Synthesis and Properties of Branched Semi-Crystalline Thermoset Resins

S YNTHESIS AND P ROPERTIES OF B RANCHED

S EMI - CRYSTALLINE T HERMOSET R ESINS

Hans Claesson

Royal Institute of Technology Fibre and Polymer Technology

Stockholm 2003

Akademisk avhandling

Som med tillstånd av Kungliga Tekniska Högskolan i Stockholm framlägges till offentlig granskning för avläggande av teknisk doktorsexamen fredagen den 5 september 2003, kl.

13.00 i kollegiesalen, administrationsbyggnaden, Valhallavägen 79, Kungliga Tekniska

Högskolan, Stockholm. Avhandling försvaras på engelska.

To my Wife

Laura

Abstract

This thesis describes the synthesis and characterization of branched semi-crystalline polymers. Included in this work is the SEC characterization of a series of dendrimers.

The branched semi-crystalline polymers were synthesized in order to investigate the concept of their use as powder coatings resins. This concept being that the use of branched semi-crystalline polymers in a UV-cured powder coating system may offer a lower temperature alternative thus allowing the use of heat sensitive substrates and the added benefit of a reduced viscosity compared to linear polymers.

A series of branched poly(ε-caprolactone)’s (PCL) (degree of polymerization: 5-200) initiated from hydroxyl functional initiators were synthesized. The final architectures were controlled by the choice of initiator structure; specifically the dendritic initiators yielded star- branched PCL’s while the linear initiator yielded comb-branched PCL’s. The dendritic initiators utilized were: (1) a 3

-generation Boltorn

H-30, commercially available hyperbranched polyester with approximately 32 hydroxyl groups, (2) a 3

-generation dendrimer with 24 hydroxyl groups, and (3) a 3

-generation dendron with 8 hydroxyl groups.

Linear PCL was synthesized for comparison. All dendritic initiators are based on 2,2- bis(methylol) propionic acid. The comb-branched polymers were initiated from a modified peroxide functional polyacrylate. The resins were end-capped with methylmethacrylate in order to produce a cross-linkable system. The polymers and films were characterized using

H NMR,

C NMR, SEC, DMTA, DSC, FT-IR, FT-Raman, rheometry and a rheometer coupled to a UV-lamp to measure cure behavior.

The star-branched PCL’s exhibited considerably lower viscosities than their linear counterparts with the same molecular weight for the molecular region investigated (2-550 kg mol

). It was also found that the zero shear viscosity increased roughly exponentially with M.

The PCL star-branched resins are semi-crystalline and their melting points (T

) range from 34-50°C; films can be formed and cured below 80°C. The viscoelastic behaviour during the cure showed that the time to reach the gel point, a few seconds, increased linearly with molecular weight. The crossover of G’ and G’’ was used as the gel point. Measurement of mechanical properties of films showed that the low molecular weight polymers were amorphous while those with high molecular weight were crystalline after cure.

The polymerization of 5,5-dimethyl-1,3-dioxane-2-one (NPC) from oligo- and multifunctional initiators was evaluated utilizing coordination and cationic polymerization.

Two tin based catalysts, stannous(II) 2-ethylhexanoate and stannous(II) trifluoromethane sulfonate, were compared with fumaric acid. Fumaric acid under bulk conditions resulted in lower polydispersity and less chance of gelling. The synthesis of star-branched polymers was confirmed by SEC data. The star polymers exhibited a T

at 20-30°C and a T

at about 100°C.

All semi-crystalline resins exhibited a fast decrease in viscosity at T

. Blends of comb- branched semi-crystalline resins and amorphous resins exhibited a transition behavior in- between that of pure semi-crystalline resins and that of amorphous resins.

The SEC characterization of a series of dendrimers with different cores and terminal groups showed that the core had an impact on the viscosimetric radius of the core while the terminal groups appeared to have no effect.

Keywords: star-branched, semi-crystalline, comb-branched, ring-opening polymerization,

poly(ε-caprolactone), dendritic, thermoset, low temperature curing, powder coating, UV-

curing, poly(5,5-dimethyl-1,3-dioxane-2-one), size exclusion chromatography, rheology,

dendritic aliphatic polyester

Sammanfattning

Avhandlingen beskriver syntesen och karakteriseringen av en serie förgrenade delkristallina polymerer. Inkluderat i detta arbete är också SEC karakteriseringen av en serie av dendrimerer.

De förgrenade delkristallina polymererna syntetiserades och karakteriserades för att undersöka ett koncept för deras användning som pulverbindemedel. Konceptet är att en delkristallin struktur tillsammans med UV-härdning kan resultera i en lägre härdtemperatur och på så vis möjliggöra användningen av pulverfärger på värmekänsliga substrat. En grenad struktur kan också ge en lägre viskositet.

En serie av förgrenade poly(ε-kaprolakton)er (PKL) (polymerisationsgrad: 5-200) initierad från hydroxyfunktionella initiatorer syntetiserades. Den slutliga arkitekturen styrdes av valet av initiator. De dendritiska initiatorerna resulterade i stjärnförgrenad PKL och den linjära initiatorn resulterade i kamförgrenad PKL. De dendritiska initiatorerna som användes var: (1) en 3

-generations Boltorn

H-30, kommersiellt tillgänglig hyperförgrenad polyester med cirka 32 hydroxylgrupper, (2) en 3

generations dendrimer med 24 hydroxylgrupper, (3) en 3rd-generations dendron med 8 hydroxylgrupper. Linjär PKL syntetiserades för jämförelse.

Alla dendritiska initiatorer är baserade på 2,2-bis(metylol) propionat syra. Den kamförgrenade polymeren initierades från en modifierad epoxid funktionell polyakrylat. Polymererna funktionaliserades med metylmetakrylat för att ge en tvärbindningsbar polymer. Polymererna och filmerna karakteriserades med

H NMR,

C NMR, SEC, DMTA, DSC, FT-IR, FT- Raman, reometer, och en reometer utrustad med en UV-lampa för att mäta härdbeteende.

De stjärnförgrenade polymererna uppvisade en betydligt lägre viskositet än de linjära polymererna med samma molekylvikt (M) (i området 2-550 kg mol

). Det observerades också att gräns skjuvviskositeten (η

) ökade exponentiellt med M, vilket var förväntat.

PKL stjärnpolymererna är delkristallina och smältpunkten (T

) är 34-50°C; filmer kan framställas och härdas under 80°C. Det viskoelastiska beteendet under härdning av de stjärnförgrenade polymererna visade att tiden till att nå gelpunkten, några sekunder, ökade linjärt med molekylvikten hos polymeren. G’=G’’ användes for att bestämda gelpunkten.

Mätningen av de mekaniska egenskaperna hos härdade filmerna av de stjärnförgrenade polymererna visade att de filmer tillverkad från polymerer med låg molekylvikt var amorfa medan de med hög molekylvikt var delkristallina.

