International Journal of Pavement Engineering

ISSN: (Print) (Online) Journal homepage: https://www.tandfonline.com/loi/gpav20

Recycled plastics as synthetic coarse and fine

asphalt aggregate

Dai Xuan Lu & Filippo Giustozzi

To cite this article: Dai Xuan Lu & Filippo Giustozzi (2022): Recycled plastics as synthetic
coarse and fine asphalt aggregate, International Journal of Pavement Engineering, DOI:
10.1080/10298436.2022.2068550

To link to this article: https://doi.org/10.1080/10298436.2022.2068550

Published online: 02 May 2022.

Submit your article to this journal

Article views: 247

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=gpav20
INTERNATIONAL JOURNAL OF PAVEMENT ENGINEERING
https://doi.org/10.1080/10298436.2022.2068550

Recycled plastics as synthetic coarse and fine asphalt aggregate

Dai Xuan Lu and Filippo Giustozzi
Civil and Infrastructure Engineering, RMIT University, Melbourne, Australia

ABSTRACT ARTICLE HISTORY

This study investigates the performance of asphalt concrete containing recycled plastics as synthetic Received 29 January 2022
aggregates in the dry mixing process. Two types of recycled plastics were used, namely, acrylonitrile Accepted 15 April 2022
butadiene styrene (RABS) and Polyethylene Terephthalate (RPET), based on the high-temperature
KEYWORDS
properties of the plastics that make them suitable to be treated as synthetic aggregates (i.e. do not Recycled plastic; synthetic
melt during the asphalt mixing phase). The proportions of recycled plastics were chosen at 0.5%, 1%, aggregate; ABS; PET; dry
2% and 4% of the mix mass. The laboratory testing programme includes the evaluation of process; asphalt road
compactability properties, moisture resistance, cracking resistance via IDEAL-CT test, fatigue resistance pavements
via four-point bending beam test and rutting resistance via Hamburg wheel tracking (wet) test. It was
found that the inclusion of recycled plastics did affect the compactability of the mixes depending on
their relative size and shape. On the other hand, the inclusion of recycled plastics did not significantly
vary the moisture resistance of asphalt mixes. The cracking and rutting behaviour of asphalt mixes
that incorporated recycled plastics were also modified and some benefits were observed depending
on the plastic content and size.

1. Introduction
Recycling is being increasingly seen as one of the main aims to and films, among various others. This type of waste generally
achieve sustainable development by the asphalt pavement yields a lower rate of recyclability due to contamination of
industry. Besides using reclaimed asphalt pavement (RAP) the waste with paper, glass, soil, chemicals and organics. On
material, the industry also aims to incorporate recycled wastes the other hand, post-industrial plastics consist of plastics col-
into the asphalt mixes to help reduce the amount of waste lected from companies manufacturing plastics items and it is
landfilled and mitigate the detrimental effect of various waste mostly made of single-polymer plastic. Usually, post-industrial
materials on the environment (Mallick et al. 2004, Ahmadinia plastic is cleaner than post-consumer plastic due to various
et al. 2011). Among recycled wastes, plastics are gaining methodologies for contamination control and quality proto-
momentum as experimental candidates to be recycled in cols, which also make it less variable over time. Thus, the
asphalt mixes in Australia and New Zealand, and worldwide. recycling rate of post-industrial plastics is usually higher
Plastic contributes to a significant portion of the waste gener- than post-consumer plastics.
ated in Australia and New Zealand (Giustozzi and Boom There are two main approaches to incorporate plastic into
2021). For example, in 2018, Australia consumed 3,407,300 asphalt mixes: wet and dry processes (Ranieri et al. 2017).
tonnes of plastics. Among those, only 320,000 tonnes (9.4%) The wet process melts and mixes plastic with bitumen before
were recycled and the remaining (90.6%) were landfilled or adding it into the aggregates; in this case, the plastic is used
exported (O’Farrell 2019). The significant volume of material as a polymer for bitumen modification (Dalhat and Al-
landfilled increases the burden on the environment by leaving Abdul Wahhab 2017, Diab et al. 2019, Padhan et al. 2020).
material on the ground that takes hundreds of years to decom- On the other hand, the dry process incorporates plastic directly
pose. It is clear that landfill is one of the temporary solutions into the hot aggregate and mixes them for a certain period of
but not one that is sustainable. time before adding bitumen (Lastra-González et al. 2016, Mar-
One of the possibilities to recycle waste plastics is to incor- tin-Alfonso et al. 2019). Compared to the wet process, the dry
porate them into bituminous materials (Costa et al. 2013, process has the potential to incorporate a greater portion of
Nizamuddin et al. 2020, Willis et al. 2020, Nizamuddin et al. recycled plastic and the asphalt production process is easier
2021). The challenge of adding waste plastic to asphalt pave- and faster. Since the properties of different waste plastics are
ments is the variability of the material, which can come from not similar, each type of production process is suitable for cer-
multiple sources with different compositions and properties tain plastics only but not for others.
(Ranieri et al. 2017, Joohari and Giustozzi 2020). Waste plas- It is important to note that in the dry process, two possible
tics can be categorised into two main streams: post-consumer alternative methods can be used. One way is to add plastic
and post-industrial. Post-consumer stream plastics come from into the hot aggregates and use the plastic as both aggregate
consumers’ kerbside collection and commonly include water coating material, i.e. when low melting point plastics are
bottles, milk containers, household items, commingled plastics added. Then, the bitumen is added into the mix to blend with

CONTACT Filippo Giustozzi fi[email protected]

© 2022 Informa UK Limited, trading as Taylor & Francis Group
2 D. X. LU AND F. GIUSTOZZI

