Effect of Fines Content on Calcium Carbonate Precipitation and Thermal Properties of Biocemented Sand Pinar Gunyol, Mohammad Khosravi, A. J. Phillips, Kathryn Plymesser, Albert E. Parker This material may be downloaded for personal use only. Any other use requires prior permission of the American Society of Civil Engineers. This material may be found at https://doi.org/10.1061/ JGGEFK.GTENG-11925 Made available through Montana State University’s ScholarWorks Effect of Fines Content on Calcium Carbonate Precipitation and Thermal Properties of Bio-cemented Sand by Pinar Gunyol, S. M. ASCE Department of Civil Engineering Montana State University, Bozeman, MT 59717 Mohammad Khosravi, Ph.D., M. ASCE Department of Civil Engineering Montana State University, Bozeman, MT 59717 (Corresponding author: email: mkhosravi@montana.edu, phone: 7202890497) Adrienne Phillips, Ph.D. Department of Civil Engineering Montana State University, Bozeman, MT 59717 (email: adrienne.phillips@montana.edu) Kathryn Plymesser, Ph.D., P.E. Department of Civil Engineering Montana State University, Bozeman, MT 59717 (email: kathryn.plymesser@montana.edu) Albert Parker, Ph.D. Center for Biofilm Engineering Department of Mathematical Sciences Montana State University, Bozeman, MT 59717 (email: parker@math.montana.edu) Submitted May 2023 ABSTRACT 1 In this study, the impacts to soil thermal properties during and after bio-cementation via 2 Microbially-Induced Calcite Precipitation (MICP) method on silty silica sand specimens with 3 varying fines content (0%, 5%, and 15%) were investigated. Firstly, calcium conversion was 4 measured after each pulse, then the MICP treated specimens were tested for cementation 5 uniformity. The evolution of thermal conductivity of silty soils with the MICP treatment was 6 assessed using a TR-3 thermocouple probe. The results show that thermal conductivity of silty 7 sands increased by 17% for specimens treated to 9.7% CaCO3. The improvement in thermal 8 conductivity was attributed to the formation of calcium carbonate bridges binding the soil grains 9 together. The results suggested that the thermal conductivity of silty soil depends on water content, 10 the number of treatment pulses, and the treatment uniformity through the soil specimen. Presence 11 of fines content in the soil was found to play an important role in the distribution and uniformity of 12 biocementation through the soil specimen. However, no statistically significant difference in the 13 thermal conductivity values of MICP treated specimens with different fines content was observed 14 (p > 0.05). The average calcium carbonate content ranged between 10.7% and 7.2% for the soils 15 with 0% and 15% fines content, respectively. The findings of this research could be used to improve 16 the efficiency of geothermal boreholes and other energy geo-structures using MICP by improving 17 thermal conductivity of dry and partially saturated soil. 18 Introduction 19 In recent years, there has been an increasing interest in the use of biological technologies in 20 geotechnical engineering to improve thermo-hydro-mechanical properties of geomaterials 21 (Mitchell et al. 2013). Certain microbes can contribute to altering their surrounding chemical 22 environments to induce the precipitation of calcium carbonate, a form of bio-cement (Van Paassen 23 2009; DeJong et al. 2010; Phillips et al. 2013). One mechanism that has been researched extensively 24 is the microbially or plant produced enzyme urease, which promotes the hydrolysis of urea to alter 25 water content and saturation conditions. In the presence of calcium, this change in saturation 26 induces the precipitation of calcium carbonate (Eqn 1) (Phillips et al. 2016; Hommel et al. 2020; 27 Mitchell et al. 2013; Phillips et al. 2013; Ebigbo et al. 2012). 28 (NH2)2CO + 2H2O + Ca2+ → 2NH4 + + CaCO3 (Eqn. 1) 29 In this process, the precipitation of CaCO3, commonly in the stable mineral form of calcite, can 30 bind together the soil particles that make up porous media. Bio-cemented sand treated by MICP has 31 been reported to provide a reduction in soil settlement (DeJong, Fritzges, and Nüsslein 2006; Van 32 Paassen 2009), an increase in soil shear strength (Ismail et al. 2002; DeJong, Fritzges, and Nüsslein 33 2006; Chou et al. 2011), and an increase in soil stiffness (Montoya and DeJong 2015; Feng and 34 Montoya 2016; Lin et al. 2016). 35 MICP technique has been recently proposed as an alternative for improving the thermal 36 properties of soil (Venuleo et al. 2016; Martinez et al. 2019; Xiao et al. 2021; Wang et al. 2020). 37 Heat transfer in soil takes place primarily through conduction, and flow of heat is largely controlled 38 by the contacts between the solid particles. Precipitation of calcite results in an increase in the solids 39 content and therefore, an increase in the average dry density of the soil through the process of 40 binding particles together at their contacts by calcite bridges (Whiffin, van Paassen, and Harkes 41 2007). Calcite crystals formed around soil particles increase the number of particle contacts and 42 enhance heat transfer paths through solid particles resulting in a higher thermal conductivity of the 43 treated soil (Venuleo et al. 2016). MICP treatment of sand could lead to a significant increase in 44 soil thermal conductivity by up to 250% (Venuleo et al. 2016) particularly in dry and partially 45 saturated soils where the calcite crystals replace the air. The same increase in calcium carbonate led 46 to 330% increase in thermal conductivity at the degree of saturation, Sr = 0%, while the increase in 47 thermal conductivity was 37% at Sr = 100% (Martinez et al. 2020). Previous studies on the effect 48 of bio-cementation on soil thermal behavior have mainly focused on thermal properties of medium 49 sand (e.g., median size of 0.2 mm). The presence of fine-grained soil (e.g., silt) in the soil matrix 50 adds complexity to the applicability of MICP techniques due to smaller pore sizes and larger 51 specific surface areas (Zamani and Montoya, 2017). Thus, the thermal properties of bio-cemented 52 soil as a function of the amount of CaCO3 formed and fines content require further investigation. 