Geotextiles and Geomembranes 46 (2018) 869–881

Contents lists available at ScienceDirect

Geotextiles and Geomembranes

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

Role of soil inherent anisotropy in peak friction and maximum dilation T

angles of four sand-geosynthetic interfaces
Aliyeh Afzali-Nejad, Ali Lashkari∗, Benyamin Farhadi
Shiraz University of Technology, Shiraz, Iran

A R T I C LE I N FO A B S T R A C T

Keywords: Using a modified direct shear apparatus, an extensive experimental investigation is conducted into the influence
Geosynthetics of the inherent anisotropy of sand on the mobilization of the peak and critical state friction angles as well as the
Sand maximum dilation angle of the interfaces between an inherently anisotropic crushed sand and two woven
Inherent anisotropy geotextiles, one nonwoven geotextile, and one geomembrane. Experimental findings confirm that both peak and
Interface
maximum dilation angles of sand-geosynthetic interfaces are affected from soil inherent anisotropy depending
Direct shear
Friction
on the bedding plane inclination with respect to the shear plane. However, a unique critical state (residual)
Dilation friction angle is attained for each interface type irrespective of the bedding plane inclination angle. Compiling
results of a total of 141 tests, it is shown that a unique rule describes stress-dilation relationship of four different
dense crushed sand-geosynthetic interfaces. The experimental data indicate that the ϕp vs. θ and ψmax vs. θ
curves are symmetrical with respect to θ = 90° for the sand-woven geotextile and sand-geomembrane interfaces.
Finally, it is shown that a constitutive equation by Pietruszczak and Mroz (2001) can predict the variation of ϕp
with θ for the sand-woven geotextile and sand-geomembrane interfaces.

1. Introduction
Polymeric geotextiles and geomembranes have been applied ex-
tensively in practice as civil engineering materials for geotechnical,
geoenvironmental, and pavement engineering purposes in the recent
years. The design and performance of geosynthetic reinforced earth
structures depend strongly on the mechanical behavior of the thin in-
terfaces transmitting forces between soil and geosynthetics at their
contact area. Experimental studies have indicated that the mobilization
of friction and dilation at the interfaces between nonplastic granular
soils and geosynthetics largely depend on the average soil particle
shape characteristics, soil relative density, normal stress, degree of sa-
turation, and soil cementation on one hand, and the geosynthetics
polymer category, surface texture, tensile strength, and surface rough-
ness on the other (e.g., Giroud et al., 1985; Negussey et al., 1989;
Athanasopoulos, 1993; Lee and Manjunath, 2000; Anubhav and
Basudhar, 2010; Khoury et al., 2011; Frost et al., 2012; Esmaili et al.,
2014; Anubhav and Wu, 2015; Ferreira et al., 2015; Hatami and
Esmaili, 2015; Martinez et al., 2015; Vieira et al., 2015; Choudhary and
Krishna, 2016; Lashkari and Kadivar, 2016; Liu et al., 2016; Vangla and
Gali, 2016a, 2016b; Vieira and Pereira, 2016; Afzali-Nejad et al., 2017;
Fowmes et al., 2017; Martinez and Frost, 2017; Punetha et al., 2017).
The mechanical behaviors of natural and man-made deposits of
granular soils are usually considered anisotropic because such soils
behave differently with rotation of the principal stress axes with respect
to their bedding plane. The anisotropic behavior of granular soils from a
micromechanical view is attributed to the preferred orientation of
particles and contact forces between them (Oda, 1993; Guo and Zhao,
2013; Jiang et al., 2017). The inherent (say initial) anisotropy is in-
stituted through soil sedimentation and each sedimentation process
possesses its own initial fabric (see Papadimitriou et al., 2005). Con-
sequently, inherent anisotropy may play a central role in the initial
stiffness, drained peak friction and maximum dilation angles. Testing
with true triaxial and hollow cylindrical torsional shear apparatuses can
offer a broad perspective of different impacts of inherent anisotropy on
the mechanical behaviors of granular soils (e.g., Nakata et al., 1998;
Yoshimine et al., 1998; Rodriguez and Lade, 2013; Xiong et al., 2016;
Al-Rkaby et al., 2017; Xiao et al., 2014, 2017). However, the true
triaxial and hollow cylindrical torsional shear apparatuses are still in
relatively limited use owing to their high operation costs. Therefore,

∗
Corresponding author.
E-mail addresses: A.afzali@sutech.ac.ir (A. Afzali-Nejad), lashkari@sutech.ac.ir, lashkari_ali@hamyar.net (A. Lashkari), B.farhadi@sutech.ac.ir,
benyamin.farhadi68@gmail.com (B. Farhadi).

https://doi.org/10.1016/j.geotexmem.2018.08.003
Received 21 January 2018; Received in revised form 28 July 2018; Accepted 2 August 2018
0266-1144/ © 2018 Elsevier Ltd. All rights reserved.
A. Afzali-Nejad et al. Geotextiles and Geomembranes 46 (2018) 869–881

Nomenclature X X = 1 − 3 cos2 θ (see Eq. 6)

θ bedding plane inclination angle (see Figs. 4 and 5)
a1, a2 parameters of Eqs. (3) and (5) σn normal stress in direct shear test
A, B, C, Dconstants of the third order polynomial function given in τ shear stress in direct shear test
Eq. (6) τp peak shear strength in direct shear test
Cc coefficient of curvature τcs critical state shear strength
Cu coefficient of uniformity ϕp peak friction angle [ = tan−1 (τp/ σn ) ]
d50 mean particle size ϕp (θ = 0∘) the peak internal friction angle for θ = 0∘
d85 particle size at which 85% of the soil is finer by weight ϕp∗ the lowest value of ϕp within the range 0°≤θ < 180°
dmax maximum particle size ϕcs critical state friction angle [ = tan−1 (τcs / σn ) ]
e0 initial void ratio prior to shear ψ dilation angle [ = tan−1 (δv / δu) ]
emax maximum void ratio ψmax maximum dilation angle [ = tan−1 (δv / δu)max ]
emin minimum void ratio ∗
ψmax the lowest value of ψmax within the range 0°≤θ < 180°
Gs average specific gravity of sand particles Ω (= Ωij) second-order tensor describing the bias in the spatial dis-
ni ith component of the unit vector normal to the shear plane tribution of ϕp
R average roundness of particles Ω1, Ω2, Ω3 principal values of Ω for naturally deposited transversely
S average sphericity of particles isotropic soils
u shear displacement
v vertical displacement

