Tailor-made poly(meth)acrylate-based networks ... · Tailor-made poly(meth)acrylate-based networks - Preparation, Structure and Properties - Vincent Lima

Tailor-made poly(meth)acrylate-based networks :preparation, structure and propertiesLima, V.G.R.

DOI:10.6100/IR587324

Published: 01/01/2005

Document VersionPublisher’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 author's version of the article upon submission and before peer-review. There can be important differencesbetween the submitted version and the official published version of record. People interested in the research are advised to contact theauthor for the final version of the publication, or visit the DOI 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 page numbers.

Link to publication

Citation for published version (APA):Lima, V. G. R. (2005). Tailor-made poly(meth)acrylate-based networks : preparation, structure and propertiesEindhoven: Technische Universiteit Eindhoven DOI: 10.6100/IR587324

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 ?

Take down policyIf you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediatelyand investigate your claim.

Download date: 30. Jun. 2018

https://doi.org/10.6100/IR587324

https://research.tue.nl/en/publications/tailormade-polymethacrylatebased-networks--preparation-structure-and-properties(0e025e90-5560-4c0e-87ab-0688425eeda4).html

Tailor-made poly(meth)acrylate-based networks

- Preparation, Structure and Properties -

Vincent Lima

CIP-DATA LIBRARY TECHNISCHE UNIVERSITEIT EINDHOVEN Lima, Vincent Tailor-made poly(meth)acrylate-based networks : preparations, structure and properties / by Vincent Lima. – Eindhoven: Technische Universiteit Eindhoven, 2005. Proefschrift. – ISBN 90-386-2996-6 NUR 913 Subject headings: radical polymerization / crosslinking / chain transfer agents; RAFT / mass spectrometry; MALDI-TOF / liquid chromatography / telechelic acrylic polymers / coating materials / Young’s modulus; micro-indentation Trefwoorden: radicaalpolymerisatie / vernetting / ketenoverdracht ; RAFT / Massaspectroscopie; MALDI-TOF / vloeistofchromatografie / lineaire acrylaatpolymeren / deklagen / Young’s modulus; microhardheid © 2005, Vincent Lima Printed by PrintService Ipskamp, The Netherlands. Cover designed by Vincent Lima, Martien Frijns This research forms part of the research programme of the Dutch Polymer Institute (DPI), Technology Area Coating Technology, DPI project #205. An electronic copy of this thesis is available from the site of the Eindhoven University Library in PDF format (www.tue.nl/bib).

Tailor-made poly(meth)acrylate-based networks

- Preparation, Structure and Properties -

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 mandag 23 mei 2005 om 16.00 uur

door

Vincent Lima

geboren te Noisy-le-Sec, Frankrijk

Dit proefschrift is goedgekeurd door de promotoren:

prof.dr. R. van der Linde en prof.dr.ir. P. J. Schoenmakers Copromotor: dr. J.C.M. Brokken-Zijp

A mes parents

1

CHAPTER 1 Introduction I) POLYMERS ............................................................................................................................... 3 II ) COATINGS.............................................................................................................................. 4 III) OUTLINE OF THIS THESIS ..................................................................................................... 5 CHAPTER 2 Literature review I) CONTROLLED RADICAL POLYMERIZATION .......................................................................... 9 I.1 INTRODUCTION ....................................................................................................................... 9 I.2 NITROXIDE-MEDIATED POLYMERIZATION ........................................................................ 10 I.3 ATOM TRANSFER RADICAL POLYMERIZATION (ATRP) ................................................... 11 I.4 THE REVERSIBLE ADDITION-FRAGMENTATION CHAIN TRANSFER (RAFT) POLYMERIZATION ...................................................................................................................... 12 II) METHODS OF SYNTHESIS FOR α,ω−TELECHELIC POLYMERS ........................................... 16 II.1 POLYCONDENSATION .......................................................................................................... 16 II.2 IONIC POLYMERIZATION .................................................................................................... 17 II.3 RADICAL POLYMERIZATION............................................................................................... 17

II.3.1 Introduction ............................................................................................... 17 II.3.2 Iniferters .................................................................................................... 18 II.3.3 Nitroxide-Mediated Polymerization .......................................................... 18 II.3.4 Atom Transfer Radical Polymerization..................................................... 20 II.3.5 Reversible Addition-Fragmentation chain Transfer polymerization ........ 21

II.3 CHARACTERIZATION OF TELECHELIC POLYMERS ........................................................... 21 II.4 CONCLUSION ....................................................................................................................... 22 CHAPTER 3 Synthesis of hydroxyl-telechelic polymethacrylates by RAFT I) INTRODUCTION...................................................................................................................... 29 II) MATERIALS .......................................................................................................................... 29 III) CHARACTERIZATION TECHNIQUES................................................................................... 31 IV) RESULTS AND DISCUSSION ................................................................................................. 33 V) CONCLUSION ........................................................................................................................ 51 CHAPTER 4 Synthesis of carboxy-telechelic polyacrylates by RAFT I) INTRODUCTION...................................................................................................................... 55 II) MATERIALS .......................................................................................................................... 55 III) CHARACTERIZATION TECHNIQUES................................................................................... 56 IV) RESULTS AND DISCUSSION ................................................................................................. 58 IV) CONCLUSION....................................................................................................................... 68 CHAPTER 5 MALDI-TOF-MS analysis of RAFT polymers I) INTRODUCTION...................................................................................................................... 71 II) MALDI-TOF-MS: EXPERIMENTAL PROCEDURE ............................................................. 71 III) ANALYSIS OF THIOCARBONYL THIO-CONTAINING POLYMERS BY MALDI-TOF-MS... 72 IV) INFLUENCE OF THE LASER INTENSITY ON THE FRAGMENTATION .................................. 79

2

V) ANALYSIS OF TRITHIOCARBONATE-CONTAINING POLY(BUTYL ACRYLATE) BY MALDI-TOF-MS .................................................................................................................................... 82 V) CONCLUSIONS ...................................................................................................................... 85 CHAPTER 6 Networks applications: synthesis and properties I) INTRODUCTION...................................................................................................................... 87 II) EXPERIMENTAL.................................................................................................................... 88 II.1) THE ATTENUATED TOTAL REFLECTION FOURIER-TRANSFORM INFRARED SPECTROSCOPY .......................................................................................................................... 88 II.2) TRANSMISSION FOURIER-TRANSFORM INFRARED SPECTROSCOPY.............................. 89 II.3) TEMPERATURE MODULATED DIFFERENTIAL SCANNING CALORIMETRY (TMDSC) ... 89 II.4) MICRO-INDENTATION........................................................................................................ 89 II.5) SOLID-STATE NUCLEAR MAGNETIC RESONANCE ........................................................... 93 III) STARTING COMPOUNDS ..................................................................................................... 94 IV) RESULTS AND DISCUSSION................................................................................................. 96 IV.1) INFRA-RED RESULTS ......................................................................................................... 96

IV.1.1) Curing mechanism .................................................................................. 96 IV.1.2) Kinetics differences: random vs. telechelic .......................................... 102

IV.2) DETERMINATION OF E-MODULUS ................................................................................. 105 IV.2.1) Telechelic poly(butyl acrylate) micro-indentation measurements........ 105 IV.2.2 Influence of mono-functional chains on the mechanical properties of poy(butyl acrylate)networks.............................................................................. 109 IV.2.3) Random poly (butyl acrylate-co-acrylic acid) polymers micro-indentation measurements.................................................................................................... 110

VI.3) MTDSC MEASUREMENTS ............................................................................................. 112 VI.4) SOLID-STATE NUCLEAR MAGNETIC RESONANCE RESULTS ....................................... 115

VI.4.1) T1- relaxations experiments .................................................................. 115 VI.4.2) T2- relaxation experiments.................................................................... 119

VII) CONCLUSION ................................................................................................................... 122 CHAPTER 7 Conclusion and recommendations I) CONCLUSION........................................................................................................................ 127 II) RECOMMENDATIONS ......................................................................................................... 129

Introduction

3

CHAPTER 1

INTRODUCTION

I) Polymers

Polymer is a word invented in the twentieth century from Greek words πολυ (poly) and µερ

(mer) that can be literally translated by “many parts”. This term is employed to describe

macromolecular entities based on the repetition of one or several moieties. The definition of a

polymer was introduced by Staudinger1, who suggested that polymers were large molecules

containing long sequences of chemical units linked by covalent bonds. Interestingly, this

statement was issued 10 years after the first manufacturing of synthetic polymer (polymer made

out of phenol and formaldehyde under the trade name Bakelite2). Man knew how to synthesize

polymers before understanding what he was doing!

Polymeric materials are an important part of our modern life. They are generally regrouped in

the common language under the improper term of plastics (derived from thermoplastic), and

have generally to fight a negative image of oil-derivative products usually associated with

pollution images in the society’s collective mind. The truth is rather different, as recycling is

possible for many types of polymers. The incineration of them is also a way to generate energy.

Misunderstanding takes place often in the public opinion: polymers are generally not bio-

degradable, which does not mean that they are polluting. Efficient waste disposal and self-

responsibility from the citizens can greatly reduce the impact of the polymers on our

environment.

The often forgotten part of polymeric material is the natural polymers. Compounds like wool,

silk, cellulose belong as well to the polymer family. The genetic code of every living species

(including humans) is stored in DNA (Desoxyribose Nucleic Acid) molecules. All the fine

organic chemistry occurring in living organisms is regulated through equilibrium reactions

involving proteins, which are nothing less than aminoacids copolymers.

Chapter 1

4

Few levels below the exceptional complexity of life mechanisms, mankind tried to develop its

own high-tech systems. Polymers are an integrant part of those inventions. Display panels,

memory chips, and semiconductor technologies based exclusively on polymers are nowadays

taking increasing importance in product technologies in detriment of metals. Pharmaceutical

applications have been found as well in the “drug delivery systems”, which show that

biocompatible polymers can be used for the controlled delivery of medication in vivo. The many

application areas of polymers extend from bulk materials, adhesives, foams and packaging

materials to textile (synthetic polymers such as polyesters or polyacrylics to replace natural ones

like silk, fur or wool), industrial fibers, composites…

It is only in the last few decades of the past millennium that polymers have been recognized as

materials that can truly form unique and intelligent materials, applicable to areas where metals

and ceramics would not be usable. The reason behind it was the shift from a research for new

exotic monomers or combination of monomers, to a research for gaining control over the

microscopic structure of the polymer. The chain-length distribution, the monomer sequence

distribution (for copolymers), the tacticity, the functionality distribution and the degree of

branching for cross-linked materials, were for instance recognized as major factors influencing

the macroscopic properties of the polymeric material. It is this understanding of these

(micro)structure – properties relationships that will certainly be the base for polymer

applications that will be discovered in the future.

II ) Coatings

In order to protect the materials from the degradation induced by the exterior environment

(atmospheric moisture, UV-light, everyday life use…), or for decorative purposes (paints,

lacquers) the great majority of products have to be topped with a thin layer at their surfaces.

This layer is defined as a coating. A coating must be a completely continuous film in order to

fulfill its function. Any imperfection will lower its efficiency (worst chemical protection or

esthetic outlook).

The polymer-based coatings generally contain several components:

- The binder: this is the material that forms the continuous film that adheres to the

substrate. For polymeric coatings, the polymer can be prepared and incorporated in

the formulation before the application, or the polymerization can occur after the

application. The binder governs, to a large extent, the properties of the film.

Introduction

5

- The pigments: they are insoluble powders dispersed in the coating mixture. Their main

function is to give color to the coating and to provide its eventual hiding power. They

can also stabilize the coating, as they are partially absorbing or reflecting the visible

light, suppressing the damaging effect that the latter can have on the binder.

- The solvents: incorporated in many formulations, their role is to adjust the viscosity of

the coating mixture for the required application mode. The solvents evaporate during

and after application. Due to environmental regulations, the importance of organic

solvents and in a more general term of Volatile Organic Components (VOC) in coatings

formulations is constantly reduced. The trend in industrial research is to develop new

systems based on the use of water as a solvent or even to suppress the use of any solvent

(powder coatings).

- The additives: they are used to avoid defects (e.g. foam bubbles, bad leveling,

sedimentation) or to provide special properties (e.g. UV stability, improved surface slip)

to the coating.

Concerning the future of coatings, the tendency is to use them more and more as “smart”

material and not only as passive barriers. Biologically-active coating, designed to release

biocides (such as organotin compounds to prevent marine fouling) or electricity-conductive

coatings for electronic display applications are good examples of what “smart” coatings can be.

III) Outline of this thesis

Conventional poly(meth)acrylate-based coating technology relies on random

copolymerizations of acrylic, methacrylic and functional comonomers. Thermosetting acrylics

have typically a number average molecular mass of 10,000-20,000 g/mol and a polydispersity

index between 2 and 33. As an example, the hydroxyl-functional comonomers employed are

usually 2-hydroxyethyl methacrylate (HEMA) and 2-hydroxypropyl methacrylate (HPMA). The

polymers are then reacted with a multifunctional crosslinker, e.g. a triisocyanate in the case of

hydroxyl-functional resins, in order to obtain the three-dimensional network. The resulting film

obtained from such systems can be qualified as ‘non-ideal’, as the network contains dangling

ends, rings and even free polymer chains. These defects in the network structure have in general

a negative effect on material properties such as the Young modulus, the hardness and the scratch

resistance. Highly-functional chains are responsible for early gelation of the system and

brittleness.

Chapter 1

6

In the 1980’s, the use of chain-transfer agents with functional groups was developed. For

example adding 2-mercaptoethanol4 or 2-mercaptopropionic acid5 in the polymerization mixture

will result, after transfer reactions involving those compounds, in initiating radicals bearing

hydroxyl or carboxyl groups. This will lead to a high fraction of molecules with a hydroxyl or a

carboxyl function on one end, reducing dramatically the fraction of non- and mono- functional

chains. The polydispersity index was also reduced, dropping to about 1.74. After cross-linking,

the films obtained exhibited improved mechanical properties but the strong odor of thiol-

containing compounds caused handling problems.

In order to investigate the influence of the functionality per polymer chain distribution and the

molecular weight distribution (MWD) on the formations and properties of poly(meth)acrylate

networks, we employed the Reversible Addition-Fragmentation chain Transfer (RAFT)6

polymerization technique to synthesize linear well-defined telechelic poly(meth)acrylates. The

obtained polymers will have to be telechelic and with a polydispersity index below 1.2. After the

synthesis of those molecules, the network formation and network properties will be studied,

establishing differences in networks made out of poly(meth)acrylates synthesized via random

radical copolymerization and the ones made out of telechelic poly(meth)acrylates.

A general introduction on polymers and coatings has been given in Chapter 1. The main

objectives of the investigation described in this thesis have also been detailed.

Chapter 2 will give a general overview about the controlled radical polymerization technique

and the ways to produce telechelic polymers, based on a literature survey.

Chapter 3 will treat the synthesis and characterization of linear hydroxyl-functional

polymethacrylates.

Chapter 4 will deal with the synthesis and characterization of linear carboxyl-functional

polyacrylates.

The discussion in Chapter 5 will develop around the Matrix Assisted Laser Desorption-

Ionisation Time Of Flight Mass Spectrometry (MALDI-TOF-MS) analysis of the polymers

prepared by RAFT polymerization.

In Chapter 6, the kinetics and mechanisms of the cross-linking reactions will be studied via

Fourier Transformed InfraRed (FT-IR) spectroscopy. Comparison between polyacrylates

synthesized via random radical copolymerization and telechelic polyacrylates with low

polydispersities will be made. The properties of the cross-linked networks will be investigated

using solid-state Nuclear Magnetic Resonance (NMR) spectroscopy, Temperature-modulated

Differential Scanning Calorimetry (TMDSC) and micro-indentation experiments.

Introduction

7

Chapter 7 will summarize the main findings of this thesis and consider the potential future

work that can be done on this subject.

Chapter 1

8

Reference List

1. Staudinger, H.; Ber. 1920, 53B, 1073.

2. Baekeland, L. H.; J. Ind. Eng. Chem 1909, 1, 149.

3. Wicks Jr., Z. W., Jones, F. N., Pappas S.P.; Organic Coatings: Science and Technology

Volume I: Film Formation, Components and Appearance. New York: John Wiley &

Sons, Inc.; 1992

4. Gray, R. A.; J. Coating. Technol. 1985, 57, 83.

5. Buter, R.; J. Coating. Technol. 1987, 59, 37.

6. Mayadunne, R. T. A., Rizzardo, E., Chiefari, J., Chong, Y. K., Moad, G., Thang, S. H.;

Macromolecules 1999, 32, 6977.

Literature review

9

Initiation

I 2 Rkd

R + Mki

RM

Propagation

RM + n M kp

RMn+1

Termination

2 RMkt

dead chains

Transfer

RM + R'-Xktr RMX + R'

CHAPTER 2

LITERATURE REVIEW

I) Controlled radical Polymerization

I.1 Introduction

In the past half-century, radical polymerization has enjoyed considerable interest from

industry. It still is, particularly due to its high tolerance for impurities. However, one major

drawback is the poor control over the characteristics of the synthesized polymers, in particular

over the polymer molecular weight (distribution) and functionality. The main reactions

occurring in the free-radical polymerization are depicted in Figure 2.1.

Figure 2.1. Schematic representation of the free-radical polymerization

In Figure 2.1, the term “dead chains” refers to chains that have undergone termination reaction

(either by combination or by disproportionation), and that can not add further monomer in

Chapter 2

10

propagation steps. Those dead chains are created by the reaction of two propagating radicals,

thus the termination rate can be expressed by the following equation:

2].[2 MkR tt =

The propagating step is governed by the following kinetic equation:

]][.[ MMkR pp =

In the past twenty years, numerous research groups have devoted their efforts towards the

control of this polymerization process. The early techniques employed were based on reversible

termination of the propagating polymeric radicals, in Nitroxide-Mediated Polymerization

(NMP)1 and Atom Transfer Radical Polymerization (ATRP)2,3. Compared to classical radical

polymerizations, controlled radical polymerizations exhibit a low fraction of terminated polymer

chains. This is explained by the introduction of a reversible deactivation reaction. A small

number of radicals and a large number of ‘dormant’ chains are present in the system. An

increase of the polymer chain lifetime in controlled radical polymerization is observed, the

growing polymer chains being reversibly deactivated by either a nitroxide (NMP) or a halide

(ATRP). In both cases, most of the polymer chains are reversibly deactivated, so a large number

of dormant chains and a small number of growing chains are obtained. As the termination

reaction between radicals is a second order in radical concentration and propagation is a first

order reaction, the effect of the radical concentration decrease will be much more pronounced on

the irreversible termination events. At full monomer conversion, the majority of the chains will

bear a nitroxide or a halide as an end-group that may allow a possible reactivation. Fresh

addition of monomer will result in a second growth of the chains.

I.2 Nitroxide-Mediated Polymerization

The first report of nitroxide-mediated polymerization was published in 1984 by Solomon et

al.1 NMP is related to the reversible trapping of propagating carbon-centered radicals by stable

nitroxide species (Figure 2.2). This produces an alkoxyamine with a carbon-oxygen bond that

can break at relatively high temperatures (typically around 120˚C). The chemical structure of the

nitroxide group employed will influence the equilibrium between propagating radicals and

dormant species. Nitroxide radicals are remarkable in many ways, as they react only with

carbon-centered radicals and do not dimerize.

Practically, two experimental procedures may be employed to conduct NMP. The first one is

the use of a classical initiator in combination with a nitroxide, e.g. 2,2,6,6-tetramethylpiperidine-

Literature review

11

O

O CH2 CH O N

O

O CH2 CH + O N

O

O CH2 CH + n CH2 CHkp

O

O CH2 CH

n+1

O

O CH2 CH

n

+ O N

O

O CH2 CH O N

n

TEMPO

kd

ka

N-oxyl (TEMPO)4. The presence of TEMPO in the system will ensure the control over the

polymerization under high temperature conditions (typically at 130˚C for styrene

polymerizations). The second experimental procedure is the addition of a selected alkoxyamine

to a monomer mixture. At high temperature, the carbon-oxygen bond will break, forming a

carbon-centered radical suitable for initiating the polymerization, and a nitroxide which will

control the radical polymerization by reversible termination. The situation is then identical to the

first one described.

Figure 2.2. Example of styrene nitroxide-mediated polymerization

The control over the polymerization will depend on many parameters, including the activation

(ka) - deactivation (kd) rate constants, the concentration of alkoxyamines, the nature of the

monomer and the temperature. Although remarkable achievements have been reported in the

polymerizations of styrene monomer, the NMP technique showed some limitations with the

control of acrylic or methacrylic monomers polymerizations with the TEMPO nitroxide. The

low activation constant for the acrylates5 and the H-abstraction of the β - hydrogen from the α-

methyl group for the methacrylates6 are the two major factors that prevent the production of

materials with low polydispersities by NMP of those monomers. However, newly designed

nitroxides have recently been used to control the polymerizations of 1-3 dienes and acrylates7-9.

I.3 Atom Transfer Radical Polymerization (ATRP)

Atom Transfer Radical Polymerization was introduced in 1995 by Matyjazewski3 and

Sawamoto2. The principle is based on the Karash reaction: the Atom Transfer Radical Addition,

which is widely used by organic chemists for carbon-carbon bond formation.

Chapter 2

12

R X + Mtn

kact

kdeactR + Mt

n+1X

kp

+ M

The control in an ATRP system is induced by the presences of an organic halide initiator and a

transition metal complex in a monomer solution. The reversible exchange of the halogen atom

between the growing polymer chain and the transition metal complex (in its higher oxidation

state) ensure the control over the polymerization (Figure 2.3).

Figure 2.3. General scheme of reversible termination in ATRP with R-X an alkyl halide, Mtn a

metal complex, kact the activation kinetic constant, kdeact the deactivation kinetic constant, kp the

propagating kinetic constant

The most commonly used transition metal couple (activator/deactivator) is the Cu(I)/Cu(II)2,10

system, although Ru(II)/Ru(III)3,11,12, Mo(V)/Mo(VI)13, Fe(II)/Fe(III)14-16, Ni(II)/Ni(III)17-19 and

Rh(I)/Rh(II)20 combinations are also reported in literature. The catalyst has to be complexed in

order to be active and be able to control the polymerization. This is achieved using some

nitrogen-containing ligands, e.g. (substituted) 2,2’-bipyridines2,16, Schiff bases21,22, multidentate

tertiary amines23. The role of the ligand is to solubilize the metal in order to create a

homogeneous system and to establish a proper equilibrium between active and dormant species.

ATRP appeared to be more versatile than NMP, since it is applicable to the polymerization of

styrenic10,16,24 and (meth)acrylic monomers3,16,25-27. Major limitations are still encountered in

polymerizations of 1,3-diene, vinyl acetate28 and monomers that can complex and poison the

metal, e.g. acid group containing monomers.

The application of ATRP in industry is still hindered by the post-polymerization purification

(in order to remove the metal complex), often necessary for esthetic, environmental and stability

reasons.

I.4 The Reversible Addition-Fragmentation chain Transfer (RAFT) polymerization

The origin of Reversible Addition-Fragmentation chain Transfer polymerization can be found

in 1998 with the studies of Rizzardo et al.29,30 on polymerizations of acrylates, methacrylates

Literature review

13

PnS

Z

SR+

PnS

Z

SR

kadd

PnS

Z

S+ R

R + m (M) Pm

Pm

M

+

PnS

Z

SR

PmS

Z

SPn

S

Z

SPm

Pn

M

+

Z

SPn

S

I + n (M)ki, kp

kadd

k-add

k-add

kβ

kβ

ki, kp

and styrene, in the presence of thiocarbonyl thio compounds. The obtained polymers exhibit low

polydispersities and predictable molecular weights.

The mechanism was partially elucidated via ESR (Electron Spin Resonance) spectroscopy31

and the extra reactions that are added to a classical radical polymerization are depicted in Figure

2.4.

Figure 2.4. Extra reactions occurring in a RAFT polymerization process

After a classical initiation step, a growing oligomeric chain reacts with a thiocarbonyl thio

compound (RAFT agent), via addition to the sulfur-carbon double bond. A carbon-centered

intermediate radical is then formed. This intermediate radical has the ability to regenerate by

fragmentation the initial oligomeric radical/RAFT agent couple or to form a new propagating

radical R. and an oligomeric thiocarbonyl thio compound (dormant polymer species). The R.

radical should reinitiate the polymerization, forming a new propagating chain. The newly-

formed thiocarbonyl thio-terminated chain acts as a chain transfer agent, like the initial RAFT

agent. This will ensure by repetitive addition-fragmentation steps on the polymeric thiocarbonyl

thio compounds, the exchange of the radical among all the chains present in the polymerization

system.

The leaving group of the thicarbonyl thio compounds (R, figure 2.4) has to be chosen in such

a way that it is a more stable radical than the oligomeric polymer chain in order to shift the

RAFT pre-equilibrium (Figure 2.4) towards the release of R radicals. The latter must rapidly

reinitiate the polymerization. In general, R is a substituted alkyl group, whose stability can be

increased with stabilizing functional groups32.

The activating group of the thicarbonyl thio compound (Z, figure 2.4) influences the stability

of the intermediate radical and thus the rate of addition of the growing radical to the

thiocarbonyl thio compounds.

Chapter 2

14

It has to be emphasized that transfer reactions do not influence the concentration of

propagating radicals. This concentration is dictated by the steady state approximation, assuming

that the production of radicals by initiation balances their disappearance by termination

reactions. In typical RAFT recipes, this concentration is in the order of magnitude of 10-8 mol.L-

1. The concentration of dormant species is equal to the initial concentration of RAFT agent

(conservation of the thiocarbonyl thio moiety), in the order of magnitude of 10-3 mol.L-1. Thus, it

can be estimated that there is only one chain “active” for 100.000 dormant chains in a typical

RAFT polymerization system.

The rapid exchange of a small number of radicals over a large number of polymeric chains

allow all the chain to grow at the same rate. This is the reason why the RAFT process produces

polymers with masses that increase linearly with conversion.

One could argue that the constant production of radicals, creating new chains, via the initiator

decomposition is not compatible with the definition of a living process. It is true that a living

process implies a constant number of chains throughout the polymerization. However, if the

total amount of decomposed initiator is kept at a low level compared to the RAFT agent

concentration, the number of chains can be assumed to remain constant throughout the

polymerization since:

[ ] [ ] [ ] [ ]( )0 02 -t tchains RAFT f I I= +

In typical RAFT recipes, the second term in the previous equation can be neglected and the

number of chain can be considered constant throughout the reaction.

The scientific community does not unanimously agree on the reasons of the reaction rate

retardation which is sometimes witnessed in the RAFT polymerization. The stability and fate of

the intermediate radical is the central point of the argument. A slower rate of polymerization

compared to the rate of the equivalent conventional free radical polymerization is due to a lower

concentration of propagating radicals in the presence of the RAFT agent.

A first group of scientists described the intermediate radical as a “radical sink” 33,34. A low

constant of fragmentation kβ is the explanation given to justify the lower concentration of

propagating radicals. After addition to the carbon-sulfur double bond of the RAFT agent, the

radical is stored in the intermediate radical and fragments very slowly, lowering the propagating

radical concentration.

The other theory explaining the retardation in the RAFT polymerization was introduced by

Monteiro et al.35 An additional reaction was added to the original RAFT mechanism scheme

from Rizzardo et al.29, i.e. the intermediate radical termination (Figure 2.5). According to this

Literature review

15

theory, the carbon-centered radical can terminate in a similar way as a propagating chain. Due to

the presence of intermediate radicals and propagating radicals in the reactive system,

termination events between two intermediate radicals or between an intermediate radical and a

propagating radical might occur. Three-arm and four-arm shaped polymers might then be

formed. These events reduce the concentration of propagating radicals in the polymerization

system, decreasing the overall polymerization rate, causing retardation.

Figure 2.5. Possible reactions for he intermediate radical termination, forming 3-arm and 4-

arm polymer chains

The aim of this introduction on RAFT is not to solve this problem of retardation (which is not

of great importance in this thesis, as the RAFT polymerizations we will use for our synthesis of

polymers for network formation studies are not prone to retardation), but to give an introduction

about the hottest debate in the RAFT literature. Arguments were exchanged in papers (involving

computer simulations33, ab initio calculation on molecular orbitals36, SEC separation on non-

polymerizing model systems37,38), without a definitive answer that would convince the whole

community of scientists working on RAFT. One of the major problems is to find a way to access

all the kinetic parameters of the reactions involved in RAFT.

Compared to the numerous advantages of the RAFT process (possibility to control the

polymerization of numerous monomers, application to water-borne emulsion39-43, easy synthesis

of complex architectures…), the problem of retardation occurring with specific RAFT agents

has to be considered as a minor problem. This is of course of interest for a complete

understanding of the mechanism, but retardation can be easily avoided by selecting the

appropriate activating group. It has been found that replacing the phenyl activating group by a

Pi + PmS S

Pn

ZPm

S SPn

Z

Pi

PmS S

Pn

Z

PiS S

Pj

Z+ Pi

S SPj

Z

SSPnPm

Z

Chapter 2

16

benzyl in thiocarbonyl thioesters44 suppresses the retardation in styrene polymerization. This

replacement reduces the stability of the intermediate radical, increasing its fragmentation rate

constant and thereby reducing its concentration.

II) Methods of synthesis for α,ω−telechelic polymers

End-functional polymers are extremely important for their possible application as building

blocks for instance as block copolymers, surfactants, macromonomers45-47. Functionality at

every chain-ends of a macromolecular chain defines a telechelic polymer. In the polymer

science community, the term “telechelic” commonly refers to a linear chain having the same

functionality at both chain-ends. Those materials are of great interest for their ability to form tri-

block polymers or tailor-made polymer networks by reaction with cross-linkers. It is for the

latter application that the first part of this thesis will focus on synthesis and characterization of

linear α,ω−telechelic polymers, with narrow molecular weight distributions.

