N.
. 2

if... .
.14.?

.-

.1
3:3: 1‘ t
...D..~ i1
1 ‘t is: a

$1913.

2&1:-

a: mo
21...»: t 31.2....
.. x {31:233.}.
.21: 5.: .z‘ .
v $

 

 

2.
5:
5.. £2 1.

.3 .
I!- a .51..
3.1:...

155...... .
Pith.

.. a
l.

3...}!

.1 9?

. {2.5}; x:
2...! a .v
.. {Evil}.

 

. :2. . 4..
338.. é.

7 ‘
.l._b«mm..ma,
3&5... : ;.

n3”;

L.»
.z, «.1... ‘53.:
(.2 6 . 3..
... .. £252.53
2 {2&5}. 3!.
. s 2.

v»!

I
.V)1v7.l:

5::
Alt:

. , “I;
. P, 93.1V’Iiv

.r

 

WIS

Jlifliiliiiiiiiiifiliiiliiimfiliflflfiiflit a

3 1293 01046 0495

This is to certify that the

dissertation entitled

CATALYTIC UPGRADING 0F LACTIC ACID
OVER SUPPORTED SALT CATALYSTS

presented by

Garry Clyde Gunter

has been accepted towards fulfillment
of the requirements for

Ph .D . degree in Chemical Engineering

 

 

Qiiioma’ea

M ajorgrofessor

Date 10-6—94

 

MS U is an Affirmative Action/Equal Opportunity Institution 0- 12771

 

 

LIBRARY |
Michigan State i
University

 

 

 

PLACE ll RETURN BOXtomnwombchoekMfiomywncord.
TO AVOID FINES mum on or bdoro dato duo.

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU IsAn Affirmative Action/Emu Opportunity lm
Was-9.1

 

CATALYTIC UPGRADING OF LACTIC ACID
OVER SUPPORTED SALT CATALYSTS

BY

Garry Clyde Gunter

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY

Department of Chemical Engineering

1994

ABSTRACT

CATALYTIC UPGRADING OF LACTIC ACID
OVER SUPPORTED SALT CATALYSTS

BY

Garry Gunter

Liquid phase conversion of lactic acid over zeolites
was performed in a batch reactor at 240-260°C and 590-960
psig for residence times of 2—3 hours. Products include
acetaldehyde and ethanol at 30-42% selectivities, and lesser
amounts of propanoic acid, acetic acid and acrylic acid.
Conversions are 3.7—5.3% at 240°C based on lactic acid fed.

Lactic acid conversion over supported and neat sodium
phosphate catalysts was performed in a packed bed reactor at
1.4—10 atmospheres, 280-3SO°C and contact times of 0.4-5.0
seconds. Five major reaction pathways were observed: 1)
decarboxylation to acetaldehyde, 2) decarbonylation to
acetaldehyde, 3) dehydration to acrylic acid, 4) reduction
to propanoic acid and S) condensation to 2,3—pentanedione.
Minor products include ethanol, acetic acid, acetone and
acetol. The high—value flavoring ingredient, 2,3-
pentanedione, has never been previously observed as a
product of lactic acid condensation.

Production of acrylic acid is favored at 350°C and 0.4

seconds residence time. 2,3—Pentanedione formation is

p
\r
.
A s
‘3-»1

‘

avarea a

r~v
v-1

R843

C

. .i . 1 .£ , .. . .. .\ .. p ~ ~ . 1.. A. .3 CL r. ..
3 ”J C. .3 I c C t i T. .3 f. i C E S r C 4 a S F.
b; . . .nu r“ w“ .C A.» A: n} w-.. rm J.‘ 4‘ ha h» »L my 54
.3 2“ I .nC ..C Q; J. v. .3. A o .C T 4“ a. S. v; up; .s‘
C . .. e n... C. .3 . .3 A c T. E d c . LL Ci 3 ‘3
.3 ~ _ . . _ wt. . .V AC “i. . A u. «D L . L ,. Au C» u v s Q» .
A C a C 1; .3 w .. .1 1A. a . ”.3 AL S . n .. sh .. a v I P l. «D
C. M“ I mu. .6 . A: .in a». t F. e .3 Q‘
. e 4 I «19 4.» .. a a . in.“ 5L Q. Q» «C AC Li. I c . t
V L w . .~.u J . A v r“ .L T A a: j v \ a .4 wfld .3 Li. 5.. .3.
.3 .~ . «u Pu A C v. .5 J. .3 v . an 2. Lu . .. 2.. is Q. a.
a: Q» .3 s . «G n J a L. n.” :v “L. .5 .‘nH “9. Zu v x . r y n .5 axu

favored at 280°C and 4.0 seconds residence time. Basic
phosphate salts favor acrylic acid and 2,3—pentanedione.
Relatively acidic salts produce more acetaldehyde and
propanoic acid.

A catalyst survey compared various supported sodium
salt catalysts including tetraborate, carbonate, nitrate,
silicate, phosphate, sulfate, chlorate, arsenate, bromate
and molybdate. Very basic salts produce higher selectivities
to 2,3-pentanedione, acidic salts produce more propanoic
acid and moderately basic salts produce more acetaldehyde.
Group V salts (nitrate, phosphate and arsenate) produce 2,3—
pentanedione selectivities up to 64%. Conversion increases
with increasing size of the central atom of the catalyst
salt.

A mixture of condensed sodium phosphate salts is
present on the surface of supported phosphate salt catalysts
and was analyzed using MAS—“P—NMR.

Chemical species adsorbed on model catalyst surfaces
were studied using transmission FTIR. The more basic sodium
phosphate Salts alter the equilibrium between the gas phase
and the surface reactants to produce a greater surface
reactant concentration. This leads to greater conversion and
better product selectivity.

Proposed reaction mechanisms to acrylic acid and 2,3-
pentanedione both involve formation of cyclic transition

states.

Copyright by
Garry Clyde Gunter

1994

r—a
C).
m
L)-
|

-
"av-v vyr -
Uh...er, “Al‘y

, "A ub‘h
I: EVLAVAlu

 

I dedicate this work to my parents, Clyde and Geneva
Gunter, who provided a strong foundation for development in

my youth.

Q'V‘
45L— '

Ie

«v

YP-W A ‘-
as Raul..—

F‘s/‘0.
-y'

g

\.

5-H
.

5
\
5..

I
...;

‘

VA

"J

““3

9-1.:

Y
‘Kc

Acknowledgements

I express my thanks to Dr. Dennis Miller, my major
professor, for help and guidance throughout the development
of this work. Thanks to Dr. James Jackson for helpful
feedback in numerous group meetings. Thanks to Dr. Carl
Lira, Dr. Kris Berglund and Dr. Ramani Narayan for their
service as members of my doctoral committee.

Thanks to Prashant Kokitkar for performing GCMS
analysis related to this work. Thanks to Robert Langford,
Doug Wadley, Lisa Benton and Mike Lussier for excellent
suggestions and collaboration on several projects. Thanks to
Kermit Johnson for teaching me the use of the NMR
spectrometer.

Thanks to the Chemical Engineering Department at
.Michigan State University, the State of Michigan Research
Excellence Fund and the United States Department of
.Agriculture for financial support during the development of

this work.

vi

List of Tat!
List of Fig

Chapter 1.

'_J

1
0L.

‘

‘l A
. ,
¢.‘_"

 

f._l

J

 

TABLE OF CONTENTS

List of Tables xiv
List of Figures xvii
Chapter 1. Introduction 1
1.1. Background 1

1.2. Literature Review 2
1.2.1. Lactic Acid 2

1.2.1.1. Manufacture 2

1.2.1.2. Properties 4

1.2.1.3. Primary Conversion Pathways 4

1.2.1.4. Thermodynamics 8

1.2.2. Manufacture of Acrylic Acid 9

1.2.3. Reactions Related to Acrylate
Production 10

1.2.3.1. Pyrolysis of
‘ d-Acetoxypropionates 10

1.2.3.2. Preparation of Methacrylic Acid 12

1.2.3.3. Oxydehydrogenation of Isobutyric

Acid to Methacrylic Acid 14

1.2.4. Dehydration of Alcohols 15

1.2.4.1. Mechanism of Acid Catalysis 15
1.2.4.2. Dehydration over Zeolite

Catalysts 17

1.2.5. d—Diketones 19

vii

1.2.5.1. Properties and Uses
1.2.5.2. Reactions
1.2.6. Catalysts
1.2.6.1. Properties of Phosphate Salts

1.2.6.2. Studies of Phosphate Salts
using MAS—“P—NMR

1.2.7. Studies of Reaction Systems using
IR Spectroscopy

Chapter 2. Lactic Acid Conversion over Zeolites in a

2

2

2

.1.

.2.

.3.

Batch Reactor

Introduction
Methods

2.2.1. Materials

2.2.2. Calcination Furnace
2.2.3. Catalyst Calcination
2.2.4. Batch Reactor

2.2.5. Batch Reactor Operation
2.2.6. Product Analysis

2.2.6.1. Reproducibility of Product
Analysis

Results
2.3.1. Definitions
2.3.2. Effect of Temperature in Blank Runs

2.3.3. Effect of Y Zeolite Catalyst on
Product Selectivities

2.3.4. Effect of Catalyst Mass on Conversion
2.3.5. Effect of Feed Concentration

2.3.6. Comparison of Catalysts at 240°C

viii

19
20
21

21

24

29

3O
3O
3O
3O
31
32
32
35

38

41
42
42

42

43
45
45

46

2.5.

Chapter 3.

< '3
'4!-
1'2.
‘1"
w
J

 

2.4. Discussion

2

.5.

Summary

Chapter 3. Lactic Acid Conversion over Phosphate Salt

3.

3

3

3

l.

.2.

.3.

.4.

Catalysts

Introduction

Methods

3

.2

.1.

Materials

Catalyst Preparation

3.2.3. Vapor—phase Reactor

3.2.4. Product Collection

3.2.5. Reactor Operation

3.2.6. Product Analysis

Results

3.3.1. Definitions

3.3.2. Reactor Characterization

3.3.3. Catalytic Effect of Phosphate Salts

3.3.4. Studies of Catalyst Loading over
Supported Salts and Neat Salts

3.3.5. Selectivity over Supported
Orthophosphate Salts

3.3.6. Reactions over Bovine Dental Enamel
and Trisodium Phosphate

3.3.7. Effect of Temperature and Residence
Time on Product Yields

Discussion

3.4.1. Mechanisms

3

3.4.3.

.4

.2.

Mass Transfer Limitations

Kinetics

ix

48

49

51
51
51
51
52
53
57
57
59
59
6O
6O

63

68

71

71

73
78
78
85

85

LA)
Lf‘

Chapter 4.

“x
’

J‘-
f~_)

LA.)

Chapter 5

‘UI
J

(.n

 

3

.5.

3.4.4. Catalyst Stability over Extended
Reactor Run

3.4.5. Mass Balance Accuracy

Summary

Chapter 4. Lactic Acid Conversion over Supported Sodium

4

4

4

4

4.

.1.

.2.

.3.

.4.

5.

Salt Catalysts-A Catalyst Screening Study
Introduction
Methods

4.2.1. Materials

4.2.2. Vapor—phase Reactor
Results
Discussion

4.4.1. Catalyst Activity
4.4.2. Acrylic Acid

4.4.3. Propanoic Acid
4.4.4. 2,3—Pentanedione
4.4.5. Acetaldehyde

4.4.6. Acetol

4.4.7. Other Products

4.4.8. Effect of Catalyst Salt Acidity
on Selectivity and Conversion

4.4.9. Periodic Considerations
4.4.10. Carbon Balance

Summary

(:hapter 5. Characterization of Sodium Phosphate Salt

5_.

5

1.

.2.

Catalysts using MAS-”P-NMR
Introduction

Methods

87
93
93

95
95
97
97
97
97
103
103
107
107
107
108
108

109

109
114
118
119
120

120

120

 

L."

(I,

(.11

(J1

U!
I

 

5

5.3. Results

5

.2.

.3

.3.

5.

5.

.1.

Materials

Sample Preparation

NMR Spectrometer

.2

.2

.3.1. Spectrometer hardware and
Data Acquis1tion
2.3.2. Spectrometer Setup
.3.3. Sample Loading
.3.4. Sample Spinning
.3.5. Collection of Free Induction
Decays
.3.6. Reference Sample
Experiments
.4.1. Data Quality
.2.4.2. Variation of Spinning Frequency
.4.3. Determination of Chemical
Shift Anisotropy
2.4.4. Sideband Suppression
2.4.5. Proton Decoupling
Neat Salts
.3.1.1. Calcium Hydroxyapatite
.3.1.2. Neat Monosodium Orthophosphate
.3.1.3. Neat Disodium Orthophosphate
.3.1.4. Neat Trisodium Orthophosphate
.3.1.5. Neat Sodium Pyrophosphate

Supported Salts

.3.2.1.

Supported Monosodium
Orthophosphate

xi

120
121

122

122
123
124

125

126
127
128
128

130

131
131
132
135
135
135
136
145
146
149

149

149

U1
“:3
l J

{I}

(.11

ChaDter 6.

ON
}
0

O‘\
t
n

(37‘

6.3.

f“

( TN

 

5.3.2.2. Supported Disodium
Orthophosphate

5.3.2.3. Supported Trisodium
Orthophosphate

5.3.2.4. Supported Sodium Pyrophosphate

5.4. Discussion

Chapter 6.

6.

6

6.

1.

.2.

3.

5.4.1. Thermal Transformations of Sodium
Phosphate Salts

5.4.2. Effects of Catalyst Support on
Thermal Transformations

5.4.3. Comparison of NMR and FTIR Results
Identification of Surface Species on
Phosphate Salt Catalysts using
Transmission FTIR

Introduction

Methods

6.2.1. Materials

6.2.1.1. Chemicals
6.2.1.2. IR Sample Disks
6.2.2. Sample Preparation
6.2.3. Post—reaction Studies
6.2.3.1. Apparatus
6.2.3.2. Procedure
6.2.4. FTIR Spectroscopy
6.2.4.1. IR Spectrometer
6.2.4.2. Collection of Spectral Data
6.2.4.3. Data Reproducibility
Results

6.3.1. FTIR of Pure Substances

xii

155

158
160

163

163

166

167

169
169
169
169
169
170
170
173
173
176
177
177
178
179
180

180

(j \
“is
a

Chapter 7 .

\J

'._1

(TN

 

6.3.2. FTIR Post-reaction Studies
6.3.2.1. Monosodium Phosphate
6.3.2.2. Disodium Phosphate
6.3.2.3. Trisodium Phosphate

6.4. Discussion
6.4.1. Reactions on Catalyst Surfaces

6.4.2. Application of Findings to Results of
Lactic Acid Conversion

6.4.3. Equilibrium of Proton Exchange
Chapter 7. Summary and Recommendations
7.1. Summary
7.2. Recommendations
Appendix A Calculation of Macropore Size in Y Zeolites

Appendix B Calculation of Chemical Shift Anisotropy
from Spinning Sideband Intensities

List of References

xiii

183
183
188
190
192

192

193
194
196
196
198

202

206

208

 

I . I . . . .

.. .. . . . ... .. .y... _... . . A. 7. .1 a. a. .3 a. . 4;... . e .2 A.

e c C C E at r C n1. : 2. CL 2.. a. r E a . ..... C n1... r a .. a 5 DH QC

3 ; 3 C. .. . .l e 5.. C. c L c . c L E : r 3 an .d 7. .C t _ QC LL c a

e .d C c . .C .n .u a PC E f r. c . c l V. c r .1. C out O my a .1 a : s c t

n . U...“ mi CL ob C :U Fe B“ E .1 .2 a: O E no C 3.. Pt n! C Y C C E
a h ale. «.3.. I} «I. fill.— 1‘v In». ~Na F0 H I: #6 fl ~ fi/w «.14 4. ~\Jo
“ .. l. at-“ l 2 fire «x.— a/u a {L «In .14 . 19 6.4 3.. 3 «J v. «44.

LIST OF TABLES

Heats of Reaction and Equilibrium Constants for
Major Reaction Pathways

Topology of Y and ZSM-S Zeolites

Effect of Phosphate Unit on Difference in
Shielding Components 022 and 033

Chemical Shifts and Tensors Reported in Literature
Batch Reactor Operating Conditions

GC Operating Parameters

Reproducibility of GC Analysis

Effect of Temperature on Product Distribution
in Blank Runs

Effect of Y—82 Zeolite on Product Yields

Effect of Catalyst Mass on Conversion
over Y Zeolites

Effect of Feed Concentration on Product
Distribution over Y-62 Zeolite

Conversion over Zeolites at 240°C
Reactor Operating Conditions

Comparison of Results in Stainless Steel and
Quartz Liner Tubes

Catalytic Effect of Si/Al Support on Product
Yields for Lactic Acid Conversion

Catalytic Effect on Product Yields of Addition
of Nafixh to Si/Al Support

Effect of Catalyst Loading and Support

xiv

18

25
27
37
4O

42

44

45

46

47
48

55

62

64

65

69

n
p
\
z-
\
x.
V.
s
§
"‘
\A‘

 

. u A dull u 4 A 9 .V‘ -‘ .
> . .n . . _ . r. AV ~ . «o C“. . . .
v o . t ‘ A_v > u u ‘ .“ Vs r.
C .r. 4. A; at v. .3 l w. «u . : . .
a e C > c A» . c .«u e . vH. «v “V ”at“ NC. “J m . AFC». . «a: C a. by». «Wu; m».
G» .0 Av .3 "In my .3 A C u o a « ~ I . . ~ .
Sn 2. l” T. E u“. C D. l.. 2. E C l. T. t. S E :. 2. S
6 «M U 1i 0 n v n o 0 a a n u 0
37 11 . . . . Ala fix. .s. .3. 4.. «L .4.. In :J raw
.1. :4 1.3 1.... 4.... 4s. as. :o :4 :J :o C . :c

11d

ht

.10.

.11.

.10.

.11.

Product Yields over Phosphate Salts

Conversion of Lactic Acid over Nafixh on Si/Al and
Bovine Dental Enamel

Lactic Acid Conversion over NafiXh on Si/Al

Reaction Rate Constants of Lactic Acid Conversion
Pathways

Activation Energies and Frequency Factors for
Lactic Acid Conversion

Extended Lactic Acid Run over Calcium
Hydroxyapatite Catalyst

Catalyst Salts Surveyed

Periodic Relationship of Catalyst Salt Central
Atoms

Product Distribution over Sodium Salt Catalysts

Effect of Electronic State of Catalyst Salt
Central Atom on Conversion at 280°C

Analysis of Carbon Balance Error

Typical Parameters for Normal Data Acquisition
Reproducibility for NMR Spectra of NaHfiKh
Signal to Noise Ratios (S/N) for NMR Spectra

Effect of Peakwidth on Chemical Shift
Reproducibility

Parameters for TOSS Pulse Sequence

Spin-Spin Coupling Constants

Chemical Shifts of Calcium Hydroxyapatites
Linewidths of 31P Signals

Chemical Shifts of Monosodium Orthophosphates

Conversion of Phosphate Units to Thermally
Stable Species for Neat NaHZPO4-HZO

Chemical Shifts of Disodium Orthophosphates

XV

72

74

8O

89

89

91

96

96

104

117
119
123
129

129

130
133
135
138
138

143

145

148

 

I p
‘ ._ .. > \. .. ». \. . rt. 49V ‘v. a:
w“ mw v. .n“ r. e c :1 hr n-.. to v; r.
e e r“ A t n C v . C n-.. r C
. . .n“ O #1 Q“ .d Ex. CA a. h“ K C a“
C Pt C an r l . o . . Ti 3.. D. . . Ci 7 ..
fil/U 3 in, u I c a 0 u
1 l . .. 1 . «1o .1. Ir». 1.1 34
Pu .- :4 -3 fva Fnrv rhu {RM 1H“. \n...

.12.
.13.

.14.

Chemical Shifts of Trisodium Orthophosphates
Chemical Shifts of Sodium Pyrophosphates

Conversion of Supported Monosodium Phosphate
with Calcination

IR Frequencies and Band Assignments of
Lactic Acid and Sodium Lactate

IR Frequencies of Sodium Orthophosphates
IR Frequencies of Condensed Sodium Phosphates

Percentage of Sodium Lactate on Trisodium
Phosphate

Incremental Pore Volumes Greater than 20%
for Y-82 Zeolite

Incremental Pore Volumes Greater than 10%
for Y-62 Zeolite

xvi

151

153

158

182
186
186

192

205

205

 

at .C F; a; at .3 :
n v fly by nu . ~ . la
a r ml a 3 n C E
L «U. VhU YL .. .n 1.5. “aw.-
4’. A14 3‘... In“. 03 PO
A1 . 4|. 1 11¢ . .‘

. x r g
.. . r a C
‘fiq C... 3 .
do L n .3
n5» n A PC
all ORV 4|;
«:4 . § Al/L

Q.» Cu
:4 .w«
.1 V.
DC m .
«‘J
«J;

‘ QC
a...

I A.
2 O
h:

«14

n? a C
C C
r a
BC Tu

w. \6 L V

A-» a.»

.r .C.

P TL
NI]
\. w

LIST OF FIGURES

Lactic Acid Synthesis via Lactonitrile

Products of Lactic Acid Conversion Reported
by Odell

Lactic Acid Reaction Pathways Observed by
Mok, Antal and Jones

Routes for the Production of Acrylic Acid
Acid Catalyzed Alcohol Dehydration Mechanism
Reactions of d—Diketones

Sodium Phosphate Salts

Thermal Transitions in the System NaJLIyO§IgO
Calcination Furnace

Batch Reactor

Vapor—phase Reactor

Reactor Flow Diagram

Effect of Temperature on Product Yields over
Trisodium Phosphate on Si/Al

Acrylic Acid Yield over Trisodium Phosphate
on Si/Al

2,3-Pentanedione Yield over Trisodium Phosphate
on Si/Al

Proposed Mechanism for the Dehydration of
Lactic Acid to Acrylic Acid

Proposed Mechanism for the Condensation of
Lactic Acid to 2,3—Pentanedione

xvii

11
16
20
22
23
33
36
56

58

75

77

79

81

82

«(J

14 1 A
.2 9
-~ nxu
l
.u.

i. . .
c. e
—__ axe
2
9:».

C . . l
C. E
a: Cu

.13
3;».

i a.
£2 a .
.I a.
f. C.»

.2 Au
.2 T.
c. C
a: “3
r3
1.

\v
‘1‘

«Ii

:J

.3
‘1.

m.

«IQ
E.

\l‘ 3"
uno‘

‘ . .
«\w «\u mC
Q . o
~
Q. “L Q
‘ a I. a I.
.k v rho n/
:9 C). :4

 

 

 

 

. . C. . I. .. 2.. . S .I_ .3 _
a a 1 FL
1 3 l S C S No t i 3. out.
«C C Cw Q «C C Q. 11 Cr
1. L. 1 .F 1. .7. .P. d C. 2.
n}. E. v.2 C. v.2 P m 1. S 1.. T ..
o ‘ FYIIJ *1 \ I I I
.‘Hv nuu. a. .« 1 i a s «,4 a.“
:4. :9. E . .3 [C an . CU

.10.

.11.

Reaction Pathways of Lactic Acid Conversion
over Na3PO4/ Si /Al

Arrhenius Plot for Primary Reaction Pathways of
Lactic Acid

Effect of Catalyst Acidity on Maximum
Selectivity to 2,3-Pentanedione

Effect of Catalyst Acidity on Maximum
Selectivity to Propanoic Acid

Effect of Catalyst Acidity on Maximum
Selectivity to Acetaldehyde

Effect of Catalyst Acidity on Product
Selectivities at 300°C

Effect of Size of Catalyst Salt Central Atom on
Bond Angles in Cyclic Transition States

Normal Data Acquisition Sequence

Total Sideband Suppression (TOSS) Pulse Sequence
MAS-“P-NMR Spectra of Calcium Hydroxyapatites
MAS—“P-NMR Spectra of Neat Monosodium Phosphates
MAS-“P—NMR Spectra of Neat Disodium Phosphates
MAS-“P—NMR Spectra of Neat Trisodium Phosphates
MAS—“P—NMR Spectra of Sodium Pyrophosphates

MAS-”P-NMR Spectra of Supported Monosodium
Phosphates

MAS-nP-NMR Spectra of Supported Disodium
Phosphates

MAS—”P-NMR Spectra of Supported Trisodium
Phosphates

Thermal Transformations of Calcined Phosphate
Salts

Apparatus for Post—reaction Studies
IR Spectra of Lactic Acid and Sodium Lactate
IR Spectra of Sodium OrthOphosphates

xviii

88

90

110

112

113

115

117
123
133
137
141
147
150

152

156

159

161

164
174
181

184

a:

Du
. l

A“.

:g

R
..

All

p. _
n v
hr
ob
f9.

9.

«V

1A“

 

 

IR Spectra
IR Spectra
IR Spectra

IR Spectra

of Condensed Sodium Phosphates
of Lactic Acid on Monosodium Phosphate
of Lactic Acid on Disodium Phosphate

of Lactic Acid on Trisodium Phosphate

Proton Exchange and Condensation Reaction Schemes

Pore Size Distribution in Y—82 Zeolite

Pore Size Distribution in Y-62 Zeolite

xix

185
187
189
191
193
203

204

Chapter 1 .

1.1. Backg:

~ .
"VA”. A ‘
,a

..— ‘
\r¢¢~...~-v-A- g
-

1'
’l.

()
I
,
(I)

.
as.
\—

v-x. C;

’44)
£1.

c
(4
(J

1

(I)

H..‘
a

v-\, r-

f‘
(I..
I

 

Chapter 1. Introduction

1.1. Background

Lactic acid holds great potential as a feedstock in the
chemical process industry. Lactic acid is produced by
fermentation of corn starch hydrolysate. Increased demand
for biodegradable polymers manufactured from lactic acid
monomers has led to improved technology for production of
lactic acid from starch hydrolysate at high yield and low
cost. These developments enhance the prospects for use of
lactic acid as a renewable-resource feedstock for production
of traditionally petroleum derived chemicals. Lactic acid
may be converted to several valuable products using
inorganic catalysis. These products include acrylic acid,
2,3-pentanedione, propanoic acid, acetol and acetaldehyde.

Acrylic acid is the starting material for the
production of acrylic and acrylate polymers, and is
currently produced from petroleum by oxidation of propylene.
One of the keys to the production of polymers from renewable
resources is the conversion of lactic acid to acrylic acid.
Lactic acid may be converted to the condensation product
2,3—pentanedione with greater than 50% selectivity over
inorganic salt catalysts.

We have attempted to determine catalysts and conditions

which favor the various product pathways and to achieve high

 

 

. . a t a :..... . T. .. ..
.1 l E .3 r C M w... .2. C i .
u.) d C C .. e a . .l To
5 a .3 A 3 t L . a. . . . 1
t u, o 5 . b .1 l .p.. DC «3 4“
C C. C I L . . p . . I C
u 2 4.. .1.. m3 1 1 .7» .1 r
ufi1¢ A: s. v Cu Wu 0 o u my C.
A V .. . n o .1 2 2 2 2.. a.
r . . .3 C r 2. o . . .n.. v . Q»
.G 1‘. P.» ”.14 1 1 l u y by. .4M

w . .

 

 

\

2
product yields using inorganic salt catalysts. We have also
attempted to understand the chemistry of the conversion of
lactic acid. We have gained insight into mechanistic aspects
of surface reactions by determination of surface species

using both NMR and FTIR.
1.2. Literature Review

1.2.1. Lactic Acid
1.2.1.1. Manufacture

The market for lactic acid is about 50 million lb/yr in
the U.S. (1,2). Its uses include pharmaceutical
preparations, leather and textiles, food additives, as well
as production of inks, solvents, lacquers and plastics (3).
This is a rather small market at present, but lactic acid
has much potential. Its possible uses include production of
lactic acid polymers, acrylic polymers, and small molecules
such as ethanol and propylene glycol for which large markets
already exist.

Presently, about 85% of the lactic acid sold is
produced by synthesis. The most important synthetic route is
via lactonitrile, as shown in Figure 1.1. Lactonitrile is a
by—product of acrylonitrile synthesis, but can also be

prepared from acetaldehyde and hydrogen cyanide (4).

I n u-n
béijsov
-

I i
‘naun‘r‘r‘ra-
II P V .—
onavvuauvt.v

-
Avv no'rAvv.
( ~| h.
‘v¢§‘,\ve¢ ‘ue-
.
'-

““vvv“-\,- -

F

t‘””rc

.-~.¢ '
J“"

F“
f

F,-
_{21?A

‘“»¢«ol

fiy~:“"
\1- ‘2‘“-

Yflnh.‘
‘9‘ >4 ‘v‘ ‘
"v° a“:

VA
‘...":‘v~..‘.
"ec‘v' :r
d ‘4.
V“
‘_5
t'“:. : “A
‘vv w
VKJ
.
‘.
a
V arsn
y‘u‘.~c .f" ‘ ~
"“

 

 

 

I“;~i
a A
““u ~V‘ ~‘
'x.
s.“
‘-
.."\“y
.V A O
"»~:d‘r-
‘
‘1‘
_‘
‘A
:~*~
~b~va .
u ‘1‘
~ ‘ ‘
M .“ya
“
‘Q
‘
s“ _
s: ‘.
\u A
‘V fifl‘
\‘V
‘
'(
fJ‘u-W
‘ ‘l \
»‘. Q 1
..
vf'
.
‘N
< .
a vk ,7‘
V: V"
‘u \r
. I
“
w."_
{'s
. .‘
"Y- ‘
‘\ N 1

3
CH3CHO + HCN —> CH3CH(OH)CN
Acetaldehyde Lactonitrile
CH3CH(OH)CN + 2 H20 + HCl -+ CH3CH(OH)COOH + NHqu
Lactonitrile Lactic Acid

Figure 1.1. Lactic Acid Synthesis via Lactonitrile

In the past, lactic acid was manufactured almost
exclusively from sugar by fermentation using various micro-
organisms. The present trend is shifting away from synthetic
routes and back to fermentation pathways. Plans have been
recently announced by DuPont, ConAgra and Cargill to build
plants to make lactic acid from corn starch. These new
plants will produce higher yields of lactic acid at lower
cost utilizing new technology. This technology includes
improved fermentation processes with lactic acid yields
greater than 95% theoretical and better purification using
nontraditional separation techniques such as membrane
separation. Current projections predict a tripling of the
lactic acid market to greater than 150 million lb/yr in 1995
with a lowering of the price from $0.60—1.00/lb to $0.25—
0.35/lb. Most of this new market is targeted for lactic acid
polymer production, but the lower price of the feedstock

will make its use for other products more viable.

 

 

 

P

 

f‘
v

l

l

L l a C. A. V .3 r . g _ .1
P I .... E . . ... 3 l: . . d.
a. .3. .3 a. a. F... p .. C .. ‘ I
. ... .1 C .3 5.. Y .1 l .l I
3 2.. S C L. Q. A . S 2 C; .1. E
. «.3 r CL .7. I CV r .1 .1 a C y a
l .2. E . Q. r r a . .1
. . . a. .1 T. .... A. a. a. . .
2 r.. ...,. . . A .. n .. rd t .. . a ..x..
. Q 2. :.. I .. .3 S a. .. ..
l a. L.. LVL : . a . n . i. I. ~ ....

1.2.1.2.

'

ru-
v'-

.3

1.2.1.2. Properties

Lactic acid has rather interesting chemistry because it
contains both hydroxy and carboxylic acid functional groups.
Lactic acid self—esterifies to form lactide and polylactic
acid (5). Lactic acid forms esters easily with alcohols. The
boiling point of lactic acid is about 190°C at 1 atm
pressure, so in many cases the high boiling point of lactic
acid derivatives makes separation by distillation difficult

because of unwanted side reactions.

1.2.1.3. Primary Conversion Pathways

Several studies have been performed which examine the
conversion pathways of lactic acid. Many of these studies
have focused on acrylic acid formation since acrylic acid is
believed to be one of the keys to production of polymers
from renewable resources.

Some of the challenges associated with catalytic
conversion of lactic acid include: 1) difficulty in
vaporizing lactic acid, 2) coking, cracking and
esterification of the high boiling lactic acid, 3) poor
selectivity to any one product leading to a complex product
mixture, 4) difficulty in shutting down undesirable
pathways, 5) sensitivity of product selectivity to contact
materials and catalyst supports and 6) problems in

developing reliable methods of analysis.

 

 

 

 

A .. . .
.. . ‘ . A... . . A . .. ‘
“(V q . .. a a... H a.»
:u 3. 1| ‘ «.3 "WA ‘5.
V .0
any .v. A. o
C be w. C .. . C
.1. ARV v. ‘ vH‘ .. .
5.. .3 . . m. H e t
a C a. _. a .
.3 n.-. f. .1.. .3
1 a a). :4 A .c . u

L.

E

A .I....

‘3‘.

5
Holmen describes reaction of lactic acid and lactates
over sulfate and phosphate catalysts (6). Acrylic acid
yields of 68% were obtained at about 400°C using 10% aqueous
lactic acid solutions.
Odell, et al., (7) described conversion of lactic acid
to acrylic, propanoic, B-hydroxypropanoic, pyruvic and

acetic acids, as shown in Figure 1.2.

CH3CH(OH)COOH -’ CszCHCOOH + CH3CH3COOH + HOCHZCHZCOOH +

Lactic Acid Acrylic Acid Propanoic Acid B-Hydroxypropanoic
Acid

CHBCHOCOOH + CHBCOOH
Pyruvic Acid Acetic Acid
Figure 1.2. Products of Lactic Acid Conversion Reported by
Odell (7)

These reactions were carried out in sealed ampoule
experiments in aqueous solution at 220—250°C over catalysts
of group VIII metal complexes such as [PtH(PEtQ3]‘. The
major product in all cases was propanoic acid with yields of
2-50%. Acrylic acid was produced at 230°C with a yield of
4%. Significant yields of B-hydroxypropanoic acid (28%) are
obtained under the same conditions. This suggests
rehydration of acrylic acid to form B—hydroxypropanoic acid
and lactic acid.

Dehydration of lactic acid and ammonium lactate to
acrylic acid in the vapor phase over aluminum phosphate

catalysts is described by Paperizos, et al. (8,9). Aluminum

 

III I I
. . . . . v . . \ . y . . . s . .. . . .. ,p . ..
S C 3. .. . V. C R C .2 t. _ .. ..-. . ., .c er... .H. E C .. .... y...
e h a» C u C h C 4‘ 2H «dd Alv .nu 2v 1‘ by 41* .l YFA. Vs C .\|\ \*
t C b .l .l a E . . .l ....n C X e. I C ,. . . -. r. .... t. and . . . .1.
3 .1. C S A. S S .3 .1 .. E a. r P. .. .3 (I a r V. 1. .... a.
h Cw 3v h v u A A v ill. nun. Y; “Ky w( ‘ . nun CL nu 313 3A... . s ‘ . ~ . 2 Vs HIM «. Nut. V\ V \
C. h .3 .. . .3 1 l a. .. . . . a . I ,. S C .3 CL 3 4. Q.
S e . . a. 2.. . .. . _ .3 v. S .Q 4.. r0 .3. a. .G I ....v. a: c .. km. _ o.
C S r. T C o. t. 2 .C .. . .3 .. . T. .. .. .t.. Q. .1 1...... .. .. .
L H a n . . .3 a. .3 . . :. 6.. .. . 2. CL . .. . ,. f . . c. y . . .._ U .2 c.
r u. .n u .l. .3 . -. . . .3 a). 7. a... 2.. .3 v . a . .3 C . 21. A. a .3 A. $1,. .1.. L.-.

 

6
phosphate catalysts were treated with an aqueous inorganic
base such as ammonia to improve selectivity to acrylic acid.
In the best case, a 61% yield of acrylic acid is obtained at
a reaction temperature of 340°C and 4.2 seconds residence
time using ammonium lactate feed. The best yield using
lactic acid feed was 43% at 340°C. Propanoic acid,
acetaldehyde and gases were the other major products.
Catalyst calcination temperature was found to affect both
conversion and yield.

Sawicki describes a catalyst prepared by impregnation
of silica supports with monobasic sodium phosphate catalyst
and then buffering with sodium bicarbonate (10). When lactic
acid feed was introduced into a downflow Pyrex—lined
reactor, this catalyst gave a yield of 58% and selectivity
of 65% to acrylic acid at 350°C.

A thorough study of the reaction pathways of lactic
acid in supercritical water was published by Mok, Antal and
Jones (11). Three primary reaction pathways are described:
1) acid catalyzed decarbonylation to acetaldehyde, CO and
Ego, 2) decarboxylation to acetaldehyde, Cozeuuifg and 3)
dehydration to acrylic acid and H53..Acetic acid and acetone
are further reaction products of acetaldehyde, while acrylic
acid hydrogenates to form propanoic acid or decarboxylates
to form ethene. Increasing solvent density favors the
dehydration pathway. Effects of solvent concentration,

temperature, pressure and acid catalysts are described for

V
k “—‘hh’
833.3 33‘...

u...‘
-G

\l

\ov-

AVVA

AV?

.\.'c¢v‘

\oh.‘

‘firf‘A A
y..\_,‘d'

‘

Av.

y—
.“

mu
‘u

.
t.

:H NC
y.

.1 r1.
v u 4‘
\fi‘

Q .‘

my

2,.

r, n R

7
each pathway. Primary reaction pathways of lactic acid

observed by Mok, Antal and Jones are shown in Figure 1.3.

—+ CH3CHO + C0 + H20

Acetaldehyde
CH3CH(OH)COOH -> CH3CHO + co2 + H2
Lactic Acid Acetaldehyde

-> CH2=CHCOOH + H20

Acrylic Acid

CH3CHO -* CH3COOH + CH3COCH3 + CH,
Acetaldehyde Acetic Acid Acetone Methane
+H2
CH2=CHCOOH -+ CH3CH2COOH
Acrylic Acid Propanoic Acid
—’ CH2=CH2 + CO2
Ethene

Figure 1.3. Lactic Acid Reaction Pathways Observed by Mok,
Antal and Jones (11)

Lira and McCrackin studied the dehydration of lactic
acid in supercritical water using phosphate, acid and base
catalysts (12). They observed 58% acrylic acid yields using
homogeneous sodium phosphate catalysts at pressures of 310
bar and 360°C. Aging of the reactor favored higher acrylic
acid yields. This was thought to be due to passivation of

reactive sites on the reactor walls which cause undesired

[NA

1

T
4
1
2
1

 

f;
.3
L.

2..

8
side reactions. This passivation is possibly caused by
selective dissolution of metals from the alloy reactor walls
under the influence of supercritical water. The same basic
pathways were observed that had been seen before by Mok,
Antal and Jones. The water—gas shift reaction may also play

a role in product distribution.

Water-Gas Shift Reaction

co + H20 :3 co2 + H2

1.2.1.4. Thermodynamics

Heats of reaction and equilibrium constants for the
major reaction pathways were calculated using literature
values or estimated from molecular orbital calculations if
literature values were not available. As seen in Table 1.1.,
all major reaction pathways are thermodynamically favorable,

so only kinetic limitations exist for product formation.

