Construction and Building Materials 165 (2018) 264–271

Contents lists available at ScienceDirect

Construction and Building Materials

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

Durability assessment of asphalt binder

Ezequiel Angius, Haibo Ding, Simon A.M. Hesp ⇑
Department of Chemistry, Queen’s University, Kingston, Ontario K7L 3N6, Canada

h i g h l i g h t s

 AASHTO M 320 lacks sensitivity and passes many binders of low durability.
 DENT testing separates binders with a high degree of sensitivity in ductile failure.
 Grade losses due to extended PAV aging are correlated with losses in the extended BBR.
 Limiting EBBR grades are highly correlated with limiting phase angle temperatures.

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

Article history: Test results are presented for 24 asphalt binders with the aim to compare various measures of durability.
Received 1 October 2017 Surprisingly high correlations were found between disparate measures of durability but practical consid-
Received in revised form 3 January 2018 erations make certain tests preferable. A comparison is provided between AASHTO M 320 grades, critical
Accepted 5 January 2018
crack tip opening displacements, extended bending beam rheometer grades and grade losses, differences
in limiting temperatures (DTc), and simple limiting phase angle temperatures (T(d)), on aged residues.
The limiting phase angle temperature appears to be a convenient surrogate property that accurately pre-
Keywords:
dicts thermoreversible aging tendencies. Several binders performed poorly, likely because they contained
Double-edge-notched tension test (DENT)
Critical crack tip opening displacement
unwarranted reclaimed asphalt.
(CTOD) Ó 2018 Elsevier Ltd. All rights reserved.
Extended bending beam rheometer test
(EBBR)
Phase angle
Asphalt specification
Thermal and fatigue cracking

1. Introduction ( 34 °C to 36 °C) Superpave grades from six different suppliers.

Shortly after the implementation of the SuperpaveÒ specifica-
tions in the late 1990s, the Ministry of Transportation of Ontario
(MTO) became concerned that roads were aging prematurely at
significant cost to taxpayers. In 2003, MTO commissioned a pave-
ment trial on Highway 655, north of Timmins, to validate the var-
ious test methods that were being proposed to improve the
performance grading of asphalt [1]. The 3.5 km trial was built late
in the season with binders of nearly identical low-temperature
Abbreviations: BBR, bending beam rheometer; CTOD, critical crack tip opening
displacement; DCM, dichloromethylene; DENT, double-edge-notched tension; DSR,
dynamic shear rheometer; FHWA ALF, Federal Highway Administration accelerated
loading facility; LS, laboratory standard; MTO, Ministry of Transportation of
Ontario; PAV, pressure aging vessel; PPA or P31, polyphosphoric acid; PG,
performance grade; RAP, reclaimed asphalt pavement; REOB, recycled engine oil
bottoms; RTFO, rolling thin film oven; SHRP, Strategic Highway Research Program.
⇑ Corresponding author.
E-mail address: simon@chem.queensu.ca (S.A.M. Hesp).
In early 2004, the air temperature reached a record 48 °C, and
the pavement surface reached the low design of 34 °C twice. (In
2005, it dropped below 30 °C on six additional days.) In spring
2005, the pavement had cracked to moderate degrees in four sec-
tions. As of 2011, section 1 remained largely free of distress with
only few cracks (mostly through core holes) because it was made
with a durable binder from Lloydminster [1]. The worst performer,
section 4, was riddled with cracks and many potholes. Fig. 1 pro-
vides representative crack maps and photographs for both sec-
tions. Performance as seen in section 1 is rare today while the
state represented by section 4 is rather common. The other four
sections showed performance somewhere in between these two
extremes. In 2014, as measured by MTO’s Automated Road Ana-
lyzer, section 1 had remained largely free of thermal cracking dis-
tress, five sections had about 1000 m of cracks each, while the
worst performer, section 4, had 2160 m of cracks, many potholes,
and will soon be in need of significant end-of-life care (MTO,

