Wetting and Wetting Transitions on Copper-Based ... - ACS Publications


Wetting and Wetting Transitions on Copper-Based...

0 downloads 95 Views 577KB Size

Langmuir 2005, 21, 937-943

937

Wetting and Wetting Transitions on Copper-Based Super-Hydrophobic Surfaces N. J. Shirtcliffe,* G. McHale, M. I. Newton, and C. C. Perry School of Biomedical and Natural Sciences, The Nottingham Trent University, Clifton Lane, Nottingham NG11 8NS, United Kingdom Received June 3, 2004. In Final Form: October 29, 2004 Rough and patterned copper surfaces were produced using etching and, separately, using electrodeposition. In both of these approaches the roughness can be varied in a controlled manner and, when hydrophobized, these surfaces show contact angles that increase with increasing roughness to above 160°. We show transitions from a Wenzel mode, whereby the liquid follows the contours of the copper surface, to a CassieBaxter mode, whereby the liquid bridges between features on the surface. Measured contact angles on etched samples could be modeled quantitatively to within a few degrees by the Wenzel and Cassie-Baxter equations. The contact angle hysteresis on these surfaces initially increased and then decreased as the contact angle increased. The maximum occurred at a surface area where the equilibrium contact angle would suggest that a substantial proportion of the surface area was bridged.

Introduction Super- or ultra-hydrophobicity is currently the focus of considerable research.1-5 Drops of water on smooth and flat hydrophobic surfaces do not usually form angles to the solid surface of greater than 120°. The addition of roughness to the surface can increase the contact angle without altering the surface chemistry. When the angle achieved exceeds 150° this is usually termed superhydrophobicity, although other definitions of superhydrophobicity have been suggested.6 Some of these surfaces can be made so water repellent that drops will not come to rest on them and simply roll off if the sample is tilted even slightly. The rolling drop has been observed to remove contamination, and such super-hydrophobic surfaces with low contact angle hysteresis are referred to as self-cleaning. This type of surface has obvious potential uses, as water will not “stick” to it. Super-hydrophobicity is used by some plants, such as Nelumbo nucifera (L.) druce and is known as the Lotus Effect.7 Super-hydrophobic surfaces have been produced using a range of chemical and physical methods. In 1996, Onda et al.8,9 demonstrated a contact angle of 174° on a surface obtained using the crystallization of a paper-sizing agent, alkylketene dimer, to provide fractal roughness. Other recent methods include using hydrophobized glass beads,10 the sol-gel process to generate a porous and, hence, rough * Corresponding author. E-mail: [email protected]. Tel.: +44 (0)115 8486375. Fax: +44 (0)115 9486636. (1) Dambacher, G. Kunstst.-Plast. Eur. 2002, 92 (6), A18. (2) Lau, K.; Bico, J.; Teo, K.; Chhowalla, M.; Amaratunga, G.; Milne, W.; McKinley, G.; Gleason, K. Nano Lett. 2003, 3 (12), 1701-1705. (3) Nun, E.; Oles, M.; Schleich, B. Macromol. Symp. 2002, 187, 677682. (4) Rossbach, V.; Patanathabutr, P.; Wichitwechkarn, J. Fibers Polym. 2003, 4 (1), 8-14. (5) Shirtcliffe, N.; McHale, G.; Newton, M.; Perry, C. Langmuir 2003, 19 (14), 5626-5631. (6) Oner, D.; McCarthy, T. Langmuir 2000, 16, 7777-7782. (7) Barthlott, W.; Neinhuis, C. Planta 1997, 202 (1), 1-8. (8) Onda, T.; Shibuichi, S.; Satoh, N.; Tsujii, K. Langmuir 1996, 12, 2125-2127. (9) Shibuichi, S.; Yamamoto, T.; Onda, T.; Tsujii, K. J. Phys. Chem. 1996, 100, 19512-19517. (10) Fuji, M.; Fujimori, H.; Takei, T.; Watanabe, T.; Chikazawa, M. J. Phys. Chem. B 1998, 102, 10498-10504.

