Plasticity role in strength behavior of cement-phosphogypsum stabilized soils

Xia Bian,Lingling Zng,Fng Ji,Ming Xi,Zhnshun Hong

a Key Laboratory of Ministry of Education for Geomechanics and Embankment Engineering,Hohai University,Nanjing,210098,China

b State Key Laboratory for Geomechanics and Deep Underground Engineering,China University of Mining and Technology,Xuzhou,221116,China

c College of Civil Engineering,Zhejiang University of Technology,Hanzhou,310014,China

d Jiangsu Water Source Company Ltd.of the Eastern Route of the South-to-North Water Diversion Project,Nanjing,210000,China

e Fujian Yongfu Power Engineering Co,Fuzhou,350108,China

f Institute of Geotechnical Engineering,School of Transportation,Southeast University,Nanjing,211189,China

Keywords:Strength Stabilization Plasticity index Microstructure Mineralogical analysis

ABSTRACT Dredged soil and phosphogypsum are frequently regarded as wasted materials,which require further treatment to control their environmental impact.Hence,phosphogypsum is proposed as a binder to stabilize dredged soil,aiming at efficiently reducing and reusing these waste materials.In this study,the engineering properties of cement-phosphogypsum stabilized dredged soils were investigated through a series of unconfined compression tests,and the effects of plasticity index of original soils on the strength improvement were identified.Then,the microstructure test and mineralogical test were performed to understand the mechanism of physical role of original soils in strength improvement.The results revealed that the unconfined compressive strength significantly decreased with the increase in plasticity index at the same binder content.The essential factor for strength improvement was found to be the formation of cementitious materials identified as calcium silicate hydrate (CSH),calcium aluminate hydrate (CAH),and ettringite (Aft).The normalized integrated intensity of cementitious materials(CSH + CAH + Aft) by pore volume decreased with the increase in plasticity index.Consequently,the density of cementitious materials filling the soil pores controlled the effectiveness of strength improvement.More cementitious materials per pore volume were observed for the original soils with lower values of plasticity index.That is,the higher strength of stabilized soils with lower values of plasticity index was attributed to a packed structure forming by integrated fabric through denser cementitious components.It can be anticipated from the above findings that the effectiveness of stabilization treatment will significantly reduce with the increase in plasticity of origin soil.

1.Introduction

Dredged soils have been known as a material with high water content and low strength,which were often abandoned as wastes(Zeng et al.,2015; Bian et al.,2016,2018,2021a).In recent years,more than 100 million m3of dredged soils have been annually generated in China.The storage of dredged soils has caused severe eco-environmental stress.Hence,there are increasing awareness to reuse dredged soils as geological material,which can be part of the metabolism of soil aggregate inside the city,in accord with the concept of sustainable development.Chemical stabilization technique has been widely used to improve the workability and strength of dredged soils by mixing cement or lime,in order to meet the requirement for the construction of roads,embankments and airports.Due to the variation of geological conditions,the dredged soils consisted of different compositions of clay minerals and clay contents (Zeng et al.,2017).Hence,the geological complication led to the variety of physico-mechanical behaviors of dredged soils (Federico et al.,2015).The effectiveness of stabilization using the same design code varied significantly with soil types,consequently leading to the uncertainty of engineering performance of stabilized soils (Pedarla et al.,2011).

The engineering properties of cement or lime stabilized soils depended primarily on the amount of additives,curing process and period,as well as the soil type (Sivapullaiah et al.,2000; Cherian and Arnepalli.,2015).Specifically,the strength improvement of stabilized soils was mainly due to the primary and secondary cementitious products from the hydration reaction (Yi et al.,2015,2018; Wang et al.,2017).The secondary products (e.g.calcium silicate hydrate(CSH)and calcium aluminate hydrate(CAH))resulted from the pozzolanic reaction between cement or lime and clay minerals,providing the long-term strength,stiffness and durability of stabilized soils(Bell,1996;Le Runigo et al.,2011;Choobbasti and Kutanaei,2017;Atahu et al.,2019;James,2020).Such improvement of engineering properties depended largely on the physical properties of original soils.The Atterberg limits (i.e.liquid limit,plastic limit and plasticity index) have been used as a powerful index to characterize the soil type,as well as engineering property of soils.Miller and Azad (2000) suggested that the strength of stabilized soils significantly decreased with the increase in plasticity index of original soils.Pedarla et al.(2011) investigated the clay mineralogical influence on the stabilization behavior.Sivapullaiah et al.(2000) found that the fabric changes of stabilized soils were controlled by the predominant clay mineralogical type (i.e.montmorillonite or kaolinite).It has been well reported that the Atterberg limits are strongly affected by clay minerals(Miller and Azad,2000).These previous studies indicated that the engineering performance of stabilized soils was associated with the Atterberg limits of original soils.Note that,the mechanism of soil plasticity role on strength improvement of stabilized soils and the corresponding microstructure variation are still less understood.

