Professional Documents
Culture Documents
Research paper
Centre de Recherche sur la Matire Divise, CRMD and Laboratoire PRISME, Universit d'Orlans, France
Mosul University, College of Engineering, Civil Engineering Department, Al-Majmooah Street, Mosul, Iraq
a r t i c l e
i n f o
Article history:
Received 17 March 2014
Received in revised form 4 October 2014
Accepted 24 October 2014
Available online 13 November 2014
Keywords:
Gypseous soils
Lime treatment
Swell potential
Wettingdrying
Curing conditions
Microstructural changes
a b s t r a c t
Lime treatment generally enhances the geotechnical properties of the soils. However, some studies have
reported an adverse effect, especially in the presence of gypsum. To develop an understanding of the underlying mechanisms, a study of ne-grained soil with different amounts of gypsum was carried out. The
inuence of curing conditions (curing period and curing temperature) and wetting-drying cycles on the
swell potential of ne-grained soil with (0, 5, 15 and 25%) gypsum was investigated to assess the longterm behavior of this type of soil (i.e. gypseous soil) when used in the construction of road and pavement
structures. Mineralogical and microstructural studies (porosimetry, X-ray diffraction and scanning electron microscopy) were carried out to evaluate the cementitious materials and the presence of ettringite.
The results show that increasing the gypsum content decreased the swell potential of the untreated soil
samples. On the other hand, the swell potential values of soil samples started decreasing after the rst
wettingdrying cycle then reached equilibrium. Lime treatment results show that the suppression of
swelling in lime-treated soil samples without gypsum was partly lost after the rst cycle, especially for
short curing periods and for both temperature values (20 C and 40 C), while long-term curing periods
(more than 28 days) gave good results with respect to the suppression of swell potential values. Lime
treatment of gypsum-containing samples showed an adverse effect on swell potential. The swell potential
values increased with curing period and curing temperature, due to the formation of ettringite. Lastly,
wettingdrying cycles caused an increase in the swell potential values of gypsum-containing samples,
cancelling out the benecial effect of lime treatment to control swell.
2014 Elsevier B.V. All rights reserved.
1. Introduction
Gypseous soils are the most common type of soils, and exhibit a
variety of engineering properties and behaviors, such as excessive
settlement (collapsing) and low shear strength, especially upon
wetting. This behavior is mainly due to gypsum dissolution induced
by water ow, thus increasing the porosity and permeability of the
soil. Therefore, the construction of engineering structures founded on
gypseous soils is highly risky as they can suffer from several engineering
problems such as cracks, tilting or differential settlement and structure
collapse (Aibn et al., 1998; Cooper, 1998, 2008).
In order to improve the engineering behavior of these soils,
several techniques are available in geotechnical engineering practice such as mechanical stabilization, chemical stabilization, or a
combination of these two methods, to create an improved soil
material that possesses the desired engineering properties (Little,
1995; Puppala et al., 2004). Chemical soil stabilization by lime
(CaO or Ca(OH)2) is a proven technique for improving the strength
of gypseous soils (Kota et al., 1996; Puppala et al., 2003; Yong and
Corresponding author.
E-mail address: muzahim.al-mukhtar@univ-orleans.fr (M. Al-Mukhtar).
http://dx.doi.org/10.1016/j.clay.2014.10.015
0169-1317/ 2014 Elsevier B.V. All rights reserved.
94
The gypsum used in this study, supplied by the Merck KGaA company,
Germany, is a very ne gypsum and passes through an 80 m sieve
opening, and with a purity of more than 99%.
2.2. Sample preparation
A standard Proctor compaction effort (ASTM D-698) was adopted in
the preparation of the soil samples. To prepare the soil samples, the
oven-dried soil (2 days at 60 C) was rst crushed and passed through
a 4 mm sieve. After grinding, the required amount of soil was mixed
with gypsum under dry conditions. Then the required amount of water
corresponding to the optimum moisture content of natural soil (11%)
was added to the soilgypsum mixtures. All mixing was done manually,
and proper care was taken to prepare homogeneous mixtures. The mixture was then placed in plastic bags for a mellowing time of 24 h.
