Electronic Properties and Memristive Behavior in Undoped Diamond Foils for Artificial Neurons-On-A-Chip Application

The electronic properties of chemical vapor deposited undoped polycrystalline diamond foils are investigated via continuous and pulsed mode measurements. In continuous mode, our diamond thin foil operates as a bipolar memristor with a variable onset voltage and internal nanobattery effect. We attribute the memristive properties to a filamentary redox-based resistive switching mechanism with ionic and electronic charge transport within our diamond foil. In order to explain the occurrence of this dip, we consider the hydrogen redox-based switching model. In this model, hydrogen ions are separated from the conductive filament and migrate towards the negatively charged electrode through atomic vacancies and produce a build-up internal nanobattery voltage. This diffusion process is accelerated by local high temperatures produced by the current carrying filament. While current compliance is not necessary for this effect to occur, it is certainly beneficial because a fixed part of the voltage drop occurs due to electrons travelling towards the positive electrode and the rest of the voltage drop occurs due to protons migrating and accumulating at the negative electrode. After the pulse is turned off, a current in reverse direction flows due to the build-up internal voltage as well as repulsion of accumulated positive charge carriers. In pulsed mode operation, our diamond foils exhibit synaptic behavior, similar to the interface between two neurons. Our memristors exhibits high retention times with high on/off ratios up to 10⁴ required for logic operations. As an inorganic synapse, our foil exhibits sensory, short-term, and long-term memory effects according to a psychological human memory model. Spike-timing dependent plasticity supports our findings. Furthermore, we observe functionalities commonly found in neurons, such as a mimicked refractory period where our conductive filament corresponds to an axon. We anticipate that our study paves the way for future artificial-neurons-on-a-chip application. Our research findings constitute an important step towards artificial-neurons-on-a-chip application, as well as organic-inorganic brain-electronics interfaces.