Cold Regions Science a nd Technology 12 4 ( 2 0 16 ) 6 7– 7 6

Contents lists available at ScienceDirect

Cold Regions Science and Technology

journal h omepag e: ww w . e l se v ie r . c o m /l o c a t e/c o l d r e g i o n s

An experimental study on the influence of cooling rates on salt expansion

in sodium sulfate soils
Yuanming Lai a,c,⁎, Xusheng Wan b , Mingyi Zhang a
a
State Key Laboratory Frozen Soil Engineering, Cold and Arid Region Environmental and Engineering Institute, Chinese Academy of Sciences, Lanzhou, Gansu 730000, China
b
School of Civil Engineering and Architecture, Southwest Petroleum University, Chengdu, Sichuan 610513, China
c
School of Civil Engineering, Lanzhou Jiaotong University, Lanzhou, Gansu 730070, China

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

Article history: This experimental study examined salt expansion in sodium sulfate soils. The results indicated that salt expan-
Received 1 May 20 15 sion was caused by t he crystallization of sodium sulfate decahydrate w h e n t he supersaturation ratio of the
Received in revised form 4 December 2015 sodi-um sulfate solution in t he soil reached a maxi mum value. The maxi mu m supersaturation ratio of t he
Accepted 31 December 2 015
sodium sulfate solution decreased as t he sodium sulfate content increased, and t he supersaturation ratio
Available online 2 4 January 201 6
decreased as crystals continued to precipitate. The total amount of crystal generation was related to t he
crystal growth rate. The rate of crystal growth decreased exponentially as te mperature decreased. The
Keywords:
Sodiu m sulfate saline soil formation of crystals was in-fluenced by the rate of cooling. When mirabilite began to precipitate, t he
Salt expansion supersaturation rate was reduced ver-sus t he rate of cooling, whereas t he salt expansion rate of t he soil
Supersaturation ratio increased. When t he salt content of sodium sulfate was less than 1.5%, salt crystals began to form after t he
The r at e of crystal growth phase change of t he solution from water to ice, but salt expansion did not occur. Whe n salt content of
Phase change sodium sulfate was less than 0.7%, t h e precipitation of salt crystals was minimal or did not occur.
Qinghai–Tibet plat eau © 2016 Elsevier B.V. All rights reserved.

1. Introduction
Saline soils cover 2% of the total land area in China, and are widely
distributed in Sinkiang, Qinghai, Gansu, Ningxia, Inner Mongolia, and
other locales (Luo, 1980). These soils are problematic to the develop-
ment of farms, and also to the construction and engineering of roads,
channels, and airports, because of the swelling of salt crystals when
the sodium sulfate soil freezes. Studies on salt expansion in sodium sul-
fate soils in China have mainly focused on measuring the expansion ra-
tios of soils with different salt contents (Chen et al., 1989; Wang et al.,
2006; Zhang and Gao, 2010). Factors that affect salt expansion include
water content, salt content, the initial dry density, and the overlying
load (Gao et al., 1996; Gu et al., 2009). The calculation of salt expansion
and the precipitation of salt crystals in soils are usually based on the
phase diagram for the corresponding solution (Niu and Gao, 2008).
However, the calculation by Niu and Gao (2008) may have had a
lower degree of accuracy because there was little discussion of the
solution's properties in the unsteady-state, and no consideration of the
metastable phase. In addition, there was very little investigation into
the process and timing of salt crystal precipitation. When the supersat-
uration rate of the sodium sulfate solution causes crystal precipitation,
⁎ Corresponding aut hor at: State Key Laboratory Frozen Soil Engineering, Cold an d Arid
Region Environment al a nd Engineering Institute, Chinese Academy of Sciences, Lanzhou,
Gansu 7300 00 , China. Tel.: + 8 6 931 4967 288; fax: + 8 6 931 827996 8.
E-mail address: ymlai@lzb.ac.cn (Y. Lai).
salt heaving of the soil will occur, and sodium sulfate decahydrate or
heptahydrate is formed as the sodium sulfate combines with either
ten or seven water molecules, respectively (Steiger and Asumssen,
2008). Very few studies have discussed the influences of the rate of
cooling on the supersaturation rate of sodium sulfate solutions when
crystals initially precipitate at low temperatures.
Most studies concerning salt expansion from outside China have ex-
amined the supersaturation rate of solutions in brick, concrete, rock, and
other building materials, and have discussed the salt heaving force gen-
erated when different salt solutions begin to crystallize (Robert, 2002;
Steiger, 2005; Espinosa et al., 2008; Koniorczyk, 2010). However, few
studies have investigated the initiation and growth of salt crystals in
natural soils, and when soil has been involved, salt expansion has only
been studied at room temperature (Mokni et al., 2010). Salt expansion
at lower and negative temperatures remains uninvestigated.
Sulfate saline soils are widespread in the western cold region of
China. The damage to roads caused by salt expansion and frost heave re-
mains an ongoing problem. Research on sulfate saline soil in cold re-
gions faces urgent questions due to the large temperature variations
in these environments, such as: at what supersaturation rate will salt
crystallization occur in soils, how does the maximum supersaturation
rate of solutions in soil vary at different cooling rates, and what influ-
ences will salt crystallization exert on salt expansion? This paper focus-
es on the influence of different cooling rates on salt crystallization and
expansion in sodium sulfate saline soils, which can cause serious
damage to infrastructure on Qinghai–Tibet Plateau.

