This article introduces an innovative methodology for the design of energy selective surfaces (ESS) using machine learning. The methodology differs from traditional ESS design approaches that rely on equivalent circuit models. Instead, it involves the training of a forward prediction neural network for the electromagnetic performance of ESS and utilizing an optimization algorithm to achieve the optimal ESS structure. The optimization objective is to align the transmission pole in the transmitted state and transmission zero in the protection state with the target frequency point, achieving a minimal insertion loss (IL) and a maximal shielding efficiency (SE) at the same time. The designed ESS structure consists of a metal frame, a nonlinear induction region, and a resonant region formed by a randomly distributed 0/1 matrix, whose electromagnetic performance varies with the distribution of the 0/1 matrix. Through multiple simulations, 6,000 sets of training data are obtained and used to train a fully connected neural network. Subsequently, the particle swarm optimization algorithm is employed to optimize the objective function. Following these procedures, several ESS structures are successfully obtained. After prototype fabrication and testing, all samples demonstrate consistency with expected results by achieving a low IL and a high SE near the target frequencies.

In this paper, we design a multifunctional micro-nano device with a hybrid metamaterial-waveguide system, which leads to a triple plasmon-induced transparency (PIT). The formation mechanisms of the three transparent peaks have their own unique characteristics. First, PIT-I can be switched into the BIC (Friedrich–Wintge bound state in continuum), and the quality factors (Q-factors) of the transparency window of PIT-I are increased during the process. Second, PIT-II comes from near-field coupling between two bright modes. Third, PIT-III is generated by the near-field coupling between a low-Q broadband bright mode and a high-Q narrowband guide mode, which also has a high-Q transparent window due to the guide mode. The triple-PIT described above can be dynamically tuned by the gate voltage of the graphene, particularly for the dynamic tuning of the Q values of PIT-I and PIT-III. Based on the high Q value of the transparent window, our proposed structure can be used for highly sensitive refractive index sensors or devices with prominent slow light effects.