Download as pdf or txt
Download as pdf or txt
You are on page 1of 10

Applied Clay Science 88–89 (2014) 39–48

Contents lists available at ScienceDirect

Applied Clay Science


journal homepage: www.elsevier.com/locate/clay

Research paper

Geotechnical properties of lime-treated gypseous soils


A. Aldaood a,b, M. Bouasker a, M. Al-Mukhtar a,⁎
a
Research Center on Divided Matter – CRMD, FRE CNRS 3520 1b rue de la Férollerie, 45100 Orléans Cedex 2, France
b
Civil Engineering Department, College of Engineering, Mosul University, Iraq

a r t i c l e i n f o a b s t r a c t

Article history: This paper presents the results of an experimental research study on the behavior of gypseous soils treated with
Received 25 July 2013 lime. The aim of this work was to investigate the effect of a wide range of gypsum content and curing conditions
Received in revised form 12 December 2013 on two important geotechnical properties: the mechanical strength and swell potential of lime-treated fine-
Accepted 15 December 2013
grained soil. For this purpose, soil samples were prepared with different gypsum content (0%, 5%, 15% and
Available online 8 January 2014
25%), treated with lime and assessed at different curing times and temperatures. The results for untreated gyp-
Keywords:
seous soil showed that the unconfined compressive strength increases and the swell potential enhances with
Gypseous soil the presence of gypsum. The geotechnical properties of the lime-treated gypseous soil depend not only on the
Lime stabilization gypsum content but also on the curing conditions. An approximate doubling of the unconfined compressive
Unconfined compressive strength strength of soil samples cured for 28 days was observed when the curing temperature was increased from
Swelling potential 20°C to 40°C. Moreover, the swell potential increased with curing times and decreased with curing temperature
Texture and microstructure in the lime-treated gypseous soil. The tests carried out at the microscopic level showed the formation of calcium
silicate hydrates (CSH) and calcium aluminate hydrates (CAH), which are responsible for strength development
in the treated soil samples. Further, the ettringite mineral, which induces swelling, was also observed in the treat-
ed soils.
© 2014 Elsevier B.V. All rights reserved.

1. Introduction and scientific background The geotechnical properties of gypseous soils can be improved by
various methods, such as chemical stabilization to avoid damage due
Gypsum (CaSO4.2H2O) is one of the most soluble of the common to gypsum dissolution. Improvement with lime treatment takes place
minerals that are found in soils. Gypseous soils are widespread in through two basic chemical reactions: short-term and long-term reac-
many areas of the world (FAO, 1993), and the gypsum content varies tions. The short-term reactions include the cation exchange reaction,
widely from low (less than 5%) to very high (about 50%) (Adams flocculation/aggregation and carbonation, which result in a decrease
et al., 2008; Al-Dabbas et al., 2012). These soils are encountered in con- in the plasticity of the soil and an increase in its workability. The long-
struction and civil engineering works, often in the road pavement layers term reactions include the pozzolanic reaction, in which calcium from
and subgrades due to the lack of economic alternative materials (Adams the lime reacts with the alumina and silica from the clay to produce sta-
et al., 2008; Ahmed, 1985; Aibn et al., 1998; Hunter, 1988; Sultan, 1995; ble calcium silicate hydrate (CSH), calcium aluminate hydrates (CAH)
Wang et al., 2003). Smith and Roberson (1962, as cited in FAO, 1990) and calcium alumino-silicate hydrates (CASH), which are responsible
found that a gypsum content of less than 10% does not significantly af- for the gain in strength, improvements in compressibility and volume
fect the soil characteristics (structure, texture and water retention). In change properties of the soil (Al-Mukhtar et al., 2010a; Bell, 1996; Ingles
contrast, an increase in the gypsum content (more than 10%) causes sig- and Metcalf, 1972; Little, 1995). Immediately after treatment, the stabi-
nificant distress on the geotechnical properties of soil samples (Ahmed, lized gypseous soil exhibits an acceptable range of engineering proper-
1985; Al-Dabbas et al., 2012). Gypseous soils show a rapid and signifi- ties and behaviors. Unfortunately, with time and in the presence of
cant settlement (collapse) when water is added, because the loose water, gypseous soil becomes less durable and new problems can be ini-
structure of their particles is cemented by soluble minerals and/or by tiated. These problems relate to the formation of expansive minerals,
small amounts of clay. Unfortunately, civil engineers have encountered which in turn causes soil heave and then the cracking of pavement
numerous problems with gypseous soils: hydraulic structure failures, structures; these problems induce a reduction in stability and loss
residential building collapse and cracking, and the settlement of pave- of bearing capacity (Hunter, 1988; Little et al., 2010; Mitchell and
ment layers (Aibn et al., 1998; Cooper, 1998; James and Lupton, 1978; Dermatas, 1992; Puppala et al., 2005; Yong and Ouhadi, 2007).
Taha, 1979; Van Alphen and Romero, 1971). The current study explored the influence of lime treatment and gyp-
sum content on the strength and free swell behavior of finely-grained
clayey soil. In this laboratory study, the gypseous soil samples (artificial
⁎ Corresponding author. Tel.: +33 2 38 25 78 81; fax: +33 2 38 25 53 76. samples) were obtained by adding different amounts of gypsum
E-mail address: muzahim@cnrs-orleans.fr (M. Al-Mukhtar). (5%, 15% and 25%) to the natural soil. The natural and gypseous soil

0169-1317/$ – see front matter © 2014 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.clay.2013.12.015
40 A. Aldaood et al. / Applied Clay Science 88–89 (2014) 39–48

samples were treated with 3% lime (based on the pH test method Table 2
suggested by Eades and Grim, 1966). The lime-treated gypseous soil Chemical analysis of natural soil.

