Shi, J. et al. (2021). Géotechnique 71, No. 4, 305–319 [https://doi.org/10.1680/jgeot.18.P.

348]

Anisotropic small-strain stiffness of calcareous sand affected by sample

preparation, particle characteristic and gradation
JINQUAN SHI
, WIM HAEGEMAN† and VEERLE CNUDDE‡

Sands exhibit a directional diversity in shear modulus at small strains, reflecting the anisotropic stiffness
behaviour in the structure. In this paper, sample preparation method, particle shape and particle size are
considered as parameters to produce samples with various initial fabric. Five preparation methods,
namely, air and water pluviation, dry and moist tamping and dry funnel deposition are used to
reconstitute the samples. The stiffness anisotropy of calcareous sand and Mol silica sand is quantified
by the small-strain modulus G0 measured in horizontal and vertical planes with the bender element
technique in triaxial tests. Test results show that calcareous sands behave as an elastic homogeneous
continuum material at small strains. Calcareous samples prepared by the air pluviation method possess
the highest anisotropic ratios. The lowest stiffness anisotropy exists in the samples prepared by moist
tamping and dry funnel deposition methods. Compared with the anisotropic ratios of the air pluviation
samples from the literature, the notably higher values for the calcareous sand in this study are attributed
to the lower sphericity. In addition, less prominent stiffness anisotropy is found in calcareous sands with
smaller particle sizes (D50).

KEYWORDS: anisotropy; laboratory tests; sampling; sands; stiffness

INTRODUCTION respectively, are quantified by equation (1). Commonly, the

Sand often exhibits directivity during its deposition process.
When the geometry of a sand particle is not regular, the long
edge tends to move to become parallel to the horizontal
direction and the short side to the vertical direction. This
deposition characteristic results in an anisotropic fabric of
sandy soils, which spontaneously introduces significant
variations in their engineering properties, for example, the
small-strain modulus G0.
The assessment of this stiffness anisotropy is easily
achieved by conducting shear wave measurement tests,
whose reliability has been proven during recent decades
(Bellotti et al., 1996; Pennington et al., 1997; Fioravante,
2000; Piriyakul & Haegeman, 2007; Ezaoui & Di Benedetto,
2009; Escribano & Nash, 2015). For a homogeneous elastic
medium, the numerical relationship between the shear
modulus and shear wave velocity becomes
G0 ¼ρVS2 ð1Þ
where G0 ðM1 L1 T2 Þ is the maximum shear modulus;
ρðM1 L3 Þ is the mass density of the medium; and
VS ðL1 T1 Þ is the shear wave velocity. In order to evaluate
the stiffness anisotropy, three shear moduli expressed as GHH,
GHV and GVH, where the first and second subscript indicate
the directions of signal propagation and polarisation,
Manuscript received 19 December 2018; revised manuscript
accepted 13 December 2019. Published online ahead of print
22 January 2020.
Discussion on this paper closes on 1 August 2021, for further details
see p. ii.

Department of Civil Engineering, Ghent University,
Technologiepark, Zwijnaarde, Belgium
(Orcid:0000-0003-1700-9127).
† Department of Civil Engineering, Ghent University,
Technologiepark, Zwijnaarde, Belgium
(Orcid:0000-0003-1400-7783).
‡ Department of Geology, Ghent University, Proeftuinstraat, Gent,
Belgium; also Department of Earth Sciences, Utrecht University,
Princetonlaan, CB Utrecht, the Netherlands.
ratio GHH/GHV is a representative parameter to demonstrate
the stiffness anisotropy of soils. Bellotti et al. (1996)
performed shear and compression wave tests on Ticino
sands with geophones in a calibration chamber. The results
under isotropic stress conditions showed a ratio GHH/GHV in
the range from 1·14 to 1·21, corresponding to specimens with
medium and very high densities accordingly. Fioravante
(2000) investigated the anisotropic stiffness behaviour of
Kenya sand and Ticino sand with bender elements. The
averaged values of GHH/GHV are computed as 1·04 and 1·11,
respectively. Other research results for the anisotropic
stiffness ratio are summarised in Table 1. So it can be
concluded that the shear modulus in the horizontal plane is
higher than that in the vertical plane. However, this stiffness
anisotropy ratio is possibly lower than 1, as discussed by
Ishibashi et al. (1991) on Ottawa sand and glass beads. They
tested samples prepared by dry tamping and a lower GHH
compared with GHV was reported.
Previous research reveals that fabric plays a vital role in the
stiffness anisotropy of sand. However, it is quite difficult to
obtain a sample with undisturbed fabric from the site.
Therefore, developing applicable methods for reconstituting
sand samples in the laboratory becomes worthwhile. Several
techniques, such as pluviation, tamping and vibration, are
used to reconstitute granular soil specimens. Studies show
that different sample reconstitution methods result in
different sand fabric and thus the stiffness behaviour at
small strains. Ezaoui & Di Benedetto (2009) showed that
Hostun sand samples prepared by air pluviation and
vibration methods exhibit higher Young’s modulus in the
horizontal direction than in the vertical direction and an
inverse tendency is observed within moist tamping samples.
Escribano & Nash (2015) found that the ratio of GHH/GHV is
slightly smaller in sand specimens prepared by moist tamping
than those reconstituted by air pluviation. Scanning elec-
tronic microscope analysis on Toyoura sand samples
conducted by Yang et al. (2008) confirmed that moist
tamping specimens possess a lower inherent anisotropy
than dry deposition specimens.

305

Downloaded by [ University of Hong Kong] on [03/11/22]. Copyright © ICE Publishing, all rights reserved.
 306 SHI, HAEGEMAN AND CNUDDE
In this study, a calcareous sand and a reference silica sand

1·09–1·16
1·06–1·14
0·87–0·93
EHH/EHV
are used to reconstitute samples at two relative densities in a
triaxial cell. Van Impe et al. (2015) showed the grain

1·28
morphology of the calcareous sand and a very irregular
grain shape was observed. They found that the maximum
and minimum void ratios are beyond the typical range of

1·17–1·23 (86% , DR , 92%)

Anisotropy ratios

1·08–1·14 (30% , DR , 41%)

silica sand. Giang et al. (2017) obtained the sphericity of a
calcareous sand and Mol sand by X-ray tomography and
they found the sphericity of the calcareous sand is lower than
that of the Mol sand. Research conducted by Payan et al.
(2016) and Giang et al. (2017) reveal that sands with lower
sphericity are stiffer due to the stronger intercontact and
rotation resistance between particles. In addition, a lot of
GHH/GHV

1·14–1·21

research has been carried out on the effect of mean grain size
(D50) and the coefficient of uniformity (Cu) on the small-
1·21

1·28

1·04
1·1

1·1
1·1

strain stiffness of sands and it has been demonstrated that Cu

is an important factor affecting the small-strain stiffness,
while the effect of D50 is insignificant (Iwasaki & Tatsuoka,
conditions

Saturation
Saturation

Saturation
1977; Wichtmann & Triantafyllidis, 2009, 2010; Enomoto,
2016; Payan et al., 2016). The research presented above
Dry

Dry

Dry
Dry

Dry
Test

indicates that the investigation of the small-strain stiffness

anisotropy of sandy soils, taking the particle characteristics
and gradation into consideration, is interesting and worth-
Calibration chamber,

Calibration chamber,

while, especially for calcareous sand with a morphology of

Triaxial cell, BE

Triaxial cell, BE
Triaxial cell, BE

Triaxial cell, BE
Triaxial cell, BE

Triaxial cell, BE

low sphericity, as reported by Giang et al. (2017).

