Vous êtes sur la page 1sur 15

Evaluation of the Performance of an Asphalt Mix with Crushed Glass

Éric Lachance-Tremblay, ing.jr.

Masters Student
École de Technologie Supérieure
Montréal, Québec

Michel Vaillancourt, Ph.D., ing.

École de Technologie Supérieure
Montréal, Québec

Daniel Perraton, Ph.D., ing.

École de Technologie Supérieure
Montréal, Québec


Glass recycling is problematic because the glass that breaks during the collection process can’t be used to
make new glass products. The idea of using glass in asphalt mixes was recently studied at the Laboratoire
des Chaussées et Matériaux Bitumineux (LCMB) of the École de Technologie Supérieure (ÉTS).

The study was separated in two parts. The optimal proportion of glass was found according to the
Laboratoire des Chaussées (LC) of the Ministère des Transports du Québec (MTQ) formulation method.
The second part of the project was to compare the performance of a conventional asphalt mixture with a
asphalt mix containing glass. To do so, the complex modulus test and thermal stress restrained specimen
test were done, as well as a test to evaluate the stripping susceptibility.

The results show that it is possible to use glass in asphalt mix without decreasing performance. On the
other hand, it is not possible to simply replace a proportion of conventional aggregates with glass. Other
parameters such as the asphalt binder proportion must be adapted.


Le recyclage de verre est problématique puisque le verre qui se brise durant le processus de collecte ne
peut être réutilisé dans la fabrication de nouveaux produits de verre. L’idée d’incorporer du verre post-
consommation dans les enrobés bitumineux est reprise par le Laboratoire des Chaussées et Matériaux
Bitumineux (LCMB) de l’École de Technologie Supérieure (ETS).

L’étude réalisée s’est fait en deux étapes. Le dosage optimal en verre a été trouvé selon la méthode de
formulation du Laboratoire des Chaussées du Ministère des Transport du Québec. Par la suite, des essais
ont été réalisés afin de comparer les caractéristiques d’un enrobé bitumineux standard et d’un enrobé
bitumineux avec verre. Les essais de détermination du module complexe et de résistance au retrait
thermique empêché ont été réalisés ainsi que l’essai de tenue à l’eau afin de vérifier la durabilité des

Les résultats obtenus montrent qu’il est possible d’incorporer du verre post-consommation dans les
enrobés bitumineux sans trop affecter les performances. Par contre, il ne suffit pas de simplement
remplacer une partie des granulats conventionnels par due verre. Un travail d’optimisation doit être réalisé
afin de permettre l’utilisation de ce matériau dans les enrobés bitumineux.


Glass products can be recycled over and over again. It is possible to melt glass and reuse it to manufacture
new glass products. In order to do so, glass must be separated and classified according to the color and
type. This part of the recycling process is really important because all glass types have a specific chemical
composition. If different types of glass are mixed together to manufacture new products, the quality is not
guaranteed [1].

Glass recycling is problematic because all the glass that cannot be classified according to the type can’t be
used to manufacture new products. This category is called mixed glass and represents 59 percent of all the
glass collected [1]. Mixed glass can be reused in different applications such as sandblasting, but the
amount of mixed glass used is not sufficient considering the total quantity of glass collected.

The idea of using crushed waste glass into asphalt mix came up at the beginning of the 1970s [2]. Around
20 trial sections were completed between 1969 and 1971 in different locations of Canada and the United
States. Soon after that, the idea was abandoned due to the high cost of glass, even though the overall
performance was quite good [2]. During the 1980s, the cost of waste glass disposal was high. The idea of
using crushed waste glass into asphalt mixes was then brought back in order to save money. At least 45
trials sections were completed between 1969 and 1988 in Canada and the United States [3]. Even though
the idea of using glass into asphalt mixes has been studied before, this project is different in two ways.
First, none of the studies in the past used the Superpave Gyratory Compactor (SGC) mix design method.
The projects were all realized according to the Marshall method. Furthermore, pavement design methods
have evolved a lot since that time. The testing procedures used in this project were selected in order to
verify the thermo-mechanical properties of an asphalt mix with glass which has never been done before.

