Article pubs.acs.org/Langmuir
Effect of Surface Depressions on Wetting and Interactions between Hydrophobic Pore Array Surfaces Petra M. Hansson,† Yashar Hormozan,‡ Birgit D. Brandner,† Jan Linnros,‡ Per M. Claesson,†,§ Agne Swerin,†,§ Joachim Schoelkopf,∥ Patrick A. C. Gane,∥,⊥ and Esben Thormann*,§ †
YKI, Ytkemiska Institutet AB/Institute for Surface Chemistry, Box 5607, SE-114 86 Stockholm, Sweden Material Physics, ICT School, KTH Royal Institute of Technology, Electrum 229, SE-164 40 Kista, Sweden § Department of Chemistry, Surface and Corrosion Science, KTH Royal Institute of Technology, SE-100 44 Stockholm, Sweden ∥ Omya Development AG, CH-4665 Oftringen, Switzerland ⊥ School of Chemical Technology, Department of Forest Products Technology, Aalto University, P.O. Box 16300, FI-00076 Aalto, Finland ‡
ABSTRACT: The surface structure is known to significantly affect the long-range capillary forces between hydrophobic surfaces in aqueous solutions. It is, however, not clear how small depressions in the surface will affect the interaction. To clarify this, we have used the AFM colloidal probe technique to measure interactions between hydrophobic microstructured pore array surfaces and a hydrophobic colloidal probe. The pore array surfaces were designed to display two different pore spacings, 1.4 and 4.0 μm, each with four different pore depths ranging from 0.2 to 12.0 μm. Water contact angles measured on the pore array surfaces are lower than expected from the Cassie−Baxter and Wenzel models and not affected by the pore depth. This suggests that the position of the three-phase contact line, and not the interactions underneath the droplet, determines the contact angle. Confocal Raman microscopy was used to investigate whether water penetrates into the pores. This is of importance for capillary forces where both the movement of the three-phase contact line and the situation at the solid/liquid interface influence the stability of bridging cavities. By analyzing the shape of the force curves, we distinguish whether the cavity between the probe and the surfaces was formed on a flat part of the surface or in close proximity to a pore. The pore depth and pore spacing were both found to statistically influence the distance at which cavities form as surfaces approach each other and the distance at which cavities rupture during retraction.
■
INTRODUCTION Forces acting between hydrophobic surfaces across aqueous media are significantly more long-range compared to the expected van der Waals force between such surfaces, with interaction distances exceeding 10 or even 100 nm. This was first investigated by Israelachvili and Pashley, soon followed by other studies where forces of even greater magnitude were observed.1−4 Further studies indicate that the interaction is a consequence of cavitation or bridging bubbles between the surfaces, and today, this theory is the most accepted explanation for this force.5−11 However, the formation, stability, and shape of the nanoscopic, or possibly microscopic, gas bubbles are still a source of debate even though there are studies showing compelling evidence for both presence and long-term stability of bubbles on hydrophobic surfaces.8,12−18 Other experiments indicate a thermodynamically driven cavitation process at small separations but show no evidence of nanobubbles.4,19−21 Capillary evaporation as being the main cause for the observed hydrophobic interaction is further emphasized by comparing force measurements for hydrophobic surfaces in water to hydrophilic surfaces in humid air.22 There are also studies © 2012 American Chemical Society
showing spectroscopic evidence, as judged from evanescentwave atomic force microscopy (AFM) or confocal Raman microscopy, for capillary evaporation and solvent segregation close to hydrophobic or superhydrophobic surfaces.23,24 The latter example describes the accumulation of air or possibly water vapor close to a superhydrophobic surface, which resulted in extremely long-range interaction distances of several micrometers as measured by AFM. Surface wettability and the nonwetting phenomenon of superhydrophobicity are of importance for many plants and animals, among which the lotus flower is the most wellknown.25,26 Lately, the difference between a lotus leaf with a rolloff angle close to 0° and a rose leaf with a high adhesion between the liquid and the surface has been brought to attention. The difference is thought to be a consequence of no water penetration for the lotus effect and penetration of water into the microstructures, but not the nanostructures, in the rose petal Received: May 18, 2012 Revised: July 2, 2012 Published: July 6, 2012 11121
dx.doi.org/10.1021/la302036d | Langmuir 2012, 28, 11121−11130
Langmuir
Article
forces. A schematic figure of the surfaces and the colloidal probe is displayed in Figure 1.