samples were cured for 2, 7 and 28 days at 20°C and 40°C. The second Compound Value (%)
curing temperature of 40°C was considered for this research work in
SiO2 68.77
order to study the geotechnical behavior for periods of longer than Al2O 8.35
28 days, since numerous studies have shown that increasing the curing Fe2O3 3.53
temperature accelerates reactions in the lime-treated soils (Al-Mukhtar MnO 0.05
MgO 0.76
et al., 2010b; Mooney and Toohey, 2010; Thompson, 1967). Finally, the
CaO 5.86
evolution of the microstructure was also studied using electronic scan- Na2O 0.61
ning microscopy, X-ray diffraction and mercury porosimetry analysis. K2O 1.46
TiO2 0.73
P2O5 0.08
2. Materials PF 8.62
Total 98.82
2.1. Soil

The soil used in this study was obtained from a site near the Jossigny This lime content was used for the treatment of all soil samples in this
region, to the east of Paris, France. The soil samples were collected at study. To prepare the soil sample, the natural soil was firstly oven-
depths varying between 1.5 to 2.0 m below the ground surface and dried for two days at 60°C; then it was pulverized and sieved (4 mm),
the natural in situ water content was determined to be approximately after which it was mixed with a predetermined amount of gypsum or
18.5%. The liquid limit was 29% and its plasticity index was equal to gypsum and lime, and thoroughly mixed in a dry state until the mixture
8%. The main index and the physical and chemical properties of the nat- had a homogeneous and uniform appearance. In order to investigate
ural soil are summarized in Tables 1 and 2. The grain size distribution only the effect of gypsum content, all the materials were added using
analysis was 17% sand, 64% silt and 19% clay. The X-ray diffraction anal- the complementary substitution method. For example, to prepare the
ysis showed that the soil contained mainly quartz and calcite, while the soil samples with 5% gypsum content, 95% of soil was mixed with 5%
clay mineral was predominantly composed of kaolinite. Based on the of gypsum. In the same way, to prepare the lime-treated soil samples,
Atterberg Limits and according to the Unified Soil Classification System 92% of soil was mixed with 5% and 3% of gypsum and lime, respectively.
(USCS), the soil was classified as low plasticity clay (CL). In the remain- After that, the required amount of water was added to the mixture
der of this paper this soil, which does not contain gypsum, is termed which was then remixed thoroughly. The mixing continued until the
natural soil. final mixture gained a uniform moisture distribution. The wet mixture
was then kept in plastic bags and the untreated soil was left for
24 hours and the lime-treated soil for one hour, to allow it time for
2.2. Hydrated lime and gypsum mellowing. Finally, the soil samples were compacted at optimum mois-
ture content and maximum dry unit weight represented by the stan-
The quicklime used in this study, supplied by the French company dard compaction curve of natural soil (Table 1) according to an ASTM
LHOIST, is a very fine lime which can pass through an 80-μm sieve open- D-698 procedure.
ing. The activity level of the lime used was 94% (Al-Mukhtar et al.,
2010a). 3. Testing methods
The gypsum used in this study, supplied by the German company
Merck KGaA, is very fine gypsum which can pass through an 80-μm 3.1. Unconfined compression and wave velocity tests
sieve opening, and with a purity of more than 99%.
To evaluate the strength characteristics of the lime-treated and
2.3. Sample preparation and compaction test untreated soil, an unconfined compression test was conducted using
cylindrical samples (50 mm diameter × 100 mm height). The cylin-
In order to conduct a precise parametric study, different gypsum drical soil samples were prepared by static compaction at a rate of
contents were added to natural soil (5%, 15% and 25%). The natural 1 mm/min. The soil samples were compacted at the moisture content
and the prepared gypseous soil samples were treated with 3% lime. and dry unit weight that represented the optimum moisture content
The quick test suggested by Eades and Grim (1966) was used to deter- of the standard compaction curve of natural soil, corresponding to the
mine the amount of lime required for the stabilization of the natural soil. maximum dry unit weight. All prepared samples containing lime were
immediately wrapped with cling film and coated with paraffin wax to
reduce moisture loss. The soil samples were cured for 2, 7 and 28 days
Table 1
at 20°C and 40°C.
Some physical and index properties of natural soil.
At the end of each curing period, the wave velocity of the soil sam-
Properties Values ples was determined. A PUNDIT instrument and two transducers
Liquid limit (%) 29 (a transmitter and a receiver) with a frequency of 82 kHz were used.
Plasticity index (%) 8 The direct transmission method, which is more sensitive than other
pH 8.01 methods, was preferred for measuring the wave velocity of the soil
Electrical conductivity (μS/cm) 400
Natural moisture content in situ (%) 18.5
samples. The wave velocity was calculated from the ratio of the travel
Specific gravity Gs 2.66 distance to travel time of the wave through the soil sample. The uncon-
Standard compaction Max. dry density (kN/m3) 17.7 fined compressive strength was then determined according to
Optimum moisture content OMC (%) 11 the ASTM D-5102 procedure (ASTM, 1994), using a Universal Testing
Modified compaction Max. dry density (kN/m3) 19.5
Machine (UTM-INSTRON 4485) at a strain rate of 0.1 mm/min.
Optimum moisture content OMC (%) 10
Grain size distribution Sand (%) 17
Silt (%) 64 3.2. One-dimensional free swell test
Clay (%) 19
USCS Group symbol CL To evaluate the free swell potential of the untreated and lime-
Group name Sandy Lean Clay
treated soil, a free swell test was performed using a standard one-
A. Aldaood et al. / Applied Clay Science 88–89 (2014) 39–48 41