According to a report by the Connecticut Department of Transportation, the following items were
identified as potential problems: 1) loss of adhesion between bitumen and glass, 2) potential of losing skid
resistance, 3) breakage of crushed glass, 4) higher stripping sensibility and 5) potential of higher raveling
[4]. To prevent those negative effects, the glass content is limited to 15 percent for a surface course [5].
For a base course, the maximum glass content may go up to 25 percent because this part of the road is less
exposed to the aggressive agents such as water and cars [6].

In all cases, the dimension of crushed glass is really important to take into account. Some research
concluded that the use of small crushed glass (100 percent passing the 9.5mm sieve) can reduce the impact
the potential problems presented in the last paragraph [7].Others studies concluded that the crushed glass
must be smaller than 4.75 mm to prevent any problems regarding particle fragmentation in service and
asphalt performances [5, 6].

The low absorption value of glass particles might be an advantage in comparison with conventional
aggregate. It might be possible to reduce the amount of bitumen used in the asphalt mix without
compromising field application and overall performance. This would be an advantage from an economical
and environmental point of view. However, this low absorption is perhaps responsible to the high
stripping sensibility between bitumen and glass [6]. Using an anti-stripping agent decreases the stripping
sensibility of asphalt mix and increases the stiffness modulus [8]. When using glass in base course mix, it
is not mandatory to use anti-stripping agent as the mix is less water-exposed, but it is well recommended
when using glass in a surface course [9].

Many researchers have shown that the use of glass particles decrease the high temperature performance
(rutting resistance) of asphalt mixes due to the smooth texture of glass [6, 10, 11]. However, according to

Wu and Yang [6], the use of glass in asphalt increases the resistance to low temperature cracking. Also, Li
and Lu [10] showed that using glass in asphalt mix decrease the Marshall stability.


The main objective of this research was to verify the possibility to use crushed glass into asphalt mix
without compromising the overall performance. This project was done in order to find new alternatives for
recycling crushed glass. This project had two specific objectives:

1. Select an optimal proportion of crushed glass; and

2. Evaluate and compare the performance of asphalt mix with crushed glass with a conventional
asphalt mix.

Our experimental program was concentrated on the ESG14 asphalt mix. This mix is standardized by the
Québec Ministry of Transportation (MTQ) and commonly used in the province as surface and base course
in bituminous pavement structure.


3.1 Overview

The optimal proportion of crushed glass was selected according to the Laboratoire des Chaussées (LC) of
Ministère des Transports du Québec (MTQ) formulation method LC 26-004: Formulation des enrobés à
l’aide de la presse à cisaillement giratoire selon la méthode du laboratoire des chaussées. This formulation
method has specifications regarding effective volumetric binder contents (Vbe %) and voids content (Vi
percent). As it was reported in the literature review that crushed glass decrease the rutting resistance, it
was decided to do the rutting resistance test. The optimal proportion of crushed glass was selected
according to the following criteria: 1) effective volumetric binder (Vbe percent), 2) voids content (Vi
percent) and, 3) rutting resistance. The test methods are described in the following sections.

Once the optimal proportion of crushed glass was found, thermo-mechanical properties and durability
were evaluated and compared with the reference mix. The following three characteristics were studied: 1)
thermal behaviour at low temperature based on the Thermal Stress Restrained Specimen Test (TSRST), 2)
rheological behaviour in the small strain domain by measuring complex modulus (E*) and, 3) stripping
susceptibility. Those tests are described in the following sections.