This study aims at investigating plasticity role of dredged soils in strength behavior of cement-phosphogypsum stabilized soils.Four naturally sedimentary soils with different values of plasticity index were mixed with water to generate the dredged soils at different initial water contents.The prepared soils were mixed with different contents of cement and phosphogypsum for preparing testing specimens.A series of laboratory tests was carried out for evaluating the variations of physical properties,strength and microstructure.The pH value and water content reduction of stabilized soils were investigated with respect to different values of soil plasticity.Accordingly,the effect of soil plasticity on unconfined compressive strength (UCS) of stabilized soils was proposed.The variation of UCS was then linked with the change in physicochemical behavior.Finally,the mechanism responsible for the physico-chemical-strength relationship was discussed along with the stabilization-induced change in microstructure.

2.Materials and methods

2.1.Materials

Four naturally sedimentary soils (termed as Soils 1-4) were used in this study,as shown in Table 1.Soil 1 was a high plasticityclay with plasticity index of 67.9%from Water Park,Fuzhou,China.Soil 2 was a high plasticity clay with plasticity index of 48.6%from Pudong river,Fuzhou,China.Soil 3 was a low plasticity clay with plasticity index of 15.2% from Xinjing river,Fuzhou,China.Soil 4 was a low plasticity clay with plasticity index of 12.5% from Fengban river,Fuzhou,China.Based on the Unified Soil Classification System(ASTM D2487-11,2011),Soils 1-4 were classified as CH,CH,CL and ML,respectively.The mineralogical compositions of the four dredged soils were reported by Wang et al.(2021),as presented in Table 2.It can be observed that the predominant clay mineral was kaolinite,with a proportion from 46% to 59%,followed by illite between 22% and 33%and chlorite ranging from 19% to 30%.

Table 1 Physical properties of four soils used in this study.

Two additives were used as the binder: cement and phosphogypsum.Their chemical compositions are listed in Table 3.Ordinary Portland cement(32.5R/N)was a commercial product from Nanjing Conch Cement Co.,Ltd.,China.Phosphogypsum was retrieved from a landfill of waste phosphogypsum in Nanjing.Based on the chemical analysis obtained from an independent laboratory test,phosphogypsum mainly contained the oxides SO3and CaO of 41.7%and 27.8%,respectively.Meanwhile,the ratio of CaO to SiO2for the cement used in this study was 2.64 (Zeng et al.,2021).

2.2.Experimental methods

The test program for stabilized soils is shown in Table 4.The stabilized soil samples were prepared from the slurry clay.First,the dredged soils retrieved from the riverbed were passed through a 2-mm seize to eliminate the large size particle.The soil slurry at an initial water content of 1 or 1.5 times the liquid limit was prepared by mixing the seized soil with predetermined water using a mixer to obtain a homogeneous slurry paste.Then,the cement and phosphogypsum powder was poured into the mixer and mixed for about 5-10 min to achieve uniformity.The amount of cement was set as 100 kg/m3,representing 1 m3slurry using 100 kg cement for stabilization.The amount of phosphogypsum ranged from 0 to 40 kg/m3.Afterwards,the mixture was transferred into plastic cylindrical mold(39.1 mm in diameter,80 mm in height).After 1 d of curing,the cylindrical samples were extruded out of the molds.The cylindrical specimens were wrapped in plastic bags separately and cured in a controlled environment((20 2) C and 95%of relative humidity).The photos of slurry,plastic mold and sample after demolding are shown in Fig.1.

