Construction and Building Materials 93 (2015) 695–702

Contents lists available at ScienceDirect

Construction and Building Materials

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

Characterizing the cracking and fracture properties of geosynthetic

interlayer reinforced HMA samples using the Overlay Tester (OT)
Lubinda F. Walubita a,b, Abu N.M. Faruk b, Jun Zhang c,⇑, Xiaodi Hu d
a
PVAMU – The Texas A&M University System, Prairie View, TX 77446, USA
b
TTI – The Texas A&M University System, College Station, TX 77843, USA
c
Department of Civil Engineering, Texas A&M University, College Station, TX 77843, USA
d
Wuhan Institute of Technology, Wuhan, Hubei Province, China

h i g h l i g h t s

 Reflective cracking is one of distresses occurring in the hot-mix asphalt (HMA) overlays.
 Crack-impeding interlayer materials are used to improve the life of the HMA overlays.
 The Overlay Tester (OT) is used to evaluate cracking resistance of HMA samples in the laboratory.
 Different geosynthetic interlayer materials embedded in HMA samples are evaluated by the OT.
 Crack resistance of geosynthetic interlayer reinforced samples are substantially improved.

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

Article history: Reflective cracking is one of the undesirable distresses occurring in hot-mix asphalt (HMA) overlays;
Received 19 February 2015 costing highway agencies millions of tax payer dollars in maintenance and rehabilitation activities. To
Received in revised form 9 June 2015 mitigate this distress, crack-impeding interlayer materials such as geosynthetic interlayers are specified
Accepted 12 June 2015
to protect the HMA overlays as part of maintenance and rehabilitation strategies. Currently however,
Available online 27 June 2015
there is no universally standardized laboratory crack test method to aid in the selection of the most
appropriate geosynthetic interlayer material for maximum crack resistance and performance. This study
Keywords:
was undertaken to evaluate the laboratory cracking-resistance and fracture performance of different
HMA overlays
Reflective cracking
geosynthetic interlayer materials embedded in HMA samples. As a means to investigate its applicability
Geosynthetic interlayer for testing interlayer materials, the Overlay Tester (OT), in a monotonic tensile loading mode
Geogrids (3.375 mm/min) at 0 °C, was explored as the study test method. Eight different geosynthetic interlayer
Paving fabric products with different properties were compared to a ‘control specimen’ using a dense-graded HMA
Paving mat mix. Although field validation is still warranted, the study results indicated a substantial improvement
Composite grid (over 20%) in the laboratory crack and fracture performance of the geosynthetic interlayer reinforced
Overlay Tester (OT) samples over the Control samples; suggesting that use of these interlayer materials may be beneficial
Fracture
in mitigating reflective cracking in HMA overlays. For the test conditions considered, the OT test in mono-
tonic tensile-loading mode also exhibited promising potential as a rapid crack test method for testing
geosynthetic interlayer materials. The sample crack failure mode, test repeatability, and statistical
variability in the test data were generally within reasonable expectations.
Ó 2015 Elsevier Ltd. All rights reserved.

1. Introduction mitigate reflective cracking in existing pavements, various meth-

Reflective cracking is one of the undesirable structural dis-
tresses occurring in hot-mix asphalt (HMA) overlays over flexible
and concrete pavements; costing highway agencies millions of
tax payer dollars in maintenance and rehabilitation activities. To
⇑ Corresponding author. Tel.: +1 (979) 845 9291.
E-mail address: j-zhang@tamu.edu (J. Zhang).
ods including application of crack-impeding geosynthetic interlay-
ers such as paving grids, paving mats, or paving fabrics are often
used in maintenance and/or rehabilitation projects as part of the
HMA overlay construction; see Fig. 1 [1–3].
As illustrated in Fig. 1, the primary role of a geosynthetic inter-
layer is to arrest the upward propagation of cracks from an existing
pavement to the surface. The geosynthetic interlayer is used to
mitigate the cracks from reflecting through the HMA overlay to

