Thesis

Synthesis and Properties of
Starch Based Biomaterials

Asaf Kleopas Sugih

The author thanks the University of Groningen for the financial support through
an Ubbo Emmius Scholarship.

RIJKSUNIVERSITEIT GRONINGEN

Synthesis and Properties of
Starch Based Biomaterials

Proefschrift

ter verkrijging van het doctoraat in de
Wiskunde en Natuurwetenschappen
aan de Rijksuniversiteit Groningen
op gezag van de
Rector Magnificus, dr. F. Zwarts
in het openbaar te verdedigen op
vrijdag 12 december 2008
om 13.15 uur

door

Asaf Kleopas Sugih
geboren op 4 juli 1975
te Bandung, Indonesië

Promotores : Prof. dr. ir. H.J. Heeres
Prof. dr. ir. L.P.B.M. Janssen
Prof. dr. F. Picchioni

Beoordelingscommissie : Prof. dr. A.A. Broekhuis
Prof. dr. A.J. Minnaard
Prof. dr. L. Moscicki

ISBN : 978‐90‐367‐3592‐6

ISBN : 978‐90‐367‐3593‐3 (electronic version)

to:
Tresna, my parents, and my sister
Table of Contents

Chapter 1: Introduction
1.1. Starch 2
1.1.1. Starch production processes 3
1.1.2. Structure and properties of starch 4
1.2. Biomaterials from starch 5
1.2.1. Plastic applications and waste issues 5
1.2.2. The potential of biodegradable plastics 7
1.2.3. Biodegradable plastics from starch 9
1.3. Starch modifications to improve product properties 9
1.3.1. Thermoplasticized starch 10
1.3.2. Cross‐linked starch 11
1.3.3. Starch Esters 11
1.3.4. Starch ‐ biopolymer blends and graft co‐polymers 13
1.3.4.1. Starch‐based blends by melt mixing 14
1.3.4.2. Starch‐based blends by in situ polymerization 16
1.4. Thesis Outline 16
1.5. References 17

Chapter 2: Experimental Studies on the Ring Opening
Polymerization of p‐dioxanone using an Al(OiPr)3‐
Monosaccharide Initiator System
2.1. Introduction 26
2.2. Materials and Methods 27
2.2.1. Materials 27
2.2.2. Methods 27
2.2.2.1. Typical example for the synthesis of polydioxanone end‐capped
with 1,2;3,4‐di‐O‐isopropylidene‐α‐D‐galactopyranose (2) 27
2.2.3. Product analyses 28
2.2.4. Calculation of average degree of polymerization 28
2.3. Results and Discussions 29
2.3.1. Screening experiments 29
2.3.1.1. Product analyses 30
2.3.1.2. Mechanistic aspects 36
2.3.2. Systematic studies 38
2.3.2.1. Product yield 39
2.3.2.2. Effects of process conditions on the average chain length ( X n exp )
and end group distribution 40
2.4. Conclusions 42
2.5. Nomenclature 43
2.6. References 44

Chapter 3: Synthesis of Poly‐(ε)‐caprolactone Grafted Starch
Co‐polymers by Ring Opening Polymerisation using
Silylated Starch Precursors
3.2.1. Materials 49
3.2.2. Methods 50
3.2.2.1. Typical example of the starch silylation procedure 50
3.2.2.2. Typical example of in situ polymerization of ε‐CL with silylated
starch 50
3.2.2.3. De‐silylation of poly‐(ε)‐caprolactone grafted silylated starch
co‐polymers 51
3.2.2.4. Peracetylation of silylated starch 51
3.2.3. Analytical methods 52
3.2.3.1. Nuclear Magnetic Resonance (NMR) 52
3.2.4. Calculations 52
3.3. Results and Discussions 53
3.3.1. Synthesis of silylated starch 54
viii

3.3.2. In situ ring opening polymerization of ε‐caprolactone with
silylated starch 57
3.3.3. Deprotection of silylated‐starch‐g‐PCL 63
3.4. Conclusions 64
3.6. References 65

Chapter 4: Synthesis of Higher Fatty Acid Starch Esters using Vinyl
Laurate and Stearate as Reactants
4.2.1. Materials 71
4.2.2. Analytical equipment 71
4.2.3. Methods 72
4.2.3.1. Typical example of the synthesis of laurate and stearate esters
of corn starch 72
4.2.3.2. Peracetylation procedure 73
4.2.3.3. Determination of the Degree of Substitution (DS) 74
4.3. Results and Discussion 75
4.3.1. Exploratory experiments 75
4.3.2. Product characterisation 76
4.3.2.1. 1H‐ and 13C‐NMR analyses 76
4.3.2.2. FT‐IR measurements 78
4.2.3. Systematic studies 79
4.3.3.1. Effect of vinyl ester to AHG ratio on the product DS 80
4.3.3.2. Effect of the addition of toluene as a co‐solvent 81
4.3.3.3. Catalysts screening 81
4.4. Conclusions 82
4.6. References 84

ix
Chapter 5: Experimental and Modeling Studies on the Synthesis
and Properties of Higher Fatty Esters of Corn Starch
5.2.1. Materials 87
5.2.2. Analytical Equipment 87
5.2.3. Methods 88
5.2.3.1. Typical example of the preparation of laurate and stearate starch
Esters 88
5.2.3.2. Peracetylation procedure and Degree of Substitution (DS)
determination 88
5.2.4. Experimental Design 88
5.3. Results and Discussion 89
5.3.1. Mathematical modelling 92
5.3.2. Product properties 96
5.4. Conclusions 102
5.6. References 103

Chapter 6: Synthesis and Properties of Reactive Interfacial Agents
for Polycaprolactone‐Starch Blends
6.2.1. Materials 109
6.2.2. Methods 109
6.2.2.1. Compatibilizer synthesis 109
6.2.2.2. Work‐up of PCL‐g‐GMA products 110
6.2.2.3. Work‐up of PCL‐g‐DEM 110
6.2.2.4. Preparation of PCL‐starch blends with the reactive compatibilisers 110
6.2.3. Analytical Methods 110
x

6.2.3.1. Calculation of the Degree of Functionalization (FD) of the
reactive compatibilizers 111
6.2.4. Statistical mdeling 111
6.3. Results and Discussions
6.3.1. Preparation of the ractive compatibilizers 111
6.3.1.1. Effect of substrate (GMA/DEM) to PCL Ratio on the FD 116
6.3.1.2. Effect of the BPO intake on the product FD 117
6.3.1.3. Modeling of the combined effects of the GMA/ DEM and BPO
intakes on the FD 118
6.3.1.4. Thermal properties of the compatibilizers 120
6.3.2. Synthesis and Properties of Starch‐ PCL Blends 121
6.3.2.1. Binary blends of starch and PCL 122
6.3.2.2. Ternary blends compatibilized with PCL‐g‐DEM 123
6.3.2.3. Ternary blends compatibilized with PCL‐g‐GMA 125
6.4. Conclusions 128
6.6. References 129

Summary 133

Samenvatting (Dutch Summary) 137

Acknowledgements 141

List of Publications 143

xi
Chapter 1
Introduction

Abstract
A general overview of starch properties will be given and the potential use of
starch as a starting material for a wide range of green biomaterials will be
reviewed and discussed. Different routes to modify starch to improve the product
properties and to extend the application range will be provided. Finally, an outline
of this thesis is given.

Chapter 1
1.1. Starch
75% of all organic material on earth is present in the form of polysaccharides.
[1]. An important polysaccharide is starch. Plants synthesise and store starch in
their structure as an energy reserve. It is generally deposited in the form of small
granules or cells with diameters between 1‐100 µm [2]. Starch is found in seeds
(i.e. corn, maize, wheat, rice, sorghum, barley, or peas) and in tubers or roots (i.e.
potato or cassava) of the plants [2‐3]. Most of the starch produced worldwide is
derived from corn, but other types of starch such as cassava, sweet potato, potato,
and wheat starch are also produced in large amounts [2, 4‐5]. Most starch crops
are very productive. Potato accumulates starch to approximately 75 % of the dry
weight in the tubers with a yield up to 21 ton starch per hectare, while corn seeds
consist of 65‐80% starch by weight, with an average yield of 4.9 ton starch per
hectare [6].
The worldwide production of starch in 2008 is estimated to be around 66
million tons [7]. Most of the starch is produced in the USA. The second and third
starch producer regions are Europe and Asia [4‐5]. Past, current and future
estimations of worldwide starch production are given in Figure 1.1. [7]. The
current price of corn starch is around $0.45 (estimated from the total volume and
value of the US corn starch export [8]).

80
70
60
Starch Production (Million Tons)
50
40
30
20
10
0
1995 2000 2005 2010
Year

Figure 1.1. Past, present and a forecast of starch production [7]
o : Europe □ : USA
∆ : the Rest of the World : Total
2

Introduction
1.1.1. Starch production processes
Starch is generally extracted from the plant by wet milling processes [9‐10]. The
plant material is grounded in water, the debris is filtered from the slurry, and
starch granules are obtained after centrifugation from the suspension. As an
example, a typical corn starch production process is given in Figure 1.2.a. The
starch is present in the endosperms (floury and horny), and is embedded in a
proteinaceous cellular matrix as is shown in Figure 1.2.b. After initial cleaning to
remove cob, sand, and other foreign materials, the corn kernel is softened by
steeping in warm water containing SO2 until the volume of the kernel increases
with 55‐65%. After a coarse milling, the mixture is fed to a hydrocyclone to
separate the germ and the rest of the kernel, which is again fed to a second milling
process. The resulting suspension from the mills contains fiber, gluten, and starch.
The fiber is removed using washing screens. The low density gluten is separated
from the starch suspension by centrifugation. The resulting starch is further
washed in a cyclone and finally dried.

a. b.

Figure 1.2. Corn starch production [9]
a. Typical corn‐milling operation
b. Cross‐sectional view of a corn kernel
3
Chapter 1
1.1.2. Structure and properties of starch
Starch is a polymer consisting of anhydroglucose (AHG) units (see Figure 1.3.a.)
[2]. Two types of AHG polymers are usually present in starch: amylose and
amylopectin [2‐3, 6]. Amylose is essentially a linear polymer in which AHG units
are predominantly connected through α‐D‐(1,4)‐glucosidic bonds. The molecular
weight of amylose is a function of the plant source and processing method, but
usually in the range of 1.6‐7 x 105 Da [9]. Amylopectin is a branched polymer,
containing periodic branches linked with the backbones through α‐D‐(1,6)‐
glucosidic bonds [2]. Each branch contains about 20‐30 anhydroglucose units. The
molecular weight of amylopectin is higher than that of amylose and is typically 4‐5
x 108 Da [9]. The content of amylose and amylopectine in starch varies and largely
depends on the starch source. Typically, the amylose content is between 18‐28%
[2]. The amylose content of several common starches is given in Table 1.1. [3].

H2COH H2COH H2COH H2COH

O O O O
H H H H H H H H
H H H H
... OH H ... OH H OH H OH H
O ... O O O ...
H OH H OH H OH H OH
a. b.
H2COH H2COH
O O
H H H H
H H
... OH H OH H
O
H OH H OH O
H2COH H2COH CH2

O O O
H H H H H H
H H H
... OH H OH H OH H
O O O ...
H OH H OH H OH
c.

Figure 1.3. Chemical Structure of Starch
a. Anhydroglucose (AHG) unit
b. Amylose
c. Amylopectin

Starch is insoluble in cold water, but it is very hygroscopic and binds water
reversibly. Heating a starch solution leads to loss of hydrogen bonding in the
interior of the starch granule and the starch will start to gelatinize. The starch
granules will swell rapidly to many times of its original volume. The linear
4

Introduction
amylose molecules leach out of the granules into the solution. The resulting
suspension contains a mixture of linear amylose molecules, swollen granules, and
granule fragments, and, depending on the amount of water present, will form a
thick paste or gel. The gelatinization temperature range can be defined as the
temperature at which granular swelling begins until the temperature when nearly
100% of the granules are gelatinized [9]. The gelatinization temperature range of
various starch sources is given in Table 1.2.

Table 1.1. Amylose content of common starches [3]
Starch Amylose (%)
Arrowroot 20.5
Corn 28
Hybrid amylomaize Class V 52
Hybrid amylomaize Class VII 70‐75
Oat 27
Manioc 15.7
Potato 20
Rice 18.5
Sago 25.8
Sweet potato 17.8
Tapioca 16.7
Wheat 26

Table 1.2. Starch gelatinization temperature range [9]
Gelatinization
Starch
Temperature Range [o C]
Potato 59‐68
Tapioca 58.5‐70
Corn 62‐72
Waxy corn 63‐72
Wheat 58‐64

1.2. Biomaterials from starch
1.2.1. Plastic Applications and Waste Issues
Plastic is the general term for a wide range of synthetic or semisynthetic
polymerisation products. Plastics are used in a wide range of applications and the
demand is still increasing every year [11]. The first generation of commercial
plastics was derived from cellulose nitrate and is known as celluloid [12‐14].
Cellulose nitrate was first prepared by A. Parker in 1838, and celluloid was
5
Chapter 1
patented by J. Hyatt in 1870 [13]. While celluloid is derived from a natural
polymer (cellulose), the oldest purely synthetic plastic is Bakelite, discovered by
Baekeland in 1907 [12, 15]. A dramatic increase in demand for plastics began after
World War II, when polyethylene (PE) was invented. PE is a very versatile plastic
because it can be shaped easily into various forms, for instance to be used in
packaging and paper coatings [13].
Plastics are very attractive materials. They have a low density and can be
shaped in thin films while maintaining good properties. The latter is important
when using the material for packaging purposes to save weight, space, and energy
during transportation of goods. Plastics have lower melting temperatures
compared to glass and metals, and therefore need less energy to shape it into
useful materials [13]. The production and consumption of plastics has increased
significantly with a rate of almost 10% every year since 1950. In 2006, the
worldwide plastics production has reached 245 million ton per annum [11] (see
Figure 1.4.). The largest application of plastics is for packaging purposes. About
29% of the total plastics produced in the USA [1], and 37% of the total plastics
demand in EU [11] is used as packaging materials. Important polymers used for
packaging are polyethylene (HDPE and LDPE), polypropylene (PP), polystyrene
(PS), polyvinyl chloride (PVC), polyethylene terephtalate (PET), and
polycarbonates (PC). Plastics are also used for building materials and automotive,
electrical, and consumer products (see Figure 1.5.)

250
Included:
Thermoplastics, Polyurethanes, Thermosets, Elastomers, 2006: 245
Adhesives, Coatings, Sealants, and PP-Fibers
200
2002: 200
Not Included:
PET-, PA-, and PolyAcryl- Fibers
Plastics Production (Million Tons)
150
1989: 100
100
1976: 50
50
1950: 1.5
0
1950 1960 1970 1980 1990 2000 2010
Year

Figure 1.4. Worldwide Plastics Production (1950‐2006) [11]
6

Introduction
Plastic waste, however, is causing serious environmental problems. It has a
high volume to weight ratio and is resistant to degradation. Plastics have been
polluting sea [1, 16], soil, rivers, and lakes, threatening fishery, ship navigation,
hydropower plants operation, irrigation, and other public works [17]. Plastic litter
is hazardous to wildlife [1, 6], and accumulation of its residues in soil cause
significant reductions in agricultural yields.
The disposal of plastics materials in municipal solid waste (MSW) is a serious
issue in many parts in the world. In the USA, the amount of plastics in MSW
increased from less than 1% in the 1960 to 11.7% in 2006 [18]. The volume fraction
of plastic in MSW is much larger due to the low density of plastics, and may be
more than twice the weight fraction [13]. In 1996, plastics waste was ranked as the
second major source of MSW after paper and paperboard, consisting of 25%‐v of
the total waste [13]. Recycling (part of) the plastics to reduce the amount of MSW
also has limitations. It is not applicable for thermoset resins [19], and is only
effective for single plastic sources or simple plastic formulations. Co‐mingled
plastics, which are usually found in waste streams, are not easy to recycle [19]. In
Europe (2006), only 20% of the plastic is recycled. Most of the waste is still
disposed (50%) by landfill or incinerated to recover the energy (30%) [11].

Figure 1.5. Plastic applications and waste treatment in Europe [11]

1.2.2. The potential of biodegradable plastics
The application of biodegradable plastics could be an attractive solution for the
problems related to the use of conventional plastics (vide infra). Biodegradable
plastics are polymeric materials capable of decomposing when given an
7
Chapter 1
appropriate environment and sufficient amount of time [15]. Biodegradable

plastics have gained considerable interest since the 1980s. Nowadays new types of
biodegradable plastics with improved properties and lower costs have been
developed [13]. Biodegradable plastic waste may be treated in composting
facilities, together with food and yard waste as well as paper. It may also be
treated in sewage sludge water treatment plants or buried in the soil [20‐21]. The
considerable growth of interest in composting as a means to replace landfill due to
the decreasing disposal spaces (especially in Europe) may also help the progress of
biodegradable plastics development [13].
Several authorities have provided definitions for biodegradable plastics [22],
and these are shown below:
ISO 472: 1998 ‐ A plastic designed to undergo a significant change in its chemical
structure under specific environmental conditions resulting in a loss of some
properties that may vary as measured by standard test methods appropriate to the
plastics and application in a period of time that determines its classification. The
change in chemical structure results from the action of naturally occurring micro‐
organisms.
ASTM sub‐committee D20.96 proposal ‐ Degradable plastics are plastic materials
that undergo bond scission in the backbone of a polymer through chemical,
biological and/or physical forces in the environment at a rate which leads to
fragmentation or disintegration of the plastics.
Japanese Biodegradable Plastic Society draft proposal ‐ Biodegradable plastics
are polymeric materials which are changed into lower molecular weight
compounds where at least one step in the degradation process is through
metabolism in the presence of naturally occurring organisms.
DIN 103.2 working group on biodegradable polymers ‐ Biodegradation of a
plastic material is a process leading to naturally occurring metabolic end products.

Biodegradable plastics may be classified into two general groups, biopolymers
from nature (from plants, animals, and microorganisms) and biodegradable
synthetic polymers [1]. Common natural biopolymers are carbohydrates, proteins
(present abundantly in plants and animals), and polyesters from micro‐organisms
[1]. Biopolymers are inherently biodegradable, because they take part in nature’s
cycle of renewal.
The market price of biodegradable plastics (mainly from starch and PLA) has
decreased the last years, and since 2000 they have become competitive with
traditional materials in some applications [20‐21]. The price of these biodegradable
plastics is expected to be reduced considerably in the next decade [23]. For
instance, the price of PLA based plastics in 2010 is estimated to be around € 1.5/
8

Introduction
kg, half of its price in 2003 (around 3 Euro/kg). The price of starch‐based
biodegradable plastics (€ 1.50‐4.50/ kg in 2005) is also expected to be lower in the
future due to considerable reduction in modification costs [23]. Biodegradable
plastics are especially very useful for single‐use applications, when recycling is not
practical or uneconomical [20‐21], or when environmental impacts have to be
minimized. Examples are the use as packaging material for carrier bags, consumer
products and food products. Biodegradable plastics are also used for food
servicewares [20], towels, utensils [15], agricultural products (mulch films, pots),
hygienic materials (diapers, napkins) [24], breathable fabrics, biofillers for tires,
and chewable items for pets [1, 20].

1.2.3. Biodegradable Plastics from Starch
Starch is a very attractive source for the development of biodegradable plastics.
[6]. The price of starch in 2007 was about $ 0.45 per kg [8]. This price is much
lower than conventional plastics derived from oil, such as polyethylene (PE, € 1.2‐
1.4 or $ 1.8‐ 2.4 per kg in 2007‐2008 [25]). Starch may become an attractive raw
material for plastics in the future, because the price of oil based polymers may still
increase due to the rise in the crude oil prices [25]. Bastioli [20‐21] showed that
nearly all biodegradable plastics available in the market are derived from starch,
either from starch‐based materials (slightly modified starch, alone or complexed
with natural or synthetic biodegradable polymers) or from polylactic acid which
originates from the fermentation of a starch feedstock. The global production
capacity of starch‐based bioplastics in 2010 is estimated to increase to 200‐300
kiloton per year from 77‐200 kiloton in 2003 [23].
Virgin starch is not suitable as a packaging material. It cannot be shaped in
films with adequate mechanical properties (high percentage elongation, tensile
and flexural strength) and is too sensitive to water [26‐27]. Consequently, starch
must be modified, either by plasticization, blending with other materials, chemical
modification or combinations of them [26], before they can be applied as
biodegradable plastics.

1.3. Starch modifications to improve product properties
Several techniques may be applied to develop starch based biomaterials with
improved properties. These are summarized briefly in the next sections.

9
Chapter 1
1.3.1. Thermoplasticized starch
Virgin starch is brittle and difficult to be processed into articles due to its
relatively high glass transition temperature (Tg, approximately 230 oC) [28], which
is even above the thermal degradation temperature. [29]. The brittleness is known
to increase in time due to free volume relaxation and retrogradation. This problem
is mainly caused by the presence of strong inter‐ and intra‐molecular hydrogen
bonds between the starch macromolecules [28, 30].
Starch can be modified to obtain materials which melt below the decomposition
temperature [31], and therefore are processable by conventional polymer
processing techniques such as injection, extrusion, and blow moulding [28, 30, 32‐
36]. The modified products are known as thermoplastic, destructed, plasticized, or
melted starch [29, 33]. The modification involves break down of the starch
granular structure by the use of plasticizers at high temperatures (90‐180 oC) and
shear, which will result in a continuous phase in the form of a viscous melt [29‐30,
32, 37]. The thermoplasticization process will decrease the interactions of the
molecular chain and destruct the structure of the starch [38]. As the result, the
semicrystalline structure of starch and its granular form are lost and the starch
polymers are partially depolymerized, resulted in the formation of an amorphous
mass [35, 39].
There are several substances used as plasticizer for the preparation of
thermoplastic starch (TPS), such as water and polyols (glycerol, glycol, sorbitol,
sugars) [30, 40]. The use of water as a plasticizer is not preferable, because the
resulting product will be brittle when equilibrated with ambient humidity [35].
The use of other plasticizers (for example glycerol) results in a rubbery material,
with better properties than virgin starch in various applications.
Yu, et al [28] discovered that the elongation of break of the thermoplastic starch
is significantly improved by plasticization with glycol, glycerol, and hexylene
glycol. The plasticized starch properties may be tuned by changing the
temperature of processing, water content, and the properties and amount of
plasticizers. For instance, the thermal properties of glycerol‐plasticized starch are a
function of water content [31, 35, 40]. At intermediate water levels, phase
separation may still occur. A biodegradation study according to ISO/CEN 14852
and ASTM D5209‐92 standards [41] on films made from starch–glycerol–water
mixtures confirmed that the films are easily biodegradable.
Although thermoplasticization seems to be a promising method, TPS
synthesized from polyol and sugar plasticizers have the tendency to re‐crystallize
(retrogradation) after being stored for a period of time, which results in
embrittlement. Another issue is the poor water resistance and low strength, which
still limits their use. The plasticizers are also usually hydrophilic and can be
10

Introduction
washed out by water [29]. Solutions to improve the properties of TPS are blending,
or coating with hydrophobic polymers [33, 39].

1.3.2. Cross‐linked starch
An anhydroglucose molecule of starch contains two secondary and one primary
hydroxyl group. These hydroxyls can react easily with a wide range of
compounds such as acid anhydrides, organic chloro‐compounds, aldehydes,
epoxy, and ethylenic compounds. Chemicals of these classes having two or more
of the reactive groups may react with two or more hydroxyls of the starch
molecules. The products are called cross‐linked starches. Cross‐linking results in a
reduction of the solubility in water and to thickening, leading to higher viscosities
[42‐43]. The cross‐linked starches have found many applications, especially in the
food, paper, textile, and adhesive industry.
Silva, et al [43] studied the mechanical properties of films from maize starch
cross‐linked by sodium trimetaphosphate (SMTP). Higher processing
temperatures generally led to higher cross‐linking levels, resulting in an increase
of the Young’s modulus and tensile strength of the products and a decrease in
elongation at break. The cross‐linked products are therefore more rigid (and less
elastic) materials than virgin starch [43].

1.3.3. Starch Esters
The development of starch esters started in the mid 19th century [44, 46].
Already in 1865, Schuetzenberger acetylated starch with acetic anhydride [44].
Most studies dealt with the synthesis of starch esters of C1‐C4 carboxylic acids,
and particularly with acetic acid [45].
The simplest starch ester is starch formate (C1), synthesized by direct addition
of formic acid to starch at room temperature. The esterification reaction is
catalyzed by H+ and proceeds with the formation of water [46]. Breakdown of the
starch and the formation of low molecular weight products occur to a significant
extent [47]. As a result, the use of formyl esters of starch is much lower than that of
acetate esters.
The most popular starch ester is starch acetate [46]. Three different types of
starch acetates may be distinguished, differing in Degree of Substitution (DS). The
DS is defined as the moles of substituents per mole of AHG units [48]. Low DS
products (0.01‐0.2) are commercially available and used as food additives and in
the textile industry. Medium DS starch acetates (0.3‐1) are still soluble in water,
while highly substituted starch acetate (DS of 2‐3) are soluble in organic solvents.
Substitution of the hydroxyl groups of starch with acetate groups makes the esters
more hydrophobic than native starch. High DS starch acetate can be easily casted
11
Chapter 1
into films using organic solvents. The degree of acetylation of starch acetates can
be easily controlled, allowing the polymer to be produced with a range of
hydrophobicities [22]. High DS starch acetates are thermoplastic materials suitable
to be used as biodegradable plastics [45].
Several synthetic routes have been developed for starch esters. Esterification
may be performed using acid anhydrides in aqueous media or organic solvents
(pyridine, DMSO, xylene, DMF, or isopropanol) with acidic (hydrochloric or
sulfuric acid) or alkaline (NaOH or triethylamine) catalysts [26]. Starch triacetate
has been successfully synthesized using acetic anhydride in combination with
pyridine‐gelatinized starch [44]. Starch esters have also been synthesized using
alkanoyl chlorides [49‐53]. Another attractive route involves the use of vinyl esters
as reagents [54]. The kinetics of the reaction between gelatinized aqueous potato
starch and vinyl acetate was studied by De Graaf, et al [55‐56], and is shown in
Scheme 1.1. Acetylation of starch in water and DMSO using vinyl acetate has been
studied lately by Mormann and Al‐Higari [57] as well as Dicke [58]. Neutral and
weak acid/ alkaline catalysts result in regioselective substitution at the C2
hydroxyl groups of starch, while alkaline catalysts (such as carbonate,
hydrogencarbonate, acetate, and phosphate) will result in C‐2, but also C‐6 and C‐
3 substitution [58].

- -
Starch OH + OH Starch O + H2O
-
O O
-
Starch O + H3C C O CH CH2 H3C C O CH CH2
O Starch
O O
H2O -
H3C C O Starch + H3C C + OH
H
-
O O
-
H3C C O Starch + OH H3C C O Starch
OH
O
H2O -
H3C C OH + Starch O

Scheme 1.1. Mechanism of starch acetylation using vinyl acetate

12

Introduction
The synthesis of long chain fatty acid ester of starch has attracted much interest
lately. The introduction of longer acid chain is expected to reduce the brittleness of
virgin starch and to increase its hydrophobicity [48]. Fatty‐acid starch esters have
been synthesized using fatty acid (octanoyl, dodecanoyl, octadecanoyl) chlorides
[49‐53]. The fatty acid chloride reactants are, however, relatively expensive and
rather corrosive. The use of methyl and glyceryl esters of fatty acid (in the absence
of solvent) to synthesize starch fatty acid esters has also been studied [59], but only
relatively low‐DS (0.34‐0.61) products could be obtained using this approach.

1.3.4. Starch ‐ Biopolymer Blends and Graft Co‐Polymers
Blending of different polymers is an established method to obtain products
with improved properties. However, polymers are rarely miscible with each other
[60] so that, in the simplest case of a binary blend, one component will be
dispersed into the other. The degree of adhesion (binding) between the dispersed
phase and the matrix is dependent on the molcular interactions between the two
components and represents a crucial factor in determining the morphology of the
blends and eventually the product performance [60]. Two main methodologies are
applied for the production of polymer blends. The first involves simple melt
mixing of the two components for example by extrusion. By working at
temperatures above the melting point or glass transition temperature of the two
components, the latter are mixed together. If the right combination of chemical
groups on the two components is present along the polymers backbone, a
chemical reaction might take place upon processing (reactive extrusion). The
second methodology for producing polymer blends involves the in situ
polymerization of one component (thus originally present in the blend in
monomeric form) in the presence of the second one. The classical example of such
process is represented by the production of High Impact Poystyrene (HIPS)
obtained by styrene polymerization in the presence of polybutadiene [61].
Although the in situ polymerization process is not as technologically
straightforward and economically convenient as melt‐mixing, it is frequently used
in order to chemically graft the polymerized component on the other one
(polystyrene on polybutadiene in our example above [61]). As a result, the two
components are chemically linked to each other, which in turn provide a very
strong adhesion at the molecular level between the dispersed phase and the
matrix. Despite this advantage, melt‐mixing remains the preferred route to
polymer blends mainly because of very practical reasons: low costs, availability of
mixing equipment and no necessity to use any organic solvent (often employed for
the in situ polymerization). However, in order to achieve also a good molecular
adhesion between the phases by melt mixing, interfacial agents (e.g.
compatibilizers) might be used. Their role is comparable to the one of a surfactant
in emulsion formation [61], i.e. they locate themselves at the interface between the
13
Chapter 1
two components stabilizing the dispersion (most probably by a steric repulsion
mechanism [62]) and providing an improved adhesion at the interface. A suitable
interfacial agent for the blends of two polymeric materials is a block copolymer for
which the chemical structure of every block is the same (or very similar) to the one
of the individual components to be blended [63] (illustrated in Figure 1.6.).
Interfacial agents already available on the market can be used as such or can be
produced upon mixing (in situ) by chemical reaction of the two components.
In the past, there have been efforts to blend as well as to graft synthetic
polymers onto starch. The synthesis of compatibilizer and its use for starch/
synthetic polymer blending has also been studied. The products have been
synthesized in the lab as well as on industrial scale [32, 39, 64‐98]. The blending
and grafting of starch with synthetic polymers is usually performed to achieve
higher hydrophobicity and to improve the mechanical and thermal properties as
well as to obtain cheaper and more biodegradable products.

Matrix
Dispersed
Phase

Figure 1.6. Illustration of the role of interfacial agent in compatibilizing blends

1.3.4.1. Starch‐based blends by melt mixing
Starch‐synthetic polymer blending has been studied as early as in 1973 [64‐65].
The most often used synthetic polymer for blending with starch is polyethylene
[27, 66‐81]. The starch/polyethylene blends are used for agricultural mulch [27, 80‐
81] or food packaging [78]. The uncompatibilized blends of starch and
14

Introduction
polyethylene show a coarse phase separation due to differences in polarity of
starch (hydrophilic) and polyethylene (hydrophobic). The mechanical properties
of these blends (tensile strength and elongation at break) decrease at higher starch
content. Polyethylene‐g‐maleic anhydride (PE‐g‐MA) and polyethylene‐g‐glycidyl
methacrylate (PE‐g‐GMA) have been used as reactive compatibilizer for starch/PE
blends. PE‐g‐MA contains reactive anhydride sites, while PE‐g‐GMA posseses
epoxide groups, which both can react in situ with the hydroxyl groups of starch
[70, 72‐73]. As a result a graft copolymer (PE‐g‐Starch), i.e. the compatibilizer, is
formed upon mixing, thus improving the dispersion of starch in the PE matrix.
The mechanical properties of the blend are also improved. The tensile strength of
the uncompatibilized blends is drastically reduced when starch content is
increased, while the tensile strength of the compatibilized ones decreases only
slightly with the starch content.
The use of conventional synthetic polymer such as PE for blending with starch
will only result in a partially biodegradable material, since the conventional
synthetic polymers are usually poorly or non‐biodegradable. To obtain completely
biodegradable products, synthetic biodegradable polymers have been applied,
among which synthetic polyesters are considered very promising materials [82].
Examples of these biodegradable polyesters are poly‐glycolide, poly‐dioxanone,
poly‐lactides, and poly‐lactones (such as poly‐butyrolactone, poly‐valerolactone,
and poly‐caprolactone). The ester bonds of these polymers are susceptible to
attack by water and this leads to enhanced biodegradability. These biodegradable
polyesters will finally decompose into non‐toxic products [73]. Some of these
polyesters also have very good mechanical, thermal and water/gas permeability
properties that are even comparable to bulk non‐biodegradable polymers such as
PE and PP, and EVOH [26].
Polycaprolactone (PCL) is a well‐known synthetic biodegradable polyester,
which combines excellent biodegradability with acceptable mechanical properties.
Studies on the blending of starch with PCL have been already described [32, 39,
84‐98]. As is the case of blends with PE, the uncompatibilized blends of starch with
PCL give coarse phase separation and a reduction in the mechanical properties
when the starch content is increased. The use of reactive compatibilizer precursors
PCL‐g‐glycidyl methacrylate (PCL‐g‐GMA) [89‐90], PCL‐g‐pyromellitic anhydride
[94‐95], PCL‐g‐maleic anhydride (PCL‐g‐MA) [91, 96], dextran‐g‐PCL [92], and of a
premade starch‐g‐PCL [93] for starch/PCL blends resulted in a better dispersion
between the phases and in turn in improved mechanical properties compared to
the uncompatibilized blends. Despite these good results, systematic studies on the
compatibilizer precursor synthesis as well as on the influence of the chain
topology and chemical reactivity (for the in situ compatibilizer formation) have not
yet been reported.

15
Chapter 1
1.3.4.2. Starch‐based blends by in situ polymerization
The synthesis of starch based graft copolymers by in situ polymerization
represents not only an alternative route to melt blending (vide supra) for the
production of novel biomaterials, but it can also provide an efficient synthetic
methodology for the production of a compatibilizer to be used for melt blending
of starch with a biopolymer [93].
In the past, starch‐g‐PCL has been synthesized using toxic materials such as
isocyanates [93]. Another approach consists of the Ring Opening Polymerization
(ROP) of ε‐caprolactone monomer in the presence of starch. In this reaction, the
hydroxyl groups of starch are supposed to function as initiating sites. Previous
studies showed that common ROP catalysts such as tin octoate or aluminium
isopropoxide gave low (0‐14%) grafting efficiencies (GE, defined as the percentage
of grafted polyester to starch compared to the total amount of homopolymer and
grafted polyester) [97]. Starch is a very hydrophilic material that always contains
moisture. The water in starch granules competes with the hydroxyl groups of
starch in the initiation step of the polymerization reaction, leading to formation of
PCL homopolymers rather that starch‐g‐PCL, thus resulting eventually in low GE
values [97]. Another possible cause for the low GE values is the heterogenous
nature of the reaction. Starch is insoluble in the typical organic solvents used for
ROP (such as toluene or THF), and the presence of liquid‐solid reaction system
leads to reduced reaction rates between starch and ε−caprolactone. The highest GE
value (up to 90%) has been achieved when using triethylaluminium as catalyst
[97‐98], which however is extremely air‐ and water‐sensitive, and difficult to
handle since it releases ethane, a very flammable by‐product, during the reaction.
A new strategy for the in situ ROP of ε−caprolactone on starch with the use of
common ROP catalyst is therefore highly desirable.

