Saturation and texture geometry effects on plastron longevity on superhydrophobic surfaces Authors: Md. Abir Hasan Jilan, Estefania Solano-Calderon, Shabnam Mohammadshahi Published: 2025 ABSTRACT Superhydrophobic surfaces (SHSs), when immersed in water, sustain an underwater interfacial air layer (plastron) whose longevity governs performance. Under quiescent conditions this longevity is set by diffusive gas transfer. In this work, effects of dissolved-gas saturation and texture geometry over plastron longevity were quantified on 3D-printed micro-post arrays. Seven textures were fabricated by masked stereolithography and spray coating of nanoparticles, yielding apparent water contact angles more than 150°. Experiments were performed in quiescent deionized water at dissolved-air saturation of S = 20%, 40%, and 60%. Plastron evolution and wetting transition from the Cassie-Baxter to Wenzel states were tracked using two optical setups (top- and side-view). Across all geometries, the plastron lifetime τ increased with S. Geometry exerted a systematic influence. Increasing the inter-post gap g shortened τ, whereas increasing the post height h increased τ. Variations in the diameter d produced a comparatively weak effect over the tested range. Side-view imaging resolved stepwise depinning with transient pinning at multiple elevations along the post sidewalls. A compact response-surface model captured dependence on geometry and saturation within the measurement domain was proposed. These results provide insight and quantitative guidance for designing SHS textures that maximize plastron persistence under quiescent conditions.Supplementary InformationThe online version contains supplementary material available at 10.1038/s41598-025-28322-w. Introduction Super-hydrophobic surfaces (SHS) were first inspired by the lotus plant Nelumbo nucifera, where water droplets remain nearly spherical and roll off without wetting the leaf. This behavior, now reproduced in engineered SHSs, enables a range of functionalities including self-cleaning1,2, drag reduction3,4, anti-icing5–7 and anti-biofouling8–10 These characteristics arise from high water contact angles (WCA > 150°) and low sliding angles11–13, typically achieved through surface roughness and low surface energy regions14. Young15 introduced the earliest theory of SHSs, cosθ = (γsv − γsl)/γlv, where θ is the static contact angle, and γsv, γsl, γlv are the solid–vapor, solid–liquid, and liquid–vapor surface tensions. However, real surfaces are chemically and topographically heterogeneous. To account for roughness, Wenzel proposed a new model16, cosθ∗ = rw cosθ, where θ∗ is the apparent contact angle and rw is the surface roughness factor. This relation still cannot explain the high contact angles of SHSs, for which Cassie and Baxter, CB, introduced another model16, cosθCB = φs cosθ + (φs − 1), where φs is the solid-area fraction. When SHSs are submerged, entrap air within their microstructures, forming a plastron (Cassie-Baxter state). Sustaining this plastron is challenging and transitioning to the Wenzel state, by liquid penetration into surface textures, degrades SHS functionality. Over the past two decades significant attention has focused on SHS longevity17–41. Numerical investigations have clarified gas dissolution pathways. Lobaton and Salamon32 showed the shape of the gas–liquid interface depending on the texture’s shape, size and pitch. For a constant shape, the gas transfer rate was found to be constant during the depinning stage where the interface slides down along the sidewalls34. Bourgoun and Ling33 showed that gas fraction primarily governs longevity and is largely independent of the advancing contact angle. Hemeda and Tafreshi21 related plastron decay to the imbalance between hydrostatic and air bubble pressure. In experimental studies, plastron longevity has been quantified using several techniques. Bobji et al.19 applied total internal reflection (TIR) imaging technique, enabling visualization and measurement of plastron decay. Lv et al.20 conducted experimental studies under pressurized conditions and proposed scaling laws for plastron collapse. Nosrati et al.17 experimentally examined the role of undersaturation using optical diagnostics and established a scaling between plastron longevity and dissolved gas saturation. In their study, SHSs with micro-holes and micro-posts were fabricated by soft lithography technique and the effect of height was investigated. In addition, the hierarchical structure on submersed surfaces is important to further improve the stability of super-hydrophobicity and delay the wetting transition35–38. Despite these advances, the effect of texture geometry (height, gap, and diameter) and dissolved-gas saturation under quiescent conditions remains insufficiently quantified. An additional open question is how SHS longevity τ scales with liquid