Mimicking the Stenocara Beetle—Dewetting of Drops from a Patterned

This paper describes the preparation of superhydrophobic surfaces that have been selectively patterned with circular hydrophilic domains. These materi...
0 downloads 0 Views 182KB Size
6154

Langmuir 2008, 24, 6154-6158

Mimicking the Stenocara BeetlesDewetting of Drops from a Patterned Superhydrophobic Surface Christian Dorrer and Ju¨rgen Ru¨he* UniVersity of Freiburg, Department of Microsystems Engineering, Laboratory for the Chemistry and Physics of Interfaces, Georges-Ko¨hler-Allee 103, D-79110 Freiburg, Germany ReceiVed January 23, 2008. ReVised Manuscript ReceiVed March 5, 2008 This paper describes the preparation of superhydrophobic surfaces that have been selectively patterned with circular hydrophilic domains. These materials mimicked the back of the stenocara beetle and collected drops of water if exposed to mist or fog. Under the effect of gravity, the drops dewetted from the hydrophilic regions once a critical volume had been reached. The surface energy in the hydrophilic regions was carefully controlled and assumed various values, allowing us to study the behavior of drops as a function of the superhydrophobic/hydrophilic contrast. We have investigated the development of drops and quantitatively analyzed the critical volumes as a function of several parameters.

Introduction Fluorination is a popular route toward man-made hydrophobic surfaces. Nature, in contrast, achieves spectacular water-repellent properties entirely without fluor. For instance, the leaves of certain plants (well-known examples are the lotus and the lady’s mantle) are so hydrophobic that water drops retain a nearly spherical shape, quickly rolling off even at very low tilting angles.1–3 As has been shown, these so-called superhydrophobic properties result from the interplay between surface chemistry and surface roughness/topography.1–4 In a second striking example from nature, the stenocara beetle generates drinking water from fog in the extremely dry Namib desert:5,6 on this insect’s elytra, small, wax-free, hydrophilic bumps are dispersed over a waxy, superhydrophobic background material. Microscopic droplets from the morning fog collect and grow on the hydrophilic bumps, dewet and quickly roll toward the beetle’s mouthparts once a critical drop size has been reached. Water harvested by this procedure is the beetle’s only supply of the precious liquid in its otherwise extremely arid habitat.5 Several authors have recently mimicked the back of the stenocara beetle by generating hydrophilic bumps on (super)hydrophobic wax films,5 plasma polymers,7 or polyelectrolyte multilayers.8 As a possible application for such patterned superhydrophobic surfaces, the collection of water in regions of the earth were fog but no running water is available has been proposed.5,7 Other options would be the removing of drops from industrial exhaust gases or the guiding of liquids in microfluidic systems. Surfaces that are (micro)structured into regions of different wettabilities have recently received some scientific interest due to potential applications in microfluidics, microfabrication, and microassembly.9–11 Common techniques for preparing such * Corresponding author. E-mail: [email protected]. (1) Neinhuis, C.; Barthlott, W. Ann. Bot. 1997, 79, 667. (2) Barthlott, W.; Neinhuis, C. Planta 1997, 202, 1. (3) Wagner, P.; Fu¨rstner, R.; Barthlott, W.; Neinhuis, C. J. Exp. Bot. 2003, 54, 1295. (4) Li, X.-M.; Reinhoudt, D.; Crego-Calama, M. Chem. Soc. ReV. 2007, 36, 1350. (5) Parker, A. R.; Lawrence, C. R. Nature 2001, 414, 33. (6) Naidu, S. G. J. Insect Phys. 2001, 47, 1429. (7) Garrod, R. P.; Harris, L. G.; Schofield, W. C. E.; McGettrick, J.; Ward, L. J.; Teare, D. O. H.; Baydal, J. P. S. Langmuir 2007, 23, 689. (8) Zhai, L.; Berg, M. C.; Cebeci, F. C.; Kim, Y.; Milwid, J. M.; Rubner, M. F.; Cohen, R. E. Nano Lett. 2006, 6, 1213. (9) Gau, H.; Herminghaus, S.; Lenz, P.; Lipowsky, R. Science 1999, 283, 46.