https://doi.org/10.1016/j.conbuildmat.2018.01.037
0950-0618/Ó 2018 Elsevier Ltd. All rights reserved.
 E. Angius et al. / Construction and Building Materials 165 (2018) 264–271
Fig. 1. Representative 2011 crack maps and photographs from 2003 Highway 655 pavement trial: (a) section 1 and (b) section 4. Note: Both crack maps are for randomly
selected 50 m stretches of pavement. Photographs were not necessarily taken in areas where cracks were mapped.
Unpublished Data, 2014). Hence, this trial has shown rather starkly
that the current AASHTO M 320 specification lacks accuracy and
needs to be improved to provide a more correct measure of
durability.
The term durability is defined by the Oxford Online dictionary
as follows: ‘‘The quality of being durable. Capability of withstand-
ing decay or wear.” [2]. In the context of asphalt science and tech-
nology this means the ability to retain certain design properties
that allow a flexible pavement to remain without significant dis-
tress for at least the design life and preferably longer. It is well doc-
umented in the literature that pavements of ostensibly the same
design, subgrade, climate and traffic can show a wide range of
lifespans. This is often due to differences in durability of material
properties that are overlooked by current specification protocols
[1]. An ideal durability test for asphalt binder can be defined as
one that measures properties showing a high degree of correlation
with performance in old age. It might be added that for practical
reasons, the test should be easy to conduct on a small amount of
material and be highly reproducible. An accurate test with a high
degree of reproducibility lessens the risk for both users and pro-
ducers of asphalt binder.
Certain asphalts high in naphthene aromatics, and low in paraf-
fin and asphaltenes, can accommodate a significant amount of
additional asphaltenes from oxidation and harden slightly over
time [3]. California Valley crude produces such binder since it is
largely naphthenic and its residue is low in asphaltenes. Binders
produced from aromatic crudes can be oxidized for long times
without much hardening. Many Western Canadian naphthenic
crudes can produce material that is highly resistant to aging [4].
In contrast, inferior paraffinic crudes typically produce materials
that are sensitive to both thermoreversible (wax crystallization,
asphaltenes and resins precipitation, phase separation, collectively
known as paraffinic demixing) and irreversible (oxidative)
hardening [5–7]. Western Texas Intermediate and certain Argen-
tinian crudes are examples that produce hardening-susceptible
binders [8].
Air blown binders are another group of materials that lack dura-
bility. While they provide impressive grades in the laboratory, their
performance in service has consistently been disappointing [9–11].
265
The air blown (oxidized) residues lack colloidal stability since addi-
tional asphaltenes introduced through the air blowing process are
poorly accommodated and form gel-type structures that hinder
stress relaxation during winter and spring. However, these binders
do well in the grading process since the structure that forms the gel
network takes time to consolidate when cold, and is therefore not
captured in most current specification tests that prescribe only
minimal conditioning.
Finally, the recent ‘‘greening” of the asphalt industry through
the increased use of reclaimed asphalt pavement (RAP), recycled
engine oil bottoms (REOB), asphalt shingle waste, and other
unknowns has introduced performance challenges. Added to the
problem is the increased use of wax-based warm asphalt technolo-
gies that facilitate compaction. Early pavement failures have
become quite common and the causes for these can nearly always
be traced back to deficient materials and processes. The incorrect
and irresponsible use of certain recycled materials is costing user
agencies more for less performance than in years past [12].
Our long-term vision is that paved roads last the lifespan
expected from quality Western Canadian asphalt instead of the sig-
nificantly reduced lifespans currently obtained in Ontario and else-
where [11–13]. Longer-lasting roads with less need for repair cause
less congestion, delay and danger for users. To achieve this, we will
develop more comprehensive, theoretically-sophisticated tests to
assess durability in asphalt. Our test methods will be available to
producers and user agencies in Ontario and beyond. Our short-
term objective is to develop a practical test protocol that can be
done in a reproducible manner, in a short amount of time, with lit-
tle material. Placing a minimum durability requirement in asphalt
binder specifications will prevent premature failures in terms of
fatigue, low temperature cracking and moisture damage.
2. Background
The two types of aging that are most relevant for long-term
pavement performance are oxidative, or non-reversible aging,
and thermoreversible aging. In the literature there are a great
number of studies related to oxidative aging. This is likely because
266 E. Angius et al. / Construction and Building Materials 165 (2018) 264–271
most asphalt is used in temperate and hot climates. A complete
review of the literature on oxidative aging is beyond the scope of
this paper. Instead, this paper will limit discussion to the way in
which the Superpave specification protocol was meant to assess
aging and how the AASHTO M 320 standard does it today, using
the rolling thin film oven (RTFO), the pressure aging vessel (PAV),
the dynamic shear rheometer (DSR), and bending beam rheometer
(BBR). It has been recognized in several studies that those binders
that suffer from oxidative aging are typically also sensitive to ther-
moreversible effects so the discussions that follow are applicable
to both [4,11].
2.1. Adoption of RTFO, PAV and BBR in AASHTO M 320
The researchers that worked under the Strategic Highway
Research Program (SHRP) to develop the Superpave asphalt binder
specification included one direct measure of durability in terms of a
maximum limit on the weight loss during RTFO aging. This had
been the norm and no significant changes were made. It is currently
set at one percent in order to avoid binders that potentially pose
environmental, health and safety hazards during construction.
A considerable amount of research focused on the refinement of
the PAV [14,15]. The SHRP researchers looked at the effect of film
