Accepted Manuscript

Processability and mechanical properties of extensively recycled high density

polyethylene

Pavel Oblak, Joamin Gonzalez-Gutierrez, Barbara Zupančič, Alexandra Aulova, Igor

Emri

PII: S0141-3910(15)00023-3
DOI: 10.1016/j.polymdegradstab.2015.01.012
Reference: PDST 7556

To appear in: Polymer Degradation and Stability

Received Date: 30 October 2014

Revised Date: 17 January 2015
Accepted Date: 20 January 2015

Please cite this article as: Oblak P, Gonzalez-Gutierrez J, Zupančič B, Aulova A, Emri I, Processability
and mechanical properties of extensively recycled high density polyethylene, Polymer Degradation and
Stability (2015), doi: 10.1016/j.polymdegradstab.2015.01.012.

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

Processability and mechanical properties of

extensively recycled high density polyethylene

Pavel Oblak*, Joamin Gonzalez-Gutierrez, Barbara Zupančič, Alexandra Aulova and

Igor Emri

PT
Center for Experimental Mechanics, Faculty of Mechanical Engineering, University of Ljubljana and
Institute for Sustainable Innovative Technology, Pot za Brdom 104, 1125 Ljubljana, Slovenia

* e-mail address of corresponding author: pavel.oblak@fs.uni-lj.si (Pavel Oblak)

RI
Keywords

SC
High density polyethylene, mechanical recycling, extrusion, processability, mechanical
properties, rheology

Abstract

U
In plastics industry it is a common practice to mechanically recycle waste material arising
AN
from production. However, while plastics are mechanically recycled, their mechanical properties
change. These changes may affect material processing conditions and quality of the end
M

products; therefore they need to be quantified. In this study, mechanical recycling of high density
polyethylene (HDPE) was simulated by one-hundred (100) consecutive extrusions cycles.
During extrusion, processability of virgin HDPE and its recyclates was studied by recording the
D

processing conditions, i.e., melt pressure and extrusion torque, while after preparation of the
TE

recyclates, melt flow index measurements (MFI), small amplitude oscillatory shear tests
(rheological properties), and differential scanning calorimetry measurements (DSC) of thermal
properties were performed. Also, mechanical properties in solid state were characterized in
EP

terms of hardness and modulus measured by nanoindentation, and finally, shear creep
compliance was measured to characterize the materials’ time-dependent mechanical properties
C

and its durability in solid state. In addition, gel permeation chromatography (GPC) and solubility
tests were implemented to study changes in the material structure.
AC

The results on rheological and MFI measurements indicate significant structural changes in
the material that occurred during the first 30 extrusion cycles. Those changes affect material
processability which is as well supported by the recorded processing pressure and torque. On
the other hand, processing did not significantly affect material thermal properties. Results on
hardness and modulus show deterioration of the material mechanical properties after 10th
reprocessing cycle. Similarly, shear creep compliance measurements showed an unfavourable
effect of mechanical recycling on the time-dependent mechanical properties, particularly after
the 30th extrusion cycle. In addition, results suggested chain branching as a dominating
 ACCEPTED MANUSCRIPT

mechanism through first 30 extrusion cycles, domination of chain scission afterwards and also
presence of cross-linking after 60th extrusion cycle.

1. Introduction
In the second half of the 20th century, plastics became one of the most universally-used and

PT
multipurpose materials in global economy. They are used in wide range of applications, for
instance, packaging, household appliances, sport accessories, electronics, medicine, and as
well in high performance components in automotive and aircraft industry. The plastics industry

RI
has benefited from 60 years of growth with a year expansion of 8.7% from 1950 to 2012 [1].

SC
With continuous growth of plastic industry also amount of generated waste is increasing, not
only as post-consumer waste but also as waste generated during the production process.
Runners, sprues and off-specification products are common in production; they are easy to

U
identify and are of high quality [2]. Instead of being rejected as waste they are granulated and
AN
reprocessed; this makes economic sense as it reduces both production waste and utilisation of
raw materials. Reprocessing of production waste is particularly important in cases where
amount of waste is high in respect to the product. As an extreme case, micro injection moulding
M

can be exposed; in its case the volume of delivery system, i.e. sprue and runner systems, which
represent the waste material, can exceed that of the injected part by a factor 100 or more [3].
D

Another example is extrusion blow-moulding technology where amount of the scrap during
production can reach 40% [4].
TE

Reprocessing of production waste is also known as in-house or primary recycling and

represents a type of mechanical recycling where plastics are grinded and then processed
EP

through a physical process [5]. In these cases material can be exposed to excessive number of
reprocessing that may affect its processability and quality of end products.
C

Properties of mechanically recycled polymers do not remain the same because of

degradation from heat, mechanical stresses and oxidation during their reprocessing.
AC

Degradation affects molecule structure and consequently material mechanical properties [6, 7]
thus the quality is the main issue when dealing with mechanically recycled products [8]. Besides
mechanical, also rheological properties are changed [9, 10, 11]. This has direct effect on the
processability and needs to be considered during the following technological processes.

In line with mentioned above, our investigation is focused on study of processability and
mechanical properties of two widely used polymers, namely LDPE and HDPE, which shared
29.5% of European plastics demand in 2012 [1]. Both, LDPE and HDPE, belong to the group of
polyethylenes, however, their behavior during mechanical recycling differs significantly, and one
cannot be predicted from another Thus they need to be studied individually. Findings on
 ACCEPTED MANUSCRIPT

extensively recycled LDPE have been already reported by our group [12], whereas here we
summarize results on mechanical recycling of HDPE material.