undersaturation level. To address this gap, textures with controlled height, gap, and diameter on identical substrates are required. Over the past two decades, different chemical and physical methods have been developed to fabricate SHSs. These techniques can be generally categorized into regular and irregular surface patterning methods. Regular patterns are typically produced via template method42,43, lithography44, and nanoimprinting45–47. More recently, additive manufacturing has emerged as a promising cost-effective fabrication approach. For example, Zhao et al.48 used fused deposition molding (FDM) followed by polydimethylsiloxane and silica coating while Sherin et al.49 employed stereolithography to create micro-pyramid structures, later coated with ZnO nanoparticles. In contrast, surfaces with random roughness have been fabricated using techniques such as chemical etching50,51, sandblasting52, spray coating53, dip coating54, plasma treatment55, phase separation56, and casting method57. In this work, SHSs with micro-pillars were fabricated using Masked Stereolithography (MSLA) 3D printing followed by spray coating of hydrophobic nanoparticle. This technique is cost-effective and scalable, avoiding the cleanroom tooling, multi-step templating, and labor-intensive processing typical of soft lithography and nanoimprinting; it enables rapid, repeatable production of well-defined textures and efficient parametric studies on identical substrates. Experiments are performed in a sealed, quiescent deionized-water environment at different controlled dissolved-oxygen saturation. The evolution of the plastron was monitored using non-intrusive optical techniques from two views (top- and side-view) enabling accurate tracking of the wetting transition. Plastron persistence was quantified as a function of gas saturation and surface geometry, and express stability in terms of image-based plastron coverage. Experimental methods Fabrication and characterization of SHSs SHSs with micro-posts were fabricated in two major steps: (i) microstructure construction by MSLA 3D printing and (ii) surface modification by hydrophobic nanoparticle coating. Microscale textures were fabricated via MSLA 3D printer (Elegoo Inc, Saturn 3 Ultra) using CHITUBOX Basic V2.2 Beta software (https://www.chitubox.com/en/article/whatsnew/chitubox-free/2.2.0). The printer has a XY resolution of 19 × 24 microns and can build a maximum volume of 218 × 122 × 260 mm3 (although the printer’s nominal pixel size is around 20 µm, it was found that features below 100 µm were not reliably reproduced after wash, UV cure, and coating, so the designs were limited to feature sizes ≥ 200 µm). The printer employs 405 nm UV light to solidify the standard photopolymer resin layer by layer. After 3D printing, samples were cleaned using compressed air and an ultrasonic cleaner (Sper Scientific, High Powered Ultrasonic Cleaner) to remove residual resin. The samples were later cured using 405 nm UV light in the curing station (Elegoo, Mercury XS Cure). Finally, hydrophobic nanoparticles (Glaco Mirror Coat Zero) were sprayed over the printed samples to modify the surface chemistry and create additional nano-scale surface roughness58–60. The coating was applied three times to ensure uniform coverage. A total of seven samples with size of 24 mm × 24 mm were and fabricated, all comprising of cylindrical micro-posts. The samples were designed to systematically vary one geometric parameter at a time, either the post diameter (d), height (h), or gap between adjacent posts (g), while keeping the others constant, Fig. 1a. Specifically, in samples #1 to #3, the gap g was varied from 200 to 600 μm, with both d and h fixed at 400 μm. In samples #4 and #5, diameter d and gap g were held constant at 400 μm, while the height h varied from 600 to 200 μm. Samples #6 and #7 had constant h and g of 400 μm, with diameter d varied from 600 to 200 μm. The uncertainty in texture dimensions of the current samples due to the 3D printer resolution was approximately ± 20 µm.Fig. 1a Schematic of a representative micro-post structure showing key geometrical parameters, b top- and side-view of the sample #2, and c water contact angle of sample #2. a Schematic of a representative micro-post structure showing key geometrical parameters, b top- and side-view of the sample #2, and c water contact angle of sample #2. The top- and side-views of baseline configuration (sample #2; d = h = g = 400 μm) are shown in Fig. 1b and were acquired using a 100 × objective on a Keyence VHX-S650E microscope. Side-view imaging shows a slight taper: the cap diameter is smaller than the base diameter (dtop < dbase), and the wall angle to vertical is about 5°. Prior work has shown that top-edge geometry and sidewall profile strongly influence plastron stability61,62. In this work, the sidewalls of all 3D printed textures exhibit a uniform