case.27,28 A rough surface structure together with a low surface energy results in a high water contact angle, but in order to achieve superhydrophobicity, a combination of micro- and nanoscale roughness on the surface is, in many cases, required. The actual contact angle on a rough surface can differ significantly from the apparent contact angle, and two basic models are commonly applied to describe extreme cases of the wetting behavior. The Wenzel model is based on the assumption that the liquid completely wets and penetrates the rough surface features.29 Therefore, a rough hydrophobic surface should appear more hydrophobic compared to a smooth surface with the same chemical composition due to the energy needed to wet the additional surface area. In contrast, the Cassie−Baxter state describes the situation where a droplet resides on an air layer between the droplet and the roughness features.30 Here, a higher contact angle is accompanied by a decrease in the contact area between the droplet and the surface and an increase in contact area between the droplet and air pockets under the droplet. Very often, the Wenzel and Cassie−Baxter models do not provide a definite solution, and a more correct description is an in-between situation with partial penetration of the surface features. Further, transition between the states depending on how the droplet is formed is possible, and it has been suggested that the activation energy for this to happen is fairly low.31 It should be noted that the validity of the Wenzel and Cassie− Baxter equations has been questioned. The debate dates long back but was recently revisited by Gao and McCarthy, showing that the contact angle of a surface is only affected by the situation at the three-phase contact line and not by the interactions occurring underneath the droplet.32 Studies showing how the Wenzel and Cassie−Baxter models have failed to predict the contact angle accurately on a heterogeneous surface with chemical or topographical heterogeneities covered by the droplet have brought this question to attention.33−35 Several experiments have suggested inadequacy of the Cassie−Baxter and Wenzel models,36,37 while others argue that this is all a consequence of incorrect use of the equations.38−40 The complicated picture for surface structure and heterogeneity in relation to wetting is also pronounced in connection with surface interaction forces. It has been proven numerous times that the hydrophobicity of a surface, as determined by the water contact angle, is not the only parameter affecting the range and strength of the measured interactions in aqueous solution.41,42 In addition, the roughness and the structure of the surface seem to have a huge impact on the measured forces.43 This is likely a consequence of nanobubbles being able to hide in crevices and irregularities on a rough surface and thereby promote the formation of air cavities bridging two nearby surfaces and thus cause capillary attraction. We have recently shown that a decrease in surface roughness length scale for surfaces covered with nanoparticles is accompanied by an increase in the magnitude of the surface forces and interaction distances.44 This is interpreted as being due to less pinning and decreased activation energy for the capillary to grow for smaller roughness length scales. In this study, pore array surfaces with two different pore spacings were prepared in silicon. These surfaces were subsequently hydrophobized and used for wetting studies as well as in AFM colloidal probe measurements to investigate how the surface forces depend on pore spacing and pore depths. The regular nature of the surface features, consisting of regularly spaced depths and no heights, allows assessment of the effects of surface depressions on both wetting characteristics and surface
Figure 1. Schematic figure of the relative sizes of the probe and the pore array surfaces, displaying two different pore separations, both with four different depths.
■
EXPERIMENTAL SECTION
Materials. The substrates used for sample preparation were silicon wafers with a measured rms roughness of 0.1 nm over a 2 × 2 μm2 area. Silica particles with a diameter of 12 μm (G. Kisker GmbH, Germany) were used as probes in the AFM force measurements. Ethanol (99.5%, Solveco, Sweden), octadecyltrichlorosilane (OTS, Alfa Aesar GmbH, Germany), hydrochloric acid (HCl, Merck, Germany), toluene (Merck), and NaCl (Merck) were all of analytical grade and used as received. An epoxy resin (JB Weld) was used to glue probes to calibrated cantilevers. The water used in all experiments was purified by means of a Milli-Q Plus Unit (Millipore) and had a resistivity of 18.2 MΩ cm. Sample Preparation. Microstructured pore array surfaces with two different pore spacings, each with four different depths, were prepared by generation of patterns on silicon wafers by photolithography followed by reactive ion etching using the cyclic Bosch process by switching between etching with SF6 and passivation with C4F8.45 Two available recipes, developed by Yun et al.,46 were applied for each pore spacing and only the number of cycles was changed to vary the pore depth. For the surfaces with 1.4 μm pore spacing, the depths were 0.3, 1.5, 4.0, and 12.0 μm, and for 4.0 μm pore spacings, the depths were 0.2, 1.0, 2.4, and 8.0 μm as seen in Figure 1. The surfaces were cleaned in hot 10% hydrochloric acid and dried in nitrogen gas followed by silanization in a 2 mM solution of OTS in toluene for 18 h. To remove excess and unreacted silane, the surfaces were sonicated, first in toluene and later in ethanol, followed by thorough rinsing in water. Reference surfaces were prepared from silicon wafers without pores by the same silanization procedure as described above. The silica probes were silanized by shaking them in a 2 mM solution of OTS in toluene for 18 h followed by rinsing in toluene. A micromanipulator (Eppendorf, Germany) was used to glue the hydrophobic probes to the cantilevers (NSC12/tipless/No Al, Mikromasch, Estonia). Surface Characterization. Scanning electron microscopy (SEM) images were recorded by a Zeiss “Ultra 55” SEM (Carl Zeiss NTS GmbH, Germany) for the side-view images and a Jeol 7 000F (Jeol Inc.) for the top-view images. AFM topographical images of the surfaces were recorded using a Nanoscope Multimode V (Bruker) operated in PeakForce mode using 11122
dx.doi.org/10.1021/la302036d | Langmuir 2012, 28, 11121−11130
Langmuir
Article
ScanAsyst software and silicon nitride cantilevers (ScanAsyst Air, Bruker). Macroscopic contact angle measurements were made using a DataPhysics OCA40 micro (DataPhysics GmbH, Germany). The system includes a high-speed CCD camera (2200 fps) with 20× magnification, automatic syringe dispenser, and a tilting table. Static contact angles were measured by placing a 4 μL droplet on the surface and recording the shape of the droplet at equilibrium. Dynamic contact angles were measured with a 30 μL droplet by the tilting method. The table was tilted at a speed of 1° s−1, and both advancing and receding contact angles were recorded at the moment the droplet started to slide. Also, the sliding angle, i.e., the tilt angle at which the droplet first started to move, was noted. To study the shape of the three-phase contact line, water droplets resting on the pore array surfaces were imaged by a light microscope (Zeiss Axio Scope with a 50× Zeiss objective (numerical aperture (NA) = 0.75)). Confocal Raman measurements were performed with a WITec alpha300 (WITec GmbH, Germany) using a 532 nm laser for excitation. An oil immersion objective with 100× magnification and NA = 0.9 was used for the measurements. The scan range for XY images was 25 × 25 μm2, for depth scans 25 × 15 μm2, and the integration time of each Raman spectrum was 2 ms. The optical parameters give a lateral resolution of 260 nm, and by assuming a refractive index of 1.3 for the sample (water inside pores), the vertical resolution can be estimated to 500 nm. A Raman spectrum was recorded at every image pixel. During a confocal Raman measurement, a droplet of the liquid was placed on the pore array surface and covered by a microscope glass slide, which gave an approximate thickness of the liquid film of 10 μm. Force Measurements. The forces between the hydrophobized pore array surfaces and the hydrophobic colloidal probe were measured using an AFM (Nanoscope Multimode III equipped with a PicoForce controller and a scanner with closed loop in the normal direction, Bruker) equipped with a fused quartz liquid cell. The cantilevers were calibrated before attachment of the colloidal probe using the method proposed by Sader et al.47 Force versus separation measurements were performed in aqueous 10 mM NaCl that had been degassed and heated to ∼50 °C just prior to use. The forces were recorded over a 10 × 10 μm2 area, on 10 by 10 regularly spaced spots. One force curve was recorded on each spot.
