Poly(urethane-isocyanurate) NetworksSynthesis, Properties and Applications

Piet Driest

Poly(urethane-isocyanurate) Netw

orks Piet D

riest

POLY(URETHANE-ISOCYANURATE)

NETWORKS

Synthesis, properties and applications

Pieter Job Driest

POLY(URETHANE-ISOCYANURATE)

NETWORKS

Synthesis, properties and applications

DISSERTATION

to obtain

the degree of doctor at the Universiteit Twente,

on the authority of the rector magnificus,

Prof. dr. T.T.M. Palstra,

on account of the decision of the graduation committee

to be publicly defended

on Friday 31 January 2020 at 12.45 hr.

by

Pieter Job Driest

born on the 6th of September, 1987

in Amsterdam, the Netherlands

This dissertation has been approved by:

Supervisors

Prof. dr. D.W. Grijpma

Prof. dr. D. Stamatialis

Cover design: Douwe Oppewal Printed by: Ipskamp Printing Lay-out: Piet Driest

ISBN: 978-90-365-4934-9

DOI: 10.3990/1.9789036549349

© 2020 Pieter Job Driest, The Netherlands. All rights reserved. No parts of

this thesis may be reproduced, stored in a retrieval system or transmitted in

any form or by any means without permission of the author. Alle rechten

voorbehouden. Niets uit deze uitgave mag worden vermenigvuldigd, in

enige vorm of op enige wijze, zonder voorafgaande schriftelijke

toestemming van de auteur.

Graduation Committee:

Chairman / Secretary Prof. dr. J.L. Herek

Supervisors: Prof. dr. D.W. Grijpma

Prof. dr. D. Stamatialis

Committee Members: Dr. D.J. Dijkstra

Prof. dr. G.J. Vancso

Prof. dr. P.J. Dijkstra

Prof. dr. K.U. Loos

Prof. dr. L.H. Koole

Imagination is not only the uniquely human capacity to envision that which

is not, and therefore the fount of all invention and innovation. In its

arguably most transformative and revelatory capacity, it is the power that

enables us to empathise with humans whose experiences we have never

shared.

J.K. Rowling

Harvard Commencement Speech │ June 5, 2008

Table of Contents

Chapter 1 General Introduction: Scope, Aims and Outline

of this Thesis

10

Chapter 2 Poly(urethane-isocyanurate)s: Chemistry,

Network Formation and Applications

18

Chapter 3 Aliphatic Isocyanurates and Polyisocyanurate

Networks

50

Chapter 4 The Influence of Molecular Interactions on the

Viscosity of Aliphatic Polyisocyanate Precursors

for Polyurethane Coatings

70

Chapter 5 The Trimerization of Isocyanate-functionalized

Prepolymers: an Effective Method for

Synthesizing Well-defined Polymer Networks

92

Chapter 6 Combinatorial Networks by Trimerization of

Mixtures of NCO-functionalized Prepolymers

110

Chapter 7 Structure-property Relations in Semi-crystalline

Combinatorial Poly(urethane-isocyanurate)

Hydrogels

138

Chapter 8 Poly(ethylene glycol) Poly(urethane-

isocyanurate) Type Hydrogels as a New Class of

Materials for Contact Lens Applications

154

Summary and Outlook 184

Acknowledgements | Dankwoord | Danksagung 194

List of Publications 204

Chapter 1 - General Introduction: Scope,

Aims and Outline of this Thesis

12

1.1 General Introduction

2019 has been pronounced the International Year of the Periodic Table, a

milestone year in chemistry in which we celebrate the 150th anniversary of

Dimitri Mendeleev’s discovery of the periodic table of the elements.1 Not

only did his ordering of the elements known at the time based on their

atomic weights and general properties lead him to (correctly!) predict which

ones were missing, it also paved the way for the creation of an international,

unified chemical language. This concept was formalized 50 years later when

the International Union of Pure and Applied Chemistry (IUPAC, celebrating

its 100th birthday in 2019) came to life. “A neutral and objective scientific

organization, IUPAC was established in 1919 by academic and industrial

chemists who shared a common goal – to unite a fragmented, global

chemistry community for the advancement of the chemical sciences via

collaboration and the free exchange of scientific information. Throughout its

long history, IUPAC has fulfilled that goal through the creation of a common

language and the standardization of processes and procedures.”2

Today, 150 years after Mendeleev first published his ordering of the

elements and 100 years after chemists worldwide united behind a shared

cause, this mission is as relevant as ever. The challenges in the field of

chemistry have obviously changed over time and presently include subjects

like climate change, the depletion of the earth’s natural reserves and the

upcoming of new scientific fields for which the ethical guidelines have not

been written yet.3 Relevant chemical topics are explicitly taken up in the

United Nations’ Sustainable Development Goals for 2030 and include the

responsible use of resources and the sustainable development and

production of materials (nrs. 9 and 12).4 Clearly, the present-day global

chemistry community formally created in 1919 will play an important role in

dealing with these issues. But where the challenges have changed over time,

the solution has not. As before, international collaboration, a free exchange

of scientific information and a common language are the key to realizing our

ambitions.

13

The Marie Curie ITN project “TheLink” is an EU-funded international,

interdisciplinary research project which aims to contribute to this ideal. As

part of this project, the work presented in this thesis was performed in the

labs of Covestro Deutschland AG (formerly Bayer MaterialScience) and was

scientifically coordinated by the University of Twente. With this approach,

we mean to illustrate the possibilities for synergetic collaboration between

industrial and academic research partners. Facilitating the free exchange of

scientific information, all research output presented in this thesis has been

(or will be) published Open Access. Finally, by bringing together researchers

with different nationalities, native tongues and cultural backgrounds, we

mean to contribute to the (further) development of an international,

common language in chemistry and material science.

1.2 Scope, Aims and Outline of this Thesis

Scope and Aims

Throughout this thesis, a recurring aim is to generate a link between

conventionally separated research communities. This may concern a

separation in the research objective (industrial / academic), the research

approach (computational / experimental) or the research discipline

(chemistry / physics / biology). We aim to establish this by performing

fundamental, scientific research based on industrially relevant chemical

processes and materials. Also, different scientific approaches and/or

disciplines are combined in each of the studies presented in this thesis.

The main subject of this work is the investigation and development

of poly(urethane-isocyanurate) (PUI) polymer networks for biomedical

applications. In the first part of this thesis, the viscosity of liquid aliphatic

isocyanurates is investigated. These compounds are important industrial

precursors in the conventional 2-component synthesis of PUI networks.5 By

isolating series of isocyanate-functional and non-functional model

compounds in high purity, the molecular interactions that dictate the

viscosity of these compounds are investigated. The experimental results are

complemented by a detailed computational study on the same series of

14

model compounds. Using this combined experimental and theoretical

approach, we aim to provide a first framework of the molecular interactions

that dictate the viscosity of liquid aliphatic isocyanurates.

In the second part of this thesis, focus is gradually shifted from raw

materials towards products. Network formation and the molecular network

structure of PUI type polymer networks are investigated. We develop a new

synthesis method for PUI type polymer networks, aiming to address some

of the drawbacks that are typically associated with the conventional 2-

component synthesis of these materials. Also here, we aim to provide

fundamental scientific understanding of industrially relevant materials and

make a correlation between the molecular structure of the PUI networks

and the physical and biological properties observed macroscopically.

In the third part of this thesis, we explore the possibilities and limits

of the new synthesis method in terms of resulting material properties and

potential applications. Specifically, new PUI type networks are synthesized,

whereby we target useful (combinations of) material properties for

biomedical applications. Relevant properties on which we report include

water uptake, crystallinity, elastic modulus, tensile strength and toughness.

These properties are investigated both in the dry state and in the water-

swollen state. Continuously, we aim to relate the observed macroscopic

material properties to the molecular polymer network structure. Finally, the

potential application of several new PUI networks as contact lens materials

is explored.

Outline

A general introduction to isocyanate- and polyurethane chemistry is

provided in Chapter 2. Emphasis is put on aliphatic poly(urethane-

isocyanurate) (PUI) networks, including their synthesis, general material

properties and typical applications. Conventional and newly developed

synthesis strategies are reviewed. The trimerization reaction of isocyanates

is shown to play a key role throughout these processes.

In Chapters 3 and 4, the viscosity of liquid isocyanate-functional

aliphatic isocyanurates is investigated. These compounds are important

precursors in the conventional 2-component synthesis of PUI networks.

15

Using a combined computational and experimental approach, model

compounds are used to investigate which molecular interactions play a role

in the observed macroscopic trends.

In Chapter 5, a new synthesis method is presented for synthesizing

well-defined PUI networks based on the trimerization of isocyanate-

functionalized prepolymers. The molecular network structure is investigated

and compared to PUI polymer networks synthesized by conventional

synthesis methods.

In Chapter 6, combinatorial PUI networks, synthesized by the

trimerization of mixtures of isocyanate-functionalized prepolymers in

solution, are investigated aiming to achieve unique material properties.6 Of

specific interest is the polymerization of mixtures of hydrophilic and

hydrophobic network components into the same combinatorial network.7

The mechanical and thermal properties of the resulting networks are

investigated both in the dry state and in the water-swollen state. Specific

relevant material properties for biomedical applications, such as toughness

in the hydrated state, are assessed.

In Chapter 7, a structure-property-relations study is presented for

the combinatorial PUI hydrogel materials. By varying the crystallizable

hydrophobic content of the combinatorial PUI networks, the degree of

crystallinity of the hydrogel in the hydrated state is correlated to the

toughness of the hydrogel in the hydrated state.

As poly(ethylene glycol) (PEG) is a hydrophilic and low-biofouling

polymer, PUI networks based on PEG are expected to be suitable materials

for contact lens applications.8 In Chapter 8, a series of these PEG PUI type

hydrogels is synthesized and characterized. Material properties relevant for

contact lens applications are investigated, such as transparency, elastic

modulus and toughness in the water-swollen state, refractive index,

biocompatibility and protein adsorption.

16

References

1. Mendelejeff, D. Über die Beziehungen der Eigenshaften zu den Atomgewichten der Elemente. Zeitschrift für Chemie 12, 405–406 (1869).

2. IUPAC. The International Union of Pure & Applied Chemistry. Available at: https://iupac.org/who-we-are/. (Accessed: 3rd August 2019)

3. United Nations Environment. Frontiers 2018/19: Emerging Issues of Environmental Concern. (2019).

4. Transforming our world : the 2030 Agenda for Sustainable Development. (United Nations General Assembly, 2015).

5. Avtomonov, E., Danielmeier, K., Driest, P. J. & Al., E. Chapters 3 and 4: Chemical Principles and Coating Technology Principles. in Polyurethanes − Coatings, Adhesives, Sealants (eds. Meier-Westhues, U., Danielmeier, K., Kruppa, P. & Squiller, E. P.) (Vincentz-Network Verlag, 2019).

6. Zant, E., Bosman, M. J. & Grijpma, D. W. Combinatorial synthesis of photo-crosslinked biodegradable networks. J. Appl. Biomater. Funct. Mater. 10, 197–202 (2012).

7. Zant, E. & Grijpma, D. W. Tough biodegradable mixed-macromer networks and hydrogels by photo-crosslinking in solution. Acta Biomater. 31, 80–88 (2016).

8. Musgrave, C. S. A. & Fang, F. Contact Lens Materials: A Materials Science Perspective. Materials (Basel). 12, 261 (2019).

17

Chapter 2 - Poly(urethane-

isocyanurate)s: Chemistry, Network

Formation and Applications

Adapted from:

E. Avtomonov1, K. Danielmeier1, P. Driest1,2 et al.; Chapters 3 and 4 in:

Polyurethanes – Coatings, Adhesives and Sealants (Editors: U. Meier-

Westhues et al.); Vincentz-network Verlag; Hannover, Germany, 2019

1Covestro Deutschland AG, CAS-Global R&D, 51373 Leverkusen, Germany

2Technical Medical Centre, Faculty of Science and Technology, Department

of Biomaterials Science and Technology, University of Twente, P.O. Box 217,

7500 AE Enschede, The Netherlands

20

2.1 Introduction

“A world without polyurethanes is hard to imagine these days. Polyurethane

chemistry has become an established technology worldwide in various

industrial applications. For innovative coatings, adhesives and sealants, it

plays a significant role and has been in many cases the key to new technology

implementation. Polyurethanes enable innovation and technological

advancement and help to develop products that meet market needs,

devising efficient and environmentally friendly manufacturing processes and

asserting a position in the global competitive environment.”

From the foreword of Polyurethanes – Coatings, Adhesives and Sealants (U.

Meier-Westhues et al.)1

2.2 Isocyanate- and Polyurethane Chemistry

Synthesis, structure and reactivity of the isocyanate group

As one of the oldest known functional groups in organic chemistry, the

isocyanate functional group has been studied extensively.2–4 Many different

synthesis routes have been described over the years for obtaining

isocyanates, as well as reactions involving them. Conventionally, the

industrial production of isocyanates is performed via the phosgenation of

the corresponding amines. The phosgenation reaction is sketched in Scheme

2.1. Although alternative production pathways of isocyanates are

increasingly investigated, these are generally limited to use in the academic

field. The most commonly adopted alternative synthesis pathways (i.e. not

involving phosgene) are by the Curtius, Hofmann and Lossen

rearrangements, which all involve carbonyl nitrene intermediates.5

21

Scheme 2.1. Synthesis of isocyanates by phosgenation.

Although the isocyanate functional group has been known for over

170 years, still new details are being discovered on a regular basis regarding

its chemical structure and reactivity. With two adjacent double bonds

connecting a nitrogen-, carbon- and oxygen atom, the isocyanate group falls

in the class of heterocumulenes. Resulting from the high electronegativity

of the adjacent heteroatoms, illustrated by the resonance structures shown

in Scheme 2.2, the central carbon atom possesses a relatively large positive

charge. Consequently, this carbon atom is a strong electrophile. In fact, the

high electrophilicity of the central carbon atom is the main rationalization

that isocyanates are employed so widely.

Scheme 2.2. Resonance structures of the isocyanate functional group.

Adapted from Delebecq et al.6

Surprisingly, despite extensive studies, still an active debate is

ongoing regarding the exact molecular structure of the isocyanate functional

group.6,7 For example, Sacher demonstrated that the isocyanate group of

phenyl isocyanate is positioned perpendicular to the benzene ring plane and

demonstrates distinctive cis/trans conformational isomerism (shown in

Scheme 2.3).8 Borisov et al. on the other hand described a completely planar

structure for the phenyl isocyanate molecule.9

22

Scheme 2.3. Perpendicular conformation and cis/trans conformational

isomerism of phenyl isocyanate reported by Sacher et al. Adapted from

Delebecq et al.6

These structural differences are significant, since they influence both the

spatial availability of the carbon centre as well as the electronic structure of

the functional group. The latter is a result of the (lack of) possible

interactions between the nitrogen lone-pair and the connecting moiety (e.g.

the aromatic ring system). Thereby, both the relative reactivities of the two

double bonds (the N=C and C=O double bond, respectively) as well as the

absolute reactivity of the functional group as a whole depend on the specific

electronic and spatial structure of the isocyanate group.

Aside from the specific structure of the isocyanate group itself, the

reactivity of isocyanates is strongly dependent on the moiety to which the

group is attached. Generally speaking, electron-withdrawing groups reduce

the electron density on the carbon atom and thereby enhance the reactivity

of the isocyanate group. This results in the following trend in reactivity:

ClSO2NCO > RSO2NCO (R = alkyl, aryl) > O=P(NCO)3 > aryl-NCO (p-NO2C6H4−

> p-ClC6H4− > p-CH3C6H4− > p-CH3OC6H4−) > alkyl-NCO.6,10 Most relevant in

this regard is the differentiation between aromatic and aliphatic

isocyanates, of which aromatic isocyanates are generally the more reactive

species. The higher reactivity of aromatic isocyanates can be considered

both an advantage (in terms of reaction rate and conversion) and a

disadvantage (in terms of selectivity and reduced control over (un)desired

side reactions). As a result, aromatic and aliphatic isocyanates are typically

used in different application fields and are rarely interchangeable. An

23

overview of commonly used industrial isocyanates is shown in Scheme 2.4:

methylenediphenyl diisocyanate (MDI) and toluene diisocyanate (TDI) are

the most common aromatic diisocyanates, while hexamethylene

diisocyanate (HDI) and isophorone diisocyanate (IPDI) are the most common

aliphatic diisocyanates.

Scheme 2.4. Examples of commonly used industrial aromatic and aliphatic

isocyanates: methylenediphenyl diisocyanate (MDI); toluene diisocyanate

(TDI); hexamethylene diisocyanate (HDI) and isophorone diisocyanate (IPDI).

Reactions of isocyanates with active hydrogen containing compounds:

formation of urethanes, ureas and equivalents

As mentioned previously, the electrophilicity of the central carbon atom is

one of the reasons that isocyanates are so useful. However, ultimately, the

combination of the electrophilic carbon atom with the adjacent nucleophilic

nitrogen atom is what makes isocyanates unique. It is owing to this specific

structure that isocyanates readily react with active hydrogen containing

compounds.6 The exploitation of this reactivity has led to the birth of what

would become a major class of industrial materials today:

polyurethanes.11,12

The most widely-adopted organic active hydrogen containing

reaction partners of isocyanates are alcohols and amines. Thiols and

carboxylic acids are used as well, but to a lesser extent. Besides organic

reaction partners, the reaction of isocyanates with water is also widely

adopted, especially in the production of foam materials. The chemical

equations of the most common isocyanate reactions with active hydrogen

containing compounds are shown in Table 2.1.

24

Table 2.1. Common active hydrogen containing reaction partners of

isocyanates and the resulting reaction products.

Reaction Partner / Product

Chemical Equation

Alcohol / Urethane

Amine / Urea

Thiol / Thio-urethane

Carboxylic Acid / Amide

Water / Urea

Urethane / Allophanat

e

Urea / Biuret

25

The resulting carbonyl compounds are generally thermodynamically

stable compounds, owing to their resonance-stabilized structure analogous

to the carboxamide structure.13 Common examples of isocyanate reaction

products are carbamate esters (generally referred to as urethanes) in the

case of reaction with an alcohol and ureas in the case of reaction with an

amine. Reactions of isocyanates with thiols or carboxylic acids result in the

formation of thiourethanes or amides, respectively. The reaction of an

isocyanate with water results in the formation of the unstable carbamic acid

moiety. After release of CO2 a free amine is formed, which typically instantly

reacts with a second isocyanate group to form the stable urea moiety.

Some of the reaction products of isocyanates still contain active

hydrogens. Also these can react with a second isocyanate moiety. These

reactions tend to occur less readily and generally require elevated

temperatures and/or specific catalysts. The most common such reactions

are the addition of an isocynate to a urethane group resulting in an

allophanate moiety and the addition of an isocyanate to a urea group

resulting in a biuret moiety.

Owing to the huge versatility of isocyanate reaction products,

catalysis plays a very important role in isocyanate chemistry. In this regard,

a high selectivity for a specific reaction product is often a more important

property of a catalyst than a reduction in reaction time. For the reactions of

isocyanates with active hydrogen containing compounds, two types of

catalysts are commonly employed. These are (organic) tertiary amines14,15

and organometallic salts (e.g. organo-lead16, -tin17, -zirconium18,19, -

magnesium20, or -iron21,22 compounds).6 Of these, dibutyltin dilaureate

(typically abbreviated DBTL or DBTDL) is historically the most widely

adopted catalyst for the urethanization reaction and is still often used as a

benchmark catalyst. In the case of DBTL, the combination of a high

selectivity for the urethanization reaction with a high catalyst activity is

considered attractive.23 Due to its high toxicity, the industrial use of DBTL is

more and more being reduced though. Thus far, no substitute catalyst with

the same level of selectivity and activity has been reported.

26

Reactions exclusively involving isocyanates: formation of uretdiones

(dimers), isocyanurates (trimers) and equivalents

Another important class of isocyanate reactions is comprised by reactions

exclusively involving isocyanates (in other words, isocyanates reacting with

themselves). The formation of cyclic dimers (uretdiones) and trimers

(isocyanurate or iminooxadiazinedione) are industrially the most important

examples. When reactions involving the addition or release of CO2 are also

taken into account, even more species become available. These include the

formation of carbodiimid, uretonimin and oxadiazinetrione. Finally,

isocyanates can polymerize to form linear polymeric or cyclooligomeric

structures, typically referred to as 1-nylons or α-nylons.24 The most common

reactions and reactions products exclusively involving isocyanates are

shown in Table 2.2.

Of these, the most widely-used reaction is the trimerization

reaction, which results in the formation of a trifunctional isocyanurate ring.

Owing to the symmetrical ring structure and the hyperconjugated π-system

parallel to the ring plane, isocyanurates are thermodynamically highly stable

compounds.25,26 On top of that, or arguably because of that, isocyanurates

form relatively easily, especially considering that three independent reacting

moieties are required in the formation. The trifunctional ring-structure and

relative ease of formation make isocyanurates ideal structures for the

synthesis of polyurethane networks. Because of the high thermodynamic

stability of these compounds, isocyanurates are especially attractive for

applications where chemical and/or thermal resistance are required.

Examples include flame-retardant construction panels (mostly aromatic

isocyanurates), industrial coatings and biomedical applications (mostly

aliphatic isocyanurates), which will be discussed in section 2.4.

27

Table 2.2. Common reactions and reaction products exclusively involving

isocyanate groups.

Formation of Reaction Equation

Uretdione

(dimer)27

Isocyanurate

(symmetrical

trimer)28

Iminooxadiazine

-dione

(asymmetrical

trimer)2930

Carbodiimid31

Uretonimin31

Oxadiazine-

trione32

Linear polymer

(α-nylon)24

28

Because of the wide variety of possible reaction products starting

from a single chemical moiety, catalysis and the choice of reaction

conditions plays an even more important role in reactions exclusively

involving isocyanates than in the reactions with active hydrogen containing

compounds discussed previously. Generally, nucleophilic, typically anionic

catalysts are used for reactions exclusively involving isocyanates. For

example, catalysts commonly used for the trimerization of isocyanates

include (alkali) metal alkoxides3, metal carboxylates33, (hetero)carbenes34,

trialkyl phosphines35 and rare-earth metal complexes33. Several examples of

commonly used isocyanate trimerization catalysts are shown in Scheme 2.5.

Scheme 2.5. Examples of commonly used isocyanate trimerization catalysts:

a) potassium methoxide3; b) potassium acetate33; c) tin octoate33; d) (CO2-

protected) N-carbene34; e) trialkyl phosphines35

2.3 Network Formation and Polyurethane Materials

The combination of the high reactivity of the starting material and the

thermodynamic stability of the reaction products makes isocyanates ideal

reactants for the production of polymeric materials. Especially the reaction

of (poly)isocyanates with poly(alcohols), resulting in polyurethane (PU) type

polymers has been of major industrial significance during the past 80

years.11,12 Many synthesis strategies have been developed over the years for

the production of PU materials. These generally involve the formation of a

3-dimensional polymer network of some sort. A main distinction can be

made between physical PU networks (also referred to as phase segregating

thermoplastic polyurethanes (TPUs)) on the one hand and chemically

crosslinked (covalent) PU networks on the other hand.

29

In TPU materials, a 3-dimensional physical polymer network is

formed based on the phase-segregation of urethane-group-containing hard

segments.36,37 Owing to the linear nature of the polymer chains, these

materials can be (re)processed at elevated temperatures. The durability of

these materials tends to be less than that of covalent PU networks though.

Since this thesis focusses on covalent PU networks, the reader is redirected

to the literature for a more detailed discussion on TPU materials.6,36–39

Covalent PU networks (or simply, PU networks) are widely adopted

materials owing to their generally excellent mechanical, optical, thermal and

chemical properties. Multiple approaches exist to effectively synthesize PU

networks, each having its advantages and disadvantages. The most common

PU network synthesis strategies, their benefits and limitations, as well as the

material properties of the resulting PU materials will be discussed below.

Two-component mixing of (poly)isocyanates and (poly)alcohols

Historically, the most widely adopted way of synthesizing PU networks is by

a 2-component mixing approach (also referred to as a 2K approach,

abbreviated from the German “2-Komponent”). In this approach, an

isocyanate-containing component is mixed with an alcohol containing

component upon application of the PU network. Once mixed, the free

isocyanate and alcohol groups react to form urethane groups. When both

components have an average functionality ≥2 and at least one of the

components has an average functionality ≥3, an infinte 3-dimensional

covalent PU network is formed.40,41 The synthesis of a PU network by a 2-

component mixing approach is schematically sketched in Scheme 2.6.

In 2K PU formulations, the (typically aliphatic) (poly)isocyanate is

often the component with a functionality ≥3. The reason for this is that high-

functionality aliphatic (poly)isocyanate oligomers are readily synthesized

based on the isocyanate addition reactions discussed previously in section

2.2.27 Specifically the trimerization of aliphatic diisocyanates such as HDI or

IPDI has been of great industrial significance during the past decades.29 The

resulting isocyanate-functional isocyanurates are broadly applied as PU

crosslinking components.

30

Scheme 2.6. Schematic representation of the synthesis of a PU network by a

2-component mixing approach.

The advantages of a 2-component approach are mostly the

relatively straightforward synthesis of the precursors with good control over

the precursor properties. As a result, 2K formulations are often the most

economic formulations. Also, because the reactive components are kept

physically separated until application, 2K formulations generally display a

beneficial combination of long shelf-life and short curing times.

The main disadvantage of a 2K approach is the requirement of

mixing of the two components upon application. Although this may seem

trivial, it has several important implications. Firstly, the final material

properties strongly depend on the mixing of components in the correct

stoichiometry. This requires a certain level of understanding of the

technology and/or clear instructions for the person applying the

formulation. As application is typically not performed by the manufacturer

but rather by an (industrial or private) consumer, 2K PU formulations are

generally not considered the most end-user-friendly approach.

Secondly, on a more fundamental level, a 2K approach requires the

two components to be readily miscible. This limits the scope of possible

combinations of precursors and thereby the range of accessible material

properties. More importantly though, adequate mixing requires the two

components to be of comparable and preferably low viscosity. However, the

synthesis of (poly)isocyanate oligomeric precursors is generally associated

with a strong increase in viscosity. For example, at room temperature the

viscosity of an industrial hexamethylene diisocyanate (HDI) trimer mixture is

on the order of 1000-3000 mPa·s while monomeric HDI has a viscosity of

only 2 mPa·s.29,42 As a result, a search for low-viscosity high-functionality (≥3)

aliphatic (poly)isocyanates has been going on for decades.42–45 A

31

fundamental understanding of the mechanisms dictating the viscosity of

aliphatic (poly)isocyanate precursors is lacking though. A framework for the

intermolecular interactions dictating the viscosity of aliphatic isocyanurates

is presented in Chapter 3 and 4 of this thesis.46,47

Thirdly, the PU networks resulting from a 2-component crosslinking

mechanism are typically poorly defined on a molecularl level. In an industrial

setting, this is generally not considered a problem. On an acdemic level

though, accurate structure-property relations studies require molecularly

well-defined polymer networks. The development of a method for

synthsizing well-defined poly(urethane-isocyanurate) networks is presented

in Chapter 5 of this thesis.48

Alternative crosslinking mechanisms: 1- or 1.5-component formulations,

blocked isocyanates and polyaddition reactions

During the past decades, several alternative PU crosslinking methods have

been developed to improve on or replace the 2-component crosslinking

approach. The first strategy is by the use of so-called blocked isocyanates.49

In this case, the reactive isocyanate groups of a (poly)isocyanate mixture are

reacted with a monofunctional active hydrogen containing compound called

a blocking agent. The resulting compounds are thermodynamically stable at

low temperature, and can be mixed with conventional OH-functional

components and/or additives without crosslinking. At elevated

temperatures, the blocking reaction is reversed, whereby the small-

molecular blocking agent is released and the reactive isocyanate group is

regenerated in-situ. The isocyanate groups subsequently react with the OH-

functional reaction partner in the conventional manner, resulting in a 3-

dimensional covalent PU network. The blocking agent typically evaporates

from the formulation. The general synthetic approach and several common

blocking agents are shown in Scheme 2.7.

32

Scheme 2.7. Blocked isocyanates are adopted to realize 1-component PU

fomulations; a) blocking and deblocking reactions using a blocking agent (H-

Bl). The reactive isocyanate group is generated in-situ by heating and

subsequent release of the blocking agent; b) examples of common blocking

agents.

The main advantage of this technique is the possibility for 1-component PU

formulations. This means that PU raw material manufacturers can control a

larger part of the production and application process, such as the correct

mixing of components. The resulting 1-component PU systems are more

enduser-friendly as adequate application relies less on the chemical

expertise of the applier. Also, because the final formulation does not contain

any free isocyanate groups, blocked isocyanate precursors are typically

more friendly to natural environments.

There are several disadvantages associated with blocked isocyanate

technologies though. Firstly, blocked isocyanate systems are more difficult

to cure than 2K systems. They require elevated temperatures to achieve a

full curing of the PU network, which limits the scope of application. Also, the

curing times of blocked isocyanate systems are typically longer compared to

their 2K equivalents. Secondly, the release of the blocking agent upon curing

is considered a disadvantage. This can potentially have a negative impact on

the final material properties (e.g. in terms of blistering or color change).

33

Also, the amount of volatile organic compounds (VOCs) that are released

upon application is increased, which has a negative environmental impact

and potentially introduces health and safety risks. Thirdly, because of the

extra blocking step and specialty chemicals involved, blocked isocyanate

systems are typically more expensive than their 2K equivalents.

Another alternative crosslinking mechanism increasingly used in industry is

crosslinking by trimerization.50 In this case, the trimerization reaction of

isocyanates is employed for the final crosslinking step instead of for the

synthesis of (poly)isocyanate-functional oligomer precursors.51,52 The

incorporation of conventional OH-functional components is either

performed beforehand (Scheme 2.8a) or is omitted (Scheme 2.8b).

Scheme 2.8. Increasingly, isocyanate addition reactions (e.g. trimerization)

are being used as the crosslinking reaction instead of for the synthesis of

isocyanate-functional oligomeric precursors; a) synthesis of an NCO-

functional prepolymer and subsequent crosslinking by trimerization; b)

synthesis and crosslinking of an isocyanate-functional precursor by

trimerization, resulting in a dense polyisocyanurate network known as

TrimTrim. No OH-functional component is incorporated in this case, so the

final network does not contain any urethane-bonds.

34

There are many advantages associated with this type of crosslinking

mechanism. Firstly, the technology allows the formulation of 1-component

PU systems. More often though, formulations are prepared to which only a

suitable catalyst has to be added.48 These are also referred to as 1.5-

component systems. What makes crosslinking by trimerization unique, is

that no (further) addition or release of chemicals is required to achieve

network formation. This is a clear advantage over the other technologies

described previously, which all rely on either the addition or release of

chemical compounds in stoichiometric amounts. Secondly, owing to the 1-

component nature of the precursor, reaction mixtures are typically

homogeneous in every stage of the reaction. The miscibility issues that are

sometimes encountered for 2K formulations are thus less frequent when

crosslinking by trimerization. Thirdly, because the molecular weight of the

precursors (e.g. an NCO-functional prepolymer) is typically higher than those

of oligomeric (poly)isocyanate precursors, the concentration of free

isocyanate groups is relatively low.

The disadvantages to crosslinking by trimerization are two-fold.

Firstly, when incorporating an OH-functional component, urethane groups

are present in the 1(.5)-component precursor. Since viscosity strongly

increases with the presence and increasing relative content of urethane

groups, the viscosity of the 1(.5)-component precursors is typically high.53

Secondly, the relatively low concentration of free isocyanate groups reduces

activity. As a result, curing times are typically significantly longer than for 2K

PU equivalents. Also, the trimerization reaction typically requires elevated

temperatures, which limits the scope of application.

Some of the disadvantages of the various crosslinking mechanisms discussed

above can be adressed by the use of a so-called dual cure system, as shown

in Scheme 2.9. In this case, PU crosslinking chemistry (e.g. crosslinking by

the trimerization of isozyanates) is combined with a second, complementary

type of crosslinking chemistry (e.g. the free radical crosslinking of

(meth)acrylates under UV-exposure).54 Precursors for such dual cure

systems contain both types of functional groups (e.g. isocyanates and

35

(meth)acrylates). The two complementary crosslinking mechanisms can be

activated either simultaneously or sequentially.

Scheme 2.9. A dual cure mechanisms can be adopted to achieve fast initial

crosslinking whilst still obtaining the good material properties associated

with PU chemistry in a post-curing step (top). Alternatively, a dual cure

mechanism can be used to the increase the network density of a PU network

after crosslinking (bottom).

The main advantage of using a dual cure system is the possibility to optimize

between curing time and final material properties. This is possible because

the crosslinking reaction of acrylates typically occurs faster than the

crosslinking reactions used in PU chemistry (e.g. trimerization).

Consequently, a covalent polymer network can be formed in the early stages

of application (i.e. on a timescale of seconds or minutes, which is difficult to

realize when relying only on the trimerization of isocyanates). The slower

trimerization reaction can then be used as a post-curing step to enhance the

material properties. Alternatively, crosslinking by trimerization can be

performed first, after which the crosslinking of (meth)acrylates can be used

as a post-curing step to increase network density.

36

Overall, many PU crosslinking mechanisms have been developed over the

years, each having their specific advantages and disadvantages. Novel PU

crosslinking technologies are still actively investigated though. One such

novel PU crosslinking technique, which is based on the trimerization of NCO-

functionalized prepolymers with a narrow molecular weight distribution, is

presented in Chapter 5 of this thesis.48

Material properties of polyurethane networks

Owing to the broad range of possible starting materials and reaction

products, PU materials are largely characterized by versatility. Generally

speaking, PUs are considered high-performance materials, well-known for

their excellent mechanical, optical, chemical and biomedical properties.55

These include high ultimate tensile strengths and high toughness values56,

high transparency and low coloration, high chemical and thermal resistance

and good biocompatibility57. Compared to other types of polymer networks

(e.g. UV-cured (meth)acrylate networks), PU networks are generally

considered tougher and more durable.55,58

PU materials containing isocyanurate rings as the crosslinking unit

(poly(urethane-isocyanurate)s or PUIs) especially demonstrate excellent

material properties, most notably in terms of thermal stability, optical

quality, durability and toughness.59–61 The high performance of such PUI type

materials is generally attributed to the thermodynamic stability of the

isocyanurate ring structure, as discussed previously in section 2.2.25,26 As a

result, PUI networks typically demonstrate improved chemcial and thermal

resistance compared to PU networks containing other types of crosslinking

units.33 The syntheses and material properties of novel PUI networks are

presented in Chapters 5, 6, 7 and 8 of this thesis.48,62–64

To illustrate the high resilience of materials crosslinked by

isocyanurate rings, the glass transition temperatures of various dense

polyisocyanurate networks are shown in Scheme 2.10 and compared to an

analogous 2K PU network that does not contain isocyanurate rings.

37

Scheme 2.10. The thermal properties such as the glasss transition

temperature (Tg) of dense isocyanurate networks (TrimTrim) can easily be

tailored over a broad range of values by combining different types of

isocyanate-functional precursors. In this example, the relative content of

IPDI-trimer in an HDI-trimer/IPDI-trimer mixture is varied. These are

compared to a conventional 2K PU network based on mixing an industrial

HDI-trimer with an industrial polyacrylate polyol.

2.4 Applications of Polyurethane Materials

Owing to the wide versatility and widely varying material properties, PU

materials are found in a broad range of applications. PU application fields

are generally characterized by a high demand on mechanical, optical and/or

biological properties, which is where PU materials are distinctive from other

material types. Since this thesis focusses on the use of aliphatic PU networks,

the most important application fields of aliphatic PU networks will be

discussed below. For an overview of the use of other types of PU materials,

such as TPUs or networks based on aromatic PUs, the reader is directed to

the literature.39,65

Coatings, adhesives and sealants

Historically, the largest application field of aliphatic PU networks is the

coatings, adhesives and sealants industry. Of these, the coatings field is by

far the largest, with an annual global production of PU resins of

38

approximately 1.600 kilotons.66 This compares to an annual global

production of PU resins for the combined adhesives and sealants industries

of approximately 0.8 kilotons.66

The use of PU resins for coatings is widespread and includes wood

coating, metal coating (e.g. coil coating, can coating and protective marine

coating), automotive (re)finish and transportation coating, plastic coating,

glass coating, textile and leather coating, paper coating and coating of

construction materials (e.g. walls and floors). The distribution (per volume)

of coating materials over the various application fields in 2017 is shown in

Scheme 2.11.

