Functional Oxides

Switching dynamics of functional oxides

Hysteresis is typically noticed in systems/devices that possess certain inertia, causing the value of a physical property to lag behind changes in the mechanism causing it; manifesting memory. Particularly in the case of nanoscale resistive switching elements, this inertia has been ascribed to Joule heating, the electrochemical migration of oxygen ions and vacancies, the lowering of Schottky barrier heights by trapped charge carriers at interfacial states, the phase-change and the formation/rupture of conductive filaments in a device’s core. And despite the fact that these mechanisms have a more substantial effect in lamella devices, we recently showed that similar attributes can be supported on considerably larger systems, contingent upon the extent of the stimulating cause, the nature of the pertinent ions and the barrier medium that governs their kinetics.

HP’s original model hypothesised that their devices comprise a mixture of TiO2 phases, a stoichiometric and a reduced one (TiO2-x), that can facilitate distinct resistive states via controlling the displacement of oxygen vacancies and thus the extent of the two phases. We recently demonstrated that substantial RS is only viable through the formation and annihilation of continuous conductive percolation channels that extend across the whole active region of a device, shorting the top (TE) and bottom (BE) electrodes; no matter what the underlying physical mechanism is. Moreover we have shown that the underlying physical mechanism of resistive switching in TiO2 nano-cells is due to the displacement of Tin+ interstitials that combine with oxygen to form a reduced phase (less insulating) that bridges the two terminals of the device. 

As we approach the atomic-scale, phase transitions of functional oxides become quantum confined, seemingly restraining the devices' functionality into bi-stable states; similarly to quantum point contacts and atomic switches. Oscillations that are observed in various experimental studies are indeed entanglements between quantised conductance states that can be ascribed to metastable phases. These heterophase fluctuations are highly undesirable when it comes to conventional memory and/or computation elements as they tend to cause state restoration and stochasticity. On the contrary, we consider these correlation effects to be unique opportunities for realising dynamic computation elements, similarly to primitive biological elements such as ion-channels, microtubules and chemical synapses. Although such elements are inherently unreliable, they employ rate-limiting reactions to reliably regulate bio-information, with a similar argument made in-situ with Ag2S.