II.1 Polycondensation

One of the most often used methods to produce telechelic polymers is the polycondensation

technique. For example, reacting a bi-functional monomer A2 with another bi-functional

monomer B2 in slight excess will result in telechelic linear polymer chains bearing B-functional

end-groups (Figure 2.6). A slight excess of A2 will produce in a similar way A-terminated

telechelic linear polymers (Figure 2.6).

Figure 2.6. Preparation of telechelic polymers via polycondensation reactions

A A B B(1+ε)+

B A B B A B

B B (1+ε)+ A A

A B A A B A

Literature review

17

A A AMM...MM AMM...MMBInitiation Propagation End-capping

II.2 Ionic polymerization

Another well-known method to prepare telechelic polymers is living ionic polymerization.

Using a functional initiator and a terminal end-capper, telechelic polymers can be produced

(Figure 2.7). Reports of telechelic polystyrene48-50, polybutadiene51 or polyisoprene48,50 can be

found in literature.

Figure 2.7. Preparation of telechelic polymers via living ionic polymerizations

Ionic polymerization leads to well-defined telechelic polymers with very low polydispersities.

However, the stringent conditions required for ionic polymerizations coupled with the limited

choice of monomers available restrains the use of ionic polymerizations for the production of

telechelic material in the industry.

II.3 Radical polymerization

II.3.1 Introduction

The most attractive polymerization technique for the industry is radical polymerization for its

relative high tolerance towards impurities, and the wide range of monomers that can be

polymerized. In order to control the structure of the produced chains, chemicals interfering in

the chain architecture-determining events (initiation, termination and chain-transfer) have to be

employed. The simplest way one can imagine to archive the synthesis of telechelic linear

polymer is to use a functional initiator and to ensure that the only chain stopping event is

termination via combination (Figure 2.8).

The ideal case of exclusive combination as chain-stopping event is extremely rare in the

polymer world. Even poly(styrene) and poly(acrylate)s, polymers that are know to terminate via

combination, exhibit a significant percentage of chains that terminate via disproportionation52,53.

Chapter 2

18

Figure 2.8. Preparation of telechelic polymers via radical coupling following a free radical

polymerization

The resulting hydrogen-terminated and double bond-terminated chains will lower the average

functionality per chain of the final polymer, and may lead to branching, due to their reactivity as

macromonomer.

II.3.2 Iniferters

An extra technique has to be mentioned in the list of controlled radical processes. In the early

80’s, Otsu et al.54,55 obtain a certain degree of control over the polymerization of styrene and

methyl methacrylate by adding disulfide or dithiocarbamate compounds to the polymerization

media. Photochemically, those sulfur-containing agents decompose, forming sulfur-centered

radicals. The latter can act as initiators, chain transfer agent and termination agent. The term

iniferter given to those additives recalls their triple role in the polymerization. The term “living”

could not be employed to describe such systems. The blame is mainly due to the low transfer

constant of the dormant polymer species to the propagating ones, and to the occurrence of many

side reactions. The iniferter method was employed with the idea to obtain telechelic polymers.

However, a significant amount of non-functional chains were found in the reaction system56.

II.3.3 Nitroxide-Mediated Polymerization

A logical step forward towards the production of telechelic polymers via a radical process is

the use of controlled radical polymerization processes, which were described in section II.1.

Nitroxide-mediated polymerization was reported as a successful method to obtain telechelic

chains.1 One of the possible methods involves in a first step the polymerization of styrene in the

presence of a hydroxyl-functional alkoxyamine. In one step, α-functional chains are obtained,

with an alkoxyamine group in the ω position. The successful reduction of the latter group in an

alcohol, using an acetic acid / zinc mixture1,57, allows in principle the preparation of well-

defined α,ω-hydroxyl telechelic poly(styrene) with a functionality close to 2 (Figure 2.9).

A A AMM...MM AMM...MMMM...MMAInitiation Propagation Termination

Literature review

19

Figure 2.9. Preparation of telechelic polymers via NMP followed by a reduction of the

alkoxyamine

Another method combines the use of TEMPO with a hydroxyl-functional initiator (Figure

2.10). At the end of the controlled polymerization, the average functionality is slightly above 1,

probably due to termination reactions via combination occurring during the polymerization57.

This result was expected, since all the chains were initiated by a hydroxyl-functional peroxide

fragment. In a post-polymerization treatment similar to the one described previously, the

reduction of the alkoxyamine into an alcohol allows the synthesis of α,ω-hydroxyl telechelic

poly(1,3-butadiene).

Figure 2.10. Preparation of telechelic polymers via NMP using an hydroxyl-functional

initiator

Another method consists to heat up the α-functional chains to 165˚C in order to thermally

cleave the carbon-oxygen bond57. Compared to a NMP process, the absence of monomer

suppresses the propagation possibility, and the chains are forced to terminate or to recombine

with the alkoxyamine, which would lead to another potential cleavage of the carbon-oxygen

bond. The yield of obtained telechelic material depends on the proportion of termination by

combination for the polymer chains, and the occurrence of side reactions. The results published

by Pradel et al.57 indicate that the average functionality per polymer chain below 2 and that side

reactions are the probable reason for these results.

HOO N

HO

O N

NMP CH3COOH / Zn

nStyrene

HO

OH

n80oC

H O O H∆

HO1,3-butadiene

TEMPOHO CH2 CH CH CH2 O N

n

CH3COOH / Zn

HO CH2 CH CH CH2 OHn

80oC

Chapter 2

20

HO CH2 CH2 O

O

CH3

BrATRP

MMAHO CH2 CH2 O

O

CH3

O O

Br

n

H2N C4H9 OH

HO CH2 CH2 O

O

CH3

O O

NH

n

C4H9 OH

+ HBr

The methods described are in principle not restricted to the specific monomer systems

described. They are in general applicable to the monomers, whose polymerization can be

controlled via the NMP process.

II.3.4 Atom Transfer Radical Polymerization

In order to obtain α,ω-telechelic chains with ATRP, the usual first step is to employ functional

halides as initiators. End-group functionality such as hydroxyl27,58, ester2,27, phenyl59, amine59,

aldehyde59, anhydride58 or carboxylic acid58,60 can easily be introduced though the proper choice

of initiator. At the end of a classical ATRP process, the functional group and

the halide atom will be found in the α and ω positions. A modification of the halide group to

give the same functionality will result in linear telechelic polymer chains bearing the same

function at both ends. One of the main techniques used to achieve this is the nucleophilic

substitution. Using functional azides24,61, amines25,62 or phosphines63, chain-end functionality is

easily changed from halide to the desired one. In Figure 2.11 the end-chain functionalization

reaction described by Coessens et al.63 is presented.

Another possibility reported by Coessens et al.25 is an end-capping reaction with a functional

monomer that can not homopolymerize (e.g. allyl alcohol25, ethene6, C6064, maleic anhydride65)

under ATRP conditions, but can copolymerize with the monomer used in the ATRP. Only one

monomer unit adds to the already-existing chain, leading to an efficient end-capping.

Figure 2.11. Preparation of telechelic polymers via ATRP followed by a nucleophilic

substitution

Literature review

21

II.3.5 Reversible Addition-Fragmentation chain Transfer polymerization

The most promising report of synthesis of telechelic polymers via RAFT has been published

in literature by Lai et al.66 Using symmetrical trithiocarbonates, telechelic polymers can be

easily obtained in one step (Figure 2.12). One limitation of this technique is the poor results with

methacrylic monomers, due to the relatively low stability of the leaving group compared to the

one of a methacrylic radical.

Figure 2.12. Production of telechelic poly(acrylate) in one step using the RAFT

polymerization with trithiocarbonates as RAFT agents

It can be noted that the use of an initiator bearing the same functionality as the leaving group,

ensures that all the chains will be initiated by a COOH-containing group (neglecting the chain

transfer events to e.g. monomer and solvent). This method has been employed before by

Mayadunne et al.67, but the functionalities introduced at the end-chains were less attractive

(CH2Ph, CH2(CH3)Ph).

A recent article of Lui et al.68 describes the synthesis of di-hydroxyl-terminated telechelic

polystyrene and poly (methyl acrylate) using a symmetrical di-hydroxyl-functional

trithiocarbonate. The characterization of the polymers in terms of functionality was limited to IR

and NMR.

II.3 Characterization of telechelic polymers

In the literature, numerous claims about telechelic polymer synthesis have been published,

without thorough characterization works on the resulting polymer. Many examples of

affirmations based on the theoretical chemistry involved in the synthesis, or non-adapted

analytical techniques, can be found.

S

SS COOHHOOC

HOOC (Polyacrylate) S

S

S (Polyacrylate) COOH

Acrylate

+ N N

CN

COOH

CN

HOOC

Chapter 2

22

Often, titration of the functional groups is the method of choice to prove that the produced

polymers are telechelic. However, the titration technique gives an average number for the

number of functions per polymer chain. One of the most commonly encountered mistakes is the

confusion between the average functionality per polymer chain and the functionality

distribution. If a batch of polymer consists of a batch of 50% of tri-functional polymers, and

50% of mono-functional polymers, the average functionality would be estimated by titration to

be similar to the one of a sample containing 100% of bi-functional polymer. The average

functionality is the same, but the distributions are different. This problem of giving an average

functionality rather than a functionality distribution is encountered also in InfraRed (IR),

UltraViolet (UV) or Nuclear Magnetic Resonance (NMR) spectroscopies.

Other doubts about the relevance of results can rise when mass-spectrometric techniques are

used to prove the production of telechelic polymers. Matrix-Assisted Laser

Desorption/Ionization Time-Of-Flight Mass Spectrometry (MALDI–TOF–MS) and

Electrospray Ionization Mass Spectrometry (ESI–MS) provide good qualitative information, but

they can not be used for quantitative results, due to variations in the ionization, separation, and

detection efficiencies.

Liquid chromatography is nowadays one of the most popular methods of the analysis for the

Functionality-Type Distribution (FTD). In a more specific class of liquid chromatography,

Liquid Chromatography under Critical Conditions (LCCC), the retention times of polymer

chains are only based on the number of functionalities present in the polymer chains, and are

independent of their molar masses69-71. More details on LCCC will be given in Chapter III and

Chapter IV, when the methods of characterization of synthesized polymers will be discussed.

II.4 Conclusion

The need for well-defined telechelic polymers with a low polydispersity dictated our choice to

apply a CRP technique for the synthesis of our poly(meth)acrylate. The stringent conditions

required for ionic polymerizations seemed to limit its use. The nitroxide-mediated

polymerization technique was discarded, due to major drawbacks in the polymerization of

methacrylates, which would have prevented the synthesis of tailor-made polymers. The choice

between RAFT and ATRP was made primarily based on the versatility of the RAFT process,

and its ability to produce telechelic carboxyl-functional polyacrylates in one step with an

appropriate trithiocarbonate. Moreover, the absence of metal complexes in RAFT

Literature review

23

polymerization is also beneficial, as those compounds can have a major impact on the cross-

linking process that will take place after the synthesis.

In parallel with carboxyl-functional poly(acrylate), hydroxyl-functional poly methacrylate will

be also synthesized via a two-step post-polymerization procedure. Using a hydroxyl-functional

RAFT agent, hydroxyl-functional polymethacrylate can be synthesized. Changing after

polymerization the thiocarbonyl thio into a hydroxyl function should allow the production of

telechelic polymethacrylates. The application of the synthesis process described by Liu et al.68

could have been suitable for our study, but it was published in the last year of this project,

preventing its utilization due to time constraints. For both hydroxyl-polymethacrylate or

carboxyl-polyacrylate, the LCCC technique will be the preferred method to analyze the

polymers synthesized, and check the functionality-type distribution.

Chapter 2

24

Reference List

1. Solomon, D. H.; Rizzardo, E.; Cacioli, P. 84-304756 EP 135280 July 11, 1984

2. Wang, J. S., Matyjaszewski, K.; Macromolecules 1995, 28, 7901.

3. Kato, M., Kamigaito, M., Sawamoto, M., Higashimura, T.; Macromolecules 1995, 28,

1721.

4. Georges, M. K., Veregin, R. P. N., Kazmaier, P. M., Hamer, G. K.; Macromolecules

1993, 26, 2987.

5. Bon, S. A. F.; PhD Thesis 1998

Eindhoven·University·of·Technology,·Eindhoven,·The·Netherlands

6. Bon, S. A. F., Steward, A. G., Haddleton, D. M.; J. Polym. Sci. Pol. Chem. 2000, 38,

2678.

7. Benoit, D., Chaplinski, V., Braslau, R., Hawker, C. J.; J. Am. Chem. Soc. 1999, 121,

3904.

8. Benoit, D., Harth, E., Fox, P., Waymouth, R. M., Hawker, C. J.; Macromolecules 2000,

33, 363.

9. Grimaldi, S., Finet, J. P., Le Moigne, F., Zeghdaoui, A., Tordo, P., Benoit, D.,

Fontanille, M., Gnanou, Y.; Macromolecules 2000, 33, 1141.

10. Matyjaszewski, K., Patten, T. E., Xia, J.; J. Am. Chem. Soc. 1997, 119, 674.

11. Simal, F., Demonceau, A., Noels, A. F.; Tetrahedron Lett. 1999, 40, 5689.

12. Simal, F., Demonceau, A., Noels, A. F.; Angew. Chem. Int. Edit. 1999, 38, 538.

13. Brandts, J. A. M., van de Geijn, P., van Faassen, E. E., Boersma, J., Van Koten, G.; J.

Organomet. Chem. 1999, 584, 246.

14. Teodorescu, M., Matyjaszewski, K.; Macromolecules 1999, 32, 4826.

15. Ando, T., Kamigaito, M., Sawamoto, M.; Macromolecules 1997, 30, 4507.

16. Matyjaszewski, K., Wei, M., Xia, J., McDermott, N. E.; Macromolecules 1997, 30,

8161.

17. Granel, C., Dubois, P., Jerome, R., Teyssie, P.; Macromolecules 1996, 29, 8576.

18. Uegaki, H., Kotani, Y., Kamigaito, M., Sawamoto, M.; Macromolecules 1997, 30, 2249.

19. Uegaki, H., Kotani, Y., Kamigaito, M., Sawamoto, M.; Macromolecules 1998, 31, 6756.

20. Percec, V., Barboiu, B., Neumann, A., Ronda, J. C., Zhao, M.; Macromolecules 1996,

29, 3665.

21. Haddleton, D. M., Jasieczek, C. B., Hannon, M. J., Shooter, A. J.; Macromolecules

1997, 30, 2190.

Literature review

25

22. Haddleton, D. M., Duncalf, D. J., Kukulj, D., Crossman, M. C., Jackson, S. G., Bon, S.

A. F., Clark, A. J., Shooter, A. J.; Eur. J. Inorg. Chem 1998, 1799.

23. Xia, J., Gaynor, S. G., Matyjaszewski, K.; Macromolecules 1998, 31, 5958.

24. Matyjaszewski, K., Nakagawa, Y., Gaynor, S. G.; Macromol. Rapid Comm. 1997, 18,

1057.

25. Coessens, V., Matyjaszewski, K.; Macromol. Rapid Comm. 1999, 20, 127.

26. Coessens, V., Pyun, J., Miller, P. J., Gaynor, S. G., Matyjaszewski, K.; Macromol.

Rapid Comm. 2000, 21, 103.

27. Haddleton, D. M., Waterson, C., Derrick, P. J., Jasieczek, C. B., Shooter, A. J.; Chem.

Commun. 1997, 683.

28. Xia, J., Paik, H. J., Matyjaszewski, K.; Macromolecules 1999, 32, 8310.

29. Chiefari, J., Chong, Y. K., Ercoles, F., Krstina, J., Jeffery, J., Le, T. P. T., Mayadunne,

R. T. A., Meijs, G. F., Moad, C. L., Moad, G., Rizzardo, E., Thang S.H.;

Macromolecules 1998, 31, 5559.

30. Rizzardo, E.; Thang, S. H.; Moad, G. 98-AU569 WO 9905099 July 20, 1998

31. Hawthorne, D. G., Moad, G., Rizzardo, E., Thang S.H.; Macromolecules 1999, 32,

5457.

32. Chong, Y. K., Krstina, J., Le, T. P. T., Moad, G., Postma, A., Rizzardo, E., Thang, S.

H.; Macromolecules 2003, 36, 2256.

33. Vana, P., Davis, T. P., Barner-Kowollik, C.; Macromol. Theor. Simul. 2002, 11, 823.

34. Barner-Kowollik, C., Vana, P., Quinn, J. F., Davis, T. P.; J. Polym. Sci. Pol. Chem.

2002, 40, 1058.

35. Monteiro, M. J., de Brouwer, H.; Macromolecules 2001, 34, 349.

36. Coote, M. L., Radom, L.; J. Am. Chem. Soc. 2003, 125, 1490.

37. Kwak, Y., Goto, A., Tsujii, Y., Murata, Y., Komatsu, K., Fukuda, T.; Macromolecules

2002, 35, 3026.

38. Kwak, Y., Goto, A., Fukuda, T.; Macromolecules 2004, 37, 1219.

39. Butte, A., Storti, G., Morbidelli, M.; Macromolecules 2000, 33, 3485.

40. Monteiro, M. J., Sjoberg, M., Van der Vlist, J., Gottgens, C. M.; J. Polym. Sci. Pol.

Chem. 2000, 38, 4206.

41. de Brouwer, H., Tsavalas, J. G., Schork, F. J., Monteiro, M. J.; Macromolecules 2000,

33, 9239.

42. Gaillard, N., Guyot, A., Claverie, J.; J. Polym. Sci. Pol. Chem. 2003, 41, 684.

43. Smulders, W., Gilbert, R. G., Monteiro, M. J.; Macromolecules 2003, 36, 4309.

Chapter 2

26

44. Quinn, J. F., Davis, T. P., Rizzardo, E.; Chem. Commun. 2001, 1044.

45. Odian G.; Principles of polymerization. New York: Wiley; 1991

46. Liu, Y., Pan, C.; European Polymer Journal 1998, 34, 621.

47. Lou, X., Detrembleur, C., Jerome, R.; Macromolecules 2002, 35, 1190.

48. Peters, M. A., Belu, A. M., Linton, R. W., Dupray, L., Meyer, T. J., DeSimone, J. M.; J.

Am. Chem. Soc. 1995, 117, 3380.

49. Quirk, R. P., Lizarraga, G. M.; Macromolecules 1998, 31, 3424.

50. Tohyama, M., Hirao, A., Nakahama, S., Takenaka, K.; Macromol. Chem. Physic. 1996,

197, 3135.

51. Reed, S. F., Jr.; J. Polym. Sci. Pol. Chem. Ed. 1972, 10, 2493.

52. Zammit, M. D., Davis, T. P., Haddleton, D. M., Suddaby, K. G.; Macromolecules 1997,

30, 1915.

53. Vana, P., Davis, T. P., Barner-Kowollik, C.; Aust. J. Chem. 2002, 55, 315.

54. Otsu, T., Yoshida, M.; Makromol. Chem. Rapid. 1982, 3, 127.

55. Otsu, T., Yoshida, M., Tazaki, T.; Makromol. Chem. Rapid. 1982, 3, 133.

56. Otsu, T.; J. Polym. Sci. Pol. Chem. 2000, 38, 2121.

57. Pradel, J. L., Boutevin, B., Ameduri, B.; J. Polym. Sci. Pol. Chem. 2000, 38, 3293.

58. Malz, H., Komber, H., Voigt, D., Hopfe, I., Pionteck, J.; Macromol. Chem. Physic.

1999, 200, 642.

59. Haddleton, D. M., Waterson, C.; Macromolecules 1999, 32, 8732.

60. Zhang, X., Matyjaszewski, K.; Macromolecules 1999, 32, 7349.

61. Li, L., Wang, C., Long, Z., Fu, S.; J. Polym. Sci. Pol. Chem. 2000, 38, 4519.

62. Coessens, V., Matyjaszewski, K.; J. Macromol. Sci Pure 1999, A36, 811.

63. Coessens, V., Matyjaszewski, K.; J. Macromol. Sci Pure 1999, A36, 653.

64. Zhou, P., Chen, G. Q., Hong, H., Du, F. S., Li, Z. C., Li, F. M.; Macromolecules 2000,

33, 1948.

65. Koulouri, E. G., Kallitsis, J. K., Hadziioannou, G.; Macromolecules 1999, 32, 6242.

66. Lai, J. T., Filla, D., Shea, R.; Macromolecules 2002, 35, 6754.

67. Mayadunne, R. T. A., Rizzardo, E., Chiefari, J., Krstina, J., Moad, G., Postma, A.,

Thang, S. H.; Macromolecules 2000, 33, 243.

68. Liu, J., Hong, C. Y., Pan, C. Y.; Polymer 2004, 45, 4413.

69. Mengerink, Y., Peters, R., van der Wal, S., Claessens, H. A., Cramers, C. A.; Journal of

Chromatography, A 2002, 949, 337.

70. Gorbunov, A., Trathnigg, B.; Journal of Chromatography, A 2002, 955, 9.

Literature review

27

71. Philipsen, H. J. A., Klumperman, B., van Herk, A. M., German, A. L.; Journal of

Chromatography, A 1996, 727, 13.

28

Synthesis of hydroxyl-telechelic polymethacrylates by RAFT

29

CHAPTER 3

SYNTHESIS OF HYDROXYL-TELECHELIC POLYMETHACRYLATES BY RAFT

I) Introduction

The first step into the thorough investigation of the network formation mechanisms and final

network properties of poly(meth)acrylates-based coatings is the synthesis of the building blocks

of those networks. The production of well-defined telechelic polymers will be mandatory. The

previous chapter summarized the techniques available for synthesis and concludes that the

RAFT polymerization process should be used in order to achieve our purposes.

In this chapter, the synthesis and characterization of RAFT poly(meth)acrylates with a

hydroxyl end-group and a two-step post-polymerization procedure, modifying the thiocarbonyl

thio into an hydroxyl end-group, will be discussed.

II) Materials

Monomers (methyl methacrylate, butyl methacrylate) of the highest purity available were

purchased from Aldrich. They were distilled over CaH2 prior to utilization in order to remove

the inhibitor. Toluene (Biosolve) and acetone (Aldrich) were distilled over CaH2 before use.

Tetrahydrofuran (THF, Biosolve) was refluxed over LiAlH4 (Aldrich) prior to use.

Azobis(isobutyronitrile) (AIBN, Merck) was recrystallized from methanol. 4,4’-Azobis(4-

cyanopentanol) (ACP) was prepared according to the method described by Clouet et al.1 with a

purity superior to 99% (no other peaks than the expected ones were observed by 1H-NMR in

CDCl3). 1H-NMR (CDCl3): δ (ppm) 3.7 (m, 4H) 2.1-2.4 (m, 4H) 2.0 (s, 2H) 1.5-1.8 (m, 4H) 1.7

(s, 6H). 2-Cyanoprop-2-yl dithiobenzoate (RAFT-AIBN) was synthesized according to the

procedure described by Le et al.11 with a purity exceeding 99% (no other peaks than the

expected ones were observed by 1H-NMR in CDCl3) 1H-NMR (CDCl3): δ (ppm) 7.4-7.9 (m, 5H)

Chapter 3

30

1.9 (s, 6H). The synthesis of (4-cyano-1-hydroxypent-4-yl) dithiobenzoate (RAFT-ACP)

followed the route proposed by Rizzardo et al.18 A purity of 97% (1H-NMR in CDCl3) was

obtained. 1H-NMR (CDCl3): δ (ppm) 7.4-7.9 (m, 5H) 3.7 (m, 2H) 2.1-2.4 (m, 4H) 1.9 (s, 3H)

1.6 (s, 1H). The major byproduct of the synthesis is the combination product of two ACP-

derived radicals which give various peaks in the 1.0 - 3.0 ppm region. Other chemicals were

purchased from Aldrich and used without further purification.

Polymerization procedure. All polymerizations were carried out in three-neck round bottom

flasks, heated in an oil bath. The reaction solution was degassed with argon and three freeze-

pump-thaw cycles were applied. A typical polymerization was performed as follows.

Polymerization of methyl methacrylate. (Step 1 – Scheme 1) Methyl methacrylate (30.1 g, 3.01

× 10-1 mol), ACP (0.076 g, 3.02 × 10-4 mol), RAFT-ACP (0.40 g, 1.51 × 10-3 mol) and toluene

(89.4 g, 103.4 mL) were mixed in a three-neck round bottom flask. Oxygen was removed and

the flask was immersed in an oil bath that was preheated at 70 ºC. After 8 hrs of polymerization,

PMMA (PMMA1) was isolated by precipitation in heptane (conversion (GC) : 78%). 1H-NMR (CDCl3): δ (ppm) 7.4-7.9 (phenyl group of the thiocarbonyl thio ester), 3.7 (CH3 of the

methyl ester group), 1.7-2.1 (CH2 of the polymer backbone), 1.6 (CH3 of the leaving group of

the RAFT agent), 1.2-1.6 (CH2 of the leaving group of the RAFT agent), 0.9-1.1 (CH3 of the

methacrylic group).

UV-VIS: Absorbance bands 220 nm (ester group, very strong), 300nm (thiocarbonyl thio group,

strong)

Aminolysis of the thiocarbonyl thio end groups in the RAFT polymers. Thiol-functional

polymers were obtained under basic conditions by cleavage of the dithioester. Oxygen was

excluded from the reactive system by three freeze-pump-thaw cycles. In some experiments, a

few drops of an aqueous solution of Na2S2O4 were added, in order to prevent the conversion of

the corresponding thiols to disulfides.

Aminolysis of poly(methyl methacrylate). PMMA1 (Mn = 17,000g.mol-1, 5.0 g, 2.94 × 10-4 mol)

was dissolved in 50 mL of THF followed by the addition of 1-hexylamine (0.032 g, 3.24

×10-4 mol) at room temperature. The reaction was allowed to proceed during an hour. A fast and

noticeable color change took place (discoloration from pink to yellow), and the final colorless

polymer (PMMA2) was isolated by precipitation in heptane.


31

1H-NMR (CDCl3): δ (ppm) 3.7 (CH3 of the methyl ester group), 1.7-2.1 (CH2 of the polymer

backbone), 1.6 (CH3 of the leaving group of the RAFT agent), 1.2-1.6 (CH2 of the leaving group

of the RAFT agent), 0.9-1.1 (CH3 of the methacrylic group).

UV-VIS: Absorbance bands at 220 nm (ester group, very strong), 300nm (thiocarbonyl thio

group, very weak)

As will be discussed below in this chapter, in some experiments, a few drops of an aqueous

solution of Na2S2O4 were added to the solution. The procedure with addition of Na2S2O4

performed on PMMA1 lead to PMMA3

Michael addition on the thiol-functional PMMA. Thiol-functional polymers were treated

under basic conditions in the presence of hydroxylethyl acrylate and a few drops of an aqueous

solution of Na2S2O4 were added, in order to prevent the conversion of the corresponding thiols

to disulfides. Oxygen was excluded from the reactive system by three freeze-pump-thaw cycles.

Michael addition on thiol-functional PMMA PMMA3 (Mn = 17,000 g.mol-1, 5.0 g, 2.94 × 10-4

mol) was dissolved in dimethyl sulfoxide (50 mL). Benzyltrimethylammonium hydroxide (40

wt. % solution in methanol) (1.22 × 10-3 g, 7.31 × 10-6 mol) and hydroxylethyl acrylate (0.034 g,

3.23 × 10-3mol) were added. A few drops of an aqueous solution of Na2S2O4 were mixed with

the solution. The reaction was allowed to proceed at room temperature overnight under argon.

The polymer (PMMA4) was then isolated by precipitation in heptane. The removal of the

catalyst was achieved by washing with water.

III) Characterization techniques

1H-NMR and 13C-NMR spectra were recorded on a Varian-400 spectrometer using TMS as an

internal standard.

Monomer conversion was determined from the concentration of residual monomer, measured

using a Hewlett-Packard (HP 5890) GC, equipped with an AT-Wax capillary column (30 mm ×

0.53 mm × 10 µm); toluene was used as internal reference.

SEC (Size-Exclusion Chromatography) was carried out using a Waters model 510 pump and a

model 410 refractive-index detector (at 313 K). The columns used were a PL-gel guard column

(5 µm particles) 50 × 7.5 mm, followed by two PL-gel mixed-C (5 µm particles) 300 × 7.5 mm

columns (313 K). THF was used as the eluent at a flow rate of 1 mL/min. Low–polydispersity

polystyrene standards (Polymer Labs) with molecular weights ranging from 580 to 7.1 × 106

g·mol-1 were used for calibration of the columns. Molecular weights were recalculated using the

Chapter 3

32

Mark-Houwink parameters of PS, PMMA, PBMA and PBA in THF (PS: K = 1.140× 10-4 dL/g,

a = 0.716; PMMA: K = 0.944 × 10-4 dL/g, a = 0.719; PBMA: K = 1.480 × 10-4 dL/g, a = 0.664;

PBA: K = 1.220 × 10-4 dL/g, a = 0.700). The samples (1 mg/mL THF) were filtered through a

0.2 µm syringe filter prior to injection. Data acquisition and processing were performed using

Waters Millennium32 (v3.00) software.

MALDI-TOF-MS (Matrix-Assisted Laser-Desorption Ionization – Time-Of-Flight – Mass

Spectrometry) measurements were performed on a Voyager-DE STR instrument (Applied

Biosystems, Framingham, MA, USA) equipped with a 337-nm nitrogen laser. Positive-ion

spectra were acquired in the reflector mode. 2-[(2E)-3-(4-tert-butylphenyl)-2-methylprop-2-

enylidene]malononitrile was used as the matrix. All samples were hand-spotted on the target.

