Crosslinking of low density polyethylene with octavinyl polyhedral oligomeric silsesquioxane as the crosslinker

Article abstract of DOI:10.1039/c4ra04886e

We report a new strategy for the crosslinking of low density polyethylene (LDPE) by using a small amount of functional nanostructured hybrid agent, octavinyl polyhedral oligomeric silsesquioxane (OVPOSS), as the crosslinker, which dramatically decreases the amount of dicumyl peroxide (DCP) and thus avoids the chain scission, scorch, and production of small pores in the conventional peroxide crosslinking strategy. By melt blending under an extruder, OVPOSS aggregates with sizes of tens of nanometers are homogeneously distributed in the matrix of LDPE, as confirmed by SEM. FTIR, DSC, and rheometry are used to study the crosslinking process and product properties. We found that vinyl groups of OVPOSS are firstly activated by the initiator of DCP and then react with LDPE to form an integrated network. The crosslinking process is fast and highly efficient because each OVPOSS molecule has eight reactive vinyl groups. The reactions complete within 10 min at 175 °C, and speed up with the increase in the content of DCP or OVPOSS. In the presence of 0.2 phr DCP, 0.5 wt% OVPOSS can effectively crosslink the composite sample with comparable properties to that with 2 phr DCP yet without OVPOSS. The obtained crosslinked LDPE should be suitable for high voltage cable materials. We believe that the approach using functional agents is powerful to crosslink or functionalize other polymers for special properties and applications.

Full text of DOI:10.1039/c4ra04886e

RSC Advances
View Article Online
View Journal | View Issue
PAPER
Crosslinking of low density polyethylene with
octavinyl polyhedral oligomeric silsesquioxane as
the crosslinker†
Cite this: RSC Adv., 2014, 4, 44030
*
*
Jiachun Wu, Zi Liang Wu, Hongmei Yang and Qiang Zheng
We report a new strategy for the crosslinking of low density polyethylene (LDPE) by using a small amount of
functional nanostructured hybrid agent, octavinyl polyhedral oligomeric silsesquioxane (OVPOSS), as the
crosslinker, which dramatically decreases the amount of dicumyl peroxide (DCP) and thus avoids the
chain scission, scorch, and production of small pores in the conventional peroxide crosslinking strategy.
By melt blending under an extruder, OVPOSS aggregates with sizes of tens of nanometers are
homogeneously distributed in the matrix of LDPE, as confirmed by SEM. FTIR, DSC, and rheometry are
used to study the crosslinking process and product properties. We found that vinyl groups of OVPOSS
are firstly activated by the initiator of DCP and then react with LDPE to form an integrated network. The
crosslinking process is fast and highly efficient because each OVPOSS molecule has eight reactive vinyl
groups. The reactions complete within 10 min at 175 ꢀC, and speed up with the increase in the content
of DCP or OVPOSS. In the presence of 0.2 phr DCP, 0.5 wt% OVPOSS can effectively crosslink the
composite sample with comparable properties to that with 2 phr DCP yet without OVPOSS. The
obtained crosslinked LDPE should be suitable for high voltage cable materials. We believe that the
approach using functional agents is powerful to crosslink or functionalize other polymers for special
properties and applications.
Received 23rd May 2014
Accepted 3rd September 2014
DOI: 10.1039/c4ra04886e
www.rsc.org/advances
blended at a relatively low temperature, and then crosslinked in
a special set-up with high temperature and pressure.2 In
1. Introduction
contrast, warm silane crosslinking approach, in which silane
molecules gra to the main chains of LDPE and crosslink LDPE
by forming Si–O–Si bonds via hydrolysis and condensation
reactions in warm water,6,7 has a low cost of the set-up and
materials, is easy to process, and obtains products with good
quality.8,9 However, its demerit is the side reactions such as
chain extension occurred during the silane graing,10 which
limits the applications of thus obtained products.
There are many efforts to study the methods of crosslinking
and properties of crosslinked LDPE, due to the signicances in
practical applications. To overcome the shortcomings
mentioned above, researchers have made modications to the
strategies. For example, in peroxide crosslinking LDPE, the
crosslinking mechanism,11 effects on crosslinking process,3,12
and rheological behaviors of crosslinked LDPE13 have been
investigated. Furthermore, accessory crosslinker is added, to
appropriately decrease the amount of peroxide, improve the
crosslinking effect, and reduce the occurrence of side reac-
tions.14,15 On the other hand, in warm silane crosslinking
approach, internal water-release agent is used to avoid the
inefficiency of water diffusion from the external environment.
The effects of silane molecular structures and graing degree
on the crosslinking process have been systematically studied.16
Beside silane molecules, many other agents, including
Low density polyethylene (LDPE), as one of the most common
thermoplastic polymers, has been widely used as packing and
cable materials due to its excellent physical and electrical
properties. However, LDPE has a major drawback in applica-
tions, the relatively low limit of the service temperature. To solve
this problem, LDPE is usually crosslinked before use. Cross-
linked LDPE possesses excellent dielectric properties, good
resistance to high-temperature deformation, and environ-
mental stress cracking resistance,1,2 which have made cross-
linked LDPE useful for insulation medium to high voltage
cables, hot water pipes, and heat shrinkable lms.