surface,11,12 vacuum-deposited (poly)tetrafluoroethylene thin films,13,14 anodic oxidation of aluminum surfaces,9 and plasma polymerization,15,16 among others. In many of these cases the idea has been to create a high aspect ratio topographic surface and to then apply a thin (∼monolayer) hydrophobic coating. Typically the hydrophobic coating provides a contact angle of 115-120° on a flat surface, and the topography then enhances this to 150-180°. Much of the recent research is materials oriented with attempts to create hard, transparent, and self-cleaning surfaces.17,18 To describe super-hydrophobic surfaces two models known as “Wenzel” and “Cassie-Baxter” are used.19,20 Both models can be understood using minimization of the surface free energy: small drops of water on surfaces form equilibrium shapes that minimize changes in the total surface energy resulting from the solid-air, solid-liquid, and water-air interfaces.21-23 Roughening a solid surface alters the relative contribution of the interfaces involving the solid. The main difference between the two models is whether the liquid drop retains intimate contact with the solid surface at all points or whether the liquid bridges across surface protrusions, thus, resulting in a drop suspended on a composite solid and vapor surface. In the Wenzel model,19 intimate contact is maintained with the solid at all points below the drop and the observed (11) Tadanaga, K.; Kitamuro, K.; Matsuda, A.; Minami, T. J. Sol.Gel Sci. Technol. 2003, 26 (1-3), 705-708. (12) Tadanaga, K.; Morinaga, J.; Minami, T. J. Sol.-Gel Sci. Technol. 2000, 19, 211-214. (13) Miller, J.; Veeramasuneni, J.; Drelich, J.; Yalamanchili, M. Polym. Eng. Sci. 1996, 36 (14), 1849-1855. (14) Palumbo, G.; Agostino, R. D.; Lamendola, R.; Corzani, I.; Favia, P. European Patents EP0985741 and EP0985740, 2000 (Procter and Gamble). (15) Coulson, S.; Woodward, I.; Badyal, J.; Brewer, S. A.; Willis, C. J. Phys. Chem. B 2000, 104 (37), 8836-8840. (16) Shirtcliffe, N.; Thiemann, P.; Stratmann, M.; Grundmeier, G. Surf. Coat. Technol. 2001, 142, 1121-1128. (17) Takeda, K.; Sasaki, M.; Kieda, N.; Katayama, K.; Kako, T.; Hashimoto, K.; Watanabe, T.; Nakajima, A. J. Mater. Sci. Lett. 2001, 20 (23), 2131-2133. (18) Nakajima, A.; Abe, K.; Hashimoto, K.; Watanabe, T. Thin Solid Films 2000, 376 (1-2), 140-143. (19) Wenzel, R. Ind. Eng. Chem. 1936, 28, 988. (20) Que´re´, D. Physica A 2002, 313 (1-2), 32-46. (21) De Gennes, P. Rev. Mod. Phys. 1985, 57, 827-863. (22) Le´ger, L.; Joanny, J. Rep. Prog. Phys. 1992, 55, 431-486. (23) McHale, G.; Newton, M. Colloids Surf., A 2002, 206, 193-201.

10.1021/la048630s CCC: $30.25 © 2005 American Chemical Society Published on Web 01/07/2005

938

Langmuir, Vol. 21, No. 3, 2005

Shirtcliffe et al.

equilibrium contact angle is then the contact angle on the rough surface, θr, described by

cos θr ) r cos θs

(1)

where r is the specific surface area of the rough surface and θs is the contact angle on a smooth surface of the same chemical nature. In the Cassie-Baxter model the liquid drop suspends itself across surface protrusions and an average of the cosines of the angle on the solid (i.e., cos θs) and on the air (i.e., cos 180° ) -1) below the drop is used. If f is the fraction of the solid surface upon which the drop sits and 1 - f is the fraction below the drop that is air, then the Cassie-Baxter equation, eq 2,20 applies.

cos θr ) f cos θs - (1 - f)

(2)

