Geotech Geol Eng (2010) 28:413422
DOI 10.1007/s10706-010-9301-0
ORIGINAL PAPER
Determination of Swelling and Shrinkage Properties
of Undisturbed Expansive Soils
Maki Ito Shahid Azam
Received: 24 October 2009 / Accepted: 22 January 2010 / Published online: 9 February 2010
Springer Science+Business Media B.V. 2010
Abstract Expansive soils exhibit large volume
changes when their water content changes. Alternate
heave and settlement due to seasonal climatic variations
result in distress and damage in civil infrastructure
systems. This research focuses on the understanding of
swelling and shrinkage phenomenon in the surface
layer of expansive soils. Undisturbed field samples
were used to capture the effect of in situ conditions
(geologically induced fissuring and environmentally
caused saturation) on volume change properties of
Regina clay. Based on laboratory investigations, the
swelling potential and swelling pressure of the native
clay at S = 82% were found to be 1.5% and 3.5 kPa,
respectively. The swell-shrink path during progressive
soil drying followed an S-shaped curve comprising of
an initial low structural shrinkage followed by a sharp
decline during normal shrinkage and then by a low
decrease during residual shrinkage. The soil micro-
structure correlated well with the observed volume
change behaviour as well as with the consistency limits.
The presence of fissures in field samples at various
degrees of saturation confirmed that the investigated
deposit is at an equilibrium condition with respect to the
M. Ito  S. Azam (&)
Environmental Systems Engineering, Faculty
of Engineering and Applied Science, University of
Regina, 3737 Wascana Parkway Regina, Regina, SK S4S
0A2, Canada
e-mail: Shahid.Azam@URegina.CA;
shahid_azam@hotmail.com
swell-shrink phenomenon. The swelling properties at
any initial saturation state were estimated using the free
swelling test and the swell-shrink test data in conjunc-
tion. The swelling potential increased 12 times (from 2
to 24%) and the swelling pressure increased by two
orders of magnitude (from 27 to 2500 kPa) with a
change in the degree of saturation from 80% (at the
plastic limit) to 60% (at the shrinkage limit).
Keywords Expansive soils  Swelling potential 
Swelling pressure  Scanning electron microscopy
1 Introduction
Expansive soils are known to exhibit large volume
changes due to changes in water content. This class of
fine-grained sedimentary soils is frequently encoun-
tered in arid and semi-arid regions of the globe
(Nelson and Miller 1992). The Regina clay deposit in
southern Saskatchewan is a typical example of such
soils that evolved due to geologic weathering of
glacial sediments under restrained leaching in a semi-
arid environment (Christiansen and Saure 2002). The
soil is primarily composed of expansive clay minerals
(such as smectite, hydrous mica, and chlorite) and
exhibits high water adsorbing and water retention
capabilities (Ito and Azam 2009). The surface layer
of the clay, which is in direct contact with the
atmosphere and where most of the civil infrastructure
resides, experiences alternate heave and settlement
123
414
due to seasonal variations in climatic conditions
(Hu and Hubble 2005). This layer is generally
characterized by vertical hairline discontinuities that
govern the swelling and shrinkage behaviour of the
geological strata. Therefore, it is important to eval-
uate the volume change properties of such soils using
undisturbed field samples.
The main objective of this research work was to
understand the swelling and shrinkage behaviour of
expansive soil deposits. Undisturbed field samples
were used to capture the effect of in situ conditions
(geologically induced fissuring and environmentally
caused saturation) on volume change properties of
Regina clay. The free swelling test was conducted to
measure the swelling potential (SP) and swelling
pressure (Ps). To display volume changes over the
entire saturation range, a swell-shrink test was con-
ducted using a modified test procedure. The data from
this test were used in conjunction with those from the
free swelling test to estimate the swelling parameters at
various degrees of saturation. The influence of fissur-
ing was investigated by observing soil microstructure
at critical stages in the volume change process.