There are mainly three approaches commercially used for
crosslinking LDPE: irradiation crosslinking (using high-energy
radiation or electronics), peroxide crosslinking, and warm water
crosslinking.3–5 The peroxide crosslinking approach generally
includes two steps: LDPE and peroxide are rstly physically
Key Laboratory of Macromolecular Synthesis and Functionalization of Ministry of
Education, Department of Polymer Science and Engineering, Zhejiang University,
Hangzhou 310027, China. E-mail: wuziliang@zju.edu.cn; yanghongmei@zju.edu.cn;
Fax: +86 57187953075; Tel: +86 57187953075
† Electronic supplementary information (ESI) available: Synthesis and
characterizations of OVPOSS, SEM, DSC, and rheological results of the control
systems are presented in ESI. See DOI: 10.1039/c4ra04886e
44030 | RSC Adv., 2014, 4, 44030–44038
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
RSC Advances
difunctional or trifunctional electron-rich olens such as tet- relieved. Scanning electronic microscopy (SEM) is applied to
ramethyltetravinylcyclotetrasiloxane, triallyl isocyanurate, and observe the size and distribution of OVPOSS aggregates in LDPE
triallyl cyanurate, can be used as accessory crosslinkers.14 In the before and aer the crosslinking. Fourier transform infrared
existence of radicals, these accessory agents have high reactivity (FTIR) spectroscopy, differential scanning calorimetry (DSC),
and are easy to gra to polymer chains,15 which improve the and rheometer are used to monitor the crosslinking process
crosslinking efficiency and affect the structure of crosslinked and effectiveness.
A rational crosslinking mechanism is
polymer. Several groups have reported that the multifunctional formulated based on the experimental results. This work pre-
molecules form “coagent bridges” during the crosslinking senting a novel approach to prepare crosslinked LDPE should
reaction.17–19 Although there are many efforts on crosslinking merit preparing other polyolens with improved properties and
LDPE by introducing accessory agents, using functional broader applications.
organic–inorganic hybrid agent as the crosslinker is rarely
reported.
2. Experimental
Polyhedral oligomeric silsesquioxane (POSS) with molecular
2.1 Materials
formula of (RSiO1.5)m, in which m is equal to 6, 8, 10, or 12
(Scheme 1a), is one hybrid molecule containing nanostructured
Si–O skeleton and surface-anchored organic groups, R, which
can be inert such as methyl and phenyl groups or functional
such as vinyl and epoxy groups.20,21 Therefore, POSS is easily to
be composite with polymers by blending, graing, and copoly-
merization,22,23 to improve the material properties such as the
heat resistance, oxidation resistance, mechanical properties,
and dielectrical properties.
When R on the angle Si atoms is inert, POSS cannot be
incorporated into polymeric matrix by chemical bonding.
However, myriad composite materials are developed by physical
blending and possess special properties.24–27 Huang et al. have
reported the improvement in dielectrical properties of
composite materials of octavinyl polyhedral oligomeric silses-
LDPE used in this work (with a density of 0.92 g cmꢁ3, Mn ¼ 1.4
ꢂ 104, Mw ¼ 7.3 ꢂ 104, and a melting ow index of 4.2 g/10 min
ꢀ
at 2.16 kg and 190 C) was purchased from Sinopec Yanshan
Petrochemical Company (Beijing, China) and soaked in ethanol
at 50 ꢀC for 5 h before use. Dicumyl peroxide (DCP, Virtulla
Tech. Co. Ltd., China) and vinyltrimethoxysilane (Nanjing
Xiangqian Chemical Co. Ltd., China) were used as received. The
ultrapure water was made in the lab. All other solvents were
purchased from Sinopharm Chemical Reagent Co. Ltd. (China)
and used without further purication.
2.2 Sample preparation
2.2.1 Synthesis of octavinyl polyhedral oligomeric silses-
quioxane (OVPOSS, Scheme 1b) and LDPE by physical quioxane (OVPOSS). OVPOSS was synthesized according to a
blending.28 When R is active, POSS with multiple functional modied classic route (Fig. S1 in ESI†).38,39 To a 500 mL three-
groups can joint several polymer chains by graing reaction to necked bottle was added anhydrous methanol (200 mL), Milli-Q
form a chemically crosslinked network, which has been used in water (18.2 MU cm, 9 mL), and concentrated hydrochloric acid
preparing elastomers29,30 and hydrogels.31–33 Zhang and his (36 wt%, 2 mL, as the catalyst). The formed mixture (mixture 1)
colleagues have studied the preparation and properties of POSS was stirred rapidly at 40 ꢀC. To a 100 mL beaker was added
lled polypropylene by physical blending and graed poly- vinyltrimethoxysilane (20 mL) and anhydrous methanol (50
propylene by reactive blending, as well as their differences.34–37 mL). The formed homogeneous mixture (mixture 2) was trans-
However, to our best knowledge there are no studies on the ferred into a 100 mL constant pressure funnel. Then mixture 2
crosslinking process and mechanism of polyolen by using a was added dropwise to mixture 1 within 4 h, followed by addi-
small content of OVPOSS as the crosslinker.
tion of 30 mL anhydrous methanol. The resulting mixture was
ꢀ
In this paper, we report a new strategy to prepare crosslinked stirred for another 5 days at 40 C. The precipitate was ltered
LDPE by using OVPOSS, a functional organic–inorganic hybrid and puried further by dissolving in tetrahydrofuran and then
agent, as the crosslinker. In this process, only a small amount of depositing in ethanol. Thus obtained OVPOSS was a white
peroxide is needed to initiate the crosslinking reaction, so that crystal powder with high quality, as characterized by nuclear
the demerit in peroxide crosslinking approach can be properly magnetic resonance (NMR) (Fig. S2†), FTIR (Fig. S3†), and X-ray
diffraction (XRD) (Fig. S4†).
