1

2 Manuscript submitted to Journal of Petroleum Science and Engineering

3
4 Title: Biosurfactant-producing bacteria in microfluidics: Pore-scale observations for insights to
5 microbial enhanced oil recovery
6
7 Authors: Taehyung Park, Ph. D.1 , Corresponding Author
8 Postdoctoral researcher, Department of Civil and Environmental Engineering,
9 Korea Advanced Institute of Science and Technology (KAIST), Daejeon, Korea
10 Email: taehyung.park@austin.utexas.edu
11
12 Tae-Hyuk Kwon, Ph.D.
13 Associate Professor
14 Department of Civil and Environmental Engineering
15 Korea Advanced Institute of Science and Technology (KAIST)
16 Email: t.kwon@kaist.ac.kr
17
18 Sheng Dai, Ph. D.
19 Associate Professor, School of Civil and Environmental Engineering,
20 Georgia Institute of Technology, GA, United States
21 Email: sheng.dai@ce.gatech.edu
22
23
24 Corresponding Author: Taehyung Park
25
26 Number of words: 5600
27 Number of tables: 1
28 Number of Figures: 8
29

30 Keywords: Microfluidics, Biosurfactant, Wettability, Enhanced Oil Recovery, Pore-scale observation,

31 Emulsification/Demulsification.
32

33

Electronic copy available at: https://ssrn.com/abstract=4143309

34 ABSTRACT

35 Multiphase flow in porous media is ubiquitous in many natural and industrial processes, including

36 microbial enhanced oil recovery (MEOR). Biostimulation strategies for enhanced oil recovery

37 using indigenous bacteria minimize ecological disturbance and avoid competition over nutrients

38 between exogenous and indigenous microorganisms. However, the difficulties in evaluating the

39 real-time transitions of wettability and ensued impacts on multiphase flow due to biomass growth

40 continue to challenge our microscopic and macroscopic descriptions of MEOR. Here, we examine

41 the wettability alteration during bacterial growth and biosurfactant adsorption at oil-brine

42 interfaces in microfluidic chips (MFC) through high-resolution real-time imaging. The bacterial

43 cell growth and biofilm accumulations throughout bacterial cultivation periods, as well as the

44 emulsification phenomena at the brine-oil systems, are investigated using a single-channel MFC.

45 The results demonstrate that surfactin produced by B. subtilis reduces the interfacial tension (IFT),

46 which results in wettability modification from oil-wet to water-wet. The biosurfactant production

47 substantially reduces both dynamic and static contact angles. In addition, the multi-channel MFC

48 experiment allows the emulation of the surfactant-flooding process and the examination of

49 surfactant concentration on oil sweeping efficiencies. These multi-channel MFC tests show

50 reduced residual oil saturation and contact angle with increased biosurfactant concentration

51 resulting in the homogeneous and stable displacement pattern when the in situ biosurfactant

52 concentration exceeds the critical micelle concentration. Our results provide the fundamental

53 understanding and hence better prediction of the time-lapse MEOR process based on fluid

54 displacement, wettability investigation, and residual oil saturation assessment.

55

Electronic copy available at: https://ssrn.com/abstract=4143309

56 GRAPHICAL ABSTRACT

57

58

59 HIGHLIGHTS:
60  B. subtilis cultivated in a micromodel reduced oil-brine interfacial tension and modified
61 the oil-brine-PDMS wettability.
62  The real-time transitions of wettability during B. subtilis cultivation was monitored in a
63 micromodel.
64  Residual oil saturation in porous media was reduced by biosurfactant in a micromodel.
65  Biosurfactant produced by B. subtilis is an effective mediator for MEOR.

66

Electronic copy available at: https://ssrn.com/abstract=4143309

67 1. INTRODUCTION

68 Many natural and industrial processes center on multiphase flows in porous media, including

69 enhanced oil recovery (EOR). Various EOR strategies have been implemented to improve

70 hydrocarbon production from nearly depleted oil reservoirs.(Lake, 1989; Lake et al., 2014) When

71 recovering oil from a depleted reservoir, the fluid’s affinity to its host porous medium, referred to

72 as wettability, determines the fluid displacement pattern that essentially governs the oil recovery

73 rates.(Kathel and Mohanty, 2013; Wang et al., 2011; Zhang et al., 2006a) Therefore, the wettability

74 among the oil, brine, and rock-forming minerals is critical for quantitatively evaluating the oil

75 recovery efficiencies upon EOR implementations.(Kathel and Mohanty, 2013; Morrow, 1990;

76 Seethepalli et al., 2004; Zhang et al., 2006a; Zhang et al., 2006b)

77 Most tight oil reservoirs are oil-wet or mixed-wet due to long-term exposure to oil.5, 8 To improve

78 the oil mobility in tight reservoirs, surfactants are generally applied to improve the spontaneous

79 imbibition of injecting fluids by altering fluids the rock wettability into water-wet.(Liang et al.,

80 2021; Wu et al., 2008) Additive synthetic surfactants to the injection fluids have proved their

81 effectiveness in EOR practices in transitioning oil-wet reservoirs toward water-wet conditions.

82 Recently, biosurfactants have been tested as an effective alternative to synthetic surfactants due to

83 their intrinsically innocuous and biodegradable nature.(Desai and Banat, 1997; Ron and Rosenberg,

84 2001) Several biosurfactants, e.g., lipopeptides (i.e., surfactin and pumilacidin)(Jacques, 2011) and

85 glycolipids (i.e., rhamnolipids, sophorolipids and trehalolipids)(Benincasa et al., 2004; Daverey

86 and Pakshirajan, 2010; White et al., 2013) have been proved to have similar effects on the

87 interfacial properties as the synthetic chemical surfactants. Among these biosurfactants, surfactin

88 produced by Bacillus subtilis has proved its efficiency in reducing the oil-brine interfacial tension

89 (IFT, ) and the wettability of oil, brine, and rock minerals from an intermediate water-wet ( =

Electronic copy available at: https://ssrn.com/abstract=4143309

90 ~45−50°) to a strong water-wet condition ( = ~20−25°).(Park et al., 2019) The reduced capillary

91 factor (·cos improves the sweeping efficiencies of oil in a porous medium and results in oil

92 recovery enhancement. This observation was confirmed by recent experiments and pore-scale

93 simulations(Kim and Santamarina, 2014; Park et al., 2019; Park et al., 2017). However, the

94 complete range of wetting transitions during biosurfactant production is yet to be fully explored,

95 especially in the regime where the in situ produced biosurfactant attaches at the oil-water interface

96 and modifies the wettability in real-time. It is worthwhile to characterize the real-time transitions

97 of wettability at the pore-scale during biosurfactant accumulation and attachment at the oil-water

98 interface.

99 Biosurfactant aided flooding tests are typically performed at the core-scale with a very limited

100 understanding of the interfacial phenomena at the microscale. Micro-focus X-ray computed

101 tomography is one of the technologies used for pore-scale visualization of wettability alteration

102 during EOR application, such as low salinity water and nanofluids flooding(Ali et al., 2021; Chen

103 et al., 2020; Lebedeva and Fogden, 2011). However, exposure to X-ray sources is by nature lethal

104 to living cells. For this reason, microfluidic devices that are optically transparent are widely used

105 to visualize and understand the pore-scale habits of multiphase fluids flows and in situ emergent

106 phenomena (Beebe et al., 2002; Gravesen et al., 1993; Squires and Quake, 2005). (Fan et al., 2018;

107 Gogoi and Gogoi, 2019; Lifton, 2016; Mohammadi and Mahani, 2020; Saadat et al., 2020)

108 The primary attention of this study was placed on the real-time observations of wettability

109 modification during bacterial growth at the microscale to elucidate the pore-scale fluid sweeping

110 mechanisms. The model organism B. subtilis strain (ATCC21332) was cultured in a patterned

111 micromodel to quantitatively evaluate dynamic (advancing and receding) contact angle transitions

112 during the bacterial growth and byproduct productions. The impact of biosurfactant-induced

Electronic copy available at: https://ssrn.com/abstract=4143309

113 wettability modification on oil displacement efficiency in a porous medium was further evaluated

114 by conducting pore-scale oil displacement experiments using a porous micromodel to mimic the

115 biosurfactant-aided waterflooding processes.

