4310
Langmuir 2007, 23, 4310-4314
Design and Fabrication of Micro-textures for Inducing a Superhydrophobic Behavior on Hydrophilic Materials Liangliang Cao,† Hsin-Hua Hu,‡ and Di Gao*,† Department of Chemical and Petroleum Engineering, Department of Mechanical Engineering, UniVersity of Pittsburgh, Pittsburgh, PennsylVania 15261 ReceiVed December 10, 2006. In Final Form: February 13, 2007
Artificial superhydrophobic surfaces are typically fabricated by tuning the surface roughness of intrinsically hydrophobic surfaces. We report here the design and fabrication of micro-textures for inducing a superhydrophobic behavior on hydrogen-terminated Si surfaces with an intrinsic water contact angle of ∼74°. The micro-textures consist of overhang structures with well-defined geometries fabricated by microfabrication technologies, which provide positions to support the liquid and prevent the liquid from entering into the indents between the micro-textures. As a result, water is in contact with a composite surface of solid and air, which induces the observed macroscopic superhydrophobic behavior.
Introduction Many plant leaves in nature, notably lotus leaves, have superhydrophobic surfaces.1 Water on these leaves beads up with a contact angle of greater than 150° and drips off rapidly when the leaves are slightly inclined, while collecting powderlike contaminations. This phenomenon has been explained by the lotus effect, which is a combined effect of the epicuticular wax-induced hydrophobicity and the surface roughness resulting from the hierarchical structures found on the leaves.1-3 This observation has stimulated extensive research to fabricate artificial superhydrophobic surfaces with self-cleaning properties, either by creating a rough surface on a hydrophobic material or by modifying a rough surface with hydrophobic coatings.2-6 Both approaches require a surface with a low enough surface free energy and hence an intrinsic water contact angle (the water contact angle when the surface is flat, θflat) of greater than 90° to further enhance the hydrophobicity by tailoring the surface roughness. Processes for obtaining a low surface free energy typically involve modifying the surfaces with organic chemical coatings. However, the hydrophobicity of chemically treated surfaces deteriorates over time, which brings major problems to the artificial superhydrophobic surfacessaging and decay.3 A question of significant interest is whether one is able to produce a superhydrophobic surface using materials with a θflat value of less than 90° because this implies the possibility of making intrinsically hydrophilic materials superhydrophobic. Nature provides a positive answer to this question by the fact that some leaves with an absence of hydrophobic waxes also * Corresponding author. E-mail:
[email protected]. Tel: (412) 6248488. Fax: (412) 624-9639. † Department of Chemical and Petroleum Engineering. ‡ Department of Mechanical Engineering. (1) Neinhuis, C.; Barthlott, W. Ann. Bot. (London) 1997, 79, 667-677. (2) Feng, L.; Li, S. H.; Li, Y. S.; Li, H. J.; Zhang, L. J.; Zhai, J.; Song, Y. L.; Liu, B. Q.; Jiang, L.; Zhu, D. B. AdV. Mater. 2002, 14, 1857-1860. (3) Blossey, R. Nat. Mater. 2003, 2, 301-306. (4) Erbil, H. Y.; Demirel, A. L.; Avci, Y.; Mert, O. Science 2003, 299, 13771380. (5) Furstner, R.; Barthlott, W.; Neinhuis, C.; Walzel, P. Langmuir 2005, 21, 956-961. (6) Nakajima, A.; Hashimoto, K.; Watanabe, T. Monatsh. Chem. 2001, 32, 31-41.
exhibit superhydrophobicity.1 Recent research has also found that the wax on lotus leaves has a θflat value of ∼74°, in contrast to the expected contact angle of greater than 90°.7 In this paper, we report the design and fabrication of microtextures for inducing a superhydrophobic behavior on hydrogenterminated (H-terminated) silicon surfaces with a θflat value of ∼74°. The micro-textures consist of overhangs with well-defined geometries fabricated by micro- and nanofabrication technologies, which provide positions to support the liquid and prevent the liquid from entering into the indents between the textures. As a result, water is in contact with a composite surface of solid and air, which induces the observed macroscopic superhydrophobic behavior.
