Interfacial confinement of water-forming reaction at FP cavity for dynamic plasmonic coloration
The FP resonance cavity, in which a gas-permeable polymer is sandwiched between the top and bottom Pd metals, was prepared to improve the performance of the gas-induced plasmonic coloration (Fig. 1a). Only H2 and O2 penetrate and reach the polymer/bottom Pd interface where a catalytic reaction is expected. According to the pertinent literature, the dissociative adsorption of H2 and O2 occurs competitively on the Pd surface, forming H2O [21, 22, 23]. Based on DFT calculations (Supplementary Note 1), exothermic steps for H2O generation with different energy pathways are expected as a function of the surface chemistry of the Pd(111) surface (O- vs. H-covered Pd). The H-covered Pd surface shows preferred H2 dissociation over the O-covered Pd surface due to the exothermic transition from state II to III (Fig. 1b). This led us to assume that different ratios or types of H2O genesis are expected. For instance, the different H2O genesis pathways and resultant distinct H2O types (i.e. bubbles: Fig. 1c and film: Fig. 1d) may cause distinct optical phenomena (optical foggy effect: Fig. 1e and whiteout effect: Fig. 1f). Then, we can monitor the atomic scales of surface catalytic reactions in color via light-matter interactions. Consequently, real-time and dramatic color changes in response to the alteration of the physical properties of the dielectrics (polymer insulating layer and newly formed water film) of the FP etalon can be realized by coupling the light-matter interaction of the FP cavity with the gas-matter reaction, which permits the water droplets/film and concomitant structural color change through gasochromic observation.
Water bubble formation at the polymer/Pd interface
Figure 2a shows the Pd-assisted catalytic reaction that produces water bubbles. At equilibrium, as the physisorption of O2 on the Pd surface is thermodynamically more favorable than that of H2, the surface coverage of adsorbed H2 is relatively small [25]. High surface coverage of O2 on the Pd surface interferes with the dissociative adsorption of H2 involved with the correlation between the d-orbital of Pd and the s-orbital of H, degrading the efficiency of the water-forming reaction. When the coverage of dissociated H2 on the Pd surface exceeds the threshold, spontaneous genesis of water film (O2 + 2H2 = 2H2O) occurs at room temperature, experimentally proven under the critical pressure ratio of O2/H2 < 2.5 (Fig. 2a) [21]. However, for spatially H2-covered spots surrounded by overwhelming O2-covered area, the water-forming reaction will occur in the forms of ‘water bubbles’ discarding the possible scenarios of a nanometer-thick water film and their successive growths via coalescence processes. In addition, the H2O evaporation rate should be suppressed and minimized than the H2O generation rate to maintain the pre-formed H2O bubbles. For this, the Pd surface was coated with a 180-nm thick perfluoro-(butenyl vinyl ether) (PBVE) film as a sieve layer with selective gas permeability only for O2 and H2 [25]. Oxygen permeability through a Pd-covering PBVE membrane is essential to trigger the water-forming reaction. As a control experiment, we compared the gas colorimetric behavior of another sample prepared using O2-impermeable polymethyl methacrylate (PMMA) as a dielectric film. There were no water bubbles under identical H2 conditions, confirming the role of dielectric films in water formation via selective gas permeability (Supplementary Fig. S1 and Supplementary Note 2).
To investigate the formation of water bubbles by H2 adsorption at the PBVE/Pd interface, a sample with PBVE/Pd/titanium (Ti) (180 nm/10 nm/10 nm) on a silicon (Si) substrate was exposed to 10% H2 in a gas chamber (N2:O2 = 4:1 as a carrier gas), and the dynamic water bubbles formation was monitored in real-time under a microscope (Fig. 2b). A Ti layer promoted adhesion between the PBVE and Pd films and suppress abrupt phase change (from α to β phase) causing Pd film deformation such as peeling and wrinkles [26]. Randomly positioned outbreaks and growth of micron H2O bubbles, followed by coalescence between neighboring bubbles, were observed (<130 s), which varies from previous studies of nanometer-thick transient water generation without a Pd-covering membrane [22, 23, 27]. After turning the H2 supply off (>130 s), we observed a gradual decrease in the water bubble volume and subsequent disappearance (225 s) equal to the initial state (0 s).
Three-dimensional (3D) PBVE deformation during the water-forming reaction was confirmed via fluorescence interference contrast microscopy (FLIC) [28]. For fluorescence (FL) visualization, a thin layer of calcein as the fluorescent dye was coated atop the PBVE surface, followed by 10% H2 exposure (Fig. 2c). Interference-assisted alternative bright/dark patterns arise from the vertical movement of the dyes, and fluorescein isothiocyanate (FITC)-labelled calcein and the resultant fringe patterns allow an analysis of the height of water bubbles (Fig. 2d and Supplementary Video 1). Figure 2e shows the normalized FL intensities measured from Fig. 2d as a function of the water-bubble scanning length with corresponding simulation results. The intensity profiles of the fringe patterns along the dashed lines confirm that more fringes can be created as the exposure time or H2 concentration increases. Tracking FL molecules via simulations revealed that the PBVE membrane swelling owing to the water bubble growth underneath the calcein layer experienced a height alteration of 138 nm (red), 290 nm (blue), and 470 nm (orange) during 100 s (Fig. 2e and 2f; Supplementary Fig. S2 and Supplementary Note 3).