Table 1.1. Heats of Reaction and Equilibrium Constants for
Major Reaction Pathways

 

 

Reaction AHL2%,, K
(kJ/mol)
CH3CH(OH)COOH —+ CH3CHO + co + H20 87 2x106
CH3CH(OH)COOH —> CH3CHO + co2 + H2 46 7x107
CH3CH(OH)COOH —> CH2=CHCOOH + H20 27 8x103
2CH3CH(OH)COOH —+ CH3C(O)C(O)C,_H5 + —21.' 2x10”
map + co2

 

* Estimated from molecular orbital calculations.

 

 

III
‘I‘ . . K x . .. F . . A . 2 wot \ . . ‘ .. wk. . . U. .. . I g .c .. .._
1; 1 F. E. a... :. D. W. r... 3 .3 .2.. .. .3. m» A. v...
v a b. p . Orb S «.2 up. n C. fi. fi! .3 .. .. Am A... .C. R. «C .Q 2..
M m. u S C t a. a. l” 3 .l .... .. . .... L... - C . . .1 .4 S
C a . I .1 l r. . . . . 2. I .. .. . S C. v . t . .1. T. C a. E
. l“ C S a. C .3 .3 .1. “.1 1.. C .3 f. r .3 . . .. .. C .. . L .. r.
2 “D r. mw p . «my . . . . a: . . w a v. Q. 2. C. 3n .. A . ‘ . v . ‘5‘ t mu.
0 W n 1 7‘ Wu :u v.1. vi. if, Eb. WNJ N I. v w!“ ht... m 4 4“ \r s .\ ‘ w"
2 .1... . . . . 3 w. w. . 3 .3 r. .. 3 C. v . .d h. a. .p. .3 -. .1 Q.
. n .G A... .. ~ . . A. n . . . T. .. v . C .C .. . Lg .. .. y . . T. C r... L-..
1 a a u. C. .5 .d .3 .3 . .a S T... .3 T. .3 3.. . . E. C a. .2 .2 L .

 

9
1.2.2. Manufacture of Acrylic Acid

Acrylic acid and acrylates are used to prepare emulsion
and solution polymers used in coatings, finishes, binders,
paints, polishes, adhesives, fibers and specialty acrylic
polymers. The market for acrylates has grown at a rate of
about 10% for many years and is expected to continue growth
at about 5% in the near future. The world production of
acrylates was about 2 billion pounds in 1985 (13). U.S.
acrylate production in 1987 was about 1.1 billion pounds
(14). BASF of Germany operates acrylate plants in the U.S.
and Europe and produced about half of the world acrylate
supply in 1991 (15).

A very large market for acrylates has been found in
production of acrylic fibers for use in polyacrylate super—
absorbing polymers (SAP). The exceptional growth of this
market has been due largely to the use of SAPS in disposable
diapers, adult incontinence and feminine hygiene products.
About 500 million pounds of acrylic fibers were produced in
the world in 1990, which is about 25% of the world acrylate
production (16). Acrylate polymeric products have good
clarity and strength and are resistant to chemical and
environmental attack (17).

Many processes have been used commercially to produce
acrylic acid in the past. Only the propene oxidation route
and the acetylene-based Reppe process (BASE) are in use at

the present time. All new plants being built are based on

J

O

0V
,5,- .‘

‘1‘.

 

 

. .
-F’Ay'?‘
“My/JV
vv..
«5
‘ "_ ".
r::~~
~-u-.l

..-..u-.. v

if

A‘pu

 

. . n
. A c . r A .N— c
. \ ~
v. . . a... . . .3
.5. .«4 mo . J aw Q. a;
Q. 2. .u. . n_. ..u .1.
h . .. ac .. .t
. . a P.‘ A. :. ... l N h
. . s ‘ N a Y a P. \r \
1&4 V A [u y N. w vu 9%»
«v... a». . A: C . a .. M...
Q. 2.. LC NU.
a; A . . s IF. :H
VNM .fi,‘ I v ‘ .

lO
propene oxidation. This process is attractive because
propene is cheap and very selective catalysts have been
developed for the process. The overall yield of acrylic acid
is about 80% based on propene (13). Figure 1.4. lists some
of the commercial methods of acrylate production. Another
process for the production of acrylic acid involves the

oxidative carbonylation of ethene over a palladium catalyst.

1.2.3. Reactions Related to Acrylate Production
Interesting work has been published dealing with
reactions of lactic acid derivatives to acrylates and

derivatives thereof.

1.2.3.1. Pyrolysis of d-Acetoxypropionates

A series of papers in 1935 by Burns, Jones and Ritchie
described the pyrolysis of d—acetoxypropionates and d—
acetoxyisobutyrates to produce acrylates and methacrylates

(18-20).
cmcoocn (CH3 ) COOCH3 —> CH2=CHCOOCH3 + CH3COOH
Methyl d—Acetoxypropionate Methyl Acrylate Acetic Acid
CH3COOC ( CH3) 2COOCH3 -+ CH3=C (CH3) coocH3 + CH3COOH

Methyl d-Acetoxyisobutyrate Methyl Methacrylate Acetic Acid

The pyrolysis was carried out under nitrogen or COijl

a vertical Pyrex glass-lined iron tube packed with quartz

 

AvOAouA

,flv~

..Loo v'

l
\1

\rA.“

~

"A-"J‘
a. .. C4 h:
v--- -‘v.‘_'

 

 

 

 

 

 
 
 

_ \ A V 5 v v_ I V‘ _ar g g . _
.4 . A

vs ‘2 5‘ . 5k bk .W‘ 4‘ 13%.

_ . A c .. t ,

. . _,.. t .. A _ «z. .
. . .. . L.‘ “w. 2» Ge 2% Ir. «1‘
n . a c r. A. 4 . T. 3
C a .v E E E 2.. .

: 2 . i he

.. :1i 3 a. A. «a .‘ Q. :J‘
v, n a v ‘. u s

.1.‘ v N - F. A- \‘ AL» v 1‘ uv- ~N.‘ & .

h~ A y “s s v N 5 s _ . a e. In?” H- «N.»

a v P. it y. . it e . L t A . o...

if. if. .3.. n7. .1... :~ :~ A. .. 1?.

ll

Propene Oxidation Route

catalyst
CI-12=CHCH3 + 02 —’ CszCHCHO + H20
Propene Acrolein
CH2=CHCHO + 1/2 02 -+ CH22CHCOOH
Acrolein Acrylic Acid

Acetylene Route (Reppe)
4HCECH + Ni(CO)4 + 4HEO + 2HCl —’ 4CH2=CHCOOH + H2 + NiCl2
Acetylene Acrylic Acid
Acetylene Route (BASF)
Ni Catalyst
HCECH + co + H20 -+ CH2=CHCOOH
Acetylene Acrylic Acid

Acrylonitrile Route

Acrylonitrile Acrylic Acid

Ketene Route <:/:>::0

H2C=C=O + CHZO —* O + CszCHCOOH
Ketene Formaldehyde B—Propiolactone Acrylic Acid

Ethylene Cyanohydrin Route
HOCH2CH2CEN + st04 + H20 -+ CszCHCOOH + NH4HSO4

Ethylene Cyanohydrin Acrylic Acid

Condensation of Acetic Acid with Formaldehyde
CH3COOH + CHZO -+ CH2=CHCOOH + H20
Acetic Acid Formaldehyde Acrylic Acid

Figure 1.4. Routes for the Production of Acrylic Acid

V"
.AA--‘l

Am

"‘ ‘5 _.n,-,-.
QwE’L‘V u. vi;
4‘

12
chips. At 470—485°C, yields of 80-83% methyl methacrylate
were obtained. This study also addressed the effect of both
temperature and ester group on product yields.

A number of papers were published in the 19403 by the
researchers Smith, Fisher, Ratchford, Fein, Rehberg,
Filachione and Lengel (21—30). Building on the work of
Burns, Jones and Ritchie, these papers also dealt with the
pyrolysis of lactic acid derivatives to acrylates, focusing_
on the effects of contact times, contact materials and ester
groups on the acrylate yields. For pyrolysis of methyl d-
acetoxypropionate, the best methyl acrylate yields (of about
90%) were obtained at 550°C and 10 seconds contact time.
Quartz chips, aluminum strips, Pyrex, carborundum, carbon
rods, copper and iron were found to be good contact
materials, while silica gel and alumina led to coking and
the formation of gases. Clay was found to be inert and led
to lower conversions when used as a contact material. Best
results were obtained using methyl ester groups mainly
because the thermal stability of the methyl group led to

higher selectivity to the desired products.

1.2.3.2. Preparation of Methacrylic Acid

Various preparations of methacrylic acid are discussed
by White (31). These include dehydration of @-
hydroxyisobutyrates using phosphorus trichloride and

dehydrohalogenation of d—halo isobutyrates.

3-}I-Hr/‘3v-Ay-

Avu‘v‘asv
‘ ‘

V

l -—: (A, ‘-
“~‘-\/ -
—’ V

in‘ -

- -‘g
..,. g .

V“ ‘f‘L IV~ ‘

D‘Avh. u-:

 

 

 

l3

CH3C(OH)(CH3)COOR + PCl3 —+ CH2=C(CH3)COOR + H20
d-Hydroxyisobutyrate Phosphorus Methacrylate
Trichloride
CHBCX(CH3)COOR -> CH2=C(CH3)COOR + HX
d-Halo Isobutyrate Methacrylate Hydrogen Halide

Filachione, Fein, Lengel and Fisher published a paper
which discussed pyrolysis of acetoxyisobutyrates to esters
of methacrylic acid (32). This was an extension of their
work on acetoxypropionates. In this work, the ester groups
were a series of lactates with various ester groups
attached. The highest yield observed was 85% at 450°C for

pyrolysis of the acetoxyisobutyrate of methallyl lactate.

CH3COOC (CH3) 2COOCH (CH3 ) COOCHzCHCH3
Acetoxyisobutyrate of Methallyl Lactate
—> CH2=C(CHQCOOCH(CH3)COOCH=CHCH3 + CH3COOH

Methacrylate of Methallyl Lactate Acetic Acid

Abe, et al., observed a 93% yield of methacrylate at
240°C over aluminosilicate 13X molecular sieve (33). The
vaporized feed was a 50% solution of methyl d-

hydroxyisobutyrate in methanol.

v x y .
5. .r~ AVUA MOW It A“ .H. A» .. «c w. . . P2 N;
Av .‘ u . .v .7 . . 4 ..
«MW _ Cu W." 5.. J. .a. . .. 4‘ life i: Ag «w AW.
Wu .1 . r” .G .3.. r N». r at 4,9. . .n. .. .3
m». 3 .3 C .. . my. .1. .5. .5 . (A. C. _ ..
a--. .5 . n.“ t t . . J. t . ..n a a c Q: S
"I C 3 .C . c .. a m C. U 3: .A.‘ 4..“ 1.. «Q
.C v. . E E .C . . .C an. i... E
T. .3 2 v... . i A .c 2.. A c S .2... . . 3...
e T. . ... .2 S a... S .. .. i at t.
11 n.“ 1 ~11. .2 q i . . .2 .3 m P. 7.

 

14

CH3C(OH) (CH3)COOCH3 —> CH2=C(CH3)COOCH3 + CH2=C(CH3)COOH +

Methyl d— Methyl Methacrylate Methacrylic Acid
Hydroxyisobutyrate 93% 2%
CH3COOCH3 + CH3OC (CH3) 2coo<2H3

Methyl Acetate Methyl d-Methoxyisobutyrate
<1% <1%

1.2.3.3. Oxydehydrogenation of Isobutyric Acid to
Methacrylic Acid

Many studies have been performed on phosphates with
mixed cations. These catalysts have been shown to exhibit
selectivity to methacrylic acid in oxydehydrogenation of

isobutyric acid.

CH3CH(CH3)COOH + 1/2o2 -+ CH2=C(CH3)COOH + H20

Isobutyric Acid Methacrylic Acid

Watkins reported methacrylic acid yields of about 70%
at 380-490°C (34,35). The catalyst was a calcined residue of
bismuth oxynitrate, iron phosphate and lead phosphate. A
mixed phosphate catalyst with a composition of
Fe1flAgmufimePL3gk supported on silica gave yields of about
75% as described by Daniel (36,37). About the same time,
Ruszala showed that iron phosphate catalysts can give
selectivities up to 80% at 380°C (38).

Moreschini, et al., obtained yields of about 76% on
mixed phosphate catalysts with an empirical formula

FEHSMeKPlgx, where Me is Ni, Zr or Al (39). Reaction

 

4 .3 xv.
. C 3 r .. .3 S i .1..
, l .1 .1 i a .2 3
.M. r .3 n. at t .l
45%. Ad ”wv‘ .s . \l . a» h
.-- L” .C C A .3
C. E . ._ h S 2..
33 .C J .4 at fit my. Tn
.xv A(v v..." u . .. g by Av
3.. .1 T“ a. .5.. r .1
. C. C 5 mu v t. S

v‘
K.

h.

3‘“
~ ‘~—.‘.'

> §
3‘ ’W “
"fi

15
temperatures were 390—410°C.

A very interesting kinetic and catalyst screening study
of heteropoly acids was published by Haeberle and Emig in
1988 (40). Their results pointed to 10—molybdo-2-vanado—
phosphoric acid and its cesium salts as catalysts for
oxydehydrogenation of isobutyric acid at nearly 80%
selectivity at 305°C. Three major pathways were identified
which lead to formation of methacrylic acid, acetone and
propene as well as CO and CO? lese reactor experiments
showed that higher temperatures favor propene formation
while high oxygen partial pressure favors the methacrylic

acid pathway.

+ % 02
-> CH2=C(CH3)COOH + H20

Methacrylic Acid

+ 02
CH3CH(CH3)COOH -> cmcocn3 + co2 + H30
Isobutyric Acid Acetone

H+

-+ CH2=CHCH3 + co + H20

Propene

1.2.4. Dehydration of Alcohols
1.2.4.1. Mechanism of Acid Catalysis

Dehydration of alcohols to form alkenes occurs via an
E1 mechanism (41). First an acid catalyst donates a proton

to the hydroxy group. In the rate determining step, a water

 
 

2.

cl“,
\VV‘

q
_/'I‘
I'v‘

’Nv

16
molecule leaves to form a carbocation. If possible, the
carbocation rearranges via methylene or hydride shifts in
order to stabilize the positive charge. The last step
involves abstraction of a hydrogen on a carbon adjacent to
the carbocation by water to regenerate the acid catalyst and
yield the alkene. The stability of the carbocation formed
controls the product distribution. The product with the most
stable cation will be the major product. This dehydration
mechanism is presented in Figure 1.5.

Catalysts for alcohol dehydration include concentrated
Bronsted inorganic acids such as sulfuric and phosphoric
acid for primary alcohols. Secondary alcohols require less
acidic conditions and tertiary alcohols are easier still.

They react in dilute inorganic acids.

fast, 80°C
CH3C(CH3)2CH(CH3)OH + H3O‘ =1 CH3C(CH3)2CH(CH3)OH;,+ + H20

3,3—Dimethyl-2-butanol

slow
CH3C (CH3) 2CH (CH3) OHZ' a CH3C (CH3) 2CHCH3 + H20
~CH3
CH3C (CH3) 2c*HCH3 -+ CH3C“ (CH3) CH (CH3 ) CH3
secondary tertiary

fast
(CH3)2C+CH(CH3)2 +HZO -+ (CH3)2C=C(CH3)2 + CH2:C(CH3)CH(CH3)2
+H3O+
major minor

Figure 1.5. Acid Catalyzed Dehydration Mechanism

De]

-a‘_.

1.2.4.2.

.15 L“ .3 . .J. :C a. .
44 Z S T. a Pa 5: A.“

a r. E 3 a. ;... L .. p. ..

w: .3 a . ~r ... 5: E a

a; .l. C; .3 . Qe A ..

.. i Y. .J‘ O C E t .l E S

.r. C r .3. .. l 4“ in... at

J. C» A: 2. 7‘ . \ .-‘ a. A:

at Z .n a p i. «Q .l 2.

A .. Fm .. . .vu r: s34 a). .p u

.C "v . ~ .. «C L L a; . n .L Au 5:

 

17

1.2.4.2. Dehydration over Zeolite Catalysts

Zeolite catalysts were considered good candidates for
dehydration of lactic acid to acrylic acid because Lewis
acid catalysts such as zeolites and alumina have been used
to catalyze gas phase dehydration of alcohols (42). Zeolites
have both Brénsted and Lewis acid sites and are known to
catalyze hydrocarbon cracking via carbonium ion
intermediates (43—45). Several other considerations
suggested zeolite catalysts to be suitable for the
conversion of lactic acid: 1) high acidity, 2) high
activity, 3) low coke formation and 4) regenerability (45).

Y and ZSM zeolites are both synthetic aluminosilicates.
Y zeolite has a Si/Al ratio of about 2.4 and a structure
made of sodalite cages connected to form 13A faujasite
supercages. The faujasite structure is entered through a
twelve-membered ring about 8A in diameter (46). ZSM—5
zeolite has a Si/Al ratio of about 3.5 and consists of a
building unit of eight five membered rings which are joined
at the edges to form chains. These chains link to form
sheets and the sheets stack to form a three dimensional
structure known as pentasil. The structure is entered
through a ten—membered ring about 6.3A in diameter (45). The

topology of zeolite structure is presented in Table 1.2.

‘1 “
T6248 1. . .

 

 

\
“h h!!! 2.."
r‘LK-‘rvtly‘u u
-.
I-
""~V‘." ‘-
“'»A..v' ‘
“VCfi ~‘Q hay.
hvv-‘\_.v ' r‘fl
bu,

o

Y: . ,.

“~' ~'.'
"\-H

V‘-\

~(~"rx

~.. .y
3v§

VANI
‘N
‘11
~.' .
.~"y
“Jim":‘i i
.. ML“‘ ‘(~V‘l
VA
3.1.,
«Ali-an-“
V ‘y, I
a.
JI a».
L“ ‘h ‘
Juan. g
V-s
C
l
a (y-
to ~v~
L‘V ( {5‘
Par“ . k
1
q
é‘Cy~
.‘3CahAl
L 1'“, Q I
“ff" 1
.. r“.
‘V‘VIRQ‘I-R
‘”=FQ .
‘y
D.‘
‘v‘g
hc-‘,Qv~
«I 7
~~‘l
‘4 ~ .
‘\‘rlL‘

 

18

Table 1.2. Topology of Y and ZSM—5 Zeolites

 

 

Zeolite Type Y ZSM-5
Superstructure Faujasite Pentasil
Number of Tetrahedra in Ring 12 10
Pore diameter (A) 8.0 6.3
Approximate Si/Al Ratio 2.4 3.5

 

Many researchers report dehydration of alcohols using
zeolite catalysts. The acid catalyzed dehydration of 2—
propanol occurs via a carbocation intermediate to form
propene and diisopropyl ether at 150°C over H Y zeolite
(47). Methanol and ethanol are converted to hydrocarbons
over HZSM-S at 400°C (48).

Dehydration of alcohols using zeolite catalysts was
reviewed by Turkevich (49). Dehydration of 2—butanol was
reported over Ca A zeolite. Dehydration of ethanol to ethene
was observed over metallic cadmium on X zeolite. 2-propanol
dehydration over H X, Na Y and Ca A zeolites was seen at
240—270°C. Ca Y zeolite was reported to produce dehydration
of ethanol to ethene at 250°C in 100% yield, dehydration of
1-propanol to propene in 90% yield at 275°C, dehydration of
2—propanol at 230°C and dehydration of cyclohexanol to
cyclohexene in 100% yield at 170°C.

Water, alcohol and.C£5 are known to reversibly promote

activity over faujasite zeolites for a wide variety of

.. . _
Afivnw‘jd _.
\u‘VHu‘L‘k/«cs \—

, .
nvzfl arr-y
v."vl\cl 53.3

n
h! "

LIV

. v .
”)C'Y‘G ‘ Q”...
U~AAA -5. ~~v _
d ‘
‘ D
‘- 6
75:1" ‘6?“ b]
v bv-v£n VL
. . ‘
""“VA<~y. «. ..
“'Jbvakv“ D‘ _
‘ ‘
‘ ' ‘
”22‘” "55.-
-“ —<
”*VLV uuvvu‘
h ' -
“vu-,‘
D‘;-LF' e C).- .
My ~

 

 

flvv 1" .
, V1r~‘_>‘
‘ ‘
V" «I
.,
‘
VF“ ’C‘: v
“\u‘ 'I ‘ l ‘
v ,. ‘
.
C”IV\A _ ‘
"’"“':~G' 1;?”
‘ v
““H'
K h a- .
s, v.) GD g(‘__
A‘A.
:3- Al‘hfn
'9- ‘2) i
‘4‘. \‘.
v‘ .
“‘WA
~“‘:‘r::‘
.\_4
1.. .
‘V‘[~.."“.fl‘r\.
v.‘\.“.
V‘
«A
“4V, Q:
~—
A‘ A
».:r Ci‘! _
.
k..‘:‘ .
I
.N,‘
(VPA
‘Vll ~P‘h
V‘. V..‘
‘ ~ ‘
“
A,
‘1 .
\u‘.‘ ‘

;s
~‘Q‘JA
yr
I
25‘,
fi
«5- , ‘
V V’Ql
‘ K

19
conditions and reactions (50). Promotion of isooctane
cracking by CCg over ion—exchanged Y zeolites was studied
using IR spectroscopy. The promotion mechanism involves
reaction of surface species with CCE to convert basic
hydroxyl sites to acidic hydroxyl sites and destruction of
basic metal hydroxide species which inhibit cracking on the
surface of the zeolite catalyst (51,52).

Dehydration of ethanol to ethene has been demonstrated
by Cursetji, et al., over high silica zeolites containing
iron (53). Yields of 98% were obtained at 350°C. This
temperature is much lower than for more familiar catalysts
such as alumina. Yields of 98.9% were observed over alumina
at 450°C.

Ethene yields of 98.3% were reported by Dao over Zn and
Mn ion-exchanged ZSM—5 zeolites at about 400°C (54).

Mao, et al., observed ethene yields greater than 97%
over ZSM-5 zeolites for feeds with low ethanol
concentrations of less than 19% by volume. The feed also
contained other impurities present in fermentation products

(55).

1.2.5. a-Diketones
1.2.5.1. Properties and Uses

d—Diketones find most use in the pharmaceutical,
flavor, fragrance and dye industries. Aliphatic diketones

are yellow liquids with a penetrating odor at high

 

n
\r

O I F _

. l ”w! . 1; Av.“ V \iul. a L
t e, C A: V. E J
a (H p C; t c 3 +1
r t _ .1 (\ .W

FL g 3 1L

3 "Y. A,“ I . 1. rt r
e h . A l 2 b k u n.)

A t mu 1). .J g f:
on . . l 13 7 . . . A h
C 3 AJ A)“ A o e 71

.. . l .n L a Cu Th . .\.

20
concentrations but a sweet, buttery smell when diluted. This
is why they find use as butter flavor ingredients. The
boiling point of 2,3—pentanedione is 108°C. 2,3—Butanedione
and 2,3—pentanedione are both soluble in water but
solubility decreases rapidly with increasing molecular
weight (56). The market for 2,3—pentanedione is about 5
klb/yr while 2,3—butanedione has a market of about 50 klb/yr

(14).

1.2.5.2. Reactions

Tautomerization of d-diketones forms the enol. Self-
condensation of d-diketones forms first the aldol, then

quinones. d-Diketones may be reduced to d-ketols and then

to d—diols (57). These reactions are shown in Figure 1.6.
RC(O)C(O)CH2R' -+ RC(O)C(OH)=CHR'
a—Diketone Enol
2 CH3C(O)C(O)CH3, -> CH3C(O)C(OH) (CH3)CH2C(O)C(O)CH3
a-Diketone Aldol

k

CH3C(O)C(OH) (CH3)CH2C(O)C(O)CH3 “’[C(CH3)=CHC(O)C(CH3):CHC(O)J

Aldol Quinone
RC(O)C(O)CH2R' —+ RC(OH)C(O)CH3R' -+ RC(OH)C(OH)CH2R'
d-Diketone d—Ketol d—Diol

Figure 1.6. Reactions of d—Diketones

F
1.2.6. Catai

1.2.6.1. Prd

.
' V

n»
“MA"WAQWL
v--.../:,. V~‘

I l
HV‘V‘j‘Auh 2“»
A§-b’.¢vsk’(.“ '

A

s ‘ ~
var-QC Hhfl‘:h’r
"“9“,

Vuv._ V_
‘ A

.
n "v- -
“AA.‘~-“ l

 

 

VA ‘
N’M,‘ '
s \“ ~
‘UH‘LC ‘I‘! k
s ‘ ‘
‘—
mu“.
"‘3‘. h“
‘ ‘6‘! C.“
‘v‘ \ F
‘v
v
‘4
h."
y‘:OQY‘~ .
H“- ‘
‘.
7:._
v “1"
.' ‘ W ‘
L. ‘ifif‘ ‘V
MA“-« :2
L:~v~_
...1MN¥:‘“
sy‘u ,‘y‘A
~, N ‘
a‘»

21
1.2.6. Catalysts
1.2.6.1. Properties of Phosphate Salts

Phosphate salts exist in several different forms. The
orthophosphate salts of sodium, for example, can exist in
monobasic, dibasic and tribasic forms. Phosphates may
condense to form chains and rings of various sizes. The
simplest chain is called pyrophosphate and is made of two
phosphate centers joined by a common oxygen atom.
Triphosphate contains three phosphate centers. When these
three phosphate ligands are joined to form a ring, the salt
is called trimetaphosphate. Other stable forms of the
phosphate salts include a long chain form with some
crosslinking called Maddrell's salt and a rather stable six
membered ring called hexametaphosphate or Graham's salt.
Some common forms of sodium phosphate salts are presented in
Figure 1.7.

Alkali phosphate salts are prepared by heating alkali
orthophosphates (58,59). Partial dehydration of monobasic
sodium phosphate (NaHfiKfi) gives disodium.pyrophosphate
OMMHPML). Further heating to temperatures above 220°C
results in the formation of Maddrell's salt. Small
quantities of trimetaphosphate are formed also and this form
is present in substantial amounts when heated above about
350°C. Maddrell's salt changes from a low to high—

temperature form at about 300°C, as shown in Figure 1.8.

I . v'
lanascdzm
Phosphate

 

 

VF‘
V O
"A n
z..-:-3-P-CE
‘1'“ I“.
“(a Lila

H‘J
\
O n’;
\\,v
\'.'\
xth‘P
/
A
‘u‘
U
\
A
0:?
I \
A
n 2
.153 V"~—
’/
n
J
A ,

 

22

O O
H H
NaO-f-OH NaO—P-OH
I
OH ONa
Monosodium Disodium
Phosphate Phosphate

o o o o o
H H H H H
HO-P-O-P-OH HO—P-O—P—O-P—OH

ONa ONa ONa ONa ONa
Disodium Tripolyphosphate
Pyrophosphate
NaO O
\ //
.APA.
C3\ ,0 O\ IONa
NaO‘“F R520
O O
\ I
0:1? P-ONa
I \ 0’ \\
NaO OjP/ O
/ \
O ONa

Graham's Salt

Figure 1.7. Sodium Phosphate Salts

O
NaO-P-ONa
l

ONa

Trisodium
PhoSphate

Sodium

Trimetaphosphate

o o o
H H H

-O—P—O-P—O-P—O-P-O-

ONa ONa ONa ONa

Maddrell's Salt

 

\a‘oc a“
n-1, —)
h‘ h- -'_1

.

VAAAAA 1- ....
x :1 1“:

.¢V.-Vy\a -‘LAL

.

TFA‘nh—p,‘

V .[q H

~-~Vublb.§’-\.
.

 

A.
oo¢$‘\§-c-‘ :
R_‘ _ .
\A. .- vfil‘ a“-
u-A—v ‘.‘-\“.
.4

“— F A

‘~ H1)

"“Ls- ..

. v- ,_
Icy.:r“~\ ~
~~A‘..IH D
a ‘

\L,‘

N‘A“-

..

-‘-,._

‘~¢‘lr.: 1 ;
"’ v

 

:r-‘V.
' ‘4

~ A
‘ a A
\A
F“-
pfoA
‘0.“ A
v. 1,
(A
A
‘Va‘n
-‘a‘ A I
A A‘
“in ‘\
\J
\.
r: «
. us
“.4. \j. .-~
v w,
‘ \-
.‘—‘
.. A
‘.
t.
“a i

‘1'
\
..‘ 7
.‘AV‘
‘8 )I\
‘\‘t‘)‘.-‘
V‘ 1\
‘V
‘V
A
V" “fix
“1‘ 1 r'\
V‘NV
‘\
5‘
D v.‘
‘
«.- ’\
‘\‘v~“
‘ v
‘K‘
x
f“.
”smy‘
«:mk
‘ C
x
N
a
Q-
~'~\Y‘
‘s 1“
\A-
V“
A
‘ .
'vl" \
‘. N
\1
\4

23

180—240°C 220—400°C >300°C
NaHZPO4 -+ Na2H2P207 —> NaPO3 (III) -> NaPO3 (II)
Monosodium Disodium Maddrell's Maddrell's
Phosphate Pyrophosphate Salt (low T) Salt (high T)
450—600°C 625°C quench

NaPO3 (II) —+ NaPO3 (I) —* Melt —+ NasPEOI8
Maddrell's Sodium Graham's
Salt (high T) Trimetaphosphate Salt

reheat 350-450°C
Na612>éo18 -+ NaPO3 (II)
Graham's Maddrell's

Salt Salt (High T)

Figure 1.8. Thermal Transitions in the System NaJWI§O§+gO
(59)

Maddrell's salt changes to trimetaphosphate in increasing
amounts as heating continues up to the melting point of
625°C. Quenching of the melt yields formation of the glassy
Graham's salt. Other forms of salts exist in the system
Na§}IQO§%gO, but most of them are not very stable.

An excellent general review of phosphates as catalysts
was published by Moffat in 1978 (60). This article describes
various reactions catalyzed by phosphates and the effects of
promoters such as copper(II) on selectivity. Phosphate
structure and preparation as well as surface
characterization are discussed. Phosphate catalysts have
been used for both dehydration of alcohols to alkenes and

oxidation of benzene to phenol.

24
1.2.6.2. Studies of Phosphate Salts using MAS-“P-NMR

Magic Angle Spinning Phosphorus—31 Nuclear Magnetic
Resonance (MAS—“P-NMR) has been used to characterize
phosphate salts. The chemical shifts and anisotropies of
several sodium phosphates are listed as reported in the
literature in Table 1.4. Chemical shifts reported are
relative to 85% Hfixh where o = -5. The chemical shift, 5,
is reported according to IUPAC convention with 5 > 0
representing a downfield shift and deshielding.

Andrew, et al., identified three factors which
determine the 31P chemical shift of sodium triphosphate
(Nagfioufl: 1) the number and electronegativaties of ligands
on phosphorus, 2) the bond angles about phosphorus, and 3)
the occupation of n-bonding orbitals of phosphorus (61).
The observed difference in chemical shift between terminal
and central phosphorus in sodium triphosphate is attributed
mostly to increased n—bonding to terminal phosphorus atoms.
Burlinson, et al., also studied linear sodium triphosphate
and determined its chemical shift and shift tensor (62).

Duncan and Douglass studied sodium pyrophosphate.
URMPJL) and sodium polyphosphate (NaPOfln, (where n226)
(63). They found that the difference between the principal
shielding components 022 and 033 increases with bonding of
the phosphate unit. These differences are listed in Table
1.3. This effect is attributed to the change in the n

character of the phosphorus bonds.

 

 

F-‘n? l 3 2‘
{SA/$6 u t ,_
I

 

F‘ 9
"*~nn *o
{MJDr‘ua . c
*1

.q‘
.r

.1..

1 .
v. .,

“-4 -

‘3
PM
mud
-

~-‘.A..v‘.-.‘)

F. ‘ .

i":"f‘n~ ,. 3
‘vn~ I,“
...‘-.u Qua- '

" ‘ ‘ n..
3 -5 a ”V.

IL",

1.‘
L);

L“
(I)

D

.
F. .
fl)

.1. - ‘
. ‘.,~
.5. n
~V§A... p.‘CQV
M
t
kn.‘,_
(
‘1'. , n ‘ p.
.u :1“ Lcr. C
y “\

“fix
. v
'“ct‘i'C'j )‘h ,1 ~-

 

 

x.
“Ev.
K
97“" .
. ~F'V‘-‘ .
Q . _
HM.‘;‘I‘
"’r». ‘
.
5
“‘V a N
“a.“ ”:5
‘v.
.
..

.
*“~‘~“S 7'
*C
‘
Ll
J. ..,
"41‘ ”IA.
“u_ P‘v
Vt
7.
:33. ‘
i A
“‘48 fl.- ,
") g
v~ :
“.‘
.2 .
.hV‘.
"& (“Va ..
'J \ 9.“
V L (Y~_
4‘ p
.H‘

 

25

Table 1.3. Effect of Phosphate Unit on Difference in
Shielding Components 022 and 033

 

 

 

 

Phosphate Unit lofl-GD, ppm
End O3-P-O-R" 0-35
Middle O2—P-(O—R)2 135—225
Ring Oz—P—(PO3),1 195—225
Branching O-P—(O-R)3 230—295

 

* from analysis by Duncan and Douglass (63).
** R is a phosphate unit.

A thorough study of sodium phosphates was published by
Hayashi and Hayamizu in which NMR results for several ortho—
, pyro-, tri-, tetra- and metaphosphates are compared with
literature (64). Rothwell, et al., performed an extensive
study of calcium phosphates, as well as a mineralized
calcium phosphate, specifically human dental enamel (65).
Both an upfield chemical shift and increased anisotropy were
observed upon protonation of a phosphate species.

Prabhakar, et al., developed a relationship for ortho—,
pyro— and metaphosphates in which the chemical shift of
phosphorus is correlated with the effective nuclear charge
of the cation, the partial charge on the phosphorus and the
radius of the cation (66). The chemical shifts of the
various forms of phosphates fall into distinct regions.
Orthophosphates exhibit the highest chemical shift, meta-
phosphates have the lowest and pyrophosphates have

intermediate shifts.

' nev-
hrsic- I

V

' 1
a p,».,-.— n‘
— ( , (
ma..- :1. S..-

 

.

S.C-"‘V‘A Q

u-_vv ‘V‘ov’ L
. M

* '5

”ET???“ .
-Jl.~~-' J

u“

‘\

 

 

,.
A, .
“ a
V
V
n -3
.‘
\C'v
~-,Z‘.v .
_(

26

Turner, et al., developed a correlation for the
chemical shift of ortho-, pyro-, and metaphosphates which
depends on five factors: 1) the number of n—bonding
electrons, 2) the electronegativity of the next nearest
neighbor, 3) the effective nuclear charge of the cation, 4)
the radius of the cation and 5) the average deviation from
the tetrahedral O-P-O bond angle (67).

Un and Klein studied the orientation and values of the
principal components of the chemical shift tensor of several
ortho-, pyro- and metaphosphates and found each to be
sensitive to different chemical features (68). 011 is
coplanar with the shortest bond, 022 is sensitive to the
formal charge of the phosphate group and 033 lies along the
longest P—O bond and correlates with the P—O bond length and
R-O-P—O-R' bond angle. The overall isotropic chemical shift,
6., is dominated by n—bonding effects, as reported by

1

several other researchers.

Table 1.4. Chemical Shifts and Tensors Reported

in Literature

27

 

 

 

 

 

 

 

 

Compound Chemical Chemical Shift Tensor, Ref.
Shift, ppm ppm
6 011 02., (I33
Cam(QH)2HX%)6 2.8 19 4 —15 (65)
Human Dental 3.3 (65)
Enamel
NaHzPO4-HZO 2.3 77. —23. -61.1 (67)
bkuHPO4 6.6 48. 5. -74.5 (67)
bkgHPOg12H§D -16.6 69. 35. -55.4 (68)
NaBPO4 14.0 (64)
13.0 (69)
13.7 (66)
Na3PO4-10H20 7 .7 (64)
Na3PO4-12HZO 7 .8 (67)
NafifiO, 2.4 37 37 —80 (63)
1.2 37 37 —80 (63)
3.0 40 40 —89 (64)
2.2 40 40 —89 (64)
3.2 (66)
2.1 (66)
2.7 46 33 —87 (68)
1.3 50 27 -81 (68)
Na4PZO7-10HZO -1.4 (64)
-2.3 38 38 —70 (64)
Na4p207-12H20 -1.7 49 32 —76 (68)
_____ -2.0 58 26 ~78 (68)
ngjzpzo, -13.0 80 24 -65 (68)

 

 

 

 

28

Table 1.4. Chemical Shifts and Tensors Reported
in Literature (Continued)

 

 

 

 

 

 

 

 

Compound 6 G11 (3.,2 033 Ref
Naspon—I 1.2T (70)
—8.oC (70)
Naspxxo(—II) 5.6T (61)
-6.2C (61)
5.1T 62 32 -109 (64)
—5.6C 120 —24 —79 (64)
4.6 69 24 -107 (68)
-5.6 117 —20 —80 (68)
NaSPxIO—II 4.0T 62 37 -113 (62)
—5.7C 118 -22 —79 (62)
4.6T (70)
-6.6C (70)
Naspxr3 1.5T 47 47 -98 (64)
-16.4C 152 —18 —85 (64)
(Napo3)6 -6.oT (64)
—19.6C 165 -19 -87 (64)
(NaPOfln (nx26) 2.0T 45 45 —95 (63)
-19.0C 160 —13 -9o (63)
(NaPO3L, —15 6 (66)
—18.5 (66)
-19.9 (66)
-24.1 (66)
—26.8 (66)
1.0T (64)
—6.2T (64)
-19.9C 165 ‘29. ~85 (64)
E : 22:32:14?

1.2.7. Stu<

 

 

 

.1 . . r a c to .3 a t A... 42 (ix 441.. hm Aw. 1.

.3 .4 . a. .3 a u .. . 1w)“ “4).. «D 2.» Q.» A t 3w . r p Q... . 7.. r . n?
v. v. 4. n. at A16 4 wk. no .71 Y. . m. 3.» ‘Am .r‘ s. Vx a... Q.
C l at .1. . p .2. . . ( .. .J n . L .3 C. Q. 4) 4 40 2.
W. _. I“ .. . Ft Art. DC A6 rt. «4“ t 41 . ups. y» b... 4] . \ 7.: Q. . \
v)“ n w 5. .AL V: . is M). n.. A o 4» .. CW. m. 5 s s . . 4 C. s x n «C 4I\

.3 n a Co :6 .C ) - . . Q. .3 .3 . t 4 w a... u .. “71¢ . w urn. .446. mm” inn.

r; my. C... Y. L t A)... . t v . C. .N n .64 e ‘ .(I . . . ( 4

«a A: .flw .1.“. r: n, . Hum aux». . . v a :4 :

Li. Y . . . an... (.nlu 2), )

P b g. . ~ m \ A

29
1.2.7. Studies of Reaction Systems using IR Spectroscopy

Infrared spectroscopy (IR) has been used to
characterize dibasic sodium phosphate. Observed vibrational
frequencies have been assigned to the various vibrational
modes in a paper by Ramakrishnan (71). MAS-“P NMR and IR
have been applied to study of the interaction of supported
heteropoly acids with silica and silica—alumina supports
(72).