116

117 2. MATERIALS AND METHODS

118 2.1. Micromodel Fabrication

119 a. Single-Channel Micromodel

120 A single-channel micromodel (SCM) was designed to monitor the real-time transition of

121 wettability during in situ production of biosurfactant. The single-channel micromodel comprised

122 of a T-junction where two channels with the width of 50 m and the heignt of 100 m (w × h)

123 intersected, as shown in Figure 1a. This T-junction facilitated the effective droplet formations,

124 which enabled observations on the interface of oil and brine and measurement of the contact angles.

125 Furthermore, as shown in Figure 1c, each channel contained serpentine chennels before the T-

126 junction, which provided better control of fluid flow and more reliable measurements of advancing

127 and receding contact angles when the fluids moved.(Keller et al., 2007; Šikalo et al., 2005) In a

128 porous medium consists of different pore throat size, the measurement of dynamic contact angle

129 becomes unstable that the equilibrium of capillary pressure changes during drainage as a function

130 of time.(Måløy et al., 1992) For this reason, the micromodel made with a single-channel is suitable

131 for dynamic contact angle measurement. Figures 1a and 1c graphically dipect the geometry of the

132 single-channel micromodel.

133 The single-channel micromodel was fabricated by general soft lithography using

134 polydimethylsiloxane (PDMS; Sylgard 184; Dow Corning, USA) based on a conventional

Electronic copy available at: https://ssrn.com/abstract=4143309

135 photolithography process (Hassanpourfard et al., 2014; Jang et al., 2019; Jang et al., 2017). In this

136 study, we fabricated both the top and bottom layers with PDMS to ensure that all the wetted

137 surfaces had the same physicochemical characteristics. The SU-8 master mold for the PDMS

138 model was first made in accordance with the design (by MicroFIT, Korea). PDMS prepolymer

139 solution was poured on the master mold and cured at 85°C for 2 hours, then the PDMS membrane

140 was carefully detached from the mold. The thickness of this top PDMS layer was 2.5 mm. Two

141 inlets and one outlet were trimmed with a 1/16-inch circular punch for the 1/16-inch tubing

142 installation. For the bottom PDMS layer, the same procedure with the top layer was repeated on

143 top of the slide glass (75 × 26 × 1 mm) to produce a thin layer of PDMS with 2.5 mm thickness.

144 Thereafter, the top PDMS layer was placed and attached on the bottom PDMS layer. The fabricated

145 single-channel micromodel was autoclaved and dried before use.

146

147 b. Multi-Channel Micromodel

148 The effect of biosurfactant on oil displacement in porous media was evaluated with a multi-channel

149 micromodel. The width of this micromodel was 7.1 mm, and the vertical length was 3.6 mm, as

150 shown in Figure 1b. The model contained in total 780 uniform cylindrical pillars, each with 150

151 m in diameter. The center-to-center distance between two neighboring pillars was 180 m. The

152 total pore volume of the micromodel was 0.71 mm3. Similar to the single-channel micromodel,

153 two layers of PDMS were attached (5 mm thickness each) on top of the slide glass, making 11 mm

154 of thickness in total. The inlet and outlet, as well as the preparation procedures of the multi-channel

155 micromodel follow the same manner as for the single-channel micromodel.

156

157 2.2. Model Bacterium Selection and Inoculum Preparation

Electronic copy available at: https://ssrn.com/abstract=4143309

158 In this study, B. subtilis strain ATCC 21332 (acquired from the American Type Culture Collection,

159 ATCC) was selected as the model organism. The Bacillus spp. genus is omnipresent in the

160 subsurface and aquatic environments and can survive under harsh environments such as extreme

161 pH, high salinity (Simpson et al., 2011), high temperature, and high pressure (Yakimov et al.,

162 1995). Accordingly, Bacillus spp. have been popularly investigated as the model organisms for

163 evaluating the feasibility of MEOR applications (Al-Bahry et al., 2013; Al-Wahaibi et al., 2014;

164 Park et al., 2017; Peet et al., 2015). Their ability to form endospores supports high survivability

165 when subjected to elevated temperature, desiccation, or chemical disinfectants (Nicholson et al.,

166 2000). More importantly, Bacillus spp. have been frequently found in oil reservoirs(Datta et al.,

167 2018; Korenblum et al., 2005; Simpson et al., 2011), making the Bacillus strain an ideal model

168 microorganism to evaluate MEOR.

169 The mineral salt medium was used as a culture medium, as listed in Table 1.(Park et al., 2017) The

170 mineral salt medium contained adequate amounts of carbon (glucose), nitrogen (nitrate and

171 ammonium) and trace elements (manganese, potassium, and iron, etc.), and this composition was

172 optimized for surfactin production, as suggested in the previous experimental studies.(Cooper and

173 Goldenberg, 1987; Rangarajan and Clarke, 2016; Shaligram and Singhal, 2010; Wei and Chu,

174 2002; Wei et al., 2007) This mineral salt medium was filter-sterilized with a 0.2 µm sterile syringe

175 filter (Chmlab Group, Terrassa-Barcelona, Barcelona, Spain) prior to use.

176 The starter culture was prepared with the cultivation in the nutrient broth (Difco, BD, NJ, United

177 States) at 37°C. From the starter culture, 1%v/v of freshly grown aerobic culture was transferred

178 to the mineral salt medium. The culture was then incubated for 48 hours at 37°C with no shaking.

179 The single-channel micromodel tests used 1%v/v B. subtilis culture-containing brine as the

180 inoculum for in situ cultivation in the model. Meanwhile, the multi-channel micromodel tests used

Electronic copy available at: https://ssrn.com/abstract=4143309

181 an autoclaved cell-free medium as a brine phase to sweep the oil-saturated model. Particularly in

182 the surfactant flooding tests, the cell-free medium was mixed with pre-determined amounts of

183 purified surfactin to achieve different critical micelle concentrations. The critical micelle

184 concentrations of surfactin produced by B. subtilis is ~34 mol/L, which is equivalent to 35

185 mg/L(Park et al., 2019). Therefore, 0, 10.5, 17.5, 35, and 70 mg/L of surfactin solutions were

186 prepared by simply mixing a cell-free medium and purified surfactin stored at low temperature,

187 which represented 0×, 0.3×, 0.5×, 1×, and 2× critical micelle concentrations.

188 Meanwhile, pure surfactin was extracted and prepared using the acid precipitation method (Figure

189 S1 in Supporting Information). The accumulated biosurfactant in the mineral salt medium was

190 extracted, and Fourier Transform-infrared Spectroscopy (FT-IR) qualitatively assessed the class

191 of the produced biosurfactant, as shown in Figure S1b. The functional groups of the extracted

192 biosurfactant were compared with those of the standard surfactin sample acquired from Sigma-

193 Aldrich (Sigma Aldrich, MO, United States). The details of biosurfactant extraction procedures

194 are available in the authors’ previous literature(Park et al., 2017).

195

196 2.3. Experimental Procedure and Data Reduction

197 a. In situ Wettability Observation with the Single-channel Micromodel

198 The transitions of immiscible oil-brine wettability during the growth of biosurfactant-producing

199 bacteria were evaluated using the single-channel micromodel. During the tests, experiments with

200 the micromodel were performed on a DM 500B microscope (Leica, Wetzlar, Germany). The

201 microscopic images were taken every 5 min for 150 hours and with the 2, 4, 10, and 50x objective

202 lens. In addition, videos were occasionally taken to measure the advancing/receding contact angles

Electronic copy available at: https://ssrn.com/abstract=4143309

203 during experimentation periods. The micromodel is equipped with two inlet ports for the injection

204 of mineral oil and the B. subtilis inoculum (Figure 1b). In this study, mineral oil having 15 cP of

205 viscosity was used as the oil phase rather than alkane hydrocarbons which are swelling solvents

206 for PDMS channel(Dangla et al., 2010).