Design of Micro-textures There are two possible states when water contacts a rough solid surface: (i) the water is in complete contact with the solid surface (Figure 1a) or (ii) the water contacts a composite surface of solid and air and forms droplets (known as fakir droplets, Figure 1b). The relationship between apparent water contact angle on a rough surface (θrough) and its intrinsic water contact angle (θflat) has been described by the Wenzel equation for the first state8
cos θrough ) r cos θflat
(1)
and by the Cassie-Baxter equation for the second state9
cos θrough ) φs cos θflat - (1 - φs)
(2)
where r is the roughness factor, defined as the ratio of the actual surface area to the apparent one (the projection on a flat surface), and φs is the area fraction of the solid surface that contacts water. The two states have been referred to as the Wenzel state and Cassie state, respectively. The relationship between cos θrough and cos θflat for these two states is plotted in Figure 1 according to eqs 1 and 2, as previously demonstrated.10 The two lines (7) Cheng, Y. T.; Rodak, D. E. Appl. Phys. Lett. 2005, 86, 144101. (8) Wenzel, R. N. Ind. Eng. Chem. 1936, 28, 988-994. (9) Cassie, A. B. D.; Baxter, S. Trans. Faraday Soc. 1944, 40, 546-551. (10) Que´re´, D. Rep. Prog. Phys. 2005, 68, 2495-2532.
10.1021/la063572r CCC: $37.00 © 2007 American Chemical Society Published on Web 03/20/2007
Superhydrophobic BehaVior on Hydrophilic Materials
Langmuir, Vol. 23, No. 8, 2007 4311
Figure 1. Relationship of cos θrough with cos θflat. The black solid, blue solid, red dotted, and red dashed lines correspond to the Wenzel state, the Cassie state, the metastable Cassie state when θflat > 90°, and the metastable Cassie state when θflat < 90°, respectively. θc is the critical intrinsic contact angle. Inset a: schematic representation of Wenzel stateswater is in complete contact with solid. Inset b: schematic representation of Cassie stateswater is in contact with a composite surface of solid and air. Inset c: schematic representation of metastable Cassie statesthis state is energetically unfavorable as compared to the Wenzel state for the same θflat; however, water may still contact a composite surface of solid and air due to an energy barrier provided by the capillary force that prevents water from entering into the indents. In the case of θflat < 90°, overhanging structures are needed.
corresponding to the two states intersect at one critical angle θcrit
θcrit ) cos-1
φs - 1 r - φs
(3)
Because φs < 1 and r > 1, θcrit > 90°. It has been proven that the energy of the system decreases monotonically when cos θrough increases.11 Therefore, when water contacts a solid surface with θflat < θc, the Wenzel state is energetically more favorable (with smaller θrough and greater cos θrough) and hence should be preferred by the system from an energetic point of view. However, fakir droplets in the Cassie state have also been observed on such surfaces consisting of high-density textures when θc > θflat > 90°, which is energetically metastable. The presence and potential importance of the metastable Cassie state has been presented previously.12-15 Such a metastable Cassie state is represented by the dotted line as an extension of the solid line representing the Cassie state in Figure 1.10 A question that remains is if the metastable Cassie state exists when θflat < 90° (i.e., cos θflat > 0). This state is represented by the dashed line as an extension of the Cassie state into the fourth quadrant in Figure 1. Although the existence of such a state has been predicted16 and occasionally observed,17 to the best of our knowledge, it has not been demonstrated by experiments with designed surfaces. In this paper, we show the existence of the metastable Cassie state on silicon surfaces with a well-defined topography and a θflat value of ∼74°. A necessary condition for observing the metastable Cassie state is the existence of an energy barrier that separates the Wenzel (11) Bico, J.; Thiele, U.; Que´re´, D. Colloids Surf., A 2002, 206, 41-46. (12) Lafuma, A.; Que´re´, D. Nat. Mater. 2003, 2, 457-460. (13) Marmur, A. Langmuir 2004, 20, 3517-3519. (14) He, B.; Patankar, N. A; Lee, J. Langmuir 2003, 19, 4999-5003. (15) Ishino, C.; Okumura, K.; Que´re´, D. Europhys. Lett. 2004, 68, 419-425. (16) Herminghaus, S. Europhys. Lett. 2000, 52, 165-170. (17) Love, J. C.; Gates, B. D.; Wolfe, D. B.; Paul, K. E.; Whitesides, G. M. Nano Lett. 2002, 2, 891-894.