2. Material and methods

2.1 Material

PT
For the purposes of the investigation, high density polyethylene (HDPE) LANUFENE®HDI-

RI
6507UV, produced by Ras Lanuf Oil & Gas Processing Co. was used. The material has density
0.965 g/cm3 (23°C) and melt flow index 7.5 g/10min (190°C and 2.16kg) [13]. Furthermore,

SC
producer provides information on suitability of material for the manufacturing of crates,
containers, trays and other similar injection moulded objectives where good weather-ability and
mechanical properties are needed. In addition, producer gives information that material contains

U
UV light stabilizers and recommends its processing temperature to be 220 – 270°C [13].
AN
2.2 Simulation of mechanical recycling
M

The process of mechanical recycling was simulated by extensive extrusion. For this purpose,
twin-screw extruder, PolyLab PTW 16/40 OS, produced by Thermo Scientific (Germany) was
D

used. Extruder had installed horizontal rod die (type 557 - 3235) and a nozzle with diameter of
4mm. The arrangement of the co-rotating extrusion screws used for the process of simulated
TE

recycling is presented in Fig.1.

C EP
AC

Figure 1: Used configuration of extruder screws

 ACCEPTED MANUSCRIPT

Material was extruded at screw rotation of 150 min-1, and processing temperature of 240 °C
with a throughput between 1200 and 1300 g/h. After material was extruded, it was pelletized
using a Thermo Haake pelletizer (type 557-2685) and a portion was removed for the purpose of
afterwards characterization. The rest of material was submitted to a new extrusion cycle.

2.3 Sample preparation

PT
In order to perform characterization of recycled materials, samples of cylindrical shape with a
diameter of 6 mm and length of 150 mm were prepared. For the preparation, a non-commercial

RI
apparatus for the gravimetric casting was used [14]. Gravimetric casting proved to be the most
suitable method for assuring isotropy of samples. A glass tube Ø9 x Ø6 x 200 mm was placed

SC
between upper and bottom fixation elements (Fig.2). A heater at 200°C in the shape of a hollow
cylinder was positioned around the glass tube and travelled with 4 mm/min from the bottom to

U
the top of the tube. The glass tube was filled with a polymer granulate, between 20 and 30 mm
above the upper surface of the heater. Compression force was assured with the piston having a
AN
loading mass of 1100g on top of it (Fig.2).
M
D
TE
C EP
AC

Figure2: Scheme of gravimetric casting procedure

2.4 Characterization

2.4.1 Measured processing parameters

PolyLab PTW extruder used for mechanical recycling simulation allows monitoring and
recording the screw torque, M(t), as well pressure of the melt, p(t), in the die channel. Screw
 ACCEPTED MANUSCRIPT

torque was measured by built-in torque sensor (type 567-0100) while melt pressure was
recorded by pressure gauge (type 557-2571) installed in the die. During every extrusion cycle,
melt pressure and extrusion torque were recorded for 10 minutes/extrusion, with the data
acquisition sampling rate of 10 data points/minute.

2.4.2 Melt flow index (MFI) measurements

PT
The melt flow index (MFI) corresponds to the mass of polymer that passes through a
prescribed capillary in time of 10 min at given applied pressure (load). Measurements were

RI
carried out according to ASTM D 1238-04c. All the tests were performed on a D4004DE
instrument produced by Dynisco (USA). The MFI was measured at the temperature of 190°C by

SC
applying pressure with a load of 2.16 kg. For the virgin material and its recyclates, MFI was
measured in triplicate.

2.4.3 Viscosity measurements

U
AN
Viscosity measurements were performed in oscillatory mode by varying the frequency of
excitation (frequency sweep) in a Haake MARS (Modular Advanced Rheometer System) from
M

Thermo Scientific (Germany). Tests were implemented using a plate-plate measuring geometry
(diameter D = 20mm, gap h = 0.5mm) at 240°C and frequencies ω increasing from 0.1 Hz
D

(0.628 rad/s) to 100 Hz (628.32 rad/s). All measurements were performed by applying a shear
stress τ of 300 Pa, previously determined as high enough to obtain good resolution of results
TE

without leaving the linear viscoelastic region of molten HDPE. For virgin HDPE as well for its
recyclates, viscosity was measured with six repetitions. Within the current study, results on
EP

viscosity measurements are presented with the magnitude of complex viscosity, I η*I, which is
related to steady state viscosity, η , by the Cox-Merz rule:

[ ]
C

η *(ω) ⇔η(γ&) , when ω[rad/ s] ⇔γ& s−1 , (1)

AC

and I η*I is defined as[15]:

G '2 +G ' '2

η* = , (2)
ω

where G’ is the storage modulus and G’’ is the loss modulus; in other words the elastic and the
viscous properties of a material which are also presented.
 ACCEPTED MANUSCRIPT

2.4.4 Differential Scanning Calorimetry (DSC)

Differential scanning calorimetry (DSC) analysis was performed on DSC7 instrument,

produced by Perkin Elmer (USA). The measurements were carried out according to ISO 11357.
Mass of the specimens was 5.0±0.1 mg. Heating and cooling rate was 20 °C/min. All the
measurements were performed in a nitrogen atmosphere. Average values of six repetitions are
presented in this study. Melting temperature Tm and crystallization temperature Tc as well as

PT
heat of fusion Hf were evaluated from the second heating run. The degree of crystallinity Xcr
was calculated based on the next relation:

RI
∆ Hf
Xcr = , (3)
∆ H 100 %

SC
where ∆H100% of crystalline polyethylene is 293 J/g [16].

