Studying the generation of foam in the presence of nanoparticles using a microfluidic system

Journal Pre-proofs Studying the generation of foam in the presence of nanoparticles using a microfluidic system Qingjian Li, Valentina Prigiobbe PII: DOI: Reference:

S0009-2509(19)30917-0 https://doi.org/10.1016/j.ces.2019.115427 CES 115427

To appear in:

Chemical Engineering Science

Received Date: Revised Date: Accepted Date:

12 September 2019 4 December 2019 10 December 2019

Please cite this article as: Q. Li, V. Prigiobbe, Studying the generation of foam in the presence of nanoparticles using a microfluidic system, Chemical Engineering Science (2019), doi: https://doi.org/10.1016/j.ces. 2019.115427

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

© 2019 Published by Elsevier Ltd.

Studying the generation of foam in the presence of nanoparticles using a microfluidic system Qingjian Lia , Valentina Prigiobbea,b,∗ a

b

Department of Civil, Environmental, and Ocean Engineering, Department of Chemical Engineering and Material Science, Stevens Institute of Technology, Castle Point on Hudson, Hoboken 07030 New Jersey, U.S.A.

Abstract In this paper, a study on the generation of foam in the presence of nanoparticles in porous media using microfluidics is reported. Drainage and co-injection tests were carried out and monitored with a high-speed camera. Convolutional neural network model was applied to quantify foam texture over time, efficiently. Microscopy images show a generation process characterized by early snap-off, followed by lamella-division, and finally leave-behind. In the presence of nanoparticles, the latter stage is delayed due to resistance to drainage imparted by capillary forces within the liquid films. Moreover, a general relationship between the generation rate and the pressure gradient, which resembles the classical constitutive equation for foam generation, i.e., rg ∝ ∇P α , could be formulated. This indicates that the generation of a foam in the presence of partially hydrophobic nanoparticles and anionic surfactant, used for strong foam formation, follows the same mechanism of a foam stabilized only with surfactant. Keywords: Artificial Intelligence, Bubbles, Foam, Foam generation, Microfluidics, Porous media, Nanoparticles.

1

2 3 4 5 6 7 8 9 10

1. Introduction Foam is injected into the subsurface to reduce gas mobility by increasing its effective viscosity and to divert the gas to low permeability zones [1, 2, 3, 4]. It has been employed in enhanced oil recovery (EOR) and in the remediation of contaminated sites [5, 6, 7, 8]. Foam is a complex fluid where the gas phase is segregated into bubbles separated by thin liquid films (called lamellae). The density of the lamellae gives to the foam its texture and its rheological properties. The low mobility of a foam in comparison to the gas and the liquid phase alone, from which it is formulated, is due to the trapped gas saturation and the increased resistance to flow of the gas bubbles [9, 10]. The drag and the resistance to flow of the lamellae through the pores and throats impart the mobility reduction to foam flow. The ∗

Corresponding author: Email address: [email protected] (Valentina Prigiobbe)

Preprint submitted to Chemical Engineering Science

December 4, 2019

11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48

larger the value of bubbles per unit volume (nf , #/m3 ) the finer the foam texture, therefore the lower the gas mobility [11, 12]. During the transport of foam in a porous medium, the gas bubbles undergo a continuous formation and destruction, which are regulated by mechanisms of generation and coalescence at the pore-scale [2, 13, 14]. Foam is thermodynamically unstable, but its spontaneous decay can be hindered using surface-active materials such as chemical surfactants and/or nanoparticles, which make the foam metastable within the time of an operation [15, 16]. Usually, chemical surfactants are employed to stabilize bubbles within a foam. They adsorb at the gas-liquid interfaces of the lamella, repelling the two interfaces by either electrical double-layer forces, steric effects, or both [9, 17]. This repulsion is called the ”disjoining pressure”. Several experimental studies have been carried out to investigate the effect of chemical surfactants on the mobility of a foam in permeable media, and empirical and mechanistic models have been formulated to describe its behavior during transport [18, 19, 20, 21, 22]. Solely nanoparticles, or in conjunction with a surfactant, have been observed to notably improve the stability of a foam [23, 24, 25, 26, 27, 28, 29]. Nanoparticles can adsorb, irreversibly, at the gas-bubble interface reducing drainage through the lamella and providing a barrier to gas mass-transfer across the bubble wall making the films very stable against rupture and inhibiting the coarsening [30, 31, 32, 33, 34, 35]. Critical characteristics of the nanoparticles that affect foam stability are: hydrophobicity, electrical properties, and particle size [36, 37]. Therefore, surface functionalization in conjunction with solution composition can remarkably vary how the nanoparticles adsorb at the gas-liquid interface and stabilize the foam [38]. The study of the effect of nanoparticles with/without surfactant on foam flow has been a focus of research in the recent years [39, 40, 41, 42, 43]. In these works, the presence of nanoparticles was observed to improve the stability of a foam. However, it is unclear their role on bubble generation [23, 39, 44]. From a modeling prospective, it is not established, yet, if constitutive equations for the bubble generation within a surfactant-stabilized foam can still be accounted for bubbles formed in a nanoparticle-stabilized foam. In this paper, a study of the foam evolution through a porous medium using a microfluidic system is presented. Earlier works, where microfluidic systems have been employed to visualize foam dynamics, aimed to observe foam generation, coalescence, and mobility as a function of permeability and fluid composition, e.g., [45, 46, 47, 48, 3, 49, 50, 51, 52]. To the best of the authors’ knowledge, this is the first study where the observation of foam evolution is carried out in conjunction with the determination of the generation kinetics by employing an algorithm based on convolutional neural network (CNN). Following previous experimental studies at larger scale [53], a relationship between rate of bubble generation and pressure was determined. Results show the generation of a foam stabilized with either a surfactant or surfactant and nanoparticles follows the same mechanism and can be described with a similar generation rate equation.