Figure 2. Three cross-sectional profiles of water in contact with a solid indent consisting of overhanging structures. A critical parameter for these different profiles is the angle (θoverhang) formed between the sidewalls of the indent and the horizontal line. (a) When θoverhang is greater than the intrinsic water contact angle (θflat), the water-air interface (meniscus) inside the indent is concave when viewed from the water side. The net force (Fs) generated by the meniscus at the water-air interface is toward the inside of the indent, which causes water to enter into the indent and have complete contact with the solid. (b) When θoverhang ) θflat, the water-air interface is flat and stays at a circular intersection of the indent. (c) When θoverhang < θflat, the meniscus inside the indent is convex. Fs is toward the outside of the indent, which prevents water from entering into the indent.
state and the Cassie state. This energy barrier may be provided by a capillary force that prevents water from entering into the indents present on the surface. In the case of θflat < 90°, overhanging structures are needed to provide positions for the water-solid contact line to suspend water above the bottom surface. Figure 2 shows three possible cross-sectional profiles for water in contact with a solid indent consisting of overhangs. A critical parameter for these different profiles is the angle (θoverhang) formed between the sidewalls of the indent and the horizontal line. When θoverhang is greater than θflat of the sidewall surface (Figure 2a), the water-air interface (meniscus) inside the indent is concave when viewed from the water side. The net force generated by the meniscus (Fs) is toward the inside of the indent, causing water to enter into the indent and have complete contact with the solid. When θoverhang ) θflat (Figure 2b), the water-air interface is flat and stays at a circular intersection of the indent, assuming that the gravity is ignored. When θoverhang < θflat (Figure 2c), the water-air interface inside the indent is convex, and Fs is toward the outside of the indent, which prevents water from entering into the indent. Therefore, the Cassie state is only possible if θoverhang is smaller than θflat. Experimental Procedures We fabricated micro-textures consisting of overhanging structures with well-defined geometries in both the micrometer and nanometer range. Figure 3 schematically shows the process flow to fabricate the micrometer-sized Si pillars with overhangs at the top edges. We started with a Si(100) wafer coated with a 0.5 µm thermally grown silicon dioxide (SiO2) thin film. The SiO2 film was patterned by photolithography, followed by wet etching in HF aqueous solution. A Bosch deep reactive ion etching (DRIE) process using inductively coupled plasma was employed to etch the SiO2-masked Si substrate, which formed arrays of Si pillars with vertical sidewalls. By aligning the square patterns of the Si pillars with the Si(100) wafer flat, the Si(110) surfaces were exposed on the four sidewalls of each pillar after this process. The exposed Si(110) sidewalls were then etched in a potassium hydroxide (KOH) aqueous solution (30% by weight) at 75 °C for 3 min. Etching of Si by KOH was anisotropic, which exposed the (111) surfaces on the sidewalls of the Si pillars. The exposed Si(111) surface underneath the top SiO2 mask formed overhanging sidewalls at an angle (θ) of 54.7° to the top Si(100) surfaces. The SiO2 mask was removed by HF afterward, leaving a hydrogen-terminated Si surface. Figure 4 schematically shows the top view of the design for the Si pillars. The pattern is a two-dimensional array of squares. The critical parameters are the width of each square (a) and the centerto-center distance between adjacent squares (x). In the mask design
4312 Langmuir, Vol. 23, No. 8, 2007
Cao et al. Table 1. Design Parameters for Si Pillar Arrays Used in Our Experiment width of each square: a (µm)
Figure 3. Schematic process flow to fabricate the micrometersized Si pillars with overhangs at the top edges. (a) The process began with a (100) Si substrate with a 0.5 µm thick SiO2 film. (b) The SiO2 film was patterned by photolithography, followed by wet etching in HF aqueous solution. (c) A deep reactive ion etching (DRIE) process was employed to etch the SiO2-masked Si substrate, which formed arrays of Si pillars with vertical sidewalls. The Si(110) surfaces were exposed on the four sidewalls of each pillar after this process. (d) The exposed Si(110) sidewalls were anisotropically etched in a KOH aqueous solution, which exposed (111) surfaces on the sidewalls of the Si pillars. (e) The SiO2 mask was removed by HF, leaving a hydrogen-terminated Si surface. The overhang angle (θ) formed between the Si(111) sidewall and the Si(100) top surface is 54.7°.
