Reversible networks in supramolecular polymers

Citation for published version (APA):Havermans - van Beek, D. J. M. (2007). Reversible networks in supramolecular polymers. Eindhoven:Technische Universiteit Eindhoven. https://doi.org/10.6100/IR631309

DOI:10.6100/IR631309

Document status and date:Published: 01/01/2007

Document Version:Publisher’s PDF, also known as Version of Record (includes final page, issue and volume numbers)

Please check the document version of this publication:

• A submitted manuscript is the version of the article upon submission and before peer-review. There can beimportant differences between the submitted version and the official published version of record. Peopleinterested in the research are advised to contact the author for the final version of the publication, or visit theDOI to the publisher's website.• The final author version and the galley proof are versions of the publication after peer review.• The final published version features the final layout of the paper including the volume, issue and pagenumbers.Link to publication

General rightsCopyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright ownersand it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.

• Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain • You may freely distribute the URL identifying the publication in the public portal.

If the publication is distributed under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license above, pleasefollow below link for the End User Agreement:www.tue.nl/taverne

Take down policyIf you believe that this document breaches copyright please contact us at:openaccess@tue.nlproviding details and we will investigate your claim.

Download date: 26. Jun. 2020

https://doi.org/10.6100/IR631309https://doi.org/10.6100/IR631309https://research.tue.nl/en/publications/reversible-networks-in-supramolecular-polymers(15d8a1b3-17d9-4290-8ed7-0d4df6abcc12).html

Reversible Networks in Supramolecular Polymers

Reversible Networks in Supramolecular Polymers

PROEFSCHRIFT

ter verkrijging van de graad van doctor aan de Technische Universiteit Eindhoven, op gezag van de

Rector Magnificus, prof.dr.ir. C.J. van Duijn, voor een commissie aangewezen door het College voor Promoties in het openbaar te verdedigen op donderdag 13 december 2007 om 16.00 uur

door

Dimphna Johanna Maria Havermans–van Beek

geboren te Weert

Dit proefschrift is goedgekeurd door de promotoren: prof.dr. R.P. Sijbesma en prof.dr. E.W. Meijer This research has been financially supported by NanoImpuls/NanoNed, the nanotechnology program of the Dutch Ministry of Economic Affairs. Omslagontwerp: Linda Havermans–van Beek en Paul Verspaget Grafische Vormgeving – Communicatie Druk: Gildeprint Enschede A catalogue record is available from the Eindhoven University of Technology Library. ISBN: 978-90-386-1165-5

“The most fundamental and lasting objective of synthesis is not production of new compounds,

but production of properties”

George S. Hammond – 1968

Table of Contents Chapter 1. Dynamic Material Properties of Supramolecular Polymers 1

1.1 Supramolecular Polymer Chemistry 2 1.2 Hydrogen Bonded Supramolecular Polymers 2 1.3 UPy Functionalized Supramolecular Polymers 4 1.4 Dynamics in Supramolecular Polymers 6 1.4.1 Stress Relaxation via Reptation 7 1.4.2 The Sticky Reptation Model 7 1.4.3 Multiple Cross–links in Reversible Gels 10 1.5 Recent Developments 11 1.5.1 Effect of Temperature and Chain stoppers on Rheological Properties 11 1.5.2 Hydrogen Bonding Nucleobases as Cross–Linking Units 13 1.5.3 Dynamics in Metallo–Supramolecular Polymers 14 1.6 Research Aim 15 1.7 Outline of this Thesis 16 1.8 References 18

Chapter 2. Supramolecular Copolyesters with Tunable Properties 21

2.1 Introduction 22 2.2 Synthesis 23 2.2.1 Synthesis of High Molecular Weight pCL, pVL, and p(CL–co 50% VL) 23 2.2.2 Synthesis of Supramolecular (Co)Polymers 24 2.3 Thermal Properties 26 2.4 Mechanical Properties 30 2.5 Conclusions 32 2.6 Acknowledgements 33 2.7 Experimental Section 33 2.8 References 36

Chapter 3. Unidirectional Dimerization and Stacking of Ureidopyrimidinone End Groups in Polycaprolactone Supramolecular Polymers 39

3.1 Introduction 40 3.2 Synthesis 42 3.3 Degree of Functionalization 44 3.4 Thermal Analysis by Differential Scanning Calorimetry 45 3.5 Variable Temperature Infrared Spectroscopy 46 3.6 Atomic Force Microscopy 48 3.7 Oscillatory Shear Experiments 49 3.8 Discussion 53

Table of Contents 3.9 Conclusion 58 3.10 Acknowledgments 58 3.11 Experimental Section 58 3.12 References 62

Chapter 4. Control of the Degree of Polymerization of a Linear Supramolecular Polyester in the Melt 65

4.1 Introduction 66 4.2 Synthesis and Characterization of Chain Stopper 67 4.3 Oscillatory Shear Experiments 67 4.4 Discussion 70 4.5 Conclusions 73 4.6 Acknowledgments 73 4.7 Experimental Section 73 4.8 References and Notes 74

Chapter 5. Melt Behavior of Supramolecular Polyester Networks 75

5.1 Introduction 76 5.2 Synthesis and Characterization 77 5.3 Oscillatory Shear Experiments 79 5.4 Discussion and Conclusions 82 5.5 Acknowledgment 84 5.6 Experimental Section 84 5.7 References 85

Chapter 6. Phase Transitions in Supramolecular Polyesters 87

6.1 Introduction 88 6.2 Synthesis 89 6.3 Thermal Characterization of Supramolecular Polyesters 90 6.4 Fluorescence Spectroscopy Measurements 93 6.4.1 Phase Transitions in PCL 2k 93 6.4.2 Glass Transition Temperatures of Supramolecular Polyesters 95 6.4.3 Crystallization & Melting Temperatures of Supramolecular Polyesters 97 6.5 Conclusions 99 6.6 Acknowledgments 100 6.7 Experimental Section 100 6.8 References 102

Table of Contents

Chapter 7. Quadruple Hydrogen Bonding in Micellar Environments 105

7.1 Introduction 106 7.2 Aqueous–Micellar Distribution Coefficients in SDS Solution 107 7.3 UV Spectroscopy of Micellar Solutions 110 7.4 Towards Transient Networks 113 7.5 Discussion 114 7.6 Conclusions 115 7.7 Acknowledgements 115 7.8 Experimental Section 115 7.9 References 117

Chapter 8. Dendrimer–Based Supramolecular Transient Networks 119

8.1 Introduction 120 8.2 Design and Synthesis of the Modules 121 8.3 Molecular Characterization of the Supramolecular Complex 122 8.4 Dynamics of the Supramolecular Structures 124 8.5 The Three Supramolecular Structures 127 8.6 Conclusions 128 8.7 Acknowledgments 129 8.8 Experimental Section 129 8.9 References 132

Summary 135

Curriculum Vitae 139

List of Publications 141

Dankwoord 143

11 Dynamic Material Properties of Supramolecular Polymers

ABSTRACT. Non–covalent interactions between low molecular weight polymers are the basis of supramolecular polymers. In this Chapter, we will focus on the properties of hydrogen bonded supramolecular polymers, as these are the subjects of all Chapters in this Thesis. The strength and lifetime of the non–covalent, reversible interactions determine the material properties of these polymers. Due to the reversibility of the interactions between the low molecular weight polymers, stress relaxation is not exclusively determined by reptation but also the lifetime of the non–covalent interaction plays a role. This Chapter gives a brief overview of achievements in the incorporation of lifetime of the non–covalent interaction into established polymer theories, followed by the research aim and the outline of this Thesis.

Chapter 1

2

1.1 Supramolecular Polymer Chemistry

The word ‘polymer’ originates from the Greek ‘polymerus’ (‘poly’ = many & ‘meros’ = part) and was introduced as a chemical term by the Swedish chemist Jöns Jakob Berzelius (1779–1848). Berzelius is, together with John Dalton and Antoine Lavoisier, considered a father of modern chemistry, who besides the term ‘polymer’ is also credited for the identification of several chemical elements such as silicon, selenium and cerium.1 Wallace Carothers redefined the term polymer in 1929, shortly after Hermann Staudinger proved the existence of long chain ‘high molecular weight compounds’. In Carothers’ definition the monomeric unit can only exist as a part of a polymer and must be covalently linked to other monomeric units to form a polymer.2,3

Around 1990, the first intentionally designed supramolecular polymers were reported by Jean–Marie Lehn, while at the same time Takashi Kato and Jean Fréchet were working on a hybrid class of materials; side–chain branched polymers.3 The idea that polymers only consist of monomeric units that are covalently linked was thereby overthrown and the area in supramolecular chemistry, in which monomeric units self–assembly into polymeric structures, is nowadays known as supramolecular polymer chemistry. Supramolecular chemistry relies on weak and reversible non–covalent interactions, such as hydrogen bonding, metal coordination, hydrophobic forces, van der Waals forces, π–π interactions, and electrostatic effects to assemble molecules into multimolecular complexes.4

Nowadays the field of supramolecular chemistry is broad and many examples of supramolecular polymers based on the above–mentioned non–covalent interactions are known in literature.4 In this Chapter, the class of hydrogen bonded supramolecular polymers will be highlighted as all supramolecular polymers described in this Thesis are based on this non–covalent interaction.

1.2 Hydrogen Bonded Supramolecular Polymers

Due to their high directionality, hydrogen bonds are attractive interactions to create supramolecular polymers as they facilitate strong binding when used in arrays. In general, the stability of an array increases with increasing hydrogen bonds; three hydrogen bonds are stronger than two, while four are stronger than three. This is already clear for the nucleotides of DNA5; adenine and thymine assemble via two hydrogen bonds and the corresponding association constant, Ka = 102 M–1, is lower than the association constant of the assembly of guanidine and cytosine (Ka = 104–105 M–1), consisting of three hydrogen bonds.

Besides the number of hydrogen bonds, also the arrangement of the donor and acceptor functionalities in an array determines strength between complementary arrays and thereby influencing the association constant.6 In Figure 1.1, different association constants of complexes formed by hydrogen bonds between arrays of three donors/acceptors are shown to illustrate the dependence of the association constant on the

Dynamic Material Properties of Supramolecular Polymers

3

arrangement of the donors (D) and acceptors (A). The lowest association constant (102–103 M–1) is observed for the complex between molecules 1 and 2, which is due to the repulsive interactions between the DAD–ADA arrays. The presence of attractive secondary interactions next to the repulsive interactions in the DAA–ADD array of molecules 3 and 4 provides an association constant which is about 2 orders higher: 104–105 M–1. An association constant higher than 105 M–1 is observed when exclusively attractive interactions are present in an array, such as shown for molecules 5 and 6, which self–assemble according to an AAA–DDD array.6b

Figure 1.1: Stability of complexes based on triple hydrogen bonds with different hydrogen bonding motifs.6b Multiple hydrogen bonding arrays are widely used and discussed in supramolecular polymers6; a few examples and important contributions will be noted here.

