Experimental study of the shrinkage behavior of cemented paste backfill

Jiahao Qin,Jian Zheng,Li Li

Research Institute on Mines and the Environment (RIME UQAT-Polytechnique),Department of Civil,Geological and Mining Engineering,École Polytechnique de Montréal,Montréal,Québec,H3C 3A7,Canada

Keywords:Cemented paste backfill (CPB)Shrinkage Evaporation Underground mine stopes Volume variation

ABSTRACT Cemented paste backfill (CPB) is largely used in underground mine stopes worldwide.When a CPB is placed in a stope,an important task is to estimate the settlement associated with the shrinkage and selfweight consolidation of the CPB.This is closely related to the volume management to ensure the stability of barricades and tight contacts between the backfill and stope roof.Over the years,shrinkage studies were mostly performed on fine-grained soils (silts and clays),with only a few publications on the shrinkage behavior of uncemented tailings.No study has been published on the shrinkage behavior of CPB.To fill this gap,a series of shrinkage tests has been conducted on CPB with different cement contents,including zero cement content (uncemented paste backfill,uCPB).The results show that the shrinkage response of CPB is very different from that of uCPB.At a given initial water content,CPB exhibits a shorter normal shrinkage stage than uCPB.The unsaturation onset water content and void ratio,shrinkage limit and final void ratio of CPB are generally higher than those of uCPB.At a given cement content,the shrinkage behaviors of CPB and uCPB are significantly influenced by the initial water content.

1.Introduction

Cemented paste backfill (CPB) is made of full (all sizes of particles) tailings,water and binder (e.g.cement,fly ash,ordinary Portland cement) (Bussière,2007;Zhao et al.,2019;Zheng et al.,2019).It is widely used in underground mines to increase ore recovery,reduce mineral dilution and improve ground stability(Hassani and Archibald,1998;Hustrulid and Bullock,2001;Potvin et al.,2005;Belem and Benzaazoua,2008;Darling,2011;Kermani et al.,2015a,b;Koupouli et al.,2016;Lu et al.,2017;Sobhi and Li,2017).In addition,this technique allows a maximum reuse of mine tailings in underground stopes.The surface disposal of mine tailings can thus be minimized,resulting in a better preservation of environment (Bussière,2007;Ercikdi et al.,2015;Yilmaz et al.,2015).All these explain well the numerous publications on CPB.For instance,a number of studies have been conducted to investigate the thermal and mechanical properties of CPB cured under different binder types and contents,superplasticizers,initial temperatures,curing stresses and drainage conditions (Cui and Fall,2016;Xu et al.,2018;Zhang et al.,2018;Wu et al.,2019).Wu et al.(2018) experimentally studied the influence of particle size distribution of CPB on its strength development.Recently,Dalcé et al.(2019) investigated the effect of segregation of hydraulic backfill on its uniaxial compressive strength development.Some comprehensive reviews on the mathematical models to evaluate the in situ performance (e.g.the thermal properties,stress state,and volume variation) of CPB have been recently reported by Cui and Fall (2019).Qi and Fourie (2019) reported the recent development in the flocculation and sedimentation,mix design and transportation of CPB.Zhang et al.(2019) summarized different types of backfill materials and the corresponding backfilling processes in coal mines in China.

When a slurried backfill is placed in a stope,an important task is to evaluate the required volume of slurried backfill to fully fill a given stope and ensure a tight contact between the backfill and stope roof.When the top surface of placed backfill is used as part of drift floor,it is critical to well estimate the volume of backfill slurry in the mine stopes to avoid creation of excessive void between the top surface of the settled backfill and the floor of overcut drift used as passing way of vehicles and workers.When the stope is filled in two stages(i.e.plug pour and final pour),a critical concern is how to ensure the final surface of the plug pour to reach the designed level (usually 2-3 m higher than the roof level of access drift in which the barricade is built to retain the backfill in place).A too much higher plug pour than the access drift leads to a useless increase of binder consumption.With a plug pour lower than the roof level of access drift,excessive pressures can exercise on the top part of the barricade during the final pour,resulting in undesirable consequence associated with internal local instability near the upper portion of the barricade (Yang et al.,2017).