Test methods

All tests in this study are performed under isotropic

geophone

geophone

effective stress conditions excluding the effect of stress-

induced anisotropy. Shear wave velocities are measured in
vertical and horizontal directions with the bender element
arrangement method proposed by Fioravante & Capoferri
(2001), which is called a frictional bender element. Air and
Moist tamping
Reconstitution

Air pluviation

Air pluviation

Air pluviation
Air pluviation

Air pluviation
Air pluviation

Air pluviation
Dry vibration

water pluviation, dry and moist tamping and dry funnel

Dry tamping

deposition are chosen as the sample preparation method

methods

(SPM). Finally, the dependency of stiffness anisotropy on

SPM, particle characteristics and gradation of calcareous
sand is examined. In addition, a series of X-ray tomography
tests are performed on the samples prepared by the five
Sphericity

methods. The micromechanism of stiffness anisotropy,

0·732

0·732

0·705
0·743

0·743
0·723

0·705

related to fabric anisotropy, is explained.

—

TEST PROGRAMME
coefficient, Cu

Test materials
Uniformity

Mol sand and six types of calcareous sand are selected in

1·86
1·67
1·22
4·46
1·56
1·42

1·86

1·67
1·6

1·6

this study. Mol sand is a uniform silica sand from the Belgian
province of Antwerp. The calcareous sands include one
natural sand (S1 sand) taken from an artificial island in the
Table 1. Summary of the anisotropy ratios of sands from literature

Persian Gulf and five sands with artificial gradation (SMol,

size D50: mm

SD1, SD2, SCU1 and SCU2). Based on the roundness scale

Mean grain
Materials

proposed by Powers (1953), the Mol sand grain is identified

0·55

0·55
0·13
0·27
0·21
0·47
0·19
0·36

0·13

0·27

to be subangular but with high sphericity and the calcareous

sand is angular with low sphericity. Giang et al. (2017)
reported that the sphericity of the calcareous sand is around
0·5, which is lower than 0·635 of Mol sand measured in the
Hostun sand

same study. Fig. 1 presents the morphology and the particle

Kussharo

size distribution of these sands. SD1 and SD2 sands have the
Toyoura

Toyoura
Hostun
Ticino

Ticino
Kenya

Kenya

same Cu but a different D50, while SMol sand replicates

HRS
Sand

the particle size distribution of Mol sand. SCU1 and SCU2

have the same D50 but different Cu with respect to S1 sand.
The gradations of SMol, SD1, SD2, SCU1 and SCU2 sands
Kuwano & Jardine (2002)

Escribano & Nash (2015)

are adjusted using grains sieved from the original S1 sand.

Chaudhary et al. (2004)
Ezaoui & Di Benedetto

Fioravante et al. (2013)

Yamashita et al. (2005)

The physical properties of these sands are given in Table 2.

Bellotti et al. (1996)

BE, bender element

Fioravante (2000)

Triaxial and bender element tests

All consolidation tests are conducted in a hydraulic
(2009)
Authors

Bishop–Wesley stress path controlled triaxial cell. The

consolidation and back-pressure supply come from two sets
of digital pressure/volume controllers manufactured by GDS.

Downloaded by [ University of Hong Kong] on [03/11/22]. Copyright © ICE Publishing, all rights reserved.
 ANISOTROPIC SMALL-STRAIN STIFFNESS OF CALCAREOUS SAND 307
100 on the sample sides opposite to each other, at the positions of
one third and two thirds of the height. The tips of the
horizontal bender elements are glued to the metal plates
80 attached on the internal surface of the membrane, as
Percentage finer by weight: %

described by Fioravante & Capoferri (2001). The signals

60
are generated and captured by a digital oscilloscope
S1 sand
Picoscope 2206 connected to a personal computer (PC) for
Mol data acquisition. Other equipment consists of an amplifier
40 SMol and signal switch boxes, as shown in Fig. 2.
SD1
SD2
S1
Mol sand
20 SCU1
SCU2
Sample preparation method. Five commonly used sample
reconstituting methods in particular air pluviation, water
0 pluviation, dry funnel deposition, dry and moist tamping are
0·01 0·1 1 10 selected in this study. The deposition and densification steps
Grain diameter: mm in each method are significantly different, thus resulting in a
variation of specimen fabric.
Fig. 1. Particle size distribution of sands

AIR PLUVIATION. The air pluviation method is carried out

The volume change is counted by the back-pressure with a sand raining system usually assembled with a poly
controller simultaneously. Three pairs of bender elements (methyl methacrylate) (PMMA) tube consisting of a sand
with two sizes are assembled to measure the shear waves container on the top and several sieves along the drop height.
propagating vertically and horizontally. One pair 10 mm The relative density of the specimens reconstituted by this
wide and 20 mm long is mounted in the top cap and pedestal, method is controlled by the mass flow rate and dropping
penetrating 5 mm into the sample. The other two pairs, with height (Vaid & Negussey, 1984; Rad & Tumay, 1987; Lo
a smaller size of 5 mm wide and 10 mm long, are mounted Presti et al., 1992; Lagioia et al., 2006). A new sand pluviator

Table 2. Physical properties of sands

Properties S1 Mol SMol SD1 SD2 SCU1 SCU2

Specific gravity, Gs 2·81 2·64 2·81 2·81 2·81 2·81 2·81

D50: mm 0·602 0·179 0·179 0·354 0·709 0·60 0·60
Cu 3·89 1·55 1·55 1·55 1·55 1·70 7·76
Maximum void ratio, emax 1·19 0·90 1·25 1·29 1·52 1·48 1·11
Minimum void ratio, emin 0·73 0·56 0·81 0·85 0·98 0·97 0·69
Fine content (diameter , 0·063 mm): % 1·13 0·12 0·12 0 0 0 6

PC

Picoscope
(signal generator)

Data acquisition
panel

Load cell
Signal Pore-water pressure
amplifier LVDT

VH
VH HV
Switch Cell pressure
Switch HV HH box
box HH

Back-pressure/volume

Axial pressure

Fig. 2. Arrangement of the triaxial system with bender elements (LVDT, linear variable differential transducer)

Downloaded by [ University of Hong Kong] on [03/11/22]. Copyright © ICE Publishing, all rights reserved.
 308 SHI, HAEGEMAN AND CNUDDE

Funnel

PMMA tube Funnel opening

Sieves Funnel

Frame and sand Sieves

collector Drop height
PVC tube
Mould
Collected
Mould
sands
Holding frame

Triaxial Triaxial
pedestal pedestal

(a) (b)

Fig. 3. Schematic diagram of the pluviators: (a) air pluviation; (b) water pluviation (PVC, polyvinyl chloride)

is designed to fit the triaxial cell, as illustrated in Fig. 3(a). of dropping height on relative density is minor. The
Before carrying out the tests, the mass of the pluviator relative density obtained using the air pluviation method is
is determined. Then, the weighted sand is poured into the selected as the reference for samples prepared by other
funnel and a uniform sand raining is formed passing five methods.
sieves. The sand falling outside the mould is collected by
the pluviator. Finally, after levelling the surface of the
sample, the pluviator together with the collected sand are WATER PLUVIATION. Vaid & Negussey (1988) showed that
weighed so the mass of the sample is obtained. Referring to samples prepared with the water pluviation method exhibit a
research conducted by Rad & Tumay (1987), the drop height more homogeneous fabric than those prepared using other
is defined as the distance between the bottom sieve and the methods. In this study, in order to obtain a sand rain, a
top surface of the specimen, and a change in the distance smaller scale pluviator is designed in the same way as
from the funnel opening to the top sieve has no influence on described for the air pluviation method, as shown in
the relative density of the sample. In this study, no attention is Fig. 3(b). With this pluviator, sands of predetermined mass
paid to controlling the drop height, so the pluviator is fixed are pluviated into the mould filled with de-aired water
on the triaxial pedestal. The mass flow rate is controlled by beforehand. It is concluded that the relative densities of
the funnel opening size. The effects of funnel opening size water pluviated specimens are neither influenced by mass
and dropping height on the relative density for both S1 sand flow rate nor dropping height due to the reduction of impact
and Mol sand are investigated and shown in Fig. 4. It is energy (Vaid & Negussey, 1988; Lagioia et al., 2006).
found that, compared with the mass flow rate, the influence Therefore, the initial density of the water pluviated

