th

Proceedings of the 18 International Conference on Soil Mechanics and Geotechnical Engineering, Paris 2013
th
Proceedings of the 18 International Conference on Soil Mechanics and Geotechnical Engineering, Paris 2013

Engineering properties of an expansive soil

Propriétés mécaniques d'un sol gonflant

Azam S., Ito M., Chowdhury R.

Environmental Systems Engineering, University of Regina, Regina, SK, Canada

ABSTRACT: The engineering properties of an expansive soil were investigated using in situ and compacted samples. The soil water
characteristic curve was found to have a bimodal shape composing of two air entry values: an initial low value corresponding to
macroporous drainage followed by a high value related to microporous flow. Likewise, the shrinkage curve was found to be S-shaped
and included a low structural shrinkage followed by a sharp decline during normal shrinkage and then by a low decrease during
residual shrinkage.

RÉSUMÉ: Les propriétés mécaniques d'un sol gonflant ont été étudiées à l'aide d'échantillons compactés obtenues sur le terrain. La
courbe caractéristique en eau du sol s'est révélé avoir une forme bimodale avec deux valeurs d'entrée d'air: une valeur initiale
correspondant à un faible drainage macroporeux suivi d'une valeur élevée par rapport à l'écoulement microporeux. De même, la
courbe de rétrécissement est en forme de S et inclus un faible rétrécissement structurelle suivie d'une forte baisse pendant le
rétrécissement normal, suivie d’une une diminution faible durant le rétrécissement résiduel.
KEYWORDS: expansive soil, soil water characteristic curve, shrinkage curve.

1 INTRODUCTION on the correlation of these properties with morphological

observations using “cryogenic specimens” in a scanning
The capital of Saskatchewan, Canada, is founded on a glacio- electron microscope (Ito and Azam 2013). These studies
lacustrine clay deposit that exhibits significant volume changes concluded that the geologically-induced soil structure governs
due to seasonal weather variations. Alternate swelling and the water migration and the swell-shrink patterns through the
shrinkage in the expansive Regina clay has impaired civil expansive clay. The present study extends the current body of
infrastructure such as transportation networks (Kelly et al. knowledge to compacted soils thereby capturing the effect of
1995), residential, industrial, and commercial facilities (Ito and soil state on the the properties of local clays. Overall, a
Azam 2010), and water supply and sewage collection systems generalized theoretical framework is developed to understand
(Hu and Hubble 2005). Damages to engineered facilities are the behavior of expansive soils.
clearly manifested in the form of differential heave in roadways The main objective of this paper is to understand the
and sidewalks, inclined cracking in slab-on-grade basements engineering properties of Regina clay using in situ and
and masonry walls, and fatigue and breakage in underground compacted specimens. Geotechnical index properties were
storage tanks and buried pipelines. The associated repair cost is determined for preliminary soil assessment. The soil water
usually quite enormous. For example, the breakage rate in the characteristic curve (SWCC) was determined to investigate the
850 km long water supply network in the city has now reached a water retention capacity of the soil. Likewise, the shrinkage
30-year maximum of 0.27 breaks/km/year, costing more than $2 curve was determined to correlate volume changes with soil
million in annual maintenance. Furthermore, the city is saturation and desaturation.
currently going through a period of infrastructure development
including mega-projects such as the Global Transportation Hub
and the Downtown Covered Stadium with a 55000-seat 2 RESEARCH METHODOLOGY
capacity. Clearly, there is a need to study site-specific soil
properties for the continuous maintenenace and improved The expansive clay was retrieved from a local soil deposit that
design of civil infrastructure systems in Regina. was found to be desiccated in early Fall and exhibited extensive
Generally, volume changes in expansive soils are derived fissuring oriented in all directions. High quality undisturbed
from clay minerals that undergo hydration due to rainfall and samples were obtained using the ASTM Standard Practice for
dehydration due to evaporation. This process is governed by the Thin-Walled Tube Sampling of Soils for Geotechnical Purposes
attraction of bipolar water molecules to the negatively charged (D1587-08) from a depth of 0.6 m to 1.2 m. Likewise, disturbed
clay particles possessing high specific surface areas (Mitchell samples were obtained from bore cuttings according to the
and Soga 2005). However, water access to individual clay ASTM Standard Practice for Soil Investigation and Sampling
particles primarily depends on the following two factors: (i) soil by Auger Borings (D1452-09). All of the specimens were
structure (mircopores within soil peds and macropores between plastic-wrapped and wax-coated and the entire collection was
the soil peds) and (ii) soil state (void ratio and degree of transported and stored at the University of Regina as per the
saturation). These parameters are respectively governed by ASTM Standard Practice for Preserving and Transporting Rock
parent geology and construction practices prevalent in an area. Core Samples (D5079-08). The latter samples were compacted
Recent research on local expansive clays has focused on the in acordance with the ASTM Standard Test Methods for
determination of unsaturated soil properties using “undisturbed Laboratory Compaction Characteristics of Soil Using Modified
samples” from the geological deposit (Azam and Ito 2011) and Effort (D1557-12).

