J Mater Sci (2011) 46:4058–4070

DOI 10.1007/s10853-011-5335-9

An investigation of the potential of polypropylene and its blends

for use in recyclable high voltage cable insulation systems
I. L. Hosier • A. S. Vaughan • S. G. Swingler

Received: 2 December 2010 / Accepted: 25 January 2011 / Published online: 5 February 2011
Ó Springer Science+Business Media, LLC 2011

Abstract Most modern extruded high voltage cables together to identify the most suitable candidate materials for
employ cross-linked polyethylene (XLPE) as the insulation future cable applications. The use of blending as a means to
material. XLPE has excellent thermo-mechanical proper- further optimise the various material properties was also
ties, is relatively cheap and has a low dielectric loss, which explored.
make it an ideal material for this application. Unfortunately,
XLPE is not easily recycled at the end of its lifetime leading
to questions concerning its long-term sustainability. A Introduction
previous investigation in this series considered the potential
of a range of ethylene-based systems to provide suitable Extruded cables based on polymeric materials have now
recyclable alternatives to XLPE. Whilst blending could largely superseded older paper/oil designs for medium and
allow systems having similar thermo-mechanical and high voltage applications [1]. Materials such as low density
electrical properties to XLPE to be designed, it was not polyethylene (LDPE), poly-vinyl chloride (PVC), ethylene
possible to obtain better performance than XLPE using propylene rubber (EPR) and ethylene–vinyl acetate (EVA)
these systems. Polypropylene offers, potentially, a route to are easy to process and are excellent electrical insulators
improved insulation systems by virtue of its higher melting making them suitable for electrical cable designs. For
point and excellent dielectric properties. However, tradi- example, PVC is widely used for light duty equipment and
tional isotactic polypropylenes have always had the prob- power cables as it has adequate mechanical and electrical
lem of being too brittle for inclusion into practical cable properties around room temperature [2]. However, for high
designs. Recently a broad range of propylene co-polymers power transmission cables, additional factors, such as the
having improved ductility have become available, which need to retain adequate mechanical integrity and good
may prove more suitable. The current study compares tra- insulation properties at high temperatures, flexibility at low
ditional isotactic and syndiotactic polypropylenes to a range temperatures and the need to reduce transmission losses as
of commercially available propylene co-polymers and much as possible, have to be considered. For these reasons,
focuses on their morphology, thermal, thermo-mechanical cross-linked polyethylene (XLPE) remains the material of
and electrical properties. These parameters were then taken choice for the current generation of extruded high voltage
transmission cables. Compared to LDPE, XLPE has dra-
matically improved thermo-mechanical stability at high
temperatures, wrought by virtue of a cross-linked network,
combined with excellent dielectric properties [3]. Unfor-
Preliminary results from this work were presented at the International tunately, the same cross-linked network results in a system
Symposium on Electrical Insulation, Vancouver, Canada, 2008. that cannot be easily recycled at the end of its service life,
although various schemes for re-use have been proposed
I. L. Hosier (&)
A. S. Vaughan
S. G. Swingler
[4–6].
ECS, University of Southampton, Highfield,
Southampton SO17 1BJ, UK Minimising the environmental impact of human activity
e-mail: [email protected] through recycling and re-use is becoming more important

