IHIUIUIHIilllHlHllHlUlHUllhIIIHIIHMHHIHIWI

1293 01055 9767

This is to certify that the

thesis entitled
Development of a Catalytlc frocess for the
Production of Maleic Anhydride from a Fermentation Feedstock

presented by

Sanjay Krishnamurthy Yedur

has been accepted towards fulfillment
of the requirements for

M. S . degree in Chem. Engr.

 

 

7}?”‘0. W
/ ,, 1

Major professor

Date June 17, 1992

 

0-7639 MS U is an Affirmative Action/Equal Opportunity Institution

 

 

v

i‘y‘iE-‘Eiéggan Stete
Urfiversity

 

 

 

PLACE IN RETURN BOX to remove thie checkout from your record.
TO AVOID FINES return on or before date due.

DATE DUE ' DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

[—TI i

MSU le An Affirmative Action/Bone! Opportunity Institution
chS—oI

 

DEVELOPMENT OF A CATALYTIC PROCESS FOR THE
PRODUCTION OF MALEIC ANHYDRIDE FROM A
FERMENTATION FEEDSTOCK

By
Sanjay Krishnamurthy Yedur

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF SCIENCE

Department of Chemical Engineering
1 992

ABSTRACT

DEVELOPMENT OF A CATALYTIC PROCESS FOR
THE PRODUCTION OF MALEIC ANHYDRIDE
FROM A FERMENTATION FEEDSTOCK

By

Sanjay Krishnamurthy Yedur

To reduce our dependency on non-renewable energy sources, it is necessary to
develop alternate pathways for the production of various chemicals which are traditionally
produced from feedstocks derived from fossil fuels. With this aim in view, an alternate
process has been proposed for the production of maleic anhydride from a fermentation
feedstock. The process utilizes the catalytic oxydehydrogenation of fermentation derived
succinic anhydride to produce maleic anhydride. Various catalysts were synthesized and
tested for the oxydehydrogenation reaction. Iron phosphate based catalysts were found to
be the best on the basis of high conversions and selectivities obtained using these catalysts.
The effect of temperature, oxygen concentration, feed concentration, contact time, and the
total time on-stream on the performance of the catalyst was studied. A set of optimum
conditions for the operation of the reactor was developed. The catalysts were characterized
based on their bulk and surface compositions, their surface areas, and their bulk

crystallographic structure.

This work is dedicated to my parents and to my brother, without whose support this work
would not have been possible.

iii

ACKNOWLEDGEMENTS

I would like to express my deep sense of gratitude to my advisor Dr. Kris Berglund
for his constant encouragement and support throughout the course of this work.

I would also like to drank my friends and colleagues Joel Dulebohn and Todd
Werpy for their useful suggestions and help during this work. ’

Special thanks are due to Kathy Severin for her help in using X—Ray photoelectron
spectroscopy.

iv

TABLE OF CONTENTS

List of Tables ' viii
List of Figures ix

1. Introduction 1
1.1 Scope 1

1.2 Maleic anhydride and its importance 1

1.3 Specifications 4

1.4 Uses 4

1.5 Production 5

1.6 Objectives 8

2. Literature Survey

2.1 Commercial technologies 9
2.1.1 Benzene route 9
2.1.2 Butane route 13
2.1.3 Butylene route 13
2.2 Alternate technology proposed 14
2.3 Characterization of catalysts 19
2.3.1 Bulk properties 20
2.3.2 Particle properties 21
2.3.3 Surface properties 23
2.3.4 Activity 24

3. Materials and Methods
3.1 Preparation of the catalysts
3.1.1 Experimental setup
3.1.2 Catalyst preparation
3.1.3 Commercial catalysts
3.2 Testing of the catalysts
3.2.1 Reactor setup
3.2.2 Product analysis
3.2.3 Experiments
3.3 Characterization of catalysts
3.3.1 Surface area analysis
3.3.2 X-ray diffraction analysis
3.3.3 Bulk composition analysis
3.3.4 Surface composition analysis
4. Results and Discussions
4.1 Testing of commercial catalysts
4.2 Testing of synthetic molybdenum oxide based catalysts
4.3 Testing of synthetic iron phosphate based catalysts
4.3.1 Effect of temperature on conversion
4.3.2 Effect of contact time on conversion
4.3.3 Effect of feed concentration on conversion
4.3.4 Effect of oxygen concentration on conversion
4.3.5 Effect of time on-stream on conversion
4.4 Results of characterization experiments
4.4.1 Surface area measurements

4.4.2 Bulk compositions
vi

25
25
25
25
28
29
29
29
34
35
35
36
37
38

42
43
43
46

48

53

53
55

 

4.4.3 Surface compositions

4.4.4 Bulk crystal structure determination

5. Conclusions

6. Recommendations for future work

7. Appendices

A. Laboratory synthesis procedure for the catalysts

B. X-ray diffraction spectra of the catalysts

8. List of References

57
57

62

68
71

78

 

Table 1.1
Table 1.2
Table 2.1
Table 3.1
Table 3.2
Table 4.1
Table 4.2
Table 4.3
Table 4.4
Table 4.5

LIST OF TABLES

Physical properties of maleic anhydride 2
Breakdown of the end uses of maleic anhydride 5
List of major maleic anhydride producing companies 10
Summary of catalysts synthesized 27
Commercial catalysts tested 28
Performance of commercial catalysts 41
Performance of synthetic molybdenum based catalysts 43
Surface areas of catalysts 55
Bulk molar compositions of the catalysts 56

Surface molar compositions of the catalysts relative to iron content 58

I'l-h‘

Figure 1.1
Figure 1.2
Figure 1.3

Figure 2.1

Figure 2.2
Figure 3.1

Figure 3.2a

Figure 3.2b

Figure 4.1

Figure 4.2

LIST OF FIGURES

Chemical structure of maleic anhydride
Conversion of succinic acid to maleic anhydride
Conversion of isobutyric acid to methacrylic acid

Halcon Design for the manufacture of maleic anhydride from
Benzene or Butane

Idealized mechanism of oxidative dehydrogenation

Reactor setup

NMR spectrum showing 5% succinic acid and 5% maleic
acid standards

NMR spectrum of the reaction products of the oxydehydrogenation
reaction of succinic acid to maleic acid over a supported iron
phosphate catalyst

Effect of temperature on the conversion of succinic acid to
maleic acid using iron phosphate catalysts under the following
conditions: a) contact time = 4 s, b) WHSV = 0.8-0.9

gm feed/gm catalyst/hr, c) food concentration = 40 g/l,

d) oxygen to feed mol ratio = 10:1, and e) time on-strearn

= 30 minutes

Effect of contact time on the conversion of succinic acid to
maleic acid using iron phosphate catalysts under the following
conditions: a) temperature = 475 'C, b) WHSV = 0.8-0.9

gm feed/gm catalyst/hr, c) food concentration = 40 g/l,

d) oxygen to feed mol ratio = 25:1, and e) time on-stream

= 15 minutes

ix

11

15

32

33

45

47

.‘

 

Figure 4.3

Figure 4.4

Figure 4.5

Figure 4.6

Figure 4.7a

Figure 4.7b

Figure B-l

Figure B-2

Figure 8-3

Figure B-4

Effect of feed concentration on the conversion of succinic acid to
maleic acid using iron phosphate catalysts under the following
conditions: a) contact time = 4 s, b) WHSV = 0.8-0.9

gm feed/gm catalyst/hr, 0) temperature = 475 'C.

(1) oxygen to feed mol ratio = 25:1, and e) time on-stream

= 15 minutes i

Effect of oxygen concentration on the conversion of succinic acid
to maleic acid using iron phosphate catalysts under the following
conditions: a) contact time = 4 s, b) WHSV = 0.8-0.9

gm feed/gm catalyst/hr, c) feed concentration = 40 g/l,

and (1) time on-stream = 15 minutes

Effect of time on-stream on the conversion of succinic acid to
maleic acid using iron phosphate catalysts under the following
conditions: a) contact time = 4 s, b) WHSV = 0.8-0.9

gm feed/gm catalyst/hr, c) temperature = 475 ’C,

d) food concentration = 40 g/l, and e) oxygen to feed mol ratio
= 25:1

Typical BET plot for a supported and promoted iron phosphate
catalyst

X-Ray diffraction spectrum of an iron phosphate catalyst, calcined
at 110 'C

X-Ray diffraction spectrum of an iron phosphate catalyst, calcined
at 500 ‘C

X-Ray diffraction spectrum of an iron phosphate catalyst, supported

by LUDOX ® HS 40%; calcined at 110 'C

X-Ray diffraction spectrum of an iron phosphate catalyst, supported

by LUDOX ® HS 40%; calcined at 500 'c

X-Ray diffraction spectrum of an iron phosphate catalyst promoted
with cerium; calcined at 110 'C

X-Ray diffraction spectrum of an iron phosphate catalyst promoted

X

49

51

52

61

73

74

A“

with cerium; calcind at 500 'C

Figure 8-5 X-Ray diffraction spectrum of an iron phosphate catalyst supported 76
byLUDOX®HS40%andpromotedwithcerium; calcinedat
l 10 'C

Figure B-6 .X-Ray diffraction spectrum of an iron phosphate catalyst supported 77
by LUDOX ® HS 40% and promoted with cerium; calcined at
500 'C

xi

1. INTRODUCTION

1.1 Scape

With the ever increasing cost of fossil fuels, especially that of petroleum, there
is a need to develop newer and more efficient technologies which are better able to use
alternate feedstocks. Most of the chemical technologies present today are based on non-
renewable energy and feedstock sources like fossil fuels. As the consumption of these
products increase, a strain is placed on the fossil fuel sources due to increased production.
Furthermore, it is well known that the worldwide stock of fossil fuels has a finite lifetime;
therefore, it is imperative that we reduce our dependence on non-renewable feedstocks.
One option is to replace conventional non-renewable feedstocks with renewable ones. The
production of maleic anhydride, an important chemical, from a completely domestic,
renewable fermentation derived source affords one such opportunity. The development of
this technology will hopefully pave the way for other technologies to be developed starting
from a fermentation feedstock. Therefore, the aim of this project is to develop a suitable
heterogeneous catalyst for the production of maleic anhydride from a fermentation

feedstock and to characterize it.

1.2 Maleic anhydride and its importance

Maleic anhydride is an unsaturated dibasic acid anhydride used extensively in
the chemical industry (Figure 1.1). Table 1.1 gives some of the important physical
properties of maleic anhydride (Cooley et al., 1983, Robinson et al., 1983).

Table 1.1 Physical properties of maleic anhydride

 

Empirical formula
Formula weight
Melting point, °C
Boiling point, °C
Specific gravity (at 20/20 °C, solid)
Specific gravity (at 20/20 °C, molten)
Viscosity, cP, at 60 °C
at 150 °C
Heat of formation, kJ/mol
Heat of combustion, MJ/mol
Heat of vaporization, kJ/mol
Heat of fusion, kJ/mol
Heat of hydrolysis, kJ/mol
Vapor pressure, mm Hg, at 44.0 °C

at 135.8 °C

at 202.0 °C
Flash point, °C: open cup
closed cup
Flammable limits, vol%, lower
“PW
Color
Crystal structure

C4H203

. 98.06

52.85

202

1.48

1.3

1.6

0.6

-470.41
-1.390

54.8

13.65

-34.9

1

100

760

110

102

1.4 - 3.4

7.1

White crystals or colorless liquid
orthorhombic

 

H0= H
C

C3—C1

/\'/\
O O 0

Figure 1.1 Chemical structure of maleic anhydride

The anhydride and the corresponding acid are industrially important raw
materials in the manufacture of alkyd and polyester resins, surface coatings, lubricant
additives, plasticizers, copolymers, and agricultural chemicals. The primary end use of
maleic anhydride is as unsaturated polyester resins whose annual volume rose to 1.4 billion
pounds in the US. in 1988 (Irving-Manshaw et al., 1989). The Chemical Products
Synopsis (May 1988) predicts the average annual consumption of maleic anhydride in both
the world and US. markets to rise steadily into the first half of the 1990's; one estimate
predicts growth at 5.9% per year during the decade 1986-1995 (Irving-Manshaw et al.,
1989). The US annual demand for maleic anhydride was 370 million pounds in both
1987 and 1988.

The growing market for maleic anhydride has prompted a number of companies
to increase their production of maleic anhydride. The major expansions planned in the US.
itself amounts to almost 300 million pounds per year (Wood, 1990). All existing maleic
anhydride production technologies are based on non-renewable feedstocks. The need to
shift to renewable feedstocks coupled with the expansion in the market suggests the

opportunity for a fermentation derived feedstock to compete with the existing technologies.

 

l . 3 Specifications (Robinson and Mount, 1983)

Maleic anhydride is specified as a white fused mass or briquettes with a
minimum crystallizing point temperature of 52.5 °C, complete and clear solubility in water
(4 gms per 10 ml water), and a minimum assay of 99.5% (ASTM D 3504-76). The color
of the melt is 20 APHA maximum, with a maximum APHA color of 40 after 2 hours of
heating at 140 °C. ASTM method D 3366-74 describes the standard test for determining the
color of maleic anhydride in the molten state.