3.2 Test Methods

3.2.1 Effective Volumetric Binder Contents, Voids Content and Rutting Resistance

The effective volumetric binder contents (Vbe) is the total bitumen content minus the volume of bitumen
absorbed by the aggregates. The LC formulation method specifies a Vbe for each type of asphalt mix. In
this project, all the mixes studied were designed in accordance to ESG14 requirement, with a Vbe of 11.4
percent with a tolerance of ± 0.1 percent. The Vbe is found using the maximum density test. For each mix,
a minimum of three tests were done according to the LC 26-045: Détermination de la densité maximale.
The LC formulation method also specifies voids content according to a number of gyrations for SGC test

which are: 1) 10 gyrations (Vi over 11.0 percent), 2) 100 gyrations (Vi between 4.0 and 7.0 percent), 3)
200 gyrations (Vi over 2.0 percent). The voids content were evaluated with the SGC in order to evaluate
the effect of crushed glass on compaction ability. For each mix, a minimum of three test were done
according to LC 26-003: Détermination de l’aptitude au compactage des enrobes à chaud à la presse à
cisaillement giratoire. The rutting resistance test was done according to LC 26-410: Résistance à la
deformation des enrobes à l’essai d’orniérage. The rutting depth must be inferior to 10.0 percent after
30,000 cycles. The dimensions of the samples tested are the following: 500 mm (length) x 180 mm
(width) x 100 mm (height) were tested using the MLPC rutting tester as shown Figure 1.

Figure 1. MLPC Rutting Tester (left) and Sample for Rutting Test (right)

3.2.2 Thermal Stress Restrained Specimen Test (TSRST)

The Thermal Stress Restrained Specimen Test (TSRST) gives a good idea of the performance of asphalt at
low temperatures. This test was done to verify the impact of glass particles on low temperature
performance. During testing, the sample is fixed on a hydraulic press system which restrains its
movement. The sample is then submitted to a constant cooling rate which causes shrinkage of the
specimen, but as the movement is restrained, tensile stress appears inside the specimen. The sample breaks
when the tensile stress inside the sample is greater than the tensile resistance of the material (Figure 2). In
this project, the cooling rate used was 10°C/hour and the dimension of the samples were 60 mm diameter
(± 5mm) and 250 mm length (± 5mm) according to the AASHTO TP10-93. For each mix, a total of three
samples were tested.

3.2.3 Complex Modulus (E*)

For an asphalt mix, the response to loading is delayed in time and changes in respect to the testing
temperature. This is due to the viscous behavior of this material. For an asphalt mix, the complex modulus
is characterized by the norm of the modulus, which is the stiffness of the material and currently designated
dynamic modulus, and the phase angle which is corresponding to the phase lag of the strain behind the
stress under sinusoidal cyclic loading.

a) TSRST b) E*

Figure 2. Thermal Stress Restrained Specimen Test (a) and Complex Modulus (E*) Apparatus (b)

Following the different temperatures and frequencies, the norm of the modulus and the phase lag angle
varies. The phase lag angle varies between 0° for a complete elastic materials and 90° for a complete
viscous material. The results can be plotted in a Cole-Cole diagram the X axis represents the storage
modulus (real part) due to the elastic behavior and the Y axis represents the loss modulus (imaginary part)
due to the viscous behavior. To see the variation of the phase lag angle, results can be plotted in the Black
diagram which shows the phase lag angle in the X axis and the norm of the modulus in the Y axis.

There are many different ways to determine the complex modulus. In this project, cylindrical specimens of
75mm in diameter and 120mm in length were submitted to sinusoidal cycles of traction-compression. The
target applied deformation was fixed to 50µm/m. This test was done using a MTS hydraulic press and an
environmental chamber.

The complex modulus apparatus is also shown in Figure 2. The different temperatures and frequencies
applied are presented in Table 1. For each mix, a total of two samples were tested.

