.3. . a
.. :...: 1...
1.... . 1.21...
hi... , .

:......i
._ .1 .1

.... .. w.

 

:... 5 .
; 3,

 

 
 
 

1 1.3....npi1 .
:1..... 1

 
 
 
   

 

 

 

PLACE IN RETURN Box to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 cJCIRC/DdoDuo.p66-p.15

 

 

MECHANISTIC STUDIES OF CATALYTIC STOBBE CONDENSATION
By

Ancuta Cemat

A THESIS
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
MASTER OF SCIENCE
Department of Chemistry

1 999

ABSTRACT
MECHANISTIC STUDIES OF THE CATALYTIC STOBBE CONDENSATION
REACTION
By

Ancuta Cemat

We report here the results of a series of mechanistic investigations of the
condensation reaction between formaldehyde and dimethyl succinate, performed under
catalytic conditions. The objective of the project was to understand the reaction pathway
towards the final product, citraconic anhydride. To investigate the reaction course we
performed a series of experiments using deuterium-labeled compounds such as CH3OD,
CD20 and CD3OCOCH2CH2COOCD3. The hydrogen-deuterium exchange was
monitored using mass spectrometry techniques. The experiments were performed in the
vapor phase, at atmospheric pressure and high temperatures, using y-alumina as the
catalyst. Our results indicated that the reaction is catalde by the acidic sites on the
alumina surface. Also, unexpected hydrogen-deuterium exchange between the reacting
species was detected. Mechanistic routes for the condensation reaction and for the

hydrogen-deuterium exchange processes are proposed.

To my parents and to Dr. Florian A. Urseanu

ACKNOWLEDGMENTS

I have passed through many pitfalls on my way to this degree and special thanks
need to be given to those who helped me pick myself up and continue on.

First of all, I would like to thank Dr. James E. Jackson, my advisor, and Dr.
Dennis J. Miller, my co-advisor, for their great guidance throughout the various steps of
my research.

I would also like to thank Dr. Robert Maleczka and Dr. Milton Smith for serving
as members of my guidance committee.

I would like to give special thanks to my friend Dr. Kirthivasan Nagarajan, who
was patient with me and helped me very much, and to all of the group members, who
stood by me for good and for bad.

My deepest appreciation is due to my parents for their love, encouragement and

faith in me during the course of my studies.

Ancuta Cemat

August 1999

iv

TABLE OF CONTENTS

LIST OF FIGURES ........................................................................................................... vii

LIST OF ABBREVIATIONS ............................................................................................. ix

CHAPTER 1: Stobbe Condensation Reaction: A Review

1 .1 . Introduction ................................................................................................................... 2
1.2. Classic Stobbe Condensation ........................................................................................ 2
1.2.1. General Character and Mechanism ...................................................................... 2
1.2.2. Scope and Limitations .......................................................................................... 6
1.2.3. Side Reactions ...................................................................................................... 7
1.2.4. Applications of Classic Stobbe Condensation ................................................... 10
1.3. Catalytic Stobbe Condensation ................................................................................... 11
l .3.1 . Introduction ........................................................................................................ 1 1
1.3.2. Scope and Limitations ........................................................................................ 11
1.3.3. Applications of Catalytic Stobbe Condensation ................................................ 13
l .4. References ................................................................................................................... 15
CHAPTER 2: Experimental studies and results ............................................................... 17
2.1 . Introduction ................................................................................................................. 18
2.2. Catalyst Characterization ............................................................................................ 20
2.2.1. Surface Area ....................................................................................................... 20

2.2.2. Acid-Base Properties ......................................................................................... 21

2.2.3. Characteristics of y-Alumina SA 3177 .............................................................. 24
2.3. Deuterium Exchange Experiments Involving CH3OD ............................................... 25
2.3.1 . Introduction .................................................................................................. 25
2.3.2. Experimental Methods and Results ............................................................. 26
2.4. Deuterium Exchange Experiments Involving CD20 .................................................. 35
2.4.1 . Introduction ....................................................................................................... 3 5
2.4.2. Experimental Methods and Results ................................................................... 35
2.5. Deuterium Exchange Experiments Involving CD3OCOCH2CH2COOCD3 ............... 45
2.5.1 . Introduction ....................................................................................................... 45
2.5.2. Experimental Methods and Results ................................................................... 45
2.6. Proposal of a Mechanism for Catalytic Stobbe Condensation .................................. 53
2.7. Summary ..................................................................................................................... 56
2.8. References ................................................................................................................... 57

vi

LIST OF FIGURES

Figure 1.1.1. Alkylidenesuccinic acid .................................................................................. 2
Figure 1.2.1. B-Diketo-compounds ...................................................................................... 3
Figure 1.2.2. Teraconic acid ................................................................................................. 3
Figure 1.2.3. General representation for classical Stobbe condensation .............................. 4
Figure 1.2.4. B-carbethoxy-y,y-diphenylvinylacetic acid .................................................... 4
Figure 1.2.5. Mechanism for classical Stobbe condensation. Reaction sequence (1) ......... 5
Figure 1.2.6. Mechanism for classical Stobbe condensation. Reaction sequence (2) .......... 5
Figure. 1.2.7. Diethyl cyclohexane-l,4-dione-2,5-dicarboxylate ........................................ 8
Figure 1.2.8. Fulgide .......................................................................................................... 10
Figure 1.2.9. Photochromism in fulgides ........................................................................... 10
Figure 2.1.1. Set-up for catalytic Stobbe condensation between CHZO (from trioxane) and
DMS ................................................................................................................................... 18
Figure 2.1.1. Catalytic Stobbe condensation between CHZO and DMS ............................ 19
Figure 2.1.2. Itaconic anhydride ........................................................................................ 20
Figure 2.2. 1. Lewis acid and basic sites on alumina .......................................................... 24
Figure 2.3.1. H-D exchange between CH3OD and DMS ................................................... 25
Figure 2.3.2. Gradual replacement of hydrogen by deuterium .......................................... 26
Figure 2.3.3. Fragmentation pattern for DMS .................................................................. 27
Figure 2.3.4. Comparison between the mass spectra of standard DMS (top) and of a

product mixture containing mainly D4-DMS (bottom) ...................................................... 28
Figure 2.3.5. Extent of deuterium incorporation in DMS over alumina ............................ 29

vii

Figure 2.3.6. Extent of deuterium incorporation in DMS over zeolite .............................. 31
Figure 2.3.7. Extent of deuterium incorporation in DMS over hydrotalcite ...................... 33
Figure 2.3.8. Mechanism for I-I/D exchange between CH3OD and DMS on alumina ..... 34

Figure 2.4.1. Formation of deuterated citraconic anhydride from CD20 and DMS .......... 35
Figure 2.4.2. Set-up for the reaction between CD20 and DMS ......................................... 36
Figure 2.4.3. Fragmentation pattern for citraconic anhydride ........................................... 37

Figure 2.4.4. Comparison between the mass spectra of standard citraconic anhydride (top)
and of a product mixture containing deuterated citraconic anhydride species (bottom)...38

Figure 2.4.5. Extent of deuterium incorporation in CA. CDZO/DMS Experiment 1 ......... 40
Figure 2.4.6. Extent of deuterium incorporation in DMS. CDzO/DMS Experiment 1 ...... 41
Figure 2.4.7. Extent of deuterium incorporation in CA. CDzO/DMS Experiment 2 ......... 43

Figure 2.4.8. Extent of deuterium incorporation in DMS. CDZO/DMS Experiment 2 ...... 44

Figure 2.4.9. Mechanism for the loss of deuterium from Dz-CA ...................................... 42
Figure 2.4.10. Mechanism for deuterium incorporation in DMS ...................................... 45
Figure 2.5.1. Reaction between CHZO and CD3OCOCH2CH2COOCD3 ......................... 46

Figure 2.5.2. Comparison between the mass spectra of standard DMS (top) and of
CD3OCOCH2CH2COOCD3 synthesized (bottom) ............................................................ 47

Figure 2.5.3. Extent of deuterium incorporation in CA. D6-DMS/TO Experiment ........... 49
Figure 2.5.4. Extent of deuterium incorporation in DMS. D6-DMS/TO Experiment ....... 50
Figure 2.5.5. Extent of deuterium incorporation in DMS. D6-DMS Experiment .............. 52
Figure 2.6.1. Proposed mechanism for catalytic Stobbe condensation. First step ............. 53
Figure 2.6.2. Proposed mechanism for catalytic Stobbe condensation. Second step ........ 54

Figure 2.6.3. Proposed mechanism for catalytic Stobbe condensation. Third step ........... 55

viii

DMS

TO

CA

TPD

L-cat.