Studies have been performed in—situ using transmission
IR techniques. Several researchers described designs for
spectroscopic vapor phase catalytic reactors based on
transmission techniques (73—81). These spectroscopic
reactors are typically designed so that vapor is fed across
a thin catalyst disk enclosed in a chamber equipped with
windows which are transparent to IR radiation. The catalyst
disk must be thin enough that significant IR transmission
occurs.

Another useful technique is Diffuse Reflectance
Infrared Fourier Transform spectroscopy (DRIFT) (82,83).
With this technique, IR radiation is focused onto a small
holder into which powdered or solid sample is placed. The
reflected radiation is collected by mirrors surrounding the
Sanple and directed to the detector. In either technique,
gzases may flow across or through the heated catalyst sample

WTiile performing FTIR spectroscopy.

Chapter 2

2.1. mt:

AF.

.1 4...;
4

“"V‘f‘v‘ a

x (V 47

'th.\4 "A\
u

n .

.""‘)~ba~

~w “is

F O
. (9A ‘
“VP. 5.5th
V ‘
‘Au' .
~ ‘
7
“v' f‘. ‘u.

 

Chapter 2. Lactic Acid Conversion over Zeolites in a
Batch Reactor

2.1. Introduction

Dehydration of alcohols is known to be catalyzed by
strong acids and proceeds via carbocation formation (41).
Zeolites have both Bronsted and Lewis acid sites and are
also known to catalyze hydrocarbon cracking via carbonium
ion intermediates (43—45). For this reason, zeolite
catalysts were considered good candidates for dehydration of
lactic acid to acrylic acid. Several other considerations
suggested zeolites would be good catalysts for the
conversion of lactic acid: 1) high acidity, 2) high

activity, 3) low coke formation and 4) regenerability (47).

2.2. Methods

2.2.1. Materials

Lactic acid was obtained from Aldrich Chemical Company
in an 85% aqueous solution and diluted to the required
concentration with reverse osmosis water from a house
supply.

Y-82 and Y—62 zeolite catalysts were obtained in powder
,form.from Union Carbide Corporation. HZSM—S catalyst was
éicquired from Mobil Research and Development Corporation in
1.-5 mm diameter extruded cylindrical pellets of varying

3O

 

)c a“
25:1: -F-S

.
“‘ Vlfi‘f
nub-4.4L”
. e

2: "
UV; -.$\.

(“Ln
.—

fiib'-

," tn.
TPY' y-n‘v «
‘-V-V*- I
A .

.
32‘;m. a!

vvxv~‘h‘.‘l

 

A
‘5.
~ I
'“-‘
v.
‘ - .
“m; z: ,
\—
l
v. w
h.
"‘1 " - N
-l‘ a ‘
L-
1
A-‘ .
“(ANA .
$.gfiufl “
“M

 

5
me “A
-' '-‘:m
v3)
)
Eff: 4
"l‘fip “
yL‘ CI"
\1
I
'2
-
”‘~...~2-‘
‘4
w.) I
.
A.
y .
~
v x,
r 4 c
‘.“\J I“
v L

 

 

Y
h!RA\ .
‘(
. r
f?“
-“Q
\.
nV
rem-x
. ._ h
‘~—..~.v~‘
“( Q
i I

31
lengths from 3—12 mm. Carbon dioxide and nitrogen gases for
purging the reactor were obtained from Airco, Inc. and AGA
Gas, Inc. in 99.5 and 99.99% purity, respectively.

The weighted mean macropore diameter was determined by
mercury porosimetry using a Micromeritics Pore Sizer 9310
and was found to be 1310 and 1910 A for Y-82 and Y-62
zeolites, respectively. The macropore diameter was not
determined for HZSM—S zeolite and the micropore size is
assumed to correspond with the particular zeolite type, as
discussed in the introduction. Calculation of the macropore

diameter is presented in Appendix A.

2.2.2. Calcination Furnace

Catalysts were calcined in a calcination furnace. The
furnace consisted of a insulated stainless steel (SS) vessel
with a SS lid which bolts onto the top. The catalyst was
loaded into a 0.5 inch ID x 5.5 inch long ceramic tube which
screws onto the bottom of the lid and extends down inside
the vessel. The tube connects to the lid by means of a nut
which screws into threads in the lid and is installed
beneath a small flange bonded to the top end of the tube. An
O—ring forms a seal between the lid and the tube. The
Catalyst is held in the vertical tube by a plug of glass
“Kaol in the bottom end.

The tube is heated from the outside by an electric

(iLement which slides loosely around the tube. The heating

C .1
at t t
L). a.
w) .A l
n).

. 4 8

U

«PA

4
app A99-
nu KILL).

.
F‘fi un
..u .19:

"4

(”Ca
\.a

{uni ,.

n44»-

FA
I—
6.“,

F.

2'2-3. Ca1

‘G

«3
h r.»

32
element is controlled by an Omega PID controller receiving
feedback from a thermocouple which extends down the center
of the tube and into the catalyst.

The catalyst may be purged by introducing gas through
an inlet in the center of the lid. The gas flows down the
center of the tube, through the catalyst and glass wool plug
and out the end of the tube near the bottom of the vessel.
Gas then flows up through the annular space between the tube
and the vessel insulation and exits at an outlet in the lid.
The purge gas flow may be reversed simply by connecting the
gas supply to the outlet. A diagram of the calcination

furnace is presented in Figure 2.1.

2.2.3. Catalyst Calcination

Catalysts were calcined at 300-400°C for 4—16 hours
under nitrogen purge. The nitrogen flow was not measured but
a brisk flow was assured by adjusting the flow until a soft

hissing sound was heard from the gas outlet connector.

2.2.4. Batch Reactor

All reactions were performed in a 300 ml Parr Mini
Reactor Model 4561M constructed of T316 stainless steel. The
reactor consists of a SS pressure vessel with a SS lid
Esealed by a flat Teflon seal. The lid connects to the vessel
13y two half circular clamps with bolts which tighten the lid

61nd compress the seal. A SS retainer ring assures that the

 

 

~.).
‘44»-

33

 

 

 

 

J
J

7:):

Heating
-*" “'7' Element

1'__, Ceramic
”'77 Tube

Catalyst

 

 

 

 

l1
i-—-ii----«-

 

 

 

 

 

 

“Thermocouple

 

_ Glass Wool
Plug

 

 

 

 

 

 

 

 

 

- Insulation

 

45

p .... SS Vessel

 

 

 

 

 

 

Figure 2.1. Calcination Furnace

' ‘ ‘m
A
RA: 3 1‘
C:U-C L,‘
a “at
«a- a
3:32.. 3*-
‘
'Qifi‘by .
‘Vuvbud-
‘ .
nary,‘ P.

veg~5fi 1
.~\‘_
“'vwx.‘

" \
5"!“y‘rx
VV"‘"V“-

.
A
*_h~.

v.,
i~ w
~\.\.~.. . -.‘

~--“l RF‘
htdcfirtg

( v w.
o\0

. .‘I

(n
W
h

“Cpi
.

 

 

34
clamps cannot fall off or be dislodged.

The lid is equipped with a mechanical stirrer which is
sealed by a water cooled seal to prevent failure of the
seal. Gas and liquid may be added to or removed from the
reactor by way of a dip tube which extends nearly to the
bottom of the vessel and by a gas outlet in the lid.
Temperature inside the reactor is measured by a thermocouple
which also extends nearly to the bottom of the vessel. The
vessel is heated by an external electric heating mantle
controlled by a dedicated Parr Controller Model 4841
receiving feedback from the thermocouple. Vessel pressure is
measured by a pressure gage at the dip tube inlet. The whole
apparatus is supported by a reactor stand equipped with a
stirrer motor.

Products of early experiments were sometimes green or
even purplish in color and white precipitates were often
observed. This suggested that lactic acid was reacting with
the metal in the reactor walls. T316 SS is composed of about
65% Fe, 12% Ni, 17% Cr, 2.5% Mo, 2.0% Mn and 1.0% Si, so
some of these metals may account for the product color and
precipitate. Many exposed surfaces were lined with or
replaced with non—corrosive materials to alleviate this
problem. A glass vessel liner, glass stirrer, Teflon dip
tube fitting, glass dip tube and glass thermocouple sheath
Vvere improvements designed to reduce contact of lactic acid

vvith metal. These changes improved the product color

 

fimaClCE
.
64
CCSEIVCa.

at;

Olav

.
nvv‘b - THV
Max, a

-v“
.

33‘2”." tr,

“-w.)
‘

5 7

6.9,

Ni
(AA
.—
...v

ya... .9
l V’ ..
auA.. “.._,

v... 4
“a h A-
\

hug y‘ta

’Na

)-.r ARA-

‘l'I':DD:y

y’-
:—
'x.
33‘s.;
7“ 'v-g
v .y.
'I(‘A~
”\~ 1
~ ~ 1
lac »

I
k.‘Q Fr”?
\‘V...
:
fi‘
7‘»
A“:
:;~.~‘
M N n
'“‘3‘ :g
‘ V“
A
S‘JWW“.
. r
V‘U.
C;r~i
u_:&.r:.
V
.5 ~_
\ 4
V‘.
'1. " 4.)
O“ > E;
\“x.
fi‘
Mb,”
s“K
\
4 4
~: 6.
\UJ q

 

35

dramatically, but small amounts of precipitate were still

observed.

The reaction vessel is rated for 3000 psig at 350°C,

but a rupture disk limits its operation to 2000 psig as a

safety measure.

2.2.

A reactor diagram is presented in Figure

2.2.5. Batch Reactor Operation

The batch

reactor was cleaned thoroughly before each

run. The vessel was disassembled and each of the components

was cleaned individually with water and acetone, if

necessary, to remove as much residue as possible. A small

brush usually removed everything except a few stains on the

glass vessel liner. The reactor lid was cleaned by rinsing

with water and
gas inlet. The
the components

The glass

flushing residue from the dip tube inlet and
reactor was towel dried after cleaning and
assembled in the reactor lid.

vessel liner was loaded with reagents and the

catalyst, if any, was poured into the liquid reagent

solution on the benchtop and stirred briefly to disperse the

catalyst. The loaded liner was then placed inside the vessel

and the reactor was assembled.

The stirrer was started and the reactor was purged with

carbon dioxide bubbled through the dip tube and exiting

‘through the gas outlet for 10 minutes. At this time the

CDutlet valve was closed, the<1% pressure was increased to

Gas In (—

loud 0.1+-

Seal Cool.
Waterln

(D
M
I\)

Gas In OIQ

Liquid Out 0:.

Seal Cooling
Water In

fl

 

 

 

 

 

36

 

 

 

 

 

 

 

 

 

' l

IF

 

 

 

H

Seal Cooling
Water Out

/ Teflon Seal

SS Stirring
Shaft

I Heating Mantel
/

~Therrnotouple
~ Glass Sheath

- Glass Dip Tube

- Glass Stirrer

~Glass Liner

\ SS Reactor
Vessel

 

 

Figure 2.2. Batch Reactor

 

S]!
5‘)
L24
: s'
D

|
on. ..b,‘

‘1 .-
an¢~buu x.

V
(’“M'CaYQ "‘
V

i4~va4

: “A.
. \"HQH
§ “-

 

\-

‘ v-
. “ L
w.“

‘1‘
F“)-
tutor.

.N

37

about 17 psig and the reactor sat for 10 additional minutes.
The inlet valve was then closed, thus sealing the reactor,
and heating was commenced. The reactor required about 50
minutes to heat from room temperature to 240°C. Temperature
overshoot was typically less than 4°C and lasted for less
than 5 minutes. Operator intervention reduced this overshoot
to 1°C in later runs.

Upon reaching reaction temperature, the timer was
started and the pressure and temperature were recorded at
intervals until the end of the run. Typical batch reactor

operating conditions are presented in Table 2.1.

Table 2.1. Batch Reactor Operating Conditions

 

Temperature (%D 240—260
Starting Pressure (at 25°C) (psig) 16—18
Ending Pressure (at 240°C) (psig) 590-620
Ending Pressure (at 260°C) (psig) 960
Pressure after Quenching (at ~0°C) (psig) 95—140
Liquid Feed Concentration (wt%) 8.5—36
Liquid Feed Mass (g) 100.0
Reaction Time (hr) 2-3
Catalyst Mass (g) 1-3
Purge Gas CO2

 

After the run, the heater was turned off, the heating
mantle was removed from the reactor and the reaction was
quenched. The bottom of the reactor was first submersed in

lust water, then ice water, as the reactor cooled to less

{WV

h
L

fiFr‘.‘y-
vvv—

1’58

n

l

 

,u ..
‘bld

 

 

38
than about 100°C. We were able to cool the reactor from
240°C to less than 50°C within 10 minutes. The whole reactor
assembly was then removed from the stand and placed in the
refrigerator after cooling to about 10°C in ice water. The
reactor was allowed to cool to about 0°C in the
refrigerator, usually by leaving it overnight.

After cooling, the temperature and pressure in the
reactor were recorded and the gases were bled off slowly
into the fume hood to relieve the pressure. The reactor was
then opened and the liquid product was collected, measured

and analyzed.

2.2.6. Product Analysis

Crude liquid reaction products were filtered to remove
particulate using disposable syringe filters, then mixed
with a standard solution containing an internal calibration
standard, 2-propanol, and a column conditioner, oxalic acid.
Analysis was performed using a 1.0 ul sample injected into
a glass 2 mm x 2 m 4% Carbowax 20M, 80/120 Carbopack B-DA
gas chromatograph column from Supelco, Inc. The column is
installed in a Varian 3700 GC with a flame ionization
detector and helium carrier gas. Chromatograms were recorded
on a Hewlett Packard HP3394 Integrator. Products analyzed by
this system include acetaldehyde, ethanol, acetone, methyl

acetate, 2-propanol, acetic acid, 2—butanone, acetol,

 

u

. A
fic'YU‘F‘

'vlt'v~..iu -

Perkin—E

K
a i
nY‘nn“
~\4_~H
l
. (
dhA‘k‘ ‘
M y .
‘V"~"v' C.
< ‘\..
W'A
~ g
3“») h“;
dA‘~‘
Q‘-
“"u

 

 

L)
,1"
1

39
2,3-pentanedione, propanoic acid, acrylic acid and lactic
acid.

Early gas samples were analyzed using a 1/8 inch x 6 ft
Spherocarb 80—100 mesh in stainless steel GC column from
Supelco, Inc. installed in a Perkin—Elmer 8500 GC with a
thermal conductivity detector. Results were printed on a
Perkin~Elmer GP—100 Graphics Printer. This GC setup detects
nitrogen, oxygen, carbon monoxide, carbon dioxide, methane,
ethane, ethene and acetylene using helium carrier gas.
Later, gas samples were analyzed using a 1/8 inch x 5 ft
80/100 Carbosieve SII in stainless steel GC column from
Supelco, Inc. installed in a Varian 3300 GC with a thermal
conductivity detector. Results were recorded on a Sargent-
Welch XKR Recorder with an electronic integrator. Using
helium carrier gas, this GC system detects nitrogen, oxygen,
carbon monoxide, carbon dioxide, methane, ethane, ethene and
acetylene. The GC operating parameters for both liquid and
gas analysis are presented in Table 2.2. GC results were
entered into a spreadsheet in which mass balances, product
yields and selectivities were calculated.

During the early stages of our studies of lactic acid
products, we experienced considerable problems with analysis
of lactic acid in the GC column. The GC peaks were not
reproducible and calibration curves were nonlinear for

lactic acid. This led to large errors in mass balances. This

d

u q~~n
_-U.A-
.

1.
«v 4 l
p C 2‘.
a)» .‘1‘
2.. LL
.‘lu .. .o
a!” fin
T... v x

9, z ”m
{Sauna
.

nal C-

Si

a
“fi’(’\
~v-\;‘

A. o 4
L x. .—
«TC . 4.
C» b.
,7 . ..
F“ r
. . e

1;..-

-4.- y

yg-«m
“u‘-\~
.

F.

flak f \

u v
1 i a .
2h 3..
vi. . .
.71 L.
NI. \ u

4. AN.
Nho ‘ls

.h.~
... . r
AH). Hba

~i
lV‘
So

§

q
n...
\—

~ . ; 4
Y\ H~¥
A- ‘ .‘
«x»
p u «\v
. a .. ‘

‘uv‘

40

Table 2.2. GC Operating Parameters

 

Liquid Analysis Column

Supelco 4% Carbowax 20M 80/120
Carbopack B—DA in glass

 

InjeCtor Temp.
Initial Column Temp.
Ramp

Final Column Temp.

Detector Temp. (FID)

200°C
100°C/0 mins
25°C/min
200°C/9 mins
200°C

 

Gas Analysis Column

Supelco 80/100 Carbosieve SII
in Stainless Steel

 

Injector Temp.
Initial Column Temp.
Ramp

Final Column Temp.
Detector Temp. (TCD)

TCD Filament Temp.

200°C
50°C/2 mins
20°C/min
200°C/2 mins
200°C
300°C

 

problem was found to be mostly due to holdup of high-boiling

lactic acid in the dead space just beneath the septum on the

inlet port to the column. Results were much more consistent

and mass balances improved when we changed our injection

technique by taking the following steps:

1) During sample injections,

inserted,

the plunger is depressed quickly,

the injection needle is

then the needle

is left inserted in the septum for one minute before removal

to allow the high boiling lactic acid to vaporize in the

Column.

Immediate removal of the needle deposits a "drop" of

lactic acid in the dead space beneath the septum, which

Slowly elutes into the column upon vaporization,

leading to

 

‘fiA.A\'!~
3‘1 “I
Juvvwv
a

3. l‘
C’V‘I‘ ‘n‘y.
~04.-4’.~-

L_A_J

vofi‘V‘f‘lq‘. A
‘*f“‘-u“\,

1“,

V“.
l. i -
I: "”1"“

m‘
'w,
“5.:
YA“.-
~
row-.qs-’
‘¥\‘. fin
“\r
‘ \ .
2'1A a
‘~" ~ .
‘V L
r. t
(Czar
\r‘ o\ -
‘ ‘15“ ,—
\A
.
Ya,‘
u‘d.‘ I».
.
i.
" 1
"5".
“so. Vb.."‘>.
.‘VC

41
an elevated baseline, poor reproducibility and nonlinear
calibrations.

2) At startup, injection of a standard solution of 17%
aqueous lactic acid saturates the column and brings it to
equilibrium quickly.

3) Repeated injections of the same sample (usually at
least two injections).

4) Use of an internal standard, 2-propanol, insures
reproducibility.

Carbon balances are typically within 10% using this

technique.

2.2.6.1. Reproducibility of Product Analysis

The reproducibility of GC analysis was tested using 6
repeated runs of a representative sample and is presented in
Table 2.3. The standard deviations of all major product
peaks are between 1.5 and 7.5% of the mean value of the
area. The standard deviation of the lactic acid peak is

highest, at 14% of the mean value.

2.3.

2.3.1

CC CC A t A t
n. . .4 4 4 ~ L...
r .. a .. Q. 4 .
A v a). .4 4 Qt
.. n) t w. 4 nxu

 

42

Table 2.3. Reproducibility of GC Analysis

 

 

Compound Standard Deviation (% of mean)
Acetaldehyde 4.7
2—Propanol 6.8
2,3-Pentanedione 6.2
Propanoic acid 6.7
Acrylic acid 1.5
Lactic acid 14.1
CO 7.5

co2 .2

 

2.3. Results

2.3.1. Definitions
The following definitions apply for discussions in this
paper:

Conversion (BOF):
1 - (moles of reactant recovered/moles of reactant fed)

Conversion (Adj): Conversion (BOF) + Error(%C)

Yield of A =
(equivalents of A recovered/moles of reactant fed)

Selectivity of A =

(equivalents of A recovered/equivalents of all
products recovered)

2.3.2. Effect of Temperature in Blank Runs
Blank runs of 34 wt% lactic acid with no catalyst were
performed at 240, 250 and 260°C at a reaction time of 3 hrs.

The products of lactic acid conversion included

m
4;: CO

a:
~x~

V.

1.“.CI

‘vV

ea
3...

«C
R\b

any
4N24
. 4 s

n ..
:4

lik.

 

43
acetaldehyde, ethanol, propanoic acid, acrylic acid, acetic
acid, acetone and acetol. As the temperature was raised from
240 to 260°C, the conversion was seen to double from 5.2 to
9.9%. Acetaldehyde selectivity decreased from 57.3 to 25.6%
as the temperature rose from 240 to 260°C while selectivity
to ethanol rose from 20.9 to 47.9% over the same temperature
rise. Selectivity to acetone and acetic acid were also seen
to increase with temperature. Acetol selectivity dropped off
to almost zero as temperature increased while selectivities
to both propanoic acid and acrylic acid remained about the
same at 13.5—16.5 and 3.7—4.0%, respectively.

Overall, an increase in reaction temperature favors
ethanol and acetic acid formation while a lower temperature
favors acetaldehyde formation. Acrylic acid and propanoic
acid selectivities are unaffected by temperature. High
temperatures also favor acetone while low temperatures favor
acetol, but each of these two minor products contribute less
than 2% to the overall conversion. The effect of temperature
on product distribution in blank runs is presented in Table

2.4.

2.3.3. Effect of Y Zeolite Catalyst on Product Selectivities
Conversion of 34 wt% lactic acid was performed over

1.000 g of Y—82 zeolite catalyst at 240 and 250°C for a

reaction time of 3.0 hours. A comparison of results over Y

Zeolite and blank runs reveals four striking differences

 

a;

, A ""
pm“. 5
U'v‘
.

p—
p—
a v.“

 

Y‘EY‘2V‘r'N‘
..KJ:.¢K.4..V.

1‘

fl
4.4

\

 
 

01
”9‘44.

 

-J
by

 

Inn.
.14.

44

Table 2.4. Effect of Temperature on Product Distribution
in Blank Runs

 

 

 

 

 

 

Temperature (%D 240 250 260
Conversion (Adj)(%) 5.2 6.6 9.9
Yield (%)

Acrylic Acid 0.27 0.36 0.53
Propanoic Acid 0.97 1.42 2.21
Acetaldehyde 4.10 3.11 3.42
Acetol 0.07 0.06 0
Acetone 0.06 0.13 0.26
Ethanol 1.50 3.58 6.40
Acetic Acid 0.19 0.31 0.55
Selectivity (%)

Acrylic Acid 3.7 4.0 3.9
Propanoic Acid 13.5 15.8 16.5
Acetaldehyde 57.3 34.7 25.
Acetol 1.0 0.6 0.01
Acetone 0.8 1.5 l.
Ethanol 20.9 39.9 47.9
Acetic Acid 2.7 3.5 4.1

 

 

* 100.0 g of 34.0 wt% lactic acid feed, reaction time 3 hrs,
no catalyst.

upon addition of the catalyst: 1) a considerable decrease in
acetaldehyde yield, 2) an increase in ethanol yield, 3) an
increase in acetic acid yield and 4) suppression of acetol
formation. Y zeolite catalyst has no significant effect on
conversion or yields to acrylic acid, propanoic acid and
acetone. The effect of Y zeolite on product yields is seen
in Table 2.5.

As seen in blank runs, an increase in reaction
temperature favors ethanol and acetic acid formation, while

decreasing acetaldehyde selectivity.

 

‘ .
LCfiV) ‘ fl
"IV‘J“&;V

9
_A:.. fiv- A
"W vv..C

E20320)

V‘
2.3.4.
.,_.
A.
.A-Y'A -
V'.-Y r‘v‘
V« Ll
3...);
’ A
".0 1",
u.
(“‘9‘ 4
“it C‘r
'—.
~-. fr-
4 4
-]~

(I)

A
' F
L"
A
a
CA ‘
~_~..~4
‘\
k'
'7‘
Mfr" .
4- 7‘»
“‘
in
\‘I L‘
‘ l".
-‘ :3.
‘H

45

Table 2.5. Effect of Y-82 Zeolite on Product Yields'

 

 

 

 

 

 

 

Catalyst 4 none Y-82" none Y—82"
Temperature (°C) 240 240 250 250
Conversion (Adj)(%) 5 2 3.9 6.6 7.3
Yield (%)

Acrylic Acid 0.26 0.23 0.36 0 36
Propanoic Acid 1.0 0.8 1.4 1.7
Acetaldehyde 4.1 1.9 3.1 2.3
Acetol 0.07 0 0.06 0
Acetone 0.06 0.07 0.13 0.15
Ethanol 1.5 1.9 3.6 4.8
Acetic Acid 0.19 0.37 0.31 0.43
* 100.0 g of 34.0 wt% aqueous lactic acid feed, 3 hr reaction time,

** 1.000 g Y-82 zeolite calcined 400°C/16 hr under'hh purge.

2.3.4. Effect of Catalyst Mass on Conversion

The effect of tripling the catalyst mass was examined
over both Y—82 and Y—62 zeolites at 240°C for a feed of 34
wt% lactic acid. The only major effect was an increase in
the ethanol yield upon addition of more catalyst. A slight
increase in the propanoic acid yield and a slight decrease
in the acetaldehyde yield were also seen, but these effects
were small. The effects of catalyst mass are shown in Table

2.6.

2.3.5. Effect of Feed Concentration

Conversion over Y—62 zeolite was studied at lactic acid
feed concentrations of 8.5 and 36.0 wt%. Table 2.7. reveals
that increasing the feed concentration produces an increase

in the selectivity to acetaldehyde and ethanol while

h: .C 2.» Cs .

")4. n4. Y. Y. .4; rl. w A 4). w“ p... r. . o u
. l . 3.. 5.. .Au i z. 24 .3 54 «J n... . . . u .

.3 .G “N. u . . . .4. no. .. t . . r. .3 4). . . x

5 ~ 5 .v wk w u aw. ‘ A v 2.. Au.- n~,. 4» u G.

.43 :4 2.. A... .4 . at v . .4. t n t A.» s .. .. o I

A)» a. .t .( . h t ".4. 4..“ hi... 4.5 4...“ w ..4 T. a I... I !

 

 

Table 2.6. Effect of Catalyst Mass on Conversion over Y

46

 

 

 

 

 

 

 

 

 

Zeolites'
Catalyst" Y-82 Y-62
Catalyst Mass (g) 1.00 3.00 1.00 3.00
Temperature (°C) 240 240 240 240
Conversion (Adj)(%) 3.9 4.9 3.7 4.1
Yield (%)
Acrylic Acid 0.23 0.24 0.22 0.20
PrOpanoic Acid 0.82 1.05 0.98 1.16
Acetaldehyde 01.88 1.66 1.48 1.27
Acetol 0 0 0.09 0
Acetone 0.07 0.08 0.06 0.06
Ethanol 1.87 3.41 1.55 2.13
Acetic Acid 0.37 0.28 0.52 0.54
* 100.0 g of 34.0 wt% aqueous lactic acid feed, 3 hr reaction time,

** Catalysts calcined 400°C/l6 hr under Eb purge.

decreasing selectivity to propanoic acid, acetol, acetone
and acetic acid. No effect was observed on the selectivity

to acrylic acid.

2.3.6. Comparison of Catalysts at 240°C

Conversion of 34 wt% lactic acid was compared at 240°C
over a blank run as well as Y-82, Y—62 and HZSM—S zeolite
catalysts for a reaction time of 3 hours. Catalysts were
found to have little effect on conversion, which hovered at
3.7—5.3% for all runs. Acrylic acid yield was slightly
suppressed by both Y and HZSM—S catalysts, with the more
acidic HZSM—S having the greatest effect. Acetaldehyde yield

was cut in half over Y zeolites and decreased 25% over

47

Table 2.7. Effect of Feed Concentration on Product
Distribution over Y—62 Zeolite

 

 

 

 

 

 

 

Lactic Acid Feed Concentration (wt%) 8.5 34.0
Temperature (°C) 240 240
Conversion (Adj)(%) 4.8 3.7
Yield (%)

Acrylic Acid 0.28 0.22
Propanoic Acid 2.16 0.98
Acetaldehyde 0.86 1.48
Acetol 0.15 0.09
Acetone 0.12 0.06
Ethanol 1.22 1.55
Acetic Acid 1.05 0.52
Selectivity (%)

Acrylic Acid 4.7 4.6
Propanoic Acid 37.0 19.9
Acetaldehyde 14.7 30.2
Acetol 2.6 1.8
Acetone 2.1 1.1
Ethanol 20.9 31.7
Acetic Acid 18.0 10.6

 

 

* 100.0 g aqueous lactic acid feed, 3 hr reaction time,
catalysts calcined 400°C/l6 hr under N2 purge.

HZSM—S. Ethanol yield was increased by both Y and HZSM—S,
again with HZSM-S producing the greatest increase. Acetic
acid yield was also increased by a factor of two over HZSM-S
and increased by a factor of 3-4 over Y zeolite. Zeolites
were observed to have little effect on yields of propanoic
acid, acetol and acetone. Product yields and selectivities

over the catalysts are compared in Table 2.8.

3
A8 a '
‘. ‘-
IV
a‘yS".
A
“.1.
_ 11.1.1.
“1
.h L
ddijlr
”'r
“3::
Y5‘
‘N 'Vk
‘9‘».

\

h-.-
'ucp
a
,Y-‘V‘
s Lab
\
Q-ua‘u
‘ivvvv

l

‘T‘

N;

CL
A. .u

 

48

Table 2.8. Conversion over Zeolites at 240°C’

 

 

 

 

 

 

 

 

 

Catalyst" None Y—82 Y—62 HZSM-S
Conversion (Adj)(%) 5.2 3.9 3.7 5.3

Yield (%)

Acrylic Acid 0.27 0.23 0.22 0.19
Propanoic Acid 0.97 0.82 0.98 1.03
Acetaldehyde 4.10 1.88 1.48 2.99
Acetol 0.07 0 0.09 0.05
Acetone 0.06 0.07 0.06 0.08
Ethanol 1.50 1.87 1.55 2.53
Acetic Acid 0.19 0.37 0.52 0.37
Selectivity (%)

Acrylic Acid' 3.7 4.5 4.6 2.6

Propanoic Acid 13.5 15.7 19 9 14.2

Acetaldehyde 57.3 35.9 30.2 41.3

Acetol 1.0 0.0 1.8 0.6

Acetone 0.8 1.3 1.1 1.1

Ethanol 20.9 35.7 31.7 35.1

Acetic Acid 2.7 7.1 10.6 5.1

* 100.0 g of 34.0 wt% aqueous lactic acid feed, 3 hr reaction time,

** Catalysts calcined 400°C/16 hr under N2 purge.

2.4. Discussion

The reaction temperature was seen to affect the overall
conversion. Conversion increased from 5.2 to 9.9% as the
reaction temperature increased from 240 to 260°C. Addition
of catalysts, variation of catalyst mass and the feed

concentration did not affect conversion significantly.

:19 w u “an L» QC «D by
h L T. . . H Y. 1.5
n y 4...? NV. 2.
at .21 g .fiu n, .
.# . 5|» .1. an
.. a my u ‘ fib A19
Any 1...- R \v «Had n 1'

 

an‘,‘~ L.

..a

A. V1
UH:

‘nu .nu
A; ..
r. A...
hwa Edy

1

1»
~‘

49

Tripling the catalyst mass for Y—82 and Y—62 zeolites
caused an increase in the ethanol yield, with no other
effects observed.

Increasing the lactic acid feed concentration from 8.4
to 34 wt% increased yields to acetaldehyde and ethanol and
cut yields to propanoic acid, acetol, acetone and acetic
acid nearly in half. Little change was noted in the overall
conversion.

Acrylic acid yield decreased slightly with increasing
acidity of the catalyst. The yield of acetaldehyde was
suppressed by both Y and HZSM—S zeolites, with Y zeolites
producing only 36—46% as much acetaldehyde as blank runs.
Both Y and HZSM—S zeolites increase ethanol production, with
HZSM-S producing the highest yield. Acetic acid yields were
increased by a factor of 2—3 over zeolites, when compared
to no catalyst. Catalysts had little effect on yields of

propanoic acid, acetol and acetone.
2.5. Summary

Conversion of aqueous lactic acid over Y and HZSM—S
zeolites were studied in a batch reactor at 240—260°C and
were found to produce mostly acetaldehyde and ethanol at
selectivities of 30—42%, as well as propanoic acid at 14-20%
selectivity. Acetic acid and acrylic acid were formed with

selectivities of 5—11% and 2-5%, respectively. Less than 2%

 

 

“.r. -

“'1‘27-
We

My‘

' v

:9
(I)
'Y

(D

am

"u

50
selectivity to acetol and acetone were also observed.
Overall conversions were 3.7—5.3% at 240°C based on lactic
acid fed.

Zeolite catalysts produced poor selectivities to the
desirable product acrylic acid, and considerable selectivity
to the undesirable products acetaldehyde, ethanol and
propanoic acid. The lactic acid conversions observed were
less than 10% at reaction temperatures of 240—260OC. These
conversions are rather small and indicate that zeolites do
not have much catalytic activity at these temperatures and
under these batch reaction conditions. Overall, zeolites
were found to be poor catalysts for conversion of lactic

acid to valuable products.

Chapte

3.1. I

 

,‘N

1’_ v‘

— ~MJ.“
.

5.

CV
~\\

.

:A;"‘

V‘vu-

3.2

3.2

vs

1|<

.;\

Chapter 3. Lactic Acid Conversion over Phosphate Salt
Catalysts

3.1. Introduction

Reactions of lactic acid were performed in the vapor
phase in a downflow differential packed bed reactor over
both supported and neat sodium phosphate salts. Effects of
temperature and residence time on the yields and
selectivities to products were observed. Product
selectivities and conversion were compared for various

sodium phosphate salts.

3.2. Methods

3.2.1. Materials

Lactic acid was obtained from Aldrich Chemical Company
in an 85% syrup solution in water, from which aqueous lactic
acid feed solutions were prepared by dilution with HPLC
grade water. HPLC grade water was obtained from Sigma—
Aldrich, J.T. Baker, Inc. and Baxter Scientific and was
reported to have a maximum evaporation residue of 3 ppm.
Reverse osmosis (RO) water used in catalyst impregnation was
obtained from a house supply. Sodium lactate was obtained
from Aldrich in solid crystalline form. Silica alumina
(Si/Al) catalyst support (93% alumina, 7% silica) was
purchased from Johnson-Matthey in cylindrical pellets with

51

52
the dimensions 2.3 mm diameter by 2.3 mm long. NaHthng,
Nafiflxh and Napflfih were obtained from Aldrich in solid lumps
or crystals. Nagfihfljngo was obtained from EM Science in a

solid form. (NaPOfl Naspfixo and Na3pflh were obtained as

practical grade solids from Sigma. Reagent grade phosphoric
acid, Hfixh, was obtained from Mallinckrodt in an 85%
aqueous solution. Bovine dental enamel was obtained from the

meat packing laboratory at MSU. Helium carrier gas was

purchased from AGA Gas, Inc. in 99.99% purity.

3.2.2. Catalyst Preparation

Catalysts consisted of inorganic salts impregnated into
silica-alumina supports. Before impregnation, catalyst
support pellets were crushed with a mortar and pestle until
all pellets were at least broken in half. All particles
retained in a #16 USA wire mesh sieve were used as catalyst
support. The catalyst salt was added to RO water in a glass
beaker and stirred to dissolve. Heating on a hotplate was
used to speed dissolution, if necessary. Crushed catalyst
support was added to the aqueous salt solution and stirred
to wet all the support. The salt was dissolved in sufficient
water to completely cover the catalyst support. In all
cases, the aqueous salt solution was immediately taken up
into the support to replace air, as evidenced by the
vigorous expulsion of tiny air bubbles upon addition of the

support. The slurry of catalyst pellets was heated on a

.
Hnbfi
l
“va-
.

OI.
~"QF ‘
H..- J. .

r—v-I

‘
‘.
I

 

YQ‘
V‘fif‘s’x
\-
v
_.
~ 1
‘ fl.‘
\ I
J”
'h‘
4 :V‘;
A“. ‘2,
‘5’
s.
5 v

53

hotplate with occasional stirring to boil off the water.
When most of the liquid was gone, the supported catalyst was
removed from the hotplate and placed in an oven. Overnight
heating at about 100°C dried the catalyst completely.
Complete drying was checked gravimetrically. Catalyst
loading was 1.0 mmol/g support unless otherwise stated.
Takeup of catalyst salt into the support was typically
greater than 95%, based on the residue left in the
preparation vessels. After impregnation, the supported
catalyst was then weighed and loaded into a quartz liner
tube which was inserted inside the reactor tube.

Bovine dental enamel used as catalyst was prepared by
boiling in water to remove flesh, then crushing and sieving
to 16 x 30 USA wire mesh. The catalyst was then calcined at

800°C for 1.0 hr in air.

3.2.3. Vapor-phase Reactor

All reactions were performed in a downflow differential
packed bed vapor—phase reactor lined with quartz. The
reactor body consists of a stainless steel tube 19.5" long
,1.25" OD, and 0.55" ID, with a 6.5" long and 0.5" thick
copper heat sink around the heated section to heat the
reactor evenly. The bottom of the reactor is removable to
allow access to the catalyst, which is supported on a coarse
Quartz frit fused into a 19" long and 0.50" OD quartz liner

‘tube. The catalyst and liner tube is inserted into the

 
 

-1. at . . Av L .. r.. r.\
”a." a. v V 1L a .v A v . 1 . . ~ .
:v u . 21¢ ‘ . p C .d V a . a an“
.3 LC 3.. Q c ad it «It. 2 C ’1“
r . Tn ‘nd r . .3 u.“ .C

704 \hU

54
reactor from the bottom. Bypassing of the catalyst bed by
gas is prevented by a Viton O-ring which seals the annulus
between the liner and the inside reactor wall. The inside
reactor temperature is measured inside a quartz thermocouple
well which extends from the bottom of the reactor up to the
bottom of the frit supporting the catalyst. The removable
reactor end is sealed to the main body of the reactor by a
spring—loaded metal to metal seal ring manufactured by the
Helicoflex Company. Stainless steel liquid and gas feed
tubes enter the top of the reactor through a Conax fitting
and extend down inside the quartz liner tube.

Liquid feed solutions are fed from a 50 ml titration
buret with graduations of 0.1 ml, through an in—line
stainless steel filter and into an Eldex HPLC positive
displacement metering pump. Liquid is pumped into the
reactor through 1/16" stainless steel tubing. Helium carrier
gas or air for calcination may be fed to the reactor.

The reactor is heated by a clamshell electric heater.
An Omega PID temperature controller receives feedback from a
thermocouple mounted outside the reactor walls between the
heating element and the copper heat sink. The temperature
inside the reactor is measured by a thermocouple inserted
into the thermocouple well at the bottom of the reactor and
extending to just below the catalyst bed.