Table 1. Parameters of the Complex Modulus Test

Temperature (°C) -35, -25, -15, -5, +5, +15, +25, +35

Frequency (Hz) 10, 3, 1, 0,3, 0,1, 0,03, 0,01


3.2.4 Stripping Susceptibility

This test was performed in order to verify the stripping susceptibility of asphalt mix with glass particles
according to LC 26-001 Tenue à l’eau. Using Marshall samples, the Marshall stability of the two groups
of specimens is compared. The first group of specimens is tested after 30 minutes in water at 60°C and the
other group is tested after 24h in water at 60°C. The residual stability is calculated by the ratio of Marshall
stability after 24h water exposure and Marshall stability of the first group. According to LC 26-001, the
residual stability must be over 70.0 percent.

3.3 Materials

In this project, only one type of asphalt mix was studied: ESG14 with PG 64-34 bitumen. Figure 4 shows
the granular curve of this mix (blue line). According to LC formulation method, each type of mix must
respect some grading criteria such as control points and restriction zone. In Figure 4, the black dots
represent the control points and red triangles the restriction zone. The granular curve must be between the
control points but no into the restriction zone.

Figure 3. Reference mix granular curve, controls points, restriction zone and fraction of aggregate
replaced by crushed glass.

The aggregates presented in Table 2 were used in various combinations to produce the different asphalt
mixes. Two different size of crushed glass were used as shown on Figure 3 (green boxes with dot line). A
total of five different glass contents (5, 10, 15, 20, and 25 percent content by mass) and one control mix
were studied. The proportion of crushed glass was chosen in order to have the same granular curve for all
mixes. Figure 4 shows the fraction were conventional aggregate were replaced by crushed glass. The
Table 3 shows the proportion of aggregates used for all mixes.

Table 2. Aggregate Specifications

Aggregate Dimension and Type Specific Gravity
Absorption (%)
10-14 mm limestone 2.725 0.53
5-10 mm limestone 2.721 0.58
0-5 mm washed limestone 2.711 0.71
0-5 mm limestone 2.718 0.65
0.630-2.5 mm glass 2.496 0.20
0-0.315 mm glass 2.532 0.07

Table 3. Proportion of Aggregates for Mixes Studied

Proportion of Different Aggregate (%)

Glass Aggregate Dimension (mm)
Content (%)
0-5 mm 0.630-2.5 0-0.315
10-14 mm 5-10 mm 0-5 mm
washed mm glass mm glass
Reference 20 29 23 28 0 0
5 20 29 18 28 5 0
10 20 29 13 28 8 2
15 20 35 5 25 10 5
20 20 35 0 25 15 5
25 20 39 0 16 17 8

a) 0 to 0.315mm b) 0.630 to 2.5mm

Figure 4. Different Sizes of Crushed Glass Used



4.1 Optimal Content of Crushed Glass Particles

The first part of the project started with finding the volume of bitumen absorbed (Vba percent) for each
mix. After that, optimal bitumen content was selected in order to obtain the desired Vbe. As we can see in
Figure 5, the use of glass particles in asphalt mix reduce the volume of absorbed bitumen (Vba) thus
reducing the optimal bitumen content for a given Vbe. This is true for all glass content except for the 20
percent. For this mix, the Vba is slightly lower than the reference mix, but the optimal bitumen content is
greater than the reference mix. This phenomenon is contrary to the trend observed and it has not been fully
studied in this project.

Figure 5. Volume of absorbed bitumen and optimal bitumen content regarding glass content.

After this first step, the voids content for each mix were measured with the SGC test except for the mix
with 5 percent crushed glass. This mix was taken out the project because of the small amount of crushed
glass used. The results are plotted in Figure 6. The two colored zones on this figure represent the voids
content specifications. The red line represents the minimal voids content after 200 gyrations. According to
the results, the use of glass particles in asphalt mix has a big impact on voids in the mix at 10 gyrations.
We can conclude that using crushed glass in asphalt mix increase the mix maneuverability and is easier to
put in place. Regarding the other criteria (100 and 200 gyrations), the use of glass particles does not seems
to have a particular effect. In accordance to SGC test, the mix with 10 and 20 percent are really similar to
the reference mix. The 15 and 25 percent voids contents are different than the reference mix but this is due
to the amount of percent passing sieve 0,080 mm which is different from the reference mix.