taper with wall-to-vertical angle 5°, and this angle does not vary independently across the samples. Table 1 summarizes the dimensions of all 3D printed samples, and Fig. S.1 in the supplementary shows the side-view of all fabricated samples. In addition to geometric dimensions, Table 1 includes the solid fraction φs (the ratio of surface area covered by the solid). It was calculated, from measured dimensions, by φs = πd2top/4λ2, where λ = dtop + g is the texture wavelength. The Wenzel roughness, rw, (the ratio of total surface area to project surface area) is given by rw = (λ2 + πdavgh)/λ2. The table also includes the static water contact angle (θ0). These contact angles were measured optically using small water droplets carefully placed on each sample, with at least three measurements taken per sample to ensure repeatability using ImageJ software (https://imagej.net/software/imagej/). The sliding (roll-off) angle on a tilting stage was measured using the water droplet (~ 10 μl). All textures showed low sliding angles of 7–9° which indicate weak pinning and are consistent with a Cassie state during motion over these feature sizes. Figure 1c shows a water droplet on sample #2 with a contact angle larger than 150°, confirming its super-hydrophobicity. Throughout the manuscript, d denotes the cap (top) diameter, dtop.Table 1List of fabricated SHSs with corresponding texture width (d), texture height (h), texture gap (g), solid fraction φs (based on dtop), Wenzel roughness rw (based on davg), and static water contact angle (θo).Sample #dtop (μm)dbase (μm)h (μm)g (μm)φsrW\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\theta_{o}$$\end{document} (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$^\circ$$\end{document})1 4004704002000.352.52151 ± 12 4004704004000.21.85152 ± 13 4004704006000.131.55151 ± 14 4005056004000.22.54152 ± 25 4004352004000.21.38152 ± 16 6006704004000.281.79151 ± 27 2002704004000.091.82151 ± 1 Table 1List of fabricated SHSs with corresponding texture width (d), texture height (h), texture gap (g), solid fraction φs (based on dtop), Wenzel roughness rw (based on davg), and static water contact angle (θo).Sample #dtop (μm)dbase (μm)h (μm)g (μm)φsrW\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\theta_{o}$$\end{document} (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$^\circ$$\end{document})1 4004704002000.352.52151 ± 12 4004704004000.21.85152 ± 13 4004704006000.131.55151 ± 14 4005056004000.22.54152 ± 25 4004352004000.21.38152 ± 16 6006704004000.281.79151 ± 27 2002704004000.091.82151 ± 1 List of fabricated SHSs with corresponding texture width (d), texture height (h), texture gap (g), solid fraction φs (based on dtop), Wenzel roughness rw (based on davg), and static water contact angle (θo). Experimental setup Each of the seven samples was tested under three different dissolved-oxygen saturation levels: S = 20%, 40%, and 60%, all conducted at 23 °C. Each condition was repeated at least twice, resulting in a total of 42 experimental runs. The concentration of gas in the deionized (DI) water was achieved using a vacuum pump and vacuum chamber (Vevor, KQ-1K). To prepare each level of saturation, DI water was kept at negative pressure in the vacuum chamber: for S = 60% saturation, the water was held at -35 kPa for at least 24 h; for S = 40% and 20% saturation, the pressure were -50 Kpa and -70 Kpa, respectively. Before each experiment, the concentration of dissolved oxygen was verified using a dissolved oxygen meter (UIUZMAR, UMR-BLE9100). The tests were performed in a custom-designed sealed chamber to maintain the gas concentration during experimentation. The chamber consisted of a stainless-steel base and transparent acrylic windows on all sides. Samples were mounted vertically along one of the acrylic walls, allowing optical access from both the top and side for visualization. The sealed configuration minimized the influence of ambient disturbances such as air exchange and evaporation, ensuring consistent experimental conditions. Optical setup for plastron visualization The status of the trapped air pocket or plastron over SHSs was examined using two different optical setups as shown in Fig. 2. The first setup was based on the reflected light imaging system. A CMOS camera (FLIR, model GS3-U3–41C6M-C) with 2048 × 2048 pixels and maximum fps of 90 (labeled camera I) was used to capture the reflected light from the SHSs via the FlyCapture2 software (https://www.teledynevisionsolutions.com). Illumination was provided by a collimated red LED (Thorlabs, model M625L4-C1, wavelength 625 nm), directed through a beam splitter. A long-working-distance objective lens (Mitutoyo 5×, WD = 34 mm) was used to magnify the image. A similar reflected-light imaging approach was used previously by authors58,60 to examine the gas fraction on a SHS with random roughness under shear flows.Fig. 2Optical setups