Background: In state-of-the-art nanometer metal-oxide-semiconductor-field-effecttransistors
(MOSFETs), threshold voltage (VT) and drain-induced-barrier-lowering (DIBL) variations
are regarded as significant challenges in circuit analysis, design and characterization.
Method: This paper proposes the improved analytical models to correctly describe VT and DIBL
variations due to random gate length fluctuation employing the propagation of variation (POV)
Result: The presented models are validated that they can accurately capture VT and DIBL’s statistical
characteristics through Monte Carlo simulations for MOSFET devices with 22nm process technology.
Furthermore, our models and simulations both revealed that VT and DIBL variations will increase
dramatically with gate length shrinking for nanometer MOSFETs.
Conclusion: The proposed statistical modeling approach provides a useful pathway for processvariation-
aware circuit design.