Wettability control of polymeric microstructures replicated from laser-patterned stamps

In this study, two-step approaches to fabricate periodic microstructures on polyethylene terephthalate (PET) and poly(methyl methacrylate) (PMMA) substrates are presented to control the wettability of polymeric surfaces. Micropillar arrays with periods between 1.6 and 4.6 µm are patterned by plate-to-plate hot embossing using chromium stamps structured by four-beam Direct Laser Interference Patterning (DLIP). By varying the laser parameters, the shape, spatial period, and structure height of the laser-induced topography on Cr stamps are controlled. After that, the wettability properties, namely the static, advancing/receding contact angles (CAs), and contact angle hysteresis were characterized on the patterned PET and PMMA surfaces. The results indicate that the micropillar arrays induced a hydrophobic state in both polymers with CAs up to 140° in the case of PET, without modifying the surface chemistry. However, the structured surfaces show high adhesion to water, as the droplets stick to the surfaces and do not roll down even upon turning the substrates upside down. To investigate the wetting state on the structured polymers, theoretical CAs predicted by Wenzel and Cassie-Baxter models for selected structured samples with different topographical characteristics are also calculated and compared with the experimental data.

www.nature.com/scientificreports/ It has been widely reported that lotus leaves and rose petals are two typical types of superhydrophobic surfaces designed by nature with opposite effects. For instance, the lotus leaves show low adhesion to water allowing droplets to easily roll off their surfaces, whereas rose petals present high adhesion preventing water droplets from moving at any tilted angles on the surfaces 14,15 . These particular wetting characteristics are not only attributed to the surface chemistry, but also to their nano-and micro-scale surface morphology. Therefore, surface texturing is often employed together with a chemical coating to modify the wettability of technical surfaces. Numerous techniques, such as hot/UV embossing 16,17 , photolithography [18][19][20] , and laser-based structuring 21,22 , have been utilized to engineer the surface topography on a large variety of materials.
For large scale production, hot embossing is one of the most used methods as it provides a low-cost, high-resolution, and high-throughput replication of structures at the micro-and nanoscale 23,24 Due to the differences in the thermomechanical properties of the soft polymer and the hard mold, well-defined structures can be replicated from the mold to the polymers over large surfaces at one-step. Therefore, mold structuring plays a crucial role in achieving high-quality pattern replication. The most common methods to produce embossing molds include photolithography, micromachining, electroplating, etching, and laser structuring, among others [25][26][27][28] . Despite these methods are well-established, they still have some drawbacks, such as high cost, significant time consumption, multiple processing steps, limitation of feature resolution, and/or inaccurate dimension control 26,29,30 .
A one-step method providing high-resolution and low-cost processing for patterning hot-embossing molds is Direct Laser Interference Patterning (DLIP) 31 , by which a wide variety of periodic ordered patterns can be fabricated on the surfaces of various materials, including metals 32,33 , ceramics 34,35 , and polymers 36,37 . In this method, an interference pattern is transferred onto the sample surface by overlapping two or more coherent laser beams. At the intensity maxima positions, the material is directly ablated, thereby producing periodic patterns. Because of the interference phenomenon, a large number of periodic micro or sub-micro features can be generated after irradiating with a single laser pulse with an outstanding processing speed up to 1 m 2 min -138, 39 . In addition, this method offers the flexibility to adjust the spatial period (Λ) of the pattern by modifying the angle between the laser beams achieving typical lateral feature sizes from ~ 500 nm up to a few tens of micrometers 40 .
In this study, a simple and inexpensive approach is proposed to fabricate hydrophobic polymeric surfaces. With this strategy, periodic microstructures with a spatial period from 1.6 to 4.6 µm on a chromium (Cr) mold fabricated by DLIP were replicated onto PET and PMMA foils by using an electrohydraulic press. The static and dynamic wetting behaviors were studied to understand the effect of surface textures and surface roughness on the wetting performance of the two polymeric surfaces.