207 Before the tests, all the wetting parts including the PDMS micromodel, tubes and valves were

208 autoclaved. The micromodel was firstly filled with the mineral oil; and the filter-sterilized brine

209 containing B. subtilis culture was then injected into the oil-filled micromodel at a flow rate of 1

210 L/min. During the brine injection, the oil injection was halted. The mineral oil was dyed with

211 Sudan III (0.1 wt%) for better visual distinction between the oil and the brine phases. The chosen

212 flow rate produced a larminar flow with the Reynold number Re of 0.22. For injection, we used a

213 syringe pump (NE-1002X; New Era Systems Inc., NY, USA) with a Norm-ject 5mL Luer lock

214 syringe. Both fluids inevitably passed the fluid resistor intervals for flow velocity control. The T-

215 junction structure facilitated two immiscible fluids in contact each other and created the meniscus,

216 which enabled monitoring of the contact angle of the oil-brine-PDMS system.

217 Throughout microbial growth and biosurfactant production, time-lapsed images of the immiscible

218 mixed fluids (oil and brine) were acquired for ~150 hours. The static contact angles were measured

219 when the pump injection was halted and the fluid-fluid interfaces sat motionless. The dynamic

220 advancing and receding contact angles were measured by pushing and retracting respectively the

221 brine phase through the brine injection pump. Considering that the flow velocity affects the

222 dynamic contact angle, the injection flow rate was fixed at 1L/min and the internal flow velocity

223 of the fluids was kept at less than 50 m/s for reliable and consistent estimates. The static and

224 dynamic contact angles of oil-brine-PDMS in the acquired images were determined using ImageJ

225 software, by drawing a straight line tangent to the fluid meniscus and the other line coincident with

10

Electronic copy available at: https://ssrn.com/abstract=4143309

226 the flat PDMS surface.(Stalder et al., 2006) In addition, the presence of bacterial biomass such as

227 bacterial cells and bacterial biofilms affects the fluid transport. Local accumulations of biomass

228 with B. subtilis appeared in the single-channel micromodel, and this rendered the flow velocity

229 variations during advancing and receding of interfaces, which hampered reliable estimation on

230 moving velocity of the oil-brine interfaces. Therefore, it is worth pointing out that we determined

231 the dynamic contact angles only when oil-brine interfaces were located in the clear channel with

232 no microbial substances, such that there was no obstacle when the fluids advanced or receded.

233

234 b. Water Flooding Test with the Oil-filled Multi-channel Micromodel

235 The water flooding tests were carried out in the multi-channel micromodel to evaluate the role of

236 biosurfactants in decreasing the residual oil saturation in porous media. As graphically depicted in

237 Figure 1d, the top-view images of the micromodel were acquired with a high-resolution camera

238 (Canon EOS 100D, Tokyo, Japan) equipped with a macro lens (Canon 100 mm 2.8f, Tokyo, Japan).

239 Mineral oil was used as an oil phase, and it pre-saturated the multi-channel micromodel to mimic

240 the typical drainage process during EOR practice. The flooding fluid (or invading fluid) was the

241 mixtures of the cell-free brine and purified surfactin with the surfactin concentrations of 0, 10.5,

242 17.5, 35 and 70 mg/L. The flooding fluid was dyed with methylene blue (0.3 wt%) for better visual

243 distinction between the oil and the brine phases. The fluids were filter-sterilized with a 0.2 m

244 filter to remove any biomass or residual substances for better visibility in the micromodel.

245 The test procedure of the surfactant flooding followed the typical tertiary recovery process that

246 extracts the remnant oil from an oil reservoir with the thermal, gas and chemical treatments.(Lake,

247 1989) The water flooding first displaces the oil in the porous medium, and the surfactant flooding

11

Electronic copy available at: https://ssrn.com/abstract=4143309

248 further displaces additional oil. In our tests, the empty clean micromodel was initially saturated

249 with the mineral oil. The flooding fluid, i.e., brine was then injected with a syringe pump (NE-

250 1002X; New Era Systems Inc., NY, USA) at a flow rate of 0.5 L/min (corresponding Reynold’s

251 number Re = 0.004). This brine injection continued for two hours, which was equivalent to 113

252 pore volumes. At this stage, the 2 h-long brine injection rendered the residual oil saturation in the

253 micromodel levelled off. Thereafter, the biosurfactant solution was injected at the same injection

254 rate for additional two hours.

255 The image analysis of the acquired time-lapsed images of the mutli-channel micromodel enabled

256 calculation and monitoring of the pore saturations of the fluids during the biosurfactant injection.

257 The static contact angles were determined by the Tangent method, which measures the contact

258 angle at the crossing point of the three phases oil, brine and PDMS(Chau, 2009; Rotenberg et al.,

259 1983). The oil-brine-PDMS contact angle in each case was measured at least 74 locations and the

260 results were fitted with the normal distribution.

261

12

Electronic copy available at: https://ssrn.com/abstract=4143309

262 3. Results and Discussion

263 3.1. Observations from Single-channel Micromodel

264 In-situ Biosurfactant Production and The Wettability Modification

265 Contact angle measurement is one of the most common and elegant tests to quantitatively assess

266 wettability. In EOR practices, among various parameters affecting the capillary pressure and the

267 fluids’ mobility in porous media, wettability is one of the most crucial factors due to its complex

268 nature. The static contact angle is referred to as the contact angle measured when the contacting

269 interface of two different phases has no motion. Whereas, the dynamic contact angle is quantified

270 when the interface has a motion on the solid surface by either receding or advancing. In this section,

271 we examine the wettability transitions during the growth of biosurfactant-producing bacteria by

272 measuring real-time dynamic and static contact angles in the single-channel micromodel.

273 Figure 2a shows the transitions in mineral oil-brine-PDMS contact angle over the course of B.

274 subtilis cultivation. It is confirmed that biosurfactant generation and accumulation caused gradual

275 reductions in the contact angles. Due to the surfactin attachment at the oil-brine interface, the static

276 contact angle decreased from ~110−115° to ~35−41° over 144 hours, as shown in Figure 2b. The

277 contact angle began to decrease after 10 hours, remained dropping for additional 20 hours, and

278 was eventually levelled off at the lower limit. Similar to the changes in static contact angles, the

279 gradual reductions of dynamic contact angles were also observed during ~10−30 hours. The

280 advancing contact angle decreased from ~138−151° to ~48−60°, whereas the receding contact

281 angle decreased from ~90−102° to ~16−30°. All the static, advancing and receding contact angles

282 reduced by ~70−90° over experimentation periods. The lower limits of the contact angles indicate

283 the in situ concentration of surfactin reached the critical micelle concentration at around 30 hours.

13

Electronic copy available at: https://ssrn.com/abstract=4143309

284 Prcevious studies have reported that 30−40 mg/L of surfactin concentration is required to reduce

285 the contact angle to the lower limit.(Abdel-Mawgoud et al., 2008; Davis et al., 1999; Liu et al.,

286 2015; Park et al., 2019) Accordingly, our test result suggests that B. subtilis produced sufficient

287 amount of surfactin at minimum 30–40 mg/L and caused modification of wettability to the

288 maximum level.

