Soil Dynamics and Earthquake Engineering 97 (2017) 133–142

Contents lists available at ScienceDirect

Soil Dynamics and Earthquake Engineering

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

Effect of silt on post-cyclic shear strength of sand MARK

⁎
Reza Noorzad , Milad Shakeri
Faculty of Civil Engineering, Babol University of Technology, Babol, Iran

A R T I C L E I N F O A BS T RAC T

Keywords: Adding non-plastic fines to sand can greatly change its behavior. Size difference between sand and silt particles
Silty sand is the main reason which causes this change. While post-cyclic and post-liquefaction behavior of clay and clean
Triaxial test sand has been widely studied, silty sand is wrongly considered to behave like clean sand and researches usually
Non-Plastic fines do not focus on it. In silty sand, through low cohesion, applying cyclic load can displace particles and result in
Cyclic loading
heterogeneity within the mixture. Even if liquefaction does not occur, rearrangement of soil particles can affect
Post-cyclic strength
monotonic ultimate strength. This study, with a series of post-cyclic monotonic triaxial tests, has shown that in
sand with a considerable amount of silt, cyclic loading can change the ultimate state strength. In sand with 15%
silt it decreases and in sand with 30% silt it increase the ultimate state strength. Changes are negligible in clean
sand or sand with 5% silt.

1. Introduction silt contents, under consolidated undrained condition (CU). Specimens

In geotechnical engineering, the majority of experimental testing on
frictional granular media involves clean sand. However, most sands in
nature contain a considerable amount of silt or clay. Only in recent
decades, the basic differences between behavior of clean sand and sand
with fines has been understood. These differences are mainly due to
disparity in size of sand grains and fine particles. Fine particles can
occupy pore space between sand grains, leading to separation of grains
and reduction in contact points, thereby resulting in the reduction in
shear strength. In addition, when fine part has a considerable plasticity,
cohesion can also be a factor in assessment of shear strength [1]. Thus
changes in the sand behavior depend mainly on arrangement of fine
particles and their plasticity. Applying a cyclic load can impress both of
these factors because, on one hand, it changes the arrangement of soil
particles and on the other hand, it can diminish strength of cohesive
soils. The post-cyclic behavior of cohesive soils, such as clay, clayey
sand and sandy clay, has already been studied [2–4]. However,
research on non-plastic soils has been predominantly focused on soils
with uniform particles like clean sand or silt [5–7]. In these types of
soil, cyclic loading can change the soil fabric and even soil strength in
some cases [8]. Undoubtedly, in soils such as silty sand, which are
composed of two types of material with a significant size difference, this
change would be more pronounced, as the displacement of particles
due to cyclic loading exacerbates the heterogeneity within the soil mass
and affects the soil structure.
In this study, 55 triaxial tests are performed on sand with different
are loaded in two stages; first they are loaded cyclically up to a certain
pore water pressure ratio (ru). This ru should not result in failure or
liquefaction. Then a monotonic compression load is applied to samples
in undrained condition and the influence of cyclic loading on undrained
post-cyclic behavior of silty sand with different silt contents is
evaluated.
2. Background
2.1. Post-cyclic strength of sand
Studies about post-cyclic behavior of sand usually has been limited
to clean sand. Initial studies on clean sand suggested that the type of
loading, whether monotonic or cyclic, has no effect on the steady state
strength of sand [9]. However, later studies on post-cyclic strength
revealed that undrained cyclic loading can change the arrangement of
sand grains. Changes in arrangement of very loose specimens can lead
to formation of weak, metastable pores in soil structure that can be
broken easily and result in structural collapse or flow failure. During
these change, shear strength decreases abruptly and soil undergoes
large deformations [10]. In specimens with higher densities, due to
constant specimen volume, local rearrangement increases the speci-
men tendency for dilatation. Low plasticity silt also behaves like loose
sand; after applying a cyclic load, the tendency for instability and flow
failure increases in this type of soil [5].
Investigating the post-liquefaction behavior of sand has also proven

⁎
Corresponding author.
E-mail addresses: rn0864@gmail.com, rnoorzad@nit.ac.ir (R. Noorzad), milad.shakeri@live.com (M. Shakeri).