Lehn7 has reported the use of triple hydrogen bonds as hetero–dimers for the complementary groups P and U with long chain derivatives of tartic acid as depicted in Figure 1.2. The individual species are solids, while the hydrogen bonded thermotropic mesophases form a hexagonal columnar superstructure after mixing of TP2 and TU2.7

Chapter 1

4

Figure 1.2: Supramolecular polymer based on assembly via triple hydrogen bonds between TP2 and TU2.7a More examples of triple hydrogen bonds in supramolecular polymers have been published by among others Zimmerman8, interactions using melamine derivatives by Lange9 et al., interactions between poly(vinyl–2,6–diamino–diaminotriazine) with small molecules by Komiyama10 and coworkers, Würthner11, and Long12. Binder13 and coworkers reported on networks formed by end group–modified poly(ether ketone) and poly(isobutylene) with 2,6–diamino–1,3,5–triazine derivatives. Liquid crystals assembled via complementary nucleobases and bis(phenylethynylene) benzene cores were recently reported by Rowan14 and coworkers.

Corbin and Zimmerman15 published a quadruple hydrogen bonding array that self–assembles regardless of the protomeric form with a high dimerization constant. Different quadruple hydrogen bonding arrays were reported by for example Sessler16, Lüning17, and Gong18. The use of ureidotriazine19 and ureidopyrimidinone20 moieties in the creation of supramolecular polymers was introduced by our group and the use of the ureidopyrimidinone will be discussed in the next section. Arrays with more than four hydrogen bonds have been reported by, among others, Binder21, Lehn22, Zimmerman23, and Gong24.

1.3 UPy Functionalized Supramolecular Polymers

The commonly used quadruple hydrogen bonding array 2–ureido–4[1H]–pyrimidinone (UPy), was developed in our laboratory20 in the 1990’s. The DDAA array is self–complementary, with an association constant of 6·107 M–1 in chloroform25 and it is easy to synthesize. The UPy moiety can adapt three different tautomeric forms, which are solvent dependent.20a For example, the most stable tautomer, 6[1H]–pyrimidinone tautomer (keto–2) is present in DMSO. The formation of very stable dimers via hydrogen bonds induces tautomerization to the, less stable, 4[1H]–pyrimidinone tautomer (keto) in

Dynamic Material Properties of Supramolecular Polymers

5

chloroform. The 2–ureidopyrimidin–4–ol or enol tautomeric form is more prevalent in THF and in toluene. Figure 1.3 displays the chemical structure of the UPy and its tautomers.

N

N

O

R NH

H

O NH

R'

N

N

O

R NH

H

O NH

R'

N

N

O

RNH

H

ONH

R'

N

N

O

RH

NH

N OR'

H

N

N

O

RH

NH

N OR'

HN

N

O

RH

NH

NOR'

H

N

N

O

R N

H

H

O

NH

R'4[1H]keto

4[1H]enol

6[1H]keto-2

Figure 1.3: Tautomers of 2–ureido–pyrimidinone, the self–complementary keto and enol tautomers and the non–self–complementary keto–2 tautomer.

Shortly after the introduction of the UPy moiety as multiple hydrogen bonding array, low molecular weight polymers, siloxane, PEO and PPO telechelics20a,26, were functionalized with this moiety, yielding reversible polymer networks. Due to the well–defined dimerization of the UPy moiety, no additional stabilization such as phase separation or crystallization of the polymeric components is required for the formation of a network. Functionalization of poly(ethylene butylene)27, polytetrahydrofuran28, polycarbonates28,29, polyesters30,31, acrylates32, polyolefins33, and oligo–siloxanes34 followed.

Figure 1.4: (top) Master curves of UPy–T–PEB–T–UPy (black squares) and hydroxyl telechelic poly(ethylene butylene).27a (bottom) Chemical structure of UPy–T–PEB–T–UPy.

Dramatic changes in material properties by examination in solid and/or molten state between modified and unmodified polymers have been reported. Complex

Chapter 1

6

thermorheological behavior was observed by examination of functionalized poly(ethylene butylene) (UPy–T–PEB–T–UPy) in the melt. (Figure 1.4)

Differences have not exclusively been observed between non–functionalized and UPy functionalized polymers. Kautz27c et al. showed enhanced lateral interactions after replacing the functional group between the poly(ethylene butylene) backbone and the UPy moiety from an urethane to a urea (UPy–U–PEB–U–UPy). The lateral interactions resulted in fibers, which were visualized by atomic force microscopy (AFM) (Figure 1.5).

10-2 10-1 100 101 102 103 10410-1100101102103104105106

10-1100101102103104105106

G''

G'G'

G''

G"G

' & G

'' [P

a]

ω [rad/s]

UPy-PEB-UPy UPy-U-PEB-U-UPy

Figure 1.5: (left) Tapping mode AFM images of thin films of a) UPy–PEB–UPy, b) U–U–PEB–U–U, c) UPy–T–PEB–T–UPy, and d) UPy–U–UPy–U–UPy. (right) Master curves of UPy–PEB–UPy (closed symbols) and UPy–U–PEB–U–UPy (open symbols). Reference temperatures 110 and 125 °C, respectively. 27c The results were compared with a supramolecular polymer in which the UPy was attached to the polymer backbone without a functional group (UPy–PEB–UPy). The presence of lateral stacking in the urethane and urea functionalities or the absence of functionality, and thus lateral stacking, led to differences in the master curves as shown in Figure 1.4 and 1.5. Higher values for the loss and storage modulus and slower dynamics were observed for the polymers in which the urethane or urea functionalities are present.

The above–mentioned results indicate the complex relation between strength of the non–covalent interaction and the dynamics of the obtained supramolecular polymer.

1.4 Dynamics in Supramolecular Polymers

Since the appearance of the first supramolecular polymers in literature, attempts have been made to describe the dynamics of supramolecular polymers by adapting existing polymer theories. Whereas stress relaxation in covalent polymers is mainly determined by reptation of the polymer, in supramolecular polymers also the dynamics of the non–covalent interaction play an important role. The incorporation of the dynamics of these non–covalent interactions into well–established polymer laws has attracted attention from both chemists and physicists. Many contributions have been made in this field, such

Dynamic Material Properties of Supramolecular Polymers

7

as the models developed by Semenov/Rubinstein35 and Jongschaap/Wientjes36, and as a consequence it is impossible to give a complete overview. A few important and interesting examples will be highlighted in this section.

1.4.1 Stress Relaxation via Reptation

A first attempt to introduce the reversibility of non–covalent interactions into a model was made by Cates37 in the late 1980’s, when he proposed a model for the stress relaxation in worm–like micelles. In this model, simple reaction kinetics for these linear polymers are assumed; the chain breaks with equal probability per unit time per unit length at all points in the chemical sequence, furthermore, stress relaxation of the polymer chain occurs via reptation. Recombination of two chains takes place in a rate proportional to the product of end group concentrations. In addition to the reptation time (τrep), Cates defined a mean time for a chain to break in two pieces (τbreak).

ζ was introduced as τbreak/τrep and two limits for ζ were defined. In the first limit, ζ ≥ 1, the dominant stress–relaxation is reptation as τrep is significantly smaller than τbreak. In the other limit, 0 < ζ ≤ 1 (τrep ≥ τbreak), the overall relaxation time corresponds to τ = (τrep×τbreak)1/2. In this case, τ is associated with a process whereby the chain breaks and is able to relax via reptation before the chain end is lost by recombination. During the relaxation of a given tube segment the chain (occupying the tube) undergoes many scissions and recombinations and the memory of either the initial length of the chain, or the position on the chain, initially corresponding to a segment, is lost. As all segments relax at the same rate, there is no dispersion of relaxation times and thus no dispersion in the self–diffusion spectrum. Therefore, a single exponential stress relaxation is a characteristic experimental feature of many supramolecular polymers with fast–breaking non–covalent interactions compared to reptation time.

The model described above has been used to describe the rheological behavior of worm–like surfactants as reported in a clear review.38

1.4.2 The Sticky Reptation Model

An important contribution to the development of an adequate theory for the description of dynamics of reversible networks was made by Leibler, Rubinstein and Colby in 199139, who extended the modified reptation model postulated by Gonzalez.40 In the LRC model (the sticky reptation model), an entangled chain can move by coherent breaking of only a few cross–links instead of all cross–links. As a result it is possible to predict the self–diffusion coefficient as a function of the fraction of closed ‘stickers’. According to this model, the stress relaxation modulus of a thermoreversible network is characterized by four regimes, which are separated by three important time scales in the reversible gel: the Rouse relaxation time (τe), the lifetime of closed ‘stickers’ (τ), and the terminal relaxation time (Td0) as schematically represented in Figure 1.6.

Chapter 1

8

Figure 1.6: Schematic comparison of the relaxation moduli for reversible networks comparing linear chains with ‘stickers’ (solid line) and linear chains without ‘stickers’ (dashed line).39

At time scales shorter than the Rouse time of an entanglement strand (t < τe), the stress relaxation modulus is determined by local relaxations and is indistinguishable from that in a polymer without ‘stickers’. At times larger than the Rouse relaxation time, but shorter than the lifetime of closed ‘stickers’ (τe < t < τ) the behavior of the network is similar to a covalently cross–linked network with modulus G1. At τ, the timescale at which the ‘stickers’ open, the modulus decreases to the level of identical linear chains without ‘stickers’ (G2) as the stress held by the ‘stickers’ relaxes by opening. The plateau of G2 will exist until the terminal relaxation time of the reversible gel (Td0).

However, secondary relaxation processes like fluctuations in tube length and constraint release have not been included so far. These processes will lead to lowering of the modulus level at given time and will broaden the spectrum of relaxation times and as a result Td0 will shift to Td. A striking feature of truly reversible networks is the existence of two maxima in the loss modulus measured by oscillatory shear experiments at temperatures above the glass temperature.

The sticky reptation model indicates that the lifetime of a cross–link can be determined directly from linear viscoelastic measurements in the high frequency regime, using spectroscopic measurements. Moreover, this model can be used to predict the increase in viscosity and the decrease in diffusion of molecules in reversible gels.39

Figure 1.7: Thermo reversible network based on phenylurazole contacts.41

Dynamic Material Properties of Supramolecular Polymers

9

Leibler39 and coworkers validated their model by the experimental results of Stadler.41 Stadler42 in his turn used the sticky reptation model to investigate modified polybutadienes in detail. Due to the use of several experimental techniques such as oscillatory shear experiments, dielectric spectroscopy measurements (DS) (to obtain information in the high frequency domain), infrared dichroism and birefringence measurements (to obtain information about the orientation of the interactions) it is of interest to discuss this system in more detail.

