This article was downloaded by: [University Of South Australia Library]On: 01 August 2012, At: 21:20Publisher: Taylor & FrancisInforma Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House,37-41 Mortimer Street, London W1T 3JH, UK

Geomechanics and Geoengineering: An InternationalJournalPublication details, including instructions for authors and subscription information:http://www.tandfonline.com/loi/tgeo20

Drained behaviour of cemented sand in high pressuretriaxial compression testsA. Marri a , D. Wanatowski b & H.S. Yu ba Department of Civil Engineering, NED University of Engineering and Technology, Karachi,Pakistanb Nottingham Centre for Geomechanics, Faculty of Engineering, University of Nottingham,University Park, Nottingham, United Kingdom

Version of record first published: 06 Jun 2012

To cite this article: A. Marri, D. Wanatowski & H.S. Yu (2012): Drained behaviour of cemented sand in high pressure triaxialcompression tests, Geomechanics and Geoengineering: An International Journal, DOI:10.1080/17486025.2012.663938

To link to this article: http://dx.doi.org/10.1080/17486025.2012.663938

PLEASE SCROLL DOWN FOR ARTICLE

Full terms and conditions of use: http://www.tandfonline.com/page/terms-and-conditions

This article may be used for research, teaching, and private study purposes. Any substantial or systematicreproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form toanyone is expressly forbidden.

The publisher does not give any warranty express or implied or make any representation that the contentswill be complete or accurate or up to date. The accuracy of any instructions, formulae, and drug doses shouldbe independently verified with primary sources. The publisher shall not be liable for any loss, actions, claims,proceedings, demand, or costs or damages whatsoever or howsoever caused arising directly or indirectly inconnection with or arising out of the use of this material.

Geomechanics and Geoengineering: An International JournaliFirst, 2012, 1–16

Drained behaviour of cemented sand in high pressure triaxial compression tests

A. Marria , D. Wanatowskib* and H.S. Yub

aDepartment of Civil Engineering, NED University of Engineering and Technology, Karachi, Pakistan; bNottingham Centre for Geomechanics,Faculty of Engineering, University of Nottingham, University Park, Nottingham, United Kingdom

(Received 20 April 2011; final version received 30 January 2012)

In this paper, drained behaviour of cemented sand under high pressure is studied. A recently developed high pressure triaxial apparatus is used. Thetest results indicate the significance of degree of cementation and confining pressure on the isotropic compression, volumetric change, stress-strainbehaviour and stress-dilatancy relationship of cemented sand at high confining pressures. The results suggest that the influence of cementationis greater at low confining stresses and it reduces with increasing confining stress where the effect of the confining pressure becomes dominant.A Scanning Electron Microscope analysis has also been included in the paper. It revealed that the particle and cement bonding breakage at highpressure is dependent on the stress level and the cement content. The higher the confining pressure the more significant particle and cement bondingbreakage is observed. The higher the cement content the fewer sand particles and cement bonds are broken.

Keywords: high pressure; cemented sand; stress-strain behaviour; triaxial compression; shear strength

1. Introduction

Numerous studies carried out in the past highlighted the impor-tance of understanding the effect of cementation on the strengthand deformation characteristics of naturally cemented soils(e.g. Clough et al. 1981, Lade and Overton 1989, Leroueil andVaughan 1990, Airey and Fahey 1991, Airey 1993, Coop andAtkinson 1993, Lagioia and Nova 1995, Cuccovillo and Coop1997, 1999, Schnaid et al. 2001, Yu et al. 2007). In sands,the cementing agents are often found to be silicates, amor-phous silica, iron oxide, and calcium carbonate (Clough et al.1981). One of the features of naturally cemented materialsis that they have variable densities. These variable densitiesand difficulty of sampling without disturbing the cementationmake it difficult to study the fundamental behaviour of natu-rally cemented materials in the laboratory (Airey 1993, Huangand Airey 1998). Therefore, in order to avoid these difficulties,artificially cemented soils have been used in many studies (e.g.Airey 1993, Huang and Airey 1998, Consoli et al. 2000, 2006,Ismail et al. 2002, Rotta et al. 2003, Haeri et al. 2005, 2006).It has been shown that the patterns of behaviour observed inboth naturally and artificially cemented soils are similar eventhough the cementation may be produced by different pro-cesses. Therefore, artificially cemented materials are assumedto simulate the stress-strain behaviour, volumetric change andstrength characteristics of naturally cemented soils. As a result,artificially cemented soils are most frequently used in thelaboratory to investigate behaviour of cemented geomaterials.

∗Corresponding author. Email: dariusz.wanatowski@nottingham.ac.uk

Although research on cemented soils has been very activein recent years, most experimental studies in this field havebeen focused on soil behaviour at low or moderate pressureswith only a few studies considering the effects of high pres-sure (Lade and Overton 1989, Coop and Atkinson 1993, Huangand Airey 1993, 1998, Lagioia and Nova 1995, Cuccovilloand Coop 1999, Rotta et al. 2003). On one hand, it is under-standable because most geotechnical problems occur at lowconfining pressures. On the other hand, there are some caseswhere the soil behaviour must be analyzed at high pressures,e.g. deep pile foundations, particularly for offshore piling, deepmine shafts, high earth dams, tunnels, and deep oil-bearingstrata. In most of these situations, soils may be subjected topressures from approximately 5 MPa to 100 MPa. There aresome cases, however, where pressures may be much higher. Forinstance, pressures in soil under the tips of deep driven pilesmay experience pressure up to 350 MPa (Yamamuro and Lade1996).

Furthermore, the majority of experimental studies at highpressures published in the literature have been focused on theisotropic compression behaviour of uncemented sand. It hasbeen reported (Lade et al. 1996, Jefferies and Been 2000, Boppand Lade 2005) that sand under isotropic loading is relativelyincompressible at low stresses, and large volume changes onlyoccur at very high stress levels where particle crushing becomesa dominant mechanism of volume change. It is also well knownthat the behaviour of cemented soils is generally more com-plex than that of uncemented soils (Leroueil and Vaughan

ISSN 1748-6025 print/ISSN 1748-6033 online© 2012 Taylor & Francishttp://dx.doi.org/10.1080/17486025.2012.663938http://www.tandfonline.com

Geo

mec

hani

cs a

nd G

eoen

gine

erin

g do

wnl

oade

d fr

om w

ww

.tand

fonl

ine.

com

2 A. Marri et al.

1990, Airey 1993, Coop and Atkinson 1993, Consoli et al.2000, 2006, Leroueil 2001, Haeri et al. 2005, Yu et al. 2007).Therefore, using experimental data obtained at low pressure tosolve practical problems where high pressures exist may lead tosignificant errors. A reasonable approach to eliminate potentialerrors is to carry out more laboratory experiments at high pres-sures and to understand element behaviour of cemented soilssubjected to high pressures. However, laboratory equipmentcapable of high pressure testing is not commonly available insoil mechanics laboratories. As a result, our understanding ofthe behaviour of cemented soils under high pressures is stillvery limited.

Several experimental studies have demonstrated that thestress-strain behaviour of bonded geomaterials changes frombrittle to ductile as the confining stress increases (Clough et al.1981, Airey and Fahey 1991, Anagnostopoulos et al. 1991,Coop and Atkinson 1993, Lagioia and Nova 1995, Cuccovilloand Coop 1999, Schnaid et al. 2001). Since bonding tends tobe brittle, this suggests that the behaviour of a cemented soilis mainly cohesive at very low confining stresses, changing topurely frictional as the confining stress increases. This transi-tion is believed to occur because the bonding bears some ofthe confining stress and thus “weakens” its effect on the soilskeleton. As shown by Yu et al. (2007), based on the data fromSchnaid et al. (2001), this “weakening” effect can be repre-sented in a stress-dilatancy relationship by the inter-particlebonding c normalized by the minor effective principal stressσ ′

3. However, Schnaid et al. (2001) carried out all their triaxialtests at low confining pressures, p′

c ≤ 100 kPa. This leads to thequestion whether the stress-dilatancy behaviour of a cementedsoil at high confining pressures is similar or different from thatat low confining pressures.

