University of Groningen Synthesis and properties of starch based … · 2016-03-08 ·...
Transcript of University of Groningen Synthesis and properties of starch based … · 2016-03-08 ·...
University of Groningen
Synthesis and properties of starch based biomaterialsSugih, Asaf Kleopas
IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite fromit. Please check the document version below.
Document VersionPublisher's PDF, also known as Version of record
Publication date:2008
Link to publication in University of Groningen/UMCG research database
Citation for published version (APA):Sugih, A. K. (2008). Synthesis and properties of starch based biomaterials. s.n.
CopyrightOther than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of theauthor(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).
Take-down policyIf you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediatelyand investigate your claim.
Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons thenumber of authors shown on this cover page is limited to 10 maximum.
Download date: 09-04-2020
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
viii
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 ( expnX )
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
3.1. Introduction 48
3.2. Materials and Methods 49
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
ix
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
3.5. Nomenclature 64
3.6. References 65
Chapter 4: Synthesis of Higher Fatty Acid Starch Esters using Vinyl Laurate and Stearate as Reactants
4.1. Introduction 70
4.2. Materials and Methods 71
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
4.5. Nomenclature 83
4.6. References 84
x
Chapter 5: Experimental and Modeling Studies on the Synthesis and Properties of Higher Fatty Esters of Corn Starch
5.1. Introduction 86
5.2. Materials and Methods 87
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
5.5. Nomenclature 102
5.6. References 103
Chapter 6: Synthesis and Properties of Reactive Interfacial Agents for Polycaprolactone‐Starch Blends
6.1. Introduction 108
6.2. Materials and Methods 109
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
xi
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
6.5. Nomenclature 128
6.6. References 129
Summary 133
Samenvatting (Dutch Summary) 137
Acknowledgements 141
List of Publications 143
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
2
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]).
1995 2000 2005 20100
10
20
30
40
50
60
70
80
Year
Sta
rch
Pro
duct
ion
(Mill
ion
Tons
)
Figure 1.1. Past, present and a forecast of starch production [7]
o : Europe □ : USA
∆ : the Rest of the World : Total
Introduction
3
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.
Figure 1.2. Corn starch production [9]
a. Typical corn‐milling operation b. Cross‐sectional view of a corn kernel
a. b.
Chapter 1
4
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].
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
OH
O ......
OCH2
HH
OH
H
H
OH
H
OH
O
OCH2
HH
OH
H
H
OH
H
OH
O...
OCH2
HH
OH
H
H
OH
HOH
O ...
OCH2
HH
OH
H
H
OH
H
a. b.
OHO
CH2
HH
OH
H
H
OH
H
OH
O...
OCH2
HH
OH
H
H
OH
H
OH
O
OCH2
HH
OH
H
H
OH
H
OH
O...
OCH2
HH
OH
H
H
OH
H
O
O ...
OCH2
HH
OH
H
H
OH
H
c.
Introduction
5
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]
Starch Gelatinization 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
Chapter 1
6
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.)
1950 1960 1970 1980 1990 2000 20100
50
100
150
200
250
Year
Pla
stic
s P
rodu
ctio
n (M
illio
n To
ns)
1950: 1.5
1976: 50
1989: 100
2002: 200
2006: 245Included:Thermoplastics, Polyurethanes, Thermosets, Elastomers,Adhesives, Coatings, Sealants, and PP-Fibers
Not Included:PET-, PA-, and PolyAcryl- Fibers
Figure 1.4. Worldwide Plastics Production (1950‐2006) [11]
Introduction
7
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
Chapter 1
8
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/
Introduction
9
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.
Chapter 1
10
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
Introduction
11
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
Chapter 1
12
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].
+ +OHStarch OH-
O-
Starch OH2
O-
Starch +
StarchO
O
CCH3 +OH2
H
OCH3 C + OH
-
StarchO
O
CCH3 + OH-
OH
O-
StarchOCCH3
+OH2
O
OHCCH3 O-
Starch
CH2CHO
O
CCH3 CH2CHO
StarchO
O-
CCH3
Scheme 1.1. Mechanism of starch acetylation using vinyl acetate
Introduction
13
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
Chapter 1
14
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.
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
Matrix
Dispersed Phase
Introduction
15
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.
Chapter 1
16
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(OiPr)3 as the catalyst. The results of this study have been used as input for the 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
Introduction
17
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
Chapter 1
18
[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,
Introduction
19
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
Chapter 1
20
[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.
Introduction
21
[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
Chapter 1
22
[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
Introduction
23
[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.
Chapter 1
24
[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.