http://dx.doi.org/10.1016/j.coldregions.2015.12.014
0165-232X/ © 20 16 Elsevier B.V. All rights reserved.
68 Y. Lai et al. / Cold Regions Science and Technology 124 (2016) 67–76
2. Experimental design
Silty clay from the Qinghai–Tibet Plateau was sampled in the labora-
tory. The measured plastic and liquid limit for this soil was 16.4% and
27.9%, respectively (Bin and He, 2011). The sample was desalinated,
dried, and pulverized. Then, five samples with 18% gravimetric water
content and different salt contents (the ratio of the mass of salt to the
mass of water-free soil) (0%, 2.1%, 2.6%, 3.2%, 3.8%) were made and
sealed at room temperature for 24 h for uniformization. Next, the sam-
ples were placed in plastic cups and compacted in five layers to make
samples of 6.3 cm in diameter and 7.5 cm in height. The dry density
and porosity of the desalinated samples were 1.711 g/cm 3 and 0.37,
respectively.
A Stevens Hydra Probe (SDI-12) (Fig. 1a) manufactured by Stevens
Water Monitoring Systems, Inc., accurate to ± 0.0014S/m, operating
temperature: − 3 0 ~ 65 °C and measurement range: 0.01 ~ 1.5 S/m; a
thermistor with ± 0. 01 °C accuracy, and a NS-WY02 displacement
sensor (Fig. 1b) manufactured by Shanghai TM Automation Instru-
ments Co., Ltd., accurate to ± 0.001 mm and measurement range:
0.001 ~ 50 mm, were installed into the soil samples to monitor varia-
tions in conductivity, temperature, and displacement. The thermistors
and Stevens Hydra Probes were inserted into samples encased by
waterproof plastic sheets, and the displacement sensors were set to
contact the top surface of soil samples. A programmable ultra-low tem-
perature chamber (manufactured by Dongguan Terchy Test Equipment
Co., Ltd. in China, Fig. 1c) was also used to vary conditions to simulate
the temperature regime of cold regions. The data acquisition system
was a date taker CR3000 (Fig. 1d) with intervals set to 10 s, which
was manufactured by Campbell Scientific headquarters, Logan, UT,
a) Stevens Hydra Probe (SDI-12)
c) The programmable ultra-low temperature
chamber
Fig. 1. The diagr am of test sensors a n d equ ip me nt .
USA. It supports complex applications with many sensors, so the test re-
sults could be stored in CR3000. The experimental setup is shown in
Fig.2. After installing the experimental apparatus, the temperature
chamber was held at 25 °C for 30 min to eliminate measurement errors
caused by any temperature differences of the samples. The temperature
was then decreased at a rate of 0.02 °C/min and then held steady at 0 °C.
3. Crystal precipitation in sodium sulfate saline soil at
positive temperatures
3.1. Supersaturation ratio of sodium sulfate solution
The experimental data were analyzed to describe the internal
temperature variations of the soil samples (Fig. 3). One sodium sulfate
molecule and ten water molecules combine to form sodium sulfate
decahydrate when the supersaturation ratio of the solution reaches a
maximum value. When crystals began to precipitate, the temperature
increased, due to the exothermic nature of the precipitation reaction.
This caused small increases or ‘jumps’ in temperature along the cooling
curve (Wan et al., 2015), which were less pronounced at lower salt con-
tents (Fig. 3). We considered the point at which the temperature rises
as the initial temperature of salt precipitation. Therefore, the initial pre-
cipitation temperatures of different solution concentrations were deter-
mined. The experimental results showed that as salt content decreased,
the initial precipitation temperature also decreased. It is commonly the
supersaturation ratio of the solution that determines the range of meta-
stable solutions in which salt crystallization may occur. The differences
in the maximum supersaturation ratios observed when solutions with
different salt contents begin to crystallize are discussed below.
b) NS-WY02 displacement sensor
d) Date Taker CR3000
 Y. Lai et al. / Cold Regions Science and Technology 124 (2016) 67–76 69

—sodium sulfate saline soil, —plastic cup, —soil sensor, —displacement sensor, —beam(be used to fix
displacement sensor), —programmable ultra-low temperature instrument, —plastic sheet, —temperature sensor,
—closure plug, —date collection system

Fig. 2. Sketch of exp er imental apparat us.