53 The effect of grain size on thermal conductivity is primarily linked to number of physical 54 contacts among soil particles (Beziat et al. 1992). The increase in thermal conductivity with 55 increasing fines content is expected to continue until the sand-silt granular mixture reach its 56 maximum dry density. Zhang et al. (2015) mixed sands with different median particle size, Silica 57 12/20 (with median particle size, d50=1.03 mm), Ottawa 20/30 (with d50 = 0.78 mm) and ASTM 58 graded (d50 = 0.36 mm) and observed that the thermal conductivity of sand improved with fines 59 content. The results suggested that thermal conductivity of the mixed sand initially increases up to 60 a critical fines content (Fcr) due to more efficient interparticle contacts and contacts per unit volume. 61 For the coarse sand (Silica 12/20 with median particle size, d50=1.03 mm), the critical fines content 62 was 70% and a 30% increase in thermal conductivity was observed. Although presence of fines 63 could improve thermal properties of soil, previous studies have indicated that homogeneous 64 distribution of bacterial cells in fine-grained soils or soils with fines content could be harder to 65 achieve due to the reduction in pore spaces and the increase in filtration effect (Rebata-Landa and 66 Santamarina 2006; Zamani 2017). Based on Zamani (2017), cementation distribution in silty 67 Nevada sand with 15% silt is less uniform compared to the clean Nevada sand (d50 = 0.15mm). 68 This paper summarizes the results of a series of column experiments on the effect of fines 69 content on bio-cementation process and thermal properties of bio-cemented silty sand. The 70 ureolytic bacterium Sporosarcina pasteurii was used as a source of urease enzyme and calcium and 71 urea containing solutions were injected into sand-filled columns to promote MICP. Thermal 72 conductivity was monitored to evaluate the alteration in the soil matrix as mineral forms in time 73 and its effect on thermal properties of the sand. Following, an assessment of treatment homogeneity 74 (i.e., uniform distribution of CaCO3 through the treated soil matrix) was conducted. The results of 75 the experiments and analyses provide insights into the thermal behavior of bio-cemented silty sands. 76 Materials and Methods 77 The laboratory testing program characterized soil samples of Ottawa (F-65) Sand (US Silica, 78 Ottawa, Illinois) mixed with 0%, 5%, and 15% silt (Sil-Co-Sil 250 obtained from U.S. Silica in 79 2020) (by weight). Figure 2 presents the grain size distributions and SEM images for both Ottawa 80 sand and Silica silt. Ottawa sand is a uniform white silica sand classified as SP (USCS ASTM 2487) 81 with rounded grains, a quartz content of 99.7%, and minimal fines (US Silica 2016). Properties of 82 the uniform sand included the following: maximum void ratio, emax = 0.83; minimum void ratio, 83 emin = 0.51; specific gravity of solids, Gs = 2.65; median particle size, d50 = 0.20 mm. Sil-Co-Sil 84 250 is a highly granular, non-plastic silt with a quartz content of over %99.5. The material is low 85 in moisture, chemically inert, and acid resistant. Properties of the silt material included the 86 following: Gs = 2.75; plasticity index, PI=0 (Price, 2018, U.S. Silica). Table 1 summarizes the emax, 87 emin, and Gs of silty sand with silt content of 0, 10, and 15%. emax and emin were determined following 88 ASTM D4253 (ASTM 2002a) and ASTM D4254 (ASTM 2002b), while Gs was determined 89 following ASTM D792. 90 All soil specimens were prepared in columns with an inner diameter of 50-mm and a height 91 of 120 mm. The soil samples were constructed layer by layer; each layer was compacted to a target 92 relative density of Dr = 50%, up to the top. The amount of soil required for each lift was determined 93 based on the desired density. Compaction was performed in eight lifts with moist soil (a moisture 94 content of 7%) distributed evenly for each layer. To prevent excessive compaction of lower layers, 95 Ladd’s (1978) compaction method was used. A portable, battery-operated TEMPOS Thermal 96 Properties Analyzer and a TR-3 thermal probe were used to measure thermal conductivity of soil 97 samples. Figure 1 displays a schematic of the system setup and the PolyVinyl Chloride (PVC) 98 column with the thermal probe. The TR-3 sensor is a 100-mm long, 2.4-mm diameter needle that 99 was inserted into the specimen through the cap and kept inside the column throughout the 100 experiment (Figure 1). The design allowed for evaluation of thermal conductivity over time as 101 mineralization takes place in the sand matrix, without causing disturbance in the soil. A minimum 102 of 10 mm of soil material parallel to the TR-3 sensor probe in all directions was required to avoid 103 errors in measuring thermal properties of the soil. The inside surface of the PVC column was 104 covered with low-friction Teflon tape (Low-Friction FEP Tape, McMaster-Carr) to minimize the 105 attachment of the sand to the column. Two non-corrosive wire mesh filters were placed at the 106 bottom and top of the sand specimen to prevent intrusion of the sand material into the injection or 107 effluent tubing. 