ARTICLE pubs.acs.org/Langmuir
Metal Matrix Composites for Sustainable Lotus-Effect Surfaces Michael Nosonovsky,* Vahid Hejazi, Aniedi E. Nyong, and Pradeep K. Rohatgi College of Engineering & Applied Science, University of Wisconsin—Milwaukee, Milwaukee, Wisconsin 53211, United States ABSTRACT: The lotus effect involving roughness-induced superhydrophobicity is a way to design nonwetting, self-cleaning, omniphobic, icephobic, and antifouling surfaces. However, such surfaces require micropatterning, which is extremely vulnerable to even small wear rates. This limits the applicability of the lotus effects to situations when wear is practically absent. To design sustainable superhydrophobic surfaces, we suggest using metal matrix composites (MMCs) with hydrophobic reinforcement in the bulk of the material, rather than only at its surface. Such surfaces, if properly designed, provide roughness and heterogeneity needed for superhydrophobicity. In addition, they are sustainable, since when the surface layer is deteriorated and removed due to wear, hydrophobic reinforcement and roughness remains. We present a model and experimental data on wetting of MMCs. We also conducted selected experiments with graphite-reinforced MMCs and showed that the contact angle can be determined from the model. In order to decouple the effects of reinforcement and roughness, the experiments were conducted for initially smooth and etched matrix and composite materials.
1. INTRODUCTION The contact angle is the main parameter that characterizes wetting of solid surfaces by liquids. Water droplets on smooth hydrophobic surfaces do not usually form contact angles with the solid surface greater than 120°. When the contact angle exceeds 150°, this is referred to as superhydrophobicity. Superhydrophobic surfaces also usually have low contact angle hysteresis, show self-cleaning properties, and have low drag for fluid flow. The range of actual and potential applications of self-cleaning surfaces is diverse including optical (self-cleaning lenses and mirrors), building and architecture (windows, exterior paints, roof tiles), textiles, solar panels, microdevices (where the reduction of adhesion is crucial), and applications requiring antifouling from biological and organic contaminants.1,2 Two models of superhydrophobic behavior are used and are known as the Wenzel3 and CassieBaxter4 models. The main difference between the two models is whether the liquid droplet retains contact with the solid surface at all points or whether the liquid bridges only across surface protrusions, thus resulting in a droplet suspended on a composite solid and vapor surface.5 In the Wenzel model (Figure 1a), the solidliquid contact occurs at all points below the droplet and the observed equilibrium contact angle with the rough surface, θ, is given by cos θ ¼ Rf cos θ0
ð1Þ
where the roughness factor Rf = ASL/AF g 1 is the ratio of the real substrate area ASL to the projected area AF, and θ0 is the contact angle on a smooth surface of the same material. In the CassieBaxter model (Figure 1b), the droplet suspends itself across surface protrusions, and an average of the cosines of the angle on the solid (i.e., cos θ0) and on the air (i.e., cos180° = 1) below the drop is used. If fSL is the fraction of the solid surface r 2011 American Chemical Society
upon which the drop sits and (1 fSL) is the fraction below the drop that is air, then the CassieBaxter equation applies. When solidliquid interface is rough, the roughness factor should also be included cos θ ¼ fSL Rf cos θ0 ð1 fSL Þ
ð2Þ
The overall conclusion is that two factors are needed to produce a superhydrophobic surface: roughness (providing high Rf and fSL) and a certain extent of initial hydrophobicity (e.g., a coating), such that cos θ0 < 0. Surface roughness magnifies the hydrophobicity, bringing the contact angle into the superhydrophobic region, 150° e θ e 180°. Furthermore, proper surface roughness is more critical than the initial superhydrophobicity, since under certain conditions even an initially hydrophilic surface can show superhydrophobic properties.6 There is also evidence that surfaces with dual-scale roughness (nanoroughness superimposed on microroughness) make hydrophobic properties sustainable.7 Since the 1990s, when new technologies emerged to produce surfaces with designed microstructure, a significant amount of research work was done on design, fabrication, and characterization of superhydrophobic surfaces from various materials, ranging from polymers and ceramics to textiles. A significant limitation on the practical application of the lotus effect for self-cleaning is the lack of sustainability of superhydrophobic microstructured coatings, which are often extremely vulnerable even to small wear rates and contamination. Several methods of making sustainable superhydrophobic and self-cleaning materials have been recently suggested including producing hierarchical Received: May 4, 2011 Revised: October 14, 2011 Published: October 15, 2011 14419
dx.doi.org/10.1021/la201656y | Langmuir 2011, 27, 14419–14424
Langmuir
ARTICLE
suggested for concretes to prevent water penetration.20 In the present paper, we investigate wetting of MMCs with the potential for various applications where self-cleaning sustainable surfaces are needed ranging from antifouling for water industry to magnetic tape-head interfaces.10 Note also that wear is often anisotropic; furthermore, the MMC itself often has anisotropic properties (for example, as a result of orientation of fiber reinforcement), so anisotropic wetting can occur.2123
Figure 1. Wetting of a microstructured surface in the (a) homogeneous (Wenzel, solidliquid) and (b) composite (CassieBaxter, solidliquidair) regime.