100 100
Drop height: 12·4 cm

80 80
Relative density: %

Relative density: %

60 60 Funnel opening size

S1 14 mm
S1 22 mm
40 40 S1 30 mm
Mol 14 mm
Mol 22 mm
Mol 30 mm
20 Mol sand 20
S1 sand

0 0
0 10 20 30 40 50 0 10 20 30 40 50
Funnel opening size: mm Drop height: cm
(a) (b)

Fig. 4. Factors affecting relative density using the air pluviation method: (a) funnel opening size; (b) drop height

Downloaded by [ University of Hong Kong] on [03/11/22]. Copyright © ICE Publishing, all rights reserved.
 ANISOTROPIC SMALL-STRAIN STIFFNESS OF CALCAREOUS SAND
specimen is loose and the densification is achieved by
vibration. In this study, the vibration energy is given by
slightly tapping the sides of the mould with a rubber hammer.
DRY AND MOIST TAMPING. A cylindrical tamper with a
diameter matched to the split mould is employed for dry
and moist tamping. For the moist tamping method, sand
with 5% water content by weight is tamped in ten equal layers
into the mould. The layer number is decreased to five and the
tamping time is reduced in the dry tamping method because
of the lack of capillary force existing in the moist tamped
sands. The under-compaction method proposed by Ladd
(1978) is used to achieve a homogeneous compaction. In
addition, the compacting force is carefully controlled to avoid
crushing in the S1 sand, owing to its fragility.
DRY FUNNEL DEPOSITION. The dry funnel deposition is
performed by keeping the opening of a long-necked funnel
at zero height to the sand surface and raising the funnel
continuously until the target weighted sand is poured into the
mould entirely (Yamamuro & Wood, 2004; Li et al., 2018).
The densification and intended sample height are achieved
by tapping the side of the mould slightly. The benefit of this
method is that it prevents the particle segregation of sands
containing fines.
Test procedure. Samples of 50 mm dia. and 100 mm high
are prepared in a split mould. After setting up the specimens,
a vacuum of 20 kPa followed by a positive isotropic cell
pressure 20 kPa is applied to the samples. The increment of
cell pressure is applied while decreasing the vacuum
simultaneously. In order to saturate the samples efficiently,
carbon dioxide and de-aired water are percolated through the
samples from bottom to top for 30 min and 1 h successively.
Then the back-pressure and cell pressure are increased to
200 kPa and 220 kPa synchronously to maintain the effective
consolidation stress at 20 kPa. The saturation with a B-value
309
larger than 0·95 is achieved after sustaining the back-pressure
for 1 h. For specimens prepared by the water pluviation
method, the drainage system is saturated in advance.
De-aired water is percolated through the samples as
described above. However, the back-pressure is increased to
350 kPa and held for 2 h to achieve the target saturation. In
the next stage, samples are subjected to an effective
consolidation stress between 50 kPa and 400 kPa with an
increment step of 50 kPa. Bender element tests are performed
after a resting period of 30 min within each step. At the end
of the test, the tested sands are collected to examine the
particle crushing. A summary of all the tests is provided in
Table 3. The repeatability of the test results was examined by
repeating the tests at least two times. For the sake of clarity,
only one test at a time is discussed.
In order to select the optimal signal excitation frequency
and interpretation method, sinusoidal pulse waves at
frequencies of 10, 15 and 20 kHz are excited with a double
amplitude of 20 V. Three signal interpretation methods,
named start–start (SS), peak–peak (PP) and cross-correlation
(CC) are compared. Fig. 5 shows the waves and the estimated
travel times at the effective consolidation pressure of 150 kPa
in the air pluviation sample. Ogino et al. (2015) showed that
the CC method is not effective when the first peak and the
maximum peak are not coincident due to the reflected
P waves. In this study, an inconsistency between the first
peak and the maximum peak of CC is found in the HH and
HV waves. Although the SS is frequently used in the
literature, the determination of the shear wave travel time is
subjective and highly affected by the near-field effect.
Considering the double peaks of CC and the objectivity,
the PP method is finally selected for the signal interpretation.
For the signal excitation frequency, Yamashita et al. (2005)
used the averaged value from the estimated travelling times
of 10, 15 and 20 kHz. In this study, it was found that the
travel times decrease slightly with an increase in the
frequency for the HH and VH waves. For HV waves, no
difference is found. Therefore, to reach a compromise,
15 kHz finally is used as the signal excitation frequency for
all samples.

Table 3. Summary of laboratory tests

Test name Material Sample preparation methods Confining pressure: kPa Relative density: %

Series-S1-60 S1 Air pluviation 400 60

S1 Moist tamping 400 60
S1 Dry funnel deposition 400 60
S1 Dry tamping 400 60
S1 Water pluviation 400 60
Series-S1-40 S1 Air pluviation 400 40
S1 Moist tamping 400 40
S1 Dry funnel deposition 400 40
S1 Dry tamping 400 40
S1 Water pluviation 400 40
Series-Mol-60 Mol Air pluviation 400 60
Mol Moist tamping 400 60
Mol Dry funnel deposition 400 60
Mol Dry tamping 400 60
Mol Water pluviation 400 60
Series-Mol-40 Mol Air pluviation 400 40
Mol Moist tamping 400 40
Mol Dry funnel deposition 400 40
Mol Dry tamping 400 40
Mol Water pluviation 400 40
Series-SMol-60 SMol Air pluviation 400 60
Series-SD1-60 SD1 Air pluviation 400 60
Series-SD2-60 SD2 Air pluviation 400 60
Series-SCU1-60 SCU1 Air pluviation 400 60
Series-SCU2-60 SCU1 Air pluviation 400 60

Downloaded by [ University of Hong Kong] on [03/11/22]. Copyright © ICE Publishing, all rights reserved.
 310 SHI, HAEGEMAN AND CNUDDE

Peak to peak
Start to start
Peak to peak
Start to start
10 kHz
10 kHz

Start to start: 0·137456 ms

Peak to peak: 0·139528 ms First peak Start to start: 0·174944 ms
Maximum peak
First peak Peak to peak: 0·179696 ms

Cross-correlation
10 kHz 10 kHz
Cross-correlation
Maximum peak
Normalised signals

Normalised signals
15 kHz
15 kHz

Start to start: 0·138336 ms

Peak to peak: 0·139392 ms Start to start: 0·170248 ms
Peak to peak: 0·179344 ms

Cross-correlation
15 kHz
15 kHz
Cross-correlation
First and maximum peak coincide

20 kHz

20 kHz Start to start: 0·165968 ms

Peak to peak: 0·179520 ms
Start to start: 0·134992 ms
Peak to peak: 0·134992 ms 20 kHz
Cross-correlation
20 kHz
Cross-correlation First and maximum peak coincide

0 0·1 0·2 0·3 0·4 0·5 0·6 0·7 0 0·1 0·2 0·3 0·4 0·5 0·6 0·7
Time: ms Time: ms
(a) (b)