Figure 2. SEM images of selected pore array surfaces: (a) top-view of a surface with 4.0 μm pore spacing, scale bar 10 μm, (b) top-view of a surface with 1.4 μm pore spacing, scale bar 10 μm, (c) side-view of a surface with 4.0 μm pore spacing, here with a pore depth of 2.4 μm, scale bar 1 μm, (d) side-view of a surface with 1.4 μm pore spacing, here with a pore depth of 1.5 μm, scale bar 200 nm.
contact angles lower than that for the surfaces with 1.4 μm pore spacing, whereas the receding contact angles were similar for these two pore spacings. The difference in static and advancing contact angles is consistent with the smaller projected pore area, 16% for 4.0 μm pore spacing in comparison to 34% for the 1.4 μm pore spacing, and thus a larger percentage of the projected surface area resembles the flat surface. All the different models for the wetting behavior of rough surfaces, Wenzel/Cassie−Baxter/ three-phase contact line, discussed below predict a slightly higher contact angle the more pores on the surface. No variation of the contact angles with respect to pore depth can be seen for any array surface, i.e., neither with 4.0 μm nor the 1.4 μm pore spacing. In the complete wetting scenario, the contact angle, θreal, can, according to the Wenzel model, be related to the contact angle on a flat surface, θflat, by
■
cos θreal = r cos θflat
RESULTS AND DISCUSSION Here, the wetting behavior, water penetration into the pores, and force measurements between the hydrophobized pore array surfaces will be presented and discussed. Side-view and top-view images of a few selected pore array surfaces can be seen in Figure 2. The side-view images show the number of etching cycles made to obtain the desired depth. From the top-view images, the projected pore area percentage was calculated to be 34% and 16% for the 1.4 and 4.0 μm surfaces, respectively. Topographical images recorded by AFM show that the area around the pores is not perfectly flat, as seen in Figure 3, which will be of importance later when the statistical analysis of the force measurements is discussed. Along a pore edge, the variation in height can be several tens of nanometers as seen clearly in Figure 3b. Also, there is no sharp edge between the flat surface and the pore, but the area around the pore displays several small peaks and edges. However, all the pores on the same surface seem to look similar. Contact Angle Measurements: The Behavior at the Three-Phase Contact Line. The static and dynamic macroscopic contact angles of water on all the hydrophobic pore array surfaces, with two different pore spacings of 1.4 and 4.0 μm and different pore depths (see Figure 1) are displayed in Figure 4. The structures of the pore array surfaces result in slightly higher advancing and static contact angles and lower receding contact angles compared to the flat reference surface. The surfaces with the larger pore spacing, 4.0 μm, displayed advancing and static
(1)
where the roughness factor, r, gives the ratio between the real surface area and the corresponding projected surface area of a flat surface.29 For all the pore array surfaces, except the two with the smallest pore depths, the increase in surface area for the pore surfaces compared to the flat surface is so large that the roughness factor is outside the range of what can be used in the Wenzel equation. However, an extremely large roughness factor would give an extremely high contact angle (close to 180°), which is not observed in our measurements. Thus, the Wenzel model is not applicable to our pore array surfaces. If the droplet does not wet the pores but instead rests on top of air-filled pores, the contact angles can, according to the Cassie− Baxter model, be described by cos θreal = f1 cos θ1 + f2 cos θ2
(2)
where f1 and f 2 are the area fractions of solid surface area and pore area, respectively.30 By assuming a cylindrical pore structure and a contact angle on the pores (θ2) of 180°, the Cassie−Baxter model gives a static water contact angle of 123° on the 1.4 μm surfaces and 117° on the 4.0 μm surfaces. The corresponding experimental values are 115−118° and 112−113° for the 1.4 and 4.0 μm surfaces, respectively, and thus somewhat lower than predicted by the Cassie−Baxter model. It is thus clear that neither the Wenzel model nor the Cassie− Baxter model gives an accurate description of how the presence 11123
dx.doi.org/10.1021/la302036d | Langmuir 2012, 28, 11121−11130
Langmuir
Article
Figure 3. AFM images with the colored lines showing the positions for line scans displayed in panels b and d: (a) pore spacing 4.0 μm, depth 0.2 μm; (c) pore spacing 1.4 μm, depth 1.5 μm; (b) selected line scans over the surface shown in frame a; (d) selected line scans over the surface displayed in frame c.