http://dx.doi.org/10.1016/j.soildyn.2017.03.013
Received 23 April 2015; Received in revised form 9 February 2017; Accepted 9 March 2017
Available online 17 March 2017
0267-7261/ © 2017 Elsevier Ltd. All rights reserved.
R. Noorzad, M. Shakeri
Nomenclature
Fc fines content (%)
fc fines content in decimal
e void ratio
eg intergranular void ratio
e* equivalent granular void ratio
b an experimental value showing active fines
ru pore water pressure ratio
that strength and stiffness of soils are greatly reduced after liquefac-
tion. But as axial strain increases, sand grains gradually reach a stable
condition under the new arrangement. Hereafter specimen shows
dilative behavior, both strength and stiffness increase [7]. In this case,
the post-cyclic behavior of sand specimens significantly depends on
maximum strain and pore water pressure induced by cyclic loading
[11].
Generally, in uniform non-plastic soils, such as sand or silt, only the
generated pore water pressure and arrangement of grains can affect the
post-cyclic strength. Since soil arrangement depends on grain shape,
which is in turn a material parameter, both decrease [6] and increase
[11] in strength and stiffness after applying cyclic load have been
reported.
2.2. Silty sand
Role of fines in silty sand can be different, while at low silt contents
the role of fines in sustaining shear stress in soil structure is almost
negligible, at silt contents higher than 10%, they can have an active role
in soil structure. One of the first works that noticed active contribution
of fines in sand belongs to Pitman et al. [12]. To model or explain
behavior of silty sand, an idealized two-size particle packing can be
assumed. This model considers silty sand as a material with only two
grain sizes which are usually average grain size of sand and silt [13].
According to Fig. 1(a), when the silt content is very low, particles are
trapped in some voids between sand grains and are not involved in
sustaining shear stress; therefore they can be considered as voids [14].
In order to account for this phenomena, many studies redefined void
ratio for this type of soil as Intergranular Void Ratio [15]:
(e + fc )
eg =
(1−fc )
Where fc is the ratio of fines to total solids by weight in dry condition.
In this case, due to inactivity of fine grains, the shear strength of
Fig. 1. Microstructure of silty sand in: (a) low fines content (b) moderate fines content (c) high fines content.
Soil Dynamics and Earthquake Engineering 97 (2017) 133–142
CSR cyclic stress ratio
N number of cycles
CP confining pressure (kPa)
q deviatory Stress (kPa)
p’ mean Effective stress (kPa)
p'ss mean Effective stress at steady state (kPa)
εa axial strain (%)
Su ultimate state strength (kPa)
RFD resistance to further deformations
silty sand is close to that of reference sand. Indeed, increase in
confining pressure can put the sand grains into a denser packing and
consequently, a small proportion of fines can take on an active role in
soil structure, but this changes are negligible [15].
Fig. 1(b) shows when fines content increases, although the fines are
not enough to completely fill voids, some of them come into effect as
separator elements between sand grains and gradually reduce contact
force between these grains. For example, analyzing the micro-structure
of Firoozkooh silty sand has demonstrated that in sand with 15% fines,
about a quarter of fines are placed between the sand grains. These
particles form metastable sand-silt-sand contacts which can be broken
easily, thereby leading to a loss of shear strength [16]. Compared to
clean sand, confining pressure have more intense effects on steady state
strength in these silty sands; an increase in confining pressure can
noticeably increase the steady state strength [15].
As demonstrated in Fig. 1(c), with further increase in fines content,
voids are completely filled with this particles. In this case, adding more
fines results in dispersing of sand grains and their isolation in a matrix
of fines. In other words this is a threshold fines content beyond which,
silt particles constitute the main soil matrix and soil strength is closer
to that of the host silt instead of the host sand. For values beyond the
threshold fines content, amount of fines does not affect the strength
significantly, although a small increase in strength has been seen at
very high fines contents [16]. Higher pressure can compress sand
grains, potentially leading to reestablishing contacts between sand
particles, which can significantly increase the strength of soil [15].
When fines content increases, a part of fines participates in
sustaining shear stress. In this condition, soil density can be defined
as equivalent granular void ratio which equals to [15]:
e + (1−b ) fc
e* =
(1) 1−(1−b ) fc (2)
Where b is an experimental value representing a fraction of fines that
plays an active role in soil structure with a value between zero and one.
134
R. Noorzad, M. Shakeri Soil Dynamics and Earthquake Engineering 97 (2017) 133–142

Table 2
Testing program.