One important difference between the Wenzel and Cassie-Baxter models is that the ease of drop motion is expected to differ, due to the extent to which the solid surface is contacted by the droplet. Que´re´ et al.24 refer to the Wenzel state as “sticky” and the Cassie-Baxter state as “slippery”. Miwa et al.25 have shown that low tilt angles for sliding, less than 1° for a 7-mg drop, can be obtained with the Cassie-Baxter state. Que´re´ et al. have studied the dynamics of interaction of droplets on super-hydrophobic surfaces. Their work includes drops rolling down surfaces (due to the surface interactions the drop rolls faster than a “marble” rolling under gravity),26 solids with a super-hydrophobic coating bouncing on water,27 and drops rebounding from surfaces.28,29 The classification of specific super-hydrophobic surfaces into these two states is possibly an oversimplification, and quantitative experimental investigation may help to clarify the situation. In this paper, we describe two simple methods by which super-hydrophobic surfaces can be produced inexpensively using copper to form the base material and a readily available commercial coating to hydrophobize it. The methods are complementary, with one involving the removal of material and the other the addition of material. The methods allow either roughness to be the controlling factor in the wetting or a combination of roughness and surface patterning. The surfaces are used to illustrate the effects of Wenzel to Cassie-Baxter transitions and demonstrate how the interplay between surface area and aspect ratio influences the static contact angles. Methods Copper Etching Method. A copper sheet (99+%, 0.127-mm thick) was obtained from Aldrich and used as the sample base and counter electrode. Disks (22-mm diameter) were placed into a 25-mm die (Graseby) and subjected to 10 tons of pressure in a press (Graseby) for 10 min. This treatment removed some of the ridges on the surface of the copper produced during manufacture and removed any bowing due to cutting. S18-13 photoresist (Shipley) was spun onto some of the copper on glass samples using a spin coater (Electronic Micro Systems 4000) at 3000 rpm, prebaked and exposed using 80 J cm-2 of ultraviolet (I-line) in a Cobilt C-800 mask aligner. The mask used was made up of tessellating squares with a circle of onehalf the side length in one corner being open (Figure 1). As S1813 is a positive resist, after developing this produced a layer of (24) Quere´, D.; Lafuma, A.; Bico, J. Nanotechnology 2003, 14 (10), 1109-1112. (25) Miwa, M.; Nakajima, A.; Fujishima, A.; Hasimoto, K.; Watanabe, T. Langmuir 2000, 16, 5754-5760. (26) Richard, D.; Que´re´, D. Europhys. Lett. 1999, 48 (3), 286-291. (27) Aussillous, P.; Que´re´, D. Nature 2001, 411 (6840), 924-927. (28) Richard, D.; Que´re´, D. Europhys. Lett. 2000, 50 (6), 769-775. (29) Richard, D.; Clanet, C.; Que´re´, D. Nature 2002, 417 (6891), 811811.

