Construction and Building Materials 151 (2017) 83–97

Contents lists available at ScienceDirect

Construction and Building Materials

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

Potential use of reclaimed asphalt pavement and recycled concrete

aggregate in base/subbase layers of flexible pavements
Muhammad Arshad ⇑, Muhammad Farooq Ahmed
Department of Geological Engineering, University of Engineering & Technology, G.T Road, Lahore, Pakistan

h i g h l i g h t s

 Addition of RAP contents can significantly increase the MR value of the blended sample.
 Addition of RAP contents can significantly increase the residual strain of the blended sample.
 Under controlled conditions RAP can be used in base/subbase layer of the pavement.

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

Article history: In current road and pavement engineering practices, the lack of fresh natural aggregate (granular mate-
Received 9 September 2016 rial) supplies with increasing processing costs have led to use various reclaimed/recycled materials from
Received in revised form 17 May 2017 old structures as a source of construction materials. Reclaimed Asphalt Pavement (RAP), and Recycled
Accepted 2 June 2017
Concrete Aggregate (RCA) have been used as aggregates for pavement construction for some time. This
Available online 21 June 2017
study is focused on the characterization of blended materials containing 50% and 75% of RAP with fresh
granular materials and RCA to evaluate whether they are suitable for granular base/subbase layers of flex-
Keywords:
ible pavements. A series of laboratory tests was performed to determine the resilient modulus (MR) and
Pavement
Recycled material
the constrained modulus (Mc) for both fresh granular materials and their blends. Statistically, the notable
Stresses increase was found in the MR values of the blended samples containing 75% RAP material and 25% fresh
Resilient modulus granular, particularly at higher levels of bulk stresses. It was also found that the accumulative strains dur-
Constrained modulus ing cyclic loading generally increase with an increase in the percentage of RAP contents in the blended
samples. Mc test results show an increasing trend with the increasing level of axial stress, however, Mc
value decreases with increasing percentage of the RAP content. Never-the-less, the t-test showed that
accumulative strains during Mc tests were found to increase significantly with an increase in the percent-
age of RAP contents.
Ó 2017 Elsevier Ltd. All rights reserved.

1. Introduction and documented by Hansen and Copeland, [17]; Zaumanis et al.,

Use of recycled materials for the construction of roads, pave-
ments, footpaths is increasing all over the world due to their cost
effectiveness and environmentally sustainable aspects. The most
common recycled materials used in different layers of flexible
pavements are Reclaimed Asphalt Pavement (RAP) [1–3], Recycled
Concrete Aggregate (RCA) [4–6], recycled bricks [7,8], recycled
glass [9–14] and fly-ash [15].
Data provided by European Asphalt Pavement Association
(EAPA) show that in Europe, 47% of the available RAP was used
in hot mix asphalt applications during 2011 [16]. In the US, statis-
tics published by National Asphalt Pavement Association (NAPA)
⇑ Corresponding author.
E-mail address: engineerarshad@hotmail.com (M. Arshad).
[18] estimates a total of 71.8 million tonnes of RAP accepted in
2011, 84% of which were used in asphalt applications. Although
nationally this is a high re-use rate, in urbanized areas the restric-
tions on the maximum allowed RAP content in mix design and
technical capabilities of asphalt plants have created a high surplus
of RAP. Several studies have shown that the RAP mixtures in hot
mix asphalt have performed equal or better than the fresh granular
material [18–23]. RAP has been used in hot mix asphalt pavements
in various percentages that reached in some cases up to 80% [24].
Most studies have recommended the use in a range from 20–50%
[15,25,26]. Hence in addition to using recycled materials as a
replacement for fresh granular material in hot mix asphalt, there
is a demand for them to be used as granular base/subbase
materials.

http://dx.doi.org/10.1016/j.conbuildmat.2017.06.028
0950-0618/Ó 2017 Elsevier Ltd. All rights reserved.
84 M. Arshad, M.F. Ahmed / Construction and Building Materials 151 (2017) 83–97
In recent years, efforts have been made to incorporate RAP into
pavement base or subbase applications [1–4,27–32]. RAP stabilised
with cement binders has been reported toperform satisfactorily in
pavement base and subbase layers [2,3].The application of RAP and
RCA in pavement base/subbase as an aggregate, however, has been
limited due to the lack of reported laboratory testing and field-
testing results [33].Most of the existing literature concludes that
MR value increases with RAP content; however, the majority of
the studies have only considered low RAP contents in the blends.
Moreover, the effects of other fundamental properties including
accumulative deformation on the structural behaviour of the RAP
are lacking. An extensive analysis of the characteristics of RAP to
be incorporated into base/subbase layer is not available in the lit-
erature [33]. This situation demands the developing procedures/
guidelines for the successful use of RAP and RCA as unbound mate-
rials within the framework of different design methods being
applied for highway pavements.
The main objective of this research is to analyse the results of
performance-based tests on selected fresh granular materials,
RCA and RAP for the use in base/subbase layer of flexible pave-
ment. In particular, the focus is placed on studying the impact of
the variation of the resilient modulus and constrained modulus
with RAP content when different types of RAP are mixed with a
fresh granular material. In order to achieve this goal, the research
emphases on the following tasks:
1. Carry out a literature review focusing on the use, testing, and
evaluation of RAP and RCA materials in unbound pavement lay-
ers, taking into account various research findings, performance
data, current practices and related specifications from other
studies.
2. Conduct laboratory tests to determine the stiffness and com-
pressibility of blended materials prepared by mixing fresh gran-
ular materials/RCA with RAP.
3. Characterise the dependency of resilient and constrained mod-
uli and accumulative strain of blends containing RAP as a major
constituent on the applied stress states under repeated loading.
2. Reclamation of RAP material
RAP is reclaimed when an existing hot mixed asphalt (HMA)
layer is removed for reconstruction or resurfacing. The two meth-
ods generally used for RAP reclamation are (i) cold milling and (ii)
ripping & breaking [34–38]. RAP properties are governed by milling
and ripping/crushing operations, as well as the characteristics of
the asphalt binder, age of the asphalt pavement and aggregates
present from which the RAP is reclaimed [38]. RAP reclaimed from
surface courses (compared to binder courses) is usually of higher
quality because of the higher quality aggregates that were used
in the original construction.
The cold milling method which is most commonly used in engi-
neering practices generally removes up to 50 mm thick HMA in a
single pass while restoring the surface to a specified grade and
slope. A number of passes may be required, depending on the dis-
tress severity, to free the surface of any rutting, bumps or other
defectiveness. The Asphalt Recycling and Reclaiming Association
(ARRA) categorise cold milling into five classes [39]:
 Class I – Milling the existing HMA to the necessary depth to
remove surface irregularities.
 Class II – Milling the existing surface to a uniform depth.
 Class III – Milling the existing surface to a uniform depth and
cross slope.
 Class IV – Milling the entire thickness of the HMA layer.
 Class V – Milling the existing surface to variable depths.
Full-depth removal involves ripping and breaking the pavement
using a rhino horn on a bulldozer and/or pneumatic pavement
breakers. In most instances, the broken material (removed HMA
blocks) is picked up by front-end loaders and loaded into haul
trucks. The material is then hauled to a central facility for process-
ing. At this facility, the RAP is processed using a series of opera-
tions, including crushing, screening, conveying, and stacking for
later use [34,35,38].
Proper crushing and screening of RAP can yield well-graded
aggregate particles that are partially or wholly coated with asphalt
binder [35,39]. Although the majority of old asphalt pavements are
recycled at central processing plants, asphalt pavements may also
be pulverised in place and incorporated into granular or stabilised
base courses using a self-propelled pulverising machine. Cold
milling is generally more efficient than ripping and crushing
because it is usually completed in-situ without hauling RAP to a
crusher or at the central processing plant for processing. However,
it is commonly accepted that cold milling produces more fines than
ripping and crushing [35].
3. Aggregate degradation during RAP reclamation
RAP particles often consist of conglomerates of smaller aggre-
gates glued together by mastic. The maximum size of RAP gener-
ally varies from 30 mm to 50 mm but has not been categorically
defined in the literature [38]. Aggregate degradation occurs during
milling and crushing operations, which causes the RAP gradation to
be finer than the gradation of the original virgin aggregates. Aggre-
gate degradation during milling is a function of the top size and the
gradation of the aggregates used in asphalt pavements [39]. How-
ever, in most cases, RAP is well graded and slightly finer than
crushed natural aggregates [37].
The crushing of HMA pavements at a central plant is carried out
using compression and impact crushers. RAP breakers are used if
crushed RAP starts to form a flat, dense mass, particularly on warm
and humid days. Impact crushers may be one viable equipment for
RAP reclamation, as there is less chance of the equipment getting
plugged with RAP material, which sometimes happens when using
jaw crushers. An impact crusher can also be treated as a secondary
crusher when a jaw crusher is used as a primary crusher. In a com-
bination crusher, the jaw crusher reduces the HMA slabs to more
convenient size, which are then further reduced to useable sizes
using the secondary roll crusher. RAP reclaimed using ripping is
generally less fine than the RAP reclaimed using milling but the
aggregate fraction is still finer than the original granular material.
The equipment used during milling and ripping operations also
affects the gradation of RAP. The RAP gradation is further influ-
enced by the gradation of the underlying base/subbase layer if
the depth of reclamation includes parts of these layers [40].
Table 1 Summaries the range of gradation of RAP materials
obtained after milling or ripping as reported by different agencies
[41].
4. Allowable RAP content in base/subbase layer
In different studies, it has been found that addition of the RAP
contents in the fresh granular materials can change the mechanical
properties of the granular material implicitly and hence the perfor-
mance of the pavement layer (base and/or subbase layer) is also
affected considerably. Table 2 lists the general trends for five engi-
neering properties including dry density, optimum moisture con-
tent, permeability, CBR and resilient modulus as the percentage
of RAP content was increased in the prepared blend. In this table,
where the blending of RAP with the fresh granular material is
 M. Arshad, M.F. Ahmed / Construction and Building Materials 151 (2017) 83–97 85

Table 1
Range of RAP particle size distribution after milling or ripping (after Saeed [41]).