2 Literature Review
2.1 Swell-Consolidation Test
The swell-consolidation behaviour of expansive soils
is customarily determined using the ASTM Standard
Test Methods for One-Dimensional Swell or Settle-
ment Potential of Cohesive Soils (D4546-08). This
standard proposes the following three test methods
for various applications:
Method AThe soil specimen is laterally restrained
and axially loaded in a consolidometer with access to
free water. The specimen is inundated and allowed to
vertically swell at a nominal seating pressure (applied
by the weight of the top porous stone and load plate)
until primary swell is complete. The sample is
incrementally loaded thereafter to cancel the deforma-
tion due to swelling. Also known as the free swelling
test method, this two-stage procedure provides two
important swelling parameters, namely: (a) SP or
wetting-induced volume increase and (b) Ps or load-
ing-induced volume decrease.
Method BThe specimen is subjected to a vertical
pressure (such as the in situ vertical overburden
123
Geotech Geol Eng (2010) 28:413422
pressure or structural loading) exceeding the seating
pressure. Thereafter, the specimen is given access to
free water. This may result in (a) swelling; (b) swelling
then compression; (c) compression; or (d) compression
then swelling. The magnitude of volume change is
measured at the applied pressure after movement is
negligible.
Method CAlso known as the constant volume
test, this method requires the specimen height to be
adjusted by consistently increasing the vertical pres-
sures after inundation to obtain Ps. The specimen is
subsequently loaded to determine the preconsolida-
tion pressure and unloaded to use the rebound data
for heave estimation.
The above mentioned laboratory test methods are
associated with several shortcomings. The first class of
shortcomings comprises the ones that are unavoidable
with the conventional apparatus such as the following:
(a) neglecting lateral volume changes and not simu-
lating specimen confinement; (b) assuming a contin-
uous soil matrix without discontinuities; and (c)
inadequate simulation of the actual availability of
water to the soil. The first two limitations were taken
care of by Azam and Wilson (2006) using a thoroughly
instrumented large-scale odometer. However, the use
of such an expensive apparatus and time-consuming
analyses may or may not be justified by every project.
A second set of limitations includes shortcomings
that can be rectified. For example, the measurement of
SP in Method A at a token seating pressure (not the
actual field overburden pressure) and the determina-
tion of Ps in Method C by applying pressure (not a
constant field overburden pressure) while the speci-
men is inundated with water. Method B was designed
to avoid this limitation. Likewise, secondary swelling
can be added to primary swelling for soils exhibiting a
significant amount of long-term volume change.
Chemical composition of the inundating liquid can
be included in the testing thereby measuring swelling
parameters pertinent to certain geoenvironmental
applications. Finally, the test data in all of the methods
can be corrected for apparatus deformability.
Sample disturbance is unavoidable even with
careful handling and transportation. The data obtained
from Method C are generally corrected for sample
disturbance using the method described by Fredlund
and Rahardjo (1993). However, Method B and Method
A do not utilize any corrections to account for
discrepancies arising sample disturbance. This renders
Geotech Geol Eng (2010) 28:413422 415

the results from the latter two methods somewhat A typical swell-shrink test data depicted on a void
erratic, albeit at a lower degree when compared with ratio versus water content plot depicts an S-shaped
Method C. curve (Tripathy et al. 2002). Theoretical lines for
In summary, Method B and Method C can be various degrees of saturation appear as straight lines
considered to give reasonably accurate estimates of with various slope angles emanating from the origin.
the swelling parameters either due to a test procedure This plot can be used to estimate both the degree of
representing field conditions and/or because of cor- saturation and the void ratio of the soil for a given
rections applied to test data for minimizing discrep- water content. Using these index properties as the
ancies. In contrast, Method A still suffers from an initial conditions, the final void ratio can be obtained
inadequate capture of in situ conditions and a lack of from the free swelling test (Method A) as follows: (a)
corrections to apply to test data. This test has been after the completion of swelling at the seating pressure
traditionally used to estimate the swelling potential of to measure the SP; and (b) after the cancellation of
intact field samples at the in situ degree of saturation. deformation due to swelling to measure Ps. This paper
However, soil saturation depends on water availability describes the above-mentioned procedure using
that, in turn, is primarily governed by seasonal Regina clay as a typical example of expansive soils.
climatic conditions and depth in the profile. Conse-
quently, the measured data correspond to a specific
time and space only. Testing of air-dried or oven dried
3 Materials and Methods
samples of expansive soils is quite difficult and does
not represent the field setting because of shrinkage
3.1 Sample Collection
cracking and the associated fabric changes. Likewise,
testing of compacted samples at various degrees of
The soil represented an extensive deposit of the
saturation is not representative of in situ conditions
glacio-lacustrine clay in the south side of the city of
because of a complete change in soil microstructure.