108 Thermal experiments were conducted in three stages. After the soil sample was prepared in 109 the column, the TR-3 sensor was inserted into the untreated moist soil and the thermal conductivity 110 of the soil was measured (Stage 1). The samples were then saturated with deionized water and the 111 thermal conductivity of the saturated soil was recorded (Stage 2). The saturated columns were 112 subjected to MICP treatments. During the treatment phase, the thermal conductivity of the 113 specimens was recorded in 2-hour intervals between each cementation treatment and 16-hours after 114 biological treatment, while for the saturation stage, the thermal conductivity at the end of each were 115 measured (Stage 3). A control column was treated only with deionized water, without bacteria, 116 urea, or calcium sources. 117 Microbial Growth 118 Prior to injections of treatments, a starter culture was prepared by aliquoting 1 mL of Sporosarcina 119 pasteurii (ATCC 11859) from thawed frozen stock to 100 mL of filter sterilized Brain-Heart 120 Infusion broth (BHI, Becton Dickinson) amended with 2% urea (20 g/L urea, Fisher Scientific). 121 The cultures were incubated overnight at 30°C on a 150-rpm shaker. Following incubation, 1 mL 122 of this culture was added to 100 mL of growth promoting media for 24 hours for use (injection into 123 the column) the next day. This starter was then used to inoculate a bacterial solution for the 124 treatment process. 125 Media 126 The process of MICP is done using an alternating set of fluids introduced into a medium. The 127 bacterial growth solution, BHI + Urea, was used to propagate bacterial cultures for treatment (Table 128 2). The calcium-providing solutions, CMM+ (Table 2), share the same base set of ingredients, but 129 include calcium chloride. 130 Injection Strategy 131 The columns were designed such that pulsed injections with resting batch reaction periods between 132 injections of enzymes and calcium containing solutions were repeatable (i.e., flow rate and 133 hydraulic residence time) between different specimen types. The compacted specimens were 134 disinfected by injecting 1.5 pore volumes of 70% ethanol and 10% bleach. The specimens were 135 then rinsed and saturated by injecting two pore volumes of deionized water prior to MICP treatment 136 phase. Biological and cementation solutions (1.5 pore volumes) were injected using an alternating 137 method (Gunyol et al. 2022). The alternating method involved flipping the soil specimens and 138 injecting the reagents from the top and then the bottom of the column (while ports remained fixed). 139 This method of alternating injection between the column ends by flipping the specimens was shown 140 to be more effective at achieving a uniform CaCO3 distribution along the specimen height 141 compared to other injection methods such as bottom-to-top injection, for the thermal conductivity 142 testing (Gunyol et al. 2022). The volume of injection fluids per pulse were 100 mL for 0% fines 143 specimen, 95 mL for 5% fines specimen, and 85 mL for 15% fines specimen, which was based on 144 the calculated pore volumes, as summarized in Table 3. The injection rate was kept constant at 4 145 mL/min throughout the experiments. An injection rate of 4 mL/min was selected for the testing 146 because higher rates caused localized precipitation and weak cementation due to insufficient time 147 for the bacteria to hydrolyze urea, while slower rates allowed more uniform cementation along the 148 column by providing time for greater urea hydrolysis efficiency (Gunyol et al. 2022). The system 149 was not monitored for the pressure increase. A summary of the soil specimens tested in this study 150 and fines content (0%, 5%, 15%, denoted throughout as F00, F05 and F15) for each specimen is 151 presented in Table 3. Specimens denoted with P32 (32 pulses of injections) were subjected to 24 152 pulses of CMM+ and 8 pulses of biological treatments. Specimens denoted with P16 (16 pulses of 153 injections) received 12 pulses of CMM+ and 4 pulses of biological treatments. Table 4 provides the 154 details of the injection program for the treatment solutions. Following the MICP treatments, the 155 specimens were dried in a 40 ⁰C oven for further analysis. 156 Ca2+ concentration determination: Effluent fluid sampling 157 The column design allowed for assessment of the reaction progression (for example, collection of 158 effluent samples of the fluids to assess the urea hydrolysis and subsequent calcium precipitation). 159 A syringe pump (KD Scientific) was used to inject solutions into the column(s). A sample of 160 effluent was collected from the fluids that were in the soil specimens after each batch period. The 161 effluent samples were analyzed for calcium concentration via colorimetric calcium assay and urea 162 concentration via a modified colorimetric Jung assay (Jung et al., 1975; Phillips, 2013). 163 CaCO3 determination: Acid washing cemented specimen 164 CaCO3 content was determined using a gravimetric acid washing method. The specimens were 165 divided into 5 sections along the length. Each section was crushed to powder using a blender. Then 166 1 g sample was added to 10 % (V/V) HNO3 to allow for CaCO3 to be dissolved. The mass of CaCO3 167 was calculated by measuring the difference between oven-dried mass of the soil sample before and 168 after the acid washing process. 