Sieve size Range of particle size distribution (percent finer)

(mm)
FHWA** New Jersey DOT* Florida DOT* Texas DOT*
37.5 100 100 100 100
25 95–100 98 97
19 84–100 90 88 86
12.5 70–100 76 – 72
9.5 58–95 68 76 58
7.5 38–75 52 52 35
2.36 25–60 20 34 27
2 – 18 – –
1.18 17–40 11 25 10
0.6 10–35 8 17 5
0.3 5–25 4 – 2
0.25 – – 8 –
0.15 3–20 2 4 1
0.075 2–15 0.5 0.5 0
**
Federal Highway Administration.
*
Department of Transportation.

Table 2
Effect of RAP content on mechanical properties of the blended materials.

Reported Blended Dry density Moisture content Permeability CBR Resilient modulus
Bennert and Maher [42] Yes Decreased Decreased Decreased – Increased
Taha et al., [43] Yes Decreased No Change Increased Decreased –
Blankenagel and Guthrie [30] Yes Decreased Decreased – Decreased –
Cosentino et al. [44] No Decreased Increased – Decreased –
MacGregor et al., [45] Yes – – No Change – Increased
Papp et al.,[46] Yes Decreased Decreased – – Increased
Sayed et al., [47] No – Decreased – Decreased –
Trzebiatowski and Benson [48] No Decreased – Increased – –

absent, comparison of the engineering properties on the basis of
100% RAP to 100% granular material is reported.
Different agencies recommend different percentages of RAP
content with respect to the fresh granular material to achieve opti-
mum results for pavement performances. Saeed [41] indicated that
in the USA 16 state DOTs (department of transportation) allowed
the use of 100% RAP as aggregate in unbound pavement layers
and 5 DOTs restricted the use of RAP to 50% or less by weight.
Taha et al., [43] performed an experimental investigation on
well-graded RAP contents having uniformity coefficient (Cu) and
curvature coefficient (Cc) equal to 6 and 1.5, respectively, while
the fresh granular material was a mixture of well-graded sand
and gravelly sand with little or no fines. The blends were obtained
by adding 20, 40, 60, 80 and 100% of the granular material with
RAP. On the basis of California Bearing Ratio (CBR) test results, they
have suggested that up to 100% RAP in subbase courses could be
allowed but the amount of RAP in unbound granular base courses
would have to be limited to 10%.
Taha et al., [1] found that the ability of RAP contents to function
as a structural component of the pavement is more pronounced
when RAP contents are stabilised with cement rather than blend-
ing it with the only granular material. They suggested that a
100% RAP aggregate should not be recommended for use as a base
material unless stabilised with cement. Based on the results of this
study, cement stabilised RAP-granular material mixtures seem to
be a viable alternative to dense-graded aggregate used in road base
construction.
Bennert and Maher [42], have conducted testing on soils of New
Jersy, quantity of RAP contents with the fresh granular material can
be limited to 50% by weight. Results from the CBR tests revealed
that there was about a 50 and 55% decrease in CBR values when
the RAP blend was increased from 0 to 25% RAP and 50 to 75%
RAP, respectively. Little change occurred in the CBR value when
the RAP percentage was increased from 25 to 50%. In general, as
the percentage of RAP increased, the CBR values decrease. For the
resilient modulus test, as the percentage of RAP increased, the resi-
lient modulus was found to increase. Permanent deformation,
however, increased as the percentage of RAP increased. The 100%
RAP mixture experienced the most deformation. They suggested
that blends with 50% RAP contents could meet the requirements
for base course material.
The study conducted by Aljassar et al., [49] showed that addi-
tion of RAP to the fresh granular material could improve signifi-
cantly the properties of subgrade soil in terms of the maximum
unit weight and CBR. A 22% addition of milled RAP produced the
best results in terms of density and CBR value.
Mokwa and Peebles [50] conducted laboratory tests on four dif-
ferent types of granular material blended with varying percentages
of RAP (20, 50 and 75%). They found that blending of RAP with
granular material resulted in only minor changes to the engineer-
ing properties of the fresh granular material. However, they sug-
gested a limiting value of 50% RAP when used for the base course.
The departments of transportation for different states in the
USA allows blends of RAP material with the fresh granular material
in the range of 20% to 60% (by weight) in the base layer of a pave-
ment section [51].
The Ministry of Transportation Ontario (MTO) has successfully
used 100% RAP as a granular base and shoulder material on several
construction projects since the early 1970’s. It has been found that
the bearing capacity or CBR value of the granular base/subbase
materials tend to decrease when RAP content is increased. Current
MTO guidelines allow for as much as 50% RAP in base and subbase
layers when used in conjunction with full depth reclamation [52].
It is also worth to consider that application of RAP in pavements
base and subbase has limitations as it shows high water absorption
and Los Angeles abrasion and low California Bearing Ratio (CBR)
values as reported by Taha et al. [43], which do not satisfy many
local-road authority specifications. Similarly Arulrajah et al. [14]
86 M. Arshad, M.F. Ahmed / Construction and Building Materials 151 (2017) 83–97
indicated that pure RAP does not meet the CBR and repeated load
triaxial test requirements to qualify as an unbound subbase mate-
rial in Australia. Cosentino et al., [44] reported that all granular
blends containing RAP exhibited some amount of creep. The study
recommended that unstabilized RAP material be blended with a
minimum of 75% approved aggregate for use in base layer where
heavy traffic is common. Otherwise, blends should be proportioned
so that the asphalt binder content does not exceed 1.5% by weight.
Some researchers are also of the view that RAP with relatively high
asphalt content in granular courses is a wastage of materials, since
the presence of asphalt binder can actually inhibit drainage in
base/subbase layers and be detrimental to the function of the pave-
ment structure.
5. Material characterization
For this research work index properties tests, resilient modulus
tests, one-dimensional constrained modulus tests (1D) were con-
ducted to evaluate the influence of RAP content on the resilient
and deformation properties of the various blends prepared by mix-
ing RAP contents (designated as ‘RAP(1)’, ’RAP(2)’, ‘RAP(3)’ & ‘RAP
(4)’) with fresh granular samples (designated as ‘A’, ‘F’ & ‘W’) and
RCA material. Each of the fresh granular samples, as well as RAP
materials, were containing crushed limestone of angular to sub-
angular shape while flat and elongated particles in the samples/ma-
terials were limited to 6% as per ASTM D 4791. The specific gravity
of the fresh granular samples varied from 2.60 to 2.65 as per ASTM C
127. RCA sample was obtained from plain concrete aggregate
designed for 25 MPa strength after 28 days having limestone as
crushed aggregate, ordinary Portland cement and coarse sand.
Fresh granular samples (A, F and W), RCA and RAP samples were
brought into the laboratory, air-dried at laboratory ambient tem-
perature (typically 23 ± 2 °C), then thoroughly mixed before being
split into equal parts by using a standard riffle box. After propor-
tioning and blending the materials, appropriate quantities of each
blend (0, 50, and 75% RAP) were stored in sealed plastic bags until
they were required for different types of testing. Table 3 shows the
matrix of the testing program designed for resilient modulus test
and constrained modulus test.
It should be noted that procedures for characterization of RAP
materials are quite varied across the world. The only characteriza-
tion that is common across the various standards and procedures is
the particle size distribution curves (gradation analysis). However
Tebaldi et al., [37,38] have suggested that a completed characteri-
zation would require a series of testing including but not limited
to: determination of particle sizes (as RAP and of recovered aggre-
gates), determination of active binder amount in RAP, shape char-
acterization of recovered aggregates, cleanliness of RAP and
homogeneity
5.1. Particle size distribution
The particle size distribution characteristics were obtained
according to AASHTO Designation T27-99 [53] for all materials/
blends. Figs. 1 and 2 shows the gradation curves for granular mate-
rials & RCA and RAP materials respectively. Each curve corresponds
Table 3
Matrix of testing program.
Fresh granular sample/RCA RAP Minimum Number of Resilient Modulus Tests
Fresh granular sample A, F, W – 3
RCA – 1
Fresh granular sample A, F, W 4  2* 3  2  4 = 24
RCA 4  2* 24=8
Total 36
*
Blended samples were prepared by mixing 50% and 75% (by weight) of each RAP type with the virgin granular aggregates and RCA.
to the average of two independent tests. Based on the gradation
curves, granular material (A) is finer than other granular materials
(F &W) while granular (w) is the most coarser among them. Simi-
larly, RAP(3) is coarser when compared with other RAP materials
and RAP(4) being the most finer among them.
Table 4 summarises the particle size distribution characteristics
of the materials including coefficient of uniformity (Cu) and coeffi-
cient of curvature (Cc). Gradation characteristics for a specific blend
were obtained according to the proportion of different granular
materials, RCA and RAP making up that specific blend.
5.2. Physical properties of recovered bitumen binder
The RAP samples were washed with tab water to eradicate for-
eign materials and the samples were put in the oven at a temper-
ature of 140 °C to lose the aggregate from the binder agent.
Recycled binder contents for different RAP types were extracted
from the RAP samples by solvent extraction method using a rota-
tional viscometer apparatus. The recovered binder contents rang-
ing from 3.3 to 4.7% were characterised in terms of penetration
(ASTM D5-06), viscosity (ASTM D4402-06) and softening points
(ASTM D36-06), results are presented in Table 5.
5.3. Modified proctor compaction test
Results of the Modified Proctor Compaction tests (AASHTO
T180) [54] were used to get the compaction characteristics of all
tested materials. Particles larger than 19 mm in the original mate-
rials were removed prior to the compaction test. Figs. 3 and 4 pre-
sent the moisture-density curves of fresh granular materials and
RAP, respectively. The optimum moisture contents (wopt) of granu-
lar and RCA are in the range of 5.5% and 7.5%. The maximum dry
unit weight varied from 20.7 to 23.3 kN/m3, with granular W hav-
ing the highest maximum dry unit weight and RCA the lowest.
The values of wopt for the RAP are in the range of 6–9% while the
maximum dry unit weight lies between 19.5 to 21.5 kN/m3. Even
though the proctor compaction tests for blends of virgin aggregates
with RAP were not performed, it assumed that the wopt values of
the blends with different RAP contents would be in the range of
5.5–7.5%.
5.4. Resilient modulus tests
The resilient modulus, MR for pavement construction materials
which is a surrogate to the elastic modulus is defined as a ratio of
deviator stress to the resilient strain experienced by the material
under repeated loading simulating traffic load. Fig. 5 illustrates
the stress conditions for resilient modulus tests as well as a sche-
matic of the resilient modulus. It has been known that the resilient
modulus of a material is affected by details of the testing method,
including sample preparation and the nature of applied cyclic
stress among others [55]. For this study, the resilient modulus tests
were conducted following AASHTO T307-99 [56].
Minimum Number of Constrained Modulus Tests
3
1
3  2  4 = 24
24=8
36
 M. Arshad, M.F. Ahmed / Construction and Building Materials 151 (2017) 83–97 87