For lime treatment, the soil samples were stabilized by 3% lime,
which represents the optimum lime content based on the ASTM
D6276 (2006) which used pH to estimate the soillime proportion requirement for Soil Stabilization. This test is derived from the Eades
and Grim test (1966). The soillime mixtures were rst prepared by
thoroughly mixing dry predetermined quantities of soil, gypsum and
lime to obtain a uniform color. Then the required amount of water
was added and again mixed to obtain a uniform moisture distribution.
The mixture was then placed in plastic bags and left for a 1 h mellowing
time (Little, 1995).
After that, the soil samples were statically compacted in rigid stainless steel rings (71 mm in diameter and 20 mm in height), at a strain
rate of 1 mm/min, until they reached the maximum dry unit weight of
the natural soil (17.7 kN/m3). After compaction, the soil samples (i.e.
soil and the ring) were immediately placed in plastic bags to avoid signicant variations in moisture content till testing. The lime-treated soil
samples were immediately wrapped in cling lm and coated with parafn wax to avoid moisture loss, then left to cure for 2, 7, 28, 180 and
360 days at two temperature values (20 C and 40 C).
2.3. Mineralogical and microstructural analysis
The mineralogical and microstructural aspect of the lime-treated soil
samples was studied using scanning electron microscope (SEM), X-Ray
diffraction (XRD) and mercury porosimetry tests. The main objective of
the mineralogical and microstructural investigations was to determine
changes due to lime treatment and to detect the formation of cementitious materials. These tests were conducted on all the soil samples.
For the scanning electron microscopy test (SEM), the preparation of
the specimens followed the method suggested by Tessier (1984) and
described by Al-Mukhtar et al. (1996) in order to preserve the soil fabric.
Soil samples (volume of 1 cm3) were injected by epoxy x resin,
polished, gold-coated and then scanned by a high resolution scanning
electron microscope (PHILIPS XL 40 ESEM). All the soil samples were
prepared in the same manner.
For XRD observations, fractured samples produced after the unconned compression test were powdered, sieved through a 400 m sieve
and dried for 24 h at 40 C. A PHILIPS PW3020 diffractometer was
used for XRD analysis. The diffraction patterns were determined using
Cu-K radiation with a Bragg angle (2) range between 460 running
at a speed of 0.025/6 s.
Pore size distribution tests were carried out using a Pore Seizer
Porosimeter, in which the mercury pressure was raised continuously
to reach more than 210 MPa and measure the apparent pore diameter
in the range of 3.6 nm to 350 m. The method assumes a cylindrical
pore geometry using a modied YoungLaplace equation, generally referred to as the Washburn equation (Giesche, 2006).
P
1
1
r1 r2
2 cos
rpore
H
100
Hi
where H is the vertical displacement in mm, which represents the difference between initial and nal readings of the dial gauge, and Hi is the
initial height of the soil sample in mm. The precision of the displacement measurement was 0.001 mm and so the precision of the axial
strain was 0.005%, thus satisfying the ASTM D6026 recommendation
for geotechnical data.
The time required to reach the nal reading of the dial gauge (the
maximum value of vertical displacement) depends on the gypsum content, curing period and curing temperature. Thus, the swell test was
continued to (2448 h) until the dial gauge reading had stabilized.
95
D
Diall gaauge
Water
Rinng suppporrtinng
mlppe
thhe sam
Loa
L adinng pplaate
( plieed prresssuree
(App
2.75
2 5 kP
Pa)
S
Soil saampple
Porrou
us stonne
0% gypsum
I 002
CAH
15% gypsum
5% gypsum
0% gypsum
Natural soil
0
30
40
50
60
10
20
30
15% gypsum
CSH
5% gypsum
10
20
30
40
50
I003 + Q101
60
15% gypsum
I004
5% gypsum
I 001
I 002
10
20
30
40
50
60
2 ()
Fig. 2. XRD patterns of soil samples cured at 20 C.
during curing and lime treatment, indicating that the quartz was not
substantially attacked by lime.
During curing conditions (curing period and curing temperature),
gypsum addition inuenced the creation (formation) of new reections
such as Ca-hydrates in the samples. With curing, these reections seem
to be reinforced, which means crystallization of these new Ca-hydrates.
As mentioned by Al-Mukhtar et al. (2010a,b, 2012), newly formed Cahydrate cannot be observed by XRD because the phases that form during long-term curing periods (i.e. 180 and 360 days) probably do not
have a well-organized crystalline structure, and therefore X-ray reections are greatly weakened. Second, it is possible that reections from
these phases overlap with both those of primary minerals of natural
soil and/or with the reections formed during 28 days.
