Improving Surface Hydrophobicity by Microrolling-Based Texturing

It has been widely reported that microscale textures on surfaces can enhance or alter various surface properties such as optical reflectivity [1], hydrodynamic behavior [2], tribological behavior [3], as well as heat exchange efficiency [4]. Imparting desired functional properties on surfaces remains a challenging scientific and technological problem. Comprehensive reviews on structured and engineered surface applications can be found in Refs. [5,6] for a more in-depth understanding of the range of functions and properties that may be imparted via texturing, how functional properties arise from textures and on methods for fabricating such surfaces.

Wetting behavior of engineered surfaces, in particular, has recently received increased attention, as researchers have attempted to enhance hydrophobicity of surfaces via texturing [2,7]. A particular case is that of superhydrophobic surfaces that have been produced by fabricating hierarchical-type (micro- and sub-microscale) surface patterns [8–10]. By controlling the wetting behavior of a surface, the degree of a liquid spreading over the surface can be controlled to serve a specific purpose or application, such as coating/adhesive spreading improvement, water repellence, and self-cleaning. A surface is considered hydrophobic when the contact angle of a water droplet on the surface is larger than 90 deg. Moreover, if the contact angle exceeds 150 deg, the surface is classified as superhydrophobic. The repelling of liquids also brings about a so-called “self-cleaning” function to surfaces. Hydrophobicity, in general, is a property of the material, and hence, its magnitude depends on the natural surface energy of the material as well as the roughness and topography of the material's surface. Lower surface energy materials have more pronounced hydrophobic properties because their surfaces have lower energy to disrupt intermolecular bonds formed between the surface and the liquid in contact. At the same time, the roughness of the surface also influences the area for liquid adhesion to the surface, and thereby directly controlling surface wettability. With this understanding of the liquid–solid adhesion mechanism, there were primarily two techniques developed for surface hydrophobicity alterations: (1) lowering the surface energy and (2) reducing the droplet–surface contact area. Though surface energy can be reduced by applying a coating with a surface energy lower than that of the original surface, the substrates and the surrounding environment of coated parts always limit the application of coatings. Reduction of the droplet–surface contact area can be altered by micro/nanoscale texturing of the surface. Texturing of micro/nanofeatures on surfaces has been achieved by micromachining [11], laser material processing [12–15], electrochemical etching [16], sandblasting [17], chemical vapor deposition [18], photolithography [19], and anodization [20].

The μRT is an alternative method for surface texturing. It is a microforming process in which pretextured rolls are used for imprinting microfeatures onto the material being rolled via the mechanism of plastic deformation [21–23]. It is a less energy-intensive process as compared to laser texturing, and the texture fabrication procedure is relatively simple. Besides, it is also suitable for large area texturing due to its high scalability. These advantages of μRT will potentially benefit the production of hydrophobic surfaces on automobile and airplane bodies for functional behaviors such as corrosion resistance and drag reduction. In this paper, the μRT process is utilized to create textured aluminum surfaces in order to improve hydrophobic properties of the original surfaces.

Different texturing methods leave their unique characteristics in the process of imparting the fabricated textures, including edge radii, surface roughness, and topological characteristics. For instance, micromachining produces sharper edges than microrolling. The characteristics of the fabricated textures have different influences on the surface's hydrophobicity. This may relate to the concept of hierarchical roughness/structures [10] that are characterized by multiscale roughness with different scales of superimposed features, for example, submicron/nanofeatures superimposed over microscale features. Therefore, different texturing methods with different process parameters have an inherent impact on the hydrophobicity of the textured surfaces. Rukosuyev et al. utilized this unique hierarchical nature of the textures produced by femtosecond laser ablation to generate superhydrophobic surfaces, which have contact angles between 150 deg and 180 deg on steel, aluminum, and tungsten carbide [24].

In a similar manner, in this paper, the utilization of hierarchical structures produced by a two-pass μRT technique to enhance the surface hydrophobicity will be investigated. The process basically produces square-patterned textures on workpiece surfaces by using a pretextured roll with microgrooves (as shown in Fig. 1(a)), applied twice orthogonally to create perpendicular channels that resemble a square pattern. It is noteworthy here that instead of using square-patterned rolls (as shown in Fig. 1(b)) to directly create the desired features in a single pass, the rolls are patterned with linear grooves to create the square pattern in two orthogonal passes. This is because the fabrication of square-patterned textures on cylindrical surfaces increases the cost and fabrication time for the pretextured rolls. The difficulties increase greatly if hard tool materials are being textured at the microscale. Compared to the square-patterned textures, the microgrooves are relatively easier to make on the roll surface. Therefore, the two-pass μRT with a groove-patterned pretextured roll is used to alleviate the difficulties in tool fabrication while enabling the texturing of square-patterned surfaces.