LIST OF ABBREVIATIONS

Dimethyl Succinate

Trioxane

Citraconic anhydride
Temperature-Programmed Desorption
Hydrogen/Deuterium

Lewis acid site on the catalyst surface

CHAPTER 1

Stobbe Condensation Reaction : A Review

1 .1 . Introduction ................................................................................................................... 2
1.2. Classic Stobbe Condensation ........................................................................................ 2
1.2.1. General Character and Mechanism ...................................................................... 2
1.2.2. Scope and Limitations .......................................................................................... 6
1.2.3. Side Reactions ...................................................................................................... 7
1.2.4. Applications of Classic Stobbe Condensation ................................................... 10
1.3. Catalytic Stobbe Condensation ................................................................................... 11
1 .3.1. Introduction ........................................................................................................ 1 1
1.3.2. Scope and Limitations ........................................................................................ 11
1.3.3. Applications of Catalytic Stobbe Condensation ................................................ 13
l .4. References ................................................................................................................... 15

1.1. INTRODUCTION

The Stobbe condensation is the reaction of carbonyl compounds with an ester of
succinic acid to form alkylidenesuccinic acids (substituted itaconic acids) (Figure 1.1.1),
or isomers formed by a tautomeric shift of hydrogen.

R cozn

RH—COZH

Figure 1 .1 .1 . Alkylidenesuccinic acid

In the classic Stobbe condensation the condensing agent is a base and the primary
product is the half-ester of the alkylidenesuccinic acid. In the recently reported catalytic
Stobbe condensation, the main product is the anhydride of the alkylidenesuccinic acid.

The mechanism of the classic Stobbe condensation was thoroughly studied1 and is
well known today. In contrast, the mechanism of the catalytic Stobbe condensation has
not been explored. The objective of this research project was to understand the pathway
of the condensation reaction between formaldehyde and dimethyl succinate, performed at
high temperature and atmospheric pressure, under y-alumina catalysis. For this purpose,

experiments involving deuterium-labeled compounds were performed.

1.2. Classic Stobbe condensation
1.2.1. General character and mechanism

In 1893 Hans Stobbe2 demonstrated that when a mixture of acetone and
diethylsuccinate was treated with sodium ethoxide, the expected acetoacetic ester type of

condensation to give the following B-diketo compounds (Figure 1.2.1 .)

COCH2COCH3 COCHZCOCH3

[C02C2H5 |:COCH2COCH3

Figure 1 .2. 1. B-Diketo-compounds

did not take place. Instead, the main reaction product was teraconic acid (Figure 1.2.2.),
CH3 COZH
CH3 COzH

Figure 1.2.2. Teraconic acid
formed by an aldol type of condensation between the carbonyl group of the ketone and
an a-methylene group of the ester. Stobbe and his collaborators undertook an extensive
study which revealed that both aldehydes and ketones generally condense with succinic
esters in this special manner.
Each mole of ester requires one mole of alkoxide. The lactone derivative is

formed first, followed by the salt of the half-ester, which, upon acidic treatment, affords
the alkylidenesuccinic acid, in the form of either the half-ester or the dibasic acid

produced by hydrolysis (Figure 1.2.3.).

R (302(3sz
>20 + i 1.Na0C2H5 R COICZHS R C02“
C02C2H5

2. aq. HCl ' fl 0' >=L
R R COzH R €0er

Figure 1.2.3. General representation of the classic Stobbe condensation
It is striking that this facile aldol type of condensation of esters with ketones is
limited to succinic and substituted succinic esters. For example, benzophenone condenses
with diethylsuccinate3 to give pure [3-carbethoxy-y,y-diphenylviny[acetic acid (Figure
1.2.4.)

Ph C02C2H5

Ph COZH

Figure 1.2.4. B- carbethoxy-y,y-diphenylvinylacetic acid
in 90% yield. In contrast, under the same conditions4 this ketone fails altogether to react
with ethyl or t-butyl acetate.

The success of the classic Stobbe condensation is not solely attributable to the
high reactivity of the ot-methylenes of succinic esters, as shown by the failure of diethyl
malonate, which has a more reactive (It-methylene group, to condense in any appreciable
extent with benzophenone.4

The specificity of succinic esters in this reaction may be associated with the
juxtaposition of a carbethoxy group with the newly formed alkoxide site, setting it up for

ring formation, as indicated in reaction sequence (1), below (Figure 1.2.5.).

R R cozczH5 R C02C2H5
\8/ T R
( v ¢
COZCZHS 8
(L) ()0sz
11
R COZCZHS R R 002(3sz
CszOG-D '1' ¢
60 9H5

I
Figure 1.2.5. Mechanism for classic Stobbe condensation. Reaction sequence (1).
The postulation of an intermediary paraconic ester (1)"4 is reasonable in view of
the fact that such substances are isolable,6 particularly when shorter reaction periods are
employed,7 and that they are cleaved by alkoxides in excellent yield to give salts of the

unsaturated half-esters.8 This cleavage may be represented by reaction sequence (2),

(Figure 1.2.6.).

 

 

Figure 1.2.6. Mechanism for classic Stobbe condensation. Reaction sequence (2).

The combined reaction sequences ( l) and (2) thus constitute a satisfactory
rationalization of the course of the classic Stobbe condensation, the irreversibility of the
second step driving the reaction to completion.

The more obvious mechanism, in which the ketone first condenses with the
succinate eliminating water which then reacts with the alkoxide to form hydroxide ion
which in turn effects partial saponification of the diester, is not tenable in view of:

(a) the failure to isolate the postulated intermediary diester, even when a large excess of
diethyl succinate was employed in the condensation, thus afi‘ording a highly competitive
source of ester groups to react with the limited amount of hydroxide ion.9
(b) the failure of other esters4 with comparably reactive methylene groups to condense
readily.

(c) the failure of the appropriate unsaturated diester to give a good yield of half-ester on
partial saponification.”l2
(d) the fact that isomers of citraconic and mesaconic acid type, which would be expected

tautomers of certain alkylidenesuccinic diesters,l3 have never been found as products of

the classic Stobbe condensation.

1.2.2. Scope and limitations

The carbonyl compounds that undergo the Stobbe condensation include members
of the following classes of substances: aliphatic, aromatic, and a,B-unsaturated
aldehydes; aliphatic, alicyclic, and aromatic ketones; diketones; keto esters and cyano

ketones.l

The succinic esters that have been employed are diethyl, dimethyl and di-t-butyl
succinate, and also or-substituted ary1-, alkyl—, and alkylidene-succinic esters.
Among the condensing agents used, sodium ethoxide, potassium t-butoxide and

sodium hydride gave the best results.

1. 2.3. Side reactions ‘
Several side reactions may accompany the classic Stobbe condensation reaction.

When aldehydes are used in the Stobbe condensation, the side reactions which

14-16 17,18

have been observed are the Cannizzaro reaction and the aldol condensation.

 

Aromatic and aliphatic aldehydes with no or-hydrogens, give the Cannizzaro reaction
when treated with bases. In this reaction one molecule of aldehyde oxidizes another to the
acid and is itself reduced to the primary alcohol. Aldehydes with an or-hydrogen do not
give the reaction, because when these compounds are treated with base the aldol
condensation is much faster.

The failure of certain ketones containing highly active or-methyl or or-methylene
groups to give good yields in the Stobbe condensation may be due in part to a tendency
for these ketones to enolize. The anion produced may be relatively stable, as with
dibenzyl ketone, in which event the ketone is recovered unchanged. On the other hand the
anion may compete1 with the ester anion in reaction with free ketone, in which case self-
condensation of the ketone is effected.

When sodium ethoxide in ether (or ethanol) is used as the condensing agent, there

is ahnost always a significant amount of reduction of the ketone to the corresponding

carbinol.19 This reduction is effected by the ethoxide, which is converted to acetaldehyde,
which in turn is largely responsible for the formation of resinous material and darkening
usually observed with this procedure.

Sodium methoxide largely eliminates the oxidation-reduction complication, since
metal methoxides are weaker reducing agents than ethoxides.20 At the same time, sodium
methoxide is a weaker condensing agent (since it is a weaker base) than sodium ethoxide
and has therefore not found general use.

Potassium t-butoxide (in t-butyl alcohol) is a considerably stronger condensing
agent than sodium ethoxide and affords better yields of pure products in much shorter
reaction periods.21 Even potassium t-butoxide and diethyl succinate, however, cause some
reduction of the ketone. In this case the reduction is effected by the alcohol formed as a
by-product in the Stobbe condensation and as a product of the self-condensation of the
succinate to produce diethyl cyclohexane-l ,4-dione-2,5-dicarboxylate (Figure 1.2.7.).

0

C02C2H5

CszOzC

Figure. 1.2.7. Diethyl cyclohexane-l ,4-dione-2,5-dicarboxy1ate
A significant amount of reduction22 occurs only with ketones that react slowly in
the classic Stobbe condensation, thus allowing a considerable concentration of ethoxide
to build up by the competing self-condensation. This reduction could be almost

completely eliminated by the use of dimethyl instead of diethyl succinate, a result that is

 

in accord with the comparative reducing properties of methoxide and ethoxide considered
above. Dirnethyl succinate is therefore useful in conjunction with t-butoxide for
condensation with slowly reacting ketones.

A solution to the problem of the competing self-condensation of the esters lies in
the use of an ester like di-t-butyl succinate.9 which reacts in this way relatively slowly.
The Stobbe condensation itself, however, is considerably slower with di-t-butyl than with
dimethyl or diethyl succinate, presumably owing to steric resistance of the carbo-t-butoxy
group to participation in the lactonization step. Therefore, longer periods of heating are
required.

