Construction and Building Materials 351 (2022) 128322

Contents lists available at ScienceDirect

Construction and Building Materials

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

Laboratory and field evaluation of plant produced asphalt mixtures

containing RAP in hot climate: A case study from Phoenix, Arizona
Ali Zalghout a, *, Samuel Castro b, Jolina Karam b, Kamil Kaloush b
a
GMU Geotechnical, Inc., 23241 Arroyo Vista, Rancho Santa Margarita, CA 92688, USA
b
Department of Civil, Environmental, and Sustainable Engineering, Arizona State University, Tempe, AZ 85281, USA

A R T I C L E I N F O A B S T R A C T

Keywords: The use of Reclaimed Asphalt Pavement (RAP) in asphalt mixtures has been gaining wide attention over the past
RAP four decades. However, the use of RAP in asphalt mixtures in harsh hot climate areas such as Phoenix, Arizona
Hot climate has been slowly and cautiously addressed due to the aged characteristics of the RAP and lack of performance
Cracking potential
data. This study aimed at evaluating such RAP mixtures in the laboratory using field produced Hot Mix Asphalt
Field performance
Mixtures (HMA). Based on detailed mix design efforts, three test sections were constructed in Phoenix: the first
being a control (0% RAP), the second and the third sections with 15% and 25% RAP contents, respectively.
During construction, loose HMA mixtures were sampled for advanced laboratory material characterization tests.
The testing program included stiffness, permanent deformation, cracking potential, and moisture damage
resistance. In addition, asphalt binder testing was performed on extracted and recovered binders. The laboratory
results on the RAP mixtures indicated that adding 15% and 25% RAP to the mix did not have significant effect on
increasing stiffness, but improved the rutting resistance, had comparable cracking performance, and had
acceptable moisture damage resistance. The binder testing that was performed on the extracted and recovered
binders yielded results that supported the asphalt mixtures’ testing analysis. The field performance monitoring of
the three test sections showed similar performance. Based on the results of this study, it was recommended to
incorporate RAP as a sustainable paving activity up to the percentages used in this study.

1. Introduction content, and this increase was also noticed after testing mixtures sub­
jected to various freeze-thaw cycles [2]. Gardiner and Wanger found
The use of Reclaimed Asphalt Pavement (RAP) as a recycled that as the stiffness increased with the addition of RAP, the permanent
component in newly produced Hot Mix Asphalt (HMA) reduces the deformation resistance increased as well [3]. However, other studies
required amount of virgin aggregates and virgin binder, leading to cost reported differing results for RAP mixtures having similar or lower
reduction, energy conservation as well as increased sustainable pave­ stiffness compared to conventional mixtures, depending on the source of
ment practices. The interest in incorporating RAP dates to the 1970s, RAP and the base asphalt binder used [4–5].
during the oil embargo. Since that time, agencies started incorporating Other studies evaluated the fatigue cracking behavior of asphalt
more RAP in their mixtures to expand their sustainable practices. In mixtures containing RAP; the results also reported conflicting trends.
order to have a successful use of RAP in HMA, its inclusion should not One study reported that the addition of RAP at 10%, 20% and 30% using
compromise pavement performance. However, as asphalt pavements softer and stiffer binders (PG 64–22 and PG 76–22), increased the fa­
age during their service life, the stiffening effect, especially in hot cli­ tigue resistance to some extent [6]. Hajj et al. reported that the addition
mates, makes it difficult to consider RAP for future uses. of RAP resulted in either better or worse fatigue resistance, depending
Several studies focused on characterizing the properties of mixtures on the RAP source and content [7]. Another study reported that for
containing RAP and their performance. One study reported that by higher RAP contents, the mixtures became stiffer and the fatigue resis­
increasing the RAP content from 15% to 40%, the stiffness increased tance was compromised [8]. By adding a polymer modified binder or by
significantly at the higher testing temperatures [1]. Another study re­ using SBS-modified RAP, the fatigue resistance was found to be lower,
ported an increase in the dynamic modulus with the increase in RAP possibly due to an increased stiffening effect [7–8].

* Corresponding author.
E-mail addresses: [email protected] (A. Zalghout), [email protected] (S. Castro), [email protected] (J. Karam), [email protected] (K. Kaloush).

https://doi.org/10.1016/j.conbuildmat.2022.128322
Received 13 March 2021; Received in revised form 2 February 2022; Accepted 30 June 2022
Available online 2 September 2022
0950-0618/© 2022 Elsevier Ltd. All rights reserved.
A. Zalghout et al. Construction and Building Materials 351 (2022) 128322

Fig. 1. Flowchart of the Experimental Program.