To summarize, the major cementitious compounds were calcium silicate hydrates (CSH), calcium aluminate hydrates (CAH) and ettringite
600
400
Natural soil
40
50
60
360 days
25% gypsum
C
L
L
15% gypsum
5% gypsum
0% gypsum
200
Natural soil
I 001
0% gypsum
30
Intensity ( counts/s )
CAH
200
Q
20
CSH
400
10
25% gypsum
600
800
360 days
CAH
2 ()
CSH
CSH + G
G
G
0% gypsum
2 ()
800
I003 + Q101
CAH
5% gypsum
200
Natural soil
15% gypsum
I004
CAH
CSH
I 002
I 001
400
I 001
0% gypsum
200
25% gypsum
600
I004
400
60
180 days
CAH
CSH
I004
CSH + G
CAH
600
50
CAH
25% gypsum
CSH + G
G
G
CAH
180 days
I 002
800
Intensity ( counts/s )
40
2 ()
I 002
CSH + G
800
I003 + Q101
2 ()
CSH
20
CSH
10
I003 + Q101
Intensity ( counts/s )
200
Natural soil
Intensity ( counts/s )
CSH
CAH
400
I 001
200
I004
CSH
5% gypsum
28 days
CSH
15% gypsum
I004
25% gypsum
600
I 002
CAH
CSH + G
25% gypsum
600
400
28 days
Intensity ( counts/s )
I 001
Intensity ( counts/s )
800
G
CAH
CSH + G
800
I003 + Q101
96
Natural soil
0
0
10
20
30
40
50
60
2 ()
Fig. 3. XRD patterns of soil samples cured at 40 C.
mineral, which were found in all the gypseous soil samples. Ettringite
is a calcium aluminum sulfate hydrate (Ca6Al2(SO4)3(OH)12 26H2O.)
type mineral which is responsible for the early strength gain and is
known for its potential to cause swell.
In samples cured at 20 C, ettringite crystals were not observed in
SEM images for any of the gypsum percentages used. The ettringite
crystals start to appear at 180 days of curing and for 25% of gypsum.
The average length of these crystals does not exceed 5 m.
In samples cured at 40 C, the ndings are almost the same as at
20 C but an early appearance of ettringite crystals at 28 days for 25%
of gypsum can be observed. Moreover, ettringite crystals seem to be
larger than those observed at 28 and 180 days when the curing temperature was 20 C. This increase in the dimension of ettringite crystals can
be explained by the effect of thermal activation as temperature increased from 20 C to 40 C (Al-Mukhtar et al., 2010b). After 360 days,
97
Table 1
Pore size distribution from mercury porosimetry results and porosities of samples.
Results concerned the two curing temperature 2040 C
Gypsum content %
Curing period
day
Small pores
0.11 m %
Medium pores
110 m %
Large pores
N10 m %
Porosity by mercury
test %
Initial total
porosity %
28
180
360
28
180
360
28
180
360
28
180
360
2430
3736
4335
3732
4029
3429
3235
4236
3536
3231
3427
3432
4121
3416
3317
3624
3128
3927
4621
3820
5118
5117
4820
5019
3346
2647
2145
2339
2640
2542
1642
1642
1144
1348
1450
1155
33
32
22
34
33
22
53
32
42
43
43
42
26.729.3
27.028.7
25.728.0
26.630.6
30.030.5
27.927.2
28.928.5
28.829.1
29.229.0
29.630.8
30.530.3
30.130.2
33.3
15
25
31.5
30.5
28.6
98
28 days
10 m
5%G
10 m
15%G
10 m
25%G
180 days
Ettrin ite
Ettringite
10 m
5%G
10 m
15%G
10 m
25%G
360 days
Ettringite cristals
Ettringite cristals
10 m
5%G
10 m
Ettringite cristals
15%G
10 m
25%G
Fig. 4. SEM images show cementitious materials and ettringite development for soil samples cured at 20 C.
28 days
Ettringite
10 m
5%G
10 m
15%G
10 m
5%G
10 m
15%G
10 m
25%G
180 days
10 m
Ettringite
25%G
360 days
Ettringite cristals
10 m
Ettringite cristals
Ettringite cristals
5%G
10 m
15%G
10 m
Fig. 5. SEM images show cementitious materials and ettringite development for soil samples cured at 40 C.