169 Microstructure Analysis 170 Samples from MICP-treated columns with 0%, 5% and 15% fines content, and the control were 171 saved for further analysis through microscopy imaging. Scanning Electron Microscopy (SEM) and 172 Energy Dispersive Spectroscopy (EDS) were conducted using JEOL JSM-6100 and Zeiss SUPRA 173 55VP field emission systems (ICAL, Bozeman, Montana), respectively. Voltages used for SEM 174 and EDS imaging were 1.00 KV and 15 KV, respectively. The Energy-Dispersive x-ray 175 Spectrometer (EDS) was used to generate characteristic x-rays to examine individual elements 176 encountered. To ensure accurate imaging of the soil samples without any disruption or distortion 177 caused by charging, all samples were coated with a thin layer of Iridium using a sputter coater. The 178 sputter coater process involved bombarding the surface of the samples with high-energy ions, which 179 deposited a thin layer of Iridium on the surface. The samples were coated for 90 seconds at a current 180 of 20 mA. 181 Analysis of Variance (ANOVA) 182 To determine the effect of: 1) fines content on thermal conductivity, kt, 2) the effect of Ca2+ 183 conversion on the change in thermal conductivity Δkt, and 3) how the effect of Ca2+ conversion on 184 Δkt depends on fines content, an analysis of variance (ANOVA) was performed. The p-value 185 obtained by ANOVA was compared to a significance level of 0.05 to test the null hypothesis, which 186 stated that the thermal conductivities were all equal under all conditions. A significance level of 187 α=0.05 indicates a 5% risk of incorrectly concluding a difference exists (rejecting the null 188 hypothesis) when there is no actual difference. 189 Results 190 The results of a series of column experiments combined with thermal testing are presented to 191 illustrate the effects of fines content on the distribution of bio-cementation and thermal properties 192 of bio-cemented sandy silt. The results include calcium carbonate precipitation profile and the 193 evaluation of thermal conductivity of MICP treated sand specimens (Ottawa F-65) with different 194 fines content (Sil-Co-Sil 250) (0%, 5%, 15%, denoted throughout as F00, F05 and F15) and 195 different pulse treatments (16 or 32 pulses, denoted as P16 or P32). For example, results for 0% 196 fines content and 16 pulses of treatment are denoted by F00-P16 and results for 15% fines content 197 and 32 pulses of treatment is denoted by F15-P32. An ANOVA was also conducted to investigate 198 whether fines content and Ca2+ conversion resulted in a statistically significant difference in thermal 199 conductivity. 200 Effect of fines content on calcium carbonate distribution and precipitation 201 Figure 3 shows the cumulative Ca2+ (g) conversion calculated based on the concentration of calcium 202 data from the effluent samples collected during each pulse compared to the concentration in the 203 injected fluids (influent). The results show that the calculated Ca2+ conversion was the highest for 204 the specimen with 0% fines (~ 22.6 g) and lowest for the specimen with 15% fines (~ 16.3 g). The 205 results from effluent sampling method and post-treatment acid digest method were consistent. The 206 results from acid digestion method indicated that the amount of total CaCO3% present within the 207 specimen was the lowest in the specimen with 15% fines and the highest in the specimen with 0 % 208 fines. The specimen with 0%, 10%, and 15% fines had an average CaCO3 of 10.7%, 9.23%, and 209 7.16%, respectively across the five layers in the specimens that were tested. Lower precipitation in 210 the sample with fines content was expected as the samples with a higher silt content had lower 211 measured pore volume, and thus received lower amount of injection fluids (Zamani, 2017). 212 Measured pore volumes of specimens with 0%, 5%, 15% fines were 65 mL, 62 mL, 55 mL, 213 respectively. 214 The CaCO3 content distributions of specimens with 0%, 5% and 15% silt content for 215 specimens receiving 32 and 16 pulses are presented in Figures 4a and 4b, respectively. For all 216 specimens, the amount of precipitated CaCO3 gradually decreased as the distance from the top and 217 bottom injection ports increased. For the specimens that received 32 pulses, the average amount of 218 CaCO3 present in one gram samples ranged from 14.7%-11.6% for the section near the bottom port, 219 8.3%-3.23% for the mid-section, and 10.7%-8.7% for the top injection port. The observed bio-220 cementation distribution along the soil height could be attributed to the filtration of bacterial cells 221 through the soil particles, as well as encapsulation of bacteria, resulting in a higher amount of 222 CaCO3 precipitation at the top and bottom layers (Whiffin, van Paassen, and Harkes 2007). 223 Filtration of bacteria through silty soils could result in a nonuniform distribution of bacteria through 224 the soil sample with a higher concentration close to the injection port (Zamani and Montoya, 2018). 225 The CaCO3 precipitation pattern is primarily linked to the delivery and efficiency of urease activity 226 of the injected bacteria (Van Paassen 2009; Barkouki et al. 2011). Bacterial cells can act as 227 nucleation sites of mineral formation during MICP (Stocks-Fischer et al. 1999; Kirkland et al. 228 2021). Bacterial cells tend to adhere to particle-particle contacts causing a significant amount of 229 calcium carbonate to precipitate at these locations (DeJong et al. 2010). As the bacterial cells are 230 traveling through the pore spaces, they are filtered through the soil particles with generally a 231 reduction of microbe concentration along the injection path (Feng and Montoya 2016; Ginn et al. 232 2002). 