Construction and Building Materials 43 (2013) 497–505

Contents lists available at SciVerse ScienceDirect

Construction and Building Materials

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

Laboratory evaluation of hot-mix asphalt containing construction

and demolition waste
A.R. Pasandín ⇑, I. Pérez
Universidade da Coruña, E.T.S.I. Caminos, Canales y Puertos, Campus de Elviña s/n, 15071 A Coruña, Spain

h i g h l i g h t s

 We studied hot-mix asphalt containing 0%, 5%, 10%, 20% and 30% of recycled concrete aggregate in place of virgin aggregate.
 To improve their water sensitivity the mixes were cured in an oven for 4 h before compaction at mixing temperature.
 Marshall mix design procedure was used.
 Optimized mixes comply with Spanish moisture damage specifications for base course.
 Resilient modulus, permanent deformation and fatigue life of optimized mixes also showed good results.

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

Article history: A laboratory study of the use of recycled concrete aggregates (RCAs) from construction and demolition
Received 30 October 2012 waste (CDW) in hot-mix asphalt (HMA) for base courses in pavements was conducted. HMA mixes con-
Received in revised form 16 January 2013 taining 0%, 5%, 10%, 20% and 30% RCA in place of natural aggregate were evaluated. The Marshall mix
Accepted 26 February 2013
design procedure was used to develop the mixes. To improve the moisture sensitivity of the mixes, they
Available online 2 April 2013
were cured in an oven for 4 h. The results indicated that the mixes comply with Spanish moisture damage
specifications. The mechanical properties (stiffness, permanent deformation and fatigue) of the mixtures
Keywords:
were studied. The mixtures exhibited good engineering properties. Although HMA with RCA requires fur-
Hot-mix asphalt
Recycled concrete aggregate
ther investigation, the results from this study were very promising.
Mechanical properties Ó 2013 Elsevier Ltd. All rights reserved.

1. Introduction neous [6,10,11,21,23]. The tiny fissures that appear during the
crushing process [12] and the weak contact between the mortar
Concerns in recent decades about achieving sustainable growth and aggregate [22] are other factors to take into account. All of
has resulted in attempts to recycle or reuse a large portion of con- these reasons make RCA from CDW a poorer-quality aggregate
struction waste materials. This is the case for recycled concrete than natural aggregate [22,24]. Obviously, differences between
aggregate (RCA) from construction and demolition waste (CDW). the properties of RCA and those of natural aggregates are going
Its use has been growing in recent years, particularly as aggregate to influence the performance of HMA made with RCA. It is also ex-
for concrete [1,2] and unbound pavement layers [3–5]. However, pected that the RCA content (between 0% and 100%) will affect
further research is needed to diversify RCA applications and make HMA performance.
its use a habitual practice. It is particularly interesting to note that some studies have
Although RCA is most likely suitable for use as aggregate in hot- indicated that HMA mixes made with RCA have higher moisture
mix asphalt (HMA) for pavements in road building, to date only a sensitivity than those made with natural aggregates only
few studies have been carried out dealing with the use of this type [10,11,15,21,23]. Moisture sensitivity or moisture damage is a
of waste material in HMA [6–23]. deterioration process that affects HMA and is defined as the degra-
Many researchers have noted that the attached mortar (Fig. 1a), dation of mechanical properties of the mixture due to the presence
which is more porous and less dense than crushed stone, seems to of water [25]. There are many ways to improve the moisture sen-
be the principal reason for RCA being of unsatisfactory quality sitivity of conventional HMA, including the use of antistripping
[6,10–12,14,22–24], but it is not the only reason. In fact, some additives, the addition of selected fillers, avoidance of hydrophilic
studies have recommended removing impurities such as wood, aggregates, etc. [26]. Previous research conducted with the same
rubber, and gypsum with the aim of making RCA more homoge- RCA used in this investigation [18,21,23] recommended allowing
the HMA sufficient time at high temperature to complete the bin-
⇑ Corresponding author. Tel.: +34 981167000; fax: +34 981167170. der absorption by the aggregates. This could be a way to improve
E-mail addresses: arodriguezpa@udc.es (A.R. Pasandín), iperez@udc.es (I. Pérez). the moisture sensitivity of HMA made with RCA, since the bitumen

0950-0618/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.conbuildmat.2013.02.052
498 A.R. Pasandín, I. Pérez / Construction and Building Materials 43 (2013) 497–505

(a) Attached mortar (b) Gypsum impurities

CRUSHED
STONE

ATTACHED MORTAR

Fig. 1. RCA used in this investigation: (a) attached mortar and (b) gypsum impurities.