As for the cracking potential of RAP mixtures, RAP contents of 20% With the unclear expectations of the effect of adding RAP on City of
and 40% were found to greatly affect the fracture resistance (cracking Phoenix asphalt mixtures, and with the harsh hot climate that Phoenix
potential increased) of asphalt mixtures using the Semi-Circular Bending experiences every year, conducting a comprehensive study to evaluate
(SCB) fracture test [9]. Ozer et al. evaluated the cracking properties of the performance of asphalt pavements with RAP is needed. This study
asphalt mixtures containing recycled materials using the Illinois Semi­ focuses on evaluating the laboratory and field performance of City of
circular Bending (IL-SCB) test [10]. The fracture energy results (tested at Phoenix asphalt mixtures that incorporate RAP with different contents.
25 ◦ C) indicated a consistent drop in fracture energy as the recycled
materials content increased. Moreover, the Flexibility Index (FI) showed 1.1. Objective
a consistent reduction with the increase in the recycled materials con­
tent. In general, with the increase in stiffness, mixtures with high RAP The objectives of the study presented herein are to (i) identify local
contents show a higher indirect tensile strength but lower toughness RAP practices in Phoenix Area and the surrounding states, and (ii)
index and strain at peak loading [11]. investigate the effect of using two different RAP contents (15% and
Regarding the potential moisture damage of RAP mixtures, some 25%) on the City of Phoenix newly produced asphalt mixtures. This was
studies have reported an acceptable moisture damage resistance with done through evaluating the laboratory and field performance of asphalt
higher RAP contents [2,7]. However, another study showed that aged mixtures containing these two percentages.
asphalt is more susceptible to moisture damage [12].
As literature shows, there is no clear expectation on the behavior of 1.2. Methodology
RAP mixtures. Researchers reported different results most likely affected
by factors such as material variability, base binder type, climate and The overall experimental process is shown in Fig. 1, followed by
aging levels. Most studies reported a stiffening effect when RAP is added, methods and details for each block/section.
leading to better rutting resistance but worse cracking behavior. How­
ever, the performance of pavements with RAP in a harsh hot climate, 2. Materials and methods
such as in Arizona, has not been thoroughly evaluated.
According to a survey conducted by the National Asphalt Pavement 2.1. Local practices on RAP usage
Association (NAPA), Arizona (at the State level) does not use more than
15% of RAP in its asphalt mixtures [13]. However, there is an interest to To understand the current practices related to RAP usage in Arizona
study the effect of using higher RAP contents on pavement performance and the nearby states with similar climatic conditions, a survey on the
in hot climates. A research study by Arredondo et al. [14] evaluated the use of RAP by these agencies was conducted. Each agency was asked if
viability of using RAP in future pavement maintenance and rehabilita­ they use RAP in asphalt pavements, in which asphalt concrete layer,
tion projects for the City of Phoenix, Arizona. The dynamic modulus and and/or whether RAP is permitted to be used in surface layers.
flow number test results showed no statistical difference between the Overall, the State of Arizona, including its municipalities and county
control mix, 15% and 25% (with lower binder grade) RAP mixtures. The agencies, have been slow to adopt the use of RAP as a regular practice.
Tensile Strength Ratio (TSR) showed that all mixtures performed well However, many agencies within the state have specifications in place
and above the specified minimum limit of 75% required by City speci­ and/or practices for using RAP materials. The outcome of the survey
fications. The study concluded that the 15% RAP content is feasible to with additional notes and comments is summarized in Table 1.
use and will not affect pavement performance based on the preliminary RAP usage in the City of Phoenix is largely based on the City’s sup­
laboratory performance testing. However, the study did not include field plement to the 2015 Maricopa Association of Governments (MAG)
production assessment or cracking potential assessment. Standard Specification. RAP is not allowed to be used in Asphalt

2
A. Zalghout et al. Construction and Building Materials 351 (2022) 128322

Table 1
Summary of Survey Results.
Agency RAP Usage Specifications

Asphalt Concrete Unbound Other

Base
Surface Non–
Surface

Phoenix No No Conditional Dust control City of Phoenix Supplement to

Dirt street stabilization MAG Uniform Standard Specifications [15]
Tucson No Yes Yes Dust control PAG Standard Specifications [16]
Shoulders
Dirt roads
Arizona Dept. of Trans. Yes Yes Yes Miscellaneous asphaltic ADOT’s Policy and Procedure Directive No.20 for the use of RAP in
(ADOT) concrete asphaltic concrete [17]
Maricopa Assoc. of Gov. Yes Yes Yes Shoulders Revision to the MAG Uniform Standard Specifications [18]
(MAG) Dirt roads
Pima Assoc. of Gov. (PAG) Yes Yes Yes —— PAG Standard Specifications [16]
Maricopa Co. Dept. of No Yes Conditional Shoulders Maricopa County DOT Supplement to the MAG Uniform Standard
Trans. (MCDOT) Dirt roads Specifications [19]
Pima Co. Dept. of Trans. Yes Yes Yes Shoulders PAG Standard Specifications [16]
(PCDOT) Dirt roads
Scottsdale No — — Dust control City of Scottsdale Supplement to MAG Uniform Standard Specifications
Dust street stabilization [20] and EVAC [21]
Not for backfills
East Valley Asphalt Comm. No Yes No Structural backfill EVAC Hot Asphalt Mix Criteria [21]
Dust control
Dirt roads
Apache Junction No No Yes Dirt road stabilization No specifications
Shoulders on rural roads &
urban arterials
Mesa No Yes Yes Shoulders Backfills Amendments to MAG Uniform Standard Specifications and EVAC [21]
Chandler — — — Dust proofing Supplement to MAG Uniform Standard Specifications [22] and EVAC
[21]
Gilbert No No No No Town of Gilbert Supplement to MAG Uniform Standard Specifications
[23] and EVAC [21]
Queen Creek No No Yes Shoulders Revision to the MAG Uniform Standard Specifications and EVAC [21]
Dust control
Dirt road stabilization
Las Vegas (Nevada) Yes Yes Yes Structure granular backfill Uniform Standard Specifications of RTCSNV [24]
Nevada Dept. of Trans. Yes Yes Yes Shouldering Nevada DOT Standard Specifications for Road and Bridge Construction
(NDOT) [25]
Texas Dept. of Trans. Yes Yes Yes Subgrade stabilization Texas DOT Standard Specifications for Construction and Maintenance
(TXDOT) of Highways, Streets, and Bridges [26]
New Mexico Dept. of Trans. Yes Yes Yes Not for backfills New Mexico DOT Standard Specifications for Highway and Bridge
(NMDOT) Construction [27]
California Dept. of Trans. Yes Yes Yes Shoulder backing California State Transportation Agency, Department of Transportation,
(Caltrans) Caltrans [28]

Fig. 2. Pavement Structure of the Three Sections.