1.4. Thesis Outline
The objective of this thesis is to study synthetic pathways to obtain starch
derivatives with the potential to be used as bioplastic. Three routes have been
studied in detail: starch esterification, starch/ biopolymer blending, and starch‐g‐
biopolymer formation.
Chapter 2 describes the ROP of p‐dioxanone initiated by a protected
monosaccharide (1,2;3,4‐di‐O‐isopropylidene‐α‐D‐galactopyranose) using
Al(O Pr)3 as the catalyst. The results of this study have been used as input for the
i
synthesis of the starch‐g‐PCL.
In Chapter 3, the synthesis of starch‐g‐PCL is reported. The method applied
basically consists of three steps (temporary partial protection of starch‐hydroxyl
16

Introduction
by trimethylsilyl groups, ROP of ε−CL on the remaining starch‐hydroxyl groups,
and removal of the silyl groups).
A preliminary study on the effect of several process variables (reactant ratio,
addition of co‐solvent, and application of different catalysts) on starch
esterification using vinyl fatty esters (vinyl laurate and vinyl stearate) is provided
in Chapter 4.
Based on these preliminary results, a systematic experimental study on the
effect of the process variables on the starch ester DS has been performed. The
results were quantified using a statistical model. The model, together with the
mechanical and thermal properties of the synthesized starch esters are reported in
Chapter 5.
The synthesis of two interesting compatibilizers (PCL‐g‐GMA and PCL‐g‐DEM)
for starch/PCL blends is discussed in Chapter 6. The use of these compatibilizers
in starch/PCL blends, including the mechanical and thermal properties, are also
reported.

1.5. References
[1]. E.S. Stevens: Green Plastics, An Introduction to the New Science of Biodegradable
Plastics, Princeton University Press, New Jersey, USA, 2002.
[2]. O.B. Wurzburg: Introduction, in Modified Starches: Properties and Uses (Ed.
O.B. Wurzburg), CRC Press, Inc., Boca Raton, USA, 1986.
[3]. A.H. Young: Fractionation of starch, in Starch: Chemistry and Technology (Ed.
R.L. Whistler, J.N. Bemiller, E.F. Passchall), Academic Press, Inc., London,
UK, 1984.
[4]. Website of International Starch Institute, Aarhus, Denmark:
http://www.starch.dk/isi/stat/rawmaterial.html, accessed on August 01, 2008.
[5]. R.M. de Bragança, P. Fowler: Industrial markets for starch, The Biocomposites
Centre, University of Wales, Bangor, Gywnedd, UK, 2004.
[6]. C. Nawrath, Y. Poirier, C. Somerville: Mini review: Plant polymers for
biodegradable plastics: cellulose, starch, and polyhydroxyalkanoates. Mol.
Breed. 1995, 1, 105‐122.
[7]. C.A. Holmes: Interactive European Network for Industrial Crops and their
Applications, Summary Report for the European Union, 2000‐2005, Agricultural
and Rural Strategy Group, Central Science Laboratory, Sand Hutton, York,
UK, 2005.
[8]. Corn, Part of A Global Economy, Corn Refiners Annual Report 2007, Corn
Refiners Association, Washington D.C., USA, 2007
17
Chapter 1
[9]. R.L. Whistler, J.R. Daniel: Starch, in Kirk Othmer Encyclopedia of Chemical
Technology Volume 22, Fifth Edition (Ed. A. Seidel), John Wiley & Sons, Inc.,
Hoboken, USA, 2005.
[10]. S.A. Watson: Corn and sorghum starches: production, in Starch: Chemistry
and Technology (Ed. R.L. Whistler, J.N. Bemiller, E.F. Passchall), Academic
Press, Inc., London, UK, 1984.
[11]. The compelling facts about plastics: An analysis of plastics production, demand and
recovery for 2006 in Europe, Plastics Europe, Association of Plastics
Manufacturers, Brussels, Belgium, 2008.
[12]. A.B. Strong: Plastics: Materials and Processing, Prentice‐Hall, Inc.,
Englewood Cliffs, New Jersey, USA, 1996.
[13]. R.J. Hernandez, S.E.M. Selke, J.D. Culter: Plastics Packaging, Properties,
Processing, Applications, and Regulations, Hanser Publishers, Munich,
Germany, 2000.
[14]. A.J. Weith: Plastics, Ind. Eng. Chem. 1939, 31, 557‐562.
[15]. G.W. Ehrenstein: Polymeric Materials: Structure, Properties, Applications,
Hanser Publishers, Munich, Germany, 2000.
[16]. J. Fritz, U. Link, R. Braun: Environmental impacts of biobased/ biodegradable
packaging, Starch/ Starke 2001, 53, 105‐109
[17]. X. Ren: Biodegradable plastics: a solution or challenge?, J. Cleaner Prod. 2003,
11, 27‐40.
[18]. Website of Environmental Protection Agency, Washington DC, USA:
http://www.epa.gov/garbage/plastic.htm, accessed on August 19, 2008
[19]. S.E.M. Selke: Packaging and the Environment, Technomic Publishing
Company, Inc., Lancaster, USA, 1990.
[20]. C. Bastioli: Global status of the production of biobased packaging materials,
Starch/ Starke 2001, 53 ,351‐355.
[21]. C. Bastioli: Global status of the production of biobased packaging materials,
the Food Biopack Conference Proceeding (Ed. C.J. Weber), Department of
Dairy and Food Science, The Royal Veterinary and Agricultural University,
Frederiksberg, Denmark, 2000.
[22]. R. Chandra, R. Rustgi: Biodegradable polymers, Prog. Polym. Sci. 1998, 23,
1273‐1335
[23]. O. Wolf, M. Crank, M. Patel, F. Marscheider‐Weidemann, J. Schleich, B.
Husing, G. Angerer: Technical Report Series, Techno‐economic Feasibility of
Large‐Scale Production of Bio‐based Polymers in Europe, Institute for Prospective
Technological Studies and European Science and Technology Observatory,
18

Introduction
European Commission Directorate General of Joint Research Centre, Seville,
Spain, 2005.
[24]. Petersen, K.; Nielsen, P.V. and Olsen, M.B. Physical and mechanical
properties of biobased materials – starch, polylactate and
polyhydroxybutyrate, Starch/ Starke 2001, 53, 356‐361
[25]. Press Release: Major Price Hikes in Plastics, A Real Case of “Force Majeure” that
should be passed through to Packaging Prices Immediately, Chambre Syndicale
des Emballages en Matiere Plastique (French Plastic and Flexible Packaging
Association), July 15, 2008.
[26]. Biobased packaging materials for the food industry, status and perspectives,
report of Food Biopack Project (Ed. C. J. Weber), Department of Dairy and
Food Science, the Royal Veterinary and Agricultural University,
Frederiksberg C., Denmark, 2000.
[27]. F.H. Otey, R.P. Westhoff, W.M. Doane: Starch‐based blown films, Ind. Eng.
Chem. Prod. Res. Dev. 1980, 4, 592‐595.
[28]. J. Yu, J. Gao, T.Lin: Note. Biodegradable thermoplastic starch. J. Appl. Polymer
Sci. 1996, 62, 1491‐1494
[29]. X. Wang, K. Yang, Y. Wang: Properties of starch blends with biodegradable
polymers. J. Macromol. Sci., Part C: Polym. Rev. 2003, 43, 385–409
[30]. J. Yu, N. Wang, X. Ma: The effects of citric acid on the properties of
thermoplastic starch plasticized by glycerol. Starch/Stärke. 2005, 57, 494–504
[31]. A.L.M. Smits, P.H. Kruiskamp, J.J.G. van Soest, J.F.G. Vliegenthart:
Interaction between dry starch and plasticisers glycerol or ethylene glycol,
measured by differential scanning calorimetry and solid state NMR
spectroscopy. Carbohydr. Polym. 2003, 53, 409–416
[32]. A.J.F. Carvalho, A.A.S. Curvelo, A. Gandini: Surface chemical modification
of thermoplastic starch: reactions with isocyanates, epoxy functions and
stearoyl chloride. Ind. Crops Prod. 2005, 21, 331–336
[33]. O. Martin, E. Schwach, L. Avérous, Y. and Couturier: Properties of
biodegradable multilayer films based on plasticized wheat starch. Starch/
Stärke 2001, 53, 372–380
[34]. K. Bagdi, P. Muller, B. and Pukanszky: Thermoplastic starch/layered silicate
composites: structure, interaction, properties. Compos. Interfaces 2006, 13, 1–17
[35]. P.M. Forssella, J.M. Mikkilti, G.K. Moates, R. Parker: Phase and glass
transition behaviour, of concentrated barley starch‐glycerol‐water mixtures, a
model for thermoplastic starch, Carbohydr. Polym. 1997, 34, 275‐282
19
Chapter 1
[36]. P. Matzinos, V. Tserki, A. Kontoyiannis, C. Panayiotou: Processing and
characterization of starch/polycaprolactone products. Polym. Degrad. Stab.
2002, 77, 17–24
[37]. F.J. Rodriguez‐Gonzalez, B.A. Ramsay, B.D. Favis: High performance LDPE/
thermoplastic starch blends: a sustainable alternative to pure polyethylene.
Polymer 2003, 44, 1517–1526
[38]. X. Qiao, W. Jiang, K. Sun: Reinforced thermoplastic acetylated starch with
layered silicates. Starch/Stärke 2005, 57, 581–586
[39]. M. Vikman, S.H.D. Hulleman, M. van der Zee, P. Myllarinen, H. Feil:
Morphology and enzymatic degradation of thermoplastic starch–
polycaprolactone blends. J. Appl. Polym. Sci. 1999, 74, 2594–2604
[40]. F.J. Rodriguez‐Gonzalez, B.A. Ramsay, B.D. Favis: Rheological and thermal
properties of thermoplastic with high glycerol content. Carbohydr. Polym.
2004, 58, 139–147
[41]. R. Gattin, A. Copinet, C. Bertrand, Y. Coutuier: Comparative biodegradation
study of starch‐ and polylactic acid‐based materials. J. Polym. Environ. 2001, 1,
11–17.
[42]. O.B. Wurzburg: Cross‐linked starches, in Modified Starches: Properties and Uses
(Ed. O.B. Wurzburg), CRC Press, Inc., Boca Raton, USA, 1986.
[43]. M.C. Silva, E.C. Ibezim, T.A.A. Ribeiro, C.W.P. Carvalho, C.T. Andrade:
Reactive processing and mechanical properties of cross‐linked maize starch.
Ind. Crops Prod. 2006, 24, 46–51
[44]. J.W. Mullen, E. Pacsu: Starch studies: Preparation and properties of starch
triesters, Ind. Eng. Chem. 1942, 34, 1209‐1217.
[45]. J.W. Mullen, E. Pacsu: Starch studies: Possible industrial utilization of starch
esters, Ind. Eng. Chem. 1943, 35, 381‐384
[46]. W. Jarowenko: Acetylated starch and miscellanous organic esters, chapter 4
in book Modified Starches: Properties and Uses, 1986, Wurzburg, O.B.
(editor), CRC Press, Inc., Boca Raton, USA, 55‐78.
[47]. T. Divers, I. Pilin, J.F. Feller, G. Levesque, Y. Grohens: Starch modifications,
destructuration and hydrolysis during O‐formylation. 2004, 56, 389‐398
[48]. A.D. Sagar, E.W. Merrill: Properties of fatty‐acid esters of starch. Journal of
Applied Polymer Science 1995, 58 (9), 1647‐1656.
[49]. S. Thiebaud, J. Aburto, I. Alric, E. Borredon, D. Bikiaris, J. Prinos, C.
Panayiotou: Properties of fatty‐acid esters of starch and their blends with
LDPE. J. Appl. Polym. Sci. 1997, 65, 705‐721.
20

Introduction
[50]. J. Aburto, S. Thiebaud, I. Alric, E. Borredon, D. Bikiaris, J. Prinos, C.
Panayiotou: Properties of octanoated starch and its blends with polyethylene.
Carbohydr. Polym. 1997, 34, 101‐112.
[51]. J. Aburto, H. Hamaili, G. Mouysset‐Baziard, F. Senocq, I. Alric, E. Borredon:
Free‐solvent synthesis and properties of higher fatty esters of starch ‐ Part 2.
Starch‐Starke 1999, 51, 302‐307.
[52]. J. Aburto, I. Alric, S. Thiebaud, E. Borredon, D. Bikiaris, J. Prinos, C.
Panayiotou: Synthesis, characterization, and biodegradability of fatty‐acid
esters of amylose and starch. J. Appl. Polym. Sci. 1999, 74, 1440‐1451.
[53]. J. Aburto, I. Alric, E. Borredon: Preparation of long‐chain esters of starch
using fatty acid chlorides in the absence of an organic solvent. Starch‐Starke
1999, 51, 132‐135.
[54]. C.E. Smith, J.V. Tuschoff: Acylation of hydroxy compounds with vinyl esters.
US Patent 2,928,828, 1960
[55]. R. A. de Graaf, G. A. Broekroelofs, L. P. B. M. Janssen, A. A. C. M.
Beenackers: The kinetics of the acetylation of gelatinised potato starch.
Carbohydr. Polym. 1995, 28, 137‐144.
[56]. R. A. de Graaf: Ph.D. Thesis, Rijksuniversiteit Groningen, 1996
[57]. W. Mormann, M. Al‐Higari: Acylation of starch with vinyl acetate in water.
Starch‐Starke 2004, 56, 118‐121.
[58]. R. A. Dicke, A straight way to regioselectively functionalized polysaccharide
esters. Cellulose 2004, 11, 255‐263.
[59]. J. Aburto, I. Alric, E. Borredon: Organic solvent‐free transesterification of
various starches with lauric acid methyl ester and triacyl glycerides. Starch‐
Starke 2005, 57, 145‐152.
[60]. M.M.Coleman, J.F.Graf, P.C.Painter: Specific Interactions and the Miscibility of
Polymer Blends, Technomic Publishing Company, Inc., Lancaster, USA, 1991.
[61]. R.J.Stokes, D.Fennel Evans: Fundamental of Interfacial Engineering, Wiley‐VCH,
New York, USA, 1997.
[62]. C.W. Macosko, P. Guegan, A.K. Khandpur, A. Nakayama, P. Marechal, T.
Inoue: Compatibilizers for melt blending: Premade block copolymers.
Macromolecules 1996, 29, 5590‐5598
[63]. W.E. Bakker and G.H. Hu, Introduction, in Reactive Polymer Blending (Ed. W.E.
Bakker, C.E. Scott, G.H. Hu), Hanser Publishers, Munich, Germany, 2001.
[64]. E. Duquesne, D. Rutot, P. Degee, P. Dubois: Synthesis and characterization of
compatibilized poly(ε‐caprolactone)/ granular starch composites. Macromol.
Symp. 2001, 175, 33‐43
21
Chapter 1
[65]. L. Contat‐Rodrigo, A. Ribes‐Greus: Thermal and viscoelastic properties of
some commercial starch products. J. Appl. Polym. Sci. 2003, 88, 1242‐1251
[66]. R. Nawang, I.D. Danjaji, U.S. Ishiaku, H. Ismail, Z.A. Mohd Ishak: Material
properties. Mechanical properties of sago starch‐filled linear low density
polyethylene (LLDPE) composites. Polym. Testing. 2001, 20, 167–172
[67]. I.D. Danjaji, R. Nawang, U.S. Ishiaku, H. Ismail, Z.A.M. Mohd Ishak:
Material behaviour. Degradation studies and moisture uptake of sago‐starch‐
filled linear low‐density polyethylene composites. Polym. Testing. 2002, 21,
75–81
[68]. R. L. Evangelista, Z.L. Nikolov, W. Sung, J. Jane, R.J. Gelina, Effect of
compounding and starch modification on properties of starch‐filled Low‐
Density Polyethylene, Ind. Eng. Chem. Res. 1991, 30, 1841‐1846
[69]. D. Bikiaris, C. Panayiotou: LDPE/Starch blends compatibilized with PE‐g‐
MA copolymers. J. Appl. Polym. Sci., 1998, 70, 1503–1521
[70]. D. Bikiaris, J. Prinos, K. Koutsopoulos, N. Vouroutzis, E. Pavlidou, N.
Frangis, C. Panayiotou: LDPE/plasticized starch blends containing PE‐g‐MA
copolymer as compatibilizer. Polvm. Degr. Stab. 1998, 59, 287‐291
[71]. A.G. Pedroso, D.S. Rosa: Mechanical, thermal and morphological
characterization of recycled LDPE/corn starch blends. Carbohydr. Polym. 2005,
59, 1‐9
[72]. A.G. Pedroso, D.S. Rosa: Effects of the compatibilizer PE‐g‐GMA on the
mechanical, thermal and morphological properties of virgin and reprocessed
LDPE/corn starch blends, Polym. Adv. Technol. 2005, 16, 310–317
[73]. W. Liu, Y.‐J. Wang, Z. Sun: Effects of polyethylene‐grafted maleic anhydride
(PE‐g‐MA) on thermal properties, morphology, and tensile properties of
Low‐Density Polyethylene (LDPE) and corn starch blends. J. Appl. Polym. Sci.
2003, 88, 2904–2911.
[74]. E.M. Nakamura, L. Cordi, G.S.G. Almeida, N. Duran, L.H.I. Mei: Study and
development of LDPE/starch partially biodegradable compounds. J. Mater.
Proc. Tech. 2005, 162–163, 236–241
[75]. R. Mani, M. Bhattacharya: Properties of injection moulded starch/ synthetic
polymer blends. III. Effect of amylopectin to amylose ratio in starch. Eur.
Polym. J. 1998, 34, 1467‐1475,
[76]. N. St‐Pierre, B.D. Favis, B.A. Ramsay, J.A. Ramsay, H. Verhoogt: Processing
and characterization of thermoplastic starch/polyethylene blends. Polymer,
1997, 38, 647‐655
22

Introduction
[77]. S. Garg, A.K. Jana: Studies on the properties and characteristics of starch–
LDPE blend films using cross‐linked, glycerol modified, cross‐linked and
glycerol modified starch. Eur. Polym. J. 2007, 43, 3976–3987
[78]. B. Raj, U. Sankar, Siddaramaiah. Low density polyethylene/ starch blend
films for food packaging applications. Adv. in Polym. Tech. 2004, 23, 32–45
[79]. T. Jana, S. Maiti: A biodegradable film, 1. Pilot plant investigation for
production of the biodegradable film. Ang. Makromol. Chem. 1999, 267, 16–19
[80]. F.H. Otey, R.P. Westhoff: Starch‐Based Films. Preliminary diffusion
evaluation. Ind. Eng. Chem. Prod. Res. Dev. 1984, 23, 284‐287
[81]. F.H. Otey, R.P. Westhoff, W.M. Doane: Starch‐based blown films. 2. Ind. Eng.
Chem. Res. 1987, 26, 1659‐1663
[82]. M.S. Lindblad, Y. Liu, A‐C. Albertsson, E. Ranucci, S. Karlsson: Polymers
from Renewable Resources. Adv. Polym. Sci. 2002, 157, 139‐161.
[83]. K.M. Stridsberg, M. Ryner, A.‐C. Albertsson: Controlled Ring‐Opening
Polymerization: Polymers with designed Macromolecular Architecture. Adv.
Polym. Sci. 2002, 157, 42‐65.
[84]. O. S. Odusanya, D. M. A. Manan, U. S. Ishiaku, B. M. N. Azemi: Effect of
starch predrying on the mechanical properties of starch/poly(e‐caprolactone)
composites. J. Appl. Polym. Sci., 2003, 87, 877–884.
[85]. D. dos Santos Rosa, T.C. Rodrigues, C.G.F. Guedes, M.R. Calil: Effect of
thermal aging on the biodegradation of PCL, PHB‐V, and their blends with
starch in soil compost, J. Appl. Polym. Sci., 2003, 89, 3539–3546
[86]. O.S. Odusanya, U.S. Ishiaku, B.M.N. Azemi, B.D.M. Manan, H.W. Kammer:
On mechanical properties of sago starch/poly(ε‐caprolactone) composites,
Polym. Eng. Sci. 2000, 40, 1298‐1305.
[87]. L. Averous, L. Moro, P. Dole, C. Fringant: Properties of thermoplastic blends:
starch–polycaprolactone, Polymer 2000, 41, 4157–4167
[88]. A. Sen, M. Bhattacharya Residual stresses and density gradient in injection
molded starch/synthetic polymer blends Polymer 2000, 41, 9177–9190
[89]. C.H. Kim, K.M. Jung, J.S. Kim. J.K. Park: Modification of aliphatic polyesters
and their reactive blends with starch. J. Polym. Environ. 2004, 12, 179‐187.
[90]. C.H. Kim, K.Y. Cho, J.K. Park: Reactive blends of gelatinized starch and
polycaprolactone‐g‐glycidyl methacrylate. J. Appl. Polym. Sci. 2001, 81, 1507–
1516.
[91]. J. John, J. Tang, Z. Yang, M. Bhattacharya: Synthesis and characterization of
anhydride‐functional polycaprolactone, J. Polym. Sci., Part A: Polym. Chem.
1997, 35, 1139–1148.
23
Chapter 1
[92]. E. Duquesne, D. Rutot, P. Degee, P. Dubois: Synthesis and characterization of
compatibilized poly (e‐caprolactone)/ granular starch composites, Macromol.
Symp. 2001, 175, 33–43
[93]. R. Mani, J. Tang, M. Bhattacharya: Synthesis and characterization of starch‐
graft‐polycaprolactone as compatibilizer for starch/polycaprolactone blends.
Macromol. Rapid Commun. 1998, 19, 283–286.
[94]. M. Avella, M.E. Errico, R. Rimedio, P. Sadocco: Preparation of biodegradable
polyesters/high‐amylose‐starch composites by reactive blending and their
characterization, J. Appl. Polym. Sci. 2002, 83, 1432–1442.
[95]. M. Avella, M.E. Errico, P. Laurienzo, E. Martuscelli, M. Raimo, R. Rimedio:
Polymer communication: Preparation and characterisation of compatibilised
polycaprolactone/ starch composites. Polymer 2000, 41, 3875–3881
[96]. C.S. Wu: Physical properties and biodegradability of maleated‐
polycaprolactone/ starch composite. Polym. Degrad. Stab. 2003, 80, 127–134.
[97]. P. Dubois, M. Krishnan, R. Narayan: Aliphatic polyester‐grafted starch‐like
polysaccharides by ring‐opening polymerization. Polymer 1999, 40, 3091‐3100.
[98]. D. Rutot, P. Degée, R. Narayan, and P. Dubois: Aliphatic polyester‐grafted
starch composites by in situ ring opening polymerization. Composite
Interfaces, 2000, 7, 215–225.

24
Chapter 2
Experimental Studies on the Ring Opening
Polymerization of p‐dioxanone using an
Al(OiPr)3‐Monosaccharide Initiator System

Abstract
The ring opening polymerization (ROP) of p‐dioxanone using a protected
monosaccharide (1,2;3,4‐di‐O‐isopropylidene‐α‐D‐galactopyranose)/Al(OiPr)3
initiator system to yield polydioxanone with a protected monosaccharide end‐
group is described. The products were synthesized at 60‐100°C and characterized
by 1H‐ and 13C‐NMR, and MALDI‐TOF mass spectrometry. Besides the desired
polydioxanone functionalised with a monosaccharide end‐group, also
polydioxanone with an OiPr end‐group was formed (20‐30 %). Systematic studies
showed that the polymer yield is a function of the reaction temperature and the
reaction time, with higher temperatures (100°C) leading to lower yields. The
average chain length of the polymers is between 7 and 58 repeating units and may
be tuned by the monomer to monosaccharide ratio (at constant Al(OiPr)3 intake).
A statistical model has been developed that successfully describes the
experimentally observed relation between the average chain length of the
functionalized polymer and reaction parameters.

Keywords: biodegradable, polyesters, ring‐opening polymerization

Chapter 2
2.1. Introduction
Aliphatic polyesters, such as polycaprolactone, polyglycolide, and polylactides,
are interesting polymers because of their good product performance and
biodegradability [1]. Polydioxanone (poly(p‐dioxanone) or poly(1,4‐dioxan‐2‐
one)), accessible by the polymerization of p‐dioxanone (1), has interesting product
properties compared to other aliphatic polyesters. Its melting temperature is close
to 110 oC, which is a unique compromise between application and processing
temperature. This melting point is considerably higher than typically found for
polycaprolactone (60 oC) and lower than that of polylactides (at least 175°C). The
relatively low melting point of polycaprolactone limits its applicability, whereas
the high temperature for polylactides results in thermal degradation and
undesirable transfer reactions during synthesis and processing steps [2].
Polydioxanone has a tensile strength close to 48.3 MPa and an elongation at break
of 500‐600%, and is tougher than polylactides and even HDPE [3]. From a
biodegradability point of view, polydioxanone also shows good performance. It is
fully degraded in the body within a period of 180 days [4]. Nishida et al [5]
reported that polydioxanone decomposes to non‐toxic gases (CO2 and H2O) by
microorganisms. Despite its good properties, only limited information about the
synthesis and properties of polydioxanone is available in the open literature,
probably because the p‐dioxanone monomer has become commercially available
only recently [1].
Biodegradable aliphatic polyesters end‐capped with sugar molecules have been
studied extensively for use in biomedical applications, particularly for nano‐
encapsulation systems for drug delivery [6]. The synthesis of protected
monosaccharide end‐capped biodegradable polymers is usually performed via
Ring Opening Polymerization (ROP). The catalysts are metal alkoxides with Lewis
acidic character [8, 9]. The ring‐opening polymerization (ROP) of p‐dioxanone
using metal catalysts such as aluminum isopropoxide [Al(OiPr)3], stannous octoate
[Sn(Oct)2], or zinc lactate has been reported. The alkoxide group will end up as an
ester end‐group in the polymer and in this way at least one of the end‐groups may
be easily controlled and varied. Exchange of the alkoxide group by e.g. reaction of
the metal‐alkoxide with an appropriate alcohol allows the synthesis of end‐capped
poly‐lactones. Several polymers with bioactive alcoholic and phenolic end‐groups
of interest for drug‐related applications were synthesized (for example geraniol,
quinine, tocopherol, testosterone, pregnenolone, stigmasterol and ergocalciferol)
[10] and also with protected monosaccharides (galactopyranose/ glucofuranose)
[6‐7].
This chapter describes experimental studies on the catalytic ROP of p‐
dioxanone in the presence of a protected glucose molecule (1,2;3,4‐di‐O‐
isopropylidene‐α‐D‐galactopyranose, 2), which to the best of our knowledge is the
first study to functionalise polydioxanone with a monosaccharide. Besides
26

Ring Opening Polymerization of p‐dioxanone using an Al(OiPr)3‐Monosaccharide Initiator System
potential applications in biomedical products, the results of this study are also of
interest for the preparation of starch/polydioxanone polymers using ROP.
Particularly interesting in this field are starch polymers grafted with
polydioxanone. However, grafting efficiencies are difficult to determine in this
system by standard analytical techniques (e.g. NMR) due to the poor solubility of
the products in standard organic solvents. As such, the synthetic pathways and
the soluble, relatively low molecular weight compounds reported in this study
may be viewed as model systems for more complex, poorly soluble heterogeneous
systems.

2.2. Materials and Methods
2.2.1. Materials
p‐Dioxanone monomer (1, Boehringer Ingelheim, Germany) was purified
according to the procedure described by Raquez et al [2, 3]. Toluene (Labscan) was
dried and stored on molecular sieves 3 Å (Labscan) under nitrogen. 1,2;3,4‐di‐O‐
isopropylidene‐α‐D‐galactopyranose (2), 97% (Sigma) and aluminum
isopropoxide, 98+% (Aldrich) were used as received. Analytical grade
dichloroethane (Labscan), heptane (Acros), and diethylether (Labscan) were used
as received. CDCl3 was obtained from Sigma and was used as received.

2.2.2. Methods
All polymerization experiments were carried out under a protective nitrogen
atmosphere using standard Schlenk‐ and glovebox techniques.

2.2.2.1. Typical example for the synthesis of polydioxanone end‐capped with 1,2;3,4‐di‐O‐
isopropylidene‐α‐D‐galactopyranose (2)
2 (1.44 g, 5.5 mmol) was dissolved in toluene (1 ml) at 50 oC. To this solution, 0.8
ml of a solution of an aluminum isopropoxide stock solution was added. This
stock solution was prepared by adding 4.07 g, (20 mmol) of aluminum
isopropoxide to 20 ml of toluene. The resulting clear solution was stirred for 2 h at
50 oC. Subsequently, part of this solution (350 µl, containing 0.55 mmol of 2 and
0.08 mmol aluminum isopropoxide) was added to pure 1 (0.8 g, 8 mmol), which
was pre‐heated till about 60°C to obtain it in a liquid state. The polymerization
was allowed to proceed for 16 h at 100 oC. The reaction mixture was clear and
colorless during the reaction. After the pre‐determined reaction time, the mixture
was brought to room temperature and several drops of HCl (1 N) were added to
stop the reaction. Next, hot dichloroethane (20‐25 ml) was added to completely
dissolve the partly solid polymer at room temperature. The hot solution was
27
Chapter 2
precipitated in a heptane/ ether mixture (300‐400 ml, 4:1 by volume) at 4‐8 oC. The
white solid was finally separated from the liquid by decantation and dried in a
vacuum oven (5 mbar, 40 oC) until constant weight. The isolated yield at this
condition was 68%.

2.2.3. Product Analyses
NMR analyses were performed in CDCl3. 1H‐ and 13C‐NMR spectra were
recorded on a Varian AMX 400 NMR. 2D‐NMR spectra were recorded on a Varian
Unity 500 NMR. Processing of the raw data was performed using VNMR software.
MALDI‐TOF spectra were recorded on an Applied Biosystems Voyager DE‐PRO
machine using dithranol/ NaI as the matrix (linear mode).

2.2.4. Calculation of Average Degree of Polymerization
The Theoretical Average Degree of Polymerization, X n theo in terms of number of
monomer units is calculated as follows [6]:
theo [monomer ]0
Xn = monomer conversion ×
[total initiator ]
(2.1.)
[monomer ]0
= monomer conversion ×
3[ Al (O i Pr) 3 ] + [ sugar ]0
Here, it is assumed that all available initiator is used effectively. If the amount
of sugar is in excess with respect to the aluminum catalyst, the above equation
simplifies to
theo [monomer ]0
Xn = monomer conversion × (2.2.)
[ sugar ]0
H‐NMR was applied to determine the experimental average degree of
1
polymerization, X n exp of the product.

X n exp is
calculated by comparing the peak area of characteristic end‐group
protons with that of a proton of the repeating unit in the polymer (A H‐polymer). As
will be shown later, two polymers with different end‐groups are present, one
initiated on a galactopyranose molecule and the other on an isopropoxide group.
This leads to the following equation:
exp A H-repeating unit

Xn = (2.3.)
A H-galactopyranose end group +A H-isopropoxide end group
28

The NMR peaks of the repeating unit overlap partially with those of the
galactopyranose end group. To compensate for this effect, eq. 2.3. is rewritten as:
exp [ A 3.4− 4.7 ppm − 6 × A 5.6 ppm ] / 6

Xn = (2.4.)
A 5.6 ppm + A 5.1 ppm
The ratio of the two different types of polymers (either end‐capped with
galactopyranose or an isopropoxide group), R gp , is calculated using:
A H − galactopyranose end group

R gp = (2.5.)
A H − isopropoxide end group
or, in term of the NMR resonances :
A 5.6 ppm
R gp = (2.6.)
A 5.1 ppm
2.3.1. Screening Experiments
Initial experiments to synthesize protected galactopyranose end‐capped
polydioxanone (3) were performed using 1,2;3,4‐di‐O‐isopropylidene‐α‐D‐
galactopyranose (2) and Al(OiPr)3 as the catalyst precursor (Scheme 2.1.).
Typically, an Al ‐ monosaccharide 2 ‐ monomer 1 mol ratio of 1: 6.6: 96 was
applied. The polymerization reaction was performed as a two step process. In the
first step, the protected mono‐saccharide 2 reacts with Al(OiPr)3 to form the actual
catalyst for the polymerization reaction. The exchange reaction typically takes
place at 50°C for 2 h. To avoid the formation of isopropoxide end groups, an
excess of monosaccharide 2 on Al was used (see eq. 2.7.).

O α γ O γ'
β α' β'
CH2OH O H2C O [ C CH2 O CH2 CH2 O ]n C CH2 O CH2 CH2 OH
6 6
O O
O H O H
H5 α O Al(OiPr)3 H5
4 1 4 1
H
O H
O + n+1 H
O H
O
2 O γ catalyst 2
3 3
H O H O
β
(2) (1) (3)

Scheme 2.1. Schematic representation of the polymerization reaction including

atom numbering scheme
29
Chapter 2
H O
H O
O H
H
O H
CH2OH O
iPr O O H2C O
O H
H
Al O iPr + 3 Al O CH2 + 3 iPr OH
O H
iPr O H O
H2C O
H O O
O H
H
O
O H O H
H H O
O H
H O H O
H O
(2.7.)
In the next step, the in‐situ formed catalyst was reacted with p‐dioxanone
monomer 1 at 100 oC for 16 h. The off‐white solid reaction product was collected
after a dissolution‐reprecipitation process using dichloroethane and a heptane/
diethyl ether mixture. Typical isolated product yields are 68% at these conditions.

2.3.1.1. Product analyses
The products were analyzed using 1H‐ and 13C‐ NMR and MALDI‐TOF. Typical
1
H‐ and 13C‐NMR spectra of 3 are shown in Figure 2.1 and 2.2., respectively.
1
H‐NMR spectra (Figure 2.1.) are not particularly informative, although it is
evident that p‐dioxanone polymerisation occurred. The typical proton resonances
of the p‐dioxanone (3.78 to 4.40 ppm) unit are broadened and shifted up to 0.1
ppm compared to the monomer. It is difficult to determine the end‐groups of the
polymer on the basis of 1H‐NMR. Although the spectra clearly indicate the
presence of the monosaccharide 2, it is not possible to determine whether this is
truly an end‐group due to overlapping peaks with protons from the poly(p‐
dioxanone) backbone. However, of interest is the presence of a small multiplet at
about 5.1 ppm. This multiplet is characteristic for the CH proton of an
isopropoxide end‐group. It confirms that polymer initiation not only occurs with
the monosaccharide but also with the remaining isopropoxide group of the
catalyst precursor (vide infra). Proton resonances of the CH3 group of the
isopropoxide end‐group, together with the CH3 groups from the protecting groups
of the sugar appear in the range 1.15‐1.51 ppm. The OiPr : 2 end‐group ratio for the
standard experiment was 1: 2.67.

