Journal of Cleaner Production 156 (2017) 911e922

Contents lists available at ScienceDirect

Journal of Cleaner Production

journal homepage: www.elsevier.com/locate/jclepro

Warm recycling of flexible pavements: Effectiveness of Warm Mix

Asphalt additives on modified bitumen and mixture performance
Arianna Stimilli Research Associate *, Amedeo Virgili Associate Professor,
Francesco Canestrari Full Professor

Politecnica delle Marche, Via Brecce Bianche, Ancona, Italy
Universita

a r t i c l e i n f o a b s t r a c t

Article history: In pavement industry, environmental and economical sustainability stimulates technical solutions able to
Received 6 September 2016 drastically decrease pollutants and energy costs caused by high production temperatures. In this sense,
Received in revised form recent research efforts focused on innovative technologies able to significantly reduce mixing and
28 March 2017
compaction temperature. The so called Warm Mix Asphalts (WMA) are bituminous mixtures produced at
Accepted 29 March 2017
Available online 19 April 2017
reduced temperatures through specific additives based on different mechanisms. The novelty of WMA
technologies determines a lack of experience about their effectiveness, which requires dedicated
Handling Editor: Yutao Wang research activities, especially when polymer modified bitumens and Reclaimed Asphalt Pavement (RAP)
are used.
Keywords: Given this context, the present study proposes a comprehensive laboratory investigation on dense-
Warm Mix Asphalt graded mixtures produced through different WMA additives representative of the main categories
Recycling currently available on the market (i.e. chemical, organic and water-based). Mixtures were prepared
Asphalt mixture including Styrene-Butadiene-Styrene polymer modified bitumen and 25% of RAP to evaluate potential
Emission
benefits deriving from the combination of warm and recycling techniques. For a complete understanding
Binder
of WMA additive effects, mechanical tests (i.e. compactability, stiffness, fatigue) carried out on mixtures
in a broad range of loading configurations were integrated by rheological analyses on bitumens.
Results indicate that lower production temperatures allowed a significant decrease in stiffness,
effectively balancing the inclusion of RAP material without penalizing mixture performance. Compact-
ability, volumetric and fracture properties indicate the possibility to produce suitable recycled warm
mixtures when the appropriate WMA additive is selected. The chemical additive appeared able to
provide overall improved performance, whereas the organic additive made the mixture brittle and
susceptible to permanent cracking.
© 2017 Elsevier Ltd. All rights reserved.

1. Introduction cycle environmental performance of asphalt recycling. On the other

Over the last years, economic and environmental benefits
incentivized pavement industry towards the production of asphalt
mixtures including high amount of Reclaimed Asphalt Pavement
(RAP). Lo Presti et al (Lo Presti et al., 2016) and Miliutenko et al.
(2013) provided exhaustive reviews of current practices for the
use of high quantities of RAP proposing useful guidelines for
adapting the design process of binders for asphalt mixtures incor-
porating up to 100% RAP and potential ways of improving the life-
* Corresponding author.
E-mail addresses: a.stimilli@univpm.it (A. Stimilli), a.virgili@univpm.it
(A. Virgili), f.canestrari@univpm.it (F. Canestrari).
hand, hot recycling techniques usually require to superheat virgin
aggregates to use their excessive thermal energy to heat up and dry
RAP particles added at ambient temperature to the other material
components right before mixing. As well described by Brock and
Richmond (2007), this superheating implies higher energy and
fuel consumption as well as an increase in air emissions during
production, which could overweight the benefits deriving from the
re-use of RAP and compromise large-scale productions and in situ
applications of recycled mixtures. At the same time, this precaution
is essential to avert the risk of mixtures with low workability and
inadequate volumetric properties, especially with high RAP
percentage.
In this sense, Warm Mix Asphalt (WMA) represents the most
promising research frontier for avoiding the renounce of high RAP

http://dx.doi.org/10.1016/j.jclepro.2017.03.235
0959-6526/© 2017 Elsevier Ltd. All rights reserved.
912 A. Stimilli et al. / Journal of Cleaner Production 156 (2017) 911e922
amounts. They can significantly reduce mixing and compaction
temperatures compared to traditional Hot Mix Asphalts (HMA)
with remarkable advantages in terms of environment and costs. As
illustrated in many literature reviews on the subject (Rubio et al.,
2012; Dinis-Almeida and Lopes Afonso, 2015; D'Angelo et al.,
2008), WMA technologies consists of products that operate on the
basis of different mechanisms (i.e. organic additives, water-based
processes, water-containing additives, chemical additives). Due to
the recent development of these technologies, laboratory in-
vestigations or in-situ applications are still very limited, especially
with respect to recycled mixtures including polymer modified bi-
tumens. Determining potentials and drawbacks of these materials
can help to judge the feasibility of WMA mixtures and identify
proper mix design and in-plant operational precautions to guar-
antee the fulfilment of all technical requirements prescribed for
traditional HMAs.
Several aspects suggest that mixture performance could signif-
icantly benefit from the combination of recycling and warm
technologies.
In this perspective, workability represents a major issue related
to warm recycled mixtures. Production temperatures and
reclaimed aged bitumen strongly affect compactability: reduced
production temperatures induce lower bitumen aging, hence
mitigating the stiffening effect of the aged bitumen released by RAP
aggregates. Effects of WMA technologies on recycled mixtures
workability is comparable with those promoted by the use of softer
virgin bitumens, as confirmed by preliminary studies which vali-
dated the possibility to include high RAP amount (up to 90e100%)
in new warm mixtures, guaranteeing reasonable workability and
acceptable in situ compaction efforts (D'Angelo et al., 2008; Mallick
et al., 2008). Mallick et al. (2008) investigated the feasibility of
using an organic WMA additive for recycling up to 75% RAP,
obtaining successful results through a significantly lower bitumen
grade. Tao and Mallick (Tao and Mallick, 2009) showed that WMA
additives guaranteed optimum workability of 100% RAP mixtures at
temperatures as low as 110
C.
At the same time, the increased stiffness provided by RAP can
counterbalance lower aging effects due to reduced production
temperatures and mitigate the consequent predictable higher
rutting aptitude as demonstrated by Lee et al. (2009) through
specific bitumen performance tests. The long-term aged bitumen
released by RAP was effective for increasing mixture rutting resis-
tance. Also Zhao et al. (2013) detected good permanent deforma-
tion resistance investigating a wide range of different plant-
produced warm recycled mixtures (RAP content up to 40%)
through the Asphalt Pavement Analyzer and the Hamburg Wheel-
Track test.
On the other hand, reduced production temperatures could
imply lower bitumen reactivation of RAP with repercussions on mix
design (i.e. optimum quantity of virgin bitumen) and, consequently,
on mechanical performance. In fact, an overestimation of RAP
bitumen reactivation would lead to insufficient virgin bitumen to
guarantee proper volumetric and adhesion properties of the final
mixture. However, the reactivation process is also time-dependent
and not only temperature-dependent. In this sense, previous ex-
periences on WMA recycled mixtures suggest that, when sufficient
mixing and storage time is provided (i.e. 2 h at selected WMA
compaction temperatures), RAP bitumen reactivates equally to hot-
recycled mixtures (Bonaquist, 2011).
The water susceptibility is a further aspect that encourages the
combination of WMA technologies and recycling. In fact, aggregates
heated at reduced temperatures risk to retain humidity compro-
mising the adhesion between bitumen and aggregates. For recycled
mixtures, adding water-resistant aggregates (thanks to the thin
film of aged bitumen coating RAP particles) guarantees improved
adhesion properties and higher ravelling resistance, with conse-
quent benefits in terms of water susceptibility. Zhao et al. (2013),
considering various WMA technologies and RAP contents, showed
satisfactory moisture resistance for all WMA-high RAP mixtures
except in case of foam technologies. Analogously, Frigio et al. (2016)
proved that RAP significantly limits negative effects of reduced
production temperatures in terms of adhesion and ravelling resis-
tance in both open and dense-graded mixtures subjected to water
conditioning. Other studies (Frigio and Canestrari, 2016) on warm
porous asphalt layers including RAP highlighted that chemical ad-
ditives were the only able to ensure adequate water resistance.
Useful indications are reported in the review by Kakar et al. (2015)
who summarized various techniques for assessing the moisture
damage in case of WMA mixtures.
In addition, each kind of WMA technology acts based on
particular mechanisms that provide specific benefits and draw-
backs. Therefore, warm mixtures behavior cannot be generalized
and has to be referred to the particular kind of additive employed.
All abovementioned aspects are further complicated when
asphalt mixtures contains Polymer Modified Bitumens (PMB),
especially if both virgin and RAP bitumens are polymer modified. In
fact, interactions between virgin and aged polymer chains as well
as WMA additives modify rheological properties of bitumen and
alter mixture performance. Few researches on this matter (Kim
et al., 2012; Kanitpong et al., 2007) have shown potential advan-
tages in terms of indirect tensile strength, storage stability, water
and rutting resistance, whereas potential drawbacks are mainly
related to increase in stiffness and fatigue behavior, particularly
when organic additives are used (Kim et al., 2012).
Although several aspects support the combination of reduced
production temperatures and recycling procedures, a number of
open issues still require further investigations due to the recent
development of WMA technologies and the consequent lack of
consolidated experiences.
Given this background, the present study aimed at providing
useful insights on mechanical performance of dense-graded mix-
tures produced with various WMA additives representative of the
three main categories available on the market (i.e. organic, chem-
ical and water-based). Two additional mixtures were produced and
used as reference for comparison purposes: the mixture usually
employed for Italian motorway binder layers, mixed and com-
pacted at standard temperatures without any additive; the same
one prepared at reduced temperatures (i.e. 130 and 120
C, 40
C
lower than standard mixing and compaction temperature, respec-
tively) without any WMA product. The latter was used to isolate the
effect of temperature reduction from the one related to WMA ad-
ditives. All mixtures were prepared with the same composition (i.e.
content/type of bitumen and aggregate distribution) including 25%
of unfractioned RAP in accordance with the mix design currently
employed for motorway binder layers.
The main objective of the research was to evaluate the potential
of combining WMA and recycling procedures. The effectiveness of
various WMA technologies for producing bituminous mixtures at
significantly reduced working temperatures was assessed to iden-
tify the product that more efficiently preserves overall mixture
performance in terms of volumetric properties, stiffness and frac-
ture resistance. To this aim, a wide range of laboratory tests were
carried out on laboratory produced samples. To further evaluate the
effectiveness of WMA additives and gain a complete understanding
of their effects on mixture performance, additional rheological tests
on polymer modified bitumens mixed with the selected WMA
additives were performed and an estimate of energy savings and
pollutants emissions is provided.
 A. Stimilli et al. / Journal of Cleaner Production 156 (2017) 911e922 913

