Analytical Modeling of Organic–Inorganic CH3NH3PbI3 Perovskite Resistive Switching and its Application for Neuromorphic Recognition. Issue 4 (26th February 2018)

• Record Type: Journal Article
• Authors: Ren, Yanyun; Milo, Valerio; Wang, Zhongqiang; Xu, Haiyang; Ielmini, Daniele; Zhao, Xiaoning; Liu, Yichun
• Is Part Of: Advanced theory and simulations. Volume 1:Issue 4(2018)
• Page Start: n/a
• Page End: n/a
• Publication Date: 2018-02-26
• Subjects: analytical modeling -- neuromorphic computing -- organic–inorganic perovskites -- resistive switching; Science -- Simulation methods -- Periodicals; Science -- Methodology -- Periodicals; Engineering -- Simulation methods -- Periodicals; Engineering -- Methodology -- Periodicals; 507.21
• DOI: 10.1002/adts.201700035
• Languages: English
• ISSNs: 2513-0390

Abstract

Abstract: This paper presents an analytical model of resistive switching in organic–inorganic CH3 NH3 PbI3 perovskite. It is interpreted that the resistive switching phenomenon is due to the formation/rupture of iodine vacancy‐based conductive filament (CF) propagating in both vertical and lateral directions. Set and reset processes are explained in the model by the evolution of the CF length and radius driven by electrical and thermal forces. The model‐based simulation results can describe the experimental results, providing an estimate of several switching parameters such as the activation energy of iodine vacancy migration and the CF diameter. Learning in a two‐transistor/one‐resistor synapse structure is demonstrated by simulations. Finally, the neuromorphic recognition of multiple patterns is demonstrated through a two‐layer neural network consisting of 5625 presynaptic neurons and four postsynaptic neurons. Abstract : An analytical modeling of resistive switching in CH3 NH3 PbI3 RRAM is designed based on the evolution of conductive filament (CF). The parameters used in the simulation are well fitted with experimental data, including the I–V curve, CF's diameter, and activation energy of iodine vacancy migration. The two‐transistor/one‐resistor structure is then used as synapse in a neuromorphic network to demonstrate multipattern learning and recognition.