Session: 06-04 Neuromorphic Computing
Paper Number: 171895
171895 - Micron-Scale Ionic Liquid-Functionalized Geopolymer Memristors for Long-Term Neuromorphic Computing and Synaptic Plasticity
Neuromorphic computing systems require energy-efficient, scalable devices that emulate biological synapses while maintaining stable memory functions over time. In this study, we present a significant advancement in the development of geopolymer (GP)-based memristors by downscaling device size to the micron scale and enhancing longevity through ionic liquid (IL) functionalization. Traditional GP memristors, although low-cost and sustainable, were previously limited by their large size and short memory retention, primarily due to water evaporation from hydrated pores that support ionic transport. Here, we overcome these limitations by (1) employing micro-molding techniques to fabricate GP memristors as small as 200 µm in diameter—achieving a 99.998% reduction in device volume—and (2) integrating the ionic liquid 1-ethyl-3-methylimidazolium trifluoromethanesulfonate (EMIM⁺ Otf⁻) into the geopolymer’s porous network, improving memory retention by over 50%.
The IL-functionalized GP memristors exhibit all three defining characteristics of memristive devices: a pinched hysteresis in I-V curves, frequency-dependent response, and convergence to a single-valued function at high frequencies. Devices demonstrated reliable resistive switching at activation voltages as low as 0.3 V and retained performance for over 20,000 switching cycles, significantly outperforming both their larger counterparts and comparable organic/inorganic memristors. Synaptic plasticity behaviors were also observed, including paired-pulse facilitation (PPF), paired-pulse depression (PPD), post-tetanic potentiation (PTP), and spike-timing-dependent plasticity (STDP), closely mimicking key dynamics of biological synapses.
Morphological and compositional analyses using SEM, EDX, and ATR-FTIR confirmed uniform pore structures and successful IL incorporation without significant structural degradation. Electrochemical analysis revealed redox activity at the silver electrode interface, but control studies confirmed that this activity does not contribute to the observed memristive behavior, which is instead governed by electroosmosis and ionic mobility within the hydrated, IL-functionalized pore structure. ICP-MS analysis further validated that Na⁺ ions dominate the internal ionic environment, supporting the proposed electroosmotic conduction model.
Comparative testing showed that pristine GP devices lost their memristive behavior within one week due to water loss, while IL-functionalized devices maintained stable performance for over three weeks. Devices also demonstrated consistent synaptic modulation over 800,000 potentiation-depression pulses and accurately emulated Hebbian learning, with programmable excitatory/inhibitory response behavior. These characteristics position IL-functionalized GP memristors as strong candidates for reservoir computing, in-memory processing, and scalable memristor crossbar arrays.
Beyond synaptic emulation, the geopolymer platform offers multifunctionality—including piezoelectric and piezoresistive behavior—making it suitable for intelligent structural health monitoring systems where sensing, processing, and memory can coexist within the same material matrix. This work establishes a path forward for integrating low-cost, biocompatible, and environmentally sustainable materials into neuromorphic hardware and smart infrastructure technologies.
Presenting Author: Reza Montazami Iowa State University
Presenting Author Biography: Dr. Reza Montazami is a Professor of Mechanical Engineering at Iowa State University. He obtained his Bachelor of Science in Physics (Virginia Tech, 2007) and earned his Ph.D. in Materials Science and Engineering (Virginia Tech, 2011). His research focuses on the study and advanced manufacturing of multifunctional soft materials, bioelectronics, bioionics, and electrochemical systems, with applications in biomedical engineering, healthcare, and homeland security. His work has received funding from the DoD, NASA, NSF, NIH, and industry partners. Montazami has contributed to over 90 peer-reviewed articles and maintains an h-index of 33.
Micron-Scale Ionic Liquid-Functionalized Geopolymer Memristors for Long-Term Neuromorphic Computing and Synaptic Plasticity
Paper Type
Technical Presentation Only