direct shear testing on inherently anisotropic soils has been recon- detailed impact of inherent anisotropy on the peak friction and max-
sidered as an economic means in recent years (e.g., Guo, 2008; Azami imum dilation angles of sand-geosynthetic interfaces. Analyses of 141
et al., 2010; Tong et al., 2014; Oboudi et al., 2016; Farhadi and sand-geosynthetic interface and 42 sand alone direct shear tests have
Lashkari, 2017). Moreover, direct shear testing is also a suitable tool to indicated that soil inherent anisotropy may affect the peak internal
study the behavior of interfaces between soils and civil engineering friction and the maximum dilation angles of sand and sand-geosyn-
materials such as steel, concrete, geotextiles and geomembranes (e.g., thetic interfaces.
Evgin and Fakharian, 1996; Lee and Manjunath, 2000; Lings and Dietz,
2005; DeJong and Westgate, 2009; Anubhav and Basudhar, 2010;
Khoury et al., 2011; Frost et al., 2012; Hatami and Esmaili, 2015; Vieira 2. Test materials
et al., 2015; Choudhary and Krishna, 2016; Liu et al., 2016; Vangla and
Gali, 2016a, 2016b; Wang et al., 2016; Vieira and Pereira, 2016; Afzali- 2.1. Crushed sand
Nejad et al., 2017; Farhadi and Lashkari, 2017; Fowmes et al., 2017).
Experimental studies reported by Oda and Koishikawa (1979), An artificially graded quarry-run sand obtained through industrial
Siddiquee et al. (2001), Azami et al. (2010), Li et al. (2010), and Qin crushing, milling, and washing processes of parent rocks was used in
et al. (2016) have confirmed that inherent anisotropy has a remarkable this study. The particle size distribution of the sand is demonstrated in
influence on the bearing capacity of shallow foundations resting on Fig. 1 and accordingly, the uniformly graded sand can be classified as
anisotropic sands and retaining walls with anisotropic backfills. The SP based on the Unified Soil Classification System (ASTM D2487).
mechanical behavior of geosynthetic reinforced earth structures Table 1 lists the primary physical properties of the tested crushed sand,
strongly depends on the mobilization of shear strength at the interfaces and a Scanning Electron Microscope (SEM) image of the typical sand
between the soil and reinforcement. However, the impact of soil in- particles is shown in Fig. 2.
herent anisotropy on the mechanical behavior of soil-structure inter-
faces is a rather young subject of research. Recently, Farhadi and
Lashkari (2017) studied the mechanical behavior of sand-steel inter-
faces of different normalized roughness values. They reported the
combined impact of contact surface asperities and sand inherent ani-
sotropy on mobilization of shear stress and dilation in rough sand-steel
interfaces. Al-Rkaby et al. (2017) studied the mechanical behavior of
geogrid reinforced anisotropic sand specimens using a large hollow
cylinder torsional shear apparatus. Depending on the angle of principal
stress rotation with respect to the bedding plane, Al-Rkaby et al. (2017)
reported that a significant improvement of strength in geogrid re-
inforced sand specimens may be attained as a consequence of con-
finement resulting from the tension in geogrid layers.
In order to investigate the outcome of sand inherent anisotropy on
the behavior of sand-geosynthetic interfaces, an extensive series of di-
rect shear tests on the interfaces between an inherently anisotropic
crushed sand and two woven geotextiles, one nonwoven geotextile, and
one geomembrane is reported here. The testing program covers a wide
spectrum of bedding plane inclination angles in order to realize the Fig. 1. Particle size distribution of the tested sand.

870
A. Afzali-Nejad et al. Geotextiles and Geomembranes 46 (2018) 869–881

Table 1
Physical properties of the crushed sand.
General particle shape Gs [−] d50 [mm] Cu [−] Cc [−] emax [−] emin [−] R [−] S [−]

angular 2.67 0.65 1.5 0.98 1.0 0.55 0.35 0.55

Fig. 2. SEM image of the crushed sand particles.
2.2. Geosynthetics
Four different types of geosynthetics were used in this investigation:
two multifilament woven geotextiles (i.e., woven GTX1, and woven
GTX2), one nonwoven geotextile (i.e., nonwoven GTX), and one PVC
geomembrane (i.e., GMB). General characteristics of each geosynthetic
are summarized in Table 2, and the SEM images for the surface texture
of the geosynthetics are illustrated in Fig. 3.
3. Testing program
A modified direct shear apparatus setup was used to investigate the
mechanical behaviors of the interfaces between the inherently aniso-
tropic crushed sand and the geosynthetics appointed in Section 2. A
number of sand alone direct shear tests were also conducted to better
realize the anisotropic response of the tested sand. For the interface
tests, the geosynthetics were glued on 100 mm × 100 mm (=
length × width) dummy steel blocks with variable thicknesses. The
thickness of each dummy block was selected such that upon installing
the geosynthetic, the sand-geosynthetic interface was located flush with
the predetermined shear plane of the direct shear box. Then, the
dummy steel block was placed in the lower half of the direct shear box
(see Fig. 4 for a schematic configuration). The direct shear box provides
an inner dimension of 100 mm × 100 mm × 22 mm (= length × width
× depth) for soil specimens. To prepare the soil specimens with in-
clined bedding planes with respect to the shear plane, the shear box was
attached to an adjustable table. Consequently, the upper level of the
table was tilted to a desired inclination angle as illustrated in Fig. 5. The
geometrical definition of the bedding plane inclination angle, θ, in
accordance with Fu and Dafalias (2011), Tong et al. (2014), and
Farhadi and Lashkari (2017) is illustrated in Fig. 4. In the testing pro-
gram, the crushed sand-geosynthtic interface and sand alone direct
shear tests covered a wide spectrum of bedding plane inclination angles
from θ = 0°–180° (i.e., θ = 0°, 15°, 30°, 45°, 60°, 75°, 105°, 120°, 135°,
150°, 165°, and 180°) [see Fig. 5]. For preparation of the specimens, dry
sand was poured with a funnel within the upper half of the shear box in
three or four successive layers [see Fig. 5(l) as well as parts “a” to “f” of
Fig. 6]. It is worth mentioning that elimination of dummy steel blocks
in the sand alone direct shear tests almost doubled the number of soil
layers. Each layer was carefully compacted to the target density by
tamping [see Fig. 6(g)]. A step-by-step schematic illustration for pre-
paration of inherently anisotropic specimens in sand-geosynthetic in-
terface tests is illustrated in Fig. 6. The sand-geosynthtic interface and
sand alone direct shear tests were carried out in accordance with ASTM
D5321 and ASTM D3080, respectively. The tests were performed at a
constant shear displacement rate of 0.5 mm/min under σn = 57 kPa. In
all tests, the crushed sand specimens were compacted to the average
initial relative density of 78% (e0 = 0.65) prior to shearing. Using a
trial and error procedure, an under compaction scheme was planned in
order to ensure uniform distribution of void ratio within the specimens.
The testing program consists of at least one duplicate for each experi-
ment.
To reduce scale effects, ASTM D3080 and ASTM D5321 recommend
that the total specimen thickness must be greater than 6dmax, the spe-
cimen width must be larger than 15d85 of sand, and the depth of each
container must be greater than 6dmax. Considering the shear box di-
mensions mentioned above and having d50 = 0.65 mm, d85 = 0.80 mm,
and dmax = 1.18 mm from the particle size distribution curve illustrated
in Fig. 1, it is observed that the shear box dimensions fulfill the ASTM
requirements.