123
J Mater Sci (2011) 46:4058–4070 4059

both politically and technologically. Consequentially the Experimental

recyclability of cable materials is becoming a serious
concern, especially in the context of increasing landfill Materials
costs. This indicates that there is now a critical need to
move away from XLPE towards recyclable alternatives for Six different polyethylenes supplied by Sigma-Aldrich
the next generation of extruded high voltage cables. With Chem. Co. were studied in this investigation (Table 1).
this in mind, a number of notable studies into the electrical Two of these represent typical isotactic (iPP) and syndio-
and thermal properties of LDPE and its blends [7–9], tactic (sPP) polypropylenes, respectively. Three materials
HDPE and its blends [10] and finally LDPE/HDPE blends are ethylene co-polymers with co-monomer contents
[9, 11, 12] have been undertaken. Whilst many of these ranging from 2 to 40% and are designated PEx where
materials have excellent dielectric properties and are easily x indicates the ethylene content. Finally PB12 is a pro-
recyclable at the end of life using existing waste channels, pylene–butene co-polymer containing 12% butene.
it was found that HDPE is too brittle and LDPE softens at
too low a temperature for incorporation into a cable [9]. Blending
Improved thermo-mechanical performance can, however,
be achieved by blending LDPE with small amounts of Four binary blends were prepared by the process of solu-
HDPE [9] offering a recyclable composite insulation sys- tion blending according to published techniques [11], as
tem with comparable electrical and thermo-mechanical detailed in Table 2. The required pellets of each material
properties to XLPE. were carefully weighed using a digital balance (Ohaus) and
Whilst the above ethylene-based systems could be used placed into xylene (1% w/v concentration); the xylene was
to provide recyclable replacements for XLPE for the next then boiled with vigorous stirring for a period of 20 min, a
generation of high voltage cables, if enhanced dielec- reflux condenser being used to recirculate any solvent
tric properties and increased operating temperatures are vapours. Meanwhile the same volume of methanol was
required, then propylene-based materials would be more cooled to *0 °C in a freezer. After blending was complete,
suitable. Indeed, isotactic polypropylene (iPP) was con- the xylene solution was poured into the methanol with
sidered during the early stages of cable development vigorous stirring. The resulting precipitate was allowed to
[13, 14], but was not widely accepted due to its inherent cool, filtered to remove the solvent and then dried at room
rigidity and the reduced electrical strength that results from temperature to constant mass in a drying cabinet vented
the formation of large spherulites [15, 16]. Nevertheless, externally. Final degassing was carried out in aluminium
the use of syndiotactic rather than isotactic polypropylene trays for 1 h at 200 °C under dynamic vacuum.
(sPP) may overcome many of these difficulties by virtue of
its improved ductility and increased spherulitic nucleation Sample preparation
density [13, 14]. In addition, recently a wide range of
propylene co-polymers have become commercially avail- Samples of the required polymer or blend were melted at
able which also offer improved ductility and therefore, with 200 °C for at least 2 min and then pressed into the required
suitable thermal processing, may offer additional candi- form. After pressing, the samples were held at this tem-
date materials for future enhanced propylene-based cable perature for *5 min to relax and were then either quenched
systems [17]. directly into water at 25 °C or crystallised isothermally in
Here we report on a study of a range of propylene-based oil baths (Grant, accuracy ± 1 °C). Plaques nominally
co-polymers as well as traditional isotactic (iPP) and syn-
diotactic (sPP) polypropylene. These are prepared through
a traditional Ziegler–Natta (ZN) catalyst process and were Table 1 Polymers used in this investigation
obtained commercially and therefore represent a typical Material Co-unit Co-unit Mw Mn
cross-section of currently available systems. We have content type (g/mol) (g/mol)
considered their thermal, thermo-mechanical and electrical iPP – – 163,000 47,700
properties as well as their morphology to assess which of sPP – – 157,000 43,200
these materials provide the best balance of properties for PE02 2 Ethylene 283,000 36,000
use in the next generation of high performance extruded
PE12 12 Ethylene 135,000 47,700
high voltage power cables. Finally, we explore the use-
PE40 *40 Ethylene 49,000 10,400
fulness of blending as a means of further optimising
PB12 12 Butene 254,000 74,000
properties.

123
4060 J Mater Sci (2011) 46:4058–4070

Table 2 Blends investigated in this study thermal analysis (DMTA) was undertaken at temperatures
Designation Composition
between -20 and 140 °C using a Rheometrics RSA II at a
fixed frequency of 100 rad/s and an applied strain of
Blend 1 20% iPP in sPP 0.02%. A dual cantilever geometry (40 9 5 mm) was used
Blend 2 50% iPP in sPP for most of the tests. For some of the low modulus mate-
Blend 3 20% iPP in PE40 rials this was impractical and therefore a parallel plate
Blend 4 50% iPP in PE40 compression fixture (diameter 12 mm) was used in its
place and to ensure linearity, a fixed 1% compressive strain
was applied.
1.5 mm in thickness were prepared using a hydraulic Electrical breakdown tests were performed according to
workshop press and aluminium moulds; strips 50 mm in ASTM D149-87 using a standard testing protocol [3, 9, 11,
length by 5 mm width were then cut for mechanical testing. 12, 17]. Testing was undertaken on film samples placed
Film samples of 70 lm nominal thickness for electrical between opposing 6.25 mm steel ball bearings immersed in
breakdown testing were prepared using a Grasby-Specac a tank of silicone fluid (Dow 200/20 CS). An increasing
25.011 hydraulic press equipped with constant thickness AC (50 Hz) voltage was then applied at a rate of 50 V/s
film maker. until failure. The results from 20 such tests per material
Isothermal crystallisation times were inferred by melting were analysed using Weibull statistics (Reliasoft Weibull
*0.1 g of sample between a clean microscope and cover 7??).
slip (sample thickness *1 mm) and viewing visually the
crystallisation process using a polarising optical microscope
equipped with a Mettler FP82 temperature controlled hot- Results: single component systems
stage. The hotstage was calibrated using a temperature
probe, itself calibrated in ice/water mixture and boiling Crystallisation behaviour
distilled water, respectively. The time for crystallisation
was inferred by viewing three separate samples of each Figure 1 shows a plot of the time to impingement for all six
material and waiting for complete impingement of spheru- co-polymers, averaged over three separate samples and as
litic forms, once this had occurred the time was noted. The expected, the crystallisation time increases exponentially
specimen was then checked for any further changes over a with the crystallisation temperature [19]. At any given
similar timescale, if no such changes were observed then the crystallisation temperature, iPP crystallises much more
original time was taken as the crystallisation time (or time rapidly than sPP as expected [20]. Unexpectedly, PE12
to impingement) otherwise the time was revised accord- displays almost identical crystallisation behaviour to iPP
ingly and re-checked using a fresh sample. Samples which while, in PB12 and PE40 the co-unit acts in the expected
did not show evidence of observable nucleation within an manner to hinder crystallisation [21]. PE02 crystallises
hour were rejected and fresh samples were prepared, crys- significantly more quickly than iPP despite its increased
tallisation being attempted at a lower temperature.