Scheme 2.11. Formulated coating volumes per application in 2017 (industrial

and architectural). Adapted from U. Meier-Westhues et al.66

Compared to other coating technologies, PU coatings are generally

appreciated for their high versatility and performance. Especially the high

toughness, high chemical, thermal and mechanical resistance, high

durability and good optical appearance associated with PU networks are

attractive properties for coatings applications. In Scheme 2.12, the results of

a benchmark study on the performance of various industrial clear coat

39

systems for automotive applications are shown.67 Clearly, the PU-based

technologies are superior compared to the other coating technologies in the

market.

Scheme 2.12. Clear coat benchmark study showing the performance levels of

various industrial clear coat systems with regard to scratch and acid

resistance: two-component high-solid polyurethane (2K PU), one-component

PU modified thermosetting acrylics (1K PU), one-component carbamate-

modified thermosetting acrylics (1K carbamate TSA), one-component acid

epoxy (1K Acid-Epoxy), one-component silane-modified thermosetting

acrylics (1K Si-TSA) and one-component thermosetting acrylics (1K TSA)

systems were tested. 2K PU systems are the overall best-performing

materials on the marktet. Adapted from U. Meier-Westhues et al.67

High-performance coatings based on PU technologies are found

across many application fields with completely different demands on

material properties. In Scheme 2.13, several examples of PU-based coating

applicaitions are shown, illustrating the broad versatility of PUs. These

include industrial soft-feel plastic coatings, marine protective metal coatings

and decorative glass coatings. The common denominator is the high demand

on durability and optical appearance.

40

Scheme 2.13. Examples of PU coating applications include, but are not

limited to: a) industrial soft-feel plastic coatings; b) marine protective

coatings; c) decorative glass coatings. Adapted from U. Meier-Westhues et

al 67

Polyurethanes for biomedical applications

An important application field for PU materials is the field of biomedical

materials and devices.55,68 Also in this field, PU materials are recognized for

their mechanical properties and durability, this time combined with their

generally good biocompatibility.69–71 As a result, PUs are increasingly used as

biomaterials, with applications including: artificial heart valves, vascular

protheses, pericardial patches, artificial blood pumps, vascular intra-aortic

balloons, catheter and cannula coating, sutures, artificial ligaments, wound

dressings, shock absorbing elements and pacemakers.72 The high toughness

of PU materials has even made it possible to replace some conventional

load-bearing materials, such as metal (alloy)s and ceramics.73 A recently

developed aliphatic PU-based wound dressing is shown in Scheme 2.14.

41

Scheme 2.14. Schematic view and application of an aliphatic PU-based

wound dressing. Adapted from U. Meier-Westhues et al.74

In Chapter 8 of this thesis, a novel potential biomedical application

for PUI networks is explored.64 Making use of the beneficial combination of

high toughness, excellent optical properties and good biocompatibility, the

potential application of PUI network hydrogels as contact lens materials is

investigated.

2.5 Conclusions The isocyanate group has a unique chemical structure, containing both an

electrophilic and nucleophilic reaction centre. The presence of a carbonyl

structure adjacent to a nitrogen heteroatom results in the useful

combination of high reactivity towards active hydrogen containing

compounds and thermodynamic (resonance-)stability of the resulting

carbonyl reaction products. The most important examples are the reactions

with alcohols and amines, resulting in urethane- and urea moieties,

respectively. Additionally, many reactions are possible that exclusively

involve isocyanates. The most important one of these is the symmetrical

trimerization of isocyanates, yielding the thermodynamically highly stable

trifunctional isocyanurate ring structure.

Owing to the high reactivity of the starting material and the

thermodynamic stability of the resulting structures, isocyanates are ideal

reactants for the production of polymer materials. Crosslinked aliphatic

polyurethane (PU) networks comprise an industrially significant class of

materials and many strategies exist to synthesize them. A 2-component (2K)

42

mixing approach, in which high-functionality (poly)isocyanate- and

(poly)alcohol components are mixed upon application, is the most widely-

adopted. This approach offers the easiest synthesis of starting materials and

is often the most economic. The largest disadvantage is the need to mix the

two components upon application. This makes the technique less end-user-

friendly and requires the two components to be of comparable and

preferably low viscosity. Several alternative crosslinking techniques

presently exist, the most promising of which is crosslinking by the

trimerization of isocyanates.

As PU materials are characterized by their versatility, so are their

application fields. For crosslinked aliphatic PU materials, the industrially

most significant application field is the coatings, adhesives and sealants

industry. The high toughness and thermodynamic stability typically

associated with PU materials result in a superior durability and thermal,

mechanical and chemical resistance compared to other coating techniques.

The superior properties of PU materials are also recognized in emerging

application fields. An example of a more recently developed area of

application is the use of PUs as biomedical materials. Also in this relatively

new field of application, PU materials are considered high-performing and

durable compared to existing techniques.

43

2.6 References 1. Meier-Westhues, U., Danielmeier, K., Kruppa, P. & Squiller, E. P.

Foreword. in Polyurethanes − Coatings, Adhesives, Sealants (eds. Meier-Westhues, U., Danielmeier, K., Kruppa, P. & Squiller, E. P.) (Vincentz-Network Verlag, 2019).

2. Wurtz, A. Reserches sur les éthers cyanique et leurs dérivés. Compt. Rend. 27, 241−243 (1848).

3. Arnold, R. G., Nelson, J. A. & Verbanc, J. J. Recent Advances In Isocyanate Chemistry. Chem. Rev. 57, 47–76 (1957).

4. Ozaki, S. Recent Advances in Isocyanate Chemistry. Chem. Rev. 72, 457–496 (1972).

5. Rauk, A. & Alewood, P. F. A theoretical study of the Curtius rearrangement. The electronic structures and interconversions of the CHNO species. Can. J. Chem. 55, 1498–1510 (1977).

6. Delebecq, E., Pascault, J. P., Boutevin, B. & Ganachaud, F. On the versatility of urethane/urea bonds: Reversibility, blocked isocyanate, and non-isocyanate polyurethane. Chem. Rev. 113, 80–118 (2013).

7. Caraculacu, A. A. & Coseri, S. Isocyanates in polyaddition processes. Structure and reaction mechanisms. Prog. Polym. Sci. 26, 799–851 (2001).

8. Sacher, E. A re-examination of the polyurethane reaction. J. Macromol. Sci. Part B 16, 525–538 (1979).

9. Borisov, E. V., Vasyanina, L. K., Bondarenko, S. P. & Zabrodin, V. B. Electronic structure of the conformation and properties of phenyl isocyanate. Zh. Fiz. Khim. 51, 1930 (1977).

10. Ozaki, S. Recent Advances in Isocyanate Chemistry. J. Synth. Org. Chem. Japan 28, 17–40 (1970).

11. Bayer, O. Verfahren zur Herstellung von Polyurethanen bzw. Polyharnstoffen. (1937).

12. Bayer, O. Das Di-Isocyanat-Polyadditionsverfahren (Polyurethane). Angew. Chemie 59, 257–272 (1947).

13. Volkov, A., Tinnis, F., Slagbrand, T., Trillo, P. & Adolfsson, H. Chemoselective reduction of carboxamides. Chem. Soc. Rev. 45, 6685–6697 (2016).

14. Chang, M.-C. & Chen, S.-A. Kinetics and mechanism of urethane reactions: Phenyl isocyanate–alcohol systems. J. Polym. Sci. Part A Polym. Chem. 25, 2543–2559 (1987).

15. Baker, J. W. & Holdsworth, J. B. 135. The mechanism of aromatic side-chain reactions with special reference to the polar effects of substituents. Part XIII. Kinetic examination of the reaction of aryl

44

isocyanates with methyl alcohol. J. Chem. Soc. 713 (1947). doi:10.1039/jr9470000713

16. Rand, L., Thir, B., Reegen, S. L. & Frisch, K. C. Kinetics of alcohol–isocyanate reactions with metal catalysts. J. Appl. Polym. Sci. 9, 1787–1795 (1965).

17. Han, Q. & Urban, M. W. Kinetics and mechanisms of catalyzed and noncatalyzed reactions of OH and NCO in acrylic polyol-1,6-hexamethylene diisocyanate (HDI) polyurethanes. VI. J. Appl. Polym. Sci. 86, 2322–2329 (2002).

18. Blank, W. J., He, Z. A. & Hessell, E. T. Catalysis of the isocyanate-hydroxyl reaction by non-tin catalysts. Prog. Org. Coatings 35, 19–29 (1999).

19. Petrak, S., Shadurka, V. & Binder, W. H. Cleavage of blocked isocyanates within amino-type resins: Influence of metal catalysis on reaction pathways in model systems. Prog. Org. Coatings 66, 296–305 (2009).

20. Gertzmann, R. & Gürtler, C. A catalyst system for the formation of amides by reaction of carboxylic acids with blocked isocyanates. Tetrahedron Lett. 46, 6659–6662 (2005).

21. Ahmad, I., Zaidi, J. H., Hussain, R. & Munir, A. Synthesis, characterization and thermal dissociation of 2-butoxyethanol-blocked diisocyanates and their use in the synthesis of isocyanate-terminated prepolymers. Polym. Int. 56, 1521–1529 (2007).

22. Korah Bina, C., Kannan, K. G. & Ninan, K. N. DSC study on the effect of isocyanates and catalysts on the HTPB cure reaction. J. Therm. Anal. Calorim. 78, 753–760 (2004).

23. Silva, A. L. & Bordado, J. C. Recent developments in polyurethane catalysis: Catalytic mechanisms review. Catal. Rev. - Sci. Eng. 46, 31–51 (2004).

24. King, C. Cyclopolymerization of Aliphatic 1,2-Diisocyanates. J. Am. Chem. Soc. 86, 437–440 (1964).

25. Heift, D., Benko, Z., Grützmacher, H., Jupp, A. R. & Goicoechea, J. M. Cyclo-oligomerization of isocyanates with Na(PH2) or Na(OCP) as ‘p-’ anion sources. Chem. Sci. 6, 4017–4024 (2015).

26. Okumoto, S. & Yamabe, S. A computational study of base-catalyzed reactions between isocyanates and epoxides affording 2-oxazolidones and isocyanurates. J. Comput. Chem. 22, 316–326 (2001).

27. Laas, H. J., Halpaap, R. & Pedain, J. Zur Synthese aliphatischer Polyisocyanate − Lackpolyisocyanate mit Biuret-, Isocyanurat- oder Uretdionstruktur. J. Prakt. Chem./Chem.-Ztg. 336, 185−200 (1994).

45

28. Bock, M., Pedain, J. & Uerdingen, W. Process for preparing polyisocyanates with isocyanate groups and their use as isocyanate components in polyurethane varnishes. (1994).

29. Richter, F., Pedain, J., Mertes, H. & Dieris, C.-G. Isocyanate trimers and mixtures of isocyanate trimers, production and use thereof. 13–14 (1997).

30. Decostanzi, M., Auvergne, R., Darroman, E., Boutevin, B. & Caillol, S. Reactivity and kinetics of HDI-iminooxadiazinedione: Application to polyurethane synthesis. Eur. Polym. J. 96, 443–451 (2017).

31. Wagner, K., Findeisen, K., Schäfer, W. & Dietrich, W. α,ω-Diisocyanatocarbodiimides, -Polycarbodiimides, and Their Derivatives. Angew. Chemie Int. Ed. English 20, 819–830 (1981).

32. Scholl, H.-J. Process for the preparation of modified isocyanurate-containing polyisocyanates and use of them. (1990).

33. Flipsen, T. A. C., Steendam, R., Pennings, A. J. & Hadziioannou, G. A novel thermoset polymer optical fiber. Adv. Mater. 8, 45–48 (1996).

34. Bantu, B. et al. CO2 and SnII Adducts of N-Heterocyclic Carbenes as Delayed-Action Catalysts for Polyurethane Synthesis. Chem. - A Eur. J. 15, 3103–3109 (2009).

35. Richter, F. U. 1,4,2-Diazaphospholidine-3,5-diones and Related Compounds: A Lecture on Unpredictability in Catalysis. Chem. - A Eur. J. 15, 5200–5202 (2009).

36. Yilgor, I. & Yilgor, E. Structure-morphology-property behavior of segmented thermoplastic polyurethanes and polyureas prepared without chain extenders. Polym. Rev. 47, 487–510 (2007).

37. Yilgör, I., Yilgör, E. & Wilkes, G. L. Critical parameters in designing segmented polyurethanes and their effect on morphology and properties: A comprehensive review. Polymer (Guildf). 58, A1–A36 (2015).

38. Sami, S. et al. Understanding the influence of hydrogen bonding and diisocyanate symmetry on the morphology and properties of segmented polyurethanes and polyureas: Computational and experimental study. Polymer (Guildf). 55, 4563–4576 (2014).

39. Engels, H.-W. et al. Polyurethanes: Versatile Materials and Sustainable Problem Solvers for Today’s Challenges. Angew. Chemie Int. Ed. 52, 9422–9441 (2013).

40. Dusek, K. Networks from Telechelic Polymers: Theory and Application to Polyurethanes. in Telechelic Polymers: Synthesis and Applications 289–360 (CRC Press, 1989).

41. Erdödi, G. & Iván, B. Novel Amphiphilic Conetworks Composed of Telechelic Poly(ethylene oxide) and Three-Arm Star

46

Polyisobutylene. 16, 959–962 (2004). 42. Renz, H. & Bruchmann, B. Pathways targeting solvent-free PUR

coatings. Prog. Org. Coatings 43, 32–40 (2001). 43. Chattopadhyay, D. K. & Raju, K. V. S. N. Structural engineering of

polyurethane coatings for high performance applications. Prog. Polym. Sci. 32, 352–418 (2007).

44. Ludewig, M., Weikard, J. & Stockel, N. Allophanate-containing modified polyurethanes (US 2006/0205911 A1). (2006).

45. Woicik, R. T., Goldstein, S. L., Barnowski, H. G. & Chandalis, K. B. Novel Super Low Viscosity Aliphatic Isocyanate Crosslinker for Polyurethane Coatings. J. Cell. Plast. 27, 64–64 (1991).

46. Driest, P. J. et al. Aliphatic isocyanurates and polyisocyanurate networks. Polym. Adv. Technol. 28, 1299–1304 (2017).

47. Driest, P. J. et al. The Influence of Molecular Interactions on the Viscosity of Aliphatic Polyisocyanate Precursors for Polyurethane Coatings. ECS Conf. Proc. (2019).

48. Driest, P. J., Dijkstra, D. J., Stamatialis, D. & Grijpma, D. W. The Trimerization of Isocyanate-Functionalized Prepolymers : An Effective Method for Synthesizing Well-Defined Polymer Networks. Macromol. Rapid Commun. 40, 1800867 (2019).

49. Tork, L., Rottmaier, L. & Hohne, W. Preparations and processes for finishing leather and for coating textiles. (1988).

50. Samborska-Skowron, R. & Balas, A. An overview of developments in poly(urethane-isocyanurates) elastomers. Polym. Adv. Technol. 13, 653–662 (2002).

51. Sasaki, N., Yokoyama, T. & Tanaka, T. Properties of isocyanurate-type crosslinked polyurethanes. J. Polym. Sci. Polym. Chem. Ed. 11, 1765–1779 (1973).

52. Špírková, M., Budinski-Simendic, J., Ilavský, M., Špaček, P. & Dušek, K. Formation of poly(urethane-isocyanurate) networks from poly(oxypropylene)diols and diisocyanate. Polym. Bull. 31, 83–88 (1993).

53. Rimdusit, S., Bangsen, W. & Kasemsiri, P. Chemorheology and thermomechanical characteristics of benzoxazine-urethane copolymers. J. Appl. Polym. Sci. 121, 3669–3678 (2011).

54. Konuray, O. et al. State of the Art in Dual-Curing Acrylate Systems. Polymers (Basel). 10, 178 (2018).

55. Akindoyo, J. O. et al. Polyurethane types, synthesis and applications – a review. RSC Adv. 6, 114453–114482 (2016).

56. Petrović, Z. S. & Ferguson, J. Polyurethane elastomers. Prog. Polym. Sci. 16, 695–836 (1991).

47

57. Burel, F., Poussard, L., Tabrizian, M., Merhi, Y. & Bunel, C. The influence of isocyanurate content on the bioperformance of hydrocarbon-based polyurethanes. J. Biomater. Sci. Polym. Ed. 19, 525–540 (2008).

58. Zant, E., Bosman, M. J. & Grijpma, D. W. Combinatorial synthesis of photo-crosslinked biodegradable networks. J. Appl. Biomater. Funct. Mater. 10, 197–202 (2012).

59. Fabbri, P. et al. Highly stable plastic optical fibre amplifiers containing [Eu(btfa)3(MeOH)(bpeta)]: A luminophore able to drive the synthesis of polyisocyanates. Polymer (Guildf). 55, 488–494 (2014).

60. Moritsugu, M., Sudo, A. & Endo, T. Cyclotrimerization of diisocyanates toward high-performance networked polymers with rigid isocyanurate structure: Combination of aromatic and aliphatic diisocyanates for tunable flexibility. J. Polym. Sci. Part A Polym. Chem. 51, 2631–2637 (2013).

61. Preis, E., Schindler, N., Adrian, S. & Scherf, U. Microporous Polymer Networks Made by Cyclotrimerization of Commercial, Aromatic Diisocyanates. ACS Macro Lett. 4, 1268–1272 (2015).

62. Driest, P. J., Dijkstra, D. J., Stamatialis, D. & Grijpma, D. W. Tough Combinatorial Poly(urethane-isocyanurate) Polymer Networks and Hydrogels Synthesized by the Trimerization of Mixtures of NCO-Prepolymers. to be Submitt. (Chapter 6 this Thesis) (2019).

63. Driest, P. J., Dijkstra, D. J., Stamatialis, D. & Grijpma, D. W. Structure-Property-Relations in Semi-crystalline Combinatorial Poly(urethane-isocyanurate) Hydrogels. to be Submitt. (Chapter 7 this Thesis) (2019).

64. Driest, P. J., Allijn, I. E., Dijkstra, D. J., Stamatialis, D. & Grijpma, D. W. Poly(ethylene glycol) Poly(urethane-isocyanurate) Type Hydrogels as a New Class of Materials for Contact Lens Applications. to be Submitt. (Chapter 8 this Thesis) (2019).

65. Brereton, G. et al. Polyurethanes. in Ullmann’s Encyclopedia of Industrial Chemistry 1–76 (Wiley-VCH Verlag GmbH & Co. KGaA, 2019). doi:10.1002/14356007.a21_665.pub3

66. Chapter 2: Economic aspects and market analysis. in Polyurethanes − Coatings, Adhesives, Sealants (eds. Meier-westhues, U., Danielmeier, K., Kruppa, P. & Squiller, E. P.) (Vincentz-Network Verlag, 2019).

67. Chapter 5: Coatings. in Polyurethanes − Coatings, Adhesives, Sealants (eds. Meier-Westhues, U., Danielmeier, K., Kruppa, P. & Squiller, E. P.) (Vincentz-Network Verlag, 2019).

48

68. Cooper, S. L. & Guan, J. Advances in polyurethane biomaterials. (Woodhead Publishing, 2016).

69. Zhou, B., Hu, Y., Li, J. & Li, B. Chitosan/phosvitin antibacterial films fabricated via layer-by-layer deposition. Int. J. Biol. Macromol. 64, 402–8 (2014).

70. Zhou, X., Zhang, T., Guo, D. & Gu, N. A facile preparation of poly(ethylene oxide)-modified medical polyurethane to improve hemocompatibility. Colloids Surfaces A Physicochem. Eng. Asp. 441, 34–42 (2014).

71. Wang, Y., Hong, Q., Chen, Y., Lian, X. & Xiong, Y. Surface properties of polyurethanes modified by bioactive polysaccharide-based polyelectrolyte multilayers. Colloids Surfaces B Biointerfaces 100, 77–83 (2012).

72. Khatoon, H. & Ahmad, S. Polyurethane: A Versatile Scaffold for Biomedical Applications. Significances Bioeng. Biosci. 2, (2018).

73. Middleton, J. C. & Tipton, A. J. Synthetic biodegradable polymers as orthopedic devices. Biomaterials 21, 2335–2346 (2000).

74. Chapter 8: New areas of application for polyurethanes. in Polyurethanes − Coatings, Adhesives, Sealants (eds. Meier-Westhues, U., Danielmeier, K., Kruppa, P. & Squiller, E. P.) (Vincentz-Network Verlag, 2019).

49

Chapter 3 - Aliphatic isocyanurates and

polyisocyanurate networks

P.J. Driest1,2, V. Lenzi3, L.S.A. Marques3, M.M.D. Ramos3, D.J. Dijkstra1, F.U.

Richter1, D. Stamatialis2 and D.W. Grijpma2,4

1Covestro, CAS-Global R&D, 51365 Leverkusen, Germany

2MIRA Institute for Biomedical Technology and Technical Medicine,

Department of Biomaterials Science and Technology, University of Twente,

P.O. Box 217, 7500 AE Enschede, The Netherlands

3Department/Centre of Physics of University of Minho, campus de Gualtar,

4710-057 Braga, Portugal

4University Medical Center Groningen, W.J. Kolff Institute, Department of

Biomedical Engineering, P.O. Box 196, 7600 AD Groningen, The Netherlands

Published in:

Polymers for Advanced Technologies 28 (2017), 1299–1304

Abstract

The production, processing and

application of aliphatic isocyanate

(NCO) based thermosets such as

polyurethane coatings and

adhesives are generally limited by

the surprisingly high viscosity of tri-

and higher-functionality

isocyanurates. These compounds

are essential crosslinking additives

for network formation. However, the mechanism by which these high

viscosities are caused is not yet understood.

In this work, model aliphatic isocyanurates were synthesized and

isolated in high purity (>99%) and their viscosities were accurately

determined. It was shown that the presence of the NCO group has a strong

influence on the viscosity of the system. From density functional theory

(DFT) calculations a novel and significant bimolecular binding potential of -

8.7 kJ/mol was identified between NCO groups and isocyanurate rings,

confirming the important role of the NCO group. This NCO-to-ring

interaction was proposed to be the root cause for the high viscosities

observed for NCO-functional isocyanurate systems. MD simulations carried

out to further confirm this influence also suggest that the NCO-to-ring

interaction causes a significant additional contribution to viscosity. Finally,

model functional isocyanurates were further reacted into densely

crosslinked polyisocyanurate networks which showed interesting material

properties.

52

3.1 Introduction The current work is aimed at better understanding the viscosity

characteristics of isocyanate (NCO) based crosslinkers, an essential

parameter in their production and processing. Isocyanates, as primary

ingredients of polyurethanes, have been of wide interest in both industry

and academia ever since polyurethanes were invented.1-4 Aliphatic

diisocyanates such as 1,6-hexamethylenediisocyanate (HDI) and its

derivatives are produced on a large scale worldwide. They are well-known

for their high performance regarding UV resistance, as well as their excellent

mechanical, chemical, optical and thermal properties in polyurethane

coatings and adhesives. Of particular significance in the performance of

these materials are the higher-functionality trimerized isocyanurate forms

of the diisocyanates, key additives as cross-linking units for network

formation. The chemical and physical stability of these cross-linking units,

resulting from the deep thermodynamic minimum of the isocyanurate ring

structure5,6, is key in the high performance of modern polyurethane coatings

and adhesives. Furthermore, the possibility of trimerizing such trifunctional

isocyanurate additives in a pure form has been investigated. This results in

the formation of highly stable densely crosslinked polyisocyanurate

networks, which show interesting material properties. In particular the good

optical properties and high chemical and thermal stability of these materials

are attractive qualities.7-10

Given that the same basic starting materials are used in the

preparation of many isocyanate-based thermosets, including well-

established polyurethane coatings, adhesives and densely crosslinked

polyisocyanurate networks, a good understanding of the physical behavior

of these precursors would be beneficial. However, a major problem

encountered in the use of aliphatic functional isocyanurates is their

relatively high viscosity, which complicates their production, processing and

application. Consequently, the search for low-viscosity high-functionality

(≥3) aliphatic isocyanates has been a topic of continuous interest for the past

decades.11-13 Surprisingly, however, very little is known about the

fundamental causes of the increase of viscosity in such systems, and the

underlying chemistry is hardly understood. Although surprisingly high

53

viscosity values have been observed and reported for aliphatic

isocyanurates (e.g., 1000-3000 mPa∙s at room temperature for industrial

HDI-based isocyanurates, compared to 2 mPa∙s for monomeric HDI13,14), a

chemical explanation for the cause of such high values has not yet been

presented. Combining the need for a better understanding of the physical

behavior of these precursors with a further investigation of the

polyisocyanurate networks made from them, we set out with three main

objectives:

1. to quantify accurately the viscosity increase upon trimerization of

aliphatic diisocyanates into NCO-functional aliphatic isocyanurates;

essential precursors for many isocyanate-based thermosets;

2. to investigate the underlying chemistry to explain the unexpectedly

high viscosities observed for NCO-functional aliphatic isocyanurates;

and

3. to further react model isocyanurates into densely crosslinked

polyisocyanurate networks and characterize the physical and

thermal properties of the resulting materials.

To investigate the viscosity characteristics of functional (network-forming)

isocyanurates, a series of model diisocyanates with varying carbon chain

lengths based on and including HDI was prepared. Also, a similar series of

monofunctional isocyanates, yielding non-functional (non-network-

forming) isocyanurates, was included as a reference. The relevant structures

and general synthetic approach are presented in Scheme 3.1.

Scheme 3.1. Structure and general synthesis of model isocyanurates.

54

3.2 Experimental Details

Materials and methods

α-ω-Linear diisocyanates were taken from internal industrial production

(Covestro Deutschland AG) and vacuum distilled before use. n-

Butylisocyanate and n-hexylisocyanate were purchased at Aldrich and

distilled before use. n-Pentylisocyanate and n-heptylisocyanate were

synthesized by reaction of corresponding amines (Sigma-Aldrich) with 1,6-

hexamethylenediisocyanate. Dodecylbenzene sulfonic acid was supplied by

Sigma-Aldrich and used as received, as were all other chemicals described.

Viscosity measurements were performed at RT on an Anton-Paar MCR501

rheometer using either a 50 mm cone-plate setup (CP50-2, d = 0.214 mm)

or a CC27 beaker-cup setup, in rotational mode at increasing shear rate from

0.1-100 s-1 (0.1-1000 s-1 for beaker-cup). DSC measurements were

performed under nitrogen on 10 mg samples on a Perkin-Elmer Kalorimeter

DSC-7 in 2 heating runs from -20°C-200°C by 20°C/min. TGA measurements

were performed under nitrogen on 4 mg samples on a Perkin-Elmer

Mikrothermowaage TGA-7 from RT to 600°C by 20°C/min. NMR-spectra

were collected on a Bruker Avance III - 700. FTIR-spectra were recorded on

a Bruker FT-IR Spektrometer Tensor II with Platinum-ATR-unit with

diamondcrystal.

Synthesis of n-pentylisocyanate and n-heptylisocyanate

In a typical experiment: in a 1000 mL 2-neck round-bottom flask with

distillation bridge, 15,45 g of n-alkylamine was dripped into 412,7 g of pure

HDI at 80 °C under formation of white deposit. The mixture was heated to

170°C and pressure was gradually reduced to 50 mbar. After 4 h distillation

was complete. The distillate fraction was redistilled at 175°C under

atmospheric pressure, collecting the pure products as transparent liquid,

analyzed by 1H-NMR and FTIR.

Synthesis of functional isocyanurates

In a typical experiment: in a cylindrical flask equipped with a continuous

Near InfraRed (NIR) probe-head (Bruker NIR Vector 22/N), 150 g of α-ω-

55

linear diisocyanate was heated to 60°C and 0.15 wt% of catalyst solution (1

wt% of N,N-methyl-butyl-pyrrolidiniumhydroxide in isopropanol) was

added, until trimerization started, followed by continuous NIR and

temperature (for 1,4-butyldiisocyanate, N,N-dimethyltrimethylsilylamine

was used as catalyst, at 120°C). After 5-20% reduction of NCO absorption-

intensity in NIR (30-90 min), an equimolar amount, compared to active

catalyst, of dodecylbenzene sulfonic acid was added to quench the reaction

(1,4-butanediol for the reaction of 1,4-butyldiisocyanate). The mixture was

transferred to a dripping funnel and evaporated using a thin-film short-path

evaporator at 150 °C and 1∙10-1 mbar, collecting the resin as highly viscous

brown oil. Reusing the distillate as starting material and supplementing with

freshly distilled diisocyanate, this procedure was repeated 6 times,

combining all resins. The combined resin-fraction was then distilled in a thin-

film short-path evaporator at 220°C and <10-7 mbar, collecting the distillate

as a very slightly greenish transparent viscous oil. The distillate was

evaporated using a thin-film short-path evaporator at 160 °C and <10-7 mbar,

collecting the pure product as resin as a very slightly greenish transparent

viscous oil, characterized by (numbers for HDI-trimer): 1H-NMR (700 MHz,

C6D6): δ = 3.78 (2H, t), 2.54 (2H, t), 1.53 (2H, quin), 1.03-0.92 (6H, m) ppm; 13C-NMR (700 MHz, C6D6): δ = 149.01, 122.92, 42.78, 42.60, 31.07, 27.99,

26.17, 26.16 ppm and gel permeation chromatography (GPC).

Synthesis of non-functional isocyanurates

In a typical experiment: in a 100 mL round-bottom flask, 34,63 g of n-alkyl

isocyanate was heated to 60°C. Dropwise, 2,46 g of catalyst solution (1 wt%

of N,N-methyl-butyl-pyrrolidiniumhydroxide in isopropanol) was added,

until trimerisation started, followed by FTIR and temperature. In 2-4 h the

reaction was run to completion, followed by disappearance of the 2275 cm-

1 NCO-band in FTIR. The mixture was trap-to-trap distilled under reduced

pressure (<10-7 mbar) at 200°C and products were isolated as intermediate

distillate fraction as transparent oils, analyzed by (numbers for n-

hexylisocyanate trimer): 1H-NMR (700 MHz, C6D6): δ = 3.82 (2H, t), 1.63 (2H,

quin), 1.22-1.14 (6H, m), 0.82 (3H, t) ppm; 13C-NMR (700 MHz, C6D6): δ =

56

149.06, 42.98, 31.73, 28.26, 26.72, 22.91, 14.19 ppm and gel permeation

chromatography (GPC).

Synthesis of polyisocyanurate networks

In a typical experiment: for the trifunctional isocyanurate, to 100 g of

isocyanurate precursor, 3.140 g of catalyst solution (0.148 g KOAc, 0.396 g

18-crown-6-ether, 2.596 g diethyleneglycol) was added and mixed in a

Hauschild Engineering DAC 150.1 FVZ Speemixer at 2000 rpm for 60 sec. The

mixture was poured out into 6.3 cm aluminium lids in 4, 8 or 16 g samples

and kept at 180°C for 15min. Materials were collected as rigid transparent

thermosets, showing slight yellowing and in some cases formation of a few

small bubbles. For difunctional isocyanates, mixtures were allowed to react

in a glass vial under nitrogen for 1-2 h before similar pouring and annealing.

Caution: Because of the strongly exothermic character of the trimerization

reaction, care should be taken with dissipation of heat, especially during the

pretrimerisation of difunctional diisocyanates.

Quantum calculations details

All DFT calculations were performed with Gaussian ‘09 software15 using a

B3-LYP hybrid functional and the 6-31+g(d,p) basis set, and have been

carried out using the ultrafine integration grid. The accuracy of the

calculations was improved considering empirical dispersion corrections, as

prescribed by the Grimme’s D3 scheme16 with the Becke-Johnson

damping.17 For all molecules, stable geometries were obtained and

characterized in terms of the relevant geometrical parameters (bonds,

angles, dihedrals and impropers).

Free energy calculations

The free energy of binding was calculated as follows:

∆𝐺𝑏𝑖𝑛𝑑𝑖𝑛𝑔 = 𝐺𝑑𝑖𝑚𝑒𝑟 − 𝐺𝑖𝑠𝑜𝑙𝑎𝑡𝑒𝑑

57

where 𝐺𝑑𝑖𝑚𝑒𝑟 is the Gibbs free energy of a bimolecular complex and

𝐺𝑖𝑠𝑜𝑙𝑎𝑡𝑒𝑑 is twice the Gibbs free energy of an isolated molecule in its lowest

energy configuration. The exact formula for 𝐺 is:

𝐺 = 𝐸0 + 𝐸𝑍𝑃𝐸 + 𝐸𝑣𝑖𝑏 + 𝐸𝑟𝑜𝑡 + 𝑘𝑏𝑇 − 𝑇𝑆

where the different contributions are from electronic energy 𝐸0, the zero-

point energy correction 𝐸𝑍𝑃𝐸 , the vibrational and rotational energies 𝐸𝑣𝑖𝑏

and 𝐸𝑟𝑜𝑡, the thermal contribution 𝑘𝑏𝑇 and the entropic term 𝑇𝑆. All of

these terms were calculated explicitly using frequency calculations. For the

bimolecular complexes, the Basis Set Superposition Error contribution has

been calculated and removed from the final free energy.

Molecular Dynamics Simulations

1) Force field parametrization: A fully flexible all-atom representation of the

molecules was adopted and the GAFF18 force field was used and

parametrized using the data from the quantum calculations. In particular,

bond angles and lengths at equilibrium were slightly tuned to match the

results from DFT. Partial charges on each atom were obtained using the RESP

method.19 To account for charge screening effects, all the coulomb

interactions were scaled by a factor of 2. To check the validity of the force

field, the interaction energies calculated by molecular mechanics were

compared to those calculated by DFT for various configurations of the

bimolecular complexes along a specified reaction coordinate, reported in

Figure S1 (Supporting Information). There is a slight underestimation of the

well depth by molecular mechanics, but the long-range part of the

interaction energy is very well described. As further validation, the

equilibrium density obtained for the HDI trimer liquid at 300 K and 1 Atm,

using molecular dynamics simulations, has been compared to the

experimental density at 300 K and 1 atm. An agreement of up to 1% was

found.

2) MD calculations Setup: MD simulations were performed using the

LAMMPS20 software package. The Velocity Verlet integration method has

been used. A timestep of 1 femtosecond has been used, along with the

58

RESPA algorithm21, using an inner timestep of 0.125 fs to keep the energy

drift under control. A general cut-off of 13 Angstroms for pair forces

calculations was considered. Above that cut-off, long range electrostatic

interactions have been evaluated using the Ewald Particle Mesh (PPPM)

method.22 For each molecular liquid six independent runs were performed.

Preparation of the system consisted of a high temperature (480 K) run of 1

ns to randomize the initial conditions, followed by an equilibration step

towards the target temperature. This was done under constant temperature

and pressure conditions, implemented by the Nosé-Hoover23 thermostat

and Parrinello-Rahman24 barostat. Input files for data production runs were

chosen to match both the equilibrium density and the target temperature.

Production runs consisted of 5 ns runs (5 million steps) under NVE (constant

energy, no thermo- and barostatting) conditions.

Calculation of Viscosity

From the production runs stress tensor components were stored every 10

steps and further post-processed to give the viscosity. Using the Green-

Kubo25 formalism, viscosities were obtained through integration of the

stress autocorrelation function:

𝜂 =𝑉

10𝑘𝑇∫ 𝑑𝑡⟨𝑃𝑥𝑦(𝑡)𝑃𝑥𝑦(0)⟩

𝜏

0

where angle brackets denote an ensemble average. To calculate the integral,

a time window of length 𝜏 was defined and different time origins were used

to minimize the integration error.26 To improve accuracy, averages from

different pressure tensor components were considered and averages over 6

independent runs were performed.

