The work, published in Light: Science & Applications, was led by Roy Maman, a PhD student in the research group of Professor Uriel Levy, working alongside Professor Ilya Goykhman and nanofabrication specialist Noa Mazurski. Their platform is a dual-band metasurface, an engineered array of subwavelength structures that controls how thermal radiation leaves a surface. The central achievement is spectral decoupling. Conventional thermal photonics struggle to manipulate the mid-wave and long-wave bands independently at the pixel level, because the physics that governs emission in one range tends to drag the other along with it. The Jerusalem team broke that dependency using a carefully designed metal-insulator-metal architecture in which each unit cell behaves like a pair of coupled resonators, one tuned to each atmospheric window.

Each pixel of the metasurface is built from a stack in which patterned nanodisks sit above a thin dielectric spacer and a metallic backplane. This metal-insulator-metal configuration supports resonant modes that absorb and re-emit thermal radiation at selectable wavelengths. By adjusting geometric parameters such as the radius of the nanodisks and the periodicity of the array, the designers can move the resonance peak of each band and reshape its spectral width. Kirchhoff’s law of thermal radiation ties a material’s emissivity to its absorptivity, so sculpting the absorption spectrum directly sculpts the thermal emission spectrum. A pixel engineered to emit strongly in the mid-wave band but weakly in the long-wave band will glow brightly to a mid-wave camera while appearing dark to a long-wave camera, and the reverse configuration is equally available.

Getting two independent channels out of one physical structure is a delicate optimization problem, because the geometric degrees of freedom are shared between the resonances. The researchers approached it with a mathematical stability analysis rooted in control theory, mapping how each geometric parameter influences each resonance property and selecting operating points where the cross-sensitivities vanish. The team summarized the principle behind the design: using a mathematical stability analysis rooted in control theory, they leveraged fundamental geometric degrees of freedom to independently tune optical resonances across both wavelength bands simultaneously, decoupling control over key resonance parameters such as peak position and spectral width and achieving near-zero correlation between the two distinct infrared channels. In practical terms, a grayscale value written into a pixel for the mid-wave image has no measurable effect on the grayscale value the same pixel presents in the long-wave image.

The demonstration that turns this physics into something visually striking is a two-portrait encryption experiment on a single chip. The researchers encoded two completely different detailed images into the exact same pixel area. Viewed through a mid-wave infrared camera, the chip displays a grayscale portrait of Sir Isaac Newton. Viewed through a long-wave infrared camera, the identical chip reveals the face of Gottfried Wilhelm Leibniz, Newton’s historical rival in the calculus priority dispute. The choice of subjects is a fitting metaphor for the technology itself: one surface, two irreconcilable truths, and the answer depends entirely on which window the observer chooses to look through.

Encryption is only one of two headline applications. In an alternate configuration, the same platform can be programmed for thermal camouflage, projecting identical signatures in both bands so that an object blends seamlessly into an engineered background. This matters because dual-band sensors are specifically deployed to defeat single-band spoofing; a hot spot that looks like the ambient scene in one window but betrays itself in the other is a classic failure mode of conventional camouflage. By matching emissivity patterns across both atmospheric windows simultaneously, the metasurface addresses the multi-band detection strategy at its own level rather than tricking one sensor at a time.

Robustness is where the platform separates itself from laboratory curiosities. The encoded images remain clear and stable across wide viewing angles and over a broad temperature range, which means the hidden information does not degrade when the chip is heated or when the camera moves off-axis. In the reported measurements, thermal images were captured at chip temperatures ranging from 32 degrees Celsius to 112 degrees Celsius, and at observation angles from zero degrees up to 44 degrees. As the chip temperature rises, the encoded patterns grow increasingly dominant over the background thermal emission, sharpening the images rather than washing them out. The long-wave image in particular stays remarkably stable at every measured angle and temperature, while the mid-wave image remains distinguishable and clearly distinct from its long-wave counterpart, confirming that high-frequency visual data never leaks between the two channels.

The physical mechanism behind this stability lies in the resonant nature of the emission control. Because the emissivity is set by geometry rather than by temperature-dependent material properties, the spectral fingerprints of the pixels persist as the chip heats and cools. Angular stability follows from the subwavelength scale of the unit cells, which suppresses the diffraction effects that would otherwise scatter or shift the response as the viewing angle changes. This combination of thermal and angular tolerance is what makes the platform a candidate for real-world deployment rather than a fixed-conditions demonstration, since field sensors rarely enjoy the luxury of a fixed camera position or a constant target temperature.