In order to estimate the required volume of backfill to fully fill an underground stope,one has to perform shrinkage tests to obtain some required parameters (Qin,2020).In fact,when a backfill slurry is placed in a stope,excess pore water pressure can be generated (Gibson,1958;Essayad,2015;Zheng et al.,2018a,b;Essayad and Aubertin,2021).The particles of the backfill are not in contact and the effective stresses are zero.The volume reduction of the backfill is mainly associated with the dissipation of excess pore water pressure.The water consumption due to cement hydration is negligible for CPB as mentioned by Helinski et al.(2007) and also shown later in this paper.The volume reduction of the backfill is thus equal to the volume of water drainage.This process is known as“sedimentation”(Ahmed and Siddiqua,2014).The process of sedimentation is completed when the grain-grain contacts of the backfill slurry start(Dromer,2004;Pedroni,2011).The void ratio at the end of sedimentation is called“sedimentation end void ratio”,which has been shown to correspond to the unsaturation onset void ratio of shrinkage curve (Qin,2020).It is thus important to have a good understanding of the shrinkage behavior of CPB.

Over the years,studies have been mainly given on the shrinkage response of fine-grained soils,such as silt and clayey soils(Crescimanno and Provenzano,1999;Albrecht and Benson,2001;Boivin et al.,2004;Péron et al.,2009).The shrinkage response of cemented clays was analyzed by Omidi et al.(1996) while the shrinkage behavior of cemented dredged sludge was studied by Zhang et al.(2010)and Hu and Zhao(2015).A few works have been conducted to investigate the shrinkage properties of uncemented tailings(Qiu and Sego,2001;Rodríguez,2006;Fisseha et al.,2007;Rodríguez et al.,2007;Saleh-Mbemba et al.,2016;Simms et al.,2019) or the chemical shrinkage associated with the hydration of cement(Walske,2014;Wood and Doherty,2014,2018;Ghirian and Fall,2015;Cui and Fall,2017).These results can be useful to analyze the shrinkage response of tailings deposited in tailings impoundments,but inappropriate for underground mines with CPB.The study of the shrinkage behavior of CPB is thus necessary.However,most of the existing publications on CPB were devoted to the study on the mechanical and thermal properties under different curing conditions(Wu et al.,2018;Xu et al.,2018;Zhang et al.,2018;Dalcé et al.,2019).Study on the shrinkage behavior of CPB has seldom or never been reported.To fill this gap,a series of laboratory shrinkage tests has been performed on a CPB.

In this article,a brief review on soil shrinkage is given first.The results of laboratory shrinkage tests on CPB are then presented.The influences of cement content and initial water content on the shrinkage behavior of CPB are then analyzed.

2.Theory of soil shrinkage

Fig.1.A typical shrinkage curve of loose and disturbed unstructured slurried soils with Gs being the specific gravity (modified from Marinho,2017).

Fig.2.Grain size distribution curve of the tested tailings.

Shrinkage refers to the volume reduction as a function of water loss,mostly due to evaporation (Terzaghi,1925;Bowles,1984;Fredlund et al.,2002;Leong and Wijaya,2015).Fig.1 schematically shows a typical shrinkage curve of loose and disturbed unstructured soils,including the stages of normal shrinkage,residual shrinkage,and zero shrinkage (Peng and Horn,2013;Marinho,2017;Mishra et al.,2019).In the stage of normal shrinkage,the specimen always remains saturated and the water content w progressively decreases from its initial value w0to a critical value,called unsaturation onset water content (wAE).When the water content further decreases to a degree smaller than wAE,the shrinkage curve becomes nonlinear and the slope decreases significantly with further water loss because grain-grain contacts begin and air starts to replace the inter-granular voids initially occupied by water.This stage is called residual shrinkage(Peng and Horn,2005).Zero shrinkage is the final stage of the drying process,in which the grain-grain contacts become very tight.The void ratio(ef) remains unchanged despite further loss of water.The water content at the intersection between the saturation line(Sr=100%)and the horizontal line of zero shrinkage stage is called shrinkage limit ws(Marinho,1994;Fredlund et al.,2002).

Over the years,several methods have been used to obtain a precise measurement of volume change associated with water loss,and subsequently the shrinkage curve.These include direct measurement method,volume replacement method,and non-contact method (Li et al.,2019).Among them,the direct measurement is adopted in this study as it is non-destructive,fast and easy to execute.It is conducted by directly measuring the volume change of soil specimens using length measurement devices,such as Vernier calipers and linear variable differential transformers(LVDTs)(Péron et al.,2009;Saleh-Mbemba et al.,2016).