Polymerisationen av 5,5-dimetyl-1,3-dioxan-2-one (NPK) från multifunktionella initiatorer utvärderades med koordinations- och katjonpolymerisation. Två tennbaserade katalysatorer och fumarsyra jämfördes. Fumarsyra in bulkreaktion resulterade i lägst polydispersitet och mindre risk för gelning jämfört med de tennbaserade katalysatorerna. Syntesen av stjärnförgrenade polymerer bekräftades av SEC-analys. Stjärnpolymererna hade ett T

på 20- 30°C och ett T

på ca. 100°C.

Alla delkristallina polymerer uppvisade en snabb minskning av viskositeten vid T

.

Blandningar av delkristallina kampolymerer och amorfa polymerer uppvisade ett

List of Papers

This thesis is a summary of the following papers:

I “Synthesis and Characterisation of Star Branched Polyesters with Dendritic Cores and the Effect of Structural Variations on Zero Shear Rate Viscosity” H.

Claesson, E. Malmström, M. Johansson & A. Hult Polymer (2002), 43(12), 3511-3518.

II “Semi-Crystalline Thermoset Resins: Tailoring Rheological Properties in Melt using Comb Structures with Crystalline Grafts” H. Claesson, C. Scheurer, E.

Malmström, M. Johansson, A. Hult, W. Paulus & R. Schwalm Submitted to:

Progress in Organic Coatings.

III “Star-Branched Poly(neopentyl carbonate)s” P. Löwenhielm, H. Claesson, E.

Malmström & A. Hult Manuscript.

IV “Rheological Behaviour during UV-curing of a Star-Branched Polyester” H.

Claesson, M. Doyle, E. Malmström, M. Johansson, J-A. E. Månsson & A. Hult.

Progress in Organic Coatings (2002), 44(1), 63-67.

V “Synthesis and Characterization of Bis-MPA Dendrimers with Different Core and Terminal Groups” M. Malkoch, H. Claesson, P. Löwenhielm, E. Malmström

& A. Hult Manuscript.

Table of Contents

Abstract………..…….. I Sammanfattning………...……… II List of Papers………....… III

1. Purpose of the Study………...………….……… 1

2. Background………..……… 2

2.1 Powder Coatings...………..….……… 2

2.1.1 UV-Cured Powder Coatings……….………. 4

2.2 Branched Polymers...………...……… 6

2.2.1 Dendritic Polymers……… 6

2.2.1.1 Dendrimers……...………... 6

2.2.1.2. Hyperbranched Polymers..……..………. 8

2.2.2 Star-Branched Polymers……….………... 8

2.2.3 Comb-Branched Polymers………. 9

2.3 Ring-Opening Polymerization……….……… 9

2.3.1 Coordination Insertion Ring-Opening Polymerization……….. 10

2.3.2 Cationic Ring-Opening Polymerization……….... 10

2.4 Rheology in the Molten State……….. 10

2.4.1 Rheological Behavior of Linear Polymers……… 10

2.4.2 Rheological Behavior of Branched Polymers………...…… 11

3. Syntheses and Chemical Characterization………..…. 13

3.1 Monomers……… 13

3.2 Star-Branched Poly(ε-caprolactone)’s………. 13

3.2.1 Synthesis……… 14

3.2.2 NMR Characterization………..…. 15

3.2.3 SEC Characterization………..……….. 16

3.3 Star-Branched Poly(neopentyl carbonate)’s………..………....…….. 18

3.3.1 Synthesis………..….. 19

3.3.1.1 Monomer Synthesis……….. 19

3.3.1.2 Polymer Synthesis………. 19

3.3.2 NMR Characterization………...…… 20

3.3.3 SEC Characterization………..….. 22

3.3.4 Catalyst Evaluation……… 23

3.3.5 Thermal Characterization……….……. 25

3.4 Comb-Branched Poly(ε-caprolactone)’s………...…….. 26

3.4.1 Synthesis……….... 26

3.4.2 IR Characterization………..……….. 27

3.4.3 SEC Characterization………...………. 28

3.5 Bis-MPA Dendrimers……….. 29

3.5.1 SEC Characterization………...………. 30

4. Rheological Characterization………..…. 35

4.1 Zero Shear Viscosity of Star-Branched Poly(ε-caprolactone)’s…………. 35

4.2 UV-curing Rheological Behavior of Star-Branched Poly(ε-caprolactone). 36 4.3 Dynamic Viscosity from Solid to Molten State………..…. 42

4.3.1 Comb Poly(ε-caprolactone) and Blends……….………. 43

5. Film Characterization………... 45

5.1 Mechanical Properties of Star-Branched Poly(ε-caprolactone) Films…… 45

5.2 Comb Poly(ε-caprolactone) Films………... 46

5.2.1 Powder and Film Preparation……… 46

5.5.2 Film Properties……….……. 47

6. Conclusions……….……. 49

7. Suggestions of Further Work………...………… 51

Acknowledgements……….………. 52

References……….……….….. 54 Appendix A – Structures of the Different Star-Branched Polymers

Appendix B – SEC

and Viscosity Data of all Star-Branched PCL’s

Appendix C – Synthetic Scheme of Boltorn-PNPC

1. Purpose of the Study

Thermoset resins are very important in industry were high demands are set on the final properties; applications such as adhesives, molding compounds and coatings. One area were thermoset resins have a significant share of the market in comparison to thermoplastic resins is in the area of powder coatings. Since the introduction of powder coatings, the industry has been striving to improve coating technology to widen its application to new markets such as wood coatings. UV-cured powder coating systems were recently introduced as an alternative to conventional thermal curing in an effort to capitalize on this new market. The goal of this body of work was to investigate a new powder coating concept, which would result in a reduction of the curing temperature. The approach was to investigate how changes in macromolecular architecture, molecular composition, molecular weight and introduction of crystallinity affect the properties relevant to powder coating applications, which include rheological, cure, and final film properties. In addition, the effects of molecular weight and architecture on resin crystallinity was also a subject of interest as was the relationship between resin structure and resin properties before, during and after curing. To obtain better overall knowledge of the relationship between architecture, functionality and properties in a cross- field approach, this included dendritic, star and comb polymers.

The specific purpose of this work was to:

Investigate the effect of star-branching on zero shear viscosity.

Investigate the UV-curing performance of star-branched polymers.

Synthesize comb-branched polymers and evaluate the film properties of the pure resin and blends with conventional resins.

Develop a polymer better suited for a low temperature curing thermoset resin, i.e. with a T

and T

proving storage stability and good film properties.

Evaluate a series of dendrimers with different cores and terminal groups utilizing a

triple detection SEC.

2. Background

This chapter briefly reviews the areas of interest covered by this thesis. First the basic concepts, applications, and problems associated with powder coatings are introduced. This is followed by a presentation of the different polymerization techniques utilized to synthesize branched polymers. Finally, there is a review of the various macromolecular architectures and their rheological behavior.

2.1 Powder Coatings

Powder coatings in their powder form can be either thermoplastic or thermosetting. The first powder coatings, developed in the 1950’s, were thermoplastic and were applied to preheated metal substrates.