1991
 Proceedings of the 18th International Conference on Soil Mechanics and Geotechnical Engineering, Paris 2013

th
Proceedings of the 18 International Conference on Soil Mechanics and Geotechnical Engineering, Paris 2013

The geotechnical index properties were determined thereby bringing the water height back to the initial reading.
according to the ASTM test methods as follows: (i) water The displaced water mass was determined by weighing the
content (w) by the Standard Test Methods for Laboratory graduated syringe before and after water filling and recording
Determination of Water (Moisture) Content of Soil and Rock by the difference. This quantity was readily converted to water
Mass (D2216-05); (ii) dry unit weight (γd) by the Standard Test volume representing the volume of the wax-coated soil. The
Method for Density of Soil in Place by the Drive-Cylinder volume of soil was obtained from the difference of volume of
Method (D2937-10) for the in situ sample and by the above- the wax coated sample and the volume of wax (mass/0.9). A
mentioned method for the compacted sample; (iii) specific 7.4% correction was applied to account for the underestimation
gravity (Gs) by the Standard Test Methods for Specific Gravity due to air entrapment at the soil-wax interface, as suggested by
of Soil Solids by Water Pycnometer (D854-10); (iv) liquid limit Prakash et al. (2008). The sample mass was also determined to
(wl), plastic limit (wp) and plasticity index (Ip) by the Standard estimate the bulk unit weight of the soil that, in turn, was
Test Methods for Liquid Limit, Plastic Limit, and Plasticity converted to the void ratio using basic phase relationships.
Index of Soils (D4318-10); and (v) grain size distribution
(GSD) by the Standard Test Method for Particle-Size Analysis
of Soils (D422-63(2007)). The entire GSD data is not given in 3 RESULTS AND DISCUSSION
this paper.
The SWCC was determined according to the ASTM Table 1 summarizes the geotechnical index properties of the
Standard Test Methods for Determination of the Soil Water investigated soil. The water content and the dry unit weight of
Characteristic Curve for Desorption Using a Hanging Column, the in situ sample were found to be 31% and 1.34 g/cm3,
Pressure Extractor, Chilled Mirror Hygrometer, and/or respectively. In early Fall when the sample was collected, the
Centrifuge (D6836-02(2008)e2) on 10 mm thick samples soil generally experiences a net water deficit given the semi-arid
obtained from both the undisturbed core and the compacted climate prevalent in the region. This was evident from the
sample. Predetermined values of matric suction were applied unsaturated (S = 82%) state of the sample in the field: the field
using pressure plate and pressure membrane extractors void ratio was calculated to be 1.05. Similar initial conditions
manufactured by Soil Moisture Equipment Inc. These (w = 38% and d = 1.29 g/cm3) were chosen for the compacted
equipment included the following: (i) a 5 bar pressure plate sample to obtain comparable data: the corresponding saturation
extractor (Model 1600) for up to 200 kPa suction; (ii) a 15 bar and void ratio were found to be 86% and 1.18, respectively. The
pressure plate extractor (Model 1500F1) for suction values high liquid limit and plastic limit indicate the high water
ranging from 300 kPa to 500 kPa; and (iii) a 100 bar pressure adsorption capability of the clay. These values are attributed to
membrane extractor (Model 1020) for suction values between the presence of expansive clay minerals such as smectite,
2000 kPa and 7000 kPa. The porous plates and the cellulose hydrous mica, and chlorite (Ito and Azam 2009). Likewise, the