Test No. FC (%) ea e* CP (kPa) (kPa) CSR ru N

1 0 0.760 0.760 50 – – Monotonic

2 0 0.762 0.762 50 0.252 0.354 9
3 0 0.760 0.760 50 0.249 0.662 16
4 0 0.760 0.760 50 0.254 0.864 22
5 0 0.757 0.757 100 – – Monotonic
6 0 0.758 0.758 100 0.249 0.371 7
7 0 0.758 0.758 100 0.251 0.670 13
8 0 0.758 0.758 100 0.244 0.876 20
9 0 0.753 0.753 200 – – Monotonic
10 0 0.753 0.753 200 0.249 0.363 3
11 0 0.753 0.753 200 0.250 0.659 9
12 0 0.751 0.751 200 0.249 0.871 16
13 5 0.701 0.790 50 – – Monotonic
Fig. 2. Grain size distribution curves for the materials.
14 5 0.701 0.789 50 0.253 0.361 5
15 5 0.700 0.789 50 0.252 0.662 14
Table 1 16 5 0.701 0.789 50 0.256 0.869 20
Characteristics of Babolsar sand. 17 5 0.698 0.786 100 – – Monotonic
18 5 0.698 0.786 100 0.252 0.350 3
D50 (mm) D10 (mm) Cu Cc FC (%) Gs 19 5 0.697 0.785 100 0.255 0.660 10
20 5 0.697 0.786 100 0.246 0.875 17
0.195 0.135 1.8 1.3 2.43 2.75 21 5 0.694 0.782 200 – – Monotonic
22 5 0.692 0.780 200 0.248 0.367 3
23 5 0.693 0.781 200 0.248 0.672 6
24 5 0.692 0.780 200 0.245 0.882 12
25 15 0.604 0.860 50 – – Monotonic
26 15 0.603 0.858 50 0.248 0.359 4
27 15 0.602 0.857 50 0.250 0.680 5
28 15 0.602 0.856 50 0.253 0.871 7
29 15 0.597 0.851 100 – – Monotonic
30 15 0.599 0.853 100 0.248 0.376 3
31 15 0.595 0.849 100 0.248 0.658 4
32 15 0.598 0.853 100 0.251 0.870 6
33 15 0.594 0.848 200 – – Monotonic
34 15 0.593 0.846 200 0.262 0.343 3
35 15 0.593 0.846 200 0.244 0.667 4
36 15 0.589 0.841 200 0.242 0.851 5
37 30 0.576 0.993 50 – – Monotonic
38 30 0.576 0.993 50 0.148 0.369 5
39 30 0.577 0.994 50 0.151 0.659 14
40 30 0.577 0.994 50 0.149 0.871 20
41 30 0.564 0.978 100 – – Monotonic
Fig. 3. Minimum and maximum void ratio of mixtures. 42 30 0.567 0.982 100 0.150 0.371 4
43 30 0.562 0.976 100 0.147 0.672 12
44 30 0.565 0.979 100 0.151 0.881 19
45 30 0.557 0.969 200 – – Monotonic
46 30 0.554 0.965 200 0.152 0.372 4
47 30 0.558 0.970 200 0.150 0.660 9
48 30 0.556 0.967 200 0.151 0.879 17
For sand with low fines content in which the b value is zero, Eq. (2) 49 5 0.699 0.788 50 0.151 0.362 20
converts into Eq. (1). In various studies, b has been considered 50 15 0.591 0.844 200 0.146 0.321 13
constant in a range from 0.25 to 0.4 [14,17–19]. 51 15 0.596 0.850 100 0.150 0.350 17
Since amount of active fines increases as fines content increases, it 52 15 0.597 0.852 100 0.158 0.632 38
53 30 0.560 0.977 100 0.100 0.650 35
is more reasonable to consider the b value as a variable depending on
54 30 0.565 0.980 100 0.102 0.365 16
fines content and other parameters. For this purpose, the following 55 30 0.576 0.993 50 0.099 0.642 46
equation is given for b [20]:
a
After consolidation
⎡ ⎡ ⎛ f ⎞ ⎤⎤
⎢ ⎢ −m ⎜ f c ⎟ ⎥ ⎥
⎝ cth ⎠ ⎥ ⎥ ⎛ rfcth ⎞
r
⎢ ⎢
b = ⎢1 − exp ⎢ ⎥ ⎥ ×⎜ ⎟ 3. Materials and methods
⎢ k ⎝ fc ⎠
⎢ ⎥⎥
⎢⎣ ⎣ ⎦ ⎥⎦ (3) 3.1. Testing equipment and materials
-1
Where r= Rd = d/D (d and D are mean grain size for host silt and
A series of triaxial tests is conducted on different mixtures of
sand respectively), k=(1−r0.25), m is a proportion ratio equals to 2.5,
Babolsar sand and Firoozkooh silt by IPC Global® universal cyclic
fc is the fines content in decimal and fcth is threshold fines content.
triaxial system. Cylindrical specimens are tested with diameter and
As mentioned previously, the behavior of silty sand and clean sand
height of 52 and 104 mm, respectively. Detailed specifications of the
are different. This difference can be extended into many aspects. One of
apparatus has been presented by Noorzad and Fardad Amini [21].
these aspects which have not been studied yet is post-cyclic behavior.
Fig. 2 shows grain size distributions of tested materials. The dark-
From results of other studies about post-cyclic behavior of other soil
colored Babolsar sand is a natural poorly graded sand with rounded
mixtures, like clayey sand [3], it can be anticipated that silty soils would
grains, characteristics of which are given in Table 1. The Firoozkooh silt
be sensitive to cyclic loading.
that passes completely through 0.075 mm sieve is an artificial silt with

135
R. Noorzad, M. Shakeri Soil Dynamics and Earthquake Engineering 97 (2017) 133–142

Fig. 4. Failure shape: (a) in specimens with 30% fines; (b) in rest of specimens.

Fig. 6. Monotonic loading of specimens: (a) CP=200 kPa (Test No. 9, 21, 33, 45); (b)
only for sand with 30% silt (Test No. 37, 41, 45).