2.4.5 Nanoindentation
U
AN
Measurements of hardness and modulus were performed with nanoindentation using G200
apparatus, produced by Agilent Technologies (USA). The measurements were carried out using
M

continuous stiffness measurement technique (CSM) [17] with a Berkovich indenter tip while
oscillation frequency was 45 Hz and amplitude 2 nm. For virgin HDPE material as well for its
D

recyclates measurements were repeated on three samples, with 40 indentations on each

sample. In order to exclude surface effects, only data measured thought the depth of penetration
TE

from 1000 to 1500 nm were taken into account and are averaged in the presented results.
EP

2.4.6 Creep measurements

Shear creep is a process of deformation change over time, γ (t ) , occurring in a material

C

loaded with constant shear stress of magnitude τ 0 . The relation between the constant shear
AC

stress loading τ (t ) = τ 0 ⋅ h(t ) , where h(t) is a step function, and the time-dependent shear strain
response, γ (t ) , is given as [18]:

γ (t )
J (t ) = , (4)
τ0

where J (t ) is the shear creep compliance. Measurements were performed on Shear Creep
Torsiometer developed and set up by our group [19]. Measuring system (Fig. 3) consists of
three parts: (i) mechanical device enabling torsional loading of a test specimen in a solid state,
(ii) heating chamber, and (iii) measurements and regulation setup including computer and self-
 ACCEPTED MANUSCRIPT

developed software, used for data acquiring and analysis. All moving parts have been equipped
with aerostatic bearings to minimize friction in the measuring device. The smallest torque
loading available is 1 x 10-5 Nm, and the smallest measurable angle of rotation is 1.0 x 10-4 rad,
the angle measuring range is from 0 to 2.8° or 0.05 rad. The test specimen is placed in the
heating chamber built for temperatures from room temperature to 160 °C, which can maintain
the temperature at ±0.5 °C during the time interval of 104 s.

PT
RI
U SC
AN
M
D

Figure 3: Scheme of Shear Creep Torsiometer

TE

For the purpose of shear creep measurements, gravimetrically casted specimens were cut to
~40 mm in length and further fixed to metal holders specifically adapted for proper gripping in
the torsiometer; KemisKol F524 adhesive produced by KemisPlus (Slovenia) was used. Before
EP

performing creep measurements all samples were annealed (Phase i) to erase the residual
stresses, i.e. heated up at 100 °C over a period of 7 h, and then gradually cooled to 30 °C over
C

a period of 8 h (Fig. 4). After annealing, the measurement continued with loading (Phase ii),
unloading (Phase iii) and conditioning of the sample at 30, 50, 65, 75 and 90 °C, as shown in
AC

Fig. 4. Loading phase (Phase ii) lasted 3 h at a constant temperature, then the torque was
removed and the temperature was increased to the next measuring temperature within 3 h
(Phase iii). The specimen was again torque-loaded, and the procedure of exchanging the
phases of loading, unloading and conditioning was repeated over the selected measuring
temperature range (30 to 90 °C). The stresses appli ed to each specimen were selected as to
remain in the linear viscoelastic region for all materials and ranged from τ0 = 9.6 x 104 Pa at 30
°C to 3.0 x 10 4 Pa at 90 °C. Applied stresses were assured by load ing torque ranged from
0.0035 Nm at 30 °C to 0.0011 Nm at 90 °C. Schematic presentation of described procedure is
presented in Fig. 4. Two repetitions were performed at each condition for all the materials.
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
Figure 4: Schematic presentation of the temperature and mechanical profile of loading in the
torsional creep measurement
M

2.4.7 Gel permeation chromatography

D

Changes in the molecular weight were studied by implementation of high temperature gel
TE

permeation chromatography (HT GPC). Samples were introduces into the eluent, a
trichlorobenzene (TCB), at concentration 2.0 mg/ml. The analysis was carried out at 150 °C with
set of three Olexis columns of total length 900 mm. Injection volume was 200 µl and flow rate
EP

1.0 ml/min. A differential refractometer (DRI detector) was employed and data analysed using
Agilent Cirrus software. Presented results, namely molecular weight distribution, MWD, weight-
average molecular weight, Mw, number-average molecular weight, Mn, and are the result of one
C

injection per material studied.

AC

2.4.8 Solubility tests

To expand the study on changes of molecular structure, solubility test was also conducted to
define the insoluble part which confirms presence of cross-linked portion. Solubility tests were
performed by extraction in Soxhlet assembly. A portion of 1.5 g of a specimen material was
introduced into the Soxhlet thimble and extracted by xylene for 24 hours at 6 to 7 cycles per
hour. Mentioned technique provided the information on a weight percentage of the insoluble part
of a sample.
 ACCEPTED MANUSCRIPT

3. Results and discussion

3.1 Extrusion torque and pressure

Within this study, mechanical recycling was simulated with one hundred cycles of
consecutive extrusion. During implementation of extrusion, set processing conditions, i.e. barrel

PT
temperature, screw speed and material throughput, remained unchanged. Real time melt
pressure and extrusion torque were recorded at every extrusion cycle and averaged over 10
min. The results on measured pressure and torque as function of the number of extrusion cycles