233 The effect of fines content on uniformity of bio-cementation throughout the specimen is 234 evident in Figure 4, which shows that the reduction in cementation in the mid-height of the column 235 becomes more pronounced as the fines content increases. The presence of fines content reduces 236 pore volume, resulting in decreased soil permeability and potentially increased filtration of bacterial 237 cells (Zamani and Montoya 2018). The reduction in soil permeability could result in slower 238 transport of injected fluids through the specimen. Slower flow allow more bacteria to attach to 239 particle surfaces (Foppen and Schijven 2006; Harkes et al. 2010; Cheng et al. 2013; Gunyol et al. 240 2022). Therefore, a higher concentration of bacteria around the injection ports can hydrolyze a 241 greater amount of urea, resulting in localized precipitation of CaCO3 in samples with fines. The 242 experimental results suggest that uniformity in cementation distribution improved by increasing the 243 number of pulses from 16 to 32. The improvement in uniformity was more pronounced for the 244 specimen with 15% fines content. By increasing the number of treatments from 16 to 32 pulses, the 245 CaCO3% (g/g) at the mid-height of specimen F15 increased by 5.5 times in the F15-32P specimen 246 compared to F15-16P, while this increase was only 2.9 times in specimen F00-32P compared to 247 F00-16P. The higher increase of CaCO3% in the specimen with 15% fines might be due to the 248 movement of loose calcium carbonate crystals and cemented silt particles. Moreover, the 249 precipitation around the silt particles could form silt aggregates that might detach from the sand 250 grains (Zamani and Montoya 2018). 251 Effect of calcium carbonate precipitation and fines content on thermal conductivity 252 Figure 5 shows the variation of thermal conductivity (kt) throughout three different phases: 1) initial 253 condition (untreated soil sample with 7% water content), 2) after saturation and before treatment, 254 and 3) MICP treatment phase. During the treatment phase, the thermal conductivity of the 255 specimens was recorded in 2-hour intervals between each cementation treatment and 16-hours after 256 biological treatment, while for the saturation stage, the thermal conductivity at the end of each was 257 measured. The specimens remained saturated throughout the treatment stage. 258 The specimens were initially compacted with a water content of 7%. After compaction, 259 three untreated control soil specimens with fines content of 0%, 5%, and 15% were placed in an 260 oven at 40° C and their thermal conductivities were measured. The dried untreated control soil 261 specimens (0% water content) had thermal conductivity values of 0.22, 0.25, and 0.26 W/mK for 262 fines content of 0%, 5%, and 15%, respectively. Figure 5 also includes the thermal conductivity of 263 an untreated soil saturated with deionized water with kt = 2.48 W/mK as a control sample. This 264 value was similar to the thermal properties of other pore filling fluids (cementation solution with kt 265 = 2.44 W/mK and biological solution without S. pasteurii with kt = 2.52 W/mK). Therefore, the 266 results were only presented for the sample saturated with water. 267 As shown in Figure 5 and Table 5, the initial values of kt for the untreated columns with 7% 268 water content and 0%, 5%, and 15% fines were 1.81, 2.08, and 2.13 W/mK, respectively. The results 269 showed a rapid increase in thermal conductivity when the samples were saturated and the air 270 (kair~0.02 W/mK) in the pore space in the soil matrix is replaced with water (kwater~0.3 W/mK). 271 Under the saturated and untreated conditions, thermal conductivity increased with increasing fines 272 content. In the untreated saturated columns with Ca2+ = 0, it was observed that F15 had the highest 273 mean thermal conductivity with a kt value of 2.75 W/mK, followed by F05 with a value of 2.65 274 W/mK, and F00 with 2.55 (p-value = 0.120). The p-value was greater than the significance level, 275 suggesting that there was no statistically significant difference between the thermal conductivity 276 values of untreated columns with the range of fines content studied here. 277 The thermal conductivity of all specimens increased with the number of pulses compared 278 to the control saturated specimen, as shown in Figure 5. The increase in thermal conductivity was 279 attributed to the reduction in porosity caused by CaCO3 precipitation, which plays a governing role 280 (Yun and Santamarina 2008). The binding of CaCO3 crystals to the grains increases the cross-281 sectional area of contact points between the particles, resulting in a reduction in intergranular 282 thermal resistivity. SEM imaging (shown in Figure 6b and 6d for 16 and 32 pulses of MICP 283 treatment, respectively) revealed that the grains were bonded together by CaCO3 bridges that could 284 act as conductive paths (kcalcite ~5 W/mK), allowing heat to transmit through the inter-grain region. 285 Bio-cementation was observed to form around the silt particles as well as between the larger 286 sand particles in soil columns with fines content, as revealed by scanning electron microscopy 287 (SEM) (Figure 6b and 6d) and energy-dispersive X-ray spectroscopy (EDS) images (Figure 7). This 288 is due to the bacteria's ability to attach to both sand and silt particles and induce the precipitation of 289 calcium carbonate. The degree and uniformity of cementation were found to depend on several 290 factors, including soil properties (Figure 6a vs Figure 6b and Figure 6c vs Figure 6d) and the number 291 of pulses of treatment as shown in Figure 6. Furthermore, the cemented silt particles occupy the 292 spaces between the cemented sand grains, thereby improving the heat transfer between the particles. 