Peak to peak

Start to start

10 kHz

Start to start: 0·328592 ms

Peak to peak: 0·341088 ms

Cross-correlation
10 kHz
First and maximum peak coincide
Normalised signals

15 kHz

Start to start: 0·332288 ms

Peak to peak: 0·336336 ms
Cross-correlation
15 kHz
First and maximum peak coincide

20 kHz

Start to start: 0·324016 ms

Peak to peak: 0·330528 ms

Cross-correlation
20 kHz
First and maximum peak coincide

0 0·1 0·2 0·3 0·4 0·5 0·6 0·7 0·8 0·9 1·0
Time: ms
(c)

Fig. 5. Travel time evaluation at the effective consolidation pressure of 150 kPa in the air pluviation S1 sample: (a) HH waves; (b) HV waves;
(c) VH waves

Downloaded by [ University of Hong Kong] on [03/11/22]. Copyright © ICE Publishing, all rights reserved.
 ANISOTROPIC SMALL-STRAIN STIFFNESS OF CALCAREOUS SAND
X-ray tomography
The high-energy computed tomography optimised for
research (Hector) at the X-ray Tomography Centre of
Ghent University (UGCT) is used in this study for measuring
the three-dimensional (3D) fabric of the calcareous sand
(Masschaele et al., 2013). The scanned samples are prepared
by the five methods as described above and in a PMMA
mould of 50 mm dia. and 120 mm high at 60% relative
density. One intact sample image is merged by three scans
conducted along the height under the atmosphere pressure.
The commercial software Octopus, originating from UGCT,
is adopted for the image reconstruction and analysis
(Vlassenbroeck et al., 2007). To quantify the fabric aniso-
tropy, the central vertical slice from the 3D reconstruction
image is used and the inclination angles of sand grains with
respect to the horizontal plane are calculated.
TEST RESULTS
Fabric anisotropy
There is a precondition that crushing, as a vital behaviour
for calcareous sand, is not present in this study. Careful
examination of the particle size gradation of the tested sands
was carried out. The results of the S1 sand samples made by
different sample preparation methods are shown in Fig. 6. It
can be seen that the gradation lines nearly overlap, meaning
that crushing is insignificant in the current stress range and
not affected by the sample preparation methods. A similar
conclusion regarding the crushability of the calcareous sands
was also given by Giang et al. (2017).
Figures 7 and 8 plot the small-strain shear modulus GHH,
GHV and GVH against the effective consolidation stresses for
the five preparation methods on both S1 and Mol sands at
60% relative density (Dr). It is seen in Fig. 7 that, for S1 sand,
GHH is remarkably higher than GHV and GVH. The result
is consistent with the widely recognised conclusions of
Bellotti et al. (1996), Fioravante (2000), Ezaoui &
Di Benedetto (2009) and Escribano & Nash (2015) that
sands always exhibit higher stiffness in the horizontal
plane than that in the vertical plane. The almost identical
values of GHV and GVH show that S1 sand samples behave as
an elastic homogeneous continuum, which is nevertheless
not detected in Mol sand. In Fig. 8, GHH has values close
to GHV, while GVH is apparently lower. Although similar
observations on sands are reported by Kuwano (1999),
Kuwano & Jardine (2002) and Escribano & Nash (2015),
the explanation for these differences is still not fully clear.
Kuwano (1999) stated that the differences between GHV and
GVH are due to the differences in particle characteristics.
100
80
Percentage finer by weight: %
60
Dry funnel deposition
40 Water pluviation
Dry tamping
Air pluviation
20 Moist tamping
Original S1 sand
0
0·01 0·1 1 10
Grain diameter: mm
Fig. 6. Crushing examination on the tested S1 sands
Downloaded by [ University of Hong Kong] on [03/11/22]. Copyright © ICE Publishing, all rights reserved.
311
Escribano & Nash (2015) attributed the differences to the
geometric and boundary effects and the heterogeneous
fabric. From the observation in this study, it is assumed
that the inconsistency between GHV and GVH for S1 and
Mol sands can partly be attributed to the particle shape,
since S1 sand has a lower sphericity than Mol sand. In
addition, the inequality between GHV and GVH of S1 samples
in moist tamping proves that the effect of the sample
preparation method on the anisotropic behaviour of soils is
non-negligible.
Effect of SPM on the shear modulus at small strains
In order to compare the shear modulus measured on
samples reconstituted using the different methods, all GHH,
GHV and GVH are normalised by the corresponding values
obtained from air pluviation samples, as plotted in Fig. 9. It
is seen that the differences induced by the sample preparation
methods on G0 are directionally dependent. Air pluviation
produces a higher value in GHH. Both the GHH and GHV in
the moist tamping sample are in the highest range. Although
an intermediate value of GVH in the S1 moist tamping
specimens is observed, compared with other methods, it can
still be concluded that the moist tamping method results in
the stiffest fabric among the five preparation methods. It can
be seen that the small-strain stiffness given by the dry funnel
deposition sample is always in the lower ranges for both
sands.
Although noticeable differences in G0 are obtained in this
study considering SPM, Tatsuoka et al. (1979) concluded,
from resonant column (RC) tests, that the effect of SPM on
G0 is negligible. The inconsistent results can be ascribed to
the test method, as described by Gu et al. (2015). They stated
that G0 measured in RC tests is a global stiffness within the
whole sample; however, the results from bender element
testing possess high directionality. In this study, the different
stiffness behaviour of S1 sand on horizontal and vertical
planes also proves that the effect of SPM on the shear moduli
in different directions is not the same. In addition, compared
with samples prepared in dry conditions, moist tamping
samples possess higher capillarity forces between the
particles. A higher tamping force applied for a given target
density results in more stable and more numerous contacts
among particles, causing higher stiffness. Yamamuro &
Wood (2004) analysed the particle contact of dry funnel
deposition samples under a microscope and verified the
theory according to which vibration and low-energy depo-
sition produce more unstable grain contacts. In accordance
with this research, it is reasonable to obtain a lower stiffness
in the dry funnel deposition samples for both sands in this
study.
Effect of spm on stiffness anisotropy
The ratio between GHH and GHV as a result of propagation
in the same direction with different polarisation directions
can provide an insight regarding the stiffness anisotropy. The
ratio of GHH/GHV for samples prepared by different methods
is presented in Fig. 10. It is seen that samples prepared by the
air pluviation method result in the most pronounced stiffness
anisotropy compared with those reconstituted by other
methods. The ratios of GHH/GHV at different consolidation
pressures are averaged at 1·59 and 1·07 for S1 and Mol
samples at 60% relative density, respectively. It is also seen
that, for the samples at 60% relative density, the lowest
anisotropy ratios are produced by dry funnel deposition and
moist tamping for S1 sands and the averaged values of the
anisotropic ratios at different consolidation pressures for
both methods decrease to around 1·26. For Mol sand, a value
 312 SHI, HAEGEMAN AND CNUDDE
350 350

300 300

250 250

200 200
G0: MPa

G0: MPa
150 150

100 100
HH HH
HV HV
50 VH 50 VH

0 0
0 50 100 150 200 250 300 350 400 450 0 50 100 150 200 250 300 350 400 450
Effective consolidation stress: kPa Effective consolidation stress: kPa

(a) (b)

350 350

300 300

250 250

200 200
G0: MPa

G0: MPa

150 150

100 HH 100 HH
HV HV
VH VH
50 50

0 0
0 50 100 150 200 250 300 350 400 450 0 50 100 150 200 250 300 350 400 450
Effective consolidation stress: kPa Effective consolidation stress: kPa