Sodium hydride has the advantage of being inexpensive and easy to use as a
condensing agent. A small amount of alcohol23 corresponding to the succinate is usually
required to initiate the reaction. The alcohol reacts rapidly with the sodium hydride to
produce sodium alkoxide, the true condensing agent. As the reaction proceeds, more
alcohol is formed as a by-product. This reacts rapidly with the sodium hydride, producing
additional sodium alkoxide; and, as the concentration of the latter gradually increases,
there is a corresponding increase in the rate of condensation as evidenced by the rate of

hydrogen evolution.

1.2.4. Applications of classic Stobbe condensation
Besides the obvious general use for preparing many varieties of unsaturated and

(by hydrogenation) saturated substituted succinic acids, the classic Stobbe condensation

 

has found wide application1 in the synthesis of other types of substances, including
substituted lactones, naphthols, tetrahydroindanones, and tetralones.

Fulgides (Figure 1.2.8.) are derivatives of dimethylene succinic anhydride.

R] o
R
2 0
R3
0

Figure 1.2.8. Fulgide
These compounds were first investigated by Stobbe24 and he coined their name (from the
Latin fulgere - to glisten and shine) because they were frequently obtained as beautiful
reflective colored crystals. Typically, fulgides are yellow or orange crystalline
compounds which change to orange, red, or blue upon irradiation with ultra-violet light,
if at least one aryl group is present in the molecule, to participate in a photochemical ring

closure (Figure 1.2.9.).

Ph 0
Ph
O hv
Ph
h 0

 

Figure 1.2.9. Photochromism in fulgides
Photochromism has been observed in crystals, solutions, polymers and glasses over a

wide range of temperatures and conditions.

10

F ulgides have various uses: chemical actinometers25 for the near u.v. region,
optical data storage, sunlight active variable density optical filters (sunglasses are an
obvious potential commercial application of these compounds). New applications of
fulgides are in spatial light modulation,26 optical waveguide construction,27 and non-

linear optical switching.28

1.3. Catalytic Stobbe condensation
1.3.1 Introduction

Recently chemists have become interested in performing the Stobbe condensation
reaction in a catalytic fashion. Several heterogeneous catalytic processes have been
patented, all claiming the preparation of citraconic (methyl succinic) anhydride,
citraconic acid and/or its isomer, itaconic (methylidene succinic) acid. Hydrolysis of
citraconic anhydride allows the formation of citraconic acid which upon thermal

isomerization forms itaconic acid.

1.3.2. Scope and limitations

In 1974 B. E. Tate and R. G. Berg (Pfizer Inc.)29 patented for the first time a
catalytic procedure for the preparation of citraconic anhydride. Formaldehyde (from
trioxane, or a heavy slurry of paraforrnaldehyde in mineral oil, heated to liberate CHZO
gas) is reacted in the vapor phase with dimethyl or diethyl succinate or with succinic
anhydride at a temperature between 280-410 °C. The reaction time varies inversely with

the temperature and ranges from 10 to 25 seconds. Best results are obtained at

11

 

temperatures of 340-410°C. Lower temperatures require longer contact times between the
reagents and the catalyst and result in lower yields. The catalyst is selected from a group
consisting of thorium sulfate, potassium diacid phosphate, lithium carbonate on alumina
support, and lithium phosphate. An inert gas, such as nitrogen, hydrogen, helium, carbon
dioxide, carbon monoxide, was used as a carrier gas. The molar ratio of formaldehyde to
starting compound is preferred to be about 3.5:1 to 5:1 for best results. At molar ratios
close to 1:1 the yields are low and accompanied by a sharp decrease in catalyst efficiency,
while at molar ratios much above 5:1 little further improvement is obtained. Citraconic
anhydride, recovered from the gaseous effluent from the reactor by trapping in a series of
cold traps followed by distillation, is obtained in 60- 80% yield at 90-98% conversion.

The Denki Kagaku Kogyo Inc.30 also patented several processes for the synthesis
of itaconic acid, citraconic acid and their ester derivatives. The catalysts used are zeolites,
plain or impregnated with PdClz, or MnClz (molecular sieves 10x), or impregnated with
La(NO3)3, Ni(NO3)3, CoClz, (molecular sieves 13x) or silica-alumina plain or containing
group IB or [IE metal salts, such as ZnClz, CuClz, LaC13, MnClz, and NiClz. The ratio of
succinic acid (or succinate or succinic anhydride) to formaldehyde (fi'om trioxane or
formalin 37%) is preferred to be 2:1 ~ 5:1. The temperature ranges between 350 and
450°C.

Itaconic acid can be easily obtained from citraconic acid by isomerization. An
aqueous 25% citraconic acid solution, autoclaved at 160°C for 8 hours, gives about 45%

yield of itaconic acid easily recovered by evaporation and crystallization from water.3 I

12

There is no mention in the literature about a carbonyl compound other than formaldehyde
being used for the catalytic Stobbe condensation. Also, no mechanistic study of this

reaction has been reported.

1.3.3. Applications of catalytic Stobbe condensation

The interest in obtaining itaconic acid in a catalytic fashion rises from the
methods presently used for its preparation and fiom its multiple uses.

Itaconic acid is currently manufactured32 by the process of fermentation of sugars
(from molasses). A 15-25% carbohydrate solution is sterilized and incubated with a
culture of Aspergillus terreus. at 35-40°C for 3-7 days under a pressure of 10-15 psi. The
pH is maintained at 5.0 by adding lime. Itaconic acid is separated from the broth by
acidification, concentration and crystallization when the fermentation process is
complete.

Itaconic acid is a valuable monomer for polymerization because of the presence of
the two carboxyl groups and the methylene group (from the double bond). The methylene
group is able to take part in addition polymerization forming polymers with many free
carboxyl groups which confer advantageous properties on the polymer. Sodium
polyitaconate or other alkali salts may be used in detergents to improve clarity and color,
used in bleaches as a stabilizer, and used in metal cleaners for rust removal.33

Itaconic acid itself polymerizes very slowly to form low molecular weight
polymers, so it is typically used in copolymerization. Itaconic acid is a specialty

monomer that affords performance advantages to certain polymers when it is incorporated

13

in small amount. Styrene-butadiene latexes containing less than 10% itaconic acid are

”'36 and paper coating.” Emulsion stability, clarity, water

widely used in carpet backing
resistance of the coatings, and adhesion to substrates are improved by the inclusion of
itaconic acid.

Dental cements made of acrylic-itaconic acid copolymers that are cured with
polyvalent metal compounds such as aluminosilicates or oxides of zinc and magnesitun
possess a good compressive and adhesive strength as well as physiological
compatibility.38

The dimethyl, diethyl and di-n-butyl esters of itaconic acid can also be used in
copolymers, (e. g. for adhesives) and esters with long chain alcohols have been proposed
as plasticizers.

With amines, itaconic acid produces N-substituted pyrrolidinones that can be used
as thickeners for greases.39 Other pyrrolidinones made from itaconic acid and a wide

range of amines have potential uses in detergents, shampoos,40 pharmaceuticals, and

herbicides.

14

1.4. References
1. Johnson, W. S.; Daub, G. H. Organic Reactions, Wiley Inc. NY, 1951, 1-73.
2. Stobbe, H. Ber. 1893, 26, 2312-2315.
3. Johnson, W.; Petersen, J. W.; Schneider, W. P. J. Am. Chem. Soc. 1947, 69, 74-78.
4. Johnson, W. S.; McCloskey, A. L. J. Am. Chem. Soc. 1950, 72, 517-520.
5. Stobbe, H. Ann. 1894, 282, 280-283.
6. Robinson, R.; Seijo, E. J. Chem. Soc. 1941, 582-586.
7. Stobbe, H.; Vieweg, R. Ann. 1911, 380, 78-83.
8. (a) Roser, K.; Ann. 1883, 220, 258-260; (b) Fittig, R. Ann. 1890, 256, 50-52; (c)Fittig,
R. Ber. 1894, 27, 2681-2685.
9. Johnson W.S.; Miller, M. W. J. Am. Chem. Soc. 1950, 72, 511-516.
10. Stobbe, H. Ber. 1908, 41, 4350-4357.
11. Johnson, W. S.; Goldman, A. J. Am. Chem. Soc. 1944, 66, 1030-1034.
12. Johnson, W. S.; Graber, R. P. J. Am. Chem. Soc. 1950, 72, 925-929.
13. Coulson, R.; Kon, G. A. R. J. Am. Chem. Soc. 1932, 2568-2572.
14. Stobbe, H.; Noaum, P. Ber. 1904, 37, 2240-2246.
15. Fichter, F.; Scheuerrnann, B. Ber. 1901, 34, 1626-1633.
16. Stobbe, H. Ann. 1911, 380, 49-54.
17. Fittig, R.; Thron, D. Ann. 1899, 304, 288-293.
18. Stobbe, H.; Leuner, K. Ber. 1905, 3682-3685.
19. Stobbe, H. Ann. 1890, 308, 114-116.