Characterisation

Differential scanning calorimetry (DSC) of *5 mg sam-

ples was undertaken using a Perkin Elmer DSC-7; melting
traces were obtained at 10 K/min and the instrument was
calibrated using high purity indium. Samples for morpho-
logical examination were microtomed and then chemically
etched for a period of 6 h in a 1% solution of potassium
permanganate in an acid mixture composed of 2 parts water,
2 parts phosphoric acid and 5 parts sulphuric acid, accord-
ing to published techniques [18]. They were then mounted
onto standard aluminium SEM stubs, gold coated and
examined at 20 kV in a Cambridge Instruments Stereoscan
360 scanning electron microscope (SEM).
Tensile testing was undertaken at room temperature
using an Instron 4301 tensile tester at a fixed speed of Fig. 1 Measured time to impingement of all five materials from
5 mm/min (gauge length 25 mm). Dynamic mechanical optical measurements of three separate samples

123
J Mater Sci (2011) 46:4058–4070 4061

ethylene co-monomer content, increased molecular weight

and higher polydispersity (Table 1). Out of all the mate-
rials, PE40 displays the slowest crystallisation rates,
probably by virtue of its much higher co-monomer content
[21].
For convenience, a single isothermal crystallisation
temperature was selected to prepare further samples for
analysis; this was chosen to give an impingement time of
15 min (as indicated by the dotted line in Fig. 1). To ensure
that complete isothermal crystallisation occurs under these
conditions, a 1 h crystallisation time was allowed. Figure 1
shows that under these conditions a crystallisation tem-
perature of 120 °C is appropriate for iPP, PE02 and PE12,
90 °C is appropriate for sPP, 100 °C for PB12 and 80 °C
for PE40.

Differential scanning calorimetry

Figure 2a shows typical DSC melting traces for samples

quenched directly from the melt arranged in order of
decreasing enthalpy from top to bottom. Isotactic poly-
propylene has a singular melting transition located at
*160 °C and the highest enthalpy of all the materials
considered of *80 J/g, indicating a crystallinity of 38%
[22]; these values are broadly consistent with other inves-
tigations [17, 21]. sPP also displays a singular melting
transition which is located at *125 °C and a correspond-
ing enthalpy of *30 J/g (crystallinity of 14%); these val-
ues are again broadly consistent with other investigations
of this type of material [20].
PE12, by contrast, shows two distinct melting transitions.
The upper at *163 °C is close to that of the iPP and the
lower at *125 °C is close to that typically expected from
high density polyethylene [9, 11]. This, along with its
almost identical crystallisation behaviour to iPP (see
Fig. 1), suggests that phase separation of ethylene and
propylene rich sequences is occurring in this propylene/
ethylene co-polymer (i.e. the two components crystallise
essentially independently). For PE02, PB12 and PE40,
increasing co-unit concentration lowers the melting point
and enthalpy as reported elsewhere in analogous sets of
random propylene–ethylene co-polymers [17, 21]. By con-
trast, PE40 has a broad melting transition extending from 80
to 150 °C and an enthalpy of only *6 J/g indicating a
crystallinity of \5%. Comparing the observed melting
temperatures above with the selected crystallisation tem-
peratures indicates that there is a consistent undercooling of
*35 K across all materials. Fig. 2 DSC melting endotherms from a quenched samples, b iso-
thermally crystallised samples, c iPP crystallised at various temper-
Figure 2b shows analogous melting traces following atures to illustrate reorganisation effects. Curves displaced for clarity
isothermal crystallisation for 1 h as described above. In
some cases, the resulting melting transitions and enthalpies in our previous studies, which exhibited significant c
are very similar to those seen after quenching. Unlike the populations after isothermal crystallisation [17, 21], com-
sets of random propylene–ethylene co-polymers considered plimentary wide angle X-ray scattering on the current