Optical foggy effect by water bubbles induced light diffusion
This study realizes various color changes by reinforcing light-matter interactions based on the volume change of the dielectric layer. To boost the light-matter interaction, the FP resonator [29], underpinning the naked-eye observation of the water-forming reaction, was adopted by depositing another Pd layer on the PBVE surface. A metal-dielectric-metal (MDM) structure, known as the FP etalon, enables sensitive color variations as a function of physicochemical properties, such as the thickness and refractive index (RI) of the dielectric layer [30, 31, 32, 33]. Thus, their modulation induced by water bubble formation and growth at the PBVE/Pd interface would cause a dramatic color shift. In addition, hemispheric water bubbles may cause diffuse reflection of incident light as they act as individual lenses (Fig. 3a).
Designing a perforated top metal layer with efficient transreflective characteristics is essential for generating the FP resonating effect. The optimum deposition thickness of 15 nm with a volumetric filling fraction of 0.45 for the nanoparticulated Pd layer was selected through experimental optimizations and a computational simulation to advance the FP resonance effect for producing wider range of color displays (Fig. S3 and Supplementary Note 4). The top-view (Fig. 3b) and side-view (Fig. 3c) scanning electron microscopy (SEM) images show that a 15-nm-thick Pd layer was deposited on the PBVE film. The FP etalons on a 4-inch Si wafer with varying dielectric PBVE thickness, tPBVE, values from 0 nm to 590 nm at 10 nm intervals were controlled by a delicate plasma etching process, showing excellent correspondence in reflection color trends with simulation results (Fig. 3d and 3e; Fig. S4 and Supplementary Note 5). The tPBVE increases from the left-bottom corner (0 nm) to the upper-right corner (590 nm) (Supplementary Fig. S4). Figure 3e shows that reflection colors of the FP etalon turn discolored and eventually become foggy when exposed to 10% H2 under atmospheric conditions. As predicted in Fig. 3a, diffuse reflection due to newly formed micron-water bubbles interrupts the FP resonance, thus resulting in foggy colors (bottom-right photograph of Fig. 3e, at 6 min), and this transition is fully reversible (Supplementary Video 2). After establishing that color transition occurs due to water bubble formation in the dielectric layer, we quantified the critical gas values that trigger color transitions. The reflectance spectra at the center yellow area (red-dashed square in Fig. 3e) were measured as H2 concentration was gradually increased and decreased by 1% every 5 min (Fig. 3f). When H2 concentration reaches 7%, nearly corresponding to the critical ratio of O2/H2 = 2.5, the waveform of reflectance over the visible range begins to flatten (lowering peaks and raising dips shown in Fig. 3g). By plotting the peak reflectance around 400 nm wavelength depending on H2 concentration (Fig. 3g), the water-forming reaction vigorously occurred from 7% to 8% and then decreased from 4 % (Fig. 3h). The rapid reduction in the peak reflectance over 7% H2 indicates that the water bubble formation and coalescence are instantaneously accelerated at around a critical ratio of 2.5. The slow recovery of the peak reflectance with decreasing H2 concentration is probably attributed to the thermodynamically unfavorable desorption of H2O molecules at the Pd surface (endothermic transition from state V to VI in Fig. 1b). At 4% H2, the recovery rate gradually increased because minuscule water bubbles (bottom left photograph of Fig. 3e, at 16 min) only existed. An H2 concentration-dependent change in the peak reflectance (lowering peaks and raising dips) was continuously observed in multiple H2 injection/purging cycles, demonstrating the excellent repeatability (≥ 28×) of our FP etalon, considering that typical metal-hydride-assisted color displays showed limited (≤ 20×) operating cycling numbers (Fig. 3i) [34].