289 The differences in contact angle at the static and dynamic states could directly impact pore-level

290 displacement mechanisms and the macroscopic multiphase flow in an oil reservoir. Therefore, it

291 is important to choose a proper contact angle (i.e., among the static, the receding, and the

292 advancing contact angles) to better predict the wettability and the fluids’ behavior in different types

293 of rock formations. During immiscible fluid flows, the immiscible fluids and their interfaces tend

294 to have a motion, which makes dynamic contact angles more suitable for the wettability evaluation.

295 The advancing contact angle is the upper boundary of the contact angle, which can be potentially

296 used to predict the injectability and the capillary pressure of water using the Young-Laplace

297 equation for water flooding in EOR practices(Jafari and Jung, 2017; Young, 1805). Whereas, the

298 receding contact angle, the lower boundary of contact angle, can be used to determine the capillary

299 pressure and the mobility of oil.

300

301 Effects of Fluid Velocity in Measured Dynamic Contact Angles

302 The dynamic contact angle is widely used to describe wetting, spreading, and adhesion processes

303 between a liquid and a solid.(Keller et al., 2007; Šikalo et al., 2005) Earlier observations have

304 shown that a dynamic contact angle measured is dependent on various measurement parameters,

305 including the drag (viscous and capillary) forces and the surface physical/chemical

14

Electronic copy available at: https://ssrn.com/abstract=4143309

306 heterogeneity(Cassie and Baxter, 1944; Jiang et al., 1979; Johnson Jr et al., 1977; Keller et al.,

307 2007). Expectedly, the fluid velocity determines the drag force, and hence largely affects the

308 contact angle(Keller et al., 2007; Šikalo et al., 2005). This section explores the impact of flow

309 velocity on the dynamic contact angle.

310 In this study, the oil-brine interface moved as a push-pull piston, and therefore, the flow velocity

311 was identical to the moving velocity of the interface (Figure 2a). Thereby, the fluid velocity was

312 estimated using the video captured for the dynamic contact angles. Particularly, we chose two time

313 intervals. The first time frame was the first 5 hours, in which we assumed that the biosurfactant

314 production was yet negligible and had no or minimal impact on the contact angles. The second

315 time frame was the range after 32 hours, where the contact angle converged to the lower limit

316 value. Meanwhile, the interval between 5–32 hours were excluded, where the bacteria grew

317 exponentially and produced the surfactin at a significantly rapid rate, such that this phase allowed

318 insufficient time for the push and pull manipulation.

319 Figure 2c shows the variations in the contact angles with the flow velocity. The velocity of 0 m/s

320 indicates no fluid mobility and thus corresponds to the static contact angle. The positive velocity

321 represents the advancing contact angle during pushing, and the negative velocity for the receding

322 contact angle during retraction, respectively. The result reveals that the greater extent of contact

323 angle hysteresis increases with an increase in the velocity, regardless of the presence of surfactin.

324 Before surfactin production, the advancing contact angle increased from ~110−115° at 0 m/s,

325 ~140° at 10 m/s (by 24%), and to ~151° at 18 m/s (by 34%). After surfactin production, the

326 advancing contact angle increased from ~35−41° at 0 m/s, ~~48° at 10 m/s (by 26%), and to

327 ~60° at 21 m/s (by 58%). The receding angles also showed the consistent tendency.

15

Electronic copy available at: https://ssrn.com/abstract=4143309

328 Our results demonstrate that 10–20 μm/s of velocity changes are sufficient to affect the contact

329 angles in a porous medium. Furthermore, the advancing and receding contact angles were expected

330 to be levelled off at the velocity greater than ~20–30 m/s (Figure 2c). This is fairly consistent

331 with the previous study result which has reported that the contact angle becomes insensitive to

332 advancing and receding velocity above 150–200 μm/s.(Keller et al., 2007) The results highlight

333 the role of advancing/receding velocities in determining dynamic contact angles and wettability.

334 The examined flow velocity regime gains more relevance in a tight formation with low

335 permeability, where an injected fluid slowly invades and displaces pre-occupying fluids.

336

337 Pore-scale Observation of Microbial Cell Growth

338 The single-channel micromodel test allows the monitoring of bacterial cell growth, biofilm

339 production, and cell attachments at a pore scale. Figure 3 presents pore-scale observation on the

340 bacterial growths during the cultivation of biosurfactant-producing bacteria, B. subtilis, in a

341 defined medium inside the single-channel micromodel. When 16 hours elapsed, the bacterial cell

342 growth initiates. During 16 to 30 hours elapsed, bacterial cells show an exponential growth phase,

343 which is consistent with the optical density measured at 600 nm (Figure 3b). After 40 hours,

344 bacterial cells start to form biofilms, the complex structures of microbiome having bacterial cells

345 in a group and are locally agglomerated. As shown in Figure 3a (41 h and 55 h), the density of

346 bacterial cell agglomeration near biofilm structure increases as time passes. It is common in nature

347 that the bacterial cells are locally accumulated near biofilms.

348 Figure 4 depicts that the presence of bacterial cells affects the static contact angle. This image was

349 acquired at 80 h elapsed when the in situ concentration of surfactin is over critical micelle

16

Electronic copy available at: https://ssrn.com/abstract=4143309

350 concentration. Note that at over critical micelle concentration, the static contact angle is 35−41°

351 (Figure 2b). By contrast, the contact angle of oil-brine-PDMS near the accumulated bacterial cells

352 and biofilms is evidently higher (~60°) than that of the cell-free region (~35−36°). This indicates

353 that local accumulation of bacterial masses degrades the wettability of water in a confined space,

354 and in turn affect the fluid mobility in a porous medium. meanwhile, the use of biofilm or

355 biopolymers has widely been introduced to selectively plug the pore space and reduce the

356 permeability of the high permeable zone to induce a high oil recovery rate. Even in underground

357 such as an oil reservoir, it should be noted that the fluid mobility in porous media can differ when

358 biomass accumulates over time.

359 Our micromodel suggests that the monitoring of bacterial cell growth, biofilm production, and cell

360 attachments at a pore scale in real-time can be a significant asset in microbiology studies. Since

361 the cultivation of bacteria in a micromodel is performed at a microscale, no additional steps such

362 as fluorescence staining or attaching a bead for density measurements(Kim et al., 2019) are needed.

363 Micromodels also allow efficient, convenient, and continuous quantification of bacterial growth

364 in real-time without discarding samples.

365

366 Demulsification of Oil-brine During in situ Biosurfactant Accumulation

367 An emulsion is defined as the dispersion of a liquid in another immiscible liquid. In the petroleum

368 industry, the formation of a crude oil emulsion is ineluctable for several sources of mixing

369 including flow through the reservoir rock, bottomhole perforations, and the presence of emulsifiers

370 such as asphaltenes or chemical additives.(Abdulredha et al., 2020; Liu et al., 2020) Because

371 separating the water in oil emulsion needs an additional refining process, the emulsion in the

372 petroleum industry is undesirable, but at the same time, unavoidable. Therefore, it is critical to

17

Electronic copy available at: https://ssrn.com/abstract=4143309

373 understand the formation mechanism of oil emulsions as well as the demulsification process of oil

374 emulsions in porous media.

375 The surfactant production with the aid of mechanical agitation during the cyclic flow causes

376 emulsification of oil and water. Figure 5 captures the process of nano-scale oil-in-water emulsion.

377 In the first stage (Figure 5a), water and oil are immiscible with a clear interface. However, when

378 bacterial grows and surfactant is produced with time, the system changes to the bicontinuous

379 microemulsion state, where the micelles become swollen and are no longer spherical (Figure

380 5b).(Acosta et al., 2003) In this bicontinuous microemulsion stage, the volume fractions of oil and

381 water phase are approximately equal and two phases are interdispersed within the system.

382 Thereafter, the oil begins to break down into small microemulsions (Figure 5c). Finally, the droplet

383 size of the emulsion decreases to the nano-scales, as shown in Figure 5d. Surfactant is not only a

384 well-known surface-active agent, but also an emulsifier that lowers the surface tension between

385 two immiscible liquids and stabilizes emulsions. Hydrophilic and hydrophobic parts of surfactant

386 molecules facilitate absorption of surfactants at the interface between oil and water; thereby the

387 surfactant molecules coat immscible fluid droplets as an emulsion and prevent them from

388 coalescing.

