Annu. Rev. Mater. Res. 2008.38:71-99. Downloaded from arjournals.annualreviews.org by 213.160.13.186 on 07/09/08. For personal use only. Wetting and Roughness David Quéré Laboratoire de Physique et Mécanique des Milieux Hétérogènes, ESPCI, 75005 Paris, France; email: [email protected] Annu. Rev. Mater. Res. 2008. 38:71–99 Key Words First published online as a Review in Advance on April 7, 2008 microtextures, superhydrophobicity, wicking, slip The Annual Review of Materials Research is online at matsci.annualreviews.org Abstract This article’s doi: 10.1146/annurev.matsci.38.060407.132434 c 2008 by Annual Reviews. Copyright All rights reserved 1531-7331/08/0804-0071$20.00 We discuss in this review how the roughness of a solid impacts its wettability. We see in particular that both the apparent contact angle and the contact angle hysteresis can be dramatically affected by the presence of roughness. Owing to the development of refined methods for setting very well-controlled micro- or nanotextures on a solid, these effects are being exploited to induce novel wetting properties, such as spontaneous filmification, superhydrophobicity, superoleophobicity, and interfacial slip, that could not be achieved without roughness. 71 1. WETTING WITHOUT ROUGHNESS 1.1. Ideal Wetting Annu. Rev. Mater. Res. 2008.38:71-99. Downloaded from arjournals.annualreviews.org by 213.160.13.186 on 07/09/08. For personal use only. Controlling the wettability of solid materials is a classical and key issue in surface engineering. Roughly speaking, two extreme limits are often desired. The first limit is complete wetting, in which a liquid brought into contact with a solid spontaneously makes a film. In the case of a windshield, for example, this film maintains the transparency of the glass; in addition, the film flows in the gravity field (if the car is stopped) or due to air friction (when it moves), taking dust particles with it. The second limit is complete drying: Liquid drops remain spherical without developing any contact with the substrate. They are thus readily evacuated, which prevents liquid contamination of the solid surface. It is of obvious interest to determine which parameters favor both these situations. The basic laws were first established for ideal solids, which are both flat and chemically homogeneous. As understood by Young and Laplace, surfaces carry a specific energy, the so-called surface tension, that reflects the cohesion of the underlying condensed phase (either solid or liquid). This quantity, denoted as γ IJ for an interface between phases I and J (below the indices are S, L, A for solid, liquid, and air, respectively), is an energy per unit area and thus a force per unit length: This force applies along the IJ surface to minimize the corresponding (positive) surface energy. We denote the liquid/air surface energy simply as γ . Hence, we arrive at a construction first imagined by Marangoni: A film spreads from a reservoir of liquid (a drop or a bath) onto a solid, as sketched in Figure 1a, provided that the solid/air surface tension γ SA (which entrains this film) is larger than γ SL + γ , the sum of the solid/liquid and liquid/air surface tensions (which both resist the spreading because complete wetting expands the two corresponding surface areas). The sign of the spreading parameter S = γ SA − γ SL − γ will thus determine the behavior of a drop on a solid: For S > 0, a drop spreads, whereas it forms a small lens in the opposite case. This lens meets the solid with a well-defined contact angle θ, whose value is similarly given by a force balance (Figure 1b). Projecting on the solid plane the different surface tensions acting on the contact line provides the equilibrium condition of the drop (1). The balance at equilibrium can be written as γSA = γSL + γ cos θ. 1. The contact angle is thus fixed univocally by the chemical nature of the different phases. Here we show that this statement can be dramatically affected if the solid is rough. We refer below to the angle θ as the chemical or Young angle. In many common situations, this angle lies between 0◦ and 90◦ (i.e., the hydrophilic case). Very qualitatively, a solid/liquid surface tension (between a b γ γSL γSA θ γ γSA γSL Figure 1 Two classical wetting situations for an ideal material. (a) A liquid film spreads, drawn by the solid/air surface tension, despite the action of the liquid/air and solid/liquid tensions. (b) Wetting is only partial, and the balance of surface tensions determines the contact angle θ. 72 Quéré Annu. Rev. Mater. Res. 2008.38:71-99. Downloaded from arjournals.annualreviews.org by 213.160.13.186 on 07/09/08. For personal use only. two condensed phases) tends to be smaller than a solid/air one (with only one condensed phase) because the phases are less contrasted in the first case. Hence, a positive cosine in Equation 1 results, implying an acute angle θ . Conversely, we could define a drying parameter D = γ SL − γ SA − γ . If γ SL is larger than γ SA + γ , the contact line will be withdrawn by surface forces until a film of air comes between the solid and the liquid: D > 0 is the criterion for complete drying. (This criterion is also simply derived by making cos θ < −1 in Equation 1.) There is a first case in which this criterion is fulfilled: For a system in which complete wetting is achieved (for example, water on freshly cleaned glass), inverting the liquid and the air immediately provides D > 0; a bubble of air at the bottom of the same glass filled with water will completely dewet glass. In a particular circumstance, the so-called Leidenfrost effect (2), D is forced to vanish: If water (or any volatile oil) is deposited on a solid whose temperature is much larger than the boiling temperature of the liquid, a vapor film forms between the solid and the liquid, which sits on its own vapor. Considering in Equation 1 that the solid role is played by the vapor, we determine that D = 0. But the situation is quite different at standard temperatures. Although complete wetting can be achieved (for example, with most light oils on most solids), complete drying of water (or any oil) on a flat solid is never observed. On the most hydrophobic solids (fluorinated materials), the contact angle never exceeds approximately 120◦ (3), to which corresponds a negative parameter D (of approximately −γ /2). We term hydrophobic these situations in which obtuse angles are observed. Another aim of this review is to show how one can take advantage of the surface roughness for filling the (large) gap existing between 120◦ and (nearly) 180◦ , thus generating ultrahydrophobic behaviors of obvious practical interest (water repellency). 1.2. Ideal Wicking Materials are often not fully solid, yet are porous. We restrict our discussion to the (ideal) case of cylindrical pores of constant diameter and consider one of these pores. A liquid will penetrate such a tube if the surface energy of the solid is lower wet than dry (Figure 2a). We can introduce a wicking parameter W = γ SA − γ SL , whose sign indicates if liquid penetrates the tube. Wicking will occur if W > 0, that is, if the contact angle is smaller than 90◦ (as we see from Equation 1). Then, the meniscus formed by the liquid inside the tube must be curved in such a way that the Laplace pressure (associated with curved surfaces) is negative below the surface—another way to understand liquid penetration as resulting from this depression that sucks the liquid inside a b Figure 2 (a) A liquid brought into contact with a tube or a slot will penetrate it provided that the surface energy of the tube is lower wet than dry. This means that, as deduced from Equation 1, the contact angle of the liquid on the tube walls must be acute. (b) Conversely, for an obtuse contact angle, the tube tends to remain dry. If gravity is present, both the rise and descent are limited. www.annualreviews.org • Wetting and Roughness 73 Annu. Rev. Mater. Res. 2008.38:71-99. Downloaded from arjournals.annualreviews.org by 213.160.13.186 on 07/09/08. For personal use only. the tube. As stressed above, contact angles are generally acute, which means that most sponges absorb most liquids. A system for which we have S > 0 will necessarily satisfy the condition W > 0 (then, the contact angle is zero, indeed smaller than 90◦ ), and it is worth discussing carefully the mechanism of invasion in this case. Therefore, as shown in Figure 1a, a liquid film (typically of a molecular thickness) progresses along the tube walls (4) such that the meniscus behind the film advances on a prewet tube. Because this meniscus suppresses the liquid/air interface as it moves, the penetration is always favorable. The wicking parameter can be written as W = γ , which is indeed the limit of W as the contact angle vanishes, as seen in Equation 1. Conversely, wicking is not favorable for W < 0, and liquid is expelled from the pore (Figure 2b)—hence the development of the idea that a solid decorated with hydrophobic cavities can remain filled with air, even if the solid is exposed to a liquid, and thus approach the Leidenfrost limit. As we see, not only can roughness modify the wettability of a solid but also— perhaps the main message of this review—roughness can result in new and specific properties such as water repellency. We first show that the natural roughness of most solids is likely to induce pinning of the contact line and thus variability of the contact angle (apparently contrasting with what can be expected from Equation 1). Subsequently, we discuss how special kinds of roughness (well-designed microstructures) can be created at the solid surface to control both wettability and pinning and, beyond, special hydrodynamic properties such as slip. 2. ROUGH SOLIDS 2.1. Contact Angle Hysteresis Most solids are naturally rough, often at a micrometric scale. Processes of fabrication (such as lamination) may generate striations or microgrooves. Materials resulting from the compaction of grains exhibit roughness at the scale of the grains. Coating can also induce roughness, in particular when the coating film dewets, thus producing microdrops at the surface. Conversely, very few solids are molecularly flat. Most often, molecularly flat solids result from solidifying a liquid film, either free or suspended on another liquid; in such cases, the roughness can correspond to the thermal roughness of a liquid interface, generally of the order of a few angstroms. This is the case of glass, solidified from its molten state after deposition onto a bath of molten tin. Gibbs pointed out that defects on a solid can pin a contact line. As a consequence, droplets on an incline stay at rest; the front and rear contact nonwetting and wetting defects, respectively (5). The resulting asymmetry in contact angles creates a Laplace pressure difference between the front (of high curvature) and the rear (of smaller curvature) and, thus, a force able to resist gravity provided that the drop is small enough (6). Both chemical heterogeneities and roughness can act as pinning sites. It is useful to think of a single defect such as is sketched in