2

49

50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66

67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82

83 84 85 86 87 88

2. Materials and methods 2.1. Surface-active materials Two types of surface-active materials, namely, an anionic surfactant (sodium lauryl sulfate (SLS), Stepan, U.S.A.) and partially hydrophobic silica nanoparticles (DP 9711, Nyacol Nano Technologies, U.S.A.), were used in this work to stabilize a foam made of nitrogen gas (N2 , 99.99% purity, Welding Supply, U.S.A.). This surfactant and the nanoparticles were selected because successfully applied in previous studies to stabilize N2 -foam [54, 55, 44] as well as preliminary foamability tests were carried out to evaluate the performance of the selected surface-active materials with other commercial materials reported in the literature. Solutions of surfactant and suspensions (comprising surfactant and nanoparticles) were prepared using ultra-pure water (Milli-Q Reference, Germany). The concentration of SLS was fixed equal to 0.3 wt.%, which is slightly larger than the critical micellae concentration (CMC) observed to be effective for foam stability [56, 57, 58, 59]. Similarly, the concentration of nanoparticles was set equal to 0.3 wt.%, determined to be effective for foam stability upon preliminary tests on bulk and flowing foam and also selected in previous experimental works [55, 44]. Further details about the surface-active materials used in this work are reported in sections 1 through 4 of the Supporting Information (SI) document. 2.2. Microfluidic system Experiments were carried out using a microfluidic system schematically reproduced in Figure 1. It consists of a porous medium chip made of glass (Dolomite, U.K.), a piston pump (Cole-Parmer, U.S.A.), a pressure transducer (P61, Validyne Engineering, U.S.A.), a gas-flow controller (F-210CV, Bronkhorst, U.S.A.), and a microscope (E600, Nikon, U.S.A.). A high-resolution camera (MU1403, AmScope, U.S.A.) was mounted on the microscope to record the foam generation process. In particular, the porous medium chip was 60 mm long, 10 mm wide, and 4 mm tick. It consisted of repeated irregular rectangular grids. The pore-throat network was of circular shape and comprised pores of diameters equal to 0.110 mm and throats of 0.085 and 0.063 mm. An image of the chip is reported in section 5 of the SI document. Experiments of drainage and co-injection were run using this system, and the methods are described herein. Three types of compositions of the liquid-phase were considered, namely: ultra-pure water, surfactant solution, and surfactant solution with nanoparticles. The temperature in the room, where the tests were run, was always set at 20 ◦ C through the air conditioner system. 2.3. Co-injection experiments In these tests, the porous medium chip was initially saturated with either ultra-pure water, surfactant solution, or surfactant solution with nanoparticles. Then, it was flooded simultaneously with N2 -gas and a solution or a suspension of the same composition of that initially injected for saturation. Both co-injected phases, namely, the gaseous- and the liquid-phase, were supplied at constant rates changed between 0–0.7 and 0–0.5 cm3 /min, 3

Figure 1: Schematic of the microfluidic system used in the experiments. The symbols O1 through O6 indicate the six locations where the process was recorded through the microscope.

89 90 91 92

93 94 95 96 97 98 99 100 101 102 103 104

105 106 107 108 109 110 111 112 113 114

respectively. The tests were continuously monitored with the pressure transducer and the high-speed camera. The experiments were run at constant liquid phase rate while the gas rate was increased within the given range. Each test was carried out until stable pressure was reached, then the gas rate was increased. 2.4. Drainage experiments In these tests, the porous medium chip was initially saturated with either ultra-pure water, a surfactant solution, or a nanoparticle suspension with surfactant. To reach saturation the porous medium was flooded until no gas bubbles could be observed with the microscope. Then, gaseous N2 was injected at constant rate into the saturated porous medium chip. The applied gas rates (ug ) changed between 0.05 and 0.7 cm3 /min, which are representative of reservoir conditions [10]. Pressure drop and foam texture were recorded with an interval of a second. Each test was repeated six times and monitored at different locations along the porous medium chip as indicated in Figure 1. At each location, an area of 2.25×1.75 mm was observed with the microscope and recorded with the camera. To verify the reproducibility of the tests, pressure measurements were compared. An example is presented in section 6 of the SI document. 2.5. Image acquisition and processing To determine the evolution of foam texture, the bubble pattern needs to be recognized from recorded images. Current approaches are based either on the addition of aqueous dyes to the liquid solutions or on the adjustment of a contrast to distinguish the two phases [60, 61, 62]. However, it becomes inefficient to quantify flexible crowded bubbles, especially inside micro-pores. Therefore, in this work, a popular image processing method [63, 64] based on the convolutional neural network (CNN) model [65, 66] is introduced to count the number of bubbles. More in detail, during the image processing, the image is divided into S × S grid cells. Each grid cell has several bounding boxes to which location parameters are assigned. For each bounding box, local features with space structures are extracted using 4

115 116 117 118 119 120 121 122 123 124 125 126 127

a convolutional approach and compressed to multiple patterns. Then, these patterns are compared with those derived from the training sample and a confidence score of similarity is determined. The box containing the pattern associated to the highest score will be a successful recognition of the object of interest (in this case a bubble). Figure 2 illustrates the performance of CNN model after training with approximately 6,000 bubbles. Part a of the figure reports the initial saturated chip without bubbles. As it is possible to see, the CNN model identified no bubble, showing its ability to differentiate the shape of a bubble from the porous medium structure. Parts b and c of this figure display the porous medium chip with a coarse and a fine foam, respectively. In the case of the coarse foam, the CNN model was able to recognize all the bubbles, which are in total 26. In the second case instead, the model recognized only 76 over the 81 present. Resulting in an accuracy of approximately 94%, which is acceptable for the scope of the application of the CNN model in this work.

