Anisotropic Wet Etched Silicon Substrates for Reoriented and

LETTER pubs.acs.org/Langmuir

Anisotropic Wet Etched Silicon Substrates for Reoriented and Selective Growth of ZnO Nanowires and Enhanced Hydrophobicity Sisi Li,† Jie Hu,† Junjun Li,† Jinghua Tian,† Zhitao Han,† Xiongtu Zhou,† and Yong Chen*,†,‡ † ‡

Ecole Normale Superieure, CNRS-ENS-UPMC 8640, 24 rue Lhomond, 75005 Paris, France Institute for Integrated Cell-Material Sciences, Kyoto University, 606-8507 Kyoto, Japan ABSTRACT: Herein we report the fabrication of ZnO nanowires on anisotropic wet etched silicon substrates by selective hydrothermal growth. Æ100æ oriented silicon wafers were first patterned by anisotropic wet etch with a KOH solution, resulting in V-shaped stripes of different periods. Then, a thin layer of gold was deposited and annealed to promote the hydrothermal growth of ZnO nanowires. It was found that the growth rate of ZnO nanowires on Æ111æ surfaces was much higher than that on Æ100æ surfaces. As a first application of such micro- and nanostructured surfaces, we show enhanced wetting properties by measuring the contact angle of water droplets on the samples obtained under different patterning and growth conditions. Our results also demonstrated the possibility of tuning the contact angle of the sample in the range between 115 and 155, by changing either the pattern of the silicon template or the hydrothermal growth conditions.

1. INTRODUCTION During the past decade, superhydrophobic surfaces have aroused much research interest due to their potential applications in selfcleaning coating,1 microfluidics,2 biochips,3 and other biocompatible materials.4 In general, hydrophobic properties can be enhanced by increasing surface roughness5,6 or lowing surface energy.7
18 By using conventional microfabrication techniques such as photolithography, nanoimprint lithography, reactive ion etching, and soft lithography, regular pillars3,19
23 could be obtained by design, showing superhydrophobic properties after surface coating of nonwetting agent. Other surface methods24 such as chemical,25 electrodeposition,26 nanotube synthesis,27 phase separation,17 as well as electrospinning28 and nonwoven assembling of nanofibers29 could be used to produce a large variety of superhydrophobic properties and so forth. Although the most observed wetting properties including contact angle hysteresis, drop rolling angle, and sticking behavior of condensed water drops could be understood using Wenzel and Cassie models,30 a number of questions remain to be answered. Among them, the exact role of dual scaled surface roughness is of particular interest since the hierarchic and micro- and nanoscaled organization can be found in many natural systems. Such a coexistence of dual scaled features is believed to contribute significantly to the quality of the superhydrophobicity.31
38 On lotus leaves, for example, bumps at the scale of 10
50 μm are covered by fine nanostructures at the scale of 100 nm.39 On rice leaves, the dual scale feature is anisotropic so that water can flow preferentially in one direction. The hierarchic surface texture can be found in animals, but the detailed organization r 2011 American Chemical Society

is in general different from one species to another. To mimic the dual scale features of natural surfaces, the easiest approach is to combine a conventional microfabrication and surface chemical synthesis techniques. In this work, we report the fabrication of novel dual scale surfaces by anisotropic wet etching of a patterned silicon wafer (top down) and hydrothermal growth (bottom up) of reoriented ZnO nanowires (NWs). Silicon is the most commonly used material for microstructure patterning, whereas ZnO NWs can be easily grown with many interesting properties for optical, optoelectronic, and piezoelectric applications. The combination of the top-down and bottom up approaches allows us to incorporate dual scale patterns in a controllable way. By using conventional photolithography and anisotropic wet etch of a silicon wafer, we obtained V-shaped line features with both Æ100æ plane and Æ111æ sidewall surfaces (54.7 oriented). By hydrothermal synthesis with an intermediate layer of gold as catalysis, high density ZnO NWs could be obtained on Æ111æ sidewall surfaces, in contrast to the growth of much less NWs on Æ100æ surfaces. Such a selective and reoriented growth allowed us to investigate enhanced wetting properties of ZnO NWs. Our results showed a clear parametric dependence of the observed wetting properties on both template patterning and growth conditions. Received: March 29, 2011 Revised: April 22, 2011 Published: May 03, 2011 6549

dx.doi.org/10.1021/la201157m | Langmuir 2011, 27, 6549–6553

Langmuir

LETTER

Figure 1. (a) Schematic of fabrication flowchart of V-shaped silicon template and ZnO nanowires on both Æ100æ and Æ111æ surfaces. (1) Deposition of 50 nm Cr on Æ100æ silicon surface; (2) photolithography of microlines; (3) removal of unprotected Cr; (4) KOH etch of silicon wafer; (5) deposition of 10 nm Cr and 50 nm Au; (6) growth of ZnO nanowires by hydrothermal synthesis. (b) SEM image of a template obtained by anisotropic wet etching of a Æ100æ silicon wafer. (c) SEM image of ZnO NWs grown on Æ111æ surfaces of the silicon template.