59

3.3 Results and Discussion

Synthesis of Isocyanurates

All syntheses were performed based on previously described methods27,28,

using an N,N-methyl-butyl-pyrrolidiniumhydroxide catalyst, as summarized

in Scheme 3.1. Because even small amounts of impurities can cause

deviations in viscosity measurements, extreme care was taken throughout

with the purification and isolation of isocyanurate products. On top of the

already high sensitivity of isocyanates to undergo various side reactions,

difunctional isocyanates (R = NCO) will inevitably form higher molecular

weight oligomers, because of the similar reactivities of the starting material

and product. The optimal conditions for isolating the functionalized trimers

in pure form were hence obtained by early quenching of the reaction.

Specifically, at about 10% conversion of the NCO groups (determined by

continuous near-infraRed (NIR) spectroscopy) the catalyst was quenched by

addition of an equimolar amount of a suitable catalyst poison. Because

isocyanates tend to exhibit reactivity toward most column-chromatography

media, the generally preferred purification technique is distillation.

Therefore, despite the high molecular weights of the products, workup of

the reaction products was done by means of a multiple-step thin-film

distillation process under very low pressures (<10-7 mbar) using a short-path

evaporator, as described below. In a typical experiment, first all unreacted

monomer was removed by thin-film evaporation at 150 °C and 10-1 mbar.

Then, the targeted trimers were separated from higher molecular weight

oligomers by thin-film distillation at 220 °C and <10-7 mbar, thereby giving

the products as the distillate. Finally, from this distillate, any reformed

monomer and other low-boiling impurities formed during the purification

process were removed by similar thin-film evaporation at 160 °C and <10-7

mbar, yielding the product in pure form as the residue.

During trimerization of the monofunctional isocyanates (R = H) of

the reference series, oligomerization does not occur, so the reactions could

be carried out to completion (as determined followed by the disappearance

of the 2275 cm-1 NCO band in the FTIR spectrum). Subsequently, the reaction

products were purified in a fashion to that described previously, in a

60

multistep trap-to-trap distillation process under very low pressures (<10-7

mbar). The products were collected as the intermediate distillate fractions.

All products were isolated in >99% purity with the exception of the sample

with n = 4 and R = NCO (>98%), as confirmed by 700-MHz 1H NMR and 13C

NMR spectroscopies and gel permeation chromatography (GPC), in most

cases limited only by the detection level of the equipment (data not shown).

Because of the strong emphasis on purity, no yields were determined.

Quantification of the Viscosities of Isocyanurates

For the isolated trimers, viscosities were measured on an Anton Paar

MCR501 rheometer using a 50 mm cone-plate (CP50-2) setup. For

validation, several samples were also measured using a more consuming but

more precise CC27 beaker-cup setup, yielding similar results. The measured

viscosity values are depicted in Figure 3.1 as a function of the carbon chain

length of the isocyanate reactants.

Figure 3.1. Viscosities of model isocyanurates.

61

As can be clearly seen from the figure, the viscosities of all functionalized

samples (R = NCO) were significantly higher than those of the non-

functionalized samples (R = H), while also increasing for higher NCO and

isocyanurate contents (i.e. lower value of n). Within the present series, the

side chains are assumed to be too short to show any macromolecular

behavior and no significant contribution to viscosity is expected from

entanglements. Because the molecular shapes of the two series are

practically identical, differing only in the fact that the linear NCO groups are

slightly more rigid than alkyl chain ends, no significant effects are expected

based on molecular shape either. Hence, any difference between the two

series can be attributed entirely to the presence/absence of the NCO group

or, more specifically, to electronic/chemical interactions involving the NCO

group. However, it should be noted that in monomeric (di)isocyanates such

as HDI, NCO groups are abundantly present as well, although here this does

not lead to increased viscosities. The specific chemical mechanism by which

the presence of the NCO group influences viscosity in these mixtures seems

to be somewhat complicated. To investigate this influence in more detail,

density functional theory (DFT) calculations were performed on the same

series of model compounds.

DFT Calculations of Bimolecular Interaction Potentials

To investigate the underlying mechanism causing the unexpectedly high

viscosities of functional aliphatic isocyanurates, quantum mechanical

calculations and viscosity simulations were carried out. To study the

molecular binding between NCO groups and the carbonyls of adjacent

isocyanurate rings, as well as between adjacent NCO groups, several model

bimolecular assembly systems were simulated using density functional

theory (DFT). For the NCO-to-ring interaction (Figure 3.2a), surprisingly, a

stable configuration was found. A significant free energy of binding of -8.7

kJ/mol was calculated for the NCO-to-ring interaction under standard

conditions (298 K, 1 atm). This is clearly above the thermal energy threshold.

However, the free energy of binding of the NCO-NCO bimolecular complex

(Figure 3.2b) was found to be positive (+23.1 kJ/mol), clearly suggesting the

absence of any relevant interaction between NCO groups themselves.

62

Figure 3.2. Bimolecular complexes used in the free energy calculations. (a)

The NCO-to-ring interaction (b) the NCO-NCO interaction. Possible

intermolecular binding sites involving Oxygen (large spheres) and Hydrogen

(small spheres) atoms are highlighted. Corresponding distances in (a) are: a

= 2.51 Å, b = 2.55 Å, c = 2.54 Å, d = 2.38 Å.

To verify the influence of these bimolecular interactions on the viscosity of

the system, MD simulations were carried out for both model series using the

Green-Kubo25 formalism. All calculated viscosity values were in agreement

with experimental results, showing the same trends. From the MD

simulations, it was found that, indeed, the additional electronic interaction

potential between NCO groups and isocyanurate rings gives rise to an

increase in the final viscosity.

Based on the combined experimental and computational results, it

is proposed that the newly identified bimolecular NCO-to-ring interaction

potential is the main cause of the high viscosities observed in functional

aliphatic isocyanurate systems. As a result, increased viscosities should be

expected only when NCO groups and isocyanurate rings are simultaneously

present. This is perfectly supported by the experimental data. Furthermore,

viscosity is expected to increase with increasing relative contents of NCO

63

groups and isocyanurate rings. This was also effectively demonstrated

experimentally. Due to the lack of interaction between adjacent NCO

groups, as indicated by the binding free energy calculations, no additional

contribution to viscosity is expected in systems containing only NCO groups.

This was also indeed observed, noting the very low viscosities of monomeric

(di)isocyanates such as HDI. Given that these interactions involve only the

functional groups, it is expected that the proposed mechanism will be

applicable to a much wider range of materials than just these model systems

based on linear aliphatic diisocyanates. In fact, in any system where both

NCO groups and isocyanurate rings are simultaneously present, this

interaction should be considered.

Synthesis and Characterization of Densely Crosslinked Isocyanurate

Networks

For the most common functional model precursor (HDI, n = 6) initial

attempts to prepare densely crosslinked networks by complete

trimerization were carried out. The resulting thermosets were investigated

in terms of mechanical and thermal properties. A comparison was made

between the use of two different catalysts (potassium acetate (KOAc) and

tin(II) 2-ethylhexanoate (SnOct)) and the difference between using the

difunctional diisocyanate or the trifunctional isocyanurate as the starting

material was investigated. The physical properties of the obtained materials

are reported in Table 3.1. All samples were collected as rigid transparent

thermosets, exhibiting slight yellowing, and in some cases, the formation of

a few small bubbles, depending on the conditions and catalyst (see Picture

3.1).

64

Picture 3.1. Photograph of HDI-based densely crosslinked polyisocyanurate

samples.

For these preliminary results it is noted that the glass transition

temperatures Tg are somewhat lower than those reported in the literature

(140 °C7,8), and that they exhibit some slight variation between samples.

Based on both the literature and our own findings, the differences in

physical properties can most likely be attributed to the degree and quality

of network formation. Because the network density increases rapidly during

the reaction, the probability of trimerizing three NCO groups quickly

decreases and complete network formation is difficult to achieve.

Furthermore, diffusion of the catalyst quickly becomes restricted, limiting

the rate of reaction even further. Based on these preliminary results,

thermosets prepared from NCO-functional isocyanurates as the starting

material seem to show better physical properties than those prepared from

the linear diisocyanates. Future work will be needed to find optimized

processing conditions.

65

Table 3.1. Physical properties of densely cross-linked isocyanurate networks

Onset of

mass loss

[°C]

Decom-

position

[°C]

Mass loss

[%]

Tga

[°C]

E-mod

[GPa]

σb

[MPa]

εb

[%]

150

°C

250

°C

HDI

trimerb

275 440 0.1 0.5 118 2.1c 60c 5.5c

HDIb 95 285 0.6 1.7 108 – – –

HDId 95 280 1.4 2.6 122 – – –

aGlass transition temperature. bCatalyst: potassium acetate (KOAc) in diethylene glycol (DEG). cCovestro internal report. dCatalyst: tin(II) 2-ethylhexanoate (SnOct).

3.4 Conclusions Model aliphatic isocyanurates of varying carbon chain length were

synthesized and isolated in high purities (>99%). The viscosities of these

model systems were accurately quantified, and viscosities were found to be

significantly higher for NCO-functional isocyanurates than for non-

functional isocyanurates. Viscosities also increased with increasing relative

contents of NCO groups and isocyanurate rings. Through DFT calculations, a

newly identified strong intermolecular binding potential was found to be

present between the NCO groups and the isocyanurate-ring carbonyl

groups. Also a less significant interaction between the NCO groups

themselves was identified. At standard conditions (298 K, 1 atm), a binding

free energy of -8.7 kJ/mol was calculated for the NCO-to-ring interaction.

This value is clearly above the thermal energy threshold. Molecular

dynamics simulations further confirmed the influence of this NCO-to-ring

interaction as a contributor to the viscosity of the system.

Based on the combined experimental and computational results, it

is proposed that this NCO-to-ring interaction is the main cause of the high

viscosities generally observed in functional aliphatic isocyanurate systems.

66

Consequently, the presence and relative content of the combination of NCO

groups and isocyanurate rings has a significant effect on increasing the

viscosity of such systems.

Finally, densely crosslinked HDI-based isocyanurate networks were

synthesized. These were found to exhibit interesting but varying mechanical

and thermal properties depending on catalyst use and the degree of

crosslinking achieved.

3.5 Acknowledgements This project has received funding from the European Union’s Horizon 2020

research and innovation program under the Marie Skłodowska-Curie Grant

Agreement no. 642890 (http://thelink-project.eu/) and it was partially

supported by the Portuguese Foundation for Science and Technology (FCT)

in the framework of the Strategic Funding UID/FIS/04650/2013, and under

the project Search-ON2: NORTE-07-0162-FEDER-000086.

67

3.6 References 1. A. Wurtz, Compt. Rend. 1848; 27, 241−243

2. O. Bayer, Angew. Chem. 1947; 59, 257−272

3. E. Delebecq, J.-P. Pascault, B. Boutevin, F. Ganachaud, Chem. Rev.

2013; 113, 80−118

4. 4 I. Yilgor, E. Yilgor, Polym. Rev. 2007; 47, 487–510

5. 5 D. Heift Z. Benkó, H. Grützmacher, A.R. Jupp, J.M. Goicoechea, Chem.

Sci. 2015; 6, 4017–4024

6. 6 S. Okumotoj, S. Yamabe, Comput. Chem. 2001; 22, 316-326

7. 7 T.A.C. Flipsen, R. Steendam, A.J. Pennings, G. Hadziioannou, Adv.

Mater. 1996; 8, 45-48

8. 8 P. Fabbri, S. Mohammad Poor, L. Ferrari, L. Rovati, S. Borsacchi, M.

Geppi, P.P. Lima, L.D. Carlos, Polymer 2014; 55, 488−494

9. 9 M. Moritsugu, A. Sudo, T. Endo, J. Polym. Sci. Part A: Polym Chem.

2013; 51, 2631−2637

10. 10 E. Preis N. Schindler, S. Adrian, U. Scherf, ACS Macro Lett. 2015; 4,

1268−1272

11. 11 D.K. Chattopadhyay, K.V.S.N. Raju, Prog. Polym. Sci. 2007; 32, 352-

418

12. 12 M. Ludewig, J. Weikard, N. Stockel, US 20060205911 A1, 2006

13. 13 H. Renz et al., Prog. Org. Coat. 2001; 43, 32–40.

14. 14 F.U. Richter, J. Pedain, H. Mertes, C.-G. Dieris, EP 0 798 299 B1, 1997,

13-14

15. 15 Gaussian ’09, Rev E01, Gaussian, Inc., Wallingford CT 2009

(www.gaussian.com)

16. 16 S. Grimme, J Antony, S. Ehrlich, H Krieg, J. Chem. Phys. 2010;

132,154104:1-19

17. 17 S. Grimme, S. Ehrlich. L. Goerigk, , J. Comput. Chem. 2011; 32, 1456-

1465

18. 18 J. Wang, R. M. Wolf, J. M. Caldwell, P. A. Kollman, D. A. Case, J. Comp.

Chem. 2004; 25, 1157-1174

19. 19 C. S. Bayly, P. Cieplak, W. Cornell, P. A. Kollman, J. Phys. Chem. 1993;

97, 10269-10280

20. 20 S. Plimpton, J. Comp. Phys. 1995; 117, 1-19

68

21. 21 M. Tuckerman, B. J. Berne, G. J. Martyna, J Chem Phys, 1992; 97,

1990-2001

22. 22 P. Ewald, Ann. Phys. 1921; 369 (3), 253–287

23. 23 W. G. Hoover, Phys. Rev. A 1985; 31 (3),1695–1697

24. 24 M. Parrinello, A. Rahman, J. Appl. Phys. 1981; 52, 7182-7190

25. 25 J. P. Hansen and I. R. McDonald, in Theory of Simple Liquids, Elsevier

Inc., 2006, Chapter 7

26. 26 D. C. Rapaport, in The Art of Molecular Dynamics Simulaton,

Cambridge University Press, 2004, Chapter 5

27. 27 Z. Pusztai, G.Vlád, A. Bodor, I.T. Horváth, H.-J. Laas, R. Halpaap, F.U.

Richter, Angew. Chem. 2006; 118, 113 –116

28. 28 H.-J. Laas, R. Halpaap, J. Pedain, J. Prakt. Chem. 1994; 336, 196–198

69

Chapter 4 - The Influence of Molecular

Interactions on the Viscosity of Aliphatic

Polyisocyanate Precursors for

Polyurethane Coatings

P.J. Driest1,2, V. Lenzi3, L.S.A. Marques3, M.M.D. Ramos3, F.U. Richter1, D.J.

Dijkstra1, D.W. Grijpma2

1Covestro Deutschland AG, CAS-Global R&D, 51373 Leverkusen, Germany

2Technical Medical Centre, Faculty of Science and Technology, Department

of Biomaterials Science and Technology, University of Twente P.O. Box 217,

7500AE Enschede, The Netherlands

3Department/Centre of Physics, University of Minho, Campus de Gualtar,

Braga 4710-057, Portugal

Published in:

ECS Conference Proceedings 2019

Abstract For the production, processing and application of aliphatic isocyanate (NCO)

based materials such as polyurethane coatings and adhesives, the viscosity

of NCO precursors is a key property. However, despite this importance,

surprisingly little is known about the underlying intermolecular interactions

influencing the viscosity these materials. In this work, a combined

experimental and theoretical study is presented investigating the various

intermolecular interactions present in aliphatic polyisocyanates and their

influence on viscosity.

By the use of ultrahigh-vacuum (<10-7mbar) short-path thin-film

distillation, series of functional and non-functional isocyanurate model

compounds were isolated in very high purity. The viscosities of these model

compounds were accurately determined and reported. It was shown that

intermolecular electronic effects involving the NCO group have a significant

influence on the viscosity of the liquid. In consequence, three different

intermolecular interactions involving NCO groups and isocyanurate rings

were proposed: NCO-NCO interactions, Ring-Ring interactions, and NCO-

Ring interactions.

To investigate these intermolecular interactions in more detail, ab

initio calculations were performed on representative small molecular model

compounds. Based on the results of the ab initio calculations, larger scale

molecular dynamics simulations could then be performed. From these

simulations, the contribution of each of the intermolecular interactions to

the viscosity of the liquid could be identified. It was shown that multiple

cooperative NCO-Ring interactions lead to the highest intermolecular

interaction energies, and consequently also to the highest viscosities.

Finally, using the Green-Kubo formalism and the recently developed GAFF-

IC forcefield, the viscosities of the model compounds used in the

experimental study were computationally predicted and turned out to be in

excellent agreement with the experimental data. Thereby, a complete

model is provided, explaining the previously poorly understood trends

regarding the viscosity of isocyanurate-based polyisocyanates.

72

4.1 Introduction Here we present a summary of the recent advancements in the study of the

viscosity of aliphatic polyisocyanates, as well as the various underlying

intermolecular interactions affecting it. A combined theoretical and

experimental approach is taken, investigating in detail the structure-

property relations of aliphatic polyisocyanates. Aliphatic polyisocyanates

form an important class of raw materials for the coatings, adhesives and

sealants industry, with a total annual production of ca. 242.000 tons.1,2 For

the production, the processing and especially the application of aliphatic

polyisocyanates, the viscosity of the liquid precursors is a key parameter.3–5

Surprisingly, however, very little is known about the underlying chemical

principles that dictate the viscosity of aliphatic polyisocyanate liquids.

Although some (mostly empirical) know-how exists within the industry, little

is reported in the scientific literature on the chemical principles and specific

(inter)molecular interactions playing a role.6 One of the things that is

specifically poorly understood, is why isocyanurate-based polyisocyanates

(also referred to as “trimers”) systematically seem to have a significantly

higher viscosity than similar oligomeric polyisocyanates based on other

NCO-derivatives, such as uretdione, allophanate and/or

iminooxadiazinedione (the latter also being referred to as “asymmetrical

trimer”).7 However, from a materials point of view, particularly the trimer-

based oligomers are of high interest and generally provide the best material

properties when it comes to e.g. heat-resistance, chemical resistance and/or

mechanical resistance of the cured coating or adhesive.8 Therefore, a more

fundamental understanding of the intermolecular interactions found in

aliphatic polyisocyanates in general, and in trimer-based polyisocyanates

specifically, and how these affect the viscosity of the liquid, would be very

useful.

Various intermolecular interactions in aliphatic polyisocyanates have been

identified and investigated in detail, together with their influence on

viscosity.9,10 For this purpose, a series of isocyanurate model compounds

with varying carbon chain lengths was designed, based on and including the

NCO-functional hexamethylenediisocyanate (HDI) trimer. To allow for

73

quantitative comparison within and between series, the model compounds

only concern the pure trimer fractions, i.e. the structures for which m = 1 as

drawn in Figure 4.1.

Figure 4.1, top. The trimerization of diisocyanates under the formation of one

or more isocyanurate ring(s).

Figure 4.1, bottom. A sketch representing the composition of a hypothetical

crude polyisocyanate reaction mixture. The trimerization reaction is

terminated prematurely by neutralizing the trimerization catalyst, after

which the unreacted monomer fraction is completely removed by thin-film

short-path evaporation. The specific composition of the resulting

“Polyisocyanate Mixture” depends on the conversion and is typically

reported as the average NCO% of the mixture. A.u. stands for arbitrary unit.

As a comparison, the equivalent series of non-functional tris(n-

alkyl)isocyanurates (i.e. the trimers of the respective monoisocyanates), as

well as the monomeric mono- and diisocyanates, were also included in the

study. In this way it was possible to isolate and investigate separately all the

74

relevant intermolecular interactions. A distinction was made between

isocyanate-isocyanate (NCO-NCO) interactions (present in the monomeric

mono- and diisocyanates), isocyanurate-isocyanurate (Ring-Ring)

interactions (present in the non-functional tris(n-alkyl)isocyanurates), and

isocyanate-isocyanurate (NCO-Ring) interactions (together with the first

two, present in the NCO-functional isocyanurates).

The first challenge, experimentally, is to isolate the model compounds in

high purity. This is particularly difficult for the functional isocyanurate model

compounds, for reasons sketched out in the bottom part of Figure 4.1.

Because during the trimerization reaction both the starting material and the

reaction product(s) contain equally reactive NCO groups, inevitably a

distribution is formed of trimers and increasing amounts of higher order

oligomers. Therefore, the reaction needs to be terminated prematurely by

neutralizing the trimerization catalyst, after which the unreacted monomer

fraction can be distilled off by thin-film short-path evaporation. The resulting

“Polyisocyanate Mixture” consists of a distribution of oligomers, the specific

composition depending on the degree of conversion during trimerization. In

order to isolate the pure trimer fractions from these mixtures of oligomers,

ultrahigh-vacuum thin-film short-path distillation was used to distil off each

trimer fraction as the distillate, followed by further purification.6 Using the

pure trimer fractions, the viscosities of the model compounds were

accurately determined. By comparing the viscosities of the functional

isocyanurates to those of the respective non-functional isocyanurates and

monomeric isocyanates, it was possible to obtain an estimation of the

contribution of each specific intermolecular interaction to viscosity.

To quantify more accurately the contribution of each of the various

intermolecular interactions, Density Functional Theory (DFT) and Symmetry-

Adapted Perturbation Theory (SAPT) calculations were performed. To keep

computational costs reasonable and isolate the bare interactions, smaller

representative model compounds were chosen for this part of the study.

Using the results from the quantum mechanical calculations as a starting

point, larger scale Molecular Dynamics (MD) simulations could then be

performed on the original series of model compounds, using the recently

75

developed GAFF-IC forcefield.10 Thereby, a complete overview is provided of

the various intermolecular interactions at play. Finally, the viscosities of the

model compounds were predicted using the Green-Kubo formalism and

compared to viscosities that were experimentally obtained.

4.2 Experimental Details

All experimental and computational details regarding this work are

described in the literature. For the details regarding the synthesis,

purification and characterization of the model compounds, the reader is

directed here.6 For the computational details regarding the DFT and ab initio

interaction energy calculations, the liquid state potential of mean force

(PMF) calculations and the pair correlation functions, the reader is directed

here.9 For the computational details regarding the modified GAFF forcefield,

GAFF-IC, and the resulting predictions of macroscopic physical properties,

including viscosity, the reader is directed here.10

76

4.3 Results and Discussion

Synthesis, Purification and Characterization of Isocyanate / Isocyanurate

Model Compounds

As previously described, we recently succeeded in isolating a series of pure

NCO-functional isocyanurate model compounds in high purity.6 First,

polyisocyanate mixtures of varying carbon chain lengths were synthesized

by the partial trimerization of the respective diisocyanate, followed by

removal of the unreacted monomer fraction by thin-film short-path

evaporation. Next, by using an ultrahigh-vacuum (<10-7mbar) thin-film

short-path evaporator operated at 220°C as the key step, it was possible to

distill off the pure trimer fractions from the respective polyisocyanate

mixtures as the distillate. Any reformed monomer or other low molecular

species that may have formed during the harsh conditions of the distillation

process were subsequently removed under <10-7mbar and at 160°C, yielding

the pure NCO-functional trimer fractions. As a comparison, the

corresponding non-functional tris(n-alkyl)isocyanurates (i.e. the trimers of

the corresponding monoisocyanates) were also synthesized and isolated.

Although no higher order oligomers are formed during the trimerization of

monoisocyanates, the isolation process was performed under the same

conditions as for the functional isocyanurates. All model compounds were

isolated in >99% purity (with the exception of the functional isocyanurate

with n=4 (>98%)), as confirmed by 1H-NMR, 13C-NMR and Size Exclusion

Chromatography (SEC), in most cases limited only by the detection level of

the equipment.

77

Figure 4.2. The viscosities at 23°C of the high-purity functional (R=NCO) and

non-functional (R=H) isocyanurate model compounds (i.e. the pure trimers),

with carbon chain lengths varying from 4 to 7. The R=NCO; n=6 point

corresponds to the HDI trimer.6

Next, the viscosities of all model compounds were accurately determined

using an Anton Paar MCR501 rheometer and a 50mm cone-plate setup. The

measured viscosities of all model compounds are reported in Figure 4.2. As

can be clearly seen from the figure, the viscosities of the functional

isocyanurates are systematically higher than those of the non-functional

isocyanurates. Considering that all molecules in this study are of comparable

molecular weight, it is not expected that differences in molecular weight

(alone) are the cause for such large differences in viscosity. Furthermore,

counter-intuitively, the viscosity increases with decreasing molecular weight

within the functional isocyanurate series. Likely, intermolecular non-

covalent effects are playing a role as well, apparently dictated by the

presence or absence of NCO groups. However, in the case of the monomeric

isocyanates, very low viscosities are found despite the abundant presence

of NCO groups. In order to gain more insight into the specific non-covalent

interactions involving the NCO groups and isocyanurate rings, quantum

78

mechanical calculations on NCO group- and isocyanurate-ring-containing

model compounds were performed.

Interaction Energy Calculations of Various Dimer Complexes of Isocyanate

/ Isocyanurate Model Compounds

In order to identify and quantify the various intermolecular interactions

existing between NCO groups and isocyanurate rings, a series of simple,

small molecular model compounds containing at least one NCO group,

isocyanurate ring, or both, was evaluated. The molecules used in this

computational study were methyl isocyanate (MIC), ethyl isocyanate (EIC),

ethylene diisocyanate (EDI), propyl isocyanate (PIC), propylene diisocyanate

(GDI; the acronym chosen with regard to the glutaryl backbone), hexyl

isocyanate (HIC), hexamethylene diisocyanate (HDI), trimethyl isocyanurate

(3MIC), trimethyl iminooxadiazinedione (A3MIC; the “asymmetrical trimer”

of methyl isocyanate), triethyl isocyanurate (3EIC) and tris(3-

isocyanatopropyl) isocyanurate (3GDI; the symmetrical trimer of GDI). To

analyze separately the different types of intermolecular interactions, various

bimolecular (dimer) complexes of these model compounds were formed.

The composition of the various dimer complexes, together with their

calculated intermolecular interaction energies by two computational

methods are reported in Table 4.1.9 The dimer complexes are organized into

three categories, based on the type of intermolecular interaction that is

present: NCO-NCO, NCO-Ring or Ring-Ring.

79

Table 4.1: Interaction energies of model dimer complexes, calculated at the

B3LYP-D3BJ level of theory and SAPT2+ methods.9 In some cases, more than

one stable configuration is possible. Three representative dimer complexes

(underlined in the first column) are shown on the right for each category.

Dimer ID

#

Interaction

type

EDFT

[kcal/mol]

ESAPT2+

[kcal/mol]

Representative

Dimer Complex

MIC-

MIC 1 NCO-NCO −5.81 −5.89

EIC-EIC 2 −6.06 −6.47

EDI-EDI 3 −6.68 −7.17

PIC-PIC 4 −6.26 −6.78

PIC-GDI 5 −7.18 −7.61

GDI-GDI 6 −7.85 −8.21

HIC-HIC 7 −6.37 −6.98

HIC-HDI 8 −6.68 −5.84

HDI-HDI 9 −6.92 −7.45

MIC-

3MIC 10 NCO-Ring −7.08 −7.55

3MIC-

EIC 11 −7.38 −8.11

3MIC-

PIC 12 −7.71 −8.46

EIC-

3EIC 13 −7.02 −7.91

3GDI-

3GDI (a) 14 −13.49 n/a

3GDI-

3GDI (b) 15 −17.58 n/a

80

Dimer ID

#

Interaction

type

EDFT

[kcal/mol]

ESAPT2+

[kcal/mol]

Representative

Dimer Complex

3MIC-

3MIC 16 Ring-Ring −13.75 −14.18

A3MIC-

A3MIC

(a)

17 −11.64 −13.13

A3MIC-

A3MIC

(b)

18 −12.36 −13.60

A3MIC-

3MIC 19 −13.16 −14.40

3EIC-

3EIC (a) 20 −12.13 −13.72

3EIC-

3EIC (b) 21 −13.08 −15.31

3EIC-

3EIC (c) 22 −10.02 −10.84

It was found that the intermolecular interaction energies are the weakest

for NCO-NCO interactions, ranging from -5.81 kcal/mol to -7.85 kcal/mol.

The various Ring-Ring interactions are clearly stronger, showing significant

interaction energies of up to -13.75 kcal/mol. As a comparison, this is ca.

three times higher than the calculated intermolecular interaction energy of

a hydrogen bond between two water molecules in vacuum.11 The NCO-Ring

interactions are more elusive in nature. Interaction energies in the range of

-7 to -8 kcal/mol were found when a single interaction is present, i.e. on the

same order of magnitude as the NCO-NCO interactions. However,

interaction energies increasing up to -17.58 kcal/mol were found when

multiple, cooperative NCO-Ring interactions are possible, such as between

a pair of (functional) GDI-trimers: 3GDI-3GDI. Although it was not possible

to obtain SAPT results for this specific dimer complex, based on DFT

calculations it was possible to quantify the cooperative effect of the multiple

81

NCO-Ring interactions more accurately. As shown in Figure 4.3, a 3GDI-3GDI

dimer complex was modified by sequentially replacing NCO groups with

hydrogen atoms, whose positions were then relaxed while keeping all other

atoms fixed. As reported in Figure 4.3, the removal of an NCO group taking

part in an NCO-Ring binding (i.e. NCO groups labeled 1 and 3) leads to a

drastic reduction in the total interaction energy of 6.46 kcal/mol and 5.72

kcal/mol, respectively. The removal of a more distant NCO group on the

other hand, such as the one labeled 2 in Figure 4.3, reduces the total

interaction energy by only 1.52 kcal/mol. Finally, to confirm that the NCO-

Ring interactions are indeed the main cause for the high total interaction

energies, the case is shown in which the three outermost NCO groups,

labeled 3* in Figure 4.3, are replaced by hydrogen atoms. In this case, the

total interaction energy is only reduced by a mere 1.07 kcal/mol. This clearly

shows that the NCO groups taking part in NCO-Ring interactions are

responsible for the high interaction energies.

Figure 4.3. Interaction energies of a 3GDI-3GDI dimer complex as a function

of the replacement of specific NCO groups by hydrogen atoms. Full (red)

circles highlight potential NCO-Ring interactions, while dashed (black) circles

indicate the peripheral NCO groups. On the left, a representation of the

dimer complex is shown, along with the NCO groups being substituted. The

3* point represents the dimer complex in which only the three outermost

NCO groups, labelled 3*, are substituted.9

82

Towards the Liquid State: Molecular Dynamics (MD) Simulations of

Isocyanate / Isocyanurate Molecular Liquids

To make the step from model compounds in the vacuum to more realistic

conditions in which isocyanates are commonly processed, i.e. the liquid

state, larger scale molecular dynamics simulations were performed. For this,

the recently developed modified GAFF forcefield, GAFF-IC, was used,

optimized for accurately predicting macroscopic physical properties of

aliphatic (poly)isocyanates.10 To obtain estimates of the binding free

energies in the liquid state, the Potential of Mean Force (PMF) curves of

three selected dimer complexes, representative of the three types of

interactions, were calculated and are shown in Figure 4.4.8

Methylisocyanate (MIC) was chosen as the solvent in this case. In agreement

with the results obtained in vacuum, the NCO-NCO and single NCO-Ring free

energies are the smallest, with local minima less than -0.2 kcal/mol. The

3MIC-3MIC dimer complex on the other hand, capable of forming the

stronger Ring-Ring interactions, shows a strong free energy profile. The most

important features are a deep minimum around 3.6 Å of -0.81 kcal/mol, and

two shallower local minima at 5.4 Å and 8.2 Å of −0.22 kcal/mol and −0.18

kcal/mol, respectively. In other words, systems containing isocyanurate

rings can form strongly bound pairs at short intermolecular distances, but

only moderately bound pairs at longer intermolecular distances.

Unfortunately, due to the high conformational complexity, it was not

possible to obtain PMF curves for the 3GDI-3GDI dimer complex to

investigate the effect of multiple, cooperative NCO-Ring interactions.

83

Figure 4.4: PMF curves obtained at 298 K and 1 atm, in a pure MIC liquid

solution, for three different model compound dimer complexes, representing

the three possible contact types: MIC-MIC representing NCO-NCO

interactions, MIC-3MIC representing single NCO-Ring interactions and 3MIC-

3MIC representing Ring-Ring interactions.9 Due to the high conformational

complexity, it was not possible to obtain PMF curves for the 3GDI-3GDI dimer

complex.

Extending the model one step further towards more realistic liquid

(poly)isocyanate systems, the next step was to perform molecular dynamics

simulations on the larger functional and non-functional isocyanurate model

compounds representing real (poly)isocyanate liquids. Now, the same series

of model isocyanurates as presented in the experimental study were used.

First, the intermolecular ring-ring pair correlation functions g(r) were

calculated both for the functional and non-functional isocyanurates, which

are reported in Figure 4.5. It is seen that for the non-functional isocyanurate

liquids, only capable of forming Ring-Ring interactions, the pair correlation

functions approximately reflect the PMF curve of the 3MIC-3MIC dimer

84

complex reported in Figure 4.4. The most important features are the two

maxima around 4Å and 6Å, corresponding to molecular pairs involved in

direct Ring-Ring interactions. However, for the functional isocyanurates, the

pair correlation functions look very different, with a significant reduction of

the g(r) at short distances. This means that when NCO groups are present,

direct Ring-Ring contacts are less frequent. Based on the data shown

previously in Table 4.1 and Figure 4.3, it is expected that this reduction in

Ring-Ring interactions can be attributed to NCO groups undergoing NCO-

Ring interactions, thereby hindering the rings for further contacts.

Next, the ring-ring pair correlation functions were decomposed for

two representative isocyanurates (the functional and non-functional

isocyanurates for which n=5) by calculating, for each molecular pair, the

relative tilt angle of the two rings. The angular decomposition is shown in

the bottom part of Figure 4.5. For the non-functional isocyanurate, it was

found that the rings have a tendency to orient in a parallel fashion, especially

at short intermolecular distances, optimizing the orientation for direct Ring-

Ring interactions. The functional isocyanurate on the other hand, has a much

less pronounced orientation profile, with fewer rings found in a parallel

orientation. The chain length of the alkylene side-chains does not seem to

have a very significant effect on the pair correlation functions.

85

Figure 4.5. Intermolecular ring-ring pair correlation functions g(r) and an

angular decomposition g(r,θ) for the case of n=5 as a function of the ring-

ring pair distance r and tilt angle θ for non-functional (left) and functional

(right) isocyanurates.9

86

Testing the Model: Predicting the Viscosities of Isocyanate / Isocyanurate

Model Compounds

Using the recently developed modified GAFF forcefield, GAFF-IC, optimized

for accurately predicting macroscopic physical properties of aliphatic

(poly)isocyanates,10 the effect of multiple, cooperative NCO-Ring

interactions on viscosity was investigated. Due to the observed high

viscosities at 23 ºC, which prevent any meaningful calculation with the

Green-Kubo formula, the MD simulations and respective experimental

measurements were performed at 85°C in this case. The resulting predicted

viscosities are reported in Figure 4.6, together with the respective

experimentally obtained values. As seen in the figure, the newly developed

model was perfectly capable of reproducing the experimentally found data

of the model compounds. The key trends, such as the large difference

between functional and non-functional isocyanurates and increasing

viscosity for decreasing molecular weight functional isocyanurates were

confirmed.

Furthermore, the effects of the various intermolecular interactions

described in the previous sections are clearly visible, providing an

explanation for the observed trends. Firstly, in systems only capable of

forming NCO-NCO interactions (i.e. isocyanate monomers), the lowest

intermolecular interaction energies were found, and consequently the

lowest viscosities are expected (data not shown in the figure). Second, in

systems only capable of forming Ring-Ring interactions (i.e. the non-

functional isocyanurates), slightly higher interaction energies were found,

and thus slightly higher viscosities are expected. However, since the Ring-

Ring interactions were shown to not be capable of forming strong

intermolecular interactions over large distances, the contribution of this

type of binding to viscosity is expected to be only moderate. Finally, in

systems capable of forming multiple, cooperative NCO-Ring interactions (i.e.

the functional isocyanurates), the highest interaction energies were found,

and consequently the highest viscosities are expected. Moreover, because

of its cooperative nature, the presence of the NCO-Ring binding is expected

to have the highest contribution to viscosity.