Figure 2: Bubble recognition using the CNN model in a microscopy image of the porous medium chip. (a) Chip saturated with surfactant solution without bubbles, (b) chip with surfactant solution and coarse foam, and (c) chip with surfactant solution and fine foam. Rectangular shapes represent the location and the size of the bubbles recognized by the CNN model and their color is related to their size.

128

129 130 131 132 133 134 135 136 137 138

139 140 141

3. Results and discussion In this section, the results from the experiments using the porous medium chip to study the evolution of the foam stabilized with either a surfactant or a surfactant and nanoparticles are presented and discussed. In section 3.1, the co-injection tests are described and in section 3.2, the drainage tests are shown. Finally, a mechanism for bubble generation within a porous medium is proposed together with a relationship between the generation rate of foam texture and the pressure gradient. Overall, the values of the capillary number (defined as [67]: Ca = µl ug /γls , with µl the viscosity of the liquid and γls the surface tension between the liquid and solid) varies between 0.213 and 1.67 × 10−4 , as expected for foam in microfluidics [68]. Details regarding the calculations of Ca are reported in section 7 of the SI document. 3.1. Co-injection experiments Co-injection experiments were carried out as described in section 2.3 and monitored with a high-speed camera and pressure transducer. The pressure applied during these tests was 5

142 143 144 145 146 147 148 149 150

increased by adjusting the gas rate between 0 and 0.7 cm3 /min. The maximum pressure that was reached corresponded to 0.45, 0.7 and 0.75 MPa when water, surfactant solution, and surfactant solution with nanoparticles were injected as liquid phase. Figure 3 reports the measured pressure drop over time for tests run injecting the gas together with either water, surfactant solution, or surfactant solution containing nanoparticles. As it is possible to see, the pressure within the system is much smaller when no surfactant was added than in the other cases. This is due to the formation of a coarse foam with negligible viscosity. This is also evident from the microscopy images reported as insets in the figure.

0.8

, MPa

0.7 0.6 0.5 0.4

Surfactant and nanoparticles Surfactant

Water

0.3 0

1

2

Time, s

3 10

4

Figure 3: Pressure drop as a function of time for three co-injection experiments. The rate of the solution was fixed at 0.5 cm3 /min and the gas rate corresponded to: 0, 0.05, 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, and 0.7 cm3 /min. The insets are microscopy images taken at the end of the tests, i.e., when the largest gas rate was applied.

151 152 153 154 155 156 157 158 159 160 161 162

Following the approach in Alvarez et al. [69], the results from these tests were reported in diagrams with on the axes the rates of the gas- and liquid-phases (Figure 4). The red lines correspond to the foam quality (fg = ug /(ug + uw )) that changes between 0 and 1. Here, the contours of the pressure drop (blue lines) and of the foam quality (red lines) in parts a and b resemble each other, suggesting a similar behaviour of the foam in the two investigated systems. However, these trends are significantly different from the ones in part c where only water was used as liquid phase. For instance, in part a when fg increases from 0 to 0.78, ∆P changes from 0.15 to 0.55 MPa; whereas, in part c where fg increases from 0 to 0.78, ∆P increases minimally, i.e., from 0.20 to 0.22 MPa. This is because a weak coarse foam was generated when only water was used, contrarily to the cases when surface active materials were employed. Similar observations could be made by comparing data in parts c with those in part b. 6

(a) 0.9

0.8 0.7

0.6

qg, cm3/min

0.6

0.6 0.5

0.4

0.5

0.1 0.4

0.2

0.3 0.3 0.1

0

(b) 0.9

0.8

0.6

0.7

qg, cm3/min

0.6 0.6

0.4 0.5 0.5 0.1

0.2

0.4

0.3

0.3 0.1

0

(c)

qg, cm3/min

0.6

0.7

0.8

0.6

0.9

0.4

0.5

0.1 0.4

0.2

0.3

0.3

0.2 0.1

0 0

0.1

0.2

0.3

0.4

0.5

3

ql, cm /min P, MPa

fg , (-)

Figure 4: Pressure drop as a function of gas and liquid injection rates for experiments run using: (a) surfactant solution, (b) surfactant solution and nanoparticles, and (c) water without surface-active materials. 7 and the foam quality (red lines). The contour lines indicate the pressure drop (blue lines)

163 164 165 166 167 168 169 170 171

Representative microscopy images are shown in Figure 6 for all three types of tests (highlighted with blue dots in Figure 4) when ∆P was equal to 0.3 MPa and fg increased up to 0.9. The foam texture was determined using the CNN model and the number of bubbles (n) within these images is provided in the insets. As it possible to read, at given fg , the values of n in the presence of only water are much smaller than those when surface-active materials were added. The value of n increases with foam quality in all cases, indicating the formation of a stronger foam as the gas fraction increases. This change is particularly significant in the presence of surface-active materials because the bubbles generated are stabilized by either the surfactant or the surfactant and the nanoparticles.

Figure 5: Microscopy images taken during co-injection tests at ∆P equal to 0.3 MPa run using: (a) surfactant solution, (b) surfactant solution and nanoparticles, and (c) only water without surfactant. n refers to the number of bubbles.

