Accepted Manuscript

Crosslinking of an Ethylene-Glycidyl methacrylate copolymer with amine click

chemistry

Massimiliano Mauri, Nina Tran, Oscar Prieto, Thomas Hjertberg, Christian Müller

PII: S0032-3861(17)30010-1
DOI: 10.1016/j.polymer.2017.01.010
Reference: JPOL 19327

To appear in: Polymer

Received Date: 26 September 2016

Revised Date: 4 January 2017
Accepted Date: 8 January 2017

Please cite this article as: Mauri M, Tran N, Prieto O, Hjertberg T, Müller C, Crosslinking of an
Ethylene-Glycidyl methacrylate copolymer with amine click chemistry, Polymer (2017), doi: 10.1016/
j.polymer.2017.01.010.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.
 ACCEPTED MANUSCRIPT
Graphical Abstract

Efficient crosslinking of a branched statistical ethylene-glycidyl methacrylate copolymer with

bifunctional amine crosslinking agents is demonstrated. The use of click chemistry opens up a

by-product free alternative to traditional crosslinking with peroxides. A well-adjusted

processing window around 120 °C permits extrusion of the copolymer and curing agent,

PT
followed by crosslinking at 160 to 200 °C.

RI
Keywords: epoxy, glycidyl methacrylate, amine, crosslinking, polyethylene copolymer

SC
Massimiliano Mauri, Nina Tran, Oscar Prieto, Thomas Hjertberg and Christian Müller

U
Crosslinking of an Ethylene-Glycidyl Methacrylate Copolymer with Amine Click
AN
Chemistry
M
D
TE
C EP
AC
 ACCEPTED MANUSCRIPT

Crosslinking of an Ethylene-Glycidyl Methacrylate Copolymer

with Amine Click Chemistry

Massimiliano Mauri,a Nina Tran,a Oscar Prieto,b Thomas Hjertbergb and Christian Müller*a

a
Department of Chemistry and Chemical Engineering, Chalmers University of Technology, 41296

PT
Göteborg, Sweden

RI
b
Innovation & Technology, Borealis AB, 44486 Stenungsund, Sweden

SC
*e-mail: christian.muller@chalmers.se

U
AN
Abstract

Commonly used crosslinking methods for polyethylenes result in the release of harmful
M

by-products. Here, we demonstrate that an epoxy-bearing polyethylene copolymer,

D

which contains 8 wt% glycidyl methacrylate, can be efficiently crosslinked without by-
TE

product formation. Click chemistry based on multifunctional amine curing agents,

which carry at least two functional groups separated by a flexible spacer, was used to
EP

prepare thermosets. Compounding of the crosslinker and copolymer through extrusion

at 120 °C could be carried out without onset of the curing reactions. Careful adjustment
C

of the curing time and temperature, ranging from 20 to 120 min and 160 to 200 °C,
AC

resulted in a high network density of at least 2.8 crosslinks per 1000 carbons at a curing

agent stoichiometry of as little as 0.5 wt%.

Keywords: epoxy, glycidyl methacrylate, amine, crosslinking, polyethylene copolymer

1 / 27
 ACCEPTED MANUSCRIPT
1. Introduction

Polyethylenes constitute the most widely used family of commodity polymers. For

electrical and medical applications were a particularly clean material is required, low density

polyethylene (LDPE) typically is the material of choice because it can be polymerised at high

pressure with high cleanliness and purity. However, LDPE is characterised by a lower

PT
crystallinity than linear polyethylenes and features a much lower onset of melting. Thus, it is

RI
common to crosslink LDPE in order to ensure dimensional stability at elevated temperatures,

to prevent stress cracking, and to improve the resistance towards chemicals.[1, 2]

SC
Crosslinking can be achieved through a number of different agents, inculding silanes, azo-

compounds and peroxides, as well as treatment with e.g. β- or γ-radiation.[3-6]

U
For the electrical insulation of e.g. medium and high-voltage power cables crosslinking
AN
with dicumyl peroxide is the most common process, which however results in unwanted by-
M

products.[7, 8] In particular methane and water must be removed from the crosslinked

polyethylene (XLPE) insulation layers, which typically involves a time and energy intensive
D

degassing step at elevated temperature.[8, 9] Moreover, peroxides and their decomposition

TE

by-products pose a considerable health hazard, which requires a suitably adapted work

environment. Therefore, future development of power cable technology will benefit from
EP

alternative crosslinking concepts for polyethylene that do not involve volatile by-products.
C

One attractive alternative is crosslinking with the help of epoxy functional groups.
AC

Epoxy crosslinking is a versatile process since epoxy resins offer excellent bonding capability

and, once cured, display outstanding properties in terms of mechanical strength as well as

chemical and electrical resistance.[10, 11] The crosslinking process can be controlled by

combining different catalysts and curing agents with tailor-made epoxy resins, which can be

tuned according to the desired property portfolio.[12, 13] In order to convert epoxy resins into

