Construction and Building Materials 185 (2018) 684–696

Contents lists available at ScienceDirect

Construction and Building Materials

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

Influence of recycled concrete and steel slag aggregates on warm-mix

asphalt properties
F.C.G. Martinho a, L.G. Picado-Santos b, S.D. Capitão c,b,⇑
a
FM Consult & Instituto Superior Técnico, Universidade de Lisboa, Av. Rovisco Pais, 1049-001 Lisboa, Portugal
b
CESUR, CERIS, Instituto Superior Técnico, Universidade de Lisboa, Av. Rovisco Pais, 1049-001 Lisboa, Portugal
c
Instituto Politécnico de Coimbra, Instituto Superior de Engenharia de Coimbra, Rua Pedro Nunes, 3030-199 Coimbra, Portugal

h i g h l i g h t s

WMA blends with a chemical additive showed low resistance to permanent deformation.

The WMA blend with 60% of RCA and organic wax was the best performant material.

Adding 30% EAFS or 60% RCA did not notably influence mechanical performance of WMA.

Fatigue resistance and stiffness modulus of WMA with RCA or EAFS were satisfactory.

Incorporating EASF or RCA in WMA was not a problem vis-a-vis water sensitivity.

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

Article history: This paper focuses on the comparison of the mechanical performance of three warm-mix asphalt (WMA)
Received 15 March 2018 blends with recycled concrete aggregate (RCA) or electric arc furnace slag (EAFS) as substitutes of part of
Received in revised form 4 June 2018 the aggregate. A conventional hot mix asphalt (HMA) and a WMA without by-products were used as ref-
Accepted 8 July 2018
erences. The evaluation was carried out in a laboratory by means of testing specimens taken from exper-
imental pavement sections built in real production, laying and compaction circumstances. Performance
testing included wheel-tracking tests, four-point bending tests, and indirect tensile strength to assess
Keywords:
water sensitivity. An organic wax and a chemical surfactant were applied to lower handling temperatures
Additives
Electric arc furnace slag
of the WMA under study. Apart from the HMA and the WMA used as references, the study evaluated the
Performance properties influence of introducing 60% of RCA or 30% of EAFS into the WMA blends as substitutes of the aggregate.
Recycled concrete aggregate The obtained results for the WMA with by-products tested allowed to conclude that the introduction of
Resistance to fatigue EAFS or RCA into the WMA blends increases Marshall stability and may increase or decrease resistance to
Resistance to permanent deformation rutting. Findings also showed that stiffness modulus is somewhat reduced and fatigue resistance does not
Stiffness modulus change significantly. Additionally, water sensitivity is slightly reduced. Comparing these results with the
Warm-mix asphalt performance observed elsewhere for WMA without by-products revealed that the addition of RCA or
EAFS is satisfactory. The construction of experimental sections used conventional batch plant, paver
and compactors without any noticeable technical problems.
Ó 2018 Elsevier Ltd. All rights reserved.

1. Introduction
Warm-mix asphalt (WMA) are asphalt mixtures generally han-
dled at lower temperatures than conventional asphalt concrete.
Their constituents are typically mixed together at temperatures
varying from 100 to 140 °C [1,2]. This also allows lower com-
paction temperatures and longer haulage distances and additional
⇑ Corresponding author at: Department of Civil Engineering, Instituto Politécnico
de Coimbra, Instituto Superior de Engenharia de Coimbra, Rua Pedro Nunes,
3030- 199 Coimbra, Portugal.
E-mail address: capitao@isec.pt (S.D. Capitão).
time to carry out construction activities [3]. The authors have sum-
marised the production technology and properties of WMA in a
previous publication [3], in which they presented a detailed
description of additive based WMA and foamed bitumen
techniques.
The idea underlying the project partially described in this paper
is contributing to improve knowledge on more sustainable asphalt
mixtures, potentially suitable to substitute hot mix asphalt (HMA)
in many paving situations. The chosen way to achieve this goal was
studying the reuse of recycled concrete and steel slag as substi-
tutes of part of natural aggregates to produce WMA, which were
manufactured by adding additives to the blend. The use of those

https://doi.org/10.1016/j.conbuildmat.2018.07.041
0950-0618/Ó 2018 Elsevier Ltd. All rights reserved.
 F.C.G. Martinho et al. / Construction and Building Materials 185 (2018) 684–696
by-products is aligned with the goals of eliminating waste going to
landfill, contributing to the valorisation of by-products and, thus,
increasing circular economy. In addition, decreasing handling tem-
peratures is also recognized as an important way of reducing emis-
sions associated with paving construction and maintenance [3].
Although WMA show suitable volumetric properties and
mechanical performance, there is some weaknesses referred to in
the literature that can be found elsewhere [3,4]. Hence, this intro-
duction presents below a brief summary of the issues associated to
the reuse of recycled concrete aggregates (RCA) and steel slag
aggregates (SSA) in asphalt mixtures found in the literature.
Crushing concrete elements produce materials known as RCA.
Unlike natural aggregates produced in quarries, RCA have cement
mortar around the surface of the coarse natural aggregates. This
highly porous mortar is responsible for the high values of porosity
and water absorption of RCA [5,6]. Therefore, to assure enough bin-
der to involve the aggregate particles, the use of RCA generally
requires higher bitumen content in asphalt concrete (AC) because
part of the binder is absorbed by the aggregate [5,6].
Because fine particles of RCA have much more voids than coarse
aggregates, porosity of blends increases if fines are added to the
blend [6]. In addition, RCA is formed by angular and rough textured
coarse and fines particles, providing high Marshall stability and
small flow values [5,6]. This trend was not observed in a number
of studies, apparently because some of them used coarse RCA while
others also added fine fractions [7].
In addition, replacing part of the aggregates by RCA generally
degrades the resistance to moisture [5–9]. Likewise, RCA generally
lead to AC with lower stiffness [5–9] mostly when coarse RCA is
used, and leads to higher stiffness mainly if RCA is included as
the fine fraction [7,10]. When the blends are submitted to ageing
by heating them in an oven before compaction, the level of binder
absorption increases, which contributes to increase stiffness [10].
The resistance to permanent deformation is usually reported to
increase when RCA are incorporated into the blends [5–8,11]. Nev-
ertheless, several studies, referred to in [7], revealed the opposite
trend for several reasons: more RCA requires higher binder con-
tent; coarse and fine fractions of RCA tend to decrease resistance
to permanent deformation compared to those with coarse fraction
only, amongst others.
The fatigue resistance of HMA is adversely affected by the intro-
duction of RCA, decreasing fatigue resistance as the amount of RCA
increases [12]. On the contrary, several studies, reported by
Pasandín & Pérez [7], reveal that RCA tend to improve resistance
of HMA to fatigue.
Steel slag is a by-product of the steel industry, produced from
impurities separated from molten steel, which solidifies after cool-
ing (generally by adding water) [13,14]. There are three types of
slags: basic oxygen furnace (BOF), electric arc furnace (EAF) and
blast furnace (BF) [14,15]. EAF slag (EAFS) was used in the present
study. This type of slag is obtained from processes of melting recy-
cled scrap to produce different types of steel. A detailed review on
the use of EAFS in asphalt mixes can be found in [15]. EAFS is
formed by a several chemical components, such as iron oxides,
lime, silica, magnesia and alumina and other minor components
[15]. EAFS are used as aggregates in asphalt mixtures because their
physical and mechanical properties are usually suitable for that
purpose. In general, the amount of fine particles is low and the
coarser ones have good shape, rough texture and high angularity.
Although EAFS have relatively high density (3200–3800 kg/m3),
they exhibit elevated porosity and, therefore, higher water and bin-
der absorption than typical natural aggregates. Nevertheless,
expansion potential of asphalt mixes with EAFS in water is negligi-
ble [13]. Also, leaching of EAFS when used in asphalt mixes is very
low [16], showing practically no health risk [15]. EAFS have
685
generally good mechanical resistance and they reveal suitable
adhesion with bitumen [15].
The asphalt mixtures that incorporate EAFS as aggregate have
usually bulk densities 15 to 20% higher than similar asphalt mixes
without EAFS. Because of manifest angularity of the EAFS particles,
the porosity of asphalt concrete made with EAFS is also higher [15].
Generally speaking, using EAFS to substitute part of the natural
aggregate tend to improve mechanical behaviour of the resulting
asphalt mixes. This tendency is referred to in the literature for Mar-
shall stability, indirect tensile strength, stiffness, fatigue and resis-
tance to permanent deformation. This improved behaviour is
general attributed, on the one hand, to a better interlock between
aggregate particles as well as the roughness of steel slag assuring a
better adhesion binder-aggregate [13] and, on the other hand, to a
higher resistance to heavy loads and shear stress [15]. Most of the
authors state that the best results were obtained for a partial sub-
stitution of natural aggregates by EAFS (around 30%). However,
there is a number of cases reported in which the level of substitu-
tion was greater [15,17] and the mechanical behaviour remained
improved [15].
In what concerns water sensitivity, some divergent results were
found in the literature. Passeto and Baldo [18,19], for instance,
mentioned in [15], consistently observed that asphalt mixtures
with coarse and fine EAFS aggregates revealed higher resistance
to water damage in indirect tensile strength tests. They attribute
this behaviour to either a thicker bitumen film involving the aggre-
gates in mixtures with EAFS due to a higher binder content or a
good slag-binder adhesion that protect the particles against water
damage. On the contrary, some studies [20] reveal that water sen-
sitivity was higher in mixtures with coarse slag aggregates as com-
pared to similar mixtures with limestone aggregates. The observed
performance seems to be related with a worse affinity of the slag
with binders than that of limestone aggregate.
The information collected in the literature shows some oppor-
tunities and challenges related with the use of RCA and EAFS to
produce and lay WMA incorporating these by products in substitu-
tion of part of natural aggregates. It’s necessary, on the one hand, to
evaluate the expected volumetric and mechanical properties of
that type of asphalt mixtures and, on the other hand, to verify
the technological viability of these bituminous products in real
production conditions.
This paper summarises and examines the results found in a
study involving the evaluation of specimens produced in labora-
tory and sawed from experimental sections constructed in real
construction circumstances with conventional equipment. Chemi-
cal and organic additives in the form of pellets were used to allow
lowering production and compaction temperatures. The main
objective of the study was to evaluate the effect of using warm-
mix asphalt technologies, together with by-products aggregates
(EAFS and RCA), on mechanical properties of asphalt concrete.
2. Materials
2.1. Aggregates, RCA, EAFS, bitumen and additives
Three fractions of limestone crushed rock, RCA and EAFS were
used as aggregates to produce the warm-asphalt mixtures pro-
duced and evaluated throughout the study. In the early stage of
the project, two blends of limestone aggregates were used to pro-
duce the reference asphalt mixtures, one HMA with 35/50 paving
grade bitumen and three WMA with different additives to reduce
handling temperatures. These compositions were characterized
in the Marshall test and in terms of volumetric properties. The
same was carried out to the six additional WMA blends in which
686 F.C.G. Martinho et al. / Construction and Building Materials 185 (2018) 684–696

