Bull Eng Geol Environ (2013) 72:25–35

DOI 10.1007/s10064-012-0457-0

ORIGINAL PAPER

Study on swelling properties of an expansive soil deposit

in Saskatchewan, Canada
Shahid Azam • Imran Shah • Mavinakere E. Raghunandan • Maki Ito

Received: 11 January 2012 / Accepted: 6 December 2012 / Published online: 6 January 2013
Ó Springer-Verlag Berlin Heidelberg 2013

Abstract The main objective of this paper was to study (entre 8 m et 9 m de profondeur). Les fortes limites de
the swelling properties of Regina clay. The deposit com- liquidité (70 ± 15 %) et limites de plasticité (33 ± 4 %)
prises a topsoil (surface to 0.3 m depth), an expansive clay indiquaient une forte capacité de rétention d’eau et une
(0.3–8 m), and a bottom till (8–9 m). High liquid limit forte capacité d’adsorption du sol argileux. Indépendam-
(70 ± 15 %) and plastic limit (33 ± 4 %) indicated high ment du type de couverture (végétation ou route fissurée),
water retention and adsorption capacity for the clay. Irre- la teneur en eau in situ, en été, correspondait à la limite de
spective of the cover type (vegetation and cracked road), plasticité. Le sol argileux présentait un potentiel de gon-
the field water content in summer closely matched the flement moyen à élevé qui était bien prévu par la relation
plastic limit. The clay was characterized by medium-to- SP = 0,16 (Ip)1,188. Le sol argileux contenait 51 % de
high swelling that was best predicted by SP = 0.16 minéraux argileux dont des smectites (32 %), de l’illite
(Ip)1.188. The soil had 51 % clay minerals including (7 %), de la kaolinite (5 %) et de la chlorite (3 %). Avec
smectite (32 %), illite (7 %), kaolinite (5 %), and chlorite une CEC de 40 cmol(?)/kg, les ions Ca2? et Mg2? se sont
(3 %). With a CEC of 40 cmol(?)/kg, Ca2? and Mg2? révélés être les principaux cations échangeables. La surface
were found to be the main exchangeable cations: the spe- spécifique était de 50 m2/g. Le potentiel de gonflement et
cific surface area was 50 m2/g. The SP and Ps for a 1.2 m la pression de gonflement pour un échantillon prélevé à
deep sample measured 12 % and 260 kPa, respectively, 1,2 m de profondeur donnaient respectivement 12 % et
and the estimated surface heave of 180 mm gradually 260 kPa. Le soulèvement de surface était estimé à 180 mm
diminished to 3.6 m depth. These values matched well with pour diminuer progressivement à 3,6 m de profondeur. Ces
consistency-based correlations. valeurs correspondaient bien aux relations établies.

Keywords Expansive soils  Geotechnical properties  Mots clés Sols gonflants  Propriétés géotechniques 
Swelling potential  Heave prediction  Clay mineralogy Potentiel de gonflement  Prévision de soulèvement 
Minéraux argileux
Résumé L’objectif principal de cet article était l’étude
des propriétés de gonflement de l’argile de Regina. La
formation géologique comprend un niveau supérieur (de la Introduction
surface à 0,3 m de profondeur), l’argile gonflante (de 0,3 m
à 8 m de profondeur) et un niveau inférieur morainique Expansive soils are frequently encountered in the Canadian
prairies in areas where favorable environments exist. A
glacio-lacustrine geology and a semi-arid climate resulted
in the development of an expansive soil deposit in Regina,
S. Azam (&)  I. Shah  M. E. Raghunandan  M. Ito Saskatchewan (Ito and Azam 2009). The presence of spe-
Environmental Systems Engineering, Faculty of Engineering and
cific clay minerals (smectite, illite and chlorite) along with
Applied Science, University of Regina, 3737 Wascana Parkway,
Regina, SK S4S 0A2, Canada seasonal variations in atmospheric conditions (precipita-
e-mail: [Link]@[Link] tion, temperature, relative humidity, wind speed, and solar

123
26 S. Azam et al.

