Analysis and discussion of different methods of artificial ice-high specimen preparation

ShuJuan Zhang ,Wei Ma,ZhiZhong Sun,HaiMin Du

State Key Laboratory of Frozen Soil Engineering,Cold and Arid Regions Environmental and Engineering Research Institute,Chinese Academy of Sciences,Lanzhou,Gansu 730000,China

1 Introduction

Frozen soil refers to soil or rock containing ice with a temperature below 0 °C (Tsytovich,1963).It is a type of inhomogeneous,multi-phase granular material which consists of solid mineral grains,ideal plastic inclusions of ice (cement ice and ice layer),unfrozen water (membrane bound water and liquid water),and gaseous components (air and vapor).Frozen soil is roughly divided by volumetric ice content into ice-poor (0%–20%),intermediate (20%–60%),ice-rich (60%–85%),and dirty ice (80%–90%),with overlaps existing (Arenson and Springman,2005;Arensonet al.,2007).In this paper,ice-high frozen soil is specifically defined as frozen soil with volumetric ice content above 20%.

The ice in frozen soil is mainly divided into pore ice,layered or segregated ice,and reticular ice,and the corresponding cryostructure types of frozen soil are called massive,lamellar,and reticular,respectively (Jumikis,1979).The first two types are most typical for ice-high frozen soil (Cheng and Ma,2006).Previous research has reported that mechanical properties of ice-poor frozen soil are greatly affected by such external factors as temperatures,loading velocities,stress paths,stress-strain histories,stress levels,and so forth,and they are controlled by such internal factors as minerals,grain gradations,water (ice) contents,salinities,and organic contents (Tinget al.,1983;Bing and Ma,2011;Laiet al.,2013).Cryostructure also has an obvious impact on the strength and deformation of frozen soil (Pekarckaya,1963;Raddet al.,1979;Jumikis,1979;Li,2004;Matthew,2013),so all of these factors should be the subject of more attention in future research of ice-high frozen soil (Zhanget al.,2012).

Ice-high foundation soil is widely distributed in permafrost regions,and whether or not it can be suitable for foundations of buildings and permanent constructions is of concern to frozen-soil researchers (Wu,1990;Andersland and Ladanyi,1994).For satisfying the requirements of engineering development,some studies on ice-high frozen soil have been carried out in China and other countries,so the method of preparing ice-high specimens becomes extremely important.Currently,there is no standard in preparing specimens withhice-high content;all the existing methods are different.In this paper their differences are analyzed and discussed,and the influence of the specimen preparation method on testing results is also analyzed according to the strength characteristics and another testing data.

2 Analysis of existing research results

The reported specimen preparation methods can be roughly divided into two categories:in the first one,considerable crushed ice or snow is used and the whole course is conducted under the condition of negative temperature;in the second one,an intimate mixture of soil and liquid water is directly consolidated into a mold by vibration or compaction in a negative room temperature.The first method is more widely used.According to the research findings,the change characteristic of frozen-soil strength with changing ice or water content is not unique.Generally,similar results were obtained under similar methods.

2.1 The effects of increasing total water content on frozen-soil strength

Shusherinaet al.(1969) investigated the effects of total water content on three frozen soils under short-term loading at temperatures between-10 °C and-55 °C.They thought that under the condition of complete saturation,the dependence of uniaxial compression strength of frozen soils on their water content is generally the same,irrespective of the grain size composition.The general form of this dependence is:When the total water content gradually increases from the fully saturated moisture state which corresponds to the minimum porosity (the maximum density of the skeleton) of frozen soil,the compression strength initially decreases until it reaches a minimum value that is below the ice strength,and then remains constant,independent of the water content;with further increases in the water content,the compression strength begins increasing until it matches the ice strength.

This general form depends on the temperature,the composition of the soil skeleton,and the saturation degree of the sample.This sample preparation method was reported in the early literatures in Russian,so the actual details cannot be found,but crushed ice and snow were used to ensure high water or ice content and uniformity of the specimens.