center-to-center distance between adjacent squares: x (µm)
on mask
actual
on mask
actual
φs ≈ (a/x)actual2
10 10 10 10 10 10
5 5 5 5 5 5
13 15 18 20 25 30
13 15 18 20 25 30
0.148 0.111 0.077 0.063 0.040 0.028
with four Si(111) sidewalls, each of which formed a 54.7° angle with the bottom Si(100) surface. The SiO2 was removed by HF afterward. An SEM image of the fabricated Si islands is shown in Figure 6a. Au nanoclusters were deposited onto the exposed Si surfaces in an aqueous solution of HF and KAuCl4. The concentrations of HF and KAuCl4 were 0.2 and 0.01 M, respectively. The deposition time was 30 s. Si nanowires were grown by a chemical vapor deposition process.18,19 Silicon tetrachloride (SiCl4, Aldrich, 99.99%) was used as the precursor. Hydrogen gas (10% in argon by volume) was used as the carrier gas. The nanowires were grown at 850 °C via a vapor-liquid-solid mechanism.18 The Si nanowires synthesized by this method had a strong preferred growth direction along the 〈111〉 axis and therefore were vertical to the exposed Si(111) sidewalls. An SEM image taken after the growth of Si nanowires is shown in Figure 6b. The Au clusters at the tips of the nanowires were removed by an aqueous solution of HCl and HNO3, followed by treatment with HF. The sample was thoroughly rinsed in deionized water and dried in air before the contact angle was measured. All the SEM images were taken by a Philips XL-30 field emission SEM setup. The lotus leaf was purchased from a local Chinese grocery store. The leaf was dried and coated with a ∼3 nm thin gold film by sputtering before the SEM images were taken. The static water contact angles were measured according to the sessile droplet method10 using a drop shape analysis system (AST Products, Inc.) with a computer-controlled liquid dispensing system. Water droplets with a volume of 8 µL were used. The experiment was performed under normal laboratory ambient conditions, 20 °C and 40% relative humidity. The contact angles were measured 5 times on each sample.