Stadler41,42 and coworkers investigated modified thermoplastic elastomers that form temporary networks by hydrogen bonds between phenylurazole units as schematically represented in Figure 1.7. The modified polybutadienes were subjected to oscillatory shear experiments to investigate the influence of hydrogen bonds on the viscoelastic properties of the thermoreversible networks. Large, reversible, branched clusters are formed by association of the phenylurazole groups, leading to an increase in apparent effective molecular weight observed by increased moduli. The temperature dependence of the modified butadiene is explained by the presence of the hydrogen bonding groups in these molecules, which become less effective at high temperatures and as a result the activation energies at high temperatures are comparable with the ones of unmodified polymers. The absence of an equilibrium network modulus is supported by the analysis of the relaxation time spectra and the presence of Newtonian behavior at very low frequencies. It is suggested that reaching the thermodynamic equilibrium is either dependent on the concentration of hydrogen bonds or by topological restrictions imposed on the interacting groups as they are fixed to the polymer backbone. Introduction of extra hydrogen bonding possibilities by replacement of the ester group by an acid functionality on the phenylurazole groups results in failure of time–temperature superposition, broadening of the rubbery plateau, a narrow transition zone from rubbery plateau to terminal relaxation at high temperatures, and increase of the glass transition temperature.

Dielectric spectroscopy measurements on phenylurazole modified polybutadiene revealed the presence of an extra α–relaxation (α*) in addition to the α–relaxation corresponding to the dynamic glass transition and the β–relaxation correlated to the local segmental motions of polybutadiene. The α*–relaxation is assigned to the dissociation dynamics of complexed urazoles, while the uncomplexed units contribute to the α–relaxation. The strength of the α– and α*–relaxations are thereby dependent on the concentration of associating groups, while the β–relaxation is not affected by them.

Coupling coefficients were determined for these polymers using infrared dichroism and birefringence measurements and an orientational coupling between the urazole functional groups is indicated. At lower temperatures, the fraction of self–assembled dimers increases, resulting in an increase of the coupling coefficient, which is therefore related to longer lifetimes of urazole–urazole complexes.42

Chapter 1

10

1.4.3 Multiple Cross–links in Reversible Gels

In the development of a model for the description of the rheological behavior of multifunctional reversible gels te Nijenhuis43 and coworkers were inspired by the statistical cross–linking model formulated by Flory44 and Stockmayer45. The statistical network theory of Flory and Stockmayer describes the behavior of polymeric networks. These indefinite network structures develop during three–dimensional polymerizations and are characterized by gelation. These gels are formed after exceeding a critical value in the degree of branching or cross–linking. Cross–links46 can originate from different kind of interactions such as crystallization of parts of polymers, phase separation in solution and bulk, hydrogen bonding interactions, complexation of polymer chains, helix formation, and interactions between side–chains. Stockmayer derived a relationship between the cross–linking index (γ ), the average number of cross–links per primary

polymer molecule, and the sol fraction (ws). Besides the sol fraction, the system consists of a fraction of non–ideal network (wn), which is composed of an ideal network fraction (wm) with a fraction dangling ends (wf) (Figure 1.8A). Before the gel–point, the ‘network’ consists only of the sol fraction (ws=1 and wn=0).

Flory and Stockmayer’s theory was limited to modeling of tetrafunctional cross–linking polymers with a monodisperse molecular weight distribution, te Nijenhuis43 extended the model to a multiple cross–linking functionalities, f, and to a polydisperse molecular weight distribution. The model of te Nijenhuis was validated by oscillatory shear experiments on several thermoreversible gels.47

A

Figure 1.8: (A) schematic representation of gel network with cM as average mol mass between cross–links and γ as average cross–linking index, (B) chemical structure and schematic representation of discotic side chain polymer.48a

One of the systems investigated using this model is a solution of the discotic side chain polymer shown in Figures 1.8B.48 Using the equilibrium shear modulus, the complex formation enthalpy of the cross–linking process was calculated while assuming constant cross–link functionality and a slight increase in the number of cross–links per unit volume upon reduction of temperature. The formation enthalpy of a cross–link is very

Dynamic Material Properties of Supramolecular Polymers

11

small, for example, with a functionality of 4, an average value of –4.1 kJ/mol is calculated. This value is much lower than the formation enthalpy of hydrogen bonded complexes studied in this Thesis. Investigation of the network at different temperatures led to the elucidation of the number of discs present in a cross–link. Most cross–links are composed of two discs and a minority of the cross–links is build up from three or four discs while the presence of larger cross–links is unlikely.48

1.5 Recent Developments

The dynamics of reversible polymers and polymer solutions continues to receive considerable attention. In this section a few recent contributions to the field of supramolecular polymer dynamics will be highlighted.

1.5.1 Effect of Temperature and Chain Stoppers on Rheological Properties

Bouteiller49 and coworkers have intensively investigated the properties of supramolecular polymers both in solution and in bulk. A hydrogen bonding bis–urea derivative (R1 = H, Figure 1.9) displayed a steep increase in specific viscosity upon increasing concentration, while the specific viscosity of a simple analogue capable of formation of two hydrogen bonds showed less concentration dependence.49a Oscillatory shear experiments displayed viscoelastic behavior of a 2 g/L bis–urea toluene solution (R1 = H, R2 = ethylhexyl). This behavior is attributed to hydrogen bonding interactions between the bis–urea molecules with the help of variable temperature dependent infrared (VT–IR) experiments. SANS measurements confirmed the organization of these molecules into long and rigid fibers.49b,c Differences in concentration dependence of specific viscosities were observed when using different substituents for R1 and R2, indicating that the strength of the hydrogen bonding interactions can be influenced by these substituents. Moreover, steric hindrance introduced by attachment of bulky substituents hinders the formation of hydrogen bonds and thus the formation of a polymeric structure.49c,d

N NNN

H

O

R2

R1 H R1

R2

O

Figure 1.9: General structure of bis–urea used by Bouteiller49 and Knoben.50 Similar to covalent polymers, the material properties of supramolecular polymers depend on the molecular weight of the polymer. As a consequence of the reversible nature of the non–covalent bonds, the degree of polymerization (DP) in a supramolecular polymer is a function of parameters such as the overall monomer concentration, temperature and association constant. Changes in rheological behavior are observed by adjustment of these parameters. Knoben50 et al. used chain stoppers to vary the degree of polymerization of bis–urea monomers that were introduced by Bouteiller49 and coworkers. Bis–urea EHUT (Figure 1.9, R1 = H, R2 = ethylhexyl) and DBUT (Figure 1.9, R1 = R2 = C4H9) were used as monomer and chain stopper, respectively. They expected that decreasing the DP either by

Chapter 1

12

temperature elevation or by addition of chain stoppers would have the same effect on zero–shear viscosity and terminal relaxation time. However, the relations between zero–shear viscosity and terminal relaxation time and DP were found to be completely different when changing DP by temperature or by chain stoppers.

Figure 1.10: Left: Arrhenius plot showing temperature dependence of zero–shear viscosity of 5.0 g/L EHUT solution (line corresponds to η0 = 3.78·10–19exp(1.41·104/T)). Right: Effect of stopper fraction on zero–shear viscosity of 5.0 g/L EHUT solution at 20 °C (line corresponds to η0 = 4·10–83.0, = average degree of polymerization).50c Exponential dependence of the zero–shear viscosity and terminal relaxation time on temperature (Figure 1.10, left) was observed in oscillatory shear experiments. Increasing the temperature from 20 to 70 °C resulted in a decrease in zero–shear viscosity of almost four orders of magnitude (Figure 1.10, left). The Arrhenius plot was used to calculate an activation energy of 150 kJ/mol, similar to activation energies observed for breaking of worm–like micelles.

In contrast to temperature elevation, a constant value in terminal relaxation time and zero–shear viscosity (Figure 1.10, right) is observed in the presence of small fractions of chain stopper. The DP is determined by the presence of free chain ends and the contribution of chain stopper can be neglected when present in small amounts. Above a certain fraction of chain stopper, the critical chain stopper fraction, both zero–shear viscosity and terminal relaxation time decrease strongly, following power–law behavior.

Static light scattering revealed a correlation length in EHUT solutions from which an association constant of 109 M–1 in cyclohexane was determined.50b Moreover, at fixed chain stopper concentration, a constant correlation length is found above a certain monomer concentration. This implies that the chain length is exclusively determined by the chain stopper fraction resulting in the observed power–law behavior (vide supra).50b Critical chain stopper fractions of 8·10–3 and 7·10–3 were calculated from zero–shear viscosity and relaxation time data, respectively.

From the results described above, it is clear that the use of chain stoppers is a valuable tool in the quest to understand the dynamics of supramolecular polymers.

Dynamic Material Properties of Supramolecular Polymers

13

1.5.2 Hydrogen Bonding Nucleobases as Cross–Linking Units

Hydrogen bonding arrays with low association constants can reach high degrees of polymerization in combination with polymer phase segregation.14 Functionalization of mesogen alkoxy–substituted bis(phenylethynyl)benzene with synthetically modified nucleobases resulted in supramolecular liquid crystalline polymers. Despite the low association constants of the nucleobases (vide supra), these polymers display stable LC phases and polymer properties, such as fiber formation.14 Further expansion of the use of nucleobases has resulted in hydrogen bonded supramolecular polymers after the functionalization of low molecular weight polymers with modified nucleobases.51 The material properties of these polymers were investigated using oscillatory shear experiments, X–ray, and VT–IR, and the outcome of these experiments will be discussed in more detail in this section.

Rowan51 and coworkers used modified nucleobases for the following reasons: 1) aromatic amide protected nucleobases reduce the type and number of hydrogen bonding possibilities and 2) phase segregation is encouraged by extension of the ‘hard’ nucleobase segment and the ‘soft’ poly(THF) backbone. The synthesis of nucleobased telechelics is schematically shown in Figure 1.11A and self–supporting films are obtained after functionalization with adenine (A) or cytosine (C).

Functionalization of poly(THF) with modified cytosine resulted in a critical gel. The reversible association of the telechelic network, forming the critical gel, was successfully analyzed by the theoretical gel model developed by Winter.52

Unusual mechanical behavior was observed for poly(THF) functionalized with modified adenine. Rheological simple behavior was observed at temperatures between 50 and 90 °C, in accordance with time–temperature superposition. The observed plateau modulus was higher than predicted and was suggested to originate from a combination of a stretched polymer backbone and phase segregation due to π–π stacking in the adenine unit (Figure 1.11B).