chemically heterogeneous surfaces include vapor deposition,9 contact printing,12–14 and photolithography.15–17 The field of patterned superhydrophobic surfaces, in comparison, has been studied less extensively; here, as an additional parameter, the surface roughness must be carefully controlled. On superhydrophobic materials, a three-dimensional surface structure is combined with a hydrophobic surface chemistry to obtain a situation where drops are suspended on top of the roughness features, with air trapped underneath.4,18–22 The contact angles (CAs) in this composite or Cassie state of wetting are high, approaching 180° in some cases.4,18–20,22 Additionally, the CA hysteresis is often greatly decreased, leading to a quick rolling off of drops from the substrates even if only very small forces are applied. While two studies have analyzed the water-gathering ability of artificial Stenocara-type surfaces by measuring the amount of water collected over a given time span,5,7 we believe that several fundamental aspects require a more detailed description. In this work, we have prepared extremely water-repellent surfaces equipped with circular “defects” of less hydrophobic, well-defined surface chemistries. Our model surfaces cover a range of wettability contrasts between 178°/0° and 178°/120° (CAs in the superhydrophobic/defect regions). We investigate, among other things, the influence of the wettability contrast on the dewetting of drops from the surfaces. (10) Srinivasan, U.; Liepmann, D.; Howe, R. T. J. Microelectromech. Syst. 2001, 10, 17. (11) Syms, R.; Yeatman, E. M.; Bright, V. M.; Whitesides, G. M. J. Microelectromech. Syst. 2003, 12, 387. (12) Drelich, J.; Wilbur, J. L.; Miller, J. D.; Whitesides, G. M. Langmuir 1996, 12, 1913. (13) Lopez, G. P.; Biebuyck, H. A.; Frisbie, C. D.; Whitesides, G. M. Science 1993, 260, 647. (14) Pompe, T.; Fery, A.; Herminghaus, S. Langmuir 1999, 15, 2398. (15) Paterson, A.; Fermigier, M. Phys. Fluids 1997, 9, 2210. (16) Mo¨ller, G.; Harke, M.; Hotschmann, H.; Prescher, D. Langmuir 1998, 14, 4955. (17) Samuel, J. D. J. S. S.; Ru¨he, J. Langmuir 2004, 20, 10080. (18) Nakajima, A.; Hashimoto, K.; Watanabe, T. Monatsh. Chem. 2001, 132, 31. (19) Feng, X.; Jiang, L. AdV. Mater. 2006, 18, 3063. (20) Blossey, R. Nat. Mater. 2003, 2, 301. (21) Dorrer, C.; Ru¨he, J. Langmuir 2006, 22, 7652. (22) Dorrer, C.; Ru¨he, J. AdV. Mater. 2008, 20, 159.

10.1021/la800226e CCC: $40.75  2008 American Chemical Society Published on Web 05/20/2008