123
4062 J Mater Sci (2011) 46:4058–4070

series did not reveal any evidence of c crystals [23]. This

behaviour indicates that not only do the current ZN cata-
lysed materials lack the random distribution of defects
required for the formation of the c phase [21] but also that
the double melting peaks occur as a result of reorganisation
during the DSC scan. Figure 2c illustrates this behaviour in
iPP crystallised isothermally at different temperatures.
After quenching or crystallisation at low temperatures,
extensive reorganisation occurs during the DSC scan to
give the upper melting peak (arrowed)—this is the peak
that is detected in quenched samples. As the crystallisation
temperature is increased, the upper melting peak is
diminished and is replaced by the lower melting peak with
a bi-modal behaviour evident at intermediate crystallisation
temperatures (i.e. 110 °C). Such reorganisation effects
account for the lowering of the apparent melting points of
iPP and PE02 after isothermal crystallisation, as well as the
observed multi-modal melting behaviour of sPP and PB12,
shown in Fig. 2b. Increasing the DSC scan rate is one way
to overcome such problems and allow the true melting
point to be determined. However, at the high scan rates
required, instrumental thermal lag becomes a serious issue
leading to much greater inaccuracies.
With the possible exception of PE40, whose melting
behaviour occurs over a very broad temperature range (so
is difficult to quantify), and taking into account variations
in observed melting point due to the above reorganisation
effects, all the materials appear to melt at a higher tem-
perature than the 105 °C typically associated with the
softening transition in XLPE [3, 24]. Therefore, despite the
increased variability inherently associated with the current
ZN catalysed materials, compared to the previously studied
random propylene co-polymers [17, 21], at least five out of
the six investigated materials appear to be suitable as high
performance cable materials, at least from the point of view
of their melting behaviour.

Morphology

Figure 3 shows the morphology of selected quenched sam-

ples. iPP shows some evidence of spherical domains
(Fig. 3a) whereas sPP, PE02 and PB12 all show a uniform
lamellar texture devoid of spherulites, an example of which
is shown in Fig. 3b. All of these materials possess a rela-
tively high crystallinity and appreciable melting enthalpies Fig. 3 SEM micrographs of quenched samples a iPP, b PE02, c PE12
in the DSC so the presence of distinct crystal forms in these
materials is not surprising [15, 17, 25]. By contrast, PE12 physical phase separation to form broadly circular domains
(Fig. 3c) exhibits a two-phase texture which is consistent within the majority propylene rich phase. Not surprisingly,
with its bi-modal DSC melting behaviour. It is well known the domains, being composed primarily of more defective,
that in ZN catalysed polymers such as these, that the defects lower molecular weight material, have etched more strongly
are strongly concentrated in the molecules with the than the surrounding matrix [18]. Finally, PE40 exhibits a
lowest molecular weight [26]. In this case the shortest ‘‘featureless matrix’’ (at SEM resolution) which is consistent
most ‘‘defective’’ or ‘‘ethylene rich’’ molecules undergo a with its very low crystallinity as measured by DSC.

123
J Mater Sci (2011) 46:4058–4070 4063

Isothermal crystallisation results in much more organ- elsewhere in polypropylene systems which have been
ised textures as shown in Fig. 4. iPP (Fig. 4a) displays deliberately nucleated with dibenzylidene-sorbitol [27].
small, well-defined spherulites *30 lm in diameter with
prominent boundaries, whereas sPP (Fig. 4b) shows similar Tensile testing
forms (*20 lm in diameter) but with rather more dif-
fuse boundaries. By contrast, PE02 (Fig. 4c) displays a Typical stress–strain curves obtained from the Instron
uniform lamellar texture, rather similar to that observed on tensile tester are shown in Fig. 5 for all six materials
quenching the same material (see Fig. 3b) and an absence considered. The stress was proportional to the strain up to
of spherulitic forms, again PE12 (Fig. 4d) displays a two- *2% which allowed the modulus to be calculated. All of
phase texture again, rather similar to that exhibited the materials, with the exception of PE40, showed a sim-
after quenching. Given its similar crystallinity to sPP, ilarly high modulus of *109 Pa, which is in line with
PB12 (Fig. 4e) displays a texture of spherulites (diameter earlier studies [17]. Following quenching (Fig. 5a), iPP
*50 lm) with diffuse boundaries. Finally the rather low exhibits a brittle failure at *10% strain, whereas PE12
crystallinity PE40 displays no evidence of crystal forms at exhibits a similar failure at *15% strain. The remaining
SEM resolution so is not shown here. A similar progres- materials, with the exception of PE40, show a character-
sion, from well-defined spherulitic morphologies to non- istic necking behaviour above *20% strain and thereafter
crystalline disorganised textures with increasing co-unit do not recover their original shape on removal from the
content, has been reported elsewhere [17, 25] and the tester. By contrast, PE40 shows a much lower modulus
current series, with few exceptions, appears to follow this (*106 Pa) and recovers its original dimensions immedi-
morphological trend. ately after removal from the tester (i.e. it has a rubbery
Finally, the two-phase texture exhibited by PE12 over nature). Under isothermal conditions (Fig. 5b) the charac-
all crystallisation conditions, together with its bi-modal teristics are very similar, the exception being that PE12
DSC melting behaviour and its rather similar isothermal now exhibits a brittle failure at a much lower strain.
crystallisation behaviour to iPP (see Fig. 1) suggests that By contrast, similar tests on XLPE [3] and analogous
ethylene sequences are being excluded during isothermal ethylene-based systems [9] reveal a modulus at room
crystallisation (i.e. PE12 crystallises rather similarly to temperature of *108 Pa, therefore the majority of the
iPP). Furthermore, the distinctive lack of spherulites, even current propylene-based systems, with the exception of
under isothermal conditions, and the unexpectedly rapid PE40, are much stiffer than XLPE at room temperature,
crystallisation associated with PE02 suggests that this which may make them unsuitable for cable applications.
material may contain a nucleating agent. Rather similar Whilst the brittle nature of iPP is widely appreciated [13],
morphologies to that shown in Fig. 4c are reported the fact that PE12 shows a very similar behaviour was

