This is the author’s version of a work that was submitted/accepted for pub-lication in the following source:

Gallage, Chaminda & Uchimura, Taro(2016)Direct shear testing on unsaturated silty soils to investigate the effects ofdrying and wetting on shear strength parameters at low suction.Journal of Geotechnical and Geoenvironmental Engineering, 142(3), Arti-cle Number-04015081.

This file was downloaded from: https://eprints.qut.edu.au/90829/

c© Copyright 2015 American Society of Civil Engineers

Notice: Changes introduced as a result of publishing processes such ascopy-editing and formatting may not be reflected in this document. For adefinitive version of this work, please refer to the published source:

https://doi.org/10.1061/(ASCE)GT.1943-5606.0001416

Direct Shear Testing on Unsaturated Silty Soils to Investigate the Effects of

Drying and Wetting on Shear Strength Parameters at Low Suction

Chaminda Gallage1 and Taro Uchimura2

1Senior Lecturer, Faculty of Science and Engineering, Queensland University of Technology, 2, George

Street Brisbane, QLD 4001, Australia; E-mail: chaminda.gallage@qut.edu.au

2Associate Professor, Department of Civil Engineering, University of Tokyo, 7-3-1, Hongo, Bunkyo-Ku,

Tokyo 113-8656, Japan; E-mail: uchimura@geot.t.u-tokyo.ac.jp

Abstract 1

A modified conventional direct shear device was used to measure unsaturated shear strength of two silty 2

soils at low suction values (0 ~ 50 kPa) that were achieved by following drying and wetting paths of soil-3

water characteristic curves (SWCCs). The results revealed that the internal friction angle of the soils was 4

not significantly affected by either the suction or the drying-wetting SWCCs. The apparent cohesion of 5

soil increased with a decreasing rate as suction increased. Shear stress-shear displacement curves obtained 6

from soil specimens subjected to the same net normal stress and different suction values showed a higher 7

initial stiffness and a greater peak stress as suction increased. A soil in wetting exhibited slightly higher 8

peak shear stress and more contractive volume change behavior than that of soil in drying at the same net 9

normal stress and suction. 10

11

Key words: unsaturated soils, direct shear tests, suction, hysteresis of suction, shear strength, 12

friction angle, apparent cohesion 13

14

15

Introduction 16

Conventional testing equipment, such as triaxial and direct shear testing apparatus used for saturated soil 17

testing, has been modified and extensively used in the determination of the shear strength of unsaturated 18

soils (Bishop and Donald 1961; Ho and Fredlund 1982; Gan et al. 1988; Ridley 1995; Vanapalli et al. 19

1996; Caruso and Tarantino 2004; Sedano et al. 2006). The modified direct shear apparatus is particularly 20

convenient for determination of shear strength of unsaturated soils (Fredlund and Rahardjo 1993). This 21

apparatus offers an advantage over triaxial tests because the drainage path is relatively short. In the direct 22

shear test, the time required to reach equilibrium conditions under applied matric suction is approximately 23

1 ~ 2 days compared to 7 ~ 15 days (depending on the sample height) for a triaxial test. Further, the 24

design of modified triaxial apparatus for unsaturated soils is complicated, and the experimental set-up is 25

tedious. Rahnenma et al. (2003) pointed out that the failure mechanism of rain-induced soil slope failures 26

can be more appropriately modeled by simple shear tests that are less complicated, easier to set-up, and 27

less time-consuming. 28

The present study involves modification of a conventional direct shear apparatus to apply the desired 29

suction to the soil specimen. The soil sample was exposed to the atmosphere to maintain pore-air pressure 30

(ua) at zero. The negative gauge pore-water pressure (uw) in the specimen was controlled by applying a 31

constant negative head to the water phase through a ceramic disk at the base. As water cavitates in the 32

measuring system when the negative gauge pressure approaches 101 kPa, the apparatus limited the 33

applied matric suction (ua-uw) to 101 kPa (i.e., 1 atm). The modified apparatus can be used to apply and 34

maintain a desired suction value following drying or wetting paths of soil-water characteristic curves 35

(SWCCs). 36

Zhao et al. (2013) reported that most rain-induced soil slope failures are shallow-seated with sliding 37

depths varying between 0.5 m and 3.0 m and therefore the shear stresses on the failure plane are in the 38

low range. To initiate rain-induced failures, the shear strength on the failure plane should decrease below 39

these mobilised stress values. During rainfall, the strength of soil decreases as soil suction decreases. 40

Therefore, at the initiation of rain-induced slope failures, a very low suction range (0 – 50 kPa) or positive 41

pore-water pressure can be expected on the failure plane (Gofar et al. 2008; Yeh et al. 2008). Further, 42

fine-grained soils exhibit a significant hysteresis (the difference between drying and wetting SWCCs) at 43

low suction range (Gallage and Uchimura 2010) and unsaturated shear strength of soil is affected by this 44

hysteresis. Therefore, to analyze and predict the rain-induced soil slope stability, it is important to 45

understand the effects of low suction and the hysteresis of SWCCs on unsaturated shear strength 46

parameters. 47

Although numerous experimental results can be found for the hysteresis between drying and wetting 48

SWCCs and effects of high suction on unsaturated shear strength parameters, limited experimental 49

evidence exists on the effects of the hysteresis and low suction on unsaturated shear strength parameters. 50

Therefore, this paper presents the results of modified direct shear tests conducted on two silty soils for a 51

low suction range (0~50 kPa). The suction values of test specimens were achieved following wetting and 52

drying paths of SWCCs of the soil (hysteresis). Further, the test results obtained from the modified direct 53

shear apparatus were used to investigate the effects of suction and dying-wetting hysteresis on the shear 54

strength, stiffness, internal friction angle, and apparent cohesion of test materials. 55