melted plastic on the aggregate surface and cover the aggregate. tests. However, the study did not provide any final conclusions
In this process, the plastic is expected to melt and blend with on the rutting behaviour of PET mix as well as any recommen-
bitumen to produce a similar – but less homogeneous – effect dation for the percentage of PET which should be used.
to what postulated in the wet process (Ranieri et al. 2017, Mar- Dalhat and others (Dalhat et al. 2019) studied the perform-
tin-Alfonso et al. 2019, Lastra-González et al. 2021) without the ance of recycled plastic waste (RPW) in dense-graded asphalt
need for the asphalt plant to handle polymer-modified bitumen. concrete. The study used an RPW in a combined form of var-
Sometimes, this process is also considered as a mixed process ious types of plastic wastes, including approximately 17% of
(i.e. a combination of dry and wet). Throughout this paper, low-density polyethylene, 25% high-density polyethylene,
the term ‘mixed’ process will be used for this specific type of 34% PET, 11% polypropylene, 4% polyvinyl chloride and 9%
dry process. polystyrene, to replace mineral aggregate. The performance
The second way to incorporate plastics in the dry process of RPW mix (RPW to replace aggregate only) and hybrid
uses recycled plastics as a substitution for natural aggregates RPW mix (RPW to replace aggregate and using RPW-
in the asphalt mix, basically as a synthetic aggregate (Mikhai- modified binder) was evaluated using the dynamic modulus
lenko et al. 2021). In this case, the plastic is not expected to test, flow number test and rutting test. Asphalt mixes produced
melt and mix with the binder. Similar to the mixed method, with PET only (used as a natural aggregate substitution at the
this approach indicates that the recycled plastic is mixed proportion of 5% and 10% per total aggregate mass), crumb
directly with natural aggregate before adding bitumen to the rubber (CR) and a conventional mix were used for comparison
mix. This procedure will be called ‘dry’ process in the present with the RPW mixes. The study found that the hybrid mix had
study. Due to the nature of the process, the selected plastic for superior performance compared to the reference mixes.
the dry process needs to have certain specific properties (i.e. Nevertheless, this study only compared the viscoelastic behav-
high melting point or amorphous behaviour) to work as a iour of the PET asphalt mix with the RPW asphalt mix via the
real aggregate, avoiding melting and blending with bitumen, dynamic modulus test and no other performance tests such as
and to cope well with asphalt mixing, conditioning and moisture resistance, cracking and rutting were conducted on
compaction. recycled PET asphalt mix. Thus, the study could not provide
So far, there have been a limited number of studies that any optimum amount of PET to be possibly used as aggregate
evaluated the recycling of plastics in asphalt concrete via the replacement in dense-graded asphalt mixes.
dry process as intended in this article. In particular, the appli- Mikhailenko et al. (2021) investigated the potential of PET
cation of recycled ABS and PET plastics as aggregate replace- and CR as aggregate replacement. Recycled PET and CR were
ment is yet to be fully explored. Ahmadinia and co-workers incorporated in semi-dense HMA, which is also known as gap-
(Ahmadinia et al. 2012) studied the potential to incorporate graded asphalt. PET was incorporated as a partial replacement
PET in stone mastic asphalt (SMA). The study investigated of natural sand at 5.1% by mass of the mix. The study investi-
the behaviour of recycled plastic SMA via the modulus test, gated the compactability, fracture resistance via the indirect
moisture test, drain off test and rutting test. The percentage tensile and semi-circular bending tests, moisture sensitivity,
of PET added to the mix ranged from 0% to 10% of bitumen rutting resistance via axial cyclic compression test, sound
content. The study concluded that the addition of PET could absorption test and texture scanning analysis. The study con-
increase the stiffness of the mix, improve the rutting resistance, cluded that the PET mix required more energy to compact
reduce the binder drain down, yet had a lower tensile strength while having similar indirect tensile strength, fracture energy,
ratio (TSR) in the moisture test. The study recommended to indirect TSR and surface texture, compared to the control mix.
use 4% to 6% of PET by weight of the binder in SMA. In The study also found that by replacing 5.1% of PET by volume
this study, PET was expected to work as an additive for the of the mix (equivalent to 2.8% by mass), the rutting of the PET
binder rather than as a replacement for natural aggregate. mix was lower than the control mix. However, this result could
However, it should be noted that the melting temperature of be contributed by the difference in the air void of PET and
PET is approximately 260°C, well above the asphalt mixing control samples (PET samples had 2% greater air voids content
temperature. This study also did not investigate the fatigue than the control samples) because of the difficulty in compact-
cracking performance of the mixes. Yoo and Al-Qadi (2014) ing the PET samples. In addition, due to the fact that only one
studied the pre- and post-peak toughening behaviour of coarse percentage of PET was used, the study could not provide a
graded hot mix asphalt by adding PET fibre. Instead of chip- robust conclusion on how much PET should be used to replace
ping PET into small particles, this study scratched plastic bot- sand in the gap-graded HMA.
tles into the form of fibres. The study found that 0.4% of PET In terms of recycling ABS in asphalt mixes utilising the dry
by weight of the asphalt mix enhanced the cracking properties process, there are very few literature studies looking into this
of the mixes. However, due to the limitation in performance material in asphalt concrete notwithstanding the large presence
tests carried out such as rutting, fatigue and moisture resist- of waste ABS around the world. Several studies investigated the
ance, the study recommended further exploration to be done recycling aspect of ABS by melting it in the binder in the wet
in future studies. Moghaddam et al. (2014) investigated the process and treated the recycled ABS as binder replacement
rutting behaviour of asphalt mixes adding PET by aggregate (Wu and Montalvo 2020). However, since ABS is an amorphous
mass from 0% to 1.0% in SMA. The study investigated the rut- polymer (i.e. it does not have a thermal transition as this is com-
ting behaviour via static and dynamic creep tests. The study monly associated with the crystalline or ordered polymeric
concluded that the addition of PET showed different beha- phase), the melting and blending of ABS with the binder were
viours of rutting resistance in the static and dynamic creep reported to be very difficult (Costa et al. 2013). Instead, ABS
 INTERNATIONAL JOURNAL OF PAVEMENT ENGINEERING 3

plastic should be incorporated into asphalt concrete in the dry of RPET and the amorphous behaviour of RABS allow both
process as aggregate replacement. plastics not to melt during the mixing process. Due to the
From the review of the current literature, it is clear that the size of the recycled plastics, RABS was used to partially replace
studies into recycled PET and ABS as a replacement for natural the coarse aggregate fraction (e.g. 7 mm aggregate) while RPET
aggregates are very limited. Most of the studies – mainly on was used to partially replace the smaller aggregate size and, in
PET – focused on the partial performance of asphalt mixes particular, virgin sand in this study.
with recycled plastics and investigated the limited contents The virgin quarry aggregate was provided by a local asphalt
of recycled plastic in the mix. None of the studies has fully contractor in Victoria, Australia in various fractions. The
investigated the asphalt mix performance (compactability, aggregate fractions were then combined to manufacture an
moisture resistance, viscoelastic behaviour, cracking and rut- asphalt mix with a nominal maximum aggregate size
ting resistance) at different PET and ABS loadings in conven- (NMAS) of 13.2 mm, hereby named AC14. Besides natural
tional dense-graded HMA as aggregate replacement. In aggregate, hydrated lime was also used as filler at 1% by the
addition, to the best of the authors’ knowledge, no previous total aggregate weight as required by many states in Australia.
study has compared the performance of dense-graded asphalt One type of bitumen, C320 bitumen (based on the viscosity
mixes produced with recycled plastics of different sizes (i.e. grade in AS2008 (AS 2013)), was used for all mixes. This bin-
coarse vs. fine aggregate replacement) in the dry process. der grade generally corresponds to a binder with a penetration
This study investigated the possibility of recycling plastics grade of 50–70. C320 can be used for different traffic levels and
in the dry process, in which the recycled plastics were used as pavement layers in Australia depending on the nominal
a substitution for natural aggregates in the asphalt mix, basi- NMAS of the asphalt mix. The basic physical properties of
cally as a synthetic aggregate. The study incorporated two C320 bitumen are presented in Table 2.
types of recycled plastic – PET and ABS – at different con-
tents of 0.5%, 1%, 2%, and 4% by weight of the mix into
dense-graded HMA. The study investigated the compactabil- 2.2. Mix design
ity, moisture resistance, propension for rutting, cracking tol-
erance and fatigue resistance of the plastic-modified asphalt The asphalt mixes used in this study were designed using the
mixes. Based on the findings, the study concluded on the aggregate gradation curves as per the guidelines of the Victor-
optimal proportions of recycled plastic to be used for each ian Department of Transport, Australia, VicRoads Section 407
type and size. (VicRoads 2021). In addition, the gradation curves were also
assessed to avoid the restricted zone for sand particles required
in the Superpave method (AASHTO 2001). The sieve sizes of
2. Materials and mix design the two types of recycled plastics are shown in Table 3 and
those of the asphalt mixes are shown in Figure 2.
2.1. Materials
RABS particles are coarser than RPET and are mainly
Two types of recycled plastics (Figure 1) were selected for this retained on the 4.75 and 2.36 mm sieves, while a majority of
study based on the experimental outcomes of their physical RPET plastic particles are retained on the 2.36 and 1.18 mm
properties described in Table 1 and past studies by the authors sieves (Table 3). Due to the different sizes, RPET was used to
(Audy et al. 2022). The first recycled plastic is a common ther- partially replace the sand portion of the asphalt mix while
moplastic polymer, recycled acrylonitrile butadiene styrene RABS was partially substituted for the coarser aggregate. The
(RABS) plastic, which comes in granules form from post- two recycled plastics have been incorporated separately in
industrial recycling operations. The second recycled plastic is the asphalt mix, i.e. no mix in this study included both
known as recycled polyethylene terephthalate (RPET). RPET RABS and RPET. This was done to study the effect of the
is one of the most common thermoplastic polymers from size and shape of the plastic particles in addition to the
post-consumer recycling operations. The high melting point content.