Figure 3. Even on a chemically homogeneous surface, the edge of the defect (of characteristic angle φ) makes the contact angle flexible at this place. We measure a (Young) angle θ before the edge and a (Young) angle π − φ + θ after the edge, considering the horizontal as the reference (we ignore with our naked eye the existence of the defect). Hence it is possible to have any angle between θ and π − φ + θ at the edge (7). A groove can thus stop the front of a liquid drop (as if it were nonwetting), and a tip will act in the opposite way so that a solid decorated with both kinds of defects yields both small and large apparent angles. Contact angles therefore generally depend on the history of the process of liquid deposition. A drop gently deposited spreads and stops when it is surrounded by primarily nonwetting defects, which prevent it from exploring the solid further. After a while, the drop evaporates, and thus its configuration is that of a drop pinned on wetting defects. The way to quantify this contact angle 74 Quéré θ Φ Annu. Rev. Mater. Res. 2008.38:71-99. Downloaded from arjournals.annualreviews.org by 213.160.13.186 on 07/09/08. For personal use only. θ Figure 3 Apparent pinning of a contact line on an edge. The Young condition stipulates that the liquid meets the solid with a contact angle θ . Hence the contact angle at the edge can take any value (if the horizontal direction is considered as the reference one) between θ and π − φ + θ , as illustrated by the colored region. hysteresis consists of slowly increasing the volume of a drop: The contact line first remains stuck before it suddenly jumps above a critical volume (for which the line suddenly depins and moves toward a next series of pinning defects). The maximum observed angle is the so-called advancing contact angle θ a . Conversely, sucking the liquid from the drop flattens it until it depins and retracts to the next wetting series of events, which stop and pin the line; the minimum corresponding angle is the receding contact angle θ r . Contact angle hysteresis can be seen as beneficial (e.g., when it is exploited for guiding a flow along a line of defects, following a predefined route) or detrimental (e.g., water drops stuck on window panes distort their transparency and contribute to degradation of the glass). It is thus crucial to understand it, but there is still a debate about the laws that relate the microscopic picture (pinning on a single defect) to the macroscopic observations (measurement of the hysteresis, which averages on many defects). We give further an example of such a calculation. More generally, we see that the contact angle hysteresis θ = θ a − θ r varies dramatically on a rough solid, from nearly zero to a giant value, of the order of θ a itself (8). 2.2. The Wenzel Model We see above that roughness impacts the contact angle hysteresis. But it also affects the typical or apparent angle, which is (often very) different from the one expected from Equation 1. This was first appreciated by Wenzel (9), using a geometrical argument based on the roughness factor, r, the ratio between the actual surface area and the apparent surface area of a rough surface. A drop placed on a rough surface (Figure 4) will spread until it finds its equilibrium configuration, characterized by a contact angle θ ∗ (possibly different from the Young angle θ). The key dx Liquid Vapor θ Solid Figure 4 The Wenzel picture. One can obtain the apparent contact angle θ ∗ by considering a small apparent displacement of the contact line and looking at the corresponding variation in surface energy, assuming that the liquid follows the accidents of the solid surface. www.annualreviews.org • Wetting and Roughness 75 assumption of the model is sketched in Figure 4: As the contact line progresses on the dry solid, it is assumed to follow all the topological variations of the material so that each piece of liquid/air interface gets replaced by a solid/liquid interface of the same surface area. The surface energy variation dE arising from an apparent displacement dx of the line can be written, per unit length of the contact line, as Annu. Rev. Mater. Res. 2008.38:71-99. Downloaded from arjournals.annualreviews.org by 213.160.13.186 on 07/09/08. For personal use only. dE = r(γSL − γSA ) dx + γ dx cos θ ∗ , 2. where the second term on the right corresponds to the change of liquid/vapor surface area as the drop spreads. The roughness increases both the solid energies, enhanced geometrically by a factor r. The minimum of E (dE = 0) yields Equation 1 if the solid is flat (r = 1); if not, we find (9) cos θ ∗ = r cos θ, 3. where θ is the chemical angle given by Equation 1. The Wenzel relation (Equation 3) predicts that roughness enhances wettability. If the factor r is larger than 1, a hydrophilic solid (θ < 90◦ ) becomes more hydrophilic when rough (θ ∗ < θ). Conversely, a hydrophobic solid (θ > 90◦ ) shows increased hydrophobicity (θ ∗ > θ). Although these tendencies are generally (but not always) observed, agreement with Equation 3 is far from quantitative (see next section). We can guess that the Wenzel relation implies strange features. For example, there is no limitation for the effect: The roughness factor can be made arbitrarily large, which seems to imply that complete wetting (cos θ ∗ > 1) or complete drying (cos θ ∗ < −1) should be induced by large roughness (r 1). We show that such behavior is not observed because Wenzel assumptions often are not satisfied. Even when Wenzel relation is likely to be obeyed, it is difficult to check directly whether the relation is being followed. Because liquid conforms to the roughness, pinning of the contact line is particularly strong in this state, both on the edges of and along the defects. Besides, pushing a Wenzel drop leaves behind cavities filled with liquid such that the drop can also be pinned by the liquid itself. As a consequence, a Wenzel state is generally characterized by very low receding angles and thus giant hysteresis (θ ∼ θ a ). In such conditions, it is very difficult to extract the sole angle θ ∗ or to check Equation 3. Modern and more detailed discussions on the validity of the Wenzel model can be found in References 10–13, which stress in particular that drops should be much larger than the defects to use such an averaged model. In the converse limit, liquid rearranges such that the contact angle depends on the drop size (10, 11, 14). 3. MICROTEXTURED SOLIDS We show in the previous section that roughness modifies both the ideal character of the Young equation (the angle is not unique) and the value of the apparent observed angle. Therefore, wettability can be tuned by roughness. We can take advantage of roughness to modulate the surface properties of a solid and, even better, to induce properties that could not be generated otherwise, a theme that has been extremely popular during the past decade. Let us quote here three factors that contributed to the burst of this domain. (a) At the end of the 1990s, researchers from the Kao Corporation in Japan showed that extremely large angles could be obtained by the use of fluorinated rough (fractal) surfaces (15, 16). This result was not fully novel; similar results had been obtained in the 1940s (17) but somehow forgotten (18). (b) At the same time, Neinhuis and Barthlott in Germany systematically analyzed the structures on the surfaces of hydrophobic plants. These researchers showed the remarkable variety of the surface designs, suggesting that nature had optimized the patterns (19, 20). This kind of study was extended to animals, and new fascinating designs were (re)discovered and discussed (21–24). There have since 76 Quéré 3.1. The Kao Experiment Kao researchers constructed different well-characterized fluorinated substrates, either rough or flat (15, 16). They compared the contact angles θ ∗ on the rough samples with the contact angles on flat materials, which should be close to the Young angle θ. This comparison was performed through the use of several liquids to vary θ . A typical result is displayed in Figure 5, in which cos θ ∗ is plotted as a function of cos θ , showing the modification of the wetting properties generated by surface roughness. This plot provides only one angle θ ∗ (which seems to be close to the advancing angle, according to References 15 and 16), so we ignore the hysteresis associated with each data point. We first notice that the abscissa is far from exploring the complete interval [−1, 1]: cos θ is never smaller than −0.3, corresponding to an angle θ of approximately 110◦ . This data point was obtained by the use of water as a liquid; as emphasized above, contact angles on flat solids are never larger than approximately 120◦ , corresponding to the maximum existing chemical hydrophobicity (7). However, even if there are only a few data points on the hydrophobic side (cos θ < 0), we see a spectacular effect. As soon as we enter this domain, the apparent contact angle θ ∗ jumps and reaches a value of 170◦ (much larger than the chemical angle); roughness here induces a wetting behavior that could not be achieved otherwise. This state is often referred to as superhydrophobic. In the hydrophilic domain (cos θ > 0), cos θ ∗ first increases linearly with cos θ; the slope is larger than unity (approximately 3). It is tempting to interpret this behavior as a Wenzel regime (Equation 3). The material roughness deduced from micropictures is indeed in good agreement with this slope (16). It is amazing to deduce this complex (and invisible) quantity from such a simple (and cheap) experiment, in which just a few drops are used to probe the surface. However, this behavior is not obeyed when the contact angle θ becomes smaller than some critical value θ c . Instead, we observe a second linear regime (with a slope smaller than unity), which tends toward θ ∗ = 0 as θ = 0 (quite trivially, a wettable solid remains wettable if rough). We see in Section 4.2 that this Figure 5 1 cos θ * Annu. Rev. Mater. Res. 2008.38:71-99. Downloaded from arjournals.annualreviews.org by 213.160.13.186 on 07/09/08. For personal use only. been many attempts to mimic these natural patterns so as to understand their efficiency. (c) The recent development of microfabrication techniques allows us today to construct very well-defined micro- or nanostructures, which has pushed researchers, e.g., to imagine new designs and to optimize given designs. We now summarize different findings related to these three factors (a–c). Cosine of the apparent contact angle θ ∗ on a textured surface, as a function of the cosine of the Young angle θ measured on the same surface, yet flat (16). The lines show the behavior expected from Equations 3, 10, and 14. 