An ultra-compact and highly stable optical numerical comparator based on Y-shaped graphene nanoribbons

Aijun Zhu,Lei Song,Lei Cheng,Cong Hu,Rabi Mahapatra
DOI: https://doi.org/10.1140/epjd/s10053-023-00748-9
2023-09-07
Abstract:Research on artificial neural network computing based on conventional integrated circuit chips has made significant progress, but it faces technical bottlenecks such as reduced energy consumption, computing speed, and efficiency. So, it is seeking integrated chips based on optical interconnections to solve the current dilemma. Fortunately, the small size and very localized graphene surface plasmon waves offer the possibility of optical integrated chips. In this paper, we propose a graphene-based one-bit optical numerical comparator. The comparator is located on the top of a 0.4-μm 2 rectangular dielectric layer, mainly consisting of Y-shaped graphene nanoribbons. The on/off effect of the graphene nanoribbons is achieved by applying an external voltage to change the chemical potential energy of the graphene switching bands. The proposed optical numerical comparator with 9.55-μm TM mode light achieves a minimum extinction ratio of 31.12 dB and amplitude modulation of 0.77 dB, as shown by the finite-difference time-domain (FDTD) method. Compared with the current optical numerical comparators, it has the advantages of a high extinction ratio, small size, low loss, and high stability. In addition, the effect of process deviation of graphene nanoribbons on the reliability of the designed optical numerical comparator is analyzed by simulation. It is beneficial to developing integrated photonic devices and has some significance for developing ultra-high-frequency and integrating artificial neural network computing.
optics,physics, atomic, molecular & chemical
What problem does this paper attempt to address?