Sodium trifluoroacetate (CF3CO2Na) was added.

A Waters (Milford, MA, USA) 2690 Alliance liquid chromatography system was used to

perform the isocratic LC experiments. This HPLC instrument contained a built-in auto-injector

with a sample loop allowing injection of variable sample volumes, and was equipped with a

Waters 996 PDA (photodiode-array detector) and a Sedex 55 evaporative light-scattering

detector (ELSD, temperature 62 °C, N2 pressure 2.2 bar). The mobile phase was prepared in-situ

using the solvent-mixing capability of the instrument. The data collection and the data analysis

were handled by Waters Millennium 3.2 software. The columns used (150 mm x 4.6 mm i.d.)

were packed in-house with Hypersil Silica (3 µm particles; 100 Å pore size; Shandon, Runcorn,

UK). PMMA separations were performed at 25°C using the following mobile phase: 42 v/v%

acetonitrile in dichloromethane, at a flow rate of 0.5 mL/min.

Thermogravimetric analysis (TGA) traces of polymers were recorded on a Pyris 6 TGA

instrument from Perkin Elmer with a temperature ramp of 10ºC/min from 40 to 500ºC under

nitrogen atmosphere.


33

IV) Results and discussion

The results of the RAFT polymerizations of methacrylates (first step of Scheme 1) are

reported in Table 3.1. All polymerizations exhibited the characteristics of controlled systems,

e.g. a linear increase of Mn with conversion and low polydispersity throughout the reaction.

Monomer Initiator CTA Conversion Mn,exp (Mn, th) Mw/Mn

MMA AIBN RAFT-AIBN 82% 2500 (2000) 1.27

MMA ACP RAFT-ACP 78% 2400 (2000) 1.21

MMA ACP RAFT-ACP 73% 17000 (15000) 1.16

BMA ACP RAFT-ACP 78% 3200 (4000) 1.22

All polymerizations were carried out at 70ºC, for 8 hrs, in toluene under argon, with [Monomer] = 2.2 M,

[Initiator] = 0.035 M and [Initiator] : [CTA] = 1:5 ; Mn, th = (([Monomer]0/[CTA]0 ) × conversion ×

Mmonomer ) + MCTA

Table 3.1. Conversions and molecular weights of the polymethacrylates synthesized by RAFT

polymerization

In order to achieve the synthesis of hydroxyl end-functionalized telechelic polymers, the

thiocarbonyl thio group had to be modified. A possible procedure to achieve this is proposed in

Scheme 3.2. According to the existing literature, thiocarbonyl thio groups undergo fast reactions

with amines17, leading to their reduction to thiols.

After polymerization, all the polymers synthesized by RAFT were treated with a slight excess

of 1-hexylamine. The efficiency of this reaction was determined by 1H-NMR. As shown in

Figure 3.3, no more aromatic protons of the thiocarbonyl thio group were detected in the 1H-

NMR spectrum, indicating a complete cleavage of the thiocarbonyl thio function.

Chapter 3

34

S S

CN

OH

S S (Polymethacrylate)

methacrylate, Toluene, 70ºC

CN

OH

hexylamine, THF, RT

CN

OHHS (Polymethacrylate)

Michael addition with HEA

CN

OHS (Polymethacrylate)

RAFT-ACP

O

OHO

CN

OHNN

CN

HO

+

ACP

Scheme 3.2. Proposed mechanism for the synthesis of hydroxyl-telechelic polymethacrylates

This complete cleavage was confirmed by other techniques. 13C-NMR showed the

disappearance of the peak around 220 ppm characteristic of C6H5CS2. UV-absorbance spectra of

the RAFT polymer before and after aminolysis were recorded as well. The absorbance band

around 300 nm observed for the polymer prior to aminolysis, (due to the thiocarbonyl thio

group) was barely seen after reaction.


35

Figure 3.3. 1H-NMR spectra in CDCl3 of a PMMA (general formula: CNC(CH3)2(C5H8O2)n CS2

C6H5) prepared by RAFT polymerization, before (1) and after (2) aminolysis

It was also noted that this post-polymerization treatment removed the red color, which is

characteristic of the presence of a thiocarbonyl thio group in the polymer chain.

However, some questions arose following the SEC characterization of the polymers. In the

chromatograms of the polymers obtained after cleavage of the thiocarbonyl thio group,

secondary populations were detected in the higher molecular weight region (Figure 3.4).

Figure 3.4. SEC chromatograms of PMMA synthesized by RAFT before and after aminolysis

(1)

(2)

8 7 6 5 4 3 2 1 0 Chemical shifting (ppm)

8.0 7.9 7.8 7.7 7.6 7.5 7.4Chemical shifting (ppm)

8.0 7.9 7.8 7.7 7.6 7.5 7.4Chemical shifting (ppm)

2.8 3.0 3.2 3.4 3.6 3.8 4.0 4.2 4.4 4.62.8 3.0 3.2 3.4 3.6 3.8 4.0 4.2 4.4 4.6

Log M

S S (PMMA)

CN

OH

CN

OHHS (PMMA)

Chapter 3

36

This phenomenon has also been recently reported by Wang et al.20 These shoulders

correspond to polymers with molecular weights approximately twice higher than the expected

values. The presence of these species is most likely due to the oxidative coupling of thiols,

which results in the formation of a disulfide bridge between two polymer chains.

In order to confirm this hypothesis, a critical-liquid-chromatography method was developed6.

In a general sense, critical-liquid-chromatography methods allowed separation of polymer

chains according to their functionality. By definition, in critical chromatography, retention is

independant of molecular weight2-5,12-16. The application of this method to our study is governed

by the need of a polymer separation based on the hydroxyl functionality. With a suitable specific

interaction between the column and the hydroxyl groups, an elution time dependant on the

number of OH functional groups should be obtained. The method should be rather robust even

though it operates under critical conditions, and should not be dramatically dependant on

parameters such as solvent composition, flow rate or temperature, for example. Due to the

presence of thiocarbonyl thio or thiol group, the method should also be relatively insensitive to

the presence of other functional groups in the polymer chains.

The solvent mixture should typically consist of a poor eluent for the polymer, and another

solvent capable of desorbing the polymer chain from the column. For its specific interaction

with the polar group (e.g. hydroxyl), bare silica is chosen as the column material.

0

5

10

15

20

25

34 39 44 49 54 59 64 69

v/v % Acetonitrile

Ret

entio

n tim

e (m

in)

PMMA 1680PMMA 3800PMMA 6950PMMA 13900PMMA 28300

(A)


37

Figure 3.5. (A) Dependence of retention time on the mobile-phase composition for PMMA

standards with different molecular weights. (B) Dependence of retention time on the mobile-

phase composition for PMMA-OH and HO-PMMA-OH samples with different molecular

weights. ELSD Note: Data points with retention times larger than 10 min in figure A (15 min in

figure B). were recorded and used to construct the curved lines. However, they are not shown in

the figures.

Dichloromethane is known to be a poor eluent for PMMA, it will then be mixed with

acetonitrile, a much more polar solvent, for the desorption of PMMA chains from the column.

The optimal ratio of the two for operating in critical condition has to be found via a systematic

variation as shown below.

Some non-functional, mono-functional and bi-functional PMMA samples of different molar

masses were used for the determination of critical conditions. It can be seen in Figure 3.5 (A)

that the retention time is independent of the molar mass of the analyzed sample for non-

functional PMMA when the acetonitrile content is between 43% and 55%. On Figure 3.5 (B), it

is observed that around 43% of acetonitrile content, the retention time is also independent of the

molar mass for hydroxyl-mono-functional and di-functional PMMA. It is then concluded that

the critical mobile phase composition for the separation of hydroxyl-functional PMMA should

be a 43:57 ratio mix of acetonitrile:dichloromethane.

0

5

10

15

20

25

34 39 44 49 54 59 64 69v/v % Acetonitrile

Ret

entio

n Ti

me

(min

)

PMMA-OH 500PMMA-OH 3300PMMA-OH 6680PMMA-OH 13900PMMA-OH 20000PMMA-2OH-1450PMMA-2OH-5718

(B)

Chapter 3

38

Figure 3.6 (A) Calibration plots of log Mp vs. retention time for PMMA standards in

different mobile phase mixtures.

(B)Calibration plots of log Mp vs. retention time for PMMA samples with one OH end-group in

different mobile phase mixtures (The values as indicated in the figure refer to the volume

percentage of acetonitrile in dichloromethane).

3

3.2

3.4

3.6

3.8

4

4.2

4.4

4.6

2 4 6 8 10 12 14 16 18 20Retention time (min)

Log

Mp

343638404348556070

Acetonitrile (%)

(A)

2.5

2.7

2.9

3.1

3.3

3.5

3.7

3.9

4.1

4.3

2 7 12 17 22Retention time (min)

Log

Mp

343638404348556070

Acetonitrile (%)

(B)


39

Another display of the critical condition determination experiments is presented in Figure 3.6.

The same conclusion applies to this graph: non-functional and mono-functional PMMA are

separated based on their functionality, independent of their molar mass when a 43:57 ratio mix

of acetonitrile : dichloromethane is employed

Using this critical-liquid-chromatography method, we were able to confirm the presence of

polymer chains with two hydroxyl end groups in PMMA2. These can be formed by the coupling

of two thiol-terminated polymer chains during aminolysis, resulting in the formation of a

disulfide bond. As depicted in Figure 3.7 (Trace 2), the reaction of a hydroxyl-functional

PMMA synthesized by RAFT polymerization with 1-hexylamine resulted in a polymer with a

noticeable fraction of difunctional chains, confirming the hypothesis of the formation of a

disulfide bond. The small fraction of non-functional polymer might be the result of transfer by

hydrogen abstraction to solvent or polymer chain during the polymerization.

Figure 3.7. Isocratic Liquid-chromatography separations of PMMA samples based on their OH-

functionality (1) PMMA1: PMMA (Mn=17.000 g/mol, PDI=1.16), synthesised by RAFT

polymerization (2) PMMA2: PMMA1 after reaction with 1-hexylamine in the absence of

sodium bisulfite; (3) PMMA3: PMMA1 after reaction with 1-hexylamine in the presence of

sodium bisulfite; (4) PMMA4: PMMA3 after Michael addition reaction with HEA in the

presence of DABCO and Na2S2O4

0-OH 1-OH 2-OH

(1)

(3)

(2)

0 50 100 150 200 250 300

Elution time (s)

(4)

Chapter 3

40

2920 2940 2960 2980 3000 3020 3040 3060

m/z

2970.67 CN

OHHS (PMMA) CN

OHH (PMMA)

CN

OHH (PMMA)

2938.78 3038.88

Although all required precautions were taken (distillation of the solvent, removal of the

oxygen by freeze-pump-thaw cycles), all RAFT polymers made by us seemed to be prone to

such an oxidative coupling after the treatment with 1-hexylamine. The observation of shoulders

on the GPC chromatograms is the basis for this affirmation. Nevertheless, the coupling of the

thiols can be suppressed by the addition of an anti-oxidant (aqueous solution of sodium bisulfite

– Na2S2O4) to the reactive medium. This is confirmed by the suppression of the shoulder on the

SEC chromatograms (not shown) and the detection of only monofunctional polymer in the LC

(Figure 3.7, Trace 3). It has to be noted that Wang et al.20 proposed a reductive system with zinc

and acetic acid to prevent the oxidative coupling of thiols to disulfides. However, the necessity

of maintaining the alcohol functions in our system prohibits the use of a compound with a

carboxylic acid function, and therefore we had to employ a different method to suppress the

coupling reaction.

Another side reaction of the nucleophilic cleavage of thiocarbonyl thio groups was revealed

by MALDI-TOF-MS experiments on the polymers after treatment with a base (Figure 3.8) and

by ESI-QTOF-MS (Electrospray-ionization-quadrupole-TOF)7.

Figure 3.8. MALDI-TOF-MS spectra of PMMA RAFT-polymer after aminolysis; counter ion =

Na+

As expected, the family of peaks corresponding to the thiol-terminated polymers was found,

with a difference between the peaks of 100 mass units, corresponding to the MMA moiety.

However, an unexpected second family of peaks was also observed, higher than the previous

one by 32 mass units. The assignment of the peaks reveals a hydrogen-atom as an end group


41

instead of the expected thiol-group. Following the conclusions of Ladaviere et al.10, a hydrogen-

radical transfer occurring during the treatment with the base can partially explain the presence of

this family of peaks. Another possible explanation will be given below. Another family was

detected with a difference of 40 m/z relative to the hygrogen-terminated one. So far, we cannot

propose a satisfactory structure assignment for this family. The disulfide-containing chains

could not be observed by mass-spectrometry techniques. We assume that the weak sulfur-sulfur

bond was broken during the analysis.

The second end group modification step of Scheme 3.2 relies on the ability of thiol

compounds to bind to a single molecule of a vinylic compound with an activated double bond

(e.g. acrylates). This Michael addition reaction is known to be essentially quantitative for

organic compounds9,19. In our polymerizations using ACP and RAFT-ACP as initiator and

RAFT agent, respectively, almost all the chains bearing a thiocarbonyl thio function possess an

alcohol function at the other end of the chain (according to the mechanism described by

Rizzardo5, and confirmed by our analysis, Figure 3.7). After the aminolysis of these RAFT

polymers, the thiol end groups obtained can be further reacted with a hydroxyl-functional

acrylate (e.g. 2-hydroxyethyl acrylate - HEA). The side reaction of the aminolysis that results in

hydrogen-terminated chains will have a major impact on the ultimate yield of bifunctional

hydroxyl-terminated polymers: the hydrogen-terminated chains can not be converted to a

hydroxyl-functional group during the Michael addition. This results in a lower fraction of

telechelic polymers. Some dead material resulting from the RAFT polymerization itself can be

an obstacle as well to the production of these telechelic material in a high yield.

In the second step of our two-step post-polymerization procedure (Scheme 3.1), the thiol-

functional polymer reacts with a hydroxyl-functional acrylate in a Michael addition reaction.

The yield of this Michael reaction could not be determined accurately by 1H-NMR, due to the

low intensities of the signals arising from the protons of the end groups and/or their overlap with

signals corresponding to the protons of the polymer backbone. Quantitative results were

obtained with liquid-chromatography (LC) techniques. Figure 3.5 shows the LC chromatograms

of a RAFT PMMA after each of the two steps of the post-polymerization treatment. The

cleavage of the thiocarbonyl thio function in the presence of sodium hydrosulfite (Trace 3)

results in a monofunctional polymer (as described above). The Michael addition of HEA on the

thiol-functional polymer in the presence of Na2S2O4 is reflected in the chromatogram by the

appearance of a large peak in the 2-OH region. However, the yield achieved (66.7 % - Table

3.9) was relatively low compared to those obtained with model compounds (for example, the

addition of HEA on dodecanthiol was nearly quantitative).

Chapter 3

42

Sample % 0-OH % 1-OH % 2-OH

PMMA1 0.8 99.2 0.0

PMMA2 0.8 65.2 34.0

PMMA3 0.7 99.3 0.0

PMMA4 0.2 33.1 66.7

PMMA1: PMMA (Mn=17.000 g/mol, PDI=1.16), synthesised by RAFT polymerization; PMMA2:

PMMA1 after reaction with 1-hexylamine in the absence of sodium bisulfite; PMMA3: PMMA1 after

reaction with 1-hexylamine in the presence of sodium bisulfite; PMMA4: PMMA3 after Michael addition

reaction with HEA in the presence of DABCO and Na2S2O4.

Table 3.9. Quantitative differentiation of PMMA-functional chains, according to their OH-

functionality, by icocratic liquid-chromatography experiments.

This low yield can be explained by several factors. The termination by disproportionation that

occurs during the RAFT polymerization leads to dead chains without the thiocarbonyl thio

function as an end group. The two-step post-polymerization treatment will not affect those

chains, which will consequently remain monofunctional. A solution to partly circumvent this

problem would be to stop the polymerizations at a lower conversion, limiting the fraction of

dead chains in the polymerization product.

Another explanation for the low yield is the formation of hydrogen-terminated chains as a side

reaction during the aminolysis of the thiocarbonyl thio groups. These cannot be functionalized

further by the Michael addition. The accessability of the SH group may also be a point of

concern. A coiling effect may prevent the reaction from occurring quantitatively, lowering the

percentage of α,ω - telechelic polymer obtained. As a final remark, we could not accurately

estimate the yield of the Michael addition of the thiol-functional poly(methacrylate). After the

aminolysis, the hydrogen-terminated and the thiol-terminated chains could not be differentiated

by an analytical technique in a quantitative way. Some side-reactions might also occur during

this Michael addition reaction, lowering the ultimate yield of di-hydroxy telechelic chains.

As an interesting remark, we want to add that the two-step post polymerization procedure

resulted into telechelic polymers without a labile thiocarbonyl thio function. Apart from the

disappearance of the red color, the absence of this function enhances the thermal stability (the

thermogravimetric traces of a RAFT-PMMA and a telechelic PMMA are presented in Figure

3.10).


43

Figure 3.10. Thermogravimetric analysis (TGA) traces of RAFT PMMA (1)

(HO(CH2)3CNC(CH3)(C5H8O2)nCS2C6H5) and telechelic PMMA (2)

(HOCH2CH2OOCCH2CH2S(C5H8O2)nCCN(CH3)(CH2)3OH)

Due to the lack of availability of reference materials, a similar liquid-chromatography

technique could not be developed yet for PBMA. In this case, MALDI-TOF-MS was employed

to investigate the outcome of the same two-step post-polymerization procedure. The MALDI-

TOF-MS spectrum obtained for a PBMA RAFT-polymer after aminolysis and Michael addition

with HEA is presented in Figure 3.11. Although it is impossible to draw any quantitative

conclusions from the MALDI-TOF-MS, we can clearly observe that a significant amount of

telechelic polymer chains with two hydroxyl end groups, corresponding to the expected

structure, were obtained.

100 200 300 400 500

0

20

40

60

80

100

Temperature (ºC)

(1)

(2)

Chapter 3

44

Figure 3.11. Experimental (1) and theoretical (2) MALDI-TOF-MS spectra of PBMA RAFT-

polymer after reaction with 1-hexylamine and HEA; counter ion = Na+

2416.57 2558.66

(2)

(1)

2400 2420 2440 2460 2480 2500 2520 2540 2560

m/z (Da)

2400 2420 2440 2460 2480 2500 2520 2540 2560

m/z (Da)

2416.93

2559.01

2000 3000 4000

m/z (Da)

CN

OHS (PMMA)

O

OHO


45

Comprehensive 2D LCxSEC separation

In the course of synthesizing RAFT PMMA, some preliminary experiments were made using

AIBN as an initiator rather than ACP. The polymer chains obtained from a AIBN/RAFT-ACP

initiator/RAFT agent combination could have initiated during the polymerization by two types

of radicals (neglecting chain transfer or side-reactions): either the leaving group of the RAFT

agent (bearing a hydroxyl function) or the fragment derived from the AIBN (non-hydroxy

functional). The LC method was employed for the analysis of this polymer (VL37A - Mn= 2.600

g/mol, PDI = 1.29 – Figure 3.12).

Figure 3.12. Separation of functional RAFT polymers according to the number of OH groups.

Evaporative light-scattering detector, mobile phase 43 % acetonitrile in dichloromethane, home-

packed silica column (150 mm × 4.6 mm i.d.; 3-µm particles; 10nm pore size), flow rate 1

mL/min.

A step further was taken with the decision to analyze this sample in a second dimension. The

LC analysis provided useful information on the functionality distribution, the chains being

separated according to the number of hydroxyl groups they were bearing, however no

information on the molar mass averages could be collected with this method. The determination

of common characteristics of synthetic polymers such as the different molar masses (Mn, Mw,

Mz) or the polydispersity index required an extra chromatographical analysis (SEC). A method

consisting of two-dimensional liquid chromatography providing information on both the

functionality-type distribution (FTD) and the molecular weight distribution (MWD) was then

developed8.

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0 50 100 150 200 250 300 350Time (s)

ELS

D re

spon

se

VL37

VL37B

0-OH 2-OH1-OH

A

Chapter 3

46

The details on the elaboration of the method can be found in reference 8. In this chapter, the

focal point of the discussion will be given to the interpretation of the chromatograms obtained

by LCxSEC.

Figure 3.13. Two-dimensional LC×SEC chromatogram of sample VL37A. (a) UV detection at

300 nm. (b) ELSD. The LC mobile phase was 50 % ACN in DCM and the flow rate was 4

µL/min.

In Figure 3.13, the UV detection reveals the presence of 4 distinct species in polymer

VL037A. The following assignment can be made:

- Peak #1: polymers having a relatively high molecular weight and no OH-functionality.

Polymer chains initiated by a non-functional AIBN fragment are probably detected with

this peak. The growth of the chain has been controlled by the thiocarbonyl thio

compounds in a RAFT process, after the initiation by the AIBN-derived radical. Some

chains initiated by an alkyl group after a chain transfer reaction might as well be present

in the population represented by this peak. It corresponds mainly to the peak in the 0-OH

region on Figure 3.12.

- Peak #2: relatively high molar masses are observed with hydroxyl functionality. The

typical RAFT recipes result in majority into polymer chains initiated by the leaving group

of the RAFT agent. In our polymerization system, the RAFT agent bears a hydroxyl

(a)

(b)


47

CN CN

group, hence it is logical to find a peak in this position. The population responsible for this

peak is mainly the one causing the peak in the 1-OH region on Figure 3.12.

- Peak #3: low molar masses compounds with no hydroxyl-functionality are observed.

Several explanations might be given for this compound. During the reaction, the AIBN

decomposes but has only an initiation efficiency of 0.6-0.7 (depending on the exact

conditions), which means that 30 to 40 % of the produced radicals will not initiate new

growing chains, but will terminate form TetraMethyl SuccinodiNitrile (TMSN).

TMSN

However, the absence of a chromophore prevents the TMSN molecule from being

decteted. Another possibility to explain the presence of peak #3 is the termination of very

short chains (a couple of monomer units added) at the beginning of the polymerization.

Following the initiator decomposition profile, the concentration of radicals is the highest

at the beginning of a radical polymerization, before the establishment of a steady state. It

is then the moment when the termination probability is the highest. The transition region

between peak #1 and peak #3 can be explained by termination of chains throughout the

polymerization. Once again, the termination procedure results in chains without a

thiocarbonyl thio function at the end. Hence, it forbids the use of this argument as a valid

explanation.

This peak might also be caused by the recombination of a thiocarbonyl thio radical and a

radical derived from the AIBN. This forms a RAFT-AIBN compound.

- Peak #4: some low molar masses compounds with an hydroxyl function are detected. LC-

ESI-MS experiments showed the presence of unreacted RAFT agent in the final product

obtained after polymerization. Short chains with one or two monomer units were also

detected via mass spectrometry.

The presence of peak #1 and peak #2 were expected and are in line which what should be

expected from a RAFT polymerization using a hydroxyl-functional RAFT agent and a non-

hydroxyl functional initiator. The peaks #3 and #4 are more intriguing as it seems improbable

that a RAFT agent (with a transfer constant that can be estimated to be around 100) can remain

untouched after 8 hours of polymerization at 70˚C. No satisfactory explanation can be provided

so far.

Chapter 3

48

Quantitative results concerning molar masses and concentration for peak #1 and peak #2 were

determined by both ELSD and UV detection. Data for peaks #3 and #4 are not presented, due to

the low molar mass of the chains, that were below the lowest standard used for calibration of the

SEC (Mn = 620 g/mol). Their combined molar concentrations represent approximately 1% in

sample VL037A. The quantitative results are presented in Table 3.14 (ELSD calibration details

presented in Reference 8).

UV, 300nm Calibrated ELSD

Peak name Mn,

kg/mol

Mp,

kg/mol PDI

Conc.*

mol %

Mn,

kg/mol

Mp,

kg/mol PDI

Conc.*

mol %

Peak #1 2.3 2.3 1.28 16 2.1 2.5 1.24 9

Peak #2 2.7 3.2 1.28 84 2.4 2.9 1.27 91

Peaks #1 and

#2

(combined)

2.6 3.0 1.30 N/A 2.3 2.9 1.28 N/A

* Combined molar concentration of peak 3 and 4 represents less than 1 mol % of the total chains

Table 3.14. Quantitative results for two-dimensional LCxSEC separation of sample VL037A

using UV detection at 300 nm and calibrated ELSD.

Small differences in the relative concentrations of peak #1 and peak #2 can be noted by

comparing the results obtained with ELSD detection and UV detection at 300nm. A possible

explanation can be the presence in the population represented by peak #2 of chains not

terminated by the thiocarbonyl thio chromophoric group. Termination by disproportionation

occurring during the polymerization will produce hydrogen-terminated and double-bond-

terminated chains. If those chains were initiated by the RAFT leaving group, they will be

hydroxyl-functional and will contribute to the peak #2. However the absence of a chromophore

will prevent them from being detected by the UV detector, resulting in an underestimation by

ELSD of the concentration of the chains producing peak #2.

The sample VL037A was treated with 1-hexylamine in order to reduce the thiocarbonyl thio

function into a thiol. Some coupling between thiols occurs, resulting in the formation of

disulfides. Coupling of hydroxyl-functional chains will be noticed in the chromatogram by the

appearance of a peak in the dihydroxyl-functional region. (VL37B - Mn= 2.800 g/mol, PDI =


49

1.35 – Figure 3.12). The 2D-LC x SEC method was used for characterization of this sample

(Figure 3.15).

Figure 3.15. Two-dimensional LC×SEC chromatogram of sample VL37B detected by ELSD.

The analytical conditions were chosen identical to those of Figure 3.13

The detection method for the SEC analysis was restricted to ELSD due to the fact that nearly

all the chains that were bearing a chromophoric group (thiocarbonyl thio group) in VL037A

have lost this fragment in the aminolysis process. The lack of chromophores forbids the use of

UV detection in order to draw quantitative conclusions. Using ELSD, the quantitative analysis

of sample VL37B is presented on table 3.16.

Peak Name Mn, kg/mol Mp, kg/mol PDI Conc. (mol %)

Peak #1 2.4 2.5 1.25 7

Peak #2 2.6 3.0 1.31 86

Peak #3 3.2 5.2 1.35 7

Peaks #1, #2 and

#3 (combined) 2.7 3.0 1.32 N/A

Table 3.16. Quantitative results for LCxSEC separation of sample VL037B without Na2S2O4

using calibrated ELSD.

The main conclusion from the 2D-LCxSEC experiment is the straightforward observation of

the disulfide-containing chains. The double hydroxyl functionality recorded coupled with the

approximate double mass of the Mp gives in a single chromatographic experiment a double

Chapter 3

50

proof backing up the disulfide formation hypothesis. This conclusion is similar to the one drawn

after the individual SEC and LCCC experiments, as was described earlier in this chapter. After

these analytical investigations, the mechanism proposed in Scheme 3.2 can be updated with our

findings. The Scheme 3.17 acknowledges the presence of the side reactions occurring during the

aminolysis.

Scheme 3.17. Synthesis of hydroxyl-telechelic polymethacrylates

S S

CN

OH

S S (Polymethacrylate)

methacrylate, Toluene, 70ºC

CN

OH

hexylamine, THF, RT

CN

OHHS (Polymethacrylate)

Michael addition with HEA

CN

OHS (Polymethacrylate)

RAFT-ACP

O

OHO

CN

OHNN

CN

HO

+

ACP

CN

OH

CN

HO S (Polymethacrylate)S(Polymethacrylate)

H (Polymethacrylate)

CN

OH

+


51

V) Conclusion

In this chapter, a synthetic path to obtain α,ω − functional linear polymers employing the

RAFT polymerization process was proposed. With a proper choice of the initiator - RAFT agent

system, the two-step post-polymerization procedure (namely, an aminolysis of a thiocarbonyl

thio group, followed by a Michael addition on the resulting thiol) was demonstrated to yield

telechelic polymers. Liquid-chromatography separations proved the successful synthesis of

telechelic PMMA to a significant extent. The absence of the thiocarbonyl thio moiety results in

a colorless material with an enhanced thermal stability. However, the fraction of dihydroxyl-

functional PMMA chains in the final product was limited by side reactions, occurring during the

RAFT polymerization of MMA and during the aminolysis procedure. The main undesirable

reaction was found to be the formation of disulfide bridges from two thiol-functional polymer

chains. Addition of an aqueous solution of sodium bisulfite suppressed this oxidative coupling

reaction. The efficiency of this procedure was demonstrated with the help of liquid-

chromatographic separations. It was also found, using mass-spectrometric techniques, that the

reduction of thiocarbonyl thio groups into thiols is accompanied by the production of hydrogen-

terminated polymer chain. MALDI-TOF-MS spectra were used to demonstrate that RAFT

polymerization of BMA, followed by the two-step end group modification procedure, yielded

hydroxyl-functional telechelic poly(butylmethacrylate). The proposed method to synthesize

telechelic hydroxyl polymers showed disappointing results in terms of average functionality for

PMMA. The synthesized polymers will not be of satisfactory purities for the network formation

studies.

A comprehensive two-dimensional liquid chromatography technique (LCxSEC) was

developed in order to simultaneously measure the molecular-weight distribution (MWD) and

functionality-type distribution (FTD) for hydroxyl-functional PMMA. Quantitative results were

obtained for a RAFT-PMMA before and after aminolysis.

Chapter 3

52

Reference List

1. Clouet, G., Knipper, M., Brossas, J.; Polym. Bull. (Berlin) 1984, 11, 171.

2. Cools, P. J. C. H., Van Herk, A. M., German, A. L., Staal, W.; Journal of Liquid Chromatography 1994, 17, 3133.

3. Falkenhagen, J., Much, H., Stauf, W., Mueller, A. H. E.; Macromolecules 2000, 33, 3687.

4. Gorbunov, A., Trathnigg, B.; J. Chromatogr. A 2002, 955, 9.

5. Gorshkov, A. V., Much, H., Becker, H., Pasch, H., Evreinov, V. V., Entelis, S. G.; Journal of Chromatography 1990, 523, 91.

