1

2
3

TITLE:

STIFFNESS EVOLUTION OF GRANULAR MATERIALS STABILIZED

WITH FOAMED BITUMEN AND CEMENT

4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46

Authors:
Felipe Halles A., (* corresponding author)
Centro de Ingeniera e Investigacin Vial (CIIV)
Pontificia Universidad Catlica de Chile
Departamento de Ingeniera y Gestin de la Construccin. Escuela de Ingeniera.
Vicua Mackenna 4860, Edificio San Agustn 3 piso, Santiago - Chile
E.mail: fhalles@ing.puc.cl
Phone: 56-2-3544245; Fax: 56-2-3544806
Guillermo Thenoux Z.
Escuela de Ingeniera
Pontificia Universidad Catlica de Chile
Departamento de Ingeniera y Gestin de la Construccin. Escuela de Ingeniera.
Vicua Mackenna 4860, Edificio San Agustn 3 piso, Santiago - Chile
E.mail: gthenoux@ing.puc.cl
Phone: 56-2-3544245; Fax: 56-2-3544806
lvaro Gonzlez V.
Facultad de Ingeniera
Universidad del Desarrollo
Avenida La Plaza 680
Email: aagonzalez@ingenieros.udd.cl
Phone: 56-2-3279756; Fax: 56-2-9253

Transportation Research Board,

92nd Annual Meeting
January 13 - 17, 2013
Washington D.C.
Topic: Recycled Materials for Transportation Infrastructure

Submission date:
Length of text:
Number of figures and tables:
TOTAL

November 15th, 2012

4004 words
9 (2250 word equivalents)
6254 words/word equivalents

Halles F., Thenoux G. and Gonzalez A.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28

STIFFNESS EVOLUTION OF GRANULAR MATERIALS STABILIZED WITH

FOAMED BITUMEN AND CEMENT
Authors: Felipe Halles, Guillermo Thenoux and Alvaro Gonzlez

Abstract: From the literature it is possible to find two trends regarding the stiffness evolution of
the Foamed Bitumen Stabilized/Recycled mixtures. The first trend indicates that once the
foamed bitumen mix reaches a constant value due to curing process, the stiffness decreases with
time due to load cycles. The second trend indicates that stiffness remains constant after curing
process. In this research, the stiffness evolution of foamed bitumen mixes stabilized with
different bitumen and cement contents is studied. The stiffness was measured using the Indirect
Tensile Fatigue Test (ITFT). Results indicate that once the foamed bitumen mix reaches a
constant value due to curing process, the stiffness will decrease or keeps constant depending on
the stress-level applied to the foamed bitumen layer. If the stress-level is lower than a specific
value, the stiffness of the mix will remain constant at a value very close to the initial stiffness. If
the stress-level is greater than a specific value the stiffness of the mix will decrease gradually.
Also, the reduction rate of the stiffness will be greater as the stress-level is higher. The analysis
of results from mixes with different bitumen and cement contents allow identifying the effect of
both stabilizing agents in the long-term stiffness.

Keywords: Stiffness evolution, Stress-level, Foamed bitumen.

Halles F., Thenoux G. and Gonzalez A.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31

GENERAL DISCUSSION
Introduction
Stabilizing or recycling of asphalt concrete pavements using Foamed Bitumen (FB) is a widely
used rehabilitation technique worldwide. However, there is still no structural design
methodology fully validated by practitioners and researchers. One reason lies in the fact that it
has not been possible to determine the elastic modulus that FB stabilized layer develops with
time. Without a reliable modulus or stiffness value, the structural analysis of a FB pavement is
uncertain. Although there are several laboratory tests for assessing the stiffness of FB mixtures,
none of them allows identifying the observed stiffness evolution with time that has been
observed in the field.
There are two trends found in the literature regarding the in-situ stiffness evolution of the
FB mix. The first comes from studies carried out by Loizos (1) who analyzed back-calculated
stiffness from Falling Weight Deflectometer (FWD) tests from a FB recycled project built in
Greece (Figure 1). Results indicated that stiffness gradually increases from the day of
construction until it reaches a constant value after a period of approximately 12 months. A fact
that explains this behavior is the loss of moisture from the mixture that occurs during the curing
period. The loss of moisture and increase in strength or stiffness has also been observed by
Bowering (2), Jones et al. (3) and Fu et al. (4), among others.
After 12 months the backcalculated stiffness remains constant despite constant traffic of
heavy-weight vehicles (1).