389 Contrary to that, we observed that the produced surfactin functioned as a demulsifier to break

390 emulsions. To illustrate this, Figure 6 displays time-lapsed snapshots of one location within the

391 single-channel micromodel during an experiment. There are monodisperse oil droplets with an

392 initial diameter of ~40 m in the channel. These emulsion droplets undergo collisions, as they

393 move. This collision is attributable to the differences in droplet velocities originating from local

394 differences in hydraulic resistance generated by the presence of neighboring droplets.(Krebs et al.,

18

Electronic copy available at: https://ssrn.com/abstract=4143309

395 2012) Over time, several small droplets collide and coalesce to a larger single droplet having ~50

396 m.

397 This demulsification is possibly attributable to the change in fluid pH as the pH regulates the

398 activity of surfactin. It is known that surfactin looses its ability as a surface-active agent in low

399 pH. Surfactin readily precipitates under an acidic condition due to the protonation of a carboxyl

400 group.(Abdel-Mawgoud et al., 2008; Park et al., 2020) It is presumed that the pH in our system

401 may have been lowered due to the consumption of phosphate buffers after several days of

402 experimental periods, and thereby emulsified emulsion droplets were demulsified owing to the

403 surfactant precipitation. Surfactants can be an emulsifier that stabilizes emulsions but at the same

404 time, can also function as a demulsifier that migrates at the oil-water interface to neutralize the

405 effect of emulsifying agents. There have been several reports to emulsify or demulsify

406 oil/water/surfactin systems through regulating pH.(Long et al., 2017; Sana et al., 2021; Yang et

407 al., 2020)

408 It should be noted that surfactin produced by B. subtilis has a profound effect not only on

409 wettability modification but also as a promising demulsifier. When water is injected, the addition

410 of surfactin could help modify the wettability and increase the oil recovery rate. Afterwards, when

411 the injected pH buffers are consumed, precipitated surfactin could possibly trigger the

412 demulsification process. Our results suggest that simple adjustment of pH can selectively emulsify

413 or separate oil/water during surfactin-aided MEOR process.

414

415 3.2. Observations from the Multi-channel Microdel

416 Effects of Biosurfactant on Water Flooding: Reducing the Residual Oil Saturation

19

Electronic copy available at: https://ssrn.com/abstract=4143309

417 This section explores residual oil saturation after either water flooding or surfactant flooding in

418 the multi-channel micromodel. Our test condition for water injection with no surfactin represents

419 the mobility ratio of -1.5 and the capillary number of -5.1. The surfactant flooding condition has

420 the mobility ratio of -1.9 and the capillary number of -5.0, respectively, for injection of

421 biosurfactant-containing water into mineral oil-filled micromodel. The calculated mobility ratio

422 and capillary number indicate that the invasion patterns of immiscible fluids are dominantly

423 governed by capillary fingering in our porous micromodel.(Cao et al., 2016; Lenormand et al.,

424 1988)

425 Figure 7 shows the oil displacement patterns of water flooding. For the first two hours, water

426 flooding with no additives proceeded. After two hours of water flooding, biosurfactant-containing

427 water was injected to simulate surfactant flooding and to quantify the enhancement of the

428 displacement efficiency of oil in a porous medium. Without biosurfactant, the water flooding

429 yielded 82.9% of oil recovery while leaving the residual oil saturation of 17.1% within micromodel.

430 Compartively, the residual oil saturation when flooding the surfactin-containing fluid with 0.3,

431 0.5, 1 and 2 critical micelle concentration were 11.7%, 9.2%, 1.9% and 0.3%, respectively. The

432 result shows that the residual oil saturation decreases as the concentration of biosurfactant in the

433 flooding fluid increases. The lowered residual oil saturation and the greater oil displacement

434 efficiency is attributable to the reduced interfacial tension and the PDMS surface that is modified

435 to be more hydrophilic by surfactin. Interestingly, it is worth pointing out that this positive effect

436 by surfactin appears still effective even when the surfactin concentration is two times greater than

437 the critical micelle concentration. In fact, the oil displacement is considerably improved when the

438 surfactant concentration increases from 1 to 2 critical micelle concentration. This is possibly

439 because the surface-active effect of surfactin is not uniform, but show a wide distribution over the

20

Electronic copy available at: https://ssrn.com/abstract=4143309

440 micromodel.

441

442 Biosurfactant Induced Wettability Modification

443 Figure 8a shows the sub-sections within the multi-channel micromodels after either water flooding

444 or surfactant flooding. These enlarged sections depict clear contrasts in the shapes of oil-water

445 meniscii between the cases with and without biosurfactant. We analyzed the contact angle at all

446 the contacts and determined their distributions, as shown in Figure 8b. The average contact angles

447 decreases as the surfactin concentration in a flooding fluid increases. The average contact angle of

448 the water-oil-PDMS interface before surfactin treatment was 154.7°. The averag contact angle

449 decreases to 129.8° when 0.3 critical micelle concentration added, to 123.7° when 0.5 critical

450 micelle concentration added, and to 113.1° by the addition of 1 critical micelle concentration (35

451 mg/L) of biosurfactant to the flooding fluid (Figure 8b). When the concentration of surfactin is

452 doubled to 2 critical micelle concentration, the average contact angle decreases to the lowest value

453 of 105.2°. It is assumed that the reduction in the water-oil interfacial tension and the modification

454 of PDMS surface to be more water-wet, led to the increased oil sweeping efficiencies, as depicted

455 in Figure 8.

456 The standard deviation indicates the variation of the contact angle in the micromodel, and implies

457 the heterogeneity in wetting behavior, such as surfactin adsorption or surfactin effectiveness. The

458 contact angle distribution without biosurfactant shows the least variation with the lowest standard

459 deviation of 4.8° among the tested cases. The cases with surfactin (0.3, 0.5, 1, and 2 critical micelle

460 concentration) show the greater standard deviation and the wider variation in contact angle than

461 the case without surfactin. This can be explained by unevenly distributed surfactin. First, before

462 the system reaches to the critical micelle concentration, surfactin monomers can be heavily locally

21

Electronic copy available at: https://ssrn.com/abstract=4143309

463 adsorbed in some regions, such that there are other locations with insufficient amount of sufactin.

464 Particularly with 0.3 critical micelle concentration and 0.5 critical micelle concentration, before

465 micells are formed, there are some ranges on the right tails overlapping with the distribution of 0

466 critical micelle concentration. Furthermore, the surfatin presence significantly reduces the oil

467 droplet size, and thus the droplet’s relative position to the PDMS pillars heavily affects the contact

468 angle, which can cause wide variations.

469 The contact angle hysteresis is generally suppressed with the presence of surfactant that standard

470 deviation decreases with an increase in surfactin concentration.(Mirchi et al., 2014) However, the

471 results here show that the contact angle hysteresis can increase when surfactant concentration is

472 less than the critical micelle concentration. The geometries of a porous medium in an oil reservoir

473 are heterogeneous. And the heterogeneuities of a porous media leads to the higher contact angle

474 hysteresis which make fluids flow behaviors highly unpredictable. In this manner, to achieve the

475 overall uniform transitions of wettability and oil recovery enhancement, it is critical to inject

476 enough amount of surfactant that is higher than critical micelle concentration.

