. . .. «2‘ 2...}. n4...)
$.33? 94mg u.k hum.
a in ha
”.55.. .
. . .
. w ..
.fin‘mfi .
2%.”.9.

.. xetzn
232.

it

3.1 2 .~
it}... .1:
3 .

Minna

nHerf}1:
. an... .
l

.«
tau . w
r. is." turf :
. , 53.. .-
:ch...r2x..
z? .2...

.. a
i.

. . .ld .. r
. , 3t 1.3 {Fugutrt
s ._ MN.“ ‘.hfl}i¥.flfi
: 1 .{t Iv
fi” . I. 1‘ ,
n’f‘X . ‘
Erumfikgfi..." 1.
Mg»?! ha...
. ~..... 3
. ;
. 1“
. r.
2..
. .2 .
..hr.......:.
1. a... vi»
Raw:
.I.I.-’~l
:1~vx..~ 3.50:... xv
“VI-r“ <5),

1
a
3.:

33:0... .

flu}?!

_. 5...:

D! .r
.

4 ”9....
1335......ilexl’.
i» av .Wetiium .rfllu.
IIV\ I 7“ o a

v

is.
2251.
.5. :3.

lxif)
:‘ . .QL

 

:3!!!
v. J... ..r

I‘Otfilill

THESiS

‘g't/LL'

This is to certify that the

dissertation entitled

SYNTHESIS AND CHARACTERIZATION
OF SUBSTITUTED POLYLACTIDES

presented by

MAO YIN

has been accepted towards fulfillment
of the requirements for

 

DOCTOR degree in W

51

Major professor

   

  

Date /3 S/‘v/hé 3000

 

MSU is an Affirmative Action/Equal Opportunity Institution 0-12771

 

 

LIBRARY
Michigan State
University

 

 

 

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

 

 

magma

 

 

 

 

 

 

 

 

 

 

 

 

6/01 chlRC/DatoprSS-DJS

 

SYNTHESIS AND CHARACTERIZATION OF SUBSTITUTED

POLYLACTIDES

By

Mao Yin

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Department of Chemistry

2000

sites-55$

E bfcac

ABSTRACT
SYNTHESIS AND CHARACTERIZATION OF SUBSTITUTED
POLYLACTIDES
BY

Mao Yin

Because of their biocompatibility and environmental degradability,
polymers derived from lactic acid have long been important materials for
medical applications. Once used primarily as sutures and implants, polylactide
materials are now entering high volume areas such as packaging. To be
successful in these new applications, polylactides must be available that exhibit
a broad spectrum of physical properties while retaining the degradability of the
parent polymers. This objective can be achieved by altering the substituents on
the polylactide chains.

A variety of substituted Iactides were synthesized and polymerized to
high molecular weight polymers using solution and bulk polymerization. The
polymers show a wide range of physical properties: glass transition
temperatures range from —37 °C to 85 °C, hydrolytic degradation rates are 2-3
times slower than polylactide's, and degradation temperatures are around 350
°C. We established the relationship between the pendant group and the physical
properties of the polymers.

The kinetics and mechanism of lactide polymerization was studied.

Lactide polymerizations are first-order reactions in both solution and bulk

AHMAD?
:I-v't' Alibi

:5; 5w “R

v? ‘v U ii
:‘i’erenl

Bali :3

rm

“ «A1;
Dewar,

Tile a:

Wis

polymerization. In bulk polymerization, polymerization rates follow the order of
size of substituted group, the bigger the substituted group, the slower the
polymerization rate. In solution polymerizations, the polymerization rates do not
follow the order of size of substituted group, which is probably caused by
different rate-determining steps for the polymerizations. We proposed that
lactide polymerization mechanism is similar to “atom-transfer polymerization”
when initiated by Sn(Oct)2. Using it as guide, we achieved control of the
molecular weights and molecular weight distributions of the polymers.

We studied the influence of stereochemistry on the kinetics of
polymerizations and the physical properties of the polymers. We found that
racemic monomers polymerize faster than either of the optically pure
monomers. The racemic polymers are usually amorphous polymers and
optically pure polymers are usually crystalline polymers.

Copolymerization of lactide with substituted Iactides were investigated.
The architectures of polymers depend on the size difference of substituted
groups between the lactide and substituted lactide. By controlling the size of
substituted group, random, block and alternating copolymers can be

synthesized.

To my wife Tianle

ACKNOWLEDGEMENTS

I wish to express my deep appreciation to Professor Gregory L. Baker for
his guidance, assistance and constant encouragement throughout the course of
this research. A work such as this would not be possible without his assistance. It
has been my privilege to have him as my mentor.

I would also like to thank professor James Jackson, Milton Smith, Gary
Blanchard, and Thomas Pinnavaia for their guidance and assistance. My thanks
go to all Baker group members: Jun Qiao, Jun Hou, Jingpin, Yiyan, Gao, Corey,
Susan, Tara, Tianqi, J, 3,, Kirk, Fadi and all the friends in Michigan State
University, for making my stay an enjoyable one

Finally, I want to thank my wife Tianle for her love and sacrifices. I thank

my parents and sister for their love and encouragement.

138'. OF Ii
US? OF Fl
CST OF S

llTiDDU

3.1
3.2

3.2

TABLE OF CONTENTS

LIST OF TABLE ................................................................................................... ix
LIST OF FIGURES .............................................................................................. xi
LIST OF SCHEMES ............................................................................................ xvi
INTRODUCTION .................................................................................................. 1
1. Applications .............................................................................................. 1
1.1 Medical application ................................................................................ 1

1.2 Biodegradable Packaging Therrnoplastics ............................................. 3

3. Ring-opening polymerization of Iactides and Iactones ............................ 11
3.1 Polymerizability .................................................................................... 11
3.2 Polymerization Mechanism .................................................................. 15
3.2.1 Cationic Polymerization ................................................................. 15
3.2.1.1 Lactones ................................................................................... 15
3.2.1.2. Lactides ................................................................................... 16

3.2.2 Anionic Polymerization ................................................................... 21
3.2.2.1 Lactones ................................................................................... 21
3.2.2.2 Lactide ...................................................................................... 24

3.2.3 Coordination Polymerization .......................................................... 27
3.2.3.1 Aluminum Derivatives ............................................................... 27
3.2.3.2 Tin Derivatives .......................................................................... 38
3.2.3.3 Rare Earth Metal Compounds .................................................. 46

3.2.4 Other Polymerization methods ....................................................... 48

vi

42

43

52.

RESUL'

32
33
Fwy
34.
34
3.4
35.

poly;

3.2.4 Other Polymerization methods ....................................................... 49

 

Modification of Polylactide Properties ..................................................... 51
.1 Manipulation of stereochemistry. ......................................................... 51
.2 Copolymerization ................................................................................. 55
.3 Blending ............................................................................................... 63
Degradation of the polylactide and polylactones ..................................... 64
.1 . Thermal degradation ......................................................................... 64
.2. Biodegradation .................................................................................. 66
JLTS AND DISCUSSION ........................................................................... 69
Monomer synthesis ................................................................................. 69
Bulk polymerization of substituted Iactides ...................................... . ...... 74
.1 Initiator survey ...................................................................................... 75
.2 Kinetics of Bulk Polymerization ............................................................ 83
Solution Polymerization of Substituted Lactides ..................................... 96
.1 AI(OiPr)3 as Initiator ............................................................................. 96
.2 Sn(Oct)2/ROH as initiators ................................................................. 106

.3 Comparison of AI(OiPr)3 and Sn(Oct)2lROH as lnitiators in Solution

olymerization ............................................................................................ 123
4. Polymerization Mechanism ............................................................. 129
3.4.1 AI(OiPr)3 as initiator ...................................................................... 129
3.4.2 Sn(Oct)2/ROH as catalystfinitiator ................................................ 133

5. The Influence of Stereochemistry on the Kinetics of Solution

olymerization ............................................................................................ 137

vii

 

4. Pair

5. Cap:

52
5.2.
5.2

5.3

6.C

6.2

6.3

EXPERI

BIBLio-c

4. Polymer properties ..................................................................................... 150

5. Copolymerization of substituted glycolides ................................................ 161
5.1. Copolymerization through comonomers ......................................... 162
5.2 Copolymerization through asymmetric monomers. ............................ 168

5.2.1. Synthesis and Polymerization of AB Monomers. ...................... 170
5.2.2 Structure of the polymer chains. .................................................. 175
5.3 Thermal properties of the copolymer. ................................................ 179

6. Crystalline substituted polylactide. ........................................................ 183
6.1 Synthesis and polymerization of optical pure monomers ................... 183
6.2 The stereochemistry of the polymer chain ......................................... 187
6.3 The crystallinity of poly(isopropyglycolide) ......................................... 196

7. The degradation of substituted polylactides .......................................... 202

EXPERIMENTAL .............................................................................................. 212
BIBLIOGRAPHY ............................................................................................... 224

viii

Table 1.

Table 2.

Table 3.

Table I.
Table 5.
Table 6.
Table 7.
Table 8.
Table 9,
Table 1(
Table 11
Table 1;

Table 1;

Table 14

Table 15

Table Ti

Table 1.

Table 2.

Table 3.

Table 4.
Table 5.
Table 6.
Table 7.
Table 8.
Table 9.
Table 10.
Table 11 .
Table 12.
Table 13.

Table 14.

Table 1 5.

Table 16.

LIST OF TABLES

Some Basic Physical Properties Comparisons of Poly(lactic acid) and
Polystyrene .......................................................................................... 5
Comparison of Polyethylene to Poly(lactic acid) .................................. 6

Heats of combustion and Strains of Cycloalkanes per Methylene

group ................................................................................................. 13
Polymerizability of cyclic ester monomers ......................................... 14
Bulk Polymerization of Ethylglycolide ................................................ 78
Bulk Polymerization of Isobutylglycolide ............................................ 79
Bulk Polymerization of Hexylglycolide ............................................... 80
The kinetic data for bulk polymerization of substituted Iactides ......... 90
Polymerization rate constants of substituted Iactides ...................... 104
Polymerization activation energies for substituted Iactides .............. 104

Polymerization rates for ethylglycolide initiated by Sn(Oct)2/BBA....114
The activation energy for ethylglycolide initiated by Sn(Oct)2/BBA..114
The activation energy of ethylglycolide initiated by Sn(Oct)2/BBA and
AI(OiPr)3 .......................................................................................... 124
Experimental value and calculated value for activation energy
difference. ........................................................................................ 146
Polymer Properties .......................................................................... 153
Glass transition temperature of substituted polylactides and

substituted polyethylene. ................................................................. 159

Table 17.
Table 18.
Table 1 9.

Table 20.
Table 21 .

Properties of copolymers ................................................................. 164
The properties of AB polymers. ....................................................... 180
Experimental and calculated values of hexad intensities in the

carbonyl region of 13C NMR spectra of poly(rac-isopropyglycolide). 191
The crystallinity and heat of fusion of poly(isopropylglycolide) ........ 197

The degradation of substituted polylactide ...................................... 205

Figure I. E

Figura 2. E
Figure 3. I
Figural. '
Figure 5, I
Figure 6. :
Figural. '
FigureB.

F'QuraS.

FlQuail).
Flaurrerr.

FlSure 12,

FlgUTe 13.

Figure 14‘

Flgllre 15

LIST OF FIGURES

Figure 1. Structures and abbreviations of monomers and degradable polymers

............................................................................................................. 2
Figure 2. Structure of Al initiators ...................................................................... 27
Figure 3. Aggregation states for AI alkoxides .................................................... 30
Figure 4. The structure of porphyrin initiators ................................................... 37
Figure 5. Structure of Sn(Oct)2 and Bu28n(Oct)2 .............................................. 46
Figure 6. Stereochemistry of lactide ................................................................. 53
Figure 7. Tacticity of polylactide ........................................................................ 54
Figure 8. Copolymer architectures .................................................................... 56
Figure 9. Degradation models ........................................................................... 67
Figure 10.The structure of substituted glycolides ............................................... 71
Figure 11.‘H NMR of ethylglycolide and poly(ethylglycolide) ............................. 73
Figure 12.8qu polymerization of ethylglycolide with and without added t-
butylbenzyl alcohol as co-initiator ...................................................... 81
Figure 13.Molecular weight versus conversion for the bulk polymerization of
substituted glycolides ......................................................................... 82
Figure 14. Kinetics of bulk polymerization of substituted glycolides .................... 84
Figure 15. Kinetics of bulk polymerization of substituted glycolides at low
conversion ......................................................................................... 85
Figure 16.The structure of the white precipitate formed during the solution
polymerization of ethylglycolide using Sn(Oct)2/ROH as initiator ....... 90

xi

Figure 17
Figure 18

Figure 19

Figure 20

Frgura 21

Figure 22

Figure 23

Figum 24

Filure 25

Figure 25

FiWe 28

Film 29

Figure 17.Kinetics of bulk polymerization of substituted glycolides .................... 91
Figure 18.Kinetics of bulk polymerization of substituted glycolides .................... 92
Figure 19.Polymerization/depolymerization data for the bulk polymerization of
ethylglycolide ..................................................................................... 93
Figure 20.Molecular weight vesus conversion fro the solution polymerization of
substituted Iactides .......................................................................... 100
Figure 21.Kinetics for the solution polymerization of substituted Iactides at 70 °C
......................................................................................................... 1 01
Figure 22.Kinetics for the solution polymerization of substituted Iactides at 90 °C
......................................................................................................... 1 02
Figure 23.Kinetics for the solution polymerization of substituted Iactides at 100
°C ..................................................................................................... 103
Figure 24.Activation energies for polymerization of substituted Iactides .......... 105
Figure 25.Kinetics for the polymerizaiton of ethylglycolide initiated by
Sn(Oct)2/ROH and Sn(Oct)2 ............................................................ 109
Figure 26.Solution polymerization of ethylglycolide at different Sn(Oct)2/alcohol
ratios ................................................................................................ 1 10
Figure 27.Solution polymerization of ethylglycolide at different Sn(Oct)2/alcohol
ratios ................................................................................................ 111
Figure 28.Solution polymerizaiton of ethylglycolide at different monomer/initiator
ratios. ............................................................................................... 112
Figure 29.Solution polymerizaiton of ethylglycolide at different monomer/initiator

ratios .......................................................................................... . ...... 113

xii

Figure 3
Figure!
Figure 3

Figure E

Figure I

Figure

Figure

Figura

FlgUfe

Figure

Figure

FlgUre

FigUfe

e 30.Solution polymerization of ethylglycolide initiated by Sn(Oct)2/BBA 115
e 31.Solution polymerization of ethylglycolide initiated by Sn(Oct)2/BBA 116
e 32.The activation energy of ethylglycolide initiated by Sn(Oct)2/BBA117
e 33.IR spectra for white precipitate formed during the polymerization of
ethylglyolide and by mixing 2-hydroxybutyric acid and Sn(Oct)2 ...... 120
e 34. Kinetics of solution polymerization of ethylglycolide showing the result
after adding extra monomer after polymerization reached equilibrium
......................................................................................................... 1 21
e 35.1nitial polymerization rate for ethylglycolide and the decreased rate
observed after adding additional monomer ...................................... 122
e 36.Solution polymerization of substituted Iactides initiated by
Sn(Oct)2/BBA ................................................................................... 125
e 37.Solution polymerization of ethylglycolide initiated by AI(OiPr)3 and
Sn(Oct)2/BBA ................................................................................... 126
e 38.Solution polymerization of ethylglycolide initiated by AI(OiPr)3 and
Sn(Oct)2/BBA ................................................................................... 127
e 39.The activation energies for polymerization of ethylglycolide initiated by
Sn(Oct)2lBBA and Al(OiPr)3 ............................................................. 128
e 40.The structure of the transition state for polymerization of substituted
Iactides ............................................................................................ 132
e 41 .Polymerization of rac—Iactide and L-Iactide ...................................... 139
e 42.Ponmerization of rac-isobutylglycolide and D-isobutylglycolide ...... 140

e 43.The structure of complexes used to simulate polymer growth ......... 147

xiii

 

 

Figure 4

Fgural

Fgum
Figure
Figure
Figure
Fgum
figure

FMUm

FlgUTE

qu“
FQUN
FlQurr
Fl91m
Flgun

FlQUrr

FiQUIT

Figure 44.The calculated structure for complex with R configuration chain end
and 8,8 Iactide ring .......................................................................... 148

Fgure 45. The calculated structure for the complex with S configuration chain

and 8,8 Iactide ring .......................................................................... 149
Figure 46.DSC run for substituted poly(glycolide)s .......................................... 152
Figure 47.Thermal analysis results for poly(ethylglycolide) ............................. 154
Figure 48.Thermogrvimetric analysis results for poly(ethylglycolide) ............... 155

Figure 49.Thermogrvimetric analysis results for substituted poly(glycolide)s .. 156
Figure 50.The secondary interaction between polyester chains ...................... 160
Figure 51.DSC runs for Iactide and ethylglycolide copolymers. ....................... 165
Figure 52.Glass transition temperature of Iactide and ethylglycolide copolymers
......................................................................................................... 166

Figure 53.1H NM/R spectra of ethylmethylglycolide and methylphenylglycolide

......................................................................................................... 1 73
Figure 54.Structure of the diastereomers of methylphenylglycolide ................. 172
Figure 55.NOE NMR spectrum of methylphenylglycolide ................................ 174
Figure 56.Carbonyl region of ”C NMR of AB copolymers ............................... 178
Figure 57.DSC runs for AB polymers ............................................................... 181
Figure 58.Thennogrvimetn'c analysis results for AB polymers ......................... 182
Figure 59.NMR spectra of the methane and carbonyl region of poly(S-

isopropylglycolide) ........................................................................... 186
Figure 60.Schematic representation of the Bernoulli and first-order Markov

models ............................................................................................. 188

xiv

Figure 61.05

pr
Figure 5233“
figure 63}
Figure 64.8

is
Figure 65.x
Figure 66, X
Figure 67.P
Figure 68,:

is

Figure ear

Figure 70.1

C
Flgme 71.1
I

I

Figure 73.r

Figure 61.DSC runs showing the melting point for poly(isopropylgcholide)s

prepared from S and rac-isopropylgycolide ..................................... 192
Figure 62.130 NMR of poly(isopropylglycolide) ................................................ 193
Figure 63.1H NMR of poly(isopropylglycolide) ................................................. 194

Figure 64. Stereosequence assignment for 13C spectrum of poly(rac-
isopropylglycolide) ........................................................................... 195
Figure 65.X-ray diffraction pattern for crystalline poly(S-isopropylgcholide).... 198
Figure 66.X-ray diffraction pattern for amorphous poly(rac-isopropylgcholide)199
Figure 67.Peak deconvolution for crystalline poly(S-isopropylglycolide) .......... 200
Figure 68.Determination of heat of fusion for 100% crystalline poly(S-
isopropylglycolide) ........................................................................... 201
Figure 69.The weight loss of substituted glycolide during hydrolytic degradation
......................................................................................................... 207
Figure 70.The molecular weight loss of substituted glycolide during hydrolytic
degradation ...................................................................................... 208
Figure 71.The molecular weight decrease of substituted Iactides during
hydrolytic degradation fitted to random chain scission model. ......... 209
Figure 72.The decrease in the degree of polymerization of substituted Iactides
during hydrolytic degradation ........................................................... 210

Figure 73.Contact angles on substituted polylactides plotted as 7 (surface

tension of water/methanol solutions) vs cos 0 ................................. 211

Scheme '
Scheme
Schema
Scheme

Scheme

Scheme

Scheme
Scheme
Scheme
Schem‘
Schema
Schema
Schem.
Schem.
Schilmr
scheTTIr
smemr

s°hemr

LIST OF SCHEMES

Scheme 1. Synthetic route for poly(lactic acid) .................................................. 10
Scheme 2. Cherdron‘s cationic polymerization mechanism ............................... 16
Scheme 3. The cationic polymerization mechanism proposed by Penczek ....... 17
Scheme 4. Cationic polymerization of Iactide initiated by protonic acids ............ 18
Scheme 5. Polymerization of Iactide initiated by methyl triflate (traditional cationic

mechanism) ....................................................................................... 19

Scheme 6. Polymerization of Iactide initiated by methyl triflate (mechanism

proposed by Kricheldorf) ................................................................... 20
Scheme 7. Propagation routes for anionic polymerization of Iactones ............... 22
Scheme 8. Polymerization mechanism of B-PL .................................................. 22
Scheme 9. The polymerization mechanism of PL proposed by Jedlinski ........... 24
Scheme 10. Anionic polymerization of Iactide through deprotonation ................ 25
Scheme 11. Anionic polymerization of Iactide .................................................... 26
Scheme 12. Mechanism of Iactide polymerization using AI alkoxides ................ 28
Scheme 13. Dissociation of the AI alkoxide aggregates ..................................... 31
Scheme 14. lntramolecular transesterification ................................................... 33
Scheme 15. Intermolecular transesterification ................................................... 34
Scheme 16. Preparation of end-functioned Al initiators ...................................... 35
Scheme 17. Mechanism for tin alkoxide initiated polymerization Iactides .......... 39

Scheme 18. Mechanism for tin halide initiated polymerization of caprolactone..40

Scheme 1'.

Scheme 2

Scheme 2

Scheme 1
Scheme 2
Scheme L
Schema
Scheme
Scheme
SCheme
Scheme
SCheme

Scheme

SCheme

Scheme

Scheme 19. Mechanism for polymerization of Iactide initiated by Sn(Oct)2 as
proposed by Nijenhuis et aI.. ............................................................. 42

Scheme 20. Mechanism for polymerization of Iactide initiated by Sn(Oct)2 as
proposed by Kricheldorf ..................................................................... 44

Scheme 21. Mechanism for Iactide polymerization initiated by Sn(Oct)2 as

proposed by Penczek ........................................................................ 45
Scheme 22. Polymerization of Iactide initiated by rare earth catalyst ................. 48
Scheme 23. initiators made from epoxides and rare earth halides ..................... 48
Scheme 24..Zwitterionic polymerization mechanism .......................................... 50
Scheme 25. Synthesis of diblock copolymers .................................................... 59
Scheme 26. Synthesis of ABA triblock copolymers ............................................ 60
Scheme 27. Synthesis of cross-Iinkable copolymers ......................................... 62
Scheme 28. Thermal degradation via cis—elimination ......................................... 65
Scheme 29. Radical thermal degradation .......................................................... 66
Scheme 30. The synthetic routes to substituted glycolides ................................ 72

Scheme 31. The propagation step for Iactide polymerization using Sn(Oct)2 as
catalyst .............................................................................................. 95
Scheme 32. The first step for the Iactide polymerization using Sn(Oct)leOH . 108
Scheme 33. The mechanism for the Iactide polymerization initiated by Al(OiPr)3
......................................................................................................... 1 30
Scheme 34. The mechanism of Iactide polymerization using Sn(Oct)2/ROH as
catalyst/initiator ................................................................................ 135

Scheme 35. The mechanism for atom—transfer radical polymerization (ATRP) 136

Scheme 36. 5
Scheme 37. 1
Scheme 38.
Scheme 33.
Scheme 40.
Scheme 41.
Scheme 42.
Scheme 43.

Scheme 44,

81

Scheme 36. Stereochemistry of Iactide polymerization .................................... 138

Scheme 37. Kinetic scheme for Iactide polymerization .................................... 142
Scheme 38. Copolymerization of substituted glycolides .................................. 161
Scheme 39. Reactivity difference leads to “blocky” copolymers ....................... 167
Scheme 40. The comparison between two copolymerization methods ............ 169
Scheme 41. Structures of AB monomers and reactivity of different sites ......... 169
Scheme 42. Synthesis of AB glycolide monomers ........................................... 172

Scheme 43. The propagation step in the polymerization of AB monomers ...... 177
Scheme 44. The synthesis of an optically pure substituted Iactide form an amino

acid precursor....... ............................................................................ 185

xviii

r. Applic

1.1 Medic

Be:
DCTTQTycoi
in Figure
YET-all ar
3.5333110
surgical c
The bore
and lutur
"399039
bl Davis
”Wynn

T':
de‘lelOpg

A DOTOU

INTRODUCTION

1. Applications

1.1 Medical application

Because of their biodegradability and biocompatibility, polylactide,

polyglycolide and polylactones (monomer and polymer structures are displayed
in Figure 1) have been used in medical applications for wound closure,1 tissue

repair and regeneration,2'4 and drug delivery.5‘7 The most successful
application has been wound closure, where biodegradable surgical sutures and
surgical clips have been made using polylactide, polyglycolide and polylactones.
The bioresorption of sutures eliminates the possibility of foreign-body reactions
and future infection. Some commercially available biodegradable sutures such as
the polyglyconate sutures (glycolide/trimethylene carbonate copolymer) marketed
by Davis 8. Geck as Maxon and polyglecaprone 25 (glycolide/caprolactone

copolymer) have been used successfully in clinics.

Tissue repair and regeneration and drug delivery applications are less
developed than wound closure, but there has been exciting progress in the field.
A porous biodegradable membrane made from DL-PLA was approved by the
FDA for periodontal use. Clinical results show that the new biodegradable
membrane, marketed in the U. S. as Guidor by the Butler Company, gives
comparable if not better periodontal tissue regeneration than the

nonbiodegradable Gore-Tex products made from poly(tetrafluoroethylene). Many

degradable products have been developed to deliver a variety of drugs, such as
anti-cancer agents, narcotic antagonists, antibiotics, fertility drugs and vaccines.
Some products have been marketed such as Lupron Depot, an injectable
microcapsule system marketed by Takeda-Abbot using a 70:30 DL-
lactide/glycolide copolymer to deliver leuprorelin (anti-cancer drug). However, the

majority of products are still in the evaluation stage and more research needs to

be done.
0 We}
(<50 0 m
n = 1: 8-PL, 8-propiolactone n = 1: Poly(B-PL), Poly(B-propiolactone)
n = 2: y-BL, y-butyrolactone n = 2: Poly(y-BL), Poly(y-butyrolactone)
n = 3: 8-VL, 8-valeroiactone n = 3: Poly(8—VL), Poly(8-valerolactone)
n = 4: e-CL, e-caprolactone n = 4: Poly(e-CL), Poly(e-caprolactone)

.6; M.

Glycolide Poly(glycolic acid) or PGA
O
O
O
\n/ik m
0
Lactides: Polylactides or PLA:

(L-lactide, D-Iactide, DL-Iactide, mesa-Iactide) Poly(L-Iactide). Poly(D-Iactide).
Poly(DL-lactide), Poly(meso-lactidej

Figure 1. Structures and abbreviations of monomers and
degradable polymers

1.2 Bior

ararsv

USpi

plastics
become
Tia Ila
mamm
Concer
wflihar

mph

W338.

Therm
lbOgg
Se'r'er
aha
Sdlo

Waier

lTaie

Dacgi

1.2 Biodegradable Packaging Thermoplastics

There is a need for an environmentally benign packaging thermoplastic as
an answer to the tremendous amounts of discarded plastic packaging materials.

U. S. plastic sales in 1997 were 74.0 billion pounds of which 29% were listed as

plastics in packaging.8 A significant amount of this plastic is discarded and
becomes a plastic pollutant that is a blot on the landscape and a threat to marine
life. Mortality estimates range as high as 1-2 million seabirds and 100,000 marine
mammals per year. A further problem with disposal plastic packaging is the
concern for dwindling landfill space. It has been estimated that most major cities
will have used up available landfills for solid waste disposal very soon. Plastics

comprise approximately 3 percent by weight and 6 percent of the volume of solid

waste.9

The polymers and copolymers of lactic acid have been known for some

time as unique materials since they are biodegradable, biocompatible and

thermoplastic. These polymers are well-behaved thermoplastics,10-11 and are
100 percent biodegradable in an animal body via hydrolysis over a time period of
several months to a year. In a wet environment they begin to show degradation
after several weeks and disappear in about a year’s time when left on or in the
soil or seawater. The degradation products are lactic acid, carbon dioxide and

water, all of which are harmless.

Conventional plastics have good reasons for their success as packaging
materials. They provide appealing aesthetic qualities in the form of attractive

packages, which can be quickly fabricated and filled, with specified units of

products

desracie

meagre
are vac
seeder
coma-rm
rs veg

Temper;

Willa:
some a
Plastici

t

re3901

Well a
Dibvrc
Tor Var
and Dr

alarm
21

products. The packages maintain cleanliness, good mechanical properties and
desirable qualities such as transparency for inspection of contents. If lactic acid

polymers are to compete with conventional plastics, they should have similar or

better properties.12 Table 1 comparies the properties of unplasticized 90I10
copolymer of L and DL-lactide and polystyrene. Unplasticized poly(lactic acid)s
are very stiff with tensile properties similar to PS. The 90/10 copolymer is
selected because of property and processing advantages. Poly(lactic acid)s
containing 98 to 100% of the L-lactic acid have very high crystallinity. Fabrication
is very difficult because of crazing, and need for processing at very high

temperatures causes discoloration and loss of molecular weight.

By using monomer, Iactide, or Iactide oligmer as plasticizers, flexible
poly(lactic acid)s can be made. Using monomer and oligmer as plasticizers has
some advantages. First, they are perfectly safe. Second, their incorporation as
plasticizers during polymerization allows shorter cycle times in the polymerization
reactor, lowers the melt viscosity during reactor handling and during
compounding, thereby easing melt processing and preventing discoloration as
well as maintaining molecular weight. The plasticization of poly(lactic acid)s
provides a systematic attenuation of properties and a broad rang of compounds
for various application. The basic tensile physical properties of a poly(lactic acid)
and polyethylene are compared in Table 2. Both materials can be fine toned over

a range of values for specific packaging applications.

