Manufacturing of mushroom-shaped structures and its hydrophobic robustness analysis based on energy minimization approach

Hydrophobic/superhydrophobic surfaces have many attractive practical applications, which include micro/nanofluidics, drag reduction,^1,2 self-cleaning windows,³ water-proofing cloths and textiles,⁴ and antibiofouling surfaces.^5,6 The repellency of hydrophobic surfaces is mainly ascribed to the lubricating gas films generated at the interface between the solid substrate and the liquid, which results in a shear-free liquid-air area and dramatically reduces the interaction at the solid–liquid interface.⁷ For all the above mentioned applications, the robustness of a coating against complete wetting is crucial. Regrettably, hydrophobic surfaces may suddenly become wet and lose their slip effect because of external factors, such as pressure,⁸ vibrations,⁹ evaporation,¹⁰ air diffusion,¹¹ and impacts.¹² Maintaining the stability of the hydrophobicity of artificial surfaces remains a problem far from being resolved and the development of surfaces with highly robust hydrophobicity is of practical importance.

Surface hydrophobicity is governed by the chemical properties and roughness of the surface.¹³ With the combined effect of low-surface-energy materials and enhanced surface roughness, the macroscopic surface contact angle approaches 180°.¹⁴ However, in this strategy, surfaces with lower free energy are usually achieved by coating with a very thin film having low energy, and the film can be scratched or peeled off easily and the hydrophobicity reduced.

Recent theoretical and experimental research has found that the tip shape of pillars is an important factor driving hydrophobic stability and that pillars with mushroom-shaped tips, rather than those with concave, spherical or flat tips, tend to enhance the stability, which is ascribed to the generation of a net force that lifts the liquid upward. Numerous processes have been proposed to fabricate different types of microstructures with mushroom-shaped tips. Im et al.,¹⁵ for example, fabricated inverse-trapezoidal microstructures employing three-dimensional diffuser lithography, obtaining a sidewall slope of the exposed photoresist region by introducing an index-matching liquid between a source light and a mask to change the diffused angle of incident ultraviolet (UV) light. However, the fabricated inverse-trapezoidal microstructures are different from microstructures with mushroom-shaped tips. Using positive- and negative-tone photoresist materials, Sameoto et al.¹⁶ fabricated pillar arrays with mushroom-shaped tips employing photolithography and a molding process, where the unexposed negative-tone photoresist was used to generate undercut microholes. The limitation of this process is that it is necessary to precisely control the exposure dose to avoid the underneath negative-tone photoresist being exposed. By combining anisotropic dry etching and isotropic wet etching of silicon using the SiO₂ etch-stop layer to generate microholes with a bottom undercut, Kang et al.¹⁷ presented a method for the fabrication of mushroom-like micropillar arrays. However, the process is more complex and the shape of the obtained inner concave structures is limited by the Si wet etching process. Wang et al.¹⁸ introduced a simpler method for fabricating mushroom-shaped microstructures, where, by performing masked and unmasked exposures, microholes with a bottom undercut are generated and pillar arrays with mushroom-shaped tips are obtained after molding. However, it is not convenient to modulate the cap size of the mushroom-like microstructure because the cap size depends on the thickness of the exposed photoresist at the bottom and yet the cap size is believed to be important to wetting resistance. Additionally, the requirement of the light transparency of substrate materials restricts the adaptation of this process.

The present work proposes a facile method for fabricating mushroom-shaped structures based on double photolithography and cast molding and demonstrates well-defined mushroom-shaped structures in large areas. The hydrophobicity robustness of the prepared surfaces is evaluated theoretically by taking the energy minimization approach and verified experimentally using UV-curable polymer drops with low surface tension by inspecting the profiles of liquid–vapor interface deformation and tracking the trailing of the receding contact line after curing with UV exposure.

A polished glass wafer with a diameter of 70 mm and a thickness of 1.1 mm was used for the substrate. A patterned positive-tone photoresist AZ P4620 (Taiwan, ROC) was used for the template. Polydimethylsiloxane (PDMS) (Sylgard 184, Dow Chemicals) was used as a filling material to duplicate the structures from the template. NaOH $(5wt%0)$ was used as developer. All chemicals used in this study were of reagent grade and the processing solutions were prepared using deionized (DI) water.