2.2 Increasing volumetric ice content results in a reduction in the frozen-soil strength

According to testing data from oversaturated fine sand,Bake (1976a) found that as the volumetric ice content increased,the uniaxial compression strength always decreased until the soil particles no longer influenced it,which was similar to the conclusions obtained by Goughnour and Andersland (1968) and Nickling and Bennett (1984).Bake used specimens that were prepared by the rodding compaction technique and vacuum method of saturation (Bake,1976b),but it is found that this method may not be suitable for specimens with water content above 30% in mass.In the specimen preparation method of Goughnour and Andersland (1968),sand particles were first added to the ice matrix when the ice was freezing in layers,and then the frozen ice was crushed to <5.0 mm,and some more sand particles were added to the crushed ice.After the mixture of crushed ice and sand particles was put into a special mold,saturation was induced from the bottom of the mold without vacuum.In this method,the amount of water entering during the saturation could affect the cementation of the sand particles with the crushed ice,which could easily result in a non-uniform texture in the specimen.

Nickling and Bennett (1984) described the following steps during each specimen preparation:Ignoring the unfrozen water content and considering the voids to be totally filled with ice,the volume of ice required at the testing ice content was calculated according to the volume of the mold and the densities of the mineral grains and ice,and approximately 80% of the required volume of water for ice in the specimen was frozen (-10.0 °C) and thoroughly crushed to a fine powder.The finely crushed ice was then mixed with a predetermined volume of rock-glacier material(-10.0 °C) and packed into the pre-chilled mold.The remaining volume of distilled water (0.5 °C) required for ice was added uniformly to the specimen surface and was expected to enter the specimen by capillarity being held within the specimen fines and the pulverized ice.It was also pointed out that part of the ice-soil mixture was reduced to slurry in most cases,which in turn could change the texture of the specimen.

2.3 Strength of frozen soil shows a complicated trend with increasing volumetric ice content

Sayles and Carbee (1981) and Arensonet al.(2005)observed the uniaxial compressive strength of saturated silt and the triaxial compressive strength of rock glacier at about-2.0 °C,respectively.They found that the uniaxial strength of silt always increased when the volumetric ice content changed from 30% to 60%,while the triaxial strength of rock glacier reached its maximum value at a volumetric ice content of about 80% when the volumetric ice content was in the range of 20%–100%.The results reported by Arensonet al.(2005) were rather discrete.

The specimen preparation method used by Sayles and Carbee (1981) was:Some specimens with dry unit weights <1.2 g/cm3were consolidated by placing air-dried soil in the molds and then vibrating them to the desired predetermined volume;specimens with high dry unit weights were compacted in layers with a spring-loaded hammer.Finally,the specimens were saturated from the bottom of the molds under vacuum and then frozen from the top downward in an open system.Obviously,this preparation method was only appropriate for the specimens in which the mixture of soil particles and water was thick enough to ensure non-separation of water and soil before being totally frozen,which was difficult for the specimens in which the mass ratio of water to soil was >0.5.

The method of Arensonet al.(2005) generally followed these steps:(1) A crushed ice–solid particle mixture was pre-cooled to the same temperature as the ice;(2) the mixture was placed in a mold and compacted if necessary;and (3) the crushed ice–soil mixture was saturated from the bottom to top by de-aired water at 0 °C in an open freezer for about 30 min.During the saturation,the density degree of the crushed ice–soil particle mixture could affect the entering volume of de-aired water,so in the specimen the ratio of crushed ice to refrozen ice could change with different degrees of compactness.This can have an important influence on the mechanical properties of the frozen specimen,as will be discussed later.

In the method of Sayles and Carbee (1981),the specimens were prepared by adding water to a predetermined volume of dry soil,but this was difficult for the oversaturated specimens (when the volumetric ice content reached above 30%,the total water content far exceeded the saturated value).In contrast,in the method of Arensonet al.(2005),crushed ice was used for all the specimens and the saturation process was conducted,so it was possible that there were two kinds of ice in the frozen samples,and the compactness of the crushed ice–solid particle mixture would control the entering volume of the liquid water.