radiation) leads to periodic swelling and shrinkage. These Figure 1 presents the physiography of the Lake Regina
soil movements are at a maximum in the surface layer of Basin and isopach map of the expansive soil deposit
the soil, which is in direct contact with the atmosphere, and (Acton et al. 1960; Pollock et al. 1987). The basin is
gradually reduce with depth due to the overburden soil bounded by the gently sloping Moose Mountain Upland in
stress. the northeast and by the flatter Missouri Coteau Upland
Cyclic volume changes in the local soil cause severe escarpment in the southwest. Both of these pre-glacial
deformations and eventual failure in lightly loaded engi- uplands are overlain by glacial drift and are presently up
neered facilities such as buried pipelines (Hu and Hubble to 40 km wide (Simpson 2004). The escarpment helped in
2005), highway embankments (Kelly et al. 1995) and removing most of the boulders from the retreating glacier
residential foundations (Hamilton 1965). The aging civil front and the latter primarily deposited finer materials in
infrastructure in the city is particularly prone to failure the basin containing the entrapped meltwater. The pro-
and entails a high maintenance cost that is gradually glacial lake allowed some outflow in the east flowing
increasing with time. For example, the rupture frequency Qu’Appelle River and its tributaries. However, the bulk of
in the 850 km long water distribution system has doubled the drainage occurred at the same time (about 1,000 years
over the last two decades to 160 breaks/year; the associ- later) and formed the Souris River that was subsequently
ated annual repair cost is estimated to be $2 million covered by sediments (Christiansen and Saure 2002).
(Azam et al. 2009). Restrained leaching and marine conditions (Na?, K?, Cl-,
The main objective of this paper was to study the Ca2?, Mg2?, and SO42-) helped in conversion of the
swelling properties of the expansive soil deposit in Regina. deposited clays of volcanic origin to the present day
The geological evolution of the clay was studied to eval- expansive clay minerals (Ito and Azam 2009). The deposit
uate the composition and spatial variability of indigenous thickness varies from zero along the lake boundary to a
materials. Site investigations were carried out to under- 12 m depth around the city (with an elevation of
stand field conditions (cover type and seasonality) and 577 masl).
collect soil samples. Geotechnical index properties were
determined to characterize the soil profile, and soil min-
eralogy was determined to correlate the composition with Research methodology
engineering behavior. The swelling potential and the
swelling pressure were determined using undisturbed Figure 2 gives a schematic of the study area which is
samples for heave prediction in the deposit. located in a residential quarter of Regina. The site was
investigated as part of an instrumentation program for
monitoring seasonal deformations (induced by soil–
Geological development weather interactions) in asbestos cement water mains. The
study site composed of the following:
The soil deposits of southern Saskatchewan originate from
(i) a park area toward the north that is generally covered
the rising of the Rocky Mountains during the Cretaceous to
with thick grass (up to 100 mm);
the Tertiary. Material deposition and erosion occurred due
(ii) a 15 m wide paved road in the middle (with 150 mm
to tectonic movements of the Williston Basin and eustatic
thick asphalt pavement);
and glacio-eustatic sea level fluctuations. During this time,
(iii) a row of single family houses toward the south.
winds carried rhyolitic ash northeastward from eruptions in
Western Montana. This acidic volcanic glass deposited at In May 2009, a 9 m deep borehole (BH-a) was drilled
the Pierre Sea floor and converted to clays over time. through the road pavement (0.6 m inside of the road edge
During the Quaternary, the area experienced eight glacial and 0.8 m away from the water main). Two previous
advances and retreats; the last one, the Laurentide, reached boreholes also existed at the site: BH-b, under the asphalt
its maximum extent about 18,000 years B.P. (Christiansen paved road (2.8 m inside of the road edge and 1.4 m south
1979). This 1,000 m thick ice sheet gradually retreated in of the water main) which was bored in September 2006
the northeastward direction and had completely waned by (Hu et al. 2008); BH-c, in the park area (1 m outside of the
approximately 8,000 years B.P. The bedrock surfaces were road edge and 2.4 m north of the water main) which was
eroded during glacial advances whereas deposition took bored in September 2006 (Clifton Associates Ltd.,
place during ice retreats; the thickness of the glacial drift unpublished). Both BH-b and BH-c were located 15 m east
ranges from 50 to 300 m (Mollard et al. 1998). A portion of of the current borehole BH-a.
the drift was deposited as till (unsorted materials from clay Field samples to a depth of 9 m were obtained using a
size to boulder size) that covers most of the southern part of Central Mine Equipment (CME-75) together with an auger
the province. and extensions mounted on a single rear axle truck. The

123
Swelling properties of expansive soil, Canada 27

Fig. 1 Physiographic division 51 00'

and isopach map of expansive Last
soil deposit in Regina Mountain 0 25 km
Qu Lake
'A e River
pp ell
ell pp
e 'A 50 45'
Eyebrow Hills Ri
ve Qu M
Upland r oo
se

H
ig
M

he
1.5 ou

st
nt

el
3.0 4.5

ev
ai 50 30'

a
n

tio
4.5
U

n
pl

=
an

83
La d

1
Regina

m
ke 12.0

as
l
R 10.5 50 15'
eg 9.0
M in 7.5
iss a 6.0
Ba

H
ou sin

ig
4.5
ri

h es
3.0
C M

tE
ot oo 50 00'
se

le
ea

va
Ja
Old Wives u w

t io
1.5 Ri
U

n
Lake Plain pl ve
r

=
an Souris River

67
d

5
Plain

m
as
49 45'

l
106 30' 106 00' 105 30' 105 00' 104 30' 104 00' 103 30'

Fig. 2 Schematic of the N 15 m

investigated site
0.15 m

Asbestos Cement
BH-c BH-b
Water Main
(September 2006, (September 2006,
Clifton Associates Ltd. (unpublished)) Hu et al., 2008)

House

1m 1.4 m 1.4 m

Park Area

15 m House
Cross Place Road

0.6 m 0.8 m

BH-a
(May 2009, House
This study)

Grass Vegetation Asphalt Pavement Note: Not to scale

sample profile was established using alternate bore cuttings Field observations during sample acquisition indicated
of about 300 mm depth and undisturbed core samples of the absence of ground water to the investigated depth.
around 600 mm depth. A total of 20 disturbed samples Consequently, no visible signs of ice lenses were found,
were acquired according to the ASTM Standard Practice although the samples were at freezing temperatures. Most
for Soil Investigation and Sampling by Auger Borings of the soil samples could be rolled into about 3 mm threads
(D1452-09). As discussed later, an average sample depth in the palm of a hand, suggesting that the field water
was used for data presentation. content (wf) would be close to the plastic limit (wp).

123
28 S. Azam et al.