Table 2
Geosynthetics general properties.
Material Thickness [mm] Mass per unit Tensile strength Tensile strength (cross Failure elongation `Failure elongation
area [gr/m2] (MD)a [kN/m] MD)a [kN/m] (MD)a [%] (cross MD)a [%]

Woven Geotextile No. 1 (woven GTX1) 0.5 375 50 25 12 12

Woven Geotextile No. 2 (woven GTX2) 0.5 350 100 102 14 14
Nonwoven Geotextile (nonwoven GTX) 2.1 485 5.75 7.05 80 100
PVC Geomembrane (GMB) 2.0 650 18.98 17.27 318.6 354.1

a
MD stands for Machine Direction.

871
A. Afzali-Nejad et al. Geotextiles and Geomembranes 46 (2018) 869–881

Fig. 3. SEM images of the geosynthetics: (a) woven geotextile No. 1; (b) woven geotextile No. 2; (c) nonwoven geotextile; and (d) geomembrane.

Fig. 4. Definition of Tong et al. (2014) and Farhadi and Lashkari (2017) for the bedding plane inclination angle with respect to shear direction in the soil-
geosynthetic interface test [clockwise or counterclockwise]: (a) 0° ≤ θ < 90°, and (b).90° < θ ≤ 180°

872
A. Afzali-Nejad et al.
Fig. 5. Adjustment of the inclined table for specimen preparation: (a) θ = 0°; (b) θ = 15°; (c) θ = 30°; (d) θ = 45°; (e) θ = 60°; (f) θ = 75°; (g) θ = 105°; (h) θ = 120°;
(i) θ = 135°; (j) θ = 150°; (k) θ = 165°; and (l) specimen preparation at θ = 30°.
4. Test results and discussions
4.1. Influence of inherent anisotropy on the dense sand behavior
The mechanical behaviors of dense crushed sand specimens covering a
broad range of bedding plane inclination angles (i.e., θ = 0°, 15°, 30°, 45°,
60°, 75°, 105°, 120°, 135°, 150°, 165°, and 180°) are first investigated. By
excluding some tests for further clarity, that is to say by presenting data in
30° intervals for θ, the curves for mobilized shear stress and vertical dis-
placement versus shear displacement of the tests with θ = 0°, 30°, 60°, 105°,
135°, and 165° are illustrated in Fig. 7. Part “a” of Fig. 7 indicates that the
peak shear strength increases gradually from 74.6 kPa for the specimen with
θ = 0° to 99.1 kPa for the specimen with θ = 105° and thereafter, decreases
to 84.1 kPa at θ = 165°. However, all shear stress versus shear displacement
curves approach a single asymptotic shear strength in Fig. 7(a), corrobor-
ating the notion that a unique critical state fabric is asymptotically attained
in highly sheared sands (see Guo and Zhao, 2011; Gu et al., 2017; Jiang
et al., 2017; Jing et al., 2017). Fig. 7(b) also signifies that the volume
change behavior of the crushed sand specimens is affected by the bedding
plane inclination angle.
The variations of peak friction angle, ϕp, with bedding plane in-
clination angle for the inherently anisotropic dense crushed sand sam-
ples are presented in Fig. 8(a). Accordingly, ϕp remains nearly un-
changed within the range of 0° ≤ θ ≤ 45°beyond which, the peak
friction angle increases to a maximum around θ = 105°. Any further
increase in the bedding plane inclination angle within the range of
873
Geotextiles and Geomembranes 46 (2018) 869–881
105° < θ ≤ 180° leads to a gradual decline of ϕp. For the sake of com-
parison, similar curves reported by Azami et al. (2010) and Tong et al.
(2014), respectively, for crushed Limestone and Mica sand are also
superimposed on the figure. Of note, the peak friction angle for crushed
Limestone was studied by Azami et al. (2010) only for 0° ≤ θ ≤ 90°.
Inspection of Fig. 8(a) indicates that the curve for peak friction angle
versus bedding plane inclination angle for the anisotropic dense cru-
shed sand of this study agrees reasonably with the existing experi-
mental data in the literature. The peak shear strength of sands is highly
dependent on particles shapes as well as soil state. Angular and elon-
gated shapes of particles (see Fig. 2), the relatively low normal stress
applied in the experiments (i.e., σn = 57 kPa), and the high initial
density of the specimens (i.e., Dr ≈ 78%) all intensify the overall in-
terlocking between the selected crushed sand particles and justify the
high peak friction angles reported in Fig. 8(a) [e.g., Verdugo and de la
Hoz, 2007; Guo, 2008]. Variation of the maximum dilation angle, ψmax,
in Fig. 8(b) indicates that ψmax and ϕp change almost in a similar
fashion with the bedding plane inclination angle.
4.2. The behavior of inherently anisotropic crushed sand-geosynthetic
interfaces
The mobilization of shear strength and volume change response of
the interfaces between the initially anisotropic crushed sand and two
woven geotextiles (i.e., GTX1 and GTX2), one nonwoven geotextile, and
one geomembrane as contact materials are studied in this section. The
A. Afzali-Nejad et al. Geotextiles and Geomembranes 46 (2018) 869–881