2 / 27
 ACCEPTED MANUSCRIPT
solid, infusible thermoset networks it is necessary to use crosslinking agents, also referred to

as crosslinkers, hardeners or curing agents.[11] Epoxy resins can, for example, be cured with

amines,[14-16] thiols,[17-19] carboxylic acids[20] and hydrazides.[21, 22] Nucleophilic

substitution on small tensioned rings like epoxides is a well-known type of click chemistry

reaction that involves only one step and releases no by-products. Moreover, epoxy-opening

PT
click chemistry reactions (Figure 1) are insensitive toward moisture and oxygen. A high

RI
thermodynamic driving force rapidly and irreversibly leads to a high yield of a single reaction

product. The efficiency of the crosslinking process strongly depends on the temperature,

SC
curing time and stoichiometry.

In this work we study the use of different crosslinking agents in combination with a

U
AN
branched statistical ethylene-glycidyl methacrylate copolymer P(E-stat-GMA). This type of

copolymer is widely employed as a reactive compatibiliser for polymer blends.[23-26] Here,

M

we use this material as a model system to explore the possibility of click chemistry type

crosslinking of epoxy-bearing polyethylene copolymers. After a broad initial screening of

D

several amines, carboxylic acids and hydrazides, we focus on amines, evaluate their use as
TE

crosslinking agents and demonstrate that efficient network formation without the release of

undesirable by-products can be achieved.

C EP

2. Experimental
AC

Materials. A branched statistical ethylene-glycidyl methacrylate copolymer P(E-stat-GMA)

with a GMA content of 8 wt% was obtained from Arkema (Lotader series; 8 wt% GMA

correspond to a monomer ratio of about 60:1 E:GMA). The copolymer had a weight-average

molecular weight ~ 85 kg mol-1 and polydispersity index PDI ~ 7, determined with size

exclusion chromatography (SEC) in 1,3,4-trichlorobenzene at 150 ºC using an Agilent PL-

3 / 27
 ACCEPTED MANUSCRIPT
GPC 220 system equipped with a refractive index and viscosity detector; calibrated with

polystyrene standards using universal calibration. The melt flow index was MFI ~ 10 g/10

min (provided by supplier). Regular LDPE was obtained from Borealis ( ~ 117 kg mol-1,

PDI ~ 9, number of long-chain branches ~ 1.9 per 1000 carbon atoms). The crosslinking

agents 1,8-diaminooctane (DAO), N,N'-dimethyl-1,8-octanediamine (DMAO),

PT
trimethylolpropane tris[poly(propylene glycol) amine terminated] ether (TMPTA), 1-

RI
tetradecylamine (TDA), octanoic hydrazide (OAD), adipic acid dihydrazide (AAD) and 1,10-

decanedicarboxylic acid (DDA) were purchased from Sigma Aldrich and used without further

SC
purification.

Compounding, crosslinking and sample preparation. Copolymer/crosslinking agent

U
formulations were compounded through extrusion for 10 minutes at 120 °C. The extruded
AN
material was first molten at 130 °C, followed by crosslinking at 160 to 200 °C and a pressure
M

of 25 bar for 20 to 120 min, resulting in 1.5 mm thick plates. Thin films for optical

microscopy and FTIR were prepared by drop-casting from 10 g L-1 hot p-xylene solutions and
D

melt pressing, respectively.

TE

Hot set test. Dumbbell-shaped, crosslinked copolymer samples with an initial gauge length of

~ 20 mm were elongated for 10 min at 200 °C by applying a weight that corresponded to a

EP

stress of σ ~ 0.2 MPa. The final length was measured to calculate the hot set extension

= / = − 1. After the measurement was

C

ratio and hot set elongation

AC

performed, the weight was removed and the samples were allowed to recover in the oven for 5

minutes, and then at room temperature for 1 hour. The permanent elongation was used

to measure the permanent extension ratio = / and permanent elongation

= − 1.

4 / 27
 ACCEPTED MANUSCRIPT
Gel-content. The gel-content of cross-linked samples was determined gravimetrically using a

solvent extraction technique. The samples (~150 mg) were placed in pre-weighed 100 mesh

stainless steel baskets and extracted in 1.1 dm3 by refluxing decalin for 6 h. An antioxidant,

10 g Irganox 1076 from Ciba-Geigy, was added to the solvent to prevent degradation. Then,

the solvent was exchanged with 0.9 dm3 of additive free, pre-heated decalin and the extraction

PT
continued for 1 h. Finally, the samples were dried first at ambient overnight and then under

RI
vacuum for about 8 hours at 50 °C. After this period the non-soluble fraction that remained in

the basket reached a constant weight, which was used to calculate the gel-content.

SC
Polarised light microscopy. Transmission optical microscopy was carried out with a Carl

Zeiss A1 optical microscope.