56

Modified Direct Shear Apparatus 57

Modification on a conventional direct shear apparatus was undertaken to facilitate direct shear testing on 58

soils under constant suction. Fig. 1 depicts the general layout of the modified apparatus. The apparatus 59

accommodates a cylindrical specimen (60 mm diameter and 20 mm height). A ceramic disk with an air 60

entry value of 100 kPa is embedded in the sample base. The sample can either be prepared on the disk or 61

placed upon it after extruding an undisturbed sample or re-compacting the sample. 62

The compartment below the ceramic disk is connected to a water tank that maintains a water level at the 63

mid-height of the sample. The water in the tank can be subjected to negative pressure, with a pressure 64

transducer connected to the bottom of the water tank facilitating the measurement (Fig. 1). Negative 65

pressure is applied to the water phase in the sample through the saturated ceramic disk at the base of the 66

sample. The soil specimen is exposed to the atmosphere through the gap (1 mm) between the two shear 67

boxes to maintain the pore-air pressure (ua) at zero. Therefore, the matric suction (ua-uw) in the sample is 68

equal to the absolute value of negative pressure in the water tank. This apparatus can be used to apply a 69

maximum suction of 100 kPa. 70

The water tank is placed on an electronic balance that can be used to measure mass up to 4.2 kg with an 71

accuracy of 0.01 g. The balance reading is used to calculate the movement of water into and out of the 72

soil specimen. A thin layer of silicon oil is placed on the water surface in the tank to minimise water 73

evaporation. It is assumed that equalization of the pore-water pressure of the sample is achieved when the 74

balance reading is approximately constant. 75

Total vertical stress (σn) is applied by a double action bellofram cylinder mounted with a loading piston, 76

which in turn is connected to the vertical loading shaft of the direct shear apparatus. Two load cells 77

attached to the load- receiving plate measure the vertical load (Wu, 2003). The sum of the readings of two 78

load cells yields the applied vertical load. During shearing, the control feedback system changes the 79

vertical load to maintain a constant vertical normal stress by taking the change in cross sectional area of 80

the sample into account. A shaft connected to the worm wheel system allows for application of horizontal 81

shear force to the upper shear box, facilitating displacement (strain) controlled shear tests. As shown in 82

Fig. 1, a load cell is connected to the horizontal shaft to measure the applied horizontal shear load. 83

Two linear voltage displacement transducers (LVDT), each with a maximum range of 10 mm, are used to 84

measure vertical and shear displacements. Output signals of all the transducers (three load cells, two 85

LVDTs, one pore-water pressure transducer) are logged using control software at specified intervals. 86

87

Testing Materials 88

Two silty soils, both from Japan, were tested in this study: Edosaki soil procured from a natural slope in 89

Ibaraki; and Chiba soil excavated from a railway embankment in Chiba prefecture. Wet sieving analysis 90

and hydrometer tests were performed on both the Edosaki and Chiba soils as they exhibit fines (particles 91

finer than 0.075 mm) contents of 16.5 % and 36 %, respectively. Table 1 displays grain size parameters 92

and basic soil properties, obtained using JGS standard test methods, for both materials, including: specific 93

gravity, maximum void ratio, minimum void ratio, and Atterberg limits. Review of these results relative 94

to the Unified Soil Classification System (USCS) characterizes both soils as silty sand. 95

Fig. 2 depicts SWCCs obtained in the laboratory using a Tempe pressure cell for Edosaki and Chiba 96

soils. Drying and the wetting SWCCs were obtained for a sample of Edosaki soil with a dry density of 97

1.35 Mg/m3, and for Chiba soil with a dry density of 1.25 Mg/m3. Gallage and Uchimura (2010) detailed 98

the Tempe pressure cell and its testing procedure used in this study. The equation in Fredlund and Xing 99

(1994) was used to best-fit the SWCC experimental data. 100

101

Testing Program and Data Analysis 102

Unsaturated shear strength parameters for the Edosaki soil (initial dry density of 1.35 Mg/m3) were 103

obtained through a series of constant suction consolidated drained direct shear (CSCDDS) tests at 104

predetermined low suction values (e.g., 0, 10, 20, 50 kPa) achieved following drying and the wetting 105

paths. Four tests were conducted under different net normal stress (σn-ua) values (e.g., 34, 84, 135, 185 106

kPa) for each suction value. All specimens were prepared with an initial gravimetric water content of 8 % 107

to obtain an initial matrix suction greater than 50 kPa. 108

A second series of CSCDDS tests were conducted on the Chiba soil (initial dry density of 1.25 Mg/m3) 109

with similar low suction values as used for the Edosaki series. As above, for each suction value four tests 110

with different net normal stress (σn-ua) values (e.g., 25, 50, 100, 200 kPa) were conducted. All test 111

specimens in this series were prepared with an initial gravimetric water content of 10 % to obtain an 112

initial matrix suction greater than 50 kPa. 113

The test procedure for measuring unsaturated shear strength parameters in wetting and drying using the 114

modified direct shear apparatus involves: 115

(1) The sample preparation and installation in the apparatus, 116

(2) Achieving the pre-determined suction following wetting or drying, and 117

(3) Shearing the sample under the controlled suction. 118

Sample Preparation 119

Following the procedure of Huang (1994), the ceramic disk and associated system were saturated. The 120

base of the sample (surface of the ceramic disk) was then elevated 1 mm above the bottom edge of the 121

lower shear box to allow the specimen to freely expand (dilate) during shearing. Thereafter, the upper and 122

lower shear boxes were clamped with spacers in between to create a gap. The gap eliminates frictional 123

resistance between the two shear boxes during shearing. The water tank was placed on the electronic 124

balance, with the water level brought to the mid-height of the specimen. The LVDT, which measures 125

vertical displacement, was set at the position indicated in Fig. 1. Following setup, all measurements were 126

initialized to zero. The valve connecting the water tank and the compartment below the saturated ceramic 127

disk and its associated system was closed. The pore-water pressure transducer was then connected to the 128

compartment to measure initial suction (negative pore-water pressure) of the sample. 129