Figure 1. Scheme showing patterning of copper samples. resist with a pattern of circular holes in it. At this spinning speed the S18-13 has a final thickness of 1.6 µm; it was not hard-baked. Samples were masked off using clear nail varnish (Top Coat, Rimmel), and the back and edges of the samples were coated, leaving a 10 × 10 mm square in the center where the S18-13 pattern was located. Etching was carried out in a potassium persulfate solution. To approximate copper board etching, solutions of 125 g of potassium persulfate (Fisons, 98+%) were dissolved in 500 mL of deionized water to produce an approximately 0.9 M solution. This was heated to 40 °C and stirred on a hotplate; samples were held in the rapidly stirred solution using plastic tweezers. They were removed and rinsed with water periodically so that they could be checked using a microscope. The etch pits in the copper developed as expected for an anisotropic etch, starting off as flat depressions and proceeding evenly in all directions, evolving toward hemispherical shapes while undermining the photoresist. Eventually the etch pits overlapped one another, leaving diamond-shaped pillars connected by low cols. The etching agent used, and indeed most etchants, tends to produce rough surfaces as grain boundaries etch faster than the grains. Over the depth of etch used the small-scale roughness appeared to reach its maximum. In some samples chemical polishing was carried out using nitric acid (70% Fisher) and sulfuric acid (98% Fisher) mixed 1:3 just before use; fresh solution was used each time to ensure reproducibility. Samples were first etched using potassium persulfate and then chemically polished for 3 s, which was sufficient to chemically polish a flat copper sheet pre-etched in potassium persulfate in the same manner. After etching the samples were dropped into 0.1 M potassium hydrogen carbonate (97+% Fisher) to halt etching. Samples were then rinsed in deionized water, the photoresist and nail varnish were removed using acetone (Fisher p.a.), and a final rinse in ethanol (Fisher p.a.) was performed before the samples were dried in a stream of nitrogen. Copper Electrodeposition Method. Copper was electrodeposited onto clean copper samples that were prepared from the copper sheet in the same manner as those for etching. The copper samples were not patterned with photoresist but were masked with nail varnish to leave a 12-mm square section in the center and a corner bare. Electrical contact was made to the bare corner of the samples. Copper growth was carried out using 1.25 M copper sulfate (copper II sulfate hydrate 98%, Aldrich) in 0.26 M sulfuric acid (SG 1.84, 98+%, Fisher). A total of 50 mL of this solution was used in a small beaker, and the solution was changed after five samples to avoid contamination. Industrially these baths can be recycled for long periods, so the likelihood of bath aging effects was low. A piece of copper of at least four times the surface area of the sample was used as the anode and placed 20 mm from the sample. Plating was carried out after 30 s of equilibration to allow the oxide coating on the electrodes to be

Wetting on Cu-Based Super-Hydrophobic Surfaces

Langmuir, Vol. 21, No. 3, 2005 939

Figure 2. Confocal profiles of etched copper surfaces: (a) 40-µm pattern, (b) 30-µm pattern, (c) 20-µm pattern, and (d) 5-µm pattern. removed by the acidic plating bath. Copper was allowed to deposit until around 36 C cm-2 of charge had passed. This would be equivalent to around 13 µm of dense deposit if the conversion were perfect. Coating. Hydrophobization of both etched and electrodeposited samples was carried out using a wash in solution designed for waterproofing breathable fabrics (Extreme Wash In, Grangers). These preparations consist of fluorocarbons emulsified in water with a detergent. The agent was diluted with deionized water by a factor of 50, and the samples were immersed at room temperature for 20 min. They were then removed and gently rinsed in deionized water before being blown dry with nitrogen. The surfactants in the coating were removed by heating the samples to 40 °C in a drying oven for 20 h. This treatment was found to coat this particular type of sample evenly, as far as we could detect by electron microscopy. Temperatures were kept low to limit oxide growth. Sample Characterization. The surface areas of the electrodeposited samples were measured by lead underpotential deposition.30 Etched samples could not be produced with sufficiently even patterns to allow measurement in this manner. Samples were connected to a potentiostat and placed into acidic lead perchlorate [4 mM PbClO4 (Aldrich, 98% as hydrate), 0.2 M LiClO4 (Fisons), 0.32 M HClO4 (Acros ACS 70%)]. They were held at -129 mV with respect to Ag/AgCl in 3 M KCl for 1000 s before measurement. Voltammograms were then taken from -129 mV to -431 mV at 20 mV s-1: the negative peak was integrated and the charge per unit area was assumed to be 300 µC cm-2. Confocal microscopy, using a Leica DMRBE (TCS SP) with a He/Ne laser and Fluotar objective, was used to investigate the roughness of the materials on a micrometer scale. These measurements produced a three-dimensional map of the surfaces. Scanning electron microscope measurements were also used to investigate the surfaces; a JEOL JSM-840A scanning electron microscope was used with an acceleration voltage of 10 kV. Samples were viewed from perpendicular to the surface and at a 45° angle. Equilibrium contact angle measurements were made using a Kru¨ss DSA10 using 5-µL drops of distilled deionized water dropped onto the sample from a hydrophobized needle on a microsyringe. The needle usually had to be tapped to get the drop to detach. A picture of the drop was taken a few seconds after deposition to avoid any problems related to drying of the (30) Siegenthaler, H.; Ju¨ttner, J. J. Electroanal. Chem. 1984, 163, 327-343.