2.2.2 Preparation of the composite samples. LDPE (100
phr) and DCP (5 phr) were put into a ask of 250 mL and mixed
through a rotary evaporator at 80 ꢀC. DCP was coated homo-
geneously on the surface of LDPE granules to produce a DCP
carrier. The DCP carrier was then blended using a co-rotating
twin-screw extruder (PRISM TSE 16 TC, Thermo Scientic,
Waltham, UK) to obtain masterbatches of DCP. The master-
batches of OVPOSS (10 g) in LDPE (90 g) were prepared on the
twin-screw extruder by melt blending. The extrusion conditions
are shown in Table S1.†
The crosslinkable composite samples were prepared by melt
Scheme 1 Molecular structures of POSS (a) and OVPOSS (b).
extrusion of LDPE and masterbatches of DCP and OVPOSS in
This journal is © The Royal Society of Chemistry 2014
RSC Adv., 2014, 4, 44030–44038 | 44031

View Article Online
RSC Advances
Table Composition and sample code of LDPE/OVPOSS/DCP
Paper
1
DHf
Xc ¼
(1)
composites
DHmð1 ꢁ FÞ
LDPE
content (wt%)
OVPOSS
content (wt%)
DCP
content (phr)
Sample
code
and the thickness of the lamellar crystal of LDPE, Dlc, can be
determined by the empirical Gibbs–Thompson equation40
100
100
100
100
100
100
100
99.8
99.8
99.8
99.8
99.5
99.5
99.5
98
0
0
0
0
0
0
0
0.2
0.2
0.2
0.2
0.5
0.5
0.5
2.0
2.0
2.0
0
LDPE
L-0.1D
L-0.2D
L-0.5D
L-0.8D
L-1.0D
L-2.0D
L-0.2P
L-0.2P-0.1D
L-0.2P-0.2D
L-0.2P-0.5D
L-0.5P
L-0.5P-0.2D
L-0.5P-0.5D
L-2.0P
ꢀ
ꢁ
0.1
0.2
0.5
0.8
1.0
2.0
0
0.1
0.2
0.5
0
0.2
0.5
0
0.2
0.5
2s
Tm ¼ T 0m 1ꢁ
(2)
DHm$Dlc
in which, DHf is the enthalpy changes obtained from the heat of
fusion in melting; DHm is the enthalpy changes in melting of a
100% crystalline sample; F is the weight fraction of OVPOSS
and DCP in the samples; Tm is the melting point of the samples;
T0m is the equilibrium melting point of polymer; s is the chain
folding energy of polyethylene. For LDPE used in this work,
DHm, Tm0 , and s are 293 J gꢁ1, 417.15 K, and 93.4 erg cmꢁ2
respectively.41
,
2.3.4 Rheological measurements. An advanced rheometric
expansion system (ARES-G2, TA, USA), operated with 25 mm
parallel plate setup, was used to measure the angular frequency
dependence of storage modulus (G0), loss modulus (G00), loss
tangent (tan d), and complex viscosity (h*) of the compounds in
the range of 100–0.015 rad sꢁ1. Strain sweep in the range of
0.01–100% was performed to determine the linearity region of
crosslinked LDPE. The frequency sweep was tested at a strain of
1% to measure the linear viscoelasticity of crosslinked LDPE.
For rheological tests, circular disk-like samples of 25 mm in
diameter and 1.2 mm in thickness were prepared on a
compression-molding machine under 15 MPa for 10 min at
175 ꢀC, then transferred into another compression-molding
machine under 15 MPa for 5 min at ambient temperature. All
the rheological tests mentioned above were operated at 150 ꢀC,
above Tm of the samples.
98
98
L-2.0P-0.2D
L-2.0P-0.5D
suitable ratios to get the prescribed compositions in Table 1 in
the twin-screw extruder. Samples are coded as L-xP-yD, in which
x and y are the content of OVPOSS in wt% and DCP content in
phr, respectively. Samples without DCP and OVPOSS are coded
as L-xP and L-yD, respectively.
2.3 Characterization
2.3.1 SEM-EDS. Scanning electron microscopy (SEM; Hita-
chi-S4800) with energy dispersive spectroscopy (EDS) was used
to characterize the dispersion state of OVPOSS in the composite
materials before and aer crosslinking. Crosslinkable and
crosslinked samples, prepared on a compression-molding
Time sweep was used to test the crosslinking behavior, in
which the samples were prepared previously at 115 ꢀC in the
compression-molding machine. The samples were stored at
room temperature for at least 48 h before the testing at 175 ꢀC.
ꢀ
machine under 15 MPa for 10 min at 115 and 175 C, respec-
tively, were broken in liquid nitrogen and then dried in a
vacuum oven. The fracture surfaces of the samples were coated
with gold by vapor deposition using a Denton Desk-1 vacuum
sputter coater for SEM observation and X-ray element mapping
of silicon.
3. Results and discussion
3.1 OVPOSS crosslinking reaction and its mechanism
2.3.2 FTIR. Infrared spectra of the thin lm samples were To obtain crosslinked LDPE, peroxide-curing reaction based on
obtained at ambient temperature with a Fourier transform radical coupling mechanism is the most common method.
infrared spectroscopy (Vector 22 FT-IR spectrum, Bruker, Ger- However, using DCP to crosslink polyolen is always accom-
many). Single beam spectra of the samples were obtained aer panied with side reactions such as chain scission and scorch at
an average of 31 scans between 4000 and 400 cmꢁ1. All spectra high temperature and production of small pores.42,43 These
were obtained in the transmittance mode.
issues can be properly addressed by using OVPOSS as the
2.3.3 DSC. The crosslinking process, crystallization and crosslinker, in which only a small amount of DCP is needed as
melting behaviors of the samples were analyzed using a Q100 the initiator.