87

Figure 4.6. The viscosities of isocyanurate model compounds. The solid

markers represent the experimentally measured viscosities6, while the open

markers represent the calculated viscosities using the Green-Kubo formalism

and the modified GAFF-IC force field.10

When comparing these expected trends with the actual viscosity values

(both calculated and experimentally measured), it is seen that these are

exactly supported. Even the initially counter-intuitive trend of increasing

viscosity with decreasing molecular weight within the functional

isocyanurate series can be explained by the proposed model of multiple,

cooperative NCO-Ring interactions, being more concentrated in the shorter-

chain isocyanurate liquids. Thereby, a first overview is provided of the

relevant intermolecular interactions in aliphatic polyisocyanates, and how

these affect the viscosity of the liquid.

88

4.4 Conclusions For the investigation of the viscosity of aliphatic isocyanurate-based

polyisocyanates, a combined experimental and theoretical study was

performed. For this purpose, series of functional and non-functional

isocyanurate model compounds were designed. These model compounds

were synthesized and by the use of ultrahigh-vacuum (<10-7mbar) short-

path thin-film distillation were isolated in very high purity.6 The viscosities

of these model compounds were accurately determined and reported. It

was shown that intermolecular non-covalent effects involving the NCO

group have a significant influence on the viscosity of the liquid.

Subsequently, three different intermolecular interactions involving NCO

groups and isocyanurate rings were proposed: NCO-NCO interactions, Ring-

Ring interactions, and NCO-Ring interactions.

To investigate these intermolecular interactions in more detail, ab

initio and DFT calculations were performed on representative small

molecular model compounds. It was shown that NCO-NCO interactions are

the weakest, ranging from -5.81 to -7.85 kcal/mol. Ring-Ring interactions are

stronger, showing significant interaction energies of up to -13.75 kcal/mol.

Finally, NCO-Ring interactions were shown to be less predictable in nature,

with interaction energies between -7 to -8 kcal/mol for single interactions,

but with interaction energies increasing up to -17.58 kcal/mol when

multiple, cooperative interactions are possible.

Based on the results of the ab initio and DFT calculations, larger

scale molecular dynamics simulations were performed. From these

simulations, the contribution of each of the intermolecular interactions to

the viscosity of the liquid was identified. Using the Green-Kubo formalism

and the recently developed GAFF-IC force field, the viscosities of the model

compounds used in the experimental study were computationally predicted

and turned out to be in excellent agreement with the experimental data.

Thereby, a first overview is provided of the relevant intermolecular

interactions found in aliphatic polyisocyanates, by which the previously

poorly understood trends regarding the viscosity of isocyanurate-based

polyisocyanates could be explained.

89

4.5 Acknowledgements This project has received funding from the European Union’s Horizon 2020

research and innovation program under the Marie Skłodowska-Curie Grant

Agreement no. 642890 (http://thelink-project.eu/) and it was partially

supported by the Portuguese Foundation for Science and Technology (FCT)

in the framework of the Strategic Funding UID/FIS/04650/2013, by the FCT

grant SFRH/BD/128666/2017 and under the project Search-ON2: NORTE-07-

0162-FEDER-000086.

90

4.6 References 1. Meier-Westhues, U., Danielmeier, K., Kruppa, P. & Squiller, E. P.

Polyurethanes − Coatings, Adhesives, Sealants. (Vincentz Network,

Hannover, 2019).

2. Golling, F. E. et al. Polyurethanes for coatings and

adhesives - chemistry and applications. Polym. Int. (2018).

doi:10.1002/pi.5665

3. Chattopadhyay, D. K. & Raju, K. V. S. N. Structural engineering of

polyurethane coatings for high performance applications. Prog.

Polym. Sci. 32, 352–418 (2007).

4. Ludewig, M., Weikard, J. & Stockel, N. Allophanate-containing

modified polyurethanes (US 2006/0205911 A1). (2006).

5. Renz, H. & Bruchmann, B. Pathways targeting solvent-free PUR

coatings. Prog. Org. Coatings 43, 32–40 (2001).

6. Driest, P. J. et al. Aliphatic isocyanurates and polyisocyanurate

networks. Polym. Adv. Technol. 28, 1299–1304 (2017).

7. Widemann, M. et al. Structure-Property Relations in Oligomers of

Linear Aliphatic Diisocyanates. ACS Sustain. Chem. Eng. 6, 9753–

9759 (2018).

8. Behnken, G., Hecking, A. & Sánchez, B. V. High performance enabled

by nature − New polyurethane crosslinkers with significant bio-

based content. Eur. Coatings J. 46–50 (2016).

9. Lenzi, V., Driest, P. J., Dijkstra, D. J., Ramos, M. M. D. & Marques, L.

S. A. Investigation on the intermolecular interactions in aliphatic

isocyanurate liquids: revealing the importance of dispersion. J. Mol.

Liq. (2019). doi:10.1016/j.molliq.2019.01.165 (in press)

10. Lenzi, V., Driest, P. J., Dijkstra, D. J., Ramos, M. M. D. & Marques, L.

S. A. GAFF-IC: realistic viscosities for isocyanate molecules with a

GAFF-based force field. Mol. Simul. 45, 207–214 (2019).

11. Steiner, T. The Hydrogen Bond in the Solid State. Angew. Chemie Int.

Ed. 41, 48–76 (2002).

91

Chapter 5 - The Trimerization of

Isocyanate-functionalized Prepolymers:

an Effective Method for Synthesizing

Well-defined Polymer Networks

P.J. Driest1,2, D.J. Dijkstra1, D. Stamatialis2, D.W. Grijpma2

1Covestro Deutschland AG, CAS-Global R&D, 51373 Leverkusen, Germany

2Technical Medical Centre, Faculty of Science and Technology, Department

of Biomaterials Science and Technology, University of Twente P.O. Box 217,

7500AE Enschede, The Netherlands

Published in:

Macromol. Rapid Commun. 40 (2019), 1800867

Abstract

For the study of polymer networks,

having access to polymer networks

with a controlled and well-defined

microscopic network structure is of

great importance. However, typically,

such networks are difficult to

synthesize. In this work, a simple,

effective and widely-applicable

method is presented for synthesizing

polymer networks with a well-defined network structure. This is done by the

functionalization of polymeric diols using a diisocyanate, and their

subsequent trimerization.

Using hexamethylene diisocyanate and hydroxyl-group-terminated

poly(ε-caprolactone) and poly(ethylene glycol), it was shown that both

hydrophobic and hydrophilic poly(urethane-isocyanurate) networks with a

well-defined network structure can readily be synthesized. By using in-situ

ATR-FTIR, it was shown that the trimerization of isocyanate endgroups is

clearly the predominant reaction pathway of network formation, supporting

the proposed mechanism and network structure. The resulting networks

possessed excellent mechanical properties, in both the dry and in the wet

state.

94

5.1 Introduction In modern materials chemistry, chemically crosslinked polymer networks

play a significant role. To understand and control the physical properties of

these materials, the relation between the microscopic, molecular structure

of the polymer network and the macroscopic material properties is key.

Therefore the study of these structure-property relations has received much

attention over the years.[1-3] For an accurate study, it is essential to have

access to well-defined polymer networks with a variable network structure

that can be prepared in a controlled manner. However, most of the common

synthesis methods for polymer networks lead to random network structures

which, at a molecular level, are poorly defined.[4,5] Consequently, methods

for synthesizing microscopically well-defined polymer networks are of

special interest,[6-8] but typically involve complex synthesis and/or are

limited to specific systems only.[9,10] Therefore, a simple and widely

applicable method for synthesizing well-defined polymer networks would be

of great use.

Scheme 5.1. The isocyanate reactions discussed: a) reaction with an alcohol

under formation of a urethane bond (in this case uncatalyzed), b) the

trimerization of isocyanates resulting in the formation of an isocyanurate-

ring, c) the mechanism of the trimerization reaction catalyzed by a

nucleophile (Nu-).[12]

In polyurethane chemistry, a well-known means of network formation is by

the trimerization reaction of isocyanates.[11] In this reaction, shown in

95

Schemes 5.1b and 5.1c, three isocyanate-groups undergo a step-wise

addition reaction forming an isocyanurate-ring, typically catalyzed by a

potent nucleophile (Nu-).[12] Since the trimerization of isocyanates occurs

readily and controllably, forming well-defined trifunctional isocyanurate-

rings, this reaction is an ideal candidate for the synthesis of well-defined

polymer networks. Surprisingly, the full potential of this reaction towards

this end has not been exploited yet.

Scheme 5.2. Simplified graphical representation of the discussed synthesis

methods for poly(urethane-isocyanurate) networks: a) trimerization

during/after step-growth polyurethanization (also referred to as the

“prepolymer two-stage technique”, see reference 13) leads to a broad

distribution in molecular weight between crosslinks (Mc); b) 2-Component

polyurethane network formation leads to a poorly defined functionality of

the crosslinking units, as well as to stoichiometry issues; whereas c) the

trimerization of NCO-functionalized prepolymers leads to a narrow

distribution of molecular weight between crosslinks as well as a well-defined

functionality of the crosslinking units of exactly 3.

96

Currently, there are two methods that are commonly adopted to

prepare polyurethane networks on the basis of isocyanate trimerization. In

the first method, shown in Scheme 5.2a, the trimerization reaction is

introduced as an intended side-reaction during polyurethanization.[13] In this

case, a (typically oligomeric) difunctional alcohol is reacted with a slight

molar excess of monomeric diisocyanate (typical molar ratios for NCO:OH lie

between 1.1:1 and 2:1). The remaining isocyanates are then trimerized

(either simultaneously or afterwards), resulting in a poly(urethane-

isocyanurate) network.[14 In such a network, the crosslinking units are well-

defined and trifunctional. However, as a result of the step-growth-character

of the polyurethanization reaction, the molecular weight between crosslinks

(Mc) has a broad molecular-weight-distribution with a dispersity index (Đ =

𝑀𝑤/𝑀𝑛) approaching a value of 2 and is therefore poorly-defined.[15]

In the second method, shown in Scheme 5.2b, a monomeric

diisocyanate is trimerized in advance into a functional polyisocyanate

mixture (sometimes referred to as “trimer” or “hardener”). During this

reaction, both the reactants and the reaction-products contain (equally)

reactive isocyanate-groups. Therefore, this reaction can only be carried out

to low conversion. After removal of the unreacted monomer fraction, the

resulting polyisocyanate mixture consists of a distribution of trimers and

higher-order oligomers, with a corresponding distribution in functionality

that is on average larger than 3.[16] To form the poly(urethane-isocyanurate)

network, this polyisocyanate mixture is reacted with a polymeric diol as a 2-

component system. In the resulting network, the molecular weight between

crosslinks is well-defined by the characteristics of the polymeric diol used.

However, in this case the functionality of the crosslinking units is poorly-

defined, due to the distribution in functionality of the polyisocyanate

mixture. Also, the 2-component character of the network formation step

often leads to stoichiometry issues. In this light, it is worth noting that other

2-component approaches for synthesizing well-defined polyurethane

networks have been described as well.[17,18] These are typically based on

reacting a macromolecular di-/polyol with a multifunctional isocyanate, or a

macromolecular polyisocyanate with a multifunctional alcohol or amine.

Since these networks do not contain isocyanurate-rings, and hence are not

97

poly(uretahane-isocyanurate) networks, further details will not be discussed

in this paper.

Here we propose a new method for poly(urethane-isocyanurate)

network formation that combines the benefits of the two previously

described techniques, see Scheme 5.2c. First, a polymeric diol (with a narrow

molecular-weight-distribution) is functionalized using a large excess of

monomeric diisocyanate. After removal of the unreacted excess of

monomeric diisocyanate, an NCO-functionalized prepolymer with a narrow

molecular-weight-distribution is obtained. This functionalized prepolymer is

then trimerized to form the poly(urethane-isocyanurate) network. In this

case, the molecular weight between crosslinks is expected to be well-

defined (i.e. displaying a narrow molecular-weight-distribution), while at the

same time all the crosslinking units are well-defined and trifunctional.

Furthermore, since the network formation step is based on a single-

component reactive system, no stoichiometry issues arise. A simple and

effective method with a wide applicability (i.e. to any polymeric diol) for

synthesizing well-defined polymer networks is thus provided. It is worth

noting that we chose the term “well-defined” networks deliberately here, as

the terms “model” or “ideal/perfect” networks which are also found in the

literature include addressing the potential formation of trapped

entanglements and/or macrocycles.

In this paper, two well-known polymeric diols (poly(ε-caprolactone)

(PCL) and poly(ethylene glycol) (PEG)) of different molecular weights were

functionalized using hexamethylene diisocyanate (HDI), and used to

respectively prepare well-defined hydrophobic and hydrophilic polymer

networks.

98

5.2 Experimental Details

Materials and methods

Poly(ethylene glycol) (𝑀𝑛 4 and 10 kg/mol) were purchased from Sigma-

Aldrich. Poly(ε-caprolactone) (𝑀𝑛 4 and 8 kg/mol) were kindly supplied by

Perstorp Chemicals GmbH. Hexamethylene diisocyanate was supplied by

Covestro Deutschland AG. All other chemicals were purchased from Sigma-

Aldrich.

All chemical reactions were conducted under a nitrogen

atmosphere using standard Schlenck-line techniques. Differential scanning

calorimetry (DSC) measurements were performed under nitrogen on ca.

10mg samples on a Perkin-Elmer Calorimeter DSC-7 in 2 heating runs from -

100°C to +150°C with a heating rate of 20 K/min. and a cooling rate of 20

K/min. NMR spectra were collected on a Bruker Advance III-700 in C6D6.

FTIR-spectra were recorded on a Bruker FTIR Spectrometer Tensor II with

Platinum-ATR-unit with diamond crystal. Size exclusion chromatography

(SEC) was performed according to DIN 55672-1:2016-03, on 4 PSS SDV

Analytical columns (2x 100Å, 5µm; 2x 1000Å, 5µm) using an Agilent 1100

Series pump and an Agilent 1200 Series UV Detector (230 nm) with

tetrahydrofuran as the elution solvent at 40°C and 1.00 mL/min. Residual

monomeric diisocyanate contents were determined by gas chromatography

(GC) according to DIN EN ISO 10283 on an Agilent Technologies 6890N

system using a 15.0m DB17 column and tetradecane as an internal standard.

Tensile tests were performed in triplicate at room temperature using a Zwick

Z0.5 tensile tester equipped with a 500N load cell at 200mm/min. with a

grip-to-grip separation of 35mm on strips of ca 60 x 4 x 0.5mm cut from a

cured film.

For swelling and solvent uptake measurements, three samples

measuring approximately 10 x 20 x 0.5mm were cut from a cured film and

weighed (denoted 𝑚𝑖). Each sample was then extracted with chloroform for

24h while periodically refreshing the solvent. Next, the sample was dried,

initially in air and subsequently under vacuum until constant weight

(denoted 𝑚𝑒). The gel content was then quantified using:

99

𝑔𝑒𝑙 𝑐𝑜𝑛𝑡𝑒𝑛𝑡 =𝑚𝑒

𝑚𝑖× 100%

Each sample was then swollen in excess chloroform or water for at least 24h,

blotted dry and weighed. Chloroform and water uptake was defined as:

𝑠𝑜𝑙𝑣𝑒𝑛𝑡 𝑢𝑝𝑡𝑎𝑘𝑒 =𝑚𝑠 − 𝑚𝑒

𝑚𝑒× 100%

where 𝑚𝑠 is the weight of an extracted specimen after swelling in the

specified solvent.

Synthesis of NCO-functionalized prepolymers

All polymeric diols were dried prior to use by azeotropic distillation in

toluene, which was subsequently removed under reduced pressure. In a

typical experiment, 0.25g of dibutyl phosphate was added to 100 g of dried

polymeric diol (kept in the melt using a heat gun). This mixture was added

dropwise to HDI at 100°C, using a molar ratio of NCO:OH of 15:1, and

subsequently kept at 100°C for 3h. The resulting mixture was then

transferred to a short-path thin-film evaporator, which was operated at a

reduced pressure of 1∙10-2mbar at 140°C. The functionalized prepolymers

were collected as clear, colorless, viscous resins that crystallized upon

cooling. Products were analyzed by SEC, GC and 1H-NMR (600 MHz, C6D6,

representative data for PEG-4k): δ = 4.60 (2H, brs), 4.26 (4H, t), 3.50 (410H,

m), 2.97 (4H, q), 2.57 (4H, t), 1.12 (4H, t), 1.02 (4H, t), 0.88 (8H, m) ppm.

Synthesis of poly(urethane-isocyanurate) networks by the trimerization of

NCO-functionalized prepolymers

In a typical experiment, 12 droplets of Sn(II)Oct2 (corresponding to ca.

0.75wt% relative to the polymer) were added to 8g of NCO-functionalized

prepolymer at 90°C. After mixing for 30 seconds in a Hauschild Speedmixer

DAC150FVZ, the mixture was brought into a mold consisting of a 0.5mm-

thick polycarbonate frame clamped between two silanized glass plates and

kept at 90°C for 24h. Cured samples were collected as flexible transparent

films, of which PCL-8k-, PEG-4k- and PEG-10k-based films slowly turned

100

opaque upon cooling. The films were analyzed by extraction, swelling,

tensile testing and DSC experiments. From the DSC measurements, melting

enthalpy values were obtained from the area under the melting peak. The

degree of crystallinity was then defined as:

𝑑𝑒𝑔𝑟𝑒𝑒 𝑜𝑓 𝑐𝑟𝑦𝑠𝑡𝑎𝑙𝑙𝑖𝑛𝑖𝑡𝑦 =∆𝐻𝑓

∆𝐻𝑓,𝑐𝑟𝑦𝑠× 100%

where ∆𝐻𝑓 is the melting enthalpy of the sample and ∆𝐻𝑓,𝑐𝑟𝑦𝑠 is the melting

enthalpy of a 100%-crystalline sample, for which values were taken from the

literature as 139.5 J/g for PCL[19] and 197 J/g for PEG[20], respectively. For

ATR-FTIR measurements, the catalyst was added to a 50wt%-solution of

functionalized prepolymer in dichloromethane, after which the solvent was

allowed to evaporate at RT overnight. A small sample was then placed

directly onto the ATR-crystal and kept at 80°C for 3h.

101

5.3 Results and Discussion

Synthesis of NCO-functionalized prepolymers

Since commercial polymeric diols may contain traces of base as received, a

small amount of dibutyl phosphate was always added prior to reaction to

neutralize the reaction mixture. Based on 1H-NMR, it was shown that the

functionalization of the polymeric PCL and PEG diols with HDI succeeded

selectively and quantitatively. Specifically, no allophanate formation was

observed. From the 1H-NMR-spectra, the average molecular weight of the

functional prepolymers could also be calculated. In order to synthesize

functionalized prepolymers with a minimally affected polydispersity index Đ

with respect to the polymeric diol, a large molar excess of diisocyanate of

15:1 was used. Using SEC, it was shown that during the functionalization

with HDI, the polydispersity index Đ of the polymers typically only increased

by about 0.05 compared to the respective polymeric diol. This shows that

the synthesis of functionalized prepolymers with a narrow molecular-

weight-distribution close to 1 is possible; where the final polydispersity is

mostly dependent on the quality of the starting material. The resulting

polydispersity indices of the prepolymers after functionalization are

reported in Table 5.1. Finally, it was shown by GC that the unreacted excess

of monomeric diisocyanate, as well as any traces of unreacted dibutyl

phosphate, were effectively removed by short-path thin-film distillation,

resulting in functionalized prepolymers with a residual HDI content lower

than 0.2wt% in all cases. The characteristics of all functionalized

prepolymers are reported in Table 5.1.

Formation of poly(urethane-isocyanurate) networks by the trimerization

of NCO-functionalized prepolymers

To investigate the proposed mechanism of network formation, the

trimerization reaction was monitored in-situ over time using ATR-FTIR. In

Figure 5.1, the FTIR-spectra of the formation of a PEG-4k network are shown

as an example. It is seen that the absorption peak at 2270cm-1, which can be

attributed to the NCO-stretching vibration, gradually disappears over time

until completely absent. More importantly, an absorption peak

102

simultaneously appears at 1690cm-1. This peak can be attributed to the

absorption of the carbonyl group of the isocyanurate ring. Noting that in

isocyanate chemistry, typically, no free carbonyl absorptions other than that

of the isocyanurate ring appear at wave numbers lower than 1700cm-1,

trimerization of isocyanates is clearly taking place.[21] Furthermore, it should

be noted that the rest of the spectrum remains essentially unchanged

throughout the course of the reaction with regard to the shape and intensity

of the peaks. This also includes the urethane-carbonyl peak at 1721 cm-1.

Although potential side-reactions such as allophanate and/or uretdione

(dimer) formation cannot be completely ruled out, the trimerization of the

isocyanate endgroups is clearly the predominant reaction pathway of

network formation. Based on the proposed mechanism, this results in fully

crosslinked networks with well-defined trifunctional crosslinking units, as

sketched in Scheme 5.2c. It should be noted that the formation of trapped

entanglements (and to a very small extent macrocycle formation) cannot be

ruled out though.

Figure 5.1. ATR-FTIR spectra of the formation of a PEG-4k network by

trimerization. "Start. Mat." shows the spectrum of the NCO-functionalized

prepolymer without catalyst, whereas "x min." shows the spectrum of the

reaction mixture after addition of the catalyst, evaporation of the solvent

and curing at 80°C for x minutes.

Table 5.1. Characterization of the NCO-functionalized prepolymers and their corresponding networks (numbers in parentheses are standard deviations).

Polymer NCO-Functionalized

Prepolymer Network

Mn [kg/ mol]

Đ [-]

HDI Content

[wt%]

Gel Content [wt%]

CHCl3 Uptake [wt%]

Water Uptake [wt%]

Tg

[°C]

Tm,p

[°C]

Emod,dry [MPa]

σb,dry

[MPa]

εb,dry

[%]

Emod,wet [MPa]

σb,wet

[MPa]

εb,wet

[%]

PCL-4k 3.8 1.29 0.05 99.2 (0.2)

408.1 (17.7)

2.0 (0.7)

-53.0 +27.6 6.6

(0.8) 4.1

(0.4) 109 (13)

6.8 (0.5)

2.7 (0.2)

62 (7)

PCL-8k 7.7 1.43 0.17 98.9 (0.3)

507.4 (2.2)

1.0 (0.2)

-55.7 +40.6 136.6 (9.5)

20.9 (4.5)

224 (30)

142.9 (-)a

22.9 (-)a

225 (-)a

PEG-4k 4.0 1.15 0.03 100 (0.0)

598.4 (7.6)

219.7 (0.5)

-47.6 +40.6 187.7 (14.4)

13.6 (2.4)

184 (63)

3.8 (0.1)b

0.9 (0.1)b

32 (3)b

PEG-10k 10.1 1.08 0.05 96.6 (1.5)

1265.9 (155.5)

508.0 (74.3)

-49.4 +55.7 165.2 (0.4)b

30.9 (3.1)b

700 (10)b

1.0 (0.3)

0.6 (0.1)

106 (10)

an=1 bn=2

104

Material properties of well-defined poly(urethane-isocyanurate) networks

The material properties of the synthesized networks are also reported in

Table 5.1. For all networks, high gel contents close to 100% were

determined. Together with the FTIR results discussed before, this shows that

the trimerization reaction was completed to a high degree of conversion,

with fully crosslinked networks as a result. As shown by the presence of a

peak melting temperature (Tm,p) in the DSC experiments, all networks were

semi-crystalline upon cooling. The cured PCL-4k and PCL-8k networks

possessed degrees of crystallinity of 24.9% and 29.4%, respectively, while

the PEG-4k and PEG-10k networks possessed degrees of crystallinity of

39.3% and 54.4%, respectively. It was seen that the peak melting

temperatures and degrees of crystallinity decreased upon functionalization

with HDI, as well as upon crosslinking.[18] This can be attributed to the

introduction of HDI moieties upon functionalization and to a decrease in

freedom of movement of the polymer chains upon covalent crosslinking.

Despite that, the degrees of crystallinity of the networks were still relatively

high, suggesting both a homogeneous network structure and a certain

remaining degree of freedom of movement of the polymer chains and/or of

the crosslinking units within that structure. These suggestions are both

supported by the proposed network structure. As a result of the crystallinity

and high degree of crosslinking, the networks showed excellent mechanical

properties. The PCL-4k and PCL-8k networks possessed E-moduli of 6.6 and

136.6 MPa, stress-at-break values (σb,dry) of 4.1 and 20.9 MPa and strain-at-

break values (εb,dry) of 109% and 224%, respectively (see Table 5.1). The PEG-

4k and PEG-10k networks possessed E-moduli of 187.7 and 165.2 MPa,

stress-at-break values of 13.6 and 30.9 MPa and strain-at-break values of

184% and 700%, respectively.

The water uptake of the networks differed significantly. The

networks based on the hydrophobic PCL polymers barely swelled in water,

whereas the hydrophilic PEG networks strongly swelled in water and can be

classified as hydrogels. The mechanical properties of the PCL networks in the

wet state did not differ much from those in the dry state, due to the low

water uptake of these networks. In the wet state, the PCL-4k and PCL-8k

networks possessed E-moduli of 6.8 and 142.9 MPa, stress-at-break values

105

(σb,wet) of 2.7 and 22.9 MPa and strain-at-break values (εb,wet) of 62% and

225%, respectively. The mechanical properties of the PEG networks on the

other hand were clearly affected by equilibration in water, leading to high

water uptakes of 219.7% for PEG-4k and 508.0% for PEG-10k. Nevertheless,

the tensile properties of the PEG networks could be determined without

difficulty even in the water-swollen state. In the wet state, the PEG-4k and

PEG-10k networks possessed E-moduli of 3.8 and 1.0 MPa, stress-at-break

values of 0.9 and 0.6 MPa and strain-at-break values of 32% and 106%,

respectively. Both in the dry state and in the wet state, the toughness of all

the networks was significantly higher than those of comparable non-

poly(urethane-isocyanurate) based networks, e.g. obtained by a

functionalization with acrylates and subsequent crosslinking under UV

radiation.[22] Likely the improved mechanical properties can be attributed to

the introduction of the urethane groups and/or isocyanurate rings. Also, the

more homogeneous network structure could contribute to improved

mechanical properties. Based on the combination of the high water uptake

and high toughness in the wet state, the hydrophilic materials could find

application in a variety of fields, including tissue engineering, biomedical

devices and/or drug delivery.

5.4 Conclusions A simple, effective and widely-applicable method was developed for

synthesizing polymer networks with a well-defined network structure. Upon

functionalization of hydroxyl-group-terminated PCL and PEG polymers with

hexamethylene diisocyanate, it was shown that well-defined hydrophobic

and hydrophilic networks can readily be synthesized by trimerization of the

prepolymers. The proposed mechanism for network formation was

supported by in-situ ATR-FTIR, extraction and thermal measurements. It was

shown that the resulting networks possessed excellent mechanical

properties in terms of elongation-at-break, stress-at-break and toughness,

in both the dry and in the wet state.

106

5.5 Acknowledgements This project has received funding from the European Union’s Horizon 2020

research and innovation program, under the Marie Skłodowska-Curie Grant

Agreement no. 642890 (http://thelink-project.eu/).

The support of the labs of Dr. Frank Richter and the analytical

departments of Covestro Deutschland AG Leverkusen is greatly

acknowledged.

107

5.6 References

[1] M. Urban, C. Craver, Structure-Property Relations in Polymers,

American Chemical Society, 1993.

[2] T. Le, V. C. Epa, F. Burdent, D. Winkler, Chem. Rev. 2012, 112, 2889-

2919.

[3] J. Mark, B. Erman, Rubberlike Elasticity - A Molecular Primer, Wiley,

New-York, USA 1988.

[4] J. Bastide, L. Leibler, Macromolecules, 1988, 21, 2647-2649.

[5] T. Norisuye, N. Masui, Y. Kida, D. Ikuta, E. Kokufuta, S. Ito, S. Panyukov,

M. Shibayama, Polymer, 2002, 43, 5289–5297.

[6] Y. Akagi, T. Katashima, Y. Katsumoto, K. Fujii, T. Matsunaga, U.-I.

Chung, M. Shibayama, T. Sakai, Macromolecules, 2011, 44, 5817–5821.

[7] J. Mark, R. Rahalkar, J. Sullivan, J. Chem. Phys., 1979, 70, 1794.

[8] M. Malkoch, R. Vestberg, N. Gupta, L. Mespouille, P. Dubois, A. Mason,

J. Hedrick, Q. Liao, C. Frank, K. Kingsbury, C. Hawker, Chem. Comm.,

2006, 26, 2774–2776.

[9] J. Mark, M. Llorente, J. Am. Chem. Soc., 1980, 102 (2), 632-636.

[10] T. Sakai, T. Matsunaga, Y. Yamamoto, C. Ito, R. Yoshida, S. Suzuki, N.

Sasaki, M. Shibayama, U.-I. Chung, Macromolecules, 2008, 41, 5379-

5384.

[11] N. Sasaki, T. Yokoyama, T. Tanaka, J. Polym. Sci. Polym. Chem., 1973,

11, 1765-1779.

[12] E. Delebecq, J.-P. Pascault, B. Boutevin, F. Ganachaud, Chem. Rev.,

2013, 113, 80-118.

[13] M. Spírkova, J. Budinski-Simendic, M. Ilavský, P. Spacek, K. Dusek,

Polym. Bull., 1993, 31, 83-88.

[14] R. Samborska-Skowron, A. Balas, Polym. Adv. Technol., 2002, 13, 653-

662.

[15] P. Flory, J. Am. Chem. Soc., 1936, 58, 1877-1885.

108

[16] P.J. Driest, V. Lenzi, L.S.A. Marques, M.M.D. Ramos, D.J. Dijkstra, F.U.

Richter, D. Stamatialis, D.W. Grijpma, Polym. Adv. Technol., 2017, 28,

1299-1304.

[17] K. Dusek, Chapter 12 in Telechelic Polymers: Synthesis and

Applications, 1989, 289-360.

[18] G. Erdödi, B. Iván, Chem. Mater., 2004, 16, 959-962.

[19] G. Crescenzi, G. Manzini, G. Calzolari, C. Borri, Eur. Polym. Mater.,

1972, 8, 449.

[20] B. Wunderlich, Thermal Analysis, Academic Press, 1990.

[21] A. Mishra, D. Chattopadhyay, B. Sreedhar, K. Raju, Prog. Org. Coat.,

2006, 55, 231-243.

[22] E. Zant, M.J. Bosman, D.W. Grijpma, J. Appl. Biomater. Function.

Mater., 2012, 10, 197-202.

109

Chapter 6 - Tough Combinatorial

Poly(urethane-isocyanurate) Polymer

Networks and Hydrogels Synthesized by

the Trimerization of Mixtures of NCO-

Prepolymers

P.J. Driest1,2, D.J. Dijkstra1, D. Stamatialis2, D.W. Grijpma2

1Covestro Deutschland AG, CAS-Global R&D, 51373 Leverkusen, Germany

2Technical Medical Centre, Faculty of Science and Technology, Department

of Biomaterials Science and Technology, University of Twente P.O. Box 217,

7500AE Enschede, The Netherlands

Resubmitted

Abstract The development of tough hydrogels

is an essential but challenging topic in

biomaterials research that has

received much attention over the

past years. By the combinatorial

synthesis of polymer networks and

hydrogels based on prepolymers with

different properties, new materials

with widely varying characteristics

and unexpected properties may be identified. In this paper, we report on

the properties of combinatorial poly(urethane-isocyanurate) (PUI) type

polymer networks that were synthesized by the trimerization of mixtures of

NCO-functionalized poly(ethylene glycol) (PEG), poly(propylene gylcol)

(PPG), poly(ε-caprolactone) (PCL) and poly(trimethylene carbonate) (PTMC)

prepolymers in solution. The resulting polymer networks showed widely

varying material properties.

Combinatorial PUI networks containing at least one hydrophilic PEG

component showed high water uptakes of >100wt%. The resulting hydrogels

demonstrated elastic moduli of up to 10.1 MPa, ultimate tensile strengths

of up to 9.8 MPa, elongation at break values of up to 624.0% and toughness

values of up to 53.4 MJ·m-3. These values are exceptionally high and show

that combinatorial PUI hydrogels are among the toughest hydrogels

reported in the literature. Also, the simple two-step synthesis and wide

range of suitable starting materials make this synthesis method more

versatile and widely applicable than the existing methods for synthesizing

tough hydrogels.

Finally, it is shown that the presence of a hydrophobic network

component significantly enhances the toughness and tensile strength of a

combinatorial PUI hydrogel in the hydrated state. This enhancement is the

largest when the hydrophobic network component is crystallizable in

nature. It is shown that the PUI hydrogels containing a crystallizable

hydrophobic network component are semi-crystalline in the water-swollen

state.

112

6.1 Introduction During the past decades, hydrogels have played an increasingly important

role in the development of biomaterials.1 The interesting material properties

of hydrogels, including a high water uptake (>100wt%), typical

biocompatibility and physical properties resembling the natural extracellular

matrix, have been exploited in a wide range of biomedical applications, such

as drug-delivery, tissue engineering, injectable fillers and many more.2

However, the poor toughness typically associated with hydrogel materials is

still a drawback and has limited the exploitation of the full potential of

hydrogel materials so far. As a result, promising applications as load-bearing

biomaterials largely remain out of reach. Consequently, the development of

tough hydrogels has received much attention over the past years.3–5

Presently, several approaches exist for synthesizing tough

hydrogels, which include (interpenetrating) double network (DN)

hydrogels3,6,7, combinatorial hydrogels8–10, thermoplastic polyurethane

(TPU) hydrogels11,12, (nano)composite hydrogels13–15, rotaxane-based

hydrogels16 and physical interaction enhanced hydrogels17,18. Of these, the

interpenetrating double network (DN) approach, schematically sketched in

Scheme 6.1a, is presently the most widely-studied technique leading to the

best-performing materials.7,19–25 Although the double network approach has

its advantages in terms of mechanical performance, there are still some

drawbacks associated with the technique. These are mostly found in terms

of practicality and applicability. For example, typically a multi-step

sequential crosslinking-swelling-crosslinking procedure is needed to obtain

the interpenetrating double network structure that is required to achieve

the desired material properties.26 In general, in terms of type and molecular

weight of possible starting materials and the resulting control over the final

material properties, the range of possibilities is considered limited in a DN

approach.

113

Scheme 6.1. Schematic representation of the discussed synthesis methods

used for obtaining tough hydrogels: a) interpenetrating Double Network

(DN) formation by repeated photocrosslinking and swelling steps; b) the

combinatorial photocrosslinking of mixtures of (meth)acrylate-

functionalized prepolymers; c) the combinatorial trimerization of mixtures of

NCO-functionalized prepolymers.

A promising alternative approach for synthesizing tough hydrogels is based

on the combinatorial (photo)crosslinking of mixtures of hydrophobic and

hydrophilic (meth)acrylate-functionalized prepolymers, as sketched in

Scheme 6.1b.8–10 Although these combinatorial hydrogels did not yet

demonstrate the same level of toughness or tensile strength as the best-

performing DN hydrogels, they did match those of various biological

materials, most notably cartilage.10 Meanwhile, the relative simplicity of the

synthesis method, the wide range of possible starting materials and the

range and control over the final material properties are a clear advantage of

a combinatorial synthesis approach.27

114

We recently developed a method to synthesize well-defined

poly(urethane-isocyanurate) (PUI) type polymer networks, that is based on

the trimerization of NCO-functionalized prepolymers.28 Similarly to

photocrosslinked combinatorial networks, it was shown that these PUI type

polymer networks are readily synthesized as well. Also, the synthetic

approach allows for the use of a wide range of starting materials (i.e. any di-

or polyol) and allows for a wide range and control over final material

properties. Interestingly, it was found that the hydrophilic single-network

PUI materials based on poly(ethylene glycol) (PEG) prepolymers already

demonstrated unexpectedly high mechanical strength and toughness in the

hydrated state.28 Therefore, we set out to investigate the properties of

combinatorial PUI networks and hydrogels, synthesized by the trimerization

of mixtures of NCO-prepolymers as illustrated in Scheme 6.1c.