Temperature control is achieved by adjusting the

Setpoint of the controlled outer temperature until the

55
temperature inside the reactor achieves the desired
temperature. The outside temperature is about 30°C higher
than inside at an inside reactor temperature of 300°C. An
attempt was made to control the temperature using feedback
from the thermocouple inside the reactor, but the lag time
was too great and instability resulted.

The cooler ends of the reactor are heated to 150-180°C
by heat tape controlled by an Omega controller. The cool
ends of the reactor are heated to prevent condensation and
collection of products in these zones. A diagram of the
reactor setup is presented in Figure 3.1. Typical reactor

operating conditions are given in Table 3.1.

Table 3.1. Reactor Operating Conditions

 

Temperature (°C) 280—350

Pressure (psig) 60

Liquid Feed Concentration (wt%) 36

Liquid Feed Flowrate (ml/min) 0.05-0.50

Residence Time (s) 0.2—8.0

Catalyst Weight (g) 2—6

Catalyst Bed Length (cm) 2.5-7.6

Feed Composition (mol%) 8% Lactic Acid
77% H20

15% He

56

Gas Feed Liquid Feed

  
 
 
 
  
 
 
  
 
  
    
 
 

Conax Fitn'ng

SS Reactor Tube

Copper
Heatsink

Clamshell
Heater

Firebn'ck

 

 

 

 

 

Thermowell
Quart/aw; 'lMetal Seal
Liner .2.: ......... :2 .. :-:

Tube E 2;] ‘—'\i . .
" L Vlton O-Rlng

 

 

 

 

 

‘ +— Effluent Out

 
 
  

Teflon Spacer

 

 

<—lnside Quartz Thermowell

E‘igure 3.1. Vapor—phase Reactor

57
3.2.4. Product Collection

Reactor effluent passes to a pair of stainless steel
ice traps in parallel. One trap accommodates a 40 ml volume
and is used for product collection during reactor
equilibration between runs. The other trap is used for
product collection during a run and accommodates only 10 ml.
It is cleaned between each run. Both traps are immersed in
ice water while in use. Trap selection is made by a set of
two—way valves.

After passing through the ice trap, reactor effluent
gases flow through a filter, gas metering valve and a ball
flowmeter. Gases may then either be collected in a evacuated
gas collection bag for analysis, or allowed to flow through
a bubblemeter and out to the vent. A reactor flow diagram is

shown in Figure 3.2.

3.2.5. Reactor Operation

Each catalyst is loaded into the reactor then heated
under helium flow until the temperature approaches 280°C,
when the lactic acid feed solution is fed at a high flowrate
of 0.5 ml/min. After product flow is established, helium and
lactic acid feed flow rates as well as reactor temperature
are adjusted to the desired values. The reactor is operated
at these conditions for about 15 minutes to allow
equilibration. The reactor run is then commenced with the

collection of liquid product in the ice trap and gas product

58

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

P P
P P
was” HPLC Liquid
pump Feed
Buret
Air He .
Filter
Reactor
Vent
Ice
Bath
Bubble APP
Meter

 

Gas Bag

 

 

IIIIIYIIIIIIIIIIIII

 

 

 

 

 

Vacuum Pump

Figure 3.2. Reactor Flow Diagram

Control
Valve

   

Bali
F lowrneter

  

 

59
in Saran gas bags for a length of time required to collect
2—7 ml of liquid product. This process is repeated at each

set of reaction conditions.

3.2.6. Product Analysis
Methods of product analysis and the reproducibility

thereof are discussed in Section 2.2.6.

3.3. Results

Conversion of lactic acid over supported and neat
phosphate salt catalysts yields the major products acrylic
acid, propanoic acid, 2,3—pentanedione and acetaldehyde.
Minor products include ethanol, acetone, acetic acid,
hydroxyacetone (acetol), 2-butanone, pyruvic acid and
unknowns.

The condensation product, 2,3—pentanedione, was first
observed as a product of lactic acid condensation in our
laboratory. The identity of 2,3—pentanedione was confirmed
by GCMS and by matching the residence time in CC analysis.
The combination of elevated pressures, low temperatures and
basic catalysts leads to observed 2,3-pentanedione
selectivities approaching 60% based on conversion of lactic

acid.

60
3.3.1. Definitions
The definition of "other" products includes all minor
products listed above except acetol, which is listed
separately, and unknowns, which are not included.
Conversion, yield and selectivity are defined in Section

2.3.1.

3.3.2. Reactor Characterization

A series of experiments was performed to establish the
catalytic effect of the phosphate salts on conversion of
lactic acid. Reactions were first carried out in a tubular
pressure reactor made of 316 stainless steel with a
stainless steel liner in a horizontal flow configuration.
This configuration led to poor mass balances due to
excessive coking caused by incomplete vaporization of the
high boiling lactic acid (b.p. 190°C). The reactor was
changed to a vertical configuration to insure complete
lactic acid vaporization.

Experiments were performed on 0.43 mmol NafiKh/g
silica—alumina support in this vertical configuration to
determine the effect of the stainless steel and quartz
liners. Results indicate that the quartz liner has a
dramatic effect on the selectivity to all desired products.
Conversion (BOF) at 280°C increased from 10% in quartz to
71% in stainless steel. Selectivity to acrylic acid triples

and the 2,3—pentanedione selectivity increases by a factor

61
of four in quartz. In contrast, the propanoic acid
selectivity decreases an order of magnitude and the
acetaldehyde selectivity is cut in half in quartz. Formation
of 2,3—pentanedione and acetaldehyde account for most of the
CO and.CI5 formed with the quartz liner, but in stainless
steel, formation of these two products accounts for only
about 70% of the gases. This indicates either significant
cracking or coking of lactic acid or secondary reactions. A
comparison of results in the stainless steel and quartz
liner tubes is presented in Table 3.2. Unless otherwise
stated, all reactions reported below were performed with the
quartz liner in place.

The effect of the reactor liner and support was
examined by performing experiments in the empty quartz liner
tube and over blank Si/Al support with no salt impregnated.
Conversion for the empty blank liner at 0.7 seconds
residence time was less than 5% even for high temperature
runs at 350°C. The only products observed in significant
quantities were propanoic acid and acetaldehyde. At 300°C,
there was virtually no reaction with an adjusted conversion
of less than 1%. The Si/Al support gave conversions from 4%
at 280°C to 19% at 350°C. As with the empty blank, the Si/Al
support exhibited selectivity to propanoic acid and
acetaldehyde with acetaldehyde selectivity increasing up to

62% at higher temperatures of 350°C. Measurable yields of

62

Table 3.2. Comparison of Results in Stainless Steel and
Quartz Liner Tubes

 

 

 

 

 

 

 

 

Catalyst Na:.,PO4 on Si/Al Na3PO4 on Si/Al
(SS) (Quartz)
Temperature (°C) 280 280
Residence Time (s) 1.60 1.53
Error (%C) -32.8 7.29
Conversion (BOF)(%) 70.6 10.4
Conversion (Adj)(%) 37.9 17.7
Yield (%)
Acrylic Acid 1.4 3.4
Propanoic Acid 9.3 0.6
2,3—Pentanedione 2.3 8.2
Acetaldehyde 6.3 2.5
Acetol 1.1 1.6
Other 1.1 0.9
co' 1.6 1
co; 8.7 4.4
Selectivity (%)
Acrylic Acid 6.4 19.8
Propanoic Acid 43.4 3.4
2,3-Pentanedione 10.9 47.4
Acetaldehyde 29.3 14.7
Acetol 5.0 9.4
Other 5.1 5.4

 

 

* Yields of gases are reported

as moles gas/moles lactic acid fed.

63
all the major products were observed at this temperature.
The catalytic effect of the blank Si/Al support and the

empty quartz tube are presented in Table 3.3.

3.3.3. Catalytic Effect of Phosphate Salts

Studies at three different residence times revealed
that the addition of Nafixh to the Si/Al support produced a
catalytic effect with strong selectivity to acrylic acid and
2,3~pentanedione. The yields of acrylic acid and 2,3-
pentanedione were increased by an order of magnitude over
the phosphate catalyst while the yield of propanoic acid
remained about the same. The yield of acetaldehyde stayed
the same or even decreased at higher temperatures of 320—
350°C.

The phosphate catalyst produced double the conversion
of the blank support at 280-320°C and contact times of 0.4—
4.0 seconds. As the temperature increased to 350°C, the
phosphate catalyst continued to affect more conversion than
the support, but the difference was not as pronounced.
Product yields for lactic acid reaction over NafiKh
supported on Si/Al and over blank Si/Al support as a
function of temperature are shown in Table 3.4. for three

different residence times.

64

Table 3.3. Catalytic Effect of Si/Al Support on Product
Yields for Lactic Acid Conversion

 

 

 

 

 

 

 

 

 

 

Catalyst Empty Quartz Tube Blank Si/Al Support
TW°C> 300 350 300 350
Residence Time (s) 0.69 0.62 0.46 0.42
Error (%C) 9.1 2.4 0.8 1.3

Conversion (BOF)(%) —8.4 1.7 5.7 17.4

Conversion (Adj)(%) 0.7 4.2 6.5 18.6

Yield (%)

Acrylic Acid 0 0.11 0.07 0.79
Propanoic Acid, 0.09 1.26 0.52 1.21
2,3—Pentanedione O 0 0.22 0.28
Acetaldehyde 0.12 1.47 1.49 9.00
Acetol 0 0 0.09 0.37
Other 0.25 0.19 0.41 2.87
co' 0 1.01 1.11 10.57
co; 0.43 2.56 0.24 1.95
Selectivity (%)

Acrylic Acid 0 3.7 2.4 5.5

Propanoic Acid 20.0 41.6 18.6 8.3

2,3-Pentanedione 0 0 7.9 1.9

Acetaldehyde 25.6 48.5 53.3 62.0

Acetol 0 0 3.3 2.5

Other 54.4 6.2 14.5 19.8

 

 

 

* Yields of gases are reported as moles gas/moles lactic acid fed.

.Uwu been ofluoca meoE\mmm meoE mm Uwuuomwu who momma mo mcamflw 4*
.m no.0imm.o mEflu mocmpflmwm «

 

 

 

 

 

 

 

 

 

 

 

N.w b.v m.w m.w w.mH b.ma m.va m.mfi umnuo
m.oa m.m b.m h.H m.m o m.m o Houmom
b.ma o.bH b.wa m.mH o.mm v.mv m.mm D.Hm mphnwpaoumom
m.mm m.mm m.mv m.mv m.H m.m m.b o mcoflbmcoucmmim.m
m.m m.m m.m H.NH m.w m.bm m.mH o.mm Ufiofl UHOQMQOHQ
b.mm o.mm m.mH m.mH m.m m.m w.m b.m Uflod UHH>HU<
vahbfl>fluomamm

mH.OH He.m em.6 mm.m mm.e me.o em.o mm.o .woo
Room.m ov.H mm.o mv.o hm.oa ha.m Ha.a mm.o :OU
ava.H No.0 mw.o mv.o bw.m hm.o av.o hm.o Hwnuo
Hm.m mm.H wm.o mo.o rm.o o mo.o o Houmom
ha.m Hm.m mm.H mb.o oo.m mH.N mv.H mm.o mc>£mpamumod
vo.h mw.m wm.¢ mm.m wm.o ma.o mm.o o mcofipmcobcmmim.m
bv.H hm.o hm.o Hm.o HN.H mm.a mm.o om.o baud UHOammonm
mm.m oo.v mm.H Hm.o mb.o cm.o no.0 mo.o Uflod Uflaxhod
choamfls

m.mm m.ma H.¢H b.m m.wa o.m m.m m.m AwVAwpdv Cofimum>co0
m.mm m.ma m.mH m.mi ¢.ha N.H b.m bw.oi AwVAmomv coflmum>cou
m.bi m.v m.mi o.NH m.H m.h m.o m.v AUwV Hounm
omm omm oom 0mm omm 0mm com 0mm Auov mHSDMHMQEmB
amid coJOmmmz ”Condom Hmim xcoam b93638

 

 

.ouOQQSm H<\flm 0o :Pflaz no aofluaooa no moamfl» nosooum so powwow

oaosamnao .a.m wanes

.pmw U000 000000 mmHOE\mmm mmHOE mm pmuuoamu mum mmmcm mo mcamflw 00

.0 00.0-00.0 0200 000000000 4

 

 

 

 

 

 

 

 

 

 

0.0 0.0 0.0 0.0 0.00 0.00 0.00 0.0 00000
0.0 0.00 0.00 0.0 0.0 0.00 0.0 0.0 000000
0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 000000000000
0.00 0.00 0.00 0.00 0.0 0.0 0.0 0.0 000000000000-0.0
0.0 0.0 0.0 0.0 0.0 0.0 0.00 0.00 0000 000000000
0.00 0.00 0.00 0.00 0.0 0.0 0.0 0.0 0000 0000000
000 >00>0000000

00.00 00.00 00.00 00.0 00.0 00.0 00.0 00.0 .000
00.00 00.0 00.0 00.0 00.00 00.00 00.0 00.0 :00
600.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00000
600.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 000000
00 00 00.0 00.0 00.0 00.00 00.00 00.0 00.0 000000000000
00.0 00.00 00.00 00.0 00.0 00.0 00.0 00.0 000000000000-0.0
00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 0000 000000000
00.0 00.00 00.0 00.0 00.0 00.0 00.0 00.0 0000 0000000
000 00000

0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 00010000 000000>000
0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 00000000 000000>0oo
0.00- 0.00- 0.00- 0.00 0.00- 0.0 0.00- 0.0 0000 00000
000 000 000 000 000 000 000 000 are 00000000000
04\00 00 #0002 0000000 0<\00 00000 00000000

 

 

 

0cmscflucouv
.0000000 0<\0m 00 :Pfl0z 00 00000000 00 000000 0000000 00 000000 00000000o .0.0 00000

.000 0000 000000

mmHOE\mmm mmHOE mm pmuuoamu mum mmmmm uo mcaw0w 00

.0 00.0-00.0 0000 000000000 0

 

 

 

 

 

 

 

 

 

 

 

 

0.0 0.0 0.0 0.00 0.0 0.0 00000
0.0 0.00 0.0 0.0 0.0 0.0 000000
0.00 0.00 0.00 0.00 0.00 0.00 000000000000
0.0 0.00 0.00 0.0 0.0 0.0 000000000000-0.0
0.0 0.0 0.0 0.0 0.0 0.00 0000 000000000
0.0 0.00 0.00 0.0 0.0 0.0 0000 0000000
000 000>0000000

00.00 00.00 00.00 00.0 00.0 00.0 .000
00.00 00.0 00.0 00.00 00.0 00.0 :00
~1.00.0 00.0 00.0 00.0 00.0 00.0 00000
600.0 00.0 00.0 00.0 00.0 00.0 000000
00.00 00.00 00.0 00.00 00.00 00.0 000000000000
00.0 00.00 00.00 00.0 00.0 00.0 000000000000-0.0
00.0 00.0 00.0 00.0 00.0 00.0 0000 000000000
00.0 00.0 00.0 00.0 00.0 00.0 0000 0000000
000 00000

0.00 0.00 0.00 0.00 0.00 0.00 00000000 000000>000
0.00 0.00 0.00 0.00 0.00 0.00 00000000 000000>0oo
0.0- 0.0- 0.0 0.0- 0.00- 0.0 .0000 00000
000 000 000 000 000 000 00.0 00000000000
00\00 00 00002 0000000 00\00 00000 00000000

 

Aconcflucouv
.uuommsm 043% 00 000002 00 0000000000. mo $00.00.». 00500000 so uommmm 000.000.0000 .00.m 0090;...

68

3.3.4. Studies of Catalyst Loading over Supported Salts
and Neat Salts

Conversion of lactic acid was performed over various
loadings of NafiKh supported on Si/Al and over neat Nafixh.
The effects of catalyst loading and support are seen in
Table 3.5.

No significant trends were observed in either the
yields or selectivities to liquid products over Nafixh
supported on Si/Al at loadings of 0.43, 0.86 and 1.00 mmol
salt/g support. A slight increase in the conversion was
observed upon increased loading at 280°C, but the effect is
small. Conversion increased from 17.7 to 26.3% upon more
than doubling the loading at 280°C. Selectivities to
propanoic acid change significantly, but the changes appear
to be unrelated to catalyst loading.

Significantly lower conversions were observed over the
neat salt, compared to the supported salt, at all
temperatures studied. Conversions over the neat salt were
typically half of those observed over the supported salt,
although this difference became less pronounced at the high
temperature of 350°C.

Employing a support enhanced the catalyst activity, but
the selectivity to products remained unchanged. This
suggests that the support serves only to increase the
surface area of the catalyst and is not involved
mechanistically in the catalysis. The neat salt was found to

have melted onto the liner walls and frit upon removal of

.Umu @000 000000 mm0OE\mmm mmHOE mm Umunommu mum mmmmm wo mpawflw 00

.m 00.onmm.o 0E0u

wocmoflmwm 0

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0.m 0.0 0.0 0.0 0.0 0.0 0.0 m.m 0.00 0.0 0.0 00000
0.00 0.0 0 0.00 0.0 0.0 0.0 0.0 o 0.0 0.0 000000
m.om 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 000000000000
0.m0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 000000000000|m.m
0.00 m.m 0.0 m.m 0.0 0.00 0.00 0.0 0.0 0.00 0.m 0000 000000000
0.00 0.00 0.00 0.0m m.00 0.00 0.00 0.00 0.00 0.00 0.00 0000 0000000
000 000>000000m

mm.0 00.0 0m.o 00.00 00.0 00.0 mm.00 mm.m 00.0 00.00 00.0 .m00
00.0 00.0 00.0 00.0 00.0 m0.o 0m.m 00.0 00.0 00.m o :00
mm.0 00.0 00.0 00.0 00.0 00.0 00.0 0m.0 mm.o 00.0 00.0 00000
0,oo.0 00.0 o 00.0 00.0 00.0 00.0 00.0 o 00.0 00.0 000000
0000.00 00.0 00200 00.0 00.0 0:200 00.00 00.0 000:0 00.0 00.0 000000000000
00.0 00.0 00.0 00.0 00.0 00.0 00.0 m0.m 00.0 00.0 00.0 000000000000-m.m
00.0 00.0 00.0 00.0 00.0 00.0 00.00 00.0 00.0 00.0 00.0 0000 000000000
00.0 00.0 00.0 mm.0 00.0 00.0 00.00 00.0 00.0 00.00 00.0 0000 0000000
000 00000

0.0m 0.0 0.0 0.0m 0.00 0.0 0.00 m.00 0.0 0.00 0.0 00000000 000000>000
0.00 0.0- 0.00- 0.00 0.00 0 mm. 0.00 0.00 mr0:. 0.00 0.00 00000000 000000>000
0.00- 0.00 0.00 0.0- 0.0- 0.00 0.000 0.00- 0.00- m.mm- 0.00- 000000000
omm 00m 000 00m 00m 000 omm com 000 omm com 30000000000000
AuquQSm m\aoEEV

0\z 00.0 00.0 m0.o 0000000

000002 0002 0000 00 0.00002 00000000

 

.000QQ5m paw 0:0cmoq um>00000 mo bomwwm .m.m 00305

.000 0000 000000 mm0oE\m00 00008 00 00000000 000 00000 00 00000» *0

.0 00.0-00.0 0000 000000000 0

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0.0 0.0 0.0 0.0 0.0 0.0 m.m 0.0 0.0 m.m 0.m 0.0 00000
0.00 0.00 0.0 0.0 0.00 0.0 0.0 0.0 m.m m.0 0.00 0.0 000000
0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 000000000000
0.0 0.00 0.00 0.m0 0.00 0.00 0.0 0.00 0.mm 0.0 0.00 0.00 000000000000-m.m
0.0m 0.0 0.0 0.0 0.0 0.0 0.00 0.00 0.0 0.0m 0.00 0.0 0000 000000000
0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0000 0000000
000 000>000000m

00.0 00.0 00.0 00.00 00.00 00.0 00.00 00.00 00.0 00.00 00.0 00.0 .m00
00.0 00.0 00.0 m0.mm mm.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 :00
00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 mm.m 00.0 mm.o 00000
0000.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 m0.0 000000
P100.000 00.0 00.0 00.00 00.0 00.0 00.00 00.00 00.0 00.00 00.0 00.0 000000000000
00.0 00.0 mm.m mm.m 00.00 00.0 00.0 mm.m0 00.00 00.0 00.00 00.0 000000000000-m.m
0m.00 00.0 mo.o mm.m 00.0 00.0 00.00 00.0 00.0 00.00 00.0 00.0 0000 000000000
00.0 00.0 00.0 00.0 0m.0 00.0 mm.m 00.0 mm.0 00.0 00.0 00.0 0000 0000000
000 00000

0.00 0.00 0.0 0.00 0.00 m.0m 0.00 0.00 m.mm 0.00 0.00 0.00 00000000 000000>000
0.00 0.0m 0.0 0.00 0.00 0.00 0.00 0.mm m.00 0.00 0.00 0.00 00000000 000000>000
m.00- 0.0- 0.0- m.00- m.00- 0.00 0.00- 0.00- 0.00- 0.00. 0.00. 0.0 0000 00000
omm com 000 000 com 00m omm com 000 0mm com 000 00.0 00000000009
0000Q030 o\0OEE0

0\z 00.0 00.0 m0.o 0000000

100002 0002 0000 00 0.00002 00000000

 

00050000000 .0uommsm 000

0000000 00000000 00 000000 .m.m 0000B

71
the catalyst from the reactor.

Nitrogen BET results indicate that the surface area of
the Si/Al support is about 5 HE/g. At our catalyst loadings
of 0.4 to 1.0 mmol/g, there is plenty of salt present to
cover the support surface with many molecular layers,
assuming uniform dispersion of the salt on the support.
Because the neat salt melted and flowed at reaction
conditions, we expect the catalyst to be in a hydrated,
liquid phase on the support during reaction. The dispersion

of the salt on the support has not been measured.

3.3.5. Selectivity over Supported Orthophosphate Salts
Reactions of lactic acid were performed over monobasic,
dibasic and tribasic sodium phosphate salts all supported on
Si/Al. Conversion was found to increase with basicity of the
salt. Increased basicity also increases selectivity to
acrylic acid and 2,3—pentanedione while lowering selectivity
to acetaldehyde and other products. The effect of catalyst
basicity on propanoic acid" and acetol selectivity showed no
clear trend. The effect of catalyst basicity on product

yields may be seen in Table 3.6.

23.3.6. Reactions over Bovine Dental Enamel and Trisodium
Phosphate

Comparisons of reactions over tribasic sodium phosphate
Eind calcium hydroxyapatite in bovine dental enamel showed

ffrom three to seven times greater conversion over Nafixx.

.Umu Uflom ofiuomH mmHoE\mmm mmHoE mm pounommu mum momma mo mpamflw *«

.m mv.o-mm.o wee» moamvflmmm v

 

 

 

 

 

 

 

 

 

 

 

 

 

 

m.e m.v w.m s.e m.m m.mm .m.mfi m.efi umnuo
0.0H s.m m.ofi m.m e.m ¢.s m.m m.m Seamus
s.ma. s.mfi H.0m e.om m.om o.sm o.mm m.mm mamgmnambmom
m.mm m.me H.mm m.me m.mH m.mH m.H m.s mcoflmmamucmm-m.m
m.m w.m m.m m.e N.Hm a.» m.m m.mfi seem unocmmoum
s.mm m.mfi m.fim m.sfi H.mfi m.m m.m e.m aflom Unasuua
Awe >Ufl>fluomamm

mH.OH so.e aw.m em.o mo.m em.o ma.H em.o .moo
sm.m mm.o efi.m mm.o em.a sm.o mm.oa HH.H zoo
MumH.H ma.o NH.H mH.o mm.o em.o sm.m He.o umnuo
Hm.m mm.o mm.m HH.O we.o OH.O sm.o mo.o HOSmom
sfi.m mm.H em.m sm.o s¢.m sm.o oo.m aw.H musnmvamumua
eo.s em.w oa.m mm.H mm.H sfi.o mm.o mm.o mcofivmcmscmm-m.m
ne.fi sm.o mm.m mfi.o ms.fi oa.o HN.H mm.o wave ofloamaoum
mm.a mm.H mm.m mm.o mm.fi mo.o as.o no.0 save unamuum
Awe same»

m.mm H.¢H a.mm w.m s.m m.a s.mH m.m screams coflmum>aoo
m.mm m.mH m.mw m.m m.sa A m.m- e.sfi s.m inclmomv aoflmum>coo
m.s- m.m- H.NH- m.m- a.m- m.e m.H m.o lows uouum
0mm oom 0mm com 0mm oom 0mm com loos musumumaame
22.5 no .862 318 no .oamdmz :28 so .onmmmz xcflm 23m 3338

 

.muamm mumsmmogm um>o mpamfls nonwoum .m.m magma

73
Since the nitrogen BET surface areas of the sodium phosphate
and the hydroxyapatite are S and 6.2 nf/g, respectively,
this indicates NafiXh has an inherently greater activity
since the surface areas of the two catalysts are comparable.
Selectivities to both acrylic acid and propanoic acid were
similar for the two catalysts over all conditions studied.
Naflxh exhibited selectivities to 2,3—pentanedione nearly
four times greater than hydroxyapatite at 280°C and 1.0
second residence time. Raising either the temperature or the
residence time quickly erodes this advantage. Selectivity to
acetaldehyde over Nafixh is one half to two thirds the
selectivity over hydroxyapatite for all conditions studied.
Results of the reactions over tribasic sodium phosphate and

calcium hydroxyapatite are presented in Table 3.7.

3.3.7. Effect of Temperature and Residence Time on
Product Yields

Lactic acid conversion over NafiXL on Si/Al was
performed at a series of temperatures of 280—3SO°C and a
series of residence times from 0.39—4.08 seconds. Yields of
acrylic acid, 2,3—pentanedione and acetol were all found to
go through a maximum at a temperature of 300-320°C for
residence times of two and four seconds. Yields of
acetaldehyde, propanoic acid and other products were found
to increase with temperature for all conditions studied. The
effect of temperature on product yields may be seen in

Figure 3.3.

.pmu Ufium ofiuomH mmHOE\mmo mmHOE mm Umuuommu mum momma mo mpawflw

t.

 

 

 

 

 

 

 

 

 

 

m.m m.m s.m m.HH s.e m.e w.m o.ma m.om e.mm umnuo
m.HH m.m H.mH o.oa m.m s.m R.H o.m m.m o Houmoa
o.Hm m.s >.mm e.Hm o.sfi s.ma m.mfi H.em s.mm m.mm maszmwfimumoa
m.em H.H m.mm e.sm m.©m m.me m.me m.mm m.mfi a.ma mcoflcwamscma-m.m
o.m a.e s.e w.m m.m m.m H.NH m.e H.m m.m anus oflocmaoum
©.Hm m.s m.om o.mH o.mm m.mH m.mH m.sH a.mH s.a nave onesnoa
Awe >Ufl>fluomamm
H©.PH ss.m mm.w om.a H¢.m sm.e mm.m mH.H mm.o mm.o moo
mm.e me.a No.0 mm.o oe.fi mm.o me.o am.o ma.o mm.o .oo
«H.m mm.o mm.o we.o ms.o ae.o me.o mm.o mm.o mm.o umsuo
4.mm.m mo.H me.a He.o mm.H mm.o mo.o mfi.o mo.o o Heumoa
7.mfi.s mo.m mm.m om.H Hm.m am.fi ms.o Hm.o mm.o mq.o maszmaamumum
Hm.fifi so.m me.m mH.H mm.m em.e mm.m mm.o om.o om.o mmoflamchcmm-m.m
HR.H mm.o be.o mH.o sm.o sm.o H©.o NH.o mo.o mo.o Baum UHOCMQOMQ
mm.s HH.m am.H mm.o oo.¢ mm.H Hm.o .Hw.o em.o efl.o Baum uflasuoa
“we came»
m.mv m.mm H.NH m.m m.mH H.efi s.m m.m o.m m.m Aweihwav coflmum>coo
o.mm m.ea m.m H.m- m.mH a.mfi m.m- m.H- e.m- m.m- Aweimomv aoflmum>coo
m.me- o.mH m.m m.s m.e m.m- o.me s.e m.s m.m lows House
oom 0mm com 0mm 0mm oom 0mm 0mm oom 0mm loos upsumHmQEme
mo.m-eo.m ma.m-aa.m we.o-oe.o so.a-mo.fi Ame mane uomucoo
H<\flm CO ngmcm H4\flm CO ngmflm
.862 3.569 .862 $38 $338

 

 

 

 

 

~28ch Hoodoo mean/om cam H4\Hm so sommmz Hm>o Uflom oeuomq mo coamum>cou .b.m manna.

75

H<\Hm co mumcdmocd esflcomflne um>o moamfi» bosooum co musumquEwe mo uowumm .m.m ousoflm

SE: 8538. m N ~< . 8. 2392.80...

omm 0mm can can own own can omN omN Ohm

 

 

 

 

 

 

1 q q d a d J u o
n N
n .v
. .. 1 e V
_opmo< ulol. \ m.
I II-
ou>€288< .11. L m m
. A
mcofimcficmdfid Incl... . or .M
1 p
Scanned loll , my.
. . - «F m
2.22. Isl. .%
P a;
1 m—.
1 w—.

 

 

 

76

This suggests that acrylic acid, 2,3-pentanedione and
acetol are destroyed by secondary reactions which convert
them to other products. No conversion of 2,3—pentanedione
was observed at 2800C, as discussed in Section 3.4.1., but
conversion may take place at higher temperatures. Acrylic
acid may be hydrogenated to propanoic acid, but this does
not account for all the propanoic acid formed. Some
propanoic acid results from direct reduction of lactic acid.
Acetaldehyde and propanoic acid are terminal products which
do not break down to a significant extent.

The highest acrylic acid yield of 15.6% is obtained at
a moderate retention time of one second and a high
temperature of 350°C. The maximum acrylic acid selectivity
is observed at a contact time of 0.4 seconds and 350°C.
Effects of temperature and contact time on acrylic acid
yield are presented in Figure 3.4.

A long contact time of four seconds and a low
temperature of 280°C favors the formation of 2,3-
pentanedione, with a maximum yield of 13.4%.

2,3-Pentanedione selectivity is very sensitive to
temperature. Selectivities of 45-52% are seen at 280°C for
all contact times studied. These selectivities drop to less
than 26% upon raising the temperature to 350°C.

At 280°C, the residence time has little effect on 2,3-
pentanedione selectivity. However, at 350°C, the selectivity

(decreases from 26 to 6% upon increasing the residence time

H<\Hm co mumzmmonm Edwpomflue Hm>o came» baud Ufla>uo< .v.m musmflm

.0. 2322.50...

com omm Gem omm ONm O Fm com omN 63.. Ohm

 

 

77

 

 

 

 

 

 

- q q - q q q _ O
. - N
I l .v
. .. m V
mvElxl m
m.
mNEIaI . m
n m .A
mPEIOI m
. D.
$.oElol . . 1 S w
1 NF
- 3
. m:

78
from 0.4 to 4 seconds. Effects of temperature and residence
time on 2,3-pentanedione yields are seen in Figure 3.5.

The maximum yield and maximum selectivity to acetol are
5.3 and 12.4%, respectively, and both occur at 320°C and 2
seconds contact time.

Propanoic acid selectivity decreases from 12 to 5% with
increasing temperature at a short contact time of 0.4
seconds. This trend reverses at a long contact time of 2—4
seconds, with selectivity increasing from 4.9 to 9.0% upon
raising the temperature from 280 to 350°C.

At temperatures of 280-320OC, acetaldehyde selectivity
is insensitive to contact time, remaining at 15-25%. Upon
raising the temperature to 350”C, the selectivity shows a
strong dependence on contact time. Acetaldehyde selectivity
increases from 19 to 68% upon increasing the contact time
from 0.4 to 4 seconds. Results of lactic acid conversion

over Nafixh supported on Si/Al are presented in Table 3.8.

3.4. Discussion

3.4.1. Mechanisms

Acrylic acid dehydration is proposed to occur via
cycloelimination. Lactic acid condenses with sodium
phosphate to yield lactic acid—2—phosphate which
subsequently cycloeliminates with the phosphate oxygen

attacking the B—hydrogen of lactic acid. With the

H¢\Hm do mumnmmonm ESflUOmHHB Hm>o mama» mcoflpmcmucmmnm.m .m.m mhdmflm

.0. 239an05.

com own 35 0mm own a PM com omN omN CNN

 

 

79

 

 

 

 

 

 

« u u q q q q a O

I .1 N

1 e
V_
W? Fm IIOIII . a . m
l w IOI
mm E larl . m
I o A
2 E |?| m...
1 m D.
med .E loll w/m

1 or

. L we

0
3

 

.Uwu Uaom UauUda mmaoE\mmm mmaoE mm Umuuommn mum mommm uo mpamaw 1

 

 

 

 

 

 

 

 

 

 

 

 

m.m N.m v.m m.m 0.x. m.w m.m ma» NJV man way may m.m umnuo
a.m m.aa ma. \1.m nv.ma maa m.m m.m m.m w.oa m.m 5m ha a0u®0<
mab m.mm a.am 0.2V m.om o.am mica m.mm b.ma b.wa o.>a Fma m.ma mphzmpamumod.
m.m a.mm v.2» w.ma a.mm ijm a.am m.ma mam mlmm m.mm m.mv w.mv mCOaUmCmbcmdum.m
o.m m.m mlw o.m m.h o.m mjv \1.m m6 m.m m.m m.m a.ma page. oaocmmoum
m.m m.ma mama o.am 0.2” m.am m.\1a m.om m.ma 5mm o.wm m.ma N.ma baud oaaxaHU/w
Gav >ua>auomamm

mm.omvm.ma mo.wawv.maam.amam.ba\1b.m bmama hm.\1 ma.oa av.m hméu mm.m .NOU
om.am moa. \Lm.m av.mm oo.m mmau av.a Mafima mm.a hm.m ova mw.o mv.o .OU
WENv ma.m mm.a mm.m mm.m wa.m mm.o Smim mvé ma.a N50 mic male Hmnuo
mm.a mmaw ba.m mm.m emim mm.m mo.a mmaw Sm.o am.m mm.a mm.o mo.o aoumod.
vv.mmmm.oa mm.m mv.ma ow.m wafw mo.m v0.5a m>.a bah. am.m mw.a mb.o mpxanocamnbmom
ha.m av.ma mv.ma mm.m aa.aaam.aa\10.m mm.m mm.m flea. mw.m ama. mm.m chapmcmucmmnm.m
amau Sim mw.a mm.m mo.m ab.a mm.o maim mm.o Zia hm.o >90 am.o Uao< Uaocmaowm
33m ma.w maze mm.m om.aa mm.\1 aa.m m.ma mm.a mm.m ooaw mm.a am.o baud Uaaaod
Aoav Ua®a>

a4; mimm m.mm m.mm m.mm m.maw m.mm whom o.ba w.mm m.ma ajva 52m AwVAhU4VCOamHm>QOU
o.mm mom mam worm m.m> o.mm m.va m.mb man m.mm m.ma m.ma m.mqu:mOmvc0ame>COU
m.b| m4? m.m m.ma|o.valm.mal o.ma a.ma| h.ma m.h| may m.m1 o.ma A061 HOHHm
0mm com owm 0mm 0mm com owm omm owm 0mm omm oom 0mm AUov wuzumquEmB
mo.vumm.m mo.mumm.a ma.a|mm.o vv.olmm.o Amv mEaB uomucou

 

 

 

 

 

diam so godmmz um>o COampm>coo baud oabomq .m.m magma

81
abstraction of the B-hydrogen, a double bond is formed
between the d~ and B-carbons of lactic acid, and the
phosphate group leaves. This process produces acrylic acid,
as well as regenerating the phosphate salt catalyst, which
loses a water molecule and repeats the cycle. This mechanism
is shown in Figure 3.6.

The proposed mechanism for condensation of lactic acid
to 2,3—pentanedione is shown in Figure 3.7. The phosphate
first combines with the acid group of lactic acid to form an
ester. The d-hydrogen of the ester is abstracted by a base
with the resulting enolate anion being stabilized by
resonance. The enolate anion attacks the carbonyl carbon of
another ester resulting in the substitution of the phosphate
group. This substitution product then decarboxylates and

rearranges to yield 2,3—pentanedione.

NoO\ ,ONo
’/P\
0“ NOSPOA (5 O OH
H H OH
0 _ _» Ar + Nc12HPO‘1
0 350°C 0 0

Figure 3.6. Proposed Mechanism for the Dehydration of Lactic
Acid to Acrylic Acid

82

a.
O (Aqueous N PO 0 0"
Ho solution) _:3__4__ HO ’9032- _’_ HO / /p032-
OH 280 °C 0 0

0"s

. 5" a
H‘CL \ J ‘ HO coo- EOH. HOWCOOPOJZ’

  

O
0 Ca or Na :/:~O \
Phos hates
HO p ——+- 0 /
0H 280°C 0 0'
H * Chkal (3
' o

‘H

9032- 0’H Po3 -

’ o
\ o
o- 'C02 0::W coo-h 0
2. H 0

Figure 3.7. Proposed Mechanism for the Condensation of
Lactic Acid to 2,3-Pentanedione

83

In a study of this condensation mechanism, an aqueous
feed solution of acetaldehyde and propanoic acid yielded no
conversion over supported Nafixh catalyst at 280°C and 1.5
seconds residence time. This rules out the condensation of
acetaldehyde and propanoic acid to form 2,3-pentanedione and
demonstrates that there is no simple decomposition of either
acetaldehyde or propanoic acid under these conditions.
Reaction of an aqueous solution of 2,3-pentanedione under
identical conditions produced virtually no products, thus
eliminating the possibility of acetaldehyde and propanoic
acid resulting from 2,3—pentanedione decomposition.

Glycolic acid,<3L(OH)COOH, an analog of lactic acid,
was also studied in an effort to determine if this
condensation mechanism could be applied to an entire class
of d-hydroxy carboxylic acids. If glycolic acid reacts by a
condensation mechanism analogous to that of 2,3—
pentanedione, one would expect to form 2—oxo-propanaldehyde.

A 28.7 wt% aqueous glycolic acid feed solution was
chosen so the molar concentration was identical to the
concentration in the 34 wt% lactic acid feed solutions used
in lactic acid studies. Reaction of the glycolic acid
solution at 280°C and residence times of 0.3 and 1.5 seconds
gave no yield of 2—oxo—propanaldehyde over Nafixh on Si/Al.
The conversion of glycolic acid was less than 5% for both

runs and products were mostly acetaldehyde, acetic acid, CO

84
and C02,.as well as minor amounts of 2,3—butanedione and
unidentified products.

A mixed condensation between lactic acid and glycolic
acid, following an analogous mechanism, would yield 2—oxo—
butanaldehyde and or 2,3—butanedione. An aqueous feed
solution of 17.0 wt% lactic acid and 14.4 wt% glycolic acid
was chosen to provide a combined molar concentration
identical to that used in lactic acid studies. This solution
was reacted under identical conditions as the glycolic acid
feed. The reaction produced no 2—oxo—butanaldehyde and only
minor amounts of 2,3—butanedione.