2. EXPERIMENTAL SECTION Figure 1a shows a schematic of the fabrication process flow. First, a Æ100æ oriented silicon wafer was cleaned in acetone, ethanol, isopropyl alcohol (IPA), and deionized (DI) water consecutively, each for 10 min. Then, it was dried with nitrogen gas and baked at 200 C for 5 min on a hot plate to eliminate any absorbed moisture. A 50 nm thick Cr layer was deposited on the surface of the silicon wafer using an e-beam evaporator (Edward Auto 500). Afterward, a thin layer of AZ5214E photoresist was spin-coated on the substrate and then exposed with a UV light and a photomask. After development with AZ726MIF developer, the substrate was submerged in chrome etchant (chrome-etch 3144, Honeywell) for Cr etching. Finally, the remaining photoresist was removed by acetone, resulting in Cr features for wet etch of the under layer silicon substrate. To ensure the wet etch performance, the sample was treated by reactive ion etching (RIE) to remove the thin layer of native silicon dioxide. A fresh KOH solution with 30 wt % KOH, 62.5 wt % DI water and 7.5 wt % IPA was prepared and warmed to 80 C in a water bath. Then, the sample was immersed in the KOH solution. Typically, the etch rate of Æ100æ oriented silicon wafer under such conditions is about 1 μm/ min. Once the etching depth was reached, the silicon wafer was rinsed in DI water and the Cr pattern was removed. Before hydrothermal growth, a 10 nm thick Cr layer and a 50 nm Au layer were sequentially deposited on the silicon template with the e-beam evaporator. Here, the Cr layer was used to enhance the adhesion of the Au layer on silicon substrate and the Au layer was used as an “intermediate layer” to assist the growth of ZnO NWs.40 Finally, the sample was annealed at 300 C for 1 h allowing improvement of the crystalline structure of the gold layer, which is critical to the oriented growth of aligned ZnO NWs. ZnO NWs were grown on the patterned silicon substrate in a solution composed of 25 mM zinc nitrate hexahydrate (Zn(NO3)2 3 6H2O, 98%, Sigma-Aldrich), 12.5 mM hexamethylenetetramine (HMTA, C6H12N4, g99.5%, Sigma-Aldrich), and 0.35 M ammonium hydroxide (NH3 3 H2O, 28.0
30.0%, Sigma-Aldrich). The sample was put face down in the solution to avoid deposition of any precipitates from the solution. Growth of ZnO NWs was conducted in a universal oven (loading model 100-800, Memmert) at 80 C. Finally, the samples were rinsed with DI water and dried in

Figure 2. (a,b) ZnO NWs grown on flat surfaces of a Æ111æ and Æ100æ oriented silicon wafers. (c,d) SEM images of ZnO NWs grown on V-shaped trenches for 22 h; (e,f) ZnO NWs grown on the same type of substrate with a much longer growth time, where (f) is taken from the Æ100æ area. air after the growth. The morphology of both etched silicon substrate and ZnO NWs grown on different silicon substrates was characterized by scanning electron microscopy (SEM) operated at 10 KV (Hitachi S-800). In order to study the wetting properties of micro- and nanostructured surfaces, a water droplet of 2 μL was deposited on the surface of the sample and the contact angle was recorded with a dedicated experimental setup. The deposition of the 2 μL water droplet was controlled with a digital syringe pump, and the image was taken with a CCD camera equipped with an objective for 30 magnification. 6550

dx.doi.org/10.1021/la201157m |Langmuir 2011, 27, 6549–6553

Langmuir

3. RESULTS AND DISCUSSION 3.1. Growth on V-Shaped Silicon Substrate. The mechanism of ZnO NW growth in a solution was recently explained by the propagation of axial screw dislocation on at the tip of the NWs.41,42 In this work, we were mainly interested in the growth selectivity on etched silicon facets of different orientations. Figure 1b shows a SEM image of a typical V-shaped pattern, which was obtained by selective wet etching of a Æ100æ silicon substrate with a Cr mask of gratings of 2 μm line width and 72 μm period. Figure 1c shows a SEM image of the corresponding ZnO

Figure 3. SEM images of ZnO NWs grown on V-shaped (a) and flat Æ111æ oriented (b) silicon template. Insets are photos of 2 μL water droplets deposited on the surface of the two substrates, showing a contact angle of 155 (a0 ) and 123 (b0 ) respectively.

