STIFFNESS MODULUS AND FATIGUE PROPERTIES OF CEMENT STABILIZED SAND

WITH USE OF A SYNTHETIC MODIFIED-ZEOLITE ADDITIVE

P. Wu
Department of Pavement Engineering, Delft University of Technology
Delft 2600GA, the Netherlands
Tel :+31 152784014, Fax: +31 152783443, E-mail: P.Wu@tudelft.nl

L.J.M. Houben, Associate professor

Department of Pavement Engineering, Delft University of Technology
Delft 2600GA, the Netherlands
Tel.:+31 152784917, Fax: +31 152783443, E-mail: l.j.m.houben@tudelft.nl

A. Scarpas, Professor
Department of Pavement Engineering, Delft University of Technology
Delft 2600GA, the Netherlands
Tel.:+31 152784017, Fax: +31 152787313, E-mail: A.Scarpas@tudelft.nl

C.E.G. Egyed
PowerCem Technologies B.V.
4782 SK Moerdijk, the Netherlands
Tel: +31 168409440, Fax : +31 168358600, E-mail: christophe@powercem.com

R. de La Roij
PowerCem Technologies B.V.
4782 SK Moerdijk, the Netherlands
Tel: +31 168409440, Fax : +31 168358600, E-mail: robin@powercem.com

2015 Annual Meeting of Transportation Research Board

January 11-15, 2015. Washington, D.C.
Word count: 4250+ (3 Tables+ 9 Figures) * 250 = 7250 words
Wu, Houben, Scarpas, Egyed, de La Roij 1

ABSTRACT

Cement stabilized material, obtained by blending cement, water and soil or aggregate, forms a durable
bonded material that can be attractive for pavement construction. However, cement stabilized material
is brittle and sensitive to overloading. This study investigates the engineering properties of using a
synthetic modified-zeolite additive in cement stabilized sand materials. The concept of applying a
non-traditional additive into the cement stabilization may provide potential economic and
environmental benefits for the road construction and maintenance. In this paper, stiffness and fatigue
properties were evaluated on a variety of mix proportions by varying cement and additive contents.
Stiffness and fatigue properties are two important parameters for pavement design. The higher
stiffness of the cement stabilized base enables the loads to distribute over a wider area and hence
considerably reduces stresses on the subgrade. Better fatigue resistance results in a higher ability to
prevent cracks caused by accumulated traffic loads. In this study, a cyclic four-point bending test was
conducted which simulates the cement stabilized base layer subjected to the repeated tensile strain and
stress. Based on the test results, adding certain amount of this additive in cement stabilized sand
mixtures can enable the beam specimen withstand higher strain at failure than the mixture without this
additive. The fatigue relations are obtained for all the tested mixtures by plotting the load cycles to
failure as a function of the applied stress or initial strain levels.
Wu, Houben, Scarpas, Egyed, de La Roij 2

INTRODUCTION
Cement stabilized material is generally defined as a mixture of soil or aggregates with addition of
measured amounts of cement and water compacted to a high density (1). As cement reacts with water,
the soil or aggregate particles are bonded by cement paste, resulting in a durable hardened material
with enhanced strength and stiffness as well as improved durability. Cement stabilized material is
primarily used as a base material underlying asphalt or concrete pavements (1). Traditional cement
stabilized materials have been extensively evaluated in literature studies and the specifications and
construction techniques are also well documented (2-4). However, the traditional cement stabilized
soil or aggregate is known as a material that exhibits brittle behavior such as sensitivity to overloading
and a low resistance to repetitive loading of high strain and stress levels (5; 6). Use of non-traditional
additive can be a potential alternative to reduce or eliminate the disadvantages of a cement stabilized
road base. This paper investigates one type of non-traditional additive, referred to as Rc, to evaluate
the potential engineering benefits of its application in cement stabilized materials. This additive is a
fine-grained powder, composed of synthetic zeolite, alkali earth metals, oxides complemented with
complex activators. Based on the previous academic studies (7-9), adding Rc additive in cement
stabilized materials can improve the mechanical properties and increase the flexibility of cement
stabilized materials. In construction of road base, Rc additive is typically applied by mixing it with in-
situ soil or crushed existing pavement materials prior to applying cement. In cementitious process, Rc
additive influences the pozzolanic reaction. This additive contains reactive Al2O3 and SiO2 which will
react with Ca(OH)2 (released from the cementitious reaction) and forms into additional cementitious
products (C-S-H and C-A-H hydrate gels). Besides, this fine-grained additive exhibits high specific
surface area so that adding Rc additive in the cementitious mixtures can increase the contact area of
cement with water and thus increases the degree of cement hydration.
The objective of this study is to evaluate the effect of this additive on the stiffness and fatigue
properties of cement stabilized sand materials. In the test program, a cyclic four-point bending test
was conducted to evaluate these two properties on 9 variable mix compositions. The stiffness was
investigated at various flexural strain levels (described as strain-sweep test) and thereby the stiffness
as a function of the strain level can be determined to indicate how the material resists the applied
strain. In fatigue test the beam specimen was subjected to various stress levels and the fatigue relation
for each mixture is obtained by means of plotting the number of load cycles to failure against the
applied stress level. During these two tests, a bending load with a fixed frequency was applied to a
beam specimen, simulating the manner in which a pavement layer is deflected by a wheel-loading (2)
and cyclic loading can simulate the repeated traffic loads.