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
26
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
Ring Opening Polymerization of p‐dioxanone using an Al(OiPr)3‐Monosaccharide Initiator System
27
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
Chapter 2
28
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, theonX in terms of number of
monomer units is calculated as follows [6]:
03
0
0
][]Pr)([3][
] [][
sugarOAlmonomer
conversionmonomer
initiatortotalmonomer
conversionmonomerX
i
theon
+×=
×= (2.1.)
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
0
0
][][
sugar
monomerconversionmonomerX theo
n ×= (2.2.)
1H‐NMR was applied to determine the experimental average degree of polymerization, exp
nX of the product. exp
nX 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:
groupendxideH-isopropogroupendyranoseH-galactop
unitgH-repeatinexpn
A A A
X+
= (2.3.)
Ring Opening Polymerization of p‐dioxanone using an Al(OiPr)3‐Monosaccharide Initiator System
29
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:
ppmppm
ppmppmexpn
A A A A
X1.56.5
6.57.44.3 6/]6[+
×−= − (2.4.)
The ratio of the two different types of polymers (either end‐capped with galactopyranose or an isopropoxide group), gpR , is calculated using:
groupenddeisopropoxiH
groupendanosegalactopyrHgp A
A R
−
−= (2.5.)
or, in term of the NMR resonances :
ppm
ppmgp A
A R
1.5
6.5= (2.6.)
2.3. Results and Discussions 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
CH2OH
HO
H O
H
H
O
H
O +
(1)(2) (3)
O
O
O
1
23
456
α
β
γ
1
23
456
α β γ α' β' γ'n OHCH2CH2OCH2
O
] COCH2CH2OCH2
O[ C
O
CH2
HO
H O
H
H
O
H
O
O
n+1Al(OiPr)3
catalyst
Scheme 2.1. Schematic representation of the polymerization reaction including atom numbering scheme
Chapter 2
30
+iPr
iPr
iPrO
O
O
Al
O
CH2OH
HO
H O
H
H
O
H
O3
O
CH2
HO
H O
H
H
O
H
O
O
O
O
Al
O
CH2
HO
H O
H
H
O
H
O
O
CH2
HO
H O
H
H
O
H
O
+ 3 OHiPr
(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 1H‐ and 13C‐NMR spectra of 3 are shown in Figure 2.1 and 2.2., respectively.
1H‐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.
Ring Opening Polymerization of p‐dioxanone using an Al(OiPr)3‐Monosaccharide Initiator System
31
Figure 2.1. 1H‐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
13C‐NMR (Figure 2.2.) is more informative and clearly shows the presence of a
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’).
Chapter 2
32
Figure 2.2. 13C‐ 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).
2D‐NMR (HSQC) was applied for complete peak assignment of the product. A 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.
Ring Opening Polymerization of p‐dioxanone using an Al(OiPr)3‐Monosaccharide Initiator System
33
Figure 2.3. HSQC 13C‐1H spectra of 3
Table 2.1. 1H‐NMR peak assignments a
Reactant peaks (ppm) Product peaks (ppm)
Galactopyranose (2)
p‐dioxanone Monomer
(1)
Poly(p‐dioxanone) end‐capped with protected galactopyranose
(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.
Chapter 2
34
Table 2.2. 13C‐NMR peak assignments
Reactant Peaks (ppm) Product Peaks (ppm)
Galactopyranose (2)
p‐dioxanone monomer
(1)
Poly(p‐dioxanone) end‐capped with protected galactopyranose
(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:
nzm 10226023/ ++= (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.
Ring Opening Polymerization of p‐dioxanone using an Al(OiPr)3‐Monosaccharide Initiator System
35
600 800 1000 1200 1400 1600 1800 2000 2200 24000
10
20
30
40
50
60
70
80
90
100
+102
Mass (m/z)
%-In
tens
ity
1098.5
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:
nzm 1026023/ ++= (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+.
nzm 1021823/ ++= (2.10.)
nzm 102181/ ++= (2.11.)
Chapter 2
36
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
600 700 800 900 1000 1100 12000
10
20
30
40
50
%-In
tens
ity
Mass (m/z)
792.7
796.7
858.3
836.1
Ring Opening Polymerization of p‐dioxanone using an Al(OiPr)3‐Monosaccharide Initiator System
37
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.
iPrOiPr
iPr
O
OAl
ROO
OAl
+ R OH
iPr OH+
+
(in excess)
ROO
OAl
+
R-OH
iPrOO
OAl
O
O
O
+ HX (termination)
+
XO
OAl
product
+O
O
O
+ m
O
O
O
by-product
n OR]
O
CCH2OCH2CH2[OO
OAl
n+m OR]
O
CCH2OCH2CH2[OO
OAl
nH OR]
O
CCH2OCH2CH2[O n+m RO]H
O
CCH2OCH2CH2[Ol iPrO]H
O
CCH2OCH2CH2[O
n l
Figure 2.6. Simplified reaction scheme for the ROP of p‐dioxanone catalyzed by Al(OiPr)3. (R = monosaccharide 2)
Chapter 2
38
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 Sample
2/Al ratio (mol/mol)
t (h)
T (°C)
dioxanone/ 2 ratio
(mol/mol)
Avg. Chain Length
( expnX )
Yieldb (%)
2/OiPr ratio Rgp
(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 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
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 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
2
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 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
3
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
Ring Opening Polymerization of p‐dioxanone using an Al(OiPr)3‐Monosaccharide Initiator System
39
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.