Further, the soil samples near the surface were observed to To correlate mineralogy with cation exchange, 100 mg
be relatively coarser grained compared to the deeper clay soil and 1 M NH4OAc were mixed in a tube and agitated
samples. Finally, structural cracks were found on the road for 5 min in a reciprocal shaker. After 24 h, the tube was
surface which could promote water infiltration under the re-agitated for 15 min and the contents filtered through
pavement. Whatman No. 42. Cations (Na?, K?, Ca2?, and Mg2?)
To preserve the field water content, the soil samples were measured by the ammonium acetate method (Hen-
were sealed in multiple, properly marked, plastic bags and dershot and Lalande 2008). The filtrate, washed with 1 M
transported to the University of Regina where they were NH4OAc and 250 mL leachate, was analyzed in a volu-
stored at a constant temperature of 23 °C. The field water metric flask. The ion types were quantified using suitable
content was determined according to the ASTM Standard dilution factors and blanks’ concentrations. The sum of
Test Methods for Laboratory Determination of Water ions was reported as cation exchange capacity (CEC).
(Moisture) Content of Soil and Rock by Mass (D2216-05) To evaluate the physicochemical phenomena at the
and the dry unit weight (cd) using the ASTM Standard Test particle surface, Brunauer Emmett Teller (BET) analysis
Method for Shrinkage Factors of Soils by the Wax Method was used to determine the specific surface area of the clay.
(D4943-08), as modified by Ito and Azam (2010). The Accelerated Surface Area and Porosimetry (ASAP
Geotechnical index properties were used for soil clas- 2010) System was used for the analysis. The tests were
sification and evaluation of swelling properties. The spe- conducted by the liquid N2 physio-sorption method, that is,
cific gravity (Gs) was determined by the ASTM Standard the physical adsorption of gas molecules due to polariza-
Test Method for Specific Gravity of Soil Solids by Water tion of multi-molecular adsorption (Brunauer et al. 1932).
Pycnometer (D854-10), the grain size distribution (GSD) in To simulate the oven drying commonly used in geotech-
accordance with the ASTM Standard Test Method for nical engineering practice, the sample was degassed at a
Particle-Size Analysis of Soils [D422-63 (2007)] and the temperature of 105 °C for 24 h.
consistency limits according to the ASTM Standard Test The swelling properties were determined according to
Methods for Liquid Limit, Plastic Limit, and Plasticity the ASTM Standard Test Methods for One-Dimensional
Index of Soils (D4318-10). Finally, the soil samples were Swell or Collapse of Cohesive Soils (D4546-08). A Dead
classified according to Unified Soil Classification System Weight Consolidation Load Frame (Model S-449) that
using the ASTM Standard Practice for Classification of contained a 64 mm internal diameter fixed ring consoli-
Soils for Engineering Purposes (D2487-10). dometer (Model S-455) was used. Identical field samples
X-ray diffraction (XRD) analyses were conducted to from an average 1.2 m depth were carefully trimmed to fit
determine mineral composition of the soil using the Bruker the consolidation ring. The samples were placed in the
D8 Advance diffractometer which has an incident beam consolidometer between two porous stones at a seating
parabolic mirror (Co K-alpha), a 0.2 mm exit slit to min- pressure of 1.2 kPa and inundated with de-ionized water.
imize intensity aberrations, a 25 mm sample diameter, and Three tests were conducted: (1) free swelling test to
a VANTEC-1TM linear detector. About 30 mg of each determine swelling potential as the change in vertical
sample was powdered using methanol as a grinding aid. height of the sample with respect to the original height; (2)
Randomly oriented bulk samples and oriented clay size constant volume test to determine swelling pressure as the
samples were prepared for overall and clay mineral iden- pressure required to cancel the swelling deformations; (3)
tification (at RH = 54 % along with Ca2?-saturation or consolidation test to determine the compressibility and
ethylene glycol solvation), respectively. XRD patterns rebound properties and to correct the measured swelling
were collected over an angle 2h range of 3°–85° for the pressure using incremental loading. The last test was
bulk samples and 3°–40° for clay samples. The Inorganic conducted according to the ASTM Standard Test Methods
Crystal Structure Database (ICSD) was used for mineral for One-Dimensional Consolidation Properties of Soils
identification and the Rietveld analysis software (AUTO- Using Incremental Loading (D2435-11). All volume
QUANTM) for material quantification. changes were measured at regular intervals of time using a
Thermo-gravimetric analyses (TGA) along with differ- liner displacement transducer (Model E-311) connected to
ential scanning calorimetry (DSC) were conducted to iden- a data logger and then to a portable computer for digital
tify water loss and phase transitions (from heat flow data recording.
variations) in the soil. About 50 mg of powdered sample was
placed in the analyzer (SETARAM TG-DSC 111) and the
temperature was gradually increased from ambient (30 °C) Results and discussion
to about 800 °C at a rate of 10 °C/min. The set-up was ini-
tially purged with nitrogen (N2) and gas flow was maintained Table 1 gives a summary of the measured geotechnical
throughout the test at a standard rate of 30 cm3/min. index properties at the investigated site (BH-a). The field

123
Swelling properties of expansive soil, Canada 29

Table 1 Summary of measured geotechnical index properties at the investigated site (BH-a)
Field condition Soil classification
Average depth Water content Dry unit weight Specific -0.075 mm -0.002 mm Liquid limit Plastic limit USCS
(m) (%) (kN/m3) gravity (%) (%) (%) (%) symbol