Thermoplastic powder coatings, by definition, are melted to form a film at elevated temperature and solidify upon cooling. Film formation is a result of the melting and coalescence of powder particles. In order for thermoplastic powder coatings to achieve good mechanical properties, high molecular weight of the resin is required.

Coalescence and leveling are mainly surface tension driven and the level of viscosity the main property affecting the rate of film formation.

The manufacture of thermoplastic powder coatings is relatively simple and raw materials are normally commodity polymers with overall acceptable properties such as polyvinyl chloride, polyolefines, nylons and polyesters.

The main disadvantages with these coatings are high fusion temperature, low pigmentation levels, and poor adhesion to metal substrates. In spite of these general shortcomings, some of them display outstanding properties such as solvent resistance (polyolefines), outstanding weathering resistance (polyvinylidene chloride) and exceptional abrasion resistance (nylon).

Low price and ease of handling are other advantages.

A few years after the development of thermoplastic powder coatings, Shell Chemicals developed the first thermosetting powder coatings. During film formation of traditional thermosetting powder coatings, a thermally activated reaction takes place, resulting in the formation of a cross-linked polymer. Their introduction solved many of the problems associated with thermoplastic powder coatings. In the late 1960’s and early 1970’s when new laws and regulations in industrialized countries gave powder coating technology a “bump”

forward.

The main components of thermosetting polymers are a primary resin and a cross-linker.

Film formation is traditionally accompanied by chemical cross-linking (curing) that commences when the system is heated. Curing affects both the viscosity and the flow. Caution must be taken so that curing does not restrict film formation and leveling, i.e. curing must start well above the glass transition temperature (T

) of the resin. During curing a three- dimensional network is formed by the low molecular weight resin and the cross-linker.

Network formation is dependent on the average degree of functionalization of the resin/cross-

stability against coalescence during storage; (3) coalescence, degassing, and leveling, i.e. film formation, at the lowest temperature possible; and (4) cross-linking at the lowest temperature possible.

Coalescence of the powder during storage can be avoided if the resin has a high T

. On the other hand, a low T

allows coalescence and leveling at a lower temperature since the rule of thumb is that the lowest feasible curing temperature is 70-80°C above the T

of the powder.

Viscosity controls the flow and leveling of a powder coating. A low viscosity promotes leveling while a high viscosity impedes leveling. The main driving force for leveling is the surface tension, which is similar for most resins. Of the remaining parameters affecting leveling, such as mean film thickness, and particle size, shape and distribution, viscosity is the only parameter controlled by molecular architecture, weight, and chemical composition.

During leveling the shear rate is very low.

Thus, the zero shear viscosity critically influences leveling.

Over the years the terminology of the different systems has grown and become confused.

For clarity, table 2.1 shows some of the major classes of powder coatings along with some details about each.

Table 2.1 Overview of the most common powder coating systems.

Common

name Primary resin Cross-linker Curing temp. (°C)

Epoxy Bis-phenol A epoxy Polyamines, anhydrides or phenolics 180

Hybrid COOH-functional

polyester Bis-phenol A epoxy 160-200

COOH-functional

polyester

Triglycidylisocyanurate or hydroxyalkylamin Polyester

OH-functional

polyester Blocked-isocyanate or amino acid

180-200

Epoxy-functional

acrylic Dibasic acid Acrylic

OH-functional acrylic Blocked-isocyanate or amino resin

130-180

Acrylate-functional

resin Free radical 120

UV*-cure

Epoxy-functional

resin Cationic 120

* Ultra Violet

The main advantages associated with powder coatings compared to solvent- or water-borne

coatings include: near zero volatile organic emission, high application speed, reduced energy

consumption, easy clean-up, recycling of over spray (>95% utilization), durable finishes,

possibility to apply thick films, and electrostatic application of 3D substrates. The powder

coating system also requires less skill and training to operate, substantially reduced

flammability and low toxicity.

These advantages have led to continuous growth of the

powder coating market during the last few decades. However, powder coatings do have some limitations, namely: increased risk of dust explosions, inability to coat large or heat sensitive substrates, some appearance limitations, major components must be solid which results in material limitations, low production and application flexibility due to the clean-up needed between color changes.

In addition, the application techniques of thermoplastic and thermoset powder coatings are commonly fluidized bed and electrostatic spray, respectively, thus limiting their use to industrial settings.

2.1.1 UV-Cured Powder Coatings

Traditionally, the powder coating cross-linking reaction has been controlled by temperature. Normal curing temperatures are 160-200°C, and are thus not suitable for application on heat sensitive substrates such as wood and plastic. Curing temperature is determined by the combination of storage stability and the film formation process. Powder coatings must be storable at 30°C without the resin particles fusing, thus the glass transition temperature, T

, of the polymer must be 60°C or higher. Film formation for amorphous resins requires a temperature at least 50°C above T

, giving a minimum curing temperature of about 110°C.

One way to lower the curing temperature is by the introduction of semi-crystalline material with a suitable melting temperature.

The advantage of using a crystalline resin is the rapid melting, versus the slow softening of an amorphous resin (figure 2.1).

Leveling Onset

of Cure

Heat

UV Cured System UV-curing

Leveling/

Curing Leveling/

Curing

Thermally Cured System

the onset of cure to be controlled by the use of ultra violet (UV) initiation, which may result in smoother coatings when desired (figure 2.2).

In recent years UV-curing of powder coatings has obtained increased attention in industrial research

as this technique allows fast curing at lower temperatures than conventional powder coatings.

The research was triggered by the possibility of coating heat sensitive substrates. Other advantages include shorter cycle times, improved storage stability, no premature reaction during manufacturing, and better leveling since viscosity does not increase until UV-irradiated.

Film formation can be performed at low temperatures (90-140°C) and cured in a matter of seconds with UV.

Although UV-curing offers many advantages, it does have limitations such as working film thickness, interferences due to UV absorbance by pigments, and difficulty in curing complex shapes.

Though pigmented coatings can be cured, the clear coatings are the most interesting. Powder coatings with bis phenol A epoxy as the binder are cationically UV cured. Acrylated epoxy resins, with or without acrylated polyesters or unsaturated maleic resins as binders, are cured via a free radical mechanism.

Crystalline resin UV-cured

Temperature/Time

Log(V iscosity )

Onset of UV-cure

Amorphous resin- thermally cured Crystalline resin

UV-cured

Temperature/Time

Log(V iscosity )

Temperature/Time

Log(V iscosity )

Onset of UV-cure Onset of UV-cure

Amorphous resin- thermally cured

Figure 2.2 A schematic diagram of viscosity as a function of temperature/time for a thermally

cured amorphous resin and an UV-cured crystalline resin. The onset of cure can be

controlled when using an UV-curing system while the thermally cured system starts to cure

almost immediately.

2.2 Branched Polymers

There are a many different types of branched polymers. Dendritic (extremely branched), pom-pom, H-, comb- and star-branched polymers are just some examples. Due to the complex architecture of branched polymers, their properties differ from their linear counterparts. The rheological properties of branched polymers have been studied extensively, especially the star polymers where they have served as models to increase the general understanding of branched macromolecules.