3
A. Zalghout et al. Construction and Building Materials 351 (2022) 128322

Table 2
Binder Properties: PG 70–10 and PG 64–16.
Test Test Temperatures for PG 70–10/PG 64–16 Test Results Specification

PG 70–10 PG 64–16

Tests on Original Binder Flash Point, T48 >230 ◦ C >230 ◦ C Min. 230 ◦ C

Apparent Viscosity, AASHTO T 316 135 ◦ C/135 ◦ C 0.565 Pa-s 0.428 Pa-s Max. 3 Pa-s

175 C/175 C
◦ ◦
0.101 Pa-s 0.082 Pa-s

Dynamic Shear, T315, G*/sin δ 70 ◦ C/70 ◦ C 1.19 kPa 1.62 kPa Min. 1.00 kPa

Tests on Residue from RTFO Mass Change − 0.143 − 0.106 Max 1.0 %

Dynamic Shear, T315, G*/sin δ 70 ◦ C/64 ◦ C 3.05 kPa 3.85 kPa Min. 2.20 kPa

Tests on Residue from PAV PAV Aging Temperature 110 C/100 C

◦ ◦

Dynamic Shear, T315, G*sin δ 34 ◦ C/28 ◦ C 3840 kPa 3790 kPa Max. 5000 kPa

Creep Stiffness, T313 0 ◦ C/− 6◦ C 93.0 MPa 117 MPa Max. 300 MPa

m-value, T313 0 ◦ C/− 6◦ C 0.312 0.335 Min. 0.300

Fig. 3. Aggregate Gradation of the Three Asphalt Base Mixtures.

Concrete and its use in other type of fill requires prior approval by the
Table 3
Engineer. RAP is not allowed to be used as base material without
Volumetric Properties of the Three Base Mixtures.
approval from the City of Phoenix. Based on anecdotical information,
the main use of RAP is for road dust control and dirt street stabilization. 0% RAP 15% RAP 25% RAP
(Control)

2.2. Mix design and sections construction Total Binder Content (%) 5.00 5.00 5.00
Contributed RAP Binder Content (%) 0.00 0.59 0.99
Added Virgin Binder Content (%) 5.00 4.41 4.01
2.2.1. Pavement structure Marshall Bulk Density (kg/m3) 2,371 2,381 2,389
In this project, a pavement structure that is typically used by the City Theoretical Maximum Specific 2.478 2.481 2.486
of Phoenix was utilized for the test sections, except that RAP was used in Gravity
Theoretical Maximum Density (kg/ 2,476 2,480 2,485
the asphalt base layer. These sections were constructed using a mixture
m3)
with a Nominal Maximum Aggregate Size (NMAS) of 19 mm (¾ inches) Stability (N) 22,285 23,975 23,175
in the base layer, and a Polymer Modified Asphalt Rubber (PMAR) Marshall Flow (mm) 2.794 2.54 2.794
mixture with NMAS of 12.7 mm (½ inches) in the surface layer. Fig. 2 Air Voids (%) 4.3 3.9 3.8
shows a schematic of the three sections’ pavement structure. Section 1 Voids in Mineral Aggregates (%) 14.5 14.5 14.2
Air Voids Filled with Asphalt (%) 70.5 72.7 72.8
was a control, where Section 2 had a RAP replacement of 15% in the Effective Asphalt Total Mix (%) 4.39 4.52 4.41
asphalt base layer, and Section 3 had a 25% RAP replacement in the Film Thickness (micro) 9 9 9
asphalt base layer. Dust/Bitumen Ratio 1.1 1 1.1

2.2.2. Materials
RAP, and PG 64–16 for the 25% RAP mixture. The binders’ properties
The materials used in this project were the most widely used ag­
are shown in Table 2.
gregates and binders in Phoenix. The RAP used was millings processed
and stockpiled at a local asphalt plant.
2.2.2.2. Aggregates. The aggregates used in the base layer mixtures had
a NMAS of 19 mm (¾ inches), whereas the surface PMAR mix had a
2.2.2.1. Asphalt binder. The asphalt binders used were a Superpave
NMAS of 12.7 mm (½ inch). Fig. 3 shows the gradation of the control
performance-graded binders, PG 70–10 for mixtures with 0% and 15%

4
A. Zalghout et al. Construction and Building Materials 351 (2022) 128322

Fig. 4. Sections Construction and Material Sampling.

base mixture and those with RAP. 3. Theory/calculations

2.2.3. Mix designs 3.1. Description of mixture testing methods

The Marshall mix design method was used to design the asphalt
mixtures. The Maricopa Association of Governments (MAG) guidelines 3.1.1. Dynamic modulus test
were followed during the mix design process. MAG specifications The dynamic modulus test was conducted at four different temper­
require a drop of one binder grade at both the low and high temperatures atures (4.4, 21.1, 37.7 and 54.4 ◦ C) and six different frequencies (25, 10,
when RAP content higher than 15% is used [15]. Thus, in the current 5, 1, 0.5 and 0.1 Hz) following the AASHTO TP 62 standard [29]. The
mix design, PG 64–16 was used for the 25% RAP mix instead of PG time-temperature superposition principle was used to determine the
70–10. The mixtures were designed with 75 blows on each side. Table 3 temperature-dependent mechanical properties at a reference tempera­
shows the volumetric properties of the three base mixtures. ture. The master curve of each mix was developed at a reference tem­
perature of 21.1 ◦ C. These master curves were constructed by shifting
2.2.4. Project location, description, and sampling the isothermal curves horizontally using Eqs. (1), (2) and (3).
The test sections were constructed on 15th Avenue from Roeser Road α
to Broadway Road in Phoenix, Arizona. The total project length was log|E* | = δ+ 1
(1)
1 + eβ+γ(logf
approximately 2,575 feet. These test sections were constructed in
r)