http://dx.doi.org/10.1016/j.conbuildmat.2015.06.028
0950-0618/Ó 2015 Elsevier Ltd. All rights reserved.
696 L.F. Walubita et al. / Construction and Building Materials 93 (2015) 695–702
Fig. 1. Geosynthetic interlayer and HMA pavement construction.
the surface. Many types of geosynthetic interlayer materials are
presently available in the commercial market and are widely used
in HMA overlay projects during maintenance and rehabilitation
projects [3].
Currently however, there is no universally standardized labora-
tory crack performance test method to aid in the selection of
geosynthetic interlayer materials. Engineers therefore have to rely
on the index properties of the geosynthetic interlayer products
alone, as there is minimal or no crack performance data on geosyn-
thetic interlayers when embedded in HMA. Consequently, a funda-
mental challenge exists to provide a laboratory crack performance
test method that can evaluate geosynthetic interlayer materials
that would best improve the crack performance of any asphalt
layer. Therefore, for a given overlay, on a pavement (PVMNT) with
a cracked existing surface, how does one select the best geosyn-
thetic interlayer material to effectively mitigate the reflective
cracking? Which of the laboratory crack test methods could one
use to fundamentally characterize the fracture properties and
accurately evaluate the crack performance potential of the geosyn-
thetic interlayer materials when embedded in HMA? With this
background, this laboratory study was initiated to address the fol-
lowing primary objectives:
(1) To investigate if the Overlay Tester (OT), when run in mono-
tonic tensile-loading mode, could satisfactorily serve as a
laboratory crack test method for characterizing the fracture
properties and evaluating the interactive cracking resistance
potential of geosynthetic interlayers when embedded in
HMA samples.
(2) To comparatively evaluate the fracture properties of various
geosynthetic interlayer materials using the OT test method
and rank them in their order of superior cracking resistance
performance; in comparison to Control samples with no
interlayer.
To achieve these objectives, eight different commercial geosyn-
thetic interlayer materials with different properties were evaluated
alongside Control samples in the OT test using a dense-graded
HMA mix. As presented in this paper, the study also addressed
the following key aspects: (a) the fracture performance improve-
ment of each geosynthetic interlayer type/material over Control
samples; (b) a comparison of the fracture performance of different
geosynthetic interlayer types; (c) the potential of the OT test and
the measured fracture parameters to screen and effectively differ-
entiate between geosynthetic interlayer types; and (d) the OT test’s
repeatability and statistical variability in the lab test data when
run in monotonic tensile-loading mode.
In terms of the paper organization, the experimental test matrix
including the geosynthetic interlayers is discussed in the first sec-
tion, which is followed by the monotonic OT test setup and loading
configuration. Thereafter, the data analysis models are discussed;
followed by the test results and synthesis of the findings. The paper
then concludes with a summary of key findings and
recommendations.
2. Experimental plan and materials
Eight different geosynthetic interlayer products, referenced as
GIM1 thru to GIM8, were evaluated using a single dense-graded
HMA mix with 12.5 mm nominal maximum aggregate size (gran-
ite/quartzite) and PG 64-22 asphalt-binder (ffi5.0% by weight of
the aggregates). HMA samples, reinforced with each of the geosyn-
thetic interlayers, were tested using the OT to get the fracture
properties and crack resistance potential versus the Control HMA
samples. All geosynthetic interlayer types were installed using a
PG 64-22 hot asphalt tack coat. A minimum of three replicate spec-
imens were tested for each geosynthetic interlayer type.
HMA samples (150-mm diameter by 100-mm in height) con-
structed with the geosynthetic interlayer material were molded
using the Superpave Gyratory Compactor (SGC) in two layers,
namely the ‘‘bottom 50-mm thick HMA’’ plus tack-coat (PG
64-22) plus interlayer material plus the ‘‘top 50-mm thick HMA’’.
Then the geosynthetic interlayer reinforced HMA samples were
fabricated/cut to the final test specimen dimensions as shown in
Fig. 2.
The HMA samples were basically molded in two steps (bottom
50-mm thick layer first and top 50-mm thick layer second) follow-
ing the SGC volumetric procedures for fabrication of 150-mm
diameter HMA samples using a gyratory compactor. One hundred
and fifty-millimeter diameter samples of geosynthetic interlayers
were then cut and placed into the measured (by weight) amounts
of hot tack coat that had been placed on the base 50-mm thick
HMA layer. The tack coat rates were reflective of the individual
interlayer requirements by their manufacturer. The interlayer ori-
entation was marked on the sides of the base and then on the top
of the final 50-mm thick HMA layer thickness. As indicated in
Table 1, only one interlayer, GIM5 (a biaxial product), was mea-
sured off angle at 45° from the machine or cross-machine direction
and the product provided equivalent resistance regardless of the
orientation.
As shown in Fig. 2, all the HMA samples were cut to a total
thickness of 62.5 mm and then notched to 18.75-mm depth
(3.125-mm wide) from the bottom in the direction of the SGC com-
paction to simulate a reflective crack on an existing cracked pave-
ment (PVMNT) surface. While the geosynthetic interlayer was
located at 25-mm from the bottom in the direction of the SGC com-
paction, the notch depth was only 18.75-mm (instead of 25-mm)
so as to avoid cutting the interlayer material during the notching
process. With the notching, the effective thickness of the test spec-
imens was 43.75-mm over the notch; see Fig. 2. At an asphalt mix
specific gravity (Rice) value of 2.656 (Gmm), the average density of
the composite HMA samples (with the interlayer material) prior to
notching was measured to be fairly consistent at 87.45% with a
 L.F. Walubita et al. / Construction and Building Materials 93 (2015) 695–702 697