MATERIALS AND TEST PROGRAM

Material properties and mix design

For the soil type, sand material is selected to be stabilized with cement. In the Netherlands, due to the
lack of natural aggregates, sand is an important material for road construction. This type of sand with
uniform grain size distribution (97% passing 2 mm sieve; 52.2% passing 0.425 mm sieve, 4.3%
passing 0.18 mm sieve) is classified as SP (poorly-graded sand) according to the USCS classification
system. Portland blast furnace slag cement (CEM III/B 42.5 N LH, containing 70% blast furnace slag
and 30% Portland Clinker) is chosen as the binder. This type of cement is frequently used in the
Netherlands for its low costs and chemical resistance.
The mix design in this study concentrates on the variable factors of cement and Rc additive
contents. The central composite design is employed to indicate these two factors. This mix design
method is considered as a useful tool to evaluate the correlations between parameters of the mixtures
and the tested properties and is also able to reduce the number of trials to achieve a balance among the
mix variables, which has been used in many studies (10-12). Table 1 shows the mix design and the
coded terms for each level. Trial 1~4 are called star point representing the independent variables (1
Wu, Houben, Scarpas, Egyed, de La Roij 3

and +1). Trial 5-8 are called corner points and the application level is denoted as (set as 2 ). The
correlation between each point is based on the equations (1) and (2). The mix proportions are
expressed as the absolute values, related to the dry mass of the sand. Rc and C represent Rc additive
and cement, respectively.
TABLE 1 Coded terms for mix variables

Mix Rc additive Cement Mixture (Rc, C),

Trial kg/m3
code Rc C
1 Rc-1C-1 -1 -1 (0.56, 159)
2 Rc-1C+1 -1 +1 (0.56, 251)
3 Rc+1C-1 +1 -1 (3.2, 159)
4 Rc+1C+1 +1 +1 (3.2, 251)
5 Rc-aC0 - 0 (0, 205)
6 Rc+aC0 + 0 (3.8, 205)
7 Rc0C-a 0 - (1.9, 140)
8 Rc0C+a 0 + (1.9, 270)
9 Rc0C0 0 0 (1.9, 205)

2 +1 0 + (1)

2 +1 0 + (2)

Where, +1 0 represents the scaling factor of cement content; +1 0 represents the scaling
factor of Rc additive content; + and are the maximum and minimum cement content (kg/m3);
3
and are the maximum and minimum Rc additive content (kg/m ).

As shown in Table 1 the cement content ranges from 140 to 270 kg/m3. The cement
proportion is in a range from 7.9% to 15.1% of the dry mass of the sand. The amount of cement used
in this study is relatively high. That is because an excessive amount of cement may make cementitious
material brittle, so that the changes in the properties caused by using Rc additive can be expected to
be efficiently observed. The Rc content is in a range of 0 to 3.8 kg/m3, which is from 0 to 2.10% of
the cement content (by mass). From the standard Proctor test (according to ASTM D698), the
optimum moisture content 8% (by mass), obtained from the moisture content-dry density curve, was
used in preparing samples for all the mixtures. The constant moisture content chosen for all these
mixtures is aimed to concentrate on the factors of cement and Rc contents and compare the
contribution of these two factors in the mechanical properties.