32
33
34
35

FIGURE 1: Elastic modulus back-calculated by Loizos (1) for the FB layer.

36
37
38
39
40
41
42

Conversely, research studies carried out in South Africa (5) indicated that FB layers
affected to traffic load show a gradual decrease in stiffness. Figure 2 shows results of the
accelerated pavement testing performed in a recycled pavement with a FB layer using 1.8%
bitumen and 2.0% cement content. After pavement construction a 40 kN traffic load was applied
using the South African Heavy Vehicle Simulator HVS (6,7). Multidepth Deflectometer (MDD)
were installed in the pavement structure to measure deflection and to calculate the effective
stiffness of each layer.

Halles F., Thenoux G. and Gonzalez A.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38

FIGURE 2 Stiffness evolution of the FB layer Stiffness, adapted from (8).

Stiffness values measured during the period of analysis indicated that FB layer has two
phases. The first phase corresponds to a decrease in stiffness until a constant stiffness state,
without having a physical manifestation on the pavement layer (cracks or deformations). The
second phase is represented by a constant stiffness of the FB layer. During the test at
approximately 1,000,000 load cycles, the traffic was increased to 80 kN. With the 80 kN load,
the stiffness of the FB layer showed a similar trend to that observed during the first 300,000 load
cycles with 40 kN, with stiffness decreasing gradually until plateaus. Results provided by the
accelerated pavement test in the long term also indicated that during the second phase the
material behaves as a granular material, accumulating permanent deformation due to cyclic loads
applications. The TG2 first 2002 guide (8) proposed that the constant stiffness state can be
comparable to a granular material only in the effective elastic modulus and behavior, but not in
the physical composition of the materials.

Influence of the Active Filler in the Mechanical Properties of FB Mixes

Active fillers are normally added to FB mixes to improve the mechanical properties of the
mixture. While some researchers and practitioners have reported mixes without any active filler
(4,9), others have reported the use of cement (10-13), lime (14-16), fly-ash (17,18) and other
active fillers (18-20). Nevertheless, studies carried out in the last years have demonstrated that
stiffness of FB mixtures is dependent on both: type of active filler and content. Halles and
Thenoux (18) reported differences of up to two times between mixes with and without cement,
for the same FB content, measured using the repeated load triaxial test. The use of lime also
significantly improves the resilient modulus measured in the triaxial test (TxRM) of the FB mix.
Hence, in addition to determining the test that better characterizes the mechanical properties of
the FB mixture, it is necessary to quantify the relative contribution of the active filler used, in
particular type and percentage added to the FB mix.

Halles F., Thenoux G. and Gonzalez A.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32

Analysis of the In-Situ Stiffness FB Mixtures

Stiffness evolution observed in pavement structures using FB layers can be explained by
studying the microstructure of the FB mixture. The tiny bitumen particles as well as cement
particles disperse throughout the aggregate by adhering to the finer particles (fine sand and
smaller) forming a mastic that bonds the large aggregates. The bonds formed by the stabilizing
agents (bitumen and active filler) are not continuous throughout the microstructure of the
aggregates matrix; the physical bond may be described as spot welded (21). When traffic loads
are applied into the pavement structure, the FB layer is deformed producing strains and stresses
in the microstructure. These stresses could break some of the bonds upon depending on the
magnitude of the stresses, orientation and concentration, producing a decrease in cohesion and
therefore a decrease in stiffness. If the stresses or strains are smaller than the limiting stress of
the mastic, then the FB mix will maintain its original cohesion and therefore its stiffness.
It is well known that FB mixes can bear tensile stress, but also could be stated that they
are not able to withstand fatigue cracks. Most of the literature have indicated that failure mode
observed in the field is plastic deformation instead of fatigue cracking (8,22); i.e. when high
stresses are produced in the bottom of the layer, then bonds in that area are broken and the mix
continues behaving as an untreated material in that specific area.
Based on these observations it is possible to state that the different trends of the stiffness
evolution could be explained by the stress state of the FB layer. To support this assumption, a
simple multi layered linear elastic model was used to calculate stresses and strains in the FB
layer in the Greek and South African projects. Dual truck tires were modeled with two 20 kN
loads using a separated by 350 mm and contact pressure of 700 kPa. In both projects, stresses
and strains were calculated at one quarter of the thickness of the FB layer, measured from the
bottom of the FB layer. Table 1 shows details of the materials elastic properties and thicknesses
for each pavement structure evaluated as well as stresses and strains calculated at one quarter of
the FB layer thickness. The indirect tensile strength (ITS) of the FB mix was assumed to be 300
kPa in both projects, which is considered a representative ITS of a mix with a 1.0% cement
content.
TABLE 1 Pavement structure characteristics and stresses/strains found in projects from
Greece and South Africa
Layer