6. Jiang, X., Lima, V., Schoenmakers, P. J.; J. Chromatogr. A 2003, 1018, 19.

7. Jiang, X., Schoenmakers, P. J., van Dongen, J. L. J., Lou, X., Lima, V., Brokken-Zijp, J.; Anal. Chem. 2003, 75, 5517.

8. Jiang, X., van der Horst, A., Lima, V., Schoenmakers, P. J.; J. Chromatogr. A 2005, accepted.

9. Kharasch, M. S., Fuchs, C. F.; J. Org. Chem. 1948, 13, 97.

10. Ladaviere, C., Doerr, N., Claverie, J. P.; Macromolecules 2001, 34, 5370.

11. Le, T. P.; Moad, G.; Rizzardo, E.; Thang, S. H. 97-US12540 WO 9801478 January 15, 1998

12. Macko, T., Hunkeler, D., Berek, D.; Macromolecules 2002, 35, 1797.

13. Mengerink, Y., Peters, R., de Koster, C. G., van der Wal, S., Claessens, H. A., Cramers, C. A.; J. Chromatogr. A 2001, 914, 131.

14. Mengerink, Y., Peters, R., van der Wal, S., Claessens, H. A., Cramers, C. A.; J. Chromatogr. A 2002, 949, 337.

15. Pasch, H., Rode, K., Chaumien, N.; Polymer 1996, 37, 4079.

16. Philipsen, H. J. A., Klumperman, B., van Herk, A. M., German, A. L.; J. Chromatogr. A 1996, 727, 13.

17. Rizzardo, E., Chiefari, J., Chong, B. Y. K., Ercole, F., Krstina, J., Jeffery, J., Le, T. P. T., Mayadunne, R. T. A., Meijs, G. F., Moad, C. L., Moad, G., Thang, S. H.; Macromol. Symp. 1999, 143, 291.


53

18. Rizzardo, E.; Thang, S. H.; Moad, G. 98-AU569 WO 9905099 July 20, 1998

19. Szabo, J. L., Stiller, E. T.; J. Am. Chem. Soc. 1948, 70, 3667.

20. Wang, Z., He, J., Tao, Y., Yang, L., Jiang, H., Yang, Y.; Macromolecules 2003, 36, 7446.

54

Synthesis of carboxy-telechelic polyacrylates by RAFT

55

CHAPTER 4

SYNTHESIS OF CARBOXY-TELECHELIC POLYACRYLATES BY RAFT

I) Introduction

As described in the previous chapter, the quantitative preparation of hydroxyl-functional

telechelic PMMA was unsuccessful. Another strategy was investigated in order to prepare

telechelic polymers by RAFT polymerization with a higher purity. The previous study

concluded that the end-group modification procedure has to be avoided in order to obtain high

yields. With the use of bi-functional trithiocarbonates in conjugation with functional initiators, a

high percentage of telechelic polyacrylate is expected to be produced readily after the

polymerization, without any further modification needed. This expectation is based on a

publication from Lai et al.1 Production of telechelic polyacrylates and polystyrene was claimed

by them although the quantitative analytical characterization of those polymers was lacking. In

this chapter, we will reinvestigate the synthetic route proposed by Lai et al.1, and a particular

attention will be dedicated to the quantitative characterization for the end-groups of the

produced carboxyl-functional telechelic polymers, using liquid chromatography under critical

conditions.

II) Materials

Butyl acrylate (99 %) was purchased from Aldrich and was distilled over CaH2 prior to

utilization in order to remove the inhibitor. Toluene (Biosolve) and acetone (Aldrich) were

distilled over CaH2 before use. Azobis(isobutyronitrile) (AIBN, Merck) and 4,4’-azobis(4-

cyanovaleric acid) (ACVA, Aldrich) were recrystallized from methanol. S,S’-Bis(α,α’-

dimethyl-α’’-acetic acid)trithiocarbonate and S-1-dodecyl-S'-(α,α’-dimethyl-α’’-acetic

Chapter 4

56

acid)trithiocarbonate were prepared according to the route described by Lai et al., yielding

purities above 99% (1H-NMR in CDCl3).

Polymerization procedure. All RAFT polymerizations were carried out in three-neck round

bottom flasks, heated in an oil bath. The reaction solution was degassed with argon and three

freeze-pump-thaw cycles were applied. The final product was isolated by rotary evaporation.

Polymerization of butyl acrylate (Figure 4.1) Butyl acrylate (42.5 g, 3.32 × 10-1 mol), ACVA

(0.72 g, 2.56 × 10-3 mol), S,S’-bis(α,α’-dimethyl-α’’-acetic acid)trithiocarbonate (6 g, 2.12 ×

10-2 mol) toluene (41.1 g, 47.5 mL) and acetone (37.6 g, 47.5 mL) were mixed in a three-neck

round bottom flask. Oxygen was removed and the flask was immersed in an oil bath that was

preheated at 80 ºC. After 4 hrs of polymerization, PBA was isolated by rotary evaporation.

III) Characterization techniques

Monomer conversion was determined from the concentration of residual monomer, measured

using a Hewlett-Packard (HP 5890) GC, equipped with an AT-Wax capillary column (30 m ×

0.53 mm × 10 µm); toluene was used as internal reference.

SEC (Size-Exclusion Chromatography) was carried out using a Waters model 510 pump and a

model 410 refractive-index detector (at 313 K). The columns used were a PL-gel guard column

(5 µm particles) 50 × 7.5 mm, followed by two PL-gel mixed-C (5 µm particles) 300 × 7.5 mm

columns (313 K). THF was used as the eluent at a flow rate of 1 mL/min. Low–polydispersity

polystyrene standards (Polymer Labs) with molecular weights ranging from 580 to 7.1 × 106

g·mol-1 were used for calibration of the columns. Molecular weights were recalculated using the

Mark-Houwink parameters of PS, PMMA, PBMA and PBA in THF (PS: K = 1.140× 10-4 dL/g,

a = 0.716; PMMA: K = 0.944 × 10-4 dL/g, a = 0.719; PBMA: K = 1.480 × 10-4 dL/g, a = 0.664;

PBA: K = 1.220 × 10-4 dL/g, a = 0.700). The samples (1 mg/mL THF) were filtered through a

0.2 µm syringe filter prior to injection. Data acquisition and processing were performed using

Waters Millennium32 (v3.00) software.

MALDI-TOF-MS (Matrix-Assisted Laser-Desorption Ionization – Time-Of-Flight – Mass

Spectrometry) measurements were performed on a Voyager-DE STR instrument (Applied

Biosystems, Framingham, MA, USA) equipped with a 337 nm nitrogen laser. Positive-ion

spectra were acquired in the reflector mode. 2-[(2E)-3-(4-tert-butylphenyl)-2-methylprop-2-

enylidene]malononitrile was used as the matrix. All samples were hand-spotted on the target.

Sodium trifluoroacetate (CF3CO2Na) was added.


57

A Waters (Milford, MA, USA) 2690 Alliance liquid chromatography system was used to

perform the isocratic LC experiments. This HPLC instrument contained a built-in auto-injector

with a sample loop allowing injection of variable sample volumes, and was equipped with a

Waters 996 PDA (photodiode-array detector) and a Sedex 55 evaporative light-scattering

detector (ELSD, temperature 62 °C, N2 pressure 2.2 bar). The mobile phase was prepared in-situ

using the solvent-mixing capability of the instrument. The data collection and the data analysis

were handled by Waters Millennium 3.2 software. The columns used (150 mm x 4.6 mm i.d.)

were packed in-house with Hypersil Silica (3 µm particles; 100 Å pore size; Shandon, Runcorn,

UK). PBA separations were performed at 25°C using the following mobile phase: 6 v/v%

acetonitrile, 1 v/v % acetic acid (added as 10 v/v % acetic acid in dichloromethane) in

dichloromethane, at a flow rate of 0.5 mL/min.

The comprehensive two-dimensional LC experiments (LC×SEC) were carried out using a

Shimadzu µ-LC 10ADvp pump at a flow rate of 20 and 40 µl/min (Shimadzu, s’ Hertogenbosch,

The Netherlands). A Rheodyne two-position six-port injection valve (Berkeley, CA, USA)

equipped with a 20 µl injection loop was used for the LC separation. The stationary phase used

in the LC was Hypersil silica (Shandon, Runcorn, UK), particle size 3-µm, pore diameter 120 Å,

column dimensions were 150 × 4.6 mm I.D., column temperature was kept constant at 55ºC in a

(Millipore) Waters temperature-control module. The 2D-SEC separation system consists of a

Kratos Spectroflow 400 pump (ABI, Ramsey, NJ, USA) equipped with two 50 × 4.6 mm I.D.

PL-columns (Polymer Laboratories, 5-µm particles with 100 Å pore size and/or 6 µm oligoPore

particles with 100 Å pore size, Church Stretton, Shropshire, UK). The SEC was coupled to a

Kratos Spectroflow 757 UV-absorbance detector (ABI), λ = 230 nm, and a Varex II A

evaporative light-scattering detector (ELSD) (Burtonsville, Maryland, USA), using N2 (3 bar) as

nebulizer gas at 65 ºC, the flow rate used in the SEC was 0.9 ml/min. For comprehensive

LC×SEC, the LC and SEC systems were coupled by an air-actuated VICI two-position ten-port

injection valve (Valco, Schenkon, Switzerland). This valve was operated using a high-speed

switching accessory (switching-time of 20 ms using 5 bar N2) and dual injection loops of equal

volume (40 or 80 µl depending on the LC flow rate, t0 SEC ≈ 2 min.). Data was collected using

a Keithley KNM-DCV 12 Smartlink interface (Cleveland, OH, USA). Two-dimensional plots

and distribution data were calculated with an in-house program written in a Matlab (Natick,

MA, USA) software environment. This program allowed us to extract LC and SEC

chromatograms at any position in the LC×SEC contour plot. The Matlab software also

controlled the valve switching and recorded exact switching times for accurate slicing.

Chapter 4

58

S

SS COOHHOOC

HOOC (Polyacrylate) S

S

S (Polyacrylate) COOH

AcrylateToluene/acetone80 C 0

+ N N

CN

COOH

CN

HOOC

ACVA

IV) Results and discussion

The procedure described by Lai et al.1 was employed to synthesize telechelic carboxyl-

functional polybutylacrylate. This route has the advantage that it leads in principal to telechelic

material in one step (Figure 4.1).

Figure 4.1. Synthesis of carboxyl-telechelic polyacrylates

We decided to polymerize butyl acrylate rather than methyl methacrylate. This choice was

dictated by the poor control over MMA polymerization using S,S’-bis(α,α’-dimethyl-α’’-acetic

acid)trithiocarbonate and S-1-dodecyl-S'-(α,α’-dimethyl-α’’-acetic acid) trithiocarbonate as

RAFT agents. In MMA polymerizations using these RAFT agents, the stability of the leaving

group is not sufficient to ensure a fast fragmentation of the intermediate radical in the pre-

equilibrium of the RAFT process. This will result in polydispersities that are too high for our

purpose (approximately 1.7). The BA polymerization conditions and main results are presented

in Table 4.2.

Lai et al. claimed to have synthesized linear telechelic polymers1 with COOH end-groups.

However, the analytical characterization was limited to SEC analysis, which does not prove the

production of telechelic polymers. In order to determine the actual fraction of telechelic chains

in our polymers, we developed a new liquid-chromatography method, suitable for separation of

poly(butyl acrylates) according to their carboxyl end group-functionality, under critical

conditions (i.e. retention independent of molecular weight)2-6. The procedure to establish the

critical conditions was similar to the one exposed in Chapter III. Systematic variation of the

mobile phase composition was used in order to find the dichloromethane : acetonitrile ratio

suitable for performing liquid chromatography in critical conditions.


59

HOOC S

S

S COOH S

S

S COOHC12H25

(a)

Sample Initiator CTA [RAFT]/[Ini] Conversion Mn,exp (Mn, th) Mw/Mn

VL103 AIBN A 8 95 % 2100 (2200) 1.13

VL068 AIBN A 20 99 % 2200 (2200) 1.12

VL123 ACVA A 8 97 % 2800 (3200) 1.11

VL127 ACVA A 8 99 % 2200 (2200) 1.11

VL131 ACVA B 20 99 % 2300 (2200) 1.14

All polymerizations were carried out at 80ºC, in a toluene/acetone (1:1 vol/vol) mixture under an argon

atmosphere

A = B =

Table 4.2. Conversions and molecular weights characteristics of the poly(butyl acrylate)

samples synthesized by RAFT polymerization

The critical conditions were more troublesome to find than in the case of hydroxyl-functional

PMMA. Using pure acetonitrile (known as a good desorbing solvent for PBA on silica columns)

as mobile phase resulted in full retention of mono and di-functional chains on the bare silica

column. We circumvent this problem by adding some acid (10 v/v % formic acid or acetic acid

in dichloromethane) to the mobile phase.

2

3

4

5

6

7

4 5 6 7 8 9 10 11Acetonitrile (%)

Ret

entio

n tim

e (m

in)

PBA 600

PBA 7390

PBA 32333

Chapter 4

60

2

3

4

5

6

7

4 5 6 7 8 9 10 11Acetonitrile (%)

Ret

entio

n tim

e (m

in)

PBA 1-COOH2610

PBA 1-COOH13260

PBA 1-COOH19090

2

2.5

3

3.5

4

4.5

5

3 4 5 6 7Retention time (min)

LogM

p

PBA 0-COOHPBA 1-COOHPBA 2-COOH

3

4

5

6

7

8

9

10

4 5 6 7 8 9 10 11Acetonitrile (% )

Ret

entio

n tim

e (m

in)

PBA 2-COOH 2400

PBA 2-COOH 3200

PBA 2-COOH 5540

PBA 2-COOH 11450

(c)

(b)

(d)

Figure 4.3. Dependence of retention time on the mobile-phase composition (at 0.5 % HAc) for

PnBA samples with different molecular weights.

(a) Non-functional PnBAs. (b) Mono-functional carboxyl PnBAs. (c) Di-functional carboxyl

PnBAs. (d) Molecular-weight effect on retention time for non-, mono- and di-carboxyl

functional PnBAs under near-critical conditions (6 % ACN and 0.5 % HAc in DCM).


61

ELSD detector, home-packed Hypersil silica column (150 mm x 4.6 mm i.d.; 3-µm particles;

10nm pore size), flow rate 0.5 mL/min, column temperature 55˚C.

It can be noticed in Figure 4.3 that the critical point, where the retention time is independent

of the molar mass of the poly(butyl acrylate) chains (determined by the crossing points of the

different curves for non-, mono- and di-COOH-functional chains), is slightly different for the

three types of polymer. For the non-functional chains, the critical conditions are observed for 6.6

v/v % acetonitrile in DCM, while it is 5.8 v/v % for mono-functional and 5.5 v/v % for di-

functional. This fact is consistant with the observations of Gorbunov et al.7 and Skvortsov et al.8

They stated that strong interactions between a functional polymer and a stationary phase will be

dependant of the molar mass, the interaction decreasing with an increasing molar mass, resulting

in a lower retention. We encountered this situation in our separations, with the strong interaction

between the carboxyl group and the silica column. To counterbalance the effect of the molecular

weight on the retention of carboxyl-functional poly (butyl acrylate), the mobile phase mixture

should contain less desorbing solvent, acetonitrile in our case. That explains that the amount of

acetonitrile required decreases with increasing number of end-groups. It can be seen in Figure

4.3 that the retention time for nonfunctional poly (butyl acrylate) increased with increasing

molecular weight, which is the opposite of the behaviour observed for mono- and di-functional-

COOH-functional poly (butyl acrylate). As a final note, we observed a much clearer separation

for low molar mass functional polymers than for high molar mass polymers. In this chapter, we

will focus on polymers with Mn < 5,000 g/mol. The established conditions (6 v/v % acetonitrile,

0.5 v/v % acetic acid in dichloromethane) are then very suitable for separations of poly (butyl

acrylate) chains based on carboxyl functionality, independent of the molar masses of the chains.

The chromatograms acquired are depicted in Figure 4.4, the quantitative results are presented

in Table 4.5. The ELSD calibration curves for mono - and di-carboxyl-functional PBAs were

established using known injected amount of polymers VL131 and VL127 (Table 4.2),

respectively. The calibration curves fit very well with a power law.

As seen in Figure 4.4, the products of polymerizations VL103 and VL068 (Table 4.2) using

AIBN as an initiator showed major peaks in the 2-COOH region. However, a significant peak in

the 1-COOH region can also be observed. The peak in the 2-COOH region can be attributed to

structure 1 (Figure 4.6, expected RAFT polymer structure) or 4 (Figure 4.6, dead chain resulting

from the combination of two chains initiated by the leaving group of the RAFT agent). The peak

in the 1-COOH region can be explained by the structures 2 (Figure 4.6, RAFT polymer

structure, one chain initiated by AIBN) and 5 (Figure 4.6, dead chain resulting from the

Chapter 4

62

combination of two chains, one initiated by the leaving group of the RAFT agent, and the other

by AIBN). The relative intensity of the 1-COOH peak compared with the 2-COOH one is higher

in the case of polymer VL103 as for VL068. This is consistent with the amount of initiator in

VL103, which is more than twice as large than in VL068. No signals in the 0-COOH region is

observed in either chromatogram. This indicates that in both experiments, the amounts of

materials 3 (Figure 4.6, RAFT polymer, with two chains initiated by AIBN) and 6 (Figure 4.6,

dead chain resulting from the combination of two chains initiated by AIBN) in the samples are

negligible. This was expected, considering the relatively low concentrations of initiator used.

Figure 4.4. Liquid-chromatography separations of PBA samples bases on their COOH-

functionality

0 1 2 3 4 5 6 7 8 9 10

Time (min)

ELSD

resp

onse

0-COOH 2-COOH1-COOH

VL131

VL127

VL123

VL068

VL103


63

(PBA)

CN

(PBA)

CN

HOOC COOH

S

SS (PBA)(PBA) COOHHOOC

S

SS (PBA)(PBA)

CN

HOOC

(PBA)HOOC (PBA) COOH (PBA)HOOC (PBA)

CN

(PBA)

CN

(PBA)

CN

(PBA)

CN

(PBA) COOHHOOC

S

SS (PBA)(PBA) COOH

CN

HOOC

S

SS (PBA)(PBA)

CNCN

HOOC COOH

S

SS (PBA)(PBA)

CN CN

Table 4.5. Quantitative analysis of carboxyl-functional PnBAs by LC-ELSD. Isocratic

conditions and column as in Figure 4.3 (samples as indicated in Table 4.2).

1 2

3

4 5

6

7 8

9 10

Figure 4.6. Possible structures resulting from the butyl acrylate-polymerizations using S,S’-

Bis(α,α’-dimethyl-α’’-acetic acid)trithiocarbonate as RAFT agent and AIBN or ACVA as

initiator.

Sample % Mono-functional % Di-functional Initiator used

VL103 3 97 AIBN

VL068 1 99 AIBN

VL123 1 99 ACVA

VL127 1 99 ACVA

VL131 1 99 ACVA

Chapter 4

64

MALDI-TOF-MS was used as an analytical technique to study the RAFT polymer VL068

despite considerable debate in the literature (notably concerning chain fragmentation during

analysis). The spectra that were acquired for trithiocarbonate-containing polymers showed no

evidence of chain fragmentation during the analysis (Figure 4.7). In this respect, analysis by

MALDI-TOF-MS seems easier for the RAFT polymers with trithiocarbonate groups than for

those with thiocarbonyl thio functions. In the spectrum, groups of three peaks separated by 22

mass units can be seen, each “triplet” being separated from the next one by a mass of 128.2

mass units (corresponding to the butyl acrylate monomer mass). The separation of each peak in

the triplet by 22 mass unit can easily be explained, considering the experimental procedure

employed for the sample preparation in the MALDI-TOF-MS analysis. The samples are mixed

with a sodium salt, in order to facilitate the ionisation of the macromolecule. The sodium can

easily exchange with the hydrogen atom of the carboxylic acid group, leading to an increase in

mass for the polymer chain of 22 g/mol. The majority of the polymer chains in VL068 contains

two carboxyl groups. The exchange of zero, one or two hydrogen(s) of the carboxylic acid

groups explains the presence of a family of three peaks separated by 22 mass units. MALDI-

TOF-MS, despite its non-quantitative character, indicates the overwhelming presence of

polymer chains with structure 1 (Figure 4.6), confirming the results obtained by liquid

chromatography.

Figure 4.7. MALDI-TOF-MS spectra of PBA RAFT-polymer VL068. Counter ion= Na+

1500 2000 2500 3000 3500

m/z (Da)


65

The liquid-chromatography separations performed on samples VL123 and VL127 (Table 4.2)

revealed the dominant presence of dicarboxyl polymer chains (Table 4.5). Assuming that

termination of propagating poly(butyl acrylate) chains occurs only by combination, the expected

structures in experiments VL123 and VL127 are 1 (Figure 4.6, expected RAFT polymer

structure), 4 (Figure 4.6, dead chain resulting from the combination of two chains initiated by

the leaving group of the RAFT agent), 7 (Figure 4.6, RAFT polymer structure, one chain

initiated by ACVA), 8 (Figure 4.6, RAFT polymer, with two chains initiated by ACVA), 9

(Figure 4.6, dead chain resulting from the combination of two chains, one initiated by the

leaving group of the RAFT agent, and the other by ACVA) and 10 (Figure 4.6, dead chain

resulting from the combination of two chains initiated by ACVA). All those structures are

dicarboxyl-functional and cannot be differentiated with the liquid-chromatography method that

was employed. The presence of small amounts of monofunctional chains is probably the result

of hydrogen-transfer reactions. Since such reactions are inherent to any kind of radical-

polymerization process, the presence of this peak does not require too much attention. It has to

be noted that these transfer reactions also provide a reasonable explanation for the presence of

monofunctional peaks in the chromatograms of VL103 and VL068.

S-1-Dodecyl-S'-(α,α’-dimethyl-α’’-acetic acid)trithiocarbonate was also used as a RAFT

agent in the polymerization of butyl acrylate. The use of this particular RAFT agent was dictated

by the desire to obtain monofunctional carboxyl chains, with a controlled molecular weight and

a low polydispersity. The possible structures of the polymer-chains at the end of this

polymerization (VL131 – Figure 4.2) are presented in Figure 4.8, and the analysis of the

polymer obtained by liquid-chromatography in Figure 4.4, Trace 5.

The liquid chromatogram showed that the RAFT polymerization resulted exclusively in

monofunctional carboxyl chains. These monofunctional polymers can have the following

structures: 1’ (Figure 4.8, expected RAFT polymer structure) or 2’ (Figure 4.8, RAFT polymer,

with a chain initiated by ACVA). If the termination reaction is again assumed to occur only by

combination, the chromatogram shows no terminated product in the case of VL131. This was

unexpected, as termination reactions are not suppressed or limited in RAFT polymerization,

contrary to ATRP or NMP. The structures that result from termination by combination are: 3’

(Figure 4.8, dead chain resulting from the combination of two chains initiated by the leaving

group of the RAFT agent), 4’ (Figure 4.8, dead chain resulting from the combination of two

chains, one initiated by the leaving group of the RAFT agent, and the other by AIBN) and 5’

(Figure 4.8, dead chain resulting from the combination of two chains initiated by ACVA)

Chapter 4

66

S

SS (PBA) COOHC12H25

S

SS (PBA)

CN

C12H25COOH

(PBA)HOOC (PBA) COOH

CN

(PBA)HOOC (PBA)COOH

CNCN

(PBA) (PBA)COOHHOOC

(1’) (2’)

(3’) (4’)

(5’)

Figure 4.8. Possible structures resulting from the butyl acrylate-polymerization using S-1-

Dodecyl-S'-(α,α’-dimethyl-α’’-acetic acid)trithiocarbonate as RAFT agent and ACVA as

initiator.

The absence of a small signal in the dicarboxyl region for sample VL131 may be partially

explained by the chromatographic broadening at high elution times. Apparently the fraction of

dead chains is below the detection limit of the present method and equipment. Some modelling

work has been conducted in order to determine the theoretical amount of dead chains that should

be present in our polymerization system. This modelling work was extremely tough to conduct,

due to the impossibility to get access to all the kinetic constants involved in the RAFT

mechanism, through literature or simple experiments. With all the precautions needed for the

results of this modelling work, it seems that at the end of the polymerization the proportion of

dead chains (difunctional - terminated by combination) accounts for less than 1 % of the total

amount of chains. This would explain the non-detection of difunctional chains in sample VL131.

Nevertheless, even with the question mark surrounding the absence of difunctional chains in

VL131, the quantitative production of monofunctional chains is very suitable for our purpose.

Comprehensive 2D LCxSEC separation

Similarly to what was performed in Chapter 3 on hydroxyl-functional poly (methyl

methacrylate), the LC separation technique was coupled to a SEC separation in order to


67

determine in a single experiment both the functionality-type distribution and the molecular

weight distribution. Due to time constraints, no quantitative results were obtained. The

following will then only provide qualitative informations.

(a) (b)

(c)

Figure 4.9. Two-dimensional LC×SEC chromatogram of sample VL103. (a) UV detection at

230 nm. (b) UV detection at 230 nm. (zoom on the 2-3-4 peaks) (c) ELSD. The LC mobile

phase was 50% ACN in DCM and the flow rate was 40 µL/min.

The two-dimensional separation spectra exhibits 4 main peaks as can been seen in Figure 4.9.

Peaks 2, 3 and 4 are respectively peaks corresponding to polymer chains with 0, 1 and 2 COOH

functionality. The explanations for the presence of those 3 peaks are the same as the ones

provided for the one-dimensional LC-separation of VL103:

- Peak #2: polymers having a relatively high molecular weight and no COOH functionality.

Two polymer chains initiated by a non-functional AIBN fragment are probably detected

with this peak. The growth of the chain has been controlled by the thiocarbonyl thio

compounds in a RAFT process, after the initiation by the AIBN-derived radical. Some

chains initiated by an alkyl radical after a chain transfer reaction might as well be present

1

2 3 42 3 4

2 3 4

Chapter 4

68

in the population represented by this peak. It corresponds mainly to the peak in the 0-

COOH region on Figure 4.4.

- Peak #3: relatively high molar masses are observed with 1 COOH functionality. The

polymers chains might have been initiated by a non-functional AIBN-derived radical or

during the RAFT polymerization some transfer reaction have occurred, resulting in

trithiocarbonate-containing chains with one COOH-functional chain on one side of the

CS3 moiety and hydrogen- or AIBN- terminated chain on the other side.

- Peak #4: relatively high molar masses are observed with 2 COOH functionalities. The

typical RAFT recipes result in majority into polymer chains initiated by the leaving group

of the RAFT agent. Having employed di-functional trithiocarbonate, this peak belongs to

the expected population after a RAFT experiment. Both sides of the chains bear a COOH-

functionality.

However, similarly to what was found in the previous chapter, a peak at low molar masses is

detected (Peak #1). Although more investigations would be required, it is logical to assume that

it may be some remaining RAFT agent. The surprising finding of the last chapter seems to occur

with the trithiocarbonate as RAFT agent as well. Further work will be required to explain fully

the presence of this low molecular weight compound in the final polymer.

IV) Conclusion

In this chapter, a synthetic path to obtain α,ω − carboxyl functional linear poly(butyl acrylate)

employing the RAFT polymerization process was proposed. An existing procedure proposed in

literature was employed, and a new advanced chromathographic characterization was employed

to confirm the purity of the expected structures. The use of trithiocarbonate-containing RAFT

agents in combination of functional initiator was demonstrated to yield telechelic polymers and

monofunctional polymers with a very high purity concerning the end-group functionality (above

99%). MALDI-TOF-MS analyses were conducted on telechelic polymers, that confirmed the

structure of di-functional polymers. Contrary to the PMMA RAFT polymers discussed in the

previous chapter, no evidence of fragmentation during the analysis was observed (see Chapter 5

for more details).


69

The highly pure mono- and di-functional poly(butyl acrylate) will be the materials of choice

for the network formation and network properties studies that will be described in Chapter 6.

The high purity and the stability of those compounds are the main reasons behind this choice.

Chapter 4

70

Reference List

1. Lai, J. T., Filla, D., Shea, R.; Macromolecules 2002, 35, 6754.

2. Cools, P. J. C. H., Van Herk, A. M., German, A. L., Staal, W.; Journal of Liquid Chromatography 1994, 17, 3133.

3. Gorshkov, A. V., Much, H., Becker, H., Pasch, H., Evreinov, V. V., Entelis, S. G.; Journal of Chromatography 1990, 523, 91.

4. Pasch, H., Rode, K., Chaumien, N.; Polymer 1996, 37, 4079.

5. Falkenhagen, J., Much, H., Stauf, W., Mueller, A. H. E.; Macromolecules 2000, 33, 3687.

6. Macko, T., Hunkeler, D., Berek, D.; Macromolecules 2002, 35, 1797.

7. Gorbunov, A., Trathnigg, B.; Journal of Chromatography, A 2002, 955, 9.

8. Skvortsov, A. M., Fleer, G. J.; Macromolecules 2002, 35, 8609.

MALDI-TOF-MS analysis of RAFT polymers

71

CHAPTER 5

MALDI-TOF-MS ANALYSIS OF RAFT POLYMERS

I) Introduction

Matrix Assisted Laser Desorption-Ionisation Time Of Flight Mass Spectrometry (MALDI-

TOF-MS) is a powerful technique for the qualitative end-group characterization of polymers1-3.

During the course of this project, the determination of end-groups was of primary importance.