The process flow for the fabrication of the mushroom-shaped microstructures is depicted in Fig. 1 and can be divided into the following steps.

Substrate cleaning (Fig. 1(a)). Glass substrates were cleaned in ultrasonic baths of acetone (10 min), ethanol (10 min), and DI water (10 min). Each substrate was then dried with pure nitrogen and baked in an oven (100 °C) for 10 min to remove residual water.

Photoresist spin-coating and pre-baking (Fig. 1(b)). The first layer of photoresist (AZ P4620, 400cP) was spin-coated on the wafer and the thickness of the first layer depended on the cap thickness of the mushroom-like microstructures. The characteristic spin-coating parameters correspond to a low speed of 500 rpm for 10 s and a high speed of 3000 rpm for 30 s resulting in a film thickness of about 6 μm. Spin-coating was followed by pre-baking of the resist at 95 °C for 5 min to harden the applied film.

Flood exposure (Fig. 1(c)). The AZ P4620 film was exposed using an exposure machine (ABM/6/350/NUV/DCCD, USA) without a mask and the dosage was adjusted to ensure a complete exposure of the photoactive compound in the photoresist. For the 6-μm thickness of the AZ P4620 coating, an exposure time of 10 s guaranteed complete exposure.

Second layer of photoresist coating, baking and exposure (Fig. 1(d) and (e)). The spin-coating parameters for the second layer of photoresist were a low speed of 500 rpm for 10 s and a high speed of 1500 rpm for 30 s resulting in a thickness of the second layer of film of about 25 μm. The baking parameters were the same than those for the first layer. Additionally, UV lithography was performed for 25 s.

Development and post-baking (Fig. 1(f)). The exposed photoresist was developed in 1–3 min using NaOH solution as the developer, with a mass fraction $5%0−9%0$⁠. After rinsing in DI water and drying with N₂ gas, post-baking of the resistance was carried out at 95 °C for 5 min to harden the photoresist film.

Cast molding with PDMS (Fig. 1(g) and (h)). After thoroughly mixing a prepolymer and a curing agent at a weight ratio of 10:1, the mixture was degassed in a vacuum environment for 10 min to eliminate trapped air bubbles and then poured onto the glass master. The mixture was evacuated in a vacuum environment for 10 min to help the liquid fill the channels of the master and was finally cured at a temperature of 60 °C, in an oven under normal atmosphere pressure for 4 hours. The resulting PDMS with replicated patterns on its surface was peeled from the master. Details of the cast molding process are given a previously published paper.¹⁹ 

The wettability of a prepared surface is usually characterized by measuring the contact angle of a drop of DI water resting on the surface. In this study, the usually used DI water drop was replaced with a UV-curable polymer. This choice was motivated by the low surface tension of the polymer and the ability of the polymer to be ‘frozen’ by UV exposure, offering the possibility of inspecting the profiles of liquid–vapor interface deformation. In addition, the use of a nonvolatile polymer drop avoided the effect of water evaporation on the measurements. The apparent contact angle of the photoresist on the prepared surface was measured with a contact-angle system (OCA20, Dataphysics, Germany) and the photoresist was then cured. Both the contact angle and three-phase contact line were employed to trace changes. The trace of the receding contact line, resulting from contraction during the resist curing and solidification of the liquid–substrate interface, was compared between surfaces with concave microstructures and those with vertical-sidewall pillars using a scanning electron microscope (SEM) (SU-180, Hitachi, Japan). The wettability robustness of the prepared surfaces was then evaluated.

