Wafer scale WS2 based ultrafast photosensing and memory computing devices for neuromorphic computing

Abstract Integration of optical sensors with a memristors can establish the bridge between photosensing devices and memory devices for Internet of Things (IoT) applications. This paper presents the realization of integrated sensing and computing memory (ISCM) devices using tungsten disulfide (WS2) and their application for neuromorphic computing. The ISCM device fabrication process is scalable as microfabrication steps followed on 2” wafer, ISCM device testing and image classification for neuromorphic computing. The photosensing/memory tests were conducted using electrical and optical stimulations (broadband spectrum). The fabricated photosensing device offers a higher responsivity (8 A/W), higher detectivity (2.85✕1011 Jones) and fast response speed (80.2/78.3 μs) at 950 nm. The memory device has shown a set/reset time of 51.6/73.5 μs respectively. Further, the repeatability, stability and reproducibility tests were conducted by stimulating the device with different modulating frequencies. The frequency modulation tests confirm that the ISCM devices are stable and perfect candidate for real-time IoT applications. Moreover, the device’s potentiation and depression results were used for image classification with the accuracy of 98.27%. These demonstrated device’s test results provide possibilities to fabricate the smart sensors with integrated functionalities.