0.2 25 20.1 2.65 56 33 25 24 MH

0.5 33 11.5 2.67 86 72 63 34 MH
1.0 26 13.6 2.71 91 78 63 34 MH
1.4 29 13.2 2.70 93 81 77 32 MH
1.8 34 11.6 2.63 91 72 83 36 CH
2.3 36 12.3 2.71 97 82 90 33 CH
2.8 35 11.2 2.70 94 72 79 32 CH
3.2 37 12.0 2.71 94 75 66 37 MH
3.7 35 12.1 2.71 89 79 65 29 CH
4.2 36 12.9 2.71 91 52 64 33 MH
4.6 37 11.3 2.72 96 70 71 32 CH
5.0 34 11.5 2.75 94 65 60 32 MH
5.5 36 11.7 2.68 92 68 76 29 CH
6.0 35 12.8 2.61 95 71 89 32 CH
6.4 36 11.1 2.74 97 70 65 32 MH
6.9 31 13.5 2.71 96 66 65 32 MH
7.3 34 13.1 2.68 92 71 55 32 MH
7.8 27 10.5 2.73 94 48 58 31 MH
8.3 19 13.8 2.72 75 28 41 24 CL
8.7 17 14.7 2.64 50 21 32 20 CL

Fig. 3 Profiles of geotechnical Water Content (%) Dry Unit Weight (kN/m3) Fines Content (% )
index properties at the 0 25 50 75 100 5 10 15 20 25 0 25 50 75 100

Topsoil
investigated site
1 Undisturbed
Sample

Expansive clay
3
Depth (m)

Till
9
BH-a BH-b BH-c BH-a BH-a BH-b
wf BH-b Clay
wl BH-c Fines

wp

water content from the surface to 7.8 m depth was found to site. Likewise, the specific gravity for most of the samples
be around 30 %, which is typical for local expansive soils was found to be above 2.7 indicating the dominance of clay
and can lead to swelling, if increased (Fredlund and minerals in the profile. The prevalence of fines (material
Rahardjo 1993). With the exception of the surface and finer than 0.075 mm) and clay size fraction (material finer
bottom samples, the dry unit weight measured around 12 ± 1 than 0.002 mm) was confirmed by grain size distribution
kN/m3 and suggested the presence of clayey soils at the data. With the exception of one top and two bottom

123
30 S. Azam et al.

Range of Swelling Potential 100

Q C: Chlorite
(Dakshanamanty and Raman, 1973)
100 Ca: Calcite
No Low Med High V High Extra High 80 D: Dolomite

Intensity (x 1000)
ne
I: Illite

e
80
Plasticity Index (%)

Li
High wl

in
U-

-L
Low wl 60 K: Kaolinite

A
P: Plagioclase
60
Illite 40 Q: Quartz
Q/C S: Smectite
40 Ca
20 S/I Q
Kaolinite K P P D S/I/C S
20 Sand Top Soil and Till
Halloysite
0
Chlorite Expansive Clay 0 10 20 30 40 50 60 70 80 90
0
0 40 80 120 160 200
Angle 2θ
Liquid Limit (%) (a) Bulk Sample
(a) Modified Plasticity Chart
100
S (001) 2+
K: Kaolinite
1.6 Range of Swelling Potential (Seed et al., 1962) 17.1 Å Ca saturated EG solvated S: Smectite
Low Med High Very High 80

Intensity (x 1000)
Top Soil and Till S (001)
1.2 15.2 Å
SP

Expansive Clay 60
Illite Ca2+ saturated RH 54%
=
25

(A=0.9)
Activity

SP
40 K (003)
SP

0.8 =
5% 2.38 Å
=
1.5

Kaolinite K (001) K (002)

%

(A=0.4) 20 3.58 Å
0.4 7.14 Å
0
0 5 10 15 20 25 30 35 40
0
0 20 40 60 80 100 Angle 2θ
Clay Content (%) (b) Clay Sample
(b) Activity vs Clay Content Chart
Fig. 5 X-ray diffraction patterns

Range of Swelling Potential

100 (Van Der Merwe, 1964) 0 -16
TGA
Low Med High Very High
DSC -20
Weight Loss (%)

80
Plasticity Index (%)

Na-Montmorillonite Top Soil and Till -2

DSC (mW)
(A=7.2) Ca-Montmorillonite Expansive Clay -24
60 (A=1.5)
-4
Illite -28
40 (A=0.9)
-6
-32
20 Kaolinite
(A=0.4) -8 -36
0 0 200 400 600 800
0 20 40 60 80 100 Temperature (oC)
Clay Content (%)
Fig. 6 Thermo-gravimetry and differential scanning calorimetry
(c) Modified Activity Chart results
Fig. 4 Volume-change predictions at the investigated site for BH-a
indicated the high water retention and adsorption capacity
of the soil. Most of the samples in the profile were either
samples, the fines generally measured 86–97 % whereas classified as inorganic clays of high plasticity (CH) or
the clays were found to be at least 48 %. The liquid limit inorganic elastic silts (MH). Overall, the geotechnical
(wl) generally varied in the range of 70 ± 15 % with index properties indicated three distinct layers at the site: a
samples at the top and bottom of the borehole exhibiting topsoil (from surface to 0.3 m depth), an expansive clay
lower values because of their coarse grained nature. Similar (from 0.3 to 8 m), and a bottom till (from 8 to 9 m).
trends were observed for the plastic limit (wp) that mea- Figure 3 compares the geotechnical index properties at
sured 33 ± 4 %. The high liquid and plastic limit values the site. The field water contents for all three boreholes