Results and Discussion
Figure 4. Design schematics for the Si pillar pattern. The critical design parameters are the width of each square (a) and the centerto-center distance between adjacent squares (x). for photolithography, a is fixed at 10 µm, and x is varied from 13 to 30 µm. When the pattern of the photoresistor was transferred to the underneath SiO2 layer, the size of the squares decreased due to the isotropic etching of SiO2 by the HF aqueous solution. The actual size of the squares was determined from the scanning electron microscopy (SEM) images taken afterward. The center-to-center distance of the squares remained unchanged during the pattern transfer. These parameters and the corresponding φs (the area fraction of the solid surface in contact with water) are listed in Table 1. The entire array was 1 cm wide and 1 cm long. Figure 5 shows the schematic process flow for fabricating the structures consisting of Si nanowires on Si islands. We started with a Si(100) substrate coated with a thin SiO2 film. After the SiO2 film was patterned by photolithography and wet etching, the sample was etched by KOH. The anisotropic etch by KOH formed Si islands
The micrometer-sized Si pillars with overhangs at the top edges fabricated by micromachining techniques are shown in Figure 7a,b. The top surfaces of the Si pillars were 5 µm × 5 µm squares and 15 µm above the bottom surfaces. The overhanging Si(111) sidewalls formed an angle (θoverhang) of 54.7° with the top Si(100) surfaces (Figure 3e). We fabricated a series of such Si pillar arrays by varying the center-to-center distance (d) between adjacent pillars from 13 to 50 µm. The θflat value of the hydrogen-terminated Si surface was measured to be 74 ( 3, which was consistent with previous reports.20 Therefore, these Si pillars provide us with an ideal system to study wetting properties of surfaces consisting of overhanging structures with θoverhang < θflat < 90°. The static water contact angles of these surfaces were measured (Figure 7c) and plotted as a function of φs in Figure 7d. It was observed that θrough increased from ∼139 to ∼162° as φs decreased from ∼0.15 to ∼0.04 until d reached ∼30 µm. When d was further increased, the measured contact angles were scattered and unrepeatable, owing to the tendency of drops to fall in the texture for such large gaps. When d was less than 30 µm, the relationship between θrough and φs agreed well with the CassieBaxter equation (Figure 7d). (18) Westwater, J.; Gosain, D. P.; Tomiya, S.; Usui, S.; Ruda, H. J. Vac. Sci. Technol., B 1997, 15, 554.
Superhydrophobic BehaVior on Hydrophilic Materials
Langmuir, Vol. 23, No. 8, 2007 4313
Figure 5. Schematic process flow to fabricate the structures consisting of Si nanowires on Si islands. (a) The process began with a Si(100) substrate coated with a 0.5 µm thick SiO2 film. (b) The SiO2 film was patterned by photolithography and wet etching. (c) The anisotropic etch by KOH formed Si islands with four Si(111) sidewalls, each of which formed a 54.7° angle with the bottom Si(100) surface. (d) The SiO2 was removed by HF. (e) Au nanoclusters were deposited onto the exposed Si surfaces. (f) Si nanowires were grown by a chemical vapor deposition process. The Si nanowires synthesized by this method had a strong preferred growth direction along the 〈111〉 axis and therefore were vertical to the exposed Si(111) sidewalls. (g) The Au clusters at the tips of the nanowires were removed. The angle (θ) formed between the nanowires grown vertically to the Si(111) sidewalls and the Si(100) horizontal surface is 35.3°.
Figure 6. SEM images. (a) Si islands fabricated by etching a SiO2-masked Si(100) substrate in KOH. (b) Si nanowires grown on the Si islands with Au clusters on the tips of the nanowires. The scale bars are 5 µm.
The fabricated surfaces with overhanging structures by growing Si nanowires on micromachined Si islands are shown in Figure 8a. The top surfaces of the Si islands were 5 µm × 5 µm squares with a 15 µm center-to-center distance. The height of the islands was ∼8 µm. The Si(111) sidewall of the islands formed a 54.7° angle with the bottom Si(100) surface. The Si nanowires were ∼3 µm long with varied widths ranging from 100 to 500 nm. The Si nanowires grown vertically on the sidewalls of two adjacent Si islands faced each other and formed an overhanging structure with a θoverhang value of 35.3° (Figure 8b). The hierarchical structures consisting of Si nanowires on Si islands provide another model system that satisfies the condition of θoverhang < θflat < 90°. The static water contact angle on the surface shown in Figure 8a was measured to be ∼160°. Superhydrophobicity has been observed on plant leaves where hydrophobic wax is absent.1 Recent experiments7 have also found that θflat of the wax on the superhydrophobic surfaces of lotus (19) Gao, D.; He, R.; Carraro, C.; Howe, R. T.; Yang, P.; Maboudian, R. J. Am. Chem. Soc. 2005, 127, 4574-4575. (20) Houston, M. R.; Maboudian, R. J. Appl. Phys. 1995, 78, 3801-3808.