2. MODELING OF WETTING OF COMPOSITE MATERIALS For a composite interface built of two fractions with the fractional areas of f1 and f2 (so that f1 + f2 = 1), the contact angle is given by the Cassie equation cos θ ¼ f1 cos θ1 þ f2 cos θ2
8
9
surfaces, porous materials infused with liquid, and various, in particular, metallic composite materials.10 Metals tend to have higher surface energies than polymers and ceramics. This makes producing a superhydrophobic metallic material a much more difficult task than a polymer- or ceramicbased one.1114 In the area of metallic superhydrophobic materials, a number of advances have been made. In the 1950s, Bikerman15 investigated wetting of stainless steel plates with different finishes with the contact angles around 90° and proposed that the surface roughness provides resistance for the sliding of water droplets. Since then, few studies of nonwetting metallic materials have been conducted. Qian and Shen16 studied the effect of the surface roughness induced by the chemical etching on wetting of metallic materials. They used Al, Cu, and Zn specimens immersed into an etchant (a mixture of HCl, H2O, and HF) at room temperature for time periods from 5 to 15 s. Shirtcliffe and McHale17 studied the wettability of Cu-base superhydrophobic surfaces. They used Cu to form the base material and a coating to hydrophobize this material. The removal or addition of material roughened the surface to control wetting by combining roughness with surface patterning. Sommers and Jacobi18 achieved anisotropic wettability on an Al surface by controlling its surface microtopography. Typical superhydrophobic surfaces have carefully designed microstructure facilitating superhydrophobicity. However, these structures are easily destroyed by wear or deterioration. An alternative approach involves inserting reinforcement in the bulk of the material, rather than at the surface, so that the reinforcement creates roughness and modifies hydrophobic properties of the material. Composite materials are usually made of bulk matrix material with reinforcement particles or fibers inside them. Metal matrix composites (MMCs) are composite materials which have a metallic matrix and a reinforcement of another metallic or nonmetallic (ceramic or polymer) material.19 MMCs with hydrophobic reinforcement can provide much broader opportunities than pure metals for design and fabrication of composite surfaces, and readily supply the reinforcement hydrophobic fraction and surface roughness due to the reinforcement. However, superhydrophobic MMCs have not yet been explored in the literature. Furthermore, in a composite material, the hydrophobic reinforcement is in the bulk of the material rather than at the surface and thus wear does not necessarily lead to the deterioration of the hydrophobic coatings, making these materials appropriate to the situations where traditional lotus-effect coatings cannot be used. The use of composite materials with hydrophobic reinforcement in the bulk has already been
ð3Þ
where θ1 and θ2 are the contact angles of the fractions. If a composite material has a matrix and reinforcement with the volume fractions of fm and fr (so that fm + fr = 1) forming a rough surface, the contact angle is then given by cos θ ¼ Rfm ð1 fr Þ cos θm þ Rfr fr cos θr
ð4Þ
where θm and θr are the contact angles for the matrix and reinforcement materials, and Rfm and Rfr are corresponding roughness factors. Note that for spherical reinforcement particles the roughness factor is equal to the ratio of half of the sphere’s area 2πR2 to the cross-sectional area πR2 or Rfr = 2. Solving for the reinforcement fraction yields the volume of the reinforcement fraction providing the desired contact angle θ fr ¼
cos θ Rfm cos θm Rfr cos θr Rfm cos θm
ð5Þ
Further assuming Rfr = 2, Rfm = 1 (no roughness expect from the reinforcement particles), and θ = 180° (the superhydrophobic limit) yields fr ¼
1 cos θm 2cos θr cos θm
ð6Þ