The main objective of this paper is to characterize themechanical behaviour of cemented sand at high confining pres-sures. A newly developed high-pressure triaxial cell capable oftesting various geomaterials including uncemented or cementedgranular soils and soft rocks is introduced. Several drainedtriaxial compression tests are presented and discussed. Finally,micromechanics of cemented specimens is analysed using SEMphotographs taken after the tests.

2. High pressure triaxial apparatus

A high pressure testing system, developed at the University ofNottingham in the United Kingdom in conjunction with GDSInstruments Ltd. was used in this study. A layout of the testingsystem is shown in Figure 1(a) and a schematic cross-sectionand a photograph of the high pressure cell are presented inFigure 1(b).

As shown in Figure 1(a), displacement (or load) in the high-pressure system was applied from the bottom of a loadingframe via a displacement controller. A 100 kN submersibleload cell was used to measure the vertical load at the top ofspecimen. The cell pressure was applied through a GDS digital

pressure/volume controller (DPVC). Another DPVC was usedto control the back pressure from the top of the specimen andmeasure the volumetric change. The DPVCs used in this studyhad a capacity of 64 MPa. The resolutions of the DPVC were0.5 kPa for pressure and 1 mm3 for volume change. A highcapacity pressure transducer was also used to measure the porewater pressure at the bottom of the specimen.

The high-pressure cell has the balanced ram installed at thetop of cell the pressure chamber, which ensures zero upliftforce and zero volume/pressure change inside the cell whenthe specimen is loaded at a high cell pressure. The chamberinside the balanced ram is linked to the cell pressure chamberusing the stainless steel tube connected to ports A and B, shownin Figure 1(b). When the cell pressure was applied during thetriaxial test, the uplift pressure on the end of the ram was equalto the downward pressure applied by the upper chamber pres-sure on the top of the piston. Due to the balanced pressuresacting on the triaxial cell ram, the vertical force applied to theloading frame was reduced significantly and the loading framewith a relatively low capacity of 100 kN could be used in all thehigh pressure tests.

A 100 kN loading frame that can function as an independentcompression machine was used in this study (Figure 1c). Theloading frame is equipped with a virtual infinite stiffness (VIS)system, in which both the measurement and control of platendisplacement are automatically corrected so that it correspondsto the deformation that occurs between the base platen and theload button of the load cell. In this way, the displacement ofplatens is corrected for strain in the load cell and side columnsof the frame, bending flexure of the cross beam, and distor-tion within the motorized mechanical transmission. Therefore,the vertical displacement could be measured without additionalexternal transducers. As small strain behaviour was not withinthe scope of this study, local strain measurement instrumen-tation was not used. Another reason for not using any localstrain measuring devices was that they could have been dam-aged when the specimen failed and shear banding occurred athigh pressures.

3. Material tested

Well-graded, medium quartz sand from Sheffield in the UnitedKingdom, so called Portaway sand, was used as the base mate-rial for the cemented specimens. Portaway sand was previouslyused at the University of Nottingham in two other experimentalprojects on evaluation and extension of a critical state modelfor sand (Yu et al. 2005) and on validation of the shakedownconcept for pavement design and analysis (Brown et al. 2008).The index properties of Portaway sand, determined by BritishStandard methods (BS 1377), are given in Table 1. The particlesize distribution of Portaway sand is shown in Figure 2. Thesand grains are mainly sub-angular and sub-rounded. It shouldbe noted that the index properties of Portaway sand reported inthis study are slightly different from those reported by Yu et al.(2005), because of different batches of the tested sand.

Geo

mec

hani

cs a

nd G

eoen

gine

erin

g do

wnl

oade

d fr

om w

ww

.tand

fonl

ine.

com

Geomechanics and Geoengineering: An International Journal 3

Pore pressure control

Back pressure/ volume control

Porous stones

Top capTop drainagevalve

Bottom drainagevalve

Cell chamber

Load cell

Pore pressure valve

Spacers

Computer

(a)

(b) (c)

Cell pressure/ volume control

Displacement control

Spec

imen

Cross beam

Ram slip coupling

Balanced ram assembly

Cell top

Clamping ring

Retaining ring

Cell base

Base platen

Top cap

Extension deviceSpacer 2

LoadcellSpacer 1

B

A

Figure 1. High-pressure triaxial testing system: (a) schematic diagram; (b) schematic cross-section and photograph of the cell; (c) photograph of the cell in theloading frame.

4. Specimen preparation and testing procedures

All the specimens tested in this study (50 mm diameter ×100 mm height) were prepared by mixing relevant amounts ofdry Portaway sand and 5%, 10%, and 15% ordinary Portlandcement by weight of dry sand. Mixing of dry materials wascontinued until a uniform appearance of the sand-cement mix-ture was obtained. Water was then added to the mixture in

accordance with the optimum moisture content of 10% andfurther mixing was performed until a homogeneous appearanceof the moist sand-cement mixture was achieved. The mixturewas then stored in an airtight container to avoid any moistureloss before subsequent compaction. The specimens were com-pacted in layers into a 50 mm diameter and 100 mm high splitmould to a target dry unit weight of 17.4 kN/m3. To improvesafe extrusion of cemented specimens, the split mould was

Geo

mec

hani

cs a

nd G

eoen

gine

erin

g do

wnl

oade

d fr

om w

ww

.tand

fonl

ine.

com

4 A. Marri et al.

Table 1. Index properties of Portaway sand

Property Value

Effective grain size D10: mm 0.19D30: mm 0.29Mean grain size D50: mm 0.39D60: mm 0.42Specific Gravity, Gs 2.65emax 0.79emin 0.46

0

10

20

30

40

50

60

70

80

90

100

0.01 0.1 1 10

(%)

pass

sing

Grain size (mm)

Figure 2. Grain size distribution curve of Portaway sand.

provided with transparency sheet to avoid sticking of thecement to the walls of the mould. To achieve a greater uni-formity of specimens the undercompaction method, proposedby Ladd (1978) was used. After compaction, the specimenswere allowed to cure inside the mould for 24 hours. Themoulds were then dismantled and the specimens were storedin a humid room to cure for 14 days before testing. The curedspecimens were characterized as moderately to very stronglycemented soils (average unconfined compressive strength of565, 2210 and 5540 kPa, respectively) using the classificationcriteria proposed by Rad and Clough (1985).

After curing, a specimen was placed on the base pedestal ofthe high pressure triaxial cell. The surface of the specimen wascovered with a thin film of a mixture of kaolin clay and finesand to minimize membrane penetration effects. Two 0.6 mmthick neoprene membranes and two O-rings were fitted aroundthe specimen and the pedestal. The cell was then assembledand filled with distilled water. After the specimen was set up,a cell pressure of 50 kPa was applied. For saturation, the spec-imen was flushed with de-aired water from the bottom to thetop for 60 minutes under a back pressure slightly lower thanthe cell pressure. After flushing was completed, a saturationramp was applied in which both the cell pressure and the backpressure were increased simultaneously with the effective stressof 50 kPa maintained constant until the back pressure reached

2000 kPa. At this stage, Skempton’s pore water pressure param-eter (B-value) greater than 0.95 for uncemented and 0.90 forcemented sand was obtained. It is worth noting that the reasonfor the high back pressure value of 2000 kPa was to assure thedegree of saturation above 0.90 for cemented specimens. Thespecimen was then isotropically consolidated to the requiredmean effective stress and sheared under drained conditions,i.e., with �σ ′

3 = 0. All the consolidated isotropically drained(CID) tests were carried out under a deformation-controlledloading mode at a constant deformation rate of 0.02 mm/min(1.2%/h). For the calculation of the applied vertical stress,the conventional cross section area correction specified by BS1377(1990) was adopted before the peak state. After the peak,the correction of the cross section area, as measured at the endof test was applied proportionally with strain from the peak tothe end of test using a method proposed by La Rochelle et al.(1988).

5. Uniformity of specimens

To characterize the nature of cementation and cement-particleinteraction a microscopic analysis was undertaken. A scan-ning electron microscope (SEM) was used to analyze selectedspecimens before and after triaxial testing. Typical photomicro-graphs obtained from cut up sections of specimens with cementcontent of 5%, 10% and 15% are shown in Figure 3(a), (b) and(c), respectively. The photomicrographs were taken after curingbut before triaxial testing.