20. Adkins,H.; Elofson, R. J. Am. Chem. Soc. 1949, 71 , 3622-3625.

15

 

21. Johnson, W. S.; Johnson, H. C. E.; Petersen, J. W. J. Am. Chem. Soc. 1945, 67, 1360-
1364.

22. Johnson, W. S.; Petersen, J. W. J. Am. Chem. Soc. 1947, 69, 2942-2947.

23. Daub, G. H.; Johnson, W. S. J. Am. Chem. Soc. 1950. 72, 501-504.

24. Stobbe, H. Ber. 1904, 37, 2236-2241.

25. Wintgens, V.; Johnston, L. J.; Scaiano, J. C. J. Am. Chem. Soc. 1988, 110, 511-515.
26. Ilge, H. D.; Langbein, H.; Reichenbacher, M. J. Prakt. Chem. 1981, 323, 367-371.
27. Piggot, R. D. PhD Thesis, Aberystwyth, 1975.

28. Cush, R.; Trundle, C.; Kirkby, C. J. G. Electron. Letters, 1987, 23, 419-423.

29. Tate, B. E.; Berg, R. G. (Pfizer Inc.) US. Pat. 3,835,162, 1974.

30. Tsunehiko, S.; Chiyuki, F. (Denki Kagaku Kogyo Inc.) Jpn. Pat. 49,101,327, 1974.
31. Linstead, R. P.; Mann, T. J. W. J. Chem. Soc. 1931, 726-729.

32. Yocum, R. H.; Nyquist, E. B. Functional Monomers, Dekker Inc. NY, 1974.

33. Carter, R. P.; Irani, R. R. (Monsanto) US. Pat. 3,405,060, 1968.

34. (Dunlop Rubber Co. Ltd.), Neth. Pat. Appl. 6,405,106, 1964.

35. Philip, D. H.; Madison, N. L. (Dow Chemical Co.) US. Pat. 3,928,695, 1975.

36. Peaker, C. R. (Uniroyal Inc.) US. Pat. 3,575,913, 1971.

37. Megee, J. F.; Nickerson, R. G. US. Pat. 3, 793,244, 1974.

38. Crisp, S.; Wilson, A. D. (National Research Development Corp.) Ger. Oflen.
2,439,882, 1975.

39. Gordon, A. A.; Coupland, K. (Exxon Research and Engineering Co.) Ger. Pat.
3, 001, 000, 1980.

40. Christiansen, A. ( Miracol Chemical Company Inc.) Br. Pat. 1,5 74, 916, 1980.

16

CHAPTER 2

Experimental studies and results

2.1. Introduction ................................................................................................................. 1 8
2.2. Catalyst Characterization ............................................................................................ 20
2.2.1. Surface Area ....................................................................................................... 20
2.2.2. Acid-Base Properties ......................................................................................... 21
2.2.3. Characteristics of y-Alumina SA 3177 .............................................................. 24
2.3. Deuterium Exchange Experiments Involving CH3OD ............................................... 25
2.3.1 . Introduction .................................................................................................. 25
2.3.2. Experimental Methods and Results ............................................................. 26
2.4. Deuterium Exchange Experiments Involving CD20 .................................................. 35
2.4.1 . Introduction ....................................................................................................... 3 5
2.4.2. Experimental Methods and Results ................................................................... 35
2.5. Deuterium Exchange Experiments Involving CD3OCOCH2CH2COOCD3 ............... 45
2.5.1 . Introduction ....................................................................................................... 45
2.5.2. Experimental Methods and Results ................................................................... 45
2.6. Proposal of a Mechanism for Catalytic Stobbe Condensation .................................. 53
2.7. Summary ..................................................................................................................... 56
2.8. References ................................................................................................................... 57

17

2.1. Introduction

Stimulated by the great economic value of itaconic acid, our group started a
project focused on the development of a continuous process for the production of itaconic
acid via the catalytic Stobbe condensation. The reaction studied1 was between dimethyl
succinate (DMS) and formaldehyde, using y-alumina as the catalyst. Reactions were run
at 380°C and atmospheric pressure in a fixed bed tubular reactor placed in the center of a

furnace to ensure even heating (Figure 2.1.1.).

'—I ,: ' :‘
11/ rotarmter
T7
N2

 

 

 

 

 

 

 

 

 

 

burette—p .” 47—- fixed-bed reactor
.” q .”
t,’ ’I' «— finmce
S 2 <— air condemer
? [4— product collection vial
pe . 1'
PW!)

Figure 2.1.1. Set-up for catalytic Stobbe condensation between CHZO (fi'om trioxane) and
DMS

Trioxane, (99%, purchased from Aldrich Chemical Co.) which is soluble in DMS
(solubility = 1 mol of trioxane/mo] of DMS)l was used as the formaldehyde source.
DMS (98%) was purchased from Aldrich Chemical Co. The molar ratio of the reagents

was CHZO : DMS = 2 : 1. The reagent mixture was stored in a burette and pumped

18

 

dropwise into the reactor at a rate of 0.3 mL/min., by means of a Masterflex (Cole-
Farmer) peristaltic pump. Nitrogen was used as the can'ier gas at a flow rate of 20
mL/min., controlled by a Cole-Parmer rotameter.

y-Alumina (SA 3177, purchased as 1/8 inch pellets from Norton Inc.) was used in
a fixed quantity of 5 grams. The pellets were ground and sieved to the size of 30 - 60

mesh. The catalyst was heated in the reactor at 380-400°C for at least 3 hours to eliminate

“A \

moisture before the reagent mixture was added.

The product mixture was passed through an air condenser for cooling purposes

 

and collected every 10 min. (up to four collections) in glass vials. Analysis by GC-Mass
spectrometry used a Hewlett Packard 5890 Series 11 gas cromatograph (SPB-l non-polar
column) connected with a VG Trio-1 mass spectrometer.

The catalytic “Stobbe condensation” product of formaldehyde with DMS was

obtained as citraconic anhydride, (Figure 2.1 .1.).

O
COZCH3 CH3
H2C=o + E ___.A1203 0
COZCH3 380 C
citraconic anhydride

Figure 2.1.1. Catalytic Stobbe condensation between CHZO and DMS

Citraconic anhydride is a double bond isomer of itaconic anhydride (Figure 2.1.2.).

19

CH

Figure 2.1.2. Itaconic anhydride
Beside citraconic anhydride, the product mixture contained monomethyl succinate,
succinic anhydride, and unreacted DMS .
For a better understanding of the reaction, the catalyst was characterized by
various methods. Exchange experiments involving deuterium-labeled reagents were

performed to probe the mechanism of the reaction.

2.2. Catalyst characterization

As heterogeneous catalysis is a surface phenomenon, the determination of surface
properties plays a large part in catalyst characterization. The investigated surface
pr0perties of alumina SA 3177 were total surface area, surface concentration of acidic
sites, and surface concentration and strength of basic sites. The methods used for these
investigations are all based on the physical adsorption of a gas (adsorbate) on the catalyst

(adsorbent) surface.
2.2.1. Surface area

The BET dynamic method2 is a rapid method of surface area determination.

Nitrogen is used as adsorbate and helium (a gas not adsorbed significantly by most

20

solids) is employed as carrier gas and diluent. The volume of gas passed through the
sample (adsorbent bed brought to 77 K in a liquid nitrogen bath) is adsorbed and
equilibrium between gas phase and solid phase is rapidly established. The change in
nitrogen concentration is recorded.
The main assumptions of this method3 are:
- the adsorption is supposed to be localized on well defined sites; all of the sites have the
same energy (homogeneous surface) and each of them can accommodate one adsorbate
molecule.
- an adsorption-desorption equilibrium is established between molecules reaching and
leaving the solid surface.

The surface area can be calculated with the formula: SBET = 4.37 x 106 vm
where: s = specific surface area (surface area of the mass unit of solid) (mZ/kg)

vm = the amount of adsorbate just sufficient to cover with a complete monolayer
the whole surface developed by the unit mass of adsorbent
4.37 x 106 = constant which includes the surface area occupied by a N2 molecule

(at 77 K), the Avogadro constant (6.025x1023 mole!) and the molar volume of N2

(22.414x10'3 m3x mole")

2.2.2. Acid-base properties
A complete description of acidic and basic properties on solid surfaces requires
the determination of the acid and base strength, and of the amount and nature (Bronsted

or Lewis type) of the acidic and basic sites.

21

2.2.2.1. Acidity

The acid strength of a solid is defined as the ability of the surface to convert an
adsorbed neutral base into its conjugate acid. The amount of acid on a solid is usually
expressed as the number or mmol of acid sites per unit weight or per unit surface area of
the solid, and is obtained by measuring the amount of a base which reacts with the solid
acid. The strength can be determined using indicators. For the determination of the
density of acid sites, there are two main methods: an amine titration method using
indicators and a gaseous base adsorption method.3
A. Amine titration method using indicators

The qualitative determination is easily made by placing the sample (0.5 grams) in
powder form into a test tube, adding non-polar solvent, such as benzene, containing 1%
indicator, and shaking briefly. The color of suitable indicators adsorbed on a surface will
give a measure of its acid strength (expressed by the Hammett acidity function H0). If the
color is that of the acid form of the indicator, then the value Ho of the surface is equal to
or lower than the pK, of the conjugate acid of the indicator. Lower values of H0
correspond to greater acid strengths. This method determines the acid strength of a
catalyst relative to that of the conjugate acid of the indicator. The amount of acid sites can
be measured by titration of a solid acid suspended in benzene with n-butylarnine using an

indicator, immediately after the determination of the relative acid strength.