The specimen was directly prepared on the ceramic disk by moist tamping at a pre-specified moisture 130

content to achieve a predetermined sample density. The top reaction frame and the modified load cells for 131

vertical load (Wu, 2003) were mounted and the spacers between shear boxes were carefully pulled out 132

minimizing disturbance to the specimen. 133

Data logging was initiated prior to stabilization of initial suction in the soil sample. The clamp on the 134

vertical loading shaft was then released and a predetermined vertical stress was applied at a constant rate 135

of increase of 2 kPa/min. At the same time, the valve connecting the water tank and the base of the 136

sample was opened allowing the water, which was subjected to a predetermined negative pressure, to 137

flow into the specimen through the ceramic disk. 138

139

Wetting and Drying Processes 140

The specimen in the wetting path 141

The initial water content of soils (8% or 10%) was determined in order to achieve initial suction greater 142

than 50 kPa. Suction was then reduced to a predetermined value, less than 50 kPa, by wetting the soil 143

sample. Negative air pressure was first applied to the water tank until the predetermined negative water 144

pressure was achieved. The valve was then opened allowing water to flow into the specimen through the 145

ceramic disk. During this water supply, the normal stress was gradually increased to a predetermined 146

value. The change in the mass of the water tank was monitored during the wetting process. Pore-water 147

pressure equalization of the specimen was assumed to have occurred when water flow stopped. At this 148

point, the reading of the electronic balance was approximately constant. It generally took between 24 to 149

30 hours to obtain pore-water pressure equalization in the sample. Fig. 3 shows the typical time histories 150

of sample pore-water pressures (measured and controlled) and the change in the mass of the water tank 151

until the pore-water pressure equalization was achieved. 152

153

The specimen in the drying path 154

Once the specimen was prepared on the ceramic disk and setup in the apparatus, it was saturated 155

allowing water, which was in the tank under atmospheric pressure conditions (zero pressure), to flow into 156

the specimen through the ceramic disk. The saturation of the specimen was assumed to have occurred 157

when the change in the mass of the water tank was negligible. At the same time, the normal stress was 158

gradually increased to a predetermined value. 159

After sample saturation, the valve was closed and negative air pressure was gradually applied to the 160

water tank until a predetermined negative pore-water pressure value was achieved. When the valve was 161

re-opened, water flowed into the tank from the specimen until pore-water pressure equalization was 162

achieved. It took between 12 to 24 hours for saturation and another 24~30 hours for de-saturation. Fig. 4 163

shows the typical time histories of the pore-water pressure and the change in the mass of the water tank 164

until the pore-water pressure equalization was achieved. 165

166

Shearing the Specimen 167

Once the predetermined suction (e.g., 0 ~ 50 kPa), either by drying or wetting, and vertical normal stress 168

(e.g., 25 ~ 200 kPa) were achieved, the specimen was sheared with the shear displacement rate of 0.00008 169

mm/sec, which is the minimum rate of shear displacement that this apparatus can apply. Gan et al. (1988) 170

reported that shear displacement rates smaller than 0.00022 mm/sec would satisfy the drainage shearing 171

assumption for unsaturated silty clay. Therefore, the shear displacement rate (0.00008 mm/sec) used in 172

this testing program would ensure the drained shearing conditions for the tested soils. 173

The vertical and shear displacements were measured in mm. Negative and positive values of the vertical 174

displacement represented contraction and dilation of the sample, respectively. The shear stress was 175

calculated by using the measured shear force and corrected cross sectional area of the specimen. During 176

shearing, the output data from the measuring system were logged at every 2 sec. The final water content 177

of the specimen was measured after the test was completed. 178

To obtain saturated shear strength properties of test materials, the specimen was first saturated following 179

the procedure explained in the section of “the specimen in the drying path” while increasing the normal 180

stress. Then, the specimen was sheared maintaining zero pore-water pressure. 181

182

Data Analysis 183

The shear stress (τ) and the normal vertical stress (σn) are calculated as follows: 184

τ = SF/Ac (1) 185

σn = NF/Ac (2) 186

where SF is the shear force, NF is the normal force to the specimen, and Ac is the corrected area of the 187

specimen being sheared. For a cylindrical specimen of internal diameter D, 188

189

−= θδθ sin

2

2

DDAc (3) 190

191

Where

= −

Dδθ 1cos in radians and δ is the relative displacement between the lower and upper shear 192

boxes. δ is equal to shear displacement (d) as the lower shear box is fixed. Shear displacement (δ ) can 193

be used to calculate shear strain (%) = (shear displacement /sample diameter) x 100). However, the shear 194

strain is not uniform around the perimeter of the sample when a circular sample is used. Therefore, the 195

shear displacement is used to interpret direct shear results when a circular sample is used. 196

197

Results and Discussion 198

The results of a series of unsaturated direct shear tests on Edosaki and Chiba soils are presented and 199

discussed in this section. 200

201

Effects of Net Normal Stress on Shear Strength and Volume Change Behaviors of Soil 202

Fig. 5 show the effects of the net normal stress (σn –ua) on shear stress-shear displacement and volume 203

change behaviors of Edosaki soil with constant suction of 20 kPa, which was achieved by following a 204

wetting path. It can be seen that the shear strength and the initial stiffness of unsaturated soils increase 205

with increasing net normal stress (σn –ua). Further, the volume of unsaturated soils becomes contractive 206