Figure 1. Physical appearance of recycled PET (a) and recycled ABS (b) plastics.
4 D. X. LU AND F. GIUSTOZZI

Table 1. Physical properties of RABS and RPET plastics. Table 3. Sieve size analysis of RABS and RPET.
Physical properties RABS RPET Sieve size (mm) RABS plastic, passing (%) RPET plastic, passing (%)
Composition ABS 100% PET 100% 19 100.0 100.0
Melting Point (°C) None, amorphous 252 13.2 100.0 100.0
Density (g/mm3) 1.03 1.35 9.5 100.0 100.0
Water contact angle (°) 93.4 95.3 6.7 100.0 100.0
Melt flow index (g/10 min) 3.81 83.51 4.75 71.8 100.0
2.36 8.5 100.0
1.18 1.9 26.5
0.6 0.5 1.4
Table 2. Basic physical properties of the binder. 0.3 0.2 0.0
Physical property Testing method Results 0.15 0.1 0.0
0.075 0.0 0.0
Penetration @ 25°C (0.1 mm) AS 2341.12 45
Viscosity @60°C (Pa.s) AS/NZS 2341.2 335
Dynamic Viscosity @135°C (Pa s) AS/NZS 2341.4 0.50
Softening point (°C) AGPT T-131 51.4
Density @ 15°C (g/cm3) AS 2341.7 1.04

Figure 2. Sieve size distribution of asphalt mixes: (a) HMA and RABS mixes, (b) HMA and RPET mixes.
 INTERNATIONAL JOURNAL OF PAVEMENT ENGINEERING 5

Table 4. Volumetric properties of asphalt mixes.

Mixes HMA 0.5% RABS 1% RABS 2% RABS 4% RABS 0.5% RPET 1% RPET 2% RPET 4% RPET Limit
Binder film index (FI) 9.3 9.2 9.1 9.0 8.8 9.2 9.2 9.2 9.1 8.0a
VMA (%) 14.6 14.6 14.5 14.4 14.1 14.6 14.5 14.4 14.2 14
VFB (%) 72.7 72.6 72.4 72.2 71.6 72.6 72.5 72.3 71.8 65–75
Dust/binder 0.96 0.96 0.96 0.96 0.96 0.96 0.96 0.96 0.95 0.6–1.2
a
This limit is based on the guidelines of Victorian Department of Transport, Australia, VicRoads Section 407 (VicRoads 2021).

RABS and RPET were incorporated at 0.5%, 1%, 2% and was determined as 5.1% by the weight of the mix.
4% by weight of the mix. The virgin aggregate fraction to The volumetric properties of the asphalt mixes are shown in
be replaced was selected in order for the final gradation of Table 4. Figure 3 shows typical cut surfaces of compacted
the plastic-modified mixes to be as similar as possible to specimens of RABS and RPET asphalt mixes. It can be seen
the control HMA made with a virgin aggregate only, i.e. 0% that the RABS and RPET form an integral part of the aggregate
recycled plastics. This approach aimed to minimise the per- structure of the asphalt samples.
formance variability associated with the aggregate gradation
in asphalt mixes.
In this study, the mixing temperature of 155°C was used in
accordance with the joint Australian and New Zealand stan- 3. Mechanical behaviour
dard AS/NZS 28911.2.1 (AS/NZS 2014a) for AC14 mixes that
3.1. Compactability
use C320 bitumen. During the mixing procedure, the plastic
particles were added at ambient temperature directly into the To evaluate the densification properties of the asphalt mixes
preheated aggregate and dry mixed for approximately 20 s. during compaction, the compactability test was carried out.
After that, the preheated binder was added to the aggregate The test was conducted on 1200 g of loose asphalt mix in a
and mixed for 5–10 min at the mixing temperature. For com- 100 mm diameter mould with the number of gyrations set
paction, the loose mix was reheated at the compaction temp- to 350. The compaction curves were built to present the
erature of 150°C for 2.5 h (AS/NZS 2014b) on the day of relationship between air voids in the sample and the number
compaction. This procedure allowed the loose mix to reach of gyrations. Three replicates were tested for each mix and the
the compaction temperature while avoiding the several heating average value of the three replicates was used to build the
cycles of the same material and excessive aging of the loose mix. compaction curves. The compactability of the mixes was eval-
A gyratory compactor was used with the following compac- uated based on the value of air voids after one gyration
tion settings (AASHTO 2001): 600 kPa of loading pressure, q(1) and the compactability K defined in Equation (1)
compaction angle of 1.25 degrees and gyration speed of 30 based on EN 12697-10.
rounds per minute. The optimum binder content was first
q(ng) = q(1) − (K × ln(ng)) (1)
designed for the control HMA (i.e. no plastic added). Then,
the same value was used for the other asphalt mixes because where q(ng) is the percentage of air voids for a number of
the main aim of this study was to investigate the effect of add- gyrations ng (%), q(1) is the percentage of air voids calculated
ing recycled plastic to asphalt. The optimum binder content after one gyration (%); K is the compactability; ng is the num-
was chosen at 4% air void after 100 gyrations. The compaction ber of gyrations.
conditions were selected to design mixes that can be used for In Equation (1), q(1) is determined as the air voids of the
sustaining traffic loading of up to 30 million equivalent stan- sample after one gyration and K is determined as the slope
dard axle loading (ESAL) over 20 years (Austroads, 2008; of the linear regression analysis by fitting the compaction
Rebbechi and Petho, 2014). The optimum binder content data points from 20 to 200 gyrations.

Figure 3. Typical cut surface of cylinders of recycled plastic mix (RABS) (a) and typical cut surface of beam of recycled plastic mix (RABS) (b).
6 D. X. LU AND F. GIUSTOZZI

