Scalable In2S3 based optical memristor devices as artificial synapse for logic realization and neuromorphic computing

• Published in: Materials science in semiconductor processing 2025-01, Vol.185, p.108985, Article 108985
• Main Authors: B, Sharmila, Divyashree, P, Sharma, Sumit, Dwivedi, Priyanka, Das, Samaresh
• Format: Article
• Language: English
• Subjects: In2S3; Neuromorphic computing; Optical memristor; Process scalability
• ISSN: 1369-8001
• DOI: 10.1016/j.mssp.2024.108985

This paper presents the scalable production of indium sulfide (In2S3) based optical memristors as an artificial synapse for logic realization and neuromorphic computing applications. The optical memristor device's performance was tested using the electrical and optical stimulation (ultraviolet (UV) to near infrared (NIR)) at room temperature. The memristor has shown the improved performance metrics for optical stimulation as compared to electrical stimulation. The In2S3 based optical memristor offers a resistance switching ratio of ∼103 at room temperature. In addition, the memristor has ultra-fast program and reprogram capability (38.8/39 μs) at 880 nm. Moreover, stability/scalability of the device was tested using the cycle-to-cycle variability test at different modulated frequencies. The electrical and optical signals were used as an input for logic gate realization. The neuromorphic computing model was implemented to predict the digits and alphabets with an accuracy of 98.19 % and 99.18 % respectively. These demonstrated optical memristors can be the key technology for future memory devices that surpass the limitations of von-Neumann bottleneck.