Regina. At the sampling site, the surface layer of the
Further, the fabric of compacted samples of the same
indigenous clay appeared desiccated and exhibited
soil at dry of optimum is known to be quite different
soil discontinuities in the order of 12 mm. A visual
from that at wet of optimum. Therefore, supplemen-
survey of the exposed cutting showed large variation
tary tests need to be conducted to correlate the
in the dip angles of the fissures with respect to the
swelling properties of intact field specimens of
horizontal. The samples were retrieved in September
expansive clays with the degree of saturation.
2007 from a depth of 0.51.0 m using the ASTM
Standard Practice for Soil Investigation and Sampling
2.2 Swell-Shrink Test
by Auger Borings (D1452-2009) for disturbed sam-
ples and the ASTM Standard Practice for Thin-
Shrinkage in a soil specimen occurs due to an increase
Walled Tube Sampling of Soils for Geotechnical
in capillary stress causing a reduction in inter-particle
Purposes (D1587-08) for undisturbed samples. To
distances thereby causing a decrease in the overall
preserve the field water content (w), the latter samples
volume of soil. The determination of swell-shrink
were wrapped with plastic sheets and painted with
behaviour of expansive soils is based on shrinkage
molten wax as per the ASTM Standard Practice for
limit that is measured using the ASTM Standard Test
Preserving and Transporting Rock Core Samples
Method for Shrinkage Factors of Soils by the Wax
(D5079-08). All of the samples were transported to
Method (D4943-08). Slurried, compacted, or undis-
the Geotechnical Testing Laboratory at the Univer-
turbed samples are desiccated from an initially
sity of Regina and stored at 24C in a humidity
saturated state without applying any surcharge
chamber (RH & 90%).
pressure. The void ratio at different water contents
is determined using the above method whereas the
ASTM Standard Test Methods for Laboratory Deter- 3.2 Geotechnical Index Properties
mination of Water (Moisture) Content of Soil and
Rock by Mass (D2216-05) is used to measure the The geotechnical index properties were determined
water content. for preliminary soil assessment according to standard

123
416
ASTM test methods as follows: (a) field water content
(w) by the Standard Test Methods for Laboratory
Determination of Water (Moisture) Content of Soil
and Rock by Mass (D2216-05); (b) field dry unit
weight (cd) by the Standard Test Method for Density
of Soil in Place by the Drive-Cylinder Method
(D2937-04); (c) specific gravity (Gs) by the Standard
Test Methods for Specific Gravity of Soil Solids by
Water Pycnometer (D854-06); (d) liquid limit (wl),
plastic limit (wp) and plasticity index (Ip) by the
Standard Test Methods for Liquid Limit, Plastic
Limit, and Plasticity Index of Soils (D4318-05); (e)
shrinkage limit (ws) by the Standard Test Method for
Shrinkage Factors of Soils by the Wax Method
(D4943-08); and (f) clay size fraction (material finer
than 0.002 mm) by the Standard Test Method for
Particle-Size Analysis of Soils (D422-63(2007)).
3.3 Free Swelling Test
The swelling and consolidation properties of the soil
were determined by conducting the free swelling test
according to the ASTM Standard Test Methods for
One-Dimensional Swell or Settlement Potential of
Cohesive Soils (D4546-08). A Dead Weight Consol-
idation Load Frame (Model S-449) that contained a
64 mm internal diameter fixed ring consolidometer
(Model S-455) was used; both devices were manu-
factured by Durham Geo Slope Indicator Inc. To
mimic precipitation, de-ionized water was used for
specimen inundation and the sample was allowed to
undergo swelling at a seating pressure of 1.5 kPa.