thickness, aging temperature, time and pressure on the perfor-
mance properties as measured with the penetration test and DSR
[14,15]. The general finding confirmed earlier research that had
shown all these variables to be of importance, especially in the case
of lesser quality binders [15]. However, for practical reasons, the
PAV was standardized using only a single film thickness of 3.2
mm (50 g/pan), temperature of 100 °C, time of 20 h, and pressure
of 2.08 MPa, and hence lacks a clear assessment of durability [4].
The original BBR specification contained provisions for creep
testing of the PAV residue after one and 24 h of cold conditioning
but the latter never appears to have caught on for reasons that
are not well documented. Supposedly, the industry lobbied against
the use of BBR grading data after 24 h of conditioning at the test
temperature, claiming it to be impractical [16]. However, a rather
significant amount of work was done during the SHRP on the
effects that extended cold conditioning has on low temperature
BBR properties. It generally confirmed earlier findings that differ-
ent crude sources produce binders that show a wide range of ther-
moreversible aging tendencies (age hardening, physical hardening,
physical aging, reversible aging, etc.) [17]. So, the BBR specification
that ended up in AASHTO M 320 [18], as used in most jurisdictions
today, also lacks a comprehensive assessment of durability.
2.2. Modified pressure aging vessel (MTO LS-228)
Our group took the original PAV research one step further and
used materials sampled during the construction of a 2003 Ministry
of Transportation of Ontario (MTO) pavement trial to show how
important the actual PAV conditions can be in helping to explain
relative performance differences in the real world [4]. Aging in
thinner films or for longer times was shown to significantly
enhance the performance grading. In spite of this, the modified
PAV protocols published by the MTO as Laboratory Standard 228
[19] have yet to find traction in Ontario pavement design. A high
degree of inertia makes it difficult to implement changes, no mat-
ter how much they will save the user agencies in the long term.
2.3. Double-edge-notched tension test (MTO LS-299, AASHTO TP 113-15)
The double-edge-notched tension (DENT) test was developed to
determine an approximate critical crack tip opening displacement
(CTOD), which reflects a binder’s critical strain tolerance in the
ductile state yet close to the ductile-to-brittle transition zone
[9,10,12,13,20]. The DENT test is modelled after the original ductil-
ity test developed by Dow [21]. He observed with a keen eye that
for binders that flow well, performance is generally satisfactory,
while for those that fail abruptly the experience is rarely accept-
able. The method was published by the MTO as Laboratory Stan-
dard 299 [20], and it has now been implemented on all Ontario
asphalt paving contracts ($300 million annual asphalt cost). The
DENT was adopted by AASHTO as provisional standard TP 113-15
in 2015 [22], and is currently being evaluated by several research
groups around the world.
2.4. Extended bending beam rheometer (MTO LS-308, AASHTO TP 122-16)
Along the same lines, a considerable amount of research was
done to show how variations in thermoreversible aging can help
to explain relative performance differences [1,9–13]. Materials
from the 2003 MTO pavement trial near Timmins [1,9,11], along
with materials from other trials and regular contracts [10,12,13],
were able to show that extended conditioning separates the satis-
factory from the unsatisfactory performing materials with a high
degree of success. The extended bending beam rheometer (EBBR)
protocol conditions samples for one, 24 and 72 h at 10 and 20
degree Celsius above the lower pavement design temperature, Td
+ 10 and Td + 20, which is where most thermoreversible aging
occurs. The 24 h is often too short a time to obtain much significant
hardening except for the very worst materials (round robin testing
has shown the reproducibility to be around ± 1 °C (MTO, Perfor-
mance Graded AC Correlation Results, April 2015 [23]). Hence,
72 h was selected as a compromise between getting sufficient
levels of hardening and keeping the test method practical. As such,
the EBBR provides a discriminating assessment of low temperature
durability. The method was published by the MTO as Laboratory
Standard 308 [24], and it has now been implemented on all Ontario
asphalt paving contracts. The EBBR was adopted by AASHTO as
provisional standard TP 122-16 in 2016 [25].
2.5. Difference in limiting BBR temperatures (DTc)
One drawback of the EBBR protocol is that it takes three days to
obtain a grade and a grade loss. So, our group and other researchers
have continued efforts to find alternative, less time consuming
ways to measure durability (e.g., [26–29], others). The difference
between limiting stiffness (S) and m-value temperatures, DTc,
being a measure of the degree of gelation in the binder, has been
investigated [26,27,29]. Binders that start as a gel-type will suffer
most from the effects of additional conditioning in both the PAV
and BBR. However, the issue is not straightforward in that there
are materials that actually improve upon additional PAV aging
[30]. There are now suggestions to condition binders for longer
in the PAV and then measure the difference in the limiting BBR
temperatures, T(S = 300 MPa) and T(m = 0.3), after one hour of con-
ditioning. The DTc obtained can be limited as another approach to
prevent the use of binders that lack durability.
The objective of this study is to compare most of the current
tests for durability using a set of 24 binders obtained during and
after the construction of a number of pavements.
3. Experimental details
3.1. Materials
Binders investigated for this study were obtained by extraction
and recovery from hot-mix asphalt, sampled 1 and 2 years after
 E. Angius et al. / Construction and Building Materials 165 (2018) 264–271
construction (core samples from two pavements) and during the
construction of a number of new pavements (loose mix for the
remaining samples). All binders had been specified as PG 64–28
in the contract documents, several according to regular AASHTO
M 320 specifications but most according to DENT and EBBR crite-
ria. The asphalt binders were obtained from three different com-
mercial suppliers and the paving work was performed by five