1 . 4 Uses

The largest and most important use of maleic anhydride is in the manufacture of
unsaturated polyester resins. On being reinforced with glass fiber, these can be used in the
construction of boats, automobiles, trucks, buildings, piping, and electrical goods. The
advantages of using reinforced resins as the material of construction are their corrosion

resistance, light weight, and mechanical strength.

Maleic anhydride is the principal raw material for the production of fumaric
acid, used primarily in the manufacture of sizing resins for newspapers. Fumaric acid is
also used in the manufacture of unsaturated polyester resins, alkyd resins, quick-setting
inks, and as a food acidulant. Maleic anhydride is also used in the manufacture of malic
acid which is used as a beverage and food acidulant.

Maleic anhydride is used to manufacture lube-oil additives like ashless
dispersants. COpolymer products of maleic anhydride are made by reaction with styrene,
ethylene, and methyl vinyl ether.

The use of maleic anhydride in agricultural products includes its use in the
manufacture of herbicides, fungicides, insecticides, and various plant growth regulators.

Table 1.2 gives the breakdown of its various end uses (Robinson and Mount, 1983).

 

Table 1.2 Breakdown of the end uses of maleic anhydride

 

 

 

USES . %
Unsaturated polyester resins 55
Fumaric and Malic acids 9
Agricultural chemicals 9
Lube oil additives 9
Maleic co-polymers 5
Other uses 13

l . 5 Production

Traditionally, maleic anhydride is produced from petroleum feedstocks.
Benzene and butane are the most common feedstocks. However, as outlined before, the
aim of this project is to develop a process which uses a fermentation feedstock to produce
maleic anhydride. One such possible feedstock is succinic acid which can be produced

from a fermentation process.

The Michigan Biotechnology Institute possesses a proprietary process for the
manufacture of succinic acid using carbohydrates as a feedstock. The carbohydrates may be
derived from corn or other agricultural materials which represent an abundant, renewable,
low cost feedstock. The conversion to succinic acid utilizes an anaerobic fermentatiOn
process with an approximate product yield of 90% based on carbohydrate feedstock. Also,

the separation and purification processes involved are relatively low cost .

It has been estimated at the Michigan Biotechnology Institute that the

manufacturing cost of succinic acid is 29 cents per pound for a plant producing 150 million

6

pounds of succinic acid a year. This is based on 8 cents per pound carbohydrate, chemical
costs as listed in the Chemical Marketing reporter, $3 per million BTU steam cost and 3.2

cents per kwhr electricity cost. Fixed costs including maintenance and administration are

 

2.7 cents per pound of succinic acid. The capital cost estimated for a production plant of
this capacity is 40 million dollars. On the other hand, the Chemical Product Synopsis of
May, 1988 quoted the list price for maleic anhydride as 55 cents per pound, while the
average sale price was estimated to be 47 cents per pound. The spot market price in
August, 1990 was 57 cents per pound. With these estimates, it is realistic to develop a

process which utilizes succinic acid as a starting material for the production of maleic

 

anhydride.

The overall reaction from succinic acid to maleic anhydride that is proposed in
this work is depicted in Figme 1.2.

O 0 H20 _ 032 HC = H
II II I

- H20 I l - H2
HOCCH20H200H ——->

C C C
O / \ / \ catalyst / \ / \
0 0 O o O

succinic acid succinic anhydride maleic anhydride

 

0—0

Figure 1.2 Conversion of succinic acid to maleic anhydride

Succinic acid can be readily dehydrated to succinic anhydride by the application
of heat Figure 1.2 shows the dehydration of succinic acid to succinic anhydride, which is
believed to be an intermediate, followed by the dehydrogenation of succinic anhydride to
maleic anhydride. This reaction is believed to occur in the presence of a catalyst and it is the

object of this work to develop a catalyst suited for this reaction. The melting and boiling

7

points of each of the compounds named are sufficiently high that a vapor phase reaction is
feasible for the proposed conversion.

An extensive literature search provided no information on the conversion of
succinic anhydride to maleic anhydride as proposed. In light of this, it was decided that an
analogous system be studied in which the conversion was similar to the reaction of interest.
The analogous reaction scheme chosen was the conversion of isobutyric acid to methacrylic
acid via oxydehydrogenation, depicted in Figure 1.3. This particular reaction was chosen
due to several reasons. The similar dehydrogenation of a carbon-carbon bond adjacent to a

CH3 CH3
. - H2
CH3 — CH2 — COOH —O-> CH2 = CH — COOH + H20
+ 2
Isobutyric acid Methacrylic acid

Figure 1.3 Conversion of isobutyric acid to methacrylic acid

carboxylic group and sufficient success in the development of appropriate heterogeneous
catalysts as discussed in detail in the next chapter are some of the important attractive
features. Also, as is evident in the literature, the temperature range involved is moderate,
vapor phase reaction is possible, catalysts used (usually iron phosphate based) are
relatively inexpensive, and the conversions and selectivities achieved are reasonably good.

It was thus decided to base the proposed reaction on this analogous system.

 

 

l . 6 Objectives
The main objective is to determine whether the production of maleic anhydride
from succinic anhydride is technically feasible, and if so, to determine the best catalyst for

the purpose and to characterize it.

2. LITERATURE SURVEY

2 . 1 Commercial technologies

Maleic anhydride is not found in nature. It was first prepared by Pelouze in
1834 by heating malic acid (Cooley et al., 1983). In more modern times, benzene had been
the predominant feedstock from the start of corrrmercial production until 1974 with a switch
to n-butane as the feedstock of choice. This change was brought about by more favorable
economics for butane and also by the recognition of health risks associated with benzene.
Although all plants in the US. have switched over to butane, many plants across the world
still run on benzene. Mirror amounts of maleic anhydride are made from butene fwdstocks
or are recovered as a by-product of phthalic anhydride manufacture. Of present installed
world capacity, 51% is based on'benzene, 41% on butane, 5% on butenes, and 3% is by-
product. Table 2.1 shows the major world manufacturers (Cooley et al., 1983).

2.1 .1 Benzene route (Robinson and Mount, I 983):

Halcon (Scientific Design Co.) technology currently accounts for over 60% of
the world capacity which uses the classical benzene route. Figure 2.1 shows the flow
diagram of the process. A plant typically consists of a tubular reactor, an absorber, a partial
condenser, a dehydrating unit, a refining column, and may or may not have a benzene
recovery unit. The tubular reactor typically is a multitubular, fixed bed reactor which uses a
molten salt mixture for cooling. Usually, the reactor is made of mild steel and can contain,
depending on the capacity, over 15,000 tubes, each being about 25 mm in diameter and up

to 4 m in length.

 

ooouoem as coach. £83qu .34 3536.: 303—25 need—am.
2.33 a 83. .35 .34 do .8336 £32
3355 ma 85....— ..oom aomoméuo—Er £820 .55
.3 Be: a. Sea. .5233: aaafiaeo £2333
cocoon R .<.m.D .653 2:52 .5332
cocoon mu .<.m.D 6333* .980 comaofiohm .2302
23:5 E .<.m.D :32 .e0 19325 2832
2.355 8 be“: 68¢»qu can: Samoan—4
cocoon 8 .<.m.D Jew—2. do £83830 825‘
ooouoom 8 canon .mmofimm OHM «Boson—w 5&2
oooaoom 9. ASE—no.0 .ooufiom U< 25mm
335mm F <.m.D .Soooucom 60 353.32
92.2 accuses:
d388,.— eeeaeo 8:83 .3858

 

ecu—39:8 macs—Xv:— 3535 30—55 no? no 05 ad 035—.

2.3.5 .3 2.85: Ea...— oELFE... 30.2: ._e 9.38.35... 2: he :38: :83: flu 9:3...—

Hz<h=n
._o

mZM—NZm—n
0:28:

~=<

        

 

eaten—Ea 6.9.2:
0.. .5

O O
.23an

o. .633 i
.033 28 e. '

I s...

8.9.2.65 o ._.

 

.225 28

r

l

:53: oh

ecu—ax

 

.285 28

12

A l-l.4 mol% concentration of benzene in air is passed through the catalyst at
60—130 grams benzene per liter of catalyst per hr. The molten salt cooler produces steam
which maintains the temperature of the salt from 350 °C to 400 °C. Pressure is maintained
at 1 to 2 atm. Steam is produced in heat exchangers that cool the effluent gas from the

reactors. Yields as high as 75 mol% have been reported.

The majority of commercial benzene oxidation catalysts are supported. The
support may be metal shapes or low surface area supports, for example, ceramically
bonded alumina, silica, or carborundum. The active component of the catalyst, which
contains a number of vanadium oxides, usually is mudded onto the carrier by concentration
from a solution. This gives a catalyst with about 1-2 mzlgm of surface area. The promoting
element is usually molybdenum. Other elements that have been tested include phosphorus,
tin, boron, silver, titanium, tungsten, nickel, cobalt, and various alkali and alkaline earth

metals.

An important part of the process is the recovery and purification of the final
product from the reactor gas, which typically contains about 1% maleic anhydride. The gas
is cooled to below 200 °C in heat exchangers; then it is further cooled to near the dew point
of water which is 55-60 °C under these conditions. About 40 to 60% of the maleic
anhydride is condensed which is then separated from the gas stream. A major problem is
that the maleic anhydride absorbs water, forming maleic acid which tends to isomerize to
the more stable isomer, fumaric acid. This clogs the system and limits the amount that can
be condensed initially. The remaining maleic anhydride is then recovered as the acid by

absorption in water.

The maleic acid solution is concentrated and dehydrated to crude maleic
anhydride directly or by the use of entraining agents like o—xylene. Here, yield losses and

downtime can occur as a result of blockages with fumaric acid and tar. Effluent gases

13

contain unreacted benzene, carbon monoxide, and trace impurities. These contaminants
maybeoxidized thermallyorcatalyticallyandcarbonabsorptionmay beusedtorecoverthe
benzene for recycling. The condensed or dehydrated crude maleic anhydride is fractionated
under reduced pressure to the finished product by either batch or continuous distillation.

2.12 Butane route (Robinson and Mount, 1983):

Monsanto was the first commercial producer of maleic anhydride from a butane
feedstock in 1974. The processing methods, the conditions, and restraints are generally
governed by the similar factors as has been described for the benzene route. The greater
ratio of water-to-anhydride formed in the butane route causes lesser amounts of maleic

anhydride to be recovered by condensation as a liquid than in the benzene route.

The major point of interest here is the catalyst used. Phosphorus-vanadium-
oxygen (PVO) is the the most widely claimed type. Numerous patents using different
promoters have been issued. The elements experimented with as a promoter include iron,
chromium, titanium, cobalt and/or nickel, zirconium, zinc, molybdenum, copper, and

many others.

2.1.3 Butylene route:

A butylene containing stream from naphtha cracking is used as the feedstock
which involves using a modified PVO catalyst system. A fluidized bed reactor can be used
for the production of maleic anhydride which offers certain advantages. These include
efficient rerrroval of reaction heat, easy maintenance of optimum reaction temperatures, long
catalyst life, and more economical and easy to construct reactors on a larger scale. In this
process, the slow rate of dehydrogenation of n-butane to butylenes is the rate determining

step.

The catalyst involved here is, as for the other processes, a modified PVO

system Modifying elements tested include titanium and/or tungsten, molybdenum, copper

14

and/or niobium, uranium, lithium and other alkali metals, silver, silicon, and zirconium.
Other systems studied include vanadium-arsenic, vanadium-titanium, vanadium-
molybdenum and/or cobalt, phosphorus-tungsten, phosphorus-molybdenum,

molybdenum, and molybdenum-tungsten-antimony (Robinson and Mount, 1983).

2 . 2 Alternate technology proposed

As mentioned in the previous chapter, an alternate pathway to the production of
maleic anhydride from fermentation derived succinic acid is proposed in this work. An
idealized reaction for this pathway is depicted in Figrne 1.2.

An extensive literature search provided no citation for the proposed reaction. In
light of this, it was decided that an analogous system be studied in which the conversion
was similar to the reaction of interest. The analogous reaction scheme chosen was the
conversion of isobutyric acid to methacrylic acid via oxydehydrogenation. This reaction is
depicted in Figure 1.3. The specific reaction scheme of converting isobutyric acid to
methacrylic acid was developed by Ashland Chemical and is believed to have been brought
on line in 1989 (Szmant, 1989). This reaction was chosen as the model on which to base
the desired reaction due to several reasons. The similar dehydrogenation of a carbon-carbon
bond adjacent to a carboxylic group, sufficient success in the development of appropriate
heterogeneous catalysts, and its commercial feasibility are some of the important attractive
features. The reason that oxidative dehydrogenation is chosen over conventional
dehydrogenation is that a study of the thermodynamics reveals a much more favorable
pathway for the reaction in the presence of oxygen as the equilibrium is shifted. Also, the
free energy change tends to be negative in the presence of oxygen and positive in its
absence, which is favorable. In some cases, the presence of oxygen also greatly increases
the equilibrium constant. Although the exact mechanism of oxidative dehydrogenation is
not completely understood, an idealized mechanism for the reaction is presented in Figure
2.2 (Hightower, 1990). The example chosen is that of the oxydehydrogenation of

5.353.523 25.2.8 he =55...er e933..— ud 9...»:—

0. 9

 

 

 

 

\...../ \ : \..../ \ d \....../ \
m.. .3.“ III! e... P. ..c on. ...o..
® ..U... .. ...n.. a ..Q.. ..
.llllllll ozoo
.
..on. ..on. so“.
\ / \ \ / \ \ /o\
o — em ‘I m oo ‘IIIIII 30/ es.
2 so ooo \s ooo es .
ed .... re. s u .
.3

16

butane to butene over an iron based catalyst. The first step shows the adsorption of the
reactant molecule onto an active site on the catalyst. The next two steps show the releasing
of the hydrogens and their subsequent reaction with surface oxygen resulting in
dehydrogenation. The last two steps show the removal of water and the regeneration of the
oxygen sites on the catalyst surface. This idealized sequence indicates possible avenues for
the optimization of the process.