After the completion of the swelling stage, the soil
specimen was incrementally loaded to determine the
consolidation properties of the expansive clay. This
stage of the test was conducted in accordance with
the ASTM Standard Test methods for One-Dimen-
sional Consolidation Properties of Soils Using Incre-
mental Loading (D2435-04). All of the volume
changes were measured at regular intervals of time
using a liner displacement transducer (Model E-311)
manufactured by Durham Geo Slope Indicator Inc.
The transducer was connected to a data logger and
then to a portable computer for digital recording.
3.4 Swell-Shrink Test
The swell-shrink test was conducted in accordance
with the ASTM Standard Test Method for Shrinkage
123
Geotech Geol Eng (2010) 28:413422
Factors of Soils by the Wax Method (D4943-08). To
obtain the void ratio, the volume of soil specimens
was determined using the water displacement method.
Each specimen was coated with molten microcrystal-
line wax (Gs = 0.91) and allowed to cool down at
room temperature. After wax solidification, the sam-
ple was submerged in a 250 ml graduated cylinder
that was filled with distilled water. The water height in
the cylinder was carefully recorded using a Vernier
calliper before and after sample submersion in the
cylinder. A graduated syringe was used to remove the
increased amount of water displaced by the sample
thereby bringing the water height back to the initial
reading. The displaced water mass was determined by
weighing the graduated syringe before and after water
filling and recording the difference. This quantity was
readily converted to water volume representing the
volume of wax-coated soil. The volume of soil was
obtained from the difference of volume of the wax-
coated sample and the volume of wax (mass/0.91). A
7.4% correction was applied to account for the
underestimation due to air entrapment at soil-wax
interface in this method, as suggested by Prakash et al.
(2008). The mass of the sample was also determined
to estimate the bulk unit weight of the soil. Using
basic phase relationships, the void ratio was deter-
mined from a knowledge of the bulk unit weight of the
soil.
3.5 Scanning Electron Microscopy
To correlate soil microstructure with volume change,
scanning electron microscopy (SEM) was performed
using the microscope (JSM-6301FXV). The cryogen-
ically prepared samples ensured an intact fabric and
minimal grain re-adjustment (Lilly and Sargent
1990). The specimen (5 mm 9 5 mm 9 5 mm) was
cut from the soil and immediately immersed in
nitrogen slush at -208C for instantaneous freezing.
The SEM chamber was flushed with argon to remove
water vapour and vacuum cleaned with a rotary
pump. Samples were placed in the chamber, subli-
mated at -40C, and examined at 2.5 kV till
complete ice removal. An ablation depth of approx-
imately 10 lm was obtained in about 45 min. Sam-
ples were gold sputtered in the cryogenic system at
-155C to minimize ionization and transferred back
to the chamber for visual examination. The vacuum
pressure and scanning and recording speeds were
Geotech Geol Eng (2010) 28:413422
8 9 10-8 kPa, 450 frame/sec and 40 s/frame, respec-
tively. The 5 kV incident beam was kept at a vertical
distance of 30 mm during photomicrograph capture.
Soil morphology was observed through photomicro-
graphs at representative locations.
4 Results and Discussion
4.1 Geotechnical Index Properties
Table 1 gives the geotechnical index properties of
Regina clay. The field water content (weight based)
and the dry unit weight were determined to be 31.4%
and 13.4 kN/m3, respectively. From the measured
specific gravity (Gs = 2.75) and degree of saturation
(S = 82%), the field void ratio was calculated to be
1.05. The surface soil is highly affected by seasonal
variations in water availability (snow melt in spring
and rainfall in summer) and water deficiency (low
rainfall and freezing in fall and winter). It undergoes
periodic saturation (S = 100%) and desaturation
(S = 60%) particularly in the top 1.0 m deep layer.