2. Experimental program conducted. In the case of short-term aging, to properly simulate

oxidation effects undergone during transportation and laydown,
2.1. Materials bituminous blends were subjected to RTFOT carried out at tem-
peratures consistent with WMA productions. Hence, 130
C was
Five dense graded bituminous mixtures were produced in lab- selected as RTFOT test temperature. Additionally, a blend made
oratory according to the mix design prescribed by Italian technical with bitumen H and 25% of long-term aged bitumen without any
standards for motorway binder layers. All mixtures included 25% by additive was tested as representative of the control mixture. When
aggregate weight of unfractioned RAP (0/16 mm) combined with required, this bituminous blend was short-term aged at standard
various virgin limestone aggregate fractions (0/4, 8/16, 14/25 and temperature (i.e. 163
C) to reproduce the common asphalt plant
filler) such that the Nominal Maximum Aggregate Size was 20 mm production process.
(Table 1). A PMB (coded as H), containing 3.8% of Styrene-
Butadiene-Styrene (SBS), was used as virgin bitumen for all mix-
tures (Table 2). A total bitumen content of 4.8% by aggregate weight 2.2. Mixtures specimens preparation and testing
was adopted, accounting for the virgin bitumen and the bitumen
coating RAP particles (i.e. 5.23% by RAP weight). RAP came from the For preparing the investigated asphalt mixtures, virgin aggre-
milling of bituminous base and binder motorway layers, containing gates and RAP where mixed with the specific bitumen by means of
the same SBS modified bitumen H themselves. Four out of five a laboratory planetary equipment. To adequately reproduce what
mixtures (hereafter named as “WMA mixtures”) were mixed and usually followed at the asphalt plant, RAP was stored at ambient
compacted at temperatures equal to 130
C and 120
C, respec- temperature and directly added into the mixer with the virgin
tively. The fifth mixture, adopted as control for comparison pur- aggregates pre-heated in a forced draft-oven at the appropriate
poses, was mixed at 170
C and compacted at 160
C according to mixing temperature (i.e. 130
C for WMA mixtures, 170
C for the
the practice currently employed at the asphalt plant. control mixture). Then, the hot bitumen, prepared with or without
In the case of WMA mixtures, the bitumen was previously WMA additives as specified in the previous paragraph, was added
blended for 10 min through a high shear mixer with WMA additives at 170
C and mixed for few minutes with aggregates until a uni-
dosed in quantity as prescribed by the producers, and then added to form mix was obtained. Before preparing testing specimens, the
the aggregates. Three WMA additives, representative of the main material was stored for 1 h in the oven at the specific compaction
categories available on the market, were selected: temperature (i.e. 120
C for WMA mixtures, 160
C for the control
mixture) in order to simulate the short-term aging effects related to
- chemical additive (coded as C), dosed as 0.5% by bitumen transportation and lay-down. A Superpave Gyratory Compactor
weight; (SGC) was used to prepare cylindrical specimens, which were
- organic additive (coded as O), dosed as 3% by total bitumen subsequently cored and cut to obtain suitable dimensions for per-
weight; forming a wide range of mechanical tests as following specified:
- zeolite (coded as Z), dosed as 0.3% by total mixture weight.
* Indirect Tensile Strength (ITS) tests carried out at 25
C by
The fourth WMA mixture was prepared at reduced tempera- imposing a deformation of 25 mm/min along a diametrical line
tures without any WMA additive and used to distinguish temper- load until the maximum failure strength was reached (EN
ature reduction and WMA additives effects on mixture workability. 12697-23);
In order to evaluate the effects of WMA additives on rheological * cyclic uniaxial compression tests to determine the complex
properties of the bituminous component, bitumens representative modulus (E*) and the phase angle (d). Tests were run at five
of each asphalt mixture investigated in this study were concur- temperatures (10, 20, 30, 40 and 50
C) and six frequencies (0.1,
rently prepared in laboratory. Dosed as previously reported, each 0.3, 1, 3, 10 and 20 Hz) by applying sinusoidal load waves. Tests
WMA additive was blended with the virgin bitumen H by means of were carried out in controlled strain mode, setting 50 mε as
a high shear mixer. To simulate the presence of RAP, 25% by target strain level (AASHTO TP 79-09);
bitumen weight (i.e. RAP content in the corresponding mixture) of * cyclic indirect tension tests at 20
C for the assessment of the
long-term aged bitumen H was added to each bituminous blend Indirect Tensile Stiffness Modulus (ITSM) according to EN
(hereafter named as “WMA bitumens”). The long-term aged 12697-26;
bitumen was obtained through Rolling Thin Film Oven Test (RTFOT) * Semi-Circular Bending (SCB) tests at a temperature of 10
C to
and Pressure Aging Vessel (PAV). Bitumens were unaged, short- evaluate the low temperature fracture propagation resistance by
term and long-term aged, depending on the type of test using a constant vertical deformation rate equal to 5 mm/min
(EN 12697-44);
* Indirect Tensile Fatigue Test (ITFT) for the estimation of the
Table 1 resistance to crack initiation, carried out at 20
C in controlled-
Aggregate and mixture grading distribution. stress mode with a rise-time of 124 ms (BS DD ABF, 1997). Failure
Sieve opening [mm] Passing [%] was established when the complete fracture of the specimen
was obtained. A minimum of 3 stress levels for each mixture
14/25 8/20 0/4 Filler RAP 0/16 Mixture
were applied.
31.5 100 100 100 100 100 100
20 52.9 100 100 100 100 92.5
A servo-pneumatic testing machine equipped with an envi-
14 0.3 78.9 100 100 96.9 77.6
10 0.3 31.6 100 100 81.1 60.9 ronmental chamber was used to complete the mechanical
6.3 0.3 0.8 100 100 61.3 47.6 characterization.
2 0.3 0.8 62.4 100 41.2 31.7 Moreover, compactability and volumetric properties were
0.5 0.3 0.8 14.3 100 25.6 13.8
investigated through the calculation of the Compaction Energy
0.25 0.3 0.8 4.4 99.2 17.2 8.8
0.063 0.3 0.8 1 89.8 8.8 5.4
Index (CEI) and the air void content derived from SGC data.
Concurrently, rheological tests were performed on bituminous
Combination [%] 16 27 29 3 25 e
blends as follows:
914 A. Stimilli et al. / Journal of Cleaner Production 156 (2017) 911e922