These results may be rationalized by noting that the
enol which is formed after hydrogen abstraction from the
lactyl phosphate is stabilized by the methyl group, causing
the d—hydrogen to be acidic. This causes the condensation
step of the mechanism to be favorable for lactic acid, but
not for glycolic acid, leading to no formation of the
glycolic acid condensation product, 2—oxo—propanaldehyde. By
electron donation, the methyl group also stabilizes the
anion formed after nucleophilic attack of the enolate anion
on the carboxyl group. The absence of this methyl group in
glycolic acid thus impedes the formation of the enol from
the glycolic acid-phosphate ester and impedes the
condensation pathway.

The observed product of the mixed condensation between

lactic and glycolic acids, 2,3—butanedione, arises from the

85
attack of the lactic enol on the carboxylate carbon of the
glycolic acid ester. Since the glycolic enol is not
stabilized by the methyl group, it is not formed and thus
the product of the reverse condensation, 2-oxo-

butanaldehyde, is not observed.

3.4.2. Mass Transfer Limitations

The reactions in this catalytic system may be
considered free from mass transport limitations according to
the Weiss-Prater criterion. The observable modulus is about
0.04 for the worst case in which a high feed rate and high

temperature are used.

3.4.3. Kinetics

Product yields from the conversion of lactic acid over
Nafixh on Si/Al were analyzed to determine rates and orders
of reaction. Rate constants and kinetic parameters were
calculated by integrating the molar balances of each product
using an in—house software package (KINFIT). Kinetic
analysis indicates that the formation of 2,3-pentanedione is
second order with respect to lactic acid. A second order fit
was much better than an attempted first order fit, which was
nonlinear. Kinetic data suggests a secondary reaction which
converts the dione to unspecified products. This secondary
pathway was included in the kinetic analysis, but the

reaction rate constant was not able to be calculated

86
accurately as a function of temperature.

As discussed in Section 3.4.1., no reaction was
observed when an aqueous solution of 2,3-pentanedione was
fed over NafiX% on Si/Al at 280°C and 1.5 seconds residence
time. This indicates that the secondary reaction of 2,3-
pentanedione is not a simple decomposition at this
temperature. However, we cannot rule out the possibility
that the dione is reacting with another compound, or more
likely, that simple decomposition occurs to a significant
extent only at elevated temperatures.

The formation of acrylic acid was found to be first
order with respect to lactic acid. A secondary reaction
which destroys acrylic acid is occurring and was accounted
for in the kinetic analysis, but as in the case of 2,3—
pentanedione, the rate at which it is occurring cannot be
accurately calculated. Propanoic acid is formed by
hydrogenation of acrylic acid, but this does not account for
all the secondary reaction of acrylic acid. A secondary
reaction to unspecified product is also included. Propanoic
acid is also formed directly from lactic acid via a first
order reduction pathway.

The formation of acetaldehyde was found to be first
order with respect to lactic acid and, as previously
discussed, acetaldehyde does not undergo a simple
decomposition. The reaction pathways are presented in Figure

3.8. Reaction rate constants for the various pathways are

87
given in Table 3.9. Activation energies and frequency
factors for the formation of major products are presented in
Table 3.10. An Arrhenius plot for the primary reaction

pathways is presented in Figure 3.9.

3.4.4. Catalyst Stability over Extended Reactor Run

34 wt% aqueous lactic acid was reacted over calcined
calcium hydroxyapatite catalyst at 320°C and 5.0 seconds
residence time for 102 hours in order to study catalyst
stability. The overall adjusted conversion stayed constant
around 13.7%. Acrylic acid and propanoic acid yields
increased from about 2.2 and 1.1%, respectively, at the
beginning of the run, to 2.9% and 1.4% near the end of the
102 hour run. The yield of 2,3—pentanedione remained
constant at about 2.6% while the yield to acetaldehyde
increased significantly from 1.6 to 3.2%. Both
decarboxylation and decarbonylation increased to account for
the increase in acetaldehyde yield. The average carbon
recovery was 109.2% with a standard deviation of 8.3%. The
results of this extended run are shown in Table 3.11.
Overall, the hydroxyapatite catalyst exhibited good
stability with yields to the desired products acrylic acid

and 2,3—pentanedione remaining stable.

H0

88

Decarbonylation¥

 

Decarboxylation'

Dehydration

 

OH———<

Reduction

 

 

Condensation

 

r

O
/L + co + H20
H + C02 4» H2

Acetaldehyde

MOI-I + H20
Acrylic O
ACId 1H2

WOH + 1/2 02
Propanoico
Acid

+ (302 + 2HZO

2,3-Pentanedione

Figure 3r8. Reaction Pathways of Lactic Acid Conversion over
NagPO4 on Si/Al

89

Table 3.9. Reaction Rate Constants of Lactic Acid Conversion
Pathways

 

Temperature (°C) 280 300 320 350
.021 0.052 0.12 0.49
.0024 0.011 0.010 0.030
.016 0.025 0.11 0.12
.7 18 31 66

 

K1 (Lactic e Acrylic)(84)
K2 (Lactic ~ Propanoic)(s%)

K3 (Acrylic * Propanoic)(s%)

KOOOO

K4 (Lactic ~ 2,3-Pentanedione)
(l/mols)

K5 (Lactic a Acetaldehyde)(s%) 0.017 0.043 0.065 0.13

 

Table 3.10. Activation Energies and Frequency Factors for
Lactic Acid Conversion

 

 

Rate Constant Activation Frequency
Energy Factor
(kJ/mol) (s*)

Kl (Lactic a Acrylic) 133.6 7.80x1010

K2 (Lactic ~ Propanoic) 106.4 2.42x107

K3 (Acrylic ~ Propanoic) 83.1 1.20x106

K4 (Lactic a 2,3-Pentanedione) 78.1 2.31x108'

K5 (Lactic ~ Acetaldehyde) 73.6 1.98x10S

 

* Units are (l/mol-s).

90

(s Ioum) iueisuoo 9193

o_.

oo—.

baud oauomq we m>m3cumm COauomom quEaum now anm msacmcnnd .m.m musmam
ooow x 3: .2..
ma. mhé ha. mm.—. m.—.

 

 

  
 
     

. . . _ 5.0

.953. fin H mm
op>so23oo<

lLlll

._
o
(5]) 1ueisuoo 9193

8.5.3. ¢m_. H mm

Eo< o__>._o< .
.953. 2. H mm

o:o_uo:3:on_-m.N

 

 

91

.Umu paum UauUMa waoE\mmm mmaoE mm Umuuomwn mum momma uo mpama» at

.mEau buoucou m mw.m1ar.v .Uoomm «

 

 

 

 

 

 

 

0.0 0.0 0.0 0.0 0.0 0.0 0.0 m.a. 0.0 0.m 0.0 0.0 0.0 «.0 “0:00

0.0a 0.0m 0.0H 0.0a 0.0a 0.0a p.0a 0.0a 0.0a 0.0a 0.0m 0.0a 0.0a 0.0a Hoomca

m.mm H.0 0.0a 0.0m 0.mm m.0m 0.0 0.HH 0.0a 0.0a 0.0 m.0a 0.0m 0.0H acacmcacocua

m.mm 0.mm v.00 0.Hm 0.0N 0.0m N.0N N.mm 0.0m N.am m.mm 0.0m ~.am H.0N cacaccccccma1m.m

0.HH 0.ma 0.NH m.aa H.0H 0.0a 0.NH «.ma m.aa 0.0 m.ma N.HH N.0 0.0a cacc caoccaoua

0.Nm 0.0m 0.0m 0.0m 0.0m m.mm 0.Hm 0.00 0.NN 0.0a 0.0m m.mm 0.0a 0.0m caca caaaoca

lav soa>aocmamm

Hm.H 1 00.H 1 a0.a 1 00.H 1 00.0 1 N0.H 1 0H.H 1 .moo
00.0 1 00.0 1 am.0 1 H0.0 1 00.0 1 00.0 1 0H.0 1 :00
00.0 00.0 00.0 00.0 00.0 mm.0 00.0 00.0 00.0 00.0 00.0 H0.0 00.0 0m.0 cacao
mm.ml mm.m a0.a 0H.m 00.H 00.H am.a 0N.N am.a 00.H NH.N 00.0 a0.a b0.a Hoomoc
ma.m 00.0 ma.m 00.0 00.0 ma.m 00.0 Hm.a 00.a mm.m 00.0 00.2 00.0 mm.a acacccacomca
Ha.m mm.m 00.N H0.m m0.m ~0.N Hm.m 00.m 00.N 00.m 0m.m Hm.m 00.0 mm.m occacmccocma1m.m
00.H ma.H ma.a mm.a am.a ma.a 0H.H am.a HH.H am.a mm.a 00.H 00.0 00.0 .caca caoccacta
0H.m «0.0 m0.m Ha.m .aa.m 00.0 m0.~ 00.N_ NN.N 00.0 m0.m Ha.m .00.m NH.N caca oaasoca
lac camaw

m.ma 0.NH m.ma H.ma_ 0.0a m.aa. 0.Ha 0.0H 0.0a 0.0a H.0a 0.0a m.aa 0.Han laciacaocOamLm>coo

0.m 0.0 0.0a H.0H _a.01 m.0 0.0m 0.N_ 0.0a H.m H.0 0.m 0.0 0.0 Aweiaom0c0a600>coo

0.ma a.m 0.0 0.H H.0H H.m 0.0H1 0.0a H.01 0.0 0.0 0.0 0.0 0.0 2000 Locum

H.00 0.00 H.0m 0.00 0.0m 0.0m .H.mm 0.0m _H.ma 0.~H .0.0 0.x. 0.H ~.H Loco mafia

 

 

om>aouoo muabomo>xono>m Edaoaou uo>o cam oao< cabooq UoUGoBXm .aa.m canoe

92

 

 

 

 

 

 

 

 

.Umu Uaom UauUMa maOE\mmm mwaOE mm Umuuoawn mum momma wo mpamaw at .wEau uumucoo m ov.w1ah.v .Uoomm «
m.m m.m o.m m.m m.m m.v o.m m.m o.m m.m v.v N.m m.m Hmnuo
h.ma w.ma m.ma m.ma m.ma m.om m.va m.ma N.ma o.ma o.ma N.ma N.Na aoumod
a.mm b.0N o.mm m.ma m.mm m.aa o.vm m.ba m.vm b.mm m.bm m.va m.om mvhnmUaMuwom
v.ma m.mm a.ma m.om a.ma m.ma m.ma m.am m.om m.am a.ma b.am v.ma oCOaUmCMUcmm1m.m
m.oa m.aa m.oa m.ma m.oa a.¢a m.m a.ma m.oa v.0a o.aa m.ma m.oa baud Uaocmaoum
v.mm m.vm m.am m.mm o.mm m.mm m.om m.vm o.mm v.am o.mN m.bm m.om baud Uaa>uud
va >ua>auomamm
1 vN.N 1 aa.a 1 no.a 1 aa.a 1 mm.a 1 mb.a 1 .WOU
1 am.a 1 am.o 1 mm.o 1 mm.o 1 mm.o 1 mm.o 1 :00
ab.o mv.o mv.o wm.o mw.o am.o mv.o aw.o bh.o «v.0 mm.o hm.o vm.o Hmnuo
mm.a mm.a mo.m mv.a aa.a va.m mo.m ma.m mo.N aa.a mm.m aa.a hm.a aOuwod
ab.m Nv.m aa.a hm.a ab.m mm.a mm.v mm.a ma.m om.m mm.m om.a am.v mfixzmpamumom
mv.m mp.m am.m ma.m mv.m bo.m mb.m av.m mm.m mb.m on.m mm.m mm.m mCOaUmCMucmm1m.N
mm.a mm.a mm.a mm.a mm.a bv.a Nv.a mm.a am.a mm.a mm.a mv.a mm.a Uao< Daocmaoum
mm.m am.m am.m ab.m mm.m oo.m cm.m mm.m am.m Nb.m aa.m mm.m am.m vaoé Uaa>904
va Uamaw
m.va m.ma m.ma a.va m.va v.ma m.wa m.ma a.va m.va w.ma m.va m.va vaahwdvc0amum>cou
b.b o.m v.a a.o o.m1 m.m1 m.>1 a.01 v.0 m.m1 a.v1 N.ma v.m vanmomVCOamHm>cou
m.m m.oa a.va o.va v.ma m.ma o.vm o.va b.ma N.ba m.ma m.01 m.oa Aowv Houuw
m.aoa N.¢m b.mm b.mm N.mm m.mb o.mb m.ab o.ab ¢.vm m.mm b.vm m.¢m Auzv wEaB

 

 

nowadaucouv .umhaouou wuauomo>xoupxm Edaoaou uo>o com baud oauomq ooccmbxm .aa.m manna

93

3.4.5. Mass Balance Accuracy

To determine carbon loss due to coking of the catalyst,
reactions were performed over two batches of blank Si/Al
support and two batches of Nafixh on Si/Al catalyst. These
used catalysts were then calcined at 550°C in air for 24
hours and the weight loss of the samples was measured.
Weight loss accounted for 1.7 to 4.4% of the carbon fed over
Si/Al support, while accounting for 6.7 to 7.6% over Nafixh
on Si/Al catalyst. For reactions over supported trisodium
phosphate, the amount of carbon deposited on the catalyst

accounts for most of the unrecovered carbon.

3.5. Summary

Lactic acid is converted to acrylic acid, propanoic
acid, 2,3-pentanedione and acetaldehyde as well as other
minor products, over sodium phosphate salts supported on
silica alumina. Acrylic acid formation is favored at a high
temperature of 350°C and short residence time of 0.4
seconds, while 2,3-pentanedione is favored at a low
temperature of 280°C and long residence times of 4.0
seconds. The more basic trisodium phosphate catalyst
exhibited more activity and selectivity to both acrylic acid
and 2,3-pentanedione than the more acidic sodium

orthophosphate salts, monosodium and disodium phosphate.

94

All primary reaction products result from first order
decomposition of lactic acid except for the second order
condensation of lactic acid to form 2,3—pentanedione.
Propanoic acid is formed by both a first order reduction of
lactic acid as well as a hydrogenation of acrylic acid.

The condensation product, 2,3—pentanedione, is proposed
to result from a Claisen condensation between two cyclic
lactyl phosphate molecules, with subsequent decarboxylation
and rearrangement. Acrylic acid dehydration is proposed to
occur via a cycloelimination in which phosphate is cleaved
from lactic acid—2-phosphate, leaving acrylic acid.

Kinetic analysis suggests that 2,3-pentanedione
selectivity may be enhanced by increasing contact time,
greater lactic acid concentration and lower temperature.

Sodium phosphate salts possess significant selectivity
to several valuable products including acrylic acid and 2,3—
pentanedione. Manipulation of the reaction conditions may be
used to enhance the selectivity to a particular product.
Catalyst activity of sodium phosphates is also high, with
conversions greater than 80% at 350°C under certain

conditions.

Chapter 4. Lactic Acid Conversion over Supported Sodium Salt
Catalysts-A Catalyst Screening Study

4.1. Introduction

The proposed mechanism for 2,3—pentanedione formation
includes the combination of the phosphate catalyst with the
acyl group of lactic acid to form an acyl phosphate. The
proposed mechanism for acrylic acid formation depends on the
phosphate forming an ester linkage with the hydroxyl group
on the C2 of lactic acid to yield a cyclic structure. The
formation of cyclic intermediates suggested a catalyst
survey to study the catalytic performance of various
catalyst salts with different central atoms and oxygen
ligands similar to phosphates in structure. Sodium catalyst
salts were systematically chosen by the central atoms which
form an adjacent series on the periodic table. The
performance of these sodium salts are directly comparable to
sodium phosphate salt catalysts. The sodium salt catalysts
surveyed are listed in Table 4.1. The relationship of the

central atoms on the periodic table is shown in Table 4.2.

96

Table 4.1. Catalyst Salts Surveyed

 

Sodium Tetraborate Nafix07
Sodium Carbonate NagXL
Sodium Nitrate NaNO3
Sodium Metasilicate NaQSiO3
Trisodium Phosphate Nafixh
Sodium Sulfate 1%5804
Sodium Chlorate NaClO3
Dibasic Sodium Arsenate Heptahydrate IkgHAsO¢7HJD
Sodium Bromate NaBrO3
Sodium Molybdate rkgMoO3

 

Table 4.2. Periodic Relationship of Catalyst Salt

Central Atoms

VIB
2
Period 3
4

5 42

MO

 

 

 

 

 

 

Group

I I IA IVA VA VIA VI IA

5 6 7

B C! N
14 15 16 17
_ Si P s 01
33 35
A8 Br

 

 

 

 

 

 

97

4.2. Methods

4.2.1. materials

All catalyst salts except trisodium phosphate (EM
Science) were obtained in solid lumps or crystals from
Aldrich Chemical Company. The origin of all other materials
and catalyst preparation methods are discussed in Sections

3.2.1. and 3.2.2.

4.2.2. Vapor-phase Reactor

All reactions were performed in a downflow differential
packed bed vapor—phase reactor lined with quartz.
Descriptions of the reactor and operating conditions are
discussed in Section 3.2.3. Product analysis techniques are

discussed in Section 2.2.6.

4.3. Results

Conversion of lactic acid was performed over each of
the catalysts supported on Si/Al support. Reactions were
performed at 280—350°C and a residence time of 1.5—3.5
seconds.

Silicate is a moderately active catalyst with
conversions (BOF) from 26.3% at 280°C to 81.8% at 350°C.
Silicate produces the highest selectivity to acrylic acid

which increases with temperature to 35.9% at 350°C. The

98
propanoic acid selectivity is one of the lowest of the
catalysts surveyed increasing from 3.7% to a maximum of 9.9%
at 350°C. The selectivity to 2,3—pentanedione for silicate
is very similar to that for trisodium phosphate. This
selectivity decreases from 52.9% at 280°C to 22.5% at 350°C.
The selectivity to acetaldehyde remains low and fairly
constant at 12—15% for all temperatures studied while the
selectivity to acetol is the highest of all the catalysts at
10-l4%. Selectivity to other products is moderate, hovering
at 5—7%. Overall, silicate shows high conversion and high
selectivity to acrylic acid as well as low selectivity to
propanoic acid and acetaldehyde. Silicate also exhibits the
highest selectivity to acetol of 10-14%.

Trisodium phosphate is an active catalyst with
conversions (BOF) from 14.2—85.4% as temperature increases
from 280 to 350°C. The selectivity to acrylic acid is high
with a maximum of 27.0% at 320°C. The selectivity of
phosphate to propanoic acid is low, increasing from 4.9% to
9.0% at 350°C. Phosphate also has a high selectivity to 2,3—
pentanedione which is surpassed only by arsenate and
silicate. This selectivity has a maximum of 51.1% at 280°C.
Phosphate has a moderate acetaldehyde selectivity which
increases from 17.2% at 280°C to a maximum of 42.0% at
350°C. A moderate selectivity to other products increases

from 3.3 to 8.8%. Overall, trisodium phosphate exhibits good

99
selectivity to acrylic acid and 2,3—pentanedione and also a
low selectivity to the undesirable product propanoic acid.

The activity of arsenate is second only to bromate and
molybdate. This is especially evident at low temperatures
where the conversion starts at 42.2% at 280°C and increases
to 95.9% at 350°C. Arsenate exhibits a moderate selectivity
to both acrylic acid and propanoic acid which hover at 15—
24% and 5-13%, respectively. The selectivity to 2,3-
pentanedione is outstanding with a maximum of 63.8% at
280°C. This selectivity decreases precipitously with
increasing temperature to 22.5% at 350°C. The acetaldehyde
selectivity increases from 13—30% with a selectivity to
other products of 2—7%. Overall, arsenate proves a very
active catalyst with the highest selectivity to 2,3—
pentanedione of all catalysts surveyed.

Carbonate is a moderately active catalyst. Conversion
based on lactic acid recovery ranges from 24.3% at 280°C to
88.1% at 350°C. Carbonate has a high selectivity to acrylic
acid ranging from 15.1% at 280°C to 25.4% at 320°C.
Propanoic acid selectivity for carbonate is moderate,
increasing from 4.0% at 280°C to 14.4% at 350°C. Carbonate
shows a moderate selectivity to 2,3—pentanedione of 47.0% at
280°C. This selectivity decreases with increasing
temperature to 18.5% at 350°C. The moderate selectivity to
acetaldehyde increases from 20.6% at 280°C to 32.3% at

350°C. Carbonate selectivity to other products is low at 4—

100
8%. Overall, carbonate shows high conversion and moderate
selectivity to desirable products.

Borate is a moderately active catalyst with conversion
(BOF) increasing from 17.4% at 280°C to 91.5% at 350°C.
Acrylic acid selectivities are low, ranging from 9—22%,
going through a maximum near 320°C. Borate and sulfate share
the lowest selectivities to acrylic acid of all the
-catalysts surveyed except molybdate. Propanoic acid
selectivities are moderate at 5—14% for borate. Borate has
moderate selectivity to 2,3—pentanedione. The maximum 2,3—
pentanedione selectivity of 43.2% for borate occurs near
320°C. All catalysts except borate show a maximum 2,3—
pentanedione selectivity near 280°C. Borate has a high
acetaldehyde selectivity which decreases with increasing
temperature from 59.6% at 280°C to 22.8% at 320°C. The
selectivity to other products is 4—10%. Overall, borate
shows poor conversion and poor to moderate selectivity to
desirable products.

The catalyst with the lowest activity is sulfate. The
conversion (BOF) increases from 4 to 52% as the temperature
rises from 280 to 350°C. Sulfate has low selectivity to
acrylic acid and 2,3—pentanedione. Acrylic and propanoic
acid selectivities hover at 10—14% over the temperature
range studied. The maximum 2,3-pentanedione selectivity is
24.5% at 280°C. The acetaldehyde selectivity of sulfate is

one of the highest of all the catalysts surveyed with a

101
maximum of 53.5% at 350°C. The selectivity to other products
is high at 7-12%. Overall, sulfate is a poor catalyst with
low conversion and low selectivity to desirable products and
high selectivity to undesirable acetaldehyde and propanoic
acid.

Bromate is an active catalyst with conversions from 45—
96%. Bromate has a high selectivity to acrylic acid which is
second only to silicate and increases with temperature from
19 to 33% from 280 to 350°C. The propanoic acid selectivity
is also high at 17—20% and the low 2,3—pentanedione
selectivity stays at 18—26%. Bromate has a moderate
acetaldehyde selectivity which hovers at 19—26%. Bromate
shows a low selectivity to other products at 4-7%. Overall,
bromate is an active catalyst with high selectivity to
acrylic acid and propanoic acid.

Chlorate produces low conversion increasing from 14% at
280°C to 90% at 350°C. The selectivity to acrylic acid is
also moderate, increasing to a maximum at 350°C of 23.6%.
The high propanoic acid selectivity follows a similar trend
with a maximum at 350°C of 19.2%. The 2,3-pentanedione
selectivity is moderate with the maximum of 29.4% occurring
at 280°C. The moderate acetaldehyde selectivity hovers at
22-26% while the high selectivity to other products is 5—
11%. Overall, chlorate produces no outstanding conversions
to any of the major products, but all products are formed in

rather even proportions.

102

Nitrate is an active catalyst with conversion (BOF)
which increases from 25% at 280°C to 92% at 350°C. Nitrate
has a high selectivity to acrylic acid which hovers at 17-
23%. The selectivity to propanoic acid is moderate,
increasing from 6.2% at 280°C to 15.1% at 350°C. Nitrate has
a high 2,3-pentanedione selectivity which decreases from
50.1% at 280°C to 24.3% at 350°C. Nitrate shows a moderate
selectivity to acetaldehyde which ranges from 16—29% while
the selectivity to other products ranges from 4—9%. Overall,
nitrates exhibit high conversions and good selectivity to
acrylic acid and 2,3—pentanedione.

Molybdate, the only salt included in this study with a
transition metal central atom, behaves differently than all
other catalysts surveyed. Molybdate produced mostly
propanoic acid in selectivities from 70—84%. Propanoic acid
selectivity decreases with increasing temperature. Molybdate
is the most active catalyst with very high conversion (BOF)
of 87.2% at the lowest temperature of 280°C. This conversion
increases slightly to 92.6% at 350°C. The acrylic acid and
2,3—pentanedione selectivities were the lowest of all the
catalysts at less than 2.7% and 1.7%, respectively. The
acetaldehyde selectivity was also low at 10—20%. Selectivity
to other products was less than 5%. Unlike the other
catalysts, molybdate affected much cracking to produce much
(IL and some CO accounting for 14-17% of the carbon balance.

Overall, molybdate is the most active catalyst surveyed but

103
produces mostly propanoic acid and C02. Product distribution

over catalyst salts surveyed is presented in Table 4.3.

4.4. Discussion

The various catalysts do not always attain the maximum
selectivity to a product at the same temperature. An
excellent example of this occurrence is the fact that borate
achieves a maximum 2,3—pentanedione selectivity at 320°C,
while all other catalysts have a maximum near 280°C. For
this reason, it was instructive to compare the maximum
selectivity attained by a catalyst at any temperature
instead of considering the selectivities at only one
temperature. In the following discussion, this selectivity

will be referred to as the "maximum selectivity".

4.4.1. Catalyst Activity

Sulfate has the lowest activity of all catalysts
surveyed while molybdate has the highest. Arsenate, bromate,
nitrate and phosphate are also active catalysts. Molybdate
shows little effect of temperature on conversion with
conversions of 86.5-92.6% for temperatures from 280-350°C.
Generally, the greater the activity of the catalyst, the

higher the cracking to produce CO and C02.

104

.Uwu Uaom cauoca mmaoE\mmm mmaoE mm Uwuuoawu mum mommm mo mcamaw a

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0.0 0.0 0.0 0.0 0.0 0.0 0.0 «A 0.0 0.0 0.0 0.0 0.0 0.0 T0 0.0 0050
m.0 0.0a 0.3 0.0 20 0.0 0.0 TH 0.0 0.3 0.: 0.0 0.: 0.3 0.3 0.0a 3000a
~13 0.0m Ham 0.0m 0.0m m.mm 0.: 0.3 0.0.0 ~00. 0.0m 0%: TE 0.: 0.3 0.3 00330300004
0.0a 0.0m 0.00 0.00 01mm 0.: 0.0.0 0.00 0.3 H00 0.00 H40 0.mm 0.00 0.00 0&0. 06600200501010.
«.3 0.0 0.0 0.0 BS 0.: 0.0 0.0 0.0 mg. 0.0 0.0 0.0 0.0 0.0 0.0 304 06500000
Tom 0.0m 0:3 H.0H mam m.mm 0:00. 0.: 0.0m 0.0m 0.0m 0.: 0.00 0.0m 0:2 0.0a 39a. 02304
100 >0a>aoo0a0m

05:02:00.0 3.10 00.0200 2.: 00.: 00.030.33.00 00.0 8.300.330 00.0 .60
8.0 00.0 00:0. 004 00.0 00.m mm.m 00; 3.2.00.0 3.0 00; 00:0. 004 00.0 00.0 .8
00.0 Sim 0004 aa.a 00.0 aa.a H00 00.0 00.0 00.m 3:0. 00.0 22m aim 2.4 mm; 0050
00.0 00.0 00.0 00:0. 004 00.0 00.0 $0 mm.m 2.0 00.0 004 00.0 3.0 $0 aa.a 3009a
00.85.2073 0m.0 00.33.2073 ~00 07300.0 3.0 00.0 00.0 2.0 00.0 00:0. 093033000
00.: 00.2 3.: 00.3 00.33.0Hm0.0m00.0m 00.0 3.: 3.: 00.0 30:10:04 3.3 00.: 0c0ac0ccoc0a1m§
3.0 0N0 Sim mm; 00.0 00.0 8.0 004 00.0 00.0 :14 00.0 00.0 02m 00; $0 304 06800018
00.3 .2113 No.3 8.0 02300.: 0.0.: 3.0 Nata 00.3002. 3.0 3.300.: 00.0 «0.0 30a 03304
Awe pamaw

0.00 0.8 040 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.3 0.0m 0.m0 N00 0.0m 0.0m 3:340 03000280
H00 0.00 0.3 0.: 0.00 0:00 0.00 23 0.00 0.2. 0.00 mi: 0.5 0.8 0.00 0.332000106300600
0.01 0.31 0.0 0.0 0.00-0.0m1m.31 0.0 A11.070.070.07 0.2 2310401231 0.7 600180000
000 0.0.0 000 00m 000 0.0.0 000 00.0. 000 0mm 000 00m 00m 0mm 000 00m 6: 0030000200
00.0100.m H0.N1H0.a 00.m1mm.a mm.m100.m Ame 0eae 00:00am0m

0.HH 0.NH 0.NH 0.ma mo omaacuco
QUMCOQHMU mumcwmh< mUMEQmOSQ quUHaflm um>HMuMU
mummamboo oaom Edacom uo>o COaDSQanomao nonconm .m.v magma

105

.Umw Dado oaDUMa moaoE\mmm mwaoE mm Umuuommu who mmmmm uo mpamaw

.1

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0.0 0.0 0.0 0.0a 0.0 0.0 0.0 0.0 0.: «.00 0.0 as 0.0 0.0 0.0 0.0 0050
0.0 0.0 0.00 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 TN 30004
0.0m 0.8 0.2 0.00 0.00 m.0m 0.00 H00 0.00 0.00 0.0.0 0.3.. 0.00 0.2. 0.00 0.00 0020300000
0.00 0.0m 0.0m 0.0m 0.: Tam H.0m 0.00. 0.0 0.0 0.00 0.0m 0.0m ~00 0.00 0.00 056000200100
m.0a 0.0a 0.2 0.2 0.: 0.00 0.00 N00 0.00 32 0.2 0.2 0.3 0:0 0.0 0.0 304 380080
0.0m Ham 0.0m 0.0a 0.00 T00 0.00 0.00 0:04 0.3. 0.: 0.2 0.0 0.8 0.: 0.00 33 020.80
A00 00a>aou0a0m

00.00030 00.0 004 2.3000 00.0 0.0.0 0m.m 00.0 00.0 00.0 00.200 004 00.0 .60
00.0 00 004 00.0 00.0 00.0 00.0 00.0 30 00.0 34 00.0 3.0 00:0 2.0 00; .8
00.0 00:0. 00.0 :4 00.0 004 004 31.0 00.0 004 00.0 00.0 00.0 004 00.0 00.0 0050
00.0 00.0 2:0. 0.0.0 00.0. 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 3.0 8004
00.2000 00.0 00.0. 00.0 3.23.0 3.0 00.00000 8.0 00.0 00.200.22.00 2.0 062033090
00.0 00.0 00:0 00.0 00.0 00.0 N04. 00:0 00.0 00.0 004 004 2.3000300 00.0 0006003000120
8.30.0.0 00.0 00;. 3.0 00.0 2.0 020 00.0 Sim 00.0 00.0 0N0 00.0 00.0 0.0.0 304 0300080
00.3000 00.0 004 00.003.200.00 00.0 00.0 00.0 00.0 00.0 00.0 00020.0 00.0 39a 020090
- 100 0H00>

0.00 0.00 0.00 ~13 0.00 0.00 0:00 0.00.0. 0.00 0.0m 0.0a 0.0 0.00 0.00 0.: 0.3310640000000200
0.00 0.00 0.0m 0.3 0.00 0.2. 0.00 N00 0.3 H30 0.00 0.0 0.0 0.2. 0.00 0.2 3:000: 303000280
0.001 0.01 0.0 0.7 0.0m-0.070.210.0793- TS 0.0 0.0 H.001 0.810.001 0.01 83 0920
00m 0mm 000 000 000 000 000 00m 000 0N0 000 00.0. 000 00.0 000 000. 6.0 0030000200.
0H.m100.~ 00.010~.m 00.010~.m 00.0100.H Am. 0500 0uc0cammm

0.0 0.0 0.0 «.0 mo 003300

OUMHOHAAU OUMEOHm OUMMHDM muwhom umxaamumu
newscaocoov mom>a0000 oacm Esapom 00>o cansnauomao oospoum .m.e manna

.Umu caum UaDUMa mwaoE\mmm mmaoE mo Umuuommu who momma uo mcamaw 1
was v.m vim ma m.m aim a6 m.m nonuo
m.o m.o v.0 m.o m.m o.aa m.ma ma. acuood.
m.ma m.ma v.1: ¢.oa m.mm \1.ma m.ma b.ma mc>£mpa0umo<
o.m ha «a o.a mam mém m.mm a.om oCOancoucwm1m.m
m.mm «.mb v.2. m.mm a.ma m.ma ma. mim
w.m min. ANN min. m.ma m.mm 5mm m.ma

 

baud oaocomoum

 

 

caoa caaxocc

 

.00 >00>a000fl0m

 

106

roam mm.mmmm.bm No.00 vm.mamo.mamb.m who .NOU
ow.m mo.w om.m on.m mm.m owlm ma.m ma.a .OU
ma.m mma ma.a mm.o mo.m mm.m aa.a aN.a anuo
mv.o ww.o ma.o oa.o mm.m mafim mm.m Mvm.m aoumod
mm.m mm.m mm.v mb.m mm.wamm.oa trim aa.a wp>£mcamD004
mm.o mm.o ww.o mm.o vmécawméa mo.ma hm.ma wCOHU026ucwm1m.m

mm.mm mv.ommm.mmmm.om mm.m om.m Naiv mm.a
mma No.a mm.o mm.o aa.a m>.aavm.aamm.m

baud oaocmmoum
cao< caa>uo<
Awe camaw

m.mm >.mw m.mm m.m> mink. m.m> v.vm N.bm vaxncdv c0amuo>coo
m.mm m.mm m.mm Ngm mam o.mm cam mém 3:059 ceamum>coo

 

 

 

 

m.m1 a.b1 0&1 m.m1 m.va1m.ma1 o.m m.ma 50$ Houum
0mm 0mm oom 0mm omm omm oom 0mm AOL mHSDMHmQEmwe
m>.m1om.m mv.m1mm.~ Amy msae wocmcamwm

0.0 0.0 so omsacoco
mumcntaaoz wumuuaz umhamufimu

 

 

 

AcmSCaucouv mumxacuou Daom Edacom Ho>o COaDSQauumaQ nonpoum .m.v magma

107

4.4.2. Acrylic Acid

Silicate and bromate exhibit high selectivities to
acrylic acid. These catalysts have maximum selectivities at
350°C of 35.9 and 33.3%, respectively. Molybdate shows the
lowest acrylic acid selectivity of 2.5—2.7%. Sulfate is
second lowest with selectivities of 11.5-14.3%.

Acrylic acid yields generally increase with
temperature, but for some catalysts, the yield goes through
a maximum near 320°C, indicating secondary reaction of

acrylic acid.

4.4.3. Propanoic Acid

Propanoic acid production is highest over molybdate
with selectivities of 70-84%. Bromate comes in second with
propanoic acid selectivities of 17—20%. Chlorate also
produces much propanoic acid at 13—19% selectivity. Silicate
and phosphate produce the least amounts of propanoic acid
with selectivities of 3—10% and 5—9%, respectively.

Yields to propanoic acid increase with temperature for
all catalysts surveyed, indicating that it is a terminal

product.

4.4.4. 2,3-Pentanedione
2,3—Pentanedione selectivity is highest for arsenate at
63.8%. Silicate, phosphate and nitrate produce almost

identical amounts of 2,3—pentanedione with selectivities of

108
52.9, 51.1 and 50.1%, respectively, at 280°C. Carbonate has
a selectivity of 47.0%. All of these catalysts have a
maximum selectivity at 280°C which falls off with increasing
temperature to less than 25% at 350°C. Borate is the only
catalyst which has a maximum selectivity at a temperature
higher than 280°C. Borate goes through a maximum of 43.2% at

320°C. Molybdate exhibits the lowest selectivity of 1-2%.

4.4.5. Acetaldehyde

Production of acetaldehyde is highest over borate and
sulfate with selectivities of 23—60% and 35—54%,
respectively. Silicate and molybdate have the lowest
selectivities at 12-15% and 10—20%, respectively.

Acetaldehyde yield increases with temperature in all
cases except for bromate, where it goes through a maximum
near 320°C. Acetaldehyde is believed to be a terminal

product.

4.4.6. Acetol

Selectivity to acetol is greatest over silicate at 10-
14%. Nitrate, phosphate and carbonate produce maximum
selectivities of 13.2, 12.4 and 11.6%, respectively.
Molybdate produces the lowest acetol selectivity of 0.3-

0.9%.

109
4.4.7. Other Products
Selectivities to other products vary from 7—12% for
sulfate to 2—5% for molybdate. Other products observed are

listed in Section 3.3.

4.4.8. Effect of Catalyst Salt Acidity on Selectivity
and Conversion

The catalyst salts were analyzed on the basis of their
acidity in an aqueous solution. The pH of a 0.1 M aqueous
solution of each of the catalyst salts was determined either
from measured values reported in literature or by
calculation from published acid dissociation constants, Ka
(84-87). No correlation was observed between the acidity of
the salts and their activity as catalysts or their
selectivity to acrylic acid.

The selectivity to 2,3-pentanedione follows a strong
trend with respect to catalyst acidity. With the exception
of nitrate, the 2,3—pentanedione selectivity increases as
the catalyst becomes more basic. Basic catalysts (pH > 9)
exhibit the highest 2,3—pentanedione selectivities of 43—
64%. As the catalyst becomes more acidic, the maximum 2,3—
pentanedione selectivity decreases to 25—29% for chlorate,
bromate and sulfate. The basic arsenate (pH = 12.0) produces
the greatest 2,3—pentanedione selectivity at 63.8%. The
trend in maximum 2,3—pentanedione selectivities may be seen

in Figure 4.1.

110

81201113

aiequoqd

 

 

   
 

91941718
3112qu
312mm I

 

aieiomo

A l
U

o o o o o
(D to <1- 0) N
(

%) 1111111100193 w1w1xew

 

1O

12

11

10

Catalyst Salt pH

Effect of Catalyst Acidity on Maximum Selectivity to 2,3-Pentanedione

Figure 4.1.

111

Nitrate does not follow the trends observed by the
other catalysts. One reason that has been suggested is that
sodium nitrate may decompose or condense to form another
species after heating to reaction conditions. This may
change the acidity of the actual surface species and affect
the selectivity to 2,3—pentanedione. Efforts are currently
being made to test this hypothesis using transmission FTIR
studies similar to those discussed in Chapter 6.

Propanoic acid selectivity increases with catalyst
acidity. The neutral chlorate and bromate catalysts produce
the most propanoic acid with maximum selectivities of 20.2
and 19.2%, respectively, while the most basic phosphate and
silicate produce the least at 9.0 and 9.9%, respectively.
Again, nitrate does not fit this trend, presumably for the
reasons stated above. The effect of catalyst acidity on
propanoic acid selectivity is presented in Figure 4.2.