drop. Tangent measurements were made on the profiles of the drops, and three images (six angles) were taken to allow removal of the occasional rogue point, caused by contamination of the surface. Advancing and receding angle measurement, the maximum and minimum contact angles possible on a drop without the contact line moving, were made using two techniques; different techniques were necessary due to the nature of “slippery” and “sticky” super-hydrophobic surfaces. The advancing and receding angles are commonly measured by increasing and decreasing the volume of a drop sitting on the surface or by tilting a sample with a drop on it until the drop begins to slide.31-33 On a drop about to slide, the angle on the lower edge is the advancing angle, whereas that on the upper edge is the receding angle. Both techniques were implemented carefully to ensure that the angles measured were as high and low as possible without moving the contact line; the drops were filmed while tilting the surfaces, and the contact angles in the frame just before movement occurred were recorded. For this work, samples with low and medium receding angles were measured using the drop filling method and 10-µL drops. This method is very effective for measuring advancing contact angles, but problems occur when receding angles are very high. In this case the drops tend to detach from the surface when the volume is reduced, as they stick more strongly to the needle than to the surface. Very large drops could be used to prevent this from happening, but the weight of the drop can cause it to infiltrate patterned surfaces, thus, forcing a Cassie-Baxter to Wenzel transition. In addition, on patterned surfaces, increasing or decreasing drop volume causes motion with a quantized stickslip step characteristic and this creates uncertainties in the measurement of advancing and receding contact angles. For drops with high receding contact angles the sliding angles method was used with 15-µL drops on a tilting table. The tilting method is applicable when small drops can be tipped off the surface by inclining it; this occurred for all samples measured here that could not be measured by varying the volume of the drop. There is some debate in the literature as to whether the (31) Good, R.; Stromberg, R. Surface and Colloid Science; Plenum Press: New York, 1979; Vol. 11. (32) McDougall, G.; Ockrent, C. Proc. R. Soc. London, Ser. A 1942, 180, 151. (33) Adamson, A. Physical Chemistry of Surfaces, 5th ed.; Wiley: New York, 1990. (34) Herminghaus, S. Europhys. Lett. 2000, 52 (2), 165-170. (35) Shirtcliffe, N.; McHale, G.; Newton, M.; Chabrol, G.; Perry, C. Adv. Mater. 2003, 16 (21), 1929-1932.

940

Langmuir, Vol. 21, No. 3, 2005

Shirtcliffe et al.

Figure 3. Electron micrographs of etched copper samples with spacings of 30 µm viewed at 45°: (a) chemically polished; (b) rough etched top and pits; and (c) small pillars produced by longer etching.

Figure 4. Dual roughness with the larger-scale roughness being bridged by the drop and the smaller-scale roughness either wetted (a) or bridged (b). two methods are compatible, but it proved impossible to accurately measure advancing and receding angles on all of the surfaces prepared for this report using just one technique; where measurements are presented the method used in each case is indicated.

Results and Discussion Etched Copper Patterns. Figure 2 shows confocal microscope image reconstructions of patterns produced using various sized masks; the pattern size refers to the diameter of the circle on the mask, and the center of each circle is spaced at twice its diameter from the center of the next. The patterns are geometrically identical to one another except that the scale of the roughness induced by the etchant did not vary so the scales of the two roughnesses varied relative to one another. The equilibrium contact angle of water observed on all of these samples was approximately 152 ( 3°. Assuming that the flat tops of the pattern are in contact with the water and the rest of the pattern is not, Cassie and Baxter’s equation (eq 2) should apply. The tops of the pattern in contact with the drop will have a fractional area equivalent to that of a unit square minus a circle with unit diameter, which gives a value for the solid fraction, f, of (4 - π)/4. Putting this into eq 2 predicts a contact angle of 151°, which agrees very well with the observed value and suggests that water is indeed excluded from the etch pits. When a sample produced using a 30-µm pattern was etch-polished after persulfate etching (Figure 3a), the contact angle of water on it after hydrophobization was lowered to around 142 ( 3°. This lowering of the contact angle suggests that the drop is in contact with parts of the sides of the pits. If we assume total penetration into the pits and model the surface as touching hemispherical depressions in a square array, the solid surface area will be increased by a factor of (π/4) + 1 over a flat surface. Putting this into eq 1 as the roughness factor r then predicts a contact angle of 139°. The slightly higher value measured on the real sample is unsurprising, given that the chemically polished surfaces are not totally flat. An additional roughness factor of 1.05 over the whole surface