RI
are presented in Fig. 5. Both curves exhibit similar trend; during the first 30 extrusion cycles
pressure and torque are linearly rising and start to decrease after the 35th extrusion. At 30th cycle

SC
pressure increases for more than 500% in respect to initial value while torque increases for
128% only. After the 100th extrusion cycle, pressure is still about 300% higher than that

U
measured during the 1st extrusion. At the same time this ratio for the torque is only 108%.
AN
Since set processing parameters were not changed trough the complete implementation of
extrusion process, it is possible to assume that all recorded changes in pressure and torque
result from changes in material inherent structure. Several authors reported that during the
M

reprocessing of HDPE, two competing processes take place, i.e., chain scission and chain
branching [20, 21, 22]. It has also been reported that cross-linking might be also present during
D

reprocessing [23, 24]. Continuous rising of extrusion pressure and torque through the first 30
extrusion cycles indicates reduction of a material flowability at a given shear rate. The reduction
TE

in flowability could be a consequence of cross-linking and/or chain branching process. However,

taking into account the obtained results on solubility test, which are presented in continuation,
EP

we might say that chain branching is the mechanism that prevails over cross-linking. On the
other hand, decreasing of pressure, which is evident after 35th extrusion, shows a slight
improvement of material flowability which results from higher mobility of the molecules and
C

indicates the presence of chain scission.

AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
Figure 5: Recorded extrusion pressure and torque in dependence of the number of extrusion
AN
cycles

Changes of melt pressure and extrusion torque are directly presenting changes in material
M

processing. Hence, in practice resetting the set processing parameters might be required when
using recycled HDPE. Moreover, production process becomes more energy-demanding and
D

potentially burdensome for the machinery when using recycled HDPE. Also, recycled material
may become unprocessable with certain type of machine, due to equipment capacity limitation.
TE

3.2 Melt flow index

EP

Melt flow index is one of the most commonly used values for quality control in industry and it
is an indicator of material changes. Melt indexer generally operates at low shear rates that are
C

not representative of shear rates usually encountered in the extrusion process. It is possible to
AC

estimate the shear rate at the capillary wall, which represents rough estimation of shear rate
[25]. In this particular case a shear rate of 14 s-1 for measurements on the virgin material and
even less than 0.5 s-1 for measurements on recyclates reprocessed more than 20 times, has
been estimated.

Several authors have investigated the effect of HDPE reprocessing on its MFI, however in all
cases number of reprocessing did not exceed 10 cycles. They presented various results;
increasing of MFI with number of reprocessing cycles [6, 26], decreasing [20, 23, 27, 28], or they
did not observe significant change [21, 24, 28]. Many authors concluded that intensity of MFI
change can be related to the content of stabilizers [6, 23, 28].
 ACCEPTED MANUSCRIPT

Results of our investigation, presented in Fig. 6, show MFI value of virgin material to be 7.5
g/10 min, which is the same as reported by the producer [13]. Furthermore, significant drop of
MFI is revealed in first extrusions, thereby the maximum drop of 98.8% is observed at 30th
extrusion cycle. After 30th cycle MFI practically does not change. Negligible increase after 30th
cycle can be seen, and seemingly negligible decrease in MFI value after 60th cycle can be
attributed to the experimental error, as evident from the magnified results in the upper-right

PT
corner in Fig. 6.

MFI index shows changes in material flowability which indicates changes in material

RI
structure. Drastic decrease of MFI shows lowering of chain mobility which suggests a branching
process present throughout the first 30 reprocessing cycles, similarly as suggested by

SC
processing parameters presented previously. Changes in previously presented processing
parameters (pressure and torque) also indicate cross-linking or chain branching processes after
35th cycle. If this was significant, MFI would increase, but increase of MFI after 35th cycle is

U
minor, that is because of the different shear rates experienced at MFI and at extrusion process.
AN
The effect of different shear rates will be observed better in the viscosity results.
M
D
TE
C EP
AC

Figure 6: Melt flow index (MFI) in dependence of the number of extrusion cycles

3.3 Viscosity

Rheological characterization namely viscosity measurements were performed on virgin

HDPE as well on its recyclates. Since simulative extrusion was carried out at 240°C,
measurements of viscosity were implemented at the same temperature, performing small
 ACCEPTED MANUSCRIPT

amplitude oscillatory shear test varying the frequency of excitation from 0.628rad/s to 628.32
rad/s.

Results are presented in Figs. 7 and 9 to 12. Complex viscosity in dependence of execution
frequency is shown in the Fig.7. As expected, viscosity is decreasing with increase of frequency
(shear thinning effect) for virgin material as well for all the presented recyclates. With the
increased number of recycling the shear thinning effect (inclination of the curve) substantially

PT
increases, Fig. 7. In Figure 8 the intensity of shear thinning effect is presented though the ratio
of complex viscosities, i.e., complex viscosity at frequency 0.629 rad/s divided by that at

RI
frequency 291.6 rad/s. Rising trend of shear thinning intensity is evident between 10th and 20th
extrusion cycles. Then, between the 20th and 60th extrusion cycles, one observes a plateau,

SC
followed by roughly linear increase of shear thinning with the number of recycling all the way up
to the 100th extrusion cycles.