The formed microfeatures have the characteristic of having round edges, as can be noticed in Fig. 2. Moreover, the hierarchical nature of the texture created by μRT comes from the material pile-ups formed on the top of the textures, similar to “valleylike” formations, as depicted in Fig. 3. These hierarchical features, i.e., valleylike features in this case, serve as air pockets that resist contact between the liquid and the surface, thereby further enhancing hydrophobicity as compared to plain textured surfaces (without the material pile-ups). It is critical to independently study the effects of hierarchical features on improving hydrophobicity. Also, since the μRT process always leads to the formation of hierarchical textures, it becomes difficult to study in isolation the impact of the hierarchical nature of the texture as compared to nonhierarchical textures. Therefore, another process, LIPMM, was utilized to replicate the texture in both its nonhierarchical (square pillars) and hierarchical (square pillars with valleys on top) forms. Subsequently, the hydrophobicity of these surfaces was tested to establish the role of the hierarchical geometry.

To investigate the role of texture geometry of the μRT-produced surfaces on hydrophobicity, samples with different textures were produced by μRT. Moreover, to study the effects of the hierarchical feature found on μRT-produced textures, LIPMM was used to replicate the μRT-produced textures.

A desktop microrolling mill (DμRM), as shown in Fig. 4 [23], was designed and machined for the surface texturing of sheet metal workpieces. The machine frame and flexure hinges were machined from a single high-strength steel block by wire electrical discharge machining to ensure compactness as well as minimize backlash and hysteresis in the mill's structure. The DμRM has a pair of piezoelectric actuators and capacitive position sensors at the both ends of the roll shafts. The capacitive position sensors mounted on both ends of the lower roll enable the real-time measurement of the roll gap. Each position sensor has a measurement range of ±2 mm, with a resolution of 10 nm. Load cells with 5 kN/μm rigidity and 30 kN load were utilized for measuring roll forces. Figure 5 shows a tungsten carbide roll (27.6 mm diameter) pretextured by wire electrical discharge machining with three sections of different micropattern geometries. Each section is 10 mm width. The textured sections of the roll are separated by 3 mm wide flat (nontextured) sections. The teeth in all the sections have a tapered shape profile, as shown in Fig. 6. The texture dimensions in these three sections are tabulated in Table 1. In this work, the middle and right sections of the pretextured roll were used for texturing. These two sections were selected because they had significant differences in both tooth top width and spacing so that the produced patterns would have noticeable differences in their dimensions. The pretextured roll was coupled with a smooth roll (65.6 mm diameter) to imprint microgrooves on one side of an aluminum sheet during the texturing process. The smooth roll is made of D2 tool steel and is heat-treated to the hardness of HRC 60.

To perform the LIPMM process, a commercially available Nd-YVO₄ laser (Lumera Lasers, Inc., Santa Clara, CA) with 8 ps pulse duration operating at its second harmonic (532 nm wavelength) was used. The pulse repetition frequency can be varied between 10 kHz and 50 kHz. Maximum pulse energy was measured by an external power meter (Gentec Solo 2(R2)), and a maximum value of 6 μJ was obtained after reflection and transmission losses within the beam delivery system. The Gaussian beam was focused by a 25 mm focusing lens to a 10.5 μm spot size (1/e²). The substrate was mounted on a five-axis motion stage with a translation resolution of 10 nm. The pulsed laser beam was brought to focus in a dielectric medium to create a localized thermal plasma, which was subsequently brought in contact with the workpiece (also immersed into the dielectric) to machine features via thermomechanical ablation [25–27] (see Fig. 7). A crater or dimple is produced on the workpiece surface when the workpiece is kept stationary with respect to the beam focus. A single microchannel is created when craters machined by consecutive plasma discharges are overlapped by moving the workpiece at a feed rate of 0.4 mm/s. These microchannels can be arranged in an orthogonal grid to create a square pattern and can also be combined with dimples to create hierarchical structures.