Fig. 5. Samples of cyclic loading: (a) without phase transformation; (b) with phase
silt by weight. Maximum and minimum void ratio of mixtures are
transformation.
separately obtained using the ASTM D4253 and D4254 standards
respectively (Fig. 3). Fig. 3 is used to guess the threshold silt content;
specific gravity of 2.70, plasticity index of 3 and liquid limit of 25. So it
the vertex of minimum void ratio in this graph is placed before 30% silt
can be considered as a non-plastic silt. Because of its red color,
content so threshold silt content has a value less than 30% and
Firoozkooh silt is conspicuously distinguishable from Babolsar sand
consequently, specimens with 30% silt are beyond the threshold silt
within the mixture.
content. Test results for sand with 30% silt support this claim, too.
Tests are performed on sand, mixed with 0%, 5%, 15% and 30% of

136
R. Noorzad, M. Shakeri Soil Dynamics and Earthquake Engineering 97 (2017) 133–142

Fig. 7. Post-cyclic loading for: (a) FC=0%, CP=200 kPa (Test No. 9–12); (b) FC=5%, CP=200 kPa (Test No. 21–24); (c) FC=15%, CP=100 kPa (Test No.29–32); (d) FC=30%, CP=100
kPa (Test No. 41–44).

Table 3 poured into a plexiglass tube with two rubber stoppers at the ends.
Silt content in different parts of a specimen, before and after cyclic loading for CP=100 After placement of stoppers, the tube is rotated continuously and
kPa, ru=0.85. vigorously for 20 min so that sand and silt are mixed well. Then one of
Specimen Condition Silt content Silt content Silt content
the stoppers is removed and a rubber membrane is lined over the tube.
part in FC=5% in FC=15% in FC=30% Subsequently, a porous disk and a filter-paper are placed on the open
end of the tube and the membrane is pulled over them. At this step,
Upper one- After 5.05% 16.52% 31.94% while porous disk is held firmly, the specimen is rotated and then being
third consolidation
After cyclic load 5.08% 19.80% 40.73%
placed on a base plate. After sealing the membrane, a split mold is
Lower one- After 4.81% 13.10% 29.11% assembled around the specimen. Since the tube has a diameter a few
third consolidation millimeters less than the mold, mold can easily place around it. In
After cyclic load 4.89% 10.26% 22.43% order to remove the tube, a low vacuum is applied to draw the
membrane off from the tube and tighten it to the mold wall. The tube
is then pulled slowly out of the mold and the mixture is densified by
3.2. Sample preparation
vibrating the base with a soft hammer. Then filter-paper, porous disk
and top cap are placed on top and the specimen is sealed by pulling the
Since homogeneity plays an important role in results of this study,
membrane on top cap. Finally a vacuum of about 15 kPa is applied
specimen preparation method that provides the most homogenous
within the specimen and loading cell filled up with water.
mixture is employed. Segregation is the most prevalent issue in
preparing silty sand specimens. Methods such as wet sedimentation
or air pluviation can induce significant segregation to these specimens 3.3. Testing procedure
[22]. Also in other methods like wet tamping or dry tamping, the
homogeneity is not at an acceptable level [23] and specimen behave To simplify the saturation process, carbon dioxide is percolated
quite different from undisturbed samples [24]. Dry deposition gives throughout specimens for an hour under the pressure of 1–5 kPa,
better results compared to previous methods, nonetheless the slurry depending on permeability of mixture. Subsequently, deaired water is
method creates much more homogeneous specimens [25,26]. Despite flushed through specimens as much as twice of its volume and
widespread use of slurry method for preparing silty sand specimens, a substituted for the CO2 within voids. In all tests, in order to decrease
recent research has shown that using a similar method but in dry the effect of back pressure on homogeneity [27], specimens are
condition, produces specimens with greater homogeneity and with a saturated in equal final back pressure of 100 kPa, with a B value
better control on sample density [8]. In this method, called mixed dry greater than 0.95. Finally specimens are consolidated isotropically to
deposition, an appropriate amount of oven-dried sand and silt are desired confining pressure of 50, 100 or 200 kPa.

137
R. Noorzad, M. Shakeri Soil Dynamics and Earthquake Engineering 97 (2017) 133–142

Fig. 8. Macrograph of specimens with 15% silt at CP=100 kPa, ru=0.85 for (a) Upper
one-third after cyclic load (b) Lower one-third after cyclic load.

Table 2 presents testing matrix. Specimens are loaded in two stages;

first they are subjected to a cyclic sinusoidal load in stress-controlled
condition with frequency of 1 Hz up to a certain pore water pressure
ratio which is defined as:

∆uc
ru =
σc′ (4)

Where Δuc is excess pore water pressure induced by cyclic loading and
σc’ is initial effective confining stress. Tests are conducted in ru=0, 0.35,
0.65 and 0.85. Pore water pressure ratios have been chosen so as to
prevent failure or liquefaction in samples. Since silty sand is capable of
liquefying in a pore water pressure ratio of less than 1 [28], the
maximum ru is limited to 0.85. In this study, ru is selected as a
parameter to evaluate cyclic loading instead of cyclic shear strain
because in silty sand, in a rather short range of shear strain, pore water
pressure increases dramatically [29]. This can complicate the inter-
pretation of results. Tests are conducted mostly at cyclic stress ratio
(CSR) of 0.25, however due to low strength of specimens with 30% silt,
tests on these specimens are performed at a lower CSR to hinder
occurrence of flow failure during cyclic loading.
Subsequently, specimens are loaded monotonically in strain-con-
trolled condition with a strain rate of 0.2%/min. Loading is continued
up to axial strain of 20%, but for specimens in which steady state did
not occur by the strain of 25%, strength at 25% strain is considered as
Fig. 9. Variation of normalized post-cyclic undrained shear strength versus pore
pressure ratio in different fines contents for: (a) CP=50 kPa; (b) CP=100 kPa; (c)
CP=200 kPa.
the ultimate state strength.
In 12 tests, first stage has been eliminated and specimens only
experience monotonic loading to simulate the condition ru =0. Failure
shape in specimens was similar, although some wrinkles are observed
in specimens with 30% silt (Fig. 4).
Cyclic loading can cause different phenomena in specimens; in ru of
about 0.35, despite the existence of excess pore water pressure, soil
behavior does not change. On the other hand, when the ru approaches
to 0.65, although no phase transformation has been occurred in most of
specimens, at monotonic stage, specimens start to show dilation shortly