for recording plastron status: a schematic view and b image of the setup. Optical setups for recording plastron status: a schematic view and b image of the setup. Figure 3a shows the initial plastron status of all samples using this optical setup. In this figure, the trapped air appears bright due to stronger reflection, whereas wetted areas appear dark due to absorption by the resin. This brightness contrast arises from the difference in refractive indices at the material interfaces. When light from the source strikes the SHS, a small portion is reflected and detected by the camera. The reflected light intensity to the incident light intensity is governed by the Fresnel equation:1\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$I_{r} /I_{i} = \left[ {\left( {n_{2} - n_{1} } \right)/\left( {n_{2} + n_{1} } \right)} \right]^{2}$$\end{document}where \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${I}_{r}$$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${I}_{i}$$\end{document} are the reflected and incident light intensities, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${n}_{1}$$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${n}_{2}$$\end{document} are the refractive indexes of the media the light is going through. Here, the refractive indexes for air, water and resin 1.0, 1.33, and 1.455, respectively. As a result, the reflected intensities from the water-resin and resin-air interfaces are approximately 0.002 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${I}_{i}$$\end{document} and 0.034 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${I}_{i}$$\end{document}, respectively. Thus, the light reflected from the air trapped areas becomes larger than the light reflected from regions where there is no air.Fig. 3a top-view and b side-view of the initial status of the plastron over SHSs. a top-view and b side-view of the initial status of the plastron over SHSs. The second optical setup was used to visualize the plastron from the side of the SHS samples. Another identical CMOS camera, labeled camera II in Fig. 2, was equipped with a variable magnification lens (Infinity 3x, working distance infinity to 16 mm) and recorded the images via the FlyCapture2 software (https://www.teledynevisionsolutions.com). Illumination was provided by a broadband Quartz Tungsten-Halogen lamp (THORLABS, model QTH10/M). This configuration allowed clear observation of the plastron along the sample edge. Figure 3b shows side-view images of the plastron using this optical method. This configuration resolves the vertical position and shape of the air–water interface along the post sidewalls. The last image in Fig. 3a and b corresponds to the fully Wenzel state. Results Figure 4a and b shows the gas dissolution process of sample #2 in stationary water with S = 20% saturation at room temperature. Initially, bright regions in top-view images are observed across the surface, indicating the presence of air pockets trapped between the micro-posts, showing the Cassie–Baxter state. As explained earlier, these air-filled regions reflect more light, producing higher image intensity due to the refractive index contrast at the air–water interface. As immersion time increases, the bright regions gradually diminish, indicating the progressive dissolution of the trapped air into the surrounding water. This results in partial wetting and a metastable state, followed by complete wetting. The final dark images correspond to the Wenzel state, where water fully penetrates the surface texture and no trapped air remains. In this condition, the transition from Cassie–Baxter to Wenzel state occurred by approximately 120 min, driven by gas diffusion from the plastron into the undersaturated bulk liquid.Fig. 4a top-view and b side view of the plastron for sample #2 at S = 20%, and c normalized gas-covered area vs. time for sample #2 at S = 20%, 40%, 60%. Lines are the mean value across two independent runs for each condition, shaded band show the range (min–max) across experiments, the side view confirmed the top-view metric within uncertainty and is not counted toward. a top-view and b side view of the plastron for sample #2 at S = 20%, and c normalized gas-covered area vs. time for sample #2 at S = 20%, 40%, 60%. Lines are the mean value across two independent runs for each condition, shaded band show the range (min–max) across experiments, the side view confirmed the top-view metric within uncertainty and is not counted toward. Side-view images in Fig. 4b further show the gas dissolution dynamics from the edge of the SHS samples. At the beginning of the experiment, a flat and continuous air pocket was clearly visible, covering the gaps between micro-posts, confirming the Cassie–Baxter state. As the immersion time increased, the air–water interface began to recede, and the height of the plastron decreased progressively. This led to