30

Figure 2.1. 1
H‐NMR Spectra of:
(a). 1,2;3,4‐di‐O‐isopropylidene‐α‐D‐galactopyranose, 2
(b). p‐dioxanone monomer, 1
(c). 1,2;3,4‐di‐O‐isopropylidene‐α‐D‐galactopyranose‐end‐capped
polydioxanone, 3

C‐NMR (Figure 2.2.) is more informative and clearly shows the presence of a
13
polydioxanone polymer backbone and a monosaccharide end group. The carbon

resonances of the polydioxanone backbone are present at δ = 63.8, 68.17 and 69.2
ppm. Carbon resonances arising from the monosaccharide end‐group are present
between δ = 66.2 and 96.2 ppm. Particularly the C2‐C6 carbons in the range δ =
66.2‐70.3 ppm are shifted considerably. For instance, C‐6 is shifted from δ = 62.28
ppm in 2 to δ 66.2 ppm in product 3. In addition to the mono‐saccharide end‐
group, characteristic resonances of an ‐O‐C(=O)‐CH2‐O‐CH2‐CH2‐OH end group
are present at δ = 61.61 (γ’) and 73.52 ppm (β’).

31
Chapter 2
Figure 2.2. 13
C‐ NMR Spectra of:
(a). 1,2;3,4‐di‐O‐isopropylidene‐α‐D‐galactopyranose, 2
(b). p‐dioxanone monomer, 1
(c). 1,2;3,4‐di‐O‐isopropylidene‐α‐D‐galactopyranose‐end‐capped
polydioxanone, 3

Carbon resonances of the isopropoxide groups from the by‐product
(isopropoxide end‐capped polydioxanone) are present at δ = 68.48 ppm (‐CH‐
(CH3)2) and between δ = 21.76‐25.97 ppm (‐CH‐(CH3)2). The resonances of the
protecting group of the sugar appears at δ = 108.73 and 109.70 (>C‐(CH3)2) and
between δ = 21.76‐25.97 (>C‐(CH3)2).
D‐NMR (HSQC) was applied for complete peak assignment of the product. A
2
typical example of a part of the 2D‐NMR spectra is given in Figure 2.3. An
overview of the data is given in Tables 2.1. and 2.2., the numbering scheme of
carbons and protons is given in Scheme 2.1.

32

Figure 2.3. HSQC 13C‐1H spectra of 3

Table 2.1. 1
H‐NMR peak assignments a
Reactant peaks (ppm) Product peaks (ppm)
Galactopyranose p‐dioxanone Poly(p‐dioxanone) end‐capped

Monomer with protected galactopyranose
(2) (1) (3)
H‐1 5.54 H‐1 5.50
H‐2 4.31 H‐2 4.32 *
H‐3 4.59 H‐3 4.60

H‐4 4.25 H‐4 4.09
H‐5 3.74 H‐5 4.16*
H‐6 3.83 H‐6 4.00
H‐α 4.26 H‐α 4.17
H‐β 3.77
H‐β 3.78
H‐β’ 3.68

H‐γ 4.301
H‐γ 4.40
H‐γ’ 3.732
H‐OH broad, around 1.8
a
all values were determined using 1H NMR, except values with *, which were determined from HSQC
spectrum due to overlapping resonances.
33
Chapter 2
Table 2.2. 13
C‐NMR peak assignments
Reactant Peaks (ppm) Product Peaks (ppm)
Galactopyranose p‐dioxanone Poly(p‐dioxanone) end‐capped

monomer with protected galactopyranose
(2) (1) (3)
C‐1 96.28 C‐1 96.21
C‐2 70.75 C‐2 70.31
C‐3 70.58 C‐3 70.63

C‐4 68.10 C‐4 68.33
C‐5 71.57 C‐5 70.92
C‐6 62.28 C‐6 66.2
C‐α 61.92 C‐α 68.17
C‐β 69.2
C‐β 68.09
C‐β’ 73.52
C‐γ 63.80
C‐γ 65.54
C‐γ’ 61.61
C‐carbonyl 170.00
C‐carbonyl 166.21
C‐carbonyl‐iPr/gal 171.11/ 170.07

NMR analyses also allow calculation of the molecular weight of the products.
For this purpose, the ratio of the intensity of the end groups and the polymer
backbone peaks is determined. A detailed procedure is given in the experimental
section. The product obtained at screening conditions (16 h reaction time at 100°C)
contains on average 13 monomer units, corresponding with an average number
molecular weight (Mn) of about 1600.
MALDI‐TOF was also applied to characterize the products. An example of a
MALDI‐TOF spectrum of 3 recorded in a dithranol/NaI matrix is given in Figure
2.4. A typical molecular weight distribution is observed. The difference in
molecular weight between the main peaks is 102 g/mol, which is the molecular
weight of a repeating dioxanone unit. The molecular weight distribution of the
major peaks may be represented by the following relation:
m / z = 23 + 260 + 102n (2.8.)
This series represents a dioxanone polymer end capped with 2 and an additional
Na ion. The latter likely stems from the matrix used to ionize the sample.
34

100 1098.5
90 +102
80
70
60
%-Intensity
50
40
30
20
10
0
600 800 1000 1200 1400 1600 1800 2000 2200 2400
Mass (m/z)
Figure 2.4. Typical MALDI‐TOF spectra of 3
A second molecular weight distribution is also clearly visible when enlarging
the spectra, see Figure 2.5. for details. This series may be described by the
following relation:
m / z = 23 + 60 + 102n (2.9.)
This relation is indicative for the presence of a polydioxanone polymer
containing an isopropoxide end‐group, in line with the NMR data.
Furthermore, two other distributions are present, although both with a very
low intensity (Figure 2.5.). These distributions may be represented by equation
(2.10.) and (2.11.) and imply the presence of dioxanone polymers with carboxylic
end groups, ionized with either Na+ or H+.
m / z = 23 + 18 + 102n (2.10.)
m / z = 1 + 18 + 102n (2.11.)
35
Chapter 2
50
792.7
40
796.7
30
%-Intensity
20
858.3
10
836.1
0
600 700 800 900 1000 1100 1200
Mass (m/z)
Figure 2.5. Enlarged MALDI‐TOF spectra for 3

Although MALDI‐TOF clearly demonstrated the presence of various types of
end‐groups in the product, it proved not suitable to determine the average
molecular weight of the products. Various samples with, according to NMR,
different molecular weights were analysed. The observed differences in the
molecular weight distributions of the various samples were only marginal. Most
likely the matrix (dithranol/NaI), although successfully applied for
galactopyranose‐end‐capped polycaprolactone [6], is not particularly suitable for
polydioxanone. Various other matrices were tested (e.g. 2‐(4′‐
hydroxybenzeneazo)benzoic acid (HABA)), but in all cases poor quality, low
resolution spectra were obtained.

2.3.1.2. Mechanistic aspects
Both NMR and MALDI‐TOF measurements imply that the main product is
indeed the desired monosaccharide end‐capped polydioxanone. In addition, small
but significant amounts (20‐30%‐mol) of polydioxanone chains with an OiPr end
group are present. A mechanistic proposal for the ROP of p‐dioxanone with
Al(OiPr)3 as the catalyst precursor leading to the desired monosaccharide end‐
capped polydioxanone is given in Figure 2.6. [6,13]. In the first step, the catalyst
precursor is treated with monosaccharide 2 resulting in an alcohol exchange
36

reaction and the formation of the desired active catalyst with preferentially all
three OiPr exchanged with 2. Next, a p‐dioxanone molecule will coordinate to the
Lewis acidic aluminium center followed by an insertion step. Subsequent
coordination and insertion of dioxanone molecules leads to the formation of a
polymer chain with a monosaccharide end‐group. During the reaction,
termination of the chain growth may occur by reaction with an alcohol. The
resulting Al‐alkoxide may again initiate a polymerisation reaction. The
termination reaction is known to be reversible and the formed polymer may again
react with an aluminum center and continue to grow [18]. Irreversible termination
of the polymerisation is performed at the end of reaction period by adding dilute
acid to the polymerisation mixture. The termination reactions lead to the
formation of ‐O‐C(=O)‐CH2‐O‐CH2‐CH2‐OH end groups.

iPr O
Al O iPr + R OH (in excess)
iPr O
O O
Al O R + iPr OH Al O iPr
O O
+ +
O O
O O
n l
O O
O O
Al [O CH2 CH2 O CH2 C ] OR
n
O
O
O
+ m
O
O O
R-OH
Al [O CH2 CH2 O CH2 C ] OR
n+m
O
+ HX (termination)
O O
Al O R Al X
O O
+ +
O O O
H [O CH2 CH2 O CH2 C ]n OR H [O CH2 CH2 O CH2 C ] n+m OR H [O CH2 CH2 O CH2 C ] l O iPr
product by-product

Figure 2.6. Simplified reaction scheme for the ROP of p‐dioxanone catalyzed by
Al(OiPr)3. (R = monosaccharide 2)
37
Chapter 2
The minor product, OiPr‐end‐capped poly(p‐dioxanone), will be formed when

the polymerization starts with an aluminum alkoxide with a remaining Al‐OiPr
group (eq. 2.7.). These may be present in the reaction mixture because the
exchange reaction between Al(OiPr)3 and 2 was incomplete, despite the excess of 2.
However, the end group may also be formed by a termination reaction with free
isopropanol, formed in the first step of the polymerisation reaction (eq. 2.7.).

2.3.2. Systematic Studies
The effect of important process variables (temperature, time and the mol ratio
of monomer to monosaccharide) on the yield, degree of polymerisation of the
product and the end group distribution was determined. A total of 15 experiments
were performed at two polymerization temperatures (60 and 100 oC), two different
reaction times (1.5 hrs and 16 hrs) and a p‐dioxanone to monosaccharide 2 molar
ratio ranging between 3.3 and 62.5. An overview of the experiments and the
results are given in Table 2.3. In all cases, a (nearly) fixed Al(OiPr)3 :
monosaccharide ratio of 1 : 6.3‐6.6 was applied.

Table 2.3. Overview of experiments a
Processing Condition Product Properties
Set dioxanone/ Avg. Chain 2/OiPr ratio

2/Al ratio t T Yieldb
Sample 2 ratio Length Rgp
(mol/mol) (h) (°C) exp (%)
(mol/mol) (Xn ) (mol/mol)
S111 6.6 16 100 3.29 7.22 30.5 2.8
S112 6.6 16 100 14.45 13.10 67.6 2.7
1 S113 6.6 16 100 19.93 15.28 81.6 2.4
S114 6.6 16 100 37.53 33.21 80.5 3.1
S115 6.6 16 100 59.78 54.49 81.3 3.1
S211 6.3 16 60 10.22 13.74 81.6 2.5
S212 6.3 16 60 16.74 18.50 86.0 2.4
2 S213 6.3 16 60 23.12 25.16 87.6 2.6
S214 6.3 16 60 42.70 46.74 91.6 2.4
S215 6.3 16 60 58.93 52.85 96.2 2.8
S221 6.3 1.5 60 8.81 14.36 91.5 3.3
S222 6.6 1.5 60 16.11 17.04 92.5 3.1
3 S223 6.6 1.5 60 22.50 19.08 84.1 2.8
S224 6.6 1.5 60 35.30 30.22 86.6 2.9
S225 6.6 1.5 60 62.55 57.96 89.4 2.9
a
For each experiment, the exchange reaction between the catalyst and the protected sugar was performed
for 2 hours at 50°C
b
The yield is the isolated yield of product 3
38

2.3.2.1. Product yield
The isolated yields of the reactions are all but one between 67 and 96% (see
Table 2.3.). One of the experiments (S111) resulted in a very low yield (31%)
compared to the other reaction. For this particular experiment, a low dioxanone:
monosaccharide 2 ratio was applied, leading to a low average chain length of the
dioxanone polymer (7.2, see Table 2.3.). It is likely that this relatively low
molecular weight compound dissolves partly during isolation/ purification
procedure leading to lower isolated yields.
The effect of the reaction temperature (60 and 100°C) on the product yield at
three different dioxanone: 2 ratios is given in Table 2.3. Clearly, the yields for
reactions conducted at 60 oC are always higher than those performed at 100 oC.
The effect of temperature on the yield of the bulk polymerization of dioxanone
using Sn(Oct)2 and Al(OiPr)3 as catalysts (without the use of a second alcohol) has
been studied by Nishida et al [5] and Esteves, et al [11]. Higher polymerisation
yields were obtained at lower temperatures. For instance, the equilibrium
conversion of dioxanone at 80oC was about 80%, which reduced to 75% when
increasing the temperature to 120oC. These results were explained by assuming
that the reaction is an equilibrium polymerisation and that the equilibrium is
shifted to the monomer side at higher temperatures. The latter is due to the slight
exothermicity of the reaction [14], with an enthalpy of polymerization of
approximately ‐15 kJ/ mol [5]. Another possibility is the occurrence of polymer
crystallization, which is expected to be more pronounced at lower temperatures.
On the basis of our experimental data and in line with literature data, we conclude
that equilibrium conversion is achieved after 16 h and that the lower polymer
yields at higher temperatures are due to the slight exothermicity of the reaction.
The effect of the reaction time on the product yield may be derived from the
data provided in Table 2.3. and particularly when comparing the data in set 2 and
3 (60oC, 1.5 and 16 h reaction time). For the two experiments with a dioxanone: 2
ratio higher than 16 (≅ 23 and 60), the yields are higher when performing the
reaction at 16 h reaction times. Evidently, equilibrium conversion is not yet
achieved within 1.5 h. However, when the reaction is performed at a low
dioxanone: 2 ratio (≅ 16), the yield at 1.5 h is higher than the yield at 16 h. Similar
observations were made by Raquez et al [3] and Kricheldorf et al [1] for dioxanone
polymerisations using Al(OiPr)3 in the absence of a external alcohol or using
benzyl alcohol. It was shown that at lower monomer to catalyst ratios, the
monomer conversion reaches a maximum value before going down to the
equilibrium monomer conversion. To the best of our knowledge, no detailed
explanation has been put forward to explain this anomalous behaviour. A more
detailed analysis on the actual nature and composition of the polymerization
products, as suggested by Raquez et al [3], will be required.
39
Chapter 2
2.3.2.2. Effects of process conditions on the average chain length ( X n exp ) and end group

distribution
The effects of the p‐dioxanone: 2 ratio, reaction time and temperature on the
X n exp of the products is shown in Figure 2.7. The X n exp increases linearly with
respect to the dioxanone : 2 ratio, as expected for a typical ring opening
polymerisation [18]. In Figure 2.7., the X n theo (eq. 2.2.) as a function of the
dioxanone : 2 ratio at 90% and 100% monomer conversion is also provided. Most
experimental points are scattered along these lines, in line with the theoretical
predictions.

60 Temperature: 100oC, time: 16 hr
Temperature: 60oC, time: 16 hr
50 Temperature: 60oC, time: 1.5 hr

theo
Xn (90%-conversion)
theo
40 Xn (100%-conversion)
Xn exp
30
20
10
0
0 10 20 30 40 50 60 70
dioxanone : 2 ratio (mol/mol)

Figure 2.7. Average Chain Length ( X n exp ) of the product as a function of the
dioxanone : 2 mol ratio

The end group distribution was determined using NMR and is expressed in
terms of R gp (eq. 2.5. and 2.6.). The R gp values (Table 2.3.) for all experiments are
scattered randomly between 2.4 and 3.3. A clear trend between R gp and the
process conditions (temperature, time, and dioxanone/ 2 mol ratio) is absent. A
possible strategy to increase the Rgp values i.e. to reduce the number of OiPr
endgroups in the product may be the removal of isopropyl alcohol formed in step
40

1 of the polymerisation process (eq. 2.7.) from the reaction mixture by e.g. vacuum
distillation before adding the dioxanone monomer [6].

2.3.2.3. Statistical Data Analysis
Quantification of the influence of the experimental factors (temperature, time
and dioxanone/ 2 ratio) on the X n exp has been performed by multivariable linear
regression [12] on the data given in Table 2.3. The 2/ Al ratio was not included in
the model as the experimental range (6.3‐6.6) was too limited to draw sound
conclusions. A linear model proved adequate to describe the effects of the
independent variables on the X n exp :
exp
Xn = 8.49 + 0.854 × (dioxanone : 2 ratio) + 0.177 × (t ) − 0.097 × (T ) (2.12.)
where t and T are respectively the polymerization time and temperature.
The analysis of variance for the model is given in Table 2.4. The low P‐value
clearly indicates that the model is statistically significant. The R2 value for the
model is 0.977 (with an adjusted R2 value of 0.974), indicating that the model
describes the experimental data reasonably well. A parity plot of the modeled
versus experimental values of the average chain length Xn confirms this statement
(Figure 2.8.).

exp
Table 2.4. Analysis of variance (ANOVA) for linear model of X n as a
function of experimental parameters
SS DF MS F P‐value

Model 4103 3 1337 173.543 3.459*10‐7
Error 93 12 7.703
Total 4010 15
41
Chapter 2
60
50
40
Xn model
30
20
10
0
0 10 20 30 40 50 60
exp
Xn

Figure 2.8. Modeled versus experimental values for the average chain length Xn
The model predicts that the X n is a clear function of the p‐dioxanone : 2 mol
ratio, with high ratios leading to a higher average chain length. In addition, the
model predicts that the X n exp is positively influenced by the polymerization time,
which is in agreement with the available data on ring opening polymerization [15,
16, 17]. Furthermore, the X n exp is negatively influenced by temperature. This is in
line with literature data [5, 11] and due to the fact that the reaction is an
equilibrium polymerization with a slight exothermic effect. Within the
experimental ranges, the model allows determination of the process variables to
obtain a product with the desired degree of polymerization.

2.4. Conclusions
The ROP of p‐dioxanone in the presence of a monosaccharide (1,2;3,4‐di‐O‐
isopropylidene‐α‐D‐galactopyranose, 2) with Al(OiPr)3 as the catalyst is reported.
The isolated yields of the off‐white solid products were between 30 and 96%.
Molecular weights (NMR) of the product were between 970 and 6200 and are a
clear function of the p‐dioxanone/2 ratio (at constant Al(OiPr)3 intake), with higher
ratios leading to higher molecular weights. Both NMR and MALDI‐TOF
measurements indicate that the products are mixtures of polymers and significant
42

amounts of p‐dioxanone polymers with an isopropoxide end group (20‐30%) were
present. A statistical model has been developed to quantify the effects of process
variables (time, temperature and monomer: monosaccharide ratio) on the X n exp .
Moreover, the findings of this study have proven to be valuable input for synthetic
studies on the preparation of novel biodegradable polymers consisting of
polydioxanones and polycaprolactones grafted on oligo‐ and polysaccharides (e.g.
starch). These studies will be reported in the next chapter.

2.5. Nomenclature
AH : peak area of certain proton in 1H‐NMR spectra [‐]
A x − y ppm : peak area of certain peak at x until y ppm in 1H‐NMR spectra [‐]
R gp : mol ratio of galactopyranose end‐capped polydioxanone and
isopropoxide‐end‐capped polydioxanone [‐]
n : number of monomer unit in the polymer products [‐]
t : time [hour]
T : temperature [oC]
X n exp : experimental average degree of polymerization the polymer products
[monomer units]
X n theo : theoretical average degree of polymerization of the polymer products
[monomer units]

2.6. References
[1]. H.R. Kricheldorf, D.O. Damrau: Polylactones, 42. Zn‐lactate‐catalyzed
polymerizations of 1,4‐dioxan‐2‐one. Macromol. Chem. Phys. 1998, 199, 1089‐
1097.
[2]. J.M. Raquez, P. Degee, R. Narayan, P. Dubois: ʺCoordination‐insertionʺ ring‐
opening polymerization of 1,4‐dioxan‐2‐one and controlled synthesis of
diblock copolymers with epsilon‐caprolactone. Macromol. Rapid Commun.
2000, 21, 1063‐1071.
[3]. J.M. Raquez, P. Degee, R. Narayan, P. Dubois: Some thermodynamic, kinetic,
and mechanistic aspects of the ring‐opening polymerization of 1,4‐dioxan‐2‐
one initiated by Al((OPr)‐Pr‐i)(3) in bulk. Macromolecules 2001, 34, 8419‐8425.
[4]. H.L. Lin, C.C. Chu, D.J. Grubb: Hydrolytic degradation and morphologic
study of poly‐p‐dioxanone. Biomed. Mater. Res. 1993, 27, 153‐166.
43
Chapter 2
[5]. H. Nishida, M. Yamashita, T. Endo, Y. Tokiwa: Equilibrium polymerization
behavior of 1,4‐dioxan‐2‐one in bulk. Macromolecules 2000, 33, 6982‐6986.
[6]. T. Hamaide, M. Pantiru, H. Fessi, P. Boullanger: Ring‐opening
polymerisation of epsilon‐caprolactone with monosaccharides as transfer
agents. A novel route to functionalised nanoparticles. Macromol. Rapid
Commun. 2001, 22, 659‐663.
[7]. K. Bernard, P. Degee, P. Dubois: Regioselective end‐functionalization of
polylactide oligomers with D‐glucose and D‐galactose. Polym. Int. 2003, 52,
406‐411.
[8]. H.R. Kricheldorf: Syntheses and application of polylactides. Chemosphere
2001, 43, 49‐54.
[9]. H.R. Kricheldorf, M. Berl, N. Scharnagl: Poly(lactones). 9. Polymerization
mechanism of metal alkoxide initiated polymerizations of lactide and various
lactones. Macromolecules 1988, 21, 286‐293.
[10]. H.R. Kricheldorf, I. Kreisersaunders: Polylactones. 30. Vitamins, hormones
and drugs as co‐Initiators of AlEt3‐initiated polymerizations of lactide.
Polymer 1994, 35, 4175‐4180.
[11]. L.M. Esteves, L. Marquez, A.J. Muller: Optimization of the coordination‐
insertion ring‐opening polymerization of poly(p‐dioxanone) by programmed
decrease in reaction temperatures. J. Appl. Polym. Sci. 2005, 97, 659‐665.
[12]. D.C. Montgomery: Design and Analysis of Experiments, 5th Edition, John Wiley
& Sons Inc., New York, USA, 2001.
[13]. P. Dubois, R. Jerome, P. Teyssie: Aluminum alkoxides ‐ A family of versatile
initiators for the ring‐opening polymerization of lactones and lactides.
Makromol. Chem. Macromol. Symp. 1991, 42/43, 103‐116.
[14]. A. Duda, A. Kowalski, J. Libiszowski, S. Penczek: Thermodynamic and
kinetic polymerizability of cyclic esters. Macromol. Symp. 2005, 224, 71‐84.
[15]. A. Kowalski, A. Duda, S. Penczek: Polymerization of L,L‐lactide initiated by
aluminum isopropoxide trimer or tetramer. Macromolecules 1998, 31, 2114‐
2122.
[16]. A. Duda, A. Kowalski, S. Penczek, H. Uyama, S. Kobayashi: Kinetics of the
ring‐opening polymerization of 6‐, 7‐, 9‐, 12‐, 13‐, 16‐, and 17‐membered
lactones. Comparison of chemical and enzymatic polymerizations.
Macromolecules 2002, 35, 4266‐4270.
[17]. A. Duda: Polymerization of epsilon‐caprolactone initiated by aluminum
isopropoxide carried out in the presence of alcohols and diols. Kinetics and
mechanism. Macromolecules 1996, 29, 1399‐1406.
44

[18]. S. Penczek, T. Biela, A. Duda: Living polymerization with reversible chain
transfer and reversible deactivation: The case of cyclic esters. Macromol. Rapid
Commun. 2000, 21, 941‐950.
45
Chapter 3
Synthesis of Poly‐(ε)‐caprolactone Grafted Starch
Co‐polymers by Ring Opening Polymerisation
using Silylated Starch Precursors

Abstract
Poly‐(ε)‐caprolactone grafted corn starch co‐polymers were synthesized using a
hydrophobised silylated starch precursor. The silylation reaction was performed
using hexamethyl disilazane (HMDS) as the reagent in DMSO at 70°C. Silylated
starch with a degree of substitution (DS) between 0.45‐0.7 was obtained. ε‐
Caprolactone is grafted to silylated starch by a ring‐opening polymerisation
catalysed by Al(OiPr)3 in THF at 50oC. The grafting efficiency varies between 28
and 58%, the remainder being homopolymers of ε‐caprolactone. The DS of the
polycaprolactone graft is between 0.21 and 0.72. The poly‐(ε)‐caprolactone side
chains consist of 40‐55 monomer units and is a function of the reagent intakes.
Experiments with native starch under similar conditions do not result in the
desired poly‐(ε)‐caprolactone grafted corn starch co‐polymers and unreacted
starch was recovered after work‐up. Removal of the silyl groups of the poly‐(ε)‐
caprolactone grafted starch co‐polymers is possible using a mild acid treatment
with diluted hydrochloric acid in THF at room temperature.

Keywords: starch, biodegradable polymers, grafting, polycaprolactone, silylation

Chapter 3
3.1. Introduction
Worldwide, 245 million tons of plastics are produced per year, and this value
increases with about 10% per year [1]. These plastics are mainly synthetic
polymers from fossil resources, which are known to degrade with difficulty and
cause serious environmental problems [2]. The development of green
biodegradable polymers for e.g. the future generation of packaging materials is
highly desirable.
Starch, a natural biopolymer, is one of the potential candidates for future
biodegradable polymer products. Starch is abundantly available. Global
production of starch is 60 million ton per year in 2004 [3]. Starch is present in the
body of many plants (tubers, roots) as granules or cells with typical particle sizes
between 1‐100 µm. The polymeric structure of starch consists of repeating
anhydroglucose units. There are two types of biopolymer in starch, amylose (a
linear polymer of anhydroglucoses with α‐D‐1,4‐glucosidic bonds) and
amylopectin (a branched polymer with α‐D‐1,6‐glucosidic bonds besides α‐D‐1,4‐
glucosidic bonds). The content of amylose in starches depends on the plant and
typically varies between 18‐28%. The amylose‐amylopectin ratio in native as well
as modified starches has a strong impact on the product properties.
Starch films are known to have good oxygen barrier properties. However, as
starch is highly hydrophilic, it is water sensitive, and the mechanical properties of
starch‐based films are generally inferior to those derived from synthetic polymers
[4, 5]. Starch modification is therefore needed to meet the product properties in a
number of application areas. Various modification strategies have been explored,
for instance grafting of monomers (like styrene and methyl methacrylate) to the
starch backbone [6, 7]. However, in almost all cases, the used monomers and the
corresponding grafted chains are not easily biodegradable. Starch has also been
thermoplasticized with the help of plasticizers such as glycerol and other
polyalcohols [8]. However, the product properties are in most cases still not up to
standards and blending with other polymers is required [9].
A wide variety of synthetic biodegradable polymers have been prepared. Well
known examples are polyesters derived from cyclic lactones (polycaprolactone,
polyvalerolactone, and polybutyrolactone). Polycaprolactone (PCL) is easily
degraded by micro‐organisms [10]. Aerobic soil‐burial experiments showed that
the mechanical properties of PCL films decreased rapidly in time and were fully
degraded after 4 weeks [11]. PCL has gained much interest for possible
applications in the medical field as well as in the area of packaging materials [12‐
13]
Several studies to combine the properties of starch and PCL have been
performed to obtain fully biodegradable materials with improved product
properties. Blends of thermoplastic starch and PCL are not fully miscible, resulting
48

Synthesis of Poly‐(ε)‐caprolactone Grafted Starch Co‐polymers using Silylated Starch Precursors
in undesirable phase separation [14]. To increase the miscibility of starch and
polycaprolactone, it has been proposed to chemically graft caprolactone onto the
hydroxyl groups of starch using ring‐opening polymerisation [15]. Common Ring
Opening Polymerization (ROP) catalysts such as tin octoate or aluminium
isopropoxide gave low grafting efficiencies (GE, 0‐14%). The highest GE (90%) was
achieved using triethylaluminium as catalyst [15]. This catalyst is extremely air
and water sensitive, therefore difficult to handle and releases ethane, a very
flammable by‐product, during the reaction. All available data indicate that the
presence of water reduces the GE. This is rationalised by assuming that water
competes with the hydroxyl groups of starch in the initiation step of the
polymerization reaction, thus leading to the formation of PCL homopolymers
rather that starch‐g‐PCL [15]. Another possible cause for the low grafting
efficiencies is the heterogenous nature of the reaction. Starch is insoluble in the
typical organic solvents used for ROP (such as toluene or THF), leading to a
liquid‐solid system. This is expected to lead to reduced reaction rates between
starch and CL compared to CL homopolymerisation, thus to a reduction in the GE.
In this chapter, an alternative method to synthesize poly‐(ε)‐caprolactone
grafted starch co‐polymers (starch‐g‐PCL) is reported. The starch source is made
less hydrophilic and thus more soluble in organic solvents by substituting part of
the OH groups of starch by a bulky silyl group [16‐18]. In this way, the ring
opening polymerisation occurs solely in the liquid phase and this is expected to
lead to higher GE values. This approach has also been applied successfully to graft
PCL and polylactide on dextran [19‐20].

3.2.1. Materials
Corn starch (Sigma) was dried at high vacuum (~1 mbar) at 100 oC for one day
before use. Hexamethyldisilazane (HMDS, Acros) and methanol (Labscan) were
used as received. DMSO (Acros) and toluene (Labscan) were dried overnight over
molecular sieves 3 Å (Merck) and stored under a protective nitrogen atmosphere.
Dry tetrahydrofuran (THF) and toluene for polymerization experiments were
obtained in closed vessels from Aldrich and were used as received. Hydrochloric
acid (HCl) 1 N was prepared from Titrisol concentrated hydrochloric acid solution
(Merck) and distilled (Milipore) water. ε‐Caprolactone monomer (Fluka) was
dried over Calcium Hydride (CaH2) for 48 h, distilled under reduced pressure at
100 oC and stored under a protective nitrogen atmosphere. Aluminium
triisopropoxide [Al(OiPr)3] (Acros) was used without further purification. A stock
solution was prepared by dissolving 1.67 gram (8 mmol) Al(OiPr)3 in 50 ml of dry
toluene in a glove box.
49

Chapter 3

3.2.2. Methods
All reactions and manipulations with air‐sensitive materials were carried out
under a protective nitrogen atmosphere either using standard Schlenk techniques
or in a glove box.

3.2.2.1. Typical example of the starch silylation procedure
The procedure for corn starch silylation was adapted from that published for
dextran [19‐23]. For each experiment, pre‐dried corn starch (6 g) and dry DMSO
(75 ml) were stirred at 70oC for about 3 h until a clear solution was formed. The
pre‐determined amount of HMDS (typically 24 ml, 0.111 mol) was added to the
gelatinized mixture to initiate the silylation reaction. The reaction was carried out
at 70 oC. After 2 and 4 h reaction time, toluene (40 ml) was added to solubilize the
precipitated (partially silylated) starch. After 6 h, another portion of toluene (20
ml) was added. After 24 h, the solvents were removed from the silylated starch
product under reduced pressure (~ 20 mbar) at 70 oC. Traces of DMSO trapped in
the product were removed by dissolving the product in a small amount of toluene
and re‐precipitation in methanol. This procedure was repeated three times. After
removal of the solvents under reduced pressure (0.1 mbar, 80 oC), the silylated
starch (1) product was dried in a vacuum oven (~5 mbar, 40 oC) until constant
weight. The white‐to‐transparent solid products were stored under vacuum in a
desiccator at 6‐8 oC. The samples were characterized by 1H‐NMR.
Silylated Starch (1, before peracetylation, Sample SN‐3, DS = 0.60):
1
H‐NMR (CDCl3, 50 oC): δ 0.12 (m, silyl‐CH3), 3‐6 ppm (m, broad peaks, starch).
Silylated Starch (1, after peracetylation, Sample SN‐3, DS = 0.60):
1
H‐NMR (CDCl3, 50 oC): δ 0.12 (m, silyl‐CH3), 1.7‐2.5 (m, acetate‐CH3), 3‐6 ppm (m,
broad peaks, starch).

3.2.2.2. Typical example of in situ polymerization of ε‐CL with silylated starch
The silylated product from the first step was dissolved in THF (0.6 mL) at 50 oC
(1‐2 h). The intake of 1 depended on its DS and was adjusted to obtain a solution
with 5 x 10‐5 mol free‐OH groups/ml of THF. To this homogenous solution, THF
(4.5 ml) and a predetermined amount of the stock solution of Al(OiPr)3 in toluene
were added. A mol ratio of ‐OH groups to catalyst of 10:1 was used. The mixture
was stirred at 50 oC for 4 h to promote the exchange reaction between the
isopropoxide groups of Al(OiPr)3 and the free ‐OH groups of starch. Subsequently,
a predetermined amount of ε‐caprolactone monomer (molar ratio of monomer to
50

OH‐groups was 13 : 1 for a standard experiment) was added and the ring opening
polymerization reaction was allowed to proceed for 24 h at 50 oC. The reaction was
stopped by cooling down the mixture to room temperature and the addition of 2‐3
drops of 1 N HCl to deactivate the catalyst. The silylated starch‐g‐PCL (2) product
was precipitated from the solution by the addition of heptane (about 250 ml) at ‐18
o
C. The solid precipitate was filtered and dried under vacuum (~ 5 mbar) at 40 oC
for 48 h. The total isolated yield at this condition (see Table 3.2.) was >99 %. The
yield was measured gravimetrically and is based on the weight of the product and
the total weight of reactants charged to the reactor. The samples were
characterized by 1H‐NMR.

Silylated Starch‐g‐PCL (2, Sample SN1CL1, DSsilylation = 0.68, DSPCL=0.21):
1
H‐NMR (DMSO d‐6, 60 oC): δ 0.12 (s, silyl‐CH3), 1.16 (d, ‐CH3 , iPr), 1.31 (m, γ‐
PCL), 1.54 (m, β and δ‐PCL), 2.25 (t, α‐PCL), 3.37 (t, ε’‐PCL), 3.98 (t, ε‐PCL), 3.5‐
3.75, 4.3‐4.5, and 5‐5.4 (m, broad peaks, starch), 4.88 ppm (m, ‐CH, iPr).

3.2.2.3. De‐silylation of poly‐(ε)‐caprolactone grafted silylated starch co‐polymers
Desilylation of the 2 was performed using a procedure described by Ydens et al
[20] for desilylation of silylated dextran‐g‐PCL. The silyl group was removed by
adding a slight excess (with respect to the number of the silyl functionalities) of
1N HCl to a solution of ‐starch‐silylated‐g‐PCL in THF (10 % w/v). After stirring
for 2 h, the desilylated starch‐g‐polycaprolactone product (3) was precipitated
using heptane, filtrated, and vacuum dried at 40 oC for 24 h. The product was
collected as a white solid and characterised by 1H‐NMR.
Starch‐g‐PCL (3, Deprotection product of Sample SN1CL2, DSsilylation = 0.68,
DSPCL=0.34):
1
H‐NMR (DMSO, 60 oC): δ 1.16 (d, ‐CH3, iPr), 1.31 (m, γ‐PCL), 1.54 (m, β and δ‐
PCL), 2.25 (t, α‐PCL), 3.37 (t, ε’‐PCL), 3.98 (t, ε‐PCL), 3.5‐3.75, 4.3‐4.5, and 5‐5.4 (m,
broad peaks, starch), 4.88 ppm (m, ‐CH, iPr).