The supersaturation ratio of the sodium sulfate solution is defined
as:
c ship between the initial concentration and the maximum value of
S¼
sat
where c is the actual concentration of salt solution and c s a t is the
saturation concentration of the solution.
The following formula can be used to calculate the solubility of
solution when the temperature is below 32.4 °C (Xu et al., 1995):
η ¼ 4:4943 1:0773T
where T is the temperature (°C), and η is the solubility, defined as the
mass of salt (g) that can be dissolved by 100 g of pure water. The
relation between the saturation molar concentration of sodium sulfate
and temperature can be obtained from Eq. (2):
csat ¼ 0:3165 1:0773T
According to the initial concentration of sodium sulfate saline soil,
Eqs. (1) and (3) can be used to calculate the maximum supersaturation
ratio of different concentrations of sodium sulfate solution. The relation-
ð1Þ
supersaturation ratio is shown in Fig. 4. It can be seen from Fig. 4
that the higher the initial concentration of the sodium sulfate saline
soil, the lower the maximum supersaturation ratio becomes and,
hence, the possibility for salt crystals to precipitate increases. The
volume of sodium sulfate decahydrate is 3.1 times that of anhydrous
sodium sulfate (Chen et al., 2006), so the salt crystallization and expan-
sion leads to soil deformation (Fig. 5). The timing of initial soil displace-
ð2Þ ments corresponded to the temperature jumps on the cooling curve,
confirming that salt crystals begin to precipitate at those moments
(Figs. 3 and 5). In the cooling experiments, it was also found that the
displacement values were larger as salt contents were higher. Fig.5
also shows that initially, the displacement of the soil samples changes
very quickly, and then decreases. This means that the rate of crystalliza-
tion is initially faster, and slows subsequently.
ð3Þ The soil conductivity is closely related to the water and ion content
in the soil. The conductivity reflects the degree of ionic activity when
the water content in soil is constant. The measured soil conductivity

F ig . 4. The relat io nship b e t w e e n t h e m a x i mu m sup er sat ur at io n rat io a n d t h e initial