Moderately basic catalysts exhibit the greatest
selectivity to acetaldehyde. Sulfate and borate produce
selectivities of 35-54 and 23~60%, respectively. The most
basic silicate catalyst produces the lowest acetaldehyde
selectivities of 12-15%. Acetaldehyde selectivity also falls
off for the relatively acidic catalysts chlorate, bromate
and nitrate. The effect of catalyst acidity on acetaldehyde

selectivity may be seen in Figure 4.3.

112

61901113 .

aieuaSJV aiequoqd ‘-

aieuoqieo

319108

 

eieilns
aiewmg

912101143 SEDAN ‘ "

 

0 LO 0 LO 0
N

F F

(%) anoelas

12

11

10

Catalyst pH

Effect of Catalyst Acidity on Maximum Selectivity to Propanoic Acid

Figure 4.2.

113

mpxnopamuood OD >ua>auomaom ESEaxmz co huapao<.um>amumu mo Doowwm .m.v opsmam

 

  

  
   

 

 

Id 52300
Nw Z or m m h m
1 u 1 u 1 u 1 u 1 u 1 u 1 o
8 O 1
V m N :3
01. w m .
a mw m. Q0
mm W a no 11.0N mm
H 1.
0 a . 0.
m N no om m.
m ........ . A
m- m gr CV w
.0 0 a a . .1
0 W m .0 om
s 8
Au 8 1+
u. e 1
m. LaAUQ

91911113

114

The selectivity to acetol is greatest over the most
basic catalyst silicate. However, acetol selectivity
exhibits no clear trend with respect to catalyst acidity.

Selectivity to other products is greatest over
relatively acidic chlorate and sulfate at 10.6 and 11.5%,
respectively, but no trend with catalyst acidity is
observed. The effect of catalyst salt acidity on product

selectivities at 300°C is illustrated in Figure 4.4.

4.4.9. Periodic Considerations

Catalyst salts with central atoms from Group IVa and Va
produce the highest selectivities to 2,3—pentanedione.

The activity of the catalyst is related to the highest
occupied electronic shell of the central atom. This trend
becomes obscured as the reaction approaches completion at
high temperature, but is very clear at the low conversions
of 280°C, as shown in Table 4.4. Molybdate exhibits the
greatest conversion of 87.2% with a partially filled O shell
(551) for ground state molybdenum. Arsenate and bromate
produce conversions of 42.2 and 45.2%, respectively, with
partially filled N shells (4p? and 4p5) for ground state
arsenic and bromine. All other central atoms with highest
occupied atomic orbitals of L and M produce less than 27%
conversion at 280°C. No correlation was observed between

conversion and the highest occupied shell L and M.

115

oooom um. mmaugfluomaom uosposm :0 032025 63300 do bootm

alequow
31211!“
312.10qu
awwoxg
aieilns
318108
aieuoqieg
aleuaSJV
elequoud

o=>8< D
0:050:9c0m1md I
20:32.“. E
035025004. 2%

.005 g

 

.q.v ossmam

.10

ml...

3

m
o
S
om
on m
00 m
00 w”
00 .m).
00 m/w
00
00
000

.9502 Baum

116

We believe both acrylic acid and 2,3—pentanedione to
arise by mechanisms in which the formation of a five
membered cyclic transition state or intermediate is
required. Formation of charged transition states are
required by these mechanisms. Oxygen ligands of the catalyst
salt must also be readily eliminated in both of these
schemes, as discussed in Section 3.4.1. Since the reaction
mechanisms are similar in these respects, the same factors
may affect the rates of both mechanisms.

Three factors may be responsible for this phenomenon of
increased conversion with the increase in highest occupied
atomic orbital.

l) Steric factors may be preventing the smaller "hard"
central atoms from allowing bond angle deformation to easily
form rings while larger, more "soft" central atoms with
loosely held bonding electrons allow more flexibility for
bond angle deformation.

2) The simple geometric steric consideration of the
size of the central atom may allow larger atoms to form
larger rings with larger bonding angles on the opposite side
of the ring from the central atom. Smaller atoms must form
smaller rings with tighter bond angles thus making the
transition state less favorable. Figure 4.5. illustrates how
the size of the central atom may affect the bond angles of a

transition state.

117

Table 4.4. Effect of Electronic State of Catalyst Salt
Central Atom on Conversion at 280°C

 

Catalyst Highest Occupied Shell of Central Conversion BOF

 

Salt Atom Electronic Ground State at 280°C (%)
Borate L (ZpW 17.4
Carbonate L (2pfl 24.3
Nitrate L (ZpW 24.9
Silicate M (3p2) 26.3
Phosphate M (3p3) 14.2
Sulfate M (3pfi 3.7
Chlorate M (3p5) 14.2
Arsenate N (4pH 42.2
Bromate N (4p5 45.2
Molybdate 0 (5s‘) 87.2

 

 

o<>B

Figure 4.5. Effect of Size of Catalyst Salt Central Atom on
Bond Angles in Cyclic Transition States

118
3) Atoms with large, loosely held electron clouds may
delocalize electronic charge easier, thus facilitating

formation of charged transition states.

4.4.10. Carbon Balance

Carbon balance errors generally increase with
increasing temperature as shown in Table 4.5. Mean carbon
balance errors vary from 0.2% at 280°C to nearly 22% at
350°C. Borate and arsenate exhibit the highest observed
carbon balance errors of —37.1 and —37.5%, respectively.

Gravimetric calcination studies were performed on
selected lots of used catalyst and discussed in Section
3.4.5. These studies indicate that about 7% of carbon fed
over a series of runs at temperatures of 280—350°C is
deposited on the catalyst in the form of coke. We know from
mass balances that much coking occurs at high temperatures
and almost none at low temperatures, so this accounts for
most of the carbon balance error observed.

The standard deviation of the carbon balance error is
almost identical to the reproducibility of the lactic acid
peak in product analysis. As mentioned in Section 2.2.6.1.,
the standard deviation of the lactic acid peak area observed
in CC analysis is 14.1% of the mean. This compares very well
with the standard deviation of the carbon balance errors in

the catalyst survey, which is 10—15%. This indicates that

119
most of the deviation observed in the carbon balance is

caused by deviation in the analysis method.

Table 4.5. Analysis of Carbon Balance Error

 

 

 

Temperature (°C) 280 300 320 350
Mean Error (%) 0.2 -6.6 -15.1 —21.8
Standard Deviation of Error (%) 9.6 12.0 15.0 9.6

 

4.5. Summary

Group V catalyst salts (nitrate, phosphate and
arsenate) show the greatest selectivities to 2,3-
pentanedione, with selectivities up to 64%.

Catalyst salt acidity affects product selectivity.
Basic salts (pH > 9) exhibit higher selectivity to desirable
2,3-pentanedione. Relatively acidic catalysts promote
formation of propanoic acid. Moderately basic catalysts (7.4
5 pH < 10) promote production of acetaldehyde.

Conversion increases with increasing level of occupied
orbitals in the catalyst salt central atom. Salts with
central atoms having N and O shell electrons produce much

higher conversions than those with only L and M shells.

Chapter 5. Characterization of Sodium Phosphate Salt
Catalysts using MAS-”P-NMR

5.1. Introduction

Sodium phosphate salts are known to undergo dehydration
and condensation upon heating (58). The sodium phosphate
salts which serve as catalysts in our reactors are heated to
temperatures up to 400°C before and during reactions, which
leads to a need to identify which chemical species are
present on the catalyst surface. The purpose of this study
is to determine the form of the phosphate catalyst under
reaction conditions employing Magic Angle Spinning-“P-
Nuclear Magnetic Resonance (MAS—”P—NMR). The conclusions
which are made have also been compared to results of a study
of surface species using transmission FTIR.

Chemical shifts reported are relative to 85% phosphoric
acid, Hfixh, where o = -6 and the chemical shift, 6, is
reported according to IUPAC convention with 6 > 0

representing a downfield shift and deshielding.
5.2. Methods
5.2.1. Materials
The origin of materials used is discussed in Section

3.2.1.

120

121
5.2.2. Sample Preparation

Neat salts were used as received without purification.
Salts which appeared clumpy were ground with a mortar and
pestle to achieve uniform consistency.

Before impregnation of catalyst supports with salt, the
support was crushed with a mortar and pestle and sieved to
30x60 USA wire mesh. The catalyst support was crushed finer
for use in NMR spectroscopic studies than for use as a
catalyst support in the vapor phase reactor. The finer
consistency of the NMR samples was necessary for better
balance and higher spinning speeds for Magic Angle Spinning
(MAS) spectroscopy.

Crushed supports were impregnated with phosphate salts
using the same method described in Section 3.2.2. Catalyst
loading was 1.0 mmol/g support for all samples.

Bovine dental enamel was prepared by boiling for two
hours to remove organic material, then crushing, grinding
and sieving to 60x100 mesh.

Samples were heated in a Thermolyne 1300 furnace
controlled by an Omega PI controller. Neat or supported
salts were prepared as described above, placed in glass
beakers or ceramic crucibles, then placed in the furnace
which had been preheated to a designated temperature.
Samples were heated for 4.3—5.6 hours at constant
temperature, then removed from the furnace and allowed to

cool in air. The furnace door was opened for as short a time

122
as possible upon addition or removal of samples. At
temperatures less than 500°C, the furnace temperature
dropped less than 15°C upon addition of the sample and
returned to within 5° of the setpoint within 5 minutes after
the start of the calcination run. For calcination of
biomineral hydroxyapatite at 800°C, the temperature dropped
50°C but returned to within 20° of the setpoint within 5
minutes. After heating, samples were stored in glass vials

with plastic screw—on lids tightly shut until ready for use.

5.2.3. NMR Spectrometer

5.2.3.1. Spectrometer Hardware and Data Acquisition

Magic Angle Spinning Phosphorus-31 Nuclear Magnetic
Resonance (MAS—“P-NMR) spectra were obtained at room
temperature by weighted Fourier transformation of free
induction decays (FIDs) observed using quadrature detection
at 161.903 MHz. Hardware consisted of a Varian VXR—4OOS
spectrometer with an Oxford cryomagnet generating a magnetic
field of 9.395 T. Appendix B details calculation of the
magnetic field strength. A Varian 400VT CP/MAS sample probe
was used with a 7 mm diameter rotor and a top speed of about
8 kHz. The computer operating system was VNMR Version 4.1
run on a Sun workstation network.

Normal data acquisition was accomplished using a single

pulse sequence shown in graphical form in Figure 5.1.

123
Typical operating parameters for normal data acquisition are

presented in Table 5.1.

 

 

d1 observe

pw

 

 

Figure 5.1. Normal Data Acquisition Sequence

Table 5.1. Typical Parameters for Normal Data ACquisition

 

Observed Nucleus 31P
Transmitter Frequency (MHz) 161.903
Spectral Width (kHz) 100
Number of Data Points 16,000
Acquisition Time (s) 0.080
Number of Scans 64
First Delay, d1 (s) 10.0
Pulse Width, pw (us) 4.0

 

5.2.3.2. Spectrometer Setup

The account was started on the workstation and a setup
procedure was run as outlined in the operating manuals to
set the acquisition parameters for the observe nucleus, 31P.
The transmitter was turned on in preparation for tuning the

probe.

124

The spectrometer probe required tuning at the start of
every daily session on the spectrometer. The tuning wand for
31P was inserted into the sample probe and the probe
assembly was installed in the magnet. The observe port of
the probe was connected to the tuning port of the
spectrometer junction box located on the leg of the magnet.
The "tune" mode was selected on the selector knob of the
junction box. The probe was tuned to the minimum signal
possible using both coarse and fine tuning knobs on the
bottom of the probe. The tune range was changed several
times before a minimum was found at the most sensitive range
setting. The probe was disconnected from the junction box,
the probe was removed from the magnet and the tuning wand
was removed from the probe in preparation for sample

loading.

5.2.3.3. Sample Loading

Visible residue was loosened from the rotor using a
wooden toothpick or similar sharp, soft object to prevent
scratching and grooving the rotor. The rotor was tapped
lightly with the mouth down on a solid surface or blown into
in order to dislodge residual particles. Any residual powder
was wiped out using a lintless wipe rolled into a
cylindrical shape. This procedure was repeated until the

rotor was clean by visual inspection.

125

The rotor was filled with loose sample and tapped
lightly on a solid surface to pack the sample evenly. This
was repeated until the rotor was full up to the level at
which the end cap contacts the sample. The sample was
checked for evenness on top before the end cap was firmly
pressed into the rotor. Handling the rotor by the end cap,
any residual material was wiped from the rotor, being
careful to look for grease from fingers and powder from
sample loading since any foreign material on the outside of
the rotor may decrease spinning frequency or affect sample
balance.

The access door to the sample compartment of the probe
was opened and the rotor was inserted into the spinner
housing until it was firmly seated. The access door was
closed and the probe was installed in the magnet. The
observe port of the probe was connected with the observe
port of the junction box. The decoupler transmitter cable
from the spectrometer was connected to the decouple port of

the probe if performing proton decoupling experiments.

5.2.3.4. Sample Spinning

The two quick-connect fittings of the spinner air
control unit, located on the leg of the magnet, were
connected with the respective connectors on the probe. One
is labeled "bearing" for the air bearing of the spinner and

the other is labeled "drive" denoting the air drive or speed

126

control. The fittings were tightened firmly, taking care to
not break the glass connectors. The fiber—optic tachometer
connector was plugged into the respective port in the probe.
The air bearing pressure on the air control unit was
increased before increasing the drive pressure. The sample
started spinning as evidenced by the accompanying whine of
the rotor and the readout of the tachometer. The complete
procedure for bringing the spinner up to speed is given in
the instruction manual. Once the spinner had reached the
appropriate frequency for the experiment, the drive pressure
required frequent observation and periodical adjustment
since the spinning frequency had a tendency to drift. The
spinning frequency was typically maintained within 15 Hz of
the designated frequency.

Spectra were obtained at spinning frequencies, U}, of
4—7 kHz. A constant spinning frequency was difficult to
maintain at frequencies greater than 7 kHz, resulting in
peak broadening. At frequencies lower than 4 kHz, broad
peakwidths and multiple isotropic peaks led to sideband

overlap, thus complicating interpretation of the spectra.

5.2.3.5. Collection of Free Induction Decays

The parameters to set up the spectrometer for the
particular observe nucleus had already been selected as
described above, but it was necessary to select the

acquisition parameters for each experiment. These parameters

127
included the spectral width, number of scans, block size,
delay and pulse width. Several of these parameters are
listed in Table 5.1.

The experiment was started and free induction decays
(FIDs) were collected for at least 32 scans. At this point
the data was examined by performing a weighted Fourier
transform of the FIDs to generate an NMR spectrum. Phasing
was adjusted to produce a flat baseline and improve the
appearance of the spectrum. If the data exhibited a
reasonably high signal to noise ratio (S/N) by visual
inspection and the phasing was good, data collection was
terminated and the data was saved. If S/N or the phasing of
the spectrum was poor, additional scans were collected until
either the spectrum improved or no improvement was realized

with many additional scans.

5.2.3.6. Reference Sample

The spectrometer required referencing to a known
standard at the start of each daily session. Synthetic
calcium hydroxyapatite, CanOH)2(POn6, (Aldrich) was chosen
as a reference because it is a solid at room temperature and
may be analyzed in the same probe setup as the other solid
samples. The reference of choice for many NMR spectra is 85%
phosphoric acid, but it was deemed inappropriate for our
experiments since use of a liquid reference required the use

of a different probe. This has the possibility of

128
introducing error because of inconsistent sample placement
in the magnetic field, magnetic field inhomogeneity and
differences in probe tuning.

A spectrum of calcium hydroxyapatite was collected and
the transmitter offset was adjusted to place the chemical
shift of the isotropic central peak of calcium
hydroxyapatite at the literature value of 2.8 ppm : 0.2
(65). The hydroxyapatite chemical shift of 2.8 ppm is

relative to the standard reference of 85% phosphoric acid.

5.2.4. Experiments
5.2.4.1. Data Quality

The chemical shift of isotropic peaks is reproducible
to i 0.50 ppm for a typical spectrum with a signal to noise
ratio (S/N) of 44 for the largest peak. Typical
reproducibility of chemical shifts in spectra of identical
samples repeated during the same session is i 0.21 ppm.
Typical reproducibility of chemical shifts in spectra
repeated during subsequent sessions is i 0.50 ppm.
Fluctuations in the magnetic field over several days between
sessions may be responsible for the lesser reproducibility
for samples repeated during different sessions than for
samples repeated during the same session. Table 5.2. shows
typical and worst—case reproducibility and S/N for NaHfiXL
spectra. S/N for spectra of various compounds are presented

in Table 5.3.

Table 5.2. Reproducibility for NMR Spectra of NaHgXL

 

 

Typical Worst Case
Chemical Shift (ppm)' i 0.21 i 1.1
S/N' 38 11
Chemical Shift (ppm)" i 0.50 i 1.1
S/N" 38 11

 

 

* Identical samples repeated during same session.

** Identical samples repeated during subsequent sessions.

Table 5.3. Signal to Noise Ratios (S/N) for NMR Spectra

 

 

 

Sample S/N Range Typical S/N
Hydroxyapatite 150-760 507
NaHIZPO4 11—166 38
NaZHPO4 38-378 162
Na3PO4 68—740 232
Na4PgL 76-300 184
Peakwidths (FWHM) for the spectra cover a range from

1.05—18.3 ppm with a typical peakwidth of about 2.5 ppm.

Error in chemical shift values is attributed largely to

uncertainty in identifying the center of the isotropic peak

due to noise and large peakwidth. The average error

associated with the chemical shift increases with an

increase in peakwidth. This trend is clearly seen in Table

5.4. A wider peak has a broader apex and thus a less easily

established center.

Transmitter drift is believed to be

insignificant compared to other errors.

130

Table 5.4. Effect of Peakwidth on Chemical Shift
Reproducibility

 

Peakwidth (FWHM), ppm Error in Chemical Shift, ppm

 

.5 0.06
.35
.43
.56
.67
.91
.15
.30

CDCDCDQNUTNN
wtokowoopoo
I—‘l—‘OOOOO

H

 

5.2.4.2. Variation of Spinning Frequency

For each NMR signal in a sample, MAS spectroscopy
produces one band at the isotropic chemical shift frequency
accompanied by spinning sidebands flanking the isotropic
peak on each side. These sidebands are evenly spaced at
integer multiples of the spinning frequency. Variation of
the spinning frequency may be used to discriminate between
isotropic peaks and spinning sidebands. The chemical shifts
of isotropic peaks are independent of the spinning
frequency, while the apparent chemical shifts of spinning
sidebands change with spinning frequency. This technique
involves acquiring spectra at several different spinning

frequencies and observing which peaks change position.

131

5.2.4.3. Determination of Chemical Shift Anisotropy

The intensities of the sidebands relative to the
isotropic peak depend on several factors including the
magnetogyric ratio of the observe nucleus, the chemical
shift, anisotropy of the chemical shift tensor, the strength
of the magnetic field and the spinning frequency. A
graphical method by Herzfeld and Berger was used to
determine the three primary values of the chemical shift
tensor from the intensities of the sidebands and other
system parameters (88). A detailed example of the method is
presented in Appendix B.

The anisotropy of the chemical shift tensor may be used
to determine if two signals with nearly identical chemical
shifts are really identical signals, or merely have similar

chemical shifts.

5.2.4.4. Sideband Suppression

Heated samples of NaHfiKh exhibit at least six
isotropic peaks with a large anisotropy associated with four
of the signals to produce intense sidebands. This produces
very complicated spectra in which it is difficult to
distinguish overlapping sidebands from isotropic peaks.

VNMR Version 4.1 software contains a method for
suppression of spinning sidebands called Total Sideband
Suppression (TOSS). This method uses a cross—polarized 180°

pulse sequence to create timed spin-echoes to suppress

132

spinning sidebands. The method employs the principle that as
the echoes interact with the nucleus from opposite sides,
the spinning sideband signals cancel each other while the
isotropic signal is reinforced. The method requires setting
the initial pulse width and the spinning frequency. A recipe
is then invoked within the program to calculate the pulse
widths and delay times of the echo pulses. The method
requires optimization of the initial pulse width for the
particular sample by performing several runs at various
pulse widths to optimize sideband suppression. The optimum
pulse width was about 6.0 us for the heated samples of
NaHfiKh. A diagram of the TOSS pulse sequence used for
heated NaHfiKh is shown in Figure 5.2. and parameters are
listed in Table 5.5.

The TOSS sequence worked well for suppression of
sidebands in NaHgXN samples, but S/N is sacrificed. Typical
values of S/N decreased from 38 for the standard acquisition

sequence to 8.2 for the TOSS sequence.

5.2.4.5. Proton Decoupling

AX coupling is caused by oscillation of the coupled
nucleus X between spin states at the frequency of the
coupling constant J”. The energy driving this oscillation
is derived from the observed nucleus A causing an identical
splitting JDiin the spectrum of A. In ADO decoupling

experiments, the sample is irradiated with two RF fields

 

 

d1 d2

 

p1

 

 

 

 

 

d3

P" 2W]

d4

133

 

 

dS'
ZIIM

 

 

2w

 

Figure 5.2. Total Sideband Suppression

Table 5.5. Parameters for TOSS Pulse Sequence'

20M

 

 

d7

(TOSS)

Pulse Sequence

 

p1
pw
d1
d2
d3
d4
d5
d6
d7

0
6.
10.

0
4.
l.
25.
163.
113.

O

6
0
5
0
9

us

us
us
us
us
us

 

* Calculated for 6 kHz spinning frequency and optimized for

heated NaHZPO, .

134
simultaneously. The observe transmitter irradiates the
sample with an RF field B3.at the observe frequency of A,
1%: while the decoupler transmitter irradiates the sample
with another strong RF field B5:near the resonant frequency
of the decoupled nucleus X,‘Qv If Bzieastrong enough that
the frequency of the oscillations induced in X by Bz‘are
greater than JM, the energy driving the oscillations is
then derived from B2 instead of from the coupling with A and
the spin states are no longer seen separately by A, thus
suppressing the splitting (89).

A range of magnitudes of selected coupling constants
for 3fiflH and 31PMP coupling are presented in Table 5.6. The
second order coupling constant, 53%, is expected to be of
the same order of magnitude as 53% which is about 0.1 ppm
(90). With typical peakwidths (FWHM) of 2.5 ppm, it would be
very difficult to observe any second order coupling since
2JPH is an order of magnitude smaller than the peakwidth.
Several heteronuclear 3IP’C‘I—fiproton decoupling experiments
were performed to determine if some of the peaks observed in
complicated spectra were caused by splitting of phosphorus
signals due to coupling with hydrogen. No 31P1H coupling was
observed and all signals were found to be either distinct
31P signals or spinning sidebands. These observations are

developed further in Section 5.3.

135

Table 5.6. Spin-Spin Coupling Constants’

 

if

 

Coupling Type Functional Group J (Hz) J (ppm)

‘JPH R2PH=O 490—760 3.0—4.7
1.7,... R2PPR2 220—400 1.4-2.5
‘Jpp R2P(=O)P(=O)R2 330-500 2.0—3.1
2.7,...“ RQPCHL, 1—5 0 . 006—0 .031
2.7,...” R2P(=O)CH3 7-15 0.04—0.09
2JP,,P R2P(=O)OP(=O)R2 15-20 0.09-0.12
3qu POCH3 6-15 0.04-0.09

 

* From Emsley and Hall (90)
** At u0 = 161.903 Mhz

5.3. Results

5.3.1. Neat Salts
5.3.1.1. Calcium Hydroxyapatite

Synthetic calcium hydroxyapatite, CaHMCflU2(POn6, was
used as a reference standard and its chemical shift was set
at the reference value of 2.8 i 0.2 ppm (65). Synthetic
calcium hydroxyapatite was observed at spinning frequencies
of 4—7 kHz and exhibited one narrow isotropic peak. The
principal values of the chemical shift tensor agree with the
literature values within the error of the experiment.

Bovine dental enamel was studied after heating to 100°C
and 800°C. At 100°C, one signal was observed at 3.0 ppm.
This signal has the same chemical shift, within experimental
error, as synthetic calcium hydroxyapatite and the principal

values of the chemical shift tensor also agree within

136
experimental error, except the literature value for 033 is
outside of experimental error. At 800°C, three signals were
observed at 2.9, 0.5 and -4.5 ppm. The peak at 2.9 ppm
accounts for 78% of the 31P signal intensity. The two small
peaks probably correspond to a pyrophosphate which is formed
upon heating and possibly a non—stoichiometric
hydroxyapatite species. Both of these have been reported to
exhibit small upfield peaks in heated hydroxyapatites (65).
The anisotropy of the 2.9 ppm peak remains unchanged upon
heating to 800°C.

We conclude that calcium hydroxyapatite is the sole
component of both synthetic calcium hydroxyapatite and
bovine dental enamel heated to 100°C. Bovine enamel remains
78% hydroxyapatite upon heating to 800°C, with the remainder
probably composed of calcium pyrophosphate, CaZPHL. Spectra
of calcium hydroxyapatites are shown in Figure 5.3. Chemical
shifts and anisotropies for calcium hydroxyapatites are
presented in Table 5.7. Linewidths (FWHM) of both biomineral
and synthetic hydroxyapatite were typically 2.0—3.0 ppm, as

seen in Table 5.8.

5.3.1.2. Neat Monosodium Orthophosphate

Neat monosodium orthophosphate monohydrate, NaHJKhng,
exhibits a single isotropic chemical shift at 1.8 i 0.4 ppm.
The literature value of 2.3 ppm lies just outside the

experimental error (67). The principal values of the

137

C)

 

k L A

b) C)

 

 

 

 

 

,_____:_,:_ FL’A_A—-—- 4 LJ' A A
o-u-o """"""'.v-

P
C . . . " fl - C .-

a) Commercial sample synthetic Cam(OH),(POJ,. v, = 6 kHz. b) Bovine

dental enamel boiled at 100°C, VP 8 6 kHz. c) Bovine dental enamel
calcined in air at 800°C, v, = 6 kHz. Bars indicate isotropic peaks.

Figure 5.3. MAS—“P—NMR Spectra of Calcium Hydroxyapatites

138

Table 5.7. Chemical Shifts of Calcium Hydroxyapatites

 

 

 

 

 

 

 

 

 

Compound Chemical Principal Values of Ref.

Shift, ppm Chemical Shift Tensor, ppm

5 011 022 033
Cam(OH)2(PO,),5 2 8 18 i 9 —3 i 8 —24 i 10 *
Cam(OH)2(PO4),5 2.8 i 0.2 19 i 3 4 i 6 —15 i 1 (65)
Bovine Dental 3 0 30 i 17 6 i 22 -46 i 19 *
Enamel, 100°C
Bovine Dental 2.9 (78)" 30 i 11 -3 i- 12 -35 i 11 *
Enamel, 800 C 0.5 (21).

-4.5 (1) '
Human Dental 3.3 (65)
Enamel
* This work
** Parenthetical values are percentages of the total signal.
Table 5.8. Linewidths of 3'lP Signals
Compound Linewidth FWHM, ppm Ref.
Cam(OH)2(PO4L) 1.9-2.6 *
Bovine Dental Enamel, 100°C 3.1 *
Bovine Dental Enamel, 800°C 1.7 *
Human Dental Enamel 3 (65)

 

 

 

* This work

chemical shift tensor also lie just outside experimental
error.

Upon heating to 300°C, six peaks become visible at 2.0,
-8.3, —15.6, —19.4, -24.1 and -26.3 ppm. Analysis by both
spinning frequency variation and spinning sideband
suppression (TOSS) confirm that all six peaks are unique

signals and are not spinning sidebands. 3Tfififl proton

139
decoupling experiments confirm that there is no splitting
into multiple peaks due to BWVH coupling.

Upon further heating to 450°C, five signals appear at
—5.4, —15.3, -19.0, -24.0 and —26.2 ppm. Again, additional
experiments confirm that all peaks are unique signals and
not sidebands or split peaks, as described above. TOSS
analysis indicated an additional signal at about 9.0 ppm,
but spinning frequency variation revealed that this was due
to either noise or a lingering spinning sideband in the TOSS
spectrum.

Linewidths (FWHM) were typically 2.5—4.0 ppm for neat
heated and unheated NaHfiKh samples. Reproducibility of
chemical shifts was typically better than i 0.5 ppm.

The signal at 2.0 ppm is assigned to NaHfiXL and
decreases in size as the temperature increases, indicating
NaHfiKN is converting to other species. This signal has
virtually disappeared by 450°C. A signal appears at —8.3 ppm
at 300°C, but disappears again by 450°C, to be replaced by
another peak at —5.4 ppm. These peaks match the central
phosphate units of the linear sodium triphosphates in
literature, Nangxo—I at —8.0 ppm and NaEPfixo-II at —5.6 ple
(61,62,64,68,70). However, the signal near —5.6 ppm is
believed to arise from the trimeric ring, trimetaphosphate,
NaBPfh, because of several considerations detailed in
Section 5.4.1. The linear sodium polyphosphates, (NaPO3)6

and (NaPOfl also show a small peak near -6 ppm which is

n'

140
assigned to a terminal phosphate unit in the literature
(64). We assign the peak at ~8.3 ppm to NafifiOm—I and the
peak at —5.4 ppm to NagPfih. There is a strong possibility
that the -6 ppm peak which was assigned to (NaPO3)n in the
literature is actually due to small amounts of NaJfiOw-II or
bthfih in the polyphosphate sample.

We were able to determine the anisotropy of the —8.3
ppm peak, but the anisotropy was not reported in the
literature. Conversely, the anisotropy of the -5.6 ppm peak
was reported in literature, but we were unable to determine
an anisotropy because the sidebands were too small.
Therefore, no comparison of anisotropy was able to be made
in order to confirm the peak assignments.

We assign all the peaks at —15.3, -19.0, —24.0 and
—26.2 ppm to linear sodium polyphosphate, (NaPOfln. Each of
these signals was observed by Prabhakar, et al., while the
peaks near —16 and —19 were also observed by other
researchers (63,64,66). The MAS—“P—NMR spectra from normal
acquisition, spinning frequency variation, TOSS and 31PPH}
decoupling of neat monosodium phosphates after calcination
are presented in Figure 5.4. The chemical shifts and tensors
are listed in Table 5.9.

Neat NaHfiKhflgO converts partly to NagfiOHpI and (NaPO3)n

at 300°C and then completely to Naysrr and (NaPOfl at 450°C.

n

The four polyphosphate peaks correspond to various phosphate

units with different thermal stabilities. The units giving

141

 

 

 

 

 

 

 

 

 

 

'v V v- v v r I V V T * T‘fiw—V
Id I. I. O I C C 0 -I -‘ -. -. 0|. a. - 3' . Q -” q” ... *
II
I
I
I
I
I

' v r v v r I V'r r r rVT v fV fY' VrV' r'ff 7V v rrrrr Y—v ' 'fi
I. C C d D I -. -Q -. ‘ H. H. 4‘ fl. - I. ' O -. -- fl. '

I
I ' f)

 

 

.1 all: '.I III] LIL I‘LJI'LO. ‘4 Id“...

0 . . O 'N " '. . “I” 'I. '0. -

 

 

a) Commercial sample NaHzPOpHZO, v, = 7 kHz. 6) NaH,PO,-H,O calcined in
air at 300°C, v, = 4 kHz. c) :300'C, v, = 6 kHz. d) :300°C, v, = 7 kHz.
e) :300'C, TOSS, v, = 6 kHz. f) :300'C, without ’Wflflu decoupling, v, =

6 kHz. Bars indicate isotropic peaks.

Figure 5.4. MAS-“P-NMR Spectra of Neat Monosodium
Phosphates

142

 

 

 

 

(E -

.e r r r T v 1 v r v v v w r r i T v v y w . , . r , v
I. . fl ‘ . 0 °” “ '. ‘- 'I“ 'I‘ ‘0‘ 'I. * I. . fl . a 0 " ‘C '. '. '0“ “C 'I. 'I. -

 

I)

 

 

 

 

k)

 

g) Naagwndgo calcined in air at 300°C, with “PPHIdecoupling, v,==6 kHz.
h) ;450°C, v, = 7 kHz. 1) :450°C, T038, 0, = 6 kHz. j) :450’C, without

“PPHIdecoupling, v,= 6 kHz. k) :4SOTL with ”PPHIdecoupling, v,=
6 kHz. Bars indicate isotropic peaks.

Figure 5.4. MAS-“P—NMR Spectra of Neat Monosodium
Phosphates (Continued)

143

Table 5.9. Chemical Shifts of Monosodium Orthophosphates

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

EEEEEE=== Chemical Fraction Principal Values of Ref.
Compound Shift,ppm <of Chemical Shift Tensor,ppm
Signal,%

5 j 01: G22 03:
15163229431132 —
lfifihPOgHg) 1.8:0.4 100 66:10 10:12 -81:10 *
lfiflLEQFJMMEQ
NaHzPO4-HZO 2 . 0:0 . 7 13 *
NagfiOKpI —8.3 19 89:13 10:2 -73:13 *
(NaPO3),1 —15.6 9 137:15 65:24 —154:17 *
(NaPO3)n ~19.4 23 128:7 60:7 -129:9 *
(NaPO3L, -24.1 23 131:13 64:14 —121:16 *
(NaPO3),1 —26.3 13 141:18 65:16 -126:21 *
lfifibEthfiflfifl I
Na3p3o9 -5.4 21 *
(NaPO3),1 ’—15.3 26 129:21 68:22 —141:25 *
(NaPO3),1 —l9.0 42 130:8 66:9 —138:9 *
(NaPOB),1 —24.0 6 140:11 62:8 —130:l4 *
(NaPO3),1 -26.2 5 *
hEfibBQlJELLAL_lQQBZ
NaHZPO4-HZO 2 . 4 7 9 *
Na3PBOq -5.9 21 *
bkfikEQnJELU£L4flMEIL
Na¥%OKfI —8.8 44 102:13 16:14 —92:l3 *
(NaPO3)n —19.5 56 122:10 45:13 -108:10 *
DkflbEanfiJ£d_A§fl£C
Nafiaog —5.8_ 100 88:6 9:7 —81:6 *

 

 

 

 

 

144

Table 5.9. Chemical Shifts of Monosodium Phosphates

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

(Continued)
Sample Chemical Fraction Principal:Values of Ref.
Compound Shift of Chemical Shift Tensor,ppm
Signal,%
— 6 01 C22 ‘03: __
W
Monosodiumfirthonhosnhate
NangrqgoI 2.3 I | 77.3 -23.1 -61.1l(67)
1' .1 1
Naspgim-I 1.2T (70)
-8.oC (70)
NaspgiO-II 5.6T (61)
—6.2” (61)
Naspgio—II 4.0T 62 37 —113 (62)
—5.7“ 118 -22 —79 (62)
Nasper-II 5.1T 62 32 -109 (64)
-5.6” 120 -24 —79 (64)
Nasper—II 4.6 69 24 —107 (68)
—5.6 117 -20 —80 (68)
NaspxiO-II 4.6T (70)
—6.6“ (70)
W
Na6P$L3 1.5T 47 47 —98 (64)
-16.4“ 152 —18 —85 (64)
(Napo3)6 -6.0T (64)
-19.6" 165 -19 —87 (64)
(NaPO3LU 2.0T 45 45 -95 (63)
“”25 —19.0° 160 -13 —9o (63)
(NaPO3L. —15 6 (66)
—18.5 (66)
-19.9 (66)
—24.1 (66)
—26.8 (66)
(NaPO3L1 1.0T (64)
-6.2T (64)
-19.9° 165 -20 —85 (64)

 

 

 

 

 

* This work
C Central phosphate unit
T Terminal phosphate unit

145
rise to signals at ~15.3 and -19.0 ppm are more thermally
stable at 450°C than those at —24.0 and -26.2 ppm. Table
5.10. shows that the less stable species convert to the more
stable species as the temperature rises so that 68% of the

signal is due to the -15.3 and —19.0 ppm peaks at 450°C.

Table 5.10. Conversion of Phosphate Units to Thermally
Stable Species for Neat NaHfiKhng

 

 

 

Phosphate Chemical Percentage of 31P Signal
Species Shift, ppm at Temperature T, %

6 T = 25°C 300°C 450°C
NaH2 PO4-H20 2 . 0 l 0 0 l3
Na5P3OlO-I —8.3 19
Na3P3O9 —5.4 20
(NaPO3),1 -15.3 9 26
(NaPOB),1 -19.0 23 42
(NaPO3),1 -24.0 23 7
(NaPO3),1 -26.2 13 5

 

 

 

5.3.1.3. Neat Disodium Orthophosphate

A single peak is present at 6.3 ppm in the spectrum of
neat disodium orthophosphate, Nafifixh. This chemical shift
is within 0.3 ppm of the literature value (67) and the
anisotropy also compares well with literature.

After heating to 300°C, complete conversion occurs to
form sodium pyrophosphate with a signal at 1.9 ppm and a
small amount of unknown at 10.8 ppm, which only makes up

1.3% of the signal. Further heating to 450°C again produces

146
mostly pyrophosphate at 2.2 ppm and a small amount of
unknown at 13.6 ppm. This unknown makes up only 2.8% of the
signal and has a chemical shift that corresponds with
trisodium orthophosphate, Nafixh. The peak was too small to
calculate the chemical shift anisotropy, so it is impossible
to ascertain whether its anisotropy is comparable with that
of Nafixh. The spectra of neat disodium phosphate are
presented in Figure 5.5. and the chemical shifts and tensors
are tabulated in Table 5.11.

Both the chemical shift and anisotropy of Nafiuwh
compare well with values reported in the literature. The
chemical shifts of pyrophosphate also agree with literature,
but the anisotropies do not. The chemical shift anisotropies
are self-consistent with samples calcined at other
temperatures and also with commercial samples of
pyrophosphate, but our results for pyrophosphate chemical
shift tensors differ from literature results by typically 23
ppm for both 0.11 and 033 after accounting for experimental
error. A possible explanation for this discrepancy is

suggested in Section 5.3.2.4.

5.3.1.4. Neat Trisodium Orthophosphate

The only transformation observed upon calcination of
neat RBQPOg12HgD is a simple dehydration to the anhydrous
form, NafiXh. One peak is present at 7.5—7.9 ppm for a

commercial sample of NafiXhfihfigO at 25°C. This chemical

147

a) b)

 

 

 

 

 

''''''''''

C) d)

 

 

 

 

11L 'Ulll_ -ltiL 'LJL..IL__~

tommuuauo-u-d-o-o- I.I.I.....O-.-Q-flc.—

 

 

 

a) Commercial sample NazHPO“ v, = 6 kHz. b) :v, = 7 kHz. c)-Na2HPO.
calcined in air at 300°C, v, = 6 kHz. d) :450°C, v, = 6 kHz. Bars
indicate isotropic peaks.