477

22

Electronic copy available at: https://ssrn.com/abstract=4143309

478 4. Conclusion

479 Wettability plays a vital role in the displacement of fluids in a porous medium. This study

480 investigated the potential of using micromodels to understand wettability transitions during

481 bacterial surfactant accumulation at the pore scale. With different types of micromodels,

482 experiments were conducted to evaluate the biosurfactant-induced MEOR process by altering the

483 wettability. A single-channel micromodel was fabricated to observe the biosurfactant-induced

484 wettability modification in real-time. It enabled direct observations of bacterial growth and biofilm

485 accumulations in a porous medium. Also, the time-lapse stages of dynamic-static contact angle

486 changes and the emulsification/demulsification processes occurring at the water-oil interface were

487 directly observed over the course of bacterial cultivations. With the single-channel micromodel, it

488 was observed that both dynamic and static contact angles decreases as biosurfactant concentration

489 increases over time. Also, dynamic contact angles varied owing to the internal velocity of fluids.

490 Specifically, the receding contact angle decreased and the advancing contact angle increased as

491 the velocity increased.

492 The multi-channel micromodel fabricated for the surfactant-flooding experiment was used to

493 investigate wettability transitions with the addition of different concentrations of biosurfactants.

494 The results showed that the imbibition rates and the residual oil saturation are directly related to

495 the surfactant concentration, and the concentration of surfactant should be over critical micelle

496 concentration to achieve homogeneous wettability transition and high oil-sweeping performance.

497

498 Acknowledgement

499 This work is supported by the Korea Agency for Infrastructure Technology Advancement (KAIA)

500 grant funded by the Ministry of Land, Infrastructure and Transport (Grant No.: 21CTAP-C163693-

23

Electronic copy available at: https://ssrn.com/abstract=4143309

501 01).

502

503 References

504 Abdel-Mawgoud, A.M., Aboulwafa, M.M. and Hassouna, N.A.-H., 2008. Characterization of
505 surfactin produced by Bacillus subtilis isolate BS5. Appl. Biochem. Biotechnol. 150(3):
506 289-303.
507 Abdulredha, M.M., Aslina, H.S. and Luqman, C.A., 2020. Overview on petroleum emulsions,
508 formation, influence and demulsification treatment techniques. Arab. J. Chem. 13(1):
509 3403-3428.
510 Acosta, E., Szekeres, E., Sabatini, D.A. and Harwell, J.H., 2003. Net-average curvature model for
511 solubilization and supersolubilization in surfactant microemulsions. Langmuir, 19(1): 186-
512 195.
513 Al-Bahry, S. et al., 2013. Biosurfactant production by Bacillus subtilis B20 using date molasses
514 and its possible application in enhanced oil recovery. Int. Biodeterior. Biodegradation 81:
515 141-146.
516 Al-Wahaibi, Y. et al., 2014. Biosurfactant production by Bacillus subtilis B30 and its application
517 in enhancing oil recovery. Colloids Surf. B 114: 324-333.
518 Ali, H. et al., 2021. Experimental investigation and two-phase flow simulation of oil and
519 nanofluids on micro CT images of sandstone for wettability alteration of the system. J. Pet.
520 Sci. Eng. 204: 108665.
521 Beebe, D.J., Mensing, G.A. and Walker, G.M., 2002. Physics and applications of microfluidics in
522 biology. Annu. Rev. Biomed. Eng. 4(1): 261-286.
523 Benincasa, M., Abalos, A., Oliveira, I. and Manresa, A., 2004. Chemical structure, surface
524 properties and biological activities of the biosurfactant produced by Pseudomonas
525 aeruginosa LBI from soapstock. Antonie Van Leeuwenhoek 85(1): 1-8.
526 Cao, S.C., Bate, B., Hu, J.W. and Jung, J., 2016. Engineering behavior and characteristics of water-
527 soluble polymers: Implication on soil remediation and enhanced oil recovery.
528 Sustainability 8(3): 205.
529 Cassie, A. and Baxter, S., 1944. Wettability of porous surfaces. Trans. Faraday Soc. 40: 546-551.
530 Chau, T., 2009. A review of techniques for measurement of contact angles and their applicability
531 on mineral surfaces. Miner. Eng. 22(3): 213-219.
532 Chen, Y. et al., 2020. Geochemical controls on wettability alteration at pore-scale during low
533 salinity water flooding in sandstone using X-ray micro computed tomography. Fuel 271:
534 117675.
535 Cooper, D.G. and Goldenberg, B.G., 1987. Surface-active agents from two Bacillus species. Appl.
536 Environ. Microbiol. 53(2): 224-229.
537 Dangla, R., Gallaire, F. and Baroud, C.N., 2010. Microchannel deformations due to solvent-
538 induced PDMS swelling. Lab Chip 10(21): 2972-2978.
539 Datta, P., Tiwari, P. and Pandey, L.M., 2018. Isolation and characterization of biosurfactant
540 producing and oil degrading Bacillus subtilis MG495086 from formation water of Assam
541 oil reservoir and its suitability for enhanced oil recovery. Bioresour. Technol. 270: 439-
542 448.
543 Daverey, A. and Pakshirajan, K., 2010. Sophorolipids from Candida bombicola using mixed
544 hydrophilic substrates: production, purification and characterization. Colloids Surf. B

24

Electronic copy available at: https://ssrn.com/abstract=4143309

545 79(1): 246-253.
546 Davis, D., Lynch, H., Varley, J.J.E. and Technology, M., 1999. The production of surfactin in
547 batch culture by Bacillus subtilis ATCC 21332 is strongly influenced by the conditions of
548 nitrogen metabolism. Enzyme Microb. Technol. 25(3-5): 322-329.
549 Desai, J.D. and Banat, I.M., 1997. Microbial production of surfactants and their commercial
550 potential. Microbiol. Mol. Biol. Rev. 61(1): 47-64.
551 Fan, Y., Gao, K., Chen, J., Li, W. and Zhang, Y., 2018. Low-cost PMMA-based microfluidics for
552 the visualization of enhanced oil recovery. Oil. Gas. Sci. Technol. 73: 26.
553 Gogoi, S. and Gogoi, S.B., 2019. Review on microfluidic studies for EOR application. J. Pet.
554 Explor. Prod. Technol. 9(3): 2263-2277.
555 Gravesen, P., Branebjerg, J. and Jensen, O.S., 1993. Microfluidics-a review. J. Micromech.
556 Microeng. 3(4): 168.
557 Hassanpourfard, M. et al., 2014. Protocol for biofilm streamer formation in a microfluidic device
558 with micro-pillars. J. Vis. Exp. JOVE (90).
559 Jacques, P., 2011. Biosurfactants. Springer.
560 Jafari, M. and Jung, J., 2017. Direct measurement of static and dynamic contact angles using a
561 random micromodel considering geological CO2 sequestration. Sustainability 9(12): 2352.
562 Jang, H. et al., 2019. Traction microscopy with integrated microfluidics: responses of the multi-
563 cellular island to gradients of HGF. Lab Chip 19(9): 1579-1588.
564 Jang, H. et al., 2017. Homogenizing cellular tension by hepatocyte growth factor in expanding
565 epithelial monolayer. Sci. Rep. 7(1): 1-10.
566 Jiang, T.-S., Soo-Gun, O. and Slattery, J.C., 1979. Correlation for dynamic contact angle. J.
567 Colloid Interface Sci. 69(1): 74-77.
568 Johnson Jr, R.E., Dettre, R.H. and Brandreth, D.A., 1977. Dynamic contact angles and contact
569 angle hysteresis. J. Colloid Interface Sci. 62(2): 205-212.
570 Kathel, P. and Mohanty, K.K., 2013. EOR in tight oil reservoirs through wettability alteration,
571 SPE Annual Technical Conference and Exhibition, Society of Petroleum Engineers.
572 Keller, A., Broje, V. and Setty, K., 2007. Effect of advancing velocity and fluid viscosity on the
573 dynamic contact angle of petroleum hydrocarbons. J. Pet. Sci. Eng. 58(1-2): 201-206.
574 Kim, S., Lee, S., Kim, J.-K., Chung, H.J. and Jeon, J.S., 2019. Microfluidic-based observation of
575 local bacterial density under antimicrobial concentration gradient for rapid antibiotic
576 susceptibility testing. Biomicrofluidics 13(1): 014108.
577 Kim, S. and Santamarina, J.C., 2014. Engineered CO2 injection: The use of surfactants for
578 enhanced sweep efficiency. Int. J. Greenh. Gas Control 20: 324-332.
579 Korenblum, E. et al., 2005. Production of antimicrobial substances by Bacillus subtilis LFE‐1, B.
580 firmus H2O‐1 and B. licheniformis T6‐5 isolated from an oil reservoir in Brazil. J. Appl.
581 Microbiol. 98(3): 667-675.
582 Krebs, T., Schroen, K. and Boom, R., 2012. A microfluidic method to study demulsification
583 kinetics. Lab Chip 12(6): 1060-1070.
584 Lake, L.W., 1989. Enhanced oil recovery. Prentice Hall Inc.
585 Lake, L.W., Johns, R.T., Rossen, W.R. and Pope, G.A., 2014. Fundamentals of enhanced oil
586 recovery. Society of Petroleum Engineers.
587 Lebedeva, E.V. and Fogden, A., 2011. Micro-CT and wettability analysis of oil recovery from
588 sand packs and the effect of waterflood salinity and kaolinite. Energy Fuels 25(12): 5683-
589 5694.
590 Lenormand, R., Touboul, E. and Zarcone, C., 1988. Numerical models and experiments on