0 θc –1 0 1 cos θ www.annualreviews.org • Wetting and Roughness 77 regime results from the penetration of the liquid inside the microtextures; this liquid surrounds the drop on which the contact angle is measured. Then, the lens of liquid sits upon a mixture of solid and liquid, at odds with the Wenzel hypothesis, which assumes a dry solid beyond the drop, as shown in Figure 4. We call this second regime superhydrophilic, and we describe this regime in Section 4. Annu. Rev. Mater. Res. 2008.38:71-99. Downloaded from arjournals.annualreviews.org by 213.160.13.186 on 07/09/08. For personal use only. 3.2. Natural Microtextures Microtextures are also found on the surfaces of many plants and animals (Figure 6). In his Natural History, Pliny the Elder noticed that water on a leaf forms perfect spheres, provided that the leaf surface is woolly (25, p. 32). The old literature (and poetry) sporadically reports the special wetting behavior of plants and animals, such as a review by Dufour in 1833 (26, pp. 68–74), a paper by Fogg in 1944 (27), and a comment on Fogg’s paper by Cassie & Baxter (28). More systematic studies performed (only) in the past decade generated a remarkable collection of microtextures, of which Figure 6a displays a few examples. On plant leaves, we often see bumps at the scale of 10–50 μm (Figure 6a,b). For the most popular of these hydrophobic plants, namely the lotus, and many other ones, these papillae are covered by fine nanostructures at the scale of 100 nm. The coexistence of two scales of roughness contributes significantly to the quality of the superhydrophobicity (29–36). However, despite many a b 50 µm c 50 µm d 50 µm 500 nm Figure 6 A few examples of natural superhydrophobic materials, as revealed by SEM. (a) Leaf of the so-called elephant’s ear (Colocasia esculenta). From Reference 39 (courtesy of Peter Wagner and Christoph Neinhuis). (b) Lotus leaf. Courtesy of Barthlott & Neinhuis (20). (c) Leg of a water strider. From Reference 23 (courtesy of Lei Jiang). (d ) Surface of a mosquito (Culex pipiens) eye. From Reference 40 (courtesy of Lei Jiang). Note the difference in scale between panels a–c and panel d. 78 Quéré Annu. Rev. Mater. Res. 2008.38:71-99. Downloaded from arjournals.annualreviews.org by 213.160.13.186 on 07/09/08. For personal use only. stimulating hypotheses, there is no real understanding for this hierarchical structure, which is not necessary for reaching very high degrees of hydrophobocity (37). Actually, such structures often provide both a high contact angle and a low contact angle hysteresis. For lotus, for example, the advancing angle is approximately 160◦ , and the receding angle is larger than 150◦ , which confers to water drops a high mobility on these leaves. In some cases, such as the rice leaf, the arrangement of the papillae at the surface can be anisotropic, and thus wetting and adhesion are also anisotropic (38). On such materials, water will flow preferentially along certain directions. Feathers of many birds [such as those of pigeons (41) and ducks] are hydrophobic and/or superhydrophobic, as are insects such as cicada, butterflies, and of course water striders (24, 42). Insects’ cuticles are covered with a layer of epicuticular wax (of typical thickness of 250 nm), which prevents the intrusion of water into the body (a serious threat for the insect) and protects the animals from dessication. Without this protection, the insect rapidly dies if exposed to dry air. But the most impressive superhydrophobic properties are related to the presence of setae on the body or on the legs (Figure 6c), allowing some animals to float on water or even to live underwater owing to the air spread on their body (43–48); see details in the recent and comprehensive review by Bush et al. (24). The setae often consist of tapered hairs with a length of 30 μm, a diameter of 1–10 μm, and an angle of inclination of typically 30◦ (Figure 6c). As for plants, there is a secondary texture, namely nanogrooves, whose exact role is still questioned (23). Other structures can be very different: Figure 6d shows the pattern that decorates the eye of Culex pipiens, the classical mosquito. It is very simple and well-ordered, at an impressively small scale (posts of size and height of approximately 100 nm) (40). We show further that some applications indeed require reduction of the pattern size. 3.3. Synthetic Microtextures As is pointed out above, many recent papers are devoted to the creation of microtextured surfaces with particular wetting properties (most often, superhydrophobic ones). Many techniques exist for producing such materials, even primitive ones: Approaching a piece of glass from a flame generates soot, which quickly darkens the glass. If you put water on this substrate, you will see it behaving as if it were a soft solid, rolling and bouncing off the surface! More generally, template synthesis, phase separation, all kinds of etching, crystallization, and electrospinning of microfibers were proposed to construct more elaborate materials (49–50). As a result, many different textures, from highly disordered or fractal to ordered and well-defined, were obtained (Figure 7). Most of these surfaces provide specific wetting properties, and we still have to understand which surfaces are the “best.” The answer depends on the required properties, which we now discuss. 4. HEMIWICKING Patterns on a hydrophilic solid at a scale much smaller than the capillary length (above which gravity dominates surface tension effects) can induce superhydrophilicity. We discuss above the Wenzel effect, in which the roughness enhances hydrophilicity, provided that liquid fits in the pattern (Figure 4), leaving dry the rest of the solid as in usual partial wetting (Figure 1b). However, the structures may also guide the liquid within the array they form, in a manner similar to wicking. The phenomenon that occurs here is not classical wicking but hemiwicking: As the film progresses in the microstructures, it develops an interface with air, leaving (possibly) a few dry islands behind it. We examine the conditions for observing hemiwicking, starting with the case of a single groove. We discuss how this phenomenon impacts the wetting laws and conclude with a few considerations of the dynamics of these films. www.annualreviews.org • Wetting and Roughness 79 Annu. Rev. Mater. Res. 2008.38:71-99. Downloaded from arjournals.annualreviews.org by 213.160.13.186 on 07/09/08. For personal use only. a b 10 µm c 500 nm d 5 µm 20 µm Figure 7 Different examples of synthetic microtextured surfaces. (a) The simplest possible surface, with regular micropillars. (Courtesy of M. Reyssat.) (b) A surface decorated with nanofibers. From Reference 51 (courtesy of L. Gao and T.J. McCarthy). (c) A surface planted with carbon nanotubes. From Reference 52 (courtesy of J. Bico). (d ) Mushroom pattern (with a flat hat). From Reference 53 (courtesy of G. McKinley). 4.1. Grooves As stressed above, many solids are naturally striated by grooves. Such defects can also be etched for specific purposes, such as directional wetting. We consider, for example, a rectangular groove of width w and depth δ, as sketched in Figure 8, in which we ignore the detail of the different menisci. For observing a spontaneous invasion of the groove, the solid must lower its energy by being wet (γ SL < γ SA ). But this is not enough because a liquid/vapor interface also develops at the top. dx δ w Figure 8 A liquid (in blue) invading a rectangular groove on a solid. Here, we ignore the menisci (at the liquid front and along the corners, ahead of it) and consider a progression of the liquid by a quantity dx. 80 Quéré We expect the interface to be flat (as represented in Figure 8), which minimizes the corresponding surface area. Hence, for a liquid progression in the groove by a quantity dx, the surface energies change by an amount (ignoring gravity effects) dE = (γSL − γSA )(2δ + w) dx + γ w dx. 4. Annu. Rev. Mater. Res. 2008.38:71-99. Downloaded from arjournals.annualreviews.org by 213.160.13.186 on 07/09/08. For personal use only. Using the Young equation, we find that the liquid progression is favorable (dE < 0) if we have θ < θc , 5. with cos θc = w/(2δ + w). 6. Whatever the values of w and δ, the latter quantity (which depends only on the aspect ratio δ/w) defines a number between 0 and 1, from which the angle θ c can be made explicit. If the groove is narrow and/or deep (w is small and/or δ is large), we recover the criterion (discussed above, in the context of Figure 2) of spontaneous penetration in a classical porous medium: θ c = 90◦ . In a general case (δ ∼ w), θ c is somewhere between 0◦ and 90◦ : It is more demanding to impregnate a groove than a 3-D porous medium. This is all the more true because the groove is shallow: As δ/w tends toward 0, so does θ c , meeting the criterion of complete wetting on a flat solid. 4.2. Assembly of Pillars We can have similar arguments for a solid decorated with microposts (such as in Figure 7a). We characterize such a surface by its pillar density φ S and roughness r. We show in Figure 9 the top view of an ethanol drop on/in a forest of pillars (with φ S = 0.05 and r = 2) (54). The drop is a lens, which deforms the colors generated by the regular array of microposts, and it is clearly surrounded by a film of ethanol; in this situation hemiwicking takes place. In some cases, the film conforms to the micropattern, so the film can take a square shape on a square array of microposts (55). Figure 9 Top view of an ethanol drop (with a diameter of a few millimeters) on a silicon wafer decorated by silicon microposts. Here, hemiwicking takes place, as deduced from the observation of a thick film ahead of the drop. From Reference 54 (courtesy of Chieko Ishino and Mathilde Reyssat). www.annualreviews.org • Wetting and Roughness 81 Air dx p b h Liquid Solid Figure 10 Annu. Rev. Mater. Res. 2008.38:71-99. Downloaded from arjournals.annualreviews.org by 213.160.13.186 on 07/09/08. For personal use only. Liquid film (in blue) propagating on a solid, within a forest of microposts of height h, mutual distance p, and radius b. The condition for the progression is deduced from the variation of surface energy associated with it. The impregnating front should propagate as pictured in Figure 10: The solid is coated by the liquid on a surface area proportional to r − φS , whereas the (flat) liquid/vapor interface develops on a surface area proportional to 1 − φS . For a film progressing by a distance dx (larger than the scale of the defects), the variation in surface energy per unit length perpendicular to the figure can be written as dE = (γSL − γSA )(r − φs )dx + γLV (1 − φs ) dx. 7. The progression is favorable provided (once again) that the Young angle is smaller than a critical value θ c , which depends only on the design of the solid (56): cos θc = (1 − φs )/(r − φs ). 