100

90

80

70

Material passing by mass (%) 60

50

40

30
Granualr (A)
20
Granualr (F)

10 Granular (W)
RCA
0
0.01 0.1 1 10 100
Diameter (mm)

Fig. 1. Particle size distribution curves for granular sample A, F, W and RCA.

In order to determine the resilient modulus of unbound materi-

als, cyclic stress of fixed magnitude for 0.1 s is applied to the spec-
imen followed by a 0.9 s rest period. The specimen is subjected to a
confining stress provided by means of a triaxial pressure chamber.
Fig. 6 shows typical repetitive load pulses that were generated by
the testing system.

5.4.2. Sample preparation method and testing procedure

All samples were compacted using the vibration device which
meets the requirement of AASHTO T307-99 [56]. The specimens
used for resilient modulus tests were fabricated using a split mould
with an inside diameter of 152 mm (6 in.) and the vibratory impact
hammer without kneading action. The height of test specimens
was 305 mm (12 in.). According to AASHTO T307-99 [56], particles
larger than 37.5 mm (i.e., 1/4 of the specimen diameter) should be
Fig. 2. Particle size distribution curves for RAP materials.
removed from the test specimens prior to compaction. A split
mould lined with a rubber membrane was mounted on the base
of the triaxial cell and a vacuum was applied to make an intimate
5.4.1. Loading system and data acquisition for resilient modulus test
contact of the membrane with the inner wall of the mould. The
The loading device used in the study is a closed-loop, servo-
moist test specimen was compacted in 5 layers in the mould. At
controlled electro-hydraulic material testing system (MTS) having
the end of the compaction process, a second membrane was
a function generator that is capable of applying repeated cycles of
applied to the sample to ensure that the correct confining pressure
haver-sine shapes load pulse and following a particular loading
could be maintained on the test specimen during the test.
history supplied by the microcomputer software.
Optimum moisture contents (OMC or wopt) obtained from mod-
For an accurate measurement of the applied load and the result-
ified proctor test results were used as a guide to estimate appropri-
ing axial deformation of the test specimen, the load cell of 11.2 kN
ate moisture for compacting the MR test specimen. It is fact that
and linear variable displacement transducers (LVDTs) of 10 mm
both the compaction technique and the compaction energy for
capacity was used.
fabricating the MR test specimens were changed from that

Table 4
Aggregate type and particle size distribution characteristics of the given materials.

Material D10 (mm) D30 (mm) D50 (mm) D60 (mm) Cu Cc % Fines % Sand Size % (4.75–9.5) mm
Granular sample A 0.2 0.45 0.9 1.75 8.5 0.6 5 70 10
Granular sample F 0.15 1.7 9 15 100 1.3 5 35 10
Granular sample W 0.6 5 15 18 33.3 2.3 5 25 12
RCA 0.25 1.5 6.5 10 36.0 0.9 5 40 18
RAP(1) 0.3 1.2 2.75 4 13.3 1.20 3 64 22
RAP(2) 0.3 1.2 2.75 4 13.3 1.20 3 60 30
RAP(3) 1.5 5 6.5 7.5 5.0 2.22 3 27 50
RAP(4) 0.3 1.2 2 2.75 9.2 1.75 3 82 10
88 M. Arshad, M.F. Ahmed / Construction and Building Materials 151 (2017) 83–97

Table 5
Physical properties of recovered bitumen binder.

RAP 60 °C viscosity 25 °C penetration Softening point

type (poise) (dmm) (°C)
(ASTM D4402) (ASTM D5-06) (ASTM D36-76)
RAP4 23,500 52 62
RAP1 32,300 34 64
RAP2 38,100 28 65
RAP3 46, 700 20 67

Fig. 3. Moisture-density relationship for granular A, F, W and RCA.

Fig. 6. Typical load pulses generated during the resilient modulus test.

Fig. 4. Moisture-density relationships for RAP materials.

Fig. 7. Apparatus for the constrained modulus test. Key: 1, loading frame; 2, load
measuring gauge; 3, deflection gauge; 4, mould containing test specimen; 5,
hydraulic pump.

corresponding to the proctor tests and some aggregates had diffi-

Fig. 5. Definition of resilient modulus [1].
culty to hold the water contents when subjected to vibrational
effect for the preparation of samples. This fact may lead to a slight
reduction in the actual target moisture content of specimen pre-
pared for MR test. However, it was assumed that consequences of
this situation will be negligible on the test results.
 M. Arshad, M.F. Ahmed / Construction and Building Materials 151 (2017) 83–97 89

Table 6
Axial stress for different stages of unloading during
(a)400
constrained modulus test. y = 9.7688x0.5711
R² = 0.9631
Stages Axial stress (kPa)
300
1st 205
2nd 275 y = 5.1885x0.6365
y = 10.23x0.5478
R² = 0.9926
3rd 310 R² = 0.9686

MR (MPa)
4th 410 200

450 100 Granular(W)

400 50%Granular(W)+50%RAP(1)
25%Granular(W)+75%RAP(1)
350
0
Axial stress (kPa)

300 0 100 200 300 400 500 600 700 800

Bulk stress (kPa)
250

200 (b) 450

150 y = 10.521x0.5538
400
Cycle 1 R² = 0.9681
100 Cycle 2 350
y = 13.614x0.4978
50 Cycle 3 300 R² = 0.9412
Cycle 4

MR (Mpa)
0 250 y = 5.1683x0.6369
R² = 0.9923
0 0.002 0.004 0.006 0.008 0.01
200
Strain (%)
150
Fig. 8. Typical stress-strain loop for the estimation of constrained modulus. Granular (W)
100
50%Granualr(W)+50%RAP(2)
After a test specimen had been fabricated and placed in the tri- 50
25%Granualr(W)+75%RAP(2)
axial chamber, the recording devices for the LVDT’s and load cell 0
were adjusted as needed before conditioning the specimen. There- 0 100 200 300 400 500 600 700 800
after, the confining pressure was set to 103.4 kPa and 750 repeti- Bulk stress (Kpa)

tions of a load corresponding to a maximum axial stress of (c) 500

103.4 kPa and corresponding cyclic axial stress of 93.1 kPa was
applied for sample conditioning. The purpose of sample condition- 400
ing is to remove the effect of the interval between compaction.
Conditioning also aids in minimising the effects of originally
MR (MPa)

300
imperfect contact of the test specimen between the sample cap
and base plate. During the cyclical loading stage, the drainage valve
was kept open to minimise the buildup of excess pore pressure. 200
Following the conditioning stage, resilient modulus tests were car-
ried out according to the loading sequences described in AASHTO 100 Granular(W)
T307-99 [56] for base/subbase materials. 50%Granular(W)+50%RAP(3)
This protocol (AASHTO T307-99) completes the MR testing with 25%Granular(W)+75%RAP(3)
0
15 loading series consisting of different combinations of confining 0 100 200 300 400 500 600 700 800
stress, cyclic axial stress and contact stress. For the first series, the Bulk stress (kPa)
maximum axial stress was set at 3 psi (21.0 kPa) and the confining (d) 400
pressure was 3 psi (21.0 kPa). One hundred repetitions of load
cycles were applied during each series of the loading. The readings
of each LVDT and the load cell for the last five cycles of each load- 300
ing sequence were considered to calculate the corresponding resi-
lient moduli.
MR (MPa)