Fig. 2. HMA samples and example of One-Cut test specimen.

Table 1
Geosynthetic interlayer types.

# Test sample notation Geosynthetic interlayer type Tensile strength Replicates Comment
1 GIM4 Paving mat–fiberglass/polyester 25 kN/m P3
2 GIM5 Same material as GIM4, but off-angle at 45° ASTM D5035 P3
3 GIM1 Paving mat–fiberglass/polyester 50 kN/m P3 Oriented in the machine
ASTM D5035 direction within the HMA samples
4 GIM2 Composite grid–fiberglass grid / polyester 50 kN/m P3
ASTM D4595
5 GIM3 Composite grid–fiberglass grid / polypropylene 50 kN/m P3
ASTM D4595
6 GIM6 Paving mat–fiberglass/polyester blend 200 N/50 mm P3
ASTM D5035
7 GIM8 Paving fabric–nonwoven polypropylene 450 N P3
ASTM D4632
8 GIM7 Composite grid–polyester 50 kN/m P3
ASTM D5035
9 Control None P3
Total P27

coefficient of variation (COV) of 1.06%. The cut specimen dimen- Table 2

sions were also consistent in terms of length (150-mm), width OTM test parameters and loading configuration.
(75-mm), and thickness (62.5-mm) with zero standard deviation # Item Parameter or characteristic
(Stdev) and COV; see Figs. 2 and 4.
1 Monotonic tensile 3.375 mm/min (0.135 in./min) in tension
As per the Tex-248-F specification, high strength 2-ton 2-part loading rate mode
epoxy (14 ± 2 g) was used for gluing the specimen to the OT plates 2 Sample relaxation time 10 min  necessary to relieve residual stresses
[4]. The epoxy was allowed to set and cure for a minimum period prior to testing in the specimen after bolting/screwing it to
of 12 h (overnight) with a 44.5 N (10-lb) curing weight at room the machine
3 Test temperature 0 °C (to simulate cold-weathercracking/
temperature (i.e., 25 °C). To minimize the effects of oxidative aging
fracture issues) [3]
during storage, the samples were kept in a cold room at around 4 Sample temperature P12 h (overnight) at 0 °C
15 °C in plastic zip-lock bags [5]. conditioning
5 Temperature Via a thermocouple wire inserted into a
conditioning dummy OT sample put in the same chamber
3. OTM test setup and loading configuration monitoring as test specimens
6 Data capturing Every 0.04 s so as to capture sufficient data for
The Overlay Tester (OT) was conducted in monotonic tensile subsequent analysis
7 Test time 610 min for one specimen
loading mode, denoted as OTM. The OTM test setup and loading 8 Test specimen 150-mm long by 75-mm breadth 62.5-mm
conditions are listed in Table 2. The 3.375 mm/min (0.135 in./min) dimensions thickness
loading rate at 0 °C was established after several trial testing with 9 Notch dimensions 18.75-mm long by about 3.125 mm width; see
numerous dummy OT samples interlayered with the same geosyn- Fig. 6
10 Effective test specimen 43.75-mm; see Figs. 2 and 4
thetic interlayer materials listed in Table 1 [6]. The framework for
thickness
establishing this loading rate was for the OTM output to yield rea-
sonably acceptable load–displacement response curves and mean-
ingfully interpretable data [7,8]. The OTM machine and sample namely the material tensile strength and strain. The analysis mod-
setup are exemplified in Fig. 3. els for computing these fracture parameters are discussed in the
subsequent text.