Results and discussion
Surface topography. Periodic hole-like micropatterns were directly fabricated onto a Cr sheet by using four-beam DLIP in a one-step process. The morphology of the laser-treated Cr surfaces, shown in Fig. 1a-c, resembles regular ordered arrays of microholes formed after laser irradiation. By controlling the overlapping angle between the individual beams, the spatial period of the pattern arrays was set to (Fig. 1a) 4.6, (Fig. 1b) 2.7 and (Fig. 1c) 1.6 µm. Accordingly, the mean diameter of the craters also decreased from 3.2 to 1.1 µm. The rims of the craters raise above the surface and form conical protrusion arrays, which is attributed to the displacement of molten material around the holes 41,42 .   Fig. 1d-f show the surface morphology of PET replicas as an example. Uniformly distributed pillar-like structures with a period between 1.6 and 4.6 µm were successfully obtained. The size of replicated microstructures was nearly the same as that of the holes in the Cr mold, demonstrating an excellent fidelity to the Cr mold dimensions.
The number of laser pulses N applied to the Cr sample was varied from 10 to 200 to gain a deeper insight into the influence of the laser processing conditions on the ablation efficiency and microstructures formation. The average depth of the laser-produced hole-like structures on Cr molds, and of the replicated pillar-like patterns on the polymers, were measured using confocal optical microscopy and are shown as a function of the number of laser pulses in Fig. 2. The results reveal a monotonic increase of the structure depths as the number of laser pulses increases, which can be well described by photo-thermal ablation behavior 43,44 . For the 4.6 µm period, the structure height tends to saturate after ~ 120 pulses (see Fig. 2a). This ablation depth saturation behavior has also been observed before 41 , which can be explained by recast material around the intensity maxima and multiple reflections in the keyhole evolution 45 . In addition, it can be observed that the structure depths for different periods differ significantly. For the larger interference period Λ = 4.6 µm, a maximum depth up to 2.6 µm was achieved, while for the 2.7 and 1.6 µm period, the maximum achieved depths were 1.5 and 0.62 µm, respectively.
After the hot embossing process, the heights of the pillar-like structures on PET and PMMA surfaces were also measured from their surface profiles. As shown in Fig. 2a, the mean structure height of the microfeatures with a period of 4.6 µm on PET and PMMA replicas are very similar to that on the corresponding molds (99.1% for PET and 98.2% for PMMA), indicating a complete filling of the soften polymers into the microcavities. On the other side, the maximum height of the pillars with a smaller period of 2.7 µm (see Fig. 2b) on PET and PMMA is 1.3 and 1.2 µm, respectively, compared to the structure depth of 1.5 µm on the Cr surface irradiated with 200 laser pulses (86.7% and 80.0% for PET and PMMA, respectively). This difference in structure height of plastic substrates and Cr mold is caused by incomplete filling during the hot embossing process, probably due to smaller cavities limiting the softer polymer to flow into them. This deviation in the height of the embossed structures from the mold is only observed on the 2.7 µm period samples irradiated with a higher number of pulses (> 150 pulses). In addition to the structure height, the aspect ratio (AR), defined as the quotient between the height and period of the micropillars (AR = depth/Λ), is used to describe the surface topographies. By increasing the number of laser pulses, the AR of the micropillars with a period of 4.6 µm increases from 12 to 57% on both PET and PMMA. Differently, AR up to 49% was observed for the 2.6 µm-period, while for the 1.6 µm patterns, the AR is smaller than 40%.

Static wetting behavior.
To evaluate the wettability of the imprinted polymeric surfaces, the static water contact angle of 7 µl-water droplets on of the PET and PMMA samples was measured before and after the hot embossing step. As shown in Fig. 3, the static contact angle of untreated flat PET (Fig. 3a) and PMMA (Fig. 3c) is 81 ± 1.8° and 64 ± 0.5°, respectively, demonstrating a hydrophilic behavior of pristine PET and PMMA without any surface texturing or chemical modification. After the periodic microstructures consisting of pillars with a depth of 2.6 µm and distributed with a period of 4.6 µm were replicated by hot embossing at 85 °C for 5 min, the wettability of PET is turned to a hydrophobic state with a maximum CA up to 139 ± 1.4° (Fig. 3b). Likewise, pillar-like features with exactly the same periodicity and depth imprinted onto PMMA films induced an increase in the CA to 120 ± 3° (Fig. 3d). These results demonstrate that the imprinted periodic microfeatures are capable to significantly increase the static CA on PET (or PMMA) for over 71% (or 88%) compared to its flat reference surfaces without any chemical modification.