123
Swelling properties of expansive soil, Canada 31

Table 2 Summary of soil composition 1

Corrected
Property Value Swelling
Pressure
Minerals (%) Smectite (32), Quartz (22), Plagioclase 0.95

Void Ratio
(9), Illite (7), Kaolinite (5), Calcite
(4), Feldspar (4), Chlorite (3) Cs = 0.08
? ? 2?
Exchangeable cation Na (1.5), K (0.9), Ca (23.0),
0.9
(cmol(?)/kg) Mg2? (13.9) Cc = 0.16
Cation exchange capacity, 39.4
CEC (cmol(?)/kg)
Specific surface area 49.9 0.85
(m2/g) 100 101 102 103 104
Applied Pressure (kPa)

Fig. 8 Void ratio versus applied pressure

12.5
September for BH-b and BH-c). Further, the water content
Swelling Potential (%)

10 values were found to be close to the plastic limit which is

typical for the local unsaturated expansive soil (Ito and
7.5 Azam 2010). The high and similar wl and wp values for all
three boreholes confirmed that the soil has a high affinity
5
for water. The dry unit weight data showed a similar
2.5 agreement among the three boreholes and generally varied
around 12 kN/m3 for the expansive clay. The fines content
0 profile corroborated well with the three distinct layers at
10-1 100 101 102 103 104 the site. In the clay layer, the fines were more than 90 %,
Time (min) and the amount of clay size material generally measured
(a) Swelling Potential around 75 %.
Figure 4 presents the volume-change predictions at the
150 investigated site for BH-a. Figure 4a shows the plasticity
Swelling Pressure (kPa)

chart, originally developed by Cassagrande (1948) and

125
modified by highlighting swelling potential (SP) ranges
100 (Dakshanamanty and Raman 1973) and clay mineral zones
75 (Holtz and Kovacs 1981). The samples from the expansive
clay which plotted close to the A-line primarily belonged to
50 the illite and kaolinite zones and were characterized as
25 materials falling in the medium-to-high SP range. Con-
versely, the samples from the topsoil and till possessed low
0 swelling potential values. Figure 4b gives soil activity
10-1 100 101 102 103 104
(A = Ip/C) as a function of clay content showing the
Time (min)
activity of common clay minerals (Mitchell and Soga
(b) Swelling Pressure 2005) and ranges of swelling potential (Seed et al. 1962).
Again most of the expansive clay specimens were in the
Fig. 7 Variation of swelling properties with time
high SP zone, while the specimens belonging to the topsoil
and till layers plotted in the low swelling potential zone.
were found to be very close to each other irrespective of the Figure 4c gives the activity chart, which was modified by
cover types and seasonality. It means that the water infil- superimposing the activity of clay minerals (Mitchell and
tration pattern through a 100 mm thick vegetative cover Soga 2005) and identifying the ranges of swelling potential
(BH-b and BH-c) is similar to that through a cracked (Van der Merwe 1964). The clay samples plotted in the
asphalt pavement cover (BH-a). This is partly attributed to zones of illite and kaolinite, with high and very high ranges
high ponded water (as observed in May for BH-a) due to of swelling potential, whereas the topsoil and till samples
melting of snow accumulated over the preceding several were characterized by low-to-medium swelling potential.
winter months. Once infiltrated, the soil deposit tends to Overall, these charts show that the clay deposit can be
retain the water in the dry summer months (as observed in described as a material with medium-to-high swelling

123
32 S. Azam et al.