3.2. Moisture susceptibility li is the load–displacement line (mm) at the i step, li+1 is the
load–displacement line (mm) at the i+1 step.
The moisture susceptibility was carried out to evaluate the
effect of adding plastic on the moisture resistance of asphalt
mixes. The testing procedure followed the stripping test stan-
dard in Australia and New Zealand AG:PT/T232 (AG:PT/ 3.4. Flexural stiffness and fatigue resistance
T232 2007). Cylindrical samples with a diameter of 100 mm
The fatigue and flexural stiffness tests were conducted using
and a height of 65 mm were produced with air voids in the
the four-point bending beam loading apparatus. The tests
range of 8.0+1.0%. Six samples were prepared for each
were performed according to Austroads test method AGPT/
mix, in which three samples were dry tested and three others
T274 (Austroads 2016). Note that the standard AGPT/T274
were wet tested using the indirect tensile test (IDT). The
regulates both flexural stiffness and fatigue test in the same
moisture resistance was evaluated based on the TSR between
standard. Tests were carried out on beams with air voids of
the wet tested samples and the dry tested samples of each
5 ± 0.5% and dimensions of 390 ± 5 mm in length, 50 ± 5
mix. For the dry test, the samples were conditioned in a
mm in depth and 63.5 ± 5 mm in width.
temperature-controlled chamber at 25°C for 2 h before test-
For the stiffness test, samples were tested at four different
ing. For the wet tested samples, they were vacuum saturated
temperatures of 0, 10, 20 and 30°C. At each temperature, the
at 50 °C to achieve 55% to 80% degree of saturation first.
samples were tested at different frequencies; 0.1, 0.5, 1, 2, 5
Then, they were conditioned in a water bath at 60°C for 24
and 10 Hz. The sinusoidal load waveform was used for this
h. Finally, the samples were conditioned in water at 25°C
study. All the samples were tested at a strain level of 50
for 2 h before testing. The load was applied at 50 mm/min
micro-strain for the stiffness test. The results of the flexural
during the IDT test.
stiffness test were used to develop a flexural modulus master
curve following Equations (5–7).
a
3.3. Fracture properties log10 |E∗ | = d + (5)
1+ eb+glog10 fr
The fracture test was performed using the IDT testing setup to
evaluate the CT index according to ASTM D8225−19 (ASTM fr = a T × f (6)
2019). The CT index was used to evaluate and compare the In Equation (5), E∗ is the flexural modulus; d, a, b, g are
cracking resistance of asphalt mixes utilising the fracture pro- fitting parameters; and fr is reduced frequency. The reduced
cess. The samples with a diameter of 150 mm and a thickness frequency is a product of time–temperature shift factor aT
of 62 mm were produced with air voids of 7 ± 0.5%. The and the actual loading frequency (Equation (6)). The determi-
samples were conditioned in a temperature control chamber nation of time–temperature shift factor is fitted using a
for 2 h at 25°C before testing by applying a monotonic loading second-order polynomial relationship as per the following
of 50 mm/min. At least five replicates were tested for each equation:
asphalt mix and the final CT index was calculated as the aver-
age of the tested samples for each mix. log10 (aT ) = a(T − Tref )2 + b(T − Tref ) (7)
The CT index was calculated as per the below equation:
where a, b are the fitting parameters, Tref is the reference
t l75 Gf temperature.
CTindex = × × (2)
62 D |m75 | The fatigue test was conducted at 20°C using the strain-con-
trolled mode at three different micro-strains of 400, 250 and a
where t is the thickness of the sample (mm), l75 is the displace-
variable lower strain value to achieve at least 1 million cycles.
ment at 75% of the peak load after the peak (mm), D is the
At least six beams for each strain level were tested. A loading
sample diameter (mm), Gf is the fracture energy or failure
frequency of 10 Hz and sinusoidal loading were applied. The
energy (Joules/m2), |m75 | is the absolute value of the post-
results of the test were used to build fatigue curves of asphalt
peak slope (kN/mm).
mixes at different strain levels. The fatigue resistance of the
The fracture energy or failure energy is determined as below:
mix was determined as the cycle at which 50% of reduction
Wf in the sample’s stiffness was achieved. The strain level at
Gf = × 106 (3)
D×t which the beam gave fatigue life of exactly 1 million cycles
was determined via Equation (8) as a linear relationship in
where Wf is the fracture work, which is determined as the area
the log-log scale between Nf and the strain level 1 according
under the load–displacement curve up to the value of post-peak
to ASTM D8237-21 (ASTM 2021):
load of 0.1 kN.
n−1 

 f (Nf ) = A0 1A6 1 (8)
1
Wf = (li+1 − li ) × Pi + × (li+1 − li ) × (Pi+1 − Pi )
2 In Equation (8), f (Nf ) is a fitting function of fatigue curve
i=1
based on the fatigue results determined at three different strain
(4)
levels; A0 and A1 are model regression parameters; Nf is the
where Pi is the applied load (kN) at the i load step appli- number of cycles that cause fatigue cracking (50% reduction
cation, Pi+1 is the applied load (kN) at the i+1 load step of initial flexural stiffness) and 16 is the strain at which the
application, cycles reaching the fatigue criterion are equal to 1 million.
 INTERNATIONAL JOURNAL OF PAVEMENT ENGINEERING 7

3.5. Rutting resistance result and illustration of fitting the rutting curves with the
Franken model are presented in Figure 4.
The rutting test was conducted using the Hamburg wheel
tracking apparatus. The samples preparation and testing pro- ep (N) = AN B + C(eDN − 1) (9)
cedure followed AASHTO T 324–17 (AASHTO 2018).
Samples with a diameter of 150 mm and a height of 62 mm d ep
= ABN (B−1) + CDeDN (10)
were prepared. The air voids target was 7 ± 0.5%. The test dN
was conducted using the wet method, in which the samples
∂ 2 ep
are submerged in water during the test. The tests were con- = AB(B − 1)N (B−2) + CD2 eDN (11)
ducted at 50°C and the samples were conditioned for 45 min ∂N 2
in the water at 50°C prior to the test. For each mix, at least where ep (N) is the permanent deformation or rut depth; N is
four samples were tested. During the test, the rut depth in the number of loading passes; A, B, C, D are the fitting
accordance with the loading cycle was recorded. The load parameters.
was simulated by metal rolling wheels at the speed of 52 passes After identifying the stripping point, the second step applies
(26 cycles) per minute. The test terminated either at the pass the Tseng–Lytton model (Equation (12)) to fit the rutting
number of 20,000 or when the maximum rut depth of curve from the onset of loading to the stripping number.
25 mm was reached. Once the model parameters are calibrated, the rutting curve
It should be noted that the Hamburg wheel tracking test due to the mechanical loading only can be drawn (Figure 4).
was conducted in the wet condition, the total rut depth is    
a l
therefore a combination of the rutting generated by the mech- ep = e1 p exp − (12)
anical load and the moisture (water) effect (Yin 2014). Due to LC
the effect of moisture, the rut depth is accelerated as stripping where e1
p is the saturated rut depth; a, l are the model
happens due to the bond loss of aggregate particles. Yin et al. coefficients.
(2014) proposed a method including two main steps to separ-
ate the rut depth caused by the mechanical load from the effect
of the moisture. The first step involves fitting the rut depth- 4. Results and discussion
number of passes curve by using the Franken model (Biligiri,
4.1. Compactability
Kaloush, Mamlouk, & Witczak 2001) (Equation (9)) to deter-
mine the stripping point. To associate the calculation of the The densification properties of the asphalt mixes are shown in
stripping point, the first and second derivatives of the Franken Figure 5 and Table 5. It can be seen that the addition of RABS –
model were analysed as per Equations (10 and 11). The strip- used as coarse aggregate replacement in the mix – tends to help
ping point is defined at the load pass at which the sign of the compaction. In particular, 0.5% of RABS shows statistically
second derivative changes from negative to positive. A typical similar values of q(1) and K to the reference HMA mix,

Figure 4. Illustration of typical results from the Hamburg wheel tracking test fitting models to separate the rut depth due to the load from the moisture effect.
8 D. X. LU AND F. GIUSTOZZI

Figure 5. Compaction curves of asphalt mixes.