The high liquid limit (wl = 83%) and plastic limit
(wp = 30%) along with a low shrinkage limit (ws =
15%) suggested the high water absorbing and retaining
characteristics of the investigated clay. The clay size
fraction (material finer than 0.002 mm) was found to
be 66%. Using the plasticity index (Ip = 53%), the
Table 1 Geotechnical index properties of Regina clay
Property ASTM test
method
Field water content, w (%) D2216-05
Field dry unit weight, cd (kN/m3) D2937-04
Specific gravity, Gs D854-06
Field void ratio, e*
Degree of saturation, S (%) 82
Liquid limit, wl (%) D4318-05
Plastic limit, wp (%) D4318-05
Plasticity index, Ip (%) D4318-05
Shrinkage limit, ws (%) D4943-08
-0.002 mm, C (%) D422-63(2007)
Activity, A = Ip/C
* e = (Gs/cd) - 1
S = w Gs/e
417
activity of the soil was found to be 0.8. These data
indicate that Regina clay is a typical expansive soil.
Recent investigations have revealed that the native soil
is primary composed of expansive clay minerals such
as smectite, hydrous mica, and chlorite (Ito and Azam
2009).
4.2 Swell-Consolidation Behaviour
Figure 1 gives the change in void ratio in relation to the
applied pressure during the free swelling test. The test
data were corrected for apparatus deformability. The
sample increased from an in situ void ratio of 1.05
(point A) to 1.08 (point B) at the seating pressure.
Thereafter, the sample was incrementally loaded up to
the maximum load of 3840 kPa (point C) to obtain the
compression curve. The pre-consolidation pressure
was graphically determined using the Cassagrandes
procedure, that is, by drawing horizontal, tangential,
and bisector lines through the point of maximum
curvature on the compression curve. The intersection
of the compression line with the bisector line repre-
sented the pre-consolidation pressure. The pre-consol-
idation pressure for the investigated clay was found to
be 140 kPa. This represents an over-consolidated clay
as the field overburden pressure was calculated to be
about 10 kPa. According to Fredlund et al. (1980),
clays in the study area have moderate compressibility
owing to a long evaporation history and desiccation
induced pre-consolidation. The compression index
(Cc) was determined to be 0.21 that fell within the
range typical of inorganic silty clays, which is between
Value
31.4
13.4
2.75
1.05
83
30
53
15
66
0.8
Fig. 1 Void ratio versus applied pressure in free swelling test
123
418
0.15 and 0.3 (Mitchell and Soga 2005). The swelling
potential was computed using the following equation:
SP De=1 e 1
The SP was found to be quite low (1.5%). This
value was confirmed by the actual volume change
measurements over time data that indicated the
completion of swelling after 5 days.
The low SP of the investigated sample is attributed
to the in situ conditions (fissures and degree of
saturation) and unavoidable disturbances during
sample collection, transportation and preparation
(Nelson and Miller 1992). At the seating pressure
of 1.5 kPa, the volume increase was governed by
changes in soil suction. The high initial degree of
saturation (82%) led to a low water intake by the
investigated sample. The use of de-ionized water that
may have diluted the pore water (pH = 7.5) did not
result in any significant overestimation of volume
increase. Likewise, vertical fissuring (hair-line dis-
continuities) in the top 1.0 m depth of the soil
resulted in consuming part of the swelling movement
in the lateral direction. Consequently, the SP mea-
sured in the vertical direction recorded a lower value.
Figure 2 gives the soil microstructure under in situ
conditions (point A in Fig. 1). The photomicrograph
shows randomly oriented flaky smectite clay particles
forming multi-particle aggregates. The random parti-
cle orientation and a flocculated fabric are attributed
to the geological deposition of the clay in a proglacial
lake environment with high alkalinity and the pres-
ence of large amounts of multivalent ions (Ito and
Azam 2009). The aggregates are formed because of
Fig. 2 Photomicrograph of in situ sample
123
Geotech Geol Eng (2010) 28:413422
water adsorption onto clay particle surfaces. A small
number of fissures with widths in the sub-micron
range indicate that the discontinuities originate in the
microstructure and are subsequently enlarged due to
sustained desiccation in the field. The micro-scale
discontinuities correlate well with the soil plastic limit
(water content at which the soil begins to crumble).
