Metasurfaces are ultrathin and flat layers of subwavelength nanostructures composed of metallic or high‐refractive‐index materials. They can alter lightwave properties effectively and show significant application potential in various nanophotonic technologies. The subwavelength meta‐atoms are generally carved by electron beam lithography or focused ion beam. It is challenging to produce large‐scale metasurface devices at low cost. Herein, the fabrication of low‐cost and large‐area plasmonic and dielectric metasurfaces through a combination of soft nanoimprint lithography and reactive ion etching is reported and the dimension of meta‐atoms by controlling the etching condition carefully and implementing an iterative imprint and etch process is tuned. Such an approach is effective to alter the metasurface resonances and reproduce new structures with minimum cost for wafer‐scale nanophotonics.

The merge between nanophotonics and a deep neural network has shown unprecedented capability of efficient forward modeling and accurate inverse design if an appropriate network architecture and training method are selected. Commonly, an iterative neural network and a tandem neural network can both be used in the inverse design process, where the latter is well known for tackling the nonuniqueness problem at the expense of more complex architecture. However, we are curious to compare these two networks’ performance when they are both applicable. Here, we successfully trained both networks to inverse design the far-field spectrum of plasmonic nanoantenna, and the results provide some guidelines for choosing an appropriate, sufficiently accurate, and efficient neural network architecture.

The lateral geometry and material property of plasmonic nanostructures are critical parameters for tailoring their optical resonance for sensing applications. While lateral geometry can be easily observed by a scanning electron microscope or an atomic force microscope, characterizing materials properties of plasmonic devices is not straightforward and requires delicate examination of material composition, cross-sectional thickness, and refractive index. In this study, a deep neural network is adopted to characterize these parameters of unknown plasmonic nanostructures through simple transmission spectra. The network architecture is established based on simulated data to achieve accurate identification of both geometric and material parameters. We then demonstrate that the network training by a mixture of simulated and experimental data can result in correct material property recognition. Our work may indicate a simple and intelligent characterization approach to plasmonic nanostructures by spectroscopic techniques.

Dissolvable and transient devices are important for environment-friendly disposal and information security. Similar to transient electronic devices, photonic devices use dissolvable metals such as magnesium and zinc to enable tunable plasmonic resonances. However, functional nanostructured substrates made of a common photoresist and a soft substrate are not dissolvable. In this study, we report the large-area, dissolvable polylactic-co-glycolic acid nanostructures formed by nanoimprint lithography and discuss the impact of the imprinting temperature and ambient conditions on the formed nanostructures. The deposition of a thin layer of metal can yield a quasi-3D plasmonic device, and the choice of zinc metal can result in an all-dissolvable device in water over a few days. Consequently, the transmission spectra of these plasmonic devices could be tuned after placement in water. This strategy yields a truly transient nanophotonic device that can be degraded after performing its function for a specific period.