Fig. 3. The cooling curve of sodium sulfate saline soil at cooling r at e of 0.02 °C/min. concentration.
70 Y. Lai et al. / Cold Regions Science and Technology 124 (2016) 67–76
Fig. 5. The displacement of sodium sulfate saline soil at cooling r at e of 0.02 °C/min.
was essentially constant before salt crystallization began (Fig. 6). How-
ever, once the salt began to precipitate, the concentration of the sodium
sulfate solution and water content in soil decreased, reflected by the de-
creasing trend in conductivity (Fig. 6). Conductivity decreased rapidly
early during salt crystallization, but the decrease was reduced as the
precipitation of salt crystals subsequently stabilized (Fig. 6).
Although the salt content remained unchanged before and after
crystallization (the salt content is defined as the proportion of the
mass of salt to the mass of dry soil), the concentration of the salt solution
varied during the whole process. It is therefore important to investigate
the variation in concentration of the solution during salt crystallization.
The relation between the solution and soil conductivity is discussed in
the following section.
3.2. Calibration test of Hydra Probe SDI-12
The soil conductivity was measured using the Hydra Probe SDI-12.
This three-parameter sensor determines the dielectric permittivity by
analyzing the signal response of radio waves propagated through the
soil. The soil electrical conductivity is calculated from a temperature-
corrected imaginary dielectric permittivity Ei using the following
formula:
SEC ¼ 2πfε 0 E i ð4Þ
Fig. 6. The conductivity of sodium sulfate saline soil at cooling r at e of 0.02 °C/min.
where SEC is the conductivity, f is the frequency (50 MHz for the Hydra
Probe), ε 0 is the vacuum dielectric constant, and Ei is the imaginary
part of the soil's dielectric permittivity.
The influence of soil salinity and water content on conductivity is
significantly larger than that of the other factors (Rhoases et al., 1999).
Relations between soil conductivity, solution concentration, and water
content were examined using regression to obtain a more precise
measurement of the solution concentration in the soil.
Four soil samples with water contents of 15%, 18%, 20% and 25% were
prepared as described in the experimental design above. The salt probe
was then completely inserted into the soil samples, and an average
conductivity value was recorded once the readings had stabilized. The
relations between soil conductivity and solution concentration for the
samples of different water content are shown in Fig. 7. The relation
between the solution concentration and conductivity can be described
by a third order polynomial function:
3
c ¼ f 3ðθÞSEC ðθÞ þ f 2ðθÞS2EC þ f 1ðθÞSEC þ fðθÞ ð5Þ
where SEC is the conductivity, and f i (θ) is a function of water content.
The calibration data show that conductivity can only be determined at
low solution concentrations when the water content is higher, because
higher concentration values are beyond the measurement range of the
instrument. The influence of water content on conductivity is reduced
when water content is less than 18%, and the variation in concentration
attenuates as the conductivity increases.
For the four samples, the polynomial functions derived from Eq. (5)
are:
f 3ðθÞ ¼ 294:57θ−21:73; f 2ðθÞ ¼ −192:93θ þ 17:747;
f 1ðθÞ ¼ 28:377θ−1:197; fðθÞ ¼ −2:4067θ þ 0:2331:
Nuclear magnetic resonance (NMR) was used to measure the varia-
tion of water contents of the soil samples under different temperatures,
because the soil water content sensor employed would have been dis-
turbed due to the high ion content of the soil. The same water contents,
salt contents, and dry densities were used as in the previous experi-
ment. The soil samples were put into a test tube and kept at the temper-
ature of 25 °C in NMR. The cooling regime was:
25 C—20 C—15 C—10 C—5 C—1 C—0 C — −1 C — −2 C
— − 3 C — −4 C —−5 C — −8 C—−10 C—−1 5 C—−20 C:
The temperature was kept constant for 3 to 4 h at each step and the
NMR signal values were transformed into water content values so that
Fig. 7. The relationship b et w e e n solution concentrat ion of sod iu m sulfate saline soil a n d
soil conductivity.
 Y. Lai et al. / Cold Regions Science and Technology 124 (2016) 67–76
the relation between water content and temperature was obtained
(Fig. 8). The water contents in the soil samples with salt contents from
2.1% to 3.8% decreased due to the precipitation of aqueous sodium sul-
fate crystals when the temperatures were above 0 °C (Fig. 8). However,
for samples with salt contents between 0% and 1%, the water contents
began to decrease at negative temperatures when water changed
phase from liquid to ice. The water content variation in sodium sulfate
saline soil could be obtained from the temperature points of crystal pre-
cipitation at the cooling rate of 0.02 °C/min.
3.3. Salt crystal precipitation
The variation in concentration of the sodium sulfate at different tem-
peratures was determined by substituting the measured values of con-
ductivity and water content into Eq. (5) (Fig. 9). The concentrations of
sodium sulfate in the soil are always greater than the saturated concen-
tration of the sodium sulfate solution at the same temperature. The
greatest variation occurs as the precipitation of salt crystals initiates,
similar to the variations observed in conductivity and water content.
The concentration change slows until it nearly reaches the curve of
the corresponding saturated concentration (Fig. 9).
It is thought that the driving force of salt crystal growth results from
the supersaturation ratio of the solution in silty clay. The rate of crystal
growth in the supersaturation solution can be described by the non-
equilibrium Freundlich isotherm (Koniorczyk, 2012), the degree
of saturation of the precipitated salt Sc can be expressed as:
¼ Sw KðM s c−AM s c sat Þp ð6Þ
dt
where K is a rate constant, p is the parameter of process, and A is a pa-
rameter for primary crystallization A ≥ S. When salt crystals are present
in the soil pores, further crystallization occurs when A ≥ 1. c refers to the
