Geo-Frontiers 2011 © ASCE 2011 4563

 
 

Evaluation of the Cantabro Durability Test for Dense Graded Asphalt

Jesse D. Doyle1, A.M. ASCE and Isaac L. Howard2, Ph.D., A.M. ASCE
Downloaded from ascelibrary.org by Universidad Militar Nueva Granada on 05/09/23. Copyright ASCE. For personal use only; all rights reserved.

1
Graduate Research Assistant, Department of Civil and Environmental Engineering,
Mississippi State University, P.O.Box 9546, Miss State, MS 39762; PH (662) 325-
3050; email: [email protected]
2
Assistant Professor, Department of Civil and Environmental Engineering,
Mississippi State University, Box 9546, Miss State, MS 39762; PH (662) 325-7193;
email: [email protected]

ABSTRACT

The Cantabro durability test is typically used for open graded asphalt mixtures
and has seen little use with dense graded mixtures. This paper presents durability
data from the Cantabro test for a number of dense graded mixes meeting current low
Reclaimed Asphalt Pavement (RAP) specifications with the goal of evaluating the
test method for use with dense graded mixtures. The purpose is to assist with
developing a baseline of durability performance of current low RAP dense graded
mixtures for future comparison to dense graded mixtures with elevated RAP contents.
Eight mixtures were tested for this paper; the focus was on low RAP content mixture
types commonly used in Mississippi, especially those used for rehabilitation
applications. Data from mixes with of a variety of aggregate types, gradations, and
binder grades are presented. Test data from laboratory-mixed laboratory-compacted,
plant-mixed field-sampled laboratory-compacted, and plant-mixed quality control
specimens are included. The effects of changes in binder content on durability results
are investigated for 0% RAP mixtures. Overall, results were found to be repeatable
and the Cantabro durability test was found to be a candidate for future comparison of
dense graded mixture durability performance of conventional mixes to high RAP
content mixes.

INTRODUCTION AND LITERATURE REVIEW

As part of an ongoing research effort, the authors have investigated

application of the Cantabro test to evaluate durability of dense graded asphalt surface
mixtures containing high RAP contents. Use of elevated levels of RAP (up to 50%)
in warm mixed surface course rehabilitation mixtures is an economically attractive
alternative to current low RAP content mixtures. Mixtures of this nature are currently
under investigation in Mississippi. Current specifications restrict the level of RAP
allowed in surface mixtures. The increased amount of aged RAP bitumen contained
in high RAP content mixtures is an area of potential concern for surface mixtures in
terms of durability and the potential for raveling. Increasing the percentage RAP of
in asphalt surface mixtures presents a number of unique challenges due to the large

Geo-Frontiers 2011
 Geo-Frontiers 2011 © ASCE 2011 4564
 
 

contribution of the aged RAP asphalt to the properties of the recycled mixture. The
Cantabro durability test is used as an indirect measure of the potential for raveling
and durability problems in Open Graded Friction Course (OGFC), also known as
Porous Friction Course (PFC), asphalt mixtures. It has gained acceptance as a
screening tool for OGFC and PFC but no published literature was identified
documenting its use for conventional dense graded asphalt mixtures. To utilize the
Cantabro test for screening high RAP mixtures, a range of acceptable values for dense
Downloaded from ascelibrary.org by Universidad Militar Nueva Granada on 05/09/23. Copyright ASCE. For personal use only; all rights reserved.

graded surface mixtures designed according to current (low RAP) specifications is

needed for comparison of performance.
The Cantabro test is often used in design of OGFC mixtures as a measurement
of durability and of the potential for aggregate loss from mixtures (Watson et. al.
2003). An upper limit of aggregate loss for un-aged OGFC mixture specimens of
20% has been recommended (Watson et. al. 2004). Use of polymer-modified binders
was found to noticeably reduce the aggregate loss compared to specimens made with
an un-modified binder and the same aggregate type and gradation (Watson et. al.
2004). No information was found in the literature regarding use of the Cantabro test
for dense-graded asphalt mixtures.
Alvarez et. al. (2008) found the Cantabro test to be somewhat sensitive to
changes in fundamental binder properties due to aging but stated that test results
might be more influenced by aggregate properties of the mixtures tested than by the
binder properties. The same authors did suggest use of the Cantabro test as a tool to
aid in selection of the most promising material combinations in terms of durability for
OGFC mixtures. Based on limited data set, Kraus (2008) provided evidence of a
possible relationship between Cantabro aggregate loss for mixtures and results of
DSR testing on the polymer-modified binder components of the mixtures; testing was
performed on both un-aged and laboratory aged binders and mixtures.