33
34
35
36
37
38
39

Elastic Modulus
(MPa)

Greece

South Africa

T (cm) (kPa) (m) SR T (cm) (kPa) (m) SR

4000
9
3
3
Surface Layer
60
45
0.2
165
100 0.55
1200
25
25
FB Layer
1000 (*)
10
----Other Layer
250
15
25
25
Subbase
90 (**)
------Subgrade
(*) Cemented treated layer, (**) Estimated by the authors, SR= Stress divided by the Indirect Tensile Strength, T =
Thickness of the layer, = stress, = strain

Based on the results obtained from the pavement structural analysis, it can be seen that
the FB layer of the pavement structure in Greece was subjected to a SR of 0.20 while the FB
layer of the pavement structure in South Africa to a SR of 0.55. Both values are very different
and could explain the behavior observed using the concepts for the thickness design of perpetual

Halles F., Thenoux G. and Gonzalez A.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46

pavements (23). The perpetual pavements concept states that if the deformations that occur at
critical points are below a specific value (endurance limit), then the asphalt layer will not suffer
fatigue and therefore its structural lifetime could be extended to periods of even 50 years. Under
this scenario, only functional maintenance actions should be conducted for keeping the standard
expected. Results of studies carried out to validate this affirmation have suggested that the
limiting tensile strain at the bottom of the asphalt concrete layers should be no greater than 60
and that at the top of the subgrade the vertical strain should be limited to 200.
Using a similar approach to FB layers, it may be assumed that if SR is lower than the
endurance limit, then the FB will maintain its stiffness.
Based on the above discussion, it is expected that after a period the FB stiffness plateaus,
which will depend of the stress/strain levels applied to the layer depending on the traffic loads
applied in the road, similarly to the behavior observed in the pavements analyzed in this article.
In the case of the pavement structure of Greece, it is probable that a SR of 20% is lower than the
SR required for generating some damage to the bonds or mastic produced by the bitumen and
cement particles.
This paper presents results of a research work carried out to represent and quantify the
stiffness evolution of FB mixes subjected to different SR levels. This information was used for
two purposes:

For establishing the effect of the different bitumen and cement content on the long term
performance of the FB mixtures, in order to determine the appropriate content of each
one.

Define the maximum stress level that must be accepted in the FB layer in order to
maintain constant its stiffness.

LABORATORY TESTING PROGRAM

Materials
Reclaimed Asphalt Pavement and Granular Materials (RAG) Properties
The RAP (Recycled Asphalt Pavement) used was collected from the Huequn - Los Sauces
recycling project, located south of Chile (IX Region). Only the RAP was pulverized by the
recycler machine, (without the addition of foamed bitumen or active fillers). The RAP was then
mixed with reclaimed granular materials in the laboratory, simulating the same conditions found
in recycled projects in Chile. The RAP and reclaimed granular materials are the Recycled
Asphalt Granular (RAG) material used in this study. One gradation was constituted from the
original material by sieving the RAP into three fractions and recombining them with reclaimed
granular base materials for laboratory testing. Also, inert baghouse dust collected from an asphalt
concrete plant was used for correcting the final RAG mix. The main properties of the RAG
materials were:

Maximum aggregate size:

Material Passing #200 sieve (0.075 mm):
Material Passing #40 sieve (0.425 mm):
Material Passing #4 sieve (4.75 mm):

19 mm
6%
13%
46%

Halles F., Thenoux G. and Gonzalez A.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43

Crushed/Fractured Particles:
Plasticity Index:
Optimum Moisture Content:
Maximum Density:

100%
Non Plastic
6.0%
2190 kg/m3

Stabilizing Agents Properties

Bitumen properties were: Absolute Viscosity at 60 C: 3210 poises, Ductility at 25 C higher
than 150 cm at 5 cm/min, Penetration Index: -0.9 and the Flash Point: 360 C using the
Cleveland Open Cup. Also, Viscosity at 60 C on residual bitumen from TFOT: 9180 poises and
ductility at 25 C higher than 150 cm at 5 cm/min. Portland cement type II was used in the
preparation of all the mixes.
Mixing Preparation, Compaction and Curing
The RAG material was foamed using the Wirtgen WLB-10 laboratory at 165 C with 2.5% of
foaming water by mass. The expansion ratio was approximately between 12 and 15 and the halflife was 10 to 12 seconds. RAG was preconditioned placing buckets with 20 Kg of material at
uncontrolled laboratory temperature (between 18 C and 25 C) and relative humidity (between
35% and 55%).
Previously to mixing, the optimum moisture content (OMC) of the RAG material was
strictly controlled and in average was near to 75% of the optimum moisture content (OMC) of
the RAG mix. Several Proctor tests were conducted to define the OMC value.
RAG materials and stabilizing agents were mixed using a twin shaft pugmill mixer.
During mix production, RAG materials were mixed in dry with the active filler, followed by the
addition of water. After one minute of mixing, FB was injected while the RAG material was
being agitated.
150 mm diameter and 60 mm in high ITS specimens were prepared and compacted using
gyratory compaction. The compaction procedure was adjusted to obtain the same density in each
specimen. Specimens were extruded from the molds immediately after compaction and cured in
a forced air oven at 40 C during 72 hrs.
Laboratory Tests
The stress controlled Indirect Tensile Fatigue test (ITFT) was used to evaluate the stiffness
evolution of the FB mixes. This test uses the same setup of the Indirect Tensile Strength (ITS)
test with cyclic loads. LVDTs were installed diametrically in order to measure the horizontal
tensile strains (Figure 3), which is used to calculate the dynamic/elastic modulus, as well as the
fatigue performance of the mix. The stress level applied in the ITFT was represented using the
Stress-Ratio concept (SR) which is defined as the quotient between the tensile stress (x) and the
Indirect Tensile Strength (ITS). The strains measured by the LVDTs were used for calculating
the elastic modulus/stiffness (S) of the mix. The following equations were used:

44
45

(1)
Where; P is the vertical load, t is the thickness of the specimen and d is the diameter.

Halles F., Thenoux G. and Gonzalez A.

1
2

(2)

3
4
5

Where; S is the elastic modulus, x the effective tensile stress and the measured tensile strain.

Where; x is the effective tensile stress and ITS is the maximum tensile stress.

7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43

(3)

Figure 3 shows how the stiffness of the mix decreases due to the progressive increase in plastic
and elastic deformations under cyclic loading.

FIGURE 3 Example of Indirect Tensile Fatigue Test.

Experimental Design
The laboratory test program using the ITFT was designed to study the impact of the bitumen and
cement contents. In addition, the ITFT results attempt to define the maximum stress-level that
the FB layer is capable to withstand without significantly reducing the stiffness.
Four stress levels, defined using the Stress-Ratio (SR), were applied to each specimen.
Five thousand load cycles were applied at each SR, for a total of 20,000 load cycles. ITFT were
carried out at 25 C using a temperature controlled cabinet. The experimental design for this
study is summarized in Table 2.
From the twelve mixes of the factorial design presented, only five were selected for
testing. Mixes FB1C1, FB2C1 and FB3C1were tested for evaluating the effect of the FB content
and mixes FB2C0, FB2C1 and FB2C2 were tested for evaluating the effect of the cement
content.

Halles F., Thenoux G. and Gonzalez A.

2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24

TABLE 2 Experimental design for stiffness evolution measurement.

Variable
# of levels
Values
RAG source
1
Bitumen grade
1
AC-24
Foaming properties
1
165C, 2.5% foaming water
Foamed bitumen content
4
1%, 2% and 3%
Cement content
3
0% - 1% - 2%
Compaction Effort
1
Gyratory compaction
Curing Period
1
72 hrs at 40 C
Water Conditioning
2
Dry
ITFT replicates
3
Test carried out at 25 C
Stress-State Levels
4
Stress Ratio equal to (Effective Stress / ITS)
Total number of tests (average)
100
From 9 (3 x 3) scenarios, only 5 were used
Mixes Evaluated in this Research Work
FB1C1
FB2C1(*)
FB3C1
FB2C0
FB2C1(*)
FB2C2
(*) same mixes/samples

FB Content (%)
1.0
2.0
3.0
2.0
2.0
2.0

Cement Content (%)