Hence, MALDI-TOF-MS was used on the different synthesized polymers in order to get

information on the nature of the end-groups. In this technique, the polymer is dispersed in a

solid matrix, desorbed by a UV-laser, and ionized by the presence of alkali salts. The main

advantages of the MALDI-TOF-MS technique versus the other mass spectrometric techniques

are the almost exclusive production of single-charged polymer molecules after ionization and

the high resolution of the obtained spectra (especially for low molar masses polymers, typically

with Mn < 10.000 g/mol). The main disadvantage is the sensitivity for fragmentation of certain

polymer chains during the analysis. Several reports in literature present RAFT polymers as

compounds prone to fragmentation under laser irradiation4-9.

In this chapter, MALDI-TOF-MS analysis of poly(butyl methacrylate) and poly(butyl

acrylate) RAFT polymers will be described. Conclusions will be drawn on the eventual

fragmentation during the analysis. The influence of the employed laser intensity has been

investigated. Also, the effect of the monomer nature and the type of RAFT agent used during the

synthesis will be highlighted.

II) MALDI-TOF-MS: Experimental procedure

Measurements were performed on a Voyager-DE STR (Applied Biosystems, Framingham,

MA, USA) instrument equipped with a 337 nm nitrogen laser. Positive-ion spectra were

Chapter 5

72

S S

CN

OH

CN

OHNN

CN

HO

acquired in reflector mode. DCTB (trans-2-[3-(4-tert-butylphenyl)-2-methyl-2-

propenylidene]malononitrile) was chosen as the matrix. Sodium trifluoroacetate (CF3CO2Na)

was added as the cationic ionization agent. The matrix was dissolved in THF at a concentration

of 40 mg/mL. Sodium trifluoroacetate (CF3CO2Na) was added to THF at a concentration of 1

mg/mL. The dissolved polymer concentration in THF was approximately 1 mg/mL. In a typical

MALDI experiment, the matrix, salt, and polymer solutions were premixed in the following

ratio: 5 µL of sample: 5 µL of matrix: 0.5 µL of salt. Approximately 0.5 µL of the obtained

mixture was hand spotted on the target plate. For each spectrum, 1000 laser shots were

accumulated.

III) Analysis of thiocarbonyl thio-containing polymers by MALDI-

TOF-MS

PBMA (Mn = 2.500 g/mol, PDI = 1.26) was analyzed using the experimental procedure

described above. The influence of the laser intensity was investigated, acquiring spectra with

intensities of 1950, 2100 and 2300 Lux.

The RAFT agent used for the synthesis was (4-cyano-1-hydroxypent-4-yl) dithiobenzoate

(RAFT-ACP) (purity of 97% determined by 1H-NMR in CDCl3) and the initiator was 4,4’-

azobis(4-cyanopentanol) (ACP) with a purity superior to 99% (1H-NMR in CDCl3). The

polymerization was carried out in toluene at 70˚C and the final product was isolated by rotary

evaporation of the solvent and the unreacted monomer.

(4-cyano-1-hydroxypent-4-yl) dithiobenzoate 4,4’-Azobis(4-cyanopentanol)

RAFT-ACP ACP


73

609.0 1707.4 2805.8 3904.2 5002.6 6101.0Mass (m/z)

0

64

0

10

20

30

40

50

60

70

80

90

100

% In

tens

ity

Voyager Spec #1[BP = 2423.2, 64]

2423.1970

2707.59362565.4097 3417.5312

1710.22273133.14921995.6326

3560.76621569.0460

1558.12032980.05602268.1144 3702.92351141.4652 2125.9027

3270.91292560.2762823.1484 4271.75022695.70691414.9229

4965.5708751.9689 3406.6591 4265.64841983.73092843.3222 3696.4157656.0670 4412.84021841.5657

1271.7540801.1282 3138.2636 3855.18652295.9647 4592.1421 5233.13061177.4519664.7383 2871.79952142.9179 4666.12704028.52723423.7266 5486.60061471.1112880.4635 2817.0926 3340.90712301.0723 4822.42034034.1308 5507.29711685.9418674.8573 2830.37661170.4264 4204.47323680.0091 4876.96592212.9217 5454.86941350.9061 2782.7518

1035.0 2028.6 3022.2 4015.8 5009.4 6003.0Mass (m/z)

0

564.0

0

10

20

30

40

50

60

70

80

90

100

% In

tens

ity


2408.62382551.7106

2693.82072124.4516 3410.79222841.77992136.3077

3552.97421851.1275 2835.91752563.5916 3694.7244

3131.95932558.66021566.95183405.1979 4121.3522

3547.33031424.8460 2273.86304406.06143830.52141282.7464

4548.01871989.7402 3842.47644690.78931248.8904 4126.65382425.61151846.8712 3004.8400 4832.2468

1036.7138 1705.0853 2284.5100 3846.73162852.9151 4303.3715 4837.05882039.39191298.7091 3581.68682857.4012 4095.7210 4699.2350 5260.95702109.38801088.5244 1639.1373 3226.44932671.7338 3820.4002 5326.92864333.99291893.25321117.1203 2448.0141

Figure 5.1. MALDI spectra of PBMA RAFT polymer with Laser Intensity = 1950 Lux

2193.0 2252.2 2311.4 2370.6 2429.8 2489.0Mass (m/z)

0

64

0

10

20

30

40

50

60

70

80

90

100

% In

tens

ity


2423.19702280.0118

2422.20862281.0024

2409.2851

2278.9991

2424.23642268.1144

2410.30062411.2970

2267.1308 2417.19132426.2525

2275.02702439.22772310.98082274.5880 2386.55132236.7972

2430.61822276.5727 2309.7541 2394.12322241.06722205.9354 2461.51582421.63782352.64722204.0262 2257.2180 2315.9166 2381.49482288.1599 2457.66202348.3605 2420.31922258.2832 2384.53762305.42632215.4963 2448.95012341.0117

A1C1

A1

B1C1

B1

Chapter 5

74

2193.0 2252.8 2312.6 2372.4 2432.2 2492.0Mass (m/z)

0

564.0

0

10

20

30

40

50

60

70

80

90

100

% In

tens

ity


2408.6238

2266.5291

2267.53492420.5002

2278.3967

2265.5355 2419.48012268.5560 2407.6231

2416.23522273.86302422.50522272.8476 2417.6869

2280.41422423.5060

2414.02872275.98342271.8271 2411.1850

2324.57552296.4215 2428.57252363.46392209.4955 2251.4823 2334.31022288.2618 2440.10452404.17542367.55892211.5340 2248.0058 2335.87832308.0850 2461.5393

853.0 1767.2 2681.4 3595.6 4509.8 5424.0Mass (m/z)

0

2468.0

0

10

20

30

40

50

60

70

80

90

100

% In

tens

ity

Voyager Spec #1[BP = 2124.4, 2468]

2124.3861

2550.64691698.1298 2693.7392

2835.82701555.0391

1413.9500 3119.9821

2841.66753268.6154986.6792 2699.6007

3694.88623546.27752273.5968

3979.5186

1989.5840 4262.96523832.3479895.4131 2136.27081709.9998 4116.6408900.6172 2846.7628 4546.98042425.55151282.7132 3971.48043415.0685906.8252 2993.93391897.2506 4542.82832436.45961440.8240 3564.31383000.35871936.1144 3991.3680 4976.04272387.4460 4422.78691475.4604 3518.37002899.6739 4860.41603957.4466

2154.0 2222.6 2291.2 2359.8 2428.4 2497.0Mass (m/z)

0

2468.0

0

10

20

30

40

50

60

70

80

90

100

% In

tens

ity

Voyager Spec #1[BP = 2124.4, 2468]

2266.4674 2408.56282267.4773

2409.5687

2410.56032265.4754

2415.6256

2416.59312273.59682274.5804

2412.53812270.5064 2413.67922271.7413 2419.4738

2276.7117 2422.51302182.4138 2324.54072286.45022237.4612 2431.66962380.52132179.5015 2464.65222318.56892255.4847



Figure 5.1, Figure 5.2 and Figure 5.3 represent the spectra acquired at various laser intensities

and enlargements in the 2200 – 2400 Da region. It is obvious, from the enlargements of the

spectra that the laser intensity strongly influences the observed peak patterns. In the three

A2

B2

C2

A2

B2

C2

A3

B3

C3

A3

B3

C3


75

n

, Na+CH2 C

CH3

O OC4H9

CN

OHSS

n

, Na+CH2 C

CH3

O OC4H9

CN

OHCH

CH2

OO

C4H9

spectra, three patterns can be observed that are repeated every 142.2 mass units (this value

corresponds to the mass of a single butyl methacrylate unit). Although the spectrum of Figure

5.1 suffers from a general poor signal-to-noise ratio, due to the low laser intensity used for the

measurement, a striking difference in resolution can still be noticed. While the peaks with an A

or C label have a relatively good resolution, the B peaks in the three spectra exhibit very poor

resolution, with non-resolved isotopic distributions.

Assigning the A and C peaks leads to the following structures:

A =

C =

The assignment for the C cluster corresponds to the expected structure after a RAFT

polymerization. The thiocarbonyl thio moiety is attached to the polymer chain. The structure

related to the A cluster can be related to several phenomena that may have happened to the

polymer chain prior to or during the analysis:

- Termination by disproportionation is the preferred termination mode for methacrylic

monomers. Simulation performed on these systems indicates that roughly 10% of the

polymer chains present at the end of the polymerization are the result of termination

reactions. It is then logical that we observe the peaks of the A family.

- Fragmentation during the MALDI-TOF-MS analysis is known to happen for certain

RAFT agents/polymers8,10. The cleavage of the weak carbon-sulfur bond of the

thiocarbonyl thio compound may result in chains terminated by a double bond following

the reaction scheme displayed in Figure 5.4.

Chapter 5

76

HO

CN

CH2

CH3

O OC4H9

CH2

CH3

O OC4H9

S S

n

n

S S

O OC4H9

HO

CN

CH2

CH3

O OC4H9

CH2

CH2

H

nO O

C4H9

HO

CN

CH2

CH3

O OC4H9

CH2

CH2S SH

+

2264.77807 2266.08587 2267.39366 2268.70145 2270.00925 2271.31704Mass (m/z)

0

100

0102030405060708090

100

% In

tens

ity

ISO:Na(CH2CCH3COOC4H9)15CCH3CNC3H6O2266.5517

2267.55502265.5483

2268.5581

2269.56122270.5642

2264.77807 2266.08587 2267.39366 2268.70145 2270.00925 2271.31704Mass (m/z)

0

2468.0

0102030405060708090

100

% In

tens

ity

Voyager Spec #1[BP = 2124.4, 2468]2266.4674

2267.4773

2265.4754 2268.4743

2269.47252270.5064

Figure 5.4. Fragmentation of the thiocarbonyl thio-containing polymer during MALDI-TOF-

MS experiments

Inspection of the isotopic distribution reveals a noticeable deviation from the theoretical

isotopic distribution (Figure 5.5).

Figure 5.5. MALDI-TOF-MS spectrum of PBMA RAFT polymer with laser intensity of 2300

Lux (a) and the simulated spectrum corresponding to a polymer A with n=15 (b)

(a)

(b)


77

HO

CN

CH2

CH3

O OC4H9

H

n

Despite a small shift of the baseline, it seems that the peaks at positions 2267.5 ± 0.1 Da and

2268.5 ± 0.1 Da are more intense in the acquired spectrum than in the theoretically calculated

spectrum. The reason is probably the presence underneath those peaks of another family of

peaks (higher by 2 Da) corresponding to the hydrogen-terminated chains with the following

formula:

D =

These hydrogen-terminated chains can have several origins:

- Termination by disproportionation will result, as discussed above, not only in double

bond-terminated chains, but also in hydrogen-terminated chains.

- Transfer reaction during the radical polymerization may occur. Toluene is known to

have a relatively low transfer constant for methacrylates (e.g. Ctr = 5.10-5 for methyl

methacrylate at 70˚C11, the transfer constant for butyl methacrylate is expected to be in

the same order of magnitude). In the eventuality of transfer to toluene, some polymer

chains should be found with a benzyl end-group. In the MALDI-TOF-MS spectra, those

chains should be displayed around 2247.0 Da. Evidence for the presences of those peaks

could not be found. One may wonder whether the benzyl radicals are suitable for butyl

methacrylate initiation. According to Fischer and Radom12, the addition of benzyl

radicals to methyl methacrylate is faster than the one of 2-cyano-2-propyl (the AIBN-

derived radical). Assuming a minor difference in reactivity between the methyl

methacrylate and the butyl methacrylate, we can conclude that the absence of benzyl-

terminated polymer chains in the spectrum suggests that the hypothesis of transfer to

solvent can be discarded. Branched polymer chains, resulting from transfer to polymer,

can not be detected directly by MALDI-TOF-MS. However, these transfer reactions

should not be occurring during the polymerizations13 (Ctr = 2.48 10-4), considering the

relatively low molecular weight of the final polymer. Transfer to monomer can also be

rejected as a cause of the formation of H-terminated chains, as the transfer constant to

methyl methacrylate is Ctr = 3.10-5 11.

- Fragmentation during the MALDI-TOF-MS analysis produces chains with a terminal

radical. Two of those radicals may terminate by disproportionation, in a similar fashion

to what occurs in radical polymerization. However, the proximity of a thiocarbonyl thio

Chapter 5

78

2265.00381 2266.25206 2267.50031 2268.74856 2269.99681 2271.24506Mass (m/z)

0

290.0

0102030405060708090

100

% In

tens

ity

ADD (ISO:Na(CH2CCH3COOC4H9)15CCH3CNC3H6O , ISO:NaH2(CH2CCH3COOC4H9)15CCH3CNC3H6O)2267.5593 2268.5647

2266.5517 2269.56892265.5483

2270.5726

2265.00381 2266.25206 2267.50031 2268.74856 2269.99681 2271.24506Mass (m/z)

0

2468.0

0102030405060708090

100

% In

tens

ity

Voyager Spec #1[BP = 2124.4, 2468]2266.4674

2267.4773

2265.4754 2268.4743

2269.47252270.5064

radical after the carbon-sulfur bond cleavage will ensure a predominant occurrence of

the double-bond forming mechanism shown in Figure 5.4.

The termination by disproportionation of growing chains during the polymerization may be an

acceptable explanation for the combined presences of double-bond-terminated chains and

hydrogen-terminated chains. However, making the reasonable assumption that the ionization

efficiencies for both types of chains are similar, and assuming a 1:1 ratio between H-terminated

and double bond-terminated chains, the isotopic distribution in the 2264-2271 Da region should

be as shown in Figure 5.6 (b), which is clearly not the case.

Figure 5.6 MALDI-TOF-MS spectrum of PBMA RAFT polymer with laser intensity of 2300

Lux (a) and the simulated spectra corresponding to a 1:1 ratio of polymer A with n=15 and

polymer D with n=15 (b)

This evidence shows that the disproportionation-termination mechanism, either during the

polymerization or during the analysis, is not the main explanation for the detection of polymer

chains with a terminal double bond. The fragmentation of the weak carbon-sulfur bond of the

thiocarbonyl thio-containing chains and the hydrogen abstraction by the thiocarbonyl thio

radical formed (Figure 5.4) must be the main cause. The spectrum displayed does not match

with the hypothetical spectrum expected from the structure predicted by the RAFT

polymerization theory. The fragmentation during the analysis is responsible for the appearance

of the extra peaks corresponding to polymer chains of general structure A.

The main reason that can be proposed for the high tendency of thiocarbonyl thio-containing

polymers to undergo fragmentation during the MALDI-TOF-MS analysis is the absorbance

band around 300 nm that the PBMA polymer exhibits. The laser used in the MALDI-TOF-MS

(a)

(b)


79

to irradiate the matrix in which the polymer is dispersed, has a wavelength of 337 nm, close to

this absorbance band. It is easy to imagine that the thiocarbonyl thio end-group can absorb this

wavelength and reach a higher energy level. The compound can then lower its energy by

fragmentation, yielding a relatively stable tertiary carbon-centered radical and a thiocarbonyl

thio radical, which is heavily stabilized through resonance with the carbon-sulfur double bond

and the aromatic ring.

Another explanation for the fragmentation is the local heating that the polymer sample

undergoes at the moment of the irradiation. It is known that the temperature can rise up to more

than 500˚C for a short period of time (in the order of a nanosecond) at the location of the laser

impact14,15. This temperature can be sufficient to induce a thermal decomposition in the case of

RAFT polymers, by breaking up the weak carbon-sulfur bond.

IV) Influence of the laser intensity on the fragmentation

Despite the impossibility to draw absolute quantitative conclusions from mass spectrometric

analysis, it is legitimate to compare the analysis of the same sample recorded with various laser

intensities. In Figures 5.1, 5.2 and 5.3, it is obvious that the relative intensity of the cluster A

increases with an increase of the laser intensity and, that the intensity of the cluster C follows an

opposite trend, decreasing with increasing intensity. The hypothesis of the laser-induced

fragmentation is thus supported by the variation of the peak patterns with intensity.

The bad resolution of the B cluster in the three spectra presented is rather unusual and seems to

be the result of a fragmentation during the flight of the polymeric ions. The measured masses

cannot be ascribed to any of the possible structures expected from the synthesis. The better

resolution of the cluster C seems to point toward a fragmentation during the ionization. An

explanation can be based on the hypothesis of the excitation of the carbon-sulfur bond to a

higher energy level. It seems that some chains relax by fragmenting during the ionization (peaks

C), while some other ones fragment during the flight (peaks B). The remaining chains are

unaffected by any fragmentation steps and are detected with the thiocarbonyl thio moiety still

attached to one of their chain ends (peaks A).

Poly(butyl acrylate) (Mn = 1.800 g/mol, MWD = 1.29) was analyzed following the

experimental protocol described above. The mode of synthesis was a RAFT polymerization with

2-cyanoprop-2-yl dithiobenzoate as RAFT agent and AIBN as initiator. The polymerization was

performed in toluene at 80˚C and the final product was isolated by evaporation of the solvent

Chapter 5

80

n

, Na+CH2 CH

O OC4H9

CN

SS

1005 1535 2065 2595 3125 3655Mass (m/z)

0

272.0

0

10

20

30

40

50

60

70

80

90

100

% In

tens

ity

Voyager Spec #1[BP = 2295.3, 272]

2295.3378

2167.2151

2039.0867 2423.4625

2551.60041782.8271

2679.7278

1653.7040 2808.8565

2936.97621525.5779

1397.4435 3065.11972426.47422682.72211269.3055 3193.17411656.6941 2170.25391141.1709 1913.9565 3321.31252939.90592687.74461629.7985 2439.42921886.0568 2183.2293 3199.19521386.57631129.3114

1887.0 1899.2 1911.4 1923.6 1935.8 1948.0Mass (m/z)

0

100

0102030405060708090

100

% In

tens

ity

ISO:CCH3CH3CN(C7H12O2)13NaCS2C6H51910.1143

1909.11101911.1164

1912.1180

1913.1196

1887.0 1899.2 1911.4 1923.6 1935.8 1948.0Mass (m/z)

0

215.0

0102030405060708090

100

% In

tens

ity

Voyager Spec #1[BP = 2295.3, 272]1910.9545

1909.9603

1911.9652

1912.9689

1913.9565 1926.98461888.0525 1898.0548 1919.9893

and the unreacted monomer. The spectrum acquired with a laser intensity of 1400 Lux is

presented in Figure 5.7

Figure 5.7. MALDI spectrum of PBA RAFT polymer with Laser Intensity = 1400 Lux

The main population can clearly be assigned to the structure expected after a RAFT

polymerization:

E =

Figure 5.8. MALDI-TOF-MS spectrum of PBA RAFT polymer with laser intensity of 1400 Lux

(a) and the simulated spectrum corresponding to polymers E with n =13 (b)

(a)

(b)


81

1034 1529 2024 2519 3014 3509Mass (m/z)

0

1.0E+4

0

10

20

30

40

50

60

70

80

90

100

% In

tens

ity

Voyager Spec #1[BP = 1001.0, 12663]

1129.1329

1257.2584

1385.3829

1513.49781141.05681397.3080

1653.57201132.1405 1910.84372167.12401125.0842

1770.75581388.38361058.9389 2423.40841898.89701413.3500 1644.6053 2680.63662148.65121894.82331626.4803 2404.6710 2936.99502143.1502

1735 1757 1779 1801 1823 1845Mass (m/z)

0

100

0102030405060708090

100

% In

tens

ity

ISO:CCH3CNCH3(C7H12O2)12CS2C6H5Na1782.0306

1783.0325

1784.0340

1785.0355

1735 1757 1779 1801 1823 1845Mass (m/z)

0

2964.1

0102030405060708090

100

% In

tens

ity

Voyager Spec #1[BP = 1001.0, 12663]1782.7124

1783.70911770.7558

1784.71261771.76381766.7004 1785.6851 1798.64491755.7363 1813.6874

A small difference in the peak positions between the theoretical and experimental spectra (less

than 1 mass unit) can be noticed in Figure 5.8. It was assumed that the calibration of the

spectrometer was the main reason for this difference and no further attention was paid to it.

Similarly to the previous series of experiments with thiocarbonyl thio-containing poly(butyl

methacrylate), the laser intensity was increased. Figure 5.9 represents a spectrum of the same

poly(butyl acrylate) recorded with a laser intensity of 2400 Lux.

Figure 5.9. MALDI spectrum of PBA RAFT polymer with Laser Intensity = 2400 Lux

Figure 5.10. MALDI-TOF-MS spectrum of PBA RAFT polymer with laser intensity of 2400

Lux (a) and the simulated spectrum corresponding to polymers E with n =12 (b)

Figure 5.10 shows evidence that some thiocarbonyl thio end-functionalized poly(butyl

acrylate) chains are still detected, despite the high laser intensity used. The other peaks recorded

could not be assigned to any structures expected from a RAFT polymerization. Their absence in

the spectrum recorded at low laser intensity indicates that they are probably the result of

(a)

(b)

Chapter 5

82

1307.0 1841.4 2375.8 2910.2 3444.6 3979.0Mass (m/z)

0

1.8E+4

0

10

20

30

40

50

60

70

80

90

100

% In

tens

ity

Voyager Spec #1[BP = 2400.6, 18377]

2400.5872

2528.7406

2144.28282656.8945

2785.04912016.1332

2913.20132378.64191887.9856 2484.8373

2228.5150 3042.35452531.73652100.3575

2275.4292 2869.3050 3170.50441972.20152916.20361866.0154 3298.64472509.7984 2766.1167 3148.56382253.4794 3426.79481630.7084 1890.9837 2662.9136 3404.83271608.7380 2919.15762407.64551318.5452 3683.08632138.2850

complex fragmentation/rearrangement reactions. The main conclusion of this experiment is that

the nature of the monomer greatly influences the overall stability of the chain. The thiocarbonyl

thio-containing poly(butyl methacrylate) chains could no longer be detected at high laser

intensity, while their equivalents in a poly(butyl acrylate) sample could be observed. The radical

stability of the carbon-centered poly(butyl acrylyl) radical (secondary radical) is lower than that

of the poly(butyl methacrylyl) radical (tertiary radical). It is then logical that less fragmentation

occurs after the laser irradiation in the acrylate case compared to the methacrylate case.

V) Analysis of trithiocarbonate-containing poly(butyl acrylate) by

MALDI-TOF-MS

As mentioned in Chapter 4, the polymerization of methacrylic monomers is troublesome with

S,S’-Bis(α,α’-dimethyl-α’’-acetic acid)trithiocarbonate, which was the preferred RAFT agent

employed in this project, due to the stabilities of the propagating radical and the leaving group

radical. The study on trithiocarbonate-containing polymers was therefore limited to poly (butyl

acrylate).

Poly(butyl acrylate) (Mn = 2.100 g/mol, MWD = 1.12) was analysed by MALDI-TOF-MS.

The RAFT agent used for the synthesis was S,S’-Bis(α,α’-dimethyl-α’’-acetic

acid)trithiocarbonate (purity superior to 99% determined by 1H-NMR in CDCl3), and the

initiator used was azobis(isobutyronitrile) (recrystallized from methanol). The polymerization

was performed in an acetone/toluene 1:1 mixture at 80˚C and the final product was isolated by

rotary evaporation of the solvent and the unreacted monomer. The spectrum acquired with a

laser intensity of 2200 Lux is presented in Figure 5.11


83

HOOC CH2 CH

CO O

C4H9

S

S

S CH2 CH

CO O

C4H9

COOH

n m

2311 2364 2417 2470 2523 2576Mass (m/z)

0

1.8E+4

0

10

20

30

40

50

60

70

80

90

100

% In

tens

ity

Voyager Spec #1[BP = 2400.6, 18377]

2400.5872

2528.7406

2401.5854 2529.7421

2399.5867

2527.7399

2402.5812 2530.7364

2378.6419 2506.80112507.80172379.6406

2377.6367 2531.73652403.5744 2505.79882380.6374 2486.8326

2532.73392404.5738 2487.83522359.6740 2533.71572405.5909 2488.82452360.66982324.6014 2522.75622452.8054

2477.0 2490.8 2504.6 2518.4 2532.2 2546.0Mass (m/z)

0

199.9

0102030405060708090

100

% In

tens

ity

ADD (ISO:Na2OOCNaCCH3CH3(C7H12O2)17CS3COOHCCH3CH3 , ISO:HOOCNaCCH3CH3(C7H12O2)17CS3COOHCCH3CH3) , ISO:NaOOCNaCCH3CH2483.4214 2528.39312505.4033

2529.39522484.4235 2506.4054

2504.40012482.4181 2527.3899

2486.4267 2508.4086 2531.3984

2509.4100 2532.39982487.42812489.4309 2511.4128 2534.4025

2477.0 2490.8 2504.6 2518.4 2532.2 2546.0Mass (m/z)

0

1.6E+4

0102030405060708090

100

% In

tens

ity

Voyager Spec #1[BP = 2400.6, 18377]2528.7406

2527.7399

2506.80112484.83732531.73652505.79882483.8358

2532.73392509.79842487.83522511.7926 2522.7562

Figure 5.11. MALDI spectra of PBA RAFT polymer with Laser Intensity = 2200 Lux

The assignments of the peaks have already been discussed in Chapter 4 and the conclusion

was that the three families of peaks corresponded to one original compound, with the following

structure:

F =

Figure 5.12. MALDI-TOF-MS spectrum of PBA RAFT polymer with laser intensity of 2200

Lux (a) and the simulated spectrum corresponding to polymers F with the substitutions of 0, 1

and 2 hydrogen atoms from the carboxylic groups by sodium atoms with n + m =17 (b)

(a)

(b)

Chapter 5

84

S

S

S(PBA)HOOC +

O

CH CH2

O

C4H9

(PBA) COOH

A difference of 1.4 mass unit is present between the theoretical spectra and the experimental

one in Figure 5.12. This difference was assumed to be caused by a calibration problem and no

further attention was paid to it. MALDI-TOF-MS analysis of the trithiocarbonate-containing

poly(butyl acrylate) showed no evidence of fragmentation. The laser intensity was raised to

2500 Lux, without any fragmentation patterns observed.

The explanation for this low tendency towards fragmentation is again based on the radical

stabilities of the compounds that might be created after fragmentation. Breakage of the carbon-

sulfur bond between the trithiocarbonate and the polymer would produce the following radical

species:

A comparison of the stabilities of the radicals produced upon fragmentation of

trithiocarbonate-containing poly(butyl acrylate) with the ones of thiocarbonyl thio-containing

poly(methyl methacrylate) leads to two major observations:

- The sulfur-centered radical is stabilized much more in the case of thiocarbonyl thio-

containing polymer, due to the delocalization of the unpaired electron on the aromatic

ring. Resonance delocalization occurs in the trithiocarbonate radical as well, but it is

limited to the carbon-sulfur double bond.

- The carbon-centered radical is more stable in the case of the polymethacrylate (tertiary

carbon) than in the case of polyacrylate (secondary carbon)

These two observations explain that the trithiocarbonate-containing poly(butyl acrylate) is less

prone to fragmentation than the thiocarbonyl thio-containing poly(butyl methacrylate) and

poly(butyl acrylate). In terms of energetic levels, the polymer molecules will reach a higher

energy level after irradiation by the laser, but contrary to the polymethacrylates, they will not

relax via a fragmentation reaction, but will return to their initial state.


85

V) Conclusions

Thiocarbonyl thio-containing poly(butyl methacrylate), poly(butyl acrylate) and

trithiocarbonate-containing poly(butyl acrylate) were analyzed by MALDI-TOF-MS. The

thiocarbonyl thio-containing poly(butyl methacrylate) showed evidence of chain fragmentation

during the analysis, with almost complete disappearance of the peaks corresponding to the

chains bearing a thiocarbonyl thio end-group at high laser intensities. A mechanism relying on

abstraction of a hydrogen from the carbon-centered poly(methacrylyl) radical by the

thiocarbonyl thio radical was proposed. This mechanism was supported by isotope-distribution

considerations. Thiocarbonyl thio-containing poly(butyl acrylate) chains showed less

fragmentation. Even at high laser intensities, some intact thiocarbonyl thio-containing chains

were observed.

The trithiocarbonate-containing poly(butyl acrylate) proved to be much more stable than the

two other polymers. No evidence was found for fragmentation even at high laser intensities. The

lower stability of the poly(acrylyl) radical versus the poly(methacrylyl) radical and the lower

stability of trithiocarbonate radical in comparison with the thiocarbonyl thio radical are the main

causes for the greater stability of the trithiocarbonate-containing poly(butyl acrylate).