Obviously, the most important step in the whole reaction scheme is the catalytic
oxidative dehydrogenation of isobutyric acid. Numerous patents have been issued for

different catalysts developed for this reaction.

Watkins (1974) developed a catalytic process for the manufacture of unsaturated
acids and esters wherein lower aliphatic acids like isobutyric acid and esters like methyl
isobutyrate are dehydrogenated in the presence of oxygen and a solid dehydrogenation
catalyst at a range of temperatures from 250 'C to about 600 'C. The catalyst used was a
calcined residue of the mixed phosphates of iron and lead. Watkins (1975) also developed
a catalyst which was a calcined residue of a mixture of bismuth oxynitrate, iron phosphate,

and lead phosphate.

Statz and coworkers (1980) patented a catalyst containing alkali metal,
chromium, iron, lead, phosphorus, and oxygen which was used for the oxidative
dehydrogenation of isobutyric acid. It was based on the discovery of a phosphate catalyst
containing the aforementioned elements to be particularly well suited for dehydrogenation
reactions used to produce unsaturated acids and esters like methacrylic acid. Both fixed bed
and fluidized bed were tried with success. For the reaction of isobutyric acid to methacrylic
acid, the catalysts were tested at 400 'C at different contact times and conversions between
20% and 60% were achieved. The selectivity to methacrylic acid was between 40% to
85%.

17

Daniel and Brusky (1981a) of Ashland Oil, Inc., developed a vapor phase
reaction wherein isobutyric acid or a functional equivalent, for instance, a lower alkyl ester
was oxidatively dehydrogenated to the corresponding «,8- ethylenically unsaturated
derivative by contact with a heterogeneous catalyst in the presence of molecular oxygen.
The catalyst used was composed of calcined phosphates of iron containing silver as a
modifier or dopant component. An optimum temperature range of 375 'C to 475 ‘C with
contact times of less than a second provided extremely high conversions of isobutyric acid
from 74% to 96% with a selectivity to methacrylic acid of over 60%. Daniel and Brusky
(1981b) developed another catalyst composed of calcined phosphates of iron containing
tellurium as a modifier. These catalysts were tested at about 400 'C and a contact time of
about 0.3 seconds and gave a conversion of about 77% to 87% with a selectivity of about
70%. Daniel (1982) also developed a catalyst composed of calcined phosphates of iron,
silver, and niobium. At about 400 'C and a contact time of about 0.5 second, conversions
of40% to80% with a selectivity tomethacrylic acidofovcr70% wereobtained.

Ruszala (1983a) developed a catalyst for the same reaction which was
composed of the calcined oxides of iron and at least two members selected from the group
consisting of antimony, niobium, tantalum, and tungsten. A support was used which
constituted anywhere between 5% and 95% by weight of the finished catalyst. A host of
supports were tried which included alumina, pumice, silicon carbide, zirconia, titania,
silica, alumina-silica etc. At a reaction temperature of around 415 'C and a contact time of
less than a second, conversions from 15% to 32% were obtained while the selectivity to

methacrylic acid was around 50%.

Daniel and Brusky (1983a) used a catalyst composed of calcined phosphates of
iron containing cobalt or lanthanum as a modifier. Using similar conditions of temperature

and contact time as Ruszala, they achieved extremely high conversions of up to 99% with a

18

selectivity of around 60%. It was noticed that a high conversion also implied a lower

selectivity towards the desired product.

Ruszala (l983b) used a catalyst composed of the calcined oxides of uranium
and tlmgsten. The highest conversion achieved with this catalyst was 47% with a selectivity
of 40% at a reaction temperature of 475 'C. Daniel (1983b) used a catalyst composed of the
calcined phosphates of chromium, molybdenum, and tungsten. A silica support was also
used. A reaction temperature of around 300 'C and a contact time of 0.5 seconds provided
an optimum set of conditions for achieving conversions up to 90% and selectivities to
methacrylic acid of up to about 70%. He also used a catalyst composed of the calcined
phosphomolybdate of potassium and cerium (Daniel, l983c). Supports tried included
silica, alumina, quartz, titanium dioxide, carbon, silicon carbide, etc. A reaction
temperature of 300 'C and a contact time of less than a second gave a conversion of
isobutyric acid of about 79% and a selectivity to methacrylic acid of about 66%. Another
catalyst tried was composed of the calcined phosphates of iron, silver, and chromium and a
support material consisting of any one of silica, alumina, quartz, titanium dioxide, carbon,
and silicon carbide (Daniel, l983d). A reaction temperature of about 400 'C and a contact
time of less than a second was used. Conversions of isobutyric acid from 84% to 96% with
a selectivity to methacrylic acid from 66% to about 75% were achieved. Calcined
phosphates of iron and cerium were also tried (Daniel, l983e). Similar supports as
indicated before were tried with success. A reaction temperature of 400 'C and a contact
time of less than a second provided a conversion of 84% isobutyric acid with a selectivity
towards methacrylic acid of about 85%.

Ruszala (1984) developed a catalyst composed of the calcined mixtures of salts
of titanium and iron. Various supports were tried including colloidal silica or any other
form of silica, alumina, pumice, zirconia, quartz, carbon, silicon carbide, etc. Reaction

temperatures from about 400 ‘C to 450 'C were used with a contact time of less than a

19

second to achieve conversions of around 35% and selectivities around 45%. Daniel (1984)
1 developed a catalyst composed of iron, phosphorus, silicon, vanadium, and molybdenum
for the conversion of isobutyric acid to methacrylic acid. This catalyst also had a support as
one of silica, alumina, quartz, titanium dioxide, carbon, silicon carbide, etc. A reaction
temperature of 280 'C to 320 'C was considered optimum. With a contact time of about a
second, a conversion of about 81% was achieved with a selectivity towards methacrylic

acid of about 74%.

Pedersen and co-workers ( 1984) provide a process for the preparation of iron
phosphorus mixed oxide catalysts with a promoter and a support. The iron and phosphorus
containing compounds and promoter elements containing compounds, if any, are
introduced into a substantially organic liquid medium, heated, separated and then dried and
calcined. These catalysts are useful for the oxidative dehydrogenation of organic

compounds such as aldehydes or carboxylic acids.

McGarvey and Moffat (1991) have studied the reaction over a series of ion
exchange modified lZ-heteropoly oxometalate catalysts. watzenberger and co-workers
(1990) have studied the deactivation of such heteropolyacid catalysts and suggest a way of
eliminating deactivation of the catalyst.

2 . 3 Characterization of catalysts

Characterization of catalysts is of paramount importance in the development of
any catalytic process. At each stage of the development, characterizing the catalyst not only
checks the effectiveness of the catalyst, but also provides a standard for future products.
According to Hanseal and Hanseal (1989), the goal of catalyst characterization should not
only be an attempt to understand the catalytic act on a molecular level, but also to study the
nature of the individual catalyst sites and their interdependence and interaction with one

anotheraswellaswiththereactants.

Z)

The characterization process is usually divided into four parts. Bulk property
measurements, particle property measurements, surface property measmements and activity

measurements. Each will be discussed separately.

2.3.1 Bulk properties:

Complete qualitative and quantitative component identification within the
catalyst is the first attribute that should be determined (Richardson, 1989). Besides the
principal components added, contaminants should also be quantified. A host of techniques
are available for the determination of bulk compositions. The most popular are
spectroscOpic methods which include x-ray fluorescence (Cullity, 1956), atomic absorption
spectroscopy, inductively coupled plasma spectroscopy, and analytical electron microscopy
(Richardson, 1989).

Phase structure identification is another important aspect which is
relatively much more difficult to measure. X-ray diffraction is quite well developed for this
application. Since each crystal phase gives a characteristic pattern, it is possible to identify
all the phases in the sample. There are, however, some problems associated with this
technique. A minimum amount of sample, depending on the atomic weight, is required and
lines from different phases often occur at the same position, interfering with and
overlapping each Other (Richardson, 1989). Millet and co-workers (1990) have specifically
studied the phase structure of iron phosphate catalysts used in the conversion of isobutyric
acid to methacrylic acid. They studied the phase structure before and after the reaction to
determine the phases present and confirmed that certain new and as yet unidentified phases
had been formed. X-ray diffraction along with Mossbauer spectroscopy can also help in
understanding the role of promoters in a catalyst. Millet and Verdine (1991) have used
those techniques to explain the role of cesium in iron phosphate catalysts. Mbssbauer

spectroscopy can also be used to determine the oxidation state of iron before and after the

21

catalytic reaction as has been performed with iron phosphate catalysts to determine the most
selective phase for the reaction (Millet et al., 1989).

Delannay (1984) has indicated that electron diffraction to identify phases is also
possible during electron microscopy. The advantage of this technique is that the resolution
is very high and so individual crystallites can be identified.

Thermal analysis for characterizing solid materials is based on the fact that, at a
proper temperature interval, any solid will undergo characteristic phase transformations
(Lemaitre, 1984). Differential thermal analysis (DTA) and thermal gravimetric analysis
('I‘GA) are two of the most widely used therml analytical techniques. The former measures
energy changes as the sample is scanned through phase changes, while the latter records
weight loss or gain (Richardson, 1989). The characteristic temperature at which a thermal
change occurs in a given sample depends on the phase composition of the sample and the
composition of the surrounding atmosphere as well as on any factor which affects the

kinetics of the transformation.

2.3.2 Particle properties:

Density is an important particle property that needs to be evaluated. Density
grading is a technique used to segregate catalyst particles based on their individual densities
(Seamans et al., 1989). Richardson (1989) mentions four different ways in which densities
may be defined. The theoretical density is the ratio of the mass of a collection of discrete
pieces of solid catalyst to the sum of the volumes of each piece, if the solid catalyst has an
ideal regular arrangement at the atomic level. X-ray diffraction can be used to measure the
theoretical density. In defining skeletal density, the volume is taken as the sum of the
volume of the solid material and any closed pores within the solid. A displacement

pychometer is used to measure the particle density in which the volume is taken as the sum

22

of volume of the solid, closed pores and accessible pores within the particle. The packing
density, or the bulk density also considers the void space between the particles.

Particle size distribution can be determined, depending on the shape of the

particles, by sieving, optical and electrical imaging, light scattering etc.

Mechanical properties can sometimes supersede outstanding catalytic
activity and often are the limiting factors for using and regenerating industrial catalysts
(Bertolacini, 1989). To evaluate the mechanical properties of catalysts, several test methods
have been standardized by the ASTM D-32 Committee on Catalysts. These include tests on
attrition and abrasion, single pellet crush strength, vibrated packing density, and particle

size analysis (Bradley et al., 1989).

Surface area is perhaps the most important particle parameter specified
irrespective of the type of surface (Richardson, 1989). The principles of physical
adsorption are employed to calculate the specific surface area which is usually given in
square meters per gram of catalyst. The linearized form of the famous Brunauer, Emmett,
and Teller equation is used to calculate the surface area. Nitrogen adsorption experiments
are the principal methods used. Surface areas of catalysts can range from 0.01 m2/gm to
1000 mzlgm or more. For example, some zeolite catalysts have extremely high surface
areas of about a 1000 m2/gm while an iron alumina catalyst used for ammonia synthesis
has a surface area of only 10 mzlgm (Richardson, 1989). Sing (1980) has recommended a

general procedure for the analysis of physisorption measurements on catalysts.

Pore size distribution has become an essential feature of particle
characterization (Lecloux et al., 1981). These are required to properly estimate pore mouth
poisoning, pore diffusional resistances, and deactivation control. Usually, macropores are
measured with mercury porosimeters, whereas mesopores and micropores are measured

with nitrogen adsorption-desorption isotherms (Richardson, 1989).

23

Although measuring the diffusivity of catalysts with a high degree of
accuracy is quite difficult, nevertheless it is important to know the diffusion coefficient for
the calculations of effectiveness factors. The classical method of measuring diffusivity is
the Wicke-Lallenbach technique, although cinematographic techniques have also been used
(Richardson, 1989). i

2 .3 .3 Surface properties:

Since the surface of the catalyst is the active part in any catalytic reaction,
measurement of surface attributes is of paramount importance to understand the behavior of
the catalyst. The important features to be characterized include morphology and

composition, structure, dispersion, and acidity. These are further discussed.