Table 2 Clay mineral compositions of four dredged soils (after Wang et al.,2021).

Table 3 Oxide composition (%) of ordinary Portland cement and phosphogypsum (after Zeng et al.,2021).

According to ASTM D4972-11 (2011),HORIBA D-54 pH meter was used to measure the pH value of stabilized specimens.The soil specimens at different curing times were air-dried and passed through a 2-mm sieve.The solution of 10 g sieved soil and 50 mL of distilled water were mixed for 3 min.After 30 min,the pH value was determined by pH meter using the solution.Triplicate measurements were applied for each specimen and the average pH values were reported.

The unconfined compression tests (UCTs) were performed on specimens at different curing times(7 d,28 d and 90 d),following ASTM D4219-11 (2011).The rate of vertical displacement in UCTswas 1 mm/min.Triplicate measurements of UCS (qu) were conducted and the average value was adopted.The water contents at different curing times were determined by oven-drying the specimens at 105 C for 24 h,using the specimens immediately after UCT.A Rigaku D/Max-2500 was used to conduct the X-ray diffraction(XRD) analysis using the lyophilized samples.The lyophilized samples were prepared through a freeze-dried procedure,as reported in Bian et al.(2019,2020).The two-theta(2θ)value used in this study ranged from 5 to 60 with a step length of 0.02 and a scanning rate of 2 per min.A NOVA NanoSEM 230 microscope was used to perform the scanning electron microscopy (SEM) analysis,in which the lyophilized specimen was coated with a gold layer to induce conductivity.

3.Results and discussion

3.1.pH value

Fig.2 shows the typical pH value of cement-phosphogypsum stabilized soils at initial water content ratio w0/wL=1 and 1.5 after 28 d of curing.It can be observed that the pH values decreased with the increase in phosphogypsum content for all stabilized soils.This can be attributable to the acidic substance of phosphogypsum,with a pH value of 3.2(Zeng et al.,2021).However,it is interesting to note that the pH values of all stabilized soils in this study were higher than 10,as shown in Table 4.Such high pH values provided the required alkaline environment for the pozzolanic reaction between additives,pore water and soil minerals.Moreover,it appears that the pH values showed a significant variation with the soil types.In Fig.1,the pH curves of Soil 4(IP=12.5%)lied above others with the highest pH values,following by the order of Soil 3(IP=15.2%),Soil 2 (IP=48.6%) and Soil 1 (IP=67.9%) along the vertical axis.This indicates that at given binder content and water content,the pH value decreased with the increase of plasticity index.

Table 4 Test program for UCS and pH value.

Fig.3 shows the variation of pH value with curing time.It can be observed that the pH value of the stabilized soils decreased with the increase in curing time.This can be attributable to the longterm reaction between portlandite and water during pozzolanic reaction.It has been reported that the OH level decreased during the long-term reaction,consequently resulting in the decrease in pH value(Chew et al.,2004;Du et al.,2014).Within 90 d of curing,the pH values of specimens with lower values of IPwere higher than those with higher values of IP.This is consistent with the variation of pH value with soil plasticity shown in Fig.1.Moreover,at a given phosphogypsum content,the pH value decreased with the increase in water content.The average change in pH value between w0/wL=1 and 1.5 was about 0.1.Such a difference was mainly due to the lower OH concentrations for the original soils with higher initial water contents at the same binder content.

To highlight the relationship between the pH value and the soil plasticity,the pH value at 20 kg/m3of phosphogypsum content is plotted against the soil plasticity index,as shown in Fig.4.Note that the pH value at 20 kg/m3phosphogypsum content was selected due to the fact that the stabilized soil with phosphogypsum content at 20 kg/m3was conducted at all the soil types and all the water contents.It can be observed that the pH value correlated well with the values of plasticity index of the original soils.It showed a clear decrease tendency,irrespective of water content and curing time.This is most probably due to that the soil with higher Ipgenerally possessed higher cation exchange capacity with higher clay fraction(Miller and Azad,2000).Hence,a greater chemical reaction was expected to be reacted.As a result,the pH value became lower for the original soils with higher Ip.Note that the most significant change in pH value occurred when Ipwas smaller than 50%.