1.0
1.0
1.0
0.0
1.0
2.0

ANALYSIS OF RESULTS
Effect of the Foamed Bitumen Content on the Stiffness Evolution
Figure 4, shows results of the ITFT carried out on mixes FB1C1, FB2C1 and FB3C1 in order to
evaluate the effect of the bitumen content on the stiffness evolution. Results correspond to the
average of two specimens evaluated.
Trend lines were fitted and extrapolated to each mix for each SR with the objective of
estimating the stiffness for additional load cycles. Examples of the trend lines for FB2C1are
shown in Figure 4. Table 3, shows details of trend lines of each mix as well as the stiffness
expected at100,000 and 1,000,000 load cycles. In addition, the quotient between the Initial and
Long Term Stiffness (RSIE) is presented in the same Table 3.
It must be noted that a couple of test were carried out using more than 100,000 cycles in
order verify the extrapolation made.

Halles F., Thenoux G. and Gonzalez A.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22

FIGURE 4 ITFT Stiffness evolution of mixes with different bitumen content.

TABLE 3 Details of trend lines for each mix and estimated stiffness.
Long Term Stiffness and RSIE for i cycle loads
R2
Stress
y = a*xb
MixFB1C1
(1% FB+1%Cem)

MixFB2C1
(2% FB+1%Cem)

MixFB3C1
(3% FB+1%Cem)

Ratio

(%)

20
30
40
20
30
40
20
30

c
c
c
c
c
c
c
c
c

-0.013
-0.136
-0.755
-0.031
-0.140
-0.384
-0.019
-0.157

50.7
83.4
96.6
88.5
85.3
94.5
61.2
90.1

1098 (92%)
654 (55%)
131 (11%)
1310 (82%)
828 (52%)
407 (25%)
1220 (87%)
682 (49%)

1066 (89%)
478 (40%)
23 (2%)
1220 (76%)
600 (38%)
168 (11%)
1168 (83%)
475 (34%)

-0.581

96.3

228 (16%)

60 (4%)

40

Mix FB2C0

23
24

i = 100,000 (RSIE)

i = 1,000,000 (RSIE)

20
c
-0.031
70.2
849 (85%)
791 (79%)
(2% FB+0%Cem)
30
c
-0.261
95.1
341 (34%)
187 (19%)
40
Failure
Mix FB2C1
20
c
-0.031
88.5
1310 (82%)
1220 (76%)
(2% FB+1%Cem)
30
c
-0.140
85.3
828 (52%)
600 (37%)
40
c
-0.384
94.5
407 (25%)
168 (10%)
Mix FB2C2
20
c
-0.013
61.2
1538 (90%)
1493 (88%)
(2% FB+2%Cem)
30
c
-0.139
90.1
844 (50%)
613 (36%)
40
c
-0.576
96.3
324 (19%)
86 (5%)
y: Stiffness of the Indirect Tensile Fatigue/Modulus Test; x: Load cycles; c: Constant value not used in
the analysis.

Halles F., Thenoux G. and Gonzalez A.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33

34
35
36
37

10

Based on the results shown on Figure 4 and Table 3, the following observations can be made:

In general, all mixes show similar trends in terms of stiffness evolution at different stress
ratios. According to results presented in Table 3, curves were fitted to measured values.
The curves are described by a power relationship with a,b regression coefficients. Table 3
shows that b coefficients are very similar. For example, when a SR of 30% was
applied, b coefficient for mixes FB1C1, FB2C1 and FB3C1 were -0.136, -0.140 and
-0.157 respectively. When a SR of 20% and 40% were applied, a larger variability was
observed, but b values were within the same range.
The stiffness of the FB mixes plateaus after a certain number of load cycles when the
stress ratio is in the order of 20%. For example, Mix FB2C1 showed a relative constant
stiffness of 1200 1300 MPa for the long term, which represents almost 80% of the
initial Stiffness (1600 MPa). In the case of SR equal to 30%, the mix showed a good
evolution of stiffness, with a constant decreasing line with a low slope. In this case the
stiffness of the mix was 600 MPa after 1 million load cycles, which represents the 37.5%
of the initial stiffness (1600 MPa). This fact showed that after all the load cycles were
applied, the mix still is able to keep cohesion due the effect of the stabilizing agents.
Similar results are observed in mixes FB1C1 and FB3C1. In the case of a SR equal to
40%, mixes showed a clear reduction in stiffness with load cycles. All of them showed
almost null stiffness after 1 million cycle loads.
When mixes were subjected to a stress ratio of 50%, specimens collapsed in a relative
short period indicating that stresses and strains applied are much larger than the cohesion
provided by the FB and cement to the mix.
Although trends of the stiffness for each mix were very similar, results show that there is
an optimum bitumen content that maximizes the stiffness. In this case, the better results
in terms of stiffness evolution were observed in Mix FB2C1 with 2% FB and 1% cement.
Table 4 shows the slope of the trend line for each mix when SR is 50%.It is interesting to
note that when mixes were subjected to a SR of 50%, the stiffness rate of change for each
mix were very similar, which means that the FB content did not significantly affect the
behavior of the mix under these stress conditions.