December 2018. Road Sections were constructed following City of log(fr ) = log(f ) + log(aT ) (2)
Phoenix specifications. Loose asphalt mixtures were sampled at the
plant. Trucks were stopped randomly before going out to the site and Log(aT ) = aT 2 + bT + c (3)
metal buckets were used to sample the asphalt mixtures. Loose mixtures
were sampled from three random trucks for each mix to ensure that the Where
sampled material was representative to the total produced asphalt |E*| = dynamic modulus;
mixture. These buckets were transported to the laboratory for further f = loading frequency;
processing and testing. Fig. 4 shows views of the construction and fr = reduced frequency of loading;
sampling process. α, β, δ, γ = fitting parameters;
aT = temperature shift factor;
2.2.5. Reheating protocol and sample preparation T = Temperature; and
The buckets were processed by placing it in the oven for 90 minutes a, b, and c = shift factor coefficients determined by optimization.
at 135 ◦ C. After that, the buckets from the three trucks were mixed
together for each mix, and then split into bags of uniform gradations as 3.1.2. Flow Number (FN) test
much as possible, minimizing segregation. These bags were stored at a The Flow Number (FN) test was conducted following the AASHTO TP
controlled temperature condition to prevent any potential aging, and 79 standard [30]. The test was performed at a constant temperature of
were used to prepare samples for the testing process. 50 ◦ C, which is similar to the effective pavement temperature at the
Samples were prepared by placing the loose mixtures (removed from studied location. The Francken model (Eq. (4)) was used to model the
bags) in the oven at the compaction temperature for 2 hours, then filling permanent strain curve. The parameters a, b, c, and d were determined
the material in the compaction mold. The mold was placed back in the using the nonlinear regression analysis. To determine the flow number
oven, and samples were compacted using the gyratory compactor when (FN) value (i.e., inflection point), the second derivative (Eq. (5)) is set to
the asphalt mixture reached the compaction temperature. zero.

εp (N) = a⋅N b + c(ed⋅N − 1) (4)

∂2 εp
= ab(b − 1)N b− 2 + cd 2 edN (5)
∂N 2

5
A. Zalghout et al. Construction and Building Materials 351 (2022) 128322

Where: 3.1.5. Uniaxial fatigue test

εp(N) = Permanent strain at N cycles; The uniaxial fatigue test was conducted to evaluate the effect of
N = Number of cycles; and adding RAP on the fatigue life of the asphalt mixtures. The test was
a, b, c, d = Regression coefficients. conducted at 18 ◦ C by subjecting a cylindrical specimen of 150 mm
height and 75 mm diameter to sinusoidal displacement. The failure
3.1.3. Indirect Tensile Asphalt Cracking Test (IDEAL-CT) criteria used in this test is the drop in phase angle. The test was con­
The IDEAL-CT [31] was conducted to evaluate the cracking potential ducted per AASHTO TP 107 [34].
of the three mixtures. The test is a typical Indirect Tensile Test (IDT) but The test was conducted at four strain levels, which were estimated
with different analysis. The test was conducted at 25 ◦ C and at a loading such that the sample will fail in less than 10,000 cycles, between 10,000
rate of 50 mm/min following the standard procedure. and 50,000 cycles, between 50,000 and 100,000 cycles, and greater than
The output of this test is a load-displacement curve. The slope at the 100,000 cycles. The Simplified Viscoelastic Continuum Damage (S-
point corresponding to 75% of the maximum load is determined and VECD) model was used to analyze the test results for the potential fa­
considered as a post peak behavior. Studies have determined that this tigue performance of the mixtures. The result of this analysis is the curve
point is typically the inflection point of the post peak behavior and can (C vs. S) where C is the secant pseudo stiffness and S is the damage
identify the brittleness or ductility of the material [31]. Based on all parameter. The power function shown in Eq. (10) was used to fit the
these parameters, the Cracking Tolerance Index (CTI) is calculated using curves.
Eq. (6). As it can be seen, the higher the fracture energy, the higher the C = 1 − C11 SC12 (10)
work needed to fracture the material, thus the higher the CTI. From the
other side, the lower the slope, the better ductility or post peak behavior, Where,
the higher the CTI. Thus, higher CTI values correspond to better cracking C = pseudo stiffness;
resistance. C11 and C12 = fitting coefficients; and
t I75 Gf S = damage parameter.
CTI = × × × 106 (6) Although the C vs. S is a good indication of the performance and can
62 D m75
indicate the level of damage sustained before failure, it does not directly
Where: indicate the fatigue performance. In order to predict the number of cy­
CTI = Cracking Tolerance Index; cles needed for fatigue failure at different strain levels, Eq. (11) was
W used:
Gf = failure energy, Gf = D*tf , where Wf is the area below the load
( ) − αC12 +1
displacement curve; (fr ) 23α Sαfailure
m75 = absolute value of the post-peak slope; Nfailure = [ ( )( ⃒ ) ]2α (11)
(α − αC12 + 1)(C11 C12 )α (β + 1) ε0.pp ⃒E* | K1
I75 = displacement at 75% the peak load after the peak; LVE