dimensional oedometer device in accordance with the ASTM standard various amounts of gypsum. Soil samples were lyophilized using an
(D-4546). The soil samples were statically compacted at a strain rate ALPHA 1–2 LD plus – GmbH apparatus before applying mercury tests
of 1 mm/min in rigid stainless steel rings that were 71 mm in diameter to minimize micro-fabric changes (Al-Mukhtar et al., 1996).
and 20 mm in height. The lime-stabilized soil samples with different
gypsum contents were subjected to 2, 7 and 28 days of curing at 20°C 4. Results and analysis
and 40°C, as mentioned previously in Section 3.1.
At the end of each curing period, the compacted soil samples in the 4.1. Compaction characteristics
oedometer rings were placed in a consolidation cell between two
dried porous stones, and a sensitive dial gauge was fixed on top of the Table 3 summarizes the compaction characteristics of untreated and
consolidation cell to measure the vertical soil displacement. After cali- lime-treated soil samples. The analysis of the test results depended on
bration with an initial vertical pressure equal to 2.75 kPa (representing the same energy of compaction (the same standard compaction effort),
the weight of the loading plate which is part of the consolidation cell), which gave different initial structures of gypseous soil samples due to
an initial reading was taken to estimate the swell potential, following the gypsum addition. The results show that, as the gypsum content
which the soil samples were soaked with tap water and allowed to increased, the Proctor compaction curves trended towards a slightly
swell under the vertical pressure (2.75 kPa) at a constant laboratory lower maximum dry unit weight (γdmax) but higher water content.
temperature (25°C ± 2°C). After that, time–swell readings were con- The reduction in the γdmax was due to the lower density value of
tinuously noted during the process. The final reading of the dial gauge gypsum particles compared with soil particles. Conversely, the opti-
(which represents the highest reading) was used to calculate the free mum moisture content (OMC) increased with the addition of gypsum.
swell potential together with the initial height of the soil samples, The value of OMC increased from 11% for natural soil to 14.4%, 15.6%
using the equation below: and 16.4% for the soil samples treated with 5%, 15% and 25% gypsum, re-
spectively. This was due to the large specific surface area of gypsum,
ΔH meaning that more water absorption is required for gypsum than for
Swell potential ¼  100 ð1Þ
Hi soil particles. The introduction of a small amount of gypsum tends to re-
duce particle separation (i.e. it causes more dispersion) due to the ex-
where ΔH is the vertical displacement in mm, which represents the dif- change of Ca++ ions between the soil and gypsum, and this results in
ference between initial and final readings of the dial gauge, and Hi is the a lowering of the γdmax and raising of the OMC.
initial height of soil sample in mm. For lime-treated soil samples, laboratory experiments showed an in-
The time required to reach the final reading of the dial gauge crease in the OMC and a decrease in γdmax with lime addition. This be-
(the maximum value of vertical displacement) depends on the gypsum havior can be attributed to the tendency of lime to absorb water in
content, curing time and curing temperature. Thus, the swell test was order to complete its hydration, causing an increase in the OMC. Further,
continued until the dial gauge reading had stabilized for at least 24 h the immediate reactions represented by flocculation and agglomeration
to allow for any residual swelling of the soil samples. resulted in an increase of the OMC and decrease of the γdmax. These var-
iations depended on the mix proportions: the maximum OMC and the
3.3. Mineralogical and micro-structural analysis minimum γdmax were recorded with the soil samples of 25% gypsum
content.
The mineralogical and micro-structural aspect of the lime-treated In summary, the compaction properties of the gypseous soil revealed
soil samples was studied using a scanning electron microscope (SEM), that as gypsum content increases, the γdmax reduces and the optimum
X-Ray diffraction (XRD) and porosity tests. The main objective of the moisture content increases.
mineralogical and micro-structural investigations was to determine
changes due to lime treatment and to find out about the formation of 4.2. Mineralogical and micro-structural changes
cementitious materials and ettringite mineral. These tests were con-
ducted on the soil samples at the end of the 28 days of curing at 20°C Mineralogical and micro-structural studies were carried out to in-
and 40°C. vestigate the mineralogy and micro-structure of the cured samples.
For the scanning electron microscopy test (SEM), the soil samples These specifically focused on determining whether cementitious min-
(volume 1 cm3) were injected by epoxy fix resin, polished, gold- erals such as calcium silicate hydrates (CSH) and calcium aluminate hy-
coated and then scanned by a high resolution scanning electron micro- drates (CAH) were present within the samples, and on analyzing the
scope (PHILIPS XL 40 ESEM). Several digital images at different magnifi- action of these materials on the texture and structure of soil samples.
cations were recorded in order to examine the formation of ettringite Additionally, attention was paid to searching for ettringite, which is
and cementitious materials. It is worth noting that all the soil samples known for its potential to cause swell.
were prepared in the same manner. XRD analyses of natural untreated and lime-treated soil samples
For the X-Ray diffraction test (XRD), fractured samples produced cured for 28 days at 20ºC and 40ºC are shown in Fig. 1. The XRD
after the unconfined compression test were powdered and sieved pattern of the natural (untreated) soil samples indicated that the soil
through a 400-μm sieve to serve as samples for the XRD test. Before test- was composed mainly of kaolinite and illite clay minerals, and
ing, the soil sample was dried for 24 hours at 40°C. A PHILIPS PW3020 contained quartz, calcite and feldspars. Identification revealed that,
diffractometer was used for XRD analysis. The diffraction patterns after the treatment with lime, the formation of CSH and CAH were
were determined using Cu–Kα radiation with a Bragg angle (2θ)
range of 4°–60° running at a speed of 0.025/6 sec.
Table 3
A pore size distribution assessment was carried out to determine the
Compaction characteristics of untreated and lime treated soil samples.
fabric of the soil samples using a Pore Seizer Porosimeter (9320), in
which the mercury pressure is raised continuously to reach more Gypsum Maximum dry unit weight Optimum moisture content
content (%) (kN/m3) (%)
than 210 MPa and measure the apparent pore diameter in the range
3.6 nm–350 μm. In a mercury intrusion porosimetry test, the mercury Without lime With lime Without lime With lime
is forced into the soil samples; the applied mercury pressure and the in- 0 17.7 16.9 11 13
truded volume of mercury are monitored during the test. Mercury in- 5 17.73 17.2 14.4 15.5
trusion porosimetry tests were conducted on the soil samples after 15 17.2 16.6 15.6 17
25 16.5 16.2 16.4 18
28 days of curing at 20°C and 40°C for all soil mixtures containing
42 A. Aldaood et al. / Applied Clay Science 88–89 (2014) 39–48

