|M.Sc Student||Liraz-Lidji Ron|
|Subject||Mixed Ionic Electronic Conductor Model of the Resistive|
|Department||Department of Electrical Engineering||Supervisors||Professor Dan Ritter|
|Professor Emeritus Ilan Riess|
|Full Thesis text|
The dissertation describes modeling of valence change resistive switching memory elements. The model includes : a) solution of drift diffusion equations for mobile ions and electrons, b) stoichiometry changes by exchange of oxygen with the ambient, c) temperature changes induced by the current, d) Mott transition at high vacancy concentration, and e) switching due to tunneling through a thin insulating layer.
The model agrees very well with experimental SET and RESET characteristics, as well as with the electroforming step. We reproduce the I-V curves as well as the temperature rise that was previously measured.
Using our model, we have carried out a study of the impact of parameters such as ion mobility, activation energy of ions, exchange current at the electrodes and voltage ramp slope. The retention time of the devices and the filament modification during the SET and RESET processes was evaluated. Finally, we have taken advantage of our model to study the performance of memristors in artificial neural systems. Our model reproduces the STDP characteristics of real neurons. By choosing materials with parameters such as the ion mobility energy activation, one can control the reactivity of the artificial neuron.
Side by side with the simulations and the theoretical model, we have performed an experiment using conductive AFM, on thin HfO2 layer deposited on InP of it. We demonstrated morphological changes of the surface that point out the high temperatures during the switching process.