Tablr
‘ a

litre

Table 1. Some Basic Physical Properties Comparisons of Poly(lactic acid) and

 

 

 

 

 

 

 

 

 

Polystyrene
Poly(lactic acid) a Polystyrene ”
Physical properties Oriented Unoriented Oriented Unoriented
Ultimate Tensile Strengthc
psi 14,700 6,900 7,400 7,015
MPa 101 48 51 48
Elastic Modulus 564,000 221,000 450,000 240,000
psi 3,889 1 ,524 3,103 1 ,655
MPa 15.4 5.8 4.0 -
Impact Strength
Ft—lbfrn - 0.44 0.4 0.24
MN/m - 23 21 13
Deflection temperature
°F 1 81 127 200 -
°c 83 53 93 -
Specific Gravity 1.25 1.25 1.05 1.05
Melt Flow Rate
200°C - 46 - 3.5
1 55°C - 2 - -

 

 

 

 

 

" 90/10 L-lactide/DL-lactide copolymer. b Crystal PS, Amoco R3. ° ASTM D882

 

Table 2. Comparison of Polyethylene to Poly(lactic acid)

 

 

 

 

 

 

 

 

 

 

Properties Low density polyethylene Poly(lactic acid) ‘
Tensile Strength:
Kpsi 2.1 8 3.19
MPa 1 5 22
Elongation, % 261 280
Tangent modulus
Kpsi 54.9 36.6
MPa 379 252
100% modulus:
kpsi 1.77 0.74
MPa 12 5
200% modulus
kpsi 1.82 1.2
MPa 13 8
HDT
264 psi, °F 95 122
1.82 Mpa, °C 35 50

 

a L-Lactide plasticizer equals 19.5 wt%

 

for lar;
5053671?
these i
$1 303
the cos
are limi
polgme

DTDCESE

lldmxy

WlThOUg C

The poly(lactic acid)s have some deficiencies that inhibit their acceptance

for large-scale applications”.14 These deficiencies include high price, poor
solvent resistance, insufficient physical properties, and various combinations of
these factors. The selling prices of lactic acid polymers have been estimated at
$1.0011b. Compared to $0.30-0.60 for polystyrene, poly(lactic acid)s are 2-3 times
the cost of the conventional plastics. The major deficiencies of poly(lactic acid)s
are limited physical properties. The technological development of biodegradable
polymers is restricted presently by the range of these polymers that can fulfill
processing and property requirements for many applications in which
biodegradability would be an important materials property. Acceptance of
biodegradable polymers will depend on five unknowns: (1) customer response to
costs that are considerably higher than conventional polymers; (2) possible
legislation (particularly water-soluble polymers); (3) the achievement of total
biodegradability; (4) wider range of available physical properties; and (5) the
development of an infrastructure to collect, accept, and process biodegradable

polymers as a generally available option for waste disposal.

2. Synthesis of Biodegradable Polyesters

Polyesters such as poly(lactic acid) can be synthesized by
polycondensation from the hydroxy acid (Scheme 1). The polycondensation of
hydroxy acids is generally performed by removal of water by distillation, with or

without catalyst, and application of vacuum as the temperature is progressively

increase
mgmi
moecuk
polyconc
Wm

debnno
dkmd

Mend

ogemng
bwmde
"”Q‘ODEr
beam
dmgm,
hcemg

Rir
oirheSe
muchras.
mohOlTier.
high 18mg
ConCentra.

The SISTEn

increased.6v15 This rather simple step leads to oligomers with low molecular
weight (<5,000 Daltons, Mw/Mn close to 2). There have been reports that high
molecular weight poly(lactic acids) have been synthesized using

polycondensation. The methods include postcondensation in organic solvents
using dicyclohexylcarbodiimide (DCC),16 condensation catalyzed by
distannoxane,17 and refluxing in high boiling point solvents such as diphenyl

ether catalyzed by various metals and protonic acids.18 However, these methods

have not been developed industrially.

A better way to make high molecular weight polyesters is through ring-
opening polymerization. As shown in Scheme 1, lactic acid was condensed to
low molecular weight oligomers, the oligomer was cracked to the Iactide, and

ring-opening polymerization of Iactide produced high molecular weight polymers.

This is the method used to make poly(lactic acid) commercially.19 The drawback

of this method is its high cost, which has been reduced tremendously by Cargill,

Inc. using continuous processes to manufacture poly(lactic acid)s.20

Ring-opening polymerization can be carried out in bulk or in solution. Both
of these methods have advantages and disadvantages. Bulk polymerization is
much faster than solution polymerization because of the higher concentration of
monomers and higher reaction temperature. High monomer concentration and
high temperature also cause some problems. Because of high monomer
concentration, a bulk polymerization system has a high viscosity, which makes

the system difficult to stir. Because of concerns about viscosity and melting point

oi the

causes
solutror
scict'ror
grader
very lor
p'bvrde
disrnb'.
such a
amoun
an ind.
GT Their

i0 mak

of the monomers, bulk polymerizations are conducted around 200 °C, which
causes some side reactions such as transesterification and discoloration. In
solution polymerization, the monomers are dissolved in organic solvents, so
solution polymerizations do not have the same viscosity and melting point
problems as bulk polymerizations. Solution polymerizations can be conducted at
very low temperature. Compared to bulk polymerization, solution polymerizations
provide much better control over molecular weight and molecular weight
distribution. However, solution polymerization uses large amounts of solvent
such as toluene. Not only the solvents are expensive, but also dealing with large
amounts of waste solvent is an environmental problem and costs money. From
an industrial point of view, solution polymerizations are not economical because
of their slow rates and the high cost of solvents. Thus, most industrial processes

to make poly(lactic acid) use bulk polymerization.

Scheme 1. Synthetic route for poly(lactic acid)

Polycondensation:

H 0
CH -C-COOH 330—»
3 I A O
OH m

Ring-opening Polymerization

H O
' -H20
Waco... ____, EV}
I A
OH m

low molecular weight oligomer

l cracking

O

O
{10)} 4 $0
m ring-opening O\n/i\

high molecular weight polymer polymerization

 

10

 

3.l

3.1 I

mus‘
have

act:

PCiy

exa:
for ;
nor

Com
war.
of th
for 3

lnCie

3. Ring-opening polymerization of Iactides and Iactones

3.1 Polymerizability

For the ring-opening polymerization of Iactides and Iactones, one question
must first be answered: are the monomers polymerizable or not? Many attempts
have been made to correlate the ring-opening polymerizability of Iactones and
Iactides with some basic parameter, such as ring strain, hydrolysis rate, or
monomer basicity. No acceptable clear correlation has been found to date, but
these attempts provide some insight into the polymerization process, and as a

result some generalizations can be made.

Whether a cyclic monomer can be polymerized to high molecular weight
polymer depends on thermodynamic and kinetic factors of the polymerization.
For a cyclic monomer to be polymerizable, the free energy of the polymerization
AG must be negative under the conditions used. The effect of ring size on AG of
polymerization can be readily illustrated by the use of the cycloalkanes as an
example. The substitution of a heteroatom for CH2 does not significantly alter AG

for polymerization as shown for various cycloalkanes, provided the heteroatom is

not too dissimilar in size and bond angle, such as for O or N.21 Table 3 22
compares the heats of combustion per methylene group in these ring compounds
with that of a methylene in an open chain alkane. This yields a general measure
of the thermodynamic stabilities of rings of different sizes. The strain is very high
for 3- and 4- membered rings, decreases sharply for 5-, 6-, 7- membered rings,

increases for 8- to 11- membered rings, and then decreases again for larger

11

TEST

oi:

rem

rings. Of course, while the presence of strain in a ring compound does not
necessarily make it polymerizable, within a given class, the higher the strain, the
more exothermic the polymerization, the more likely it is to be polymerizable. The
thermodynamic feasibility does not guarantee the actual polymerization of
Iactones. Polymerization requires that there be a kinetic pathway for the ring to
open and undergo reaction at a reasonable rate. Table 4 lists the experimental
results for the polymerizability of a large number of Iactones. It shows that none
of the 5-membered Iactones investigated polymerized, and that substitution
rendered some of the 6-membered Iactones unpolymerizable; however, hearty all

of the 4-, 7-, and 8-, membered Iactones were polymerizable.

Hall and coworkers23 investigated the postulate of Carothers that
hydrolysis rate and polymerization tendencies should correspond with each
other.24 However, they found this suggestion not to be generally valid for
alkaline hydrolysis because B-propiolactone and y—butyrolactone hydrolyzed at
comparable rates, while the former polymerized and the latter did not. The

attempt to relate polymerizability of Iactones with basicities of Iactones by

Lundberg and 00x25 also did not produce satisfactory results.

12

Table 3. Heats of combustion and Strains of Cycloalkanes per Methylene

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

group
n-member Heat of combustion per Strain per methylene
cycloalkanes methylene group group
(KJ/mole) (KJ/mole)
3 697.6 38.6
4 686.7 27.7
5 664.5 5.5
6 659.0 0.0
7 662.8 3.8
8 664.1 5.1
9 664.9 5.9
10 664.1 5.1
1 1 663.2 4.2
12 660.3 1 .3
13 660.7 1 .7
14 659.0 0.0
15 659.5 0.5
16 659.5 0.5

 

 

 

 

 

13

Table 4. Polymerizability of cyclic ester monomers

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Compound Ring size Polymerizability
B-propiolactone 4 +
B-butyrolactone 4 +
y—butyrolactone 5 -
y-valerolactone 5 -
8-valerolactone 6 +

glycolide 6 +

Iactide 6 +
or-n-propyl-S—valerolactone 6 -
tetramethylglycolide 6 -
e-caprolactone 7 +
3-oxa-e-caprolactone 7 +
cis-disalicylide 8 +
di-o-cresotide 8 +
trisalicylide 1 2 +

 

+ represents polymerizable, - represents unpolymerizable.

14

 

3.2 Pc

3.2.1

3.2.1.‘

laclbrr

‘vi

3ch

3.2 Polymerization Mechanism

3.2.1 Cationic Polymerization

3.2.1.1 Lactones

The main cationic initiators used in ring-opening polymerization (ROP) of
Iactones can be divided into four subgroups: (a) protonic acids (HCI, RCOOH,
RSO3H, etc.), (b) Lewis acids (AICI3, BF3, FeClg, etc.), (c) alkylating agents

(stabilized carbocations, e.g. CF3803CH3, Et308F4), and (d) acylating agents

(eg. CH3COOCI3).26 In addition to these traditional acids and electrophiles,

precursors of carbocations have also been considered, e.g., ammonium or

phosphonium salts stabilized by complex counterions,27 which are transformed
into the active species by thermal or photochemical processes. It is also worth

mentioning diaryliodonium salts, which release an active cationic species upon

reaction with a reducing agent.28

Many mechanisms have been proposed for the different cationic initiators.

The polymerization mechanism proposed by Cherdron et 31.29 (Scheme 2) was
accepted for a long time. It consists of an electrophilic attack on the endocyclic
oxygen of the lactone and the subsequent rupture of the acyl-oxygen bond with

formation of an acyl carbonium ion prone to propagate.

Penczek et al. questioned Cherdron‘s mechanism. From chain-end

analyses and kinetics data, they proposed another mechanism30r31 (Scheme

3), in which the electrophilic attack was on the exocyclic oxygen of lactone.

15

Kricheldorf et al.32 and Okamoto33 also proposed mechanisms for cationic

polymerization, which are very similar to what Penczek has proposed.

Scheme 2 Cherdron‘s cationic polymerization mechanism

0 o
(\O Ké‘)
R + ____> _>
O 0
ll e-CL ll

RO—(CH2)5

 

O
+ ———> R0— C —C
C ( H2)s \Cii) Polymer

3.2.1.2. Lactides

The secondary carbon atom in Iactide is less sensitive to nucleophilic
attack compared to the primary carbon atom in lactones. At present, only triflic
acid and methyl triflate initiate controlled cationic Iactide polymerizations; HBr
and HCI also initiate cationic Iactide polymerizations, but yield low molecular
weight polymers. There also are reports where Sn(ll) and Sn(lV) halogenides
were used to polymerize Iactides, but it is believed that the actual initiating
species are HBr and HCI, which are generated by Sn(ll) and Sn(lV) halogenides

reacting with hydroxyl-containing impurities.

16

Scheme 3.The cationic polymerization mechanism proposed by Penczek.

R— 05 o
i 0 d} v/E‘)
R + ——> L
Ti 0
' ’ RO—C—(CHQS— 0:0 ———’ Polymer

A mechanism proposed by Kohn34 for the initiation and polymerization of
Iactide with strong acids such as HBr and HCI is depicted in Scheme 4.
Protonation at one of the two available carbonyl oxygens on Iactide is followed by
a proton shift and an SN2-type transesterification reaction. In my opinion, this
mechanism has flaws. The exocyclic oxygen is more nucleophilic than the
endocyclic oxygen, so it is more reasonable that it should the exocyclic oxygen

that attacks the carbonyl group, and not the endocyclic oxygen.

Methyl triflate (gives very interesting results in Iactide polymerization.
Kricheldorf found that the polymerization of L-Iactide gave polylactide that was
100% optically pure. If the polymerization proceeded with complete Walden
inversion like traditional cationic polymerizations (Scheme 5), the resulting

polymer must contain 50% D- and 50% L- units. The polymers should not be

17

 

Optic
the

lacti

Scheme 4. Cationic polymerization of Iactide initiated by protonic acids

optically active. Kricheldorf proposed a different mechanism35 (Scheme 6) for
the methyl triflate initiated polymerization. The initially formed positively charged
Iactide ring is opened by an SN2 attack of triflate anion accompanied by Walden
inversion. The triflate ester end-group reacts with another Iactide again in an 8N2
fashion, so that a second Walden inversion occurs. The net result is perfect

retention of the original configuration.

Scheme 5. Polymerization of Iactide initiated by methyl triflate (traditional

cationic mechanism)

 

0
I". S
O
CH: 9H:
o-cre—o—co—cH—o—cr-r3
4»: R . S
o’ -----
——>
0 _..l\n/o
O
/ Monomer

Polymer

19

Scheme 6. Polymerization of Iactide initiated by methyl triflate

(mechanism proposed by Kricheldorf)

.OCH3
0’? ac“3 9H3

F3C—303‘ +/l\n/° __, Fac—SOZO—CH—fit—O—CH—CO—OCl-la
0

V o

‘5”3 9‘3
O + F3C— SOZO—CH— Ic— o— cr-r—co— OCH3 _>

so .

CH3
<1}ng
t: o 2CH3
- 0’
Fae—sowo , F30— SOZOMéT/CHQ

0 CH3

&

20

3.2.2 Anionic Polymerization

3.2.2.1 Lactones

Many kinds of anionic initiators have been used in the ring-opening
polymerization of lactones. The most commonly used initiators are alkali metal
alkoxides, such as potassium methoxide, and potassium butoxide. To improve

the solubility and efficiency of the alkali alkoxides, mixtures of alkali alkoxides
and crown ethers have been used as initiators. In addition to alkali alkoxides,36'

38 metal carboxylates,36 alkali metals, and cyclopentadienyl sodium39 also have

been considered as initiators.

The mechanism of anionic polymerization of Iactones depends on the ring
size and substituents. There are two possible propagation routes in the anionic
polymerization of Iactones (Scheme 7): (A) propagation with alkyl-oxygen bond
scission, resulting in an active carboxylate center or (B) propagation with acyl-

oxygen bond scission leading to an active alcoholate cente.

e-Caprolactone (e-CL) polymerization is initiated exclusively by
alcoholates. In the case of B—propiolactone (B-PL), the situation is much more
complicated. According to Penzek et. al., B-PL can be initiated by alcoholates or
carboxylates. When B-PL is initiated by carboxylate, the carboxylate ion is the
only initiation species. When initiated by alcoholate ion, both carboxylate and
alcoholate ions were produced. The two ions coexist during the initial period of
polymerization. The carboxylate ions reproduce themselves every propagation

step, but 50% of the alcoholate ions were changed to carboxylate ions at every

21

propagation step. After several steps, almost all the alcoholate ion vanished

(Scheme 8).

Scheme 7. Propagation routes for anionic polymerization of Iactones

O
A ,
\rla‘lzln ___) .W—(CHflmr—C/e

J \‘0

o\ e
(CH2)n __
”('1‘ j" l —_s t/vw—E (or-r2),,,,——oe

Scheme 8. Polymerization mechanism of B-PL.

 

o

i, A /’O

W(CH2)2— C,\\e ___, W.(CH2)2_ c:\e

A ‘0
o / /

l .\
c c o B
W—n—( Hzlz- 9 ——> fi—(CHQz—Oe

22

 

Alcoholate anions are much stronger nucleophiles than carboxylate
anions. s—CL (AHp = -13.9 KJ/mol) is less strained than B-PL (AHp = -82.3
KJ/mol), so carboxylate anions only initiate polymerization of B-PL and can not
initiate polymerization of e-CL. The alcoholate anions are stronger and smaller
sized nucleophiles. Thus they initiate polymerization of both e-CL and B—PL, but
are less discriminating. They attack at the carbonyl carbon and the carbon next
to the endo-oxygen on the ring and thus both alcoholate and carboxylate ions

exist during the polymerization of B-PL.

Penczek’s mechanism37 seems to explain the experimental data well, but
Jedlinski38r40 gave a different account about how the polymerization of B-PL
proceeds. During the polymerization of B-PL, Jedlinski found unsaturated double
bonds as the end groups and no traces of the initiator in the resulting polymers.

This finding was supported by data from Dale and Kricheldorf“. From his data,

Jedlinski proposed the mechanism shown in Scheme 9.

The results above are very confusing. Both authors used the same initiator
system, potassium methoxide to polymerize B-PL, but their results were totally
different. The only difference in the experimental protocol was the solvent.
Penczek used DMF and Jedlinski used THF. Maybe the polarity of solvents can

explain the difference in the initiation of polymerization.

23

Scheme 9. The polymerization mechanism of PL proposed by Jedlinski

0

R0 010‘ R0 \
. RO'K“ —-> ———> KOH + \ll/\
(3
<3

0

f HO O'K+ -
+ KOH ____’ \n/\/ ___) *KO\n/\/OH
0 o
- H:O/ 0
1 El

Polymer

Polymer

O
D: e.
<-———
0

3.2.2.2 Lactide

The mechanism of anionic polymerization of Iactide is very complex. The
polymerization conditions, such as structures of initiators, counterion,
temperature, and solvent, have a large influence on the mechanism of

polymerization.

Krichedorf and Krieser-Saunders42 found that only strong bases such as
potassium tert-butoxide (pKa = 18) and butyllithium polymerize Iactide and weak
bases such as the benzoate ion (pKa = 5.7), and phenoxide ion (pKa = 9.5) were

not able to initiate polymerization. In addition, polymer was racemized regardless

24

of the initiator, temperature and solvent. They also were unable to find the
initiator fragment in the isolated polymer chain. From these facts, they proposed

that initiation mainly involves deprotonation of Iactide (Scheme 10).

Jedlinski et al.43 used potassium methoxide to initiate the polymerization
of Iactide. They found the initiator fragment in the polylactide and very little
raoemization. Thus, they proposed a different mechanism (Scheme 11). The
methoxide anion attacks the carbonyl carbon in the monomer, with acyl-oxygen

bond scission and formation of the methyl ester end group and the active

Scheme 10. Anionic polymerization of Iactide through deprotonation

0 P
9’!
o ’ O
\n/g\ + RO‘ ———> ROH + 0%
o O

 

25

alcoholate center. Propagation proceeds via acyl-oxygen bond scission and

regeneration of alcoholate anion each propagation step.

Scheme 11. Anionic polymerization of Iactide

O

. “he

The contradictory results above may be caused by the polymerization
conditions. Krichedolf et al. used a tertiary alkoxide (potassium tert-butoxide) and
Jedlinski used a primary alkoxide (potassium methoxide). The difference in the
nuclephilicity of anions might explain the difference in the initiation mechanism.
The groups also used different solvents. The polarity of solvent might play a role

in the initiation mechanism.

Anionic polymerizations of Iactides and Iactones are very fast processes,
but side reactions such as racemization and back-biting make it difficult to obtain
high molecular weight polymer with desired properties. It is unlikely that anionic

polymerization of Iactides and Iactones will have important industrial applications.

26

3.2.3 Coordination Polymerization

As previous discussed, the anionic polymerization of Iactide and Iactones
have side reactions such as transesterification and racemization. These
undesired side reactions result from the high reactivity of the alcoholate. A
decrease in the reactivity, eg. by modification of the counterion, is a possible

way to eliminate, or at least delay these side reactions.

During the last several decades, many research groups have worked on
initiating polymerization using the alkoxides and organometallic derivative of
different metals, aluminum,44r45 tin,46v47 zinc,48r49 iron50 and some rare earth

metals such as yttrium,51 lanthanum.52 The most frequently used metals are

aluminum and tin.

3.2.3.1 Aluminum Derivatives

The most widely used aluminum initiators are aluminum triisopropoxide37,

and soluble bimetallic u-oxo-alcoholates, expecially zinc and aluminum53

(Figure 2).

Y Y 04

\Al-—O\

DEAF? T6 Zn\o’A(
o
T )1

Figure 2.Structure of Al initiators

27

 

Aluminum alkoxides polymerize Iactide and Iactones through a

“coordination-insertion” mechanism,54‘56 which involves coordination of initiator
with the exo-oxygen of the monomer and insertion of the monomer into an Al-O

bond of the initiator, follwed by cleavage of the acyl-oxygen bond of the cyclic

monomer (Scheme 12).

Scheme 12 Mechanism of Iactide polymerization using Al alkoxides

 

28

The polymerization of Iactide and Iactones using Al alkoxides is a living
polymerization. If all impurities are excluded, there is no termination reaction and
the molecular weight of polymer increases linearly with increases in the
conversion of monomer to polymer. The polymerization is first order in
concentration of monomer and concentration of initiator, and has the kinetic

expression shown below.

—M = kperrIr
dt

Polymerization kinetics are influenced by polymerization conditions such

as monomer and solvent.57 When using aluminum triisopropoxide to polymerize
e-caprolactone (e-CL), only one of three alkoxide groups initiates polymerization.
In polymerization of Iactide (LA), all three alkoxide groups initiate polymerization.
This difference is caused by aggregation of the AI(OrPr)3 in the solvent. AI(OrPr)3
has two aggregation states in nonpolar solvents: a tetramer and a tn’mer, called
A4 and A3 respectively.53'52(Figure 3) The exchange rate between the tetramer
and trimer is small, and almost can be neglected. The two aggregates react with
s-CL with different rates. Therefore, when a mixture of A4 and A3 is used to
initiate the polymerization e-CL, A3 is consumed completely whereas A4 remains
unreacted. This observation explains why it is assumed that only one alkoxide
group from Al(OrPr)3 participates in the polymerization of s—CL. Actually, an IVA
mixture was used in such a ratio that on average, one alkoxide of the initial
amount of AI(OIPr)3 initiates polymerization. A3 provides all of the alkoxide

groups, while A4 is inactive. Lactide is less reactive than s-CL, which gives A4

29

enough time to dissociate before the monomer polymerizes completely.
Furthermore, monomer addition causes the aggregated alkoxides to dissociate to
a trisolvated six-coordinate “AI(OIPr)3 - monomer” complex. The only difference
between s—CL and Iactide polymerization is that for s—CL, only A3 dissociates and

for Iactide, both A3 and A4 dissociate (Scheme 13).

 

 

l
O———Al

OR

 

 

OR

I
74%“
O 0
RO— N/Fl l2\\Ar—0R
RO A3 OR

Figure 3 Aggregation states for AI alkoxides

30

Scheme 13. Dissociation of the AI alkoxide aggregates

/ \
A3 > R0 l OR —> Polymer

 

.?
+
.P
O
l

 

T
O

31

The use of different solvents does not modify the average number of

active alkoxide groups per AI, but according to Teyssie,53 the polymerization rate
in THF is significantly smaller than that in toluene. When s—CL is polymerized, the
half polymerization time 0112) is 3.5 min in toluene and 8.7 min in THF ([s-CL]=1

moi/L, [M]o/[l]o=200 and T = 0°C).

The reason for the rate decrease is that THF is a much better solvating
agent than toluene. The ether oxygen of THF is able to compete with the
carbonyl group of Iactones and Iactide for coordination to aluminum. Because the
ether oxygen on THF is more nucleophilic than the carbonyl group, it makes the

formation of AI(OrPr)3o(lactone)3 unfavorable.

AI(OIPr)3 is a very clean initiator for lactones. The molecular weight
increases linearly with conversion and the molecular weight distribution is very
narrow. But this linearity is not observed at very high conversions, or at high

polymerization temperatures. The deviation from linearity is caused by

transesterification reactions.55 Transesterification reactions can be
intramolecular (“back-biting” Scheme 14) or intermolecular. Backbiting produces
a shorter polymer chain and a cyclic oligomer from original polymer chain, and is
responsible for the broadening of the molecular weight distribution and for the
decrease in the number-average molecular weight. lntennolecular
transesterification (Scheme 15) does not change the number-average molecular

weight of the polymer, but it does increase the polydispersity of the polymer.

32

Scheme 14. lntramolecular transesterification

Y VS 4 ..__._
013%. . We ..__.

/ \OR

0:2}.

9M2};

cyclic °"90mer

33

Scheme 15. Intermolecular transesterification

The rate of transesterification depends on the temperature, and the kind of

metal center in the initiator. Kricheldorf et al.64 compared the transesterification
activity of different kinds of metal alkoxides. He found that aluminum alkoxides
are least likely to cause transesterification and the tin alkoxides are most likely to
cause transesterification. The order for transesterification activity is AI(OIPr)3 <

Zr(OnPr)4 < Ti(OnBu)4 < Bu3SnOMe < Bu26n(OMe)2.

Polymerizations of Iactide and Iactones initiated by aluminum alkoxides
are perfect living polymerizations and the alkoxide groups are the end-groups of
the polymers. If the alkoxide group contains some functionality, the polymers will
have a functional end-group. Halogens, tertiary amines, carbon-carbon double
bonds, and methacrylates are typical example of functional groups that can be
placed at the chain end. The initiators are easy to synthesize by mixing Al(Et)3

with the corresponding alcohol (Scheme 16). By controlling the amount of

Scheme 16. Preparation of end-functioned AI initiators

AllEt)a+ n HOCH2CH2X —-> StenAIrocrrzcrrzxrn 1sns3

0

9hr M .
Et3-nAI(OCH2CH2X)n+/Kl( _>
O O n

O

0
ll
X= Br, -CH=CH2, ~NEt2, -OCCH=CH2

35

alcohol, the structures of the initiators can be controlled. End-functionalized

polylactides and polylactones have been used in many fields. Teyssie et al.55‘69
used them to synthesize crosslinkable polylactides and polylactones, graft
copolymers and star-branched polyesters. The great versatility of end-
functionalized aluminum alkoxides opens the way to the macromolecular
engineering of polylactides and polylactones. This approach will dramatically

increase the range of the physical properties for polylactide and polylactones.

lnoue et al. reported that derivatives of tetraphenylporphinato—aIuminum
are efficient and versatile initiators for the living polymerization of Iactones and
Iactide. The polymerization mechanism agrees with a “coordination-insertion“
mechanism that involves acyl-oxygen cleavage. The polymerization rate can be
remarkably increased by addition of sterically hindered Lewis acids. Coordination
of the Lewis acid to the carbonyl oxygen of lactone or Iactide monomers makes

them prone to nucleophilic attack. These porphyrin derivatives polymerize not
only Iactones and Iactide (Figure 4, X=OR),70'72 but also epoxide 73:74(X=Cl),
methacrylates,75 and methacylonitrile76 (X=SR, R=alkyl). The corresponding

copper and iron complexes were investigated by Kricheldorf et al.77 The iron or
copper porphrin complexes polymerize Iactide but produce relatively low

molecular weight polymer in low yield.

 

X = OR, Cl, Alkyl, SR, OZCR

Figure 4. The structure of porphyrin initiators

37

3.2.3.2 Tin Derivatives.

Many kinds of tin compounds have been used in the polymerization of
Iactide and Iactones such as SnO,78 SnOz,79 SnX4, San (X = Br, Cl),80'82

Sn(OR)2. Sn(OR)4,83»86 and (RCOO’)28n.46v47337'39 The mechanism for
polymerization using tin compounds is very controversial, and anionic, cationic
and coordination mechanisms have all been proposed. Because of a lack of

experimental evidence, none of the mechanism is very convincing.

Kricheldorf90:91 investigated tin oxides and oxides of many other metals
such as Mg, Sb, Pb and Ge for L-Iactide polymerization. These catalysts are
heterogeneous, so the repeatability of polymerization is poor. Also, these
catalysts cause monomer racemization and transesterification. There is no
detailed mechanisric study of the metal oxide catalyzed polymerizations, but
people believe that hydroxyl-containing impurities such as water and alcohol

initiate the polymerization.

Tin alkoxides perform just like aluminum alkoxides. The polymerization
mechanism involves monomer coordination to the tin alkoxide, followed by
insertion of monomer into the Sn-O bond (Scheme 17). Tin alkoxides are very
efficient initiators. They do not cause racemization of monomer up to 150 °C, but
since tin alkoxides are good transesterification catalysts for noncyclic ester
groups, they may cause extensive back-biting degradation of polylactones and
polylactide at elevated temperatures. The loss of molecular weight control is the

reason that tin alkoxides are not as widely used as the aluminum alkoxides.