U
Differential Scanning Calorimetry (DSC). DSC was carried out under nitrogen between -50
AN
to 160 ºC at a scan rate of 10 ºC min-1 using a Mettler Toledo DSC2 calorimeter equipped
M

with a HSS7 sensor and a TC-125MT intracooler. Second heating scans were recorded after

cooling at 10 ºC min-1 from 160 °C. The sample weight was 2 to 3 mg.
D

Thermogravimetric Analysis (TGA). TGA was carried out under nitrogen between 40 and
TE

500 ºC at a scan rate of 10 ºC min-1 using a Mettler Toledo TGA/DSC 3+.

FTIR spectroscopy. Measurements with a resolution of 4 cm-1 were performed on thin films
EP

with a Perkin Elmer 2000 FTIR equipped with a beam condenser.

C

Dynamic mechanical analysis (DMA). Samples with a size of 15×5×1.25 mm were cut from
AC

melt-pressed films. DMA was carried out between 40 to 150 ºC at a heating rate 3 ºC min-1, a

frequency of 0.5 Hz and 1% strain using a TA Q800 DMA.

5 / 27
 ACCEPTED MANUSCRIPT
3. Results and Discussion

The epoxy groups of the here studied ethylene-glycidyl methacrylate copolymer (Fig.

1a) can undergo a polyaddition reaction with e.g. amine, carboxyl and hydrazine functional

groups. In order to convert the copolymer to a thermoset network it is necessary to use a

PT
multifunctional crosslinking agent that can link adjacent polymer chains through reaction with

two or more epoxy groups. We chose candidates that facilitate a sufficiently large processing

RI
window of the resin without the risk of crosslinking. A convenient temperature range for

extrusion of LDPE is 120 to 140 °C. Therefore, we aimed for crosslinking agents that only

SC
react above 160 °C. Moreover, we required candidates to be (1) non-toxic, (2) liquid at the

U
extrusion temperature and curing temperature (to enable efficient compounding and avoid
AN
sublimation, respectively), (3) equipped with two or more epoxy-reactive functionalities, and

(4) likely to react rapidly with the oxirane ring without by-product formation. The selected
M

curing agents and their properties are summarised in Table 1. The amine-based curing agents

DAO, MDAO and TMPTA are most promising since they meet every requirement. DAO and
D

MDAO carry two primary and secondary amines, respectively, and the melting temperature
TE

can be adjusted by changing the length of the aliphatic spacer with an octyl segment resulting
EP

in a well adjusted ~ 40 °C and 50 °C. TMPTA carries three primary amines and is liquid at

room temperature. Mono-functionalised TDA was selected despite a low boiling point for
C

comparison with similar two-headed crosslinking agents like DAO or MDAO. Among
AC

carboxylic acid curing agents DDA was selected, which featured a slightly higher melting

point of 130 °C compared to amine-based curing agents. The choice of suitable and

commercially available hydrazide-based curing agents was more limited. We selected OAD

and AAD, which meet close to all requirements, for initial screening.

6 / 27
 ACCEPTED MANUSCRIPT
In a first set of experiments we evaluated the crosslinking efficiency of the selected

curing agents by comparing the elongation of copolymer samples at 200 °C. For this hot-

set test the copolymer was crosslinked with 1 wt% curing agent at 200 °C for two hours. The

network density was calculated based on the molecular weight between crosslinks

according to the affine network model that is applicable to elastomers:[27]

PT
!
2
= +
( − )
(1)

RI
where = 0.754 g cm-3 is the density of LDPE at 200 °C, R is the gas constant, T is the

SC
temperature, is the true stress and = / is the hot set extension ratio at 200 °C.

Here it is interesting to note that industrial standards for electrical insulation demand a hot set

elongation of = − 1 < 75 % for

U
= 0.2 MPa.[28] We find that the amine-based
AN
curing agents DAO, MDAO and TMPTA give rise to a low < 25 % (Fig. 2a), which
M

corresponds to 13, 10 and 4 crosslinks per 1000 carbons, respectively ( ~ 1.1, 1.3 and 3.5

kg mol-1). The carboxylic acid-based curing agent OAD gave rise to an elongation of just over
D

54% (~3 crosslinks per 1000 carbons), whereas the use of hydrazine-based AAD resulted in
TE

rupture. We ascribe the inferior curing efficiency of OAD and in particular AAD to the

relatively high (cf. Table 1), which for both materials lies above the extrusion temperature
EP

of 120 °C and complicates homogenisation.

C

Encouraged by the promising behaviour of the amine-based curing agents, we chose to

AC

focus on DAO, MDAO and TMPTA and carried out a detailed analysis of their crosslinking

performance. For all amine-crosslinked samples we note that long exposure to high

temperatures and ambient atmosphere gives rise to a yellow colour, which may arise due to

oxidation of the curing agents (Fig 3.). Compared to the neat copolymer the crosslinked

samples appeared more transparent. Polarised light microscopy of drop-casted copolymer

7 / 27
 ACCEPTED MANUSCRIPT
films revealed a less well-developed spherulitic microstructure upon curing with TMPTA

(Fig. 4a). We rationalise this observation with a reduction in crystal growth rate due to the

presence of a large number of network points, which leads to smaller domains that give rise to

less light scattering and thus increased transparency.