3.2.UCS

Figs.5 and 6 depict the relationship between UCS (qu) and phosphogypsum contents at 3 and 28 d of curing,respectively.At short curing time (3 d),quincreased with the increase in phosphogypsum content to a peak value,followed by a decrease in quwith further increase in phosphogypsum content.The phosphogypsum content for the peak quranged from 10 kg/m3to 20 kg/m3.On the other hand,qupositively increased with the increase in phosphogypsum content at 28 d of curing.This behavior mainly resulted from the swelling natural of ettringite(Aft),which was the main cementitious product from the pozzolanic reaction between cement,phosphogypsum and clay minerals.When the curing period was short (＜7 d),the increase in strength with phosphogypsum was due to the compression of soil void with the swelling of Aft.However,when phosphogypsum content exceeded the optimal value,over-swelling of Aft would lead to the disturbance of soil bonding,resulting in the further decrease of qu(Zeng et al.,2021).Meanwhile,at longer curing period (＞28 d),the strong cementation bonding was formed with an increase in quwith cementitious products(i.e.CSH or Aft).Therefore,quincreased with the increase in phosphogypsum content after 28 d of curing.

Moreover,the results of UCS also significantly varied with the soil type.The value of quat the same phosphogypsum content was significantly higher for the soil with lower Ip.The slopes of strength curves also showed a decrease trend with the increase in Ip.This suggested that the change in quof stabilized soils as a function of phosphogypsum content depended significantly on the soil plasticity of the original soils.The effectiveness of phosphogypsum on the strength improvement diminished for the soils with higher Ipat the same binder content.

To illustrate the effect of soil plasticity on strength gain due to phosphogypsum,the relationship between quand plasticity index at 20 kg/m3of phosphogypsum is plotted in Fig.7.It is evident that the change in qushowed a good correlation with Ip.Irrespective of curing time and initial water content,qudecreased significantly with the increase in plasticity index of original soils until Ip=50%.When Ipexceeds 50%,qushowed a slight reduction with soil plasticity index.Obviously,for a given Ip,quincreased significantly with the increase of curing time,and decreased with the increase of water content.This behavior agreed in general with the previous findings of strength variation of cement stabilized soils(Miller and Azad,2000; Chew et al.,2004; Lorenzo and Bergado,2004).

Comparing the result in Fig.4 with that in Fig.7,it can be found that both pH value and qushowed strongly correlation with Ip.Hence,it is logical to correlate quwith pH value for a given phosphogypsum content.Fig.8 demonstrates the typical relationship between quand pH value at 20 kg/m3of phosphogypsum.It can be observed that for a given phosphogypsum content,quat a given curing time increased significantly with pH value of stabilized soils,irrespective of soil plasticity.The relationship between quand pH value can be fitted by an exponential function as

Fig.1.Photos of (a) slurry,(b) plastic mold,and (c) specimen after demolding.

where a,b and c are the fitting parameters.The coefficients of determination (R2) are 0.91,0.97 and 0.95 for curing time t=7 d,28 d and 90 d,respectively.

From Eq.(1),it is logical to deduce that there is a correlation between UCS and pH value for cement-phosphogypsum stabilized soils.That is,pH value can be used as an indicator to predict the strength of cement-phosphogypsum stabilized soils.This provides a useful and rapid first-order concept to evaluate the effectiveness of phosphogypsum on the strength gain of stabilized soils using the pH value,because the examination of pH value is easy to conduct.

3.3.Water content reduction

For the soil slurry with high water content,the water content reduction is also directly related with strength growth of stabilized soils.Fig.9 shows the typical relationship between water content reduction and phosphogypsum content at 28 d of curing.In the figure,Δw represents the water content reduction,equalling to w0wt,where wtis the water content of stabilized soil after curing.It can be seen from the figure that water content reduction for stabilized soils increased remarkably with the increase in phosphogypsum content.This was mainly attributable to the consummation of water due to the chemical reaction between phosphogypsum and hydration products (i.e.CAH) (Zeng et al.,2021).For a given phosphogypsum content and initial water content ratio w0/wL,it is evident that the water content reduction for soil with higher Ipwas significantly higher than that with lower Ip.This corresponded to the soil with higher Ippossessing a higher initial water content at the same w0/wL.In addition,for the same soil,the water content reduction obviously increased with the increase in w0/wL.