EXPERIMENTAL PROGRAM

Results of the Cantabro durability test are used in this study as a relative
measurement of durability between different mixtures. The propensity for aggregate
loss for a given mixture was utilized as a measure of the brittleness and potential for
raveling of that mixture. Cantabro test durability data was collected for typical
Mississippi Department of Transportation (MDOT) approved dense-graded asphalt
mixtures as well as for mixtures designed in the laboratory for this study. Typical
Mississippi asphalt mixtures will contain RAP whenever it is available to reduce cost;
relatively few MDOT mixtures do not contain any RAP. The maximum percentage
of RAP currently allowed in MDOT approved surface mixtures is 15%. Many
Mississippi mixtures contain local crushed gravel as well as imported crushed stone.
The focus of this study was to establish the range of durability properties of typical
dense-graded surface mixtures.
Specimens for durability testing were prepared by three different methods: 1)
laboratory-mixed and laboratory-compacted; 2) plant-mixed, field-sampled,
laboratory-compacted specimens; and 3) plant-mixed quality control (QC) specimens.
Specimens prepared by method 1) were prepared according to standard laboratory
methods. Specimens prepared according to method 2) were produced with asphalt

Geo-Frontiers 2011
 Geo-Frontiers 2011 © ASCE 2011 4565
 
 

mixture which was field sampled in bulk from overlay paving projects located near
the campus of Mississippi State University and returned to the laboratory in insulated
containers. The surface overlay projects were located on U.S. Highway 45 between
Tupelo and Meridian Mississippi. Specimens were compacted immediately from the
field sampled material without any additional heating. Specimens prepared by
method 3) were produced at an asphalt plant laboratory as production quality control.
Downloaded from ascelibrary.org by Universidad Militar Nueva Granada on 05/09/23. Copyright ASCE. For personal use only; all rights reserved.

Mixtures Tested

A total of eight mixtures were tested for this study; all were produced at
normal hot mix temperatures and their properties are presented in Table 1. Mixtures
1 and 2 did not contain RAP and were designed in the laboratory as a baseline for
comparison to mixtures containing RAP. Mixtures 1 to 6 contained aggregates from
the same sources, both local crushed gravel and imported crushed limestone.
Mixtures 3, 4, and 5 all had identical gradations; the only difference was the design
traffic level and subsequently different asphalt contents. Mixture 5b and c was
produced with the same asphalt content as Mixture 5a but neat PG 67-22 binder was
substituted to reduce cost when paving non-trafficked locations such as shoulders.
Mixture 7 contained crushed gravel from a different geographic location within
Mississippi that has lower water absorption (roughly 2.3%) than the gravel used in
Mixtures 1 through 6 (roughly 3.7%). Mixture 8 (19.0 mm NMAS) would not be
used as a surface mixture but was included for comparison; it did not contain any
crushed gravel and the source of limestone was the same as for Mixtures 1 to 6.
The asphalt content (AC) of the mixtures varied on some occasions from the
design asphalt content (Pb). This was noted when it occurred. All mixtures contained
neat PG 67-22 binder except Mixture 5a which contained PG 76-22.

Table 1. Properties of Mixtures Tested

Mix Preparation Aggregate Components1 (%) NMAS
ID Method Gravel Limestone Sand RAP (mm) Ndes Pb2
1 1 67 22 10 0 9.5 65 5.7
3
2 1 81 8 10 0 12.5 75 6.6
3 3 37 37 10 15 9.5 50 6.2
4 3 37 37 10 15 9.5 65 6.0
5a 1 37 37 10 15 9.5 85 5.8
5b 2 37 37 10 15 9.5 85 5.8
5c 1 37 37 10 15 9.5 85 5.8
6 2 39 35 10 15 12.5 65 5.4
7 1 63 15 6 15 12.5 85 5.6
8 1 0 74 10 15 19.0 65 4.1
1) All mixtures contained hydrated lime as anti-strip comprising 1% of the aggregate component.
2) Total design asphalt as percentage of mixture including contribution from RAP.
3) Mixture was designed according to specification for airfields with higher VMA requirements.