Textures were fabricated with a two-pass μRT process at room temperature with a rolling speed of approximately 86 mm/min. The 0.4 mm aluminum sheets (AA3003) were rolled to imprint microgrooves on their surfaces in the first rolling pass. The rolled part was then cut into smaller pieces and rotated by 90 deg to perform the second rolling pass. The length of the cut portion after the first pass was equal to the sample width in the second pass. As a result, microgrooves were imprinted in both the longitudinal and the transverse directions to form square patterns. Because the workpiece material, i.e., aluminum, is soft comparing to the roll materials, there were no observable wears found on the rolls between passes. Both fine and coarse square patterns were produced on the aluminum specimens by using the middle and right sections of the pretextured roll. Example images of coarse and fine square patterns are shown in Figs. 8 and 9, respectively. Dimensions and images of channels were obtained with an optical three-dimensional (3D) microcoordinate system, Alicona InfiniteFocus, using a 20 × magnification with a 0.05 μm vertical and 2 μm lateral resolution. Due to the imperfections of the teeth on the pretextured roll, certain inevitable variations exist in groove width and depth as shown in Fig. 9(b). Hence, the averaged channel width and depth are taken for investigation. Four textures of different dimensions were produced (C1, C2, F1, and F2), as summarized in Table 2, including coarse and fine textures denoted by “C” and “F,” respectively. Because of the tapered teeth profiles, the average channel widths are larger than the tooth top width, as stated in Table 1. Note that the teeth in the pretextured roll were not fully indented into the material, and therefore, pile-ups were formed at the channel openings. The pile-up effect will be discussed in Sec. 4.1. The roughness of the unrolled AA3003 sheet was Ra = 0.3 μm.

It has been indicated that valleylike hierarchical structures were formed on the square pillars produced by μRT, which will be shown in Sec. 4.2. The same features were replicated by LIPMM in order to verify their role in increasing the contact angle. Three-dimensional views of flat-top and valley-top square patterns are shown in Figs. 10 and 11. Aluminum (AA3003), the same material as used in μRT, was used in LIPMM. However, a thicker sheet of 1 mm was used in this process, and its surface roughness had a slight difference than that used in μRT. The roughness of the original surface was Ra = 0.5 μm. Note that although LIPMM can also perform texturing, μRT has a higher production rate than LIPMM. Therefore, LIPMM is used here only as a prototyping tool while μRT is suggested for large area texturing.

Static contact angle tests were performed on the textured surfaces to observe their hydrophobic behavior. The contact angle is the angle between the liquid droplets relative to the contact interface and reflects the hydrophobicity of a surface. Surfaces with contact angles larger than 90 deg are defined as hydrophobic.

Figure 12 shows a comparison of the mean contact angles of the different textured surfaces listed in Table 2, while Fig. 13 shows the corresponding images of droplets showing the contact angles. The contact angle is defined as shown in Fig. 13. After fabrication, the surfaces were cleaned with acetone and then air dried after rinsing with water prior to each test. A single dose of a 3 μl water droplet was placed onto the sample surface for contact angle measurement. The contact angle tests were performed with the Drop Shape Analyzer DSA30 (Kruss, Germany). This device is particularly designed for contact angle measurement and droplet imaging, as shown in Fig. 13. Five measurements have been repeated on each surface and the average value is taken for analysis. The error bars in the figures represent the minimum and maximum contact angles in the five measurements. The average contact angle of the nontextured surface is 100 deg. Compared to the smooth surface, i.e., nontextured surface, the textured surface gives larger contact angle values ranging from 110 deg to 130 deg, except for the C1 surface. The C1 surface has a coarse square pattern with a shallow channel depth, which is about 1 μm. The channel depth of the C1 surface is comparable to the initial roughness of the specimen. As a consequence, there is no difference in the resulting contact angle. When a comparison is made between the C1 and C2 surfaces, it is evident that deeper channels lead to a larger contact angle. A similar observation can also be made from the fine texture patterned (F1 and F2) surfaces.

The widths of the square pillars formed by the fine and coarse square patterns are 100 μm and 200 μm, respectively. The pillar surface fraction of the fine texture is about 64%, while in the coarse case, it is about 57%. The surface fraction is defined as the ratio of the area on top of the pillar to the total surface area. However, results in Fig. 12 do not reflect how the pillar surface fraction impacts the contact angle. This is because the channel depths vary in the different cases. In order to conduct further investigations on the impact of the pillar surface fraction, the production of textures should be precisely controlled, and this will be the subject of future work.