In order to gain additional insight with regard to the crystallinity and microstructure of

PT
crosslinked copolymer formulations we carried out differential scanning calorimetry (DSC).

RI
The crystallinity X of P(E-stat-GMA), was calculated from second DSC heating thermograms

according to the total enthalpy method:[29]

SC
∆$%
" = , )* (+
∆$% − &, -(' ( −' )/
(2)
.

U
where ∆$% is the enthalpy of fusion calculated from the integrated area under DSC melting
AN
endotherms, ∆$% = 290 J g-1 is the enthalpy of fusion of 100 % crystalline PE and ' (
M

)* (+
and ' are the heat capacity of the amorphous crystalline phase, respectively. The limits

of integration were chosen to be ! = 20 °C and = 120 °C. For all amine curing agents, we
D

found that crosslinking does not strongly impact the crystallinity, which decreases from 40 %
TE

for the neat copolymer to a slightly lower value of 32 to 35 % after a curing time of two hours
EP

at 200 °C.

We observe a decrease in peak melting temperature from ~ 111 °C for the neat
C

copolymer to about 100 °C after crosslinking at 200 °C for two hours (Fig. 4). The associated
AC

reduction in the thickness of crystalline polyethylene lamellae can be calculated according to

the Gibbs-Thomson equation:

2
0 = ∙
∆$% −
(3)

8 / 27
 ACCEPTED MANUSCRIPT
where = 90.4 mJ m-2 is the fold surface energy, = 1.0 g cm-3 and = 418.6 K denotes

the equilibrium melting temperature of polyethylene. We find that the lamellar thickness that

corresponds to the peak decreases from 0 ~ 7.4 nm for the neat copolymer to e.g. 0 ~ 5.4

nm for DAO-cured material after crosslinking at 200 °C for two hours.

Thermogravimetric analysis (TGA) allowed us to test for volatile by-products. We

PT
selected a TGA heating profile that mimics the entire curing process (two hours at 180 °C)

RI
and compared the results with standard dicumyl peroxide (DCP) crosslinking of regular

LDPE using 2 % DCP (Fig. 4). All samples were dried in vacuum at 50 °C prior to TGA

SC
measurements. For neat LDPE and P(E-stat-GMA), as well as P(E-stat-GMA) that contained

U
DAO, MDAO and TMPTA, we observe a steady TGA thermogram up to 300 °C with a
AN
baseline drift of not more than 0.4 wt% during the isotherm at 180 °C. Evidently, no release

of by-products could be detected in case of amine curing. In contrast, DCP crosslinking gave
M

rise to a weight loss of 1.4 % due to the release of by-products that had formed during the

crosslinking reaction, namely water, methane, acetophenone, cumyl alcohol and α–methyl
D

styrene.[7]
TE

We then went on to study the influence of the reaction conditions on the crosslinking
EP

process with regard to curing time (10 to 120 minutes), temperature (160, 180 and 200 °C)

and the stoichiometry of reactants. It can be anticipated that an even ratio of reaction sites
C

leads to efficient crosslinking. For DAO, MDAO and TMPTA a 1:1 epoxy:N-H stoichiometry
AC

corresponds to 2, 4.5 and 4 wt% of the curing agent, respectively. Thus, we chose to vary the

curing agent concentration from 0.1 to 1 wt% for DAO and MDAO and from 0.15 to 2 wt%

for TMPTA.

To investigate the degree of crosslinking of the cured samples, gel content

measurements and hot set tests were carried out for all formulations and crosslinking

9 / 27
 ACCEPTED MANUSCRIPT
parameters. Initially, we explored the impact of curing temperature for a constant curing time

of two hours (Fig. 5a). A gel content G > 70 % was reached regardless of the temperature

(160 to 200 °C) when using at least 0.5 wt% DAO and MDAO or 1 wt% TMPTA. However,

the same formulations performed poorly when subjected to the hot set test, and mostly

ruptured. This behaviour can be explained by considering the epoxy:N-H stoichiometry,

PT
which has a value of 4:1 for 0.5 wt% DAO and 1 wt% TMPTA, and 9:1 for 0.5 wt% MDAO.

RI
The resulting network is evidently not dense enough to prevent significant elongation at 200

°C. A doubling of the epoxy:N-H ratio led, upon curing for two hours, to a further increase in

SC
gel content that was accompanied by a considerable decrease in hot set elongation.

In a further set of experiments we studied the impact of curing time on formulations

U
AN
containing 1wt% DAO, 1 wt% MDAO and 2 wt% TMPTA, using a constant curing

temperature of 200 °C (Fig. 5b). We observed a rapid increase in gel content and decrease in
M

hot set elongation, owing to the high reactivity of epoxies and amines. For instance, DAO and

MDAO formulations displayed a G > 80 % and ~ 20 % after a curing time of 20 minutes.