(in the figures, the negative vertical displacement represents the volume reduction of the specimen) as the 207

net normal stress increases. Similar results can be observed from four tests conducted under constant 208

suction of 20 kPa, which was achieved by following a drying path (Figs. 6(a) and (b)). Tests conducted on 209

both Edosaki and Chiba soils with constant suction of 0 (saturated samples), 10 kPa, and 50 kPa achieved 210

by following either wetting or drying paths showed similar results. It is seen from the results that the 211

shear strength and the initial stiffness of unsaturated soils increase and the volume of soils becomes 212

contractive as the net normal stress increases. These behaviors are independent of the degree of saturation, 213

the method of suction achievement (drying or wetting), and soil type. Further, similar results have been 214

observed by Huang (1994); Blight (1967); Gan and Fredlund (1996); Mashhour et al. (1995). 215

According to Newton’s friction law, frictional resistance increases with increasing the normal stress 216

acting on the plane. The same concept is applied to the shear plane when a direct shear test is conducted 217

on soil. In a direct shear test, the applied vertical stress is normal to the shear plane and therefore when 218

the vertical stress (net normal stress) is increased, the shear strength (the failure shear stress) of soil 219

increases (Fig. 5(a) and Fig. 6(a)). Further, a greater vertical stress induces a higher confinement to the 220

direct shear sample. As a result, the shear stiffness (the initial slope shear stress – shear displacement 221

curve) of soil increases with the vertical net normal stress (Fig. 5(a) and Fig. 6(a)). 222

Initial dry density of soil samples tested in this study is 1.35 g/cm3 or 1.25 g/cm3. These densities are 223

relatively low as no clear peak stress is observed in the shear stress – shear displacement plots. Therefore, 224

when these samples are sheared the net normal stress may force soil particles to close-packed 225

arrangement (volume contraction) during shearing. As a result, samples become more contractive as the 226

net normal stress increases (Figs. 5(b) and 6(b)). This volume contraction makes the sample denser and as 227

a result the shear resistance and the initial stiffness increase. However, when the net normal stress is low 228

and it is smaller than the peak (maximum) shear stress of the samples, the vertical normal stress is not big 229

enough to force soil particles into close-packed arrangement during shearing. In such cases, the samples 230

tend to show dilation (volume increase) during shearing as shown in Fig. 6(b) by the specimen with net 231

normal stress of 34 kPa. It can be concluded that under constant suction, the increase in vertical stress 232

causes a progressive sample volume change from dilatancy to compression until dilatancy completely 233

disappears. 234

235

Effects of Suction on Shear Stress-Shear Displacement and Volume Change Behaviors 236

Figs. 7(a) and (b) show the shear stress-shear displacement and the volumetric behaviors, respectively, 237

of four tests conducted on Edosaki soil specimens subjected to the same net normal stress of 34 kPa and 238

different suction values (e.g., 0, 10, 20, and 50 kPa). The suction values were achieved by wetting. The 239

results suggest that the shear strength and stiffness of the soil increase with the increase in soil suction. 240

The soil exhibits more dilative volumetric behavior as suction is increased. Similar results were observed 241

for every group of four tests with the four different suction values: 0, 10 kPa, 20 kPa, and 50 kPa that 242

were achieved by following either drying or wetting path. Tests were conducted on both Edosaki and 243

Chiba soils. Each group of four tests was conducted by maintaining a constant net normal stress (e.g., 34, 244

84, 135, 185 kPa for Edosaki soil, and 25, 50, 100, 200 kPa for Chiba soil). The test results strongly 245

demonstrate that suction has a great influence on the shear characteristic of unsaturated soil. Independent 246

of:soil type, net normal stress, andhow the suction was achieved (i.e., following drying and wetting paths), 247

the shear resistance and stiffness of soils increases, and the specimens exhibit more contractive behaviors 248

as suction increases. These phenomena are consistent with the fact that inter-particle forces in soils 249

increase as suction increases. 250

To clarify the physics of unsaturated particulate media, two spherical particles of radius R in contact are 251

considered. The water meniscus between them is bound by the two particles and by an imaginary torus. 252

The small radius of this doughnut-shaped torus is r1 and the distance from the centre to the inside wall of 253

the torus is r2. Therefore, the local contact force, F, which the meniscus imposes on the particles, 254

contributed by the pressure of the fluid acting on the cross-sectional area of the meniscus and the surface 255

tension (Ts) acting along the perimeter of the meniscus, can be expressed as (Cho and Santamarina 2001): 256

257

)2())(( 22

2 rTruuF swa ππ +−= (4) 258

259

This force is the only one arising from meniscus water and increases as suction increases. Therefore, the 260

effects of matric suction result in a greater normal force holding the particles together and limiting 261

slippage strength (Sawangsuriya 2006). As a result, the stiffness and the strength of unsaturated soils 262

increase with increasing matric suction. However, this effect is not unbounded, since the contact force (F) 263

tends toward a limiting value due to progressive reduction in the meniscus radius (r1) as suction increases 264

(Mancuso et al. 2002). 265

As shown in Fig. 7(a), a residual state is achieved at 5 - 6 mm shear displacement. At this shear 266

displacement, pore-water paths could be discontinued at the shear plane and pore-water pressure in the 267

shear plane will be equal to atmospheric pressure. Therefore, no suction effect can be observed at the 268

residual state. 269

Another effect of suction on the unsaturated soil behavior can be clearly seen in the volumetric 270

deformation during shearing. As shown in these figures, the soil is more contractive at zero suction and 271

becomes less contractive with the increasing value of suction prior to shear, regardless of the value of the 272

net confining pressure. As soil structures become stronger and less deformable with an increase in suction, 273

it is expected that they will exhibit dilative (less contractive) behavior with an increase in suction. 274

However, it is worth mentioning that the effect of suction on volumetric behavior of unsaturated soil 275

reduces as the net confining stress increases. Similar effects of suction on stress-strain and volumetric 276

behaviors have been reported by Huang (1994); Blight (1967); Gan and Fredlund (1995); Mashhour et al. 277