(non-Cassie−Baxter behavior). To shed light on these results, we have obtained optical images of the three-phase line for a water droplet (see Figure 5). It is observed that the three-phase line
Figure 4. Static, advancing, and receding water contact angles for all the different pore array surfaces and a flat reference surface. d = pore depth (μm). Figure 5. A water droplet on a hydrophobic pore array surface with a pore spacing of 4.0 μm and a pore depth of 2.4 μm. Notice how the three-phase contact line is extended and how it avoids moving across the pores.
of surface depressions on our surfaces alters the contact angles, even though the contact angle values for the pore array surfaces with 4.0 μm pore spacing are fairly close to the values predicted by the Cassie−Baxter equation. However, as discussed by Gao and McCarthy32 and others,34,35 a correct interpretation of the Wenzel and Cassie−Baxter equations should only consider the wetting at the three-phase contact line and exclude the situation beneath the droplet. The lack of effect of pore depth, and the relatively small effect of pore spacing, seen in our experiments suggests that the three-phase line does not enter into the pores (non-Wenzel behavior) and does not rest on air filled pores
avoids covering the pores, creating a jagged line. This means that the surface structure under the three-phase line does not represent the surface structure covered by the droplet which explains why the contact angles are lower than the values predicted by Wenzel and Cassie−Baxter models.32,35 It should be noted that while these results tell us that the three-phase line 11124
dx.doi.org/10.1021/la302036d | Langmuir 2012, 28, 11121−11130
Langmuir
Article
every pixel, are indeed from the position marked with an arrow in the figures, whereas the signals from silicon, such as those in Figure 7b, arise from the outermost surface layer. The fact that water is wetting the pores is inconsistent with capillary theory49 that shows that no water will be able to penetrate pores of micrometer size provided they have as high contact angles as measured for our surfaces. This suggests that the pore walls are less hydrophobic than the flat part of the surface and in fact have contact angles below 90°. A possible reason is incomplete removal, during the HF treatment, of the organic passivation layer applied during the etching of the pore array structure. The alternative reason that the toluene solution used for silanazation did not fully penetrate the pore structure, hence causing less efficient hydrophobization of the pore wall surfaces, seems less likely considering that the confocal Raman images of pore array surfaces covered with toluene verify that toluene is able to pass into the pores (Figure 7c,f). A parameter more difficult to control is the number of free OH groups on the top surface in relation to the number of such groups on the wall surface in the pores. A different amount of accessible OH groups will affect the amount of silanes on the surface and, hence, the hydrophobicity. In conclusion, water does penetrate the pores, and this is due to a lower hydrophobicity of the pore walls, and two possible reasons for this have been suggested. Force Measurements: Shape of Force Curves. A typical force curve measured between a hydrophobized flat reference sample and a hydrophobic probe is displayed in Figure 8a. This force curve has the shape normally observed for interactions between flat hydrophobic surfaces in water.4,10,11 On approach, a sudden attraction is observed from where the probe jumps into contact with the surface. This phenomenon is assigned to the formation of an air cavity that pulls the surfaces together, and we refer to the distance at which the attraction appears as the jump-in distance. Upon separation, a strong long-range attractive force corresponding to extending the bridging air cavity is observed. The force decreases until a separation where the cavity ruptures and the force returns to zero. We refer to this distance as the rupture distance. The solid line in Figure 8a represents a theoretical fit to the restoring force of stretching of an air capillary, calculated using
avoids the pores, it does not provide any information on the wetting state below the droplet. In the force measurements, which will be described later, not only the behavior of the threephase line but also the wetting state, that is the eventual presence of air cavities in the pores beneath the droplet, is of importance. From the dynamic contact angle measurement, the advancing angles are shown to be only marginally larger than the static ones. This is in agreement with the picture of the three-phase line resting on a flat part of the substrate, avoiding the pores. However, the significantly smaller receding contact angles as well as the larger contact angle hysteresis on the pore array surfaces compared to on the flat surface suggest that dewetting is counteracted by droplet pinning at the pores. Since low contact angle hysteresis is expected for the Cassie−Baxter state and larger contact angle hysteresis for the Wenzel state, this result indicates at least partial wetting of the pores,48 even though a direct pinning effect where the drop is stretched under strain prior to jumping over the pore during receding also may contribute. Further, the sliding angles when tilting the surfaces varied from 30° for the pore array surfaces to 15° for the flat reference surface, which means that the droplet on the pore array surface indeed seems to be pinned around the edges of the pores. During measurements it was also observed that droplets on a flat surface move smoother, i.e. more continuously, compared to the situation on a pore array surface where the droplet displays more discrete, steplike motions. Confocal Raman Microscopy: Air or Water in the Pores? To elucidate whether air or water is present in the pores, confocal Raman microscopy images of a pore array surface completely covered with water were recorded. In Figure 6,
⎛ ⎜ F = 4πγcR ⎜1 − ⎜ ⎝ Figure 6. Confocal Raman spectra for the silicon surface, water, and toluene. The arrows indicate the peaks for the respective substance that were used to achieve contrast in the images shown in Figure 7.
⎞ ⎟ 2 ⎟ + D ⎟⎠
D V πR
(3)
where c=
spectra for the different components are shown, together with arrows indicating which peaks are used for creating the Raman images. A top-view image with water illustrated in yellow, as shown in Figure 7a, clearly demonstrates accumulation of water in the pores. This image is from a cross section a few micrometers down into the sample, which clearly demonstrates that water indeed has penetrated into the pores. Also a depth scan, displayed in Figure 7d, shows apparent evidence for water presence in the pores. However, the resolution of the image is not good enough to determine whether water penetrated all the way to the bottom of the pores or if there is some air present underneath. It should be noted that the silicon surface acts like a mirror and does not allow recording of Raman spectra in its bulk. Hence, the water spectra, where one spectrum was recorded in
cos(θ1 + β) + cos θ2 2
(4)
Equation 3 is based on the assumption that the cavity volume is constant during separation of the surface and the probe.50 It describes how the capillary force, F, between a sphere, with radius R, and a surface can be calculated. D is the distance between the probe and the surface, V is the volume of the cavity, θ1 and θ2 are the contact angles of the cavity against the surface and the probe, respectively, and β is the angle between the line starting from the center of the spherical probe going in the normal direction and toward the contact point of the cavity edge at the probe. In this case β ≪ θ1, and it is therefore omitted. For a full description of the geometrical arrangement the reader is referred to an article by Kappl and Butt.50 From the volume of the cavity (V ∼ 5 × 10−21 m3) used to calculate the theoretical curve, the radius of the cavity 11125
dx.doi.org/10.1021/la302036d | Langmuir 2012, 28, 11121−11130
Langmuir
Article
Figure 7. Confocal Raman images of a pore array surface with a pore spacing of 4.0 μm and a pore depth of 8 μm: (a−c) XY scans where (a) shows water, (b) silicon, and (c) toluene as yellow. Notice how water and toluene have accumulated in the pores. (d−f) XZ scans where (d) shows water, (e) silicon, and (f) toluene in yellow. The arrows represent the positions where the XZ (in the XY images) or the XY (for the XZ images) images were recorded. Scale bars: 5 μm.