Sample preparation and test methods

Prismatic specimens with size of 400 by 50 by 50 mm were prepared. During mixing, Rc additive was
added to sand and mixed thoroughly prior to applying cement. After mixing all the components
uniformly, the homogeneous mixture was filled in steel moulds and compacted to the maximum
proctor density. After compaction and hardening for 24 hours, all the specimens were demoulded and
cured in an environmental chamber with a temperature of 202oC and a relative humidity of 90% for
28 days.
The stiffness modulus was measured under an increasing strain level from 50, 60 until 400
m/m at a fixed frequency of 10 Hz, which is described as strain-sweep test. In a strain-controlled
mode, 200 load cycles were applied at each strain level. The fatigue test was performed by subjecting
Wu, Houben, Scarpas, Egyed, de La Roij 4

prismatic specimen to various stress levels at fixed frequency 10 Hz and at each applied stress level
the stress-controlled mode was applied and hence the amplitude of the applied load was held constant
during testing. For each mixture, at least three stress levels (less than the flexural strength obtained
from the monotonic flexural test which is not included in this paper) were applied and 3 replicate
specimens were tested at each stress level. The resulting number of load repetitions till failure of the
specimen is determined. A four-point bending fatigue machine was used for these two tests, shown in
Figure 1. The fatigue machine was placed in an environmental-controlled chamber and the
temperature was constantly 20C, which ensures the consistency of the testing conditions for all the
mixtures .

FIGURE 1 Four-point bending test setup

ANALYSIS OF LABORATORY RESULTS

Stiffness modulus of cement stabilized sand

In the strain-sweep test, at each strain level 200 cyclic loadings were applied and the stiffness
modulus was taken as the average value of the last 10 cycles. The stiffness is plotted as a function of
the applied flexural strain levels.

Stiffness development under various applied strain levels

Figure 2 presents the plot of stiffness against various applied strain levels, based on the mixtures with
variable Rc additive contents. Each curve is the mean value of three tested samples. Figure 2
demonstrates that the stiffness modulus of cement stabilized sand is strongly influenced by the applied
strain. When the applied strain increases, the stiffness modulus significantly decreases until failure.
For the cement content 205 kg/m3, the stiffness modulus of the mixture with Rc 1.9 kg/m3 steadily
decreases with increasing strain level and at the end of the test the material didnt suddenly break. In
contrast, the mixtures without Rc sharply decreases at a rather low strain around 140 m/m, which
indicates that failure cracking occurred. And for the mixture with Rc content 3.8 kg/m3, in which the
ratio of Rc with cement content is relatively high, the stiffness also sharply decreased. In the second
graph it can be clearly seen that as the Rc content increases from 0.56 to 3.2 kg/m3, the mixtures show
much more flexibility and failed at much higher strain levels. These results imply that adding Rc in
cement-sand mixtures enables the specimens to withstand higher strain levels at failure. It might
indicate that in the pavement structure, when the high flexural strain is induced in base layer due to
the heavy loads, by using Rc additive could potentially reduce the chance of cracking.
Wu, Houben, Scarpas, Egyed, de La Roij 5

20000 (0;205)

Stiffnessmodulus(MPa)
(1.9;205)
15000 (3.8;205)

10000

5000

0
50 100 150 200 250 300 350
Strain(m/m)
20000 (0.56;251)
Stiffnessmodulus(MPa)

(3.2;251)
15000 (0.56;159)
(3.2;159)
10000

5000

0
50 100 150 200 250 300 350
Strain(m/m)
FIGURE 2 Influence of Rc contents on the stiffness development.