Chapter 2
40
2.3.2.2. Effects of process conditions on the average chain length ( expnX ) and end group
distribution
The effects of the p‐dioxanone: 2 ratio, reaction time and temperature on the exp
nX of the products is shown in Figure 2.7. The expnX increases linearly with
respect to the dioxanone : 2 ratio, as expected for a typical ring opening polymerisation [18]. In Figure 2.7., the theo
nX (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.
0 10 20 30 40 50 60 700
10
20
30
40
50
60
dioxanone : 2 ratio (mol/mol)
Xn ex
p
Temperature: 100oC, time: 16 hr
Temperature: 60oC, time: 16 hr
Temperature: 60oC, time: 1.5 hr
Xn theo (90%-conversion)
Xn theo (100%-conversion)
Figure 2.7. Average Chain Length ( expnX ) 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 gpR (eq. 2.5. and 2.6.). The gpR values (Table 2.3.) for all experiments are scattered randomly between 2.4 and 3.3. A clear trend between gpR 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
Ring Opening Polymerization of p‐dioxanone using an Al(OiPr)3‐Monosaccharide Initiator System
41
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 exp
nX 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 exp
nX :
)(097.0)(177.0)ratio :dioxanone(854.049.8exp TtX n ×−×+×+= 2 (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.).
Table 2.4. Analysis of variance (ANOVA) for linear model of expnX 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
Chapter 2
42
0 10 20 30 40 50 600
10
20
30
40
50
60
Xn exp
Xn m
odel
Figure 2.8. Modeled versus experimental values for the average chain length Xn
The model predicts that the nX 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 exp
nX is positively influenced by the polymerization time, which is in agreement with the available data on ring opening polymerization [15, 16, 17]. Furthermore, the exp
nX 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
Ring Opening Polymerization of p‐dioxanone using an Al(OiPr)3‐Monosaccharide Initiator System
43
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 exp
nX . 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
HA : peak area of certain proton in 1H‐NMR spectra [‐]
ppmyxA − : peak area of certain peak at x until y ppm in 1H‐NMR spectra [‐]
gpR : 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] exp
nX : experimental average degree of polymerization the polymer products [monomer units]
theonX : 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.
Chapter 2
44
[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.
Ring Opening Polymerization of p‐dioxanone using an Al(OiPr)3‐Monosaccharide Initiator System
45
[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.
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
48
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
Synthesis of Poly‐(ε)‐caprolactone Grafted Starch Co‐polymers using Silylated Starch Precursors
49
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. Materials and Methods 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.
Chapter 3
50
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): 1H‐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): 1H‐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
Synthesis of Poly‐(ε)‐caprolactone Grafted Starch Co‐polymers using Silylated Starch Precursors
51
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 oC. 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): 1H‐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): 1H‐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
Chapter 3
52
(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) 1H‐ NMR spectra were recorded in CDCl3 at 50oC or in DMSO d‐6 at 60 oC on a
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.
ppm
ppm
protonstarch
silylCH
AA
AA
8.53
6.06.0
silylation 27
77/
9/31DS 3
−
−−×=×= (3.1.)
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:
1ACL4.33.3
2.48.3
'
'
2
22 +=+
=−
−
−
−−
ppm
ppm
CH
CHCH
AA
AAA
ε
εε (3.2.)
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.)
Synthesis of Poly‐(ε)‐caprolactone Grafted Starch Co‐polymers using Silylated Starch Precursors
53
( )⎥⎥⎦⎤
⎢⎢⎣
⎡
−×
+
−=
−−
−−
silylation
24834333
94844333PCL
DS31
ratio CL/OH-ε).(5.0
.5.0DS
....
....
AA
AA (3.3.)
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:
% 1002/
1 % 1002/
1
% 1002/
2/ GE
4.33.3
9.48.4
'
'
'
2
2
2
×⎟⎟⎠
⎞⎜⎜⎝
⎛−=×⎟
⎟⎠
⎞⎜⎜⎝
⎛−=
×−
=
−
−
−
−
−
−−
AA
AA
AAA
CH
isoCH
CH
isoCHCH
ε
ε
ε
(3.4.)