138
R. Noorzad, M. Shakeri
Fig. 10. Variation of normalized post-cyclic undrained shear strength versus pore
pressure ratio in sand with 30% silt.
Fig. 11. Stress path for monotonic and post-cyclic loading of sand: (a) with 5% silt at
confining pressure of 200 kPa (Test No. 21–24) (b) with 30% silt at confining pressure of
100 kPa (Test No. 41–44).
after the stage begins. Finally, in ru of 0.85, phase transformation
occurs in soil due to cyclic loading, and behavior changes from
compressive to dilative. Fig. 5 exemplifies behavior of samples with
or without occurrence of phase transformation in cyclic stage.
139
Soil Dynamics and Earthquake Engineering 97 (2017) 133–142
4. Results
4.1. Monotonic tests
In monotonic tests, as demonstrated in Fig. 6(a), behavior of
Babolsar sand is dilative even at low densities. Such behavior has been
already seen in Froster sand and seems to be related to sand fabrics
[11]. Adding silt to sand, can reduce shear strength and decline the
tendency for dilation. Insofar as in specimens with 30% silt, in which
the sand grains are floating in the silt matrix, limited flow and quasi
Steady State happens or even in confining pressure of 50 kPa, flow
failure can be observed (Fig. 6(b)). This phenomenon in silty sand was
usually considered as reverse behavior because through increase of
confining pressure, the static liquefaction potential declines in these
soils [30,31]. However, some researchers did not observe reverse
behavior [32]. But a recent study shows there is no reverse behavior
in silty sand. In fact, the observed phenomenon is due to high
compressibility of sand with high amount of silts during consolidation.
If this compressibility change the location of steady state line with
respect to isotropic consolidation line, mentioned behavior happens
[33]. In the current study, this behavior has been observed in sand with
30% silt.
4.2. Post-cyclic strength
After cyclic loading, strength at low strains decreased in all speci-
mens. However, as shown in Fig. 7(a–d), this reduction is more
appreciable in samples with higher silt content and/or higher ru.
Especially, specimens which have experienced phase transformation
at the cyclic stage. For example, in specimens with 30% fines, peak
strength decreases as ru increases, inasmuch as there is no peak
strength for specimens in which phase transformation have occurred
at cyclic loading stage. In fact, these specimens have begun to dilate
prior to initiation of monotonic loading.
Cyclic loading, in addition to its influence on the strength at low
strains, can also affect the ultimate state strength of soil even more
significantly. Fines content is the most important parameter in assess-
ment of change in ultimate state strength. Fig. 7(a) proves that, in clean
sand, cyclic loading has negligible effect on the ultimate state strength
and specimens with any ru reach to a rather equal strength. Such
behavior has also been detected in sand with 5% silt (Fig. 7(b)).
Generally, the behavior of clean sand and sand with 5% silt are
analogous.
Unlike sand with low fines contents, in specimens with 15% silt,
cyclic loading can affect the ultimate state strength. In Fig. 7(c),
through increase of ru, the ultimate state strength decreases. This
change represents the alternation in propagation of silt particles within
the soil structure, due to cyclic loading.
In sand with 30% silt, as ru increases, in addition to transition from
quasi steady to dilative behavior, the ultimate state strength also
changes. Fig. 7(d) shows that dilative behavior of the specimens
continues up to 25% strain.
5. Discussion
5.1. Mechanism of changes
In order to analyze changes in silty sand strength after cyclic
loading and its mechanism in samples with various silt contents, a
binary packing model is utilized. In this model the sand-silt mixture is
assumed to contain two materials with only two different particle size
which are usually mean grain of sand and silt. Cyclic loading creates a
tendency for displacement in all particles within the specimens. Since
the silt particles are smaller, friction between them is less than sand
R. Noorzad, M. Shakeri
Fig. 12. Variation of RFD versus confining pressure in sand with: (a) 0% fines; (b) 5% fines; (c) 15% fines; (d) 30% fines.
grains and due to negligible mass, these particles can easily float and
migrate within the specimen as ru increases. In contrast to these
particles, in sand grains, contact forces as well as relatively massy
grains, usually bound the movement.
Post-cyclic rearrangement of silt particles within the specimen
depends on their quantity. For example, in clean sand and sand with
5% silt, most particles are located in voids between sand grains. So
when specimens are subjected to cyclic loading, despite the tendency to
displace, silt particles are confined and cannot migrate. Therefore,
change in ultimate state strength of these soils after cyclic loading is
negligible. It should be noted that in these cases, confining pressure
which is directly applied to sand grains, limits their displacement
during cyclic loading and does not allow silt particles to escape from
the voids. Moreover, post mortem examination of the samples have
revealed a uniform distribution of red-color silt in sand matrix
confirming that in low silt contents, the sample remains rather
homogeneous after cyclic loading.
In sand with 15% silt, although still voids are not filled completely
with silt, a considerable amount of silt particles are located between
sand grains in a metastable condition (Fig. 1(b)). The smallest
displacement in the specimen due to cyclic loading can result in
releasing and floating of silt particles within the specimen.
Consequently, some of silt particles migrate upward whereas massive
sand grains stay at their place. So the arrangement of soil changes. The
more the pore water pressure ratio generated (through more number of
cycles or higher CSR), the greater the changes are. Observation of these
specimens after testing, reveals that color of mixture in upper third of
samples, tends to red due to deposit of silt, even after removing excess
140
Soil Dynamics and Earthquake Engineering 97 (2017) 133–142
fines slurry from the top of specimens, while in lower third, the mixture
is slightly darker. In fact, if the number of cycles was large enough,
specimens would be more like two-layered sand-silt samples instead of
mixed ones. Consequently, in upper part of specimen, sand grains are
held apart by excess silt. So the strength in upper part of specimen and
therefore the total ultimate state strength decreases. This reduction
proves that after cyclic loading, specimens tend to behave more like
layered samples with the same silt content and density which have
lower ultimate state strength in comparison with mixed ones [34].
In sand with 30% silt, voids are completely filled and sand particles
are isolated in the silt matrix. Hence, after applying the cyclic load, silt
particles can move easier without any confinement due to sand grains.
Their movement upwards reestablishes sand grains contact in lower
parts of specimens. In this case, as discussed previously, increase in silt
content at upper parts of specimens, does not create any considerable
changes in mechanism of sustaining shear stress. In this part of
specimen, shear bands keep passing through silt particles, so the
strength difference is not substantial. Conversely, at lower parts, if the
contact between sand grains reestablishes, soil strength increases
dramatically and therefore, strength of whole specimen increases as
well. However the reestablishment of contacts also depends on other
parameters like confining pressure which at higher values, can bound
the movement of particles. Moreover, with higher pore water pressure
ratio (and more loading cycles), changes in soil arrangement will be
more intense. As a result, in sand with 30% silt, at confining pressure of
50 and 100 kPa, cyclic loading reestablishes the contact between sand
grains and increases the ultimate state strength. Whereas higher
confining pressure of 200 kPa limits the displacement of silt particles
R. Noorzad, M. Shakeri Soil Dynamics and Earthquake Engineering 97 (2017) 133–142