3.Shrinkage tests of CPB

3.1.Test materials

The CPB is made of tailings,cement and water.The tailings used in this study were collected from a hard rock mine in Quebec,Canada.Before specimen manipulation,the tailings were stored in sealed barrels.The specific gravity of the tailings was measured to be Gs=2.73.The grain size distribution of the used tailings was analyzed through sieve and hydrometer tests.Fig.2 shows the grain size distribution curve of the tested tailings with D10=3.3 μm,D30=14.2 μm,D60=43.2 μm,Cu=13.02(coefficient of uniformity),and Cc=1.4 (coefficient of curvature).It can be classified as low plastic silt (ML) according to the unified soil classification system(USCS)(Holtz et al.,1981).As the tested tailings contain 40%of fine particles smaller than 20 μm,the requirement that at least 15% of fine particles should be smaller than 20 μm as defined in Potvin et al.(2005) for a backfill to be considered as a paste backfill is largely satisfied.The saturated hydraulic conductivity ksatwas measured to range from 1×10-6m/s to 5×10-6m/s while the void ratio from 0.46 to 0.83 (Rodet,2019).These values are somewhat higher than typical values of tailings (Bussière,2007).Ordinary Portland cement of type 10 was used as the binder material with a specific gravity of 3.15.

The CPB specimens were obtained by mixing given masses of oven-dried tailings,de-aired water and ordinary Portland cement in a bowl.The initial(gravimetric)water contents w0are 40%,50%,60%,80%,and 100%(or 71.4%,66.7%,62.5%,55.6%and 50%in terms of initial solid contents by mass,S0),respectively.The CPB was prepared with cement contents C0of 3%,5%,and 7%,respectively.These cement contents are in the range of commonly used CPB in underground mines in Canada (Kesimal et al.,2005;Fall et al.,2007).Uncemented paste backfill (uCPB) has also been tested for the sake of comparison.After the mixing process,the CPB specimens were vibrated for 30 s to remove any air bubbles in the slurry.

3.2.Instrumentation and testing procedures

Fig.3.Photograph of the testing instrumentation with a Vernier caliper and a testing mold placed on a balance.

Table 1Parameters of the prepared specimens at the initial state.

The testing instrumentation was developed from that of Péron et al.(2009) and Saleh-Mbemba et al.(2016).Fig.3 shows the photograph of the testing instrumentation,which consists of a polyvinyl chloride (PVC) mold having an internal opening of 199.8 mm long,30 mm wide and 50.2 mm high,a balance(accuracy:0.1 g)and a Vernier caliber(accuracy:0.01 mm).In this study,two more PVC molds have been used,one having an internal opening of 205.9 mm long,31.1 mm wide and 50 mm high and another having an internal opening of 199.9 mm long,29.7 mm wide and 50.1 mm high.These dimensions are the results of a compromise,considering test time,accuracy and representativeness of test results and oven capacity(Li et al.,2013;Pedroni,2011;Zheng et al.,2020a,b).

The tests started by measuring the mass and sizes of the internal opening of the empty mold having its internal walls and base applied with grease and Teflon.A prepared tailings specimen was then manually placed in the mold with a flat spoon until the mold was filled up with specimen.During the placement,the mold was gently tapped on the four sides to eliminate any air bubbles between the fill and side walls.A soil knife and a wet towel were used to carefully remove the excess material on top of the mold.The initial total mass of each specimen (Mt0) was measured while the initial volume (V0) was determined from the sizes of the internal opening covered with grease and Teflon.The initial total mass of solids (Ms0.tot) and initial mass of water (Mw0) are calculated from the initial water content (w0) and total mass (Mt0) as follows:

In order to prevent sticking between the specimen and the inside walls of the mold,the base and lateral walls of the mold were first applied with a thin layer of silicone grease and then covered by a layer of Teflon.The measured dimensions of the internal opening above are the dimensions after Teflon and grease are applied.This treatment is necessary and efficient to avoid cracking of the cemented tailings during the shrinkage.

The initial masses of cement solids (Ms0.c) and tailings solids(Ms0.tail)can then be calculated as follows from the cement content(C0):

Fig.4.Shrinkage curves of uncemented and cemented tailings with different cement contents under the initial water contents of (a) 100%,(b) 80%,(c) 60%,(d) 50%,and (e) 40%.

The initial total volume of solids(including tailings and cement),Vs0.tot,is obtained as follows:

where ρtailings(=2730 kg/m3) and ρcement(=3150 kg/m3) are the densities of the tailings and cement solids,respectively.