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]:
2/1
⎟⎟⎠
⎞⎜⎜⎝
⎛ −∆=
m
v
VRTH
δ
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)
3.3. Results and Discussions 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.
Chapter 3
54
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].
OHO
CH2
HH
OH
H
H
OH
HDMSO
50 deg. C
R
R
R
OO
CH2
HH
O
H
H
O
H
[R = H or Si(CH3)3]
+
Starch
CH3
CH3
CH3
SiCH3
CH3
CH3
NHSi
Hexamethyldisilazane
+ NH3
1
23
4 5
6
1
23
4 5
6
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.
Synthesis of Poly‐(ε)‐caprolactone Grafted Starch Co‐polymers using Silylated Starch Precursors
55
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.
Chapter 3
56
1.5 2 2.5 3 3.5 4 4.50.4
0.45
0.5
0.55
0.6
0.65
DS
sily
latio
n
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
Experiment HMDS intake (ml, mmol)
HMDS: AHG mol ratio
Product DS (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
Synthesis of Poly‐(ε)‐caprolactone Grafted Starch Co‐polymers using Silylated Starch Precursors
57
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.
R
R
R
OO
CH2
HH
O
H
H
O
H
(R = H or Si(CH3)3
Silylated Starch
(1)
+O
O
ε-CaprolactoneMonomer
Catalyst: Al(OiPr)3
Toluene, 50 oC
[R = H or Si(CH3)3 ]
n
Silylated Starch-g-Polycaprolactone
(2)
HOOC (CH2)5
RO
R
O
OCH2
HH
O
H
H
H
(3.6.)
Chapter 3
58
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.
Synthesis of Poly‐(ε)‐caprolactone Grafted Starch Co‐polymers using Silylated Starch Precursors
59
O
CH2
HH
OR
H
H
OR
H
[
OO
α β χ δ ε
[iPr O]nCH2CH2CH2CH2CH2
O
O C OHCH2CH2CH2CH2CH2
O
C
Silylated Starch-g-PCL
(2)
PCL homopolymer
α β χ δ ε
OH]nCH2CH2CH2CH2CH2
O
O C CH2CH2CH2CH2CH2
O
O Cα' β' χ' δ' ε'
α' β' χ' δ' ε'
iPr
iPrO
O
O
Al
O
CH2
HH
OR
H
H
OR
H
OO
O
O
OH
O
CH2
HH
OH
H
H
OH
H
OO
iPr
iPr
iPrO
O
O
AlO
O
+
+
Silylated Starch
(1)
+
+ OHiPriPr
iPr
iPrO
O
O
Al
iPr
iPrO
O
O
Al
O
CH2
HH
OR
H
H
OR
H
OO
1. Exchange Reaction:
2. Polymerization Reaction:
n+1
n+1
[R = H or Si(CH3)3 ]
[R = H or Si(CH3)3 ]
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).
[H O]nCH2CH2CH2CH2CH2
O
O C OHCH2CH2CH2CH2CH2
O
C
PCL homopolymer with carboxylic end-group
α β χ δ ε α' β' χ' δ' ε'
iPr
iPr
iPrO
O
O
Al
O
O
+
iPr
iPr
OH
O
O
Al +
OH2 + OHiPriPr
iPr
OH
O
O
Al
1. Exchange Reaction:
2. Polymerization Reaction:
(3.7.)
Chapter 3
60
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
Experiment ε‐CL/ OH [mol/ mol]
Total Yieldb [%]
Avg. Chain Length [mon. units]
DSPCL Grafting Efficiency [%]
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).
Synthesis of Poly‐(ε)‐caprolactone Grafted Starch Co‐polymers using Silylated Starch Precursors
61
12 14 16 18 20 22 24 26 28 3030
35
40
45
50
55
ε-CL : Starch-OH Ratio [mol/ mol]
Ave
rage
Pol
ymer
Len
gth
[mon
. uni
ts]
12 14 16 18 20 22 24 26 28 30
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
DS
PCL
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.
Chapter 3
62
RO]n
O
C[(CH2)5OAl + R OH RO]nOC[(CH2)5OHROAl +
ROAl +O
O
n RO]n
O
C[(CH2)5OAl
kp
kd
ktr(a)
ktr(a)
iPrOAl + R OH ROAl + iPr OH a.
b.
c.
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].
Synthesis of Poly‐(ε)‐caprolactone Grafted Starch Co‐polymers using Silylated Starch Precursors
63
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. 1H‐ NMR Spectrum of Starch‐g‐PCL (reaction product after desilylation of sample SN1CL2, DSsilylation =0.68, DSPCL=0.34), in DMSO‐d6
n
THF, 2 hr
n
Starch-g-Polycaprolactone
(3)
H+
Silylated Starch-g-Polycaprolactone
(2)
HOOC (CH2)5
OH
H
O
OCH2
HH
O
H
H
H
HOOC (CH2)5
RO
R
O
OCH2
HH
O
H
H
H
[R = H or Si(CH3)3 ]
(3.8.)