2.2.1 Dendritic Polymers

Dendritic polymers, comprised of dendrimers and hyperbranched polymers (figure 2.3), are synthesized from AB

monomers. An AB

monomer consists of two different functional groups, A and B, where there are two or more B’s for every A (figure 2.4). The final structure is dependent upon the growth process. A controlled growth yields a dendrimer while an uncontrolled growth yields a hyperbranched polymer.

Dendrimer Hyperbranched polymer

Figure 2.3 Schematic representations of a dendrimer and a hyperbranched polymer.

2.2.1.1 Dendrimers

The term “dendrimer” was first coined by Tomalia et al. to describe a large family of

regularly branched poly(amidoamines).

Dendrimers contain bonds that converge to a single

point with each repeating unit containing a branch junction and with the final molecule

featuring a very large number of identical chain ends (figure 2.3). The interest in dendrimers

molecules to a single building block. In a final reaction, the completed dendrons or arms are attached to a multi-functional core. While it has been shown that the two approaches can give exactly the same structure, the growth process are essentially opposite. One disadvantage of the divergent approach is that it generates a large number of functional end groups which may not all react to form the next completely substituted layer (generation). This makes separation of fully functionalized and almost fully functionalized species impossible. On the other hand, the convergent growth approach offers better control over each step of the synthesis. With each reaction, fewer functional groups are involved and purification is simplified due to the large differences in molar mass between product and by-products. Multiple synthetic steps are required to produce high molecular weight polymers with either approach making the synthesis of dendrimers tedious and expensive.

However, recently Fréchet et al. developed a method for the divergent synthesis of aliphatic polyester (bis-MPA) dendrimers utilizing an anhydride building block.

This approach proved to be highly efficient and circumvented the previous time-consuming purification problems.

2.2.1.2 Hyperbranched Polymers

The main features distinguishing hyperbranched macromolecules from dendrimers are the ability to synthesize these structures in one step and that they contain linear units. The linear units produce characteristics such as broad molecular weight distribution and irregular branching (figure 2.4).

Hyperbranched macromolecules, first reported by Kim and Webster

, have been studied in detail since 1989. However, Flory discussed the fundamental concepts underlying their synthesis more than 40 years ago.

Flory predicted that AB

monomers with one reactive group of type A and x reactive groups of type B would polymerize readily and give a soluble, easy to process (low viscosity), three-dimensional structure free of cross-links.

Figure 2.4 Schematic of the synthesis, structure and different repeating units of a hyperbranched polymer.

Hyperbranched macromolecules contain three types of repeating unit dependent on degree of substitution; dendritic, linear and terminal (figure 2.4). The dendritic unit is composed of fully substituted AB

monomers, the linear unit has one reacted and one unreacted B group, and the terminal unit has two unreacted B groups. The degree of branching (DB), used to characterize hyperbranched polymers, defined by Fréchet et al

, follows:

Core molecule

B

B

B B

B

B B

B B

Terminal unit

B

Linear unit Dendritic unit

B

B

B B B B

B

B B

B

B B

B

A B B B B

B

A B

+

Core molecule

B

∑ ∑ ⁺ ∑

= All repeating units

units Terminal units

Dendritic DB

Dendrimers have a DB of 1 since they contain only dendritic and terminal units. The comparison of two hyperbranched polymers with the same chemical composition but different DB has shown that solubility increases with the degree of branching while the melt viscosity is inversely related.

Hyperbranched polymers are usually prepared in a single-step polymerization and are thus not as tedious to synthesize as dendrimers. Even though theoretically one-step growth of a hyperbranched macromolecule could lead to a “perfect” dendrimer, it has never been encountered due to the kinetics of the polymerization techniques used. A living polymerization may lead to more dendrimer-like structures. To obtain a high molecular weight hyperbranched polymer, several conditions must be fulfilled e.g. the reactive groups, A and B, should only react with each other and side reactions should be kept to a minimum preventing deactivation and cross-linking.

Hult et al.

used p-toluene-sulfonic acid to catalyze the bulk synthesis of hyperbranched aliphatic polyester in the molten state. 2,2- Bis(methylol)propionic acid (bis-MPA) was used as the monomer and 2-ethyl-2- (hydroxymethyl)-1,3-propanediol as the core molecule. Voit et al. have performed additional work with the same monomer.

2.2.2 Star-Branched Polymers

Star polymers are one of the simplest forms of branched polymers. They consist of a core molecule onto which linear polymers are coupled or grafted from. Synthesis can be divided into three general approaches described in figure 2.5. The first route is the core-first method where polymer chains are grown directly from a multifunctional core (route A). The other two routes utilize the arm-first method, where preformed linear polymers are linked to a multifunctional coupling agent or a diene (routes B and C).

One of the most common approaches is anionic polymerization of monodisperse arms, then attachment to a chlorosilane functional core molecule, which can vary in degree of functionalization (route B), permitting good control of molecular weight and polydispersity.

Other synthetic approaches include atom transfer radical polymerization,

nitroxide- mediated polymerization,

ring-opening polymerization (ROP) utilizing coordination insertion,

ring-opening metathesis polymerization

and radical addition-fragmentation chain-transfer polymerization.

Some of the uses for star polymers include various coating applications such as antifouling

coatings, conductive coatings and low volatile organic content coatings.

Star polymers

are also used as colloidal stabilizers, additives to improve impact resistance and reduce

Living prepolymer

Monomer

+ +

R R Multifunctional

core moiety Multifunctional

initiator

Polymerization

Coupling reaction

Linking reaction

A

Living prepolymer

B

C

Diene

Figure 2.5 Schematic illustrations of the three approaches for the synthesis of a star- branched polymer.

2.2.3 Comb-Branched Polymers

Comb-branched polymers consist of a backbone with polymer side chains.

The backbone and side chains can either be of the same or different chemical composition (graft copolymers). Some of the first graft copolymers were acrylonitrile-butadiene-styrene copolymer (ABS), high impact polystyrene (HIPS) and non-ionic emulsifiers. In the case of ABS and HIPS, the copolymerization yielded a better result than merely physical blending due to the low compatibility between the components.

The synthetic approaches for comb polymers are the same as for star polymers.

2.3 Ring-Opening Polymerization

Ring-opening polymerization (ROP) can proceed through a number of different

mechanisms depending on the type of monomer and catalyst involved. Of the wide range of

polymers produced via ROP, some have gained industrial significance; for example

poly(caprolactam) and poly(ethylene oxide).

ROP of cyclic esters, such as ε-caprolactone

(CL) and L,L-dilactide, is gaining industrial interest due to their degradability.

The

mechanisms of interest of interest in this work are coordination insertion and cationic.

2.3.1 Coordination Insertion Ring-Opening Polymerization

Coordination insertion ring-opening polymerization is an effective route to obtain well- defined polyesters. Stannous(II) 2-ethylhexanoate (Sn(Oct)

) is a common catalyst for the ROP of lactones and lactides.

There are two main proposed mechanisms for the ROP of cyclic esters using Sn(Oct)

and a hydroxyl functional co-initiator; complex formation between the monomer and hydroxyl group prior to ROP and formation of a tin-alkoxide prior to initiation.