This is worth to mention that similar testing procedure for the 200
resilient modulus test including apparatus, sample preparation
method and data acquisition system has been documented by
other researchers as well [57,58]. 100 Granular(W)
50%Granular(W)+50%RAP(4)
25%Granular(W)+75%RAP(4)
5.5. Constrained modulus tests
0
0 100 200 300 400 500 600 700 800
Owing to the limit of specimen dimensions in the resilient mod- Bulk stress (kPa)
ulus testing, particles larger than 37.5 mm have to be removed and
replaced. As such, the tested specimen tends to have a different Fig. 9. Effect on resilient modulus of blends containing granular W and: (a) RAP(1);
(b) RAP (2); (c) RAP (3); (d) RAP(4).
structure when compared with that of the original material. Such
practice could lead to a certain effect on the accumulative (irrecov-
erable) strain under repeated loading. This is because large size induced by removing large size particles, a series of 1D compres-
particles may form locally stable structure and have different sion tests were carried out using a 300 mm diameter and
residual deformation. In order to minimise the potential effect 300 mm high mould (net dimensions to accommodate a particular
90 M. Arshad, M.F. Ahmed / Construction and Building Materials 151 (2017) 83–97

(a) 500 strained modulus, which reflects the compressibility and hence
the stiffness of a material, can be obtained from these tests.

400
5.5.1. Loading system and data acquisition
Fig. 7 shows the test setup of 1D compression for the con-
MR (MPa)

300 strained modulus. The deformation of the sample is measured

using three dial gauges fixed at 120° relative to each other. Obvi-
200 ously, such an arrangement of the dial gauges provides ‘‘best” esti-
mate for the average deformation even when the loading plate has
RCA
a slight tilt. A load cell is used to measure the load that is controlled
100
50%RCA+50%RAP(1) manually using a hydraulic pump. During the 1D compression test,
25%RCA+75%RAP(1) the sample is unloaded at different stress levels to obtain the
0 unloading characteristics and the constrained modulus of the
0 100 200 300 400 500 600 700 800
Bulk stress (kPa) material. Table 6 shows the four distinct stress levels at which
unloading takes.
(b) 500
5.5.2. Sample preparation method, testing procedure and
400 determination of constrained modulus
Samples for constrained modulus tests were prepared in a man-
ner similar to that used for resilient modulus test specimens. More
MR (MPa)

300
specifically, specimens were compacted using the same vibratory
compaction device at the target moisture contents determined by
200
modified compaction test. Similar to the resilient modulus testing,
the specimens were conditioned by applying a number of loading-
100 RCA unloading cycles. The maximum vertical pressure in the condition-
50%RCA+50%RAP(2)
ing stage is selected as 205 kPa, which is, in fact, equal to the axial
25%RCA+75%RAP(2)
0 stress in the conditioning stage of the resilient modulus testing. 10
0 100 200 300 400 500 600 700 800 loading-unloading cycles were applied to the specimen by the
Bulk stress (kPa)
manual control at the average rate in the range of 100–150 kPa/
(c) 500 min before the data of load and settlement were recorded. Fig. 8
shows typical loading-unloading cycles conducted at four different
levels of axial stress. The slope of the average trend line drawn for
400
each loop gives an estimated value of constrained modulus at
respective stress level.
300
MR (MPa)

6. Effect of RAP contents on resilient modulus (MR)

200

Resilient modulus tests for different blends of granular materi-

100 RCA als and RCA with different types of RAP materials were conducted
50%RCA+50%RAP(3) in accordance with the procedure described in AASHTO T307-99
25%RCA+75%RAP(3)
0 [56] for base/subbase materials as discussed earlier.
0 100 200 300 400 500 600 700 800
Bulk stress (kPa)
6.1. Effect of RAP content on measured MR of granular sample W
(d)
500 The four types of RAP materials were mixed with granular sam-
ple W in proportion to prepare eight blends with RAP content of
400 50% and 75% respectively. Figs. 9a–d show the effect of the addition
of these RAP on the measured values of MR at different bulk stres-
ses. The resilient moduli of blends containing granular sample W
300
MR (MPa)

and RAP (1) generally increase with RAP content. On comparing

the data from the respective trend lines, it reveals that the resilient
200 modulus of the blend with 75% RAP is approximately 40 Mpa to
80 Mpa higher than that of granular W, while the increase in resi-
100 RCA lient modulus is about 35 Mpa on the average at 50% RAP content.
50%RCA+50%RAP(4) It is also observed that at a low level of bulk stress (h) (say up to
25%RCA+75%RAP(4) 300 kPa) the increase in MR value due to 50% increase in RAP(1)
0
0 100 200 300 400 500 600 700 800 that is from 50% to 75% is less than 20 Mpa. When the bulk stress
Bulk stress (kPa) is beyond 300 kPa, the resilient modulus of blends having 75% RAP
content is noticeably higher than that of blend with 50% RAP.
Fig. 10. Effect on resilient properties of blends containing granular RCA and: (a) In case of blends with RAP(2), there is no clear separation
RAP(1); (b) RAP(2);(c) RAP(3);(d) RAP(4), between the trend lines for blends with 50% and 75% RAP content
up to bulk stress (h) of 300 kPa as shown in Fig. 9b, although they
test specimen). This mould can be used for the test materials with a both have higher values of MR (by approximately 40 Mpa) than the
maximum particle size of approximately 40–50 mm. The con- granular sample W.
 M. Arshad, M.F. Ahmed / Construction and Building Materials 151 (2017) 83–97 91

Table 7
t-Test analyses for the effect of RAP contents on resilient modulus of the materials.

Comparison of material type/blend T-statistic T-critical (one tail) Comments

Granular (W) versus 50% Granular (W) + 50% RAP(1) 1.25 1.70 Insignificant
Granular (W) versus 25% Granular (W) + 75% RAP(1) 1.88 1.70 Significant
50% Granular (W) + 50% RAP(1) versus 25% Granular (W) + 75% RAP(1) 0.69 1.70 Insignificant
Granular (W) versus 50% Granular (W) + 50% RAP(2) 1.71 1.70 Significant
Granular (W) versus 25% Granular (W) + 75% RAP(2) 1.26 1.70 Insignificant
50% Granular (W) + 50% RAP(2) versus 25% Granular (W) + 75% RAP(2) 0.51 1.70 Insignificant
Granular (W) versus 50% Granular (W) + 50% RAP(3) 0.75 1.70 Insignificant
Granular (W) versus 25% Granular (W) + 75% RAP(3) 2.43 1.70 Significant
50% Granular (W) + 50% RAP(3) versus 25% Granular (W) + 75% RAP(3) 1.71 1.70 Significant
Granular (W) versus 50% Granular (W) + 50% RAP(4) 0.79 1.70 Insignificant
Granular (W) versus 25% Granular (W) + 75% RAP(4) 2.0 1.70 Significant
50% Granular (W) + 50% RAP(4) versus 25% Granular (W) + 75% RAP(4) 1.05 1.70 Insignificant
RCA versus 50% RCA + 50% RAP(1) 1.71 1.70 Significant
RCA versus 25% RCA + 75% RAP(1) 5.5 1.70 Significant
50% RCA + 50% RAP(1) versus 25% RCA + 75% RAP(1) 3.16 1.70 Significant
RCA versus 50% RCA + 50% RAP(2) 1.03 1.70 Insignificant
RCA versus 25% RCA + 75% RAP(2) 2.87 1.70 Significant
50% RCA + 50% RAP(2) versus 25% RCA + 75% RAP(2) 1.90 1.70 Significant
RCA versus 50% RCA + 50% RAP(3) 0.93 1.70 Insignificant
RCA versus 25% RCA + 75% RAP(3) 3.07 1.70 Significant
50% RCA + 50% RAP(3) versus 25% RCA + 75% RAP(3) 2.22 1.70 Significant
RCA versus 50% RCA + 50% RAP(4) 2.71 1.70 Significant
RCA versus 25% RCA + 75% RAP(4) 3.28 1.70 Significant
50% RCA + 50% RAP(4) versus 25% RCA + 75% RAP(4) 0.49 1.70 Insignificant