leaves is approximately 74°, which is contrary to the expected values of greater than 90°. These observations cannot be understood by the previous explanation of the lotus effect (i.e., a wax-induced hydrophobicity (θflat > 90°) enhanced by the surface roughness). The results obtained in our experiments provide a plausible explanation of these phenomena. The hierarchical structures on the surface of the lotus leaves consist of microscale bumps and nanoscale hair-like protrusions (Figure 8c). They form a similar surface topography (Figure 8d) as the surfaces constructed using Si nanowires and Si islands. The θflat value of the carnauba wax on the surface of the lotus leaves is approximately 74°,7 which is also similar to the θflat value of the hydrogen-terminated Si surface. Therefore, the superhydrophobicity of these lotus leaves may be induced by the same mechanism that has been suggested to explain the superhydrophobicity of the hierarchical Si nanowire structuressthe overhanging structures formed by the nanometer-sized protrusions cause water to contact a composite surface of solid and air in a metastable Cassie state, resulting in an apparent superhydrophobicity on surfaces with a θflat value of less than 90°.
4314 Langmuir, Vol. 23, No. 8, 2007
Cao et al.
in this paper are superhydrophobic to water droplets during the contact angle measurement, they are completely wetted after immersion in water. The superhydrophobicity recovers after they are dried in air. This reversible process can be repeated multiple times. Similar phenomena have been observed on lotus leaves.7,16 One explanation of these phenomena may be proposed from our experiments as the following. The system is in a metastable Cassie state with θflat 1, θrough will be less than θflat according to eq 1. This typically leads to the observed hydrophilic surfaces with a small θrough value due to the large roughness factors (r) on these surfaces. Figure 7. Superhydrophobic surfaces consisting of Si pillars with overhangs at the top edges. (a) SEM image of the Si pillars. (b) A close-up SEM image of panel a. (c) An optical image of a water droplet on the Si pillars during the contact angle measurement. (d) Static water contact angles measured as a function of φs. The solid line is plotted according to the Cassie-Baxter relation (eq 2).
Figure 8. Superhydrophobic surfaces with hierarchical structures. (a) SEM image of Si nanowire arrays grown on micrometer-sized Si islands. (b) A schematic cross-sectional profile of panel a. The Si nanowires grow vertically to the Si(111) sidewalls and therefore form an angle (θ) of 35.3° with the horizontal Si(100) surface. (c) SEM image of the surface of a lotus leaf. (d) A schematic crosssectional profile of panel c. The hierarchical structure on the surface of the lotus leaf consists of microscale bumps and nanoscale hairlike protrusions. The scale bars in panels a and c are 5 µm.
Another interesting phenomenon we have observed is that although all the surfaces with overhanging structures presented
Conclusion In summary, we have fabricated micro-textures consisting of overhanging structures of Si with well-defined geometries in both micrometer and nanometer scales. Although the intrinsic water contact angle of the H-terminated Si surface is about 74°, the constructed surfaces induce superhydrophobic behavior. This phenomenon and the fact that these surfaces are completely wetted when immersed in water resemble what has been observed on lotus leaves. Plausible explanations for these phenomena are provided based on our experiments and analysis. Certainly, problems such as the quantitative evaluation of the stability of the metastable Cassie state, the hydraulic pressure required to wet the surfaces fabricated in our work, and the transition between the Cassie and the Wenzel states in terms of contact angle hysteresis still need to be studied further. Nonetheless, it is expected that this paper could provide additional insight into the interesting superhydrophobic phenomena found in nature and that the principle could be applied to make intrinsically hydrophilic materials superhydrophobic by tailoring the surface topography. Acknowledgment. This work was partially supported by the National Science Foundation and the University of Pittsburgh Mascaro Sustainability Initiative. LA063572R