The main advantage of ROP of cyclic esters (and amides) versus a condensation reaction of the equivalent hydroxy carboxylic acid is absence of water formation.

2.3.2 Cationic Ring-Opening Polymerization

A broad range of heterocyclic compounds can undergo cationic ring-opening polymerization (CROP).

CROP propagates either through an activated monomer or an activated chain end mechanism. In both cases the propagation involves the formation of a positively charged species.

A major drawback of CROP is the occurrence of unwanted side reactions, thus limiting the molecular weight of the final product.

2.4 Rheology in the Molten State

Rheology is defined as the science of the deformation and flow of matter. In the molten state linear polymers often exhibit pronounced viscoelastic properties, such as shear thinning, extension thickening, viscoelastic normal stresses, and time-dependant rheology. All of these properties are due to the physical nature of most polymers, which is long and easily distorted.

Viscosity is a measure of a fluid’s resistance to flow and describes the internal friction of a moving fluid. Viscosity is an extremely important property of polymer melts and solutions since it controls the possibility of processing. If the viscosity is too high, processing will be difficult or impossible. In the case of powder coatings (and other types of coatings), viscosity controls leveling.

2.4.1 Rheological Behavior of Linear Polymers

The main factor affecting viscosity in the molten state is the molecular weight. Figure 2.6

shows the well-known plot of how zero shear viscosity (η

) relates to molecular weight (M).

η

depends on M as η

∝ M below the critical molecular weight (M

) and η

∝ M

above

M

.

Log(M) M

η

∝M

η

∝M

L og ( η

)

Figure 2.6 The relationship between zero shear viscosity ( η

) and molecular weight (M) for a linear polymer with low polydispersity.

The steep increase in viscosity above M

is due to entanglements. Entanglements restrict molecular motion by preventing chains from passing one another and moving perpendicular to their own molecular contour. In order to relax stress, individual polymer chains have to move along their own contour in a snake-like fashion. This snake-like motion is called reptation.

For linear polymers the onset of entanglement starts at their M

, which for highly flexible polymers ranges between 300-600 atoms in the main chain.

2.4.2 Rheological Behavior of Branched Polymers

Dendritic polymers are essentially free of entanglements and exhibit a spherical shape at higher generations due to sterical and branching considerations. This results in unique rheological properties, which is one of their most prominent features.

In contrast to chain- type macromolecules, which exhibit a non-Newtonian shear thinning behavior at a critical shear, dendritic macromolecules exhibit Newtonian flow behavior in solution and bulk regardless of the number of generations.

The rheological properties of star-branched polymers have also been extensively investigated. Due to their simplicity, they have served as models for the development of new or refinement of existing rheology theories.

The rheological behavior of star polymers differs from linear and dendritic polymers, since their relaxation modes are different.

Entangled star polymers cannot relax through reptation since one end is attached to a core moiety. The arms relax through primitive path fluctuations

and constraint release.

During primitive path fluctuations, some times called “breathing modes”, a polymer arm is drawn back and then re-extended into a new tube. Constraint release occurs when surrounding arms fluctuate. Tube widening by constraint release is analogous to the addition of a low molecular weight solvent and is therefore called “dynamic dilution”.

As previously mentioned, η

depends on M as η

∝ M below M

and η

∝ M

above M

.

However, for star polymers the viscosity increases exponentially with molecular weight

and

hence no M

is found to coincide with the onset of entanglement. Also, the η

of a star

polymer is not dependent on the total M, but on the arm M.

This makes it possible to

increase the molecular weight without changing the viscosity by increasing the number of arms.

Comb polymers are normally less defined than star polymers but studies have shown that

the η

of comb polystyrenes is lower than that of a linear polymer with the same molecular

weight.

In addition, Sherrington et al. prepared a large number of branched poly(ethylene

terephthalate)s (PET) by addition of a branching agent and a chain stopper. The branched

PET’s exhibited both lower solution and melt viscosities in spite of a significantly higher

molecular weight than the linear model polymer.

3. Syntheses and Chemical Characterization

This chapter covers the synthesis and chemical characterization of the different branched polymer structures, which appear in this work. The chemical characterizations include nuclear magnetic resonance (NMR), size exclusion chromatography (SEC), Fourier-transform infrared spectroscopy (FT-IR) and FT-Raman. Triple detection (SEC

) and universal calibration (SEC

) were utilized in the SEC characterization. Also covered in this chapter is the SEC characterization of a series of dendrimers.

3.1. Monomers

The monomers used as the starting material in the synthesis of the different architectures are shown in figure 3.1. 2,2-Bis(methylol)propionic acid (bis-MPA), a crystalline dihydroxy carboxylic acid, was used as the monomer in the synthesis of the dendron and dendrimer. It is also the monomer for the hyperbranched polymer Boltorn H-30 (commercially produced by Perstorp).

ε-Caprolactone (CL), a liquid cyclic ester, was used to synthesize the arms of both the star-branched and comb poly(ε-caprolactone)’s (PCL). 5,5-Dimethyl-1,3-dioxane-2-one, also known as neopentyl carbonate (NPC), was the crystalline cyclic carbonate used in the synthesis of the star-branched poly(neopentyl carbonate) (PNPC). The hydroxyl functional initiators used in the synthesis of the PNPC stars and the comb polymer backbone are not included in figure 3.1, however the structures of the initiators are described in section 3.3.1.2 and the composition of the backbone is described in section 3.4.

O O

O O

O

O H

O OH

OH

Bis-MPA ε-Caprolactone Neopentylene carbonate

Figure 3.1 Monomers used in the synthesis of the various macromolecular architectures.

3.2 Star-Branched Poly(ε-caprolactone)’s

Variations were made in the architectures of a group of star-branched PCL’s in order to

investigate the effect of branching on zero shear viscosity. ε-Caprolactone was chosen as the

monomer due to its ease of polymerization and to the semi-crystallinity of the formed

polymer.

3.2.1 Synthesis

The synthesis of star polymers can be divided into three parts: synthesis of the core molecules, grafting, and end capping. The core molecules include the third generation bis- MPA dendron and dendrimer which were synthesized according to a procedure by Hult et al.

The final core moiety was a pseudo third-generation, bis-MPA based, hyperbranched molecule, similar to a polymer developed by Hult et al.,

now commercially available under the trade name Boltorn

and kindly supplied by Perstorp AB.

Grafts were accomplished by ring-opening polymerization (ROP) of CL onto hydroxy- functional cores. The reactions were performed in bulk at 110°C with Sn(Oct)

as the catalyst with the degree of polymerization (DP) controlled by the monomer to initiator feed ratio.

Finally, the hydroxy-functional end groups of poly(ε-caprolactone) (PCL) were end-capped with methacrylate groups (scheme 3.1).

O O

O

O n OH O

O

O O

O H

n

O

O

O n O

O O

O n O

O O O

n OH O

O

O O

O H

n

O

O

O

O

OH

n

O O

O

O O H

n O

O

O

O

O H

n O

O O

O O

H

n O

H O H

O

H OH

OH OH OH OH OH O H

O

O Sn(Oct)₂

110°C

+

traces of water would initiate homo-polymerization. The structures of the different star polymers are presented in Appendix A.