membranes were submerged in distilled and de-aired water for clay size fraction (material finer than 0.002 mm) was found to
24 hours to expel air bubbles. Thereafter, the specimens along be around 65%. The fine grained nature of the soil suggests a
with the retaining ring were placed on their respective porous high water retention capacity. The calculated soil activity (A =
plate or cellulose membrane and allowed to saturate. Next, the Ip / C) of about 0.8 is associated with moderate swelling.
excess water was removed and each plate or membrane was Overall, the soil was classified as CH (clay with high plasticity)
placed in the designated extractor. For each suction value, the according to the Unified Soil Classification System (USCS).
expelled water from the samples was monitored in a graduated
Table 1. Summary of geotechnical index properties
burette. When two consecutive readings nearly matched over a
24 hour period, the test was terminated and the sample water Property In situ Compacted
content was determined.
The dew point potentiameter (WP4-T) was used for suction Water Content, w (%) 31 38
measurement at low water content corresponding to total
Dry Unit Weight, d (g/cm3) 1.34 1.29
suction values greater than 7000 kPa. The sampling cup was
half filled with soil to ensure accurate suction measurement Specific Gravity, Gs 2.75 2.74
(Leong et al. 2003) by using about 5 mg of material with a
known water quantity. The unsaturated sample was forwarded Void Ratio, e* 1.05 1.18
to the head space of the sealed measurement chamber, set at
Degree of Saturation, S (%)† 81 86
25°C temperature, through a sample drawer and was allowed to
equilibrate with the surrounding air. Equilibration was usually Liquid Limit, wl (%) 83 77
achieved within 10 min to 20 min, as detected by condensation
on a mirror and measured by a photoelectric cell. From Plastic Limit, wp (%) 30 27
knowledge of the universal gas constant, R (8.3145 J/mol°K), Plasticity Index, Ip (%) 53 50
sample temperature, T (°K), water molecular mass, X (18.01
kg/kmol), and the chamber relative humidity, p/po, soil suction Clay Size Fraction, C (%) 66 64
was calculated (ψ = RT/X ln (p/po)) and displayed on the
potentiameter screen. The water content of the soil was USCS Symbol CH CH
measured as described earlier. * e = (Gs w / d) - 1
The shrinkage curve was determined in accordance with the †
S = w Gs / e
ASTM Standard Test Method for Shrinkage Factors of Soils by
the Wax Method (D4943-08). To obtain the void ratio, the Figures 1 shows SWCC with gravimetric water content. The
volume of soil specimens was determined using the water samples were put in a water tub for one week and the water
displacement method. Each specimen was coated with molten content measured 38% for the in situ sample and 46% for the
microcrystalline wax (Gs = 0.9) and allowed to cool down at compacted sample. Irrespective of the initial water content, the
room temperature. After wax solidification, the sample was SWCC data fitted well to bimodal distributions with two air
submerged in a 250 mL graduated cylinder that was filled with entry values: a lower value (10 kPa) corresponding to drainage
distilled water. The water height in the cylinder was carefully through fissures followed by a higher value (300 kPa and 100
recorded using a Vernier caliper before and after sample kPa for the two samples, respectively) associated with seepage
submersion in the cylinder. A graduated syringe was used to through the soil matrix. When the samples were gradually
remove the increased amount of water displaced by the sample desaturated, air first entered into the fissures at low suction.