22

B. Gaseous base adsorption method

When gaseous bases are adsorbed on acid Sites, a base adsorbed on a strong acid
site is more stable than one adsorbed on a weak acid site, and is more difficult to desorb.
As elevated temperatures stimulate liberation of the adsorbed bases from the acid sites,
those at weaker sites will be released preferentially. Thus, the proportion of adsorbed base
at various temperatures can give a measure of acid strength. The amount of a gaseous
base which a solid acid can adsorb chemically from the gaseous phase is a measure of the
amount of acid on its surface.

In temperature-prograrnmed desorption (TPD) studies a solid previously
equilibrated with a basic adsorbing gas (ammonia, pyridine etc.) under well-defined
conditions is subjected to a programmed temperature rise and the amount of desorbing
gas is continuously monitored in a spectrum. In the ammonia case, no distinction can be
made between Bronsted and Lewis acidity, as the NH3 coordinatively bondes to the
catalyst giving only one peak in the spectrum. The distinction between Bronsted and
Lewis acidity is possible when pyridine is used as a gaseous base, because pyridine
coordinatively bonded to the catalyst gives peaks at different positions for Brtinsted acid
sites and for Lewis acid sites.
2.2.2.2. Basicity

The basic strength of a solid surface is defined as the ability of the surface to
convert an adsorbed neutral acid to its conjugate base. The density of basic sites is usually

expressed as mmol of basic sites per unit weight or per unit surface area of the solid.

23

There are two main methods for the measurement of strength and amount of basic sites:
benzoic acid titration method using indicators and gaseous acid adsorption method3.
A. Benzoic acid titration method using indicators

The principle of this method is similar to that for the determination of acidic
strength, using acid indicators. This method determines the basic strength of a catalyst,
relative to that of the conjugate base of the indicator. Similarly, the number of basic sites
can be measured by titrating a catalyst sample with benzoic acid using an indicator.
B. Gaseous acid adsorption method

The principle of this method is the same as that of the gaseous base adsorption
method. As adsorbates, acidic molecules such as carbon dioxide or nitric oxide can be
used. The amount of carbon dioxide irreversibly adsorbed is a good measure of the

amount of basic sites on solid surfaces.

2.2.3. Characteristics of y-Alumina SA 3177
Hindin proposed a simple schematic description for the generation of acid and basic sites

on the surface of alumina by dehydration5 (Figure 2.2.1 .).

OH OH 9 09
—o—Ar—o_Ar—o— 13—2. —O—Al—O—Al—O—

Lewis acid Basic
site site

Figure 2.2.1. Lewis acid and basic sites on altunina

24

The Lewis acid site is visualized as an incompletely coordinated aluminum atom
formed by dehydration while the basic site is considered to be a negatively charged
oxygen atom.

y-Alumina SA 3177 was tested by the TPD methods using NH3, pyridine and C02
and the presence of acid sites of the Lewis type and of basic sites was proven.4 The TPD
experiments showed that the density of basic sites is 0.049 mmol/g and the density of T
acid sites is 0.232 mmol/g. The base indicator method showed that the acid sites on
y-alumina SA 3177 have a pK,l ranging between (+1.1) and (-0.2). The surface area

determination by BET method showed that y-alumina SA 3177 has a surface area of

 

106.94 mZ/g.

2.3. Hydrogen-deuterium exchange experiments involving CH3OD
2.3.1. Introduction

The first step in the mechanism of the classic Stobbe condensation reaction is the
formation of the enolate anion at C2 in DMS by abstraction of a proton by a strong base.
In order to determine whether the anion is also generated under heterogeneous catalytic
conditions, experiments using DMS and CH3OD as reagents were performed. The H/D

exchange can be schematically represented as follows (Figure 2.3.1.).

H
COZCH, D COZCH
Al 0 3
+ CH3OD —2—2»
COZCH3 C02CH3

Figure 2.3.1. H-D exchange between CH3OD and DMS

25

The experiments demonstrated the enolization of the DMS. Also, by performing the
same experiments with a highly acidic catalyst and with a highly basic catalyst, it was

shown that the reaction is catalyzed by the Lewis acid sites on the alumina surface.

2.3.2. Experimental methods and results

The experiments were performed in the set-up presented previously. The molar
ratio of the reagents was DMS : CH3OD = 1 : 40 (exchangeable H : D = 1 : 10). CH3OD
(99.5%) was purchased from Aldrich Chemical Co.. The flow rate of the reagent mixture
was 0.3 mL/min. Nitrogen was used as carrier gas at a flow rate of 20 ml/min, controlled
by a Cole-Parmer rotameter. The catalyst, alumina SA 3177, 30-60 mesh in size, was
heated in the reactor at 380-400°C for at least 3 hours to eliminate moisture before the
reagent mixture started being added. The product mixture was passed through an air
condenser for cooling purposes and collected in glass vials for 10 minutes at each
temperature.

The experiments covered a large range of temperatures, from 190°C to 430°C, in
increments of 40°C.

The gradual H/D exchange follows the scheme (Figure 2.3.2.) :
H 4-DMS 911,92, H3D-DMS C_H30_D, HzDz-DMS M, HD3-DMS C_H,0_D, D 4-DMS

Figure 2.3.2. Gradual replacement of hydrogen by deuteritun

The product mixture was analyzed by GC-Mass spectrometry using a Hewlett

Packard 5890 Series 11 gas cromatograph (SPB-l non-polar column) connected with a

VG Trio-1 mass spectrometer.

26

The monitored compound was DMS (M=146.14 g/mole), whose fragmentation

pattern is‘5 shown in Figure 2.3.3.

 

 

l o ' 4 O
HzC’E‘OCHs EEC/E6) HZ?)
-CE30- -CO HZ \ ’0CH3
I-IthJ\C,OCI-13 thcmcua ' E
r r. 0
[M+] m/z=115 m/z=87

Figure 2.3.3. Fragmentation pattern for DMS

The EI+ mass spectrum6 of DMS does not show the molecular ion peak. The base
peak is at m/z = 115, the fragment which corresponds to the loss of a methoxy group. The
corresponding fragments containing 1, 2, 3 and 4 deuterium atoms will thus have
m/z = 116, m/z = 117, m/z = 118 and m/z = 119, respectively. A comparison between the
mass spectrum of standard DMS and of a product mixture containing mainly D4-DMS is
shown in Figure 2.3.4.

The fact that the m/z = 87 fragment in standard DMS mass spectrum corresponds
to a fragment with m/z = 91 in the second mass spectrum proves that the four deuterium
atoms in D4-DMS are located at C2 and C3 and no deuterium is incorporated in the
methoxy groups.7

Using8 the mass spectra of recovered DMS with various degrees of deuteration,
the change in the content of Dl-DMS, Dz-DMS, D3-DMS, D4-DMS and H4-DMS with
temperature was examined. The result is shown in Figure 2.3.5.

Over alumina, the extent of conversion of DMS to DINA-DMS, represented by

the m/z = 115 curve, was quite rapid and it begins to equilibrate from 350°C.

27

 

UG LAB-BASE The TRIO-1 GC-HS Data System
Sampleilou molecular ueight CC

 

 

 

 

 

 

 

 

 

 

. lnstrunenttTrio-l
0158814 4137(6.884)
1001 11?.1 2670592
55.1
59.1
zrs-
114.1\
87.1
116.1
er.2\45.2 86 1- 88.1 113.1 ’

a . v 7111; v 1 '1 v v 1' v AD'l/v vggltll V V l l V 7 ‘fi' T V V Y

"’33“ 49 59 Q9. 19 .19 99 199___112 12% 139, 142__4Ufll_

 

 

 

 

 

 

 

 

 

UG LAB-BASE The TRIO-1 GC-HS Data System
Samplgtlou molecular weight GC lnstrunent:Trio-1
jDICF1814 413 (6.884)
109* 119.1 487424
39.2
118.1
\
58.2
\
ZFS‘
117.1
37.2
32.3 91.1
9.2 116.1
44.3”‘3 ’61.: 291
.I.-,.1I.n.l . . . .....l.. .1, ......
"’z'IQ 4L 59 69 7.9. .62 9Q 1é9___119_ur9__1.aa_139_u9_

 

 

 

Figure 2.3.4. Comparison between the mass spectra of standard DMS (top) and of a

product mixture containing mainly D4-DMS (bottom)

28

evacuees”. mzaaofu
«28:? BS mEQ 5. 5580985 Estfizuc .8 Eucam .m.m.m Sam?—

 

 

omv

 

 

 

 

 

 

 

 

2282?:
8.. 8m 8m 8m 8m SF
. 0 t> _ 0 . r c
1r OP
-- om
on
.- 9.
.- om
mzoeolnl .- 8
229.8 1:1.
298+ .. 2.
295+
988+
.r on
.4 om
8.

