Memristor devices are electrical resistance switches that can retain a state of internal resistance based on the history of applied voltage and current. In this thesis memristor models are presented which bring together a current versus voltage tunnel barrier formalism with a physical electronics based predictor for time rate change in the tunnel barrier width. Elements of the predictor are based on the physical principles for drift dynamics of the oxygen vacancies in TiO2 unit cells which are created in the electroforming process. Previous memristor models incorporated a tunnel barrier model but they employed a time rate change in tunnel width which was strictly based on empirical fitting with memristor data. This thesis leads to three models. They are hybrid model (HB), hybrid model 2 (HB2) and heuristic model (HTM). The HB2 model is an extension to HB and evaluates the memristor with better physical accuracy. This model makes use of complete domain equations described by Simmons. The HTM is also an extension to the HB model. To achieve nonlinear device switching indicated by experimental data. Simulations are presented for the proposed hybrid drift-tunnel models. Results demonstrate typical memristor current versus voltage profiles can be obtained using triangle and sine wave applied voltages with adjustable amplitudes. Matlab simulations confirm the operation of the models. Spice compatible models were also developed, so that these models can be incorporated into application oriented use of memristor e.g. cross bar simulations, neural network simulations. The HTM model was compared with the HP experimental data and is in agreement with this data.

Date of publication

Spring 5-21-2014

Document Type

Thesis (Local Only Access)



Persistent identifier