Fig. 6. Step-by-step procedure for preparation of samples with inclined bedding angle: (a) horizontal positioning of the adjustable table; (b) bolting of the lower half
of the shear box to the adjustable table; (c) placement of the dummy steel block in the lower half of the shear box; (d) bolting of the upper half of the shear box; (e)
placement of the side walls; (f) setting of the adjustable table to the desired inclination angle considering Fig. 5 and raining dry sand into the shear box; (g) tamping of
the sand to the desired density; and (h) horizontal positioning of the adjustable table and surface finishing using a spatula.

testing program covers bedding plane inclination angles of θ = 0°, 15°,
30°, 45°, 60°, 75°, 105°, 120°, 135°, 150°, 165°, and 180° for all of the
sand-geosynthetic interface tests. For the crushed sand-geosynthtic in-
terfaces sheared under θ = 0°, 30°, 60°, 105°, 135°, and 165°, the curves
representing mobilization of the shear strength with shear displacement
and also the vertical displacement against shear displacement are illu-
strated in Fig. 9. A comparison between parts “a” and “c” of Fig. 9 in-
dicates that the peak shear strengths of the crushed sand-woven GTX1
interfaces are generally greater than those for the crushed sand-woven
GTX2 interfaces. However, the average critical state shear strengths of
the interfaces between GTX1 and GTX2, and the dense crushed sand are
practically the same. Comparison of Fig. 9(b) with Fig. 9(d) shows that
the crushed sand-woven GTX1 interfaces possess a greater tendency
towards dilation than the crushed sand-woven GTX2 interfaces. The
extent for variations of peak shear strength at the interfaces between
the crushed sand and nonwoven geotextile with the bedding plane in-
clination angle is almost similar to that for the crushed sand-woven
GTX2 interfaces; however, the greatest critical state (say residual) shear
strengths among all of the interface types are definitely achieved for the
crushed sand-nonwoven geotextile interfaces. Moreover, once subjected
to shear stress, the crushed sand-nonwoven geotextiles demonstrate a
greater tendency towards contraction at the beginning, which is sub-
sequently recovered partly by dilation. The crushed sand-geomembrane
interfaces demonstrate a gentle peak shear strength [see Fig. 9(g)], and
their mild dilation does not terminate until the end of the experiments.
Both the peak and critical state shear strengths of the crushed sand-
GMB interfaces are significantly lower than those for the crushed sand-
geotextile interfaces, which may be attributed to the initial smoothness