3.2.2 NMR Characterization

The

H NMR spectra of the polymers were used to calculate the DP of the PCL grafts and the molecular weight (table 3.1). The DP was calculated by comparing the integrals of the protons on the methylene adjacent to the hydroxyl end-group (A) relative to those on the methylene next to the carbonyl carbon (C) (figure 3.2). Although it is also possible to perform the calculation by comparison with the methylene next to the oxygen in the repeat unit (B), this method of calculation is inferior due to the overlapping shift of the core molecules. The accuracy of the DP value obtained for polymers with high DP is reduced due to the reduced relative size of the peak from the protons next to the hydroxyl group.

32.5 1.0 34.6

Integral

(ppm)

0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5

32.5 1.0

Integral

(ppm)

3.5 3.6 3.7 3.8 3.9 4.0 4.1

O OH

O

O

R n

A C B

C

Figure 3.2

H NMR spectrum of a dendron-PCL with an average DP of about 35.

It has been shown that the

C NMR shift of the quaternary carbon in the repeat unit of the initiating species is sensitive to the degree of substitution.

This quaternary carbon resonates at 50.6 ppm if both hydroxyl groups are unreacted, at 48.8 ppm if one hydroxyl group remains, and at 46.8 ppm if both hydroxyl groups are reacted. This resonance change is useful in assessing the success of the grafting reaction.

C NMR spectra of the polymers showed that those with shorter arms still contained unreacted hydroxyl groups (figure 3.3).

A B

C

(ppm)

44 45 46 47 48 49 50 51

Figure 3.3

C NMR spectrum of the quaternary carbons from the bis-MPA repeating unit in the hyperbranched core of Boltorn-PCL, DP8. The peak at 46.8 ppm shows the presence of fully functionalized bis-MPA hydroxyl groups. The peak at 48.8 ppm shows the presence of partially functionalized bis-MPA end groups. The absence of a peak at 50.6 ppm shows that there are no end groups with both hydroxyl groups unreacted. The integral calculation of functionalization shows that approximately four hydroxyl groups remain unreacted. (The peaks are shifted downfield 0.8 ppm due to an uncalibrated spectrum.)

Full functionalization was observed at varying DP for the various structures analyzed (figure 3.4). The dendron and dendrimer were fully reacted at a DP of about 13-15 while the hyperbranched polymer was fully substituted at a DP of about 20. The spectra further showed that in all cases only one of the hydroxyl groups, 48.8 ppm, was unreacted. This lack of functionalization was probably due to the statistical nature of the reaction since the star polymers with longer arms showed no trace of incomplete functionalization of end-groups.

(ppm)

44 45 46 47 48 49 50 51

Figure 3.4

C NMR spectrum of Boltorn-PCL, DP of 32. Only one peak at 46.8 ppm

indicates that all hydroxyl groups have reacted. (The peak is shifted downfield 0.8 ppm due to

an uncalibrated spectrum.)

SEC

is based on the fact that V

∝ [η]M ,where is the hydrodynamic volume and [η] is the intrinsic viscosity, for a wide range of polymers and complex architectures including e.g.

dendritic-, H-, star-, comb- and co-polymers.

This makes SEC

a useful tool for the determination of molecular weight since there is no influence from the architecture or chemical composition of the analyzed compounds.

The molecular weights, especially the weight average molecular weight (M

), obtained from SEC

, are in general, very close to those from SEC

, which is to be expected since both methods are independent of architecture and chemical composition. However, light scattering (SEC

method) is dependent on the refractive index increment, which is low for PCL in tetrahydrofuran (the mobile phase), resulting in poor signal to noise ratio, especially for the low molecular weight fraction. The agreement of the SEC

data with the number average molecular weight (M

) from

H NMR data is, in general, good at low DP. This agreement is reduced at high DP since accurate end-group analysis becomes more difficult with increasing DP (table 3.1). In addition, the SEC determination of M

is very sensitive to errors and difficult o determine exactly. Polydispersity generally increased with increasing molecular weight. This is due to intra- and intermolecular transesterification that occurs at high conversion and molecular weight.

Table 3.1 SEC and

H NMR data of the star-branched PCL characterized with SEC

and SEC

. Mark-Houwink (MH) equation: [ η ] = kM

. The MH α value is a polymer conformation parameter. The α value decreases with the compactness of the structure.

(degree of polymerization DP, polydispersity index PDI)

DP

DP

M

1H NMR

M

SEC³

PDI

SEC³

M

SEC^UC

PDI

MH α value

Branching average Boltorn-PCL

50 51 189 900 239 000 1.58 236 300 9.13 0.263 39 70 79 292 200 303 800 1.54 308 700 7.86 0.204 46 Dendrimer-PCL

12 14 41 100 41 300 1.44 42 000 1.04 0.771 18 15 14 41 100 42 300 1.39 43 700 1.03 0.775 19 20 24 68 500 67 300 1.31 72 100 1.01 0.742 24 35 42 117 700 99 900 1.60 104 700 1.14 0.758 22 60 51 142 300 109 200 1.47 122 300 1.30 0.787 23 Dendron-PCL

18 15 14 600 15 800 1.27 16 000 1.01 0.822 6.8 40 46 42 900 47 800 1.39 43 000 1.27 0.836 8.2 80 81 74 900 68 000 1.30 66 400 1.71 0.817 8.5 Linear PCL

15 17 2 200 3 000 1.23 2 300 1.01 0.888 -

50 45 5 400 6 400 1.24 5 200 1.03 0.860 -

80 82 9 600 13 300 1.53 10 300 1.46 0.885 -

A 17 1 900 2 800 1.20 2 200 1.02 0.865 -

B 39 4 400 6 600 1.53 5 100 1.46 0.923 -

C 117 13 300 16 900 1.38 14 200 1.41 0.908 -

a

g mol

The polydispersity index (PDI) values obtained from SEC

are similar for all samples whereas there is a broad distribution/range of PDI’s obtained from SEC

. This can be attributed to the insensitivity of the RALLS detector (SEC

) to low molecular weight polymers with a low refractive index increment, resulting in too narrow distribution of the disperse polymers (Boltorn-PCL). In addition, when considering the multifunctional initiator moieties, Boltorn a polydisperse hyperbranched polymer, the monodisperse dendron and dendrimer it is obvious that the SEC

PDI’s are incorrect. The PDI of Boltorn H-30 in N,N-dimethylformamide and THF utilizing SEC

is, according to a study performed by Månson et al. at least 2.

The Mark-Houwink (MH) data permits the calculation of the number of arms on a star polymer. Polymers with long chain branches such as star- and comb-branched polymers have reduced hydrodynamic radius and intrinsic viscosity [η] compared to linear polymers of the same molar mass. By comparing the intrinsic viscosity of a branched polymer and a linear reference polymer it is possible to estimate the degree of branching. The number of arms was calculated using the Zimm-Stockmayer equations.