Table 2
Hard modified bitumen characteristics.

Bitumen characteristics Aging Standard Unit Virgin bitumen H

SBS polymer content by weight e e % 3.8

Penetration [25
C; 100 g; 5s] Virgin ASTM D5 0.1 mm 54
RTFOT 27


Ring and ball softening point Virgin ASTM D36 C 71
RTFOT 77


Dynamic viscosity @ 135 C Virgin ASTM D4402 Pa$s 1.24
Mass loss after RTFOT RTFOT ASTM D2872 % 0.05
G*/sind @ 10 rad/s, 64
C Virgin ASTM D7175 kPa 1.90
RTFOT 8.32
G*$sind @ 10 rad/s, 25
C PAV ASTM D7175 kPa 1295.50
Creep stiffness @ 6
C PAV ASTM D6648 kPa 65190
m-value @ 18
C PAV ASTM D6648 e 0.33

PG AASHTO M 320 64e16

* viscosity tests carried out at 5 temperatures (i.e. 110, 130, 150,
170 and 180
C) through a Brookfield viscometer (EN 13302-2);
* frequency sweep tests by means of a Dynamic Shear Rheometer
(DSR) in plate-plate configuration to characterize the blends in
terms of complex modulus (G*) and phase angle (d). Tests were
performed in controlled strain mode (deformation level ¼ 0.5%)
on unaged blends by applying sinusoidal load waves in a fre-
quency range from 1 to 30 Hz and in a temperature range from 4
to 82
C (plate diameter ¼ 8 mm, gap ¼ 2 mm for test temper-
ature between 4 and 34
C; plate diameter ¼ 25 mm,
gap ¼ 1 mm for test temperature between 40 and 82
C);
* oscillatory loading tests performed on unaged and short-term
aged materials (AASTHO M 320). Tests were run by means of
the DSR (plate diameter ¼ 25 mm, gap ¼ 1 mm) at 5 tempera-
tures (52, 58, 64, 70, 76 and 82
C) with a constant strain
amplitude of 0.5% and a constant frequency of 1.59 Hz;
* oscillatory loading tests performed on long-term aged materials
at 20
C. Tests were run by means of the DSR (plate
diameter ¼ 8 mm, gap ¼ 2 mm) with a constant strain ampli-
tude of 0.5% and a constant frequency of 1.59 Hz.
The summary of the experimental program for both mixtures
and bitumens is reported in Table 3 along with the number of
replicates carried out. For each test, a one-way ANOVA analysis at
95% confidence level was performed to verify the statistical sig-
nificance of results.
3. Results and analysis
3.1. Workability and ITS results
Workability was evaluated in terms of compactability and
volumetric properties of laboratory compacted specimens. The
Compaction Energy Index (CEI) and the air void content were
adopted as main parameters to compare materials. CEI represents
the in-situ energy necessary to obtain a certain density during
construction. For dense-graded mixtures, it is calculated as the area
under the densification curve between the 8th gyration and the
gyration corresponding to the 92% of the maximum density
(Mahmoud and Bahia, 2004). Therefore, the higher the CEI, the
lower the compactability. Mixtures analyses were integrated by
viscosity tests on bitumens, each one representative of one inves-
tigated mixture as specified in paragraph x2.1.
Although bitumens analysis did not show significant difference
in terms of viscosity between materials (Fig. 1), the evaluation of
compactability for the corresponding mixtures (Table 4a) demon-
strated WMA additives effectiveness. Warm mixtures were able to
achieve similar workability compared to the control HMA despite
the remarkable gap in production temperatures and the conse-
quent higher bitumen viscosity.
The effectiveness of WMA additives is particularly appreciable
when WMA mixtures are compared with the mixture prepared at
lower production temperatures without adding any WMA additive
(M_W): a drastic increase in CEI value was detected for the latter
compared to both the control HMA and WMA mixtures.
Bitumens and mixtures results suggest that WMA additives
provided good workability promoting better interactions between
aggregates and bitumen rather than acting at the viscosity level.
Such a consideration can be related to lubricity mechanisms, which
enhance the achievement of improved internal skeleton and opti-
mum volumetric properties. Results obtained in terms of air voids
(Vm), Voids in Mineral Aggregates (VMA) and Voids Filled with
Bitumen (VFB) measured at Ndes (i.e. 120 gyrations) supports this
finding (Table 4a). For all mixtures, except the one realized at
reduced temperatures without WMA additives, the voids content
respected the limit range prescribed by technical standards (i.e.

Table 3
Experimental program.

Mixture Code Test type (number of replicates)

ITS Complex modulus ITSM SCB ITF

Control mixture (HMA, no WMA additive) M_H 4 2 6 4 6

Warm mixture WITHOUT WMA additive) M_W 4 2 6 4 6
Warm mixture þ CHEMICAL WMA additive M_WC 4 2 6 4 6
Warm mixture þ ORGANIC WMA additive M_WO 4 2 6 4 6
Warm mixture þ ZEOLITE WMA additive M_WZ 4 2 6 4 6

Bitumen Code Test type (number of replicates)

Viscosity Frequency sweep Oscillatory tests

Hard (WITHOUT WMA additive) B_H 2 6 4

Hard þ CHEMICAL WMA additive B_WC 2 6 4
Hard þ ORGANIC WMA additive B_WO 2 6 4
Hard þ ZEOLITE WMA additive B_WZ 2 6 4
 A. Stimilli et al. / Journal of Cleaner Production 156 (2017) 911e922 915

5 expected limited bitumen reactivation due to lower mixing and

compaction temperatures. The effective working bitumen, param-
4
eter defined by Stimilli et al. (2015), acting in warm mixtures is
likely composed by less stiff aged bitumen than traditional hot
recycled mixtures and, consequently, is characterized by lower
Log(η) [mPa ∙ s]

3 viscosity and better workability.