Mimicking the Stenocara Beetle

Langmuir, Vol. 24, No. 12, 2008 6155

Experimental Section Micromachining. Silicon nanograss was fabricated according to the “black silicon method”.23 In this technique, reactive ion etching is used to generate silicon surfaces with high degrees of roughness. The resulting spike structure traps light rays, the material therefore appearing completely black. The processed substrates were submitted to an oxygen plasma treatment to remove remnants of the passivation layer. Surface Modification. After micromachining, a monolayer of a benzophenone-based silane, (4-(3′-chlorodimethylsilyl)propyloxybenzophenone), was immobilized at the bare silicon according to procedures described previously.17,21,22,24 The samples were then dipped into a solution of poly(heptadecafluorodecylacrylate) (PFA, c ) 3 mg/mL in 1,1,2-trichlorotrifluoroethane) and withdrawn at a speed of 1 mm/s. After evaporation of the solvent, the polymercoated surfaces were exposed to UV radiation (5 min at 265 nm), leading to the covalent attachment of a polymer monolayer to the silicon.17,21,22,24,25 In a final step, any unbound polymer was removed in a Soxhlet extraction procedure using again 1,1,2-trichlorotrifluoroethane as a solvent. On polished silicon slides, this procedure led to the deposition of thin films of around 5 nm thickness, as determined by ellipsometry (assuming a refractive index for the polymer of n ) 1.337). Surface Patterning. To generate bumps, 50 mg/mL solutions of the following polymers were prepared: poly(dimethylacrylamide) (PDMAA) in ethanol, poly(styrene) (PS) in toluene, and PFA in 1,1,2-trichlorotrifluoroethane. Defined volumes of these solutions (V ) 0.5-1 µL) were dispensed onto the nanograss surfaces using a pipet. After evaporation of the solvent by placing the substrates on a hotplate (T ) 80°), this process was repeated until circular bumps of material were obtained. In an alternative procedure, an aluminum mask was prepared with circular openings 500 µm, 1 mm, 2 mm, and 5 mm in diameter. The mask was placed on a nanograss substrate and the ensemble was irradiated with shortwavelength UV radiation (λ ) 190 nm, T ) 2 h, pen-ray lamp from LOT Oriel), leading to a photovolatilization of the PFA surface coating in those regions not protected by the photomask. Surface Characterization. Contact angles (CAs) were measured using an OCA20 system from Dataphysics. To determine the wettabilities of the polymers, thin layers were dip-coated onto polished silicon slides. Then 2-µL drops of water were deposited and analyzed. The advancing/receding angles were measured by piercing sessile drops with a fine needle and dispensing/ withdrawing liquid at a rate of 0.1 µL/s. For a determination of the critical volumes, the substrates were first inclined at a defined angle. A tiny drop was then deposited on a hydrophilic spot, and the dispensing needle was pulled out of the droplet. A 1-µL drop was dispensed from the needle; drops of this size were so small that they clung to the needle. Lowering the needle onto the drop that was already resting on the hydrophilic spot led to the coalescence of both drops. The volume of the drop on the hydrophilic spot was thus increased in 1-µL increments. Eventually, following a coalescence event, a rolling off of the drop occurred. The drop volume at this point was taken as the critical volume, VCrit.

Results and Discussion Patterned Superhydrophobic Surfaces. In a first step, superhydrophobic nanograss was prepared and characterized. As we have recently reported, the roll-off angles on this type of material are extremely low (below 1° for a 2-µL droplet).16 For the contact angle of a sessile drop, we measured 178° ( 2°, with no appreciable hysteresis. Hydrophilic to moderately hydrophobic spots were subsequently generated following two alternative (23) Jansen, H.; de Boer, M.; Legtenberg, R.; Elwenspoek, M. J. Micromech. Microeng. 1995, 5, 115. (24) Prucker, O.; Naumann, C. A.; Ru¨he, J.; Knoll, W.; Frank, C. W. J. Am. Chem. Soc. 1999, 121, 8766. (25) Dorrer, C.; Ru¨he, J. Langmuir 2007, 23, 3820.

Figure 1. Side view micrograph of an artificial stenocara beetle’s back with hydrophilic bump and hydrophobic background material. Insets show an electron microscopy image of the nanograss structure and the chemical formulas of the polymers that were used for coating the nanograss structure and as a bump material.

procedures. In the first case, we made use of the fact that, on rough surfaces, depending on the interplay between surface roughness and chemistry, two distinct wetting states may appear. In the composite state, as explained above, drops rest on top of the surface features; their footprint is partially in contact with air. Generally speaking, this wetting state is preferred if the CA of the respective liquid on smooth layers of the surface coating, θS, is above a critical value, θCrit. In the Wenzel state, in contrast, drops penetrate the surface structure. Wenzel wetting is energetically preferred if θS < θCrit. From theory, θCrit is given by the following equation:26,27

r cos θe,Crit ) φ cos θe,Crit + φ - 1

(1)