when the probe is in close proximity to the surface is estimated to be 800 nm. Repeated force curves obtained at different positions on the flat reference surface showed variation in the magnitudes of the jump-in and rupture distances due to the stochastic nature of cavity formation and rupture, as discussed previously.5,44 However, the shape of the force curve stayed the same and could in all cases be described by eq 3. This is in contrast to the force curves observed between the probe and the pore array surfaces, which showed a larger variety in shape. Many force curves retained the flat surface characteristics as can be seen in Figure 8b, while two other examples of force curves are displayed in Figures 8c,d. The one in Figure 8c has an unusually large rupture force, as defined as the force where the cavity breaks, and the other displayed in Figure 8d shows several steps in the curve instead of a smooth extension of the cavity. Since these features only occur with the pore array surfaces, they should be related to the position of the cavity in relation to the pores on the surface. All the force curves measured for the pore array surfaces were evaluated in terms of their shape and sorted into one of two groups. The first group contains force curves with the normal constant capillary volume shape (Figure 8b), consistent with eq 3. All the other force curves, with appearances like those in Figures 8c or 8d, were grouped into the second category. Interestingly, the fraction of force curve belonging to group 2 was 0.4 for the surfaces with 1.4 μm separations between pores and 0.11 for the surfaces with 4.0 μm pore separation, i.e., similar to the area fractions covered by pores, 0.34 and 0.16, on their respective surfaces. This provides strong evidence that type 1 force curves correspond to cavities with their center on the solid surface (i.e., the colloidal probe center is on top of the solid surface), whereas type 2 curves are associated with positions where the center of the colloidal probe is above a pore. The situation with the center of the probe residing above a pore is rather complex. In case the pore was completely filled with water, no air capillary is expected to form, and this does not agree with our results. In case the pores were completely filled with air, the measured forces should be similar to those measured between a hydrophobic probe and an air bubble. This type of interaction has been measured numerous times and shown to result in extremely long-range interactions on approach, far beyond the shape and distances seen here.51−53
We have already seen in Figure 5 that the water/air contact line preferentially resides between the pores rather than on top of them. Similarly, the border between the pores and the flat surface will impose a constraint on the position of the three-phase line and constitute a barrier for its motion during retraction. In such a situation, eq 3 does not hold, and thus the shape of the retraction curve is expected to change. A typical feature of the type 2 curves (Figures 8c,d) is one or several regions where the attraction is roughly constant with separation. Such a situation can be reached if the volume of the cavity is increased by diffusion of air molecules from air pockets present in the pores. Because of pinning at the pore−solid contact line, when the cavity is extended to sufficiently large distances (above 100 nm), the shape of the cavity becomes too distorted to be maintained, and the cavity breaks while still having a large base area. This results in a higher than normal attractive force at the rupture distance as seen in Figure 8c. The shape of the force curve in Figure 8d indicates movement of the cavity three-phase line around pore edges, each motion starting by pinning of the cavity at a pore edge followed by a sudden jump to a place away from it. Consistent with this interpretation, these force curves with special features are more common for the surfaces with smaller pore spacing, 1.4 μm, where the area fraction covered by pores is larger and where the three-phase line thus has more difficulties avoiding the holes. Figure 9 displays histograms of the rupture forces for the surfaces with 1.4 and 4.0 μm pore separation, where it is clear that the higher rupture forces are significantly more common for the surface with 1.4 μm pore separation. Because of the larger area fraction of pores for the surface with 1.4 μm pore separation, the likelihood for a cavity to be pinned and break when in contact with a pore is of course increased. No correlation between the frequency of force curve with different shapes and pore depth could be detected, which suggests that the density of pores is the sole determinant for the frequency of type 2 curves. The radius of the cavity when the probe is close to the surface was estimated to be around 800 nm using eq 3. This indicates that, for 1.4 μm pore spacing, the cavity three-phase line will always be in contact with a pore, while for the 4.0 μm surfaces, there are occasions when the cavity would experience the surface as flat. Since many force curves for both types of surfaces are 11126
dx.doi.org/10.1021/la302036d | Langmuir 2012, 28, 11121−11130
Langmuir
Article
Figure 9. Rupture force, defined as the force when the cavity breaks on retraction, for measurements on the pore array surfaces, divided into two groups depending on the size of the pore spacing. (a) Rupture forces for all the force curves recorded with the pore array surface having 4.0 μm pore spacing and (b) force curves from the pore array surface with 1.4 μm pore spacing. Note that the larger rupture forces occur exclusively on the 1.4 μm surfaces.
