Phase Change Memory Model Development Based on Cell Geometry

Non-volatile memory (NVM) plays an important role as a data storage element for mobile devices in this era of information. The market of NVM has been dominated by Flash memory for many years due to its low fabrication cost. However, it is starting to approach its limit in scaling for future development which has raised the need to investigate alternatives. Phase change memory (PCM) which offers many advantages over other new type of NVMs is considered to be the potential candidate for next generation mainstream NVM. In order to facilitate the development of PCM, a PCM model that can simulate its circuit level behavior is important. In research literature, some preliminary models are available. However, without considering the cell geometry and properties of the PC material, the PCM behavior cannot be accurately predicted and most of the physics and effects are still missing. A complete PCM model that can correctly predict and explain the circuit behavior is needed.;The behavior of the PCM cell should depend mainly on the properties of PC material and its structure. As the operation of PCM consists of different mechanisms, the whole PCM model is divided into several parts, reading, writing and secondary effects. The corresponding working mechanism is explained by the physical theory and described using mathematical equations in the model. All the sub-models are implemented into a complete PCM compact model by using Verilog-A platform for circuit simulation. The calculation result of the models is compared with the experimental data reported in the literature to verify its correctness.;Using this model, a more realistic PCM behavior in reading and writing can be simulated. By including the secondary effects, the performance of PCM cell in array configuration can be estimated accurately. Also, the PCM cell dimension can be optimized according to the design specifications in a systematic way.