Fig.10 depicts the normalized water content reduction (Δw/Ip)of stabilized soil at 28 d of curing.In this figure,the water content reduction was normalized with plasticity index of the original soils,to illustrate the effect of soil plasticity.It should be pointed out that the change in normalized water content reduction ratio Δw/Ipwith Ipshowed an opposite trend in comparison with that of water content reduction.That is,at a given phosphogypsum content and w0/wL,the normalized water content reduction Δw/Ipdecreased significantly with the increase in Ip.In other words,although the water retaining capacity was higher for the soil with higher Ip,the proportion of water consummation to the natural water retaining capacity (i.e.Ip) during the chemical treatment was lower for the soil with higher Ip.Hence,the variation of Δw/Ipin Fig.10 was similar to the change in quin Fig.5.That is,both Δw/Ipand qudecreased with the increases in Ipand w0/wL,but increased with the increase in phosphogypsum content.

Fig.2.Measured pH values of stabilized soil(28 d of curing):(a)w0/wL=1,and(b)w0/wL=1.5.

The above findings indicate that there is a good correlation between the normalized water content reduction (Δw/Ip) and the UCS (qu),as shown in Fig.11.The results showed that there was a highly significant positive correlation between quand Δw/Ip,expressed as

Fig.3.Variation of pH value with curing time.

Fig.4.Relationship between pH value and plasticity index at 20 kg/m3 of phosphogypsum content.

where α and β are the material constants.The coefficients of determination (R2) are 0.97 and 0.79 for w0/wL=1 and 1.5,respectively.At the same condition,quexponentially increased with the increase in Δw/Ip,irrespective of soil plasticity.Meanwhile,it is clear that the initial water content of slurry showed a significant influence on the quof stabilized soil.To obtain the same magnitude of qu,the normalized water content reduction (Δw/Ip)for soil with w0/wL=1.5 was about 1.5-2.4 times that of soil with w0/wL=1.This means that the effectiveness of soil stabilization in terms of quvalue decreased rapidly with the increase in initial water content.In other words,to stabilize slurry with higher initial water content,extra amounts of binders were required to consume more water,meeting the same engineering standard,which will increase the total cost.

3.4.Modulus of elasticity

Fig.5.Relationship between qu and phosphogypsum content (3 d of curing): (a) w0/wL=1,and (b) w0/wL=1.5.

Fig.6.Relationship between qu and phosphogypsum content(28 d of curing): (a) w0/wL=1,and (b) w0/wL=1.5.

The evaluations of modulus of elasticity (E50),as indicated in Figs.12 and 13,have a similar trend to the strength development curve.E50was defined as the slope of the stress-strain curve at 50%of peak axial stress,which was commonly used as an input parameter for calculating the deformation of stabilized soil (Bian et al.,2021b).At short curing times (＜7 d),E50rapidly increased with the increase in phosphogypsum content to the peak,and then decreased with further increase of phosphogypsum content.For the case of longer curing period (＞28 d),E50monotonously increased with the increasing phosphogypsum content within the studied range.The variation of E50with Ipfollowed the similar trend as that of qu.At a given phosphogypsum content,it is found that E50decreased with the increase in Ip,which was consistent with the change in qu.

Commonly,E50can be expressed as a function of quto predict the compressibility of cement stabilized soil(Lorenzo and Bergado,2004;Du et al.,2014;Zeng et al.,2021).The relationships of E50and qufor different stabilized soils are illustrated in Fig.14.The ratios of E50to quranged approximately from 40 to 450,with an average value of 233 for the different stabilized soils in this study.It should be emphasized that the wide range of E50/quratio may result from the variation of plasticity of the original soils.As shown in Fig.14,the experimental data for soil with different values of Ipis plotted,where the average value of E50/quratio was obtained for each soil.It is clear that the average E50/quincreased from 83 to 331 for Soil 1 to Soil 4,following the decreasing order of Ip.This suggested that the modulus of stabilized soil also decreased with the increase in soil plasticity at a given binder content.Fig.14 also depicts that the average E50/qufor Fuzhou soil with Ip=43.9%was 111 as reported in Zeng et al.(2021).This value was close to that of Soil 2 with Ip=48.6%in this study.Due to the fact that the stabilization binder was similar in this study and Zeng et al.(2021),this confirmed that the modulus of stabilized soil was dependent on the Ipof the original soil.

