Publication date: 22nd July 2026
The rapid growth of artificial intelligence demands new computing hardware capable of overcoming the energy and data-transfer limitations of conventional von Neumann architectures. Memristive devices based on mixed ionic-electronic materials have emerged as promising candidates for neuromorphic and in-memory computing by simultaneously storing and processing information. Among them, halide perovskites offer unique advantages owing to their low ion migration barriers, tunable optoelectronic properties, and solution-processability. However, the instability and toxicity of lead-based perovskites remain major obstacles for practical implementation.[1] Therefore, we develop lead-free two-dimensional halide layered double perovskites (LDPs) as a new platform for memristive and neuromorphic devices. The materials are based on Ruddlesden-Popper (BNA)4AgBiX8) and Dion-Jacobson (PDMA)2AgBiX8 (X = I or Br, BNA = benzylammonoium, PDMA = 1,4-phenyldimethylammonium) compositions synthesized through a combined mechanochemical and solution-processing strategy, enabling phase-pure thin films with controlled crystallization and orientation.[2] The influence of supramolecular spacer design, halide composition, and film processing on structural, optical, and electrical properties will be discussed.[3] The LDPs are integrated into three complementary device platforms: asymmetric metal-insulator-metal (MIM) memristors, photovoltaic full-stack devices, and crossbar architectures. The MIM devices exhibit filamentary resistive switching governed by Ag migration and halide-vacancy dynamics, delivering high ON/OFF ratio along with reliable endurance and retention. The photovoltaic-inspired full-stack architecture further demonstrates the possibility of self-powered memristive operation, where the built-in electric field assists ionic transport and switching while electron- and hole-transport layers regulate carrier injection, suppress leakage currents, and improve switching linearity and energy efficiency.[4] Finally, crossbar devices exhibit artificial synaptic behaviors, representing a significant advancement in the development of environmentally benign memristive materials and providing a versatile platform for next-generation low-power neuromorphic hardware and self-powered artificial intelligence systems.
M.G., P.K., and J.V.M. appreciate the support from the Research Council of Finland (Project No. 362642). M.R.G. and J.V.M. were supported by the European Research Council (ERC) under the European Union’s Horizon research and innovation programme (Grant agreement No.101114653,SmartHyMat). M.R.G. and H.M.U. acknowledge SUSMAT Bluesky University of Turku grant (project no. 26004782).