absorbed by RCA makes the whole aggregate surface be coated by is polishable and has a poor affinity with bitumen, so it would be wise to remove
it before using RCA in HMA production. The natural aggregate used was a hornfels
the binder, leaving no fissures through which water could pene-
that was supplied by a local contractor and is typically used in HMA production in
trate. Furthermore, it reduces the porosity decreasing at the same Spain. The compositions of the natural and recycled aggregates were determined
time water accessible voids. using X-ray fluorescence tests. The results indicated that the RCA (61.46% SiO2)
Moisture sensitivity is not the only HMA property that could be and hornfels (62.30% SiO2) are siliceous aggregates. Consequently, both of them
affected by the use of CDW as recycled aggregate. Some research- were expected to exhibit poor stripping performance.
The RCA and natural aggregate properties were evaluated according to the Span-
ers have stated that the use of RCA in HMA production leads to
ish General Technical Specifications for Roads, also known as PG-3 [27]. The results
mixes with less stiffness than conventional mixes [10,15], while (shown in Table 1) indicate that the RCA had a lower bulk specific gravity (qa) than
others suggest the opposite [8,11]. The literature on the permanent the natural aggregate as well as a higher water absorption (W24). This is due to the
deformation of HMA made with RCA yields varied conclusions. mortar on the RCA surface. The sand equivalent (SE) values of the RCA and the natural
aggregate complied with the PG-3 for HMA as a base course material. The RCA’s Los
Most researchers consider mixtures made with RCA to comply
Angeles (LA) abrasion coefficient only complied with the PG-3 for HMA as a base
with specifications related to permanent deformation course material in low-volume roads in heavy traffic category T4. The LA abrasion
[9,11,15,18,21,23] and exhibit permanent deformation behavior coefficient of the hornfels aggregate complied with the PG-3 for HMA as a base course
similar to conventional mixes [10] or better [7,8,13,20]. However, material in roads in heavy traffic category T00. The LA abrasion coefficient of a mix of
other researchers indicate that despite meeting the specifications, RCA and natural aggregate was also determined. The results showed that for mixes of
0%, 5%, 10%, 20% and 30% replacement of natural aggregate by RCA, the combined
permanent deformation performance worsens as the percentage
(RCA + natural) LA abrasion coefficient complied with the PG-3 (LA < 25%) for HMA
of RCA in the mix increases [15]. Others indicate that permanent as a base course material in roads in heavy traffic category T00.
deformation performance is also influenced by the mixture grada-
tion [14]. The fatigue life of HMA containing RCA has been studied 2.1.2. Binder and filler
less. The studies to date indicate that HMA with RCA are similar in A B50/70 binder from Venezuela was chosen for use in this study. Its engineer-
fatigue life to conventional mixtures [11,17]. There is one excep- ing properties are presented in Table 2. Gray Portland cement (CEM II/B-M (V-L)
32.5 N) was obtained from a commercial source for use as mineral filler. Its Blaine
tion: when RCA is used as filler, fatigue life improves [20].
surface area was equal to 3134 cm2/g and its specific gravity was equal to 3.10 g/
This paper presents an investigation of the mechanical proper- cm3.
ties of HMA made with RCA from CDW for base courses. The aim of
the investigation is to design HMA with RCA and achieve good 2.2. Testing program
mechanical properties and good moisture sensitivity performance.
Following the example of previous studies [18,21,23], to improve 2.2.1. Marshall mix design
the moisture sensitivity of the asphalt mixes used, they were cured The Marshall mix design procedure, as specified in NLT-159/86 [28], was used
in this investigation. The laboratory mixing temperature was 170 °C and the com-
in an oven at the mixing temperature for 4 h after mixing and be-
paction temperature was 165 °C. Percentages of 0%, 5%, 10%, 20% and 30% of RCA
fore compaction. This made it possible for, the aggregate, particu- by weight of total aggregate were studied. The aggregate gradation, an AC 22 base
larly the RCA, to absorb a greater amount of bitumen. Not only G (Fig. 2), was chosen in accordance with the PG-3. The aggregate gradation had a
improved mortar resistance but also less water absorption, and
thus better moisture damage performance, is expected as a result. Table 1
Characterisation of aggregates.
The moisture sensitivity, stiffness, permanent deformation and fa-
tigue life of HMA mixes containing RCA were studied. Aggregate Standard RCA Hornfels PG-3 Specificationsa
T00–T1 T3–T2 T4
2. Materials and methods qa (g/cm3) EN-1097-6 2.63 2.79 – – –
WA24 (%) EN 1097-6 5.08 1.08 – – –
2.1. Basic materials SE (%) EN 933-8 67 61 P50 P50 P50
LA abrasion (%) EN 1097-2 32 14.1 625 630 –
2.1.1. Aggregates
Two types of aggregates were used: RCA and natural aggregate. RCA was ob- Traffic category T0 refers to 4000 > AADHT P 2000.
tained from demolition waste from residential buildings of different origins and Traffic category T1 refers to 2000 > AADHT P 800.
qualities in Madrid (Spain) and was supplied by a CDW recycling plant. Aggregate, Traffic category T2 refers to 800 > AADHT P 200.
concrete, and similar materials constituted 89.3% of the mass of the RCA obtained Traffic category T3 refers to 200 > AADHT P 50.
for use in this study, and bituminous materials constituted 6.5%. The remainder Traffic category T4 refers to AADHT < 50.
a
of the constituents were impurities (ceramics, wood, rubber, gypsums, etc.) that Traffic category T00 refers to AADHT (Annual Average Daily Heavy
could introduce variation in the test results. Gypsum (Fig. 1b) is an impurity that Traffic) P 4000.
 A.R. Pasandín, I. Pérez / Construction and Building Materials 43 (2013) 497–505 499

Table 2
Properties of asphalt cement.

Test Standard B50/70 PG-3 specification

Original
Penetration (100 g, 5 s, 25 °C), 0.1 mm NLT-124 52 50–70
Softening point, °C UNE-EN 1427 54.9 48–57
Flash point, °C ISO 2592 >290 >235
Density (25 °C), g/cm3 NLT-122 1.009 >1.0
After rolling thin-film oven test
Penetration (100 g, 5 s, 25 °C), 0.1 mm NLT-124 68 >50
D Softening point, °C NLT-125 6.5 69

2.2.3. Stiffness
Gradation curve of a AC 22 base G The indirect tensile stiffness modulus test (ITSM) was used to determine the
100
resilient modulus of the HMA mixes in accordance with UNE-EN 12697-26 Annex
90 C [31], using a Cooper NU 14 tester. In each test, ten conditioning haversine pulses
were applied along the vertical diameter of a cylindrical specimen. Afterward, five
80 Upper limit
haversine test pulses were applied. The repetition period of the impulse was
70 Lower limit 3 ± 0.1 s. The maximum load was selected to achieve a maximum horizontal strain
Percent passing

60 Selected gradation curve of 0.005% of the specimen diameter. The rise time was 124 ± 4 ms. The average stiff-
ness modulus of a specimen was determined from the five test pulses. After rotating
50
the specimen 90°, the test sequence was repeated. The average stiffness from the
40 two tested diameters was recorded as the stiffness modulus of the HMA specimen.
The specimens were cured in an oven at 170 °C for 4 h before compaction. Mar-
30
shall specimens were formed (at 75 blows on each side) at the optimum binder con-
20 tent for each RCA percentage. The indirect tensile modulus of the specimens was
10 tested in a controlled-temperature cabinet at temperatures of 0 °C, 10 °C and 20 °C.
The resilient modulus was determined using the following equation:
0
0.063 0.25 0.5 2 4 8 16 22 32 F  ðm þ 0:27Þ
MR ¼ ð2Þ
Sieve size (mm) zh