At bulk stress of 670 kPa, increase in the MR value is approxi-
mately 6% and 18% for the blends containing 50% and 75% RAP
(2), respectively, when compared with the corresponding value
for the granular W.
The blend with 75% RAP(3) has a significant increase in MR
value, in the range of 60 Mpa to 80 Mpa, when compared with
the corresponding value for the granular W; as shown in Fig. 9c.
However, the blend having 50% of RAP(3) only have marginally
higher MR values than the corresponding value for the granular W.
Mixing RAP(4) with the granular W also resulted in higher MR
values as shown in Fig. 9d. The MR values of the blend with 75%
RAP are approximately 55 Mpa higher than those of the granular
W on the average. However, no significant difference in the MR val-
ues was observed for blends with 50% and 75% RAP(4). Almost sim-
ilar trends were observed in the variation of MR values for the other
blended samples prepared by mixing the granular samples A, F and
RAP. For all tested materials, the resilient modulus increases with
the applied bulk stresses. The relation between the resilient mod-
ulus and the bulk stress of a specific material can be approximately
expressed via a power law.
6.2. Effect of RAP content on measured MR of RCA
Similar to the cases for granular sample W, eight blends were
prepared by mixing the RCA, with the RAP content being 50%
and 75% in the blends. In general, all RCA-RAP blends have higher
resilient moduli than clean RCA for given range of bulk stress; as
shown in Fig. 10a–d.
It can be inferred from Fig. 10a that the addition of RAP(1) into
clean RCA induces remarkable increase in the resilient modulus,
even more, pronounced than that for granular W. More specifically,
the blend with 75% RAP content results in an average increase in
MR by 120 Mpa at h (bulk stress) = 85.5 kPa to 138 Mpa at
h = 671.5 kPa. Compared with MR = 97.8 Mpa and 281.3 Mpa for
the clean RCA at the corresponding bulk stress levels, the relative
increase of MR is as high as 127% and 49%. It is also noted from
Fig. 10a that the trend line for the blend with 75% RAP(1) is almost
parallel to that of clean RCA material when the bulk stress
increases from 85.5 kPa to 671.5 kPa. The trend line for the blend
with 50% RAP(1) has slightly higher slope than that of the virgin
RCA material and the blend with 75% RAP, which implies a slightly
higher pressure sensitivity of this blend. For example, comparing
with the clean RCA, the increase of MR of the blend with 50% RAP
content is DMR = 34 Mpa at h = 85.5 kPa and DMR = 79 Mpa at
h = 510 kPa, respectively.
Fig. 10b presents the variations of MR with the bulk stress for
different blends of RCA with RCA(2). At the RAP content of 50%,
only a small increase in MR is observed. The blend with 75% RAP
content has much higher MR values than the clean RCA. The
increase of MR varies with stress level: DMR = 53 Mpa (equivalent
to 54%) at h = 85.5 kPa and DMR = 135 Mpa (or 51%) at
h = 510 kPa. Recall the blend with 75% RAP(1), the increase in MR
is almost a constant (DMR = 125 Mpa to 138 Mpa) in the same
range of the bulk stress.
The addition of RAP(3) into RCA also results in higher resilient
moduli for the blends, as shown in Fig. 10c. It is interesting to
note that the blends with 50% and 75% RAP content have practi-
cally the same resilient moduli, which are approximately 35 MPa
(or 36%) at h = 85.5 kPa and DMR = 133 Mpa (or 40%) at
h = 670 kPa higher than those of the virgin RCA at the correspond-
ing bulk stress levels.
The effect of RAP(4) on the resilient moduli of RCA blends was
found to be almost similar to that of RAP(3) for the same range
of bulk stress, as shown in Fig. 10d. Only marginal different is
observed when the RAP content is changed from 50% to 75%, with
DMR = 47 Mpa (equivalent to 48%) at h = 85.5 kPa and DMR = 106 -
Mpa (or 38%) at h = 670 kPa.
In conclusion, the addition of RAP to RCA induces the more sig-
nificant increase in the resilient moduli when compared with an
increase in MR value when RAP was added to granular samples.
These observations are in accordance with the findings docu-
mented by many researchers including MacGregor et al. [45], Alam
et al. [33], Bennert and Maher [42], Kim and Labuz [15].
6.3. Statistical analysis to understand the general effect of RAP content
on MR
Figs. 9 and 10 show that MR values of the blended samples are
higher when compared with the corresponding values of the fresh
granular material over a range of bulk stress (83–671 kPa).
92 M. Arshad, M.F. Ahmed / Construction and Building Materials 151 (2017) 83–97
However, to this end, statistical analysis is performed to deter-
mine, whether the increase in MR values is significant or not when
compared with the corresponding values of the fresh granular
material. For this purpose one-tail t-test was performed using
unequal variances of the two data sets as determined from f-test
(a)
(b)
(c)
(d)
Fig. 11. Accumulative strain under repeated loading at different cell pressures for:
(a) granular A and its blends; (b) granular F and its blends; (c) granular W and its
blends; (d) RCA and its blends.
and at an alpha value of 0.05 which corresponds to 95% confidence
level. On comparing the t-statistic with the t-critical it is found that
almost all the comparisons involving blends containing 75% RAP
contents have higher value of t-statistic when compared with the
t-critical (see Table 7). The interpretation of this outcome assumes
that the null hypothesis is rejected for these comparisons i.e., the
addition of 75% RAP contents has increased the MR values signifi-
cantly when compared with the corresponding values for the fresh
granular material. However, for almost all the blends containing
RAP content up to 50%, null hypothesis is acceptable i.e. statisti-
cally there is no significant effect of RAP content on increasing
the MR values of these blends when compared with the corre-
sponding MR values for the fresh granular materials.
6.4. Effect of RAP content on accumulative strain under cyclic loading
during resilient modulus test
In the Mechanistic-Empirical design of pavement, the resilient
moduli of the materials is an important material property required
for pavement performance analysis. On the other hand, pavement
distress is closely related to permanent deformation of pavement
materials. More specifically, the accumulative residual deforma-
tion of materials dominates rutting potential of the pavement
under repeated traffic loading. Consequently, it is equally impor-
tant to understand the effect of RAP on the residual deformation
of granular–RAP blends.
This section examines the effect of RAP on accumulative defor-
mation of various blends during resilient modulus testing. Even
though only 100 cycles of loading-unloading were applied under
each stress state, the effect of RAP content on residual deformation
can still be considered as qualitatively representative.
6.4.1. General observations and findings
Fig. 11a presents the evolution of vertical (axial) residual strain
in granular A and its blends with 50% and 75% RAP(1) and RAP(3),
respectively, during the resilient testing. In this figure, the horizon-
tal axis stands for the total number of load repetition in 15 loading
sequences described in AASHTO T307-99 [51]for base/subbase
materials. It should be noted that the curves in the figure only
reflects the history of residual strain in resilient modulus testing,
and the stress state varies at each load sequence. The following
observations are obtained from scrutinising the data in Fig. 11a:
1. The addition of RAP into the granular causes a significant
increase in the residual strain for the same stress history. The
higher the RAP content, the larger the residual strain.
2. For a given confining (cell) pressure r3, the residual strain after
100 load repetition increases with the cyclic deviator stress
amplitude. For example, when r3 = 68.9 kPa, the residual/incre-
mental (ratio) residual strain of the specimen having 75% RAP
(1) were 0.25%/0.03%, 0.42%/0.17%, and 1.0%/0.58% at the devi-
ator stress of 62.0 kPa, 124.1 kPa and 186.1 kPa respectively.
3. In addition to the stress state and RAP content, the residual
strain is affected by the properties of the RAP material. In other
words, blends having the same RAP content but different RAP
materials have completely different residual strain.
The above observations are generally applicable for all other
combinations of granular samples and RAP; as shown in Fig. 11b–d
for blends of granular sample F, W and RCA, respectively.
6.5. Statistical analysis to understand the general effect of RAP
contents on the residual strain during resilient modulus test
Statistical analysis of the Fig. 11is presented in Table 8 on the
basis of t-test (as discussed in Section 6.2.1). The t-critical and
 M. Arshad, M.F. Ahmed / Construction and Building Materials 151 (2017) 83–97 93

Table 8
t-Test analyses for the effect of RAP contents on accumulative strain during resilient modulus test.

Comparison of material type/blend T-Statistic T-Critical (one tail) Comments

Granular (A) versus 25% Granular (A) + 75% RAP(1) 3.4 1.76 Significant
Granular (A) versus 25% Granular (A) + 75% RAP(3) 2.13 1.73 Significant
Granular (A) versus 50% Granular (A) + 50% RAP(1) 1.06 1.74 Insignificant
Granular (A) versus 50% Granular (A) + 50% RAP(3) 0.82 1.72 Insignificant
25% Granular (A) + 75% RAP(1) versus 50% Granular (A) + 50% RAP(1) 2.04 1.72 Significant
25% Granular (A) + 75% RAP(3) versus 50% Granular (A) + 50% RAP(3) 1.43 1.73 Insignificant
Granular (F) versus 25% Granular (F) + 75% RAP(2) 2.94 1.76 Significant
Granular (F) versus 25% Granular (F) + 75% RAP(3) 2.61 1.75 Significant
Granular (F) versus 50% Granular (F) + 50% RAP(2) 2.74 1.74 Significant
Granular (F) versus 50% Granular (F) + 50% RAP(3) 2.34 1.74 Significant
25% Granular (F) + 75% RAP(2) versus 50% Granular (F) + 50% RAP(2) 0.43 1.70 Insignificant
25% Granular (F) + 75% RAP(3) versus 50% Granular (F) + 50% RAP(3) 0.30 1.68 Insignificant
Granular (W) versus 25% Granular (W) + 75% RAP(2) 2.33 1.75 Significant
Granular (W) versus 25% Granular (W) + 75% RAP(4) 2.17 1.74 Significant
Granular (W) versus 50% Granular (W) + 50% RAP(2) 1.88 1.74 Significant
Granular (W) versus 50% Granular (W) + 50% RAP(4) 2.35 1.57 Significant
25% Granular (W) + 75% RAP(2) versus 50% Granular (W) + 50% RAP(2) 0.57 1.70 Insignificant
25% Granular (W) + 75% RAP(4) versus 50% Granular (W) + 50% RAP(4) 0.18 1.70 Insignificant
RCA versus 50% RCA + 50% RAP(2) 3.13 1.75 Significant
RCA versus 25% RCA + 75% RAP(2) 3.45 1.77 Significant
RCA versus 50% RCA + 50% RAP(4) 3.14 1.76 Significant
RCA versus 25% RCA + 75% RAP(4) 3.14 1.77 Significant
25% RCA + 75% RAP(2) versus 50% RCA + 50% RAP(2) 0.70 1.71 Insignificant
25% RCA + 75% RAP(4) versus 50% RCA + 50% RAP(4) 0.78 1.70 Insignificant