A

NH2OOH2N n

HOR

O

HNO

ONH n

RO

OR

NN

NN

NH

OOMe

R=

A

NN

NH

OO C B Figure 1.11: (A) Synthetic procedure to obtain ditopic macromonomers, (B) The schematic shows the segregation of A–THF–A of the nucleobase hard segments (discs) connected by chains of poly(THF) (lines) from a linear system to a gellike material.51b

Chapter 1

14

Due to the low cross–linking density, the system displayed simple thermorheological behavior similar to a linear polymer at temperatures below 90 °C. However between 70 and 90 °C, an increase in cross–links occurred due to a decrease in π–π stacking between the nucleobases with an affective increase in hydrogen bonding as result (Figure 1.11B). Gel behavior is observed at temperatures above 100 °C, characterized by parallel moduli with respect to the frequency. Similar to the cytosine modified polymer, loss of polymer–like properties is observed when the temperature is increased above 130 °C.

The polymer backbone size and composition or the strength of interactions between end– groups are most likely not exclusively responsible for the transitions observed in the gel phase, but also the dissociation rate of the chain end from the network plays an role.

1.5.3 Dynamics in Metallo–Supramolecular Polymers

Over the last few years, Craig53,54 and coworkers investigated the dynamics in metallo–supramolecular polymers, polymers in which metal coordination4 accounts for the non–covalent, reversible interaction. The formation of a three–dimensional associative polymer network based on specific metal–ligand coordination between bis–Pd(II) and Pt(II) organometallic cross–linkers and poly(4–vinylpyridine) (PVP) in DMSO (Figure 1.12) will be discussed briefly in this section.

These networks display viscoelastic responses and several parameters such as the dynamic viscosity and elastic storage modulus were determined. The dynamic mechanical properties were controlled by the choice of metal (Pt(II) or Pd(II)) used for coordination, which determines the dissociation rates of the cross–links. To summarize and conclude their findings the authors stated: ‘strong means slow’. This statement is related to the low frequency of dissociation when using a strong coordinating metal.

Recently, the same authors54b investigated the transition from a sol of discrete polymer aggregates to a percolated polymer network by oscillatory shear experiments on gels with different concentrations of pincer–metal cross–linkers. A critical concentration was determined from the sharp change in viscosity over a narrow cross–linker percentage, which marks the transition between the two phases. The authors suggest an equilibrium

Figure 1.12: Schematic representation of reversible cross–links between poly(4–vinyl pyridine) (PVP) through coordination with bis(MII–pincer) complexes: OTf = [CF3SO3]¯.53

Dynamic Material Properties of Supramolecular Polymers

15

between intra– and intermolecular cross–linkers because the total weight percent of the network determines the critical concentration. Changing the metal–pincer resulted in changes in viscosity over several orders of magnitude, which is therefore directly related to the lifetime of the cross–linking interactions.

1.6 Research Aim

Material properties of supramolecular polymers are highly dependent on the nature, strength, and lifetime of the non–covalent interaction, which ‘glues’ the low molecular weight monomeric precursors together to form high molecular weight polymers. In this Thesis, the relationship between material properties, dynamics, and network structure of hydrogen bonded supramolecular polymers is investigated in detail.

Ureidopyrimidinone (UPy) functionalized polymers were studied with techniques such as oscillatory shear experiments to establish the lifetime of the UPy moieties in the melt. The lifetime of this non–covalent interaction is important as it influences the stress relaxation of these supramolecular polymers. A short lifetime leads to a stress relaxation that is a combination of reptation and lifetime of the non–covalent interaction.37,38

It is known that the number of cross–links in a supramolecular system determines whether the system displays simple material or complex network behavior.48,51,53,54 The relationship between cross–link density and material properties, such as stress relaxation and phase transitions, of UPy polymers were investigated. The cross–link density can be tuned by small chemical modifications of the UPy moiety or in close vicinity of the UPy moiety. For example, the cross–link density is enhanced by the presence of an urethane functionality, inducing lateral stacking, while the presence of a bulky group on the UPy moiety will disrupt lateral interactions, reducing the cross–link density. The relation between dynamics and the environment (melt, bulk, and aqueous) of the UPy moiety were explored.

The results obtained and described in this Thesis will be useful in the development of new supramolecular polymers that need to fulfill specific demands concerning material properties.

Chapter 1

16

1.7 Outline of this Thesis

In contrast to covalent chemistry, supramolecular chemistry allows for the creation of (segmented) block copolymers by simple mixing of different macromonomers equipped with associating moieties on the chain ends resulting in the spontaneous assembly of high molecular weight polymers. In Chapter 2, the enzymatic ring–opening polymerizations of homopolymers and random copolymers of polycaprolactone and polyvalerolactone are described. Both the homopolymers and the random copolymers are functionalized with the ureidopyrimidinone moiety. A segmented supramolecular copolymer is obtained after mixing of two different homopolymers in different ratios. Supramolecular random block copolymers are formed via hydrogen bonding between the UPy moieties of the functionalized random copolymers. The material properties can be tuned by varying the ratio of the different polymer components in the polymeric material.

The dynamics of supramolecular linear polyester melts is investigated by oscillatory shear experiments in Chapter 3. The influence on the dynamics of the functional group between the polycaprolactone backbone and the UPy moiety is highlighted in this Chapter. The presence of additional hydrogen bonding interactions will strengthen the hydrogen bonding array with an increase in cross–link density and resulting in completely different rheological behavior.

In Chapter 4, the degree of polymerization of linear supramolecular polymers is tuned by the use of chain stoppers to obtain information about the relation between the degree of polymerization and material parameters such as zero–shear viscosity and terminal relaxation time. While in Chapters 3 & 4 the emphasis is on linear polymers, in Chapter 5 the role of polymer architecture on the dynamic mechanical properties is investigated.

The temperatures of phase transitions of UPy polymers are determined by the extend of the specific interactions between the polymer backbone and the UPy end groups. Increase in cross–link density by lateral interactions leads to shifts in phase transition temperatures. In Chapter 6, these phase transitions are investigated in detail by fluorescence spectroscopy using a supramolecular fluorescent probe.

From the macroscopic and microscopic properties of UPy functionalized polymers as discussed in Chapters 2–5, the focus is shifted to the behavior of UPy moieties in an aqueous environment in Chapter 7. In water, the solvent molecules strongly compete for hydrogen bond donor and acceptor sites with synthetic molecules and the formation of UPy dimers is thereby hampered. Hydrophobic microenvironments such as micelles can be used to avoid interaction with water molecules. Therefore, these microenvironments are able to accommodate the formation of hydrogen bonded supramolecular polymers and networks in water. UPy dimerization via hydrogen bonds in micelles is investigated by UV–vis titrations in the presence of a naphthyridine, a hydrogen bonding moiety complementary to the UPy moiety.

Dynamic Material Properties of Supramolecular Polymers

17

In Chapter 8 the association between telechelic polymers and fifth generation dendrimers equipped with adamantyl groups on the periphery in chloroform is discussed. Complexation between the dendrimer and telechelic polymer leads to the formation of single entities structures at low concentrations, while at high concentrations a transient network is developed. The results of dynamic light scattering experiments over a broad concentration range are described and reveal the characteristic features of the formed transient network.

Chapter 1

18

1.8 References

1. Wisniak, J. Chem. Educator 2000, 5, (6), 343–350. 2. Carothers, W. H. J. Am. Chem. Soc. 1929, 51, (8), 2548–2559. 3. Zimmerman, N.; Moore, J. S.; Zimmerman, S. C. Chem. Ind. 1998, 604–610. 4. Ciferri, A., Supramolecular Polymers. 2 ed.; CRC Press, Taylor & Francis Group: Boca Raton, 2005. 5. Sivakova, S.; Rowan, S. J. Chem. Soc. Rev. 2005, 34, 9–21. 6. See these recent reviews for detailed description of multiple hydrogen bonding arrays (and references therein): a) Zimmerman, S. C.; Corbin, P. S., Heteroaromatic Modules for Self–Assembly Using Multiple Hydrogen Bonds. Springer–Verlag: Berlin, 2000; Vol. 96, pp 63–94. b) Brunsveld, L.; Folmer, B. J. B.; Meijer, E. W.; Sijbesma, R. P. Chem. Rev. 2001, 101, (12), 4071–4097. c) Wilson, A. J. Soft Matter 2007, 3, 409 – 425. d) Binder, W. H.; Zirbs, R. Adv. Polym. Sci. 2007, 207, 1–78. e) Bouteiller, L. Adv. Polym. Sci. 2007, 207, 79–112. 7. a) Fouquey, C.; Lehn, J. M.; Levelut, A. M. Adv. Mater. 1990, 2, (5), 254–257. b) Gulik–Krzywicki, T.; Fouquey, C.; Lehn, J.–M. PNAS 1993, 90, 163–167. c) Brienne, M.–J.; Gabard, J.; Lehn, J.–M.; Stibor, I. J. Chem. Soc., Chem. Commun. 1989, 24, 1868–1870. 8. Murray, T. J.; Zimmerman, S. C. J. Am. Chem. Soc. 1992, 114, 4010–4011. 9. Lange, R. F. M.; Meijer, E. W. Macromolecules 1995, 28, (3), 782–783. 10. a) Asanuma, H.; Ban, T.; Gotoh, S.; Hishiya, T.; Komiyama, M. Macromolecules 1998, 31, (2), 371– 377. b) Asanuma, H.; Ban, T.; Gotoh, S.; Hishiya, T.; Komiyama, M. Supramol. Sci. 1998, 5, (3 – 4), 405–410. 11. Würthner, F.; Thalacker, C.; Sautter, A. Adv. Mater. 1999, 11, (9), 754–758. 12. a) Yamauchi, K.; Lizotte, J. R.; Hercules, D. M.; Vergne, M. J.; Long, T. E. J. Am. Chem. Soc. 2002, 124, (29), 8599–8604. b) Yamauchi, K.; Lizotte, J. R.; Long, T. E. Macromolecules 2002, 35, (23), 8745–8750. 13. a) Kunz, M. J.; Hayn, G.; Saf, R.; Binder, W. H. J. Polym. Sci., Part A: Polym. Chem. 2004, 42, 661– 674. b) Binder, W. H.; Kunz, M. J.; Ingolic, E. J. Polym. Sci., Part A: Polym. Chem. 2004, 42, 162–172. 14. a) Sivakova, S.; Wu, J.; Campo, C. J.; Mather, P. T.; Rowan, S. J. Chem. Eur. J. 2006, 12, 446–456. b) Sivakova, S.; Rowan, S. J. Chem. Comm. 2003, 2428–2429. 15. Corbin, P. S.; Zimmerman, S. C. J. Am. Chem. Soc. 1998, 120, (37), 9710–9711. 16. Sessler, J. L.; Wang, R. Angew. Chem. Int. Ed. 1998, 37, (12), 1726–1729. 17. Lüning, U.; Kühl, C. Tetrahedron Lett. 1998, 39, 5735–5738. 18. Gong, B.; Yan, Y.; Zeng, H.; Skrzypczak–Jankunn, E.; Kim, Y. W.; Zhu, J.; Ickes, H. J. Am. Chem. Soc. 1999, 121, (23), 5607–5608. 19. Hirschberg, J. H. K. K.; Brunsveld, L.; Ramzi, A.; Vekemans, J. A. J. M.; Sijbesma, R. P.; Meijer, E. W. Nature 2000, 407, 167–170. 20. a) Sijbesma, R. P.; Beijer, F. H.; Brunsveld, L.; Folmer, B. J. B.; Hirschberg, J. H. K. K.; Lange, R. F. M.; Lowe, J. K. L.; Meijer, E. W. Science 1997, 278, (5343), 1601–1604. b) Beijer, F. H.; Sijbesma, R. P.; Kooijman, H.; Spek, A. L.; Meijer, E. W. J. Am. Chem. Soc. 1998, 120, (27), 6761–6769. c) Beijer, F. H.; Kooijman, H.; Spek, A. L.; Sijbesma, R. P.; Meijer, E. W. Angew. Chem. Int. Ed. 1998, 37, (1/2), 75–78. 21. a) Zirbs, R.; Kienberger, F.; Hinterdorfer, P.; Binder, W. H. Langmuir 2005, 21, (18), 8414–8421. b) Binder, W. H.; Kluger, C.; Josipovic, M.; Straif, C. J.; Friedbacher, G. Macromolecules 2006, 39, (23), 8092–8101. 22. Kolomiets, E.; Buhler, E.; Candau, S. J.; Lehn, J.–M. Macromolecules 2006, 39, (3), 1173–1181. 23. Corbin, P. S.; Zimmerman, S. C.; Thiessen, P. A.; Hawryluk, N. A.; Murray, T. J. J. Am. Chem. Soc. 2001, 123, (43), 10475–10488. 24. a) Bialecki, J. B.; Yuan, L.–H.; Gong, B. Tetrahedron 2007, 63, 5460–5469. b) Li, M.; Yamato, K.; Ferguson, J. S.; Gong, B. J. Am. Chem. Soc. 2006, 128, (39), 12628–12629. 25. Söntjens, S. H. M.; Sijbesma, R. P.; van Genderen, M. H. P.; Meijer, E. W. J. Am. Chem. Soc. 2000, 122, (31), 7487–7493.