Fig. 2. Gradation curve. where MR = the resilient modulus (MPa), F = the maximum applied load (N), z = the
horizontal deformation (mm), h = the sample thickness (mm) and m = Poisson’s ratio
(a Poisson’s ratio of 0.35 was assumed for the HMA mixes for all of the test
maximum aggregate size of 22 mm and a 4% mineral filler content. It is important to
temperatures).
take into account that RCA fine fraction is more sensitive to adhesion phenomena
Spanish specifications [27] require a minimum resilient modulus of 11,000 MPa
and exhibits greater water absorption. Furthermore, the smaller the size of RCA par-
at 20 °C in the case of high-modulus HMA. However, there are no requirements for
ticles is, the higher the content of mortar cement is, which negatively affects the
conventional mixtures.
properties of RCA [24]. Also, it is more difficult to remove impurities in the fine frac-
tion. For these reasons, to manufacture HMA, it was considered more suitable to
make the replacement of RCA in the coarse fractions 8/16 mm (RCA replacement 2.2.4. Permanent deformation
of 5%, 10%, 20% and 30%) and 4/8 mm (RCA replacement of 30%). Substitution at lar- To evaluate the resistance of the mixes to permanent deformation, repeated-
ger fractions could introduce greater dispersions in the results due to the heteroge- load axial testing (RLAT) without confinement was conducted, in accordance with
neity of the material. DD 226:1996 [32], using a Cooper NU 14 tester. In the test, axial square pulses
After mixing, the loose mixtures were cured in an oven at 170 °C for 4 h. Cylin- are applied to cylindrical samples. The same Marshall specimens used in the resil-
drical specimens 101.6 mm in diameter and 63.5 mm in height were formed using ient modulus tests were used in the repeated-load axial tests. Each specimen was
the Marshall compactor (with 75 blows on each side). In order to determine the held at a test temperature of 30 °C in a controlled-temperature cabinet overnight
optimum asphalt content (Bo), the bitumen percentage was selected, inside limits, and then was placed between two load platens. A preload of 10 kPa of axial stress
that allow to achieve the maximum Marshall stability and thus, the highest traffic was applied for 600 ± 6 s. The sample was then subjected to 1800 load applications.
category possible. Also, flow (F), air voids (Va) and voids in the mineral aggregate These tests were performed under the following load conditions: an axial stress of
(VMA) were selected in accordance with the PG-3 specifications. RCA is a porous 100 ± 2 kPa, a load application period of 1 s and a rest period of 1 s. The following
aggregate, so it is interesting to determine not only the optimum asphalt content equation was used to calculate the axial permanent strain:
but the effective binder content (Pbe) and the absorbed bitumen content (Pba) [29].
Dh
edðn;TÞ ¼ ð3Þ
h0
2.2.2. Moisture damage
UNE-EN 12697-12 [30] was used to evaluate the moisture sensitivity (stripping where ed(n,T) = the axial permanent strain (in le) after n load applications at temper-
potential) of HMA made with RCA by measuring the loss of indirect tensile strength, ature T in °C, h0 = the initial distance between the two load platens (mm) and h = the
expressed in terms of the tensile strength ratio (TSR). In this test, a set of cylindrical axial deformation (mm).
samples is subdivided into two subsets with the same number of specimens in each
subset. One subset was kept dry at room temperature, while the other subset was 2.2.5. Fatigue
saturated and was held in a water bath for 3 days at 40 °C. After that time, the two Fatigue life was evaluated using the repeated indirect tensile fatigue test (ITFT),
subsets were left a minimum of 2 h at 15 °C, the dry specimens in air and the wet in accordance with UNE-EN 12697-24 Annex E [33], using a Cooper NU 14 tester.
specimens in water. The tensile strength of the specimens in each subset was then This is a controlled-stress fatigue test in which cylindrical specimens are subjected
determined. to repeated haversine loads along a vertical diameter. Controlled-stress tests are
Five sets (0%, 5%, 10%, 20% and 30% RCA) of ten samples each were evaluated. adequate to evaluate the fatigue life of thick bituminous pavement layers with high
The samples were produced at the optimum bitumen content for each RCA percent- stiffnesses [34]. Marshall specimens at the optimum bitumen content were com-
age. The specimens were left in an oven at 170 °C for 4 h after mixing. Marshall pacted at 75 blows per side after 4 h in an oven. For every RCA content, a minimum
specimens were formed (with 50 blows on each side). Moisture sensitivity was of three specimens were tested at each of three constant tension levels (350 kPa,
evaluated as follows: 450 kPa and 550 kPa were chosen). The tests were conducted at a reference temper-
ature of 20 °C in a controlled-temperature cabinet. The Whöler equation was used
ITSw to obtain regression equations for fatigue life:
TSR ¼  100 ð1Þ
ITSD
e0 ¼ k  Nf n ð4Þ
where TSR = the tensile strength ratio (%), ITSW = the average tensile strength of five
conditioned (wet) specimens (MPa) and ITSD = the average tensile strength of five where Nf = the number of load cycles to fatigue failure, k and n = material constants
unconditioned (dry) specimens (MPa). TSR P 80% is required by PG-3 specifications obtained from the ITFT and e0 = the initial tensile horizontal strain at the center of
[27] for HMA for use in base courses. the specimen in le.
500 A.R. Pasandín, I. Pérez / Construction and Building Materials 43 (2013) 497–505

