A Lignin‐Based Stable and Highly Performing Resistive Switching Device for Multilevel Data Storage and Neuromorphic Applications

Abstract This work demonstrates a low‐cost, environment‐friendly, stable, and forming‐free resistive‐switching device using lignin, extracted from coconut husk (CH), embedded in poly(methyl methacrylate) (lignin@PMMA) as the active layer, aluminum (Al) as the top electrode (TE), and fluorine‐doped tin oxide (FTO) as the bottom electrode (BE). By varying the value of −V min in the DC sweep 0 V→2 V→0 V→−V min →0 V, both write‐once‐read‐many times (WORM) memory characteristics and bipolar Resistive Random Access Memory (RRAM) behavior with multilevel resistance in the high‐resistance state (HRS) are observed. The RRAM demonstrates excellent endurance (≈5.5 × 10 4 cycles), long data retention (≈1.02 × 10 7 s), and minimal cycle‐to‐cycle (C2C), device‐to‐device (D2D), and batch‐to‐batch (B2B) variability. The low‐temperature and solution‐based fabrication method aligns well with the CMOS back‐end‐of‐line (BEOL) process. In addition, the device also shows great potential for use as an artificial synapse as it successfully emulates different neuromorphic functions such as excitatory postsynaptic current (EPSC), inhibitory postsynaptic current (IPSC), potentiation‐depression (P‐D), paired‐pulse facilitation (PPF), spike‐timing‐dependent plasticity (STDP), and the transition from short‐term memory (STM) to long‐term memory (LTM). Simulation results of a memristive neural network utilizing the parameters derived from the neuromorphic performance of the device show ≈99% MNIST digit recognition accuracy after minimal epochs.