Overall viscosity and compactability analyses lead to the
2 conclusion that the equiviscous principle usually employed for
B_H determining optimum mixing and compaction temperatures is not
B_WC suitable for warm recycled mixtures prepared with polymer
1
B_WO modified bitumens, validating what already observed through
B_WZ laboratory viscosity tests by previous researches focused on WMA
0 mixtures with unmodified bitumens (Bonaquist, 2011; Zeng et al.,
100 120 140 160 180 200
2014). Bitumen viscosity analysis is not sufficient to properly
Temperature [°C]
evaluate overall workability of WMA mixtures especially when
Fig. 1. Dynamic viscosity of laboratory prepared bituminous blends. polymer modified bitumens and RAP aggregates are used. Different
methods based on direct evaluation of mixture workability should
be implemented for warm productions.
Table 4 Although small differences characterized the various WMA
Compactability, volumetric properties and ITS: (a) average results of all laboratory mixtures, CEI values as well as air voids suggest that the chemical
compacted samples (diameter: 150 mm); (b) ANOVA results.
additive (M_WC) is able to ensure the most promising workability
(a) compared to the control mixture. Such a result can be related to the
Mixture CEI [-] Air voids Vm [%] VMA [%] VFB [%] ITS [MPa] enhanced interactions at the aggregate-bitumen interface devel-
oped thanks to the surfactants contained within the chemical
M_H 139 4.00 14.36 74.71 1.52
M_W 593 5.20 17.37 59.55 1.00 additive.
M_WC 197 4.40 14.7 72.2 1.55
M_WO 260 4.80 15.79 69.18 1.43 3.2. Stiffness
M_WZ 246 4.70 14.32 75.11 1.69

(b) The analysis of the master curves of the complex modulus norm
Compared CEI Air voids Vm ITS for both bitumens and mixtures allowed a comprehensive evalua-
materials tion of materials in terms of stiffness. In both cases, frequency
p- significant? p- significant? p- significant?
value value value sweep tests were performed in a wide range of temperature and
frequency for verifying the analytical model assumed for the
M_H/M_W 2.71E- YES 4.82E- YES 5.78E- YES
12 03 06
master curve determination at a reference temperature of 30
C.
M_H/M_WC 0.94 NO 0.74 NO 0.99 NO Due to the presence of significant amount of aged bitumen as well
M_H/M_WO 0.10 NO 0.12 NO 0.26 NO as high content of elastomeric polymers, the investigated materials
M_H/M_WZ 0.77 NO 0.46 NO 0.57 NO are not thermo-rheologically simple. Therefore, the Partial Time-
M_WC/M_WO 0.79 NO 0.78 NO 0.35 NO
Temperature Superposition Principle as defined by Di Benedetto
M_WC/M_WZ 0.94 NO 0.84 NO 0.12 NO
M_WZ/M_WO 0.92 NO 0.88 NO 0.04 YES et al (Di Benedetto et al., 2011) was exploited to build the master
curves: for bitumens, the modified CAM (Christansen, Anderson
and Marasteanu) model was assumed for the master curve of jG*j
(Christensen and Anderson, 1992), whereas the four-parameters
3%  Vm  5%). VMA and VFB values showed a trend consistent
sigmoidal function developed by Pellinen et al. (2004) was
with air voids, confirming the feasibility of WMA additives for
applied for mixtures complex modulus jE*j. In both cases, free
producing warm mixtures without compromising volumetric
variation of shift factors was allowed for achieving model
properties.
optimization.
The goodness of WMA additives was further strengthened by
Within the temperature/frequency range investigated, the
Indirect Tensile test results expressed in Table 4a in terms of Indi-
analysis of bitumen master curves confirmed viscosity results: no
rect Tensile Strength (ITS) which is usually adopted, along with
significant difference exists between warm bitumens and the
volumetric properties, as main control and acceptance parameter
modified bitumen H without any additive (Fig. 2a). Therefore, based
for motorway binder layers. ITS showed that warm mixtures ach-
on the collected results, the rheological properties of the bitumi-
ieved mechanical properties similar to the control mixture (M_H),
nous phase was not altered by adding WMA additives. Only the
with a trend strictly related to the air voids content, as also
bitumen blended with organic additive (B_WO) appeared slightly
demonstrated by the mixture M_W characterized by considerably
stiffer compared with the other warm bitumens according to the
lower performance in accordance with the higher Vm. The signifi-
crystallization of the molecular structure of bitumen when the
cance of above considerations is corroborated by ANOVA results
material cools. As the temperature drops down the melting point of
summarized in terms of p-value in Table 4b for the main perfor-
the organic wax, the additive solidifies into microscopic uniformly
mance parameters.
distributed particles, which can increase binder stiffness in a way
A further aspect that surely provided a fundamental contribu-
comparable to fiber-reinforced materials (Rubio et al., 2012).
tion in terms of workability is the less relevant oxidation under-
On the contrary, the analysis of asphalt mixtures master curves
gone by warm mixtures thanks to lower production temperatures,
(Fig. 2b) highlighted a sharp difference between control and all
which determine lower stiffness of the overall bituminous phase, as
warm mixtures, with significant higher stiffness for the former.
shown in details in the next paragraph.
In order to interpret overall results, it is useful to bear in mind
Moreover, the interpretation of the good results achieved in
that bitumens used for performing frequency sweep tests where
terms of workability for WMA mixtures should also consider the
unaged whereas mixtures were prepared at the selected
916 A. Stimilli et al. / Journal of Cleaner Production 156 (2017) 911e922

1,E+07 WMA additive type, as shown in the Black Diagrams of Fig. 3 in

B_H Tref = 30°C
terms of complex modulus norm and phase angle. A unique
B_WC
continuous line describes all measured data preventing a clear
1,E+05 B_WO
distinction between mixtures. Therefore, the adding of WMA ad-
B_WZ
ditives did not cause variations in the proportion between elastic
and dissipative component of the bituminous blends investigated.
1,E+03 This finding, combined with the lower stiffness of WMA mix-
|G*| [kPa]

tures, is particularly important in the light of cracking aptitude,

hereafter discussed, and further confirms the possibility to produce
1,E+01 mixtures at reduced temperatures without penalizing the material
response under traffic loadings.
Finally, it is worth noting that the temperature susceptibility of
1,E-01 the mixture prepared with the organic additive (M_WO) was much
more remarkable than all other mixtures, as demonstrated by
higher glassy modulus (Eg) and lower rubbery modulus (Er)
1,E-03
1,E-08 1,E-06 1,E-04 1,E-02 1,E+00 1,E+02 1,E+04 1,E+06 1,E+08
calculated through the model adopted for the master curve
(Fig. 2b). Such a behavior is consistent with the crystallization
Reduced frequency [Hz]
phenomena that occurs when organic waxes are involved in the
(a)
production process of warm asphalt mixtures. In fact, at low tem-
1.E+05
M_H perature (i.e. high frequency) the molecular crystallization causes
Tref = 30°C
M_W an increase in material stiffness. On the contrary, once the tem-
M_WC perature approaches the melting point of the additive (i.e. low
1.E+04
M_WO frequency) the mixture loses its consistency with a consequent
M_WZ

1.E+03
|E*| [kPa]

1,E+06

1,E+05
1.E+02
Mixture Er [MPa] Eg [MPa] 1,E+04 B_H
M_H 34.4 33618.7 B_WC
1.E+01 M_W 15.8 25041.0 1,E+03 B_WO
M_WC 19.7 26926.5
B_WZ
|G*| [MPa]

M_WO 6.8 34397.7 1,E+02

M_WZ 40.9 24555.9
1.E+00 1,E+01
1.E-08 1.E-06 1.E-04 1.E-02 1.E+00 1.E+02 1.E+04 1.E+06 1.E+08

Reduced frequency [Hz] 1,E+00

(b)
1,E-01
Fig. 2. Complex modulus master curves at 30
C: (a) bitumens, (b) mixtures.
1,E-02