potential possessing a water adsorption capacity that is The third peak was observed between 600 and 800 °C with
similar to the kaolinite and illite clay minerals. about 2 % weight reduction. This is due to calcite (CaCO3)
Figure 5 gives the X-ray diffraction patterns of the clay. decomposition into solid calcium oxide (CaO) and gaseous
The bulk sample (Fig. 6a) showed the presence of dis- CO2 (Shahraki et al. 2011).
tinctive quartz peaks at 2h = 24.3°, 31.0°, 42.7°, 46.1°, Table 2 summarizes the mineralogical composition of
and 58.9°, thereby ensuring the accuracy of the analysis for Regina clay. The soil was found to contain 51 % clay
other mineral phases (Moore and Reynolds 1997). Quartz minerals comprising of smectite (32 %), illite (7 %), kao-
originated from sandstones whereas the presence of calcite linite (5 %), and chlorite (3 %). The major exchangeable
and dolomite is attributed to calcareous depositions during cations were found to be Ca2? and Mg2? along with a total
the various marine transgressions (Braman et al. 1999). CEC of 40 cmol(?)/kg. These multi-valent ions stabilize
The presence of Na? plagioclase (originating from crys- interactions with pore water in the basal planes of the clay
tallization of molten magma) is indicative of restrained layers, thereby restricting expansion. The measured spe-
leaching in the soil (Mitchell and Soga 2005). This is cific surface area of 50 m2/g indicated a tendency toward
supported by the presence of clay minerals (smectite, illite, moderate water adsorption of the clay; typical values range
chlorite, and kaolinite) around 2h = 24° and 41° in the from 15 m2/g for kaolinite to 800 m2/g for montmorillonite
soil. Derived from weathering of volcanic ash in a marine (Cerato and Lutenegger 2002).
environment, these clay minerals were deposited in a Figure 7a gives the swelling properties of the undis-
proglacial lake that facilitated further conversion of clay turbed clay sample in the form of swelling potential (Dh/h)
minerals to more expansive species (Weaver 1989). The and Fig. 7b in the form of swelling pressure. The evolution
data depicted in Fig. 5b confirmed the presence of a large of both parameters with respect to time followed the same
amount of smectite in the sample. The figure shows that the trend, showing the three habitual swelling stages (Azam
smectite (001) peak with a basal spacing of 15.2 Å at and Wilson 2006): (1) initial low swelling because of low
2h = 8.03° (RH = 54 and Ca2?-saturation) shifted to a unsaturated hydraulic conductivity; (2) intermediate pri-
basal spacing of 17.1 Å at 2h = 6.10° (at RH = 54 % and mary swelling due to an established wetting front; (3)
ethylene glycol solvation) whereas the kaolinite (001) peak secondary low swelling owing to near saturation condi-
did not show any shift. The kaolinite mineral in the tions. The tests were complete in about 1 week and the
investigated soil deposit is associated with the Whitemud final measured values of swelling potential and swelling
Formation rather than in situ alteration due to a low pressure were found to be 12 % and 135 kPa, respectively.
leaching rate (Pruett and Murray 1991). Figure 8 plots void ratio versus the applied pressure for
Figure 6 shows the result of the TGA and DSC tests for the natural expansive clay: the test data duly corrected to
the expansive clay. The data indicated that the entire pro- account for apparatus deformation. Although not presented
cess was endothermic (DSC \ 0) with a total weight loss in the figure, the pre-consolidation pressure (210 kPa) was
of 7.2 % (TGA). The expansive clay exhibited three main higher than the overburden pressure (20 kPa). This over-
endothermic peaks along with distinct water reduction. The consolidation is attributed to the removal of the glacial
first peak occurred between 50 and 350 °C with an asso- mass and excessive desiccation thereafter (Ito and Azam
ciated weight reduction of about 3 %. This was related to 2010). The compression index (Cc) and the swelling index
the loss in adsorbed or interlayer water and confirmed the (Cs) were determined to be 0.16 and 0.08, respectively.
presence of expansive clay minerals in the sample (Azam These values fall within the typical range for inorganic
et al. 2000). The second endothermic peak between 400 clays (Mitchell and Soga 2005). More importantly, the
and 600 °C and the corresponding weight reduction of up swelling pressure required to maintain a constant void ratio
to 2 % was due to the removal of hydroxyl (OH) ions in the measured 135 kPa. This value was corrected to account for
hydroxide sheet of kaolinite (Moore and Reynolds 1997). sample disturbance using the graphical procedure

Table 3 Relationships used to predict swelling potential and swelling pressure

Equations C (%) Relationship References
 1:188
(1) 10–40 SP ¼ 0:16 Ip Azam (2007)
 2:44
(2) 8–65 SP ¼ 0:00216 Ip Seed et al. (1962)
 
(3) 21–93 Ps ¼ 135 þ 2 C þ Ip  wi Sabtan (2005)
n h  1:12  2 2 i o
(4) 30–80 Ps ¼ 6:89 3:5817 102 Ip C =wi þ 3:7912 Nayak and Christensen (1971)

SP swelling potential, Ps swelling pressure, C clay content, Ip plasticity index, wi initial water content

123
Swelling properties of expansive soil, Canada 33

Fig. 9 Swelling properties of Swelling Potential (%) Swelling Pressure (kPa) Cumulative Heave (mm)
the soil deposit

Top Soil
0 10 20 30 40 50 0 100 200 300 400 500 0 100 200 300

Expansive Clay
3

Depth (m)
5

7
Overburden
Pressure

Till
9
Equation (i) Equation (iii) Equation (iii)
Equation (ii) Equation (iv) Equation (iv)
Undisturbed Sample Undisturbed Sample Based on Undisturbed Sample

expansive soils. The test conditions for obtaining all of the

Table 4 Comparison of vertical heave in the deposit
equations used a token surcharge load.
References Calculation basis Depth Vertica Figure 9 summarizes the profiles of swelling properties
(m) heave at the investigated site for BH-a. Portraying similar
(mm)
swelling potential trends because of an identical power law
This study Laboratory 1.2 110 form, equation (1) in Table 3 was found to be more con-
measurement 1.2 260 servative whereas equation (2) in Table 3 was quite sen-
Equation (3) in Table 3 1.2 85 sitive to plasticity index. The former equation estimated an
Equation (4) in Table 3 SP ranging between 5 and 20 % for the investigated
Shah (2011) Field measurement 0.7 75 deposit. Despite being independent of the field water
2.2 55 content, this is a reasonable range of SP variation in Regina
Yoshida et al. Field measurement 0.0 110 clay during summer when wf approaches wp. The predicted
(1983) 0.9 45 SP values from equation (1) in Table 3 are close to those
1.1 35 previously determined through oedometer testing. Ito and
2.7 10 Azam (2010) found the SP to be as low as 2 % for
wi = 30 % and can increase up to 24 % for wi = 15 %.
described by Fredlund and Rahardjo (1993). Horizontal, The SP of 12 % measured for the undisturbed clay sample
tangential, and bisector lines were drawn through the point from 1.2 m depth matched with the predictions from both
of maximum curvature on the compression branch of the equations.
curve. The intersection of the compression line with the The estimated swelling pressure profiles showed similar
bisector line represents the corrected swelling pressure of patterns, such that equation (3) in Table 3 predicted Ps
260 kPa, which was twice the measured value. This is values ranging from 170 to 340 kPa whereas equation (4)
attributed to an increased sample compressibility that in Table 3 computed Ps values to be between 60 and
prevented sample swelling to the in situ void ratio (Azam 170 kPa. The two equations can be considered to provide
and Wilson 2006). an upper bound and lower bound for Ps. This is in agree-
Table 3 provides a list of published semi-empirical ment with measured test data for Regina clay such as those
relationships for determining the swelling properties of the reported by Yoshida et al. (1983) who found that the Ps
investigated deposit. The swelling behavior of soils linearly decreases from 350 kPa at 0.7 m depth to 22 kPa
depends on the clay content (material finer than at 2.2 m depth. The overburden pressure ðr0v Þ required to
0.002 mm), as clay particles have a high specific surface overcome the swelling pressure was found to occur at
area that enables soils to attract water. As explained earlier, 4.6 m depth for equation (3) in Table 3 and at 2.8 m for
the water adsorption capability of a clay depends on its equation (4) in Table 3. The swelling pressure of 260 kPa
mineralogical composition that is determined based on the determined in this study using the undisturbed clay sample
consistency limits and given as Ip. The actual amount of was close to the Ps predicted using equation (3) in Table 3.
water adsorbed by an expansive clay is governed by the The cumulative heave (H) was calculated by adding the
initial water content (wi). The selected equations cover all heave in each layer, with thickness (ho) and void ratio (eo)
geological and field conditions governing swelling of derived from Table 2, and using an average value for