172 173 174

Figure 6 reports representative microscopy images for tests with fg equal to 0.5 and increasing pressure drop. As it is possible to see, the increase in ∆P and n is more significant in the presence of surface-active materials (parts a and b) than with just water (part c), as 8

175 176

expected, and there is little difference between the two cases. This suggests the formation of a foam with similar texture, regardless of the surface-active material used to stabilize it.

Figure 6: Microscopy images taken during co-injection tests at fg equal to 0.5 run using: (a) surfactant solution, (b) surfactant solution and nanoparticles, and (c) only water without surfactant. n refers to the number of bubbles.

177 178 179 180 181 182 183 184 185 186

3.2. Drainage experiments Drainage experiments were carried out as described in section 2.4 and monitored with a high-speed camera and pressure transducer. Examples of pictures taken during these tests are reported in Figure 7. In parts a and b of this figure, foam generation occurring during two experiments using surfactant, and surfactant and nanoparticles, respectively, is shown. In both cases, the porous medium chip was flooded at gas rate of 0.1 cm3 /min. The time when gas started to invade the microchip was set as initial time, i.e., 0 s. Each row of the figure corresponds to a certain time. Soon after the gas starts to invade the chip, foam forms and then travels through the medium in 26–28 s, displacing almost entirely the liquid-phase. No significant difference in the foam behavior can be observed here between 9

187 188 189 190 191

the two experiments. Contrarily to the test where no surface-active materials were added to the initial solution. In this case, fingering and formation of preferential paths along the walls of the medium are created (Figure 11 of the SI document). This is because of both the instability and the low viscosity of the gas-phase in comparison to the water, which can be overcome by the formation of foam. (a) 0 s

(b) 0 s 1

2

3

4

5

6

5s

5s

11 s

9s

14 s

15 s

18 s

20 s

22 s

23 s

26 s

28 s

Figure 7: Pictures of the porous medium chip taken during drainage experiments. The foam was stabilized with (a) surfactant, and with (b) surfactant and nanoparticles. The gas rate was fixed at 0.1 cm3 /min. The white shadow refers to the foam. 192 193 194 195 196 197 198 199 200 201 202 203 204

To gain an insight into the evolution of the foam texture, tests were repeated six times at the same conditions and monitored with a microscope at the different locations, 1 through 6, along the chip as indicated in Figure 1. An example of the images taken during runs where the gas rate was set at 0.1 cm3 /min is reported in Figure 8. Each row of this figure corresponds to a different location and each column to a different time. The images where selected in the early period of the generation when foam appeared for the first time within the observation area. As it is possible to see, there is a similar delay in the appearance of the foam in each location, making the whole image assemble almost a single test. The reproducibility of these experiments was confirmed by the pressure profiles, which are reported in section 6 of the SI document. The lines in the diagrams correspond to the average value with the error bar indicating the variability observed in the series of repetitions. It is evident that the difference between the pressure profiles is larger than the variability of the measurements, confirming the reproducibility of the tests at each fixed gas rate. 10

205 206 207 208

As it is possible to see in Figure 8, the gas bubbles can be distinguished from the liquidphase. To quantify their number and therefore gain an insight into the dynamics of the foam an image processing method based on artificial intelligence, explained in section 2.5, was applied.

Figure 8: Images of the porous medium chip taken during drainage experiments at different locations, namely O1 through O6. The foam was stabilized with (a) surfactant, and with (b) surfactant and nanoparticles. The gas rate was fixed at 0.1 cm3 /min.

209 210

Figure 9 reports the results upon the application of the CNN model to the recorded images. Here, the number of bubbles at each ith -observation area (ni ) as a function of time 11

211 212 213 214 215 216 217

is shown at given gas rate. In each quadrant, ni grows initially sharply and then reaches a plateau. However, a clear delay exists between locations, which decreases with the gas rate. The maximum value of ni increases with the gas rate, even though not monotonically throughout the tests. Overall, the maximum values reached by n1 , i.e., in the first observation area (O1), is lower than in the ith -locations downstream. This is because of the continuous generation of new bubbles during foam migration within the chip and the negligible coalescence at these conditions.

Figure 9: Number of bubbles at each ith -observation area (ni ) over time and at different gas rate. (a) through (e) foam stabilized with surfactant, (f) through (j) foam stabilized with surfactant and nanoparticles.

218 219 220

221 222 223 224 225 226 227 228 229

Assuming that the number of bubbles within the observation area resembles the entire location where it is centered, the total number of bubbles (N (t), #) within the medium writes, 6 X ni (t) N (t) = A, (1) a i=1 where a and A indicate the areas of the observation cell of the microscope (e.g., the area of location O1) and its extension to the corresponding region, which are equal to 4×10−2 and 1 cm2 , respectively (Figure 1). Figure 10 reports the results of this calculation, where N (t) as a function of time is given at each investigated gas rate. Overall, N increases at almost constant slope and reaches a plateau within 75 s, where the maximum foam textures change between 8×107 and 13×107 #/m2 . No significant difference can be observed between the tests run with either surfactant or surfactant and nanoparticles. The rate of generation and the maximum values of N increase with gas rate until 0.3 cm3 /min. Above this rate, a less pronounced difference among the profiles can be noticed. 12

Figure 10: Total number of bubbles (N ) as a function of time at given gas rate. Foam stabilized with (a) surfactant, and (b) surfactant and nanoparticles.