It can be seen from Figure 3 that in all three specimens thesand grains are well coated by the cement. However, the thick-ness and size of the bonds increases with increasing cementcontent. Furthermore, it was observed from several micro-graphs at various magnifications that the number and sizeof inter-particle voids reduced with higher cement content.As shown in Figure 3, on a microscopic level, some non-uniformity in the specimens could be noticed. This is becausewhen the cement content increases, the cement not only bondsparticles together (with some sand particles completely sur-rounded by the cement) but also fills some of the pores asinclusions. Nonetheless, on a macroscopic scale, the cementedspecimens could be assumed as homogeneous.

6. Behaviour in isotropic compression

The effects of initial void ratio and cement content on theisotropic compression characteristics of Portaway sand werefirst investigated. Then the effects of high confining pressureon particle crushing and cement bond breakage were examinedin further detail with the help of SEM analysis carried out onthe specimens retrieved after the isotropic compression tests.

Typical results of the isotropic compression tests carriedout on Portaway sand with 0% and 10% cement content areplotted in Figures 4 and 5, respectively. The specimens were

Geo

mec

hani

cs a

nd G

eoen

gine

erin

g do

wnl

oade

d fr

om w

ww

.tand

fonl

ine.

com

Geomechanics and Geoengineering: An International Journal 5

(a) (b)

(c)

Figure 3. SEM photographs of cemented specimens before testing: (a) 5% cement; (b) 10% cement; (c) 15% cement.

isotropically consolidated from various void ratios to meaneffective stress of 50 MPa. The solid lines and symbols inthe figures indicate measured data whereas the dashed linesindicate extrapolated parts of compression curves. It can beseen from Figures 4 and 5 that the e-log(p′) curves obtainedfor uncemented and cemented sand are similar. In both typesof specimens there is a very little compression up to yield-ing points. However, after the yielding a significant changein void ratios can be observed. It can also be observed fromFigures 4 and 5 that there is a tendency for all the compres-sion curves to approach a unique line regardless of the initialvoid ratio. It should be noted that in this paper, the primaryyielding was determined as the state where the isotropic com-pression curve on the e-p′ plot (i.e. in the linear scale) deviatedfrom the initial linear behaviour (Cuccovillo and Coop 1997,Rotta et al. 2003, Consoli et al. 2006, Dalla Rossa et al. 2008).This point can be considered as the state at which breakage ofthe cement bonds commences (Consoli et al. 2006, Wang andLeung 2008). The results presented in Figures 4 and 5 agreewell with those published by other researchers (e.g. Bopp andLade 2005, Rotta et al. 2003, Consoli et al. 2005, dos Santoset al. 2010).

Selected SEM micrographs of specimens with 10% cementcontent are shown in Figure 6. The specimens had initial

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.01 0.1 1 10 100

e

p' (MPa)

e = 0.495

e = 0.631

Cement = 0 %

Isotropic compression of uncemented Portaway sand

Figure 4. Isotropic compression curves of uncemented Portaway sand.

void ratios of 0.657 (Figure 6a), 0.584 (Figure 6b), and 0.523(Figure 6c). These values correspond to the relative densities of40% (medium loose state), 62% (medium dense state) and 81%(dense state), respectively.

Geo

mec

hani

cs a

nd G

eoen

gine

erin

g do

wnl

oade

d fr

om w

ww

.tand

fonl

ine.

com

6 A. Marri et al.

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.01 0.1 1 10 100

e

p' (MPa)

e = 0.657e = 0.584

e = 0.523

Cement = 10%

Isotropic compression of cemented Portaway sand

Figure 5. Isotropic compression curves of cemented Portaway sand.

It can be seen from Figure 6 that the extent of parti-cle and bond damage is relatively high at the loose state(Figure 6a) and gradually diminishes with increasing relativedensity (Figures 6b and 6c). Eventually, the particle and bondbreakage becomes insignificant at dense state (Figure 6c). This

suggests that there is significant effect of initial void ratio (orrelative density) on particle crushing and bond breakage ofcemented sand. The lower the void ratio (or higher relative den-sity) the lower level of damage in the cemented specimens athigh confining pressures.

The effect of cement content on the compressibility ofPortaway sand was also investigated. Figure 7 shows typicalcompression curves for specimens with 0%, 5%, 10%, and15% cement content. All the specimens had the same initialvoid ratio of 0.49 at the beginning of compression. It shouldbe noted that all the tests presented in Figure 7 were termi-nated at the confining pressure of 20 MPa because the loadcell used in the high pressure triaxial system reached its capac-ity. It can be observed from Figure 7 that the amount and rateof compression decrease with the increase in cement content.As a result, different void ratios were obtained at the end ofeach compression test. In other words, the compressibility ofPortaway sand reduces with increasing cement content, whichis consistent with observations made by Lade and Overton(1989), Rotta et al. (2003), Consoli et al. (2005) and dos Santoset al. (2010). This might be due to the reduction in particlecrushing because of increase in the cement content, as thelarge volume changes can only occur when particle crushingbecomes dominant, as reported by Lee and Farhoomand (1967),

(a) (b)

(c)

Figure 6. SEM photographs of cemented sand specimens after isotropic compression at high pressures with initial relative densities of: (a) Dr = 40%; (b) Dr =62%; (c) Dr = 81%.

Geo

mec

hani

cs a

nd G

eoen

gine

erin

g do

wnl

oade

d fr

om w

ww

.tand

fonl

ine.

com

Geomechanics and Geoengineering: An International Journal 7

0.3

0.4

0.5

0.6

0.01 0.1 1 10 100

e

p' (MPa)

C = 0%

C = 5%

C = 10%

C = 15%

e = 0.49 ~

Figure 7. Isotropic compression curves of specimens with different cementcontents.

McDowell et al. (1996), McDowell and Bolton (1998), Mesriand Vardhanabhuti (2009).

7. Behaviour in drained shearing

In this study, isotropically consolidated drained (CID) triaxialtests were carried out on both uncemented and cemented sandspecimens. The specimens were consolidated to mean effectivestresses, p′

c = 1, 4, 8, and 12 MPa and sheared under drainedconditions with σ ′

3 maintained constant. A summary of all theCID tests carried out in this study are given in Table 2. Typicalresults obtained from the tests carried out on cemented speci-mens with cement contents 5%, 10%, and 15%, are plotted asq-εa and εv-εa curves in Figure 8.

It can be seen from the q-εa curves that during shearing atσ ′

3 = 1 MPa, the deviatoric stress firstly reached a peak, andthen reduced gradually to a constant ultimate value for all threecement contents. In other words, strain softening behaviour wasobserved in all the three tests. However, for the tests carriedout under the confining pressure 12 MPa, the peak deviatoricstresses could not be clearly defined and there was practicallyno strain softening observed.

Figure 8 also shows that the stress-strain behaviour obtainedfrom tests carried out at 4 MPa and 8 MPa was in between those

Table 2. Summary of isotropically consolidated drained compression tests

Initial conditions Failure state Ultimate state

Test σ ′3(MPa) C (%) ec p′

peak (kPa) qpeak (kPa) φ′peak (o) p′

ult (kPa) qult (kPa) φ′ult (o)