200
 Technical Committee 101 - Session I / Comité technique 101 - Session I

th
Proceedings of the 18 International Conference on Soil Mechanics and Geotechnical Engineering, Paris 2013

Although the fissures are sealed due to hydration of clay data depicted a bimodal function with a fissure air entry value
minerals, these discontinuities have much lower tensile of 10 kPa. For the in situ sample, the average degree of
strengths than the soil aggregates (Azam and Wilson 2006). saturation decreased due to drainage through fissures and
This led to a quick drainage through these paths of least reached about 70% when most of the fissures were filled with
resistance. Subsequent application of suction affected the soil air. The corresponding saturation for the compacted sample was
aggregates and eventually forced air to enter into the pore found to be around 85% indicating the presence of relatively
system of the aggregate. The upward SWCC shift of the smaller and less frequent discontinuities. Once the fissures were
compacted sample with respect to the in situ sample is attributed desaturated, water flow had to occur through voids in the soil
to the high initial water content of the former sample. Upon aggregates. Microporous drainage required a high suction (6000
water inundation, this sample favored particle hydration because kPa) for air to enter into the soil matrix. In contrast to Figure 1,
of a comparatively looser state (e = 1.18). Likewise, the the significantly higher matrix air entry value in this figure is
relatively homogeneous structure of the compacted sample attributed to the gradual decrease in soil volume with increasing
correlated well with its smaller difference between the two air suction. As explained later, this volume reduction is primarily
entry values. The effect of soil structure was eliminated when due to reduced fissure sizes and, as such, was not captured in
the two curves merged at higher suction. Desaturation occurred Figure 2 that is based on an average void ratio.
at an increased rate up to residual suction of 2000000 kPa (w =
5%) and the curves finally joined the abscissa at 106 kPa. 100

50
Fissure AEV 80
Fissure

Degree of Saturation (%)

Matrix AEV AEV
40
Gravimetric Water Content (%)

60
Matrix
AEV
30
Fissure 40
AEV
20 Matrix
AEV 20
In situ soil
10 Compacted soil
In situ soil 0
Compacted soil 100 101 102 103 104 105 106
0 Soil Suction (kPa)
100 101 102 103 104 105 106
Figure 3. SWCC with degree of saturation
Soil Suction (kPa)
Figure 1. SWCC with gravimetric water content The SWCC given in the form of water content versus matric
suction is the most accurate representation for expansive soils.
Figure 2 presents the SWCC in the form of void ratio as a This is because gravimetric water content (measured for each
function of soil suction. The measured data closely followed suction value) captures the water drainage through fissures and
unimodal distributions showing a single air entry value (300 is independent of volume changes due to water adsorption by
kPa for the in situ sample and 100 kPa for the compacted clay particles. Likewise, the SWCC represented in the form of
sample). The latter sample plotted at a higher void ratio at low degree of saturation versus soil suction is most suitable for
suction and merged with the in situ sample at 500 kPa. understanding volume decrease in fissures due to suction
Thereafter, the singular curve exhibited a sharp decrease in void application. Marinho (2005) reported that for plastic soils,
ratio up to the residual condition, became asymptotic to the drainage through water filled pores is associated with pore
abscissa after the residual suction and never reporting to 106 kPa compressibility due to capillarity and this phenomenon renders
on complete drying. This is because void ratio of a soil pertains such soils to remain saturated over higher suction values. This
to an average value for all voids and does not differentiate representation implies that the expansive soil aggregates remain
between inter-aggregate fissures and intra-aggregate pores. saturated over a wider range of suction generally prevalent in
Clearly, SWCC representation in the form of void ratio is not the field, as postulated by Fityus and Buzzi (2008).
suitable for expansive soils. Figure 4 shows the shrinkage curve for the investigated
expansive clay. Theoretical lines representing various average
1.4 saturation degrees were obtained from basic phase relationships
AEV and using Gs = 2.75. The initially unsaturated samples were first
wetted to achieve close to saturation conditions and
1.2
subsequently desaturated by applying different suction values.
The void ratio and water content of each sample were
1 determined as described earlier in this paper. The data depicted
Void Ratio