Geo-Frontiers 2011
 Geo-Frontiers 2011 © ASCE 2011 4566
 
 

Test Methods

The Cantabro durability test procedure used in this study is based on the
Mississippi test method for OGFC mixture abrasion loss (MT-85) (MDOT 2005) for
un-aged specimens and is described as follows. A 150 mm diameter gyratory
compacted specimen of nominal 115 mm height (controlled by mass of mixture) is
placed in the drum of an LA Abrasion testing machine without the charge of steel
Downloaded from ascelibrary.org by Universidad Militar Nueva Granada on 05/09/23. Copyright ASCE. For personal use only; all rights reserved.

spheres and subjected to 300 revolutions. The largest portion of the tested specimen
is removed from the drum, lightly brushed off with a cloth and weighed. The mass of
the specimen is recorded before and after the test and the loss in sample mass during
the test is reported as a percentage of the original mass. The variable ML is used to
designate percentage of mass loss in the Cantabro test.
Prior to testing, specimens were conditioned in an environmental chamber
overnight (minimum of twelve hours) at 25 C. A dummy specimen of similar
properties with an embedded thermocouple was exposed to the same conditions to
verify that the internal temperature of all specimens equilibrated to 25 C. The
temperature of the LA Abrasion drum was checked before every test and required to
be 25 ± 2 C before testing. All debris leftover from the previous test was removed
from the LA Abrasion drum before each test to ensure that there was no variability
introduced to the results due to cushioning of the test sample. Figure 1 visually
compares an untested specimen to several tested specimens of the same mixture at
varying binder contents and subsequently varying levels of mass loss.

Untested AC = 5.0 Pb = 6.6 AC = 7.5

Specimen ML = 22.3 ML = 9.1 ML = 6.8

Figure 1. Specimens of Mix 2 with Varying AC%

TEST RESULTS

Mixtures without RAP at Varying Asphalt Contents

Mixtures 1 and 2 containing no RAP were tested at a range of asphalt

contents to investigate the effect of asphalt content on durability; results are presented
in Figure 2. As the binder content is increased, or conversely as the air voids
decrease, the mass loss is reduced. At low asphalt contents the mass loss is
noticeably higher. From Figure 2a it appears that the trends of the 12.5 mm and 9.5
mm mixtures are similar with the trend of the 9.5 mm mix being lower than that of
the 12.5 mm mix. Design asphalt contents of the 9.5 mm and 12.5 mm mixtures are

Geo-Frontiers 2011
 Geo-Frontiers 2011 © ASCE 2011 4567
 
 

5.7% and 6.6% respectively. A linear regression that includes data from both
mixtures is shown in Figure 2a (solid line) with a coefficient of determination of 0.70.
Figure 2b re-plots the durability results with air voids as the independent
variable; data from the two mixtures is interspersed and separate trends are no longer
readily apparent. A regression containing data from both mixtures is shown in Figure
2b (solid line); the coefficient of determination is improved from 0.70 to 0.86
compared to the regression in Figure 2a using all data and asphalt content as the
Downloaded from ascelibrary.org by Universidad Militar Nueva Granada on 05/09/23. Copyright ASCE. For personal use only; all rights reserved.

independent variable. Based on these limited results, use of air voids appears more
promising than asphalt content as an indicator of durability results and may allow for
better comparisons of mixtures with widely varied design asphalt contents. However,
for the conditions tested the two terms are highly interrelated and further investigation
is needed to better quantify the results.

30.0

Mix 2
25.0 y = -7.036x + 56.7 R² = 0.84

20.0
Mix 1 and 2
y = -6.283x + 50.6 R² = 0.70
ML

15.0

10.0
9.5 mm Mix 1
5.0 12.5 mm Mix 2
Mix 1
y = -8.484x + 61.4 R² = 0.76
0.0
4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0
AC
a) Mass loss as a function of asphalt content

30.0
Mix 2
25.0 y = 2.176x + 1.52 R² = 0.89

20.0
Mix 1
ML

y = 1.762x + 3.86 R² = 0.78

15.0

10.0

5.0 9.5 mm Mix 1

Mix 1 and 2 12.5 mm Mix 2

y = 2.072x + 2.22 R² = 0.86
0.0
0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0
Va
b) Mass loss as a function of air voids

Figure 2. Durability Results of 0% RAP Mixtures at Varying Asphalt Contents

Geo-Frontiers 2011
 Geo-Frontiers 2011 © ASCE 2011 4568
 
 

Comparison of Mixtures at Design Asphalt Content

Specimens of all eight mixtures prepared at their respective design asphalt

contents (Pb) were tested and results are summarized in Table 2. Figure 3a presents
a boxplot of all the data. Figure 3b presents the result of an analysis of means for the
data. There is little practical difference in the mean mass loss results for Mixtures 1
Downloaded from ascelibrary.org by Universidad Militar Nueva Granada on 05/09/23. Copyright ASCE. For personal use only; all rights reserved.

to 6. Mean mass loss of plant-mixed (QC) specimens is somewhat lower than that for
specimens prepared according to ordinary laboratory methods or specimens prepared
from plant-mixed field-sampled mixture. Mixture 7 is somewhat higher than for the
other 9.5 mm and 12.5 mm NMAS surface mixtures; however note that for all unaged
(other than short term oven aging during specimen preparation) surface mixtures, the
mean mass loss is no greater than 15%. Mass loss for the 19.0 mm NMAS binder
mixture is significantly higher than for the surface mixtures; this result is supported
by Alvarez et al. (2008).