concentration of sodium sulfate solution, while c sa t represents the satu-
rated concentration of sodium sulfate solution. Ms is the molar mass of
sodium sulfate, and S w is the degree of saturation of the liquid phase.
The supersaturation ratio of sodium sulfate solution in the soil in-
creases as temperature decreases, and when it reaches A, salt
crystals rapidly begin to precipitate by the largest volume, which
will then slow down because of the decrease of the concentration
caused by the
F ig . 8. The relat io nship b e t w e e n w a t e r c o nt e nt of sodiu m sulfat e saline soil a n d
t e mp er at ure.
71
Fig. 9. The variation of solution concentration of sodiu m sulfate saline soil u nd er different
t emp er at u res at t he cooling r at e of 0.02 °C/min.
precipitation itself. The parameter of process p can be represented by
the function:
rffiffiffiffi
ffi
p¼ ð7Þ
0
where c 0 is the initial concentration of the solution, and c is the actual
concentration. When p = 0, the reaction has terminated.
It was assumed that the crystal growth interval is in accordance
with experimental date acquisition interval, which was set to 10 s,
and that the temperature in this time interval was stable. Eq. (6) can
be rearranged to yield:
p
ΔS c;i ¼ S w;i K i M s c i −A i M s c sat;i Δt i ð8Þ
where the subscript i is the order of a certain period of time (e.g. 1, 2,
3 … ), S w,i is the degree of saturation of the liquid phase in i-th period
of time, and the other notations in the equation are these defined in
Eq. (6) in i-th period of time.
In the process of temperature cooling, the crystals' growth of ten
water sodium sulfate causes the change of displacement of the soil
samples, so the change in the salt expansion rate can be described by
the following expression:
Δn ¼ Δnc −Δnl −Δ na ð9Þ
where △n refers to the change of salt expansion rate, △nc stands for the
change of porosity made by precipitated salt, △nl represents the change
of porosity generated by liquid and △na refers to the change of porosity
induced by air. The soil is deformed due to the pressure generated in the
salt crystallization process. Since the volume of the pores occupied by
the salt crystals is very small, the change in air porosity in Eq. (9) is
not considered in this paper, and we assume that the volume of sodium
sulfate solution is approximately equal to the volume of pore water. The
relation between porosity and saturation degree can be expressed by
the following formula (Zhou et al. 1998; Koniorczyk 2012):
n j ¼ S jn ð10Þ
where nj refers to porosity of the j-th phase and Sj is the degree of pore
saturation with the j-th phase(j = c, l, a-precipitated, liquid, gas). There-
fore, changes of both the saturation degree of precipitated salt (S c ) and
the supersaturation ratio of the sodium sulfate solution (S N a2S O4 ) under
different temperatures can be calculated using experimentally-obtained
values of the salt expansion ratio, water content, and concentration of
sodium sulfate solution (Fig. 10). The supersaturation ratio of the
72 Y. Lai et al. / Cold Regions Science and Technology 124 (2016) 67–76
sodium sulfate solution in the soil increases as the temperature de-
creases early in the cooling process. When the supersaturation ratio
reaches a maximum value, crystals of ten water sodium sulfate begin
to precipitate and the saturation degree of precipitated salt rises from
0. As the salt crystallization occurs, the supersaturation ratio of the
solution decreases, and the crystal precipitation and rate of increase in
saturation degree of the precipitated salt decrease (Fig. 10). Fig. 10
also shows that the higher the salt content in the soil, the earlier the su-
persaturation ratio reaches a maximum value and the earlier the salt
crystallization occurs in the soil solution. When the cooling terminates,
the saturation degree of salt crystallization reaches its maximum value.
Since the temperature is assumed to be constant in each time inter-val,
the crystal growth rate remains unchanged. However, the crystal
growth rate is considered to be influenced by the temperature and
can be calculated using Eq. (8). Fig. 11 shows the relation between the
temperature and crystal growth rate calculated from Eq. (8). The crystal
growth rate decreases as temperature decreases. There is a distinct in-
terval in which the crystal grows dramatically; the onset of this interval
is typically within 7–8 °C of the temperature at which salt crystallization
begins. The initial crystal growth rate is higher for solutions with higher
concentrations. The crystal growth rate is associated with the reaction
process, and its relation with temperature can be described by:
K ¼ AeβT þ k0 ð11Þ
where A is an empirical coefficient, dependent on the initial solution
concentration, β is related to the medium of crystal growth, and k 0 is a
correction factor.
4. The influence of the cooling rate on crystallization
Additional soil samples with the same salt (0%, 2.1%, 2.6%, 3.2%, 3.8%)
and water contents (18%) as those in the above test were formed, the
size dimension, dry density and porosity of soil samples are the same
as the above test, and the cooling rates were set to 0.1 and 1 °C/min, re-
spectively, in order to study the influence of different cooling rates on
crystallization and salt expansion. The temperature, conductivity, and
displacement of the samples were measured during the cooling period
as in the above tests. The experimental setup is also the same as the ex-
periment mentioned above (Fig. 2) and the experimental results are
shown in Figs. 12 and 13. When the salt content of sodium sulfate soil
is 3.8%, the final salt expansion ratios of soil are 1.542, 1.205 and 0.511
Fig. 10. The supersaturation ratio of sodium sulfate solution a nd sat urat ion deg ree of salt
crystallization change u nd er different t emper at ur es.
Fig. 11. The relationship b et we e n crystal growth r at e an d t emper at u re.
under the cooling rate of 0.02, 0.1 and 1 °C/min, respectively. The
displacement/salt expansion ratios of other soil samples also decrease