Figure 3 illustrates the clay fabric under fully
saturated conditions (point B in Fig. 1). As expected,
the photomicrograph is quite similar to the previous
one (S = 82% and e = 1.05) as depicted by the
presence of clay aggregates and the comparable
number and size of micropores. This is because an
18% increase in saturation is associated with a change
in void ratio of only 0.03 thereby confirming the low
SP (1.5%) of the investigated clay. Further, the
random particle orientation suggests that swelling
deformations be experienced in all directions as
opposed to the vertical direction used for measure-
ment in the free swelling test.
Figure 4 shows the soil morphology during consol-
idation under full saturation (point C in Fig. 1). The
photomicrograph reveals a compressed soil fabric with
minimal pore spaces under an applied pressure of
3840 kPa that decreased the void ratio to 0.65 (almost
40% reduction with respect to the previous sample). The
less frequent micropores in this sample indicate that
most of the pore water has dissipated from the clay
matrix. Conversely, the presence of clay aggregates
means that the adsorbed water is still present in the
specimen and ensures complete saturation. This electro-
chemically bonded water is difficult to remove by the
application of external loads (Mitchell and Soga 2005).
Fig. 3 Photomicrograph of saturated sample
Geotech Geol Eng (2010) 28:413422
4.3 Swell-Shrink Behaviour
Figure 5 shows the results of the swell-shrink test in
the form of void ratio versus water content. Theoret-
ical lines representing various degrees of saturation
values were obtained from basic phase relationship
and using Gs = 2.75. Several field samples were first
wetted to achieve complete saturation (point B) from
an initially unsaturated state (point A). Thereafter,
applying different suction values in a pressure plate
extractor desaturated the specimens: the sample at
point D was air-dried (RH = 50 5%). The void
ratio and water content of each sample was deter-
mined as described earlier in this paper. The initial
data points in this study corroborate those reported by
Shuai (1996) who conducted free swelling tests on a
compacted sample of Regina clay with similar initial
conditions (w = 26% and e = 0.97).
Fig. 4 Photomicrograph of consolidated sample
Fig. 5 Swell-shrink curve
419
The progressive drying of the undisturbed expan-
sive soil was found to follow an S-shaped curve.
The curve comprises of an initial low structural
shrinkage followed by a sharp decline during nor-
mal shrinkage and then by a low decrease during
residual shrinkage (Haines 1923). During structural
shrinkage, some of the larger and relatively stable
voids are emptied such that the decrease in soil volume
is less than the volume of water lost. During normal
shrinkage, volume decrease in soil is equal to the
volume of water lost thereby leading to a 45 straight
line parallel to the 100% saturation line. During
residual shrinkage, air enters the pores close to the
shrinkage limit and pulls the particles together due to
suction. This leads to a further decrease in soil volume
albeit lower than the volume of water lost.
Figure 5 further indicates that the transition
between the various shrinkage stages correlate well
with consistency limits as follows: (a) structural
shrinkage from the S = 100% line to the plastic limit;
(b) normal shrinkage from the plastic limit to the
shrinkage limit; and (c) residual shrinkage from the
shrinkage limit to complete desiccation. As described
later in this paper, the middle linear portion of the
swell-shrink curve representing only 20% change in
saturation (from S = 80% to S = 60%) is associated
with bulk of the volume changes in the expansive
Regina clay.