different contractors. Extracted and recovered binders were tested
in order to deal with the use of RAP and other unknowns at the hot
mix plant.
3.2. Binder extraction, recovery and aging
The mixes were left to soak overnight in reagent grade dichlor-
omethylene (DCM) or toluene and rinsed several times to extract
nearly all the binder. The 4–6 L of asphalt solution was left to
sediment overnight in order to remove the fines. The bulk of the
solvent was evaporated at moderate temperatures and vacuum
(50–80 °C and 500 mbar) under a dry nitrogen gas atmosphere in
a rotary evaporator. Once no more solvent was visibly being dis-
tilled, the temperature was increased in steps of 20 °C to a final
160 °C. After the temperature was stabilized, the flask was left to
rotate at 10 rpm and a vacuum below 100 mbar for an additional
one hour.
The binders recovered from core samples were tested without
further aging while the materials from loose mix were PAV-aged
according to standard procedures (AASHTO R 28-09 [31]) prior to
performance grading.
3.3. Rheological and failure testing
Recovered binders were tested in the DSR for high and interme-
diate temperature performance properties according to standard
procedures (AASHTO R 29-08 [32]). Both complex modulus, G⁄
(Pa), and phase angle, d (°), were measured at 10 rad/s and various
temperatures.
Aged binders were poured in silicone molds with aluminum
inserts and tested in a ductilometer after 0.5–2 h of condition-
ing at 15 °C prior to DENT testing. The displacement rate was
kept constant at 50 mm/min. Specimens were tested with 5,
10 and 15 mm ligaments at a thickness of 10 mm in 30 mm
width. The essential work of failure, we (J.m 2), was calculated
as the specific total work (i.e., area under the force-
displacement curve) at zero ligament. The we was determined
by extrapolation to zero from ligament lengths of 5, 10 and
15 mm. An approximate critical crack tip opening displacement
was calculated as the essential work divided by the average net
section stress in the smallest ligament specimen, CTOD = we/
rnet section. The CTOD is used for specification grading of asphalt
binders in the ductile state and provides a measure of durabil-
ity as it has been shown to correlate with field performance
[12,13].
The PAV residues were finally poured into slender beams and
tested according to Ontario’s EBBR method (LS-308 [24]). A total
of 12 beams were placed in ethanol baths kept at 8°C and 18 °C
(six beams in each bath). Three samples each were tested at pass
and fail temperatures after one, 24 and 72 h. The warmest of the
limiting temperatures where the creep stiffness, S(60 s), reached
300 MPa or the m(60 s)-value reached 0.3 set the EBBR grade.
The EBBR grade was then subtracted from the one hour test result
in order to obtain a 72 h grade loss, which is a measure of
durability that has shown high correlation with old age pavement
performance [1,10,11].
267
4. Results and discussion
4.1. Standard AASHTO M 320 assessment of durability
The findings for the standard AASHTO M 320 [18] high (XX),
intermediate (II) and low temperature (YY) grades, LS-299 CTOD
[20], and LS-308 EBBR [24] grade and grade loss durability assess-
ments are in Table 1.
The two core samples C1 and C2 were extracted with both DCM
and toluene, in order to compare the effect of the solvent-
extraction method on performance-based properties. It is evident
that the solvent does not have any significant effect on the find-
ings. The average grade temperature difference of 0.9 °C, and an
average CTOD difference of 1.55 mm between the samples are well
within the reproducibility levels for both test methods and agree
with results obtained for a significantly larger set of materials [33].
From the AASHTO M 320 XX and YY grading results it can be
concluded that all binders are expected to perform well since they
achieved the required 64 °C high temperature and 28 °C low tem-
perature grades set out in the contract documents.
The intermediate environmental temperature for the grade
specified in the contract documents is (64 + 28)/2–28 + 4 or
22 °C, which is above all the intermediate grades in Table 1. For
the climatic zone it is (58 + 28)/2–28 + 4 or 19 °C, which is above
all but three of the grades in Table 1. For the three that failed, they
do so by only marginal amounts (0.6 °C to 2.0 °C). Hence, the
intermediate grading criteria also suggest these binders should
all perform well.
4.2. Double-edge-notched tension (DENT) analysis
In contrast with the AASHTO M 320 findings, the LS-299 DENT
protocol, which was adopted on selected contracts from which the
samples originated, failed a considerable number of samples. The
specification required a minimum CTOD of 14 mm to pass in a
28 °C zone. Out of the sample set tested, Table 1 shows that a
total of nine failed the minimum criterion. Three of the nine fail-
ures were for samples extracted and recovered from loose mix
used on a contract for which both LS-299 CTOD and LS-308 EBBR
acceptance criteria had been specified and thus should not have
failed. However, acceptance was based on binder supply tank sam-
ples, rather than material extracted and recovered from the hot-
mix asphalt, so the contractor was not penalized. The remaining
six failures came from contracts which had asphalt cement speci-
fied according to conventional AASHTO M 320 criteria. These mixes
were likely contaminated with RAP where it was not sanctioned.
The acceptance is now based on LS-299 CTOD and LS-308 EBBR
properties of extracted and recovered asphalt cement.
4.3. Extended bending beam rheometer analysis
Table 1 shows that the LS-308 EBBR standard also managed to
pass only the 15/24 (62.5%) binders that did well in the LS-299
CTOD test. There appears to be a moderate correlation between
EBBR and DENT results, and both should therefore be good indica-
tors of durability. The extended conditioning in the LS-308 EBBR
method penalizes certain binders more than others, which is
reflected in its greater sensitivity compared to AASHTO M 320.
The grade range covered over 11.8 °C for the EBBR compared to
9.3 °C for the BBR, a difference of nearly 27%.
The LS-308 EBBR grade loss was positive for all binders, imply-
ing the extended BBR protocol produced limiting grades warmer
than those found using the AASHTO M 320 low temperature
standard by an average of 3.8 °C.
268 E. Angius et al. / Construction and Building Materials 165 (2018) 264–271