(c) (d)

350

300

250

200
G0: MPa

150

100
HH
HV
50 VH

0
0 50 100 150 200 250 300 350 400 450
Effective consolidation stress: kPa

(e)

Fig. 7. G0 plotted against effective consolidation stress: S1 sand, Dr = 60%: (a) dry tamping; (b) dry funnel deposition; (c) air pluviation; (d) water
pluviation; (e) moist tamping

lower than 1 is observed for the moist tamping sample at Differences in the stiffness anisotropy produced by
60% relative density and only the air pluviation and dry different SPM are discussed in several previous research
tamping methods give ratios higher than 1 at 40% relative studies. It is recognised that the moist tamping method
density. results in higher stiffness in the vertical plane and reduces the

Downloaded by [ University of Hong Kong] on [03/11/22]. Copyright © ICE Publishing, all rights reserved.
 ANISOTROPIC SMALL-STRAIN STIFFNESS OF CALCAREOUS SAND 313
300 300

250 250

200 200
G0: MPa

G0: MPa
150 150

100 100
HH HH
HV HV
50 VH 50 VH

0 0
0 50 100 150 200 250 300 350 400 450 0 50 100 150 200 250 300 350 400 450
Effective consolidation stress: kPa Effective consolidation stress: kPa

(a) (b)

300 300

250 250

200 200
G0: MPa

G0: MPa

150 150

100 100
HH HH
HV HV
50 VH 50 VH

0 0
0 50 100 150 200 250 300 350 400 450 0 50 100 150 200 250 300 350 400 450
Effective consolidation stress: kPa Effective consolidation stress: kPa

(c) (d)

300

250

200
G0: MPa

150

100
HH
HV
50 VH

0
0 50 100 150 200 250 300 350 400 450
Effective consolidation stress: kPa

(e)

Fig. 8. G0 plotted against effective consolidation stress: Mol sand, Dr = 60%: (a) dry tamping; (b) dry funnel deposition; (c) air pluviation;
(d) water pluviation; (e) moist tamping

inherent anisotropy (Yang et al., 2008; Ezaoui & Di moist tamping samples of Ham River sand, while, with the
Benedetto, 2009; Escribano & Nash, 2015). Escribano & same sand, a higher value of about 1·10 is proposed by
Nash (2015) obtained a value of GHH/GHV of about 1·04 from Kuwano & Jardine (2002) for air pluviation samples.

Downloaded by [ University of Hong Kong] on [03/11/22]. Copyright © ICE Publishing, all rights reserved.
 314 SHI, HAEGEMAN AND CNUDDE
1·4 1·4
1·3 1·3
1·2 1·2
1·1 1·1
1·0 1·0
0·9 0·9
G0/G0 AP

G0/G0 AP
0·8 0·8
0·7 Dry tamping 0·7 Dry tamping
Dry funnel deposition Dry funnel deposition
0·6 0·6 Air pluviation
Air pluviation
0·5 Water pluviation 0·5 Water pluviation
Moist tamping Moist tamping
0·4 0·4
0·3 0·3
0 50 100 150 200 250 300 350 400 450 0 50 100 150 200 250 300 350 400 450
Effective consolidation stress: kPa Effective consolidation stress: kPa
(a) (b)

1·4 1·4

1·3 1·3
1·2
1·2
1·1
1·1
1·0
1·0 0·9
G0/G0 AP

G0/G0 AP
0·9 0·8
Dry tamping 0·7 Dry tamping
0·8 Dry funnel deposition
Dry funnel deposition
Air pluviation 0·6 Air pluviation
0·7
Water pluviation 0·5 Water pluviation
0·6 Moist tamping Moist tamping
0·4
0·5 0·3
0 50 100 150 200 250 300 350 400 450 0 50 100 150 200 250 300 350 400 450
Effective consolidation stress: kPa Effective consolidation stress: kPa
(c) (d)
1·4 1·4

1·3 1·3

1·2 1·2

1·1 1·1
G0/G0 AP

1·0 1·0
G0/G0 AP

0·9 0·9
Dry tamping Dry tamping
0·8 Dry funnel deposition 0·8
Dry funnel deposition
Air pluviation Air pluviation
0·7 0·7
Water pluviation Water pluviation
0·6 Moist tamping Moist tamping
0·6

0·5 0·5
0 50 100 150 200 250 300 350 400 450 0 50 100 150 200 250 300 350 400 450
Effective consolidation stress: kPa Effective consolidation stress: kPa
(e) (f)

Fig. 9. Effect of SPM on G0 for S1 sand and Mol sand at Dr = 60%: (a) HH S1; (b) HH Mol; (c) HV S1; (d) HV Mol; (e) VH S1; (f) VH Mol

Yang et al. (2008) found the particle orientations from the
moist tamping samples are more random and an isotropic
fabric was observed. Combining the results from this study
with previous conclusions, the variation in stiffness aniso-
tropy can be explained as follows. In this study, it was
observed that the uniform sand rain avoids the cone-
shaped build-up in the air pluviation method, providing a
plane for the particle horizontal alignment. During the
deposition of the pluviated sands, the grains reach a stable
state under gravity. The continuous deposition process
provides sufficient time and space for the particle rearrange-
ment and an approximately horizontal bedding plane is
formed, as illustrated in Fig. 11(a). Based on the conclusion
of Yamashita et al. (2005) that shear waves propagate faster in
the plane parallel to the bedding plane than in the normal
one, combined with equation (1), the highest stiffness
anisotropy of the air pluviation sample in this study can be
explained. Although the pluviation technique is carried out in
the water pluviation method, it is deemed that the deposited
particles are not aligned horizontally ideally compared with
the grains in the air pluviation sample. In this study, in order
to deposit sands in a similar way in the air and water
pluviation methods, the usual operation in the water pluvia-
tion method by keeping the funnel opening beneath the water
surface and lifting it upward slowly is not adopted. The
uniform sand depositing at a height generates a certain

Downloaded by [ University of Hong Kong] on [03/11/22]. Copyright © ICE Publishing, all rights reserved.
 ANISOTROPIC SMALL-STRAIN STIFFNESS OF CALCAREOUS SAND 315
1·8 1·2
1·7
1·6 1·1

1·5
1·0
1·4
1·3 0·9
1·2
GHH/GHV

GHH/GHV
1·1 0·8
1·0
0·9 Dry tamping 0·7 Dry tamping
Dry funnel deposition Dry funnel deposition
0·8
Air pluviation 0·6 Air pluviation
0·7 Water pluviation
Water pluviation
0·6 Moist tamping 0·5 Moist tamping
0·5
0·4 0·4
0 50 100 150 200 250 300 350 400 450 0 50 100 150 200 250 300 350 400 450
Effective consolidation stress: kPa Effective consolidation stress: kPa
(a) (b)

1·8 1·2
1·7
1·6 1·1

1·5
1·0
1·4
1·3 0·9
1·2
GHH/GHV

GHH/GHV

1·1 0·8
1·0
0·9 Dry tamping 0·7
Dry tamping
0·8 Dry funnel deposition Dry funnel deposition
Air pluviation 0·6 Air pluviation
0·7
Water pluviation Water pluviation
0·6 Moist tamping 0·5 Moist tamping
0·5
0·4 0·4
0 50 100 150 200 250 300 350 400 450 0 50 100 150 200 250 300 350 400 450
Effective consolidation stress: kPa Effective consolidation stress: kPa
(c) (d)