CD-0C0.05M 0.05 0 0.49 109 180 40.4 94 137 35.8CD-0C0.1M 0.1 0 0.49 205 318 38.1 178 237 33.1CD-0C0.2M 0.2 0 0.48 411 624 37.3 389 473 32.0CD-0C0.3M 0.3 0 0.48 592 881 36.6 516 662 31.9CD-0C0.5M 0.5 0 0.48 964 1395 35.6CD-0C1M 1 0 0.47 1836 2363 32.4 1732 2144 30.9CD-0C4M 4 0 0.44 7779 9815 31.9 7768 9769 30.6CD-0C8M 8 0 0.43 13690 17188 31.3 13690 17188 30.3CD-0C10M 10 0 0.43 18450 22505 30.6 17450 22505 30.0CD-0C12M 12 0 0.41 20093 24361 30.4 19879 23718 29.8CD-0C20M 20 0 0.39 32924 38945 29.4 31517 34712 28.1CD-5C0.05M 0.05 5 0.50 331 779 58.2 145 249 41.9CD-5C0.1M 0.1 5 0.50 445 654 52.6 202 303 37.1CD-5C0.3M 0.3 5 0.50 763 1385 44.2 555 761 34.3CD-5C1M 1 5 0.50 2288 3800 40.6 1815 2519 33.5CD-5C2M 2 5 0.49 4188 6244 37.2 3525 4670 32.9CD-5C4M 4 5 0.48 7537 10637 34.8 7099 9332 32.6CD-5C6M 6 5 0.47 11086 15265 34.1 10486 13482 32.0CD-5C8M 8 5 0.47 14470 19367 33.2 14149 18409 31.7CD-5C10M 10 5 0.46 17775 23160 32.6 17622 22907 31.2CD-5C12M 12 5 0.45 21192 27662 32.1 21072 27305 30.7CD-5C20M 20 5 0.43 33282 39896 29.7 31868 33930 28.5CD-10C0.05M 0.05 10 0.52 701 1949 71.9 143 272 46.2CD-10C0.5M 0.5 10 0.52 2585 6125 59.0 1190 1963 40.3CD-10C1M 1 10 0.51 3887 8484 52.4 2221 3601 38.5CD-10C4M 4 10 0.51 8972 14882 40.5 7280 9932 33.8CD-10C6M 6 10 0.51 11616 16990 36.0 10591 13904 33.1CD-10C8M 8 10 0.50 14880 20750 34.5 14174 18643 32.7CD-10C10M 10 10 0.49 18242 24726 33.6 17491 22479 32.4CD-10C12M 12 10 0.49 21886 29630 33.3 21345 28015 31.8CD-10C20M 20 10 0.47 33946 41868 30.5 31371 34124 29.1CD-15C1M 1 15 0.53 4943 11124 56.6 2139 3393 39.5CD-15C4M 4 15 0.52 10943 20776 46.2 8309 12952 36.2CD-15C8M 8 15 0.51 16585 25857 38.2 14913 20844 34.5CD-1510M 10 15 0.51 17566 25655 36.5 17055 21227 34.0CD-15C12M 12 15 0.51 22221 30715 35.2 21602 28863 33.2CD-15C20M 20 15 0.50 34584 43761 31.2 31003 35003 30.0

Geo

mec

hani

cs a

nd G

eoen

gine

erin

g do

wnl

oade

d fr

om w

ww

.tand

fonl

ine.

com

8 A. Marri et al.

0

10

20

30

40

q (M

Pa)

ε v (

%)

1 MPa 4 MPa8 MPa 12 MPa

c = 5%

–8

–4

0

4

8

120 5 10 15 20 25 30

(a) (b)

1 MPa 4 MPa8 MPa 12 MPa

c = 5%

εa (%)

0

10

20

30

40

q (M

Pa)

1 MPa 4 MPa8 MPa 12 MPa

c = 10%

–8

–4

0

4

8

120 5 10 15 20 25 30

1 MPa4 MPa8 MPa

c = 10%

ε v (

%)

εa (%)

(c)

0

10

20

30

40

q (M

Pa)

1 MPa 4 MPa8 MPa 12 MPa

c = 15%

–8

–4

0

4

8

120 5 10 15 20 25 30

1 MPa4 MPa8 MPa12 MPa

c = 15%

ε v (

%)

εa (%)

Figure 8. Stress-strain and volume change behaviour of cemented sand with the cement contents: (a) 5%; (b) 10%; (c) 15%.

obtained at 1 MPa and 12 MPa. The specimens with 5% and10% cement content displayed very little or no strain softening.Similar behaviour was also observed for the specimens with15% cement content. However, peak deviatoric stresses fol-lowed by strain softening could be clearly identified. Deviatoricstresses in all the tests approached constant at the end of eachtest, as shown in Figure 8. Therefore, it can be concluded thatthe ultimate state was reached in all the tests. It can be seenfrom the q-εa curves that the higher the p′

c, the higher thedeviatoric stress at the peak state and at the ultimate state.

An examination of the q-εa curves plotted in Figure 8 showsthat the behaviour of cemented sand is dependent on the cementcontent. In general, the increase in the peak deviatoric stressand the reduction in the peak axial strain can be observed withthe increase in cement content. In other words, the increasedamount of cementation changes the stress-strain behaviour ofcemented specimens from ductile to brittle. However, it canbe observed that the effect of cementation reduces with theincrease in confining pressure. For example, the specimen with

15% cement content sheared at σ ′3 = 1 MPa exhibited stiff

stress-strain behaviour up to the peak deviatoric stress, thensoftened, and approached a constant ultimate state at the endof test. On the other hand, the peak deviatoric stress followedby strain softening was not clearly observed for the specimenwith the same cement content sheared at σ ′

3 = 12 MPa.The εv-εa curves of the tests on cemented sand are also pre-

sented in Figure 8. The volumetric contraction was observed inall the tests carried out at σ ′

3 = 8 and 12 MPa. For almost allthe tests carried out at σ ′

3 of 1 and 4 MPa, an initial volumetriccontraction was followed by a subsequent volumetric dilation,except for the specimen with 5% cement content sheared atσ ′

3 = 4 MPa, which exhibited only volumetric contraction.It can be observed from Figure 8 that, regardless of the cementcontent, the higher the p′

c, the more contractive behaviourof the specimens was obtained from the tests. It can also beobserved from the εv-εa curves shown in Figure 8 that thevolumetric strain curves were approaching a constant value atthe end of each test. Therefore, the ultimate state obtained from

Geo

mec

hani

cs a

nd G

eoen

gine

erin

g do

wnl

oade

d fr

om w

ww

.tand

fonl

ine.

com

Geomechanics and Geoengineering: An International Journal 9

the q-εa curves could also be considered to be very close tothe critical state for the cemented sand tested in this study.However, it should be pointed out that either single or multi-ple shear bands, often combined with bulging, have occurred inall the CID tests carried out in this study. This could affect anaccurate measurement of critical state parameters because afterthe occurrence of shear bands, the stresses and strains measuredby the load cells and displacement transducers might no longerrepresent the element behaviour of soil (e.g. Chu et al. 1996,Chu and Wanatowski 2008, 2009, Wanatowski and Chu 2008).

The εv-εa curves plotted in Figure 8 also show that increas-ing the cement content reduces the contraction of the cementedsand for a given confining pressure. However, the effect ofcementation on the volumetric strain behaviour reduces athigher confining pressures, where all the specimens exhibitcontractive behaviour and the volumetric strain curves getcloser to each other.

The effects of confining pressure and cement content on thestress-strain behaviour of soil, illustrated in Figure 8, agree wellwith other experimental data published in the literature (Cloughet al. 1981, Lagioia and Nova 1985, Lade and Overton 1989,Airey and Fahey 1991, Coop and Atkinson 1993, Yamamuroand Lade 1996, Cuccovillo and Coop 1999, Schnaid et al.2001, Haeri et al. 2005). In general, the increase in confiningpressure increases the peak deviatoric stress, the axial strainto the peak and the amount of contraction during shearing.Therefore, the stress-strain behaviour becomes increasinglyductile with increasing confining pressure. On the other hand,the increase in cement content increases the peak deviatoricstress but reduces the axial strain to the peak and the amount ofcontraction during shearing (i.e. increases the amount of dila-tion). In other words, the stress-strain behaviour changes fromductile to brittle with increasing cement content.

In order to examine the fabric changes, particle crushingand cement bond breakage of Portaway sand due to shear-ing, SEM analysis of the specimens was carried out after thetests. Figure 9 shows the micro features of an uncemented

Figure 9. SEM photographs of an uncemented specimen after shearing.

(i.e. 0% cement content) specimen of Portaway sand shearedat σ ′

3 = 20 MPa. It can be seen that significant particlecrushing occurred during shearing. The close-up of the speci-men in Figure 9 confirms that a significant particle break-age has occurred in this test altering the gradation and fabricsof the specimen. Similar observations were made for otheruncemented specimens tested at high pressures in this study.Furthermore, it was observed that the amount of particlescrushed during shearing increased with increasing confiningpressure. This is consistent with previous experimental stud-ies carried out on uncemented granular materials (e.g. Lee andFarhoomand 1967, Lade and Yamamuro 1996, Lade et al. 1996,McDowell and Bolton 1998, Yamamuro and Lade 1996).

Figure 10 shows micro features of a cemented specimenwith 10% cement content. It can be seen that the particlecrushing and cement bond breakage have occurred. However,SEM photos and visual inspection of specimens with differentcement contents revealed that the amount of particles crushedand bonds broken during shearing reduced with an increase incement content.

