RESUMEN
All-optical diffractive neural networks, as analog artificial intelligence accelerators, leverage parallelism and analog computation for complex data processing. However, their low space transmission efficiency or large spatial dimensions hinder miniaturization and broader application. Here, we propose a terahertz spoof plasmonic neural network on a planar diffractive platform for direct multi-target recognition. Our approach employs a spoof surface plasmon polariton coupler array to construct a diffractive network layer, resulting in a compact, efficient, and easily integrable architecture. We designed three schemes: basis vector classification, multi-user recognition, and MNIST handwritten digit classification. Experimental results reveal that the terahertz spoof plasmonic neural network successfully classifies basis vectors, recognizes multi-user orientation information, and directly processes handwritten digits using a designed input framework comprising a metal grating array, transmitters, and receivers. This work broadens the application of terahertz plasmonic metamaterials, paving the way for terahertz on-chip integration, intelligent communication, and advanced computing systems.
RESUMEN
Coherent perfect absorption (CPA) and amplification of electromagnetic waves are converse phenomena, where incoming radiations are coherently dissipated or amplified by structured incidences. Realizing such two phenomena simultaneously in a single device may benefit various applications such as biological sensing, photo detection, radar stealth, solar-thermal energy sharing, and wireless communications. However, previous experimental realizations of CPA and amplification generally require precise controls to the loss and gain of a system, making dynamic switching between the absorption and amplification states a challenge. To this end, we propose a nonlinear approach to realize CPA and parametric amplification (PA) simultaneously at the same frequency and demonstrate experimentally dynamic switch from the CPA to PA states in a judiciously designed nonlinear spoof plasmonic waveguide. The measured output signal gain can be continuously tuned from -33 dB to 22 dB in a propagation length of 9.2 wavelengths. Compared to the traditional linear CPA, our approach relaxes the stringent requirements on device dimensions and material losses, opening a new route to actively modulate the electromagnetic waves with giant amplification-to-absorption contrast in a compact platform. The proposed nonlinear plasmonic platform has potential applications in on-chip systems and wireless communications.
