Geo-material surface modification of microchips using layer-by-layer (LbL) assembly for subsurface energy and environmental applications

Y. Q. Zhang ab, A. Sanati-Nezhad *bc and S. H. Hejazi *a
aSubsurface Fluidics and Porous Media Laboratory, Chemical and Petroleum Engineering, University of Calgary, Calgary, AB T2N 1N4, Canada. E-mail: shhejazi@ucalgary.ca
bBioMEMS and Bioinspired Microfluidic Laboratory, Department of Mechanical and Manufacturing Engineering, University of Calgary, Calgary, AB T2N 1N4, Canada. E-mail: amir.sanatinezhad@ucalgary.ca
cCentre for Bioengineering Research and Education, University of Calgary, Calgary, AB T2N 1N4, Canada

Received 28th June 2017 , Accepted 13th November 2017

First published on 17th November 2017


A key constraint in the application of microfluidic technology to subsurface flow and transport processes is the surface discrepancy between microchips and the actual rocks/soils. This research employs a novel layer-by-layer (LbL) assembly technology to produce rock-forming mineral coatings on microchip surfaces. The outcome of the work is a series of ‘surface-mimetic micro-reservoirs (SMMR)’ that represent multi-scales and multi-types of natural rocks/soils. For demonstration, the clay pores of sandstones and mudrocks are reconstructed by representatively coating montmorillonite and kaolinite in polydimethylsiloxane (PDMS) microchips in a wide range of channel sizes (width of 10–250 μm, depth of 40–100 μm) and on glass substrates. The morphological and structural properties of mineral coatings are characterized using a scanning electron microscope (SEM), optical microscope and profilometer. The coating stability is tested by dynamic flooding experiments. The surface wettability is characterized by measuring mineral oil–water contact angles. The results demonstrate the formation of nano- to micro-scale, fully-covered and stable mineral surfaces with varying wetting properties. There is an opportunity to use this work in the development of microfluidic technology-based applications for subsurface energy and environmental research.


1 Introduction

Rocks and soils, often referred to as subsurface porous media, are made up of rock-forming minerals. They have an abundance of pores and microfractures ranging from nanometers to centimeters in size. Within the pore spaces and microfractures, formation fluids such as oil, water and gas form, migrate and are stored; external fluids like carbon dioxide (CO2), oil displacing agents and groundwater containing radionuclides invade, react and remain. These authigenic and extraneous fluids have been investigated in many crucial energy and environmental applications: fossil fuel resources exploitation,1,2 geological storage of carbon dioxide,3,4 radioactive waste disposal5 and soil science.6,7

Lab-on-a-chip or microfluidic technology, which performs laboratory operations in miniaturized flow cells, has exhibited numerous functionalities and potentials to address subsurface challenges.3,5–13 Other than the most well-known advantages such as real-time observation, microfluidics presents a unique capability of reconstructing two- and three-dimensional (2D and 3D) pore network geometries.14–20 However, the pore surface features normally have poor reflections in microchips made from conventional polymeric or glass materials. The pore structure is a dominant factor that controls the physics of flow in porous rocks and soils, the pore surfaces, which act as the flow boundaries or reaction sites, also play a critical role in the flow and transport processes.

Rock/soil surface chemistries affect the subsurface mass transfer phenomena in a variety of ways. From a macroscopic perspective, mineral surfaces have complex affinities to formation fluids and demonstrate distinct wetting properties. In oil and gas recovery research, fluid distribution and displacement efficiency are closely linked to the rock's surface wettability. At the microscale, the molecules of contiguous fluids could interact with a solid surface by adsorption and desorption. Examples are the adsorption of certain oil components to clay minerals,21,22 physiochemical reactions between injected CO2 and rocks,23 the adsorption and retardation of repository-released radionuclides on geologic media5 and adsorption storage of methane in coal-bed and shale strata.24,25 Thus, the rock/soil surface chemistries can be placed at a pivotal position in subsurface porous medium research. This is mostly due to a ‘size reduction effect’,26 with the ubiquity of the micro- and nano-scale pore spaces and the large specific surface area in porous media.

There are many scientific questions related to subsurface fluid–solid interactions that have not been considered in microfluidic systems due to the poor surface representation within microchips. For instance, current pore-scale investigations on CO2 enhanced oil recovery (EOR) or CO2 geological storage, using conventional or microfluidic methods, mostly focus on multi-scale CO2 invasion processes,27–29 solubility trapping mechanisms,3,30 and CO2-brine mass transfer predictions.31–33 Even though there is a recognition that the reactivity between CO2 and silicate minerals is a significant CO2 trapping mechanism,3,29–31,34,35 the microscale observation for the fluid–solid interactions is rarely performed through the microfluidic technique. This may be due to the lack of constitutive surface properties in microfluidic chips. There is a need to develop rock- and soil-mimetic surface chemistries in microchips using advanced microfabrication and surface modification techniques.

Various strategies have been employed to analyze the microscale fluid–solid interactions in microfluidic systems. The approaches are categorized based on three principal paths for modeling the rock/soil surface properties in microchips. The first approach is to pack mineral particles into microchannels and then use an X-ray tomography technique to visualize mineral–fluid interactions in three dimensions.36–38 The second technique incorporates geomaterials into the microfabrication of micromodels. In this method, natural rocks like shale5,12 and rock-forming minerals such as quartz,39 calcite16 and granite5 are used as substrates to fabricate microchips. These rocks and minerals are cut and ground into thin slices and then subjected to an etching process to form flow channels. The advantage of this method is the inclusion of the rock chemical composition in the microchips. Unfortunately, the etching process on these materials is costly and the minimum channel size may not match those created by soft lithography with polymeric materials. The third approach is to modify the surfaces of microchips made of conventional materials such as glass or polymers. The modification can chemically change the microchip surface wettability10,40–42 or physically coat the surface with geomaterials.43–45 Chemical modifications are effective for creating single or mixed wettability on the surfaces for biomedical or material synthesis applications. However, chemical modification approaches normally fail to replicate the naturally occurring heterogeneous wettability in microchips.40,46

Kaolinite and montmorillonite are the commonly used minerals to coat micromodel surfaces by particle aggregation and deposition.43–45,47–49 The clay-functionalized micromodels or substrates are equipped with certain mineral properties. They are used to study low-salinity water flooding, wettability alteration in the presence of clay and clay-related formation damage in hydrocarbon production processes mainly. A major limitation for this approach is that the sealed microchip channel walls are typically partially covered by clay minerals and the coatings have poor stability under fluid flow conditions.43–45 Even though the geomaterial coatings on glass substrates demonstrate stability under slight flooding conditions,47–49 the case for microchips must still be verified. Corresponding to this, there are methods employed to reduce the mobility of coated mineral particles in microchip channels. One such method is to fix kaolinite particles on silicon substrates by thermal treatment.43 However, what is still required is versatile and simple protocols for the geomaterial coating modification of microchips in order to achieve real rock/soil surface chemistries.

Layer-by-layer (LbL) assembly is a prevalent method for the assembly of coatings. It uses alternating depositions of two or more oppositely-charged materials onto charged substrates.50–52 Over the past two decades, the LbL assembly has received increased attention from many fields of study: material synthesis,53,54 energy,55 optics,56 biomedicine,57–60 membranes,61,62 and nanoelectronics.63 Theoretically, every substrate that can have surface charges can be modified using the LbL assembly method. This includes substrates with arbitrary shapes, e.g., planar, particulate, irregular and 3D structures. Modification using the LbL assembly is expected to provide the capability to build rock-mimetic mineral coatings because of its versatility with the inherent properties of geomaterials.

There are three factors which ensure the feasible and universal modification for the geomaterial coatings in microchips using the LbL assembly technique. First, microchips made of microfabrication materials (polymers, glass, silicon and PMMA) can be handily treated to possess charges by exposing them to plasma or oxidizing chemicals like a piranha solution. The second factor is the geomaterials' charging properties. Most geomaterials, including clays, silica and organics, are naturally charged under formation conditions. The third factor is that microchannels can be easily exposed to assembly materials with their interconnectivity feature.

In this study, the LbL assembly technique is applied to microfluidic chips to develop subsurface mimetic porous systems. As representatives, two clays of montmorillonite and kaolinite are coated on polydimethylsiloxane (PDMS) and glass surfaces. The coated microchannels act like clay pores in sandstones and mudrocks, demonstrating the validity of the developed method. The morphological, structural, stability and wetting properties of the formed mineral coatings are characterized by a scanning electron microscope (SEM), optical microscope, profilometer and contact angle meter. Microchips composed of channels (width of 10–250 μm and depth of 40–100 μm) are coated to demonstrate the applicability of this method to multiphase flow experiments. The output of this work is a series of custom-built ‘surface-mimetic micro-reservoirs (SMMR)’, providing new tools for the field of subsurface energy and environmental research.

2 Experimental section

2.1 Geological bases

During the diagenesis process of sedimentary rocks, the matrix and cement substances cover the detrital grains either as pore-lining or pore-bridging materials. They can also occur as pore-filling materials in interstitial spaces amongst the framework grains.64,65 Clay minerals are common components that exist in both cement and matrix materials.66 The clay pores are formed during the sedimentary history. They can be macro intergranular pores where clay particles arrange tangentially on the bedding surfaces of framework grains,67 or intracrystalline pores within well-crystallized clay floccules.

In mudrocks, clay pores are abundant as the fine-sized clays are normally the primary mineral groups. Clay minerals may take an average proportion of 60% or more of the rock surface area by weight,22 especially in continental shale formations. In Fig. 1, a backscattered electron (BSE) imaging and the matching quantitative evaluation of minerals by scanning electron microscopy (QEMSCAN) of an illitic shale sample are provided.68 As seen from the figures, a large body of intracrystalline pores are likely to be woven into a lot of flocculated clay minerals, as the illite floccules can be indicative of pore spaces.69 In sandstones, clay pores are not as abundant as in mudrocks due to the less content of clay minerals. However, the clay pores in sandstones can primarily impact hydrocarbon storage and migration, the formation response to the injection of water-based fluids (flooding and fracturing fluids) and thus, a range of oil recovery issues.


image file: c7lc00675f-f1.tif
Fig. 1 Backscattered electron (BSE) image (a) and the matching quantitative evaluation of minerals by scanning electron microscopy (QEMSCAN) image (b) of an illitic shale sample. Reproduced from ref. 68 with permission from the American Association of Petroleum Geologists.

In this study, the microchip surfaces are coated with clay minerals that exist in common sedimentary rocks, sandstones and mudrocks. For brevity, montmorillonite and kaolinite are used as the geomaterials. The technique can be easily applied to other minerals.

2.2 Materials

All of the materials utilized in this work are from Sigma-Aldrich (St. Louis, MO, USA). There is one exception, the fluorescent red dye (Kingscote 506[thin space (1/6-em)]250-RF4) was purchased from Cole-Parmer (Canada Inc.). The minerals used in this work are montmorillonite (surface area of 240 m2 g−1 and mean diameter of 303 nm) and kaolinite (surface area of 9 m2 g−1 and mean diameter of 678.6 nm). The salts used for the brine preparation include NaCl, KCl, CaCl2·2H2O and MgCl2·6H2O.

The protocol used to prepare stable montmorillonite and kaolinite dispersions is as follows: 1) disperse clay minerals in deionized (DI) water, with concentrations ranging from 0.3–0.5 wt%; 2) sonicate the dispersions for 60 minutes (Q125, QSonica); 3) preserve the dispersions for 10 hours under room conditions; and 4) decant the dispersions and keep the supernatants for usage or for further treatment by repeating the sonication, preservation and decantation processes as needed. To prevent the aggregation and deposition of the particles, freshly-made clay dispersions are always used for instant LbL assembly procedures.

Poly(diallyldimethylammonium chloride) (PDDA) is a commonly used polyelectrolyte for LbL assembly53,57 and is used in this work. In a previous study,70 the concentration of PDDA had a significant effect on the coating thickness and surface roughness. However, no quantified results are available. In our study, PDDA, with a concentration of 0.2 wt%, provided the desired coatings in the microchips.

2.3 Layer-by-layer (LbL) assembly routes

The mechanism for the LbL assembly technology relies on the reverse charges of the substrate surface and coating materials and the resulting electrostatic fields. PDMS substrates are prepared by mixing the PDMS base and the crosslinking agent in a ratio of 10[thin space (1/6-em)]:[thin space (1/6-em)]1 and curing the mixture in an oven at 80 °C for eight hours. It has a strong negative charge after oxygen plasma oxidation. Rock-forming minerals are primarily negatively-charged under aqueous formation conditions. In this study, the geomaterials and PDDA molecules, which are positively-charged, are alternately deposited on the plasma-treated PDMS substrate. They follow proper assembly routes as outlined below. Under the electrostatic forces, the desired multilayer mineral coatings are built on the substrates by repeating the assembly procedure for multiple cycles.

To date, five broad categories of LbL assembly routes have been developed: immersive, spin, spray, electromagnetic and fluidic assembly.71–78 These assembly routes cover a variety of different materials, geometries and scales of coatings for diverse applications. Microchips' target surfaces to be modified are usually microscale flow channels or complex fluidic networks within confined spaces. In consideration of the geometries specific, we adopted the fluidic and immersive assembly techniques for the microchips modification.

The schematic of the fluidic LbL assembly on the PDMS substrate is illustrated in Fig. 2a. Oxygen plasma treatment is used for the tight bonding of the PDMS fluidic network and the glass slide. After plasma treatment, the PDMS surfaces carry strong negative charges (Fig. 3a). They are ready to adsorb the positively-charged PDDA molecules. As the charges remain temporarily on the PDMS surface, the primary injection of the PDDA solution should begin within 10 minutes after the plasma bonding.


image file: c7lc00675f-f2.tif
Fig. 2 Schematics of the layer-by-layer (LbL) assembly routes in microchips. (a) Fluidic LbL assembly in microchips. Poly(diallyldimethylammonium chloride) (PDDA) solution, deionized (DI) water and the clay dispersion are independently and alternately injected into the microchips through three inlets. Three programmable pumps are used to automate the assembly process. (b) Immersive LbL assembly in microchips. A microchip is dipped into the PDDA solution and clay dispersion alternately; a thorough DI water rinse is conducted between the two steps for removing the loosely-attached particles or molecules and facilitating the rearrangement of the adsorbed materials. The remaining liquid on microchips in each step is dried to speed up the procedures.

image file: c7lc00675f-f3.tif
Fig. 3 Schematic of the LbL assembly of clay minerals on PDMS surfaces. (a) PDMS surface carries negative charges after oxygen plasma. (b) PDDA molecules are adsorbed on the PDMS surface by electrostatic forces. (c) Clay particles are adsorbed by PDDA molecule chains via electrostatic forces. (d) Multi-layer mineral coating is formed after n cycles of the LbL assembly.

The PDDA molecule chains carry multiple positive charges and interact with the negatively-charged PDMS surface under the electrostatic forces (Fig. 3b). When the negatively-charged clay particles are induced, one or more clay particles are adsorbed by each long PDDA molecule chain via electrostatic force. They form an imbricate-like arrangement (Fig. 3c). A thorough DI water rinse is conducted between every two assembly processes to remove the loosely-attached materials and facilitate the rearrangement of the adsorbed particles/molecules. After one cycle of the LbL assembly (PDDA injection–DI water rinse–clay dispersion injection–DI water rinse), a bilayer of the coating, labeled as (P/C)1, is created. Repeating the cycles for n times produces a mineral coating with n bilayers, (P/C)n (Fig. 3d). Using a specially designed 3-inlet microchip, the PDDA solution, clay dispersion and the flushing agent (DI water) are injected independently and alternately into the microchip through the inlets. Three programmable syringe pumps (PHD ULTRA 70-3007, Harvard Apparatus) are used to automate the injection processes with a consistent injection rate. For each injection, the injection volume is greater than three times of the channel/network volume to ensure a complete coating.

The immersive LbL assembly route within the microchips is shown in the Fig. 2b. Different from the fluidic assembly, the immersive LbL assembly modifies the surfaces before bonding the fluidic layers. This route shares the same assembly mechanism as the fluidic assembly, as illustrated in Fig. 3. In this method, the PDMS substrate is immediately dipped into a PDDA solution for 15 minutes after the plasma treatment, forming the first PDDA layer. To shorten the assembly time, the coatings are dried using a heat lamp for one minute before rinsing with DI water. Other drying strategies may be adopted as needed. After that, the substrate is dipped into the clay dispersion for 15 minutes and then undergoes the drying and flushing processes to form one bilayer of the coating. Multilayer coatings, (P/C)n, are made by repeating the procedure for n times. The undesired coating area on the substrate (outside of the channel/pore area) is cleaned using tape before sealing the fluidic layers.

3 Results and discussion

The topmost layer of the created coatings is composed of rock-forming minerals. Thus, chemical mimetics of rock surfaces are established in the enclosed microchips to achieve SMMR systems. To characterize the properties of the SMMR, four steps are taken: 1) detect the morphological and structural properties of the mineral coatings using a scanning electron microscope (SEM), optical microscope and profilometer; 2) test the coating stability under flooding experiments; 3) characterize the mineral surface wettability by measuring the mineral oil–water contact angles; and 4) conduct two-phase fluid flow experiments in the built SMMR systems to observe the impact of the formed mineral coatings in the microchips. By performing the LbL assembly on glass substrates, we also demonstrate the applicability of the method for other materials.

The LbL assembly approach is used to coat the surfaces of the microchips, of which the rectangular channels have a width in the range of 10 to 250 μm and depth in the range of 40 to 100 μm. These ranges cover the most commonly used channel sizes. Fig. 4 depicts the cross-section view of a bare and coated simple channels and also an example of the designed flow networks. Fig. 4a and b represent the conventional bare and coated PDMS microchannels with a simple geometry (channel width: 250 μm and 150 μm, depth: 100 μm). Fig. 4c represents a coated complex flow network (channel width: 10 μm, depth: 40 μm), which is designed based on the Voronoi algorithm.79 Both of the chips in Fig. 4b and c have two-bilayers of montmorillonite. All the walls of the channels in the simple and the complex networks are covered with the coating material. Fig. 4d is a 3D view of the complex flow network (area: 5 mm × 5 mm). These microchips are applied in section 3.4 for flow experiments.


image file: c7lc00675f-f4.tif
Fig. 4 Scanning electron microscopy (SEM) photomicrographs of the cross-sections of a simple microchip channel/network and the three-dimensional (3D) microscope image of the flow network. (a) Cross-section of the bare PDMS microchip channel. (b) Cross-section of the montmorillonite-coated (two bilayers) microchip channel and the enlarged view of the coating. (c) Cross-section of the microchip flow network and an enlarged view of a coated channel. (d) 3D view of the network (image taken using a Nikon A1R laser confocal microscope system).

3.1 Morphology and structure of coated surfaces

The surface-modified PDMS slabs are used to characterize the morphological and structural properties of the mineral coatings. In Fig. 4b and c, microchips with channel scales from 10 μm to 250 μm are ideally coated by the minerals with two cycles of the LbL assembly action. Generally, mineral coatings with 2–3 bilayers present full-cover appearance on the PDMS substrates. However, larger layer counts like 5 to 6 produce more uniform surface morphologies. As the characterization operations are easier and more reliable with a more uniform surface, larger layer counts are preferred for the measurements. However, in this study, six-bilayers are the maximum applied to the on-chip modification. This limit is set to minimize the overall thickness of the coated layer and decrease the effect of coated layers on flow patterns. Thus, coatings with six bi-layers are observed by utilizing SEM and the measured surface profiles of the samples have a layer count from zero to six.

Fig. 5 shows the SEM images taken from PDMS slabs coated with six-bilayers of montmorillonite/kaolinite using the immersive LbL assembly. In Fig. 5a and b, montmorillonite particles of irregular shapes and varying sizes are adsorbed on the PDMS surface. The surface is demonstrated to be fully covered by the minerals when the image Gamma value is adjusted and clay textures are revealed in the dark spots of the SEM images.


image file: c7lc00675f-f5.tif
Fig. 5 SEM photomicrographs of the PDMS surfaces coated with six-bilayer montmorillonite/kaolinite by the immersive LbL assembly route. (a) and (b) Montmorillonite-coated PDMS surface (scale bars: 20 μm, and 2 μm, respectively). (c) and (d) Kaolinite-coated PDMS surface (scale bars: 20 μm and 2 μm).

Compared to the montmorillonite coating, the kaolinite coating has a tighter texture and reduced space amongst the clay particles (Fig. 5c and d). Different surface morphologies are present when different rock-forming minerals are employed as LbL assembly materials.

Two types of profilometer are used to measure the structural properties (coating thickness and surface profiles) of the LbL-based mineral films. The 3D surface profile of a PDMS slab with both coated and uncoated surfaces is measured by the Wyko NT1100 Optical Profiling system (Fig. 6a). The left side reveals a bare and smooth PDMS surface. The right side is a six-bilayer montmorillonite/PDDA coating on the PDMS where a rough morphology is observed. Fig. 6b shows the quantitative thickness measurement of the coatings with 0–6 layer counts (measured by using a P-6 Stylus Profiler). The thickness ranges from a few hundred nanometers to about one micrometer. Within the range of measurements (0–6 layer counts), a linear relationship between the coating thickness and the layer counts is observed for both montmorillonite and kaolinite coatings. Even though a slight reduction of channel size is observed after the LbL assembly modification, an amendment on the small channel sizes based on the thickness of coating promotes the accurate experimental flow analysis. The thinness and the easy-to-get uniformity (within six bilayers) of the coatings on PDMS surfaces assure the wide applications of the LbL assembly technique on microchips.


image file: c7lc00675f-f6.tif
Fig. 6 Surface profile measurements of the mineral coatings on PDMS. (a) 3D surface profile display of the six-bilayer montmorillonite/PDDA coating on PDMS (measured by using a Wyko NT1100 optical profiling system). The left side: bare PDMS surface; right side: rough mineral coating. (b) Thickness of the montmorillonite/kaolinite coatings with layer counts ranging from zero to six (measured by using a Tencor P-6 stylus profiler). Data = mean ± standard deviation.

3.2 Stability of coated surfaces

Stability of the established mimetic surfaces in the microchips is another critical factor in the development of SMMR. As SMMRs are used for subsurface fluid flow systems, the stability of the created surfaces is initially tested by exposing the surfaces to flooding of formation brine and reduced-salinity brines. The brines have varying salinities ranging from 15[thin space (1/6-em)]000 ppm to 0 ppm (DI water) and are prepared by mixing sodium chloride with DI water.

For the dynamic tests, the coated slabs are aged in formation brine (15[thin space (1/6-em)]000 ppm) for three days firstly. The aging step simulates the initial brine-saturated state of the rocks. Then, the aged surfaces are flooded with the synthesized formation brine (15[thin space (1/6-em)]000 ppm) and low-salinity brines (12[thin space (1/6-em)]000 ppm, 10[thin space (1/6-em)]000 ppm, 8000 ppm, 5000 ppm, 3000 ppm and 0 ppm) separately. Each flooding test is continued for six hours, with a high flow velocity of 60 m per day. The samples used for the stability testing are the same as the ones used for SEM observations in section 3.1. All flooding operations are conducted by using a self-assembled flow cell composed of a high-pressure pump and Teflon tubing.

The coatings of both clay minerals provide excellent dynamic stability in the flooding experiments. Fig. 7 depicts the SEM images of the surfaces coated with montmorillonite before and after the flooding tests. When compared to the original mineral coatings (Fig. 5a and b), no obvious morphological change is observed on the aged surfaces (Fig. 7a and b). The coatings remain intact after high-velocity (60 m per day) brine flooding (Fig. 7c and d) and DI water flooding (Fig. 7e and f) as well. The kaolinite coating images are not provided for the similarity.


image file: c7lc00675f-f7.tif
Fig. 7 Morphological observation and thickness measurements of the LbL modified PDMS surfaces (with six bilayers of montmorillonite), before and after dynamic flooding tests. (a) and (b) The mineral coating morphology after aging in NaCl brine (15[thin space (1/6-em)]000 ppm) for three days (scale bars: 20 μm and 2 μm, respectively). (c) and (d) Surface morphology after aging for three days and after brine-flooding for 6 h (flow velocity of 60 m per day; scale bars: 20 μm and 2 μm, respectively). (e) and (f) Surface morphology after aging for three days and after DI water-flooding for 6 h (flow velocity of 60 m per day; scale bars: 20 μm and 2 μm, respectively). (g) Thickness measurements of two LbL assembly multilayers (six-bilayer of montmorillonite and kaolinite on PDMS substrates, labelled PDMS (P/M)6 and PDMS (P/K)6, respectively) after different treatments: no treatment, aged (7 days) and aged (7 days) + flooded (1 h or 6 h). The flooding medium is a synthesized brine prepared according to ref. 80. Data = mean ± standard deviation.

The dynamic stability of the LbL coatings exposed to different ionic species is examined by conducting flooding tests with a synthesized brine containing four types of common ions in formations (Na+, K+, Ca2+, and Mg2+). The brine is prepared using an available recipe.80 The stability of the six-bilayer montmorillonite/kaolinite LbL coating is examined by measuring the thickness change after seven days of aging the surfaces in the synthesized brine and one to six hours of high-velocity (60 m per day) flooding over the coated surfaces. The measured thickness, displayed in Fig. 7g, confirms a negligible change to the coating structure after aging and/or flooding treatments. The values show a discrepancy within the experimental error limitation (SD = 0.0137 and 0.0322, respectively).

Hence, the developed SMMR systems can be fully utilized in many subsurface research problems that may include various ionic conditions and high flow rates.

3.3 Wetting properties of coated surfaces

The wetting properties of the coated surfaces are characterized by measuring the mineral oil–brine (15[thin space (1/6-em)]000 ppm of NaCl in DI water) contact angles (θob). PDMS slabs with different mineral coatings are immersed in bulk mineral oil. Then a brine droplet is placed on the surface. After reaching equilibrium, an optical contact angle goniometer (OCA 15EC, DataPhysics Instruments) automatically captures the image of the drop shape and calculates the contact angle. Twelve samples are prepared by submerging three types of surfaces (bare PDMS, PDMS with six bilayers of montmorillonite or kaolinite) into mineral oil and brine, respectively, and subjected to aging for varying periods (0 day, 3 days and 30 days). Three samples are prepared for repeating each of the measurement. Each measurement is repeated four times on the same sample at different locations.

The mean value and standard error of the collected data are found in Fig. 8. The bare PDMS surfaces without surface modification have hydrophobic properties with θob = 162° (mean value). The wetting properties remain constant after aging the PDMS surfaces in mineral oil for 30 days (θob = 163°). A slight decrease in the contact angle is observed after aging the samples in brine for three days (θob = 141°) and then 30 days (θob = 143°). There is a tendency for the hydrophobicity reduction with the aging in brine. Overall, PDMS provides a stable hydrophobic nature. While a slight alteration may exist when treating the PDMS with mineral oil or brine, it is not significant.


image file: c7lc00675f-f8.tif
Fig. 8 Contact angle measurements of the bare PDMS surface and the LbL assembled mineral surfaces on PDMS. The multilayers are six-bilayer of montmorillonite/kaolinite. Data = mean ± standard error.

The modified PDMS surfaces with montmorillonite and kaolinite have distinct wetting properties when compared to bare PDMS. In Fig. 8, the montmorillonite-coated PDMS substrate has similar wettability properties to the bare PDMS. Both are originally hydrophobic. This wetting state remains unchanged after aging in mineral oil for 30 days. After aging in brine for three days and 30 days, a remarkable drop in hydrophobicity was observed. The contact angles decreased from θob = 159° to θob = 101° and θob = 111°, respectively. Different from the bare PDMS and montmorillonite-coated PDMS surfaces, the kaolinite coating has hydrophilic properties all the time (θob = 77° for original surface). After aging in mineral oil for 30 days, a slight wettability change is observed. Moderate changes are detected when aging the samples in brine for different periods of time (e.g., contact angle changes from θob = 77° to θob = 49° after 30 days of aging). All the samples under aging are remained immersed before the measurements. They are naturally dried for the same time in ambient conditions for the measurement.

Factors like the distinct surface morphology, different microscale coating structures (roughness and penetrability) and the inherent chemical differences of the surfaces all have an impact on the different wettability displays and wettability changes of the bare PDMS and montmorillonite/kaolinite coatings.

3.4 Surface-mimetic micro-reservoir (SMMR)

The developed SMMR with the complex network shown in Fig. 4c and d has a porosity value of 0.15 ± 0.01. The permeability of the SMMR with the same network is measured by controlling the constant pressure gradient and measuring the flow rate in the network while applying Darcy's law. The FLUIGENT microfluidic flow control system (MFCS™-EZ) is coupled with a flow-rate platform (FRP) to measure the change of the flow rate by the pressure gradient. The measurements are conducted using DI water at 20 °C under three different inlet pressures (20, 30 and 50 kPa). Each measurement is repeated three times. The mean value of the measured permeability of the complex microchip with two kaolinite bilayers is 0.0773 Darcy (SD = 0.0033).

The effect of the mineral coatings on microflow behaviors is observed in the two-phase flow experiments in bare (Fig. 9a) and LbL modified microchips (Fig. 9b). The microchip is composed of one primary channel and side branches with widths ranging from 150 μm to 250 μm. The depth of the microchannel is 100 μm everywhere in the fluidic network (Fig. 4a and b). Two kaolinite bilayers are uniformly coated on the surfaces.


image file: c7lc00675f-f9.tif
Fig. 9 DI water (black colour in a and b and white colour in c–f) displacing mineral oil (red colour, dyed with a fluorescent red dye (Kingscote 506[thin space (1/6-em)]250-RF4)) in bare and LbL modified SMMRs. (a) and (b) show the displacement patterns in the bare and coated PDMS microchips with a simple channel geometry, respectively (scale bar: 100 μm). (c) and (e) show the displacement patterns within the bare PDMS microchip with a complex designed network at the time of breakthrough and 15 pore-volume (PV) injection, respectively. (d) and (f) show the displacement patterns within the LbL modified microchip (two bilayers of kaolinite) at the time of breakthrough and 15 PV injection, respectively (scale bar: 500 μm).

Fig. 9a and b, respectively, depict the flow patterns during the displacement of oil in the bare and coated microchannels. The red color in the images depicts mineral oil, which is dyed with a fluorescent red dye (Kingscote 506[thin space (1/6-em)]250-RF4). The mineral oil has a viscosity of 0.174 Pa s and density of 0.893 g ml−1. The DI water is indicated using a black color. In both instances, a long intruding water finger displacing the oil phase is formed. The uniform oil–water interface is observed in the bare PDMS chip (Fig. 9a). There are many tiny oil droplets remaining close to the PDMS surface, where the image is taken. The droplets vanish gradually with continued water flooding. In contrast, a non-uniform oil–water interface is formed in the near-mineral coating zone in the SMMR (Fig. 9b). Much of the residual oil is distributed on the mineral-coated surface and remain immobile for a relatively long period of the displacement.

The two-phase flow experiments are also conducted in the SMMR microchips with the complex pore network as shown in Fig. 4c and d. In these experiments, DI water is injected into the network at a flow rate of 0.01 ml h−1 displacing the oil. Fig. 9c and d reveal the flooding patterns in the uncoated and coated chips immediately following the breakthrough times. The patterns emerging after 15 pore-volume (PV) injection of DI water are shown in Fig. 9e and f. DI water is injected at the top right corner of the network and fluids are produced from the bottom left corner. FIJI software® is used to identify the colorless water phase distribution by subtracting the real-time displacement pattern from the original saturated network. The red color in Fig. 9c–f represents mineral oil and the white color is the water phase.

Fig. 9c and e depict the water–mineral oil displacement pattern in an uncoated microchip. At the time of breakthrough (Fig. 9c), a narrow preferential flow path connects the inlet and outlet of the network. The displacement efficiency defined as the ratio of the swept area to the total channel area is determined to be 34% (SD = 1% over three experiments). After 15 PV injections of DI water (Fig. 9e), the displacement efficiency increases to 59% (SD = 5%). There is almost no oil production from the network with more injections. Fig. 9d and f demonstrate two-phase flow patterns in a microchip coated with two bilayers of kaolinite clay. The injected water is observed to be widely spread across the coated channels, indicating the changes in the wetting properties of the surface. This modified chip has a higher displacement efficiency than the unmodified chip. At the time of breakthrough (Fig. 9d), the displacement efficiency is 48% (SD = 4%). The displacement efficiency increases to 64% (SD = 9%) after 15 PV injections of DI water (Fig. 9f). The differences in flow patterns may be attributed to the modifications applied to the channel surfaces, which alter the wetting characteristics and surface roughness properties. A comprehensive analysis is necessary to fully characterize multiphase fluid dynamics in SMMRs.

3.5 Application of the LbL assembly on other microfabrication materials

Subsurface porous systems normally encompass elevated pressures and temperatures. The coated PDMS chips can withstand a relatively high temperature (∼100 °C). However, PDMS material transformation occurs under very low pressures. In microfluidics technology, rigid materials like glass and quartz are generally employed for high-pressure and high-temperature testing conditions. Here, the feasibility of coating glass surfaces by the LbL assembly is considered.

Glass sheets are negatively-charged after an oxygen plasma treatment. This is followed by implementing self-assembly procedures used for the PDMS material in this study. Fig. 10 depicts the SEM images of glass surfaces coated with montmorillonite and kaolinite. Like PDMS, a full surface coverage by clay particles is produced on the glass substrate. The stability tests for the mineral coatings over the glass substrate also confirm that the coating has remained unchanged under the same testing conditions as used on the PDMS surfaces (see section 3.2).


image file: c7lc00675f-f10.tif
Fig. 10 SEM photomicrographs of the glass substrates coated with six-bilayer montmorillonite/kaolinite by the immersive LbL assembly route. (a) and (b) Montmorillonite coated glass surface (scale bars: 20 μm and 2 μm). (c) and (d) Kaolinite coated glass surface (scale bars: 20 μm and 2 μm).

The proposed LbL assembly is a universal technique that can be applied to modify the microchip surfaces of different materials for subsurface flow system. The solid surfaces of microchips made of glass, polymers, silicon and PMMA can be easily treated to achieve the required charging properties. This is done by exposing them to plasma or oxidizing chemicals (e.g., piranha solution). Thus, these materials can be subjected to the coating process described for the PDMS and glass substrates. It is noted that due to the different types and volumes of surface charges, differences exist in the thickness of multilayers formed by coating the same minerals on different substrates. Thus, similar stability, chemical characterization and multiphase flow experiments performed on PDMS substrates need to be implemented for other materials.

4 Conclusions

A surface-mimetic micro-reservoir (SMMR) was developed by combining the advanced microfluidics and layer-by-layer (LbL) assembly surface-modification technologies. Desired rock/soil surface properties were custom-built in microchips by coating rock-forming minerals on the microchannel walls.

For demonstration, the clay pores/microfractures in two common sedimentary rocks, sandstones and mudrocks, were reconstructed. As representatives, montmorillonite and kaolinite were used to coat the PDMS microchips with a wide range of channel sizes (width of 10–250 μm and depth of 40–100 μm). The microscale observation by a scanning electron microscope, an optical microscope and two types of profilometer demonstrated the formation of the nano- to micro-scale, fully-covered and rough mineral coating on the solid surfaces. The thickness of the mineral coating showed a linear relationship with the layer counts under the experimental conditions in this study. The coatings had excellent dynamic stability under high-velocity flooding tests (60 m per day) with complex ionic conditions (salinity ranging from 0 to 15[thin space (1/6-em)]000 ppm and ions of Na+, K+, Ca2+, and Mg2+). This stability validation demonstrated the capability to apply the developed SMMR to a lot of subsurface flow and transport research. The mineral-coated surfaces and the bare PDMS surface displayed distinct wetting properties: 1) the montmorillonite-coated surface and the bare PDMS surface had hydrophobic properties with or without aging treatment in mineral oil/brine; and 2) the kaolinite-coated surface remained hydrophilic. The impact of the coatings on two-phase flow dynamics in small and large microchannels was analyzed and compared to the results using uncoated microchips. As a representative for the rigid materials used for fabrication of micromodels, glass substrates were also coated with montmorillonite and kaolinite using LbL modification procedures.

It is important to note that the stability of coatings exposed to other chemicals used in enhanced oil recovery (EOR) studies must be tested for each specific agent. In our study, the LbL coatings were stable during initial tests of different chemicals, including solvents (crude oil diluted with decane, kerosene), surfactants (Span and Tween 20) and silica nanoparticles. The coating durability of other solutions such as acid and alkali solutions injected into the microchips should also be tested.

Further study is needed to optimize the assembly conditions, design various types of SMMR and explore the practical applications of SMMR on subsurface research. Even though not every rock surface property (e.g., surface texture) is studied or realized in the reconstructed mineral surfaces in the SMMR, the custom-built real-surface chemical compositions and the rough flow channels have demonstrated the tremendous potential to facilitate the application of microfluidics technology on subsurface porous systems. This is a validated robust tool to use in the exploration of subsurface energy and environmental questions, especially issues related to fluid–solid (rocks/soils) interactions.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

The authors thank the Natural Sciences and Engineering Research Council of Canada, the China Scholarship Council (CSC), the CMC Microsystems, the Canada Foundation for Innovation and the University of Calgary Global Research Initiative in Unconventional Hydrocarbon Resources-Beijing Site for the support to this work. We thank Ms. Parisa Bazazi for the contact angle measurements.

References

  1. B. Morin, Y. Liu, V. Alvarado and J. Oakey, Lab Chip, 2016, 16, 3074–3081 RSC .
  2. N. S. Gunda, B. Bera, N. K. Karadimitriou, S. K. Mitra and S. M. Hassanizadeh, Lab Chip, 2011, 11, 3785–3792 RSC .
  3. L. Zuo, C. Zhang, R. W. Falta and S. M. Benson, Adv. Water Resour., 2013, 53, 188–197 CrossRef CAS .
  4. I. C. Bourg, L. E. Beckingham and D. J. DePaolo, Environ. Sci. Technol., 2015, 49, 10265–10284 CrossRef CAS PubMed .
  5. Y. S. Oh, H. Y. Jo, J. H. Ryu and G. Y. Kim, J. Hazard. Mater., 2017, 324, 373–381 CrossRef CAS PubMed .
  6. D. Ciceri and A. Allanore, PLoS One, 2015, 10, e0139979 Search PubMed .
  7. C. E. Stanley, G. Grossmann, X. C. i Solvas and A. J. deMello, Lab Chip, 2016, 16, 228–241 RSC .
  8. V. A. Lifton, Lab Chip, 2016, 16, 1777–1796 RSC .
  9. M. H. Schneider and P. Tabeling, Am. J. Appl. Sci., 2011, 8, 927–932 CrossRef .
  10. K. Ma, R. Liontas, C. A. Conn, G. J. Hirasaki and S. L. Biswal, Soft Matter, 2012, 8, 10669–10675 RSC .
  11. D. Sinton, Lab Chip, 2014, 14, 3127–3134 RSC .
  12. M. L. Porter, J. Jimenez-Martinez, R. Martinez, Q. McCulloch, J. W. Carey and H. S. Viswanathan, Lab Chip, 2015, 15, 4044–4053 RSC .
  13. B. Bao, J. Riordon, F. Mostowfi and D. Sinton, Lab Chip, 2017, 17, 2740–2759 RSC .
  14. D. S. Park, J. Upadhyay, V. Singh, K. E. Thompson and D. E. Nikitopoulos, in ASME 2015 International Mechanical Engineering Congress and Exposition, Houston, Texas, 2015 Search PubMed .
  15. W. Yun, C. M. Ross, S. Roman and A. R. Kovscek, Lab Chip, 2017, 17, 1462–1474 RSC .
  16. W. Song, T. W. de Haas, H. Fadaei and D. Sinton, Lab Chip, 2014, 14, 4382–4390 RSC .
  17. S. A. Kelly, C. Torres-Verdin and M. T. Balhoff, Lab Chip, 2016, 16, 2829–2839 RSC .
  18. S. S. Datta, J. B. Dupin and D. A. Weitz, Phys. Fluids, 2014, 26, 062004 CrossRef .
  19. O. Vincent, A. Szenicer and A. D. Stroock, Soft Matter, 2016, 12, 6656–6661 RSC .
  20. K. Xu, T. Liang, P. Zhu, P. Qi, J. Lu, C. Huh and M. Balhoff, Lab Chip, 2017, 17, 640–646 RSC .
  21. C. Hu, J. E. Morris and R. L. Hartman, Lab Chip, 2014, 14, 2014–2022 RSC .
  22. S. Boggs, Principles of sedimentology and stratigraphy, Pearson, Boston, 5th edn, 2012 Search PubMed .
  23. S. Wang, Z. Tao, S. M. Persily and A. F. Clarens, Environ. Sci. Technol., 2013, 47, 11858–11865 CrossRef CAS PubMed .
  24. J. Xiong, K. Liu, X. J. Liu, L. X. Liang and Q. Zeng, RSC Adv., 2016, 6, 110808–110819 RSC .
  25. P. Charoensuppanimit, S. A. Mohammad and K. A. M. Gasem, Energy Fuels, 2016, 30, 2309–2319 CrossRef CAS .
  26. S. Y. Prakash and J. Yeom, Nanofluidics and microfluidics: systems and applications, William Andrew, Waltham, 1st edn, 2014 Search PubMed .
  27. S. Berg and H. Ott, Int. J. Greenhouse Gas Control, 2012, 11, 188–203 CrossRef CAS .
  28. F. Starecki, S. Morais, R. Chahal, C. Boussard-Plédel, B. Bureau, F. Palencia, C. Lecoutre, Y. Garrabos, S. Marre and V. Nazabal, Int. J. Greenhouse Gas Control, 2016, 55, 36–41 CrossRef CAS .
  29. R. Pini, S. C. M. Krevor and S. M. Benson, Adv. Water Resour., 2012, 38, 48–59 CrossRef CAS .
  30. B. X. Li, H. A. Tchelepi and S. M. Benson, Adv. Water Resour., 2013, 62, 488–498 CrossRef CAS .
  31. C. M. Oldenburg, S. W. Webb, K. Pruess and G. J. Moridis, Transp. Porous Media, 2004, 54, 323–334 CrossRef CAS .
  32. J. M. Nordbotten, B. Flemisch, S. E. Gasda, H. M. Nilsen, Y. Fan, G. E. Pickup, B. Wiese, M. A. Celia, H. K. Dahle, G. T. Eigestad and K. Pruess, Int. J. Greenhouse Gas Control, 2012, 9, 234–242 CrossRef CAS .
  33. O. A. Moultos, I. N. Tsimpanogiannis, A. Z. Panagiotopoulos and I. G. Economou, J. Chem. Thermodyn., 2016, 93, 424–429 CrossRef CAS .
  34. L. J. Zhang, Y. Kim, H. Jung, J. M. Wan and Y. S. Jun, Environ. Sci. Technol. Lett., 2016, 3, 92–97 CrossRef CAS .
  35. A. Z. Al-Yaseri, M. Lebedev, A. Barifcani and S. Iglauer, J. Chem. Thermodyn., 2016, 93, 416–423 CrossRef CAS .
  36. P. Zhu and K. D. Papadopoulos, Phys. Rev. E: Stat., Nonlinear, Soft Matter Phys., 2012, 86, 046313 CrossRef PubMed .
  37. O. Vincent, B. Marguet and A. D. Stroock, Langmuir, 2017, 33, 1655–1661 CrossRef CAS PubMed .
  38. S. Schlüter, S. Berg, M. Rücker, R. T. Armstrong, H.-J. Vogel, R. Hilfer and D. Wildenschild, Water Resour. Res., 2016, 52, 2194–2205 CrossRef .
  39. S. A. Kelly, Doctor of Philosophy Dissertation, The University of Texas at Austin, 2015 Search PubMed .
  40. H. Lee, S. G. Lee and P. S. Doyle, Lab Chip, 2015, 15, 3047–3055 RSC .
  41. B. Zhao, C. W. MacMinn and R. Juanes, Proc. Natl. Acad. Sci. U. S. A., 2016, 113, 10251–10256 CrossRef CAS PubMed .
  42. J. W. Grate, M. G. Warner, J. W. Pittman, K. J. Dehoff, T. W. Wietsma, C. Zhang and M. Oostrom, Water Resour. Res., 2013, 49, 4724–4729 CrossRef CAS .
  43. W. Song and A. R. Kovscek, Lab Chip, 2015, 15, 3314–3325 RSC .
  44. W. Song and A. R. Kovscek, J. Nat. Gas Sci. Eng., 2016, 34, 1276–1283 CrossRef CAS .
  45. W. B. Bartels, H. Mahani, S. Berg, R. Menezes, J. A. van der Hoeven and A. Fadili, SPE J., 2017, 22, 1362–1373 CrossRef .
  46. E. A. Al-Khdheeawi, S. Vialle, A. Barifcani, M. Sarmadivaleh and S. Iglauer, Int. J. Greenhouse Gas Control, 2017, 58, 142–158 CrossRef CAS .
  47. A. Fogden, Colloids Surf., A, 2012, 402, 13–23 CrossRef CAS .
  48. E. V. Lebedeva and A. Fogden, Environ. Sci. Technol., 2010, 44, 9470–9475 CrossRef CAS PubMed .
  49. E. V. Lebedeva and A. Fogden, Colloids Surf., A, 2011, 380, 280–291 CrossRef CAS .
  50. G. Decher, J. D. Hong and J. Schmitt, Thin Solid Films, 1992, 210–211, 831–835 CrossRef CAS .
  51. N. A. Kotov, T. Haraszti, L. Turi, G. Zavala, R. E. Geer, I. Dékány and J. H. Fendler, J. Am. Chem. Soc., 1997, 119, 6821–6832 CrossRef CAS .
  52. W. Sun, Z. Zhou and J.-F. Li, RSC Adv., 2016, 6, 489–494 RSC .
  53. Z. Tang, N. A. Kotov, S. Magonov and B. Ozturk, Nat. Mater., 2003, 2, 413–418 CrossRef CAS PubMed .
  54. J. Wu, N. Akhtar, R. Y. N. Gengler, T. T. M. Palstra and P. Rudolf, J. Mater. Chem. C, 2017, 5, 1782–1788 RSC .
  55. Y. Yan, D. Wang and P. Schaaf, Dalton Trans., 2014, 43, 8480–8485 RSC .
  56. Y. Lvov, S. Yamada and T. Kunitake, Thin Solid Films, 1997, 300, 107–112 CrossRef CAS .
  57. G. Mehta, M. J. Kiel, J. W. Lee, N. Kotov, J. J. Linderman and S. Takayama, Adv. Funct. Mater., 2007, 17, 2701–2709 CrossRef CAS .
  58. M. Keeney, X. Y. Jiang, M. Yamane, M. Lee, S. Goodman and F. Yang, J. Mater. Chem. B, 2015, 3, 8757–8770 RSC .
  59. L. Luo, S. H. Chung and A. Manthiram, J. Mater. Chem. A, 2016, 4, 16805–16811 CAS .
  60. Y. L. Wang, D. Wu, Y. Zhang, X. Ren, Y. G. Wang, H. M. Ma and Q. Wei, RSC Adv., 2015, 5, 56583–56589 RSC .
  61. K. J. McKenzie, F. Marken, M. Hyde and R. G. Compton, New J. Chem., 2002, 26, 625–629 RSC .
  62. S. Matosevic and B. M. Paegel, Nat. Chem., 2013, 5, 958–963 CrossRef CAS PubMed .
  63. A. Javey, S. Nam, R. S. Friedman, H. Yan and C. M. Lieber, Nano Lett., 2007, 7, 773–777 CrossRef CAS PubMed .
  64. H. Blatt, R. J. Tracy and B. E. Owens, Petrology : igneous, sedimentary, and metamorphic, W.H. Freeman, New York, 3rd edn, 2006 Search PubMed .
  65. J. W. Neasham, in SPE Annual Fall Technical Conference and Exhibition, Denver, Colorado, 1977 Search PubMed .
  66. M. Valášková and G. S. Martynková, Clay Minerals in Nature: Their Characterization, Modification and Application, InTech, 2012 Search PubMed .
  67. M. Das and N. Medhi, Int. J. Res. Eng. Appl. Sci., 2015, 5, 25–33 Search PubMed .
  68. A. M. Swift, L. M. Anovitz, J. M. Sheets, D. R. Cole, S. A. Welch and G. Rother, Environ. Geosci., 2014, 21, 39–57 CrossRef .
  69. R. M. Slatt and N. R. O'Brien, AAPG Bull., 2011, 95, 2017–2030 CrossRef CAS .
  70. V. Vertlib, M. Dietiker, M. Plötze, L. Yezek, R. Spolenak and A. M. Puzrin, J. Mater. Res., 2008, 23, 1026–1035 CrossRef CAS .
  71. K. Ariga, J. P. Hill and Q. Ji, Phys. Chem. Chem. Phys., 2007, 9, 2319–2340 RSC .
  72. J. J. Richardson, M. Björnmalm and F. Caruso, Science, 2015, 348, aaa2491 CrossRef PubMed .
  73. N. Raman, M. R. Lee, S. P. Palecek and D. M. Lynn, J. Controlled Release, 2014, 191, 54–62 CrossRef CAS PubMed .
  74. Y. Wang, Y. Liu, Y. Cheng, E. Kim, G. W. Rubloff, W. E. Bentley and G. F. Payne, Adv. Mater., 2011, 23, 5817–5821 CrossRef CAS PubMed .
  75. A. J. Mateos, A. A. Cain and J. C. Grunlan, Ind. Eng. Chem. Res., 2014, 53, 6409–6416 CrossRef CAS .
  76. E. Kharlampieva, V. Kozlovskaya, J. Chan, J. F. Ankner and V. V. Tsukruk, Langmuir, 2009, 25, 14017–14024 CrossRef CAS PubMed .
  77. J. Seo, J. L. Lutkenhaus, J. Kim, P. T. Hammond and K. Char, Langmuir, 2008, 24, 7995–8000 CrossRef CAS PubMed .
  78. B. D. Tiu, D. L. Kernan, S. B. Tiu, A. M. Wen, Y. Zheng, J. K. Pokorski, R. C. Advincula and N. F. Steinmetz, Nanoscale, 2017, 9, 1580–1590 RSC .
  79. M. Wu, F. Xiao, R. M. Johnson-Paben, S. T. Retterer, X. Yin and K. B. Neeves, Lab Chip, 2012, 12, 253–261 RSC .
  80. H. Mahani, S. Berg, D. Ilic, W. B. Bartels and V. Joekar-Niasar, SPE J., 2015, 20, 8–20 CrossRef CAS .

This journal is © The Royal Society of Chemistry 2018