 

mm

 

 

29

The Dl-DMS and Dz-DMS species started to appear at 190°C, but neither D3-DMS nor
D4-DMS were detected at this temperature. At 230°C all four deuterated species were
detected. As shown by the graph, the exchange seems to follow a certain pattern. At

23 0°C the relative concentrations of Dl-DMS and Dz-DMS are higher than D3-DMS and
D4-DMS concentrations. From 310°C onwards the exchange appears to be rapid, with the
D3-DMS and D4-DMS concentrations getting higher than the Dl-DMS and Dz-DMS
concentrations, consistent with a sequential series of H/D exchanges.

In order to determine whether the H/D exchange is due to the high temperatures or
not, the experiment was repeated with no catalyst over the same range of temperatures.
The mass spectra of recovered DMS showed no deuterium incorporation at any
temperature, which proves that the H/D exchange between CH3OD and DMS is due to
the catalyst.

The next question that comes to mind is whether the deuterium exchange is acid
or base catalyzed. As stated before, alumina has both acid and basic sites and the acid
sites are characterized by Lewis acidity.

Zeolites are known to have highly acidic sites and no basic sites. The
characteristics of zeolites 13-X were also detennined4 experimentally. The TPD
experiments showed that the density of acid sites is 1.851 mmol/g.The base indicator
method showed that the acid sites have a pK, ranging between (-8.1) and (-9.3). The
surface area is 435 m2/ g. The experiment was therefore repeated with zeolite l3-X in the
H form as catalyst, over the same range of temperature, 190°C-430°C. The results are

presented in Figure 2.3.6.

30

aoefiaxm mzeaommu
9208 5.5 mEQ E 823898": Enhance .«o ESxm .o.m.m 8:me

 

 

 

 

 

 

 

 

2321.58.
03.. can own 09.... ovw 8F
. n u b 0 . I o
QI\\|§\III _
.1 Or
. om
. om
.. ov
.. om W
.. co
wzoéolil
920.2.le .. on
m20.~o+
9295+ .. on
9208+
.- om
oo—

 

 

 

 

 

Over zeolite 13-X, the extent of conversion of DMS to Dam-DMS was greater and more
rapid than in the alumina case.

Since the H/D exchange occurred with a highly acidic catalyst, it must be
considered that in the alumina case, the exchange may likewise be catalyzed by the acidic
sites.

To verify whether the basic sites on alumina have any role in H/D exchange
between CH3OD and DMS, the experiment was repeated using a highly basic catalyst,
hydrotalcite (MgO-Ale3 cocatalyst) over the same range of temperatures, 190-430°C.

The characteristics of hydrotalcite (Mg : A1 = 75 : 25) were determined4
experimentally. The density of basic sites is 0.196 mmol/g. The surface area is 143.42
mz/g. The pK, of the basic sites is greater than (+4.8).

The results are presented in Figure 2.3.7. H/D exchange was observed, but to a
much smaller extent, even at high temperatures. Do-DMS is the main species at all of the
temperatures. D3-DMS and D4-DMS species remain at a very low concentration all the
time. Among the deuterated DMS species, Dl-DMS is predominant at all times.

Since a highly basic catalyst allowed only small extent of H/D exchange and since
alumina has only slight basic character (much less than hydrotalcite), it can be considered

that the contribution of basic sites on alumina to the IUD exchange is negligible.

32

 

Begum gen—0.5
86—5963 3.5 mEQ E cosmeoEooE Stance mo 3me find onE

 

 

omv

 

 

 

 

 

 

 

 

8.. 88 23$th Sm 8m 8,
.1 u .. I . H . 1.... 0
«III ”N. I m.
.- or
.. ow
- om
.. 9.
.. om
.. 8
mzoeolxl .. 2
9.8-8 1:1
£5,814! . 8
mzosolll
358+ . 8
8.

 

 

mm

 

 

33

As a conclusion to the above experiments, the IUD exchange between CH3OD
and DMS in the alumina case can be considered as being effected by the Lewis acid sites.
The mechanism for the hydrogen-deuterium exchange between CH3OD and DMS is

shown in Figure 2.3.8.

f‘L-cat
:0:

H ocrr3
H
cozcn3

 

CH3OUI l
t G t.
O L-Ca . vii—C3
> D

H _, H
cozcrr,

COzCH3

Figure 2. 3. 8. Mechanism for H/D exchange between CH3OD and DMS on alumina
The oxygen fi'om the carbonyl group in DMS coordinates the Lewis acid site on
alumina surface by the means of a lone pair of electrons. This oxygen will thus have a
positive charge. A methanol molecule abstracts a proton from C2 in DMS, and the
corresponding anion is stabilized as an enolate. The carbon-oxygen double bond in DMS
is then reformed and a deuterium is abstracted from a CH3OD molecule, followed by

detachment from the Lewis acid Site.

34

 

2.4. Experiments involving CD20 and DMS
2.4.1. Introduction

In order to study the behaviour of the two hydrogen atoms from formaldehyde,
experiments using CD20 and DMS were performed. If no H/D exchange takes place, then
deuterium should be present only in the methyl group of citraconic anhydride, according

to the following scheme (Figure 2.4.1.) :

0 0
D HC
D C=O + 0CH3 fl. 2 o
2 ocrr3

Figure 2.4.1. Formation of deuterated citraconic anhydride from CD20 and DMS
The results showed deuterium incorporation both in citraconic anhydride and in

the recovered DMS, which means that H/D exchange between deuterated citraconic

anhydride and DMS took place.

2.4.2. Experimental methods and results

Paraformaldehyde-dz (99%, purchased from Merck&Co., Inc.) was used as a
formaldehyde source. Since paraforrnaldehyde is not soluble in DMS, the set-up
described previously was modified (Figure 2.4.2.). Paraformaldehyde-d2 (4 grams) was
stored in a glass tube having a fi'it in the middle and was heated with a heating tape at
185-190°C to decompose paraformaldehyde-dz into CD20 gas. The tube was attached
directly to the reactor. The carrier gas, nitrogen (20 mein, controlled by a Cole-Farmer

rotameter), was passed through the tube, thus carrying the CD20 gas into the reactor. The

35

 

 

 

 

 

 

 

 

 

 

 

 

'1 V 4:
J. <— tube with
parafirrrmldehyde—dz
_—: 7 rotarmten,
burette—p .” 47-— fixed-bed reactor
I, H”, N2
,’ 3' ~— furnace
Q <— air condemer
f <— product collection vial
pe . 1 .
Pm!)

Figure 2.4.2. Set-up for the reaction between CD20 and DMS

feed of CD20 gas was regulated by the nitrogen flow. A quantity of 1.1 grams of

paraformaldehyde-dz was recovered in the tube, so only 2.9 grams (0.097 moles) were

consumed during the reaction, which gives a feed of 0.0725 g/rnin of CD20. DMS was

stored in a burette and fed at a flow rate of0. 15 mI/min into the reactor by means of a

Masterflex (Cole-Parmer) peristaltic pump. The catalyst, alumina SA 3177, 30-60 mesh

in size, was heated in the reactor at 380-400°C for at least 3 hours to eliminate moisture

before the reagents were added. The reaction was run

at 380°C and atmospheric pressure

for 40 minutes. The product mixture was passed through an air condenser for cooling

purposes and collected every 10 min (up to four collections) in glass vials. The molar

ratio of the reagents, based on DMS feed (0.046 moles total) and the quantity of CD20

36

consumed, was CD20 : DMS = 2.11 : 1. Due to the small amount of feed, the first sample
was not analyzable.

The product mixture from samples 2, 3, and 4 was analyzed by GC-Mass
spectrometry using a Hewlett Packard 5890 Series 11 gas cromatograph (SPB-l non-polar
column) connected with a VG Trio-1 mass spectrometer. Two compounds were
monitored: citraconic anhydride (M = 112.09 g/mole) and DMS (M = 146.14 g/mole).

The fragmentation pattern for citraconic anhydride is presented in Figure 2.4.3.

 

 

 

o ‘ .1
// H3C\
H3C\C /C\ CG)
0 - C02 H
H / H \C
\C \\
\\ O
o
[M+ ] m/z=68

Figure 2.4.3. Fragmentation pattern for citraconic anhydride
The EIJr mass spectrum9 of citraconic anhydride shows the molecular peak at m/z = 112.
The peak corresponding to the m/z = 68 fragment (resulting from the molecular fiagment
by the loss of a CO; molecule) was considered for calculations because this fragment
contains all four hydrogen atoms of the citraconic anhydride molecule, so any deuterium
incorporation is followed by an increase in m/z value of this fragment. A comparison
between the mass spectra of standard citraconic anhydride and of a product mixture

containing deuterated citraconic anhydride species is presented in Figure 2.4.4.

37

UG LAB-BASE The TRIO-1 GC-HS Data System

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Sam 1e:lou molecular ueight GC Instrumentttriori
PFX3§Z 375 (5.659)
100139.2 199600
68.0
'IJ'S‘
40.2
I
41.1
t ,6 -
11].»! I: ll 1+ ' '1 L. 11*"
M: 59 :2 ma rg'fl'iL'LU'"léerléi_mfi
UG LAB-BASE The TRIO-1 GC-flS Data System
Sam-letlou molecular ueizht GC Instrument11rio-1
1"' ' ' ..11 on: ‘ : . . to ' 4 - '-= to ‘1' + ' to 31- ” -‘;;2‘4
,00139.0 ‘ .
,40-9 68.0
69.0
ZI'S‘
41.0
«.0 7"“
’43.053.0 ‘70
l 52.01”".0 ' \l (71-9 113.0
0. L.- 1:--+ . .L . . . . . . . . . L-
"/249 :9 63 L 4 9L 499 ALL 159..