230 231 232 233 234 235 236 237 238 239

Pressure drop measured during selected experiments is reported in Figure 11 as a function of pore volume injected (i.e., P V = qg t/Vp , where t is time, Vp is the pore volume of the porous medium, equal to 0.038 cm3 ). Here, it is possible to see, that in all cases the pressure follows the same trend. It initially increases with a slope independently from the gas rate, reaches a maximum value around 500 PV, and then decreases stabilizing around a constant value which is approximately 0.05 MPa. Such large pressure values are due to the foam generation as in the absence of a foam the pressure within the system is negligible (dark red line). The maximum pressure drop increases with gas rate and higher pressures are observed in the experiments where surfactant and nanoparticles were used. This may due to a larger viscosity of the foam in the presence of nanoparticles, as previously observed [40, 70, 43].

13

Figure 11: Pressure drop as a function of pore volume injected (PV) at given gas rate. Foam stabilized with (a) surfactant, and (b) surfactant and nanoparticles.

240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265

Contrarily to bulk foams created by shear forces or turbulence, the generation of a foam in porous media is controlled by capillary forces. Three mechanisms of generation of a foam in porous media have been identified [2]: snap-off, lamella division, and leave-behind. A strong foam forms with snap-off and lamella division. The former occurs as the gas bubble expands decreasing capillary pressure (Pc , i.e., the pressure difference between the gas and the liquid phase at the gas-liquid interface) causing a pressure gradient moving the fluid. If Pc drops below a critical value, the accumulated liquid forms a gas bubble. Lamella division occurs when a lamella approaches a branch within its flow field, separating into two. Leavebehind is a foam formation mechanism that occurs below critical flow velocities and causes a coarse foam (or weak foam). The liquid is constrained into a lamella that might rupture or stay if stable. In these tests, all three mechanisms could be identified. Pictures taken during tests, where 0.7 cm3 /min of gas was injected, are shown in Figure 12. Foam readily forms after the injection of the gas into the pre-saturated chip with a solution containing surface-active materials. An initial coarse foam is generated due to snap-off followed by lamella division with formation of finer bubbles. While the pressure still increases within the medium (green line between 250 and 500 PV in Figure 11), residual lamellae left behind by the gas-flow start to appear. In this phase, pressure increases due to the increasing viscosity of the fluid. More in detail, it is possible to notice that between 3 and 300 PV, the foam created in the presence of only surfactant is coarser than that in the presence of nanoparticles. Moreover, more uniform size of bubbles could be observed in the latter case. The coarsening with stable lamellae left behind within the pore-throats appears after 200 and 400 PV in the case of surfactant and surfactant with nanoparticles. These differences in the two systems could be due to a more stable foam created in the presence of nanoparticles, as previously observed in other works [34, 71]. In the presence of nanoparticles, a slower kinetics of film drainage is imparted by capillary forces within the liquid films [72, 25]. Eventually, a residual pressure within the medium is established which 14

266 267 268 269 270

is approximately the same regardless of the applied surface-active material and also gas rate (Figure 11). In these tests, no minimum pressure was necessary to be overcome in order to generate a strong foam. Such an observation was also made in an earlier work by Gauglitz et al. [53] for N2 -gas foam. However, the pressure evolution and the number of bubbles generated (Figure 10) do depend on the gas rate.

Figure 12: Microscopy images of the observation area O1 during tests at qg equal to 0.7 cm3 /min and using (a) surfactant (b) surfactant and nanoparticles as surface-active materials.

15

271 272 273 274 275

Following previous formulations of generation kinetics [53, 73, 74], the bubble number as a function of measured pressure difference was determined at each applied gas rate and the results are reported in Figure 13. Overall, the generation rate increases with the pressure gradient as: rg ∝ ∇P α , with α an exponent. But, under the same conditions, a larger number of bubbles can be generated with surfactant than with surfactant and nanoparticle. 1010

5

rg, #/m 3s

4

3

2

1 Surfactant Surfactant and nanoparticles 0 0

0.3

0.6

0.9

1.2

1.5

P MPa/m Figure 13: Generation rate (rg , #/m3 ) as a function of pressure gradient (∇P , MPa/m) measured at each applied gas rate. Dashed lines are only a visual guide.

276

277 278 279 280 281 282 283 284 285 286 287 288 289 290 291

4. Conclusions In this paper, a study on the generation of foam in porous media using as surface-active materials either an anionic surfactant or an anionic surfactant with partially hydrophobic silica nanoparticles is reported. Co-injection and drainage type of tests were carried out and they suggest overall that a strong similarity exists between the generation mechanism occurring during the generation of a foam in the presence of either surfactant or surfactant and nanoparticles. Microscopy images show a generation process characterized by early snapoff, followed by lamella-division, and finally leave-behind, which results in thin lamellae left within the pore throats. In the presence of nanoparticles, the latter stage is delayed due to resistance to drainage imparted by capillary forces within the liquid films. An algorithm based on convolutional neural network (CNN) model for object recognition was employed to determine the evolution of the foam texture during the drainage tests. Following a previous experimental study, a relationship between the generation rate and pressure gradient was derived. It resembles the classical constitutive equation for foam generation [18, 19, 20, 21, 22], i.e., rg ∝ ∇P α , indicating that the generation of a foam in the presence of nanoparticles and surfactant follows the same mechanism of a foam stabilized only with surfactant. From 16

299

the modeling prospective this is important for the formulation of the constitutive equation of foam generation. It is suggested that classical constitutive equation for generation can still be regarded for a foam created in the presence of nanoparticles. To the best of the authors’ knowledge this is the first time that such a relationship is derived from the quantification of foam bubbles within a porous medium. Future work will include the investigation of foam behavior at limiting conditions of bubble rupture using surfactant, nanoparticles, and a combination of them. The results will help to formulate a constitutive equation for foam coalescence.