8. Liquid invasion on a microstructured solid can thus be tuned by the geometry of the structures. For dilute defects (small φ S ), we have cos θ c ≈ 1/r: The rougher the substrate, the larger θ c , i.e., the more likely that hemiwicking occurs. For a substrate composed of disconnected defects (such as posts), the liquid front must somehow be activated to achieve the jumps sketched in Figure 10. For wetting liquids, this is made possible via the menisci, which form around each post, allowing the liquid to reach the next row. In other cases, the contact line can remain pinned in a metastable Wenzel state, and an external source of energy (such as vibrations) must be employed to nucleate a contact with the next rows of pillars. We can even imagine equilibrium situations in which the drop coexists with a wet ring of finite extension (looking a bit like a fried egg). If, for example, the energy barrier is passed owing to the action of the Laplace pressure, the progression can stop once the drop spreads enough to make its Laplace pressure too small for inducing a further motion. We can interpret the second regime (for θ < θ c ) in Figure 5 as resulting from hemiwicking. In this experiment, the solid is very rough, with a fractal structure. Even if we do not know the value of the parameter φS , we expect it to be smaller than 1, so cos θ c should be of the order of 1/r. The second regime indeed starts close to the abscissa where the Wenzel regime (of slope r) intercepts the line cos θ ∗ = 1, that is, for cos θ ≈ 1/r. The apparent angle θ ∗ then hardly depends on θ , which can be understood qualitatively: The drop sits on a composite surface consisting mainly of liquid, apart from a few solid islands. The angle θ ∗ should be very close to 0 (the value it would take if there were only liquid), but it cannot reach this value owing to the islands on which the angle is θ > 0. The number of islands should be a function of θ, which makes it difficult to produce a general theory. The value expected for the angle θ ∗ is based on Figure 11, in which we sketch the drop coexisting with the impregnating film. We consider a displacement of the contact line by a quantity dx. The solid becomes wet on a fraction of surface φS , and liquid interfaces are eliminated on a fraction 1 − φS . Because the displacement also implies an increase of the liquid/vapor interface of the drop, the total change of surface energies eventually becomes (per unit length of the line) dE = φs (γSL − γSA ) dx − (1 − φs )γ dx + γ dx cos θ ∗ . 82 Quéré 9. dx Annu. Rev. Mater. Res. 2008.38:71-99. Downloaded from arjournals.annualreviews.org by 213.160.13.186 on 07/09/08. For personal use only. Liquid θ Solid Figure 11 Drop coexisting with a film that self-propagated within the material textures. The apparent angle is obtained by considering a displacement of the contact line and computing the corresponding variation of surface energy. The minimum of E yields the apparent angle θ ∗ , as first shown by Cassie (57, 58): cos θ ∗ = 1 − (1 − φs ) cos θ. 10. For small φS , the angle θ ∗ hardly varies with θ, as observed in Figure 5 for θ < θ c . This variation is linear when the cosines of both angles are plotted, and the slope provides φS . We would deduce, for example, from Figure 5 that φS ≈ 0.15. However, the actual behavior should not be linear for a disordered surface, for which the proportion of emerged islands should itself be a function of θ (the smaller the θ, the smaller the φS ), making the actual variation θ ∗ (θ) less simple. 4.3. Dynamics of Hemiwicking The force that drives hemiwicking can be derived from Equation 7. We get, per unit length, F = −dE/dx = γ (r − φS ) (cos θ − cos θ c ), which is fixed by the quality of wetting and by the design of the surface. If wetting is complete (θ = 0), a molecular layer propagates ahead of the impregnated film, which lowers the surface energy via the suppression of liquid/vapor interfaces only, on a surface area proportional to r − 1. Hence, the force then is γ (r − 1). It only depends on the roughness and logically vanishes if the material is flat (r = 1). Owing to the small scale of the textures, this (constant) driving force is generally balanced by viscous force. This resistance to the flow should scale as ηVx, denoting x as the impregnated distance and η as the liquid viscosity. Balancing both forces, we find that the film should progress as the square root of time, with dynamics similar to the wicking Washburn law inside a porous medium (59). For rough substrates as for grooves, the liquid films indeed progress in a Washburn fashion (60–62). This is also true within forests of microposts (56). In this case, we can calculate the coefficient characterizing the dynamics (if we note x2 = Dt, D is this coefficient), allowing us to be more precise about the way the dynamics can be tuned by the design of the posts (54). For wetting liquids and posts of height h, distance p, and radius b (b p), the last paragraphs predict a wicking force scaling as γ bh/p2 . The exact form of the viscous resistance depends on the pillar height. (a) For short pillars (h p), the dissipation is fixed by the depth h of the flow (the velocity gradients take place between the bottom solid and the free liquid interface). We deduce a viscous force (per unit length) ηVx/h and thus a dynamic coefficient D ∼ (γ /η) (bh2 /p2 ), which is efficiently tuned by the pillar height. (b) For tall pillars (h p), dissipation takes place mainly around the pillars. The viscous force per pillar is ηVh (omitting an Oseen logarithmic factor), with x/p2 pillars per unit length. Hence, there is a viscous force ηVhx/p2 , which eventually yields a very simple dynamic coefficient: D ∼ γ b/η. We thus see how the design can be optimized: If for a given application a film of height h must propagate, we select pillars of height h (hemiwicking www.annualreviews.org • Wetting and Roughness 83 is a good way for setting films of a desired thickness). Then, p is chosen to maximize the speed of propagation; it will be taken of order h because there is no dynamic benefit for having more compact networks. 5. SUPERHYDROPHOBICITY Annu. Rev. Mater. Res. 2008.38:71-99. Downloaded from arjournals.annualreviews.org by 213.160.13.186 on 07/09/08. For personal use only. 5.1. Air Trapping On hydrophobic solids, the situation is of course different from that for hydrophilic solids. If the solid is rough enough, we do not expect that the liquid will conform to the solid surface, as assumed in the Wenzel model (Figure 4). Rather, air pockets should form below the liquid [this is the so-called Cassie or fakir state (48–49)], provided that the energetic cost associated with all the corresponding liquid/vapor interfaces is smaller than the energy gained not to follow the solid. This criterion can be made more quantitative by consideration of, again, pillar-like textures. If the liquid/air interfaces are assumed to be flat (which can be justified by a condition of constant Laplace pressure in the liquid, which, for defects much smaller than the drop size, can be taken as null), the wet and liquid surface areas are proportional to (r − φS ) and (1 − φS ), respectively. Hence, air pockets are favored, provided that (63–65) (r − φs ) (γSL − γSA ) > (1 − φs )γ , 11. which (through the use of the Young formula) can be reformulated as θ > θ c , with cos θc = −(1 − φs )/(r − φs ). 12. This criterion is similar to the one established for propagating a film of liquid inside the texture. Air here replaces liquid, so the critical angle expected from Equation 12 is just π minus the critical angle below which hemiwicking takes place (Equation 8). For very rough solids (r 1), this criterion is always satisfied. Then, θ c tends toward 90◦ , and θ is indeed larger than this value, because we assumed chemical hydrophobicity. For materials decorated with long hairs, for example, the roughness factor r ≈ 2πbh/p2 can typically be 5 to 10 (as deduced from Figures 6c or 7c, for example). Figure 12 confirms that the leg of Microvelia, a small bug walking on water, does not contact the liquid, as evidenced by the distance visible between the leg and its reflection (24). a b 1 cm Figure 12 (a) A Cassie state in action: Microvelia walking on water (scale bar, 1 cm). (b) Thin (hydrophobic) hairs allow the bug to be repelled by and to skate on water. From Reference 24 (courtesy of D. Hu and J.W.M. Bush). 84 Quéré dx Liquid Air Annu. Rev. Mater. Res. 2008.38:71-99. Downloaded from arjournals.annualreviews.org by 213.160.13.186 on 07/09/08. For personal use only. θ Solid Figure 13 Displacing the contact line in the Cassie regime: The energy balance must include the creation of liquid/air interfaces below the drop, as indicated by the dotted lines. The situation is more ambiguous for modest roughness factors, such as those provided by small pillar density φS . The chemical contact angle θ is typically 100◦ to 110◦ , and its cosine is slightly negative. Thus, the criterion for air trapping is not obeyed. However, Cassie states are often observed in spite of a higher interfacial energy (64, 66, 67). The air present before we place a drop can remain trapped in a metastable state, as long as the drop does not nucleate a contact with the ground surface of the solid. We discuss in Section 5.3 the metastability of Cassie states. This Cassie regime is the one of interest because, in addition to a large contact angle, it provides a small contact angle hysteresis, owing to the presence of the air cushion. As a consequence, we term superhydrophobic the only Cassie regime, which generates amazing properties such as reduced adhesion, water repellency, and slip (partially discussed in Sections 5.2 and 6.3). Because the drop sits on a mixture of solid and air, we expect a large apparent angle θ ∗ . If there is only air, Young (Equation 1, where we replace the index S by A) predicts a “contact” angle of 180◦ (i.e., no contact). Any deviation from this value tells us the proportion of solid actually contacting the liquid. The variation of interfacial energy arising from a displacement of the contact line by a quantity dx (as sketched in Figure 13) is related to the creation of new wet solid surface and liquid/vapor interfaces. The final balance can be written as dE = φs (γSL − γSA ) dx + (1 − φs )γ dx + γ dx cos θ ∗ . 