Fig. 10. Stiffness anisotropy ratio plotted against effective consolidation stress: (a) S1 sand, Dr = 60%; (b) Mol sand, Dr = 60%; (c) S1 sand,
Dr = 40%; (d) Mol sand, Dr = 40%

dry funnel deposition method, the almost zero dropping

Funnel
increasing
Space
(a)
Fig. 11. Illustration of the deposition process: (a) air pluviation;
(b) dry funnel deposition
amount of kinetic energy and gives rise to the water
disturbance, which affected the particle settlement in
reverse, resulting in a less anisotropic fabric than that in the
air pluviation sample. During the deposition of sands in the
height limits time and space and the particle movement is
insufficient and not comparable with that in the air pluviation
method, as shown in Fig. 11(b). In addition, the vibration
applied to the dry funnel deposition samples, which is
required for densification, also plays a negative role in the
stiffness anisotropy. In the literature review given by Börzsönyi
& Stannarius (2013), they summarised that the elongated
grains tend to be vertically aligned after experiencing high
intense vibration; for example, a peak acceleration of 3g or
many thousands of taps. In the dry funnel deposition sample,
although the vibration is lower (around 100 taps), it is deemed
Movement that the grain rotation occurs and leads to the decrease in the
restricted fabric anisotropy. For the moist tamping specimen, there is a
directional and higher energy input during the sample
reconstruction, as mentioned earlier. Kuwano & Jardine
(b) (2002) indicated that granular media develop their internal
fabric to match the applied load, providing the preferred
pathways for wave propagation. Similarly, in this study, the
tamping load in the moist tamping method increases the
vertical stiffness and allows the shear waves to propagate faster
than those in the vertical planes of the other samples. In the
dry tamping sample, the tamping effort is reduced greatly due
to the lack of capillary force, resulting in a lower vertical
stiffness and a higher stiffness anisotropy compared with those
in the moist tamping sample.

Downloaded by [ University of Hong Kong] on [03/11/22]. Copyright © ICE Publishing, all rights reserved.
 316
X-ray image analysis
Looking to the more distinct stiffness anisotropy behav-
iour of S1 sands compared with that of Mol sand, the direct
visual investigation of the fabric of S1 sands at 60% relative
density was carried out based on X-ray tomography. The
fabric anisotropy was quantified by measuring the incli-
nation angle of the major axis of the sand grain. Fig. 12
shows an example vertical section from the 3D
SHI, HAEGEMAN AND CNUDDE
reconstruction image of the air pluviation sample and the
major axes of the visible grains are labelled. Six randomly
selected captures on one slice are used for quantifying the
particle orientation for a sample preparation method.
Finally, the angle distribution in the five samples is obtained,
as shown in Fig. 13. It is deemed that an even angle
distribution within 0° to 180° represents the isotropic fabric.
In this study, quantifying the grain orientation in three

Z Z

Y X

Fig. 12. Vertical section and capture from X-ray image of the air pluviation sample

100 90 80 Particle number: 1089 100 90 80 Particle number: 1166

18 110 70 18 110 70
120 60 120 60
16 130 50 16 130 50
14 14
Percentage: %

Percentage: %

140 40 140 40
12 12
10 150 30 10 150 30
8 8
160 20 160 20
6 6
4 170 10 4 170 10
2 2
0 180 0 0 180 0
(a) (b)

100 90 80 Particle number: 1858 100 90 80 Particle number: 1180

18 110 70 18 110 70
120 60 120 60
16 130 50 16 130 50
14 14
Percentage: %

Percentage: %

140 40 140 40
12 12
10 150 30 10 150 30
8 8
160 20 160 20
6 6
4 170 10 4 170 10
2 2
0 180 0 0 180 0
(c) (d)

100 90 80 Particle number: 1789 100 90 80

18 110 70 18 110 70
120 60 120 60
16 130 50 16 130 50
14 14
Percentage: %

Percentage: %

140 40 140 50°<Angle<130° 40

12 12 Vertical align
10 150 30 10 150 30
8 8
160 20 160 20
6 6
4 170 10 4 170 140°<Angle<180° 0°<Angle<40° 10
2 2 Horizontal align Horizontal align
0 180 0 0 180 0
(e) (f)

Fig. 13. The inclination angle distribution of S1 sand samples prepared by the five methods: (a) air pluviation; (b) water pluviation; (c) dry
tamping: (d) moist tamping; (e) dry funnel deposition; (f) angle ranges of the particle alignment

Downloaded by [ University of Hong Kong] on [03/11/22]. Copyright © ICE Publishing, all rights reserved.
 ANISOTROPIC SMALL-STRAIN STIFFNESS OF CALCAREOUS SAND 317
Table 4. Percentages of particles preferring horizontal and vertical alignments for S1 sand samples

Particle inclination angle Percentages: %

Air pluviation Water pluviation Dry tamping Moist tamping Dry funnel deposition

50° , Angle , 130° 14·78 24·79 31·02 32·54 43·01

0° , Angle , 40°, 140° , Angle , 180° 80·07 64·84 58·27 56·19 47·26

1·9 2·2
1·8 2·1
1·7 2·0
1·6 1·9
1·5 1·8
1·4
1·7
1·3
1·6

GHH/GHV
GHH/GHV

1·2
1·5
1·1
1·4
1·0
1·3
0·9 SMol SCU1
0·8 SD1 1·2 S1
0·7 SD2 1·1 SCU2
0·6 Mol 1·0
0·5 0·9
0·4 0·8
0 50 100 150 200 250 300 350 400 450 0 50 100 150 200 250 300 350 400 450
Effective consolidation stress: kPa Effective consolidation stress: kPa
(a) (b)

Fig. 14. Effect of D50 and Cu on the stiffness anisotropy of calcareous sand: (a) D50; (b) Cu

dimensions is unattainable since the particle segregation is
quite hard under the current scan resolution, limited by
the large sample size and the degree of gradation of the S1
sands. However, Yang et al. (2008) proved that evaluating
the fabric anisotropy by two-dimensional images is practic-
able. Under the limited resolution, labelling some grains
requires a visual judgement and the smallest detectable
grain diameter is about 0·3 mm. In addition, the differences
in the analysed particle number, as seen in Fig. 13, are
attributable to the difficulty of variation in particle recog-
nition because the image quality is different. However, after
trial and error, it is found that sufficiently objective and
consistent results are obtained when the grain number
exceeds 1000.
It is seen in Fig. 13 that particles tending to a horizontal
alignment play the dominant role in the air pluviation
sample. In the water pluviation sample, the vibration
performed for densification and the disturbance induced by
the water flow result in a slightly more isotropic fabric than
that in the air pluviation sample. It is interesting to note that
the dry funnel deposition method produces more isotropic
sample fabric than the moist tamping method, while, in
Fig. 10(a), the difference in anisotropic ratios between the
moist tamping and dry funnel deposition samples is insig-
nificant. It is deemed that the tamping force in the moist
tamping method increases the vertical stiffness and decreases
the stiffness anisotropy. From Table 4, it is seen that the
amount of particles tending to be horizontally aligned in the
dry tamping sample is slightly greater than those of the moist
tamping sample. In Fig. 10(a), the stiffness anisotropy ratio
of the dry tamping sample is higher than that of the moist
tamping sample. This is explained in that, as mentioned
above, the reduced tamping force in the dry tamping method,
compared with that in the moist tamping method, leads to a
smaller increment of the vertical stiffness, resulting in a
higher stiffness anisotropy.
Effect of particle characteristics and gradation on the stiffness
anisotropy of calcareous sand
In order to clarify the effect of the particle characteristics
and gradation on the stiffness anisotropy, identical tests are
performed on SMol, SD1, SD2, SCU1 and SCU2 sands with
the air pluviation method. Fig. 14 shows the anisotropy
ratios of these sands under isotropic consolidation pressures.
It is seen that the sands (SMol, SD1 and SD2) with larger D50
but identical Cu show higher anisotropy ratios, reflecting that
the increased particle size enhances the stiffness anisotropy.
However, the effect of Cu on the stiffness anisotropy is
insignificant.
It is interesting to note that the ratios of GHH/GHV of the six
calcareous sands are significantly higher than those of sands
in Table 1. Attention is paid to the particle shape. Aspect
ratio and sphericity are two main parameters to quantify the
particle shape. Aspect ratio is defined as the ratio between
Feret minimum and Feret maximum diameters, as illustrated
in Fig. 15. Sphericity is the ratio of the surface area of a
sphere having the same volume as a particle to the actual area
of the particle. Gao et al. (2015) mention that the low aspect
ratio of a sand particle enhances the inherent stiffness
Feret maximum diameter
Feret minimum diameter
Fig. 15. Feret diameters