4. Data analysis models and statistics

During OTM testing, the measurable parameters include the
applied load, the horizontal tensile displacement rate, time, and
test temperature. The primary output of the OTM test is the load–
displacement (L–D) response curve exemplified in Fig. 4. From
the OTM output data and the L–D response curve shown in Fig. 4,
the following load and fracture parameters were computed,
4.1. Data analysis models
As previously stated, the primary output from the OTM test is
the L–D response curve (Fig. 4) that is used to calculate the fracture
parameters of the test sample. Table 3 lists the load and calculated
fracture parameters together with their respective mathematical
models [7–9].
698 L.F. Walubita et al. / Construction and Building Materials 93 (2015) 695–702

Fig. 3. OT machine and sample setup.

Fig. 4. OTM load–displacement (L–D) response curve.

Table 3
OTM fracture parameters and analysis models.

# Fracture parameter Notation Unit Mathematical model

1 Peak failure (tensile) load Pmax kN Peak load from load–displacement response curve; see Fig. 4
2 Material tensile strength rt kPa rt ¼ CrossPeak load
section area
¼ Pmax
tb
3 Material tensile strain at peak failure load – ductility potential et (mm/mm) DPmax Do
et ¼ dp
4 Effective specimen cross-sectional area Ae mm2 tb ffi (lcr  b) ffi b(t  dn) ffi 75  (62.5–18.75) ffi 3281.25 mm2

Legend: t = specimen thickness (62.5 mm), b = specimen width (75 mm), Do = initial displacement measured at start of test, DPmax = displacement measured at peak load,
dp = opening of OT base plate (2 mm), lcr = length traversed by crack (t  dn), dn = notch length or depth (18.75 mm).

4.2. Statistical tools and methods 5. Lab test results and analysis

To authenticate the validity of the test results and the corre-
sponding findings in terms of the reliability/repeatability of the
OTM test, variability in the laboratory test data, and performance
differences among the different geosynthetic interlayer materials,
the following statistical tools were engaged:
(a) Standard MS Excel descriptive statistics such as average
(Avg), standard deviation (Stdev), and coefficient of variance
(COV).
(b) Comprehensive statistical analysis using t-Tests, ANOVA,
and Tukey’s HSD methods [10].
For OTM repeatability and variability assessment in the test
data, a COV of 30% was used as the threshold value (i.e.,
COV 6 30%) [5,7–9]. For the t-Tests, ANOVA, and Tukey’s HSD, sta-
tistical analyses were performed at both 90% and 95% confidence
levels, respectively.
Table 4 lists the OTM test results in a rank order of superiority
based on the computed fracture parameters; which are individu-
ally discussed in the subsequent text. A graphical plot of the L–D
response curves is also shown in Fig. 5. All these lab test results
and graphical plots represent an average of three replicate test
specimens per material type.
5.1. The peak tensile failure load, Pmax
In terms of the magnitude of Pmax, Table 4 shows that GIM1 is
the superior material consecutively followed GIM2, GIM3, GIM4,
GIM5, GIM6, GIM7, GIM8, and lastly, the Control. From Table 4
and Fig. 5, it is also evident that the Pmax of GIM2 and GIM3 are
insignificantly different and so, are GIM4 versus GIM5 and GIM6
versus GIM7. As theoretically expected, the least amount of tensile
force (Pmax) to cause failure was recorded in the Control samples
 L.F. Walubita et al. / Construction and Building Materials 93 (2015) 695–702 699