both cases the peak strength is decreased in comparison with homo-

genous specimens [34]. But there are some differences too, like
increase of ultimate strength in some cases. Table 3 and Fig. 8 show
an example of changes in silt distribution and soil structure of silty
sands after the tests.
Fig. 9 shows the influence of ru on normalized post-cyclic strength
for different fines content. Results are normalized with respect to
corresponding parameters from monotonic tests. As demonstrated,
changes in clean sand and sand with 5% silt are insignificant. In sand
with 15% silt, ultimate state strength decreases with ru increment. The
rate of reduction is not greatly changed in various confining pressures
because in this case, even a small displacement is enough to release the
metastable silt particles. In specimens with 30% silt, changes can be
substantial. In confining pressure of 50 kPa, after cyclic loading,
migration of silt particles cause in reconnection of sand grains. So
the mixture experiences limited flow instead of flowing and conse-
quently the strength increases in a way that cannot be shown in current
scale of the figure. Hence, the results of sand with 30% silt are
presented separately in Fig. 10. The figure shows that in sand with
30% silt, change in strength is highly dependent to confining pressure
and after cyclic loading, the ultimate state strength can decrease or
increase in different confining pressures.

5.2. Potential of dilation

Fig. 11 compares stress path of monotonic and post-cyclic tests for

sand with 5% silt in confining pressure of 200 kPa. It can be assessed
that in these specimens (like all other specimens), as ru increases,
dilation potential increases in specimens. In order to evaluate the
tendency to dilation in specimens, a parameter called resistance to
further deformations (RFD) is defined as follows [35]:
pul′ − ppt′
RFD =
pul′ (5)

Where p’ul is mean effective stress at ultimate state from corresponding

monotonic test and p’pt is mean effective stress at which phase
Fig. 13. Effect of CSR on ultimate state shear strength of sand with 15% silt: (a)
transformation has been occurred in post-cyclic monotonic loading.
CP=200 kPa, ru=0.35 (Test No. 34, 50); (b) CP=100 kPa, ru=0.65 (Test No.31, 52).
For specimens that have begun to dilate from beginning of monotonic
loading, p’pt is considered as mean effective stress at the start of
monotonic stage.
Fig. 12 shows that in all specimens except sand with 30% silt, an
increment in confining pressure, declines the dilation potential and at
the same time, increases the variation of RFD among similar specimens
with different pore water pressure ratios. In fact as fines content
increases, the effect of confining pressure and ru on RFD increases too.
Only in sand with 30% silt, RFD increases anomalously as confining
pressure increases. As mentioned before, this is due to high compres-
sibility of these specimens which leads to decrease of flow potential at
higher confining pressures. However in post-cyclic loading, this
behavior is gradually vanished.

