Random networks of disconnected nanoparticles in dielectric layers as a source of electric responsivity

Many efforts are currently focused on materials with responsive features for neural-inspired devices. Different approaches are followed, based on various mechanisms – from ferroelectric switching to structural phase changes, from magnetic tunnel junctions to metal filament formation. Here we analyze an alternative strategy based on an unconventional electrical response arising from percolative charge transport and charge trapping in discrete random networks of oxide nanostructures in a dielectric matrix. After an analysis of the mechanisms which can potentially be a source of a plastic response in this class of systems, we report evidence of this behavior in a system comprising an alkali-germanosilicate amorphous matrix with incorporated Ga-oxide nanostructures. The active material – consisting in a film 70 nm thick interfaced to p-type Si and Au electrodes – gives a responsive behavior to pulsed bias which is accompanied by bias dependent electric conduction, with resistivity changes of an order of magnitude by applying 2 V, as well as a dielectric response with hysteretic features, as expected by the model. The results represent a first proof of concept of an unexplored strategy for the design of responsive systems.