Table 4 example, use of the GIM1 interlayer material leads to over 50%
OTM test results. improvement in the rt of the HMA samples; meaning it can
Rank# HMA sample notation Pmax (kN) rt (kPa) et (mm/mm) withstand up to over 1.5 times the tensile stress sustainable by
1 GIM1 (Paving mat) 8.74 2579 3.36E01 the Control samples without the interlayer material. Overall, all
2 GIM2 (Composite grid) 8.01 2364 2.57E01 the interlayer materials exhibited over 10% improvements in the
3 GIM3 (Composite grid) 7.96 2348 2.49E01 HMA tensile strength, as compared to the Control specimens
4 GIM4 (Paving mat) 7.48 2207 2.07E01 without any interlayer materials.
5 GIM5 (Paving mat) 7.4 2184 1.91E01
6 GIM6 (Paving mat) 6.66 1966 1.63E01
7 GIM7 (Composite grid) 6.6 1948 1.58E01 5.3. The material failure tensile strain, et
8 GIM8 (Paving fabric) 5.92 1748 1.24E01
9 Control 5.38 1588 1.05E01 For the purpose of this study, tensile strain was defined as the
ductility and elongation potential of the material (or sample) prior
to tensile failure (i.e., break at peak load) – the higher in magnitude
without any interlayer material. Compared to the interlayered the better in terms of ductility. As can be seen in Table 4 and Fig. 5,
samples, this means that the Control samples would fail at a lower the interlayered samples, with comparatively higher et values in
tensile loading while almost twice the tensile loading would be magnitude, exhibited superiority over the Control samples. Like
required to cause failure in the samples interlayered with GIM1. with the preceding fracture parameters, the GIM1 thru GIM3 mate-
rials ranked at the top in terms of performance superiority based
5.2. The material tensile strength, rt on the et magnitude followed by the GIM4/5 materials. Similar to
Pmax and rt, GIM8 was the poorest asphalt-interlayer material,
Tensile strength is quantitatively defined as the maximum indicating an average of only 12% et improvement over the
amount of tensile stress that a material can be subjected to before Control samples.
fracture failure. Theoretically, the higher the rt in magnitude, the
better is the material’s crack resistance potential. As shown in 5.4. Material discriminatory ratio (DR)
Table 3, rt is primarily computed as function of Pmax and therefore,
its response trend mirrored that of Pmax; see Fig. 6 below. For the simplicity of interpretation of the results, the concept of
Just like for Pmax, the rank order of material superiority based discriminatory ratio (DR) was also introduced in the data analysis
on the rt magnitude in Fig. 6 is GIM1 > GIM2 > GIM3 > GIM4 > [7]. DR is an arithmetic ratio of two corresponding parametric val-
GIM5 > GIM6 > GIM7 > GIM8 > Control; which is somewhat consis- ues (e.g., rt) comparing a good material with a relatively poor or
tent with theoretical expectations, particularly for the Control reference material such as GIM1 versus the Control. The larger
samples. Clearly, these results highlight the significant benefits of the DR in magnitude, the greater the difference between the mate-
using these interlayer materials to enhance the HMA tensile rials and the more effective the fracture parameter is in discrimi-
strength against reflective cracking and fracture damage. For nating and differentiating the two materials. A listing of these DR

Fig. 5. OTM material L–D response curves.

Fig. 6. rt lab results in rank order of superiority.

700 L.F. Walubita et al. / Construction and Building Materials 93 (2015) 695–702

results and the corresponding performance improvement factors
over the Control samples are shown in Table 5. The interlayer crack
performance factor (ICPF) values in Table 5 were computed as fol-
lows: ICPFi = 10  (DRi  1.0). To convert the ICPF values into per-
centage, as given in parentheses in Table 5, simply multiple the
ICPF by 10, i.e., %ICPFi = 10  ICPFi or %ICPFi = 100  (DRi  1.0).
In terms of the sustainable tensile stress and strain prior to
cracking and fracture failure, the interlayer crack performance fac-
tors (ICPFs) computed (i.e., ICPFi = 10  (DRi  1.0)) in Table 5
shows a significant improvement in performance with the usage
of geosynthetic interlayer materials. The greatest performance
improvement is indicated for the GIM1 thru to GIM3 materials
with ICPF values greater than 50% and 40%, respectively; followed
by the GIM4/5 materials. Like for GIM4 versus GIM5 or GIM2 ver-
sus GIM3, the GIM6 and GIM7 materials’ performance improve-
ments are hardly different. As theoretically expected, the least
performance improvement is indicated for GIM8 with an ICPF of
only 10% for tensile strength.
When looking at the DR magnitudes, it is evident that the strain
exhibits superiority in terms of differentiating the materials. For
instance, the strain indicates that GIM1 is about 2.3 times better
than the Control while it is about 1.62 times for the stress.
Similarly, the stress suggests that GIM2 and GIM3 are insignifi-
cantly different, which is not so apparent with the strain.
6. Variability and statistical analysis
Variability and statistical analysis of the lab test data was based
on the standard MS Excel descriptive statistics, t-Tests, ANOVA,
and Tukey’s HSD methods are discussed in the subsequent text.
And like all the preceding results, these statistical analyses were
also based on three replicate test specimens per material type.
6.1. Standard MS Excel descriptive statistics
OTM test repeatability and variability in the test results was
assessed using a COV of 30% as the threshold value, i.e.,
COV 6 30% [7,9]. Fig. 7 shows a bar-chart plot of the COV results
as a function of material type and the corresponding fracture
parameters.
As can be observed from Fig. 7, the three parameters generally
indicate acceptable repeatability with very low COV values that are
less than 30%. For all the materials evaluated including the Control
samples, the COV for both Pmax and tensile strength are in fact less
than 10%; reflecting a fairly high level of OTM repeatability and sta-
tistical confidence in the results for these fracture parameters gen-
erated under monotonic testing.
Overall, these descriptive statistical analyses implicitly suggest
that OT test in monotonic tensile-loading mode is fairly a repeat-
able test based on the three evaluated fracture parameters namely
Pmax, rt, and et.
6.2. ANOVA, t-Tests, and Tukey’s HSD analysis
ANOVA, t-Tests and Tukey’s HSD analyses at 90% and 95% con-
fidence levels indicated that there was a statistically significant dif-
ference in the fracture performance among the materials for all the
parameters evaluated – that is, at least two or more materials were
statistically significantly different [10]. An example of these results
and ranking/groupings of the materials are shown in Fig. 8.
In terms of the interpretation of the statistical data shown in
Fig. 9, the materials having parametric values that are statistically
not significantly different would be categorized in the same Group,
e.g., A, B, or C. Materials categorized in Group A would have higher
numerical values or more superior than materials listed in Group B