Chapter 5

86

Reference List

1. Scrivens, J. H., Jackson, A. T.; Int. J. Mass. Spectrom. 2000, 200, 261.

2. Hanton, S. D.; Chem. Rev. 2001, 101, 527.

3. McEwen, C. N., Peacock, P. M.; Anal. Chem. 2002, 74, 2743.

4. Vosloo, J. J., Wet-Roos, D., Tonge, M. P., Sanderson, R. D.; Macromolecules 2002, 35, 4894.

5. Destarac, M., Charmot, D., Franck, X., Zard, S. Z.; Macromol. Rapid Com. 2000, 21, 1035.

6. D'Agosto, F., Hughes, R., Charreyre, M. T., Pichot, C., Gilbert, R. G.; Macromolecules 2003, 36, 621.

7. Ganachaud, F., Monteiro, M. J., Gilbert, R. G., Dourges, M. A., Thang S.H., Rizzardo, E.; Macromolecules 2000, 33, 6738.

8. Schilli, C., Lanzendoerfer, M. G., Mueller, A. H. E.; Macromolecules 2002, 35, 6819.

9. Vana, P., Albertin, L., Barner, L., Davis, T. P., Barner-Kowollik, C.; J. Polym. Sci. Pol. Chem. 2002, 40, 4032.

10. Favier, A., Ladaviere, C., Charreyre, M. T., Pichot, C.; Macromolecules 2004, 37, 2026.

11. Gopalan, M. R., Santhappa, M.; J. Polym. Sci. 1957, 25, 333.

12. Fischer, H., Radom, L.; Angew. Chem. Int. Edit. 2001, 40, 1340.

13. Morton, M., Piirma, I.; J. Am. Chem. Soc. 1958, 80, 5596.

14. Koubenakis, A., Frankevich, V., Zhang, J., Zenobi, R.; J. Phys. Chem. A 2004, 108, 2405.

15. Sadeghi, M., Wu, X., Vertes, A.; J. Phys. Chem. B 2001, 105, 2578.

Networks applications: synthesis and properties

87

CHAPTER 6

NETWORKS APPLICATIONS: SYNTHESIS AND PROPERTIES

I) Introduction

To study the influence of molecular weight distribution and functionality-type distribution of

starting polymers on the network formation, network structure and material properties in cured

thermoset resins, a comparison between end-functionalized telechelic polymers with a narrow

molecular weight distribution and random co-polymers with about the same average

functionality by number and molecular weight by number are needed. The carboxylic acid-

functional poly(butyl acrylate) polymers, whose synthesis and characterizations were described

in chapter 4 are used as telechelic starting compounds. In addition to those RAFT polymers,

several random copolymers of butyl acrylate and acrylic acid were synthesized. The number-

average molecular weight and average functionality per polymer chain of those random

copolymers were nearly identical to the telechelic ones. A comparison between the two kinds of

polymers will allow us to have a clear view on the influences of the distributions of both

parameters on the network structure and on the mechanical properties of the cured coatings

made from them.

This chapter will first focus on the mechanism of cure and the rate of cure of these polymers

using triglycidyl isocyanurate (TGIC) as a cross-linker and 1,4-diazabicyclo[2.2.2]octane

(DABCO) as a catalyst. Attenuated Total Reflection Fourier-Transform InfraRed (ATR-FT-IR)

spectroscopy will be used to study these aspects for both types of polymers (random copolymers

and telechelic ones). The elastic moduli of the final cured coatings will then be determined by

micro-indentation. A particular attention will be given to the influence of the dangling chains on

these moduli. Information about the final network structures will be provided by analysis of

these cure materials with solid-state nuclear magnetic resonance spectroscopy (NMR) and

temperature modulated differential scanning calorimetry (TMDSC).

Chapter 6

88

II) Experimental

II.1) The Attenuated Total Reflection Fourier-Transform InfraRed spectroscopy

Attenuated Total Reflection Fourier-Transform InfraRed spectroscopy technique relies on the

fact that reflection occurs when a beam radiation passes from a denser medium to a less dense

medium. Beyond a certain critical angle of incidence the reflection is almost complete. It has

been proven both theoretically and experimentally, that in fact the beam penetrates into the less

dense medium. The depth of penetration can vary from a fraction of a wavelength to several

wavelengths, depending of the wavelength used, the incidence angle, and the difference of the

density of the two media. The penetrating radiation eventually interacts with the less dense

medium, attenuating the beam at wavelengths of absorbance bands. The beam radiation after

interaction with the media is recorded, and by comparison with the original radiation, an

absorption spectrum is obtained.

Figure 6.1. Schematic representation of the trajectory an infrared beam in an ATR crystal

coated with a polymer

In the setup used here, a single reflection of an IR beam was used to determine the change in

the absorption during cure (Figure 6.1).

One difference between ATR-IR and classical absorption infrared spectroscopy is that the

spectra acquired are not thickness-dependant, as the radiation penetrates only a few micrometers

into the sample. The peaks observed are similar to the ones observed in ordinary absorption

infrared. The main advantage of ATR-IR is that absorption spectra are readily available for a

variety of samples with a minimum of preparation. However, homogenous materials are needed.

Polymer

To detector From radiation source

Heating plate Diamond

crystal


89

The curing reactions were followed by attenuated total reflection (ATR)-FTIR on a Bio-Rad

Excalibur FTS3000MX infrared spectrophotometer using the golden gate setup (20 scans per

spectrum with a resolution of 4 cm−1) equipped with an ATR diamond unit.

II.2) Transmission Fourier-Transform InfraRed spectroscopy

A thin film of the reaction mixture was applied on a zinc selenide (ZnSe) window.

During the curing reaction IR spectra were recorded using a Biorad UMA 500 infrared

microscope that was coupled to a Biorad FTS 6000 FTIR spectrometer. The microscope was

equipped with a broad band MCT detector. This window was heated to the reaction temperature

by means of a Linkam THMS 600 hot stage. The heating rate was 30 ° C/min.

The kinetic mode of the Biorad WinIR pro software was used to record the spectra during the

reaction. The reaction time was set to 30 min., the time resolution was 20 sec. Spectra were

recorded with a resolution of 4 cm-1

II.3) Temperature Modulated Differential Scanning Calorimetry (TMDSC)

TMDSC experiments were carried out employing a modified Perkin-Elmer DSC-7 apparatus

calibrated for cell constant and temperature using indium standards. The instrument was

modified using a commercial precision function generator to allow a sinusoidal temperature

modulation. The measurements were carried out under nitrogen atmosphere to prevent

degradation of the polymer samples upon heating. The masses of the pans on the reference side

and the sample side were balanced to give a zero signal for the baseline. The experiments were

carried out with a global temperature ramp of 2 K/min with an oscillatory frequency of 12.5

mHz and a temperature amplitude of 0.06 K.

II.4) Micro-indentation

Indentation experiments were carried out at room temperature and ambient atmosphere using a

home built apparatus1. In general, in one run, 25 indentations were made using a maximum

loading of 1, 2 or 4 mN with a penetration rate of 10 nm/s. The calibration procedure of Oliver

and Pharr2 was used to correct for the load frame compliance of the apparatus and the imperfect

shape of the indenter tip. The compliance of the system was determined to be 0.3 nm/mN and

Chapter 6

90

hch max

P a b c d

P

h ch resh h max

S Loading

Unloading

P max

from the projected area of the indenter A, the contact depth hc was calculated using the relation

[6.1] with a=24.5 and b=5.71 µm.

A=ahc2+bhc [6.1]

The tip radius was 0.9 µm. The area function of this indenter was calibrated using B270 glass

(Schott, Jena, Germany), with an elastic modulus that was determined independently to be

75±1 GPa using the pulse-echo method. The calibration was performed using B270 glass for a

depth range of 0.1–2.9 µm, where the maximum indentation depth was restricted by the load

limitation. A typical load-displacement curve for a viscoelastic materialobtained after a micro-

indentation experiment is presented in Figure 6.2.

The effective moduli (Eeff) of our coatings at different loadings were calculated using relation

[6.2].

A

SEeff 2π

β=

1 [6.2]

where β is a constant, related to the geometry of the indenter (β = 1.034 for a Berkovich

indenter, used here) and S the unloading stiffness.

Figure 6.2. Schematic representation of load P versus indenter displacement h. (a) Initial

surface; (b) surface profile after load removal; (c) indenter; (d) surface profile under load.


91

Ef, Hf

E, H Eapp

Es, Hs

Es, Hs

hc/hf

Happ

Eapp

Happ Eeff

c, H c

Eeffs, H

s

Eeffs, H

s

In the case of a micro-indentation experiment performed on a coating, the obtained E value

can represent not only the intrinsic properties of the top layer, but a combination of the

mechanical properties of both the coating and the substrate, especially if the penetration depth is

high compared to the coating thickness, and if a large difference exists between the elastic

modulus of the coating and the one of the substrate. It has been shown that even for rubbery

materials, the influence of the substrate on the Eeff can be neglected if the penetration depth of

the indenter is less than 1/10 of the total thickness of the coating3. The elastic modulus and the

hardness values can be typically represented as in Figure 6.3.

Figure 6.3. Schematic representation of the apparent elastic modulus Eapp and hardness Happ

versus relative indenter displacement hc/hf (Eeffc, H c, hf = effective elastic modulus and hardness

of the substrate; Eeffs, H s)

As can be seen in Figure 6.4, The Eeff measured within the accuracy of measurement is at least

not influenced by the substrate for our telechelic and random copolymer-based networks when

the total penetration depth (hc) is less than 10% of the total thickness of the coating (hf).

The elastic modulus of the cross-linked coating itself was determined using relation [6.3],

assuming that the telechelic-based networks and the random copolymer-based networks behave

like a perfectly elastic material (ν = 0.5).

i

i

eff EEE

22 -1-11 ν+

ν= [6.3]

where E and ν are Young’s modulus and Poisson’s ratio for the sample analyzed, Ei and νi

Young’s modulus and Poisson’s ratio for the indenter.

Chapter 6

92

Figure 6.4. Evolution of effective elastic modulus versus the penetration depth of the indenter :

thickness of the film for coating formulations ratio with [epoxy]:[COOH] ratio of 1.4 :1

catalyzed by DABCO (7 w/w % of TGIC) for telechelic poly(butyl acrylate) networks (a) and

random poly(butyl acrylate-co-acrylic acid) networks (b) (hf being the thickness of the coating)

In practice, in this chapter, the Eeff values were used to compare results. These values were

obtained from measurements with a total penetration depth of less than 10% of the total coating

thickness. The values presented were values averaged over three calculated unloading

stiffnesses S using loads of 1, 2 and 4 mN. The error of measurement in these average values

was ± 10%. The data obtained are presented in Figure 6.14 and 6.17. The data presented are for

different [epoxy] : [COOH] ratios in the coatings formulations. The [epoxy] : [DABCO] ratio

was always kept the same (1 : 0.07 in weight). Phenomena such as pile-up and sink-in, which

may lead to some overestimation or underestimation of the contact area and consequently to

errors for the elastic moduli and the hardness, were not taken into account.

Figure 6.5 shows that both types of cross-linked coatings are purely elastic and that the value

of 0.5 for the Poisson’s ratio could be used in equation [6.3]. The variation observed in the

loading and unloading curve can be explained by hysteresis. The larger amount of hysteresis

observed in the cross-linked random copolymers will be explained later in the Results and

Discussion part (Section IV).

(a) (b)

0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.500

1

2

3

4

5

6

Load = 1 mN Load = 2 mN Load = 4 mN

Eef

f (M

Pa)

hc / hf 0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50

0

1

2

3

4

5

6

Load = 1 mN Load = 2 mN Load = 4 mN

E eff (M

Pa)

ht / hf


93

Figure 6.5. Load-displacements curves for coating formulations ratio with [epoxy]:[COOH]

ratio of 1.4 :1 catalyzed by DABCO (7 w/w % of TGIC) for random poly(butyl acrylate-co-

acrylic acid) networks (a) and telechelic poly(butyl acrylate) networks (b)

II.5) Solid-state Nuclear Magnetic Resonance

Proton-decoupled solid-state 13C-NMR spectra were recorded on a Bruker DMX500

spectrometer, operating at a 1H and 13C frequency of 500.13 and 125.13 MHz respectively. A

4mm magic angle-spinning (MAS) probe head was used in static mode as well as with a sample

rotation rate of 6 kHz. The radio-frequency power was adjusted to obtain 5µs 90˚ pulses both for

the 1H and 13C nuclei. The 38.56 ppm resonance of adamantine was used for external calibration

of the 13C chemical shift.

Proton spin-lattice relaxation in the laboratory and in the rotating frame T1(1H) and

T1ρ(1Η), respectively, were measured for each component of the polymers separately both via

cross-polarization (CP) to the 13C nuclei and via direct excitation of the 1H nuclei. Relaxation

delays in CP-derived experiments, D1, were 1 and 3 s, and the number of experiments per

relaxation data set, NE, was 20.

Proton transverse relaxation, T2, was determined at several temperatures ranging from -100˚C

to + 180˚C, using a BVT-3000 variable temperature unit. The decay of the transverse

magnetization was measured with the Hahn-echo pulse sequence (HEPS), 90˚-τ-180˚-τ-

(acquisition), where τ ≥ 2.5 µs. An echo signal is formed after the second pulse in the HEPS

with a maximum at time t = 2τ after the first pulse. By varying the pulse spacing in the HEPS,

the amplitude of the transverse magnetization, A(2τ), is measured as a function of time t = 2τ. T2

was determined by computer, fitting each 1H-NMR decay with a bi-exponential function.

0 5 10 15 200.0

0.5

1.0

1.5

2.0

2.5

Load

, mN

Displacement, µm

0 5 10 15 20 25 30

0.0

0.5

1.0

1.5

2.0

Load

, mN

Displacement, µm

(a) (b)

Chapter 6

94

III) Starting compounds

Telechelic poly(butyl acrylate): the sample VL103 was used for the experiments involving

telechelic polymers. Its average functionality per chain was determined to be 1.985 by liquid-

chromatography separation, and its number-average molecular weight is equal to 2.100 g/mol

with a polydispersity of 1.13. More details for the synthesis and characterization can be found in

Chapter 4.

Random poly (acrylic-acid-co-butyl acrylate): In order to determine the effect of the molecular

weight distribution on the network formation and the network properties, some

poly(butylacrylate)/poly(acrylic acid) copolymers were synthesized. The recipes of the

polymerizations were designed in order to obtain number-average molecular weights and

number-average COOH-functionality as similar as possible to the ones of the poly(butyl

acrylate) synthesized by RAFT.

Butyl acrylate (20.0 g, 0.156 mol), acrylic acid (1.4 g, 2.10 × 10-3 mol) and AIBN (6.4 g, 3.89 ×

10-2 mol) in toluene (50 mL) were added dropwise to a 250 mL round-bottom flask containing

50 mL of toluene. The flask was heated to 80˚C. The reaction was allowed to proceed for 8 hrs.

Isolation of the copolymer was achieved by rotary evaporation.

Overall conversion: 99+% (GC); Mn = 2200 g/mol, MWD = 1.75 (GPC)

Based on the almost complete conversion, the average molar mass and the initial butyl acrylate /

acrylic acid ratio, the average carboxylic acid functionality per chain is assumed to be close to 2.

Triglycidyl isocyanurate (TGIC): the TGIC cross-linker was received from Aldrich and was

claimed to be 98% pure. However, Liquid-Chromatography – Mass Spectrometry (LC-MS) and 1H-NMR spectroscopy reveal the presence of impurities. It was estimated by 1H-NMR that up to

15 % of the expected epoxy functionalities were not present in the product. The interpretation of

the spectrum in order to determine the chemical structures of the impurities was not successful.

Purification by recrystallisation was known to be very difficult to perform, it was then decided

to use the TGIC as such. It the rest of this chapter the ratios [carboxylic acid] : [epoxy] are

expressed assuming a 100% purity of the TGIC. We were fully aware of the presence of

impurities, but the same batch of TGIC was applied to all the systems, which allows comparison

among the various experiments. The results presented in this chapter will then be semi-

quantitative rather than quantitative.


95

N N

N O

O

O

O

O

O

N N

TGIC

1,4-diazabicyclo[2.2.2]octane (DABCO): DABCO (98+%, Merck) was used as received.

DABCO

CH3COOH (Acetic acid, 99+%, Aldrich) and acetonitrile (HPLC grade, 99+%, Biosolve) were

used as received.

Sample preparation

In general, mixtures of polymer – TGIC – DABCO were prepared using the following ratios:

- [epoxy] : [COOH] = 1.6 : 1 (molar ratio)

- [epoxy] : [DABCO] = 1 : 0.07 (weight ratio)

Variation of the [epoxy] : [COOH] ratio were employed. The same ratios were employed for

the model experiment using acetic acid.

ATR-FT-IR experiments: the polymer, the cross-linker and the catalyst were initially dissolved in

acetonitrile in order to obtain a homogenous solution. The acetonitrile is then removed under

reduced pressure at room temperature. The solventless mixture is then directly applied on the

ATR-FT-IR crystal, preheated at the desired temperature under a nitrogen atmosphere. The time

zero point corresponds to the time of application of the sample on the golden gate. The layer

thickness was a few microns.

Transmission FT-IR experiments: the sample preparation was similar to the one of the ATR-FT-

IR samples. A thin film of reaction mixture was then applied on a zinc selenide (ZnSe) window.

The time zero point corresponds to the time when the heating up of the ZnSe window starts.

Chapter 6

96

TMDSC: Typically, for every measurement, about 2 mg of cured polymer coating was weighted

out using a precision balance. Standard aluminum pans of similar predetermined masses were

used for all the measurements. The polymer, the cross-linker and the catalyst were initially

dissolved in acetonitrile (50 w/w %) and applied on an aluminum substrate. The samples were

then cured overnight at 120˚C, and the coating was scratched out of the plate with a scalpel.

Micro-indentation experiments: the polymer, the cross-linker and the catalyst were initially

dissolved in acetonitrile (50 w/w %) and applied on an aluminum substrate. The plate is placed

under a ventilated fume hood for 30 min, in order to evaporate some of the solvent and avoid

crater formation during the cure. The aluminum plates are then placed overnight in an oven at

120˚C for cross-linking to occur. The thickness of the coatings after cure was about 100 µm.

Solid-state NMR experiments: the polymer, the cross-linker and the catalyst were initially

dissolved in acetonitrile (50 w/w %) in a 10 mL glass vial. The vial was then placed overnight in

an oven at 120˚C to cross-linking purposes. The cured material was scraped out of the vial and

about 100 mg of it was placed in a solid-state NMR rotor for analysis.

IV) Results and Discussion

IV.1) Infra-red results

IV.1.1) Curing mechanism

The acid curing of epoxy is widely used in the coatings industry. Although the reaction can

proceed at high temperature, without the addition of an extra ingredient, often catalysts are used

in order to lower the curing time and/or temperature. Through the years various catalytic

systems were employed: ammonium salts4, metal salts5-7, salts of carboxylic acids8 and tertiary

amines9-17. A lot of the studies used anhydrides as the acid component and therefore, the

mechanism proposed in some of these articles must be different from our situation where we use

carboxylic acid groups. Based on our experiments done so far and on the literature available, we

propose the following mechanism for the reaction of epoxy groups with carboxylic acids in the

presence of DABCO, for our TGIC, telechelic polymer and DABCO mixtures:


97

RO

OH2 R

O

O H

H O

OR

2 N N + RO

O H

H O

OR

k1

k-1N NH OOC R,2

N NH OOC R, +O

R'OOC R,

O

R'N NH

k-2

k2

,

Complex A

Complex B

OOC R,O

R'N NH, R

O

O CH2 CH R'

OH

N N+k3

Figure 6.6. ATR-FT-IR absorption spectra during the curing of telechelic poly(butyl acrylate)

by TGIC ([epoxy] : [COOH] = 1.6:1), catalyzed by DABCO (7 w/w % of TGIC), at a

temperature of 100˚C.

As it can be seen in Figure 6.6, a change in absorption during cure of the mixture of telechelic

polymer, TGIC and DABCO can clearly be observed by transmission FT-IR. A decrease in

absorption of a shoulder at 1700 cm-1 and a peak at 900 cm-1 are observed. Both are known to be

20 min

3500 3000 2500 2000 1500 1000

0.35

0.30

0.25

0.20

0.15

0.10

0.05

0.00

W b

Abso

rban

ce

Wavelength (cm-1)

Chapter 6

98

1700 1650 1600 1550 1500 14501700 1650 1600 1550 1500 14501700 1650 1600 1550 1500 14501700 1600 15001700 1650 1600 1550 1500 1450

40 min20 min10 min

5 min0 min

wavenumber (cm-1)

characteristic absorptions for the C=O of the COOH dimer and the C-O of the epoxy groups

respectively, which are present in the reacting mixture, showing that at the temperature used

here, the cross-linking of the components occurs. No absorption for the formation of the new

ester group was found. The C=O stretching absorption of this new ester group is expected at

about 1730 cm-1 which coincides with the large C=O stretching vibration of the acrylate group

present in the telechelic polymer.

Interestingly, following the thermal curing with ATR-FT-IR, we were able to notice a band

related to an intermediate at about 1560 cm-1 which appeared and disappeared during cure

(Figure 6.7). The following reasoning always relies on data obtained following the curing of

telechelic poly(butyl acrylate), but the same was observed for random poly(acrylic acid-co-butyl

acrylate).

Fedoseev et al.18, who studied the cure of acrylic acid and ethylene oxide observed the

formation of a band during cure. They employed pyridine as a catalyst, used NaCl and KBr

pellets as their medium in the IR experiments. They assigned the band at about 1550 cm-1 to the

product of the reaction between the carboxylic acid and the KBr medium. However, in our

studies, the ATR diamond crystal used is inert and can not react with the carboxylic acid,

forming a carboxylate anion. Hence, this band has to be attributed to an intermediate that is

produced during the reaction. The position of the intermediate band at 1560 cm-1 strongly

suggests that the intermediate formed contains a carboxylate anion. The change in absorption of

this band over cure time at various cure temperatures is presented in Figure 6.8.

Figure 6.7. ATR-FT-IR absorption spectra during the curing of telechelic poly(butyl acrylate)

by TGIC ([epoxy] : [COOH] = 1.6:1), catalyzed by DABCO (7 w/w % of TGIC), at a

temperature of 100˚C.


99

0 20 40 60 80 100 120

0

20

40

60

80

100

0 20 40 60 80 100 120

0

20

40

60

80

100

0 20 40 60 80 100 120

0

20

40

60

80

100

120oC

100oC

80oC

Abso

rptio

n ch

ange

in th

e 15

60 c

m-1 b

and

A.U

.

Time (min)

In order to get more information about the structure of this intermediate, test experiments were

performed. No band at 1560 cm-1 was observed during the heating at 100ºC of mixtures of

telechelic polymer and DABCO, TGIC and DABCO, or telechelic polymer and DABCO. The

ratios of the different components in these mixtures were kept the same as in a typical cure

reaction. No reaction at all was observed for the first two mixtures over a heating period of 1

hour. In the last mixture, a decrease in adsorption of the C=O band of the acid dimer end-group

was observed at 1700 cm-1. However, the rate of disappearance was much slower compared to a

curing experiment in the presence of DABCO. Apparently, the band at 1560 cm-1 is only

observed when the 3 components (acid dimer – cross-linker – catalyst) are present. This

suggests that a complex between the 3 species is formed during the cure reaction and that the

band at 1560 cm-1 may be attributed to the C=O band of the COO- group in complex B of the

proposed mechanism.

Figure 6.8. Change in absorption over time of the 1560 cm-1 band at different curing

temperatures of a telechelic poly(butyl acrylate), TGIC ([epoxy] : [COOH] = 1.6:1), DABCO (7

w/w % of TGIC) mixture

It is interesting to notice in Figure 6.8 that after the building up of this intermediate, this

compound is still present at the end of the cure. The level of this absorption seems to vary with

the reaction temperature. However, when it is taken into account that generally the intensity of

the C=O band of a carboxylate anion group is 2-3 times stronger compared to the one of a

carboxylic acid, the final amount left may be very small (a few percent of the original acid

Chapter 6

100

0 10 20 30 40 50 60 70 80 90 100 110 120 130

0

20

40

60

80

100

0 10 20 30 40 50 60 70 80 90 100 110 120 130

0

20

40

60

80

100

0 20 40 60 80 100 120

0

20

40

60

80

100

Abs

orpt

ion

chan

ge in

the

900

cm-1 b

and

A.U

.

Time (min)

100oC

120oC

80oC

0

20

40

60

80

100

0 20 40 60 80 100 120

0

20

40

60

80

100

0 20 40 60 80 100 120

0

20

40

60

80

100

120oC

100oC

80oC

Abs

orpt

ion

chan

ge in

the

C=O

ban

d at

170

0 cm

-1 A

.U.

Time (min)

group) and the reaction of this final amount may be slowed down considerably by the decrease

of chain mobility by network formation at the end of the cure reaction.

Figure 6.9. Decrease of the 1700 cm-1 band at different temperatures for the curing of telechelic

poly(butyl acrylate) by TGIC ([epoxy] : [COOH] = 1.6:1), catalyzed by DABCO (7 w/w % of

TGIC)

A continuous decrease is observed for the intensity of the 1700 cm-1 characteristic of the

carboxylic acid dimers (Figure 6.9), and also for 900 cm-1 band, characteristic of the C-O in the

epoxy group by ATR-FT-IR (Figure 6.10). However, it was more difficult to follow its

disappearance, as the region contains other peaks that interfere with the one belonging to the

epoxy group.

Figure 6.10. Decrease of the 900 cm-1 band at different temperatures for the curing of telechelic

poly(butyl acrylate) by TGIC ([epoxy] : [COOH] = 1.6:1), catalyzed by DABCO (7 w/w % of

TGIC).


101

0 5 10 15 20 250 5 10 15 20 250 5 10 15 20 25

3006 cm-1

1578 cm-1

Abs

orba

nce

A.U

.

Time (min)

Comparing the first part of the reaction in Figure 6.8 and 6.10, the spectra suggest that the

intermediate responsible for the band at 1560 cm-1 also influences the 900 cm-1 band. The

changes in the slopes at different curing temperatures in Figure 6.10 coincide with the maxima

observed in Figure 6.8. The proposed intermediate molecular structure contains also an epoxy

group which is expected to absorb at about 900 cm-1. The slow initial decrease in absorption at

900 cm-1 over time strongly suggests that initially the complex B is formed, while the epoxy is

slowly consumed. We suggest that when most of the catalyst has disappeared, the reaction

accelerates, provoking the changes in Figure 6.8 and 6.10. More investigation is needed to

confirm this speculation.

A model experiment was conducted, following the reaction at 100ºC of acetic acid, TGIC

([epoxy] : [COOH] = 1.6:1) and DABCO (DABCO (7 w/w % of TGIC). The IR-spectra were

easier to analyze in the C=O band region, as the massive band of the butyl acrylate of our

polymers was not present. We could clearly see that a great majority of the carboxylic acid

groups are presents as dimers at time zero of the reaction (νC=O=1700 cm-1), while the presence

of monomer (νC=O=1750 cm-1) was very small. We can see in Figure 6.11 that the intermediate

at 1560 cm-1 is also formed in this model reaction and that its concentration varies similarly to

the intermediate found in the curing of the telechelic polymer.

Figure 6.11. Variation of various bands at 90ºC for the reaction of acetic acid with TGIC

([epoxy] : [COOH] = 1.6:1), catalyzed by DABCO (7 w/w % of TGIC).

Chapter 6

102

This IR experiment was performed in the transmission mode, which allowed us to obtain a

stronger signal in interesting regions of the spectra. The variations of the following bands were

followed:

- 1578 cm-1: Symmetric stretching of the C=O band of the carboxylate anion

- 3006 cm-1: Symmetric stretching of the C-H band of the epoxy group

The disappearance of the epoxy group seems to proceed in two steps similarly to what was

found previously. Acceleration in the epoxy consumption happens after a maximum in the

intermediate concentration is reached. The same reasoning that was made on the curing of the

telechelic polymers by TGIC seems to be applicable for this model reaction as well

In the curing of random copolymers, the intermediate at 1560 cm-1 was also detected. We can

then propose the mechanism presented for our TGIC, telechelic polymer and DABCO mixtures

as a general reaction mechanism for epoxy groups with carboxylic acids in the presence of

tertiary amines.

IV.1.2) Kinetics differences: random vs. telechelic

As it can been seen in Figure 6.12, the disappearance of the acid dimer peak is much faster in

the case of telechelic polymer than for the random copolymer case. In both cross-linking

reactions, the concentration of the acid groups, epoxy groups and catalyst are the same, within

the error of measurement. Several explanations for the difference observed can be given using

steric considerations. For the telechelic poly (butyl acrylate), the distance between carboxylic

acid groups is roughly 2.000 g/mol. Once one carboxylic acid has reacted with an epoxy, the

other carboxylic acid of the chain is still available, in a reasonably mobile form, being far away

from the reacted one. For the random copolymer chains bearing 2 or more acid groups, the

distance between the two carboxylic acid groups can be much smaller. It is on average, largely

inferior to the one encountered in the telechelic polymer case. After reaction of a carboxylic acid

group, the mobility of the closest unreacted acid group is on average reduced, in a much more

pronounced fashion than for the telechelic one. When a copolymer chain contains more than two

acids groups, this effect is even more pronounced. This lack of mobility/availability of the

carboxylic acid groups may be one of the reasons behind the difference of the cure reaction rate

shown in Figure 6.12.

Another explanation for the difference in rate might be the concentration of acid dimers

between two carboxylic acid end-groups. There might be differences in the dimer formation

probability, possibly leading to a lower concentration of acid dimers for the random poly


103

(acrylic acid-co-butyl acrylate). This ultimately may lower the formation of complex B, and may

explain the rate difference observed in the FT-IR studies. Unfortunately, the ratio between

monomers of acid group and dimers of acid could not be verified, because the absorption at

about 1750 cm-1, where the band for the C=O of the acid monomer lies, was overlapped by the

very strong C=O band of the acrylate group present in the polymers.