The initial void ratio e0is then obtained as follows:

Table 1 shows the properties of specimens prepared for the shrinkage tests.

The drying process of the filled mold took place under an ambient environment with a room temperature of 20°C±3°C and relative humidity between 30% and 50% (Saleh-Mbemba et al.,2016).At predetermined time intervals,the mass of the filled mold was measured.The mass loss of water and the new water content can then be determined as follows:

where Mtis the measured total mass of the specimen during the shrinkage tests.

Meanwhile,the length and width of the specimen were directly measured with the Vernier caliper while the thickness was obtained by subtracting the measured settlement from the initial thickness.Each dimension was determined by averaging five measurements at different positions.The volume and void ratio variation of the specimen can then be calculated as follows:

where Vtis the total volume of the specimen,which can be calculated as the multiplication of the measured length L,width B,and height H of the specimen during the shrinkage tests.

The interval between two successive measurements was generally 1 h during the first day and increased to longer time from the second day.The tests were ended when the total volume of the specimen became unchanged for 2 d.The specimen was then withdrawn from the mold and placed in the oven at a temperature of 250°C for 24 h.The total mass of solids Msf.totwas then measured.The shrinkage curve can then be obtained by plotting the variation of the void ratio (Eq.(8)) as a function of the water content (Eq.(7)).

4.Test results and interpretation

Fig.4 shows the shrinkage curves of the tested CPB with different cement contents when the initial gravimetric water content w0is 100%(Figs.4a),80%(Fig.4b),60%(Figs.4c),50%(Fig.4d)and 40% (Fig.4e),respectively.The test results of uCPB have also been plotted in the figure for the sake of comparison.In all cases,the shrinkage curves of uCPB show a linear variation between the void ratio and water content,following the saturation line during the normal shrinkage stage.These results are similar to those obtained on tailings by Rodríguez (2006) and Saleh-Mbemba et al.(2016).Thereafter,the shrinkage curves exhibit a short residual shrinkage stage.These results are similar to those shown in Rodríguez (2006),but different from those reported by Saleh-Mbemba et al.(2016) who showed that the residual shrinkage stage was almost absent and the shrinkage curves were almost bilinear curve.This difference can be partly explained by the difference in the gradation and soil water retention characteristics of tested tailings.However,more work is necessary to fully understand the different shrinkage behaviors of different materials.After the residual shrinkage stage,a zero shrinkage stage (e=ef) is observed on the tested tailings,as shown by Rodríguez(2006)and Saleh-Mbemba et al.(2016).

At a given initial water content w0,the shrinkage curves of CPB and uCPB follow the same saturation line(Sr=100%) with a slope equal to Gsin the normal shrinkage stage.For the case of w0=50%,slight differences can be observed between the shrinkage curves of CPB and uCPB.This is probably due to some errors or disturbance associated with some renovation work around the laboratory.In all cases,the CPB exhibits shorter normal shrinkage stage than the uCPB.The unsaturation onset water content (wAE) and void ratio(eAE) as well as the shrinkage limit (ws) and final void ratio (ef) of the CPB are much larger than those of uCPB.They generally increase as the cement content increases even though the increasing rates are not homogeneous when the initial water content changes from 100% to 40%.These results are similar to those reported by Omidi et al.(1996) on cemented clays with a cement content of 3%-12%and those presented by Zhang et al.(2010) on the shrinkage of cemented dredged sludge with a water content of 70%-140%and a cement content of 0%-20%.The results clearly indicate the occurrence of hydration during the shrinkage tests on the CPB.

When cement meets water,a chemical reaction process known as hydration starts.According to Locher(2006),small quantities of ettringite and calcium hydroxide are generated at the very beginning.The CPB may become stiffer,but without generation of any cohesion.This stage can last for 4-6 h.After this first stage of induction,the reaction accelerates.Long fibers up to 2 m can be generated as long as the required spaces are available.With further hydration,a network of short,strip-shaped calcium silicate particles can be formed after 24 h.The resulting microstructure can exhibit a certain strength even though the porosity still remains quite high,which was also found in scanning electron microscope(SEM)analysis and pore structure study by Ouellet et al.(2008).As the tested CPB has high initial void ratios and low cement contents,the long fibers and microstructure network can freely develop and occupy more space.When the shrinkage reaches the onset point of unsaturation,the stiff and strong fibers and microstructure network resist the volume contraction associated with the suction,resulting in higher final void ratios and shrinkage limits than the uCPB.

