. Boudart, M., A.I.Ch.E.J., 2, 62(1956).

255. Kittrell, J.R., W.G.Hunter and C.C.Watson, A.I.Ch..E.J.11(6),

1051(1965).

256. Powell, M.J.D., Computer J. 3, 175(1960).

257. Powell, M.J.D., Computer J., 6, 155(1964).

258. Valstar, J.M.,"A study of the fixed bed reactor with application

to the synthesis of vinyl acetate", p.149, Dissertation, Delft

(1969).

259. Otani, S., T.Iwamura, S.Hayashiand D.Ogawa, U.S.Patent No.3, 597,

490(1971).

260. Kruijer, S., "Entropie en beweeglijkheid bij adsorptie op actieve

kool". Dissertation, Delft (1955).

261. Filers, J., Internal Report, Delft (1975).

262. Gerritsen, L.A., Internal Report, Delft (1975).

180

263. Draper, N.R.and H.Smith, "Applied Regression Analysis", Wiley,

New York (1966).

264. Perry, R.H. and C.H.Chilton, "Chemical Engineers Handbook", 5th

ed., McGraw-Hill, New York (1973).

265. Betcke, R.A., Internal Report, Delft (1973).

266. Chu, C.J., J.M.Kalil and W.A.Wetteroth, Chem.Eng.Progr.,49, 141

(1953).

267. Aneke, L.E. P.J.van den Berg and W.A.de Jong, Paper for the

6th Int.Congr.Catalysis, London (1976).

268. Aneke, L.E. and T.van Herwijnen, "A method for solving material

and heat balances of complex chemical processed'. Delft (1971) .

269. Aneke, L.E. "Preliminary Process Design of the disproportionation

of Toluene", Delft (1970).

181

SAMENVATTING

Disproportionering is een potentiele alternatieve methode voor het

benutten van een overschot aan tolueen afkomstig van de fabricage van

aromatische koolwaterstoffen. Alhoewel de reaktie zowel in een vloei-

bare als in een gasfase bedreven kan worden,is de laatstgenoemde

commercieel gezien een interessanter proces. Het vindt plaats in de

aanwezigheid van vaste zure katalysatoren. Het meeste hier vermelde

werk heeft betrekking op de bereiding en karakterisering van een

katalysator met een niveau van aktiviteit, selektiviteit en stabili-

teit dat voor een commercieel proces alsmede voor een studie van de

reaktie kinetiek vereist is.

Ten eerste zijn de bereiding van zo een katalysator, aangeduid als

ABl en met een samenstelling 72% HY/18% B-ALF2/10% Cu, en het effekt

van enige proces condities op de aktiviteit, selektiviteit en stabili-

teit beschreven. De vermelde resultaten laten zien dat de katalysator

een redelijke tolueen disproportioneringsprestatie bezit en tonen aan

dat 500 C de optimale aktiveringstemperatuur is:de aktiviteit ver-

dwijnt vrijwel als er een hogere temperatuur gebruikt wordt. Daarnaast

zijn de fysische en chemische eigenschappen bepaald door stikstof-

adsorptie, kwikpenetratie porosimetrie en zuurgraad bepalingen.

De resultaten van het textuur onderzoek bevestigen dat de meeste

konventionele methoden voor poreuze stoffen niet van toepassing zijn

voor zeolieten en zeoliet bevattende katalysatoren. Dienovereenkomstig

wordt er een nieuwe methode voorgesteld en gebruikt om waarden van de

oppervlaktes van zeolitische microporieen te verkrijgen die in over-

eenstemming zijn met de waarden berekend uit kristallografische ge-

gevens. De kwikpenetratie-experimenten geven een maat voor de ruimtes

tussen de katalysatordeeltjes en die van de porien behorende tot het

in de katalysator aanwezige B-aluminium fluoride en koper. De resul­

taten van de zuurgraad bepalingen, gemeten door n-butylamine titratie

en ammoniak adsorptie, zijn gekombineerd met die van de textuur van de

katalysator om aan te tonen dat slechts 10% van de totale oppervlakte

182

bestaat uit zure plaatsen. De resultaten van textuur- en aktiviteits-

metingen suggereren dat de tolueen disproportioneringsaktiviteit van

katalysator ABl is gelokaliseerd in de overgangsporien en dat de micro-

porien slechts dienen om de zware reaktieprodukten te verzamelen die

anders tot deaktivering zouden leiden. De resultaten van ammoniak

adsorptie gekombineerd met de invloed van de aktiveringstemperatuur

op de katalytische aktiviteit suggereren dat Bronsted zure plaatsen

verantwoordelijk zijn voor de tolueen disproportioneringsaktiviteit.

Reaktiesnelheidsmodellen, afgeleid aan de hand van een aantal

mechanismen bestaande uit eenvoudige adsorptie, oppervlaktereaktie en

adsorptiestappen worden gebruikt om de kinetische gegevens te be-

schrijven. Niet-lineaire regressie techniek wordt toegepast om een

groep van modellen te isoleren met de kleinste varianties. Een

F-test op deze varianties toont aan dat het verschil tussen die van

twee modellen met de kleinste waarden niet significant is op een 95%

betrouwbaarheidsniveau;de twee modellen zijn statistisch gezien gelijk-

waardig. Experimenten met reaktieprodukten tonen aan dat benzeen geen

meetbare invloed heeft op de tolueen disproportioneringssnelheid

terwijl xylenen een zeker vertragend effekt hebben. Het proefschrift

besluit met enige ontwerpbeschouwingen voor de realisatie van een

industrieel disproportioneringsproces.