Figure 6.12. Decrease of the 1700 cm-1 band at different temperatures for the curing of carboxy-

functional polymer by TGIC ([epoxy] : [COOH] = 1.2:1), catalyzed by DABCO (7 w/w % of

TGIC); (a) telechelic poly (butyl acrylate); (b) random poly(acrylic acid-co-butyl acrylate)

In literature, a mechanism for the catalysis by tertiary amines of the carboxylic acid – epoxy

reaction was proposed by Kucharski and Rubczak12 (Figure 6.13):

0 50 100 150 2000.0

0.2

0.4

0.6

0.8

1.0

0 50 100 150 2000.0

0.2

0.4

0.6

0.8

1.0

0 50 100 150 2000.0

0.2

0.4

0.6

0.8

1.0

Abs

orpt

ion

chan

ge in

the

C=O

ban

d at

170

0 cm

-1 A

.U.

Time (min)

100oC

120oC

80oC

(a)

0 100 200 300 400 500 600 700 8000.0

0.2

0.4

0.6

0.8

1.0

0 100 200 300 400 500 600 700 8000.0

0.2

0.4

0.6

0.8

1.0

0 100 200 300 400 500 600 700 8000.0

0.2

0.4

0.6

0.8

1.0

100oC

120oC

80oC

Time (min)Abs

orpt

ion

chan

ge in

the

C=O

ban

d at

170

0 cm

-1 A

.U.

Time (min)

(b)

Chapter 6

104

N N HOOC R N NH OOC R+ ,

N NH OOC R, +O

R'OOC R,

O

R'N NH+

OOC R,O

R'R

O

O CH2 CH R'

O

R

O

O CH2 CH R'

ON NH+ R

O

O CH2 CH R'

OH

N N+

k1

k-4

k4

k3

k-2

k2

k-1

Figure 6.13. A proposed mechanism for the catalyze by DABCO of the carboxylic acid – epoxy

reaction

The authors stated using a steady-state approximation on the carboxylic anion – quaternary

ammonium cation complex, and assuming that 1-12 , kkk << that:

O]COOH][DABC-R[K]COO-R, HDABCO,[ 1- =+ [6.4]

where 1-

11K

kk

=

Unfortunately, we can affirm that the proposed mechanism or the assumptions made were

incorrect. The disappearance of the acid groups can be expressed by:

O]COOH][DABC-R[-]COO-R,H-DABCO[dt

d[COOH]- 1-

1 kk-+= [6.5]

By substituting equation [6.5] in equation [6.4], we obtain:

0dt

d[COOH]- =

which can obviously not be true. It seems that the steady-state approximation used by the

authors is not realistic. The band at 1560 cm-1 surely represents a complex that is formed during

the cross-linking reaction. Figure 6.8 clearly shows that the steady-state approximation can not


105

be applied to this compound. We believe that this is the point where Kucharski and Rubczak’s

kinetic investigation is incorrect. However, their proposed mechanism for the epoxy-carboxylic

acid reaction is more detailed than the one we have proposed at the beginning of this chapter. It

was impossible for us to investigate thoroughly the mechanism with the help of techniques such

as FT-IR or NMR, due to time constraints.

IV.2) Determination of E-modulus

IV.2.1) Telechelic poly(butyl acrylate) micro-indentation measurements

The Eeff for coatings based on telechelic poly(butyl acrylate) cross-linked with TGIC with be

referenced as Eefftel and are presented in Figure 6.14. The effect of the [polymer] : [cross-linker]

ratio on this mechanical property appeared to be large.

Figure 6.14. Eeff tel values for VL103 telechelic polymer cross-linked with TGIC with various

[epoxy] : [COOH] ratios.

It can be noticed in Figure 6.14 that the Eefftel increases continuously when an increasing

amount of epoxy functions is present in the formulation before cross-linking.

0

5

10

15

20

25

30

Eef

f (M

Pa)

11.21.41.622.42.83.2

1mN 2mN 4mN

Loading force

Chapter 6

106

As it was shown before (equation [6.3])

i

i

eff EEE

22 -1-11 ν+

ν= [6.3]

and considering that Ei >> E and that the cross-linked polymer behaves like a perfectly elastic

material (ν = 0.5) we obtain:

0.75tel teleffE E= [6.6]

For the sake of clarity, all the reasoning will be done on Eefftel (unless otherwise mentioned).

The conclusions will of course also apply to Etel as well.

In a different form, a part of the data of Figure 6.14 is presented again in Figure 6.15.

Figure 6.15. Evolution of the effective elastic modulus with various amounts of

[epoxy]:[COOH] ratios for telechelic-based poly(butyl acrylate) networks. Loading force = 1

mN

The curve displayed in Figure 6.15 was derived from Figure 6.14, using the values obtained

with a loading force of 1 mN. It is usually typical for a coating to have values for E increasing

up to a certain [polymer]:[cross-linker] ratio (typically 1:1), and then to observe a decrease of

the E modulus when the cross-linker is in excess. This is easily understandable, as for ratios

different from 1:1, the structure of the network will be imperfect, and defects such as dangling

ends will have a tremendous effect in decreasing the E modulus. The molecular weight between

cross-links might also be different. However in our case, the effective modulus increases

0

5

10

15

20

25

0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 2.8 3 3.2 3.4[Epoxy]:[COOH] ratio

Eef

f (M

Pa)


107

continuously. For [epoxy] : [COOH] ratios above 2, the Eeff values reaches numbers that are

difficult to assign to soft poly(butyl acrylate) networks. The well-known polyetherification of

TGIC, catalyzed by bases is probably the reason behind this huge increase in the modulus at

high [epoxy] : [COOH] ratios

In order to confirm this, the curing conditions were applied to a mixture of only TGIC and

DABCO, with no polymer. After curing a yellow brittle film was obtained. The TGIC reacts

with itself in a poly-condensation reaction, forming a polyether. With three functionalities per

organic molecule, the reaction will ultimately lead to the formation of a three-dimensional

insoluble film. The film obtained contains a lot of hard isocyanurate rings, which confer a high

E modulus to the film.

Interpreting this experiment, it seems likely to conclude that a competitive reaction (auto-

polymerization of TGIC) occurs next to the carboxylic acid – epoxy reaction. When an excess of

TGIC is used, clusters of hard polyethers are present in the matrix of poly(butyl acrylate). When

the [epoxy] : [COOH] ratio was raised to 3.2, a phase separation was visually noticeable. It is

then logical to assume that for the [epoxy] : [COOH] ratios superior to 2 (when the E values rise

to improbable values), the Eefftel measured is not a parameter only influenced by the structure of

the poly(butyl acrylate) network (formed by epoxy-carboxylic acid reactions), but rather a

number relative to the modulus of a poly(butyl acrylate) network, with polymerized TGIC

included as network domains and/or clusters inside it.

It is also known that trans-esterification might occur between the alcohol group resulting from

the epoxy-carboxylic acid reaction and the butyl acrylate-side chain. Although this phenomenon

may occur, it is difficult to explain the high values (for example 25MPa for the 3.2:1 ratio) only

with this reaction, as the chemical structure of the network should be roughly the same. It is only

hard domains of polymerized TGIC that can cause the elastic modulus to rise to such values.

The principal question that remains concerns the influence of this autopolymerization side-

reaction on the values obtained for the elastic modulus for [epoxy] : [COOH] ratios below 2. We

will answer this question in the last part of this chapter using solid-state NMR experiments.

Rubber elasticity theory

In the case of perfectly alternative elastomeric networks, resulting in a Poisson’s ratio of

nearly 0.5, the modulus, E, can be expressed as the following equation using the Gaussian

rubber elasticity theory19-21.

Chapter 6

108

3 eE RTν=

[6.7]

where R is the gas constant and T the absolute temperature and νe the density of elastically

active chains.

For our telechelic polymer we can assume that the density of cross-link points can be related

to the molecular weight of the polymer, and we obtain:

cM

RT3ρ=E [6.8]

where ρ is the density of the poly(butyl acrylate) gel, Mc, the molar mass between junctions,

R, the gas constant and T, the absolute temperature. This formula is valid well-above Tg. The Tg

of our polymers was experimentally determined by DSC and are in the range of -10˚C

(experimental data will be provided in the next section of this chapter). Performing the micro-

indentation experiments at room temperature, we may assume that this rubber elasticity theory is

applicable to our coatings. It is generally assumed than a minimum of 20 monomer units in the

polymer chain is needed for the rubber elasticity theory to be applicable. We are with our

systems a little bit below this limit, but we still consider that comparing our experimental results

to the theoretical ones is relevant.

When we assume that the molar mass between cross-links is equal to the molar mass of the

telechelic polymer, that is 2000 g/mol and a value of 1 for the density, the theoretical value for

the E modulus of 3.7MPa can be calculated using equation [6.8]. From this value, and assuming

a Poisson's ratio of 0.5, we calculated a theoretical value for Eefftel

of 4.9 MPa. This theoretical

value is very close to the value observed for [epoxy] : [COOH] ratios of 1.4, 1.6 and 2. Hence,

the most “perfect” network in our coatings is probably at a much higher [epoxy] : [COOH] ratio

than 1:1. Further evidence for this will be presented in the solid-state NMR part of this chapter.

One could argue that due to the complexity of the network topology, it is not possible to

compare our experimental values with the ones obtained using the Gaussian rubber elasticity

theory. It is however interesting to notice that the obtained experimental values are very similar

with those theoretically expected for a perfectly alternating network. A much stronger difference

would have been expected beforehand. The Gaussian model used here does not take into


109

account network defects such as loose ends, loops and entanglements. The first two factors will

lower the Eefftel value whereas the last factor will enlarge it. The difference between the

experimental and the theoretical values for the Eefftel value can then be explained by the network

defaults.

The deviations at ratios above 2 might be attributed to the presence of the polyether network

produced by the autopolymerization of TGIC. The deviation with the rubber elasticity theory at

lower [epoxy] : [COOH] ratios are possibly caused by the larger Mc values between cross-links

and/or the presence of more dangling ends in the networks. The influence of these factors on the

elastic modulus of the network will be discussed in more details in the next section of this

chapter.

IV.2.2 Influence of mono-functional chains on the mechanical properties of poy(butyl

acrylate)networks

The synthesis of well-defined linear mono-functional poly(butyl acrylate) allowed us to

introduce a controlled amount of defects (dangling ends) in the network structure. We select the

[epoxy] : [COOH] ratio of 1.6 for our next experiments. This choice was governed by the global

reliable results that were obtained in our previous investigations, and the belief that the influence

of the polyether formation is not yet predominant for this ratio. The coating formulations

prepared contained 5 w/w %, 10 w/w %, 20 w/w % of the total weight of polymers of mono-

functional polymer VL131 (Mn= 2.300 g/mol, MWD = 1.14). Due to the small differences

between the molar masses of the mono- and di-functional poly(butyl acrylate), the weight

percentage can roughly be approximated to number percentage.

As shown in Figure 6.16, the Eefftel decreased continuously with an increasing amount of

dangling ends. This effect is more severe when the first few percents of mono-functional chains

are introduced.

The plasticizing effect of the dangling ends is a known phenomenon in the literature. This

gives a reasonable explanation why the Eefftel values measured for [epoxy] : [COOH] ratios of 1

and 1.2 are lower than the one predicted by the rubber elasticity theory. Impurities in the cross-

linker lower the actual amount of epoxy groups in the formulation (about 15% less), and lower

the [epoxy] : [COOH] below stoechiometry at experimental ratios below 1.2.

Chapter 6

110

3

3.5

4

4.5

5

0 5 10 15 20 25

wt. % of dangling ends added

Eeff

(MPa

)

Figure 6.16. Evolution of the effective elastic modulus with various amounts of dangling ends

for telechelic-based poly(butyl acrylate) cross-linked with TGIC ([epoxy] : [COOH] = 1.6).

This may result in non-reacted COOH functions, leaving dangling ends in the network

structure. The dangling ends can also be obtained in another manner. At high conversion of the

cross-linking reaction, the remaining COOH functional groups might be hard to access for the

epoxy groups still present, for entropic reasons. The mobility of the chains is reduced by the

cross-linking points, allowing some dangling ends to persist, although macroscopically some

epoxy and carboxylic acid groups are still present in the network.

By addition of mono-functional polymer to the coating formulation, not only the amount of

dangling ends but also Mc increases. This increase will lower the Eefftel value and is another

explanation for the lowering of the moduli reported in Figure 6.16.

IV.2.3) Random poly (butyl acrylate-co-acrylic acid) polymers micro-indentation measurements

The Eeff for coatings based on random poly (butyl acrylate-co-acrylic acid) cross-linked with

TGIC will be referred to as Eeffran and are presented in Figure 6.17. The effect of the [polymer] :

[cross-linker] ratio on the effective elastic modulus was investigated.


111

0

1

2

3

4

5

6Ee

ff (M

Pa)

11.21.41.62

1mN 2mN 4mN

Loading force

Figure 6.17. Eeffran values for random copolymer cross-linked with TGIC with various [epoxy]

: [COOH] ratios.

Similar to the telechelic poly(butyl acrylate), all the reasoning will be done on Eeffran (unless

otherwise mentioned). The evolution of Eeffran with different [polymer]:[cross-linker] ratios for

random poly(butyl acrylate-co-acrylic acid) is presented on Figure 6.18.

As seen in Figure 6.18, we observe for the effective modulus of the random copolymers-based

networks a trend similar to the one of the telechelic polymers-based networks. Unfortunately no

coating at high [epoxy] : [COOH] ratios were prepared, it was impossible to compare both

systems for high ratios. For the random copolymers, the effective modulus continuously

increases with increasing amounts of cross-linker, similar to what was observed for the

telechelic polymers. The same explanation can be given, i. e. the formation of a more complete

network with a lower Mc and the presence of less-dangling ends at [epoxy] : [COOH] ratios of

1.6:1 and 2:1.

We observed in our experiments that the values of the effective modulus are lower for the

networks based on the random copolymers compared to the telechelic polymer-based ones. This

difference can be estimated at 20%. The lower values found are in contradiction with the

theoretical predictions of Dusek et al22. Their calculations show that random copolymers-based

networks have a higher elastic modulus compared to the telechelic polymers-based networks,

the average functionality per chain being the same.

The main factors explaining these low values may be the incompletion of the cross-linking

reactions, and the larger Mc. Another factor explaining this lowering might be the “non

Chapter 6

112

-35

-30

-25

-20

-15

-10

-5

0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7

[epoxy] : [COOH]

Tg (º

C)

Telechelic polymers (Tg uncured = - 61.2ºC)

Random polymers (Tg uncured = - 46.9ºC)

randomness” of the copolymer chains. Those defects in the random copolymers-based networks

are possibly responsible for the hysteresis phenomenon that was observed on the load-

displacement curves during the micro-indentation experiments for the random copolymer-based

coatings(Figure 6.5). The adhesion between the indenter tip and the coating can be another

reason.

Figure 6.18. Evolution of the effective elastic modulus with various amounts of

[epoxy]:[COOH] ratios for for random poly(butyl acrylate-co-acrylic acid) networks and

telechelic poly(butyl acrylate) networks. Loading force = 1 mN

VI.3) MTDSC measurements

Figure 6.19. Evolution of the glass transition temperature with various amounts of

[epoxy]:[COOH] ratios for for random poly(butyl acrylate-co-acrylic acid) networks and

telechelic poly(butyl acrylate) networks.

0

5

10

15

20

25

30

0 0.5 1 1.5 2 2.5 3 3.5

[epoxy] : [COOH] ratio

Eeff

(MPa

)

Random copolymersTelechelic polymers


113

-12

-11.5

-11

-10.5

-10

-9.5

-9

-8.5

-8

0 5 10 15 20 25

wt. % of dangling ends added

Tg (º

C)

For both systems (telechelic poly (butyl acrylate) and random copolymers) an increase of the

glass transition temperature with an increasing amount of the [epoxy] : [COOH] ratio is

observed (Figure 6.19). Those measurements are in line with the conclusions drawn from the

micro-indentation experiments.

We observe a decrease of the glass transition temperature with an increasing amount of

dangling ends as expected (Figure 6.20), confirming the results of the micro-indentation

measurements. The decrease in the glass transition temperature is more severe when the first

few percents of mono-functional chains are introduced as it was observed for the Eefftel.

Figure 6.20. Evolution of the glass transition temperature with various amounts of dangling

ends for telechelic-based poly(butyl acrylate) networks ([epoxy] : [COOH] = 1.6).

When the different values obtained for the telechelic cross-linked coatings (with and without

extra-added dangling chains) were plotted against the corresponding glass transition temperature

values, a linear relation was found within the accuracy of the measurement (Figure 6.21). No

such linear relation was found for the random copolymers- based coatings (Figure 6.22).

Chapter 6

114

0

0.5

1

1.5

2

2.5

3

3.5

4

-32 -30 -28 -26 -24 -22 -20

Tg (ºC)

Eeff

(MPa

)

0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

5

-18 -16 -14 -12 -10 -8 -6

Tg (ºC)

Eeff

(MPa

)

Figure 6.21. Evolution of the effective elastic modulus with the glass transition temperature for

telechelic-based poly(butyl acrylate) networks ([epoxy] : [COOH] = 1.6).

Figure 6.22. Evolution of the effective elastic modulus with the glass transition temperature for

random-based poly(butyl acrylate) networks ([epoxy] : [COOH] = 1.6).

It is known that for cross-linked coatings, the glass transition temperature and the elastic

modulus can both be described as a function of the molecular weight between cross-links23. This

can explain the linear relation observed in Figure 6.21. These data also suggest that the

variations found in the elastic modulus and in the glass transition temperature of the telechelic

polymer-based coatings are mainly determined by variations in Mc. In general, no linear relation

between the E modulus and the glass transition temperature is found. The numerous


115

imperfections in the random copolymer-based networks (especially the presence of free-polymer

chains with a very low elastic modulus) may be the cause for the deviation from linearity shown

in Figure 6.22.

VI.4) Solid-state Nuclear Magnetic Resonance results

VI.4.1) T1- relaxations experiments

Networks based on telechelic poly (butyl acrylate) cross-linked with TGIC (with [epoxy] :

[COOH] ratios of 1:1 and 1.6:1) were analyzed with solid-state NMR. Figure 6.23 shows the 13C

spectra of the starting compounds and of the cross-linked material.

The assignments of the different peaks of the 13C NMR spectrum of the VL103 telechelic

poly (butyl acrylate) are presented in Table 6.24. Figure 6.23.c strongly suggests that in this

formulation most of the epoxy (at 50 ppm) and acid groups (at 182 ppm) have disappeared after

cross-linking.

We have investigated proton spin-lattice relaxation in the rotating frame (T1ρ) and in the

laboratory (T1) via cross-polarization of the 13C nuclei to study possible domain formation. 1H-NMR T1 and T1ρ measurments yield mobility information of the considered nuclei at

respectively the nanosecond and millisecond time scale. This information can provide

indications about the homogeneity of the networks, and the possible presence and miscibility of

polymerized TGIC in our coatings.

As a result of the proton-proton dipolar coupling, protons in a polymer continuously exchange

their polarization (“nuclear magnetization”). This process, usually referred to as spin diffusion,

tends to average out local differences in NMR properties, such as relaxation. Spin diffusion is

fast in rigid polymers with closely interspaced protons, and slow in mobile polymers with low

proton density.

At a standard spin diffusion coefficient D of approximately 1 nm2.ms-1 (although this value

can be the subject of discussion considering the high mobility of the polymer chains), proton-

proton spin diffusion averages out any T1 or T1ρ relaxation difference for average domain size in

a polymer blend smaller than about 1nm,. All protons then decay with the same effective T1 or

T1ρ. In contrast, if the domain size is larger than about 50 nm, spin diffusion is too slow to

average out such differences, and each phase will decay with its intrinsic, probably different T1

or T1ρ values. Furthermore, since T1 tends to be 10 to 100 times longer than T1ρ, the effective

Chapter 6

116

diffusion path length is longer in T1 experiments. Therefore, in the intermediate range, for

domain sizes ranging from 1 to 50 nm, different effective T1ρ values and a single effective T1 are

expected. Spin diffusion is still able to average out T1 differences, although it is not able to

homogenize T1ρ.

For the domain analysis proton using T1 and T1ρ relaxation via cross-polarization to the 13C

nuclei, the signals in the aliphatic regions (0-70 ppm) and in the carbonyl region (140-180 ppm)

were studied.

Figure 6.23. Solid-state 13C NMR spectra obtained via direct excitation for (a) telechelic

poly(butyl acrylate) (b) TGIC (c) cross-linked network with [epoxy] : [COOH] ratio of 1:1.6

(* = solvent peaks)

(a)

(b)

200 180 160 140 120 100 80 60 40 ppm

200 180 160 140 120 100 80 60 40 ppm

200 180 160 140 120 100 80 60 40 ppm180 160 140 120 100 80 60 40 20 ppm180 160 140 120 100 80 60 40 20 ppm

(c)

* *


117

Carbon number Description Chemical shift (ppm)

1 CH3 butyl group (CH3CH2CH2CH2) 14



4 CH2 butyl group (CH3CH2CH2CH2) 64.5

5 COO acrylate group 174

5’ COO acrylate group, close to CS3 170

6 CS3 trithiocarbonate Theoretically: 220

7 CH backbone 40

8 CH2 backbone 37

9 Quaternary carbon of the end-groups 41

10 CH3 of the end-groups 25.5

11 End-group COOH 182

Table 6.24. 13C NMR peak attribution for VL103

1

CH2 CH

O O

S C

S

S CHCH2 CH

O O

HOOC CH2

OO

CH CH2

OO

COOH

23

4

5

78 6

5’ 5’9

10

11

Chapter 6

118

Figure 6.25. 13C-NMR spectra obtained at different T1ρ-filter times via cross-polarization to

the 13C nuclei, for cross-linked network with [epoxy] : [COOH] ratio of 1:1 (* = sideband) (a) 1H T1ρ-filter = 10 µs (b) 1H T1ρ-filter = 6 ms

The analysis of two 13C spectra with different proton T1ρ filters, as presented in Figure 6.25,

indicates that the poly (butyl acrylate) and the TGIC decay similarly. Hence, both compounds

show identical T1ρ behavior. The homogeneity is then proven on the nanometer scale. The T1-

relaxation experiments were then unnecessary, as they are giving information on the micrometer

scale.

The main conclusion of this T1ρ-relaxation experiment is the homogeneity of the formed

network after the epoxy-carboxylic acid reaction. Unfortunately the lack of time did not allow us

to investigate T1ρ-relaxation for formulations with higher epoxy content. It would have been

interesting to perform relaxation experiments on formulations that gave unrealistic high values

for the elastic modulus (such as 3.2 : 1 for example). Nevertheless, we are able to conclude that

the polyetherification (autopolymerization of the TGIC) reaction does not occur in an extended

-50150 100 50 0 ppm

* * *

* * *

(a)

(b)


119

a 2b

2 2(2 ) exp (1 ) expA K

τ ττ ν ν

2

= − + − −

Τ Τ

fashion, as no phase of TGIC could be recorded. We can then affirm that, when there is no

excess of epoxy groups versus carboxylic acid groups, the polyetherification is not competitive

to the epoxy – carboxylic acid reaction. The rate of this polyetherification is probable orders of

magnitude lower than the one of the epoxy – COOH reaction. The data obtained so far suggest

that the polyetherification only significantly occurs when almost all the carboxylic acid groups

have reacted.

VI.4.2) T2- relaxation experiments

T2-relaxation was investigated for telechelic-based poly(butyl acrylate) networks, with

[epoxy] : [COOH] ratios of 1:1 and 1.6:1 in order to collect information on molar masses

between cross-linking points. As reported by Litvinov24,25, the T2-relaxation time for elastomer

networks is sensitive to rotational motion of the polymer chains at temperatures about 100˚C-

150˚C above the glass transition temperature. More specifically, T2 depends on the amplitude of

collective segments motions. The mobility of a chain is dependant of the temperature. The

higher the temperature, the more mobile the chain is. However, beyond a certain point, the

mobility of the chain will not vary with the temperature anymore, its motions being restricted by

physical or chemical entanglements. In our case, it is naturally the cross-linking points that will

prevent the chain from moving freely. The T2-relaxation time becomes then independent of the

temperature, and on the T2 – temperature, a plateau value is obtained. At low temperatures, the

chain motion is characterized by local segment motion. When no cooperative motion occurs, the

value of the T2-relaxation time reaches a constant value. The T2 – temperature graph exhibits

then another plateau at low temperatures determined by the static dipolar proton-proton couple.

At higher temperature, the chain motion increases, and the dipole interactions become more and

more averaged.

In our proton – T2-relaxation experiments, the observed decays of the so-called Hahn-echo as

a function of the echo time, 2τ did not follow a mono-exponential behavior. However, it was

possible to fit these decays with a bi-exponential function, described by the following

equation26:

[6.9]

where K represents a constant related to the amplitude of the chain motions, ν is the fraction

of fast-relaxing segments in the networks, T2a and T2b are the T2-relaxation values respectively

Chapter 6

120

2

2

p

rlT

ZaT

=

150 200 250 300 350 400 4501

10

100

1000

10000

150 200 250 300 350 400 4501

10

100

1000

10000 [epoxy] : [COOH] 1:1 [epoxy] : [COOH] 1.6:1

T 2A (µ

s)

Temperature (K)

uc

ZC MM

n∞=

related to fast and slow-relaxing components in the networks. Due to the structure of the poly

(butyl acrylate) chain, it is logical to assign the T2a proton relaxation to the protons of the poly

(butyl acrylate) backbone, and the T2b proton relaxation to the protons of the butyl side-chains.

The ν parameter is then independent of the temperature.

The number of statistical segments, Z, may be determined using the following equation24,25:

[6.10]

where 2pT is the value of T2 at the high temperature plateau and 2

rlT is the value of T2,

independent of the network structure, in the glassy state. A geometrical coefficient, a, is

introduced and is dependant on the angle between the segment axis and the internuclear vector

between protons in the main chain. For polymers containing aliphatic protons in the main chain,

the coefficient a is close to 6.2 ± 0.7. The presence of the trithiocarbonates CS3 in the middle of

our polymer chains probable influences the value of this coefficient.

Figure 6.26. Evolution of T2a-relaxation time with temperature for networks based on telechelic

poly (butyl acrylate) with different amount of cross-linker

Using the number of backbone bonds in the statistical segment, C∞, the molar mass of network

chains between chemical network junctions, Mc is calculated:

[6.11]


121

2

2

2

2

(1:1)(1:1)(1:1.6)

(1.6 :1)

p

rlc

pc

rl

TM T

M TT

=

(1:1.6)(1:1)(1:1.6) (1:1)

teleffc

telc eff

EMM E

=

Where Mu is the molar mass per elementary chain unit for the polymer chain and n is the

number of backbone bonds in an elementary unit.

For our polymers, the values for the a and C∞ coefficients are not known. Instead of

determining the absolute masses between cross-linking points, analyzing the T2a-relaxations

curves, we can estimate a ratio for the masses between cross-links in the two systems.

[6.12]

As it can be seen in Figure 6.26, the T2a-relaxation time curve reaches at high temperatures a

plateau value for both systems. This evidences the presences of networks points in the systems,

limiting the motions of the polymer chains, and acknowledges the formation of a three-

dimensional network in both samples. The gap between the two plateau values at high and low

temperature is wider in the case of the coating prepared with a 1:1 [epoxy] : [COOH] ratio,

indication of a looser network.

Using the values obtained in Figure 6.26 an the relation [6.12], we obtain a value of 2.8 ± 10%

for the ratio between the distances between cross-links in the sample with the [epoxy] : [COOH]

ratio of 1:1 and in the one with the [epoxy] : [COOH] ratio of 1:1.6.

Using equation [6.8] it is easy to write that:

[6.13]

The values for the elastic moduli of these coatings presented of Figure 6.15 leads to a ratio of

2.4 ± 20 %. This value is, within the error of measurement, equal to the value observed with

solid-state NMR. T2-relaxation measurements on both coatings. Hence, these solid-state NMR

experiments confirm that the differences of elastic modulus values obtained for the cross-linked

telechelic polymers using these two ratios in the formulations presented on Figure 6.14 and 6.15

can be explained by a looser network, not taking into account the dangling ends. The topological

distance between two neighboring cross-linking points is approximately three times larger in the

sample with a 1:1 [epoxy] : [COOH] ratio.

Chapter 6

122

VII) Conclusion

In this chapter, the telechelic poly (butyl acrylate), whose synthesis was described in Chapter

4 were employed to form three-dimensional networks, via an epoxy-carboxylic acid reaction

between the acid functionalities present at the ends of the linear polymer chains and the epoxy

groups of the TGIC cross-linker, catalyzed by the tertiary amine functions of DABCO. In order

to determine the influence of the functionality distribution and the molecular weight distribution

of the polymer, some random poly (acrylic acid-co-butyl acrylate) were synthesized with the

number-average molecular weight and number-average functionality very similar to the ones of

the telechelic polymer.

The mechanism of this epoxy-carboxylic reaction was studied by ATR-FT-IR using telechelic

polymers for cross-linking experiments. A band at 1560 cm-1 was detected, and its growth and

disappearance over time were recorded. We attributed this band to the stretching vibration of a

carboxylate anion group present in an intermediate complex involving the COO- group derived

from the carboxylic acid functions present in the polymer, the epoxy group and DABCO. No

presence of this band was recorded when only two of the three components were mixed and

heated up to the curing temperature. A new mechanism to explain the cross-linking reaction was

proposed. More investigations were done to check the validity of the mechanism. The structure

of the intermediate was confirmed by studying the rate of disappearance of the acid dimers, the

epoxy band at 900 cm-1 and the rate of formation and disappearance of the intermediate at 1560

cm-1 at different cure temperature. Further information was obtained by replacing the polymer

by acetic acid, while keeping the same [epoxy] : [COOH] ratio. The same variations in the band

adsorptions of the epoxy and the intermediate were found in this experiment.