25

Electronic copy available at: https://ssrn.com/abstract=4143309

591 immiscible displacements in porous media. J. Fluid Mech. 189: 165-187.
592 Liang, T. et al., 2021. Surfactant-EOR in tight oil reservoirs: Current status and a systematic
593 surfactant screening method with field experiments. J. Pet. Sci. Eng. 196: 108097.
594 Lifton, V.A., 2016. Microfluidics: an enabling screening technology for enhanced oil recovery
595 (EOR). Lab Chip 16(10): 1777-1796.
596 Liu, Q., Lin, J., Wang, W., Huang, H. and Li, S., 2015. Production of surfactin isoforms by Bacillus
597 subtilis BS-37 and its applicability to enhanced oil recovery under laboratory conditions.
598 Biochem. Eng. J. 93: 31-37.
599 Liu, Y. et al., 2020. A review on emulsification via microfluidic processes. Front. Chem. Sci. Eng.
600 14(3): 350-364.
601 Long, X., He, N., He, Y., Jiang, J. and Wu, T., 2017. Biosurfactant surfactin with pH-regulated
602 emulsification activity for efficient oil separation when used as emulsifier. Bioresour.
603 Technol. 241: 200-206.
604 Måløy, K.J., Furuberg, L., Feder, J. and Jøssang, T., 1992. Dynamics of slow drainage in porous
605 media. Phys. Rev. Lett. 68(14): 2161.
606 Mirchi, V., Saraji, S., Goual, L. and Piri, M., 2014. Dynamic interfacial tensions and contact angles
607 of surfactant-in-brine/oil/shale systems: implications to enhanced oil recovery in shale oil
608 reservoirs, SPE Improved Oil Recovery Symposium.
609 Mohammadi, M. and Mahani, H., 2020. Direct insights into the pore-scale mechanism of low-
610 salinity waterflooding in carbonates using a novel calcite microfluidic chip. Fuel 260:
611 116374.
612 Morrow, N.R., 1990. Wettability and its effect on oil recovery. J. Pet. Technol. 42(12): 1,476-
613 1,484.
614 Nicholson, W.L., Munakata, N., Horneck, G., Melosh, H.J. and Setlow, P., 2000. Resistance of
615 Bacillus endospores to extreme terrestrial and extraterrestrial environments. Microbiol.
616 Mol. Biol. Rev. 64(3): 548-572.
617 Park, T., Jeon, M.-K., Yoon, S., Lee, K.S. and Kwon, T.-H., 2019. Modification of interfacial
618 tension and wettability in oil-brine-quartz system by in situ bacterial biosurfactant
619 production at reservoir conditions: Implications to microbial enhanced oil recovery.
620 Energy Fuels 33(6): 4909-4920.
621 Park, T. et al., 2017. Biosurfactant as an Enhancer of Geologic Carbon Storage: Microbial
622 Modification of Interfacial Tension and Contact Angle in Carbon dioxide/Water/Quartz
623 Systems. Front. microbiol. 8: 1285.
624 Park, T., Yoon, S., Jung, J. and Kwon, T.-H., 2020. Effect of Fluid–Rock Interactions on In Situ
625 Bacterial Alteration of Interfacial Properties and Wettability of CO2–Brine–Mineral
626 Systems for Geologic Carbon Storage. Environ. Sci. Technol. 54(23): 15355-15365.
627 Peet, K.C. et al., 2015. Microbial growth under supercritical CO2. Appl. Environ. Microbiol. 81(8):
628 2881-2892.
629 Rangarajan, V. and Clarke, K.G., 2016. Towards bacterial lipopeptide products for specific
630 applications—a review of appropriate downstream processing schemes. Process Biochem.
631 51(12): 2176-2185.
632 Ron, E.Z. and Rosenberg, E., 2001. Natural roles of biosurfactants. Environ. Microbiol. 3(4): 229-
633 236.
634 Rotenberg, Y., Boruvka, L. and Neumann, A., 1983. Determination of surface tension and contact
635 angle from the shapes of axisymmetric fluid interfaces. J. Colloid Interface Sci. 93(1): 169-
636 183.

26

Electronic copy available at: https://ssrn.com/abstract=4143309

637 Saadat, M., Tsai, P.A., Ho, T.-H., Øye, G. and Dudek, M., 2020. Development of a microfluidic
638 method to study enhanced oil recovery by low salinity water flooding. ACS Omega 5(28):
639 17521-17530.
640 Sana, S., Sengupta, D. and Datta, S., 2021. Application of biosurfactant in the formulation of
641 petrochemical products such as demulsifying/emulsifying agents, Green Sustainable
642 Process for Chemical and Environmental Engineering and Science. Elsevier.
643 Seethepalli, A., Adibhatla, B. and Mohanty, K., 2004. Wettability alteration during surfactant
644 flooding of carbonate reservoirs, SPE/DOE Symposium on Improved Oil Recovery.
645 Society of Petroleum Engineers.
646 Shaligram, N.S. and Singhal, R.S., 2010. Surfactin–a review on biosynthesis, fermentation,
647 purification and applications. Food Technol. Biotechnol. 48(2): 119-134.
648 Šikalo, Š., Wilhelm, H.-D., Roisman, I., Jakirlić, S. and Tropea, C., 2005. Dynamic contact angle
649 of spreading droplets: Experiments and simulations. Phys. Fluids 17(6): 062103.
650 Simpson, D.R., Natraj, N.R., McInerney, M.J. and Duncan, K.E., 2011. Biosurfactant-producing
651 Bacillus are present in produced brines from Oklahoma oil reservoirs with a wide range of
652 salinities. Appl. Microbiol. Biotechnol. 91(4): 1083.
653 Squires, T.M. and Quake, S.R., 2005. Microfluidics: Fluid physics at the nanoliter scale. Rev.
654 Mod. Phys. 77(3): 977.
655 Stalder, A.F., Kulik, G., Sage, D., Barbieri, L. and Hoffmann, P., 2006. A snake-based approach
656 to accurate determination of both contact points and contact angles. Colloids Surf. A
657 Physicochem. Eng. Asp. 286(1-3): 92-103.
658 Wang, Y. et al., 2011. Surfactant induced reservoir wettability alteration: Recent theoretical and
659 experimental advances in enhanced oil recovery. Pet. Sci. 8(4): 463-476.
660 Wei, Y.-H. and Chu, I.-M., 2002. Mn2+ improves surfactin production by Bacillus subtilis.
661 Biotechnol. Lett. 24(6): 479-482.
662 Wei, Y.-H., Lai, C.-C. and Chang, J.-S., 2007. Using Taguchi experimental design methods to
663 optimize trace element composition for enhanced surfactin production by Bacillus subtilis
664 ATCC 21332. Process Biochem. 42(1): 40-45.
665 White, D., Hird, L. and Ali, S., 2013. Production and characterization of a trehalolipid
666 biosurfactant produced by the novel marine bacterium Rhodococcus sp., strain PML026. J.
667 Appl. Microbiol. 115(3): 744-755.
668 Wu, Y., Shuler, P.J., Blanco, M., Tang, Y. and Goddard, W.A., 2008. An experimental study of
669 wetting behavior and surfactant EOR in carbonates with model compounds. SPE J. 13(01):
670 26-34.
671 Yakimov, M.M., Timmis, K.N., Wray, V. and Fredrickson, H.L., 1995. Characterization of a new
672 lipopeptide surfactant produced by thermotolerant and halotolerant subsurface Bacillus
673 licheniformis BAS50. Appl. Environ. Microbiol. 61(5): 1706-1713.
674 Yang, Z. et al., 2020. Application of biosurfactant surfactin as a pH-switchable biodemulsifier for
675 efficient oil recovery from waste crude oil. Chemosphere 240: 124946.
676 Young, T., 1805. An essay on the cohesion of fluids. Philos. Trans. R. Soc. (95): 65-87.
677 Zhang, D.L., Liu, S., Puerto, M., Miller, C.A. and Hirasaki, G.J., 2006a. Wettability alteration and
678 spontaneous imbibition in oil-wet carbonate formations. J. Pet. Sci. Eng. 52(1-4): 213-226.
679 Zhang, P., Tweheyo, M.T. and Austad, T., 2006b. Wettability alteration and improved oil recovery
680 in chalk: The effect of calcium in the presence of sulfate. Energy Fuel 20(5): 2056-2062.
681