Increasing of shear thinning intensity may indicate changes in inherent material structure,

U
which are most pronounced between 10th and 20th extrusion cycle. This observation and
AN
hypothesis is well supported also by changes of mechanical properties presented in later
sections of this article.
M
D
TE
C EP
AC

Figure 7: Variation of complex viscosity in dependence of frequency

 ACCEPTED MANUSCRIPT

PT
RI
U SC
Figure 8: Ratio of complex viscosity; complex viscosity at frequency 0.629 rad/s divided by that
AN
at frequency 291.6 rad/s

Further, complex viscosity in dependence of extrusion cycle is presented in Fig.9 for low
frequency, 0.628rad/s, and in Fig.10 for high frequency, 291.6 rad/s, loading regime. In both
M

cases, significant leap of complex viscosity is evident between 5th and 20th extrusion cycle.
Increase is more evident at low frequency, Fig.9, where the value at 20th cycle exceeds 600% of
D

the initial one, while under high frequency, Fig.10, an increase of 123% was recorded. After the
20th reprocessing cycle, complex viscosity at low frequency constantly fluctuates between 400
TE

and 700%, on the other hand, decreasing trend is recorded at high frequency. Also, it can be
said that viscosity at low frequency shows similar trend as MFI previously presented while on
EP

the other hand, trend of viscosity at high frequency is similar as trends of extrusion pressure and
torque.
C

Increase of complex viscosity between 5th and 20th shows lowering the chain mobility which
may result from chain branching. Its presence at consecutive processing was observed and
AC

reported by several authors [20, 21, 22]. On the other hand, dropping of complex viscosity at
high frequency indicates that chain scission may have occurred. In the literature, chain scission
is also a mechanism suggested to run simultaneously with chain branching when HDPE material
undergoes degradation during processing [20, 21, 22].
 ACCEPTED MANUSCRIPT

PT
RI
SC
Figure 9: Variation of complex viscosity in dependence of the number of extrusion cycles at min.

U
(0.628rad/s) frequency
AN
M
D
TE
C EP
AC

Figure 10: Variation of complex viscosity in dependence of the number of extrusion cycles at
max. (291.6rad/s) frequency

Trends of storage modulus, G', and loss modulus, G'', (Figs. 11 and 12) as function of
extrusion cycles are similar as trend of complex viscosity prior described. In this case, it is
important to add information that for virgin material and recyclates from the first ten reprocessing
cycles, G'' is greater than G’ through all the frequency range. When taking into observation low
frequency, rise of G’ over G’’ is recorded between 70th and 80th extrusion cycle while at high
frequency this happens already after 10th extrusion.
 ACCEPTED MANUSCRIPT

It can be assumed that material elasticity increases due to branching process which results
from HDPE material degradation under the processing condition [20, 21, 22]. Intensive
branching process may be presented through increasing of G' between 5th and 20th cycle,
measured at high frequency. This effect is more evident at higher frequencies since the material
has no time for relaxation. Suggested branching process between 5th and 20th cycle is in line
with explanations to the recorded processing pressure and MFI previously presented.

PT
RI
U SC
AN
M
D

Figure 11: Variation of storage (G') and loss (G'') modulus in dependence of the number of
extrusion cycles at min. (0.628rad/s) frequency
TE
C EP
AC

Figure 12: Variation of storage (G') and loss (G'') modulus in dependence of the number of
extrusion cycles at max. (291.6rad/s) frequency
 ACCEPTED MANUSCRIPT

Consecutive processing increases shear thinning effect and this could in practice require
more stable processing conditions or narrower window of suitable processing conditions, when
using recycled material instead of virgin one, especially if it was recycled more than 10 times.
Furthermore, increase of viscosity, in particular after 10th extrusion cycle, may cause the need
for changing the processing parameters, while in worst case, more powerful machinery could be
needed for processing; as was already mentioned in previous sections.

PT
In case of G’ and G’’ we can say that increasing material elasticity causes the extrudate
swelling i.e. die swell [30, 31]. This should be taken into account when using material recycled

RI
more than 10 times, or in other words, some adjustments of equipment or processing conditions
may be required to prevent excessive die-swell.

SC
3.4 Thermal properties and crystallinity

U
Due to potential influence of recycling on material processability, also thermal
characterization has been implemented on virgin HDPE and its recyclates.
AN
Melting and crystallization temperatures (Tm and Tc) can be considered constant from
extrusion 0 to 30, while both temperatures continuously decrease after the 30th extrusion cycle.
M

However, the decrease is not significant since the minimum values of Tm and Tc, recorded at
100th extrusion cycle, represent a decrease smaller than 4%. Therefore Tm for all HDPE
D

materials investigated here can be reported as 133.7 ± 1.9 °C and Tc as 109.7 ± 1.1 °C.
TE

Since most of the heat necessary to melt the polymer during extrusion is generated through
shear by the screw rotation, that is 80-90% [31], it is possible to say that recorded changes of
Tm and Tc have no significant effect in practice. Similar findings, taking into account lower
EP

number of reprocessing cycles, were reported by several authors [6, 21, 24].

A curve representing the degree of crystallinity as a function of extrusion cycles is presented

C

in Fig. 13. It shows that extrusion leads to noticeable decrease in a degree of crystallinity only
AC

after 20th reprocessing cycle. An evident leap was detected between 20th and 30th extrusion
cycle while the largest drop of 18% can be observed at 90th extrusion.

Similarly, several authors who investigated the effect of HDPE reprocessing on degree of
crystallinity [6, 24, 26, 27], did not observe any significant change on crystallinity throughout 10
reprocessing cycles. This is in line with our findings from the first 20 extrusion cycles.
Decreasing of the degree of crystallinity shows the reducing of molecular ability to pack and
form crystal regions. This might be a consequence of chain branching or cross-linking in the
material, which are mechanisms present during degradation when materials are reprocessed
[20].
 ACCEPTED MANUSCRIPT

Furthermore, changes in the degree of crystallinity may result in changes of mechanical

properties, i.e. creep, modulus and hardness, presented in continuation.

