Effect of drying-wetting cycles on pore characteristics and mechanical properties of enzyme-induced carbonate precipitation-reinforced sea sand

Ming Hung ,Ki Xu ,Zijin Liu,b ,Choshui Xu ,Mingjun Cui

a College of Civil Engineering,Fuzhou University,Fuzhou,350108,China

b Xiamen Rail Transit Construction and Development Group Co.,Ltd.,Xiamen,361000,China

c School of Civil,Environment and Mining Engineering,The University of Adelaide,Adelaide,5005,Australia

Keywords: Enzyme-induced carbonate precipitation(EICP)Plant-based urease Drying-wetting (D-W) cycles Microstructure

ABSTRACT Enzyme-induced carbonate precipitation (EICP) is an emanating,eco-friendly and potentially sound technique that has presented promise in various geotechnical applications.However,the durability and microscopic characteristics of EICP-treated specimens against the impact of drying-wetting(D-W)cycles is under-explored yet.This study investigates the evolution of mechanical behavior and pore characteristics of EICP-treated sea sand subjected to D-W cycles.The uniaxial compressive strength(UCS)tests,synchrotron radiation micro-computed tomography (micro-CT),and three-dimensional (3D) reconstruction of CT images were performed to study the multiscale evolution characteristics of EICPreinforced sea sand under the effect of D-W cycles.The potential correlations between microstructure characteristics and macro-mechanical property deterioration were investigated using gray relational analysis(GRA).Results showed that the UCS of EICP-treated specimens decreases by 63.7% after 15 D-W cycles.The proportion of mesopores gradually decreases whereas the proportion of macropores increases due to the exfoliated calcium carbonate with increasing number of D-W cycles.The microstructure in EICP-reinforced sea sand was gradually disintegrated,resulting in increasing pore size and development of pore shape from ellipsoidal to columnar and branched.The gray relational degree suggested that the weight loss rate and UCS deterioration were attributed to the development of branched pores with a size of 100-1000 μm under the action of D-W cycles.Overall,the results in this study provide a useful guidancee for the long-term stability and evolution characteristics of EICPreinforced sea sand under D-W weathering conditions.

1.Introduction

With the increasing global population,the need for empty land is generally increasing in urban areas while the properties of soil on available land generally are not suitable for construction.The physical or chemical soil reinforcement methods have been reported in the literature (Radlińska et al.,2013).However,these traditional soil reinforcement methods inevitably result in environmental concerns(Hu et al.,2021).The increasing concern about global temperature alteration and the resulting influence on biological systems have pushed the quest for novel soil stabilization methods,which can accommodate modern requests with the conservation of the climate(Saif et al.,2022).

Enzyme-induced carbonate precipitation (EICP) and microbialinduced carbonate precipitation (MICP) are the emerging biotechnology that has been developed for soil stabilization (He et al.,2020;Arab et al.,2021;Jiang et al.,2022a).EICP technique achieves soil improvement through catalyzing the hydrolysis of urea to facilitate the calcium carbonate precipitation,which binds the soil particles (Martin et al.,2021).Its reaction process is similar to the mineralization that commonly occurs in natural environment using free urease (plant-and bacterial-based urease enzymes) for urea hydrolysis and the calcium carbonate crystal could precipitate subsequently when calcium ions are available.The EICP technique using the free urease instead of living organisms could reduce the cost when used in field implementations (Tirkolaei et al.,2020).Extensieve research has confirmed the adaptability of EICP technology for reinforcing the Ottawa 20-30 sand (Cui et al.,2021),desert sands(Miao et al.,2020)and other granular soils(Gao et al.,2019;Meng et al.,2021).The MICP/EICP-treated specimens generally exhibit rock-like behavior similar to sandstone (Sharma and Satyam,2021).The improvement of soil properties,via EICP/MICP technique,has been reported,such as micro-mechanical properties(Huang et al.,2021;Xiao et al.,2022),macro-mechanical properties(Cui et al.,2017;Xiao et al.,2018,2019;Hoang et al.,2020;He et al.,2022;Li et al.,2022),liquefaction resistance(Xiao et al.,2020),and freezing-thawing resistance (Jin et al.,2020).The perfect groutability and relatively low cost make the great potential of EICP in various geotechnical applications,such as soil improvement(Yasuhara et al.,2011),dust control (Sun et al.,2021),heavy metal immobilization (Xue et al.,2022),and mitigation of coastal sand dune erosion(Liu et al.,2021).