The random copolymers were found to cross-link similarly to the telechelic ones but at a

slower rate. Two main reasons were proposed to explain this phenomenon. First of all, the lack

of mobility of random copolymers chains compared to the telechelic ones, once the cross-

linking process has begun. Another possibility is that the smaller amount of COOH dimers

between two close COOH groups in the random copolymer might as well reduce the reaction

rate of the cross-linking process. The cross-linked coating made from telechelic and random

polymers always had a glass transition temperature well below room temperature and appeared

to be fully elastic.

The elastic moduli of coatings based on telechelic poly (butyl acrylate) and random poly

(acrylic acid-co-butyl acrylate) were estimated using micro-indentation measurements. The


123

evolutions of the E modulus versus the [epoxy] : [COOH] ratio used in the cross-linking

formulation were presented for both systems. It was found that with a similar [epoxy] : [COOH]

ratio in the starting formulations, the moduli of the coatings based on the telechelic polymer

were higher than the coatings prepared from random copolymers. We attributed this difference

to two different phenomena. The incompletion of the cross-linking reaction in the random

copolymer is a first possible explanation. The second reason is the difference in cross-link

density between both systems. The network based on random copolymer is probably less

“perfect” and is looser than the one based on telechelic polymers. TMDSC confirms these

results with evolutions of glass transition temperatures in line with the evolutions of the E

moduli. A linear relation between glass transition temperature and elastic modulus was found

for the telechelic polymers. This relation is not linear for random copolymers. We attributed this

difference to the presence of non-functional chains in the network.

Influence of the dangling chains on the E modulus of the coatings based on telechelic

poly(butyl acrylate) was demonstrated by micro-indentation, employing various amounts of

well-defined mono-functional chain in the curing of telechelic polymer. It has been shown that

the E modulus decreases with an increasing amount of dangling chains. This effect is more

severe when the first few percents of mono-functional chains are introduced.

For high [epoxy] : [COOH] ratios (above 2), a dramatic increase of the E modulus was

observed in the telechelic polymer-based coatings, those values being unrealistic for

polyacrylate-based networks. The reason behind it is the formation of hard clusters, whose

presence is justified by the polyetherification of the epoxy moieties of the TGIC. This side-

reaction is competitive to the epoxy-carboxylic acid one but its kinetics is probably much

slower. It occurs to a significant extent, only when the epoxy groups are in excess, as no phase

separation is observed for low [epoxy] : [COOH] ratios by solid-state NMR.

Finally, solid-state NMR was employed to investigate the homogeneity of the networks made

from telechelic polymers for the lower [epoxy] : [COOH] ratios. T1ρ-relaxation experiments

revealed the homogeneity of the formed network after the epoxy-carboxylic acid reaction on the

nanometer scale, ruling out the presence of TGIC-polymerized clusters in the final network. T2-

relaxation experiments showed that the coating with a 1:1 [epoxy] : [COOH] ratio has an

average molecular weight between cross-linking points roughly three times higher than the one

of the 1.6:1 [epoxy] : [COOH]. This ratio for molecular weight between cross-links was

compared to the one calculated from the rubber elasticity theory, using the E modulus values

determined by micro-indentation. A satisfactory agreement between the two methods of

calculation was found.

Chapter 6

124

Reference List

1. Soloukhin, V. A., Posthumus, W., Brokken-Zijp, J. C. M., Loos, J., de With, G.; Polymer 2002, 43, 6169.

2. Oliver, W. C., Pharr, G. M.; J. Mater. Res. 1992, 7, 1564.

3. Li, Z., Brokken-Zijp, J. C. M., de With, G.; Polymer 2004, 45, 5403.

4. Percy, E. J. GB 1120301

5. Bearden, C. R. US 3215731

6. FR 1485764

7. Takayama, Y. JP 44002685

8. Ito, H. JP 43026606

9. Caldwell, J. R. US 2484487

10. Simpson, A. J. GB 998394

11. Hamamoto, Y. JP 45017662

12. Kucharski, M., Lubczak, R.; J. Chem. Technol. Biot. 1998, 72, 117.

13. Shechter, L., Wynstra, J.; J. Ind. Eng. Chem. 1956, 48, 86.

14. Kakiuchi, H., Tanaka, Y.; J. Org. Chem. 1966, 31, 1559.

15. Madec, P. J., Marechal, E.; Makromol. Chem. 1983, 184, 323.

16. Matejka, L., Pokorny, S., Dusek, K.; Polym. Bull. 1982, 7, 123.

17. Matejka, L., Lovy, J., Pokorny, S., Bouchal, K., Dusek, K.; J. Polym. Sci. Pol. Chem. Ed. 1983, 21, 2873.

18. Fedoseev, M., Gurina, M., Sdobnov, V., Kondyurin, A.; J. Raman Spectrosc. 1996, 27, 413.

19. Dossin, L. M., Graessley, W. W.; Macromolecules 1979, 12, 123.

20. Pearson, D. S., Graessley, W. W.; Macromolecules 1980, 13, 1001.

21. Langley, N. R., Polmanteer, K. E.; Journal of Polymer Science, Polymer Physics Edition 1974, 12, 1023.


125

22. Dusek, K., Duskova-Smrckova, M., Lewin L.A., Huybrechts J., Barsotti R.J.; 27th FATIPEC Congress, Aix-en-Provence 2004, Congress Proceedings, Vol 1 2004, 209.

23. Porter, D.; Group interaction modeling of polymer properties. New York: 1995

24. Litvinov, V. M., Dias, A. A.; Macromolecules 2001, 34, 4051.

25. Litvinov, V. M.; Macromolecules 2001, 34, 8468.

26. Fry, C. G., Lind, A. C.; Macromolecules 1988, 21, 1292.

Chapter 6

126

127

CHAPTER 7

CONCLUSION AND RECOMMENDATIONS

I) Conclusion

The rapid development in the past few decades of ‘living’ controlled radical polymerization

allowed the possibility to synthesize well-defined polymers in a quite straightforward manner.

The molecular weights can easily be controlled, low polydispersities obtained and

functionalization of the polymer chain is not much of a problem anymore. One question arises:

is the synthesis of such well-defined structures only an academic scientific achievement or can it

leads to breakthrough findings in the field of material properties?

In order to determine the influence of the functionality type distribution and the molecular

weight distribution for polymers used as building block for the formation and properties of

poly(meth)acrylate networks, we have synthesized well-defined telechelic polyacrylates using

the RAFT polymerization technique, and random copolymer randomly introducing

functionalities in polyacrylate chains.

In this thesis, a chapter of introduction was written, in which an update on controlled radical

polymerization (with an emphasis on Nitroxide Mediated Polymerization, Atom Transfer

Radical Polymerization and Reversible Addition Fragmentation chain Transfer technique) and a

literature review on available methods to obtain telechelic polymers were given.

The first step towards the synthesis of well-defined networks concerned the synthesis of linear

end-functionalized polymers with a controlled molecular weight and a low polydispersity. The

RAFT polymerization technique was employed to obtain hydroxy end-functionalized poly

(methyl methacrylate). A two-step post-polymerization procedure was proposed in order to

modify the thiocarbonyl thio end-group, transforming it successively into a thiol end-group (via

aminolysis) and a hydroxyl end-function (via a Michael addition). Liquid-chromatography

separations proved the successful synthesis of telechelic PMMA to a significant extent.

However, attempts to quantitatively synthesize telechelic PMMA proved to be unsuccessful. We

128

demonstrated that formation of disulfide functions via oxidative coupling during the aminolysis

and hydrogen-transfer to the solvent were the main reasons lowering the final yield of α,ω-

poly(methyl methacrylate). MALDI-TOF-MS and two-dimensional LC-SEC separations

confirm these findings.

The selection of butyl acrylate as a monomer and trithiocarbonates as RAFT agents resulted in

the straightforward synthesis of well-defined carboxylic acid-functional poly (butyl acrylate).

Liquid-chromatography separations proved the nearly quantitative production of mono- and di-

functional poly(butyl acrylate) chains, those compounds being suitable for network formation

studies.

The end-group analysis by MALDI led to interesting observations on fragmentation

phenomena during the analysis of thiocarbonyl thio-containing poly(meth)acrylate. The

importance of the monomer was proven, as well as the effect of the stabilizing group of the

RAFT agent in the fragmentation probability. The effect of the laser intensity on the

fragmentation was investigated as well.

In the final part of this thesis, the mechanism for the reaction of the carboxylic acid groups of

the polymers with a tri-functional epoxy-crosslinker is described. The cure reaction was studied

with FT-IR and a new mechanism is proposed to explain the experimental data. The network

structure and several properties of the cured networks were then determined by micro-

indentation, DSC and solid-state NMR. Comparison between networks based on the telechelic

polymers that have a narrow functionality distribution and a narrow molecular weight

distribution, and random copolymers with a broad functionality distribution and a broad

molecular weight distribution was made. It was found that the presence of dangling ends has an

effect on the properties. Even only a few percent of dangling chains in a network structure

lowers the elastic moduli of the cured coatings. However, the main factor influencing the elastic

modulus and the glass transition parameters seems to be the molecular weight between cross-

links. A linear relationship between those two parameters was found for telechelic polymers.

The elastic modulus values for coatings based on telechelic polymers and prepared with high

[epoxy]: [COOH] values (above 2:1) suggest the presence of local domains of high elastic

moduli, formed by the autopolymerization of TGIC.

129

II) Recommendations

The production of telechelic poly (butyl acrylate) was relatively easy once trithiocarbonates

were used as RAFT agents. The paper describing the synthesis of those trithiocarbonates and

their successful application as controlling agents for the polymerization of butyl acrylate was

published one and a half year after the start of this project. Chapter 3 relates the struggles that

were encountered trying to use thiocarbonyl thio compounds as RAFT agents in order to obtain

telechelic polymers. Contrary to what was claimed at this time in the literature, the aminolysis

reaction is far from being quantitative. We evidenced at least two side-reactions occurring at the

same time as the reduction of thiocarbonyl thio functions into thiols: formations of disulfide

bridges, and hydrogen-abstraction from the solvent. If end-group functionalities are of primary

importance for polymer chains, we advice to avoid any manipulation on the polymer after the

RAFT polymerization. The use of di-functional RAFT agent, with functionality on both the

leaving group and the activating group is encouraged in order to obtain telechelic polymers after

one step. The trithiocarbonates are a specific class of such compounds, as the activating and

leaving group are degenerated.

Numerous claims in the literature concerning functionalities of telechelic polymers suffer

from a general lack of quantitative analytical work. The structures are often confirmed by mass

spectrometry analysis, without the required quantitative aspect of the characterization. We have

demonstrated in Chapter 3 and 4 that the combination of synthetic and analytical works are

required to seriously confirm the synthesis of telechelic chains. In order to complete the

analytical work, more work will be performed on the two-dimensional LC-SEC separation of

the telechelic poly (butyl acrylate) in order to quantify the amounts of the different populations

detected (Chapter 4).

The time spent on the synthesis and characterization of telechelic chains was lacking at the

end of this project, when the comparison between telechelic polymers and random copolymers

as building blocks for networks was made. Several points could not to be investigated due to

time constraints:

- Influence of the molecular weight on the E modulus: following the synthetic path

presented in Chapter 4, telechelic polymers with various molecular weights could easily

be synthesized. The determination of the elastic modulus by micro-indentation could bring

information of the influence of the cross-link density on the mechanical properties.

130

Comparison of experimental values with the ones obtained by various theoretical models

could then be made.

- Nature of the cross-linking agent: after synthesizing carboxylic acid functional polymers,

the choice was made to use the epoxy chemistry in order to obtain tri-dimensional

networks. The TGIC cross-linker was up to our knowledge the tri-functional epoxy

compound with the most documentation in the literature. Unfortunately the analysis

performed on it revealed the presence of impurities, making the investigation on the cross-

linking itself and on the properties of the final coatings impossible to be quantitative. The

amount of epoxy group can be determined, but the functionality distribution was

unknown. Two choices in order to improve this study can be proposed: either the TGIC

can be purified by preparative HPLC for example, or the epoxy – COOH chemistry has to

be replaced by another one. That can imply a change of cross-linking agent, or the

modification of the COOH-functionality of the polymer end-groups. As we wrote it in the

previous paragraph, this modification should preferentially be performed on the RAFT

agent itself before the polymerization, rather than after it.

- Mechanism of the carboxylic acid – epoxy reaction, catalyzed with tertiary amines: the

mechanism we have proposed in chapter 6 is probably over-simplified, although it should

be sufficient to describe the cross-linking reaction with an acceptable agreement with

experimental data. Reactions at fixed [epoxy] : [COOH] ratios and variable DABCO

amount should be performed in order to have a look at the catalyst concentration

influence. In the same way, curing formulations with a fixed amount of DABCO and

different [epoxy] : [COOH] ratios should be prepared. These experiments should validate

the mechanism we have proposed and possibly could gain access to the kinetic constants.

- Solid-state NMR: networks with formulations containing high amounts of TGIC should

be looked at with T1-and T1ρ-relaxations techniques, in order to prove the presence of hard

clusters in the networks, formed by the autopolymerization of the TGIC. We proposed this

theory in order to explain the amazingly high values obtained for the elastic modulus of

coatings based on telechelic poly (butyl acrylate) when high TGIC contents were used in

the preparation.

Summary

The research described in this thesis is a part of project #205 of the DPI (Dutch Polymer

Institute). One of the challenges was to synthesize telechelic poly(meth)acrylayes with well-

defined masses and low polydispersities, in order to evaluate the potential benefit that can be

obtain in the coating formulations by using such polymers. Complete characterization of the

synthesized polymers was done, via several analytical techniques. The second part of the project

concerned the comparison between the newly-designed polymers and the more classical random

copolymers, both types of polymers being used as building blocks for three-dimensional

networks.

The RAFT (Reversible Addition-Fragmentation chain Transfer) polymerization technique has

been chosen for the production of polymer chains with predicable molar masses and low

polydispersities. A hydroxyl-functional thiocarbonyl thio RAFT agent was initially used. In

order to obtain an hydoxy functionality at both ends of a linear polymethacylate, a new two-step

post-polymerization procedure, involving an aminolysis of the thiocarbonyl thio group of the

RAFT polymer, followed by a Michael addition on the resulting thiol was proposed. Contrary to

what was claimed in the literature at the time the research was conducted, the aminolysis

reaction was found to be prone to side-reactions that lowered its yield. We evidenced transfer to

solvent and oxidative coupling of thiols as side reactions by using MALDI-TOF-MS and liquid

chromatography under critical conditions as analytical techniques. A polymer batch containing

66% of telechelic OH-functional chains was produced, following the proposed method, but

attempts to increase this percentage were unsuccessful.

Using symmetrical COOH-functional trithiocarbonates as RAFT agents, it was possible to

synthesize telechelic poly(butyl acrylate) in one step. The molar masses of these polymers were

predictable and low polydispersity indexes were obtained. MALDI-TOF-MS and liquid

chromatography under critical conditions confirmed the structure of the polymer, and

quantitative results proved that some batches of poly(butyl acrylate) synthesized via this route

contain more than 99% of bi-functional chains. Considering the high purity of those polymers in

terms of functionality, those polymers were chosen as building blocks for the network formation

and networks properties studies.

The MALDI-TOF-MS analysis of the RAFT polymers pointed out the tendency for RAFT

polymers to fragment under laser irradiation. It was found that three main parameters influence

the fragmentation probability of a RAFT polymer chain during a MALDI analysis:

- the laser intensity

- the type of RAFT agent used during the synthesis (thiocarbonyl thio-containing

compounds are more prone to fragment than the trithiocarbonate-containing ones)

- the chemical nature of the backbone (polymethacrylates fragment more easily than

polyacrylates)

Finally, a comparison between telechelic polymers and random copolymers as building blocks

for networks was made. The chemistry used for the cross-linking reaction was the carboxylic

acid – epoxy reaction, catalyzed by tertiary amine. A mechanism was proposed to describe the

reaction. FT-IR measurements prove the existence of a intermediate involving the 3 components

(COOH of the polymer – cross-linker – catalyst) during the reaction. A characteristic band at

1560 cm-1 was attributed to the C=O band of the COO- group of an intermediate in the proposed

mechanism. The telechelic polymers reacted faster than their random equivalent.

The elastic moduli of the networks were determined using micro-indentation technique.

Contrary to what could be expected from a theoretical model, the telechelic-based coatings

exhibited higher elastic modulus values than the random co-polymers-based ones. A linear

relation between the elastic modulus and the glass transition temperature of the networks was

established for networks based on telechelic polymers.

Solid-state NMR T1ρ−relaxation experiments showed that at low [epoxy} : [COOH] ratio in

the formulation of telechelic chains-based polymers, the network obtained after cross-linking

was homogeneous at nanometer scale. T2−relaxation experiments confirmed the conclusions

obtained from micro-indentation, showing that the molar mass between cross-kinks is the major

parameter influencing the elastic modulus, in the case of telechelic chains-based networks.

Samenvatting

Het onderzoek, uitgevoerd in het kader van dit proefschrift, was een onderdeel van project 205

van het Dutch Polymer Institute. Een van de uitdagingen van dit onderzoek was de synthese van

lineaire (meth)acrylaat polymeren met een goedgekarakteriseerde eindstandige reactieve

groepen (functionaliteit = 2), een voorspelbaar moleculairgewicht en een nauwe

moleculairgewichtsdistributie. Een tweede uitdaging van het onderzoek was om deze polymeren

te gebruiken als uitgangsmateriaal voor het maken van doorgeharde coatings en een aantal

eigenschappen van deze coatings te bepalen. Een derde uitdaging van dit onderzoek was om

deze coatings te vergelijken met coatings die gemaakt waren van meer klassieke polymeren met

een zeer vergelijkbaar gemiddelde moleculair gewicht en reactieve groep functionaliteit maar

met een veel bredere moleculair gewicht distributie en een veel bredere en hangende

functionaliteit van de reactieve groepen over de ketens. Het doel van dit deel was om na te gaan

in hoeverre het gebruik van de hierboven beschreven eindstandig functionele polymeren

aantrekkelijker eigenschappen aan de doorgeharde coatings geven door de aanwezigheid van

een ander driedimensioneel netwerkstructuur.

Voor het synthetiseren van polymeer ketens met een voorspelbare moleculair gewicht en een

smalle moleculair gewichtsdistributie werd gebruik gemaakt van de RAFT (Reversible

Addition-Fragmentation chain Transfer) polymerisatiemethode. Initieel werd in dit onderzoek

een hydroxy–gefunctionaliseerde dithioester RAFT agens gebruikt. Om een hydroxy-

functionaliteit te verkrijgen aan beide uiteinden van een lineair polymethacrylaat, werd er een

nieuwe twee-staps synthese toegepast na de polymerisatie reactie. De eerste stap was een

aminolyse van de dithioester groep van het RAFT polymeer. In de tweede stap werd de

verkregen eindstandige thiol groep omgezet in een HO- groep (Michael addition). In

tegenstelling tot wat werd beweerd in de literatuur op het moment dat dit onderzoek werd

uitgevoerd traden er nevenreacties op tijdens de aminolyse, die de uiteindelijke de opbrengst

verlaagde. Het verkregen polymeer product bestond slechts voor 66% uit polymeren met aan

beide uiteinden een hydroxy groep. Door gebruik te maken van MALDI-TOF-MS en vloeistof

chromatografie onder kritische condities werd aangetoond dat oxidatieve koppeling van de thiol

groepen en de reactie met het oplosmiddel verantwoordelijk waren voor deze nevenreacties.

Pogingen om een zuiverder product te krijgen mislukten.

Door gebruik te maken van een symmetrisch trithiocarbonaat RAFT agens met aan beide

zijden een eindstandige COOH-functionaliteit werden er in een stap poly(butyl)acrylaat

polymeren gemaakt met aan beide zijden een eindstandige COOH groep. De verkregen

moleculairgewichten van deze polymeren kwamen overeen met de berekende waarden.

Bovendien was de variatie in moleculairgewicht over de verschillende ketens zeer laag.

MALDI-TOF-MS en vloeistof chromatografie onder kritische condities bevestigden de structuur

van deze polymeren en kwantitatieve resultaten toonden aan dat sommige poly(butyl)acrylaten,

die langs deze weg verkregen waren, voor meer dan 99% bestonden uit polymeren met aan

beiden zijden een -COOH groep. Door deze hoge zuiverheid in functionaliteit bleken deze

polymeren geschikte uitgangsstoffen te zijn voor het maken van goed gedefinieerde

driedimensionele moleculaire netwerken en voor het bestuderen van de eigenschappen van de,

na doorharding, uit deze polymeren verkregen coatings.

Uit onze MALDI-TOF-MS analyse bleek dat de RAFT polymeren geneigd zijn te

fragmenteren onder invloed van het laserlicht dat gebruikt wordt tijdens de analyse. . De

volgende drie parameters beïnvloeden de waarschijnlijkheid van fragmentatie tijdens de analyse:

- de licht intensiteit

- het type RAFT agens dat gebruikt werd gedurende synthese van deze polymeren

(polymeren verkregen met behulp van dithiocarbonyl agentia zijn meer gevoelig voor

fragmentatie dan trithiocarbonyl verbindingen).

- De chemische samenstelling van de polymeer keten (polymethacrylaten fragmenteren

makkelijker dan polyacrylaten)

Tenslotte werden de goed gekarakteriseerde en gedefinieerde lineaire poly(butyl)acrylaten met

COOH eindgroepen die hierboven beschreven zijn doorgehard met behulp van een trifunctionele

epoxydoorharder onder invloed van een tertiaire aminekatalysator. Zowel de

doorhardingsreactie als enkele materiaaleigenschappen van de verkregen coatings werden

bestudeerd. De verkregen resultaten werden vergeleken met de resultaten die verkregen werden

door in plaats van deze polymeren lineaire polybutyl copolymeren te gebruiken met eenzelfde

gemiddelde functionaliteit en moleculair gewicht maar met een veel bredere

moleculairgewichtsdistributie en een veel bredere en hangende functionaliteit van de COOH

groepen over de polymeerketens. Deze laatste polymeren werden gesynthetiseerd door gebruik

te maken van standaard radicaal polymerisatietechnieken.

Voor de doorhardingsreactie van beide typen van polymeren werd eenzelfde nieuw

mechanisme voorgesteld. Metingen met behulp van FT-IR toonden het bestaan van een

intermediair complex aan tussen de eindgroep van het polymeer de doorharder en de katalysator.

Een karakteristieke absorptie bij 1560 cm-1 werd waargenomen voor dit intermediair. Deze band

werd toegewezen aan de COO- groep van het polymeer. De snelheid van de doorhardingsreactie

was langzamer voor de polymeren met de hangende COOH functionaliteit.

De elasticiteitsmoduli van de coatings gemaakt uit beide typen hierboven besproken

polymeren werden bepaald met behulp van microindentatie metingen. In tegenstelling tot wat

verwacht kon worden op basis van de theorie bleken de elasticiteitsmoduli lager uit te vallen

voor de coatings gemaakt uit polymeren met hangende functionaliteit. Alleen voor de coatings

gemaakt uit de polymeren met eindstandige COOH functionaliteit werd er een lineaire relatie

gevonden tussen de E-moduli en de Tg, zelfs als het percentage loseindige groepen in het

netwerk groot was.

Vaste stof NMR T1ρ relaxatie experimenten lieten zien dat de netwerkstructuur na

doorharding van de eindstandige COOH polymeren homogeen is op nanometer schaal wanneer

de [COOH]/[epoxy] gelijk is aan een. De vaste stof NMR T2 relaxatiemetingen bevestigden de

conclusies van de microindentatie metingen namelijk dat het moleculairgewicht tussen de

crosslinks de belangrijkste parameter is voor de bepaling van de grootte van de E-modulus van

de coatings gemaakt uit eindstandige COOH polymeren

Acknowledgments

Completing a Ph.D. thesis on its own is a utopia. During the four years spent in Eindhoven, I

was lucky to have great people on my side always friendly and ready to help me.

I deeply thank Prof. Rob van der Linde, my first promoter, for the opportunity he gave me to

start a Ph.D. research in his group. Rob, despite your retirement in 2002, you have always

followed closely my struggles and progresses during these years. Your energy and positive

attitude was always of great help when things seemed to go badly. I am extremely proud to have

you as a promoter.

My second promoter, Prof. Peter Schoenmakers is acknowledged for the great help he

provided. Peter, I have very much appreciated my frequent meetings with you in Eindhoven or

in Amsterdam. There were always a lot a fun and a lot of science. Your extreme attention to

every word in scientific discussions in particular made me more precise/accurate about my

speaking and my writing. Thanks a lot for everything.

I would like to thank my co-promoter Dr. José Brokken for the daily supervision she provided

me. José, your endless efforts to improve the quality of my various manuscripts resulted in

better papers and a better thesis. We confronted on numerous occasions our views on different

aspects of my research. The long discussions often gave me new ideas and extra motivation. I

have always had the feeling that if I could convince you, I could convince anyone.

I would like to thank all the members of my examination committee for their presence and

their valuable comments during the revision of this thesis.

The Dutch Polymer Institute is thanked for the financial support that was provided for this

project. I would like especially to thank the industrial contact people whose remarks were of

great help throughout this project.

During those 4 years, I met a lot of people who help me with analytical techniques that

seemed a little bit difficult to me at first sight. Xulin Jiang was “the master of Liquid

Chromatography” for me. All my synthetic work would have been useless without a proper

characterization. Xulin, thanks to your great work, we were able to show that the goals assigned

to both our projects were reached. It has been a pleasure to collaborate with you. After Xulin’s

project was over, Aschwin van der Horst was of great help for the characterization of COOH-

functional chains. Thank you Aschwin !

Wieb Kingma and Marion van Straten are thanked for their help with SEC and MALDI-TOF-

MS. Bastiaan Staal is also thanked for the amazing amount of time he spent helping me. Bas,

without your help, Chapter 5 would never have been as it stands now.

Chapter 6 essentially exists due to the help of five people. Without proper help, I would never

have been able to write it. Otto van Asselen, I am greatly grateful for all the time you have tried

to help me determining a proper mechanism for my cross-linking reaction, using FT-IR. Zhili

Li, your expertise in micro-indentation experiment combined with an extreme rapidity

producing results was a true relief when time was winding down. Prof. Guenther Hoehne is

thanked for his precious help with TMDSC measurements. Dr. Pieter Magusin and Brahim

Mezari, thanks a lot for the solid-state NMR experiments. That was a short but fruitful and

enjoyable collaboration.

Many thanks to Dr. Mariëlle Wouters for all the rheology experiments performed at TNO.

Unfortunately, none of them could appear in this booklet, this is how science is sometimes.

I would also like to thank Patrick Terregino, whose help during three months resulted in the

finishing of the RAFT polymer synthesis. The VL103 sample you have synthesized Pat is the

central part of the network formation study. Thanks a lot!

The STO 1.28 office was a place of fun and scientific arguments through the years. Thanks to

Auke, Huiqi and Willem-Jan for having contributed to the nice atmosphere during working

hours. The SPC group has always considered me as one of them despite the groups shuffling

that happened at the beginning of my stay at the TU/e. I would like to thank particularly Wouter,

Rajan, Jelena, Jens, Raf, Ellen, Maarten and Robin for their friendship and good moments. Dr.

Bert Klumperman is also acknowledged for the guidance he provided me in the RAFT world.

The old SVM gang will always have a special place for me. Marshall, Willem, Okan, Chouaïb

Gabriëlle, Jos, Victor and Robert, it was a lot of fun during my stay in Eindhoven, and I hope we

will keep in touch in the future.

Despite living outside France, I have never felt lonely. Merci à tous mes amis français qui ont

toujours répondu présent quand le besoin s’en faisait sentir. Julien, Laurent(s), Francis,

Michael(s), Emmanuel(s), Ronan, Stéphane(s), Gérald, Pierre et Alexandre, merci pour votre

amitié.

Merci également à ma famille et à ma belle-famille pour tout le soutien nécessaire à

l’accomplissement d’une thèse à l’étranger. Merci tout particulièrement à mes parents qui m’ont

toujours donné la possibilité de suivre la voie que j’avais choisie. Sans l’environnement dans

lequel j’ai été élevé, mon cursus n’aurait pas pu être ce qu’il a été.

Et au final, je remercie grandement Delphine pour tout ce qu’elle m’a apporté au quotidien

pendant ces 4 ans. Vivre ensemble et travailler ensemble n’est pas toujours facile, mais ton

soutien était indispensable pour la réussite de cette thèse. Le meilleur reste à venir…

Curriculum Vitae

Vincent Georges Robert Lima was born on the 9th of June 1977 in Noisy-le-sec, France. He

obtained in 1994 his high school diploma (Baccalauréat) in Lycée Gustave Monod in Enghien-

les-bains (France). He then graduated in 1999 from the Institut de Science et Technologie –

Chimie des Matériaux (Université Pierre et Marie Curie – Paris, France) after completing an

graduation project in the North-American research center of ATOFINA in King of Prussia (PA,

USA), working on “Synthesis of core-shell poly(acrylate)s impact modifiers in emulsion, using

the ATRP technique”. After fulfilling his military obligations, he started in 2001 a Ph.D. project

under the guidance of Prof. Rob van der Linde at the Technological University of Eindhoven

(The Netherlands). This thesis is the results of 4 years of research. Since March 2005, he has

been working in the research department of Océ Technologie in Venlo (The Netherlands).

Tailor-made poly(meth)acrylate-based networks ... · Tailor-made poly(meth)acrylate-based networks - Preparation, Structure and Properties - Vincent Lima

Documents