874
A. Afzali-Nejad et al.
Fig. 7. Shear strength vs. shear displacement and vertical displacement vs.
shear displacement curves for dense (e0 = 0.65) crushed sand samples with
σn = 57 kPa sheared under θ = 0°, 30°, 60°, 105°, 135°, and 165°.
Fig. 8. Variation of peak friction and maximum dilation angles with the bed-
ding plane inclination angle in the inherently anisotropic dense crushed sand of
this study: (a) ϕp vs. θ; (b) ψmax vs. θ
875
Geotextiles and Geomembranes 46 (2018) 869–881
of the geomembrane outer surface. However, scratches on the outer
surface of the geomembrane were easily discernible by naked eye at the
end of each crushed sand-GMB interface test.
For the interfaces between the inherently anisotropic crushed
sand and two woven geotextiles, one nonwoven geotextile, and one
geomembrane, variations of the peak friction angle, the critical state
friction angle, and the maximum dilatancy angle are demonstrated
through parts “a”, “b”, and “c” of Fig. 10, respectively. For both types
of the crushed sand-woven geotextiles interfaces as well as the in-
terface between the crushed sand and geomembrane, the peak fric-
tion angle descends from θ = 0° and reaches a minimum in between
θ = 15° and 30°. Then, the peak friction angle begins to increase
towards a maximum around θ = 90° and subsequently, the peak
friction angle gradually drops to reach the second minimum in be-
tween θ = 150° and 165°. For the crushed sand-nonwoven geotextile
interface, minimums for the peak friction angle are, respectively,
attained around θ = 75° and 165° and a transient maximum peak
friction angle is observed in between θ = 120° and 135°. For each
interface type, a single critical state friction angle can reasonably fit
the experimental data in Fig. 10(b). The latter observation is in
agreement with the recent DEM simulations of Gu et al. (2017), and
Jing et al. (2017). As expected, the crushed sand-nonwoven geo-
textile possesses the highest critical state friction angle, stemming
from the intrinsically greater entanglement of the crushed sand
particles in the nonwoven geotextile filaments at the contact region.
Furthermore, the crushed sand-GMB interface possesses the lowest
critical state friction angle as a direct consequence of the smoothness
of the gemembrane outer surface. For each sand-geosynthetic type in
Fig. 10(c), it is observed that the measured trend for the variation of
the maximum dilation angle nearly resembles the corresponding
curve for the peak friction angle with the bedding plane inclination
angle in Fig. 10(a).
4.3. Direction-dependency of ϕp, and ψmax
Rose diagrams of (ϕp − ϕp∗) vs. θ and (ψmax − ψmax ∗
) vs. θ drawn in
∗ ∗
the polar coordinates with (ϕp − ϕp ) and (ψmax − ψmax ) as the radial
coordination [ϕp∗ and ψmax∗
are the minimum values of ϕp and ψmax
within the range 0° ≤ θ < 180°, respectively], and θ as the angular
coordination have been applied in the literature as a convenient tool to
study the direction-dependent peak internal friction and maximum di-
lation angles of soils and soil-structure interfaces (e.g., Fu and Dafalias,
2011; Tong et al., 2014; Farhadi and Lashkari, 2017). For the interfaces
between crushed sand and woven GTX1, woven GTX2, nonwoven GTX,
and GMB, rose diagrams of (ϕp − ϕp∗) vs. θ are illustrated in Fig. 11. In
parts “a” and “b” of Fig. 11, the (ϕp − ϕp∗) vs. θ curves for the interfaces
between crushed sand and woven GTX1 and GTX2 possess two large
leaves approximately oriented towards θ = 90°, and 105° (and
θ = 270°, and 285° for the obverse leaves), respectively. On the other
hand, relatively narrow leaves are observed through parts “c” and “d”
of Fig. 11 for the crushed sand-nonwoven GTX and crushed sand-GMB
interfaces, which are, respectively, tilted towards θ = 120°, and 90°
(and θ = 300°, and 270° for the obverse leaves). In a similar fashion, the
∗
(ψmax − ψmax ) vs. θ curves for four different inherently anisotropic
sand-geosynthetic interfaces are illustrated in Fig. 12. For each inter-
face type, inspection of Figs. 11 and 12 reveals that the statistical or-
∗
ientations of the large leaves in the (ψmax − ψmax ) vs. θ , and
(ϕp − ϕp∗) vs. θ planes are almost the same.
A. Afzali-Nejad et al.
Fig. 9. Influence of the bedding plane inclination angle (i.e., θ = 0°, 30°, 60°, 105°, 135°, and 165°) on shear strength vs. shear displacement and vertical dis-
placement vs. shear displacement curves for: (a) and (b) crushed sand-woven GTX1 interfaces; (c) and (d) crushed sand-woven GTX2 interfaces; (e) and (f) crushed
sand-nonwoven GTX interfaces; (g) and (h) crushed sand-GMB interfaces.
4.4. Stress-dilation law for anisotropic sand-geosynthetic interfaces
Dove and Jarrett (2002) and Lings and Dietz (2005) reported results
of extensive experimental investigations on shear stress mobilization
and dilation at interfaces between different sands and rough aluminum
and steel plates. They suggested that the following correlation can ef-
fectively establish a linkage between stress variables and dilation in the
rough sand-aluminum and sand-steel interfaces:
ϕp − ϕcs ≈ ψmax (1)
The surface characteristics of the contact material have a strong role
in the critical state friction angle (say, the residual shear strength), the
maximum dilation, and consequently, the peak friction angle (say, the
peak shear strength) of soil-structure interfaces [e.g., Dove and Jarrett
876
Geotextiles and Geomembranes 46 (2018) 869–881
(2002)]. Recently, Afzali-Nejad et al. (2017) showed that the stress-
dilation data of various sand-woven geotextile interfaces follow Eq. (1)
reasonably. In the experiments conducted by Dove and Jarrett (2002),
Lings and Dietz (2005), and Afzali-Nejad et al. (2017), sand particles
were air-pluviated into the direct shear box (say θ = 0°) and hence, the
aforementioned experimental studies do not denote possible influence
of sand anisotropy on stress-dilation law of soil interfaces.
The (ϕp − ϕcs ) vs. ψmax data for 141 tests on the interfaces between
crushed sand-woven GTX1, crushed sand-woven GTX2, crushed sand-
nonwoven GTX, and crushed sand-GMB covering θ = 0°, 15°, 30°, 45°,
60°, 75°, 105°, 120°, 135°, 150°, 165°, and 180° are illustrated in Fig. 13.
Accordingly, it is observed that ϕp − ϕcs changes almost linearly with
ψmax for all of the interface types and bedding plane inclination angles:
ϕp − ϕcs ≈ 1.099 ψmax (2)
A. Afzali-Nejad et al.
Fig. 10. Variation of peak friction angle, critical state friction angle, and
maximum dilation angle with the bedding plane inclination angle in crushed
sand-geosynthetic interfaces: (a) ϕp vs. θ; (b) ϕcs vs. θ; and (c) ψmax vs. θ
As a result, it can be concluded that although both ϕp and ψmax vary
with θ, Eq. (1) can be applied effectively to describe the stress-dilation
linkage of the inherently anisotropic sand-geosynthetic interfaces.
5. Prediction of the variation of ϕp with θ
Based on the seminal methodology by Pietruszczak and Mroz
(2001), the orientation-dependent behavior of ϕp = ϕp (θ) in the direct
shear apparatus can be calculated through:
ϕp (θ) = ϕp (θ = 0∘) [1 + Ωij ni nj + a1 (Ωij ni nj )2 + a2 (Ωij ni nj
Geotextiles and Geomembranes 46 (2018) 869–881
where Ω describing the bias in the spatial distribution of ϕp, is a second-
order traceless tensor (tr Ω = Ω11 + Ω22 + Ω33 = Ω1 + Ω2 + Ω3 = 0 ), ni
is the ith component of the unit vector normal to the shear plane, and
ϕp (θ = 0∘) is the peak internal friction angle for θ = 0∘. Due to the effect
of the gravity field, naturally deposited sands are usually considered
transversely isotropic (i.e., Ω1 = Ω3 ≠ Ω2 where directions 1 and 3 are
defined in the deposition plane, and direction 2 is defined as normal to
the deposition plane). For such a configuration, one has (e.g.,
Pietruszczak and Mroz, 2001; Oboudi et al., 2016):
Ωij ni nj = Ω1 (1 − 3 cos2 θ) (4)
Implementation of Eq. (4) into Eq. (3) yields:
ϕp = ϕp (θ = 0∘) [ 1 + Ω1 (1 − 3 cos2 θ) + a1 (Ω1 (1 − 3 cos2 θ))2 + a2 (Ω1 (1 − 3 cos2 θ))3 + …]
(5)
where a1 and a2 are parameters depending on the interface type.
According to Oboudi et al. (2016), the first four terms in Eqs. (3)
and (4) are sufficient to calculate the variation of ϕp with θ, and the
negligible participation of higher order terms may be ignored. Thus, Eq.
(5) can be re-written as the following third-order polynomial function:
ϕp = A + BX + CX 2 + DX 3 (6)
wherein, X = 1 − 3 cos2 θ , A = ϕp (θ = 0∘) , B = A Ω1, C = a1 A Ω12
and D = a2 A Ω13 . Now, A, B, C, D [and accordingly ϕp (θ = 0∘) , Ω1, a1
and a2] can be determined consecutively by fitting a third order poly-
nomial function to the experimental data of ϕp illustrated against
X [=1 − 3 cos2 θ]. Using Eq. (6) together with the parameters obtained
from the aforementioned calibration procedure (see Table 3), the pre-
dicted curves of the peak friction angle versus the bedding plane in-
clination angle for the crushed sand-woven GTX1, crushed sand-woven
GTX2, crushed sand-nonwoven GTX, and crushed sand-GMB interfaces
are shown through parts “a” to “d” of Fig. 14, respectively. Except for
the interfaces between the crushed sand and the selected geomembrane
[see Fig. 14(d)], it is observed that Eq. (5) can realistically duplicate the
experimental measurements in Fig. 14. In this view, further researches
on the anisotropic behavior of anisotropic sand-nonwoven interfaces
appear to be compulsory. Finally, once Eq. (5) was calibrated (say
ϕp (θ = 0∘) , Ω1, a1 and a2 were determined) and ϕcs was obtained ex-
perimentally, either Eq. (1), or Eq. (2) together with Eq. (5) can be
applied for calculation of ψmax against θ.
6. Conclusions
A fundamental experimental study was undertaken to investigate
the impact of soil inherent anisotropy on the mobilization of shear
strength and the volume change behavior of the interfaces between an
inherently anisotropic sand and four different geosynthetics (i.e., two
woven geotextiles, one nonwoven geotextile, and one PVC geomem-
brane). Tests were performed using a modified direct shear apparatus
on sand-geosynthetic interfaces covering a broad range of bedding
plane inclination angles, i.e. from θ = 0°–180° with 15° steps in order to
achieve a comprehensive insight into the influence of soil inherent
anisotropy on the peak mobilized friction angle, critical state friction
angle and maximum dilation angle of the inherently anisotropic sand-
geosynthetic interfaces. Based on the analysis and interpretation of the
experimental data, the following main conclusions can be drawn:
• For the interfaces between the crushed sand and two woven geo-
textiles, the measured peak friction angles were dissimilar at
)3 + …] (3)
877
A. Afzali-Nejad et al.
Fig. 11. Rose diagrams for (ϕp − ϕp∗) vs. θ demonstrating direction-dependency of the peak internal friction angle: (a) sand-woven GTX1 interfaces; (b) sand-woven
GTX2 interfaces; (c) sand-nonwoven GTX interfaces; (d) sand-GMB interfaces.
identical bedding plane inclination angles; however, nearly identical
critical state friction angles were attained. A notable tendency to-
wards contraction was initially observed in the volume change re-
sponse of the crushed sand-nonwoven geotextile interfaces and
thereafter, peak shear strength was achieved at larger shear dis-
placements in such interfaces. The crushed sand-geomembrane in-
terfaces demonstrate gentle peak shear strengths and their mild di-
lations do not terminate at the end of the tests.
• Both the peak internal friction and the maximum dilation angles of
the pure sand specimens as well as crushed sand-geosynthetic (i.e.,
two woven geotextiles, one nonwoven geotextile, and one geo-
membrane) interfaces are affected by soil inherent anisotropy
through the changes in soil bedding plane inclination angle with
respect to shear plane.
• For pure sand specimens, ϕp and ψmax remain almost unchanged
878
Geotextiles and Geomembranes 46 (2018) 869–881
within the range 0∘ ≤ θ ≤ 45∘beyond which, both start to increase
and reach their maximums around θ = 105∘ . Thereafter, they de-
crease steadily with θ within the range 105∘ < θ ≤ 180∘.
• For the crushed sand-woven geotextile and crushed sand-geomem-
brane interfaces, variations of ϕp and ψmax with θ posses two tran-
sient minimums within the ranges θ ∈ [15∘ , 30∘] as well as
θ ∈ [150∘, 165∘]. For such interfaces, a transient maximum was
observed at θ = 90∘.
• For the sand-woven geotextile and sand-geomembrane interfaces,
the ϕp vs. θ and ψmax vs. θ curves are almost symmetrical with re-
spect to θ = 90°. However, asymmetric responses for both the ϕp vs.
θ and ψmax vs. θ curves were attained for the sand alone and sand-
nonwoven geotextile interface tests.
• For the crushed sand-nonwoven geotextile interface, minimums for
ϕp are, respectively, attained around θ = 75° and 165° and a
A. Afzali-Nejad et al. Geotextiles and Geomembranes 46 (2018) 869–881