Fig. 4 SEM micrographs of a iPP crystallised at 120 °C, b sPP crystallised at 90 °C, c PE02 crystallised at 120 °C, d PE12 crystallised
at 120 °C, e PB12 crystallised at 100 °C

123
4064 J Mater Sci (2011) 46:4058–4070

Fig. 6 DMTA test results from a quenched samples, b isothermally

crystallised samples
Fig. 5 Tensile test results from a quenched samples, b isothermally
crystallised samples

constant at *106 Pa and only decreases significantly from

initially somewhat surprising. However, it is thought that its room temperature value above 100 °C. Out of the
this is due to the two-phase morphology of this material; remaining materials, iPP shows the highest modulus and
the interfaces between the ethylene rich domains and the sPP the lowest, whereas PE02, PE12 and PB12 display
surrounding propylene rich matrix could act as sites for intermediate values with PB12 having the lowest modulus.
crack propagation. Clearly, for a cable application, a more Clearly the inclusion of a co-monomer serves to lower the
flexible, less brittle system is preferred, which clearly rules modulus and improves the mechanical flexibility [17], with
out both iPP and PE12. 1-butene having a greater effect for the same co-monomer
content. Compared with XLPE, which has a modulus
Dynamic mechanical thermal testing which falls from 108 Pa at room temperature to *107 pa at
90 °C and\106 Pa at 110 °C [3], and LDPE which exhibits
Values of the mechanical modulus measured as a function very poor mechanical properties above 60 °C [9], all of the
of temperature are shown in Fig. 6 and, as in the tensile current polypropylenes exhibit a significantly improved
testing results shown above, the crystallisation route has modulus at higher temperatures. Whilst a high modulus at
only a minor effect on mechanical modulus. The results at low temperatures may cause problems for cable installa-
room temperature confirm those obtained from tensile tion, the good high temperature performance could be a
testing; five materials show a rather similar behaviour, significant advantage at cable operating temperatures,
whereas PE40 displays a significantly lower modulus. conferring sufficient mechanical integrity to the cable
Despite the rubbery behaviour of PE40, its modulus is without the need for cross-linking.

123
J Mater Sci (2011) 46:4058–4070 4065

Electrical breakdown testing However, under conditions of isothermal crystallisation

(Fig. 7b), the breakdown performance of iPP and PE12 are
Weibull plots of the electrical breakdown test results are significantly reduced [17] whereas the other materials offer
shown in Fig. 7. Unlike the mechanical properties or comparable (within the experimental uncertainties of
melting behaviour, which show very little dependence on ±10 kV/mm) breakdown strengths to their quenched
the crystallisation route, the electrical breakdown strength counterparts. Under isothermal crystallisation conditions,
is significantly influenced by this variable. Under condi- either spherulites with prominent boundaries or two-phase
tions of quenching (Fig. 7a), all six materials with the structures are formed in iPP and PE12, respectively. It
exception of PE40 exhibit a characteristic Weibull break- therefore appears that the same boundary regions that serve
down strength of 160–170 kV/mm and a shape parameter to increase mechanical brittleness also serve to reduce
ranging between 12 and 20. By contrast, PE40 has a electrical breakdown strength [15–17]; both the mechanical
breakdown strength of only 49 kV/mm with a shape brittleness and poor breakdown strength provide us with
parameter of only 5. By contrast, identical tests on XLPE very good reasons for rejecting both iPP and PE12 as
tapes (not shown here for clarity) reveal breakdown potential candidate cable materials.
strengths of between 130 and 150 kV/mm, so under con-
ditions of a rapid quench, five of the six materials out- Discussion
perform XLPE.
Whilst all of the materials so far examined exhibit one or
more desirable properties, none possess all of the require-
ments of low temperature flexibility, high temperature
integrity and high breakdown strength required of an ideal
thermoplastic insulation system. Further optimisation of the
properties is therefore required before any of the current
materials would be suitable for future recyclable cables. Out
of the six materials considered, sPP and PE40 provide the
most desirable combination of good breakdown strength
and thermo-mechanical performance. Blends based on these
materials and iPP will be examined to determine if their
thermo-mechanical and electrical breakdown properties can
be further optimised. In a manner akin to previous HDPE/
LDPE blends [9, 11, 12], crystallisation was performed at
120 °C for 1 h to ensure complete crystallisation of the iPP
within a molten matrix of either sPP or PE40.