Fig.7.Relationship between qu and plasticity index at 20 kg/m3 of phosphogypsum.

Fig.8.Relationship between qu and pH value at 20 kg/m3 of phosphogypsum.

Fig.9.Water content reduction of stabilized soil (28 d of curing).

3.5.Mineralogy and microstructure

Fig.10.Normalized water content reduction of stabilized soil (28 d of curing).

XRD results of the original soils and stabilized soils of 20%phosphogypsum at 28 d of curing are presented in Fig.15.It can be found that illite,kaolinite and quartz were clearly apparent in the original soils.Hence,the main clay minerals for the four naturally sedimentary soils were illite and kaolinite.For the stabilized soils,the XRD patterns revealed several new peaks as compared with those of original soils.This indicated that the main cementitious products were CSH(peaks at 3.04 Å,2.77 Å and 1.87 Å),CAH(peaks at 7.65 Å,3.77 Å and 2.86 Å),and Aft (peaks at 3.87 Å,2.56 Å and 2.16 Å).These cementitious materials agreed well with previous findings for cement stabilized soils(Horpibulsuk et al.,2009;Wang et al.,2017; Zeng et al.,2021).

Fig.11.Relationship between qu and normalized water content reduction.

It has been widely accepted that the long-term strength development of stabilized soils was attributable to bonding effect of cementitious materials.Hence,it is logical to relate the variation of strength with the amount of main cementitious materials(i.e.CSH,CAH and Aft).Table 5 shows the variation of relative proportion of cementitious materials(CM),ratio of mass of cement to soil,as well as the pore volume of original soils v(=1+e),where e represents the void ratio of stabilized soil.The amount of cementitious materials corresponded to the integrated intensity of cementitious materials,CM(=CSH+CAH+Aft),where the relative proportion of CM expressed as the integrated intensity of CM for each soil to that of Soil 1.Thereby,the relative proportion of CM for Soil 1 was 100%.It should be emphasized that this semi-quantitative analysis only depicted the relationship between the strength improvement and the amount of main cementitious materials,which did not discuss the absolute quantities of these cementitious materials.Fig.16 depicts the variation of relative proportion of cementitious materials and cement ratio for Soils 1-4.At the same cement amount(expressed in kg/m3),the ratio of mass of cement to soil decreased from Soil 1,Soil 2 to Soil 3,and slightly increased to Soil 4.This was consistent with the change in density as shown in Table 5.Meanwhile,the relative proportion of cementitious materials also followed the same trend.It means that for a certain volume of soil,the amount of cementitious materials increased with the mass of cement.It can also be deduced that the clay mineral of all four soils were adequate for the pozzolanic reaction at the cement content of 100 kg/m3in this study.However,it should be pointed that the relative proportion of cementitious materials was higher for the soil with higher Ip,due to the higher mass ratio of cement.This behavior was opposite to the variation of strength with soil plasticity.One should also be aware that at the same initial water content ratio(w0/wL),the water content for soil with higher Ipwas much higher than that of lower Ip.This suggested that the pore volume for soil increased with the increase of Ipat the same w0/wL.Thus,it is interesting to normalize the amount of cementitious materials with the pore volume of soil (v=1 + e) as shown in Fig.17.It seems that the normalized cementitious materials decreased with the increase in Ip,the same tendency as the variation of qu.In other words,quincreased with the increase of normalized cementitious materials.Note that the definition of normalized cementitious materials corresponded to the amount of cementitious materials per pore volume.This means that at a given binder content,the strength of stabilized soil primarily depended on the density of cementitious materials filling the soil pores,rather than the total amount of cementitious materials.Hence,it can be conducted that when the amounts of cementitious materials were at a similar level,the soil with lower Ippossessed a lower level of pore volume,leading to a higher density of cementitious materials infilling the interparticle pores.Hence,the denser formation of clay cement clusters resulted in a stronger cementation bonding effect,thus a higher strength (Kamruzzaman et al.,2006).