where r is Wenzel’s roughness factor and φ is the fraction of the drop footprint in contact with solid. For silicon nanograss, both r and φ are unknown. However, from our experiments, the following statements may be derived: the CA of water on smooth layers of PFA, θAdv ) 120° ( 2°, was higher than θCrit, because water drops wetted hydrophobicized nanograss in the composite state. In contrast, the CAs of solutions of PDMAA in ethanol, PS in toluene, and PFA in 1,1,2-trichlorotrifluoroethane on smooth layers of the fluoropolymer, θAdv ) 22° ( 2°, 30° ( 3°, and 27° ( 2°, were below θCrit. These liquids penetrated the nanograss structure and formed Wenzel drops. Following these preliminary experiments, polymer bumps were prepared by dispensing drops of polymer solution onto the superhydrophobic surfaces. For a given CA of a drop on a surface, the diameter of the drop footprint, d, could be controlled by varying the dispensed volume. We adjusted d to a value of around 2 mm. Following the initial dispensing step, the solvent was allowed to evaporate. The receding angle of the polymer solutions on the nanograss surfaces was 0°; therefore, d remained constant during the evaporation process. The dispensing procedure was repeated until bumps of solid material with the shape of a spherical cap were obtained (Figure 1). To characterize the wettability of the polymers, thin layers were dip-coated onto polished silicon slides. The following CAs were determined using water as a test liquid: θAdv ) 51° ( 2°, θRec ) 23° ( 2° (PDMAA); θAdv ) 97° ( 2°, θRec ) 81° ( 2° (PS); and θAdv ) 120° ( 1°, θRec ) 101° ( 2° (PFA). The inset in Figure 1 shows the (26) Bico, J.; Thiele, U.; Quere, D. Colloids Surf., A 2002, 206, 41. (27) Furmidge, C. G. L. J. Colloid Sci. 1962, 17, 309.

6156 Langmuir, Vol. 24, No. 12, 2008

Dorrer and Rühe

Figure 3. Detail of the dewetting process for a substrate equipped with a PS bump (a) and a substrate equipped with a photoablated hydrophilic region (b). The images show time steps 50 ms apart. The white lines denoted by w indicate the extension of the hydrophilic domain. The black lines in part a are tangents to the bump and the drop at the contact line. The arrows in part b indicate the region where the meniscus becomes unstable. The small drop in the last image of series a is a drop that has freshly settled onto the bump from the ambient fog.

Figure 2. Development of a water drop on a patterned superhydrophobic surface (bump material: PDMAA, bump width ) 2 mm). See the text for a detailed description. The lines indicate tangents to the drop at the three-phase contact line.

chemical formulas of the polymers that were used in the fabrication process. In the second case, a lithographical technique was employed: superhydrophobic nanograss was covered with an aluminum mask into which circular holes of various diameters had been drilled. The samples were exposed to short-wavelength UV radiation, leading to a photoablation of the fluoropolymer coating in those regions not protected by the photomask; here, the bare silicon was exposed. We observed that drops were sucked into regions of the nanograss surfaces where the fluoropolymer was missing (superwetting behavior; the CAs in these domains were θAdv ) θRec ) 0°). Development of Drops on a Patterned Superhydrophobic Surface. In the experiments described in the following, we studied the development of drops on the hydrophilic to moderately hydrophobic spots qualitatively. To follow the analogy to the stenocara beetle, water was brought onto the surfaces as a mist by spraying it from a fine nozzle. In this case, microscopic droplets immediately started to collect on the samples. On the superhydrophobic parts of the surfaces, the smaller drops were initially stationary; they did not behave as if they were in contact with a superhydrophobic material. The small drops were initially on the same size scale as the roughness features and therefore experienced not a superhydrophobic contact but were immobile. However, these drops constantly coalesced with each other or were sucked up by larger drops that came rolling over the surface. Thus, in many cases, the drops eventually made their way to the hydrophilic/moderately hydrophobic spots. The development of drops on the hydrophilic patches themselves followed a well-defined mechanism, as illustrated in Figure 2 for a PDMAA bump: liquid quickly began to collect on the bump and a small droplet was formed that, however, did not yet cover the entire bump (1.5 s). The CA of the drop relative

Figure 4. Critical volumes at which drops rolled off as a function of 1/sin R for different bump materials (bump diameter ) 2 mm).

to the surface, θ, was at this stage equal to the advancing CA on the bump material. The drop grew to eventually occupy the entire bump. The meniscus was now pinned on the bump edge; with further growth in the drop volume, the CA therefore increased (3-6 s). As a result of gravity, the CA on the downhill side of the drop now became larger than that on the uphill side (9 s). On the downhill side, the advancing angle on the superhydrophobic parts of the surface was eventually reached, leading to a movement of the contact line into this region (12 s). On the uphill side, the CA simultaneously decreased (12-21 s). Finally, the drop rolled off once a critical volume had been surpassed (21-24 s). The last stages of the dewetting process were different depending on the hydrophobicity of the bump: for the more hydrophobic bumps (PFA and PS), the contact line receded into the bump, as illustrated in Figure 3a. Upon dewetting, no liquid was left behind. For the more hydrophilic substrates (photoablated spot and PDMAA bump, respectively), a different mechanism was observed: in this case, the contact line remained pinned on the uphill edge of the hydrophilic domain. As indicated by the arrows in Figure 3b, the meniscus then became unstable with increasing volume and the drop rolled off, leaving a small amount of liquid behind. Critical Volume for the Dewetting of Drops. In this section, we are concerned with a quantitative analysis of the critical volumes at which drops dewetted from the bumps for the case