Table 5
DR results and performance improvement factors.
Discrimination Ratio (DR) Interlayer Crack Performance Factor (ICPF)
Material Ranking Load (Stress) Strain Load (Stress) Strain
GIM1 1.62 3.20 ≅ 6 (62.35%) ≅ 22 (>100%)
GIM2 1.49 2.45 ≅ 5 (48.84%) ≅ 14 (>100%)
GIM3 1.48 2.37 ≅ 5 (47.80%) ≅ 14 (>100%)
GIM4 1.39 1.97 ≅ 4 (38.96%) ≅ 9 (97.14%)
GIM5 1.38 1.82 ≅ 4 (37.53%) ≅ 8 (81.90%)
GIM6 1.24 1.55 ≅ 2 (23.78%) ≅ 5 (55.24%
GIM7 1.23 1.50 ≅2 (22.61%) ≅ 5 (50.48%)
GIM8 1.10 1.18 ≅ 1 (10.03%) ≅ 1 (18.09%)
Control (Reference) 1.00 1.00 N/A (Reference)

Fig. 7. Coefficient of variation (COV) results.

 L.F. Walubita et al. / Construction and Building Materials 93 (2015) 695–702 701

Materials Compared: All Materials

Type of Analysis: t-Tests, ANOVA, & Tukey's HSD

Are the Materials Stascally Significantly Different?

Fracture Parameter GIM1 GIM2 GIM3 GIM4 GIM5 GIM6 GIM7 GIM8 Control
Peak Load, Pmax (kN) A A A B B C C D E
Tensile Strength, t (kPa) A A A B B C C D E
Failure Strain, t (mm/mm) A B B C C D D E F
Fialure MicroStrains, t A B B C C D D E F

Fig. 8. Statistical comparisons – all materials.

Fig. 9. Undesirable (GIM8#01) and desirable (GIM8#02) crack failure modes.