indicating that the addition of 0.5% RABS did not significantly reduction in the compactability of the mixes when RPET con-
affect the compactability. Whereas, adding 1% or higher tent is increased. For example, 1% and 2% RPET mixes had
RABS proportions produced an apparent reduction in the 5.2% and 6.1% air voids, respectively, after 100 gyrations.
compaction energy, indicating a noticeable reduction in Although higher than the standard asphalt mix with no plastic,
q(1) and K values. However, adding more RABS (e.g. greater the air voids content indicates that adding up to 2% RPET can
than 2%) affected the final stage of the compaction curves, in still be feasible with some adjustment to the field compaction
which the compactability tends to decrease. This phenom- procedure. Yet, incorporating 4% RPET significantly decreases
enon is observed when comparing 2% RABS with 4% RABS the compactability of the asphalt mix, with air voids around
after 100 gyrations, for example. In this comparison, it is 8% after 100 gyrations.
thought that the presence of recycled plastic introduced less The difference in the compatibility of plastic mixes is due to
resistance to the compaction force when the mix is in its their shape and size in addition to their physical properties (see
initial densification state, resulting in better compactability Figure 1). RABS had a similar shape of virgin coarse aggregate
for mixes with higher RABS content. However, at denser with substantially spherical shape and fragmented edges, while
states, the contribution of plastic particles to the resistance RPET presents as flaky and elongated particles. For that, RABS
of compaction load was more obvious. The plastic particles particles could re-adjust and rotate within the mix during the
created a bouncing effect (elastic recovery to resume the orig- compaction process more easily.
inal shape of plastic particles after deformation) after remov-
ing the load or changing the load direction due to the
4.2. Moisture susceptibility
gyratory angle. The contribution of plastic particles to the
densification of the mix becomes significant at this stage, The outcomes of the moisture susceptibility analysis are shown
hence resulting in more difficult compaction. The bouncing in Table 6. The addition of RABS tends to slightly reduce the
effect in high RABS mixes is likely due to the fact that tensile strength in the dry samples compared to the control
RABS is a heterogeneous multi-phase polymer that partially HMA although no definitive trend was identified. Similar ten-
contains polybutadiene, effectively synthetic rubber (Hasan sile strength at failure was observed in the wet samples of
et al. 2016), with very low glass transition temperature (i.e. RABS mixes (up to 2%) compared to HMA in the t-test stat-
– 90 °C). istic with the significance level of 0.05. The exception falls
In the opposite manner, the addition of RPET plastic to into the case of 4% RABS mix in which the tensile strength
partially replace virgin sand tends to hinder the compactability of the mix was significantly lower than that of HMA in the
of the mixes. In particular, the addition of 0.5% RPET did not wet condition. On the other hand, the addition of RPET
significantly affect the compactability of the mix compared to showed a clearer reduction of tensile strength values in both
the control mix. However, there is a clear trend of overall dry and wet conditions. The dry tensile strength of RPET

Table 5. Compactability parameters q(1) and K for mixes.

Mix HMA 0.5% RABS 1% RABS 2% RABS 4% RABS 0.5% RPET 1% RPET 2% RPET 4% RPET
q(1) 20.40 21.55 19.51 18.99 18.87 20.20 21.19 21.73 22.50
St. dev. 0.144 1.066 0.184 0.274 0.491 1.814 0.379 0.202 0.415
K −2.836 −3.025 −2.531 −2.551 −2.092 −2.641 −2.735 −2.722 −2.544
St. dev. 0.032 0.261 0.038 0.037 0.256 0.382 0.053 0.100 0.096
 INTERNATIONAL JOURNAL OF PAVEMENT ENGINEERING 9

Table 6. Moisture resistance results of asphalt mixes in the TSR tests.

Mix HMA 0.5% RABS 1% RABS 2% RABS 4% RABS 0.5% RPET 1% RPET 2% RPET 4% RPET
ITS (kPa) Dry 1199.6 1075.1 1091.8 1076.5 1117.7 1131.0 1100.9 1092.1 1006.5
St. dev. (kPa) Dry 53.0 82.4 51.6 134.9 82.8 26.7 41.8 23.4 13.8
ITS (kPa) Wet 1043.1 1014.4 1033.4 1008.8 964.2 1033.1 941.2 922.5 904.6
St. dev. (kPa) Wet 27.3 38.7 50.9 51.9 2.3 18.4 58.9 43.1 21.5
TSR (%) 87.0 94.3 94.7 93.7 86.3 91.3 85.5 84.5 89.9

mixes started to decrease significantly when adding 2% and 4.3. Fracture properties
more RPET. Similarly, the wet tensile strength began reducing
The results of the IDEAL-CT test are shown in Figure 6
significantly from 2% RPET while 0.5% and 1% RPET exhib-
using box and whisker plots. The cross marker ‘x’ presents
ited no statistical difference in the tensile strength compared
for the mean value, the ‘box’ shows the range of the lower
to that of the control HMA.
quartile and upper quartile data, and the extreme range of
Overall, it can be seen that the moisture susceptibility of the
the plot (error bars) shows the highest and lowest values of
asphalt mixes was not significantly affected by adding recycled
the data set. It is reported in the literature that a higher
plastics. Since most standards worldwide prescribe a minimum
CT index is indicative of better cracking resistance of the
TSR value of 80%, all recycled plastic asphalt mixes satisfy that
asphalt mix.
target. There was an increase in the TSR value of 0.5% RABS,
It can be seen that the cracking resistance of asphalt mixes
1% RABS and 2%RABS mixes compared to the control mix.
with RABS did not show significant variations in the CT index
However, the addition of 4% RABS did not significantly
compared to the control mix and among different RABS pro-
affect the TSR value. It is postulated that the bonding between
portions in the t-test with the significance level of 0.05. On the
RABS and bitumen was stronger than the bonding between
other hand, adding RPET to the mix showed a significant
bitumen and aggregate. However, when adding high pro-
increase in CT index, thus the cracking resistance of asphalt
portions of RABS (e.g. 4%), the film binder index was reduced
mixes. Higher content of RPET showed greater CT index
(Table 3), reducing the bond strength between RABS particles
values. In particular, RPET mixes with 0.5% recycled plastic
and bitumen. This resulted in a reduction in moisture resist-
had a statistically significant lower CT index than the mixes
ance of 4% RABS mix compared to lower-content RABS
with 1%, 2% and 4% RPET. However, the addition of 1% or
mixes. A similar situation was observed with RPET in which
more RPET exhibits no statistically significant change in the
the TSR values of RPET asphalt mixes were not significantly
CT index values.
different from the control mix when adding 1%, 2% and 4%
To limit test variability, five samples were tested for each
RPET, while 0.5% RPET showed a slight increase in TSR. In
mix. Theoretically, the addition of recycled materials could
conclusion, the addition of plastics as an aggregate replace-
increase variation in the test results due to the heterogeneity
ment did not have any adverse effect on the moisture resist-
of the final asphalt mixes and within the tested samples.
ance of asphalt mixes if the TSR threshold of 80% is
Recently, Habbouche et al. (2021) conducted a round-robin
considered.
testing programme to investigate the precision of the CT

Figure 6. CT index of asphalt mixes.

10 D. X. LU AND F. GIUSTOZZI

Table 7. Constitutive parameters of the CT index of asphalt mixes.

Mix HMA 0.5% RABS 1% RABS 2% RABS 4% RABS 0.5% RPET 1% RPET 2% RPET 4% RPET
Average air voids content (%) 6.8 6.7 6.9 7.0 7.5 6.5 7.2 6.9 7.8
Peak load (kN) 15.4 15.3 14.9 15.5 16.0 14.5 14.1 14.3 14.9
CoV (%) 3.6 4.9 7.6 4.0 4.8 2.0 7.9 5.8 2.3
L75 (mm) 9.2 10.0 9.9 9.8 10.0 10.2 11.0 11.0 10.7
CoV (%) 5.3 3.3 5.4 4.1 3.4 6.2 3.9 3.5 2.8
Fracture energy (J/m2) 15,843 16,555 15,834 16,468 16,250 16,000 16,572 16,689 17,325
CoV (%) 11.0 5.9 4.0 2.4 7.4 2.4 7.8 5.7 3.3
|m75| (kN/mm) 2.3 2.3 2.2 2.3 2.3 2.0 1.9 1.7 1.8
CoV (%) 10.0 9.0 15.6 10.5 13.1 7.6 10.5 9.9 4.4