13. It is crucial to assume flat liquid/vapor interfaces. Consideration of some curvature would modify the result (liquid/vapor interfaces would then have a larger surface area), suggesting that the angle should (slightly) increase as the drop gets smaller. In recent experiments, Rathgen et al. (68) analyzed the light diffracted through (transparent) Cassie materials, providing a very precise measurement of this curvature. At equilibrium, E is the minimum, which yields the apparent angle θ ∗ (69): cos θ ∗ = −1 + (1 − φs ) cos θ. 14. This description must be complemented by a (local) Young condition at each contact line (at the edge of the drop and for each liquid/vapor interface below). This condition is satisfied by the presence of edges on the posts (or more generally of large slopes on the rough material). As discussed above in the context of Figure 3 and because we have θ > 90◦ , sharp angles permit this condition. We expect stronger pinning if these edge angles are smaller: Re-entrant designs will make more robust Cassie states, and we see below (Section 5.3) that they can even induce air trapping in a hydrophilic situation (29). Equation 14 usually predicts large angles. For θ ≈ 110–120◦ and φS between 5% and 10%, we get apparent angles of 160–170◦ . In Figure 14, we see a millimetric water drop on a silicon www.annualreviews.org • Wetting and Roughness 85 Annu. Rev. Mater. Res. 2008.38:71-99. Downloaded from arjournals.annualreviews.org by 213.160.13.186 on 07/09/08. For personal use only. Figure 14 Millimetric water drop on a hydrophobic surface textured with regularly spaced micropillars. The texture acts as a diffraction grating, which induces structural colors. From Reference 70 (courtesy of M. Reyssat). substrate where silicon micropillars (similar to Figure 7a, with φS = 5%) were etched and coated with a fluoropolymer (70). The drop is like a pearl (69, 71), sitting on a solid whose iridescences reveal the regular array of “defects,” which diffracts light (structural color) (66, 72, 73). Conversely, structures much smaller than the wavelength of light yield transparency (74–76). This raises the interesting question of the smallest size generating water repellency, which remains to be solved. The smaller is φS , the larger is θ ∗ . But for a more complex topology, φS should be a function of θ (69). Conversely, the measurement of θ ∗ in a Cassie situation should provide the solid fraction φs contacting the liquid, a quantity of interest for characterizing not only wetting but also hydrodynamic slip (see Section 6.3) or any properties related to a solid/liquid contact (e.g., electrical or chemical). In the limit of small φS , we note that θ ∗ = π − ε (with ε 1), and Equation 14 rewrites to ε ≈ 2 cos (θ /2)φs1/2 , 15. whose critical behavior in φS emphasizes the difficulty for achieving a strict nonwetting situation (ε ≈ 0). However, Gao & McCarthy (51) approached, and perhaps reached, this limit (within the uncertainty of the measurements) by using nonwoven assemblies of nanofibers (Figure 7b) on which both the advancing and receding angles were 180◦ . The same authors reported similar results for pulverulent hydrophobic solids obtained by compressing commercially available lubricant (37). In both cases, the texture is submicrometric and quite regular, with smooth rounded defects, which should induce a very low hysteresis—the quantity we now discuss. 5.2. Toward Nonsticking Water The most important property of a Cassie surface is its small contact angle hysteresis. As a consequence, drops are unusually mobile, which generates novel properties, such as bouncing, that are not observed on conventional materials (77). Liquids behave to a large extent as Leidenfrost drops, for which the underlying vapor minimizes the friction. However, one generally observes on superhydrophobic materials a residual hysteresis (and thus adhesion), whose value is still debated today (78–83). A key factor in this discussion is the possibility of pinning the contact line on the 86 Quéré Annu. Rev. Mater. Res. 2008.38:71-99. Downloaded from arjournals.annualreviews.org by 213.160.13.186 on 07/09/08. For personal use only. y p 2b Figure 15 In a Cassie state, a drop is likely to pin on the edges of the defects as we displace it. The drop becomes distorted, and the energy stored in this deformation fixes the amplitude of the hysteresis. defects. Therefore, the shape of the defects or the sharpness of the edges is significant, as evidenced in a classical experiment by Öner & McCarthy (71) in which they varied the post shape. Here we restrict the discussion to the case of strong pinning on dilute defects, elucidated in 1984 by Joanny & de Gennes (84) and Pomeau & Vannimenus (85). The model is based on Figure 15, in which we see a few defects on which the line pins as we move the liquid, using a force F. The line meets φ S defects per unit area and thus, for a displacement dx, φ S dx defects per unit length. Passing each of them, an energy is stored and then released in the liquid (where it gets dissipated by viscosity) as the line depins. Hence, we have F dx = φS dx. 16. We guess that will depend on the shape of the defects (for example, complex contours will generate a higher , and thus a larger hysteresis, unlike small and rounded defects). For the sake of simplicity, we assume an equilibrium (Young) angle of 90◦ ; in addition, our defects are pillars (or disks) of radius b and mutual distance p (with b p). We thus have φ S ∼ b2 /p2 . As the line pins on a defect of size b, the drop gets distorted, as shown in Figure 15. Its tails form surfaces of zero curvature: r = b cosh (x/b); that is, for x b, r ≈ 1/2 b exp(x/b). The deformation is maximum (x = u) for the largest lateral deformation r, i.e., for the typical distance p between two defects. Hence we get u ≈ b log ( p/b). 17. The pinning force on a defect f is related to b (the line pins on the contour of each defect): f ≈ πbγ , which yields a relationship between force and deformation, f ≈ π γ u/ log ( p/b). 18. Equation 18 defines a linear spring of stiffness K = π γ /log(p/b) (84). Hence, there is an energy = 1/2Ku2 stored in the deformation. The force necessary to move the line can be written, per unit length, as F = γ (cos θr − cos θa ). 19. www.annualreviews.org • Wetting and Roughness 87 Putting together these equations, we finally get (cos θr − cos θa ) ∼ φS log(1/φS ). 20. Annu. Rev. Mater. Res. 2008.38:71-99. Downloaded from arjournals.annualreviews.org by 213.160.13.186 on 07/09/08. For personal use only. The contact angle hysteresis vanishes with the density of defects, but the presence of a logarithm in Equation 20 makes this behavior quite pathological: At small φ S , the hysteresis decreases because there are fewer defects (the linear term), but the logarithm term (slowly) diverges, making the residual hysteresis appreciable. This result seems to be in agreement with many existing data, but more remains to be done to check these models quantitatively and to extend them to more complex patterns. Hysteresis makes drops stick on solids, despite gravity field or air flow. A general calculation of the sticking force is difficult, but a simplified argument allows us to evaluate how the hysteresis enters this quantity (86). We assume that the rear half of the drop joins the solid with an angle θ r , whereas the front half meets it with an angle θ a . The capillary sticking force can be written as π γ (cos θ r – cos θ a ), denoting l as the radius of the solid/liquid contact (quasi-circular for θ 1). Assuming a geometric contact l ≈ Rε (where ε is the difference between the mean angle and π , and R is the drop radius) and using Equations 15 and 20, we find that a drop will move in the gravity field (on a vertical window pane) provided that φs3/2 log(1/φS ) < R2 κ 2 , 21. in which we introduce the capillary length κ −1 = (γ /ρg)1/2 (2.7 mm for water) and ignore all the numerical coefficients. Once again, the density of defects is crucial for driving the wetting properties (here the degree of adhesion of a drop on a solid). As we could guess, small densities are required for suppressing adhesion. However, such a limit also weakens the stability of the Cassie state, which we now discuss. 5.3. Metastable Cassie States As stressed above, drops are often observed in a Cassie state in spite of a smaller Wenzel energy. As a consequence, these metastable Cassie states are fragile (70). Figure 16 shows two millimetric water drops on a microtextured substrate (with a pillar density φ S of 1% and pillar height h of 12 μm). The first drop is placed without any impact, and the second drop is released from a few centimeters such that it meets the solid with a velocity of a few tens of centimeters per second. Figure 16 Millimetric water drops of the same volume on a superhydrophobic substrate covered with dilute pillars (φS = 0.01, h = 2 μm). The drop on the right was thrown on the substrate, whereas the one on the left was carefully deposited. As observed, Cassie and Wenzel states (left and right, respectively) can coexist. From Reference 70 (courtesy of M. Callies). 88 Quéré Annu. Rev. Mater. Res. 2008.38:71-99. Downloaded from arjournals.annualreviews.org by 213.160.13.186 on 07/09/08. For personal use only. The first globule is a Cassie drop (we even see light passing below it), whereas the second one is in a Wenzel state: On this substrate of small roughness (r ≈ 1.1), the contact angle is close to the Young angle, slightly larger than 90◦ in this example. When the Wenzel state is the less energetic one, any perturbation of a Cassie drop can provoke its transition to this state. Conversely, a Wenzel drop is firmly bound to its stable configuration. It is of practical importance to quantify the robustness of metastable Cassie states and to understand the conditions provoking impalement. Obviously, an energy barrier must be overcome to find the ground state (87–90). One can evaluate this barrier by considering the penetration of a Cassie drop. Assuming unchanged liquid/vapor interfaces as the drop sinks, the only change in surface energy corresponds to the (unfavorable) wetting of the posts’ walls. This implies a (positive) energy per unit area E = (γ SL − γ SA )(r − 1) = −γ (r − 1) cos θ (this quantity becomes negative in a hydrophilic situation). We thus find an energy barrier E ≈ 2πbh/p2 γ cos θ. It is proportional to the pillar height h, which appears as a natural parameter for tuning this quantity. The energy barrier E is generally too large to be overcome by thermal energy (we need defects of molecular dimensions to get E of the order of kT ). However, the energy can be supplied by pressing on the drop (66, 90), by vibrating the substrate (91), or by an impact (92, 93). Indeed, the higher the posts, the larger is the resistance to impalement. Once the liquid penetrates the texture, it remains strongly pinned, and the “printed” drop is even able to conform to the network of microposts (92). Sbragaglia et al. (94) described the dynamics of the Cassie/Wenzel transition as very quick and following a zipping mechanism: One row of cavities gets filled (in a time of approximately 10 μs on a length of 100 μm!) before jumping to the next row. This process is somehow reminiscent of the progression in a groove sketched in Figure 8. The surface force here involves the creation of wet surfaces and the suppression of suspended liquid/air interfaces, whereas the resisting force should be viscous. We thus expect a Washburn law for the progression (see Section 4.3), which can be extremely quick at the small scale of the phenomenon. For a pattern composed of posts whose height h is comparable to the pitch p (of approximately 10 μm), the typical time for invading a row of length x scales as ηx2 /γ h, of the order of 10 μs for water and x = 100 μm. The way in which solid/liquid contacts nucleate for triggering the transition is interesting. Interfaces above the air pockets are curved, fitting the global curvature of the drop (Figure 17). If the drop becomes small enough, the liquid can reach the underlying solid and then propagate. The size of a drop should thus impact the drop’s wetting state. Indeed, small droplets are more likely to be in a Wenzel state than are large droplets (66), which can also be evidenced by observing the evaporation of a drop. Then, the drop’s size varies continuously, and investigators have reported the existence of a critical radius below which the drop suddenly falls