In-Sensor Computing Realization Using Fully CMOS-Compatible TiN/HfOx based Neuristor Array

In-Sensor Computing Realization Using Fully CMOS-Compatible TiN/HfOx based Neuristor Array

ACS Sensors

10.1021/acssensors.3c01418

http://dx.doi.org/10.1021/acssensors.3c01418

HAIZHONG ZHANG, Peng Qiu, Yaoping Lu, Xin Ju, DongZhi Chi, Kwang Sing Yew, Minmin Zhu, Shaohao Wang, Rongshan Wei, Wei Hu

Bioengineering, Instrumentation, Fluid Flow and Transfer Processes, Process Chemistry and Technology

14 September 2023

Zhang, H., Qiu, P., Lu, Y., Ju, X., Chi, D., Yew, K. S., Zhu, M., Wang, S., Wei, R., & Hu, W. (2023). In-Sensor Computing Realization Using Fully CMOS-Compatible TiN/HfOx-Based Neuristor Array. ACS Sensors. https://doi.org/10.1021/acssensors.3c01418

With the evolution of artificial intelligence, the explosive growth of data from sensory terminals gives rise to severe energy-efficiency bottleneck issues due to cumbersome data interactions among sensory, memory, and computing modules. Heterogeneous integration methods such as chiplet technology can significantly reduce unnecessary data movement; however, they fail to address the fundamental issue of the substantial time and energy overheads resulting from the physical separation of computing and sensory components. Brain-inspired in-sensor neuromorphic computing (ISNC) has plenty of room for such data-intensive applications. However, one key obstacle in developing ISNC systems is the lack of compatibility between material systems and manufacturing processes deployed in sensors and computing units. This study successfully addresses this challenge by implementing fully CMOS-compatible TiN/HfOx-based neuristor array. The developed ISNC system demonstrates several advantageous features, including multilevel analogue modulation, minimal dispersion, and no significant degradation in conductance (@125 °C). These characteristics enable stable and reproducible neuromorphic computing. Additionally, the device exhibits modulatable sensory and multi-store memory processes. Furthermore, the system achieves information recognition with a high accuracy rate of 93%, along with frequency selectivity and notable activity-dependent plasticity. This work provides a promising route to affordable and highly efficient sensory neuromorphic systems.

Publisher Copyright

This research / project is supported by the National Research Foundation - Competitive Research Program
Grant Reference no. : NRF-CRP24-2020-0002

This work is supported by Science and Technology Major Project of Fujian Province, China (Grant No. 2022HZ027006) and Quanzhou Municipal Science and Technology Major Project, China (Grant No. 2022GZ7), National Natural Science Foundation of China (Grant No. 62274036),

This document is the Accepted Manuscript version of a Published Work that appeared in final form in ACS Sensors, copyright © American Chemical Society after peer review and technical editing by the publisher. To access the final edited and published work see doi.org/10.1021/acssensors.3c01418

https://oar.a-star.edu.sg/communities-collections/articles/19324

2379-3694

Institute of Materials Research and Engineering