Theoretically,the cementation and hydration of CPB consume pore water and transform a portion of free water into bonded water,resulting in high final void ratio and shrinkage limit.It is thus interesting to compare the final total masses of solids(including the bounded water and the initial total masses of tailings and cement)at the end of each shrinkage test (Msf.tot) with the initial total masses of solids (Ms0.tot).

Table 2 shows the total masses of solids at the beginning(Ms0.tot)and end (Msf.tot) of each shrinkage test.The shrinkage parameters,including the unsaturation onset water content(wAE)and void ratio(eAE)as well as the shrinkage limit(ws)and final void ratio(ef),are also presented in the table.

Table 2Parameters of the tested CPB and uCPB at different stages of shrinkage.

For the case of uCPB,the total masses of solids at the end(Msf.tot)and beginning(Ms0.tot)of each shrinkage test are very close to each other.These results indicate that the mass measurements and the test procedure are reliable.

When the cement content C0was 7%,the total mass of solids at the end(Msf.tot)of each shrinkage test was even smaller than that at the beginning (Ms0.tot).This is because the cement content C0was high and the bond generated at the end of each shrinkage test between the CPB specimens and Teflon became quite strong.Care had to be given to recover all the specimen of each test.Loss of specimens was unavoidably generated.

When the CPB has a cement content C0of 3% or 5%,the total masses of solids at the end (Msf.tot) and beginning (Ms0.tot) of each shrinkage test are very close to each other.These results indicate that the water consumption and bonded water are negligible during the curing process of the CPB.Similar results have been reported by Helinski et al.(2007),who indicated that the amount of water consumption by hydration and chemical shrinkage can be negligible for high water content and low cement content of CPB.If one evaluates the degree of cementation and hydration of CPB based on the consumption of water,one probably concludes that the cementation and hydration are absent during the two or three days of shrinkage.In fact,the combined water during the hydration can be divided into evaporable and non-evaporable types.The nonevaporable water represents the chemically bounded water,which is combined with cement and transferred into hydrated compounds.It does not evaporate or separate from the hydrated compound even under high temperature.The non-evaporable water during the first 3 d of curing only represents 12% of cement(Locher,2006).Table 3 shows the measured and expected final masses of solids at the end of each shrinkage tests of 2 d or 3 d.The variation of solid mass in specimens should show the mass of nonevaporable water.One can see that the measured and expected values are very close to each other.The negligible amount of nonevaporable water consumption during the shrinkage tests of 2 d or 3 d cannot be taken as a reliable indicator of hydration.

Table 3Measured and expected final total masses of solids at the end of each shrinkage test for the CPB having cement contents of 3% and 5%,respectively.

Fig.5.Shrinkage curves of tailings with different initial water contents under the cement contents of (a) 0%,(b) 3%,(c) 5%,and (d) 7%.

Fig.6.Variations of (a) shrinkage limit ws and (b) final void ratio ef as a function of cement content under different initial water contents.

Fig.7.Variations of (a) wAE and (b) eAE as a function of cement content under different initial water contents.

Fig.8.Photograph of different crack patterns of CPB with different boundary conditions:(a) No treatment;(b) Applying plastic film;and (c) Applying grease and Teflon.

Fig.5 shows the shrinkage(e-w)curves for different initial water contents when the initial cement content C0is 0% (Fig.5a),3%(Fig.5b) and % (Fig.5c) and 7% (Fig.5d),respectively.For uCPB(C0=0%),the shrinkage curves of different initial water contents follow the same saturation line in the normal shrinkage stage.However,the final void ratio and shrinkage limit increase as the initial water content increases.

For CPB,the influence of initial water content on the shrinkage behavior becomes more pronounced.The shrinkage limit and final void ratio increase as the initial water content increases.At a given cement content,the grain-grain spacing of a specimen with a higher initial water content can be larger.The development of long fibers and network microstructure can be freer and more complete.This can explain the larger final volume and larger final void ratio of the specimen at the end of each shrinkage test.

Fig.6 shows the variations of the shrinkage limit (Fig.6a) and final void ratio (Fig.6b) as a function of cement content under different initial water contents.The uCPB shows the lowest shrinkage limit,which is in the range of 30%-40%.At a given initial water content,the cement content of CPB has only a slight effect on the shrinkage limit.The value of wsincreases slightly as the cement content increases from 3% to 7%.At a given cement content,the shrinkage limit of CPB can be significantly influenced by the initial water content.A higher initial water content results in a higher shrinkage limit.Similarly,the uCPB shows the lowest final void ratio,compared to the CPB.At a given initial water content,the final void ratio of CPB increases slightly as the cement content increases from 3% to 7%.At a given cement content,higher initial water content leads to higher final void ratio.