Table 5 Relative proportion of cementitious compounds at 28 d of curing.

Fig.12.Relationship between E50 and phosphogypsum content(3 d of curing):(a)w0/wL=1,and (b) w0/wL=1.5.

Fig.13.Relationship between E50 and phosphogypsum content (28 d of curing): (a)w0/wL=1,and (b) w0/wL=1.5.

Fig.14.Relationship between modulus of elasticity (E50) and UCS (qu).

Fig.15.XRD patterns of (a) original soils and (b) stabilized soils after 28 d of curing.

Fig.18 shows the typical SEM micrographs for stabilized soil of 20% phosphogypsum after 28 d of curing.The major hydration reaction products of all the stabilized soils were CSH/CAH and Aft,which was consistent with the XRD analysis.Generally,CSH/CAH fabric was spread distribution on clay cluster and filled pore space between clay particles,forming a denser structure in comparison with the original slurry state.Moreover,the needle-like Aft crystal was also found on the clay surface and linked with the clay particle and CSH fabric to further improve the packed structure (Bahmani et al.,2016; Bian et al.,2018).The microstructure of stabilized soil significantly varied with the soil plasticity at the same binder content.For the high plasticity soil(Fig.18a and b),it seems that the pore volume was larger,with the appearance of large voids.Meanwhile,the less degree of cementitious material was observed to connect the soil particles and fill the pores,forming a looser fabric.On the other hand,the void volume decreased significantly for the soil with lower Ip(Fig.18c and d).The density of cementitious materials bridging the conjunctures of solid particles in soilcement matrix was at a higher level,leading to a denser structure.These observations confirmed with the findings of XRD in Fig.17.Hence,the soil with higher Iprequired more amount of binder content to refine the larger void volume,achieving the same extent of strength improvement.This explained the lower efficiency for stabilizing the higher plasticity soil.

Fig.16.Variations of relative proportion of cementitious materials and cement ratio.

Fig.17.Changes of normalized cementitious materials and qu.

Fig.18.SEM micrographs for stabilized soil after 28 d of curing: (a) Soil 1,(b) Soil 2,(c) Soil 3,and (d) Soil 4.

4.Conclusions

This study investigated the improvement of engineering properties of four original soils with different values of plasticity index stabilized by cement and phosphogypsum.The effect of soil plasticity index on the strength characteristic was related with the variation of physiochemical properties.The mineralogical composition and microstructure analysis was used to further understand the mechanism of soil plasticity role on the stabilization effectiveness.The following conclusions can be drawn from this study:

(1) The use of phosphogypsum can substantially improve the UCS quof soil with different values of plasticity index Ip.At a given phosphogypsum content,qusignificantly decreased with the increase in Ip.This suggested that the effectiveness of strength improvement was more significant for soil with low Ip.

(2) The pH value strongly correlated to the plasticity index,which can be used as an indicator to predict the change in UCS qu.The relationship between quand pH value can be expressed as an exponential function,i.e.qu=a exp(bpHc).

(3) The UCS qucan be normalized with the normalized water content reduction Δw/Ip,i.e.qu=α exp(βΔw/Ip).This relation was greatly influenced by initial water content w0,moving to the right side of axis with the increase of w0.This suggested that the effectiveness of soil stabilization in terms of qudecreased rapidly with the increase in initial water content.

(4) XRD analysis suggested that the main cementitious materials were CSH,CAH and Aft.The normalized integrated intensity of cementitious materials with pore volume strongly correlated with UCS qu,showing a decrease tendency with the soil plasticity index Ip.

(5) SEM micrographs for stabilized soil showed that the density of cementitious materials filling the soil pores increased with the reduction in Ip,leading to a denser structure for soil with lower Ip,thus a higher qu.This was the main reason for the lower strength improvement effectiveness for the soil with higher Ip.

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

This study is supported by the National Natural Science Foundation of China (Grant Nos.52178328 and 52178361).Partial financial support from the Open-end Research Fund of State Key Laboratory for Geomechanics and Deep Underground Engineering(Grant No.SKLGDUEK2114) is also acknowledged.