Memristive and Threshold Switching in Epitaxial Cu 2– x Se Thin Films for Neuromorphic Computing

Neuromorphic computing is based on electronic devices that emulate both synaptic memory and neuronal activation within compact, energy-efficient architectures. Materials capable of supporting these distinct neuromorphic functions within a single chemical system are therefore highly desirable. Here we report on the switching behavior of epitaxial Cu2–xSe cubic thin films as a function of the copper concentration. Films with low Cu deficiency display volatile threshold switching with sharp turn-on, analogous to artificial neuron activation. In contrast, films with higher Cu deficiency exhibit nonvolatile memristive switching with analog conductance modulation and stable potentiation and depression, consistent with synaptic operation. Our results demonstrate that synaptic and neuron-like behaviors can be obtained by adjusting the stoichiometry within the Cu2–xSe material system.