Next: 2.7.2 Minimos-NT Up: 2.7 Simulation Software Previous: 2.7 Simulation Software Contents
Parallel to the MC approach drift-diffusion (DD) and hydrodynamic (HD) multi-dimensional modeling approaches have been employed for studies of GaN based HEMTs. Some of them make use of commercial software while others rely on university-developed simulators. Some simulators are tailored for simulation of nitride-based optoelectronic devices such as TIBERCAD [120,121], although it also features drift-diffusion and hydrodynamic transport and heat balance models. The device simulation tool TCAD STUDIO [122] developed by ESEMI has been used for the simulation of SiC based heterostructures [123], but no support for GaN based devices is available yet. A simulation tool which features models for Nitrides is APSYS by Crosslight [124]. The simulator focuses on compound semiconductors with silicon as a special case. Though the main application area is the analysis of optoelectronic devices, the simulator has also been employed for studies of HEMT structures [125]. Another commercial device simulation tool is GENIUS by Cogenda [126]. While it features DD and HD transport models and also lattice heating, it does not offer built-in material models for Nitrides and consequently has yet to be applied for GaN based HEMTs.
A commercial device simulation tool which is widely used is the SENTAURUS device simulator by Synopsys. It is based on TAURUS MEDICI (which itself is derived from the PISCES software by the Stanford University) and the device simulator DESSIS by ISE. The latter has been employed in numerous theoretical studies of GaN-based HEMTs. Those include the breakdown characteristics dependent on surface defect charges [127], quantum and hot electron effects [128], and self-heating and current collapse effects [129].
Another commercial simulator based on the Stanford tools is ATLAS by Silvaco. It features HD and DD transport models and material models for Nitrides. It was used for field-plate optimization of HEMT structures [130] and drain-lag and gate-lag effects [131] (a DD transport model was considered in those works), but also for verification of analytical models [132].