∗
Fig. 12. Rose diagrams for (ψmax − ψmax ) vs. θ demonstrating direction-dependency of the maximum dilation angle: (a) sand-woven GTX1 interfaces; (b) sand-
woven GTX2 interfaces; (c) sand-nonwoven GTX interfaces; (d) sand-GMB interfaces.

transient maximum for ϕp is observed in between θ = 120° and

135°.
• For all of the inherently anisotropic sand-geosynthetic interfaces
studied here, the critical state (i.e., residual) friction angle was
nearly insensitive to the bedding plane inclination angle. Therefore,
a unique critical state fabric is finally achieved in soil-geosynthetic
interfaces irrespective of the initial (inherent) fabric.
• Compiling the data of 141 interface tests on four different interface
types covering a wide domain of bedding plane inclination angles, it
was observed that the stress-dilation linkage of all interfaces can be
effectively described by ϕp − ϕcs ≈ ψmax irrespective of the interface
type, and bedding plane inclination angle (say inherent anisotropy).
• It was observed that the methodology originally proposed by
Pietruszczak and Mroz (2001) can accurately predict variation of ϕp
with θ for the sand-woven geotextile and sand-geomembrane in-
terfaces provided that, ϕp vs. θ curve is symmetrical with respect to
Fig. 13. Correlation between ϕp - ϕcs and ψmax for the inherently anisotropic
θ = 90°.
crushed sand-geosynthetic interfaces for all of the bedding plane inclination
angles.