8. Failure characteristics

Due to a cohesive-frictional nature of cemented soils, theirfailure behaviour is influenced by the friction angle, φ′

f, andthe cohesion intercept, c. These failure characteristics can bedetermined from the peak state (i.e. failure state) data obtainedfrom CID triaxial tests. The peak points obtained from theCID tests carried out on specimens of Portaway sand withdifferent cement contents under a wide range of confining pres-sure are plotted on the q-p’ plane in Figure 11. Best-fit failureenvelopes were plotted for the failure data obtained from spec-imens with different cement contents. The failure envelopeswere also extrapolated to zero confinement.

It can be seen from Figure 11 that all the failure envelopesare curved. However, it appears that there is somehow increase

Figure 10. SEM photograph of a specimen with 10% cement content aftershearing.

Geo

mec

hani

cs a

nd G

eoen

gine

erin

g do

wnl

oade

d fr

om w

ww

.tand

fonl

ine.

com

10 A. Marri et al.

0

5

10

15

Dev

iato

ric S

tres

s at

Fai

lure

, qf (

MP

a)

20

25

30

35

40

45

0 5 10 15 20

Mean Effective Stress at Failure, p'f (MPa)

25 30 35 40 45

0% Cement

5% Cement

10% Cement

15% Cement

Uncemented Portaway sand

Figure 11. Failure envelopes of specimens with different cement contents.

in the curvature of the failure envelopes with the increasein cement content. It can also be observed that the failureenvelopes move to higher stress levels with increasing cementcontent. As a result, extrapolated deviatoric strength at zeroconfinement also increases with increasing cement content.This in turn suggests that the cohesion intercept and an inter-particle cohesion of Portaway sand would also increase withthe cement content.

Figure 11 also shows that the slope of the failure envelopesdecreases with the increase in confining pressure, which showsthat the curvature of failure envelopes is also affected by confin-ing pressure. This indicates that the shear strength of cementedspecimens is significantly reduced when specimens are shearedat high pressures. Such a reduction occurs because of the break-age of cementing bonds by consolidation and shearing at highpressures, as observed in this study and reported by Lade andOverton (1989), Malandraki and Toll (2000), and Asghari et al.(2003). This means that the influence of cement content isgreater at low confining stresses and this effect reduces withan increase in confining stress. Therefore, it is expected thatthe failure envelopes of uncemented and cemented sand wouldmerge at very high confining pressures. However, the confiningpressures applied in this study were not high enough to confirmthis observation.

A summary of all failure states obtained from CID triaxialtests on uncemented and cemented Portaway sand is shownin Figure 12, in which the failure friction angle φ′

f in plot-ted versus the effective confining stress σ ’3. In addition, thefriction angle at the critical state (φ′

cs= 29.8◦) determined forPortaway sand by Yu et al. (2005) is indicated in the figure.It should be pointed out however, that the critical state frictionangle was determined by drained and undrained triaxial tests on

25

30

35

40

45

50

55

60

65

70

75

0 2 4 6 8 10 12 14 16 18 20

Fric

tion

Ang

le a

t Fai

lure

, φf (

deg)

Effective Confining Pressure, σc' (MPa)

0% cement

5% Cement

10% Cement

15% Cement

φ'cs = 29.8° (from Yu et al. 2005)

Figure 12. Effect of confining pressure and cement content on the frictionangle of Portaway sand.

loose Portaway sand at relatively low mean effective stressesp′

c < 1 MPa.As shown in Figure 12, the friction angle at failure increases

with increasing cement content but reduces with increasingconfining pressure. Therefore, the differences between the peakstrength of specimens with different cement contents are largeat low confining pressures and reduce significantly at high pres-sures. In other words, the influence of cement content is greaterat low confining stresses and this effect reduces with an increasein confining stress. It can also be observed from Figure 12 thatthe failure friction angle of uncemented sand determined at theconfining pressure, σ ′

3 = 20 MPa coincides with the criticalstate friction angle determined for loose specimens at low pres-sures. Furthermore, failure states of specimens with cementcontent of 5%, 10% and 15% approached asymptotically thecritical state of uncemented sand (i.e. 0% cement content).As shown in Figure 12, at the confining pressure of 20 MPa, aspecimen with 5% cement content practically reached the criti-cal state. Specimens with 10% and 15% cement content shearedat the confining pressure of 20 MPa were also very close to thecritical state.

It is worth noting that the effect of confining pressure on thefailure characteristics of artificially cemented sand observed inthis study is consistent with the findings previously reportedby Yamamuro and Lade (1996) for a clean sand, Malandrakiand Toll (2000) for an artificially cemented soil, Asghari et al.(2003) for a cemented coarse-grained alluvium, Rampello et al.(1993) for a natural clay, and Lo and Wardani (2002) and Loet al. (2003) for a silt stabilized by a mixture of cement and flyash. It should be pointed out, however, that there is no general

Geo

mec

hani

cs a

nd G

eoen

gine

erin

g do

wnl

oade

d fr

om w

ww

.tand

fonl

ine.

com

Geomechanics and Geoengineering: An International Journal 11

agreement regarding the effects of cement content on the peakfriction angle. While some researchers reported increase in thefriction angle due to increasing cement content (Saxena et al.1988, Lade and Overton 1989), others reported a parallel move-ment of the failure envelope (i.e. no change in the frictionangle) with the increase of cement content (Wissa et al. 1971,Clough et al. 1981). It is important to note that in previousstudies the effects of cement content on the friction angle ofcemented soils were investigated at relatively low confiningpressures. The results presented in this paper, carried out at awide range of confining pressures, confirm that the peak fictionangle increases with the increase in cement content.

As reported by Coop and Atkinson (1993), Schnaid et al.(2001) and Haeri et al. (2005), the mode of failure is also impor-tant in analyzing the failure behaviour of cemented sand. Thisis because the mode of failure affects significantly the strengthparameters determined from the tests.

Failure modes of selected cemented specimens with differ-ent cement contents at the end of CID tests are presentedin Figure 13. In general, the failure of cemented specimenssheared at lower confining pressures was accompanied by sin-gle shear bands without any significant barrelling while thefailure of cemented specimens sheared at higher confining pres-sures was accompanied by barrelling and multiple shear bands.In other words, the increase in confining pressure changes thefailure mode of cemented sand from brittle to ductile, whichwas also demonstrated by transition of sharp to smooth peakdeviatoric stress in the stress-strain curves (see Figure 8). Thisis consistent with observations made by Asghari et al. (2003)for cemented gravelly sand.

For the microscopic study of the modes of failure, SEM pho-tos of two typical samples with the same cement contents anddry densities (c = 15% and γ = 17.4 kN/m3), sheared at dif-ferent confining pressures (20 MPa and 1 MPa) are shown inFigures 14 and 15, respectively. The failure of specimen shownin Figure 14 was accompanied by barrelling. The failure ofspecimen shown in Figure 15 occurred with the formation ofa single shear band.

Figure 14 shows typical SEM images taken from thebulging section of a specimen with 15% cement sheared atσ ′

3 = 20 MPa. From the SEM image shown in Figure 14, itcan be observed that most bonds are broken and sand particlesare detached from the sand-cement matrix. A few cracks on aparticle can also be seen, which in turn suggests that particlebreakage occurred in this test. Such particle breakage is con-sistent with that observed in uncemented specimens sheared athigh pressures (Marri 2010). However, the comparison of SEMmicrographs of uncemented and cemented sand reveals that theamount of particles crushed during drained shearing is largerfor the uncemented specimens. This suggests that the cementbonds in cemented specimens are weaker than sand particles.Therefore, when a cemented specimen is sheared at high pres-sures most of the breakage occurs within the bonds and the sandparticles are protected from breaking until a very high devia-tor stress is reached. In the case of uncemented specimens, thebonds are absent and the high shear stresses must be sustained

(a)

(b)

(c)

Figure 13. Modes of failure of cemented specimens with different cementcontents: (a) 5%; (b) 10%; and (c) 15%.

entirely by sand particles. As a result, sand particles in unce-mented specimens start to break earlier compared to those incemented specimens (Marri 2010).

