Al-Stoichiometry-Engineered Defect Profiles in InAlZnO Nanofibers for Balanced Charge Transport and Optical...

Metal oxide optoelectronic synapses are promising for neuromorphic vision, but combining a low dark current with sustained photoinduced plasticity remains challenging. Here, Al stoichiometry is engineered in electrospun In1AlxZn0.08O (x = 0.007, 0.011, and 0.015) nanofiber networks to tune sub-bandgap defects, carrier transport, and photocarrier relaxation. Optical, spectroscopic, electrical, and synaptic measurements identify 1.1 atomic % Al as optimal. Strong Al-O bonding suppresses excess oxygen-vacancy-related donor states while preserving the In 5s transport network. The optimized device shows an off-state current of 10-11 A, an on/off ratio of 107, a field-effect mobility of 14 cm2 V-1 s-1, and a paired-pulse facilitation index of 271%. Under pulsed UV light, defect-assisted trapping and slow relaxation retain 30% of the peak photocurrent after 2 h. Neural network simulations achieve 95.7% handwritten digit recognition accuracy, highlighting the value of cation stoichiometry control for low-power optoelectronic neuromorphic computing.