Morphology involves the study of the shape and size of the catalysts.
Numerous techniques are available , the primary ones being Scanning Electron Microscopy
(SEM), Transmission Electron Microscopy (TEM), Scanning Transmission Electron
Microscopy (STEM), X-ray diffraction, Extended X-ray Absorption Fine Structure
(EXAFS), Auger Electron Spectroscopy (AES), X—ray Photoelectron Spectroscopy (XPS),
and Ultraviolet Photon Spectroscopy (UPS) to name a few. Delgass and co-workers
(1979) provide an excellent description of various spectrosc0pic techniques involved.
Cohen (1990) has used x-ray diffraction to obtain statistical information on the nature of

catalysts and on the structure of micron-sized single crystal particles.

Structure identification of the surface phases is best accomplished by electron
microscopy using either an SEM or TEM or both. The advantage in using electron
microscopy is that since electrons interact much more strongly with matter than do X-rays
or neutrons, they can be scattered appreciably by extremely small atomic clusters (Howie,
1980). This technique is often used complimentary to Auger Electron Spectroscopy or X-
ray Photoelectron Spectroscopy.

24

Dispersion of the active fraction of the catalyst is defined as the ratio of the
numberofactiveatomsexposedatthe surfacetothetotalnumberofactiveatomspresentin
the catalyst. The experimental techniques used to measure dispersion include chemisorption
techniques, X-ray diffraction and scattering, Transmission Electron Microscopy, and XPS

peak intensity measurements (Lemaitre et al., 1984).

Acidity measurements include measuring the type of acidity, the acid strength,
and the distribution of acid strengths (Richardson, 1989). Jacobs (1984) has provided a

detailed account of the various techniques available to measure the acidities of catalysts.

2 .3 .4 Activity:

Kinetic activities are based on the rate equations for the reaction under study.
For this reason, it is essential to conduct kinetic experiments to determine the exact rate
equation first. Once the rate equation is available, the activity can be expressed either as a

rate, a rate constant or an Arrhenius constant (Richardson, 1989).

To measure the kinetic rate data, different types of experimental reactors may be
used. Tubular reactors can be operated in either a differential or an integral mode. Spinning
basket or internal recycle reactor is very commonly used for measuring kinetic data
(Kenney, 1980). Batch reactors are usually employed for liquid phase reactors using
sllu'ried catalysts. Pulse reactors are simple and fast reactors that offer economy of feed and

provide protection against process deactivation (Richardson, 1989).

Milller-Erlwein and Hofmann (1988) have considered the problem of evaluating
kinetic parameters from dynamic experiments for the heterogeneously catalyzed oxidative
dehydrogenation of isobutyric aldehyde to methacrolein.

3. MATERIALS AND METHODS

The project was essentially divided into three stages: a) preparation of the

catalysts, b) testing of the catalysts, and c) characterization.

3 . 1 Preparation of the catalysts
3.1.1 samurai Setup:

The catalysts were prepared in accordance with the experimental techniques
mentioned in the various patents describing the conversion of isobutyric acid to methacrylic
acid, discussed in detail in the last chapter. A wide mouthed flask was used as the reaction
vessel with a refluxing column and a condenser attached to it. The reaction mixture was
constantly stirred and the vapors refluxed using cooling water in the reflux column.
Refluxing was continued for about 24 hours after which the reaction was considered to be
complete. The excess water in the system was distilled off. The resulting thick paste
obtained was dried in an oven overnight at about 110 'C. The catalyst was removed from
the reaction vessel and calcined in an oven. Calcination was carried out in air at 450 'C for

six hours and further calcined in a mixture of nitrogen and oxygen at 500 'C for two hours.

3 . I .2 Catalyst preparation:

Two general categories of catalysts were prepared. The first were catalysts
which had a molybdenum oxide base whereas the second had an iron phosphate base. In
the latter category, four different types of catalysts were prepared, each having a different
composition: the first was the base iron phosphate catalyst, the second had a support
(LUDOX® HS 40%) added to the base, the third had a promoter (lanthanum or cerium)
added to the base, and the fourth had both the support and the promoter added to the base.

5

3

Table 3.1 gives a summary of the various catalysts prepared and their
constituents. All chemicals used were bought fiom Aldrich Chemical Co., except for the
silica support LUDOX® HS 40% (DuPont Chemicals), iron nitrate nonahydrate
(Columbus Chemical Industries, Inc.), and heavy water (Cambridge Isotope Laboratories).

A typical iron phosphate catalyst preparation which has all the ingredients: the
base catalyst, a support, and a promoter is described as an example. The preparation

follows the techniques available in patent literature (Daniel. 1983c).

Forty eight and a half grams of iron nitrate nonahydrate was mixed with 15 ml
phosphoric acid, 120 ml distilled water, 10 ml LUDox® HS 40% as the support, and 2.6
grams of cerium nitrate hexahydrate as the promoter. The chemicals were put in a reaction
vessel and refluxed while being continuously stirred. Refluxing was continued for almost
24 hours after which the reaction was deemed complete. It was assumed that all the iron
nitrate would have reacted with the phosphoric acid to produce the phosphate while the
silica support (LUDox® HS 40%) and the cerium promoter would have embedded into
the crystal structure of the catalyst. The slurry obtained was heated for about two hours in
order to distill off the excess water. The thick paste formed was dried for about 24 hours at
110 ’C. The catalyst obtained was calcined at 450 ’C in air at a furnace available at the
Composite Materials and Structures Center, Michigan State University. The catalyst was
further calcined in the presence of a mixture of nitrogen and oxygen at 500 'C in the reactor

at the Michigan Biotechnology Institute.

The other iron phosphate catalysts were prepared in exactly the same way,
except that the silica support and the cerium or lanthanum promoter was added only when

so desired. Detailed preparations for all catalysts are given in Appendix A.

. m

27

Table 3.1 Summary of catalysts synthesized

 

Catalyst

Constituents

 

Iron phosphate based

1) Base catalyst

2) Base catalyst with support

3) Base catalyst with promoter

4) Base catalyst with promoter

5) Base catalyst with support
and promoter

6) Base catalyst with support
and promoter

Molybdenum oxide based

1) Molybdate catalyst

2) Molybdate catalyst

3) Molybdate catalyst

Iron, phosphorus, and oxygen
Iron, phosphorus, oxygen, and silicon
Iron, phosphorus, oxygen, and
lanthanum as promoter

Iron, phosphorus, oxygen, and
cerium as promoter

Iron, phosphorus, oxygen, silicon,
and cerium

Iron, phosphorus, oxygen, silicon,
and lanthanum

Molybdenum, oxygen, phosphorus,
vanadium, and copper.

Molybdenum, oxygen, phosphorus,
vanadium, tungsten, and hydrogen.
Molybdenum, oxygen, phosphorus

cerium, and potassium

 

23

The molybdenum oxide based catalysts were prepared using a molybdate
compound instead of the iron phosphate. A typical preparation of a molybdate catalyst is
described next. This preparation follows that described in the patent literature discussed
earlier (Daniel, 1983c). Other preparations are given in Appendix A.

Thirty milliliters of distilled water was used as a solvent into which 1.72 grams
of potassium hydroxide was dissolved. A thick slurry was made by mixing 27 grams of
(NH4)2M07024.4H20, 8.1 grams of (NH4)2Ce(NO3)6, and 3.2 ml of phosphoric acid.
30 ml of concentrated hydrochloric acid was added slowly to the slurry, which caused it
change to a red color and then back to yellow. The yellow slurry was placed on a magnetic
stirrer and heated at 125 'C for five hours. A thick paste was obtained which was placed in
an oven todry at 110 'C for 18 hours.

3.1.3 Commercial catalysts:
Commercially available standard dehydrogenation catalysts were also procured
from United Catalysts Inc., Louisville, Kentucky. Table 3.2 lists these catalysts along

with their constituents and form. Exact compositions are proprietary information.

Table 3.2 Commercial catalysts tested

 

 

 

 

 

 

 

CATALYST CONSTITUENTS and FORM
United Catalyst G-84C Iron, Potassium, Chrome ; 1/8" extrusions
United Catalyst G-65 Nickel on Refractory ; 1/8" extrusions
United Catalyst G-13 Copper Chromite; 3/16" x 3/16" tablets
United Catalyst G-22 Copper Chromite; 3/16" x 3/16" tablets
United Catalyst 04] Cluomium on Alumina; ll8" extrusions

 

 

 

3 . 2 Testing of the Catalysts
All the catalysts were tested in a reactor located at the Michigan Biotechnology

Institute. The reactor setup is shown in Figure 3.1.

3.2.1 Reactor setup:

The reactor was a simple quartz tube with a glass fiit attached at the center
which served to act as a support for the catalyst. The reactor tube was enclosed inside a
three zone furnace (series 3210, manufactured by Applied Test Systems, Inc.). The
temperature of each zone of the furnace was independently controlled using Digi-Sense
temperature controllers (model 2186-10, marketed by Cole-Parmer Instrument Company).
A syringe infusion pump with adjustable flow rates, (model 11, manufactured by Harvard
Apparatus), at the top of the reactor served to inject the succinic acid feed solution into the
reactor. Glass wool was used to prevent any contamination on the flit. Glass beads were
loosely packed on the glass wool up to the top of the reactor to provide a uniform flow of
the feed on to the catalyst. Nitrogen was used as an inert carrier gas through the reactor. A
provision was made to introduce oxygen into the reactor along with the nitrogen. The flow
of the two gases was regulated by Brooks 5850B series mass flow controllers. The product
gases from the reactor were condensed into a collection flask placed in an ice bath. The
reaction products, collected on the sides of the collection vessel were dissolved in water
and taken for analysis. Carbon dioxide production was monitored by routing the effluent

gases to a flask containing a buffer solution of pH 5 into which an ORION (model 95-02)
C02 electrode was immersed.

3.2.2 Product analysis:
Product analysis was done on a VXNMR 300 machine located at the Max T.

Rogers facility at Michigan State University. 1H nuclear magnetic resonance was the

 

 

 

 

 

 

 

 

 

 

l.Nirrogengascylinder 6PM

ZOxygengascylinder 7.Collectionvessel
3.8yringepumpwithfeed 8.1cebath
4.Reacta'tube 9.Terrrperlttrrecontrollers
5.Catalyst

Figure 3.1 Reactor setup

31

technique of choice because of the ease in identifying products and determining their
composition. This technique permits the identification of all compounds regardless of their
functionality. In this case however, maleic anhydride cannot be detected because of the
large amounts of water in the system, which causes the hydration of the anhydride to the
acid. So, the final detectable product was maleic acid. As the technique required a presence
of a deuterated solvent to lock on, the succinic acid was dissolved in heavy water, D20,
instead of regular water. Typical NMR spectra are shown in Figure 3.2. Figure 3.2a shows
succinic and maleic acid standards at a concentration of 5% in D20; while Figure 3.2b

shows the reaction products from a typical run over an iron phosphate catalyst. D20 is

highly hygroscopic, absorbing atmospheric water to form HDO. This shows up in the
NMR spectra as the large solvent peak. Decoupling techniques were used to reduce the size
of the large solvent peaks expected in all analyses. As is evident from comparing the two

spectra, the formation of maleic acid is confirmed.

1H NMR, as the name suggests, is actually a technique to measure the number
of protons present in the sample (Delgass et al., 1979). As maleic acid has two protons less
than succinic acid, the peak of maleic acid will be half that of succinic acid at the same
concentrations. Therefore, to determine the relative amounts of the two, the area of the
maleic acid peak had to be multiplied by two to equate it with the succinic acid peak. In
Figure 3.2b, the numbers alongside the peaks indicate the area as obtained by the analysis.
Since there was initially no maleic acid, all maleic acid present is as a result of conversion

of succinic acid. As an example, the conversion to maleic acid is given as

% Conversion (O.l9*2)/ {(0.l9*2)+1.09} * 100

25.8 %

The absence of any other detectable peak in the spectrum also indicates the high

selectivity of the reaction. The only other reaction product possible which is not detectable

3.3:!» 22. 22a:— $m E.- Eua 9.53... 93 2:32.» .5592.» :22 and «Sara

see or _. . m m e m m
_ _

r—bp-b—bbbh—p-b- hbpb—Fbpp—Lppn—th- h-pb—npbb—p

 

IL

one a: use:

 

 

Boa 258.6

 

328.8 82.38.... .3... 93.3.3.6 a .55 29.. 39......
e. 22. 9.53.: 2. 5.8.8.. 5.353323»: 2.. 3 52.3.... 5.3.3.. o... 3 5.9.32... :22 3..” 9...»...—

 

 

 

 

eooot.m m e m m a m «.2222
p.rr...r:r....._r.rr............r:......_.C...:.....r..c.._.......:......
8.. a...
2832.:
one

 

22. 03.8.6

34

by 1H NMR is carbon dioxide which could be produced if the succinic acid was being
directly decarboxylated to C02 in the presence of oxygen. This possibility was monitored

by the 002 electrode. The presence of 002 is indicated by a large change in the voltage
reading by the electrode.

3.2 .3 Experiments:
Experiments were conducted with commercially available dehydrogenation
catalysts to determine if the conversion of succinic acid to maleic acid was possible with

reasonable conversion and selectivities.

Subsequently, experiments were conducted with the catalysts prepared in the
laboratory to determine their effectiveness in the proposed conversion. Thereafter, the
effect of various parameters on the conversion of succinic acid to maleic anhydride were
studied. These parameters included temperature, contact time, oxygen concentration, and
water concentration. Also, the effect of time on-stream on catalyst performance was

evaluated.