Multichromatic change by the water-film formation and following whiteout effect
O2-predominated Pd surface promotes random water bubble formation under H2 gas exposure (Fig. 2 and 3). Figure 1 shows that the Pd-dominated gas can determine the surface chemistry and resultant water formation types. We investigated the colorimetric changes in the FP etalon after preparing the H2-predominated Pd surface before the O2 gas supply and examined whether such different surface properties promote distinct forms of water products. In accordance with previous studies that a nanoscale water film could be formed when a Pd surface is entirely preoccupied with H2 before the O2 gas supply [23], the FP etalon was exposed to 10% H2 for 5 min to induce full β-phase Pd hydride (PdHX) and H2-predominated Pd surface (Fig. 4a) [35], followed by 20% O2 gas supply. The uniform formation of nano-water film at the PBVE/Pd interface is useful in the optical interpretation since the RI of water (n = 1.33–1.34 in the visible range) is similar to that of PBVE (n =1.34 at 550 nm and 1.333 at 1.55 um). Thus, generating a nanoscale water film underneath the PBVE layer can be simply interpreted as thickening the dielectric layer of the FP etalon, enabling multichromatic changes as a function of the water film thickness. Figure 4b shows reversible and drastic color changes in the 10% H2 pre-exposed etalon under 20% O2 gas exposure. Compared with the color changes above (Fig. 3), a rapid and dramatic color change was observed, producing an exotic optical whiteout effect in 10 min. In addition, reversible color recovery was completed within 20 min after cutting off the O2 supply.
To elucidate the water-film thickness-based color variation on the FP etalon, we analyzed the time-lapse spectra at the center yellow area (red-dashed squares in Fig. 4b) with increasing O2 concentrations up to 20% by 2% every 1 min (Fig. 4c and Supplementary Video 3). The red shift of the measured waveform was initiated at 12% O2 concentration (8 min) and maximum before the gas cut-off (12 min). The red shift in the visible range is attributed to the increased effective thickness of the dielectric film (PBVE + water). A simulation was conducted by varying tdielectric to confirm the effect of dielectric film thickening on the color shift of the FP etalon (Supplementary Fig. S5 and Supplementary Note 6). Simulations verified that the measured spectral shifts arose from the dielectric thickening effect, that is, additional H2O film formation below the PBVE layer. The second resonance peaks (Fig. 4d), extracted from the reflection spectra (Fig. 4c), show not only a red shift but also an overall decrease in magnitude (Fig. 4d). From an optical analysis perspective, the thickening of the dielectric film caused the coexistence of higher-order resonances in the visible range, resulting in the whiteout of the FP etalon (Fig. 4e).
Scalable and transparent display for gas detector via water-forming reaction
Implementing the water-forming reaction at the gas-permeable polymeric membrane and Pd interface can cause substantial innovations in various hydrogen applications. Forthcoming commercial coloration device technology must fulfil both aesthetic elements and outstanding performance and must be prepared in a facile fabrication process over various solid supports (such as a transparent and flexible substrate) in a scalable manner. Furthermore, it would be beneficial if the MDM could comprise various dielectric and metal components with identical functions. Respectively, our proposed technology can realize various theme colors by simply controlling the thickness of the metal and polymer layers. Furthermore, delicate and aesthetic drawings or patterns of a few microns can be easily fabricated via compatible semiconductor processes, meeting demanding industrial needs with high prospects for commercialization and technological standardization.
By depositing all metal layers through a shadow mask, a flower-patterned etalon (f-etalon) was prepared on a 2.5 cm × 2.5 cm glass substrate (Fig. 5a). Because tdielectric determines the overall reflection colors, spatially selective PBVE etching by the RIE process produces a wide color range of flower patterns (Fig. 5b and Supplementary Fig. S6). The top Pd, as a trans-reflective layer, can be replaced by other perforated metal films, including gold (Au) or silver (Ag), as they maintain color changes in response to the gases (Supplementary Fig. S7). For accomplishing water-forming-assisted color changes, a perforated top metal film is essential because the opening at the top metal layers allows gas to penetrate across the top metal layer. Figure 5c represents that the optical foggy effect (upper line; Fig. 5c) and whiteout effect (bottom line; Fig. 5c) were still available in the f-etalon having10-nm-thick Ag top metal, compared with the same color-changing behavior at the f-etalon with 10-nm-thick top Pd. For instance, the reference color of the f-etalon (red square of yellowish flower petals in Fig. 5c) gradually faded and then shrouded by a pale gray color at 600 s (foggy effect). Meanwhile, the yellow color of the identical zone turned white within 240 s (whiteout effect) after O2 injection (Supplementary Video 4).
Structural color intrinsically produces angle-dependent color variation due to its increased optical path length in the resonating dielectric film. The etalon color varied with increasing viewing angle (from 0° to 75°), resembling the effect of thickening the optical cavity. Because the diffuse reflection by water bubbles is angle-independent, the foggy effect can occur at any viewing angle (Fig. 5d), making the proposed etalon structure applicable to gas leakage warning smart windows. As schematically demonstrated in Fig. 5e, the f-etalons installed in a hydrogen gas storage facility can instinctively recognize the leakage of H2 gas by dimming the original colors of the flower patterns. This scenario was demonstrated using an actual miniature model by placing four f-etalon samples inside a miniature gas tank storage room (Fig. 5f and Supplementary Video 5).