Mimicking the Stenocara Beetle

Langmuir, Vol. 24, No. 12, 2008 6157

Figure 5. Critical volumes as a function of the diameter of the hydrophilic spot (photoablated substrate, substrate tilting angle R ) 23°).

where the substrates were tilted at a fixed angle. The conditions for the sliding of drops on inclined, chemically homogeneous surfaces have been the subject of several scientific publications.27–33 In general, for a given surface and tilting angle, a drop starts to slide once its volume is increased above a critical volume, VCrit. As a condition for the drop to remain stuck on a surface of homogeneous wettability, the pinning force must be larger or equal to the gravitational pull experienced by the drop:27,28

wγ(cos θRec - cos θAdv) g FgV sin R

(2)

where w is the width of the solid-liquid contact, γ is the liquid-vapor interfacial energy, F is the liquid density, g is the gravitational acceleration, and R is the substrate tilting angle. For a drop on a homogeneous surface, w is not an independent parameter because the size of the liquid-solid interface changes with the drop volume V and the CA of the drop on the surface:

w(V,θ)γ(cos θRec - cos θAdv) g FgV sin R

(3)

For the case of a chemically homogeneous surface, VCrit is therefore a relatively complicated function of the tilting angle and the wettability contrast. For the patterned superhydrophobic surfaces, we determined VCrit as a function of the bump diameter, the substrate inclination angle, and the wettability contrast. In Figure 4, VCrit is plotted against 1/sin R for various surface chemistries (R was between 18° and 32°). The critical volumes were lowest for the surfaces where PFA was used as bump material; for a tilting angle of 32°, drops rolled off at a volume of 19 ( 1 µL. The highest volumes were observed for the surface with the photoablated spot at the lowest tilting angle; here, a value of 74 ( 2 µL was recorded. It is more interesting to look at the dependence of the critical volume on the bump diameter, d (Figure 5). Here, a relatively strong dependence was observed, with values ranging from 12 ( 1 to 104 ( 4 µL as d was increased from 0.5 to 5 mm. Since the relation between VCrit and d was approximately linear, we can write

VCrit ∼ d

(4)

where ∼ means “proportional to”. This result implies that the width of the solid-liquid contact, and thus the force with which the droplet is pinned, is determined by the size of the bump. Figure 6 shows the development of a drop on a moderately (28) Miwa, M.; Nakajima, A.; Fujishima, A.; Hashimoto, K.; Watanabe, T. Langmuir 2000, 16, 5754. (29) Berejnov, V.; Thorne, R. E. Phys. ReV. E 2007, 75, 66308. (30) Roura, P.; Fort, J. Phys. ReV. E 2001, 64, 011601. (31) Kawasaki, K. J. Colloid Sci. 1960, 15, 402. (32) Dussan, V.; Chow, R. J. Fluid Mech. 1983, 137, 1. (33) McDougall, G.; Ockrent, C. Proc. R. Soc. London Ser. A 1942, 180, 151.

Figure 6. Development of a drop on a moderately hydrophobic bump in the top view (bump material, PS; bump width, w ) 2 mm). The drop grows from a size of 15 µL (a) to 40 µL (d). The white circles indicate the bump; the dashed line in part d indicates the position of the contact line. The drop reflects the microscope illumination beam.

hydrophobic bump in the top view. In Figure 6a, liquid covered the entire bump. In Figure 6d, in contrast, the contact line had receded into the bump, and a rolling off of the drop was imminent. We note that, between parts a and d of Figure 6, the width of that part of the drop’s liquid-solid interface that was in contact with the bump remained approximately constant. Probably, there was an additional fraction of the liquid-solid interface over which the drop was in contact with the superhydrophobic parts of the surface. However, due to the extremely high receding angle on hydrophobicized nanograss, the pinning force exerted by this part of the drop’s contact area was negligible. For the structured surface, our results suggest that w(V, θ) may be replaced by the bump diameter in eq 3:

dγ(cos θRec - cos θAdv) ) FgVCrit sin R

(5)