A combination of permanent deformation and fatigue mechanisms occur during absorption. The values obtained for VMA indicate that not consid-
the ITFT [35]. Thus, two failure criteria were used to determine the fatigue life of a
ering asphalt absorption overestimates this parameter. When
specimen: the total number of load cycles to complete splitting of the specimen
along the vertical plane or the total number of cycles to 10% vertical deformation
using RCA in HMA, this consideration could be of great importance,
of the specimen along the vertical plane, whichever came first. particularly if a high percentage of RCA is used.
As shown in Table 3, according to PG-3, all the samples tested
contain air voids and voids in mineral aggregate suited to meet a
3. Test results and discussion heavy traffic category T00 [27,36].
Table 3 shows that the same occurs with VFA: not considering
3.1. Marshall mix design asphalt absorption overestimates VFA. If absorbed bitumen is ta-
ken into account, VFAabs decreases as the RCA content increases,
The optimum asphalt content (Bo), effective binder content reflecting the absorptivity of RCA.
(Pbe) and bitumen absorption content (Pba) of the HMA mixes The stability (S) values are very high: greater than 15 kN in all
are shown in Table 3, along with other mix properties, such as cases. This means that HMA made with RCA in percentages ranging
the unit weight (UW), air voids (Va), voids in the mineral aggregate from 0% to 30% comply with PG-3 specifications for traffic category
with and without consideration of binder absorption (VMAabs and T00 [36].
VMA, respectively), voids filled with asphalt with and without con- Most of the results indicate an increasing trend in flow (F) as the
sidering binder absorption (VFAabs and VFA, respectively), stability RCA percentage increases. Thus, it can be concluded that an in-
(S) and flow (F). crease in RCA content produces mixtures that are more susceptible
As Table 3 shows, a general pattern exists among the Bo, Pba, to deformation. However, it should be noticed that F values are in-
Pbe and RCA contents: Bo and Pba increase with increasing RCA side the range specified for traffic category T00 by the PG-3 [36].
content, while Pbe decreases. On the one hand, this indicates that The Marshall S/F ratio may be used as an indicator of stiffness:
the porosity of the mortar attached to the RCA makes the RCA ab- mixtures with high S/F ratios are more resistant to permanent
sorb binder in proportion to the percentage of RCA in the HMA. deformation. Although it can be observed that the S/F ratios ob-
Thus, HMA containing RCA requires more binder as the RCA con- tained do not follow a clearly defined pattern, it seems that
tent increases. On the other hand, the decreasing trend in Pbe with increasing the RCA content decreases the S/F ratio, so a general
increasing RCA indicates that the bitumen film thickness decreases conclusion can be drawn that mixtures with higher RCA content
with increasing RCA content, which could affect HMA performance, will perform worse in terms of permanent deformation.
especially for high RCA contents. However, there are exceptions to
this general pattern. For example, 20% RCA is an exception to this 3.2. Moisture damage
trend. It could be explained by the heterogeneity of RCA used in
this investigation. As said before, the RCA used in this investigation The results of the moisture sensitivity tests for samples made
was obtained from the demolition of residential buildings, hence, it with optimum binder content are shown in Fig. 3. On the one hand,
has particles of other materials such as gypsum, wood, and asphalt. as the results show, a certain pattern exists: in general, ITS de-
Moreover, it has concrete from different origins and thus, with dif- creases as RCA increases. This is true in both the dry and wet states.
ferent qualities. There is only one exception: the samples with 5% of RCA did not
As was expected and Table 3 shows, UW decreases as RCA in- follow this trend. This exception could be explained by the heter-
creases. This is because RCA is less dense than natural aggregate. ogeneity characteristics of the RCA used. On the other hand, all
Furthermore, the results show that Va exhibits a slight increase of the TSR values obtained were greater than 80%, which is the
as the RCA content increases. Va is determined as a function of unit requirement established in the Spanish specifications for HMA in
weight, so a decrease in UW leads to an increase in Va. However, an base courses. However, it is surprising that the worst TSR value
increase in Va can be explained by the difficulty of compacting was measured for an RCA percentage of 0%. This suggests that
HMA containing RCA, due to the roughness of the attached mortar. although the moisture sensitivity of RCA is not good, it is better
The results in Table 3 also show that when asphalt absorption is than the moisture sensitivity of the natural aggregate used. This
not considered, VMA increases slightly as the RCA percentage in- is consistent with the results obtained from the X-ray fluorescence
creases. This result is obvious because Bo increases as well, and tests: as the SiO2 percentage increases, the stripping performance
not considering asphalt absorption implies that all of the bitumen worsens. As mentioned previously, the literature review indicates
is part of VMA. In contrast, if asphalt absorption is considered, different results for the stripping performance of HMA made with
VMAabs decreases slightly as the RCA percentage increases. This RCA. This discrepancy can be explained by the fact that the strip-
means that an increase in RCA content produces greater asphalt ping performance of HMA made with RCA depends not only on

Table 3
HMA properties at optimum asphalt contents.

HMA oven-cured for 4 h at 170 °C before compaction PG-3 specifications

0% RCA 5% RCA 10% RCA 20% RCA 30% RCA T00–T0 T1–T2 T3 T4
Bo (%) 4.00 3.80 3.90 4.40 4.30 >3.5
Pbe (%) 3.24 2.86 2.85 2.95 2.52 –
Pba (%) 0.79 0.97 1.09 1.51 1.86 –
Va (%) 7.25 7.57 7.27 7.87 7.71 5–8 6–9 5–9 –
VMA (%) 16.84 16.66 16.54 18.14 17.71 P14
VMAabs (%) 14.98 14.45 13.96 14.75 13.61 –
VFA (%) 56.82 53.96 55.85 56.59 56.44 –
VFAabs (%) 51.41 46.94 47.76 46.61 43.21 –
S (kN) 17.31 19.41 16.83 16.78 17.66 >15 >12.5 >10 8–12
F (mm) 2.02 2.22 2.51 2.14 2.54 2–3 2–3.5 2.5–3.5
UW (t/m3) 2.41 2.40 2.40 2.35 2.34 –
S/F (kN/mm) 8.57 8.75 6.71 7.85 6.96 –
 A.R. Pasandín, I. Pérez / Construction and Building Materials 43 (2013) 497–505 501