for the same fracture parameter whereas, materials in Group B
would have higher numerical values than materials in Group C,
etc., and that the difference in their numerical values are statisti-
cally significant [7]. Additionally, these statistical methods also
takes into account the variability associated with each parameter
and the sample replicates when grouping and ranking the materi-
als [7,10]. Based on these statistical interpretations, the following
can be inferred from Fig. 9:
 With the exception of the tensile strain parameter, GIM1, GIM2,
and GIM3 all fall in the same Group A ranking and therefore,
have an insignificant difference in their fracture performance
for the HMA mix evaluated under OTM testing.
 The GIM4 and GIM5 are not statistically significantly different
in terms of fracture performance under OTM testing and fall in
the same group ranking for each respective fracture parameter.
 Unexpectedly, GIM6 and GIM7 are categorized in the same sta-
tistical Group C; indicating a statistically insignificant difference
in fracture performances based on the Pmax, tensile strength,
and strain parameters, respectively.
 For all the fracture parameters considered in Fig. 8, the Control
ranks differently as the poorest material; which is consistent
with theoretical expectations. The second poorest is GIM8,
which according to Fig. 8 is statistically significantly different
from all the other interlayer materials.
Overall, it can be concluded from Fig. 8 that the GIM1 thru to
GIM3 materials (in Group A) exhibited similar superior statistical
fracture performance, consecutively followed by the GIM4 and
GIM5 materials (in Group B), then GIM6, GIM7, GIM8, and lastly,
the Control (poorest).
7. Discussion and synthesis of the results
As expected, the interlay reinforced samples out-performed the
Control samples, sustaining more tensile load and strain and
therefore, withstanding more energy prior to fracture damage
and crack failure. Compared to the Control samples, these results
and findings support that the usage of geosynthetic interlayers in
HMA overlays may serve to minimize reflective cracking and frac-
ture damage. In general, the following can be inferred from the
results/findings and observations made in this study:
 The Pmax and tensile strength parameters exhibited the least
variability, with COV values less than 10%. These two parame-
ters also indicated a similar fracture performance differentia-
tion and ranking of the materials evaluated. This was not
unexpected because, as indicated in Table 3, tensile strength
is primarily computed as a function of Pmax. Between the two
parameters however, the tensile strength (rt) should be pre-
ferred over Pmax for comparatively characterizing the fracture
performance of interlayer materials.
 With respect to the GIM1 thru to GIM3 versus the GIM4/5 mate-
rials, all the fracture parameters indicated a similar discrimina-
tory ranking, with the former being superior to the later.
 Likewise and as theoretically expected, the Control samples
turned out to be the poorest.
Almost all the samples tested exhibited the theoretically
expected fracture failure mode with the cracks propagating
through the interlayer material up to the sample surface. One out-
lier, GIM8 sample#01, exhibited an undesirable shear failure mode
with the sample shearing off at the layer interface without propa-
gating through the interlayer material; see Fig. 9. Where material
was available, some additional samples were remolded and
retested in the case of undesirable crack failures as exemplified
in Fig. 9.
As well as being credited to good workmanship (i.e., being
meticulous and consistent, adhering to procedures, use of similar
operators/technicians, etc.), this observation may suggest that the
OTM has promising potential as a good candidate test method for
evaluating interlayer materials and characterizing their fracture
properties in the lab to mitigate reflective cracking in HMA
702 L.F. Walubita et al. / Construction and Building Materials 93 (2015) 695–702
pavements. However, it should be noted that these results and
findings pertain only to the HMA mix, interlayer materials, and
OTM test conditions defined in this study. Therefore, the overall
conclusions may not be exhaustive.
8. Summary and recommendations
In this laboratory study, eight different geosynthetic interlayer
products were comparatively evaluated for their crack-resistance
potential and fracture performance using the OTM test in mono-
tonic tensile loading mode (i.e., 3.375 mm/min); at a test temper-
ature of 0 °C. Numerous composite HMA samples reinforced with
geosynthetic interlayer products were monotonically tested in
the lab and compared with Control samples, but fabricated from
the same dense-graded HMA mix. The key findings, conclusions,
and recommendations drawn from this study are summarized as
follows:
 With the exception of three outlier (i.e., one out of 27 samples)
that failed in shear at the geosynthetic interlayer layer interface,
all the other samples failed in fracture mode with the crack
neatly propagating through the geosynthetic interlayer material
up to the surface; as theoretically expected. The outlier (GIM8
Sample# 01) represents only 3.7% of the total sample pool,