The manufactured mushroom-like microstructures, with a pitch of 80 μm and various cap diameters, are shown in Fig. 2. When the PDMS film with mushroom-shaped patterns on its surface overlapped the photoresist master, clearly visible interference fringes were seen as shown in Fig. 2(b), which indicates the good uniformity of the fabricated microstructures in a large area. Figure 2(a) shows that the tip diameter of the mushroom-shaped pillars mainly depends on the development time. In addition to the cap diameter, θ_overhang, shown schematically in Fig. 3(a), is an important geometric parameter of surface wettability. Indeed, Fig. 3(b)–(c) provides the quantitative relation between the cap diameter, θ_overhang, and development time, where the cap diameter increases with the development time and the maximum diameter is limited by the pattern pitch; in general, θ_overhang decreases as a function of the development time, reaching a minimum value of 40°. By controlling the development time, the main geometric parameters driving the wettability of the fabricated patterns can be modulated conveniently. The proposed simple and high-yield methodology can therefore be used satisfactorily to manufacture robust hydrophobic surfaces.

When a liquid droplet rests on a textured surface, which wetting state (either a fully wetted Wenzel state or a Cassie–Baxter state) will be realized is determined by considering the overall free energy of the system. If a Cassie–Baxter state corresponds to the overall free energy for the system reaching a local or global minimum, it means that the hydrophobic surface is stable. On the basis of the overall free-energy minimization of the system, Patankar²⁰ investigated the energy of a drop on a rough substrate in the extreme wetting state and pointed out that a stable Cassie–Baxter state can be realized by optimizing the geometric feature of surface roughness. Marmur²¹ examined the problem in more detail and theoretically derived the energy change of a drop on a rough substrate during wetting state transition, stating that a local minimum in the Gibbs energy may exist depending on the surface topography characteristics. Tuteja²² built on Marmur’s work and calculated the change in the Gibbs free energy density with the evolution of the solid–liquid interface on a substrate with various geometric morphological features. Using an approach similar to that used by Tuteja,²² we calculate the areal Gibbs free energy density (G* = ( free energy of the system)/(original surface area of drop)) variation for a UV-curable polymer (γ_lv = 26.5 mN/m, θ = 70°) on propagating of a liquid–air interface of different surfaces. The results of these calculations can be used to estimate the hydrophobic stability of the prepared mushroom-like patterned surfaces.

From the thermodynamic point of view, the areal Gibbs free energy density of a drop with given volume in equilibrium on a substrate is given by

where G* is the Gibbs free energy density, θ_app is the apparent contact angle, θ_instri is the intrinsic contact angle on a smooth substrate, f is the ratio of the actual area of liquid–solid contact to the projected area on the horizontal plane, and r_f is the ratio of the actual wetted area to the total solid area of the textured surface. To facilitate the description of the calculation, a Cartesian coordinate system as shown in Fig. 4 is built and the side profile function of the patterns is fitted according to the topological surface feature of the fabricated microstructures. We can then obtain the geometrical parameters on propagation of the liquid–air interface as

where f(x) is the fitted side profile function, $f′$(x) is the derivative function of f(x), l is the pitch of the microstructures with a unit of μm, and s is the cap diameter with a unit of μm.

To compute the Gibbs free energy density, we developed Matlab code and adopted the following calculation procedure.

1. For a given liquid–air interface location (given x/h in Fig. 4), the geometrical parameters of f and r_f are calculated using eq. 2 and eq. 3.

2. Using eq. 1 and f and r_f calculated above, the Gibbs free energy density for a given liquid–air interface location can be obtained.

3. Changing the liquid–air interface location and the apparent contact angle, the Gibbs free energy density at all liquid–air interface locations of the side profile of the fabricated patterns are obtained.

4. Three-dimensional figures are plotted; the figures show the minimum Gibbs free energy density and allow the stable thermodynamic state to be deduced.

In our calculations, for any given distance x of the liquid–air interface from the top of the textured substrate (normalized with respect to the maximum height of the surface h; see Fig. 4), we assume a number of different values for the temporary apparent contact angle θ_app such that 0° < θ_app < 180°. We then compute the areal Gibbs free energy density (G*) of the liquid drop for a given θ_app and x/h, with respect to a reference state G* = 0 at x/h = 0. The calculation results are as shown in Fig. 5, where it is seen that for mushroom-like microstructures, the global minimum of the areal Gibbs free energy density appears at the location x/h = 0 for the UV-curable polymer, implying that a Cassie–Baxter state is the stable state of the system. By comparison, for microstructures with vertical-sidewall pillars, the global minimum of the areal Gibbs free energy density appears at the location x/h = 1, corresponding to the fully wetted Wenzel state.

