# Defect-Driven Neuromorphic Plasticity in Planar ZnO Optoelectronic Synapses

## TL;DR

This paper explores how defects in ZnO affect the performance of optoelectronic synapses, enabling better artificial vision hardware.

## Contribution

A cross-time scale design strategy linking atomic defects to neuromorphic performance in ZnO synapses is introduced.

## Key findings

- Long-lived defect states in ZnO slow PPC decay and regulate synaptic plasticity.
- Optimized ZnO synapses achieve 90.8% accuracy in handwritten digit recognition.
- Synaptic weight can be predictively tuned by controlling defect kinetics.

## Abstract

Understanding how
atomic-scale defect dynamics influence system-level
neuromorphic behavior is crucial for the rational design of oxide-based
optoelectronic synapses. In this study, a planar ZnO synapse has been
introduced where the nanosecond-scale oxygen-vacancy carrier lifetime
is directly linked to second-scale persistent photoconductivity (PPC)
decay and key synaptic plasticity parameters. By combining steady-state
and time-resolved spectroscopies with electrical measurements, a dynamic
framework that spans multiple time scales has been developed: long-lived
defect states slow PPC decay, which in turn regulates paired-pulse
facilitation retention and the efficiency of short-to-long-term plasticity
transitions. This framework allows for predictive tuning of the synaptic
weight by controlling defect occupation and release kinetics. The
optimized ZnO synapse operating at 0.1 V demonstrates robust long-term
potentiation and achieves 90.8% recognition accuracy in handwritten
digit recognition. This work presents a cross-time scale design strategy
that bridges atomic-level defect engineering with neuromorphic system
performance, paving a route toward artificial vision hardware.