102.5 96.5 98.7

only PG-3 specification concerning resilient modulus is that high-
2.50 100 modulus HMA must have a minimum resilient modulus of
2.40 11,000 MPa at 20 °C. As Fig. 4 shows, the HMA mixtures produced
94.2
2.30 with optimum binder content and with 10–30% RCA almost satisfy
82.8 80
ITSd
the 11,000 MPa requirement that would classify them as high-
2.20
ITSw
modulus HMA mixtures.
2.10
ITS (MPa)

60 Thus, despite HMA mixtures made with RCA not meeting the

TSR (%)
TSR
2.00 requirement to be classified as high-modulus HMA mixtures, they
1.90 40 are nonetheless very stiff mixtures. High-resilient-modulus mix-
1.80 tures are stiffer and have longer fatigue lives when they are used
in thick layers [34]. However, at low temperatures, high-resil-
1.70 20
ient-modulus mixtures are more susceptible to thermal and fati-
1.60 gue cracking [37]. Thus, mixtures used in surface courses with
1.50 0 elastic moduli above 3100 MPa at 20 °C must be used with caution
0 5 10 20 30
in cold regions [37]. As Fig. 4 shows, all the mixtures tested in this
RCA (%)
investigation had resilient modulus values much higher than
Fig. 3. Moisture sensitivity test results. 3100 MPa at 20 °C. Although the mixes studied are intended for
use in base courses and not surface courses, the results indicate
that HMA with RCA is suitable for surface course use in temperate
the type of RCA used (origin, impurities, crushing process, etc.) but regions. However, the high stiffness values obtained confirm that
also on the affinity of the natural aggregate used with bitumen. the mixtures can be used in thinner layers while maintaining their
A two-way analysis of variance (ANOVA) was conducted to structural integrity [15] if they are used in temperate regions. HMA
determine the effect of RCA percentage (0%, 5%, 10%, 20% and used in Spain has a typical mean resilient modulus of 5000 MPa at
30%) and state (dry or wet) on indirect tensile strength (ITS). The 20 °C [38]. The resilient modulus values of the mixes studied were
model explains 89.88% of the total variance. The RCA percentage higher than 5000 MPa, which indicates that the studied mixtures
is statistically significant at the 95% confidence level are stiffer than conventional mixtures. Fig. 4 shows that the resil-
(p = 0.039 < 0.05). The other main effect, i.e., dry or wet state, is ient modulus of HMA made with RCA is strongly dependent on
not significant (p = 0.221). On the basis of the ANOVA, it can be de- the test temperature. As expected, the mixtures were stiffer at low-
duced that the most influential factor in ITS is the RCA percentage. er temperatures. In Spain, a winter resilient modulus 1.5 times
That the wet or dry state is not statistically significant means that higher than the mean value is often adopted [38]. The resilient
the applied treatment reduces the effect of water on the indirect modulus of HMA with RCA 0 °C is 2.1–4.9 times higher than that
tensile strength. It should be noted that at 5% RCA, the wet strength at 20 °C, while at 10 °C, the resilient modulus is 1.5–3.7 times high-
is higher than the dry strength. er. Therefore, the resilient modulus of HMA made with RCA exhib-
A one-way ANOVA was conducted to determine the effect of its much more dependence on in-service temperatures than
RCA percentage (0%, 5%, 10%, 20% and 30%) on the tensile strength conventional mixtures. Moreover, there is no clear pattern be-
ratio (TSR). The ANOVA results indicate that the RCA percentage tween RCA content and resilient modulus, due to the heterogeneity
has a significant effect on TSR (p = 0.000). Thus, a variation in the of the RCA used. Only at 20 °C does the resilient modulus seem to
RCA content will influence the stripping performance of HMA. increase with increasing RCA content. However, this is only true for
RCA contents up to 20%.
3.3. Stiffness A two-way ANOVA was conducted to determine the effect of
RCA percentage (0%, 5%, 10%, 20% and 30%) and test temperature
As stated previously, there are no Spanish specifications for the (0 °C, 10 °C and 20 °C) on stiffness (MR). The model explains
acceptance of conventional mixtures in terms of resilient modulus 97.85% of the total variance. The test temperature is statistically
(MR), so it is useful to compare the resilient modulus results ob- significant at the 95% confidence level (p = 0.000 < 0.05). As ex-
tained for the different mixtures tested in this investigation. The pected in view of the test results obtained, the ANOVA confirms

30,000
26,673.5

22,857.5 24,117.5
25,000 22,277.5

23,407.0 20,167.0
20,000
MR (MPa)

17,425.0
15,632.0 0ºC
18,001.5
15,000 10ºC
15,492.5
20ºC
9,517.5
10,000
10,826.5 10,527.0
10,316.0
9,532.0
5,000
0 5 10 15 20 25 30

RCA (%)

Fig. 4. Resilient modulus.

502 A.R. Pasandín, I. Pérez / Construction and Building Materials 43 (2013) 497–505

11,000

10,000

9,000
Axial permanent strain (με)
8,000

7,000

6,000
0% RCA
5,000 5% RCA

10% RCA
4,000
20% RCA

3,000 30% RCA

2,000
0 200 400 600 800 1,000 1,200 1,400 1,600 1,800
Number of cycles at 30ºC

Fig. 5. Accumulated permanent deformation through 1800 cycles.