To study the effect of surface topography on the wetting characteristics of plastic substrates, the measured CAs on patterned PET and PMMA surfaces were plotted as a function of the aspect ratio for different spatial periods ranging from 1.6 to 4.6 µm (Fig. 4a,b). The results show that the contact angle on both polymers presents an upward trend with the aspect ratio, independently of the structure period. In the case of the patterns with a period of 4.6 µm, the CAs on PET range from 91 ± 2° to 139 ± 1.4° as the aspect ratio increases from 12 to 57%. A www.nature.com/scientificreports/ similar trend was also observed for the patterns with a spatial period of 1.6 and 2.7 µm, although the maximum aspect ratio for the 1.6 and 2.7 µm-period is only 39% and 49%, respectively. However, it is worth noting that the periodic patterns with the smallest spatial periods of 1.6 µm show a higher contact angle than the surfaces with a spatial period of 2.7 and 4.6 µm, for comparable values of aspect ratios. A similar trend was also observed on the imprinted PMMA surfaces. In this case, for the 2.7 µm and 4.6 µm periodic surface patterns with a height of 0.41 and 0.55 µm, respectively, the CAs do not exceed 90° indicating a hydrophilic surface, although the CAs were increased by 33% compared to its flat reference. However, in case of the 1.6 µm period, a hydrophobic surface with a mean CA of 92 ± 1.8° can be achieved even with features with a low structure height of 0.18 µm. The periodic patterns with the smallest spatial period (Λ = 1.6 µm) also present higher contact angles compared to the larger ones, which is consistent with the result observed on imprinted PET (Fig. 4a). Similar results were reported on polyurethanes (PU) structured directly by DLIP method with line-like patterns featuring spatial periods between 2 and 5 µm 46 . The results showed that the structures with 2 − 3 µm periods present higher contact angles compared to the patterns with 4 -5 µm. This effect was explained by the differences in the surface topography of the structured samples. In the present experiments, the surface profiles in Fig. 4c-h also show that the periodic microfeatures on PET and PMMA have different topography shape for different periods. For instance, the profiles with Λ = 2.7 µm and Λ = 4.6 µm (Fig. 4c-f) exhibit a shape similar to half-spheres with narrow trenches separating them, which yields a relatively large area where the water droplets can sit. In contrast, the samples with the smallest period (Fig. 4g,h) have a triangular-like profile with sharp peaks and thus with a smaller droplet seating area. This evident different profile shape might cause the difference in the wetting behavior of micropatterns with similar aspect ratio but different spatial periods.
As described above, surface topography on the polymeric surfaces significantly influences the wettability, because the changes in the texture pattern and roughness cause variations in the actual liquid-solid contact area. In general, two models are used to describe the interactions of liquids and solids with textured surfaces: where θ w is the CA of the rough surface, W is the roughness factor defined as the ratio between the real surface area in contact with the liquid and its projected area, and θ 0 is the CA measured on a flat surface. The Wenzel model then predicts that an originally hydrophilic surface becomes more hydrophilic as the surface roughness increases, whereas a hydrophobic character increases when the starting surface is hydrophobic. In the case of the Cassie-Baxter model, the contact angle θ c is described by: (1) cos θ w = W cos θ 0 ,  (2) for better understanding the influence of the topographical characteristics in the wetting state of water droplets in contact with polymer surfaces. The detailed calculations can be found in the supplementary information. For comparison, Fig. 5a,b present the measured and calculated CAs as a function of surface texture aspect ratio for PET and PMMA substrates, respectively. In the case of PET, the predicted Wenzel angles θ w varies from 76° to 81° depending on the surface aspect ratio and the spatial period, while the Cassie-Baxter contact angles θ c show a hydrophobic wetting state ranging from 146° to 162°. Regarding PMMA, the θ w is between 45° and 62°, while the θ c ranges from 147° to 159°.