Table 1
Conventional Durability Measures.

Sample Code AASHTO M 320, °C LS-299, mm LS-308, °C

XX II YY CTOD Grade Grade Loss
C1T 77 17 29 10.3 25 3.7
C1D 78 17 28 12.4 25 3.1
C2T 79 17 31 8.9 26 4.8
C2D 80 19 30 7.9 24 6.2
L3T 69 14 33 14.0 31 2.7
L4T 77 14 34 18.0 31 2.8
L5T 74 16 33 15.1 28 5.0
L6T 67 10 36 36.5 34 2.0
L7T 71 20 28 8.8 22 5.2
L8T 71 21 28 9.7 24 3.6
L9T 66 12 35 15.9 32 3.0
L10D 70 7 37 22.3 34 2.7
L11D 68 13 32 15.1 29 2.7
L12D 69 13 33 15.7 29 3.9
L13D 72 11 35 18.3 32 3.0
L14D 72 13 33 16.7 30 3.1
L15D 77 21 29 8.7 23 6.2
L16D 75 19 30 9.7 25 5.5
L17D 72 19 31 12.1 26 5.3
L18D 72 15 34 16.1 31 2.6
L19D 71 13 33 21.4 30 3.4
L20D 67 12 34 23.4 32 1.8
L21D 71 13 33 22.6 30 3.5
L22D 69 15 33 18.3 28 4.1
Averages: 72 15 32 15.8 28 3.8
Ranges: 66 to 80 7 to 21 28 to 37 7.9 to 36.5 22 to 34 1.8 to 6.2
Pass Criteria: >64 >19 < 28 >14 < 28 <6

Note: C = core, L = loose mix, D = dichloromethylene (DCM), and T = toluene. XX = AASHTO M 320 high grade temperature based on G*/sin d, II = AASHTO M 320 intermediate
grade temperature based on G*sin d, and YY = AASHTO M 320 low grade temperature based on S(60 s) and m(60 s) (18). Grades are rounded to the nearest degree.

4.4. Modified pressure aging vessel analysis
All the binders were treated for an additional 20 h in the PAV for
a total of 40 h, according to the LS-228 Modified PAV protocol [19],
to measure changes in rheological properties. Binders aged in this
way were once more tested according to both standard DSR and
BBR protocols at pass and fail temperatures for all specification
properties [18]. The results were compared with conventional
properties.
A comparison for the LS-308 EBBR grade after 20 h of PAV aging
and the limiting one hour BBR grade after 40 h of aging is provided
in Fig. 2(a). The graph shows that there is a rather strong correla-
tion between the limiting grade temperatures (R2 = 0.91), in agree-
ment with earlier findings [4]. Additionally, the average difference
in grade temperature between the two methods was found to be
only 0.90 °C. As the LS-308 EBBR standard has been proven to accu-
rately predict binder performance in the real world [1,5,10–13], it
can be concluded that extending the pressure aging treatment to
40 h will likely produce limiting grades which are less prone to
error when predicting asphalt durability. This approach would
accomplish the grading in less time. However, caution is warranted
in that remarkable binders exist that appear to improve their BBR
performance upon additional PAV aging [30]. Hence, it appears the
good correlation seen for this set of materials might not reflect all
commercial binders that are currently available.
The grade losses for AASHTO M 320 at 40 h versus 20 h of PAV
aging are compared to the standard LS-308 grade losses from
Table 1 in Fig. 2(b). There again appears to be a general correlation
and the average difference between both grade losses was found to
be only 0.90 °C.
It should be noted that the grade loss correlation appears
weaker as suggested by a lower R2 of 0.49. This can be attributed
to the small range of values seen in both the LS-308 EBBR grade
loss (1–72 h conditioning) and the AASHTO M 320 (40–20 h PAV)
grade loss. Unlike the high correlation found in Fig. 2(a), the ranges
in Fig. 2(b) do not provide enough sensitivity for any meaningful
conclusions to be drawn.
4.5. Differences in limiting temperatures (DTc)
The difference in limiting BBR temperatures, TS = 300 MPa–Tm = 0.3,
commonly referred to as DTc, has gained a significant amount of
attention as a potential specification property to assure a minimal
level of durability (e.g., [4,27,29]). Binders that deviate from the
line of equality in a plot of TS versus Tm are more prone to form
gel-type structures with reduced relaxation ability, and have
shown to provide inferior performance [1,4,5,9–13,26,27]). The
correlations between DTc values for each BBR/PAV protocol and
the LS-308 EBBR grade loss are in Fig. 3(a)–(c). It is obvious that
there is little to no correlation after one hour in the BBR, but some-
what of a trend after 72 h of cold conditioning prior to pass/fail
testing. This makes sense since most of the binders included for
this study start off close to a DTc of zero (Fig. 3(a)) and after 72 h
some become m-controlled by moderate amounts (Fig. 3(c)).
Hence, DTc correlates with the grade loss albeit not strongly. Once
again, the sample size used is limited in binder performance vari-
ation, therefore future DTc studies would benefit from a sample
pool containing a greater range of materials with more varied
durability.
4.6. Limiting phase angle and loss modulus analysis
The final comparison involves the AASHTO M 320 YY, EBBR
grades, limiting phase angle temperatures and intermediate M
320 II grades. The latter is determined by the temperature at which
the binder loss modulus, G⁄sin d, reaches 5.0 MPa. The relevant
graphs are provided in Fig. 4.
The findings are most interesting in several respects. First, the
comparison between Fig. 4(a) and (b) shows how the EBBR test is
more sensitive than the regular BBR test. The grades are spread
 E. Angius et al. / Construction and Building Materials 165 (2018) 264–271 269