 

 

 

Figure 2.4.4. Comparison between the mass spectra of standard citraconic anhydride (top)

and of a product mixture containing deuterated citraconic anhydride species (bottom)

38

Using8 the mass spectra of citraconic anhydride (CA) with various degrees of
deuteration, from the product, the change in the content of Dl-CA, Dz-CA, D3-CA and
Do-CA (non-deuterated CA) with time was examined. The result is shown in Figure 2.4.5.
Although the expected product should have contained only D2-CA, the Do-CA species
was the most prevalent at all times (about 48-55%), followed closely by the Dl-CA
species (about 35-40%). Surprisingly, the Dz-CA species has a low concentration, about
10%. The D3-CA always has a very low concentration. No D4-CA has been detected.

Using8 the mass spectra of recovered DMS with various degrees of deuteration, as
previously described ( see 2.3.2), the change in the content of D,-DMS, Dz-DMS,
D3-DMS, D4-DMS and Do-DMS (non-deuterated DMS) with time was examined. The
result is shown in Figure 2.4.6. The Do-DMS species, which should have been the only
species present, has the highest concentration all the time (about 48-55%), followed
closely by Dl-DMS (about 30-3 5%). The D2-DMS species has a low concentration
(about 10-15%), while D3-DMS and D4-DMS have a very low concentration.

Since the paraformaldehyde-dz was used directly from the box, without any
drying, the detected H/D exchange was first suspected to have been the result of traces of
water possibly present in the paraformaldehyde-dz. Therefore, the experiment was
repeated in a slightly modified set-up. The CD20 gas was passed through a tube
containing anhydrous P205 and a tube containing molecular sieves, previously dried at
400°C for 2 hrs. Both tubes were heated with heating tape. The molar ratio of the reagents

was CD20 : DMS = 2.17 : 1. All the other parameters were the same.

39

e .5583 mzeoaou
<0 5 5230985 83.53% mo 82me .m.v.m ocswi

 

 

 

 

 

 

 

 

 

 

 

 

SE 0::
on A . Wm .._. om mw or
In 1 II II .1 IV”. - I..ll..l'.l! I J
Ll
1+
.- ow
(0&0 1.»...
<0-No+ .. on
(0.51.1
(0.00191 A. ow
.. om

 

 

 

 

40

°/o le

 

_ 1:68:35 mEQoaou
mEQ E catacofioofi Estasoe mo Eme .céN oSmE

 

 

 

 

 

 

 

 

 

 

 

r— E OE
0v mm on A _ MN :- ON mr
" H .r .P _
Ia 111111111
K ri‘r‘r
113 VI
.f
H
A\

 

m20-vo+
msaéolwrl
m20-~o+
w_>_0;o+

 

 

 

 

41

°/o z/W

 

 

 

 

The products were analyzed the same way as previously described and the results
are presented in Figure 2.4.7. for citraconic anhydride, and in Figure 2.4.8. for DMS. As
Shown by the graphs, the H/D exchange took place again. In the citraconic anhydride
case, the concentration of various deuterated species, especially Do-CA and Dl-CA,
varies over a larger range than in the previous experiment. In the DMS case, the
variations are about the same, except for the Do-DMS, which varies on a larger scale than
previously.

Since any moisture from paraformaldehyde-dz was eliminated, and since the H/D
exchange still took place, it is clear that the H/D exchange between CD20 and DMS is
mainly due to the catalyst. A possible explanation for the loss of deuterium from Dz-CA

(Figure 2.4.9.) and for deuterium incorporation in DMS (Figure 2.4.10.) is proposed.

CHsOH g
0 ' C H
D... “D 0 c a. ”3“ 0
2 0 DH H31) DHC
<— 0 2 0
.0. I
K» L-cat. O\Ci),.cat. (‘9‘?431-
CHsOH 4%”
O O
DHzC L—cat. DHZC
O 3 O
G
O ®O\ L-cat.

Figure 2.4.9. Mechanism for the loss of deuterium from Dz-CA

L-cat. represents the Lewis site, on the catalyst surface.

42

N 26853.0 £9050
<0 5 59216985 850838 mo ueoaxm 6.: oSwE

 

 

 

 

 

 

 

 

 

 

€25 05:.
9. mm on an cm 9 or
. m... 0 1 011111 o
IIIIIIIIvII.IIlIII ‘I‘IIIIIIIII
I .. or
- om
_
.. on
.. 8
.. om
.- 8
._
<98 1T
4. on
5.8+
<95 IT 8
58+
. 8

 

 

 

°/o z/Lu

 

 

43

 

N Bataan mEQodu
mED 5 5:22:85 Enhance we 886nm .wéd 0.53m

 

0».

mm

 

 

 

 

 

 

 

III“
«i
ll

wioéo +
wioéo it.-.
wZQ.No+
win—-po 1.1
min—.00 1.1

 

 

 

 

 

°/o Z/'~u

 

 

r‘L-cat. - t. - .
:O: CH3QH 6b)?” ca CH3§IH CH3DH )1 cat

H OCH; P111 OCH, 2 H OCH3

COZCH3 C02CH3 COZCH3

CH3OHJ l

 

E”Loam
0 be... 68/
1:, OCH3 D OCH3
H
—>
COZCH3 COZCH3 g.

Figure 2.4.10. Mechanism for deuterium incorporation in DMS
The methanol Obtained during the condensation is responsible for the IUD exchange
between Dz-CA and DMS. It abstracts a deuterium from Dz-CA and then donates this
deuterium to a DMS molecule, thus serving as a carrier for the deuterium between Dz-CA
and DMS. This proves that for the regular catalytic Stobbe condensation (no deuterated
Species involved) the hydrogens fiom citraconic anhydride are continuously exchanged

with the ones from C2 and C3 in DMS.

2.5. Experiments involving CHZO and D3COCOCH2CH2COOCD3
2.5.1. Introduction

In classic Stobbe condensations, in the first sequence of the mechanism, a
methoxy group is liberated and in the second sequence it abstracts a proton from the

intermediary paraconic acid. There is no report in the literature about hydrogen atoms

45

from this methoxy group being incorporated in the condensation product. In order to
verify whether this is also true for Stobbe-like condensation under heterogeneous
catalytic conditions, experiments using CHZO and DMS deuterated at the ester groups,
D3COCOCH2CH2COOCD3 (D6-DMS), were performed. The reaction is expressed by the
following scheme (Figure 2.5.1.) :

O O

H3C

_ OCD3
HZC—O + OCD3 A1203 0

Figure 2.5. 1. Reaction between CHZO and D3COCOCH2CH2COOCD3
The results showed unexpected deuterium incorporation in citraconic anhydride.
Also, the analysis of the recovered DMS revealed the incorporation of deuterium in

positions other than the ester groups.

2.5.2. Experimental methods and results

Since D3COCOCH2CH2COOCD3 is not commercially available, it was
synthesized10 from CH3OD and succinic acid. The D3COCOCH2CH2COOCD3 product
was analyzed by GC-Mass spectrometry using a Hewlett Packard 5890 Series 11 gas
cromatograph (SPB-l non-polar column) connected with a VG Trio-1 mass spectrometer.
A comparison between the mass Spectra of standard DMS and of the synthesized
D3COCOCH2CH2COOCD3 is presented in Figure 2.5.2. The deuterium incorporation is

proved by the peak corresponding to the m/z = 118 fragment (Obtained by loss of a

46

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

UG LAB-BASE The TRIO-1 GC-flS Data System
9002101100 molcgular ueight GC Instrumcrrtil’rio-l
07611 14 41 .884)
,“ul 115.1 2670592
55.1
59.1
zrs-
114.1
\
07.1
116.1
41.2\4sl.2 ll 86-1\’ea.1 113.1 I’
0 . .Hl. . .L. . - 391:}. .‘n .
wing at; 59 69 10 09 90 100 119 120 ' 130 ' 159 ' 1:0
06 LAB-BASE The TRIO-1 GC-HS Data System
SanFletlou molecular Ucifiht GC lnstmmentflrio-l
H . H N : to - o 0 0 fl .
1001 55-9 62.. 3409792
1111.0
90.0
zrs‘
54 0 117.0
. \ \
19.0
40.0 /‘
l 09 0 101.0
99.0
,, 1.1111111 11;- , 491311.. . r ,1 , , , . ,
"I: 40 '50 62 ' 70 00 32 Q0 “0 15:0 0
"/2 29 «a 6'9 '43! M 15.9. AMA—13.0..

 

 

 

 

Figure 2.5.2. Comparison between the mass Spectra of standard DMS (top) and of

D3COCOCH2CH2COOCD3 synthesized (bottom)

methoxy group from D3COCOCH2CH2COOCD3), which replaces the peak
corresponding to the m/z=115 fragment.