D

The network density continues to increase linearly with time, reaching a value of more than
TE

10 crosslinks per 1000 carbons after two hours. Although DAO offers twice as many reactive

hydrogens than MDAO, their crosslinking performance is similar for the concentration of
EP

curing agent. We propose that upon reaction of a primary amine with an oxirane ring, the thus
C

generated secondary amine is more likely to react with a closeby oxirane ring from the same
AC

polymer chain due to a combination of steric hindrance and proximity. Therefore, curing

agents that carry primary or secondary amines yield a similar number of network points.

Crosslinking with TMPTA proceeded more slowly, reaching ~ 20 % only after 90

minutes, which corresponds to a network density of less than 4 crosslinks per 1000 carbons.

The lower curing rate of TMPTA formulations can be rationalised with the the larger size of

10 / 27
 ACCEPTED MANUSCRIPT
this crosslinking agent as well as increased steric hindrance of the amine group due to the

adjacent methyl group at the α position. We argue that the difference in reactivity between the

primary and secondary amine does not depend on the basicity because steric hindrance is a

bigger contributor in an SN2 reaction type, and it affects the strength of the nucleophile.

Primary amines are known to be more reactive with epoxy rings compared to secondary

PT
amines, despite being less nucleophilic.14,38,39 Furthermore, taking into consideration the two

RI
different primary amines that are the object of this study, the positive inductive effect of the

α-methyl chain gives TMPTA more basic and nucleophilic features compared to DAO. These

SC
properties however do not lead to a faster reaction rate, showing how the steric hindrance is

once again the major contributor.

U
AN
To demonstrate that a multi-headed curing agent is needed to efficiently crosslink the

here investigated ethylene-glycidyl methacrylate copolymer, we prepared formulations with

M

1-tetradecylamine (TDA), which only carries one primary amine. We compared TDA with

DAO for a 4:1 epoxy:amine stoichiometry, i.e. 3 wt% TDA and 1 wt% DAO. Curing for 45
D

minutes at 200 °C resulted in a high G ~ 90 % in case of both TDA and DAO. However, the
TE

mono-functionalised crosslinker displayed a significantly larger hot set elongation of ~

120 %, indicating a much lower network density of ~ 2.6 crosslinks per 1000 carbons. These
EP

results indicate that both hydrogens of a primary amine are able to react with an epoxy group,
C

although the reactivity of the second hydrogen is limited by the considerable steric hindrance
AC

that arises once the first hydrogen has reacted with an epoxy.

To gain additional insight into the rate of epoxy-ring opening, we followed the curing

process of all three amine-based formulations with Fourier transform infrared spectroscopy

(FTIR). We observe two characteristic absorptions by the oxirane ring at 3050 and 911 cm-1,

which we attribute to C-H tension of the methylene bridge of the epoxy ring and C-O

11 / 27
 ACCEPTED MANUSCRIPT
deformation of the oxirane group, respectively. Since the first band is located close to a more

prominent O-H absorption band, we used the C-O signal to follow the curing reaction. We

calculated the number of remaining epoxy rings, using the carbonyl peak at 1750 cm-1 for

normalisation, while the number of unreacted N-H group was extrapolated from the

stoichiometry of the reaction (Fig. 6). For a curing temperature of 200 °C both DAO and

PT
MDAO formulations initially feature a rapid drop in the number of epoxy groups by 40 % and

RI
20 %, respectively. Since DAO contains twice as many reactive N-H hydrogens than MDAO,

the former initially consumes twice as many epoxy rings (crosslinking time ~ 20 to 40 min).

SC
For longer crosslinking times we observe a more gradual decrease and eventually

consumption of most reactive N-H functionalities after two hours of curing. The nonlinear

U
consumption of epoxy rings is in contrast to the linear increase in network density as
AN
calculated from hot set elongation experiments (cf. Fig. 5). Permanent network points can
M

arise due to either an inter-chain crosslink or a combination of inter- and/or intra-chain

crosslinks that trap an entanglement (Fig. 7). We note that P(E-stat-GMA) contains
D

approximately 8 glycidyl methacrylate monomers per 1000 carbons. Samples cured for two
TE

hours at 200 °C with either DAO or MDAO feature a much higher number of about 12 and 10

network points per 1000 carbons, which exceeds the number of oxirane rings (cf. Fig. 5b).
EP

This strongly indicates the presence of a substantial number of trapped entanglements.

Therefore, we explain the linear increase in the number of network points despite the initially
C
AC

rapid consumption of epoxy rings with the need for a sufficiently high number of crosslinks in

order to trap entanglements. Intra-chain crosslinks only start to contribute to the network

density at longer curing times. Instead, TMPTA consumes oxirane rings more slowly but

linearly, with 30 % reactive N-H hydrogens remaining after two hours of curing. The lower

epoxy:N-H reaction rate likely arises due to the larger size of the curing agent as well as

increased steric hindrance.