Through the comparison of the intrinsic viscosity of the branched polymer ([η]

) and a linear equivalent ([η]

), the branching index g is calculated:

[ ] [ ]

ε

η η

 

 



= 

l

g

where ε is a form factor dependent on the type of branched polymer.

The form factor is generally considered to be ∼0.75. This definition of g is less exact than the ratio of the radius of gyration (R

), however it is more practical to use due to the problems associated with measuring R

over the whole molecular weight distribution. The number of arms is then calculated using

( 3 f 2 ) / f

g = −

where f is the number of arms. The calculated number of arms is in good agreement with the theoretical number of arms for the star polymers with a dendron and dendrimer core moieties.

Their number of arms is 8 and 24 respectively. At lower DP the measured number of arms is lower than the theoretical value. This was expected since the

C NMR showed that functionalization at low DP’s was incomplete. The Boltorn-PCL with theoretically 32 arms, however, displayed a higher number of arms. This can be contributed to the nature of Boltorn, a hyperbranched polymer with a lower degree of structural uniformity.

Appendix B lists SEC as determined by SEC

and viscosity data of all star-branched PCL.

3.3 Star-Branched Poly(neopentyl carbonate)’s

One of the goals of this work was to identify a solid material suitable for low temperature

curing. Since the thermal properties of branched poly(ε-caprolactone)’s are unsuitable for

3.3.1 Synthesis

3.3.1.1 Monomer Synthesis

Preparation of NPC from diethyl carbonate and neopentyl glycol as previously described by Sarel et al.

References to production of NPC are also found in patent literature.

The reaction proceeded in two steps. In the first step, oligomer or prepolymer was formed which was subsequently subjected to pyrolysis and ring closure resulting in a cyclic monomer. The synthetic route used in this work was similar to the patented procedures in that it omits an extraction step. Sn(Oct)

was used as catalyst to ensure formation of prepolymer and increase the rate of ring closing depolymerization. All volatile by-products and reagents including residual neopentyl glycol were removed under reduced pressure and NPC was obtained by pyrolysis at 210°C followed by distillation (scheme 3.2). This elaborated method does not require extraction and the monomer could be conveniently prepared in 200 g scale in one pot with a 70% yield.

O

H OH O O

O

O O

O Sn(Oct)₂,130°C

+

Prepolymer

Scheme 3.2 Synthetic route for NPC.

3.3.1.2 Polymer Synthesis

An experimental series was designed in order to find favorable conditions for the synthesis of star-branched PNPC’s with cores consisting of various polyols. Polymers with three and four arms were synthesized from trimethylolpropane (TMP) (3OH), di-TMP and etoxylated pentaerythritol (PP50) (4OH). TMP and di-TMP represent polyols with primary hydroxyl groups in neopentylic positions. PP50, on the other hand, has approximately 5 ethylene oxide- repeating units per molecule and the hydroxyl groups in this polyol are less sterically crowded. Star-branched polymers were also synthesized from a Boltorn H30 (3

pseudo generation) (Appendix C). In addition, linear PNPC by initiation from n-BuOH was synthesized and used as a reference compound.

O

H OH

OH

O H

OH O

O H

OH

O O OH

O OH O

O H

O

O

TMP Di-TMP H PP50

Figure 3.5 Three and four functional core molecules.

Previously reported work suggests tin(II) compounds and mild organic acids as promising

catalysts for realization of the target structures. Kricheldorf et al. reported the synthesis of

linear PNPC with a molecular weight of up to 250 kg mol

by a coordination insertion mechanism utilizing Sn(Oct)

.

Hedrick et al. have reported stannous(II) trifluoromethane sulfonate (Sn(OTf)

) as an efficient catalyst for polymerization of lactones under milder conditions than Sn(Oct)

.

NPC has also been polymerized cationically. There are two proposed mechanisms, the activated monomer mechanism and the activated chain end mechanism. Early work on cationic polymerization of NPC reports side reactions causing decarboxylation that result in formation of ether groups in the growing chains. Reported initiator systems consist of strong acids, such as triflic acid or boron trifluoride,

and more recently living systems, such as triethyl borate/HCl

Et

O.

However, the use of strong acids can increase the risk of side reactions during polymerization, and the nature of the alkyl borate initiating species prevents initiation from a multifunctional scaffold. The catalysts chosen for this study were Sn(Oct)

, Sn(OTf)

and fumaric acid (pK

= 3.02). The choice of fumaric acid was inspired by the work of Endo et al. who recently synthesized ε-caprolactone star-branched polymers utilizing this catalyst.

3.3.2 NMR Characterization

1

H NMR was employed to monitor conversion and estimate the degree of polymerization

(DP

) (figure 3.6). Conversion was calculated by comparing the peak integral at 1.12 ppm

(b) with the peak at 0.99 ppm (b’). The peak 1.12 ppm (b) corresponds to the protons of the

methyl groups in the monomer and the peak at 0.99 ppm (b’) represents the methyl groups of

the polymer backbone. The DP

was calculated by comparing the integrals of the repeating

unit peak at 0.99 ppm (b’) and the peak at 0.93 ppm (d), corresponding to the methyl groups

in the terminal repeat unit. The DP

is thus equal to I

/I

+1. This calculation can also

be performed on the integral at 3.34 ppm (c) that corresponds to the CH

in the α position to

the terminal hydroxyl group. In this case DP

is equal to I

/(3×I

)+1. If di-TMP is

present as the initiating polyol, the protons located next to the ether bond in di-TMP yield a

resonance peak in the vicinity of 3.3 ppm, which makes it impossible to calculate DP using

the peak integral at 3.34 ppm (c). However, the peaks arising from the ether bond in di-TMP

indicate where peaks would likely emerge if ether bonds were formed during the

polymerizations. Since decarboxylation and the formation of undesirable ether bonds is

known to occur in cationic polymerization of cyclic carbonates.

The fact that the integral

of the 3.3 ppm resonance peak did not increase during the polymerization from di-TMP and it

was not found in any of the other prepared polymers, it is evident that no ether bonds were

formed during polymerization.

(ppm)

0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0

O O

O O O O

R O O

n O O H

a b a´

b´

c d

a d

b a´

b´

c HBP

HBP

Figure 3.6

H NMR spectra of a Boltorn-PNPC containing unreacted monomer.

The degree of substitution of the hydroxyl groups of Boltorn was investigated using

C NMR. As previously mentioned, Hult et al. have shown that the

C NMR resonance corresponding to the quaternary carbon of the bis-MPA repeat unit is dependent on the degree of substitution.

The resonance shifts for the fully substituted, mono-substituted and un- substituted Boltorn are 46, 48 and 50 ppm respectively. A distinct peak was detected at 46 ppm accompanied by a smaller peak at 48 ppm, suggesting that the majority of the hydroxyl groups of Boltorn were substituted (figure 3.7).

(ppm)

10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160

(ppm) 46 48 50

Figure 3.7

C NMR spectra of a Boltorn-PNPC. The enlarged area shows the quaternary carbons.