D = specimen diameter; and

Where:
t = specimen thickness.
Nfailure = predicted cycles to failure;
fr = reduced frequency;
3.1.4. C* fracture test
|E*| = dynamic modulus at the frequency and temperature of
In a viscoelastic material, C* can be defined as the energy rate line
loading simulated;
integral that describes the stress and strain rate field surrounding the
α = viscoelastic damage rate;
crack tip. The C* can be measured experimentally due to the relationship
β = load form factor, taken as 0 in this work to simulate reversed
between the J-integral and C* parameter. J can be defined as the energy
sinusoidal loading;
difference between two specimens subjected to the same load yet have
ε0.pp = the peak-to-peak strain magnitude for the simulated loading
different crack length [32–33] as expressed in Equation (7).
history;
dU K1 = loading shape factor; and.
J= − (7)
da Sfailure = damage level at failure (defined from the experimental
results).
Where,
U = Potential energy; and 3.1.6. Indirect Tensile Strength/Modified Lottman test
a = Crack length. Moisture damage is caused by the loss of adhesion between binder
Thus, C* can be calculated as power or energy rate difference be­ and aggregate in the presence of moisture. The Modified Lottman testing
tween two specimens, under same loading, with incrementally different method was conducted for the three mixtures following the AASHTO T
crack length. Mathematically, C* can be expressed by Eq. (8) below. 283 procedure [35]. The Indirect Tensile Test (IDT) was performed on
dU * each sample. The TSR is the ratio of the tensile strength of the condi­
C* = − (8) tioned set to that of the unconditioned one.
da
The unconditioned subset consisted of samples stored at room tem­
Where, perature, whereas the conditioned one consisted of samples that
U* = Power or energy rate for a given load P. received partial vacuum saturation under a partial pressure of 26 in Hg
The rate of work done (U*) is defined as the area under the load “P” for 5 to 10 minutes. The conditioned samples were then submerged into
vs. displacement “u” curve. It is calculated using the Eq. (9) below: water for another 5 minutes. After that, this subset was conditioned at
∫u − 16 ◦ C for at least 16 hours, followed by placing it in a hot water bath at
U* = Pdu (9) 60 ◦ C for 24 hours.
0 The Indirect Tensile Test (IDT) was then performed on each sample.
The test was conducted at 10 ◦ C, by applying a load on multiple It consisted of applying a compressive load at a controlled deformation
asphalt disks at constant displacement rates, and the crack length over rate of 50 mm/min. Finally, the Tensile Strength was determined.
time was measured for each disk. The testing was performed at five
different displacement rates (0.15, 0.23, 0.3, 0.38, and 0.42 mm/min) in
which each disk was tested at one displacement rate.

6
A. Zalghout et al. Construction and Building Materials 351 (2022) 128322

Fig. 5. E* Dynamic Modulus Results: (a) Semi-Log Space and (b) Log-Log Space.

3.2. Description of binder testing methods

10g
|G* | = ( ( )k )mke (13)
During the mixing process, RAP is mixed with virgin aggregates and 1+ ωc

virgin binder at high temperatures. The aged RAP binder affects the
ωR

properties of the total binder in the mix, but the extent of this effect is
dependent on the degree of blending. Thus, the properties of the binder C1 (T − TR )
logaT = (14)
in the final mix cannot be assumed unchanged. In order to determine the C2 + T − TR
properties of these binders, binder extraction and recovery were
Where:
performed.
|G*| = the dynamic shear modulus;
10 g = binder glassy modulus (determined through optimization);
3.2.1. Extraction and recovery
ωc = crossover frequency;
The extraction process was conducted using the centrifuge method,
ωR = reduced frequency;
per AASHTO T 164 standard [36]. The recovery process was conducted
me and k = fitting coefficients;
using a Rotary Evaporator equipment. In this process, the solution of
T = test temperature;
Trichlorethylene (TCE) and asphalt was distilled by immersing the
TR = Reference Temperature; and
rotating flask partially in a heated oil bath while the solution is subjected
C1 and C2 = Time-Temperature Shift Factor Function Fitting
to a partial vacuum and a flow of nitrogen gas to prevent binder
Coefficients.
oxidation. The process was done according to ASTM D5404 standard
[37].
3.2.4. Multiple Stress Creep and Recovery (MSCR) test
The MSCR test was performed on the binder recovered from the three
3.2.2. Performance Grading (PG)
mixtures according to AASHTO T 350 [39]. The test temperature was
The high-temperature PG was determined to evaluate the stiffening
maintained at 64 ◦ C for all the binders, and creep and recovery pa­
effect of RAP on the binder present in the final mix. The PG was deter­
rameters were determined. The two test parameters determined from
mined based on AASHTO M 320 [38], in which the recovered binders
the MSCR tests are the percentage recovery (R), and the nonrecoverable
were considered short-term aged, since it was extracted and recovered
creep compliance (Jnr) at 0.1 kPa and 3.2 kPa. The Jnr parameter is a
from a plant produced mixture. The AASHTO M 320 high-temperature
binder parameter that correlates well with mixture rutting and is the
PG RTFO criteria is presented in Equation (12).
basis of the MSCR test. A lower value of the Jnr at 3.2 kPa was found to
G* correlate to an improvement of the mixture rutting. As for the percent
≥ 2.2 kPa (12)
Sinδ recovery, it represents the recovered strain value at the end of the 9-sec­
Where: ond recovery period.
G* = the dynamic shear modulus; and
Sinδ = Sine of phase angle. 4. Results
A 25 mm diameter plate geometry was used for this test since the
binder testing took place at a high temperature. The test was set to start 4.1. Mixture testing results
at 64 ◦ C for both the control and 15% RAP mixture, since the virgin
binder was PG 70–10. For the 25% RAP mix, the test was set to start at 4.1.1. Dynamic modulus testing results
58 ◦ C since the virgin binder was PG 64–16. The dynamic modulus master curves for each mix are shown in Fig. 5
(a) and (b) in both semi-log and log-log spaces, respectively. As it can be
3.2.3. Time-temperature sweep seen, adding 15% RAP to the mix did not have a stiffening effect when
To determine the stiffening effect that the RAP binder induces on the compared to the control mix. Adding 25% RAP to the mix while drop­
blended binder, the complex shear modulus was performed on the ping one binder grade (PG 64–16) resulted in almost similar stiffness to
extracted and recovered binders. The test was conducted at five different the control and 15% RAP mixtures. Fig. 5(a) shows that at lower tem­
temperatures (10, 20, 30, 40, and 54 ◦ C) and at nine different fre­ peratures, the effect of dropping one binder grade at both high and low
quencies ranging from 30 Hz to 0.1 Hz. Similar to the dynamic modulus performance grading temperatures compensated for the addition of 25%
testing, the time-temperature superposition principle was used to shift RAP. It also shows that the 15% RAP mixture with higher binder grade
the isothermal curves into a final master curve. The Christensen- had a slightly higher stiffness than 25% RAP with lower binder grade.
Anderson-Marasteanu (CAM) model presented in Eq. (13) was used to The statistical analysis performed using ANOVA showed that the dif­
develop the master curves. The William-Landel-Ferry (WLF) equation ference in the dynamic modulus values at all temperatures and fre­
presented in Eq. (14) was used to model the shift factors. quencies is not significant, except at 37.7 ◦ C in which there were some

7
A. Zalghout et al. Construction and Building Materials 351 (2022) 128322

Fig. 6. Flow Number Testing Results: (a) Accumulated Strain Curves and (b) Flow Number.