Published results have confirmed the potential of EICP and MICP techniques in geotechnical applications (Jiang et al.,2020).However,few studies have been reported on the effect of long-term exposure to the atmosphere and hydrosphere on the biocemented sands.Research on the long-term performance and durability of MICP/EICP-treated soil in various environments is limited.The alternation of natural environmental factors generally affects the mechanical properties of soil used in practical engineering (Huang et al.,2022).Upon these climatic changes,cyclic drying-wetting (D-W) has become a serious concern in geotechnical engineering.The cyclic D-W plays a vitally important role in influencing the strength and deformation properties of rock and geomaterials.Meanwhile,the cyclic D-W significantly affects the physical parameters of geomaterials (Dehestani et al.,2020;Zeng et al.,2020),and the evolution of micropores in response to D-W cycles is related to the deterioration of geomaterials (Zhao et al.,2020).The soil reinforced by the EICP technique inevitably faces the challenge of climatic changes such as rainfall,evapotranspiration and tide in the field implementation.The periodical moisture changes due to the cyclic D-W may affect the stability of various field application sites(Zhou et al.,2018;Sharma and Satyam,2021).Therefore,more attention should be paid to the physical and mechanical properties of EICP-treated specimens subjected to D-W cycles.The evolution of pore characteristics and mechanical properties of EICP-treated specimens subjected to different numbers of D-W cycles should be always studied before field implementation.Sharma and Satyam (2021) studied the weight loss rate and variation of macro-mechanical properties of MICP-treated specimens under the action of cyclic D-W treatment.Huang et al.(2022)investigated the modification effects of MICP-treated cohesive soils under D-W treatment.These studies mainly focus on the changes in mechanical properties,which ignore the influence of DW cycles on the changes in microstructures.To the authors’acknowledgment,the effect of D-W cycles on the development of pore characteristic within EICP-treated specimens remains poorly understood,and there is no systematic study on the correlations between microstructure characteristic parameters and macromechanical properties of EICP-treated specimens subjected to DW cycles.Therefore,more efforts should be made to investigate the multiscale evolution characteristics of EICP-reinforced sea sands prior to field applications.

The objectives of this study are to explore the evolution characteristics of EICP-treated specimens under the action of cyclic D-W treatment in terms of mechanical properties and pore characteristics.In this study,a series of tests with different numbers of D-W cycles was conducted on the sea sands reinforced by EICP technique using soybean husk urease.The accumulate weight loss and the evolution characteristics of surface morphology of EICP-treated specimens subjected to different numbers of D-W cycles were studied.The mechanical properties of EICP-treated specimens undergoing different numbers of D-W cycles were measured.The synchrotron radiation micro-computed tomography (micro-CT)was utilized to obtain three-dimensional (3D) reconstruction models of EICP-treated specimens for investigating pore morphologies and their evolution characteristics with the effect of D-W cycles.In addition,the gray relational analysis (GRA) was introduced in this study,and the gray relational co-efficiency(GRC)and gray relational grade (GRG) were calculated as the indices to disclose the potential correlations between microstructure characteristics and macro-mechanical properties of EICP-reinforced sea sands.

2.Materials and methods

2.1.Sands

The experimental sea sands were taken from the northeastern fringe of Xiamen,Fujian Province,China.The sea sands were desalinated before being used and the pH value was measured to be 7.15.The chloride ion content complied with the requirement of Chinese codes JGJ 52-2006 (2006) and GB/T 14684-2011 (2011).Sea sands with a grading size range of 0-1 mm were used in this study.Fig.1 shows the particle size distribution of sea sands.The void ratio of the used sea sand is 0.51,and its density is 1.49 g/cm3.The sea sands used in this study is same as that used in the field implementation of EICP technique as the backfill materials for the bedding layer beneath an underground cable duct in China (Xu et al.,2022).

Fig.1.Sea sands used in this study: (a) Collection site of the sea sands;and (b) Grain size distribution curve of the used sea sands.