Fig.7 shows the variations of unsaturation onset water content wAE(Fig.7a)and void ratio eAE(Fig.7b).One can see that the values of wAEand eAEfor CPB are higher than those of uCPB.The mechanism to explain these results is the same as that for shrinkage limit and final void ratio.At a given cement content,the values of wAEand eAEfor CPB increase significantly as the initial water content increases.However,the values of wAEand eAEfor CPB at a given initial water content increase only slightly as the cement content increases.

5.Discussion

The molds used in this study are 200 mm long,30 mm wide,and 50 mm deep.These sizes were chosen to study the volume change properties of CPB and uCPB only associated with water loss by evaporation.However,large dimension testing molds,especially in depth,should be avoided.Otherwise,the processes of segregation,self-weight consolidation and arching effect may take place during the shrinkage tests.This may render the result interpretation very complex.Moreover,a very thick specimen may not only need longer time to evaporate,but also result in dried specimen near the top surface and saturated specimen near the base (Nahlawi and Kodikara,2006).

In this study,the drainage and evaporation were only possible in the upward direction.In practice,lateral drainage can take place in a mine stope through the surrounding fractured rock walls(Belem et al.,2016) or permeable barricades (Rankine,2005).The lateral drainage can accelerate the water loss and increase the volumetric shrinkage rate (ratio of decreased volume to initial volume).More work is required to consider more representative drainage conditions.

Another limitation of this study is related to the drying process conducted in laboratory under an ambient environment with a room temperature of 20°C ± 3°C and relative humidity between 30% and 50%.These environmental conditions may not always be representative of underground mine stopes,in which the temperature can be as high as 35°C in 3 km deep mines or as low as subzero in permafrost regions(Wu et al.,2013;Jiang et al.,2016).As the main object of shrinkage tests is to analyze the variation of volume as a function of water loss,the results of this study can be assumed to remain valid as long as the temperature is higher than 0°C or lower than 35°C.However,more work is needed to verify the influence of temperature and humidity on the shrinkage behavior of CPB.Moreover,the shrinkage tests were conducted with only one type of tailings and binder.The water reducer admixture (e.g.superplasticizer) is also not used in the CPB in order to avoid its influence on the shrinkage of CPB.Therefore,the water content of 40%is the minimum value that can be prepared to ensure the CPB to be in a saturated state.The initial water contents of 50%,60%,80%and 100% are higher than that used in practice,which cannot be strictly representative of the real CPB.They were used here to investigate the influence of different water contents on the shrinkage behavior of CPB.More work is also required to investigate the shrinkage behavior of CPB made of different tailings and binders as well as with the addition of superplasticizer.

When shrinkage tests were conducted without applying lubricants or plastic film on the base and inner walls of the mold,cracks would appear during the shrinkage tests,as shown in Fig.8.Cracking can be avoided when the mold is treated with grease and Teflon.More work is required to consider the influence of cracks on the shrinkage behavior in the future.

6.Conclusions

The shrinkage behaviors of CPB and uCPB have been investigated.The shrinkage behavior of CPB is very different from that of uCPB.The normal shrinkage stage of the CPB is much shorter than that of uCPB.The unsaturation onset water content and void ratio,shrinkage limit,and final void ratio of CPB are higher than those of uCPB.The shrinkage curves of CPB can be significantly influenced by the initial water content.At a given cement content,the unsaturation onset water content and void ratio,shrinkage limit and final void ratio of CPB increase significantly as the initial water content increases.However,the shrinkage behavior of CPB is only slightly influenced by the cement content.At a given initial water content,the increase of cement content from 3%to 7%can only lead to slight increase in unsaturation onset water content,void ratio,shrinkage limit,and final void ratio.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgments

The authors would like to acknowledge the financial support from the Natural Sciences and Engineering Research Council of Canada(Grant No.NSERC 402318),Fonds de recherche du Québec-Nature et Technologies(Grant No.FRQNT 2015-MI-191676),Mitacs Elevate Postdoctoral Fellowship(Grant No.IT12573),and industrial partners of the Research Institute on Mines and the Environment(RIME UQAT-Polytechnique).

Appendix A.Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.jrmge.2021.01.005.