For any given liquid, some of above parameters can be combined into the dimensionless capillary length, λC, which is normally defined as

λC )

 Fgγ

(6)

The capillary length is a characteristic length scale for a liquid onto which gravity acts as a body force and interfacial energy as a surface force. In the present case, a modified capillary length λC,m can be introduced by assuming that the effect of gravity is mitigated by a factor of sin R:

λC,m )

 Fg sinγ R

(7)

Substituting eq 7 into eq 5 and solving for VCrit yields

VCrit ) λC,m2d(cos θRec - cos θAdv)

(8)

Equation 8 allows for the calculation of the critical volume from the geometrical and chemical parameters of the surface. In Figure 7, the dimensionless critical volume VCrit/(λC,m2d) is plotted against cos θRec - cos θAdv for the data from Figure 3. For θRec, the receding angle on the bump material was used; the advancing angle on the superhydrophobic nanograss surface was substituted for θAdv. As Figure 7 shows, the agreement between theory (straight line in Figure 7) and experiment was good for lower wetting contrasts, indicating that eq 8 adequately described the dewetting behavior in this regime. We note that, for a nonhysteretic

6158 Langmuir, Vol. 24, No. 12, 2008

Dorrer and Rühe

regions, we take a value of 170°. The diameter of the hydrophilic patches on the beetle’s back is around 500 µm.5,7 For a tilting angles between 30° and 45°, our model predicts critical volumes between 7 and 10 µL. These values correspond to a range of drop radii between 2 and 4 mm, which is fairly close to the drop diameters that have been observed for the real beetle (4-5 mm).5 Discrepancies might for one be due to the fact that, in the desert, wind exerts an additional force on the drops. Second, the hydrophilic spots on the beetle’s back are not perfectly circular; this feature might influence the dewetting behavior because the effectiVe diameter of the hydrophilic spot might vary depending on the orientation of the spot relative to the gravitational force.

Conclusions Figure 7. The dimensionless critical volume is plotted against the wettabity contrast for the data from Figure 3. The straight line is the theoretical shape of the curve according to eq 8 (assuming a bump diameter of 2 mm).

surface (cos θRec - cos θAdv ) 0), eq 8 predicts a critical volume of zero; in this case, the pinning power of the defect is zero. For high wetting contrasts (i.e., bumps with θRec ) 23° and 0°), VCrit/(λC,m2d) was lower than predicted by eq 8. We attribute this behavior to a mechanism where the drop did not recede uniformly over the entire bump, as for the more hydrophobic substrates (Figure 3a). Rather, due to the instability of the liquid meniscus (Figure 3b), the drop rolled off before the full pinning force of the hydrophilic domain was experienced. A direct comparison of our results with the real beetle is difficult because, for instance, the precise contact angles on the stenocara beetle’s back are not known. For a rough calculation using eq 8, we assume that the receding CAs on the beetle’s hydrophilic patches are in a range from 50° to 70°, which are reasonable values for a smooth, waxy surface.3,34 For the superhydrophobic

In this work, we have prepared various superhydrophobic surfaces equipped with smooth, circular patches of hydrophilic to moderately hydrophobic polymers. These materials acted to mimic the back of the stenocara beetle, collecting water if exposed to a foggy atmosphere. For a range of wetting contrasts, patch diameters, and tilting angles, we have investigated the critical volumes at which drops dewetted from the circular defects and rolled off from the substrates. Our data was analyzed in terms of a simple model where the pinning force exerted by a bump was opposed to the gravitational pull experienced by the drops. As our results revealed, the pinning force for a given bump is constant and does not depend on the drop volume. Interestingly, for the surfaces discussed here, the complex surface structure thus actually leads to a simplification of the theoretical description compared to surfaces of homogeneous wettability. LA800226E (34) Cheng, Y.-T.; Rodak, D. E.; Wong, C. A.; Hayden, C. A. Nanotechnology 2006, 17, 1359.