2.2.Crude urease solution,cementation solution and bio-cemented sand column preparation

Fig.2 shows the procedure for the bio-cemented sand column preparation,including the following six steps:

Fig.2.Process of extraction of plant-derived urease and the EICP treatment of sea sands.

(1) Soybean husks were completely dried in the drying oven and ground into powder.Subsequently,the No.100-mesh was used to sieve the inadequately crushed soybean husks.

(2) Every 100 g soybean husk powder was mixed with 1 L of distilled water containing 30% ethanol,and the turbid solution was stirred at 800 rpm for 30 min.

(3) The turbid solution was placed in a 4°C refrigerator for 10 h,and then centrifugally rotated at a velocity of 10,000 rpm at this temperature for 15 min (initial urease activity of the crude urease was 5.5 IU/mL).

(4) EICP solution was prepared by mixing the crude urease solution with cementation solution (equal mole masses of CaCl2and urea)to achieve a final concentration of 0.75 mol/L.

(5) 100 mL EICP solution(slightly more than one pore volume of sand column)was injected into the sand column from top to bottom by a peristaltic pump at a rate of 1 mL/min,and several EICP treatments were performed at a 24-h interval.Nine cycles of EICP treatment were conducted in this test.The EICP solution was allowed to freely drain from polyvinyl chloride (PVC) tubes during the bio-grouting.

(6) Distilled water was used to flush the bio-cemented sea sand column to remove the residual substances.All specimens were flushed by 100 mL distilled water for five times.

After specimen preparation was completed,the sea sand columns were placed in a natural environment for one week and then dried in the oven at 60°C for 48 h.

2.3.Test procedure for D-W durability of bio-cemented sea sand columns

The test procedure for D-W durability of bio-cemented sea sand columns is shown in Fig.3.The D-W cycle adopted in the test consists of a wetting process and a drying process.Up to date,there is still no standard testing method for EICP-treated specimens under D-W cycles.In this study,Chinese national code GB/T 50266-2013 (2013) was adopted for D-W cycles of EICP-treated sea sand columns.These specimens were treated with 1-15 D-W cycles.The EICP-treated sea sand columns were placed in a tank filled with distilled water with a pH vale of 5.72.The distilled water was injected into the tank at a 2-h interval,and the water level rose by 1/4 of specimen height after each injection (i.e.specimens were completely submerged by the distilled water after 8 h).The specimens were submerged in the distilled water for 42 h and then oven-dried for 48 h to permit the excess moisture dissipation.The specimens subjected to 0,5,10 and 15 cycles were firstly scanned by the micro-CT with a resolution of 6.5 μm.Thereafter,uniaxial compressive strengths of EICP-treated specimens were tested,and their micromorphology was observed by 3D X-ray microtomography (V|tome|x S240 CT).The vertical tomographic interval of each scanning was set to 25 μm.A total of 4000 images of CT scanning can be obtained for each bio-cemented sea sand column.The schematic diagram of the experimental procedure is shown in Fig.3.

Fig.3.Schematic diagram of the experimental procedure.

2.4.Preprocessing of images obtained by micro-CT

The bio-cemented sea sand columns appear to have multiphase,such as pore,sand and calcium carbonate.To quantitatively characterize the volume of each phase,the original CTscanning images need to be properly digitalized for the later 3D reconstructions (Pare et al.,2016;Zhang et al.,2017;Carmignato et al.,2018).Additionally,the original CT-scanning images also need to be preprocessed for better image clarity and to eliminate irrelevant information (Xiong et al.,2020).The preprocessing of images in this study was composed of three steps: (1) Histogram equalization of original CT-scanning images to improve image contrast;(2) smooth filtering implemented on the image that has been processed by histogram equalization;and(3)segmentation of images with the Otsu method(Otsu,1979;Houston et al.,2013).The Otsu algorithm would determine the segmentation threshold between the pores and solid phase(sea sands and calcium carbonate).The segmentation results of the CT-scanning image are shown in Fig.4.It should be noted that four typical smooth filtering methods were implemented in this study.It can be observed from Fig.4 that the clarity of images processed by bilateral filtering is the best among these four methods.Therefore,the bilateral filtering method was used in this study.The three steps of image preprocessing were conducted using MATLAB software.

Fig.4.The procedure of CT-image preprocessing.