Fig. 7. IDEAL-CT Analysis for the Three Mixtures: (a) Tensile Strength Values, (b) Fracture Energy Values and (c) CTI Values.

significant differences at the low frequencies. Overall, the addition of different, which indicates that the aged binder in the RAP improved the
RAP at both percentages did not affect the stiffness. rutting resistance.

4.1.2. Flow Number (FN) testing results 4.1.3. IDEAL-CT results

The flow number test results agreed with different studies found in As it can be seen from Fig. 7(a), adding 15% RAP increased the
literature, where there is an expected improvement in the rutting maximum tensile strength of the asphalt mix, and adding more RAP
resistance of the asphalt mix when RAP is incorporated. The change in (25%) resulted in further increase. The statistical analyses performed
the accumulated strain percentage with the increased number of loading using ANOVA and the t-test at the 95% confidence level indicated that
cycles is shown in Fig. 6(a). The graph shows that with the increase in these values are not statistically different, which means RAP did not
RAP content, the number of cycles needed to reach a certain strain affect the tensile strength of the mix. Moreover, the fracture energy
percentage gets higher, which indicates an increase in rutting resistance. calculated from the area under the load-displacement curve shows that
It is also noted that the stiffening effect of the RAP in the 25% RAP mix the addition of RAP resulted in a slight increase in the energy required to
dominated the softening effect of the PG 64–16 binder. To better clarify fracture the material (Fig. 7(b)). However, these values are not different
the results, the average FN value of each mix are compared in Fig. 6(b). statistically, which means adding RAP in both percentages (15% and
The addition of 15% RAP had a stiffening effect on the mix, which led to 25%) did not have significant effect on the fracture energy. On the other
a better rutting resistance, as the 15% RAP mix reached tertiary flow hand, the CTI combines both the energy parameter and post-peak slope
around 1400 cycles. Moreover, the addition of 25% RAP improved the parameter in its calculation. The CTI shows a slight decrease while
rutting performance, in which the 25% RAP mixture had the best rutting adding RAP, but the difference between these values is statistically
resistance with flow cycle around 1863. Statistically, these values are insignificant, meaning that the addition of 15% and 25% (with softer

8
A. Zalghout et al. Construction and Building Materials 351 (2022) 128322

Fig. 8. C* Testing Results.

Fig. 9. Uniaxial Fatigue Testing Results: (a) Damage Characteristic Curves and (b) Fatigue Life Curves.

binder) RAP does not have a significant effect on the cracking potential
of the mixtures (Fig. 7(c)).

4.1.4. C* fracture testing results

The C* versus the crack growth rate (a*) are plotted in Fig. 8. The
higher the slope, the more energy needed to propagate the crack. As it
can be seen from Fig. 8, the 15% RAP mix had a better crack propagation
property than both the control and the 25% RAP. This behavior could be
caused by the higher effective binder content the 15% RAP mix has
(Table 3). The 25% RAP mix had the poorest crack propagation poten­
tial, yet it was similar to the control mixture behavior.

4.1.5. Uniaxial fatigue testing results

The damage characteristic curves of the three mixtures are presented
in Fig. 9(a). As expected, the control mixture sustained more damage
before failure, while the 15% RAP and 25% RAP mixtures sustained less
damage. However, these curves are not enough to tell the full story; thus,
the number of cycles until fatigue failure were determined for each mix
and presented in Fig. 9(b). As it can be seen from the figure, the three
Fig. 10. TSR Results.
curves almost overlap on top of each other. Therefore, for a certain
strain level, the number of cycles required to cause fatigue failure is
almost identical, leading to the fact that using 15% and 25% (with softer 4.1.6. Modified Lottman test results
binder) RAP contents does not have a significant effect on the fatigue life The TSR results are presented in Fig. 10. As can be seen, adding 15%
of the mixtures. and 25% RAP to the mix did not have any significant effect on the
moisture damage resistance, as the TSR values are comparable.

9
A. Zalghout et al. Construction and Building Materials 351 (2022) 128322

control mix just to ensure that the extraction, recovery, and testing
procedures were properly done. The results are presented in Fig. 11. The
resulted PG of the binder extracted from the control mix was expected to
be at high temperature of 70 ◦ C. This is because there was no RAP in the
control mix, so there was no stiffening effect. The test stopped at 71 ◦ C
after the failure criteria was met. The testing on the binder recovered
from the 15% RAP mix showed that the addition of 15% RAP had a slight
stiffening effect, in which the binder failed at 74 ◦ C. The slight increase
in failure temperature reflects the aged binder presence; However, the
high-temperature PG would not change. For the 25% RAP binder, which
was initially with high-temperature PG of 64, the addition of 25% RAP
yielded to an increase from 64 to 76. The test was completed at a tem­
perature of 80 ◦ C, which means the binder was close to be graded as 82.
The reason behind this is the very stiff RAP binder, which was shown to
have a high-temperature PG of 108.

4.2.2. Time-temperature sweep

The master curves for the three binders are shown in Fig. 12(a) and
(b) for different spaces. As the figures illustrate, the addition of 15% RAP
to the mix did not have a significant stiffening effect on the binder, in
which the two master curves corresponding to the control and 15% RAP
Fig. 11. Recovered Binders High-Temperature PG Results.
binders are very close. This is also in agreement with the PG results. For
the 25% RAP binder, the stiffening effect of the highly aged RAP is
Statistically, the three TSR values are not different. Moreover, all the
clearly shown, in which the master curve is above the two other binders
mixtures passed the City of Phoenix specifications of minimum TSR
at all combinations of temperatures and frequencies. This result con­
value of 75%. It is worth noting that hydrated lime was used as anti­
trasts the mixtures’ dynamic modulus results, mainly since thorough
stripping agent in the three mixtures, which generally improves the
blending would have occurred while extracting and recovering the
moisture resistance properties.
binder.