Results: blend systems

Differential scanning calorimetry

Figure 8a contains typical DSC melting traces from the

four quenched blends. In all of the blend systems, the
melting peak at *160 °C can be associated with the iPP
component and this is larger in Blends 2 and 4 as antici-
pated (see Table 2). In Blends 1 and 2, a second melting
peak associated with the sPP component is easy to locate at
*125 °C whereas the low crystallinity PE40 component is
much more difficult to detect in Blend 4, manifesting itself
as a broad ‘‘hump’’ in Blend 3 (arrowed). The compositions
of all the blends were checked by determining the enthalpy
of the peak associated with the iPP and comparing it to the
results obtained by crystallising this material in isolation
Fig. 7 Electrical breakdown results from a quenched samples, over at least three separate samples from the degassed
b isothermally crystallised samples blends. In all blends the composition was accurate to ±5%

123
4066 J Mater Sci (2011) 46:4058–4070

across the samples; that is in the blends, the iPP peak thermodynamic stability of the iPP crystals that are formed
showed an enthalpy of *16 and *40 J/g for blends con- isothermally within a molten matrix composed of an
taining 20% and 50% iPP, respectively. increasing concentration of PE40 (which is effectively an
Figure 8b shows analogous data for the isothermally amorphous material).
crystallised blends. Blends 1 and 2 show an almost iden- The similarity of the melting behaviour of the PE40 and
tical melting behaviour to their quenched counterparts with sPP components to those manifested after quenching
the exception of a small peak at *135 °C, which may
indicate that a small fraction of the sPP has undergone
isothermal crystallisation. However, the behaviour of the
iPP component is quite different in Blends 3 and 4. The
double melting behaviour again indicates that significant
reorganisation of the iPP is occurring during the DSC scan
itself, as wide angle X-ray analysis did not reveal any
corresponding c population. The reorganisation which
serves effectively to ‘‘anneal’’ crystals from the lower
melting to the upper melting peak is clearly more pro-
nounced in Blend 3 which has the lowest in iPP content.
This would indicate a progressive reduction of the

Fig. 8 DSC results for the blends a quenched, b isothermally Fig. 9 SEM micrographs of quenched blends a Blend 1, b Blend 2,
crystallised c Blend 4

123
J Mater Sci (2011) 46:4058–4070 4067

indicates that these components undergo very little crys- more apparent and better organised following isothermal
tallisation at 120 °C, as required. crystallisation, the crystallisation of iPP is clearly impeded
in Blends 3 and 4 as suggested by the DSC data.
Morphology
Tensile testing
Figure 9 contains SEM micrographs of the quenched blend
systems. Blends 1 and 2 exhibit crystals of iPP in a matrix Figure 11a shows tensile testing data for the quenched
of sPP; in Blend 1 (Fig. 9a) these crystals are isolated and blends. Blends 1 and 2 exhibit the characteristic ‘‘necking’’
do not possess any long range order whereas, in Blend 2 behaviour that was previously associated with several of the
(Fig. 9b), a texture of open spherulites is formed. In Blend co-polymers but, crucially, the two blends do not show the
3 very little lamellar detail can be resolved whereas, in brittle behaviour associated with iPP. The modulus of both
Blend 4 open, tenuous spherulitic structures are observed systems at small strains is *109 Pa and is therefore similar
(Fig. 9c). to sPP. By contrast, Blends 3 and 4 exhibit a rubbery
Similar morphologies are exhibited after crystallisation behaviour, with no signs of necking, but with a higher
at 120 °C (Fig. 10), Blend 1 contains individual crystals of modulus than PE40, *107 and 108 Pa, respectively; clearly
iPP in an sPP matrix (Fig. 10a) and in Blend 2 (Fig. 10b) the iPP has a stiffening effect.
these crystals are organised into spherulitic forms con- After isothermal crystallisation (Fig. 11b), all four
taining short ‘‘blocky’’ crystals. In Blend 3 (Fig. 10c) the blends show the brittle fracture behaviour associated with
organisation is such that tenuous spherulitic forms can be iPP. Blends 1 and 3, having the least iPP content, show the
resolved in the SEM and finally, in Blend 4 the spherulites highest strain at failure, *20 and 40%, respectively, whilst
become more space filling and are better resolved at SEM the corresponding blends with 50% iPP show less favour-
resolution (Fig. 10d). able properties. Again more iPP leads to a more brittle
With both sPP and PE40 as matrix, blends with iPP material. This indicates that isothermal crystallisation of the
contents of only 20% tend to form individual crystals within iPP under conditions where large spherulites are formed
that matrix, whereas increasing the iPP content to 50% leads (see Fig. 10) has detrimental effects on the mechanical
to open spherulitic forms. Whilst the structures are clearly properties, even within a blend system.