TABLE 4 Slope of the trend lines for SR equal to 50%

Mix

Stress Ratio (%)

Mix FB1C1
Mix FB2C1
Mix FB3C1
Mix FB2C0
Mix FB2C1
Mix FB2C2

50
50
50
40(*)
50
50

(*) mixes FB2C0 collapsed for SR equal to 40%

Slope (m)
(y = m*x + b)
-0.1116
-0.1207
-0.1189
-0.0643
-0.1207
-0.2183

Halles F., Thenoux G. and Gonzalez A.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46

11

Effect of the Cement Content on the Stiffness Evolution

Figure 5, shows results of the ITFT carried out on mixes FB2C0, FB2C1 and FB2C2 in order to
evaluate the effect of the cement content on the Stiffness evolution. Results correspond to the
average of two specimens.
Trend lines were fitted and extrapolated to the modulus measured for each SR with the
aim of estimating the stiffness if additional load cycles were applied. Examples of the trend lines
for mix FB2C2 loaded at different SRs are depicted in Figure 5. Table 3 shows details of the
trend lines of each mix as well as the extrapolated stiffness at100,000 and 1,000,000 load cycles.
In addition, the RSIE was included in Table 3.

Figure 5 ITFT Stiffness evolution of mixes with different cement content.

Based on results of Figure 5 and Table 3, the following observations are made:

It can be observed that there are significant differences between mixes without cement
(Mix FB2C0) and with cement (Mixes FB2C1 and FB2C2) for the same bitumen content.
Mixes with cement (FB2C1 and FB2C2) have larger stiffness compared to Mix FB2C0
(with FB only). While stiffness of mixes FB2C1 and FB2C2 are 1310 MPa and 1538
MPa respectively for 100,000 cycles, stiffness for mix FB2C0is 849 MPa. These results
indicate that a pavement structure with FB2C1 and FB2C2 layers will have a better
structural capacity than using mix FB2C0.
Additionally, Mixes FB2C1 and FB2C2 were able to withstand more cycles than mix
FB2C0 for the same test loading sequence. While Mix FB2C0 collapsed for a SR equal to
40%, mixes FB2C1 and FB2C2 collapsed for a SR equal to 50%.

Halles F., Thenoux G. and Gonzalez A.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46

12

The stiffness of the FB mixes plateaus after a certain number of load cycles when the
stress ratio is in the order of 20%, providing guarantee that the mix will keep its cohesion
in the long-term.
For the case of a SR equal to 30%, mixes with cement (FB2C1 and FB2C2) showed an
acceptable stiffness evolution. It is interesting to note that although mix FB2C2 initially
has a greater stiffness than mix FB2C1, when extrapolated to 100,000 cycles, both
stiffnesses tend to be the same. Stiffness of Mix FB2C1 is equal to 828 MPa for 100,000
extrapolated cycles while stiffness of mix FB2C2 for the same conditions is 844 MPa.
For the case of a SR equal to 40%, mix FB2C1 showed a better stiffness evolution than
mix FB2C2. In this case the stiffness rate of change for mix FB2C2 was significantly
higher than mix FB2C1, which gives lower stiffnesses in the long term. When comparing
trend lines fitted to the data, it can be seen that values of the variable b are -0.384 for
mix FB2C1 and -0.576 for mix FB2C2. If the fitted equations are used to extrapolate the
stiffness, at 1 million load cycles the stiffness of mix FB2C1 is 168 MPa while stiffness
of mix FB2C2 is 86 MPa. In contrast, mix FB2C0 (only with FB) collapsed for a SR
equal to 40%.
For the case of a SR equal to 50% in mixes FB2C1 and FB2C2 as well as SR equal to
40% in mix FB2C0, trend lines were fitted to the data (see Table 4). The slopes for each
equation (m) indicated that as the cement content increases the rate of change of the
stiffness also increases. This means that when increasing the cement content, mixes with
the worst behavior are obtained for a SR of 50%.
Overall, the use of 2.0% cement in a FB layer will be beneficial only if the FB layer is
loaded to a SR equal or lower than 30%. If the FB layer is loaded to a SR equal or higher
than 40%, it is recommended only to use 1.0% cement together with FB.