Regarding the influence of the cement content on the stiffness development, Figure 3 shows the
stiffness modulus of three mixtures with different cement contents and the same Rc content 1.9 kg/m3.
Again, each curve is derived from the average value of three replicate specimens.
20000
(1.9;140)
Stiffnessmodulus(MPa)

16000 (1.9;205)

12000 (1.9;270)

8000

4000

0
50 100 150 200 250 300 350 400
Strain(m/m)
FIGURE 3 Influence of cement content on the stiffness development.
As shown in Figure 3, when the cement content increases to 270 kg/m3 or decreases to 140 kg/m3, the
stiffness of these two mixtures both reduce significantly at a relatively low strain around 150 m/m.
That means the mixture with relatively low amount of cement is not able to withstand high flexural
strain while the excessive amount of cement may make the materials become rather brittle. This
implies that there is an optimum mixture with appropriate amounts of cement and Rc additive which
can achieve high flexibility, under the optimum moisture content 8%. Therefore, herein it can be
roughly concluded that, to increase the strain at failure, the optimum cement content can be ranging
from 205 to 251 kg/m3. Meanwhile, based on these cement content levels, the optimum Rc content
seems to be at 10.2%, related to the mass of cement.
 Wu, Houben, Scarpas, Egyed, de La Roij 6

Influence of mix variables on the stiffness modulus

As discussed above, the stiffness modulus of sand-cement may decrease under increased strain levels
which means micro-cracking will occur as the applied strain exceeds a certain limit or the fatigue
damage may happen under repeated loadings. Thus, the initial stiffness which is obtained at the first
strain level 50 m/m, is adopted to evaluate the influence of mix variables on this parameter. Table 2
summarizes the initial stiffness of all the tested mixtures. For each mixture, results of three test
repetitions are given.
TABLE 2 Test data of initial stiffness for all the tested mixtures
Mix design Mix design
Sample Density Stiffness Sample Density Stiffness
Rc Cement Rc Cement
code code
kg/m3 kg/m3 kg/m3 MPa kg/m3 kg/m3 kg/m3 MPa
Rc0C0-1 1.9 205 2021 16944 Rc0Ca-1 1.9 270 2055 19939
Rc0C0-2 1.9 205 1999 16271 Rc0Ca-2 1.9 270 2046 18936
Rc0C0-3 1.9 205 2030 17513 Rc0Ca-3 1.9 270 2047 19810
Rc0C-a-1 1.9 140 1924 11938 Rc-1C1-1 0.56 251 2014 19627
Rc0C-a-2 1.9 140 1954 12600 Rc-1C1-2 0.56 251 2017 19063
Rc0C-a-3 1.9 140 -- -- Rc-1C1-3 0.56 251 1987 17538
Rc1C-1-1 3.2 159 1971 12860 Rc-aC0 -1 0 205 1963 13131
Rc1C-1-2 3.2 159 1947 12021 Rc-aC0 -2 0 205 1978 15414
Rc1C-1-3 3.2 159 1929 11816 Rc-aC0 -3 0 205 1957 14250
RcaC0 -1 3.8 205 1945 16540 Rc-1C-1-1 0.56 159 1940 13400
RcaC0 -2 3.8 205 1918 14876 Rc-1C-1-2 0.56 159 1977 14964
RcaC0 -3 3.8 205 1955 17005 Rc-1C-1-3 0.56 159 1924 12154
Rc1C1-1 3.2 251 2033 18659
Rc1C1-2 3.2 251 1990 16906
Rc1C1-3 3.2 251 2014 18581
The initial stiffness modulus values of all the mixtures are compared in Figure 4 which illustrates the
average stiffness value together with the standard deviation indicating the variation of test results.

FIGURE 4 Stiffness of all the tested mixtures.

In Figure 4 it can be observed that a higher cement content generally yields a higher initial stiffness
modulus. On the other hand, the effect of Rc content on the stiffness appears to be influenced by the
cement content. For instance, in the mixtures with cement content 159 and 251 kg/m3, the lower
amount of Rc content contributes to higher stiffness, while for the mixtures with cement content 205
Wu, Houben, Scarpas, Egyed, de La Roij 7

kg/m3, the higher Rc content generates higher stiffness. This can be seen in the mixtures with the
same cement content and different Rc contents (marked in red). In order to quantify the effect of these
two factors on the stiffness, estimation models are developed for the 28-day stiffness modulus of
cement stabilized sand materials, shown in Figure 5. Figure 5 illustrates the estimated data derived
from statistical models and the real test data, including all the mixtures. Two statistical models are
considered, with combined effect of cement and Rc contents as well as density of the specimen. The
cement content and density of the specimen are evaluated in a power function which is obtained by
observing the results in Table 2. The Rc content is evaluated in two functions: exponential and linear.
25000 25000