Chapter 3
64
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 [‐]
ppmyxA − : peak area of certain peak at x until y ppm 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 [‐]
silylationDS : DS of silyl group substituents [‐]
PCLDS : 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]
mV : molar volume [cm3/ mol]
Synthesis of Poly‐(ε)‐caprolactone Grafted Starch Co‐polymers using Silylated Starch Precursors
65
Greek symbols:
vH∆ : heat of vaporization [kJ/mol]
δ : Hildebrand solubility parameter [cal1/2cm‐3/2]
3.6. References [1]. 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.
[2]. H. Yavuz, C. Babac: Preparation and biodegradation of starch/polycaprolactone films. J. Polym. Environ. 2003, 11, 107‐113.
[3]. Website of International Starch Institute, Aarhus, Denmark: 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.
Chapter 3
66
[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
Synthesis of Poly‐(ε)‐caprolactone Grafted Starch Co‐polymers using Silylated Starch Precursors
67
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.
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 1H‐, 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
70
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
Synthesis of Higher Fatty Acid Starch Esters using Vinyl Laurate and Stearate as Reactants
71
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. Materials and Methods 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 1H‐ 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.
Chapter 4
72
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): 1H‐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). 1H‐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). 13C‐NMR (before peracetylation, CDCl3): δ 14.0 (C12), 22.7 (C11), 24.9 (C3), 28‐32 (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) 1H‐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)
Synthesis of Higher Fatty Acid Starch Esters using Vinyl Laurate and Stearate as Reactants
73
1H‐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) 13C‐NMR (before peracetylation, CDCl3): δ 14.0 (C18), 22.7 (C17), 25.0 (C3), 26‐32 (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.
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
1
2
3
4
5
6
7
8
9
10
11
12
R
R
O
O
O
O
O
O
O
CH3
1S
2S3S
4S5S
6S
Starch-Laurate
1
2
3
4
5
6
7
8
9
10
11
12
OCH3
R
O
O
O
O
O
O
O
CH3
1' 2'
Starch-Laurateperacetylated
1S
2S3S
4S5S
6S
O
R
R
O
O
O
O
O
O
CH31
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
1S2S
3S
4S5S
6S
Starch-Stearate
Chapter 4
74
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.).
acetateesterfatty DS3DS −= (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.:
starchH
acetateH
AA
−
−=acetateDS (4.2.)
76.56.3 −
− =A
A starchH (4.3.)
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.).
326.28.1 C
acetateHAA
A−
= −− (4.4.)
fattyesterHC AA −×= 22 (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.:
Synthesis of Higher Fatty Acid Starch Esters using Vinyl Laurate and Stearate as Reactants
75
21218.15.0123 −−
− ==AA
A CCfattyesterH (laurate) (4.6.)
33338.15.0183 −−
− ==AA
A CCfattyesterH (stearate) (4.7.)
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.
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.
R
O
O CH2
CH3 O
H
O
OHOH
OH
n
O
OH
OH
OR
On
+ +
(1) R = laurate (C12)
(2) R = stearate (C18)
Chapter 4
76
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.
Synthesis of Higher Fatty Acid Starch Esters using Vinyl Laurate and Stearate as Reactants
77
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
Chapter 4
78
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.
5001000150020002500300035004000
C-O
C-O
C-O
C=O
C=O
C-H
C-H
C-H-OH
Wavenumber (cm-1)
a. native starch
b. starch laurate
c. starch stearate
Abso
rban
ce
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
Synthesis of Higher Fatty Acid Starch Esters using Vinyl Laurate and Stearate as Reactants
79
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
Exp. Vinyl Ester Catalyst Vinyl Ester: AHG ratio [mol/ mol]
Amount of Toluene added
[ml] DS
1 laurate Na2HPO4 2 ‐ 1.13 2 laurate Na2HPO4 3 ‐ 1.23 3 laurate Na2HPO4 6 ‐ 0.90 4 laurate Na2HPO4 2 5 0.99 5 laurate Na2HPO4 3 5 1.07 6 laurate Na2HPO4 6 5 0.90 7 stearate Na2HPO4 2 ‐ 1.08 8 stearate Na2HPO4 3 ‐ 1.05 9 stearate Na2HPO4 4 ‐ 0.91 10 stearate Na2HPO4 6 ‐ 0.60 11 stearate Na2HPO4 2 5 1.01 12 stearate Na2HPO4 3 5 0.57 13 stearate Na2HPO4 4 5 0.68 14 stearate Na2HPO4 6 5 0.24 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.