Besides, a significant discrepancy between the experimentally measured θ and the calculated θ w and θ c angles can also be observed. Furthermore, for both materials the measured contact angles θ lie between the CAs calculated with the models, and they increase with the surface aspect ratio. The results also show that the wetting state depends on further topographical characteristics of the imprinted microstructures, such as spatial period and depth. For example, for lower aspect ratios ~ 0-10% and in particular for larger spatial periods (e.g. 4.6 µm), the experimental CAs for both materials are closer to the predicted Wenzel CAs. However, clearly the measured contact angles θ do not show a decreasing trend as the Wenzel (θ w ) condition. Differently, for higher aspect ratios, the measured CAs are better described by the Cassie-Baxter condition. For example, the periodic structure produced on PET with Λ = 4.6 µm and AR = 57% (rightmost image of Fig. 5a), has a measured contact angle of (2) cos θ c = C cos θ 0 − (1 − C), www.nature.com/scientificreports/ 139 ± 1.4°, which is closer to the predicted Cassie-Baxter angle of 158° (compared to the corresponding Wenzel angle of 76°). Thus, deeper structures seem to be more efficient to prevent the water drops from penetrating into the structure valleys. Similar is the case for PMMA, where the highest CA of 120°, obtained for a pattern with the same aspect ratio, was only 24% lower than the calculated Cassie-Baxter angle (θ c = 157°). According to previous results 49, 50 , pure Cassie-Baxter or Wenzel wetting situations rarely occur, as the wettability of real surfaces is much more complicated. Since our experimental CAs lie between the Wenzel and Cassie-Baxter calculated angles, it can be assumed that the rough surfaces produced on both PET and PMMA adopt neither Wenzel nor Cassie-Baxter states, but an intermediate state. Many researchers have observed this behavior on microstructured surfaces as well, which is called composite or mixed wetting state [49][50][51][52][53] . In this state, the water droplet partly sits on air pockets as well as it partly penetrates the valleys of the rough solid surface. For example, He et al. measured a composite wetting state on a rough PDMS surface patterned with perfect square pillar arrays 52 . In this case, the measured contact angle of 152.5°, lies between the theoretical calculated Cassie-Baxter (165.7°) and the Wenzel angles (121°).
The reason for the composite wetting is still under discussion. According to Whyman et al., wetting of inherently hydrophilic surfaces is always energetically favorable, while Cassie-Baxter wetting is possible only when air is trapped into the rough surfaces 54 . Therefore, patterned superhydrophobic surfaces sometimes lack stability on Cassie-Baxter wetting state due to the low energy barrier separating the Cassie-Baxter and the Wenzel states. Lafuma et al. reported that the mixed state usually happens when the surface is decorated with a single pattern, which is not very rough, and its hydrophobicity is moderate (e.g. θ = 110°) 55 . Therefore, the Cassie-Baxter state of air trapping is metastable and irreversible transitions towards other states can be easily induced. However, recent studies reported that this composite wetting also exists on some complicated hierarchical surfaces, such as in rose petals 49,56 . Wetting transitions can also occur spontaneously or induced by external stimuli such as gravity, pressure, vibration, or bouncing of the droplet 49,54 . As the laser-induced surface morphology is typically complex, which usually lacks homogeneity due to the Gaussian profile of laser beams or due to additional surfaces features (e.g. nanoscaled laser-induced periodic surface structures (LIPSS)) 31 , it is difficult to obtain an accurate characterization of the microscopic wetting state. Therefore, for a better understanding of the wetting phenomenon at the microscale level, further investigation needs to be done in the future.

Dynamic wetting behavior.