Figure 15 shows typical SEM images taken along the shearband of a specimen with 15% cement content sheared at σ ′

3 =1 MPa. It can be seen from Figure 15 that there is detachment ofthe sections of shear plane due to shear force and the deforma-tion of cement bonding is only rigorous along this plane withoutsignificant crushing of particles. This in turn suggests that atlower confining pressures the thrust of the deformation during

Geo

mec

hani

cs a

nd G

eoen

gine

erin

g do

wnl

oade

d fr

om w

ww

.tand

fonl

ine.

com

12 A. Marri et al.

Figure 14. SEM photograph of bulging section of cemented specimen with15% cement content after shearing at σ ′

3 = 20 MPa.

Figure 15. SEM photograph of shear plane section of a cemented specimenwith 15% cement content after shearing at σ ′

3 = 1 MPa.

shearing occurs along the shear plane. However, further SEManalysis of cemented specimens consolidated and sheared todifferent stress and strain levels is required to fully understandthe way in which the micro fabric changes in response to thestress and strain imposed on the soil. This would constitute aseparate research task in the future.

9. Stress-dilatancy relationship

When a dense sand element is sheared, an increase in vol-ume, so-called dilation, occurs due to geometrical constraintsimposed by the fabric against applied stresses. One of the inter-esting aspects of volumetric behaviour of granular materials ishow dilation is gradually suppressed during shearing at higherconfining pressures. Examples of such behaviour for dense

Portaway sand sheared at a wide range of confining pressures(50 kPa to 20 MPa) are shown in Figure 16. Volumetric changecurves of uncemented specimens are shown in Figure 16(a)whereas the volume change of specimens with 5% cement isillustrated in Figure 16(b). It can be seen from Figures 16(a) and16(b) that in both cases the behaviour changes from volumetricdilation at lower confining pressures to compression at higherconfining pressures. As shown in Figure 16, at high confin-ing pressures, dense uncemented and cemented Portaway sandexhibits compression throughout the entire shearing, which issimilar to the volumetric behaviour of loose sand. Thus, thedilation is gradually suppressed by the increasing confiningpressure.

A great deal of attention has been given to the stress-dilatancy relationship of soils (e.g. Rowe 1969, Nova 1982,Bolton 1986, Wood 1990, Li and Dafalias 2000, Been andJefferies 2004). Although there are several stress-dilatancyrelationships in the literature, they retain the general formof Rowe’s original equation, derived from minimum energy

–15

–10

–5

0

5

10

150 10 20

(a)

(b)

30 40

0.05 MPa

0.1 MPa

0.3 MPa

0.5 MPa

1 MPa

4 MPa

8 MPa

12 MPa

20 MPa

Cement = 0% γd = 17.4 kN/m3

Cement = 5% γd = 17.4 kN/m3

–15

–10

–5

0

5

10

150 10 20 30 40

0.05 MPa

0.1 MPa

0.3 MPa

0.5 MPa

1 MPa

4 MPa

8 MPa

12 MPa

20 MPa

ε v (

%)

ε v (

%)

εa (%)

εa (%)

Figure 16. Suppression of the dilatancy by increasing confining pressure forthe specimens with cement contents of: (a) 0%; (b) 5%.

Geo

mec

hani

cs a

nd G

eoen

gine

erin

g do

wnl

oade

d fr

om w

ww

.tand

fonl

ine.

com

Geomechanics and Geoengineering: An International Journal 13

considerations of particle sliding (Rowe 1962). Rowe’s stress-dilatancy relationship is commonly expressed as R = KcD,where R is the ratio of principal stresses, Kc is a constant and D= (1-dεv/dεa) is the rate of dilatancy related to volumetric andmajor principal strain rates.

Following the work of Rowe (1962), the following stress-dilatancy relationship can be used to describe the rate ofdilatancy of a cemented (or bonded) soil:

D =σ ′

1σ ′

3

tan2(

π4 + φ′

f

2

)+ 2c

σ ′3

tan(

π4 + φ′

f

2

) (1)

where, σ ′1 and σ ′

3 are the major and minor principal stresses,respectively, c is the interparticle cohesion, and φ′

f is thefriction angle of soil.

It can be seen from Equation (1) that the dilatancy of acemented soil is influenced by both the interparticle cohesionc and the angle of friction φ′

f, which shows that dilatancy,is inhibited by the presence of cohesion or bonding betweenthe particles of a cemented soil. This has been illustratedby a number of experimental studies (Anagnostopoulos et al.1991, Coop and Atkinson 1993, Cuccovillo and Coop 1999,Schnaid et al. 2001). For example, triaxial compression testson artificially cemented sand carried out by Schnaid et al.(2001) demonstrated that the dilation of artificially cementedsand is also inhibited by interparticle bonding. In addition,experimental results presented in this study as well as otherexperimental studies (Clough et al. 1981, Airey and Fahey1991, Anagnostopoulos et al. 1991, Coop and Atkinson 1993,

Lagioia and Nova 1995, Cuccovillo and Coop 1999, Schnaidet al. 2001) have demonstrated that the stress-strain behaviourof bonded geomaterials changes from brittle to ductile as theconfining stress increases. Since bonding tends to be brittle, thissuggests that the behaviour of a cemented soil is mainly cohe-sive at very low confining stress, changing to purely frictionalas the confining stress increases. This transition is believed tooccur because the bonding bears some of the confining stressand thus ‘weakens’ its effect on the soil skeleton. As shown byYu et al. (2007), based on the data from Schnaid et al. (2001),the stress-dilatancy relationship in Equation (1) can representthis ‘weakening’ effect as the interparticle bonding c is normal-ized by the minor effective principal stress σ ′

3 (which is alsothe confining stress in the conventional triaxial test). However,Schnaid et al. (2001) carried out all their triaxial tests at lowconfining pressures, pc

′ < 100 kPa. This leads to the ques-tion whether the stress-dilatancy relationship can capture thebehaviour of a cemented soil at high pressures.

The stress-dilatancy relationships of Portaway sandare plotted in Figure 17. For the sake of clarity onlyselected results are shown to illustrate typical behaviourat high pressures. Figure 17(a) shows three sets of stress-dilatancy plots for specimens with different cementcontent sheared at confining pressures σ ′

3 of 4, 8 and12 MPa, respectively. Figure 17(b) presents three setsof stress-dilatancy plots for specimens with cement con-tents of 5%, 10% and 15% sheared at different confiningpressures.

It can be seen from Figure 17(a) that for specimens shearedat the same effective confining stress, the effect of extra cement

0.0

0.5

1.0

1.5

2.0C = 5%C = 10%C = 15%

0.0

0.5

1.0

1.5

2.0

–2.0 –1.5 –1.0

(–δεv/δεq)

–0.5 0.0 0.5

C = 5%C =10%C =15%

0.0

0.5

1.0

1.5

2.0

(q/p

′)(q

/p′)

(q/p

′)

C = 5%C = 10%C = 15%

(a) (b)

σ′3 = 4 MPa

σ′3 = 8 MPa

σ′3 = 12 MPa

0.0

0.5

1.0

1.5

2.04 MPa8 MPa12 MPa

Cement = 10%

0.0

0.5

1.0

1.5

2.0

–2.0 –1.5 –1.0 –0.5 0.0 0.5

4 MPa8 MPa12 MPa

Cement = 15%

0.0

0.5

1.0

1.5

2.04 MPa8 MPa12 MPa

Cement = 5%(q/p

′)(q

/p′)

(q/p

′)

(–δεv/δεq)

Figure 17. Typical stress-dilatancy relationships obtained for cemented specimens at high pressures: (a) effect of cement content; (b) effect of confining pressure.

Geo

mec

hani

cs a

nd G

eoen

gine

erin

g do

wnl

oade

d fr

om w

ww

.tand

fonl

ine.

com

14 A. Marri et al.

content is to move the plot of q/p′ versus (-dεv/dεq) upward.As a result, the stress ratios at the peak and ultimate statesincrease with increasing cement content. This means that theeffect of cementation, which introduces the component ofcohesion into the stress-dilatancy relationship, is to increasebonding between particles at a given stress. For specimens withthe same cement content sheared at different σ ′

3, the plot isgenerally shifted to the left as the effective confining stressincreases (Figure 17b). This means that the effect of confin-ing pressure is to reduce or prevent dilatancy at a given stress,as shown previously in Figure 16.

