Hybrid Halide Perovskite-Based Dual-Gated Electrochemical Transistors for Switchable Optoelectronics and Neuromorphic Computing
Implementing Organization
Indian Institute Of Technology Roorkee
Principal Investigator
Dr. Monojit Bag
Indian Institute Of Technology Roorkee
mbagfph@iitr.ac.in
Project Overview
Metal halide perovskites have been the materials of the decade, as tremendous progress in perovskite solar cells, perovskite light-emitting diodes, perovskite-based detectors, and memristor applications has been observed. Due to high charge carrier mobility (theoretically), halide perovskites are also used in field-effect transistor (FET) applications. However, the maximum carrier mobility from the FET measurement reported in the literature is ~ 40 cm2/V.s, which is well below the theoretical limit (over 2000 cm2/V.s) for the perovskite materials. The primary concern in Perovskite-FETs is the understanding of electronic mobility in the presence of ionic transport. Therefore, decoupling electronic-ionic transport in perovskites may improve the overall charge carrier mobility towards the theoretical limit. Typically, high-k dielectrics are used to fabricate high-performance FETs. However, not all the perovskite materials show high electronic/hole mobility in solid-state gate geometry due to strong electronic-ionic coupling in these devices. Alternatively, electrolytes as the gate insulator can produce 10 - 100 times more capacitance than solid dielectrics; therefore, FET can be operated at a lower voltage with improved efficiency and charge carrier mobility. Unfortunately, there are not many aqueous or non-aqueous electrolytes that can be used for perovskite-based FETs, as halide perovskites are highly unstable with most of these electrolytes. The solvent from the electrolytes diffuses into the perovskite active layers, degrading perovskite morphology and optoelectronic properties. In the current proposal, we address these critical issues in electrolyte-gated perovskite FETs and propose remedies to improve the overall stability of the active layer. We also propose a dual-gated perovskite FET configuration to control the electronic-ionic coupling in the channel. Moreover, these electrolyte-gated FETs could be promising candidates for next-generation optoelectronic devices and neuromorphic computing. Our project focuses on the optimization of hybrid perovskite-based FETs that can be used with quasi-solid-state gel electrolytes as dielectric materials. As electrolytes contain ions and have a high dielectric constant, they can function as dielectrics. Traditionally, metal oxide dielectric has been used as the gate electrode in perovskite field-effect transistors, but here we will use electrolyte as the gate electrode. Also, decoupling electronic-ionic transport in perovskites may improve the overall charge carrier mobility. The intrinsic semiconductor parameters, such as charge carrier mobility, bulk resistance, interfacial contact resistance, and capacitance, also affect the device performance. Finally, the response of these devices is much slower than conventional FETs due to the ionic contribution. This slow ion kinetics is similar to that of the synapses. Therefore, these devices can store the previous state data, and they can be used in neuromorphic computing. The objective of this proposal is to fabricate efficient and stable field-effect transistors using hybrid halide perovskite (HHPs) and quasi-solid-state gel electrolyte. To stabilize the structure, we introduce an ion-permeable membrane separator, which would only allow ions to pass through while other molecules would be blocked at the interface. We also propose fabricating a 4-terminal device for tunable FET parameters, where one of the gate dielectrics could be oxide-based materials. The status of HHP-based FET is of the order of 40-50 cm2 V-1 s-1 device mobility. We propose here several objectives to achieve high mobility as well as smart applications in optoelectronics and neuromorphic computing. Therefore, the basic characteristics of these memory devices will be implemented in fully connected neural networks (FCNN) or reservoir computing (RC).