as the cooling rate increases. The conductivities show a little change.
For sodium sulfate soil with 3.8% salt content, the conductivities are
0.218, 0.208 and 0.204 at the cooling rate of 0.02, 0.1 and 1 °C/min,
respectively, which means that the solution concentrations increase a
little as cooling rates increase. Also, we can judge the initial crystallizing
temperature of salt by temperature jumping points on the cooling
curves, and it can be found that the initial crystallizing temperature
decreases as the cooling rate increases. Therefore, the cooling rate not
only affects the maximum amount of salt expansion, but also the initial
supersaturation ratio of solution, which was calculated at three different
cooling rates of 0.02, 0.1 and 1 °C/min, respectively (Fig. 14). The effect
of the cooling rate on crystallization occurs because with the cooling
rate slowing, the supersaturation ratio of the solution decreases at the
same temperature, and crystallization then occurs more readily. In
addition, the supersaturation ratio of solution will increase as the initial
temperature of crystallization decreases, implying that the crystalliza-
tion of a supersaturation solution at low temperature is more difficult
than that at higher temperature under the same cooling rate.
The linear relation between the supersaturation ratio of crystalliza-
tion and the initial temperature of crystallization under different cooling
rates can be represented as:
S0 ¼ λT 0 þ χ ð12Þ
Where
χ ¼ 0:0349 ln þ 1:927
dt
where S 0 is the crystallization supersaturation ratio, T 0 represents the
initial temperature of crystallization (°C), λ is a constant related to soil
properties, which can be founded from Fig. 13 that the value of λ is
around 0.03, and χ is a function of the cooling rate dT/dt.
Because the salt expansion rate is related to the saturation degree of
salt crystals (S c ), and the crystal growth rate is a function of tempera-
ture, the crystal growth rates under different cooling regimes are equiv-
alent at the same temperature. Eq. (8) indicates that in the same time
interval, the crystal growth rate will decrease as the cooling rate in-
creases, causing a decrease in the saturation degree of salt crystals. As
a result, the salt expansion rate decreases. In addition, the decrease of
supersaturation ratio will make the salt crystals precipitate more easily
to approach the saturation of solution. This means that as the cooling
 Y. Lai et al. / Cold Regions Science and Technology 124 (2016) 67–76
a) Temperature
c) Conductivity
Fig. 12. The variations of t emperat ur e, displacement a nd conductivity of sodiu m sulfate saline soil versus t ime at t h e cooling r at e of 1 °C/min.
rate decreases, salt crystal growth changes from an unsteady state to a
more stable state (in Fig. 15).
From Fig. 15, it can be found that the metastable interval of sodium
sulfate solution increases as the cooling rate increases, when the cooling
rate is very small, the phase change curve of sodium sulfate solution will
be close to its saturation curve. The increase of metastable interval raises
the supersaturation ratio of the solution, and it is more difficult for salt
crystals to precipitate.
According to Eq. (12), the absolute value of λ increases as the rate of
cooling increases and the influence of different rates of cooling on crys-
tallization at low temperatures is apparently strengthened. The degree
of supercooling depends on the rate of cooling, implying that greater
supercooling limits salt crystallization. Although crystallization at
positive temperatures has been previously investigated and discussed
in this paper, it remains uninvestigated at negative temperatures.
Therefore, an experiment on low-concentration sodium sulfate saline
soil was performed at negative temperatures (note that when the
temperature decreases to 0 °C, the concentration of soil solution with
high salt content will decrease).
5. The crystallization of sodium sulfate saline soil at
negative temperature
Four soil samples were used in this experiment. Their water contents
were all 18%, and their salt contents were 0%, 0.5%, 1% and 1.5%, respec-
tively. Their height and dry density are the same as those in the
73
b) Displacement
experiment above. The experiment was held constant at 20 °C for half
an hour, then decreased to − 2 0 °C at a cooling rate of 0.02 °C/min,
and all the samples were kept under this temperature. The changes in
temperature, conductivity, and displacement of the samples are
shown in Fig. 16.
In Fig. 16a, region A is considered to be the region of initial phase
change between water and ice, where substantial heat is released, caus-
ing a perturbation in the temperature curve. The shape of temperature
curve for the sample with 0.5% salt content is similar to the curve of
the sample with no salt content, implying that there is no salt crystalli-
zation in the sample with the lowest salt concentration, or that the heat
generated by any salt crystallization is too limited to cause a noticeable
change to the temperature curve. For the sample with 1.0% salt content,
after the phase transition occurs, salt crystallization was apparent at
− 3 . 7 °C in region B, and the heat released from salt crystallization
caused the temperature to stabilize for a short time. However, the
phase transition to ice only happens after salt crystallization at 0.05 °C
(region C) for the sample with 1.5% salt content.
The changes in conductivity shown in Fig. 16b also support the inter-
pretation of the above observation. Free water freezes first when the
phase transition from water to ice begins in the sample with no salt con
-tent, leading to a rapid decrease in conductivity with a steep
slope. When the free water has frozen, additional freezing of adsorbed
water occurs more slowly, due to the constraint of the free energy of
soil par-ticle surface toward the bound water and the slope of the
conductivity trace begins to decrease. When most of the bound
water has been
74 Y. Lai et al. / Cold Regions Science and Technology 124 (2016) 67–76