part of the natural aggregate was replaced by RCA, EAFS or both. summarizes the aggregates’ characteristics measured according
These blends were mixed and compacted in the laboratory to to the Portuguese specifications.
choose, based on the properties achieved, those that would be The EAFS and RCA aggregates were examined in terms of their
applied in the trial sections to be built in the later stage of the potential to release dangerous compounds, according to EN
project. 12457-4 [21] and EN 1744-3 [22]. The obtained analysis revealed
The blends selected were mixed in asphalt plants, laid and com- that both used by-products easily satisfy all the requirements indi-
pacted in trial sections of pavement in real production conditions. cated in the European regulations for leaching limits [23].
Furthermore, taking into account the laboratory results, the three Conventional 35/50 paving grade bitumen, with a penetration,
chosen blends suffered some adjustments prior to their use in at 25 °C, of 45  0.1 mm, and a ring & ball softening point of 56
the field. The grading Portuguese envelope for dense graded mix- °C, was the binder used for all the blends evaluated throughout
tures (AC 20 base) was the reference to derive the aggregate the project.
blends. Fig. 1 shows the blends’ grading curves applied in the trial One organic wax (SasobitÒ) and one chemical additive (Redis-
sections as well as the grading curves of RCA and EAFS, and two etTM) were used to lower the handling temperatures of the WMA.
images to show the appearance of the two by-products. Table 1 The additives were added to the blends in the form of solid pellets.

100
99.84
90 EAFS RCA
100.0
(Electrical Arc Furnace Slag) 87.24 89.0
(Recycled Concrete Aggregate)
80
% of material passing

EAFS 0/16 mm
RCA 0/20 mm
70

60 60.0
57.02
50

40

30 32.01 32.0

22.0
20

10 7.98 1.2 9.0

1.76
0

10
0.01

0.1

100
0.01

10

100
0.1

Particles size (mm)

Particles size (mm)
100

90 TWc(EAFS)3550
80 TWc(RCA)3550 = TWw(RCA)3550
% of material passing

70 Grading envelope
60

50

40

30

20
10

0
0.01

10
0.1

100

Particles size (mm)

Fig. 1. Grading curves of the aggregate blends, EAFS and RCA, and general view of EAFS and RCA.

Table 1
Properties of aggregates.

Properties Symbol Standard Limestone Aggregates RCA EAFS

10/20 (mm) 4/10 (mm) 0/4 (mm) 0/20 (mm) 0/16 (mm)
Sand Equivalent (%) SE EN 933-8 – – 62 30 78
Methylene Blue (g/kg) MBF EN 933-9 – – 2.3 3.7 0.2
Flakiness Index (%) FI EN 933-3 12.8 13.6 – 6.0 2.0
Los Angeles (%)1 LA EN 1097-2 27 – – 43 25
Density after drying (Mg/m3) qa EN 1097-6 2.6 2.6 2.5 2.3 3.6
Water absorption (%)1 WA EN 1097-6 1.74 1.09 0.99 6.1 1.7
Floating material (cm3/kg) FL EN 933-11 – – – 0.0 0.0
Concrete, conc. products and mortars Rc EN 933-11 – – – 84.0 0.0
Unbound aggr., natural stone and aggr. treated with hydraulic binders Ru EN 933-11 – – – 9.0 99.0
Bituminous materials Ra EN 933-11 – – – 0.7 0.0
Masonry elements of clay materials, calcium silicates and cellular concrete; Rb EN 933-11 – – – 5.0 0.0
Plastics, rubbers, metals, non-floating wood and stucco X EN 933-11 – – – 0.6 0.9
1
Note: Requirements of LA and WA24 commonly used in Portugal for natural aggregates to be applied in asphalt mixtures: 1. LA40 and WA24  2 (for base layers); 2. LA20
and WA24  2 (for binder layers).
 F.C.G. Martinho et al. / Construction and Building Materials 185 (2018) 684–696 687

Table 2
Composition of the studied blends.