PT
RI
U SC
AN
Figure 13: Degree of crystallinity in dependence of the number of extrusion cycles

3.5 Hardness and instant modulus

M

Mechanical properties of virgin HDPE as well as its recyclates, i.e. hardness and instant or
D

short-term modulus, were measured using nanoindentation. Results are presented in Figs. 14
and 15. It can be seen that both, hardness and instant modulus, exhibit similar trend. For both
TE

properties, one observes four different regions: (Region I) Through the first 10 extrusion cycles,
no significant changes are detected while further, (Region II) between 10th and 20th extrusion
cycle, “quick” drop is evident. Afterwards (Region III), both values are gradually decreasing up to
EP

70th extrusion cycle where (Region IV) both properties stabilise at about 80% of initial values.

When taking into account first 10 extrusion cycles, we can say that our results of mechanical
C

behaviour investigation are in line with those reported by other authors [21, 24], which showed
AC

no significant changes in mechanical properties; while none of the previous studies extend their
investigation above 10 reprocessing cycles, thus these results cannot be compared with data
from the available literature.

Changes of hardness and modulus can be attributed to the changes in the degree of
crystallinity as previously presented in Fig. 13.
 ACCEPTED MANUSCRIPT

PT
RI
SC
Figure 14: Hardness in dependence of the number of extrusion cycles

U
AN
M
D
TE
C EP

Figure 15: Modulus in dependence of the number of extrusion cycles

AC

Information on changes as described should be taken into account when designing products
from recycled HDPE. However, it can be said that mechanical properties of HDPE recycled up
to 10 times does not differs from those of virgin material and even after 100 recycling cycles
material retains 80% of its initial mechanical properties.

3.6 Shear creep compliance

Time-dependent (long term) properties of virgin HDPE and its recyclates were characterized
performing shear creep compliance measurements. To present long-term product mechanical
stability, first, master curves of virgin HDPE and its recyclates were generated and shifted to
 ACCEPTED MANUSCRIPT

reference temperature 30°C using time-temperature s uperposition principle, which was

executed with the closed form shifting (CFS) methodology [32]. Furthermore, isochronal values
of creep compliance at time equal to 3 and 10 years were taken from individual master curves;
results are presented in Fig. 16. It can be seen that extrusion did not significantly affect shear
creep compliance through first 30 reprocessing cycles. After 30 cycles an increasing trend is
evident after 3 years as well after 10 years. At 100th extrusion cycle, creep compliance rises for

PT
about 16% after 3years and 20% after 10 years in comparison to virgin material. In addition we
might say that difference in creep between 3 and 10 years remains more or less constant.

RI
U SC
AN
M
D
TE

Figure 16: Shear creep compliance as function of the number of extrusion cycles after 3 and 10
years.

Since shear creep compliance is measured in solid state we can relate increasing shear
EP

creep with extrusion cycles to the degree of crystallinity, presented in Fig. 13, which shows the
opposite, decreasing trend. We may conclude that products made of HDPE recycled more than
C

30 times, have lower long-term mechanical stability or in other words, using those materials
represents the risk that a product will have lower performance in long-term applications.
AC

3.7 Molecular weight and structure

Finally, to support suggested changes in the material structure, gel permeation

chromatography technique was used. Measurements were performed on virgin HDPE as well on
its recyclates. The results are presented in Figs. 17 to 19.

In Fig. 17 normalized molecular weight distribution curves are presented for the virgin
material and five extrudates; 20, 30, 50 70 and 100 times extruded material. It can be seen that
sequential reprocessing affects the molecular weight distribution in two ways; (i) molecular
 ACCEPTED MANUSCRIPT

weight distribution is broaden with number of extrusion cycles, this is evident from the beginning
of the reprocessing and (ii) peak of the molecular weight distribution curve shifts towards smaller
values, this becomes evident after 30th extrusion cycle. Recorded broadening of the molecular
weight distribution can be attributed to chain branching and cross-linking process while drifting
of the molecular weight distribution to smaller values suggests that there are processes which
reduce the molecular weight, like chain scission. From the results we may assume that chain

PT
branching process, as defined in continuation, took the place from the beginning of the
reprocessing, meanwhile chain-scission process is simultaneously happening after 30th
extrusion. Identical findings were suggested by the results on other characterization methods

RI
previously presented in this report.

SC
100
virgin HDPE
90 extrusion 20
80

U
extrusion 30
70 extrusion 50
AN
norm. w (%)

60 extrusion 70
50 extrusion 100
40
M

30
20
D

10
TE

0
100 1000 10000 100000 100000010000000
Mw (g/mol)
EP

Figure 17: Normalized molecular weight distribution for virgin HDPE and its extrudates.
C

Weight-average molecular weight, Mw, and number-average molecular weight, Mn, in

dependence of number of extrusion cycles are presented in Figs. 18 and 19. It is evident that
AC

Mw increases though first 30 extrusion cycles and continuously drops after. Mn has similar
trend, but the rise is not so obvious, and there is not a clear peak. Mw is large-molecule
sensitive that is why its rise in the first 30 extrusion cycles supports assumption on presence of
branching process. Contrary, decreasing of Mn after 30th cycle suggests an increase in the
amount of low molecular material which is indication of chain scission.
 ACCEPTED MANUSCRIPT

240000
220000
200000
180000
160000
140000
Mw (g/mol)

120000
100000

PT
80000
60000
40000

RI
20000
0

SC
0 10 20 30 40 50 60 70 80 90 100
Number of Extrusion Cycles

U
Figure 18: Weight -average molecular weight as function of the number of extrusion cycles.