To further investigate the wetting characteristics of the imprinted polymers, the dynamic contact angles of water on textured PET and PMMA with a periodicity of 2.6 μm were determined as shown in Fig. 6. Only one period is shown here since the results show negligible dependence on the period www.nature.com/scientificreports/ compared to the surface aspect ratio. The contact angle hysteresis Δθ is defined as the difference between the advancing θ a and receding θ r contact angles 57 : It can be seen in Fig. 6, that the advancing contact angles of water droplets on flat PET and PMMA surfaces are 81 ± 0.4° and 76 ± 0.6°, respectively, while the receding angles are 63 ± 1.3° and 57 ± 1.5°. The average hysteresis Δθ on PET and PMMA was calculated using the Eq. 3, obtaining the same value of 19° for both materials. H. Y. Erbil et al. reported receding contact angles θ r = 64°-66° for PET and θ r = 54°-64° for PMMA with an average hysteresis of 19.5 ± 1.5° (PET) and 23.5 ± 1.5° (PMMA) 58 , respectively, which is consistent with our measurements.
Regarding the textured PET and PMMA surfaces, the results show that both the advancing and receding angles increase with the texture aspect ratio. The hysteresis Δθ for textured PET varies in the range between 12 ± 9.6° and 23 ± 1.6° without showing a clear trend with the aspect ratio of the microstructure. In case of PMMA, a slight increase of the hysteresis was observed as a function of the aspect ratio, reaching values up to 38 ± 3.7°. According to the previous report by K.-Y. Yeh et al., when a droplet on a patterned surface falls into the Wenzel state, the advancing contact angles increase with an increase in the surface roughness, whereas the receding angle decrease 59 . On the other hand, if a droplet on a surface belongs to Cassie-Baxter state, the advancing and receding contact angles are independent of the surface roughness. In addition, they also reported that hysteresis Δθ remains constant independently of the surface roughness values. As observed in Fig. 6, any of these two wetting behaviors are consistent with our findings. Discontinuity of the receding contact angle and the contact angle hysteresis can be observed in Fig. 6, which further confirmed that the structured PET and PMMA surfaces adopt a composite wetting state.
The contact angle hysteresis can be used for estimating a measure of the "stickiness" of a surface 60 . D. Quéré et al. 57 calculated the droplet sticking condition according to the numerical calculation of the sliding of a liquid drop on a solid surface 61 : where l is the radius of a droplet which is in contact with a substrate, σ is the liquid/vapor surface tension, ρ is the density of the liquid, g is the gravitational acceleration, ζ is the drop volume (ζ = 4π/3R 0 3 ), R 0 is the radius of curvature of the droplet, and γ is the critical angle of inclination. By introducing an average angle θ (θ r = θ-Δθ/2 and θ a = θ + Δθ/2), the drop will remain stuck, when: where κ -1 is the capillary length, which is in the range of millimeters for most systems 57 . Small droplets (R 0 < κ -1 ) will stick when the substrate is inclined by a moderate angle, whereas a droplet that is significantly larger than the capillary length will move down because of the gravity 57 . For a 7 µl-droplet with a radius R 0 of 1.19 mm, the drop is found to roll only when Δθ < 4 -5° or θ > 160°; otherwise, it will stick to the surface 57 . However, when the tilting angle of the substrates is larger than the angle of sliding (or angle of repose), the droplet will move down.