-20 4

ΔTc (20 h PAV, 1 h BBR), oC

y = 1.00x 2
M 320 (40 h PAV), °C

R² = 0.91
-25
0

-2
-30

-4 y = -0.33x + 1.40
R² = 0.14
-35 -6
-35 -30 -25 -20 (a) 0 3 6 9
(a) LS-308 EBBR (20 h PAV), °C LS-308 Loss, °C

9 4

y = 0.93x
y = -0.68x + 0.06

ΔTc (40 h PAV, 1 h BBR), oC

2
R² = 0.29
R² = 0.49
M 320 Loss (PAV), °C

0
6

-2

-4
3

-6

-8
0 (b) 0 3 6 9
0 3 6 9
LS-308 Loss, °C
(b) LS-308 EBBR Loss, °C
4
Fig. 2. (a) Correlation between AASHTO M 320 grade (40 h PAV and 1 h condition-
ing at the test temperature prior to BBR pass/fail testing) and LS-308 grade (20 h y = -0.87x + 1.37
2
ΔTc (20 h PAV, 72 h BBR), oC

PAV and 72 h conditioning prior to BBR pass/fail testing), and (b) Correlation
between PAV grade loss (BBR 40 h PAV with 1 h conditioning minus BBR 20 h PAV
R² = 0.51
with 1 h conditioning) and the EBBR grade loss (BBR 20 h PAV with 1 h conditioning
minus BBR 20 h PAV with 72 h conditioning). 0

out over a range of 11.8 °C for the EBBR tests versus 9.3 °C for the
regular BBR test (27% better for the EBBR). Second, more revealing
is that the limiting 30° phase angle temperature has even higher
sensitivity covering a range of 13.3 °C (43% better than the BBR).
This goes to show how the non-equilibrium state in which the bin-
der is tested in the BBR and EBBR creates performance issues due
to the under-design of the pavement for thermal cracking. A higher
degree of sensitivity with equal or better reproducibility lowers
risk and thus benefits both users and producers of asphalt cement.
Third, there appear to be very high correlations between the EBBR
grade, measured after 72 h of cold conditioning, and the limiting
phase angle temperatures, measured after 15 min of conditioning
at more moderate temperatures (R2 of 0.94–0.96). This suggests
that it is possible to get nearly the same information from a simple
phase angle measurement as what is obtained from a rather differ-
ent, and some would say cumbersome, extended BBR protocol.
Phase angles are determined in a DSR on very small amounts of
material after minimal conditioning, which would provide for a
much improved and user-friendly test method. Finally, the
AASHTO M 320 II grade appears to be less sensitive than the limit-
ing phase angle temperature. It suffers from reproducibility issues
related to the complex modulus measurement changing during
equilibration [23,37], and it appears to pass nearly all binders sold
in the Ontario market, making the acceptance criterion somewhat
questionable.
-2
-4
-6
-8
(c) 0 3 6 9
LS-308 Loss, °C
Fig. 3. Correlation between LS-308 grade loss (BBR 20 h PAV with 1 h conditioning
minus BBR 20 h PAV with 72 h conditioning) and DTc property for regular and
extended conditioning protocols.
Our group and others have previously considered the interme-
diate and low temperature phase angles for the performance rank-
ing of asphalt binders [11,26,34–36]. Extracted and recovered
materials were tested from a number of Eastern and Northeastern
Ontario trial sections and contracts and it was found that the phase
angle (tan d) provided a very accurate measure of pavement per-
formance after only minimal conditioning [11,26]. Binders with
phase angles >28° at 10 °C showed little cracking while those
below 28° had mostly cracked prematurely and severely. This find-
ing agreed well with results previously and independently
270 E. Angius et al. / Construction and Building Materials 165 (2018) 264–271

-10 64-28) produced by three suppliers and used by five contactors.

A wider assortment of binders with different modifiers would need
R² = 0.85
to be investigated before the EBBR protocol can be replaced with a
simple DSR test. Further, the DENT test will likely continue to be
M 320 YY Grade, oC

R² = 0.87 needed in order to provide a second safeguard against premature

-20
cracking distress. It has been reported by Glover et al. [38] that
there is not always a clear relationship between low strain rheo-
R² = 0.86
logical properties and ductility. Hence, for the time being it is pru-
dent to use both rheological and failure tests to control cracking
-30 distress.

5. Summary and conclusions

-40
Given the results presented, the following summary and con-
-15 -5 5 15
(a) clusions are provided:
T(δ = 25o, 30o and 35o), oC
1) The regular AASHTO M 320 specification appears to lack sen-
sitivity and passes a significant number of asphalt binders
-10
deficient in durability.
R² = 0.95 2) The approximate critical crack tip opening displacement
(CTOD) as measured with the double-edge-notched tension
R² = 0.96
(DENT) test provides a significant degree of differentiation
LS-308 Grade, oC

-20 for binders in the high strain regime.

R² = 0.94 3) For binders investigated in this study, the extended pressure
aging vessel (PAV) conditioning for 40 h provides low tem-
perature grades that are very similar to those obtained by
-30 extended bending beam rheometer (EBBR) testing of the
20 h PAV residues.
4) The difference in limiting temperatures from the BBR mea-
surement (DTc parameter), as measured on 20 h PAV resi-
due, shows little to no correlation with the grade loss in
-40
the EBBR. It does show a modest correlation for 40 h PAV
-15 -5 5 15
residues.
(b) T(δ = 25o, 30o and 35o), oC 5) There appears to be a very high correlation between the
EBBR grade and limiting temperatures where the phase
25 angle reaches 25°, 30° and 35°. Hence, these parameters
are worthy of further investigation for future thermal and
y = 0.83x + 16 fatigue cracking specifications for asphalt binders.
R² = 0.87 6) For the time being it remains prudent to use both rheological
20
M 320 II Grade, oC

and failure tests to control thermal and fatigue cracking dis-

tress in old age.