A discussion on the main differences between them and the
expected influence on WMA’s mechanical characteristics can be
found elsewhere [3].
2.2. Compositions of blends
The information concerning the constituents and compositions
of all the studied blends are summarized in Table 2. Although the
data presented refers to all the blends studied, the objective of this
paper is highlighting the results derived from the blends applied in
real production conditions, whose compositions were adjusted
based on a previous evaluation of Marshall specimens produced
from analogous blends in the laboratory. The rates of additives
considered to produce the mixtures followed the manufacturers’
recommendations.
With the aim of making easier the identification of the blends,
the following key is used throughout the text: L – laboratory
blends; T – trial sections of pavement; W – warm-mix asphalt;
H – hot-mix asphalt; c – chemical additive; s – surfactant additive;
w – wax additive; RCA – recycled concrete aggregates; EAFS –
electric arc furnace slag aggregates; D – upper size sieve of the blend
(D = 14; 16; 20 mm); bitumen penetration @ 25 °C in 0.1 mm –
(10/20 = 1020; 35/50 = 3550; 50/70 = 5070; 160/220 = 160220).
A direct match was considered between some of the laboratory
mixtures and the blends produced in an asphalt plant, such as
LH3530 and TH3550, LWc3550 and TWc3550, LWc(EAFS)3550
and TWc(EAFS)3550, LWc(RCA)3550 and TWc(RCA)3550, LWw
(RCA)3550 and TWw(RCA)3550 (Table 2). The remaining labora-
tory blends LWw3550, LWw(EAFS)3550, LWw(RCA + EAFS)3550
and LWc(RCA + EAFS)3550 were not considered for testing in real
production conditions as the observed preliminary characteristics
were not considered promising.
2.3. Production processes and handling temperatures
Throughout the early stage of the testing program, the produc-
tion and compaction processes of the blends were established in
the laboratory. This was done by moulding cylindrical specimens
with (101.6 ± 0.1) mm in diameter. Mixing was carried out accord-
ing to EN 12697-35 [24] and compaction was accomplished
according to EN 12697-30 [25]. The laboratory blends were evalu-
ated in what concerns some volumetric properties as well as in the
Marshall test. In the second phase of the project, the blends with
better results for those properties were mixed in a batch plant, laid
with a common road paver and compacted with conventional com-
pactors. For further evaluation of mechanical properties of
TWc3550, TWc(EAFS)3550, TWc(RCA)3550 and TWw(RCA)3550
in the laboratory, slabs were sawed from the trial sections to pro-
duce the needed specimens for mechanical testing.
Following the guidelines from the additive manufacturers and
found in the literature [3], the WMA blends were mixed to a target
temperature of 120 °C whereas the HMA blends were prepared at
165 °C.
Compaction in the laboratory was accomplished between 100
and 110 °C for the WMA mixtures and between 120 and 150 °C
for the HMA. In what concerns the experimental sections, all the
trial sections were compacted in less than 45 min after laying,
allowing compaction above the minimum temperature of 80 °C
required for the WMA applied.
2.4. Volumetric properties, Marshall stability and flow of mixtures
The cylindrical specimens were characterized in terms of Mar-
shall stability and flow as well as the most common volumetric
properties: porosity, voids in mineral aggregates (VMA) and voids
filled with binder (VFB). These volumetric properties were mea-
sured on the laboratory specimens and on the samples collected
from the experimental sections, whereas stability and flow were
determined for laboratory specimens only (see Table 2: LH3550
to LWc(RCA + EAFS)3550) in the first stage of the project. Fig. 2
shows the average results measured for the aforementioned prop-
erties. For ease of comparison, the results of matching blends pro-
duced in laboratory and cut from the trial sections pavements are
adjacent to each other in the graphs.
Fig. 2 shows that the WMA blends LWw3550 and LWc3550
returned low stability values, below the minimum limit currently
used in Portugal. Although in terms of stability/flow ratio, all the
blends satisfied the minimum of 2 kN/mm typically required for
AC 20 base in Portugal, LWw(RCA)3550, LWc(RCA)3550, LWw
(RCA + EAFS)3550 and LWc(RCA + EAFS)3550 revealed higher flow
values than the Portuguese requirements. Typically, AC 20 has
porosity values from 3 to 6%. In this study, blends LWw(RCA)
3550, LWc(RCA)3550, LWw(RCA + EAFS)3550 and LWc(RCA +
EAFS)3550 revealed low porosity.
As shown in Table 2, no similar materials to LWw3550, LWw
(EAFS)3550, LWw(RCA + EAFS)3550 and LWc(RCA + EAFS)3550
were applied in the experimental sections. Based on the results
and some practical reasons, such as the blends’ workability
688 F.C.G. Martinho et al. / Construction and Building Materials 185 (2018) 684–696

Porosity (%) VMA (%)

9% 18%
8% 17%
7% 16%
6%
15%
5%
4% 14%
3% 13%
2% 12%
1% 11%
0% 10%
LH3550

LWc3550

LWw(RCA+EAFS)3550

TWc(RCA)3550

TWw(RCA)3550
TWc(EAFS)3550
TH3550
LWw3550

TWc3550

LWc(RCA+EAFS)3550
LWc(EAFS)3550

LWw(EAFS)3550
LWc(RCA)3550

LWw(RCA)3550

LH3550

LWc3550

LWw(RCA+EAFS)3550

TWc(RCA)3550

TWw(RCA)3550
TH3550
LWw3550

TWc3550

LWc(RCA)3550
LWc(RCA+EAFS)3550
LWc(EAFS)3550
TWc(EAFS)3550
LWw(EAFS)3550

LWw(RCA)3550
Stability (kN) Flow (mm)
16 6 MAX:
14 MAX: Portuguese 5 Portuguese
12 specificaons limit specificaons
10 4
8 Min. 3
6 2
4
2 1
0 0
LWc3550
LH3550
LWw3550

LWw(RCA)3550
LWc(RCA)3550
LWw(EAFS)3550
LWc(EAFS)3550
LWw(RCA+EAFS)3550
LWc(RCA+EAFS)3550

LWc3550
LH3550

LWw3550

LWw(RCA)3550

LWc(RCA)3550

LWw(EAFS)3550

LWc(EAFS)3550

LWw(RCA+EAFS)3550

LWc(RCA+EAFS)3550
Fig. 2. Average values of Marshall stability and flow, and volumetric properties of the studied blends.

observed and the search for better results in the performance eval- 3.2. Methods
uation stage, LWw(RCA)3550, LWc(RCA)3550 and LWc(EAFS)3550
were the compositions selected to the later stage of the project. Wheel-tracking (WT) tests, performed according to EN 12697-
The volumetric properties were somewhat different in the speci- 22 [27], were carried out to evaluate permanent deformation per-
mens cut from the trial sections and in the specimens produced formance of the blends. The wheel of WT applies a stress of about
in the laboratory. The differences observed can be mainly attribu- 700 kPa and passes over the specimen up to 10,000 times. When a
ted to compaction method (impact compaction) standardised for rut depth of 20 mm is attained the tests finish, even if the specified
the Marshall method, which does not reproduce the field com- number of cycles is not achieved. The evaluation of permanent
paction method [26]. deformation resistance of the studied blends was carried out at
50 and 60 °C.
3. Mechanical properties methods and results Four-point bending tests (EN 12697-26 [28]) were carried out
to assess the stiffness modulus and phase angle of the blends.
3.1. Specimens for testing The followed standard considers prismatic beams to perform
repetitive four-point bending tests, under controlled displacement
In the field, the studied asphalt mixtures were mixed in a batch conditions (strain level of 50 lm/m) at 20 °C. The applied strain
plant, transported in lorries, laid with a paver in the trial sections varied with time according to a sinusoidal wave. Each specimen
and compacted with tyre and metallic rollers. was submitted to loading frequencies of 30, 20, 10, 5, 3 and 1 Hz.
Slabs with 300  400  60 mm3 were cut from the trial sec- Fig. 4 presents the measured stiffness moduli and phase angles
tions’ pavements to carry out wheel-tracking tests. Some of those for the strain level of 50 lm/m.
slabs were cut again to make prismatic beams with 400x55x55 Evaluation of fatigue resistance of the blends was also per-
mm3 to submit to four-point bending tests. formed in four-point bending tests following the procedure
Samples of loose mixtures, collected during the construction of described in EN 12697-24 [29], testing eighteen specimens (with
pavements trial sections, were used to mould and compact cylin- a cross section of 55  55 mm2), six per strain level. The tests were
drical specimens with 101.6 ± 0.1 mm in diameter, aiming at sub- carried out at 20 °C in displacement control conditions, applying
mitting them to indirect tensile tests. repetitive sinusoidal loads along time, with a frequency of 10 Hz
 F.C.G. Martinho et al. / Construction and Building Materials 185 (2018) 684–696 689