Force Measurements: Statistics. Statistics, based on 100 force curves for each pore array surface, obtained from the force measurements are summarized in Figure 10. The two types of force curves were grouped into two categories based on their appearance, as described above, and evaluated separately. One of the reasons for this was to determine if the group 1 force curves, which have the normal behavior described by eq 3 and illustrated in Figure 8b, would display the same (or similar) statistical values as obtained when probing the flat reference surface. Such a conclusion was not straightforward to make, and it seems as if the pores are affecting the interactions not only through the shape of the force curve. Looking at the mean values, the trends for both pore spacings are strikingly similar to only small differences in the absolute values. In general, the trends for the jump-in distance and rupture distance with respect to pore depth are obvious but opposite. The former seems to decrease with increasing pore depth, while the latter increases. In order to form a cavity on approach, water must evaporate or air on the surface must assemble and pull the probe toward the surface. The easier this happens, the longer the jump-in distance will be. As concluded from the contact angle and Raman measurements, water penetrates the pore features. However, since the pore depth affects the measured forces some air must still be present in the pores. It is reasonable to assume that the closer the air pockets are to the surface, the more readily they facilitate formation of cavities between the probe and the pore array surface. Statistically, such air pockets are closer to the surface in the less deep pores, which explains why the jump-in distance increases with decreasing pore depth. The observation
Figure 8. Representative force curves measured between a hydrophobic colloidal probe and (a) a flat hydrophobic surface and (b−d) a hydrophobized pore array surface with a pore spacing of 1.4 μm and a pore depth of 1.5 μm. The solid red lines in (a) and (b) are theoretical fits to the retraction curve calculated from eq 3.
consistent with eq 3, it can be concluded that proximity of a pore does not always have a significant effect on the capillary force law. 11127
dx.doi.org/10.1021/la302036d | Langmuir 2012, 28, 11121−11130
Langmuir
Article
longer rupture distances are observed for deeper pores suggests that once the cavity is formed, a surface with deeper pores, i.e., having larger pore volume where air can reside, can supply the cavity with more air and hence support the existence of a more stretched cavity that leads to a larger rupture distance. Also, as the surfaces are separated, the three-phase contact line will eventually come in contact with a pore edge, which possibly introduces an asymmetry and an early cavity rupture. It seems like the amount of available air is the dominating feature for deciding the rupture distance for the deep pores while for the shallower pores, the rupture due to presence of pores has a more profound effect on the result. This is why the rupture distance mean values do no approach the value measured for the flat surface as the depth of the pores decreases. The statistics for the adhesion force, seen in Figure 10c, demonstrates that the group 1 force curves, consistent with eq 3, have higher mean values that are closer to that for the reference surface than those belonging to group 2. The standard deviation for the adhesion values is large since it is affected by many variables like contact area, presence of air, height differences of the area around the pores, the surface roughness (see Figure 3), and the structure of the pore walls. Considering this, we restrain from drawing any conclusions about a possible effect of the pore depth of the adhesion value.
■
CONCLUSIONS
■
AUTHOR INFORMATION
We have shown that wetting by water and surface interaction forces between hydrophobic surfaces in water are affected by the presence of surface depressions, here in the form of ordered microstructured pore arrays. From water contact angle measurements, it was concluded that neither the Wenzel nor the Cassie− Baxter model could describe the wetting behavior. Instead, the three-phase contact line avoids the pores, creating a jagged line that explains the higher contact angle for the pore array surfaces compared to a chemically similar flat surface. Confocal Raman microscopy images demonstrated that water penetrated into the pores, but not necessarily filling them completely. In fact, the statistical analysis of the surface force data suggests that some air pockets are present within the pores. Forces measured between the hydrophobic pore array surfaces and a hydrophobic colloidal probe showed that it is possible to distinguish between cavities formed between the pores and those formed in close proximity to a pore. The first category has the characteristic force versus distance dependence associated with stretching of a capillary of constant volume, as most often seen in measurements with a flat hydrophobic surface. In other force curves, it was evident that an early rupture of the cavity occurred, leading to an unusually large rupture force, or several steps in the force curve were observed as a consequence of movements of the cavity three-phase line around the pores. Statistical analysis of the measured force curves show that the jump-in distance decreases with an increased pore depth, while the situation is the opposite when it comes to the rupture distance. The difference in jump-in distance is suggested to be a consequence of air, statistically, being closer to the probe for the shallower pores, hence promoting cavity formation. Once the cavity has been created, the deeper pores can, statistically, supply more air, which rationalizes the increase of the rupture distance.
Figure 10. Statistics showing (a) the jump-in distance, (b) the rupture distance, and (c) the adhesion force for measurements between a hydrophobic colloidal probe and the hydrophobized pore array surfaces. d = pore depth (μm). Group 1 represents the force curves that can be fitted to eq 3 (constant capillary volume) while group 2 force curves have large rupture forces and/or several steps in the retraction curve.