that the other main effect, RCA percentage, is not significant cling, the mixtures exhibit rapid densification. Although the slopes
(p = 0.0759). On the basis of the ANOVA, it can be deduced that are very similar for all the tested percentages of RCA, the different
the dominant factor in specimen stiffness is test temperature. amounts of this rapid initial densification in the different mixes are
largely responsible for the differences in the final permanent defor-
3.4. Permanent deformation mation at 1800 load cycles. It is expected that higher bitumen con-
tents produce mixtures with greater susceptibility to plastic
Results of the permanent deformation tests indicate the rutting deformation [7]. In the case of HMA made with RCA, as the RCA
potential of the mixtures. Fig. 5 shows the accumulated permanent content increases, the effective bitumen content decreases, so the
deformation values versus the number of loading cycles. As ex- increase in the permanent deformation slope is mainly due to
pected, permanent deformation increases with the number of cy- the increase in RCA content. These results are consistent with the
cles for mixtures made with the optimum binder content. There Marshall S/F ratio results. For all of the RCA percentages tested,
is no clear pattern between the final permanent deformation at cy- the HMA mixtures exhibited ultimate permanent deformation lev-
cle 1800 and the RCA percentage. It could be explained by the het- els lower than those that conventional mixtures exhibit. For exam-
erogeneity of the RCA used in this investigation. It is well known ple, Santagata et al. [41] obtained values of the final strain between
that HMA permanent deformation occurs mainly due to densifica- 4000 and 11,000 le at 1800 load cycles for HMA made with vari-
tion (a decrease in the air void content causing a volume change) ous binders, while Aschuri et al. [42] obtained values of approxi-
and plastic flow (aggregates and binder being gradually moved mately 13,000 le for conventional mixtures. As Fig. 5 shows, all
without a volume change) [39]. The test results indicate that per- of the studied mixtures had final axial permanent strain values
manent deformation of HMA made with RCA occurs mainly due lower than 11,000 le at 1800 cycles. Thus, mixtures with RCA that
to densification. Plastic flow did not occur, so the specimens did have been oven-cured for 4 h will perform well against rutting.
not fail during the tests. Fig. 5 also shows that after a given number A one-way ANOVA was conducted to determine the effect of
of load applications, a linear relationship exists between the axial RCA content on permanent deformation at 1800 load cycles. The
permanent strain and the number of load cycles. The slope of the ANOVA results show that the RCA content has a significant effect
line reflects the trend of axial deformation, such that larger slopes on permanent deformation (p = 0.000).
indicate less resistance to permanent deformation [40]. As Table 4
shows, the slopes of the lines between cycles 600 and 1800 are 3.5. Fatigue
very similar, but indicate a certain pattern between RCA content
and permanent deformation: in general, when RCA content in- As stated previously, the ITFT is suitable for high-stiffness mix-
creases, so does permanent deformation. However, the 30% RCA tures used in thick layers. Thus, it is a suitable test for simulating
mix did not follow this pattern. At the beginning of the load cy- the fatigue performance of the high-resilient-modulus mixtures
examined in this investigation.
Table 4 A fatigue test ends when a sample is fractured by a diametral
Slopes between cycles 600 and 1800 in permanent plane and split into two parts. None of the tested specimens failed
deformation curves.
in terms of the alternate criterion, a vertical deformation of 10% of
RCA (%) Slope the specimen diameter. Therefore, it can be concluded that fatigue
0 0.29 mechanisms were predominant in the failure of all the HMA sam-
5 0.44 ples tested (Fig. 6).
10 0.46 Fig. 7 shows the initial horizontal strain versus the number of
20 0.52
cycles to failure at 20 °C on a logarithmic scale for the mixtures
30 0.37
made with the optimum asphalt content. The same figure shows
 A.R. Pasandín, I. Pérez / Construction and Building Materials 43 (2013) 497–505 503

vary a great deal for the different fatigue equations, which might
((a) S
Sam
mplees m
mad
de w
with
h 0%
% RCA
R A be due to the heterogeneity of the RCA.
A fatigue equation for the control mixture (0% RCA and without
curing oven time) has been plotted for comparison with the fatigue
equations obtained for the mixes tested in this study. In general,
HMA made with RCA that has been left in the oven for 4 h exhibits
similar fatigue life to control mixture, except the mixture made
with 30% RCA that has a fatigue resistance slightly slower than
conventional. Moreover, the slopes of the equations for the mixes
with RCA percentages between 0% and 20% are flatter than the
slopes of the equations for the control mixture, indicating that
HMA containing up to 20% RCA has a greater fatigue life than con-
ventional mixtures at higher number of cycles. For HMA containing
30% RCA, the slope of the fatigue equation is similar to that of the
conventional mixture.

4. Final remarks

The results of the tests conducted in this investigation are

encouraging. Bituminous mixtures made with RCA from CDW were
shown to exhibit good moisture sensitivity and more than accept-
able mechanical properties. With adequate pretreatment and with
a limited RCA content, HMA made with RCA can be used in warm
regions. The mixtures should be oven-cured for 4 h before compac-
tion, and a maximum of 30% of RCA appears to be adequate.
The use of HMA made with RCA promotes sustainable construc-
tion by providing numerous environmental benefits, such as re-
duced extraction of natural aggregates from quarries, and
avoidance of the visual impact of landfills, avoidance of rejection
of raw materials.
However, there are some disadvantages to the use of RCA in
HMA. The cost of removal of impurities (e.g., gypsum) is one disad-
vantage. There are also environmental disadvantages, such as an
(b
b) Sam
S mplees m
mad
de with
w h 30% R
RC
CA
increase in bitumen consumption with increasing RCA content
and the associated increase in energy consumption. Limiting the
Fig. 6. ITFT failure patterns: (a) samples made with 0% RCA and (b) samples made RCA content to 30% produces a maximum increase of 0.5% in the
with 30% RCA. bitumen content, which is an acceptable value. Further investiga-
tion is needed to implement a storage system in asphalt plants that
the fatigue equations of the tested mixtures. It is interesting to maintains the temperature of such mixtures without increased
note that mixtures with low RCA percentages (up to 20%) perform consumption of fossil fuels.
similarly to each other in fatigue, especially at high deformation
levels, while the fatigue performance of HMA with 30% RCA is a lit- 5. Conclusions
tle poorer. This is highlighted by the fact that the fatigue equation
for this mixture has a much more pronounced slope than the other HMA made with 0%, 5%, 10%, 20% and 30% RCA in the 4/8 mm
fatigue equations. However, the coefficients of determination (R2) and/or 8/16 mm coarse fractions of the aggregate gradation have

0% RCA 5% RCA 10% RCA 20% RCA 30% RCA

εo = 3,428.8xNf-0.229 εo = 5,278.8xNf-0.274 εo = 2,930.6xNf-0.214 εo = 3,526.5xNf-0.227 εo = 10,051xNf-0.35
R² = 0.9129 R² = 0.7519 R² = 0.7972 R² = 0.8814 R² = 0.7723
1,000
ε 0 (με)

100
0% RCA
5% RCA
10% RCA
20% RCA
30% RCA CONTROL MIXTURE
Control mixture εo = 10,279xNf-0.319
R² = 0.9161
10
1,000 10,000 100,000 1,000,000
Cycles to failure (Nf)

Fig. 7. Fatigue life.