Chapter 4
80
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.
0 1 2 3 4 5 60
0.5
1
1.5
Deg
ree
of S
ubst
itutio
n (D
S)
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
Synthesis of Higher Fatty Acid Starch Esters using Vinyl Laurate and Stearate as Reactants
81
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
Chapter 4
82
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.
Laurate Stearate0
0.5
1
1.5
2
2.5
3
Vinyl-Ester
Deg
ree
of S
ubst
itutio
n
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.
Synthesis of Higher Fatty Acid Starch Esters using Vinyl Laurate and Stearate as Reactants
83
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 [‐]
ppmyxA − : peak area of certain peak at x until y ppm in 1H‐NMR spectra [‐]
DS : Degree of Substitution, average value of mole of substituted –OH per mole of anhydroglucose (AHG) units [‐]
esterfatty DS : DS of fatty ester (laurate or stearate) group substituents [‐]
acetateDS : DS of acetate group substituents after peracrtylation [‐]
4.6. References [1]. 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.
[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.
[5]. 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.
[6]. A. D. Sagar, E. W. Merrill: Properties of Fatty‐Acid Esters of Starch. J. Appl. Polym. Sci. 1995, 58, 1647‐1656.
Chapter 4
84
[7]. 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.
[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 esters. Cellulose 2004, 11, 255‐263.
[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.
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
86
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.
Experimental and Modeling Studies on the Synthesis and Properties of Higher Fatty Esters of Corn Starch
87
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. Materials and Methods 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 1H‐ 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 [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
Chapter 5
88
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 ( 1x ), catalyst intake ( 2x ), and reaction temperature ( 3x ). 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
Experimental and Modeling Studies on the Synthesis and Properties of Higher Fatty Esters of Corn Starch
89
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 ( 4x ). The ranges for the individual variables ( 1x ‐ 4x ) 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:
∑ ∑ ∑∑≠
+++=i i j jk
kjjkiiiii xxxxy ββββ 20 (5.1.)
Here, βi, βii and βjk are the regression coefficients obtained by a multiple regression procedure. One of the terms, namely 2
4x , 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.
Chapter 5
90
R
O
O CH2
CH3 O
H
O
OHOH
OH
n
O
OH
OH
OR
On
+ +
(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 1x , catalyst intake 2x , and reaction temperature 3x ) 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 ( 1x ‐ 3x ) are shown in Table 5.1. An additional variable 4x (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
Full Factorial Variables Additional Variable Variable
Level Vinyl ester to AHG‐starch ratio [mol/mol]
1x
Catalyst intake [%‐w]a
2x
Reaction temperature [°C]
3x
Catalyst basicity (pKb)b
4x
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.
Experimental and Modeling Studies on the Synthesis and Properties of Higher Fatty Esters of Corn Starch
91
Table 5.2. Overview of experimental data
Experiment Vinyl Ester Catalyst Vinyl : AHGa
[mol/mol] Catalyst [%‐w]
Temperature [˚C]
DS
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 38 stearate CH3COONa 3.00 5.00 80 2.12 39 stearate CH3COONa 3.00 2.00 110 2.44 40 stearate CH3COONa 3.00 5.00 110 2.56 41 stearate CH3COONa 5.00 2.00 80 1.41 42 stearate CH3COONa 5.00 5.00 80 1.40 43 stearate CH3COONa 5.00 2.00 110 1.82 44 stearate CH3COONa 5.00 5.00 110 2.79
Chapter 5
92
45 stearate CH3COONa 4.00 3.50 95 1.98 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.
2413
2312
22114310
429328417316215443322110
xxxxx
xxxxxxxxxxxxxxDS
ββββ
ββββββββββ
++++
+++++++++=
(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
Experimental and Modeling Studies on the Synthesis and Properties of Higher Fatty Esters of Corn Starch
93
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
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
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
2
2.5
3
DS actual
DS
pre
dict
ed
0 0.5 1 1.5 2 2.5 30
0.5
1
1.5
2
2.5
3
DS actual
DS
pre
dict
ed
a. starch laurate b. starch stearate
Chapter 5
94
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.
Experimental and Modeling Studies on the Synthesis and Properties of Higher Fatty Esters of Corn Starch
95
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
a. starch laurate
b. starch stearate
Chapter 5
96
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.
-20 -10 0 10 20 30 40Temperature (o C)
End
othe
rmic
E
xoth
erm
ic Tm = 21 oC
Tc = 12 oC
heating
cooling
Figure 5.3. DSC analysis of starch stearate esters ( ) DS 1.45 (. . .) DS ( 2.39)
Experimental and Modeling Studies on the Synthesis and Properties of Higher Fatty Esters of Corn Starch
97
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.).