and three strain levels (150, 250 and 350 lm/m). As usual, fatigue
failure of specimens was considered the moment in which a 50%
loss of the initial stiffness modulus was observed.
The evaluation of water sensitivity was carried out based on
ITSR – Indirect tensile strength ratio – following the procedures
indicated in EN 12697-12 [30] and EN 12697-23 [31].
This evaluation was performed on six cylindrical specimens
with (100 ± 3) mm in diameter. They were moulded in the labora-
tory and compacted by impact in the Marshall compactor. Three of
the specimens were firstly conditioned in water at (40 ± 1) °C for
68 to 72 h and then in air at (15 ± 1) °C for a time greater than or
equal to 2 h. Finally, indirect tensile strength (ITS) was evaluated
in diametrical load compression tests. ITSR is the ratio between
ITS obtained for the conditioned specimens (ITSw) and ITS mea-
sured to the dry specimens (ITSd) not submitted to the condition-
ing procedure.
3.3. Evaluation of rutting resistance
Fig. 3 shows the results of slope in air, WTSAIR (wheel-tracking
slope), and RDAIR (rut depth, in mm), obtained for the two param-
eters considered to evaluate resistance to permanent deformation,
according to EN 12697-22 [27], determined from the time – defor-
mation curves (small device in air). The error bars show the range
of tests’ results obtained on two specimens.
The results show that permanent deformation resistance of
TWc(RCA)3550 (blend with a chemical additive and RCA) is very
sensitive to temperature. Although it has a reasonable performance
at 50 °C, its behaviour significantly worsened at 60 °C as the tests
did not achieved 10,000 loading cycles. Fig. 2 had already shown
that LWc(RCA)3550, the matching preliminary blend, revealed a
tendency to deform at high temperature in the Marshall compres-
sion test. Fig. 3 also shows that the blends produced with chemical
additive revealed more temperature sensitivity than the blends
produced with organic wax.
The low stability obtained in the Marshall test for LWc3550
(with chemical additive and limestone natural aggregate), shown
in Fig. 2, could foresee the weak resistance observed for
TWc3550, the equivalent reference blend tested in the trial
sections.
TWc(EAFS)3550 (WMA with chemical additive and EAFS) had
an intermediate performance amongst the blends tested even
though it revealed low S/f ratio (2 kN/mm) in the Marshall test.
However, the rut depth developed is quite high, suggesting a weak
resistance to permanent deformation.
The reference HMA (TH3550) and TWw(RCA)3550 (WMA
with organic additive and RCA) were the best performant blends
in terms of resistance to permanent deformation. TWw(RCA)
3550 and TWw(RCA)3550 (chemical additive and RCA) have
the same types and contents of bitumen and aggregates but
not the same additive to reduce handling temperature. This
seems to be the most important factor in what concerns perma-
nent deformation resistance at 60 °C in this case. These results
show that despite the Marshall stability and flow of LWc(RCA)
3550 and LWw(RCA)3550 were almost the same, the corre-
sponding blends applied in the test site, respectively, TWc(RCA)
3550 and TWw(RCA)3550, revealed very different permanent
deformation resistance.
Looking at the WT results it can be stated that the effect of
including EAFS as part of the aggregate in TWc(EAFS)3550 is
inconsistent in terms of permanent resistance performance.
Despite the level of rut depth reached at 60 °C is slightly lower
for TWc(EAFS)3550 than for TWc3550 (the reference WMA),
the first one revealed weaker resistance at 50 °C. Also, TWw
(RCA)3550 had an increase in rutting performance when com-
pared with TWc3550. These results reveal that adding organic
wax and RCA as part of the aggregate contributed to increase
permanent deformation performance of the WMA, allowing a
performing level similar to HMA.
Despite there are no requirements established for the WT test
results in Portugal, the high levels of rut depth reached by TWc
(EAFS)3550 and TWc(RCA)3550 at 50 °C seem to be incompatible
with a suitable permanent deformation resistance for the climatic
conditions in Portugal, where the pavements’ binder layers can
achieve up to 50 °C in the hottest regions in summer.
3.4. Stiffness
Fig. 4 presents WMA with EAFS and RCA in green. The reference
WMA, TWc3550, with chemical additives, had stiffness moduli

WTSAIR (mm/103 cycles) RDAIR (mm)

4.0 25
50°C 50°C
3.5 20
60°C
60°C 3.10 20 17.09
3.0 14.22

2.5 15 13.1

2.0
1.5 10
6.62 6.9
0.91 0.73 5.13
1.0 0.71 0.21 0.25 3.97 3.59 4.09
5
0.34
0.5 0.23 0.14 0.25

0.0 0
50
0
0

50
50

c(R 550
0

50
0

c(R 550

55
55
55
55

35

35
35

35

TW A)3
c3
T W A )3

)3
c3

TW S )3

TH
A)
TH
A)

TW

FS
TW

RC
C
F

RC
C

A
A

c(E

w(
c(E

w(

TW
TW
TW

Note: For TWw(RCA)3550, WTSAIR was determined for a number of loading cycles higher than 2000.

Fig. 3. Wheel-tracking test results: WTSAIR and RDAIR (50 & 60 °C). Note: For TWw(RCA)3550, WTSAIR was determined for a number of loading cycles higher than 2000.
690 F.C.G. Martinho et al. / Construction and Building Materials 185 (2018) 684–696

100
10000 TWc35 50 TWc(EAFS)3550 TWc35 50 TWc(EAFS)3550
TWw(RCA)3550 TWw(RCA)3550 TWw(RCA)3550 TWw(RCA)3550

Phase angle (degree)

Stiffness modulus (MPa)
TH3550 TH3550

1000 10
1 5 25 1 5 25
Frequency (Hz) Frequency (Hz)

Fig. 4. Stiffness moduli and phase angles for the studied mixtures (at 50 lm/m).

higher than the other blends. TWW(RCA)3550 is the one that 8000
revealed a deformability behaviour closest to TWc3550, with a 30 Hz
stiffness at 10 Hz around 4700 MPa and a phase angle of 19.4°. In 7000

Sffness modulus (MPa)

general, HMA have higher stiffness modulus and lower phase
angles than similar WMA. However, in this study TH3550 had
greater porosity than usually (see Fig. 2). Because of this, the stiff-
ness values for this blend were moderated in comparison with typ-
ical values for conventional HMA and lower than those of the
reference WMA blend (TWc3550).
When EAFS was introduced into the blend (TWc(EAFS)3550),
stiffness modulus has reduced close to 29% and the phase angle
slightly increased (from 23.2 to 24.5) in comparison to TWc3550.
In what concerns stiffness the results were good to TWw(RCA)
3550 (with chemical additive and RCA) in comparison with a typ-
ical hot-mix asphalt, even if the values were the lowest of the
study (reducing circa 31% as compared to the reference blend
TWc3550), with a modulus around 4000 MPa at 10 Hz.
The results for TWw(RCA)3550 showed that when RCA was
used as part of the aggregate blend the stiffness moduli are close
to those measured to TWc3550 (the WMA without by-products).
However, the values at 10 Hz or higher were slightly lower for
TWw(RCA)3550 (e.g. stiffness of 4722 and 5273 MPa for TWw
(RCA)3550 and TWc3550, respectively), although phase angles
have been around 16% lower for TWw(RCA)3550 (the coefficient
of variation of phase angle is 6.4%). This indicates that including
RCA into the WMA blends the viscous component seems to
decrease as compared to the blends without by-products and with
EAFS.
A multivariate regression analysis was also carried out with
SPSS (Statistical Package for the Social Sciences) to complement
the evaluation in terms of stiffness modulus. The used statistical
tools allowed the validation of regression analysis assumptions.
The best-fit models (higher adjusted R2), after several types of
models tried [26], were based on the independent variables deter-
mined in the laboratory. Stiffness modulus (E) was considered as
the dependent variable whereas loading frequency (f in Hz), poros-
ity (Vm in %) and strain level (e in lm/m) were used as independent
variables as they were statistically significant and led to the best fit
models [expression (1)]. The models presented in Fig. 5 only con-
sider the range of porosity observed throughout the experimental
framework. The derived regression model is useful as a comple-
ment of Fig. 4 because it allows us to observe the response of the
studied warm-mix asphalt in terms of stiffness if porosity and
loading conditions (strain and frequency) vary in the range evalu-
ated in this study.
log E ¼ a  log f þ b  V m þ c  log e þ d
10 Hz
6000
TWc(EAFS)3550
5000
30 Hz
4000 1 Hz
3000 10 Hz
TWw(RCA)3550
2000
1 Hz
1000
0
4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0
Porosity (%)
Fig. 5. Results of stiffness modulus for TWc(EAFS)3550 and TWw(RCA)3550
derived from Eq. (1).
The coefficients a, b, c and d as well as the adjusted R2 are pre-
sented in Table 3.
Fig. 5. shows the variation of stiffness modulus calculated from
Eq. (1) as a function of porosity in the range of values observed on
the specimens submitted to testing. Although porosity being
higher in TWw(RCA)3550 than in TWc(EAFS)3550, the results
show that using organic wax as additive and RCA as substitute of
part of the aggregate increased stiffness moduli of the WMA.
3.5. Resistance to fatigue
Fig. 6 presents the fatigue laws derived from the tests carried
out.
Although the results in Fig. 6 reveal that fatigue behaviour of
the blends is somehow similar, they also show that slope of the
lines is slightly higher for the WMA blends (TWc3550 to TWw
(RCA)3550) than for the HMA (TH3550). Also, TWc3550, the
WMA without by-products, has a line less sloping than the blends
with by-products. Amongst these, TWw(RCA)3550, with organic
wax and RCA, has the less sloping line. This finding can be analysed
in more detail by calculating e4, e5 and e6, which represent the
strain values to induce specimens’ failure after 10.000, 100.000
and 1 million loading cycles, respectively (Fig. 7).
Fig. 7 confirms different rankings for the tested blends depend-
ing on the number of loading cycles applied. TH3550, for example,
ð1Þ
is the worst performant blend at 10.000 cycles and is the best one
 F.C.G. Martinho et al. / Construction and Building Materials 185 (2018) 684–696 691