would be required to produce the contact angle observed using Wenzel’s equation, which is not unreasonable. The agreement between these values suggests that water contacted the entire surface of the sample under these conditions. When a sample produced using the same mask was stripped of photoresist after etching and then the tops of the pattern were roughened by etching for 30 s in the same persulfate solution, the contact angle of water on the pattern after hydrophobization increased to 155 ( 3°. If two levels of roughness are combined on a single surface, the contact angle on the surface can be calculated from the combined roughness. The Supporting Information to this paper provides first principles derivations based on surface free energy arguments that show that if the contact angle for the smaller-scale roughness is considered first; this angle can be inserted into Cassie-Baxter’s or Wenzel’s equation using the larger-scale roughness, and this result will be the same as that considering both at once. For clarity of definitions a summary of the results of the derivations are presented below. From the earlier measurements we expect the larger scale of roughness on these surfaces to be bridged by the water and to follow Cassie and Baxter’s equation. Considering the two possible extremes, the smaller-scale roughness could be wetted, leading to the situation depicted in Figure 4a combining Wenzel-type and CassieBaxter hydrophobicity, or the smaller-scale roughness could be bridged, leading to the case depicted in Figure 4b with Cassie-Baxter hydrophobicity on two scales. Equation 3 for Wenzel on Cassie-Baxter and eq 4 for dual Cassie-Baxter were derived from the combined roughnesses in these cases:

cos θ ) rfPA cos θA + (1 - fPA)

(3)

where fPA is the fractional area of the larger pattern, when projected onto a plane (i.e., fPA ) ∆APA/∆AP) and r is the roughness factor of the smaller-scale roughness. The angle on a flat surface is defined here as θA.

Wetting on Cu-Based Super-Hydrophobic Surfaces

cos θ ) fPLfPS cos θA - (1 - fPLfPS)

Langmuir, Vol. 21, No. 3, 2005 941

(4)

where fPL is the fractional surface (projected onto a plane) of just the larger-scale roughness and fPS that of the smaller-scale roughness, as defined in Figure 4. It is shown in the Supporting Information that eqs 3 and 4 are entirely equivalent to finding the angle on the smaller-scale roughness using the appropriate equation (i.e., either Wenzel eq 1 or Cassie-Baxter eq 2 with the r and f factors corresponding to the small-scale roughness or pattern) and entering the resulting angle into Cassie and Baxter’s equation, eq 2, but with the f factor corresponding to the fractional area of the larger pattern. In the case investigated experimentally the random nature of the smaller-scale roughness makes it difficult to calculate the angle expected. We measured the contact angle on an unpatterned surface etched to produce similar roughness at 122 ( 3°. Inserting this into Cassie and Baxter’s equation, eq 2, with fP ) (4 - π)/4, the value for the larger-scale roughness gives a calculated contact angle of 154°, within error of the measured value. Longer etch times reduced the area of the tops of the peaks and, thus, the factor f in eq 2; a comparison of part b with part c of Figure 3 shows that the tops of the peaks can be significantly reduced in area. The equilibrium contact angle on this surface was 161 ( 3°. Putting this value for θr and 122° for θs into eq 2 gives a value for f of 0.116, or about half of that of the other samples. The electron micrographs in Figure 3 suggest that this may indeed be the case, but the samples were not homogeneous enough to allow direct measurement of this value. These contact angle results show that changing the roughness of the tops of partially wetting samples changes the effective angle θs in the Cassie and Baxter equation whereas increasing the roughness of the base of a pattern on which water contacts completely (Wenzel-like), however, can cause switching to bridging of the depressions (Cassie-Baxter-like). A mechanism for this is suggested by Herminghaus,34 who notes that the superposition of two roughnesses with low pitch creates parts of the surface with higher inclination to the horizontal than would be the case with either alone. Bridging can occur if the local inclination angle plus the contact angle reaches 180°, as the meniscus can then be horizontal, so increasing the inclination of parts of the surface will increase the chance of bridging occurring. Other evidence suggests that combining two layers of roughness produces a stronger effect than the geometrical sum of the roughness,35 particularly when the smaller-scale roughness is at the base of larger-scale depressions. As the etch pits were hemispherical both mechanisms are possible in this case. Some information concerning whether a surface shows Cassie-Baxter or Wenzel hydrophobicity can be obtained from contact angle hysteresis measurements, which determine how easily drops can be tipped off a surface. Some reports suggest that contact angle hysteresis can be used as a measure of whether a water drop is in full or partial contact with a hydrophobic surface.25,36 Lafuma and Quere´37 have described the Wenzel super-hydrophobic state as “sticky”, and the Cassie-Baxter state as “slippery”, as the contact angle hysteresis on surfaces where Wenzel’s equation applies are generally higher than those where the Cassie-Baxter equation applies. The contact angle hysteresis of water drops was measured on the various etched samples prepared for this study. (36) Marmur, A. Langmuir 2004, 20 (9), 3517-3519. (37) Lafuma, A.; Quere´, D. Nat. Mater. 2003, 2 (7), 457-460.