Chapter 5
98
20 100 200 300 400 500 600 7000
20
40
60
80
100
Temperature (oC)
%-w
eigh
t
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.).
Experimental and Modeling Studies on the Synthesis and Properties of Higher Fatty Esters of Corn Starch
99
Table 5.6. Mechanical properties of the starch esters
Starch Ester DS Stress at break [MPa]
Elongation at break [%]
Modulus [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.
Chapter 5
100
0 5 10 15 20 25 30 35
0
1
2
3
4
5
6
strain (%)
stre
ss (M
Pa)
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.).
Experimental and Modeling Studies on the Synthesis and Properties of Higher Fatty Esters of Corn Starch
101
Table 5.7. Comparison of the mechanical properties of corn starch ester (this study) with potato starch ester [17]
Corn Starch Laurate Ester
Potato Starch Laurate Ester
Corn Starch Stearate Ester
Potato Starch Stearate Ester
Property Low DS (1.75)
High DS (2.26)
Low DS
High DS (2.7)
Low DS (1.45)
High DS (2.39)
Low DS (1.8)
High DS (2.7)
Stress at Break [MPa]
3.0 ±2.2
3.5 ±1.0
n.a.a. 0.7 ±0.4
3.2 ±0.1
2.7 ±1.0
3.7 ±0.6
1.9 ±0.3
Elongation at Break [%]
8.7 ±3.3
26.1 ±15.2
n.a.a. 1500 ±8.6
21.7 ±4.1
3.1 ±1.3
9 ±2
10 ±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.
Chapter 5
102
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]
0β : constant regression coefficients obtained by a multiple regression procedure [‐]
iβ : regression coefficients of variable xi obtained by a multiple regression procedure [various units]
iiβ : regression coefficients of quadratic variable xi2 obtained by a multiple regression procedure [various units]
jkβ : 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 [‐]
Experimental and Modeling Studies on the Synthesis and Properties of Higher Fatty Esters of Corn Starch
103
esterfatty DS : DS of fatty ester (laurate or stearate) group substituents [‐]
pKb : base dissociation constant [‐]
T : temperature [oC]
Tc : crystallization temperature [oC]
Tm : melting temperature [oC]
ix : experimental variables used for mathematical modelling using multiple regression procedure [various units]
Greek symbols:
σy : stress at yield [MPa]
5.6. References [1]. 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.
[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.
[6]. Website of International Starch Institute, Aarhus, Denmark: 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.
Chapter 5
104
[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.
[14]. 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.
[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.
[17]. 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.
[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.
[19]. 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.
[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
Experimental and Modeling Studies on the Synthesis and Properties of Higher Fatty Esters of Corn Starch
105
[23]. 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.
[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 esters. Cellulose 2004, 11, 255‐263.
[28]. M. Ashby, H. Shercliff, D. Cebon, Materials: Engineering, Science, Processing and Design, Butterworth‐Heinemann, Amsterdam, the Netherlands, 2007.
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
108
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”,
Synthesis and Properties of Reactive Interfacial Agents for Polycaprolactone‐Starch Blends
109
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. Materials and Methods 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.
Chapter 6
110
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 1H‐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
Synthesis and Properties of Reactive Interfacial Agents for Polycaprolactone‐Starch Blends
111
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:
100% (mol) backbone PCL theof units repeating ofnumber
(mol) PCL toattached GMA/DEM of moles ofnumber FD ×= (6.1.)
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.
Chapter 6
112
6.3. Results and Discussions 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.).
*O
*O
n
*O
*O
n
COOEt
COOEt
*O
*O
n
OOO
DEM
BPO
GMA
BPO
PCL-g-DEM
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.
Synthesis and Properties of Reactive Interfacial Agents for Polycaprolactone‐Starch Blends
113
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).
Chapter 6
114
Table 6.1. Overview of experiments for the PCL‐g‐GMA compatibilisers a
Intake (%‐mol) b Sample
GMA BPO
FD (%)
T cryst. (oC)
ΔH cryst. (J/ g‐PCL)
T melt. (oC)
ΔH melt. (J/ g‐PCL)
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 (%‐mol) b Sample
DEM BPO
FD (%)
T cryst. (oC)
ΔH cryst. (J/ g‐PCL)
T melt. (oC)
ΔH melt. (J/ g‐PCL)
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
Synthesis and Properties of Reactive Interfacial Agents for Polycaprolactone‐Starch Blends
115
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.
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
+ R O hydrogenabstractionn n
+
Termination
+ H or
+ R or
+ RO
+ k
n
k+1
O
O
CCH2(CH2)4 O
O
CCH-
(CH2)4
GMA
nO
O
CCH(CH2)4
CH2
CH3 C-
OCH2CHCH2O
O
C
GMA
O
O
CCH(CH2)4
CH2
CH3 C
O
CH2CHCH2O
O
C
Chapter 6
116
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.