1,E-03
production temperature (different for the control HMA and warm 1,E+01 1,E+02
mixtures). Thus, results suggests that the different short-term δ [°]
oxidation undergone by the mixtures is the major aspect respon-
sible for stiffness properties, whereas the adding of WMA additives (a)
1,E+05
did not provide any variations in the bitumen rheological behavior M_H
and, hence, any contributions in terms of stiffness. As expected,
M_W
lower production temperatures applied to warm mixtures induced
M_WC
less oxidation and, consequently, a stiffening effect due to aging
M_WO
significantly reduced compared to HMA. However, consistently
1,E+04 M_WZ
with bitumen rheological findings, the only material with slight
higher stiffness among WMA mixtures was the one prepared with
|E*| [MPa]

organic additive (M_WO), confirming that crystallization mecha-

nisms typical of these products lead to an increase in stiffness.
The major importance of aging rather than the presence of
1,E+03
WMA additives on the development of stiffness properties is
further confirmed by the mixture prepared at reduced tempera-
tures without any additive (M_W). Despite the absence of WMA
additives, M_W mixture exhibited stiffness values comparable with
the other warm mixtures (all characterized by similar air voids
1,E+02
ranging around 4% and 5%). 1,E+00 1,E+01 1,E+02
Although the remarkable difference in terms of stiffness recor- δ [°]
ded between warm mixtures and the control mixture, no signifi-
(b)
cant gap appeared when the elastic behavior was analyzed. Such a
result is detectable for both bitumens and mixtures regardless of Fig. 3. Black Diagram: (a) bitumens, (b) mixtures.
 A. Stimilli et al. / Journal of Cleaner Production 156 (2017) 911e922
reduction in the overall stiffness properties and negative effects in
terms of performance.
Stiffness properties were also evaluated in terms of Indirect
Tensile Stiffness Modulus (ITSM) at 20
C. Results obtained were
consistent with what found through the uniaxial compression test
and further confirmed what previously observed. This similar trend
is shown in Fig. 4, which relates ITSM values and complex moduli
determined through the master curves at the same temperature
(20
C) and a frequency of 2 Hz (comparable with the ITSM test rise-
time of 124 ms). ANOVA results are also reported in Fig. 4 in terms
of p-values to support what previously discussed.
Overall findings detected in terms of stiffness for WMA mixtures
are particularly interesting considering the presence of RAP: the
mitigation of stiffness due to the adoption of WMA technologies
represents an important element to propose the increase of RAP
within the mixture avoiding brittleness risks that could compro-
mise overall material performance. In this sense, the choice of the
most appropriate additive is fundamental to develop satisfactory
stiffness properties. At the same time, the different behavior
detected between warm and hot mixtures suggests that, in the
pavement design process, stiffness moduli of warm recycled mix-
tures should not be estimated through prediction models available
in literature for bituminous mixes. A specific laboratory investiga-
tion is necessary for warm mixtures, considering also the effects of
compaction and lay-down conditions as well as additional variables
that significantly affect pavement service life (e.g. pavement
structure, climate conditions, traffic).
The remarkable effect in terms of stiffness detected between
warm mixtures and the control mixture due to different production
temperatures (well demonstrated by ANOVA results summarized in
Fig. 4) can also be identified comparing unaged and short-term
aged bitumens. Oscillatory loading tests were carried out through
the DSR on unaged and short-term aged bitumens (x2.2) with and
without WMA additives. It is worth remembering that the short-
term aging was conducted by means of RTFOT at different tem-
peratures (i.e. 120
C for warm mixtures and 163
C for the control
mixture) to properly simulate the authentic oxidation effects
experienced by mixtures during transportation and lay-down. G*/
sind was used as main indicator for evaluating the behavior of the
investigated materials at high service temperatures. In particular,
the percent increases D(jG*j/sind), DjG*j and Dd, all calculated
comparing unaged and short-term aged values as
D ¼ ðRTFOT  unagedÞ=unaged, are expressed in Table 5 and Fig. 5
as percentage average of the data recorded at each testing
temperature.
The increase measured for bitumen H (without WMA additives
16000
T = 20°C
14000
12000
10000
|E*| [MPa]
8000
6000
4000
2000
0 0
M_H M_W M_WC M_WO
Fig. 4. Mixture stiffness at 20
C: ITSM Vs jE*j @ 2 Hz and ANOVA results.
917
and short-term aged at 163
C) was extremely higher (156.2%) than
that one calculated for warm bitumens (lower than 50%) which
were short-term aged at 120
C, confirming the same trend found
for mixtures stiffness. In particular, the variation of the parameter
G*/sind was mainly attributed to a stiffness increase (DjG*j) rather
than a change in the phase angle Dd as also demonstrated by the
statistical significance detected through ANOVA analysis carried
out for each parameter (Fig. 5). Such a result further confirms the
findings previously highlighted for mixtures (i.e. irrelevant differ-
ences in phase angle for all investigated materials). Moreover, the
absolute values measured for jG*j/sind ratify the tendency of the
organic additive to provide a noticeable increase in stiffness (before
and after aging at each testing temperature). This finding was
predictable based on the crystallization mechanisms typical of
organic waxes as previously discussed.
3.3. Fracture resistance
To evaluate the fracture resistance of the investigated materials,
Semi-Circular Bending (SCB) tests and Indirect Tensile Fatigue Tests
(ITFT) were performed at a temperature of 10
C and 20
C,
respectively. Moreover, since the bituminous phase of asphalt
mixtures is the main responsible for fatigue mechanisms, oscilla-
tory loading tests were run on long-term aged bitumen samples by
means of the DSR at 20
C with the aim to gain a better under-
standing of mixtures behavior and determine possible correlations
between mixture and bitumen data. For this purpose, the param-
eter G*$sind (representative of the dissipated energy during loading
cycles) was monitored as directly related to the material fatigue
resistance limit.
Fig. 6 reports SCB test results in terms of fracture toughness K
and fracture energy G (as average of four replicates) as proposed by
the fracture mechanics approach developed by Huang et al. (2013).
K expresses the cracking initiation resistance of the material,
whereas G is more related to the cracking propagation resistance
once the crack is already formed (representative of material
ductility). To better interpret test results, Fig. 6 summarizes also p-
values obtained through the ANOVA analysis.
It is possible to note that producing at lower temperatures
caused a decreased in the cracking initiation resistance as
demonstrated by lower fracture toughness K of all warm mixtures
compared to the control mixture. Such a behavior was expected,
since the more severe aging process undergone by the control
mixture was believed to determine an initially stiffer material.
Moreover, reduced production temperatures can induce lower RAP
bitumen reactivation and lower adhesion at the bitumen-aggregate
16000 Compared |E*| and ITSM
|E*|@2Hz materials p-value significant?
ITSM 14000 M_H / M_W 3.25E-04 YES
M_H / M_WC 3.96E-04 YES
12000 M_H / M_WO 0.02 YES
M_H / M_WZ 2.91E-04 YES
ITSM [MPa]
10000
M_WC / M_WO 4.09E-03 YES
8000 M_WC / M_WZ 0.34 NO
M_WZ / M_WO 2.53E-03 YES
6000 1
4000
2000
M_WZ
918 A. Stimilli et al. / Journal of Cleaner Production 156 (2017) 911e922

Table 5
Oscillatory test results.

Bitumen Aging jG*j/sind [MPa] Average D(jG*j/sind) [%]

T ¼ 52
C T ¼ 58
C T ¼ 64
C T ¼ 70
C T ¼ 76
C T ¼ 82
C

B_H Unaged 79585 46673 27335 15945 9109 5131 156.2

RTFOT 269150 94548 67871 44378 22490 11430
B_WC Unaged 19645 11198 6688 4164 2660 1660 46.6
RTFOT 31540 18020 10395 6050 3547 2059
B_WO Unaged 72545 41600 22285 11853 6649 3756 43.1
RTFOT 102350 60628 32275 17088 9346 5335
B_WZ Unaged 23185 13195 7570 4371 2560 1515 40.2
RTFOT 33505 18053 9893 6598 3551 2115

Δ (|G*|/sinδ) Δ |G*| Δδ
200
185.2
Compared Δ (|G*|/sinδ) Δ |G*| Δδ
materials p-value significant? p-value significant? p-value significant?
156.2
B_H / B_WC 1.40E-08 YES 3.50E-09 YES 0.05 YES/NO
150 B_H / B_WO 4.43E-08 YES 8.42E-09 YES 0.07 NO
B_H / B_WZ 1.29E-08 YES 5.29E-09 YES 0.06 NO
B_WC / B_WO 0.84 NO 0.35 NO 0.79 NO
B_WC / B_WZ 0.43 NO 0.76 NO 0.64 NO
Δ [%]

B_WZ / B_WO 0.61 NO 0.51 NO 0.98 NO

100

50 46.6 47.2 43.1 40.2 40.1

35.7

16.6
4.8 7.9 8.1
0
B_H B_WC B_WO B_WZ

Fig. 5. Variation of stiffness/phase angle between short-term and unaged bitumen with ANOVA results.