STELLINGEN

1. De studie door Yashima et al. over de dampfase disproportionerings-

reaktie van tolueen over H-mordeniet is een klassiek voorbeeld van

de invloed van diffusie limitering op kinetiek metingen.

Yashima, T., H.Moslehi and N.Hara, Bull.Japan Petr.Inst., 12,

106(1970).

2. Bij porievolume-gegevens afgeleid uit capillaire adsorptie en con-

densatie kunnen extrapolatie methodes zoals die van Gurvitsch en

van Dubinin tot verkeerde conclusies leiden.

3. Voor een nauwkeurige bepaling van het inwendige porievolume van

poreuze massa's met kwikporosimetrie is het nodig om kwikpenetratie

metingen uit te voeren voor een aantal monsters van varierende

deeltj esgrootte.

4. Bij complexe reakties is het niet altijd mogelijk om analytische

uitdrukkingen van opbrengst en selektiviteit af te leiden.

Dit proefschrift:hoofdstuk 2.

5. De disproportionering van xyleen op de katalysator ABl, die gebruikt

is voor de kinetische studie beschreven in dit proefschrift, ver-

loopt sneller dan de transalkylatie reaktie met tolueen.

Dit proefschrift:appendix 9.

6. Het idee van het optreden van een ternaire zadel azeotroop in het

systeem Ureum-H.O-CO_-NH_ bij chemisch evenwicht moot gezien worden

als het meest eenvoudige model dat de tot nu toe gedane fasemetingen

goed beschrijft.

Lemkowitz, S.M., Dissertatie, Delft (1973).

7. Experimentele resultaten verkregen met industriele katalysatoren

moeten niet worden gepubliceerd in wetenschappelijke tijdschriften

zonder volledige openbaring van de samenstelling van deze katalysatoren.

Dit proefschrift:referenties 84 en 175.

8. Vooruitgang in de biologische wetenschappen zal alleen gelijke tred

houden met die in de chemie en de natuurkunde indien mathematische

analyse op veel groter schaal toegepast wordt op biologische pro-

blemen.

Beck, S.D.,"The simplicity of Science",Pelican (1962).

Bronowski, J."The Ascent of Man", BBC Press, London (1973).

9. Hoewel Langmuir-type modellen nuttig zijn om de snelheden van hete-

rogene katalytische reakties te beschrijven, zijn zij zwak op

fysische en chemische gronden.

10. Politieke onafhankelijkheid van zowel "oosf'als "west", gepaard

gaande met economische zelfbekwaamheid is het enige levensvatbare

antwoord op problemen van de zogenaamde derde wereld.

11. Het zou goed zijn wanneer aanhangers van "aangepaste technieken"

zich beter op de hoogte zouden stellen van de locale ervaring en

behoefte in zich snel ontwikkelende landen.

12. De bekwaamheid van nieuwe werknemers kan verbeterd worden door een

goed gepland orientatieprogramma en een duidelijke specificatie van

hun eventuele plichten.

13. Het ontwerp van de plaat aangebracht op het ruimteschip "Pioneer 10"

waarvan men hoopt dat het onderschept zal worden door een "Intelli-

gente beschaving" elders in het heelal, is gebaseerd op twijfel-

achtige veronderstellingen.

Britannica Book of the Year, 1975, p.632.

STELLINGEN

1. The study by Yashima et al. of the vapour-phase disproportionation

of toluene over H-mordenite is a classic example of the influence

of diffusion limitation on measured kinetics.

Yashima, T., H.Moslehi and N.Hara, Bull. Japan Petr.Inst., 12,

106(1970)

2. For pore volume data derived from capillary adsorption and

condensation measurements, extrapolation methods like those of

Gurvitsch and of Dubinin may lead to misleading conculsions.

3. For accurate determination of the internal pore volume of porous

solids by mercury porosimetry it is necessary to carry out

penetration measurements on a number of samples of different par­

ticle sizes.

4. When complex reactions are involved, it is not always possible

to derive an analytical expression for quantities like yield and

selectivity.

This thesis:section 2.5

5. Xylene disproportionation is faster than transalkylation with

toluene over the catalyst ABl used in the kinetic study reported

in this thesis.

This thesis:Appendix 9

6. The idea of the existence of a ternary saddle azeotrope in the

system Urea-H^O-CO^-NH_ at chemical equilibrium must be seen as

the simplest model which describes the current experimental results

adequately.

Lemkowitz, S.M., Thesis, Delft (1975).

7. Experimental results obtained with industrial catalysts should not

be published in scientific journals without full disclosure of

the nature and composition of such catalysts.

This thesis:references 84 and 175.

8. Progress in the biological sciences will keep pace with that in

chemistry and physics only when mathematical analyses are applied

on a much greater scale to biological problems.

Beck, S.D., "The Simplicity of Science", Pelican (1962).

Bronowski, J."The Ascent of Man", BBC Press, London (1973).

9. Although Langmuir-type models are useful for representing the

reaction rates of heterogeneous catalytic reactions, they are

weak on physical and chemical grounds.

10. Political independence from both east and west accompanied by

economic self-sufficiency is the only viable answer to the

problems of the so-called third world.

11. It would be good for the adherents of "intermediate technology"

to acquaint themselves more with the local experience and needs

in the fast-developing countries.

12. The efficiency of new employees can be improved by a well-planned

orientation program and a clear spcification of their eventual

duties.

13. The design of the plaque on Pioneer 10 space craft which, it is

hoped, will be intercepted by an "intelligent civilization"elsewhere

in the universe is based on some questionable assumptions.

Britannica Book of the Year, 1975, p.632.