40000
AN
35000
30000
M

25000
Mn (g/mol)

20000
D

15000
TE

10000
5000
EP

0
0 10 20 30 40 50 60 70 80 90 100
Number of Extrusion Cycles
C

Figure 19: Number-average molecular weight as function of the number of extrusion cycles.
AC

Portions of insoluble parts defined by extraction test are presented in the Table 1. It can be
seen that insoluble part was detected only after the 60th extrusion cycle, while only for 100-times
extruded material the amount of insoluble part becomes significant, i.e. 24.5 wt%. Recorded
insoluble portions confirm presence of cross-linking mechanism after 60th reprocessing cycle
which intensity is highly increased after 100th extrusion.
 ACCEPTED MANUSCRIPT

Table.1: Insoluble portion of the virgin HDPE and its recyclates

Material Insoluble portion (wt.%)

Virgin HDPE 0
20 times recycled 0
30 times recycled 0
50 times recycled 0
60 times recycled 1.0

PT
70 times recycled 1.3
100 times recycled 24.5

RI
Results on GPC measurements and solubility test, similarly as other results presented in this
paper, suggest molecular enlargement to be present during reprocessing, and after 30th

SC
extrusion cycle also chain-scission process. Results on solubility test support the idea that
molecular enlargement is occurring through chain branching, and that the cross-linking occurs

U
only after the 60th reprocessing cycle. In other words, chain branching is the prevailing
AN
mechanism during the initial cycles of the reprocessing, but in the final cycles cross-linking
becomes more important.
M

4. Conclusions
D

In the current study, effect of mechanical recycling on processability and mechanical

properties of HDPE was investigated. Results on rheological measurements, i.e. MFI, complex
TE

viscosity, storage and loss modulus, showed unfavourable effect of mechanical recycling on
HDPE processability. With increasing number of the reprocessing cycles MFI drops practically to
EP

zero after 30th extrusion cycle. Further, viscosity significantly increases between 5th and 20th
cycle and material elasticity as well. Moreover, changes in processability were directly monitored
through the melt pressure in the die channel and extrusion torque. Both values rise in first 30
C

extrusion cycles, pressure reaches more than 500% in respect to first extrusion. On the other
AC

hand, mechanical properties do not significantly change through the first 10 extrusion cycles.
Hardness and modulus evidently drop right after 10th cycle. Similarly, shear creep compliance
measurements show an unfavourable effect of mechanical recycling on the time-dependent
mechanical properties, while in this case, evident changes are visible in particularly after the 30th
extrusion cycle. However, all those changes are well supported by the results of investigation on
the degree of crystallinity which does not show significant changes through the first 20 extrusion
cycles while after, between 20th and 30th cycle, a significant drop and gradual decreasing
afterwards is recorded.

Researchers suggested two competing mechanisms of degradation during reprocessing of

HDPE, those are chain scission and side-chain branching [20, 21, 22]; in addition to this, some
 ACCEPTED MANUSCRIPT

suggested also presence of cross-linking [23, 24]. Results presented in this report, GPC
measurements and solubility tests, also suggest that chain-scission and chain branching occur
simultaneously from the start of the reprocessing. Due to chain scission caused by the shear
forces in the extruder, randomly ''chopped up'' molecular segments begin to merge to other
molecules and form branches. Through the first 30 extrusion cycles the molecular structure is
being dominated by chain branching, while after 30th extrusion cycle chain scission becomes

PT
predominant over the chain branching. After 60th extrusion cycle, the branched molecule chains
begin to cross-link. This effect becomes more pronounced with increasing number of extrusions,
but up to 100th extrusion cycle it is still being dominated by the chain scission, even though

RI
almost a quarter of the material has been cross-linked.

SC
It can be concluded that mechanical recycling worsens HDPE processability already through
the first 30 reprocessing cycles. We can expect the need for resetting the processing
parameters, increased burden on equipment and extrudate swelling when using recycled HDPE,

U
especially if the material was recycled more than 10 times. Moreover, mechanical recycling
causes also deterioration of HDPE mechanical properties that becomes evident after 10th
AN
reprocessing cycle. However, it needs to be emphasized that even after 100 recycling material
retains 80% of its initial mechanical properties. This information should be taken into account
M

when designing products from recycled HDPE.

D
TE
C EP
AC
 ACCEPTED MANUSCRIPT

Acknowledgements
This work was conducted with the financial support of the Slovenian Research Agency – ARRS.
We would like to acknowledge the support of Odelo Slovenija for enabling us performing MFI
measurements and BASF Company for their help with implementation of GPC measurements.

