Abstract
1 min readHalide perovskites (HPs) are promising memristive materials for neuromorphic hardware because of their defect-tolerant electronic structure, low-temperature processability, and mixed ionic–electronic conductivity (MIEC). Nevertheless, the chemical/ionic instabilities of 3D HPs, moisture sensitivity, mobile halide vacancies, and grain-boundary percolation complicate reliable operation. Low-dimensional perovskites mitigate these limitations by incorporating hydrophobic organic layers and increasing formation energies while retaining switchable ionic pathways. Within the quasi-2D family, alternating cation in the interlayer (ACI) perovskites, described by A′A n − 1 B n X 3n+1 , shorten interlayer spacing and provide H-bond-assisted lattice stabilization, thereby enabling more efficient out-of-plane transport than Ruddlesden–Popper (RP) phases. Here, we fabricate a memristor using (GA)(MA) 3 Pb 3 I 10 perovskite as the active layer, deposited on ITO-coated glass via one-step spin-coating, with Au as the top electrode in a metal–insulator–metal (MIM) stack. A volatile NH 4 Cl additive is introduced to steer crystallization; chloride-mediated intermediates drive vertical alignment of the ACI slabs and yield compact, columnar grains with suppressed lateral percolation, while transient Cl - largely vacates during annealing, leaving an iodide-rich lattice with passivated interfaces. The devices exhibit bipolar, nonvolatile resistive switching with an on/off ratio up to 10 3 and, under continuous analog sweeps, display reproducible potentiation and depression with gradual conductance updates across multiple intermediate states. The electrical responses are consistent with iodide-vacancy (V I )–mediated filamentary switching preferentially along out-of-plane [PbI 6 ] 4 − columns; orientation engineering confines ion drift to the thickness direction and suppresses stochastic lateral pathways, improving the linearity and symmetry of analog weight updates. This study suggests a materials route to stable analog memristors by combining vertical texture with vacancy-controlled filaments and indicates a device platform readily integrated into dense crossbar arrays for neuromorphic computing systems.
Discussion(0)
No comments yet. Be the first to comment.