Downloaded by [ University of Hong Kong] on [03/11/22]. Copyright © ICE Publishing, all rights reserved.
 318 SHI, HAEGEMAN AND CNUDDE
Table 5. Summary of sphericity of sands used in this study
Sand S1 SMol SD1
D50: mm 0·602 0·179 0·354
Sphericity 0·454 0·485 0·472
anisotropy. From the database summarised by Altuhafi et al.
(2016), a positive correlation between aspect ratio and
sphericity is found. Therefore, it is deduced that sands with
lower sphericity produce higher stiffness anisotropy. In this
study, in order to obtain the sphericity of these sands, another
series of X-ray tomography tests was performed on the grains
sieved from S1 sands by diameter. The analysed particle
number was selected as 300 for each diameter group. Then
the sphericity of the six calcareous sands was calculated
according to the mass percentages of the sieved grains and
the values are listed in Table 5. For this calcareous sand, an
averaged value of around 0·46, which is relatively lower than
the sphericity of sands in Table 1, is obtained. In addition, in
Fig. 14, the anisotropy ratio of SMol sand is higher than that
of Mol sand, proving that a decrease in sphericity enhances
the stiffness anisotropy. Giang et al. (2017) showed that the
sphericity of the calcareous sands decreases with the
increasing of particle size, which is also observed in Table 5.
CONCLUSIONS
This paper investigates the effects of sample preparation
method, particle characteristics and gradation on the
inherent stiffness anisotropy of calcareous sands. GHH, GHV
and GVH are obtained by conducting bender element tests in
a triaxial cell under isotropic stress conditions. From the test
results, the following conclusions are drawn.
(a) The calcareous S1 sands behave as homogeneous
continuum media, showing almost identical GHV and
GVH. This is not observed in Mol sand. The stiffness
anisotropy behaviour of S1 sands is more significant
than that of the Mol sand.
(b) Among the five sample preparation methods used,
moist tamping produces the highest stiffness at 60%
relative density. The dry funnel deposition sample
always gives the lower stiffness among the five methods.
(c) For calcareous sand, air pluviation samples exhibit the
highest anisotropy ratios. Moist tamping increases the
stiffness in the vertical plane so reduces the stiffness
anisotropy. The anisotropy ratio from dry funnel
deposition is in the lowest range. Particle alignment,
measured by X-ray tomography, explains these
differences in stiffness anisotropy.
(d ) Sands with lower sphericity show higher inherent
stiffness anisotropy. The lower anisotropy ratios
observed in sands with smaller D50 are attributed to the
higher sphericity exhibited by the smaller particles. The
effect of Cu on the stiffness anisotropy is negligible.
ACKNOWLEDGEMENTS
The first author acknowledges the financial support from
the China Scholarship Council (no. 201607040079). The
assistance from Filip Van Boxstael and Jan Van der Perre of
the Geotechnical Laboratory at Ghent University and Arjen
Mascini of the Department of Geology and Soil Science
(UGCT) in Ghent University is greatly appreciated.
Downloaded by [ University of Hong Kong] on [03/11/22]. Copyright © ICE Publishing, all rights reserved.
SD2 SCU1 SCU2 Mol
0·709 0·60 0·60 0·179
0·445 0·451 0·456 0·635
NOTATION
Cu uniformity coefficient
D50 mean grain size diameter
Dr relative density
emax maximum void ratio
emin minimum void ratio
G0 small shear modulus
GHH small-strain modulus in horizontal plane
GHV small-strain modulus in vertical plane (obtained from the
signal propagating horizontally)
GVH small-strain modulus in vertical plane (obtained from the
signal propagating vertically)
Gs specific gravity
VS shear wave velocity
ρ soil density
REFERENCES
Altuhafi, F. N., Coop, M. R. & Georgiannou, V. N. (2016). Effect
of particle shape on the mechanical properties of natural
sands. J. Geotech. Geoenviron. Engng 142, No. 7, 145,
https://doi.org/10.1061/(ASCE)GT.1943-5606.0001569.
Bellotti, R., Jamiolkowski, M., Lo Presti, D. C. F. & O’Neill, D. A.
(1996). Anisotropy of small strain stiffness in Ticino sand.
Géotechnique 46, No. 1, 115–131, https://doi.org/10.1680/
geot.1996.46.1.115.
Börzsönyi, T. & Stannarius, R. (2013). Granular materials com-
posed of shape-anisotropic grains. Soft Matter 9, No. 31,
7401–7418, https://doi.org/10.1039/c3sm50298h.
Chaudhary, S. K., Kuwano, J. & Hayano, Y. (2004). Measurement of
quasi-elastic stiffness parameters of dense Toyoura sand in
hollow cylinder apparatus and triaxial apparatus with bender
elements. Geotech. Test. J. 27, No. 1, 23–35, https://doi.org/
10.1520/GTJ11267J.
Enomoto, T. (2016). Effects of grading and particle charac-
teristics on small strain properties of granular materials. Soils
Found. 56, No. 4, 745–750, https://doi.org/10.1016/j.sandf.2016.
07.014.
Escribano, D. E. & Nash, D. F. T. (2015). Changing anisotropy of G0
in Hostun sand during drained monotonic and cyclic loading.
Soils Found. 55, No. 5, 974–984, https://doi.org/10.1016/j.sandf.
2015.09.004.
Ezaoui, A. & Di Benedetto, H. (2009). Experimental measurements
of the global anisotropic elastic behaviour of dry Hostun sand
during triaxial tests, and effect of sample preparation.
Géotechnique 59, No. 7, 621–635, https://doi.org/10.1680/
geot.7.00042.
Fioravante, V. (2000). Anisotropy of small strain stiffness of Ticino
and Kenya sands from seismic wave propagation measured
in triaxial testing. Soils Found. 40, No. 4, 129–142,
https://doi.org/10.3208/sandf.40.4_129.
Fioravante, V. & Capoferri, R. (2001). On the use of multi-
directional piezoelectric transducers in triaxial testing.
Geotech. Test. J. 24, No. 3, 243–255, https://doi.org/10.1520/
GTJ11344J.
Fioravante, V., Giretti, D. & Jamiolkowski, M. (2013). Small strain
stiffness of carbonate Kenya sand. Engng Geol. 161, 65–80,
https://doi.org/10.1016/j.enggeo.2013.04.006.
Gao, Y., Wang, Y. H. & Su, J. C. P. (2015). Experimental
characterization of the influence of fines on the stiffness of
sand with inherent fabric anisotropy. Soils Found. 55, No. 5,
1148–1157, https://doi.org/10.1016/j.sandf.2015.09.015.