which is statistically insignificant at 95% confidence level.
 The OTM test generally exhibited good repeatability with a
reasonably acceptable level of statistical variability in the lab
test data. All the three computed fracture parameters had
COV values less than the 30% threshold; with the tensile
strength (rt) being one of the least variable parameter.
 On average, the tensile strength (rt) parameter exhibited the
least statistical variability with COV values less than 10% and
also exhibited a fairly strong discriminating potential among
the different interlayer products and capturing the differences
in their fracture performance potential.
 Consistent with theoretical expectations, there was a marked
improvement in the cracking resistance and fracture perfor-
mance with the introduction of the geosynthetic interlayer
materials in the HMA samples. Compared to the Control sam-
ples, for example, there was in fact over 40% improvement in
laboratory fracture performance with use of the GIM1(>50%)
and GIM2/3 (>40%) interlayer materials.
 Considering all the fracture parameters evaluated (with empha-
sis on the tensile strength [rt]) and the statistical analysis meth-
ods engaged, the rank order of superior fracture performance
for all the materials is as follows: GIM1 > GIM2 > GIM3 >
GIM4 > GIM5 > GIM6 > GIM7 > GIM8 > Control.
Although field validation is strongly warranted, this study has
provided laboratory evidence that use of geosynthetic interlayer
materials has potential to improve the fracture performance of
HMA overlays. Thus, the usage and application of these geosyn-
thetic interlayer materials should be optimized, particularly where
crack mitigation is needed.
Notwithstanding the need for additional lab testing and field
correlations, the OTM, with the test conditions and loading
parameters derived in this study exhibited promising potential as
a crack test method for characterizing the fracture properties and
crack-resistance potential of geosynthetic interlayer materials.
The sample crack failure mode, test repeatability, and statistical
variability in the OTM test data were generally within reasonable
tolerances for this study. However, it is strongly recommended
that the OTM test method should be explored further along with
correlations to other standard crack tests (i.e., the dynamic loading
OT, etc.) and validation with field performance data.
Disclaimer
The contents of this paper reflect the views of the authors who
are responsible for the facts and accuracy of the data presented
herein and do not necessarily reflect the official views or policies
of any agency or institute. This paper does not constitute a stan-
dard, specification, nor is it intended for design, construction, bid-
ding, contracting, tendering, or permit purposes. Trade names were
used solely for information purposes and not for product
endorsement.
Acknowledgements
The authors thank Saint-Gobain ADFORS for their financial sup-
port and all those who helped during the course of this research
work. In particular, the technical assistance and guidance provided
by Daniel Hunt and Tao Yu is gratefully acknowledged. Special
thanks and gratitude also go to all those who assisted with labora-
tory work, data analysis, and documentation, namely: Stanford
Nguyen, Jason Huddleston, Geoffrey S. Simate, and Raenita Hassan.
References
[1] Cleveland GS, Button JW, Lytton RL. Geosynthetics in Flexible and Rigid
Pavement Overlay Systems to Reduce Reflection Cracking. Research Report
FHWA/TX-02/1777; 2002. p. 1–297.
[2] Ali Khodaii, Shahab Fallah, Fereidoon Moghadas Nejad, Effects of geosynthetics
on reduction of reflection cracking in asphalt overlays, Geotext Geomembr 27
(1) (2009) 1–8.
[3] Mostafa Elseifi, Rakesh Bandaru. Cost effective prevention of reflective
cracking of composite pavement. Research Report FHWA/LA.11/478.
Louisiana State University, Baton Rouge, LA 70803, USA; 2011.
[4] Texas Department of Transportation. Online Manuals: The Overlay Test – Test
Designation Tex-248-F. Austin, TX, USA. Accessed March 2014.
[5] Walubita LF, Faruk ANM, Das G, Tanvir HA, Zhang J, Scullion T. The Overlay
Tester: a sensitivity study to improve OT repeatability and minimize
variability in the ot test results. Technical Research Report FHWA/TX-10/0-
6607-1. Texas A&M Transportation Institute, College Station, TX, USA; 2012.
[6] Carver C, Sprague J. Asphalt overlay reinforcement, the state of the practice,
summarized. Geotechnical Fabrics Report; March 2011.
[7] A.N.M. Faruk, X. Hu, Y. Lopez, L.F. Walubita, Using the fracture energy index
concept to characterize the HMA cracking resistance potential under
monotonic crack testing, Int J Pavement Res Technol 7 (1) (2014) 40–48.
[8] L.F. Walubita, A.N.M. Faruk, A.E. Alvarez, T. Scullion, The Overlay Tester (OT):
using the fracture energy index concept to analyze the OT monotonic loading
test data, Constr Build Mater 40 (2013) 802–811.
[9] Walubita, et al. The Overlay Tester: comparison with other crack test methods
and recommendations for surrogate crack tests. Technical Research Report#.
FHWA/TX-12/0-6607-2. TTI – Texas A&M University System, College Station,
Texas, USA; 2013.
[10] Tukey JW. The problem of multiple comparisons. Princeton, New Jersey,
Princeton University, USA; 1953 [Unpublished manuscript].