In this paper, the synthesis and characterization of a variety of

combinatorial PUI polymer networks and hydrogels is investigated, using

combinations of several hydrophilic and hydrophobic polymeric diols that

are often used in biomaterials research of different molecular weight as the

starting materials: i.e. poly(ethylene glycol) (PEG), poly(propylene gylcol)

(PPG), poly(ε-caprolactone) (PCL), and poly(trimethylene carbonate)

(PTMC). In the molecular weight range used, PEG is a crystallizable

hydrophilic polymer with a glass transition temperature (Tg) between -24°C

and -37°C and a peak melting temperature (Tm) between 58°C and 66°C.29,30

PCL is a crystallizable hydrophobic polymer with a Tg and Tm of approximately

-60°C and 59°C, respectively, while PPG and PTMC are amorphous

hydrophobic polymers with Tgs of approximately -60°C and -15°C,

respectively.31–33

115

6.2 Experimental Details

Materials and methods

Poly(ethylene glycol)s (𝑀𝑛 4 and 10 kg/mol), stannous octoate (Sn(II)Oct2),

1,6-hexanediol and propylene carbonate were purchased from Sigma-

Aldrich. Poly(ε-caprolactone)s (𝑀𝑛 4 and 8 kg/mol) were kindly supplied by

Perstorp Chemicals GmbH. Trimethylene carbonate (TMC) was kindly

supplied by ForYou Medical Devices Co Ltd. Hexamethylene diisocyanate

(HDI) and poly(propylene glycols) (𝑀𝑛 4 and 8 kg/mol) were supplied by

Covestro Deutschland AG. All other chemicals were purchased from Sigma-

Aldrich.

All reactions were conducted under a nitrogen atmosphere using

standard Schlenck-line techniques. Differential scanning calorimetry (DSC)

measurements were performed under nitrogen on ca. 10 mg samples on a

Perkin-Elmer Calorimeter DSC-7. Samples in the dry state were measured in

2 heating runs from -100°C to +150°C with a heating rate of 20 °C/min and a

cooling rate of 20 °C/min. Samples in the hydrated state were measured in

a high-pressure DSC-pan in 1 heating run from +5°C to +80°C with a heating

rate of 20 °C/min. The glass transition temperature(s) (Tg), peak melting

temperature(s) (Tm,p) and melting enthalpy value(s) (∆𝐻𝑚, obtained from

the area under the melting peak) were determined based on the last heating

run. NMR-spectra were collected on a Bruker Advance III-700 in the specified

solvent. FTIR-spectra were recorded on a Bruker FTIR Spectrometer Tensor

II with Platinum-ATR-unit with diamond crystal. Size exclusion

chromatography (SEC) was performed according to DIN 55672-1:2016-03,

on 4 PSS SDV Analytical columns (2x 100Å, 5µm; 2x 1000Å, 5µm) using an

Agilent 1100 Series pump and an Agilent 1200 Series UV Detector (230 nm)

with tetrahydrofuran as the elution solvent at 40°C and 1.00 mL/min.

Residual monomeric diisocyanate contents were determined by gas

chromatography (GC) according to DIN EN ISO 10283 on an Agilent

Technologies 6890N system using a 15.0m DB17 column and tetradecane as

an internal standard. Tensile tests were performed in triplicate at room

temperature using a Zwick Z0.5 tensile tester equipped with a 500N load cell

at 200mm/min. with a grip-to-grip separation of 35mm on strips of

116

approximately 60 x 4 x 0.5mm cut from a cured film. Elastic moduli and strain

values were derived from the initial grip-to-grip separation and are

therefore indicative only. The toughness values of the materials were

calculated from the work at break value, normalized for the dimensions of

the specimen.

For swelling measurements, three samples measuring

approximately 10 x 20 x 0.5mm were cut from a cured, extracted and dried

film and weighed (denoted 𝑚𝑖). Each sample was then swollen in excess

water for at least 24h, blotted dry and weighed again in the hydrated state

(denoted 𝑚𝑠). The water uptake was defined as:

𝑤𝑎𝑡𝑒𝑟 𝑢𝑝𝑡𝑎𝑘𝑒 =𝑚𝑠 − 𝑚𝑖

𝑚𝑖× 100%

Synthesis of an OH-functional PTMC polymeric diol

A PTMC polymeric diol was synthesized by the previously reported ring-

opening polymerization of trimethylene carbonate, using 1,6-hexandiol as

the initiator and Sn(II)Oct2 as the catalyst.8 Under dry conditions, 101.1g

(0.99 mol) of trimethylene carbonate was weighed into a dried and nitrogen-

secured 250mL 3-neck flask. Next, 2.93g (24.8 mmol) of 1,6-hexanediol was

added and the mixture was heated to 130°C. Next, 55mg of Sn(II)Oct2 was

added and the mixture was kept at 130°C for 40h. The resulting product was

collected as a viscous, transparent, colorless resin and analyzed by SEC and 1H-NMR (600 MHz, CDCl3,): δ = 4.00 (160H, brs), 3.39 (4H, brs), 1.64 (80H,

brs), 1.35 (4H, brs), 1.04 (4H, brs) ppm.

Synthesis of NCO-functionalized prepolymers

All polymeric diols were dried prior to use by azeotropic distillation in

toluene, which was subsequently removed under reduced pressure. In a

typical experiment, 0.25 g of dibutylphosphate was added to 100 g of dried

polymeric diol, which was kept in the melt using a heat gun where

applicable. This mixture was added dropwise to HDI kept at 100°C, using a

molar ratio of NCO:OH of 15:1, and the resulting mixture was subsequently

kept at 100°C for 3h. In the case of PPG polymeric diols, the procedure was

performed at 120°C and kept at 120°C for 6h. The resulting mixture was then

117

transferred dropwise into a cylindrical glass short-path thin-film evaporator

equipped with Teflon wipers with a volume of approximately 500mL, which

was operated at a reduced pressure of 1∙10-2 mbar at 140°C. All

functionalized prepolymers were collected as clear, colorless, viscous resins,

of which the PEG- and PCL-based resins crystallized upon cooling to room

temperature. Products were analyzed by SEC, GC and 1H-NMR (700 MHz,

C6D6, representative data for PEG-4k): δ = 4.60 (2H, brs), 4.26 (4H, t), 3.50

(410H, m), 2.97 (4H, q), 2.56 (4H, t), 1.12 (4H, t), 1.02 (4H, t), 0.88 (8H, m)

ppm.

Synthesis of combinatorial poly(urethane-isocyanurate) networks by the

trimerization of mixtures of NCO-functionalized prepolymers in solution

Combinatorial poly(urethane-isocyanurate) networks were synthesized in

analogy to a previously reported procedure,28 with slight modifications. In a

typical experiment, 4g of propylene carbonate was weighed into a

polypropylene cup. Next, the respective NCO-prepolymers were added in

the melt in the reported weight ratios, always totaling 4g. The mixture was

then kept under nitrogen under continuous stirring, until a macroscopically

homogenous solution was formed. Whenever crystallization was observed,

the mixture was gently heated to 50-60°C until the mixture was liquid again.

The mixture was subsequently degassed by gently applying a reduced

pressure of ca. 1-10mbar until no more bubbles were observed, whereby it

was observed that no residue was built up in the cooling trap of the vacuum

pump. Next, ca. 60 mg of Sn(II)Oct2 (corresponding to ca. 0.75wt% relative

to the solution) was added. After mixing for 15 seconds in a Hauschild

Speedmixer DAC150FVZ, the mixture was brought into a mold consisting of

a 0.5mm-thick polycarbonate frame clamped between two silanized glass

plates measuring 20 x 10cm and kept at 90°C for 48h. Unless reported

otherwise, all reaction mixtures were optically transparent within 15min

after heating to 90°C. Cured samples were collected as flexible, non-sticky

films, which were macroscopically homogeneous, unless reported

otherwise. The films were subsequently extracted by swelling in an excess

of acetone for 12-24h while periodically refreshing the solvent. Typically, 6-

8 films were extracted simultaneously using 3x 500mL acetone. Next, 5L of

118

n-hexane was slowly dripped into the 500mL of acetone over the course of

24h which was allowed to overflow, thereby slowly exchanging the acetone

with n-hexane. During this process, the films slowly contracted. The

resulting films were then dried, first in air and subsequently in vacuum, and

analyzed by FTIR, swelling, tensile testing and DSC experiments as described.

The remainder of each film was then swollen in an excess of water for at

least 24h, and analyzed in the water-swollen state by tensile testing and DSC

experiments.

119

6.3 Results and Discussion

Synthesis of NCO-functionalized prepolymers

All polymeric diols were functionalized using hexamethylene diisocyanate

(HDI), in a procedure analogous to the one described before.28 In the case of

the PPG polymeric diols, the reaction temperature and time were increased

to 120°C and 6h, respectively. This was done because the uncatalyzed

urethanization reaction typically occurs less readily for secondary alcohols

than for primary alcohols. Since as-received commercial polymeric diols may

contain traces of base, working under slightly acidic conditions was always

ensured by adding a small amount of dibutyl phosphate (DBP) to the

reaction mixture prior to reaction. This was done in order to prevent any

undesired side-reactions involving the isocyanates, such as isocyanurate

(trimer), uretdione (dimer) or allophanate formation.

Based on 1H-NMR, it was confirmed that the functionalization of all

polymeric diols with HDI succeeded selectively and quantitatively. No side

products were observed in the 1H-NMR spectra. From the 1H-NMR-spectra,

the average molecular weights of the functionalized prepolymers were

calculated. In order to synthesize functionalized prepolymers with a narrow

molecular weight distribution, a large molar excess of HDI of 15:1 was used.

Using SEC, it was shown that the polydispersity index (Đ = 𝑀𝑤/𝑀𝑛) of the

prepolymers after functionalization was relatively low and was always <1.5.

Finally, it was confirmed by GC that the unreacted excess of monomeric

diisocyanate was effectively removed by short-path thin-film evaporation.

This resulted in functionalized prepolymers with a residual HDI content

lower than 1wt% in all cases. The characteristics of all functionalized

prepolymers are reported in Table 6.1.

120

Table 6.1. Characterization of the NCO-functionalized prepolymers. Mn

represents the molecular weight of the NCO-functionalized prepolymer as

calculated from the respective 1H-NMR spectrum; Đ represents the

polydispersity index of the NCO-functionalized prepolymer; and Residual HDI

Content represents the mass percentage of residual free (unreacted) HDI in

the NCO-functionalized prepolymer after thin-film evaporation.

NCO-Functionalized

Prepolymer

Mn

[kg/ mol]

Đ

[-]

Residual HDI

Content

[wt%]

PEG-4k-diNCO 3.9 1.17 <0.01

PEG-10k-diNCO 9.4 1.20 0.09

PCL-4k-diNCO 3.9 1.30 0.21

PCL-8k-diNCO 7.8 1.43 0.17

PPG-4k-diNCO 4.1 1.30 0.02

PPG-8k-diNCO 8.0 1.13 <0.01

PTMC-4k-diNCO 4.1 1.44 0.98

Synthesis of combinatorial poly(urethane-isocyanurate) networks by the

trimerization of mixtures of NCO-functionalized prepolymers in solution

Various combinatorial poly(urethane-isocyanurate) (PUI) polymer networks

were synthesized by the trimerization of mixtures of NCO-functionalized

prepolymers. The synthesis approach was similar to the one reported

previously for single-network PUIs, with the main difference that here the

crosslinking reaction was performed in solution instead of in bulk.28 This was

done because typically NCO-prepolymers of different nature are not

homogeneously miscible as such. Propylene carbonate was selected as a

high-boiling aprotic inert solvent in which the combinatorial trimerization

reactions could be carried out at a concentration of 50wt%. In most cases,

the 50wt% solutions of combined NCO-functionalized prepolymers in

propylene carbonate were stable, homogeneous and transparent at room

temperature. In some cases, the prepolymer mixture was slightly turbid at

room temperature, indicating a certain degree of phase separation.

121

Typically, these prepolymer solutions turned transparent and homogeneous

within 15min after addition of the catalyst and heating to 90°C and remained

transparent throughout the trimerization reaction. In a few cases involving

PPG prepolymers, macroscopic phase separation into domains with a size of

ca. 0.1 – 1mm was observed during the trimerization reaction. The resulting

polymer films were heterogeneous after crosslinking. These films were not

characterized. Whenever this was the case, it is indicated in the table(s)

where the material properties are reported.

To obtain the networks in the dry state after the trimerization

reaction was completed, the high boiling propylene carbonate was extracted

from the polymer networks using acetone. In order to subsequently de-swell

the polymer films in a controlled manner, the acetone was slowly exchanged

for n-hexane over the course of 24h. Since n-hexane is a non-solvent for all

the network components used in this study, this solvent-exchange-process

resulted in the contraction of the networks. Typically, the films also turned

opaque during this process. This indicates that (micro)phase separation of

the different network components took place within the polymer networks

upon de-swelling.

After drying, the completion of the trimerization reaction was

confirmed in all cases by FTIR. This was done by confirming the

disappearance of the absorption peak in the range of 2270cm-1 associated

with the NCO groups. Simultaneously, the appearance of an extra carbonyl

peak in the range of 1690cm-1 associated with the aliphatic isocyanurate

(trimer) was confirmed, in analogy to what was reported previously.28 It was

observed that the polymer films remained opaque upon drying, indicating

phase separation.

Material properties of combinatorial poly(urethane-isocyanurate)

networks in the dry state

The thermal and tensile material properties of the polymer networks in the

dry state are reported in Table 6.2. In the tensile tests, the networks

displayed widely varying mechanical properties, ranging from relatively stiff

to more flexible and ductile elastomeric behavior. A selection of tensile

stress-strain diagrams is shown in Figure 6.1a. The wide range of mechanical

Table 6.2. Characterization of combinatorial PUI networks in the dry state. The prepolymers present in the networks (indicated

by a grey box) are present in equal weight percentages. Numbers in parentheses represent standard deviations.

# Network Component Thermal Tensile

Net

w. n

r.

PEG

-4k

PEG

-10

k

PLC

-4k

PC

L-8

k

PP

G-4

k

PP

G-8

k

PTM

C-4

k

Tg,dry

[°C]

Tm,p,dry

[°C]

ΔHm

[J·g-1]

Emod,dry

[MPa]

σb,dry

[MPa]

εb,dry

[%]

Wtensile,dry

[MJ·m-3]

1 -49.9 55.6 103.6 73.6

(-)a

17.5

(-)a

864.5

(-)a

72.79

(-)a

2 -58.8 39.5 62.4 65.0

(-)a

23.8

(-)a

1305

(-)a

150.4

(-)a

3 -58.4 1) 46.2

2) 50.6

1) 13.2

2) 57.3

184.7

(19.4)

19.5

(0.5)

871.2

(26.9)

103.0

(3.2)

4b -b -b -b -b -b -b -b

5b -b -b -b -b -b -b -b

6 -c - c - c 3.2

(0.8)

1.5

(0.2)

107.2

(23.9)

1.06

(0.28)

7 -55.3 54.5 76.4 124.9

(30.4)

20.1

(1.3)

1025

(40.5)

112.7

(3.3)

8 -58.4 54.6 81.6 221.1

(46.7)d

16.0

(0.7)d

555.4

(22.7)d

60.8

(0.3)d

9b -b -b -b -b -b -b -b

10b -b -b -b -b -b -b -b

11 -45.7 42.6 50.2 34.6

(7.4)

4.1

(0.6)

255.8

(65.0)

8.65

(1.91)

12 -57.8 53.5 53.6 105.4

(29.2)

25.1

(8.5)

586.5

(108.5)

78.2

(24.6)

13b -b -b -b -b -b -b -b

14 -67.1 36.2 24.2 15.6

(2.2)

5.2

(0.1)

339.9

(13.9)

11.03

(0.70)

15 -19.7 37 34 43.7

(2.1)

5.6

(0.4)

198.5

(18.4)

6.95

(0.65)

16 -64.3 41 26.7 21.6

(1.2)

4.7

(0.3)

170.9

(24.1)

5.84

(0.86)

17 -64.7 50.4 35 42.3

(1.8)

13.0

(0.8)

520.3

(36.9)

36.91

(3.23)

18 1) -56.1

2) -18.1 47.3 27.8

16.4

(0.8)

3.2

(0.1)

47.2

(2.6)

1.01

(0.08)

19b -b -b -b -b -b -b -b

20 1) -63.9

2) -19.5 - c - c

1.5

(0.2)

0.5

(0.1)

67.5

(12.8)

0.24

(0.06) an=1 bmacroscopic phase separation observed during crosslinking, resulting in a heterogeneous polymer film cnot observed

dn=2

124

properties is best indicated by the elastic moduli measured in the dry state

(Emod,dry), which covered a range of values from 1.5 MPa to 221.1 MPa. Also

the ultimate tensile strengths (σb,dry) and elongation at break values (εb,dry)

varied widely and ranged from 0.5 MPa to 25.1 MPa and from 47.2% to

1305%, respectively. Starting from only 4 different types of common

polymeric diols, this shows that the synthesis method allows for the

production of a wide variety of materials. The toughness values of the

networks (Wtensile,dry) were calculated from the work at break values and are

reported in Table 6.2 as well. They cover a range from a few MJ·m-3 for the

weaker networks up to 150 MJ·m-3 for the toughest network.

Analogous to the single-network PUIs reported previously, the

thermal properties of the combinatorial networks (reported in Table 6.2) are

dictated by the thermal properties of the respective polymer network

components present in the network. It was found that whenever a

crystallizable prepolymer was present in the combinatorial PUI network, a

corresponding crystallization peak was observed in the DSC curve of the

network. In fact, in the cases where two crystallizable prepolymer

components of a different nature were present in one combinatorial

network (networks 2, 3, 7 and 8), an overlapping double melting peak was

observed in the DSC measurements of the network. In the case of network

3, of which the DSC curves are shown as an example in Figure 6.1b, even two

separate melting- and crystallization peaks were observed. At this point it is

worthwhile to note that all solutions containing PEG and PCL prepolymers

were optically transparent and homogeneous throughout the crosslinking

reaction. It is therefore expected that the prepolymers were

homogeneously mixed at the molecular level during crosslinking. The fact

that separate melting peaks are then observed for the covalently crosslinked

combinatorial polymer network in the dry state is noteworthy. This indicates

that the two separate crystallizable network components of different nature

were able to crystallize within the same combinatorial polymer network.

125

Figure 6.1. Physical characterization of selected combinatorial

poly(urethane-isocyanurate) networks in the dry state: a) tensile stress-

strain curves of several PUI networks in the dry state. The numbers

correspond to the network numbers reported in Table 6.2. To illustrate the

wide range of elastic moduli, the initial part of the stress-strain diagrams is

shown in the inset; b) DSC curves of network 3 (a PEG-4k/PCL-8k

combinatorial PUI network) in the dry state, showing a double melting peak

and a double crystallization peak in both heating runs and in the cooling run,

respectively.

126

Mechanical properties of combinatorial poly(urethane-isocyanurate)

networks in the hydrated state

All combinatorial PUI networks were characterized by swelling in water. The

resulting water uptake values of the materials are reported in Table 6.3. As

expected, all networks containing at least one hydrophilic PEG component

(samples 1W-11W, marked in blue in Table 6.3) strongly swelled in water

and can therefore be classified as hydrogels. These materials all

demonstrated water uptakes >100wt%. Samples that did not contain any

hydrophilic PEG component (samples 12W-20W) were hydrophobic in

nature and therefore barely swelled in water. These networks all

demonstrated water uptakes <2.5wt% and are not considered hydrogels.

Generally, it was found that water uptake of the networks was dictated by

the molecular weight and nature of the prepolymers present in the network.

Networks based on lower molecular weight prepolymers, having a denser

network structure, typically demonstrated lower water uptake values than

their higher molecular weight counterparts.

The mechanical properties of all combinatorial networks in the

hydrated state are reported in Table 6.3. The mechanical properties of the

hydrophobic networks (networks 12W-20W) did not change considerably

compared to the dry state and will not be further discussed. The mechanical

properties of the networks classified as hydrogels (networks 1W-11W) on

the other hand changed considerably compared to the dry state. These

hydrogels demonstrated elastic moduli in the hydrated state (Emod,wet)

ranging from 0.5 MPa to 10.1 MPa. The ultimate tensile strengths of the

hydrogels in the hydrated state (σb,wet) ranged from 0.03 MPa to 7.7 MPa,

increasing up to 9.8 MPa for the best-performing sample of network 2W.

The elongation at break values in the hydrated state (εb,wet) ranged from

82.2% to 624%. The toughness values of the hydrogels in the hydrated state

(Wtensile,wet) were calculated from the work at break values and ranged from

0.02 MJ·m-3 to 33.2 MJ·m-3, increasing up to 53.4 MJ·m-3 for the best-

performing sample of network 2W. These are considered excellent

mechanical properties for hydrogels in the hydrated state for the given

amounts of water content. The tensile stress-strain diagrams of several PUI

hydrogels in the hydrated state are shown in Figure 6.2a.

127

Figure 6.2. Physical characterization of selected combinatorial PUI hydrogels

in the hydrated state; a) tensile stress-strain diagrams of several hydrogels

in the hydrated state. The numbers correspond to the network numbers

reported in Table 6.3. In the inset, a more detailed plot is shown of the curves

of hydrogels 3W, 6W, 8W and 11W; b) The DSC heating curve of hydrogel

2W (a PEG-4k/PCL-4k combinatorial PUI hydrogel) in the hydrated state,

showing a clear melting peak around 42.6°C; c) schematic representation of

the intermolecular interactions present within the hydrophobic phase of a

combinatorial hydrogel. These intermolecular interactions can be semi-

crystalline in nature (e.g. for hydrogel 2W, on the left) or amorphous in

nature (e.g. for hydrogel 6W, on the right).

Notably, ultimate tensile strength values of several MPa and

toughness values in the range of several tens of MJ·m-3 are comparable to

the toughest double network (DN) hydrogels that are well-known from the

literature.20,22,25,34,35 For comparison, the mechanical characteristics of a

selection of tough hydrogels reported in the literature are reported in Table

6.4. Both double network (DN) type hydrogels and combinatorial

photocrosslinked (CP) type hydrogels are included. The highest ultimate

tensile strengths and toughness values of the new PUI hydrogels lie in the

Table 6.3. Characterization of combinatorial PUI networks in the hydrated state. The prepolymers present in the networks

(indicated by a grey box) are present in equal weight percentages. Numbers in parentheses represent standard deviations.

Networks demonstrating a water uptake (WU) >100% are defined as hydrogels and are marked in blue.

# Network Components Swelling Thermal Tensile

Net

w. n

r.

PEG

-4k

PEG

-10

k

PLC

-4k

PC

L-8

k

PP

G-4

k

PP

G-8

k

PTM

C-4

k

WU

[wt%]

Tg,wet

[°C]

Tm,p,wet

[°C]

ΔHm

[J·g-1]

Emod,wet

[MPa]

σb,wet

[MPa]

εb,wet

[%]

Wtensile,wet

[MJ·m-3]

1W 1799.0

(128.4) -a -a -a

0.5

(0.3)

0.03

(0.01)

87.8

(32.0)

0.02

(0.01)

2W 107.5

(0.4) -a 42.6 14.0

10.1

(1.9)

7.7

(1.9)

624.0

(250.7)

33.2

(14.3)

3W 166.3

(1.3) -a 57.0 14.5

3.5

(0.01)

1.0

(0.1)

122.0

(32.5)

1.02

(0.28)

4Wb -b -b -b -b -b -b -b -b

5Wb -b -b -b -b -b -b -b -b

6W 109.4

(0.9) -a -a -a

2.9

(0.4)

1.1

(0.2)

82.2

(23.6)

0.58

(0.22)

7W 154.0

(4.4) -a 44.5 16.9

3.7

(1.7)

4.1

(0.5)

305.5

(51.6)

8.89

(2.23)

8W 287.9

(2.4) -a 56.7 10.8

0.6

(0.4)

0.6

(0.05)

208.1

(35.6)

0.96

(0.19)

9Wb -b -b -b -b -b -b -b -b

10Wb -b -b -b -b -b -b -b -b

11W 138.4

(8.8) -a -a -a

2.7

(0.1)d

0.9

(0.2)d

135.5

(49.9)d

0.86

(0.45)

12W 0.9

(0.3) -a 58.6 87.8

140.1

(113.2)d

25.8

(5.2)d

510.5

(82.3)d

76.3

(13.5)

13Wb -b -b -b -b -b -b -b -b

14W 1.1

(0.1) -a 49.9 32.8

37.9

(2.2)

7.2

(0.4)

295.4

(54.7)

14.7

(2.34)

15W 1.3

(0.1) -a 44.6 36.5

56.9

(2.1)

5.3

(0.3)

190.8

(19.0)

8.11

(0.81)

16W 1.3

(0.1) -a 55.1 36.2

38.2

(2.7)

5.9

(0.6)

162.9

(28.2)

7.57

(1.44)

17W 1.7

(0.4) -a 54.4 42.2

44.6

(5.4)

12.9

(0.6)

428.1

(19.5)

32.1

(1.58)

18W 1.6

(0.1) -a 54.2 35.3

19.5

(1.8)

3.7

(0.3)

55.9

(9.6)

1.42

(0.30)

19Wb -b -b -b -b -b -b -b -b

20W 2.4

(1.1) -a -a -a

1.6

(0.6)d

0.8

(0.1)d

134.7

(40.5)d

0.71

(0.25) anot observed bmacroscopic phase separation observed during crosslinking, resulting in a heterogeneous polymer film cn=1 ; dn=2

130

same range as the PUU-hydrogels reported by Yang et al, which are

presently the toughest hydrogels reported in the literature.12 Considering

the relative simplicity and especially the versatility of the synthesis method

reported here, combinatorial PUI hydrogels are considered a promising

candidate for the future development of tough hydrogels and related

biomaterials.

To investigate the effect of the swelling in water on the thermal

properties of the combinatorial PUI hydrogels, we also performed DSC

experiments in the water-swollen state (see Table 6.3). Interestingly, for all

combinatorial hydrogels containing a crystallizable PCL prepolymer as the

hydrophobic component (hydrogels 2W, 3W, 7W and 8W) a melting peak

was observed in the hydrated state. The melting peaks were all found in the

range of 42°C to 57°C, which corresponds to the typical melting range of PCL

domains within a PUI polymer network.28 Significant melting enthalpies of

10-20 J/g were measured for these hydrogels. As an example, the DSC

heating curve of hydrogel 2W in the hydrated state is shown in Figure 6.2b,

showing a clear semi-crystalline melting peak around 42.6°C.

The observed melting peaks suggest that the segregated

hydrophobic domains observed in the dry state remain intact when the

combinatorial network is swollen in water. The presence of such segregated

hydrophobic domains within a swollen hydrogel network has been reported

in the literature before and has been correlated to a significant

enhancement of the mechanical performance of a hydrogel.17,18,37 The high

toughness values and generally good mechanical properties of the

combinatorial PUI hydrogels in the hydrated state can likely be attributed to

this effect.

131

Table 6.4. Characteristics of several tough hydrogel materials reported in the

literature. Numbers in parentheses represent standard deviations.

Hydrogel Material

Swelling Tensile

Water Uptake

[wt%]

Emod,wet

[MPa]

σb,wet

[MPa]

εb,wet

[%]

Wtensile,wet

[MJ·m-3]

PDLLA-PCL-PEG CPa 132

(59)

0.43

(0.04)

1.08

(0.47)

320

(78)

1.46b

(0.71)

PTMC-PDLLA-PCL-

PEG CPc

181

(2)

1.49

(0.38)

1.81

(0.31)

283

(60) -d

PAMPS-PAAm DNe 900f -d 0.68 75 -d

PDMAEA-Q/PNaSS

DNg

73.9f

(7.4)

7.9

(0.6)

5.1

(0.6)

750

(80)

18.8

(1.9)

Alginate-PAAm DNh 614f 0.029 0.156 2200 -d

SA/PAMAAc JDNi 567f -d 1.8

(0.03)

655.7

(26.7) -d

PUU3-12j 186f 2.5 8.5 1000 45

Cartilagek 400 2.5 1.5 100 -d

aData on a PDLLA-10k/PCL-4k/PEG-10k combinatorial photocrosslinked

network (PDLLA-PCL-PEG CP)10 bRecalculated (and corrected) from original data cData on a PTMC-4k/PDLLA-4k/PCL-4k/PEG-4k combinatorial

photocrosslinked network (PTMC-PDLLA-PCL-PEG CP)9 dData not reported eData on a poly(2-acrylamido-2-methylpropanesulfonic acid)-

poly(acrylamide) double network (PAMPS-PAAm)22 fRecalculated from original data gData on an acryloyloxethyltrimethylammonium chloride-sodium p-

styrenesulfonate (PDMAEA-Q/PNaSS) double network25 hData on a calcium alginate/poly(acrylamide-acrylic acid) double network

(Alginate-PAAm DN)35 iData on a sodium alginate/poly(acrylamide-acrylic acid) joint double

network (SA/PAMAAc JDN)20 jData on a PEG-2k/IPDI Polyurethane-urea physical network (PUU3-12)12 kData on patella cartilage36

132

In the literature, hydrophobic phase segregation within swollen hydrogel

networks has mainly been limited to amorphous materials. In such cases,

the intermolecular interactions between the hydrophobic network

components are random and non-specific in nature, as sketched in Figure

6.2c on the right side. The current PUI hydrogels containing an amorphous

PPG- or PTMC prepolymer as the hydrophobic component (hydrogels 4W,

5W, 6W, 9W, 10W and 11W) are also examples of such hydrogels. The

mechanical performance of these hydrogels is better than that of single-

network hydrogels, but is not exceptional.

In the combinatorial PUI hydrogels containing a crystallizable PCL

prepolymer as the hydrophobic network component however, the

segregated hydrophobic domains are semi-crystalline in nature. In these

hydrogels (2W, 3W, 7W and 8W), the intermolecular interactions between

the hydrophobic network components are well-ordered and cooperative, as

sketched in Figure 6.2c on the left side. The presence of such crystalline

domains is generally associated with a strong increase in the toughness of a

material.38 In fact, the combinatorial hydrogels containing a semi-crystalline

PCL phase demonstrate the highest toughness values in the hydrated state

of all combinatorial PUI hydrogels studied. The high toughness values are

likely attributable to the semi-crystalline nature of the hydrophobic domains

of these hydrogels.

133

6.4 Conclusions Combinatorial poly(urethane-isocyanurate) (PUI) polymer networks can be

effectively synthesized by the trimerization of mixtures of NCO-

functionalized PEG-, PCL-, PPG- and PTMC prepolymers. The resulting

polymer networks demonstrate widely varying material properties in the dry

state, highlighting the wide applicability and versatility of the synthesis

method. Combinatorial networks consisting of two different crystallizable

prepolymers demonstrate a double melting peak in the DSC experiments in

the dry state, showing that two different network components are able to

crystallize simultaneously within one combinatorial PUI network.

Combinatorial PUI networks containing at least one hydrophilic PEG

component demonstrate high water uptake values of >100wt%. The

resulting combinatorial PUI hydrogels exhibit excellent mechanical

characteristics in the water-swollen state. This is indicated by elastic moduli,

ultimate tensile strengths, elongation at break values and toughness values

of up to 10.1 MPa, 9.8 MPa, 624.0% and 53.4 MJ·m-3, respectively. These

values are considered very high and show that combinatorial PUI hydrogels

are among the toughest hydrogels reported in the literature so far.

The presence of a hydrophobic network component in combinatorial PUI

hydrogel networks is shown to enhance the toughness and tensile strength

of the hydrogel in the hydrated state. This increase is the largest when the

hydrophobic network component is crystallizable in nature. The

combinatorial PUI hydrogels containing a crystallizable hydrophobic

network component were semi-crystalline in the water-swollen state.

6.5 Acknowledgements This project has received funding from the European Union’s Horizon 2020

research and innovation program under the Marie Skłodowska-Curie Grant

Agreement no. 642890 (http://thelink-project.eu/).

The support of the labs of Dr. Frank Richter and the analytical departments

of Covestro Deutschland AG Leverkusen is greatly acknowledged.

134

6.6 References

1. Hoffman, A. Hydrogels for biomedical applications. Adv. Drug Deliv. Rev. 64, 18–23 (2012).

2. Annabi, N. et al. 25th Anniversary Article: Rational Design and Applications of Hydrogels in Regenerative Medicine. Adv. Mater. 26, 85–124 (2014).

3. Liu, Y. et al. Recent Developments in Tough Hydrogels for Biomedical Applications. Gels 4, 46 (2018).

4. Creton, C. 50th Anniversary Perspective: Networks and Gels: Soft but Dynamic and Tough. Macromolecules 50, 8297–8316 (2017).

5. Gong, J. P. Materials both Tough and Soft. Science (80-. ). 344, 161–162 (2014).

6. Gong, J. P. Why are double network hydrogels so tough? Soft Matter 6, 2583–2590 (2010).

7. Chen, Q., Chen, H., Zhu, L. & Zheng, J. Fundamentals of double network hydrogels. J. Mater. Chem. B 3, 3654–3676 (2015).

8. Zant, E., Bosman, M. J. & Grijpma, D. W. Combinatorial synthesis of photo-crosslinked biodegradable networks. J. Appl. Biomater. Funct. Mater. 10, 197–202 (2012).

9. Zant, E. & Grijpma, D. W. Synthetic Biodegradable Hydrogels with Excellent Mechanical Properties and Good Cell Adhesion Characteristics Obtained by the Combinatorial Synthesis of Photo-Cross-Linked Networks. Biomacromolecules 17, 1582–1592 (2016).

10. Zant, E. & Grijpma, D. W. Tough biodegradable mixed-macromer networks and hydrogels by photo-crosslinking in solution. Acta Biomater. 31, 80–88 (2016).

11. Güney, A., Gardiner, C., McCormack, A., Malda, J. & Grijpma, D. Thermoplastic PCL-b-PEG-b-PCL and HDI Polyurethanes for Extrusion-Based 3D-Printing of Tough Hydrogels. Bioengineering 5, 99 (2018).

12. Yang, N., Yang, H., Shao, Z. & Guo, M. Ultrastrong and Tough Supramolecular Hydrogels from Multiurea Linkage Segmented Copolymers with Tractable Processablity and Recyclability. Macromol. Rapid Commun. 38, 1–6 (2017).

13. Drozdov, A. D. & Christiansen, J. D. Nanocomposite Gels with Permanent and Transient Junctions under Cyclic Loading. Macromolecules 51, 1462–1473 (2018).

14. Tang, X.-Z. et al. Tough and highly stretchable graphene oxide/polyacrylamide nanocomposite hydrogels. J. Mater. Chem. 22, 14160 (2012).

135

15. Aouada, F. A., de Moura, M. R., Orts, W. J. & Mattoso, L. H. C. Preparation and Characterization of Novel Micro- and Nanocomposite Hydrogels Containing Cellulosic Fibrils. J. Agric. Food Chem. 59, 9433–9442 (2011).

16. Okumura, Y. & Ito, K. The Polyrotaxane Gel : A Topological Gel by. Adv. Mater. 13, 485–487 (2001).

17. Argun, A., Tuncaboylu, D. C., Sari, M., Okay, O. & Sahin, M. Structure optimization of self-healing hydrogels formed via hydrophobic interactions. Polymer (Guildf). 53, 5513–5522 (2012).

18. Zhang, Y., Li, Y. & Liu, W. Dipole-dipole and h-bonding interactions significantly enhance the multifaceted mechanical properties of thermoresponsive shape memory hydrogels. Adv. Funct. Mater. 25, 471–480 (2015).

19. Yang, J., Li, Y., Zhu, L., Qin, G. & Chen, Q. Double Network Hydrogels with Controlled Shape Deformation : A Mini Review. J. Polym. Sci. PART B Polym. Phys. 1351–1362 (2018).

20. Zhang, M., Ren, X., Duan, L. & Gao, G. Joint double-network hydrogels with excellent mechanical performance. Polymer (Guildf). 153, 607–615 (2018).

21. Gong, J. P. Why are double network hydrogels so tough? Soft Matter 6, 2583 (2010).

22. Gong, J. P., Katsuyama, Y., Kurokawa, T. & Osada, Y. Double-Network Hydrogels with Extremely High Mechanical Strength. Adv. Mater. 15, 1155–1158 (2003).