(TA instruments) differential scanning calorimeter (DSC) in a
The morphology of the fracture surface of crosslinkable and
owing nitrogen atmosphere. All the samples, weighted crosslinked samples (L-0.2P-0.2D) is shown in Fig. 1 and S5.†
around 5 to 10 mg, were sealed in standard aluminum Before crosslinking, OVPOSS crystals with cuboid shape and
capsules and heated from 40 to 220 ꢀC. Then the sampleꢀs were size ranging from hundreds of nanometers to several microm-
ꢀ
cooled from 220 to 40 C and reheated from 40 to 180 C. All eters are randomly dispersed in the matrix (Fig. 1a and S5a†).
the heating and cooling rates were kept at 10 C minꢁ1. The However, aer crosslinking, OVPOSS aggregates break to small
ꢀ
crystallinity of the samples, Xc, was determined by the particles with the size of tens of nanometers (Fig. 1c, S5b and
following equation
c†).35 Elemental mapping images (Fig. 1b and d) indicate that
44032 | RSC Adv., 2014, 4, 44030–44038
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
RSC Advances
Fig. 1 SEM images (a and c) and corresponding elemental mapping
images (b and d) of the fracture surface of L-0.2P-0.2D before (a and b)
and after (c and d) crosslinking.
the particles are just OVPOSS aggregates. We found that the
breaking of micron-sized OVPOSS crystals to nano-sized parti-
cles is not due to high temperature but closely related to the
crosslinking reaction, because without DCP to trigger the
reaction, the morphology and size of OVPOSS aggregates in L-
0.2P have no changes aer the same heating process (Fig. S5d–
f†). We also found micron-sized OVPOSS particles in the
mixture of OVPOSS and DCP disaggregated into nano-sized
particles aer the thermal initiated reaction (Fig. S6†); in the
Fig. 2 FTIR spectra of the OVPOSS, LDPE, crosslinkable and cross-
linked L-0.5P-0.2D.
melt blending. Aer the crosslinking, there are several notable
changes of the peaks due to the reaction-induced variations of
chemical environment of some groups: (i) the peaks at 1604 and
969 cmꢁ1 almost disappear, indicating that most of vinyl groups
have been reacted; (ii) the peak at 1118 cmꢁ1 has evident change
from a single peak to two split peaks, mainly due to the
destruction of perfect symmetry of Si–O skeleton; (iii) the
disappearance of peak at 584 cmꢁ1 is also due to the reaction-
induced variation of chemical environment for Si–O–Si bond;
(iv) the peak at 775 cmꢁ1, which is partially hidden by the broad
peak of LDPE at 730 cmꢁ1, has a detectable change, due to the
transformation from Si–C]C to Si–C–C. Based on these results,
we conclude that most vinyl groups of OVPOSS have reacted and
been incorporated into the polymeric matrix of crosslinked
LDPE. It is expected that OVPOSS effectively links LDPE chains
to form a network.
ꢀ
reaction process, there is a small exothermic peak at ꢃ145 C
besides the main peak at ꢃ170 ꢀC (Fig. S7†). DCP starts to
decompose at ꢃ145 ꢀC, which triggers the reaction of vinyl
groups at the surface of OVPOSS aggregates. The reaction would
produce localized heat and change the crystalline morphology.
If these differences between the outer and inner crystals of
OVPOSS are big enough, the aggregates will break into small
pieces. In consequence, a large amount of OVPOSS molecules
are exposed and have the chance to connect with radicals. The
following reaction generates more heat and forms the second
ꢀ
large exothermic peak at 170 C. In the presence of DCP, the
well dispersed OVPOSS molecules and nano-sized particles with
multiple vinyl groups will act as an effective crosslinker to
crosslink LDPE.
The crosslinking process is also investigated in situ by using
DSC. The crosslinkable samples (L-xP-0.2D and L-0.2P-yD) are
sealed in the aluminum capsules; crosslinking reaction takes
place during the rst heating process up to 220 ꢀC. As shown in
Fig. 3a, the exothermic peak of composite sample with 0.2 wt%
OVPOSS (L-0.2P-0.2D) moves to low temperature, when
compared to that of L-0.2D sample without OVPOSS. As the
content of OVPOSS increases, the exotherm rapidly increases,
indicating the effective crosslinking reaction of OVPOSS.
We should note that the peroxide crosslinking process of
LDPE without OVPOSS shows an exothermic peak at an iden-
tical temperature, ca. 190 ꢀC. However, the exothermic peaks in
our systems with OVPOSS appear at relatively low temperature.
Fig. 2 shows FTIR spectra of the pure LDPE, OVPOSS, and the
composite sample (L-0.5P-0.2D) before and aer the cross-
linking process. In contrast with the pure LDPE, several char-
acteristic peaks of OVPOSS exist in the spectrum of
crosslinkable L-0.5P-0.2D (before crosslinking). The peaks at
1604, 1118, and 775 cmꢁ1 correspond to the absorption of nC]C
,
nSi–O–Si, and nSi–C]C, respectively. The peaks at 969 and 584
cmꢁ1 correspond to the absorption of the stretching vibration of
trans-vinylene and to the ngerprint spectrum of Si–O–Si skel-
eton vibration, respectively. By comparision of the two spectra,
we can see that characteristic peaks of OVPOSS appear in the
right places of the spectrum of composite material prepared by
This journal is © The Royal Society of Chemistry 2014
RSC Adv., 2014, 4, 44030–44038 | 44033

View Article Online
RSC Advances
Paper
Furthermore, the peak slightly moves to high temperature when
the content of OVPOSS increases from 0.2 to 2 wt% (Fig. 3a).
These differences suggest a different crosslinking mechanism
in our work.