2.5.3D reconstruction and parameter extraction of EICP-treated sea sand columns

Avizo? software was implemented in this study to conduct 3D reconstruction and parameter extraction of EICP-treated sea sand columns.The 3D models of pores and solid phase within the EICPtreated sea sand columns were reconstructed based on the marching cube (MC) algorithm built in the Avizo? software.The pores in each EICP-treated sea sand column were extracted and visualized to study the geometrical shape and spatial distribution of pores with different D-W cycles.The volume,diameter,shape factor,and dip angle were extracted by the Avizo?software based on the binary images derived from image preprocessing (Wei et al.,2019) to further quantify the deterioration characteristics of EICPtreated sea sand columns under the effect of D-W cycles.The parameters used in this study and their formulae are summarized in Table 1 (Li et al.,2019).

Table 1Pore parameters and formulae in the 3D reconstruction model.

2.6.GRA

GRA is a multivariate analysis method used to study the intimacy of research objects based on qualitative analysis (?inici et al.,2022;Lv et al.,2022).In this study,the gray relational degree was employed to evaluate the effect of each microstructure parameter on the macro-mechanical properties.The GRC and GRG were calculated to evaluate the relation between the macromechanical properties and microstructure parameters.The larger the correlation coefficient,the stronger the sequence relationship,and the greater the influence(Han et al.,2022).Generally,the main calculation process of GRA includes four steps(Wu et al.,2022):(1)Establishment of the reference sequence and comparison sequence according to the related research;(2)normalization of the test data;(3)calculation of the mean absolute difference;and(4)calculation of GRC and GRG.The specific calculation process and the determination of value distinguishing efficiency can refer to the previous literature (Rahiman et al.,2021;Srinivasan et al.,2021;Wu et al.,2022).

3.Results and discussion

3.1.Effect of D-W cycles on weight loss rate and surface morphology of EICP-reinforced specimens

The weight of specimens was measured after each D-W cycle to investigate the weight loss rate of specimens after undergoing different D-W cycles.The images embedded in Fig.5 show the surface morphology variations of EICP-treated specimens with the action of cyclic D-W treatments.For the specimens without D-W treatment,the average calcium carbonate content (CCC) is 14.5%-15.6% with a relatively uniform distribution along the specimen height.The calcium carbonates precipitated in the EICP treatment bond the loose sea sands together.The pores on the surface are filled with calcium carbonates and form a white protective layer for the column.For the specimens undergoing 4 D-W cycles,the surface covered by white precipitation was significantly reduced.No disintegration occurred although many pores were exposed.With the increasing number of D-W cycles,the white protective layer nearly faded away and disintegration occurred in the bottom part of the column.The result indicates that the CCC in the bottom part of EICP-reinforced columns primarily is slightly lower than that in the upper and middle parts.The lower CCC means that fewer effective calcium carbonate crystals exist in the specimens,which intensifies the disintegration of specimens.For the specimens undergoing 12 D-W cycles,significant disintegration occurred in the major part of the column and the diameter of specimens decreased from the initial average value of around 50 mm-47.6 mm.The diameters of the upper,middle and bottom parts of column were measured after each cycle of D-W treatment.The average of the three values was taken as the average diameter of column.

Fig.5.Schematic diagram of the surface morphology variation.

