Sticky or Slippery Wetting: Network Formation Conditions Can Provide

May 13, 2016 - Tailored topography: a novel fabrication technique using an elasticity gradient. Vartika Parihar , Saumyadwip Bandyopadhyay , Soumen Da...
0 downloads 0 Views 8MB Size
Download PDF
Research Article www.acsami.org

Sticky or Slippery Wetting: Network Formation Conditions Can Provide a One-Way Street for Water Flow on Platinum-cured Silicone Chenyu Wang, Sithara S. Nair, Sharon Veeravalli, Patricia Moseh, and Kenneth J. Wynne* Department of Chemical and Life Science Engineering, Virginia Commonwealth University, 601 West Main Street, Richmond, Virginia 23284, United States S Supporting Information *

ABSTRACT: In the course of studies on Sylgard 184 (SPDMS), we discovered strong effects on receding contact angles (CAs), θrec, while cure conditions have little effect on advancing CAs. Network formation at high temperatures resulted in high θadv of 115−120° and high θrec ≥ 80°. After network formation at low temperatures (≤25 °C), θadv was still high but θrec was 30−50°. Uncertainty about compositional effects on wetting behavior resulted in similar experiments with a model DVDH silicone elastomer (Pt-PDMS) composed of a vinyl-terminated poly(dimethylsiloxane) (PDMS) base and a polymeric hydromethylsilane cross-linker. Again, network formation at high temperature (∼100 °C) resulted in high CAs, while low-temperature curing retained high advancing CAs but gave low receding CAs (θrec 30−50°). These changes in receding CAs translate to strong effects on water adhesion, wp, which is the actual work required to separate a liquid (water) from a surface: wp ∝ (1 + θrec). When the values θrec 84° for high-temperature and θrec 50° for low-temperature network formation are used, wp is ∼1.5 times higher for curing at low temperature. The origin of low receding contact angles was investigated by attenuated total reflectance IR spectroscopy. Absorptions for Si−OH hydrogen-bonded to water (3350 cm−1) were stronger for low- versus high-temperature curing. This result is attributed to faster hydrosilylation during curing at higher temperatures that consumes Si−H before autoxidation to Si− OH. Sharp bands at 3750 and 3690 cm−1 due to isolated -Si−OH are more prominent for Pt-PDMS than those for S-PDMS, which may be due to an effect of functionalized nanofiller. To explore the impact of wp on water droplet flow, gradient coatings of S-PDMS and Pt-PDMS elastomers were prepared by coating a slide, maintaining opposite ends at high and low temperatures and thus forming a thermal gradient. When the slide was tilted, a droplet moved easily on the high-temperature end (slippery surface) but became pinned at the low-temperature end (sticky surface) and did not move when the slide was rotated 180°. The surface was therefore a “one-way street” for water droplet flow. Theory provides fundamental understanding for slippery/sticky behavior for gradient S-PDMS and Pt-PDMS coatings. A model for network formation is based on hydrosilylation at high temperature and condensation curing of Si−OH from autoxidation of Si−H at low temperatures. In summary, network formation conditions strongly affect receding contact angles and water adhesion for Sylgard 184 and the filler-free mimic Pt-PDMS. These findings suggest careful control of curing conditions is important to silicones used in microfluidic devices or as biomedical materials. Network-forming conditions also impact bulk mechanical properties for Sylgard 184, but the range that can be obtained has not been critically examined for specific applications. KEYWORDS: PDMS, Pt cure, curing temperature, contact angles, ATR-IR, water adhesion, gradient wetting

■

INTRODUCTION

Escherichia coli cells and biofilms were less susceptible to antibiotics.10 Emphasizing chemical approaches to hydrophilic PDMS, Brook and co-workers11 reviewed methods for the synthesis of amphiphilic copolymer networks, in which both the surface and the interior of the siloxane network had enhanced hydrophilicity. Seeking to simplify chemistry, amphiphilic poly(ethylene glycol) (PEG)-functionalized networks were pre-

Poly(dimethylsiloxane) (PDMS) has a -(CH3)2SiO- D repeat unit with a highly flexible -Si-O-Si- main chain (Tg ≈ −120 °C).1 Although Si−O bonds are polar, methyl groups on silicon shield polarity and the surface free energy for PDMS elastomers is low.2,3 The well-known resistance of PDMS (silicone) elastomers to adhesion and/or wetting is a benefit for applications such as coatings.4−9 However, this hydrophobic nature compromises usefulness in some applications. For example, in a study of hydrophobic versus hydrophilic coatings, a Sylgard 184 coating had the greatest number of adherent © XXXX American Chemical Society

Received: February 18, 2016 Accepted: May 13, 2016

A

DOI: 10.1021/acsami.6b02066 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces

ethylbenzene.44 Sylgard-184B is composed of dimethyl methylhydrogen siloxane, dimethylvinyl-terminated dimethylsiloxane, dimethylvinylated and trimethylsilylated silica, tetramethyl(tetravinyl)cyclotetrasiloxane, and ethylbenzene.45 Network formation is accomplished by a platinum catalyst. Vinyl-terminated poly(dimethylsiloxane) (MW 28 000), methylhydrosiloxane−dimethylsiloxane copolymer (trimethylsiloxy-terminated) (MW 900−1200, MeHSiO 50−55 mol %), and platinum− divinyltetramethyldisiloxane complex were purchased from Gelest, Inc. 2,4,6,8-Tetramethyl-2,4,6,8-tetravinylcyclotetrasiloxane was purchased from Tokyo Chemical Industry Co, Ltd. Hexane (99%) was obtained from Acros Organics. All chemicals were used as received. Coatings. Sylgard 184. Coatings were prepared by first handmixing the base and curing agent in a 10:1 mass ratio as per the instructions of the manufacturer. The container was then placed in a speed mixer-DAC 150FV (Flacktek Inc., Landrum, SC). High-speed (HS) mixing was then employed at 2700 rpm for 60 s. This HS mixing process was repeated two times to obtain a highly viscous, bubble-free, optically transparent resin. For dynamic contact angle (DCA) measurements, coverslips (Corning, 24 × 40 × 0.5 mm) were dip-coated in a freshly prepared two-component mixture and manipulated to give even resin distribution prior to curing. Coating thickness was determined by weighing coverslips or microscope slides before and after coating (mass difference = m). Thickness T is calculated from coating mass, area of the slide, and PDMS density d (∼1 g/cm3). For coverslips, T = m/2ad. Calculated coating thickness was typically 50 μm. For goniometry, 1 g of resin mixture was spread on each microscope slide (25 × 75 mm) to ensure constant thickness. Thickness was calculated as T = m/ad. Calculated thickness was ∼500 μm. After gelation (∼15 min), two sets of slides/coverslips were cured at 60 or 100 °C for 4 or 48 h. The resulting coatings are designated as SPDMS for Sylgard with cure temperature and time following. For this set, designations are S-PDMS-60(4), S-PDMS-60(48), S-PDMS100(4), and S-PDMS-100(48). Coatings were tacky after 4 h at ambient temperature, but curing to nontacky coatings took place after 48 h [S-PDMS-25(48)]. DVDH Elastomer. To generate Pt-cured elastomers similar to Sylgard 184 but of known composition and without a siliceous filler, DVDH samples were made in a manner analogous to that described by McCarthy and co-workers.13 In a typical preparation, 20 g of vinylterminated poly(dimethylsiloxane) (DV-PDMS), 2 g of methylhydrosiloxane−dimethylsiloxane copolymer (trimethylsiloxy-terminated) and 0.1 g of 2,4,6,8-tetramethyl-2,4,6,8-tetravinylcyclotetrasiloxane (inhibitor to retard gelation) were hand-mixed and then mixed in a speed mixer at 2700 rpm for 60 s. Platinum−divinyltetramethyldisiloxane complex solution (40 μL, 10 wt % in hexane) was added, followed by mixing at 2700 rpm for 30 s. An optically transparent, viscous resin was obtained. Coated microscope slides were prepared in the same way as described above for Sylgard 184. As for S-PDMS, two sets of DVDH samples were cured at 60 and 100 °C for 4 h and for 48 h. The resulting coatings are designated Pt-PDMS-60(4), Pt-PDMS-60(48), Pt-PDMS-100(4), and Pt-PDMS-100(48). One set was cured at ambient temperature (∼25°) for 48 h, yielding nontacky Pt-PDMS25(48). Gradient Coatings. To generate a temperature gradient, one end of a bare microscope slide was placed on a hot plate set at 200 °C while the other end was on a plate at low temperature (Figure 1). The slide was leveled and the temperature gradient was developed over 15 min. A coated microscope slide (before curing, coating thickness ∼500 μm) was then placed on the bare slide. Network formation occurred over a temperature gradient from the high end (∼200 °C) to the low end (∼0 °C for S-PDMS or ∼25 °C for Pt-PDMS). Heat conduction no doubt resulted in the hot end being 80°).13,26 However, if the mixed resin was held at ambient temperature for some time before carrying out highertemperature treatment, high advancing contact angles (CAs) were retained but low receding CAs were found (30−50°). The result was high contact-angle hysteresis (CAH = θΔ = θadv − θrec). Curious about this result, we prepared a model platinumcured silicone similar to that described by McCarthy and coworkers,13 which also gave a low receding CA for the “ambient temperature first” step for the cure process. In view of the importance of receding CAs to work of adhesion,16,27−29 electrowetting,26 and biophysical processes such as protein adsorption, a study of θrec as a function of network formation conditions was undertaken. For both Sylgard 184 and a Pt-catalyzed network prepared from divinyl poly(dimethysiloxane) (DV) and poly(hydrodimethysiloxane) (DVDH), low receding CAs were observed for network formation at ambient temperature or when an initial lowtemperature step was employed. The impact of low θrec was demonstrated by sticky wetting behavior for coatings where a low-temperature step was employed but slippery behavior without this step and only high-temperature cure. A mechanism for low θrec resulting from the low-temperature step is supported by attenduated total reflectance infrared (ATR-IR) spectroscopy.

■

EXPERIMENTAL SECTION

Materials. A Sylgard-184 kit from Dow Corning is supplied in two parts, Sylgard-184A (base) and Sylgard-184B (curing agent). According to the material safety data sheets (MSDS), Sylgard-184A is composed of dimethylvinyl-terminated dimethylsiloxane, dimethylvinylated and trimethylated silica, tetra(trimethoxysiloxy)silane, and B

DOI: 10.1021/acsami.6b02066 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces

same phase scale (z, degrees) as indicated in figure legends. The phase scale was chosen to optimize image quality and consistency with topographical images. Nanomechanical properties were obtained with a Bruker Dimension ICON atomic force microscope by use of PeakForce quantitative nanomechanical mapping (QNM) software. This mode maps nanomechanical properties including modulus and adhesion while simultaneously imaging topography. Mechanical properties are obtained from calibrated force curves. QNM software generates the reduced Young’s modulus E* by fitting the retraction curve by use of the Derjaguin−Muller−Toropov (DMT) model.31,32 Fracture surface areas (10 × 10 μm) were analyzed for S-PDMS coatings by use of a Scanasyst-air tip (Bruker) with a nominal spring constant of 0.4 N/m; the peak force set-point was 0.08 V. Attenuated Total Reflectance Infrared Spectroscopy. A Thermo Scientific Nicolet iS10 spectrometer equipped with a Smart iTR attachment and a diamond crystal was used for ATR-IR spectra. The Smart-iTR attachment employs a pressure of ∼40 psi that is achieved by placing the sample on the Dia crystal and turning a knurled knob until a click is obtained. Spectra were analyzed with Omnic software. Hydrosilylation and/or hydrolysis−condensation curing takes place depending on network-forming conditions. The influence of curing temperature on the presence or absence of -Si−OH in the wavelength range 3100−3900 cm−1 was investigated. At 3500 cm−1, the depth of penetration of the evanescent wave is ∼475 nm. A calculation is provided in Supporting Information.

Figure 1. Preparation of gradient coatings. PDMS-200, while the other end is S-PDMS-0(4)/100(0.5). To complete network formation, Pt-PDMS gradient coatings were heated at 100 °C for 4 h. These gradient coatings are designated Pt-PDMS200 for the high-temperature end or Pt-PDMS-25(48)/100(4) for the low-temperature end. These designations are for convenience, as gradient temperatures are not known accurately. Contact Angles. Dynamic Contact Angles. Dynamic contact angles (DCA) were obtained by the Wilhelmy plate method25 on dipcoated coverslips using a Cahn model 312 analyzer (no longer manufactured). Glass containers used for DCA analysis were cleaned by rinsing with Nanopure water and treatment with a gas/oxygen flame. Water surface tension (72.6 ± 0.4 dyn·cm−1) was checked before each experiment with a flamed glass coverslip. By analyzing force versus distance curves (fdc), advancing (θadv) and receding (θrec) contact angles were obtained via the relationship F = m/g = Pγ(cos θ), where F is the force derived from respective mass changes (milligrams) on immersion and emersion, g is the gravitational constant, γ is the liquid surface tension, and θ is the contact angle.30 Extrapolating the fdc to the point of immersion eliminates the need for a buoyancy correction to F. Three cycles in succession were conducted with stage speeds of 100, 50, and 40 μm/s, but no dependence on rate of immersion was found. Contact angles (±1−2°) are averages of several force−distance cycles. Water was tested for purity after each sample analysis to examine the extent of water contamination due to leached species.30 Static Contact Angles. Static contact angles were obtained by use of a Rame−Hart goniometer equipped with a liquid crystal display (LCD) camera. Deionized water (∼18.2 MΩ) was used as the probe liquid. A water drop (20 μL) was placed on the coated surface and the image was captured immediately. Captured images were analyzed and water contact angles (WCAs) were obtained by use of Dropview image software version 1.4.11. Average values were obtained from five observations. Droplet Mobility. Drop images for tilted slides provided a supplemental measure of receding contact angles. Water drop mobility tests were done by placing a 40 μL water drop on the hightemperature end of gradient coatings and subsequently following the course of travel to a point where the drop was pinned. Rotating the slide 180° confirmed adhesion of the droplet without any movement. Images of droplets were obtained after inversion (turning slides upside down) to further investigate water droplet adhesion. Atomic Force Microscopic Imaging. Morphological and nanomechanical investigations were carried out on a Dimension-3100 (Digital Instruments, CA) atomic force microscope (AFM) with a NanoScope V controller. Tapping-mode imaging was performed in air by use of microfabricated silicon cantilevers (40 N/m, Veeco, Santa Barbara, CA). Unless otherwise noted in figures, tapping force corresponded to set-point ratios rsp of 0.8 and 0.6, where rsp = Aexp/A0, A0 is free oscillation amplitude, and Aexp is experimental oscillation amplitude. Images were analyzed by using NanoScope v710r1 software. For comparisons, AFM images were normalized to the

■

RESULTS AND DISCUSSION Hydrosilylation, the catalyzed addition of Si−H to an unsaturated vinyl moiety, was made practical by discoveries of Saam and Speier,33, Speier,34 and Karstedt.35,36 Network formation via platinum cure with Karstedt’s catalyst results in C−H and Si−C bonds. An important advantage is minimal shrinkage compared to condensation curing. The Pt-catalyzed addition reaction is now one method for making biomedicalgrade silicones.37,38 Sylgard 184. Sylgard 184 is available as a convenient twopart resin system. This silicone elastomer is widely used, with numerous reports of WCAs. However, unlike any prior report, we find that a change in network-forming conditions can result in much lower receding CAs (30−50°) compared to those typically reported (e.g., θadv 117°; θrec 94°).39 In the following subsections, we discuss processing, surface morphology, and mechanical properties (AFM) before describing wetting behavior and a “one-way street for water drop flow” that emphasizes the influence of receding CAs on water droplet adhesion. Processing. Sylgard 184 base and curing agent were mixed in the usual 10:1 mass ratio. Under laboratory conditions, both bulk mechanical and surface properties can be adversely affected by bubble formation that often results from hand mixing. Removal of bubbles can be accomplished by exposure to nitrogen gas or vacuum.13,40 Use of high-speed mixing (Experimental Section) eliminates bubble formation and results in homogeneity that is obviously important but difficult to obtain with viscous resins. Nanoscale Morphology. Fillers are required to confer strength to weak PDMS elastomers. AFM images such as those reported in studies of fumed silica nanocomposites are easily obtained due to large differences in modulus between nanoparticles and silicone resin.41 However, under the acquisition conditions employed, AFM images for S-PDMS coating surfaces were featureless, regardless of cure temperature. Most of our images of thin films were featureless and C

DOI: 10.1021/acsami.6b02066 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces agreed with the findings of Vansco and co-workers,42 who reported smooth, structureless images for Sylgard 184 using network formation conditions similar to our S-PDMS-100. However, after several processing and cure options were explored, spin-coated films of S-PDMS-60(48) were imaged successfully. Figure 2 shows 1 × 1 μm height and phase images

cross-linker Si−H oxidation to Si−OH and condensation chemistry lead to a third network. Peak Force Tapping Atomic Force Microscopy. A recent study by Nair et al.46 showed that AFM nanoindentation profiles with force variation together with conventional tappingmode AFM imaging (Asylum instrumentation) and X-ray photoelectron spectroscopy provided detailed nanoscale morphological characterization of a hybrid modified polyurethane to a depth of ∼20 nm. With interest in S-PDMS solidstate morphology and nanomechanical properties, we turned to peak force tapping AFM using a Bruker instrument. For this imaging mode, tip nanoindentation with recording each force distance curve results in a modulus map of the scanned area.47−50a Figure 4 shows height and modulus (DMT) images

Figure 2. Height and phase images for S-PDMS-60(48): set-point ratio (A) 0.8 and (B) 0.6. Rq (nanometers) is shown on height images.

that are typical of several. Though faint at moderate tapping (rsp 0.8), images obtained at rsp 0.6 revealed near-surface nanoscale features. Some of these features are round (diameter ∼120 nm), but others are irregularly shaped with varying nanoscale sizes. With false color imaging, lighter features in phase imaging are associated with increased stiffness or increased modulus.43 The nanoscale features in Figure 2B are therefore assigned to siliceous nanofiller. The modest contrast for the nanoscale features is attributed to hydrosilylation grafting of the polymeric cross-linker dimethyl methylhydrogen siloxane in Sylgard 184B to dimethylvinylated silica in Sylgard 184A44,45 that forms a soft shell for the nanofiller. Optical transparency is retained, as there is negligible light scattering for these nanoscale features in the visible region of the electromagnetic spectrum (ca. 400−700 nm). Based on AFM imaging, a simple model for this nanostructured elastomer is depicted in Figure 3A, illustrating nanofiller-based and chain-based hydrosilylation chemistry. Other studies to be described later extend upon this model by recognizing that, depending on network-forming conditions,

Figure 4. Height and peak force tapping AFM images (10 × 10 μm) for (A) S-PDMS-25(48), (B) S-PDMS-60(48), and (C) S-PDMS100(48).

for Sylgard fracture surfaces. The average modulus for SPDMS-25(48) is 6.5 MPa. A slight increase in modulus to 6.6 MPa is observed for S-PDMS-60(48), while an increase to 9.8 MPa is found for S-PDMS-100(48). Several protocols have been used for curing Sylgard 184 and the determination of mechanical properties. Only representative mechanical property data are presented and discussed. In a study of human cell adhesion, Palchesko, et al.50b carried out cure at 65 °C (12−24 h) and Instron-based stress−strain modulus measurements (1.7 MPa). Volinsky, et al.50c also chose 65 °C (1 h) and reported 2.6 MPa. Johnston et al.40 assumed complete cure for 25, 100, and 150 °C. Young’s modulus increased from 1.32 to 2.05 MPa for 25 and 100 °C cure. For 95 °C cure (30 min), which approximates cure conditions for S-PDMS-100, Roy and co-workers51 reported a modulus of 1.54 MPa (Instron, ASTM D 412). From these and many other reports, cure conditions have ranged from ambient to 150 °C, resulting in modulus determinations in the 1−3 MPa range. A clear trend in increased modulus with cure temperature was identified by Johnston et al.40 This trend agrees with the increase in modulus from nanomechanical measurements described above, namely, 6.5 MPa/25 °C to 9.8 MPa/100 °C. However, the modulus determined on fracture surfaces by nanomechanical measurements is higher by a factor of ∼3 than macromeasurements. The source of the difference between nanomeasurements and macromeasurements is not clear and will be the subject of further study. Except for a few nanopeaks in the height image (light features), a featureless fracture surface is observed for S-PDMS-

Figure 3. Proposed network model for (A) filled Sylgard 184 and (B) DVDH network. Red chains, DH cross-linker bound to nanofiller; green chains, DH cross-linker bound only to DV end groups. -Si-CH2CH2moieties from hydrosilylation are depicted by filled circles. D

DOI: 10.1021/acsami.6b02066 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces

with CAs. Static WCAs are similar to those typically reported for silicones (110°).13 When the different methods and different samples are considered, there is reasonable agreement between DCA and advancing/receding drop measurements for θadv. Of note, dynamic advancing CAs are systematically ∼10° higher than static WCAs. Static or sessile drop contact angles result from an expanding drop that comes to rest after deposition (20 μL). The perimeter of the drop determines the three-phase contact line, which may be relaxed due to reorganization of polymer chains at the outermost surface or other process. In dynamic measurements the three-phase contact line is less influenced by relaxation processes, as the three-phase contact line is constantly refreshed so that higher advancing CAs are observed. By advancing/receding drop measurements (Figure 6), receding drop CAs show a trend similar to DCA measurements (Figure 3), although higher receding contact angles are found for 25 and 60 °C cure. These differences are attributed primarily to the different measurement methods. In addition, as will be discussed, receding contact angles are typically less reproducible than advancing contact angles, as the receding three-phase contact line may encounter chemical and topological inhomogeneities. Table S1 in Supporting Information provides additional measurements of θadv and θrec for the sample shown in Figure 6. A higher standard deviation was found for receding contact angles. Droplet Mobility. Figure 7 shows water drop images comparing S-PDMS-25(48) and S-PDMS-60(48). For S-

25(48). Faint heterogeneity is seen for the S-PDMS-60(48), while a plethora of nanoscale features with relatively high modulus are observed for S-PDMS-100(48). Paralleling the surface studies already discussed, high modulus features are attributed to nanofiller, as it is likely that the presence of nanofiller vinyl functionality results in covalent binding of DH cross-linker (Figure 2A). The trend of increasing cross-link density with increasing curing temperature will be discussed in the context of a model for wetting behavior. Dynamic Contact Angles. Figure 5 shows DCA force− distance curves (fdc) for S-PDMS-25(48), S-PDMS-60(48),

Figure 5. DCA contact angles for (A) S-PDMS-25(48), (B) S-PDMS60(48), and (C) S-PDMS-100(48).

and S-PDMS-100(48). The advancing CA of ∼123° is about the same for all network-forming conditions. However, θrec is 34° for S-PDMS-25(48) compared to 85° for S-PDMS100(48). An intermediate value is observed for cure at 60°. Contact angle hysteresis θΔ is ∼90° for S-PDMS-25(48) while that for S-PDMS-100(48) is 38°. Goniometry was investigated to augment these results. Goniometry. S-PDMS coatings are nanoscopically smooth (Figure 2) and are assumed not to have a contribution to contact angles from surface roughness.52 Figure 6 shows water drop images for Sylgard coatings cured at three temperatures. Static drops and advancing/receding drops are shown along

Figure 7. Water drop adhesion for (A) S-PDMS-25(48) and (B) SPDMS-60(48).

PDMS-25(48), the droplet remained stationary, maintaining the three-phase contact line during tilt (Figure 7A). In contrast, the drop moved down the surface rapidly on S-PDMS-60(48) (Figure 7B). In order to capture the latter images, a tilt of 22° was used to slow droplet movement. Images for slides turned upside down show that the threephase contact line is greater for S-PDMS-60(48); in fact, the drop shown in Figure 7B fell off by gravity a few seconds after the image was obtained. This confirms the impact of a high receding CA on water drop adhesion. Gradient Sylgard 184 Coating. In 1992, Chaudhury and Whitesides53 studied a gradient surface prepared by diffusing the vapor of decyltrichlorosilane over a silicon wafer. For the hydrophilic end θadv was ∼25°, but for the hydrophobic end θadv was ∼98°. For the hydrophilic end θrec was ∼0°, while it was ∼88° for hydrophobic end. Driven by low advancing and

Figure 6. Contact angles by goniometry for (A) S-PDMS-25(48), (B) S-PDMS-60(48), and (C) S-PDMS-100(48). E

DOI: 10.1021/acsami.6b02066 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces receding CAs, when inclined with the hydrophilic end higher, a water drop “ran uphill” from the hydrophobic end. For Sylgard coatings, only differences in the receding CA present a gradient. However, results shown in Figure 7 suggested an experiment testing water drop adhesion. A coated microscope slide was cured over a thermal gradient from high (∼200 °C) to low (≤ambient) temperature. After 4 h the lowtemperature end was tacky, so curing was completed at 100 °C for 30 min. The notation S-PDMS-0(4)/100(0.5) indicates the low-temperature end was held at ∼0 °C for 4 h and at 100 °C for 30 min. An advancing CA of ∼120° was observed over the entire surface (not shown), while a receding CA gradient was obtained, reflecting S-PDMS network formation at three different areas (Figure 8). The high-temperature end (S-PDMS-200) had a receding CA of 88°, while the low-temperature end [S-PDMS0(4)/100(0.5)] had a CA of 38°. The receding CA for the middle was 51° (Figure 8A). Thus, this coating has a gradient receding CA and a constant advancing CA.

A water droplet roll-off test was used to characterize this distinctive gradient surface. Three water droplets of equal volume (40 μL) were placed at the high-temperature end (SPDMS-200), at the low-temperature end [S-PDMS-0(4)/ 100(0.5)], and in the middle of the coated microscope slide. The coated slide was slowly tilted lengthwise (Figure 8B). The water droplet at the high-temperature end rolled off at a tilt angle of ∼60°, followed by the droplet in the middle, which rolled off at a tilt angle of ∼80°. The droplet at the lowtemperature end remained pinned to the surface even at 90°. This result can be explained by differences in water adhesion strength. The work of Gao and McCarthy54,55 suggested that the work of water adhesion could be quantified by the relationship wp ∝ (1 + θrec), where wp represents the actual work required to separate a liquid (water) from a surface. To estimate the effect of differences in receding contact angle, the ratio of wp was calculated with θrec values for high- and low-temperature ends (Supporting Information). Receding water CAs were those reported in Figure 8A. By this method, wp for the lowtemperature end is ∼1.7 times higher than that for the end cured at high temperature. This difference in water adhesion accounts for the difference in water droplet mobility. In order to explore further the influence of receding contact angles on water droplet adhesion, a water droplet of 25 μL was placed on the end of the Sylgard 184 coating cured at high temperature. The slide was slowly tilted lengthwise. As shown in Figure 9, the water drop started to move at a tilt angle of ∼60° (a) and continued moving with increasing tilt angle (b and c). The droplet stopped at the end cured initially at low temperature with a tilt angle of 90°. The slide was then turned 180° while the tilt angle was maintained at 90° (d). The droplet did not move back to the end cured at high temperature. Rather, the droplet was pinned on the end cured initially at low temperature. Because of the gradient in receding CA, the coating surface constitutes a “one-way street” for water drop movement. Hydrophobic recovery is a well-studied phenomenon that occurs after high-energy treatment of silicone coatings, including Sylgard 184.13 After optimum plasma oxidation (30 s) of Sylgard 184, Gao and McCarthy54,55 reported an increase in receding CAs from ∼5° to >50°. Concern about the stability of CAs led to repeating the one-way street experiment ∼2 months after gradient preparation. No visible change was observed. We conclude that the low/high receding CA gradient has good stability at ambient temperature. This stability is attributed to high -(MeHSi-O)- content of the copolymer− cross-linker and autoxidation to near-surface -[MeSi(OH)-O]that is network-stabilized. DVDH Elastomer. As discussed, the receding CA for Sylgard 184 can be easily controlled by using different curing protocols. However, Sylgard 184 is a commercial product with a composition that is only generally known. The issue of uncertain effects on wetting behavior from unknown constituents is thus raised. As a result, similar experiments to those already described were carried out on a synthesized DVDH silicone elastomer with known composition. This DVDH elastomer, Pt-PDMS, is composed of a 28 kDa vinyl-terminated PDMS base and a polymer cross-linker with Si−H functionality. Network formation was Pt-catalyzed; a small amount of inhibitor was added to retard gelation. Gradient DVDH Elastomer. A gradient surface for Pt-PDMS was prepared by a process similar to that for S-PDMS, except

Figure 8. (A) Receding CAs for gradient Sylgard coatings. (B) Frame grabs for droplet roll-off test at inclination angles shown at left (⊗, no motion). F

DOI: 10.1021/acsami.6b02066 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces

Figure 9. Screen-captured images of a video showing one-way street behavior for the gradient S-PDMS coating. H = high-temperature end; L = lowtemperature end. (a−c) Droplet movement from H to L at tilt angles shown; (d) slide shown in panel c rotated 180° (⊗, no motion).

the curing time was extended due to a slower rate of network formation. Advancing CAs were ∼115° by goniometry irrespective of cure temperature. Figure 10 shows receding

Figure 10. Receding contact angles for gradient DVDH elastomer. Figure 11. Screen-captured images of a video showing one way street behavior of Pt-PDMS. H = high-temperature end; L = lowtemperature end; slide at 90°. (a−c) Progress of droplet movement from H to L; (d) slide rotated 180° (⊗, no motion).

CAs decreased from 84° for Pt-PDMS-200 at the hightemperature end to 50° for the low-temperature end, PtPDMS-25(48)/100(4). For this gradient, the receding contact angle for low-temperature cure was about the same or somewhat higher than comparable measurements on the SPDMS gradient (Figure 8A). The Pt-PDMS gradient was placed vertically with the end cured at high temperature on top, and a water drop was placed on this surface. As shown in Figure 11, the water droplet rolled down immediately. When it reached the area near the lowtemperature cured end, the droplet stopped. The droplet did not move when the slide was rotated 180° in the plane. Thus, one-way street behavior was demonstrated for the Pt-PDMS gradient. Again, the effect of curing temperature on receding CA was stable with time. The receding CA for the Pt-PDMS gradient showed no change 1 month after preparation. Attenuated Total Reflectance Infrared Spectroscopy. Low receding CAs resulting from network formation at low temperatures (≤25 °C) suggested the presence of near-surface Si−OH. ATR-IR spectroscopy was employed for characterization, as Si−OH absorptions occur in the 3200−3800 region. Isolated Si−OH groups on silica show a sharp band at 3750 cm−1, while hydrogen bonding to Si−OH lowers the O−H stretching frequency to ∼3400 cm−1.56−58

S-PDMS and Pt-PDMS elastomers were obtained after network formation for 48 h at low (≤25 °C), medium (∼60 °C), and high (100 °C) temperatures. Figure 12 shows spectra for S-PDMS-25(48) and Pt-PDMS-25(48). Strong, broad absorptions at 3350 cm−1 are assigned to Si−OH hydrogenbonded to water. Autoxidation of poly(hydrosilanes) has been previously investigated.59,60 Kurian and Kennedy61 assigned a broad band at ∼3500 cm−1 to Si−OH groups arising from autoxidation of SiH. For S-PDMS-25(48) and Pt-PDMS25(48), no absorption is seen at 2140 cm−1 for Si−H, indicating that this autoxidation is complete.61 Spectra are shown in Figure S1 in Supporting Information, along with spectra after network formation for 4 and 48 h at 60 and 100 °C. The area of the 3350 cm−1 absorption decreases with increasing temperature, at which network formation takes place (Figure 12). This result is attributed to faster hydrosilylation for curing at higher temperatures that consumes Si−H before autoxidation to Si−OH. Condensation of Si−OH to Si−O−Si at higher temperatures also decreases Si−OH. The combination G

DOI: 10.1021/acsami.6b02066 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces

reactions proceed simultaneously: hydrosilylation and autoxidation. For cure at high temperature, hydrosilylation is rapid with network formation by Si−C and C−H bonds (Figure 13A,B). Autoxidation of Si−H presumably occurs, but Si−OH groups are trapped in the bulk as isolated entities. The latter is clear from ATR-IR spectra, where 3690 and 3750 cm−1 peaks are seen for S-PDMS and Pt-PDMS after 48 h cure at 100 °C but hydrogen-bonded Si−OH is weak or absent. After network formation at ambient temperature for 4 h, SPDMS and Pt-PDMS coatings were tacky initially and required 48 h for a tack-free state. Figure 13C,D depicts the result of network formation near or below ambient temperature. Autoxidation of Si−H is faster than hydrosilylation, resulting in a network with low -HC−C−Si- cross-link density. In Figure 13C, formation of -HC−C−Si- is represented by the small black ovals. ATR-IR demonstrates that autoxidation at low temperature in air results in a broad absorption at 3350 cm−1, characteristic of near-surface Si−OH and represented by blue features on the perimeter of the green oval representing the cross-linking copolymer (Figure 13C,D). In air, near-surface SiOH groups are unfavored compared to low-surface-energy PDMS chains, and the advancing CA is high. However, SiOH groups are strongly hydrophilic and the Si−O−Si main chain is highly flexible (Tg < −100 °C). As a result, near-surface SiOH groups are enthalpically driven to the outermost surface upon immersion in water and a low receding CA is observed (