25%G
0
0
10
20
30
2 days
7 days
28 days
180 days
360 days
2
1
0
0
10
20
30
2 days
7 days
28 days
180 days
360 days
99
2
1
0% G
5% G
15% G
25% G
0
0
10
20
30
0
0
2
3
No. of Cycle
Fig. 8. Cyclic free swell of untreated soil samples with wettingdrying cycles.
100
inuence on the properties of soil. Fig. 8 shows the swell potential of untreated soil samples with different amounts of gypsum, corresponding
to the number of wettingdrying cycles. The swell potential of untreated soil samples decreased with increasing wettingdrying cycles. The
maximum signicant reduction in the swell potential values was recorded after the rst cycle and this reduction gradually reached equilibrium at the 4th wettingdrying cycle. Further, it is observed that the
effects of wettingdrying cycles on the swell potential of the untreated
soil samples depended on the amount of gypsum addition: the maximum reduction in swell potential occurred for samples with a high gypsum content (i.e. 15% and 25%). At the same time, it was observed that
the trend of reduction in swell potential of soil samples without gypsum
is similar to that of samples with gypsum. The reduction in swell potential is attributed to destruction of the soil matrix (especially the clay
structure matrix) during wettingdrying cycles, as well as to partial
breakdown of the soil particles by reconstruction of the structure of aggregates that occurred during these cycles (Yazdandoust and Yasrobi,
2010). The results obtained in this investigation are in agreement with
those reported by (Dif and Bluemel, 1991; Guney et al., 2007). During
wettingdrying cycles, the soil samples without gypsum cracked as a result of shrinkage and low tensile strength, whereas there were no obvious cracks in the gypseous soil samples, especially those with higher
gypsum content, as shown in Fig. 9. Crack propagation occurred when
the tensile stress due to the increase in suction was equal to or higher
than the cohesion forces. In the untreated samples, as the percentage
of gypsum increases, the percentage of clay in the sample decreases (replacement method) and consequently the shrinkage diminishes. Moreover, gypsum in the soils acts as a ller and participates in the reduction
of suction induced in the clayey parts of the soil. It is worth noting that
after the last wettingdrying cycles most cracks were found to be
0%G
5%G
situated at the same locations as after the rst cycle and the following
cycles enhance only these phenomena. Concerning shrinkage, the dimension can be millimetric and the visible cracks have a homogeneous
distribution and can reach a millimeter.
Figs. 10 and 11 show that the swell potential of the lime-treated soil
samples increased with increasing number of wettingdrying cycles. At
the short-term curing periods (from 2 days to 28 days) and for both
temperature values, the maximum swell potential values were recorded
after the rst wettingdrying cycle. Thereafter, swell values reached
equilibrium or decreased slightly with increasing wettingdrying cycles. Long-term curing periods (more than 28 days) suppressed the
swell potential of the soil samples without gypsum, even with increasing wettingdrying cycles. This behavior is attributed to more cementitious materials being formed during these periods. For the soil samples
containing gypsum, the swell potential values increased with wetting
drying cycles up to the 2nd cycle, then stabilized or decreased slightly.
The results obtained at short-term periods of curing and for both curing
temperatures indicate that the benecial effect of lime treatment is lost
with wettingdrying cycles. These results support the study carried out
by (Khattab et al., 2007). It can be stated that the wettingdrying cycles
caused the gradual destruction of the pozzolanic reaction of limetreated soil samples.
In the treated samples, the claylime reactions are certainly higher
than the suction developed in these soils, which explains why no
shrinkage or cracks were observed.