Figure 14 shows the topography of a μRT-textured surface, as measured by the Alicona InfiniteFocus System. Due to the pile-up effect, small cavities are formed on the top of the square as marked with a solid-lined square in the figure. These valleylike structures are nonuniform and asymmetric. The maximum trough and peak distance is about 6 μm. They may interfere with the attachment of the droplet to the specimen surface and potentially increase the contact angle for enhanced hydrophobicity. A discussion on the effects of this valleylike feature is provided in Sec. 4.2.

Nosonovsky and Bhushan suggested that hierarchical roughness promotes hydrophobicity [28]. The study stated that submicron structures superimposed over microscaled structures enable surfaces having larger contact angles and facilitate the attainment of superhydrophobic conditions. Although the valley-topped pillars do not exhibit a significant hierarchical roughness, the superimposition of microfeatures, i.e., the micropillars with the valleys on top, increases the contact angle. Surfaces with flat-top (Fig. 10) and valley-top square patterns (Fig. 11) produced by LIPMM were subjected to contact angle tests for which the results are shown in Fig. 15. Figure 16 shows the images of the droplets in the contact angle measurements. Note that the purpose of these tests is to investigate the effect of valleylike structures on the pillar tops on hydrophobicity. The LIPMM-produced structure mimics the valleylike structures inherently generated by μRT. It can be observed that surfaces with textures give larger contact angles than smooth surfaces. When the contact angle is compared with the nontextured and flat-top square patterned surfaces, the valley-topped square pattern has a larger contact angle by approximately 20 deg and 10 deg, respectively, which indicates improved hydrophobic properties. Therefore, it can be concluded that the “valleys” on the pillar tops, which are formed due to the pile-up effect in μRT, produce a hierarchical roughness effect leading to enhanced hydrophobicity.

Based on the above findings, it can be seen that a significantly larger area of a hierarchical texture such as a valley-topped square pattern can be produced by μRT to improve the hydrophobic performance of rolled parts.

The influence of the groove aspect ratio on the contact angle of rolled workpieces was investigated. Figure 17 shows the effect of the μRT-formed groove's aspect ratio on contact angle. Note that the widths of the grooves are noticeably smaller than the widths of pillars in these cases. From the results, the contact angle increases with groove aspect ratio for both coarse and fine texture patterns. When the aspect ratio is very low, i.e., 0.02, the contact angle is similar to that of the nontextured surface, as shown by the results in Fig. 12. The contact angle increases as the aspect ratio increases for both cases. When the aspect ratio is higher than 0.2, the contact angles in both cases are similar, about 125 deg. The results are consistent with the theory derived by Patanker [29], whose work was inspired by the self-cleaning phenomenon of lotus leaf. He studied the hieratical nature of the leaf and proposed a theoretical method to calculate the contact angle, which was found increasing with the increasing pillar aspect ratio. The results obtained in this paper experimentally verified the theoretical finding using two processes that uniquely isolated the effect of two geometric features (groove and bumper).

Forming a composite solid–liquid–air interface helps to increase the contact angle [28]. The grooves on the textured surface provide air pockets between the liquid and the solid to form a composite interface. However, if the gravity–capillary wave amplitude of a liquid is comparable to the groove depth, the liquid will fill the groove. The transition of the composite interface to a homogeneous liquid–solid interface reduces the contact angle [28]. From the results in Fig. 17, the contact angle is similar to that of a nontextured surface when the aspect ratio is about 0.02. This is because the grooves are not deep enough to form a composite interface, causing them to be filled with liquid. As a result, it has a contact angle as low as that of the smooth surface. As the aspect ratio increases, the contact angle increases due to the larger air pocket volume leading to a reduction in the amount of liquid filling the grooves. When the aspect ratio is high, i.e., above 0.2 in this study, the groove depth is no longer comparable to the gravity–capillary wave amplitude. Consequently, the contact angles for both coarse and fine texture patterns give similar contact angles.

Figure 18 shows the contact angles under the effect of the pillar surface area-to-volume ratio. The surface area-to-volume ratio is defined as the ratio of the total surface area of a pillar, i.e., pillar top surface and its four side surfaces, to its volume. For both coarse and fine texture patterns, larger surface area-to-volume ratios give smaller contact angles. However, their sensitivities to the ratio on the contact angle changes are different, i.e., the slopes of the contact angle change are different. Fine texture pattern is more sensitive to the ratio so that it gives a significant drop in contact angle when the ratio is slightly larger, i.e., from 0.2 to 0.3.

