Metasurfaces are composed of a two-dimensional array of carefully engineered subwavelength structures.

A standardized hybrid deep-learning model based on a combination of a deep convolutional network and a recurrent neural network is proposed to predict the optical response of metasurfaces considering their shape and all the important dimensional parameters (such as periodicity, height, width, and aspect ratio) simultaneously. It is further used to aid the design procedure of the key components of solar thermophotovoltaics (STPVs), i.e., metasurface based perfect solar absorbers and selective emitters. Although these planar meta-absorbers and meta-emitters offer an ideal platform to realize compact and efficient STPV systems, a conventional procedure to design these is time taking, laborious, and computationally exhaustive. The optimization of such planar devices needs hundreds of EM simulations, where each simulation requires multiple iterations to solve Maxwell's equations on a case-by-case basis. To overcome these challenges, we propose a unique deep learning-based model that generates the most likely optical response by taking images of the unit cells as input. The proposed model uses a deep residual convolutional network to extract the features from the images followed by a gated recurrent unit to infer the desired optical response. Two datasets having considerable variance are collected to train the proposed network by simulating randomly shaped nanostructures in CST microwave studio with periodic boundary conditions over the desired wavelength ranges. These simulations yield the optical absorption/emission response as the target labels. The proposed hybrid configuration and transfer learning provide a generalized model to infer the absorption/emission spectrum of solar absorbers/emitters within a fraction of seconds with high accuracy, regardless of its shape and dimensions. This accuracy is defined by the regression metric mean square error (MSE), where the minimum MSE achieved for absorbers and emitters test datasets are 7.3 × 10−04 and 6.2 × 10−04 respectively. The trained model can also be fine-tuned to predict the absorption response of different thin film refractory materials. To enhance the diversity of the model. Thus it aids metasurface design procedure by replacing the conventional time-consuming and computationally exhaustive numerical simulations and electromagnetic (EM) software. The comparison of the average simulation time (for 10 samples) and the average DL model prediction time shows that the DL model works about 98% faster than the conventional simulations. We believe that the proposed methodology will open new research directions towards more challenging optimization problems in the field of electromagnetic metasurfaces.

Flat optics have become capable of achieving unprecedented functionalities through electromagnetic (EM) wave manipulation by employing the metasurfaces. The most crucial part in the design of metasurface is the selection the constitutive component i.e. the meta-atom's material and structure so that it exhibits the precise operation as per the desired application. The unit-cell design calls for an iterative loop of simulations in order to explore the EM responses for intended operation. In this work, we have studied the absorption response of refractory materials under visible light radiations for their utilization in energy harvesting applications. The absorption response estimation using machine-learning techniques for the materials having very high melting-points, mechanical stabilities and inertness to the atmosphere has been carried out to investigate their performance in the broadband range. The presented regression models incorporate hybrid data format i.e. they simultaneously contend with 3-D and 1-D properties of various shapes of nano-resonators. The images' feature extraction is carried out by employing Singular Value Decomposition. The trained models are potent enough to bypass the repetitive sequence of optimization involved in conventional EM solvers. Additionally, the models are capable of predicting the optimum shape along with structural dimensions of unit-cell. For forward model, the MSEs for training and testing are 1.302×10 -2 and 3.269×10 -2 while R 2 scores are 0.9804 and 0.8764, respectively. The approach applied is so robust that, irrespective of complexity of unit-cell structure is, it serves the purpose of predicting the distinct structure with highest performance while bypassing the time-and computationally-intensive EM simulations.

No abstract