Figure 6. Voids Content of Asphalt Mixes

After the compaction ability test, two mixes were selected for rutting test. The mix with 10 and 20 percent
of crushed glass were chosen because their voids content for SGC test were valid according to the

Figure 7 shows the rutting test results performed on both mixes. In Figure 7, the red line indicates the
maximum rutting depth allowed after 30,000 cycles. We can see that the measured rut (percent) for both
mix with crushed glass are over specifications and so, after only 10,000 cycles. According to these results,
we can say that the use of crushed glass decreases the rutting resistance of asphalt mix.

With those results, it was decided to decrease the bitumen content in the 10 percent crushed glass mix.
This was done in order to get a good rutting resistance. Rutting results for this mix are indicated in Figure
7 with the dashed line. One possible reason that could explain poor rutting resistance of asphalt mix with
glass particles is the texture of glass. Glass particles have a smoother texture than conventional aggregates.
This characteristic decreases the internal friction between the glass particles and the aggregate, which
causes the mix to be more sensitive to rutting.

From this part of the project, it was decided to work with the asphalt mix with 10 percent of crushed glass
and lower bitumen content. This final mix has a bitumen content of 4.53 percent (Vbe of 10.66 percent)
compared to 4.85 percent (Vbe of 11.38 percent) for the original 10 percent crushed glass mix.

Figure 7. Rutting Depth of Asphalt Mixes

4.2 Performance Evaluation

The goal of this part of the project was to evaluate and compare the performance of the asphalt mix with
10 percent crushed glass with the reference mix. The three following characteristics were evaluated: 1)
TSRST, 2) complex modulus, 3) water stability. The results are presented in this section.

For the TSRST, three samples per mix were tested. During the test, the stress is recorded as well as the
temperature. Figure 8 shows the average results of the two mixes. As shown, the average failure stress as
well as the temperature is the same for both mixes. In both cases, the average failure temperature is very
close from the PG low temperature (-34) which is common for the TSRST. It is important to remember
that the crushed glass mix bitumen content is lower than the reference mix. Overall, we can say that using
crushed glass as well as a small diminution of bitumen content has no effect on the low temperature
performance of asphalt mix.

For the complex modulus test, two samples for each mix were tested. Samples were chosen in order to
have the similar voids content because it is known that these parameters have a big impact of complex
modulus test results. Complex modulus testing during this study has proven that the difference between
two samples for the same mix is negligible. For that reason, only one test result per mix will be presented
in this paper. The results in Cole-Cole diagram are plotted in Figure 9 and the Black diagram in Figure 10.
The voids content for the two test presented are 5.4 percent for reference mix and 5.0 percent for crushed
glass mix.

Figure 8. Thermal Stress Restrained Specimen Test (TSRST) Average Results

Figure 9. Complex Modulus Test Results Represented in the Cole-Cole Diagram


Figure 10. Complex modulus Test Results Represented in the Black Diagram

First of all, the results in Cole-Cole diagram (Figure 9) do not show any significant differences between
the two mixes.

Also, we can see that the results follow a unique curve for each mix. Figure 10 shows that the phase lag
angle decreases after reaching a maximum value. This behavior is observed when testing at high
temperature (+35°C). This is due to the bitumen because at high temperature the bitumen is really soft as
compared to the aggregate. This temperature depends on the bitumen type. We also see that the results do
not follow a unique curve, which means that the mixes do not conform to Time-Temperature
Superposition Principle [13]. This is a common behavior for polymer modified bitumen, which is the case
with the bitumen used in this project [12].

According to these results, it seems that adding crushed glass in asphalt as well as a small reduction of
bitumen content has no effect on the resulting stiffness.

For the stripping susceptibility test, a total of six samples per mix were tested (three without curing, three
with curing). It is important to take into account the variability of this test. According to ASTM D6927-
06, the coefficient of variation must not be greater than 6 percent and the acceptable range of two results is
16 percent of the mean. All the results presented in Table 4 respect those specifications.