that the reference surface, where air only can accumulate directly on the surface, has the largest jump-in distance supports this notion that a longer jump-in distance is facilitated by air pockets being close to the probe. The length of the rupture distance is decided by how far it is possible to extend the cavity on retraction, which is affected by the amount of air initially present on the surface or in the pores that can be accommodated in the cavity but also by disturbances during movements of the three-phase contact line. The fact that
Corresponding Author
*Tel + 46 8 790 9920, Fax +46 8 208 284, e-mail
[email protected]. 11128
dx.doi.org/10.1021/la302036d | Langmuir 2012, 28, 11121−11130
Langmuir
Article
Notes
(18) Yang, S.; Dammer, S. M.; Bremond, N.; Zandvliet, H. J. W.; Kooij, E. S.; Lohse, D. Characterization of nanobubbles on hydrophobic surfaces in water. Langmuir 2007, 23, 7072−7077. (19) Yushchenko, V. S.; Yaminsky, V. V.; Shchukin, E. D. Interaction between Particles in a Nonwetting Liquid. J. Colloid Interface Sci. 1983, 96, 307−314. (20) Ederth, T.; Liedberg, B. Influence of wetting properties on the long-range “hydrophobic” interaction between self-assembled alkylthiolate monolayers. Langmuir 2000, 16, 2177−2184. (21) Wallqvist, V.; Claesson, P. M.; Swerin, A.; Schoelkopf, J.; Gane, P. A. C. Interaction forces between talc and hydrophobic particles probed by AFM. Colloids Surf., A 2006, 277, 183−190. (22) Thormann, E.; Simonsen, A.; Hansen, P.; Mouritsen, O. Force Trace Hysteresis and Temperature Dependence of Bridging Nanobubble Induced Forces between Hydrophobic Surfaces. ACS Nano 2008, 2, 1817−1824. (23) McKee, C. T.; Ducker, W. A. Refractive Index of Thin, Aqueous Films between Hydrophobic Surfaces Studied Using Evanescent Wave Atomic Force Microscopy. Langmuir 2005, 21, 12153−12159. (24) Brandner, B. D.; Hansson, P. M.; Swerin, A.; Claesson, P. M.; Wåhlander, M.; Schoelkopf, J.; Gane, P. A. C. Solvent Segregation and Capillary Evaporation at a Superhydrophobic Surface Investigated by Confocal Raman Microscopy and Force Measurements. Soft Matter 2011, 7, 1045−1052. (25) Barthlott, W.; Neinhuis, C. Purity of the Sacred Lotus, or Escape from Contamination in Biological Surfaces. Planta 1997, 202, 1−8. (26) Neinhuis, C.; Barthlott, W. Characterization and Distribution of Water-Repellent, Self-Cleaning Plant Surfaces. Ann. Bot. 1997, 79, 667− 677. (27) Bhushan, B.; Nosonovsky, M. The Rose Petal Effect and the Modes of Superhydrophobicity. Philos. Trans. R. Soc., A 2010, 368, 4713−4728. (28) Feng, L.; Zhang, Y.; Xi, J.; Zhu, Y.; Wang, N.; Xia, F.; Jiang, L. Petal Effect: A Superhydrophobic State with High Adhesive Force. Langmuir 2008, 24, 4114−4119. (29) Wenzel, R. N. Resistance of Solid Surfaces to Wetting by Water. Ind. Eng. Chem. 1936, 28, 988−994. (30) Cassie, A.; Baxter, S. Wettability of Porous Surfaces. Trans. Faraday Soc. 1944, 40, 546−551. (31) Patankar, N. A. Transition between Superhydrophobic States on Rough Surfaces. Langmuir 2004, 20, 7097−7102. (32) Gao, L.; McCarthy, T. J. How Wenzel and Cassie were Wrong. Langmuir 2007, 23, 3762−3765. (33) Pease, D. C. The Significance of the Contact Angle in Relation to the Solid Surface. J. Phys. Chem. 1945, 49, 107−110. (34) Bartell, F.; Shepard, J. Surface Roughness as Related to Hysteresis of Contact Angles. II. The Systems Paraffin 3 Molar Calcium Chloride Solution Air and Paraffin Glycerol Air. J. Phys. Chem. 1953, 57, 455−458. (35) Extrand, C. Contact Angles and Hysteresis on Surfaces with Chemically Heterogeneous Islands. Langmuir 2003, 19, 3793−3796. (36) Yamamoto, K.; Ogata, S. 3-D Thermodynamic Analysis of Superhydrophobic Surfaces. J. Colloid Interface Sci. 2008, 326, 471−477. (37) Erbil, H. Y.; Cansoy, C. E. Range of Applicability of the Wenzel and Cassie- Baxter Equations for Superhydrophobic Surfaces. Langmuir 2009, 25, 14135−14145. (38) McHale, G. Cassie and Wenzel: Were They Really So Wrong? Langmuir 2007, 23, 8200−8205. (39) Nosonovsky, M. On the Range of Applicability of the Wenzel and Cassie Equations. Langmuir 2007, 23, 9919−9920. (40) Panchagnula, M. V.; Vedantam, S. Comment on How Wenzel and Cassie were Wrong by Gao and McCarthy. Langmuir 2007, 23, 13242. (41) Hato, M. Attractive Forces between Surfaces of Controlled ’’Hydrophobicity’’ Across Water: A Possible Range of ’’Hydrophobic Interactions’’ between Macroscopic Hydrophobic Surfaces Across Water. J. Phys. Chem. 1996, 100, 18530−18538. (42) Nguyen, A. V.; Nalaskowski, J.; Miller, J. D.; Butt, H. J. Attraction between Hydrophobic Surfaces Studied by Atomic Force Microscopy. Int. J. Min. Process. 2003, 72, 215−225.
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS Omya Development AG is thanked for funding part of this project and for supporting cooperation between YKI, KTH, and its industrial Mineral and Surface Chemistry R&D. Part of the project was financed by the Swedish Foundation for Strategic Research (SSF). A.S. thanks the Troëdsson foundation for support of an adjunct professorship at KTH. The Troëdsson Foundation is also gratefully acknowledged for the grant for the confocal Raman equipment.