12 / 27
 ACCEPTED MANUSCRIPT
In a final set of experiments we used dynamic mechanical analysis (DMA) to

characterise the mechanical properties of fully crosslinked samples. We compared

formulations that had been cured at 200 °C for two hourse with 1 wt% DAO, 1 wt% MDAO

and 2 wt% TMPTA as well as for 45 min with 3 wt% TDA, for which we calculate a network

density of 12, 10, 6 and 2.6 crosslinks per 1000 carbons using the affine network model

PT
(equation 1). The storage modulus for the cured formulations has a lower value of 40-50 MPa

RI
as compared to 70 MPa in case of neat P(E-stat-GMA) due to the lower crystallinity of the

samples after curing. Above the melting temperature the crosslinked samples display a

SC
distinct rubber plateau. The plateau modulus 2 is related to the molecular weight between

crosslinks by:

= /2
U (4)
AN
which we calculated at T = 140 °C using a density of = 0.789 g cm-3 for LDPE at this
M

temperature. For the DAO, MDAO, TMPTA and TDA cured samples we obtain a value of

~ 0.9, 1.3, 1.2 and 3.4 kg mol-1, which corresponds to a network density of about 15, 10, 12
D

and 4 crosslinks per 1000 carbons. We note a good agreement with the network density
TE

calculated for DAO, MDAO and TDA using the affine network model but a mismatch for
EP

TMPTA, which we explain with the three-functional nature of this crosslinker.

C

4. Conclusions
AC

We have explored the suitability of epoxy ring opening reactions for curing of an

ethylene-glycidyl methacrylate copolymer with the aim to develop a by-product free

crosslinking alternative for e.g. dicumyl peroxide. We screened three types of curing agents,

i.e. amines, carboxylic acids and hydrazides, and found that in particular amines offer a

promising crosslinking performance. Multifunctional curing agents that carry at least two

13 / 27
 ACCEPTED MANUSCRIPT
amines were necessary to obtain efficient crosslinking. The choice of amine, i.e. primary or

secondary, appeared to have little effect on network formation or the rate of crosslinking. In

contrast, the epoxy:N-H stoichiometry, curing time and temperature were found to

significantly influence the crosslinking efficiency. Overall, in case of curing agents that

featured two amines separated by an aliphatic spacer a concentration of only 0.5 wt% was

PT
needed to achieve satisfactory network formation, leading to a hot set elongation of less than

RI
100 %, i.e. a network density of more than 2.8 per 1000 carbons (Fig. 9). The cured

specimens exhibited outstanding mechanical strength, optical transparency as well as heat

SC
resistance. We conclude that epoxy-amine curing is a viable, by-product free and hence more

sustainable alternative to peroxide-based crosslinking of polyethylene based resins.

U
AN
5. Acknowledgements
M

Borealis AB is acknowledged for funding. We thank Mattias Andersson for his tireless
D

experimental support and Anders Mårtensson for help with SEC measurements.
TE
EP

References

[1] C. Beveridge, A. Sabiston, Methods and benefits of crosslinking polyolefins for industrial
C

applications, Mater. Design 8 (1987) 263-268.

AC

[2] I. Chodak, Properties of Cross-Linked Polyolefin-Based Materials, Prog. Polym. Sci. 20

(1995) 1165-1199.

[3] M. Lazar, R. Rado, J. Rychly, Cross-Linking of Polyolefins, Adv. Polym. Sci., Polymer

Physics, Springer, Heidelberg, 1990, pp. 149-197.

14 / 27
 ACCEPTED MANUSCRIPT
[4] G. Tillet, B. Boutevin, B. Ameduri, Chemical reactions of polymer crosslinking and post-

crosslinking at room and medium temperature, Prog. Polym. Sci. 36 (2011) 191-217.

[5] R. Patterson, A. Kandelbauer, U. Müller, H. Lammer, Crosslinked Thermoplastics, in: H.

Goodman, H. Dodiuk-Kenig (Eds.), Handbook of Thermoset Plastics, Elsevier, 2014.

PT
[6] M.G. Andersson, M. Jarvid, A. Johansson, S. Gubanski, M.R.S. Foreman, C. Müller, M.R.

Andersson, Dielectric strength of gamma-radiation cross-linked, high vinyl-content

RI
polyethylene, Eur. Polym. J. 64 (2015) 101-107.

SC
[7] A. Smedberg, T. Hjertberg, B. Gustafsson, Crosslinking reactions in an unsaturated low

density polyethylene, Polymer 38 (1997) 4127-4138.

U
[8] T. Andrews, R.N. Hampton, A. Smedberg, D. Wald, V. Waschk, W. Weissenberg, The
AN
role of degassing in XLPE power cable manufacture, IEEE Electr. Insul. Mag. 22 (2006) 5-

16.
M

[9] M. Severengiz, T. Sprenger, G. Seliger, Challenges and Approaches for a Continuous

D

Cable Production, Procedia CIRP 40 (2016) 18-23.

