Gradual switching in the memristor or memcapacitor devices is the key parameter for the next generation of bio-inspired neuromorphic computing. Here, we have fabricated the WOx/ZrOx dual-oxide layered device, which shows the coexistence of gradual resistive and capacitive switching arisen from the current and capacitance hysteresis curves, respectively. The expansion of hysteresis loop can be modulated by altering the oxygen content in the oxide materials. Interestingly, the presence of negative differential resistance (NDR) is dependent on the voltage sweep direction and range of applied bias, which can be reasoned by the local electric field, charge trapping/detrapping, and conduction band offset at the dual-oxide interface. This study provides the concept of the coexistence of current and capacitance hysteresis along with NDR, and it is highly potential for memristor and memcapacitor circuits to explore neuromorphic computing.

Metal oxide ZrO2 has been widely explored for resistive switching application due to excellent properties like high ON/OFF ratio, superior data retention, and low operating voltage. However, the conduction mechanism at the atomistic level is still under debate. Therefore, we have performed comprehensive insights into the role of neutral and charged oxygen vacancies in conduction filament (CF) formation and rupture, which are demonstrated using the atomistic simulation based on density functional theory (DFT). Formation energy demonstrated that the fourfold coordinated oxygen vacancy is more stable. In addition, the electronic properties of the defect included supercell confirm the improvement in electrical conductivity due to the presence of additional energy states near Fermi energy. The CF formation and rupture using threefold and fourfold oxygen vacancies are demonstrated through cohesive energy, electron localization function (ELF), and band structure. Cohesive energy analysis confirms the cohesive nature of neutral oxygen vacancies while the isolated behavior for +2 charged oxygen vacancies in the CF. In addition, Nudged Elastic Band (NEB) calculation is also performed to analyze the oxygen vacancy diffusion energy under different paths. Moreover, we have computed the diffusion coefficient and drift velocity of oxygen vacancies to understand the CF. This DFT study described detailed insight into filamentary type resistive switching observed in the experimentally fabricated device. Therefore, this fundamental study provides the platform to explore the switching mechanism of other oxide materials used for memristor device application.