Fig. 10 SEM micrographs of isothermally crystallised blends a Blend 1, b Blend 2, c Blend 3, d Blend 4

123
4068 J Mater Sci (2011) 46:4058–4070

Fig. 12 DMTA results from a quenched blends, b isothermally

Fig. 11 Tensile test results from a quenched blends, b isothermally crystallised blends
crystallised blends
Electrical breakdown testing

Dynamic mechanical thermal testing Figure 13a shows the electrical breakdown data for the
quenched blends. Blends 1 and 2 offer comparable break-
Figure 12 shows DMTA data from quenched and isother- down performance to sPP (*180 kV/mm); the addition of
mally crystallised samples. As discussed above, Blends 1 iPP to sPP clearly has a negligible effect on the electrical
and 2 possess a modulus at room temperature of *109 Pa properties. Blend 3 shows the lowest breakdown strength
so are probably too stiff for use in cables. However, Blends of the four blends (*130 kV/mm) but is significantly
3 and 4 offer a useful compromise in properties with Blend improved compared to PE40 (*50 kV/mm, see Fig. 7a);
4 (50% iPP in PE40) approaching the ‘‘ideal’’ thermo- the addition of iPP therefore overcomes the low breakdown
mechanical characteristic associated with XLPE [3]. strength associated with PE40 in isolation. The addition of
Despite these issues, all the blends appear to give sufficient further iPP (Blend 4) increases the breakdown strength, but
mechanical rigidity at temperatures exceeding 110 °C, only to *140 kV/mm.
with the high temperature stability being improved after Figure 13b shows analogous data for isothermally crys-
isothermal crystallisation (Fig. 12b). Unfortunately, such tallised samples, the relative order of the blends is preserved
blends can show brittle behaviour (Fig. 11b), particularly and Blends 1 and 2 exhibit no significant change in their
in those systems that exhibit distinct inter-spherulitic electrical breakdown strengths compared to quenching. By
boundaries, for example, Blend 4. In this regard rapid contrast, Blends 3 and 4 show markedly lower breakdown
crystallisation is preferable to isothermal crystallisation. strengths than their quenched counterparts (*120 and

123
J Mater Sci (2011) 46:4058–4070 4069

thermo-mechanical properties to XLPE, combined with a

breakdown strength which is much better than that mea-
sured from comparable XLPE samples. Rapid crystallisa-
tion of this material ensures the highest breakdown strength
combined with good mechanical flexibility at low tem-
peratures and adequate mechanical integrity up to 120 °C.
By contrast, isothermal crystallisation has deleterious
effects on mechanical properties and can reduce the elec-
trical breakdown strength to a degree.

Conclusions

1. Six propylene-based systems have been studied by

differential scanning calorimetry, scanning electron
microscopy, tensile testing, dynamic mechanical ther-
mal analysis and by electrical breakdown testing to
determine which could be suitable for inclusion into
future recyclable high voltage cables.
2. Whilst all six materials had sufficiently high melting
temperatures for cable applications, five of the materials
had a mechanical modulus that was too high and the
remaining material had a mechanical modulus that was
too low for cable applications. However, these materials
do have the advantage that their mechanical properties
are retained up to sufficiently high temperatures for
cable applications without the need for cross-linking.
3. Two of the materials, iPP and PE12 showed rather
brittle behaviour and poor electrical breakdown per-
formance when crystallised at high temperatures and
were rejected early on as candidate cable materials.
Fig. 13 Electrical breakdown results from a quenched blends, Whilst all of the materials studied exhibit one or more
b isothermally crystallised blends desirable characteristics, none possesses all of the
requirements of low temperature flexibility, high
*140 kV/mm, respectively). Despite these problems, temperature integrity and high breakdown strength
unlike iPP crystallised isothermally (see Fig. 7b) none of required by an ideal thermoplastic insulation system.
the blends exhibit a dramatic decrease in breakdown 4. The technique of blending was used to determine if
strength following isothermal crystallisation suggesting that such an optimal balance of mechanical, thermal and
the inter-spherulitic regions have been reinforced electri- electrical properties could be achieved using the
cally by the addition of the low melting component. current materials. Two pairs of blends based on sPP
and PE40 were investigated and iPP was incorporated
Discussion at concentrations of 20 and 50%.
5. A blend containing 50% iPP in PE40 (Blend 4) appears
Blends 1 and 2 clearly offer the best breakdown perfor- to provide the best balance of properties for use in a
mance out of the four blend systems, but are too stiff at future cable system; rapidly crystallising these systems
room temperature for use in a practical cable. In particular, gives sufficient mechanical flexibility combined with
after isothermal crystallisation these blends take on the adequate electrical breakdown strength whilst slow
brittle nature of iPP. crystallisation can be deleterious for mechanical
PE40 clearly has very desirable mechanical properties properties and makes the materials more brittle.
but has very poor breakdown performance. By adding iPP,
its breakdown performance can be dramatically improved Acknowledgements The GPC analysis on all the materials was
without sacrificing mechanical properties. Out of the two kindly carried out by Smithers Rapra using their Polymer Laborato-
PE40-based blends considered, Blend 4 offers very similar ries GPC220. We would like to thank Mr Luca Cozzarini of the