15 Given the scale of thermal cracking problems in North American

pavements, the issue of which durability properties to use for the
control of this distress deserves continued attention.
10
Acknowledgements and Disclaimer

5 The authors thank the Natural Sciences and Engineering

-15 -10 -5 0 5
(c)
T(δ = 30), oC
Fig. 4. Correlation between limiting phase angle temperatures (T(d), 15 min
conditioning) and (a) AASHTO M 320 YY (1 h conditioning), (b) LS-308 EBBR (72 h
conditioning), and (c) AASHTO M 320 II grades (15 min conditioning).
reported by Migliori et al. [35] who found a limit of 27° to be opti-
mal (at a slightly different oscillatory frequency).
Our findings suggest that the phase angle is worthy of further
investigation. It may well be the preferred property for the control
of pavement performance in old age. However, caution is war-
ranted to make sure that a switch would result in improved pave-
ment performance rather than a continued struggle. The set of
materials used for this study was limited to a single grade (PG
10 15
Research Council of Canada and the user agency of the asphalt
for financial support. Haibo Ding thanks the China Scholarship
Council for the financial support that was provided to study at
Queen’s University. The authors also thank user agency staff for
supplying the materials evaluated in this study.
Neither of the sponsors necessarily agree with, endorse or will
adopt the findings, conclusions, or recommendations either
inferred or expressly stated in subject data developed.
References
[1] A. Rigg, A. Duff, Y. Nie, M. Somuah, N. Tetteh, S.A.M. Hesp, Non-isothermal
kinetic analysis of reversible aging in asphalt cement, Road Mater. Pavement
Des. 18 (4) (2017) 185–210.
[2] Oxford Online Dictionary, www.oxforddictionaries.com, accessed on July 28,
2017.
 E. Angius et al. / Construction and Building Materials 165 (2018) 264–271
[3] J.C. Petersen, A review of the fundamentals of asphalt oxidation,
Transportation Research Circular E-C140, Transportation Research Board,
Washington, D.C., 2009.
[4] J. Erskine, S.A.M. Hesp, F. Kaveh, Another look at accelerated aging of asphalt
cements in the pressure aging vessel, June 13-15, Proc., Fifth Eurasphalt &
Eurobitume Congress, Istanbul, 2012.
[5] S.A.M. Hesp, S. Iliuta, J. Shirokoff, Reversible aging in asphalt, Energy Fuels 21
(2007) 1112–1121.
[6] G. Van Gooswilligen, F.T. DeBats, T. Harrison, Quality of paving grade bitumen–
A practical approach in terms of functional tests, in: Proc., Fourth Eurobitume
Congress, 1989, pp. 290–297.
[7] A. Schmets, N. Kringos, T. Pauli, P. Redelius, T. Scarpas, On the existence of wax-
induced phase separation in bitumen, Int. J. Pavement Eng. 11 (6) (2010) 555–
563.
[8] I. Kodrat, D. Sohn, S.A.M. Hesp, Comparison of polyphosphoric acid modified
asphalt binders with straight and polymer-modified materials, Transport. Res.
Rec. 2007 (1998) 47–55.
[9] S.A.M. Hesp, S.N. Genin, D. Scafe, H.F. Shurvell, S. Subramani, Five year
performance review of a northern ontario pavement trial, in: Proc., Canadian
Technical Asphalt Association, 2009, pp. 99–126.
[10] M.O. Zhao, S.A.M. Hesp, Performance grading of the Lamont, Alberta C-SHRP
pavement trial binders, Int. J. Pavement Eng. 7 (3) (2006) 199–211.
[11] J.L. Freeston, G. Gillespie, S.A.M. Hesp, M. Paliukaite, R. Taylor, Physical
hardening in asphalt, in: Proc., Canadian Technical Asphalt Association, 2015,
pp. 53–81.
[12] S.A.M. Hesp, H.F. Shurvell, Waste engine oil residue in asphalt cement August
28, Proc., Seventh International Conference on Maintenance and Rehabilitation
of Pavements and Technological Control, Mairepav7, Auckland, New Zealand,
2012.
[13] S.A.M. Hesp, A. Soleimani, S. Subramani, T. Phillips, D. Smith, P. Marks, K.K.
Tam, Asphalt pavement cracking: analysis of extraordinary life cycle
variability in Eastern and Northeastern Ontario, Int. J. Pavement Eng. 10 (3)
(2009) 209–227.
[14] J.W. Button, M. Jawle, Evaluation and development of a pressure aging vessel
for asphalt cement, Transp. Res. Rec. 1391 (1993) 11–19.
[15] H. Bahia, D.A. Anderson, The Pressure Aging Vessel (PAV): a test to simulate
rheological changes due to field aging physical properties of asphalt cement,
in: J.C. Hardin (Ed.), ASTM STP 1241, ASTM, Philadelphia, 1995.
[16] D.A. Anderson, Discussion following the paper by Rigg et al., J. Assoc. Asphalt
Paving Technol. 86 (2017).
[17] H.U. Bahia, D.A. Anderson, Glass transition behavior and physical hardening of