3.2.2.4. Peracetylation of silylated starch
Characterisation of the silylated starch by NMR proved very difficult due to the
presence of very broad and overlapping resonances arising from starch.
Peracetylation of the remaining OH groups of modified starch is a well established
procedure to improve characterisation of the products by NMR [24]. The
peracetylation procedure applied in this study was adapted from the literature [24,
25]. Typically, 1 (0.1 g) was suspended in THF (4%‐w/v) and stirred at 55 oC until
it was fully dissolved (typically 3 h). Subsequently, the peracetylating reagents
51

Chapter 3
(DMAP, acetic anhydride and pyridine in a DMAP : acetic anhydride : pyridine :
AHG molar ratio of 1: 10: 22 : 1) were added. The peracetylation reaction was
conducted for 7 h at 50 oC. The product was precipitated by the addition of
methanol and washed several times with methanol. It was finally dried overnight
in a vacuum oven at 70 oC and 5 mbar until constant weight.

3.2.3. Analytical Methods
3.2.3.1. Nuclear Magnetic Resonance (NMR)
H‐ NMR spectra were recorded in CDCl3 at 50oC or in DMSO d‐6 at 60 oC on a
1
Varian AMX 400 NMR machine.

3.2.4. Calculations
The DS of the silylated starch (DSsilylation) is defined as the average number of
silyl groups present on an anhydroglucose unit (AHG) of starch. DSsilylation may be
calculated using 1H‐NMR spectra of the products after peracetylation using eq. 3.1.
1 ACH 3 silyl / 9 7 A−0.6−0.6 ppm

DSsilylation = × = × (3.1.)
3 Astarch proton / 7 27 A3−5.8 ppm
where Ax‐y stands for the 1H‐NMR peak area in the range δ x‐y ppm.
The Average Chain Length (ACL) of the Poly‐(ε)‐caprolactone chain is defined
as the average number of CL repeating units in a grafted polymer chain. The ACL
is calculated from 1H‐NMR spectra by comparing the peak area of protons
attached to ε‐carbon atoms in a repeating CL unit with that of the characteristic ε’
protons of the last CL unit in a PCL chain [26] (see Figure 3.3.). In this calculation,
it is assumed that the average length of the grafted chain is equal to the chain
length of the homopolymer. This leads to the following equation:
ACH 2 −ε + ACH 2 −ε ' A3.8− 4.2 ppm

ACL = = + 1 (3.2.)
ACH 2 −ε ' A3.3−3.4 ppm
The degree of substitution of the PCL graft on 2 (DSPCL) is defined as the average
number of PCL polymer chains present on an AHG unit of starch. When assuming
that all ε‐CL monomer is converted, the DSPCL may be calculated using eq. 3.3. The
assumption of high conversions (>95%) was correct for all experiments (see Table
3.2.)
52

0.5. A3.3−3.4 − A4 .8− 4 .9

DS PCL = (3.3.)
⎡ 0.5.( A3.3−3.4 + A3.8− 4 .2 ) 1 ⎤
⎢ ×
⎣⎢ ε - CL/OH ratio (3 − DSsilylation )⎥⎦⎥
The grafting efficiency (GE) is defined as the percentage of PCL grafted to
starch compared to the total amount of polymerized CL (grafted and PCL
homopolymer). It is calculated by comparing the area of protons related to the
PCL grafted to starch with the area of the protons of all PCL chains present in the
product. This leads to the following equation:
ACH 2 −ε ' / 2 − ACH −iso

GE = × 100 %
ACH 2 −ε ' / 2
(3.4.)
⎛ A ⎞ ⎛ A ⎞
= ⎜1 − CH −iso ⎟ × 100 % = ⎜⎜1 − 4.8− 4.9 ⎟⎟ × 100 %
⎜ ⎟
ACH 2 −ε ' / 2 ⎠ A3.3−3.4 / 2 ⎠
⎝ ⎝
where Ax‐y stands for the peak area in the range δ x‐y ppm.
The Hildebrand solubility parameter of HMDS and DMSO were calculated
using the following equation [27]:
1/ 2
⎛ ∆H v − RT ⎞
δ = ⎜⎜ ⎟⎟
⎝ Vm ⎠
where ∆Hv stands for heat of vaporization, T stands for absolute temperature,
and Vm stands for molar volume. The values of ∆Hv and Vm were obtained from the
SciFinder Scholar database (American Chemical Society, 2007)

The overall procedure to synthesize poly‐(ε)‐caprolactone grafted starch co‐
polymers (3) consists of three steps and includes hydrophobization of starch by
silylation of part of the hydroxyl groups of starch using hexamethyl disilazane
(HMDS), followed by an in‐situ Ring Opening Polymerization (ROP) of ε‐
caprolactone monomer on the hydrophobized starch and subsequent silyl group
removal by a mild acid treatment. Although all steps have been investigated, the
focus of this chapter will be on the first two steps of the procedure.

53

Chapter 3
3.3.1. Synthesis of Silylated Starch
The silylation of corn starch was performed with HMDS as the silylating agent
(eq. 3.5.). The silylation procedure was adapted from that previously reported for
dextran [19‐23].
6 6
H2C OH H2C O R
1 DMSO O 1
O H3C CH3 4H H
4H H5
H 50 deg. C H5
+ H3C Si NH Si CH3 OR H
+ NH3
OH H
3 H3C CH3 3
2 OH H 2OR
H
[R = H or Si(CH3)3]
Starch Hexamethyldisilazane Silylated Starch
(1)

(3.5.)
Instead of using DMSO as solvent, mixtures of toluene and DMSO were applied
to avoid precipitation of the silylated starch during the reaction. In this way, a
homogeneous reaction mixture was maintained throughout the reaction. The
silylated products were characterised by NMR. Very broad peaks of starch
protons at δ 3‐5 ppm and a sharp peak of the methyl substituents of the silyl group
at about δ 0 ppm were observed (see Figure 3.1.a.).
The degree of substitution of the silyl groups (DSsilylation) was determined by 1H‐
NMR. It is defined as the average number of silyl substituted OH groups on the
anhydroglucose (AHG) unit of starch. The 1H‐ NMR spectrum of silylated starch
(Figure 3.1.a.) shows the presence of silyl groups at about δ 0 ppm. However, the
starch peaks are very broad and this feature hampers accurate determination of
the DSsilylation. An additional peracetylation procedure to substitute the free
hydroxyl groups with acetate groups was performed to improve the quality of the
NMR spectra, as suggested by Einfeldt et al [25]. The 1H‐ NMR peaks from the
AHG unit after peracetylation were indeed considerably sharper and allowed
more accurate DS calculations (Figure 3.1.b.). Using standard conditions (HMDS:
AHG molar ratio of 3, 70 oC, 24 h), a product with a DS of 0.60 was obtained.

54

Figure 3.1. Typical 1H‐NMR Spectra of Silylated Starch
(Sample SN3, DSsilylation=0.60, in CDCl3, 50 oC):
(a). not peracetylated (b). peracetylated

The effect of the HMDS to starch molar ratio (1.5‐4.5) on the product DS was
investigated by performing experiments with a constant starch and a variable
HMDS intake. The results are given in Table 3.1. and Figure 3.2.

55

Chapter 3
0.65
0.6
DSsilylation
0.55
0.5
0.45
0.4
1.5 2 2.5 3 3.5 4 4.5
HMDS : AHG Ratio [mol/ mol]

Figure 3.2. DS of the silylated products at different HMDS: AHG ratios
(DMSO, 70 oC)

Table 3.1. Effect of HMDS: AHG mol ratio on DS of the silylation product a
HMDS intake HMDS: AHG Product DS

Experiment
(ml, mmol) mol ratio (DSsilylation)
SN1 12 (56) 1.5 0.68
SN2 18 (83) 2.25 0.67
SN3 24 (111) 3 0.60
SN4 36 (167) 4.5 0.46
a
Experiments were performed in DMSO at 70°C. For all experiments, a fixed starch intake of 6 g
(37 mmol AHG) starch was applied.

Surprisingly, the DSsilylation decreases for higher intakes of HMDS. For reactions
with an order higher than zero, a positive effect of higher reagent intake on the
reaction rate and thus the DS is expected. The experimentally observed lowering
at higher HMDS intakes is likely due to a decrease in the polarity of the reaction
mixture. HMDS is a rather apolar compound (Hildebrand solubility parameter of
6.25 cal1/2cm‐3/2) due to the presence of the bulky apolar SiMe3 groups. Its presence
will reduce the polarity of the reaction medium (DMSO, solubility parameter =
11.36 cal1/2cm‐3/2) considerably. At a mol ratio of HMDS to AHG of 4.5:1, the
volumetric HMDS intake is about half of the DMSO intake. The silylation reaction
56

involves charged‐ionic species [20] and a reduction in the polarity of the reaction
medium is expected to lead to a lowering of the silylation reaction rates. Similar
reductions in the reaction rates when working at higher reactant intakes were
observed for the esterification of starch with vinyl laurate and stearate [28].
To the best of our knowledge, the silylation of starch in pure DMSO using
HMDS as the reagent has not been reported to date. The DSsilylation (0.45‐0.7) of the
products is in the range of those published for other starch silylation systems.
Petzold et al [18] reported that silylation of starch with trimethylsilylchloride
(TMSCl) in pyridine yielded trimethyl‐silyl substituted starch with DS values
between 0.3‐2.2. Silylation of starch with HMDS in formamide, DMF, DMA/LiCl,
pyridine, liquid ammonia, and DMSO/pyridine mixtures yielded silylated starch
with DS values ranging between 0.7‐3.0 [18, 29]. The use of HMDS to silylate
dextran (Mw=6000‐40000) in DMSO (HMDS to OH molar ratio of 0.25‐5.0) gave
silylated dextran with DS values between 1.1 and 3.0 [21]. The much higher DS
values obtained for dextran compared to starch may be related to differences in
molecular weights between starch and dextran and the type of AHG linkages
(mainly α‐1,6‐glucosidic for dextran).

3.3.2. In situ Ring Opening Polymerization of ε‐Caprolactone with Silylated
Starch
A number of in situ ROP experiments with ε‐caprolactone (CL) were carried out
using a typical silylated starch sample (DS=0.68, sample SN‐1) in THF at 50 oC for
24 h using Al(OiPr)3 as catalyst. A schematic representation of the reaction is
provided in eq. 3.6. After precipitation with heptane and vacuum drying, white
solid products with isolated yield > 96% were obtained. The products are soluble
in DMSO as well as in less‐polar solvent such as chloroform and THF.

O
O C (CH2)5 O H
n
O
H2C O R Catalyst: Al(OiPr)3 CH2
O O Toluene, 50 oC O
H H H H
H H
+
OR H OR H
H OR H OR
(R = H or Si(CH3)3
[R = H or Si(CH3)3 ]
Silylated Starch ε-Caprolactone Silylated

Monomer Starch-g-Polycaprolactone
(1)
(2)
(3.6.)
57

Chapter 3
The products were characterized using 1H‐NMR analysis in DMSO‐d‐6 as the
solvent. A typical spectrum is shown in Figure 3.3.
The peaks from the polycaprolactone units are clearly present in the range of δ
1.2‐4 ppm and imply that the ring‐opening polymerisation reaction of CL indeed
occurred. Resonances from the starch peaks are observable as small, broad peaks
in the region δ 3.4‐3.8 and 5.0‐5.4 ppm. However, not all of the caprolactone is
grafted to starch and large amounts of PCL homopolymers (72%) were formed.
This is clearly indicated by resonances of the iPr end‐group of the PCL
homopolymer at δ 4.9 and 1.2 ppm. Further process optimization studies allowed
the synthesis of products with less than 42% of homopolymers (vide infra). The
homopolymers are formed by direct polymerization of ε‐CL initiated on
isoproproxide moieties attached to the Al catalyst, as is shown in Figure 3.4.
Apparently, the exchange reaction between Al(OiPr)3 and the OH groups of
silylated starch is not quantitative under the conditions applied in this study. The
formation of CL homopolymers for this type of reactions has been observed before
[15, 26].

Figure 3.3. Typical 1H‐NMR spectrum of a silylated starch‐g‐PCL sample.
(Sample SN1CL1, DSsilylation =0.68, DSPCL=0.21) in DMSO‐d6 at 60 oC.
Coding of the peaks is given in Figure 3.4.

58

1. Exchange Reaction:
iPr O
Al O iPr
H2C OH H2C O
iPr O O O
H H
Al O iPr O H O H
+ OH H O O + iPr OH
H OR H
iPr O H
H OH H OR
Silylated Starch [R = H or Si(CH3)3 ]

(1)
2. Polymerization Reaction:
iPr O O O
α β χ δ ε α' β' χ' δ' ε'
Al O iPr O H2C [O C CH2 CH2 CH2 CH2 CH2 ]n O C CH2 CH2 CH2 CH2 CH2 OH
H2C O O
O H
O H
O + n+1 OR H O
H H
O H
OR H O
H H OR
H OR
Silylated Starch-g-PCL
(2)
O
iPr O O α β χ δ ε O α' β' χ' δ' ε'
O
Al O iPr + n+1 iPr [O C CH2 CH2 CH2 CH2 CH2 ]n O C CH2 CH2 CH2 CH2 CH2 OH
iPr O
PCL homopolymer
Figure 3.4. Product formation for the reaction between silylated starch and ε‐
caprolactone

Besides the ‐OH group of starch, residual water may also initiate the
polymerisation reaction. This leads to the formation of carboxylic end groups (see
eq. 3.7.). However, peaks arising from a ‐COOH unit could not be detected in 13C‐
NMR spectra (δ 175‐180 ppm).
1. Exchange Reaction:
iPr O iPr O
Al O iPr + H2O Al OH + iPr OH
iPr O iPr O
2. Polymerization Reaction:
O
iPr O O α β χ δ ε O α' β' χ' δ' ε'
O
Al OH + H [O C CH2 CH2 CH2 CH2 CH2 ]n O C CH2 CH2 CH2 CH2 CH2 OH
iPr O
PCL homopolymer with carboxylic end-group

(3.7.)
59

Chapter 3
The ratio of homopolymerisation versus grafting on starch may be obtained by
comparing the integrals of selected peaks in 1H‐NMR spectra. In the case of only
homopolymerisation, the intensity of the peak from the ‐CH2‐ end group of the
homopolymer (ε’ at δ 3.3 ppm) should be twice the intensity of the ‐CH‐ proton of
the isopropoxide end group (δ 4.9 ppm). However, in all samples, the intensity of
the resonance ε’ was considerably higher that that of the ‐CH‐ iPr peak. This
implies that grafting of caprolactone to starch also occurs to a significant extent.
The grafting efficiency (GE) for the sample given in Figure 3.3. (SN1CL1) is 28%.
Five experiments were performed to study the effect of different ε‐CL to
silylated starch ratio. The results are given in Table 3.2. and Figure 3.5. The yield
of the products was measured gravimetrically and varies between 96.5 and 100%.
This implies that the ε‐CL conversion is essentially quantitative in all cases. The
GE increases with higher ε‐CL intakes, and reaches 58% for a ε‐CL to starch–OH
ratio of 29.2.
The ACL of the polymer and the DSPCL increase almost linearly with the ε‐
caprolactone intake (Figure 3.5.). This indicates that higher ε‐CL concentrations
during the reaction lead to longer PCL grafts as well as to higher levels of
initiation of the grafting reaction on free hydroxyl group of silylated starch.

Table 3.2. Overview of results for the grafting of ε‐CL on silylated starch a
ε‐CL/ OH Total Avg. Chain Length Grafting Efficiency

Experiment DSPCL
[mol/ mol] Yieldb [%] [mon. units] [%]
SN1CL1 13.0 >99 40 0.21 28
SN1CL2 15.0 >99 43 0.34 43
SN1CL3 18.9 >99 44 0.47 48
SN1CL4 22.5 99 49 0.58 55
SN1CL5 29.2 96.5 54 0.72 58
a.
All reactions were performed using the same intake of SN1 silylated starch (DS=0.68) in THF at 50 oC with
Al(OiPr)3 as the catalyst (1 mol Al(OiPr)3 per 10 mol‐starch‐OH groups).
b.
Determined gravimetrically and defined as the total weight of the isolated product divided by the total
intake of reactants (silylated starch and CL).
60

55 0.9
0.8
50
Average Polymer Length [mon. units]
0.7
45 0.6
DSPCL
0.5
40
0.4
35 0.3
0.2
30
12 14 16 18 20 22 24 26 28 30
ε-CL : Starch-OH Ratio [mol/ mol]

Figure 3.5. Average Chain Length (ACL) and the DS for as a function of the CL‐
starch intake. ∆ : Average Chain Length (ACL); □ : DSPCL

The mechanism of ROP of cyclic esters such as caprolactone in the presence of
an alcohol (ROH, silylated starch in our case) is provided in Scheme 3.1. [30‐31].
Higher monomer intakes are expected to lead to higher polymerization rates as
shown in Scheme 3.1.b. This will result in longer PCL chains in the final product,
in line with the experimental observations (see Table 3.2. and Figure 3.5.).
At higher ε‐caprolactone intakes not only the ACL of the grafted chain increases
but also higher values for the DSPCL are observed. This finding may be rationalised
by assuming that the rate of chain transfer (Scheme 3.1.c.) with starch is increased
at higher ε‐caprolactone intakes. The rate of this reaction is expected to be a
function of both the starch and the concentrations of Al‐species with a growing
PCL chain. The starch intake for all experiments was equal, meaning that the
concentration of Al‐species with a growing PCL chain should be higher at higher
ε‐CL intakes. This is indeed predicted by the mechanism given in Scheme 3.1.b.;
higher caprolactone intakes will increase the rate of this reaction and lead to
higher concentration of Al‐species with a growing PCL chain.

61

Chapter 3
Al O iPr + R OH Al O R + iPr OH a.
O kp O
Al O R + n Al O [(CH2)5 C O]n R b.
kd
O O
ktr(a)
Al O [(CH2)5 C O]n R + R OH Al O R + HO [(CH2)5 C O]n R c.
ktr(a)

Scheme 3.1.

The observation that a higher monomer to alcohol ratio leads to higher amounts
of PCL chains with an alcohol end group and thus a higher DS was also reported
for the polymerization of p‐dioxanone with Zn‐lactate as catalyst and α‐tocopherol
as the alcohol [32].
To show the potential of our approach to use hydrophobised starch instead of
native starch for the ring opening polymerisation of cyclic esters, several ring
openings polymerisations of native starch with ε‐CL monomer were performed
either in pure ε‐CL or in a mixture of ε‐CL and toluene (80‐100oC, 24 hr). At the
start of the reaction, the starch was always insoluble in the reaction medium. After
reaction the product was isolated, washed thoroughly with toluene and dried. The
weight of the product, however, was very close to the initial starch intake.
Examination of the products by FT‐IR does not show the presence of caprolactone
vibrations. Thus, it may be concluded that solubilisation of starch is of key
importance to obtain poly‐(ε)‐caprolactone grafted starch co‐polymers.
These findings are in line with earlier studies on the ROP of ε‐caprolactone on
native granular starch using Al(OiPr)3 as catalyst [15]. Here, caprolactone
polymerization did not occur after 18‐24 h reaction time and only liquid ε‐CL was
recovered. Only when performing the reaction with high concentrations of the
aluminium catalyst, a product with a GE of about 13% was obtained. This low GE
was explained by assuming that the reaction between starch and Al(OiPr)3 is slow
and due to the heterogeneous nature of the reaction mixture.
Our study, together with the result of Dubois et al [15] indicate that
homogenous reaction conditions are required for the successful ROP of ε‐CL to
obtain poly‐(ε)‐caprolactone grafted starch co‐polymers when using Al(OiPr)3 as
the catalyst. When performing the reaction under heterogenous conditions, a high
grafting efficiency is only achievable when using triethylaluminium as the catalyst
[15].

62

3.3.3. Deprotection of Silylated‐Starch‐g‐PCL
A preliminary experiment was performed to remove the silylated groups of the
silylated starch‐g‐PCL (eq. 3.8.) using a mild acid treatment with diluted
hydrochloric acid in THF at room temperature. All silyl groups were removed
successfully, as is clearly seen from an NMR spectrum given in Figure 3.6.

Figure 3.6. H‐ NMR Spectrum of Starch‐g‐PCL (reaction product after
1
desilylation of sample SN1CL2, DSsilylation =0.68, DSPCL=0.34), in
DMSO‐d6

O O
O C (CH2)5 O H O C (CH2)5 O H
n n
H2C CH2
O THF, 2 hr O
H H H H
H H
OR H H + OH H
H OR H OH
Silylated Starch-g-Polycaprolactone Starch-g-Polycaprolactone
(2) (3)
(3.8.)
63

Chapter 3
3.4. Conclusions
The successful synthesis of poly‐(ε)‐caprolactone grafted corn starch co‐
polymers using a three step approach is reported. The key feature is the use of a
homogeneous reaction mixture for the ROP of starch with ε‐CL. This was achieved
by making the starch more hydrophobic by partial substitution of the OH groups
by trimethylsilyl groups. Silylated starch with a low‐medium DS (0.46‐0.68) was
obtained using a DMSO/toluene mixture as the solvent and HMDS as the
silylating agent. The ROP with ε‐CL was performed using Al(OiPr)3 as catalyst in
THF as the solvent. Poly‐(ε)‐caprolactone grafted silylated starch co‐polymers
with average chain length of 40‐55 monomer units (polymer molecular weight of
4500‐6300) were obtained. The DS of the PCL chains was between 0.21‐0.72,
depending on the ε‐CL‐starch ratio. Considerable amounts of ε‐CL hompolymers
with isopropyl end‐groups were also formed. The grafting efficiency varied
between 28‐58%, the highest value was obtained with a ε‐CL‐AHG ratio of 29.2.
Control ROP experiments of ε‐CL with native starch under similar conditions did
not produce the desired poly‐(ε)‐caprolactone grafted corn starch co‐polymers,
indicating that homogeneous reaction conditions are favorable for the grafting
reaction. The products may have interesting applications as compatibilizers for
starch‐polymer blends.

3.5. Nomenclature
A : peak area of certain proton in 1H‐NMR spectra [‐]
ACL : average number of CL repeating units in a grafted polymer chain or
homopolymer [monomer units]
DS : Degree of Substitution, average value of mole of substituted –OH per
mole of anhydroglucose (AHG) units [‐]
DSsilylation : DS of silyl group substituents [‐]
DSPCL : DS of PCL chain substituents [‐]
GE : Grafting Efficiency, the percentage of PCL grafted to starch compared
to the total amount (grafted and homopolymer) of polymerized CL [%]
R : gas constant, 1.986 cal mol‐1 K‐1
T : temperature [K]
Vm : molar volume [cm3/ mol]
64

Greek symbols:
∆H v : heat of vaporization [kJ/mol]
δ : Hildebrand solubility parameter [cal1/2cm‐3/2]

3.6. References
[2]. H. Yavuz, C. Babac: Preparation and biodegradation of
starch/polycaprolactone films. J. Polym. Environ. 2003, 11, 107‐113.
http://www.starch.dk/isi/stat/rawmaterial.html, accessed on August 01, 2008.
[4]. J.M. Krochta, C. De Mulder‐Johnston: Edible and biodegradable polymer
films: Challenges and opportunities. Food Technol. 1997, 51, 61‐74.
[5]. A. Tsiapouris, L. Linke: Water vapor sorption determination of starch based
porous packaging materials. Starch‐Starke 2000, 52, 53‐57.
[6]. E.B. Bagley, G.F. Fanta, R.C. Burr, W.M. Doane, C.R. Russell: Graft
copolymers of polysaccharides with thermoplastic polymers ‐ new type of
filled plastic. Polym. Eng. Sci. 1977, 17, 311‐316.
[7]. M.K. Beliakova, A.A. Aly, F.A. Abdel‐Mohdy: Grafting of poly(methacrylic
acid) on starch and poly(vinyl alcohol). Starch‐Starke 2004, 56, 407‐412.
[8]. Biobased packaging materials for the food industry, status and perspectives, report of
Food Biopack Project (Ed. C. J. Weber), Department of Dairy and Food Science,
the Royal Veterinary and Agricultural University, Frederiksberg C.,
Denmark, 2000.
[9]. X. Wang, K. Yang, Y. Wang: Properties of starch blends with biodegradable
polymers. J. Macromol. Sci., Part C: Polym. Rev. 2003, 43, 385–409
[10]. S. Karlsson, A.C. Albertsson: Biodegradable polymers and environmental
interaction. Polym. Eng.Sci. 1998, 38, 1251‐1253.
[11]. D.A. Goldberg: Review of the biodegradability and utility of
poly(caprolactone). J. Environmen. Polym. Degrad. 1995, 3, 61‐67.
[12]. R. Chandra, R. Rustgi: Biodegradable polymers. Prog. Polym. Sci. 1998, 23,
1273‐1335.
[13]. E. Chiellini, R. Solaro: Biodegradable polymeric materials. Adv. Mater. 1996,
8, 305‐313.
65

Chapter 3
[14]. L. Averous, L. Moro, P. Dole, C. Fringant: Properties of thermoplastics
blends: starch‐polycaprolactone. Polymer 2000, 41, 4157‐4167.
[15]. P. Dubois, M. Krishnan, R. Narayan: Aliphatic polyester‐grafted starch‐like
polysaccharides by ring‐opening polymerization. Polymer 1999, 40, 3091‐3100.
[16]. D. Klemm, L. Einfeldt: Structure design of polysaccharides: novel concepts,
selective syntheses, high value applications. Macromol. Symp. 2001, 163, 35‐47.
[17]. K. Petzold, D. Klemm, A. Stein, W. Gunther: Synthesis and NMR
characterization of regiocontrolled starch alkyl ethers. Des. Monomers Polym.
2002, 5, 415‐426.
[18]. K. Petzold, A. Koschella, D. Klemm, B. Heublein: Silylation of cellulose and
starch ‐ selectivity, structure analysis, and subsequent reactions. Cellulose
2003, 10, 251‐269.
[19]. C. Nouvel, P. Dubois, E. Dellacherie, J.L. Six: Controlled synthesis of
amphiphilic biodegradable polylactide‐grafted dextran copolymers. J. Polym.
Sci., Part A: Polym. Chem. 2004, 42, 2577‐2588.
[20]. I. Ydens, D. Rutot, P. Degee, J.L. Six, E. Dellacherie, P. Dubois: Controlled
synthesis of poly(epsilon‐caprolactone)‐grafted dextran copolymers as
potential environmentally friendly surfactants. Macromolecules 2000, 33, 6713‐
6721.
[21]. C. Nouvel, I. Ydens, P. Degee, P. Dubois, E. Dellacherie, J.L. Six: Partial or
total silylation of dextran with hexamethyldisilazane. Polymer 2002, 43, 1735‐
1743.
[22]. D. Nouvel, P. Dubois, E. Dellacherie, J.L. Six: Silylation reaction of dextran:
Effect of experimental conditions on silylation yield, regioselectivity, and
chemical stability of silylated dextrans. Biomacromolecules 2003, 4, 1443‐1450.
[23]. D. Rutot, E. Duquesne, I. Ydens, P. Degee, P. Dubois: Aliphatic polyester‐
based biodegradable materials: new amphiphilic graft copolymers. Polym.
Degrad. Stab. 2001, 73, 561‐566.
[24]. K. Petzold, L. Einfeldt, W. Gunther, A. Stein, D. Klemm: Regioselective
functionalization of starch: Synthesis and 1H NMR characterization of 6‐O‐
silyl ethers. Biomacromolecules 2001, 2, 965‐969.
[25]. L. Einfeldt, K. Petzold, W. Gunther, A. Stein, M. Kussler, D. Klemm:
Preparative and 1H NMR investigation on regioselective silylation of starch
dissolved in Dimethyl Sulfoxide. Macromol. Biosci. 2001, 1, 341‐347.
[26]. T. Hamaide, M. Pantiru, H. Fessi, P. Boullanger: Ring‐opening
polymerisation of epsilon‐caprolactone with monosaccharides as transfer
66

agents. A novel route to functionalised nanoparticles. Macromolecul. Rapid

Commun. 2001, 22, 659‐663.
[27]. D.L. Hertz Jr.: Solubility parameter concepts – a new look, presented at ACS
Rubber Division meeting, Mexico City, 1989
[28]. L. Junistia, A.K. Sugih, R. Manurung, F. Picchioni, L.P.B.M. Janssen, H.J.
Heeres: Synthesis of higher fatty acid starch esters using vinyl laurate and
stearate as reactants, 2008, accepted for publication in Starch‐Starke.
[29]. D. Horton, J. Lehmann: Selective 6‐O‐acetylation of amylose. Carbohydr. Res.
1978, 61, 553‐556.
[30]. A. Duda: Polymerization of ε‐caprolactone initiated by aluminum
isopropoxide carried out in the presence of alcohols and diols. Kinetics and
mechanism. Macromolecules 1996, 29, 1399‐1406.
[31]. S. Penczek, T. Biela, A. Duda: Living polymerization with reversible chain
transfer and reversible deactivation: the case of cyclic esters. Macromol. Rapid
Commun. 2000, 21, 941‐950.
[32]. H.R. Kricheldorf, D.O. Damrau: Zn‐lactate catalyzed polymerizations of 1,4‐
dioxan‐2‐one. Macromol Chem. Phys. 1998, 199, 1089‐1097.
67
Chapter 4
Synthesis of Higher Fatty Acid Starch Esters using
Vinyl Laurate and Stearate as Reactants

Abstract
This chapter describes the synthesis of long fatty esters of corn starch (starch‐
laurate and starch‐stearate) with a broad range in degree of substitution (DS=
0.24 ‐ 2.96). The fatty esters were prepared by reacting the starch with vinyl
laurate or vinyl stearate in the presence of basic catalysts (Na2HPO4, K2CO3, and
Na‐acetate) in DMSO at 110°C. The yellowish products were characterized by
1
H‐, 13C‐NMR and FT‐IR. The DS of the products is a function of the carbon
number of the fatty acid chain, vinyl ester to starch ratio and the type of
catalyst. When performing the reactions using Na2HPO4 as the catalyst, the DS
for the starch‐laurate compounds is higher than for the corresponding starch‐
stearates. For low vinyl‐ester to starch ratios, an increase in the vinyl‐ester
concentration leads to higher product DS values. At higher ratios, the DS
decreases, presumably due to a reduction of the polarity of the reaction
medium. K2CO3, and Na‐acetate are superior catalysts with respect to activity
compared to Na2HPO4 and products with DS values close to 3 were obtained.

Keywords: corn starch, esterification, vinyl laurate, vinyl stearate

Chapter 4
4.1. Introduction
Green biodegradable polymers derived from natural resources are
potentially very interesting substitutes for non‐biodegradable petroleum‐based
polymers. An attractive field of application for these polymers is the use as
packaging materials. For the current petrochemical based products recycling is
often neither practical nor economically feasible [1].
Natural polymers such as starch, cellulose or proteins are potentially very
interesting starting materials for biodegradable packaging materials. In
particular starch is attractive as it is relatively cheap and abundantly available.
However, the use of native starch for packaging materials is limited due to its
low moisture resistance, poor processibility (high viscosity), high brittleness,
and incompatibility with hydrophobic polymers. Further modification of starch
is therefore required to introduce hydrophobicity and to improve mechanical
and moisture barrier properties.
Esterification of starch with low molecular weight fatty acid derivatives is
one of the oldest modification technologies to improve starch properties. The
first paper on the acetylation of starch was already published in 1865 [2].
However, most of the studies performed to date use short chain carboxylic
acids (C1‐C4), and particularly acetic acid derivatives (C2) [2‐4].
The introduction of acetate groups on starch makes the product more
hydrophobic, and consequently, more water‐resistant products may be
obtained [3‐4]. The hydrophobicity increases with the degree of acetate
substitution (DS, defined as the moles of substituents per mole of
anhydroglucose (AHG) units) [4]. However, the mechanical properties of high‐
DS starch derivatives of low chain carboxylic acids still need considerable
improvements before large scale application as packaging materials becomes
within reach. The major obstacle is the pronounced brittleness of the materials,
even after the addition of plasticizers [5]. To improve the mechanical properties,
higher molecular weight carboxylic starch esters (C4‐C6) [6], and even fatty acid
derivatives (C12‐C18) have been used in the modification reaction [5, 7],
resulting in products with DS values up to 2.7 [1, 5]. The mechanical properties
and hydrophobicity of the products were significantly improved when using
these longer chain fatty acid precursors [1, 5]. However, the fatty ester
substituents [1, 5‐6] were introduced using fatty acid chloride reagents, which
are relatively expensive and rather corrosive [7]. An alternative method using
methyl and glyceryl laurate esters in the absence of solvent has been recently
developed [7]. Relatively low‐DS (0.34‐0.61) products were obtained using this
approach.
Recently Mormann et al [8] explored the possibility of using vinyl esters and
particularly vinylacetate as reagents for the preparation of starch esters. Their
70

Synthesis of Higher Fatty Acid Starch Esters using Vinyl Laurate and Stearate as Reactants
research focused on the synthesis of starch acetates and only two examples with
a higher molecular weight fatty acid vinyl ester were reported. The reactions
were either performed in water or in DMSO using a basic catalyst (Na2HPO4).
The maximum attainable DS of starch acetate in water was below 1 and limited
to 0.01 when using vinyl‐laurate. In DMSO, starch esters with a substantially
higher DS value (up to 1.6 for starch acetate ester) were obtained. This solvent
effect is likely caused by the higher solubility of the vinyl esters in DMSO
compared to water, leading to higher reaction rates.
We here report an investigation on the synthesis of higher fatty acid esters of
starch with an emphasis on the introduction of laurate and stearate ester side
chains. The synthesis of starch stearate esters using vinyl ester reagents has, to
the best of our knowledge, not been reported to date. The effects of the starch to
vinylester ratio on the reaction rates and DS have been explored. In addition,
the use of basic catalysts other than Na2HPO4 has been investigated. The effect
of the addition of a non‐polar solvent (toluene) to the reaction medium to
solubilise the products and thus to enhance the reaction rates has also been
studied.

4.2.1. Materials
Corn starch (approx. 73% amylopectine and 27% amylose) was purchased
from Sigma (Germany). The starch was dried before use for 48 h at 105 oC under
vacuum (approx. ~1 mbar), leading to a moisture content of 2 %‐wt (measured
gravimetrically). Analytical grade vinyl stearate (Aldrich, Japan), vinyl laurate
(Fluka, Germany) and acetic anhydride (Merck, Germany) were used without
further purification. Potassium carbonate (Boom, the Netherlands), sodium
acetate (Merck, Germany) and disodium hydrogenphosphate (Merck,
Germany) were used as received. Technical grade dimethyl sulfoxide (DMSO),
4‐N,N‐dimethylaminopyridine (DMAP), and tetrahydrofuran (THF) were
supplied by Acros (Belgium) and were also used as received.

4.2.2. Analytical Equipment
1
H‐ and 13C‐NMR spectra were recorded in CDCl3 on a 400 MHz Varian AMX
NMR machine. The spectra were recorded at 50 oC, as recommended by Laignel
et al [9]. IR spectra were recorded on a Spectrum 2000 FT‐IR Spectrometer
(Perkin Elmer). The products were placed directly on the diamond plate and 50
scans with a resolution of 4 cm‐1 were recorded.
71
Chapter 4
4.2.3. Methods
4.2.3.1. Typical example of the synthesis of laurate and stearate esters of corn starch
Corn starch (0.5 g) was first gelatinized in DMSO (5 ml) at 70 oC for 3 h,
resulting in the formation of a homogenous transparent solution. Subsequently,
vinyl laurate or vinyl stearate (3 mol/mol AHG units in starch) and potassium
carbonate catalyst (2 %‐wt with respect to starch) were added and the mixture
was stirred at 110 oC for 24 h. After cooling, the product was precipitated using
methanol (100 ml) and separated from the liquid phase by decantation. The
product was washed twice with methanol (50 and 25 ml, respectively). Finally,
the product was dried in a vacuum oven (70 oC, approximately 5 mbar) for 24 h
until constant weight.
The samples were characterized by 1H‐ and 13C‐NMR and FT‐IR. The atom
numbering scheme is given in Figure 4.1., typical spectra are given in Figure
4.2. (1H‐NMR), Figure 4.3. (13C‐NMR) and Figure 4.4. (FT‐IR).