293 Figure 8 shows the evolution of thermal conductivity with calculated cumulative Ca2+ (g) 294 conversion. The Ca2+ conversion was determined by subtracting the measured effluent Ca2+ 295 concentration from the initial Ca2+ injected (CMM+, cementation solution). The increase in Ca2+ 296 conversion implies a higher degree of CaCO3 precipitation, which results in an increase in the solid 297 mass of the soil matrix. The increase in thermal conductivity under saturated conditions is 298 calculated as Δkn = (ktreated_n – kuntreated), where n is the number of pulses, kuntreated is the thermal 299 conductivity of the saturated untreated specimen, and ktreated_n represents the thermal conductivity 300 of saturated treated specimen after n pulses. A correlation between the calcium conversion and the 301 increase in thermal conductivity was observed for all specimens with varying fines content. The 302 change in thermal conductivity (Δkt) increased by 17 W/mK for every 1000g Ca2+ converted (p-303 value < 0.0001), indicating that there was a statistically significant difference between the thermal 304 conductivity values of columns with Ca2+ conversion. 305 Effect of Fines Content on Thermal Properties of Bio-cemented Soil 306 The effect of fines content on thermal properties of bio-cemented soil was investigated using: 1) 307 the change in the mean thermal conductivity with the number of pulses (Figure 5), 2) the rates of 308 increase in thermal conductivity with the number of pulses (Figure 5), and 3) the change in mean 309 Δkt values with Ca2+ conversion (Figure 8). As shown in Figure 5 and summarized in Table 5, F15-310 P32 resulted in the highest mean thermal conductivity value of 2.88 W/mK, followed by F05-P32 311 with a value of 2.79 W/mK and F00-P32 with 2.68 (p-value = 0.118). The rates of increase in 312 thermal conductivity (kt) of the columns were also not statistically significant, with values of 0.016 313 for F15, 0.019 for F05, and 0.017 for F00 (p-value of 0.259). The mean Δkt values were calculated 314 as 0.150 W/mK for F00, 0.141 W/mK for F05, and 0.138 W/mK for F15 (p-value ≥ 0.814). In all 315 three cases, the p-value was greater than the significance level, indicating that there was not enough 316 evidence to suggest that the effect of bio-cementation on the increase of thermal conductivity (Δkt) 317 was dependent on the range of fines content studied here. 318 Although the results failed to show a statistically significant effect on the thermal 319 conductivity of treated columns with different fines content, the SEM images of the specimens 320 F15P16 (Figure 6b) after 16 pulses and F15P32 (Figure 6d) after 32 pulses of MICP treatment 321 showed that the presence of fine particles can help fill gaps between coarser grains, increasing the 322 number of mineral bridges between grains and the number of paths for heat transfer. This process 323 reduces pore size and improves cementation, as the silt particles are cemented to each other and the 324 sand particles. Additionally, the precipitation of calcium carbonate increases soil density and 325 improves soil thermal properties. Most importantly, the cemented silt particles can transfer heat 326 between sand grains without being displaced, which enhances heat flow through the system. For 327 the untreated saturated sample, the difference between the thermal conductivity of samples F00 and 328 F15 was 0.25 W/mK. For the bio-treated sample, the difference between the two (F00-32P and F15-329 P32) was only 0.14 W/mK, suggesting that the effect of fines content on thermal conductivity was 330 reduced by increasing the number of pulses which could be attributed to the reduction in porosity 331 due to CaCO3 precipitation. 332 Discussion 333 Previous studies investigating the effect of microbial-induced calcium carbonate precipitation 334 (MICP) on the thermal properties of soils have focused exclusively on sand, with contributions 335 from the authors listed in Table 6. These studies have examined the impact of MICP bonding 336 between soil particles (Venuleo et al. 2016; Wang et al. 2020), degree of saturation (Martinez et al. 337 2019; Xiao et al. 2021), and gradation and porosity (Xiao et al. 2021) on thermal properties of 338 MICP-treated soils. However, the influence of fines content on cementation uniformity and thermal 339 conductivity of MICP-treated sand specimens are not fully understood. As discussed in the previous 340 sections, fines content could influence the precipitation of calcium carbonate and uniformity of 341 MICP treated soil, the results. Nevertheless, our results show that there is no statistically significant 342 difference in thermal conductivity values among untreated columns with varying fines content. 343 Table 6 summarizes the details of previous studies including soil properties (Dr and d50), 344 thermal testing method, specimen dimension, and the method used to calculate the average CaCO3 345 content in the specimen. Figure 9 plots the changes in the thermal conductivity (Δk = ktreated – 346 kuntreated) obtained in this study versus average CaCO3% in and compares them with those obtained 347 by Martinez et al. (2019) (Figure 9a), Xiao et al. (2021), and Wang et al. (2020) (Figure 9b). ktreated 348 and kuntreated are thermal conductivities of saturated treated and saturated untreated specimens. The 349 average CaCO3% presented for this study was calculated based on the amount of CaCO3% present 350 in the soil along the thermal probe length. CaCO3 content was defined as the ratio of the mass of 351 CaCO3 to the mass of oven-dried MICP-treated soil (Figure 9a) (similar to Martinez et al. (2019)), 352 while Xiao et al. (2021) and Wang et al. (2020) defined the CaCO3 content as the ratio of the mass 353 of CaCO3 to the mass of MICP-treated soil after HCl dissolution (Figure 9b). 