Dynamic Material Properties of Supramolecular Polymers

19

26. Lange, R. F. M.; van Gurp, M.; Meijer, E. W. J. Polym. Sci., Part A: Polym. Chem. 1999, 37, 3657– 3670. 27. a) Folmer, B. J. B.; Sijbesma, R. P.; Versteegen, R. M.; van der Rijt, J. A. J.; Meijer, E. W. Adv. Mater. 2000, 12, (12), 874–878. b) Keizer, H. M.; van Kessel, R.; Sijbesma, R. P.; Meijer, E. W. Polymer 2003, 44, (19), 5505–5511. c) Kautz, H.; van Beek, D. J. M.; Sijbesma, R. P.; Meijer, E. W. Macromolecules 2006, 39, (13), 4265–4267. 28. Keizer, H. M.; Sijbesma, R. P.; Jansen, J. F. G. A.; Pasternack, G.; Meijer, E. W. Macromolecules 2003, 36, (15), 5602–5606. 29. Dankers, P. Y. W.; Zhang, Z.; Wisse, E.; Grijpma, D. W.; Sijbesma, R. P.; Feijen, J.; Meijer, E. W. Macromolecules 2006, 39, (25), 8763–8771. 30. a) Dankers, P. Y. W.; Harmsen, M. C.; Brouwer, L. A.; van Luyn, M. J. A.; Meijer, E. W. Nat. Mater. 2005, 4, (7), 568–574. b) van Beek, D. J. M.; Spiering, A. J. H.; Peters, G. W. M.; te Nijenhuis, K.; Sijbesma, R. P. Macromolecules 2007, Accepted. c) van Beek, D. J. M.; Gillissen, M. A. J.; van As, B. A. C.; Palmans, A. R. A.; Sijbesma, R. P. Macromolecules 2007, 40¸ (17), 6340–6348. 31. Yamauchi, K.; Kanomata, A.; Inoue, T.; Long, T. E. Macromolecules 2004, 37, (10), 3519–3522. 32. a) Yamauchi, K.; Lizotte, J. R.; Long, T. E. Macromolecules 2003, 36, (4), 1083–1088. b) Park, T.; Zimmerman, S. C. J. Am. Chem. Soc. 2006, 128, (44), 14236–14237. c) McKee, M. G.; Elkins, C. L.; Park, T.; Long, T. E. Macromolecules 2005, 38, (14), 6015–6023. d) Elkins, C. L.; Park, T.; McKee, M. G.; Long, T. E. J. Polym. Sci., Part A: Polym. Chem. 2005, 43, 4618–4631. 33. Rieth, L. R.; Eaton, R. F.; Coates, G. W. Angew. Chem. Int. Ed. 2001, 40, (11), 2153–2156. 34. Hirschberg, J. H. K. K.; Beijer, F. H.; van Aert, H. A.; Magusin, P. C. M. M.; Sijbesma, R. P.; Meijer, E. W. Macromolecules 1999, 32, (8), 2696–2705. 35. a) Rubinstein, M.; Semenov, A. N. Macromolecules 2001, 34, (4), 1058–1068. b) Semenov, A. N.; Rubinstein, M. Macromolecules 1998, 31, (4), 1373–1385. c) Rubinstein, M.; Semenov, A. N. Macromolecules 1998, 31, (4), 1386–1397. d) Semenov, A. N.; Rubinstein, M. Eur. Phys. J. B 1998, 1, 87 –94. e) Semenov, A. N.; Rubinstein, M. Macromolecules 2002, 35, (12), 4821–4837. 36. a) Wientjes, R. H. W.; Duits, M. H. G.; Jongschaap, R. J. J.; Mellema, J. Macromolecules 2000, 33, (26), 9594–9605. b) Jongschaap, R. J. J.; Wientjes, R. H. W.; Duits, M. H. G.; Mellema, J. Macromolecules 2001, 34, (4), 1031–1038. c) Wientjes, R. H. W.; Jongschaap, R. J. J.; Duits, M. H. G.; Mellema, J. J. Rheol. 1999, 43, (2), 375–391. d) Wientjes, R. H. W.; Duits, M. H. G.; Bakker, J. W. P.; Jongschaap, R. J. J.; Mellema, J. Macromolecules 2001, 34, (17), 6014–6023. 37. a) Cates, M. E. Macromolecules 1987, 20, (9), 2289–2296. b) Cates, M. E. J. Phys. Chem. 1990, 94, (1), 371–375. 38. Cates, M. E.; Candau, S. J. J. Phys. Condens. Matter 1990, 2, 6869–6892. 39. Leibler, L.; Rubinstein, M.; Colby, R. H. Macromolecules 1991, 24, (16), 4701–4707. 40. a) González, A. E. Polymer 1983, 24, 77 – 80. b) González, A. E. Polymer 1984, 25, 1469–1474. 41. a) Stadler, R.; de Lucca Freitas, L. L. Colloid. Polym. Sci. 1986, 264, (9), 773–778. b) Hilger, C.; Stadler, R.; de Lucca Freitas, L. L. Polymer 1990, 31, 818–823. c) de Lucca Freitas, L. L.; Stadler, R. Macromolecules 1987, 20, (10), 2478–2485. 42. a) Müller, M.; Seidel, U.; Stadler, R. Polymer 1995, 36, (16), 3143–3150. b) Müller, M.; Kremer, F.; Stadler, R.; Fischer, E. W.; Seidel, U. Colloid. Polym. Sci. 1995, 273, 38–46. c) Seidel, U.; Stadler, R.; Fuller, G. G. Macromolecules 1994, 27, (8), 2066–2072. 43. a) te Nijenhuis, K. Makromol. Chem. 1991, 192, 603–616. b) te Nijenhuis, K. Makromol. Chem. Macromol. Symp. 1991, 45, 117–126. c) te Nijenhuis, K. Polym. Gels Networks 1993, 1, 185–198. d) te Nijenhuis, K. Polym. Gels Networks 1993, 1, 199–210. e) Franse, M. W. C. P.; te Nijenhuis, K. J. Mol. Struc. 2000, 554, 1–10. 44. a) Flory, P. J. J. Am. Chem. Soc. 1941, 63, 3083–3090. b) Flory, P. J. J. Am. Chem. Soc. 1941, 63, 3091–3096. c) Flory, P. J. J. Am. Chem. Soc. 1941, 63, 3096–3100. d) Flory, P. J. J. Am. Chem. Soc. 1947, 69, 30–35.

