Molecular Engineering of Interlayer Spacings in 2D Dion-Jacobson Perovskites for High-Fidelity Neuromorphic Computing

The development of memristive devices is important for energy-efficient neuromorphic computing. While two-dimensional (2D) organic-inorganic hybrid perovskites offer structural tunability and improved stability, the correlation between their molecular-level interlayer architecture and resistive switching (RS) kinetics remains to be fully elucidated. Herein, we modulate the inorganic layer spacing in Dion-Jacobson (DJ) phase 2D perovskites, BDAPbI4, HDAPbI4, and ODAPbI4, by tailoring the alkyl chain length of diammonium ligands. The expanded interlayer environment regulates ion migration, resulting in improved RS behavior in HDAPbI4 and ODAPbI4 devices, with an ON/OFF ratio exceeding 103 and cycling endurance over 100 cycles. These devices emulate bioinspired synaptic functions, including the transition from short-term to long-term plasticity. Cross-Sim simulations of a three-layer neural network using the device conductance characteristics achieved a recognition accuracy of 96.6% for the MNIST data set, close to the software-defined benchmark of 98.2%. Overall, these results identify interlayer spacing as an important structural parameter that regulates ion migration, resistive switching behavior, and analog conductance modulation in 2D Dion-Jacobson perovskite memristors.