4.2. Binder testing results 4.2.3. Multiple Stress Creep and Recovery (MSCR) testing results
Fig. 13(a) and (b) show the Jnr and recovery percentages for the
4.2.1. Performance Grading (PG) three binders, respectively. As expected, the addition of 15% RAP
The testing was also performed on the binder recovered from the

Fig. 12. Complex Shear Modulus Master Curves of the Three Recovered Binders: (a) Semi-Log Space and (b) Log-Log Space.

Fig. 13. MSCR Results for the Three Recovered Binders: (a) Jnr Values and (b) Recovery Values.

10
A. Zalghout et al. Construction and Building Materials 351 (2022) 128322

Fig. 14. Surface Condition of the Three Sections.

decreased the Jnr values and increased the recovery values, and these In terms of cracking, the two RAP mixtures had similar cracking
changes were due to the increase in the elastic components inside the properties as the control mixture. Similarly, for fatigue and moisture
binder caused by the addition of aged RAP binder. Similarly, the addi­ resistance performance, the mixtures studied were not affected by the
tion of 25% RAP yielded a lower Jnr and higher recovery, due to the addition of RAP. The recovered binder testing showed that only the 25%
availability of more elastic binder. RAP content increased the performance grading of the binder with
higher complex shear modulus values. The MSCR test results showed
5. Field evaluation better recovery and lower non-recoverable strain values. Twenty-six
months after construction, all test sections showed no difference in the
The three test sections showed no distresses (Fig. 14) per the two site pavement distress survey, agreeing with the laboratory test results
visits in July 2020 and February 2021; that is, 19 and 26 months after trends and observations.
construction, respectively. The road is still relatively new, and the RAP Based on this study’s results, the City of Phoenix is already consid­
base layers are also protected with the surface mix (PMAR), but the road ering using 15% RAP in more pavement rehabilitation projects.
has already been subjected to moderate levels of traffic and frequent
heavy rainstorms during two Phoenix monsoon summer seasons and Declaration of Competing Interest
winters. However, future field monitoring over the next few years needs
to be continued to monitor the field performance of the RAP base layers. The authors declare that they have no known competing financial
interests or personal relationships that could have appeared to influence
6. Conclusions the work reported in this paper.

Asphalt concrete recycling can potentially save public works

agencies money and earn them sustainability credit. In this study, Acknowledgments
asphalt mixtures laboratory performance using 15% and 25% RAP in
conventional asphalt mixtures were evaluated. The test results showed The authors would like to acknowledge the help of Brandie Barrett,
that incorporating these two RAP contents in the City of Phoenix asphalt Chris Ewell, Peter Rupal, Robert Duvall, Ryan Stevens, Mark Glock, Rick
mixtures will not affect the properties or pavement performance. Evans, Chris Manno, Anthony Humphrey, Kini Knudson, and Kyle
The addition of 15% RAP into the HMA did not affect the stiffness of Vance. Moreover, the authors would like to acknowledge many con­
the mix while keeping the same binder grade. However, with 25% RAP, tributors from Fisher Industries for their support and collaboration on
one binder grade should be dropped. In this study, the binder grade for the mixture designs, materials supply and asphalt plant production and
15% RAP was PG 70–10 compared to PG 64–16 for 25% RAP. By doing sampling. In particular, Greg Groneberg, Trey Billingsley, and Austin
so, the stiffness of the mixtures did not change significantly, and the Bolze. Thanks are also due to Josh Skinner of M. R. Tanner Construction.
rutting resistance was improved. Acknowledgments also go to Ramadan Salim and Gonzalo Arredondo for
their assistance as part of their association with the ASU research team.