PT
References
[1] Plastics-the facts 2013, An analysis of European latest plastics production, demand and waste data,

RI
Plastics Europe 2013

SC
[2] L. Lundquist, Y. Leterrier, P. Sunderland, J. A. E. Månson, Life Cycle Engineering of Plastics:

Technology, Economy and Environment, Elsevier, 2000

U
[3] C. Jeggy, Micro-injection moulding: From process to modelling, UCL Presses Universitaires. de
AN
Louvain, 2004

[4] NIIR Board, The complete book on biodegradable plastics and polymers (Recent Developments,
M

Properties, Analyses, Materials & Processes), Asia Pacific Business Press Inc., Delhi, 2006

nd
[5] Goodship V, Introduction to Plastics Recycling (2 edition), Smithers Rapra Technology Limited, UK,
D

2007
TE

[6] Abad MJ, Effects of a mixture of stabilizers on the structure and mechanical properties of polyethylene

during reprocessing, Journal of Applied Polymer Science (2004). 92(6): 3910-3916

EP

[7] Kotiba H, Mosab K, Fawaz D, Effect of recycling on rheological and mechanical properties of

poly(lactic acid)/ polystyrene polymer blend, Jurnal of Materials Science (2011) 46:3013-3019
C

[8] Al-Salem SM, Lettieri P, Baeyens J, Recycling and recovery routs of plastic solid waste (PSW): A
AC

review , Waste Management 29 (2009) 2625-2643

[9] Dostal J, Kašparkova V, Yatloukal M, Muras J, Šimek L, Influence of the repeated extrusion on the

degradation of polyethylene. Structural changes in low density polyethylene, European Polymer

Jurnal, 44 (2008) 2652-2658

[10] da Costa HM, Ramos VD, Rocha MCG, Rheological properties of polypropylene during multiple

extrusion, Polymer Testing 24 (2005) 86-93

 ACCEPTED MANUSCRIPT

[11] Gonzalez-Gonzalez VA, Neira-Velazquez G, Angulo-Sanchez JL, Polypropylene chain scission and

molecular weight changes in multiple extrusion, Polymer Degradation and Stability 60 (1998) 33-42

[12] Jin H, Gonzalez-Gitierrez J, Oblak P, Zupančič B, Emri I, The effect of extensive mechanical recycling

on the properties of low density polyethylene, Polymer Degradation and Stabiliy, 97 (2012) 2262-

2272

PT
[13] Ras Lanuf Oil & Gas Processing Co, LANUFENE®HDI-6507UV technical specification,

http://www.nascopolymers.com/PDF/HDPE/HDI%206507%20UV.pdf (accessed: may 2014)

RI
[14] Kubyshkina G, Zupančič B, Štrukelj M, Marion L, Emri I, The influence of different sterilization

SC
techniques on the time-dependant behaviour of polyamides, Journal of Biomaterials and

Nanobiotechnology, 2011, 2, 361-368

U
rd
[15] Ferry JD, Viscoelastic properties of polymers (3 edition), John Willer&sons, Canada, 1980
AN
[16] Blaine RL, Thermal Application Note, Polymer Heats of Fusion, TA Instruments, NewCastle, USA,

www.tainstruments.com/library_download.aspx?file=TN048.pdf (accessed: Jun 2014)

M

[17] Li X, Bhushan B, A review of nanoindentation continious stiffness measurement technique and its

applications, Material Characterization 48, 2002

D

[18] Tschoegl NW, The phenomenological theory of linear viscoelastic behavior, an introduction,
TE

Germany: Springer, 1989

[19] Metlikovič P, Emri I, Naprava za merjenje lezenja torzijsko obremenjenih polimernih preizkušancev,
EP

Strojniški vestnik - Journal of Mechanical Engineering 1989; 35(4-6): 56-58

C

[20] Kealy T, Rheological Analysis of the Degradation of HDPE During Consecutive Processing Steps and

for Different Processing Conditions, Journal of Applied Polymer Science 112(2): (2009) 639-648
AC

[21] Pinheiro LA, The role of chain scission and chain branching in high density polyethylene during

thermo-mechanical degradation, Polymer Degradation and Stability 86(3): (2004) 445-453

[22] Loultcheva MK, Recycling of high density polyethylene containers, Polymer Degradation and Stability

(1997). 57(1): 77-81

[23] Kartalis CN, Mechanical recycling of postused high-density polyethylene crates using the

restabilization technique. I. Influence of reprocessing, Journal of Applied Polymer Science 73(9):

(1999) 1775-1785
 ACCEPTED MANUSCRIPT

[24] Apone S, Effects of thermomechanical treatments on HDPE used for TLC ducts, Polymer Testing

22(3): (2003) 275-280

th
[25] Rauwendaal C, Polymer Extrusion (4 edition), Hanser, Munich, 2001

[26] Luzuriaga S, Degradation of pre-aged polymers exposed to simulated recycling: Properties and

thermal stability, Polymer Degradation and Stability 91(6): (2006) 1226-1232

PT
[27] Altan M, Residual stresses determination in injection molded virgin and recycled HDPE blends:

RI
mechanical properties and morphology, E-Polymers, 2008

[28] Epacher E, Two-step degradation of high-density polyethylene during multiple extrusion, Journal of

SC
Applied Polymer Science 74(6): (1999) 1596-1605

[29] Bozorg-Haddad A, Compressive strength and creep of recycled HDPE used to manufacture

U
polymeric piling, Construction and Building Materials 26(1): (2012) 505-515
AN
th
[30] Van Krevelen D, Nijenhuis K , Properties of Polymers (4 edition), Elsevier, 2009

[31] Giles HF, Extrusion, The Definitive Processing Guide and Handbook, William Andrew, Inc., NY, 2005
M

[32] Gergesova M, Zupančič B, Saprunov I, Emri I, The closed form t-T-P shifting (CFS) algorithm, Journal
D

of rheology, vol.55, no.1, p[1-16] 2011

TE
C EP
AC