300

Acknowledgements

292 293 294 295 296 297 298

303

The authors would like to thank the American Chemical Society Petroleum Research Fund (ACS-PRF) under the grant number PRF# 57739-DNI9 and the Innovation & Entrepreneurial Fellowship Program at Stevens Institute of Technology for financial support.

304

References

301 302

305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334

[1] W. R. Rossen, Foams in enhanced oil recovery, Foams: Theory, Measurements and Applications 57 (1996) 413–464. [2] T. Ransohoff, C. Radke, Mechanisms of foam generation in glass-bead packs, SPE Reservoir Eng. 3 (02) (1988) 573–585. [3] C. A. Conn, K. Ma, G. J. Hirasaki, S. L. Biswal, Visualizing oil displacement with foam in a microfluidic device with permeability contrast, Lab Chip 14 (20) (2014) 3968–3977. [4] R. F. Li, W. Yan, S. Liu, G. Hirasaki, C. A. Miller, et al., Foam mobility control for surfactant enhanced oil recovery, Soc. Petrol. Eng. J. 15 (04) (2010) 928–942. [5] L. Lake, Enhanced oil recovery, Prentice Hall, 1989. [6] G. Hirasaki, R. Jackson, M. Jin, J. Lawson, J. Londergan, H. Meinardus, C. Miller, G. Pope, R. Szafranski, D. Tanzil, Field demonstration of the surfactant/foam process for remediation of a heterogeneous aquifer contaminated with dnapl, NAPL Removal: Surfactants, Foams, and Microemulsions (2000) 3–163. [7] C. N. Mulligan, S. Wang, Remediation of a heavy metal-contaminated soil by a rhamnolipid foam, Eng. Geol. 85 (1) (2006) 75–81. [8] C. Portois, C. S. Boeije, H. J. Bertin, O. Atteia, Foam for environmental remediation: Generation and blocking effect, Transport Porous Med. 124 (3) (2018) 787–801. [9] A. Kovscek, C. Radke, Fundamentals of foam transport porous med., ACS Publications, 1994. [10] L. W. Lake, R. T. Johns, W. R. Rossen, G. A. Pope, Fundamentals of enhanced oil recovery, Society of Petroleum Engineers, 2014. [11] M. J. Falls, A. H., J. Ratulowskl, The apparent viscosity of foams in homogeneous bead packs, SPE Reservoir Eng. 4 (2) (1989) 155–164. [12] G. Hirasaki, J. Lawson, Mechanisms of foam flow in porous media: apparent viscosity in smooth capillaries, Soc. Petrol. Eng. J. 25 (02) (1985) 176–190. [13] A. I. Jim´enez, C. J. Radke, Dynamic Stability of Foam Lamellae Flowing Through a Periodically Constricted Pore, 1989, Ch. 25, pp. 460–479. [14] Z. Khatib, G. Hirasaki, A. Falls, Effects of capillary pressure on coalescence and phase motilities in foams flowing through porous media, SPE Reservoir Eng. (1998) 919–926. [15] A. Aronson, V. Bergeron, M. Fagan, C. Radke, The influence of disjoining pressure on foam stability and flow in porous-media, Colloids Surf. A 83 (2) (1994) 109–120.