38

Scheme 17. Mechanism for tin alkoxide initiated polymerization Iactides

O
Bu Sn 0
—-> 3 \O/i\n/ TKOR
O

 

polymer

Tin halides such as SnClz and SnBrz are very good catalysts, and high
molecular weight polylactide and polylactones have been obtained using tin
halide catalysts. Because they are not soluble in organic solvents and
monomers, the repeatability problem still exists. The proposed mechanisms for
lactone and Iactide polymerization using tin halides are very different. The
mechanism of lactone polymerization is shown in Scheme 18. It includes
coordination of acidic tin halides with the carbonyl oxygen, ring-opening by
transfer of a halide atom from the catalyst to the (lb-carbon, followed by cleavage
of the alkyl-oxygen bond. The propagation can proceed through two routes:
through the carboxylate (A) or through alcoholate (B). The exact propagation
species has not been identified. Lactide polymerization using tin halides acts
much different. Unlike lactone polymerizations, halide end groups have not been

found by NMR and elemental analysis. Lactide polymerization is not initiated by

39

transfer of halide atoms, probably because the secondary carbon on the Iactide
is less sensitive to nucleophilic attack than the primary carbon of a lactone.
Based on the above observations, Kricheldorf proposed that tin halides react with

impurities such as alcohol or water to first form tin alkoxides, which then follow

the coordination-insertion polymerization mechanism.

Scheme 18. Mechanism for tin halide initiated polymerization of caprolactone

 

SnBr4 + o ——>
O
BraSn\ O/U\ /Br A O 0
: (CH2)5 Br
I ——> Br3Sn\ /
O> (CH2)/ (CH2)5
([3 l
Polymer
O l
O
Br ~. /u\ /Br 0
385' ° (CH2)5 B /o\ kO/lk /Br
50> Brasn (CH2)5 (cg-(2)5

40

Tin(ll) 2-ethylhexanoate (Sn(Oct)2) is the most widely used catalyst for
Iactide polymerization. Commercial polylactide is synthesized using this catalyst
since it provides high reaction rates, high conversions, and high molecular
weights even under rather mild polymerization conditions. There are extensive

studies of this catalyst system, but the mechanism is still unclear. Nijenhuis et

al.92:93 proposed the cationic polymerization mechanism shown in Scheme 19.

This is a complicated mechanism, but it is unlikely that the polymerization
follows this path. First, the dissociation of Sn(Oct); in a non-polar environment is
not energetically favorable. Second, formation of an eight membered ring during
polymerization should also be unfavorable. Finally, the strong interaction
between Sn and Iactide prior to reaction with the OH group is energetically
unfavorable. To date, there is no spectroscopic evidence for any of the

intermediates in the mechanism.

41

Scheme 19. Mechanism for polymerization of Iactide initiated by
Sn(2-ethylhexanoate); as proposed by Nijenhuis et al.

s..- 331 __. 14371;;

H\

%‘H
\xSn I Sn
I R l
Oct Oct

\(lko o ,
——> H/X\ ——> 0 9H
/S_/ S" O
i)
R T" R0
Oct

f}

o O
H o
0 OH H
0 OH
fiO—Sn-u-O i0 <———>
R0 bet 2&0 /¥0----Sn—o 3&0
RO

OH

_. #0 Hob—{2% ....

42

Kricheldorf et al.94 proposed the mechanism shown in Scheme 20. The
first step consists of coordination of alcohol or water to Sn(Oct)2. The
complexation of Iactide to the intermediate polarizes its carbonyl group, making it
susceptible to nucleophilic attack from alcohol. This mechanism also has a
shortcoming. Based on the assumption that the tin catalyst is active via free sp3d2
orbitals, Sn(Oct)2 can coordinate with both alcohol and Iactide. But this
mechanism can not explain the activity of Bu28n(Oct)2, which is six coordinate
and should not be active. In past, Bu28n(Oct)2 polymerizes Iactide, though not as

well as Sn(Oct)2. The authors rationalized their results by treating the Oct group

as a monodentate ligand, but it is well known92 that Oct is a bidentate ligand,

and Sn(Oct)2 and Bu28n(Oct)2 have the structure shown in Figure 5.

Penczek et al.95 proposed that Sn(ll) alkoxide species initiated the Iactide
and lactone polymerizations shown in Scheme 21. In their work, they identified
intermediates such as RO[M],,OSnOct. This mechanism is very similar to what

we propose based on our experimental work. The details will be discussed later.

43

Schr

Raf I
R

Scheme 20. Mechanism for polymerization of Iactide initiated
by Sn(2-ethylhexanoate)2 as proposed by Kricheldorf

(if
T“ 'T1 33 )T
2,.“
RZ—Sn—R‘ + R3OH ———> R 2?"“0—H ———>
a: R: 2
R3 H O
1 1
R2! 3 ”6—H R2. 2 I
n’ 'u. n___,0 o\ 3
Rl’l'R V‘ 0 o ——> [21’le R
O O
O

___.___) Polylactide

Scheme 21 Mechansim for Iactide polymerization initiated by

Sn(2-ethylhexanoate); as proposed by Penczek et al.

0
\ lulu“ O
Olrunnflsn“\ + ROH
O
O
O 0 O
\ + _9_, \ O\R
Ounnll“sn_OR On||||"“sn 0

2n

45

O

Bu
0 \ I |I|||\O
\ ““‘O O||IIHI|||SnQ|H‘
09|I||I|I||Sn‘|\\\| I
\ Bu 0
O
Solon); Buzsrmm)2

Figure 5. Structure of Sn(Oct)2 and Bu28n(Oct)2

There are some other proposed mechanism when Sn(2-ethylhexanoate): is

used as catalyst such as the one by Vert et al39., but there is no published
mechanism that can explain every experimental fact. More research is needed to

fully understand this reaction.

3.2.3.3 Rare Earth Metal Compounds

The 15 Ianthanide elements (La, Ce, Pr, Nd, Pm, Sm, Eu, Gd, Tb, Py, Ho,
Er, Tm, Yb, and Lu) together with Sc and Y elements are called rare earth
elements. They prefer tri-valent oxidation states in complex formation. Rare earth
metal complexes are known to initiate the living polymerization of methyl
methacrylate over a wide range of polymerization temperatures to give high
molecular weight syndiotactic polymers with extremely narrow molecular weight

distributions. Lactones and Iactide have also been successfully polymerized

46

using rare earth halide systems and rare earth complex systems. Yasuda et
al.96-98 used [SmH(CsMe5)2]2 and LnMe(C5Me5)2(THF) (Ln=Sm, Yb, Y, Lu) to

polymerize Iactide and lactones. Also, Feijen99r100 thoroughly studied the
polymerization of Iactide and lactone using yttrium alkoxides. These catalysts
initiate polymerization via a “coordination-insertion” reaction. The mechanism

(Scheme 22) is very similar to that of the aluminum alkoxides and tin alkoxides.

Shen et al.101'106 found that rare earth halides can be used to polymerize
Iactide and lactones. He also found that the polymerization rate was dramatically
increased by adding epoxides. The halides first react with expoxide to form the
alkoxides, (Scheme 23) which are much better initiators than halides. Because of
their high activity and few side reactions, rare earth catalysts will have a bright

future in Iactide polymerization.

47

Scheme 22. Polymerization of Iactide initiated by rare earth catalysts.
Me
OOTL /%Ln‘/
__> \
O
“<30 \(KO fl YLO
O
j/ko

“is __, 5:51) “by...

CEO

This rife

Polylactide

Scheme 23. Initiators made from epoxides and rare earth halides

o
C) R + erx3 __, X3,,,Ln[(OCHCH2)y]X

x = Cl, Br, I; R = H, CH3, CHZCI, CHZOCHZCH= CH2

48

3.2.4 Other Polymerization methods

In addition to the polymerization methods mentioned above, other
polymerization methods such as free-radical polymerization, active hydrogen
polymerization and zwitterionic polymerization have been used with Iactide and
lactones. Free radicals are usually ineffective for the polymerization of Iactones

and Iactide. Molecular weights are often low, and monomer conversion

limited,107 so there are few examples of radical polymerization of lactones. The
polymerizations of Iactones and Iactide initiated by active hydrogen compounds,

e.g., amines or alcohols, are relatively slow processes, and usually produce

polyesters of low molecular weight.108 There are examples of the polymerization
of Iactones initiated by ethanolamine or ethylene glycol, but polymerizations

typically need 3-4 days to reach high conversion even at 180 °C.

In contrast to the slow polymerization of Iactones initiated with active

hydrogen compounds, zwitterionic polymerizations are fast processes that give

high yields and high molecular weight polymers. Wilson and Beaman109 showed
that polymerization of pivalolactone initiated with strained cyclic amines is a
zwitterionic process (Scheme 24). Initiation corresponds to formation of a
zwitterion as a result of nucleophilic attack by the amine at the lactone methylene
group. The carboxylate anion of the zwitterion is the propagating species.
Zwitterionic polymerization usually only occurs with very strained Iactones, such
as B-PL. There are also some claims that the polymerization of e-CL proceeds
through a zwitterionic mechanism when initiated with aniline in the presence of a

protic acid.

49

Scheme 24. Zwitterionic polymerization mechanism

4;. . >$ __, 9:292:09

L Q~W%&

 

50

4. Modification of Polylactide Properties

To be successful in areas such as packaging, polylactides must be
available that exhibit a broad spectrum of physical properties while retaining the
degradability of the parent polymer. Typical approaches used to modify the
physical properties of polylactides include manipulation of stereochemistry,

copolymerization and formation of blends.

4.1 Manipulation of stereochemistry.

Lactide exists as three diastereomers: SS, RR, and RS (Figure 6). The
SS diastereomer is referred to as L-lactide, the RR as D-Iactide and the RS as
mesa-Iactide. An equimolar ratio of L—lactide and R-lactide is referred to as rac or

D, L—lactide.

Polymers prepared from different Iactides have very different physical
properties. Polymers of high purity L-lactide or D-lactide are crystalline with

melting points around 180 °C. The polymers of rec-Iactide are amorphous with a

glass transition temperature ranging from 22-65 °C.110 The different properties
are related to the tacticity of the polymers. The polymer can have three limiting
kinds of tacticity: atactic, isotactic and syndiotactic (Figure 7). If all the
Stereocenters on the polymer chain have the same configuration such as
RRRRRRRRR or 88888888, the polymer is called isotactic. A syndiotactic

polymer structure occurs when the configurations of the sterecenters alternate

51

from one repeating unit to the next such as RSRSRSRSRS. If successive

stereocenters are randomly distributed, the polymer is called atactic.

L-Iactide and D-lactide are known to polymerize to isotactic crystalline
polymers while rec-Iactide polymerizes to atactic amphours polymer. By
manipulating the stereochemistry of polymer, the properties of the polymer can
be altered. For example, poly(L-Iactide) degrades slowly because of its highly

crystallinity; by copolymerizing a small amount of rec-Iactide with L-Iactide, the

degradation rate of the resulting polymer is much faster.111r112 The added rac-
lactide introduces R stereocenters into the polymer backbone, which act as
defects and interrupt the crystallinity of the polymer. Thus, the degree of
crystallinity of the polymer is smaller, which results in a faster degradation rate.
By controlling the amount of rec-Iactide added to L-Iactide, the rate of
degradation can be controlled. Poly(L-lactide) has a melting point around 180 °C,
which makes poly(L-Iactide) very hard to process without discolorization and loss
of molecular weight. To solve this problem, a small amount of rec-Iactide is often

copolymerized with L-Iactide to decrease the melting point.

Tsuji et al.113'122 found that mixing an equimolar amount of poly(L-
Iactide) and poly(D-lactide) produces a polymer with a melting temperature and
crystal structure different from poly(L-lactide) and poly(D-lactide) homopolymer.
For example, the homopolymer of poly(L-lactide) has a melting point of about
180 °C, but the mixture melts near 230 °C. The mixture of poly(L-lactide) and
poly(D-Iactide) is termed a stereocomplex. The crystal structure of the

stereocomplex has poly(L-lactide) and poly(D-lactide) chains packed side by side

52

in a 1:1 ratio of D and L monomer units. Because poly(L-lactide) and poly(D-
Iactide) pack better together than either packs with itself, the racemic crystallites

are more stable than crystallites from the homopolymer.

O O 0
fl huh.¢ $0
0 O O
(3. 3) (R. R) (R: S)
L-lactlde D-lactide meso-Iactide

Figure 6. Stereochemistry of Iactide

53

\erwrsee'

isotactic

\ir"rb<ir°t*cjbr°rbir

syndiotactic

atactic

Figure 7. Tacticity of polylactide

4.2 Copolymerization

Copolymerization is the most important method for modifying the
properties of polymers and adjusting them to fit the needs of a particular
application. For example, poly(e-caprolactone) (PCL) is a biodegradable polymer
with a half-life in vivo of one year and that is permeable to many drugs. In
contrast, PLA is hardly permeable to most drugs, and its half-life time is much
shorter, a few weeks in vivo. Thus, combining the permeability of PCL and the
rapid biodegradation of PLA may lead to a wide range of drug delivery devices
with adjustable properties such as permeability, hydrophilicity and degradation
rate. As shown in Figure 8, copolymers can be divided into three groups: random

copolymers, block copolymers and graft copolymers.

ln random copolymers, the different repeating units are randomly
distributed along the copolymer chain. Random copolymers are usually
synthesized by initiating a polymerization in a mixture of comonomers and
catalyst. The randomness of the copolymer is controlled by many factors such as
temperature, nature of the initiators, reactivity of the monomers and ratio of the
comonomers in the monomer pool. For example, the reactivities of Iactide and s-
caprolactone are very different. When copolymerized, Iactide prefers to add to
itself and less than 50% of caprolactone has reacted by the time almost all of the

Iactide monomer is polymerized. Thus poly(e-caprolactone-co-DL-lactide) has a
blocky structure.123'126 When e-caprolactone copolymerizes with e—methyI-e-

caprolactone,127 the distribution of the monomers in the copolymer are nearly

perfectly random because of the similar reactivity of the monomers. The

55

randomness of the copolymer is also controlled by initiators. Shen et al.")2
found that poly(s-caprolactone—co-DL-lactide) produced using the NdCl3-
propylene oxide catalyst system is much more random than poly(e-caprolactone-
co-DL-lactide) produced using Sn(Oct); and Al(OiPr)3. By choosing a high

polymerization temperature, ordered copolymers can be randomized by

transesterification reactions.128

Random copolymers usually have properties intermediate between those

of the parent homopolymers and are usually amorphous polymers with a single

m ABBBAABABBAABAAA m

Random Capolymer

mAAAAAAABBBBBBm m AAAAAAAA(BBBBBBBB)nAAAAAAAm

Diblock copolymer Triblock copolymer

3
MWWWWUWE
t

Graft copolymer

Figure 8. Copolymer architectures

glass transition temperature. The glass transition temperature can be predicted
using the Fox equation.

(Mail?

3 8| g2

where T9 is the glass transition temperature of the copolymer, T91 and T92
are the glass transition temperatures of the pure homopolymer of components 1

and 2, and W1 and W2 are the corresponding weight fractions.

By altering the ratio of the comonomers, the properties of the copolymer

can be easily controlled. Besides e—caprolactone, Iactide has been copolymerized
with monomers such as glycolide]:129 8—valerolactone,1130-131

carbonates,132v133 and ethylene glycol.78 These copolymers greatly widen the

scope of the physical properties of polylactides.

The most progress has been made in block copolymers of Iactide and

lactones. Block copolymers have been made with different architectures such as

diblock,40r134v135 triblock,73v136r137 star-block,138'140 and multiblock.141 In
contrast to random copolymers, block copolymers are usually multiphase
materials. They can form multi-crystalline phases when blocks in the copolymer
crystallize separately, alternating crystalline and amorphous phases if some
blocks crystallize and some blocks stay amorphous, and a mixture of amorphous
phases if none of the block crystallize. The different morphologies of block
copolymers provide a wide range of physical properties that can be selected for

different applications.

57

There are several ways to synthesize block copolymers. The first is to take

advantage of the living nature of Al-initiated polymerization of Iactide and

lactones."42"145 As shown in Scheme 25, the comonomers are polymerized
sequentially. First, e-caprolactone is polymerized to form a poly(e-caprolactone)
block. Because there is no termination reaction, the blocks remain active toward
polymerization. Lactide monomer is then added, and a polylactide block grows
from the end of polylactone block to form the diblock copolymer. There are some
problems with this method. The first is the polymerization order. To' make
poly(lactide-bloc—e-caprolactone), the lactone monomer has to be polymerized
first. If Iactide is polymerized first, the lactone monomer will not polymerize at all,
due to the different cross-propagation rates. The PCL active chain ends react
with LA monomers 2.0x105 times faster than PLA active chain ends react with CL
monomers. Thus, it is very important to know the correct order to add monomer.
The second problem is the number of chains initiated by an initiator. For
Al(OiPr)3, only one alkoxide group is active if the monomer is e-caprolactone, but
all three alkoxide groups are active when Iactide is the monomer. This difference
leads to a mixture of the block copolymer of Iactide and lactone and the
homopolymer of Iactide. This problem can be solved by using initiators such as

(Bu)2AlOiPr that initiate only one chain for both Iactide and e-caprolactone.

58

The second way to make block copolymer is to use macroinitiators. As
shown in Scheme 26, there are two ways to use macroinitiators. In the first, a

polymer with a hydroxyl end-group such as poly(ethylene glycol) monomethyl

ether is reacted with Al(Me)3 to produce the macroinitiator.145'1"t8
Polymerization of Iactide using this initiator produces block copolymer. The

second route uses Sn(Oct)2 and a polymer with hydroxyl end-groups as the

initiator.93v133v149 This route is very flexible. Triblock and multiblock copolymers

can be synthesized using this method.

Scheme 25. Synthesis of diblock copolymers

O

0 o
0 —» o
BugAlOiPr 0 AlBu2
n
o
o o
o o
0 o
0 m

59

Scheme 26. Synthesis of ABA triblock copolymers

HO—(0420H20)n—CI-12Cl-120Me + AIMGa ——> MezAlO—(CI-I20H20)n— Cl'lzCI'EOMe

o
o ‘zzo °
0 O
, MezAI/E fio— (CH20H20),,— CHZCHQOMe (1)

 

sn(OCt)2

o
l-lO—(CH20H20)n—CHZCHQOH + oQ‘xo ____,
o
o
4"be
O—(CH20H20)n— crecreo\[\n/L 3'21 (2)
m 0
0 m

Graft copolymers of Iactide and Iactones are rare compared to random

copolymers and block copolymers. People mainly focus on using polyols such as

poly(vinyl alcohol),150 and polysaccharides (pullulan, amylose)151r152 as
initiators to polymerize Iactide and lactones. Polyols are hydrophilic, and

polylactide and polylactones are hydrophobic. Combining these two segments in

60

a copolymer will provide materials very special properties such as micelle

formation in solvents, which can be used to make drug delivery devices.

In addition to synthesizing copolymers, the crosslinking of homopolymers
and copolymers of linear polylactides and polylactones can be used to modify the
physical and mechanical properties of these materials. By introducing crosslinks
into polylactide, physical properties such as the crystallinity, the melting point and
the glass transition temperature will be influenced. The characteristics of the
degradation by hydrolysis of these biodegradable polymers will also be
influenced. Three methods that have been used to crosslink polylactide and
polylactones: high-energy electron beams, peroxide-induced radical crosslinking
and copolymerization of functional Iactide and lactone oligomers with a

multifunctional monomer.

High—energy electron beams significantly degrade polylactide and
polylactones, so it is not a suitable technique for crosslinked polylactide and
polylactones. Peroxide crosslinking of polylactide and polylactones appears to be
an effective method for affecting the thermal as well as the mechanical
properties. However, in terms of the potential biomedical application of these
materials, the method has two disadvantages. First, peroxides modified
polylactide and polylactone chains may have undesirable degradation products.

Secondly, many peroxide compounds are toxic, so crosslinked materials will

need to be thoroughly extracted to remove peroxides, which is very costly.153

61

Copolymerization of functional Iactide and lactone prepolymers with other
monomers is the most successful cross-linking method. As shown in Scheme 27,
the multifunctional prepolymers with end-groups such as methacylate were

synthesized and cured with monomers such as styrene and methyl methacrylate

to form networks.154-158

Scheme 27. Synthesis of cross-linkable copolymers

0
Ho /
\_ + \'/lko Sn(Octrz
4 °” WK
HO

M071 fl? M

“W 071 flfi

62

4.3 Blending

Copolymerization is an effective way to modify the properties of the
polymers. However, copolymers are often hard to make and are very expensive.
Blending of polymers might offer a more cost-effective way to modify polymer
properties compared to copolymerization. Blends oftem have inferior properties
compared to copolymers, but their ready availability and low-cost are very

attractive.

Blends can be made through thermal mixing and solvent mixing. In
thermal mixing, the components of the blend are melted, and mixed thoroughly to
make blends. In solvent mixing, the components of the blend are dissolved in a

common solvent, and evaporation of the solvent yields the blend.

Macroscopic properties such as impact and tensile strength, and
degradation behavior can be modified by a reasonable choice of the blend
components. The final properties will depend not only on the chemical
composition of the blend but also on its physical characteristics, such as glass
transition temperature, crystallinity and morphology, which, in turn, are a direct
consequence of the compatibility between the components in the blend. The
miscibility is the single most important factor for blends. If the components are
miscible, the blend is a homogeneous system. If the components are immiscible,
the blend will form a two-phase or multi-phase system. Most blends are phase-

separated, which hinders applications of the blends.

There are many reports of blends of biodegradable polymers. For

instance, Langer et al.159 used poly(lactide)/pluronic blends as protein-releasing

63

matrices. Eguiburu et al.160 blended amorphous and crystalline polylactides with
poly(methyl methyacrylate) and poly(methyl acrylate). The resulting blends are

miscible, and have interesting thermal properties. Poly(L-Iactide) has also been

blended with poly(e—caprolactone) and other polymers.161'165

5. Degradation of the polylactide and polylactones

There are two ways to degrade polylactide and polylactones. They can be
degraded thermally, and they can be degraded environmentally which includes

hydrolytic and enzymatic degradation.

5.1.Thermal degradation

Understanding thermal degradation is very important for polymer
processing and polymer recycling. Poly(L—lactide) has a high melting point, so
processing steps such as injection molding and extrusion must be done at
around 200 °C, a temperature where polylactide and polylactone begin to
degrade. Degradation can be prevented if we understand the mechanism for
thermal degradation. Polylactide and polylactones also can be recycled by

selecting favorable thermal degradation pathways.

Transesterification, cis-elimination and radical reactions165'172 have
been proposed as thermal degradation mechanisms. The transesterification
reaction was discussed earlier. The intra-molecular transesterification reaction
generates volatile low molecular weight oligomers, which causes the polymer to
degrade. The mechanism of the cis-elimination reaction is shown in Scheme 28.
The elimination of a proton from methyl group produces 1-alkenes and carboxylic
acids. The carboxylic acids also accelerate the degradation of the polymer. The
proton on the methyl group is not very labile, so the cis-elimination reaction is
lilely only at high temperatures, Usually the cis-elimination reaction cannot
compete with the thermal degradation of polylactides by transesterification.
Radical reactions (Scheme 29) can start with either alkyI-oxygen or acyl-oxygen
homolysis. Several types of oxygen and carbon-centered macroradicals may be
formed. From the radicals, a varietiy of volatile compounds can form, such as
carbon dioxide, carbon monoxide and aldehydes. The thermal degradation can
be influenced by many factors such as residual metal compounds and unreacted

monomers. Residual metal and monomers increase the degradation rate greatly.

Scheme 28. Thermal degradation via cis-elimination

on
0.
CH3/ \W‘sH OR C
37 ' clr 0 // ”2
\ CCHZ -——> CH3/ \¢ + crl
C OH COOR

OOR

65

Scheme 29. Radical thermal degradation

«W @3va **

carbonyl-radical carbon-radical

1

CO, CO), aldehyde, cyclic oligomers...

i

acyl-racidal oxy-radical

“OM °w**—» eds. age/k

5.2. Blodeg radation

Biodegradation mechanisms for Iactide include hydrolytic

degradation129v136r173‘175 and enzymatic degradation.176'178 The two
degradation processes are very different as shown schematically in Figure 9.
Enzymes are high molecular weight molecules, so it is very difficult for enzymes
to diffuse into polymers. Thus, enzymatic degradation almost always happens at

the surface of the polymer. The top half of the Figure 9 shows the weight loss

and molecular weight loss as a function of time for enzymatic degradation.
Because degradation is limited to the surface of the polymer, the surface of the
polymer is eroded, but interior of the polymer is unchanged. The weight of
polymer drops as polymer chains on the surface are degraded to soluble
oligomers, but the molecular weight of the polymer remains almost constant until

most of the polymer weight is lost. In contrast to enzymatic degradation, the

Figure 9. Degradation models

Weight
Molecular weight

Time Time

Enzymatic Degradation

Weight
Molecular weight

FF

Time Time

Hydrolytic Degradation

67

molecular weight drops continually and the polymer weight remains constant
during hydrolytic degradation. Hydrolytic degradation ocurrs in the bulk of the
polymer. Water molecules diffuse into polymer and hydrolyze the ester bonds of
polylactide, causing the molecular weight of the polylactide to drop. When the
molecular weight drops to where the polymer becomes soluble, the weight of
polymer begin to drop. Hydrolysis of the ester bond produces a carboxylic acid,
which catalyzes further degradation of the polylmers. This phenomenon is called

autocatalysis.

Temperature, pH of the degradation medium, the stereochemistry of the
polymer, the morphology of the polymer and the physical size of the polymer all
influence the degradation rate of the polymer. Because water diffuse at a slower
rate in crystalline polymers, and has a lower solubility due to the higher density of
chains in crystalline polymers, crystalline polylactide degrades slower than
amorphous polylactide. Enzymatic degradation is sensitive to the .
stereochemistry of the polymer, and enzymes often selectively degrade one
isomer. The degradation process is complicated, and there are many
controversial conclusions. Many people question the role of enzymes in
polylactide degradation, since hydrolytic degradation always accompanies the
enzymatic degradation. Much work needs to be done in order to fully understand

the degradation process.

68

RESULTS AND DISCUSSION

1 . Monomer synthesis

In order to synthesize substituted polylactides through ring-opening
polymerization, a series of substituted glycolides, shown in Figure 10, were
prepared. For simplicity, we used the glycolide structure as a template to name
these compounds. Substituted glycolides have been known for over a century,

with the synthesis of ethylglycolide via the dimerization of the sodium salt of 2-

bromobutyric acid reported as early as 1893.179 We prepared substituted
glycolides using the two routes outlined in Scheme 30. The first route mirrors the
standard method used to prepare the Iactide dimer. The appropriate or-hydroxy
acid was first condensed in toluene to give low molecular weight oligomers, and
cracking the oligomers under reduced pressure in the presence of a

transesterification catalyst such as ZnO directly yielded the volatile dimer. This

method has been used industrially to produce polylactide.19 In the second route,

an or-bromo acyl bromide was condensed with an or-hydroxy acid to form an

ester,180 followed by ring closure to give the cyclic dimer. In this method, it is
very important to remove the N(Et)3 in the work-up. Even trace amount of N(Et)3
can change dimers to low molecular weight oligomers in a short period of time.
Because N(Et)3 is not nucleophilic, it is likely that Et3NHBr caused the
polymerization. The first method gives the dimer in higher yield, but the second
method offers more flexibility in synthesis and can be used to synthesize

unsymmetrical dimers, which are discussed in a later section.

69

Both synthetic methods yield a mixture of diastereomers. Thermal
cracking of oligomers consistently yields a near-statistical mixture of the R,S and
R,R/S,S diastereomers. 1H NMR of a representative example, that of
ethylglycolide, is shown in Figure 11. The methine protons of the 3,6-
disubstituted glycolide ring appear as a doublet of doublets near 5 4.85. The
signal for R,S isomer appears at 4.88, while that from the RR and 8,8 isomer is
at 4.83. The near equal intensities of the peaks confirms the 1:1 mixture of
R,R/S,S and R,S diastereomers. Similar results were obtained for

isobutylglycolide and hexyglycolide prepared by the thermal cracking route.

Non-statistical mixtures were obtained when the second route was used.
The R,R/S,S isomers are about 70% of the total and the R,S isomer is about
30%, which is probably is related to the kinetics of diastereomer formation. It is
likely that the R,R/S,S isomers form faster than the R,S isomer. In addition, n'ng-
closing competes with oligomerization to form the linear dimer, trimer and
tetramer. Because of the slower rate for forming the R,S isomer, oligomerization

is probably favored to form linear oligomers rather than S,R glycolides.

70

fits rev

Lactide Ethylglycolide
3,6-dimethyl-1 ,4-dioxane-2,5-dione 3,6-diethyl-1 ,4-dioxane-2,5—dione
O O

O
Isobutylglycolide Hexylglycolide
3,6—diisobutyI-1 ,4-dioxane—2,5-dione 3,6-dihexyl-1 ,4-dioxane-2,5-dione

Figure 10. The structures of substituted glycolides

Scheme 30. The synthetic route to substituted glycolides

 

 

 

O O
HO p-TSOH. toluene I-IO H
OH * O
A
R R n
oligomer
ZnO A
O
Br\'/ll\ O
Br
R
>
O
RN
0

R : CH3. CH2CH3. CHZCH(CH3)2
(CHzlscHs

72

Iee_.oo>_m_§ex_oeoec oo__oo>_a_>£o Lo «:22 I. .: 2:9".

 

 

 

 

 

 

 

 

 

 

 

add me o; m; ozs. m.~ o.m m.m o.e We 06 m.m
PhhpLhLlplrhrlpp pbpb—pphbblbhh——__pr—hhlhthlrlFFlbbllhrpb—ppbb—»LI»P
Edd Se 2:. £4 84 3.3
LFPFFLtht»pbb—hbbphhpbrphht—prrh
0
see me o; m; 9N m.m o.m m.m o.e me 9m m.m
—~F>pr—rhblrppp—bFPF—hplubbrrbrbbbprph—pppprLrp»plphpphb»blhllerbph
f<
o J..—
I add 09m 36 31m

 

 

.I/
.I#...I.I\
I...