Figure 6 gives the photomicrograph of the air-
dried sample (point D in Fig. 5). This figure explains
the observed volume change behaviour of the inves-
tigated expansive soil when studied in conjunction
with the previous photomicrographs. When compared
with Fig. 4 (point C in Fig. 1), the clay fabric still
Fig. 6 Photomicrograph of air-dried sample
123
420 Geotech Geol Eng (2010) 28:413422

appears to be quite packed (e = 0.56) despite a low

degree of saturation of 21%. This is attributed to the
pulling together of clay particles due to suction
during residual shrinkage that starts close to the
shrinkage limit. When compared with Fig. 2 (point A
in Fig. 5) or Fig. 3 (point B in Fig. 5), the clay fabric
clearly shows the presence of numerous and rela-
tively larger pore spaces and a considerable decrease
in the aggregate size. The existence of fissures in all
of the field samples (Figs. 2, 3, and 6) is confirmed
for the investigated saturation range, that is, from
100% through 20%. The decrease in aggregate size is
attributed to the partial removal of adsorbed water
from the clay. Fig. 7 Shrinkage stages and corresponding linear equations
The swell-shrink path is reversible once the sample
reaches an equilibrium condition. Tripathy et al. (2002)
reported that this stage is usually attained after about
four cycles in compacted soils. For natural Regina clay
samples having undergone numerous swell-shrink
cycles since deposition, an equilibrium condition is
considered to have reached. This was confirmed by the
closely matching photomicrographs highlighting the
presence of fissures in field samples at various degrees
of saturation. Further, the perpendicular distance
between the normal shrinkage line and the saturation
line is known to decrease with increasing surcharge
pressure. Given the low seating pressure (1.5 kPa) in
the free swelling test and the no applied load in the
swell-shrink test, results from the two tests can be used
Fig. 8 Free swelling curve with different initial degrees of
in conjunction. Specifically, the swelling potential and saturation
swelling pressure can be determined for any initial
degree of saturation using a graphical procedure. condition and the swelling pressure was graphically
Figure 7 simplifies the swell-shrink path by draw- determined at the intersection with the compression
ing tangents to the three limbs of the S-shaped curve. curve.
These tangents corresponded to the three shrinkage Table 2 summarizes the variation in the estimated
stages and closely matched the swell-shrink curve. swelling parameters with changes in the degree of
The resulting straight-line equations were used along saturation. The data indicate that the variation in both
with the next figure (Fig. 8) to estimate the swelling SP and Ps is marginal up to the plastic limit
parameters for a given initial condition. The initial (S = 80%). The swelling potential increased 12 times
straight-line portion of the plot in Fig. 8 was obtained
from knowledge of the swell-shrink path. The void
Table 2 Variation in swelling parameters with saturation
ratio and the degree of saturation values at various
stages in soil behaviour were obtained from Fig. 7. Condition S (%) SP (%) Ps (kPa)
These stages included wet, field, plastic limit, shrink- Wet 100 0 0
age limit, and desiccated conditions. Using the initial Field 82 1.5 3.5
void ratio (from Fig. 7) and the final void ratio (from
Plastic limit 80 2 27
Fig. 8), the swelling potential for each of the critical
Shrinkage limit 60 24 2500
condition was estimated by Eq. (1). Next, horizontal
Desiccated 0 35 12000
lines were drawn for each of the above-mentioned

123
Geotech Geol Eng (2010) 28:413422
(from 2 to 24%) and the swelling pressure increased
by two orders of magnitude (from 27 kPa to
2500 kPa) when the soil was dried from the plastic
limit (S = 80%) to the shrinkage limit (S = 60%).
This wide range in swelling parameters between the
two consistency limits corresponds well with seasonal
variations in the field water content profile and
explains the engineering problems in Regina clay.
The maximum SP and Ps values are obtained under
completely dry conditions.
It is important to note that the field samples for this
research were retrieved from a depth between 0.5 and
1.0 m. This was done to highlight the effect of fissuring
on volume changes in the native expansive soil and for
economic reasons. A deeper soil sample would be more
representative in actual foundations and would depend
on the type of structure to be built. Locally, most
construction activities take place in the summer and
fall seasons when the soil experiences large variations
in saturation (from S = 100% through S = 60%) in the
top 1.0 m layer. This range in the degree of saturation
of the soil narrows down with depth and is negligible
beyond 3.0 m depth. Still, construction sites have to be
protected from possible rains and thoroughly dewa-
tered in areas of shallow ground water table.
5 Summary and Conclusions
A theoretical framework does not exist to capture the
effect of fissures (macroscopic voids) and pores
(microscopic voidsas part of a porous material)
separately. This research employed the ASTM test
methods to determine the effect of saturation and
desaturation on the volume change properties of
undisturbed Regina clay. Field samples were used to
determine the swell-consolidation and swell-shrink
behaviour of the highly expansive native soil. Com-
monly used in engineering practice, the test methods
provided reasonable approximations of the geotech-
nical behaviour of the investigated soil, albeit not
covering the whole picture. Therefore, results from
these tests were studied in conjunction with soil
microstructure to understand the effect of fissuring on
volume changes. The main conclusions of this
research are summarised below:
1. The swelling potential and swelling pressure of an
undisturbed sample of the expansive Regina clay
421
measured 1.5% and 3.5 kPa, respectively. Such
low values are attributed to a low soil affinity for
free water at the high initial degree of saturation
(82%), lateral consumption of the swelling move-
ment by vertical fissures, and unavoidable distur-
bances occurring during collection, transportation
and preparation of the field sample.
