Physical device simulation is used frequently to predict characteristics of the electronic
devices, either in the design phase before the device manufactured, or in an exploratory phase
to investigate the performance of the possible device. This simulation work evaluates the FOM
performances attained by the SiC devices shown in Fig. 4.1. The governing
current transport model used in this simulation is the the drift-diffusion (DD) model where
many of the well known characteristics and features of a semiconductor device are
addressed.
A variety of SiC specific models outlined in
Chapter 3, including band structure and bandgap narrowing;
SRH and auger recombination models incorporated with carrier life time; mobility dependencies
on impurity concentration, lattice temperature, carrier concentration, carrier energy,
parallel and perpendicular electric fields; a breakdown models, including stress dependent,
leakage current, and temperature- and field-dependent impact ionization; an incomplete
ionization model which accounts for ionization level dependence on temperature, polytype and
dopant site locations are few to mention employed during the course of this simulation. For
those model parameters which do not exist at all in SiC, the corresponding silicon parameters
available in MINIMOS-NT are often used.
One of the difficulty which can be seen in the
SiC device simulation is the reverse voltage convergence problem due to the low intrinsic
carrier concentration 
(3.69) at room temperature caused by the wide
bandgap. When the simulator has to cope with the majority carrier concentration on the order
of 10
cm
and a minority carrier concentrations of less than
10
cm, the numerical accuracy is not enough, and cancellation problems are
common. One way around this problem is to conduct the simulation at elevated temperature
ranges and lower once the avalanche condition has been achieved. A similar problem can occur
with deep traps since the filling and emptying is slow at room temperature because of the SiC
large activation energy.
For the anisotropic behavior of SiC's material parameters,
the appropriate anisotropic ratio values has been set depend on the simulated device polytype,
the orientation of the wafer surface either perpendicular or parallel to the c-axis which the
device made up of, and the device structure (vertical or lateral). The mobility and impact
ionization coefficients anisotropic values require great care during the forward and reverse
bias device simulation, respectively.
Another consideration is the channel mobility
in MOS devices. Due to material development and design difficulties, the surface
mobility in SiC is as low as 10% of the bulk mobility, contrary to 50% in Si MOSFET. An
important consideration has been taken into account for the incomplete ionization of dopants
below room temperature and heavily doped SiC devices.
Meshing is an important step
to ensure that a convergent solution can be obtained. The major concern of the meshing scheme
is to make sure the mesh is matched with the non-linear variations of semiconductor physical
quantities. The general rule therefore is to put more node in areas where physical quantities
change rapidly. For example, for MOSFET simulation more mesh lines are allocated near the two
ends of the gate electrode, at the channel-substrate interface and near the source and drain
junctions.
T. Ayalew: SiC Semiconductor Devices Technology, Modeling, and Simulation