3 35 Compared Fracture Energy G Fracture Toughness K

T = 10°C Fracture energy materials p-value significant? p-value significant?
Fracture toughness M_H / M_W 0.34 NO 3.25E-04 YES
Fracture toughness K [N/mm1,5]

30 M_H / M_WC 0.01 YES 0.32 NO

Fracture Energy G [kJ/m2]

M_H / M_WO 4.40E-05 YES 1.15E-04 YES

25 M_H / M_WZ 0.25 NO 4.89E-03 YES
2 M_WC / M_WO 2.41E-04 YES 4.72E-04 YES
20 M_WC / M_WZ 0.04 YES 0.01 YES
M_WZ / M_WO 4.53E-03 YES 0.02 YES
15
1
10

0 0
B_H B_W B_WC B_WO B_WZ

Fig. 6. SCB results at 10
C and ANOVA results.

interface, causing a decrease in the cracking initiation resistance of most brittle behavior (lowest G value) along with the lowest
warm mixtures. Such a tendency was even worsened by the pres- cracking initiation resistance K, consistently with stiffness results
ence of organic additive and zeolite, whereas the use of the and with the crystallization mechanisms typical of these products.
chemical additive allowed a performance compensation likely due The mixture containing zeolite exhibited an intermediate behavior.
to the surfactants, which promote higher adhesion at the bitumen- Indirect Tensile Fatigue Test (ITFT) results confirm SCB test an-
aggregate interface and determine better interactions and a more alyses. The mixture with the chemical additive showed the most
uniform stress distribution within the mixture. This assertion is promising behavior in terms of fatigue resistance (Fig. 7) as
confirmed by the highest energy fracture G showed by the warm demonstrated by the lowest value of the exponent (B) of the power
mixture with the chemical additive (M_WC). At the same time, the law regression adopted to describe test data in a graph which re-
mixture prepared with the organic additive (M_WO) exhibited the lates the maximum horizontal strain εx and the number of loading
 A. Stimilli et al. / Journal of Cleaner Production 156 (2017) 911e922 919

cycles at failure N (i.e. fatigue curve: εx ¼ A$N B ). The term A is in 2.5

the intercept with the vertical axis, whereas the coefficient B rep-
M_WC

Fracture Energy G [kJ/m2]

resents the slope of the regression line: the higher the slope, the 2.0 M_W M_H
lower the fracture resistance. Again, the mixture with the organic
additive had the worst performance (highest B value), whereas M_WZ
1.5
zeolite ensured intermediate resistance.
To better interpret test results, it is interesting to compare the
M_WO
control mixture and the mixture prepared at reduced temperature 1.0
without WMA additives: the latter was characterized by good fa-
tigue resistance, higher than the control mixture despite the 0.5
absence of WMA additives. This finding demonstrates that lower y = -9.51x + 3.48
R² = 0.97
production temperatures reduce oxidation effects, guaranteeing 0.0
improved performance compared to HMA. However, results also 0.10 0.12 0.14 0.16 0.18 0.20 0.22 0.24 0.26
suggest that the WMA additive type significantly alters mixture Fatigue curve power law coefficient B [-]
behavior, confirming that the choice of the most appropriate ad-
Fig. 8. SCB results Vs ITFT results.
ditive is fundamental for developing adequate performance.
The optimum correlation between SCB and ITF test results is
shown in Fig. 8. Notwithstanding different test temperatures (10
fatigue curves (B) supports the feasibility of the mechanical char-
and 20
C for SCB and ITFT tests, respectively), the cracking pro-
acterization carried out.
pensity detected with SCB tests at 10
C is verified also through
A further confirmation about the reliability of the fracture
fatigue tests at 20
C. The high value of the regression coefficient R2
resistance analysis came from bitumen test results. jG*j$sind values
in the graph which relates fracture energy values with the expo-
(parameter usually adopted for fatigue analyses as related to the
nent of the power law regression lines adopted for determining
dissipated energy at each loading cycle) was measured on long-

500 500
T = 20°C M_H T = 20°C M_H
M_W M_WC
εx [μstrain]

εx [μstrain]

y = 539.7x-0.138
y = 612.7x-0.150 R² = 0.99
R² = 0.93

y = 885.2x-0.203 y = 885.2x-0.203
R² = 0.88 R² = 0.88
50 50
500 5000 50000 500000 500 5000 50000 500000
Number of loading cycles Number of loading cycles
500 500
T = 20°C M_H T = 20°C M_H
M_WO M_WZ
εx [μstrain]

εx [μstrain]

