The intrinsic capacitances in a MOSFET can be simulated using TCAD's ac analysis solve statement. There are 16 total capacitances as provided in this link: https://www.iue.tuwien.ac.at/phd/wagner/node23.html#SECTION001041000000000000000
For the same device specs and dimensions, similar capacitance matrix can be also extracted from spice (ngspice) using the @device_id(Cij) where ij refers to the terminals. While I understand that there will be slight differences between spice and TCAD since spice uses analytical equations, I find differences up to an order of 1e-01.
These are huge differences and I am looking to the experts to shine some light into why this might be?
Dear Premkumar Vincent ,
-TCAD parasitic capacitances and the spice parasitic capacitance for the same device may differ because of the following:
The spice parameters are extracted from the measured static characteristics and small signal measurements.
- If the TCAD can simulate the static characteristics and the small signal characteristics as it is measured and then extract the spice parameters from these simulated characteristics as has been done with the measured characteristics then the capacitances extracted from the two sets will be not much different from each other especially when the simulation and experimental results belong to the same transistor.
- The deviations will be due to tolerances in the technological and physical parameters of the same device type.
- If one extracts directly the small signal capacitances by simulating the device, then it may be different from those extracted from spice which is based on the experimental results.
-The small signal parameters extracted from SPICE are an approximation of those directly measured or calculated by direct simulation.
- Another serious cause of difference is that the these small signal capacitances normally depend on the DC operating conditions.
So, you can compare them only at the same DC operating point.
Best wishes
Kindly check http://www-tcad.stanford.edu/~fmr/research/docs/thesis5p0.pdf
Just to add to what Abdelhalim said, SPICE uses any one of three MOSFET models, depending on what information you give it. Also, as far as I remember, it uses the Meyer model for voltage-variable capacitances so, if your simulator doesn't, that would be another contributory factor.
- Badges
- Science method