in Figure 4 indicate S-shaped shrinkage curves for both sample

AEV types and represent the progressive drying of the investigated
0.8 expansive soil. The curves are composed of an initial low
structural shrinkage followed by a sharp decline during normal
shrinkage and then by a low decrease during residual shrinkage
0.6 (Haines, 1923). During structural shrinkage, water within the
In situ soil
fissures and some of the larger and relatively stable voids is
Compacted soil
0.4
removed such that the decrease in soil volume is less than the
volume of water lost. Volume decrease in soil is equal to the
100 101 102 103 104 105 106 volume of water lost during normal shrinkage thereby leading to
Soil Suction (kPa)
Figure 2. SWCC with void ratio a 45o straight line, which is almost parallel to the 100%
Figure 3 gives the SWCC in the form of degree of saturation saturation line. This suggests that drainage primarily takes place
versus suction. Similar to Figure 1, the laboratory measured through the soil matrix in the normal shrinkage zone. During

201
 Proceedings of the 18th International Conference on Soil Mechanics and Geotechnical Engineering, Paris 2013

th
Proceedings of the 18 International Conference on Soil Mechanics and Geotechnical Engineering, Paris 2013

residual shrinkage, air enters the pores close to the shrinkage 4 CONCLUSIONS
limit and pulls the particles together due to suction. This leads
to a further decrease in soil volume albeit lower than the volume The engineering properties of a typical expansive soil (from
of water lost. Furthermore, the high volume change and the Regina, Saskatchewan, Canada) were investigated under in situ
closeness to the saturation line during normal shrinkage and compacted conditions. For both sample types, the clay
corresponded well with the comparatively looser state and a behavior was characterized by its internal structure comprising
relatively homogeneous structure of the compacted sample. of fissures and aggregates. The SWCC using water content on
Finally, the observed shrinkage curve is reversible because the the ordinate showed a bimodal distribution with two air entry
in situ soil has undergone numerous swell-shrink cycles since values: a lower value (10 kPa) corresponding to drainage
deposition. Likewise, Tripathy et al. (2002) reported that through fissures followed by a higher value (300 kPa for the in
equilibrium conditions are usually attained after about four situ sample and 100 kPa for the compacted sample) associated
cycles in compacted soils. with seepage through the soil matrix. Sample type became
irrelevant when the flow started to occur through the soil
1.4 matrix. The matrix air entry value was found to be about 6000
% kPa when the SWCC was plotted in the form of the degree of
50 %

%
%

80
70

0%
60

= 10
S saturation versus soil suction. Likewise, the shrinkage curve
S=
S=

S=

S=
was found to be S-shaped and included a low structural
shrinkage followed by a sharp decline during normal shrinkage
1.0
and then by a low decrease during residual shrinkage. The
extent of volume change that depends on the initial void ratio
Void Ratio

must be calculated from the reversible swell-shrink curve.