TE

[10] H.Q. Pham, M.J. Marks, Epoxy Resins, Ullman's Encyclopedia of Industrial Chemistry

Wiley-VCH, Weinheim, Germany, 2006.

EP

[11] J.-P. Pascault, J.J.R. Williams, Epoxy Polymers, New Materials and Innovations, Wiley-
C

VCH, Weinheim, Germany, 2010.

AC

[12] E.M. Petrie, Epoxy Adhesive Formulations, McGraw-Hill, New York, 2006.

[13] R. Auvergne, S. Caillol, G. David, B. Boutevin, J.P. Pascault, Biobased Thermosetting

Epoxy: Present and Future, Chem. Rev. 114 (2014) 1082-1115.

[14] C.A. May, G.Y. Tanaka, Epoxy Resins: Chemistry and Technology, Marcel Dekker,

New York, 1988.

15 / 27
 ACCEPTED MANUSCRIPT
[15] E. Girardreydet, C.C. Riccardi, H. Sautereau, J.P. Pascault, Epoxy-Aromatic Diamine

Kinetics .2. Influence on Epoxy-Amine Network Formation, Macromolecules 28 (1995)

7608-7611.

[16] M.J. Marks, R.V. Snelgrove, Effect of Conversion on the Structure-Property

Relationships of Amine-Cured Epoxy Thermosets, ACS Appl. Mater. Interfaces 1 (2009)

PT
921-926.

RI
[17] J.A. Carioscia, J.W. Stansbury, C.N. Bowman, Evaluation and control of thiol-ene/thiol-

epoxy hybrid networks, Polymer 48 (2007) 1526-1532.

SC
[18] C.E. Hoyle, A.B. Lowe, C.N. Bowman, Thiol-click chemistry: a multifaceted toolbox for

U
small molecule and polymer synthesis, Chem. Soc. Rev. 39 (2010) 1355-1387.
AN
[19] I. Gadwal, J.Y. Rao, J. Baettig, A. Khan, Functionalized Molecular Bottlebrushes,

Macromolecules 47 (2014) 35-40.

M

[20] R. Alex, P.P. De, S.K. De, Self-Vulcanizable Rubber Blend System Based on Epoxidized
D

Natural-Rubber and Carboxylated Nitrile Rubber, J. Polym. Sci. C: Polym. Lett. 27 (1989)
TE

361-367.

[21] A.M. Tomuta, X. Ramis, F. Ferrando, A. Serra, The use of dihydrazides as latent curing
EP

agents in diglycidyl ether of bisphenol A coatings, Prog. Org. Coatings 74 (2012) 59-66.
C

[22] A. Kumar, R.R. Ujjwal, A. Mittal, A. Bansal, U. Ojha, Polyacryloyl Hydrazide: An

AC

Efficient, Simple, and Cost Effective Precursor to a Range of Functional Materials through

Hydrazide Based Click Reactions, ACS Appl. Mater. Interfaces 6 (2014) 1855-1865.

[23] N. Torres, J.J. Robin, B. Boutevin, Study of compatibilization of HDPE-PET blends by

adding grafted or statistical copolymers, J. Appl. Polym. Sci. 81 (2001) 2377-2386.

16 / 27
 ACCEPTED MANUSCRIPT
[24] V. Chiono, S. Filippi, H. Yordanov, L. Minkova, P. Magagnini, Reactive compatibilizer

precursors for LDPE/PA6 blends. III: ethylene-glycidylmethacrylate copolymer, Polymer 44

(2003) 2423-2432.

[25] Q. Wei, D. Chionna, M. Pracella, Reactive compatibilization of PA6/LDPE blends with

glycidyl methacrylate functionalized polyolefins, Macromol. Chem. Phys. 206 (2005) 777-

PT
786.

RI
[26] Y.W. Xu, J. Loi, P. Delgado, V. Topolkaraev, R.J. McEneany, C.W. Macosko, M.A.

Hillmyer, Reactive Compatibilization of Polylactide/Polypropylene Blends, Ind. Eng. Chem.

SC
Res. 54 (2015) 6108-6114.

U
[27] P.J. Flory, Networks, Encyclopedia of Polymer Science and Engineering, Wiley, 1987,
AN
pp. 95-112.

[28] Svensk Elstandard. IEC 60811-507:2012. Electric and optical fibre cables. Test methods
M

for non-metallic materials – Part 507: Mechanical tests. Hot set test for cross-linked materials.
D

[29] U.W. Gedde, Polymer Physics, Kluwer Academic Publishers, 1995.

TE
C EP
AC

17 / 27
 ACCEPTED MANUSCRIPT
Figures

PT
RI
U SC
AN
M
D
TE

Fig 1 (a) P(E-stat-GMA) copolymer; (b) reaction schemes for epoxy ring opening using
EP

amines, dihydrazides and carboxylic acids; (c) amine-based crosslinkers; (d) carboxylic acid-
C

based crosslinkers; (e) dihydrazide-based crosslinkers; (f) dicumyl peroxide.