123
34 S. Azam et al.

swelling index (Cs = 0.075) in the following equation (vegetation and cracked road) and the field water
(Fredlund and Rahardjo 1993): content remained close to the plastic limit throughout
H ¼ ho Cs = ð1 þ eo ÞRLog ðPs =r0v Þ ð1Þ the summer months.
3. The clay was characterized by medium-to-high swell-
The Ps value for each layer was obtained from equations ing that was best predicted by SP = 0.16 (Ip)1.188. The
(3) and (4) in Table 3, and that determined in this study soil had 51 % clay minerals including smectite (32 %),
(Ps = 260 kPa) whereas the overburden pressure ðr0v Þ was illite (7 %), kaolinite (5 %), and chlorite (3 %). With a
calculated from knowledge of the soil weight above each CEC of 40 cmol(?)/kg, Ca2? and Mg2? were found to
layer. Showing similar trends, the corresponding be the main exchangeable cations: the specific surface
cumulative heave was found to be 300, 120, and 180 mm area measured 50 m2/g.
at the surface and curvilinearly diminished at depths of 2.8, 4. The SP and Ps for a 1.2 m deep undisturbed sample
4.6 and 3.6 m. were 12 % and 260 kPa, respectively. The correspond-
Table 4 summarizes the vertical heave in the deposit ing cumulative heave was found to be 180 mm at the
calculated through different methods. Heave calculated surface, gradually diminishing to a depth of 3.6 m.
from laboratory measurements of swelling properties was These values were found to be in general agreement
110 mm at 1.2 m depth, which is close to that obtained by with existing consistency-based correlations.
equation (4) in Table 3. At comparable depths, this value is
about 1.5–3 times more than that measured in the field, Acknowledgments The authors would like to acknowledge the
e.g., 70 mm reported by Shah (2011) and 35 mm reported logistic support from the City of Regina and the National Research
by Yoshida et al. (1983). This over-estimation is attributed Council and the financial assistance from the Communities of
to lateral confinement of the sample in the oedometer ring Tomorrow. Thanks to the University of Regina for providing labo-
ratory space and computing facilities. Thanks to the Advanced Sep-
in contrast to the field where the soil also undergoes lateral aration Technologies Division of Natural Resources Canada and the
swelling. Likewise, the oedometer sample could not cap- International Test Centre of the University of Regina for conducting
ture the effect of fissures that off-set part of the horizontal soil composition analysis.
swelling movement. This means that heave predictions
based on oedometer test results should be adjusted using
site-specific correlations and/or field-monitoring data. This References
is supported by preliminary results (data not given in this
paper) of ongoing monitoring at the investigated site. Acton DF, Clayton JS, Ellis JG, Christiansen EA, Kupsch WO (1960)
Physiographic divisions of Saskatchewan. Saskatchewan
Research Council, Saskatoon Map 1
Azam S (2007) Study on the swelling behaviour of blended clay-sand
Summary and conclusions soils. J Geotech Geol Eng 25:369–381
Azam S, Wilson GW (2006) Volume change behavior of a fissured
expansive clay containing anhydrous calcium sulfate. In:
A glacio-lacustrine geology and a semi-arid climate resulted Proceedings of the 4th international conference on unsaturated
in the development of an expansive soil deposit in Regina, soils, vol 1, Arizona, pp 906–915
Saskatchewan, Canada and govern the engineering behavior Azam S, Abduljauwad SN, Al-Shayea NA, Al-Amoudi OSB (2000)
Effect of calcium sulfate on swelling potential of an expansive
of the clay in terms of exhibiting cyclic volume changes. As clay. Geotech Test J 23:389–403
part of an instrumentation program for monitoring the sea- Azam S, Hu Y, Chowdhury R (2009) Sensor installation for field
sonal deformations in asbestos cement water mains, induced monitoring in expansive soils. In: Proceedings of 62nd Canadian
by soil-weather interactions, this study focused on investi- geotechnical conference, Halifax, pp 917–922
Braman DR, Sweet AR, Lerbekmo JF (1999) Upper Cretaceous-
gating soil properties at a selected site. The main conclusions Lower Tertiary lithostratigraphic relationships of three cores
of this research are summarized as follows. from Alberta, Saskatchewan, and Manitoba, Canada. Can J Earth
Sci 36:669–683
1. The soil profile composed of three distinct layers, Brunauer S, Emmett PH, Teller E (1932) Adsorption of gases in
namely: a topsoil (from surface to 0.3 m depth), an multimolecular layers. J Am Chem Soc 60:309–319
expansive clay (from 0.3 to 8 m), and a bottom till Cassagrande A (1948) Classification and identification of soils.
J Transp Eng 113:901–930
(from 8 to 9 m). The middle expansive clay layer was
Cerato AB, Lutenegger AJ (2002) Determination of surface area of
characterized by a fines content of 86–97 % and a clay fine-grained soils by the ethylene glycol monoethyl ether
size fraction of at least 48 %. (EGME) method. Geotech Test J 25:1–7
2. The high liquid (70 ± 15 %) and plastic (33 ± 4 %) Christiansen EA (1979) The Wisconsinan deglaciation of Southern
Saskatchewan and adjacent areas. Can J Earth Sci 116:913–938
limits indicated the high water retention and adsorption
Christiansen EA, Saure EK (2002) Stratigraphy and structure of
capacity of the expansive soil. The nature of this clay Pleistocene collapse in the Regina Low, Saskatchewan, Canada.
layer was found to be unaffected by the type of cover Can J Earth Sci 39:1411–1423