The patterns illustrated in Figure 17 are consistent withexperimental data reported by Schnaid et al. (2001), DallaRossa et al. (2008) and Consoli et al. (2009), and the computedbehaviour using the ‘Clay and Sand Model’ (Yu 1998) reportedby Yu et al. (2007) at low confining pressures. However, differ-ences in stress-dilatancy behaviours at high pressures are muchsmaller compared to low pressures. This again shows that effectof cementation diminishes when specimens are sheared at highconfining pressures.

Figure 17 also shows that most specimens sheared at highpressures displayed a compressive behaviour with the rate ofcompression reducing to reach zero at ultimate state indicatedby the vertical axis. Some specimens with cement content of10% and 15% also exhibited dilative behaviour. The amount ofdilation, however, was very small, as shown earlier in Figures 8and 16.

10. Conclusions

In this paper, results obtained from isotropic compression anddrained triaxial high pressure tests on artificially cementedPortaway sand are presented. The effects of cement contentand confining pressure on the isotropic compression, stress-strain and volume change behaviour in drained shearing, failurecharacteristics and stress-dilatancy relationship are discussed.Based on the experimental results, the following conclusionscan be drawn:

(1) The effect of initial void ratio on the behaviour ofuncemented and cemented sand in isotropic compres-sion diminishes at high pressures. Hence, the isotropiccompression curves of specimens with the same cementcontents but different initial void ratios converge at highpressures. In other words, the effect of initial void ratiobecomes less significant at high confining pressures.On the other hand, the experimental data obtained inthis study showed that the effect of cement content onthe isotropic compression of cemented sand becomesmore significant at high confining pressures where thecompressibility of specimens reduces significantly withincreasing cement content. These observations are consis-tent with those previously published in the literature (Lade

and Overton 1989, Rotta et al. 2003, Consoli et al. 2005,dos Santos et al. 2010)

(2) Isotropically consolidated drained compression testsshow that both cement content and confining pressureaffect the stress-strain and volume change behaviour ofcemented sand. An increase in cement content increasesthe peak deviatoric stress but reduces the amount ofcompression during shearing (i.e. increases the amountof dilation). On the other hand, an increase in confin-ing pressure increases the peak deviatoric stress and theamount of compression during shearing. Therefore, thestress-strain behaviour becomes increasingly ductile withincreasing confining pressure. The experimental data alsoshowed that the influence of cementation is greater at lowconfining stresses and it is gradually suppressed by theincreasing confining pressure.

(3) Cement content and confining pressure also affect failurecharacteristics of Portaway sand. It was observed that thefailure parameters of uncemented dense sand determinedat the high confining pressure coincide with the criticalstate parameters determined for loose specimens obtainedat low pressures. Furthermore, failure states of cementedspecimens at high pressures approached asymptoticallythe critical state of uncemented sand.

(4) The stress-dilatancy relationships of cemented sand areaffected by cement content and confining pressure.The presence of cementation introduces the componentof cohesion into the stress-dilatancy relationship andincreases bonding between particles at a given stress. As aresult, the stress ratios at the peak and ultimate statesincrease with increasing cement content. An increase inconfining pressure prevents dilatancy at a given cementcontent. This is consistent with experimental data reportedby Schnaid et al. (2001) and the computed behaviourreported by Yu et al. (2007) at low confining pressures.

(5) From stress-strain curves and photographs of the speci-mens it was observed that there exists strain localizationand subsequent shear banding. However, no clear trendin the formation and width of shear band was deter-mined by the increase in the cement content and confiningpressures. Complex modes of failures such as brittle, tran-sitional, and ductile were observed depending upon thecement content and confining pressure.

(6) The SEM analysis carried out in this study shows thatthe isotropic consolidation of cemented specimens tohigh pressures does not cause any significant damageof the specimen fabrics. Most crushing of sand particlesand breakage of cement bonds appear to occur duringshearing. The amount of particles broken during shear-ing at high pressures increases with increasing confiningpressure and reduces with increasing cement content.However, further study is needed to fully understand theway in which the micro fabric changes in response to thestress and strain imposed on the cemented soil under highconfining pressures.

Geo

mec

hani

cs a

nd G

eoen

gine

erin

g do

wnl

oade

d fr

om w

ww

.tand

fonl

ine.

com

Geomechanics and Geoengineering: An International Journal 15

References

Airey, D.W. and Fahey, M., 1991. Cyclic response of calcareoussoil from the northwest shelf of Australia. Géotechnique, 41(1),101–122.

Airey, D.W., 1993. Triaxial testing of naturally cemented carbonatesoil. Journal of Geotechnical Engineering, 119(9), 1379–1398.

Anagnostopoulos, A.G., Kalteziotis, N., Tsiambaos, G.K. andKavvadas, M., 1991. Geotechnical properties of the CorinthCanal marls. Geotechnical and Geological Engineering, 9, 1–26.

Asghari, E., Toll, D.G. and Haeri, S.M., 2003. Triaxial behaviour ofa cemented gravely sand, Tehran alluvium. Geotechnical andGeological Engineering, 21, 1–28.

Been, K. and Jefferies, M., 2004. Stress-dilatancy in very loose sand.Canadian Geotechnical Journal, 41, 972–989.

Bopp, P.A. and Lade, P.V. 2005. Relative density effects on undrainedsand behaviour at high pressures, Soils and Foundations, 45(1),15–26.

Bolton, M.D., 1986. The strength and dilatancy of sands.Geotechnique, 20(1), 65–78.

British Standard, 1990. Methods of test for soils for civil engineeringpurposes, BS 1377. London: British Standard Institution.

Brown, S.F., Juspi, S. and Yu, H.S., 2008. Experimental observationsand theoretical predictions of shakedown is soils under wheelloading. In: E. Ellis, H.S. Yu, G. McDowell, A. Dawson andN. Thom, eds. Advances in Transportation Geotechnics. London:Taylor and Francis, London, 707–712.

Chu, J., Lo, S.-C.R. and Lee, I.K., 1996. Strain softening and shearband formation of sand in multi-axial testing. Géotechnique,46(1), 63–82.

Chu, J. and Wanatowski, D., 2008. Instability conditions of loose sandin plane-strain. Journal of Geotechnical and GeoenvironmentalEngineering, 134(1), 136–142.

Chu, J. and Wanatowski, D., 2009. Effect of loading mode on strainsoftening and instability behavior of sand in plane-strain tests.Journal of Geotechnical and Geoenvironmental Engineering,135(1), 108–120.

Clough, G.W., Sitar, N., Bachus, R.C. and Rad, N.S., 1981.Cemented sands under static loading. Journal of the GeotechnicalEngineering Division, 107(6), 799–817.

Consoli, N.C., Rotta, G.V. and Prietto, P.D.M., 2000. The influenceof curing under stress on the triaxial response of cemented soils.Geotechnique, 50(1), 99–105.

Consoli, N.C., Casagrande, M.D.T. and Coop, M.R., 2005. Effectof fiber reinforcement on the isotropic compression behav-ior of a sand. Journal of Geotechnical and GeoenvironmentalEngineering, 131(11), 1434–1436.

Consoli, N.C., Rotta, G.V. and Prietto, P.D.M., 2006. Yielding-compressibility-strength relationship for an artificially cementedsoil cured under stress. Geotechnique, 56(1), 69–72.

Consoli, N.C., Vendruscolo, M.A., Fonini, A. and Dalla Rosa,F., 2009. Fiber reinforcement effects on sand considering awide cementation range. Geotextiles and Geomembranes, 27(3),196–203.

Coop, M.R. and Atkinson, J.H., 1993. The mechanics of cementedcarbonate sands. Géotechnique, 43(1), 53–67.

Cuccovillo, T. and Coop, M.R., 1997. Yielding and pre-failure defor-mation of structured sands. Geotechnique, 47(3), 491–508.

Cuccovillo, T. and Coop, M.R., 1999. On the mechanics of structuredsands. Géotechnique, 49(6) 741–760.

Dalla Rosa, F., Consoli, N.C. and Baudet, B.A., 2008. An experimentalinvestigation of the behaviour of artificially cemented soil curedunder stress. Géotechnique, 58(8), 675–679.

dos Santos, A.P.S., Consoli, N.C., Heineck, K.S. and Coop,M.R., 2010. High-pressure isotropic compression tests onfiber-reinforced cemented sand. Journal of Geotechnical andGeoenvironmental Engineering, 136(6), 885–890.