The disintegration is directly reflected in the weight loss of the tested specimens during the D-W cycles.Fig.6 shows the weight loss rate of specimens after each D-W cycle.Unlike the linearly accumulated weight loss of gypsiferous specimens in the D-W cycle(Jiang et al.,2021,2022b),the weight loss rate of EICPtreated sea sand column increases nonlinearly in the D-W cycle.The weight loss of gypsiferous specimens is caused by the dissolution of gypsum.Its accumulated weight loss increases linearly with increasing number of W-D cycles,meaning that the weight loss of gypsiferous specimens in each W-D cycle is relatively stable.However,the weight loss of EICP-reinforced specimens in each D-W cycle is unstable.This is because the EICP-reinforced sea sand grains are bonded together by the calcium carbonate crystal precipitated on the contact point.Once the bonding fails due to water erosion,large-size sea sand grains would fall off intensively,resulting in the significant increase of weight loss.Therefore,the accumulated weight loss of EICPtreated sea sand column increases nonlinearly with increasing number of D-W cycles.In this study,the growth rate of weight loss rate is introduced to investigate different stages of specimen weight loss.As expected,the changes in the weight loss rate can be divided into three stages.The weight loss rate reaches its maximum before 3 D-W cycles and then decreases slightly until 7 D-W cycles were implemented.The non-effective calcium carbonate precipitations on the surface of the specimen would rapidly fall into the suspension after water penetrated and inundated the material.The calcium carbonate attached to the surface of sea sands without forming effective bonding of sea sand grains would be easily eroded in the first few numbers of DW cycle,resulting in the greater weight loss rate in the early stage of D-W cycles.Subsequently,the growth rate of the weight loss rate increases again,meaning that more solid phase was eroded by the D-W cycle in stage II.As shown in Fig.5,unbound sea sands and many small sea sand aggregates have fallen.Therefore,the growth rate of weight loss increases in this stage.Ultimately,the growth rate reaches a plateau in stage III,and the accumulated weight loss of specimens increases linearly with the increasing number of D-W cycles in this stage.A similar observation has been made by Liu et al.(2021) in their study of MICPtreated sandy soils against acid rain.The results implied that the disintegration in stage III is relatively more severe than that in stage II because major sea sand cluster near the specimen surface has been eroded in stage II.The non-effective calcium carbonate crystals and unbound sea sand particles were constantly eroded by the dilute water.It should be noted that the unstable calcium carbonate near the inner part of the specimens could be more easily eroded in stage III.Deionized water will penetrate the inside of the specimen once the hard shell layer on the surface of the column is eroded.Therefore,it is reasonable that the growth rate of weight loss in stage III is higher than that in stage II.

Fig.6.Weight loss rate and growth rate of weight loss vs.the number of D-W cycles.

3.2.Variation of parameters for characterizing the pores in each DW cycle

In the 3D model reconstructed by the Avizo software,the parameters of pores can be obtained to observe the microstructure variation of EICP-treated sea sand column.As discussed in Section 3.1,the weight loss rate increases constantly and the growth rate of weight loss increases with the destruction of the “protective layer”.Therefore,the column was characterized into three parts according to the distance from the axis of the specimens,namely the inner,middle and outer parts,respectively (Fig.6a).The CCC,void volume ratio,pore density and average pore volume were measured in different parts of the column after undergoing different D-W cycles.

Fig.7a shows the variation of CCC in the three parts of the EICPtreated column in each D-W cycle.The CCC of all the three parts decreases with the increasing number of D-W cycles.For the original specimen,the outer part has the highest CCC while the inner part has the minimum CCC.The CCC of the inner part changes slightly when the number of D-W cycles is less than 10.A significant decrease in CCC was observed in the outer part of the column in the first 5 D-W cycles.However,the CCC of the middle and inner parts significantly decreases after 5 and 10 cycles,respectively.This suggests that the D-W cycle erodes the calcium carbonate in the outer part in the first few cycles.Under this condition,the weight loss and reduction of CCC in the internal structure are relatively low.The CCC of the inner and middle parts will significantly reduce after a large amount of calcium carbonate is washed with distilled water.Additionally,the variation of pore characteristics in each part was studied during the D-W cycles.The results show that the pore density of the outer part gradually increases within 5 D-W cycles because a great number of pores with smaller size could be generated in the outer part of column as the calcium carbonate crystal here could gradually be eroded in the early stage of the D-W cycles (Fig.7c).As expected,it can be noticed in Fig.7b and d that the void volume ratio and average pore volume increase with the increasing number of D-W cycles.Instead,the pore density of the whole specimen is reduced after D-W cycle.It is suggested that the pores with small sizes in each part are gradually interconnected to form large-size pores.Consequently,there are more large-size pores instead of small-size ones due to the erosion of calcium carbonate.To further study the variation of pore and microstructure after different D-W cycles,the pore size distribution and morphology of pores in EICP-treated sea sand columns are discussed in the following section.

Fig.7.Variation of pore characteristic parameters in different parts of EICP-treated specimens.