Table 2. Summary of Durability Data for Mixtures with Design Asphalt Content
Mix Preparation Va Results (%) ML Results (%)
1 2
ID Method n Avg. Std. dev. Avg. Std. dev.2
1 1 4 5.2 0.6 12.3 1.4
2 1 3 3.7 0.3 8.3 1.3
3 3 8 3.0 0.6 7.2 1.0
4 3 23 4.3 1.2 7.9 1.2
5a 1 4 5.5 0.4 9.9 0.8
5b 23 5 7.5 0.2 12.1 1.8
5c 1 5 6.0 0.4 11.0 0.9
6 23 4 5.3 0.1 11.8 0.7
7 1 4 5.3 0.5 14.9 0.8
8 1 4 5.7 0.1 21.4 3.2
1) Number of specimens tested.
2) Sample standard deviation computed with small sample statistical theory (i.e. n-1).
3) Field sampled mixture was returned to laboratory in insulated containers for immediate
compaction.

A two-tailed pooled t-test was performed at the 5% significance level to

compare Mixture 5a to Mixture 5c (PG 76-22 compared to PG 67-22 with same
gradation, asphalt content and compactive effort). The difference was not significant
but the small sample does not allow for any conclusions to be drawn. A one-tailed
pooled t-test for the same data does indicate a significant difference exists at the 5%
significance level. A two-tailed pooled t-test was performed at the 5% significance
level to compare the two specimen preparation methods used for Mixtures 5b and 5c.
The difference was not significant; the difference was also not significant for a one-
tailed pooled t-test. Pooled t-tests also did not indicate any significant difference
between the 9.5 mm and 12.5 mm NMAS gradations of Mixtures 5b and 6 with the
same preparation method and roughly the same proportions of aggregate types.

Geo-Frontiers 2011
 Geo-Frontiers 2011 © ASCE 2011 4569
 
 

30

25

20
Downloaded from ascelibrary.org by Universidad Militar Nueva Granada on 05/09/23. Copyright ASCE. For personal use only; all rights reserved.

ML

15

10

0
1 2 3 4 5
5a 6
5b 7
5c 86 97 10
8
Mixture
a) Boxplot of mass loss for all mixtures tested
30

α = 0.05
25
Significantly different
Not significantly different
20
Mean ML
ML

15
12.24
10 10.31
8.38

0
1 2 3 4 5
5a 6
5b 7
5c 86 97 10
8
Mixture
b) One-way normal analysis of means for mass loss

Figure 3. Statistical Analysis of Durability Data for all Mixtures Tested

Test Repeatability

The repeatability of the Cantabro test results was good; for the plant mixed
quality control of which the largest number of replicate specimens was available (23),
the standard deviation of the durability results (Figure 4a) was the same as the

Geo-Frontiers 2011
 Geo-Frontiers 2011 © ASCE 2011 4570
 
 

standard deviation for air voids of the same specimens (Figure 4b). Of note in Table
2, the standard deviation of ML results for Mixture 8 (19.0 mm NMAS binder mix)
was noticeably higher (s = 3.2) than for any other sets of specimens; however the
mean mass loss of those specimens was also much higher than for the surface
mixtures. With the exception of the aforementioned Mixture 8, the variability of
results from the Cantabro test does not appear to be excessive based on the limited
data set presented herein.
Downloaded from ascelibrary.org by Universidad Militar Nueva Granada on 05/09/23. Copyright ASCE. For personal use only; all rights reserved.