354 The results indicate that there is a positive correlation between the increase in thermal 355 conductivity (Δk) and the amount of CaCO3 precipitation in the soil matrix. The maximum Δk after 356 bio-cementation was 0.44 W/mK, representing a 17% increase in thermal properties, for 9.7% 357 CaCO3 in this study. The results of this study are consistent with those obtained by Martinez et al. 358 (2019) for silica sand with a mean particle size of 0.26 mm, where an increase of 0.44 W/mK was 359 observed for 6.72% increase in CaCO3 content. Direct comparison of the results with those obtained 360 by Xiao et al. (2021) and Wang et al. (2020) is challenging because different testing methods were 361 used. Xiao et al. (2021) utilized a hot disk thermal analyzer through a single-sided heat-transfer 362 process, while Wang et al. (2020) used a dual probe heat pulse (DPHP) sensor. In all cases, 363 however, an increase in CaCO3 content results in a significant increase in Δk, indicating that MICP 364 treatment could effectively enhance the thermal conductivity of sand with fines content. 365 Conclusion 366 A series of column experiments was conducted to investigate the effect of fines content on bio-367 cementation process and thermal properties of bio-cemented silty sand. Five soil columns were 368 subjected to MICP treatment with alternating injection method, resulting in a similar cementation 369 pattern within the columns for all specimens with higher concentration at the injection port and 370 lower concentration at the middle of the column. Measurements from effluent sampling and post-371 treatment acid digest methods showed less mineral with increasing fines content. The average 372 calcium carbonate content ranged between 7.2% for F15-16P with 15% fines content and 10.7% 373 for F00-32P with 0% fine. This was expected as the samples with higher silt content have lower 374 pore volume and received lower amount of injection fluids. For all specimens, the amount of 375 precipitated CaCO3 gradually decreased as the distance from the top and bottom injection ports 376 increased. This was attributed to the filtration of bacterial cells through the soil particles, resulting 377 in a higher cementation near the injection points. The effect of filtration became more prominent 378 as the fines content in the specimens increased. 379 Measurements of thermal conductivity were obtained prior to and during MICP treatment. 380 Thermal conductivity increased with the number of pulses for all specimens compared to the control 381 specimen. Under saturated and untreated conditions, a 15% increase in fines content resulted in an 382 approximately 11% increase in thermal conductivity. This increase in thermal conductivity is 383 expected to continue until the sand-silt granular mixture reaches its maximum dry mass density. 384 The difference in thermal conductivity between samples F00 and F15 was reduced by 1.8 times 385 after MICP treatment, compared to their initial saturated state. In the untreated dry and initial 386 conditions, the specimen with 15% fines (F15) had over 20% higher conductivity than the specimen 387 without fines (F0), likely due to improved particle packing. However, when saturated and treated 388 with MICP, F15 had only a 4% higher conductivity than F0, suggesting that the effect of fines 389 content on thermal conductivity can be mitigated by increasing the number of pulses. This observed 390 behavior was attributed to the reduction in porosity resulting from CaCO3 precipitation. It appears 391 that the increased filtration effect resulting from reduced pore size in the presence of fine particles 392 was found to have a negligible impact on the effect of bio-cementation on the thermal conductivity 393 of the soil. 394 The results presented herein suggest that MICP treatment can be a viable option to increase 395 the thermal conductivity of soils within the range of fines content studied here (less than 15%), 396 which can enhance the efficiency of geothermal systems. 397 Data availability statement 398 Some or all data, models or code generated or used during the study are available upon reasonable 399 request. 400 Acknowledgement 401 The authors appreciate the assistance of the staff of the center for biofilm engineering (CBE) at 402 MSU. 403 References 404 Barkouki, T. H. , B. C. Martinez, B. M. Mortensen, T. S. Weathers, J. D. De Jong, T. R. Ginn, N. 405 F. Spycher, R. W. Smith, and Y. Fujita. 2011. 'Forward and Inverse Bio-Geochemical Modeling 406 of Microbially Induced Calcite Precipitation in Half-Meter Column Experiments', 90. 407 Beziat, A., Dardaine, M., Mouche, E. 1992. Measurements of the thermal conductivity of clay-sand 408 and clay-graphite mixtures used as engineered barriers for high-level radioactive waste disposal. 409 Applied clay science, 6(4), 245-263. 410 Cheng, L., R. Cord-Ruwisch, and M.A. Shahin. 2013. 'Cementation of sand soil by microbially 411 induced calcite precipitation at various degrees of saturation', Canadian Geotechnical Journal, 412 50: 81-90. 413 Chou, Chiung-Wen, Eric Seagren, A. Asce, Ahmet Aydilek, M. Asce, and Michael Lai. 2011. 414 'Biocalcification of Sand through Ureolysis', Journal of Geotechnical and Geoenvironmental 415 Engineering, 137. 416 DeJong, Jason T., Michael B. Fritzges, and Klaus Nüsslein. 2006. 'Microbially Induced 417 Cementation to Control Sand Response to Undrained Shear', Journal of Geotechnical and 418 Geoenvironmental Engineering, 132: 1381-92. 419 DeJong, Jason T., Brina M. Mortensen, Brian C. Martinez, and Douglas C. Nelson. 2010. 'Bio-420 mediated soil improvement', Ecological Engineering, 36: 197-210. 421 Ebigbo, A., A. Phillips, R. Gerlach, R. Helmig, A. B. Cunningham, H. Class, and L. H. Spangler. 422 2012. 'Darcy-scale modeling of microbially induced carbonate mineral precipitation in sand 423 columns', Water Resources Research, 48. 424 Feng, K. , and B. M. Montoya. 2016. 'Influence of Confinement and Cementation Level on the 425 Behavior of Microbial-Induced Calcite Precipitated Sands under Monotonic Drained Loading', 426 Journal of Geotechnical and Geoenvironmental Engineering, 142: 04015057. 427 Foppen, J. W., and J. F. Schijven. 2006. 