27

Electronic copy available at: https://ssrn.com/abstract=4143309

682
683

28

Electronic copy available at: https://ssrn.com/abstract=4143309

684 List of Table
685
686 Table 1. Medium composition for Bacillus subtilis growth and surfactin production
687

29

Electronic copy available at: https://ssrn.com/abstract=4143309

688 Table 1. Medium composition for Bacillus subtilis growth and surfactin production

Compound Concentration
Carbon source
Glucose 40 g/L
Mineral salt medium
MgSO4 8.0 × 10-4 M
CaCl2 8.0 × 10-4 M
FeSO4 8.0 × 10-4 M
Na2EDTA 8.0 × 10-4 M
MnSO4 8.0 × 10-4 M
Nitrogen source
NH4Cl 0.1 M
NaNO3 0.118 M
Phosphate buffer
KH2PO4 0.03 M
Na2HPO4 0.04 M
689

30

Electronic copy available at: https://ssrn.com/abstract=4143309

690 List of Figures
691
692 Figure 1. Schematic descriptions of the geometries of (a) the single-channel micromodel, (b) the
693 multi-channel micromodel, and their corresponding experimental setups for (c) the single-channel
694 and (d) the multi-channel micromodels.
695
696 Figure 2. (a) Snapshots of examples of receding (A, D), static (B, E), and advancing (C, F) contact
697 angles before and after the bacterial cultivation. (b) The dynamic and static contact angle
698 evolutions with time. And (c) the contact angle variations by velocity. Note: Negative velocity on
699 the x-axis indicates the pump retracted the water phase for receding contact angle measurement,
700 and where the positive velocity indicates when the pump pushed the fluid for advancing contact
701 angle measurement. Zero velocity is when the fluid was standstill after the pump is paused.
702
703 Figure 3. Observations of (a) bacterial growth, biofilm accumulations and (b) optical density
704 measured from a separate batch experiment
705
706 Figure 4. Uneven contact angles at the immiscible oil-brine interfaces depend on bacterial biofilm
707 accumulations. Note: The snapshot was taken at the static condition.
708
709 Figure 5. Typical emulsification process from water/oil describing (a) before the emulsion is
710 initiated, (b) bicontinuous microemulsion phase, (c) breakdown of oil into the microemulsions,
711 and (d) the formation of nanoemulsions.
712
713 Figure 6. Demulsification process as time passes. Collision and coalesce of oil droplets in water
714 (oil-in-water emulsions) as time passes. Note that the width of the channel is 50m.
715 Figure 7. Residual oil saturation changes in multi-channel micromodel (MCM) after 2h
716 biosurfactant flooding. The figures represent the final projections of (a) 0, (b) 0.3, (c) 0.5, (d) 1
717 and (e) 2 critical micelle concentration (where 1 critical micelle concentration is equivalent to 35
718 mg/L).
719 Figure 8. (a) Contact angle comparison of non-treated and surfactin-treated microfluidics chip
720 and (b) Normal distribution of contact angles of mineral oil-water-PDMS systems. Figure A’ and
721 B’ below shows the irreducible water representing connate water in the presence of surfactants.

31

Electronic copy available at: https://ssrn.com/abstract=4143309

722

723 Figure 1. Schematic descriptions of the geometries of (a) the single-channel micromodel, (b) the multi-channel micromodel, and their
724 corresponding experimental setups for (c) the single-channel and (d) the multi-channel micromodels.
725

32

Electronic copy available at: https://ssrn.com/abstract=4143309

726

727 Figure 2. (a) Snapshots of examples of receding (A, D), static (B, E), and advancing (C, F) contact angles before and after the bacterial
728 cultivation. (b) The dynamic and static contact angle evolutions with time. And (c) the contact angle variations by velocity. Note:
729 Negative velocity on the x-axis indicates the pump retracted the water phase for receding contact angle measurement, and where the
730 positive velocity indicates when the pump pushed the fluid for advancing contact angle measurement. Zero velocity is when the fluid
731 was standstill after the pump is paused.

33

Electronic copy available at: https://ssrn.com/abstract=4143309

732

733 Figure 3. Observations of (a) bacterial growth, biofilm accumulations and (b) optical density measured from a separate batch
734 experiment
735
736
737

34

Electronic copy available at: https://ssrn.com/abstract=4143309

738
739 Figure 4. Uneven contact angles at the immiscible oil-brine interfaces depend on bacterial biofilm accumulations. Note: The snapshot
740 was taken at the static condition.
741

35

Electronic copy available at: https://ssrn.com/abstract=4143309

742
743 Figure 5. Typical emulsification process from water/oil describing (a) before the emulsion is initiated, (b) bicontinuous microemulsion
744 phase, (c) breakdown of oil into the microemulsions, and (d) the formation of nanoemulsions.
745

36

Electronic copy available at: https://ssrn.com/abstract=4143309

746
747 Figure 6. Demulsification process as time passes. Collision and coalesce of oil droplets in water (oil-in-water emulsions) as time
748 passes. Note that the width of the channel is 50m.
749

37

Electronic copy available at: https://ssrn.com/abstract=4143309

750

751 Figure 7. Residual oil saturation changes in multi-channel micromodel (MCM) after 2h biosurfactant flooding. The figures represent
752 the final projections of (a) 0, (b) 0.3, (c) 0.5, (d) 1 and (e) 2 critical micelle concentration (where 1 critical micelle concentration is
753 equivalent to 35 mg/L).
754

38

Electronic copy available at: https://ssrn.com/abstract=4143309

755

756 Figure 8. (a) Contact angle comparison of non-treated and surfactin-treated microfluidics chip and (b) Normal distribution of contact
757 angles of mineral oil-water-PDMS systems. Figure A’ and B’ below shows the irreducible water representing connate water in the
758 presence of surfactants.
759

39

Electronic copy available at: https://ssrn.com/abstract=4143309