0.6
5 ACKNOWLEDGEMENTS

In situ soil The authors acknowledge the material and financial support
Compacted soil provided by the Saskatchewan Ministry of Highways and
0.2 Infrastructure and the University of Regina for providing
0 10 20 30 40 50 laboratory space.
Gravimetric Water Content (%)
Figure 4. Shrinkage curve
6 REFERENCES
Theoretically, the shrinkage curve comprises of two straight
lines: a sloped line closely following the S = 100% line that Azam S. and Ito, M. 2011. Unsaturated soil properties of a fissured
expansive clay. Proceedings, 64th Canadian Geotechnical
joins a horizontal line at a void ratio associated with the
Conference, Toronto, Canada. 313:1-5.
shrinkage limit of the soil. This means that soils essentially Azam, S. and Wilson, G.W. 2006. Volume change behavior of a
remain saturated up to the shrinkage limit following a J-shaped fissured expansive clay containing anhydrous calcium sulfate.
curve. Due to the presence of fissures, the investigated soil (in Proceedings, 4th International Conference on Unsaturated Soils,
both in situ and compacted conditions) exhibited deviations Carefree, Arizona, USA. 1, 906-915.
from this theoretical behavior. Fityus, S. and Buzzi, O. 2008. The place of expansive soils in the
The definition the degree of saturation for expansive soils is frameworks of unsaturated soil mechanics. Applied Clay Science
not straight forward. Since such soils consist of discontinuities 43, 150-155.
and soil aggregates, the calculated degree of saturation pertains Gens, A. and Alonso, E.E. 1992. A framework for the behaviour of
unsaturated expansive clays. Canadian Geotechnical Journal 29,
to an average value for the entire soil mass. This definition was
1013-1032.
used in Figure 3 because it permits easy calculations. A more Haines, W.B. 1923. The volume change associated with variations of
accurate approach is to consider only the soil aggregates as water content in soil. Journal of Agricultural Science 13, 296-310.
saturated up to the matrix air entry value of Figure 3 and the Hu, Y. and Hubble, D.W. 2005. Failure conditions of asbestos cement
fissures as air filled cracks. This is close to an equilibrium soil water mains. Canadian Geotechnical Journal 34, 608-621.
microstructure that allows alternate swelling and shrinkage. In Ito, M. and Azam, S. 2013. Engineering properties of a vertisolic
this approach, the change of water volume in the soil mass expansive soil deposit. Engineering Geology. 152(1):10-16
equates to the volume change of the soil aggregates and that of Ito, M. and Azam, S. 2010. Determination of swelling and shrinkage
the cracks. Gens and Alonso (1992) explained the two levels of properties of undisturbed expansive soils. Geotechnical and
Geological Engineering 28, 413-422.
soil structure in their framework as follows: the micro-level is
Ito, M. and Azam, S. 2009. Engineering characteristics of a glacio-
governed by physicochemical interactions between the lacustrine clay deposit in a semi-arid climate. Bulletin of
expansive clay minerals thereby forming aggregates whereas Engineering Geology and the Environment 68, 551-557.
the macro-level includes both the aggregates and the fissures. Kelly, A.J., Sauer, E.K., Barbour, S.L., Christiansen, E.A., and Widger,
Overall, water flow through expansive soils is governed by R.A. 1995. Deformation of the Deer Creek bridge by an active
soil structure as indicated by bimodal SWCCs for both landslide in clay shale. Canadian Geotechnical Journal 32, 701-724.
undisturbed and compacted samples: sample type becomes Leong, E.C., Tripathy, S. and Rahardjo, H. 2003. Total suction
irrelevant when the flow starts to occur through the soil matrix. measurement of unsaturated soils with a device using the chilled-
Likewise, the S-shaped swell-shrink curve confirms the mirror dew-point technique. Geotechnique 53, 173-182.
Marinho, F.M.A. 2005. Nature of soil-water characteristic curve for
structure dominant volume change behavior of such soils for
plastic soils. Journal of Geotechnical and Geoenvironmental
both samples. In the absence of an adequate definition for the Engineering 131, 654-661.
degree of saturation, soil state pertains to the initial void ratio. Mitchell, J.K. and Soga, K. (2005). Fundamentals of Soil Behaviour.
This parameter governs the extent of volume change that, in 3rd ed., John Wiley and Sons, Inc. New York, NY, USA.
turn, must be calculated from the reversible swell-shrink curve Prakash, K., Shidharan, A. Baba, J.A., and Thejas, H.K. 2008.
for expansive soils. Determination of shrinkage limit of fine-grained soils by wax
method. Geotechnical Testing Journal 32, 86-89.
Tripathy, S., Subba Rao, K.S., and Fredlund, D.G. 2002. Water content-
void ratio swell-shrink paths of compacted expansive soils.
Canadian Geotechnical Journal 39, 938-959.

202