23. Haque, M. A., Kurokawa, T. & Gong, J. P. Super tough double network hydrogels and their application as biomaterials. Polymer (Guildf). 53, 1805–1822 (2012).

24. Mai, T. T., Matsuda, T., Nakajima, T., Gong, J. P. & Urayama, K. Distinctive Characteristics of Internal Fracture in Tough Double Network Hydrogels Revealed by Various Modes of Stretching. Macromolecules 51, 5245–5257 (2018).

25. Luo, F. et al. Oppositely charged polyelectrolytes form tough, self-healing, and rebuildable hydrogels. Adv. Mater. 27, 2722–2727 (2015).

26. Dragan, E. S. Design and applications of interpenetrating polymer network hydrogels. A review. Chem. Eng. J. 243, 572–590 (2014).

27. Zant, E., Blokzijl, M. M. & Grijpma, D. W. A Combinatorial Photocrosslinking Method for the Preparation of Porous Structures with Widely Differing Properties. Macromol. Rapid Commun. 36, 1902–1909 (2015).

28. Driest, P. J., Dijkstra, D. J., Stamatialis, D. & Grijpma, D. W. The

136

Trimerization of Isocyanate-Functionalized Prepolymers : An Effective Method for Synthesizing Well-Defined Polymer Networks. Macromol. Rapid Commun. 40, 1800867 (2019).

29. Read, B. E. Mechanical relaxation in some oxide polymers. Polymer (Guildf). 3, 529–542 (1962).

30. Bailey, F. E. (Frederick E. & Koleske, J. V. Poly(ethylene oxide). (Academic Press, 1976).

31. P.Pêgo, A., Poot, A. A., Grijpma, D. W. & Feijen, J. Copolymers of trimethylene carbonate and ε-caprolactone for porous nerve guides: Synthesis and properties. J. Biomater. Sci. Polym. Ed. 12, 35–53 (2001).

32. Faucher, J. A. The dependence of glass transition temperature on molecular weight for poly(propylene oxide) and poly(butylene oxide). J. Polym. Sci. Part B Polym. Lett. 3, 143–145 (1965).

33. Pêgo, A. P., Grijpma, D. W. & Feijen, J. Enhanced mechanical properties of 1,3-trimethylene carbonate polymers and networks. Polymer (Guildf). 44, 6495–6504 (2003).

34. Li, J., Illeperuma, W. R. K., Suo, Z. & Vlassak, J. J. Hybrid hydrogels with extremely high stiffness and toughness. ACS Macro Lett. 3, 520–523 (2014).

35. Sun, J.-Y. et al. Highly stretchable and tough hydrogels. Nature 489, 133–136 (2012).

36. Chin-Purcell, M. V. & Lewis, J. L. Fracture of Articular Cartilage. J. Biomech. Eng. 118, 545 (1996).

37. Przeradzka, M. A., van Bochove, B., Bor, T. C. & Grijpma, D. W. Phase-separated mixed-macromer hydrogel networks and scaffolds prepared by stereolithography. Polym. Adv. Technol. 28, 1212–1218 (2017).

38. Galeski, A. Strength and toughness of crystalline polymer systems. Prog. Polym. Sci. 28, 1643–1699 (2003).

137

Chapter 7 - Structure-Property Relations

in Semi-crystalline Combinatorial PUI

Hydrogels

P.J. Driest1,2, D.J. Dijkstra1, D. Stamatialis2, D.W. Grijpma2

1Covestro Deutschland AG, CAS-Global R&D, 51373 Leverkusen, Germany

2Technical Medical Centre, Faculty of Science and Technology, Department

of Biomaterials Science and Technology, University of Twente P.O. Box 217,

7500AE Enschede, The Netherlands

To be submitted

Abstract

The development of tough hydrogel

materials for load-bearing

biomedical applications is a

challenging topic that has received

much attention over the past years.

For polymer materials in the dry

state, a correlation between

toughness and crystallinity of the

material is well-known. However,

such a correlation has not been

extensively studied for hydrogel materials in the water-swollen state.

In this work, we aim to get better insights into the relationship

between the crystallinity and toughness of poly(urethane-isocyanurate)

(PUI) hydrogels in the water-swollen state. For this, series of model

combinatorial PUI hydrogels with increasing crystallizable and amorphous

hydrophobic content are synthesized and characterized.

Using melting enthalpy as a measure for crystallinity, it is shown that

the toughness of combinatorial PUI hydrogels in the water-swollen state

strongly increases with increasing crystallinity of the hydrogel in the water-

swollen state. This correlation is an important structure-property-

relationship for the future development of tough hydrogel materials for

load-bearing biomedical applications.

140

7.1 Introduction The development of tough, mechanically resilient hydrogels for load-bearing

biomedical applications has received much attention over the past years.1–3

At present, the most widely adopted strategy for synthesizing tough

hydrogels is by an (interpenetrating) double network (DN) approach, as

sketched in Scheme 7.1a.1,4,5 Although DN hydrogels typically demonstrate

good mechanical performance in the water-swollen state, the ease and

versatility of their synthesis is limited. A cumbersome synthesis procedure is

needed to obtain the interpenetrating double network structure required to

achieve the desired material properties.6 A simpler and more widely

applicable synthesis method for hydrogels with a comparable or better

mechanical performance in the water-swollen state would be of great use.

Scheme 7.1. Methods for synthesizing tough hydrogels; a) a sequential

crosslinking-swelling-crosslinking approach used to synthesize an

interpenetrating double network structure; b) the combinatorial

trimerization of mixtures of NCO-functionalized prepolymers.

141

Recently, a synthesis method was developed by our group to

synthesize combinatorial poly(urethane-isocyanurate) (PUI) type polymer

networks and hydrogels by the trimerization of mixtures of NCO-

functionalized prepolymers in solution (Scheme 7.1b).7 It was shown that

these combinatorial PUI type hydrogels demonstrate excellent mechanical

performance in the water-swollen state, as illustrated by high ultimate

tensile strengths of up to 9.8 MPa and high toughness values of up to 53.4

MJ·m-3. Also, the simple two-step synthesis is applicable to a wide range of

suitable starting materials and it is more versatile than the DN approach

typically used to obtain tough hydrogels.

Our previous work showed that the presence of a hydrophobic

network component in a combinatorial PUI hydrogel significantly increases

the toughness of the hydrogel in the water-swollen state. It was shown that

this increase in toughness is the largest when the hydrophobic network

component is crystallizable in nature. In fact, the toughest PUI hydrogels

were shown to be semi-crystalline in the water-swollen state at room

temperature, suggesting a correlation between the crystallinity and

toughness of the hydrogel in the water-swollen state. Such a correlation is

well-known for polymer materials in the dry state,8 but has not been

extensively studied for hydrogel materials in the water-swollen state.

In this work, we aim to get better insights into the relationship

between crystallinity and toughness of PUI hydrogels in the water-swollen

state. For this, two series of combinatorial PUI hydrogels with increasing

hydrophobic contents are synthesized. In one series, the hydrophobic

component is crystallizable in nature, while in the other series the

hydrophobic component is amorphous in nature. The resulting hydrogels

are characterized for their thermal and mechanical properties, both in the

dry state and in the water-swollen state. The melting enthalpy, obtained

from dynamic scanning calorimetry (DSC) measurements, is used as a

measure for the crystallinity of the hydrogel and is correlated to the

toughness of the hydrogel in the hydrated state.

Poly(ethylene glycol) (PEG), poly(ε-caprolactone) (PCL) and

poly(propylene gylcol) (PPG) with number-average molecular weights (𝑀𝑛)

of 4 kg/mol are used as the polymeric diols. In the molecular weight used,

142

PEG is a crystallizable hydrophilic polymer with a glass transition

temperature (Tg) between -24°C and -37°C and a peak melting temperature

(Tm) between 58°C and 66°C,9,10 PCL is a crystallizable hydrophobic polymer

with a Tg and Tm of approximately -60°C and 59°C, respectively,11 and PPG is

an amorphous hydrophobic polymer with a Tg of approximately -60°C.12

143

7.2 Experimental Details

Materials and methods

Poly(ethylene glycol) (𝑀𝑛 4 kg/mol), stannous octoate (Sn(II)Oct2),

propylene carbonate and toluene were purchased from Sigma-Aldrich.

Poly(ε-caprolactone) (𝑀𝑛 4 kg/mol) was kindly supplied by Perstorp

Chemicals GmbH. Hexamethylene diisocyanate (HDI) and poly(propylene

glycol) (𝑀𝑛 4 kg/mol) were supplied by Covestro Deutschland AG. All other

chemicals were purchased from Sigma-Aldrich.

Differential scanning calorimetry (DSC) measurements were

performed under nitrogen on ca. 10 mg samples on a Perkin-Elmer

Calorimeter DSC-7. Samples in the dry state were measured in 2 heating runs

from -100°C to +150°C with a heating rate of 20 °C/min and a cooling rate of

20 °C/min. Samples in the hydrated state were measured in a high-pressure

DSC-pan in 1 heating run from +5°C to +80°C with a heating rate of 20

°C/min. The glass transition temperature(s) (Tg), peak melting

temperature(s) (Tm,p) and melting enthalpy value(s) (∆𝐻𝑚, obtained from

the area under the melting peak) were determined based on the last heating

run.

Tensile tests were performed in triplicate at room temperature

using a Zwick Z0.5 tensile tester equipped with a 500N load cell at

200mm/min. with a grip-to-grip separation of 35mm on strips of

approximately 60 x 4 x 0.5mm cut from a cured film. Elastic moduli and strain

values were derived from the initial grip-to-grip separation and are

therefore indicative only. The toughness values of the materials (Wb) were

calculated from the work at break value, normalized for the dimensions of

the specimen.

For swelling measurements, three samples measuring

approximately 10 x 20 x 0.5mm were cut from a cured, extracted and dried

film and weighed (denoted 𝑚𝑖). Each sample was then swollen in excess

water for at least 24h, blotted dry and weighed again in the hydrated state

(denoted 𝑚𝑠). The water uptake (WU) was defined as:

144

𝑤𝑎𝑡𝑒𝑟 𝑢𝑝𝑡𝑎𝑘𝑒 (𝑊𝑈) =𝑚𝑠 − 𝑚𝑖

𝑚𝑖× 100%

Synthesis of NCO-functionalized prepolymers

All reactions were conducted under a nitrogen atmosphere using standard

Schlenck-line techniques. All polymeric diols were dried prior to use by

azeotropic distillation in toluene, which was subsequently removed under

reduced pressure. In a typical experiment, 0.25 g of dibutylphosphate was

added to 100 g of dried polymeric diol, which was kept in the melt using a

heat gun where applicable. This mixture was added dropwise to HDI kept at

100°C, using a molar ratio of NCO:OH of 15:1, and the resulting mixture was

subsequently kept at 100°C for 3h. In the case of PPG polymeric diols, the

procedure was performed at 120°C and kept at 120°C for 6h. The resulting

mixture was then transferred dropwise into a cylindrical glass short-path

thin-film evaporator equipped with Teflon wipers with a volume of

approximately 500mL, which was operated at a reduced pressure of 1∙10-2

mbar at 140°C. All functionalized prepolymers were collected as clear,

colorless, viscous resins, of which the PEG- and PCL-based resins crystallized

upon cooling to room temperature. Products were analyzed by SEC, GC and 1H-NMR (700 MHz, C6D6, representative data for PEG-4k): δ = 4.60 (2H, brs),

4.26 (4H, t), 3.50 (410H, m), 2.97 (4H, q), 2.56 (4H, t), 1.12 (4H, t), 1.02 (4H,

t), 0.88 (8H, m) ppm.

Synthesis of combinatorial poly(urethane-isocyanurate) networks by the

trimerization of mixtures of NCO-functionalized prepolymers in solution

Combinatorial poly(urethane-isocyanurate) networks were synthesized

according to a previously reported procedure.7 In a typical experiment, 4g of

solvent (propylene carbonate in the case of PEG/PCL mixtures and toluene

in the case of PEG/PPG mixtures) was weighed into a polypropylene cup.

Next, the respective NCO-prepolymers were added in the melt in the

reported weight ratios, always totaling 4g. The mixture was then kept under

nitrogen under continuous stirring, until a macroscopically homogenous

solution was formed. Whenever crystallization was observed, the mixture

was gently heated to 50-60°C until the mixture was liquid again. The mixture

145

was subsequently degassed by gently applying a reduced pressure of

approx. 10 to 100mbar until no more bubbles were observed, whereby it

was observed that no residue was built up in the cooling trap of the vacuum

pump. Next, ca. 60 mg of Sn(II)Oct2 (corresponding to ca. 0.75wt% relative

to the solution) was added. After mixing for 15 seconds in a Hauschild

Speedmixer DAC150FVZ, the mixture was brought into a mold consisting of

a 0.5mm-thick polycarbonate frame clamped between two silanized glass

plates measuring 20 x 10cm and kept at 90°C for 48h. All reaction mixtures

were optically transparent within 15min after heating to 90°C. Cured

samples were collected as flexible, non-sticky films, which were

macroscopically homogeneous. The films were subsequently extracted by

swelling in an excess of acetone for 12-24h while periodically refreshing the

solvent. Typically, 6-8 films were extracted simultaneously using 3x 500mL

acetone. Next, 5L of n-hexane was slowly dripped into the 500mL of acetone

over the course of 24h which was allowed to overflow, thereby slowly

exchanging the acetone with n-hexane. During this process, the films slowly

contracted. The resulting films were then dried, first in air and subsequently

in vacuum, and analyzed by attenuated total reflection Fourier-transformed

infrared spectroscopy (ATR-FTIR), swelling, tensile testing and DSC

experiments as described. The remainder of each film was then swollen in

an excess of water for at least 24h, and analyzed in the water-swollen state

by tensile testing and DSC experiments.

146

7.3 Results and Discussion

Synthesis of NCO-functionalized prepolymers and PUI networks by the

selective trimerization of NCO-functionalized prepolymers

Polymeric diols were functionalized using hexamethylene diisocyanate (HDI)

and subsequently trimerized as reported previously.7 It was confirmed that

the functionalization of all polymeric diols with HDI succeeded selectively

and quantitatively in all cases. The characteristics of all functionalized

prepolymers are reported previously.7

Combinatorial poly(urethane-isocyanurate) (PUI) polymer networks

were synthesized by the trimerization of mixture of the NCO-functionalized

prepolymers in solution. In this case, to prevent phase separation, toluene

was used as the inert solvent for the PPG-containing prepolymer mixtures

instead of propylene carbonate. All reaction mixtures were transparent and

homogeneous throughout the trimerization reaction. The completion of the

trimerization reaction was confirmed using attenuated total reflection

Fourier-transformed infrared spectroscopy (ATR-FTIR) as reported

previously (data not shown).7

Material properties of combinatorial PUI networks in the dry state

The thermal and tensile mechanical properties of the PUI networks in the

dry state are reported in Table 7.1. Similar to the PUI networks on which we

reported previously, the mechanical properties varied based on the

molecular weight and nature of the prepolymers present in the network.7

The tensile stress-strain diagrams of all PUI networks in the dry state are

shown in Figure 7.1. In the case of PEG/PCL combinatorial PUI networks, all

networks were semi-crystalline in nature. These networks were generally

tough materials, which is consistent with the literature.8 In the case of

PEG/PPG combinatorial networks, a melting peak was only observed in the

dry state for the networks containing PEG. In these networks, a decreasing

melting enthalpy is observed for decreasing PEG content.

Table 7.1. Mechanical and thermal properties in the dry state of combinatorial PUI networks with increasing hydrophobic

content (HC). On the left, the hydrophobic component is crystallizable in nature (PEG/PCL), on the right, the hydrophobic

component is amorphous in nature (PEG/PPG).

PEG/

PCL Tensile Thermal

PEG/

PPG Tensile Thermal

HC

[wt%]

Emod,dry

[MPa]

σb,dry

[MPa]

εb,dry

[%]

Wb,dry

[MJ·m-3]

Tg

[°C]

Tm

[°C]

ΔHm

[J·g-1]

HC

[wt%]

Emod,dry

[MPa]

σb,dry

[MPa]

εb,dry

[%]

Wb,dry

[MJ·m-3]

Tg

[°C]

Tm

[°C]

ΔHm

[J·g-1]

0 208.9

(42.3)

15.4

(3.1)

403.8

(62.0)

42.34

(9.53) -50.3 46.7 98.4 0

208.9

(42.3)

15.4

(3.1)

403.8

(62.0)

42.34

(9.53) -50.3 46.7 98.4

20 85.8

(26.2)

10.4

(1.5)

401.4

(78.2)

28.33

(5.16) -54.7 42.9 76.8 20

83.8

(-)b

11.8

(-)b

291.0

(-)b

26.55

(-)b -50 44.9 90.3

40 55.7

(8.5)

7.5

(0.6)

322.6

(41.9)

16.34

(1.45) -56.6 35.3 62.7 40

19.5

(5.4)a

3.8

(0.6)a

32.0

(3.3)a

0.78

(0.05)

1) -62.5

2) -0.3 42.4 55.1

60 37.4

(9.4)

7.8

(1.3)

327.2

(68.3)

15.00

(2.98) -58.1 30.5 58.0 60

6.2

(2.0)

2.2

(0.5)

54.0

(7.7)

0.71

(0.19)

1) -61.7

2) -22.5 42.2 36.3

80 66.0

(3.5)

11.4

(1.0)

318.2

(22.2)

23.01

(1.69) -57.4 35.3 60.2 80

4.1

(0.2)

0.9

(0.2)

33.4

(5.6)

0.18

(0.06)

1) -61.7

2) -3.2 42 15

100 111.1

(1.0)a

12.5

(1.6)a

221.4

(11.2)a

16.42

(1.40) -53.40 33.6 42.9 100

3.6

(0.4)

0.7

(0.04)

21.6

(2.0)

0.09

(0.01) -61.7 -c -c

an=2 bn=1 cnot observed

148

Figure 7.1. Tensile stress-strain diagrams in the dry state of combinatorial

PUI networks with increasing hydrophobic content (HC); a) the hydrophobic

component is crystallizable in nature (PCL); b) the hydrophobic component is

amorphous in nature (PPG).

149

Material properties of combinatorial PUI networks in the water-swollen

state

The swelling properties of the combinatorial PUI networks and their thermal

and tensile mechanical properties in the water-swollen state are reported in

Table 7.2. The tensile stress-strain diagrams in the water-swollen state are

shown in Figure 7.2. Both for the crystallizable and amorphous PUI

hydrogels, the water uptake (WU) decreased with increasing hydrophobic

content. The water uptake values ranged from 380 wt% for a completely

hydrophilic PEG PUI network to <2wt% for both of the completely

hydrophobic networks.

The tensile mechanical properties of the PUI hydrogels in the water-

swollen state vary and strongly depend on the nature of the hydrophobic

network component. The crystallizable PEG/PCL combinatorial hydrogels

were generally tough and demonstrated ultimate tensile strengths and

toughness values of up to 335 MPa and 23.8 MJ·m-3, respectively. The

amorphous PEG/PPG hydrogels on the other hand were mechanically very

poor and demonstrated ultimate tensile strengths and toughness values <1

MPa and <0.3 MJ·m-3, respectively.

As reported in Table 7.2, the PEG/PCL combinatorial hydrogels

containing more than 40wt% of PCL were semi-crystalline in the water-

swollen state. This is in accordance with what we reported previously.7 The

combinatorial PEG/PCL hydrogels demonstrated increasing melting

enthalpies for increasing PCL content, reaching up to approximately 25 J·g-1

for a 100 wt% PCL network. Melting enthalpy values were generally lower

than the ones we reported previously: the PUI hydrogel containing 60 wt%

PCL demonstrated a melting enthalpy of approximately 2 J·g-1 whereas the

PUI hydrogel containing 50 wt% PCL on which we reported previously

demonstrated a melting enthalpy of approximately 14 J·g-1.7 Finally, it was

found that all hydrogels containing an amorphous hydrophobic network

component were amorphous in the water-swollen state. In other words, no

crystalline phase attributable to PEG domains was observed in any of the

hydrogels in the water-swollen state.

Table 7.2. Mechanical properties in the water-swollen state of semi-crystalline and amorphous combinatorial networks with an

increasing hydrophobic content (HC). On the left, the hydrophobic component is crystallizable in nature (PCL-4k), on the right,

the hydrophobic component is amorphous in nature (PPG-4k).

PEG/

PCL Swelling Tensile Thermal

PEG/

PPG Swelling Tensile Thermal

HC

[wt%]

WU

[wt%]

Emod,wet

[MPa]

σb,wet

[MPa]

εb,wet

[%]

Wb,wet

[MJ·m-3]

Tg

[°C]

Tm

[°C]

ΔHm

[Jg-1]

HC

[wt%]

WU

[wt%]

Emod,wet

[MPa]

σb,wet

[MPa]

εb,wet

[%]

Wb,wet

[MJ·m-3]

Tg

[°C]

Tm

[°C]

ΔHm

[Jg-1]

0 380.5

(3.8)

1.1

(0.5)

0.4

(0.05)

34.6

(5.0)

0.08

(0.02) -a -a -a 0

380.5

(3.8)

1.1

(0.5)

0.4

(0.05)

34.6

(5.0)

0.08

(0.02) -a -a -a

20 220.8

(1.2)

2.5

(0.2)

0.7

(0.1)

63.6

(8.9)

0.29

(0.06)

1) 40.9

2) 64.5 -a -a 20

313.4

(7.5)

1.7

(0.4)

0.2

(0.1)

9.4

(3.5)

0.01

(0.01) -a -a -a

40 116.9

(1.2)

3.9

(0.1)b

1.6

(0.6)b

96.7

(66.9)b

1.11

(0.76)b

1) 25.8

2) 41.2 -a -a 40

176.6

(2.9)

2.0

(0.3)b

0.5

(0.0)b

37.2

(11.6)b

0.12

(0.03)b

1) 32.8

2) 38.1 -a -a

60 51.3

(0.6)

12.0

(0.5)

3.7

(1.5)

158.2

(120.1)

4.86

(3.65) -a 43 2.0 60

120.5

(1.9)

2.5

(0.3)

0.7

(0.1)

56.6

(19.2)

0.27

(0.11)

1) 31.0

2) 57.6 -a -a

80 16.3

(0.1)

54.7

(2.3)

10.9

(1.1)

335.4

(40.8)

23.84

(3.04) -a 45.3 19.3 80

39.6

(8.8)

1.8

(0.4)

0.7

(0.04)

52.8

(7.4)

0.24

(0.04) -a -a -a

100 0.9

(0.1)

103.6c

(4.6)

6.6c

(0.1)

77.9c

(25.4)

4.74c

(1.28) -a 44.8 25.5 100

1.8

(0.1)

2.3

(0.7)

0.7

(0.03)

28.0

(1.0)

0.12

(0.01) -a -a -a

anot observed bn=2 cpremature failure

151

Figure 7.2. Tensile stress-strain diagrams in the water-swollen state of

combinatorial PUI networks with an increasing hydrophobic content (HC).

The water uptake (WU) of each network is reported next to the respective

curve; a) the hydrophobic component is crystallizable in nature (PEG/PCL); b)

the hydrophobic component is amorphous in nature (PEG/PPG). The 100wt%

PCL film (HC:100%, top) failed prematurely.

152

Of high interest are the combined mechanical and thermal

properties of the combinatorial PUI hydrogels. Most importantly, a

considerable increase in the toughness of PUI hydrogels is shown with

increasing melting enthalpy values. As mentioned previously, such a

correlation between crystallinity and toughness has been reported for

polymer materials in the dry state.8 The results presented here form a strong

indication that the same correlation is applicable to hydrogel materials in

the water-swollen state. In other words, the toughness of hydrogel materials

can be increased considerably by increasing the crystallinity of the

hydrophobic component of the hydrogel in the water-swollen state. This is

considered an important structure-property-relationship in the future

development of tough hydrogel materials. As shown here, an effective

method to incorporate crystallinity into hydrogel materials is by the

combinatorial trimerization of mixtures of NCO-functionalized prepolymers.

7.4 Conclusions

The combinatorial trimerization of mixtures of NCO-functionalized

prepolymers is an effective method for synthesizing tough hydrogels. For

semi-crystalline combinatorial PUI hydrogels, it is shown that he toughness

of the hydrogel in the water-swollen state increases significantly with

increasing melting enthalpy. This correlation between melting enthalpy and

toughness of the hydrogel in the water-swollen state is an important

structure-property-relationship for the future development of tough

hydrogels for load-bearing biomedical applications.

7.5 Acknowledgements This project has received funding from the European Union’s Horizon 2020

research and innovation program under the Marie Skłodowska-Curie Grant

Agreement no. 642890 (http://thelink-project.eu/).

The support of the labs of Dr. Frank Richter and the analytical departments

of Covestro Deutschland AG Leverkusen is greatly acknowledged.

153

7.6 References 1. Liu, Y. et al. Recent Developments in Tough Hydrogels for

Biomedical Applications. Gels 4, 46 (2018). 2. Creton, C. 50th Anniversary Perspective: Networks and Gels: Soft

but Dynamic and Tough. Macromolecules 50, 8297–8316 (2017). 3. Gong, J. P. Materials both Tough and Soft. Science (80-. ). 344, 161–

162 (2014). 4. Gong, J. P. Why are double network hydrogels so tough? Soft

Matter 6, 2583–2590 (2010). 5. Chen, Q., Chen, H., Zhu, L. & Zheng, J. Fundamentals of double

network hydrogels. J. Mater. Chem. B 3, 3654–3676 (2015). 6. Dragan, E. S. Design and applications of interpenetrating polymer

network hydrogels. A review. Chem. Eng. J. 243, 572–590 (2014). 7. Driest, P. J., Dijkstra, D. J., Stamatialis, D. & Grijpma, D. W. Chapter 6

of this Thesis. (2019). 8. Galeski, A. Strength and toughness of crystalline polymer systems.

Prog. Polym. Sci. 28, 1643–1699 (2003). 9. Read, B. E. Mechanical relaxation in some oxide polymers. Polymer

(Guildf). 3, 529–542 (1962). 10. Bailey, F. E. (Frederick E. & Koleske, J. V. Poly(ethylene oxide).

(Academic Press, 1976). 11. P.Pêgo, A., Poot, A. A., Grijpma, D. W. & Feijen, J. Copolymers of

trimethylene carbonate and ε-caprolactone for porous nerve guides: Synthesis and properties. J. Biomater. Sci. Polym. Ed. 12, 35–53 (2001).

12. Faucher, J. A. The dependence of glass transition temperature on molecular weight for poly(propylene oxide) and poly(butylene oxide). J. Polym. Sci. Part B Polym. Lett. 3, 143–145 (1965).

Chapter 8 - Poly(ethylene glycol)

Poly(urethane-isocyanurate) Hydrogels

for Contact Lens Applications

P.J. Driest1,2, I.E. Allijn2, D.J. Dijkstra1, D. Stamatialis2, D.W. Grijpma2

1Covestro Deutschland AG, CAS-Global R&D, 51373 Leverkusen, Germany

2Technical Medical Centre, Faculty of Science and Technology, Department

of Biomaterials Science and Technology, University of Twente P.O. Box 217,

7500AE Enschede, The Netherlands

Published in:

Polym. Int. (in press), DOI: 10.1002/pi.5938

Abstract

Over the past decades, the global

use and market of contact lenses

has expanded steadily. Due to the

many demands on material

properties (e.g. mechanical,

optical and biological) the

development of novel contact lens

materials is challenging.

Specifically, the ideal combination

of high equilibrium water content (EWC), high toughness in the hydrated

state and low protein adsorption is difficult to realize.

In this work, poly(ethylene glycol) poly(urethane-isocyanurate) (PEG

PUI) type hydrogels that combine the above important properties are

presented as a new class of materials for contact lens applications. It is

shown that these PEG PUI hydrogels demonstrate high toughness values in

the hydrated state ranging from 98 to 226 kJ·m-3 and elastic moduli ranging

from 0.8 to 17.2 MPa for networks with equilibrium water contents ranging

from 76.3 to 16.1wt%. These hydrogels also demonstrate transmittance

values >90% across the visible spectrum, clarities close to 100% in most

cases and refractive indices ranging from 1.48 to 1.36. Importantly, these

hydrogels are non-cytotoxic and demonstrate lower bovine serum albumin

adsorption values than several commercial contact lenses of 0.24 to 0.65

mg/g compared to 0.55 to 1.38 mg/g after 24h, respectively.

This combination of high EWC, high toughness in the hydrated state

and low protein adsorption is exceptional. These properties can be

attributed to the PEG PUI network structure: the use of a PEG polymeric

backbone provides hydrophilicity and chemical inertness while the PUI type

crosslinking units provide high toughness in the hydrated state.

156

8.1 Introduction Since the introduction of soft contact lenses in the 1960s, the development

of contact lens materials has received much attention over the years and is

still an actively discussed topic today.1,2 At present, the market for soft

contact lenses is estimated to exceed $8.5billion globally, with an estimated

150 million users of contact lenses worldwide in 2019.3,4 Historically, the

market for soft contact lenses is dominated by 2 main classes of materials:

poly(2-hydroxyethyl methacrylate) (pHEMA) type hydrogels and silicone

type hydrogels.1 An overview of common contact lens materials and their

general physical properties is shown in Table 8.1.2

Despite the long history of the field, novel (classes of) soft contact

lens materials are still actively sought after and reported on today.5 The

development of new contact lens materials is considered a challenging field

though. This is easily appreciated when looking at the impressive list of

requirements a material has to fulfil before being considered for contact lens

applications (see Table 8.2).2,5,6

Typically, modification of the various material characteristics is not

complementary and improving one property often deteriorates another. A

specifically challenging combination of contradictory material properties is

high hydrophilicity, high toughness and low biofouling.7,8 One of the reasons

for this is that the hydrophilicity of contact lens materials is generally

realized by the introduction of hydrogen-bonding moieties (e.g. -OH, -NH2

or -COOH groups) into the polymeric backbone. However, the presence of

such hydrogen bonding moieties has been shown to play a key role in the

biofouling process of hydrogel materials.9 By facilitating biofouling, contact

lenses can hold microbes in prolonged contact with the cornea of the eye,

leading to ocular infections and complications.8 In fact, even with good

contact lens care practice, up to 80% of contact lenses and contact lens cases

were found to harbor disease-causing microbes.10 This shows that there is

still need for a low-biofouling hydrophilic hydrogel material for contact lens

applications.6,8,11–13

157

Table 8.1. General properties of some common contact lens materials.

Adapted from Musgrave et al.2

Material EWC

[%]

Emod

[MPa]

O2-

perm.a

[Dk]

Wear

timeb

[days]

Polymer structure

PMMA 0 1000 0 <1

PMMA-

Silicone 0 -c 15 -c

Silicone-

HEMA (rigid) 0 10 10-100 -c

pHEMA

Hydrogel

30-

80 0.2-2 10-50 1-7

Silicone

(PDMS)

Hydrogel

20-

55 0.2-2 60-200 ~7-28

PVA 60-

70 -c 10-30 <1

aOxygen permeability, typically reported in Dk-units (ml O2/[ml⋅mmHg])

brecommended maximum wear time before lens disposal

cdata not reported

158

Table 8.2. List of typical requirements for contact lens materials.

Material Property Requirement / Comment

Physical and

Chemical

Hydrophilicity As high as possible. Typically indicated by the

equilibrium water content (EWC)

Oxygen permeability As high as possible

Mechanical

properties in the

hydrated state

Need to be suitable. Most notably, a suitable

elastic modulus and high toughness

Chemical stability Sufficient. With respect to both shelf-life and

common sterilization techniques such as UV-

exposure or autoclaving

Optical

Transmittance As high as possible across the visible spectrum

Clarity As high as possible. Clarity is typically considered

a measure for the see-through quality of the

material

Refractive index Needs to be suitable

Biological

Biocompatibility Required. Typically indicated by non-cytotoxicity

Biofouling As low as possible. Specifically, low protein

adsorption

159

In this context, poly(ethylene glycol) (PEG) is considered a promising

material.2,5 Specifically the combination of high hydrophilicity with chemical

inertness is attractive. The main drawback associated with PEG hydrogel

materials however, is the poor mechanical performance (i.e. the poor

toughness) in the hydrated state that is typically associated with crosslinked

PEG hydrogels.14–16 Consequently, the use of PEG has thus far been limited

to surface modification procedures of existing contact lens materials.8,11,17

Hydrogels based on crosslinked PEG have not been reported in this context.

Scheme 8.1. Schematic representation of the synthesis of a PEG PUI polymer

network. First, a PEG polymeric diol is functionalized using hexamethylene

diisocyanate (HDI); next the PUI network is formed by the selective

trimerization of the NCO-functionalized PEG prepolymer.

160

In this work, we investigate the use of a recently developed new

class of PEG-based poly(urethane-isocyanurate) (PUI) type polymer

networks for contact lens applications.18 The synthesis of these networks is

based on the functionalization and subsequent trimerization of PEG

polymeric diols using hexamethylene diisocyanate (HDI), as shown in

Scheme 8.1. The resulting PEG PUI networks were shown to exhibit high

water uptakes of up to 508wt% combined with high toughness in the

hydrated state.18 Since both PEG and PUI type polymer networks are also

known for their excellent optical properties19,20, high refractive indices21–23

and biocompatibility24,25, we hypothesize that these PEG PUI type hydrogels

would be a promising class of materials for contact lens applications. PEG

PUI hydrogels of varying crosslinking density are synthesized and

characterized for properties relevant to contact lens applications (see Table

8.2) and compared to commercially available contact lenses.

161

8.2 Experimental Details

Materials and methods

Materials

Poly(ethylene glycol)s (𝑀𝑛 = 0.4, 0.6, 1, 2, 4 and 10 kg/mol), stannous

octoate (Sn(II)Oct2) and resazurin sodium salt were purchased from Sigma-

Aldrich. Hexamethylene diisocyanate (HDI) was supplied by Covestro

Deutschland AG. A human retinal pigment epithelium cell line (Arpe-19) and

Dulbecco's Modified Eagle Medium/Nutrient Mixture F-12 (DMEM/f12) with

10% fetal bovine serum (FBS) cell culture medium were purchased from

ATCC. Commercial contact lenses (Midafilcon A day lenses manufactured by

Menicon and Riofilcon A day lenses, Comfilcon A month lenses and

Somofilcon A month lenses manufactured by Cooper Vision) were kindly

supplied by a local optometrist. A DC Protein Assay kit was purchased from

Bio-Rad. All other chemicals were purchased from Sigma-Aldrich.

Hydrogel synthesis and characterization

All chemical reactions were conducted under a nitrogen atmosphere using

standard Schlenck-line techniques. NMR-spectra were collected on a Bruker

Advance III-700 in C6D6. Size exclusion chromatography (SEC) was performed

according to DIN 55672-1:2016-03, on 4 PSS SDV Analytical columns (2x

100Å, 5µm; 2x 1000Å, 5µm) using an Agilent 1100 Series pump and an

Agilent 1200 Series UV Detector (230 nm) with tetrahydrofuran as the

elution solvent at 40°C and 1.00 mL/min. Residual monomeric diisocyanate

contents were determined by gas chromatography (GC) according to DIN EN

ISO 10283 on an Agilent Technologies 6890N system using a 15.0m DB17

column and tetradecane as an internal standard. Attenuated total reflection

Fourier-transformed infrared (ATR-FTIR) spectra were recorded on a Bruker

FTIR Spectrometer Tensor II with a Platinum-ATR-unit with diamond crystal.

Physical properties

Differential scanning calorimetry (DSC) measurements were performed

under nitrogen on ca. 10 mg samples on a Perkin-Elmer Calorimeter DSC-7

162

in 2 heating runs from -100°C to +150°C with a heating rate of 20 K/min and

a cooling rate of 20 K/min. The glass transition temperature (Tg) and peak

melting temperature (Tm,p) were determined based on the second heating

run. Tensile tests were performed in triplicate at room temperature using a

Zwick Z0.5 tensile tester equipped with a 500N load cell at 200mm/min. with

a grip-to-grip separation of 35mm on strips of ca 60 x 4 x 0.5mm cut from a

cured film. Elastic moduli and strain values were derived from the initial grip-

to-grip separation and are indicative. The toughness values of the materials

were calculated from the work at break value, normalized for the

dimensions of the specimen.