MeasureedStiffness(MPa)
Measuredstiffness(MPa)

20000 20000

15000 15000
y=x y=x
R=0.902 R=0.902
10000 10000

5000 5000
5000 10000 15000 20000 25000 5000 10000 15000 20000 25000
Estimatedstiffnessfromeq.3(MPa) Estimatedstiffnessfromeq.4(MPa)

(a) Exponential function (b) Linear function

FIGURE 5 Estimation models of 28-day stiffness.
As can be seen in Figure 5 these two estimation models exhibit the same R2 and seem to approximate
the test data quite well. That means these two models can be used for the laboratory mix design which
can aid in obtaining the optimum mixture. These two estimation models (equation 3 and 4) in Figure 5
are described as follows:
125.4 0.48 . .
R2=0.902 (3)
0.48 .
1.3 96.7 R2=0.902 (4)
Where, E is the stiffness modulus at 28 days and at applied flexural strain 50 m/m, MPa; C is the
cement content, kg/m3; Rc is the Rc additive content, kg/m3; D is the density of the specimen, kg/m3.

Fatigue property of cement stabilized sand

Fatigue test data

Fatigue life under cyclic loading is the appropriate type of failure to predict the structural lifetime of
the stabilized layer in the pavement structure (13). Figure 6 gives an example of a fatigue test result at
different stress levels which shows a typical plot of stiffness modulus ratio (defined as the ratio of the
occurring stiffness modulus to the initial value) versus the number of load repetitions. The initial
stiffness modulus is obtained as the average value during the 90th to 100th load application. As shown
in Figure 6, under repeated loading the stiffness modulus ratio gradually decreases while at the end of
the test the stiffness modulus rapidly decreases until fracture occurs. Because during the stress-
controlled loading mode a constant stress is applied and the stiffness modulus is decreasing, so in such
condition the strain of the specimen is always increasing until failure. As soon as the crack is initiated,
it will rapidly propagate through the specimen. The failure point is defined as the moment when the
Wu, Houben, Scarpas, Egyed, de La Roij 8

beam gets completely fractured. Besides, it can be seen that the higher the stress level that is applied,
the smaller the number of load repetitions the beam can withstand.
100

Percentageofintialstiffness
80

60
(%)
40

20 1500 kPa 1300 kPa 1000 kPa

0
1.0E+02 1.0E+03 1.0E+04 1.0E+05 1.0E+06
Loadrepetitions
FIGURE 6 Ratio of stiffness versus load cycles for mixture (1.9, 205).
In this paper, the fatigue line is plotted as a function of the applied stress level against the number of
load repetitions to failure. According to standard NEN-EN 12697-24, the fatigue life in relation to the
initial strain (strain amplitude at 100th cycle) is suggested. Hereafter, these two types of fatigue
models are both evaluated, described as follows:
N k1 2
orN k1 2
(5)
Where, N is the number of load cycles at each applied stress level; is the initial strain (m/m); is
the applied stress (kPa); k1 and k2 are experimentally determined coefficients.
As an example, Figure 7 shows the fatigue models for mixture (1.9, 205). In Figure 6 the
number of load cycles N and applied stress level and initial strain are all presented in a log-log
scale aiming to illustrate the slope k2 of the fatigue line.
10 10
log(N)=13.62log()+47 log(N)=10.47log()+25
Log(N),loadingcycles
Log(N),loadingclycles

8 R=0.9437 8 R=0.971

6 6

4 4

2 2

0 0
0 1 2 3 4 0 1 2 3 4
Log(),appliedstresslevel(kPa) Log(),initialstrain(m/m)
FIGURE 7 Fatigue relation of the tested mixture (1.9, 205).
As demonstrated in Figure 7, good correlations between the number of load cycles and the stress level
or the initial strain are established and these two types of model show similar R2 values. The fatigue
models for all the sand-cement mixtures are given in Table 3 through the coefficients log(k1) and k2..
In Table 3 it can be seen that the fatigue relation for mixture Rc-1C-1 (0.56, 159) shows very low R2
and thus it is excluded in the following analysis. It has to be noted that the R2 of fatigue model in
some mixtures is relatively low. Molenaar (14) attributed the reason of low R2 to the brittle nature of
cemented materials, which means a small change in the applied tensile strain or stress can have a very
large effect on the number of load repetitions to failure.
Wu, Houben, Scarpas, Egyed, de La Roij 9