Table 3
Models of stiffness modulus derived from multivariate regression analysis.

Blends By-product Additives Coefficients Adjusted R2 (%)

a b c d
TWc(EAFS)3550 EAFS Chemical 0.306 0.048 0.058 3.581 94.8
TWw(RCA)3550 RCA Organic Wax 0.219 0.065 0.051 4.042 79.6

μ
μ

μ
ω

Fig. 6. Fatigue laws at 20 °C derived from four-point bending tests by regression analysis.

Failure strain ( /m) 3.6. Water sensitivity

TWc3550

ε4 TWc(EAFS)3550 Since the heating temperature of constituents in WMA blends is

TWc(RCA)3550 relatively low, some humidity can remain in the blend. This is
387
366
350 353 342
TWw(RCA)3550 referred to in the literature [3] as potentially harmful to water
TH3550 resistance of WMA. Fig. 8 shows the results for the studied blends.
ε5 All the results revealed good performance in what concerns
220 224 208 211 217 water sensitivity. In general, the specifications require ITSR values
ε6 higher that 80%. TH3550 (HMA) reached a value of 93.1% and
133 129 124 126 138 TWc3550, the reference WMA, achieved 94.2% for ITSR. When EAFS
or RCA have been incorporated into the blend as substitutes of the
natural aggregate, both the ITS and ITSR decrease as compared to
10 000 100 000 1 000 000 the value measured to TWc3550 but the observed performance
Number of loading cycles was still good. However, the incorporation of RCA resulted in lower
resistance to water damage in TWc(RCA)3550 and TWw(RCA)3550
Fig. 7. Fatigue performance in terms of strain needed to specimen’s failure.
4 000 100
ITS d (kPa) ITS w (kPa) ITSR (%)

94.2 95
at 1 million cycles. Taking into account that evaluation of fatigue 3 500 93.1
91.6
resistance aims to anticipate which mixtures will behave better 90
and worse after being submitted to a high number of loading 3 000
84.7
cycles, e6 is typically the preferred parameter to evaluate fatigue 85
ITS (kPa)

ITSR (%)

82.8
resistance of asphalt concrete. 2 500 80
2 297
The e6 values show that TWc3550, the blend without by- 2 139
products, is the best performant WMA blend (only 3% less than 75
2 000 1 831 1 870
HMA, TH3550). The introduction of EAFS or RCA slightly 1 725
1 549 70
decreased resistance to fatigue as compared to TWc3550: 3% in 1 463
1 500 1 340
TWc(EAFS)3550, 7% in TWw(RCA)3550 and 5% in TWw(RCA) 1 290 65
1 093
3550. Moreover, the results also reveal that the WMA blends,
including the ones with EAFS or RCA, resist better to higher 1 000 60
TWc35 50 TWc(EAFS)3550 TWc(RCA)3 550 TWw(RCA)3550 TH3550
strain levels that the HMA (TH3550) whereas for low strain
levels the opposite occurs. Fig. 8. Results of ITSR, ITSw and ITSd for the blends assessed in the study.
692 F.C.G. Martinho et al. / Construction and Building Materials 185 (2018) 684–696
than the use of EAFS. Therefore, the results showed that for the
additives and production temperatures used for the studied blends,
no problems were observed regarding susceptibility to moisture of
the tested WMA blends, when EAFS or RCA was added as part of
the aggregate.
3.7. Comparative performance of the studied mixtures
Fig. 9 compares the relative performance of the tested
blends in what concerns resistance to permanent deformation
(RD and WTS), stiffness, fatigue resistance and water sensitivity.
The best performant blend represented in the graph for a
parameter achieves 100% in that parameter. The remaining
mixtures have a lower percentage, which represents the rela-
tive performance of each blend as compared to the best perfor-
mant one.
As Fig. 9 shows, TH3550 and TWw(RCA)3550 (WMA with wax
and RCA) are the blends with the closest performance in most of
the assessed parameters. Moreover, it also reveals that for the
WMA blends with a chemical additive (TWc3550, TWc(EAFS)
3550 & TWw(RCA)3550) revealed similar behaviour, with a
remarkable low resistance to permanent deformation. Conse-
quently, the observed performance was apparently commanded
by the type of additive used instead of the effect of introducing
EAFS or RCA as substitutes of part of the aggregates. In what con-
cerns resistance to fatigue, stiffness and water sensitivity of the
blends, the use of EAFS or RCA did not significantly influence the
observed performance. Fig. 9 also shows that although TWw
(RCA)3550 has incorporated RCA, its general behaviour was not
comparable to the remaining WMA blends. Generally speaking,
TWw(RCA)3550 revealed good properties, comparable to TH3550
in all the parameters considered. As far as permanent deformation
is concerned, TWw(RCA)3550 revealed even better resistance than
RD
100%
Results overview
80%
60%
40%
ITSR
20%
TH3550
TWc3550
0%
TWc(EAFS)3550
TWc(RCA)3550
TWw(RCA)3550
fague sffness
Fig. 9. Relative performance of the tested blends in terms of five evaluated
parameters.
Fig. 10. Global score of the blends assessed in the study.
the observed for TH3550 (HMA). This behaviour seems to derive
more from the organic wax used as additive than from the incorpo-
ration of RCA.
Because engineers should have a procedure to decide on the
blends to use or to reject, complementarily to the visual analysis
based on Fig. 9, a process to rank the blends was also considered.
Briefly, it consists in using the results of mechanical performance
measured in the laboratory, to derive a single score. Six parameters
were considered to rank the blends: RDAIR, WTSAIR, E, phase angle
(/),e6 and ITSR. Taking into account that AC 20 is typically applied
in base or binder pavement layers, some of those parameters are
more important than the others. Therefore, the parameters were
weighted to consider different contributions for an overall perfor-
mance indicator. Although this procedure is somehow subjective, it
expresses the authors’ perspective regarding the suitable proper-
ties for a blend to be applied in base and binder layers: RDAIR
15%; WTSAIR 15%; E 20%; / 5%; e 6 40%; ITSR 5%. The first step
involves a transformation of each parameter to allow expressing
it in a scale from 0 to 100%, dividing each individual value by the
maximum (or minimum) of them. After this, these partial scores
are summed to determine a global score. Values closer to 100% rep-
resent a better global performance than lower scores. Fig. 10 shows
the scores calculated to the blends considering the aforesaid
procedure.
Taking all the evaluated parameters and the corresponding
weights into consideration, TWw(RCA)3550 was the best blend,
even better than TH3550 (HMA). These results also show that the
combination in the same blend of a chemical additive and RCA,
has slightly reduced the general performance of TWc(EAFS)3550
and TWc(RCA)3550 in comparison to TWc3550. This small score
decrease is mainly associated with resistance to fatigue and
stiffness.
3.8. Relative performance as compared with other WMA blends
An expected delivery of this study was to assess the influence of
substituting in WMA blends part of the natural aggregate by RCA
or EAFS. Apart from the results provided by this study both for con-
WTS
ventional WMA and WMA with by-products as aggregates, some
additional results are available in the literature, particularly for
WMA manufactured with natural aggregates. Notice that the com-
positions of the blends studied elsewhere can differ in some degree
from the mixtures under study. Therefore, these sources of infor-
mation should be considered only to compare the general tenden-
cies observed for conventional WMA in terms of the resulting
properties, when using RCA and EAFS. Fig. 11 presents a summary
of the comparisons carried out.
 F.C.G. Martinho et al. / Construction and Building Materials 185 (2018) 684–696 693