AC

18 / 27
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE

Fig 2 Hot set elongation and permanent elongation of P(E-stat-GMA) specimens

EP

crosslinked with 1 wt% crosslinker (DAO, MDAO, TMPTA, DDA and AAD) at 200 °C for

two hours (top), and photographs of corresponding specimens after deformation (bottom).
C
AC

19 / 27
 ACCEPTED MANUSCRIPT

PT
Fig 3 P(E-stat-GMA) dumbbell-shaped samples cured at 200 °C with (a) DAO, (b) MDAO,

RI
and (c) TMPTA for different times (from left to right: neat polymer, 20, 45, and 90 min). In

the background, passages from the 16th-century political treatise The Prince, by the Italian

SC
diplomat and political theorist, Niccolò Machiavelli.

U
AN
M
D
TE
C EP
AC

20 / 27
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
C EP
AC

Fig 4 (a) Optical micrographs of P(E-stat-GMA) thin films containing 2 % TMPTA before

(left) and after crosslinking at 200 °C for two hours (right); (b) DSC second heating

thermograms, and (c) TGA thermograms of neat P(E-stat-GMA) and LDPE, LDPE

crosslinked with DCP and P(E-stat-GMA) cured with amine-based crosslinkers (DAO,

21 / 27
 ACCEPTED MANUSCRIPT
MDAO, TMPTA) at 180 °C for two hours. The shaded area highlights the weight loss of

volatile by-products that arise from DCP decomposition.

PT
RI
U SC
AN
M
D
TE
C EP
AC

Fig 5 (a) Gel content of P(E-stat-GMA) crosslinked with 0.1 to 2 wt% amine-based curing

agents (DAO, MDAO, TMPTA) at 160 °C (green), 180 °C (red) and 200 °C (blue) for 2 h; (b)

hot set elongation (red), permanent elongation (grey), gel content (blue) and
22 / 27
 ACCEPTED MANUSCRIPT
network density (green) of P(E-stat-GMA) crosslinked with amine-based curing agents (1

wt% DAO, 1 wt% MDAO, 2 wt% TMPTA) at 200 °C for 10 to 120 min.

PT
RI
U SC
AN
M
D
TE
C EP
AC

Fig 6 Total amount of epoxy and N-H groups for P(E-stat-GMA) crosslinked with 1 wt%

DAO (top), 1 wt% MDAO (centre), and 2 wt% TMPTA (bottom) at 200 °C for two hours.
23 / 27
 ACCEPTED MANUSCRIPT

PT
RI
U SC
Fig 7 Schematic representation of (a) an intra-chain crosslink, (b) an inter-chain crosslink, (c)
AN
a permanent network point that arises due to an entanglement that is trapped by two intra-

chain crosslinks, and (d) an entanglement that is trapped by a combination of intra- and inter-
M

chain crosslinks. Note that this schematic only strictly applies for MDAO where each
D

functional group can only react once with an epoxy ring.

TE
C EP
AC

24 / 27
 ACCEPTED MANUSCRIPT

PT
RI
U SC
Fig 8 Dynamic mechanical analysis (DMA) of neat P(E-stat-GMA) and samples cured for
AN
two hours at 200 °C with 1 wt% DAO, 1 wt% MDAO and 2 wt% TMPTA, and for 45 min at

200 °C with 3 wt% TDA.

M
D
TE
C EP
AC

25 / 27
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
C EP
AC

Fig 9 Summary of temperature and crosslinking time needed to obtain a hot set elongation of

less than 100 %, i.e. a network density of more than 2.8 crosslinks per 1000 carbons, for P(E-

stat-GMA) crosslinked with DAO (top), MDAO (centre), and TMPTA (bottom).

26 / 27
 ACCEPTED MANUSCRIPT

Table 1. Molecular weight M, melting temperature , boiling point 3 4+ and flash point

%+(* ; *at ambient temperature.

Curing agent State of matter* M (g mol-1) 56 (°C) 5789: (°C) 5;:<=> (°C)

TDA solid 213 38 162 >120

PT
DAO solid 144 50 225 165

MDAO solid 172 34 249 >120

RI
TMPTA liquid 440 - - >120

SC
DDA solid 230 130 245 220

AAD solid 174 182 - 150

U
OAD solid 158 89 301 135
AN
M
D
TE
C EP
AC

27 / 27
 ACCEPTED MANUSCRIPT
- the use of click chemistry opens up a by-product free alternative to traditional

crosslinking of polyethylene with peroxides

- a branched statistical ethylene-glycidyl methacrylate copolymer can be crosslinked

with bifunctional amine crosslinking agents

- a well-adjusted processing window around 120 °C permits extrusion of the copolymer

PT
and amine curing agent, followed by crosslinking at 160 to 200 °C

RI
U SC
AN
M
D
TE
C EP
AC