TABLE 3 Parameters for fatigue relations of all the tested mixtures

Fatigue life, log(N)
Mix design
Mix Stress based relation Strain based relation
code Rc Cement log(k1) k2 R2 log(k1) k2 R2
[kg/m3] [kg/m3] [-] [-] [-] [-]
Rc0C0 1.9 205 47 13.62 0.94 25 10.47 0.97
Rc0C-a 1.9 140 43 12.45 0.93 22 8.92 0.32
Rc+1C-1 3.2 159 45 14.02 0.60 16 6.45 0.67
Rc+aC0 3.8 205 41 11.49 0.81 18 6.85 0.78
Rc+1C+1 3.2 251 55 15.84 0.97 30 13.01 0.86
Rc0C+a 1.9 270 53 15.18 0.93 28 12.01 0.89
Rc-1C+1 0.56 251 52 15.04 0.94 26 11.16 0.88
Rc-aC0 0 205 47 13.59 0.98 34 15.07 0.96
Rc-1C-1 0.56 159 10 2.29 0.10 6 1.51 0.11

Figure 8 graphically presents the fatigue relations given in Table 3, expressed by the load cycles to
failure as a function of the applied stress (or strain) levels, in log-log scale.
10 (1.9,140)
(3.2,159)
log(N),loadingcycles

8 (0,205)
(1.9,205)
(3.8,205)
6
(0.56,251)
(3.2,251)
4 (1.9,270)

0
0 0.5 1 1.5 2 2.5 3 3.5 4
log(),appliedstresslevel(kPa)
(a) Stress-based fatigue relations
10 (1.9,140)
(3.2,159)
(0,205)
log(N),loadingcycles

8
(1.9,205)
(3.8,205)
6 (0.56,251)
(3.2,251)
(1.9,270)
4

0
0 0.5 1 1.5 2 2.5 3 3.5 4
log(),initialstrain(m/m)
(b) Strain-based fatigue relations
FIGURE 8 Fatigue lines for the tested mixtures.
Wu, Houben, Scarpas, Egyed, de La Roij 10

In Figure 8 generally it can be seen that the fatigue lines are influenced by the mix compositions and
these two types of fatigue relations show different trends. For instance, the stress-based relations are
rather parallel between mixtures, in other words the variation of the slope (k2-value) of the fatigue
lines is relatively small. In contrast, the slope of the strain-based fatigue lines are relatively more
variable. However, generally there doesnt seem to exhibit clear or consistent influence of cement and
Rc contents. Moreover, as can be seen for the stress-based relations, the k2 value ranges from 12 to 16
indicating a large slope and rather brittle behavior of cement stabilized sand material, which can be
compared with k2 obtained for the asphalt concrete mixtures generally ranging from 3 to 6 in literature
(15).
Despite the variation of the k1 and k2 values, it is found that these two coefficients are well
correlated, shown in Figure 9.

18 18
15 15
12 12
K2

K2
9 9
y=0.27x+0.99 y=0.51x 2.13
6 R=0.92 6
R=0.99
3 3

0 0
0 20 40 60 0 20 40 60
Log(k1) Log(k1)

(a)Stressbasedfatiguerelations(b)Strainbasedfatiguerelations
FIGURE 9 Correlation between k1 and k2
Figure 9 includes all the k1 and k2 values in the fatigue relations of all the mixtures and k1 is indicated
by its log. It can be found that k1 and k2 are highly correlated in both types of fatigue models. k2 is
linearly correlated to log(k1) independent of the mix design. Although different mixtures alter the
values of k1 and k2, the correlation between k1 and k2 is consistent, given in the following equations:
2 0.27 log 1 0.99 R2=0.92 stress-based fatigue relations (6)

2 0.51 log 1 2.13 R2=0.99 strain-based fatigue relations (7)

Similarly, research (15) gave the k1 and k2 relation for asphalt concrete materials and
emphasized that the uniform relationship between k1 and k2 doesnt change in different mixtures
which may have a potential use in pavement design.

