Liquid Crystal-Infused Porous Polymer Surfaces: A ‘Slippery’ Soft Materials Platform for the Naked-Eye Detection and Discrimination of Amphiphilic Species

Abstract

We report the design and characterization of liquid crystal (LC)-infused porous polymer membranes that can detect and report on the presence of natural and synthetic amphiphiles in aqueous solution. We demonstrate that thermotropic LCs can be infused into nanoporous polymer membranes to yield LC-infused surfaces that exhibit slippery behaviors in contact with a range of aqueous fluids. In contrast to conventional liquid-infused surfaces (SLIPS or LIS) prepared using isotropic oils, aqueous solutions slide over the surfaces of these LC-infused materials at speeds that depend strongly upon the composition of the fluid, including the presence, concentration, or structure of a dissolved surfactant. In general, the sliding times of aqueous droplets on these LC-infused surfaces increase significantly (e.g., from times on the order of seconds to times on the order of minutes) with increasing amphiphile concentration, allowing sliding times to be used to estimate the concentration of the amphiphile. Additional experiments revealed other intrinsic and extrinsic variables or parameters that can be used to further manipulate droplet sliding times and discriminate among amphiphiles of similar structure. Our results are consistent with a physical picture that involves reversible changes in the interfacial orientation of the anisotropic LCs mediated by the interfacial adsorption of amphiphiles. These materials thus permit facile ‘naked-eye’ detection and discrimination of contaminating amphiphiles in aqueous samples using equipment no more sophisticated than a stopwatch. We demonstrate the potential utility of these LC-infused surfaces for the unaided, naked-eye detection and monitoring of amphiphilic bio-toxins in small droplets of fluid extracted directly from cultures of two common bacterial pathogens (P. aeruginosa and S. aureus). The ability to translate molecular interactions at aqueous/LC interfaces into large and readily-observed changes in the sliding times of small aqueous droplets on surfaces could open the door to new applications for anti-fouling, liquid-infused materials in the context of environmental sensing and other fundamental and applied areas.

Graphical Abstract

Introduction

Slippery liquid-infused porous surfaces (SLIPS) and liquid-infused surfaces (LIS) comprise a relatively new class of synthetic soft materials fabricated by the infusion of lubricating oily liquids into chemically compatible nanoporous, microporous, or topographically patterned surfaces.^1-9 Provided that the chemical properties of the lubricant and the underlying surfaces are suitably matched, these materials present a ‘slippery’ layer of mobile fluid at the surface that can repel other immiscible fluids or substances with which they come in contact.^{1, 2, 5, 7-9} For example, SLIPS and LIS can shed droplets of aqueous solutions at very low sliding angles (e.g., < 5°), endowing these materials with robust anti-icing,^{10, 11} self-cleaning,^{2, 12} and other anti-fouling properties.^{1, 7-9, 13, 14} Depending on the nature of the infused oil, these materials can also prevent fouling by other complex fluids, including commercial and industrial liquids and gels,^{1, 7-9, 15} and bio-fouling by microorganisms.^{4, 7, 13, 16, 17} Attention to practical issues such as long-term stability will further expand the commercial utility of these materials. Integration of new design principles that impart new functions and behaviors to SLIPS and LIS while maintaining slippery character could also open the door to exciting and entirely new applications of these antifouling materials.

Aizenberg and co-workers reported the first examples of SLIPS in 2011 by infusing perfluorinated liquids into nanoporous polytetrafluoroethylene (PTFE) membranes.¹ Since that initial report, many groups have expanded on the range of lubricating liquids and underlying porous matrices that can be used to fabricate SLIPS, improve their chemical and physical stabilities in complex environments, and design multifunctional coatings with improved anti-fouling behaviors.^{8, 12, 17-25} It is now broadly recognized that the properties of the infused oil can have substantial impacts on both the stability of the mobile liquid layer (e.g., the degree to which the infused oil can be displaced by a contacting fluid)^{9, 26, 27} and the mobility of droplets of aqueous fluid (e.g., droplets of water slide more slowly on SLIPS fabricated using higher viscosity oils, and more rapidly on coatings infused with lower viscosity liquids).^{5, 25, 28}

These latter observations have motivated recent work to explore the infusion of ‘functional’ oils with physicochemical properties that can be manipulated actively and dynamically.^{19, 20, 22, 25, 29, 30} We communicated recently that hydrophobic nanoporous polymer coatings fabricated by covalent layer-by-layer (LbL) assembly could be infused with a thermotropic liquid crystal (LC), an anisotropic and birefringent fluid, to design SLIPS that respond to chemical changes in their environments.²⁰ For example, SLIPS fabricated using a model LC in the nematic state exhibited slippery properties that changed substantially and reversibly when a contacting liquid contained (or did not contain) an amphiphilic molecule, such as a surfactant. That past work demonstrated that LC-infused SLIPS can discriminate actively based on the chemical composition of a contacting fluid, suggesting new approaches for the manipulation of droplet mobility and providing a potential basis for the design of surfaces that permit the ‘naked-eye’ detection of environmental analytes (e.g., by reducing rates of droplet migration to levels that can be discriminated amongst by eye, or with the aid of a simple stopwatch).²⁰ More recently, Wang et al. reported the design of SLIPS fabricated using ferrofluids to yield so-called ‘FLIPS’ that can be transformed using magnetic fields to present either smooth or multiscale hierarchical surface features, providing surfaces that permit active control over the self-assembly of colloidal particles at the micrometer scale or the dislodging of bacterial biofilms at centimeter length scales.²⁹ This past work, combined with a steadily growing body of research on the functionalization and/or patterning of new rough or porous surfaces that can be used to host stable films of infused oils, has significantly expanded the range of potential applications of ‘slippery’ surfaces.^{25, 31-33}

The work reported here was motivated by our past observations of responsive behaviors in slippery LC-infused layer-by-layer coatings and the potential of these materials to enable dynamic control over the mobility of immiscible fluid droplets.²⁰ This current study sought to (i) expand upon those key findings and explore the generality of this approach to the design of environmentally-responsive SLIPS, (ii) provide insight into key chemical and physical factors that govern the dynamic behaviors of these LC-infused materials and their responses to fluids of varying composition, and (iii) explore the potential of LC-infused SLIPS to enable the development of new soft material platforms for the detection of environmental agents or the discovery of new chemical and biological agents.

In this study, we demonstrate that thermotropic LCs can be infused into nanoporous PTFE thin films to yield LC-infused membranes that exhibit slippery behaviors and remain physically and functionally stable when contacted with a broad range of synthetic and biological aqueous fluids. We also show that droplets of aqueous fluids slide over the surfaces of these LC-infused materials at speeds that depend upon the composition of the fluid (e.g., the ionic strength of the fluid or the presence, concentration, and structure of natural and synthetic amphiphiles contained within it). In general, sliding times on these LC-infused SLIPS increase significantly with increasing amphiphile concentration in the droplet, consistent with a physical picture involving the adsorption of amphiphiles at aqueous/LC interfaces and permitting measurement of differences in sliding time to be used to both identify the presence and estimate the concentration of amphiphiles in a solution. Finally, we demonstrate that these materials can be used to report the presence of amphiphilic toxins in aqueous samples containing either Gram-negative or Gram-positive bacteria, providing a conceptually straightforward and practical approach to the naked-eye identification of bacterial contamination. The ability of these materials to translate molecular interactions at interfaces created between aqueous solutions and thin films of LCs into large and readily-observed changes in the sliding times of small aqueous droplets has significant implications for the application of liquid-infused materials in the context of environmental sensing and other fundamental and applied areas.

Materials and Methods

Materials.

Sodium dodecyl sulfate (SDS, ACS grade, ≥99.0%), dodecyltrimethylammonium bromide (DTAB, ≥98.0%), hexadecyltrimethylammonium bromide (HTAB, ≥98.0%), silicone oil (η = 50 cSt), Brij 30 (C₁₂E₄), sodium chloride (NaCl, ACS grade, ≥99.0%) and 3-oxo-C12-AHL were obtained from Millipore Sigma (Milwaukee, WI). The thermotropic liquid crystals 5CB and E7 were purchased from Jiangsu Hecheng Display Technology Co. (Jiangsu, China). Phosphate-buffered saline (PBS) (137 mM NaCl, 2.7 mM KCl, 10 mM phosphate; pH 7.4) was prepared from OmniPur® 10× concentrate (Millipore Sigma, Milwaukee, WI). Unlaminated PTFE membrane filters (pore size = 0.2 μm, thickness = 25-51 μm) were purchased from Sterlitech Corporation (Kent, WA). Eutrophic lake water was locally sourced from Lake Mendota, Madison, WI. Nature’s Touch skim milk was purchased from Kwik Trip (Madison, WI). Pooled human urine was purchased from Innovative Research, Inc. (Novi, MI). Luria-Bertani medium (LB), Lennox formulation, was purchased from Research Products International (Mt. Prospect, IL). Brain heart infusion (BHI) medium was purchased from Teknova (Hollister, CA). N-Butanoyl-L-homoserine lactone (C4-AHL) was purchased from Cayman Chemical (Ann Arbor, MI). 3-oxo-C12-HS was synthesized by base-catalyzed hydrolysis of 3-oxo-C12-AHL in analogy to previously reported procedures.³⁴ Dodecyl-1,12-bis(trimethylammonium bromide) (DBTAB) was a kind gift from Prof. Nicholas L. Abbott (Cornell University, Ithaca, NY). The phenol-soluble modulin PSM-α-3 was a kind gift from Prof. Samuel H. Gellman (UW-Madison, Madison, WI). Rhamnolipids (90% pure) were obtained from AGAE technologies (Corvallis, OR). 3-(3-Hydroxyalkanoyloxy)alkanoic acid (HAA) and AIP-III D4A were synthesized according to previously reported methods.^{35, 36} The use of the term ‘water’ in all sections below refers to water with a resistivity of 18.2 MΩ, obtained from a Millipore filtration system, unless otherwise noted. All materials were used as received without further purification unless otherwise noted.

General Considerations.

Scanning electron micrographs were acquired using a LEO 1550 scanning electron microscope at an accelerating voltage of 3 kV. Samples were coated with a thin layer of gold using a gold sputterer operating at 10 mA under a vacuum pressure of 50 mTorr for 1 min prior to imaging. Digital photographs and videos were acquired using a Samsung Galaxy S7 smartphone. Sliding time data were analyzed using Microsoft Excel and plotted using GraphPad Prism 7 (version 7.0h). For measurements of absorbance at 600 nm (OD₆₀₀) to monitor bacterial growth, 200 μL of culture were added to a clear-bottomed 96-well plate (Corning 3370) and absorbance was measured using a Synergy 2 plate reader (Biotek) with Gen5 1.05 software. All experiments performed using bacterial cultures, including measurements of sliding times on LC-infused SLIPS, were performed in a biological safety cabinet.

Preparation of SLIPS.

SLIPS were prepared at ambient room temperature (~20 °C) by depositing a lubricating liquid (e.g., either the LCs 5CB or E7, or silicone oil) on the top surface of a porous polymer membrane (supported on a glass slide) using a pipette. The lubricating liquid was then spread using tweezers to form a uniform over-coated layer. Samples were allowed to stand for several minutes to allow the liquid to infuse into the porous membrane (evident by a visual change in the opacity of the membrane) through capillary wicking. The excess liquid was then removed from the surface by dabbing with weighing paper.

Characterization of Droplet Sliding Times.

Characterization of the sliding times of droplets placed on the surfaces of LC-infused SLIPS was performed in the following general manner. LC-infused SLIPS were placed on a custom-made stage, and the stage was attached to the moving arm of a digital protractor using binding clips. The digital protractor was set at a specified sample angle, and a pre-determined volume of aqueous solution was placed as a droplet on the surface of the liquid-infused surface. Sliding droplets were recorded on digital video, and the time required for droplets to slide 4.0 cm along the surface was measured using a digital timer. In some cases, aqueous solutions were prepared by adding food coloring to enhance visual contrast of the sliding droplets. For characterization of the sliding times of droplets of bacterial cultures, three biological replicates were performed. After each measurement, the surface was washed by depositing multiple water droplets and allowing them to slide down the surface until the sliding time of the water droplets returned to a value of ~3 s. For each surfactant solution, the sliding times of at least 3-5 droplets were measured and used to calculate an average sliding time with standard deviation. Each experimental series was performed on one common LC-infused slippery surface, with appropriate experimental controls, in order to prevent variability in sliding time measurements between different LC-infused surfaces.

Bacteria and Culture Conditions.

All bacteria were grown at 37 °C, with shaking at 200 rpm (see Table S1 for strain and plasmids information). Staphylococcus aureus was cultured in BHI medium, and Pseudomonas aeruginosa was cultured in LB medium. For all bacterial strains, an overnight culture was grown in a 15 mL glass tube (no more than 2 mL of culture) or a 25 mL Erlenmeyer flask (no more than 5-10 mL of culture) to allow for sufficient aeration. Experiments using S. aureus proceeded as follows: An overnight culture of bacteria (strain 6390 or 9222) was diluted 1:100 in BHI medium and shaken for 24 h. The peptide AIP-III D4A (if applicable) was added when the overnight culture was diluted to achieve a final concentration of 1 μM. DMSO was added as a vehicle control (no greater than 2% final concentration) to cultures not containing added peptide. Experiments using P. aeruginosa proceeded as follows: An overnight culture of bacteria was diluted 1:100 in fresh LB medium and shaken for 6, 12, or 24 h, as specified. To induce RhlR-controlled phenotypes in PAO-SC4 (ΔlasI rhlI), C4-AHL was added when the overnight culture was diluted to achieve a final concentration of 200 μM. DMSO (no greater than 2%) was added to cultures as a vehicle control for experiments not containing added C4-AHL. For all experiments in which aliquots of bacterial culture were removed for characterization on LC-infused SLIPS, OD₆₀₀ values were measured and found to be similar for all strains at each time point and were consistent with those reported in past studies.³⁶

Results and Discussion

LC-Infused PTFE Membranes Influence the Sliding of Aqueous Droplets Containing Surfactant

Our past work demonstrated that the infusion of a thermotropic LC into hydrophobic and nanoporous polymer coatings fabricated by reactive/covalent layer-by-layer assembly can be used to design SLIPS that respond actively to changes in the chemical composition of the contacting liquid (e.g., the presence or absence of surfactants).²⁰ As a step toward investigating the broader utility of this approach and addressing practical challenges associated with the use of layer-by-layer coatings, we sought to characterize the infusion of LCs into commercially available and single-component nanoporous PTFE membranes that have been used as matrices for the infusion of conventional isotropic oils to design SLIPS in other past studies.^{1, 4, 6, 28}

We first performed a series of experiments to determine whether PTFE membranes could be infused with thermotropic LCs in the nematic state, and whether the resulting LC-infused membranes were ‘slippery’ and chemically or physically stable upon contact with a broad range of liquids. The infusion of thermotropic LCs into porous PTFE membranes with pore sizes of 200 nm and thicknesses of ~ 25-50 μm resulted in LC-infused SLIPS that allowed aqueous droplets to slide readily on the surface (the LC used in these experiments was E7 unless otherwise noted; the chemical structures of the mesogens that comprise this LC are shown in Figure S1; SEM images showing top-down views of these membranes are shown in Figure S2; see Materials & Methods for additional details). Figure 1A shows top-down views of a 50 μL droplet of PBS (colored green using food coloring to enhance visual contrast) placed on an LC-infused SLIPS tilted at 20°; the droplet was observed to slide over a length of 4 cm in ~4 s (similar sliding behavior was also observed for water droplets placed on LC-infused SLIPS). The high degree of mobility of water droplets on these LC-infused membranes is consistent with the presence of a ‘slippery’ LC lubricant phase on the porous PTFE membrane that remains stable in the presence of water droplets (S_os(w) ≥ 0; see Table S2). The LC-infused SLIPS were also stable when contacted with a broad range of chemically complex liquids, including acidic (pH 1) and alkaline (pH 11) aqueous solutions, skim milk, unfiltered lake water, and human urine. As shown in Figure S3, 50 μL droplets of these different liquids placed on LC-infused SLIPS tilted at 20° were also observed to slide over a length of 4 cm in ~4 s, similar to the behaviors of water droplets.

Figure 1.

(A,B) Images showing ‘top-down’ views at different time points of (A) a droplet of PBS (50 μL, colored green) and (B) a PBS droplet containing 100 μM SDS (50 μL, colored orange) sliding down E7-infused SLIPS tilted at 20°. (C) Plot showing sliding behaviors of droplets of PBS or PBS droplets containing SDS (100 μM) on PTFE membranes infused with silicone oil (white), thermotropic liquid crystals E7 (black) and 5CB (gray) at 23 °C, the gold bar and light blue bars show the sliding behaviors of PBS droplets containing SDS on E7-infused SLIPS equilibrated to a temperature of 70 °C and 5CB-infused SLIPS equilibrated to a temperature of 37 °C, respectively; results are expressed as the time required for a 50 μL droplet to slide 4 cm on the surface tilted at 20°. (D,E) Schematic showing proposed changes in the anchoring of LCs (yellow) from (D) planar to (E) homeotropic upon adsorption of surfactant (black) to the aqueous-LC interface formed between a surfactant-containing aqueous droplet (blue) sliding down an inclined LC-infused SLIPS (yellow). (E) Plot showing the sliding time of 50 μL DBTAB-containing water droplets (0.005–100 mM) on E7-infused SLIPS tilted at 20°.

We previously reported that water droplets containing surfactants slide more slowly on LC-infused SLIPS fabricated by the infusion of E7 into nanoporous PEI/PVDMA multilayer films as compared to droplets of water alone.²⁰ We observed similar differences in the sliding behaviors of droplets with and without surfactants on LC-infused SLIPS fabricated by infusion of E7 into the PTFE membranes investigated here. SDS-containing droplets were observed to slide very slowly compared to droplets of PBS (see Video S1 and S2 in the Supporting Information). For example, as shown in Figure 1B, a droplet of PBS containing 100 μM SDS (colored orange) slid over a length of 4 cm in ~63 s, compared to the ~4 s time required for a PBS droplet that did not contain surfactant (Figure 1A). We also measured the sliding times of SDS-containing droplets on E7-infused SLIPS maintained at 70 °C, a temperature well above the nematic/isotropic transition temperature of E7 (~60 °C). As shown in Figure 1C (gold bar) the SDS-containing droplets slid over a length of 4 cm in ~8 s, a time that is significantly faster than that observed on surfaces infused with E7 in the nematic state at ambient room temperature (~63 s). Additional experiments using SLIPS fabricated by the infusion of the thermotropic liquid crystal 5CB (in the nematic state) instead of E7 into PTFE membranes revealed similar results (the structure of 5CB is shown in Figure S1; 5CB-infused SLIPS maintained a ‘slippery’ lubricant phase in the presence of water droplets (S_os(w) ≥ 0); see Table S2). As shown in Figure 1C, a 100 μM SDS-containing droplet slid ~15 times slower (~63 s) on 5CB-infused surfaces compared to PBS droplets (~4 s). However, for 5CB-infused SLIPS maintained at 37 °C, a temperature above the nematic/isotropic transition temperature (~35 °C) of this LC, SDS-containing droplets again slid appreciably faster (blue bar; ~8 s) (Figure 1C). Lastly, as also shown in Figure 1C, we note that no differences in sliding time were observed between SDS-containing droplets and droplets of PBS on PTFE membranes infused with silicone oil, a model isotropic oil (both types of droplets slide over a length of 4 cm in ~4 s).

The results of these experiments demonstrate that the novel responsive sliding behaviors observed in past studies using layer-by-layer coatings are preserved when more well characterized and single-component commercial porous PTFE membranes are used as host substrates. When combined, these results support the hypothesis that changes in droplet sliding speeds is the result of the anisotropic nature of the infused LC and its behavior at interfaces created with aqueous media (that is, the large changes in droplet sliding times observed here occur in ways that are independent of the nature of the underlying substrate). On the basis of these results, we conclude that the infusion of LCs could likely also be used more generally to impart responsive behaviors to SLIPS fabricated using a variety of other well-known hydrophobic matrices used to fabricate slippery surfaces.

Influence of Amphiphile Structure and Other Parameters on Droplet Sliding Speeds

We speculate that the slower sliding speeds of surfactant-containing droplets shown in Figure 1B result from dynamic changes in the anchoring of the LCs (see schematic in Figure 1D,E) as aqueous/LC interfaces are formed beneath a droplet and surfactant adsorbs there. It is well understood that thermotropic LCs such as E7 and 5CB adopt so-called homeotropic anchoring when hosted at LC-air interfaces (i.e., the mesogens are generally aligned perpendicular to the interface), and that they adopt so-called planar anchoring when hosted at interfaces created between LCs and water (i.e., the mesogens are generally aligned parallel to the interface). In addition, previous studies have reported that adsorption of surfactants such as SDS at aqueous/LC interfaces can result in an orientational transition in the anchoring of LCs from planar to homeotropic orientations at the interface.^37-40 In the experiments reported above, a water droplet placed on LC-infused SLIPS results in the formation of an aqueous/LC interface under the droplet, and the LCs near that interface would be expected to exhibit planar anchoring. A water droplet containing a surfactant also results in the formation of an aqueous/LC interface, however the surfactant molecules in the droplet should also adsorb at the aqueous/LC interface and, thereby, promote homeotropic anchoring in the underlying LC. We note that we were unable to characterize the orientation of the LCs in the experiments described above using polarized light microscopy, a method commonly used to characterize the orientation of LCs at LC/aqueous interfaces,^{37, 41} because of the complexities of the system used here, including the opacity and thickness of the PTFE membranes. However, the results reported above, when combined with those of our past studies in LC-infused layer-by-layer coatings, and the results of additional experiments reported below involving surfactants with different tail lengths and head groups, are consistent with this general hypothesis.

We anticipate that any potential changes from planar to homeotropic anchoring that occur at aqueous/LC interfaces created by contact with aqueous droplets containing surfactant would occur and form continuously at that interface as the droplet slides along the surface. We did not observe changes in the speeds of droplets as they slid along LC-infused surfaces, providing general support for this hypothesis. It is, of course, possible that the concentration of surfactant in an aqueous droplet could become depleted if some of it remains bound at air/LC interfaces created in areas behind a sliding droplet (that is, sliding droplets could leave behind ‘trails’ of adsorbed surfactants as they move across a surface, which would result in a concomitant reduction in surfactant concentration in the droplet). We did not measure changes in surfactant concentration in the droplets in the studies performed here, and if surfactant depletion does occur, it did not occur to extents that resulted in significant changes in droplet sliding speed at the surfactant concentrations and path lengths evaluated in the experiments above. We do note, however, that PBS droplets placed on surfaces previously exposed to sliding SDS-containing droplets were observed to slide over a distance of 4 cm over ~7 s, a time that is slower than the sliding times of PBS droplets on fresh LC-infused PTFE membranes that were never exposed to surfactant-containing droplets (~4 s, as described above). This difference in sliding times is generally consistent with the view that surfactant from sliding droplets could remain at LC/air interfaces after surfactant-laden droplets have moved along the surface. We note further, in this context, that the sliding times of PBS droplets on ‘previously used’ LC-infused SLIPS returned to values of ~4 s and were otherwise indistinguishable from freshly-prepared surfaces after ‘rinsing’ them with 3-5 additional water droplets. This result suggests that the impact of adsorption of surfactant, to whatever extent it may occur, is reversible. In general, we found it possible to use, rinse, and reuse these LC-infused SLIPS multiple times with no observable changes in subsequent droplet sliding behaviors, with the one exception of cases in which substantially high surfactant concentrations (e.g., above CMC) were used. Under those conditions, the sliding droplets appeared to remove LCs from the membranes, resulting in an erosion of membrane performance. However, in most cases, this damage could be reversed, and the performance of the membranes could be restored, by the addition of more LC to the surface of the membrane.

To investigate further the role that homeotropic anchoring of LC may play in influencing droplet sliding speeds, we also evaluated the sliding speed of aqueous droplets containing the cationic bolaform surfactant dodecyl-1,12-bis(trimethylammonium bromide) (DBTAB; structure shown in Figure 1F). DBTAB adopts a looped configuration at oil/water interfaces and has a much higher limiting surface area (~107 Ų at an air-water interface) compared to the limiting surface area of analogous classical surfactants (e.g., 55-63 Ų for DTAB) that adopt tilted configurations at air/water interfaces.^{42, 43} Past studies report that DBTAB promotes planar, rather than homeotropic, anchoring of 5CB at aqueous/LC interfaces at concentrations ranging from 0.01 mM to 100 mM.³⁹ The contact angle (77° ± 1°) and droplet base diameter (4.0 ± 0.04 mm) of 50 μL 100 mM DBTAB-containing aqueous droplets on LC-infused SLIPS is similar to the contact angle (74.4° ± 0.7°) and droplet base diameter (4.14 ± 0.02 mm) of 50 μL 100 μM SDS-containing PBS droplets. However, we observed aqueous droplets containing between 5 μM to 100 mM DBTAB to slide over a distance of 4 cm in ~3 s (Figure 1F), a sliding time comparable to those of droplets of water alone, and a time that is substantially faster than those of droplets containing SDS (~63 s). We also note that the concentrations evaluated here encompass the CMC of DBTAB (20-50 mM) and are above the observed onset of surface activity of DBTAB at the air-water interface (< 1 mM).^{39, 42} These results thus provide further support for the view that the large differences in sliding speeds observed for droplets containing single-tailed surfactants such as SDS (above) and DTAB (as described below) result from dynamic and surfactant-induced changes in the orientation of the LC from planar to homeotropic in regions of the SLIPS that are in contact with the droplets.

The results of additional experiments characterizing the influence of surfactant concentration, salt concentration, and surfactant structure on the sliding speeds of surfactant-containing droplets on LC-infused SLIPS (Figure 2) were also consistent with the proposition that orientational changes in the anchoring of LCs can influence droplet sliding behaviors. Past studies have established that the anchoring of LCs at aqueous/LC interfaces is influenced strongly by the areal density of the surfactant molecules adsorbed at the interface and interactions between the surfactant tails and LCs.^{37, 39, 40, 44-47} For example, it has been demonstrated using aqueous-5CB interfaces that, with increasing surfactant concentration, the limiting areal density of surfactant tails at the interface increases and results in homeotropic alignment of the LC.^{37, 39, 44, 45} Similar results were obtained by increasing the electrolyte concentration, which can screen electrostatic repulsion between charged surfactant head groups.^{37, 39, 45} Finally, we mention that it has also been reported that both the nature of the hydrophilic head group and the aliphatic chain length of the surfactant can impact the orientation of LCs at aqueous-5CB interfaces.^{39, 47}

Figure 2.

(A) Plot showing the sliding time of 50 μL PBS droplets containing SDS (10–100 μM) on silicone oil-infused SLIPS (gray) and E7-infused SLIPS (black) tilted at 20°. (B) Plot showing the influence of NaCl concentration (0 to 100 mM) on the sliding time of SDS-containing droplets (0 μM (white), 10 μM (gray), and 100 μM (black)). (C) Table showing bulk concentration of various surfactants (with different head and tail group structure) in water droplets that slid off a E7-infused SLIPS tilted at an angle of 20° at average sliding times between 58 s – 62 s (~ 1 min) (see text for additional explanation).

As shown in Figure 2A (black bars), we observed that droplet sliding times increase substantially as the concentration of SDS in droplets of PBS was increased from 0 μM to 100 μM. Droplets of PBS containing 10 μM SDS slid rapidly (over ~5 s) over a distance of 4 cm on LC-infused SLIPS, whereas droplets containing 100 μM SDS exhibited sliding times of ~63 s. The sliding times of SDS-containing droplets exhibited uniform sliding times of ~3 s on PTFE membranes infused with silicone oil regardless of SDS concentration (Figure 2A; gray bars).

Experiments using droplets containing a fixed concentration of SDS with different concentrations of electrolyte (NaCl) revealed that manipulation of electrolyte concentration also impacts the sliding time of surfactant-containing droplets on LC-infused SLIPS. As shown in Figure 2B, for 100 μM SDS solutions in water (black bars), the addition of NaCl at 100 mM increases the sliding time by 15 times (~62 s) compared to NaCl at 0.5 mM. Varying the NaCl concentration of water droplets free of SDS over this same concentration range did not impact droplet sliding times (Figure 2; white bars). The sliding times of SDS-containing droplets were also insensitive to NaCl concentration at lower concentrations of SDS (e.g., 10 μM; Figure 2; gray bars).

We next investigated the influence of surfactant head and tail group structure on droplet sliding speeds using four different surfactants: SDS, DTAB, HTAB, and the non-ionic surfactant C₁₂E₄ (structures shown in Figure 2C). For these experiments, we prepared surfactant solutions in water to decouple the impact of surfactant structure from that of salt concentration. We measured the sliding times of droplets containing many different concentrations of these four surfactants and found sliding times to vary considerably as a function of concentration and structure. For simplicity and to permit general comparisons, Figure 2C reports the concentrations of each surfactant that resulted in average sliding times between 58–62 s (approximately 1 min). For surfactants with different head groups [SDS (anionic), DTAB (cationic), and C₁₂E₄ (non-ionic)] but identical aliphatic tail lengths (12 carbons), the measured concentrations were 1 mM, 3 mM, and 0.1 mM respectively. These values correlate with the concentration regime in which these surfactants adsorb strongly at both oil-water and air-water interfaces.^{39, 42, 48, 49}

We also tested the impact of changes in the aliphatic tail length (while keeping the head group constant) on the sliding time of a surfactant-containing droplet on LC-infused SLIPS. Figure 2C shows a comparison of results for DTAB and HTAB, which possess different aliphatic tail lengths (12 vs. 16 carbons, respectively) but identical cationic head groups. The concentration of HTAB (100 μM) required to achieve an average sliding time of ~60 s was found to be 30 times lower than that for DTAB (3 mM). This result is consistent with those of our past study using LC-infused layer-by-layer coatings and, more broadly, with the fact that as the alkyl chain length of a surfactant increases, the interfacial density of the adsorbed surfactant increases (compared to a shorter-tailed surfactant at a similar bulk concentration) and the fact that longer alkyl chains can penetrate deeper into the LC, which should lead to anchoring of E7 closer to the normal with respect to the aqueous/LC interface.^{20, 39, 40, 47} Finally, we note that the addition of 100 mM NaCl to solutions of C₁₂E₄ did not result in changes to droplet sliding speeds, consistent with the fact that C₁₂E₄ is a non-ionic surfactant and, thus, the interfacial density of the adsorbed surfactant should not be affected by the addition of an electrolyte.

Naked-Eye Detection of Small-Molecule Amphiphiles and Toxins in Droplets Extracted from Cultures of P. aeruginosa

Past studies have demonstrated the potential of different LC-based materials platforms, including planar aqueous/LC interfaces^{38, 40, 41, 46} and colloidal LC emulsions of free-floating micrometer-scale LC droplets in water, ^{46, 50-52} to sense and report on the presence of different environmental amphiphiles (such as lipids, proteins, and surfactants) with remarkable sensitivity. In those past studies, changes in LC orientation promoted by the adsorption of amphiphiles was generally characterized using polarized light microscopy^{41, 46, 50} or by changes in the forward- and side-scattering of light using flow cytometry.^{36, 53} While these analytical methods are effective, they require specialized and expensive instrumentation and, in general, some degree of technical knowledge to interpret the sometimes-complex results that arise from them. The LC-infused SLIPS reported here offer a new platform that translates factors that promote changes in the anchoring of LCs at aqueous interfaces (e.g., the presence of an amphiphile) to other readily observable macroscale phenomena (e.g., the readily observable rate at which a droplet of water slides across a surface).

The large and substantial differences in the sliding speeds of surfactant-containing and non-surfactant-containing droplets (see Videos S1) provide a straightforward and visual, ‘naked-eye’ approach for the detection of surfactants or other amphiphilic contaminants in aqueous environments. This approach would, in general, require no special equipment or expertise to interpret (e.g., a droplet sliding over a short distance within 4 s can be readily distinguished from a droplet that requires 1 min to traverse the same distance). In addition, because sliding speeds are also observed to vary as a function of surfactant concentration (Figure 2A), it is possible that this approach could also be used to provide estimates of the concentration of an analyte in an aqueous solution using equipment as basic as a stopwatch, or by using computer image analysis. Here, we note that in cases where differences in sliding speeds may be small and more difficult to distinguish, they can be further magnified by varying several simple extrinsic parameters, such as sliding angles or droplet volumes, that also have impacts on sliding times. For example, the difference in sliding time (Δt) between a 50 μL PBS droplet and a 50 μL droplet of 100 μM SDS (in PBS) increases from 7 ± 2 s to 86 ± 11 s when the sliding angle of the LC-infused SLIPS is reduced incrementally from 23° to 17°. Similarly, decreasing the volume of a droplet of PBS containing 100 μM SDS from 60 μL to 40 μL magnifies the difference in sliding time (Δt) compared to a PBS droplet from 7 ± 2 s to 37 ± 3 s (see Figure S4). For charged surfactants, the results discussed above (Figure 2B) also illustrate that the detection limits of LC-infused SLIPS can also be manipulated by modifying electrolyte concentration. When combined, modifications to both extrinsic and intrinsic parameters can be varied to increase or decrease droplet mobility and, in turn, influence the sensitivity of the response of an LC-infused SLIPS surface. Finally, on a practical note, we mention that these materials have the potential to be used and reused multiple times without affecting droplet sliding behaviors because changes in LC anchoring on LC-infused SLIPS are transient and reversible (see Video S2 and discussion above).

To demonstrate proof of concept and explore the potential of this approach to naked-eye detection, we performed a series of additional experiments to determine whether measurements of droplet sliding times on LC-infused PTFE membranes could be used to identify the presence of amphiphilic compounds produced by the bacterium P. aeruginosa, a common Gram-negative pathogen that uses the non-ionic amphiphilic small molecule 3-oxo-C12-AHL and the shorter-tailed, non-amphiphilic analog C4-AHL to regulate its quorum sensing (QS) system, and thereby control virulence.^{54, 55} P. aeruginosa uses QS to synchronize important behaviors once it achieves high cell densities, including biofouling and the production of amphiphilic toxins (i.e., rhamnolipid) that can be detrimental in environmental and healthcare settings.⁵⁶ These amphiphilic QS signals and virulence factors therefore represent markers of the presence of P. aeruginosa in a given environment and also provide useful information about the number of bacteria present and their virulence.

We recently reported that free-floating microscale droplets of 5CB suspended in aqueous media can be used to detect and report the presence of biologically-relevant concentrations of 3-oxo-C12-AHL and rhamnolipid, as well as an amphiphilic precursor to the biosynthesis of rhamnolipids [3-(3-hydroxyalkanoyloxy)alkanoic acid, (HAA)], using polarized light microscopy and flow cytometry (see Figure 3A for the structures of these amphiphiles).³⁶ The experiments described here sought to determine whether the amphiphilicity of these compounds could provide a basis for unaided, or naked-eye, detection of these bacterial products by simple measurement of the sliding times of droplets obtained directly from cultures of P. aeruginosa on the surfaces of LC-infused SLIPS.

Figure 3.

(A) Structures of the AHLs and bacterial biosurfactants investigated in this study (n = 3–11 for rhamnolipid and HAA). HAA was evaluated as a mixture of stereoisomers (see Materials and Methods). (B) Plot showing the sliding time of droplets of LB medium, P. aeruginosa WT (PAO1) culture, and P. aeruginosa QS-mutant (ΔrhlI lasI) culture at 6 h (black), 12 h (light gray), and 24 h (dark gray) on E7-infused SLIPS tilted at an angle of 20°. 2× and 4× denotes the number of folds of dilution of P. aeruginosa WT (PAO1) culture in LB medium before measuring the sliding time. (C) Top-down images showing 35 μL droplets of P. aeruginosa WT (PAO1) culture (4x diluted in LB medium before measuring the sliding time; blue) and QS-mutant (ΔrhlI lasI) (orange) at 0 s, 6 s, 16 s, and 36 s on LC-infused SLIPS (refer to Video S3).

We performed a series of initial experiments to measure the sliding times of 50 μL droplets containing known concentrations of 3-oxo-C12-AHL (over the range of 25–50 μM) and C4-AHL (over the range of 1-1000 μM) as well as rhamnolipid (over the range of 5–40 μg/mL), and HAA (over the range of 6.25-25 μg/mL) on E7-infused SLIPS (see Figure S5). The concentration ranges used in these experiments were selected to encompass the range of biologically relevant concentrations of these amphiphiles.^{36, 57} Solutions of C4-AHL, 3-oxo-C12-AHL, and HAA solutions were prepared in PBS containing 1% (v/v) DMSO to enhance solubility. Inspection of the results in Figure S5A-B shows that the sliding times of C4-AHL-containing and 3-oxo-C12-AHL-containing droplets are similar compared to those of control droplets (containing only PBS and 1% DMSO) over the range of concentrations used here. These results for C4-AHL are consistent with past results demonstrating that this short tail AHL does not promote changes in the configuration of free-floating LC droplet emulsions.³⁶ Similar studies of LC droplets demonstrate that 3-oxo-C12-AHL can adsorb at aqueous LC interfaces;³⁶ our current results for 3-oxo-C12-AHL on LC-infused SLIPS suggest that the concentrations used here are not sufficient to lead to large changes in the droplet sliding time. We also characterized the sliding times of droplets containing 3-oxo-C12-HS (the lactone hydrolysis product of 3-oxo-C12-AHL, an anionic amphiphile that forms naturally in aqueous solution and in cultures of bacteria)³⁴ over the range of 25–50 μM and did not observe significant changes relative to control droplets (PBS + 1% DMSO). Additional experiments using 3-oxo-C12-AHL and 3-oxo-C12-HS at concentrations that are higher than those typically reported in communities of planktonic bacteria (100 μM and 250 μM, respectively)^{36, 58} resulted in longer sliding times (~12 s). Further inspection of the results shown in Figure S5, however, reveals substantial differences in the sliding times of droplets containing HAA (at concentrations ≥ 12.5 μg/mL, Figure S5D) or rhamnolipid (at concentration ≥ 20 μg/mL; Figure S5E) compared to control droplets (PBS + 1% DMSO). For example, droplets containing 12.5 μg/mL HAA or 20 μg/mL rhamnolipid slid over a distance of 4 cm in ~18 s or ~30 s, respectively, compared to control droplets, which slid much more rapidly over the same distance (in ~4 s). These results demonstrate that measurements of droplet sliding times on LC-infused SLIPS can be used to report on the presence (or absence) of QS-controlled amphiphiles such as HAA and rhamnolipid in aqueous solutions.

We reasoned that the threshold of the surfactant concentration required to promote a change in anchoring of LCs (and reduction in droplet sliding times on LC-infused SLIPS) could be decreased by ‘spiking’ the surfactant solutions with additional amphiphilic species at lower (sub-threshold) concentrations; for example, low (sub-threshold) concentrations of SDS could potentially be added to sample droplets containing rhamnolipid to increase sensitivity. To explore the potential of this approach, we prepared solutions of different concentrations of rhamnolipid also containing 4 μM SDS (a concentration that, by itself, does not change the sliding time of an aqueous droplet; see Figure 2A and the discussion above) and measured the sliding times on E7-infused SLIPS. As shown in Figure S6, the addition of low concentrations of SDS resulted in a two-fold reduction of the limit of detection for rhamnolipid, from 20 μg/mL to 10 μg/mL. We did not further optimize the conditions used here or explore the lower limit of detection that is possible using this approach. These results do suggest, however, additional practical means by which differences in droplet sliding times can be manipulated or magnified to enhance the potential utility of LC-infused SLIPS in the context of sensing.

We next performed a series of biological experiments to determine whether LC-infused SLIPS could be used to monitor the production of bacteria-produced amphiphiles in live cultures. For these experiments, we cultured two different strains of P. aeruginosa: wild-type (WT, PA01) and ΔlasI rhlI (the latter is a genetic mutant strain of P. aeruginosa lacking genes critical to QS and that is, thus, unable to produce QS-associated factors, including rhamnolipid and HAA).³⁶ These experiments were performed in LB culture medium with shaking at 37 °C (see Materials & Methods for additional details). Aliquots of these bacterial cultures were removed at pre-determined time points (6, 12, and 24 hours), and the sliding times required for 35 μL droplets of these samples to slide 4 cm at a 20° incline were measured. Inspection of Figure 3B reveals that droplets of cultures of the ΔlasI rhlI mutant had short sliding times (~5 s) at incubation times of 6, 12, and 24 hours that were indistinguishable from the sliding time of LB medium alone. This result is consistent with the fact the ΔlasI rhlI mutant lacks genes critical to QS and, thus, is unable to produce either HAA or rhamnolipid. Further inspection of Figure 3B also reveals droplets taken from cultures of the WT mutant after only 6 hours of incubation to slide rapidly (over ~5 s). This result is consistent with the observation of low, sub-quorate populations of bacteria at this early time point that are unable to produce HAA or rhamnolipid.³⁶

In contrast, droplets taken from cultures of the WT mutant after 12- and 24-hours of incubation did not slide on LC-infused SLIPS and, instead, spread on the surfaces of these materials, likely due to the presence of a significantly higher concentration of QS-controlled surfactants. Further dilution of these samples with LB medium reduced droplet spreading and enabled meaningful measurements of sliding times. Two-fold dilution of samples taken at 12 hours of incubation of the WT strain resulted in droplet sliding times of ~21 s (Figure 3B). Samples taken after 24 hours of incubation, which would be expected to contain higher concentrations of HAA and rhamnolipid relative to earlier time points, required additional dilution; four-fold dilution with LB medium yielded sliding times of ~36 s.

These results from experiments on live bacteria are consistent with an increase in the concentration of these QS-controlled amphiphiles in WT P. aeruginosa cultures over time. Additional experiments were performed using two other mutant strains (ΔrhlA and ΔrhlB) that lack functional proteins in the rhamnolipid biosynthetic pathway (see Supporting Information, Figure S7). RhlA is upstream of RhlB, so the ΔrhlB mutant accumulates the intermediate HAA and does not convert it into rhamnolipid, while both HAA and rhamnolipid production is abrogated in ΔrhlA.⁵⁹ As expected, droplets of ΔrhlA culture collected after different incubation periods of 6, 12 and 24 hours exhibited sliding times on LC-infused SLIPS that were fast and indistinguishable from those of LB medium alone (~5 s), whereas an increase in sliding times was observed in samples collected from ΔrhlB culture at 12 and 24 hours, likely due to the presence of HAA (Supporting Information, Figure S7).

When combined, the results of these experiments demonstrate that measurements of the sliding times of droplets extracted directly from bacterial cultures on LC-infused SLIPS can be used to identify the presence of two amphiphilic factors (rhamnolipid and HAA) in cultures of P. aeruginosa and, in particular, distinguish between and monitor changes in the growth of sub-quorate and quorate populations of this human pathogen (see Figure 3C and Video S3). The results generated using the P. aeruginosa mutants described above also demonstrate that changes in droplet sliding times reported here are the result of the production of HAA and rhamnolipid, and not the result of other compounds produced by bacteria under these growth conditions. These large differences in droplet sliding times allow the presence and production of bio-surfactant toxins to be monitored visually and unambiguously in the absence of any additional specialized equipment or assays.

Detection and Monitoring of Amphipathic Peptide Toxins Produced by Cultures of S. aureus

We performed a final series of experiments to determine whether the results reported above could be used to identify the presence of amphipathic toxins produced by another common bacterium, the Gram-positive pathogen Staphylococcus aureus. It is well known that S. aureus produces a family of amphipathic α-helical peptides known as phenol soluble modulins (PSMs), also under the control of QS.⁶⁰ Amphipathic peptides differ substantially in structure from single-tailed surfactants, and it was unclear at the outset of these studies whether PSMs could adsorb at aqueous/LC interfaces and change the anchoring of LCs. We therefore prepared solutions of PSM-α3, one of several PSMs produced by S. aureus at high cell densities,⁶¹ at concentrations ranging from 12.5 mM to 100 mM in PBS and measured the sliding times of the droplets of these solutions on E7-infused SLIPS. As shown in Figure 4A, the sliding times of PSM-containing droplets increased substantially with an increase in the concentration of PSM. For example, the sliding time increased from ~16 s to ~93 s with an increase in the concentration of PSM-α3 from 25 mM to 100 mM (Figure 4A), suggesting that PSMs can interact with LCs and induce changes in the anchoring of LCs in ways that, at the least, lead to large changes in sliding behaviors that are similar to those observed above using conventional surfactants.

Figure 4.

(A) Plot showing the sliding time of 50 μL PBS droplets containing different concentrations of PSM-α3 (0 mM, 12.5 mM, 25 mM, 50 mM, and 100 mM) on LC-infused SLIPS tilted at 20°. (B) Plot showing the sliding time of droplets of BHI media and cultures of S. aureus QS mutant (lacking AgrBD, proteins critical for QS), WT (2x diluted in BHI media before measuring the sliding time), and WT cultured with AIP-III D4A at a concentration of 1 μM (refer Video S4, S5). All the S. aureus strains were cultured for 24 hours at 37 °C before measuring the sliding time.

We performed additional experiments to determine whether LC-infused SLIPS could detect the presence of PSMs in live cultures of S. aureus and, thereby, provide methods to monitor QS in communities of this pathogen in ways analogous to the studies of P. aeruginosa above. For these experiments, we cultured a S. aureus WT strain and a QS mutant strain (agr-nulllacking proteins critical for QS) for 24 hours and measured the sliding times of droplets of these cultures on LC-infused SLIPS (additional details of these experiments can be found in the Materials & Methods). Droplets obtained from cultures of the WT strain slid significantly more slowly (over ~27 s) compared to droplets obtained from cultures of the QS mutant strain (~7 s), consistent with the expected presence of PSMs in the WT culture (see Figure 4B and also Video S4). To provide support for this conclusion, we also performed experiments in which we added a known inhibitor of S. aureus QS (AIP-III D4A, at a concentration of 1 μM) to cultures of the WT S. aureus strain and measured the sliding time of the culture after 24 hours. AIP-III D4A has been demonstrated to fully inhibit S. aureus QS at concentrations ≥ 1 nM.⁶² We observed the sliding times of droplets of cultures incubated in the presence of this inhibitor to be comparable (~7 s) to those of the QS mutant (see Figure 4B and Video S5). Taken together, these results demonstrate that readily observed changes in the sliding times of droplets of S. aureus cultures can be used to not only detect the presence of PSMs and quorate populations of bacteria, but also identify the presence (or absence) of synthetic chemical inhibitors of QS. These results thus also suggest a potential basis for the development of straightforward droplet-based, bio-analytical screening assays that could be used as a tool to identify new synthetic inhibitors of bacterial QS. Experiments to this end are currently underway and will be reported in due course.

Summary and Conclusions

We have demonstrated that thermotropic LCs can be infused into nanoporous PTFE membranes to design slippery liquid-infused surfaces that can detect, monitor, and report on the presence of natural and synthetic amphiphiles in aqueous solution. In contrast to the behaviors of aqueous droplets on the surfaces of conventional slippery surfaces infused with isotropic oils, aqueous droplets slide on LC-infused SLIPS at speeds that depend strongly upon the presence, concentrations, and/or structures of dissolved amphiphiles. Sliding times of droplets on the LC-infused PTFE membranes reported here increase substantially—from times on the order of several seconds to times on the order of a minute—with increasing concentration of amphiphile. These large differences permit straightforward measurements of droplet sliding times to be used to estimate the concentration of an amphiphile in the droplet. Our results also reveal several other intrinsic and extrinsic parameters that can be used to further manipulate (e.g., speed up or slow down) droplet sliding times and, thereby, increase sensitivity or discriminate among amphiphiles of similar structure. Overall, our results are consistent with a physical picture that involves transient and reversible changes in the interfacial orientation of the anisotropic LCs at air/water interfaces mediated by the adsorption of amphiphiles to the LC/water interface. Our results also suggest that this approach has the potential to be general, and that LCs could likely be used in combination with other porous or rough surfaces typically used to prepare other SLIPS and LIS to introduce new and useful functions.

The materials reported here are straightforward to prepare, can be applied or transferred to a variety of secondary surfaces, and permit the unaided or ‘naked-eye’ detection and discrimination of amphiphilic contaminants in aqueous environments without the need for additional equipment or assays (in the case of large differences in sliding speeds) or with equipment that is no more sophisticated than a stopwatch (in cases where smaller differences in sliding speed may be observed). These features, combined with the low cost and ease of preparation of these materials, suggest opportunities to deploy these materials in the field and in low resource environments (e.g., ranging from clinics to water sampling studies to school science classes). To explore the feasibility of this approach and provide proof of concept in an applied context, we demonstrated the utility of these LC-infused surfaces for the naked-eye detection and monitoring of the production of small-molecule and peptidic amphiphilic bio-toxins in small droplets of fluid collected directly from cultures of P. aeruginosa and S. aureus, two clinically important bacterial pathogens. The ability of these LC-infused materials to translate molecular interactions at aqueous/LC interfaces into large and readily-observed, unambiguous changes in the sliding times of small aqueous droplets could open the door to new applications for anti-fouling, liquid-infused materials in the context of environmental sensing and in many other fundamental and applied areas.