Starch‐laurate (Sample 17, Table 4.1., DS = 2.52):
1
H‐NMR (before peracetylation, CDCl3): δ 0.9 (t, 3H, 12), 1.1 (m, broad peaks,
16H, C4‐11), 1.5 (m, 2H, C3), 2.4 (m, broad peaks, 2H, C2), 3‐6 ppm (m, broad
peaks, 7H, C1S‐6S).
1
H‐NMR (after peracetylation, CDCl3): δ 0.9 (t, 3H, C12), 1.3 (m, broad peaks,
16H, C4‐11), 1.5 (m, 2H, C3), 1.8‐2.6 (m, broad peaks, 3H, C2’), 2.3 (m, 2H, C2),
3‐6 ppm (m, 7H, C1S‐6S).
C‐NMR (before peracetylation, CDCl3): δ 14.0 (C12), 22.7 (C11), 24.9 (C3), 28‐32
13
(C4‐9), 31.9 (C10), 34.1 (C11), 61.9 (broad, C6S), 68‐74 (broad, C2S, 3S, 5S), 76‐78
ppm, overlap with CDCl3 (C4S), 95.4 (broad, C1S), 172‐174 ppm (C=O, attached
to O‐C2S, O‐C3S, and O‐C6S).
FT‐IR (cm‐1): 2920 (C‐H stretching), 2850 (C‐H stretching), 1740 (C=O), 1455
(CH2), 1410 (C‐H bending), 1370 (C‐H bending), 1350 (C‐H bending), 1295, 1230
(C‐O stretching), 1150 (C‐O stretching), 1110 (C‐O stretching), 1020 (C‐O
stretching), 935 (C‐O stretching), 760, 720.

Starch‐stearate (Sample 19, Table 4.1., DS = 2.96)
1
H‐NMR (before peracetylation, CDCl3): δ 0.9 (t, 3H, C18), 1.0 (m, broad peaks,
28H, C4‐C17), 1.5 (m, 2H, C3), 2.3 (m, broad peaks, 2H, C2), 3‐6 ppm (m, broad
peaks, 7H, C1S‐6S)
72

1
H‐NMR (after peracetylation, CDCl3): δ 0.9 (t, 3H, C18), 1.3 (m, broad peaks,
28H, C4‐C17), 1.5 (m, 2H, C3), 1.8‐2.6 (m, broad peaks, 3H, C2’), 2.4 (m, 2H, C2),
3‐6 ppm (m, 7H, C1S‐6S)
C‐NMR (before peracetylation, CDCl3): δ 14.0 (C18), 22.7 (C17), 25.0 (C3), 26‐32
13
(C4‐15), 32.0 (C16), 34.2 (C2), 61.4 (broad, C6S), 68‐74 (broad, C2S, 3S, 5S), 75.7
(C4S), 95.5 (broad, C1S), 172‐174 ppm (C=O, attached to O‐C2S, O‐C3S, and O‐
C6S)

FT‐IR (cm‐1): 2920 (C‐H stretching), 2850 (C‐H stretching), 1740 (C=O), 1455
(CH2), 1410 (C‐H bending), 1370 (C‐H bending), 1350 (C‐H bending), 1295, 1150
(C‐O stretching), 1100 (C‐O stretching), 1020 (C‐O stretching), 950 (C‐O
stretching), 865, 760, 720.

6S O R
4S
O 5S O
R O 1S
2S O
O 2 4 6 8 10 12
3S
CH3
1 3 5 7 9 11
O
Starch-Laurate
O
1' CH3 2'
6S O
4S O
O 5S
R O 1S
2S O
O 2 4 6 8 10 12
3S CH3
1 3 5 7 9 11
O
Starch-Laurate
peracetylated
6S O R
4S
O 5S O
R O 1S
2S O
O 2 4 6 8 10 12 14 16 18
3S
CH3
1 3 5 7 9 11 13 15 17
O
Starch-Stearate
Figure 4.1. Numbering scheme for carbon atoms of products

4.2.3.2. Peracetylation procedure
The presence of remaining hydroxyl groups in the products resulted in broad
and overlapping starch resonances in 1H‐NMR spectra [10] and hampered
calculation of the DS. A peracetylation reaction to substitute all of the remaining
73
Chapter 4
hydroxyl groups with acetate groups was applied to obtain reliable DS data.
The peracetylation procedure by Einfeldt et al [11] was applied. Typically, the
starch ester (0.1 g) was suspended in THF (4%‐w/v) and stirred at 55 oC until
the starch was fully dissolved (typically 3 h). Subsequently, the peractylating
reagents (DMAP, acetic anhydride and pyridine in a DMAP : acetic anhydride :
pyridine : AHG molar ratio of 1: 10: 22 : 1) were added. The peracetylation
reaction was conducted for 7 h at 50 oC. The product was precipitated by the
addition of methanol and washed several times with methanol. It was finally
dried overnight in a vacuum oven at 70oC and 5 mbar until constant weight.

4.2.3.3. Determination of the Degree of Substitution (DS)
The DS of the product was calculated using 1H‐NMR spectra of the products
after peracetylation. The DS of the fatty acid esters was calculated from the DS
of the products after peracetylation (eq 4.1.).
DS fatty ester = 3 − DS acetate (4.1.)
The DS of the acetate groups of the products may be calculated by
comparing the unit area of the acetate protons (AH‐acetate) to the unit area of the
starch protons (calculated from the intensity of the starch peaks at δ 3.6‐5.6
ppm). The procedure was described earlier by Elomaa et al [10] and the relevant
equations are shown in equation 4.2. and 4.3.:
AH − acetate
DSacetate = (4.2.)
AH − starch
A3.6−5.6
AH − starch = (4.3.)
7
Determination of the unit area of the acetate H‐atoms (AH‐acetate) is hampered
by peak overlap with the H‐atoms attached to C2 (AC2) of the fatty acid chains (δ
1.8‐2.6 ppm range) and a correction has to be made (eq 4.4.).
A1.8− 2.6 − AC 2
AH − acetate = (4.4.)
3
AC 2 = 2 × AH − fattyester (4.5.)
The AC2 values of both the laurate and stearate side chains were calculated
from the peak intensity in the range δ 0.5‐1.8 ppm (protons attached to the fatty
acid carbons C3‐C12 for laurate and C3‐C18 for stearate) using eq. 4.5., 4.6. and
4.7.:
74

AC 3−C12 A0.5−1.8
AH − fattyester = = (laurate) (4.6.)
21 21
AC 3−C18 A0.5−1.8
AH − fattyester = = (stearate) (4.7.)
33 33
In eq. 4.3.‐4.7., Ax‐y stands for the peak area in the range δ x‐y ppm, while ACx‐Cy
is the area of the H‐atoms attached to carbons in the range Cx‐Cy (carbon
numbering scheme is given in Figure 4.1).

4.3. Results and Discussion
4.3.1. Exploratory Experiments
A number of exploratory experiments were carried out with vinyl‐laurate
and vinyl stearate (vinyl ester: AHG molar ratio of 1 : 3) at 110°C for 24 h in
DMSO using K2CO3 as the catalyst. A schematic representation of the
esterification reaction of starch with the vinyl esters is provided in Scheme 4.1.

OH OH
O
O O H3C O
+ +
R O CH2
OH O
HO HO H
n R n
O
(1) R = laurate (C12)

(2) R = stearate (C18)
Scheme 4.1.

The reaction was performed in two discrete steps. Initially, the starch was
gelatinised in DMSO at 70°C for 3 h to make the hydroxyl groups of starch
more accessible for reaction. Subsequently, the vinylester and the catalyst were
added and the reaction mixture was heated to 110°C. After 2‐3 h, the esterified
starch started to separate from the medium in the form of a gel. After 24 h, the
brownish gel was precipitated with methanol and the product was collected
after vacuum drying in the form of a transparent, light yellow solid. The
products of these exploratory reactions are insoluble in water and DMSO, but
swell in organic solvents such as toluene and THF.
75
Chapter 4
The DS of the products was determined by using NMR (vide infra). When
using a vinyl‐laurate : AHG molar ratio of 3 and K2CO3 as the catalyst, a
product DS of 2.52 was obtained. A reaction with vinylstearate at similar
conditions resulted in a stearate starch ester with a DS of 2.96.

4.3.2. Product Characterisation
4.3.2.1. 1H‐ and 13C‐NMR analyses
The solubility of the products in common NMR solvents (DMSO‐d6 or
CDCl3) is a function of the product DS. Medium DS starch laurate and starch
stearate (1 < DS < 2) dissolve poorly in DMSO‐d6 and CDCl3, even at higher
temperatures (50 oC). Higher DS products have a higher solubility in CDCl3 and
good quality 1H‐ and 13C‐NMR spectra could be obtained (Figure 4.2. and 4.3.).

Figure. 4.2. Typical 1H‐NMR spectrum of:
(a) native starch in DMSO d‐6 at 60oC.
(b) starch laurate, DS = 2.52 (Sample 17, Table 4.1.) in CDCl3 at
50oC.
(c) peracetylated starch laurate, DS = 2.52 (Sample 17, Table 4.1.) in
CDCl3 at 50oC.
For atom numbering scheme: see Figure 4.1.
76

A typical 1H‐NMR spectrum of starch laurate is shown in Figure 4.2. Clearly
visible are the peaks arising from starch and the aliphatic hydrogen atoms of
the fatty acid chain (δ 0.8‐2.5 ppm). The starch peaks (δ 3‐5.5 ppm) are broad
and overlapping [10]. This feature hampers the DS determination by NMR, and
therefore a peracetylation procedure to substitute all of the remaining OH
groups with acetate groups was applied [8, 10‐11]. The 1H‐NMR spectrum of a
typical peracetylated starch laurate is shown in Figure 4.2.c. NMR spectra of the
peracetylated products are considerably improved in terms of peak resolution
and allow a more reliable calculation of the DS. The proton signals of the acetate
methyl group, required for DS determinations, are together with the CH2
groups of the acid chain adjacent to the ester moiety in the range δ 1.8‐2.3 ppm.

Figure 4.3. Typical 13C‐NMR spectra of:
(a) native starch, in DMSO d‐6 at 60oC.
(b) starch laurate, DS = 2.52 (Sample 17) in CDCl3 at 50oC.
(c) starch stearate, DS = 2.96 (Sample 19) in CDCl3 at 50oC.
For atom numbering scheme: see Figure 4.1.

Typical 13C‐NMR spectra of the products are given in Figure 4.3. Clearly
visible are the carbon resonances of the fatty ester chains (δ 10‐35 ppm) and C
atom of the ester group (δ 170‐175 ppm). The resonances arising from the
anhydroglucose unit of starch are broadened. Two of the carbon resonances (1S
and 4S) are considerably shifted compared to native starch. The same
77
Chapter 4
phenomenon was observed by Dicke for starch acetate [12]. The shift of the
starch peaks and the presence of peaks arising from the fatty ester chains clearly
indicate that the esterification reaction with vinyl laurate and vinyl stearate was
successful.

4.3.2.2. FT‐IR measurements
The FT‐IR spectra of starch laurate and starch stearate are shown in Figure
4.4.b. and 4.4.4.c. For comparison, a spectrum of native starch (Figure 4.4.a.) is
also included.
c. starch stearate
C-H C-O
C=O
b. starch laurate C-O

Absorbance
C-H C=O
C-O
a. native starch
-OH
C-H
4000 3500 3000 2500 2000 1500 1000 500

Wavenumber (cm-1)
Figure 4.4. FT‐IR Spectra of starch laurate (DS= 2.52, Sample 17, Table 4.1.),
starch stearate (DS= 2.96, Sample 19, Table 4.1.) and native starch

FT‐IR spectra of both starch laurate and starch stearate (Figure 4.4.b. and
4.4.c.) show characteristic bands of the carbonyl group of the fatty esters in the
1750‐1700 cm‐1 region. In addition, the C‐H stretching vibrations of the alkyl
groups of the fatty ester chain are clearly present at 2920 and 2850 cm‐1.
Characteristic peaks of the polysaccharide backbone are visible in the 1250‐900
cm‐1 region (C‐O stretching) [13]. The near absence of remaining hydroxyl
78

vibrations in the range 3000‐3600 cm‐1 and at 1640 cm‐1 indicates that the DS of
the product is high, in line with the NMR data.

4.2.3. Systematic Studies
The effect of important process variables like the vinyl‐ester to AHG ratio,
type of catalyst and the effect of the addition of co‐solvents on the product DS
was studied in more detail. Most of the experiments (14) were performed using
Na2HPO4 as the catalyst. In addition, four experiments were performed with
two alternative basic catalysts (K2CO3 and Na‐acetate). The results are shown in
Table 4.1.

Table 4.1. Overview of the esterification of starch using vinyl‐esters and
basic catalysts a
Vinyl Ester: Amount of
Exp. Vinyl Ester Catalyst AHG ratio Toluene added DS
[mol/ mol] [ml]
1 laurate Na2HPO4 2 ‐ 1.13
4 laurate Na2HPO4 2 5 0.99
7 stearate Na2HPO4 2 ‐ 1.08
11 stearate Na2HPO4 2 5 1.01
15 laurate K2CO3 3 ‐ 2.52
16 laurate CH3COONa 3 ‐ 2.54
17 stearate K2CO3 3 ‐ 2.96
18 stearate CH3COONa 3 ‐ 2.44
a
All experiments were performed at 24 h at 110°C in DMSO with a catalyst intake of 2 wt% based on
starch.
79
Chapter 4
4.3.3.1. Effect of vinyl ester to AHG ratio on the product DS
The effect of the vinyl‐ester to AHG molar ratio on the product DS was
determined for both types of vinyl‐esters with Na2HPO4 as catalyst (samples 1‐
3, 7‐10). The results are presented in Figure 4.5. The highest DS value was 1.23
for vinyl‐laurate at an intermediate vinyl‐ester‐AHG ratio of 3.

1.5
Degree of Substitution (DS)
0.5
0
0 1 2 3 4 5 6
Vinyl Ester: AHG Starch Ratio [mol/mol]

Figure 4.5. DS of the product as a function of the type of vinyl ester and the
vinyl ester‐starch ratio (24 h reaction time, 110 oC, 2 wt% catalyst
intake on starch).
○ : starch‐laurate □ : starch stearate
(lines for illustrative purposes only)

The DS of the products is a clear function of the vinyl laurate and stearate
ester intakes (Figure 4.5.). The DS values are increasing with higher vinyl‐
ester/AHG molar ratio until a certain maximum. A further increase leads to a
reduction in the DS. This behaviour is likely the result of two opposing effects.
Higher concentrations of vinyl esters are expected to lead to higher
esterification reaction rates. At low to medium vinyl ester/AHG ratios (0‐3), this
positive effect dominates the reaction rate and the DS of the products will
therefore increase at higher vinyl ester intakes. A further increase in the vinyl
ester intake leads to a reduction in the DS. This is likely due to a reduction of
the polarity of the reaction medium. At a ratio of 1:6, the ester intake is equal on
a weight basis to the DMSO intake. This reduced polarity is expected to lead to
80

a lowering of the reaction rates due to a reduction of the solubility and degree
of ionization of the starch reactant as well as the base catalysts. These negative
effects dominate the reaction performance at higher vinyl ester/AHG ratios and
lead to a reduction in the DS values.
When using Na2HPO4 as the catalyst, the starch laurate esters display higher
DS values than the starch stearates. This effect is particularly evident at higher
vinyl ester/ AHG ratios (>3) (Figure 4.5.). Thus, the DS of the product is also a
function of the chain length of the fatty acid, with high carbon numbers leading
to a reduction in the DS. Aburto, et al [14] reported the synthesis of fatty esters
of starch using alkanoyl chloride reactants (C8‐C18) with reactant ratio of 6 mol
alkanoyl chloride/mol AHG. A similar trend in reactivity pattern was observed
in this study and the DS decreased from 1.7 for lauroyl chloride to 0.8 for
stearoyl chloride. The difference was explained by assuming that the reaction
rate is reduced for larger reagents due to steric effects and this explanation
likely also holds for the reactions with the vinylesters [14].

4.3.3.2. Effect of the addition of toluene as a co‐solvent
A number of reactions were performed using a co‐solvent. In this case, the
reactions were initiated in DMSO and toluene was added after 12 h reaction
time to re‐dissolve the poorly soluble partially‐esterified starch products (entry
4‐6, 11‐14 in Table 4.1.). A similar procedure was proposed by Nouvel, et al [15]
for the silylation of starch. Here, the addition of co‐solvents (toluene/ THF) led
to an increase in the DS. These findings were rationalised by assuming that the
co‐solvents increase the solubility of the silylated products, leading to enhanced
reactivity.
The addition of toluene for the esterification of starch with vinyl‐esters
surprisingly did not lead to improved DS values. The products have about the
same DS value for vinyl‐laurate when using only DMSO and even reduced DS
values for vinyl‐stearate (see Table 4.1.). Although toluene may positively effect
the reaction by (partly) re‐dissolving starch ester precipitates, it also results in a
dilution of the reaction mixture and a reduction in the polarity. The latter
factors appears to have a strong effect on reaction rates (vide supra), with
reductions in polarity leading to lower reaction rates.

4.3.3.3. Catalysts screening
A number of alternative basic catalysts for Na2HPO4, i.e. K2CO3 and Na‐
acetate were tested. The results are given in Table 4.1. and illustrated in Figure
4.6. It is clear that Na‐acetate and K2CO3 are considerably more active than
Na2HPO4 and products with a significantly higher DS were obtained. For starch
81
Chapter 4
laurate esterification, the two catalysts are equally effective and products with a
DS of about 2.5 were obtained. For starch stearate, K2CO3 gave products with a
significantly higher DS (2.96) compared to Na‐acetate (DS=2.44). Thus, the DS of
the product is also tunable by proper catalyst selection.
The DS of the laurate ester when using Na‐acetate is higher than for the
stearate ester (Table 4.1. and Figure 4.6.), in line with the findings for NaH2PO4.
However, when using K2CO3 as the catalyst, the DS for the laurate ester is lower
than the stearate ester. Apparently, the statement that the DS for the laurate
esters is always higher than for the stearate esters is not generally valid and a.o.
function of the type of catalyst.
2.5
2
Degree of Substitution
1.5
0.5
0
Laurate Stearate
Vinyl-Ester

Figure 4.6. Comparison of DS values with different catalysts (Vinylester:

AHG molar ratio = 3:1, catalyst amount = 2%‐w, 110oC, DMSO).
gray: Na2HPO4 black: K2CO3 white: Na‐acetate

4.4. Conclusions
A study on the synthesis of corn starch fatty acid esters with high DS values
is reported. The products were synthesised in DMSO using vinyl‐esters in the
presence of basic catalysts (Na2HPO4, K2CO3, and Na‐acetate). The yellow
products were characterized by 1H‐ and 13C‐NMR, and FTIR and confirm the
presence of chemically bound fatty acid chains. The DS of the products is a clear
function of the chain length of the fatty ester and the type of catalyst. K2CO3 and
Na‐acetate are superior with respect to activity when compared with Na2HPO4.
82

With these catalysts, products with a DS > 2.4 could be obtained for both laurate
and stearate esters.
The DS of the products may also be tuned with the vinyl ester/AHG molar
ratio. At low vinyl ester/ AHG ratio, the DS of the product increases at higher
vinyl ester intakes. A maximum was observed at a vinyl ester: AHG ratio
between 2 and 4. Higher ratios led to a reduction in the DS, presumably due to
a reduction of the polarity of the reaction medium. Important product
properties will be described in the next chapter.

4.5. Nomenclature
A : peak area of certain proton in 1H‐NMR spectra [‐]
DS : Degree of Substitution, average value of mole of substituted –OH
per mole of anhydroglucose (AHG) units [‐]
DS fatty ester : DS of fatty ester (laurate or stearate) group substituents [‐]
DSacetate : DS of acetate group substituents after peracrtylation [‐]
4.6. References
[2]. J.W. Mullen and E. Pacsu: Starch Studies: Preparation and Properties of
Starch Triesters. Ind. Eng. Chem. 1942, 34, 1208‐1217.
[3]. M. Bengtsson, K. Koch, and P. Gatenholm: Surface octanoylation of high‐
amylose potato starch films. Carbohydr. Polym. 2003, 54, 1‐11.
[4]. Y. X. Xu, Y. Dzenis, M. A. Hanna: Water solubility, thermal characteristics
and biodegradability of extruded starch acetate foams. Ind. Crops Prod.
2005, 21, 361‐368.
[6]. A. D. Sagar, E. W. Merrill: Properties of Fatty‐Acid Esters of Starch. J. Appl.
Polym. Sci. 1995, 58, 1647‐1656.
83
Chapter 4
various starches with lauric acid methyl ester and triacyl glycerides.
Starch‐Starke 2005, 57, 145‐152.
[8]. W. Mormann, M. Al‐Higari: Acylation of starch with vinyl acetate in
water. Starch‐Starke 2004, 56, 118‐121.
[9]. B. Laignel, C. Bliard, G. Massiot, J. M. Nuzillard: Proton NMR
spectroscopy assignment of D‐glucose residues in highly acetylated starch.
Carbohydr. Res. 1997, 298, 251‐260.
[10]. M. Elomaa, T. Asplund, P. Soininen, R. Laatikainen, S. Peltonen, S.
Hyvarinen, A. Urtti: Determination of the degree of substitution of
acetylated starch by hydrolysis, H‐1 NMR and TGA/IR. Carbohydr. Polym.
2004, 57, 261‐267.
[11]. L. Einfeldt, K. Petzold, W. Gunther, A. Stein, M. Kussler, D. Klemm:
Preparative and H‐1 NMR investigation on regioselective silylation of
starch dissolved in dimethyl sulfoxide. Macromol. Biosci. 2001, 1, 341‐347.
[12]. R. Dicke: A straight way to regioselectively functionalized polysaccharide
[13]. J.F. Mano, D. Koniarova, R.L. Reis: Thermal properties of thermoplastic
starch/synthetic polymer blends with potential biomedical applicability. J.
Mater. Sci., Mater. Med. 2003, 14(2) 127‐135.
[14]. J. Aburto, H. Hamaili, G. Mouysset‐Baziard, F. Senocq, I. Alric, and E.
Borredon: Free‐solvent Synthesis and Properties of Higher Fatty Esters of
Starch – Part 2, Starch‐Starke 1999, 51, 302‐307.
[15]. C. Nouvel, I. Ydens, P. Degee, P. Dubois, E. Dellacherie, J. L. Six: Partial or
total silylation of dextran with hexamethyldisilazane. Polymer 2002, 43,
1735‐1743.
84
Chapter 5
Experimental and Modeling Studies on the Synthesis
and Properties of Higher Fatty Esters of Corn Starch

Abstract
This chapter describes a systematic study on the synthesis of higher fatty esters of
corn starch (starch‐laurate and starch‐stearate) by using the corresponding vinyl
esters. The reactions were carried out in DMSO using basic catalysts (Na2HPO4,
K2CO3, and Na‐acetate). The effect of the process variables (vinyl ester to starch
ratio, catalyst intake, reaction temperature and type of the catalyst) on the Degree
of Substitution (DS) of the starch laurate and starch stearate esters was determined
by performing a total of 54 experiments. The results were adequately modeled
using a non‐linear multivariable regression model (R2≥0.96). The basicity of the
catalyst and the reaction temperature have the highest impact on the product DS.
The thermal and mechanical properties of some representative product samples
were determined. High‐DS products (DS= 2.26‐2.39) are totally amorphous
whereas the low‐DS ones (DS= 1.45‐1.75) are still partially crystalline. The thermal
stability of the esterified products is higher than that of native starch. Mechanical
tests show that the products have tensile strength (stress at break) between 2.7‐3.5
MPa, elongation at break of 3‐26%, and modulus of elasticity of 46‐113 MPa.

Keywords: starch esters, trans‐esterification, vinyl laurate, vinyl stearate

Chapter 5
5.1. Introduction
Plastics are very useful materials and are used in large amounts (245 million ton
per annum) [1‐2]. The largest application area of plastics is the use as packaging
material (37%) [1]. However, most of the plastic waste is not biodegradable, and
this causes serious environmental problems [1, 3‐4]. The development of novel
biodegradable plastic materials capable of decomposing when given an
appropriate environment and time is of utmost importance [5].
Starch is an attractive feedstock for the synthesis of novel biodegradable
plastics. It is cheap and abundantly available [3]. The global starch production was
estimated at 60 million tons in 2004 [6]. The use of native starch as a building block
for novel biodegradable polymers, however, is rather limited to date. Examples
are agricultural mulch [7], packaging material, and food service‐ware [8]. The
limited use is mainly because of a number of unfavourable properties of starch
such as low moisture resistance, high brittleness, and incompatibility with
hydrophobic polymers [9]. Chemical or physical modification is required for
successful applications as biodegradable polymers. Typical examples are blending
with polyvinyl alcohol [3], polyethylene [7] and polycaprolactone [10] and
chemical modification by esterification with organic acids [11‐19].
Esterification is one of the oldest methods used to improve starch properties
[11]. Most early studies mainly focused on the use of short‐chain carboxylic acid
(C1‐C4), and particularly on the synthesis of starch acetate (C2) [11‐12]. Mullen
and Pacsu [11, 12] studied the synthesis and properties of of C1‐C6 esters of starch.
The mechanical properties of products with longer ester chains (C4 and C6)
products and plasticized acetate esters were considerably improved compared to
native starch. Sagar and Merill [13] studied the synthesis and properties of C4‐C6
esters of high amylose starch. The products were biodegradable, however, the
mechanical properties were still not satisfactorily. Thiebaud, et al [14] and Aburto,
et al [15‐19] synthesized longer chain fatty esters (C8‐C18) of potato starch and
corn amylose using fatty acid chlorides and pyridine. The high DS esters showed
interesting properties. The products were hydrophobic and the mechanical
properties were considerably improved compared to native starch. Alternative
routes to avoid the use of acid‐chloride/pyridine combinations have been
developed. The use of methyl and glyceryl esters to prepare starch laurate (C12)
ester was investigated but only yielded products with relatively low DS values
(max. 0.65) [19]. The use of vinylesters has also been explored [20‐22]. However,
the research activities were mainly limited to the use of vinylacetate. Two
examples using higher vinyl esters (vinyl laurate) were reported by Mormann, et
al [20]. Unfortunately, the physicochemical properties of the products were not
mentioned.
86

Experimental and Modeling Studies on the Synthesis and Properties of Higher Fatty Esters of Corn Starch

We have recently performed exploratory studies on the synthesis of starch
esters of higher fatty acids using vinyl laurate and vinyl stearate as the reagents
[23]. Long fatty esters of corn starch with a broad range in degree of substitution
(DS= 0.24 ‐ 2.96) were prepared by reacting the starch with vinyl laurate or vinyl
stearate in the presence of basic catalysts in DMSO. This chapter describes
systematic studies using Design of Experiments (DOE) to gain insights in the effect
of process variables (temperature, vinyl ester to anhydroglucose ratio, catalyst
type and intake) on the DS of the products. The experiments were modeled using
non‐linear multivariable regression. In addition, the thermal and mechanical
properties of representative examples of the highly hydrophobic materials are
described and discussed.

5.2.1. Materials
Corn starch (approx. 73% amylopectine and 27% amylose) was purchased from
Sigma (Germany). It was dried for 48 hour at 105 oC under vacuum (~1 mbar) to
reduce the moisture content below 2 % before use. Analytical grade vinyl stearate
(Aldrich, Japan), vinyl laurate (Fluka, Germany), and acetic anhydride (Merck,
Germany) were used without further purification. The catalysts for the
esterification reaction were analytical grade potassium carbonate (Boom, the
Netherlands), sodium acetate, and disodium hydrogenphosphate dodecahydrate
(both Merck, Germany). Technical grade dimethyl sulfoxide (DMSO), 4‐N,N‐
dimethylaminopyridine (DMAP), and tetrahydrofuran (THF) were supplied by
Acros (Belgium). Analytical grade methanol, pyridine, and toluene were obtained
from Labscan (Ireland). All these chemicals were used as received.

5.2.2. Analytical Equipment
H‐ and 13C‐NMR spectra were recorded in CDCl3 on a 400 MHz Varian AMX
1
NMR machine. The spectra were recorded at 50 oC, as recommended by Laignel et
al [24]. TGA measurements were performed using a Perkin Elmer TGA 7
Thermogravimetric Analyzer. The samples were heated to 700 oC in a nitrogen
atmosphere with a heating rate of 10°C/min. DSC analyses were performed on a
TA Instruments DSC 2920. The samples (about 10 mg) were placed in sealed
aluminum cells. After a first heating run from room temperature up to 200 oC to
delete the thermal history of the material, each sample was cooled to ‐50°C and
then heated again to 200 oC (heating rate 10 oC/min). The thermal properties (Tm,
Tc, and Tg) of each sample were determined from the spectra related to the cooling
run and the second heating one. T‐Bones samples (with thickness of 2 mm) for
determination of the tensile properties were prepared using a melt press
87
Chapter 5
apparatus (Fontijne, Holland), operated at 150 oC and 150 bar for 3 minutes. The

tensile tests were performed using an Instron Series IX Automated Materials
Testing System 1.09 at 20oC and a crosshead speed of 30 mm/min.

5.2.3. Methods
5.2.3.1. Typical example of the preparation of laurate and stearate starch esters
The corn starch (0.5 g) was gelatinized in DMSO (5 ml) at 70oC for 3 h. After the
mixture became a homogenous, colourless solution, vinyl laurate or vinyl stearate
(3‐5 mol/mol with respect to the AHG units) and the catalyst (potassium
carbonate, or sodium acetate, or disodium hydrogenphosphate, 2‐5%‐w with
respect to starch) were added. The reactor content was kept at 80‐110oC for 24 h.
The product was precipitated using methanol (100 ml), and the liquid phase was
removed by decantation. This precipitation‐decantation procedure was repeated
twice using 50 and 25 ml of methanol, respectively, to purify the reaction product.
Finally, the product was dried in a vacuum oven (70oC, 5 mbar) for 24 h or until
constant weight.

5.2.3.2. Peracetylation procedure and Degree of Substitution (DS) determination
The presence of un‐substituted hydroxyl groups in the products resulted in
broad and overlapping starch resonances in 1H‐NMR spectra. A peracetylation
reaction to substitute all of the remaining hydroxyl groups with acetate groups
was therefore applied to obtain reliable DS values. Typically, the starch ester (0.1
g) was added to THF (4%‐w/v) and stirred at 55 oC until dissolution (typically 3 h).
Subsequently, the peracetylating reagent (1 mol DMAP, 10 mol acetic anhydride,
and 22 mol pyridine per mol AHG units) were added. The peracetylation reaction
was conducted for 7 h at 50 oC. The product was precipitated by the addition of
methanol and washed several times with methanol before finally dried overnight
in a vacuum oven at 70oC and 5 mbar. The DS of the products was calculated
using a procedure given in previous work [23].

5.2.4. Experimental Design
The experiments were carried out in 6 blocks. The three variables used in each
block were the vinyl ester to AHG starch mol ratio ( x1 ), catalyst intake ( x 2 ), and
reaction temperature ( x3 ). All experiments in a block were conducted using a 3‐
variable, 2‐level Full‐Factorial Experimental Design with one center point, giving a
total of 23+1 = 9 experiments per block. For each block, a given type of catalyst
(Na2HPO4, K2CO3 or Na‐acetate) and vinyl ester (either vinyllaurate or
vinylstearate) was applied. This gave a total of 54 experiments (2 types of vinyl
88


esters x 3 types of catalysts x 9 experiments per block). In a later stage, the type of
catalyst was quantified using the pKb of the catalyst ( x 4 ). The ranges for the
individual variables ( x1 ‐ x 4 ) are shown in Table 5.1.
The mathematical analysis of the experimental data was performed with the
software package “Matchad 13” (Mathsoft). The response (y, DS of the products)
was modelled using the following expression:
y = β 0 + ∑ β i xi + ∑ β ii xi2 + ∑∑ β jk x j x k (5.1.)
i i j k≠ j
Here, βi, βii and βjk are the regression coefficients obtained by a multiple
2
regression procedure. One of the terms, namely x 4 , caused singularity of the
matrix used for the modelling, and was therefore excluded. A t‐statistic was used
to rank the individual regression coefficients (βi, βii, or βjk) according to their
relative importance [25].
An analysis of variance (ANOVA) was performed to check the adequacy of the
model. The applied procedure is well described in the literature [25] and consists
of calculating the sum of squares (SS) for the model and the error, together with
the total sum of squares. In combination with the relative degrees of freedom (DF)
it is possible to calculate the mean square (MS) for the model and the
corresponding error. On the basis of the latter values, the F‐value for the model is
calculated. With this information the P‐value for the model is determined. The
latter value is related to the statistical significance of the model.

5.3. Results and Discussion
A schematic representation of the esterification reaction of starch with a vinyl
ester is provided in eq. 5.2. The starch was gelatinized before the addition of the
vinylester to make the starch OH groups more accessible for reaction. The
reactions were carried out in DMSO for 24 h using three different basic salts
(Na2HPO4, K2CO3, or Na‐acetate) as the catalysts. The products of the reaction
were brownish gels which became lighter in colour after product precipitation and
washing with methanol. After drying, the products were isolated as transparent,
yellowish solids. The products were soluble in organic solvents such as toluene
and THF.
89
Chapter 5
OH OH
O
O O H3C O
+ +
R O CH2
OH O
HO HO H
n R n
O
(1) R = laurate (C12)

(2) R = stearate (C18)
(5.2.)
The products were characterized by 1H‐NMR in CDCl3. The peaks of starch (δ 3‐
5.5 ppm) and aliphatic hydrogen atoms of the fatty acid chain (δ 0.8‐2.5 ppm) were
clearly present. The starch peaks were broad and overlapping, and only after the
starch esters were peracetylated, the resolution of these peaks was considerably
improved. The DS of the products was determined by using 1H‐NMR spectra of
the peracetylated starch esters.
The effect of three reaction parameters (molar ratio of vinyl ester to AHG units
of starch x1 , catalyst intake x2 , and reaction temperature x3 ) on the DS of starch
was studied using a full factorial experimental design with one center point. The
ranges of the values of the independent variables ( x1 ‐ x3 ) are shown in Table 5.1.
An additional variable x 4 (related to the basicity of the catalyst) was also included
in order to obtain a general model for starch esterification. A total of 54
experiments were conducted.