(1995). 278

279

Effects of Drying and Wetting on Shear Stress –Shear Displacement and Volume Change Behaviors 280

The results shown in Fig. 8 were obtained by shearing two Edosaki sand specimens subjected to the 281

same constant value of suction of 10 kPa and an identical net normal stress of 84 kPa. The only difference 282

was the method used to achieve the suction: one was achieved by wetting and the other was by drying. It 283

can be observed from the results that the specimen subjected to wetting exhibits slightly greater peak (or 284

failure) shear stress than the specimen subjected to drying. Further, the specimen subjected to wetting is 285

more contractive than that subjected to drying. The similar behaviors were observed independent of soil 286

(e.g: Edosaki or Chiba soil), suction value, and net normal stress used. Though, the difference between 287

wetting and drying stress-strain behavior is small, a significant difference in volume change during 288

shearing can be observed. 289

The results shown in Fig. 8 suggest that the specimen subjected to wetting (lower water content) has 290

slightly higher peak shear strength than that subjected to drying (higher water content). These results 291

contradict the previous finding that the higher water content resulting from the drying process results in a 292

higher strength in the soil sample (Lamborn 1986; Vanapalli et al. 1996; Oberg and Sallfors 1997). 293

Therefore, to investigate the possible reasons for these contradictory results, detailed analyses were 294

conducted on data of two identical specimens (one subjected to drying and the other subjected to wetting) 295

measured during shearing. 296

Fig. 9 shows the vertical displacement, degree of saturation, and suction change during shearing of two 297

identical soil samples of Chiba soil re-compacted to an initial dry density of 1.25 g/cm3 at a gravimetric 298

water content of 10%. Although both specimens were subjected to the same suction of 10 kPa, one was 299

subjected to drying process and the other was subjected to wetting process. Both specimens were sheared 300

under a constant vertical net stress of 50 kPa. As shown in Fig. 9(a), the vertical displacement suggests 301

that the specimen subjected to wetting is more contractive than the specimen subjected to drying. No 302

water content change in the sample was observed in either specimen during shearing and the final water 303

contents (just after the test) were measured as 20.9 % and 18.7% for the sample subjected to drying and 304

wetting, respectively. The change in degree of saturation of the specimens during shearing was calculated 305

using the final water content and the sample volume change. The sample volume change was calculated 306

using the measured vertical displacement and assuming a constant cross sectional area which is equal to 307

the cross sectional area of a shear box. Figure 9(b) shows the change in degree of saturation of the two 308

specimens during shearing. As both specimens exhibited contractive behavior, the degree of saturation 309

increased as shown in Fig. 9(b). Since the specimen subjected to drying is more contractive than the one 310

subjected to wetting, the difference between the degree of saturation of the specimens subjected to drying 311

and wetting decreased as shown in Fig. 9(b). It was observed by Wu et al. 2007 and Airey (1987) that the 312

volumetric strain in the shear band in the direct shear test is significantly greater than the average 313

volumetric strain of the specimen. Hence, the volume change in the shear band of the specimen from 314

wetting and from drying could be more contractive and more dilative, respectively, than the calculated 315

values. Especially at peak stress, this volume change behavior could make the degree of saturation of the 316

shear band of the specimen from wetting to be greater than that from drying. This phenomenon may give 317

higher strength to the specimen subjected to wetting than to the subjected to drying. The experimental 318

results herein may then agree with the results of Lamborn (1986); Vanapalli et al. 1996; Oberg and 319

Sallfors (1997). 320

The suction in the specimens was controlled by applying negative pressure to the water supplying tank 321

and therefore, no change in suction in the specimens was observed during shearing (Fig. 9(a)). 322

323

Effects of Suction and Wetting-Drying Hysteresis on Shear Strength Parameters 324

The direct shear test results on unsaturated soils demonstrated the significant effect of suction on the 325

shearing resistance, of which a more detailed and quantitative analysis will be made in this section. In this 326

analysis, the failure of a specimen refers to its peak (maximum) shear stress. The correction for the 327

dilatancy effect was not made on the measured shear stresses at failure since the majority of direct shear 328

tests exhibited contractive behavior. 329

To obtain the shear strength parameters (the apparent cohesion (c’) and the internal friction angle (φ’)) 330

corresponding to a particular suction value, four soil specimens subjected the same suction and four 331

different net normal stresses (i.e., 34, 84, 134, and 184 kPa) were sheared. Implying the mentioned failure 332

criterion, the failure shear stress for each test was then obtained. Plotting these failure shear stresses with 333

the corresponding net normal stress and performing best linear-fitting on the plotted data, the apparent 334

cohesion(c’) and the internal friction angle (φ’) corresponding to the particular suction value were 335

obtained. Figs. 10(a) and (b) depict the variation of φ’ and c’, respectively, with the increase in suction, 336

for Edosaki and Chiba soils. Furthermore, these figures demonstrate the effect of wetting–drying on φ’ 337

and c’. 338

It can be seen from the Fig. 10(a) that φ’ is relatively independent of the increase in suction, and the 339

calculated φ’ values are quite consistent with the effective internal friction angle of the saturated soil. It 340

can be observed that the effective friction angle of saturated soil makes the same contribution to the shear 341

strength of unsaturated soil. Further, Fig. 10(a) shows that the effect drying-wetting hysteresis of the 342