Figure 5.5. MAS—“P—NMR Spectra of Neat Disodium Phosphates

148

Table 5.11. Chemical Shifts of Disodium Orthophosphates

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Sample Chemical Fraction Principal Values of Ref.
Compound Shift,ppm <1f Chemical Shift Tensor,
Signal,% ppm
5__ 011 O22 O33
DEQHEqulifil
NaZHPO4 | 6.3:O.3| [ 67:10 ~19:11 ~67:10 ]*
Nafiflih_3flflic
Unknown 10.8 1.3 *
Na4P207 1 . 9 98 . 7 *
NaZHBQ4_4.iQ:C
Na3PO4 (?) 13.6 2.8 ~ *
NafifiO7 2.2 97.2 84:10 -34:16 —57:9
HQQHEQ4 . O
NazHPO4 6.5 89
Na4PZO7 1.3 11
DEQHEQnifi42fl_jflD£C
Na4PZO7 [1 2.9 | 100 1 1*
Ndfifiih ' °
NafifiO7 ~2.3 91 78:5 -29:5 ' -55:4
Na4PZO7-nHZO -2 .0 9 *
Wines
. 1. 1 1 1
NaZHPO4 6.6 48.9 5.7 —74.5 (67)
NaZHPO4 -16 6 69.8 35 5 —55.4 (68)
.12H20
. j' ] 1 ]
Na3PO4 14.0 (64)
Na3PO4 13.0 (69)
Na3PQA 13.7 (66)
W
1061,1320, L 2.4 37 37 ~80 (63)
1.2 37 37 -80 (63)
1061,1220, 3.0 40 40 ~89 (64)
2.2 40 40 —89 (64)
Na4P207 3.2 (66)
2.1 (66)
Na4PgL 2.7 46 33 —87 (68)
1.3 50 27 —81 (68)
Na,,PZO7 —1.4 (64)
'10H20 ~2.3 38 38 ~70 (64)
Na4PJL -1.7 49 32 —76 (68)
°12H20 ~2.0 58 26 ~78 (68)

 

* This work

 

 

 

149

shift agrees with literature results within 0.3 ppm (64,67).

By 300°C, complete dehydration has occurred, yielding
anhydrous Nafixh, indicated by a single shift at 14.2 ppm.
No further change is seen up to 450°C, with a single signal
observed at 13.9 ppm. The anisotropy changes very little
upon dehydration, as seen by the results in Table 5.12.
Spectra of neat trisodium phosphates are presented in Figure

5.6.

5.3.1.5. Neat Sodium Pyrophosphate

Commercial samples of neat sodium pyrophosphate,
DEHPJL, exhibit only one peak corresponding to anhydrous
pyrophosphate at a chemical shift of 2.3 : 0.3 ppm. The
same signal is observed at calcination temperatures of 25,.
300 and 450°C. The spectra are shown in Figure 5.7. and the
chemical shifts and anisotropies are presented in Table
5.13. Data quality, chemical shift anisotropy and
comparisons with literature results are discussed in Section

5.3.2.4.

5.3.2. Supported Salts
5.3.2.1. Supported Monosodium Orthophosphate

Monosodium orthophosphate supported on silica alumina
also undergoes changes in composition as the calcination

temperature increases. At 100°C, two signals are present at

150

 

 

 

I uj

 

 

 

 

 

 

 

 

I I I O I' I I. v u u -, , , , Y a,
«O -I OI O -I -‘I I I I C I I I. 0 u. -I -I 0Q - -. .-
I I
1 ii _..
v v I - v - I v I V I W V 7' I V V r"' '
I I I. I O -I «I at a u- I I I. I I I. I I. I I a no

a) Commercial sample Na,PO.-12H,O, v, = 6 kHz. b) :v, = 7 kHz.

c) Na,PO.-12H20 calcined in air at 300°C, v, = 6 kHz. d) :450°C, 0, =
6 kHz. Bars indicate isotropic peaks.

Figure 5.6. MAS-”P—NMR Spectra of Neat Trisodium Phosphates

 

 

151

Table 5.12. Chemical Shifts of Trisodium Orthophosphates
Ref.

 

Sample Chemical Fraction Principal Values of
Compound Shift,ppm of Chemical Shift Tensor,
Signal, % ppm

5 011 O7? 033

a-

 

 

 

 

 

Na3RQ4J5E
Na3PO4-12HZOI 7.7:0.2 I 100
112.313.04.10019‘.
NalPOL [14 .2
Na3EQ4_45_(KC
Nafixh [13.9 [ 100 | 11:22 -14:23 ~39:221*
Na3BQ4_SiLAl_lll(r_C

Na3PO4 1 *
Na3PO4012H2O
NaiP3Om-II
Na3PQ4_SiLAl_3D_0_°C

Na3PO4 14.0 96.8 *
Unknown 6 . 6

Na3PQ4_S_i_LAl_45_0£
Na3PO4 13.7 100 9:19 -14:21 —37:19 *

literature MEJHES
. 1' 1 1 ]
NaJPO4 14 .
Na3PO4 13 .
NasPO‘1 13.
Na3PO4-10HZO 7 .
NaBPO4-12HLO 7 .
3' . 1 ]

NaSPBOlo-II 5.

[ 22:16 -8:22 —38:16 I*

1 100 I 12:21 -14:23 -40:21I*

 

 

*

*

WCDUJ

\ONO‘I
H
H
DJ

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

CD\I\IOO

 

 

 

’3

'3

(

(

62 37 -113 (
118 -22 -79 (
62 32 -109 I
(

(

(
(

*3

Nasp3om—II 4.

“J

.3

Na5P3Om-II 5.
120 —24 —79

69 24 -107

Nasp3olo-II 4.
117 ~20 ~80

v-l

 

NaSPBOlo—II 4.

' l

 

 

I
m
mmmmqwqowm

 

 

* This work
T Terminal phosphate unit
C Central phosphate unit

 

 

 

 

 

 

 

a) b)
l
I I
c) d)
i
l
y I
l_~_
u I
2)

 

I- I- I" l‘ - I C I I s. ‘C -. ~- 'I- - I. l‘ u. I. U - C I O " " .- >- -I- -

kHz. b) Na.P,O. calcined in air at

a) Commercial sample Na,P,O,, v, = 7
300°C, v, = 6 kHz. c) ;450°C, v, = 6 kHz. d) Na,P,O, on Si/Al calcined in
air at 100°C, v, = 6 kHz. e) :300’C, v, = 6 kHz. f) :450’C, v, = 6 kHz.

Bars indicate isotropic peaks.

Figure 5 .7 . MAS-3lP—NMR Spectra of Sodium Pyrophosphates

 

153

Table 5.13. Chemical Shifts of Sodium Pyrophosphates

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Sample Chemical Fraction Principal Values of Ref.
Compound Shift,ppm of Chemical Shift Tensor,ppm
Signal,%
5 Oil 0.22 033
1316422071122
Na4P207 2.6 100 84:12 —37:13 —55:10 *
M42297J121‘1C
Na4PZO7 2.0 100 *
15342297415922
Na4PZO7 2.4 100 85:9 --38:12 -54:8 *
M4BZQ7_S.1LAJ._J.QD_£° °
Na4PZO7 2.4 100 *
N34220,_51131_300:c
Na4PZO7 2.2 100 *
M4BZQ7_SiLAl_450:C
Na4PZO7 2.4 100 84:12 —35:16 -57:11 *
W
W
Na4PZO7 2.4 37 37 —80 (63)
1.2 37 37 -80 (63)
Na4PgO~. 3.0 40 40 —89 (64)
2.2 40 40 —89 (64)
Na4PZO7 3.2 (66)
2.1 (66)
NaqPEO7 2.7 46 33 —87 (6 )
1.3 50 27 —81 (68)
Na4P207 —1.4 (64)
'lOHZO ~2.3 38 38 ~70 (64)
Na4P207 ~1.7 49 32 ~76 (68)
'lZHZO ~2.0 58 26 -78 (68)
Na2H2P207 t-13.0 _ 80 24 -65 (68)

 

 

 

 

 

* This work

 

154

2.4and-5.9 ppm, which are assigned to NaHfiKbng and

Nmpfig,respectively.
At300°C, two peaks are observed at chemical shifts of

-8.86md —19.5 ppm. These are assigned to NaJ§OE-I and

(NaR%)M respectively. The principal values of the chemical
shifl:tensor also agree within experimental error with both

literature and experimental values except for (333 for -19.5
but

ppm,vdmch is slightly outside of experimental error,

A third signal at about 17.2 ppm was indicated

still close.
but we were unable to confirm the peak by

by TOSS analysis,
spinning speed variation due to sideband overlap by the
other strongly anisotropic signals. No corresponding signals

were observed in our experiments or in literature which
suggest a possibility for the identity of this peak. The

peak is located at an interval of one spinning frequency

—19.5 ppm, suggesting the

downfield from the signal at

‘possibility that it is a lingering spinning sideband. In any

case, time signal is weak and only contributes 12% to the

tCfiXil signal in the TOSS spectrum. 31PVH}decoupling
expmnxunents demonstrate that both signals at —8.8 and -19.5

ppm are unique signals.
(hie signal is observed at 450°C at a chemical shift of

-5.53 ppun which is attributed to NagRfik. This signal is very

lxroeui (IUNHM 18 ppm), the spectrum is noisy and a large

aruiscfl:rcmmy contributes to large sidebands. These factors

CCHnbille t:o make it difficult to determine whether there is

 

 

155

onh/omesignal or several signals in close proximity. The

tmahfimpes suggest the latter. Spinning frequency variation

and3HWH}decoupling experiments show that if there is

indeainpre than one peak, they are within about 5 ppm of

eachcxher. The principal values of the chemical shift

tenanrare close to the literature values, but on and 022

arerxm within experimental error. The spectra for supported

 

rmxmpodium phosphate are presented in Figure 5.8. and the

chemical shifts and tensors are shown in Table 5.9.
Supported monosodium phosphate converts to different
phosphate species as the temperature rises. The percentage

of the total 31P signal that is assigned to each species is

The reproducibility of the chemical

shown in Table 5.14.
(FWHM)

1.1 ppm at 300°C and 450°C. The peakwidth

shifts is :
9 ppm at 300°C and 18 ppm at

is typically 5.4 ppm at 100°C,

450°C.

Supported Disodium Orthophosphate

5.3.2.2.
disodium

Dflaen supported on silica alumina,

ortlmxflmosphate undergoes the same basic transformations as

Iieat: sanuples. The peakwidths (FWHM) for both neat and
two

:supmxortexi disodium phosphate are 2.1—3.5 ppm. At 100°C,
ssigruals eare seen at 6.5 and 1.3 ppm, which correspond to
complete

NaZHPO4 and Na4P207, respectively. By 300°C,

ccnivearsixon to pyrophosphate has been achieved. At 450°C,

 

156

 

 

 

 

'IC H. fl. *

dI

 

 

 

 

 

 

 

 

v I r I r ' I I r I’
. fl . u . ‘. ’. '“ ’. ". "I "‘ ‘I. -

a) NaHfiKLdfiO on Si/Al calcined in air at 100°C, v, = 6 kHz. b) :300°C,
v, =- 4 kHz. c) :300'C, v, - 6 kHz. d) :300°C, v, = 7 kHz. e) :300‘C, TOSS,
v, =- 6 kHz. )5) :300°C, without ”PI‘HI decoupling, v, = 6 kHz. Bars
indicate isotropic peaks. '

Figure 5.8. MAS-31P-NMR Spectra of Supported Monosodium
Phosphates

 

 

 

157

9)

 

 

 

I+
I
‘1
I
o
I

 

 

 

k)

'''''''''

g) NaH,PO,-H,O on Si/Al calcined in air at 300°C, with nPI‘H) decoupling,
V, = 6 kHz. h) :450'C, v - 6 kHz. i) :450°C, v, =- 7 kHz. j) :450'C,

without ”Pl‘m decoupling: v, =- 6 kHz. k) :450°c, with "em decoupling,
v, =- 6 kHz. Bars indicate isotropic peaks.

Figure 5.8. MAS—3‘P-NMR Spectra of Supported Monosodium
Phosphates (Continued)

 

 

 

158

Table 5.14“ Conversion of Supported Monosodium Phosphate

with Calcination

 

 

 

Phosphate Chemical Percentage of 31P Signal
Species Shift, ppm at Temperature T, %

O T = 1000C 300°C 450°C
NaHZPO4-HZO 2 .4 79
Nasppm-I ~8.8 44
Na3P3OQ —5.9 21 100
(NC-11:0.)n —19.5 56

 

 

 

 

anhydrous pyrophosphate is observed with a small impurity of
hydrated pyrophosphate. Both hydrated pyrophosphates,
Na4P207-10HZO and Na4PZO7-12HZO, have shifts within 0.3 ppm of
the observed signal and the peak was too small to measure
the anisotropy, so it is impossible to distinguish which
hydrated species is present. The spectra of supported
disodium orthophosphates are presented in Figure 5.9. and
the chemical shifts and anisotropies are tabulated in Table

5.11.

5.3.2.3. Supported Trisodium Orthophosphate

Spectra of NafiXh supported on silica alumina at 100°C
reveal both hydrated and anhydrous trisodium phosphate as
well as a small amount of sodium tripolyphosphate, NaSPJNO-
II, with a shift of 3.9 ppm. Heating to 300°C results in
complete dehydration of trisodium phosphate, as well as

destruction of the triphosphate and the appearance of a

 

159

 

 

 

3)
I I
b) C)
, , MJ‘.' .1....

a) Na,HPO. on Si/Al calcined in air at 100°C, v, = 6 kHz. b) :300‘C,
v, = 6 kHz. C) ;450°C, 0, = 6 kHz. Bars indicate isotropic peaks.

Figure 5.9. MAS—“P-NMR Spectra of Supported Disodium
Phosphates

160
small unknown at 6.6 ppm. Both triphosphate and the unknown
contribute less than 4% to the total 31P signal. By 450°C,
the only species observed is anhydrous trisodium phosphate.
Each of the principal values of the chemical shift tensor

for anhydrous Nafifih varied by less than 3 ppm between the

neat and supported samples, which is well within the

experimental error of 20 ppm. Reproducibility for trisodium
phosphate spectra is : 0.2 ppm and peakwidths (FWHM) of

signals are 4.0-6.6 ppm. NMR spectra of supported trisodium

phosphate are presented in Figure 5.10. Chemical shifts and

tensors are shown in Table 5.12.

5.3.2.4. Supported Sodium Pyrophosphate

Spectra of supported sodium pyrophosphate show pure

pyrophosphate at calcination temperatures of 100, 300 and

450°C. This demonstrates the thermal stability of

pyrophosphate. The chemical shift of pyrophosphate agrees

with literature results within : 0.4 ppm. The chemical shift

anisotropy of the various samples agrees within : 2 ppm and
peakwidths (FWHM) are 2.0—3.0 ppm. The spectra of neat and
supported pyrophosphates are presented in Figure 5.7. and
the chemical shifts are tabulated in Table 5.13.

A significant difference between our results and those
of literature is our observance of only one peak for
pyrophosphate. Several researchers reported two distinct

chemical shifts for pyrophosphate, with 0.8-1.4 ppm

161

 

 

 

 

 

 

 

 

 

 

 

I I I I I I ' I v I 'I I v ' v I ' I V
i. I I C I n I. I 'l. °. '. 0‘ '. '. —
- vv v W v v rfi fiv '7 v v r .1
, I I- v , v- I ,.v I .v 1 I" v v I I I I
N I I C I I I. I OI. °. -I -. a -. - I I I . i I. I '0. I . I ’

 

a) NafiKLdfiO on Si/Al calcined in air at 100°C, v, = 6 kHz. b) :300°C,
0, = 6 kHz. c) :450°C, 0, = 6 kHz. Bars indicate isotropic peaks.

Figure 5.10. MAS-“P—NMR Spectra of Supported Trisodium
Phosphates

162
difference between the two signals (63,64,66,68). Our
results do not show two distinct peaks, but the lineshapes
of many of the spectra suggest that there may be more than
one peak. The difference between the two peaks is estimated
to be 0.3-0.5 ppm, which is considerably smaller than that
reported in literature.

As seen before with pyrophosphate formed from calcined
disodium phosphate, the chemical shift anisotropies from our
results are self-consistent with other samples containing
pyrophosphate. But our results for pyrophosphate chemical
shift tensors differ from literature results by typically 23
ppm for both 011 and 033 after accounting for experimental
error. A possible explanation for this error is the
existence of multiple peaks with almost identical chemical
shifts but dissimilar anisotropies. This leads to the
possibility that the observed sideband intensities arose
from different anisotropies, depending upon which signal
gave the largest sideband intensity at that particular
sideband. This could lead to an error in the calculated
anisotropy, since the method used to calculate the
anisotropy assumes that all the spinning sidebands belong to

the same signal with an identical anisotropy.

 

 

163

5 . 4 . Discussion

5.4.1. Thermal Transformations of Sodium Phosphate Salts

Calciunlhydroxyapatite forms about 20% pyrophosphate
qun1 heating to 800°C. Neat monosodium orthophosphate
inonohydrate converts to mostly linear polyphosphate and
atmnn: 20% trimetaphosphate upon calcination at 450°C while

supported monosodium orthophosphate forms exclusively

trimetaphosphate under the same conditions. Both neat and

supported disodium orthophosphate condense to form >98%
pyrophosphate after calcination to only 300°C. Neat
trisodium orthophosphate dodecahydrate undergoes a simple
dehydration upon calcination to yield the anhydride.
Supported trisodium orthophosphate also yields >96%
anhydrous trisodium phosphate upon calcination at 300°C.
Pyrophosphate appears to be one of the most thermally stable
of the phosphates studied with only pure pyrophosphate
observed after calcination at 450°C for both supported and

neat salts. The transformations of phosphate salts upon

calcination are outlined in Figure 5.11.

A review of the thermal transitions in the monosodium
phosphate system was reported by Greenfield and Clift (58).
Our results agree with theirs for the phosphate species
formaiimmn calcination with the exception of one point.

rI‘heyreport the formation of the ring trimetaphosphate,

Nafigow as opposed to the linear tripolyphosphate, NaSPJLO.

164

O 'te

100°C 800°C
Cam(OI-I)2(PO,)6 -’ Calo(OH).-_,(PO,)O -' Ca10(OH)—_,(PO4),, + Ca._,P._,O7
Hydroxyapatite Pyrophosphate
79% 21%
s ’u os te
300°C 450°C
NaHzPOpHZO '* (NaPOJn + NasP,Om-I + NaHJPOng -‘ NaBPDQ + (NaPOfln
Monosodium Polyphosphate Tripolyphosphate Trimetaphosphate
Phosphate 68% 19% 13% 21% 79%
Sppported Moposodipm Phosphate
100°C 300°C 450°C
NaH3P0,-H20 + Na,P,Oq ~ (Napoj) n + NaSP,Om-I ~ Na3P3OQ
Monosodium Trimetaphosphate Polyphosphate Tripolyphosphate
Phosphate
79% 21% 56% 44%
Meat Dispdipm Phosphate
300°C 450°C
Na._,HPO4 -° Na4PJO7 + Unknown -° Na,P._,O1 + Na4PZO7-nHZO
Disodium Pyrophosphate
Phosphate 98.7% 1.3% 97.2% 2.8%
L e ‘s " os te
100°C 300°C 450°C
NazHPO, + Na4P:O7 -* Na4PJO7 -° Na4P._,O7 + Na4P307-nI-IEO
Disodium Pyrophosphate '
Phosphate
89% 11% 91% 9%
e t '50 ' os te
300°C 450°C
Na,PO,-12H30 -’ Na3PO4 “ Na_.PO4
Trisodium
Phosphate
SU‘ ‘t 's "L hos ate
100°C 300°C 450°C
Na3PO, + Na3PO,-12H_,O + NaSPlom-II -’ NajPO, + Unknown -' Na.,PO4
Trisodium Tripolyphosphate
Phosphate
85% 11% 4% 97% 3%
e d S o e \ 0 OS te
300°C 450°C
Na.,P-_,O7 -' Na4P._,O7 -' Na4PJO7
Pyrophosphate

Figure 5.11. Thelermal Transformations of Calcined Phosphate
Sa ts

 

 

165
Chemical analysis was used in the study and the methods used
included ion exchange, qualitative analysis using various
reagents, colorimetry, solubility and gravimetric analysis.
The work appears to be very thorough; there is no reason to
believe that tripolyphosphate was misidentified as
trimetaphosphate in the study.

Using NMR, we observed a signal which corresponded to
one of the two signals reported in literature for
tripolyphosphate, NaJfiOw—II. This chemical shift matches
that for the central phosphate unit. The peak for the
terminal phosphate was never seen, although it should be the
more intense of the two signals. This points to the
conclusion that the central phosphate peak from
tripolyphosphate has a chemical shift that is very close to
that of trimetaphosphate, since the two phosphorus atoms are
in similar environments between two phosphate units. Also,
the stoichiometric product of condensation of three
monosodium phosphate equivalents is trimetaphosphate, not

tripolyphosphate.

3 NaHZPO4 —> Na-JrP3O3 + 3 H30

Since we observe nearly pure product at 450°C for supported
monosodium phosphate, it is not possible to form
tripolyphosphate without abstracting the required excess

sodium ions from another source, either on the support

166
surface, or from another species. These considerations lead
us to the conclusion that the observed signal at -5.8 ppm
arises from trimetaphosphate rather than tripolyphosphate as

we had first thought to be the case.

5.4.2. Effects of Catalyst Support on Thermal
Transformations

Only in the case of monosodium phosphate does the
support have a major effect on the phosphate species formed
after calcination. Neat monosodium phosphate forms mostly
polyphosphate while the supported salt forms only
trimetaphosphate, Nafigog.

The role of the support in which product is formed may
be explained by several considerations. The support may
isolate the phosphate ions, causing it to be more difficult
for more than three phosphate units to come into close
proximity in order to react. This may be caused by spreading
out of the phosphate on the surface, resulting in a physical
distance between the phosphate units. This may also be
caused by the attraction of the phosphate ions around
specific sites on the support surface. The latter
explanation is very plausible since we know the support to
possess acidic sites, which would conceivably be attractive
to the basic phosphate ions. The acidic sites may also be
combining with the phosphate ions during reaction, thus

changing the mechanism and the product of the reaction.

167
Another explanation may be the fact that the support
places the phosphate units on a more or less two dimensional
surface with respect to each other. This orientation may
make the trimetaphosphate product more attractive either
because of the reaction stereochemistry or simply because
trimetaphosphate is more stable than polyphosphate on a flat

surface.

5.4.3. Comparison of NMR and FTIR Results

Our results using FTIR to characterize phosphate
catalysts indicate the formation of polyphosphate upon
heating of monosodium phosphate above 250°C, in agreement
with our NMR results in Section 6.3.2.1. FTIR spectroscopy
demonstrated the conversion of disodium phosphate to
tetrasodium pyrophosphate above 250°C, also in agreement
with results obtained using NMR. The one difference between
the two studies is the fact that FTIR indicated the
formation of pyrophosphate from trisodium phosphate above
200°C, whereas NMR showed simple dehydration of the hydrated
trisodium phosphate. This difference is explained by noting
that the FTIR study was performed on a system in which
lactic acid was deposited on the surface of the trisodium
phosphate, in contrast to the NMR study in which phosphate
was observed with no lactic acid present. The lactic acid
abstracts sodium ions from the trisodium phosphate to form

sodium lactate and disodium phosphate. The disodium

168
phosphate then condenses to form pyrophosphate in the FTIR
study, while in the NMR study, the trisodium phosphate
remains unconverted.

The findings of this study, when combined with results
in the presence of lactic acid in the FTIR study, help
explain some of the trends observed in the conversion of
lactic acid over sodium phosphate catalysts. These trends

are covered in detail in Section 6.4.

Chapter 6. Identification of Surface Species on Phosphate
Salt Catalysts using Transmission FTIR

6.1. Introduction

In this study, IR-transparent disks were coated with
sodium phosphate catalyst salts and then exposed to lactic
acid vapors at 25—3SOOC. The treated disks were then cooled
and Fourier Transform Infrared spectroscopy (FTIR) was used
to determine the chemical species present on the disk. This
study increases our knowledge of the species present on the
surface of phosphate salt catalysts when treated with lactic
acid at reaction temperatures. This leads to better
understanding of why certain trends in conversion and
selectivity are observed for actual reactions of lactic acid

over phosphate salt catalysts.

6.2. Methods

6.2.1. Materials

6.2.1.1. Chemicals

The origin of materials used is discussed in Section

3.2.1.

169

170

6.2.1.2. IR Sample Disks

Silicon IR sample disks were obtained from Spectra Tech
Incorporated with the dimensions of 9 mm diameter and 2 mm
thickness. Silicon was chosen as the supporting material for
these transmission FTIR studies because it is insoluble in
water; is mechanically, chemically and thermally stable; and
has a wide IR—transparent window from 8300-660 cm"3 (91).
Although silicon has considerable IR absorption below about
660 cm*, our experience shows useful results down to about
550 cm”. Brittleness is one drawback to use of silicon

since the edges of silicon disks tend to chip easily.

6.2.2. Sample Preparation

After exposure to phosphate salts, small IR peaks
corresponding to strong phosphate IR bands were observed on
the silicon disks even after thorough cleaning with both
water and acetone. This indicates that phosphates are
impossible to completely remove from the silicon disk
surface. For this reason, one sample disk was dedicated to
use with samples which contained phosphate species. Another
disk was dedicated to use with samples which contained only
organics.

Aqueous solutions were prepared by diluting the solute
to 3.74 x 10‘6 mol/ml with water. To avoid impurities, HPLC

grade water with less than 3 ppm evaporation residue was

171
used in all experiments. Care was taken to clean all
glassware used in sample preparation very thoroughly of
dust, lint and mineral deposits.

The silicon sample disk was placed in a metal clip-
style sample holder which secures the disk by the edges,
leaving both surfaces open. 25 ul of aqueous solution was
applied to each face of the silicon sample disk in
volumetrically measured amounts using disposable 25 ul
glass pipets.

Vacuum drying was used to remove water from sample
disks. A disk held in a clip-style holder was placed in a
glass vacuum desiccator. Air was evacuated from the
desiccator using a rotary vacuum pump. The progress of water
evaporation was timed and observed. When the moisture was
gone, as determined by visual inspection, air was readmitted
to the desiccator immediately and the sample removed. This
process was then repeated for a second loading on both sides
of the disk. If sample freezing was observed during drying,
the sample was immediately exposed to air for a few seconds
until the ice melted. The desiccator was evacuated again and
drying was resumed. The time under vacuum required to dry a
typical sample loaded with a total of 100 ul of solution
was 10—13 minutes.

After vacuum drying, the sample disk was removed from
the clip-style holder and transferred to a magnetic sample

holder (Spectra Tech) which fits into the spectrometer. The

172
edges of the sample disk were sealed using Teflon tape so
infrared light could not leak between the edges of the disk
and the magnetic sample holder. The magnetic holder was
pressed to fit tightly around the edges of the sample disk.
The disk and holder assembly was then held up to the light
and visually inspected to check for light "leakage" around
the disk edges. The disk and holder assembly was then placed
in a glass jar containing a small vial of desiccant to
remove moisture while being transported to the spectrometer.
Transfer of the sample disk from the vacuum desiccator to
the jar was performed as fast as reasonably possible to
minimize sample exposure to moist ambient air. The time
required for this transfer was typically two to four
minutes.

A method of trial and error was used to determine the
amount and concentration of solutions required to obtain
good results. Too little solute on the sample disk results
in little absorption by the sample and obscures small
spectral features. Excess Solute results in "saturation" or
too much absorption and causes peaks with distorted features
and flat or rounded apexes (as opposed to pointed ones).

The volume of solution applied to the disk face is also
important. Insufficient liquid volume results in incomplete
coverage of the disk face while excess liquid results in

spillage of liquid from the disk face during handling. A

173
solution volume of 25 ul was found to be ideal in our experiments.

Since we were required to load each disk twice, double
the solute concentration should be used in future
experiments to deposit a sufficient amount of solute in only
one loading.

The use of heat to evaporate water from sample disks
was attempted, but did not work well and was abandoned. Heat
tended to "bake on" the phosphates, making them harder to
remove from the disks. Heat also made organics evaporate too
fast. When organics evaporated quickly along with the water,
the reproducibility of amounts deposited during sample

preparation was diminished.

6.2.3. Post-reaction Studies
6.2.3.1. Apparatus
An apparatus was designed to expose prepared sample
disks to a flow of lactic acid vapor at temperatures of 100—
350°C. A diagram of this apparatus is shown in Figure 6.1.
The apparatus consists of a 200 m1 three—neck round
bottom flask equipped with a thermometer in one neck which
extends down into the flask and a dip tube in another neck
for a helium carrier gas inlet. The third middle neck
connects the flask to a 1.0 inch ID vertical column about 6
inches long with three nibs to support a coarse glass frit
halfway up the column. The frit supports the silicon sample

disk during runs. The column is connected to a 75°

174

mononum

 

as“.
882.8
2%:

3.50
80

cowuommucumom “Ow msumnmamc .H.w musmam

         

6292
9.82..
so 8.0m
o_d:ouo§o5 o .2
.252 88..
\
\ .
\ \ I
225 \
.235
a: __ooo
x
x z 8.5 80 “III...
/ , Eu $0.0 J
xmfi 29:...“
835952:
”may” / _ 238855
_ ; ;E o
mafiao :Uo

 

175
distilling head with a quartz thermocouple sheath entering
from the top and extending down to near the frit. The
sidearm is connected to a Liebig condenser with a 1050
vacuum adapter on the lower end. The adapter connects with a
250 ml Erlenmeyer flask. The whole apparatus is made of
glass except for Teflon adapters with rubber O—ring seals
for the thermometer, dip tube, and quartz thermocouple
sheath.

The round bottom flask is heated on the bottom by a
heating mantel, and by heating tape wrapped around the top
and the three necks. Both the mantel and heat tape are
controlled by an Omega PI controller connected to a control
thermocouple inserted between the flask and the mantel. A
thermometer measures the temperature inside the flask
independently. The column and distilling head are heated by
two heating tapes controlled by another Omega PI controller
receiving feedback from the thermocouple near the glass frit
inside the column. Use of two separate heating tapes allows
the column and head to be disassembled for easy access to
the sample.

The exterior plumbing allows helium carrier gas to flow
either direction through the apparatus by turning two three-
way valves. The effluent flows through a bubblemeter to

allow flow measurement.

176
6.2.3.2. Procedure

The apparatus was prepared by first adding about 100 ml
of 85% aqueous lactic acid solution to the round bottom
flask, then driving off most of the water present in the
solution. This involved heating the contents of the flask to
about 140°C under a helium carrier gas flow of 500 ml/min.
The column was heated to 160°C to prevent lactic acid
condensation in the column. Holding these conditions for
about one hour distilled off most of the water in the lactic
acid solution.

After water removal, the column was preheated to the
temperature at which the subsequent run was to be performed
and the apparatus was allowed to come to steady state. At
this point, a sample disk was prepared by vacuum drying as
described in Section 6.2.2. When the sample disk was ready,
the carrier gas flow was turned to the reverse direction (so
that no lactic acid is carried through the column), the
column heat was turned off and the column was cooled by a
fan blowing ambient air across the column.

When the column temperature fell below 100°C, the
distilling head was removed from the column, the sample disk
was placed on the glass frit inside the column and the head
was replaced as quickly as possible. This operation was
performed with reverse helium carrier gas flow of about 500
ml/min. The column was then heated to the required

temperature of the run. Upon equilibration at the proper

177
temperature, the carrier gas flow was turned to the forward
direction to expose the disk to lactic acid and the timer
was started. After exposure to the lactic acid vapor stream
for 10 minutes, the carrier gas flow was reversed to avoid
further exposure to lactic acid, the column was cooled to
less than 100°C and the sample disk removed. The disk was
then loaded into the magnetic sample holder and placed in
the holding jar, as described in Section 6.2.2. The time
required to cool the column from 350°C to less than 100°C is
about 17 minutes. Less time is required for lower
temperatures.

For samples at 25°C, sample preparation involved vacuum
drying of an aqueous solution containing both the phosphate
salt and lactic acid without vapor deposition of lactic acid
in the heated column. In this respect, the preparation of
samples at 25°C and those at all higher temperatures is
different: Lactic acid was vacuum dried at 25°C, while it

was vapor deposited at temperatures 2 100°C.

6.2.4. FTIR Spectroscopy
6.2.4.1. IR Spectrometer

Fourier Transform Infrared spectra were collected on a
Nicolet IR/42 spectrometer with a broadband helicoil globar
source and a deuterated triglycine sulfate (DTGS)
pyroelectric detector with a KBr window. The spectrometer

used a Michelson interferometer with a KBr beamsplitter. The

178
hardware was operated by Nicolet PC/IR Version 3.00 menu-
driven software run on a Dell System 200 personal computer.
Spectra were collected in the mid-IR region of 400-4800 cm"1

with 32 scans and a resolution of 4 cm4.

6.2.4.2. Collection of Spectral Data

Before collecting data on the spectrometer, the
computer was rebooted and the PC/IR program started anew.
This set all parameters to their default values to ensure
consistency between spectra. After purging the sample
compartment with house nitrogen, a background spectrum was
taken with no sample holder in the sample holder frame. The
magnetic sample holder was then loaded with the clean
silicon sample disk. The sample compartment was opened and
the magnetic sample holder was inserted into the sample
holder frame inside the spectrometer. The sample compartment
was closed as soon as possible and purged, before taking a
reference spectrum of the clean sample disk. Reference
spectra were typically used for about five sample spectra
before the baseline drift became large enough to require
taking a new reference spectrum. After taking a reference
spectrum, the sample disk was prepared with a sample and
loaded into the magnetic sample holder, as described above,
then placed in the spectrometer. A sample spectrum was taken
after purging the sample compartment. The spectrometer

sample compartment was purged by house nitrogen for at least

179
three minutes before collection of all background, reference
and sample spectra to assure the absence of water peaks due
to atmospheric moisture.

When a background, reference and sample spectrum were
acquired, the reference spectrum was then subtracted from
the sample spectrum and the baseline of the resulting
spectrum was corrected using a third order method. In a few
cases the baseline had extreme curvature or error. In most
of these cases, the sample was either prepared again after
taking a new reference spectrum, or in some cases the
baseline was corrected using a fourth order method.

The sample, result and reference spectra were all
annotated with information about their preparation and saved
on microdisks. The spectra were plotted in two windows, a
wide window of 400—4800 cm‘1 and a closeup of the

l

fingerprint region of 600—1800 cm".

6.2.4.3. Data Reproducibility

The reproducibility of both the sample preparation and
spectroscopic techniques was confirmed by comparing repeated
runs of samples. The frequencies of the major bands for all

1 resolution of the

repeated runs agreed within the 4 cm"
spectrometer. The ratios of the magnitudes of the major
peaks agreed within 1 23% in the worst case, but typically

within : 10%.

180

6.3. Results

6.3.1. FTIR of Pure Substances

Some notable differences are observed between the
spectra of lactic acid and sodium lactate. Lactic acid shows
a very strong carbonyl stretching band [UC=O] at 1728 cm“1
which shifts to 1591 cm‘1 for sodium lactate. This shift
shows a decrease in the calculated force constant from 12.2
N/cm for lactic acid to 10.3 N/cm for sodium lactate. This
represents a 16% decrease in the C=O bond strength. Lactic
acid has a strong combined stretching and bending band
[UMCO + 6NOH] at 1221 cm'1 which does not appear in the
sodium lactate spectrum. A clearly defined set of six peaks
appears at 1265, 1314, 1364, 1422, 1456, and 1473 cm‘1 for
the lactate salt. Some peaks in this "fingerprint" are
present, but less defined, in the lactic acid spectrum. A

l in the lactic

moderate stretching band [UC—COOH] at 826 cm'
acid spectrum shifts to 855 cm” for sodium lactate. Table
6.1. shows an overview of the spectral features and band
assignments of the lactic acid and sodium lactate spectra.
Lactic acid and sodium lactate spectra are presented in
Figure 6.2. All band assignments are based on the results of
Cassanas, et al. (92).

Certain features are common to all sodium

orthophosphates. Sodium orthophosphates have a strong UaSPO3

band at 1020—1080 cm‘1 and a moderate 6agrp band at

181

ace

 

ecu—

oumuomq ssflcom ccm choc ofluomq wo muuuwdm mH

=35": «baa—55:25;

ecc— ===~ ==e~ ==a~ ==~m

 

:2 as:

 

 

 

 

, 323.. 5.5.5

.N.6 666624

acum

aoueugulsueu

182

Table 6.1. IR Frequencies and Band Assignments of
Lactic Acid and Sodium Lactate

 

Lactic Acid

Sodium Lactate
1

Band Assignments'

 

 

(cnfl) (cnf )
3600—2300 vs,b 3600-2500 vs,b ILO
2992 s 2977 s UCH
2946 s 2934 s UCH
1728 vs 1591 vs UC=O
1458 s 1473 m,sh oasCH3
1458 s 1456 m (SasCH3
1420 m,sh 1422 s PQWCO + 5MOH]
1375 s 1390 m 6%0H”
1375 s 1364 m 63yCH3
1340 m 1314 m BCH
1275 s,sh 1265 m 5MOHH'
1221 s - [umco + 6MOH]
1127 s 1123 s aflco”
1090 m,sh 1088 m 6mco"'
1046 s 1042 m uC-CH3
924 w 925 w rCH3
826 m 855 m UC-COOH
745 w 779 w OOCO
652 w 677 w OC-COH
530 w 530 w UOCO
428 w 415 w omCCO
* based on results of Cassanas, et al.

** intermolecular hydrogen bond

*** solvated

vs = very strong, 3
sh =
sy = symmetric, as

r = rocking

strong,

m : moderate,
shoulder, AC : acid, AL = alcohol,
asymmetric, U = stretching,

weak, b = broad,

6 = bending,

183
470-570 cm‘1 (71). Two strong, broad water bands appear at
2000—3700 cm‘1 and 1600—1680 cm”. Our results agree within
about 10 cm'1 for the frequencies of the major peaks when
compared to literature data (71,93). IR spectra and spectral
features of sodium orthophosphates are presented in Figure
6.3. and Table 6.2.

Common features of condensed sodium phosphates are two
strong, broad water bands at 2000—3700 cm'1 and 1655—1670
cm'1 and a moderate band at 485—550 cm4, presumably due to
bending vibrations (ofiPO3). Frequencies of the major peaks
of sodium pyrophosphate (Napng) and sodium tripolyphosphate
(NasPfixo) agree with literature values within about 10 cm?

(93). Features of some condensed sodium phosphates are shown

in Figure 6.4. and Table 6.3.