3.3.Effect of D-W cycles on pore size distribution characteristics

Fig.8 shows the pore size distribution of EICP-treated columns subjected to different D-W cycles.It can be observed that the pore size is mainly concentrated in the range of 10-1000 μm.The pore size in this study was classified according to Huoduote’s classification method (Huoduote 1966;Liu et al.,2017),namely micropores (1-10 μm),mesopores (10-100 μm) and macropores (100-1000 μm).The peak in the pore size range of 10-100 μm is lower than that in 100-1000 μm,indicating that the pores of EICP-treated specimens are mainly macropores.The variation of the pore volume is mainly caused by the change in the proportion of macropores and mesopores.As shown in Fig.8,after 15 D-W cycles,the percentage of macropores increases by 12.04% and that of the mesopores reduces by 2.6%.This suggests that with the increasing D-W cycle number,the mesopores gradually change into macropores due to numerous calcium carbonate precipitation being eroded.Additionally,it is interesting that the mesopores and macropores increase after 5 D-W cycles.However,in the following D-W cycles,the mesopores increase and macropores reduce with the increasing number of D-W cycles.This result implies that in the first 5 D-W cycles,the mesopores formed because many unbound calcium carbonate precipitations were produced by water erosion.For the micropores,the erosion of unbound calcium carbonate results in the connection of micropores and more mesopores formed correspondingly.This result demonstrates that disintegration has not occurred in EICP-treated specimens yet in the first 5 D-W cycles,as discussed in Section 3.1.The newly generated mesopores and micropores result in the variation of pore porosity.With the increasing number of D-W cycles,the proportion of mesopores and macropores increases,results in the occurrence of disintegration(as shown in Fig.5).

Fig.8.Pore size distribution of EICP-treated sea sand column after undergoing different numbers of D-W cycles.

3.4.Effect of D-W cycles on morphology of pores

In addition to the porosity,the morphology of pores also affects the macro-mechanical properties of EICP specimens.Therefore,the shape factor of pores within specimens after undergoing different numbers of D-W cycles was determined in the Avizo software.In the 3D reconstruction model,four kinds of pore morphology,i.e.ellipsoidal,columnar,branched and irregular,were observed.As shown in Fig.9b,ellipsoidal pores are the main pore type in EICPtreated specimens,followed by columnar and branched pores.With the increasing number of D-W cycles,the ellipsoidal pores reduce by 10.8%,and the columnar and branched pores increase by 5.3% and 2.7%,respectively.To study the mechanism of morphology variation,pores in the section at half the height of specimen were extracted from the 3D reconstructed model for the investigation of pore characteristic.As shown in Fig.9c,few connected pores were observed in the original specimen,and numerous ellipsoidal pores were observed during the calcium carbonate precipitation.With the increasing number of D-W cycles,the calcium carbonate around the ellipsoidal pores was generally eroded.The adjacent disconnected pores would be connected due to the effect of D-W cycles.The void left by the eroded calcium carbonate makes ellipsoidal pores change into columnar pores,which increases the moisture in the specimens and intensifies the erosion of calcium carbonate within specimens.With the increasing number of D-W cycles,the columnar pores change into branched pores and further influence the mechanical properties of EICP-reinforced specimens.

Fig.9.Morphology of pores and variation of pores after undergoing D-W cycles: (a) Morphology of pores;(b) Shape factor of pores;and (c) Different morphologies of pores.

Pore dip angle can also be used as an intuitive parameter to characterize the pore development direction (Li et al.,2019).The dip angle is the most intuitive parameter to characterize the pore orientation because pores with different characteristics would have different orientations.The variation of dip angle in this test intuitively indicates the change of pore orientation as well as pore characteristic (Li et al.,2018;Wei et al.,2019).To investigate the pore development direction in each part of EICP-treated specimens after undergoing different numbers of D-W cycles,the pore dip angle of EICP-treated specimen was extracted from the 3D reconstruction model in Avizo software.As shown in Fig.10,before the DW cycle,the number of pores with a dip angle of 22.5°-45°accounts for about 60% of the total pores in the entire specimen.In the first 5 cycles,the number of pores with a dip angle of 22.5°-35°reduces and that of the pores with a dip angle of 75°-90°increases from 12% to 18%.However,with the increasing number of D-W cycles,the number of pores with a dip angle of 75°-90°increases constantly and that of the pores with a dip angle of 30°-37.5°reduces.This result indicates that the development direction of pores gradually deflects to the vertical direction.Similar changes were observed in the inner,middle and outer parts,namely the number of pores with a large dip angle increases and that of the pores with a small dip angle reduces.The increasing dip angle of pores validates that the ellipsoidal pores gradually change into columnar and branched pores.