\.f

 

73

2. Bulk polymerization of substituted Iactides

Bulk polymerizations are solvent-free polymerizations, which are usually
run at temperatures higher than the melting point of the monomer. Because of
the high concentration of the monomer and high polymerization temperature
typically found in bulk polymerizations, the bulk polymerization rate and polymer
yield is very high. However, because of the high polymerization temperature, bulk
polymerizations are usually accompanied by side reactions which limit control of
the polymerization. In contrast, in solution polymerization the monomers are
dissolved in organic solvents and the polymerization can be conducted at very
low temperatures. Compared to bulk polymerization, solution polymerization
provides better control over molecular weight and molecular weight distribution.
For example, Iactide has been polymerized using both bulk polymerization and
solution polymerization. The bulk polymerizations are usually run from 110 °C to
180 °C and polymerizations finish within several hours. Side reactions such as
transesterification, racemization and discoloration always accompany
polymerization, and thus the control over molecular weight and molecular weight
distribution is poor. Solution polymerizations are usually run from 50 °C to 90 °C
and polymerizations may take days to finish. Solution polymerization provides
excellent control over the molecular weight and molecular weight distribution of

the polymers.

74

 

Zilnni

hdpe
ofidesi
variety
WOWdr
showr
bmad
POWnr
show
hexyl
Oflen
conv
533k
mole
265
mor
90h
bfler
aye
fro,

17‘

2.1 Initiator survey.

Many compounds have been used as catalysts in bulk polymerizations of
Iactide including metal halides,82.181.182 metal carboxylates,46 metal

oxides50»78 and many kinds of organometallic compounds.183v184 We tested a
variety of catalysts for their ability to polymerize substituted glycolides. 1H NMR
provides a convenient method for following the polymerization reaction. As
shown in Figure 11 for ethylglycolide, the methine peak at 4.85 evolves into a
broad peak at 5.05 ppm during polymerization, and thus the conversion to
polymer can be calculated by integration of the two signals. Tables 5, 6, and 7
show results from the polymerization of ethylglycolide, isobutylglycolide and
hexylglycolide respectively. As shown in Table 5, runs without alcohol initiators
often gave near-quantitative conversion to polymer, but SnO and SnBrz gave low
conversion and low molecular weights. Both SnO and SnBrz are heterogeneous
systems and the poor activity is not surprising. However, the number average
molecular weight obtained from soluble initiators varied greatly, ranging from
26,000 for Sn(2-ethylhexanoate)2 to > 100,000 for Ph4Sn. Given a 100:1
monomer to initiator ratio and complete participation by all initiator added to the
polymerization, the Mn for these polymerization should be near 17,000.
Membrane osmometry results obtained in toluene indicate that the number
average molecular weight of poly(ethylglycolide) and polystyrene determined
from GPC data are comparable, and thus number-average molecular weights >

17,000 are consistent with incomplete initiation. The effect of adding alcohol as

75

 

minu:

COINS

polvm
pohnr
ethyll

Thus.

of lac
inhal
Conn
adde
are l

impL

initiator (second row of each entry) is clear. All of initiators surveyed gave high

conversions and molecular weights close to the theoretical values for M...

Sn(2-ethylhexanoate)2 is the most efficient catalyst known for the bulk
polymerization of Iactide. Lactide polymerizations using this catalyst have high
polymerization rates and require small catalyst loads. Also, Sn(2-
ethylhexanoate); produces high molecular weight polylactides in high yield.

Thus, Sn(2-ethylhexanoate); was our primary choice of catalyst.

Because most proposed mechanisms for solution and bulk polymerization
of Iactide using Sn(2-ethylhexanoate)2 invoke participation of water or alcohol as
initiators, we added t-butylbenzyl alcohol (BBA) to polymerizations to gain better
control of the molecular weight and the molecular weight distribution. Without the
added alcohol, the polymerization is initiated by residual moisture or alcohols that
are present as impurities in the polymerization, and because the amount of such
impurities vary from run to run, the initiation efficiency should also vary. The
polymerization of ethylglycolide shows the importance of the added alcohol
(Figure 12). The results from six Sn(2-ethylhexanoate)2 initiated polymerizations
are shown, three without an alcohol initiator, and three where one equivalent of t-
butylbenzyl alcohol was added. Runs with added alcohol show a linear
relationship between molecular weight and conversion up to >60% conversion.
The data from the three runs without alcohol are more scattered, especially at
high conversion, showing that added alcohol improved the efficiency of initiation

and provides better control of molecular weight.

76

poll
molr
thec
for I
one
beha
has

attrit

and I

Figure 13 shows the evolution of molecular weight with conversion for the
polymerization of Iactide, ethylglycolide, isobutylglycolide and hexylglycolide. The
molecular weight increases linearly with conversion and nearly reaches the
theoretical values, 14,000 for polylactide, 17,000 for poly(ethylglycolide), 23,000
for poly(isobutylglycolide) and 28,000 for poly(hexylglycolide). These data show
one characteristic expected of a living polymerization, a linear relationship
between Mn and conversion. The drop in molecular weight at the end of the run
has been observed previously in Iactide polymerizations,92u185 and is usually

attributed to intramolecular transesterification reactions that form cyclic products

and decrease the number average molecular weight.

~4-

Enlr

Table 5. Bulk Polymerization of Ethylglycolide

 

 

 

Entry Catalyst %converiona Mnx10'3 Mwan
1 Sn(Oct)2 99 26.2 2.00
95 16.0 1.73
2 Ph4Sn 95 114 1.33
96 15.5 1.88
3 SnO 7 5.4 1.15
94 14.9 1.86
4 PhD 96 31.7 1.65
97 13.9 1.90
5 SnBr4 98 37.7 1.78
98 15.2 1.89
6 SnBl'z 61 13.3 1.64
97 13.9 1.82

 

[ethylglycolide]:[catalyst] = 100: polymerization mn at 180 °C for 2.5 hours

a. determined by 1H NMR; top and bottom entries for each catalyst are
data for polymerization with no alcohol added and with neopentyl alcohol
([ethylglycolide1/[aIcohol]=100) as the initiator, respectively.

78

En

hOurs

data

(Usab

Table 6. Bulk Polymerization of Isobutylglycolide

 

 

Entry Catalyst %converiona Mnx10'3 Mw/Mn
1 Sn(Oct); 97 37.1 2.00
96 22.2 1.65
2 Ph4Sn 95 50.2 1.97
96 21 .3 1.88
3 SnO 45 10.9 1.23
94 20.8 1 .84
4 PhD 96 41 .9 1.65
97 22.9 1.90
5 SnBr4 98 57.1 1.88
98 20.5 1 .78
6 SnBrz 76 31 .2 1 .84
97 21.8 1.89

 

[isobutylglycolide]:[catalyst] = 100: polymerization run at 180 °C for 2.5
hours

a. determined by 1H NMR; top and bottom entries for each catalyst are
data for polymerization with no alcohol added and with neopentyl alcohol
([isobutylglycolide]l[alcohol]=100) as the initiator, respectively.

79

Table 7. Bulk Polymerization of Hexylglycolide

 

 

Entry Catalyst %converiona Mnx10'3 Mw/Mn
1 Sn(Oct); 95 53.2 1 .92
97 26.8 1.93
2 Ph4Sn 93 43.4 1.83
96 25.5 1.78
3 SnO 31 13.5 1.19
97 27.9 1.81
4 PhD 94 67.7 1.95
98 25.9 1.82
5 SnBu 95 47.1 1.88
99 26.9 1.89
6 SnBrz 32 17.3 1.54
96 23.2 1.87

 

[hexyllglycolide]:[catalyst] = 100: polymerization run at 180 °C for 2.5

hours

a. determined by 1H NMR; top and bottom entries for each catalyst are
data for polymerization with with no alcohol added and with neopentyl alcohol

([hexylglycolide]l[alcohol]=100) as the initiator, respectively.

80

 

 

 

 

35
A
30 -
25 - 4
L
«9° 20 A
._
x
r:
2 15 - f
f”’
A x
10 " A 1’
x6
1”
5 ' I.’
,4
fl
0 l L 1 1
0 20 4O 60 80 100

conversion (%)

Figure 12. Bulk polymerization of ethylglycolide with (O) and without (A)
added t-butylbenzyl alcohol as co-initiator. Polymerization conditions: 130 °C,

[Sn(Oct)2]/[t-butylbenzyl alcohol] = 1, [Monomer]l[lnitiator] = 100.

81

 

30

 

 

 

 

 

O 20 40 60 80 100
Conversion (%)

Figure 13. Molecular weight versus conversion for the bulk polymerization of
substituted glycolides. (A) Iactide, (I) ethylglycolide, (O) hexylglycolide, (O)
isobutylglycolide. Polymerization condition: 130 °C, [Sn(Oct)z]/[t-butylbenzyl

alcohol] = 1, [Monomer]/[lnitiator] = 100.

82

2.2 Kinetics of Bulk Polymerization

Ring-opening polymerizations of Iactides and Iactones typically follow first
order kinetics that can be expressed by Eq. (1)

__d[M]=
R— T: kp[M][l] .......................... (1)

where [M] and [I] are the concentration of monomer and initiator, and k,, is the
rate constant for propagation. For the case of a living polymerization, [I] is

constant and integration of Eq. (1) yields:

[M L

— ln(
[M 10

):kp[1]0t ............................. (2)

where [M], is the concentration of the monomer at time t, [Mjo is the initial
monomer concentration (at t=0) and [ljo is the initial concentration of initiator. For
living polymerizations, [I] is constant and plots of -ln([M],/[M]o) vs. t are linear.
Figure 14 shows kinetic data for several substituted Iactides. After a fast initial
polymerization, the rate slows as the conversion exceeds 80% (-
In([M]./[M]o)=1.6). The data for low conversions are shown in Figure 15, and from
each data set, kp[l] values were extracted from slopes of the linear portion of the
the plots. However, the data in Figure 14 also show strong deviation from
linearity at conversions above 80%, and that the conversion saturated near 97%

conversion.

83

 

 

 

 

4
‘ A
A . I
3- ‘ - ’
I C
3 o
5
$5
22“ 2
‘c’ ‘-
' o
11 'O
o
‘:
I
0 .
o 20 40

Time (min)

Figure 14. Kinetics of bulk polymerization of substituted glycolides. (A)
Iactide, (I) ethylglycolide, (O) hexylglycolide, (O) isobutylglycolide.
Polymerization conditions: 130 °C, [Sn(Oct)2]/[t-butylbenzyl alcohol] = 1,

[Monomer]l[lnitiator] = 100

 

 

 

 

 

3
E. 2'
,2
E
E

1.

A
O I I
o 5 1o 15

Time (min)
Figure 15. Kinetics of bulk polymerization of substituted glycolides at low
conversion. (A) Iactide, (I) ethylglycolide, (O) hexylglycolide, (O)
isobutylglycolide. Polymerization conditions: 130 °C, [Sn(Oct)2]/[t-butylbenzyl

alcohol] = 1, [Monomer]l[lnitiator] = 100

85

 

From Figure 14 and Figure 15, we Ieamed that bulk polymerizations of
substituted Iactides are first order in the concentration of monomer and first order
in the concentration of initiator at low conversion. At high conversion, the
polymerization rate begins to drop and finally polymerization stops without

reaching 100% conversion. This phenomenon has been reported in the

literature.92.93 Witzke et 31.37 explained it using an equilibrium polymerization
model. However, we think there are two possibilities that can influence the
polymerization kinetics. First, the polymerization is an equilibrium reaction,
second, the initiator continuously degrades during the polymerization. The
possibility of catalyst decay came from the solution polymerization of
ethylglycolide, where we recovered a precipitate from the polymerizations. We
believe it has the structure shown in Figure 16. This precipitation may represent

catalyst lost from the solution polymerization (discussed later).

We used these two models to fit the kinetic data that we obtained from

bulk polymerization. If the polymerization is an equilibrium reaction, the kinetics

of the polymerization can be expressed using following equations.87

[M], = [MM +([M]0 -[M]eq)e_kpm’ ............................... (1)

86

 

where [Mleq is the monomer concentration at equilibrium, X is the conversion
(([Mo-[M],)/[M]o), and kp is rate constant for polymerization. We fit the data using

equation (2). The results shown in Figure 17 show that the fit was good.

If the initiator continuously degrades during the polymerization, the

kinetics of the polymerization can be expressed as:

_d[M]

= -kdr
dt kp[M][I]e ................................... (3)

p:

where kd is the rate constant for catalyst decay. Integration of equation (3) yields

[M]: kP 4"
-ln——— =—I (1- d .................................. 4
([M]0) kd[ ]o e ) ()

We also used equation (4) to fit the kinetic dat the result is shown in
Figure 18. Unfortunately, the data also fits equation (4) well. Thus, we still can
not decide which model correctly describes the kinetics of the Iactide

polymerization.

To decide which model is appropriate, we studied the depolymerization
reaction. If the Iactide polymerization is indeed an equilibrium reaction,
depolymerization should give us the same [Mleq as polymerization. We
scrupulously purified the polymers to remove all monomer in the polymer, and we

conducted polymerization and depolymerization using the same conditions. The

87

results are shown in Figure 19. The polymerization and depolymerization indeed
reached the same [M]9q, about 3%. The white precipitate recovered from solution
polymerization was tested as a catalyst for the bulk polymerization of
ethylglycolide. The polymerization also reached 97% conversion. These

experiments proved that the equilibrium model is the right model.

At low conversion, the polymerization is first order in monomer
concentration with characteristics of a living polymerization. For a first order

reaction, the kinetics can be expressed as:

d M
R =-—[-—]-=kp[M][1] ............................ (5)
dt
where [M] and [I] are the concentration of monomer and initiator, and kp is the

rate constant for propagation. For the case of a living polymerization, [l] is

constant and integration of Eq. (5) yields

_ LML _
111([M]0 ) _ k p[I]0t ........................ (6)

where [M], is the concentration of the monomer at time t, [M10 is the initial

monomer concentration (at t=0) and [I] is the concentration of initiator

From Figure 15, we know that the plot of -In([M],/[M]o) vs. tis linear at low
conversion. The slope is kp[I], which is the rate constant for Iactide

polymerization. The correct fit to the kinetic model should produce the same kp[l].

88

The results (Table 8) show that the kp[I] values obtained from the equilibrium
model are consistent with those obtained from the fit to the first order reaction,
and that the catalyst degradation model dramatically overestimates the k,,[l].
These k,,[l] data indirectly support that equilibrium model as the correct model to

describe the kinetics of Iactide polymerization.

89

H—0\ 0
C—C’ \' .x O§C C’
/ \\ "s- \ / \
R 0 I: O O’H
C.) I
H’ O I ,0
\c_c< \,Sn: >C C\’
R/ \O/ i O O—H
5
Sn

Figure 16. The structure of the white precipitate formed during the solution

polymerization of ethylglycolide using Sn(Oct)2/ROH as initiator

Table 8. The kinetic data for bulk polymerization of substituted Iactides

 

 

 

 

 

 

Kpm (sec‘1)><103
Monomer from from fit to from fit to catalyst
conversion equilibrium degradation
vs. time data model model

Iactide 6.2 6.7 15.1
ethylglycolide 4.8 4.7 10.8
hexylglycolide 3.7 3.7 6.7
isobutylglycolide 3.0 3.0 5.3

 

 

 

 

 

 

90

 

100

 

 

 

 

 

 

80
§
3 60
.9
9
(D
t): 40
8
A Poly(lactide)
I Poly(ethylglycolide)
20 e Poly(hexylglycolide)
O Poly(isobutylglycolide)
0 l l l
0 1 0 20 30 40

time (min)

Figure 17 Kinetics of bulk polymerization of substituted glycolide. (A)
Iactide, (I) ethylglycolide, (O) hexylglycolide, (O) isobutylglycolide.
Polymerization condition: 130 °C, [Sn(Oct)2]l[t-butylbenzyl alcohol]=1,

[Monomer]l[lnitiator]=100. The data were fitted using an equilibrium model.

91

 

 

 

 

.. / .
e
3
2
5‘1.
2. 2 ’
1?
1 _ A Poly(lactide)
I Poly(ethylglycolide)
,, e Poly(hexylglycolide)
I O Poly(isobmylglycolide)
0 .
0 20 40

Time (min)

Figure 18. Kinetics of bulk polymerization of substituted glycolide. (A) Iactide, (I)
ethylglycolide, (I) hexylglycolide, (O) isobutylglycolide. Polymerization condition:
130 °C, [Sn(Oct)2]/[t-butylbenzyl alcohol]=1, [Monomer]l[lnitiator]=100. The data

were fitted using the initiator degradation model.

92

conversion

100

95

90

 

 

 

 

 

_ o

.e- .p_e. - _ _ _ 3 ______ o
. ‘ . g I
9
_ e polymerization
- o depolymerization
0 10 20

time (minutes)

Figure 19. Polymerization/depolymerization data for the bulk polymerization of

ethylglycolide. Conditions: 200 °C [Sn(Oct)z]l[BBA]=1 [monomer]l[catalyst]=100

93

Ring opening polymerization of substituted Iactides is facile, though a bit
slower than Iactide itself. The propagation step of glycolide polymerization is
depicted in Scheme 31. The glycolide ring is arbitrarily drawn in a planar
conformation with the R group in an equatorial position, although a more realistic
representation would need to account for the boat-like conformation of the ring
seen from x-ray studies and the mixture of glycolide diastereomers used in the
polymerization. The polymerization rate of glycolide is faster than Iactide (R =
CH3), because the steric bulk of the methyl group hinders nucleophilic attack at
the ring carbonyls. Increasing the size of the ring substitutent should decrease
the polymerization rate further. Returning to Figure 15, we see that the rates of

polymerization follow the expected trend: lactide> ethylglycolide> hexylglycolide>

isobutylglycolide. A survey of Iactide monomers by Hall186 showed that ring
substitution plays a major role in defining the polymerizability of Iactide. For
example, 3,3,6,6-tetramethyl-1,4-dioxane-2,5-dione, obtained by adding two
methyl groups to the Iactide ring, does not undergo ring-opening polymerization.
Presumably, nucleophilic attack by either the initiator or the growing polylactide

chain is too hindered to lead to polymer.

Scheme 31. The propagation step for Iactide polymerization using

Sn(2-ethylhexanoate); as catalyst

R' O

\o
I ,o
Snz'V

0/5 “"0 R
b

95

3. Solution Polymerization of Substituted Lactides

Solution polymerizations exhibit different polymerization behavior than
bulk polymerizations. Because of the added solvent, the concentration of the
monomer is lower and the propagation rate, klele'], is reduced. Also, solution
polymerizations are run at lower temperatures than bulk polymerization, usually
lower than boiling point of the solvent. Because of the lower polymerization
temperature, control of molecular weight and the molecular weight distribution is
much better and there are fewer side reactions. The solvents used in solution
polymerizations of Iactide include toluene, THF and 0112012. The most commonly
used solvent is toluene. We choose Al(OiPr)3 and Sn(2-ethylhexanoate)2/alcohol
as initiators since Al(OiPr)3 is known to be a good initiator for solution
polymerization of Iactide, while the Sn(2-ethylhexanoate)2/alcohol system is
commonly used for bulk polymerization. We wanted to evaluate the Sn system
under solution polymerization conditions, since studying the more controlled
solution polymerization can help us understand more about the polymerization
mechanism when Sn(2-ethylhexanoate); is used as catalyst for bulk

polymerization.

3.1 Al(OiPr); as Initiator

Al(OiPr)3 is well-known to initiate living Iactide polymerization, where the
molecular weight grows linearly with conversion and the polymerization follows

first order kinetics. Plots of —In([M],/[M]o) versus t should be linear. The

substituted Iactides have structures similar to Iactide, so we should expect the
polymerization of substituted Iactides also to be living and follow first-order
kinetics. The molecular weight versus conversion data are shown in Figure 20
and a plot of —In([M],/[M]o) versus t is shown in Figure 21. As expected, the
polymerization of substituted Iactides using Al(OiPr)3 is a living polymerization
and follows first order kinetics. From Figure 20 we can see that Al(OiPr)3 initiates
three chains. The molecular weight of polymers can be predicted from the
equafion:

= [M] xM
" 3x[1] °

 

where M, is the number-average molecular weight of polymer, [M] is
concentration of the monomer, [I] is the concentration of the initiator, M0 is
molecular weight of monomer and n is the number of initiating species for the
Al(OiPr)3 complex. Since the observed molecular weight is one third that

predicted by the WW] ratio, each isopropoxide must initiate one polymer chain.

Based on the bulk polymerization results, we expected that the rates of
the polymerization should follow the order: Iactide > ethylglycolide >
hexylglycolide > isobutylglycolide, because the steric bulk of the ring
substitutents increase in that order. However, the solution polymerization rates
do not follow the expected order. Lactide, ethylglycolide and hexylglycolide do
follow the expected trend, but isobutylglycolide has the bulkiest side group and

the fastest polymerization rate.

97

To understand this phenomenon, we calculated the activation energy for
solution polymerization of substituted Iactides. The activation energy can be

calculated from the Arrhenius equation:

 

-152

k = Ae RT
Ink = lnA— Ea
RT

where k is the rate constant, A is a constant of integration called the frequency
factor, E, is the activation energy, R is the gas constant and T is the temperature
in Kelvin. When —In R is plotted against 1/T, the activation energy and frequency

factor can be extracted from slope and intercept, respectively.

Polymerizations were run at three temperatures: 70 °C, 90 °C and 100 °C.
The results are shown in Figure 21, Figure 22 and Figure 23 respectively. The
rate constants are shown in Table 9. The activation energies were obtained from
the plots shown in Figure 24 and the results are listed in Table 10. The results
are totally unexpected. We believed that bulkier side groups should hinder attack
at the carbonyl group on the Iactide ring, and lead to a higher the activation
energy. Our expectation that the bulkier side groups should result in the higher
activation energies is based on the assumption that the nucleophilic attack on the
carbonyl group is the rate-determining step. This might not be the case. The
polymerization of Iactide includes several mechanistic steps: coordination,

nucleophilic attack, and ring opening. Nucleophilic attack might be rate limiting

98

for linear alkyl substituents, but may not be the rate-determinating step for more

bulky or highly substituted Iactides.

99

 

10

\I’ .3

/A
O I T I I
0 20 40 60 80 100

conversion (%)

 

 

 

Figure 20. Molecular weight versus conversion for the solution
polymerization of substituted Iactides. (A) Iactide, (I) ethylglycolide, (O)
hexylglycolide, (O) isobutylglycolide. Polymerization conditions: toluene

as solvent, Al(OiPr)3 as initiator, 90 °C, [Monomer]l[lnitiator] = 100

100

 

1.2 -

 

 

 

E
E 0.8 f
E
E
E A

0.4 ’ ‘

I
I
o 41 I l l
0 5 10 15 20 25

time (h)
Figure 21 Kinetics for the solution polymerization of substituted Iactides at 70 °C.
(A) Iactide, (I) ethylglycolide, (O) hexylglycolide, (O) isobutylglycolide.
Polymerization conditions: toluene as solvent, Al(OiPr)3 as initiator,

[Monomer]l[lnitiator] = 100

101

 

-|n([M]t/[M]0)

 

 

 

 

Time (h)
Figure 22 Kinetics for the solution polymerization of substituted Iactides
at 90 °C, (A) Iactide, (I) ethylglycolide, (I) hexylglycolide, (O)
isobutylglycolide. Polymerization conditions: toluene as solvent, Al(OiPr)3

as initiator, [Monomer]l[lnitiator] = 100

102

 

1.6

’3

'— 1.

a 2

2.

E

.E 0.8
0.4

 

 

 

 

Time (h)

Figure 23 Kinetics for the solution polymerization of substituted Iactide
100 °C. (A) Iactide, (I) ethylglycolide, (O) hexylglycolide, (O)
isobutylglycolide. Polymerization condition: toluene as solvent,

Al(OiPr)3 as initiator, [Monomer]l[lnitiator] = 100

103

Table 9. Polymerization rate constants of substituted Iactides

 

kpx1 03 (LImoI-s)

 

 

 

 

 

 

Monomers 70 °C 90 °C 100 °C
Iactide 10.5 38.5 73.8
ethylglycolide 8.03 25.0 54.2
hexylglycolide 6.15 18.2 32.6
isobutylglycolide 22.2 40.7 56.6

 

 

 

 

Table 10. Polymerization activation energies for substituted Iactides

 

 

 

 

 

 

Monomers E. (kJImol) lnA
Iactide 68.8 21 .6
ethylglycolide 66.2 20.4
hexylglycolide 58.6 17.4
isobutylglycolide 32.9 9.7

 

 

 

 

 

 

 

 

 

 

-1.5 -
’3
5
E
-2.5 -
_3.5 L l 1
0.0026 0.0027 0.0028 0.0029 0.003

1IT

Figure 24. Activation energies for polymerization of substituted Iactides

(A) Iactide, (I) ethylglycolide, (O) hexylglycolide, (O) isobutylglycolide

105

3.2 Sn(Oct)2IROH as initiators

The Sn(Oct)2/ROH system is a very good catalyst/initiator system for bulk
polymerization, but Its use in solution polymerizations is rarely mentioned in the
literature. Also, solution polymerizations have a lower polymerization rate than
bulk polymerizations, so it is relatively easy to study the kinetics and mechanism

of the Sn(Oct)2/ROH system in solution.

That ROH is the initiator and Sn(Oct); is the catalyst often has been
claimed in the literature, but direct evidence is lacking. We compared solution
polymerization of ethylglycolide initiated by Sn(Oct)2 and by Sn(Oct)leOH,
where ROH is neopentyl alcohol. The results are shown in Figure 25. The
diamonds of Figure 25 are the data for polymerization initiated by only Sn(Oct);
where moisture and impurities have been scrupulously excluded. The rate of
polymerization (slope of curve) is very low, and the polymerization barely
reached 10% conversion. If moisture and impurities are not carefully excluded,
polymerizations often showed appreciable conversion. The squres at the top of
Figure 25 show results for a polymerization where one equivalent of neopentyl
alcohol was added as initiator. Both higher polymerization rates and conversions
are routinely obtained when ROH is used. Hydroxyl-containing compounds such
as ROH and H20 are truly the initiator for the polymerization. The 10%
conversion achieved in the bottom trace is likely caused by impurities. No matter
how hard you try to exclude moisture and impurities, small-scale polymerizations
always are affected by trace impurities. In the bottom trace, the impurities were

carefully excluded, which caused the low polymerization rate and low conversion.

106

To further understand the role of ROH in polymerization, we ran solution
polymerizations at different Sn(Oct)2/ROH ratios, while keeping the concentration
of ethylglycolide and Sn(Oct)2 constant. The results are shown in Figure 26 and
Figure 27. There are two important results. The polymerization follows first order
kinetics up to about 70% conversion, and the higher the ROH/Sn(Oct)2 ratio, the
faster the polymerization. However, as shown in Figure 27, the polymerization
rate did not increase by five when the ROH/Sn(Oct)2 ratio was increased by five,
which means that the kinetics of the polymerization are not first order in ROH.
We ran two solution polymerizations with ROH/Sn(Oct)2=1, and
[M]/[Sn(0ct)2]=100 and 200, respectively. The results are shown in Figure 28
and Figure 29. Figure 29 is the linear part of Figure 28. The results show that
when the amount of both Sn(Oct); and ROH double, the rate of polymerization
also doubles, which proves that Sn(Oct)2 and ROH together decide the kinetics

of the polymerization.

An explanation for the results is that the first step of the polymerization is
an equilibrium between Sn(Oct)2 and ROH as shown in Scheme 32. The true
initiatior is intermediate (A). The concentration of (A) can be calculated using

equation below:

K ___ [0ctSn0R][H0ct]
[Sn(Oct)2][R0H]

 

where K is the equilibrium constant.

107

When [Sn(Oct2]/[ROH]=1, the concentration of initiator (OctSnOR) can be

calculated using following equation:

 

[OctSnOR] = qusnmct), ][ROH]

so the concentration of initiator (OctSnOR) will double if [Sn(Oct)2] and [ROH]
both double. If [ROH] increases, [OctSnOR] will increases, but at a level defined

by equation above.

Scheme 32. The first step for Iactide polymerization using Sn(Oct)2/ROH

o
O—Sn—O + ROH
o
0 HO
4.
O—Sn— OR 0
A

108

 

 

 

 

0001:1013
DO
U
22" D
E 1:
2 1::
g a
5.
l1.
0
D
O§'M”OOOO 1
0 100 200 300 400

Figure 25 Kinetics for the polymerization of ethylglycolide initiated by
Sn(Oct)2/ROH and Sn(Oct)2. (Cl) Sn(Oct)2/neopentyl alcohol (9) Sn(Oot)2.

Polymerization temperature is 90 °C. [Ml/[Sn(Oct)2]=100, [ROH]/[Sn(0ct)2]=1

109

2.5

 

A
QIAI
O.‘A
21- .. ‘
O
. A
A
5°45- .'A
l—l .‘
2. -
$1..
E I
' .5
0.5-

 

 

0 f l l i

50 100 150 200
time (h)

0

Figure 26 Solution polymerization of ethylglycolide at different
Sn(Oct2)/alcohol ratios. Polymerization temperature is 90 °C. (0)
[Sn(OctzlllROH]=1/10, (I)[Sn(0ct2]/[ROH]=1/5, (A) [Sn(Octfl/[ROH]=1 .

ROH is neopentyl alcohol. [M]=0.2 mollL, [Ml/[Sn(Oct)2]=100.