Manufacturing considerations strengthen the case further. The design is fully compatible with standard photolithography, the workhorse patterning process of the semiconductor industry, so the metasurfaces can be scaled to large areas at low unit cost without exotic fabrication steps. That compatibility matters for the security applications the researchers envision, where anti-counterfeiting tags and encryption elements are only useful if they can be produced cheaply and in volume. The team forecasts that the platform could redefine next-generation anti-counterfeiting tags, secure optical data encryption, and multi-sensor military stealth, positioning the chip at the intersection of commercial authentication technology and defense-grade signature management.

The researchers also point toward a dynamic future for the technology. Incorporating active materials such as graphene, electro-optic layers, or phase-change alloys could eventually unlock real-time, reconfigurable dual-band thermal animations, effectively turning static camouflage into dynamic, adaptable thermal displays. In such a scheme, the two portraits hidden in a single surface would no longer be fixed at fabrication time but could be rewritten electrically or thermally, allowing a camouflaged object to update its false signature on demand or an encryption element to cycle through authentication states. For now, the passive version already delivers something previously out of reach: two uncorrelated thermal images coexisting on one chip, each invisible to the camera assigned to the other, built with the same fabrication tools that already produce the world’s microelectronics.

Subject of Research: Dual-band metasurfaces for independent emissivity control in mid-wave and long-wave infrared imaging

Article Title: Multi-channel metasurface reshapes thermal camouflage and security encryption

Article References: Multi-channel metasurface reshapes thermal camouflage and security encryption. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: metasurface, thermal camouflage, infrared imaging, MWIR, LWIR, emissivity control, optical encryption, anti-counterfeiting, photolithography, thermal photonics, nanofabrication, Light Science and Applications

The central obstacle is thermodynamic. Gold’s surface free energy is roughly 1.12 joules per square metre, an order of magnitude higher than that of graphene under ambient conditions. When gold atoms are deposited onto ordinary substrates such as silicon dioxide, this high surface energy drives them to clump together into three-dimensional islands rather than spreading out into a flat layer, a behaviour known as Volmer-Weber growth. The result is a discontinuous scatter of metallic droplets rather than the uniform, layer-by-layer Frank-van der Merwe film needed for a usable ultrathin conductor. Previous attempts to force flat growth relied on seed layers or adhesion promoters that permanently bonded the gold to its growth substrate, destroying the transferability that device integration demands, or produced films whose roughness was inherited from an imperfect underlying surface.

The team’s solution begins with the growth template itself. Instead of evaporating copper onto a support, where the columnar grain structure of the deposited metal imposes its own roughness, the researchers created a freestanding copper foil by template stripping. A 500-nanometre copper layer was first evaporated onto a single-crystal silicon (111) wafer, electroplated to a self-supporting thickness of 30 to 60 micrometres, and then peeled away. The gold-facing surface of the foil inherits the atomic flatness of the silicon wafer beneath it. Atomic force microscopy revealed terraced, essentially featureless copper with a root-mean-square roughness of just 0.24 nanometres over a four-micrometre scan, dropping to 0.16 nanometres at higher magnification. Because the deposited gold faithfully replicates the copper relief, this suppression of template corrugation proved to be the decisive factor enabling continuous gold films down to 1.5 nanometres, compared with a previous lower bound of 3.5 nanometres.

The choice of copper is not merely a matter of smoothness but of energetics. The adsorption energy of gold atoms on copper, approximately 7.78 joules per square metre, comfortably exceeds twice the surface free energy of gold, satisfying the thermodynamic criterion for two-dimensional, layer-by-layer growth. On silicon dioxide, by contrast, the adsorption energy is far too low, and islanding is inevitable. Molecular dynamics simulations made this contrast vivid: gold atoms deposited onto the copper template self-assembled into a uniform, highly crystalline monolayer even under deliberately anisotropic deposition conditions, while the same atoms landing on silicon coalesced into disordered, discrete islands. High-resolution electron microscopy of three-nanometre films confirmed the picture experimentally, showing a continuous, featureless surface on copper-derived films against percolated networks on molybdenum disulfide and classic island morphologies on bare oxide.

Direct evidence for layer-by-layer growth came from a strikingly simple test. The researchers deposited a nominal one-nanometre gold layer, a regime where island formation is most pronounced on conventional substrates, and re-examined the surface. The roughness barely changed, rising only from 0.16 to 0.18 nanometres, with no discrete islands visible. After transfer to silicon dioxide substrates, films ranging from two to eight nanometres in thickness remained continuous and uniform, with roughness below 0.4 nanometres and only rare pinholes in the thinnest sample. Transmission electron microscopy of a suspended three-nanometre membrane revealed a polycrystalline structure with crystalline domains often exceeding 20 nanometres laterally, several times larger than the film itself, and lattice fringes spaced about 0.24 nanometres apart, the signature of close-packed gold (111) planes.