Temperature dependence was studied by changing the temperature of the middle
zone of the furnace which actually enclosed the catalyst. Contact time was varied by
changing the height of the bed of catalyst keeping the flow rates constant. A range of
oxygen concentrations was achieved by controlling oxygen flow rate though the reactor by
means of a mass flow controller. The effect of feed concentration was studied by using
different succinic acid concentrations. Long time runs were conducted to study the possible
deactivation of the catalysts with time. In order to keep all the reactions on an equal footing,
a constant weight hour space velocity, defined as the mass of feed per mass of catalyst per

hour (Robinson, 1989) was maintained throughout in all the runs.

5

3 . 3 Characterization of the catalysts
In order to characterize a catalyst sufficiently, it is necessary to determine the
specific surface area, surface and bulk compositions, and the bulk crystal structure of the

catalyst.

3 .3 .1 Surface area analysis:

The specific surface area, 88 (ngm'l), is perhaps the singular most important
particle parameter specified without regard to the type of surface. Measurement of surface
area is based upon the principles of physical adsorption. Surface area of the catalysts is
measured by nitrogen adsorption experiments done on a Micrometrics PulseChemisorb
2700 instrument. The Brunauer, Emmett, and Teller (BET) analysis (Robinson, 1989) is

used to determine the actual surface area. This approach uses the BET equation shown

 

 

below:
V _ CP
VM _ [pa-pl [1+ (c-1)p/po] (3.1)
Where,
p = pressure

po = saturation pressure
V = volume adsorbed

VM = monolayer volume

C = constant

The parameter c includes the heat of adsorption and liquefaction and is fairly constant for a

given class of materials (like oxides and metals), with values below 100 (Robinson, 1989).
The BET equation is valid only for low pressures, up to a value of 0.3 for p/po. At higher

$

pressures, liquid condensation begins in the micropores and progresses through the

mesopores as the pressure approaches the saturation pressure, i.e. p/po = 1.0, at which

point the equation breaks down.

Experimentally, it is impractical to measure VM since it is valid in such a small
region. However, a very practical method employed is to linearize the BET equation by

rearranging the variables to obtain the following equation:

 

 

P 1 (c-1)
—— = + (p/po)
V (pa-p) Vatc ch (3.2)

This equation gives rise to linear plots whose slope and intercept can easily be measured.

The value of VM can be calculated as the inverse of the sum of the slope and the intercept.
With this information, and knowing the weight of the sample used, the specific surface area
(S) of the catalyst is easily determined by using equation (3.3).

S = VM.A.N/M (3.3)
where A = Avogadro's number
M = molar volume, and

= area of each adsorbed gas molecule; for Nitrogen, N = 16.2 nm2

3 .3 .2 X -Ray difi’raction analysis:

In X-ray diffraction, monochromatic x-rays are reflected from the sample with
diffraction lines produced from the repetitive dimensions of crystal planes. Each crystal
type gives a characteristic pattern, so that the position of the lines gives an idea of the

particular phase of the compound that might be present in the sample (Richardson,1989).

37

X-ray diffraction analysis was used in this work to determine the crystallinity of
the catalyst. Once that was established. the exact crystalline phase of the material was
determined by comparing the peaks obtained to the standard spectra available. Of special
interest to this project was to determine whether a change in calcining conditions gave rise
to any additional phases. Diffraction spectra of catalysts calcined at different temperatures
were compared to determine the effect ofcalcining temperature on the catalysts.

The X-ray diffraction experiments were conducted on a Rigaku X-ray
diffractometer having a rotating anode. Copper Ka X-rays were generated by the machine

which was used for the analysis.

3 .3 .3 Bulk composition analysis:

Elemental analysis of the catalyst was done in order to determine the bulk
composition. Standard assays were employed to quantify percentages of iron, phosphorus,
cerium, silicon, nitrogen, and any other elements that might be present. However, due to
the inherent nature of the assays, it was not possible to determine the concentration of
oxygen in the presence of metals. Bulk composition analyses of all catalysts were obtained
from Galbraith Laboratories Inc., Tennessee.

Kjeldahl nitrogen was determined volumetrically. Decomposition was done in a
Carius furnace. Steam distillation of the sample from the Kjeldahl flask was done after
making it alkaline with excess of 50% NaOH. The liberated ammonia was trapped in
saturated boric acid solution. Titration was carried out with 0.01N H2804 using methyl
red-methylene blue indicator. Detection limit was 0.18 mg nitrogen. Quantitation limit was
0.3 mg nitrogen. Possible interferences were from N-N and NO linkages.

Cerium was determined by inductively coupled plasma emission spectroscopy.
Decomposition of the sample was done in a Carius furnace. Determination was by emission

$

at 413.77 nm. Detection limit was 1.6 ppm. Only spectral interferences were possible in

this assay.

Silicon was determined by inductively coupled plasma emission spectroscopy.
Decomposition was by fusion. Determination was by emission at 251.6 nm. Detection limit

was 0.10 ppm.

Iron was determined by inductively coupled plasma emission spectroscopy.
Decomposition was done either by wet digestion or by lithium metaborate fusion.

Determination was by emission at 248.3 nm. Detection limit was 0.30 ppm.

Phosphorus was determined by plasma emission spectroscopy. Sample
preparation involved decomposition by digesting in an acid mixture or by lithium
metaborate fusion. Determination was by primary emission at 178.2 nm. Detection limit

was 0.05 ppm.

3.3 .4 Sufiace composition analysis:

The surface compositions of the catalysts is determined by using X-ray
photoelectron spectroscopy (XPS). In XPS (or Electron Spectroscopy for Chemical
Analysis, ESCA), use is made of the principle that a soft x-ray photon incident on a sample
ejects any electrons with less binding energy. A sample is bombarded with monoenergetic
photons like Al Ka (1486.6 eV) or Mg Ka (1253.6 eV), which ejects electrons from the
core and valence shells in which the ionization potential, or binding energy, is less than the
primary photon energy (Edmonds, 1980). These ejected electrons can be detected and
counted to obtain a spectrum characteristic of the elements present in the sample. As
sensitivity is possible only to depths of one to twenty layers, this technique gives
information only on the weighted average surface composition of the sample. However,

this is extremely desirable, since it is the surface of the catalyst that is the active part.

$

XPS data were obtained on a Perkin-Elmer Surface Science instrument
equipped with a model 10-360 precision energy analyzer and an ornnifocus small spot lens.
X-rays were generated by an Al (1486.6 eV) anode operated at 15 kV and 20 mA. XPS
binding energies were referenced to the C ls line (284.6 eV) and were measured with a

precision of :l: 0.2 eV or better.

4. RESULTS AND DISCUSSIONS

4 . 1 Testing of commercial catalysts

Five commercially available catalysts were tested under different conditions in
order to ascertain their suitability to the reaction being studied. These catalysts were
obtained from United Catalyst Incorporated and were recommended by the manufacturers
as proven dehydrogenation catalysts. Table 3.2 lists these catalysts and their constituents.

Exact compositions are proprietary information.

Each of the catalyst was tested at two different temperatures. The temperatures
chosen were 375 °C and 475 'C. Each run was conducted under the conditions specified
below.

a) contact time = 4 seconds,

b) weight hourly space velocity (WHSV) = 0.8-0.9 grns reactant per gm of catalyst
per hour,

c) feed concentration = 40 grams per liter,

(1) oxygen to succinic acid mole ratio = 10:1, and

e) time on-stream = 15 minutes.

The conversions and selectivities obtained are shown in Table 4.1. As is evident
from the numbers in Table 4.1, the commercial catalysts showed little promise as a catalyst
for the conversion of succinic acid to maleic anhydride. Although these catalysts gave
conversions of over 90%, as evidenced by the relative size of the peaks on the NMR
spectra, the selectivity to maleic acid was less than 1%. The poor selectivity was confirmed

by the C02 elecn'ode measurements which showed an extremely high percentage of 002
40

Table 4.1 Performance of commercial catalysts

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

CATALYST TEMPERATURE CONVERSION SELECTIVITY

(° C) % %

G-84 C 375 > 90 < 1
475 > 90 < 1

G65 375 > 90 < 1
475 > 90 < 1

GB 375 > 90 < 1
475 > 90 < 1

G-22 375 > 90 < 1
475 >90 < 1

G4] 375 > 90 < l
475 > 90 < l

 

 

.‘Vs

4.2

formation during the reaction. It was thus evident that most of the reactants were being
directly decarboxylated over the catalyst. The extremely small intensity of the peaks on the

NMR spectrum made it impossible to measure the conversion with any greater accuracy

than that reported.

The presence of a large amount of C02 is indicative of an acid catalyzed
reaction. This can be explained on the basis of the metal oxides present in these catalysts.
Metal oxides are active in the dissociation of steam at high temperatures. This leads to the
dissociation of water and the subsequent formation of acid and base sites on the catalyst
surface which significantly increases decarboxylation. Thus, it is unlikely that any acidic or
basic catalyst will be suitable for the oxidative dehydrogenation reaction. ‘

4 . 2 Testing of synthetic molybdenum oxide based catalysts
The three molybdenum oxide based catalysts synthesized in the laboratory were

analyzed under the following conditions:

a) contact time = 4 seconds,

b) WHSV = 0.8 to 0.9 gm succinic acid per gram catalyst per hour,

c) temperature = 475 'C,

d) feed concentration = 40 gms per liter,

e) oxygen to succinic acid ratio = 10: 1, and

0 total time on-stream = 30 minutes.

Table 4.2 shows the performance of these catalysts. All the molybdenum based
catalysts yielded conversions of less than 10%; however, the selectivity towards maleic
anhydride was always greater than 95%. This is confirmed by the NMR spectra obtained
and by the 002 sensor which did not indicate the production of any measurable amount of

carbon dioxide. The lack of 002 formation indicates that the postulated mechanism is

operating and that, in this case, decarboxylation is not the predominant pathway.

 

43

Unfortunately, the molybdenum based catalysts possess an inherent low thermal stability
which reduces their potential for commercial application to operating temperatures of less
than about 400 'C. An irreversible phase change occurs at this temperature which makes it
impossible to use these catalysts at higher temperatures, which is necessary to obtain higher

conversions.

Based on these observations, it was decided to discontinue further testing of
molybdenum based catalysts and to concentrate on those which would be better suited for
commercial application.

4 . 3 Testing of synthetic iron phosphate based catalysts

Preliminary studies were done in order to test the ability of the iron phosphate
based catalysts prepared in the laboratory to convert succinic acid to maleic anhydride.
Initial testing revealed an improved conversion using catalysts promoted by cerium
compared to those promoted by lanthanum under identical conditions. This observation

resulted in limiting fin'ther testing to only those catalysts promoted by cerium.

4 .3 .1 Efl’ect of temperature on conversion:
All the catalysts were tested for the conversion obtained and the selectivity

towards maleic acid. Three different temperatures of 375 ‘C, 425 'C, and 475 'C were

Table 4.2 Performance of synthetic molybdenum based catalysts

 

 

 

 

 

CATALYST CONVERSION SELECTIVITY
CONSTITUENTS % %
Mo, P, V, Cu, O 3 >95
Mo, P, V, W, O, H 9 >95
Mo, Ce, K, P, O 5 >95

 

 

 

 

44

tested. These temperatures were the broad ranges described in patent literature for the
conversion of isobutyric acid to methacrylic acid (Daniel, 1983a,b,c etc.). The reactions
were carried out under the following conditions:

a) contact time = 4 seconds,

b) WHSV = 0.8-0.9 grams reactant per gram of catalyst per hour,

c) feed concentration = 40 grams per liter,

d) oxygen to succinic acid mole ratio = 10:1, and

e) total time on-stream = 30 minutes.

Figure 4.1 shows the results obtained for these catalysts. The results show a
very promising trend for the iron phosphate based catalysts. From the results obtained, it is
evident that temperature is a very important variable in the performance of the catalysts.
For all the catalysts, at 375 'C, the conversions obtained were between 16% and 33%;
whereas, at 475 'C, the conversions obtained were between 49% and 59%. This
establishes that a higher temperature is more favorable for the reaction within the chosen

range.

Also, the selectivities obtained were all greater than 95%. Almost no production
of carbon dioxide or any other by-products was observed. Less than 5% of succinic acid
feed was converted to acrylic acid and C02. This high selectivity obtained with the iron
phosphate based catalysts is more than that achieved for the analogous reaction of

isobutyric to methacrylic acid reported in patent literature.

These results conclusively prove that succinic acid can be converted to maleic
anhydride using the iron phosphate based catalysts. The high selectivity obtained is
believed to be a consequence of the inherent stability of the five membered ring of succinic
anhydride. The absence of any free succinic acid sites offers little opportunity for

decarboxylation and subsequent production of carbon dioxide to occur.

 

 

 

 

 

 

 

 

m-
50.
a 1
ii 40"
4% .
8
.E 30-:
0
> t
a
O
U 20-
10"
l
0 j ‘ 7. ' ' ‘ ' l

 

360 380 400 420 440 460 480

Temperature (°C)

Figure 4.1 Effect of temperature on the conversion of succinic acid to
maleic acid using iron phosphate catalysts under the following conditions:
a) contact time = 4 s, b) WHSV = 0.8-0.9 gm feed/gm catalyst/hr, c) feed
concentration = 40 g/I, (1) oxygen to feed mol ratio = 10:1, and e) time on-

stream = 30 minutes

48

Based on these results, it was decided to conduct further studies on these four
catalysts only at 475 'C in order to optimize other conditions for the best operation of the

reactor.