110

 

-|n([Mlt/[M]0)

 

 

 

 

0 4L l

0 20 40 60
time (h)
Figure 27 Solution polymerization of ethylglycolide at different
Sn(Oct2)/alcohol ratios. Polymerization temperature is 90 °C. (0)
[Sn(Oct21/[ROH]=1/10, (I)[Sn(Oct2]/[ROH]=1/5, (A) [Sn(Oct2]l[ROH]=1.
ROH is neopentyl alcohol. [M]=0.2 mollL, [Ml/[Sn(Oct)2]=100.

111

 

 

 

 

3
09 e
.9 D a 1!?
o. D
22 ' . DD
.5... e P
2. 0 DD
.2.. .
C D
_'1 - 1:]
CI
QC]
El
OP 1 L L
0 100 200 300 400
time(h)

Figure 28 Solution polymerization of ethylglycolide at different
monomer/initiator ratio. ([3) [M]![I]=200 (Q) [M]/[l]=100. Polymerization

temperature is 90 °C [M]=0.2 mollL, [ROH]/[Sn(0ct)2]=1

112

 

 

 

 

1.5 -
E
E.
E
.— 1 -
E
—? o

0.5 -

O l A l

o 20 4o 60 80

time (h)

Figure 29 Solution polymerization of ethylglycolide at different
monomer/initiator ratio. (0) [M]/[l]=200 (Q) [M]/[|]=100. Polymerization

temperature is 90 °C [M]=0.2 moI/L, [ROH]/[Sn(0ct)2]=1

113

We also ran solution polymerizations of ethylglycolide at three
temperatures 70 °C, 90 °C and 110 °C. The results are shown in Figure 30 and
Figure 31. Figure 31 shows the linear portion of Figure 30. These figures
provide polymerization rates for ethylglycolide initiated by Sn(Oct)2/BBA at three
temperatures (Table 11). From these rates, we calculated the activation energy

for ethylglycolide polymerization (Figure 32 and Table 12).

Table 11 Polymerization rates for ethylglycolide initiated by Sn(Oct)2IBBA

 

 

 

k, x 103 (mollL-s)
Monomer 70 °c 90 °c 110 °c
ethylglycolide 0.31 2.43 13.1

 

 

 

 

 

 

[Sn(Oct)2]/[BBA]=1, BBA is t-butyl benzyl alcohol. [M]/[Sn(0ct)2]=100

Table 12 The activation energy for ethylglycolide initiated by Sn(Oct)2/BBA

 

monomer E. (KJImoI) lnA

 

ethylglycolide 75.68 21 .3

 

 

 

 

 

114

 

 

 

 

AA
2.5 - A
I I I
AA I.
’3 2 ' A I.
H I
E '5 .A I . e ...
U I
C I . O
-. .. u
1 I ...
I O
0.5 0'
O
.0
o l l
0 200 400 600
time (h)

Figure 30 Solution polymerization of ethylglycolide initiated by Sn(Oct)2/BBA.

[Sn(Oct)2]/[BBA]=1, BBA is t-butyl benzyl alcohol. (A) 110 °C (I) 90 °C (0)

70 °C [Ml/[Sn(Oct)2]=100

115

1.2

 

0.8 -

-|n([Mlt/[M]0)

 

 

 

O
0.4 o
0.2
O
0 1 1
0 50 100 150
time (h)

Figure 31 Solution polymerization of ethylglycolide initiated by Sn(Oct)2/BBA.

[Sn(Oct)2]/[BBA]=1, BBA is t-butyl benzyl alcohol. (A) 110 °c (I) 90 °C (9)

70 °c [Ml/[Sn(Oct)2]=100

116

 

ln(kp)

 

 

_5.5 l l I l
2.5 2.6 2.7 2.8 2.9 3

1rrx103

 

Figure 32 The activation energy of ethylglycolide initiated by Sn(Oct)2lBBA

117

As mentioned above, we isolated a white precipitate during the solution

polymerization of ethylglycolide. Vert et al.89 reported a similar precipitation
during polymerization of Iactide using Sn(Oct)2.This white precipitate does not
dissolve in organic solvents, so we could not directly identify it using NMR.
Indirect methods were used to identify this white precipitate. Dissolution of the
precipitation in dilute HCI solution, followed by extracting with ether, gave one
organic compound, 2-hydroxybutyric acid. Also, when 2-hydroxylbutyric acid was
added to Sn(Oct)2, a white precipitation formed instantly. These two precipitates

have almost identical IR spectra (Figure 33).

How does this white precipitate influence the kinetics of the solution
polymerizations? Is it umimportant, as was shown for bulk polymerizations, or
does it remove active catalyst from the polymerization and have a significant
influence on the polymerization kinetics? To answer these questions, we set up
the following experiment. A solution polymerization of ethylglycolide was run until
polymerization stopped. A second volume of ethyllgycolide solution equal to that
initially used was added to polymerization system. If formation of the precipitate
does not influence the kinetics of the solution polymerization, the polymerization
should return to its equilibrium state with polymerization rate halved. If all of the
catalyst has degraded to the white precipitate, and can not polymerize
ethylglycolide, then the polymerization would not return to its equilibrium state
and will remain near 50% conversion. The result is shown in Figure 34. The
initial polymerization reached a plateau at 93% conversion at about 54 hours.

Additional monomer solution was added, and the conversion dropped to about

118

50%. The polymerization then slowly retured to the plateau at >90% conversion.
Thus, we can say that the polymerization is indeed an equilibrium reaction.
However, we can not ignore the effect of catalyst degradation. As shown in
Figure 35, the polymerization rate at the beginning of the polymerization and
after the monomer solution was added are very different. The initial
polymerization rate is almost ten times faster than the polymerization rate after
the monomer was added. As mentioned before, if there is no effect from catalyst
degradation, the polymerizationrate after adding monomer should be half the
initial polymerization rate. Since the observed rate is 5 times lower than
expected, we conclude that ~80% of the catalyst had precipitated by the time the
second portion of monomer was added, and does not participate in

polymerization.

119

 

 

    

 

 

 

white precipitate from
2-hydroxy acid
:
.2
a
% white precipitate from
.3 polymerization
«l
4500 3500 2500 1 500 500

wave number (cm'1)

Figure 33. IR spectra for white precipitates formed during polymerization of

ethylglycolide and by mixing 2-hydroxylbutyric acid and Sn(Oct)2.

120

 

 

 

 

3
- <— monomeradded
2.5 - 1:1
I
E 2 ' :1
a - a
a :1
H15 P
E
E . D
I 1 b: [553
I
0.5 '
0 1 1 1 I
0 50 100 150 200 250
time (h)

Figure 34. Kinetics of solution polymerization of ethylglycolide showing the

result after adding extra monomer after the polymerization reached equilibrium.

Polymerization temperature is 90 °C. [M]/[Sn(0ct)z]=100, [ROH]I[Sn(Oct)2] = 1

121

1.6

 

 

 

 

1.2 ..
E.
g
F! 0.8 '-
E.
E

0.4 b

O l l

0 50 100 150

Time (h)
Figure 35. Initial polymerization rate for ethylglycolide and the decreased
rate observed after adding addtional monomer. Polymerization temperature

is 90 °C. [Ml/[Sn(Oct)2]=100, [ROH]I[Sn(Oct)2] = 1

122

Earlier, we saw that for solution polymerizations using Al(OiPr)3, the
polymerization rates for substituted Iactides did not follow the order predicted by
the steric hindrance of the substituent on the glycolide ring (Figure 21, 22 and
23). To determine whether the effect is initiator-related, we ran solution
polymerizations of substituted Iactides using Sn(Oct)2/BBA instead Al(OiPr)3. The
results are shown in Figure 36. The results are slightly different, but again the
polymerization rates did not follow the order expected based on steric
hinderance. For polymerizations using Al(OiPr)3, at 90 °C, the order of the
polymerization rate is isobutylglycolide > Iactide > ethylglycolide >hexylglycolide,
but for polymerization using Sn(Oct)2/BBA, the order is Iactide > isobutylglycolide
> ethylglycolide > hexylglycolide. These results reflect activation energy

differences for the two initiator systems.

3.3 Comparison of Al(OiPr)3 and Sn(Oct)2lROH as Initiators in Solution

Polymerization

Sn(Oct)2/ROH is the best initiator for bulk polymerization of Iactides, but it
is less effective in solution polymerizations. Figure 37 shows a comparison of the
Al(OiPr)3 and Sn(Oct)2/ROH initiators for solution polymerization of ethylglycolide
at 70 °C. The polymerization initiated by Al(OiPr)3 is almost 10 times faster than
the polymerization initiated by Sn(Oct)2/ROH. Considering that an Al(OiPr)3
molecule initiates three chains and Sn(Oct)2IBBA only initiates one chain, Iactide
chains in the Al(OiPr)3 system grow almost three times faster at 70 °C than those

initiated by Sn(Oct)2/BBA. Figure 38 shows that the same rate difference holds

123

at 90 °C. The activation energies calculated from the solution polymerizations are
shown in Figure 39 and Table 13. The activation energy for Al(OiPr)3 catalyzed
polymerization is ~10% less than for the Sn(Oct)2/BBA system. Theoretically, lnA
from both initiators should be the same because lnA represent the steric factors
of the monomer. The values of lnA calculated from the experiments are fairly

close.

Table 13 The activation energy of ethylglycolide initiated by Sn(Oct)2/BBA and

 

 

 

Al(OiPr)3,
initiator E. (KJlmol) lnA
Al(OiPr)3 66.2 20.4
Sn(Oct)2/BBA 75.7 21 .3

 

 

 

 

 

124

1.4

 

-|n([M]lI[M]0)

 

 

 

 

80

time (h)

Figure 36 Solution polymerization of substituted Iactides initiated by
Sn(Oct)2/BBA. Polymerization temperature is 90 °C, (A) Iactide, (I) ethylglycolide,
(I) hexylglycolide, (O) isobutylglycolide [Sn(Oct)2]l[BBA]=1, BBA is t-butyl benzyl

alcohol. [Ml/[Sn(Oct)2]=100. Toluene is the solvent.

125

1.4

 

-In([M]tI[M]0)

 

 

0 50 100
time (h)

 

Figure 37. Solution polymerization of ethylglycolide initiated by Al(OiPr)3 (O)
and Sn(Oct)2/BBA (I). [M]o=0.2 mol/L. [M]![l]=100, [Sn(Oct)2]/[BBA]=1

polymerization temperature is 70 °C and toluene is the solvent.

126

 

 

 

 

1.6 b
A 1.2 "
,2.
E
2'.
g 0.8 -
5 u
I

0.4

0 [ 1 1 1
0 20 40 60 80
time (h)

Figure 38. Solution polymerization of ethylglycolide initiated by Al(OiPr)3 (O)
and Sn(Oct)2/BBA (I). [M]o=0.2 mol/L. [M]![l]=100, [Sn(Oct)2]/[BBA]=1

polymerization temperature is 90 °C and toluene is the solvent.

127

 

ln(kp)

 

 

_6 J L l l
2.5 2.6 2.7 2.8 2.9 3

1rrx103

 

Figure 39. The activation energy for polymerization of ethylglycolide

initiated by Sn(Oct)2/BBA (I) and Al(OiPr)3 (9),

128

3.4. Polymerization Mechanism

3.4.1 Al(OiPr); as initiator

We think there are two possible scenarios that cause abnormal kinetic
behavior. First, the polymerization is a three-step process as shown in Scheme
33. It includes coordination (A), nucleophilic attack (B) and ring-opening (C).
When we expect polymerization rate to drop with increasing size of substituted
group, we assume that the nucleophilic attack step (B) is the rate-determining
step. However, the nucleophilic attack step (B) may not be the rate-determining
step and identification of the rate-determining step has not been addressed in
literature. It is possible that step B and step C are rate-determining. In step C, the
ring opens up to form a linear chain. In a ring system, the substituents on the ring
and the initiator are more crowded than that on a linear chain. The larger the
group, the more crowded the ring and the easier it is to open the ring. If step C is
rate-determining, larger substituents on the ring should lead to faster
polymerization. From the activation energy data from Table 10, we see that the
activation energy decreased with increases in the size of the substituents. We
believe that step C is the rate-determining step for isobutylglycolide
polymerization. The frequency factor lnA, which represent the steric effect of the
reaction, also is in the right order. The larger the size of the substituted group,

the smaller the value of lnA.

129

Scheme 33. The mechanism for the Iactide polymerization

initiated by Al(OiPr)3

 

R1 R1
Al(OR2)3
#0 ———> 0
R A
3
R20
R20\A
0 O ‘—C—_ R1
R20 0
R1 R20

 

A second scenario is based on the coordination number of Al during
Iactide polymerization. As mentioned in the Introduction, Al(OiPr)3 exists as a
trimer and a tetramer (shown in Figure 3) in solution. During Iactide
POlymerization, the aggregated complex is broken up by monomer to form
Complex (A) shown in Figure 40. The coordination number on Al is 6. During the
POlymerization of Iactide, ethylglycolide and hexylglycolide, complex (A) is
f0"fled. In the case of isobutylglycolide, it is may not possible to accomodate
three monomers on Al because of the large size of substituted group. It is may be

Possible to form the 4-coordinate complex (B) shown in Figure 40, with only one

130

monomer molecule coordinated to the Al atom. The 4-coordinate Al is more
reactive than 6-coordinate AI, and perhaps that is why isobutylglycolide
polymerized faster than Iactide, ethylglycolide and hexylglycolide. However, in
polymerizations of substituted Iactides initiated by Sn(Oct)2/ROH, the
polymerization rates also do not follow the order of the size of subsitituents, and

Sn(Oct)2/ROH does not aggregate. Thus, it is unlikely that this hypothesis is right.

131

Figure 40. The structure of the transition state for polymerization of

substituted Iactides

132

3.4.2 Sn(Oct)2IROH as catalyst/initiator

Based on the kinetic data, we propose the mechanism shown in Scheme
34. The first step is an equilibrium between Sn(Oct)2 and ROH to form RO-Sn-
Oct and HOct. Because of the non-polar environment, it is unlikely that either
Sn(Oct); or ROH will dissociate appreciably. We were unable to observe the RO-
Sn-Oct intermediate using low temperature NMR. One reason for our failure
could be that the equilibrium favors Sn(Oct)2 and thus the equilibrium
concentration of RO-Sn-Oct is below the detection limits of the NMR. A different
method must be developed to identify the intermediate. Recently, Penczek et al.
reported detection of RO-Sn-Oct during polymerizaiton using MALDI Mass

spectrometry.

The second step is initiation. The actual initiator is RO-Sn-Oct, a tin
alkoxide. Initiation by tin alkoxides has been well documented in the literature.
Though the equilibrium amount of RO-Sn-Oct is small, a fast equilibrium leads to
fast initiation. We know from kinetic data that polymerizations using
Sn(Oct)2/ROH are living polymerizations below 80% conversion. One

characteristic of living polymerization is a fast initiation reaction.

The first two steps of the polymerization are much faster than the third
step of the polymerization: propagation. The propagation of Iactides initiated by
metal alkoxides have been well studied in systems such as Al(OiPr)3 and
Sn(OMe)4. It is a coordination-insertion process. The fourth step of the

polymerization is the regeneration of catalyst and initiator. HOct reacts with

133

active chains to regenerate Sn(Oct); and a polymer chain with a hydroxyl chain
end. The hydroxyl chain end can re-initiate polymerization. The last step of the
polymerization is catalyst degradation. As mentioned before, Sn(Oct)2 degrades
to an insoluble white powder, the salt of Sn and the 2-hydroxy acid. In bulk
polymerization, this white powder can also polymerize Iactide, so it does not
influence the kinetics of the polymerization. However, in solution polymerization,
this white powder is insoluble, which means the catalyst is no longer in the

polymerization system and causes the polymerization rate to drop dramatically.

This reaction mechanism is similar to atom-transfer polymerization (ATRP)
(Sheme 35). The initiator reacts with catalyst to generate an active species,
followed by initiation and propagation, and then initiator and catalyst are
regenerated. ATRP is a living polymerization. Lactide polymerization initiated by
Sn(Oct)2/ROH is also a living polymerization at low conversion, and both have

termination steps.

134

Scheme 34. The mechanism of Iactide polymerization using Sn(Oct)2/ROH

as catalystflnitiator

ROH
SnOctz
ROMnOH
\J HOct
M, kd
V
HOct ROSnOct
M, k,
ROMSnOct
(SnM)x polymer
(insoluble)

135

Scheme 35. The mechanism for atom-transfer radical polymerization

 
  

(ATRP)
R'——Cl
+
’i‘
R'~[ACH2—C‘}CH2—CXY =R-c1
Y n ('3'

CUC'Z H2C:CXY

X
I
Y n

kdxi
X I_ I_
Y n

136

3.5. The Influence of Stereochemistry on the Kinetics of Solution

Polymerization

The stereochemistry of the monomer has a big influence on the kinetics of
solution polymerizations of substituted Iactides. As shown in Figure 41 and
Figure 42, racemic Iactide polymerized twice as fast as L-lactide, and racemic

isobutylglycolide polymerized three times as fast as D-isobutylglycolide.

Lactide monomers have two stereocenters, so racemic Iactides include an
equimolar amounts of the (R,R) and (S,S) stereoisomers. As shown in Scheme
36, when L-lactide or D-lactide are polymerized, there is only one propagation
reaction, an R active center reacting with (R,R) monomer with a rate constant
ka/RR, or an S active center reacting with (S,S) monomer with a rate constant
MKS/3,3). When racemic Iactides polymerize, four propagation reactions must be
considered, the active center with R stereochemistry reacting with (R,R)
monomer with a rate constant kmmg), the active center with R stereochemistry
reacting with (S,S) monomer with a rate constant mpg/3,3), the active center with S
stereochemistry reacting with (R,R) monomer with a rate constant kmsmm and
the active center with S stereochemistry reacting with (S,S) monomer with a rate
constant kp(s,s,s,. Usually, kmmg) is equal to MKS/5,3) and kpayss, is equal to
kp<S/R.R)- lf kp(S/R,R) and kp(R/S,S) > kp(s,s.s) and kpmm R), the racemic monomer will
polymerize faster than the pure D or L monomers, which is called syndiotactic

preference.

137

Scheme 36. Stereochemistry of Iactide polymerization

o
O 1
R2 O R o kMRIRJRI
\ -———>
‘0 A110R3>2 + o
R‘ R1
0
R (R. R)
o
O 1.
R2 o R “(mo ”M's,
‘0 ‘AI<0R3>2 + o —"
R‘ 'R1
0
R (S. S)

O

S (R. R)

O

O 1
RM.
R2 /u\/ iii/[LO was,”
‘0 ‘AI(0R3)2 + (kn/l. —__’

- "'R1
R1

0
s (5.3)

138

 

 

 

 

0.8 -
’a
"5" 0.6 L
E
E.
:5 0.4 -
O
0.2 -
0 1 1
0 2 4 6

time (h)
Figure 41. Polymerization of rec-Iactide (I) and L-lactide (O).

Polymerization temperature is 90 °C. [M]0=0-2 mollL, initiator is Al(OiPr)3,

[M]/[l]=100. Toluene is the solvent.

139

 

 

 

 

I
I
1.5 ~
I
E
E
E
H 1 l-
E
2.5
0.5 -
0 1
0 5 10
time (h)

Figure 42. Polymerization of rec-isobutylglycolide (I) and D-isobutylglycolide
(O). Polymerization temperature is 90 °C. [M]o=0.2 mollL, initiator is Al(OiPr)3,

[M]l[|]=100. Toluene is the solvent.

140

As shown in Figure 41 and Figure 42, polymerization rates of L-lactide
and rac-lactide are different. Similar results were obtained for the polymerization
rate of L-isobutylglycolide and rec-isobutylglycolide. Because of the
stereochemistry of the monomer and the chain end, there are four different
reactions during the polymerization rec-polylactides as shown in Scheme 36.
The kinetics of rac-lactide polymerization can be expressed as shown in Scheme
37. Equations A and C in Scheme 37 produce isotactic chains, so k5,“) =
[QR/RR) = k,~, while B and D produce syndiotactic chains, so k(s,R,R) = kHz/3,3) =ks. It

has been found that k. < ks in Iactide polymerization, which means syndiotactic

chain placement is favored in Iactide polymerization. Munson et al.187 found k.

lks =0.6 for the bulk polymerization of Iactide initiated by Sn(Oct); at 180 °C and

Kasperczyk188 found k; lks = 0.32 for solution polymerization of Iactide initiated
by tert-butoxide at 20 °C in THF. The method used to calculate k; Iks was
monitoring ratios of stereosquences by NMR. Because the NMR assignment of
the stereosquence is still a controversial topic, and the resolution of NMR is not
sufficent to clearly separate all stereosquences and account for side reactions
such as transesterification and racemization which will alter the stereosquence,
the ki lks ratio reported by Munson and Kasperczyk may not be very accurate. By
studying the kinetic behavior of rac-lactide and L-lactide polymerization, we

obtained more accurate measures of kilks.

141

Scheme 37. Kinetic scheme for Iactide polymerization

“918135)

—s' + $3 —_> —sss‘ A

kMSIRR)
—S' + RR ——> —SRR B

, kpima) ,
—— R + RR ——-> — RRR c
kp(RIS.S)
— R. 1' SS —-> —SRR' D

142

From Scheme 37, we can write the kinetic equation as:

d M 6 0 I I
“157] = k....,.,is iiSi+k..,..,.,[S 11R1+k....,..[R ][R]+k(R/s_.-,[R 1181-...(7)

Since

k(S/S.S) = k(R/R,R) = ki

k(R/S,S) = k(S/R,R) = ks

We re-write equation (7) as:

—i"[—M—i = k,[Si][S]+ks[Si][R]+k,-[Ri][R]+ kS[R'][S] ................................... (8)

In rac-lactide polymerization:

_ JM
[SI—[RI- 2

Equation (8) can be re-written as

—i[(%4—] = O.5k,-[Si][M]+ 0.5kS[Si][M]+ O.5k,-[Ri][M]+ 0.51., [MM] ........... (9)
— gig—i = 0.5k,([5‘]+ [Ri])[M] + 0.5k,([5* ] + [Ri])[M] .................................. (10)
Also:

[S']+[R']=[1]

143

Equation (10) can be re-written as:

_ aw]

.11 = (051-, + 0.51.5 )[I][M] ........................................ (11)

lntergration of (11) yields:

[M].
-1 —= 0.5k.+0.5k It .......................................... 12
“Wk; ( . .)[] ( )

Plots of -ln([M],/[M]o) vs tgive (0.5K,- + 0.5 ks)[I] as the slope.

In L-Iactide polymerization, there is only one monomer, so the kinetics of

the polymerization can expressed using equation (13):

M
— 111% = k,[1]t ............................................................. (13)

k,- [I] can be obtaied from plot of —ln([M],/[M]o) vs t. If a L-lactide and a rac-
lactide polymerization is run under the same conditions, [I] is the same for both
polymerizations. From these two slopes, k. lks can be solved. For polymerization
of Iactide at 90 °C, k. lks = 0.29, when initiated by Al(OiPr)3 in toluene, which is
close to the value (0.32) reported by Kasperczyk. For polymerization of
isobutylglycolide, k; lk8 =0.24 under the same conditions. From k. Iks, the

difference in activation energy can be calculated using the Arrhenius equation.

144

polyn

iacti
the
con
an
Syl
ple

Fi

ic

—(AE,.—A_ES)

_i_:e RT
k

S

(AE,-AES)=—RT1n%

S

For polymerization of Iactide, the energy difference is 3.7 KJ/mol. For

polymerization of isobutylglycolide, the energy difference is 4.2 kJ/mol.

We also used semi—empirical molecular mechanics calculation to simulate
Iactide polymerization. The calculations were done using SPARTAN. We set up
the two complexes shown in Figure 43. Complex A has a chain end with R
configuration and a coordinated S,S Iactide ring. Complex B has a chain end with
an S configuration and a coordinated S,S Iactide ring. Complex A leads to
syndiotactic chain placement and complex B will produce isotactic chain
placement. The optimized structures of complex A and complex B are shown in
Figure 44 and Figure 45 (Images are presented in color). In Figure 44, the O1
and AI bond is almost perpendicular to the Iactide ring, so it is in the right position
for nucleophilic attack. We drove 01 close to C10 on the Iactide ring until the new
O1-C10 bond was formed and monitored the energy of the complex along the
pathway. The transition state energy of nucleophilic attack can be obtained from
this energy trajectory. From the energies of complex A and the transition state,
the activation energy of nucleophilic attack can be calculated, which is the
activation energy for syndiotactic chain placement. Using the same method, the

activation energy for producing isotactic chain placement was also calculated.

145

betwee

dose.

Ta

user
0011
the

diff.

is

Ste

between calculated value and experimental value is shown in Table 14. It is fairly

close.

Table 14. Experimental value and calculated value for activation energy

 

 

 

 

difference.
AE, - AE. (KJ/mol)
Iactide isobutyglycolide
Experimental value 3.7 4.2
Calculated value a 4.5 5.1

 

 

 

 

 

a. Calculated using SPARTAN.

The SPARTAN calculation has strong limitations. First, the calculation is
used to simulate gas phase, which is very different from actual reaction
conditions. Second, since a semi-empirical method was used in the calculation,
the accuracy of calculation may not be good enough to judge the small energy

difference shown in Table 14.

The preference for syndiotactic chain placement for Iactide polymerization
is cause by small differences in activation energy. which is probably caused by

steric hindrance at polymer growing site.

146

O
H H
0

CH5. \ O ""‘CH3

0 -.
W‘- O ,.I' CH3 CH3
l
O

H
O

A

R chain end and S S Iactide ring

\ o
O
H: H
' 0
CH3 \ O “HICHG
O’AIIIIHIO O
_ O

o I H
/OH H3 H
° ‘1’

S chain end and S S Iactide ring

Figure 43. The structure of complexes used to simulate polymer growth

147

as 07 0 H27
2
H25 ,
‘21.
111- ,
llz . 4 f
01
c3 H1
02 1
- H3
()9 "23
\ 011
I 10
7 H12 llfl "‘..._~ 0‘
'9. ' (:13 013
H22 32
HS
a0 v H30
c5
r c1
" \ 0” .,. ‘ in?
H28
cs
H5
H15

Figure 44. The calculated structure for the complex with R configuration chain

end and 8,8 Iactide ring.

148

 

H15
11
c5
"7 I‘- 05
I
.3
H22
3 1121
03
)10 i 0,2
cm 013
011 . ’
'
H [:0 1117
,.
c17
. ' ' H18
C15
' H27

Figure 45. The calculated structure for the complex with S configuration

chain end and 8,8 Iactide ring.

149

4. Pol

give 1
0111111
0013
5qu
mi
pol)
wei

$111

of

ii]

4. Polymer properties

To determine the properties of the polymers, polymerizations were run to
give molecular weights near 50,000 g/mol (Table 15). The crude polymers were
purified by precipitation into methanol and dried to constant weight. All polymers
obtained from the polymerization of substituted glycolides are colorless, and are
soluble in solvents ranging from toluene to CHCI3. Flexible films can be either
melt pressed at 150 °C or cast from solvent. Because of the low Tgs for the
polymers, the films tend to be somewhat tacky regardless of the molecular
weight, and characterization of the polymers by polarizing optical microscopy
showed that none of the polymers is crystalline. Given that the monomers used
in the polymerizations is a mixture of RR, 8,8, and R,S diastereomers, the lack

of crystallinity is not surprising.

DSC scans were used to measure the glass transition temperatures for
the polymers, and the results appear in Figure 46 and Table 15. Overall, thermal
analysis data show that Tgs of the substituted glycolides range from —37 °C for
poly(hexylglycolide) to 66 °C for polylactide. For polymers substituted with linear
alkyl groups, The Tgs decreased as length of the alkyl group increased. For these
polymers, the flexible pendant group reduces T9 by acting as “internal diluent”,
lowering the frictional interaction between chains. Conversely, the branched
pendant group of poly(isobutylglycolide) hinders rotation of the polymer

backbone, resulting in a higher T9.

150

and

010

line

We further investigated the properties of poly(ethylglycolide) using thermal
and dynamic mechanical analyses. DSC runs show a T9 near 12 °C, with no sign
of crystallinity. A T9 of 12 °C is lower than that of polylactide itself (66 °C), and in
line with our expectation that increasing the length of the side chain should lead
to a decrease in T9. DMA runs support the Tg assignment made from DSC data.
As shown in Figure 47, the tan 5 trace for the polymer shows a peak near 12 °C
which is nearly identical to the baseline inflection seen in the DSC data. In
addition, the DMA probe position trace shows the expected behavior for a
polymer heated to above T9, expansion followed by penetration of the probe

through the sample.

The decomposition of polymers measured using TGA define the limiting
use temperature of the polymer. Figure 48 shows the TGA plots for the
decomposition of poly(ethylglycolide) in air and under N2. The two data sets
show only slight differences, an indication that the decomposition is probably
dominated by depolymerization of the polymer to monomer. As shown in Figure
49, the TGA plots for all polymers are similar, with the onset for decomposition
shifting to higher temperatures as the size of the alkyl group increases. We found
that all polymers decomposed to their monomers, and thus the shift likely reflects
the lower volatility of monomers that have large substituents. In trial
depolymerization reactions run on 0.1 g scales, we recovered over 93% of

polymer mass as monomers.

151

 

Poly(lactide)
_.——-—'—'/‘-—

Poly(isobutylglycolide)

Poly(ethylglycolide)

Poly(hexylglycolide)

heat flow (endo) —>

 

 

 

l j l l l L l I

-80 -60 -40 -20 0 20 40 60 80 100

 

temperature (°C)

Figure 46 DSC runs (second heating after flash quenching from 100 °C) for

substituted poly(glycolide)s. Heating rate: 10 °C/min under helium.