Moreover, ettringite formation and water absorption also resulted in
increasing the swell potential of the lime-treated soil samples containing gypsum. Swelling in gypseous soil samples is normally associated
with absorption of water during ettringite formation (Hunter, 1988;
Mitchell and Dermatas, 1992). Wild et al. (1993) reported that swelling
15%G
Untreated
25%G
15% G
25% G
2
3
No. of Cycle
2
3
No. of Cycle
0% G
7 days
15% G
5% G
15% G
1.5
25% G
0.5
0% G
7 days
5% G
1.5
15% G
0.5
0.5
25% G
0.5
0
0
2
3
No. of Cycle
2
3
No. of Cycle
28 days
5% G
5% G
15% G
1.5
15% G
1.5
25% G
0% G
0% G
28 days
Free Swell (%)
5% G
1.5
25% G
25% G
0.5
0.5
0
0
2
3
No. of Cycle
0% G
5% G
1.5
180 days
0% G
5% G
15% G
0.5
2
3
No. of Cycle
180 days
15% G
1.5
0% G
2 days
5% G
1.5
0% G
2 days
101
25% G
0.5
25% G
0
0
2
3
No. of Cycle
2
0% G
360 days
0% G
360 days
5% G
1.5
15% G
5% G
15% G
1.5
2
3
No. of Cycle
25% G
0.5
25% G
0.5
0
0
No. of Cycle
Fig. 10. Cyclic free swell of lime treated soil samples cured at 20 C with wettingdrying
cycles.
2
3
No. of Cycle
Fig. 11. Cyclic free swell of lime treated soil samples cured at 40 C with wettingdrying
cycles.
102
conditions and clay content must be made to counter the adverse effects
of lime addition. Such measurements aim to reduce the action of
ettringite formation.
Table 2
Residual water content of lime treated soil with curing conditions.
Gypsum (%)
0
5
15
25
a
2 days
7 days
28 days
180 days
360 days
10.39.8
9.89.4
9.69.2
9.59.1
9.69.5
9.39.1
9.19.0
9.08.8
8.67.7
8.07.1
7.56.9
7.26.1
7.46.4
6.75.7
6.25.0
5.94.8
6.85.7
6.15.1
5.54.5
5.14.0
References
Adams, A.G., Dukes, O.M., Tabet, W., Cerato, A.B., Miller, G.A., 2008. Sulfate induced heave
in Oklahoma soils due to lime stabilization. Geo-congress, Conference Proceedings.
ASCE, pp. 444451.
Aibn, S.A., Al-Abdul Wahhab, H.I., Al-Amoudi, O.S.B., Ahmed, H.R., 1998. Performance
of a stabilized marl base: a case study. Construction and Building Materials 12,
329340.
Al-Mukhtar, M., Belanteur, N., Tessier, D., Vanapalli, S.K., 1996. The fabric of a clay soil
under controlled mechanical and hydraulic stress states. Applied Clay Science 11
(2), 99115.
Al-Mukhtar, M., Lasledj, A., Alcover, J.F., 2010a. Behaviour and mineralogy changes in
lime-treated expansive soil at 20 C. Applied Clay Science 50, 191198.
Al-Mukhtar, M., Lasledj, A., Alcover, J.F., 2010b. Behaviour and mineralogy changes in
lime-treated expansive soil at 50 C. Applied Clay Science 50 (2), 199203.
Al-Mukhtar, M., Khattab, S., Alcover, J.F., 2012. Microstructure and geotechnical properties
of lime-treated expansive clayey soil. Engineering Geology 139, 1727.
ASTM D4546: Standard Test Methods for One-Dimensional Swell or Collapse of Soils.
ASTM D6026: Standard Practice for Using Signicant Digits in Geotechnical Data.
ASTM D6276, 2006. Standard Test Method for Using pH to Estimate the SoilLime Proportion Requirement for Soil Stabilization.
ASTM D698: Standard Test Methods for Laboratory Compaction Characteristics of Soil
Using Standard Effort.
Bell, F.G., 1996. Lime stabilization of clay minerals and soils. Engineering Geology 42,
223237.
Bergaya, F., Lagaly, G., 2006. General introduction: clays, clay minerals and clay science.
In: Bergaya, F., Theng, B.K.G., Lagaly, G. (Eds.), Developments in clay science 1. Handbook of Clay Science 1, pp. 118 (Chapter 12.10).
Bergaya, F., Vayer, M., Lagaly, G., 2006. Cation and anion exchange. In: Bergaya, F., Theng,
B.K.G., Lagaly, G. (Eds.), Developments in Clay Science 1. Handbook of clay science 1,
pp. 9791000 (Chapter 12.10).
Boyadgiev, T.G., Verheye, W.H., 1996. Contribution to a utilitarian classication of gypsum
in soils. Geoderma 74, 321338.