879
A. Afzali-Nejad et al. Geotextiles and Geomembranes 46 (2018) 869–881

Table 3
Parameters used in simulation of ϕp vs. θ curves.
Interface type ϕp (θ = 0∘) , [deg.] Ω1, [−] a1, [−] a2, [−]

Crushed sand-woven GTX1 47.193 0.0432 1.792 −167.950

Crushed sand-woven GTX2 45.647 0.0505 −4.390 −125.996
Crushed sand-nonwoven GTX 46.249 0.0079 −322.96 −24698.98
Crushed sand-GMB 39.189 0.0469 12.546 −27.357

Fig. 14. Predicted ϕp vs. θ curves against the experimental data for: (a) sand-woven GTX1 interfaces; (b) sand-woven GTX2 interfaces; (c) sand-nonwoven GTX
interfaces; (d) sand-GMB interfaces.

• The SEM image presented in Fig. 2 indicates that the tested sand
consists of angular and somewhat elongated particles. The specimen
preparation method in conjunction with the particle shape char-
acteristics may impact the orientation of the particles and their as-
sociated contacts. Therefore, the influence of particle shape on the
anisotropic behavior of sand-geotextile interfaces must be further
investigated in future researches.
Acknowledgement
The constructive comments by the chief editor of Geotextiles and
Geomembranes and anonymous reviewers are acknowledged. The au-
thors would like to express their gratitude to Narvin-Gostar-Parsian Co.
and Kia-Pars-Layer (KPL) for providing geotextiles and geomembranes
used in this study.
References
Afzali-Nejad, A., Lashkari, A., Shourijeh, P.T., 2017. Influence of particle shape on the
shear strength and dilation of sand-woven geotextile interfaces. Geotext.
Geomembranes 45, 54–66.
Al-Rkaby, A.H.J., Cheginzadeh, A., Nikraz, H.R., 2017. Anisotropic strength of large scale
geogrid-reinforced sand: experimental study. Soils Found. 57 (4), 557–574.
Anubhav, Basudhar, P.K., 2010. Modeling of soil-woven geotextile interface from direct
shear test results. Geotext. Geomembranes 28, 403–408.
Anubhav, Wu, H., 2015. Modeling of non-linear shear displacement behavior of soil-
geotextile interface. Int. J. Geosynth. Ground Eng. 1, 19.
Athanasopoulos, G.A., 1993. Effect of particle size on the mechanical behavior of soil-
geotextile composites. Geotext. Geomembranes 12 (3), 255–273.
Azami, A., Pietruzczak, S., Guo, P., 2010. Bearing capacity of shallow foundations in
transversely isotropic granular media. Int. J. Numer. Anal. Meth. GeoMech. 34 (8),
771–793.
Choudhary, A.K., Krishna, A.M., 2016. Experimental investigation of interface behavior of
different types of granular soil/geosynthetics. Int. J. Geosynth. Ground Eng. 2, 4.
DeJong, J.T., Westgate, Z.J., 2009. Role of initial state, material properties, and con-
finement condition on local and global soil-structure interface behavior. ASCE J.
Geotech. Geoenviron. Eng. 135 (11), 1646–1660.
Dove, J.E., Jarrett, J.B., 2002. Behavior of dilative sand interfaces in a geotribology
framework. ASCE J. Geotech. Geoenviron. Eng. 128 (1), 25–37.
Esmaili, D., Hatami, K., Miller, G.A., 2014. Influence of matric suction on geotextile re-
sinforcement-marginal soil interface strength. Geotext. Geomembranes 42, 139–153.
Evgin, E., Fakharian, K., 1996. Efffect of stress path on the behavior of sand-steel inter-
face. Can. Geotech. J. 33, 853–865.
Farhadi, B., Lashkari, A., 2017. Influence of soil inherent anisotropy on behavior of
crushed sand-steel interfaces. Soils Found. 57 (1), 111–125.
Ferreira, F.B., Vieira, C.S., Lopes, M.L., 2015. Direct shear behavior of residual soil-geo-
synthetic interfaces- influence of soil moisture content, soil density and geosynthtic
type. Geosynth. Int. 22 (3), 257–272.
Fowmes, G.J., Dixon, N., Fu, L., Zaharescu, C.A., 2017. Rapid prototyping of geosynthetic
interfaces: investigation of peak strength using direct shear tests. Geotext.
Geomembranes 45 (6), 674–687.
Frost, J.D., Kim, D., Lee, S.-W., 2012. Microscale geomembrane-granular material inter-
actions. KSCE J. Civ. Eng. 16 (1), 79–92.
Fu, P., Dafalias, Y.F., 2011. Study of anisotropic shear strength of granular materials using
DEM simulation. Int. J. Numer. Anal. Meth. GeoMech. 35, 1098–1126.
Giroud, J.P., Arman, A., Bell, J.R., Koerner, R.M., Milligan, V., De Mello, V.F.B., 1985.
Geotextiles in geotechnical engineering- practice and research. Geotext.
Geomembranes 2 (3), 179–242.
Gu, X., Chen, Y., Huang, M., 2017. Critical state shear behavior of the soil-structure in-
terface determined by discrete element modeling. Particulogy. https://doi.org/10.
1016/j.partic.2017.02.002. (in press).
Guo, P., 2008. Modified direct shear test for anisotropic strength of sand. ASCE J. Eng.
Mech. 134 (9), 1311–1318.
Guo, N., Zhao, J., 2013. The signature of shear-induced anisotropy in granular media.
Comput. Geotech. 47, 1–15.