I 003 + Q101
800
G

CSH + G
G
G A
G
E E 25% gypsum

600

CAH

CAH
C

Intensity ( counts/s )
L
L 15% gypsum

I 004
CSH
400
5% gypsum

K 001
0% gypsum
200 I 001

I 002
Q
K F C Q
K C Natural soil
0
0 10 20 30 40 50 60
2θ (°)

800 I 003 + Q101


G
CSH + G

G
G B
CSH
G
E E E 25% gypsum

600
Intensity ( counts/s )

CAH

CAH
C
L
L 15% gypsum

I 004
CSH

400
5% gypsum

K 001
0% gypsum
200
I 001

I 002

Q
K F C Q
K C Natural soil
0
0 10 20 30 40 50 60
2θ (°)

Fig. 1. XRD patterns of the soil samples after 28 days of curing (A) at 20 °C (B) at 40 °C (G: Gypsum; L: Lime; E: Ettringite; Q: Quartz; K: Kaolinite; I: Illite. C: Calcite; F: Feldspar).

observed. Ettringite was also found in all the gypseous soil samples. increased with an increasing curing temperature, as will be discussed
Ettringite is a calcium aluminum sulfate hydrate (CASH) type of mineral in the next section.
which is responsible for the early strength gain. The high intensity of Gypsum addition affects the pore size distribution of soil samples by
the ettringite mineral was observed within the soil samples having 5% increasing the amount of micropores with a diameter of (0.04–0.05 μm),
gypsum content (especially for the soil samples cured at 40ºC). The which increased with high gypsum content (25%) and when cured at
presence of ettringite within gypseous soil samples can cause structural 20ºC; while the amount of macropores with a diameter of (2–3 μm)
distress. However, the ettringite may not have weakened the soil by did not seem to be significantly affected by this addition. The gypseous
swelling, but instead may have strengthened it (Wild et al., 1998). soil samples showed a new reflexion centered on 6 μm which was not
Fig. 2 illustrates the effect of curing temperatures and gypsum con- present in soil samples without gypsum. This reflexion may be due to
tent on the pore size distribution of lime-treated soil samples. It can the ettringite formations, as shown in Fig. 1. Moreover, ettringite forma-
be seen that, the soil samples without gypsum content exhibited a bi- tions influence the microstructural properties of lime-treated gypseous
modal pore size distribution, with a less pronounced reflexion of pore soil. Such an influence is of great interest since the effect of ettringite on
size centered on 1 μm when cured at 20ºC; they exhibited a tri-modal the geotechnical properties of soil is a function of its microstructural
distribution, with a less pronounced reflexion of pore size centered on changes. Porosimetry measurements provided some additional insight
2 μm when cured at 40ºC. In addition, micropores of less than 0.01 μm into the gypsum addition effects, as exhibited by soil samples treated
seem not to have been significantly affected by the curing temperatures. with lime. It was found that the gypseous soil samples exhibited the
Further, the amount of micro- and macropores reduced with curing same pore size distribution (tri-modal) with gypsum addition at all cur-
temperature due to the pozzolanic reaction products during curing pe- ing temperatures, as shown in Fig. 2. Moreover, the mineral ettringite
riods. The pozzolanic products (CSH and CAH) not only enhanced the fills the pores within the soil matrix, thus leading to a decrease in the
inter-cluster bonding strength but also filled the pore space. As a result, void ratio of the gypseous soil samples. This assumption is in agreement
the unconfined compressive strength of the soil samples significantly with the results of a scanning electron microscope (SEM) analysis
A. Aldaood et al. / Applied Clay Science 88–89 (2014) 39–48 43

Cumultative intrusion (mL/g)


0.025 0.25

Incrimental Intrusion (mL/g)


0% G Curing Curing 0% G

0.02 at 20°C 0.2 at 20°C 5% G


5% G
15% G
15% G
0.015 0.15 25% G
25% G

0.01 0.1

0.005 0.05

0 0
0.001 0.01 0.1 1 10 100 0.001 0.1 10
Entrance Diameter (µm) Entrance Diameter (µm)

Cumultative intrusion (mL/g)


0.25
Incrimental Intrusion (mL/g)

0.025
Curing 0% G Curing 0% G

0.02 at 40°C 5% G 0.2 at 40°C 5% G

15% G 15% G
0.015 0.15 25% G
25% G

0.01 0.1

0.005 0.05

0 0
0.001 0.01 0.1 1 10 100 0.001 0.01 0.1 1 10 100
Entrance Diameter (µm) Entrance Diameter (µm)

Fig. 2. Incremental and cumulative pore size distribution of the soil samples cured for 28 days at 20°C and 40°C.