Table 1. Contact Angle Hysteresis on Etched Patterns and Smooth Surfaces

sample smooth copper etched unpatterned chemically polished patterned (Figure 3a) pattern smooth tops rough pits (Figure 2) pattern rough tops and pits (Figure 3b) a

equilibrium contact angle, deg ((2°)

contact angle hysteresis, deg ((6°)

115 122 142

47 94 83

152

37a

155

22a

Hysteresis measured by tilting the sample.

Figure 5. Electron micrographs of electrochemically deposited copper at various current densities: (a) 100 mA cm-2 viewed from 45°, (b) 200 mA cm-2 viewed from 45°, (c) 100 mA cm-2 viewed from the perpendicular, 5× higher magnification, and (d) 200 mA cm-2 viewed from the perpendicular, magnification as in part c.

Table 1 shows that the contact angle hystereses on the chemically polished and flat etched samples are higher than those on a flat surface, whereas those on rough etched patterns were lower. This agrees with the calculation made on the equilibrium contact angles, indicating that the smooth chemically polished pits were filled with water and the rough etched pits were not. Copper Electrodeposition. Electrodeposition of copper onto flat copper electrodes produced randomly rough surfaces. During copper electrodeposition the resistive voltage drop across the cell prevented potential-dependent deposition so current control was used. The roughness of the copper deposit varied with the current density at the electrode. At low current densities the copper deposits were slightly rough, but increasing the current increased the roughness dramatically. Figure 5 shows electron micrographs of copper deposited at two different current

942

Langmuir, Vol. 21, No. 3, 2005

Shirtcliffe et al.

Figure 6. Surface area of copper deposits measured by lead underpotential deposition plotted against current density during deposition.

Figure 7. Equilibrium water contact angle on hydrophobized surfaces plotted against specific surface area of electrodeposited copper. The solid curve was added only as a guide, and the dotted curve shows the angle predicted by Wenzel’s equation.