Avg. = 7.9
6 Std. Dev. = 1.2
n = 23

5
Frequency

0
6 6.5 7 7.5 8 8.5 9 9.5 10 10.5
ML

a) Frequency histogram of mass loss

7

Avg. = 4.3
Std. Dev. = 1.2
6
n = 23

5
Frequency

0
Less 3 3.5 4 4.5 5 5.5 6
Va

b) Frequency histogram of air voids

Figure 4. Frequency Histograms of Mixture 4 Data

Geo-Frontiers 2011
 Geo-Frontiers 2011 © ASCE 2011 4571
 
 

SUMMARY AND CONCLUSIONS

The durability of surface mixtures containing 25% RAP or more is a concern

due to the contribution of RAP bitumen to overall mixture stiffness. The Cantabro
durability test is being evaluated for use with dense-graded asphalt mixtures as a step
toward the goal of this research to evaluate the performance of high RAP content
surface mixtures. The research presented in this paper is a first step in evaluating use
Downloaded from ascelibrary.org by Universidad Militar Nueva Granada on 05/09/23. Copyright ASCE. For personal use only; all rights reserved.

of the Cantabro test for this purpose. For the Cantabro test to be useful for this
purpose it must be repeatable and to be capable of differentiating between different
mixtures. Based on the results from the limited dataset of mixtures with low or no
RAP presented in this paper, the Cantabro test is worth investigating further for
dense-graded mixtures. Other preliminary work with a 50% RAP mixture that was
not included in this paper indicated mass loss results on the order of 25% for un-aged
specimens.
To establish a range of durability results for current practice dense-graded
mixtures in Mississippi, a much larger set of data is being collected by the authors of
the next few months that will include specimens from a greater number of asphalt
plants, a wider variety of aggregate types, and increased replication. As the data is
collected linear regression or other appropriate statistical techniques will be utilized
to establish a criteria for comparison to high RAP surface mixtures that are being
designed and evaluated as part of the overall research.
Results of Cantabro durability testing of dense-graded asphalt mixtures was
presented for hot mixed asphalt mixtures encompassing a variety of design traffic
levels, binder grades, nominal maximum aggregate sizes and aggregate types present
in Mississippi. Gyratory compacted specimens were prepared or obtained according
to three different methods including standard laboratory mixing and compaction,
plant mixed field sampled mixture that was compacted immediately without
additional heating, and plant mixed quality control specimens. No significant
differences were found between results for specimens of the same mixture prepared in
the laboratory according to standard methods compared to specimens prepared from
plant-mixed field-sampled material. For the two laboratory designed 0% RAP
mixtures tested, air voids may be a better indicator of Cantabro results than asphalt
content alone; however a larger data set will be required to make any conclusions.
The mean mass loss of specimens was found to be no greater than 15% for the seven
un-aged surface mixtures tested. The variability of the Cantabro test for dense-graded
asphalt mixtures was found to be reasonable for the seven surface mixtures tested.
Overall, use of the Cantabro durability test to evaluate performance of high RAP
content surface mixtures compared to conventional low RAP surface mixture appears
promising.

ACKNOWLEDGEMENTS

The authors would like to thank Mr. Scott Glusenkamp and Mr. Jimmy Ivy of
APAC-Mississippi Inc. as well as Mr. Huey Stockstill Jr. of Huey Stockstill Inc. for
their assistance in identifying material sources and donation of materials for this
project. The authors would also like to thank Mr. Garrett Elliot of APAC-Mississippi

Geo-Frontiers 2011
 Geo-Frontiers 2011 © ASCE 2011 4572
 
 

Inc. for his assistance in obtaining plant quality control specimens. Finally the
authors would like to acknowledge the Mississippi Department of Transportation for
providing the funding of State Study 212 which made this research possible.

REFERENCES

Alvarez, A. E., Martin, A. E., Estakhri, C., and Izzo, R. (2008). “Evaluation of
Downloaded from ascelibrary.org by Universidad Militar Nueva Granada on 05/09/23. Copyright ASCE. For personal use only; all rights reserved.

Durability Tools for Porous Friction Courses.” Transportation Research

Board 87th Annual Meeting Compendium of Papers, (DVD-ROM), Paper No.
08-2370, TRB, Washington D.C.
Kraus, Z. R. (2008). The Morphology of Polymer Modified Asphalt and its
Relationship to Rheology and Durability. MS Thesis, Texas A&M University.
Mississippi Department of Transportation (2005). Materials Division Inspection,
Testing, and Certification Manual. Mississippi Department of Transportation.
Jackson, MS.
Watson, D. E., Cooley Jr., L. A., Moore, K. A., and Williams, K. (2004). “Laboratory
Performance Testing of Open-Graded Friction Course Mixtures,”
Transportation Research Record, (1891), 40-47.
Watson, D. E., Moore, K. A., Williams, K., and Cooley Jr., L. A. (2003).
“Refinement of New-Generation Open-Graded Friction Course Mix Design,”
Transportation Research Record, (1832), 78-85.

Geo-Frontiers 2011