CONCLUSIONS
An analysis of the stiffness degradation of foamed bitumen mixtures was done using the indirect
tensile fatigue test (ITFT) aiming to identify the evolution of the elastic modulus or stiffness of
the mixture in long term. The experimental study was designed to obtain the maximum stress
level that the FB layer is capable of withstanding without significantly reducing its stiffness as
well as to study the effect of the bitumen and cement contents in the long term stiffness.
Based on the results presented the main conclusions may be summarized as follow:

The stiffness of FB mixes will evolve according the stress-level applied to the FB layer.
If the stress level is lower than a specific value the stiffness of the mix will remain
constant at a value very close to the initial stiffness. If the stress level is greater than a
specific value the stiffness of the mix will gradually decrease.
If the Stress-Ratio of the mix is lower than 20%, then the mix will have a stiffness within
a range of 75% - 90% of the Initial Stiffness after one million load cycles. The value will
keep relatively constant during the life of the pavement.
If the Stress-Ratio of the FB mix layer is between 20% and 40%, the stiffness will
gradually decrease. The higher the Stress-Ratio, the higher will be the reduction rate of
stiffness.

Halles F., Thenoux G. and Gonzalez A.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46

13

If the Stress-Ratio is around 50% then the mix will collapse and the cohesion provided
by stabilizing agents will reduce to zero in a relative short period. In that case the
stiffness will be equivalent to the elastic/resilient modulus of the reclaimed material
without stabilizing agents.
Analysis of the bitumen content effect showed very little influence on the stiffness
evolution, but it was possible to find an optimum content that maximizes the stiffness; in
this case, Mix FB2C1 followed by mixes FB3C1 and FB1C1. Conversely, the effect of
cement content was significant on the stiffness evolution as well as in the absolute value
of stiffness. In addition, it is possible to state that cement must always be incorporated to
FB mixes for guaranteeing minimum short and long term stiffness.
Results showed that the use of 2% of cement in the FB layer will be beneficial only if the
FB layer is loaded to a SR equal or lower than 30%. If the FB layer is loaded to a SR
equal or higher than 40%, it is recommended to add only 1% of cement and FB.

The data and analysis provided in this research work can be used to estimate an effective elastic
modulus (EMM) of the FB mix in the long term based on the stress-state expected at the FB
layer. The EMM may be defined as a percentage of the initial stiffness and a shift factor must be
developed to adapt the elastic modulus provided by the ITFT to the elastic modulus that better
represents the mechanical properties of the FB mix in the field.
It must be noted that the laboratory work presented in this paper was carried out using
only one aggregate source, one bitumen source, a single active filler and a single temperature.
All these factors contribute significantly to the FB mix performance and therefore results
obtained are limited. Conclusions must be validated using a larger experimental design.
REFERENCES
1.

2.

3.

4.

5.
6.

7.

Loizos A. In-situ characterization of foamed bitumen treated layer mixes for heavy-duty
pavements. International Journal of Pavement Engineering, Vol. 8, N 2, June 2007, 123135.
Bowering R.H. Foamed bitumen: Production and application of mixtures, evaluation and
performance of pavements. In Proceedings of Association of Asphalt Paving
Technologists, Vol 45: 45373, 1976.
Jones D., Fu P., Harvey J.T., and Halles F. Full-Depth Reclamation with Foamed
Asphalt: Final Report. Davis and Berkeley, CA. University of California Pavement
Research Center. (Research Report UCPRC-RR-2008-07), 2008.
Fu P., Jones D., Harvey J.T. and Halles F. Investigation of the Curing Mechanism of
Foamed Asphalt Mixes Based on Micromechanics Principles. Journal of Materials in
Civil Engineering, Vol. 22, pp.29-38, 2010.
Long F. M. The development of structural design models for foamed bitumen treated
pavement layers. Contract Report CR-2001/76, Gauteng CSIR, South Africa, 2001.
Verhaeghe B, Sadzik E. and Visser A. Three Decades of Development and Achievements:
The Heavy Vehicle Simulator in Accelerated Pavement Testing. 10th International
Conference on Asphalt Pavements (ICAP), 2006.
Du Plessis L., Coetzee NF and Burmas N. Heavy vehicle simulator in accelerated
pavement testing: a historical overview and new developments. 3rd International
Conference on Accelerated Pavement Testing (APT), Madrid, Spain, October 2008.