performed on soil samples cured for 28 days at 20°C and 40°C, as shown silicate hydrate (CSH) and calcium aluminate hydrates (CAH), which
in Fig. 3. Ettringite was observed to have formed and precipitated in the act as cementitious materials (Fig. 1) so that they in turn contribute to
pores of the soil matrix, especially of the soil samples with a higher the strength development of the soil samples.
amount of gypsum content. It can be noted from Fig. 2 that the The unconfined compressive strength values of soil samples contain-
soil samples cured at 20°C and 40°C showed a higher proportion of ing gypsum were higher than those of the zero gypsum samples. This
macropores centered at around 2 μm. The soil samples with a gypsum behavior can be attributed to the faster hydration rate (greater reaction
content of 5% had a macropore reflexion that was slightly higher than between soil, lime and gypsum) and the formation of ettringite by
for other types of soil. This reflexion is at 3 μm and 5 μm for a tempera- the reaction of sulfate ions with CSH or CAH, as shown in Fig. 1. It is
ture of 20°C and 40°C respectively. The same figure also shows that at known that gypsum improves soil strength by altering the course of
20°C the micropore reflexion is more pronounced for the soil containing the hydration of calcium silicate, which is predominantly formed in
gypsum: it is centered at around 4 nm. At 40°C, this pore mode is more the early stages of hydration. The addition of gypsum leads to the
flattened, meaning that the micropores are present in the range release of sulfate ions which react with the alumina phase in the soil.
4–20 nm. Gypsum accelerates the chemical reaction between soil and lime
(Holm et al., 1977; Kujala and Nieminen, 1977). Silicate hydrate is
4.3. Unconfined compressive strength properties formed as well as calcium aluminate hydrate, which favors stronger
soil samples.
The unconfined compressive strengths of natural and gypseous soil Gypsum has a significant activation effect on soil-lime, with the
samples compacted at OMC and γdmax were measured. The soil samples main hydration products being CSH, CAH and ettringite. Furthermore,
were found to have unconfined compressive strengths of 0.19, 0.23, 0.27 the formation of CSH and CAH are not the only factors contributing to
and 0.32 MPa for 0%, 5%, 15% and 25% gypsum, respectively. The in- the higher strength of the gypseous soil samples, as ettringite also
crease in strength values with gypsum content is explained by the re- plays an important role.
duction in the void ratio of the soil samples, as shown in Table 4. The effect of the gypsum content was most pronounced in the soil
Indeed, gypsum particles not exceeding 80 μm in diameter will fill the samples treated with 5% gypsum, which demonstrated the largest un-
pore space between soil particles, therefore increasing the compactness confined compressive strength (Fig. 4). The soil samples treated with
of the matrix and consequently the unconfined compressive strength of 15% gypsum were the next strongest, while the soil samples treated
the soil samples. Further, decreasing the void ratio in tested soil samples with 25% gypsum exhibited the lowest strength. The presence of
with the same moisture content (i.e. 11%) may introduce increased unreacted gypsum particles, which represent light-weight materials
suction pressure which results in increased unconfined compressive with a specific gravity equal to 2.3, was the cause of this reduction
strength. in the unconfined compressive strength of the soil samples. Also, the in-
The results for the unconfined compressive strength of the lime- clusion of gypsum content affects the texture of the soil samples: as the
treated soil samples with various amounts of gypsum content and sub- gypsum content increased the size of the macropore increased (Fig. 2)
jected to different curing conditions are displayed in Figs. 4 and 5. The resulting in a reduction of the unconfined compressive strength.
unconfined compressive strength of all mixes, with or without gypsum, Added to this, the reactions forming the cementitious materials are
increased with increased curing temperature. This behavior is attribut- also affected by gypsum content, causing these changes in the uncon-
ed to the acceleration of chemical reactions in the mixtures due to the fined compressive strength.
increase in temperature. In fact, a higher temperature promotes the The unconfined compressive strength of the soil samples increased
pozzolanic reaction within the mixture and the formation of calcium with curing time and curing temperature (Fig. 4). The increased
44 A. Aldaood et al. / Applied Clay Science 88–89 (2014) 39–48

0% G

Cementing
Cementing
Materials
Materials

5% G
Cementing
Materials

Cementing
Materials

15% G

Ettringite
Ettringite

25% G

Ettringite Ettringite

Fig. 3. Cementing materials development during 28 day of curing at 20°C at left and 40°C at right.

strength ranged from 2 to 6 and 2 to 10 times the original natural


strength for the soil samples cured at 20°C and 40°C, respectively. More-
Table 4
over, the results showed an approximate doubling in the unconfined
Variation of specific gravity and void ratio with gypsum addition.
compressive strength of the soil samples cured for 28 days when the
Gypsum (%) Specific gravity (Gs) Void ratio (%) curing temperature was increased from 20°C to 40°C.
0 2.66 0.5 Fig. 4 makes it clear that the trend in the rate of gain in the strength
5 2.6 0.46 of the soil samples cured at 40°C was sharper than that for the soil
15 2.56 0.44 samples cured at 20°C. The continuous reaction between soil, lime and
25 2.49 0.4
gypsum with increased temperature, as well as the formation of CSH,
A. Aldaood et al. / Applied Clay Science 88–89 (2014) 39–48 45

2.5 4.4. Wave velocity measurements

2 The wave velocity test method is one of the non-destructive test


methods used to evaluate the stiffness properties of materials. A wave
velocity test was conducted to evaluate the untreated and lime-
UCS (MPa)