Essentially, the peroxide crosslinking of LDPE is a multiple-
step reaction. The DCP decomposes into free radicals at an
appropriate temperature, which capture hydrogen atoms from
the polyethylene chains. The polymer chains then become
macroradicals, which couple to form a crosslinked polymeric
network.11,44 However, in our work, the radicals from decom-
position of DCP preferentially react with the active vinyl groups
of OVPOSS, which should have a lower energy than taking a
hydrogen from the main chain of LDPE (Fig. 4). This is sup-
ported by the DSC results (Fig. 3a and S7†), i.e. the exothermic
peaks of crosslinking LDPE with OVPOSS move to low temper-
ature. The exothermic peak of L-0.2P-0.2D is ꢃ170 ꢀC, 20 ꢀC
lower than that of L-0.2D. The radical of OVPOSS can take a
hydrogen from LDPE chain, leading to the formation of poly-
ethylene network. We should note that part of DCP radicals also
can capture a hydrogen from LDPE main chain to form mac-
roradicals, which couple each other or with OVPOSS radicals to
form a network (Fig. 4). By introducing OVPOSS as the main
Fig. 4 Reaction scheme of OVPOSS crosslinked LDPE.
crosslinker, we can dramatically decrease the amount of DCP
and therefore avoid the possibility of chain scission and the
amount of side product from DCP decomposition.
The crosslinking of LDPE by OVPOSS results in rational
modication of the product properties. DSC is also applied to
investigate the melting temperature, Tm, and crystallinity, Xc, of
crosslinked LDPE (Fig. 3b). Two serials of samples, L-xP-0.2D
and L-0.2P-yD, are used to study the effect of OVPOSS and DCP
on Tm and Xc. The typical melting and crystalline parameters
including Tm, crystallization temperature Tc, the thickness of
the lamellar crystal Dlc, and Xc obtained from the melting
endotherms are given in Tables 2 and 3. The crosslinking leads
to slight decrease in Tm, Dlc, and Xc, less than 2 ꢀC, 0.4 nm, and
3%, respectively, due to the restricted chain mobility and
folding to form crystal domain. Only when the sample contains
a relatively large content of OVPOSS (2 wt% in L-2.0P-0.2D), the
melting peak shows notable change, corresponding to 5 ꢀC Tm,
1.13 nm Dlc, and 5% Xc lower than those of L-0.2D.
3.2 Dynamic rheology of samples during and aer
crosslinking
The crosslinking process of LDPE with OVPOSS, leading to the
formation of an integrated network and variation of mechanical
properties, is investigated by dynamic rheology.
3.2.1 Strain sweep of crosslinked LDPE. Strain sweep with
strain amplitude, g, ranging from 0.01% to 100% is performed
on the samples of pure LDPE and crosslinked samples at 150 ꢀC,
above their Tm. Fig. 5 shows the normalized storage modulus
(G0/Gp0 ) as a function of g. Here, G0 and Gp0 are the storage
Table
2 Melting and crystalline parameters of crosslinked LDPE
samples with 0.2 phr DCP, L-xP-0.2D
Samples
Tc (ꢀC)
Tm (ꢀC)
Dlc (nm)
DHf (J gꢁ1
)
Xc (%)
LDPE
L-0.2D
L-0.2P-0.2D
L-0.5P-0.2D
L-2.0P-0.2D
98.8
98.3
97.6
96.8
93.6
112.0
111.8
111.0
110.5
107.0
8.90
8.83
8.61
8.50
7.70
114.5
109.3
107.1
105.4
97.0
39.1
37.4
36.7
36.2
33.9
Fig. 3 DSC thermograms of crosslinkable LDPE during the heating
process. (a) L-xP-0.2D, during the crosslinking; (b) L-xP-0.2D, after the
crosslinking. Heating rate: 10 ꢀC minꢁ1
.
44034 | RSC Adv., 2014, 4, 44030–44038
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
RSC Advances
Table
3 Melting and crystalline parameters of crosslinked LDPE
samples with 0.2 wt% OVPOSS, L-0.2P-yD
Samples
Tc (ꢀC)
Tm (ꢀC)
Dlc (nm)
DHf (J gꢁ1
)
Xc (%)
LDPE
L-0.2P
L-0.2P-0.1D
L-0.2P-0.2D
L-0.2P-0.5D
98.8
98.4
97.9
97.6
96.9
112.0
111.8
111.2
110.9
110.1
8.90
8.83
8.68
8.61
8.40
114.5
114.5
109.8
107.1
104.7
39.1
39.1
37.6
36.7
36.0
Fig. 5 Normalized storage modulus, G0/Gp0 , as a function of strain
amplitude, g, for pure LDPE and crosslinked LDPE with different
content of OVPOSS and DCP at 150 ꢀC. Frequency, u, is 3.14 rad sꢁ1
.
modulus and the value in the linearity region. At frequency, u,
of 3.14 rad sꢁ1, G0/Gp0 was found to be stable at g < 33% for pure
LDPE, whereas the critical g for onset of nonlinearity decreases
to 16%, 16%, and 6% for L-2.0D, L-0.2P-0.5D, and L-2.0P-0.5D,
respectively. All the following dynamic tests are performed at
g ¼ 1%.
Fig. 6 G0 as a function of crosslinking time, t, at a temperature
of 175 ꢀC for samples of LDPE, L-xP-0.2D (a), and L-0.2P-yD (b).
Corresponding G0 of L-yD is also shown for a comparision.