4 .3 .2 Efl'ect of contact time on conversion:

The effect of contact time on the conversion of succinic acid to maleic acid over
two iron phosphate based catalysts was studied in order to establish the optimum contact
time to be used. The experiments were carried out under the following reaction conditions:

a) Feed concentration = 40 g/L,

b) WHSV= 0.8 to 0.9 grams succinic acid per gram of catalyst per hour,
c) Temperature = 475 'C,

d) Oxygen to succinic acid mole ratio = 25:1, and

c) Total time on-stream = 15 minutes.

The results of the experiments are shown in Figure 4.2. For both the pure base
catalyst and the base catalyst with support and promoter, the optimum contact time was
determined to be 4 seconds. All of the iron phosphate catalysts tested showed
approximately 15% less conversion at a contact time of 2 seconds as compared to a contact
time of 4 seconds. A contact time of 6 seconds shows virtually no increase in the
conversions obtained. It is to be noted that under all conditions, the selectivities obtained

for the iron phosphate based catalysts were higher than 95%.

This is an indication that equilibrium is achieved in four seconds. Based on this,

it was decided to conduct all futme experiments at a contact time of 4 seconds.

4 .3 .3 Efi‘ect of feed concentration on conversion:
In a typical fermentation broth, the concentration can be expected to fluctuate

somewhat depending on the the conditions in the fermenter and on the efficiency of

 

 

 

 

 

 

 

 

 

7O
1
60.
E ..
(D 50
9 .
Q)
3- 40-
C
.9
e 30- “2“
0
> 1
5 20
0
‘ —0— mm
10. "'fi'“ mammal”:
O . u . . . . .
1 2 3 4 5 6 7

Contact time (seconds)

Figure 4.2 Effect of contact time on the conversion of succinic acid to
maleic acid using iron phosphate catalysts under the following conditions:
a) temperature = 475 ‘C, b) WHSV = 0.8-0.9 gm feed/gm catalyst/hr, c)
feed concentration = 40 g/l, d) oxygen to feed mol ratio = 25:1, and e)- time

on-stream = 15 minutes

48

separation. This implies that the concentration of feed obtained can vary in concentration. It

is thus imperative to study the effect of feed concentration on the catalyst performance.

The effect of the feed concentration on conversion and selectivity was tested by
using two concentrations of 40 g/L and 80 g/L. The results of the experiments are shown in
Figure 4.3. All the experiments were carried out under the following reaction conditions:

a) Contact time = 4 seconds,

b) WHSV: 0.8 to 0.9 grams succinic acid per gram of catalyst per hour,
c) Temperatm'e = 475 'C,

d) Oxygen to succinic acid mole ratio = 25:1, and

e) Total time on-stream = 15 minutes.

From the results obtained, it is evident that the catalysts tested performed
equally well even when the feed concentration was doubled from 40 g/L to 80 g/L. Again,
selectivities obtained were greater than 95%. This indicates the versatility of the catalyst
towards the range of feed concentrations expected from a typical fermentation product.

4 .3 .4 [fleet of oxygen concentration on conversion:

The amount and purity of oxygen required for the oxidative dehydrogenation
reaction to take place is an important parameter since the cost of oxygen will be a major
factor in the economics of the process. Experiments were carried out with two catalysts
varying the relative concentrations of oxygen to the feed. The experiments were conducted
under the following conditions:

a) Contact time = 4 seconds,
b) WHSV: 0.8 to 0.9 grams succinic acid per gram of catalyst per hour,
c) Total time on-stream = 15 minutes, and

d) Feed concentration = 40 g/L.

 

 

 

 

 

 

 

 

 

 

 

w-
50-
O I
G
o
4o-
a.
v c
8
.E 30-:
2 .
t:
o
C) 20'
"C}-'lnnmaflnn
—O- sumac-um:
10- "A" pro-mam
“U“ aupporteddtpromotedcatalyat
O ' I ' I r I t 1 w I v

30 ' 40 50 60 70 80 90

Feed concentration (grams/liter)

Figure 4.3 Effect of feed concentration on the conversion of succinic acid
to maleic acid using iron phosphate catalysts under the following
conditions: a) contact time = 4 s, b) WHSV = 0.8-0.9 gm feed/gm
catalyst/hr, c) temperature = 475 °C, d) oxygen to feed mol ratio = 25:1,

and e) time on-stream = 15 minutes

50

The results of the experiments are shown in Figure 4.4. In all experiments, the
selectivities obtained were over 95%. The results indicate that when the mole ratio of
oxygen to succinic acid was greater than 10:1, there was no substantial increase in the
conversions obtained. If the proposed mechanism for oxidative dehydrogenation was
working, the total oxygen requirement for conversion of succinic acid to maleic anhydride
would be stoichiometric with respect to succinic acid. In all the experiments conducted. the
oxygen concentration used was always above at least ten times the minimum required; so
the result obtained, i.e., no change in conversion with increase in oxygen concentration is
only to be expected and further substantiates the proposed mechanism. This low oxygen
requirement strongly suggests that pure oxygen is not required for large scale production
and that air should be an adequate source of oxygen, thereby reducing the operating cost

substantially.

4 .3 .5 Efi'ect of time on-stream on conversion:

Experiments were conducted on two catalysts to test their performance after
different time intervals. This information is of interest because it would determine whether
the catalysts need to be regenerated after a certain amount of time. The experiments were
carried out under the following reaction conditions:

a) Contact time = 4 seconds,

b) WHSV: 0.8 to 0.9 grams succinic acid per gram of catalyst per hour,
c) Temperature = 475 'C,

d) Oxygen to succinic acid mole ratio = 25: 1, and

e) Feed concentration = 40 g/L.

The effect of time on-stream on catalyst performance is shown in Figure 4.5.
Selectivities achieved were greater than 95%. From the results, it is seen that the base iron
phosphate catalyst shows a slight decrease in conversion with time; whereas, the catalyst

promoted with cerium and supported with silica shows a slight increase in conversion after

51

 

 

 

 

 

 

 

 

7o
* 475°C
60‘ 4:. {I
I E, .
o 50‘
f:
8 J ,
- —n—
5% 40- manly:
g ‘ + mm
g 30-
>
= I
O
u 20.
. tr 4 375°C
10« ‘
o . . . .
5 , 15 25

Moles of oxygen per mole of succinic acid

Figure 4.4 Effect of oxygen concentration on the conversion of succinic
acid to maleic acid using iron phosphate catalysts under the following
conditions: a) contact time = 4 s, b) WHSV = 0.8-0.9 gm feed/gm
catalyst/hr, c) feed concentration = 40 g/l, and d) time on-stream = 15

minutes '

 

 

 

 

 

 

 

m.
c 50‘ a
G 5" u
0
a I
0
e- 40.
fi
.9.
i”.
‘>’
g 30‘
U I
—'D— bane-tam:
201
—A— thppatedeflyrt
10 ' I ' I 1 I V 1 1 I

 

10 20 30 40 50 60 70

Time (minutes)

Figure 4.5 Effect of time on-stream on the conversion of succinic acid to
maleic acid using iron phosphate catalysts under the following conditions:
a) contact time = 4 s, b) WHSV = 0.8-0.9 gm feed/gm catalyst/hr, c)
temperature = 475 ’C, d) feed concentration = 40 g/l, and e) oxygen to feed

mol ratio = 25:1

53

60 minutes. However, there is no indication of any significant change in the catalyst
performance with time. Although the runs need to be conducted for a longer length of time,
indications are that catalyst poisoning and regeneration are not significant problems for

these catalysts.

4 . 4 Results of characterization experiments

To characterize the catalysts under study, it was decided to find the surface area
of each of the catalysts, to determine the bulk and surface compositions and to try and
determine the crystal structure information of each of the catalysts. The techniques used

were physisorption, X-ray diffraction and X-ray photoelectron spectroscopy.

4 .4 . 1 Surface area measurements:
As discussed in the previous chapter, the linearized BET relation (eq.3.2)
reproduced below, was used to calculate the surface areas of the catalysts. Multipoint

analysis was used to ensure better accuracy. Nitrogen partial pressures of 5, 10 and 20

 

 

percent were chosen.
P 1 (ed) (p/ )
—" = + P

A graph of p/(V*(po-p)) vs. p/po gives a straight line with a slope of (c-1)/(VMc) and an
intercept of 1/(VMc). As described in the previous chapter, the slope and intercept can be
utilized to give the surface area of the catalyst in ngm'1 using equation 3.3.

Figure 4.6 shows a typical BET plot for a base iron phosphate catalyst with a
silica support and with cerium as a promoter. The slope and the intercept together give the
surface area by using equation 3.3. Table 4.3 shows the calculated surface area for each of
the four catalysts.

P/[V(Po-P)]

 

0.14

0.12‘

0.10“

0.08 ‘

0.06 "

0.04 '

0.02 '

 

 

 

 

0.00 . ' .
0.00 0.10 0.20

P/Po

Figure 4.6 Typical BET plot for a supported and promoted iron phosphate

catalyst

55

Table 4.3 Surface areas of catalysts

 

 

 

 

 

 

 

 

CATALYST SURFACE AREA, m2/gm
Base iron phosphate 3 2.66
Base iron phoghate-ILUDOX‘D HS 40% 2.73
Base iron phosphate+Ce promoter 3.20
Base iron phosphate+Ce+LUDOX® HS 40% 6.38

 

It is to be noted that the surface area of all the iron phosphate catalysts are
extremely low, less than 10 m2/gm. Even the catalyst containing a silica support did not
show an increase in the surface area. This shows that it is not possible to make catalysts

with high surface areas.

4 .4 .2 Bulk compositions:

Bulk compositions of the four iron based catalysts were obtained by analyses
conducted at Galbraith Laboratories, Inc. The assays used are described in the previous
chapter. The only drawback of the assay used is that it is impossible to obtain the

concentration of oxygen in the presence of other metals like iron.

Table 4.4 gives the bulk composition of the catalysts. As expected, the
proportions of iron and phosphorus are approximately the same, whereas the amount of
nitrogen is quite low. This indicates that the catalyst preparation reaction went almost to
completion and that all the nitrates were substituted by phosphates forming nitric acid.
Since no other elements were introduced initially, it is reasonable to assume that the

remaining percentage is all oxygen.

 

*9. 3 0593.60

 

 

 

 

 

8.8 .. 8.» 2a :5 v 3.3 8.8 + 3298.... .35 use
snag oo

3.: .. - 8; 8 v 3.: 2.8 +23%...“ :3 §m
$9. a: Guam—DA

3.3. ... one - 3 v 8.3 2.: + 32%.... as use

8:. a - - 3 v 2.3 2.2 ragga :3 8.5

s s e s s s
2838 zoos—m See—mo 252.22 3.558.: 29: 55:5

 

 

 

 

 

 

 

3:33 2: he 33:83:89 .52: 3...: v... 035—.

 

4.4.3 ' Surface compositions:

Surface compositions were measured using X-ray photoelectron spectroscopy.
XPS data were obtained on a Perkin-Elmer Surface Science instrument equipped with a
model 10-360 precision energy analyzer and an omnifocus small spot lens. X-rays were
generated by an Al (1486.6 eV) anode operated at 15 kV and 20 mA. xps binding energies
were referenced to the C Is line (284.6 eV) and were measured with a precision of :t 0.2
eV or better. The samples were outgassed at 400 ‘C for 24 hours to remove any surface

moisture accumulated during storage.

Table 4.5 gives the relative surface compositions of each of the four catalysts
relative to the amount of iron on the surface. It is to be noted that in almost all cases, the
surface composition is quite different from the bulk composition. The concentration of iron
sites on the surface is much less as compared to the other elements. The concentration of
phosphorus atoms is at least five times greater than that of iron at the surface, whereas the
concentration of oxygen is more than ten times that of iron. It is to be noted that in all
catalysts, there is a significant amount of carbon observed. This is carbon adsorbed onto
the surface from hydrocarbons present in air, a general contaminant in XPS. The fact that
surface compositions differ so widely from the bulk compositions suggests that either a
better control needs to be implemented during the preparation of the catalyst in order to
ensure a homogeneous distribution of all the elements throughout the catalyst, or that it is

simply not possible for iron to be in larger quantities on the surface.

4 .4 .4 Bulk crystal structure determination:

X-ray diffraction techniques were used to try and identify the various crystal
phases believed to be present in the bulk catalyst. Although the catalysts proved to be
extremely crystalline, no single phase was uniquely present in any of the catalysts. As the

spectra obtained show, a mixture of phases is present in all catalysts.

 

as. a: axon—3+6

 

 

 

 

 

 

 

 

 

 

 

 

3. m. S 35 and n _ + seaweed 8: 3.8
88:55 no
2.. an - a... :3 : _ + engage 8a can
*8 m: axon—3
a... 3 5 - an... a. _ + 22.5.... 8: use
3 2 - - and a _ 32%.... .5: can
22:20 2828 2835 2250 23222 @3853...— zS: $555

 

58:3 :9: e. 03.20.. 89:83 2: he 23.6353 3.2: coatam m... 035.