2.4 Strength characteristic of frozen soil as affected by temperature range and total water content

Maet al.(2008) and Laiet al.(2008) prepared specimens in a similar way:In a negative surrounding(about-7.0 °C),predetermined weights of both crushed ice and dry soil were mixed together,and then distilled water at 0 °C was uniformly added to the mixture,and the specimens were quickly compacted in layers (Maet al.,2008) or by two ends (Laiet al.,2008)to the required size.In their studies,the uniaxial compressive strength characteristics of frozen clay were similar to the triaxial strength of frozen sand at different temperatures.When the total water content increased from 30%,at the temperature of-5.0 °C to-6.0 °C,the strengths initially decreased and then gradually became constant with further increases in the water content;at the temperature of-1.0 °C the strengths were at their minimum values,which was consistent with the conclusion of Arensonet al.(2004).

However,it is an interesting phenomenon that the uniaxial strength of clay always exceeds or approximates the triaxial strength of sand under the same temperatures and strain rates,which is in conflict with the description of Wuet al.(2002).What causes this?A close analysis of the methods of Maet al.(2008) and Laiet al.(2008) shows that the difference is in the size of the crushed ice and the amount of liquid water adopted:Maet al.(2008) used ice <5.0 mm and about 30% of the total water content,while Laiet al.(2008)used <2.0 mm ice and 10% of the total water content.Following,these two factors will be discussed in detail.

3 Influence of crushed-ice size and liquid water content on frozen soil strength

In order to analyze the influence of crushed ice size and liquid water content on the strength of frozen soils,we conducted a series of uniaxial compressive tests,and some typical results are shown in figure 1.When the size of the crushed ice was <2.0 mm and the corresponding liquid water contents were 15% and 10%,respectively,the stress-strain curves show strain-hardening behavior,and the stresses on curve 1 exceeded the corresponding ones on curve 2 except for the initial stage.When the size of the crushed ice was<5.0 mm,the stress-strain curve shows strain-softening behavior,as shown with curve 3.Thus,the strength and stress-strain behavior of frozen soil were different when the size of the crushed ice and the amount of liquid water were different.

4 Influence of crushed ice on frozen soil strength and deformation mechanism

To thoroughly assess the influence of crushed ice on frozen soil strength and deformation mechanism,we used silty sand and performed the following tests:

1) The particle size of the silty sand was <2.0 mm.To ensure a uniform specimen texture,the size of the crushed ice was the same as the silty sand.The preparation method was similar to Laiet al.(2008).The total water content of the specimen was calculated by equation(1):

wherewtotalrepresents the total water content of the specimen,andmi,mlwandmsrepresent the weights of crushed ice,liquid water and dry soil,respectively.

2) During the specimen preparation,the temperature of the cryogenics lab was about-7.0 °C to prevent the crushed ice from melting and the ice-soil-water mixture being frozen in a short time.The weight of liquid water (mlw) was determined according to the weight of unfrozen water (mu) at-7.0 °C (shown in figure 2),but in the course of preparation specimens,the value ofmlwwas intentionally increased to twice as large as that ofmuin order to ensure that the mixture of ice-soil-water was moist enough.

Figure 2 shows the relationships between unfrozen water content and temperature at different total water contents.For the frozen soil specimens containing crushed ice,the amount of crushed ice increased as the total water content increased,so the unfrozen water content (wu) was absorbed by both crushed ice and soil particles.The results of figure 2 are explained by equation(2):

In preparing a specimen,the weight of crushed ice far exceeded the amount of 0 °C water,so the reproduced ice was considered as the crushed one in equation(2)after the specimen was frozen.

For the above specimens of determined volumes,even though the total water content increased and the corresponding dry density decreased,the liquid water added to the ice-soil mixture per unit mass was constant.The typical results from our uniaxial compressive tests are presented in figures 3–5.

Figure 3 indicates that the stress-strain curves changed from strain-softening to strain-hardening as the total water contents increased.Obviously,when the total water content was close to pure ice,the stress-strain curve still showed strain-hardening characteristics.

Figure 4 shows that the uniaxial compression strength of silty sand initially increased and then tended to be stable when the total water content increased from 30% to 316%.After the tests were done,the low-water-content specimens were drum-shaped with cracks (Figure 5a),which was similar to the deformation of dense sand (Zhaoet al.,2004);the high-water-content samples tended to be evenly compacted (Figures 5b and 5c),which was different from the conclusions by Zhang (2009).