17

335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385

[16] A. S. Elhag, C. Da, Y. Chen, N. Mukherjee, J. A. Noguera, S. Alzobaidi, P. P. Reddy, A. M. AlSumaiti, G. J. Hirasaki, S. L. Biswal, Q. P. Nguyen, K. P. Johnston, Viscoelastic diamine surfactant for stable carbon dioxide/water foams over a wide range in salinity and temperature, J. Colloid Interface Sci. 522 (2018) 151–162. [17] R. Pugh, Bubble and foam chemistry, Cambridge University Press, 2016. [18] A. Kovscek, T. Patzek, C. Radke, A mechanistic population balance model for transient and steadystate foam flow in boise sandstone, Chem. Eng. Sci. 50 (23) (1995) 3783–3799. [19] S. Kam, W. Rossen, et al., A model for foam generation in homogeneous media, Soc. Petrol. Eng. J. 8 (04) (2003) 417–425. [20] S. I. Kam, Q. P. Nguyen, Q. Li, W. R. Rossen, et al., Dynamic simulations with an improved model for foam generation, Soc. Petrol. Eng. J. 12 (01) (2007) 35–48. [21] S. I. Kam, Improved mechanistic foam simulation with foam catastrophe theory, Colloids Surf. A 318 (1) (2008) 62–77. [22] E. Ashoori, D. Marchesin, W. Rossen, Roles of transient and local equilibrium foam behavior in porous media: traveling wave, Colloids Surf. A 377 (1) (2011) 228–242. [23] Z. Du, M. P. Bilbao-Montoya, B. P. Binks, E. Dickinson, R. Ettelaie, B. S. Murray, Outstanding stability of particle-stabilized bubbles, Langmuir 19 (8) (2003) 3106–3108. [24] A. C. Martinez, E. Rio, G. Delon, A. Saint-Jalmes, D. Langevin, B. P. Binks, On the origin of the remarkable stability of aqueous foams stabilised by nanoparticles: link with microscopic surface properties, Soft Matter 4 (7) (2008) 1531–1535. [25] F. Carn, A. Colin, O. Pitois, M. Vignes-Adler, R. Backov, Foam drainage in the presence of nanoparticlesurfactant mixtures, Langmuir 25 (14) (2009) 7847–7856. [26] M. Azmin, G. Mohamedi, M. Edirisinghe, E. Stride, Dissolution of coated microbubbles: The effect of nanoparticles and surfactant concentration, Mater. Sci. Eng. C 32 (8) (2012) 2654–2658. [27] M. Lotfollahi, I. Kim, M. R. Beygi, A. J. Worthen, C. Huh, K. P. Johnston, M. F. Wheeler, D. A. DiCarlo, Foam generation hysteresis in porous media: Experiments and new insights, Transport Porous Med. 116 (2) (2017) 687–703. [28] N. Yekeen, M. A. Manan, A. K. Idris, E. Padmanabhan, R. Junin, A. M. Samin, A. O. Gbadamosi, I. Oguamah, A comprehensive review of experimental studies of nanoparticles-stabilized foam for enhanced oil recovery, J. Petrol. Sci. Eng. 164 (2018) 43–74. [29] M. Zhao, R. Wang, C. Dai, X. Wu, Y. Wu, Y. Dai, Y. Wu, Adsorption behaviour of surfactantnanoparticles at the gas-liquid interface: Influence of the alkane chain length, Chemical Engineering Science 206 (2019) 203–211. [30] R. Aveyard, J. Clint, D. Nees, Small solid particles and liquid lenses at fluid/fluid interfaces, Colloid Polym. Sci. 278 (2) (2000) 155–163. [31] B. P. Binks, Particles as surfactantssimilarities and differences, Curr. Opin. Colloid Interface Sci. 7 (1-2) (2002) 21–41. [32] G. Kaptay, On the equation of the maximum capillary pressure induced by solid particles to stabilize emulsions and foams and on the emulsion stability diagrams, Colloids Surf. A 282 (2006) 387–401. [33] A. B. Subramaniam, C. Mejean, M. Abkarian, H. A. Stone, Microstructure, morphology, and lifetime of armored bubbles exposed to surfactants, Langmuir 22 (14) (2006) 5986–5990. [34] B. P. Binks, M. Kirkland, J. A. Rodrigues, Origin of stabilisation of aqueous foams in nanoparticle– surfactant mixtures, Soft Matter 4 (12) (2008) 2373–2382. [35] S. Guignot, S. Faure, M. Vignes-Adler, O. Pitois, Liquid and particles retention in foamed suspensions, Chemical Engineering Science 65 (8) (2010) 2579–2585. [36] B. P. Binks, J. H. Clint, Solid wettability from surface energy components: Relevance to pickering emulsions, Langmuir 18 (4) (2002) 1270–1273. [37] G. Kaptay, Interfacial criteria for stabilization of liquid foams by solid particles, Colloids Surf. A 230 (1-3) (2003) 67–80. [38] A. J. Worthen, S. L. Bryant, C. Huh, K. P. Johnston, Carbon dioxide-in-water foams stabilized with nanoparticles and surfactant acting in synergy, AIChE J. 59 (9) (2013) 3490–3501.

18

386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436

[39] D. A. Espinoza, F. M. Caldelas, K. P. Johnston, S. L. Bryant, C. Huh, et al., Nanoparticle-stabilized supercritical co2 foams for potential mobility control applications, in: SPE Improved Oil Recovery Symposium, Society of Petroleum Engineers, 2010. [40] A. J. Worthen, H. G. Bagaria, Y. Chen, S. L. Bryant, C. Huh, K. P. Johnston, Nanoparticle-stabilized carbon dioxide-in-water foams with fine texture, J. Colloid Interface Sci. 391 (2013) 142–151. [41] J. Yu, M. Khalil, N. Liu, R. Lee, Effect of particle hydrophobicity on co2 foam generation and foam flow behavior in porous media, Fuel 126 (2014) 104–108. [42] I. Kim, A. J. Worthen, K. P. Johnston, D. A. DiCarlo, C. Huh, Size-dependent properties of silica nanoparticles for pickering stabilization of emulsions and foams, J. Nanoparticle Res. 18 (4) (2016) 82. [43] V. Prigiobbe, A. J. Worthen, K. P. Johnston, C. Huh, S. L. Bryant, Transport of nanoparticle-stabilized co2 -foam in porous media, Transport Porous Med. 111 (1) (2016) 265–285. [44] R. Singh, K. K. Mohanty, Foam flow in a layered, heterogeneous porous medium: A visualization study, Fuel 197 (2017) 58–69. [45] A. M. Ga˜ n´ an-Calvo, J. M. Gordillo, Perfectly monodisperse microbubbling by capillary flow focusing, Phys. Rev. Lett. 87 (27) (2001) 274501. [46] P. Garstecki, M. J. Fuerstman, H. A. Stone, G. M. Whitesides, Formation of droplets and bubbles in a microfluidic t-junctionscaling and mechanism of break-up, Lab Chip 6 (3) (2006) 437–446. [47] J. I. Park, D. Jagadeesan, R. Williams, W. Oakden, S. Chung, G. J. Stanisz, E. Kumacheva, Microbubbles loaded with nanoparticles: a route to multiple imaging modalities, Acs Nano 4 (11) (2010) 6579–6586. [48] H. Chen, J. Li, W. Zhou, E. G. Pelan, S. D. Stoyanov, L. N. Arnaudov, H. A. Stone, Sonication– microfluidics for fabrication of nanoparticle-stabilized microbubbles, Langmuir 30 (15) (2014) 4262– 4266. [49] Q. Sun, Z. Li, S. Li, L. Jiang, J. Wang, P. Wang, Utilization of surfactant-stabilized foam for enhanced oil recovery by adding nanoparticles, Energ. Fuel 28 (4) (2014) 2384–2394. [50] M. M. Almajid, A. R. Kovscek, Pore-level mechanics of foam generation and coalescence in the presence of oil, Adv. Colloid Interfac. 233 (2016) 65–82. [51] K. Osei-Bonsu, P. Grassia, N. Shokri, Investigation of foam flow in a 3D printed porous medium in the presence of oil, J. Colloid Interface Sci. 490 (2017) 850–858. [52] D. Vecchiolla, V. Giri, S. L. Biswal, Bubble–bubble pinch-off in symmetric and asymmetric microfluidic expansion channels for ordered foam generation, Soft Matter 14 (46) (2018) 9312–9325. [53] P. A. Gauglitz, F. Friedmann, S. I. Kam, W. R. Rossen, Foam generation in homogeneous porous media, Chem. Eng. Sci. 57 (19) (2002) 4037–4052. [54] X. Duan, J. Hou, T. Cheng, S. Li, Y. Ma, Evaluation of oil-tolerant foam for enhanced oil recovery: laboratory study of a system of oil-tolerant foaming agents, J. Petrol. Sci. Eng. 122 (2014) 428–438. [55] R. Singh, K. K. Mohanty, Synergy between nanoparticles and surfactants in stabilizing foams for oil recovery, Energ. Fuel 29 (2) (2015) 467–479. [56] Y.-F. Chen, C.-H. Yang, M.-S. Chang, Y.-P. Ciou, Y.-C. Huang, Foam properties and detergent abilities of the saponins from camellia oleifera, Int. J. Mol. Sci. 11 (11) (2010) 4417–4425. [57] Q.-W. Yang, Z.-L. Wu, Y.-L. Zhao, Y. Wang, R. Li, Enhancing foam drainage using foam fractionation column with spiral internal for separation of sodium dodecyl sulfate, J. Hazard. Mater. 192 (3) (2011) 1900–1904. [58] N. Yekeen, A. K. Idris, M. A. Manan, A. M. Samin, Experimental study of the influence of silica nanoparticles on the bulk stability of sds-foam in the presence of oil, J. Disper. Sci. Technol. 38 (3) (2017) 416–424. [59] N. Yekeen, A. K. Idris, M. A. Manan, A. M. Samin, A. R. Risal, T. X. Kun, Bulk and bubble-scale experimental studies of influence of nanoparticles on foam stability, Chinese J. Chem. Eng. 25 (3) (2017) 347–357. [60] K. Ma, R. Liontas, C. A. Conn, G. J. Hirasaki, S. L. Biswal, Visualization of improved sweep with foam in heterogeneous porous media using microfluidics, Soft Matter 8 (41) (2012) 10669–10675. [61] K. Osei-Bonsu, P. Grassia, N. Shokri, Investigation of foam flow in a 3d printed porous medium in the