Giang, P. H. H., Van Impe, P. O., Van Impe, W. F., Menge, P. &
Haegeman, W. (2017). Small-strain shear modulus of calcareous
sand and its dependence on particle characteristics and
 ANISOTROPIC SMALL-STRAIN STIFFNESS OF CALCAREOUS SAND
gradation. Soil. Dyn. Earthq. Engng 100, 371–379, https://doi.
org/10.1016/j.soildyn.2017.06.016.
Gu, X. Q., Yang, J., Huang, M. S. & Gao, G. Y. (2015). Bender
element tests in dry and saturated sand: signal interpretation
and result comparison. Soils Found. 55, No. 5, 951–962,
https://doi.org/10.1016/j.sandf.2015.09.002.
Ishibashi, I., Chen, Y. & Chen, M. (1991). Anisotropic behavior of
Ottawa sand in comparison with glass spheres. Soils Found. 31,
No. 1, 145–155, https://doi.org/10.3208/sandf1972.31.145.
Iwasaki, T. & Tatsuoka, F. (1977). Effects of grain size and grading
on dynamic shear moduli of sands. Soils Found. 17, No. 3,
19–35, https://doi.org/10.3208/sandf1972.17.3_19.
Kuwano, R. (1999). The stiffness and yielding anisotropy of sand.
PhD thesis, Imperial College, London, UK.
Kuwano, R. & Jardine, R. J. (2002). On the applicability of
cross-anisotropic elasticity to granular materials at very small
strains. Géotechnique 52 No. 10, 727–749, https://doi.org/
10.1680/geot.2002.52.10.727.
Ladd, R. S. (1978). Preparing test specimens using undercompac-
tion. Geotech. Test. J. 1, No. 1, 16–23, https://doi.org/10.1520/
GTJ10364J.
Lagioia, R., Sanzeni, A. & Colleselli, F. (2006). Air, water and
vacuum pluviation of sand specimens for triaxial apparatus. Soils
Found. 46, No. 1, 61–67, https://doi.org/10.3208/sandf.46.61.
Li, Y., Yang, Y., Yu, H. & Roberts, G. (2018). Effect of
sample reconstitution methods on the behaviors of granular
materials under shearing. J. Test. Eval. 46, No. 6, 2718–2725,
https://doi.org/10.1520/JTE20170126.
Lo Presti, D., Pedroni, S. & Crippa, V. (1992). Maximum dry density
of cohesionless soils by pluviation and by ASTM D 4253-83:
A comparative study. Geotech. Test. J. 15, No. 2, 180–189,
https://doi.org/10.1520/GTJ10239J.
Masschaele, B., Dierick, M., Van Loo, D., Boone, M. N.,
Brabant, L., Pauwels, E., Cnudde, V. & Van Hoorebeke, L.
(2013). HECTOR: a 240 kV micro-CT setup optimized for
research. J. Phys. Conf. Ser. 463, No. 1, 012012, https://doi.org/
10.1088/1742-6596/463/1/012012.
Ogino, T., Kawaguchi, T., Yamashita, S. & Kawajiri, S. (2015).
Measurement deviations for shear wave velocity of bender
element test using time domain, cross-correlation, and frequency
domain approaches. Soils Found. 55, No. 2, 329–342,
https://doi.org/10.1016/j.sandf.2015.02.009.
Payan, M., Khoshghalb, A., Senetakis, K. & Khalili, N. (2016).
Small-strain stiffness of sand subjected to stress anisotropy. Soil
Dyn. Earthq. Engng 88, 143–151, https://doi.org/10.1016/
j.soildyn.2016.06.004.
Pennington, D. S., Nash, D. F. T. & Lings, M. L. (1997). Anisotropy
of G0 shear stiffness in Gault Clay. Géotechnique 47, No. 3,
391–398, https://doi.org/10.1680/geot.1997.47.3.391.
Piriyakul, K. & Haegeman, W. (2007). Void ratio function for elastic
shear moduli for Boom Clay. Géotechnique 57, No. 2, 245–248,
https://doi.org/10.1680/geot.2007.57.2.245.
Downloaded by [ University of Hong Kong] on [03/11/22]. Copyright © ICE Publishing, all rights reserved.
319
Powers, M. C. (1953). A new roundness scale for sedimentary
particles. J. Sedimentary Petrology 23, No. 2, 117–119, https://
doi.org/10.1306/D4269567-2B26-11D7-8648000102C1865D.
Rad, N. S. & Tumay, M. T. (1987). Factors affecting sand specimen
preparation by raining. Geotech. Test. J. 10, No. 1, 31–37,
https://doi.org/10.1520/GTJ10136J.
Tatsuoka, F., Iwasaki, T., Yoshima, S., Fukushima, S. & Sudo, H.
(1979). Shear modulus and damping by drained tests on
clean sand specimens reconstituted by various methods.
Soils Found. 19, No. 1, 39–54, https://doi.org/10.3208/
sandf1972.19.39.
Vaid, Y. P. & Negussey, D. (1984). Relative density of pluviated sand
samples. Soils Found. 24, No. 2, 101–105, https://doi.org/
10.3208/sandf1972.24.2_101.
Vaid, Y. P. & Negussey, D. (1988). Preparation of reconstituted sand
specimens. In Advanced triaxial testing of soil and rock (eds
R. T. Donaghe, R. C. Chaney and M. L. Silver), ASTM STP
977, pp. 405–417. Philadelphia, PA, USA: American Society for
Testing and Materials.
Van Impe, P. O., Van Impe, W. F., Manzotti, A., Mengé, P., Van den
Broeck, M. & Vinck, K. (2015). Compaction control and
related stress–strain behaviour of off-shore land reclamations
with calcareous sands. Soils Found. 55, No. 6, 1474–1486,
https://doi.org/10.1016/j.sandf.2015.10.012.
Vlassenbroeck, J., Dierick, M., Masschaele, B., Cnudde, V., Van
Hoorebeke, L. & Jacobs, P. (2007). Software tools for quanti-
fication of X-ray microtomography at the UGCT. Nuclear
Instruments Methods Phys. Res. A 580, No. 1, 442–445,
https://doi.org/10.1016/j.nima.2007.05.073.
Wichtmann, T. & Triantafyllidis, T. (2009). Influence of the
grain-size distribution curve of quartz sand on the small
strain shear modulus Gmax. J. Geotech. Geoenviron. Engng
135, No. 10, 1404–1418, https://doi.org/10.1061/(ASCE)GT.
1943-5606.0000096.
Wichtmann, T. & Triantafyllidis, T. (2010). On the influence of the
grain size distribution curve on P-wave velocity, constrained
elastic modulus Mmax and Poisson’s ratio of quartz sands. Soil
Dyn. Earthq. Engng 30, No. 8, 757–766, https://doi.org/10.1016/
j.soildyn.2010.03.006.
Yamamuro, J. A. & Wood, F. M. (2004). Effect of depositional
method on the undrained behavior and microstructure of sand
with silt. Soil Dyn. Earthq. Engng 24, No. 9–10, 751–760,
https://doi.org/10.1016/j.soildyn.2004.06.004.
Yamashita, S., Hori, T. & Suzuki, T. (2005). Effects of initial
and induced anisotropy on initial stiffness of sand by
triaxial and bender elements tests. In Geomechanics: testing,
modeling, and simulation (eds J. A. Yamamuro and J. Koseki),
GSP 143, pp. 350–369. Reston, VA, USA: American Society of
Civil Engineers.
Yang, Z. X., Li, X. S. & Yang, J. (2008). Quantifying and
modelling fabric anisotropy of granular soils. Géotechnique 58,
No. 4, 237–248, https://doi.org/10.1680/geot.2008.58.4.237.