152

Table 15. Polymer Properties

 

 

 

 

 

 

Polymer Mnx10'3 3' MwIMn T, °C b

polylactide 35.2 1 .89 22-65
Poly(ethylglycolide) 45.6 1.78 15
Poly(hexylglycolide) 43.2 1 .91 ~37
Poly(isobutylglycolide) 47.3 1 .83 22

 

 

 

 

a. measured by GPC in THF using polystyrene as standard

b. measured by DSC under He at a rate of 10 °C/min

153

 

 

DSC

probe position

 

 

 

 

 

-55 -35 -15 5 25
temperature (°C)

Figure 47. Thermal analysis results for poly(ethylglycolide). The bottom 3
traces are DMA results for a sample in a parallel plate geometry. Heating rate:

10 °C/min under helium (DSC) or N2 (DMA).

154

 

100 -

 

80-

sample weight (%)

 

 

 

 

0 100 200 300 400 500 600
temperature (°C)

Figure 48 Thermogravimetric analysis results for poly (ethylglycolide).

Heating rate: 40 °C/min.

155

 

 

100

 
  
       

\" Poly(hexylglycolide)

30 r "
S
-o-0 Poly(lactide) —>
I: 60 -
.9
o
3
.23 40 r Poly(ethylglycolide)
E
:0
Cl)

20 -

Poly(isobutylglycolide)

 

 

0 l L
200 250 300 350 400 450 500

 

temperature (°C)

Figure 49 Thermogravimetric analysis results for substituted poly(glycolide)s

run in air. Heating rate: 40 °Clmin.

156

 

00ml
trent
chai
ternl

brar

cha
than
incr
9101
dist
cha
res
sut
Qro
Sn1
gka
the

Ski

Ser

9r0

The glass transition temperatures of the substituted polylactides and
comparable substituted polyethylenes are shown in Table 16. There is a similar
trend in the glass transition temperature with changes in the structure of the side
chain. For polymers substituted with linear alkyl groups, the glass transition
temperature decreased as length of the alkyl group increases. Conversely,

branched pendant groups result in higher glass transition temperatures.

The glass transition temperature is dependent on the flexibility of polymer
chain and secondary forces between polymer chains. When the group changes
from methyl to ethyl to hexyl group, the cross-sectional area of the group
increases slightly. However, because of the linear nature of these substituted
groups, they are very flexible. They act as “internal plasticizer" to increase the
distance between polymer chains and lower the frictional interaction between
chains. The net effect is to reduce the rotational barriers of the backbone, which
results in a decrease of the glass transition temperature. For branched
substituents, such as isobutyl and isopropyl, the cross sectional size of the
groups is larger than the in linear counterparts, and the flexibility of the group is
smaller, which result in a higher rotational barrier for the backbone and a higher
glass transition temperature. Compared to isopropyl, the extra methylene unit of
the isobutyl group increases the flexibility of the group, which decreases the

glass transition temperature.

There is one important difference between the substituted polylactide
series and the polyethylene series. Polylactides with branched substituted

groups, poly(isobutylglycolide) and poly(isopropylglycolide), have lower glass

157

transiti01
poiy(isol
tempera
betweer
strong 0'
This int
transitio
strong,

further

iSOprop
interact
DOIy(isc
barrier

weaker
transitii
the int;
of ihe

the

p°iypri

transition temperatures than polylactide. In the substituted polyethylene series,
poly(isobutylethylene) and poly(isopropylethylene) have higher glass transition
temperatures than poly(propylene). The difference lies in the secondary forces
between polymer chains in both series. Polylactides are polyesters, which have
strong dipole-dipole interactions between polymer chains as shown in Figure 50.
This interaction hinders rotation of the backbone, which increases the glass
transition temperature of polymer. When R is small, the chain-chain interaction is
strong, but when R is large as in the isobutyl and isopropyl cases, the chains are
further separated and the interaction is weak. The larger size of isobutyl and
isopropyl groups increases the rotational barrier, but the weakened secondary
interaction between polymer chains decreases the rotational barrier. Thus, for
poly(isopropylglycolide) and poly(isopropylglycolide), the increase in rotational
barrier gained from increasing the size of the substituted group can not offset the
weakened interaction between polymer chains, which results in a lower glass
transition temperature than polylactide. For the substituted polyethylene series,
the interaction between polymer chains is very weak because of the low polarity
of the polymer backbone. The rotational barrier is mainly decided by the size of
the substituted group. That is why poly(isopropyethylene) and

poly(isobutylethylene) have higher glass transition temperatures than

polypropylene.

158

Table 16. Glass transition temperature of substituted polylactides and

substituted polyethylene.

 

 

 

 

 

 

 

 

 

 

polymer T9 (°C) polymer To (°C)

polylactide 45-65 Poly(propylene) -15 - -3
Poly(ethylglycolide) 12 Poly(1-butene) -50
Poly(hexylglycolide) -37 Poly(hexylethylene) -65
Poly(isobutylglycolide) 23 Poly(isobutylethylene) 29
Poly(isopropylglycolide) 50 Poly(isopropyethylene) 50

 

159

 

   

Figure 50. The secondary interaction between polyester chains

160

5. Cop<

twor0'
togefli
asymi

copol

5. Copolymerization of substituted glycolides

The copolymerization of substituted glycolides has been done using the
two routes shown in Scheme 38. In route 1, the two comonomers were mixed
together and heated up with catalyst to make copolymers. In route 2, the
asymmetrically substituted glycolides were synthesized and polymerized to make

copolymers.

Scheme 38 Copolymerization of substituted glycolides

o o
R‘ #TLO Sn(Oct)2/ROH
R1 + \n/kRz A
o o
RTE]: Sn(Oct)2/ROH
\n/iRz ROW/RC0 (2)

 

    

161

5.1.

copc
PIOI
E0.
001'

MC

inl

si

5.1. Copolymerization through comonomers

Ethylglycolide (EG) and Iactide (LA) have very similar structures. Their
copolymers should retain the biodegradability, but have different physical
properties than their homopolymers. A series of copolymers (EG/LA=1I5,
EG/LA=1/3, EG/LA=1, EG/LA=3, EG/LA=5) were synthesized by bulk
polymerization at 130 °C catalyzed by Sn(Oct)2/BBA. All polymers have
molecular weights >40,000 g/mol. The DSC plots are shown in Figure 51. The
DSC runs show one glass transition temperature for each copolymer, an
indication that the copolymer is not phase separated. Considering the structural
similarity of the monomers, it is not surprising that copolymers are single phase.
The glass transition temperatures listed in Table 17 fall those of polylactide and
poly(ethylglycolide) and increase with an increase the mole fraction of the Iactide
comonomer. The glass transition temperatures of copolymers can be predicted

using the Fox equation:

where T9, Ty. and T93 are glass transition temperature of the copolymer, and the
pure homopolymers derived from monomers A and B respectively, and w
represents the mass fraction of the polymer. The Fox equation successfully
predicts the glass transition temperatures of random copolymers and plasticized
systems. An underlying assumption is that the polymer is homogeneous in

composition, and that the comonomers are distributed randomly along the chain.

162

As sh
equa‘
or nr

5805

and

far

As shown in Figure 52, the glass transition temperature of copolymers fit the Fox
equation fairly well. However, whether the copolymer is truly a random copolymer
or not must be defined by a characterization technique such as NMR, which is

sensitive to structure at the molecule level. Based on copolymerizations of Iactide

and glycolide]:189 the distribution of monomer units in the polymers is probably
far from random because differences in reactivity for the monomers. As
illustrated in Scheme 39, these effects are likely due to differences in the steric
bulk at the carbon or to the carbonyl. Lactide and glycolide share the same ring
structure, but the absence of lactide’s methyl groups makes glycolide far more
reactive and thus “blocky” polymers are obtained from copolymerizations.
However, transesterification reactions can increase the randomness of

copolymers.

163

Table 17. Properties of copolymers

 

 

 

 

 

 

 

 

 

Polymers M.,x10'a 3 "Jun T9 °C b
Polylactide 35.2 1 .89 66
EGILA=1/5 41.5 1.85 42.
EGILA=1I3 47.8 1.76 38
EGILA=1 41.7 1.88 30
EGILA=3 50.8 1 .92 23
EGILA=5 43.6 1.85 19
Poly(ethylglycolide) 45.6 1 .78 15

 

 

 

 

a. measured by GPC in THF using polystyrene as standard
b. measured by DSC under He at a rate of 10 °Clmin

164

 

 

Polylactide

_.—-v—'—"—'/—_—

EGILA=1/5
._.__._»/—‘_‘—_

EGILA=1/3
Jm

EGILA=1 f/h—

 

endo —>

Poly(ethylglycolide)

f

l l

-50 0 50 100
Temperature ( °C)

 

 

 

Figure 51. DSC runs (second heating after flash quenching from 100 °C) for

Iactide and ethylglycolide copolymers. Heating rate: 10 °Clmin under helium.

165

 

70

 

 

 

 

50 1 curve from
A Fox uation
0 eq
0
‘5
I'-

30 5

10 I U U I

0 0.2 0.4 0.6 0.8 1
W1/(W1'i'W2)

Figure 52. Glass transition temperatures of Iactide and ethylglycolide copolymers.

166

Scheme 39. Reactivity difference leads to “blocky” copolymers.

RO‘

less more
hindered hindered

polymerizes
preferentially

glycolide rich Iactide rich

 

167

5.2

5.2 Copolymerization through asymmetric monomers.

Simple copolymerization has some shortcomings. If one of the
comonomers is not polymerizable, the copolymer cannot be made. Also, the
mismatch in reactivity between the comonomers leads to inhomogeneous
incorporation of the comonomers into the polymer chain. These problems can be
solved by designing the AB monomers shown in Scheme 40. A variety of AB
monomers have been synthesized. As shown in Scheme 41, each monomer
contains two sites for ring opening, and to a first approximation the two carbonyls
of the ring can be treated independently. One site should have a reactivity similar
to Iactide, while the other should more closely resemble that of a substituted
Iactide. Thus we expect that the active end of growing Iactide chain will attack the
least hindered site on the AB monomer (the Iactide carbonyl) at a rate somewhat
slower than for Iactide, and ring will open to give the substituted lactic acid

residue at the growing chain end.

AB monomers also have limitations. AB monomers only produce
copolymers whose composition is 50% A and 50% B. Using mixtures of

comonomers can produce copolymers with any composition of A and B.

168

Sci

Scheme 40. The comparison between two copolymerization methods

111$. 2.311941.

Because monomer A is not polymerizable, the compolymer can not be synthesized

O
O
_.__’ O
0%
O
n
0
Using an AB monomer, the copolymer can be synthesized

Scheme 41 . Structures of the AB monomers and reactivity of different sites

0 o
O
O O
O
\[I/‘V o O
O O
tflmgthngchofldg ethylmethylglycolide methylphenylglycolide

hindered site
( for ring -opening

 

0 1L
0 kg—

 

preferred site
of ring-openning

169

5.2.1

CODI

syn

trar

(I)

5.2.1. Synthesis and Polymerization of AB Monomers.

As shown in Scheme 42, the AB monomers were prepared by
condensation of a—bromopropionyl bromide and the desired lactic acid. In this
synthesis, it is important to remove all of the NEta during the work-up, since even
trace amounts of NEta will polymerize the monomer to low molecular weight
oligomers. Racemic a-bromopropionyl bromide and racemic lactic acids were
used in the syntheses. As shown in Figure 53, the monomers are not statistical
mixtures of R,R, 8,8 and R,S diastereomers. In the case of ethylmethylglycolide,
the diastereomers form in a 3:1 ratio. For methylphenylglycolide, the selectivity is
so high that the minor diastereomer is barely seen in the 1H NMR spectrum. After
recyrstallization, the pure diastereomer can be obtained. To identify the major
isomer, we obtained the Nuclear Overhauser Effect (NOE) difference NMR
spectra of methyphenylglycolide. As shown in Figure 54, methines H8 and Hi,
can be attached to the same side of the ring or opposite site of ring. If Ha and H,
are on the same side of ring, the signal of the Hb will be enhanced if H. is
radiated. As shown in Figure 55, the intensity of H), increased when H, was
radiated. Thus, the major diastereomer is a mixture of the R,R and 8,8 isomers,
and the minor isomer is the R,S, or SR isomer. The kinetics of ring-closure
probably favor the formation of the R,R and 8,8 isomers, and ring-closing
competes with formation the oligomers. The selectivity can directly result from
the ring closing step, or the selectivity can arise from different rates for

conversion of the linear dimers to Iactides or oligomers.

170

high
stow)
knov
ope
trim

poi

Ethylmethylglycolide and methylphenylglycolide can be polymerized to
high molecular weight polymer easily. However, trimethylglycolide polymerizes
slowly. Even after 24 hours at 180 °C, the conversion is only about 75%. It is
known that the more substituents ring has, the more difficult it is for the ring to
open. Lactide and other substituted Iactides all are di-substituted, and
trimethylglycolide is tri-substituted. One extra substituent makes it hard to

polymerize.

171

Scheme 42. Synthesis of AB glycolide monomers

O
OH + Br R2
R1)\[r Rz/Kn/ —Z—’ >—-"/kRi
O 0 Br 0
acetone, NEt3
O
RK‘JK
O
0\n/i\
R1
0
O
H O 0 HI, 0 CH3
O—goicm 0—<\ 011.,
(R, R) (R, S)

Figure 54 Structure of the diastereomers of methylphenylglycolide

172

 

 

Minor iSt

Ha

Major isomer

Hb /
i J Minor isomer
__., i a 2’ - 111.

 

 

 

 

Major isomer

Minor isomer 0

' ii 1 ‘i I 0%
11 . ‘.
i

" 1 .
Ii ’ ,I:
1": . ii. 15‘”. .' ll ii 0

~ V: L ’ r r. ‘1. l
iv \/ =4. ”"J;J J / ".J

,i 1'11
1
WflJw/“L-I'

 

TrTTYTIFTTrIITfFFIT—Tl

T I I T l T—TTT—Tfi'T—fl

5.1 5.0 4.9 4.8

 

 

 

T T I I T T T T l T I T Y i T I I l T T T 1 T I T T T I Y I I I

Figure 53. 1H NMR spectra of ethylmethylglycolide and

methylphenylglycolide

173

ou__oo>_m_>cocq_>£oE ho E4583 mzz m0: .3 239....

 

 

I'ti bl? 1') [7' L It Iii t iii |

 

174

5.2.2 Str

Who
reactions 3

reiative siz

 

 

 

smaII, 101
°°P0iymer
dominates
If either F
reaction I

form.

is not Ia
know [I
po‘Yietr
R1 is m
imp.

mande
DOIy(e1

dEieCt

piliyml

reSidu

5.2.2 Structure of the polymer chains.

When AB monomers polymerize, there are four possible propagation
reactions as shown in as Scheme 43. Which reaction dominates depends on the
relative size of the R1 and R2 side group. If the size difference of R1 and R2 is
small, four reactions co-exist in the polymerization leading to random
copolymers. If the size difference of R1 and R2 is large, one of the four reactions
dominates, and an alternating (AB)n polymer with head to head defects will form.
If either R1 or R2 is so large that attack from active center is totally blocked, one

reaction exists in the polymerization, and a perfectly alternating copolymer will

form.

For ethylmethylglycolide, R1 is methyl and R2 is ethyl. The size difference
is not large, and from the homopolymerization of Iactide and ethylglycolide, we
know the reactivity difference of two monomers is not large. We expect that
poly(ethylmethylglycolide) is a regio-random polymer. For methylphenylglycolide,
R1 is methyl and R2 is phenyl. The phenyl group is much larger than the methyl

group, but from studies of the polymerization of mandelide, we know that the

mandelic acid residue should have some reactivity. We expect that

poly(ethylmethylglycolide) is an alternating polymer with some head to head

defects. For trimethylglycolide, we know that tetramethyglycolide is not
polymerizable, so we expect that active center can only attack the lactic acid

residue and give poly(trimethylglycolide) as a perfectly alternating polymer.

175

 

 

 

The
Tra
001
ad;
an
16
till
T1

si

tc

fr

We used 13C NMR to determine the chain structure of the AB copolymers.
The carbonyl region of the 13C NMR of AB copolymer is shown in Figure 56.
Trace A is the 13C NMR spectrum of poly(ethylmethylglycolide), which is very
complicated. There are signals from two different carbonyl groups: one
adjacent to methyl group with chemical shifts from 169 ppm to 169.8 ppm
and one adjacent to the ethyl group with chemical shifts from 168.2 ppm to
169 ppm. Both carbonyl groups show multiple peaks, which are caused by
the randomness of the chain structure and the stereochemistry of the chain.
Trace B is the 130 NMR spectrum of poly(trimethylglycolide), which shows a
simpler pattern. Again, there are signals from two different carbonyl groups:
one adjacent to the methyl group with chemical shifts from 168.7 ppm
to169.6 ppm and one adjacent to the dimethyl groups with chemical shifts
from 170.3 ppm to 171.3 ppm. Since tetramethylglycolide is not
polymerizable, we think that the nucleophilic attack ocurred exclusively at the
Iactide residue and that poly(trimethylglycolide) may be a perfectly alternating
polymer. Although the carbonyl resonances also show multiple peak
patterns, this may be caused by the stereochemistry of the chain, because
racemic monomer was used. Trace C is the 13C NMR of
poly(methylphenylglycolide). Both carbonyl groups have multiple peak
patterns, although the resolution is not very good. Because the monomer
used in synthesis of AB copolymers are not optically pure, the spectra
contains the information of both stereochemistry of chain and structure of

chain. It is very difficult to determine the structure of chains. To

176

U03

are

Sch

unambiguously determine the structure of the chain, optically pure monomers

are needed. The synthesis of optically pure monomers is still ongoing.

Scheme 43. The propagation step in the polymerization of AB monomers

1
___> head to tail

——2—> head to head

 

O
4
O R‘ ——-> head to head
R2 w 0—2:
R2

_:L, head to tail
[Mi]: 3

177

 

 

 

 

 

 

 

 

 

 

T I l V V V I T f fi T I V V V I ' V I T V V V l V fi V V I" f r I V V l
170.5 169.5 168.5 167.5 ppm
it
. i
l T". I
,1; i j t. 1
i i " a r. i
l _\" \ ’ v’ \ \ "x
W,va.. A A/J VA] \w‘—/ \WW ’J’J \W N/ J M) \\m
l I I l I 1 l I I Y I Y I I I Y I ' ' ‘
171.5 170.5 169.5 168.5 ppm
.1 C
i ‘1
I
l . J
1 . I
i / ‘v’ .
i \\ a" i
«I: it i i
I \_ ‘ii .‘ VA~
f] j " (I '1 rj/ \
H’l 2" J 1‘
x ,r \m
.nrvx/‘mm_~~ Maw" ‘ww/‘N/V'W’r'm ‘NvW-I‘MAW'KNM/y \‘\k.-.V~W.AN
T T I T T Y T T T T 1' Y I I I Y Y I I f f T I I I I T I T 1 7 T I 7 I I Y 7 7 Y I I I Y Y 1
170.5 169.5 168.5 167.5 166.5 ppm

 

 

Figure 56. Carbonyl region of 13C NMR of AB copolymers. A is
poly(ethylmethylglycolide), B is poly(trimethylglycolide) and C is

Dolv(methvl0henvlolvcolide)

178

5.3 The

E
copolyr
greatly
Poly(rr
Aithou
its sh
4,000
transi
expe1
isom
How.
from
Optic
the

is v.
or i
W
Ch:
de
dif
tht
Ti

5.3 Thermal properties of the copolymer.

DSC was used to measure the glass transition temperatures of the
copolymers. The results are shown in Figure 57 and Table 18. Bulky side chains
greatly increase the glass transition temperature of polymers.
Poly(methylphenylglycolide) has a glass transition temperature at 85 °C.
Although, poly(trimethylglycolide) shows a glass transition temperature at 50 °C,
its should be higher, since the molecular weight of polymer was low (about
4,000). Polymers with molecular weights over 20,000 should have higher glass
transition temperatures. None of the polymers were crystalline, which was
expected, since the ethylmethylglycolide was a mixture of 8,8, R,R, and R,S
isomers. Trimethylglycolide and methylphenylglycolide were also racemic.
However, we made (S,S)-methylphenylglycolide and (R,R)—methylphenylglycolide
from L-mandelic acid and D-mandelic acid, and the copolymers from these
optically pure AB monomers are still amorphous. There are two possibilities for
the lack of crystallinity in these copolymers. First, the proton or to the phenyl ring
is very reactive, and at high temperature, this proton may be lost to form a radical
or ion. In either case, the polymers are going to be racemized, which inhibits
crystallization. During bulk polymerization at 180 °C, the reaction mixture
changed from colorless to dark brown very quickly. We think it was caused by the
deprotonation. Second, because of transesterification and because the reactivity
difference for attack at the hindered and less hindered side is not large enough,
there are many head-to-head and head to tail placements in the polymer chain.

Thus, the polymer chain is not ordered enough to crystallize.

179

The
Polyttrimeti;
the low m
about 6% l
the impurit

thermally é

Poly

 

 

 

 

Poly

The decomposition temperatures of the polymer are shown in Figure 58.
Poly(trimethylglycolide) has the lowest the decomposition temperature, due to
the low molecular weight of the polymer. Poly(methylphenylglycolide) shows
about 6% residue at 370 °C, which then disappears at 600 °C. It is not caused by
the impurity such as catalyst, it is well-known that aromatic polymers often form

thermally stable compounds when thermally decomposed.

Table 18. The properties of AB polymers.

 

 

 

 

 

 

 

 

 

 

Polymer M..x10'3 a Min/Mn T, °C ”
Poly(ethylmethylglycolide) 31 .2 1 .81 85
Poly(trimethylglycolide) 35.6 1 .85 50
i Poly(methylphenylglycolide) 4.2 1.25 22

 

a. measured by GPC in THF using polystyrene as standard
b. measured by DSC under He at rate of 10 °Clmin

180

 

methyl phenyl glycolide

trimethylglycolide

J

ethyl methyl glycolide

T I T

-10 30 70 110
temperature ( °C)

’uhH’

endo —-.

 

 

 

Figure 57. DSC runs (second heating after flash quenching from 100 °C) for

A8 polymers. Heating rate: 10 °Clmin under helium.

181

 

100 - -— Poly(methylphenylglycolide)

\

_Poly(ethylmethylglycolide)

(D
O
t

O)
O

A
O

- Poly(trimethylglycolide)

weight percentage (%)

N
O
‘r

 

v

 

 

 

O ' “
0 200 400 600

temperature ( °C)

Figure 58. Thermogravimetric analysis results for AB polymers run in air.

Heatina rate: 40 °Clmin.

182

6.1 Syr

I
polyme
polyme
crystal
availal
amino
optica
Synth
with t
This
meth
litera
fOUn
88p;

qua-

Was
last
180'
rag

She

6. Crystalline substituted polylactide.

6.1 Synthesis and polymerization of optical pure monomers

All of the substituted polylactides mentioned above are amorphous
polymers because the monomers used to make the polymers are racemic. If we
polymerize optically pure monomers, the resulting polymers should show some
crystallinity. Since most optically pure 2-hydroxyacids are either not readily
available or are very expensive, we prepared optically pure 2-hydroxyacids from
amino acids. Optically pure amino acids are easily available and cheap, and the
optically pure monomers can be obtained through simple chemical reactions. The

synthesis of a monomer from valine is shown in Scheme 44. Valine was treated

with NaNOz in acidic aqueous solution to yield the corresponding 2-hydroxyacid.

This reaction proceeds with retention, and thus L-valine yields L-2-hydroxy-3-
methylbutyric acid. The synthetic procedure used is a modified version of a
literature preparation. The literature preparation uses HCI as the acid, but we
found that about 10% of the product is the 2-chloroacid, which is very difficult to
separate from the 2—hydroxyacid. Switching to dilute H2804 gave nearly
quantitative yields of the 2-hydroxyacid.

The dimer of 2-hydroxy-3-methylbutyric acid is isopropylglycolide, which
was synthesized using the same method used to synthesize other substituted
lactides such as ethylglycolide. Because we wanted to obtain optically pure
isopropylglycolide, the high temperature cracking process used in making
racemic monomers is inappropriate since it leads to some epimerization of the

stereocenter. 2-Hydroxy-3-methylbutyric acid was condensed in toluene using p-

183

toluenesulfonic acid as catalyst to give a mixture of the dimer and low molecular
weight oligomers. The resulting glycolide was crystallized directly from toluene to
give pure S-isopropylglycolide. rac—2-Hydroxy-3-methylbutyric acid gave a
mixture of R,R, 8,8 and R,S isomers, and after recyrstallization from ether, only

pure rac-isopropylglycolide (1 :1 mixture of R,R and 8,8 isomers) was obtained.

Melt polymerizations carried out at 180 °C using Sn(Oct)2/BBA as the
catalyst/initiator reached 95% conversion after about one hour. At the beginning
of the polymerization, the mixture was an easy to stir liquid, but after several
minutes, the mixture became very viscous. After 10 minutes, the polymerization
mixture solidified and was impossible to stir. The solidification was caused by
crystallization of the polymer, which has a melting point >180 °C. The details of
crystallization will be discussed later. The molecular weight distribution of the
resulting polymer is very broad, with a polydispersity about 3, probably due to
crystallization of polymer. Because of crystallization, the polymerization reaction
can become diffusion controlled. If the diffusion rate of monomer is much slower
than the polymerization rate, monomer is not evenly distributed in the mixture.
Thus, polymerization sites have different propagation rates leading a mixture of

long and short polymer chains and a broad molecular weight distribution.

NMR spectra of poly(S-isopropylglycolide) are shown in Figure 59. The
top trace shows the methine proton region for the homo decoupled 1H NMR
spectrum, and the bottom trace shows the carbonyl region of the 13C NMR
spectrum. We conclude that there is almost no racemization during the

polymerization, because only one peak is seen in both spectra.

184

Scheme 44. The synthesis of an optically pure substituted Iactide from an

amino acid precursor

 

185

 

 

 

 

 

 

TT'UIITTTTTTITWTTTTVTII'YUTl'UI'T'lrl’Ill'TrUllU

. . I . . . .
5.2 5-1 5.0 4.9 4.8 ppm

 

 

 

 

 

 

VIVWTTYI‘FTYV [ I V

H... ......, “1-... ....,...f,
168.9 168.7 168.5 168.3 168.1 ppm

 

Figure 59. NMR spectra of the methine (top) and carbonyl (bottom) region

of poly(S-isopropylglycolide)

186

 

6.2 The stereochemistry of the polymer chain

A number of the physical properties of poly(lactide) are linked to its
stereosequence distribution in the polymer chain. For example, pure isotactic
poly(L-Iactide) crystallizes at faster rate and to a larger extent than when L-
lactide is polymerized with small amount of either D—lactide or meso—lactide."90
in the 1H and 13C NMR spectra of polylactides, the observed resonances can be
assigned to various stereosequence combinations in the polymer. The
assignments are designated as various combination of “i” isotactic pair-wise
relationships (-RR- and -SS-) and “s” syndiotactic pair relationship (-RS- and -

SR-). In the NMR spectra, the -RR- and -SS- diads and -RS- and -SR- diads are

indistinguishable and have identical chemical shifts.

The assignment of resonances for polylactides up to hexads has been
made by several research groups.137:191'194 Bernoulli and first-order Markov
models have been used to rationalize the assignments. In Figure 60 are shown
the building of a polymer chain by a Bernoulli and a first-order Markov process. In
the Bernoulli model, the stereochemistry of chain end is not important, which

means that stereochemistry of the chain end does not influence the addition of
the monomer to the chain end. In the first—order Markov model, the

stereochemistry, which may be i or s is strongly influenced on the adding

monomers to the chain end.

187

 

 

 

 

My.

' ‘-

 

i.0 /d|>
6' \
.

 

 

 

First-order Markov model

Figure 60. Schematic representation of the Bernoulli and first-order

Markov model.

188

The physical properties of poly(isopropylgycolide) are also dependent on
the stereosequence of the polymer chain. As shown in Figure 61, pure poly(S-
isopropyglycolide) has a melting point around 230 °C. The melting point of
poly(isopropyglycolide) decreases when S-isopropylglycolide is polymerized with
a small amount of rec-isopropyglycolide. From measurements of the heat of
fusion, we found that the extent of crystallization also decreased. As was done
with polylactide, we would like to assign the stereosquences of
poly(isopropyglycolide) to understand the relationship between the physical

properties of poly(isopropyglycolide) and stereochemistry of the polymer chain.

The stereochemical assignments were made by comparing the trends
observed in the spectra of a number of poly(isopropylglycolide)s to the probability
distribution expected on the basis of their Iactide feed composition. Shown in
Figure 62 and Figure 63 are the carbonyl region of the 13C NMR of
poly(isopropyglycolide) and the methine region of 1H NMR of
poly(isoproplyglycolide). Trace A in each figure is spectrum of the polymer
prepared from 100% S-isopropyglycolide. The single peak in both spectra can be
identified as resulting from an iiiii sequence. The B traces are spectra for the
polymers prepared from 85% R-isopropylgycolide and 15% racemic
isopropyglycolide. In additon to the iiiii peak, several new peaks appeared.
Because the amount of S-isopropyglycolide is small, probability dictates that
these peaks should be isiii and iiisi sequences. By comparing the spectra from
samples with different ratios of S-isopropylglycolide and rac-isoproylglycolide, we

assigned all 11 hexads of the polymer chain. The result is shown in Figure 64.

189

 

Because the resolution of 1H NMR spectrum is not as good as for the 13C NMR,

we only assigned stereosequences in the 130 NMR spectrum.