y = 1 488.5x-0.247
R² = 0.96 y = 851.25x-0.189
R² = 0.74

y = 885.2x-0.203 y = 885.2x-0.203
R² = 0.88 R² = 0.88
50 50
500 5000 50000 500000 500 5000 50000 500000
Number of loading cycles Number of loading cycles
Fig. 7. ITFT results at 10
C.
920 A. Stimilli et al. / Journal of Cleaner Production 156 (2017) 911e922
term aged bitumens at 20
C subjected to oscillatory loading tests.
Relating these values to the maximum horizontal deformation level
calculated for each mixture through the fatigue curve of Fig. 7 for a
conventional number of loading cycle equal to 106, a good corre-
lation was found as depicted in Fig. 9. When jG*j$sind increases
(higher dissipated energy) the strain level to ensure a resistance of
106 cycles decreases due to less ability to resist the applied loading.
Such an observation demonstrates the importance of the bitu-
minous component for developing good fracture resistance that
can be adequately predicted by studying bitumen behavior. More-
over, it further confirms that the different aging degree undergone
by conventional HMAs and warm mixtures determines differences
in terms of rheological properties, which in turn affect overall
mixture response. When long-term aged materials are considered,
the effect of WMA additives is much more significant than on
unaged bitumens.
3.4. Environmental impact
To emphasize the positive environmental impact of warm
recycling techniques, the present paragraph provides information
about the potential energy savings and emissions reduction related
to an extensive use of WMA technologies in road construction. For a
complete evaluation of environmental effects, the overall life-cycle
of each single mixture component accounting for corresponding
direct and indirect costs and emissions (e.g. extraction, production,
transportation, milling, final disposal) should be considered.
However, comparing warm and hot productions at equal boundary
conditions (e.g. aggregate and binder type, plant and construction-
site distance, climate) can provide a meaningful estimate of the
potential impact of warm mixtures solely due to the difference in
working temperatures. The only additional energy required for
warm mixtures would be related to warm additives manufacturing
and transportation. However, based on previous specific studies
which analyzed the overall environmental impact associated to
WMA mixtures (Leon and Jensen, 2012; Mazumder et al., 2016;
Zaumanis et al., 2012), this contribution can be neglected since it
is minimal compared to energy variations caused by working
temperatures reduction. A more detailed methodological approach
could be followed for a complete Life Cycle Analysis (LCA).
In this case, a simplified approach was adopted to estimate the
energy saving related to mixture production (Zaumanis et al.,
2012). The energy Q required to heat a certain amount of aggre-
gates (m) was calculated as follows:
Q ¼ c$m$DT
90
80 M/B_WC
[microstrain]
70
M/B_WZ
60
M/B_H
50
ε @ Nf = 106
40
30
20
10
0
0 500 1000 1500
|G*|∙sinδ [MPa]
Fig. 9. ITFT results of mixtures Vs oscillatory loading test results of bitumens and ANOVA results in terms of jG*j$sind.
where c is the specific heat measure in J/kg$K (assumed equal to
920 J/kg$K) and DT the variation from ambient to the desired
working temperature (105 and 145
C for warm and hot mixtures,
respectively). The bitumen was not considered for calculating po-
tential energy savings since it is kept at the same temperature for
both warm and hot mixtures. Assuming an average annual pro-
duction of asphalt concrete equal to 25 million tons (annual
quantity in Italy for maintenance activities), Q is 3.34$1015 J for hot
mixtures and 2.42$1015 J for warm mixtures, which means an en-
ergy saving around 27.5%. This contribution is only related to the
energy required for producing the mixture. To quantify overall
positive effects in terms of environmental impact, pollutants
emitted during and after in-plant productions should also be
accounted for.
In this sense, a monitoring of air emissions was undertaken
during large-scale in-plant productions of both warm and hot
mixtures corresponding to the same materials investigated in this
study (Stimilli et al., 2017). Measurements were performed at equal
boundary conditions (e.g. wind intensity and direction, air tem-
perature, humidity, asphalt plant configuration) for effectively
isolating the impact of working temperatures. Through an in-situ
Continuous Emissions Monitoring System (composed of a sample
probe, a filter, a sample line, a gas conditioning system, a calibration
gas system, and a series of gas analyzers) and a pollutant concen-
tration monitor apparatus (based on a “hot dry” extractive method)
gases and dusts emissions were measured (i.e. dusts, oxygen, car-
bon monoxide, nitrogen oxide, sulfur oxide, Volatile Organic
Compound, Polycyclic Aromatic Hydrocarbon). After a specific
conditioning process, the air sample enters a sampling manifold
where individual gas analyzers (which reflect the parameters being
monitored) extract each investigated component and measure the
corresponding concentration by means of various techniques (e.g.
infrared and ultraviolet adsorption, chemiluminescence, fluores-
cence, chromatography and spectrography). The monitoring of
emissions was extended for several hours to guarantee that the
asphalt plant was working at full capacity. For each detection, the
measurement point was chosen in correspondence of the stack
connected with the aggregate drying drum in order to identify the
most significant emission rate. Table 6 summarizes the main pa-
rameters considered relevant in terms of environmental impact
(results are expressed as average of three independent detections).
All data are expressed in terms of concentration of each sub-
stance within the investigated gas medium and are corrected in
relation to the reference oxygen concentration prescribed by the
technical standards as follows:
[1]
Compared |G*|·sinδ
materials p-value significant?
B_H / B_WC 0.02 YES
B_H / B_WO 7.46E-03 YES
B_H / B_WZ 0.03 YES
B_WC / B_WO 4.84E-04 YES
B_WC / B_WZ 0.34 NO
B_WZ / B_WO 5.61E-04 YES
M/B_WO
2000 2500 3000 3500
 A. Stimilli et al. / Journal of Cleaner Production 156 (2017) 911e922
Table 6
Pollutant emissions expressed as corrected concentration E.
Parameters Standard
Dusts UNI EN 13284
Carbon monoxide (CO) Italian DM 25.8.2000
Nitrogen oxide (NOx) Italian DM 25.8.2000
Sulfur oxide (SOx) Italian DM 25.8.2000
Volatile Organic Compound (VOCs) UNI EN 12619
Polycy. Aromatic Hydrocar. (PAHs) Italian DM 25.8.2000
Temperature at the stack [
C] e e
21  O
E¼ $EM
21  OM
where E is the corrected concentration, EM the measured concen-
tration, O the reference oxygen concentration and OM the
measured oxygen concentration.
The average values measured for warm productions were
generally lower than those recorded for HMA mixtures. As direct
consequence of lower temperatures at the fumes stack during
WMA productions, a significant reduction (around 18%) in the ni-
trogen oxide (NOx) concentration was detected in case of WMA
productions (i.e. 52.4 mg/Nmc) with respect to HMA productions
(i.e. 61.8 mg/Nmc). Such a finding shows a direct correspondence
between NOx concentration and in-plant operational temperatures.
The sulfur oxide (SOx) amount was also lower for WMA mixtures
than for the reference HMA. Again, the temperature achieved
during the combustion process is the main parameter to control
SOx emissions since the other two main variables that affect SOx
(i.e. content of volatile sulfur in raw materials and type of fuel) were
equal for all mixtures investigated. As far as the VOC concentration
is concerned, the same positive trend observed for warm pro-
ductions in terms of NOx and SOx is confirmed (VOC reduction
around 23%). Finally, it is worth noting that the monitoring was
performed only at the asphalt plant during mixtures production
and did not involve lay-down and compaction phases, aspect that
could further improve environmental benefits especially in terms
of operators' health.
Such findings demonstrate the potential of warm technologies
for reducing pollutant emissions concurrently providing significant
energy savings. The overall positive impact in terms of environ-
mental sustainability also corresponds to significant cost savings,
which improve even more the advantages of in-plant productions
at reduced working temperatures.
4. Conclusions
The effectiveness of WMA technologies for producing recycled
dense-graded bituminous mixtures containing polymer modified
bitumens was evaluated by adopting various WMA additives. Both
mechanical tests on mixtures and rheological analyses on bitumens
were performed to determine the main performance required
against typical flexible pavement distresses. Compactability and
volumetric properties, stiffness characteristics and cracking resis-
tance were evaluated leading to the following main conclusions:
- although WMA additives did not cause any bitumen viscosity
reduction, they provided good workability compared to the
reference mixture, despite the remarkable gap in production
temperatures. Rather than bitumen viscosity changes, better
interactions between bitumen and aggregate particles as well as
reduced oxidation phenomena undergone during production
921
Standard limit [mg/Nmc] Production temperatures
130
C 170
C
20 11.5 11.8
e 1081.0 1021.0
500 52.4 61.8
1700 113.2 129.0
e 8.4 10.4
0.1 <0.067 <0.067
92.7 107.0
and laydown and potential lower reactivation degree of RAP