450
(a) 140 50%Granular(A)+50%RAP(3)
Agg. A, 25% RAP 3
400
25%Granular(A)+75%RAP(3)
Agg. A, 75% RAP3
120 350 Granular(A)
Aggregate A
Axial stress (kPa)

100 300
Mc (MPa)

80 250

200
60
150
40 Granular(W)
Agg W
50%Granular(W)+50%RAP(1)
100
50% Agg. W+50%RAP(1)
20
25% Agg.W +75%RAP(1)
25%Granular(W)+75%RAP(1) 50
0
0
150 200 250 300 350 400 450
0 0.5 1 1.5 2
Axial stress (kPa) Axial strain (%)

(b) 140 Fig. 13. Influence of RAP content on residual settlement in 1D compression test on
granular A and its blends with RAP(3).
120

100
comparisons i.e. accumulated axial strain values of the blended
samples were statistically significant on the higher side when com-
MC (Mpa)

80
pared with the corresponding values for the virgin material at a
60
particular level of cell pressure and the number of load repetition
40 RCA during the resilient modulus test. This conclusion is more pro-
50%RCA+50%RAP(2)
nounced for the blends containing 75% of the RAP material and
20 25% virgin material because the difference between t-statistic and
25%RCA+75%RAP(2)
0
t-critical is more prominent for such materials when compared
150 200 250 300 350 400 450 with those blends containing 50% of the RAP material and 50% vir-
Axial stress (kPa) gin material.

Fig. 12. Variation in Mc values due to change in axial stress and percentage of RAP
content: (a) Granular W and its blends with RAP(1); (b) RAC and its blends with
7. Effect of RAP content on the constrained modulus (Mc) of
RAP(2). blended samples

t-statistic values for each t-test were calculated using unequal vari-
ances of the two data sets (as determined from f-test) and at an
alpha value of 0.05 which match to 95% confidence level. The
results of t-tests show that in general t-statistic is higher when
compared with the t-critical for all the comparison involving virgin
(fresh) materials and blended materials. This finding in return
leads to the interpretation that null hypothesis is rejected for these
The testing procedure and method for the determination of con-
strained modulus (Mc) has been described in Section 5.5. Typical
stress–strain loops are shown in Fig. 8 corresponding to different
levels of axial stress. For this research, Mc values are calculated at
four distinct levels of axial stress of 205, 275, 310 and 410 kPa
for different types of granular samples and their blends with RAP
materials.
94 M. Arshad, M.F. Ahmed / Construction and Building Materials 151 (2017) 83–97

(a) 3.0 value increases with axial stress and is affected by RAP content
Granular A
as well. For instance, in Fig. 12a, when the axial stress is increased
25% Granular (A)+75%RAP(1) from 205 kPa to 410 kPa, the value of Mc for test specimen pre-
2.5
pared from granular sample W increases from 73.8 Mpa to
Accumulative axial strain (%)

25% Granular (A)+75%RAP(2)

50% Granular (A)+50%RAP(1)
115 Mpa, which corresponds to 57.8% increase. For the same axial
2.0 50% Granular (A)+50%RAP(2)
stress variation (205 kPa–410 kPa), the value of Mc for the blend
prepared by mixing 50% granular W and 50% RAP (1) increases
1.5
by 32%, approximately. On the other hand, for the blend prepared
1.0
by mixing 25% granular W and 75% of RAP (1) the corresponding
increase in the Mc value remained limited to 18%. In other words,
0.5
205 kPa 275 kPa 310 kPa 410 kPa the variation of Mc tends to decrease as the amount of RAP content
is increased in the blend, however, Mc value certainly increases
0.0 with increase in axial stress. A similar trend in variation of Mc value
0 5 10 15 20 25 30 was observed for almost all the sample tested.
Number of load cycle (N)
Recall that the resilient modulus (MR) generally increases with
(b) 5.0 RAP content. The variation of Mc with RAP content shows a some-
4.5 Granular W what different pattern. This difference can be attributed to the
25% Granular (W)+75%RAP(1) effect of loading rate and the cohesion induced by RAP particles.
Accumulative axial strain (%)

4.0
25% Granular (W)+75%RAP(2)
3.5
RAP particles have some coating of asphalt as a binder agent, which
50% Granular (W)+50%RAP(1)
50% Granular (W)+50%RAP(2)
is, in fact, a viscous material. Since the resilient modulus test is a
3.0
dynamic test, the viscous asphalt binder coating around aggregate
2.5
particles tends to have higher resistance to deformation. However,
2.0 the 1D compression test (constrained modulus test) is a static load-
1.5 ing test in which viscous RAP particles offer less resistance to the
1.0 205 kPa 275 kPa 310 kPa 410 kPa external loading. Consequently, MR value tends to increase but
0.5 Mc value tends to decrease with the increase in RAP content. How-
ever, from the authors’ perspective, this trend should be further
0.0
0 5 10 15 20 25 30 investigated at a higher percentage of the binder contents with
Number of load cycle (N) varying physical properties (penetration, viscosity, flash point
etc.) in the RAP material. This is due to the fact that asphalt coating
(c) 6.0
Granualr F may provide some cohesion among the particles and this inter-
5.0
25% Granualr(F)+75%RAP(3) particle cohesion may dominate over the effect of loading rate.
25% Granualr(F)+75%RAP(4)
Accumulative axial strain (%)

50%Granualr(F)+50%RAP(3)
4.0 50% Granualr(F)+50%RAP(4) 7.2. Effect of RAP content on accumulative strain due to cyclic loading
during constrained modulus test
3.0
Similar to the resilient modulus test, the effect of RAP content
2.0 on residual deformation was also observed in the 1D compression
205 kPa 275 kPa 310 kPa 410 kPa test. A few unloading-reloading cycles were applied to the speci-
1.0 men at different stress levels (approximately 205, 275, 310 and
410 kPa) in these tests; as illustrated in Fig. 8.
0.0
Fig. 13 shows the stress-strain relationship for granular A and
0 5 10 15 20 25 30
Number of load cycle (N) its blends with 50% and 75% of RAP (3). It is obvious that the resid-
(d)8.0 ual vertical strain increases significantly with RAP content. The
100% RCA
majority of the residual strain developed in the initial loading.
7.0 25% RCA+75%RAP(3) However, scrutinising the measured data revealed that the accu-
25% RCA+75%RAP(4)
Accumulative axial strain (%)

6.0 50% RCA+50%RAP(3)

mulative deformation (strain) may become significant at higher
50% RCA+50%RAP(4) RAP content. This observation implies that for granular blends with
5.0 low RAP content, heavy initial compaction or preloading may sta-
4.0 bilise the material and reduce the accumulative settlement in the
future under the reptitive wheel loading. For increased RAP con-
3.0
tent, large accumulative settlement may take place, which is con-
2.0 sistent with observations about the residual deformation in the
205 kPa 275 kPa 310 kPa 410 kPa
resilient modulus testing.
1.0
Figs. 14a–d show the development of accumulative (residual)
0.0 axial strain in 1D compression tests (constrained modulus test)
0 5 10 15 20 25 30
for different blends as well as for fresh granular materials. From
Number of load cycle (N)
all of these figures, it can be interpreted that the residual strain
depends not only on the RAP content in the blend but also the type
Fig. 14. Effect of RAP content on accumulative strain during constrained modulus
of RAP for the respective number of load cycle corresponding dif-
test for blends containing: (a) granular A; (b) granular W; (c) granular F; (d) RCA.
ferent axial stress levels.
Fig. 14a summarizes the development of residual strain for
7.1. Variation of Mc value with stress level granular A and its blends containing RAP(1) and RAP(2). For blends
with 50% RAP content, the residual strain at different stress levels
Fig. 12a and b shows a variation of Mc value with axial stress for are higher than those of the fresh granular, and its rate of evolution
different materials. From this figure, it can be interpreted that Mc with axial stress is almost the same as the fresh granular. However,
 M. Arshad, M.F. Ahmed / Construction and Building Materials 151 (2017) 83–97 95

Table 9
t-Test analyses for the effect of RAP contents on accumulative strain during constrained modulus test.