■
REFERENCES
(1) Israelachvili, J.; Pashley, R. The Hydrophobic Interaction Is LongRange, Decaying Exponentially with Distance. Nature 1982, 300, 341− 342. (2) Israelachvili, J.; Pashley, R. Measurement of the Hydrophobic Interaction between Two Hydrophobic Surfaces in Aqueous Electrolyte Solutions. J. Colloid Interface Sci. 1984, 98, 500−514. (3) Pashley, R. M.; McGuiggan, P. M.; Ninham, B. W.; Evans, D. F. Attractive Forces between Uncharged Hydrophobic Surfaces: Direct Measurements in Aqueous Solution. Science 1985, 229 (4718), 1088. (4) Christenson, H. K.; Claesson, P. M. Cavitation and the Interaction between Macroscopic Hydrophobic Surfaces. Science 1988, 239, 390− 392. (5) Wallqvist, V.; Claesson, P. M.; Swerin, A.; Ostlund, C.; Schoelkopf, J.; Gane, P. A. C. Influence of Surface Topography on Adhesive and Long-Range Capillary Forces between Hydrophobic Surfaces in Water. Langmuir 2009, 25, 9197−9207. (6) Thormann, E.; Simonsen, A.; Hansen, P.; Mouritsen, O. Interactions Between a Polystyrene Particle and Hydrophilic and Hydrophobic Surfaces in Aqueous Solutions. Langmuir 2008, 24, 7278− 7284. (7) Meyer, E.; Rosenberg, K.; Israelachvili, J. Recent Progress in Understanding Hydrophobic Interactions. Proc. Natl. Acad. Sci. U. S. A. 2006, 103, 15739. (8) Attard, P. Nanobubbles and the Hydrophobic Attraction. Adv. Colloid Interface Sci. 2003, 104, 75−91. (9) Christenson, H.; Claesson, P. Direct Measurements of the Force between Hydrophobic Surfaces in Water. Adv. Colloid Interface Sci. 2001, 91, 391−436. (10) Carambassis, A.; Jonker, L. C.; Attard, P.; Rutland, M. W. Forces Measured between Hydrophobic Surfaces Due to a Submicroscopic Bridging Bubble. Phys. Rev. Lett. 1998, 80, 5357−5360. (11) Parker, J. L.; Claesson, P. M.; Attard, P. Bubbles, Cavities, and the Long-Ranged Attraction between Hydrophobic Surfaces. J. Phys. Chem. 1994, 98, 8468−8480. (12) Ishida, N.; Inoue, T.; Miyahara, M.; Higashitani, K. Nano Bubbles on a Hydrophobic Surface in Water Observed by Tapping-Mode Atomic Force Microscopy. Langmuir 2000, 16, 6377−6380. (13) Ishida, N.; Sakamoto, M.; Miyahara, M.; Higashitani, K. Optical Observation of Gas Bridging between Hydrophobic Surfaces in Water. J. Colloid Interface Sci. 2002, 253, 112−116. (14) Sakamoto, M.; Kanda, Y.; Miyahara, M.; Higashitani, K. Origin of Long-Range Attractive Force between Surfaces Hydrophobized by Surfactant Adsorption. Langmuir 2002, 18, 5713−5719. (15) Simonsen, A. C.; Hansen, P. L.; Klosgen, B. Nanobubbles Give Evidence of Incomplete Wetting at a Hydrophobic Interface. J. Colloid Interface Sci. 2004, 273, 291−299. (16) Zhang, X. H.; Maeda, N.; Craig, V. S. J. Physical Properties of Nanobubbles on Hydrophobic Surfaces in Water and Aqueous Solutions. Langmuir 2006, 22, 5025−5035. (17) Zhang, X. H.; Quinn, A.; Ducker, W. A. Nanobubbles at the Interface between Water and a Hydrophobic Solid. Langmuir 2008, 24, 4756−4764. 11129
dx.doi.org/10.1021/la302036d | Langmuir 2012, 28, 11121−11130
Langmuir
Article
(43) Butt, H. Capillary Forces: Influence of Roughness and Heterogeneity. Langmuir 2008, 24, 4715−4721. (44) Hansson, P. M.; Swerin, A.; Schoelkopf, J.; Gane, P. A. C.; Thormann, E. Influence of Surface Topography on the Interactions between Nanostructured Hydrophobic Surfaces. Langmuir 2012, 28, 8026−8034. (45) Chang, C.; Wang, Y. F.; Kanamori, Y.; Shih, J. J.; Kawai, Y.; Lee, C. K.; Wu, K. C.; Esashi, M. Etching Submicrometer Trenches by Using the Bosch Process and Its Application to the Fabrication of Antireflection Structures. J. Micromech. Microeng. 2005, 15, 580. (46) Yun, S. H.; Lee, H. S.; Ha Kwon, Y.; Gothelid, M.; Mo Koo, S.; Wagberg, L.; Karlsson, U. O.; Linnros, J. Multifunctional Silicon Inspired by a Wing of Male Papilio Ulysse. Appl. Phys. Lett. 2012, 100, 033109−033109−3. (47) Sader, J. E.; Chon, J. W. M.; Mulvaney, P. Calibration of Rectangular Atomic Force Microscope Cantilevers. Rev. Sci. Instrum. 1999, 70, 3967−3969. (48) Sheng, Y. J.; Jiang, S.; Tsao, H. K. Effects of Geometrical Characteristics of Surface Roughness on Droplet Wetting. J. Chem. Phys. 2007, 127, 234704. (49) Evans, D. F.; Wennerström, H. The Colloidal Domain; WileyVCH: New York, 1999. (50) Butt, H.; Kappl, M. Normal Capillary Forces. Adv. Colloid Interface Sci. 2009, 146, 48−60. (51) Ally, J.; Vittorias, E.; Amirfazli, A.; Kappl, M.; Bonaccurso, E.; McNamee, C. E.; Butt, H. J. Interaction of a Microsphere with a SolidSupported Liquid Film. Langmuir 2010, 26, 11797−11803. (52) Preuss, M.; Butt, H. J. Direct Measurement of Particle-Bubble Interactions in Aqueous Electrolyte: Dependence on Surfactant. Langmuir 1998, 14, 3164−3174. (53) Preuss, M.; Butt, H. Measuring the Contact Angle of Individual Colloidal Particles. J. Colloid Interface Sci. 1998, 208, 468−477.
11130
dx.doi.org/10.1021/la302036d | Langmuir 2012, 28, 11121−11130