5.3. Effect of CSR

In physical perspective, as CSR decreases, more loading cycles

should be needed to approach a certain ru. In order to show the effect of
CSR on post-cyclic strength of silty sands, a few tests are conducted on
specimens with lower CSR. Similar specimens are loaded cyclically with
different CSR values up to an equal target ru. Then, their post-cyclic
Fig. 14. Effect of CSR on ultimate state shear strength of sand with 30% silt for ru=0.35,
strength are examined. As expected, in clean sand and sand with 5%
CP=100 kPa (Test No. 42, 54).
silt, changes were negligible. In sand with 15% silt although, the
ultimate state strength varies while CSR changes. Fig. 13(a) shows
so the contact does not establish and ultimate state strength decreases
specimens of two similar tests on sand with 15% fines in confining
with increasing ru, just like sand with 15% silt. Sand with 30% silt that
pressure of 200 kPa. In order to achieve a pore water pressure equal to
has been subjected to dynamic load also has some similarities with
0.35, first specimen subjected to a dynamic load with a CSR of 0.25 and
layered samples with the same silt content and density. For example, in
reached the desired value in 3 cycles. But in the second specimen, it

141
R. Noorzad, M. Shakeri
takes 13 cycles to approach the same ru in CSR of 0.15. Substantial
difference, indicates the impact of number of loading cycles on the
ultimate state strength. Nevertheless, this difference would be less
severe in further ru of 0.65 and 0.85 (Fig. 13(b)). Because in these
ratios all specimens have already been subjected to a rather high
number of cycles and soil structure approaches to its eventual rear-
ranged post-cyclic condition. So after a certain number of cycles, CSR
would lose its effect on ultimate state strength.
In the sand with 30% silt, also a strength reduction has been
observed in specimens after changing the CSR in a procedure similar to
sand with 15% silt (Fig. 14).
So the post-cyclic strength in silty sand depends on fines content,
confining pressure, ru, CSR and also density of specimens.
6. Conclusion
A series of consolidated undrained triaxial tests are conducted on
different sand-silt mixtures in order to evaluate the post-cyclic behavior
of silty sand. Results show that after cyclic loading:
• Shear strength at low strains (lower than 1%) decreases in all
specimens. This reduction is more obvious in higher fines contents
(i.e. 15% and 30% silt). In fact an increment in ru, directly results in
a decline in strength at low strains. Zero strength shows the
occurrence of liquefaction.
• Ultimate state strength at low fines contents (0% and 5%) does not
change after cyclic loading, because in these specimens, silt particles
are confined within voids and cannot move. In sand with 15% silt,
cyclic loading results in displacement of silt particles which are in a
metastable condition. These particles can migrate to upper parts of
the specimens and decrease the strength. In specimens with 30% silt
which are beyond the threshold fines content, cyclic loading can
either increase or decrease the shear strength at ultimate state. If
migration of silts in these specimens, result in reestablishment of
contacts between sand grains, ultimate state strength will increase
dramatically, otherwise it will decrease.
• Changes of ultimate state strength in sand with 15% fines is not
dependent on confining pressure, but in sand with 30% fines it is
highly dependent on confining pressure.
• As ru increases, the dilation potential increases in all specimens. At
higher confining pressures, dilation potential decreases, except in
sand with 30% fines in monotonic loading which has a different
behavior due to compressibility.
•
When number of cycles increases in order to approach a certain ru,
changes in shear strength will be more intense. The more the
number of cycles are, the less the effect of CSR is.
Acknowledgment
The authors would like to acknowledge Dr. Mehdi Omidvar for his
kind helps in preparing the manuscript.
References
[1] Ovando-Shelley E, Pérez BE. Undrained behavior of clayey sands in load controlled
triaxial tests. Géotechnique 1997;47(1):561–9.
[2] Yasuhara K. Postcyclic undrained strength for cohesive soils. J Geotech Eng
142
Soil Dynamics and Earthquake Engineering 97 (2017) 133–142
1994;120(11):1961–79.
[3] Soroush A, Soltani-Jigheh H. Pre- and post-cyclic behavior of mixed clayey soils.
Can Geotech J 2009;46(2):115–28.
[4] Erken A, Ulker BMC. Effect of cyclic loading on monotonic shear strength of fine-