504 A.R. Pasandín, I. Pérez / Construction and Building Materials 43 (2013) 497–505

been studied. To improve their moisture sensitivity, the mixtures The authors would like to thank Nynas Bitumen for supplying
were oven-cured for 4 h after mixing and before compaction at the binder used in this study.
170 °C. The following conclusions are drawn from this research:
References
 RCA has a high LA abrasion coefficient, which is a weakness of
HMA made with RCA. However, combined with suitable natural [1] González-Fonteboa B, Martínez-Abella F. Recycled aggregates concrete:
aggregate and mix properties. Mater Constr 2005;55(279):53–66.
aggregate in suitable proportions, RCA can be used to produce [2] Kwan Wai Hoe, Ramli Mahyudin, Kam Kenn Jhun, Sulieman Mohd Zailan.
an HMA mix with acceptable abrasion resistance. Influence of the amount of recycled coarse aggregate in concrete design and
 Asphalt absorption is another property of RCA that must be durability properties. Constr Build Mater 2012;26:565–73.
[3] Poon Chi Sun, Chan Dixon. Feasible use of recycled concrete aggregates and
taken into account. In general, the optimum asphalt content
crushed clay brick as unbound road sub-base. Constr Build Mater
and bitumen absorption of HMA made with RCA increases with 2006;20:578–85.
increasing RCA content, while the effective binder content [4] NCHRP. Report 598. Performance-related test of recycle aggregates for use in
unbound pavement layers. Transportation Research Board; 2008.
decreases slightly. This is due to the absorptivity of the attached
[5] Vegas I, Ibáñez JA, Lisbona A, Sáez de Cortazar A, Frías M. Pre-normative
mortar on the surface of RCA. As a consequence, as the RCA con- research on the use of mixed recycled aggregates in unbound road sections.
tent increases, more binder is needed. In addition, as the effec- Constr Build Mater 2011;25:2674–82.
tive binder content decreases, so does the thickness of the [6] Cupo-Pagano M, D’Andrea A, Giavarini C, Marro C. Use of building demolition
waste for asphalt mixes: first results. Energy, environment and technological
binder film, so it is particularly important to limit the RCA con- innovation. In: Proceedings of III international congress; 1994. p. 203–8.
tent in HMA production. [7] Shen D, Du J. Evaluation of building materials recycling on HMA permanent
 Asphalt absorption is an important parameter influencing the deformation. Constr Build Mater 2004;18:391–7.
[8] Shen D, Du J. Application of gray rational analysis to evaluate HMA with
volumetric properties of HMA made with RCA. Not considering reclaimed building materials. J Mater Civil Eng 2005;17:400–6.
asphalt absorption leads to overestimating VFA and VMA. The [9] Aljassar AH, Al-Fadala KB. Recycling building demolition waste in hot-mix
differences between the volumetric properties determined con- asphalt concrete: a case study in Kuwait. J Mater Cycles Waste 2005;7:112–5.
[10] Paranavithana Sumeda, Mohajerani Abbas. Effects of recycled concrete
sidering asphalt absorption versus not considering it become aggregates on properties of asphalt concrete. Resour Conserv Recy
more pronounced as the RCA content increases, which is 2006;48:1–12.
another reason to limit the RCA content. [11] Pérez I, Toledano M, Galego J, Taibo J. Mechanical properties of hot mix asphalt
made with recycled aggregates from reclaimed construction and demolition
 HMA made with RCA exhibits high Marshall stability values, but debris. Mater Constr 2007;57(285):17–29.
the indirect tensile strength decreases with RCA content, so the [12] Tam Vivian WY, Tam CM, Le KN. Removal of cement mortar remains from
RCA content must be limited. recycled concrete aggregate using pre-soaking approaches. Resour Conserv
Recy 2007;50:82–101.
 Moisture damage is a key concern. The RCA content has been
[13] Wong YD, Sun DD, Lai D. Value-added utilisation of recycled concrete in hot-
demonstrated to be an important factor in the stripping perfor- mix asphalt. Waste Manage 2007;27:294–301.
mance of HMA made with RCA. The moisture susceptibility of [14] Gul Waqar Ahmed. Effect of recycled cement concrete content on rutting
HMA made with RCA may be better or worse than conventional behaviour of asphalt concrete. Doctoral thesis. Middle East Technical
University, Ankara; 2008.
HMA, depending on the natural aggregate with which the RCA is [15] Mills-Beale J, You Z. The mechanical properties of asphalt mixtures with
combined. recycled concrete aggregates. Constr Build Mater 2010;24:340–5.
 The tested mixtures exhibited very good stripping performance. [16] Mukhopadhyay Anol K, Geiger Brian J, Button Joe. Use of alkali–silica reaction-
affected recycled concrete aggregate in hot-mix asphalt. J Transp Res Board
HMA mixes made with RCA and oven-cured for 4 h before com- 2010;2179:1–9.
paction comply with Spanish specifications for use in pavement [17] Pérez I, Gallego J, Toledano M, Taibo J. Asphalt mixtures with construction and
base courses in road construction. demolition debris. Proc Inst Civil Eng Transp 2010;163:165–74.
[18] Pérez I, Pasandín AR, Gallego J. Impact of construction and demolition waste
 The tested HMA mixtures containing RCA exhibited very high aggregates on stripping in hot-mix asphalt. In: Presented at 89th annual
stiffness and a strong dependence on temperature. Such mix- meeting of the transportation research board, Washington (DC); 2010.
tures should not be used in cold regions. [19] Bhusal Sushanta, Li Xiaojun, Wen Haifang. Evaluation of effects of recycled
concrete aggregate on volumetrics of hot-mix asphalt. Transp Res Rec J Transp
 Although the RCA content seems to have no influence on the Res Board 2011;2205(3):36–9.
stiffness of the mixture, at 20 °C, the results indicate that an [20] Meizhu Chen, Juntao Lin, Shaopeng Wu. Potential of recycled fine aggregates
increase in the recycled aggregate content slightly stiffens the powder as filler in asphalt mixture. Constr Build Mater 2011;25:3909–14.
[21] Pérez I, Pasandín AR, Gallego J. Stripping in hot mix asphalt produced by
mixture, although these particular results were not statistically
aggregates from construction and demolition waste. Waste Manage Res
significant. 2012;30(1):3–11.
 Although the permanent deformation test results indicate that [22] Lee Cheng-Hsiao, Du Jia-Chong, Shen Der-Hsien. Evaluation of pre-coated
HMA with up to 30% RCA exhibit good rutting performance, per- recycled concrete aggregate for hot mix asphalt. Constr Build Mater
2012;28:66–71.
manent deformation increases with increasing RCA content, so [23] Pérez I, Pasandín AR, Medina L. Hot mix asphalt using C&D waste as coarse
mixtures with higher RCA percentages should be expected to aggregates. Mater Des 2012;36:840–6.
be more susceptible to deformation. [24] de Juan Marta Sánchez, Gutiérrez Pilar Alaejos. Study on the influence of
attached mortar content on the properties of recycled concrete aggregate.
 The fatigue life of HMA made with RCA is satisfactory. In partic- Constr Build Mater 2009;23:872–7.
ular, mixes with up to 20% RCA exhibit similar fatigue life to [25] Caro S, Masad E, Bhasin A, Little DN. Moisture susceptibility of asphalt
conventional mixture. For mixture made with 30% RCA fatigue mixtures, Part 1: mechanisms. Int J Pavement Eng 2008;9(2):81–8.
[26] Abo-Qudais Saad, Walid Mulqi Mohammad. New chemical antistripping
life is slightly slower than for conventional mixture. additives for bituminous mixtures. J ASTM Int 2005;2(8):1–11.
 Although the results of this investigation are encouraging, HMA [27] MINISTERIO DE FOMENTO. O.C. 24/08 sobre el Pliego de Prescripciones
made with RCA requires further investigation, particularly to Técnicas Generales para obras de carreteras y puentes (PG-3). Artículo: 542-
Mezclas bituminosas en caliente tipo hormigón bituminoso; 2008.
assess its performance in the field. It is also necessary to deter- [28] Ministerio de Obras Públicas y Transportes. Normas NLT. Ensayos de
mine the maximum percentage of RCA that can be used in HMA. carreteras. Dirección General de Carreteras, 2nd ed. Madrid, Spain; 2002.
[29] Asphalt Institute. Mix Design methods. for asphalt concrete and other hot-mix
types. MS-2. 6th ed.; 1997.
[30] AENOR. Asociación Española de Normalización y Certificación. UNE-EN 12697-
Acknowledgments 12. Métodos de ensayo para mezclas bituminosas en caliente. Parte 12:
Determinación de la sensibilidad al agua de las probetas de mezcla
The authors wish to acknowledge the Spanish Ministry of Edu- bituminosa. Madrid, Spain; 2006.
[31] AENOR. Asociación Española de Normalización y Certificación. UNE-EN 12697-
cation and Science for sponsoring this research through Project 26. Métodos de ensayo para mezclas bituminosas en caliente. Parte 26:
BIA2010-17751. Rigidez. Madrid, Spain; 2006.
 A.R. Pasandín, I. Pérez / Construction and Building Materials 43 (2013) 497–505 505