The unusual intensitry of the iiiii hexad in the poly(rac-isopropyglycolide)
spectrum suggests that the polymerization follows non-Bernoullian statistics and
stereoselection occurs during the polymerization of rec-isopropyglycolide. We
used the first-order Markov model to describe such a process. We assume that
the probabilities for an RR monomer adding to an RR chain end and an SS
monomer adding to an SS chain end are p, and the probabilities of an RR
monomer adding to an SS chain end and an SS monomer adding to an RR chain
end are (1-p). It is possible to calculate the expected intensity values of the

individual sequences as shown below:

iiiii = p3 + 0.5p2(1-p)

iiiis = siiii = iisii = o. 5p2(1-p )

iiisi = isiii = 0.5p2(1-p) + 0.5p(1-p)2
iisis = siiis = sisii = o. 5p(1-p)2

isisi = 0.5p(1-p)2 + 0.5(1-p)3

sisis = 0.5(1-p)3

We compared the experimental values and calculated values using
equation 2[(lca.-lexp)/lexp]2, where lea. is calculated peak intensity and lexp is
experimental peak intensity. We found that when p=0.61, the lea. is closest to law.

The result is shown in Table 14. It indicates a preference for addition RR

190

monomer to the RR chain end and SS monomer addition to SS chain end, which
enhances contribution of isotactic segments in poly(isopropylglycolide). This
result is very different from that obtained for polylactide, where several authors
reported that the Iactide polymerization favors syndiotactic chain segments

(p<0.5). The reason for this difference is not clear and is still under investigation.

 

Table 19. Experimental and calculated values of hexad intensities in the

 

carbonyl region of 13C NMR spectra of poly(rac-isopropygcholide)

 

 

 

 

 

 

 

 

 

 

 

 

iiiii+siii iiisi sisii isiii-I-iisii iisis+sisis isisl
i+iiiis +siiis
Experimental 46.2 11.8 4.4 22.3 7.07 8.25
values
Calculated
values 44.4 11.9 4.6 23.8 7.61 7.61
(p=0.61)
191

 

 

100:0

 

93:7

endo ——>

85:15

 

 

 

100 150 200 250
temperature (°C)

Figure 61. DSC runs showing the melting point for poly(isopropyglycolide)s
prepared from S and rac-isoproplyglycolide. The ratios are the Szrac content in

the polymers.

192

 

 

t.’
C O
C rac-isopropylglycolide H >§\
n

 

B 85% S-isopropyglycollide
15% rac-isopropyglyoolide

 

A 100% S-isopropyglycolide

A

 

 

168.9 168.7 168.5 168.3 168.1 ppm

 

 

Figure 62. 13C NMR of poly(isopropylglycolide)

193

 

 

O
C rac-is0propylglycolide {L
H n

z-.- ,4,-_-_._J”\/x_ ,,- ._

ht r __..fi

8 85% S-isopropylglyoolide
15% rac-isopropylglycolide

 

 

 

A 100% S-isopropyglycolide

A L

 

 

 

 

 

lelllTTlTlTTITTTTITITTIFTTTITTTTITTIIITIIIIITTl—I

5.20 5.15 5.10 5.05 5.00 4.95 4.90 4.85 4.80 ppm

 

 

Figure 63. ‘H NMR of poly(isopropylglycolide)

194

_ L

$2_oo>_m_>ao.aom_.oflv>_oq Co 828on on. .8 “cascgmmm 85:389on .3 23m.“—

EQQ m.me odoa h.mmfi m.me m.mmH 9me .7me N.moH m.mmH v.mmH m.mmH odofl ~2me m.mwa «3de

LhP-Frbphpp—Pblhh—i—i—pp—--h—PPPL.—bL~P_F-PFF-uph—phhp—thh—rhb-p—PF>»~»FL——bpthpph

 

 

 

EocoQEoo .m:u_>__uc_

ME»

.5wa 9w:

 

 

528on 8.936%

830on .mBom

 

195

6.3 The crystallinity of poly(isopropyglycolide)

We used X-ray diffraction to determine the crystallinity of poly(R-

isopropyglycolide). This method has been used by Herrnans and Weidinger to
measure the crystallinity of polyethylene,195 polypropylene,196 and

polystyrene.197 The X-ray diffraction pattern of crystalline
poly(isopropylglycolide) is shown in Figure 65. A valid procedure for subtracting
the amorphous scattering from the total scattering is always the first and most
essential step in any attempt to derive a measure of crystallinity from X-ray
diffraction. However, unavoidably, it always involves certain uncontrollable
assumptions. The X-ray diffraction pattern for amorphous poly(isopropyglycolide)
made from rac-isopropyglycolide is shown in Figure 66. The amorphous sample
shows a broad peak with maximum at about 12 degrees. We assumed that the
amorphous fraction in the crystalline sample should have the same maximum
and the same line-shape. To make the problem easier, we selected a 10-degree
window, 6-16 degrees, for the calculation. PeakSolve software was used to
deconvolute the diffraction pattern into amorphous and crystalline components.
The result is shown in Figure 67. The area under amorphous peak is Os.m and

the area under three crystalline peaks is O”.

The degree of crystallinity is calculated by comparing the results from two
or more samples of the same polymer with crystalline fraction X1 and X2, where
X1-X2 is as large as possible. Then, X1 and X2 can be described using following

equafion:

X1/X2=Ocr1/Ocr2

196

(1‘X1)/(1'X2)=Oam1/Oam2

The crystallinity of several different poly(isopropylglycolide) samples are
shown in Table 20. We combined the crystallinity data and the heat of fusion
measured using DSC to estimate the heat of fusion for a 100% crystalline
sample. By plotting the heat of fusion vs. crystallinity of sample, we use data from
Figure 68 to estimate the heat of fusion for 100% crystallinity. The result,

43.6.J/g, is about half that of polylactide’s 100 J/g.

Table 20. The crystallinity and heat of fusion of poly(isopropylglycolide)

 

 

 

 

Samples Crystallinity ' (%) Heat of fusion (AH Jig)
1 46.9 22.1
2 35.2 17.9
3 27.8 14.3

 

 

 

 

 

a. estimated from X-ray analysis of samples

197

4500

 

4000 -

3500 -

3000 -

2500 -

 

2000 -

1500 -

intensity/second

1000 -

500 -

 

 

0 l l I I
0 10 20 3O 40 50

26

 

Figure 65. X-ray diffraction pattern for crystalline poly(S-isopropylglycolide)

198

 

350

300 -

250 -

200 -

150 F

intensitry/second

100 -

50-

 

 

o l l l
0 1 0 20 30 40

26

Figure 66. X-ray diffraction for amorphous poly(rac-isopropylglycolide)

199

 

 

4000 h

(D
O
O
O
I

 

2000

’

amorphous peak .

Intensity/second

1000

 

 

26

Figure 67. Peak deconvolution for crystalline poly(S-isopropylglycolide)

200

 

 

 

 

25
AH1oo = 43.6 Jlg

3
3 20 -
r:
.2
in I
.2
h
o
H
a
0 15 r
I

10 l 1

20 30 40 50
crystallinity (%)

Figure 68. Determination of heat of fusion for 100% crystalline

poly(isopropylglycolide)

201

7. The degradation of substituted polylactides

The degradability of the substituted polylactides was evaluated by
hydrolytic degradation in phosphate buffer solution (0.01M KHzPOJKzHPO4
solution, pH=7.4) at 55 °C. To exclude the influence of autocatalysis. we
chopped the polymer samples into small pieces. The auotcatalytic effect is
caused by the increase in acidity when the degradation products are acidic and
have limited diffusion. If polymer samples are small enough, the diffusion rate of
buffer solution into the bulk is higher than the polymer chain scission and the
hydrolytic degradation can proceed uniformly in the bulk of the specimen. The
morphology of the polymer also influences the degradation of polymer. The
crystalline polymer degrades much slower than the amorphous polymer. The
much tighter packing of the crystalline phase makes water much harder to
penetrate into the bulk of the polymer. To study the influence of chemical
structure of polymer on degradability, we have to eliminate the influence of
morphology. To do that, we used amorphous substituted polylactides prepared
from racemic monomers. All polymers were tested using DSC to ensure that the

polymers were amorphous.

The weight loss and molecular weight decrease with degradation time are
two of most important properties of biodegradable polymers. The weight and
molecular weight loss versus time are shown in Figure 69 and Figure 70,
respectively. The sample weights were constant for a period of time, and then
began to drop. The molecular weight of polymer dropped immediately once the

degradation begins. These are typical behavior for hydrolytic degradation.

202

To calculate the degradation rate constant, we used the random scission
model to fit the data. For the random scission model, the average number of

bond cleavages per polymer molecule (N) was calculated according to:

N=[Mn(0)/Mo(t)-1]= Kail"n(0)t

where Mn(0) and mm are the number-average molecular weights of polyesters
at time 0 and time t. kd is the rate constant for hydrolytic degradation, and P..(0) is

the number-average of degree of polymerization at time 0.

K, was obtained from plots of [Mn(0)/Mo(t)-1]/P,,(0) versus t. The slopes of
the plots are the rate constants for hydrolytic degradation. The results are shown
in Figure 71. From the figure, we can see that polylactide degrades fastest,
poly(ethylglycolide) and poly(hexylglycolide) have almost the same degradation
rate and slower than poly(lactide), and poly(isobutylglycolide) degrades the

slowest.

An interesting phenomenon showed up when remaining weight fraction
was plotted vs. degree of polymerization, as shown in Figure 72. The weights of
all polymers began to drop at almost the same degree of polymerization, which is
about 30. We think this is critical molecular weight for entanglement. When the
degree of polymerization is higher than 30, almost all of the polymer chains are
entangled, so polymers will not dissipate into the water and there is almost no

weight loss for the polymers. When the degree of polymerization of polymer is

203

lower than 30, some of polymer chains are not entangled, so these chains will

dissipate into water and there is weight loss for the polymers.

How the substituent influences the degradation rate of substituted
polylactides is the question we need to answer. We think there are two factors
that influence the degradation rate: the glass transition temperature and the
surface energy of the polymers. At the degradation temperature, the free volume
will be larger in polymers with a relatively low glass transition temperature than
those with high glass transition temperatures. Thus, low T9 polymers will have a
higher water content, which causes the higher degradation rate. If a polymer has
a very low surface energy (hydrophobic), the solubility of water will be low and

the degradation rate will be low.

The glass transition temperatures of the substituted lactides are shown in
Table 21. The order of glass transition temperature is Iactide > isobutylglycolide
> ethylglycolide > hexylglycolide. Based on glass transition temperature, we
would expect the order of the degradation rate is hexylglycolide > ethylglycolide >
isobutylglycolide > Iactide. However, besides the glass transition temperature, we

also need to consider the surface energy of the polymers.

We used contact angle method measurements to characterize the
surface energy of the substituted polylactides. We make a series of
water/methanol solutions with different water/methanol ratios. The surface
tensions of these solution were measured, and then these solution were dropped

on the polymer films. The contact angle between the solution and polymer film

204

were measured, and the surface energy of the film were calculate using the

Zisman equation:

0089 = 1 + marc)

where 9 is contact angle, 71. is surface tension of testing solution (water/methanol
solution). yo is the surface energy of the polymer surface, and B is constant.
Values of cost) vs yL were plotted for each substituted polylactide shown in

Figure 73, and then the lines were extend to cost) =1. The intercept 7., value is

the surface energy of polymer (Table 21).

Table 21 . The degradation of substituted polylactide

 

 

 

 

 

polymer Glass transition Surface energy Degradation
temperature(°C) y (dynlcm) rate (day'1)
x103
polylactide 65 21 .6 3.7
Poly(ethylglycolide) 1 5 19.9 2.6
Poly(isobutylglycolide) 23 18.5 2.4
Poly(hexylglycolide) -37 16.2 1 .0

 

 

 

 

 

 

205

The length of linear alkyl substituted group increases from Iactide to
ethylglycolide to hexylglycolide, which causes the glass transition temperature
and surface energy to decrease. However, these two factors have opposite
effects. Decreasing glass transition temperature increases the degradation rate,
and decreasing the surface energy decrease the degradation rate. So a
combination these two factors may explain the observed order of degradation
For poly(isobutylglycolide), the branched the side chain gives the polymer a
relatively high glass transition temperature and low surface enerQY; these two
factors work in concert to slow the degradation rate. It is not surprising that

poly(isobutylglycolide) has slowest the degradation rate.

206

 

 

 

 

1 1 . o O O
A . ' O I I O
t s
c \a
.3 0.8 . \\
0
t!
5
IL
:5, 0.6 -
0
3
g) A“
N
8
, o
9‘ 0.2 - e\
0 1 ‘ 1
0 20 40 60
Time (days)

Figure 69. The weight loss of substituted glycolide during hydrolytic
degradation. (A) Iactide, (I) ethylglycolide, (O) hexylglycolide, (O)

isobutylglycolide.

207

 

 

 

 

 

7. 1501;01:286me T‘
‘ [Poly(etl'iylglycolide) ;
0.8 - oPoly(isobutylglycolide) ‘
o Poly(hexylglycolide) l
a 0.6 -
‘E’
2
2. O
t;
c
2 0.4
A
0.2 ~
0
\‘ ‘u
o 1 ME
0 10 20 30 40
Time (days)

Figure 70. The molecular weight loss of substituted glycolide during
hydrolytic degradation. (A) Iactide, (I) ethylglycolide, (O) hexylglycolide, (O)

isobutylglycolide.

208

 

 

0.1

 

 

 

I

0.08 -
8
‘c’
a /‘
c 0.06 - ‘ I
,L
a:
E: e
a 004 ’ /
7; I
a A

0.02 - ‘

0 l L l
0 10 20 30 40
Time (days)

Figure 71. The molecular weight decrease of substituted lactides during
hydrolytic degradation fitted to the random chain scission model. (A) Iactide,

(I) ethylglycolide, (O) hexylglycolide, (O) isobutylglycolide

209

 

"=30
A :0 I ‘0”-

.0
oo

 

F’ 5:
-b O)

Remaining Weight Fraction
(3
iv

 

 

 

 

1000 100 10 1
Pn

Figure 72. The decrease in the degree of polymerization of substituted Iactides
during hydrolytic degradation. (A) Iactide, (II) ethylglycolide, (I) hexylglycolide,
(O) isobutylglycolide

210

60

 

40-

   

10-

 

 

0 I l I l
0 0.2 0.4 0.6 0.8 1

cos(6)

 

Figure 73. Contact angles on substituted polylactides plotted as 7 (surface
tension of water/methanol solutions) vs cos 9. (A) Iactide, (I) ethylglycolide, (O)

hexylglycolide, (O) isobutylglycolide

211

EXPERIMENTAL

1. General

Unless otherwise specified, ACS reagent grade starting materials and
solvents were used as received from commercial suppliers without further
purification. Proton nuclear magnetic resonance (‘H NMR) and carbon nuclear
magnetic resonance (13C NMR) analyses were carried out at room temperature in
deuterated chloroform (CDCI3) on a Varian Gemini-300 and Varian Vax-SOO
spectrometers with the solvent proton signals being used as chemical shift
standards. Mass spectral analyses were carried out on a VG Trio-1 Benchtop

GC-MS.

The molecular weights of polymers were determined by gel permeation
chromatography (GPC) using a PLgel 20m Mixed A column and a Waters R401
Differential Refractometer detector at room temperature. THF was used as the
eluting solvent at a flow rate 1 mL/min, and monodisperse polystyrene standards
were used to calibrate the molecular weights. The concentration of the polymer
solutions used for GPC measurements was 1 mg/mL. Differential scanning
calorimetry (DSC) analyses of the polymers were obtained using a Perkin Elmer
DSC 7. Samples were run under a helium atmosphere at a heating rate of 10
°Clmin, with the temperature calibrated with an indium standard.
Thermogravimetric analyses (T GA) were run in both air and under nitrogen at a

heating rate of 10 °Clmin using a Perkin Elmer TGA 7. Measurements of the

212

mechanical properties of polymer samples were made using a Perkin Elmer DMA

7 Dynamic Mechanical Analyzer at a heating rate of 10 °Clmin.

X-ray Powder Diffraction (XRD) patterns were recorded on a Rigaku
rotaflex 2008 diffractometer equipped with a rotating anode, Cu Kgx-ray radiation
(k: 1.541838A) and a curved crystal graphite monochromator. The x-ray was
operated at 45 KV and 100 mA. Diffraction patterns were collected at 001°
intervals between 5 and 45° values of 20 at a scanning rate of 01° per minute
and DS and SS slit widths of 1/2. Powder samples were prepared by spreading
solid samples on the window of the glass sample holder with a spatula, or by
taping polymer film made from solvent casting on to the window of the sample

holder.

2. Synthesis

Symmetric Monomer Synthesis, general procedure. (Route 1 in
Scheme 3) A mixture of 10 g of the appropriate a-hydroxy acid and 0.2 g of p-
toluenesulfonic acid in 700 mL of toluene was heated at reflux for four days, with
the water removed azeotropically using a Barrett trap. The toluene Solution was
then cooled, washed with sat. NaHCOa, and dried over M9804. After removing

the toluene, about 0.1 g of ZnO was added and the residue was distilled under

reduced pressure using a Kugelrohr distillation apparatus.19

3,6-Diethyl-1,4-dioxane-2,5-dione (Ethylglycolide) (1). Ethylglycolide

was collected at 180 °C (100 mtorr), and was dissolved in the minimum amount

213

of ether needed to dissolve the product. The solution was cooled to —30 °C, and
petroleum ether was added drop-wise until the solution turned cloudy. The
colorless crystals were collected by cold filtration and dried under vacuum to give
5.2 g (63%) of ethylglycolide as a colorless oil. 1H NMR indicates that the
product is a statistical mixture of diastereomers. 1H NMR (300 MHz, CDCI3): 6
4.88 (dd), 4.83 (dd, 1H total for the signals at 4.88 and 4.83), 2.08 (m, 2H), 1.15

(tt, 3H). 130 NMR (75 MHz, cocna) 5: 167.62, 166.58, 76.4, 75.6. mp 19.5 - 20.5
°c Iit.179 21-22 °c; MS (EI) m/z = 173.4 (M+1). Preparation via or-bromoacyl

bromides.180 Under a nitrogen atmosphere, 0.90 g (8.7 mmol) of 2-
hydroxybutyric acid and 2.1 g (8.7 mmol) of 2-bromobutyryl bromide were heated
at 80 °C until HBr evolution ceased (0.5-2 h). The solution was cooled and 100
mL of dry acetone were added followed by the drop-wise addition of 2.2 mL (17
mmol) of triethylamine. The solution was heated to reflux for 3 h, and then
cooled to room temperature. After removing the acetone under reduced
pressure, the residue was washed with sat. NaHCOa, extracted with ether, and
the ether layer was dried over M9804. After removing the ether, the dimer was
purified by recrystallization as described above to give 0.62 g (41%) of
ethylglycolide as a colorless oil. The product is approximately a 4:1 mixture of

the R,R/8,8 and R,S diastereomers.

3,6-Diisobutyl-1,4-dioxane-2,5-dione (isobutylgycolide) (2).
Isobutylglycolide was collected at 120 °C (50 mtorr), and was recrystallized from

ether. The white crystals were collected by filtration and dried under vacuum to

give 6.1g (71%) of isobutylglycolide as white solid. 1H NMR indicates that the

214

product is a statistical mixture of the R,R/8,8 and R,S diastereomers. 1H NMR
(300 MHz, CDCI3): 5 4.91 (dd), 4.89 (dd), (1H total for the signals at 4.91 and
4.89), 1.8-2.0 (m, 3H), 0.95 (m, 6H). 13C NMR (75 MHz, CDCI3) 8: 167.28,

166.20, 74.88, 74.09, 40.37, 38.82, 24.07, 23.84, 23.02, 22.84, 21.29, 21.25. MS

(mlq) 229.4 (M+1); mp 167-167.5 °c ("L198 169-170 °C).

3,6-Dihexyl-1,4-dioxane-2,5-dione (hexylgycolide) (3). Hexylglycolide
was collected at 145 °C (50 mtorr), and was recrystallized from ether. The white
crystals were collected by filtration and dried under vacuum to give 5.8 g (65%)
of hexylglycolide as a white solid. 1H NMR indicates that the product is a
statistical mixture of the R,R/S,S and R,S diastereomers. 1H NMR (300 MHz,
CDCI3): 8 4.88 (dd), 4.83 (dd), (1 H total for the signals at 4.88 and 4.83), 2.01 (m,
2H), 1.4-1.6 (br m, 2H), 1.2-1.4 (br m, 6H), 0.85 (t, 3H). 13C NMR (75 MHz,
CDCI3) 6: 166.95, 165.85, 76.42, 75.65, 31.99, 31.46, 31.41, 30.15, 28.74, 28.56:
MS (m/z) 285.4 (M+1), mp 78-80 °C .

3,6-DiisopropyI-1,4-dioxane-2,5-dione (isopropylgycolide) (4).
lsopropylglycolide was synthesized from DL-2-hydroxy-3-methyl butyric The
crude product was recrystallized directly from toulene after removing of toluene.
The white crystal was collected by filtration and dried under vacuum. Yield 47%.
1H NMR (300 MHz, CDCI3): 6 4.95 (d, 1H), 2.30 (m, 1H), 1.03 (d, 3H), 1.01 (d,

3H). 13C NMR (75 MHz, CDCI3) 5 166.38, 79.55, 29.36, 19.55, 15.81. MS (m/z)

201.4 (M+1), mp 137-138 °c [iit.199 136 °C].

215

 

3S,6$-DiisopropyI-1,4-dioxane-2,5-dione (S-isopropylgycolide) (5). S-
isopropylgycolide was synthesized from L-2-hydroxy-3-methyl butyric The S-
isopropylgycolide was recrystallized from toluene. The white crystals were
collected by filtration and dried under vacuum. Yield 43%, [alzo = -264.0 (c=1,
THF). 1H NMR (300 MHz, CDCI3): 84.95 (d, 1H), 2.30 (m, 1H), 1.03 (d, 3H), 1.01
(d, 3H). 13C NMR (75 MHz, CDCI3) 8: 166.38, 79.55, 29.36, 19.55, 15.81. MS
(m/z) 201.4 (M+1), mp 140-141 °C. The S-isopropylglycolide is 99% ee was
determined by hydrolyzing S-isopropylglycolide to 2-hydroxy-3-methyl butyric
acid and comparing the optical rotation value the acid with known value

([a]20=+19.0 c=1 CHCI3) from Aldrich catalog.

2-hydroxy-3-methylbutyric acid. (6) A solution of 60 mL concentrated
H2SO4 in 1L r water was cooled in an ice-bath. To the cooled solution was added
60 g of valine follwed by the dropwise of 144 g NaNOz dissolved in 1L water.
After the solution was stirred at 0 °C for overnight, the solution was extracted with
ether (6 x 300 mL), and the ether layer was dried over M9304. After removing
the ether, the product was purified by recrystallized from toluene. The white
crystals were collected by filtration and dried under vacuum to give 45 g (75%) of
2-hydroxy-3-methyl butyric acid as a white solid. DL-2-hydroxy-3-methyl butyric
acid was obtained from DL-valine and L-2-hydroxy-3-methyl butyric acid was
prepared form L-valine. The L-2-hydroxy—3-methyl butyric acid was obtained in

99% ee based on measurement of its optical rotation ([or]20=+18.9 c=1 CHC'a).

216

Unsymmetric Monomer synthesis, general procedure. Under a
nitrogen atmosphere, one equivalent of a 2-hydroxy acid and one equivalent of a
2-bromoacyl bromide were heated at 80 °C until HBr evolution ceased (0.5-2 h.).
The solution was cooled and 200 mL of dry acetone was added for each gram of
acid, followed by the dropwise addition of two equivalent of triethylamine. The
solution was heated to reflux for 6 h., and then was cooled. After removing the
acetone, ethyl acetate was added to dissolve the residue. The solution was
washed with 2N HCI, water, then washed with sat. NaHCOa, and the organic
layer was dried over MgSO4. After removing solvent, the monomers was purified

by recrystallization and distillation.

3-ethyl-6-methyI-2,5-dloxane-1,4-dione (7) was synthesized from 2-
hydroxybutyric acid and 2-bromopropionyl bromide. It was purified by distillation
(45 °C/50 mtorr) to give a colorless oil. Yield: 53%. 1H NMR (CDCla) 8 5.03 (q),
4.99 (q, 1H for signal at 5.03 and 4.99 ppm), 4.86 (dd), 4.83 (dd, 1H for signal at
4.86 and 4.83), 1.90-2.20 (m, 2H), 1.69 (d), 1.66 (d, 3H for signals at 1.69 and
1.66 ppm). 1.15 (t). 1.14 (t, 3H for signals at 1.15 and 1.14ppm). 13C NMR
(CDCI3) 8 167.6, 166.8, 166.3, 165.6, 77.7, 76.6, 72.5, 72.2, 25.3, 23.3, 17.5,
15.7, 9.1, 8.7. MS (m/q) 159.4 (M+1), The product is approximately a 3:1 mixture

of the R,R/S,S and R,S diastereomers.

3-dimethyl-6-methyI-2,5-dioxiane-1,4-dione (8) was synthesized from 2-
hydroxyisobutyric acid and 2-bromopropionyl bromide. Recrystallization from

ether give white crystal. mp 66-69°C. Yield: 43%. 1H NMR (CDCI3) 8 5.07 (q, 1H),

217

 

1.68 (s, 6H). 1.65 (d, 3H). 130 NMR (00013) 8 168.59, 166.64, 6055,7292,

26.25, 25.31, 17.45 MS (m/q) 159.4 (M+1)

3-methyI-6-phenyl-2,5-dioxane-1,4-dione (9) was synthesized by using
mandelic acid and 2-bromopropyionyl bromide. Recrystallization from toluene
gave white crystals. mp 153-156 °C. Yield: 33%. 1H NMR (CDCI3) 8 7.43 (m, 5H),
5.92 (s, 1H) 5.17 (q, 1H), 1.63 (d, 3H). 130 NMR (00013) 5 166.90, 165.55,

131.24, 129.95, 128.92, 127.44, 77.74, 72.77, 16.42. MS (m/q) 207.4 (M+1),

S-3-methyl-S-6-phenyl-2,5-dloxiane-1,4-dione (10) was synthesized
from S-mandelic acid and 2-bromopropionyl bromide. Recrystallization from
toluene gave white crystals. mp 160-163 °C, yield: 41%. [0020 = +301.0 1H NMR
(CDCI3) 8 7.43 (m, 5H), 5.92 (s, 1H) 5.17 (q, 1H), 1.63 (d, 3H). 13C NMR (CDCla)
8166.90, 165.55, 131.24, 129.95, 128.92, 127.44, 77.74, 72.77, 16.42. MS (m/q)
207.4 (M+1).

R-3-methyl-R-6-phenyl-2,5-dioxiane-1,4-dione (11) was synthesized
from R-mandelic acid and 2-bromopropionyl bromide. Recrystallization from
toluene gave white crystals. mp 160-163 °C, yield: 35%. [alzo = -302.0. 1H NMR
(CDCla) 8 7.43 (m, 5H), 5.92 (s, 1H) 5.17 (q, 1H), 1.63 (d, 3H). 13C NMR (CDCI3)
8166.90, 165.55, 131.24, 129.95, 128.92, 127.44, 77.74, 72.77, 16.42. MS (m/q)

207.4 (M+1).

218

 

3. Bulk polymerization of substituted glycolides.

Solvent-free polymerizations were carried out in sealed tubes prepared
from 3/8 inch diameter glass tubing. A representative polymerization is
described below. In an oxygen and moisture-free dry box, a solution of initiator in
toluene (z 0.01 M) and 0.19 of monomer were added to the tube. The amount of
initiator solution added was determined by the desired monomer/initiator ratio.
For runs using initiators that are insoluble in toluene, the initiator was added
directly to the tube and the walls of the tube were washed with solvent to ensure
that all of the initiator was added to the monomer. The solvent was removed in
vacuum, and the tube was sealed and immersed in an oil bath at 130 °C. At the
end of the polymerization, the tube was cooled, opened, and the polymer was
dissolved in THF. A portion of the sample was evacuated to dryness and
analyzed by NMR for conversion. After removal of the solvent, the polymer was
dissolved in toluene and precipitated into methanol to remove residual initiator.
Typical yields of poly(ethylglycolide) were >85%. For kinetic runs, multiple tubes
were prepared and individual tubes were removed from the heating bath at
predetermined intervals and were cooled in ice, opened, and the contents
dissolved in THF. A portion of the sample was analyzed by GPC for molecular
weight, and the remainder was evacuated to dryness and analyzed by NMR for
conversion. Polymerizations that used alcohols as co-initiators were set up as
described above, except that the appropriate amount of alcohol was added to the

toluene solution of initiator just prior to adding the initiator solution to the tube.

219

For insoluble initiators, tubes were first loaded with monomer and initiator, and

the alcohol co-initiator was directly added to the tube as a toluene solution.

4. Solution polymerization of substituted glycolides.

The reaction flask was charged with 2 mmol of monomer and dried under
vacuum (diffusion pump) at room temperature overnight. Toluene (10 mL ) was
added to the solvent flask through a rubber septum with a syringe, and the
toluene was purified by initiating an anionic polymerization of styrene. The
solvent was transferred under vacuum to the reaction flask, and the initiator
Al(OiPr)3 or Sn(Oct); was added into reaction flask with a syringe through a
rubber septum. The amount of initiator solution added was determined by the
desired monomer/initiator ratio The polymerization was carried out at 70 °C, 90
°C and 100 °C. After the polymerization finished, the reaction was terminated
with 1 mL 2NHC| solution then washed with distilled water until PH=7. The
polymer was precipitated into cold methanol, filtered, and dried under vacuum.
For kinetic studies, small samples were removed at predetermined times using a
syringe through the rubber septum. The samples were analyzed by NMR for

conversion and GPC for molecular weight.