SWCCs on φ’ is insignificant. This φ’ represents the frictional resistance at inter-particle contacts, and 343

may be affected by surface roughness of the particles (Christopher et al. 2008), particle crushing during 344

loading (Hamidi et al. 2009; Bolton 1986), and the density (dilatancy effect) (Bolton 1986), but not by 345

suction or water content (Cokca et al 2004; Mouazen et al. 2002). 346

Fig. 10(b) shows that the apparent cohesion (c’) increases at a decreasing rate as the suction increases. A 347

number of researchers (Escario & Saez 1986; Fredlund et al. 1987; Wheeler 1991; Ridley 1995; Ridley et 348

al. 1995) found a linear increase of c’ with suction until the air-entry value of soil, followed by a non-349

linear increase (decreasing rate). The air-entry values of soils used in this study are relatively small, 3 – 4 350

kPa (Fig. 2), and therefore the effect of the air-entry on linear and non-linear variation of c’ with suction 351

cannot be observed. The increase in suction may increase the inter-particle bonding force which can 352

contribute to the apparent cohesive force between particles. Furthermore, these figures reveal that the c’ 353

from wetting is greater than that from drying, and this is consistent with the finding of Han et al. (1995). 354

A possible reason for this behavior has been discussed under the sub-topic of “Effects of the drying and 355

the wetting on shear stress–shear displacement and volume change behaviors” 356

357

Conclusions 358

The shear behavior of unsaturated silty soils subjected to low suction (0 ~50 kPa) and low net normal 359

stress was investigated in this study by using a modified direct shear apparatus. A series of shear tests was 360

conducted on Edosaki and Chiba soils under various combinations of net normal stress and suction. The 361

suction was achieved by either wetting or drying in order to investigate the effect of wetting-drying (the 362

hysteresis of SWCC) on the shear behavior of unsaturated soils. The main conclusions of this study are 363

follows: 364

• The shear resistance increases and volume change becomes more contractive with the increase in 365

the net normal stress. 366

• Shear stiffness and peak shear stress increase as suction increases. Further, a soil having a higher 367

suction shows less contractive volume change during shearing. 368

• The internal friction angles of tested materials are independent of the suction and the wetting-369

drying hysteresis of SWCC. 370

• The results indicated that the apparent cohesion increases with suction in a non-linear fashion at a 371

decreasing rate with increasing suction. 372

• Soil with suction achieved by wetting exhibits higher apparent cohesion at the same suction than 373

soil with suction achieved by drying. 374

• During shearing, soil from wetting is more contractive than the soil from drying at the same 375

suction and the net normal stress. 376

377

Acknowledgement 378

The authors gratefully acknowledge the Promoting Fundamental Transport Technology Research of the 379

Japan Railway Construction, Transport and Technology Agency (JRTT) and Grants-in-Aid for Scientific 380

Research of the Japan Society for the Promotion of Science (JSPS) for the financial support for this study. 381

This paper is based on the first author’s research completed as partial fulfillment for the Ph.D 382

requirements at the University of Tokyo, Japan. Therefore, the first author acknowledges the scholarship 383

received from the Ministry of Education, Science, Sports, and Culture, Government of Japan 384

(MONBUSHO) for undertaking his doctoral degree. 385

References 386

Bishop, A. W., and Donald, I. B. (1961). “The experimental study of partially saturated soils in the 387

triaxial apparatus.” Proc., 5th Int. Conf. Soil Mech. Found. Eng., Vol. 2, Dunod, Paris, France, 13-21. 388

Blight, G. E. (1967). “Effective stress evaluation for unsaturated soils.” J. Soil Mech. Found. Eng. Div., 389

93(2), 125-148. 390

Bolton, M. D. (1986). “The strength and dilatancy of sands.” Geotechnique, 36(1), 65-78. 391

Caruso, A., and Tarantino, A. (2004). “A shear box for testing unsaturated soils from medium to high 392

degree of saturation.” Geotechnique, 54(4), 281-284 393

Cho, G. C., and Santamarina, J. C. (2001). “Unsaturated particulate materials – particle-level studies.” J. 394

Geotech. Geoenviron. Eng., 127(1), 84-96. 395

Christopher, A.B., Tuncer, B. E., Craig, H. B., and David, M. M. (2008). “Geological and physical factors 396

affecting the friction angle of compacted sands.” J. Geotech. Geoenviron. Eng., 134(10), 1476-1489. 397

Cokca, E., Erol, O., and Armangil, F. (2004). “Effects of compaction moisture content on the shear 398

strength of unsaturated clay.” Geotech. Geol. Eng. J., 22(2), 285-297. 399

Escario, S., and Saez, J. (1986). “The shear strength of partly saturated soils.” Geotechnique, 36(4), 453-400

456. 401

Fredlund, D. G., and Rahardjo, H. (1993). Soil mechanics for unsaturated soils, John Wiley & Sons, New 402

York. 403

Fredlund, D. G., Rahardjo, H., and Gan, J. K. M. (1987). “Non-linearity of strength envelope for 404

unsaturated soils.” Proc. 6th Int. Conf. Expansive Soils, Vol. 1, Taylor & Francis, New Delhi, India, 49-405

54. 406

Fredlund, D. G., and Xing, A. (1994). “Equation for the soil-water characteristic curve.” Can. Geotech. J., 407

31(4), 521-532. 408

Gallage, C. P. K., and Uchimura, T. (2010). “Effects of dry density and grain size distribution on soil-409

water characteristic curves of sandy soils.” Soil Found., 50(1), 161 -172. 410

Gan, J. K. M., Fredlund, D. G., and Rahardjo, H. (1988). “Determination of the shear strength parameters 411

of an unsaturated soil using the direct shear test.” Can. Geotech. J., 25(8), 500-510. 412

Gan, J. K. M., and Fredlund, D. G. (1996). “Shear strength characteristics of two saprolitic soils.” Can. 413

Geotech. J.,33(4), 595-609. 414

Gofar, N., Lee, M. L., and Kassim, A. (2008). “Response of Suction Distribution to Rainfall Infiltration in 415

Soil Slope.” Elect. J. Geotech. Eng., 13(E), 1-13. 416

Hamidi, A., Alizadeh, M., and Soleimani, S. M. (2009). “Effect of particle crushing on shear strength and 417

dilation characteristics of sand gravel mixtures.” Int. J. Civil Eng., 7(1), 61-71. 418