Fig.10.Rose diagram of pore dip angle of EICP-treated specimen after undergoing different numbers of D-W cycles (N).

3.5.Effect of D-W cycles on macro-mechanical properties of EICPreinforced sea sands

The UCS results of EICP-treated specimens subjected to different D-W cycles are discussed in this section.The axial stress vs.axial strain curves of the specimens experiencing different numbers of D-W cycles are presented in Fig.11.It can be found that the D-W cycle significantly affects the stress-strain curve of the EICPtreated specimens.The results show that the UCS of EICP-treated specimens decreases gradually with the increasing number of DW treatments.The UCS decreases rapidly when 5 D-W cycles were implemented.As discussed in Section 3.1,the weight loss rate is relative greater in the early stage of D-W cycles.This result indicates that a high number of unbonded crystals are eroded by the D-W cycles to form more pores,resulting in high weight loss rate and reduction of UCS.However,with a further increasing number of D-W cycles,the decrease of UCS is relatively slower.In this stage,the macro-mechanical properties of EICP-reinforced sea sands are controlled by the effective crystals precipitated at the particleparticle contact.This type of crystals generally has stronger bonds than others,which has good resistance ability to the water erosion during the D-W cycles.Therefore,a slight strength loss was observed in this stage of D-W cycles.The prominent decrease in UCS of EICP-treated specimens and the significant increase of micropores (Fig.8) in the first 5 D-W cycles indicate that the decreasing UCS is mainly associated with the micropores generated in the EICP-treated specimens.Failure of effective bond at the particle-particle contact contributes to the reduction in the UCS of EICP-reinforced sea sand,especially for the stages after 5 D-W cycles.

Fig.11.(a) Unconfined compression stress-strain curves of EICP-treated specimens with the effect of different cyclic D-W treatments;and (b) Failure model of EICP-treated specimens after undergoing several cyclic D-W treatments.

In addition,the failure strain of EICP-treated specimens decreases with the increasing number of D-W cycles.The results indicate that the EICP-treated specimens becomes more brittle with the increasing number of D-W cycles,and the peak strength of specimens decreases and tends to be stable after 10 D-W cycles.From the stress-strain curves,it can be found that the failure curves of EICP-treated specimens after undergoing different numbers of D-W cycles are essentially the same,and the post-peak behavior of all specimens tends to be ductile.Similar results were also reported by Liu et al.(2019).It is concluded that the service life of EICP-treated specimens is relatively short and it is critical to improve the specimen durability against cyclic D-W treatment when used in field implementation.

3.6.Effect of variation in pore characteristic parameters on UCS and weight loss rate

The GRC,GRG and gray relational order of each pore characteristic parameter with the weight loss rate and UCS as the reference sequences were determined separately,as presented in Table 2.The results in Table 2 indicate that the branched pores have the most obvious effect on the weight loss rate and the macropores have the most obvious effect on UCS.As discussed in Section 3.4,the volume of branched pores is generally larger than that of columnar pores.The morphology of pores changes from ellipsoidal pores into columnar and branched pores with the action of cyclic DW treatment.The calcium carbonate around the ellipsoidal pores in the original specimen is exfoliated under the action of cyclic D-W treatment,which results in the generation of columnar and branched pores.Therefore,it is reasonable to suggest that the weight loss rate is due to the generation of columnar and branched pores.In addition,the gray relational order of average volume ranks third among all the pore characteristic parameters.It means that changes in the average volume could also have a relatively significant effect on the weight loss rate of EICP-treated specimens.The average volume directly reflects the degree of pore development.A similar calculation process was implemented to obtain the gray relational order of each pore characteristic parameter with UCS as the reference sequence.Table 2 shows that the gray relational order of macropore,columnar pore and mesopore rank top three respectively among all the pore characteristic parameters.From the variation of pore morphology observed in the 3D reconstruction model,the frequency counts of branched pores increase with the increasing number of D-W cycles.The deterioration of UCS is obvious in the cyclic D-W treatment.Therefore,it can be concluded that the generation of branched pores with sizes ranging in 100-1000 μm is the main cause of the deterioration of UCS.The gray relational order of each pore characteristic parameter obtained in the study further validates the results and conclusions obtained in the CT scanning,3D reconstruction model and UCS test.

