Modeling the Short-Channel Effects
Short-channel effects are bias dependent and subject to process fluctuations in channel length, channel doping, oxide thickness, junction depth, mobility, series resistance, ...
- The important question is how to model them, or the sequence in which these effects should be modeled.
- Should the 2D SCE be included in the effective channel length? Or effective LDD lateral diffusion? Or effective channel doping? Or S/D series resistance?
Short-channel effects demonstrate themselves as a gradual effect as the gate length alone is decreased
- Modeling of the individual SCE must be separated from (and can be calibrated by) its long-channel counterpart.
- Empirical fitting should be performed over a wide range of gate lengths.
Saturation-current modeling depends on accurate threshold-voltage modeling as well as mobility and series-resistance modeling
- Bias dependencies should be included in the threshold-voltage modeling.