1.5 treated soil mixtures and to determine the variation and correlation of
the wave velocity with gypsum content and curing conditions. The
wave velocity of the soil samples was measured before performing the
1
unconfined compression test.
The wave velocity values of natural and gypseous soil confirmed
0%G at 40°° C 0%G at 20°° C
0.5 the results for unconfined compressive strength: 617, 642, 664 and
5%G at 40°° C 5%G at 20°° C
15%G at 40°° C 15%G at 20°° C 755 m/sec for 0%, 5%, 15% and 25% gypsum, respectively. The increase
25%G at 40°° C 25%G at 20°° C in the wave velocity values with the addition of gypsum was due to
0 the reduction in the void ratio of soil samples, as discussed previously.
0 10 20 30
Moreover, in a three-phase system (solid, liquid and gas) such as
Curing Period (day) compacted soil, wave transmission occurs through all the phases. Gen-
Fig. 4. Unconfined compressive strength variation Vs curing time.
erally, wave velocities in solids are higher than velocities in liquids,
which in turn are higher than velocities in gases. Therefore, a lower
amount of voids in the soil samples gives higher wave velocity values
CAH and ettringite minerals, caused the soil samples cured at 40°C to compared to the other samples.
consistently grow more quickly. Additionally, the refinement of the The results of the wave velocities of lime-treated soil samples are
pore structure and decrease of the porosity of the soil samples with in- presented in Figs. 6 and 7. The results show similar trends to the results
creased curing temperature contributed to the higher rate of increase in for the unconfined compressive strength. For all levels of gypsum con-
the unconfined compressive strength of the soil samples when com- tent, the data demonstrates that the wave velocity essentially increased
pared to those samples cured at 20°C, because of the higher amount of linearly with the curing period (Fig. 6) and for both curing tempera-
hydration products that were formed. This argument is consistent tures. After 28 days, the soil samples treated with 0%, 5%, 15% and 25%
with the results of the mercury intrusion porosimetry test, as shown gypsum attained maximum wave velocity values of the order of 1483,
in Fig. 2. 1609, 1533 and 118 m/sec for curing temperatures of 20 °C, and 1602,
Finally, many attempts were made to directly correlate the uncon- 1716, 1643 and 1225 m/sec for curing temperatures of 40 °C, which
fined compressive strength values with the water content after the cur- gave for the same curing temperatures an improvement ratio of 2.4,
ing (residual water content) of the soil samples, as shown in Table 5. 2.5, 2.3 and 1.4, and 2.5, 2.6, 2.4 and 1.6 times the wave velocity of the
Residual water content is the ratio of the weight of water to the weight natural soil (950 m/sec), respectively.
of the solid after curing periods. The residual water content was ob- The variation of the wave velocity with the unconfined compressive
served to decrease progressively with the length of curing period and strength values for all soil mixtures is illustrated in Fig. 8. Generally,
curing temperatures, and it also decreased with increasing gypsum con- the wave velocity values increased with unconfined compressive
tent. As expected, the lower value of residual water content was related strength values and the relationship was linear with the coefficient of
to the higher gypsum contents. This further reinforces the fact that the determination (R2 = 0.65). However, a significant amount of scatter
addition of gypsum had a greater influence on the residual water con- was observed in the data, which could be attributed to the unconfined
tent than the curing period, due to the high affinity of these materials compressive strength of the soil being the property of the material
to water. The reduction in the residual water content is attributed to that is most affected by its composition, including cementing material,
the hydration process of lime. Also, as the curing period increased, the voids filled with air and water and other parameters (fissures, cracks
amount of pozzolanic compounds increased, resulting in the increase etc.). Thus, the unconfined compressive strength of the soil samples is
of the weight of solids per unit volume. Further, ettringite formation not directly related to the velocity of the wave propagation; although,
leads to a further decrease in residual water content due to the chemical in general, a higher unconfined compressive strength can be expected
consumption of water during ettringite nucleation and subsequent to be associated with a higher velocity.
growth. The reduction of water content and the formation of ettringite
minerals negatively influenced the strength properties of the soil sam- 4.5. One-dimensional free swell properties
ples, especially the soil samples with higher gypsum content.
The swell potentials of the natural and gypseous soil samples,
2.5
2 days at 40° C compacted at OMC and γdmax, were measured. The soil samples had
7 days at 40° C swell potentials of 1.9, 1.4, 0.8 and 0.4 for 0%, 5%, 15% and 25% gypsum
28 days at 40° C
2 2 days at 20° C content, respectively. The decrease in the swell potential values of the
7 days at 20° C soil samples was due to the addition of non-expansive materials
28 days at 20° C
(gypsum) and the reduction in the soil matrix, especially clay content
UCS (MPa)

1.5 (due to the complementary substitution method used in the samples'


preparation). Also, the void ratio decreased with gypsum content, as
1 presented in Table 4, since gypsum acts mainly as an inert filler and
tends to reduce the swell of the soil samples. Similar observations
were reported by Yilmaz and Civelekoglu (2009) during their study
0.5 on the effect of gypsum addition on the behavior of expansive soil
(bentonite soil). The swell potential decrement was larger when the
0 addition of gypsum was larger, and the reduction in value of the swell
0 10 20 30 potential reached 70% for the soil samples with 25% gypsum content.
Gypsum Content (%) Figs. 9 and 10 illustrate the effects of the curing conditions and gyp-
sum content on the free swell potential of lime-treated soil samples.
Fig. 5. Unconfined compressive strength variation Vs gypsum content. From these two figures it can be seen that the free swell potential of
46 A. Aldaood et al. / Applied Clay Science 88–89 (2014) 39–48

Table 5
Residual water content and unconfined compressive strength of soil samples.

Curing temperature (°C) Gypsum (%) Curing period (day)

2 7 28

w/c (%) UCS (MPa) w/c (%) UCS (MPa) w/c (%) UCS (MPa)

20 0 10.3 0.42 9.6 0.68 8.6 0.76


5 9.8 0.79 9.3 1.26 8 1.4
15 9.6 0.75 9.1 1.12 7.5 0.98
25 9.5 0.57 9 0.93 7.2 0.89
40 0 9.8 0.6 9.5 0.74 7.7 1.45
5 9.4 1.16 9.1 1.77 7.1 2.44
15 9.2 0.96 9 1.54 6.9 1.8
25 9.1 0.78 8.8 1.26 6.1 1.47