a) Resistance to permanent deformaon

WTSAIR (@50oC) T20Wc(EAFS)3550
M22 T20Wc(EAFS)3550
M22 WTSAIR (@60oC)
0.0 0.2 (mm/103 cycles)
8: 14Ww5070
H 0.2
T20Wc(RCA)3550
M23 0.6
1.0
0.3 1.4
0.5 1.8
G
7: 14Ww5070 M25 2.2
0.6 T20Ww(RCA)3550 2.6
0.8 3.0
3.4
0.9 3.8
9: 20WwI M23
T20Wc(RCA)3550
6: 14Wc5070F A 14Wc5070
1:

5: 14Ww5070E B 14Ww5070
2:

4: 20Ww1020D C 14Ww160220
3:
M25
T20Ww(RCA)3550

RDAIR (@50oC) M22

T20Wc(EAFS)3550 T20Wc(EAFS)3550
M22 RDAIR (@60oC)
2 0 (mm)
4 2
6 4
4: 20Ww1020 T20Wc(RCA)3550 6
D 8 M23 8
10 11: Ww5070K 10 M23
T20Wc(RCA)3550
12 12
14 14
16 16
18 18
20 20

3: 14Ww160220C M25
T20Ww(RCA)3550

10: 20WwJ M25

T20Ww(RCA)3550

2: 14Ww5070B A 14Wc5070
1:
I
9: 20Ww

b) Resistance to fague and sffness

ε6 (20oC) T20Wc(EAFS)3550
M22 T20Wc(EAFS)3550
M22 Sffness (MPa)
(μm/m) 180 Q 8000
17: 16Ww5070 M23
T20Wc(RCA)3550
160 7000
18: 16Ww5070
R 140 T20Wc(RCA)3550
M23
16: 16Ws5070 6000 T20Ww(RCA)3550
120 P M25
100 5000
80 4000
60
40 15: 16Ws5070
O 3000 1: 14Wc5070
A
20 2000
D
4: 20Ww1020 0 T20Ww(RCA)3550
M25 1000

N
14: 16Ws5070 E5: 14Ww5070

M
13: 16Ww5070 F6: 14Wc5070
3: 14Ww160220C A 14Wc5070
1:
12: 16Ws5070L G
7: 14Ww5070
B
2: 14Ww5070 H
8: 14Ww5070

c) Indirect tensile strength and water sensivity

ITSdry (kPa) T20Wc(EAFS)3550
M22 T20Wc(EAFS)3550 ITSR (%)
M22
2500 18: 16Ww5070R 100 T20Wc(RCA)3550
M23
13: 16Ww5070 2250 T20Wc(RCA)3550 90
M 2000 M23 16: 16Ws5070P 80 T20Ww(RCA)3550
M25
1750 70
1500 60
1250 15: 16Ws5070O 50 1:
A 14Wc5070
1000 40
750 30
12: 16Ws5070L 500 T20Ww(RCA)3550
M25 20
250 14: 16Ws5070N 10 2:
B 14Ww5070
0 0

M
13: 16Ww5070 3:
C 14Ww160220
D
4: 20Ww1020 1:
A 14Wc5070
12: 16Ws5070L 4:
D 20Ww1020

8: 14Ww5070H E 14Ww5070
5:
3: 14Ww160220C B 14Ww5070
2: 7: 14Ww5070 F6: 14Wc5070
G

Fig. 11. Comparison of performance parameters observed in the study with results collected from the literature. 1: AC 14 surf 50/70: chemical [32], 2: AC 14 surf 50/70:
organic wax [33], 3: AC 14 surf 160/220: organic wax [33], 4: AC 20 base 10/20: organic wax [33], 5: AC14 surf 50/70 mix@145 °C: organic wax [34], 6: AC14 surf 50/70
mix@130 °C: chemical [34], 7: AC14 surf 50/70 mix@145 °C: organic wax [33], 8: AC14 surf 50/70 mix@125 °C: organic wax [33], 9: AC20, mix@147.5 °C: organic wax [34],
10: AC20, mix@125 °C: organic wax [35], 11: 50/70: organic wax [36], 12: AC16 surf 50/70: surfactant [37], 13: AC16 surf 50/70: organic wax [37], 14: AC16 surf 50/70:
surfactant A [38], 15: AC16 surf 50/70: surfactant B [38], 16: AC16 surf 50/70: surfactant C [38], 17: AC16 surf 50/70: organic wax [38], 18: AC16 surf 50/70: organic wax [39].
694 F.C.G. Martinho et al. / Construction and Building Materials 185 (2018) 684–696
In terms of rutting resistance (Fig. 11a), TWw(RCA)3550 per-
formed very well either at 50 or 60 °C in comparison with a sub-
stantial number of other WMA without by-products. This allow
us to anticipate good performance for a combination of wax &
RCA. The performance obtained for WTSAIR at 50 °C, for instance,
is quite close to the obtained for a WMA produced with a hard
bitumen 10/20 (blend 4), which is known to perform very well
against permanent deformation.
In what concerns TWc(EAFS)3550 (chemical & EAFS) and TWc
(RCA)3550 (chemical & RCA) the permanent deformation perfor-
mance was quite irregular. At 50 °C TWc(RCA)3550 almost
achieved a resistance similar to the best performant blends pro-
duced with 50/70 paving grade bitumen but at 60 °C its perfor-
mance was weak. Although for TWc(EAFS)3550 the results were
inconsistent when analysing the two measured rutting perfor-
mance parameters, the comparison with the other blends consid-
ered reveals that combining a chemical additive and EAFS as
substitute of aggregates has a tendency to originate weak resis-
tance to rutting.
Looking at resistance to fatigue (Fig. 11b) all the studied blends
performed quite well in comparison with the results taken from
the literature, even better than the observed for the blends 1, 2
and 18, all of them produced with a bitumen softer than the stud-
ied blends and, therefore, with increased expectations in terms of
fatigue resistance. For the combination of additives (wax or chem-
ical) and by-products (RCA or EAFS) used as substitutes of part of
the aggregate, resistance to fatigue seems not to be a concern.
In terms of stiffness moduli at 20 °C and 10 Hz, TWc(EAFS)3550,
TWc(RCA)3550 and TWw(RCA)3550 revealed lower values than
practically all the blends used for comparison. Nevertheless, the
values of stiffness modulus measured for the studied blends are
approximately in the range 4000–5000 MPa. According with the
authors experience as consultants, this is satisfactory for typical
pavement asphalt layers. Therefore, stiffness level of these WMA
produced with RCA and EAFS is not a problem.
Although TWc(EAFS)3550 and TWc(RCA)3550 have achieved
values of ITSdry generally lower than those observed for the other
WMA without by-products considered for assessment, their sensi-
tivity to water was good, with ITSR values above 80%. Also, TWw
(RCA)3550 reached good performance in terms of ITSdry and ITSR
as compared to all the WMA mixtures used for comparison. Based
on this analysis, it can be concluded that the incorporation of EAFS
or RCA in WMA is not a problem vis-a-vis resistance to water of
WMA.
As a corollary of the comparisons performed above, a procedure
close to that proposed in [40] was followed to summarise the
trends referred to in the literature (already mentioned at the
Table 4
Tendency of change in properties of bituminous mixtures (adapted from [40]).
Parameter Marshall stability
HMA with RCA1
HMA with EAFS1
WMA technology (at HMA temperature)1
WMA technology (with temperature reduction)1
WMA with RCA (in this study: with temperature reduction)2
WMA with EAFS (in this study: with temperature reduction)2
usually increases.
may increase.
usually unchanged.
may decrease.
usually decreases.
1
Comparing with conventional HMA.
2
Comparing with a similar WMA blend without by-products as aggregates.
beginning of this paper) as well as the general tendencies observed
in this study (Table 4). They are shown together to highlight the
effects on the properties of WMA blends derived from substituting
part of the aggregates by RCA or EAFS.
4. Conclusions
The use of by-products in WMA blends was the main focus of
this study aiming at verifying, on the one hand, the feasibility of
these techniques in real production conditions and, in the other
hand, to evaluate the influence on the WMA properties resulting
from the incorporation of RCA or EAFS as substitutes of part of
the aggregates.
With the exception of handling temperatures of the mixtures,
lower than the commonly applied for HMA, the construction pro-
cedures used in the field to lay and compact the WMA blends with
60% of RCA or 30% of EAFS did not require any modification of the
traditional construction techniques. These percentages of aggre-
gate substitution were derived from preliminary testing carried
out in laboratory. The specimens collected from the field allowed
the development of a mechanical characterisation program in the
laboratory.
The trends mentioned in the literature for HMA incorporating
the same types of by-products considered in this study were not
completely confirmed (Table 4). Although the blends produced
with RCA, for instance, had the same binder content that the other
blends, the resulting porosity was not consistently higher. Only
TWw(RCA)3550 (wax and RCA) revealed porosity considerably
higher but the same did not happen for the similar laboratory
blend (LWw(RCA)3550). This shows that the porous cement mor-
tar of the RCA did not consistently absorb the blends’ binder.
The results produced this study as well as the information gath-
ered from the literature led to the following conclusions:

The general performance of WMA with 60% of RCA was very
good (even better than HMA used as reference) when the addi-
tive was an organic wax, but rutting resistance was weak when
a chemical additive was used instead of wax;

Because all the WMA blends with chemical additive tested had
rather weak permanent deformation resistance, apparently the
effect of adding a chemical additive overlapped the effect of
incorporating RCA into the blend;

The comparison of results of this study with the obtained for
other blends evaluated elsewhere reveals that combining chem-
ical additives and EAFS as substitute of aggregates has a ten-
dency to originate mixtures with weak resistance to rutting;
Porosity Rutting resistance Stiffness Fatigue resistance Water sensitivity
 F.C.G. Martinho et al. / Construction and Building Materials 185 (2018) 684–696

Although the use of by-products as substitutes of part of the
aggregate decreased the stiffness moduli achieved in compar-
ison with the blend without by-products, typical values, at 20
°C, were about 4000–5000 MPa, which is satisfactory for a con-
ventional base/binder layer;

Fatigue life of WMA with RCA or EAFS was very satisfactory and
close to the performance observed for the blends (HMA and
WMA) used as references; also, comparing the results obtained
with other determined elsewhere on conventional WMA led to
the conclusion that resistance to fatigue seems not to be a con-
cern for WMA with RCA or EAFS;

Sensitivity to water slightly decreased for the WMA blends with
by-products as compared with the conventional WMA and HMA
and the performance achieved by ITSR was over 80%, which is
the typical lower level requirement for asphalt concrete; so,
water sensitivity was not a problem when EAFS or RCA are
added to the blends to produce WMA;