Table 5.1. Experimental Design Variables
Independent Variables
Additional
Full Factorial Variables
Variable Variable
Level
Vinyl ester to AHG‐ Catalyst Reaction Catalyst
starch ratio [mol/mol] intake [%‐w]a temperature [°C] basicity (pKb)b
x1 x2 x3 x4
Low (‐1) 3 2.0 80 K2CO3 (3.66)
Middle (0) 4 3.5 95 Na2HPO4 (6.8)
High (+1) 5 5.0 110 CH3COONa (9.25)
a.
in %‐wt based on starch intake, in the model (eq. 5.3.) it is transformed to 105 x catalyst mol amount.
b.
taken from ref [26].

The results of the experiment are given in Table 5.2.
90


Table 5.2. Overview of experimental data
Vinyl : AHGa Catalyst Temperature

Experiment Vinyl Ester Catalyst DS
[mol/mol] [%‐w] [˚C]
1 laurate Na2HPO4 3.00 2.00 80 0.62
2 laurate Na2HPO4 3.00 5.00 80 0.75
3 laurate Na2HPO4 3.00 2.00 110 1.46
4 laurate Na2HPO4 3.00 5.00 110 1.12
5 laurate Na2HPO4 5.00 2.00 80 0.27
6 laurate Na2HPO4 5.00 5.00 80 0.87
7 laurate Na2HPO4 5.00 2.00 110 1.27
8 laurate Na2HPO4 5.00 5.00 110 1.32
9 laurate Na2HPO4 4.00 3.50 95 1.11
10 laurate CH3COONa 3.00 2.00 80 2.28
11 laurate CH3COONa 3.00 5.00 80 2.44
12 laurate CH3COONa 3.00 2.00 110 2.54
13 laurate CH3COONa 3.00 5.00 110 2.78
14 laurate CH3COONa 5.00 2.00 80 2.24
15 laurate CH3COONa 5.00 5.00 80 2.42
16 laurate CH3COONa 5.00 2.00 110 2.59
17 laurate CH3COONa 5.00 5.00 110 2.67
18 laurate CH3COONa 4.00 3.50 95 2.17
19 laurate K2CO3 3.00 2.00 80 2.23
20 laurate K2CO3 3.00 5.00 80 2.52
21 laurate K2CO3 3.00 2.00 110 2.52
22 laurate K2CO3 3.00 5.00 110 2.94
23 laurate K2CO3 5.00 2.00 80 2.54
24 laurate K2CO3 5.00 5.00 80 2.88
25 laurate K2CO3 5.00 2.00 110 2.72
26 laurate K2CO3 5.00 5.00 110 2.84
27 laurate K2CO3 4.00 3.50 95 2.50
28 stearate Na2HPO4 3.00 2.00 80 0.48
29 stearate Na2HPO4 3.00 5.00 80 0.47
30 stearate Na2HPO4 3.00 2.00 110 1.53
31 stearate Na2HPO4 3.00 5.00 110 1.35
32 stearate Na2HPO4 5.00 2.00 80 0.07
33 stearate Na2HPO4 5.00 5.00 80 0.09
34 stearate Na2HPO4 5.00 2.00 110 0.96
35 stearate Na2HPO4 5.00 5.00 110 0.93
36 stearate Na2HPO4 4.00 3.50 95 0.27
37 stearate CH3COONa 3.00 2.00 80 1.70
91
Chapter 5

46 stearate K2CO3 3.00 2.00 80 2.64
47 stearate K2CO3 3.00 5.00 80 2.21
48 stearate K2CO3 3.00 2.00 110 2.96
49 stearate K2CO3 3.00 5.00 110 2.64
50 stearate K2CO3 5.00 2.00 80 2.24
51 stearate K2CO3 5.00 5.00 80 2.55
52 stearate K2CO3 5.00 2.00 110 2.59
53 stearate K2CO3 5.00 5.00 110 2.90
54 stearate K2CO3 4.00 3.50 95 2.41
a.
mol ratio of the vinyl ester to AHG units of starch

5.3.1. Mathematical Modeling
Modelling of the DS data for both vinylesters was performed using non‐linear
multivariable regression based on the 4 independent variables (Table 5.1.). The
type of catalyst was quantified using the basicity constant (pKb) in water [26‐27].
The experimental data for both starch laurate and starch stearate esterification are
best described with a quadratic model including interaction terms (eq. 5.3.).The
modelled values for βij for the two esters are given in Table 5.3.
DS = β 0 + β1 x1 + β 2 x 2 + β 3 x3 + β 4 x 4 + β 5 x1 x 2 + β 6 x1 x3 + β 7 x1 x 4 + β 8 x 2 x3 + β 9 x 2 x 4

+ β10 x3 x 4 + β11 x 22 + β12 x32 + β13 x 42
(5.3.)

Table 5.3. Values for the coefficients of the DS model for starch laurate and
stearate
Coefficient Starch Laurate Starch Stearate
β0 9.1748 17.1082

β1 0.2183 ‐0.1104
β2 0.0732 ‐0.0857
β3 ‐0.0327 ‐0.1568
β4 ‐2.5650 ‐2.6939
β5 0.0027 0.0120
β6 ‐0.0008 0.0010
β7 ‐0.0249 ‐0.0470
β8 ‐0.0006 0.0001
β9 ‐0.0066 0.0017
β10 0.0016 0.0023
β11 0.0011 0.0009
β12 0.0003 0.0008
β13 0.1979 0.1947
92


The result of the analysis of variance is given in Table 5.4. The very low P‐
values indicate that the models are statistically significant. The R2 values for the
models are 0.970 (starch laurate) and 0.967 (starch stearate), respectively and it
may be concluded that the models describe the experimental data well. The parity
plots of both models are given in Figure 5.1. and confirm this statement.

Table 5.4. Analysis of variance for DS models for starch esterification
Starch Laurate
SS DF MS F P‐value R2 values

Model 16.3399 13 1.2569 35.2903 1.96 x 10‐8 R2 0.970
Error 0.4986 14 0.0356 R2adjusted 0.945
Total 16.8386 27 R2press 0.875
Starch Stearate
SS DF MS F P‐value R2 values

Model 20.9531 13 1.6118 27.629 1.026 x 10‐7 R2 0.967
Error 0.8167 14 0.0583 R2adjusted 0.939
Total 21.7699 27 R2press 0.805
3 3
2.5 2.5
2 2
DS predicted
DS predicted
1.5 1.5
1 1
0.5 0.5
0 0
0 0.5 1 1.5 2 2.5 3 0 0.5 1 1.5 2 2.5 3
DS actual DS actual
a. starch laurate b. starch stearate

Figure 5.1. Parity plots of the DS models for starch laurate and stearate

Moreover, the adjusted‐R2 values are very close to the R2 ones, which indicate
[25] that all significant variables are included in the model. Finally, we also
performed a PRESS analysis [25] (see corresponding R2 values in Table 5.4.), which
represents an “internal” validation method for the model. Also in this case the
reasonable R2 values (0.805‐0.875) indicate that the model correctly predict the
93
Chapter 5
products DS as function of the process variables within the range of experimental
variables.
To evaluate the effect of each variable on the product DS, two plots showing the
dependence of the DS on reaction temperature and catalyst amount (in %‐w), for
both starch laurate and starch stearate are given in Figure 5.2. Of all the variables
studied, the type of catalyst has the largest effect on the product DS for both the
laurate and stearate esters (Figure 5.2.). The highest product DS values were
obtained using K2CO3. Catalyst performance of CH3COONa was slightly less than
for K2CO3 whereas the lowest DS products were produced when using Na2HPO4
as the catalyst. These results may be rationalised by considering the role of the
catalyst in the modification reaction. It is assumed that the first step in the reaction
sequence is activation of the starch OH groups by deprotonation by a base [21, 22].
The resulting anion will react with the vinyl ester to from the product. On the
basis of the sequence, it can be rationalised that the rate of the reaction (and thus
the product DS) will be higher when using a stronger base. This is indeed the case
when comparing the performance of K2CO3 with CH3COONa. K2CO3 is a stronger
base (pKb 3.66) than CH3COONa (pKb 9.25) and this leads to higher product DS
values for K2CO3. However, the performance of Na2HPO4 does not follow this
trend. The pKb for the latter (6.8) is intermediate between that of the other two
catalysts, whereas catalyst performance is considerable lower. A similar trend was
observed by Dicke [27] for the acetylation of Hylon VII starch using vinylacetate.
The product DS for Na2HPO4 was considerably lower (1.00) than for K2CO3 (2.18)
and CH3COONa (1.82). A possible explanation of this peculiar behaviour of
Na2HPO4 is the fact that the pKb values in water are used for quantification
whereas the actual solvent for the reaction is DMSO. Unfortunately, the base
strengths of the catalysts in DMSO are not known. Another explanation may be
related to the regio‐chemistry of the reaction. Dicke [27] showed that Na2HPO4 has
a strong tendency to selectively deprotonate the OH group at the C2 position of
starch leading to C2 substituted acetate esters. This was not the case for alkaline
catalysts, such as carbonate or acetate salts leading to higher DS esters. The
explanation for the high preference of C2 substitution for the Na2HPO4 catalyst is
not yet known and needs to be established by mechanistic studies.

94


a. starch laurate

b. starch stearate
Figure 5.2. 3D contour plot of the DS as a function of reaction temperature and
catalyst intake (at constant vinylester to AHG mol ratio of 4)

Besides the type of catalyst, the catalyst intake and reaction temperature also
affect the product DS (Figure 5.2.), although to a lesser extent. As expected and in
95
Chapter 5
line with studies on starch acetates [27], higher reaction temperatures and catalyst
intakes lead to higher product DS values. Of all variables studied, the vinyl ester
to AHG ratio has the smallest effect on the product DS.

5.3.2. Product Properties
The thermal and mechanical product properties for two starch‐laurate and
stearate samples with medium (DSlaurate = 1.75, DSstearate = 1.45) and high DS values
(DSlaurate = 2.26, DSstearate = 2.39) were determined. Representative DSC spectra for
the starch stearate are given in Figure 5.3. and the results for all samples are
summarised in Table 5.5.

heating
Tm = 21 oC
Exothermic
cooling
Endothermic
Tc = 12 oC
-20 -10 0 10 20 30 40
o
Temperature ( C)

Figure 5.3. DSC analysis of starch stearate esters
( ) DS 1.45 (. . .) DS ( 2.39)

96


Table 5.5. Melting point (Tm) and crystallization temperature (Tc) of
representative starch laurate and stearate samples
Starch ester DS Tm [˚C] Tc [˚C]

Starch laurate 2.26 n.d.a n.d.a
Starch laurate 1.75 27 22
Starch stearate 2.39 n.d.a n.d.a
Starch stearate 1.45 21 12
a.
n.d.= not detectable

The thermal behavior of the products is a strong function of the DS. In
particular the behaviour at relatively high temperatures (>0 oC) is further
discussed. At moderate DS values the products still display a melting temperature
(at 20‐30 oC) and a crystallization temperature (at 10‐25 oC). These values are
reduced considerably compared to virgin corn starch, which is known to contain
crystalline areas, especially in the amylopectin part [15, 22]. Unfortunately the
thermal transitions for the virgin material lie above the degradation temperature
and can therefore not be accurately determined [2]. In any case, esterification of
starch with vinyl laurate and stearate reduces the crystallinity of the starch
considerably and leads to a lowering of the transition temperatures. Thermal
transitions at relatively high temperature (melting, crystallization, or glass
transition above 0 oC) were even absent for the high DS products, implying that
these products are fully amorphous. To the best of our knowledge, these are the
first examples of completely amorphous starch esters.
Changes in the transition temperatures of native starch by esterification have
been reported in the literature [13‐14, 17]. For instance a Tm of 32 oC was reported
for a potato starch stearate ester with a DS of 1.8 [17], close to the value obtained in
our study for the medium DS starch stearate (21 °C). The melting point of the
potato starch stearate was close to that of model compounds for the stearate side
chains (methyl stearate, 40‐42 oC; octadecane, 28‐30 oC) and the authors concluded
that the transition temperatures are solely determined by the side chain without
any significant contribution of the starch backbone. However, in the present work,
the Tm for the medium DS starch laurate (27 oC) is much higher than the
corresponding side chain model compounds (‐10 oC for dodecane and 5 oC for
methyl laurate). This clearly indicates that the observed thermal transitions are not
only determined by the side chains but are an interplay between that of the starch
backbone and ester side chains.
TGA analysis was performed to study the thermal degradation behavior of the
starch laurate and stearate samples (Figure 5.4.).
97
Chapter 5
100
80
%-weight
60
40
20
0
20 100 200 300 400 500 600 700
o
Temperature ( C)

Figure 5.4. TGA analysis of various starch samples.
( ): native starch
( ): starch laurate (DS 2.26)
(. . .): starch stearate (DS 2.39)

The TGA (Figure 5.4.) curves clearly show starch esterification with either
laurate or stearate results in products with enhanced thermal stability. Native corn
starch degrades at lower onset temperatures (250‐300 oC) than the corresponding
esters (300‐350 oC). The difference in thermal stability between the laurate and
stearate sample is limited. Similar trends were reported for amylose octanoate‐
stearate‐laurate esters [17] and potato starch octanoate‐ laurate esters [14].
The mechanical properties (stress at break, elongation at break and elasticity
modulus) of the products were determined (Table 5.6.).
98


Table 5.6. Mechanical properties of the starch esters
Stress at break Elongation at break Modulus

Starch Ester DS
[MPa] [%] [MPa]
Starch laurate 2.26 3.5 ± 1.0 26.1 ± 15.2 59.9 ± 14.1
Starch laurate 1.75 3.0 ± 2.2 8.7 ± 3.3 82.7 ± 4.0
Starch stearate 2.39 2.7 ± 1.0 3.1 ± 1.3 112.7 ± 14.9
Starch stearate 1.45 3.2 ± 0.1 21.7 ± 4.1 46.3 ± 2.8

All materials generally show tensile strengths between 2.7‐3.5 MPa, elongation
at break between 3‐26%, and modulus of elasticity between 46‐113 MPa. For starch
laurate, an increase in product DS results in lower tensile strengths and modulus
of elasticity, but higher elongation at break. This is in agreement with previous
studies regarding the mechanical properties of starch esters with different chain
length of the fatty acid moiety [13, 17]. An inverse behaviour is, however,
observed for the high DS starch stearate sample. This product is more rigid
compared to the product with medium DS value (higher tensile strength and
elasticity modulus, and lower elongation at break).
A deeper understanding of the mechanical behavior can be gained by
investigating the shape of the stress‐strain curves (typical examples in Figure 5.5.).
Both starch laurate samples display, independently of the DS values, a clear plastic
behavior with a maximum in the stress‐strain curves. The values of the yield
strain, calculated as the ratio between the stress at yield (σy, determined from the
stress‐strain curves as the stress at which the material ceases to be linearly elastic)
and the modulus, is roughly 0.05 for both laurate samples. This indicates that their
mechanical behavior is comparable to that of typical engineering polymers [28].
Starch stearate with a relatively low DS (1.45) displays a very similar behavior
with respect to the laurate samples (i.e. plastic deformation and yield strain of
about 0.05). On the other hand, starch stearate at relatively high DS values (2.39)
does not show any plastic behavior but only an elastic one. The corresponding
yield strain (roughly 0.025) lies still in the typical range of engineering polymers
but also close to that typical of composite materials [28], thus indicating a
relatively more rigid material with respect to all other samples. These
considerations strongly point out the fact that the mechanical behavior can be
coarsely (plastic vs elastic behavior) but also finely (stress and strain at break but
also modulus values) tuned by the chemical structure of the fatty acid chains as
well as the DS values.

99
Chapter 5
4
stress (MPa)
0 5 10 15 20 25 30 35
strain (%)

Figure 5.5. Stress‐strain curves for starch esters.
( ): starch laurate (DS 1.75) ( ): starch laurate (DS 2.26)
(. . .): starch stearate (DS 1.45) (‐ . ‐): starch stearate (DS 2.39).

Although there are many differences in the synthetics methods as well as in
testing conditions and procedures, a rough comparison can be made between the
mechanical properties of the esters described in this chapter and those for reported
for related starch esters synthesized with alkanoyl chloride as reagent and
pyridine as catalyst [17] (Table 5.7.).

100


Table 5.7. Comparison of the mechanical properties of corn starch ester
(this study) with potato starch ester [17]
Corn Starch Potato Starch Corn Starch Potato Starch

Laurate Ester Laurate Ester Stearate Ester Stearate Ester
Property
Low High Low High Low High Low High
DS DS DS DS DS DS DS DS
(1.75) (2.26) (2.7) (1.45) (2.39) (1.8) (2.7)
Stress 3.0 3.5 0.7 3.2 2.7 3.7 1.9

n.a.a.
at Break [MPa] ±2.2 ±1.0 ±0.4 ±0.1 ±1.0 ±0.6 ±0.3
Elongation 8.7 26.1 1500 21.7 3.1 9 10
n.a.a.
at Break [%] ±3.3 ±15.2 ±8.6 ±4.1 ±1.3 ±2 ±2
a.
n.a.= not available

Inspection of the mechanical properties of the starch esters as shown in Table
5.7. confirms that the corn starch esters synthesized in our research are relatively
rigid materials as compared to the potato starch esters. In this respect the
difference in stress and elongation at break for the high‐DS starch laurate sample
is striking. While for corn starch (DS 2.26) relatively high stress at break (3.5 MPa)
is coupled with a low elongation (26 %), for potato starch exactly the opposite is
observed: relatively low stress (0.7 MPa) coupled with a significant (1500%)
elongation. The same considerations are valid for starch stearate at high DS values
(higher stress and lower elongation when using corn starch instead of potato one).
On the other hand at low DS values comparable stress at break (3.2 MPa for corn
starch esters, 3.7 MPa for potato starch ones) are coupled with significant
differences in the elongation (22% for corn starch esters, 9% for potato starch). In
agreement with our own data, an unexpected behavior was also observed for the
mechanical properties of the starch esters from potato starch. Their medium or
high DS stearate esters of potato starch are more rigid compared to the starch
esters from lower chain fatty acids (octanoate and laurate) [17]. They related this
inverse property of starch stearate to crystallization of C18 side chains, as
confirmed by DSC. Here, we show that although no crystallization of C18
occurred (see DSC data of the high DS material in Table 5.5.), the high DS material
is still more rigid, merely because of the structural property of the C18 side chain
with respect to the starch backbone.
Comparison with the literature data implies that the mechanical behaviour of
starch esters of higher fatty acids is also a clear function of the type of starch used.
Moreover, this comparison also confirms our conclusion derived from the analysis
of thermal properties (vide supra): it is not only the individual factors (kind of
starch and fatty acid, DS) which mainly determines the thermal and mechanical
behaviour, but the interplay between those factors.
101
Chapter 5
5.4. Conclusions
A systematic study, including statistical modelling, on the synthesis of corn
starch esters with long chain fatty acids is described. The starch esters were
synthesized by reacting gelatinized starch with vinyl laurate or vinyl acetate in
DMSO in the presence of basic catalysts (Na2HPO4, K2CO3, and Na‐acetate).
Statistically adequate (R2 ≥0.96 and P‐value of ≤10‐7) second‐order mathematical
models correlating the effect of process variables (vinyl to AHG‐starch mol ratio,
reaction temperature, catalyst intake, and catalyst basicity) to the DS of the starch
ester products were developed. The DS of the products is a strong function of the
basicity of the catalyst. The use of K2CO3 and CH3COONa catalysts resulted in
medium‐high DS products (2.1‐2.9 for starch laurate, 1.4‐3 for starch stearate),
while the use of Na2HPO4 resulted in low‐medium DS products (0.3‐1.5 for starch
laurate, 0.07‐1.5 for starch stearate). Reaction temperature and catalyst intake also
affect the product DS although to a lesser extent than the type of catalyst. The
models may be used to determine the appropriate process conditions to obtain a
product with a pre‐defined DS.
Thermal and mechanical behaviour of the samples of different DS values clearly
shows that the DS represents one of the crucial factors affecting the final product
properties. Comparison with literature data indicates that the properties may be
(fine) tuned also by the starch source. Thus, the chemical composition of the
starting materials (either starch or the ester precursor) as well as the processing
parameters affect the final product DS values and as such provide an effective
toolbox to modulate the desired product properties for a given application.

5.5. Nomenclature
β : regression coefficients obtained by a multiple regression procedure
[various units, depend on the unit of related variable x]
β0 : constant regression coefficients obtained by a multiple regression
procedure [‐]
βi : regression coefficients of variable xi obtained by a multiple regression
procedure [various units]
β ii : regression coefficients of quadratic variable xi2 obtained by a multiple
regression procedure [various units]
β jk : regression coefficients of interaction variable xjxk obtained by a
multiple regression procedure [various units]
DS : Degree of Substitution, average value of mole of substituted –OH per
mole of anhydroglucose (AHG) units [‐]
102


DS fatty ester : DS of fatty ester (laurate or stearate) group substituents [‐]
pKb : base dissociation constant [‐]
Tc : crystallization temperature [oC]
Tm : melting temperature [oC]
xi : experimental variables used for mathematical modelling using
multiple regression procedure [various units]

Greek symbols:
σy : stress at yield [MPa]

5.6. References
[2]. R. J. Hernandez, S. E. M. Selke, J. D. Culter: Plastics Packaging, Properties,
Processing, Applications, and Regulations, Hanser Publishers, Munich,
Germany, 2000.
[3]. E. S. Stevens: Green Plastics, An Introduction to the New Science of Biodegradable
Plastics, Princeton University Press, New Jersey, USA, 2002.
[4]. A. Marsden: The challenge of domestic waste disposal, in Packaging in the
Environment (Ed. G. M. Levy), Blackie Academic & Professional, Glasgow,
UK, 1986.
[5]. G.W. Ehrenstein: Polymeric Materials, Structure – Properties ‐ Applications,
Hanser Publishers, Munich, Germany, 2000.
http://www.starch.dk/isi/stat/rawmaterial.html, accessed on August 01, 2008
[7]. F. H. Otey, R. P. Westhoff, W. M. Doane: Starch‐based blown films. Ind. Eng.
Chem. Prod. Res. Dev. 1980, 19, 592‐595.
[8]. C. Bastioli: Global Status of the Production of Biobased Packaging Materials,
in The Food Biopack Conference: Foodstuffs, Packaging, and Biopolymers,
Conference Proceedings, Copenhagen, Denmark, 27‐29 August 2000 (Ed. C.J.
Weber), Department of Dairy and Food Science, The Royal Veterinary and
Agricultural University, Frederiksberg C, Denmark, 2000.
103
Chapter 5
[9]. Biobased packaging materials for the food industry, status and perspectives,
report of Food Biopack Project (Ed. C. J. Weber), Department of Dairy and
Food Science, the Royal Veterinary and Agricultural University,
Frederiksberg C., Denmark, 2000.
[10]. X. L. Wang, K. K. Yang, Y. Z. Wang: Properties of starch blends with
biodegradable polymers. J. Macromol. Sci., Part C: Polym. Rev. 2003, C43, 385–
409.
[11]. J. W. Mullen, E. Pacsu: Starch studies: preparation and properties of starch
triesters. Ind. Eng. Chem. 1942, 34, 1209‐1217.
[12]. J. W. Mullen, E. Pacsu: Starch studies: possible industrial utilization of starch
esters. Ind. Eng. Chem. 1943, 35, 381‐384.
[13]. A. D. Sagar, E. W. Merrill: Properties of fatty‐acid esters of starch. J. Appl.
Polym. Sci. 1995, 58, 1647‐1656.
[15]. J. Aburto, S. Thiebaud, I. Alric, E. Borredon, D. Bikiaris, J. Prinos, C.
Panayiotou: Properties of octanoated starch and its blends with polyethylene.
Carbohydr. Polym. 1997, 34, 101‐112.
[16]. J. Aburto, H. Hamaili, G. Mouysset‐Baziard, F. Senocq, I. Alric, E. Borredon:
Free‐solvent synthesis and properties of higher fatty esters of starch ‐ Part 2.
Starch‐Starke 1999, 51, 302‐307.
[18]. J. Aburto, I. Alric, E. Borredon: Preparation of long‐chain esters of starch
using fatty acid chlorides in the absence of an organic solvent. Starch‐Starke
1999, 51, 132‐135.
various starches with lauric acid methyl ester and triacyl glycerides. Starch‐
Starke 2005, 57, 145‐152.
[20]. W. Mormann, M. Al‐Higari: Acylation of starch with vinyl acetate in water.
Starch‐Starke 2004, 56, 118‐121.
[21]. R. A. de Graaf, G. A. Broekroelofs, L. P. B. M. Janssen, A. A. C. M.
Beenackers: The kinetics of the acetylation of gelatinised potato starch.
Carbohydr. Polym. 1995, 28, 137‐144.
[22]. R. A. de Graaf: Ph.D. Thesis, Rijksuniversiteit Groningen, 1996
104


[23]. L. Junistia, A. K. Sugih, R. Manurung, F. Picchioni, L. P. B. M. Janssen, H. J.
stearate as reactants, 2008, accepted for publication in Starch‐Starke.
[24]. B. Laignel, C. Bliard, G. Massiot, J. M. Nuzillard: Proton NMR spectroscopy
assignment of D‐glucose residues in highly acetylated starch. Carbohydr. Res.
1997, 298, 251‐260.
[25]. D. C. Montgomery: Design and Analysis of Experiments 5th edition, John Wiley &
Sons Inc., New York, USA, 2001.
[26]. R. A. Day, A. L. Underwood: Quantitative Analysis 5th edition, Prentice‐Hall,
New Jersey, USA, 1986.
[27]. R. A. Dicke, A straight way to regioselectively functionalized polysaccharide
[28]. M. Ashby, H. Shercliff, D. Cebon, Materials: Engineering, Science, Processing
and Design, Butterworth‐Heinemann, Amsterdam, the Netherlands, 2007.
105
Chapter 6
Synthesis and Properties of Reactive Interfacial
Agents for Polycaprolactone‐Starch Blends

Abstract
The synthesis of two reactive interfacial agents for starch‐PCL blends, PCL‐g‐
glycidyl methacrylate (PCL‐GMA) and PCL‐g‐diethyl maleate (PCL‐g‐DEM) is
decribed. The compounds were prepared by reacting a low molecular weight PCL
(Mw 3000) with glycidyl methacrylate or diethyl maleate in the presence of
benzoylperoxide at 130°C. The effect of important process variables (intiator and
monomer intakes as well as estimated solubility of the monomer in molten PCL)
on the degree of grafting (FD) of the GMA and DEM units to the PCL backbone
was explored in detail and quantified using multivariable linear regression.
Highest values of the FD (up to 45 %) were observed for PCL‐g‐GMA, at relatively
high GMA and BPO intakes. The FD values for PCL‐g‐DEM were considerably
lower (up to 7 %). The reactive interfacial agents were tested for their performance
in PCL‐starch blends. Both act as compatibilizers for PCL/starch blends by
improving the interfacial adhesion between the starch particles and the PCL
matrix. As a result, the mechanical behavior of the compatibilized blends is in
general different from that of pure PCL and of the corresponding
uncompatibilized blends. In particular the elastic modulus for the compatibilized
blends is significantly higher than that of uncompatibilized ones. At relatively low
starch intakes, PCL‐g‐DEM has at least a comparable performance with respect to
PCL‐g‐GMA, despite the expected differences (favorable to PCL‐g‐GMA) in the in
situ formation of the compatibilizers. This discrepancy could be explained on the
basis of the functional groups (GMA or DEM) distribution along the PCL
backbone.

Keywords: starch‐PCL blend, compatibilizer, glycidyl methacrylate, diethyl maleate,
grafting

Chapter 6
6.1. Introduction
Starch is a cheap and abundantly available natural polymer with very good
application perspectives in the area of biodegradable plastics. Unfortunately, the
native material is very hydrophilic and important mechanical properties are
inferior compared to most synthetic polymers and this hampers its direct use as
packaging materials. Starch modifications to improve the product properties like
enhanced hydrophobicity and mechanical properties were already reported in the
early 19th century [1, 2]. One of the well known modification strategies is blending
the starch with polymers displaying a stronger hydrophobic character and better
mechanical properties, such as polyethylene (PE) or polystyrene (PS) [3‐6].
Unfortunately, these synthetic polymers are poorly or non‐biodegradable. To
overcome this issue, synthetic biodegradable polymers have been applied. Among
these, polyesters are considered very promising alternatives [7]. The ester bonds
are susceptible to attack by water and this leads to enhanced biodegradability. A
well‐known biodegradable polyester, polycaprolactone (PCL), is known to be
degraded with ease by microorganisms widely distributed in nature [8]. Aerobic
soil‐burial experiments showed that the mechanical properties of PCL films
decrease rapidly in time [9]. As a consequence, PCL has gained considerable
interest for possible applications in the fields of packaging materials and medical
applications [10‐11].
Blending of starch and synthetic biodegradable polyesters has been widely
applied for the synthesis of fully‐biodegradable products. However, blends of
hydrophilic starch and hydrophobic biodegradable polyesters exhibit phase
separation [12] due to differences in polarity of the building blocks. This
phenomenon is highly undesirable and limits the application range considerably
[13]. To reduce the tendency for phase separation, a compatibilizer (i.e. an
interfacial agent) may be used to improve the interfacial association between the
two polymer phases. In general a compatibilizer is a block‐copolymer where each
block displays a chemical structure equal or very similar to that of the polymers to
be mixed. This leads, for starch/PCL blends, to an ideal compatibilizer having both
PCL and starch blocks. Such structure is rather difficult to achieve by simple
copolymerization methods and it is usually prepared in situ (i.e. upon mixing) by
using a functionalized PCL. The latter displays the presence, along the backbone,
of polar groups (usually epoxides or anhydrides [13‐18]) able to react with the –
OH groups along the starch backbone. It must be stressed here that the word
“compatibilizer” is correctly used only when the block copolymer is actually able
to significantly influence the dispersion of the polymers to be mixed (most
probably through a steric stabilization mechanism [19]). When using
ungelatinized starch as a component in the blend, as in this study, it would be
actually more accurate to define the block copolymer as an “interfacial agent”,
108

Synthesis and Properties of Reactive Interfacial Agents for Polycaprolactone‐Starch Blends
which is able to mainly improve the interfacial adhesion between the polymer and
starch itself.
This chapter describes a systematic study on the synthesis of two reactive
interfacial agents for starch‐PCL systems, PCL‐g‐glycidyl methacrylate (PCL‐g‐
GMA) and PCL‐g‐diethyl maleate (PCL‐g‐DEM). The effect of important process
variables on the degree of grafting of the GMA and DEM units to the PCL
backbone has been explored in detail and quantified using multivariable linear
regression. The various compatibilizers have been tested for their performance in
PCL‐starch blends. Exploratory studies on the synthesis of PCL‐g‐GMA and its
applications for starch/PCL blends have been published [13, 16], however,
systematic studies and subsequent quantification of the functionalization reaction
has not been reported to date. The synthesis and application of PCL‐g‐DEM is, to
the best of our knowledge, an absolute novelty of the present research.

6.2.1. Materials
Polycaprolactone (PCL) CAPA 2304, Mw=3000 from Solvay Caprolactones, UK
was used for the preparation of the interfacial agents. This low molecular weight
PCL grade was used without further purification. Glycidyl methacrylate 97%
purity (Aldrich), diethyl maleate ≥97% purity (Fluka), and benzoyl peroxide 75%
(Merck, Germany) were used as received. Tetrahydrofuran (THF, >99%) was
obtained from Acros, Belgium, xylene (99.8%) from Merck, Germany and
methanol (99.8%) from Labscan, Ireland. Corn starch (with approx. 73%
amylopectin and 27% amylase) was obtained from Sigma and high molecular
weight PCL (CAPA 6503, Mw=50000) from Solvay Caprolactone, UK. The starch
was dried for at least 24 h at 110oC under vacuum (approx. ~1 mbar) prior to use.

6.2.2. Methods
6.2.2.1. Compatibilizer synthesis
The compatibilizers were prepared in a Brabender Plasticorder PL2000 batch‐
kneader (chamber volume 35 cm3). The intake of reagents was maximally 75‐80%
of the volumetric volume to ensure proper mixing. The kneader was heated to
130 oC and PCL (CAPA 2304) was added while maintaining a rotational speed of
80 rpm. After the PCL was melted (1‐2 minutes), a solution of BPO in GMA or
DEM was added drop by drop over a period of 5 minutes. The materials were
mixed for another 5 minutes, after which the equipment was stopped and the
chamber was opened to collect the samples. Intakes for each experiment are given
in Table 1 and 2.
109
Chapter 6
6.2.2.2. Work‐up of PCL‐g‐GMA products [13]
To remove unreacted GMA monomer and GMA homopolymer, PCL‐g‐GMA (5
g) was dissolved in 50 ml THF, stirred for 1.5 h, and then filtered. Methanol (450
mL) was added to the filtrate, and the product was precipitated at 6‐8 oC. The
solvent was decanted and the solid product was dried in a vacuum oven (40 oC, 5
mbar).

6.2.2.3. Work‐up of PCL‐g‐DEM products
Purification of the PCL‐g‐DEM product was performed according to a modified
procedure for PCL‐g‐maleic anhydride [17]. PCL‐g‐DEM (5 g) was dissolved in
xylene and refluxed at 150 oC for 2 h. The resulting suspension was filtered and
precipitated using methanol (450 mL) at 6‐8°C. The solvent was decanted and the
solid product was dried in a vacuum oven (40 oC, 5 mbar).

6.2.2.4. Preparation of PCL‐starch blends with the reactive compatibilisers
The PCL‐starch blends were prepared in a Brabender Plasticorder PL2000
batchkneader (chamber volume 35 cm3). An operation temperature of 170°C and a
rotation speed of 80 rpm were applied [18]. PCL was added to the chamber
followed by the addition of the starch and the reactive compatibilizer. The content
was blended for 15 minutes. Subsequently, the chamber was opened and the
resulting material was collected.

6.2.3. Analytical Methods
1
H‐NMR measurements were performed using a 400 MHz Varian AMX Oxford
NMR apparatus with CDCl3 (99.8%, Aldrich) as the solvent. Digital Scanning
Calorimetry (DSC) measurements were performed using a Q1000 TA Instruments
equipped with a TA Instruments DSC cooling system. Each sample was first
heated from 0 oC to 100 oC (heating rate 10 oC/min) to remove the thermal history
of the material. The transition temperatures of each sample were further
determined by first cooling down the samples from 100 oC to 0 oC and
subsequently heating up back to 100 oC (cooling and heating rate were 10 oC/min).
The error on the transition temperature is assumed to be ± 1 oC and 5 % of the
calculated values for the corresponding enthalpies. Scanning Electron Microscopy
(SEM) was performed using a Jeol 6320 F Scanning Electron Microscope. Before
analysis, the samples were covered with a paladium/platinum conductive layer of
3 µm thickness, created using a Cressington 208 sputter coater. Infrared spectra
were collected with a FT‐IR apparatus in the ATR mode using a Spectrum 2000
instrument from Perkin Elmer. Tensile tests were performed using an Instron 4301
110

machine. The T‐bone samples were prepared using a Fontijne Holland TH 400 hot‐
press. For a given sample/blend, 8 different T‐bones were used. For every T‐bone,
strain at break (ε), stress at break (σ) and modulus (E) were measured. The
corresponding value for every blend was calculated as an average of the 8
measurements while the standard deviation was taken as absolute error on the
average values.