lime-stabilized soil samples without gypsum decreased with increased the formation of ettringite nuclei which have the ability to grow, as
curing time, and that a greater reduction was observed for soil samples the necessary chemical compounds for this are available. Moreover, al-
cured at 40°C compared to samples cured at 20°C. The free swell poten- though no crystalline ettringite was observed in the soil samples with
tial became suppressed at seven days for soil samples cured at 40°C and 5% gypsum content (Fig. 3), it could that the presence of an amorphous
at 28 days for samples cured at 20°C. This reduction is attributed to the ettringite product causes the swell, as mentioned by Wild et al. (1999).
addition of lime, which is known to be an effective stabilizer for reduc- Further, X-ray diffraction test results emphasized the ettringite forma-
ing the swell potential of clay soils. When lime is added to the soil, there tion of the soil samples with 5% gypsum content, and the highest inten-
is an immediate effect on the soil properties caused by cation exchange, sities of ettringite mineral were found in these samples. At higher levels
which results in a reduction in the diffuse double layer and soil plastic- of gypsum content (more than 5%), the soil samples exhibited less
ity, thus causing flocculation and agglomeration. The other reaction is swelling in comparison to samples with a 5% gypsum content, even
the pozzolanic reaction, which is time- and temperature-dependent though there were ettringite crystals, as shown in Fig. 3. This was due
(Al-Mukhtar et al., 2010a,b; Little, 1995), and which results in cementi- to unreacted gypsum particles, which act as inert filler materials and re-
tious materials (Fig. 2) that bind the soil particles together. Thus the strict the swell. Also, these samples appeared to contribute less to
magnitude of free swell is decreased. ettringite formation than did the pozzolanic reactions, and a small
For the soil samples with added gypsum, the free swell potential swell potential was therefore observed with an increased gypsum
increased with curing time (Fig. 9). The free swell potential values content.
increased from 0% for soil samples without gypsum to 0.85%, 0.5% and Finally, increasing the temperature from 20°C to 40°C led to a
0.35% for soil samples cured for 28 days at 20°C, and to 0.75%, 0.45% greater reduction in swell potential. This may be attributed to the
and 0.25% for soil samples cured for 28 days at 40°C, with 5%, 15% and greater formation of cementitious materials during curing at 40°C
25% gypsum content, respectively. During two days of curing at both than at 20°C. In addition, soil samples cured at 20°C show a higher
temperature values, the swell potential of the soil samples decreased proportion of macropores, a coarser pore structure and higher poros-
with the addition of gypsum (Fig. 9). This behavior was due to the pres- ity values, when compared with soil samples cured at 40°C. Thus the
ence of unreacted lime and gypsum particles, which act as filler mate- small pores and lower porosity give little liberty, if any, for the clay
rials and tend to reduce the swell potential. particles to expand because of the cementitious materials which
After two days of curing, the swell potential of the soil samples formed during the curing process. Moreover, smaller pore size and
increased up to 5% gypsum content, and then decreased with more lower porosity decreases the soil's ability to retain further water
swelling occurring for soil samples cured for 28 days. In the case of upon wetting, and this in turn reduces the swell potential.
the lime-treated soil samples with 5% gypsum content, the increase in
the swell potential can be explained by the chemical reactions caused 5. Conclusions
by the presence of gypsum, alumina and lime. These reactions lead to
The aim of this work was to investigate the effect of gypsum content
and curing conditions on two important geotechnical properties: the

1750
1750

1500
Wave Velocity (m/sec)

Wave Velocity (m/sec)

1500

1250
1250

1000 2 days at 40° C


1000
7 days at 40° C
0%G at 40° C 0%G at 20° C 28 days at 40° C
750 2 days at 20° C
5%G at 40° C 5%G at 20° C 750
15%G at 40° C 15%G at 20° C 7 days at 20° C
25%G at 40° C 25%G at 20° C 28 days at 20° C
500 500
0 10 20 30 0 10 20 30
Curing Period (day) Gypsum Content (%)

Fig. 6. Wave velocity variation Vs. curing conditions. Fig. 7. Wave velocity variation Vs. gypsum content.
A. Aldaood et al. / Applied Clay Science 88–89 (2014) 39–48 47

2.5 1
2 days at 40° C 2 days at 40°° C
y = 0.0017x - 1.0543
7 days at 40° C 7 days at 40° C
R² = 0.65
2 28 days at 40° C 28 days at 40°° C
0.75 2 days at 20°° C
2 days at 20° C

Free Swell (%)


7 days at 20°° C
UCS (MPa)

1.5 7 days at 20° C


28 days at 20°° C
28 days at 20° C
0.5
1

0.5 0.25

0
0
750 1000 1250 1500 1750 0 10 20 30
Wave Velocity (m/sec) Gypsum Content (%)
Fig. 8. Unconfined compressive strength and wave velocity relationship of all soil samples. Fig. 10. Free swell potential vs. gypsum content.

mechanical strength and swell potential of lime-treated fine-grained enhancing their mechanical properties, but the amount of lime
clayey soil. In order to conduct a precise parametric study, all tested added must be adjusted according to the mineralogical composition
samples were prepared in the laboratory. Moreover, studies in the liter- and the gypsum content of the soil.
ature have shown that the gypsum content in nature varies widely. - Tests at the microscopic level (porosimetry, XRD and SEM) are the
Thus, to control the exact gypsum content in the different samples test- key to better insight into the lime stabilization reactions of gypseous
ed and in order to study an important range of gypsum soil behavior, soils and the evolution of the geotechnical properties of these soils.
samples were prepared with three gypsum contents: 5%, 15% and 25% Analyses of lime-treated gypseous soils revealed the formation of
by dry mass of soil. The following conclusions can be drawn from this new hydrates (CSH) and (CAH) and also ettringite. The high intensi-
study: ties of the new minerals were observed within the soil samples
cured at 40ºC. An understanding of the changes occurring in
- The presence of gypsum enhanced the mechanical properties but the mineralogy and at the microstructural levels (porosities and
the swelling potential decreased as the gypsum content increased, pore size distribution) is essential for a better interpretation of the
with a more pronounced effect with a higher gypsum content. modifications in the geotechnical properties of the lime-treated gyp-
- It is necessary to determine the gypsum content before lime stabili- seous soil.
zation in order to investigate the negative or positive impact on the
effectiveness of soil stabilization.
- For the tested soil treated with 3% lime, the presence of gypsum en- References
hanced the strength properties of lime-treated soil up to a certain
Adams, A.G., Dukes, O.M., Tabet, W., Cerato, A.B., Miller, G.A., 2008. Sulfate induced heave
limit (5% gypsum), and as the gypsum content further increased in Oklahoma soils due to lime stabilization. Geo-congress, Conference Proceedings,
the strength properties decreased for all curing conditions. ASCE, pp. 444–451.
- The contribution of lime to stabilization was found to be more prom- Ahmed, A.A., 1985. Lime stabilization of soils containing high soluble salts contents. (M.Sc.
Thesis) Civil Engineering Department, College of Engineering, University of Mosul,
inent for the soil samples with lower gypsum content (5%). This im- Iraq.
plies that lime stabilization plays an important role in the formation Aibn, S.A., Al-Abdul Wahhab, H.I., Al-Amoudi, O.S.B., Ahmed, H.R., 1998. Performance of a
of a compact matrix for mixes with a lower gypsum content. stabilized marl base: a case study. Constr. Build. Mater. 12, 329–340.
Al-Dabbas, M.A., Schanz, T., Yassen, M.J., 2012. Proposed engineering of gypsiferous soil
- The beneficial effect of lime stabilization in controlling the swell po- classification. Arab. J. Geosci. 5, 111–119.
tential of gypseous soil is partially lost due to the ettringite forma- Al-Mukhtar, M., Belanteur, N., Tessier, D., Vanapalli, S.K., 1996. The fabric of a clay soil
tion, especially when the soil is subject to long curing times. under controlled mechanical and hydraulic stress states. Appl. Clay Sci. 11 (2),
99–115.
- Lime treatment can be effective in stabilizing gypseous soils and in Al-Mukhtar, M., Lasledj, A., Alcover, J.F., 2010a. Behaviour and mineralogy changes in
lime-treated expansive soil at 20 °C. Appl. Clay Sci. 50, 191–198.
Al-Mukhtar, M., Lasledj, A., Alcover, J.F., 2010b. Behaviour and mineralogy changes in
0%G at 40°° C 0%G at 20°° C lime-treated expansive soil at 50 °C. Appl. Clay Sci. 50 (2), 199–203.
1 ASTM, 1994. Annual book of ASTM standards. Soil and Rock, vol. 04.08. American Society
5%G at 40°° C 5%G at 20°° C
15%G at 40°° C 15%G at 20°° C
for Testing and Materials, Philadelphia.
Bell, F.G., 1996. Lime stabilization of clay minerals and soils. Eng. Geol. 42, 223–237.
25%G at 40°° C 25%G at 20°° C
Cooper, A.H., 1998. Subsidence hazards caused by the dissolution of Permian gypsum in
0.75 England: geology, investigation and remediation. Engineering Geology Special Publi-
cation, 15. Geological Society, London 265–275.
Free Swell (%)