Comparison of material type/blend T-Statistic T-Critical (one tail) Comments

Granular (A) versus 25% Granular (A) + 75% RAP(1) 10.24 1.83 Significant
Granular (A) versus 25% Granular (A) + 75% RAP(2) 9.35 1.86 Significant
Granular (A) versus 50% Granular (A) + 50% RAP(1) 4.32 1.80 Significant
Granular (A) versus 50% Granular (A) + 50% RAP(2) 8.87 1.78 Significant
25% Granular (A) + 75% RAP(1) versus 50% Granular (A) + 50% RAP(1) 9.21 1.81 Significant
25% Granular (A) + 75% RAP(2) versus 50% Granular (A) + 50% RAP(2) 7.42 1.82 Significant
Granular (W) versus 25% Granular (W) + 75% RAP(1) 9.68 1.85 Significant
Granular (W) versus 25% Granular (W) + 75% RAP(2) 9.31 1.85 Significant
Granular (W) versus 50% Granular (W) + 50% RAP(1) 11.23 1.83 Significant
Granular (W) versus 50% Granular (W) + 50% RAP(2) 7.91 1.83 Significant
25% Granular (W) + 75% RAP(1) versus 50% Granular (W) + 50% RAP(1) 3.64 1.78 Significant
25% Granular (W) + 75% RAP(2) versus 50% Granular (W) + 50% RAP(2) 3.16 1.76 Significant
RCA versus 50% RCA + 50% RAP(3) 10.96 1.83 Significant
RCA versus 25% RCA + 75% RAP(3) 9.54 1.86 Significant
RCA versus 50% RCA + 50% RAP(4) 15.09 1.87 Significant
RCA versus 25% RCA + 75% RAP(4) 10.15 1.85 Significant
25% RCA + 75% RAP(3) versus 50% RCA + 50% RAP(3) 0.47 1.75 Insignificant
25% RCA + 75% RAP(4) versus 50% RCA + 50% RAP(4) 5.90 1.75 Significant
Granular (F) versus 25% Granular (F) + 75% RAP(3) 10.22 1.85 Significant
Granular (F) versus 25% Granular (F) + 75% RAP(4) 8.17 1.76 Significant
Granular (F) versus 50% Granular (F) + 50% RAP(3) 6.25 1.83 Significant
Granular (F) versus 50% Granular (F) + 50% RAP(4) 14.55 1.81 Significant
25% Granular (F) + 75% RAP(3) versus 50% Granular (F) + 50% RAP(3) 5.34 1.81 Significant
25% Granular (F) + 75% RAP(4) versus 50% Granular (F) + 50% RAP(4) 6.42 1.85 Significant

when RAP content is increased to 75%, the accumulative strain was
10–16 times higher than that of the fresh granular under the same
axial stress.
The test results for other granular samples and their blends
with different RAP as presented in Figs. 14b–d are generally consis-
tent with the findings shown in Fig. 14a. For all cases, the residual
strain of blends increases with RAP contents. The value of the
residual strain may vary over a wide range, depending on the prop-
erties of both the granular samples and RAP. At the RAP content of
50%, however, the blends of granular F have much smaller residual
strain than the others.
According to the Fig. 14, the following conclusions can be
drawn:
1. The residual strain in blends is generally sensitive to the RAP
content. RAP content higher than 50% tends to induce very large
accumulative strain under cyclic loading;
2. The residual strains in blends of low RAP content can be sta-
bilised by initial loading to minimise the later accumulative
strain under cyclic loading;
3. It might be difficult to establish a relation between the residual
strain and RAP content since the properties of the aggregate
(possibly the internal structure of the aggregate) and the RAP
have a significant effect on the overall settlement of aggregate
blends.
the virgin material at a particular level of axial stress and number
of load repetition during the constrained modulus test.
8. Limitations of the study
1. The experimental results presented and discussed in this study
are based on only one type to fresh granular materials (crushed
limestone) although their gradation characteristics were
different.
2. RAP/RCA materials included in this study were of also limited
verity.
3. Asphalt binder contents of the RAP materials considered were
restricted from 3.3 to 4.7% although variation in their other
physical properties was broad.
4. The analysis of accumulative (residual) strain in based on 1500
number of load repetitions, however, practically a road/pave-
ment structure may be subjected to millions of load repetitions
during its lifetime.
5. A unique relation between the residual strains and RAP content
could not be obtained.
6. During the constrained modulus test only four distinct levels of
axial stress were considered with few number of load cycles.
9. Summary and conclusions