Halles F., Thenoux G. and Gonzalez A.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45

8.
9.

10.

11.

12.

13.

14.

15.
16.
17.

18.

19.

20.

21.

22.

23.

14

Asphalt Academy Transportek. South African Interim Technical Guidelines. Design and
Use of Foamed Bitumen Treated Materials - TG2. CSIR Transportek, South Africa, 2002.
Sunarjono Sri. Tensile Strength and Stiffness Modulus of Foamed Asphalt Applied to a
Grading Representative of Indonesian Road Recycled Pavement Materials.
DinamikaTeknikSipil, Vol.7, pp 1 10, January 2007.
Thenoux G., Gonzalez A. and Jamet A. Aspectos Constructivos del Primer Proyecto de
Reciclado en Fro In-Situ con Asfalto Espumado en Chile. Revista Ingeniera en
Construccin Vol. 18, 2003, pp. 148-156, Pontificia Universidad Catlica de Chile.
Twagira M.E., Jenkins K.J. and Ebels L.J. Characterization of Fatigue Performance of
Selected Cold Bituminous Mixes. Proceedings of the 10th International Conference on
Asphalt Pavements, Quebec, Canada, 2006.
Jenkins K.J.; Long F.M. and Ebels L.J. Foamed Bitumen Mixes = Shear Performance?.
International Journal of Pavement Engineering, Volume 8, Issue 2, June 2007, pages 85
98.
Long F. M. and Theyse H.L. Mechanistic-Empirical Structural Design Models for
Foamed and Emulsified Bitumen Treated Materials. Conference on Asphalt Pavements
for Southern Africa CAPSA, 2004. South Africa.
Nataatmadja A. Some Characteristics of Foamed Bitumen Mixes. In Transportation
Research Record. Journal of the Transportation Research Board, N 1767, TRB, National
Research Council, Washington D.C., 2001, pp. 120-125.
Ramanujam J.M. and Jones J.D. Characterization of Foamed-Bitumen Stabilization.
International Journal of Pavement Engineering, Vol. 8, Issue 2, June 2007, pp. 111-122.
Leek C. Insitu Foamed Bitumen Stabilisation The City of Canning Experience.
Proceedings from 20th Australian Road Research Board Conference, Melbourne, 2001.
Saleh M. Effect of Rheology on the Bitumen Foamability and Mechanical Properties of
Foam Bitumen Stabilised Mixes. International Journal of Pavement Engineering, Volume
8, Issue 2, June 2007, pages 99 110.
Halles F. and Thenoux F. Degree of influence of active fillers on the properties of
recycled mixes with foamed asphalt. In Transportation Research Record 2095: 127-135,
2009.
Lancaster J., McArthur L. and Warwick R. Vicroads Experience with Foamed Bitumen
Stabilisation. Proceedings 17th Australian Road Research Board Conference, Volume 16,
Part 3, pp. 193-211, 1994.
Hodgkinson A. and Visser, A.T. The role of fillers and cementitious binders when
recycling with Foamed Bitumen or Bitumen Emulsion. Proceedings of the 8th Conference
on Asphalt Pavements for Southern Africa. September, 2004.
Jenkins, K.J. Mix design considerations for cold and half-cold bituminous mixes with
emphasis on foamed bitumen. Ph.D. Thesis. University of Stellenbosch, South Africa,
2000.
Gonzalez, A, Cubrinovski, M, Pidwerbesky, BD, and Alabaster, D 2009, Full scale
experiment on foam bitumen pavements in CAPTIF accelerated testing facility,
Transportation Research Record 2094, Transportation Research Board, Washington,
D.C., 21-29.
Committee on General Issues in Asphalt Technology. Perpetual Bituminous Pavements.
Circular Number 503. A2D05 Committee, Transportation Research Board, 2001.