An impression of the stickiness of water on the flat hydrophilic and structured hydrophobic PET surfaces can be grasped by looking at the photographs in Fig. 7. For better visualization, water was dyed by 0.1% KMnO 4 solution to create a colorful contrast. As it can be seen in Fig. 7a, the water droplets slide down on smooth untreated PET surface when the foil is tilted to a vertical position. On the contrary, the drops preserve their spherical cap geometries on the imprinted PET foil in each of the 9 × 9 mm 2 textured areas with pillar-like microstructures (Fig. 7b). The laser-patterned stamp was processed with different parameters corresponding to each of the πℓσ (cos θ r − cos θ a ) = ρgς sin γ ,   Figure 7c shows droplets remaining immobile and stuck to the microstructured areas on PET even when the sample is turned upside down (rotated by 180°), independently of the surface roughness. These results confirm a high adhesion of water to the structured surfaces, and a negligible influence of the surface roughness on the stickiness. Similar results were also observed on PMMA surfaces. Surfaces with a high contact angle and very strong adhesion to wetting fluid are termed parahydrophobic surfaces 62 . Rose petal is a prime example exhibiting this parahydrophobic behavior, known as "rose petal effect". The reason for the unique adhesive superhydrophobicity in this natural example is its special hierarchical microstructure, which is composed of periodic micropapillae arrays of 16 μm diameter and nanoscale cuticular folds of 730 nm width, decorating the top of the micropapillae 15 . The wetting state of water on rose petals was recently determined using confocal microscopy, showing that water does not fully wet all the surface features, but only partially wets the gaps between micropapillae in a composite wetting 56 . Likewise, the composite wetting is assumed to be the reason for the adhesive hydrophobic surfaces in this study.

Conclusions
In this study, we have reported a rapid approach for producing periodic pillar-like microarrays on PET and PMMA foils by utilizing Direct Laser Interference Patterning combined with the hot embossing method. Under optimized processing conditions, micropillar arrays with spatial periods from 1.6 to 4.6 µm and structure heights between 180 nm and 2.6 µm were successfully fabricated. The measured surface aspect ratio of micropillar arrays ranged between 12 and 57%. Static water contact angle measurements of the PET and PMMA replicas confirmed that the micropillar arrays can increase CAs leading to hydrophobic surfaces. The theoretical predictions of CAs were compared to systematic matching experiments. The results showed that the measured CAs lied between the calculated Wenzel and Cassie-Baxter CAs, indicating that a composite wetting behavior occurred on the rough surfaces. Both advancing and receding angles were found to increase with the surface roughness on both materials. The contact angle hysteresis slightly varied between 12° and 23° for PET and between 29° and 38° for PMMA. Furthermore, the water droplets presented a high adhesion to the polymer surfaces, which indicates that further modifications of the polymer surfaces are necessary to reach the super-hydrophobic conditions, which will be investigated in the future.

Experimental section
Materials. PET polymer foils were purchased from Pütz Folien GmbH + Co. KG and PMMA foils (PLEXI-GLAS@) from Evonik Performance Materials GmbH. According to the manufacturers, PET and PMMA have glass transition temperatures (T g ) of 85 °C and 109 °C, respectively. The thickness of both PET and PMMA were 200 µm, which were much thicker than the deepest structure formed. The used mold consisted of a chromium (Cr) sheet with a size of 87.4 × 66.7 mm 2 and a roughness (R a ) of 50 nm, provided by Sächsische Walzengravur GmbH.
Picosecond-direct laser interference patterning. A self-developed DLIP workstation (VIS/IR-DLIP μFAB, Fraunhofer IWS, TU Dresden) consisting of a solid-state picosecond-pulsed laser system (NeoLASE GmbH) with a wavelength of 532 nm and a pulse duration of 70 ps was utilized for texturing the Cr mold by overlapping four Gaussian beams (TEM00 mode) on the mold surface. The angle between the laser beams was varied to obtain hole-like structures with spatial periods of 1.6, 2.7 and 4.6 µm. The evolution of the topography on the Cr surfaces was studied as a function of the applied number of pulses N, namely 10, 40, 80, 120, 160, and 200 pulses at a constant fluence of 3.5 J cm -2 and a fixed repetition rate of 10 kHz. A detailed description of the used setup and the fabrication steps can be found elsewhere. 31 .
Plate-to-plate hot embossing. The as-prepared chromium sheets were utilized to replicate the laserproduced micropatterns on PET and PMMA films by using an electrohydraulic press (Paul-Otto Weber GmbH). The hot embossing process involves four steps 31 . First, the polymer foil was placed between two heated plates www.nature.com/scientificreports/