A cross section of the specimen in figure 5a is shown in figure 6a.Figure 6b is a cross section of another specimen prepared by directly adding the total water to dry soil and then freezing it.Clearly,the texture in figure 6b is quite dense whereas in figure 6a it is rather loose.In figure 6b,the soil particles are fully mixed with liquid water before frozen,and in the freezing course almost all the water are frozen into ice,so the soil particles are cemented with ice together very well;while in figure 6a,only 4% of the total water content participants in the supply of cementing force,and the crushed ice itself also needs to extra water to be cemented together in freezing course.So the effective water content that could produce cementing force in freezing course is much larger in figure 6b than in figure 6a,which mainly results in the difference of the texture.

Figure 1 Stress-strain behavior of frozen sand with the total water content of 60% and strain rate of 6.67×10-4 s-1 at-1.0 °C.(Liquid water content and ice particle size:①15.0%and <2.0 mm;②10.0% and <2.0 mm;③15.0% and <5.0 mm)

Figure 2 Relationship between unfrozen water contents and temperature of frozen soils.(Total water content and corresponding dry densities:I:557.1% and 0.15 g/cm3;II:150.0%and 0.5 g/cm3;III:50.0% and 1.07 g/cm3;IV:15.0% and 1.88 g/cm3;V:ice (particle size <2.0 mm) and 0.89 g/cm3 )

Figure 3 Stress-strain curves with different water contents

Figure 4 Relationships between uniaxial compaction strengths and total water contents

Figure 5 Deformation of silty sand samples after uniaxial compression strength testing.

According to figures 3–5,the crushed ice had hardly any cementing force in the freezing course and was more like a special kind of soil composition,and it could only be cemented together by using extra liquid water.Thus,when a liquid water was added into the ice-soil mixture and the volume of the sample was constant,the mass ratio of the dry soil to the crushed ice decreased with the increasing total water content.This means that the soil particles were more powerful than crushed ice in absorbing water,and ultimately the amount of refrozen water in the ice-soil mixture per unit mass increased as the total water content increased,which caused the results shown in figures 3 and 4.This demonstrates that in the artificially prepared specimens the role of crushed ice was different from that of ice formed by the natural way.

In preparing ice-high or water-high specimens,the use of crushed ice can ensure the ultimate required ice or water content,but the amount of crushed ice will affect the strength value and alter the deformation mechanism.This means the influence of crushed ice on the mechanical properties of ice-high specimens is different from that of ice formed naturally.

Figure 6 Cross sections of specimens with 30% total water content under different preparation methods.(a) 96% of the total water content was directly replaced with crushed ice during sample preparation;(b) the required water was directly added to soil particles during sample preparation

5 Conclusions

By analyzing and comparing several methods of preparing ice-high specimens,we found that using crushed ice can ensure the required ice or water content when a specimen is prepared,but the following problems will be introduced:

? In the freezing course,the crushed ice particles cannot be totally cemented without adding more liquid water,that is,the crushed ice itself has no cementing force;

? When mixing crushed ice,dry soil,and liquid water to prepare a specimen,the dry soil particles cannot be fully wetted in a short time,which in turn will affect the texture of the specimen.

? Crushed ice may increase the strength and change the deformation mechanism of a frozen specimen.

It is therefore strongly recommended that in the course of preparing an ice-high specimen,the ice should not be replaced with crushed ice and it should be frozen in the natural way.

This work is financially supported by the Excellent National Key Laboratory Special Fund of China (No.41023003),the National Natural Science Foundation of China (No.41101068),the National Key Basic Research Program of China (973 Program) (No.2012CB026102),and the project of the State Key Laboratory of Frozen Soil Engineering (No.SKLFSE-ZT-07).

Andersland OB,Ladanyi BL,1994.An Introduction to Frozen Ground Engineering.Chapman &Hall,New York,USA.

Arenson LU,Johansen MM,Springman SM,2004.Effects of volumetric ice content and strain rate on shear strength under triaxial conditions for frozen soil samples.Permafrost and Periglacial Processes,15(3):261–271.

Arenson LU,Springman SM,2005.Triaxial constant stress and constant strain rate tests on ice-rich permafrost samples.Canadian Geotechnical Journal,42(2):412–430.