10 20 30 40 50 600
5
10
15
20
25
30
Substrate Intake (mol% on CL units)
FD (%
)
GMA, BPO=0.6% mol/mol-CL unitsGMA, BPO=0.3% mol/mol-CL unitsDEM, BPO=0.6% mol/mol-CL unitsDEM, BPO=1.1% mol/mol-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.
Synthesis and Properties of Reactive Interfacial Agents for Polycaprolactone‐Starch Blends
117
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.).
0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.20
5
10
15
20
25
30
35
40
45
50
BPO Intake (mol% on CL units)
FD (%
)
GMA, GMA=24% mol/mol-CL unitsGMA, GMA=12% mol/mol-CL unitsDEM, DEM=30% mol/mol-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
Chapter 6
118
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:
( )δ,, im nnfFD = (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:
δδδ imimim nnnnnnFD 2729.06022.15431.00325.08875.1 −+++−= (6.3.)
Synthesis and Properties of Reactive Interfacial Agents for Polycaprolactone‐Starch Blends
119
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 ( imnn ), the interaction between (PCL‐monomer) intake and solubility parameter difference ( δmn ) and the interaction between the latter factor and the initiator intake ( δin ).
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.
Figure 6.5. Graphical representation of the FD model for PCL‐g‐GMA. a. 3D plot. b. Contour plot.
1020
3040
50
0.20.4
0.6
0.810
20
40
60
80
GMA Intake (mol% on CL units)
BPO Intake (mol% on CL units)
FD (%
)
10
10
10
20
20
20
20
30
30
30
40
40
40
50
50
60 70
GMA Intakes (mol% on CL units)
BP
O In
take
(mol
% o
n C
L un
its)
10 15 20 25 30 35 40 45 500.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
FD (%)
a. b.
Chapter 6
120
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
1020
3040
50
0.20.4
0.6
0.810
2
4
6
8
DEM Intake (mol% on CL units)
BPO Intake (mol% on CL units)
FD (%
)
1
2
2
3
3
3
44
4
55
56
6
6
7
7
DEM Intake (mol% on CL units)
BP
O In
take
(mol
% o
n C
L un
its)
10 15 20 25 30 35 40 45 500.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
FD (%)
a. b.
Synthesis and Properties of Reactive Interfacial Agents for Polycaprolactone‐Starch Blends
121
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
Sample σ (MPa)
ε (%)
E (MPa)
Tc (oC)
∆Hc (J/gPCL)
Tm (oC)
∆Hm (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.
Chapter 6
122
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.).
0 5 10 15 20 25 3035
40
45
50
55
60
65
70
75
Starch content (%)
∆ H
(J/
g-P
CL)
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
Synthesis and Properties of Reactive Interfacial Agents for Polycaprolactone‐Starch Blends
123
matrix. This is visually confirmed by the morphology of the prepared blends (Figure 6.8.).
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.).
a. b. c.
Chapter 6
124
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.
a. b. c.
Synthesis and Properties of Reactive Interfacial Agents for Polycaprolactone‐Starch Blends
125
0 1 2 3 4 5
320
330
340
350
360
370
380
390
PCL-g-DEM amount (wt%)
E (
MP
a)
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
Chapter 6
126
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.).
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.
a. b. c.
Synthesis and Properties of Reactive Interfacial Agents for Polycaprolactone‐Starch Blends
127
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
O O
OH
OH
PCL-g-GMA
Starch particle
OH
OH
PCL-g-DEM
Starch particle
Chapter 6
128
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
in : amount of initiator intake [mol% on CL units]
mn : amount of monomer GMA or DEM intake [mol% on CL units]
Synthesis and Properties of Reactive Interfacial Agents for Polycaprolactone‐Starch Blends
129
T : temperature [oC]
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.
[4]. R. Mani, M. Bhattacharya: Properties of injection moulded starch/ synthetic polymer blends – III. Effect of amylopectin to amylase ratio in starch, Eur. Polym. J. 1998, 34, 1467‐1475.
[5]. R. Mani, M. Bhattacharya: Properties of injection moulded starch/ synthetic 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.
Chapter 6
130
[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.
Synthesis and Properties of Reactive Interfacial Agents for Polycaprolactone‐Starch Blends
131
[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.
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
134
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
Summary
135
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‐
Summary
136
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.
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
138
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(OiPr)3 concentratie). Een hogere verhouding leidde tot hogere molecuul 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 1H‐ 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.
Samenvatting
139
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
Samenvatting
140
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).
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
142
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.
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. Heeres: Synthesis of higher fatty acid starch esters using vinyl laurate and 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).