the formation of partial wetting and a metastable intermediate state. Eventually, the trapped air fully dissolved into the water, and the surface transitioned into Wenzel state, marked by the complete disappearance of the air layer. This side-view observation confirms that the plastron was fully lost by approximately 120 min, consistent with the timeline observed in the top-view images. Side-view images also reveal that due to the hierarchical micro/nanostructures, the plastron did not collapse uniformly. Initially trapped at the top of the micro-posts, the air–water interface progressively receded and became pinned at multiple intermediate heights along the post sidewalls. This stepwise depinning indicates that the multiscale roughness promoted metastable states during the transition from Cassie–Baxter to Wenzel. To quantify the plastron longevity, the air-pocket coverage of each SHS over time was analyzed using the top-view images. The percentage of bright surface area was calculated and normalized as the intensity area fraction:2\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$I_{n} \left( t \right) = A\left( t \right)/A_{0}$$\end{document}where A(t) is the bright-region area at time t and A0 is the bright-region area at the beginning of the experiment (from top-view images). Values of In ≈ 1 indicate the full air coverage between micro-posts while the plastron between different posts is connected and creates a single initially large dry region. As In → 0, the surface reaches the Wenzel state (complete wetting). Figure 4c shows the time evolution of In for sample #2, under three levels of S (20%, 40%, and 60%). To present variability clearly and avoid over-plotting, the mean curve is reported with a shaded band. At the initial period of gas diffusion process, In remains nearly constant at 1. As the immersion time increased, In is reduced until wetting happens. Clearly, by reducing S (i.e., increasing the undersaturation level), In decreases faster with time, indicating larger gas diffusion rates. This trend is consistent among different SHSs as shown in the supplementary figure S.2 (values of In versus time for all seven samples). Instead of raw images, time-series plots that capture the overall plastron evolution are showed. Because interface height is not uniform during depinning, In was used as the primary scalar; the side-view images qualitatively confirm interface recession, wetting, and the timing of plastron loss. Although the texture is regular, some regions get wet earlier than others, consistent with prior reports19,63. Figure 5a summarizes the plastron lifetime τ (time for the Cassie–Baxter state to transition to the Wenzel state) for all samples as a function of dissolved air saturation S (20%, 40%, 60%). As expected, the S = 60% level yielded the longest plastron lifetimes across all samples, whereas the S = 20% level creates the fastest transitions. At lower S, the water can absorb more gas, accelerating dissolution; at higher S, the driving force is reduced, so the plastron persists longer and is more stable. Consequently, τ is largest at S = 60%, intermediate at S = 40%, and smallest at S = 20%. Figure 5b shows geometric sensitivity using Sample #2 (d = h = g = 400 μm) as the baseline. Varying the gap g at fixed d = h = 400 μm (Samples #1, #2, and #3) yields a decrease of lifetime τ with increasing g across all S. Wider spacing lowers the capillary pressure, weakening pinning and facilitating earlier wetting, despite the larger gas fraction at higher g. Changing the height h at fixed d = g = 400 μm (Samples #5, #2, and #4) produces the opposite trend: τ increases with h. Taller posts provide greater trapped-air volume and more intermediate pinning levels along the sidewalls, thereby slipping the air–water interface due to sublevel structures and sustaining Cassie/partial-Cassie configurations for longer before collapse to Wenzel. These trends are consistent with prior reports39–41. By contrast, when the diameter d at fixed h = g = 400 μm (Samples #7, #2, and #6) changes, no significant changes are observed in τ. Notably, Sample #4 (square symbol) exhibited the longest plastron stability among all cases across all three levels, consistent with its larger reservoir. In addition, the geometry is parameterized by the gap g rather than the wavelength (d + g) because the capillary pressure (pc = γ/g) governing pinning depends on the gap g, giving g direct physical meaning; moreover, a single pitch value can map to multiple (d, g) pairs. This choice aligns with the gap-based impalement literature64,65.Fig. 5a lifetime τ (mean value) of all samples versus saturation level, and b effects of surface textures on τ. Symbols are mean value across n independent runs per condition (n = 2–4; Table S1 for counts by condition). Error bars: range (min–max) across runs. Top- and side-view measurements were obtained