'Evaluation of data from the literature on the transport and 428 survival of Escherichia coli and thermotolerant coliforms in aquifers under saturated 429 conditions', Water Res, 40: 401-26. 430 Ginn, Timothy R., Brian D. Wood, Kirk E. Nelson, Timothy D. Scheibe, Ellyn M. Murphy, and T. 431 Prabhakar Clement. 2002. 'Processes in microbial transport in the natural subsurface', Advances 432 in Water Resources, 25: 1017-42. 433 Gunyol, P., M. Khosravi, A. Phillips, and K. Plymesser. 2022. 'Thermal Properties of Bio-434 Cemented Sand.' in, Geo-Congress 2022. 435 Harkes, Marien P, Leon A Van Paassen, Jacco L Booster, Victoria S Whiffin, and Mark CM van 436 Loosdrecht. 2010. 'Fixation and distribution of bacterial activity in sand to induce carbonate 437 precipitation for ground reinforcement', Ecological Engineering, 36: 112-17. 438 Hommel, Johannes, Arda Akyel, Zachary Frieling, Adrienne Phillips, Robin Gerlach, Alfred 439 Cunningham, and Holger Class. 2020. 'A Numerical Model for Enzymatically Induced Calcium 440 Carbonate Precipitation', Applied Sciences, 10: 4538. 441 Ismail, Mostafa, H. A. Joer, Mark Randolph, and A. Meritt. 2002. 'Cementation of porous materials 442 using calcite', Géotechnique, 52: 313-24. 443 Jung, D., Biggs, H., Erikson, J., Ledyard, P.U. 1975. New colorimetric reaction for end-point, 444 continuous-flow, and kinetic measurement of urea. Clinical Chemistry, 21(8), 1136-1140. 445 Kirkland, Catherine M., Arda Akyel, Randy Hiebert, Jay McCloskey, Jim Kirksey, Alfred B. 446 Cunningham, Robin Gerlach, Lee Spangler, and Adrienne J. Phillips. 2021. 'Ureolysis-induced 447 calcium carbonate precipitation (UICP) in the presence of CO2-affected brine: A field 448 demonstration', International Journal of Greenhouse Gas Control, 109: 103391. 449 Lin, H., M.T. Suleiman, D. G. Brown, and E. Kavazanjian. 2016. 'Mechanical Behavior of Sands 450 Treated by Microbially Induced Carbonate Precipitation', Journal of Geotechnical and 451 Geoenvironmental Engineering, 142: 04015066. 452 Martinez, A., L. Huang, and M. G. Gomez. 2019. 'Thermal conductivity of MICP-treated sands at 453 varying degrees of saturation', Géotechnique Letters, 9: 15-21. 454 Martinez, Alejandro, Lin Huang, and Michael Gomez. 2020. 'Enhancement of the thermal 455 conductivity of sands via microbially-induced calcite precipitation', E3S Web of Conferences, 456 205: 09011. 457 Mitchell, Andrew C., Adrienne Phillips, Logan Schultz, Stacy Parks, Lee Spangler, Alfred B. 458 Cunningham, and Robin Gerlach. 2013. 'Microbial CaCO3 mineral formation and stability in 459 an experimentally simulated high pressure saline aquifer with supercritical CO2', International 460 Journal of Greenhouse Gas Control, 15: 86-96. 461 Montoya, B. M., and J. T. DeJong. 2015. 'Stress-Strain Behavior of Sands Cemented by 462 Microbially Induced Calcite Precipitation', Journal of Geotechnical and Geoenvironmental 463 Engineering, 141: 04015019. 464 Phillips, A.J. 2013. Biofilm-induced calcium carbonate precipitation application in the subsurface, 465 Bozeman, Montana : Montana State University. Bozeman, Montana. 466 Phillips, Adrienne J, Alfred B Cunningham, Robin Gerlach, Randy Hiebert, Chiachi Hwang, 467 Bartholomeus P Lomans, Joseph Westrich, Cesar Mantilla, Jim Kirksey, and Richard Esposito. 468 2016. 'Fracture sealing with microbially-induced calcium carbonate precipitation: A field 469 study', Environmental science & technology, 50: 4111-17. 470 Phillips, Adrienne J., Robin Gerlach, Ellen Lauchnor, Andrew C. Mitchell, Alfred B. Cunningham, 471 and Lee Spangler. 2013. 'Engineered applications of ureolytic biomineralization: a review', 472 Biofouling, 29: 715-33. 473 Rebata-Landa, Verónica, and J. Santamarina. 2006. 'Mechanical limits to microbial activity in deep 474 sediments', Geochemistry Geophysics Geosystems - GEOCHEM GEOPHYS GEOSYST, 71. 475 Stocks-Fischer, S., Galinat, J.K., Bang, S.S. 1999. Microbiological precipitation of CaCO3. Soil 476 Biology and Biochemistry, 31(11), 1563-1571. 477 Van Paassen, Leon Andreas. 2009. 'Biogrout, ground improvement by microbial induced carbonate 478 precipitation'. 479 Venuleo, Sara, Lyesse Laloui, Dimitrios Terzis, Tomasz Hueckel, and Mounir Hassan. 2016. 480 'Microbially induced calcite precipitation effect on soil thermal conductivity', Géotechnique 481 Letters, 6: 39-44. 482 Wang, Zhaoyu, Nan Zhang, Jinhua Ding, Qi Li, and Junhao Xu. 2020. 'Thermal conductivity of 483 sands treated with microbially induced calcite precipitation (MICP) and model prediction', 484 International Journal of Heat and Mass Transfer, 147: 118899. 485 Whiffin, Victoria S., Leon A. van Paassen, and Marien P. Harkes. 2007. 'Microbial Carbonate 486 Precipitation as a Soil Improvement Technique', Geomicrobiology Journal, 24: 417-23. 487 Xiao, Yang, Yifan Tang, Guoliang Ma, John S McCartney, and Jian Chu. 2021. 'Thermal 488 conductivity of biocemented graded sands', Journal of Geotechnical and Geoenvironmental 489 Engineering, 147: 04021106. 490 Yun, Tae, and J. Santamarina. 2008. 'Fundamental study of thermal conduction in dry soils', 491 Granular Matter, 10: 197-207. 492 Zamani, A., and B. M. Montoya. 2018. 'Undrained Monotonic Shear Response of MICP-Treated 493 Silty Sands', Journal of Geotechnical and Geoenvironmental Engineering, 144: 04018029. 494 Zamani, Atefeh. 2017. "Liquefaction Mitigation of Silty Sands with Microbial Induced Calcium 495 Carbonate Precipitation." In. 496 Zhang, Nan, Xinbao Yu, Asheesh Pradhan, and Anand J. Puppala. 2015. 'Effects of Particle Size 497 and Fines Content on Thermal Conductivity of Quartz Sands', Transportation Research Record, 498 2510: 36-43. 499 500 Introduction Materials and Methods Analysis of Variance (ANOVA) Results Effect of fines content on calcium carbonate distribution and precipitation Conclusion Copy Cover Page.pdf Blank Page Blank Page