If we look at the results, it can be seen that the reference mix has excellent water stability. For the crushed
glass mix, the residual stability is lower than the reference mix, thus being over the specification.

Table 4. Stripping Susceptibility Test Results

Average Average Average

Mix type Condition voids deformation Marshall
Stability (%)
content (%) (mm) stability (N)
No curing 2.2 4.8 15,441.7
Reference 99.3
With curing 2.5 5.2 15,337.5

10 % Crushed No curing 3.6 3.4 14,833.6

Glass With curing 4.7 4.7 13,011.2


The main objective of this study was to verify the possibility to use crushed glass in asphalt mix without
compromising the overall performances. The tests done were: maximal density, gyratory compactor,
rutting resistance, TSRST, complex modulus and stripping sensibility.

According to LC 26-004, two ESG14 mixes with 10 and 20 percent of crushed glass were prepared using
the Superpave gyratory compactor. But those mixes approved to have very poor rutting resistance. It was
then decided with the 10 percent mix to decrease the bitumen content to obtain acceptable rutting
resistance. The characteristics were then evaluated and compared with the reference mix. The use of
crushed glass as well as a small reduction of the bitumen content has no impact on the low temperature
performance (TSRST) and the complex modulus (E*). However, a small difference has been noted
regarding the stripping susceptibility; the crushed glass mix having a higher sensitivity than permitted in
the standards.

Overall, this project has shown that it is possible to use crushed glass in asphalt mix, but special attention
must be taken to the following points:

• Dimension of the crushed glass particles;

• Optimal bitumen content may decrease compared to a conventional mix; and

• Rutting resistance may decrease.


[1]. Gagné, L., "Le Verre : Fiches informatives", (2010)

[2]. Watson, T. "When the tire hits the glasphalt", Journal, Resource Recyling, 18-21 (1988.).

[3]. Dembickl, M. "Glasphalt Paves an Alternate Route". Journal, Waste Age Magazine, 23, 87-92

[4]. Kandhal, P.S., "Waste Materials in Hot Mix Asphalt: An Overview", National Center for Asphalt
Technology: Auburn University, Alabama, (1992).

[5]. Federal Highway Administration, U.S.D.o.T., "User Guidelines for Waste and Byproduct
Materials in Pavement Construction Waste Glass", (2012).

[6]. Wu, S., W. Yang, and Y. Xue, "Preparation and Properties of Glass-asphalt Concrete", Wuham
University of Technology: Key Laboratory for Silicate Materials Science and Engineering of
Ministry of Education, (2004).

[7]. Hughes, C.S., "Feasibility of Using Recycled Glass in Asphalt", Commonwealth of Virgina:
Richmond, (1990).

[8]. Arabani, M. and N. Kamboozia, "The linear visco-elastic behaviour of glasphalt mix under
dynamic loading conditions", Journal, Construction and Building Materials, 41, p. 594-601,

[9]. Day, D. and R. Schaffer, "Glasphalt Paving Handbook", University of Missouri-Rolla, (1989).

[10]. Li, K., et al., "Glasphalt Mixs' Performance Research and Analysis", in ICTE 2011, (2011).

[11]. Airey, G.D., A.C. Collop, and N.H. Thom, "Mechanical Performance of Asphalt Mixs
incorporation Slag and Glass Secondary Aggregates", in 8th Conference on Asphalt Pavements
for Southern Africa, Nothingham Centre for Pavement Engineering: Sun City South Africa,

[12]. Di Benedetto, H. and J.-F. Corté, "Matériaux routiers bitumineux 2 : constitution et propriétés
thermomécaniques des mélanges", Mécanique et Ingénierie des Matériaux. Vol. 1ère éd. (2005).

[13]. Di Benedetto, H., et al., "Linear viscoelastic behaviour of bituminous materials: From binders to
mixes", Journal, Road Materials and Pavement Design, 5, p. 163 - 202, (2004).