880
A. Afzali-Nejad et al.
Hatami, K., Esmaili, D., 2015. Unsaturated soil-woven geotextile interface strength
properties from small-scale pullout and interface tests. Geosynth. Int. 22 (2),
161–172.
Jiang, M., Zhang, A., Fu, Z., 2017. 3-D DEM simulations of drained triaxial tests on in-
herently anisotropic granulates. Eur. J. Environ. Civ. Eng. https://doi.org/10.1080/
19648189.2017.1385541.
Jing, X.-Y., Zhou, W.-H., Zhu, H.-X., Yin, Z.-Y., Li, Y., 2017. Analysis of soil-structural
interface behavior using three-dimensional DEM simulations. Int. J. Anal. Numer.
Meth. Geomech. https://doi.org/10.1002/nag.2745. (in press).
Khoury, C.N., Miller, G.A., Hatami, K., 2011. Unsaturated soil-geotextile interface beha-
vior. Geotext. Geomembranes 29, 17–28.
Lashkari, A., Kadivar, M., 2016. A constitutive model for unsaturated soil-structure in-
terfaces. Int. J. Numer. Anal. Meth. GeoMech. 40, 207–234.
Lee, K.M., Manjunath, V.R., 2000. Soil-geotextile interface friction by direct shear tests.
Can. Geotech. J. 37, 238–252.
Li B, Zeng X, Ming H (2010). Seismic response of retaining wall with anisotropic backfills.
ASCE Geotechnical Special Publication 384 (Proc. of Earth Retention Conference),
Reston, VA, 688–695.
Lings, M.L., Dietz, M.S., 2005. The peak strength of sand-steel interfaces and the role of
dilation. Soils Found. 45 (6), 1–14.
Liu, F.-Y., Wang, P., Geng, X., Wang, J., Lin, X., 2016. Cyclic and post-cyclic behavior
from sand-geogrid interface large-scale direct shear tests. Geosynth. Int. 23 (2),
129–139.
Martinez, A., Frost, J.D., 2017. The influence of surface roughness form on the strength of
sand-structure interfaces. Géotech. Lett. 7, 1–8.
Martinez, A., Frost, J.D., Hebeler, G.L., 2015. Experimental study of shear zones formed at
sands/steel interfaces in axial and torsional axisymmetric tests. ASTM Geotech. Test.
J. 38 (4), 409–426.
Nakata, Y., Hyodo, M., Murata, H., Yasufuku, N., 1998. Flow deformation of sands sub-
jected to rotation of principal stress rotation. Soils Found. 38 (2), 115–128.
Negussey, D., Wijewickreme, W.K.D., Vaid, Y.P., 1989. Geomembrane interface friction.
Can. Geotech. J. 26, 165–169.
Oboudi, M., Pietruszczak, S., Razaqpur, A.G., 2016. Description of inherent and induced
anisotropy in granular media with particles of high sphericity. ASCE Int. J. Geomech.
16 (4). https://doi.org/10.1061/(ASCE)GM.1943-5622.0000635.
Oda, M., 1993. Inherent and induced anisotropy in plasticity theory of granular soils.
Mech. Mater. 16, 35–45.
Oda, M., Koishikawa, I., 1979. Effect of strength anisotropy on bearing capacity of
shallow footing in dense sand. Soils Found. 19 (3), 15–28.
881
Geotextiles and Geomembranes 46 (2018) 869–881
Papadimitriou, A.G., Dafalias, Y.F., Yoshimine, M., 2005. Plasticity modeling of the effect
of sample preparation method on sand response. Soils Found. 45 (2), 109–123.
Pietruszczak, S., Mroz, Z., 2001. On failure criteria for anisotropic cohesive-frictional
materials. Int. J. Numer. Anal. Meth. GeoMech. 25 (5), 509–524.
Punetha, P., Mohanty, P., Samanta, M., 2017. Microstructural investigation on mechan-
ical behavior of soil-geosynthetic interface in direct shear test. Geotext.
Geomembranes 45 (3), 197–210.
Qin, J., Zeng, X., Ming, H., 2016. Centrifuge modeling and the influence of fabric ani-
sotropy on seismic response of foundations. ASCE J. Geotech. Geoenviron. Eng. 142
(3). https://doi.org/10.1061/(ASCE)GT.1943-5606.0001414.
Rodriguez, N.M., Lade, P.V., 2013. True triaxial tests on cross-anisotropic deposits of fine
Nevada sand. ASCE Int. J. Geomech. 13 (6), 779–793.
Tong, Z., Fu, P., Zhou, S., Dafalias, Y.F., 2014. Experimental investigation of shear
strength of sands with inherent fabric anisotropy. Acta Geotech. 9, 257–275.
Vangla, P., Gali, M.L., 2016a. Effect of particle size of sand and surface asperities of
reinforcement on their interface shear behavior. Geotext. Geomembranes 44 (3),
254–268.
Vangla, P., Gali, M.L., 2016b. Shear behavior of sand-smooth geomembrane interfaces
through micro-topographical analysis. Geotext. Geomembranes 44, 592–603.
Verdugo R, de la Hoz K (2007). Strength and stiffness of coarse granular soils. In: Ling HI,
Callisto L, Leshchinsky D, Koseki J (Eds.). Soil Stress-strain Behavior: Measurement,
Modeling and Analysis. Geotechnical Collection of Papers of the Geotechnical
Symposium in Rome, Springer.
Vieira, C.S., Pereira, P.M., 2016. Interface shear properties of geosynthetics and con-
struction and demolition waste from large-scale direct shear tests. Geosynth. Int. 23
(1), 62–70.
Vieira, C.S., Lopes, MdL., Caldeira, L., 2015. Sand-nonwoven geotextile interfaces shear
strength by direct shear and simple shear tests. Geomech. Eng. 9 (5), 601–618.
Xiao, Y., Liu, H., Chen, Y., Chu, J., 2014. Influence of intermediate principal stress on the
strength and dilatancy behavior of rockfill material. ASCE J. Geotech. Geoenviron.
Eng. 140 (11), 04014064.
Xiao, Y., Xiang, J., Liu, H., Ma, Q., 2017. Strength-dilatancy relation of sand containing
non-plastic fines. Géotech. Lett. 7 (2), 1–7.
Xiong, H., Guo, L., Cai, Y., Yang, Z., 2016. Experimental study of drained anisotropy of
granular soils involving rotation of principal stress direction. Eur. J. Environ. Civ.
Eng. 20 (4), 431–454.
Yoshimine, M., Ishihara, K., Vargas, W., 1998. Effect of principal stress direction and
intermediate principal stress on undrained shear behavior of sand. Soils Found. 38
(3), 179–188.