6.2.3.1. Calculation of the Degree of Functionalization (FD) of the reactive compatibilizers
The number of moles of GMA or DEM present on the PCL backbone was
quantified using the degree of functionalisation (FD). The FD is defined as:
number of moles of GMA/DEM attached to PCL (mol)

FD = × 100% (6.1.)
number of repeating units of the PCL backbone (mol)
The FD was calculated using 1H‐NMR by comparing the area of protons
belonging to the GMA (‐CH< proton at δ 3.2 ppm) or DEM (‐CH2‐ protons at δ 4.2
ppm) side chains with that of a characteristic proton resonance of the PCL
backbone (‐CH2‐ protons at δ 4.0 ppm [13, 15‐16, 19]. A 5% relative error in the
peak area of the NMR spectra was assumed, leading to a 10% relative error in the
FD values.

6.2.4. Statistical Modeling
The influence of different processing parameters on the FD values has been
determined by performing a multivariable regression procedure on the available
data. As a result we were able to obtain a model for the FD of the reaction. The
validity of the model was determined by performing an analysis of variance
(ANOVA, Table 3). This procedure is described in detail in the literature [23] and
consists of calculating the sum of squares (SS) for the model and the error. When
the relative degrees of freedom (DF) are known, it is possible to calculate the mean
square (MS) for the model and the error. On the basis of the latter values, the F‐
value for the model is determined followed by the P‐value. The latter is a measure
of the statistical significance of the model.
111
Chapter 6
6.3.1. Preparation of the Reactive Compatibilizers
Twelve compatibilizers were prepared by either reacting glycidyl methacrylate
(GMA) or diethyl maleate (DEM) with low molecular PCL as the starting polymer
and benzoyl peroxide (BPO) as the radical initiator (Scheme 6.1.).

COOEt
COOEt
O
* n *
DEM
O
PCL-g-DEM
BPO
O
* n *
O
O
GMA O O
BPO
O
* n *
O
PCL-g-GMA

Scheme 6.1. Functionalization reactions (only showing reactivity for the >CH2 in
α position on PCL backbone)

Typical 1H‐NMR spectra for the products are shown in Figure 6.1.

112


Figure 6.1. Typical 1H‐NMR spectra of PCL‐g‐GMA (top) and PCL‐g‐DEM
(bottom)

Peak assignments were based on available data reported for related products
[13, 15‐16, 20]. The FD values and the thermal properties of the products are
shown in Table 6.1. (PCL‐g‐GMA) and Table 6.2. (PCL‐g‐DEM).

113
Chapter 6
Table 6.1. Overview of experiments for the PCL‐g‐GMA compatibilisers a
Intake
FD T cryst. ΔH cryst. T melt. ΔH melt.
Sample (%‐mol) b
(%) ( C)
o (J/ g‐PCL) ( C)
o (J/ g‐PCL)
GMA BPO
PCL ‐ ‐ ‐ 27 75 51 79

PCL‐g‐GMA 1 12 0.6 5.7 22 72 47 72
PCL‐g‐GMA 2 24 0.6 15.2 19 68 47 69
PCL‐g‐GMA 3 36 0.6 27.9 29 67 48 68
PCL‐g‐GMA 4 6 0.3 3.2 25 73 47 73
PCL‐g‐GMA 5 24 1.1 45.2 c 20 66 46 68
PCL‐g‐GMA 6 12 0.3 6.3 28 71 47 67
a Experiments were carried out at 130 °C.
b %‐mol with respect to the CL repeating units in the PCL
c based on the soluble fraction of the compatibilizer

Table 6.2. Overview of experiments for the PCL‐g‐DEM compatibilisers a
Intake
FD T cryst. ΔH cryst. T melt. ΔH melt.
Sample (%‐mol) b
(%) ( C)
o (J/ g‐PCL) ( C)
o (J/ g‐PCL)
DEM BPO
PCL ‐ ‐ ‐ 27 75 51 79

PCL‐g‐DEM 1 15 0.6 3.5 27 70 44 72
PCL‐g‐DEM 2 30 0.6 2.1 30 68 47 62
PCL‐g‐DEM 3 45 0.6 7.2 21 64 45 66
PCL‐g‐DEM 4 30 1.1 3.6 26 61 42 60
PCL‐g‐DEM 5 60 1.1 6.5 25 63 41 66
PCL‐g‐DEM 6 7.5 0.3 0.9 30 70 47 69
a Experiments were carried out at 130 °C
b %‐mol with respect to the CL repeating units in the PCL

One of the PCL‐g‐GMA compatibilizers (PCL‐g‐GMA 5, see Table 6.1.), was
only partly soluble in CDCl3, and therefore the FD is based on the soluble fraction
of the compatibilizer only. The presence of an insoluble fraction, combined with a
relatively broad molecular weight distribution (as shown by GPC, but not shown
here for brevity), suggests that cross‐linking occurred during this experiment.
In general, the FD of the PCL‐g‐GMA (3.2‐45.2%) products is much higher than
those of PCL‐g‐DEM (0.9‐7.2%). This difference may be either due to the difference
in mutual solubility of the GMA and DEM in PCL or differences in the molecular
114

mechanism of the grafting reaction. The mutual solubility may be expressed in
terms of the differences of solubility parameters of PCL and the reagents [18]. The
difference between the solubility parameters of GMA and PCL is 0.29 cal1/2cm‐3/2,
while it is much higher (6.3 cal1/2cm‐3/2) for DEM and PCL [20]. Thus, GMA is likely
better soluble in PCL, leading to higher values for FD of the products, as
confirmed by our experiments. However, the higher experimental grafting
efficiencies for PCL‐GMA may also be rationalised by considering the reaction
mechanism for the preparation of the compatibilisers. GMA molecules may either
react directly with a radical at the PCL backbone or with a radical present on an
already coupled GMA molecule. The latter leads to longer GMA grafts on a PCL
backbone [13]. A simplified representation of the reactivity of GMA is shown in
Figure 6.2.

O O
hydrogen -
(CH2)4 CH2 C O + RO (CH2)4 CH C O
abstraction
n n
+ GMA
O
(CH2)4 CH C O
n
H2C
-
H3C C C O CH2 HC CH2
O O
Termination
+ k GMA
+ H or
+ R or O
+ RO
(CH2)4 CH C O
n
H2C
H3C C C O CH2 HC CH2
k+1 O O
Figure 6.2. Simplified scheme of the GMA Grafting Reaction Mechanism [12]

The reactivity of DEM in radical reactions is expected to be different from that
of GMA. Previous studies on maleic anhydride (MA), a compound resembling the
115
Chapter 6
chemical structure of DEM, showed that MA reacts easily with a radical on the
PCL backbone. However, subsequent reactions of MA to an already grafted MA
molecule hardly occur. Hence, the length of a MA graft is always unity whereas
longer grafts are possible for GMA. Our experimental findings, i.e. higher FD
values for GMA than for DEM are in line with this explanation and support the
proposed molecular reaction mechanisms.

6.3.1.1. Effect of substrate (GMA/DEM) to PCL ratio on the FD
The effect of the substrate (GMA or DEM) to PCL ratio on the FD was studied
by changing the GMA/DEM intake at different BPO amounts. The results are
graphically provided in Figure 6.3.
30
25
20
GMA, BPO=0.6% mol/mol-CL units

FD (%)
15 GMA, BPO=0.3% mol/mol-CL units

DEM, BPO=0.6% mol/mol-CL units
DEM, BPO=1.1% mol/mol-CL units
10
0
10 20 30 40 50 60
Substrate Intake (mol% on CL units)

Figure 6.3. Effect of GMA and DEM to CL‐units ratio (mol/mol) on the FD of the
products (constant PCL intake, 130°C)

Higher GMA intakes lead to higher FD values. This trend is independent of the
BPO amount and matches with data reported by other groups [13, 22]. It is most
probably related to the fact that GMA is able to propagate to longer grafted chains
when reacted with PCL (Figure 6.2.). Thus an increase in the GMA intake will
provide more monomer available for the growing of the grafted chains leading to
higher FD values.
116

The experimental trend for DEM is different. The FD values are within a rather
narrow range, although a slight increase in the FD values might be appreciated.
Such behavior is slightly in contrast to what observed for maleic anhydride [13, 15,
17], for which a levelling off and eventually a decrease of the FD values for
relatively higher MA amounts has been observed. Such discrepancy is probably
related to the relatively low intake of DEM as well as to the different reactivity of
DEM compared to MA [22].

6.3.1.2. Effect of the BPO intake on the product FD
The effect of the BPO intake on the FD was studied by using different intakes of
BPO (Figure 6.4.).
50
45
40
35
30
FD (%)
25
20
GMA, GMA=24% mol/mol-CL units

15
GMA, GMA=12% mol/mol-CL units
DEM, DEM=30% mol/mol-CL units
10
0
0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2
BPO Intake (mol% on CL units)

Figure 6.4. Effect of the BPO intake on the product FD (130°C, constant substrate
to PCL ratio)

For high GMA to CL ratio (24%‐mol/ mol CL units), doubling the amount of
initiator results in considerable higher product FD. These results are in line with
earlier work [15, 17]. The use of higher initiator concentrations will result in an
increase in the number of formed radicals. This will lead to a higher proportion of
PCL radicals by hydrogen abstraction from the polymer backbone and thus to
higher FD values. However, at relatively lower GMA intakes, no detectable
influence of the BPO amount on the FD is observed (Figure 6.4.). Apparently, there
117
Chapter 6
is an optimum ratio between the BPO concentration and available monomer
(GMA in this case) on the FD. If many macroradicals are created in the system
(relatively high BPO intakes) at low GMA intake, the possibility of side reactions
will become significant. In particular the occurrence of “cage effects”, i.e.
recombination of (macro) radicals, as well as several transfer reactions might be
responsible for the observed trend [22].
The data for PCL‐g‐DEM shows a similar trend as for the experiments with a
high GMA to CL ratio (24%‐mol/ mol CL units), namely an increase in the BPO
intake results in higher product FD. The effect is however much less pronounced
than for GMA. A doubling of the initiator intake for PCL‐g‐DEM only results in a
70% increase in the FD (compared to 300% for PCL‐g‐GMA). This phenomenon is
likely related to the different nature of the radical grafting mechanism of GMA
and DEM on PCL as mentioned before.

6.3.1.3. Modeling of the combined effects of the GMA/ DEM and BPO intakes on the FD
Previous studies showed the importance of the initiator to monomer ratio on
the FD values [15, 17]. However, these investigations focused on a better
understanding of the individual variables by studying the effect of higher
peroxide and monomer intakes while keeping for example their ratio constant.
The results described in the previous paragraph imply that the mutual interaction
between these variables and not the absolute value itself determines the final FD
values to a great extent. To quantify synergic effects of monomer and initiator
molar intakes on the FD of GMA and DEM on the PCL backbone, a statistical
model has been developed by performing a multivariable linear regression on the
data reported in Table 6.1. and 6.2. Here, the BPO and monomer intakes are
considered as independent parameters. In addition, the mutual solubility of DEM
in PCL and DMA in PCL was included in the model by using a parameter δ,
defined as the difference in solubility parameters between PCl and the substrates.
This leads to the following equation:
FD = f (nm , ni , δ ) (6.2.)
where nm is the molar amount of monomer in the feed, ni the molar amount of
initiator in the feed and δ the difference in solubility parameters calculated using
group contributions [21].
The model provided in eq. 6.3. gives the best description of the experimental
data:
FD = −1.8875 + 0.0325nmδ + 0.5431niδ + 1.6022nm ni − 0.2729nm niδ (6.3.)
118

The analysis of variance data are given in Table 6.3. The very low P‐value
implies that the model is statistically significant. This was also confirmed by
inspection of the residue distribution by a normal probability plot (not reported
here for brevity) [23]. The R2 value for the model (0.957) and its closeness to the
adjusted R2 (0.941) also suggests that all important variables have been included in
the model.

Table 6.3. Analysis of variance for the FD model provided in eq 6.3.
SS DF MS F P‐value

Model 1828 4 456.908 44.928 <10‐9
Error 81.358 8 10.17
Total 1909.358 12

The value of the coefficients in the model imply that the FD is positively
influenced by the interaction between monomer and initiator intake ( nm ni ), the
interaction between (PCL‐monomer) intake and solubility parameter difference
( nmδ ) and the interaction between the latter factor and the initiator intake ( ni δ ).
Graphical representation of the FD models for PCL‐g‐GMA and PCL‐g‐DEM are
given in Figure 6.5. and Figure 6.6., respectively.

1
60 70
40
50
0.9
80
0.8
20
50
40
30
60
0.7
10
FD (%)
40
0.6 FD (%)
20 30 40
20
0.5
10 30
0
0.4 20
1
0.8 50
40 0.3 20
0.6 10
30
0.4 20 0.2
BPO Intake 0.2 10 10 15 20 25 30 35 40 45 50
GMA Intake
(mol% on CL units) GMA Intakes (mol% on CL units)
(mol% on CL units)
a. b.

Figure 6.5. Graphical representation of the FD model for PCL‐g‐GMA.
a. 3D plot. b. Contour plot.
119
Chapter 6
4
3
6
0.9
8
0.8

6
0.7
FD (%)
4
4
5
FD (%)
6
0.6
7
2
0.5
0
0.4
1
0.8
3
50
6
5
1
40 0.3
0.6
30
0.4
7
20
2
0.2
BPO Intake 0.2 10 10 15 20 25 30 35 40 45 50
DEM Intake
(mol% on CL units) DEM Intake (mol% on CL units)
(mol% on CL units)
a. b.

Figure 6.6. Graphical representation of the FD model for PCL‐g‐DEM.
a. 3D plot. b. Contour plot.

Remarkable is once more the different trends for the two substrates. While for
GMA an increase in BPO or monomer intakes invariably leds to higher FD values,
for DEM a clear transition is observed. For relatively high DEM intakes (> 30 %
mol/mol), a higher BPO intake leads to a reduction of the FD values while an
opposite trend is observed at lower DEM intakes. These differences in dependency
of the FD values on the DEM and GMA intakes may be explained on the basis of
the different grafting mechanisms as previously discussed. From a more practical
point of view, the statistical model provides besides a reliable description of the
experimental data also a good mathematical description to modulate the FD
values of the two substrates by changing the chemical composition (monomer,
radical initiator and PCL intakes).

6.3.1.4. Thermal properties of the compatibilizers
The thermal properties (Table 6.1. and 6.2.) of the compatibilizers were
determined by DSC. For all samples, the melting temperature and the relative
enthalpy decrease with respect to pure PCL. Furthermore, the crystallization
temperature and enthalpy are not a clear function of the FD values, although both
properties are significantly lower than those of pure PCL. Such changes in the
thermal properties compared to pure PCL may be caused by the introduction of
grafts on the PCL chains. This induces irregularities and is expected to result in a
lowering of the Tc and the Tm. Similar observations have also been made by Kim, et
al [13] working with PCL‐g‐GMA. The random behavior of the Tc as function of
the FD values is probably the result of two concurring effects: the presence of
irregularities, which is expected to lead to a Tc reduction, and a favored nucleation
120

of the PCL chains in the compatibilizers compared to virgin PCL (higher Tc)
induced by the presence of polar groups.
The lower crystallization and melting enthalpy of the compatibilizers compared
to virgin PCL is possibly caused by the presence of side chains on the PCL
backbone which are expected to have a negative effect on the crystallinity of the
products.

6.3.2. Synthesis and Properties of Starch‐ PCL Blends
The two compatibilizers (PCL‐g‐GMA and PCL‐g‐DEM) were further used as
interfacial agents in blends of PCL with starch. A total of 12 blends were prepared:
a series of binary ones (not containing any compatibilizers) constituting our
reference points, a series with PCL‐g‐DEM (different intakes of the latter at fixed
starch/PCL ratio) and two series with PCL‐g‐GMA (one with different intakes of
PCL‐g‐GMA at a fixed starch/PCL ratio and one with a fixed compatibilizer intake
at three different starch/PCL ratios). An overview of all prepared blends together
with their thermal and mechanical (tensile tests) properties is given in Table 6.4.

Table 6.4. Thermal and mechanical properties of starch/PCL blends a
σ ε E Tc ∆Hc T m ∆Hm

Sample
(MPa) (%) (MPa) (oC) (J/gPCL) (oC) (J/gPCL)
PCL 16.3 640.5 270.2 35 55 57 62
S/PCL 10/90 15.3 489.5 321.2 36 69 57 68
S/PCL 20/80 10.5 425.4 337.0 36 51 57 53
S/PCL 30/70 7.1 230.0 341.8 36 40 57 43
S/PCL/PCL‐g‐DEM 20/80/1 11.1 401.8 371.2 30 52 57 46
S/PCL/PCL‐g‐DEM 20/80/2 10.9 384.9 342.8 30 54 57 50
S/PCL/PCL‐g‐DEM 20/80/5 11.0 379.5 329.9 30 55 57 52
S/PCL/PCL‐g‐GMA 20/80/1 9.2 357.2 368.3 29 53 57 51
S/PCL/PCL‐g‐GMA 20/80/2 9.6 343.4 380.8 31 58 58 54
S/PCL/PCL‐g‐GMA 20/80/5 11.7 431.6 372.6 30 57 57 50
S/PCL/PCL‐g‐GMA 20/80/10 10.1 305.9 386.0 29 63 57 60
S/PCL/PCL‐g‐GMA 10/90/2 13.6 424.6 332.0 28 54 58 51
S/PCL/PCL‐g‐GMA 30/70/2 5.5 168.9 430.3 29 55 56 48
a
PCL‐g‐DEM has an FD of 1.7 %. PCL‐g‐GMA has an FD of 9.6%.

Our experimental design allows comparisons of thermal and mechanical
properties as function of the starch content for binary blends (no compatibilizer)
and intake of PCL‐g‐DEM and of PCL‐g‐GMA.
121
Chapter 6
6.3.2.1. Binary blends of starch and PCL
For binary blends (no compatibilizer is added) a monotonous decrease of the
stress and strain at break is observed as function of the starch intake, respectively
from 16.43 MPa and 640.5 % for pure PCL to 7.1 MPa and 341.8 % at 30 % starch
content. The simultaneous increase in the modulus (from 270.2 MPa up to 341.8
MPa) clearly indicates that the rigidity of the blends increases at higher starch
contents. This is in agreement with the data previously reported on PCL/starch
blends and related to the lack of interfacial adhesion between the starch particles
and the PCL matrix [18]. The thermal behavior is characterized by no significant
changes in the Tm and Tc values but unreported trends of the corresponding
enthalpies as function of the starch content (Figure 6.7.).

75
70
65
60
∆ H (J/ g-PCL)
55
50
45
40
35
0 5 10 15 20 25 30
Starch content (%)

Figure 6.7. Melting and crystallization enthalpy as function of the starch content
for binary blends with PCL.
○ : ∆Hc □ : ∆Hm

The two trends are remarkably mirroring each other and are both characterized
by an increase of ∆Hm and ∆Hc with respect to pure PCL at 10 % starch content
followed by a monotonous decrease of both quantities as function of the starch
intake. This is most probably due to a nucleation effect of the starch on PCL as
observed also for other kinds of polymer blends [27]. The thermal and mechanical
properties indicate no or very little interaction of the starch particles with the PCL
122

matrix. This is visually confirmed by the morphology of the prepared blends
(Figure 6.8.).
a. b. c.

Figure 6.8. Morphology of starch/PCL binary blends.
(a) S/PCL 10/90 (b) S/PCL 20/80 (c) S/PCL 30/70

Indeed starch particles are clearly dispersed in the PCL matrix but no
interaction (adhesion) between the two phases is actually detected. Thus, the
starch particles are simply inserted into voids of the PCL matrix. The presence of
these voids was also observed for sago starch/PCL blends [25, 26] and explained
by assuming that the voids are formed by water in the starch. Evaporation of the
water during blend preparation, either by heating or in combination with
mechanical stress [24], and the lack of interface adhesion cause void formation.

6.3.2.2. Ternary blends compatibilized with PCL‐g‐DEM
PCL‐g‐DEM, whose synthesis is described for the first time in this work, was
used as compatibilizer for the preparation of ternary blends with starch and PCL.
The thermal behavior is characterized (Table 6.4.) by constant values of the Tm
(with respect to pure PCL) while the Tc is slightly lower (30 vs 35 oC) than that of
pure PCL and it is independent of the compatibilizer content. The latter trend is
also valid for the corresponding enthalpies: i.e. a decrease with respect to pure
PCL is observed and then a substantial invariance as function of the PCL‐g‐DEM
content. Such behavior is consistent with the hypothesis that the starch particles
hinder the melting and crystallization processes of the PCL chains. However, in
order to fully understand the role of PCL‐g‐DEM, the corresponding binary blend
(S/PCL 20/80) contitutes a better reference point compared to pure PCL. In this
respect the thermal properties remain substantially unchanged with the exception
of the Tc, for which a 6 oC drop is observed when using PCL‐g‐DEM. As
consequence one might expect a slight different structure of these ternary blends
as compared to the corresponding binary one. This is confirmed by examination of
the blend morphology by SEM (Figure 6.9.).
123
Chapter 6
a. b. c.

Figure 6.9. Morphology of ternary blends compatibilized with PCL‐g‐DEM.
(a) S/PCL/PCL‐g‐DEM 20/80/1
(b) S/PCL/PCL‐g‐DEM 20/80/2
(c) S/PCL/PCL‐g‐DEM 20/80/5

Indeed starch particles in the ternary blends display a smoother interface with
the PCL matrix compared to the corresponding binary blends (Figure 6.9.). A
closer inspection of the SEM pictures reveals that the starch particles are clearly
embedded in the PCL matrix with almost no void spaces at the interface. As a
result of the different morphology also the mechanical behavior display
differences with the virgin PCL and the binary blend. The stress and strain at
break remain constant as function of the compatibilizer intake while the modulus
display an optimum as function of the compatibilizer intake (Figure 6.10.).
At 1 %‐wt of PCL‐g‐DEM in the blends the modulus increases with respect to
the binary blend (0 %‐wt compatibilizer in Figure 6.10.). This can be explained by
the improved interfacial adhesion [14] between PCL and starch, which will hinder
the flowability and fibre forming ability of PCL matrix under cold drawing,
resulting in more rigid material with higher modulus. At higher PCL‐g‐DEM
intakes the lower average molecular weight of the compatibilizer as compared to
the one of the unmodified PCL used in the blends (3000 vs 50000) is probably
responsible for the observed decrease in the E values.

124

390
380
370
360
E (MPa)
350
340
330
320
0 1 2 3 4 5
PCL-g-DEM amount (wt%)

Figure 6.10. Modulus of ternary blends S/PCL/PCL‐g‐DEM as function of PCL‐g‐
DEM amount.

6.3.2.3. Ternary blends compatibilized with PCL‐g‐GMA
When using PCL‐g‐GMA as compatibilizer two possible comparisons can
actually be made: one at fixed starch/PCL ratio and changing the amount of PCL‐
g‐GMA and one at different starch/PCL ratios but with fixed amount of PCL‐g‐
GMA (2 %‐wt).
The first comparison as function of the PCL‐g‐GMA intake (ternary blends with
starch/PCL ratio of 20/80) results in quite similar considerations as the ones made
for PCL‐g‐DEM. In particular also in the case of PCL‐g‐GMA, by taking as
reference the corresponding binary blend (S/PCL 20/80), it can be observed that:
1) the melting temperature as well as the crystallization and melting
enthalpies do not change significantly (discrepancies in the values are
within the boundary defined by the experimental error);
2) the crystallization temperature experiences a drop of about 5‐6 oC;
3) the stress and strain at break are lowered and hardly a function of the
PCL‐g‐GMA amount;
4) the modulus is higher and hardly a function of the compatibilizer amount.
Remarkable is that in the case of PCL‐g‐GMA no optimum is found in the
modulus as function of the compatibilizer intake. However, in all cases (i.e. at all
PCL‐g‐GMA contents) there is a clear increase of the modulus with respect to the
125
Chapter 6
binary blend, clearly indicating that also PCL‐g‐GMA (like PCL‐g‐DEM) acts as
compatibilizer in the blends Such hypothesis is partially and qualitatively
confirmed by the blends morphology (Figure 6.11.).
a. b. c.

Figure 6.11. Morphology of ternary blends S/PCL/PCL‐g‐GMA.
(a) S/PCL/PCL‐g‐GMA 20/80/1
(b) S/PCL/PCL‐g‐GMA 20/80/2
(c) S/PCL/PCL‐g‐GMA 20/80/3

In this case we observe structural features (partially smooth interface between
the components, presence of voids, partial interfacial adhesion) which render
these ternary blends a kind of “intermediate” case between the binary ones and
those compatibilized with PCL‐g‐DEM. This is surprising if one takes into account
the different FD values: 9.8 % for PCL‐g‐GMA against 1.7 % for PCL‐g‐DEM. Thus,
despite a much more favourable FD value for PCL‐g‐GMA with respect to PCL‐g‐
DEM and even despite a relative low reactivity of ester groups towards the –OH
groups on starch and nucleophilic groups in general [27, 28, 29], PCL‐g‐DEM is at
least as active as compatibilizer (compare modulus values in Table 6.4. and
morphologies in Figures 6.9. and 6.11.) than PCL‐g‐GMA for blends containing 20
%‐wt of starch. These differences in compatibilizing effect are not yet fully
understood. However, one might speculate that the longer length of the GMA
grafts compared to DEM (vide supra) plays a negative role in the in situ formation
of the compatibilizer. As given schematically in Figure 6.12. (left) the structure of
PCL‐g‐GMA is inhomogeneous at the molecular level with long poly(GMA)
branches pending from the PCL backbone. This confines all reactive GMA groups
in relatively “isolated” spots along the PCL backbone. As a result, upon reaction of
PCL‐g‐GMA with one of the –OH groups on the surface of the starch particles the
remaining GMA groups are actually scarcely available for further reaction with
other –OH groups spatially distant from the reacted one.
126

PCL-g-DEM
PCL-g-GMA
OH OH
O O
OH Starch particle Starch particle

OH

Figure 6.12. Schematic representation of the in situ reaction between
functionalized PCL and starch.

On the other hand (right of Figure 6.12.) PCL‐g‐DEM has much less reactive
groups (lower FD) than PCL‐g‐GMA; but since DEM is preferentially grafted as
monomer, the distribution of the reactive group along the PCL backbone is more
“homogenous”. As consequence, once PCL‐g‐DEM has reacted with one –OH
groups on the starch particle, other groups will presumably remain available
(arrows in Figure 6.12.) for further reaction, thus probably ensuring a better
coverage of the surface.
The proposed explanation implies however that at relatively higher starch
contents (> 20 %‐wt) the segregation of the poly(GMA) chains in “isolated” spot
along the PCL backbone would become less important. Indeed, at higher starch
contents more –OH groups would be available for reaction with the GMA groups,
thus attenuating the effect discussed above. In order to check this we compared
blends with the same amount of PCL‐g‐GMA (2 %‐wt) but with different starch
intakes (10, 20 and 30 %‐wt respectively). Concerning the thermal behavior (Table
6.4.), the Tc decreases (with respect to the corresponding binary blends) while all
other factors (Tm and enthalpies) remain virtually unchanged with respect to the
binary blends and also as function of the starch intake. Moreover, for all starch
amounts (Table 6.4.), when comparing ternary blends (compatibilized with PCL‐g‐
GMA) with the binary ones, the stress and strain at break decrease while the
modulus is unchanged at low starch contents and increases significantly for S/PCL
blends with 30 %‐wt starch. The latter result (from a modulus of 341.8 MPa for
S/PCL 30/70 up to 430.3 MPa for S/PCL/PCL‐g‐GMA 30/70/2) clearly indicates that
the efficiency of PCL‐g‐GMA as compatibilizer becomes more relevant at
relatively higher starch contents (>30 %) [25]. This is in agreement with the
hypothesis made above (Figure 6.12.) and relating the “coverage” of the starch
particle surface upon reaction with PCL‐g‐GMA with the compatibilization
efficiency. We can therefore conclude that the efficiency of PCL‐g‐GMA in the
127
Chapter 6
compatibilization of starch/PCL blends can be significantly improved by changing
the blend composition.

6.4. Conclusions
A systematic study, including statistical modeling, has been performed on the
synthesis of two reactive compatibilisers, PCL‐g‐glycidyl methacrylate (PCL‐g‐
GMA) and PCL‐g‐diethyl maleate (PCL‐g‐DEM). The proposed model to quantify
the effects of process variables (monomer and initiator intake, mutual solubility of
the monomer into the molten polymer) on the FD adequately described the
experimental results (R2=0.957, P‐value ≤10‐9). The most important variable for the
final product FD is the interaction between the amount of monomer and initiator
used. This represents an unreported mathematical confirmation of the fact that
these kinds of functionalization reactions are mainly governed by the synergy
between the different process variables and only slightly by their individual
values. The PCL‐g‐GMA and PCL‐g‐DEM compatibilizers display lower melting
temperatures and melting enthalpies compared to virgin PCL.
The reactive compatibilizers were used in blends of starch with PCL. At a fixed
starch content (20 %‐wt) PCL‐g‐DEM seems to have sligthly more efficient
compatibilizing effect than PCL‐g‐GMA as shown by blends morphology and
elastic modulus values. This is in contrast with chemical reactivity and amount of
chemical groups along the PCL backbone (both factors favorable to GMA as
compared to DEM) but it is explainable on the basis of the group distribution
along the PCL backbone. The latter hypothesis is indirectly confirmed by the
observation that PCL‐g‐GMA becomes more efficient at relatively higher starch
contents in the blends. From a more practical point of view it can be concluded
that the newly synthesized PCL‐g‐DEM, firstly reported in this work, can replace
PCL‐g‐GMA as compatibilizer at relatively low starch contents offering at the
same time the advantage of a less pronounced modification of the polymer
backbone and a reduced consumtpion of polar groups to be grafted on PCL.

6.5. Nomenclature
E : initial modulus [MPa]
FD : functionalization degree, moles of GMA or DEM present per mole of
CL repeating units
ni : amount of initiator intake [mol% on CL units]
nm : amount of monomer GMA or DEM intake [mol% on CL units]
128

Tc : crystallization temperature [oC]
Tm : melting temperature [oC]

Greek symbols:
∆Hc : enthalpy of crytallization [J/g‐PCL]
∆Hm : enthalpy of melting [J/g‐PCL]
δ : solubility parameter [cal1/2cm‐3/2]
σ : stress at break [MPa]
ε : strain at break [%]

6.6. References
[1]. W. Jarowenko: Acetylated starch and miscellanous organic esters, in Modified
Starches: Properties and Uses (Ed. O.B. Wurzburg), CRC Press, Inc., Boca Raton,
USA, 1986.
[2]. J.W. Mullen, E. Pacsu: Starch Studies: Preparation and properties of starch
triesters. Ind.Eng.Chem. 1942, 34, 1209‐1217.
[3]. F.J. Rodriguez‐Gonzales, B.A. Ramsay, B.D. Davis: High performance LDPE/
thermoplastic starch blend: a sustainable alternative to pure polyethylene,
Polymer 2003, 44, 1517‐1526.
polymer blends – III. Effect of amylopectin to amylase ratio in starch, Eur.
Polym. J. 1998, 34, 1467‐1475.
polymer blends – IV. Thermal and morphological properties, Eur. Polym. J.
1998, 34, 1477‐1487.
[6]. E. Psomiadou, A. Ioannis, C.G. Biliaderis, H. Ogawa, N. Kawasaki:
Biodegradable films made from low density polyethylene (LDPE), wheat
starch and soluble starch for food packaging applications. Part 2. Carbohydr.
Polym. 1997, 33, 227‐242.
[7]. M.S. Lindblad, Y. Liu, A‐C. Albertsson, E. Ranucci, S. Karlsson: Polymers
from Renewable Resources. Adv. Polym. Sci. 2002, 157, 139‐161.
[8]. S. Karlsson, A‐C. Albertsson: Biodegradable Polymers and Environmental
Interaction. Polym. Eng. Sci. 1998, 38, 1251‐1253.
129
Chapter 6
[9]. D. Goldberg: A review of the biodegradability and utility of
poly(caprolactone). J. Environ. Polym. Degrad. 1995, 3, 61‐67.
[10]. R. Chandra, R. Rustgi: Biodegradable polymers. Prog. Polym. Sci. 1998, 23,
1273‐1335.
[11]. E. Chiellini, R. Solaro: Biodegradable polymeric materials. Adv. Mater. 1996, 8,
305‐313.
[12]. L. Averous, L. Moro, P. Dole, C. Fringant: Properties of thermoplastic blends:
starch‐polycaprolactone. Polymer 2000, 41, 4157‐4167.
[13]. C.H. Kim, K.Y. Cho, J.K. Park: Grafting of glycidyl methacrylate onto
polycaprolactone: preparation and characterization. Polymer 2001, 42, 5135‐
5142.
[14]. M. Avella, M.E. Errico, P. Laurienzo, E. Martuscelli, M. Raimo, R. Rimedio:
Preparation and characterization of compatibilised polycaprolactone/ starch
composites. Polymer 2000, 41, 3875‐3881.
[15]. R. Mani, M. Bhattacharya, J. Tang: Functionalization of polyesters with
maleic anhydride by reactive extrusion. J. Polym. Sci., Part A: Polym. Chem.
1999, 37, 1693‐1702.
[16]. C.H. Kim, K.Y. Cho, J.K. Park: Reactive blends of gelatinized starch and
polycaprolactone‐g‐glycidyl methacrylate. J. Appl. Polym. Sci. 2001, 81, 1507‐
1516.
[17]. J. John, J. Tang, Z. Yang, M. Bhattacharya: Synthesis and characterization of
anhydride‐functional polycaprolactone. J. Polym. Sci., Part A: Polym. Chem.
1997, 35, 1139‐1148.
[18]. C.H. Kim, K.M. Jung, J.S. Kim, J.K. Park: Modification of Aliphatic Polyesters
and Their Reactive Blends with Starch. J. Polym. Environ. 2004, 12, 179‐187.
[19]. C.W. Macosko, P. Guegan, A.K. Khandpur, A. Nakayama, P. Marechal, T.
Inoue: Compatibilizers for melt blending: Premade block copolymers.
Macromolecules 1996, 29, 5590‐5598.
[20]. C.S. Wu: Physical properties and biodegradability of maleated‐
polycaprolactone/ starch composite. Polym. Degrad. Stab. 2003, 80, 127–134.
[21]. M.M. Coleman, J.F. Graf, P.C. Painter: Specific Interactions and the Miscibility of
Polymer Blends, Technomic Publication, Pennsylvania, USA, 1991.
[22]. E. Passaglia, M. Marrucci, G. Ruggeri, M. Aglietto: Chemical reactions
affecting the free‐radical grafting of diethyl maleate onto ethylene polymers.
Gazz. Chim. It. 1997, 127, 91‐95.
[23]. D. C. Montgomery: Design and Analysis of Experiments 5th edition, John Wiley &
Sons Inc., New York, USA, 2001.
130

[24]. O.S. Odusanya, D.M.A. Manan, U. S. Ishiaku, B.M.N. Azemi: Effect of starch
predrying on the mechanical properties of starch/ poly(ε‐caprolactone)
composites. J. Appl. Polym. Sci. 2003, 87, 877‐884.
[25]. O.S. Odusanya, U.S. Ishiaku, B.M.N. Azemi, D.M.A. Manan, H.W. Kammer:
On mechanical properties of sago starch/ poly(ε‐caprolactone) composites.
Polym. Eng. Sci. 2000, 40, 1298‐1305.
[26]. U.S. Ishiaku, K.W. Pang, W.S. Lee, M. Ishak: Mechanical properties and
enzymic degradation of thermoplastic and granular sago starch filled poly(ε‐
caprolactone). Eur. Polym. J. 2002, 38, 393‐401.
[27]. E. Passaglia, M. Aglietto, G. Ruggeri, F. Picchioni: Formation and
compatibilizing effect of the grafted copolymer in the reactive blending of 2‐
diethylsuccinate containing polyolefins with poly‐ε‐caprolactam (nylon‐6).
Polym. Adv. Technol. 1998, 9, 273‐281.
[28]. K.D.Safa, M.H.Nasirtabrizi: Ring opening reactions of glycidyl methacrylate
copolymers to introduce bulky organosilicon side chain substituents. Polym.
Bull. 2006, 57, 293‐304.
[29]. A. Tedesco, R.V. Barbosa, S.B.M. Nachtigall, R.S. Mauler: Comparative study
of PP‐MA and PP‐GMA as compatibilizing agent on polypropylene/nylon 6
blends. Polym. Test. 2002, 21, 11‐15.