index, which was not developed in ASTM standard D8225−19 4.4. Flexural stiffness and fatigue resistance
(ASTM 2019). The study reported that the CT index had rela-
The flexural modulus master curves at a reference temperature
tively high variation but a final comprehensive assessment on
of 20°C are shown in Figure 7. The results show that the
the variability of the CT index was not conducted yet. In this
addition of plastic as aggregate replacement of a certain aggre-
study, the variation associated with the CT index (Figure 6)
gate fraction did not significantly change the flexural stiffness
was assessed by looking at changes in the fracture energy Gf,
of the asphalt mixes compared to the control asphalt mix,
displacement at 75% peak load after the peak l75, and the absol-
especially when low plastic contents (i.e. 0.5% and 1% RABS
ute value of the post-peak slope abs(m75). In addition, peak
or RPET) were used. A more obvious change in stiffness was
load was also investigated to further provide information of
observed when adding a greater plastic content (e.g. 2% and
the test variability. As it can be observed in Table 7, the values
4%), at which the flexural modulus of recycled plastic mixes
of peak loads were rather consistent for all mixes, with CoV
was significantly lower than that of the control HMA. The
ranging from 5% to 8%. Another contributor of the CT
reduction in stiffness of recycled plastic mixes is related to
index, l75, also showed highly consistent results, with CoV ran-
the substantial lower stiffness of plastic compared to the virgin
ging from 5% to 7% (Table 7). RPET mixes tend to give greater
aggregate. Overall, there are no significant upward/downward
values of l75 than RABS mixes and the control mix, suggesting
movements of the master curves along the reduced frequency
greater plastic deformation tolerance developed in RPET
domain (i.e. most of the curves overlapping), meaning recycled
samples in the post-peak fracture process. In terms of fracture
plastics used as a partial aggregate replacement does not par-
energy, high consistency in the results was observed, with CoV
ticularly affect the temperature susceptibility of the mix at
varying from just above 2% to 12%. The post-peak slope abs
the particular content used in this study.
(m75) showed a slightly greater variation with CoV from
The fatigue resistance was evaluated based on the fatigue
7.5% to 15.5%, which is an acceptable range of CoV for asphalt
curves at different strain levels and the strain value ɛ6, at
testing (West et al. 2018). The post-peak slopes values of RPET
which the beam lasted for 1 million cycles before reaching
mixes were lower than those of RABS, indicating more ductile
the 50% stiffness failure criterion, was recorded. The results
behaviour of RPET samples during the fracture process at
showed a modest increment of fatigue life when adding plastic,
intermediate temperature.
from 1.3% to 38.3% compared to that of the control mix

Figure 7. Flexural stiffness master curves of asphalt mixes at a reference temperature of 20°C.
 INTERNATIONAL JOURNAL OF PAVEMENT ENGINEERING 11

(Figures 8 and 9). The proportion and type of plastic affected The results in this section prove the effect of the plastic
the fatigue behaviour of the mixes and the mixes with RABS aggregate shape and size on the fatigue behaviour of the
(coarse aggregate replacement) generally show better fatigue modified asphalt mixes. Compared to RPET, RABS is larger
performance than the mixes with RPET (fine aggregate repla- in size with sharp edges and spherical in shape, hence working
cement) at the same plastic content. like an effective part of the aggregate skeleton in the mix. Con-
For RABS, all mixes showed minimal difference in fatigue versely, RPET particles have a much smaller size, hence only
life at 400 micro-strain, which is the highest strain level used contributing to part of the fine aggregate in the mix structure.
in this study. The fatigue life becomes more different at In addition, RPET had more elongated and flaky particles,
lower strain levels. In terms of ɛ6, higher proportion of which could result in a lower ability to carry a load and
RABS showed greater increment of ɛ6; 9.8%, 17.2%, 23.9%, could produce a drawback in the fatigue resistance of the
and 30.4% compared to the ɛ6 of the control mix for 0.5%, mix compared to RABS particles which possess angularity in
1%, 2% and 4% RABS, respectively. In the case of RPET shape. The dissimilar shape and size of RABS and RPET par-
mixes, the fatigue curves tend to be parallel in the log-log ticles produced differences in the internal structure of mixes,
scale (Figure 8(b)). This indicates the RPET mixes have a simi- leading to variations in fatigue resistance of RABS and RPET
lar trend of fatigue life regardless of the applied strain levels. mixes. When comparing the fatigue resistance of the control
RPET mixes with 0.5% and 1.0% plastic showed minimal mix with plastic-modified mixes, it is clear that the mastic in
improvements in fatigue life; the ɛ6 value increased by 1.3% the virgin mix sustains the strain applied onto the beam
and 7.6% compared to the control mix for 0.5% and 1% because of the significant difference in stiffness of aggregate
RPET, respectively. Adding higher RPET contents promoted with mastic. However, in the case of the plastic-modified
greater resistance to fatigue, with 11.7% and 38.8% increase mixes, the stiffness of plastic is significantly smaller than the
for 2% and 4% RPET mixes, respectively. virgin aggregate, hence its capacity to partially contribute to

Figure 8. Fatigue curves of asphalt mixes: (a) HMA and RABS mixes, (b) HMA and RPET mixes.
12 D. X. LU AND F. GIUSTOZZI

Figure 9. ɛ6 of asphalt mixes.

sustaining the strain applied onto the beams. This resulted in rut depth reduction trend remained unchanged. Adding
enhanced fatigue resistance of plastic-modified mixes although RABS gave a maximum reduction of 24% in rut depth com-
more refined fatigue analysis approaches can be used to evalu- pared to the control mix.
ate the energy dissipation during fatigue tests of plastic asphalt Similar to the case of RABS, RPET mixes also showed
mixes. improvements in rutting resistance compared to the control
mix. The addition of 0.5% and 1% RPET resulted in 25%
and 16% reduction in the rut depth, respectively and 2% and
4.5. Rutting resistance
4% RPET mixes provided 22% and 29% reduction in the
The final rut depth results, after discounting the relative effect final rut depth compared to the control mix.
of moisture, are presented in Figure 10. In general, plastic- The results of the rutting test show that the addition of plas-
modified mixes showed better rutting resistance than the tics in the asphalt mix tends to increase the rutting resistance
control mix. However, a higher variation was observed for although the improvement is rather modest, with an average
the rutting depth values of asphalt mixes that contained plas- reduction in the rut depth ranging from 10 to 25%. It is well
tics, especially in the case of RABS at high content of plastic. acknowledged that the aggregate skeleton and the mastic
The addition of RABS showed a modest improvement in mainly contribute to the rutting resistance of asphalt concrete.
rutting resistance. Adding more than 1% RABS resulted in a Since the mixes in this study used the same binder and filler
greater variability of the rutting results although the overall proportion, the contribution of the mastic is deemed to be

Figure 10. Rut depth of asphalt mixes, HWTT – wet method, 20,000 passes and 50°C.
 INTERNATIONAL JOURNAL OF PAVEMENT ENGINEERING 13