1. Mixing RAP materials with fresh granular samples results in

7.3. Statistical analysis to understand the general effect of RAP
contents on the residual strain during constrained modulus test
Statistical analysis of Fig. 14 is presented in Table 9 on the basis
of t-test. The t-critical and t-statistic values for each t-test were cal-
culated using unequal variances of the two data sets (as deter-
mined from f-test) and at an alpha value of 0.05 which equate to
95% confidence level. The results of t-tests show that t-statistic is
clearly higher when compared with the t-critical for all the com-
parison involving virgin materials and blended materials. This
result in return guide to the understanding that null hypothesis
is rejected for these comparisons i.e. accumulated axial strain val-
ues of the blended samples were statistically significant on the
higher side when compared with the corresponding values for
higher resilient moduli than obtained for fresh granular sam-
ples under the same loading conditions. These results are
encouraging, especially at such high RAP contents of 75%. How-
ever, development of higher residual strains under repeated
loading due to the addition of RAP contents may be detrimental
to the life of road pavement structure.
2. For all the tested materials, the resilient modulus increases with
the applied bulk stresses. The relation between the resilient
modulus and the bulk stress of a specific material can be
approximately expressed via a power law having a coefficient
of correlation values in the range of 0.96–0.99.
3. It is also found that the increase in the resilient modulus is more
significant at higher bulk stress levels for the blends containing
75% of the RAP contents. Additionally, statistical analysis
96 M. Arshad, M.F. Ahmed / Construction and Building Materials 151 (2017) 83–97
revealed that no significant increase in resilient modulus
occurred when RAP contents were increased from 50% to 75%
in the blended samples.
4. The increase in the accumulative residual strains during resili-
ent modulus test was significant at RAP contents of 75% on
comparing with relevant values for the fresh granular samples.
In most of the cases, the accumulative residual strains increase
was insignificantly when the RAP contents varied from 0 to 50%,
similarly in some of the cases no remarkable variation in the
accumulative residual strains could be identified when the
RAP contents increased from 50% to 75%.
5. The constrained modulus tests provide another approach to
investigate the stiffness and deformation properties of blends
containing granular samples and RAP materials with larger size
particles up to 50 mm. The increase of residual strains owing to
the addition of RAP contents into the fresh granular samples is
also observed in constrained modulus tests.
6. For the blended samples, the value of constrained modulus in
general decreases with the increase in RAP contents at a partic-
ular level of axial stress but this value certainly increases with
increase in applied axial stress for fresh granular and blended
samples.
7. The residual strains of the blended samples have higher pres-
sure sensitivity than that of the fresh granular sample during
the constrained modulus test.
8. More specifically, during the constrained modulus test (1D
compression tests) the addition of RAP into granular sample
induce more pronounced residual strain when compared to
those occurred during resilient modulus test at the same per-
centage of the RAP contents.
Acknowledgement
The first author gratefully acknowledges the research training
and assistantship provided by the Dr. Peijun Guo, Professor,
Department of Civil Engineering, McMaster University, Hamilton,
Canada. Authors also acknowledge the valuable comments and
suggestions of the reviewers for the improvement of this article.
References
[1] R. Taha, A. Al-Harthy, K. Al-Shamsi, et al., Cement stabilization of reclaimed
asphalt pavement aggregate for road bases and subbases, J. Mater. Civ. Eng. 14
(2002) 239–245.
[2] L.R. Hoyos, A.J. Puppala, C.A. Ordonez, Characterization of cement-fiber-treated
reclaimed asphalt pavement aggregates: preliminary investigation, J. Mater.
Civ. Eng. 23 (2011) 977–989.
[3] A.J. Puppala, L.R. Hoyos, A.K. Potturi, Resilient moduli response of moderately
cement-treated reclaimed asphalt pavement aggregates, J. Mater. Civ. Eng. 23
(2011) 990–998.
[4] C.S. Poon, D. Chan, Feasible use of recycled concrete aggregates and crushed
clay brick as unbound road sub-base, Constr. Build. Mater. 20 (2006) 578–585.
[5] A. Azam, D. Cameron, Geotechnical properties of blends of recycled clay
masonry and recycled concrete aggregates in unbound pavement construction,
J. Mater. Civ. Eng. 25 (2012) 788–798.
[6] A. Gabor, D. Cameron, Properties of recycled concrete aggregate for unbound
pavement construction, J. Mater. Civ. Eng. 24 (2012).
[7] A. Arulrajah, M. Ali, M. Disfani, et al., Geotechnical performance of recycled
glass-waste rock blends in footpath bases, J. Mater. Civ. Eng. 25 (2012) 653–
661.
[8] A. Arulrajah, J. Piratheepan, M. Disfani, et al., Geotechnical and
geoenvironmental properties of recycled construction and demolition
materials in pavement subbase applications, J. Mater. Civ. Eng. 25 (2012)
1077–1088.
[9] J. Wartman, D.G. Grubb, A. Nasim, Select engineering characteristics of crushed
glass, J. Mater. Civ. Eng. 16 (2004) 526–539.
[10] L.T. Lee Jr, Recycled Glass and Dredged Materials. 2007, DTIC Document.
[11] M. Ali, A. Arulrajah, M. Disfani, Suitability of using recycled glass-crushed rock
blends for pavement subbase applications, in Geo-Frontiers, et al., Advances,
Geotechnical Engineering. 2011 (2011) 1325–1334.
[12] M. Disfani, A. Arulrajah, M. Bo, et al., Environmental risks of using recycled
crushed glass in road applications, J. Clean. Prod. 20 (2012) 170–179.
[13] M.A. Imteaz, M.Y. Ali, A. Arulrajah, Possible environmental impacts of recycled
glass used as a pavement base material, Waste Manage. Res. 30 (2012) 917–
921.
[14] A. Arulrajah, J. Piratheepan, M. Disfani, et al., Geotechnical and
geoenvironmental properties of recycled construction and demolition
materials in pavement subbase applications, J. Mater. Civ. Eng. 25 (2013)
1077–1088.
[15] W. Kim, J. Labuz, Resilient modulus and strength of base course with recycled
bituminous material, Final report for the Minnesota Department of
Transportation project (2007) (Subcontract No. 2005-xx).
[16] E.A.P.A. (Eapa), Asphalt in Figures 2011. 2011.
[17] K.R. Hansen, A. Copeland, 2nd Annual Asphalt Pavement Industry Survey on
Reclaimed Asphalt Pavement, Reclaimed Asphalt Shingles, and Warm-Mix
Asphalt Usage: 2009–2011. (2013).
[18] M. Zaumanis, R.B. Mallick, R. Frank, 100% recycled hot mix asphalt: A review
and analysis, Resour. Conserv. Recycl. 92 (2014) 230–245.
[19] I.L. Al-Qadi, M. Elseifi, S.H. Carpenter, Reclaimed asphalt pavement—a
literature review (2007).
[20] Y. Huang, R.N. Bird, O. Heidrich, A review of the use of recycled solid waste
materials in asphalt pavements, Resour. Conserv. Recycl. 52 (2007) 58–73.
[21] X. Li, M. Marasteanu, R. Williams, et al., Effect of reclaimed asphalt pavement
(proportion and type) and binder grade on asphalt mixtures, Transport. Res.
Rec. (2008) 90–97.
[22] M. Zaumanis, R.B. Mallick, Review of very high-content reclaimed asphalt use
in plant-produced pavements: state of the art, Int. J. Pavement Eng. 16 (2015)
39–55.
[23] R. Izaks, V. Haritonovs, I. Klasa, et al., Hot mix asphalt with high RAP content,
Procedia Eng. 114 (2015) 676–684.
[24] D. Bloomquist, G. Diamond, M. Oden, et al., Engineering and environmental
aspects of recycled materials for highway construction; appendix 1: final
report (1993).
[25] M. Solaimanian, E. Savory, Variability analysis of hot-mix asphalt concrete
containing high percentage of reclaimed asphalt pavement, Transport. Res.
Rec. (1996) 89–96.
[26] A. Hussain, Q. Yanjun, Evaluation of Asphalt Mixes Containing Reclaimed
Asphalt Pavement for Wearing Courses, Proceedings. International Conference
on Traffic and Transportation Engineering, School of Civil Engineering (2012).
[27] M. Maher, N. Gucunski, W. Papp Jr, Recycled asphalt pavement as a base and
sub-base material, American Society for Testing and Materials, West
Conshohocken, PA (United States), 1997.
[28] T. Park, Application of construction and building debris as base and subbase
materials in rigid pavement, J. Transp. Eng. 129 (2003) 558–563.
[29] R. Taha, Evaluation of cement kiln dust-stabilized reclaimed asphalt pavement
aggregate systems in road bases, Transp. Res. Rec. (2003) 11–17.
[30] B. Blankenagel, W. Guthrie, Laboratory characterization of recycled concrete
for use as pavement base material, Transp. Res. Rec (2006) 21–27.
[31] Y.-H. Cho, T. Yun, I.T. Kim, et al., The application of recycled concrete aggregate
(RCA) for hot mix asphalt (HMA) base layer aggregate, KSCE J. Civil Eng. 15
(2011) 473–478.
[32] J. Piratheepan, A. Arulrajah, M. Disfani, Large-scale direct shear testing of
recycled construction and demolition materials, Adv. Civil Eng. Mater. 2 (2013)
25–36.
[33] T.B. Alam, M. Abdelrahman, S.A. Schram, Laboratory characterisation of
recycled asphalt pavement as a base layer, Int. J. Pavement Eng. 11 (2010)
123–131.
[34] W.H. Chesner, R.J. Collins, M. Mackay, User guidelines for waste and by-
product materials in pavement construction. (1998).
[35] R. Davio, Lessons learned: TxDOT’s efforts to increase the use of recycled
materials, Public Roads. 64 (2000).
[36] A.T. Papagiannakis, E.A. Masad, Pavement design and material, Wiley, New
York, 2007.
[37] G. Tebaldi, E.V. Dave, P. Marsac, et al., Synthesis of standards and procedures
for specimen preparation and in-field evaluation of cold-recycled asphalt
mixtures, Road Mater. Pavement Des. 15 (2014) 272–299.
[38] G. Tebaldi, E. Dave, P. Marsac, et al. Classification of recycled asphalt pavement
(RAP) material, in: Proceedings of the 2nd international symposium on asphalt
pavements and environment, Fortaleza, Brazil (2012).
[39] P. Kandhal, R. Mallick, Pavement recycling guidelines for state and local
governments: Participants reference book. Final report, September 1995–
December National Center for Asphalt Technology, Auburn, AL (United
States); Federal Highway Administration, Washington, DC (United States)
(1997).
[40] D. Roddier, C. Cermelli, A. Aubault, et al., WindFloat: A floating foundation for
offshore wind turbines, J. Renewable Sustainable Energy. 2 (2010) 033104.
[41] A. Saeed, N.C.H.R. Program, N.R.C.T.R. Board, Appendixes to NCHRP Report 598.
2007: National Cooperative Highway Research Program, Transportation
Research Board of the National Academies (2007).
[42] T. Bennert, A. Maher, The development of a performance specification for
granular base and subbase material. Publication FHWA-NJ-05-003.
Washington, DC: FHWA, US Department of Transportation (2005).
[43] R. Taha, G. Ali, A. Basma, et al., Evaluation of reclaimed asphalt pavement
aggregate in road bases and subbases, Transp. Res. Rec. (1999) 264–269.
[44] P.J. Cosentino, E.H. Kalajian, A.M. Bleakley, B.S. Diouf, T.J. Misilo, A.J. Petersen,
R.E. Krajcik, A.M. Sajjadi, Improving the properties of reclaimed asphalt
pavement for roadway base applications, Florida Institute of Technology,
Melbourne (2012).
 M. Arshad, M.F. Ahmed / Construction and Building Materials 151 (2017) 83–97
[45] J. MacGregor, W. Highter, D. DeGroot, Structural numbers for reclaimed
asphalt pavement base and subbase course mixes, Transp. Res. Rec. 1687
(1999) 22–28.
[46] W.J. Papp, M.H. Maher, T.A. Bennert, N. Gucunski, in: Behavior of Construction
and Demolition Debris in Base and Subbase Applications, Geotechnical Special
Publication, 1998, pp. 122–135.
[47] S.M. Sayed, J.M. Pulsifer, R.C. Schmitt, Construction and performance of
shoulders using unrap base, J. Mater. Civ. Eng. 5 (1993) 321–338.
[48] B. Trzebiatowski, C. Benson, Saturated hydraulic conductivity of compacted
recycled asphalt pavement, Geotech. Test. J. 28 (5) (2005) 1–6.
[49] A.H. Aljassar, K.B. Al-Fadala, M.A. Ali, Recycling building demolition waste in
hot-mix asphalt concrete: a case study in Kuwait, J. Mater. Cycles Waste
Manage. 7 (2005) 112–115.
[50] R.L. Mokwa, C.S. Peebles, Evaluation of the engineering characteristics of RAP/
aggregate blends. Publication FHWA-MT-05-008. Washington, DC: FHWA, US
Department of Transportation (2005).
[51] E.J. Mcgarrah, Evaluation of current practices of reclaimed asphalt pavement/
virgin aggregate as base course material. No. WA-RD 713.1 (2007).
[52] Senior, B. Gorman, S. Szoke. Recycling road building materials and experience
with full depth reclamation in the Ontario provincial highway system, in:
97
Proceedings of the Transportation Association of Canada Annual Conference
recycled materials and recycling processes for sustainable infrastructure
session, Toronto, Ont. (2008).
[53] AASHTO Designation T 27 (99), Sieve Analysis of Fine and Coarse Aggregates,
American Association of State Highway and Transportation Officials (1999).
[54] AASHTO T (180), Modified Method of Test for the Moisture-Density Relations
of Soils, American Association of State Highway and Transportation Officials
(2006).
[55] R.D. Barksdale, J. Alba, N.P. Khosla, et al., Laboratory determination of resilient
modulus for flexible pavement design. No. NCHRP Project 1-28 (1997).
[56] AASHTO T307-99, Determining the Resilient Modulus of Soils and Aggregate
Materials, American Association of State Highway and Transportation Officials
(2007).
[57] G. Peijun, E. John, Importance of strain level in evaluating resilient modulus of
granular materials, Int. J. Pavement Eng. 12 (2011) 187–199.
[58] S. Dieter, G. Peijun, L. Ying, Resilient modulus properties of granular highway
materials, Can. J. Civ. Eng. 36 (2009) 639–654.