Table 4 Equilibrium modulus, Ge0 , and gelation time, tg, of L-yD and
L-xP-yD
3.2.2 Rheologic behaviors during the crosslinking process.
Fig. 6a shows G0 as a function of the crosslinking time, t, at
175 ꢀC for crosslinkable samples of L-xP-0.2D. G0 increases with
t and achieves a constant value aer 10 min, indicating the
completion of crosslinking reaction. With a constant content of
DCP, 0.2 phr, the equilibrium modulus, Ge0 , increases with the
content of OVPOSS. With incorporation of 0.5 wt% OVPOSS, Ge0
of L-0.5P-0.2D is 6 times that of L-0.2D (Table 4).
Fig. 6b shows G0 as a function of t at 175 ꢀC for crosslinkable
samples of L-0.2P-yD. G0 also increases with t and achieves a
constant value aer 10 min. With a constant content of
OVPOSS, 0.2 wt%, G0 increases with the content of DCP. Even a
small amount of OVPOSS can dramatically increase Ge0 . The
incorporation of 0.2 wt% OVPOSS (L-0.2P-0.2D) leads to Ge0 3
times that without OVPOSS (L-0.2D). Ge0 of L-0.2P-0.1D is around
2 times that of L-0.2D; Ge0 of L-0.2P is almost the same as that of
pure LDPE (Table 4). These results indicate that OVPOSS
effectively crosslinks the polyethylene chain, which can only
been triggered in the presence of DCP (Fig. 6b). Ge0 of L-0.5D is
Samples
Ge0 (kPa)
tg (s)
LDPE
L-0.2P
4.4
4.5
—
—
L-0.2D
7.8
—
L-0.5D
17.3
14.7
19.4
45.7
151.5
38.1
138
128
80
44
35
24
L-0.2P-0.1D
L-0.2P-0.2D
L-0.5P-0.2D
L-2.0P-0.2D
L-0.2P-0.5D
between the values of L-0.2P-0.1D and L-0.2P-0.2D (Table 4). For
a better crosslinking of LDPE, an optimal combination between
the contents of OVPOSS and DCP is needed.
During the crosslinking, G0 and the loss modulus, G00,
increase with t in different speeds and cross each other aer a
This journal is © The Royal Society of Chemistry 2014
RSC Adv., 2014, 4, 44030–44038 | 44035

View Article Online
RSC Advances
Paper
certain time. The gelation time at G0 ¼ G00, tg, is extracted as a rheological behavior of a polymer could be characterized by
special parameter to characterize the crosslinking speed, which “self-similar relaxation pattern”. For critical gel, its viscoelastic
is also listed in Table 4. tg for L-0.5D is 138 s, however, with parameters are shown to be related as
incorporation of 0.2 wt% OVPOSS, tg for L-0.2P-0.5D is short-
G00ðuÞ
G0ðuÞ ¼
¼ SGð1 ꢁ nÞcosðnp=2Þun
(3)
ened to 24 s, indicating high crosslinking efficiency of OVPOSS.
For L-xP-0.2D, tg decreases from 80 s to 35 s with the increase in
OVPOSS content from 0.2 to 2.0 wt%. For L-0.2P-yD, tg shortens
from 128 s to 24 s with the increase in DCP content from 0.1 to
0.5 phr. It is interesting to note that tg for L-2.0P-0.2D is longer
than that for L-0.2P-0.5D, while the Ge0 is almost 4 times that of
the latter. These results suggest that the crosslinking speed and
tg are mainly determined by the amount of DCP (the initiator),
whereas the crosslinking degree and Ge0 are mainly determined
by the content of OVPOSS (the crosslinker).
tanðnp=2Þ
where S is gel strength, n is the relaxation exponent whose value
lies between 0 and 1, G(1 ꢁ n) is the gamma function.
At the critical gel point, when 0 < u < 1/l0, tan d becomes
frequency independent, i.e.
ꢀ
ꢁ
G00
G0
np
tan d ¼
¼ tan
(4)
2
l0 is the relaxation time of the transition from rubbery plateau
to the reputation regime in the low-frequency zone. This typical
characteristic of the critical gel provides us with an easy tool to
determine the gel point using rheological methods, i.e., in a
tan d versus u plot, the gel point manifests itself by the
appearance of a plateau with zero slope. Typically, for polymer
melts the slopes of curves in a tan d versus u plot are negative
3.2.3 Rheologic behaviors of crosslinked LDPE. The reac-
tions in the presence of DCP and OVPOSS with different content
produce branched chains or crosslinked network of LDPE with
different extent. According to Winter's theory,45–47 at gel point
with the rst appearance of an integrated network, the
Fig. 7 G0 (a) and tan d (b) as a function of u for crosslinked samples of L-xP-0.2D at 150 ꢀC. The solid lines in (a) are drown according to eqn (3).
Fig. 8 G0 (a) and tan d (b) as a function of u for crosslinked samples of L-yD at 150 ꢀC. The solid lines in (a) are drown according to eqn (3).