CONCLUSIONS

This paper investigates the effect of a synthetic modified-zeolite additive on the properties of cement
stabilized sand material. Stiffness and fatigue properties are evaluated on a variety of mix proportions
by varying the cement and additive contents. The principle findings reported in this paper are
concluded as follows:
The stiffness modulus of cement stabilized sand is strongly influenced by the applied strain.
As the strain increases, the stiffness modulus significantly decreases until failure. Adding
certain amount of Rc additive enables sand-cement material to withstand higher strain before
failure under cyclic loading.
Wu, Houben, Scarpas, Egyed, de La Roij 11

Estimation models are developed to predict the stiffness modulus of cement stabilized sand
materials, with combined effect of cement content, Rc content (ranging from 0 to 2.0% of
cement content, by mass) and density of the specimen. The obtained models show a good fit
for the test data.
Fatigue relations are obtained for all the mixtures and these relations are influenced by
different mixtures. The k2 value of the stress-based fatigue models ranges from 12 to 16
indicating a large slope and rather brittle behavior of cement stabilized sand materials.
This paper focuses on the mechanical properties, e.g., stiffness and fatigue properties. Since cement
stabilized base is susceptible to shrinkage cracking. Further research will be conducted to evaluate the
shrinkage behaviour of cement stabilized materials with Rc additive and develop a simplified method
to estimate the transverse cracking performance for cement stabilized sand materials which can be
beneficial for the pavement design.

REFERENCES

1. American Concrete Institute Committee 230. State-of-the-Art Report on Soil Cement. ACI
Materials Journal, Vol. 87, No. 4, pp. 395417, 1990.
2. Williams, R.I.T. Cement-treated pavements: Materials, Design and Construction. London:
Elsevier Applied Science Publishers LTD, 1986.
3. TRH 13. Cementitious Stabilizers in Road Construction South Africa. Pretoria, South Africa,
1986.
4. Little, D.N. Recommended Practice for Stabilization of Subgrade Soils and Base Materials.
NCHRP web-only document 144. Texas A&M University, Texas, 2009.
5. Molenaar. A.A.A. Lecture Notes CT 4860 Structural Pavement Design, Design of Flexible
Pavements. Delft University of Technology, the Netherlands, 2007.
6. Syed, I.M. and Scullion, T. Performance evaluation of recycled and stabilized bases in Texas.
In Transportation Research Record: Journal of the Transportation Research Board, 1757(1),
pp.14-21, 2001.
7. Wu, P., Houben, L.J.M., Egyed, C.E.G. and de La Roij, R. Modified cement binder for
cement stabilized materials and concrete pavements, 8th International DUT-Workshop on
Research and Innovations for Design of Sustainable and Durable Concrete Pavements, Prague,
Czech Republic, 2014.
8. Moloisane, R.Y. Evaluation of the long-term strength behaviour of unpaved roads stabilized
with non-traditional stabilizers. Master thesis, University of Pretoria, South Africa, 2009.
9. Birgisson, B and Egyed, C.E.G. New Nano crystalline structure leads to visco-elastic
behaviour of cement based materials, 2008.
10. Xuan, D.X. Cement Treated Recycled Crushed Concrete and Masonry Aggregates for
Pavements. PhD Thesis, Delft University of Technology, Delft, the Netherlands, 2012.
11. Robinson, G.K. Practical Strategies for Experimenting. Chichester; New York: Wiley, 2000.
12. Medani, T.O. Estimation of fatigue characteristics of asphaltic mixes using simple tests. PhD
Thesis, Delft University of Technology, the Netherlands, 2000.
13. Crockford, W.W. and Little, D.N. Tensile fracture and fatigue of cement-stabilized soil.
Journal of transportation engineering, 113(5), 520-537, 1987.
14. Molenaar, A.A.A., and Pu, B. Prediction of fatigue cracking in cement treated base courses.
Proceedings of 6th RILEM International Conference on Cracking in Pavements (pp. 191-199),
2008.
15. Ghuzlan, K.A. and Carpenter, S.H. Traditional fatigue analysis of asphalt concrete mixtures.
Urbana, 51, 61801, 2002.