Although conventional equipment was used to build the exper-
imental sections of pavement, no specific problems were
observed in the field for the WMA blends with RCA or EAFS in
any of the necessary operations: mixing, transport, laying and
compaction.
Concluding, it can be stated that based on the authors’ experi-
ence throughout this study WMA produced with a high percentage
of RCA or EAFS is practicable but some combinations of them with
additives should be evaluated with caution. In fact, in situations
where good performance to permanent deformation is required
the combination in a WMA of RCA and a chemical additive should
be avoided. Since the convenient evaluation is carried out, solu-
tions based on WMA with RCA or EAFS are suitable alternatives
to conventional HMA. It must also be emphasized that the use of
RCA or EAFS as part of the aggregates fulfil the European goals of
increasing circular economy, reducing the waste going to landfill
and increasing valorisation of by-products in different value
chains.
Conflicts of interest
The authors declare no conflicts of interest.
Acknowledgements
The authors want to thanks to the technical personnel working at
the Laboratory of Highways and Transportation (LVTC) of the
CERIS, Instituto Superior Tecnico, Universidade de Lisboa.
This research did not receive any specific grant from funding
agencies in the public, commercial, or not-for-profit sectors.
References
[1] F. Olard, C. Noan, Low Energy Asphalts. Routes Roads 336/337, PIARC (World
Road Association), 2008, pp. 131–145.
[2] EAPA, The Use of Warm Mix Asphalt – EAPA Position Paper, European Asphalt
Pavement Association, Brussels, 2010.
[3] S. Capitão, L. Picado-Santos, F. Martinho, Pavement engineering materials:
review on the use of warm-mix asphalt, Constr. Build. Mater. 36 (2012) 1016–
1024, https://doi.org/10.1016/j.conbuildmat.2012.06.038.
[4] F. Martinho, L. Picado-Santos, S. Capitão, Mechanical properties of warm-mix
asphalt concrete containing different additives and recycled asphalt as
constituents applied in real production conditions, Constr. Build. Mater. 131
(2017) 78–89, https://doi.org/10.1016/j.conbuildmat.2016.11.051.
[5] J. Zhu, S. Wu, J. Zhong, D. Wang, Investigation of asphalt mixture containing
demolition waste obtained from earthquake-damaged buildings, Constr. Build.
Mater. 29 (2012) 466–475, https://doi.org/10.1016/j.conbuildmat.2011.09.
023.
[6] S. Paranavithana, A. Mohajerani, Effects of recycled concrete aggregates on
properties of asphalt concrete, Resour. Conserv. Recycl. 48 (2006) 1–12,
https://doi.org/10.1016/j.resconrec.2005.12.009.
695
[7] A. Pasandín, I. Pérez, Overview of bituminous mixtures made with recycled
concrete aggregates, Constr. Build. Mater. 74 (2015) 151–161, https://doi.org/
10.1016/j.conbuildmat.2014.10.035.
[8] J. Mills-Beale, Z. You, The mechanical properties of asphalt mixtures with
recycled concrete aggregates, Constr. Build. Mater. 24 (2010) 230–235, https://
doi.org/10.1016/j.conbuildmat.2009.08.046.
[9] A. Zulkati, Y. Wong, D. Sun, Mechanistic performance of asphalt-concrete
mixture incorporating coarse recycled concrete aggregate, J. Mater. Civ. Eng. 25
(9) (2013) 1299–1306.
[10] A. Pasandín, I. Pérez, Effect of ageing time on properties of hot-mix asphalt
containing recycled concrete aggregates, Constr. Build. Mater. 52 (2014) 284–
293, https://doi.org/10.1016/j.conbuildmat.2013.11.050.
[11] W. Gul, M. Guler, Rutting susceptibility of asphalt concrete with recycled
concrete aggregate using revised Marshall procedure, Constr. Build. Mater. 55
(2014) 341–349, https://doi.org/10.1016/j.conbuildmat.2014.01.043.
[12] S. Wu, B. Muhunthan, H. Wen, Investigation of effectiveness of prediction of
fatigue life for hot mix asphalt blended with recycled concrete aggregate using
monotonic fracture testing, Constr. Build. Mater. 131 (2017) 50–56, https://
doi.org/10.1016/j.conbuildmat.2016.11.045.
[13] A. Kavussi, M. Qazizadeh, Fatigue characterization of asphalt mixes containing
electric arc furnace (EAF) steel slag subjected to long term aging, Constr. Build.
Mater. 72 (2014) 158–166, https://doi.org/10.1016/j.conbuildmat.2014.08.
052.
[14] J. Xie, W. Shaopeng, J. Lin, J. Cai, Z. Chen, W. Wei, Recycling of basic oxygen
furnace slag in asphalt mixture: material characterization and moisture
damage investigation, Constr. Build. Mater. 36 (2012) 467–474.
[15] M. Skaf, J. Manso, Á. Aragón, J. Fuente-Alonso, EAF slag in asphalt mixes: a brief
review of its possible re-use, Resour. Conserv. Recycl. 120 (2017) 176–185,
https://doi.org/10.1016/j.resconrec.2016.12.009.
[16] J. Xie, W. Shaopeng, L. Zhang, Y. Xiao, W. Ding, Evaluation the deleterious
potential and heating characteristics of basic oxygen furnace slag based on
laboratory and in-place investigation during large-scale reutilization, J.
Cleaner Prod. 133 (2016) 78–87, https://doi.org/10.1016/j.jclepro.2016.05.
106.
[17] M. Zhao, W. Shaopeng, C. Zongwu, L. Chao, Function investigation of stone
mastic asphalt (SMA) mixture partly containing basic oxygen furnace (BOF)
slag, J. Appl. Biomater. Funct. Mater. 14 (Suppl. 1) (2016) S68–S72, https://doi.
org/10.5301/jabfm.5000307.
[18] M. Pasetto, N. Baldo, Experimental evaluation of high performance base course
and road base asphalt concrete with electric arc furnace steel slags, J. Hazard.
Mater. 181 (2010) 938–948, https://doi.org/10.1016/j.jhazmat.2010.05.104.
[19] M. Pasetto, Baldo, Laboratory investigation on foamed bitumen bound
mixtures made with steel slag, foundry sand, bottom ash and reclaimed
asphalt pavement, Road Mater Pavement Des. 13 (2012) 691–712, https://doi.
org/10.1080/14680629.2012.742629.
[20] M. Ameri, S. Hesami, H. Goli, Laboratory evaluation of warm mix asphalt
mixtures containing electric arc furnace (EAF) steel slag, Constr. Build. Mater.
49 (2013) 611–617, https://doi.org/10.1016/j.conbuildmat.2013.08.034.
[21] CEN. EN 12457-4, Characterization of Waste – Leaching – Compliance Test for
Leaching of Granular Waste Materials and Sludges – Part 4 – One Stage Batch
Test at a Liquid to Solid Ratio of 10 l/kg for Materials with Particle Size Below
10 mm (without or with size reduction), European Committee for
Standardization, Brussels, 2002.
[22] CEN. EN 1744-3, Tests for Chemical Properties of Aggregates – Part 3 –
Preparation of Eluates by Leaching of Aggregates, European Committee for
Standardization, Brussels, 2002.
[23] The Council of The European Union, Council Decision of 19 December 2002
establishing criteria and procedures for the acceptance of waste at landfills
pursuant to Article 16 of and Annex II to Directive 1999/31/EC, Official Journal
of the European Communities (2003). 16.1.2003.
[24] CEN. EN 12697-35, Bituminous Mixtures – Test Methods for Hot Mix Asphalt
– Part 35 – Laboratory Mixing, European Committee for Standardization,
Brussels, 2016.
[25] CEN. EN 12697-30, Bituminous Mixtures – Test Methods for Hot Mix Asphalt
– Part 30 – Specimen Preparation, Impact Compactor, European Committee for
Standardization, Brussels, 2012.
[26] F. Martinho, Misturas betuminosas temperadas com incorporação de
subprodutos industriais: caracterização e validação tecnológica (Warm mix
asphalt with incorporation of industrial by-products: characterisation and
technological validation), Universidade de Lisboa (in Portuguese), Instituto
Superior Tecnico, 2014. http://files.isec.pt/DOCUMENTOS/SERVICOS/BIBLIO/
Teses/Tese_Dout_Fernando-Martinho.pdf.
[27] CEN. EN 12697-22, Bituminous Mixtures – Test Methods for Hot Mix Asphalt
– Part 22 – Wheel Tracking, European Committee for Standardization,
Brussels, 2003/2007.
[28] CEN. EN 12697-26, Bituminous Mixtures – Test Methods for Hot Mix Asphalt
– Part 26 – Stiffness, European Committee for Standardization, Brussels, 2012.
[29] CEN. EN 12697-24, Bituminous Mixtures – Test Methods for Hot Mix Asphalt
– Part 24 – Resistance to Fatigue, European Committee for Standardization,
Brussels, 2012.
[30] CEN. EN 12697-12, Bituminous Mixtures – Test Methods for Hot Mix Asphalt
– Part 12 – Determination of the Water Sensitivity of Bituminous Specimens,
European Committee for Standardization, Brussels, 2008.
[31] CEN. EN 12697-23, Bituminous Mixtures – Test Methods for Hot Mix Asphalt
– Part 23 – Determination of the Indirect Tensile Strength of Bituminous
Specimens, European Committee for Standardization, Brussels, 2017.
696 F.C.G. Martinho et al. / Construction and Building Materials 185 (2018) 684–696
[32] J. Oliveira, H. Silva, L. Abreu, J. Gonzalez-Leon, The role of a surfactant based
additive on the production of recycled warm mix asphalts – less is more,
Constr. Build. Mater. 35 (2012) 693–700, https://doi.org/10.1016/
j.conbuildmat.2012.04.141.
[33] H. Silva, J. Oliveira, J. Peralta, J. Zoorob, Optimization of warm mix asphalts
using different blends of binders and synthetic paraffin wax contents, Constr.
Build. Mater. 24 (2010) 1621–1631, https://doi.org/10.1016/
j.conbuildmat.2010.02.030.
[34] H. Silva, J. Oliveira, C. Ferreira, P. Pereira, Assessment of the performance of
warm mix asphalts in road pavements, Int. J. Pavement Res. Technol. 3 (3)
(2010) 119–127.
[35] Bolzan P, Balige M, Mezclas asfalticas en caliente elaboradas y compactadas a
menores temperaturas, in: XXXVI Reunión del Asfalto Bicentanário de los
Andes. Buenos Aires, Argentina, 2010.
[36] E. Toraldo, C. Brovelli, E. Mariani, Laboratory investigation into the effects of
working temperatures on wax-based warm mix asphalt, Constr. Build.
Mater. 44 (2013) 774–781, https://doi.org/10.1016/j.conbuildmat.2013.03.
085.
[37] E. Sanchez-Alonso, A. Vega-Zamanillo, D. Castro-Fresno, M. DelRio-Prat,
Evaluation of compactability and mechanical properties of bituminous mixes
with warm additives, Constr. Build. Mater. 25 (5) (2013) 2304–2311, https://
doi.org/10.1016/j.conbuildmat.2010.11.024.
[38] S. Redondo, J. León, A. González, Sánchez E, Estudio del ángulo de contacto de
betunes aditivados y su efecto sobre distintas propiedades mecánicas de las
mezclas asfálticas, Revista Carreteras (2011) 178.
[39] Oliveira J, Silva H, Fonseca P, Kim Y, Hwang S, Pyun J, Lee H, Laboratory and
field study of a WMA mixture produced with a new temperature reduction
additive, in: Proceedings of the 2nd International Warm-mix Conference, St.
Louis, USA, 2012.
[40] M. Zaumanis, L. Poulikakos, M. Partl, Performance-based design of
asphalt mixtures and review of key parameters, Mater. Des. 141
(2018) 185–201.