into the Wenzel regime (95, 96). Following the notations in Figure 17, the interface curvature scales as δ/p2 (for h < p). Equating it with the drop curvature 1/R yields the depth of penetration δ of the interface inside the texture: δ ∼ p2 /R; the smaller the drop, the larger δ is. When it becomes of the order of the pillar height δ h p Figure 17 The liquid/vapor interface is curved, owing to the curvature of the drop, but the interface also can be curved if we apply pressure to the drop. δ denotes the lowering of the interface below the top of the posts. www.annualreviews.org • Wetting and Roughness 89 h, a solid/liquid contact can nucleate on the bare substrate and propagate if the Cassie state is metastable. This implies a critical radius for a Cassie drop scaling as (96) R∗ ∼ p 2 / h. 22. Annu. Rev. Mater. Res. 2008.38:71-99. Downloaded from arjournals.annualreviews.org by 213.160.13.186 on 07/09/08. For personal use only. Note that, as deduced from Figure 3, depinning from the edge will also occur if the drop radius is smaller than p/|cos θ|, a limiting condition for modest chemical hydrophobicity. Here we assume that this second condition is screened by the first one, i.e., p > h/|cos θ|. The radius R ∗ can be much larger than p if h < p. The Cassie state will be all the more robust because this critical radius is small (no drops fall in the Wenzel drops, except invisible ones). One can achieve this in two ways: either by making h large, using micro- or nanofibers for decorating the solids (see Figures 6c and 7c) (96), or by reducing both p and h. Miniaturizing the pattern size enhances the resistance of the Cassie state, which may explain the existence of such small scales in many natural materials. Jiang and coauthors (40) recently reported that the eye of C. pipiens apparently remains dry even if exposed to tiny drops, as encountered in the foggy and moist environments where these mosquitoes usually circle. Figure 18 is a close-up of C. pipiens after the mosquito passed through an aerosol of water droplets. Water condenses on most of the animal, but the eyes remain dry, which of course preserves its vision (renowned as excellent). Figure 6d displays a microphotograph of the textures observed on the surface of the eye; these are remarkably small, with p ∼ h ∼ 100 nm. With these values, we get R ∗ ∼ 100 nm: A drop at this scale not only is invisible but also evaporates quasi-instantaneously. In a cloud, drops are quite polydisperse, with a typical radius of 10 μm—such small drops would impale on most microtextured surfaces but might resist the Wenzel transition for the nanopattern worn by the mosquito’s eye. Other promising metastable Cassie states are those obtained on hydrophilic materials with a particular design. Oils having contact angles of the order of 40◦ can be suspended on special textures, producing an increase of the angle by approximately 100◦ (the superoleophobic effect) (53, 97–102). As shown by Herminghaus, fakir wetting drops are possible on overhangs or reentrant angles, that is, sites where a hydrophilic Young condition can be satisfied (29). Fibers and defects with overhangs do provide quite robust oleophobicity and are able to resist a Wenzel transition by pressing on the liquid with the Laplace pressure related to the size of the “holes” Figure 18 Close-up of Culex pipiens after exposure to water aerosol. Droplets condense on the antennas, but the black eyes remain dry—a condition for preserving the eyesight of the insect. From Reference 40 (courtesy of L. Jiang). 90 Quéré Annu. Rev. Mater. Res. 2008.38:71-99. Downloaded from arjournals.annualreviews.org by 213.160.13.186 on 07/09/08. For personal use only. at the surface. With structures with pronounced re-entrant profiles, such as those displayed in Figure 7d (where the pattern evokes a mushroom, or a hoodoo), Tuteja et al. (53) spectacularly reported that even a liquid as wetting as octane can be suspended in a Cassie state with advancing and receding contact angles as high as 160◦ and 140◦ , respectively. 6. SPECIAL PROPERTIES As we see above, hydrophobic Cassie materials generate high contact angles and small hysteresis, ideal conditions for making water drops very mobile. We conclude this article by discussing a few special properties potentially generated by these surfaces, such as anisotropy, wettability switches, and slip. 6.1. Anisotropy Many natural (Figure 6a,b,d ) and synthetic (Figure 7a,b,c,d ) textures are isotropic. However, it can be interesting to design directional structures, such as arrays of parallel grooves or microwrinkles, that consequently generate anisotropic wetting, in particular, in the Cassie regime (69, 103–105). Owing to a differential pinning of contact lines, the contact angles (and the hysteresis) are quite different along and perpendicular to the grooves. Axial motion is preferred, and such designs are appropriate when liquid must be guided. There are examples of such patterns in nature (38, 41, 106), as in Figure 19, which shows the scales covering the wings of the butterfly Papilio ulysses. Both the arrangement and microtexture of the tiles contribute to the directionality of this material. Another kind of anisotropy is exploited by water striders (see its inclined hairs in Figure 6c): Striders strike the surface perpendicular to the grooves, which generates a large contact force, before swinging the legs by 90◦ to align them in the direction of the motion for skating. Motion will arise from alternating pinning and gliding events (24). Figure 19 The wings of Papilio ulysses. The way the tiles are displayed together with the detail of the texture confer anisotropy to the texture. From Reference 106 (courtesy of S. Berthier). 100 µm www.annualreviews.org • Wetting and Roughness 91 6.2. Wettability Switches Annu. Rev. Mater. Res. 2008.38:71-99. Downloaded from arjournals.annualreviews.org by 213.160.13.186 on 07/09/08. For personal use only. Textured surfaces undergo a brutal change of wettability as the contact angle exceeds 90◦ (Figure 5). This behavior can be exploited to achieve superhydrophobic/superhydrophilic switches. Different physicochemical effects affecting the solid wettability, such as light on photocatalytic textures (107, 108) or heat for temperature-sensitive coatings (109), can be used for triggering the transition. The comprehensive review by Feng & Jiang (50) provides details. Electric field, as we have learned from Lippman, also affects wettability. Applying a voltage across a drop lowers its contact angle, and this effect is amplified on a textured surface. A drop with an angle of approximately 160◦ can nearly spread under the action of modest voltages (approximately 10 V) (73). However, the liquid gets irreversibly pinned in the superhydrophilic state (or in any state in which it intimately contacts the rough solid), contrasting with our expectations for a switch. Krupenkin et al. (110) proposed to use a short and intense pulse of current (through a thin conductive layer on the sample), which evaporates the liquid close to the surface, hence restoring a Cassie-suspended state. More generally, there is today no clear example of a Wenzel state (even potentially metastable) spontaneously transforming into a Cassie state. This situation is detrimental as a vapor condenses: This naturally forces a Wenzel situation, which often evolves toward mixed and ambiguous Cassie/Wenzel situations (111–115). Much remains to be done to achieve genuine antidew materials. 6.3. Giant Slip Experiments confirm that superhydrophobic materials can provide slippage as water flows on them, provided that these materials are in the Cassie state. The amplitude of the phenomenon is captured by the so-called slip length, which is the extrapolated distance on which the liquid velocity vanishes (Figure 20). First introduced by Navier, this length is generally molecular. However, it can become of the order of 10 nm on flat hydrophobic solids, as experiments using a surface force apparatus show (116). Similar to what we saw for wetting, this hydrophobic behavior can be dramatically enhanced if the solid is rough (117). Ou et al. (118) reported micrometric slip lengths from measurements of the pressure necessary to drive a given flux of water along a square channel striated with microgrooves. For a Poiseuille flow, the flux varies as W 4 ∇p/η, denoting ∇p as the pressure gradient along the flow and W as the width and depth of the channel. For a large slip (λ > W ), the flux instead scales as λW 3 ∇p/η. Slip at the wall reduces the pressure gradient by a factor λ/W. Ou et al. found pressure reduction by approximately 40%, suggesting slip lengths of the order of 10 μm (see also References 119–121). λ Figure 20 The slip length λ is the distance inside the solid for which the velocity profile of a flowing liquid vanishes. 92 Quéré Figure 21 2 Slip length on superhydrophobic patterns of nanotubes of constant density, but with a varying pitch. In the Cassie state (circles), the slip length is micrometric and increases with the pitch. In the Wenzel state (squares), there is no measurable slip. (µm) Annu. Rev. Mater. Res. 2008.38:71-99. Downloaded from arjournals.annualreviews.org by 213.160.13.186 on 07/09/08. For personal use only. 1.5 1 0.5 0 0 2 4 p (µm) 6 Using microparticle imaging velocimetry, Joseph et al. (122) directly measured slip lengths for water flowing on hydrophobic carbon nanotubes. As shown in Figure 21, the slip length dramatically depends on the nature of hydrophobicity. In a Wenzel state (induced by pressing on the liquid), there is no measurable slip, in agreement with the paper of Richardson (123), who stipulated that roughness kills any (potential) slippage if the liquid conforms to it. Conversely, micrometric slip lengths are observed in the Cassie state, and increase linearly with the post distance p, in these experiments performed at constant φ S . This slip can be dramatically damped if liquid/air interfaces are curved (as in Figure 17), owing, for example, to a pressure exerted on the liquid (124). It is natural to expect a large slip in a Cassie situation: Liquid glides on air, owing to the large viscosity ratio between water and air (typically a factor of 100). However, part of the liquid contacts the top of the posts, which limits the total slip on the surface. We can make this argument more quantitative, following a recent analysis by Ybert et al. (125). For a flow of typical velocity V, the size “affected” by the presence of a post should scale as b, the post radius (b p, the pitch of the post array). The friction force per pillar, and thus per surface area p2 , should scale as ηVb, denoting η as the liquid viscosity. This yields a viscous stress σ ∼ ηVb/p2 . This stress dominates the one arising from the underlying air flow, provided that ηVb/p2 > ηa V/h, in which we introduced the air viscosity ηa . Hence there is a geometric requirement, bh/p2 > ηa /η, which can be achieved by adjusting the post height h. With water, ηa /η is of the order of 10−2 , and the latter criterion will be satisfied with posts of characteristics b = 1 μm, h = 10 μm, and p = 10 μm, for which the factor bh/p2 is 10−1 . As seen in Figure 20, the stress σ can also be written as ηV/λ, from which we deduce an effective slip length λ: 1/2 λ ∼ p 2 /b ∼ p/φS . 