Table 2GRC,GRG and gray relational order of each pore characteristic parameter.

3.7.Water erosion mechanism of EICP-reinforced sea sands subjected to D-W cycles

For the clear expression of our point of view,the schematic diagram of water erosion mechanism in D-W cycles of EICPreinforced sea sands is shown in Fig.12.The loose sea sand particles are bonded together by the CaCO3precipitation at the particleparticle contacts.A great number of CaCO3crystals are also precipitated in the pore space without forming the non-effective bonding at contacts due to the lack of nucleation site (Ahenkorah et al.,2022).At the early stages of D-W cycles,the powder and non-effective bonds rapidly fall into suspension when the water penetrates into the EICP-reinforced sea sands,leaving a great number of ellipsoidal pores within the sea sand column.With the increases of D-W cycles,some of the weak bonds may also be broken due to the increasing matric suction induced by the reduced water content after the EICP-reinforced sea sands are dried.Therefore,the ellipsoidal pores gradually increase in size and change into the columnar pores.In this stage,only the effective bonds at particle-particle contacts work as cementing agent.At the later stages of D-W cycles,the water destroys the effective bonds at particle-particle contacts,which work as the last line of defense to the water erosion.In this context,the fine sea sand particles and effective bonds are eroded during the D-W cycles,resulting in the generation of branched pores.As discussed in Section 3.6,it can be hypothesized that the greater reduction of mechanical properties of EICP-reinforced sea sands is due to the presence of branched pores.The greater pore volume and the less number of effective bonds result in the disintegration and strength loss of EICP-reinforced sea sand column.

Fig.12.Schematic representation of water erosion mechanism in D-W cycles on EICP-reinforced sea sands.

4.Conclusions

In this work,a series of D-W treatments was conducted to investigate the durability of EICP-reinforced sea sand.The effect of D-W cycles on pore characteristics and mechanical properties of EICP-reinforced sea sand was investigated using 3D reconstruction,CT scanning technique,and UCS test.The gray relational degree was introduced to reveal the potential correlations between microstructure characteristics and macro-mechanical properties of EICPreinforced sea sands.The conclusions of the present study are summarized as follows:

(1) The weight loss of EICP-treated specimen is the greatest at the beginning of cyclic D-W treatment because the noneffective calcium carbonate would be eroded easily by the deionized water.With the increasing number of D-W cycles,the growth rate of weight loss decreases before 8 D-W cycles and increases slightly until reaching a platform after 11 D-W cycles.The pores left by the exfoliated calcium carbonate connect and result in the disintegration of the EICP-treated specimen,which leads to the increasing growth rate of weight loss in stage II.

(2) The pore with sizes ranging in 100-1000 μm and four kinds of pore morphologies were observed in the EICP-treated specimens.D-W cycles could lead to the decrease of pore density and increase of average volume.The mesopores gradually change into macropores due to the exfoliated calcium carbonate in response to the D-W treatment.The development direction of pores gradually deflects to the vertical direction with the increasing number of D-W treatments.The increasing dip angle of pores and decreasing ellipsoidal pores indicate that the small-size pores are gradually connected and then form large pores,such as columnar and branched pores.

(3) The UCS and the peak strain decrease gradually with the increasing number of D-W treatments.The EICP-treated specimens were found to become more brittle with the increasing number of D-W cycles.The strength deterioration of EICP-reinforced sea sand is mainly associated with the macropores generated inside the rock in response to the D-W cycles.

(4) The GRC,DRD and gray relational order of each pore characteristic parameter with the weight loss rate and UCS as reference sequences indicate that the deterioration of the UCS is closely related to the evolution of branched pores with sizes ranging in 100-1000 μm.The branched pores in the EICP-reinforced sea sand facilitate the generation of failure surface in UCS test,resulting in the deterioration of sand mechanical properties.

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 gratefully acknowledge the financial support of National Natural Science Foundation of China(Grant No.41972276),Natural Science Foundation of Fujian Province,China (Grant No.2020J06013),and "Foal Eagle Program" Youth Top-notch Talent Project of Fujian Province,China (Grant No.00387088).