Eades, J.L., Grim, R.E., 1966. A quick test to determine lime requirements for soil stabiliza-
tion. Highw. Res. Rec. 139, 61–72.
FAO, 1990. Management of gypsiferous soil. Bulletin No. 62, Rome, Italy.
0.5
FAO, 1993. World soil resources: An Explanation of the FAO World Soil Resource Map at a
Scale of 1:25000000. FAO World Soil Resource Report 66, Rome, Italy.
Holm, G., Trank, R., Ekstrom, A., 1977. Improving lime column strength with gypsum.
Proceeding of the 9th International Conference on Soil Mechanics and Foundation
0.25 Engineering. 3, pp. 903–907.
Hunter, D., 1988. Lime-induced heave in sulfate-bearing clay soils. J. Geotech. Eng. ASCE
114 (2), 150–167.
Ingles, O.G., Metcalf, J.B., 1972. Soil Stabilization Principles and Practice. Butterworth,
0 Sydney.
0 10 20 30 James, A.N., Lupton, A.R.R., 1978. Gypsum and anhydrite in foundations of hydraulic struc-
tures. Geotechnique 28 (3), 249–272.
Curing Period (day) Kujala, K., Nieminen, P., 1977. On the reactions of clays stabilized with gypsum lime. Pro-
ceedings of the 9th International Conference on Soil Mechanics and Foundation Engi-
Fig. 9. Free swell potential vs. curing conditions. neering. 3, pp. 929–936.
48 A. Aldaood et al. / Applied Clay Science 88–89 (2014) 39–48

Little, D.N., 1995. Handbook for Stabilization of Pavement Sub Grade and Base Courses Thompson, M.R., 1967. Factors influencing the plasticity and strength of lime-soil mix-
with Lime. National Lime Association, Kendall Hunt Publishing Company, Iowa, USA. ture. Bulletin 492, Engineering Experiment Station, Univeristy of Illinois, Urbana.
Little, D.N., Nair, S., Herbet, B., 2010. Addressing sulfate-induced heave in lime treated Van Alphen, J.G., Romero, F.D.R., 1971. Gypsiferous soils. Bulletin 12, International Insti-
soils. J. Geotech. Geoenviron. 136 (1), 110–118. tute for Land Reclamation and Improvement, Wageningen, Netherlands.
Mitchell, J.K., Dermatas, D., 1992. Clay soil heave caused by lime-sulfate reactions: innova- Wang, L., Roy, A., Seals, R.K., 2003. Stabilization of sulfate-containing soil by cementitious
tions and uses of lime. ASTM Spec. Tech. Publ. 1135, 41–64. mixtures: mechanical properties. Transp. Res. Rec. 1837, 12–19.
Mooney, M.A., Toohey, N.M., 2010. Accelerated curing and strength-modulus correlation Wild, S., Kinuthia, J.M., Jones, G.I., Higgins, D.D., 1998. Effects of partial substitution of lime
for lime-stabilized soils. Report No. CDOT-2010-1; Final Report Colorado Dept. of with ground granulated blast furnace slag (GGBS) on the strength properties of lime-
Transportation DTD. Applied Research and Innovation Branch (56 pp). stabilized sulfate-bearing clay soils. Eng. Geol. 51, 37–53.
Puppala, A.J., Intharasombat, N., Vempati, R.K., 2005. Experimental studies on ettringite- Wild, S., Kinuthia, J.M., Jones, G.I., Higgins, D.D., 1999. Suppression of swelling asso-
induced heaving in soils. J. Geotech. Geoenviron. 131 (4), 325–337. ciated with ettringite formation in lime stabilized sulphate bearing clay soils by partial
Sultan, S.A., 1995. Design of low volume rural roads in level Iraqi terrain. (PhD. Thesis) substitution of lime with ground granulated blast furnace slag. Eng. Geol. 51, 257–277.
Department of Civil Engineering, College of Engineering, University of Baghdad, Iraq. Yilmaz, I., Civelekoglu, B., 2009. Gypsum: an additive for stabilization of swelling clay soil.
Taha, S.A., 1979. The effect of leaching on the engineering properties of Qayiara soil. Appl. Clay Sci. 44, 166–172.
(M.Sc. Thesis) Civil Engineering Department, College of Engineering, University of Yong, R.N., Ouhadi, V.R., 2007. Experimental study on instability of bases on natural and
Mosul, Iraq. lime/cement-stabilized clayey soils. Appl. Clay Sci. 35, 238–249.

You might also like