23. For a constant pillar density φ S , λ is linear with the pillar spacing p, as seen in Figure 21. For b p (φ S 1), we expect very large slip lengths, compared with the values found on flat solids. λ will typically be between a few p, i.e., 1–10 μm (as reported experimentally), 1000 times larger than the slip length on a flat hydrophobic solid! If air friction dominates the pillar friction (h < ηa p2 /ηb), the slip length becomes λ ∼ ηh/ηa , which can be very large as well. The contact angle θ ∗ is also determined by the density φ S (Equation 14), so θ ∗ and λ should be correlated. At small φ S , we note θ ∗ = π − ε. Introducing Equation 22 in Equation 15 [where www.annualreviews.org • Wetting and Roughness 93 2 cos (θ /2) is taken of order unity] yields λ ∼ p/ε, 24. Annu. Rev. Mater. Res. 2008.38:71-99. Downloaded from arjournals.annualreviews.org by 213.160.13.186 on 07/09/08. For personal use only. which directly connects the quality of the nonwetting with the amplitude of the slip. It is of obvious practical interest to maximize the slip and thus to work with low post densities. However, we see that the liquid in this limit is likely to sink inside the texture, provoking a complete failure of the slip properties. There again, there is an optimum to find, and the design to be given to the microstructures might itself be questioned. The design can also induce slip anisotropy. If grooves are considered, for example, slip is expected to be larger along the grooves than perpendicular to them. One can prove that the slip length (which is again of the order of p, the distance between grooves) differs by a factor of two in both directions (126–128), reflecting the factor-of-two difference of the viscous force on a slender object (such as a cylinder) in both directions. 7. CONCLUSION Textured surfaces, which provide superwetting, superslip, and superhydrophobicity, are supersurfaces. But so what? After a decade of intense research, we have hundreds of materials for which drops behave in the eccentric ways we described, but there is no real large-scale application (contrasting with many other smart surfaces, such as self-cleaning ones, for example). The fragility of these materials (in both a mechanical sense and a thermodynamic sense) seems to limit industrialization. Conversely, the textures in the natural world can be repaired and protected from contamination (insects spend an appreciable part of their existence grooming their legs). However, there may be short-term applications for synthetic materials, such as disposable devices (for microfluidics) or temporary surface treatments with microbeads (sticking on the material and forming there the desired microstructures)—both cases in which aging is not a major obstacle. On a more fundamental point of view, many interesting questions remain unsolved, among which we select (a) the question of the reduction of size—it would be useful to quantify how the wetting properties vary as the size of microtextures vanishes; (b) the question of optimization: What is the “best” of the microstructures, according to the searched application? We need to define a battery of tests to build a classification of the existing textures, before going further to more detailed models; (c) the search for new properties, such as antidew, for which the use of flexible microfibers seems promising; and (d ) a quantitative understanding of contact angle hysteresis, which remains to be fully done. As we see, wetting and roughness should continue to play their entertaining game with each other in the coming years! DISCLOSURE STATEMENT The author is not aware of any biases that might be perceived as affecting the objectivity of this review. ACKNOWLEDGMENTS It is a real pleasure to thank Anne-Laure Biance, José Bico, Aurélie Lafuma, Mathilde Reyssat, and Denis Richard for the many exchanges on pearl drops. I also thank Yong Chen, Chieko Ishino, Anne Pépin, and Ko Okumura for stimulating collaborations. I am very grateful to Hervé Arribart, Serge Berthier, Lydéric Bocquet, John Bush, Christophe Clanet, Lichao Gao, Stephan Herminghaus, L. Mahadevan, Thomas McCarthy, Glen McHale, Gareth McKinley, Lei Jiang, 94 Quéré Christoph Neinhuis, and Julia Yeomans for precious discussions and providing documents for this report. Finally, the help of A. Dechy and V. Dolies regarding the pictures was greatly appreciated. Annu. Rev. Mater. Res. 2008.38:71-99. Downloaded from arjournals.annualreviews.org by 213.160.13.186 on 07/09/08. For personal use only. LITERATURE CITED 1. Young T. 1805. An essay on the cohesion of fluids. Philos. Trans. R. Soc. A 95:65–87 2. Leidenfrost JG. 1996. On the fixation of water in diverse fire. Int. J. Heat Mass Trans. 9:1153–66 3. Shafrin EG, Zisman WA. 1964. Upper limits to the contact angles of liquids on solids. In Contact Angle, Wettability and Adhesion: Advances in Chemistry Series, Vol. 43, ed. RF Gould, pp. 145–57. Washington, DC: Am. Chem. Soc. 4. De Gennes PG. 1985. Wetting: statics and dynamics. Rev. Mod. Phys. 57:827–63 5. Furmidge CGL. 1962. Studies at phase interfaces. 1. The sliding of liquid drops on solid surfaces and a theory for spray retention. J. Colloid Sci. 17:309–314 6. Dussan VEB, Chow RTP. 1983. On the ability of drops or bubbles to stick to nonhorizontal surfaces of solids. J. Fluid Mech. 137:1–29 7. Oliver JF, Huh C, Mason SG. 1977. Resistance to spreading of liquids by sharp edges. J. Colloid Interface Sci. 59:568–81 8. Johnson RE, Dettre RH. 1964. Contact angle hysteresis. I. Study of an idealized rough surface. In Contact Angle, Wettability and Adhesion: Advances in Chemistry Series, Vol. 43, ed. RF Gould, pp. 112–35. Washington, DC: Am. Chem. Soc. 9. Wenzel RN. 1936. Resistance of solid surfaces to wetting by water. Ind. Eng. Chem. 28:988–94 10. Swain PS, Lipowsky R. 1998. Contact angles on heterogeneous surfaces: a new look at Cassie’s and Wenzel’s laws. Langmuir 14:6772–80 11. Wolansky G, Marmur A. 1998. The actual contact angle on a heterogeneous rough surface in three dimensions. Langmuir 14:5292–97 12. Wolansky G, Marmur A. 1999. Apparent contact angles on rough surfaces: the Wenzel equation revisited. Colloids Surf. A 156:381–88 13. Borgs C, de Coninck J, Kotecký R, Zinque M. 1995. Does the roughness of the substrate enhance wetting? Phys. Rev. Lett. 74:2292–94 14. Checco A, Guenoun P, Daillant J. 2003. Nonlinear dependence of the contact angle of nanodroplets on contact line curvature. Phys. Rev. Lett. 91:186101 15. Onda T, Shibuichi S, Satoh N, Tsujii K. 1996. Super-water-repellent fractal surfaces. Langmuir 12:2125– 27 16. Shibuichi S, Onda T, Satoh N, Tsujii K. 1996. Super water-repellent surfaces resulting from fractal structure. J. Phys. Chem. 100:19512–17 17. Norton FJ. 1945. U.S. Patent No. 2,386,259 18. Gao L, McCarthy TJ. 2006. “Artificial lotus leaf ” prepared using a 1945 patent and a commercial textile. Langmuir 22:5998–6000 19. Neinhuis C, Barthlott W. 1997. Characterization and distribution of water-repellent, self-cleaning plant surfaces. Ann. Bot. 79:667–77 20. Barthlott W, Neinhuis C. 1997. Purity of the sacred lotus, or escape from contamination in biological surfaces. Planta 202:1–8 21. Wagner T, Neinhuis C, Barthlott W. 1996. Wettability and contaminability of insect wings as a function of their surface sculptures. Acta Zool. 77:213–25 22. Parker AR, Lawrence CR. 2001. Water capture by a desert beetle. Nature 414:33–34 23. Gao X, Jiang L. 2004. Biophysics: water-repellent legs of water striders. Nature 432:36 24. Bush JWM, Hu D, Prakash M. 2008. The integument of waterwalking arthropods: form and function. Adv. Insect Physiol. 34:117–92 25. Pliny the Elder. 2004. Natural History: A Selection, transl. ed. JF. Healey. London: Penguin. Rev. ed. 26. Dufour L. 1833. Recherches Anatomiques et Physiologiques sur les Hémiptères Accompagnées de Considérations Relatives à l’Histoire Naturelle et à la Classification de Ces Insectes. Extr. Mém. Savans Etrang., Tome IV. Paris: Impr. Bachelier www.annualreviews.org • Wetting and Roughness 95 27. 28. 29. 30. 31. Annu. Rev. Mater. Res. 2008.38:71-99. Downloaded from arjournals.annualreviews.org by 213.160.13.186 on 07/09/08. For personal use only. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 96 Quéré Fogg GE. 1944. Diurnal fluctuation in a physical property of leaf cuticle. Nature 154:515 Cassie ABD, Baxter S. 1945. Large contact angles of plant and animal surfaces. Nature 155:21–22 Herminghaus S. 2000. Roughness-induced nonwetting. Europhys. Lett. 52:165–70 Minko S, Muller M, Motornov M, Nitschke M, Grundke K, et al. 2003. Two-level structured self-adaptive surfaces with reversibly tunable properties. JACS 125:3896–900 Patankar NA. 2004. Mimicking the lotus effect: influence of double roughness structures and slender pillars. Langmuir 20:8209–213 Shirtcliffe NJ, McHale G, Newton MI, Chabrol C, Perry CC. 2004. Dual-scale roughness produces unusually water-repellent surfaces. Adv. Mater. 16:1929–32 Shirtcliffe NJ, McHale G, Newton MI, Perry CC. 2005. Wetting and wetting transitions on copper-based superhydrophobic surfaces. Langmuir 21:937–43 Gao L, McCarthy TJ. 2006. The “lotus effect” explained: two reasons why two length scales of topography are important. Langmuir 22:2966–67 Bhushan B, Jung YC. 2006. Micro- and nanoscale characterization of hydrophobic and hydrophilic leaf surfaces. Nanotechnology 17:2758–72 Nosonovsky M. 2007. Multiscale roughness and stability of superhydrophobic biomimetic interfaces. Langmuir 23:3157–61 Gao L, McCarthy TJ. 2007. A commercially available perfectly hydrophobic material. Langmuir 23:9125– 27 Wang R, Hashimoto K, Fujishima A, Chikuni M, Kojima E, et al. 1999. Photogeneration of highly amphiphilic TiO2 surfaces. Adv. Mater. 10:135–38 Wagner P, Furstner R, Barthlott W, Neinhuis C. 2003. Quantitative assessment to the structural basis of water repellency in natural and technical surfaces. J. Exp. Bot. 54:1295–1303 Gao XF, Yan X, Yao X, Xu L, Zhang K, et al. 2007. The dry-style antifogging properties of mosquito compound eyes and artificial analogues prepared by soft lithography. Adv. Mater. 19:2213–15 Bormashenko E, Bormashenko Y, Stein T, Whyman G, Bormashenko E. 2007. Why do pigeon feathers repel water? Hydrophobicity of pennae, Cassie–Baxter wetting hypothesis and Cassie–Wenzel capillarityinduced wetting transition. J. Colloid Interface Sci. 311:212–16 Lee W, Jin MK, Yoo WC, Lee JK. 2004. Nanostructuring of a polymeric substrate with well-defined nanometer-scale topography and tailored surface wettability. Langmuir 20:7665–69 Thorpe WH. 1950. Plastron respiration in aquatic insects. Biol. Rev. 25:344–90 Noble-Nesbitt J. 1963. Transpiration in Podura aquatica L. (Collembola, Isotomidae) and the wetting properties of its cuticle. J. Exp. Biol. 40:681–700 Heckman CW. 1983. Comparative morphology of arthropod exterior surfaces with the capability of binding a film of air underwater. Int. Rev. Gesamten Hydrobiol. Hydrogr. 