131
Summary

Plastics made from fossil resources (a.o oil and gas) are very attractive materials
for a broad range of applications. Examples are the use as packaging and
construction materials. Its production has increased significantly since 1950, with a
rate of almost 10% every year. After use, plastics may end up in the environment
and cause serious environmental problems. For instance, it has a high volume to
weight ratio and is generally resistant to degradation. In combination with the
current high prices for petrochemical products, there is a strong need for
renewable alternatives for plastics from fossil resources.
Starch‐based biodegradable materials are considered interesting candidates to
replace certain types of conventional plastics. Starch is relatively cheap and
available from a broad range of plants. Starch is a polymer consisting of
anhydroglucose (AHG) units. There are two types of polymers present in starch:
amylose and amylopectin. Amylose is essentially a linear polymer in which AHG
units are predominantly connected through α‐D‐(1,4)‐glucosidic bonds, while
amylopectin is branched polymer, containing periodic branches linked with the
backbones through α‐D‐(1,6)‐glucosidic bonds. The content of amylose and
amylopectine in starch varies and largely depends on the starch source.
The use of virgin starch for packaging materials is restricted because it cannot
be shaped to films with adequate mechanical properties (high percentage
elongation, tensile and flexural strength). Starch is also too sensitive to water.
Starch must therefore be modified before it can be applied as a biodegradable
plastic. There are several starch modification methods available, such as
thermoplasticization, blending with other materials, chemical modification or
combinations thereof. The application of a number of prospective methods for
developing starch‐based biomaterials will be described in this thesis.
Chemical grafting of biodegradable polyesters on the starch backbone is
expected to result in less hydrophilic and thus less water sensitive materials with
improved mechanical properties. The synthesis of such grafted products by the in
situ ring opening polymerization (ROP) of the monomers on the hydroxyl groups
of starch is unfortunately not very straightforward. The main reasons are the
water sensitivity of common catalysts and the fact that starch is poorly or even
insoluble in the common organic solvents used for ROP. The application of an
alternative method involving hydrophobisation of starch by the introduction of
large hydrophobic SiMe3 groups followed by a ROP with a polyester precursor
and subsequent removal of the SiMe3 groups is expected to result in higher
grafting efficiencies. To gain insight in the potential of this approach, we have
Summary
initially performed research on the ROP with a simpler model system involving a
simple protected mono‐saccharide instead of starch.
The results for the model system involving the ROP of p‐dioxanone initiated by
hydroxyl groups of a protected monosaccharide (1,2;3,4‐di‐O‐isopropylidene‐α‐D‐
galactopyranose) using Al(OiPr)3 as the catalyst, are described in Chapter 2. The
polymerizations were performed at 60‐100°C, and off‐white solid products with
isolated yields of 30‐96% were obtained. The yield of the polymers was a function
of the reaction temperature and the reaction time, with higher temperatures
(100°C) leading to lower yields. Average molecular weights of the products were
between 970 and 6200 (7 – 58 monomer units) and were a clear function of the p‐
dioxanone/ 1,2;3,4‐di‐O‐isopropylidene‐α‐D‐galactopyranose ratio (at constant
Al(OiPr)3 intake), with higher ratios leading to higher molecular weights. A
statistical model has been developed to quantify the effects of process variables
(time, temperature and monomer: monosaccharide ratio) on the average degree of
polymerization. Characterization of the products using 1H‐ and 13C‐NMR as well
as MALDI‐TOF mass spectrometry showed the presence of significant amounts of
p‐dioxanone polymers with an isopropoxide end group (20‐30%)
The knowledge obtained from the model system was applied for the synthesis
of starch‐g‐poly‐ε‐caprolactone using hydrophobised silylated starch. The results
are described in detail in Chapter 3. The synthetic procedure may be divided into
three steps: hydrophobisation of starch by the introduction of SiMe3 groups
followed by in situ ROP, and subsequent removal of the silyl groups. The
silylation reaction was performed using hexamethyl disilazane (HMDS) in DMSO/
toluene mixtures at 70°C. Silylated starch with a low to medium DS (0.46‐0.68) was
obtained. The grafting of ε‐caprolactone to the silylated starch by a ring‐opening
polymerisation catalysed by Al(OiPr)3 was performed in THF at 50oC. Poly‐(ε)‐
caprolactone grafted silylated starch co‐polymers with average chain length of 40‐
55 monomer units (molecular weight of 4500‐6300) were obtained. The DS of the
PCL chains was between 0.21‐0.72, depending on the ε‐CL to starch ratio.
Considerable amounts of ε‐CL homopolymers with isopropyl end‐groups were
also formed. The grafting efficiency for the desired reaction was 28‐58%. The silyl
groups of the poly‐(ε)‐caprolactone grafted starch co‐polymers were finally
successfully removed using a mild dilute hydrochloric acid treatment in THF at
room temperature.
Esterification of starch with carboxylic acid derivatives is one of the oldest
strategies to improve starch properties. Most of the previous starch esterification
studies involved the use of short chain carboxylic acids (C1‐C4), and particularly
acetic acid derivatives (C2) have received considerable attention. The
hydrophobicity of starch acetates is higher than virgin starch, but the products are
still very brittle, even in the presence of plasticizers. The use of higher carboxylic
(fatty) acid to esterify starch resulted in products with significantly improved
134

Summary
mechanical properties and hydrophobicities. The synthesis of these fatty starch
esters is, however, usually performed using fatty acid chlorides which are
relatively expensive and rather corrosive. The use of methyl and glyceryl esters
results in products with only relatively low DS. In Chapter 4‐5, an alternative
method for higher fatty starch esters by using fatty vinyl esters is reported.
Chapter 4 describes the results of a preliminary study on the synthesis of long
fatty esters of corn starch (starch‐laurate and starch‐stearate) using vinyl fatty
esters. The starch esters were prepared by reacting starch with vinyl laurate or
vinyl stearate in the presence of basic catalysts (Na2HPO4, K2CO3, and Na‐acetate)
in DMSO at 110°C. The yellowish products were characterized by 1H‐, 13C‐NMR
and FT‐IR. The products have a broad range of degree of substitution (DS = 0.24 ‐
2.96). The DS of the products was a clear function of the chain length of the fatty
ester, vinyl ester to starch ratio, and the type of catalyst. For low vinyl‐ester to
starch ratios, an increase in the vinyl‐ester concentration led to higher product DS
values. At higher ratios, the DS decreased, presumably due to a reduction of the
polarity of the reaction medium. K2CO3 and Na‐acetate are superior with respect
to activity when compared with Na2HPO4. With these catalysts, products with a
DS > 2.4 could be obtained for both laurate and stearate esters.
In Chapter 5, a systematic study, including statistical modelling on the
synthesis of long fatty esters of corn starch (starch‐laurate and starch‐stearate)
using corresponding vinyl fatty esters is reported. The thermal and mechanical
properties of some representative product samples are also described. The starch
esters were synthesized by reacting gelatinized starch with vinyl laurate or vinyl
acetate in DMSO in the presence of basic catalysts (Na2HPO4, K2CO3, and Na‐
acetate). Statistically adequate (R2≥0.96 and P‐value of ≤10‐7) second‐order
mathematical models correlating the effect of process variables (vinyl to AHG‐
starch mol ratio, reaction temperature, catalyst intake, and catalyst basicity) to the
DS of the starch ester products were developed. The DS of the products is a strong
function of the basicity of the catalyst. Reaction temperature and catalyst intake
also affect the product DS but only to a lesser extent. The use of Na2HPO4 resulted
in low‐medium DS products (0.3‐1.5 for starch laurate, 0.07‐1.5 for starch stearate),
while the use of K2CO3 and CH3COONa catalysts resulted in medium‐high DS
products (2.1‐2.9 for starch laurate, 1.4‐3 for starch stearate). High‐DS products
(DS= 2.26‐2.39) are totally amorphous whereas the low‐DS ones (DS= 1.45‐1.75) are
still partially crystalline. The thermal stability of the esterified products is higher
than that of native starch. Mechanical tests show that the products have tensile
strength (stress at break) between 2.7‐3.5 MPa, elongation at break of 3‐26%, and
modulus of elasticity of 46‐113 MPa.
The last part of this thesis deals with reactive blending of starch and
biodegradable polymers using reactive interfacial agents. In Chapter 6, the
synthesis of two polycaprolactone (PCL)‐based reactive interfacial agents, PCL‐g‐
135
Summary
glycidyl methacrylate (PCL‐GMA) and PCL‐g‐diethyl maleate (PCL‐g‐DEM), and
their use in compatibilising starch‐PCL blends is described. The PCL‐based
compatibilisers were prepared by reacting a low molecular weight PCL (Mw 3000)
with glycidyl methacrylate or diethyl maleate in the presence of benzyolperoxide
initiator at 130°C. A statistically adequate model (R2=0.957, P‐value ≤10‐9) has been
developed to quantify the effects of process variables (monomer and initiator
intake, mutual solubility of the monomer into the molten polymer) on the
functionalisation degree (FD) of the GMA and DEM units to the PCL backbone.
Highest values of the FD (up to 45 %) were observed for PCL‐g‐GMA, at relatively
high GMA and BPO intakes. The FD values for PCL‐g‐DEM were considerably
lower (up to 7 %). Both of the PCL‐based compatibilisers improve the interfacial
adhesion between the starch particles and the PCL matrix in starch/ PCL blends.
As the result, the mechanical behavior of the compatibilised blends is in general
different from that of pure PCL and of the corresponding uncompatibilised
blends. In particular the elastic modulus for the compatibilised blends is
significantly higher (reaching up to 430 MPa) than that of uncompatibilised ones
(320‐340 MPa). At a fixed starch content (20 %‐wt), PCL‐g‐DEM seems to have
slightly more efficient compatibilising effect than PCL‐g‐GMA as shown by blends
morphology and elastic modulus values. This is in contrast with chemical
reactivity and amount of chemical groups along the PCL backbone (both factors
favorable to GMA as compared to DEM) but it is explainable on the basis of the
group distribution along the PCL backbone. The latter hypothesis is indirectly
confirmed by the observation that PCL‐g‐GMA becomes more efficient at
relatively higher starch contents in the blends.

136
Samenvatting

Kunststoffen gemaakt van fossiele grondstoffen als olie en gas zijn zeer
aantrekkelijke materialen voor een breed scala aan toepassingen. De productie van
kunststoffen is sinds 1950 met bijna 10% per jaar gestegen. Na gebruik kunnen de
materialen in het milieu eindigen en grote problemen veroorzaken. Dit, in
combinatie met de huidige hoge prijzen voor de grondstoffen, heeft geleid tot een
grote interesse voor biologisch afbreekbare alternatieven gemaakt uit
hernieuwbare grondstoffen.
Biologisch afbreekbare polymeren op basis van zetmeel kunnen mogelijk
bepaalde conventionele kunststoffen vervangen. Zetmeel is relatief goedkoop en
te winnen uit een breed scala aan planten. Zetmeel is een polymeer bestaande uit
anhydroglucose (AHG) eenheden. De AHG eenheden kunnen op een aantal
manieren aan elkaar gekoppeld worden. Twee type polymeren zijn dominant
aanwezig in zetmeel: amylose en amylopectine. Amylose is een lineair polymeer
waarin de AHG eenheden hoofdzakelijk via α‐D‐(1,4)‐glucoside bindingen
verbonden zijn. Het amylopectine is een vertakt polymeer waarbij de
vertakkingen via α‐D‐(1,6)‐glucoside bindingen met de hoofdketen zijn
verbonden. De verhouding amylose en amylopectine in zetmeel varieert en hangt
grotendeels af van de plant waaruit het zetmeel gewonnen wordt.
Het gebruik van zetmeel voor verpakking materialen is momenteel beperkt,
mede omdat het niet mogelijk is om films met goede mechanische eigenschappen
te maken. Het zetmeel is daarnaast ook erg gevoelig voor water. Modificatie van
zetmeel kan het aantal toepassingen sterk vergroten. Er zijn verschillende
methodes beschikbaar om zetmeel te modificeren. Voorbeelden zijn thermo‐
plastificatie, menging met andere materialen, chemische modificaties en
combinaties van deze methoden. Dit proefschrift beschrijft een aantal methodes
om nieuwe polymere materialen op basis van zetmeel te maken.
Het chemisch grafting van biologisch afbreekbare polyesters op de zetmeel
ketens zou kunnen leiden tot minder hydrofiele en dus minder water gevoelige
materialen met betere mechanische eigenschappen. De synthese van dergelijke
producten met een in situ katalytische ringopening polymerisatie (ROP) van
geschikte monomeren als b.v. ε‐caprolactone op de hydroxy‐groepen van zetmeel
is helaas niet eenvoudig. De belangrijkste redenen zijn de watergevoeligheid van
de katalysatoren en het feit dat zetmeel slecht of zelfs onoplosbaar is in de
typische organische oplosmiddelen die voor de ROP gebruikt worden. De
toepassing van een alternatieve methode waarbij zetmeel eerst minder hydrofiel
gemaakt wordt door de introductie van grote hydrofobe SiMe3 groepen, gevolgd
door de ROP met een polyester precursor en verwijdering van de SiMe3 groepen
Samenvatting
zou in theorie moeten leiden tot een hogere grafting efficiency. Om inzicht in het
potentieel van deze benadering te krijgen is onderzoek verricht naar de ROP met
een eenvoudig modelsysteem bestaande uit een beschermde mono‐saccharide in
plaats van zetmeel.
De resultaten voor het modelsysteem waar de ROP van p‐dioxanone wordt
geïnitieerd door de hydroxy‐groepen van een beschermde mono‐saccharide
(1,2;3,4‐di‐O‐isopropylidene‐α‐D‐galactopyranose) met Al(OiPr)3 als katalysator
worden in Hoofdstuk 2 beschreven. De polymerisaties werden tussen 60‐100°C
uitgevoerd en leverde vaste producten in geïsoleerde opbrengsten tussen de 30 en
96%. De opbrengst van de polymeren was een functie van de reactietemperatuur
en de reactietijd, waarin hogere temperaturen (100°C) tot lagere opbrengsten
leiden. De gemiddelde molecuul massa van de producten lag tussen de 970 en
6200 (7 ‐ 58 monomeereenheden) en was een functie van de verhouding p‐
dioxanone/1,2;3,4‐di‐O‐isopropylidene‐α‐D‐galactopyranose (bij constante
Al(O Pr)3 concentratie). Een hogere verhouding leidde tot hogere molecuul
i
massa’s. De effecten van procesvariabelen (tijd, temperatuur en monomeer:

monosaccharide verhouding) op de gemiddelde polymerisatie graad zijn
gekwantificeerd met een statistisch model. Karakterisering van de producten met
1
H‐ en 13C‐NMR en MALDI‐TOF massaspectrometrie toonden de aanwezigheid
aan van de gewenste producten en daarnaast significante hoeveelheden van p‐
dioxanone polymeren met een isopropoxide eindgroep (20‐30%).
De kennis verkregen met het modelsysteem is toegepast om ε‐caprolactone te
graften op zetmeel. De resultaten worden in Hoofdstuk 3 in detail beschreven. De
gevolgde procedure bestaat uit drie stappen: hydrofobisering van het zetmeel
door de introductie van SiMe3 groepen, gevolgd door in situ ROP met ε‐
caprolacton en daarna de verwijdering van de silyl groepen. De silylerings reactie
werd bij 70°C uitgevoerd met hexamethyl disilazane (HMDS) in DMSO/tolueen
mengsels. De substitutiegraad van de SiMe3 groepen varieerde van laag tot
middelhoog (degree of substitution, DS = 0.46‐0.68). De grafting van ε‐
caprolactone aan het gesilyleerde zetmeel werd in THF bij 50°C uitgevoerd met
Al(OiPr)3 als katalysator. De polycaprolacton eenheden (PCL) van het verkregen
poly‐ε‐caprolactone gesilyleerde zetmeel hadden een gemiddelde ketenlengte van
40‐55 monomeereenheden (molecuul massa 4500‐6300). De DS van de PCL ketens
lag tussen de 0.21‐0.72 en was afhankelijk van de ε‐caprolacton/zetmeel
verhouding. Er werden ook aanzienlijke hoeveelheden homopolymeer van ε‐
caprolacton met isopropyl eindgroepen gevormd. De grafting efficiency van de
gewenste reactie was 28‐58%. De silyl groepen van het poly‐ε‐caprolactone
gesilyleerde zetmeel werden uiteindelijk succesvol verwijderd door middel van
een milde behandeling met verdund zoutzuur in THF bij kamertemperatuur.

138

Samenvatting
De esterificatie van zetmeel met carbonzuur derivaten is één van de oudste
strategieën voor het verbeteren van zetmeeleigenschappen. In de meeste zetmeel
esterificatie studies wordt gebruikt gemaakt van kleine carbonzuren (C1‐C4), en in
het bijzonder azijnzuur derivaten (C2). De acetaat esters zijn in het algemeen
minder hydrofiel dan zetmeel. Echter de producten zijn nog steeds zeer bros, zelfs
in aanwezigheid van plastificeermiddelen. Het gebruik van hogere (vet‐) zuren
resulteerde in producten met beduidend betere mechanische eigenschappen en
hydrofobiciteit. De synthese van deze esters wordt in het algemeen uitgevoerd
met vetzuurchloriden die vrij duur en sterk corrosief zijn. Het gebruik van methyl
en glyceryl esters resulteert in producten met een lage substitutiegraad. In
Hoofdstuk 4‐5, wordt een alternatieve methode voor de synthese van zetmeel
esters met lange vetzuren beschreven.
Hoofdstuk 4 beschrijft de resultaten van een voorstudie naar de synthese van
lange vetzuur esters uit maïszetmeel (zetmeel laureaat en zetmeel stearaat) met de
vinyl esters van de vetzuren al reagens. De zetmeel esters werden bereid door het
zetmeel met vinyl laureaat of vinyl stearaat in aanwezigheid van basische
katalysatoren (Na2HPO4, K2CO3 en Na‐acetaat) in DMSO bij 110°C te laten
reageren. De producten werden met 1H‐, 13C‐NMR en FT‐IR gekarakteriseerd. De
substitutiegraad varieert van 0.24 tot 2.96. De DS van de producten was een
duidelijke functie van de keten lengte van de vetzure vinylester, de vinyl‐
ester/zetmeel verhouding en het type katalysator. Bij lage vinyl‐ester/zetmeel
verhoudingen leidde een verhoging van de vinyl‐ester concentratie in eerste
instantie tot hogere DS waarden. Echter, bij verdere verhoging van de verhouding
verminderde de DS, vermoedelijk als gevolg van een verlaging van de polariteit
van het reactiemengsel. K2CO3 en Na‐acetaat vertonen superieure activiteiten
vergeleken met Na2HPO4. Met deze katalysatoren worden voor zowel de laureaat
als stearaat esters producten met een DS van boven de 2.4 verkregen.
In Hoofdstuk 5, wordt een systematische studie, inclusief statistische
modellering, naar de synthese van lange vetzuur esters uit maïszetmeel (zetmeel
laureaat en zetmeel stearaat. De thermische en mechanische eigenschappen van
sommige representatieve producten zijn uitgebreid bepaald. De zetmeel esters
werden bereid door zetmeel met vinyl laureaat of vinyl stearate in DMSO in de
aanwezigheid van basische katalysatoren (Na2HPO4, K2CO3, en Na‐acetaat) te
laten reageren. De effecten van belangrijke procesvariabelen (vinyl ester/zetmeel
molverhouding, reactietemperatuur, katalysator concentratie en katalysator
basiciteit) op de DS zijn gekwantificeerd met een statistisch model (R2 ≥ 0.96 en P‐
waarde van ≤ 10‐7). De DS van de producten is een sterke functie van de basiciteit
van de katalysator. De temperatuur en de katalysator concentratie beïnvloeden de
DS ook maar in mindere mate. Het gebruik van Na2HPO4 resulteerde in producten
met een laag tot middelhoge DS (0.3‐1.5 voor zetmeel laureaat, 0.07‐1.5 voor
zetmeel stearaat), terwijl het gebruik van de K2CO3 en Na‐acetaat katalysatoren
139
Samenvatting
resulteerde in producten met een middelhoog tot hoge DS waarden (2.1‐2.9 voor
zetmeel laureaat, 1.4‐3.0 voor zetmeel stearaat). De producten met een hoge DS
(2.26‐2.39) zijn volledig amorf terwijl de producten met een DS van 1.45‐1.75 nog
gedeeltelijk kristallijn zijn. De thermische stabiliteit van de geësterificeerde
producten is hoger dan die van zetmeel. Mechanische tests tonen aan dat de
producten een treksterkte hebben variërend van 2.7‐3.5 MPa, een uitrekking tot
breuk van 3‐26%, en een elasticiteit modulus van 46‐113 MPa.
Het laatste hoofdstuk van dit proefschrift (Hoofdstuk 6) behandelt de menging
van zetmeel en biologisch afbreekbare polymeren met reactieve compatibilisers.
Als reactieve compatibilisers zijn twee polycaprolactone derivaten, PCL‐g‐glycidyl
methacrylaat (PCL‐GMA) en PCL‐g‐diethyl maleaat (PCL‐g‐DEM), gebruikt. De
synthese van deze verbindingen wordt uitgebreid beschreven. De PCL gebaseerde
compatibilisers werden bereid door reactie van laag moleculair PCL (Mw 3000)
met glycidyl methacrylaat of diethyl maleaat in de aanwezigheid van een
benzyolperoxide initiator bij 130°C. Er is een statistisch model (R2 =0.957 en P‐
waarde van ≤ 10‐9) ontwikkeld waarmee de effecten van procesvariabelen
(monomeer en initiator concentratie, wederzijdse oplosbaarheid van het
monomeer in het gesmolten polymeer) op de functionalisatie graad (FD) van de
GMA en DEM eenheden aan de PCL keten adequaat gekwantificeerd kunnen
worden. De hoogste FD waarden (tot 45%) werden gevonden voor PCL‐g‐GMA bij
relatief hoge GMA en BPO concentraties. De FD waarden voor PCL‐g‐DEM waren
aanzienlijk lager (tot 7%). Beide reactieve compatibilisers verbeteren de interactie
tussen de zetmeeldeeltjes en de PCL matrix. Dit leidt tot verschillen in
mechanische eigenschappen tussen gecompatibiliseerde mengsels en van de niet‐
gecompatibiliseerde mengsels van zetmeel en PCL. Zo is bijvoorbeeld de
elastische modulus van de gecompatibiliseerde mengsels beduidend hoger (tot
430 MPa) dan die van niet‐gecompatibiliseerde mengsels (320‐340 MPa).

140
Acknowledgements

The synthesis of biodegradable materials is currently receiving a lot of
attention, not in the last place because of the problems related with conventional
plastics. These include the current high oil prices, the raw material for fossil
derived plastics, together with difficulties in managing plastic waste. This thesis is
a summary of my research on developing novel biodegradable polymers from
starch. Although it is indeed a very interesting topic, I found out that developing a
better fully‐biodegradable‐material from starch is not an easy job, since the
mechanical properties as well as hydrophobicity of this starting material is
relatively poor.
First of all, I want to thank my three promoters, Prof. Erik Heeres, Prof. Leon
Janssen, and Prof. Francesco Picchioni. I want to thank Erik Heeres for being more
than a good supervisor. Thank you Erik, for your consistent support, for the
inspiring ideas during the discussions, for upgrading my writing quality, and for
helping me arranging many things when I was not able to do it from Indonesia. I
want to thank Leon Janssen for his hospitality, for his efforts to make every foreign
students feel at home in Groningen, and for always encouraging me to finish this
thesis. I want to thank Francesco Picchioni for being always there when I needed
him to discuss anything, especially on his opinions about the statistical modelling
and mechanical properties. He has also been a good friend and always supported
me during the hard times. I also want to thank Ineke Ganzeveld, who was my first
daily supervisor for about two months, for valuable discussions during my early
PhD period and for handing me over to the right hands.
I also want to give my deep appreciation to my reading committee, Prof. A.A.
Broekhuis, Prof. A.J. Minnaard, and Prof. L. Mosiscki. Thank you very much for
spending your valuable time reading my thesis and for the constructive comments
and corrections you have made, which improved the quality of this thesis.
I want to thank the persons of Building 18 in Nijenborgh 4 for making me feel at
home in Groningen and for their help during my research. I want to thank Marya
van der Duin‐de Jonge, for her hospitality and solving administrative matters,
Erwin Wilbers, Anne Appeldoorn, Marcel de Vries, and Laurens Bosgra for being
always helpful with technical issues and Jan Henk Marsman for the help with
analytical equipment. People from other groups have also been very kind and
supportive to my research. Gert Alberda van Ekenstein and Harry Nijland helped
me a lot with mechanical and thermal property measurements, as well as taking
the SEM pictures. Anno Wagenaar and Wim Kruizinga taught me how to operate
the NMR machines and to interpret the spectra. Anno was also always there when
Acknowledgements
I had questions or needed specific equipments. Albert Kieviet helped me with the
MALDI‐TOF measurements. The chemical and glassware magazijn as well as the
FWN library personnel were always patient, friendly, and helpful. Annette
Korringa, Anneke Toxopeus, and the ISD staff helped me a lot with visum and
residence permit arrangements as well as tax papers.
My special thanks are for Laura Junistia and Jan‐Pieter Drijfhout, master
students who worked with me. Our collaboration has also become part of this
thesis. Many master and PhD students, visiting researchers, and assistant
professors have shared the days with me in Groningen, in the office, in the lab,
and on many occassions. I want to thank Edwin A. Hofman, Farchad H. Mahfud,
Taufik, Iwan Harsono, Niels van Vegten, R. Sari Fabianti, Francesca Gambardella,
Jasper Huijsman, Francesca Fallani, Anna Nizniowska, Hans Heeres, Nidal
Hammoud Hassan, Gerard Kraai, Boelo Schuur, Jelle Wildschut, Judy Retti
Witono, Marcel Wiegman, Diana Santangelo, Oscar Rojas, Abdul Osman, Nadia
Gozali, Diana Jirjis, M. Chalid, Wahyudin, Asal Hamarneh, Yao Jie, Zhang
Youchun, Agnes R. Ardiyanti, Teddy, Wita Sondari, August Kurniawan, Wuri
Raspati, Ais Jenie, C.B. Rasrendra, Ignacio Melian Cabrera, Jaap Bosma, Louis
Daniel, Erna Subroto, Arjan Kloekhorst, and Henk van de Bovenkamp for
cheering my days and being my friends during summer and winter, spring and
autumn in Groningen.
Groningen is also a place where I found true friendship. A lot of thanks for
Henk Stegeman and Poppy Sutanto, Buana Girisuta and Rina Karina, Henky
Muljana (my paranymph) and Anindhita Widyadhana, Vincent Nieborg and Fesia
Lestari, and Iwan Kustiawan for becoming my family. Thanks also for Marcel and
Connie, Ronny Sutarto, bu Ida Susanti, Bima and Wisnu (also my paranymph),
mas Pandu, Tiara, Puspita Kencana, Yongki, and Mahesa for being my brothers
and sisters in Groningen. I have met a lot of Indonesian students in Groningen,
and although we seldom met afterwards, our memories will never fade away.
Thank you AW, Teguh, Kuslan, Taufik, Samuel, Thyara, Mita, Togi, Endy, Tenny,
Dado, Egi, Ita, kak Atha, Wangsa, pak Harry, mbak Mia, Patrick and Oppie, kak
Roga, and many others for the good old times! Thank you Tante Ietje, Tante
Caroline, Tante Oppie, mbak Ika, Tante Smith, and Tante Alma for being like my
parents and family in Groningen.
Thank you God, for never leaving me alone, for your never‐ending blessings,
and for my wife and family. Thank you Tresna, for your love, support, patience,
and passion. Thank you my parents and sister, for everything. You all have always
been my inspiration and motivation.

142
List of Publications

1. A.K. Sugih, F. Picchioni, H.J. Heeres, Experimental Studies on the ring opening
polymerization of p‐dioxanone using an Al(OiPr)3‐monosaccharide initiator
system, 2008, Eur. Polym. J. (in press, DOI: 10.1016/j.eurpolymj.2008.10.010).
2. L. Junistia, A.K. Sugih, R. Manurung, F. Picchioni, L.P.B.M. Janssen, H.J.
stearate as reactants, 2008 (accepted for publication in Starch‐Starke).
3. L. Junistia, A.K. Sugih, R. Manurung, F. Picchioni, L.P.B.M. Janssen, H.J.
Heeres, Experimental and modeling studies on the synthesis and properties of
higher fatty esters of corn starch, 2008 (accepted for publication in Starch‐
Starke).
4. A.K. Sugih, F. Picchioni, L.P.B.M Janssen and H.J. Heeres, Synthesis of poly‐
(ε)‐caprolactone grafted starch co‐polymers by ring‐opening polymerisation
using silylated starch precursors, 2008 (submitted to Carbohydr. Polym.).
5. A.K. Sugih, J.P.Drijfhout, F.Picchioni, L.P.B.M.Janssen, H.J.Heeres, Synthesis
and properties of reactive interfacial agents for polycaprolactone‐starch blends,
2008 (submitted to J. Appl. Polym. Sci.)
6. A.K. Sugih, L. Junistia, F. Picchioni, and H.J. Heeres: Green, starch‐based
innovative product: The synthesis of grafted polycaprolactone‐starch by ring
opening polymerisation, presented in 11th Asia Pacific Confederation of Chemical
Engineering (APCChE) Congress, Kuala Lumpur (Malaysia), 27‐30 August 2006
(oral presentation)
7. J.P. Drijfhout, A.K. Sugih, H.J. Heeres, A.A. Broekhuis, and F. Picchioni:
Functionalization and blending of polycaprolactone, presented in International
Workshop: “From Polymer Modification to Multicomponent Polymer Systems”,
Bratislava (Slovak Republic), 26‐28 November 2006 (oral presentation).
8. J.P. Drijfhout, A.K. Sugih, H.J. Heeres, A.A. Broekhuis, and F. Picchioni:
Biodegradable polymer materials based on polyesters and starch/proteins,
presented in “Polymers for Advanced Technologies”, 9th National Conference
organized by the Society for Polymer Science, India, Pune Chapter, Mumbai (India),
17‐20 December 2006 (oral presentation).
9. L. Junistia, A.K. Sugih, F. Picchioni, and H.J. Heeres: Synthesis and properties
of novel green biopolymers derived from starch and fatty acids, presented in
the 1st International Symposium on Sustainable Chemical Product and Process
Engineering, South China University of Technology, Guangzhou (China), 25‐28
September 2007 (oral presentation).

Thesis

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Thesis

Uploaded by

Copyright:

Available Formats

H2COH H2COH H2COH H2COH

H2COH H2COH CH2

appropriate environment and sufficient amount of time [15]. Biodegradable

polymerization, X n exp of the product.

exp A H-repeating unit

exp [ A 3.4− 4.7 ppm − 6 × A 5.6 ppm ] / 6

A H − galactopyranose end group

(2) (1) (3)

Scheme 2.1. Schematic representation of the polymerization reaction including

polydioxanone polymer backbone and a monosaccharide end group. The carbon

Galactopyranose p‐dioxanone Poly(p‐dioxanone) end‐capped

Galactopyranose p‐dioxanone Poly(p‐dioxanone) end‐capped

The minor product, OiPr‐end‐capped poly(p‐dioxanone), will be formed when

Set dioxanone/ Avg. Chain 2/OiPr ratio

2.3.2.2. Effects of process conditions on the average chain length ( X n exp ) and end group

60 Temperature: 100oC, time: 16 hr

Temperature: 60oC, time: 16 hr

50 Temperature: 60oC, time: 1.5 hr

SS DF MS F P‐value

1 ACH 3 silyl / 9 7 A−0.6−0.6 ppm

ACH 2 −ε + ACH 2 −ε ' A3.8− 4.2 ppm

0.5. A3.3−3.4 − A4 .8− 4 .9

ACH 2 −ε ' / 2 − ACH −iso

Starch Hexamethyldisilazane Silylated Starch

HMDS intake HMDS: AHG Product DS

Silylated Starch ε-Caprolactone Silylated

Silylated Starch [R = H or Si(CH3)3 ]

ε‐CL/ OH Total Avg. Chain Length Grafting Efficiency

Silylated Starch-g-Polycaprolactone Starch-g-Polycaprolactone

agents. A novel route to functionalised nanoparticles. Macromolecul. Rapid

DS fatty ester = 3 − DS acetate (4.1.)

(1) R = laurate (C12)

b. starch laurate C-O

4000 3500 3000 2500 2000 1500 1000 500

Figure 4.6. Comparison of DS values with different catalysts (Vinylester:

apparatus (Fontijne, Holland), operated at 150 oC and 150 bar for 3 minutes. The

(1) R = laurate (C12)

Vinyl : AHGa Catalyst Temperature

45 stearate CH3COONa 4.00 3.50 95 1.98

Coefficient Starch Laurate Starch Stearate

β0 9.1748 17.1082

SS DF MS F P‐value R2 values

SS DF MS F P‐value R2 values

Starch ester DS Tm [˚C] Tc [˚C]

Stress at break Elongation at break Modulus

Corn Starch Potato Starch Corn Starch Potato Starch

Stress 3.0 3.5 0.7 3.2 2.7 3.7 1.9

number of moles of GMA/DEM attached to PCL (mol)

PCL ‐ ‐ ‐ 27 75 51 79

PCL ‐ ‐ ‐ 27 75 51 79

H3C C C O CH2 HC CH2

GMA, BPO=0.6% mol/mol-CL units

15 GMA, BPO=0.3% mol/mol-CL units

GMA, GMA=24% mol/mol-CL units

FD = −1.8875 + 0.0325nmδ + 0.5431niδ + 1.6022nm ni − 0.2729nm niδ (6.3.)

SS DF MS F P‐value

BPO Intake (mol% on CL units)

σ ε E Tc ∆Hc T m ∆Hm

a. b. c.

a. b. c.

a. b. c.

OH Starch particle Starch particle

massa’s. De effecten van procesvariabelen (tijd, temperatuur en monomeer: