Highly Transparent Superhydrophobic Surfaces from the Coassembly

Feb 28, 2011 - ... interest in applications such as solar cell panels and window treatments. ..... As seen in Movie S2, a smaller water droplet, 5 μL...
0 downloads 0 Views 1MB Size
ARTICLE pubs.acs.org/Langmuir

Highly Transparent Superhydrophobic Surfaces from the Coassembly of Nanoparticles (e100 nm) Raghuraman G. Karunakaran, Cheng-Hsin Lu, Zanhe Zhang, and Shu Yang* Department of Materials Science and Engineering, University of Pennsylvania, 3231 Walnut Street, Philadelphia, Pennsylvania 19104, United States

bS Supporting Information ABSTRACT: We report a simple and versatile approach to creating a highly transparent superhydrophobic surface with dual-scale roughness on the nanoscale. 3-Aminopropyltrimethoxysilane (APTS)-functionalized silica nanoparticles of two different sizes (100 and 20 nm) were sequentially dip coated onto different substrates, followed by thermal annealing. After hydrophobilization of the nanoparticle film with (heptadecafluoro-1,1,2,2-tetrahydrodecyl)trichlorosilane for 30 min or longer, the surface became superhydrophobic with an advancing water contact angle of greater than 160° and a water droplet (10 μL) roll-off angle of less than 5°. The order of nanoparticles dip coated onto the silicon wafer (i.e., 100 nm first and 20 nm second or vice versa) did not seem to have a significant effect on the resulting apparent water contact angle. In contrast, when the substrate was dip coated with monoscale nanoparticles (20, 50, and 100 nm), a highly hydrophobic surface (with an advancing water contact angle of up to 143°) was obtained, and the degree of hydrophobicity was found to be dependent on the particle size and concentration of the dip-coating solution. UV-vis spectra showed nearly 100% transmission in the visible region from the glass coated with dual-scale nanoparticles, similar to the bare one. The coating strategy was versatile, and superhydrophobicity was obtained on various substrates, including Si, glass, epoxy resin, and fabrics. Thermal annealing enhanced the stability of the nanoparticle coating, and superhydrophobicity was maintained against prolonged exposure to UV light under ambient conditions.

’ INTRODUCTION A surface that exhibits a water contact angle of 150° or greater with very little flow resistance, such as observed for lotus leaves1 and the legs of water striders,2 is considered to be superhydrophobic. Such surfaces are self-cleaned by the enrolling water droplet and therefore are of interest in numerous applications, including selfcleaning coatings, impermeable textiles, microfluidics, and lab-on-achip devices. Generally, the wetting behavior is dependent on both the surface chemistry (i.e., surface energy) and surface topography (i.e., physical roughness). Surfaces with the same solid-liquid interfacial area but smaller feature sizes have smaller effective contact areas and less-stable three-phase contact lines and thus higher water droplet mobility.3 Roughness on two or more length scales has been suggested to be responsible for the observed superhydrophobicity observed on aquatic plant leaves and insect legs and wings.1,2,4-6 For example, the lotus leaf surface is composed of 3D epicuticular wax crystals that have micrometer-sized papillose epidermal cells with sub-micrometer-sized randomly oriented tubules on them.1 Recent studies suggest that nanoscale roughness has the added benefit of enhancing the roughness, especially when the feature size of the microstructures is large and the asperity is small.7,8 Fascinated by discoveries pertaining to natural surfaces, many researchers have attempted to fabricate multiscale roughness in synthetic materials together with chemical functionalization to achieve superhydrophobicity.3,9-20 Sub-micrometer-sized colloidal particles12,19 and microstructured pillar arrays fabricated by photolithography3,21 and soft r 2011 American Chemical Society

lithography techniques22,23 are often used to provide a predefined microroughness. Other techniques, such as chemical vapor depositio,24 and electrochemical deposition,25 have also been employed to generate random microroughness. Nanoparticles (NPs), however, are commonly used to control the surface nanoroughness12,19,26-28 because they are readily available commercially and can be synthesized by St€ober chemistry29 with uniform size and tunable surface chemistry. In addition, their deposition on a microstructured surface by spin coating and dip coating is straightforward and inexpensive. Despite these advances, how small the nanoscale roughness should be to achieve superhydrophobicity remains unclear. Meanwhile, little attention has been paid to the optical transparency of the obtained rough surfaces until recent interest in applications such as solar cell panels and window treatments. Surface roughness and transparency are generally competitive properties. The blurring of the surface due to roughness is mainly caused by a Mie scattering effect. Rayleigh scattering is applied when the radius of the scattering particle is much smaller than the wavelength of the incident light. The optical quality of the film increases when the roughness dimension is much less than the wavelength of light. The intensity of the Rayleigh scattering radiation increases rapidly as the ratio of particle size to Received: October 9, 2010 Revised: February 7, 2011 Published: February 28, 2011 4594

dx.doi.org/10.1021/la104067c | Langmuir 2011, 27, 4594–4602

Langmuir

ARTICLE

Figure 1. Schematic illustration of APTS functionalization of a silica nanoparticle.

wavelength increases. Therefore, sub-100-nm roughness is preferable for achieving high transparency in visible light. A few groups have attempted to create transparent superhydrophobic surfaces.16,18,20,30-34 The fabrication often either requires expensive lithographic tools to achieve small feature sizes or involves rather complicated chemical synthesis procedures. Ling et al.35 recently demonstrated a simple approach by dip coating a single layer of 60 nm SiO2 NPs onto an amine-terminated selfassembled monolayer (SAM)-coated glass/silicon oxide substrate, followed by perfluorosilane deposition. The authors suggested that the random, moderately dense coverage of NPs, in contrast to that of close-packed NPs obtained by convective assembly, was essential to achieving superhydrophobicity. The results are different from the literature with respect to the necessity of dual-scale roughness for superhydrophobicity as well as our previous study when coating a monolayer of aminosilanefunctionalized SiO2 NPs onto a swollen poly(styrene-ran-acrylic acid) film. In the latter, Cassie-Baxter wetting behavior was observed on close-packed (70% coverage) NPs (15, 50, and 106 nm in diameter) whose surfaces were wetted by polystyrene segments.36 A transition from Cassie-Baxter state wetting to Wenzel state wetting was observed when the particle size was increased to 230 nm in diameter with 30% coverage. These observations motivate the present studies, which are used to determine whether single-scaled NPs can achieve superhydrophobicity as observed by Ling et al.35 and if not, the maximum size of NPs and the ratio between larger to smaller NPs used to achieve superhydrophobicity without sacrificing film transparency. Herein, we report a simple yet versatile strategy for creating a highly transparent superhydrophobic surface via the coassembly of amine-functionalized silica NPs of two different sizes (100 and 20 nm). The NPs were sequentially dip coated onto a Si wafer, followed by thermal annealing (200 or 400 °C for 2 h) to enhance the adhesion and stability of the NP film on the Si wafer. After vapor deposition of the NP film with (heptadecafluoro-1,1,2,2,-tetrahydrodecyl)trichlorosilane for more than 30 min, the surface became superhydrophobic with an advancing water contact angle (θadv) of greater than 160° and a water droplet (10 μL) roll-off angle of less than 5°. In contrast, surfaces coated with monoscale nanoparticles (20, 50, and 100 nm) were very hydrophobic, with an advancing water contact angle of up to 143°, but also possessed large contact angle hysteresis. Because the NP size was e100 nm, the obtained particle films on glass substrates were found to be both superhydrophobic and highly transparent: nearly 100% transmission in the visible region, which is similar to that of bare glass. We further demonstrated the versatility of our approach by coating dual-scale nanoparticles on other substrates, such as epoxy and fabrics. The obtained superhydrophobicity is robust against prolonged exposure to UV light under ambient conditions.

’ EXPERIMENTAL SECTION Materials. All chemicals were used as received, including 3(triethoxysilyl)propyl succinic anhydride (TESPSA) (95%, Gelest, Inc.), 3-aminopropyltrimethoxysilane (APTS) (99%, Aldrich), octadecyltrichlorosilane (OTS) (Gelest Inc.), and (heptadecafluoro-1,1,2,2,-tetrahydrodecyl)trichlorosilane (99%, Gelest, Inc.), namely, perperfluorosilane. Silica NPs with three different diameters (100 ( 3, 50 ( 3, and 20 ( 3 nm; 30 wt % in isopropanol) were purchased from Nissan Chemicals. Polyester fabric was provided by Bigsky Technologies (Pittsford, NY). Surface Functionalization of Silicon Substrates. The Si wafers were precleaned in a 1% solution (v/v) of Detergent 8 (a lowfoaming phosphate-free soap solution from Alconox) in deionized (DI) water at 65 °C for 1 h, followed by sonication in DI water, isopropanol, and acetone for 20 min, respectively. After being dried, the substrates were treated with oxygen plasma (30 W, 0.2 Torr, Harrick plasma cleaner PDC-001) for 60 min. The oxidized Si wafers were then silanized immediately by immersing them in a 0.01 M solution of TESPSA in ethanol for 18 h. The excess silane was removed by sonicating the substrates in ethanol and acetone for 30 min, followed by drying with compressed air. A 3.2 ( 1 nm layer thickness was measured by ellipsometry. Surface Functionalization of Silica Nanoparticles. The asreceived silica NPs were pelletized by centrifugation at 11 000 rpm overnight, followed by drying in vacuo for 3 h. The NPs were then functionalized with APTS by slightly modifying the procedure reported in the literature.36 In a typical experiment, an excess of a 5% (v/v) solution of APTS was added to the nanoparticle suspension in anhydrous ethanol (200 proof, from Decon) under a nitrogen atmosphere and allowed to react at 80 °C for 8 h (Figure 1). The APTS-functionalized nanoparticles (APTS-SiO2 NPs) were then purified via centrifugation at 11 000 rpm overnight, followed by drying in vacuo for 3 h. Dip Coating of APTS-SiO2 Nanoparticle Films and Their Hydrophobilization. TESPSA-silanized Si wafers were immersed in a well-dispersed ethanolic solution of APTS-SiO2 NPs at different concentrations for 10 s and withdrawn at a rate of 4 cm/min. For the dual-scale assembly, after the deposition of the first layer of the NP film (20 or 100 nm), the substrate was dried and coated with the second NP solution. To improve the adhesion between the NPs and the substrate and their mechanical robustness, we annealed the film at 200 or 400 °C for 2 h. The obtained nanoparticle films (both monoscale and dual scale) were subsequently treated with oxygen plasma (30 W, 0.2 Torr) for 60 s to introduce surface hydroxyl groups and placed in a desiccator containing 100 μL of perfluorosilane on a separate glass slide for vapor deposition in vacuo for variable time periods. The same procedure was employed to coat nanoparticles on polyester fabric except that the fabric was annealed at 160 °C for 3 h after coating the first layer of NPs. Characterization. Static and dynamic water contact angles were measured with a model 290 Rame-Hart standard automated goniometer. The static contact angle was measured from a 5.0 μL water droplet. Advancing and receding water contact angles were measured by automatically adding and removing water from the substrate, respectively. All contact angle values were averaged over three different spots on each 4595

dx.doi.org/10.1021/la104067c |Langmuir 2011, 27, 4594–4602

Langmuir

ARTICLE

Figure 2. SEM images of monoscaled APTS-SiO2 nanoparticles dip coated onto Si wafers. (a-c) 20 nm nanoparticles. (a) 0.1, (b) 0.5, and (c) 1.0 wt %. (d-f) 50 nm nanoparticles. (d) 0.1, (e) 0.5, and (f) 1.0 wt %. (g-i) 100 nm nanoparticles. (g) 0.1, (h) 0.5, and (i) 1.0 wt %. The 500 nm scale bar is applicable to all images. sample. For roll-off angle measurement, the substrate was placed on a custom-designed stage with a protractor attached to it, and a 10 μL water droplet was used. All roll-off angle values were averaged over three different measurements on each sample. The morphologies of the NP films, which were sputter-coated with gold, were imaged with an FEI Quanta 600 FEG environmental scanning electron microscopy (ESEM). The surface topography of the samples was imaged with a Dimension 3000 atomic force microscope from Digital Instruments with a Si3N4 cantilever in tapping mode. The rms roughness values were calculated from 5 μm  5 μm images using Nanoscope VII software. The optical transparency of the glass substrates was measured using a Varian UV-vis-NIR Cary 5000 spectrophotometer. The thickness of TESPSA and perfluorosilane deposited on a Si wafer was measured using a Rudolph Research AutoEL-II null ellipsometer, and the values were averaged over three different spots. The refractive index of TESPSA was assumed 1.5.

’ RESULTS AND DISCUSSION Surfaces with dual-scale roughness have been suggested essential to achieve superhydrophobicity. Recently, Ling et al.35 demonstrated superhydrophobicity from a monolayer of singlesized SiO2 NPs (60 nm in diameter) dip coated on glass from the aqueous solution (0.6 mg/mL, pH 7), followed by the vapor deposition of perperfluorosilane, 1H,1H,2H,2H-perfluorodecyltriethoxysilane. Because the NP size and coating thickness were much smaller than the wavelength of visible light, the resulting superhydrophobic surface was optically transparent. Although the fabrication procedure is quite simple, the results are in sharp contrast to the requirement of dual-scale roughness reported in

the literature. Furthermore, it was suggested that the random, moderately dense coverage of NPs was the main reason to warrant the superhydrophobicity and low hysteresis (the difference between advancing and receding contact angles). This is different from our observation of enhanced hydrophobicity, although not superhydrophobicity, from a monolayer of closepacked APTS-SiO2 NPs deposited on a swollen poly(styreneran-acrylic acid) film when the particle size is e100 nm.36 In addition, the approach reported by Ling et al. required a very high annealing temperature (1100 °C) to enhance the stability of the nanoparticle coating and a rather long duration (at least 5 h) to deposit perfluorosilane, which might generate additional roughness that contributes to superhydrophobicity. To address these concerns, we chose to use the same dipcoating approach to deposit amine-functionalized SiO2 NPs from their ethanol solution, a common solvent used in the dipcoating process because of its low boiling point, on TESPSAsilanized Si wafers for covalent bonding. The NP film was then vapor deposited with the same type of perfluorosilane (Figure S1) at variable time durations. Our focus here is to study the influence of particle size and coverage on hydrophobicity. To prevent agglomeration of the NPs in solution, which could lead to a nonuniform coating and irreproducible wetting results, we functionalized the SiO2 NPs with APTS (Figure 1), thus introducing positive charges on the particle surface. We employed three different sizes of APTS-SiO2 NPs (20, 50, and 100 nm in diameter) and varied their concentrations in ethanol (0.1-1.0 wt %) and the fluorosilane deposition time (0.5-10 h) to investigate whether any of these parameters would contribute 4596

dx.doi.org/10.1021/la104067c |Langmuir 2011, 27, 4594–4602

Langmuir

ARTICLE

Table 1. Water Contact Angles of Si Wafers Coated with Single-Sized APTS-SiO2 Nanoparticles with Variable Nanoparticle Concentrations and Fluorosilane Treatment Time measured water contact angle (deg) size of SiO2

conc of SiO2

fluorosilane deposition

NPs (nm)

NPs (wt %)

time (h)

20

50

100

Wenzel contact angle, θsta

θadv

θrec

roughness factor, r

θw(deg)

0.1

6.0

118.5 ( 3

129.4 ( 2

99.2 ( 4

1.14

115

0.5 1.0

0.5 6.0

148.8 ( 5 126.3 ( 1

161.4 ( 2 138.9 ( 2

120.3 ( 5 101.9 ( 2

1.76 1.26

140.5 118.2

0.1

6.0

111.9 ( 2

119.1 ( 2

104.5 ( 1

1.08

113.9

0.5

0.5

132.2 ( 2

150.8 ( 5

107.7 ( 5

1.39

127.5

1.0

6.0

135.7 ( 1

153.3 ( 1

132 ( 1

1.52

131.8

0.1

6.0

119.9 ( 1

133.1 ( 2

106.3 ( 2

1.16

118.4

0.5

0.5

128.1 ( 5

143.1 ( 5

108.9 ( 1

1.27

123.8

1.0

6.0

135.6 ( 1

147.4 ( 3

115.7 ( 2

1.52

131.9

to the superhydrophobicity. As seen in Figure 2, a monolayer of SiO2 NPs was randomly deposited on a Si wafer, and their coverage was varied depending on the NP size and concentration in ethanol. When the concentration was low, 0.1 wt %, NPs were sparsely deposited on the Si wafer. When the concentration was increased to 0.5 wt %, a dramatic increase in NP coverage was observed for all particle sizes, and the density of 20 and 50 nm NPs was similar to the observation reported by Ling et al.,35 which coated 60 nm SiO2 NPs at 0.6 mg/mL. Further increases in NP concentration with respect to dip coating did not seem to increase the NP coverage for these two particles. This suggests that access to succinic anhydride groups on Si wafers by aminefunctionalized SiO2 NPs (20 and 50 nm) reached its maximum at ∼0.5 wt % NP concentration, where excess NPs were rinsed off during dip coating. In contrast, the coverage of 100 nm NPs continued to increase until the concentration reached 1.0 wt %. The water contact angle values from monoscale NP films are summarized in Tables 1 and S1-S3. The static water contact angles (θsta) increased (up to 150°) with increased NP concentration and more or less leveled off when the NP concentration was greater than 0.5 wt %, which was consistent with the NP coverage observed in SEM images. However, the highly hydrophobic surfaces also presented large water contact angle hysteresis (θadv - θrec = 10-30°), thus the water droplets were immobile on substrates. The prolonged perfluorosilane treatment time did not seem to have a significant effect on hydrophobicity, and SEM and AFM images did not show any changes in surface roughness. On a rough surface, there are two distinct models by Wenzel37 and Cassie-Baxter38 to explain the surface roughness effect on the contact angle. In Wenzel’s model, roughness effectively increases the actual surface area. The apparent Wenzel contact angle, θw, on a rough surface is defined as cos θw ¼ rcos θ0

ð1Þ

where r is the roughness factor and is defined as the ratio of the actual surface area to the apparent surface area and θ0 is the equilibrium contact angle on a flat surface or the Young’s contact angle. In our experiments, r for monoscale NPs is calculated from AFM measurements (Table 1). If θ0 > 90°, then θw is increased by roughness. When the substrate is intrinsically hydrophilic (θ0 < 90°), the solid-liquid interaction is favored; the Wenzel

Figure 3. Schematic illustration of the coassembly of APTS-SiO2 nanoparticles of different sizes.

contact angle will be decreased by roughness, resulting in spontaneous spreading on the rough surface. In the CassieBaxter model, however, the surface is considered to be a composite surface of solid and air, which has an apparent Cassie contact angle, θc, defined as cos θc ¼ f ðcos θ0 þ 1Þ - 1

ð2Þ

where f is the fraction of the liquid-solid contact. According to eq 2, θc increases with decreasing f as more air is trapped between the grooves of the rough surface. As seen in Table 1, the measured static water contact angles, θsta, agree well with the calculated θw with an error of less than 8%, suggesting that the monoscale roughness is not sufficient to achieve highly mobile superhydrophobicity. We note that the roughness factor for higher water contact angles in our system is determined by two factors in the NP assembly: (1) the surface coverage of NPs and (2) the surface coverage of fluorosilane, which is deposited in between and on top of NPs. When the particle size is small (e.g., 20 nm) and the concentration of NPs is low, the surface coverage of NPs is also low. Therefore, more fluorosilane could be coated between the NPs. Over long deposition times (e.g., 6 h), the surface roughness generated by NPs can be smoothed out. This might not be the case when 4597

dx.doi.org/10.1021/la104067c |Langmuir 2011, 27, 4594–4602

Langmuir

ARTICLE

Table 2. Water Contact Angles of Si Wafers Coated with Two Differently Sized APTS-SiO2 Nanoparticles (100 nm First and 20 nm Second) with Variable Nanoparticle Concentrations and Fluorosilane Treatment Time measured water contact angle (deg) conc of 100 nm

conc of 20 nm

fluorosilane

roughness

theoretical contact

SiO2 NPs (wt %)

SiO2 NPs (wt %)

deposition time (h)

θsta

θadv

roll-off angle (deg)

factor, r

angle, θ (deg)

0.5 0.5

0.5 0.5

0.0 0.5

48.9 ( 2 162.3 ( 1

55.4 ( 5 168 ( 2

4(1

1.52 1.69

166.2

0.5

0.5

1.0

163.3 ( 2

170.1 ( 2

4( 1

1.65

165.9

0.8

0.5

0.5

151.4 ( 2

166.9 ( 2

6(1

1.59

158.6

0.8

0.5

1.0

150.2 ( 1

165.5 ( 3

7(1

1.6

158.7

1.0

0.5

0.5

151.3 ( 1

165.9 ( 2

7(2

1.55

157.1

1.0

0.5

1.0

150 ( 1

162.9 ( 5

10 ( 1

1.54

156.9

the particle size becomes larger as the spacing between NPs for fluorosilane deposition becomes smaller. Therefore, our intention here is to search for an optimum combination of NP size, concentration, and fluorosilane treatment time to achieve a larger roughness factor and higher water contact angles. To enhance the nonwettability of the NP surface, we then prepared dual-sized NP films by the successive dip coating of 20 and 100 nm APTS-SiO2 NPs, followed by perfluorosilane deposition as depicted in Figure 3. The ratio of smaller to larger particles was chosen to be less than 1/4 according to Ming et al.12 on the basis of free-energy modeling of dual-scale roughness and its contribution to the surface wettability. Two dip-coating sequences were investigated, that is, 20 nm first (Figure 3a) and 100 nm first (Figure 3b), to study the effect of dipping sequences on surface roughness and thus superhydrophobicity. First, we coated 100 nm NPs followed by 20 nm NPs (both 0.5 wt % in ethanol). The resulting water contact angle was 48.9 ( 2° (Table 2). After mild oxygen plasma treatment and perfluorosilane deposition, the static water contact angle was increased dramatically to >150° and exhibited superhydrophobic nonwetting behavior with an advancing contact angle >162° (Table 2). On such a surface, it was very difficult to measure the receding contact angle because the water droplet tended to stick to the goniometer needle rather than the substrates as explained by the Cassie-Baxter model.38 We varied the concentration of APTSSiO2 NPs and perfluorosilane treatment time separately to investigate the individual contribution of nanoroughness and chemistry to superhydrophobicity. As seen in Tables 2 and S4, a perfluorosilane treatment time of as little as 30 min was sufficient to produce superhydrophobicity, illustrating the self-cleaning property (Supporting Information Movie S1). To confirm the superhydrophobicity and quantify the flow resistance of the surface, we measured the water droplet roll-off angle, sometimes also called the sliding angle, which is defined as the tilt angle when the liquid drop starts to move on a surface.39 The substrate was placed on a custom-designed stage, and the water droplets (10 μL) rolled off easily at a very small tilt angle (150° with advancing contact angle values of >160° (Table 3). Different from the SEM images shown in Figure 4, 100 nm particles were found to sit on top of a nanocarpet of 20 nm particles, for which the coverage could be tuned by the respective nanoparticle concentration in ethanol for dip coating (Figure 5). After the deposition of perfluorosilane, the morphology of NP assemblies remained unchanged. The rms roughness of the dual-scale film (both from 0.5 wt %) was 30.4 nm as measured by AFM (Figure S2), corresponding to a 20% increase in surface area over the smooth surface. We varied the concentration of 20 nm NPs from 0.5 to 1.0 wt % before the deposition of 100 nm nanoparticles. In all cases, the coated surfaces were found to be superhydrophobic. Although the 20 nm NP film alone did not provide high nanoroughness, subsequent deposition of 100 nm NPs

Figure 5. SEM images of dual-sized APTS-SiO2 nanoparticles successively dip coated onto Si wafers with 20 and 100 nm nanoparticles at different concentrations. (a, b) Deposition of 20 nm SiO2 nanoparticles (0.5 wt %) followed by 100 nm SiO2 nanoparticles (0.5 wt %) (a) without and (b) with fluorosilane treatment for 1 h. (c, d) Deposition of 20 nm SiO2 nanoparticles (0.5 wt %) followed by 100 nm SiO2 nanoparticles (0.8 wt %) (c) without and (d) with fluorosilane treatment for 1 h. Scale bars are 500 nm.

enhanced the fraction of the air pocket at the composite interface (Cassie-Baxter state) and thus the nonwettability. We hypothesize that this was because after 20 nm NP deposition, the surface was no longer smooth, therefore affecting the subsequent deposition of 100 nm particles that appeared to be less dense than the monoscale 100 nm NPs deposited on a flat Si wafer. Overall, the surface roughness was increased from 4.2% (20 nm only) to 20% (100 on 20 nm NPs), which was also larger than that of 100 nm NPs alone (10%) at the same dip coating concentration. The study also suggests that it is not necessary to have a smaller nanoroughness on top of a larger-scale rough surface, as many have reported, to generate superhydrophobicity as long as a composite surface is rough enough to prevent water penetration between the grooves. To improve the adhesion of NP films to the substrate and their mechanical stability, we used two annealing temperatures, 4599

dx.doi.org/10.1021/la104067c |Langmuir 2011, 27, 4594–4602

Langmuir

ARTICLE

Figure 6. Optical transparency of the dual-sized nanoparticle film. (a) Optical photograph of water droplets (7 μL) on glass coated with 100 nm SiO2 nanoparticles (0.5 wt %), followed by the deposition of 20 nm SiO2 nanoparticles (0.5 wt %) and fluorosilane treatment for 3 h. (b) UV-vis spectra of superhydrophobic nanoparticle films dip coated in different sequences and compared to bare glass.

Figure 7. Optical photographs of (a) 5 μL water droplets (mixed with malachite green dye) and (b) 10 μL paraffin oil droplets on polyester fabric (left) with and (right) without dual-sized nanoparticles.

200 and 400 °C, each for 2 h, before perfluorosilane deposition. NP films annealed at 200 °C were stable against solvent rinsing and sonication. However, most particles came off after molding to the poly(dimethylsiloxane) (PDMS) film. In comparison, NP films annealed at 400 °C remained on the Si wafer against PDMS molding. As seen in Table S4, this annealing treatment (400 °C for 2 h) did not change the apparent water contact angle or the rms roughness. The film remained superhydrophobic when left on the laboratory bench for a month or under continuous UV exposure (200 mW/cm2) for a week. One distinct advantage of our superhydrophobic films obtained from the coassembly of dual-sized NPs with diameter e100 nm is the high optical transparency. As seen in Figure 6a, when the nanoparticle film was coated onto a glass substrate, the letters underneath the glass were very visible. The UV-vis spectra showed that the superhydrophobic film maintained its optically transparency (relative transmission >99%) in the visible wavelength with respect to the bare glass (Figure 6b). For particles much smaller than the wavelength of the scattered light, the Mie scattering is rather weak. Meanwhile, Rayleigh scattering is applied when the radius of the scattering particle is much smaller than the wavelength of the incident light. According to Raleigh’s theory, the Raleigh scattering intensity is given by   6! 1 d I=I0 µ R2 λ4

ð4Þ

where I0 is the incident light intensity, d is the diameter of the particle, λ is the wavelength of the incident light, and R is the distance between the particle and the detector, typically a few centimeters or more. In a simple calculation, we assume that the diameter of the particle on a dual-scale surface is ∼120 nm and the wavelength of the incident beam is 532 nm. The Raleigh scattering is   1 ð5Þ I=I0 µ ð37:28 nm2 Þ 2 R Clearly, the contribution of Raleigh scattering in the visible region is negligible. Finally, to demonstrate the versatility of our coating approach in creating superhydrophobic surfaces, we dip coated the dualsized nanoparticles onto various substrates, including epoxy and fabrics in addition to Si wafers and glass substrates. Specifically, on a polyester fabric, we dip coated 20 nm particles first, followed by 100 nm particles and perfluorosilane deposition for 1 h. The NP film was annealed at 160 °C for 3 h before perfluorosilane treatment to improve the adhesion to the fabric. As shown in Figure 7a, the nontreated fabric surface was very hydrophilic, where water containing malachite green dye completely wetted the fabric, leaving a stain. The fabric became superhydrophobic after being coated with NPs and surface hydrophobilization, as clearly evident in the left image in Figure 7a, where the water droplets bead up. In household applications, it will be ideal to have both water- and oil-repellent fabrics. Designing such a 4600

dx.doi.org/10.1021/la104067c |Langmuir 2011, 27, 4594–4602

Langmuir superoleophobic surface is more challenging than designing a superhydrophilic surface because the surface tension (γ) of organic liquids is much smaller than that of water (72.4 mN/m). It has been suggested that re-entrant structures with metastable wetting is necessary to achieve superoleophobicity.41 As a first attempt, we applied a droplet of paraffin oil (γ = 48 mN/m) to the polyester fabric, and an oil contact angle of >110° was obtained (Figure 7b). In comparison, the oil completely wetted an untreated fabric. A more detailed study will be reported elsewhere.

’ CONCLUSIONS We demonstrated a simple yet efficient approach to preparing superhydrophobic surfaces on various substrates, including Si, glass, epoxy, and fabrics, by sequentially dip coating nanoparticles of two different sizes, followed by the deposition of perfluorosilane. The surface exhibited a self-cleaning property with a water droplet (10 μL) roll-off angle of less than 5°. We found that the dual-scale nanoroughness was essential to achieving superhydrophobicity on randomly non-close-packed NPs. When the substrate was dip coated with monoscale nanoparticles (20, 50, and 100 nm), a very hydrophobic surface was obtained and the degree of hydrophobicity was dependent on the particle size and surface coverage. Because of the small particle size, the glass coated with dual-scale nanoparticles (20 and 100 nm) showed nearly 100% transmission in the visible region, similar to that of the noncoated glass. Furthermore, we showed that thermal annealing enhanced the stability of the nanoparticle film such that superhydrophobicity was maintained against prolonged exposure to UV light under ambient conditions. Understanding the size and coverage of nanoroughness with respect to surface hydrophobicity will help to improve the fabrication of optically transparent superhydrophobic surfaces. ’ ASSOCIATED CONTENT

bS

Supporting Information. Schematic description of the creation of monoscale roughness by nanoparticle assembly. Water contact angle values of monoscale NP films. Annealing treatment effect (400 °C for 2 h) on the water contact angle and rms roughness after the deposition of the first layer of SiO2 NPs. AFM images of dual-scale NP films deposited in different sequences. Videos of the self-cleaning effect on the superhydrophobic surface and the 5 μL water droplet rolling off of the surface. This material is available free of charge via the Internet at http://pubs.acs.org.

’ AUTHOR INFORMATION Corresponding Author

*E-mail: [email protected].

’ ACKNOWLEDGMENT The work was supported by NSF CAREER award DMR0548070. We thank Xuelian Zhu for help with calculating the Raleigh scattering values of the nanoparticle films. The PENN Regional Nanotechnology Facility (PRNF) and the Nano/Bio Interface Center (NBIC) are acknowledged for access to SEM and AFM instruments. We thank Bigsky Technologies for providing the polyester fabric.

ARTICLE

’ REFERENCES (1) Barthlott, W.; Neinhuis, C. Planta 1997, 202, 1. (2) Gao, X. F.; Jiang, L. Nature 2004, 432, 36. € (3) Oner, D.; McCarthy, T. J. Langmuir 2000, 16, 7777. (4) Wagner, T.; Neinhuis, C.; Barthlott, W. Acta Zool. 1996, 77, 213. (5) Autumn, K.; Liang, Y. A.; Hsieh, S. T.; Zesch, W.; Chan, W. P.; Kenny, T. W.; Fearing, R.; Full, R. J. Nature 2000, 405, 681. (6) Parker, A. R.; Lawrence, C. R. Nature 2001, 414, 33. (7) Extrand, C. W. Langmuir 2005, 21, 10370. (8) Gao, L. C.; McCarthy, T. J. Langmuir 2006, 22, 2966. (9) Onda, T.; Shibuichi, S.; Satoh, N.; Tsujii, K. Langmuir 1996, 12, 2125. (10) Herminghaus, S. Europhys. Lett. 2000, 52, 165. (11) Shirtcliffe, N. J.; McHale, G.; Newton, M. I.; Chabrol, G.; Perry, C. C. Adv. Mater. 2004, 16, 1929. (12) Ming, W.; Wu, D.; van Benthem, R.; de With, G. Nano Lett. 2005, 5, 2298. (13) Li, X. M.; Reinhoudt, D.; Crego-Calama, M. Chem. Soc. Rev. 2007, 36, 1350. (14) Lim, J.-M.; Yi, G.-R.; Moon, J. H.; Heo, C.-J.; Yang, S.-M. Langmuir 2007, 23, 7981. (15) Puukilainen, E.; Rasilainen, T.; Suvanto, M.; Pakkanen, T. A. Langmuir 2007, 23, 7263. (16) Bravo, J.; Zhai, L.; Wu, Z.; Cohen, R. E.; Rubner, M. F. Langmuir 2007, 23, 7293. (17) Dorrer, C.; Ruhe, J. Adv. Mater. 2008, 20, 159. (18) Levkin, P. A.; Svec, F.; Frechet, J. M. J. Adv. Funct. Mater. 2009, 19, 1993. (19) Qian, Z.; Zhang, Z. C.; Song, L. Y.; Liu, H. R. J. Mater. Chem. 2009, 19, 1297. (20) Im, M.; Im, H.; Lee, J. H.; Yoon, J. B.; Choi, Y. K. Soft Matter 2010, 6, 1401. (21) Furstner, R.; Barthlott, W.; Neinhuis, C.; Walzel, P. Langmuir 2005, 21, 956. (22) Zhang, Y.; Lo, C. W.; Taylor, J. A.; Yang, S. Langmuir 2006, 22, 8595. (23) Nosonovsky, M.; Bhushan, B. Nano Lett. 2007, 7, 2633. (24) Liu, H.; Feng, L.; Zhai, J.; Jiang, L.; Zhu, D. B. Langmuir 2004, 20, 5659. (25) Bok, H. M.; Kim, S.; Yoo, S. H.; Kim, S. K.; Park, S. Langmuir 2008, 24, 4168. (26) Leng, B. X.; Shao, Z. Z.; de With, G.; Ming, W. H. Langmuir 2009, 25, 2456. (27) Li, X. M.; He, T.; Crego-Calama, M.; Reinhoudt, D. N. Langmuir 2008, 24, 8008. (28) Lin, P. C.; Yang, S. Soft Matter 2009, 5, 1011. (29) St€ober, W.; Fink, A.; Bohn, E. J. Colloid Interface Sci. 1968, 26, 62. (30) Artus, G. R. J.; Jung, S.; Zimmermann, J.; Gautschi, H. P.; Marquardt, K.; Seeger, S. Adv. Mater. 2006, 18, 2758. (31) Kim, M.; Kim, K.; Lee, N. Y.; Shin, K.; Kim, Y. S. Chem. Commun. 2007, 2237. (32) Zimmermann, J.; Artus, G. R. J.; Seeger, S. J. Adhes. Sci. Technol. 2008, 22, 251. (33) Li, G. X.; Liu, Y.; Wang, B.; Song, X. M.; Li, E.; Yan, H. Appl. Surf. Sci. 2008, 254, 5299. (34) Latthe, S. S.; Imai, H.; Ganesan, V.; Kappenstein, C.; Rao, A. V. J. Sol Gel Sci. Technol. 2009, 53, 208. (35) Ling, X. Y.; Phang, I. Y.; Vancso, G. J.; Huskens, J.; Reinhoudt, D. N. Langmuir 2009, 25, 3260. (36) McConnell, M. D.; Yang, S.; Composto, R. J. Macromolecules 2009, 42, 517. (37) Wenzel, R. N. Ind. Eng. Chem. 1936, 28, 988. (38) Cassie, A.; Baxter, S. Trans. Faraday Soc. 1944, 40, 546. (39) Dettre, R. H.; Johnson, R. E. J. Colloid Interface Sci. 1969, 31, 568. 4601

dx.doi.org/10.1021/la104067c |Langmuir 2011, 27, 4594–4602

Langmuir

ARTICLE

(40) Jeong, H. E.; Lee, S. H.; Kim, J. K.; Suh, K. Y. Langmuir 2006, 22, 1640. (41) Tuteja, A.; Choi, W.; Ma, M. L.; Mabry, J. M.; Mazzella, S. A.; Rutledge, G. C.; McKinley, G. H.; Cohen, R. E. Science 2007, 318, 1618.

’ NOTE ADDED AFTER ASAP PUBLICATION This article was published ASAP on February 28, 2011. Figure S1 in the Supporting Information has been modified. The correct version was published on March 21, 2011.

4602

dx.doi.org/10.1021/la104067c |Langmuir 2011, 27, 4594–4602