bitumen (due to reduced mixing and compaction temperature)
[2]
simultaneously concurred to determine such a result.
- The equiviscous principle employed to calculate working tem-
peratures seems not suitable for warm recycled mixtures. The
analysis of bitumen viscosity is not sufficient to properly eval-
uate overall workability of WMA mixtures, especially when
polymer modified bitumens and RAP aggregates are used.
Different methods based on direct evaluation of mixture wok-
ability should be implemented for warm productions.
- WMA mixtures complied with all acceptance requirements
prescribed by technical standards for motorway binder mixes
(i.e. air void content and Indirect Tensile Strength).
- Adding WMA additives did not cause changes on rheological
properties of unaged bitumens. On the contrary, short and long-
term aged bitumens demonstrated different behavior depend-
ing on the type of WMA additive. The organic additive showed
the worst performance (in terms of stiffness and cracking apti-
tude), whereas surfactants contained within the chemical ad-
ditive promoted better adhesion between aggregate particles
and bitumen compensating reduced production temperatures
and lower RAP bitumen reactivation.
- Mixture stiffness was considerably lower in case of WMA ma-
terials thanks to the reduced short-term oxidation process. Only
the organic additive achieved higher stiffness than the other
warm mixtures due to crystallization mechanisms occurring
within the molecular structure. The mixture prepared with the
organic additive was also characterized by the highest temper-
ature susceptibility. No change in the elastic component of all
WMA mixtures was recognized.
- Warm mixtures prepared with the chemical additive guaran-
teed improved performance in terms of cracking resistance and
ductility even compared to the control mixture: surfactants
enhance adhesion at the bitumen-aggregate interface, so
enabling a more uniform distribution of external loads. The
good agreement between results collected through various
mechanical tests demonstrated the feasibility of the laboratory
characterization carried out.
- Good accordance was also found between bitumens and mix-
tures in terms of fracture resistance, demonstrating the possi-
bility to adequately predict mixture fatigue properties by
measuring the dissipated energy on long-term aged bitumens.
- According to a basic prediction model applied for quantifying
energy consumptions and the direct monitoring of air emissions
at the asphalt plant, it was demonstrated that a temperature
reduction up to 40
C can successfully reduce pollutants as well
as induce energy savings.
In summary, overall results indicate the possibility to produce
adequate dense-graded mixtures at significant reduced tempera-
tures, concurrently suggesting that the choice of the most
922 A. Stimilli et al. / Journal of Cleaner Production 156 (2017) 911e922
appropriate WMA additive is fundamental for the development of
satisfactory performance. The analysis indicates that the use of
chemical additives including surfactants and adhesion enhancers
can result the most proper product to balance potential drawbacks
deriving by reduced production temperatures in case of recycled
mixtures with polymer modified bitumens.
The findings detected for WMA mixtures are particularly
interesting considering the presence of RAP within the materials
investigated. The satisfactory performance found with the adoption
of suitable WMA technologies can represent an important element
to propose a further increase in RAP within mixtures allowing an
even more sensible reduction of pollutant emissions and energy
savings.
In this regard, future studies should also deepen the analysis of
the degree of bitumen released by RAP aggregates because it could
likely be lower in case of warm productions. Concurrently, in plant
production of WMA mixtures and full scale experimental sections
are already scheduled in order to validate laboratory results, find
suitable correlations between laboratory and in-situ measurements
and determine possible operational issues of the asphalt plant
related to reduced mixing temperatures.
Acknowledgements
This research was sponsored by Pavimental SpA. This support is
greatly acknowledged. Data analysis and opinions are those of the
authors and do not necessarily reflect those of the sponsoring
agency.
References
Bonaquist, R., 2011. Mix Design Practices for Warm Mix Asphalt. Transportation
Research Board. NCHRP Report 691.
Brock, J.D., Richmond Sr., J.L., 2007. Milling and Recycling, Technical Paper T-127.
Christensen, D.W., Anderson, D.A., 1992. Interpretation of dynamic mechanical test
fata for paving grade asphalt cements. Assoc. Asph. Pav Technol. 61, 67e116.
Di Benedetto, H., Sauzeat, C., Bilodeau, K., Buannic, M., Mangiafico, S., Nguyen, Q.T.,
Pouget, S., Tapsoba, N., Van Rompu, J., 2011. General overview of the time-
temperature superposition principle validity for materials containing bitumi-
nous binder. Int. J. Roads Airp 1 (1), 35e52.
Dinis-Almeida, M., Lopes Afonso, M., 2015. Warm mix recycled asphalt e a sus-
tainable solution. J. Clean. Prod. 107, 310e316.
D'Angelo, J., Harm, E., Bartoszek, J., Baumgardner, G., Corrigan, M., Cowsert, J.,
Harman, T., Jamshidi, M., Jones, W., Newcomb, D., Prowell, B., Sines, R.,
Yeaton, B., 2008. Warm-mix Asphalt: European Practice. Report FHWA/PL-08/
007.
Frigio, F., Canestrari, F., 2016. Characterisation of warm recycled porous asphalt
mixtures prepared with different WMA additives. Eur. J. Environ. Civ. Eng.
http://dx.doi.org/10.1080/19648189.2016.1179680 published online.
Frigio, F., Stimilli, A., Bocci, M., Canestrari, F., 2016. Adhesion properties of warm
recycled mixtures produced with different WMA additives. In: Proceedings of
the 4th CEW on “Functional Pavement Design”.
Huang, B., Shu, X., Zuo, G., 2013. Using notched semi circular bending fatigue test to
characterize fracture resistance of asphalt mixtures. Eng. Fract. Mech. 109,
78e88.
Kakar, M.R., Hamzah, M.O., Valentin, J., 2015. A review on moisture damages of hot
and warm mix asphalt and related investigations. J. Clean. Prod. 99, 39e58.
Kanitpong, K., Sonthong, S., Nam, K., Martono, W., Bahia, H.U., 2007. Laboratory
study on warm mix asphalt additives. In: Transportation Research Board 86th
Annual Meeting.
Kim, H., Lee, S.-J., Amirkhanian, S.N., 2012. Influence of warm mix additives on PMA
mixture properties. J. Transp. Eng. 138 (8), 991e997.
Lee, S.J., Amirkhanian, S.N., Park, N.W., Kim, K.W., 2009. Characterization of warm
mix asphalt binders containing artificially long-term aged binders. Const. Build.
Mater 23 (6), 2371e2379.
Leon, J.G., Jensen, P.H., 2012. Environmental aspects of warm mix asphalts produced
with chemical additives. In: Proceedings of the 5th Eurasphalt & Eurobitume
Congress, Istanbul, 13e15th June.
Lo Presti, D., Jime nez del Barco Carrio n, A., Airey, G., Hajj, E., 2016. Towards 100%
recycling of reclaimed asphalt in road surface courses: binder design method-
ology and case studies. J. Clean. Prod. 131, 43e51.
Mahmoud, A.F.F., Bahia, H.U., 2004. Using the gyratory compactor to measure
mechanical stability of asphalt mixtures. In: Wisconsin Highway Research
Program 05e02.
Mallick, R.B., Kandhal, P.S., Bradbury, R.L., 2008. Using warm-mix asphalt technol-
ogy to incorporate high percentage of reclaimed asphalt pavement material in
asphalt mixtures. Transp. Res. Rec. 2051, 71e79.
Mazumder, M., Sriraman, V., Kim, H.H., Lee, S.-J., 2016. Quantifying the environ-
mental burdens of the hot mix asphalt (HMA) pavements and the production of
warm mix asphalt (WMA). Int. J. Pavement Res. Technol. 9, 190e201.
Miliutenko, S., Bjo € rklund, A., Carlsson, A., 2013. Opportunities for environmentally
improved asphalt recycling: the example of Sweden. J. Clean. Prod. 43, 156e165.
Pellinen, T.K., Witczak, M.W., Bonaquist, R., 2004. Asphalt mix master curve con-
struction using sigmoidal fitting function with non-linear least squares opti-
mization. In: Recent Advances in Material Characterization and Modeling of
Pavement Systems. Geotechnical Special Publication 123.
Rubio, M.C., Martìnez, G., Baena, L., Moreno, F., 2012. Warm mix asphalt: an over-
view. J. Clean. Prod. 24, 76e84.
Stimilli, A., Virgili, A., Canestrari, F., 2015. New method to estimate the “re-acti-
vated” binder amount in recycled hot-mix asphalt. RMPD 16, 442e459.
Stimilli, A., Frigio, F., Sciolette, S., Canestrari, F., 2017. In-plant production of warm
recycled mixtures produced with SBS modified bitumen: a case study. In:
Transport Infrastructure and Systems proceedingss of the Aiit International
Congress on Transport Infrastructure and Systems (Tis, Rome 2017),
pp. 143e152. http://dx.doi.org/10.1201/9781315281896-21. Print ISBN: 978-1-
138-03009-1.
Tao, M., Mallick, R.B., 2009. Effects of warm-mix asphalt additives on workability
and mechanical properties of reclaimed asphalt pavement material. Transp. Res.
Rec. 2126, 151e160.
Zaumanis, M., Jansen, J., Haritovonos, V., Smirnovs, J., 2012. Development of
calculation tool for assessing the energy demand of Warm Mix Asphalt. Proc.
Soc. Behav. Sci. 48, 163e172.
Zeng, M., Wu, C., Zhang, J., 2014. Determining mixing and compaction temperatures
of Evotherm® warm mix asphalt using 100% reclaimed asphalt pavement. Int. J.
Pavement Res. Technol. 7 (6), 389e396.
Zhao, S., Huang, B., Shu, X., Woods, M., 2013. Comparative evaluation of warm mix
asphalt containing high percentages of reclaimed asphalt pavement. Const.
Build. Mater 44, 92e100.