identical. This means that the difference in rutting resistance one from the other. Therefore, two main performance par-
comes from the internal aggregate structure of the mixes. ameters, cracking and rutting resistance, should be evaluated
Because the plastic particles have lower stiffness than the virgin against the performance of the control mix. The perform-
aggregate, they can sustain more strain and thus energy ance-based evaluation is illustrated in Figure 11. It should be
applied on it. Once the load is removed, they tend to bounce noted that the IDEAL-CT test is an ASTM standard and has
back rather than leading to further densification like in the not yet adopted by Australian and New Zealand testing sys-
case of virgin aggregate which has much higher stiffness. How- tems. The introduction of this test aims to evaluate whether
ever, the behaviour under loading of each plastic type is differ- a more rapid test using inexpensive equipment is suitable for
ent depending on the mechanical properties of the plastic assessing the cracking resistance of the asphalt mixes. The per-
material and the physical appearance of the plastics, i.e. size formance-based graphs in Figure 11 create four quadrants with
and shape. This study found that the larger plastic particles different performance expectations by using the rutting depth
(i.e. RABS) help improve the rutting resistance better than and ɛ6 or CT index values of the control mix. In quadrant 1,
finer plastic aggregates (i.e. RPET). one can expect the alternative mix containing recycled
material has better rutting and cracking resistance than the
control mix. In quadrant 2, asphalt mixes with greater cracking
4.6. Performance and sustainability discussion
performance but lower rutting resistance will appear. Asphalt
It has become normal practice to justify the application of any mixes falling in quadrant 2 could be considered for use in
recycled materials in asphalt if the asphalt mix incorporating pavements where the main failure type is associated with
recycled material performs better than or equivalent to the cracking. Quadrant 3 contains mixes with better rutting resist-
conventional HMA. In that sense, the evaluation of the appli- ance than the control mix, but lower cracking resistance. This
cability of recycled plastic in asphalt mixes is recommended to quadrant shows mixes that are possibly suitable for pavements
being assessed by a performance-based method. For that, the with rutting as a primary concern but not cracking. On the
plastic-modified mixes were checked against different per- other hand, quadrant 4 includes asphalt mixes with lower per-
formances of the control HMA, including the compactability, formance in both rutting and cracking. Those mixes should
moisture susceptibility, cracking and rutting resistance. not be used as they provide no performance benefit. Finally,
In respect to the compactability, various Australian states the mixes laying on the two axes (having similar rut depth
and New Zealand regulated different minimum air void values or similar cracking resistance compared to the control mix)
at the end of the compaction process, which are normally 2% are considered iso-performance mixes. In this study, all
or 2.5% for heavy duty pavement. The cap in the minimum air recycled plastic mixes lay in quadrant 1, meaning that they
void thresholds aims to avoid excessive compaction which had better rut and cracking resistance than the control mix.
could result in the instability of the asphalt mix and lead to rut- This indicates the suitability of the recycled plastic mixes for
ting susceptibility. Currently, there is no maximum air voids use in road pavements in Australia and New Zealand.
threshold for the compactability evaluation in Australia and In conclusion, the suitability of the addition of recycled
New Zealand. This feature is ultimately important to avoid plastic in asphalt should consider several factors, from the
possible difficulties during the compaction stage when adding compaction aspects to moisture susceptibility, stiffness, crack-
excessive content of recycled materials, specifically recycled ing and rutting resistance of the mix. Recycled plastic should
plastics in this study. A particular example of the compaction only be considered for use in asphalt mixes if its addition
difficulty that can possibly be experienced is the case of 4% could benefit the final performance of the modified mix com-
RPET mix which had rather low compactability and high air pared to a conventional mix. Recycling various materials in
voids at the end of the compaction process. Low compactabil- asphalt per se is not a quantifiable measure of sustainability
ity causes difficulties during compaction in the field and pro- and its effect can be detrimental if the final durability of the
motes the use of additional compaction energy that may road asset is reduced. A simple way to compare and assess
result in fragmentation of the aggregate. High air voids content pros and cons could be through the adoption of the perform-
after compaction could lead to premature damage of the road ance graphs in Figure 11. It is recommended that plastic-
pavement. modified mixes in quadrant 1 should be considered as being
Regarding the moisture susceptibility test, a minimum TSR ready for implementation and field testing. Whereas plastic-
limit of 80% is set by the Australian states and New Zealand. In modified mixes in quadrants 2 and 3 could be considered for
this study, all mixes satisfied the moisture resistance particular situations depending on the type of distress that is
requirement. of major concern and the particular pavement design. Evi-
The resilient modulus and the flexural stiffness test are two dently, any mix assessed up to this level should have received
types of stiffness tests used in parallel by Australian states and prior approval to what relates to compactability and moisture
New Zealand. The flexural stiffness test was used in this study resistance. A particular example of this is the case of 4% RPET,
and the test was conducted via the four-point bending beam in which the mix showed better performance than the control
apparatus. The results from the stiffness test are generally mixes in both cracking and rutting, yet the mix had very poor
not regulated as a pass or fail criterion but are reported mainly compactability. For that, the 4% RPET mix is considered not to
for pavement design purposes. be suitable for construction. In addition to this example, the
The above criteria are mostly based on thresholds with mixes with 2% and 4% RABS had less than 2% air voids
many different mixes being able to comply with these after 160 Superpave gyrations (e.g. corresponding to 250 gyra-
thresholds although offering a very different performance tions in the New Zealand compaction standard, and 250 or 350
14 D. X. LU AND F. GIUSTOZZI

Figure 11. Performance graphs considering rutting and cracking of asphalt mixes (a) rutting and ɛ6, and (b) rutting and CT index.
 INTERNATIONAL JOURNAL OF PAVEMENT ENGINEERING 15

gyrations in the Australia system depending on the states), RPET should be limited to maximum 1% to partially replace
thus these mixes should not be used for heavily trafficked pave- the finer or sand fraction. The present study also offers some
ments. Although the findings from this study are for recycled user-friendly evaluation method based on the performance
plastics in asphalt, the same concept could be applied to inves- graphs developed to quickly discern among mixes that incor-
tigate the applicability of other recycled materials. porate recycled materials. A sustainable solution should be
one that not only looks at the saving of raw materials per se
but also ensures a certain level of durability is obtained com-
5. Conclusions
pared to the mixes already in use.
This study investigated the suitability and the relative effects of
adding recycled plastics PET and ABS in dense-graded HMA
using the dry process to partially replace natural aggregate at Disclosure statement
different proportions of 0.5%, 1%, 2% and 4% (by weight of No potential conflict of interest was reported by the author(s).
the mix). Based on the results, the following conclusions can
be drawn;
Funding
− The addition of recycled plastics as a fraction of natural
aggregate in dense-graded HMA is a feasible solution that The authors would like to acknowledge the financial contribution of Aust-
can potentially use a much larger quantity of recycled roads Ltd through the funding of the project APT6305 – Use of road-
grade recycled plastics for sustainable asphalt pavements.
material compared to the wet process (bitumen modifi-
cation with plastic).
− The presence of plastics affected the compactability of the
asphalt mixes. The extent of the effect depends on the pro- ORCID
portion, size and shape of the recycled plastic added to the Filippo Giustozzi http://orcid.org/0000-0002-6956-4686
mix. It was found that coarse plastic particles with spherical
and fragmented shape did not reduce the compactability of
asphalt mixes. References
− The moisture resistance of asphalt mixes was observed not AASHTO, 2001. AASHTO provisional standards: interim edition.
to be affected by the addition of the recycled plastics since Washington, DC: American Association of State Highway and
all mixes passed the 80% limit for TSR value. No stripping Transportation Officials (AASHTO), Northwest.
was observed for recycled plastic mixes in the TSR test. AASHTO, 2018. AASHTO T 324-17. Standard method of test for Hamburg
However, the indirect tensile strength of the recycled asphalt wheel-track testing of compacted asphalt mixes. American Association
of State Highway and Transportation Officials (AASHTO),
mixes tends to be similar to or slightly lower than the con- Northwest, Washington DC.
trol mix. More advanced moisture tests should be used to AG:PT/T232, 2007. Stripping potential of asphalt – tensile strength ratio.
assess moisture damage. Austroads Manual of Test Methods.
− The stiffness of asphalt mixes changed slightly by adding Ahmadinia, E., et al., 2011. Using waste plastic bottles as additive for stone
plastics. The master curves of the mixes were close to the mastic asphalt. Materials & Design, 32 (10), 4844–4849.
Ahmadinia, E., et al., 2012. Performance evaluation of utilization of waste
control mix with many of them overlapping. A minimal polyethylene terephthalate (PET) in stone mastic asphalt. Construction
difference in the master curve was only observed at high and Building Materials, 36, 984–989.
recycled plastic contents such as 2% and 4% RABS. AS 2008-2013, 2013. Bitumen for pavements. Standard Australia.
− The addition of recycled plastic helps enhance the cracking AS/NZS (Australian/New Zealand Standard), 2014a. Methods of
resistance of asphalt mixes. The extent of the improvement sampling and testing asphalt. Part 1: Sample preparation – mixing,
quatering and conditioning of asphalt in the laboratory. AS/NZS
relies on the content and the type of plastic used. Quantitat- 2891.2.1. Sydney, Australia: AS/NZS.
ively, the improvement ranged from minor to significant. AS/NZS (Australian/New Zealand Standard), 2014b. Methods of
The general trend indicates that higher recycled plastic con- sampling and testing asphalt. Part 2: Sample preparation – compaction
tent provides some enhancement in the cracking resistance. of asphalt test specimens using a gyratory compactor. AS/NZS
− The plastic-modified asphalt mixes had better rutting resist- 2891.2.2:2014, Sydney, Australia: AS/NZS.
ASTM, 2019. “D8225 − 19. Standard test method for determination of
ance than the conventional mix. The rutting performance of cracking tolerance index of asphalt mix using the indirect tensile crack-
the mixes was also observed to depend on the type and con- ing test at intermediate temperature.” ASTM International.
tent of the recycled plastic used. On average, adding plastic ASTM, 2021. “D8237 − 21. Standard test method for determining fatigue
could reduce the rut depth by 20%. failure of asphalt-aggregate mixtures with the four-point beam fatigue
device.” ASTM International.
Audy, R., et al., 2022. Selection of recycled waste plastic for incorporation
Based on the comprehensive evaluation reported in this in sustainable asphalt pavements: a novel multi-criteria screening tool
study that accounted for compactability, moisture resistance, based on 31 sources of plastic. Science of the Total Environment, 829,
cracking performance tests and rutting, it is suggested that 154604.
up to 1% of RABS can be used to partially replace the coarse Austroads, 2008. A review of Austroads gyratory compaction research.
aggregate fraction for heavy traffic pavements in Australia Sydney, NSW: Austroads. APT-T94-08.
Austroads, 2016. AGPT/T274. Characterisation of flexural stiffness and
and New Zealand. Up to 2% RABS can instead be used for fatigue performance of bituminous mixes.
medium or low traffic pavements according to the Australian Biligiri, K P, et al., 2001. Rational modeling of tertiary flow for asphalt
and New Zealand definition of traffic levels. The addition of mixtures. Transportation Research Record, 63–72.
16 D. X. LU AND F. GIUSTOZZI