a) Temperature b)Displacement

c) Conductivity
Fig. 13. The variations of t emper at ure, displacement a nd conductivity of sodiu m sulfate saline soil versus t i me at t he cooling r at e of 0.1 °C/min.

frozen and there is cooling in the soil without latent heat, the conductiv-
ity decreases exponentially. The change in conductivity of the sample
with 0.5% salt content was similar to that of the soil with no salt content.
For the sample with 1% salt content, the conductivity decreased when
freezing initiated. However, due to the phase change from water to
ice, the concentration of sodium sulfate solution increases gradually to
the extent that salt precipitation begins to occur, during which an
amount of water will be carried out, and as a result, the conductivity
slope of P 2B (through analyzing the temperature curve, P2 B is consid-
ered to be the conductivity when the salt crystallization and phase
change between water and ice are co-existing) remained basically the
same as prior to the phase change. The precipitation of crystals occurred
at 0.05 °C in the sample with 1.5% salt content. This was reflected by the
range of P1C on the conductivity curve. In this process, water molecules

F ig . 14. The relati onship be t w e e n ma x i m u m sup er sat u r at io n rat io a n d t h e initial

precipitated t e mp er at ur e u nd er different cooling rates. Fig. 15. The effect of cooling r at e on t he solubility diagram of sodium sulfate.
 Y. Lai et al. / Cold Regions Science and Technology 124 (2016) 67–76
a) Temperature
c) Displacement
Fig. 16. The changes of t e mp er at ure, displacement a nd conductivity of sodium sulfate saline soil versus t ime at cooling r at e of 0.02 °C/min.
are carried by sodium sulfate (due to the decreases of both the concen-
tration of sodium sulfate and water content), leading to a rapid decrease
in conductivity. In the post rang of P1C, the variation in conductivity is
mainly caused by the phase change between water and ice in the soil
and a small amount of salt crystals precipitate, which causes a relative
decrease in the slope of the conductivity curve.
It can be seen from Fig. 16c that, as crystallization began in the
sample with 1.5% salt content, the displacement of the soil began to
increase and the air volume decreased. During the phase change, the
soil porosity decreases as water is transformed into ice crystals, causing
the displacement to decrease. In the stage of bound water freezing, the
displacement increases under the forces of salt expansion and frost
heave. Finally, the displacement stabilizes once most of the adsorbed
water has frozen.
The precipitation of salt from the unsaturated sodium sulfate saline
soil below the freezing temperature did not cause salt expansion
when the salt content of the soil was low (0.5%–1.0%), and salt crystal-
lization only occurred as adsorbed water froze (Fig. 16c). In this case,
the temperature at which initial salt crystallization occurs decreased
as the salt content of soil decreased. Since the temperature point of
initial salt crystallization can be observed on the cooling curve, addition-
al soil samples with different salt contents were prepared for another
temperature cooling experiment to illustrate this. The salt contents
were between 0.5% and 1.4% and the cooling rate was 1 °C/min.
The cooling bath with a temperature accuracy of ± 0.0 1 °C was
used to cool samples. Thermistors were used to measure internal
75
b) Conductivity
temperatures of soil samples. The data acquisition system was also
CR3000 with intervals set to 10 s. The thermistors were inserted into
samples encased by waterproof plastic sheets. The samples were put
into the cooling bath at a constant temperature of 25 °C for half an
hour, then the temperature was decreased to − 2 0 °C and the samples
were kept at this temperature for 2 h. The experimental results are
shown in Fig. 17.
Salt precipitation occurred in the samples with salt contents be-
tween 0.8% and 1.4% at low temperature. The samples with greater
salt content had higher temperatures of crystal initiation (Fig. 17). Salt
crystal precipitation occurred at the same time as the phase change
from water to ice in the sample with 1.4% salt content. The dashed
line in Fig. 17 represents the initial salt crystallization at low tempera-
tures for the samples. But for sodium sulfate saline soils with the salt
contents from 0.5% to 0.7%, no obvious salt crystallizations appear at
low temperature because the heat released from crystallization is less
than that from phase change between water and ice, which is the reason
that there are not obvious jumps or changes of temperature on the
cooling curve.
For saline soil with low salt content, if salt crystals want to
generate, the supersaturation ratio must reach the maximum at
corresponding temperature. The ice–water phase change continuously
occurs to increase the concentration of solution, and the salt crystals
generate. And the supersaturation ratio increases as salt content of soil
decreases, so it is difficult for saline soil with low salt content to crystal-
lize, especially for saline soil with salt content below 0.7%.
76 Y. Lai et al. / Cold Regions Science and Technology 124 (2016) 67–76
Fig. 17. t he cooling curve of sodium sulfate saline soils wit h low salt contents.
6. Conclusions
The following conclusions were drawn from the above experimental
results and analyses:
(1) The precipitation maximum supersaturation ratios are different
for sodium sulfate saline soils with different initial concentra-
tions. In general, the maximum supersaturation ratio increases,