Boltorn Boltorn

3.3.3 SEC Characterization

A universal calibration (SEC

) method was used to study the obtained star polymers while triple detection including light scattering was excluded due to a poor signal to noise ratio. All polymers were synthesized using fumaric acid as the catalyst.

The relationship between intrinsic viscosity the degree of branching is discussed in section 3.2.3. This relationship can also be used to rank polymers according to branching by comparing the Mark-Houwink (MH) plots of the respective polymers.

The MH plots for linear, 4-arm and multi-arm star-branched polymers are depicted in figure 3.8. A trend is clearly seen with decreasing slope in the order linear, four arm and multiarm polymer. This suggests successful grafting of PNPC onto the multifunctional initiator molecules.

2.91

-1.60

-1.26

-0.92

-0.59

-0.25

Linear

4 arm star

Hyperbranched

2.91

-1.60

-1.26

-0.92

-0.59

-0.25

Linear -PNPC

PP50-PNPC

Boltorn-PNPC

Log[M]

Figure 3.8 Mark-Houwink plot of linear and star-branched PNPC.

Relative response

that the low molecular weight region has a different slope. The four arm star and the linear polymers are unsymmetrical, an indication of side reactions.

3.3.4 Catalyst Evaluation

Linear PNPC initiated from n-BuOH were synthesized in order to evaluate the catalytic performance of Sn(OTf)

in toluene under mild conditions. The degree of polymerization (DP) controlled by the monomer to initiator feed ratio The DP aimed for (DP

) were 20 and 100. The polymerizations were performed in toluene at 50°C in order to prevent n-BuOH from boiling. The polymerizations were run to 90% conversion; the yields after precipitation in cold methanol were 70 and 60% for the DP

20 and 100 respectively (table 3.2). The low yield is the result of side reactions that result in a fraction of soluble oligomers, and this decrease in yield was most evident at the higher monomer to initiator ratio. The SEC trace was bimodal at the high DP

with moderate PDI in both cases.

The efficiency of the catalysts to form star polymers was evaluated by reacting NPC with di- TMP in presence of the respective catalysts: Sn(Oct)

, Sn(OTf)

and fumaric acid (table 3.2).

The polymerizations were performed in bulk at 130°C in order to maintain low viscosity at high conversions. The polymerizations utilizing Sn(Oct)

and Sn(OTf)

exhibited high PDI’s, indicating poor control. The PDI was significantly lower for the polymer synthesized using fumaric acid. As fumaric acid resulted in the lowest PDI it was selected for further study. A linear relationship between conversion and molecular weight was observed for conversion up to 80-90% (figure 3.10). Polydispersities were fairly low (1.2) at conversions below 50% and increased gradually up to 2-2.5 above 90% conversion. The observed increase in M

and broadening of PDI show an increase in side reactions at high conversion. However, observed side reactions in this reaction were significantly lower than observed for the tin catalysts (table 3.2). All polymerizations showed M

values lower than the theoretical values. The deviation of M

from the theoretical value was greater at higher DP’s, and may to some extent be attributed to spontaneous thermal polymerization or initiation from impurities. An experiment was therefore performed to determine the extent of thermal polymerization or decomposition of monomer; NPC was heated at 130°C for 24 hours resulting in 6%

conversion according to

H NMR.

0 1000 2000 3000 4000 5000 6000

0 20 40 60 80 100

Conversion (%) M

(g mol

)

1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2

PDI

Mn, NMR Mn, SEC PDI

Figure 3.10 M

obtained from

H NMR, SEC and PDI as a function of conversion for the

polymerization of NPC with a DP

of 40 by fumaric acid at 130 ° C.

Table 3.2 Data of the ROP of NPC with hydroxyl functional initiators in the presence of different catalysts.

Initiator Catalyst DP

M

theo.

Temp.

(°C) Time (h)

Conv.

(%)

Yield (%)

M

NMR

M

SEC

M

SEC PDI α n-BuOH Sn(OTf)

20 2 700 50 48 93 70 1 800 2 400 2 700 1.2 0.69 n-BuOH Sn(OTf)

100 13 100 50 48 91 60 2 100 3 800 4 700 1.2 0.70 Di-TMP Sn(Oct)

10 5 400 130 16 85 60 4 700 4 700 18 400 3.9 0.43 Di-TMP Sn(OTf)

10 5 400 130 4 90 70 4 800 4 900 13 800 2.8 0.30 Di-TMP Fumaric acid 10 5 400 130 9 70 55 3 400 3 100 4 300 1.4 0.39 TMP Fumaric acid 5 2 100 130 5 83 68 1 900 1 800 2 900 1.6 0.30 TMP Fumaric acid 20 7 800 130 9 47 35 3 400 3 400 4 500 1.3 0.12 PP50 Fumaric acid 10 5 600 130 9 75 60 3 900 4 700 6 000 1.3 0.46 PP50 Fumaric acid 20 10 700 130 20 90 75 7 600 6 600 9 000 1.3 0.46 Boltorn Fumaric acid 5 23 000 130 4 73

22 000 - - - - Boltorn Fumaric acid 10 42 000 130 6 80 65 30 000 8 500 52 000 6 0.12 Boltorn** Fumaric acid 10 42 000 130 24 80

30 000 25 000 99 000 4 0.28 Boltorn Fumaric acid 30 120 000 130 24 92 75 47 000 4 000 20 000 5 0.12

a g mol^-1. * No precipitation. ** Data in this row is from a fractional precipitation of the sample above it.

The polymers derived from PP50 showed lower polydispersities at high conversion (90%) compared to polymers based on initiators with neopentylic hydroxyl groups (table 3.2). The SEC

values obtained for the three and four arm star polymers were close to those obtained by

H NMR. Polymerizations initiated from Boltorn in the presence of fumaric acid had aliquots taken for

H NMR analysis. The analysis of synthesis with a DP

of 10 showed that the molecular weight increased up to a DP of 7-8. Additional polymerizations were performed with DP

of 5, 10 and 30. Conversions were kept below 90% to avoid gelling.

3.3.5 Thermal Characterization

The precipitated polymers were subjected to differential scanning calorimetry (DSC) analysis that comprised two melting/crystallization cycles with heating and cooling rates of 10°C/min. In the first cycle the sample was heated from 25 to 140°C and then cooled to - 30°C. The sample was then heated to 140°C to complete the procedure. On the second heating, melting endotherms were only observed for three and four arm star polymers with an arm length of 8-10 repeating units or more (figure 3.11). A one-hour annealing segment at 50°C was therefore added prior to the second heating but no or little effect was observed. A T

was also observed between 20-30°C. Some endotherms were bimodal or very broad on the second heating. Boltorn-PNPC DP

=30 was the only multiarm star that exhibited an endotherm in the second heating. The two distinct melting endotherms that may be attributed to the presence of linear polymer (figure 3.11).

-50 0 50 100 150

Linear PNPC Di-TMP-PNPC PP50-PNPC Boltorn-PNPC

E ndo

Temperature (°C)

Figure 3.11 DSC traces for linear and star polymers, 2

heating after annealing 50 ° C, 1 h.