asphalt binders, J. Assoc. Asphalt Paving Technol. 62 (1993) 93–129.
[18] AASHTO (American Association of State Highway and Transportation Officials).
M 320-09, Standard Specification for Performance-Graded Asphalt Binder,
AASHTO, Washington, D.C., 2009.
[19] Ministry of Transportation of Ontario. LS-228 Modified Pressure Aging Vessel
Protocol. Revision 29 to MTO Laboratory Testing Manual, 2015.
[20] Ministry of Transportation of Ontario. LS-299 Method of Test for Asphalt
Cement’s Resistance to Fatigue Fracture Using Double-Edge-Notched Tension
Test (DENT). Revision 23 to MTO Laboratory Testing Manual, 2007.
[21] A.W. Dow, Discussion on ductility test, Proc. Assoc. Asphalt Paving Technol. 5
(1936) 139–141.
[22] AASHTO (American Association of State Highway and Transportation Officials).
Determination of Asphalt Binder Resistance to Ductile Failure Using the
271
Double-Edge-Notched Tension (DENT) Test. TP 113-15, AASHTO, Washington,
D.C., 2015.
[23] Bituminous Section, Materials Engineering and Research Office, Ministry of
Transportation of Ontario. Performance Graded AC Correlation Results for
April 2015 – Round 1, Downsview, Ontario, July 15, 2015.
[24] Ministry of Transportation of Ontario. LS-308 Determination of Performance
Grade of Physically Aged Asphalt Using EBBR Method. Revision 29 to the MTO
Laboratory Testing Manual, 2015.
[25] AASHTO (American Association of State Highway and Transportation Officials).
Determination of Performance Grade of Physically Aged Asphalt Binder Using
the Extended Bending Beam Rheometer (BBR) Method. TP 122-16, AASHTO,
Washington, D.C., 2016.
[26] A. Soleimani, S. Walsh, S.A.M. Hesp, Asphalt cement loss tangent as surrogate
performance indicator for control of thermal cracking, Transp. Res. Record
2162 (2009) 39–46.
[27] S.A.M. Hesp, S. Subramani, Another look at the bending beam rheometer for
specification grading of asphalt cements July 8-10, Proc., Seventh International
Conference on Maintenance and Rehabilitation of Pavements and
Technological Control, Mairepav6, Torino, Italy, 2009.
[28] G. Rowe, Evaluation of the relationship between asphalt binder properties and
non-load related cracking. prepared discussion, J. Assoc. Asphalt Paving
Technol. 80 (2011) 649–663.
[29] R.M. Anderson, G.N. King, D.I. Hanson, P.B. Blankenship, Evaluation of the
relationship between asphalt binder properties and non-load related cracking,
J. Assoc. Asphalt Paving Technol. 80 (2011) 615–663.
[30] H. Xu, A. McIntyre, T. Adhikari, S.A.M. Hesp, P. Marks, S. Tabib, Quality and
durability of warm rubberized asphalt cement in Ontario, Transp. Res. Rec.
2370 (2013) 26–32.
[31] AASHTO (American Association of State Highway and Transportation Officials).
Standard Practice for Accelerated Aging of Asphalt Binder Using a Pressurized
Aging Vessel (PAV), AASHTO R 28-09, AASHTO, Washington, D.C., 2009.
[32] AASHTO (American Association of State Highway and Transportation Officials).
Standard Practice for Grading or Verifying the Performance Grade (PG) of an
Asphalt Binder. R 29-08, AASHTO, Washington, D.C., 2008.
[33] H. Ding, J. Gyasi, S.A.M. Hesp, P. Marks, Y. Nie, M. Somuah, S. Tabib, N. Tetteh, I.
Ubaid, Performance grading of recovered asphalt binders from recent Ontario
paving contracts (Submitted for Publication, October 2017).
[34] F. Migliori, M. Pastor, G. Ramond, Etude statistique de quelques cas de
fissurations thermiques, in: Proc., Fifth Eurobitume Congress, Stockholm,
1993, pp. 724–728.
[35] F. Migliori, G. Ramond, M. Ballie, B. Brule, C. Exmelin, B. Lombardi, J. Samanos,
A.F. Maia, C. Such, S. Watkins, Correlations between the thermal stress
cracking of bituminous mixes and their binders’ rheological characteristics
Paper No. 046, Luxemburg, 3-6 May, in: Proc., Eurobitume Workshop on
Performance Related Properties for Bituminous Binders, 1999, pp. 1–5.
[36] I. Widyatmoko, M.W. Heslop, R.C. Elliott, The viscous to elastic transition
temperature and the in situ performance of bituminous materials, J. Inst.
Asphalt Technol. 14 (2005) 3–7.
[37] Kriz, P. DSR-PAV Test Improvement. Task Force Status Update, Asphalt
Institute Technical Advisory Committee Meeting, December 7, 2016.
[38] Glover, C.J., Davidson, R.R., Domke, C.H., Ruan, Y., Juristyarini, P., Knorr, D.B., Jung,
S.H. Development of a New Method to Assess Asphalt Binder Durability with
Field Validation. Report No. FHWA-TX-05/1872-2. Texas Transportation Institute,
Texas A&M University System, College Station, Texas 77843-3135, 2005.