Han, K. K., Rahardjo, H., and Broms, B. B. (1995). “Effects of hysteresis on the shear strength of a 419

residual soil.” Proc., 1st Int. Conf. Unsat. Soils, E. E. Alonso and P. Delage, ed., Vol. 1, A. A. Balkema, 420

Paris, France, 499-504. 421

Ho, D. Y. F., and Fredlund, D. G. (1982). “A multi-stage triaxial test for unsaturated soils.” Geotech. Test. 422

J., 5(1/2), 18-25. 423

Huang, Y. (1994). “Effects of suction on strength and deformation behavior of unsaturated collapsible 424

soils.” Ph.D. thesis, University of Tokyo, Tokyo, Japan. 425

Lamborn, M. J. (1986). “A micromechanical approach to modeling partly saturated soils.” M.Sc. thesis, 426

Texas A&M University, Texas, USA. 427

Mancuso, C., Vassallo, R., and d’Onofrio, A. (2002). “Small strain behaviour of a silty sand in controlled 428

suction resonant column – torsional shear tests.” Can. Geotech. J., 39(1), 22-31. 429

Mashhour, M. M., Ibrahim, M. I., and El-Eemam, M. M. (1995). “Variation of unsaturated soil shear 430

strength parameters with suction.” Proc., 1st Int. Conf. Unsat. Soils, E. E. Alonso and P. Delage, ed., Vol. 431

3, A. A. Balkema, Paris, France, 1487-1493. 432

Mouaza, A. M., Ramon, H., and Baerdemaeker, J. D. (2002). “Effect of bulk density and moisture content 433

on selected mechanical properties of sandy loam soil.” Biosys. Eng., 83(2), 217-224. 434

Oberg, A. L., and Sallfors, G. (1997). “Determination of shear strength parameters of unsaturated silt and 435

sands based on the water retention curve.” Geotech. Test. J., 20(1), 40-48. 436

Sawangsuriya, A. (2006). “Stiffness-suction-moisture relationship for compacted soils.” PhD. thesis, 437

University of Wisconsin, Madison, USA. 438

Rahnenma, A., Habibagahi, G., and Ghahramani, A. (2003). “A new simple shear apparatus for testing 439

unsaturated soils.” Iran. J. Sci. Tech., 27(B1), 73-80. 440

Ridley, A. M., Burland, J. B., and Monroe, A. S. (1995). “Unconfined compression tests with pore 441

pressure measurements.” Proc., 11th Afric. Reg. Conf. Soil. Mech. Found. Eng., Egyptian Geotechnical 442

Society, Cairo, Egypt, 332-345 443

Ridley, A. M. (1995). “Strength-suction-moisture content relationships for Kaolin under normal 444

atmospheric conditions.” Proc., 1st Int. Conf. Unsat. Soils, E. E. Alonso and P. Delage, ed., Vol. 2, A. A. 445

Balkema, Paris, France, 645-651. 446

Senado, J. A. I., Vanapalli, S. K., and Garga, V. K. (2006). “Modified ring shear apparatus for unsaturated 447

soils testing.” Geotech. Test. J., 30(1), 1-9. 448

Vanapalli, S. K., Fredlund, D. G., Pufahl, D. E., and Clifton, A.W. (1996). “Model for the prediction of 449

shear strength with respect to soil suction.” Can. Geotech. J., 33(3), 379-392. 450

Wheeler, S. J. (1991). “An alternative framework for unsaturated soil behavior.” Geotechnique. 41(2), 451

257-261. 452

Wu, P., Matsushima, K., and Taysuoka, F. (2007). “Effects of specimen size and some other factors on 453

the strength and deformation of granular soil in direct shear tests.” Geotech. Test. J., 31(1), 45-64 454

Wu, P. (2003). “Particle size effects on soil-reinforcement interaction in direct shear tests.” Ph.D. thesis, 455

University of Tokyo, Tokyo, Japan. 456

Yeh, H. F., Lee, C. C., and Lee, C. H. (2008). “A rainfall infiltration model for unsaturated soil slope 457

stabilisation.” J. Environ. Eng. Manage., 18(4), 261 – 268. 458

Zhao, H.F., Zhang, L.M., and Chang, D. S. (2013). “Behavior of coarse widely graded soils under low 459

confining pressures.” J. Geotech. Geoenviron. Eng., 139(1), 35-48 460

Table 1. Physical Properties of Test Materials Properties Edosaki Soil Chiba Soil

Specific gravity, Gs 2.75 2.72

Mean Grain size, D50 [mm] 0.22 0.14

Coefficient of uniformity, Cu= D60/D10 16.4 54.40

Coefficient of gradation, Cc=(D30)2/(D10*D60) 3.97 1.95

Sand content, [%] (0.075 – 4.75 mm) 83.50 64.00

Fines content, [%] (smaller than 0.075 mm) 16.50 36.00

Clay content, [%] (smaller than 0.002 mm) 2.00 5.00

Maximum void ratio, emax 1.59 1.74

Minimum void ratio, emin 1.01 1.11

Liquid limit [%] NP 25.78

Plastic limit [%] NP 23.52

Plastic index NP 2.26

Fig. 1. Schematic diagram of modified direct shear apparatus Fig. 2. Soil-water characteristic curves for Edosaki and Chiba soils Fig. 3. Sample preparation and bringing the sample to 10 kPa suction following the wetting path Fig. 4. Sample preparation and bringing the sample to 10 kPa suction following the drying path Fig. 5. Effects of the net normal stress on, (a) shear stress-shear displacement behavior, (b) volumetric behavior of unsaturated Edosaki soil (suction 20 kPa achieved by following the wetting) Fig. 6. Effects of the net normal stress on, (a) shear stress-shear displacement behavior, (b) volumetric behavior of unsaturated Edosaki soil (suction 20 kPa achieved by following the drying) Fig. 7. Effects of the suction on, (a) shear stress-shear displacement behavior, (b) volumetric behavior of Edosaki soil (the net normal stress is 34 kPa for all tests and the suction was achieved by wetting) Fig. 8. Effects of drying-wetting on shear stress-shear displacement and volumetric behaviors of Edosaki soil (the net normal stress is 84 kPa, the suction is 10 kPa) Fig. 9. Behaviors of volume change (vertical displacement), suction, and degree of saturation during shearing of two identical soil samples with 10 kPa suction achieved by drying and wetting in Chiba soil Fig. 10. Effects of the suction and drying-wetting on (a) the internal friction angle and (b) the apparent cohesion of Edosaki soil and Chiba soil