which has a solution (fr < 1) if θr >120°. Thus, it is difficult to produce a composite interface by only using the reinforcement roughness If water forms partial contact with the solid (composite or CassieBaxter) interface with the fractional solid liquid contact areas fSLm and fSLr, the contact angle is given by cos θ ¼ Rfm ð1 fr ÞfSLm cos θm þ Rfr fr fSLr cos θr 1 þ fr fSLr þ ð1 fr ÞfSLm
ð7Þ
Solving for the reinforcement fraction yields the volume of the reinforcement fraction providing the desired contact angle θ fr ¼
cos θ Rfm fSLm cos θm þ 1 fSLm Rfr fSLr cos θr Rfm fSLm cos θm þ fSLr fSLm
ð8Þ
Making the assumptions of Rfr = 2, Rfm = 1, fSLr = 1, and θ = 180° yields fr ¼ 14420
fSLm fSLm cos θm 2cos θr fSLm cos θm þ 1 fSLm
ð9Þ
dx.doi.org/10.1021/la201656y |Langmuir 2011, 27, 14419–14424
Langmuir
ARTICLE
Table 1. Chemical Composition of the Samples sample material
composition
copper base alloy
Cu (79.0- 82.0%), Sn (2.5- 3.5%), Pb (6.37.7%), Zn (7.010%), P (0.02%), Al (0.05%), Si (0.005%)
copper-graphite composite
Cu (81%), Ni (5%), Fe (4%), Al (9%), Mn (1%), 60% vol of graphite.
aluminum base alloy
Al (88%), Si (12%)
aluminum-graphite composite
Al (35%), Si (5%), 60% vol of graphite
Figure 2. SEM images of polished samples: (a) copper base alloy, (b) copper graphite MMC, (c) aluminum base alloy, and (d) aluminum graphite MMC.
The effect of wear on a composite material is twofold. First, the matrix roughness factor, Rfm, can be changed due to material removal and evolve to a certain “equilibrium value”.24 This can affect the solidliquid fractional area, fSLm. Second, the reinforcement particles can be removed as matrix surface layers are removed due to the deterioration. However, new particles come in contact, so it is expected that the values of Rfr and fSLr do not change significantly. To decouple the effect of reinforcement and matrix roughness, we investigated experimentally wetting of composite materials with initially smooth surfaces and with the matrix roughness by etching, as described in the next section.
3. EXPERIMENTAL SECTION In order to verify experimentally the models presented in the preceding section, we prepared eight samples of MMCs, four with a relatively smooth surfaces and four roughened by etching, and measured their roughness and contact angles. By applying the etchant, we simulated corrosive wear of the samples. 3.1. Sample Preparation. The MMC samples were produced using previous methods for the Cu-graphite25 and Al-graphite26 composites.
The Cu-graphite samples were produced through the stir mixing of graphite particles, 60 vol %, into the base Cu alloy melt. This was followed by centrifugal casting, which involved the pouring of the Cu alloy melts containing graphite particles into the rotating mold in a horizontal centrifugal casting machine at 800 rpm. The Al-graphite sample production involved the low-pressure infiltration of graphite particles in a quartz tube (35 mm diameter and 266 mm length) with the Al base alloy in order to form the alloy with 60 vol % graphite. The Al base alloy was pre heated to 800 °C, and gas nitrogen at 758 kPa was applied on the resulting melt. The four samples of Al- and Cu-based alloys and their graphite composites were sectioned to 2.0 cm 1.5 cm 0.2 cm pieces. We used Al and Cu because they are standard materials which have been used as a matrix in the literature due to their suitability for making MMCs. Furthermore, Al- and Cu-based samples are easy to work with and are inexpensive as well. Table 1 presents the chemical composition of the samples. To investigate the effect of roughness, both smooth and rough samples were studied. To produce a smooth surface, the samples were grinded successively with 400, 600, and 1200 grit SiC paper. After that, the samples were polished with a soft cloth impregnated with 1 μm alumina. The scanning electron microscope (SEM) images of the polished samples are as shown in Figure 2. 14421
dx.doi.org/10.1021/la201656y |Langmuir 2011, 27, 14419–14424
Langmuir
ARTICLE
Table 2. Measured and Calculated Surafce Roughness and Contact Angles measured sample
etching time
Ra (μm)
calculated
measured contact angle
Rfm
contact angle (Wenzel, fSLm = 1)
contact angle (CassieBaxter, fSLm = 0)
Al base alloy
0