Chapter 1

20

45. a) Stockmayer, W. H. J. Chem. Phys. 1943, 11, (2), 45–55. b) Stockmayer, W. H. J. Chem. Phys. 1944, 12, (4), 125–131. 46. te Nijenhuis, K. Polym. Bull. 2007, 58, 27–42. 47. te Nijenhuis, K. Polym. Gels Networks 1996, 4, 415–433. 48. a) Franse, M. W. C. P.; te Nijenhuis, K.; Picken, S. J. Rheol. Acta 2003, 42, 443–453. b) Franse, M. W. C. P.; te Nijenhuis, K.; Groenewold, J.; Picken, S. J. Macromolecules 2004, 37, (20), 7839–7845. 49. a) Boileau, S.; Bouteiller, L.; Lauprêtre, F.; Lortie, F. New J. Chem. 2000, 24, 845–848. b) Lortie, F.; Boileau, S.; Bouteiller, L.; Chassenieux, C.; Demé, B.; Decouret, G.; Jalabert, M.; Lauprêtre, F.; Terech, P. Langmuir 2002, 18, (19), 7218–7222. c) Simic, V.; Bouteiller, L.; Jalabert, M. J. Am. Chem. Soc. 2003, 125, (43), 13148–13154. d) Lortie, F.; Boileau, S.; Bouteiller, L. Chem. Eur. J. 2003, 9, 3008–3014. 50. a) Knoben, W.; Besseling, N. A. M.; Bouteiller, L.; Cohen Stuart, M. A. Phys. Chem. Chem. Phys. 2005, 7, 2390–2398. b) Knoben, W.; Besseling, N. A. M.; Cohen Stuart, M. A. Macromolecules 2006, 39, (7), 2643–2653. c) Knoben, W.; Besseling, N. A. M.; Cohen Stuart, M. A. J. Chem. Phys. 2007, 126, 024907: 1–9. 51. a) Rowan, S. J.; Suwanmala, P.; Sivakova, S. J. Polym. Sci., Part A: Polym. Chem. 2003, 41, 3589– 3596. a) Sivakova, S.; Bohnsack, D. A.; Mackay, M. E.; Suwanmala, P.; Rowan, S. J. J. Am. Chem. Soc. 2005, 127, (51), 18202–18211. 52. a) Winter, H. H. Polym. Eng. Sci. 1987, 27, (22), 1698–1702. b) Vallés, E. M.; Carella, J. M.; Winter, H. H.; Baumgaertel, M. Rheol. Acta 1990, 29, 535–542. 53. a) Yount, W. C.; Loveless, D. M.; Craig, S. L. Angew. Chem. Int. Ed. 2005, 44, 2746–2748. b) Yount, W. C.; Loveless, D. M.; Craig, S. L. J. Am. Chem. Soc. 2005, 127, (41), 14488–14496. 54. a) Loveless, D. M.; Jeon, S. L.; Craig, S. L. J. Am. Chem. Soc. 2005, 38, (24), 10171–10177. b) Loveless, D. M.; Jeon, S. L.; Craig, S. L. J. Mater. Chem. 2007, 17, 56–61.

22 Supramolecular Copolyesters with Tunable Properties

ABSTRACT. The effect of chain structure (supramolecular random copolymer vs supramolecular segmented copolymer) on material properties of supramolecular polymers was studied, using polyesters end–functionalized with quadruple hydrogen bonding ureidopyrimidinone (UPy) units. Mixing of miscible UPy homopolymers led to supramolecular segmented copolymers while functionalized random copolymer diols resulted in supramolecular random copolymers. The (co)polymers were prepared by (co)polymerization of ε–CL and δ–VL using Novozym 435, followed by end– functionalization with UPy. Thermal analysis of the functionalized (co)polymers showed two melting transitions. With variable temperature IR, the lower transition was attributed to the melting of the polyester part, while the higher transition corresponded to melting of UPy moieties. The materials can therefore be considered as supramolecular thermoplastic elastomers with a hard phase of microphase separated UPy dimers, giving mechanical strength to the material. Mixing of UPy functionalized homopolymers gave better control over the mechanical properties than UPy functionalized copolymers as a correlation was found between the Young’s modulus and the fraction of δ–VL polymer in the material.

This work has been published: D. J. M. van Beek, Martijn A. J. Gillissen, Bart A. C. van As, Anja R. A. Palmans, Rint P. Sijbesma; Macromolecules, 2007, 40, (17), 6340–6348.

Chapter 2

22

2.1 Introduction

Polyesters are commonly used in biomaterials because of their biocompatibility, biodegradability and tunable mechanical properties.1–4 Poly(ε–caprolactone) (PCL)4 is a particularly well studied aliphatic polyester receiving much attention due to its potential use in biomedical and pharmaceutical applications. However, due to its high crystallinity,5 PCL degrades slowly,2–4,6 which is a major drawback for a successful application as a biomaterial. Copolymerization1,4,7 of ε–caprolactone (CL) with other monomers, as well as blending8–13 with other polymers substantially improved the degradation rates. For example, copolymers of CL with ethylene glycol4,14, D– and L–lactide4,10, ω–pentadecalactone,15,16 1,5–dioxepan–2–one4,5, 1,4–dioxepan–2–one17 or δ–valerolactone7,18–20 all showed enhanced hydrolysis rates due to a lower crystallinity in the copolymer. Moreover, the thermal and mechanical properties of the copolymers could be tuned by varying the copolymer composition.19 However, if small modifications of the copolymer are required such as a different comonomer ratio or the incorporation of a slightly more polar monomer, the complete polymerization procedure needs to be repeated. Moreover, minor chemical modifications may require a radical change of the synthetic procedure such as the need of protection and deprotection steps. Here, we report a method based on reversible supramolecular interactions, which allows for a simple modification of the copolymer without the need of elaborate synthesis.

Figure 2.1: Supramolecular random copolymers (first copolymerize then functionalize) and supramolecular segmented copolymers (mixing of homopolymers). Supramolecular polymerizations based on quadruple hydrogen bonding of the ureidopyrimidinone (UPy) moiety (Figure 2.1, Scheme 2.1) was developed and elaborately studied in our laboratory.21,22 Upon end–functionalization with the UPy moiety of low molecular weight polymers with low tensile strength, the macromonomers polymerize via hydrogen bonding to form elastic materials with useful mechanical properties.22–24 Dramatic changes in mechanical properties were found for a wide range of polymers such as polyethers, polycarbonates and polyesters. Microphase separation between the UPy moiety and the polymer backbone was shown to contribute significantly to the improvement of mechanical properties in, e.g. UPy functionalized poly(ethylene

Supramolecular Copolyesters with Tunable Properties

23

butylene).22,25,26 Supramolecular polymerization allows the formation of segmented copolymers, which has been reported for non–miscible diblock copolymers27–29 and non–miscible segmented copolymers28,30–33. In this Chapter we will focus on the preparation of miscible segmented copolymers by simply mixing of functionalized homopolymers.24

Two methods of preparation of supramolecular copolymers as depicted in Figure 2.1 are compared in this study. First, low molecular weight random copolymers are discussed, which after functionalization with UPy moieties result in supramolecular random copolymers. These are compared with supramolecular segmented copolymers, which are obtained by mixing of functionalized homopolymers. Homo– and copolymers of ε–caprolactone (CL) and δ–valerolactone (VL) were selected to prepare the supramolecular copolymers because of vast knowledge of their chemical structure, thermal properties, and their potential as biomaterials.4,7,18–20

Ring–opening polymerizations of lactones can be performed chemically4,34–36 or enzymatically.37–40 All polymerizations presented here were performed enzymatically, using the immobilized Lipase B from Candida antarctica (Novozym 435) as it is then possible to work under mild conditions. Moreover, it is known that the copolymerization of two lactones using Novozym 435 results in random copolymers.16,41–43 In the present work, the possibility to tune both thermal and mechanical properties of copolymers by using supramolecular chemistry is investigated, and the method which leads to materials with the best tunable material properties is identified.

2.2 Synthesis

The syntheses of the polymers discussed in this Chapter are based on enzymatic ring–opening polymerizations. The ring–opening polymerizations of high molecular weight polyesters and of low molecular weight polyester diols will first be discussed. The polyester diols are subsequently functionalized with UPy moieties.

2.2.1 Synthesis of High Molecular Weight pCL, pVL and p(CL–co 50% VL)

High molecular weight polyesters (pCL, pVL, and p(CL–co 50% VL)) were synthesized as reference materials to compare the thermal and mechanical properties of supramolecular (co)polymers with "traditional" polymers. The polyesters were obtained by enzymatic ring opening polymerization of the corresponding lactones employing Novozym 435 as the catalyst. To reduce the amount of water, which acts as initiator and limits the molecular weight, all reagents and Novozym 435 were rigorously dried as previously described.37 Although the drying procedure significantly slows down the reaction rate, polyesters of good molecular weight are accessible in this way.40 For example, pCL was obtained with a Mn of 82.6 kg/mol and a PD of 2.0 with a reaction time of 6 days. In contrast to pCL and pVL, which were white powders, p(CL–co 50% VL) was obtained as an oil.

Chapter 2

24

2.2.2 Synthesis of Supramolecular (Co)Polymers

Low molecular weight prepolymers, poly(caprolactone) diol (pCL–diol), poly(valerolactone) diol (pVL–diol), and random copolymers of CL and VL with varying VL content (entries 8, 10, 12), were obtained by modification of the method to obtain high molecular weight polyesters (Scheme 2.1). A bifunctional initiator, 1,6–hexanediol, was used in a 1:20 I/M ratio, ensuring that all polymer chains have two diol end groups.

Scheme 2.1: Ring opening polymerization of lactones catalyzed by Novozyme 435. Using this procedure, polymer diols with a Mn of approximately 2 kg/mol were obtained as determined by NMR and GPC. From 1H NMR, the ratio of CL (n) to VL (m) was determined by comparison of the integrated peak areas of the signals at δ = 1.44 ppm and δ = 2.37 ppm (See Experimental Section). The observed and feed ratios are in good agreement for all polymer diols (Table 2.1).

Table 2.1: Feed ratio and Observed Molar Ratio of CL (n) and VL (m) in Random Copolymers before and after Functionalization with the UPy Moiety obtained from 1H NMR.

Random Copolymer Molar Feed Ratio CL / VL

(n:m)

Observed Molar Ratio CL / VL

(n:m) p(CL–co 20% VL)diol 80 : 20 81 : 19 p(CL–co 14% VL)diUPy 81 : 19 86 : 14 p(CL–co 50% VL)diol 50 : 50 52 : 48 p(CL–co 33% VL)diUPy 52 : 48 67 : 33 p(CL–co 80% VL)diol 20 : 80 18 : 82 p(CL–co 80% VL)diUPy 18 : 82 20 : 80

All prepolymers were subsequently functionalized with the ureidopyrimidinone (UPy) group by reaction with the UPy–NCO synthon (Figure 2.2) according to the literature procedure.23 In the preparation of p(CL–co 33% VL)diUPy, the excess of UPy–NCO was removed by reaction with a aminomethyl polystyrene resin after which precipitation in heptane followed. Table 2.1 summarizes the CL/VL ratio in the UPy functionalized

Supramolecular Copolyesters with Tunable Properties

25

polymers after workup. For both p(CL–co 14% VL)diUPy and p(CL–co 80% VL)diUPy the CL/VL ratio slightly increased upon functionalization. On the other hand, for p(CL–co 33% VL)diUPy the CL/VL ratio significantly and reproducibly increased from 50/50 to 67/33. Functionalization of the diols was performed in the presence of dibutyl tin dilaurate (DBTDL), a Lewis acid which is known to depolymerize pVL segments.19, 45 In order to investigate the influence of DBTDL on the copolymer composition, samples of pVL–diol, pCL–diol, and p(CL–co 50% VL)diol were refluxed in chloroform with and without DBTDL for 72 h and monitored with GPC (Figure 2.3). All supramolecular (co)polymers (entries 5, 7, 9, 11, 13) were obtained in yields varying from 34 to 77% (Table 2.2).