The experiments were performed at 380°C in the set-up presented in Figure 2.1.1.
The molar ratio of the reagents was CHZO : D3COCOCH2CH2COOCD3 = 2 : 1. The flow
rate of the reagent mixture was 0.3 mL/min. Nitrogen was used as the carrier gas at a flow
rate of 20 mL/min., controlled by a Cole-Partner rotameter. The catalyst, alumina SA
3177, 30-60 mesh in size, was heated in the reactor at 380-400°C for at least 3 hours to
eliminate moisture before the reagent mixture was added. The product mixture was
passed through an air condenser for cooling purposes and collected in glass vials every 7
minutes for 35 minutes.

The product mixture was analyzed by GC-Mass spectrometry using a Hewlett
Packard 5890 Series 11 gas cromatograph (SPB-l non-polar column) connected with a
VG Trio-1 mass spectrometer.

The monitored compounds were: citraconic anhydride and recovered DMS . Using their
mass spectra obtained for each sample, the change in the content Of various deuterated
species was obtained and is presented in Figure 2.5.3. for citraconic anhydride and in
Figure 2.5.4. for DMS.

In the citraconic anhydride case, the non-deuterated species, which should have
been the only CA species present, has the highest concentration all the time, except for
the first sample, where is overtaken by the Dl-CA species. The Dl-CA species has a

considerable concentration all the time, between 45% and 25%. The Dz-CA species has a

48

 

.5683: 0529.0

<0 3 858983 8:08:96 00 Swim .m.m.m 2:»E

 

 

 

 

mm on ma om AEEV OS: 0.
. . I . .
AT, I I
-
A
<08 :01
<98 IT
<96 Iil
<98+

 

 

 

 

.. ON

.. om

.. ov

 

 

8F

°/o Z/ur

 

 

49

mED 5 00:82:85 8:08:00 .8 Eme {Wm 8:wE

26883.”: 0.529.:

 

 

 

om

fir—

mm ow

u
_ q

+1

 

 

msfiéo II
mzofio IT
£20-00 1.: .

 

 

8:5 65:9

_
_

ill?!

    

.0.

Tom

wow

wow

wow

.on

row

row

 

 

8..

"/6 2/w

 

 

50

significant concentration (20%) only for the first sample, then its concentration becomes
low (about 10%). The D3-CA species has a very low concentration all the time.

In the deuterated DMS case, the D6-DMS has the highest concentration all the
time. The D7-DMS and Ds-DMS species have a low concentration.

The deuterium incorporation in citraconic anhydride and D6-DMS is surprising. A
possible explanation comes from the fact that a small amount of DMS is usually
hydrolyzed to monomethyl succinate, thus liberating methanol, in this case CD3OH. The
oxidation of CD3OH allows the formation of CD20, which is then incorporated in
citraconic anhydride, which in turn can be involved in a H/D exchange with DMS, as
shown previously.

For a better understanding, the experiment was repeated in the same conditions, but using
only D3COCOCH2CH2COOCD3 as a reagent.

The product samples were analyzed by GC-Mass spectrometry using a Hewlett
Packard 5890 Series 11 gas cromatograph (SPB-l non-polar column) connected with a
VG Trio-1 mass spectrometer. The results are presented in Figure 2.5.5.

The D6-DMS species has the highest concentration all the time. The D7-DMS and
Ds-DMS have a low concentration, showing a low extent of H/D exchange.

Since only Dé-DMS was used, and still, the recovered DMS composition showed that
H/D exchange took place, then somehow, deuterium from the methoxy group replaced
the hydrogen from D6-DMS. A possible explanation comes from the hydrolysis of a small
amount of D6-DMS to the corresponding monomethyl succinate, thus liberating

methanol, in this case CD3OH. As stated before, the oxidation of CD3OH allows the

51

2025qu 333 mzosa
mEQ 5 "8920985 Estfisou mo :.me .m.m.m oSwE

 

 

 

 

 

 

 

 

 

 

 

 

mm on mm ONE-5 25.—m— 2 m o
0 h a p ._ .. “ plLllll o
4 i 9 l
a
.. o.
. om
. on
. 8
9298+ w
/
$5.51!! .. on J
o/
wzoeoif. .. 8
.. oh
\\*\‘\\\.‘Tl!l’lll/ .- on
,_..lillxl\\\ .. 8
8.

 

 

52

formation of CD20, which is then incorporated in citraconic anhydride, which in turn can
be involved in a H/D exchange with DMS.

Since the GC-Mass spectrometry analysis did not detect any citraconic anhydride,
an HPLC analysis was tried. The HPLC column used was BIO-RAD 5, connected to a
Waters 410 refractometer. The eluent used was 20% acetonitrile in 0.005 M H2804.
Oxalic acid was used as internal standard. The HPLC results showed, indeed, the
presence of citraconic anhydride in very small amount in the product mixtures: 0.2% in
the third sample, 0.3% in the first and fifth samples and 0.5% in the second and fourth

samples. This proves the proposed explanation stated previously.

2.6. Proposal of a mechanism for catalytic Stobbe condensation.

Based on the above experiments, a mechanism for the Stobbe condensation
between formaldehyde and DMS, using y-alumina as catalyst, is proposed.

The Lewis acid site catalyzes the loss of a proton from DMS (Figure 2.6.1.) to a
methanol molecule, followed by the attack of the corresponding anion on the carbon of a
formaldehyde molecule. A carbon-carbon bond is formed and the oxygen from
formaldehyde abstracts the proton from the CH3O+H2 species, to form a hydroxy diester.

The lone pair of electrons from the hydroxy! group attacks the carbonyl group
(Figure 2.6.2.) and closes a ring, which is stabilized as a lactone derivative by the loss of
a methoxy group and of a proton. The loss of a proton from the resonance form of the
lactone derivative and detachment from the Lewis acid site allows the formation of

monomethyl itaconate.

53

 

(+3 (i-Cat
L-cat
OCH3 L. OCH3
‘—
H C02CH3 H C 02CH3

Figure 2.6.]. Proposed mechanism for catalytic Stobbe condensation. First step.

 

Figure 2.6.2. Proposed mechanism for catalytic Stobbe condensation. Second step.

54

 

-cat.

 

O L-cat. 0
OCH3 (31130H H2C
<+>o\ o
H
U H
o <'
CH3oH “a"
o
CH3OH‘Ji
CO:\[_,-- cat.

Figure 2.6.3. Proposed mechanism for catalytic Stobbe condensation. Third step.
The attack of the lone pair of electrons from the oxygen of the hydroxyl group, on
the carbonyl group, followed by loss of the methoxy group and of a proton, allows the
formation of itaconic anhydride which is then isomerized to citraconic anhydride, a

molecule of methanol serving as carrier for the proton (Figure 2.6.3.).

55

 

2.7. Summary

Ever since it was first reported, the classic Stobbe condensation reaction proved
to be special, because it was limited to succinate esters. Chemists struggled and
succeeded in solving the mechanism of this condensation. The classic Stobbe
condensation is a stoechiometric reaction, which uses a strong base as a condensing
agent. It is used for the synthesis of many useful compounds.

Recently chemists became interested in performing this reaction in a catalytic
fashion. Up to now, the only carbonyl compound that offered good results was
formaldehyde. The condensation product, citraconic anhydride, can be easily converted
into itaconic acid, a valuable monomer.

Since no information about the mechanism under catalytic conditions was
available, we decided to investigate it. For this purpose, experiments involving
deuterium-labeled compound were performed, using y-alumina as the catalyst. The
products were analyzed by GC-Mass spectrometry and the extent of deuterium
incorporation was determined. The experiments proved the enolization of DMS and
showed that the Lewis acid sites on the catalyst surface are responsible for the catalytic
activity, in this case. Mechanisms were proposed to explain each H/D exchange process.

Finally, a mechanism for the condensation of formaldehyde with DMS, using

y-alumina as catalyst, was proposed.

56

2.8. References :
1. Dushyant, S., Ph.D. Thesis, (tentatively 2000), MSU. He determined the conditions
that afforded the best results for the reaction between formaldehyde and DMS over y-

alumina.

2. Thomas, J. M.; Thomas, W. J. Principles and Practice of Heterogeneous Catalysis,
VCH, Weinheim, 1997, 266-268.

3. Anderson, J. R.; Boudart, M. Catalysis, Science and Technology, Springer-Verlag,
Berlin, 1981, 2, 182-184.

4. Nagarajan, K., MSU, unpublished results.

5. Moffat, J. B. Theoretical Aspects of Heterogeneous Catalysis, Van Nostrand Reinhold,
N.Y., 1990, 508-510.

6. Fort, A.W.; Patterson, J. M. J. Org. Chem, 1976, 41, 23, 3697-3671.
7. Love, C. J .; Mcqiullin, F. J. J. Chem. Soc. Perkin I, 1973, 2512-2515.

8. Lambert, J. B.; Shurvell, H. F. Organic Structural Spectroscopy, Prentice Hall, 1998,
448-449.

9. Grimminger, W.; Kraus, W. Liebigs Ann. Chem, 1979, 1575-1578.

10. Vogel, A. Vogel ’s Textbook of Practical Organic Chemistry, John Wiley&Sons, Inc.,
NY, 1978, 508-509.

57

llliltlll’lllllllflllflllllll