densities. At 100 mA cm-2 a slightly rough, undulating structure with occasional spherical protrusions was produced (Figure 5a,c). At 200 mA cm-2 structures consisting of clumps of spheres were present (Figure 5b,d). These were generally taller than their width. Higher resolution micrographs showed that the larger structures were covered with smaller hemispherical protrusions (Figure 5d). Brady and Ball38 showed that electrodeposition of copper on a pointlike electrode produces fractal structures under diffusion-limited conditions. This occurs as the concentration of copper ions at the surface can be considered to be zero and protrusions on the deposit will intercept inwardly, diffusing ions otherwise bound for other parts of the surface. In this way protrusions grow at the expense of lower parts of the surface. On a flat plate the diffusion geometry will be different and competition between neighboring growths may affect their tendency to branch. The surface areas of the copper deposits were measured by lead underpotential deposition. Figure 6 shows how increasing the deposition current above 100 mA cm-2 increased the surface area of the deposit. At higher current densities the samples exhibited very high specific surface areas. When hydrophobized, these surfaces showed contact angle enhancement dependent upon their surface area and hence, roughness, up to very high contact angles. Figure 7 shows the contact angles against the specific surface areas of the samples. The steep line on the plot was calculated using Wenzel’s equation, eq 1, which assumes that the water is in contact with the entire surface (38) Brady, R.; Ball, R. Nature 1984, 309, 225-229.

Figure 8. Contact angle hysteresis of water on electrodeposited copper plotted against the specific surface area. The last two points at high specific surface area measured by the tilting plate method.

of the copper; the solid curve is a trend line as a guide to the eye for the data. It is evident from the measured curve that the data does not follow the Wenzel equation. This suggests that the water is not in contact with the entire surface of any of the electrodeposited copper samples. When the contact angle hysteresis on the electrodeposited samples was plotted against the specific surface area it increased before decreasing to a very small value (Figure 8). The dotted line is a trend line to aid the eye. There are various theories about how contact angle hysteresis behaves on a super-hydrophobic surface, but if the contact angle hysteresis were controlled by the contact area, this would suggest that the contact area of the drops on electrodeposited copper initially increased with specific surface area. This would imply that at low roughness values and up to a specific surface area somewhere between 3 and 5 the contact area increased. Although some of the surface was almost certainly not wetted (Figure 7), the increase in surface area of the wetted regions could have been greater than that hidden in bridged areas. As the roughness increased further the contact angle hysteresis dropped again, possibly as the water bridging exceeded the increase in roughness of the contacted surface, and somewhere around a specific surface area of 6 it crossed the value measured on a flat surface. Conclusions Etched copper patterns were produced with diamondshaped pillars separated by hemispherical pits. When the

Wetting on Cu-Based Super-Hydrophobic Surfaces

pits were relatively smooth the equilibrium contact angles were close to those predicted by Wenzel’s equation. When the pits were rough the water did not seem to enter them and the contact angles observed matched those predicted by Cassie and Baxter’s equation; assuming that water was only in contact with the tops of the pillars. When the tops of the patterns were etch-roughened the contact angle increased again and matched a modified Cassie and Baxter’s equation using a measured value of the contact angle on the rough tops of the pattern. This suggests that the water is in contact with the pattern when the depressions are smooth but cannot remain in contact when it is roughly etched. Electrodeposited copper formed rough surfaces that were like low-order fractals. They took the form of globular deposits with globular texture. Contact angles on these surfaces increased with their surface area and reached very high values. Comparing the measured contact angles with values expected from Wenzel’s equation revealed that water drops were not in contact with the whole surfaces even on surfaces giving relatively low angles. Contact angle hysteresis increased with surface area of the deposits before decreasing to a very low value.

Langmuir, Vol. 21, No. 3, 2005 943

The etched copper surfaces followed the Wenzel and Cassie-Baxter equations, giving contact angles as calculated and showing “stickiness” or “slipperiness” depending upon their state. The electrodeposited samples showed gradual changeover from Wenzel- to CassieBaxter-type wetting, and contact angle hysteresis reached a maximum after a considerable amount of the surface was bridged by the drop. This suggests that the measurement of contact angle hysteresis is not entirely indicative of the type of wetting present. Acknowledgment. The financial support of the UK Engineering and Physical Sciences Research Council (EPSRC) and the MOD Joint Grant Scheme under Grant GR/02184/01 is gratefully acknowledged. Supporting Information Available: Derivation of contact angle equations, showing equivalence of considering two scales of roughness separately or concurrently. This material is available free of charge via the Internet at http://pubs.acs.org. LA048630S