Arenson LU,Springman SM,Sego DC,2007.The rheology of frozen soils.Applied Rheology,17(1):1–14.

Baker THW,1976a.Compressive Strength of Some Frozen Soils.M.Sc.Thesis,Dept.of Civil Engineering,Queen’s University,Kingston,Ontario.

Baker THW,1976b.Preparation of artificially frozen sand specimens.N.R.C.C.15349,National Research Council of Canada,Division of Building Resources,pp.1–5.

Bing H,Ma W,2011.Laboratory investigation of the freezing point of saline soil.Cold Regions Science and Technology,67(1–2):79–88.

Cheng GD,Ma W,2006.Frozen soil engineering problems in construction of the Qinghai-Tibet Railway.Chinese Journal of Nature,28(6):315–320.

Goughnour RR,Andersland OB,1968.Mechanical properties of a sand-ice system.ASCE Journal of Soil Mechanics,94(SM4):923-950.

Jumikis AR,1979.Cryogenic texture and strength aspects of artificially frozen soils.Engineering Geology,13(1–4):125–135.

Lai YM,Xu XT,Dong YH,et al.,2013.Present situation and prospect of mechanical research on frozen soils in China.Cold Regions Science and Technology,87:6–18.

Lai YM,Zhang Y,Zhang SJ,et al.,2008.Experimental study of strength of frozen sandy soil under different water contents and temperatures.Chinese Journal of Rock and Soil Mechanics,30(12):3665–3670.

Li HP,2004.Effect of Texture and Structure of Frozen Soil on Its Strength and Deformation Behavior.Ph.D.Dissertation,Chinese Academy of Science,Lanzhou,China.

Matthew TB,2013.Secondary creep approximations of ice-rich soils and ice using transient relaxation tests.Cold Regions Science and Technology,88:17–36.

Ma XJ,Zhang JM,Chang XX,et al.,2008.Experimental research on strength of warm and ice-rich frozen clays.Chinese Journal of Rock and Soil Mechanics,29(9):2498–2502.

Nicking WG,Bennett L,1984.The shear strength characteristics of frozen coarse granular debris.Journal of Glaciology,30(106):348–357.

Pekarskaya NK,1963.Shear Strength of Frozen Ground and Its Dependence on Texture.Izd-Vo Akad Nauk,SSSR,Moscow,U.S.Army Cold Regions Research and Engineering Laboratory,Draft Translation,115:60-77.

Radd FJ,Wolf LH,1979.Ice lens structures,compression strength and creep behavior of some synthetic frozen silty soils.Engineering Geology,13(1–4):169–183.

Sayles HF,Carbee DL,1981.Strength of frozen silt as a function of ice content and dry unit weight.Engineering Geology,18(1–4):55–66.

Shusherina EP,Bobkov YP,1918.Effect of Moisture Content on Frozen Ground Strength.Technical Translation,National Research Council Canada,pp.1–19.

Ting JM,Martin RT,Ladd CC,1983.Mechanisms of strength for frozen sand.ASCE Journal of Geotechnical Engineering,109(10):1286–1302.

Tsytovich HA,1985.Frozen Soils Mechanics (transl.by Zhang CQ,Zhu YL).Science Press,Beijing,China,pp.19.

Wu ZW,1990.Thermal heave properties of ice-rich frozen soil.Chinese Journal of Glaciology and Geocryology,12(3):1587–1592.

Wu ZW,Ma W,2002.Strength and Creep of Frozen Soils.Lanzhou University Press,Lanzhou,China,pp.54–64.

Zhao CG,Bai B,Wang YX,2004.Theoretical Soil Mechanics.Tsinghua University Press,Beijing,China,pp.202.

Zhang LM,Li ZJ,Jia Q,2009.Experimental study on uniaxial compressive strengths of artificial fresh ice.Chinese Journal of Hydraulic Engineering,11(40):1392–1396.

Zhang SJ,Sun ZZ,Du HM,2012.A preliminary analysis of main factors affecting stress-strain behaviors of frozen soil with high-water content.Applied Mechanics and Materials,204–208(C):128–134.