For swelling measurements, three samples measuring

approximately 10 x 20 x 0.5mm were cut from a cured film and weighed in

the dry state (values denoted 𝑚𝑑). Each sample was then swollen in excess

water for at least 24h, blotted dry and weighed again in the hydrated state

(values denoted 𝑚𝑤). The water uptake of the material was then estimated

as:

𝑤𝑎𝑡𝑒𝑟 𝑢𝑝𝑡𝑎𝑘𝑒 =𝑚𝑤−𝑚𝑑

𝑚𝑑× 100% (eq. 1)

The equilibrium water content (EWC) was estimated as:

𝐸𝑊𝐶 =𝑚𝑤−𝑚𝑑

𝑚𝑤× 100% (eq. 2)

Oxygen permeabilities were estimated according to the model proposed by

Morgan and Efron26 using the equation:

𝐷𝑘 = 1.67𝑒0.0397𝐸𝑊𝐶 (eq. 3)

163

where Dk is the oxygen permeability of the material in Dk-units (ml

O2/[ml⋅mmHg]) and EWC is the equilibrium water content.

Optical properties

Wavelength-dependent transmission experiments were performed using a

Konica Minolta CM-5 spectrophotometer. Scattering experiments were

performed using a BYK-Gardner Haze-Gard Plus haze meter after norm

ASTM D 1003_11. Scattering characteristics are divided into three

parameters: transmittance (T), haze (H) and clarity (C). Of these,

transmittance is defined as the ratio of transmitted light compared to the

incident light. Haze indicates the amount of light diffused under wide angle

(>2.5°) scattering. Clarity indicates the amount of light diffused under

narrow angle (<2.5°) scattering. Refractive index measurements were

performed at 20.0°C using a Rudolph Research J357 automatic

refractometer equipped with a temperature-controlled sample presser for

gels. Six independent measurements were taken for each sample, of which

the average value is reported.

Biological properties

Sterilization by UV-exposure was performed for 1h under four 9 Watt UV-

lamps at 365 nm and approx. 6 mW/cm2. Sterilization by autoclaving was

performed for 20min at 120°C in a PBI International pressure cooker from

Salm and Kipp BV. Fluorescence measurements were performed on a Perkin

Elmer Wallac Victor3 spectrophotometer.

Synthesis and characterization of NCO-functionalized PEG prepolymers

All polymeric diols were dried prior to use by azeotropic distillation in

toluene, which was subsequently removed under reduced pressure. In a

typical experiment, 0.25 g of dibutylphosphate was added to 100 g of dried

polymeric diol (where necessary, kept in the melt using a heat gun). This

mixture was added dropwise to HDI at 100°C, using a molar ratio of NCO:OH

of 15:1, and subsequently kept at 100°C for 3h. The resulting mixture was

then transferred dropwise into a short-path thin-film evaporator, which was

operated at a reduced pressure of 1∙10-2 mbar and 140°C. Products were

164

collected as transparent, colorless, viscous resins, of which the PEG-1k-,

PEG-2k-, PEG-4k- and PEG-10k-based resins crystallized upon cooling to

room temperature. Products were analyzed by SEC, GC and 1H-NMR (700

MHz, C6D6, representative data for PEG-4k): δ = 4.60 (2H, brs), 4.26 (4H, t),

3.50 (410H, m), 2.97 (4H, q), 2.56 (4H, t), 1.12 (4H, t), 1.02 (4H, t), 0.88 (8H,

m) ppm.

Synthesis and characterization of PEG PUI polymer networks by the

selective trimerization of NCO-functionalized PEG prepolymers

Poly(urethane-isocyanurate) (PUI) polymer networks were synthesized in

analogy to a previously reported procedure, with slight modifications.18 In a

typical experiment, 8g of NCO-prepolymer was weighed into a

polypropylene cup, heated to 90°C and subsequently degassed by gently

applying reduced pressure until no more gas formation was observed. Next,

ca. 60 mg of Sn(II)Oct2 (corresponding to ca. 0.75wt% relative to the

prepolymer) was added at 90°C. After mixing for 15 seconds in a Hauschild

Speedmixer DAC150FVZ, the reaction mixture was brought into a mold

consisting of a 0.5mm-thick polycarbonate frame clamped between two

silanized glass plates measuring approximately 20 x 10cm and kept at 90°C

for 24h. Products were collected as flexible transparent homogeneous films,

of which the films based on PEG-4k and PEG-10k prepolymers slowly turned

opaque upon cooling to room temperature. The films were analyzed in the

dry state by ATR-FTIR, water swelling measurements, tensile testing and DSC

experiments and subsequently swollen in excess water for at least 24h.

Finally, the films were analyzed in the hydrated state for physical properties

(tensile testing and mass-loss after sterilization by UV-exposure or

autoclaving), optical experiments (transmittance, scattering and refractive

index) and biological experiments (cytotoxicity and protein adsorption,

described below).

165

Indirect cytotoxicity measurements of PEG PUI hydrogels and commercial

contact lenses

Cytotoxicity measurements were performed on the PEG PUI hydrogels and

several commercial contact lenses as described previously, with minor

modifications.27 In short, PEG PUI hydrogels were swollen in milliQ water,

cut in discs with a diameter of 12mm (n=3), blotted dry and weighed. Next

the hydrogels were transferred to a phosphate buffered saline (PBS) solution

and autoclaved for 20min at 120℃ in a pressure cooker. The PEG PUI

hydrogels and the commercial contact lenses (n=3) were then incubated in

DMEM/f12 with 10% FBS (0.1 g/mL) at 37℃ in a humidified incubator

containing 5% CO2. Meanwhile, Arpe-19 cells were seeded on a polystyrene

tissue culture well plate with a concentration of 12.5·104cells/mL and left for

24h at 37℃ in a humidified incubator containing 5% CO2. After 24h, the

medium was removed from the Arpe-19 cells and replaced with conditioned

medium from the PEG PUI hydrogels or commercial contact lenses,

respectively. A control cell culture was kept in which the medium was

replaced with fresh medium that had not been in contact with hydrogel

materials. After 48h the cell viability was measured using an Alamar blue

solution (440µM resazurin sodium salt in phosphate buffered saline (PBS))

and normalized with respect to the control cell-culture. To measure the cell

viability, a 10:1 (v/v) mixture of cell-medium and Alamar blue solution was

added to the cells. After incubating for 2h at 37℃, the fluorescence emission

of the solution at 590nm was measured using an excitation wavelength of

560nm.

Protein adsorption measurements of PEG PUI hydrogels and commercial

contact lenses

Protein adsorption measurements were performed on the PEG PUI hydrogel

films and several commercial contact lenses as described previously, with

minor modifications.6 In short, the PEG PUI hydrogels were swollen in milliQ

water and cut in discs with a diameter of 12mm (n=3). Commercial contact

lenses were used as supplied. All samples were incubated for 24h in 2mL

PBS, containing 1.5mg bovine serum albumin (BSA) per mL PBS, under gentle

shaking at 37°C. Afterwards, the samples were extracted three times with

166

neat PBS for at least 1.5h per extraction step and subsequently dried. The

dry samples were then swollen in PBS containing 1% (v/v) sodium dodecyl

sulfate (SDS) and kept at 37°C under gentle shaking for three days. The

concentration of BSA in the supernatant was then determined using a Bio-

Rad DC Protein Assay kit according to the supplier’s instructions.

167

8.3 Results and Discussion

Synthesis of NCO-functionalized PEG prepolymers and PEG PUI polymer

networks by the selective trimerization of NCO-functionalized PEG

prepolymers

Poly(ethylene glycol) (PEG) polymeric diols were functionalized using

hexamethylene diisocyanate (HDI) and subsequently trimerized as reported

previously.18 Using 1H-NMR, SEC and GC, it was confirmed that the

functionalization of all PEG polymeric diols with HDI succeeded selectively

and quantitatively. The characteristics of all functionalized PEG prepolymers

are reported in the supporting information in Table S1. Next, the NCO-

functionalized prepolymers were crosslinked by the selective trimerization

of the terminal NCO groups. This resulted in the formation of PEG

poly(urethane-isocyanurate) (PUI) type polymer networks. The completion

of the trimerization reaction was confirmed using attenuated total reflection

Fourier-transformed infrared spectroscopy (ATR-FTIR) (supporting

information, Figure S1).

Physical properties of PEG PUI networks and hydrogels: equilibrium water

content, oxygen permeability, mechanical properties and stability

Water uptake, EWC and oxygen permeability

Due to the hydrophilicity of PEG, the water uptake and resulting equilibrium

water content of the PEG PUI networks were generally high. The shortest-

chain PEG PUI network, synthesized from a 0.4 kg/mol PEG prepolymer,

demonstrated a water uptake of 19.2wt%. This corresponds to an EWC of

16.1wt% in the hydrated state. The longest-chain PEG PUI hydrogel on the

other hand, synthesized from a 10 kg/mol PEG prepolymer, demonstrated a

water uptake of 321.6wt%, corresponding to an EWC of 76.3wt%. All water

uptakes and EWCs are reported in Table 8.3. Owing to the well-defined

network structure on which we reported previously, the water uptakes and

EWCs of the hydrogels could easily be controlled over a wide range.18

Notably, the range of EWCs measured for the PEG PUI hydrogels

168

encompasses almost the entire range of EWCs typically reported for contact

lens hydrogels (i.e. approx. 20 to 80wt%, see Table 8.1).2 This includes both

soft and rigid contact lens hydrogels.

From the EWCs, the oxygen permeabilities of the PEG PUI hydrogels

were approximated using the model proposed by Morgan and Efron (eq. 3)

and are reported in Table 8.3.26 This semi-empirical model correlates oxygen

permeability (the product of oxygen diffusivity and solubility, typically

referred to as “Dk” in contact lens literature) to EWC and is used for

conventional, non-silicon hydrogels.28 The model assumes that oxygen

transport only takes place through the water phase of the hydrogel while no

oxygen transport takes place through the solid phase of the hydrogel.

Consequently, the oxygen permeability values estimated using this model

should be considered as minimal values rather than the actual oxygen

permeability values. The difference between the predicted and actual

oxygen permeability values evidently depends on the specific oxygen

transport properties of the hydrogel´s solid phase. In this respect, it is

worthwhile to note that PEG polymer networks have been shown to exhibit

a non-zero oxygen permeability in the dry state and actually demonstrate a

certain level of oxygen affinity.29 Therefore, we expect that the PEG PUI

hydrogels have higher actual oxygen permeability values than the

approximated values reported in Table 8.3. However, an experimental

assessment will be needed to determine the exact oxygen permeabilities of

the PEG PUI hydrogels.

Table 8.3. Physical properties of the PEG PUI networks and hydrogels in the dry state and in the hydrated state. WU stands for

water uptake. Numbers in parentheses are standard deviations.

PEG PUI Hydrogel

Prepolymer

WU

[%]

EWC

[%]

Tg

[°C]

Tm,p

[°C]

Emod,dry

[MPa]

σb,dry

[MPa]

εb,dry

[%]

Emod,wet

[MPa]

σb,wet

[MPa]

εb,wet

[%]

Wb,wet

[kJ·m-3]

O2-perm.c

[Dk]

PEG-0.4k-diNCO 19.2

(0.1)

16.1

(0.1) -20.2 - a

17.1

(2.4)

2.7

(0.8)

13.9

(2.6)

17.0

(0.8)

2.0

(0.4)

9.8

(2.3)

110

(40) 3.2

PEG-0.6k-diNCO 26.1

(0.3)

20.7

(0.2) -23.2 - a

18.7

(1.3)

2.8

(0.4)

13.6

(1.6)

17.2

(0.5)

2.0

(0.3)

9.4

(1.6)

99

(27) 3.8

PEG-1k-diNCO 87.1

(0.6)

46.5

(0.2) -47.2 - a

10.9

(0.6)

1.4

(0.2)

11.5

(2.1)

9.6

(0.3)

1.6

(0.8)

14.1

(8.5)

142

(123) 10.6

PEG-2k-diNCO 150.6

(0.7)

60.1

(0.1) -42.9 21.6

6.6

(1.1)

1.6

(0.4)

25.1

(10.1)

5.9

(0.4)

1.1

(0.1)

17.0

(2.8)

98

(20) 18.2

PEG-4k-diNCO 222.6

(0.7)

69.0

(0.1) -49.5 39.3

171.5

(6.1)

19.3

(0.5)

250.7

(26.2)

4.0

(0.2)b

0.8

(0.1)b

21.7

(1.0)b

104

(5)b 25.9

PEG-10k-diNCO 321.6

(4.2)

76.3

(0.2) -49.4 53.8

309.2

(30.0)

26.4

(2.7)

517.0

(40.9)

0.8

(0.3)

0.4

(0.1)

88.9

(23.8)

226

(94) 34.5

anot observed bn=2 capproximated value calculated from the EWC assuming zero transport of oxygen through the solid phase of the hydrogel. Reported in Dk-

units (ml O2/[ml⋅mmHg])

170

Mechanical properties in the dry state and in the hydrated state

The mechanical properties of all polymer networks were assessed both in

the dry state and in the hydrated state and are reported in Table 8.3. It was

found that the dry polymer networks were generally tough materials, with

varying mechanical properties dependent on network density and

crystallinity. In the dry state, the elastic moduli (Emod,dry) of the networks

ranged from 6.6 to 309.2 MPa, while the stress at break (σb,dry) and

elongation at break (εb,dry) values ranged from 1.4 MPa to 26.4 MPa and from

11.5% to 517.0%, respectively. The PUI networks synthesized from PEG

prepolymers with a molecular weight of 2kg/mol and higher were semi-

crystalline in nature in the dry state, displaying melting peaks in the range of

20-60°C.

Table 8.3 also reports the mechanical properties of the hydrogels in

the hydrated state, which are important for their application as contact

lenses. The PEG PUI hydrogels were generally tough in the hydrated state,

covering a wide range of material properties. In the hydrated state, the

elastic moduli of the materials (Emod,wet) ranged from 0.8 to 17.2 MPa, the

stress at break (σb,wet) values ranged from 0.4 to 2.0 MPa and the elongation

at break (εb,dry) values ranged from 9.4% to 88.9%. The hydrogels

demonstrated toughness (Wb,wet) values in the hydrated state ranging from

98 to 226 kJ·m-3.

The tensile stress-strain diagrams of the PUI hydrogels in the

hydrated state are shown in Figure 8.1a, illustrating the high toughness of

the gels. Importantly, the stress at break values were >400 kPa for all

hydrogels and the toughness values were approximately 100 kJ·m-3 or more.

This illustrates the high mechanical resilience of these hydrogels in the

water-swollen state. In practice, it was observed that in the water-swollen

state all hydrogels could easily be handled without fracture or rupture. It

should be noted that this is non-trivial for crosslinked PEG hydrogels. As far

as we know, these PEG PUI hydrogels are the toughest crosslinked PEG

hydrogels reported in the literature, clearly outperforming other typical

crosslinked PEG hydrogels.14–16 The high toughness of the PEG PUI type

hydrogels can likely be attributed to the poly(urethane-isocyanurate) type

network structure on which we reported previously.18

171

To illustrate the wide range of mechanical properties accessible by

the synthesis method used, the tensile elastic moduli of the hydrogels are

plotted together with the EWCs in Figure 8.1b. The measured range of elastic

moduli covers most of the range of elastic moduli typically reported for soft

contact lens materials (0.2-2MPa) and also includes typical values reported

for rigid contact lens materials (ca. 10 MPa) (see Table 8.1).2 It should be

noted that the current span of elastic moduli is obtained simply by varying

the molecular weight of the starting material. Other strategies typically used

to manipulate the mechanical properties of hydrogel materials may also be

considered, such as the incorporation of mono-functional prepolymers

and/or prepolymers with a different chemical nature.

Hydrogel stability during sterilization

A first assessment of the stability of the PEG PUI hydrogels was done by

exposing the hydrogels to two common sterilization procedures. The

hydrogels were analyzed gravimetrically to detect any changes in mass and

visually by inspecting each sample for changes in color, size and/or opacity.

No mass loss was determined for any of the hydrogels after 1h of UV-

exposure nor after autoclaving at 120°C for 20min. Also, no changes in

color, size or opacity of the hydrogels were observed. Although changes in

other material properties, such as mechanical properties, degree of

swelling or optical properties, have not been investigated, these results

indicate that the PEG PUI hydrogels are stable under the sterilization

procedures used.

Figure 8.1. Mechanical properties of PEG PUI hydrogels in the water-swollen state; a) tensile stress-strain diagrams of PEG PUI

hydrogels in the hydrated state. The molecular weights of the respective prepolymers as well as the equilibrium water contents

are reported next to the curves; b) tensile elastic moduli (Emod; black squares; approximated as the tangent of the linear regime

of the tensile stress-strain dependence) and equilibrium water content (EWC; blue circles) of the PEG PUI hydrogels as a function

of the number average molecular weight between crosslinks Mc of the polymer network, approximated by the number average

molecular weight of the respective prepolymer Mn.

173

Optical properties of PEG PUI hydrogels: transmittance, haze, clarity and

refractive index

Transmittance, Haze and Clarity

The transparency of the PEG PUI hydrogels was investigated by two

methods. Firstly, the transmittance of the hydrogels was determined as a

function of wavelength. The transmission spectra of all hydrogels are shown

in Figure 8.2a. Typically, an average transmittance of 90% is considered a

minimum requirement for contact lens applications,2,30 although minimum

values of 95% are also reported.5 As seen in the spectra, the transmittance

of the hydrogels was generally >90% throughout the visible spectrum and

on average around 95% for most hydrogels. It should be noted that these

measurements were performed on hydrogel films with a thickness of 0.5-

1mm, whereas the typical center-point-thickness of a contact lens is around

50-250µm.31 In other words, even in non-optimized form the PEG PUI

hydrogels already meet the general transmittance requirements for contact

lens applications.

Secondly, the scattering characteristics of the transmitted light were

investigated (see results in Table 8.4). All hydrogels demonstrated high

transmittance values of >90%, with the PEG-1k, PEG-2k and PEG-4k

hydrogels demonstrating transmittance values of around 95%. This is

consistent with the wavelength dependent measurements discussed

previously. More importantly, the PEG PUI hydrogels generally showed very

high clarities, in most cases with values close to 100%. Clarity is also referred

to as the “see-through quality” of the material and describes how well fine

details and sharp images can be perceived when looking through the

material.32 Finally, the haze values, considered as a measure for the loss of

contrast32, were in most cases generally low and in the range of 1-4%.

Figure 8.2. Optical properties of PEG PUI hydrogels; a) wavelength-dependent transmission spectra of the PEG PUI hydrogels in

the UV-visible spectrum. A dotted grey line is included at the 90%-transmittance threshold as a guide for the eye; b) refractive

indices (nD20) and equilibrium water contents (EWC) of the PEG PUI hydrogels as a function of the molecular weight between

crosslinks of the PEG PUI networks. Marked by grey lines is the range of refractive indices described by Caló et al. as “ideal

hydrogels should have a refractive index value matching the range 1.372–1.381”.5

175

Table 8.4. Scattering characteristics (transmittance, haze and clarity) and

refractive indices (nD20) of PEG PUI hydrogels.

PEG PUI

Hydrogel

Prepolymer

Transmittance

[%]

Haze

[%]

Clarity

[%]

Refractive

Index (nD20)

[-]

PEG-0.4k-diNCO 93.1 3.4 96.2 1.4796

PEG-0.6k-diNCO 93.1 1.8 99.5 1.4707

PEG-1k-diNCO 95.2 1.9 99.4 1.4010

PEG-2k-diNCO 95.1 3.0 99.4 1.3768

PEG-4k-diNCO 95.7 2.2 98.7 1.3611

PEG-10k-diNCO 90.5 22.8 86.3 1.3562

Refractive index

The refractive indices of the PEG PUI hydrogels in the hydrated state

are reported in Table 8.4 and plotted in Figure 8.2b together with their

EWCs. The refractive indices ranged from ca. 1.48 for the shortest-chain

(PEG-0.4k) hydrogel to ca. 1.36 for the longest-chain (PEG-10k) hydrogel.

Indicated by the trends shown in Figure 8.2b, the refractive index seems to

be mostly dependent on the EWC of the hydrogel. This is consistent with

what is reported in the literature.33

The literature regarding the preferred value of the refractive index

of contact lens materials is inconsistent. On the one hand, it is reported that

the refractive index value of a contact lens material should preferably be as

high as possible, in order to allow the manufacture of the thinnest possible

contact lenses.34,35 On the other hand, it is also reported that the refractive

index of a contact lens material should preferably match that of the human

cornea as closely as possible, values ideally lying between 1.372 and

1.381.5,36 Typical values reported in the literature for commercial and novel

scientific contact lens materials range from approximately 1.38 to 1.45.33–35

The PEG PUI hydrogels span a wide range of refractive index values,

encompassing all values mentioned in the literature. In this regard, it is safe

to say that the range of refractive index values of the PEG PUI hydrogels is

176

suitable for contact lens applications. The intrinsically high refractive indices

of the materials comprising the hydrogel networks are considered beneficial

in this context. For example, a refractive index value of 1.48 is reported for

a pure HDI polyisocyanurate network22 and a value of 1.47 is provided by the

manufacturer for a liquid PEG-0.6k polymeric diol. Furthermore, the

refractive index of PEG polymeric diols has been shown to increase for

increasing molecular weight PEG.21 These high values allow for hydrogel

formulations with a higher EWC and/or thinner contact lenses, both of which

are associated with improved wearing comfort.34 Also the broad range and

easy control over the refractive index of the PEG PUI hydrogel materials is

considered beneficial for further optimization.

Biological properties of PEG PUI hydrogels: biocompatibility and protein

adsorption

Biocompatibility

The biocompatibility of the prepared and extracted PEG PUI hydrogels was

investigated in an indirect cytotoxicity study in analogy to a procedure

reported previously.27 A human retinal pigment epithelium cell line (Arpe-

19) was used as a model cell line for this. The viabilities of the cells after 48h

is reported in Table 8.5. As can be seen, all PEG PUI hydrogels were non-

cytotoxic under these conditions, indicating good biocompatibility. This is in

line with our expectations, considering that all the chemical moieties

present in the hydrogel network are known to be biocompatible.24,25 In most

cases, PEG PUI hydrogels had a slight stimulating effect on cell growth,

indicated by cell viabilities >100% compared to the control cell culture.

Table 8.5. Biological properties of the PEG PUI hydrogels and a selection of commercial contact lenses. Numbers in parentheses

are standard deviations.

PEG PUI Hydrogel

Prepolymer

EWC

[%]

Cell viability

after 48h

[%]

Adsorbed BSA

after 24ha

[mg/g]

Contact Lens Material

(manufacturer, wear time)

EWC

[%]

Cell

viability

after 48h

[%]

Adsorbed

BSA after

24ha

[mg/g]

PEG-0.4k-diNCO 16.1

(0.1)

111

(3)

0.24

(0.09)b

Midafilcon A

(Menicon, day) 56c

99

(4)

0.85

(0.07)b

PEG-0.6k-diNCO 20.7

(0.2)

110

(3)

0.45

(0.05)

Riofilcon A

(Cooper Vision, day) 58c

92

(3)

1.38

(0.33)

PEG-1k-diNCO 46.5

(0.2)

109

(2)

0.39

(0.18)

Comfilcon A

(Cooper Vision, month) 48c -d

0.55

(-)e

PEG-2k-diNCO 60.1

(0.1)

106

(1)

0.64

(0.19)

Somofilcon A

(Cooper Vision, month) -f -d

0.71

(-)e

PEG-4k-diNCO 69.0

(0.1)

102

(1)

0.65

(0.08)

PEG-10k-diNCO 76.3

(0.2)

104

(3)

0.64

(0.14)

aquantified as the total mass of adsorbed BSA in 24h, normalized for the dry weight of the hydrogel sample; bn=2; cas provided by the manufacturer; dnot determined; en=1; fnot provided

178

Protein Adsorption

To investigate the biofouling properties of the PEG PUI hydrogels, a protein

adsorption study was performed in analogy to a procedure reported

previously.6 Bovine serum albumin (BSA) was used as a model protein. The

amounts of adsorbed BSA after 24h, normalized for the dry weight of the

hydrogel, are reported in Table 8.5 for all PEG PUI hydrogels and several

commercial contact lenses. The PEG PUI hydrogels demonstrated BSA

adsorption values ranging from 0.24 mg/g to 0.65 mg/g. These values are

significantly lower than those of the commercial contact lenses tested,

which ranged from 0.55 mg/g to 1.38 mg/g. Especially the disposable

contact lenses intended for single-day use demonstrated high BSA

adsorption values of 0.85 mg/g and 1.38mg/g. This indicates that PEG PUI

hydrogels are suitable for the production of contact lenses with a wear time

of ≥1 month, at least as far as biofouling properties are concerned.

The low biofouling properties of the PEG PUI hydrogels can likely be

attributed to the inert chemical structure of the PEG PUI hydrogel networks.

The new PEG PUI hydrogels are unique in this context, in the sense that they

are hydrophilic but do not contain any free -OH, -NH2, -COOH or charged

moieties. This results in the desirable combination of high EWC hydrogels

with low biofouling properties, which is challenging to achieve using the

currently used contact lens materials. This is considered a significant

improvement, since biofouling is at present one of the main issues

associated with contact lens materials.6,8,11–13 In combination with the high

toughness in the hydrated state discussed previously, this makes PEG PUI

hydrogels an attractive new class of materials for contact lens applications.

179

8.4 Conclusions PEG PUI hydrogels comprise an attractive novel class of hydrogel materials

for contact lens applications, meeting all requirements on material

properties reported for the application (see Table 8.2). Owing to the specific

PEG PUI network structure, a desirable combination of high EWC, high

toughness in the hydrated state, excellent optical properties, good

biocompatibility and low biofouling properties is achieved. Also the oxygen

permeability and the chemical stability with respect to common sterilization

techniques of the PEG PUI hydrogels are suitable for the application in mind.

This particular combination of properties can be attributed to the

PEG PUI polymer network structure comprising the hydrogel networks. The

use of a PEG polymeric backbone provides the unusual combination of

hydrophilicity and chemical inertness. As a result, high EWC hydrogels that

demonstrate a low protein adsorption are obtained. The PUI type

crosslinking units provide the high toughness needed for contact lens

applications, which conventional crosslinked PEG hydrogels typically lack.

8.5 Acknowledgements This project has received funding from the European Union’s Horizon 2020

research and innovation program under the Marie Skłodowska-Curie Grant

Agreement no. 642890 (http://thelink-project.eu/) and from ZonMW under

Grant Number MKMD 114022507. The support of the analytical

departments of Covestro Deutschland AG and Currenta AG is greatly

acknowledged. Duco Schriemer and 20Med Therapeutics BV are kindly

acknowledged for supplying an Arpe-19 cell line and cell culture medium.

180

8.6 References

1. Nicolson, P. C. & Vogt, J. Soft contact lens polymers : an evolution. 22, 3273–3283 (2001).

2. Musgrave, C. S. A. & Fang, F. Contact Lens Materials: A Materials Science Perspective. Materials (Basel). 12, 261 (2019).

3. Nichols, J. J. & Fischer, D. Contact Lenses 2018. Contact Lens Spectr. 34, 18–23 (2019).

4. Moreddu, R., Vigolo, D. & Yetisen, A. K. Contact Lens Technology: From Fundamentals to Applications. Adv. Healthc. Mater. 1900368 (2019). doi:10.1002/adhm.201900368

5. Caló, E. & Khutoryanskiy, V. V. Biomedical applications of hydrogels : A review of patents and commercial products. Eur. Polym. J. 65, 252–267 (2015).

6. Zhang, W., Li, G., Lin, Y., Wang, L. & Wu, S. Preparation and characterization of protein-resistant hydrogels for soft contact lens applications via radical copolymerization involving a zwitterionic sulfobetaine comonomer. J. Biomater. Sci. Polym. Ed. 28, 1935–1949 (2017).

7. Annabi, N. et al. 25th Anniversary Article: Rational Design and Applications of Hydrogels in Regenerative Medicine. Adv. Mater. 26, 85–124 (2014).

8. Xiao, A. et al. Strategies to design antimicrobial contact lenses and contact lens cases. J. Mater. Chem. B 6, 2171–2186 (2018).

9. Elder, M. J., Stapleton, F., Evans, E. & Dart, J. K. G. Biofilm-related infections in ophthalmology. Eye 9, 102–109 (1995).

10. Dart, J. The inside story: why contact lens cases become contaminated. Contact Lens Anterior Eye 20, 113–118 (1997).

11. Thissen, H. et al. Clinical observations of biofouling on PEO coated silicone hydrogel contact lenses. Biomaterials 31, 5510–5519 (2010).

12. Lord, M. S., Stenzel, M. H., Simmons, A. & Milthorpe, B. K. The effect of charged groups on protein interactions with poly ( HEMA ) hydrogels. 27, 567–575 (2006).

13. Korogiannaki, M., Samsom, M., Schmidt, T. A. & Sheardown, H. Surface-Functionalized Model Contact Lenses with a Bioinspired Proteoglycan 4 ( PRG4 ) -Grafted Layer. ACS Appl. Mater. Interfaces 10, 30125–30136 (2018).

14. Zant, E., Bosman, M. J. & Grijpma, D. W. Combinatorial synthesis of photo-crosslinked biodegradable networks. J. Appl. Biomater. Funct. Mater. 10, 197–202 (2012).

181

15. Gong, J. P. Why are double network hydrogels so tough? Soft Matter 6, 2583–2590 (2010).

16. Sakai, T. et al. Design and fabrication of a high-strength hydrogel with ideally homogeneous network structure from tetrahedron-like macromonomers. Macromolecules 41, 5379–5384 (2008).

17. Moser, T. et al. Hydrophilic Organic Electrodes on Flexible Hydrogels. (2016). doi:10.1021/acsami.5b10831

18. Driest, P. J., Dijkstra, D. J., Stamatialis, D. & Grijpma, D. W. The Trimerization of Isocyanate-Functionalized Prepolymers : An Effective Method for Synthesizing Well-Defined Polymer Networks. Macromol. Rapid Commun. 40, 1800867 (2019).

19. Meier-Westhues, U., Danielmeier, K., Kruppa, P. & Squiller, E. P. Polyurethanes - Coatings, Adhesives and Sealants. (Vincentz-Network Verlag, 2019).

20. Driest, P. J. et al. Aliphatic isocyanurates and polyisocyanurate networks. Polym. Adv. Technol. 28, 1299–1304 (2017).

21. Kolská, Z., Valha, P., Slepička, P. & Švorčík, V. Refractometric study of systems water-poly(ethylene glycol) for preparation and characterization of Au nanoparticles dispersion. Arab. J. Chem. (2016). doi:10.1016/J.ARABJC.2016.11.006

22. Flipsen, T. A. C., Steendam, R., Pennings, A. J. & Hadziioannou, G. A novel thermoset polymer optical fiber. Adv. Mater. 8, 45–48 (1996).

23. Fabbri, P. et al. Highly stable plastic optical fibre amplifiers containing [Eu(btfa)3(MeOH)(bpeta)]: A luminophore able to drive the synthesis of polyisocyanates. Polymer (Guildf). 55, 488–494 (2014).

24. Liu, G. et al. Cytotoxicity study of polyethylene glycol derivatives. RSC Adv. 7, 18252–18259 (2017).

25. Burel, F., Poussard, L., Tabrizian, M., Merhi, Y. & Bunel, C. The influence of isocyanurate content on the bioperformance of hydrocarbon-based polyurethanes. J. Biomater. Sci. Polym. Ed. 19, 525–540 (2008).

26. Morgan, P. B. & Efron, N. The oxygen performance of contemporary hydrogel contact lenses. Contact Lens Anterior Eye 21, 2–6 (1998).

27. Ulu, A. et al. Poly(2-hydroxyethyl methacrylate)/boric acid composite hydrogel as soft contact lens material: Thermal, optical, rheological, and enhanced antibacterial properties. J. Appl. Polym. Sci. 135, 46575 (2018).

28. Seitz, M. E. et al. Influence of silicone distribution and mobility on the oxygen permeability of model silicone hydrogels. Polymer (Guildf). 118, 150–162 (2017).

182

29. Lin, H. & Freeman, B. D. Gas permeation and diffusion in cross-linked poly(ethylene glycol diacrylate). Macromolecules 39, 3568–3580 (2006).

30. Tummala, G. K., Rojas, R. & Mihranyan, A. Poly(vinyl alcohol) Hydrogels Reinforced with Nanocellulose for Ophthalmic Applications: General Characteristics and Optical Properties. J. Phys. Chem. B 120, 13094–13101 (2016).

31. Lira, M., Pereira, C., Real Oliveira, M. E. C. D. & Castanheira, E. M. S. Importance of contact lens power and thickness in oxygen transmissibility. Contact Lens Anterior Eye 38, 120–126 (2015).

32. Optical Properties of Transparent Materials: Transmission - Haze - Clarity. (BYK-Gardner Catalog 2010/2011).

33. Gonzalez-Meijome, J. M. et al. Refractive index and equilibrium water content of conventional and silicone hydrogel contact lenses. Ophthalmic and Physiological Optics 26, (Elsevier Science, 2006).

34. Childs, A. et al. Fabricating customized hydrogel contact lens. Sci. Rep. 6, 34905 (2016).

35. Varikooty, J., Keir, N., Woods, C. A. & Fonn, D. Measurement of the Refractive Index of Soft Contact Lenses During Wear. Eye Contact Lens Sci. Clin. Pract. 36, 2–5 (2010).

36. Patel, S., Marshall, J. & Fitzke, F. W. Refractive index of the human corneal epithelium and stroma. J. Refract. Surg. 11, 100–5

183

Summary and Outlook

185

Summary

This thesis describes our research on aliphatic poly(urethane-isocyanurate)

(PUI) networks, their synthesis, material properties and applications.

Generally speaking, PUI networks are considered promising materials for

biomedical applications. Specifically, the high toughness (both in the dry

state and in the water-swollen state), excellent optical properties and good

biocompatibility are highly attractive characteristics for biomedical

applications. In this thesis, first the viscosity characteristics of liquid aliphatic

isocyanurates are investigated. These are important industrial precursors for

the production of PUI networks. Next, a novel synthesis method for the

formation of well-defined PUI networks is presented. Using the newly-

developed synthesis method, various single and combinatorial PUI networks

are synthesized and characterized, demonstrating promising material

properties for biomedical use. Finally, the potential use of a specific type of

PUI networks based on poly(ethylene glycol) is explored for contact lens

applications.

Chapter 1 provides a brief description of the perspective in which the work

was performed. The scope, aims and outline of this thesis are presented.

In Chapter 2, a general introduction to isocyanate (NCO) and

polyurethane chemistry is provided. The unique reactivity of the isocyanate

functional group is discussed, as well as common reaction partners and

products. Next, various synthesis methods for the production of PUI

networks are reviewed. A general distinction can be made between the

conventional 2-component mixing approach and more recently-developed

1- and 1.5-component techniques. Finally, the general material properties

and typical applications of aliphatic PUI networks are discussed. It is shown

that PUI networks are generally appreciated for their versatility and high

toughness and durability. Owing to these properties, the largest application

field of aliphatic PUI networks is the coatings, adhesives and sealants

industry. At present, PUI networks are increasingly used for biomedical

applications and devices as well.

186

Isocyanate-functional aliphatic isocyanurates are important

industrial precursors for the 2-component synthesis of PUI networks. The

viscosity of these compounds is an important property for their production,

processing and application, but the underlying chemical interactions

dictating the viscosity of these liquids are poorly understood. In Chapter 3,

we report on the high-purity synthesis and isolation (>99%) of aliphatic

isocyanurate model compounds. The viscosities of these model compounds

are accurately quantified. By varying the carbon chain-length and comparing

isocyanate-functional isocyanurates to non-functional isocyanurates, the

effect of their molecular structure on viscosity is investigated. It is shown

that the viscosity of the liquid is significantly higher when NCO groups and

rings are simultaneously present and that viscosity increases for increasing

relative NCO- and ring content. Using density functional theory (DFT), the

presence of a strong bimolecular binding potential between the NCO group

and the isocyanurate-ring is identified for the first time. Using molecular

dynamics (MD) simulations, the viscosities of the model compounds were

accurately predicted, providing a more complete picture of the structure-

property-relations that dictate the viscosity of these liquids. The presence of

the newly-identified NCO-Ring interaction is shown to give rise to a strong

increase in viscosity.