6.3.2. FTIR Post-reaction Studies
6.3.2.1. Monosodium Phosphate

Lactic acid was deposited on a sample disk loaded with
monosodium phosphate (NaHfiKh) at temperatures of 25—350“C as
described in Section 6.2. Figure 6.5. shows the spectral
results of post-reaction studies on monosodium phosphate.
Neither sodium lactate (UC=O at 1590 cm*) nor phosphoric
acid is observed, indicating no significant amount of
proton exchange over the temperature range studied. A small,

broad shoulder peak appears at 1593 cm'1 but is attributed

184

monogamocdocuuo Ezflcom mo manomdm mH

.25 2352325

aea— acu—

ace—

cau—

\

cam—

 

 

{I
q
1
1
1|
1

3‘

32.7.2... .555“:

 

 

.m.o muomwm

anuengmsueu

185

98V
[[9
’ 8175 295

E3

 

09L

£68
516

wmumcdmocm anoom oomcmccoo

co muuomdm mH .v.o wusmflm

.ese. ses—

32.32.58?—

 

Swineifioéfi

 

i‘

«zingers—E

 

3.53 mh0=E==Q>Gg
92:. E—N— eav—
u
8
m...
6 Ir Iv
8 z
00 mm M“
w .. l
.V I-
n a
z
’—

BIO
260
59“

anuemmsuml

Suing—:3;

 

186

 

 

 

 

 

 

 

 

Table 6.2. IR Frequencies of Sodium Orthophosphates

Phosphoric Monosodium Disodium Trisodium

Acid Phosphate Phosphate Phosphate

H5PO4 NaHzpo4 NaZHPo4 Na3PO4
(cm4) (cm7) (cm7) (cm*)

1169 s 1279 m 1080 s 1430—1460 m

1019 s 1165 s 990 m 1028 s
880 w 1049 s 843 w 936 w
803 w 947 s 650—696 w 839 w
702 w 858 m 534 m 708 w
471 m 801 w 569 m

704 w
505 m

s = strong, m = moderate, w = weak

Table 6.3. IR Frequencies of Condensed Sodium Phosphates

Sodium Trisodium Sodium Sodium
Pyrophosphate Trimeta- Tripoly- Polyphosphate

phosphate phosphate

Na4PZO7 Na3P309 Na5P3Om (NaPO3)
(cm4) (cm7) (cm4) (cm”)
1123 s‘ 1289 s 1227 s 1265 s
1024 w' 1161 m 1142 s 1154 m
993 w 1100 s 1024 w 1092 s
980 w’ 988 s 974 m 1019 s
915 s' 760 m 895 m 934 m
893 s" 687 w 745 w 876 s
710 w 640 w 702 w 779 w
650 w 517 m 548 m 706 w
552 m' 486 m

* most notable features, ** likely a water peak,

3 : strong, m = moderate, w : weak

187

 

wumcdmozm Esflcomocoz co baud ofluomq wo muuoodm mH .m.m madman
=35": mhu£E==a>flg
E; can can =2: .32 2:; =2: 25—
1 I I I I I I I m I “in" I w I I I I I I I I {M I
M % 8 M I“ 5 mN
5 I. I- Z
0 5

 

 

aauengulsueu

 

188
t:o a broad water peak at about 1650 cm’1 for monosodium
phosphate.

Monosodium phosphate is present at and below 2000C,
indicated by the observed bands 1280, 1165, 1057, 992, 874,
and 706 cm”. Lactic acid leaves the surface before 150°C,
indicated by the peaks 1744, 1458, 1408, 1377, 1219, 1132,

1096, 1049 and 837 cm“.

Sodium polyphosphate, (NaPOfl forms at 250-350°C,

indicated by peaks at 1242—1273, 1165, 1092, 1046, 938,
856—885 and 706 cm7. These peaks become stronger and more
well—defined as the temperature increases, which possibly

indicates an increasing degree of condensation of the

phosphate as the temperature increases.

6.3.2.2. Disodium Phosphate

Infrared spectra of post—reaction studies over
disodium phosphate are presented in Figure 6.6. Lactic acid
converts partly to sodium lactate at 25—100°C. The
respective carbonyl bands appear at 1740 and 1609 ch,
indicating that both sodium lactate and lactic acid are
present. Lactic acid peaks include 1740, 1456, 1414, 1372,
1219, 1134, 1090 and 1047 ch. Peaks attributed to sodium
lactate are 1609, 1456, 1414, 1372, 1134, 1090, 1047, 862
and 530 ch. In the acid-base exchange, disodium phosphate
is partly converted to monosodium phosphate. Observed

disodium phosphate peaks include 988, 835 and 530 cm'1

189

mumsmmonm Esficomflo co cao< ofluomq wo muuowam mH

 

 

 

 

.o.o apnoea
:35": mamas—525;
ace can can cac— camp aee— acc— can—
I I I I 0 I I I U I I I U I” L u 0 III! I I U I I I U I III I
0 0 Ir
5 m I 8 u R w. w. m.
l I_. I
an m“ .u L |.na nu It Mu cu mw
IL Ibnu .l I» - no
c. an nunu . no mun“ nu
a v . I 9 e
nun. as M” av nu
.v '8! B n S
11.. 9 a: m.
9 6 6 m... . . w
.7 cc 0 8 z a
an cu nu nonb amN
I; can
1 .. {and
Z
I. E
I

 

 

190
Vuhile monosodium phosphate peaks include 1171, 1047, 943,
882 and 862 cm”.

Both lactic acid and sodium lactate leave the surface
eat 150°C and above, indicated by the disappearance of both
carbonyl bands at 1740 and 1609 cm*.

All the phosphates convert to an unidentified species
which we believe to be an intermediate to pyrophosphate
formation at 150-2000C. This unidentified species has IR
bands at 1395, 1358, 1142, 1057, 995—945, 858, 739, 608 and
542 cm4. These bands do not correspond to any
orthophosphates or condensed phosphates which we studied. At
250°C, about 78% of the unknown phosphate has converted to
sodium pyrophosphate (Na4ng), and at BOO—350°C, all
unidentified phosphate has converted to pyrophosphate.
Pyrophosphate IR bands observed were 1146, 1123, 1030, 988,
899, 733, 552 and 490 tm*.

6.3.2.3. Trisodium Phosphate

Spectra from post-reaction studies of lactic acid on
trisodium phosphate are seen in Figure 6.7. These spectra
reveal that lactic acid forms sodium lactate and that either
lactic acid or sodium lactate is present at 25-350°C. At
room temperature, lactic acid and trisodium phosphate
undergo a virtually complete proton exchange to form sodium
lactate and disodium phosphate. As the temperature

increases, the percentage of sodium lactate changes on the

191

mumcdmocm Esaoomfiue co cflo< ofluomq mo manomdm mH .r.m madman

:35": «mania—Saw;

 

   

 

93‘ 2.5 can :2: :92. 23— cam— 2::
I I I I 0 I I I U I I I U I LlI ” I I I n I I I u I I I
l
mmz
n: no MW canUIL
a g o u l I!
I. 5 7.. I.
l.- mN I.
I ll I.
D.
- .. w
is s m
cm n
B
U
a. m
n: a
6
an _u

192
surface, as seen in Table 6.4. The IR bands due to sodium
lactate are 1597, 1475, 1460, 1424, 1364, 1316, 1271, 1121,
1050, 853, 779 and 679 cm”. The disodium phosphate bands
observed are 1090, 990, 845 and 650—696 cm“. As the
temperature increases to ZOO-350°C, disodium phosphate
condenses to sodium pyrophosphate, identified by the peaks

at 1127, 1034, 1000, 988 and 934 cm4.

Table 6.4. Percentage of Sodium Lactate on Trisodium

 

 

Phosphate
Temperature (%D Lactate in Sodium Salt Form (%)

25 100
100 65
150 74
200 85
250 88
300 90
350 83

 

6.4. Discussion

6.4.1. Reactions on Catalyst Surfaces

The post—reaction studies show that the more basic is
the phosphate salt loaded onto the surface, the more sodium
lactate and lactic acid is retained on the surface,
especially at elevated temperatures above 150°C. The overall

reaction scheme pertaining to lactic acid—sodium phosphate

193
proton exchanges and the subsequent phosphorus-to-phosphorus

condensations is outlined in Figure 6.8.

W
(gone 2150°C) (no rxn.)
Lactic Acid + lfifiLPO4 -4 Sodium Lactate + HBPO4
(2250°C)
NaHZPO4 —+ (NaPO3),1
. 1' 1 1
(gone 2150°C) (<150°C)
Lactic Acid + Nagfixh —4 Sodium Lactate + NaHfiKh
(150-200°C) (2250°C)
NaZHPO4 —+ Unidentified Phosphate —+ Nafi§O7
(23331: flfgi sodium PhQsphaf‘e
Lactic Acid + NaflXL —4 Sodium Lactate + Nafifixh
(2200°C)
Na2HPO4 —> Na4PzO7

Figure 6.8. Proton Exchange and Condensation Reaction
Schemes

6.4.2. Application of Findings to Results of Lactic Acid
Conversion

The findings of this study may explain some of the
trends observed in conversion of lactic acid over sodium
orthophosphate catalysts. We observed that monosodium
phosphate condenses to sodium polyphosphate at reaction
conditions above 250°C, while both disodium and trisodium

phosphates form sodium pyrophosphate above about 250°C. This

194
explains why selectivities to products are similar for
disodium and trisodium phosphate catalysts since the actual
catalyst is the same.

To explain the increased conversion over trisodium
phosphate when compared to disodium phosphate, we employ the
following scenario. Excess sodium ions are present on the
surface after condensation of trisodium phosphate to
pyrophosphate. These ions are present either in the form of
sodium lactate or in the form of sodium ions passivated on
the surface. The sodium ions shift the equilibrium of lactic
acid between the gas phase and the surface. This shift in
equilibrium raises the concentration of both sodium lactate
and lactic acid on the surface which increases the
conversion.

In contrast, few excess sodium ions are left on the
surface after condensation of disodium phosphate to
pyrophosphate since there are stoichiometric amounts of
sodium ions and phosphate ions present. The few sodium ions
do not shift the equilibrium between the gas phase and the
surface, resulting in lower lactic acid and sodium lactate

concentrations on the surface and thus lower conversion.

6.4.3. Equilibrium of Proton Exchange
The equilibrium constants for the first proton exchange
.between lactic acid and the various phosphate species

present at reaction conditions were calculated from acid

195
dissociation constants obtained from literature (84,86). The
only consistent data available was for aqueous solutions at
near room temperature, which makes application of these
results to surface reactions at reaction conditions (280—
350°C) only qualitative, at best.

The equilibrium constants for proton exchange between
lactic acid and sodium polyphosphate and also between lactic
acid and sodium pyrophosphate are 2 x 104 and 2 x 103
respectively. The equilibrium for formation of sodium
lactate is favorable in both cases, but is an order of
magnitude greater for pyrophosphate than for polyphosphate.
Since lactic acid is much more volatile than sodium lactate,
the greater formation of sodium lactate over pyrophosphate
indicates more sodium lactate would be present on the
surface of a pyrophosphate catalyst. In other words, the
pyrophosphate has a greater ability to "pull down" lactic
acid onto the catalyst surface (in the form of sodium
lactate) than does polyphosphate. This supports the results
of our catalytic studies where higher conversion is observed
over disodium phosphate, which forms pyrophosphate at
reaction conditions, than over monosodium phosphate, which
forms polyphosphate. We conclude that the equilibrium
between the gas phase reactant, lactic acid, and the surface
phosphate species affects the conversion by altering the

concentration of the reactant on the surface.

Chapter 7. Summary and Recommendations

7.1. Summary

Catalytic conversion of lactic acid over zeolites in a
batch reactor produced mostly acetaldehyde and ethanol at
30—42% selectivities as well as lesser amounts of propanoic
acid, acetic acid and acrylic acid. Conversions based on
lactic acid fed were 3.7—5.3% at 240°C.

Lactic acid conversion over supported and neat sodium
phosphate catalysts was performed in a downflow differential
packed bed reactor. Five major reaction pathways were
observed: 1) decarboxylation to form acetaldehyde, CIb and
Pg, 2) decarbonylation to form acetaldehyde, CO and HgL 3)
dehydration to form acrylic acid, 4) direct reduction to
propanoic acid and 5) condensation to 2,3-pentanedione.
Minor products included ethanol, acetic acid, acetone and
acetol. The condensation of lactic acid to 2,3—pentanedione
was first observed in our laboratory.

Production of acrylic acid was favored at high
temperatures of 350°C and short residence times of 0.4
seconds, while 2,3—pentanedione formation is favored at
280°C and long residence times of 4.0 seconds. Increasing
the contact time leads to an increase in decarbonylation to
acetaldehyde. Decarboxylation is not strongly affected by
contact time. Basic phosphate salts favor acrylic acid and

2,3—pentanedione, while more acidic salts produced greater

196

197
amounts of the undesired products propanoic acid,
acetaldehyde, acetic acid, acetone and acetol.

A catalyst survey was performed to compare various
supported sodium salt catalysts with different central atoms
but with oxygen ligands similar to sodium phosphate. The
sodium salts studied include tetraborate, carbonate,
nitrate, silicate, phosphate, sulfate, chlorate, arsenate,
bromate and molybdate. Catalyst salt acidity affects product
distribution with basic salts producing higher selectivities
to 2,3-pentanedione, relatively acidic salts producing more
propanoic acid and moderately basic salts producing more
acetaldehyde. Group V salts (nitrate, phosphate and
arsenate) exhibit 2,3—pentanedione selectivities up to 64%.
Conversion increased strongly with increasing level of
occupied molecular orbitals in the central atom of the
catalyst salt.

Chemical species present on supported sodium phosphate
salt surfaces were determined using MAS—“P-NMR. Sodium
phosphate salts undergo complex dehydration reactions upon
heating to reaction temperatures, so a mixture of condensed
sodium phosphate salts was present on the surface of the
supported catalysts and was readily analyzed using NMR.

Adsorbed chemical species on catalyst salt surfaces
were studied using transmission FTIR after exposure to
lactic acid at reaction temperatures. The more basic sodium

phosphate salts affect greater conversion by altering the

198
equilibrium between the gas phase reactants and the surface
reactants. A greater surface concentration of lactic acid
and sodium lactate is observed over the more basic salts
leading to greater conversion and better product
selectivity.

Using FTIR and NMR, we show that catalyst salt acidity
changes the distribution of chemical species present on a
model catalyst surface at reaction conditions. The size of
electronic cloud of the central atom of the catalyst salt
has a strong effect on the catalyst activity. These
considerations were employed to propose reaction mechanisms
to acrylic acid and 2,3—pentanedione which both involve the

formation of cyclic transition states.

7.2. Recommendations

Our FTIR results suggested that excess sodium ions on a
catalyst surface will increase the concentration of
reactants on a catalyst surface and produce higher
conversions in reactions over real catalysts. Studies may be
performed to determine if the concentration of sodium ions
on the surface of a catalyst may be manipulated. This may be
achieved perhaps by addition of NaOH in increasing
concentrations along with a constant phosphate catalyst
concentration. From our observations in this work, one would
expect this increase in concentration may be correlated with

conversion and selectivity over a supported salt catalyst.

199
One may also study supported sodium ions by themselves to
determine their catalytic activity and selectivity.

In light of the fact that we know disodium and
trisodium phosphate to form pyrophosphate upon heating to
reaction conditions, another interesting study may be to
synthesize a supported catalyst made of pyrophosphate to
simulate condensed disodium phosphate. To simulate trisodium
phosphate, equimolar amounts of pyrophosphate and NaOH may
be used. A comparison of the performance of these two
catalysts may prove an interesting comparison with disodium
and trisodium phosphate catalysts.

The catalyst survey results indicate that increasing
the number of occupied molecular orbitals in the central
atom of the catalyst salt (alternatively, the "size" of the
central atom) leads to an increase in the conversion over
the salt. We also know Group V salts produce greater
selectivities to desirable products. These observations lead
to the suggestion that the salts of the "larger" Group V
elements antimony and bismuth may prove interesting
catalysts. Future studies may determine the catalytic
properties of sodium pyroantimonate, Nafingpr6HfD, or
sodium metaantimonite, NaSbOf3HgD, and sodium
metabismuthate, NaBiO3.

The possibilities of using NMR to make mechanistic
determinations are great. Our use of NMR to determine the

chemical species present on supported catalyst surfaces only

200

scratches the surface of the utility of NMR. NMR observation
is not obscured by high temperature emissions as FTIR has a
tendency to be, so NMR may be used to examine reactions in—
situ at reaction temperatures. NMR may by used to observe
the mechanisms of the reactions by several different, but
complimentary approaches. 1H of the lactic acid reactant on ’
a catalyst surface may be observed perhaps by using a D4)
solvent or by evaporation of HgDZbefore observation. 13C E
labeled positions of the lactic acid reactant may be H
observed on a catalyst surface. Lactic—l-HC acid, sodium
salt, is available from Aldrich Chemical Company. The
central atoms of many catalyst salts may be observed using
NMR. The group V elements L%L 31P, 75As, 121Sb, 123Sb and 209Bi
all have substantial natural abundances and may be observed
using NMR (84). The feasibility of getting good results
using available equipment and methods is unknown, but the
possibilities exist.

FTIR studies of reactions in-situ may provide
information not available by post-reaction experiments in
which the samples have been cooled prior to observation. In-
situ reactions may be studied using an infrared
transmittance cell in which gas phase reactants are passed
over a thin IR-transparent catalyst disk which is observed
using FTIR as the reaction proceeds. Several examples and
designs for this type of reactor are presented in literature

(73-81).

201

In—situ reactions may also be studied using Diffuse
Reflectance Infrared Fourier Transform spectroscopy (DRIFT).
This method analyses infrared radiation reflected from a
solid sample and does not require tedious sample or catalyst
disk preparation, is not sensitive to sample thickness, may
be used to examine non-transmitting samples, and may be used
to observe samples in—situ. Spectra Tech markets a diffuse
reflectance cell called the Collector which may be
configured as an in—situ reactor cell with a temperature
range up to 900°C and a pressure range from 10'5 Torr to 1500
psig (91). Examples of the applications of diffuse

reflectance may be found in literature (82-83).

APPENDICES

Appendix A Calculation of Macropore Size in Y Zeolites

Mercury porosimetry was performed on Y—82 and Y—62
zeolite samples using a Micromeritics Pore Sizer Model 9310.
After mercury intrusion into the catalyst samples, pressure
data as a function of intrusion volume was used by dedicated
software to produce a plot of incremental pore volume per
gram of catalyst as a function of pore diameter, as seen in
Figures A.1. and A.2. Incremental pore volumes are tabulated
in Tables A.1. and A.2.

The weighted mean pore diameter was calculated from the

pore size distribution for each catalyst as follows:

n

23vmdi
121
Q :
n
X wi
121
where d = weighted mean pore diameter
vq 2 weight of data point i
(L 2 pore diameter at data point i
n = total number of data points in consideration

202

203

P1]ZC:FQC3P4EEFQ.I'T'IICZES FDCDFQEE
01.04
Y-82 Zeolite
GCG

PENETROMETER NUMBER 161

INCREMENTAL PORE VOLUME PER GRAN 2 vs.

Maximum Incremental Volume

7

20! a0! 00! 501 601
con-o! ooooooooo I ooooooooo 8 ooooooooo to... ..... It

01 101
IOOOIOOOOOIOOI.

233 -ttlttl!

183 '33830833188

143 -I!!!8!!!I!888!!
112 '888388883388838818
87.7 '888888881338388
68.6 -!ll!!t!
53.6 '83!
02.0 '88

5115

(at)

0.392 '3888

0.307 -l!3!!
0.240 '8188!!!

0.188 '3!litttlltttttltlttlttlIt!8!!!3311!!!!8333lillltltlltlltlllt3!
0.147 '833838it!I!81I138818818888881188!!!I!{liltllttltlti33188888!!!
0.115 -II3188!!!l!!!!l!88!!!l!881183!ll38818181320183!!!3381388131!!!
0.0877 -II!III!!I!3ltlttltlltlltlltttt18881181338!!!818188831188818
0.070: -llt!t!!!!l!llt!8!3!l18813883!!!
0.054? -!ll!!l!lt!

0.04:? '31313188!!!

0.033 'lltl!!!

0.0262 -!!tt

0.0205 '11!!!

0.0160 -t!

0.0125 .333

0.00981 -|ll!
0.00767 '81!!!

0.00600 ‘I!!!I!l!l!l

I ......... 1 ......... I ......... I ......... I ......... z ......... L
02 10: :oz 302 $01 so: so:
Figure A.1. Pore Size Distribution in

ESJEZEEZF? C?ZS:|<3
PAGE 6
7/14/90
19:00-

PORE DIAMETER (micrometers)
0.0800 cc/g

701 801 901 1001
........ ‘00....OOOIOOOOIOOIOIIIOIOOICCI
°N0
'23
'1“
'1“

0
u—
.—
N

O
:1

.....~uaaseuz
C O O C C O O O
“GINO-NOOO-o- N

C O

38%3322333

“0‘
OJ—
NM

°°°°~——NNH&MN-fl-HN
. o l O 3

13188188888811!
8883!K8188888333888l!3381883888!
88!8313888883188!!!3188881813!!!

Y-82 Zeolite

204

P1JZC3FQCJPIEEFQZE‘FIECZES F’CJFQEE
V1.04
Y-62 Zeolite
GCG

PENETROMETER NUMBER 161

INCREMENTAL PURE VOLUME PER GRAN Z vs.
Maximum Incremental Volume =

SSJIIEEEF? ERCSILCD
PAGE 6
7/14/90
23:40

PORE DIAMETER
0.0946 cc/g

(micrometers)

01 101 201 301 101 501 601 701 801 901 1001
I... ...... 1.666 ..... 1 ooooo .6661 ooooooooo 1 ......... 1 ooooooooo Ioooooooool oooooooo oloooqoooonloaoooooool
300 ' ° 300
233 ' ' 233
183 '11 ' 183
143 '1111 ‘ 163
112 ‘111111 ' 112
37.7 -!t!ttl ‘ - 37.7
68.6 '11111 ‘ 68.6
5306 -0000 . 330‘
020° ‘00 ' 026°
32.8 ‘11 ' 32.8
2301 -00 - 2567
20.1 ‘11 ‘ 20.1
13.7 ’1 ' 13.7
12.3 '1 ' 12.3
9.39 '1 ' 9. 9
7.30 '1 ° 7.50
5.87 '1 c 5.87
1.59 '1 - 1.39
3.59 ’11 ' 3.39
2009 ‘000 - 201'
5 1 1 E 1.72 “111 ' 1.72
1.34 ‘1111 ‘ 1.31
1 0| 1 1.03 ‘1111 ‘ 1.03
0.820 ‘1111 ° 0.820
0.642 ‘111111 ' 0.642
0.502 “11111111 ‘ 0.302
0.392 '11111111111 ‘ 0.392
0.307 '11111111111111111111111111111111111111 ° 0.307
0.200 '1111111111111111111111111111111111111111111111111111111111111111111111111111 ‘ 0.210
0.100 '1111111111111111111111111111111111111111111111111111111118111111111111111111111111111111111111111 ° 0.108
0.107 '11111111111111111111111111111111111111111111111111111111111111111111 ' 0.117
0.113 ‘111111111111111111111111111111111111 ' 0.11
0.0897 -:::::txtxz:Ixttxxzztxtxts - 0.3397
0.0702 -txxxzxxtxxtxxxx _ a 97 2
0.0309 '111111 ' 0.0219
0.01:? -1313311 - 0.042°
0.033 '111111 ' 0.0330
0.0262 ‘11111 ° 0.025:
0.0203 '1111 - 0.0203
0.0160 '11 - 0.0160
0.0123 '111 ' 0.0122
0.00981 '111 ° 0.00? 1
0.00767 ‘11111 ' 0.00767
0.00600 '1111111111 ‘ 0 30600
.... ..... 1 ......... 1 ......... 1 ......... 1 ......... 1 ......... 1 ......... 1 .................. 1 ..........
01 101 201 301 401 501 601 701 8 1 901 100?

Figure A.2. Pore Size Distribution

in Y-62 Zeolite

205

Table A.1. Incremental Pore Volumes Greater than 20%
for Y-82 Zeolite

 

 

 

 

i (% (um) w.1 ng (um)
1 0.188 0.76 0.143

2 0.147 0.93 0.137

3 0.115 0.93 0.107

4 0.0897 0.58 0.0520

5 0.0702 0.31 0.0218
23 3.51 0.4608
d = 0.4608/3.51 = 0.131 um = 1310 A

Table A.2. Incremental Pore Volumes Greater than 10%
for Y-62 Zeolite

 

 

 

i di (um) wi widi (mm)

1 0.392 0.11 0.0431

2 0.307 0.38 0.1167

3 0.240 0.76 0.1824

4 0.188 0.96 0.1805

5 0.147 0.68 0.1000

6 0.115 0.35 0.0403

7 0.0897 0.25 0.0224

8 0.0702 0.15 0.0105
3.64 0.6959

 

XI
d

= 0.6959/3.64 = 0.191 um = 1910 A

Appendix B Calculation of Chemical Shift Anisotropy from
Spinning Sideband Intensities

Using Herzfeld and Berger method (88).

Calculate ratios of the intensities of each of the nth
order sidebands to the isotropic peak intensity, In/Lr
Using the graphs in the paper by Herzfeld and Berger, plot
the contour of each of the intensity ratios. Find the point
where the contours intersect and note the values for u and

p. Using u and p, solve for the principal values of the

chemical shift tensor, 0”, on and On, as detailed below.

Calculation of Magnetic Field Strength, Ho

 

1% = hU/gfln

H0 2 (6.62618x10'“ Js) (161.903x10gs‘111107 ergZJ)
(2.261) (5.0505x10'“ erg/G) (104 G/T)

H = 9.395 T

Calculation of Principal Values of the Chemical Shift Tensor

(o33 — on) : ZHDJLflfig a a
-C511 + 0022 + 033 = a (eqn. 1)

O 2 (011 '1' G33 ' 20221/1033 - G11) : (011 + 033 ’ 2022)/a
Oll - 2022 + 033 : an a b (eqn. 2)

-6 = 03‘: (011 + 022 + on)/3

011 + 022 + O33 = -35 a c (eqn. 3)

3 Equations and Unknowns

o11 — 2022 + 033 b

3
—o11 + 0022 + 033 = a

206

207

Solve for on, 022 and 033

—l O
l —2

l l

where

1 011 a
1 022 = b
1 033 c

 

Q = b

= magnetic field strength, T
Planck constant = 6.62618x10'34 Js
observe frequency = 161.903 MHz
nuclear g—value = 2.261 for 31P, (ref. 84)
nuclear magneton = 5.0505><10‘24 erg/G
= magnetogyric ratio = 10.8394x107 rad/sT
= isotropic chemical shift
= principal values of chemical shift tensor
2 sample spinning frequency, Hz

0
II II

H

2

P'

1111

r

w ooo<tocsm

p..-

-1/2 1/6 1/3
0 -l/3 1/3
1/2 1/6 1/3

LIST OF REFERENCES

10.

11.

12.

LIST OF REFERENCES

McKetta, J.J., Ed., "Encyclopedia of Chemical
Processing and Design", 28, 1-12, Marcel Dekker,
N.Y.(1988).

Lipinsky, E.S., R.G. Sinclair, Chem. Engr. Prog.,
82(8), 26, Aug. (1986).

Simpson, F.J., "McGraw Hill Encyclopedia of Science and
Technology", 5th Ed., 7, 513-5, McGraw Hill Pub. Co.,
N.Y. (1982).

Mark, H.F., Ed., "Kirk-Othmer Concise Encyclopedia of
Chemical Technology", 630-1, John Wiley and Sons, N.Y.
(1985).

Kirk, R.E., D.F. Othmer, Eds., "Encyclopedia of
Chemical Technology", 8, 167-80, Interscience
Encyclopedia, N.Y. (1952).

Holmen, R.E., (Minnesota Mining and Mfg.), USP
2,859,240, Nov 4 (1958).

Odell, B., G. Earlam, D.J. Cole-Hamilton, J.
Organometallic Chem., 290, 241-8 (1985).

Paperizos, C., W.G. Shaw, S.R. Dolhyl, (Standard Oil
Co.), EP 0,181,718 (1986).

Paperizos, C., S.R. Dolhyl, W.G. Shaw, (Standard Oil
Co.), USP 4,786,756 (1988).

Sawicki, R.A., (Texaco Inc.), USP 4,729,978 (1988).

Mok, W.S., M.J. Antal, Jr., M. Jones, Jr., J. Org.
Chem., 54, 4596-602 (1989).

Lira, C.T., P.J. McCrackin, Ind. Eng. Chem. Res.,
32(11), 2608-13 (1993).

208

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

209

Mark, H.F., N.M. Bikales, C.G. Overberger, G. Menges,
J.I. Kroschwitz, Eds., "Encyclopedia of Polymer Science
and Engineering", 1, 211-234, John Wiley and Sons, N.Y.
(1985).

Szmant, H.H., "Organic Building Blocks of the Chemical
Industry", 297-9, John Wiley and Sons, N.Y. (1989).

Heylin, M., Ed., Chem. Engr. News, 69(25), 8, Jun 24
(1991).

Heylin, M., Ed., Chem. Engr. News, 69(25), 34, Jun 24
(1991).

Mark, H.F., Ed., "Kirk-Othmer Concise Encyclopedia of
Chemical Technology", 24-6, John Wiley and Sons, N.Y.
(1985).

Burns, R., D.T. Jones, P.D. Ritchie, J. Chem. Soc.,
400-6 (1935).

Burns, R., D.T. Jones, P.D. Ritchie, J. Chem. Soc.,
714-7 (1935).

Ritchie, P.D., J. Chem. Soc., 1054-61 (1935).

Smith, L.T., C.H. Fisher, W.P. Ratchford, M.L. Fein,
Ind. Engr. Chem., 34(4), 473-9 (1942).

Fisher, C.H., C.E. Rehberg, L.T. Smith, J. Am. Chem.
Soc., 65, 763-7 (1943).

Fein, M.L., W.P. Ratchford, C.H. Fisher, J. Am. Chem.
Soc., 66, 1201-3 (1944).

Filachione, E.M., J.H. Lengel, C.H. Fisher, J. Am.
Chem. Soc., 66, 494-6 (1944).

Fisher, C.H., W.P. Ratchford, L.T. Smith, Ind. Eng.
Chem., 36(3) 229-34 (1944).

Ratchford, W.P., C.E. Rehberg, C.H. Fisher, J. Am.
Chem. Soc., 66, 1864—6 (1944).

Rehberg, C.E., C.H. Fisher, J. Am. Chem. Soc., 66,
1203-7 (1944).

Ratchford, W.P., C.H. Fisher, Ind. Engr. Chem., 37(4),
382-7 (1945).

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

44.

210

Rehburg, C.E., C.H. Fisher, J. Am. Chem. Soc., 67, 56-7
(1945).

Filachione, E.M., J.H. Lengel, C.H. Fisher, J. Am.
Chem. Soc., 68, 330-3 (1946).

White, T., J. Chem. Soc., 238-9 (1943).

Filachione, E.M., M.L. Fein, J.H. Lengel, C.H. Fisher,
J. Am. Chem. Soc., 70, 526-9 (1948).

Abe, T., S. Hieda, (Mitsubishi Gas Chemical Co., Inc.),
EP 379,691 (Aug. 1990).

Watkins, W.C., (Eastman Kodak Co.), USP 3,855,279
(1974).

Watkins, W.C., (Eastman Kodak Co.), USP 3,917,673
(1975).

Daniel, C., P.L. BruskY. (Ashland Oil Inc.), USP
4,299,980 (1981).

Daniel, C., (Ashland Oil'Inc.), USP 4,410,729 (1983).

Ruszala, F.A., (Ashland Oil Inc.), USP 4,410,727
(1983).

Moreschini, L., L. Dalloro, E. Cavaterra, G. Petrini,
R. Covin, (Montedison, S.p.A.), USP 4,499,302 (1985).

Haeberle, T., G. Emig, Chem. Eng. Technol., 11, 392-402
(1988).

Solomons, T.W.G, "Organic Chemistry", 3rd Ed., John
Wiley & Sons, N.Y. (1984).

McKetta, J.J., Jr., D.F. Othmer, H.F. Mark, Eds.,
"Kirk-Othmer Encyclopedia of Chemical Technology", 2nd
Ed., 1, 531-41, Interscience Pubs., Div. of John Wiley
6 Sons, N.Y. (1963).

Guisnet, M., G. Perot, "Zeolites: Science &
Technology", ed. F. Ram oa Ribeiro, p.397-420, NATO
Advanced Study Institute on Zeolites-Science and
Technology, Alcabideche, Portugal (1983).

Kerr, G.T., Catal. Rev.-Sci. Eng., 23(1,2), 281-91
(1981).

45.

46.

47.

48.

49.

50.

51.

52.

53.

54.

55.

56.

57.

58.

59.

211

Bhatia, S., "Zeolite Catalysis: Principles and
Applications", 239-44, CRC Press, Boca Raton, Florida,
(1990).

Haynes, H.W., Jr., Catal. Rev.-Sci. Eng., 17(2),
273-336 (1978).

Jacobs, P.A., K.U. Leuven, Catal. Rev.-Sci. Eng.,
24(3), 415-40 (1982).

Anderson, J.R., K. Foger, T. Mole, R.A. Rajadhyaksha,
J.V. Sanders, J. Catal., 58, 114-30 (1979).

Turkevich, J., Catal. Rev., 1, 24-5 (1968).

Minachev, Kh.M., Ya.I. Isakov, ed. W.M. Meier, J.B.
Uytterhoeven, Advances in Chemistry Ser. 121, 451-60,
ACS, Wash., D.C. (1973).

Mirodatos, C., P. Pichat, D. Barthomeuf, J. Phys.
Chem., 80(12), 1335—42 (1976).

Mirodatos, C., A.A. Kais, J.C. Vedrine, P. Pichat, D.
Barthomeuf, J. Phys. Chem., 80(21), 2366-71 (1976).

Cursetji, R.M., A.N. Singh, A.V. Deo, Chemical Age of
India, 37(6), 407—10 (1986).

Dao, L.H., R.L. Mao; D. Klass, Ed., Energy From Biomass
and Wastes XI., 11, 877-91, Inst. of Gas Tech.,
Chicago, (1988).

Mao, R.L.V., T.M. Nguyen, G.P. McLaughlin, Appl.
Catal., 48, 265—77 (1989).

Mark, H.F., D.F. Othmer, C.G. Overberger, G.T. Seaborg,
Eds., "Kirk-Othmer Encyclopedia of Chemical
Technology", 3rd Ed., 13, 894-940, John Wiley and Sons,
N.Y. (1981).

Rodd, E.H., Ed., "Chemistry of Carbon Compounds", 1(A),
719-20, Elsevier Pub. Co., N.Y. (1951).

Greenfield, 5., M. Clift, "Analytical Chemistry of the
Condensed Phosphates”, Pergamon Press, N.Y. (1975).

Mark, H.F., J.J. McKetta, Jr., D.F. Othmer, Eds.,
"Kirk-Othmer Encyclopedia of Chemical Technology", 2nd
Ed., 15, 232-76, Interscience Pubs., Div. of John Wiley
& Sons, N.Y. (1968).

60.

61.

62.

63.

64.

65.

66.

67.

68.

69.

70.

71.

72.

73.

74.

75.

76.

212

Moffat, J.B., Catal. Rev.-Sci. Engr., 18(2), 199-258
(1978).

Andrew, E.R., D.J. Bryant, E.M. Cashell, B.A. Dunell,
Chem. Phys. Lett., 77(3), 614 (1981).

Burlinson, N.E., B.A. Dunell, J.A. Ripmeester, J. Magn.
Reson., 67, 217 (1986).

Duncan, T.M., D.C. Douglass, Chem. Phys., 87, 339
(1984).

Hayashi, S., K. Hayamizu, Bull. Chem. Soc. Jpn., 62,
3061 (1989).

Rothwell, W.P., J.S. Waugh, J.P. Yesinowski, J. Am.
Chem. Soc., 102(8), 2637 (1980).

Prabhakar, S., J.K. Rao, C.N.R. Rao, Chem. Phys. Lett.,
139(1), 96 (1987).

Turner, G.L., K.A. Smith, R.J. Kirkpatrick, E.
Oldfield, J. Magn. Res., 70, 408 (1986).

Un, S., M.P. Klein, J. Am. Chem. Soc., 111, 5119
(1989).

Mudrakavoskii, I.L., V.P. Shmachkova, N.S. Kotsarenko,
V.M. Mastikhin, J. Phys. Chem. Solids, 47, 335 (1986).

Haubenreisser, U., G. Scheler, A.-R. Grimmer, Z. Anorg.
Alleg. Chem., 532, 157 (1986).

Ramakrishnan, V., G. Aruldhas, J. Raman. Spectrosc.,
18, 147—9 (1987).

Rao, K.M., R. Gobetto, A. Iannibello, A. Zecchina, J.
Catal., 119, 512-6 (1989).

Eischens, R.P., W.A. Pliskin, S.A. Francis, J. Chem.
Phys., 22, 1786 (1954).

Eischens, R.P., S.A. Francis, W.A. Pliskin, J. Phys.
Chem., 60, 194-201 (1956).

Eischens, R.P., W.A. Pliskin, Adv. Catal., 10(1), 1-56
(1958).

Penninger, J.M.L., J. Catal., 56, 287-9 (1979).

77.

78.

79.

80.

81.

82.

83.

84.

85.

86.

87.

88.

£39.

9C).

9]..

922.

213

Rebensdorf, 8., R. Larsson, Z. Anorg. Allg. Chem., 453,
127-38 (1979).

Rebensdorf, B., T. Lindblad, S.L.T. Andersson, J.
Catal., 128, 293-302 (1991).

Rebensdorf, 8., T. Lindblad, J. Catal., 128, 303-10
(1991).

Hicks, R.F., C.S. Kellner, B.J. Savatsky, J. Catal.,
71, 216-8 (1981).

Worly, S.D., J.P. Wey, W.G. Neely, Surf. Sci. Catal.,
ACS Symp. Ser. #482, D.J. Dwyer, F.M. Hoffman, Eds.,
ACS, Wash., D.C. (1992).

Hamadeh, T.M., D. King, P.R. Griffiths, J. Catal., 88,
264-72 (1984).

Van Every, K.W., P.R. Griffiths, Appl. Spectrosc.,
45(3), 347-59 (1991).

Weast, R.C., Ed. "CRC Handbook of Chemistry and
Physics", 662 ed., CRC Press, Boca Raton, Florida
(1985).

Shugar, G.J., R.A Shugar, L. Bauman, "Chemical
Technician's Ready Reference Handbook", McGraw Hill,

N.Y. (1975).

Perrin, D.D., "Ionisation Constants of Inorganic Acids
and Bases in Aqueous Solution", Pergamon Press, Oxford
(1982).

Nemeth, B., "Chemical Tables", Halsted Press, John
Wiley and Sons, N.Y. (1975).

Herzfeld, J., A.E. Berger, J. Chem. Phys., 73(12), 6021
(1980).

Martin, M.L., G.L. Martin, J.-J. Delpuech, "Practical
NMR Spectroscopy", 201, Heyden, Philadelphia (1980).

Emsley, J., D. Hall, "The Chemistry of Phosphorus",
Wiley, N.Y. (1976).

"The Complete Guide to FT-IR", Spectra Tech Catalog,
Stamford, Conn. (1993).

Cassanas, G., M. Morssli, E. Fabregue and L. Bardet, J.
Raman Specrocs., 22, 409-13 (1991).

214

93. "The Infrared Spectra Handbook of Inorganic Compounds",
Stadler Research Labs, Philadelphia (1984).