Experiments were conducted to verify the theoretical analysis of the hydrophobic stability of the prepared surface. The apparent contact angle of the liquid polymer drop on the prepared surface was measured with a contact-angle system and the polymer was then cured. The trace of the receding contact line of depinning during UV exposure, which resulted from contraction during the polymer curing and solidification of the liquid–substrate interface, was tracked. The trace was then compared between surfaces with concave microstructures and those with vertical-sidewall pillars using an SEM.

The wetting properties of the surface with mushroom-like structures and the surface with a normal pillar array are here compared in terms of hydrophobic stability as shown in Figs 6 and 7, where Figs 6(a) and 7(a) show the manufactured mushroom-like structures and regular arrayed pillars. Firstly, the static contact angles of the drops before and after curing were measured; results are shown in Fig. 6(b) and 7(b). It is seen that the static contact angles of the mushroom-like patterns before and after curing (139° and 133.1°, respectively) were much larger than those of the normal arrayed pillars (70.2° and 65.5°, respectively). A possible reason is that different contact states form. Liquid polymer drops spontaneously adopt a so-called Cassie–Baxter state (i.e., the drop sits on top of the microposts, trapping air pockets below) in the concave pattern surface while Wenzel contact appears for the normal arrayed pillar surface. Figures 6(c) and 7(c) present SEM images of resin droplets before curing on surfaces with mushroom-like structures and regular arrayed pillars, respectively. Figure 6(d) shows the trace of the three-phase contact line after depinning and the meniscus at the bottom of the solidified resin after peeling off from the substrate. From the interface on the mushroom-like pattern surface (Fig. 6(d)), one can imagine the liquid photoresist meniscus at the liquid–solid interface before curing and conclude that the photoresist does not wet the bottom of the microstructures. For comparison, after curing and peeling off the substrate with normal arrayed pillars, deep holes and even broken PDMS pillars were seen on the surface of the solidified resin droplet as shown in Fig. 7(d), thus corroborating the Wenzel contact assumption and confirming a pinning effect and strong adhesion.

For further verification, the change in the three-phase contact line trace due to contraction during resist curing was inspected. Figure 6(e) and 6(f) shows that for the surface with mushroom-like structures, the three-phase contact line moved toward the drop bulk, indicating that there is no pinning formation. Although no pinning occurred on the whole, there is adhesion of the cured resin on top of the partial mushroom-like cap, which may be ascribed to the smoothness of the microstructure cap top. The construction of dual-level structures may be a feasible solution to this problem. For normal arrayed pillars, no three-phase contact line movement was observed, indicating that the Wenzel contact leads to resist pinning formation.

We proposed a convenient method with which to manufacture robust hydrophobic surfaces with mushroom-shaped microstructures and analyzed the hydrophobic stability by taking an energy minimization approach. Double photoresist coatings and double photolithography were adopted to fabricate the mushroom-like microstructure array, in which the cap thickness, cap diameter and stem height of the structures can be conveniently modulated by respectively controlling the thickness of the first layer of the photoresist film, the development time, and the thickness of the second layer of the photoresist film. Using this method, mushroom-like microstructures with controllable cap thickness, cap diameter, and stem height of the structures can be conveniently produced. Well-defined mushroom-shaped microstructures in a large area were demonstrated and the hydrophobic stability of the prepared surface was investigated theoretically and experimentally. Experimental results show that, in comparison with normal pillar arrays with vertical sidewalls, the mushroom-like structures can effectively enhance the hydrophobic stability, which agrees well with theoretical prediction. The proposed process provides a useful method with which to manufacture robust hydrophobic surfaces in a cost-effective and convenient manner.

This work was financially supported by the National Natural Science Foundation of China (No. 51375381,51575427) and the National Key Research and Development Program of China (No.2016YFB1101000).