11
A. Zalghout et al. Construction and Building Materials 351 (2022) 128322

Funding [18] Maricopa County Department of Transportation, Uniform Standard Details for
Public Works Construction, 2021.
[19] Maricopa County Department of Transportation, Supplement to the Maricopa
This study was funded by the City of Phoenix under grant number Association of Governments’ Uniform Standard Specifications and Details for
138901-0. The authors greatly acknowledge the support of the City of Public Works Construction, Phoenix: Maricopa County, 2017.
Phoenix. [20] City of Scottsdale, “Supplement to MAG Uniform Standard Specifications for Public
Works Construction,” Scottsdale, 2020.
[21] East Valley Asphalt Committee, “Hot Asphalt Mix Criteria,” East Valley Asphalt
References Committee, November, 2019.
[22] City of Chandler. “Supplement to MAG Uniform Standard Specifications and
[1] Shah, A., R.S., McDaniel, G.A., Huber, & V.L. Gallivan. Investigation of properties Details for Public Works Construction.” Chandler. 2020.
of plant-produced reclaimed asphalt pavement mixtures. Transportation research [23] Town of Gilbert. “Public Works and Engineering Standards.” Gilbert. 2021.
record, 03-, 2007. [24] Regional Transportation Commission of Southern Nevada, Uniform Standard
[2] L. Loria, E.Y. Hajj, P.E. Sebaaly, M. Barton, S. Kass, T. Liske, Performance Specifications, RTCSNV, Las Vegas, Nevada, 2017.
evaluation of asphalt mixtures with high recycled asphalt pavement content, [25] Nevada Department of Transportation, Standard Specifications for Road and
Transportation Research Record 2208 (1) (2011) 72–81. Bridge Construction, NDOT, Carson City, Nevada, 2014.
[3] M.S. Gardiner, C. Wagner, Use of Reclaimed Asphalt Pavement in Superpave Hot- [26] Texas Department of Transportation, “Standard Specifications for Construction and
Mix Asphalt Applications, Transportation Research Record 1681 (1999) 1–9. Maintenance of Highways, Streets, and Bridges,” Texas DOT, November 2014.
[4] A.K. Apeagyei, B.K. Diefenderfer, S.D. Diefenderfer, Rutting resistance of asphalt [27] New Mexico State Department of Transportation, “Standard Specifications for
concrete mixtures that contain recycled asphalt pavement, Transportation Highway and Bridge Construction,” NMDOT, 2019.
Research Record 2208 (1) (2011) 9–16. [28] C.S.T. Agency, Standard Specifications, Department of Transportation, Caltrans,
[5] I. Widyatmoko, Mechanistic-Empirical Mixture Design for Hot Mix Asphalt 2018.
Pavement Recycling, Journal of Construction and Building Materials. 22 (2008) [29] Standard Method of Test for Determining Dynamic Modulus of Hot Mix Asphalt
77–87. HMA. AASHTO, TP 62-07. American Association of State Highway and
[6] Huang, B., B.K., Egan, W.R., Kingery, Z. Zhang, and Z. Gang. Laboratory Study of Transportation Officials, Washington, DC: 2007.
Fatigue Characteristics of HMA Surface Mixtures Containing RAP. CD-ROM, [30] Standard Method of Test for Determining the Dynamic Modulus and Flow Number
Transportation Research Board of National Academies, Washington, D.C., 2004. for Hot Mix Asphalt (HMA) Using the Asphalt Mixture Performance Tester (AMPT),
[7] E.Y. Hajj, P.E. Sebaaly, R. Shrestha, Laboratory evaluation of mixes containing in: AASHTO TP79, American Association of State Highway and Transportation
recycled asphalt pavement (RAP), Road Materials and Pavement Design 10 (3) Officials, Washington, DC, 2009.
(2009) 495–517. [31] F. Zhou, S. Im, L. Sun, T. Scullion, Development of an IDEAL Cracking Test for
[8] Li, X., and N. Gibson. Analysis of RAP with Known Source History and Influence on Asphalt Mix Design and QC/QA, Road Materials and Pavement Design 18 (sup4)
Fatigue Performance. Compendium of Papers, 92nd Annual Meeting of the (2017) 405–427.
Transportation Research Board, Washington, D.C., 2013. [32] J.R. Rice, A path independent integral and the approximate analysis of strain
[9] X. Li, M.O. Marasteanu, R.C. Williams, T.R. Clyne, Effect of Reclaimed Asphalt concentration by notches and cracks, Journal of applied mechanics 35 (2) (1968)
Pavement (Proportion and Type) and Binder Grade on Asphalt Mixtures, 379–386.
Transportation Research Record 2051 (2008) 90–97. [33] Begley, J., and J., Landes. 1972. The J integral as a fracture criterion Fracture
[10] H. Ozer, I.L. Al-Qadi, P. Singhvi, T. Khan, J. Rivera-Perez, A. El-Khatib, Fracture Toughness: Part II: ASTM International.
Characterization of Asphalt Mixture with RAP and RAS Using the Illinois Semi- [34] Determining the Damage Characteristic Curve of Asphalt Concrete from Direct
Circular Bending Test Method and Flexibility Index.“, Transportation Research Tension Cyclic Fatigue Tests. AASHTO TP 107. American Association of State
Record 2575 (2016) 130–137. Highway Transportation Officials (AASHTO). Washington DC: 2014.
[11] X. Shu, B. Huang, D. Vukosavljevic, Laboratory Evaluation of Fatigue [35] Standard Method of Test for Resistance of Compacted Asphalt Mixtures to
Characteristics of Recycled Asphalt Mixture. Journal of, Construction and Building Moisture-Induced Damage. AASHTO T 283. American Association of State
Materials 22 (2008) 1323–1330. Highway and Transportation Officials, Washington, D. C.: 2014.
[12] J. Pan, R.A. Tarefder, M.I. Hossain, A Study of Moisture Impact on Asphalt before [36] Standard Method of Test for Quantitative Extraction of Asphalt Binder from Hot
and after Oxidation Using Molecular Dynamics Simulations, TRB 95th Annual Mix Asphalt (HMA), in: AASHTO T164, American Association of State Highway
Meeting Compendium of Papers, Transportation Research Board of the National and Transportation Officials, Washington, D.C, 2016.
Academies, Washington, D.C., 2016. [37] Standard Practice for Recovery of Asphalt from Solution Using the Rotary
[13] Williams, B.A., A. Copeland, and T.C. Ross. Asphalt Pavement Industry Survey on Evaporator. ASTM D5404. American Society for Testing and Materials, West
Recycled Materials and Warm-Mix Asphalt Usage: 2017. 2018. Conshohocken: ASTM International, 2012.
[14] Gonzalo A., S. Giannakorous, and K. Kaloush. Reclaimed Asphalt Pavement Project [38] Specification for Performance-Graded Asphalt Binder. AASHTO M320. American
Phase I. City of Phoenix, November 2017. Association of State Highway and Transportation Officials, Washington, D.C.:
[15] Street Transportation Department, 2015 City of Phoenix Supplement to the 2015 AASHTO. 2016.
MAG Uniform Standard Specifications and Details for Public Works Construction, [39] Standard Method of Test for Multiple Stress Creep Recovery (MSCR) Test of
Phoenix: City of Phoenix, 2015. Asphalt Binder Using a Dynamic Shear Rheometer (DSR). AASHTO T 350.
[16] Pima Association of Governments, Standard Specifications and Details for Public American Association of State Highway Transportation Officials, Washington, D.C.,
Improvements, Pima Association of Governments, Tucson, 2015. 2014.
[17] Arizona Department of Transportation, Material Group, Policy and Procedure
Directive: Guidance on the use of Reclaimed Asphalt Pavement (RAP) in Asphaltic
Concrete, Phoenix: Arizona Department of Transportation, 2010.

12