220

5. The surface energy of polymers

Polymer films. The polymers were dissolved in small amount of toluene,
then the solutions were spread on the clean glass slides. The glass slides were

dried in air for 3 days, then dried under vacuum for one day.

Surface tension of solutions. A series of HzolCHaoH solutions were

made. The densities of solutions were measured using hydrometer. The surface

tensions of solutions were measured using pendant drop method200 by FTA 200
contact angle analyzer, which was calibrated using de-ionized water.

Surface tension of polymers. The H20ICH30H solutions were dropped
on the polymer films. The contact angles between the solutions and polymer
films were measured using FTA 200 contact angle analyzer, which was
calibrated using de-ionized water. From change of contact angles with change of
surface tensions of solutions, the surface energies of polymers can be calculated

by Zisman equation:

6089 = 1 1' BWLJYc)

where 9 is contact angle, yL is surface tension of testing solution (water/methanol

solution). yo is the surface energy of the polymer surface, and [3 is constant.

221

6. The degradation of polymers.

About 50 mg of polymer sample was accurately weighed and chopped to
1mm by 1mm pieces. The chopped sample was immersed in to 15 mL 0.1M
phosphate buffer solution (pH=7.4). The degradation was carried out at 55 i 0.1
°C. After a predetermined period of time, the sample was taken out, washed
thoroughly with distill water and dried under vacuum. The sample was weighed to

determine the weight loss and molecular weight loss was determined by GPC.

222

BIBLIOGRAPHY

223

10.

11.

12.

13.

14.
15.

BIBLIOGRAPHY

Zhang, X. C.; Wyss, U. P.; Pichora, D.; Goosen, M. F. A.; J. Macromol. Sci.-
Pure Appl. Chem. 1993, A30, 933-947.

Han, D. K.; Hubbell, J. A.; Macromolecules 1997, 30, 6077-6083.

Kricheldorf, H. R.; KreiserSaunders, l.; Macromol. Symp. 1996, 103, 85- It:
102. '

Chen, X. H.; McCarthy, S. P.; Gross, R. A.; Macromolecules 1997, 30,
4295-4301 . *

 

Nobes, G. A. R.; Marchessault, R. H.; Maysinger, D.; Dmg Deliv. 1998, 5,
167-177.

Fukuzaki, H.; Yoshida, M.; Asano, M.; Kumakura, M.; Biomaten'als 1990,
11 , 441446.

Breitenbach, A.; Kissel, T.; Polymer 1998, 39, 3261-3271.
McCoy, M.; C&EN 1998, 76, 71-72.
Sinclair, R. G.; J. M. S. - Pure Appl. Chem. 1996, A33, 585-597.

Meinander, K.; Niemi, M.; Hakola, J. S.; Selin, J. F.; Macromol. Symp.
1997, 123, 147-153.

Grijpma, D. W.; Nijenhuis, A. J.; Vanwijk, P. G. T.; Pennings, A. J.; Polym.
Bull. 1992, 29, 571 -578.

Spinu, M.; Jackson, C.; Keating, M. Y.; Gardner, K. H.; J. Macromol. Sci.-
Pure Appl. Chem. 1996, A33, 1497-1530.

Vert, M.; Schwarch, G.; Coudane, J.; J. Macromol. Sci.-Pure Appl. Chem.
1995, A32, 787-796.

Bastioli, C.; Macromol. Symp. 1998, 135, 193-204.

Fukuzaki, H.; Yoshida, M.; Asano, M.; Kumakura, M.; Polymer 1990, 31,
432.

224

16.
17.

18.

19.
20.
21.

22.

23.

24.

25.

26.

27.
28.

29.
30.

31.

Buchholz, B.,German patent 443,542,A2 (1991)
Otera, J.; Kawada, K.; Yano, T.; Chem. Lett. 1996, 225-226.

Yamaguchi, A.; Suzuki, K.; Enomoto, K.; Ajioka, M.; Bull. Chem. Soc. Jpn.
1995, 68, 2125-2131.

Deane, D. D.; Hammond, E. G.; J. Dairy Sci. 1960, 43, 1421.
Guber, P. R.; Benson, R. D.; lwen, M. L.,U. S. patent 5,141,023 (1992)

Johns, D. B.; Lenz, R. W.; Luecke, A. Ring-Opening Polymerization; lvin, K.
J. and Saegusa, T., Ed.; Elsevier Applied Science Publishers, 1984; Vol. 1,
pp 461.

Odian, G. Principle of Polymerization; Third ed.; John Wiley & Son:, 1991.

Hall, H. K., Jr.; Brandt, M. K.; Mason, R. W.; J. Am. Chem. Soc. 1958, 80,
6420.

Carothers, W. H.; Borough, G. L.; Van Natta, R. W.; J. Am. Chem. Soc
1932, 54, 761.

Lundberg, R. 0.; Cox, E. F. Ring-Opening Polymerization; Marcel Dekker:
New York, 1969.

Kricheldorf, H. R.; Michael Jonte, J.; Dunsing, R.; J. Macromol. Soc. 1986,
A23(4), 495.

Rozenberg, D.; Makromol. Chem, Macromol. Symp. 1990, 32, 267.

Crivello, J. V.; Lockhart, T., P.; J. polym. Sci, Polym. Chem. Ed. 1983, 21 ,
97.

Cherdron, H.; Korte, F.; Makromol. Chem. 1962, 56, 179.

Penczek, S.; Hofman, A.; Szymanski, R.; Slomkowski, S.; Makromol. Chem.
1984, 185, 655-667.

Hofman, A.; Stomkowski, S.; Penczek, S.; Makromol. Chem. 1987, 188,
2027-2040.

225

32.

33.

35.
36.

37.

38.

39.

40.

41.

42.

43.

44.

45.

Kricheldorf, H. R.; Michael Jonte, J.; Dunsing, R.; Makromol. Chem. 1986,
187, 771-785.

Okamoto, Y.; Makromol. Chem, Macromol. Symp. 1991, 42/43, 117.

Kohn, F. E.; Van Ommen, J. G.; Feuen, J.; European Polymer Journal
1983, 19, 1081-1088.

Kricheldorf, H. R.; Dunsing, R.; Makromol. Chem. 1986, 187, 1611-1625.

Jedlinski, Z.; Kurcok, P.; Kowalczuk, M.; Matuszowicz, A.; Dubois, P.;
Jerome, R.; Kricheldorf, H. R.; Macromolecules 1995, 28, 7276-7280.

Penczek, S.; Duda, A.; Libiszowski, J.; Macromol. Symp. 1998, 128, 241-
254.

Kurcok, P.; Kowalczuk, M.; Hennek, K.; Jedlinski, Z.; Macromolecules
1 992, 25, 201 7-2020.

Yuan, M. L.; Xiong, C. D.; Deng, X. M.; J. Appl. Polym. Sci. 1998, 67, 1273-
1276.

Kurcok, P.; Penczek, J.; Franek, J.; Jedlinski, Z.; Macromolecules 1992, 25,
2285-2289.

Kricheldorf, H. R.; Scharnagl, N.; Jedlinski, 2.; Polymer 1996, 37, 1405-
141 1.

Kricheldorf, H. R.; Kreiser-saunders, l.; Makromol. Chem-Macro. Chem.
Phys. 1990, 191, 1057-1066.

Jedlinski, Z.; Walach, W.; Kurcok, P.; Adamus, G.; Makromol. Chem.-
Macro. Chem. Phys. 1991, 192, 2051-2057.

Emig, N.; Nguyen, H.; Krautscheid, H.; Reau, R.; Cazaux, J. B.; Bertrand,
G.; Organometallics 1998, 17, 3599-3608.

Leborgne, A.; Vincens, V.; Jouglard, M.; Spassky, N.; Makromol. Chem,
Macromol. Symp 1993, 73, 37-46.

226

46.

47.

48.

49.

50.

51.

52.

53.

55.

56.

57.
58.
59.

60.
61.
62.

Kowalski, A.; Duda, A.; Penczek, S.; Macromol. Rapid Commun. 1998, 19,
567-572.

Witzke, D. R.; Narayan, R.; Abstr. Pap. Am. Chem. Soc. 1998, 216, U11-
U12.

Schwach, G.; Coudane, J.; Engel, R.; Vert, M.; Polym. Int. 1998, 46, 177-
182.

Kricheldorf, H. R.; Damrau, D. 0.; Macromol. Chem. Phys. 1997, 198,
1753-1766.

Sodergard, A.; Stolt, M.; Macromol. Symp. 1998, 130, 393-402.

Simic, V.; Girardon, V.; Spassky, N.; Hubert-Pfalzgraf, L. G.; Duda, A.;
Polym. Degrad. Stabil. 1998, 59, 227-229.

McLain, S. J.; Ford, T. M.; Drysdale, N. E.; Abstr. Pap. Am. Chem. Soc.
1992, 204, 188-POLY.

Bero, M.; Kasperczyk, J.; Adamus, G.; Makromol. Chem, Macromol. Symp
1993, 194, 907-912.

Degee, P.; Dubois, P.; Jerome, R.; Macromol. Symp. 1997, 123, 67-84.

Dubois, P.; Jacobs, C.; Jerome, R.; Teyssie, P.; Macromolecules 1991, 24,
2266-2270.

Montaudo, G.; Montaudo, M. S.; Puglisi, C.; Samperi, F.; Spassky, N.;
Le Borgne, A.; Wisniewski, M.; Macromolecules 1996, 29, 6461 -6465.

Duda, A.; Macromolecules 1996, 29, 1399-1406.
Kowalski, A.; Duda, A.; Penczek, S.; Macromolecules 1998, 31, 2114-2122.

Ropson, N.; Dubois, P.; Jerome, R.; Teyssie, P.; Macromolecules 1993, 26,
6378-6385.

Duda, A.; Penczek, S.; Macromolecules 1995, 28, 5981-5992.
Duda, A.; Penczek, S.; Macromol. Rapid Commun. 1995, 16, 67-76.
Dubios, P.; Jerome, R.; Teyssie, P.; Polymer Perprint 1994, 35, 536-537.

227

 

 

67. E

68. '

69.

70.
71

72

63.

64.

65.

66.

67.

68.

69.

70.
71.
72.

73.
74.

75.
76.
77.

78.

Ropson, N.; Dubois, P.; Jerome, R.; Teyssie, P.; Macromolecules 1995, 28,
7589-7598.

Kricheldorf, H. R.; Boettcher, C.; Tonnes, K. U.; Polymer 1992, 33, 2817-
2824.

Degee, P.; Dubois, P.; Jerome, R.; Teyssie, P.; Macromolecules 1992, 25,
4242-4248.

Dubois, P.; Jerome, R.; Teyssie, P.; Abstr. Pap. Am. Chem. Soc. 1990,
199, 16-POLY.

Barakat, l.; Dubois, P.; Jerome, R.; Teyssie, P.; J. Polym. Sci. Pol. Chem.
1993, 31, 505-514.

Tian, D.; Dubois, P.; Jerome, R.; Teyssie, P.; Macromolecules 1994, 27,
4134-4144.

Barakat, |.; Dubois, P.; Jerome, R.; Teyssie, P.; Goethals, E.; J. Polym. Sci.
Pol. Chem. 1994, 32, 2099-3110.

Trofimoff, L.; Aida, T.; lnoue, 8.; Chem. Lett. 1987, 991-994.
Endo, M; Aida, T.; lnoue, S.; Macromolecules 1987, 20, 2982-2988.

lsoda, M.; Sugimoto, H.; Aida, T.; lnoue, S.; Macromolecules 1997, 30, 57-
62.

Sugimoto, H.; Aida, T.; lnoue, S.; Macromolecules 1990, 23, 2869-2875.

Sugimoto, H.; Kawamura, C.; Kuroki, M.; Aida, T.; lnoue, S.;
Macromolecules 1994, 27, 2013-2028.

sugimoto, H.; Aida, T.; lnoue, S.; Macromolecules 1993, 26, 4751-4755.
Mogstad, A. L.; Waymouth, R. M.; Macromolecules 1994, 27, 2313-2315.

Kricheldorf, H. R.; Boettcher, C.; Makromol. Chem-Macro. Chem. Phys.
1993, 194, 463-473.

Kricheldorf, H. R.; Meierhaack, J.; Makmmol. Chem-Macro. Chem. Phys.
1993, 194, 715-725.

228

79.
80.
81.
82.
83.

84.

85.
86.
87.

88.

89.

90.

91.

92.

93.

94.

95.

Sandner, B.; Steurich, 8.; Gopp, U.; Polymer 1997, 38, 2515-2522.
Kricheldorf, H. R.; Sumbel, M.; Eur. Polym. J. 1989, 25, 585-591.
Kricheldorf, H. R.; Mahler, A.; Polymer 1996, 37, 4383-4388.
Kricheldorf, H. R.; WeegenSchulz, B.; Polymer 1995, 36, 4997-5003.

Kricheldorf, H. R.; Lee, S. R.; Scharnagl, N.; Macromolecules 1994, 27,
3139-3146.

Kricheldorf, H. R.; Sumbel, M. V.; Kreisersaunders, l.; Macromolecules
1991, 24, 1944-1949.

Kricheldorf, H. R.; Eggerstedt, S.; Macromolecules 1997, 30, 5693-5697.
Kricheldorf, H. R.; Lee, S. R.; Macromolecules 1995, 28, 6718-6725.

Witzke, D. R.; Narayan, R.; Kolstad, J. J.; Macromolecules 1997, 30, 7075-
7085.

Leenslag, J. W.; Pennings, A. J.; Makromol. Chem-Macro. Chem. Phys.
1987, 188, 1809-1814.

Schwach, G.; Coudane, J.; Engel, R.; Vert, M.; J. Polym. Sci. Pol. Chem.
1997, 35, 3431-3440.

Kricheldorf, H. R.; Boettcher, C.; J. Macromol. Sci-Pure Appl. Chem. 1993,
A30, 441-448.

Kricheldorf, H. R.; Damrau, D. 0.; J. Macromol. Sci-Pure Appl. Chem.
1998, A35, 1875-1887.

Nijenhuis, A. J.; Grijpma, D. W.; Pennings, A. J.; Macromolecules 1992, 25,
6419-6424.

Du, Y. J.; Lemstra, P. J.; Nijenhuis, A. J.; Vanaert, H. A. M.; Bastiaansen,
C.; Macromolecules 1995, 28, 2124-21 32.

Kricheldorf, H. R.; Kreisersaunders, l.; Boettcher, C.; Polymer 1995, 36,
1253-1259.

Kowalski, A.; Duda, A.; Penczek, S.; Macromolecules 2000, 33, 689-695.

229

 

96. Yasuda, H.; lhara, E.; In Advances in Polymer Science; Advances in
Polymer Science 133, 1997.

97. Yasuda, H.; lhara, E.; Bull. Chem. Soc. Jpn. 1997, 70, 1745-1767.

98. Yamashita, M.; Takemoto, Y.; lhara, E.; Yasuda, H.; Macromolecules 1996,
29, 1798-1806.

99. Stevels, W. M.; Ankone, M. J. K.; Dijkstra, P. J.; Feijen, J.; Macromolecules
1996, 29, 8296-9303.

100. Stevels, W. M.; Ankone, M. J. K.; Dijkstra, P. L.; Feijen, J.; Macromol.
Chem. Phys. 1995, 196, 1153-1161.

101. Huang, O. H.; Shen, Z. 0.; Zhang, Y. F.; Shen, Y. Q.; Shen, L. F.; Yuan, H.
2.; Polym. J. 1998, 30, 168-170.

102. Shen, Y. Q.; Zhu, K. J.; Shen, Z. 0.; Yao, K. M.; J. Polym. Sci. Pol. Chem.
1996, 34, 1799-1805.

103. Shen, Y. Q.; Shen, Z. 0.; Zhang, F. Y.; Zhang, Y. F.; Polym. J. 1995, 27,
59-64.

104. Shen, Y. Q.; Shen, Z. 0.; Zhang, Y. F.; Yao, K. M.; Macromolecules 1996,
29, 8289-8295.

105. Shen, Y. Q.; Shen, Z. 0.; Shen, J. L.; Zhang, Y., F.; Yao, K. M.;
Macromolecules 1996, 29, 3441 -3446.

106. Shen, Y. Q.; Sheri, Z. 0.; Zhang, Y. F.; Hang, G. H.; J. Polym. Sci. Pol.
Chem. 1997. 35, 1339-1352.

107. Ohse, H.; Cherdron, H.; Korte, F.; Makromol. Chem. 1965, 86, 312.

108. Saotome, K.; Kodaira, Y.; Makromol. Chem. 1965, 82,41.

109. Wilson, D. R.; Beaman, R. G.; J. Polym. Sci, Part A-1 1970, 8, 2161.
110. Perego, G.; Cella, G. D.; Bastioli, C.; J. Appl. Polym. Sci. 1996, 59, 37-43.

111. Macdonald, R. T.; McCarthy, S. P.; Gross, R. A.; Macromolecules 1996, 29,
7356-7361.

230

112.

:113.

114.
115.
'116.
117.

°118.

" 119.
.120.

121.
122.
123.

124.

125.

126.

127.

128.

Grijpma, D. W.; Pennings, A. J.; Macromol. Chem. Phys. 1994, 195, 1633-
1647.

Tsuji, H.; Horii, F.; Hyon, S. H.; lkada, Y.; Macromolecules 1991, 24, 2719-
2724.

lkada, Y.; Tsuji, H.; Macromolecules 1993, 26, 6918-6926.
lkada, Y.; Hyon, S. H.; Tsuji, H.; Macromolecules 1991, 24, 5651-5656.
Tsuji, H.; lkada, Y.; J. Appl. Polym. Sci. 1994, 53, 1061-1071.

Tsuji, H.; lkada, Y.; Hyon, S. H.; Kimura, Y.; Kitao, T.; J. Appl. Polym. Sci.
1994, 51, 337-344.

Tsuji, H.; lkada, Y.; Macromolecules 1993, 26, 6918-6926.
Tsuji, H.; lkada, Y.; Macromolecules 1992, 25, 5719-5723.
Tsuji, H.; Hyon, S. H.; lkada, Y.; Macromolecules 1992, 25, 2940-2946.
Tsuji, H.; Hyon, S. H.; lkada, Y.; Macromolecules 1991, 24, 5651-5656.
Tsuji, H.; Hyon, S. H.; lkada, Y.; Macromolecules 1991, 24, 5657-5662.

HiljanenVainio, M. P.; Drava, P. A.; Seppala, J. V.; J. Biomed. Mater. Res.
1997, 34, 39-46.

Lostocco, M. R.; Huang, S. J.; Abstr. Pap. Am. Chem. Soc. 1996, 212, 245-
PMSE.

Stevels, W. M.; Ankone, M. J. K.; Dijkstra, P. J.; Feijen, J.; Macromol. Symp.
1996, 102, 107-113.

Veld, P.; Velner, E. M.; VanDeWitte, P.; Hamhuis, J.; Dijkstra, P. J.; Feijen,
J.; J. Polym. Sci. Pol. Chem. 1997, 35, 219-226.

Duda, A.; Biela, T.; Libiszowski, J.; Penczek, S.; Dubois, P.; Mecerreyes, 0.;
Jerome, R.; Polym. Degrad. Stabil. 1998, 59, 215-222.

Gallardo, A.; Roman, J.; Dijkstra, P. J.; Feijen, J.; Macromolecules 1998,
31, 7187-7194.

231

129.
130.

131.

132.
133.
134.

135.

136.
137.
138.
139.

140.

141.
142.

143.

144.

145.

King, E.; Cameron, R. E.; Polym. Int. 1998, 45, 313-320.

Storey, R. F.; Herring, K. R.; Hoffman, D. C.; J. Polym. Sci. Pol. Chem.
1991, 29, 1759-1777.

Nakayama, A.; Kawasaki, N.; Maeda, Y.; Arvanitoyannis, |.; Aiba, S.;
Yamamoto, N.; J. Appl. Polym. Sci. 1997, 66, 741 -748.

Cal, J.; Zhu, K. J.; Yang, S. L.; Polymer 1998, 39, 4409-4415.
Ruckenstein, E.; Yuan, Y. M.; J. Appl. Polym. Sci. 1998, 69, 1429-1434.

Kim, Y. J.; Adamson, D. H.; Abstr. Pap. Am. Chem. Soc. 1998, 216, U24-
U24.

Lee, D. S.; Choi, S. W.; Jeon, B. M.; Kim, S. W.; Abstr. Pap. Am. Chem.
Soc. 1998, 216, U878-U878.

Li, S. M.; Anjard, S.; Rashkov, l.; Vert, M.; Polymer 1998, 39, 5421-5430.
Song, C. X.; Feng, X. D.; Macromolecules 1984, 17, 2764-2767.
Choi, Y. R.; Bae, Y. H.; Kim, S. W.; Macromolecules 1998, 31, 8766-8774.

Grijpma, D. W.; Joziasse, C. A. P.; Pennings, A. J.; Macromol. Rapid
Commun. 1993, 14, 155-161.

Joziasse, C. A. P.; Veenstra, H.; Topp, M. D. C.; Grijpma, D. W.; Pennings,
A. J.; Polymer 1998, 39, 467-474.

Li, Y. X.; Kissel, T.; Polymer 1998, 39, 4421-4427.
Eguiburu, J. L.; Berridi, M. J. F.; Sanroman, J.; Polymer 1995, 36, 173-179.

Jerome, R.; Mecerreyes, D.; Tian, D.; Dubois, P.; Hawker, C. J.; Trollsas,
M.; Hedrick, J. L.; Macromol. Symp. 1998, 132, 385-403.

Jacobs, 0.; Dubois, P.; Jerome, R.; Teyssie, P.; Macromolecules 1991, 24,
3027-3034.

Vanhoorne, P.; Dubois, P.; Jerome, R.; Teyssie, P.; Macromolecules 1992,
25, 37-44.

232

146.

147.

148.

149.

150.

151.
152.

153.

154.
155.

156.

157.

158.

159.

Reeve, M. S.; McCarthy, S. P.; Gross, R. A.; Macromolecules 1993, 26,
888-894.

Xiong, C. D.; Cheng, L. M.; Xu, R. P.; Deng, X. M.; J. Appl. Polym. Sci.
1995, 55, 865-869.

Tortosa, K.; Miola, C.; Hamaide, T.; J. Appl. Polym. Sci. 1997, 65, 2357-
2372.

Wang, H.; Dong, J. H.; Qiu, K. Y.; J. Polym. Sci. Pol. Chem. 1998, 36, 695-
702.

Huang, S. J.; Onyari, J. M.; Nicolais, L.; DelNobile, S.; Mensitieri, G.;
Ambrosio, L.; Abstr. Pap. Am. Chem. Soc. 1997, 213, 126-CELL.

Ohya, Y.; Maruhashi, S.; Ouchi, T.; Macromolecules 1998, 31 , 4662-4665.

Ohya, Y.; Maruhashi, S.; Ouchi, T.; Macromol. Chem. Phys. 1998, 199,
2017-2022.

Nijenhuis, A. J.; Grijpma, D. W.; Pennings, A. J.; Polymer 1996, 37, 2783-
2791.

Aoyagi, T.; Miyata, F.; Nagase, Y.; J. Control. Release 1994, 32, 87-96.

Grijpma, D. W.; Kroeze, E.; Nijenhuis, A. J.; Pennings, A. J.; Polymer 1993,
34, 1496-1503.

Sandner, B.; Steurich, S.; Wartewig, S.; Macromol. Symp. 1996, 103, 149-
162.

Storey, R. F.; Warren, S. 0.; Allison, C. J.; Puckett, A. D.; Polymer 1997,
38, 6295-6301.

Grijpma, D. W.; Pennings, A. J.; Macromol. Chem. Phys. 1994, 195, 1649-
1663.

Cook, A. D.; Pajvani, U. B.; Hrkach, J. S.; Cannizzaro, S. M.; Langer, R.;
Biomaterials 1997, 18, 1417-1424.

233

160.

161.

162.

163.
164.
165.

166.

167.

168.

169.
170.

171.

172.
173.
174.
175.

176.

Eguiburu, J. L.; lruin, J. J.; Fernandez-Berridi, M. J.; Roman, J. S.; Polymer
1998, 39, 6891-6897.

Lostocco, M. R.; Borzacchiello, A.; Huang, S. J.; Macromol. Symp. 1998,
130, 151-160.

Yang, J. M.; Chen, H. L.; You, J. W.; Hwang, J. C.; Polym. J. 1997, 29,
657-662.

Tsuji, H.; lkada, Y.; J. Appl. Polym. Sci. 1996, 60, 2367-2375.
Blumm, E.; Owen, A. J.; Polymer 1995, 36, 4077-4081.

Liu, X.; Dever, M.; Fair, N.; Benson, R. S.; J. Environ. Polym. Degrad.
1997, 5, 225-235.

Wachsen, 0.; Platkowski, K.; Reichert, K. H.; Polym. Degrad. Stabil. 1997,
57, 87-94.

Albertsson, A. C.; Sjoling, M.; J. Macromol. Sci-Pure Appl. Chem. 1992,
29, 43-54.

Zhang, X. C.; Wyss, U. P.; Pichora, D.; Goosen, M. F. A.; Polym. Bull.
1992, 27, 623-629.

Sodergard, A.; Nasman, J. H.; Polym. Degrad. Stabil. 1994, 46, 25-30.

Albertsson, A. C.; Palmgren, R.; J. Macromol. Sci-Pure Appl. Chem. 1993,
A30, 919-931.

Wachsen, 0.; Reichert, K. H.; Kruger, R. P.; Much, H.; Schulz, G.; Polym.
Degrad. Stabil. 1997, 55, 225-231.

Kopinke, F. D.; Mackenzie, K.; J. Anal. Appl. Pyrolysis 1997, 40-1, 43-53.
Li, S. M.; McCarthy, 8.; Biomaterials 1999, 20, 35-44.
Hyon, S. H.; Sen-I Gakkaishi 1998, 54, 527-531.

Li, S. M.; Garreau, H.; Vert, M.; Petrova, T.; Manolova, N.; Rashkov, |.; J.
Appl. Polym. Sci. 1998, 68, 989-998.

lwata, T.; Doi, Y.; Macromolecules 1998, 31 , 2461-2467.

234

177. Abe, H.; Doi, Y.; Hori, Y.; Hagiwara, T.; Polymer 1998, 39, 59-67.

178. Gajria, A. M.; Dave, V.; Gross, R. A.; McCarthy, S. P.; Polymer 1996, 37,
437-444.

179. Bischoff, C. A.; Walden, P.; Chem. Ber. 1893, 26, 262-264.
180. Schollkopf, V. U.; Hartwig, W.; Sprotte., U.; Angew. Chem 1979, 91.

181. Kricheldorf, H. R.; Weegenschulz, B.; J. Polym. Sci. Pol. Chem. 1995, 33,
2193-2201.

182. Albertsson, A. C.; Eklund, M.; J. Polym. Sci. Pol. Chem. 1994, 32, 265-279.

183. Deng, X. M.; Zhu, Z. X.; Xiong, C. D.; Zhang, L. L.; J. Appl. Polym. Sci.
1997, 64, 1295-1299.

184. Kricheldorf, H. R.; Lee, S. R.; Macromolecules 1996, 29, 8689-8695.

185. Kasperczyk, J.; Bero, M.; Makromol. Chem-Macro. Chem. Phys. 1993,
194, 913-925.

186. Hall, H. K.; Schneider, A. K.; J. Am. Chem. Soc. 1958, 80, 6409-6412.

187. Thakur, K. A. M.; Kean, R. T.; Hall, E. S.; Kolstad, J. J.; Munson, E. J.;
Macromolecules 1998, 31, 1487-1494.

188. Kasperczyk, J. E.; Macromolecules 1995, 28, 3937-3939.

189. Liu, G.; Fang, Y. E.; Shi, T. Y.; Chem. J. Chin. Univ.-Chin. 1997, 18, 486-
488.

190. Huang, J.; Lisowski, M. S.; Runt, J.; Hall, E. S.; Kean, R. T.; Buehler, N.;
Lin, J. S.; Macromolecules 1998, 31, 2593-2599.

191. Thakur, K. A. M.; Kean, R. T.; Zell, M. T.; Padden, B. E.; Munson, E. J.;
Chem. Commun. 1998, 1913-1914.

192. Zell, M. T.; Padden, B. E.; Paterick, A. J.; Hillmyer, M. A.; Munson, E. J.;
Thakur, K. A. M.; Kean, R. T.; Abstr. Pap. Am. Chem. Soc. 1998, 215,
U310-U310.

235

 

 

193. Thakur, K. A. M.; Kean, R. T.; Hall, E. S.; Doscotch, M. A.; Munson, E. J.;
Anal. Chem. 1997, 69, 4303-4309.

194. Thakur, K. A. M.; Kean, R. T.; Zupfer, J. M.; Buehler, N. U.; Doscotch, M. A.;
Munson, E. J.; Macromolecules 1996, 29, 8844-8851.

195. Weidinger, A.; Hermans, P. H.; Makromol. Chem. 1961, 44-46, 24.
196. Weidinger, A.; Hermans, P. H.; Makromol. Chem. 1961, 50, 98.
197. Weidinger, A.; Hermans, P. H.; Makromol. Chem. 1962, 52, 169.
198. Takayama, Y.; Bull. Chem. Soc. Jpn. 1933, 8, 173-176.

 

199. Schmidt, E.; Sachtleben, R.; Justus Liebigs Ann. Chem. 1878, 193, 112.

200. Adamson A. 6., Physical Chemistry of Surfaces; Fifth ed.; John Wiley &
Son:, 1990. 28-30.

236