Xing Zhou and Khee Yong Lim
School of Electrical & Electronic Engineering, Nanyang Technological
University, Nanyang Avenue, Singapore 639798
Phone: (65) 790-4532. Fax: (65) 791-2687. Email:
exzhou@ntu.edu.sg
Fig. 1 | Fig. 2 | Fig. 3 | Fig. 4 | Fig. 5 | Fig. 6 | Fig. 7 | Fig. 8 | Fig. 9 | Fig. 10
Fig. 1 Channel field from quasi-2D solution for three gate lengths at Vds = Vgs = 2.5 V. VSR lengths (Dl) are all calculated based on the extracted model of the actual devices.
Fig. 2 CLM model with or without Dl fitted to the measured Idsat - Ldrawn data (right axis) using the effective Early voltage (left axis). The extracted x = 0.05354 (without Dl) is a best fit for all Ldrawn devices.
Fig. 3 Measured (symbol) and modeled Ids - Vds curves with (line) and without (cross) CLM for the three devices indicated.
Fig. 4 Modeled average field Eav (left axis) and average velocity [based on <v> = µeff0Eav/2] (right axis) against gate length.
Fig. 5 Length-dependent average field, saturation field (left axis), and effective Early voltage (right axis) at four Vgs values.
Fig. 6 Ids - Vds characteristics for the 0.2-µm device. Symbol: measured; Solid line: model (with CLM); Cross: model (without CLM); Dotted line: model (with x increased by three times).
Fig. 7 Calculated VSR length Dl for all gate lengths and biases (Vds, Vgs) as indicated.
Fig. 8 Calculated channel field and modeled average field for the 0.24-µm device at Vgs = 2.5 (solid line) and 1 V (dotted line). The bias-dependent VSR length Dl is shown in the inset.
Fig. 9 Length- and bias-dependent effective potential drops across Rsd (open/dotted), VSR (open/dashed), and linear channel (solid symbol/line) in saturation.
Fig. 10 Length- and bias-dependent effective potential drops
across Rsd (open/dotted), VSR (open/dashed), and linear
channel (solid symbol/line) in linear region.