Login / Signup

Picoampere Dark Current and Electro-Opto-Coupled Sub-to-Super-linear Response from Mott-Transition Enabled Infrared Photodetector for Near-Sensor Vision Processing.

Mohit KumarSeokwon LimJisu KimHyungtak Seo
Published in: Advanced materials (Deerfield Beach, Fla.) (2023)
Light-intensity selective superlinear photodetectors with ultralow dark current can provide an essential breakthrough for the development of high-performing near-sensor vision processing. However, the development of near-sensor vision processing is not only conceptually important for device operation (given that sensors naturally exhibit linear/sublinear responses), but also essential to get rid of the massive amount of data generated during object sensing and classification with noisy inputs. Therefore, achieving the giant superlinear photoresponse while maintaining the picoampere leakage current, irrespective of the measurement bias, is one of the most challenging tasks. Here, Mott material (vanadium dioxide) and silicon-based integrated infrared photodetectors are developed that show giant superlinear photoresponse (exponent >18) and ultralow dark current of 4.46 pA. Specifically, the device demonstrates an electro-opto-coupled insulator-to-metal transition, which leads to outstanding photocurrent on/off ratio (>10 6 ), a high responsivity (>1 mA W -1 ), and excellent detectivity (>10 12  Jones), while maintaining response speed (τ r  = 6 µs and τ f  = 10 µs). Further, intensity-selective near-sensor processing is demonstrated and night vision pattern reorganization even with noisy inputs is exhibited. This research will pave the way for the creation of high-performance photodetectors with potential uses, such as in night vision, pattern recognition, and neuromorphic processing.
Keyphrases
  • working memory
  • machine learning
  • high intensity
  • deep learning
  • big data
  • high speed
  • climate change
  • rare case
  • artificial intelligence
  • low cost