19

437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462

presence of oil, Journal of colloid and interface science 490 (2017) 850–858. [62] M. J. Shojaei, K. Osei-Bonsu, S. Richman, P. Grassia, N. Shokri, Foam stability influenced by displaced fluids and by pore size of porous media, Industrial & Engineering Chemistry Research 58 (2) (2018) 1068–1074. [63] J. Redmon, A. Angelova, Real-time grasp detection using convolutional neural networks, CoRR abs/1412.3128. [64] J. Redmon, S. Divvala, R. Girshick, A. Farhadi, You only look once: Unified, real-time object detection, in: Proceedings of the IEEE conference on computer vision and pattern recognition, 2016, pp. 779–788. [65] Y. LeCun, Y. Bengio, G. Hinton, Deep learning, Nature 521 (7553) (2015) 436. [66] I. Goodfellow, Y. Bengio, A. Courville, Deep learning, MIT press, 2016. [67] S. Khodaparast, M. K. Kim, J. E. Silpe, H. A. Stone, Bubble-driven detachment of bacteria from confined microgeometries, Environmental Science & Technology 51 (3) (2017) 1340–1347. [68] A. Huerre, V. Miralles, M.-C. Jullien, Bubbles and foams in microfluidics, Soft Matter 10 (2014) 6888– 6902. [69] J. Alvarez, H. Rivas, W. Rossen, et al., Unified model for steady-state foam behavior at high and low foam qualities, Soc. Petrol. Eng. J. 6 (03) (2001) 325–333. [70] J. Yu, M. Khalil, N. Liu, R. Lee, Effect of particle hydrophobicity on CO2 foam generation and foam flow behavior in porous media, Fuel 126 (2014) 104–108. [71] Q. Sun, Z. Li, S. Li, L. Jiang, J. Wang, P. Wang, Utilization of surfactant-stabilized foam for enhanced oil recovery by adding nanoparticles, Energ. Fuel 28 (4) (2014) 2384–2394. [72] T. S. Horozov, Foams and foam films stabilised by solid particles, Current Opinion in Colloid Interface Science 13 (3) (2008) 134–140. [73] T. Dicksen, G. J. Hirasaki, C. A. Miller, Conditions for foam generation in homogeneous porous media, Soc. Petrol. Eng. J. [74] S. I. Kam, Improved mechanistic foam simulation with foam catastrophe theory, Colloids Surf. A 318 (1) (2008) 62–77.

20

•

Foam generation in porous media in the presence of nanoparticle was studied

•

A microfluidic system was employed to visualize the bubble formation

•

A convolutional neural network model was used to quantify the bubbles

•

Microscopy images show early snap-off followed by lamella-division and leave-behind

•

A general relationship between the generation rate and the pressure was formulated

Declaration of interests ☒ The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. ☐The authors declare the following financial interests/personal relationships which may be considered as potential competing interests:

Author Contribution Statement Qingjian Li: Investigation, Methodology, Software, Validation, Visualization. Valentina Prigiobbe: Funding acquisition, Resources, Supervision, Project administration. Both authors contribute to Writing.