for all runs; side-view verified the top-view data and is not counted toward n. a lifetime τ (mean value) of all samples versus saturation level, and b effects of surface textures on τ. Symbols are mean value across n independent runs per condition (n = 2–4; Table S1 for counts by condition). Error bars: range (min–max) across runs. Top- and side-view measurements were obtained for all runs; side-view verified the top-view data and is not counted toward n. These empirical trends admit a simple theoretical interpretation. In stationary water, based on Fick’s first law66, loss of the plastron is diffusion-limited:3\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$J = - D_{G} \;\partial c/\partial x \approx \left( {c_{i} - c_{\infty } } \right)D_{G} /\delta$$\end{document}where, J is the gas flux, DG the diffusivity in water, δ an effective stagnant layer thickness, ci the interfacial concentration (csat based on Henry equilibrium), and c∞ = S csat the bulk concentration at dissolved-air saturation S. Hence Δc = csat (1 − S), thus4\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\left| J \right| \approx D_{G} c_{sat} \left( {1 - S} \right)/\delta$$\end{document} And the plastron lifetime scales as the trapped gas volume per unit projected area divided by the flux5\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\tau \approx V_{g} /\left( {A \times \left| J \right|} \right)$$\end{document}where the initial Vg/A, trapped gas volume per unit projected area, for the current micro-posts is6\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\left. {V_{g} /A} \right|_{t = 0} \approx \varphi_{g} h = h\left[ {1 - \left( {\pi d^{2} } \right)/4\left( {d + g} \right)^{2} } \right]$$\end{document}where φg = 1 − φs. Combining these equations yields the leading scaling:7\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\tau \propto \left[ {\varphi_{g} h/\left( {1 - S} \right)} \right] \times \left[ {\delta /D_{G} c_{sat} } \right]$$\end{document} The same scaling can also be written using an averaged mass flux J = mo/Aτ; where mo/A ∝ Vg/A is the initial gas per unit projected area. If the diffusion layer thickens in time, so J ∝ t-1/2 and mass balance yields:8\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\tau \propto (\varphi_{g} h)^{2} /\left[ {D_{G} c_{sat}^{2} \left( {1 - S} \right)^{2} } \right]$$\end{document}which predicts the same qualitative trends as Eq. 7. These explain two key observations: at fixed h, decreasing g strengthens capillary pinning (higher capillary pressure PCa = γ/g tending to lengthen τ); but it also reduces the gas fraction and the initial Vg (tending to shorten τ); the net trend indicates pinning dominates over reservoir changes. At fixed g and ϕs, τ grows almost linearly with h since it enlarges the initial Vg and therefore lengthens the diffusion-limited lifetime τ. Figure 6a shows that the lifetime increases with the nondimensional height h/λ for all saturation levels and for all fabricated samples. To test the theoretical models, the dependence of lifetime on undersaturation level was examined. Consistent with Eq. (8), plotting τ against (1 − S) on log–log axes yields an exponent close to -2, Fig. 6b. These results are also in agreement with previous results of Bourgoun and Ling33 and Nosrati et al.17. All the data analysis were performed using MATLAB software (https://www.mathworks.com).Fig. 6a plastron’s lifetime τ vs. normalized texture height h/λ for all samples, and b τ as a function of 1 − S for all SHSs. a plastron’s lifetime τ vs. normalized texture height h/λ for all samples, and b τ as a function of 1 − S for all SHSs. To further explain the geometry-lifetime trends, the free-energy difference between Wenzel and Cassie-Baxter states per unit projected area nondimensionalized by the liquid–vapor surface tension γ was evaluated:9\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\Delta E/\gamma A = \left( {E^{Wenzel} - E^{Cassie} } \right)/\gamma A = \left[ {\left( {\varphi_{s} - r_{w} } \right)\cos \theta - 1 + \varphi_{s} } \right]$$\end{document}where θ is water contact angle on an ideally flat surface of the same material (here θ = 150.5°, Supplementary Fig. S1). In addition, ΔE/γA is computed considering only interfacial energies21. A positive ΔE/γA indicates that the Cassie-Baxter state has lower interfacial energy than the Wenzel state and is therefore thermodynamically stable. Table 2 listed the values of ΔE/γA. Increasing the gap g at fixed d and h (Samples #1, #2 and #3), enlarges the texture’s wavelength λ, reducing both φs and rW (listed in Table 1). The net effect is a decrease of ΔE/γA with g, from 1.13 to 0.32 for samples #1 to #3, a weaker Cassie preference and faster transition, consistent with the shorter lifetimes at wider spacing. In contrast, increasing the post height h at fixed d and g increases free-energy difference from 0.24 to 0.91 for