68:715–36 Vogel S. 2006. Living in a physical world. VIII. Gravity and life in water. J. Biosci. 30:309–22 Marmur A. 2006. Underwater superhydrophobicity: theoretical feasibility. Langmuir 22:1400–2 Shirtcliffe NJ, McHale G, Newton MI, Perry CC, Pyatt AJ. 2006. Plastron properties of a superhydrophobic surface. Appl. Phys. Lett. 89:104600 Nakajima A, Hashimoto K, Watanabe T. 2001. Recent studies on superhydrophobic films. Monatshefte Chim. 132:31–41 Feng X, Jiang L. 2006. Design and creation of superwetting/antiwetting surfaces. Adv. Mater. 18:3063–78 Gao L, McCarthy TJ. 2006. A perfectly hydrophobic surface. JACS 128:9052–53 Lau KKS, Bico J, Teo KBK, Chhowalla M, Amaratunga GAJ, et al. 2003. Superhydrophobic carbon nanotube forests. Nano Lett. 3:1701–5 Tuteja A, Choi W, Ma M, Mabry JM, Mazzella SA, et al. 2007. Designing superoleophobic surfaces. Science 318:1618–22 Ishino C, Reyssat M, Reyssat E, Okumura K, Quéré D. 2007. Wicking within forests of micropillars. EPL 79:56005 Courbin L, Denieul E, Dressaire E, Ajdari A, Roper M, et al. 2007. Imbibition by polygonal spreading on microdecorated surfaces. Nat. Mater. 6:661–64 Bico J, Tordeux C, Quéré D. 2001. Rough wetting. Europhys. Lett. 55:214–20 Annu. Rev. Mater. Res. 2008.38:71-99. Downloaded from arjournals.annualreviews.org by 213.160.13.186 on 07/09/08. For personal use only. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. 88. 89. 90. Cassie ABD, Baxter S. 1944. Wettability of porous surfaces. Trans. Faraday Soc. 40:546–50 Cassie ABD. 1948. Contact angles. Discuss. Faraday Soc. 3:11–16 Washburn EW. 1921. The dynamics of capillary flow. Phys. Rev. 17:273–83 Cazabat AM, Cohen-Stuart MAC. 1986. Dynamics of wetting: effects of surface roughness. J. Phys. Chem. 90:5845–49 Apel-Paz M, Marmur A. 1999. Spreading of liquids on rough surfaces. Colloids Surf. A 146:273–79 Rye RR, Mann JA, Yost FG. 1996. The flow of liquids in surface grooves. Langmuir 12:555–65 Bico J, Thiele U, Quéré D. 2002. Wetting of textured surfaces. Colloids Surf. A 206:41–46 Patankar N. 2003. On the modeling of hydrophobic contact angles on rough surfaces. Langmuir 19:1249– 53 Alberti G, DeSimone A. 2005. Wetting of rough surfaces: a homogenization approach. Proc. R. Soc. London Ser. A 461:79–97 Lafuma A, Quéré D. 2003. Superhydrophobic states. Nat. Mater. 2:457–60 Marmur A. 2004. The lotus effect: superhydrophobicity and metastability. Langmuir 20:3517–19 Rathgen H, Sugiyama K, Ohl CD, Lohse D, Mugele F. 2007. Nanometer-resolved collective micromeniscus oscillations through optical diffraction. Phys. Rev. Lett. 99:214501 Bico J, Marzolin C, Quéré D. 1999. Pearl drops. Europhys. Lett. 47:220–26 Callies M, Quéré D. 2005. On water repellency. Soft Matter 1:55–61 Öner D, McCarthy TJ. 2000. Ultrahydrophobic surfaces. Effects of topography length scales on wettability. Langmuir 16:7777–82 Gu ZZ, Uetsuka H, Takahashi K, Nakajima R, Onishi H, et al. 2003. Structural color and the lotus effect. Angew. Chem. Int. Ed. 42:894–97 Krupenkin TN, Taylor JA, Schneider TM, Yang S. 2004. From rolling ball to complete wetting: the dynamic tuning of liquids on nanostructured surfaces. Langmuir 20:3824–27 Tadanaga K, Katata N, Minami T. 1997. Formation process of superwater-repellent Al2 O3 coating films with high transparency by the sol-gel method. J. Am. Ceram. Soc. 80:1040–42 Nakajima A, Fujishima A, Hashimoto K, Watanabe T. 1999. Preparation of transparent superhydrophobic boehmite and silica films by sublimation of aluminum acetylacetonate. Adv. Mater. 11:1365–68 Nakajima A, Hashimoto K, Watanabe T, Takai K, Yamauchi G, et al. 2000. Transparent superhydrophobic thin films with self-cleaning properties. Langmuir 16:7044–47 Richard D, Quéré D. 2000. Bouncing water drops. Europhys. Lett. 50:769–75 Extrand CW. 2002. Model for contact angles and hysteresis on rough and ultraphobic surfaces. Langmuir 18:7991–99 McHale G, Shirtcliffe NJ, Newton MI. 2004. Contact-angle hysteresis on superhydrophobic surfaces. Langmuir 20:10146–49 Extrand CW. 2006. Designing for optimal water repellency. Langmuir 22:1711–14 Gao LC, McCarthy TJ. 2006. Contact angle hysteresis explained. Langmuir 22:6234–37 Kusumaatmaja H, Yeomans JM. 2007. Modeling contact angle hysteresis on chemically patterned and superhydrophobic surfaces. Langmuir 23:6019–32 DeSimone A, Grunewald N, Otto F. 2007. A new model for contact angle hysteresis. Netw. Heterog. Media 2:211–25 Joanny JF, de Gennes PG. 1984. A model for contact angle hysteresis. J. Chem. Phys. 81:552–62 Pomeau Y, Vannimenus J. 1985. Contact angle on heterogeneous surfaces: weak heterogeneities. J. Colloid Interface Sci. 104:477–88 Quéré D, Azzopardi MJ, Delattre L. 1998. Drops at rest on a tilted plane. Langmuir 14:2213–16 Patankar NA. 2004. Transition between superhydrophobic states on rough surfaces. Langmuir 20:7097– 102 Ishino C, Okumura K, Quéré D. 2004. Wetting transitions on rough surfaces. Europhys. Lett. 68:419–25 Barbieri L, Wagner E, Hoffmann P. 2004. Water wetting transition parameters of perfluorinated substrates with periodically distributed flat–top microscale obstacles. Langmuir 23:1723–34 Carbone G, Mangialardi L. 2005. Hydrophobic properties of a wavy rough substrate. Eur. Phy. J. E 16:67–76 www.annualreviews.org • Wetting and Roughness 97 Annu. Rev. Mater. Res. 2008.38:71-99. Downloaded from arjournals.annualreviews.org by 213.160.13.186 on 07/09/08. For personal use only. 91. Bormashenko E, Pogreb R, Whyman G, Bormashenko Y, Erlich M. 2007. Vibration-induced CassieWenzel wetting transition on rough surfaces. Appl. Phys. Lett. 90:201917 92. Reyssat M, Pépin A, Marty F, Chen Y, Quéré D. 2006. Bouncing transitions on microtextured materials. Europhys. Lett. 74:306–12 93. Bartolo D, Bouamrirene F, Verneuil E, Buguin A, Silberzan P, et al. 2006. Bouncing or sticky droplets: impalement transitions on superhydrophobic micropatterned surfaces. Europhys. Lett. 74:299–305 94. Sbragaglia M, Peters AM, Pirat C, Borkent BM, Lammertink RGH, et al. 2007. Spontaneous breakdown of superhydrophobicity. Phys. Rev. Lett. 99:156001 95. McHale G, Aqil S, Shirtcliffe NJ, Newton MI, Erbil HY. 2005. Analysis of droplet evaporation on a superhydrophobic surface. Langmuir 21:11053–60 96. Reyssat M, Yeomans JM, Quéré D. 2008. Impalement of fakir drops. EPL 81:26006 97. Morra M, Occhiello E, Garbassi F. 1989. Contact angle hysteresis in oxygen plasma treated poly(tetrafluoroethylene). Langmuir 5:872–76 98. Chen W, Fadeev AY, Hsieh MC, Öner D, Youngblood J, et al. 1999. Ultrahydrophobic and ultralyophobic surfaces: some comments and examples. Langmuir 15:3395–99 99. Teare DOH, Spanos CG, Ridley P, Kinmond EJ, Roucoules V, et al. 2002. Pulsed plasma deposition of superhydrophobic nanospheres. Chem. Mater. 14:4566–71 100. Woodward I, Schofield WCE, Roucoules V, Badyal JPS. 2003. Super-hydrophobic surfaces produced by plasma fluorination of polybutadiene films. Langmuir 19:3432–38 101. Otten A, Herminghaus S. 2004. How plants keep dry: a physicist’s point of view. Langmuir 20:2405–8 102. Ahuja A, Taylor JA, Lifton V, Sidorenko AA, Salamon TR, et al. 2008. Nanonails: a simple geometrical approach to electrically tunable superlyophobic surfaces. Langmuir 24:9–14 103. Yoshimitsu Z, Nakajima A, Watanabe T, Hashimoto K. 2002. Effects of surface structure on the hydrophobicity and sliding behavior of water droplets. Langmuir 18:5818–22 104. Chen Y, He B, Lee JH, Patankar NA. 2005. Anisotropy in the wetting of rough surfaces. J. Colloid Interface Sci. 281:458–64 105. Chung JY, Youngblood JP, Stafford CM. 2007. Anisotropic wetting on tunable microwrinkled surfaces. Soft Matter 3:1163–69 106. Berthier S. 2007. Iridescences: The Physical Colors of Insects. New York: Springer 107. Zhang XT, Sato O, Fujishima A. 2004. Water ultrarepellency induced by nanocolumnar ZnO surface. Langmuir 20:6065–67 108. Feng X, Feng L, Jin M, Zhai J, Jiang L, et al. 2004. Reversible superhydrophobicity to superhydrophilicity transition of aligned ZnO nanorod films. J. Am. Chem. Soc. 126:62–63 109. Sun T, Wang G, Feng L, Liu B, Ma Y, et al. 2004. Reversible switching between superhydrophilicity and superhydrophobicity. Chem. Int. Ed. 43:357–60 110. Krupenkin TN, Taylor JA, Wang EN, Kolodner P, Hodes P, et al. 2007. Reversible wetting-dewetting transitions on electrically tunable superhydrophobic nanostructured surfaces. Langmuir 23:9128–33 111. Nahre RD, Beysens DA. 2004. Nucleation and growth on a superhydrophobic grooved surface. Phys. Rev. Lett. 93:076103 112. Cheng YT, Rodak DE. 2005. Is the lotus leaf superhydrophobic? Appl. Phys. Lett. 86:144101 113. Cheng YT, Rodak DE, Angelopoulos A, Gacek T. 2005. Microscopic observations of condensation of water on lotus leaves. Appl. Phys. Lett. 86:194112 114. Wier KA, McCarthy TJ. 2006. Condensation on ultrahydrophobic surfaces and its effect on droplet mobility: Ultrahydrophobic surfaces are not always water repellant. Langmuir 22:2433–36 115. Dorrer C, Rühe J. 2007. Condensation and wetting transitions on microstructured ultrahydrophobic surfaces. Langmuir 23:3820–24 116. Cottin-Bizonne C, Cross B, Steinberger A, Charlaix E. 2005. Boundary slip on smooth hydrophobic surfaces: intrinsic effects and possible artifacts. Phys. Rev. Lett. 94:056102 117. Cottin-Bizonne C, Barrat JL, Bocquet L, Charlaix E. 2003. Low-friction flows of liquid at nanopatterned interfaces. Nat. Mater. 2:237–40 118. Ou J, Perot B, Rothstein JP. 2004. Laminar drag reduction in microchannels using ultrahydrophobic surfaces. Phys. Fluids 16:4635–43 98 Quéré Annu. Rev. Mater. Res. 2008.38:71-99. Downloaded from arjournals.annualreviews.org by 213.160.13.186 on 07/09/08. For personal use only. 119. Ou J, Rothstein JP. 2005. Direct velocity measurements of the flow past drag-reducing ultrahydrophobic surfaces. Phys. Fluids 17:103606 120. Truesdell R, Mammoli A, Vorobieff P, van Swol F, Brinker CJ. 2006. Drag reduction on a patterned superhydrophobic surface. Phys. Rev. Lett. 97:044504 121. Roach P, McHale G, Evans CR, Shirtcliffe NJ, Newton MI. 2007. Decoupling of the liquid response of a superhydrophobic quartz crystal microbalance. Langmuir 23:9823–30 122. Joseph P, Cottin-Bizonne C, Benoit JM, Ybert C, Journet C, et al. 2006. Slippage of water past superhydrophobic carbon nanotube forests in microchannels. Phys. Rev. Lett. 97:156104 123. Richardson S. 1973. On the no-slip boundary condition. J. Fluid Mech. 59:707–19 124. Steinberger A, Cottin-Bizonne C, Kleimann P, Charlaix E. 2007. High friction on a bubble mattress. Nat. Mater. 6:665–68 125. Ybert C, Barentin C, Cottin-Bizonne C, Joseph P, Bocquet L. 2007. Achieving large slip with superhydrophobic surfaces: scaling laws for generic geometries. Phys. Fluids 19:123601 126. Philip JR. 1972. Flows satisfying mixed no-slip and no-shear conditions. Z. Angew. Math. Phys. 23:353–70 127. Philip JR. 1972. Integral properties of flows satisfying mixed no-slip and no-shear conditions. Z. Angew. Math. Phys. 23:960–68 128. Lauga E, Stone HA. 2003. Effective slip in pressure-driven Stokes flow. J. Fluid Mech. 489:55–77 www.annualreviews.org • Wetting and Roughness 99
© Copyright 2026 Paperzz