Hydrophobicity has a close relationship with the area over which contact occurs with the liquid as well as groove depths, which depends on the height of the pillars. In this study, the variations of the surface area-to-volume ratio were due to the changes of pillar heights for each case, i.e., fine and coarse patterns. The changes in pillar surface and the height vary the ratio and affect liquid adhesion on a pillar. When the ratio is large, the area per volume provided for liquid attachment is large and makes the surface more hydrophilic, and therefore, a smaller contact angle is exhibited. Additionally, the pillar surface fraction of the fine texture pattern, which is 64%, is higher than that of the coarse texture pattern (57%). The higher pillar surface fraction of the fine texture pattern combines with the surface area-to-volume ratio resulting in a sharper drop in contact angle. By compiling the findings in Sec. 4.3, one can conclude that finer and taller pillars help in increasing the contact angle. However, higher pillar density significantly impacts the hydrophobicity of a surface.

Contact angles were measured in both longitudinal and transverse directions on surfaces with unidirectional grooves. All samples were textured with μRT. The coarse pattern has grooves with a width of 65 μm and a depth of 17 μm while the fine pattern has grooves with a width of 25 μm and a depth of 7 μm. Figure 19 shows the contact angles of surfaces with grooves in different directions as well as the surfaces with square pattern textures. The contact angles of fine surfaces with grooves in the transverse direction do not exhibit any improvement as compared to smooth surfaces. However, they show larger contact angles in the longitudinal direction (reaching up to 140 deg) than the square pattern textured surfaces (C and F cases) shown in Fig. 20. In the longitudinal direction, the existence of grooves disrupts the continuity of the surface for water droplet attachment. As a result, larger contact angles were obtained than in the transverse direction. From the results, the anisotropy of the textures gave a significant contact angle improvement in the selected direction, i.e., longitudinal direction in this case.

The μRT is a high-yield forming process that can potentially texture large areas in a relatively short time, as compared to tool-based and beam-based methods. One of the applications of μRT is to create hydrophobic surfaces. Understanding that different texturing methods result in different inherent impacts on the hydrophobicity of a surface, surfaces textured with different patterns by μRT were subjected to contact angle tests to investigate their surface hydrophobicity.

• With a two-pass μRT process, textures with square patterns were imprinted on surfaces. The rolled square pattern has proven its effectiveness in improving the hydrophobic performance of μRT-textured surfaces. Larger texture depths gave more significant improvements. However, further investigation is needed to establish the effect of pillar surface fraction on hydrophobic performance improvement.

• Valley structures were found on the top of the rolled surfaces with square patterns, which are formed because of the pile-up effect of the imprinted grooves. The effect of valley-topped pillar geometry on hydrophobic improvement has been also proven with LIPMM-textured samples. Valleys on the top of the pillars gave larger contact angles as compared to flat-top pillars. Uniform production of larger valleys by μRT will be investigated in the future.

• The groove's aspect ratio and pillar surface area-to-volume ratio also affect the contact angle. The contact angle increases with increasing groove aspect ratio while it decreases with increasing pillar surface area-to-volume ratio. The knowledge of the effects of this micron-range geometry on hydrophobicity can help in the design of microsurface features and of their subfeatures to generate surfaces that provide composite liquid–solid–air interfaces for the achievement of hydrophobic/superhydrophobic properties. Moreover, the hydrophobic sensitivity of a surface increases with increasing pillar density, i.e., pillar surface fraction.

• The increase in contact angle depends on the combined effect of the aspect ratio and surface area-to-volume ratio. Further study is required to estimate the best combination of these two factors in order to have a larger contact angle to reach the superhydrophobic state.

• The anisotropy of textures also affects the contact angles in different groove directions.

In conclusion, the μRT method can be utilized to produce microrolled textured parts with better hydrophobic performance. Moreover, the μRT technique offers extremely high productivity in surface texture generation. With precise process control and integration with other manufacturing techniques, hydrophobic surfaces can be manufactured at a relatively low cost and at a fast rate.

Supports from the National Science Foundation (CMMI-1100507/1100787), Metal Industries Research Development Centre of Taiwan, and Dr. Edward F. Smith, III from Deringer-Ney, Inc., are gratefully acknowledged. The help of Professor Q. Jane Wang of the Tribology Laboratories at Northwestern University, Evanston, IL, was essential in performing the contact angle measurements.