Cooper, A.H., 1998. Subsidence hazards caused by the dissolution of Permian gypsum in
England: geology, investigation and remediation. Geological Society, London, Engineering Geology Special Publication 15, 265275.
Cooper, A.H., 2008. The classication, recording, data basing and use of information about
building damage caused by subsidence and landslides. Quarterly Journal of Engineering Geology and Hydrogeology 41 (3), 409424.
Dif, A.E., Bluemel, W.F., 1991. Expansive soils under cyclic drying and wetting. Geotechnical Testing Journal 14 (1), 96102.
Eades, J.L., Grim, R.E., 1966. A quick test to determine lime requirements for soil stabilization. Highway Research Record 139, 6172.
FAO, 1990. Management of gypsiferous soil. Bulletin, No. 62, Rome, Italy.
FAO (1993): World soil resources Report, 66, Rome, Italy: An Explanation of the FAO
World Soil Resource Map at a Scale of 1:25000000.
Giesche, H., 2006. Mercury porosimetry: a general (practical) overview. Particle & particle
systems characterization 23 (1), 919.
Guney, Y., Sari, D., Cetin, M., Tuncan, M., 2007. Impact of cycling wettingdrying on swelling behavior of lime-stabilized soil. Building and Environment 42, 681688.
Hunter, D., 1988. Lime-induced heave in sulfate-bearing clay soils. Journal of Geotechnical
Engineering, ASCE 114 (2), 150167.
Khattab, S.A.A., Al-Mukhtar, M., Fleureau, J.M., 2007. Long-term stability characteristics of
a lime-treated plastic soil. Journal of Materials in Civil Engineering, ASCE 19 (4),
358366.
Kota, P.B.V.S., Hazlett, D., Perrin, L., 1996. Sulfate-bearing soils: problems with calcium
based stabilizers. Transportation Research Record 1546, 6269.
Little, D.N., 1995. Handbook for Stabilization of Pavement Sub Grade and Base Courses
with Lime. Kendall Hunt Publishing Company, Iowa, USA (by National Lime
Association).
Little, D.N., Nair, S., Herbet, B., 2010. Addressing sulfate-induced heave in lime treated
soils. Journal of Geotechnical and Geoenvironmental Engineering, ASCE 136 (1),
110118.
Mathew, P.K., Rao, S.N., 1997. Effect of lime on cation exchange capacity of marine clay.
Journal of Geotechnical and Geoenvironmental Engineering, ASCE 123 (2), 183185.
Mehta, P.K., 1983. Mechanism of sulphate attack on Portland Cement. Cement and Concrete Research 13, 401406.
Mitchell, J.K., Dermatas, D., 1992. Clay soil heave caused by limesulfate reactions. Innovations and uses of lime. ASTM Special Technical Publication 1135, pp. 4164.
Ouhadi, V.R., Yong, R.N., 2008. Ettringite formation and behavior in clayey soils. Applied
Clay Science 42, 258265.
Petry, M.T., Little, M.D., 1992. Update on sulfate induced heave in treated clays; problematic sulfate levels. Transportation Research Record 1362, 5155 (Washington D.C.).
Puppala, A.J., Wattanasantichatoen, E., Intharasombat, L., Hoyos, L.R., 2003. Studies to understand soil compositional and environmental variables effects on sulfate heave
problems. Proceeding of 12th Pan-American Conference on Soil Mechanics and Geotechnical Engineering, Cambridge, Massachusetts, pp. 833839.
Puppala, A.J., Grifn, J.A., Hoyos, L.R., Chomtid, S., 2004. Studies of sulfate-resistant cement
stabilization methods to address sulfate-induced soil heave. Journal of Geotechnical
and Geoenvironmental Engineering 130 (4), 391402.
103
substitution of lime with ground granulated blast furnace slag. Engineering Geology
51, 257277.
Yazdandoust, F., Yasrobi, S.S., 2010. Effect of cyclic wetting and drying swelling behavior
of polymer-stabilized expansive clays. Applied Clay Science 50, 461468.
Yilmaz, I., Civelekoglu, B., 2009. Gypsum: an additive for stabilization of swelling clay soil.
Applied Clay Science 44, 166172.
Yong, R.N., Ouhadi, V.R., 2007. Experimental study on instability of bases on natural and
lime/cement-stabilized clayey soils. Applied Clay Science 35, 238249.