0.1 1 10 1000.0

0.1

0.2

0.3

0.4

0.5

Chiba Soil (ρd=1.25 g/cm3)Edosaki Soil (ρd= 1.35 g/cm3)

V

olum

etric

wat

er c

onte

nt, θ

w

Suction [kPa]

Drying (Chiba) Wetting (Chiba) Drying (Edosaki) Wetting (Edosaki)

Lines are F&X fitting

0 200 400 600 800 1000 1200-80

-60

-40

-20

0

20000 40000 60000 80000-80

-60

-40

-20

0

2670

2672

2674

2676

2678

Sample preparationP

ore-

wat

er p

ress

ure

[kP

a]

Time [sec]

She

arin

g th

e sp

ecim

en

Control pore-water pressure to bring the sample to a constant suction of 10 kPa

Change in the weight of the tank

constant suction (ua-uw) =10 kPa

Time [sec]

Por

e-w

ater

pre

ssur

e [k

Pa]

Bal

ance

read

ing

[g]

0 40000 80000 120000 160000-60

-50

-40

-30

-20

-10

0

10

2660

2662

2664

2666

2668

2670

2672

2674

2676

Por

e-w

ater

pre

sure

[kP

a]

Time [sec]

Bal

ance

read

ing

[g]

shea

ring

the

spec

imen

Balance reading

Equalization of suctioninside the specimen

Application of the constant suctionof 10 kPa by following the drying processSample preparation

and saturation

0

40

80

120

160

0 2 4 6 8 10 12-0.6

-0.4

-0.2

0.0

0.2(b)

(a)

(σn-ua)=34 kPa

(σn-ua)=84 kPa

(σn-ua)=135 kPa

(σn-ua)=185 kPa

Edosaki soilρd=1.35 g/cm3

(ua-uw)=20 kPa

Wetting

S

hear

stre

ss [k

Pa]

Contraction

Dilation

(σn-ua)=84 kPa

(σn-ua)=135 kPa

(σn-ua)=34 kPa

(σn-ua)=185 kPa

Wetting

Ver

tical

dis

plac

emen

t [m

m]

Shear displacement [mm]

0 2 4 6 8 10 12-0.6

-0.4

-0.2

0.0

0.2

0.4

0.60

40

80

120

160

200

(σn-ua)=185 kPa

(σn-ua)=135 kPa

(σn-ua)=84 kPa

(σn-ua)=34 kPa

Ver

tical

dis

plac

emen

t [m

m]

Shear displacement [mm]

(b)

(a)

(σn-ua)=185 kPa

(σn-ua)=135 kPa

(σn-ua)=84 kPa

(σn-ua)=34 kPa

Edosaki soilρd=1.35 g/cm3

(ua-uw)=20 kPa

(Drying)

S

hear

stre

ss [k

Pa]

0

10

20

30

40

50

0 2 4 6 8 10 12-0.8

-0.6

-0.4

-0.2

0.0

0.2

0.4

0.6(b)

(a)

ua-uw = 50 kPa

ua-uw = 20 kPa

ua-uw = 10 kPa ua-uw = 0 kPa

Wetting

Edosaki soilρd=1.35 g/cm3

(σn-ua)=34 kPa

S

hear

stre

ss [k

Pa]

ua-uw = 50 kPa

ua-uw = 20 kPa

ua-uw = 10 kPa

ua-uw = 0 kPa

Ver

tical

dis

plac

emen

t [m

m]

Shear displacement [mm]

0 2 4 6 8 10 12

0

20

40

60

80

V. dis. Drying

V. dis. Wetting

DryingWetting

Edosaki soilρd=1.35 g/cm3

(σn-ua)=84 kPa(ua-uw)=10 kPa

She

ar s

tress

[kP

a]

Shear displacement [mm]

Contraction

Dilation

-0.6

-0.4

-0.2

0.0

0.2

Ver

tical

dis

plac

emen

t [m

m]

0 2 4 6 8 10 12

44

46

48

50

3

4

5

6

-0.8

-0.6

-0.4

-0.2

0.0

Wetting

Drying

D

egre

e of

sa

tura

tion

[%]

Shear displacement [mm]

Difference of degree of saturation of two samples (suction achieved by drying and wetting) during shearing

Diff

eren

ce o

f deg

ree

of s

atur

atio

n [%

]

(b)

(a)

Suction in Wetting

Suction in Drying

V. displacement in Drying

V. displacement in Wetting

ρd=1.25 g/cm3

(σn-ua)=50 kPa(ua-uw)=10 kPa

V

ertic

al

disp

lace

men

t [m

m]

8

9

10

11

12

13

Suc

tion

[kP

a]

0 10 20 30 40 5005

10152025

30323436384042

App

aren

t co

hesi

on, C

[kP

a]

Suction, (ua-uw) [kPa]

(b)

Chiba Soil (ρd=1.25 g/cm3) Wetting Drying

Inte

rnal

fri

ctio

n an

gle,

φ' [

deg.

]Edosaki Soil (ρd=1.35 g/cm3)

Wetting Drying(a)