44036 | RSC Adv., 2014, 4, 44030–44038
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
RSC Advances
while for the solids (gels) it is positive. Hence, the transition homogeneously in LDPE matrix aer melt blending and cross-
from melt to gel can be easily identied if a change from linking. The crosslinking process between OVPOSS and LDPE
negative to positive slope appears in a tan d versus u plot.
has been monitored by FTIR, DSC, and rheometer. We have
Fig. 7 shows G0 and tan d as a function of u for crosslinked rationally contented a crosslinking mechanism: the radicals
samples of L-xP-0.2D at 150 ꢀC. Pure LDPE exhibits a non- from DCP decomposition preferentially react with the vinyl
terminal ow characterized by G0 ꢃ u1.50 and G00 ꢃ u0.86 at groups of OVPOSS and then gra to LDPE chains to from a three
frequency u < 0.2 rad sꢁ1 due to the wide molecular weight dimensional network. The crosslinking speed increases with
distribution of LDPE (Mw/Mn ¼ 5.14). The crosslinking not only DCP content, whereas crosslinking degree increases with
elevates G0 and depresses tan d over the whole u range, but also OVPOSS content. Therefore, an optimal combination of
leads to the uid-to-solid transition, as shown by the gradually OVPOSS and DCP is needed for better crosslinking. The cross-
diminished u-dependence of G0 with the increase in OVPOSS linking leads to increase in storage modulus, G0, and decrease
content (Fig. 7a). The tan d curves of crosslinked LDPE exhibit in the thickness of the lamellar crystal of polyethylene due to
negative slopes when the content of OVPOSS is less than 0.2 the restricted chain mobility and folding to form crystal
wt%, and positive slopes when the content OVPOSS is more domain. The sample with 0.5 wt% OVPOSS and 0.2 phr DCP (L-
than 0.2 wt% (Fig. 7b). For L-0.2P-0.2D, tan d with a constant 0.5P-0.2D) exhibits the same rheological behaviors as that with
value of 0.6 is independent of u. Therefore, 0.2 wt% of OVPOSS 2 phr DCP prepared by the conventional peroxide crosslinking
is the critical value for the samples of L-xP-0.2D to form an (L-2.0D). This work describes a new strategy to crosslink LDPE
integrated network.
by using functional hybrid agents as the crosslinker, which
In comparision, the curves of G0 and tan d versus u for should be suitable for crosslinking or functionalization of other
samples of L-xP show that OVPOSS physically modied LDPE polymers towards specic applications.
can only slightly elevate G0 (Fig. S8a†). The slopes of the tan d
versus u curves keep negative for the physical blending samples
(Fig. S8b†), indicating no network existed in these samples.
Acknowledgements
Fig. 8 shows G0 and tan d as a function of u for crosslinked This research was nancially supported by Natural Science
samples of L-yD. G0 of the samples increases with DCP content Foundation of Zhejiang Province, China (no. Y14E030021),
at low u, yet its u-dependence gradually decreases as u Industrial Research Projects of the Science and Technology Plan
increases (Fig. 8a). Without OVPOSS, 0.8 phr of DCP is the of Ningbo, China (no. 2010B10022), and Fundamental Research
critical value to form an integrated network (Fig. 8b); its tan d Funds for the Central Universities of China (no. 2014QNA4036).
keeps an constant value of 0.5. The sample with 2 phr DCP (L-
2.0D) exhibits the same rheological behaviors to those of L-0.5P-
0.2D, indicating a similar crosslinking effect. These results
Notes and references
prove further the high efficiency of OVPOSS as the crosslinker in
crosslinking LDPE.
1 S. Dadbin, M. Frounchi, M. H. Saeid and F. Gangi, J. Appl.
Polym. Sci., 2002, 86, 1959.
Different from other composite materials by using POSS as
the llers to improve the mechanical or electrical properties, in
our work OVPOSS is used as a crosslinker and incorporated to
the polymer matrix. Therefore, the amount of OVPOSS as
crosslinker (0.2–2 wt%) is much less than other POSS as llers
(usually >15 wt%). The corresponding properties of POSS such
as oxidation resistance and ame resistance are not
pronounced in our systems. However, as we presented above,
OVPOSS as a special crosslinker is of high efficiency in cross-
linking LDPE. This strategy can be generalized to other systems,
by using different functional agents such as carbon nanotube
and graphene to modify polymers, elastomers, hydrogels,
etc.48–50
2 A. Smedberg, T. Hjertberg and B. Gustafsson, Polymer, 1997,
38, 4127.
3 L. H. U. Andersson and T. Hjertberg, Polymer, 2006, 47, 200.
4 A. Smedberg, T. Hjertberg and B. Gustafsson, Polymer, 2003,
44, 3395.
5 Y. T. Shieh and T. H. Tsai, J. Appl. Polym. Sci., 1998, 69, 255.
6 H. C. Kuan, J. F. Kuan, C. C. M. Ma and J. M. Huang, J. Appl.
Polym. Sci., 2005, 96, 2383.
7 P. Pollet, C. L. Liotta, C. A. Eckert, M. Verma, E. Nixon,
S. Sivaswamy, R. Jha, F. Momin, L. Gelbaum,
B. I. Chaudhary and J. M. Cogen, Ind. Eng. Chem. Res.,
2011, 50, 12246.
8 S. Dalai and W. X. Chen, J. Appl. Polym. Sci., 2002, 86, 1296.
9 Z. H. Li, G. H. Hu, J. P. Corriou, S. Hoppe, C. Fonteix,
R. Laine, J. Habimana and D. Deheunynck, Polym. Eng.
Sci., 2013, 53, 1571.
4. Conclusions
In this paper, we present the crosslinking of LDPE with a small 10 C. M. Jiao, Z. Z. Wang, X. M. Liang and Y. Hu, Polym. Test.,
amount of functional nanostructured hybrid molecule, 2005, 24, 71.
OVPOSS, as the crosslinker, which effectively decreases the 11 B. Likozar and M. Krajnc, Polym. Eng. Sci., 2009, 49, 60.
content of DCP and avoids the side reaction and production of 12 A. Smedberg, T. Hjertberg and B. Gustafsson, Polymer, 2004,
small pores in the peroxide crosslinking approach. SEM and
elemental mapping results show that OVPOSS aggregates, with 13 X. H. Zhang, H. M. Yang, Y. H. Song and Q. Zheng, Polymer,
the content of 0.2–2 wt% and the size ranging from tens of 2012, 53, 3035.
nanometers to several micrometers, are dispersed 14 A. L. Logothetis, Prog. Polym. Sci., 1989, 14, 251.