grained soils. Eng Geol 2007;89(4):243–52.
[5] Yasuhara K, Murakami S, Song B, Yokokawa S, Hyde AFL. Postcyclic degradation of
strength and stiffness for low plasticity silt. J Geotech Geoenviron Eng
2003;129(8):756–69.
[6] Yasuhara K, Murakami S, Komine H, Unno T. Effect of initial static shear stress and
principal stress reversal on cyclic and post-cyclic undrained shear of sand. 16th
ICSMGE, Osaka, Japan; 459-463; 2005.
[7] Yoshida N, Yasuda S, Kiku M, Masuda T, Finn L. Behavior of sand after
liquefaction. Proceedings from the fifth U.S.-Japan workshop on earthquake
resistant design of lifeline facilities and countermeasures against soil liquefaction,
Buffalo, N.Y., U.S; 1994. p.181–98.
[8] Wood FM, Yamamuro JA, Lade PV. Effect of depositional method on the undrained
response of silty sand. Can Geotech J 2008;45(11):1525–37.
[9] Poulos SJ, Castro G, France JW. Liquefaction evaluation procedure. J Geotech Eng
1985;111(6):772–92.
[10] Alarcon-Guzman A, Leonards GA, Chameau JL. Undrained monotonic and cyclic
strength of sands. J Geotech Eng 1988;114(10):1089–109.
[11] Vaid YP, Thomas J. Liquefaction and postliquefaction behavior of sand. J Geotech
Eng 1995;121(2):163–73.
[12] Pitman TD, Robertson PK, Sego DC. Influence of fines on the collapse of loose
sands. Can Geotech J 1994;31(5):728–39.
[13] Chang WJ, Hong ML. Effect of clay content on liquefaction characteristics of gap-
graded clayey sands. Soils Found 2008;48(1):101–14.
[14] Thevanayagam S, Martin GR. Liquefaction in silty soils—screening and remediation
issues. Soil Dyn Earthq Eng 2002;22(9–12):1035–42.
[15] Thevanayagam S. Effect of fines and confining Stress on undrained shear strength
of silty sands. J Geotech Geoenviron Eng 1998;124(6):479–91.
[16] Bahadori H, Ghalandarzadeh A, Towhata I. Effect of non plastic silt on the
anisotropic behavior of sand. Soils Found 2008;48(4):531–45.
[17] Rahman MM, Lo SR, Gnanendran CT. On equivalent granular void ratio and steady
state behavior of loose sand with fines. Can Geotech J 2008;45(10):1439–56.
[18] Yang SL, Sandven R, Grande L. Instability of sand-silt mixtures. Soil Dyn Earthq
Eng 2006;26(2–4):183–90.
[19] Ni Q, Tan TS, Dasari GR, Hight DW. Contribution of fines to the compressive
strength of mixed soils. Géotechnique 2004;54(9):561–9.
[20] Rahman MM, Lo SR, Gnanendran CT. Reply to the discussion by Wanatowski and
Chu on “On equivalent granular void ratio and steady state behavior of loose sand
with fines”. Can Geotech J 2009;46(4):483–6.
[21] Noorzad R, Fardad Amini P. Liquefaction resistance of Babolsar sand reinforced
with randomly distributed fibers under cyclic loading. Soil Dyn Earthq Eng
2014;66:281–92.
[22] Ishihara K. Liquefaction and flow failure during earthquakes. Géotechnique
1993;43(3):351–415.
[23] Carraro JAH, Prezzi M, Salgado R. Shear strength and stiffness of sands containing
plastic or nonplastic fines. J Geotech Geoenviron Eng 2009;135(9):1167–78.
[24] Huang AB, Huang YT. Undisturbed sampling and laboratory shearing tests on a
sand with various fines contents. Soils Found 2007;47(4):771–81.
[25] Kuerbis R, Vaid YP. Sand sample preparation – the slurry deposition method. Soils
Found 1988;28(4):107–18.
[26] Carraro JH, Prezzi M. A new slurry-based method of preparation of specimens of
sand containing fines. Geotech Test J 2008;31(1):1–11.
[27] Xia H, Hu T. Effects of saturation and back pressure on sand liquefaction. J
Geotech Eng 1991;117(9):1347–62.
[28] Ishihara K. Soil Behaviour in Earthquake Geotechnics, First ed.. New York: Oxford
University Press; 1996. p. 220.
[29] Erten D, Maher MH. Cyclic undrained behavior of silty sand. Soil Dyn Earthq Eng
1995;14(2):115–23.
[30] Yamamuro JA, Lade PV. Steady-state concept and static liquefaction of silty sands.
J Geotech Geoenviron Eng 1998;124(9):868–76.
[31] Yamamuro JA, Lade PV. Static liquefaction and reverse behavior of silty sand.
GeoCongress 2012;118(1):829–38.
[32] Chu J, Leong WK. Effect of fines on instability behaviour of loose sand.
Géotechnique 2002;52(10):751–5.
[33] Rahman MM, Lo SR. Undrained behavior of sand-fines mixtures and their state
parameter [ASCE]. J Geotech Geoenviron Eng 2014;140(7):04014036.
[34] Naeini SA, Baziar MH. Effect of fines content on steady-state strength of mixed and
layered samples of a sand. Soil Dyn Earthq Eng 2004;24(3):181–7.
[35] Thevanayagam S, Shenthan T, Mohan S, Liang J. Undrained fragility of clean
sands, silty sands, and sandy silts. J Geotech Geoenviron Eng
2002;128(10):849–59.