Haeri, S.M., Hamidi, A. and Tabatabaee, N., 2005. The effect of gyp-sum cementation on the mechanical behaviour of gravely sands.Geotechnical Testing Journal, 28(4), 380–390.

Haeri, S.M., Hamidi, A., Hosseini, S.M., Asghari, E. and Toll, D.G.,2006. Effect of cement type on the mechanical behaviour of agravely sand. Geotechnical and Geological Engineering, 24(2),335–360.

Huang, J.T. and Airey, D.W., 1993. Effects of cement and den-sity on an artificially cemented sand. In: A. Anagnostopoulos,F. Schlosser, N. Kalteziotis and R. Frank, eds., GeotechnicalEngineering of Hard Soils-Soft Rocks, Rotterdam: Balkema,553–560.

Huang, J.T. and Airey, D.W., 1998. Properties of artificially cementedcarbonate sand. Journal of Geotechnical and GeoenvironmentalEngineering, 124(6), 492–499.

Ismail, M.A., Joer, H.A., Randolph, M.F. and Meritt, A., 2002.Cementation of porous materials using calcite. Géotechnique,52(5), 313–324.

Jardine, R.J., Symes, M.J. and Burland, J.B., 1984. The measurementof soil stiffness in the triaxial apparatus. Géotechnique, 34(3),323–340.

Jefferies, M. and Been, K., 2000. Implications for critical state the-ory from isotropic compression of sand, Géotechnique, 50(4),419–429.

Ladd, R.S., 1978. Preparing test specimens using undercompaction.Geotechnical Testing Journal, 1(1), 16–23.

Lade, P.V. and Overton, D.D., 1989. Cementation effects in frictionalmaterials. Journal of the Geotechnical Engineering Division,115(10), 1373–1387.

Lade, P.V. and Yamamuro, J.A., 1996. Undrained sand behaviour inaxisymmetric tests at high pressures. Journal of GeotechnicalEngineering, 122(2), 120–129.

Lade, P.V., Yamamuro, J.A. and Bopp, P.A., 1996. Significance ofparticle crushing in granular materials. Journal of GeotechnicalEngineering, 122(4), 309–316.

Lagioia, R. and Nova, R., 1995. An experimental and theoreticalstudy of the behaviour of a calcarenite in triaxial compression.Géotechnique, 45(4), 633–648.

La Rochelle, P., Leroueil., S., Trak, B., Blais-Leroux, L. and Tavenas,F. 1988. Observational approach to membrane and area correc-tions in triaxial tests. In: R.T. Donaghe, R.C. Chaney, and M.L.Silver, eds., Advanced Triaxial Testing of Soil and R, ASTM STP977, Philadelphia: American Society for Testing and Materials,715–731.

Lee, K.L. and Farhoomand, I., 1967. Compressibility and crushingof granular soils in anisotropic triaxial compression. CanadianGeotechnical Journal, 4(1), 68–86.

Lee, K.L. and Seed, H.B., 1967. Drained strength characteristics ofsands, Journal of the Soil Mechanics and Foundations Division,93(6), 117–141.

Leroueil, S. and Vaughan, P.R., 1990. The general and congruenteffects of structure in natural soils and weak rocks. Géotechnique,40(3), 467–488.

Geo

mec

hani

cs a

nd G

eoen

gine

erin

g do

wnl

oade

d fr

om w

ww

.tand

fonl

ine.

com

16 A. Marri et al.

Leroueil, S., 2001. Natural slopes and cuts: movement and failuremechanisms. Géotechnique, 51(3), 197–243.

Li, X.S. and Dafalias, Y.F., 2000. Dilatancy for cohesionless soils.Geotechnique, 50(4), 449–460.

Lo, S.-C.R. and Wardani, S.P.R., 2002. Strength and dilatancy ofa silt stabilized by a cement and fly ash mixture. CanadianGeotechnical Journal, 39, 77–89.

Lo, S.-C.R., Lade, P.V. and Wardani, S.P.R., 2003. An experimen-tal study of the mechanics of two weakly cemented soils.Geotechnical Testing Journal, 26(3), 328–341.

Marri, A., 2010. The mechanical behaviour of cemented granu-lar materials at high pressures, Thesis (PhD), University ofNottingham, UK.

Mesri, G. and Vardhanabhuti, B., 2009. Compression of granularmaterials. Canadian Geotechnical Journal, 46(4), 369–392.

Malandraki, V. and Toll, D.G., 2000. Drained probing triaxial tests ona weakly bonded artificial soil. Géotechnique, 50(2), 141–151.

McDowell, G.R., Bolton, M.D. and Robertson, D., 1996. The frac-tal crushing of granular materials. Journal of the Mechanics andPhysics of Solids, 44, 2079–2102.

McDowell, G.R. and Bolton, M.D. 1998. On the micromechanics ofcrushable aggregates. Geotechnique, 48(5), 667–679.

Nova, R., 1982. A constitutive model under monotonic and cyclicloading. In: G. Pande and O.C. Zienkiewicz, eds., SoilMechanics- Transient and Cyclic Loads, New York: John Wiley& Sons Ltd., 343–373.

Rad, N.S. and Clough, G.W., 1985. Static behaviour of variablycemented beach sands. In: R.C. Chaney and K.R. Demars, eds.,Strength Testing of Marine Sediments: Laboratory and In-SituMeasurements, ASTM STP 883. Philadelphia: ASTM, 306–317.

Rampello, S., Viggiani, G. and Georgiannou, V.N., 1993. Strengthand dilatancy of natural and reconstituted Vallerica clay. In: A.Anagnostopoulos, F. Schlosser, N. Kalteziotis and R. Frank, eds.,Geotechnical Engineering of Hard Soils-Soft Rocks, Rotterdam:Balkema, 761–768.

Rotta, G.V., Consoli, N.C., Prietto, P.D.M., Coop, M.R. and Graham,J., 2003. Isotropic yielding in an artificially cemented soil curedunder stress. Géotechnique, 53(5), 493–501.

Rowe, P.W., 1962. The stress dilatancy relationship for static equilib-rium of an assembly of particles in contact. Proceedings of theRoyal Society, A269, 500–527.

Rowe, P.W., 1969. The relation between the shear strength ofsands in triaxial compression, plane strain, and direct shear.Géotechnique, 19(1), 75–86.

Saxena, S.K., Reddy, K.R. and Avramidis, A.S., 1988. Liquefactionresistance of artificially cemented sand. Journal of GeotechnicalEngineering, 114(12), 1395–1413.

Schnaid, F., Prietto, P.D.M. and Consoli, N.C., 2001. Characterizationof cemented sand in triaxial compression. Journal ofGeotechnical and Geoenvironmental Engineering, 127(10),857–868.

Wanatowski, D. and Chu, J., 2008, Effect of specimen prepara-tion method on the stress-strain behavior of sand in plane-strain compression tests. Geotechnical Testing Journal, 31(4),308–320.

Wang, Y.H. and Leung, S.C., 2008. A particulate-scale investigation ofcemented sand behaviour, Canadian Geotechnical Journal, 45,29–44.

Wissa, A.E.Z., Christian J.T., Davis E.H. and Heiberg, S.,1971. Consolidation at constant rate of strain. Journal ofthe Soil Mechanics and Foundations Division, 97(SM10),1393–1411.

Wood, D.M., 1990. Soil behaviour and critical state soil mechanics.Cambridge: Cambridge University Press.

Yamamuro, J.A. and Lade, P.V., 1996. Drained sand behaviour inaxisymmetric tests at high pressures. Journal of GeotechnicalEngineering, 122(2), 109–119.

Yu, H.-S., 1998. CASM: a unified state parameter model for clayand sand. International Journal for Numerical and AnalyticalMethods in Geomechanics, 22(8), 621–653.

Yu, H.S., Khong, C.D., Wang, J. and Zhang, G., 2005. Experimentalevaluation and extension of a simple critical state model for sand.Granular Matter, 7, 213–225.

Yu, H.S., Tan, S.M. and Schnaid, F., 2007. A critical state frame-work for modelling bonded materials. Geomechanics andGeoengineering, 2(1), 61–74.

Geo

mec

hani

cs a

nd G

eoen

gine

erin

g do

wnl

oade

d fr

om w

ww

.tand

fonl

ine.

com