N

NN

NO

OO

O

O

O

NN

N

O

O

NNO

HO

O

N O

H

n

n

H H

H

H

H H

N

NN

NO

O

N O

H

N

NN

ONNO

HO

O OO O

O

Om

m

HH

H

H

HH

N

NN

ON N O

HO

N

NN

ONNO

HO

O OO

O

OO

O

OO

On

m

n

m

H H

H

H

H H

N

N NH

NH

O

HO

N C O

pCLdiUPy

pVLdiUPy

UPy-NCO

p(CL-co 14% VL)diUPy / p(CL-co 33% VL)diUPy / p(CL-co 80% VL)diUPy

Figure 2.2: Functionalized supramolecular (co)polymers. The retention time of DBTDL is shorter than the retention time of the VL monomer (Figure 2.3A). No signals of the VL monomer were observed for the pVL–diol (Figure 2.3B) in chloroform without DBTDL. Besides the signal of DBTDL and the polymer itself no extra signals could be observed for pCL–diol (Figure 2.3C). For the combination of DBTDL and chloroform, two signals could be observed for pVL–diol after 16 h; the first one is attributed to DBTDL, while the second one is attributed to the VL monomer (Figure 2.3D). A similar spectrum was obtained for p(CL–co 50% VL)diol (Figure 2.3E). Therefore, the shift in CL/VL ratio after functionalization with UPy moieties is attributed to the action of DBTDL.

Chapter 2

26

A6 8 10 12 14

0

1x103

2x103

3x103

4x103

5x103A

.U. [

-]

Time [min]

DBTDL VL

B

6 8 10 12 14

0

1x103

2x103

3x103

A.U

. [-]

Time [min]

t = 0 h t = 16 h t = 72 h

C6 8 10 12 14

0

1x103

2x103

3x103

4x103

DBTDL

A.U

. [-]

Time [min]

t = 0 t = 16 h t = 72 h

D6 8 10 12 14

0

1x103

2x103

3x103

VLDBTDL

A.U

. [-]

Time [min]

t = 0 h t = 16 h t = 72 h

E

6 8 10 12 14

0

1x103

2x103

3x103

VLDBTDL

A.U

. [-]

Time [min]

t = 0 h t = 16 h t = 72 h

Figure 2.3: (A) GPC traces of DBTDL and VL monomer in chloroform, GPC traces after t = 0 (no DBTDL added), 16 and 72 h after addition of DBTDL and refluxing in chloroform for (B) pVL–diol without addition of DBTDL, (C) pCL–diol, (D) pVL–diol, and (E) p(CL–co 50% VL)diol.

2.3 Thermal Properties

The thermal properties of the polymers were investigated using differential scanning calorimetry (DSC) and variable temperature attenuated total reflection infrared spectroscopy (VT–IR). The melting temperatures around 55 °C and glass transitions at –60 °C observed for pCL (entry 1) and pVL (entry 2) (Figure 2.4A), are in good agreement with values previously reported in literature.5,12,18,36 In IR (Figure 2.4B), a shift from 1724 cm–1 (crystalline νs(C=O)) to 1737 cm–1 (amorphous νs(C=O)) was observed around 56 °C, indicating melting of PCL12,13,46,47 and PVL.12 The signal at 1724 cm–1 decreased for both polymers, while a new and much stronger signal appeared at 1737 cm–1, represented in Figure 2.2B. Melting was furthermore confirmed by the disappearance of the stretch

Supramolecular Copolyesters with Tunable Properties

27

vibration at 1293 cm–1, corresponding to the C–O and C–C stretch vibrations of the crystalline fraction of the material. (Figure 2.4B, inset) The thermal behavior of p(CL–co–VL) (entry 3) differs distinctly from the homopolymers pCL and pVL as the melting temperature is approximately 40 °C lower (Table 2.2, Figure 2.4A). A lowering in the melting temperature of copolymers of VL and CL has been observed previously by Storey et al.19 and Faÿ et al.18. Furthermore, Gruvegård5 and coworkers suggested that suppression of melting points observed for copolymers originates from a mismatch in the sequence length of the crystal. The conformations in the crystals of CL and VL are both planar zig–zag, however, the sequence length of CL (17.0 Å) is slightly bigger than for VL (15.7 Å), a plausible reason for a suppressed melting point. The work of Furuhashi et al.48 supports this latter work as they concluded that due to the odd number of methylene units in CL the carbonyl ester groups are distributed regularly on both sides of the chains, in contrary to even–numbered polymers like VL.

Table 2.2: Composition, Molecular Weight and Thermal Properties of HMW (co)Polymers, (co)Polymer Diols, Supramolecular (Co)Polymers, and Mixtures of Supramolecular Polymers. Entry Polymer CL

[%] VL [%]

Yield[%]

Mn* [kg/mol]

PDI* [–]

Tg [°C]

Tm [°C]

ΔH [J/g]

1 pCL 100 0 34 82.6 2.0 –58 53 62 2 pVL 0 100 50 22.3 6.3 –50 55 61 3 p(VL–co–CL) 50 50 66 6.9 4.3 –60 15 64 4 pCL–diol 100 0 94 3.4 1.5 –63 46 95 5 pCLdiUPy 100 0 60 3.8 1.6 –55 39 / 63 32a

6 pVL–diol 0 100 74 2.3 1.8 –61 42 75 7 pVLdiUPy 0 100 77 3.0 1.7 –50 34 / 77 8 / 8 8 p(CL–co 20% VL)–diol 81 19 96 2.9 1.5 –62 28 / 33 57a 9 p(CL–co 14% VL)diUPy 86 14 45 4.6 1.5 –55 22 / 52 24 / 610 p(CL–co 50% VL)–diol 50 50 86 2.9 1.7 –88b 6 53 11 p(CL–co 33% VL)diUPy 67 33 79 2.7 1.8 –56 73 6 12 p(CL–co 80% VL)–diol 18 82 98 3.0 1.5 –84 13 48 13 p(CL–co 80% VL)diUPy 20 80 34 8.9 1.3 –49 23 / 34 49a 14 Mix (pCLdiUPy /

20% pVLdiUPy) 80 20 n.a. n.a. n.a. –55 38 / 65 21a

15 Mix (pCLdiUPy / 50% pVLdiUPy)

50 50 n.a. n.a. n.a. –54 30 / 70 13a

16 Mix (pCLdiUPy / 77% pVLdiUPy)

23 77 n.a. n.a. n.a. –52 34 / 74 15a

* determined by GPC in THF, n.a.: not applicable, a: total for 2 transitions, b: determined from second heating run. The homopolymers pCL–diol and pVL–diol display single melting points at 46 and 42 °C (Figure 2.4C), respectively. This is substantially below the melting temperatures found for pCL and pVL of 53 and 55 °C, respectively. The differences are attributed to the significantly lower molecular weights of pCL–diol and pVL–diol. Furthermore, the random copolymer diols (entries 8, 10, 12) also display single melting temperatures, indicating that no macrophase separation between VL and CL occurred. However, an additional melting point at 28 °C is observed for p(CL–co 20% VL)–diol, but the origin of

Chapter 2

28

this transition is not clear. The random copolymer diols all show lower melting temperatures than the corresponding homopolymer diols as a consequence of the random character of the copolymers (vide supra). Glass transition temperatures were observed below –60 °C for all prepolymer diols (Table 2.2). All UPy functionalized materials showed similar glass transition temperatures around –55 °C, but significant differences in the melting points were observed (Table 2.2). Furthermore, the polymers changed from oils or waxy solids into elastic solids upon functionalization.22,23

A

0 20 40 60 80 10020

30

40

50

60

70

pVL

pCL

p(VL-co-CL)

Hea

t flo

w [m

W] (

endo

up)

Temperature [ºC] B

25 35 45 55 65 750.1

0.2

0.3

0.4

0.5

0.6

1250 1275 1300 1325 1350

70 - 80 ºC

25 ºC

60 ºC65 ºC

Frequency [cm-1]

1293 cm-1

pCL

Inte

nsity

[A.U

.]

Temperature [ºC]

pCL pVL

C

0 20 40 60 80 1000

10

20

30

40

50

Mel

ting

Tem

pera

ture

[ºC

]

Fraction δ-VL [%] D0 20 40 60 80 100

0

10

20

30

40

50

60

70

80

Mel

t Tem

pera

ture

[ºC

]

Fraction δ-VL[%]

random & segmented copolymers

E

20 30 40 50 60 70 800.1

0.2

0.3

0.4

0.5

1710 1720 1730 1740

55 - 80 ºC50 ºC25 - 45 ºC

Frequency [cm-1]

1293 cm-1

1668 cm-1

Inte

nsity

[A.U

.]

Temperature [ºC] F

-100 -75 -50 -25 0 25 50 75 100 12510

15

20

25

30

35

40

45

Hea

t flo

w [m

W] (e

ndo

up)

Temperature [ºC]

p(CL-co 33% VL)diUPy Mix(pCLdiUPy / 50 %pVLdiUPy)

Figure 2.4: (A) DSC traces of pCL, pVL and p(VL–co–CL). (B) ATR–IR of pCL (triangles) and pVL (squares) at 1737 cm–1, inset: VT–IR of pCL at 1293 cm–1. (C) Melting temperatures of prepolymers diols vs δ–VL fraction. (D) Melting temperatures of supramolecular random copolymers (squares) and supramolecular segmented copolymers (circles) vs δ–VL fraction. (E) VT–IR of pCLdiUPy vs temperature, inset: VT–IR at 1724 cm–1. (F) DSC traces of p(CL–co 33% VL)diUPy (black), Mix (pCLdiUPy / 50% pVLdiUPy) (gray).

Supramolecular Copolyesters with Tunable Properties

29

Besides p(CL–co 33% VL)diUPy, all polymers functionalized with UPy moieties displayed two melting temperatures. (Table 2.2 and Figure 2.4D) Variable temperature IR was performed to investigate the nature of the melting temperatures in the supramolecular polymers. Figure 2.4E shows the infrared spectrum for pCLdiUPy as a function of temperature, which is representative for the other polymers. The intensities of the absorption bands at 1724, 1668, and 1293 cm–1 were evaluated as a function of temperature. The bands at 1724 and 1293 cm–1 are typical for the crystalline C=O vibration of polyesters and their intensities are expected to decrease when pCL and pVL become amorphous.12,46,47 Around 45 °C a decrease in intensity at 1293 cm–1 and a shift from 1724 cm–1 to higher wavenumbers were indeed observed (Figure 2.2E and inset), corresponding to the lowest melt transition at 39 °C. The band at 1668 cm–1 originates from the UPy moiety, although we cannot assign it to a specific vibration.49,50 A closer look to the intensity at 1668 cm–1 showed a decrease around a temperature of 60 °C, corresponding to the highest melting transition at 63 °C. We conclude that the lowest melting temperatures can be attributed to the melting of the polyester backbone and the highest melting temperature can be attributed to melting of the UPy moiety. Therefore, microphase separation of the UPy moiety and polymer part is indeed the origin of the appearance of two melting peaks. Although functionalized with UPy moieties, the melting temperatures of the polyester backbone show a δ–VL fraction dependency, similar to the melting temperatures as observed for the (co)polymer diols (Figure 2.4C).