and it is difficult for ten water sodium sulfate crystals to precipi-
tate as the initial concentration is decreasing.
(2) The amount of initial salt crystallization and expansion are
influenced by the cooling rate. When the cooling rate is lower,
the supersaturation ratio of the sodium sulfate solution is
reduced, and crystals precipitate more readily. This causes the
saturation degree of precipitated salt to increase, intensifying
the salt expansion. When the sodium sulfate solution reaches
the crystallization supersaturation ratio, the supersaturation
ratio begins to decrease toward a value of 1. Crystallization
ceases when the supersaturation ratio of the solution is less
than or equal to 1.
(3) As the temperature decreased, the crystal growth rate in the soil
decreased exponentially. For the same salt content, decreasing
the cooling rate resulted in greater growth of salt crystals and
an increase in salt expansion of the soil. Crystallization and salt
expansion did not occur before frost heave in sodium sulfate
saline soil when the salt content was b1.5%.
(4) Salt crystallization only occurred after the initial frost heaving
for soils with low concentrations of sodium sulfate, and under
the same cooling rate, the initial precipitation temperature
decreased as the salt content decreased. The salt crystallization
reduced to a very weak point when the salt content was less
than 0.7%. For the sodium sulfate saline soil of low salt content,
the salt crystals precipitated at a temperature below the freezing
point will fill the pores of the unsaturated soil, making it more
compacted.
Created by the trial version of PDF Focus .Net 6.9.7.6!
The trial version sometimes inserts "trial" into random places.
Get the full version of PDF Focus .Net.
This research did not consider the effect of salt migration during the
cooling and crystallization process. Salt migration may cause local salt
accumulation in the soil, and could affect crystallization and the
resulting soil expansion, so further studies on sodium sulfate saline
soil are suggested to account for these potential effects.
Acknowledgments
This research was supported by the National Natural Science
Foundation of China (41230630), the Western Project Program of the
Chinese Academy of Sciences (KZCX2-XB3-19), the National Key Basic
Research Program of China (973 Program No. 2012CB026102), and
the foundation of the State Key Laboratory of Frozen Soil Engineering
(SKLFSE-ZY-03).
References
Bin, H., He, P., 2011. Experiment al st udy of w a t e r a nd salt redistribut ions of saline soil
wit h different freezing modes. Rock Soil Mech. 3 2 ( 8), 230 7– 23 12 .
Chen, X.B., Qiu, G.Q., Wang, Y.Q., et al., 1989. Salt redistribution a n d heave of saline soil
during cooling. J. Glaciol. Geocryol. 1 1 ( 3), 231 – 238 .
Chen, X.B., Liu, J.K., Liu, H.X., et al., 2006. Frost Action of Soil a nd Foundation Engineering.
Science Press, Beijing.
Espinosa, R.M., Franke, L., Deckelmann, G., 2008. Phase changes of salts in porous ma t e -
rials: crystallization, hy drat i o n a n d deliq uescence. Constr. Build. Mat er . 22,
17 5 8–1 77 3.
Gao, J.P., Wu, J.H., Deng, Y.S., et al., 1996 . A multi-factor study of salt expansion of sulfate
salty soil. J. Glaciol. Geocryol. 1 8 ( 2) , 17 0 –1 77.
Gu, Q.K., Wu, A.H., Li, N., 2009. The exp er ime nt study of salt heaving behaviors of sulfate
saline soil. J. Xi'an Univ. Technol. 2 5 ( 3), 283 –287.
Koniorczyk, M., 2010. Modelling the phase change of salt dissolved in pore water —equilibrium
and non-equilibrium approach. Constr. Build. Mater. 24, 1119–1128.
Koniorczyk, M., 2012. Salt transport an d crystallization in non-isothermal, partially satu-
r at ed porous materials considering ions interaction model. Int. J. Heat Mass Transf.
55, 655 – 6 79 .
Luo, W.F., 1 980. Highway Engineering in Saline Soil Regions. China Co m mu nicat io n s
Press, Beijing.
Mokni, N., Olivella, S., Alonso, E.E., 20 10. Swelling in clayey soils ind uc ed by t h e presence
of salt crystals. Appl. Clay Sci. 47, 1 05– 1 12.
Niu, X.R., Gao, J.P., 2008. Deduction of salt expansion expression during p ur e salt expan-
sion period of sulfate saline soil. Chin. J. Geotech. Eng. 3 0 ( 7), 105 8 –1 06 1.
Rhoases J. D., Chanduv I. F., Lesch S., Soil salinity assessment. FAO Imigation a nd Drainage
Papers, 1999, 57, 3–7.
Robert, J.F., 2005. Salt d amage in porous materials: h ow high supersat urat ions ar e gen er -
at ed. J. Cryst. Growth 282, 4 35– 469 .
Steiger, M., 2005. Crystal gr owt h in porous materials—I: t he crystallization pressure of
large crystals. J. Cryst. Growth 282, 455–469.
Steiger, M., Asumssen, S., 20 0 8. Crystallization of so diu m sulfate p has es in porou s
materials: t he p hase diagr am Na2SO4–H2O a nd t he generat ion of stress. Geochim.
Cosmochim. Acta 72, 4 291–4 30 6 .
Wan, X.S., Lai, Y.M., Wang, C., 2015. Experime ntal study on t h e freezing t e mp er at u re s of
saline silty soils. Permafr. Periglac. Process. http://dx.doi.org/10. 10 02/ pp p. 183 7.
Wang, J.C., Li, J.S., Wang, C.M., 20 0 6. A study on single-cycle salt a nd frost heavin g of
sulfate saline soil. J. Jilin Univ. 3 6 ( 3), 410 –41 6.
Xu, X., Wang, J., Zhang, L., et al., 1995. Mechanisms of Frost Heave a nd Soil Expansion of
Soils. Science Press, Beijing.
Zhang, J.Y., Gao, J.P., 2010 . Analysis of salt heaving process in sulphuric acid salinized soil.
Technol. Highway Transport. 6, 8–11.
Zhou, Y., Rajapakse, R.K.N.D., ASCE, M., et al., 1998. Coupled heat–mo isture–air t ransfer in
deformable u nsat urat ed media. J. Eng. Mech. 124, 1 09 0–1099.