2. The swell-shrink path during progressive drying
of undisturbed samples of Regina clay followed
an S-shaped curve. The curve comprised of an
initial low structural shrinkage (S = 100% up to
the plastic limit saturation of 80%) followed by a
sharp decline during normal shrinkage (S = 80%
up to the shrinkage limit saturation of 60%) and
then by a low decrease during residual shrinkage
(S = 60% up to complete desiccation).
3. The soil microstructure at critical stages in the
two oedometer tests correlated well with the
observed volume change behaviour (swelling,
consolidation, and shrinkage) as well as with the
consistency limits (plastic limit and shrinkage
limit). The presence of fissures in field samples at
various degrees of saturation confirmed that the
investigated deposit is at an equilibrium condition
with respect to the swell-shrink phenomenon.
4. The swelling properties at a given initial degree
of saturation can be estimated using the results of
the free swelling test and the swell-shrink test in
conjunction. The swelling potential increased
12 times (from 2 to 24%) and the swelling
pressure increased by two orders of magnitude
(from 27 to 2500 kPa) with a change in the
degree of saturation from 80% (at the plastic
limit) to 60% (at the shrinkage limit).
Acknowledgments The authors acknowledge the financial
support provided by the Natural Science and Engineering
Research Council of Canada and the Mathematics of
Information Technology and Complex Systems Inc., Canada.
Sincere thanks to the University of Alberta and the University
of Regina for providing laboratory space and computing
facilities.
References
Azam S, Wilson GW (2006) Volume change behavior of a
fissured expansive clay containing anhydrous calcium
sulfate. In: Proceedings, 4th international conference on
unsaturated soils, Carefree, Arizona, USA. 1: 906915
123
422
Christiansen EA, Saure EK (2002) Stratigraphy and structure
of pleistocene collapse in the regina low, Saskatchewan,
Canada. Can J Earth Sci 39:14111423
Fredlund DG, Rahardjo H (1993) Soil mechanics for unsatu-
rated soils, 1st edn. Willy, New York
Fredlund DG, Krahn J, Hasan JU (1980) Variability of an
expansive clay deposit. In: Proceedings, 4th international
conference on expansive soils, Denver, Colorado, USA. 1:
322338
Haines WB (1923) The volume changes associated with varia-
tions of water content in soil. J Agric Sci 13:296310
Hu Y, Hubble DW (2005) Failure conditions of asbestos
cement water mains. Can Geotech J 34:608621
Ito M, Azam S (2009) Engineering characteristics of a glacio-
lacustrine clay deposit in a semi-arid climate. Bull Eng
Geol Environ 68:551557
Lilly MR, Sargent JA (1990) Cryogenic scanning electron
microscopy research application in arctic and polar
123
Geotech Geol Eng (2010) 28:413422
regions. In: Proceedings, 9th international conference of
offshore mechanics and arctic engineering, Houston,
Texas, USA. 4: 5160
Mitchell JK, Soga K (2005) Fundamentals of soil behavior, 3rd
edn. Willy, New York
Nelson JD, Miller DJ (1992) Expansive soils. Willy, New York
Prakash K, Shidharan A, Baba JA, Thejas HK (2008) Deter-
mination of shrinkage limit of fine-grained soils by wax
method. Geotech Test J 32:8689
Shuai F (1996) Simulation of swelling pressure measurements
on expansive soils. Ph.D. thesis, University of Saskatch-
ewan, Saskatoon, SK, Canada
Tripathy S, Subba Rao KS, Fredlund DG (2002) Water con-
tentvoid ratio swell-shrink paths of compacted expan-
sive soils. Can Geotech J 39:938959