LETTER

NWs after 22 h growth. As can be seen, the length of the fabricated ZnO NWs can be as large as 20 μm, keeping a diameter of about 1 μm. Clearly, the density of the ZnO NWs on the Æ111æ surface (54.7 oriented) is much higher than that on the Æ100æ surface. Moreover, the ZnO NWs of Æ111æ oriented facets are much more regular and vertical than that on Æ100æ surfaces. Such remarkable differences are originated from different crystalline surfaces of the silicon, deserving a more detailed investigation as discussed below. 3.2. Substrate Orientation Dependences. To verify the growth dependence of the ZnO NWs growth on different substrates, we used two native silicon wafers without etching, one Æ111æ oriented and another Æ100æ oriented. Under the same Cr/Au coating and growth conditions, the two samples resulted in completely different ZnO NWs. High density, vertical, and ultralong ZnO NWs arrays were obtained on Æ111æ surfaces (Figure 2a), whereas much less and disoriented ZnO NWs could be observed on Æ100æ surfaces (Figure 2b). The density of ZnO NWs produced by hydrothermal synthesis is essentially dependent on the number of nuclei formed at the beginning of the growth, and it was believed that the arrival of more ions on the substrate may not initiate new nuclei at a later stage.40,43 It is also known that the atomic density of the silicon Æ111æ surface is higher than that of the Æ100æ surface, which would favor the c-axis oriented ZnO NW growth. The hydrothermal synthesis of our ZnO NWs was assisted by an intermediate gold layer which was relatively thick (50 nm) and annealed at 300 C.

Figure 4. SEM images of ZnO NWs grown on silicon templates of V-shaped stripes of different periods but the same growth condition (a
c) or different growth times but the same stripe period (d
f). Insets are photos of 2 μL water droplets deposited on the surfaces of the corresponding substrates, showing different contact angles, as reported in (g) and (h). 6551

dx.doi.org/10.1021/la201157m |Langmuir 2011, 27, 6549–6553

Langmuir Besides, this gold layer was added to the sample after coating a thin layer of Cr on the silicon substrate to improve the gold adhesion. Such a process for sample preparation complicates the interpretation of observed growth selectivity. In the past, a large number of investigations have been shown that atomically flat and large terraces could be formed on the surface of gold layers deposited on Æ111æ silicon substrates.44 Such atomically flat terraces should be helpful for the growth of high density and vertical ZnO NWs. Figure 2c and d shows another example of ZnO NWs grown on Æ111æ oriented facets of a V-shaped silicon wafer for 22 h. Here, the growth mainly took placed in the 25 μm deep trenches obtained with a silicon template of 45 μm line-and-width gratings. Again, only few and disoriented ZnO NWs could be found in the top Æ100æ areas, but high density and well-oriented ones fully filled the template trenches. Increasing the growth time further led to the change of the growth regime on both silicon Æ111æ and Æ100æ surfaces. Indeed, we observed well populated ZnO NWs on both Æ111æ and Æ100æ surfaces only after a few hours of more incubation, as shown in Figure 2 e and f. 3.3. Enhanced Superhydrophobicity. It is known that the wettability of a surface can be enhanced by increasing the surface roughness. In our case, both flat and etched silicon substrates are hydrophilic. After ZnO NWs growth, the flat sample becomes hydrophobic, whereas the etched sample could become either hydrophobic or superhydrophobic, depending on both the pattern geometry and the growth conditions. Figure 3 shows a comparison between the etched and nonetched samples, both patterned with high density ZnO NWs. As can be seen, the etched sample resulted in a wavy structure of ZnO NWs with hierarchic organization of dual features and a contact angle of about 155. In contract, the ZnO NWs grown on flat Æ111æ silicon shows a contact angle of only about 123. These results unambiguously demonstrate the change of wetting property of as-grown ZnO NWs from a hydrophobic to a superhydrophobic state by changing the morphology of the template. Additional experiments were performed by varying either the grating period or the growth time. Figure 4a
c shows ZnO NWs after 22 h growth on three etched silicon substrates of the same line width (2 μm) and depth (15 μm) but different periods (144, 72, and 36 μm). The measured contact angles were 143, 151, and 155, indicating that reducing the grating period of the etched substrate led to an increased surface roughness so that its hydrophobicity (Figure 4g). Similarly, we observed an increase of the surface hydrophobicity of as grown ZnO NWs with the increase of growth time. Figure 4e
g shows the results obtained with the same type of etched substrates (V-shaped features of 72 μm period, 2 μm line width, and 15 μm depth) after 5, 12, and 22 h growth. The measured contact angles were 115, 126, and 151, respectively as reported in Figure 4h. The observed increases in surface hydrophobicity are function of both pattered micro and nano features. It is known that in either Wenzel or Cassie state, the contact angle of a droplet increases with the increase of the surface roughness but they have different drop rolling angles. In the range of our investigation, the drop rolling angle on all ZnO NWs surfaces remain small (