The optical and electrical performance places these films among the best ultrathin transparent conductors yet reported. Peak transmittance climbed monotonically as thickness decreased, reaching 86.6 percent for the 1.5-nanometre film, up from 80 percent in the team’s earlier work, while sheet resistance for the thicker films fell to 14.9 ohms per square at five nanometres and 7.8 ohms per square at eight nanometres, approaching the theoretical limits for bulk crystalline gold. Spectroscopic ellipsometry showed that the dielectric function of films from three to eight nanometres closely tracks that of bulk gold, with a plasma frequency within six percent of the bulk value, quantitative evidence of a near-bulk free-electron density and minimal defects. Benchmarked against previous ultrathin gold, multilayer chemical-vapour-deposition graphene and commercial indium tin oxide, the films occupy a leading position on the standard transparency-conductivity figure of merit.

Electrical transport in the thinnest films is governed by the physics of reduced dimensionality. Because the film thickness is far smaller than the roughly 40-nanometre electron mean free path in bulk gold, surface and grain-boundary scattering dominate the resistivity. A combined Fuchs-Sondheimer and Mayadas-Shatzkes model, with physically reasonable parameters for specular surface scattering and grain-boundary reflection, matched the measured data across the two-to-eight-nanometre range. Flexibility testing showed that a five-nanometre film on plastic survived 1,000 bending cycles at a radius of about three millimetres with only a minimal increase in sheet resistance, a resilience that underpins the device demonstrations that followed.

Those demonstrations span photonics, thermal management and biomedicine. An eight-nanometre gold film served as the transparent anode in solution-processed organic light-emitting diodes, which outperformed identical devices built on commercial indium tin oxide, likely because gold presents a lower energy barrier of 0.1 electron-volts to the hole-injection layer, and because the atomically smooth anode eliminates the protrusions that can short-circuit the roughly 100-nanometre organic stack. The team also fabricated an eight-by-eight matrix-addressable transparent OLED display on flexible plastic at six-inch scale. For thermal camouflage, the same films exploit a dual property: high visible transparency preserves the object’s appearance, while high infrared reflectivity suppresses thermal emission, in accordance with Kirchhoff’s law. A five-nanometre gold coating reduced the apparent temperature seen by an infrared camera to about a third of the actual background temperature, a roughly 30 percent improvement over the team’s earlier rougher films, and a gold-polymer laminate concealed a satellite-shaped object from thermal imaging while remaining visually transparent.

Perhaps the most striking application is on human skin. Transferred as electronic tattoos, five- and twenty-nanometre gold electrodes conformed to the forearm and maintained stable impedance across repeated measurements, with values at 10 kilohertz of 20.3 and 10.5 kilohms respectively, comparable to state-of-the-art graphene electronic tattoos. When used to record electrocardiograms and electromyograms, the gold tattoos achieved signal-to-noise ratios of 14.9 and 69.5 decibels, exceeding the 12.7 and 62.4 decibels obtained from clinical silver-silver-chloride gel electrodes. The researchers attribute this advantage to the exceptional mechanical conformability of the ultrathin films, which yields intimate skin contact, minimises motion artefacts and outweighs the electrodes’ higher intrinsic impedance. Looking forward, the team notes that the same template-stripping and transfer paradigm should extend to other quasi-two-dimensional metals such as silver, palladium and platinum, while single-crystal copper templates and alternative stabilising interfaces based on sulfur-containing materials may push continuous metallic films toward the one-nanometre threshold, opening a broad new library of atomically thin metals for next-generation technologies.

Subject of Research: Wafer-scale synthesis and transfer of atomically thin continuous gold films for optoelectronics, thermal management and bio-interfaces

Article Title: Wafer‐Scale Atomically‐Thin Gold: Transferable Platform for Flexible Optoelectronics, Thermal Management and Bio‐Interfaces

Article References: Yakubovsky, D., Mironov, M., Ermolaev, G., Khrebtov, A., Slavich, A., Narute, P., Chowdhury, P., Kirtaev, R., Khramtsov, I., Minnekhanov, A., Tselikov, G., Grudinin, D., Sidorova, M., Pak, N., Maslov, P., Vyshnevyy, A., Zhou, H., Kravets, V. G., Grigorenko, A. N., … Volkov, V. (2026). Wafer‐Scale Atomically‐Thin Gold: Transferable Platform for Flexible Optoelectronics, Thermal Management and Bio‐Interfaces. Advanced Science, 13(55), Article e76495. https://doi.org/10.1002/advs.76495

Image Credits: AI Generated

DOI: 10.1002/advs.76495

Keywords: ultrathin gold, 2D materials, template stripping, transparent electrodes, flexible electronics, OLEDs, thermal camouflage, electronic tattoos, epidermal electronics, Frank-van der Merwe growth, sheet resistance, wafer-scale fabrication