 

$

X-ray diffraction patterns were obtained using a Rigaku difiractometer equipped
with a rotating anode and using Cu-Ka radiation. An operating voltage of 45kV and an
operating current of lOOmA was used. The slit sizes used were: DS = 0.5, SS = 0.5, RS =
0.3, and RSm = 0.45. X-ray diffraction spectra were obtained for all the four catalysts
calcined at two different conditions. One set was calcined at 110 'C while the other was
calcined at 500 'C as described earlier. Figure 4.7a shows the diffraction spectrum for a
catalyst calcined at 110 'C and Figure 4.7b shows the same catalyst calcined at 500 'C. An
inspection of the two indicates that at a higher temperature, although the degree of
crystallinity does not change much, new phases are created which replace existing
crystallographic phases. This indicates that calcination temperature is an important
parameter in the preparation of the catalyst. The diffraction spectra for all the other catalysts
are included in Appendix B.

An attempt was made to identify the various crystallographic phases present by
comparing the d-spacings obtained from the spectrum to those of known crystallographic
phases available in literature. The one phase common in all the catalysts is Fe(PO3)3; Iron
(III) metaphosphate. Other forms of Fe(PO3)3 are also present along with a number of as
yet unidentified phases. A study of these diffraction spectra is evidently important as it will
provide a yardstick to check the reproducibility of future catalyst preparations.

U. a: .a .3523 .8228 8233.... .3... an .e 5.58% 5.325.. 533" a: 9:53

SQFN

a... “a... "3... a .. ha.” a n.3,... “a... “a... Ella... r; j

 

 

 

 

 

 

 

U. can .a 92:28 .3383 8233.... .5... an be 5:33..» 5.82:... ~33" 5.... 9...»:—

8EN

an; {iii an; .35; ..atra 6;
l n t a, . a ~ ~ . ~ -
E. l. l ll 0. t. I- t. I ‘

 

 

 

l
puoocs rad stunoj

 

 

5. CONCLUSIONS

Conventional commercial dehydrogenation catalysts offer little promise in
achieving the conversion of succinic acid to maleic anhydride under oxidative
dehydrogenation conditions. The selectivities achieved are extremely poor which makes

these catalysts non-competitors in the attempt to find a catalyst suited for commercial

production.

This work demonstrates that unlike the commercial catalysts, the molybdenum
based catalysts show a high degree of selectivity, although the conversions obtained are
very low; none of them exhibiting a conversion greater than 10%. Also, the inherent
instability of these catalysts at higher temperatures render them unsuitable for use at

temperatures in excess of 4(1) °C, which may be necessary.

The results obtained with the iron phosphate catalysts are very promising.
Experiments were conducted using the base iron phosphate catalyst, the base catalyst
promoted by cerium, the base catalyst with a silica support and the base catalyst with both
the promoter and the support. In all cases a conversion of at least 40% has been achieved.
High selectivities (>95%) were also achieved which makes these catalysts extremely viable
for the commercial production of maleic anhydride from fermentation feedstocks.

The high conversions and selectivities obtained are attributed to the stability of
the five membered ring structure of succinic anhydride. This stable structure minimizes the

opportunity for decarboxylation to occur.

63

From the experiments conducted, the iron phosphate based catalysts indicate
high conversions and selectivities even in the presence of copious amounts of water. Also, I
it is evident that large amounts of pure oxygen are not required for the reaction to take place
and that on an industrial scale, atmospheric air should provide sufficient oxygen for the

reaction, thus favoring the economics of the entire process.

The experiments also indicate that the base iron phosphate catalysts achieve as
good conversions and selectivities as the ones supported by silica, indicating that the
support does not play a significant role. There seems to be an advantage in using the cerium

promoter but further studies are required for confirmation.

Characterization experiments were conducted in order to detemrine the physical
attributes of the iron phosphate based catalysts. The surface area of the catalysts were
measured which indicates the extremely low active surface area available for the reaction,
on the order of less than 10 ngm'l. Bulk composition analyses were conducted which
were then compared with the surface composition data obtained by XPS. In all cases, the
surface composition was quite different compared to the bulk compositions. X-ray
diffraction experiments were conducted in order to determine the crystalline phases present
under different calcining conditions. New phases were created when the catalysts were
calcined at a higher temperature of 500 'C. A mixture of different phases was present in
each catalyst. Although some of these phases could be identified by comparing with
standard known phases, there were some new phases present which could not be

identified.

6. RECOMMENDATIONS FOR FUTURE
WORK

Ithasbeenesmbfishedindrisworkthatthenonphosphatescamlystsamindeed
suitable for the production of maleic anhydride fiom succinic acid. Certain other operating
parameters like temperature, contact time etc. have also been optimized. However, for the
catalytic process to be adapted to industry, it is also necessary to optimize the conditions
under which the catalysts are prepared and characterized as well as all the design and
operating parameters in the overall process.

Out of the four iron phosphate based catalysts tested in this work, it was not
conclusively proven which was the best suited for the reaction under study. Additional
experiments will have to be carried out to determine the most suitable catalyst. The
conversions obtained using a more expensive promoted catalyst have to be evaluated vis-a-
vis the conversions obtained from an unpromoted catalyst to decide whether the increase in
costisjustifiedbyan increaseirrproductyield. Otherpromoterslikecesiumwillhavetobe
tested instead of cerium to see if a higher conversion can be achieved. At present, the use of
a silica support does not indicate any increased performance compared to an unsupported
but promoted catalyst.

Once the catalyst that gives the highest conversions and selectivities is
determined, it is essential to find out out the rate equation for the catalyzed reaction. Simple
kinetic experiments will be needed to determine the rate equation and the constants therein.
The rate equation may nrrn out to be of a complex order, but in any case, it will provide

valuable information on the theoretical conversions attainable under different conditions.
64

Theparticle sizeand shapeofthecatalysttobeusedisanimportantparameter
to be studied. These decide the extent of diffusion and mass transfer resistances on the

smfaceandalsothepressuredropacrossthebedofcatalysts.

Certain conditions timing the synthesis of the catalysts need to be optimized.
Although the conditions used in this work appear to be reasonably well suited, further
catalyst preparations under different conditions will be neccessary to determine the best
possible set of conditions. Of specific interest will be the conditions under which the
catalystsarecalcined. Ithasbeenestablisheddratcalciningatahighertemperauueresultsin
achangein thecrystal structureofthecatalyst. Furtherexperimentson thesamelineswill
havetobedonetodeterminethemostfavorablecrystal str'ucurreofthecatalyst.

Bulk elemental analysis as well as surface composition analysis using XPS, as
was done in this work, will be an invaluable tool providing direct information on the
distribution of the various elements throughout the catalyst material

Further characterization work will have to be done in order to ensure the
reproducibility of the catalyst synthesis. Powder X-ray diffraction experiments will indicate
whether every crystal phase is reproduced each time or not. These experiments can also
determine the metal crystal sites as well as identify any contaminant phase in spent

catalysts.

The technique of Mossbauer (or nuclear gamma resonance) spectroscopy
provides information on such parameters as isomer shift, quadrupole splitting and magnetic
hyperfine splitting. These parameters allow a detailed analysis of the chemical state of any
desired atoms within a material and as such will be utilized to provide a more complete
picture of the iron phosphate catalysts. For example, the oxidation states of iron before and
after the catalytic reaction can be determined. Mossbauer spectroscopy can also provide a

viable means of investigating the exact role of promoters. Jones ( 1980) has discussed some
of these applications with iron oxide catalysts as an example.

Phosphorus NMR (nuclear magnetic resonance) can also provide a very

accmate measure of the concentration of phosphorus on the surface and in the bulk catalyst.

Although catalytic activity, selectivity and the deactivation rate are important
criteria for developing and evaluating commercial catalysts, mechanical properties such as
catalyststability,attritionresistanceandcrushingstrength arejustasimportant(Bertolacini,
1989). For this reason it is essential that the catalysts under study he tested for single pellet
and bulk Crushing strengths. The exact form in which the catalysts are to be manufactured,
that is, whether they should be extruded, pelletized, etc. has to be decided upon. Also, if
pellets are to be made, it has to be decided if the pellets will be spherical or cylindrical.
Standardized procedures need to be developed for the industrial scale manufacture of the

catalysts.

As discussed previously, the ultimate aim of this project is to produce maleic
anhydride from a fermentation derived succinic acid. All the experiments carried out in this
study utilized pure succinic acid procured from the market. It is therfore necessary to test
these catalysts on succinic acid obtained directly from a fermentation broth. The broth will
contain certain impurities like amino acids, other proteins and acetic acid, and the
performance of the catalyst in the presence of these impurities will have to be evaluated. In
this work, feed concentrations of 40 g/L and 80 g/L were tested. These are the expected
range of concentrations in a typical fermentation broth. Of special interest will be the
performance of the catalyst at much higher concentrations. If significantly higher yields are
achieved, the ability of these catalysts to perform at higher concentrations may provide an
economic advantage in which case the feed will have to be concentrated before being fed

intothereactor.

67

Atmosphericairhastobetestedtoseeifitcanbeusedinsteadofpureoxygen
inordertocutcosts.Thefeedtoairratiotobeusedhastobeoptimized.

It is also essential to determine the catalyst life without the need for
regeneration. In this work. the catalysts showed virtually no deactivation at the end of one
hour. However, further tests will have to be carried out to determine the lifetime of the
catalystinacontinuousoperation ofthereactor.Oncethatisknown,testswillhavetobe
carried out to determine if the catalysts can be regenerated to their original state.
Reoxidation and steam treatment will be the two primary methods to regenerate the

catalysts.

Finally, the reactor design itself has to be optimized. While a fixed bed
operation seems to work reasonably well, it might be more advantageous to use a fluidized
bed, especially if frequent regeneration is required during a continuous operation. In that
case a parallel can be drawn from the fluidized catalytic cracking (FCC) column widely
used in the petroleum industry, wherein regular regeneration is practiced in a continuous
process. A large pilot scale reactor will be required to make a realistic economic survey of
the overall process. An economically favorable survey will be an indication of an optimum
reactor design which has to integrate the fermentation process to produce succinic acid and
the catalystic conversion technology developed in this work to convert the succinic acid to

rmleic anhydride.

 

APPENDICES

 

APPENDIX A

LABORATORY SYNTHESIS PROCEDURE FOR THE

CATALYSTS

Iron phosphate based catalysts:

Iron phosphate

The base iron phosphate catalyst was prepared by dissolving 48.5 grams of
Fe(NO3)3-9H20 (iron nitrate nonahydrate) and 15 ml of H3PO4 (85%) in 120 ml of
water. The solution was refluxed with constant stirring and heating; after 2 hours of
refluxing, a creamish slurry was formed. Refluxing was carried on for about 20
hours, after which the slurry was distilled to form a thick paste. The paste was
dried in an oven at 110 °C for about 24 hours. The catalyst was calcined at 450 °C
for 6 hours in a forced air oven. A second calcination was performed in a quartz
reactor tube at 500 °C for 2 hours with an oxygen flow rate of 10 ml/min and a

nitrogen flow rate of 10 ml/min.

Iron phosphate with silica support

The base iron phosphate catalyst with a silica support was prepared exactly as in
preparation 1, but with the addition of 10 ml of Ludox" HS 40% to the aqueous
solution. The catalyst was calcined in a similar fashion.

Cerium promoted iron phosphate
The base iron phosphate catalyst with a promoter (5% cerium) was prepared exactly

as in preparation 1, but with the addition of 2.5 grams of Ce(N03)3~6H20 (cerium
$

to

nitrate hexahydrate) to the aqueous solution. The catalyst was calcined in a similar

fashion.

Lanthanum promoted iron phosphate
The base iron phosphate catalyst with a lanthanum promoter was prepared exactly
as in preparation 1, but with the addition of 9.9 grams of lanthanum pentaoxide to

the aqueous solution. The catalyst was calcined in a similar fashion.

Cerium promoted iron phosphate catalyst supported on silica

The base iron phosphate catalyst with a promoter (5 % cerium) and a silica support

was prepared exactly as in preparation 1, but with the addition of 2.5 grams of
Ce(NO3)3-6H20 and 10 ml of Ludox" HS 40% to the aqueous solution. The

catalyst was calcined in a similar fashion.

Lanthanum promoted iron phosphate catalyst supported on silica

The base iron phosphate catalyst with a lanthanum promoter and a silica support
was prepared exactly as in preparation 1, but with the addition of 2.5 grams of
lanthanum pentaoxide and 10 ml of Ludon HS 40% to the aqueous solution. The
catalyst was calcined in a similar fashion.

Molybdenum oxide based catalysts:

The first catalyst was prepared by adding 0.71 ml of H3PO4 (85%), 1.17 grams of
NH4VO3, 19.5 grams of (NH4)2M0207, and 0.48 grams of Cu(NO3)°2.5H20 to
300 ml of water. A fine suspension was formed. The water was evaporated off to
form a thick paste. The paste was dried in air in an oven for 12 hours at 155 °C.
The catalyst was calcined at 400 °C in a forced air oven for six hours. A second
calcination was performed in a quartz reactor tube at 400 °C for 2 hours with an

oxygen flow rate of 10 ml/min and a nitrogen flow rate of 10 ml/min.