Costa, L. M., et al., 2013. Incorporation of waste plastic in asphalt binders Moghaddam, T. B., Soltani, M., and Karim, M. R., 2014. Experimental
to improve their performance in the pavement. International Journal of characterization of rutting performance of polyethylene terephthalate
Pavement Research and Technology, 6 (4), 457–464. modified asphalt mixes under static and dynamic loads. Construction
Dalhat, M., and Al-Abdul Wahhab, H., 2017. Performance of recycled and Building Materials, 65, 487–494.
plastic waste modified asphalt binder in Saudi Arabia. International Nizamuddin, S., et al., 2020. Recycled plastic as bitumen modifier: The
Journal of Pavement Engineering, 18 (4), 349–357. role of recycled linear low-density polyethylene in the modification
Dalhat, M., Al-Abdul Wahhab, H., and Al-Adham, K., 2019. Recycled plas- of physical, chemical and rheological properties of bitumen. Journal
tic waste asphalt concrete via mineral aggregate substitution and binder of Cleaner Production, 266, 121988.
modification. Journal of Materials in Civil Engineering, 31 (8), 04019134. Nizamuddin, S., Boom, Y. J., and Giustozzi, F., 2021. Sustainable poly-
Diab, A., Enieb, M., and Singh, D., 2019. Influence of aging on properties mers from recycled waste plastics and their virgin counterparts as
of polymer-modified asphalt. Construction and Building Materials, bitumen modifiers: A comprehensive review. Polymers, 13 (19),
196, 54–65. 3242.
Giustozzi, F., and Boom, Y., 2021. Use of road-grade recycled plastics for O’Farrell, K, 2019. 2017-18 Australian plastics recycling survey – national
sustainable asphalt pavements: overview of the recycled plastic industry report. Australian Government Department of the Environment and
and recycled plastic types. Austroads 2021. Energy [Online]. Available from: https://bit.ly/3a4hoxs.
Habbouche, J., et al., 2021. Round Robin Testing Program for the indirect Padhan, R. K., Sreeram, A., and Gupta, A., 2020. Evaluation of trans-poly-
tensile cracking test at intermediate temperature: Phase I. octenamer and cross-linking agents on the performance of waste poly-
Hasan, M. R. M., et al., 2016. A simple treatment of electronic-waste plas- styrene modified asphalt. Road Materials and Pavement Design, 21 (4),
tics to produce asphalt binder additives with improved properties. 1170–1182.
Construction and Building Materials, 110, 79–88. Ranieri, M., et al., 2017. Asphalt surface mixes with improved perform-
Joohari, I. B., and Giustozzi, F., 2020. Chemical and high-temperature ance using waste polymers via dry and wet processes. Journal of
rheological properties of recycled plastics-polymer modified hybrid Materials in Civil Engineering, 29 (10), 04017169.
bitumen. Journal of Cleaner Production, 276, 123064. Rebbechi, J., and Petho, L., 2014. Guide to pavement technology: part 4B.
Lastra-González, P., et al., 2016. Comparative analysis of the performance Asphalt.
of asphalt concretes modified by dry way with polymeric waste. VicRoads, 2021. VicRoads Section 407:2021, dense graded asphalt.
Construction and Building Materials, 112, 1133–1140. West, R., et al., 2018. Development of a framework for balanced mix
Lastra-González, P., et al., 2021. Analysis of replacing virgin bitumen by design. Project NCHRP, 20–07.
plastic waste in asphalt concrete mixtures. International Journal of Willis, R, Yin, F., and Moreas, R., 2020. Recycled plastics in asphalt Part A:
Pavement Engineering, 1–10. state of the knowledge. NAPA-IS-142 report. National Asphalt
Mallick, R. B., et al., 2004. Evaluation of use of synthetic lightweight Pavement Association.
aggregate in hot-mix asphalt. Transportation Research Record: Wu, S., and Montalvo, L., 2020. Repurposing waste plastics into cleaner
Journal of the Transportation Research Board, 1891 (1), 1–7. asphalt pavement materials: a critical literature review. Journal of
Martin-Alfonso, J., et al., 2019. Use of plastic wastes from greenhouse in Cleaner Production, 124355.
asphalt mixes manufactured by dry process. Road Materials and Yin, F., etal, 2014. Novel method for moisture susceptibility and rutting
Pavement Design, 20 (sup1), S265–S281. evaluation using hamburg wheel tracking test. Transportation
Mikhailenko, P., et al., 2021. Effect of waste PET and CR as sand replace- Research Record: Journal of the Transportation Research Board, 2446
ment on the durability and effect of waste PET and CR as sand repla- (1), 1–7. https://doi.org/10.3141/2446-01.
cement on the durability and acoustical properties of semi dense Yoo, P., and Al-Qadi, I., 2014. Pre-and post-peak toughening behaviours
asphalt (SDA) mixes. Sustainable Materials and Technologies, 29, of fibre-reinforced hot-mix asphalt mixes. International Journal of
e00295. Pavement Engineering, 15 (2), 122–132.