In Chapter 4, we investigate the intermolecular interactions in

aliphatic isocyanurate liquids in more detail. We provide a first framework

of which specific intermolecular interactions are present in these liquids and

how they affect viscosity. Using DFT, we identify and quantify several

bimolecular interaction potentials involving NCO groups and isocyanurate

rings. It is shown that NCO-NCO interactions are relatively weak, with

interaction energies ranging from 5.8 to 7.9 kcal/mol. Ring-Ring interactions

are stronger, showing significant interaction energies of up to 13.8 kcal/mol.

NCO-Ring interactions are less predictable: interaction energies between 7

and 8 kcal/mol are found for single interactions while high interaction

energies increasing up to 17.6 kcal/mol are found when multiple,

cooperative interactions are possible. Using MD simulations, we show that

these multiple, cooperative NCO-Ring interactions have a strong increasing

effect on the viscosity of the liquid state. Thereby we provide a first

187

explanation for why the viscosity of aliphatic isocyanate-functional

isocyanurates is typically higher than that of non-functional isocyanurates or

other types of (poly)isocyanate oligomers.

In Chapter 5, a new synthesis method is presented for synthesizing

both hydrophilic and hydrophobic PUI polymer networks with a well-defined

network structure. The newly-developed synthesis method is based on the

functionalization of common polymeric diols using hexamethylene

diisocyanate (HDI) and their subsequent trimerization. It is shown that by

trimerizing NCO-functionalized prepolymers with a narrow molecular

weight distribution, PUI networks with a well-defined network structure can

be obtained. In contrast to conventional PUI networks, our PUI networks

demonstrate both a well-defined molecular weight between crosslinks (Mc)

and a well-defined functionality of the crosslinking units of exactly 3. These

PUI networks were shown to be flexible, tough materials both in the dry

state and in the water-swollen state.

In Chapter 6, we report on the synthesis of various combinatorial

PUI networks by the trimerization of mixtures of NCO-functionalized

prepolymers in solution. We show that combinatorial PUI networks

containing at least one hydrophilic PEG component show high water uptakes

of >100wt%. After swelling in water, the resulting hydrogels have elastic

moduli of up to 10.1 MPa, ultimate tensile strengths of up to 9.8 MPa,

elongation at break values of up to 624.0% and toughness values of up to

53.4 MJ·m-3. These values are exceptionally high and show that

combinatorial PUI hydrogels are among the toughest hydrogels reported in

the literature. Next, we show that the presence of a hydrophobic network

component significantly enhances the toughness and tensile strength of a

combinatorial PUI hydrogel in the hydrated state. This enhancement is the

largest when the hydrophobic network component is crystallizable in

nature. Finally, using dynamic scanning calorimetry, we show that PUI

hydrogels containing a crystallizable hydrophobic network component are

in fact semi-crystalline in the water-swollen state.

In Chapter 7, structure-property relations in semi-crystalline

combinatorial PUI hydrogels are investigated. Model combinatorial PUI

hydrogel networks, containing varying amounts of a crystallizable

188

hydrophobic network components, are synthesized and characterized.

These are compared to model combinatorial PUI hydrogel networks

containing varying amounts of an amorphous hydrophobic network

component. It is shown that the toughness of combinatorial PUI hydrogels

in the water-swollen state strongly increases with an increasing degree of

crystallinity of the hydrogel in the water-swollen state.

In Chapter 8, PUI networks based on poly(ethylene glycol) (PEG) are

presented as a new class of materials for contact lens applications. After

swelling in water, the PEG PUI hydrogels demonstrate high toughness values

ranging from 98 to 226 kJ·m-3 and elastic moduli ranging from 0.8 to 17.2

MPa for networks with equilibrium water contents (EWCs) ranging from 76.3

to 16.1wt%. The hydrogels demonstrate transmittance values >90% across

the visible spectrum, clarities close to 100% in most cases and refractive

indices ranging from 1.48 to 1.36. We show that these hydrogels are non-

cytotoxic, indicating good biocompatibility. Using bovine serum albumin as

a model protein, we show that PEG PUI hydrogel films demonstrate low

protein adsorption values, ranging from 0.24 mg/g to 0.65 mg/g after 24h.

These values are significantly lower than those of several commercial

contact lenses tested, which ranged from 0.55 mg/g to 1.38 mg/g after 24h.

The reported combination and range of material properties is considered

highly attractive for contact lens applications. Specifically, the contradictory

combination of high EWC, high toughness in the hydrated state and low

protein adsorption is exceptional and beneficial. These properties are

attributable to the PEG PUI network structure: the use of a PEG polymeric

backbone provides hydrophilicity and chemical inertness while the PUI type

crosslinking units provide high toughness in the hydrated state.

Outlook

Aliphatic PUI networks comprise a promising class of materials for

biomedical applications. The high toughness in both the dry state and the

water-swollen state associated with these materials is attractive for their

potential use as load-bearing biomaterials. Also, in combination with their

189

excellent optical properties, good biocompatibility and anti-biofouling

properties, PUI hydrogels are a promising class of materials for contact lens

applications. Before these materials can be applied in a biomedical context

though, further investigation and development is needed. This concerns

both industrial PUI precursors such as liquid aliphatic isocyanurates and PUI

networks and hydrogels in their final form.

Viscosity of liquid aliphatic isocyanurates

For a complete understanding of the structure-property-relations that

dictate the viscosity of liquid aliphatic isocyanurates a more detailed

experimental study is needed. Using specifically designed model

compounds, the molecular interactions and bimolecular complexes

proposed in Chapters 3 and 4 of this thesis could be experimentally

identified. For example, a detailed nuclear magnetic resonance (NMR) study

on high-purity model compounds can be considered to identify and quantify

the proposed NCO-NCO, NCO-Ring and Ring-Ring interactions

experimentally. Also, an X-ray diffraction study on crystalline small-

molecular NCO-Ring compounds can be considered to quantify the

intermolecular distances between the chemical moieties.

Regarding the computational part, the scope of the study could be

extended to include more (types of) (poly)isocyanates. Other industrially

relevant (poly)isocyanates such as methylenediphenyl diisocyanate (MDI),

toluene diisocyanate (TDI) and isophorone diisocyanate (IPDI) were not yet

included in this study and can be considered a good starting point. The

including of various other structure types is of high interest as well. For

example, allophanate, uretdione and iminooxadiazinedione type structures

are known to have significantly lower viscosities than the equivalent

isocyanurate type structures.12 A complete overview of all the relevant

intermolecular interactions present in these structures and their relative

strengths would be of great significance for the development of industrial

low-viscosity polyurethane precursors.

190

Network formation in aliphatic PUI networks

For an accurate study of PUI materials, having access to molecularly well-

defined PUI networks is important. Using the newly-developed synthesis

method reported in Chapter 5 of this thesis, the synthesis of PUI networks

with a variable, well-defined network structure is now possible. As this

synthesis method primarily provides a means for performing more accurate

structure-property-relations studies on PUI networks, an obvious next step

is to perform these studies.

An interesting example would be to investigate and quantify to what

extent the isocyanurate type crosslinking unit contributes to the mechanical

performance of PUI type networks.3 By comparing well-defined PUI

networks to PU networks containing other types of crosslinking units, the

specific contribution of the isocyanurate ring to the mechanical

performance of PUI networks can be targeted.

Aliphatic PUI networks for biomedical applications

At present, we consider biomedical materials and devices the most

promising novel application field for aliphatic PUI networks. Before these

materials can be applied in a biomedical context though, further

investigation and development of the materials is needed. In Chapter 8 of

this thesis, we already reported that a good biocompatibility of these

materials is indicated based on an indirect cytotoxicity test. This study can

be expanded to a more detailed cytotoxicity study using varying model cell-

lines and circumstances. Also, the biodegradation products should be

considered in such an experiment. Ultimately, the biocompatibility of these

materials should be investigated using an in-vivo assessment.

In Chapters 6 and 7 of this thesis, we have shown that the synthesis

of combinatorial PUI networks based on the trimerization of mixtures of

NCO-prepolymers can be used to produce exceptionally tough hydrogels

demonstrating toughness values in the range of several tens of kJ·m-3. Such

hydrogels can be applied as load-bearing biomaterials. For example, the use

of PUI type hydrogels for tissue engineering, artificial cartilage or artificial

implants can be considered. Before application in this context is possible, a

191

more detailed investigation of the physical properties of these combinatorial

PUI hydrogels is needed. An investigation of the compressive stress-strain

behaviour and the (repeated) load-bearing characteristics of these materials

can be considered. Also, the biodegradation process and rate of aliphatic PUI

networks is presently unknown, as are the biodegradation products and

their biocompatibility. A thorough in-vitro study, potentially followed by an

in-vivo and/or clinical study, should be performed before these materials

can be applied inside the human body.

Making use of the expertise and knowledge of PUI networks present

within the industrial coatings, adhesives and sealants industry, synergy can

be found with the field of biomedical materials and devices. For example,

PUI type materials can be optimized for use as (lubricious) coatings for

biomedical devices such as catheters and cannulas. The material properties

of the PUI networks reported in Chapters 6, 7 and 8 of this thesis are

promising characteristics in this respect. These include controllable water

uptake values, high toughness values (both in the dry state and in the

hydrated state), good biocompatibility, low biofouling and high stability with

respect to common sterilization techniques.

Finally, in Chapter 8 of this thesis we have demonstrated that PEG

PUI type hydrogels comprise a promising class of materials for contact lens

applications. Before commercialization in this context is possible though, the

detailed formulation of PEG PUI hydrogels should be optimized for this

specific application. This includes the optimization of the elastic modulus of

the hydrogel, optimization of the equilibrium water content and

optimization of the industrial production. Methods that are typically used to

manipulate the material properties of hydrogels may be considered, such as

the incorporation of mono-functional prepolymers and/or prepolymers with

a different chemical nature.

192

References 1. Ludewig, M., Weikard, J. & Stockel, N. Allophanate-containing

modified polyurethanes (US 2006/0205911 A1). (2006). 2. Widemann, M. et al. Structure-Property Relations in Oligomers of

Linear Aliphatic Diisocyanates. ACS Sustain. Chem. Eng. 6, 9753–9759 (2018).

3. Sasaki, N., Yokoyama, T. & Tanaka, T. Properties of isocyanurate-type crosslinked polyurethanes. J. Polym. Sci. Polym. Chem. Ed. 11, 1765–1779 (1973).

193

Acknowledgements

Dankwoord

Danksagung

195

Het tot stand komen van dit proefschrift was niet mogelijk geweest zonder

de hulp van vele mensen. Die wil ik hier graag bedanken voor hun rol, groot

of klein, in het proces.

Fistly, I would like to thank my promotors Dirk Grijpma and Dimitrios

Stamatialis. Your help and supervision are greatly acknowledged; without it

this thesis would not have existed.

Beste Dirk, heel erg bedankt voor al je hulp, advies, begeleiding,

sturing en geduld bij het tot stand komen van dit proefschrift. Als jij na het

eerste jaar niet vriendelijk doch dringend het onderzoek in de richting van

de materiaalontwikkeling had gestuurd was ik denk ik nu nog steeds

trimeren aan het synthetiseren. Ook wil ik je graag bedanken voor je

waardevolle adviezen omtrent het structureren van wetenschappelijk

onderzoek en wetenschappelijke teksten. Dat zijn inzichten om een leven

lang plezier van te hebben. Tot slot wil ik je nog een belofte doen. Ik. Zal.

Kortere. Zinnen. Maken. De niet-meer-dan-één-komma-per-zin-maatregel

heeft de leesbaarheid van dit boekje gered. Daar ben ik je dankbaar voor; en

de rest van de lezers ook.

Dear Dimitrios, thank you very much for all your help in the

structuring and fine-tuning of the manuscript. Your eye for detail and feeling

for organizing scientific text have been of great value. Also, your active

participation in conferences, colloquia and project meetings, including those

of TheLink, is much appreciated. I know that your family is not always spilling

with joy when you travel for work, but for us students your presence at these

events is very valuable. Please thank them as well for sharing you with us.

Secondly, I would like to thank the members of the graduation

committee for taking the time to read my thesis. Your effort to evaluate my

work are much appreciated.

Next I would like to thank my supervisors at Covestro, Leverkusen,

foremostly Dick Dijkstra and Frank Richter.

Beste Dick, heel erg bedankt voor de goede begeleiding en prettige

samenwerking gedurende mijn 4 jaar bij de Rheologie afdeling. Een

onuitputtelijke bron van kennis in de polymeerfysica aan het einde van de

196

gang is een luxe waar maar weinig PhD-studenten van kunnen genieten.

Bedankt voor het feit dat ik altijd kon aankloppen voor uitleg; of het nu ging

over de mechanische eigenschappen van polymeren, over wat de leukste

wijken en kroegen van Keulen zijn of over het doorgronden van de Duitse

cultuur vanuit een Nederlands perspectief.

Dear Frank, thank you very much for all your help and supervision,

both inside the lab and out. Your enthusiasm for chemistry (“Chemie muss

auch im freien Fall klappen” and “Never talk yourself out of an experiment”

are wisdoms I am not likely to forget soon) has been highly contagious and

has often motivated me to go the extra mile. Especially after having gelled

another HDI-trimerization or ruining one of your flasks again (sorry about

those)… But mostly I enjoyed our conversations that started with “Frank, do

you have any experience in working with …”, followed by some dodgy, dark

and forgotten corner of organic chemistry I had come across in an attempt

to realize one of my fantasy ideas. I was always amazed when you were able

to give me the exact state-of-the-art in that specific field from the top of

your head and direct me to the right papers and people I would need to find

my answer. Thank you for that; I am proud of the papers that we co-

authored on the chemistry of isocyanate-derivatives together.

Zunächst möchte ich mich gerne bei verschiedene Abteilungsleiter

bedanken für die persönliche Beratung und für die Unterstützung bei

meinem Projekt. Lieben Dank, dass eure Türe immer auf standen und ihr

euch immer bereit erklärt habt Zeit für mich zu finden.

Raul Pires, danke für deine immer gute Beratung und dein

Vertrauen in mich. Das hat mich stark motiviert und wird ich nicht schnell

vergessen. Dirk Achten, lieben Dank für deine positive Mentalität und gute,

kreative Ideen. Es hat mich immer Spaß gemacht bei dir vorbei zu kommen

um über Chemie (oder sonstiges ☺) zu quatschen. Péter Krüger, herzlichen

Dank für die Hilfe und Unterstützung in der Physikabteilung. Die gute

Atmosphäre innerhalb der Physik ist zum großen Teil zurückzuführen auf

deine Anerkennung des Personals. Ich (und vielen mit mich) habe mich

dementsprechend immer wohl gefühlt auf der 4. Etage in Q24. Karsten

Danielmeier, danke, für die Möglichkeit mal einen größeren internen

197

Projekt intensiv mitzumachen. Die Projektmanagement Skills die ich bei dir

abschauen konnte haben sich sehr nützlich erwiesen beim Fertigschreiben

der Doktorarbeit. ☺

Offensichtlich gibt es noch viele andere Covestro-kollegen die ich

erkenntlich bin. Lieben Dank an alle, für die super Atmosphäre, gute Ideen,

schöne Chemie, leckerer Kuchen, gute Witzen, schlechte Witzen,

Weihnachtsfeier, Sommerfeier, etc. etc. Insbesondere danke ich die

Kollegen aus der NCO-Runde, bzw. aus Q1, unter anderem: Dieter Mager,

Hans-Josef Laas, Christoph Eggert, Sebastian Doerr, Jan Suetterling, Florian

Golling, Rainer Trinks, Anne Kutz, Robert Maleika (Danke für al deine Hilfe

im Land des Patentrechts!), Heike Heckroth, Bernd Garska, Hans Grablowitz

und Serguei Kostromine. Natürlich gehören auch die Kollegen aus Q24 dazu,

unter anderem: Christoph Thiebes, Thomas Büsgen, Saskia Beuck, Andreas

Hecking, Yvonne Reimann, Daniel Raps, Walther Rauth, Mathias Matner,

Richard Meisenheimer und Joerg Tillack.

Auch vielen Dank an alle Laboranten in Q1, Q18 und Q24 die mich

die vergangene Jahren geholfen haben bei der Analyse (ich weiß, dass meine

Proben nicht immer die Einfachsten wahren… ☺). Insbesondere danke ich:

Steven Thiering für alle Hilfe im Labor Richter, Dirk Witschonke für die

vielen GPCs und GCs, Thomas Cüper, Edgar Eimermacher und Mischa

Simons für die rheologie, Brigitte Stechschulte, Kathrin Kemper und

Michaela Mallon für die Zugprüfungen, Dennis Turowski für deine

wunderschöne IR Messungen, Frank Ortmann und Sebastian Pinkert für die

thermische Messungen und die Kollegen bei Currenta in Q18 für die NMR

Messungen und Kristallstrukturanalyse. Auch war die administrative

Unterstützung von Marlis Beenen-Fuchs und Silke Köster von unschätzbare

Wert: vielen lieben Dank euch!

Of course I would also like to thank my colleagues at the University of

Twente for their input, ideas and help. Our conversations on the red couches

about anything but chemistry were always enjoyable and a nice distraction

from the scientific meetings. Ilaria Geremia and Kasja Skrzypek (I really

hope that that is spelled correctly… ☺), thanks for all the fun times that we

had together and for a great Christmas dinner. My visits to Twente were

198

always worth it when you guys were around. ☺ Also, thanks to Natalia

Chevtchik, Denys Pavlenko, Iris Allijn, Duco Schriemer and the rest of the

BST-group for the good times.

The same applies to my partners-in-crime within Covestro: the

group of “Covestro Doktoranden”. Thomas Prenveille, thanks for the good

talks and the bad jokes (yours mostly of course) and for several

(un)forgettable Covestro Weihnachtsfeier. Veronika Eilermann, es hat mich

sehr viel Spaß gemacht mit dir zusammen an dem PU-Buch zu Arbeiten.

Dank dafür und viel Erfolg in deinem Projekt. Anne Hansen, lieben Dank für

deine Interesse in meinem Projekt und deine Humor. Du hast gute Ideen

über Chemie und ich wünsche auch dich viel Erfolg bei deinem Projekt.

Also, many thanks to my fellow-PhD students in the consortium of

TheLink. I’ve really enjoyed our biannual meetings; both the project dinners

and the scientific progress meetings.

Wenn man sich mit ca. 20 verschwitzten deutsche Männer in eine Umkleide

antrifft, ist normalerweise irgendwas ganz schief gelaufen. Aber die 2.

Herren der RTHC Bayer Leverkusen sind darauf der Ausnahme: ich hätte

mich kein bessere verschwitzten deutsche Männer für sowas aussuchen

können! Ich danke euch für alle nette, sportliche, lustige, intensive,

lehrreiche, mit Kölsch angereicherte und immer angenehme

Hockeymomente, sowohl auf dem Platz als daneben. Es hat mich sehr viel

Spaß gemacht, euch die vergangene Jahren mal ein bisschen richtiges

Hockey beizubringen! ;)

Lieber Heike Heuser, Mischa Simons und Edgar Eimermacher, vielen Dank

für eure tägliche gute Laune und die angenehme Atmosphäre im Raum 424.

Ich hab mich vom Anfang bis Ende des Projekts bei euch im Büro zuhause

gefühlt und die vier Jahr sind vorbei geflogen. Es hat mich sehr viel Spaß

gemacht das Büro mit euch zu teilen und ich bedanke mich für die viele

angenehme Momente zusammen.

Das gleiche gilt natürlich für die PEAChT-bande: Tanja Hebestreit,

Andreas Hebestreit, Christoph Moethe und Edgar Eimermacher. Lieber

Tanja, vielen Dank für das hochhalten des Niveaus wenn es nötig war und

199

für die leckere Kaffee die wir immer bei dir genießen konnten (natürlich viel

leckerer als die in H4..!). Lieber Chris, vielen Dank für deine immer gute

Laune und für al deine Hilfe im Labor. Lieber Andreas, Tanja habe ich

bedankt für das Hochhalten des Niveaus wenn es nötig war; dich danke ich

für das Absenken wenn es nötig war! Das gemeinsame Essen in H4 und die

Kaffeekränzchen in Café Ratz haben mich immer viel Spaß gemacht. Ich wird

niemals mehr eine Zitronenkuche essen, ohne an dich zu denken.

Zum Schluss: für einen erfolgreichen Abschluss einer Doktorarbeit

ist neben der Doktorvater noch eine Person sehr wichtig: der Bürovater. Ich

bin sehr froh, dass ich es mit meine Bürovater nicht besser treffen haben

konnte. Lieber Edgar, vielen lieben Dank für deine Freundschaft und deine

Hilfe während die vergangene vier Jahren. Von alle Höhepunkte bei Covestro

waren die tägliche Höhepunkte in der Kantine von H4 doch immer noch die

Beste. Same table. Same time. Same jokes. Unter dem Genuss von den

typisch deutschen Nudeln mit Spinat und Fisch hast du mich mit viel Humor

die deutsche (zum Teil Eifeler?) Sprache und Kultur beigebracht und dafür

bin ich dich sehr dankbar. Insbesondere war dein Unterricht zum Thema

Flexibilität und Pünktlichkeit (Überpünktlichkeit?) beim Essen sehr lehrreich.

Auch das Bescheid Geben, dass man im Bayer-gefängnis landet, wenn man

zwei Schnitzels unter dem Reis versteckt, hat mich viele Monate Zeit

gespart. Also, mach es gut, viel Erfolg bei deinem letzten Jahr schuften (aber

lass bitte auch noch ein paar Proben für Mischa übrig, du hast mich selbst

gesagt: „es ist kein Kameradschaft, wenn nur den Kamerad schafft!“). Und

nachdem: genieß deine wohlverdiente Rente. Wenn du nochmal unterwegs

bist von oder nach Vrouwenpolder, schau dann gerne mal in Utrecht vorbei.

Für dich und deins steht der Tür immer auf und ein leckeres Hertog-Jan

bereit.

Ook al speelde het uitvoerende deel van dit onderzoek zich vooral in het

buitenland af, dit boekje had er niet kunnen liggen zonder de afleiding en

ondersteuning van vrienden vanuit Nederland. Velen van jullie hebben in de

afgelopen jaren de moeite genomen om mij (later ons) in Keulen op te

komen zoeken en dat waardeer ik enorm. De schnitzels en de braadworsten

kwamen me op een gegeven moment wel bijna de neus uit als we wéér een

200

weekendje Duitse tour van Brauhaus naar Brauhaus op de planning hadden.

Maar aangezien ze bij een enkeling van jullie letterlijk weer de neus uit

kwamen (ik zal geen namen noemen; die worst zal misschien niet helemaal

gaar geweest zijn denk ik. Of was het toch de Kölsch?) zal ik zeker niet klagen.

We hebben in ieder geval een hoop pret gemaakt in de weekenden dat we

bezoek hadden en ik bedank eenieder die langs is geweest voor het afleggen

van de trip.

Daarnaast dank aan mijn (oud)huisgenoten van DKP (o.a. Michiel

Marck, Sander Sinninghe Damsté; Marco van der Schoot; Harmen Booij,

Martijn Bekke; Jan-Wigbold Sluiter, Antonie van Eijk, Philip van Heemstra;

Jan-Willem de Bakker, Duco Klynstra; Bob ter Haar, Derk-Willem Duynstee;

Marijn Gelders) voor de geslaagde OHD’s, Oktoberdiners etc. Dank aan mijn

“vriendjes van thuis-thuis”: de C4 (Joris Wielders, Rogier Gundlach en Leon

Unk) voor de langlopende vriendschap. Dank aan mijn clubgenoten (Aus,

Arwin, Benni, Bob, Dré, Eijsvogel, Haas, Hans, Harmen, Jap, Jip, Joep, JW,

Langenberg, Michiel en Sloof) voor de vele huwelijken / feestjes /

housewarmings / Lumstrumthemakleurenbekendmakingsborrels en andere

belangrijke evenementen die de moeite waard waren om een weekend voor

op en neer te komen. Hetzelfde geldt natuurlijk voor vrienden van de ON

Hoogh / Mannendag (Ischa, Maurits, Pieter DJ, Kluis, Remmelt, Pang, Prick

en Ap). Bedankt voor alle activiteiten en mooie weekenden, met als kers op

de taart de Lumstrumeditie van de Mannendag op Chateau de la Noue de

Coq. Ik kijk nu al weer uit naar de volgende 10 jaar touwtrekken, vuur

maken, bijlen gooien, ver-plassen, ad-estafettes, …, …, etc., etc., etc.

Speciale dank gaat natuurlijk uit naar mijn paranimfen, Jeroen van Dorp en

Benjamin Thijs. Dit is echter het enige bedankje in dit dankwoord dat onder

voorbehoud geldt; eerst nog even zien hoe die verdediging straks loopt.

Maar dat komt vast goed, en natuurlijk bedank ik jullie wel graag nu al voor

jullie rol in de aanloop daar naar toe.

Jeroen, wij begonnen ons gezamenlijke studie-avontuur meteen

Samen op het Matje: we hadden in de eerste week van onze studie een of

ander natuurkunde-practicum gemist en mochten ons direct melden bij

Prof. Knoester voor uitleg. Al snel bleek dat wij met z’n tweeën ook de

201

voltallige intersectie vormden tussen onze studentenvereniging en de studie

scheikunde (wiskundig ook wel: VAP ∩ SK), wat toch een extra band schept.

Ik vind het erg leuk dat we door de jaren heen elkaars scheikundige

avonturen zijn blijven volgen en ik hoop dat we dat in de toekomst ook zo

blijven doen. Soortgenoten met dezelfde dubbele interesse voor sociale

interactie én chemie zijn per slot van rekening zeldzaam en ik ben blij dat we

elkaar daar in vinden.

Bennie, a.k.a. het Orakel van Groningen, de man die het antwoord

weet op alle vragen. Geen slechte eigenschap voor een paranimf. Maar dat

is natuurlijk niet de (enige) reden dat ik je aan mijn zijde gevraagd heb.

Tijdens onze vele gesprekken aan de bar in de Pintelier in Groningen was het

me een genoegen om samen met jou de ins & outs (en de tergend langzame

voortgang) van onze beide scripties te bespreken. Ik kijk met veel plezier

terug op onze discussies, waarin we samen de grenzen van de wetenschap

en van de bierkaart hebben verkend. Inmiddels liggen er nieuwe spannende

wetenschappelijke projecten in het verschiet en ik kijk er naar uit om ook

die samen te fileren onder het genot van een LaChouffe-Delirium-combo.

Lieve familie, ook al was ik de afgelopen jaren fysiek verder weg, jullie zijn

voor mij dichterbij geweest dan ooit. In mijn eentje in een vreemde stad met

een andere taal en een andere cultuur om me heen was het zeker niet altijd

makkelijk en op die momenten heb ik jullie steun gevoeld. Dankjewel dat

jullie zo lief hebben meegeleefd, meegedacht en meegevoeld.

Ik begon dit kopje aanvankelijk met “Lieve (schoon)familie”, maar ik

realiseerde me dat, zeker nu June er is, ik jullie met recht ook gewoon familie

mag noemen. Lieve Jacob en Sietske, dankjewel voor jullie gastvrijheid en

de warmte waarmee we altijd bij jullie thuis worden ontvangen. Ik voel me

inmiddels aan de Texellaan bijna net zo thuis als bij mijn eigen ouders. Dank

natuurlijk ook aan Pieter, Jelte, Haitske, Gesan en Dennis voor de vele

gezellige spelletjesmiddagen, Sinterklazen, kersten, oud-en-nieuws, etc.

Juul en Marthe, bedankt voor jullie gezelligheid, lekkere eten en alle

keren dat we bij jullie op de bank mochten slapen in Amsterdam als we

vanuit Keulen op bezoek waren! Mart, bovendien bedankt voor je gute

Laune en dito grappen (ook als de rest daar soms anders over denkt; niet

202

iedereen snapt wat humoor is). Bedankt, für die Blumen. Bedankt voor je

volledige vertrouwen in mij/ons, altijd en overal. Bedankt voor je interieur-

en mode-advies; gevraagd en ongevraagd. ☺ Bedankt voor al je support. Ik

zou me geen beter zusje kunnen wensen; en June zich geen betere tante.

Lieve Mars, dankjewel dat je altijd klaar staat. Je hebt altijd tijd om

even te bellen, ook wanneer het duidelijk is dat je het zelf druk hebt. Je bent

er altijd om even iets te overleggen. We kunnen altijd bij je terecht voor

advies, dat ook nog eens altijd goed is (jij zou consultant moeten worden ☺).

Dit boekje was nooit afgekomen zonder jouw suggesties over hoe ik het

project beter kon structureren in de beginfase. Dankjewel voor al het

meedenken, meekijken, nalezen en suggesties geven: altijd opbouwend, op

basis van en met ruimte voor eigen inbreng, maar altijd met verbetering tot

gevolg. The best kind.

Lieve pap, dankjewel dat je altijd de goede dingen weet te zeggen.

Wanneer ik worstel met het verwoorden van een ervaring of een gevoel,

weet jij soms in een paar zinnen te vangen wat ik bedoel. Dat heeft al een

hoop zaken in perspectief geplaatst en het helpt altijd bij het maken van

keuzes, of het reflecteren daarop. De gezellige (en lekkere!) etentjes waarbij

dit meestal plaatsvindt maken het plaatje compleet. Laten we er daar nog

veel van doen.

Lieve mam, dankjewel dat je altijd trots bent. En wat zou opa het

ook leuk gevonden hebben; die kijkt vast ergens aandachtig mee. Dankjewel

dat je altijd zo enthousiast bent over de dingen die wij doen en ons daarin

ondersteunt. De energie is aanstekelijk en motiveert om nog een stapje

extra te doen (daarin geef je natuurlijk zelf ook het goede voorbeeld). Lekker

ondernemen, geloven in en handelen vanuit eigen kracht, het ijzer smeden

wanneer het heet is en hoezo genoegen nemen met óf-óf zolang er

misschien ook nog én-én in het vat zit?? Het zijn credo’s om naar te leven en

ze zijn de output en kwaliteit van dit boekje zeker ten goede gekomen.

Lieve schat. Ik zou een tweede boek kunnen schrijven als ik alles zou

opschrijven dat ik je graag wil zeggen. Dat kan niet, maar ik geloof dat je het

meeste - de belangrijke dingen - ook onuitgesproken al wel weet. Toch zijn

er ook een paar dingen waar ik je graag nog expliciet voor wil bedanken.

203

Bedankt voor al je support de afgelopen jaren. Het was geen

makkelijke keuze om ineens voor vier jaar naar Duitsland te verhuizen en ik

heb me, zeker in het begin, regelmatig afgevraagd of ik er wel goed aan had

gedaan. Toen we na het eerste jaar merkten dat er door de afstand toch ook

wat druk op de relatie kwam te staan was het besluit heel dichtbij om ermee

te stoppen en terug te komen naar Nederland. Onze relatie was een offer

dat ik niet wilde maken, hoe leuk het werk ook, en daar sta ik nog steeds

volledig achter. Maar dat ik zou stoppen wilde jij niet, en dat jij toen besloten

hebt om jouw leven in Utrecht op pauze te zetten en voor mij naar Duitsland

te verhuizen is het liefste dat iemand ooit voor mij gedaan heeft. Zonder dat

besluit was dit boekje (dit hele project) er niet geweest. Dankjewel.

Vanaf dat jij naar Duitsland kwam werd alles leuker en ik heb enorm

genoten van al onze avonturen. Van het stapelbedje van Felix tot de Breite

Strasse (“I’m not famous around here..!”) tot ons fijne huis aan de

Rothgerberbach, wat hebben we een hoop pret gemaakt samen. Lekkere

avondjes borrelen thuis (meer oog voor de snacks dan voor de ring die ook

nog op tafel stond ☺), weekendjes weg, heel veel spellen spelen, lunchen bij

het graantentje, een kast bouwen, een prachtige vakantie naar Afrika, alles

is leuk zolang we het samen doen. Van alle mensen die ik ken kan ik met jou

het hardste lachen en ik ben superblij dat wij bij elkaar horen.

Het grote Duitsland-avontuur is inmiddels achter de rug, maar het

volgende is alweer begonnen. Lieve kleine June, wat fijn dat jij er bent; weet

dat er veel van je gehouden wordt.

Piet Driest

Januari 2020

204

List of Publications

• P.J. Driest, V. Lenzi, L.S.A. Marques, M.M.D. Ramos, D.J. Dijkstra, F.U.

Richter, D. Stamatialis and D. W. Grijpma; Aliphatic isocyanurates

and polyisocyanurate networks; Polym. Adv. Technol. 28 (2017),

1299-1304

• M. Widemann, P.J. Driest, P. Orecchia, F. Naline, F.E. Golling, A.

Hecking, C. Eggert, R. Pires, K. Danielmeier and F.U. Richter;

Structure−Property Relations in Oligomers of Linear Aliphatic

Diisocyanates; ACS Sustainable Chem. Eng. 6 (2018), 9753-9759

• C. Thiebes, (…) and P.J. Driest; Wasserfrei Härtende Klebstoffe auf

Polyisocyanatbasis, submitted (2018), EP18201522.2

• P.J. Driest, D.J. Dijkstra, D. Stamatialis and D.W. Grijpma; The

Trimerization of NCO-functionalized Prepolymers: an Effective

Method for synthesizing Well-defined Polymer Networks;

Macromol. Rapid Commun. 40 (2019), 1800867

• V. Lenzi, P.J. Driest, D.J. Dijkstra, M.M.D. Ramos and L.S.A. Marques;

GAFF-IC: Realistic Viscosities for Isocyanate Molecules with a GAFF-

Based Force Field; Mol. Simulat. 45 (2019), 207-214

• E. Avtomonov, K. Danielmeier, P.J. Driest et al.; Chemical Principles;

Chapter 3 in: Polyurethanes – Coatings, Adhesives and Sealants (U.

Meier-Westhues et al.); Vincentz-network, Hannover, 2019

• P.J. Driest, F.E. Golling, I.C. Latorre-Martinez, R. Pires and D. Achten;

Crosslinking technologies based on the polyaddition reactions of

polyisocyanates; Chapter 4.6 in: Polyurethanes – Coatings,

Adhesives and Sealants (U. Meier-Westhues et al.); Vincentz-

network, Hannover, 2019

205

• V. Lenzi, P.J. Driest, D.J. Dijkstra, M.M.D. Ramos and L.S.A. Marques;

Investigation on the Intermolecular Interactions in Aliphatic

Isocyanurate Liquids: Revealing the Importance of Dispersion; J.

Mol. Liq. 280 (2019), 25-33

• P.J. Driest, D. Kleinschmidt, R. Meisenheimer and D. Achten;

Storage-stable Thermolatent Catalysts for the Polymerization of

Isocyanates; “Erfindungsmeldung“ submitted (2019)

• P.J. Driest, V. Lenzi, L.S.A. Marques, M.M.D. Ramos, F.U. Richter and

D.J. Dijkstra; The Influence of Molecular Interactions on the Viscosity

of Aliphatic Polyisocyanate Precursors for Polyurethane Coatings;

ECS Conference Proceedings (2019)

• P.J. Driest, I.E. Allijn, D.J. Dijkstra, D. Stamatialis and D.W. Grijpma;

Poly(ethylene glycol) Poly(urethane-isocyanurate) Hydrogels for

Contact Lens Applications; Polym. Int. (in press), DOI:

10.1002/pi.5938

(To be) submitted

• P.J. Driest, D.J. Dijkstra, D. Stamatialis and D.W. Grijpma; Tough

Combinatorial Poly(urethane-isocyanurate) Polymer Networks and

Hydrogels Synthesized by the Trimerization of Mixtures of NCO-

Prepolymers; resubmitted

• P.J. Driest, D.J. Dijkstra, D. Stamatialis and D.W. Grijpma; Structure-

Property Relations in Semi-crystalline Combinatorial PUI Hydrogels;

to be submitted