A

1250 1275 1300 1325 13500.0

0.1

0.2

0.3

0.4

0.5

Inte

nsity

[A.U

.]

Wavenumber [cm-1] B

1650 1660 1670 1680 1690 17000.0

0.1

0.2

0.3

0.4

0.5

1668 cm-1

Inte

nsity

[A.U

.]

Wavenumber [cm-1]

Figure 2.5: ATR–IR of p(CL–co 33% VL)diUPy at (A) 1293 cm–1 and (B) 1668 cm–1. Unexpectedly, p(CL–co 33% VL)diUPy displayed only one melting temperature at 73 °C. Infrared spectroscopy at room temperature (Figure 2.5A) showed the absence of a vibrational band at 1293 cm–1, indicating an amorphous polyester backbone at room temperature. This is in line with the maximum in melting point depression previously described by Storey19 et al. for a CL/VL ratio of 67/33. The characteristic vibration corresponding to the crystalline form of the UPy moiety at 1668 cm–1 (Figure 2.5B) is present at room temperature. The melting point observed at 73 °C is therefore attributed to melting of microphase–separated domains of UPy dimers. High melting points for UPy moieties have been observed before for polymers with an amorphous backbone. Thermal analysis of UPy telechelic (amorphous) poly(ethylene butylene)25 showed a single melting

Chapter 2

30

transition at 69 °C in the first heating run, attributed to the melting of the microphase separated UPy moieties. The slight differences in melting temperature of the UPy moieties can be explained by the enhanced UPy crystallite formation in an amorphous polymer matrix due to higher mobility of the UPy groups, resulting in larger crystallites with a higher melting temperature.

The melting transitions corresponding to melting of the UPy moieties observed for the supramolecular random copolymers do not show any relation with increasing VL ratio (Figure 2.4D). Whereas melting point suppression is absent in the supramolecular segmented copolymers, higher values for the melting transitions of the polymer backbones (entries 14, 15, 16) were observed, compared to the supramolecular random copolymers (Figure 2.4D). The presented supramolecular segmented copolymers show a constant melting temperature (34 °C) similar to the blended CL/VL oligomers studied by Storey19. In addition, a linear increase of the melting temperature of the UPy moiety from 63 °C to 77 °C with increasing VL content was observed.

2.3 Mechanical Properties

Functionalization of low molecular weight polymers with UPy groups leads to improved mechanical properties.22,23 The mechanical properties of the copolymers were studied by tensile testing and representative stress–strain curves are depicted in Figure 2.6 whereas the results are summarized in Table 2.3. The mechanical properties of the pure polymers, pCLdiUPy and pVLdiUPy, represent the two extremes. pCLdiUPy is slightly more brittle (Table 2.3), but is more flexible than pVLdiUPy as indicated by Young’s moduli of 28 and 45 MPa, respectively. As the maximum applied stress for pCLdiUPy (2.9 MPa) is lower than for pVLdiUPy (4.7 MPa), pCLdiUPy is also weaker. The supramolecular polymers differ in a similar manner from each other as the high molecular weight counter parts (entries 1 and 2). pCL is also more tough, more flexible but weaker than pVL (Table 2.3). Although the Young’s modulus seems to increase with increasing VL content, no relation between the stress–strain curves and the ratio of VL to CL in the copolymers can be observed for the random copolymers (Figure 2.6A). Because of limited availability, p(CL–co 80% VL)diUPy was only measured once and its results can not be interpret with great certainty. All tensile tests were preformed at room temperature and when a closer look is taken to Figure 2.2D, it becomes clear that the polymer backbones of the supramolecular random copolymers are all molten at room temperature. Remarkably, despite the molten polymer backbone, the materials have sufficient mechanical properties to undergo mechanical testing. Even p(CL–co 33% VL)diUPy displays good mechanical properties. The polymer properties of the supramolecular random copolymers are not influenced anymore by the composition of the polyester backbone but the material properties are mainly determined by the UPy dimers. Microphase separation of the polymer backbone and the UPy hydrogen bonding moieties (melting temperatures above 30 ºC) results in an elastomeric network, responsible for the strength and flexibility of the materials.

Supramolecular Copolyesters with Tunable Properties

31

A0 5 10 15 20 25 30 35

0

1

2

3

4

5

6

σ [M

Pa]

ε [%]

pCLdiUPy p(CL-co 14% VL)diUPy p(CL-co 33% VL)diUPy p(CL-co 80% VL)diUPy pVLdiUPy

B0 5 10 15 20 25 30 35

0

1

2

3

4

5

6 pCLdiUPy Mix (pCLdiUPy / 20% pVLdiUPy) Mix (pCLdiUPy / 50% pVLdiUPy) Mix (pCLdiUPy / 80% pVLdiUPy) pVLdiUPy

σ [M

Pa]

ε [%]

Figure 2.6: Representative stress–strain curves of (A) Supramolecular random copolymers and (B) Supramolecular segmented copolymers. In contrast to the stress–strain curves of the supramolecular random copolymers, the stress–strain curves of the supramolecular segmented copolymers were found to lie between the two extremes. Moreover, a correlation between the stress–strain curves and the ratio CL:VL was observed (Figure 2.6B). Shifting the ratio of CL:VL from 20:80 to 80:20 by increasing pVLdiUPy, leads to more brittle but stronger and stiffer materials (Figure 2.6B) as evidenced by the increasing Young’s modulus (Table 2.3). Apparently, the polyester mainly present in the mixture, determines the material properties of the polymer film.

Table 2.3: Mechanical Properties of Supramolecular Random Copolymers and Supramolecular Segmented Copolymers with Data Derived from Engineering σ–ε Curves.

Polymer E

[MPa] σmax

[MPa] εbreak [%]

pCLdiUPy 28.1 ± 0.6 2.9 ± 0.1 24.0 ± 1.9 pVLdiUPy 45.4 ± 1.2 4.7 ± 0.1 20.7 ± 1.1 p(CL–co 14% VL)diUPy 20.9 ± 0.5 2.4 ± 0.1 31.0 ± 1.4 p(CL–co 33% VL)diUPy 38.2 ± 4.2 4.4 ± 0.1 24.4 ± 1.2 p(CL–co 80% VL)diUPy* 29.0 1.5 12.8 Mix (pCLdiUPy / 20% pVLdiUPy) 33.4 ± 0.5 3.4 ± 0.1 22.9 ± 1.1 Mix (pCLdiUPy / 50% pVLdiUPy) 38.2 ± 4.2 4.4 ± 0.1 24.4 ± 1.2 Mix (pCLdiUPy / 77% pVLdiUPy) 39.7 ± 2.5 3.8 ± 0.2 17.8 ± 1.7 pCL 255 ± 25 15 ± 0.7 685 ± 112 pVL 241 ± 158 22 ± 5.2 792 ± 163

*n = 1 While tensile testing is performed at room temperature and allows for the determination of Young’s modulus, dynamic mechanical thermal analysis (DMTA) permits us to investigate the mechanical properties from below the glass transition temperature up to the melting point. Comparison of the properties of materials with similar CL/VL ratios is of particular interest, and therefore, DMTA was measured for p(CL–co 33% VL)diUPy

Chapter 2

32

and Mix (pCLdiUPy / 50% pVLdiUPy). Figure 2.7 depicts the dynamic moduli and tanδ of p(CL–co 33% VL)diUPy (Figure 2.7A) and Mix (pCLdiUPy / 50% pVLdiUPy) (Figure 2.7B) against temperature.

-80 -60 -40 -20 0 20 40 60 800

20

40

60

80

100

120

140

160

E'' [

MPa

]

Temperature [ºC]

10-1

100

101

102

103

104

E dyn

amic [M

Pa]

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

0

20

40

60

80

100

120

140

160

E'' [

MPa

]

Temperature [ºC]

10-1

100

101

102

103

104

E dyn

amic [M

Pa]

Figure 2.7: Dynamic mechanical analysis curves of (left) supramolecular random copolymer p(CL–co 33% VL)diUPy and (right) supramolecular segmented copolymer Mix (pCLdiUPy / 50% pVLdiUPy). For both polymers a gradual decrease in dynamic modulus (Edynamic) is observed from glass to rubber and finally to melt. The dynamic moduli are very similar, indicating that below the glass transition temperature the polymers have similar properties. Glass transition temperatures of –55.7 and –55.4 ºC for the random copolymer and segmented copolymer, respectively, were determined from the maxima in loss modulus (E”). These values correspond very well to the glass transition temperatures determined by DSC (Table 2.2). After the glass transition, a rubber plateau is reached for p(CL–co 33% VL)diUPy at a value of 40 MPa between –20 and 40 °C. The sample yielded around the melting temperature of the UPy moiety (73 °C). A higher value for the dynamic modulus is observed after the glass transition for Mix (pCLdiUPy / 50% pVLdiUPy). However, the dynamic modulus decreased with increasing temperature and no rubber plateau can be observed. Melting of the polymer backbone is visible by the drop in modulus around 40 ºC. The sample of Mix (pCLdiUPy / 50% pVLdiUPy) began to yield after reaching the melting temperature of the UPy dimers.

2.5 Conclusions

The results presented here show that the strong and directional hydrogen bonds in UPy telechelic polyesters can be used to prepare supramolecular copolymers in two fundamentally different ways. Functionalizing a random copolymer leads to supramolecular random copolymers, while mixing miscible telechelic homopolymers yields supramolecular segmented copolymers. In the functionalization of VL–containing diols, attention should be paid to the amount of catalyst used as it may cause depolymerization of VL segments, changing the original CL/VL ratio in the final product. The mechanical and thermal properties of the random copolymers and the segmented copolymers are different. In the random copolymer, the melting point of the polyester backbone is dependent on composition and is below room temperature when the monomer ratio is

Supramolecular Copolyesters with Tunable Properties

33

close to 1, negatively affecting the mechanical properties of the material. In the supramolecular segmented copolymer, the melting point of the crystalline polyester parts is nearly independent of composition, and the modulus of the materials can be tuned by varying the composition. Both types of materials show an additional melting point ascribed to crystallinity of the UPy groups. The materials can therefore be described as supramolecular thermoplastic elastomers51 in which the UPy moiety is mi