[32] BSi. British Standards Institution. Draft for Development DD 226:1996. [38] Ministerio de Fomento. CEDEX. IECA. Manual de firmes con capas tratadas con
Method for determining resistance to permanent deformation of bituminous cemento; 2003.
mixtures subject to unconfined dynamic loading; 1996. [39] Verhaeghe BMJA, Myburgh PA, Denneman E. Asphalt rutting and its
[33] AENOR. Asociación Española de Normalización y Certificación. UNE-EN 12697- prevention. In: Proceedings of 9th conference on asphalt pavements for
24: 2006+A1. Métodos de ensayo para mezclas bituminosas en caliente. Parte southern, Africa; 2007.
24: Resistencia a la fatiga. Madrid, Spain; 2006. [40] He Gui-ping, Wong Wing-gun. Laboratory study on permanent deformation of
[34] Khalid HA. A comparison between bending and diametral fatigue test for foamed asphalt mix incorporating reclaimed asphalt pavement materials.
bituminous materials. Mater Struct 2000;33:457–65. Constr Build Mater 2007;21:1809–19.
[35] Cocurullo A, Airey GD, Collop AC, Sangiorgi C. Indirect tensile versus two-point [41] Santagata FA, Canestrari F, Pasquini E. Mechanical characterization of asphalt
bending fatigue testing. Proc ICE – Transp 2008;161(4):207–20. rubber-wet process. In: Proceedings of 4th international SIV congress.
[36] Dirección General de Carreteras. Pliego de prescripciones técnicas generales Palermo. Italy; 2007.
para obras de carreteras y puentes PG-3. Ediciones Liteam, Madrid, Spain; [42] Aschuri Imam, Woodward David, Woodside Alan. Permanent deformation
2002. characteristics of asphalt concrete containing reclaimed materials. In: Sixth
[37] American Association of State Highway and Transportation Officials (AASHTO). international conference on maintenance and rehabilitation of pavements and
Guide for design of pavement structures, Washington (DC); 1993. technological control, vol. 1. Italy: Politechnico di Torino; 2009 p. 232–42.