45, 4867.
This journal is © The Royal Society of Chemistry 2014
RSC Adv., 2014, 4, 44030–44038 | 44037

View Article Online
RSC Advances
Paper
15 S. J. Oh and J. L. Koenig, J. Polym. Sci., Part B: Polym. Phys., 31 K. Zeng, L. Wang and S. X. Zheng, J. Phys. Chem. B, 2009, 113,
2000, 38, 1417. 11831.
16 H. Azizi, J. Morshedian, M. Barikani and M. H. Wagner, Adv. 32 J. F. Mu and S. X. Zheng, J. Colloid Interface Sci., 2007, 307,
Polym. Technol., 2011, 30, 286. 377.
17 H. G. Dikland, T. Ruardy, L. Vanderdoes and A. Bantjes, 33 L. Wang, K. Zeng and S. X. Zheng, ACS Appl. Mater. Interfaces,
Rubber Chem. Technol., 1993, 66, 693. 2011, 3, 898.
18 H. G. Dikland, L. Vanderdoes and A. Bantjes, Rubber Chem. 34 Z. Y. Zhou, L. M. Cui, Y. Zhang, Y. X. Zhang and N. A. W. Yin,
Technol., 1993, 66, 196. Eur. Polym. J., 2008, 44, 3057.
19 Z. H. Murgic, J. Jelencic and L. Murgic, Polym. Eng. Sci., 1998, 35 Z. Y. Zhou, L. M. Cui, Y. Zhang, Y. X. Zhang and N. W. Yin, J.
38, 689. Polym. Sci., Part B: Polym. Phys., 2008, 46, 1762.
20 Y. Kawakami, Y. Kakihana, A. Miyazato, S. Tateyama and 36 Z. Y. Zhou, Y. Zhang, Z. Zeng and Y. X. Zhang, J. Appl. Polym.
M. A. Hoque, Silicon Polymers, ed. A. M. Muzafarov, 2011,
vol. 235, p. 185.
21 G. Z. Li, L. C. Wang, H. L. Ni and C. U. Pittman, J. Inorg.
Organomet. Polym., 2001, 11, 123.
Sci., 2008, 110, 3745.
37 Z. Y. Zhou, Y. Zhang, Y. X. Zhang and N. W. Yin, J. Polym. Sci.,
Part B: Polym. Phys., 2008, 46, 526.
38 P. A. Agaskar, J. Am. Chem. Soc., 1989, 111, 6858.
22 D. B. Cordes, P. D. Lickiss and F. Rataboul, Chem. Rev., 2010, 39 P. A. Agaskar, Inorg. Chem., 1991, 30, 2707.
110, 2081.
40 E. James, Physical Properties of Polymers Handbook, Springer,
Ohio, 2nd edn, 2006.
23 L. Zhang, H. C. L. Abbenhuis, Q. H. Yang, Y. M. Wang,
P. Magusin, B. Mezari, R. A. van Santen and C. Li, Angew. 41 J. D. Hoffman and R. L. Miller, Polymer, 1997, 38, 3151.
Chem., Int. Ed., 2007, 46, 5003.
24 M. Joshi, B. S. Butola, G. Simon and N. Kukaleva,
Macromolecules, 2006, 39, 1839.
25 H. W. Milliman, H. Ishida and D. A. Schiraldi,
Macromolecules, 2012, 45, 4650.
26 K. Nusser, G. J. Schneider, W. Pyckhout-Hintzen and
D. Richter, Macromolecules, 2011, 44, 7820.
42 X. H. Zhang, H. M. Yang, Y. H. Song and Q. Zheng, J. Appl.
Polym. Sci., 2012, 126, 939.
43 X. H. Zhang, H. M. Yang, Y. H. Song and Q. Zheng, J. Appl.
Polym. Sci., 2012, 126, 1885.
44 M. Lazar, A. Kleinova, A. Fiedlerova, I. Janigova and
E. Borsig, J. Polym. Sci., Part A: Polym. Chem., 2004, 42, 675.
45 J. C. Scanlan and H. H. Winter, Macromolecules, 1991, 24, 47.
27 X. Wang, Y. Hu, L. Song, W. Y. Xing, H. D. Lu, P. Lv and 46 H. H. Winter, Polym. Eng. Sci., 1987, 27, 1698.
G. X. Jie, J. Polym. Res., 2011, 18, 721. 47 H. H. Winter and F. Chambon, J. Rheol., 1986, 30, 367.
28 X. Y. Huang, L. Y. Xie, P. K. Jiang, G. L. Wang and Y. Yin, Eur. 48 H. P. Cong, P. Wang and S. H. Yu, Small, 2014, 10, 448.
Polym. J., 2009, 45, 2172.
29 C. B. Cong, C. C. Cui, X. Y. Meng and Q. Zhou, J. Appl. Polym.
Sci., 2013, 130, 1281.
30 D. Zhang, Y. Liu, Y. Shi and G. Huang, RSC Adv., 2014, 4,
6275.
49 Y. C. Jung, N. G. Sahoo and J. W. Cho, Macromol. Rapid
Commun., 2006, 27, 126.
50 Y. Y. Shao, S. Zhang, M. H. Engelhard, G. S. Li, G. C. Shao,
Y. Wang, J. Liu, I. A. Aksay and Y. H. Lin, J. Mater. Chem.,
2010, 20, 7491.
44038 | RSC Adv., 2014, 4, 44030–44038
This journal is © The Royal Society of Chemistry 2014