70

The resulting catalyst was composed of molybdenum, phosphorus, vanadium,
copper, and oxygen.

The second molybdenum oxide based catalyst was prepared by adding 1.6 ml of
H3PO4 (85%), 30 grams of M003, 211 grams of vzos, and 10.7 grams of we, to
200 ml of distilled water. The suspension formed was refluxed for 72 hours. It was
filtered after being cooled to room temperature. The filtrate was evaporated to
dryness at 120 °C, and dried in an oven at 115 °C for 16 hours. The catalyst was
calcined at 400 °C in a forced air oven for six horns. A second calcination was
performed in a quartz reactor tube at 400 °C for 2 hours with an oxygen flow rate

of .10 ml/min and a nitrogen flow rate of 10 ml/min.

The resulting catalyst was composed of molybdenum, phosphorus, vanadium,

tungsten, oxygen, and hydrogen.

The third molybdenum based catalyst was prepared by adding 27 grams of
' (NI-I4)6Mo7024-4H20, 8.1 grams of (NI-I4)zCe(NO3)5, 1.73 grams of KOH. 3.2
ml of H3PO4 (85%) , and 30 ml of concentrated HCl to 30 ml of distilled water.
The resulting slurry was heated at 125 °C to a thick paste. The paste was dried in an
oven at 115 °C for 18 hours. The catalyst was calcined at 400 °C in a forced air
oven for six hours. A second calcination was performed in a quartz reactor tube at

400 °C for 2 hours with an oxygen flow rate of 10 ml/min and a nitrogen flow rate
of 10 ml/min.

The resulting catalyst was composed of molybdenum, cerium, potassium,

phosphorus, and oxygen.

APPENDIX B
X-RAY DIFFRACTION SPECTRA OF THE CATALYSTS

The X—ray diffraction spectra of the base iron phosphate catalyst calcined at 110 °C
and at 500 'C were shown in Chapter4. Thediffraction spectraforalltheothercatalysts
are compiled in the following pages.

D. e: .a 90538
3.3 m: 6 x25.— 3 3.32... 3:58 8358.... .3... .3 .e 5.58... 5.82:... ~33" :— 0.52...-

ESPN

a”. a $3... a .. .3”. all! ... .3.” “a... pa”. 213.. 5:! III.

 

 

1
puooes rad srtrnog

 

 

0. Sm .. 3.5.8
3.3 m: a x25.— 3 2:83... .538 22.38.... :2. a. 2. 5.58... 5.89.5... 5: a... 8......—

505k

 

 

 

 

pm 19:! smnog

 

 

74

U. a:
.a .8523 “Eaton 5.: 3.2.3.... 2288 32.38.... .5... ea 3 5.592.- 539.5529 533" 9.: 9...»...—

805. N

.. . .. . fiaéllifi a . a:
anar.agaa§.; .. L L.
a

 
 

 

 

I
puooas 13d smnog)

 

 

 

0. can
.a .3528 2.5.3.. 5... 90.2.3... 35.3.3 0.2.58...— ....... .3 .c .558... 5.88:... ha...” v... 9.3....

Sofia

a... a .. “a... a .. .3”. .3 .3... “a .. “a... 51'! p. .3

l
. pnooos .xad sumog

 

 

 

U. 3. .a .8528 2......8 a...» 3.3.8.. 9...
.2... a: e .89.... 3 3.8.... .558 3.....8... .8... 5. .e 5.58... 5.8.5.... 5.... n... 8......

80.58

a... “a ... “a” a .. .au .3 ... .3” “a... Na... . are; 3

 

 

 

pnooas 13d sumog

 

 

 

 

 

.0. can .a .3523 2.5.3.. 5.: 99.2.8... 9..-
.2... a: e .82... 3 33...... .538 22.8.... .3... 5. 2. E332... 5.8.5.... 5.... n... 2......
$2.... a

a... a 4.3.; a. a .35; cant-II! SI!
3 . v . . . s . .
II ‘0 I ll ll ll II I. I I

a

 

puooas 39d sumog

 

 

LIST OF REFERENCES

LIST OF REFERENCES

Bertolacini, R.J. (1989). Mechanical and Physical Testing of Catalysts.
Characterization and Catalyst Development- An Interactive Approach. (Bradley,
S.A., M.J.Gattuso, and R.J.Bertolacini, Eds.), p. 380, ACS Symposium Series
411, Washington DC.

Bradley, S.A., E.Pitzer, and W.J.Koves. (1989). Bulk Crush Testing of Catalysts.
Characterization and Catalyst Development- An Interactive Approach. (Bradley,
S.A., M.J.Gattuso, and R.J.Bertolacini, Eds.), p. 398, ACS Symposium Series
411, Washington DC.

Cohen, LB. (1990). X-ray Diffraction Studies of Catalysts. Ultramicroscopy, 34:
41-46.

Cooley, SD. and J.D.Powers. (1983). Maleic Acid and Anhydride. Encyclopaedia
of Chemical Processing and Design, Vol. 29 (McKetta 1.1., and WA.Cunningham
Eds.), p. 35-55, Marcel Dekker Inc., New York.

Cullity, 3D. (1956). Elements of X-Ray Difi"raction, Addison-Wesley, Reading,
Massachusetts.

Daniel, C. (1982). Oxydehydrogenation Process for Preparing Methacrylic acid and
its Lower Alkyl Esters. U .5 .Patent 4,355,176.

Daniel, C. (l983b). Oxydehydrogenation of lsobutyric acid and its Lower Alkyl
Esters. U .S.Patent 4,391,989.

Daniel, C. (l983c). Oxydehydrogenation Process. U .5 .Patent 4,410,723.

Daniel, C. (1983d). Oxydehydrogenation Process for Preparing Methacrylic acid
and its Lower Alkyl Esters. U .S Patent 4,410,729.

Daniel, C. (l983c). Oxydehydrogenation Process. U .5 .Patent 4,410,728.
78

79

Daniel, C. ( 1984). Oxydehydrogenation Process. U.S.Patent 4,439,621.

Daniel, C., and P.L.Brusky. (1981b). Catalytic Oxydehydrogenation Process.
U .3 .Patent 4,299,980.

Daniel, C., and P.L.Brusky. (1983a). Catalytic Oxydehydrogenation Process.
U.S.Patent 4,374,268. '

Daniel, C., and P.L.Bursky. (1981a). Catalytic Oxydehydrogenation Process.
USPatent 4,298,755.

Delannay, F. (1984). Transmission Electron Microscopy and Related
Microanalytical Techniques. Characterization of Heterogeneous Catalysts.
(Delannay, F., Ed.), p. 71, Marcel Dekker, New York.

Delgass, W.N., G.Haller, R.Kellerman, and J.H.Lunsford. (1979). Spectroscopy
in Heterogeneous Catalysis, Academic Press, New York.

Edmonds, T. (1980). The Characterization of Industrial Catalysts with ESCA.
Characterization of Catalysts. (Thomas, J.M., and R.M.Lambert, Eds.), p. 30,
John Wiley and Sons.

Haensel, V., and H.S.Haensel. (1989). The role of Catalyst Characterization in
Process Development. Characterization and Catalyst Development— An Interactive
Approach. (Bradley, S.A., M.J.Gattuso, and R.J.Bertolacini, Eds.), p. 2, ACS
Symposium Series 411, Washington DC.

Hightower, LW. (1990). Kinetics and Mechanism. Lecture notes from
'Applications of Heterogeneous Catalysis', University of Houston, Houston, TX.
December 3-7, 1990. p. C-34.

Howie, A. (1980). The Study of Supported Catalysts by Transmission Electron
Microscopy. Characterization of Catalysts. (Thomas, J.M., and R.M.Lambert,
Eds.), p. 12, John Wiley and Sons.

Irving-Monshaw, S., and A.Kislin. (1989). Lively Markets Brighten Maleic
Anhydride Horizons. Chemical Engineering, 96(3): 35.

so

Jacobs, RA. (1984). The Measurement of Surface Acidity. Characterization of
Heterogeneous Catalysts. (Delannay, F., Ed). P. 367, Marcel Dekker, New York.

Jones, W. (1980). Use of Mossbauer Spectroscopy for Catalyst Characterization.
Characterization of Catalysts. (Thomas, J. M., and R. M. Lambert, Eds ). P. 114,
John Wiley and Sons.

Kenney, C.N. (1980). Concluding Remarks - Comments by a Chemical Engineer.
Characterization of Catalysts. (Thomas, J.M., and R.M.Lambert, Eds.), p. 274,
John Wiley and Sons.

Lecloux, AJ. (1981). Texture of Catalysts. Catalysis, Science and Technology,
Vol. 2 (Anderson, J.R., and M.Boudart, Eds.), p. 171, Springer-Verlag, New
York.

Lemaitre, J .L. (1984). Temperature Programmed Methods. Characterization of
Heterogeneous Catalysts. (Delannay, F., Ed.), p. 29, Marcel Dekker, New York.

Lemaitre, J.L., P.Govind Menon, and F.Delannay. (1984). The Measurement of
Catalyst Dispersion. Characterization of Heterogeneous Catalysts. (Delannay, F.,
Ed.), p. 299, Marcel Dekker, New York.

McCarvey, G.B., and J.B.Moffat. ( 1991). The Oxidative Dehydrogenation of
lsobutyric acid to Methacrylic acid on Ion Exchange Modified 12-Heteropoly ‘
Oxometalates. Journal of Catalysis, 132: 100-116.

Millet, J.M., C.Virely, M.Forissier, P.Bussiere, and J.C.Vedrine. (1989).
Mossbauer Spectroscopic Study of Iron Phosphate Catalysts used in Selective
Oxidation. Hyperfine Interactions, 46: 619-628.

Millet, Jean-Marc M., and J.C.Vedrine. (1991). Role of Cesium in Iron
Phosphates used in lsobutyric acid Oxidative Dehydrogenation. Applied Catalysis,
76: 209-219.

Millet, Jean-Marc M., J.C.Vedrine, and G. Hecquet. (1990). Proposal for Active
Sites of Iron Phosphates in lsobutyric Oxidative Dehydrogenation Reaction. New
Developments in Selective Oxidation. (Centi, G. and F.Trifiro, Eds.) Elsevier
Science Publishers B.V., Amsterdam, p. 833-840.

81

Muller-Erlwein, E., and H.Hofmann. (1988). Determination of Kinetic Parameters
from Dynamic Investigations of Heterogeneously Catalyzed Reactions. Chemical
Engineering Science, 43(8): 2245-2250.

Pedersen S.E., N.J.Bremer, and J.L.Callahan. (1984). Iron - Phosphorus Mixed
OxiderCatalysts and Process for their Preparation. U.S.Patent 4,427,792.

Richardson, J.T. (1989). Principles of Catalyst Development. Plenum Press, New
York.

Robinson, W.D., and R.A.Mount. (1983). Maleic Anhydride, Maleic and Fumaric
Acid. K irk-Othmer Encyclopaedia of Chemical Technology, Vol. 14 (Mark et al.,
Eds.), John Wiley & Sons, p. 770-793. '

Ruszala, F.A. (1983b). Oxydehydrogenation of lsobutyric acid and its Lower
Alkyl Esters. U.S.Patent 4,391,990.

Ruszala, RA. (1984). Oxydehydrogenation of lsobutyric acid and its Lower Alkyl
Esters. U .S.Patent 4,434,298.

Ruszala, F.A., and T.J.Weeks. (1983a). Oxydehydrogenation Process for
Preparing Methacrylic acid and its Lower Alkyl Esters. U .S .Patent 4,374,270.

Seamans, J.D., J.G.We1ch, and C.A.Vuitel. (1989). Improved Regeneration
Quality with Length and Density Grading. Characterization and Catalyst
Development- An Interactive Approach. (Bradley, S.A., M.J.Gattuso, and
R.J.Bertolacini, Eds.), p. 2, ACS Symposium Series 411, Washington DC.

Sing, K.S.W. (1980). The Use of Physisorption for the Determination of Surface
Area and Pore Size Distribution. Characterization of Catalysts. (Thomas, J .M., and
R.M.Lambert, Eds.), p. 12, John Wiley and Sons.

Statz, R.J., and J.K.Doty. (1980). Catalyst and Dehydrogenation Process.
U.S.I’atent 4,232,174.

Szmant, H.H. (1989). Organic Building Blocks of the Chemical Industry. John
Wiley and Sons. p. 312.

8‘2

Watkins, W.C. (1974). Catalytic Process for the Manufacture of Unsaturated Acids
and Esters. U .S.Patent 3,855,279.

Watkins, W.C. (1975). Catalytic Process for the Manufacture of Unsaturated Acids
and Esters. U .S.Patent 3,917,673.

Watzenberger, 0., G. Emig, and D.T.Lynch. (1990). Oxydehydrogenation of
lsobutyric acid with Heteropolyacid Catalysts: Experimental Observations of
Deactivation. Journal of Catalysis, 124: 247-258.

Wood, Andrew. (1990). Maleic Anhydride's Growing Pains. Chemical Week,
146(16): 6.

 

MICHIGAN STATE UNIV. LIBRARIES
l“WIWIWIWW““W”WWW
31293010559767