123
Swelling properties of expansive soil, Canada 35

Dakshanamanty V, Raman V (1973) A simple method of identifying geology of Canadian cities. Geological Association of Canada,
an expansive soil. Soils Found Jpn Soc Soil Mech Found Eng Special Paper, Issue, vol 42, pp 147–170
13:97–104 Moore DM, Reynolds RC Jr (1997) X-ray diffraction and the
Fredlund DG, Rahardjo H (1993) Soil mechanics for unsaturated identification and analysis of clay minerals, 2nd edn. Oxford
soils. Wiley, New York University Press, New York
Hamilton JJ (1965) Shallow foundations on swelling clays in Western Nayak NV, Christensen RW (1971) Swelling characteristics of
Canada. In: Proceedings of the international research and compacted expansive soils. Clays Clay Miner 19:251–261
engineering conference on expansive clay soils, vol 2. Texas Pollock DH, Oosterveen J, Jones B, Jaques D, Lebedin J (1987)
A&M University, pp 183–207 Geology field tour of the Regina area. In: 40th Canadian
Hendershot WH, Lalande H, Duquette M (2008) Ion exchange and geotechnical conference, Regina
exchangeable cations. In: Carter MR (ed) Soil sampling and Pruett RJ, Murray HH (1991) Clay mineralogy, alteration history, and
methods of analysis. Canadian Society of Soil Science, Lewis economic geology of the Whitemud Formation, Southern
Publishers, Canada Saskatchewan, Canada. Clays Clay Miner 39:586–596
Holtz WG, Kovacs WD (1981) An introduction to geotechnical Sabtan AA (2005) Geotechnical properties of expansive clay shale in
engineering. Prentice-Hall Inc., Englewood Cliffs Tabuk, Saudi Arabia. J Asian Earth Sci 25:747–757
Hu Y, Hubble DW (2005) Failure conditions of asbestos cement Seed HB, Woodward RJ, Lundgren R (1962) Prediction of swelling
water mains. Can Geotech J 34:608–621 potential for compacted clays. J Soil Mech Found Div 90:53–87
Hu Y, Vu HQ, Lotfian K (2008) Instrumentation of asbestos cement Shah I (2011) The performance of a steel basement house in Regina. Report
pipe in expansive soil. In: Proceedings of the international submitted to the Environmental Systems Engineering, Faculty of
pipelines conference, Atlanta, pp 1–10 Engineering and Applied Science, University of Regina, SK
Ito M, Azam S (2009) Engineering characteristics of a glacio- Shahraki BK, Mehrabi B, Gholizadeh K, Mohammadinasab M (2011)
lacustrine clay deposit in a semi-arid climate. Bull Eng Geol Thermal behavior of calcite as an expansive agent. J Min Metall
Environ 68:551–557 47:89–97
Ito M, Azam S (2010) Determination of swelling and shrinkage Simpson MA (2004) Geology and groundwater resources of the
properties of undisturbed expansive soils. Geotech Geol Eng J Regina area, Saskatchewan. Saskatchewan Research Council,
28:413–422 Publication No. 10420-1E04
Kelly AJ, Sauer EK, Barbour SL, Christiansen EA, Widger RA Van der Merwe DH (1964) The prediction of heave from the
(1995) Deformation of the Deer Creek Bridge by an active plasticity index and percentage clay fraction of soils. Civil Eng S
landslide in clay shale. Can Geotech J 32:701–724 Afr 6:103–107
Mitchell JK, Soga K (2005) Fundamentals of soil behaviour, 3rd edn. Weaver CE (1989) Clays, muds, and shales. In: Developments in
Wiley, New York sedimentology, vol 44. Elsevier, New York
Mollard J, Kozicki P, Adelman T (1998) Some geological, ground- Yoshida RT, Fredlund DG, Hamilton JJ (1983) The prediction of total
water, geotechnical and geoenvironmental characteristics of the heave of a slab-on-grade floor on Regina clay. Can Geotech J
Regina Area, Saskatchewan, Canada. In: Karrow PF (ed) Urban 20:69–81

123