123
4070 J Mater Sci (2011) 46:4058–4070

DMRN, University of Trieste, Italy, for kindly carrying out the wide 12. Hosier IL, Vaughan AS, Swingler SG (2000) J Polym Sci: Part B
angle X-ray scattering. The work is funded through the EPSRC 38:2309
Supergen V, UK Energy Infrastructure (AMPerES) grant in collab- 13. Yoshino K, Ueda A, Demura T, Miyashita Y, Kurahashi K,
oration with UK electricity network operators working under Ofgem’s Matsuda Y (2003) In: Proceedings of the 7th international con-
Innovation Funding Incentive scheme; full details can be found on ference on properties and applications of dielectric materials, vol
http://www.supergen-amperes.org. 1–3, pp 175–178
14. Kurahashi K, Matsuda Y, Miyashita Y, Demura T, Ueda A,
Yoshino K (2004) Electr Eng Jpn 146:18
15. Zhao Y, Vaughan AS, Sutton SJ, Swingler SG (2001) Polymer
References
42:6599
16. Katsunami K, Ishii K, Tanaka Y, Ohki Y (1991) In: 3rd Inter-
1. Vaughan AS, Zhao Y, Barré LL, Sutton SG, Swingler SG (2003) national conference on properties and applications of dielectric
Eur Polym J 39:355 materials, vol 1–2, pp 999–1002
2. Marzinotto M, Mazzanti G, Mazzetti C, Pornpili M, Santulli C, 17. Hosier IL, Cozzarini L, Vaughan AS, Swingler SG (2009) J Phys:
Schiaffino P (2007) In: 2007 Annual report of conference on Conf Ser 183(012015):1
electrical insulation and dielectric phenomena, pp 49–52 18. Olley RH, Bassett DC (1982) Polymer 23:1707
3. Martin CP (2004) The impact of mechanical stress on the 19. Gedde UW (2001) Polymer physics. Kluver Academic Publish-
mechanical integrity of XLPE cables. PhD thesis, University of ers, Dordrecht
Southampton, UK 20. Wang ZG, Phillips RA, Hsiao BS (2001) J Polym Sci Part B:
4. White CC, Wagenblast J, Shaw MT (1999) In: 57th Annual Polym Phys 39:1876
technical conference SPE I-III, pp 3193–3197 21. Hosier IL, Alamo RG, Esteso P, Isasi JR, Mandelkern L (2003)
5. Tokuda S, Horikawa S, Negishi K, Uesugi K, Hirukawa H (2003) Macromolecules 36:5623
Furukawa Rev 23:88 22. Isasi J, Mandelkern L, Galante MJ, Alamo RG (1999) J Polym
6. Ashihara S, Goto T, Yamazaki T, Tomimatsu K, Ito H (2008) Sci Part B: Polym Phys Ed 37:323
In: Proceedings of 2008 international symposium on electrical 23. De Rosa C, Auriemma F, Circelli T, Waymouth RM (2002)
insulating materials, pp 522–525 Macromolecules 35:3622
7. Ray I, Khastgir D (1994) Plast Rubb Compos Process Appl 22:37 24. Hosier IL, Vaughan AS, Campus A, Nilsson U (2007) In: Pro-
8. Jana RN, Nando GB, Khastgir D (2003) Plast Rubb Compos ceedings of international conference on solid dielectrics,
32:11 pp 227–230
9. Hosier IL, Vaughan AS, Swingler SG (2010) J Mater Sci 25. Hosier IL, Alamo RG, Lin JS (2004) Polymer 45:3441
45:2747. doi:10.1007/s10853-010-4262-5 26. Lotz B, Graff S, Straupe C, Wittman JC (1991) Polymer 32:2902
10. Heo J, Lim KJ, Lim JN, Jung EH, Lee HK (2009) In: Proceedings 27. Zhao Y, Vaughan AS, Sutton SJ, Swingler SG (2001) Polymer
of international conference on electrical engineering 2009, pp 1–4 42:6587
11. Hosier IL, Vaughan AS, Swingler SG (1997) J Mater Sci
32:4523. doi:10.1023/A:1018617200285

123