samples #5 and #4 showing a stronger Cassie preference and longer persistence. Changing the diameter d at fixed g and h shows a comparatively small ΔE/γA and therefore lifetime changes. It should be noted that energetics alone (Eq. 8) does not set the time to transition, and lifetime also follows a diffusion-limited scaling (Eq. 7). So, geometry affects both energy (ΔE/γA) and kinetics. This resolves the observed case where Sample #4 outlives.Table 2Non-dimensional energy barrier of different samples.Sample#1#2#3#4#5#6#7\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\Delta E/\gamma A$$\end{document}1.130.590.320.920.240.560.49 Table 2Non-dimensional energy barrier of different samples.Sample#1#2#3#4#5#6#7\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\Delta E/\gamma A$$\end{document}1.130.590.320.920.240.560.49 Non-dimensional energy barrier of different samples. Conclusion In this work, effects of dissolved-oxygen saturation and texture geometry over the plastron longevity of 3D-printed superhydrophobic surfaces with micro-post arrays were experimentally investigated. Samples were fabricated by MSLA 3D printing which allowed independent control of inter-post gap g, height h, and diameter d. Hydrophobic nanoparticle coating was applied on the 3D printed samples. Then, samples were tested in quiescent water, without any interference of hydrostatic pressure and shear flow, at different saturation levels of S = 20%, 40%, and 60%. Two different optical setups were developed (top-view and side-view) to visualize plastron status and track the transition from Cassie-Baxter to Wenzel state. The most important findings are summarized as follows:Using MSLA followed by hydrophobic coatings, post arrays exhibiting water contact angles more than 150°, confirming superhydrophobic behavior.Across all geometries, the plastron lifetime τ increased with saturation level and wetting transition took longer time at higher saturation levels.The geometry trends arise from coupled effects of capillary entry pressure and the initial gas volume which together set the diffusion-limited timescale.Geometry exerted a systematic influence: enlarging the inter-post gap g shortened τ; increasing the post height h increased τ; and varying the post diameter d produced a comparatively weak response over the tested range.Side-view imaging resolved stepwise depinning of the air–water interface, with temporary pinning at multiple elevations along the post sidewalls, behavior that is characteristic of hierarchical micro/nano-roughness. Using MSLA followed by hydrophobic coatings, post arrays exhibiting water contact angles more than 150°, confirming superhydrophobic behavior. Using MSLA followed by hydrophobic coatings, post arrays exhibiting water contact angles more than 150°, confirming superhydrophobic behavior. Across all geometries, the plastron lifetime τ increased with saturation level and wetting transition took longer time at higher saturation levels. Across all geometries, the plastron lifetime τ increased with saturation level and wetting transition took longer time at higher saturation levels. The geometry trends arise from coupled effects of capillary entry pressure and the initial gas volume which together set the diffusion-limited timescale. The geometry trends arise from coupled effects of capillary entry pressure and the initial gas volume which together set the diffusion-limited timescale. Geometry exerted a systematic influence: enlarging the inter-post gap g shortened τ; increasing the post height h increased τ; and varying the post diameter d produced a comparatively weak response over the tested range. Geometry exerted a systematic influence: enlarging the inter-post gap g shortened τ; increasing the post height h increased τ; and varying the post diameter d produced a comparatively weak response over the tested range. Side-view imaging resolved stepwise depinning of the air–water interface, with temporary pinning at multiple elevations along the post sidewalls, behavior that is characteristic of hierarchical micro/nano-roughness. Side-view imaging resolved stepwise depinning of the air–water interface, with temporary pinning at multiple elevations along the post sidewalls, behavior that is characteristic of hierarchical micro/nano-roughness. These findings advance the understanding of wetting transitions on textured SHSs and provide quantitative design guidance for maximizing plastron persistence in practical applications (under the present undersaturated conditions, no persistent nanoscale plastron was resolved after the Cassie-Wenzel transition within the resolution of the imaging). Future work should extend to non-quiescent conditions to quantify the roles of hydrostatic head, imposed shear, or mechanical vibrations to assess robustness of samples. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Supplementary Material 1