Memristors are nanosized electronic devices that can be used to fabricate next generation memories, and to build up electronic synapses for neuromorphic computing. A memristor consists on a metal-insulator-metal nanocell, in which electrical impulses are applied between the electrodes to modulate the resistivity of the insulator. In this way, a high and a low resistivity state can be intentionally and cyclically induced, which can be used to simulate the ones and zeros of the binary code. The resistivity changes are generated due to local atomic rearrangements produced by the electrical field applied, but understanding this phenomenon is very challenging because i) it takes place in very small areas, and ii) it happens at the insulating stack, which is buried in the top electrode.
The group lead by Prof. Paul C. McIntyre at Stanford University has recently developed a new methodology to observe in situ these local conductivity changes. The method consists on replacing the top electrode by a conductive liquid electrolyte, which can be polarized to stress the insulator. After the stress, the electrolyte can be rinsed and the surface of the insulator is scanned via conductive atomic force microscopy. The nanoscale studies have been carried by Prof. Mario Lanza's group, and reveal the formation of local spots with diameters below 4 nm that are responsible for the conductivity change. By tuning the conductivity of the liquid electrolyte, this method even allows distinguishing the contribution of electrical field and thermal heat into the currents generated.
This work was recently presented in the first China RRAM International Workshop (see http://www.chinarram.org), which was held at Soochow University on June 12th-14th of 2017. This event, hosted by Prof. Mario Lanza, aggregated most world leaders in the field of memristors, including Prof. Philip Wong (Stanford University), Prof. Wei Lu (Michigan University) and Prof. Tony Kenyon (University College London), among many others. The workshop is expected to hold its second edition in 2019 and become a biannual meeting for the memristors community.