Effects of Scaling on MOS Device Performance

IOSR Journal of VLSI and Signal Processing (IOSR-JVSP)
Volume 5, Issue 1, Ver. III (Jan - Feb. 2015), PP 25-28
e-ISSN: 2319 – 4200, p-ISSN No. : 2319 – 4197
www.iosrjournals.org
DOI: 10.9790/4200-05132528 www.iosrjournals.org 25 | Page
Effects of Scaling on MOS Device Performance
Manish Kumar
Department of ECE, North Eastern Regional Institute of Science & Technology, India
Abstract:This paper presents effects on MOS transistor performance due to scaling of its dimensions. Scaling
theory deals with the change in the device characteristics with the decrease in the dimensions of a MOS
transistor. MOS transistors are continuously scaled down due to the desire for high density and high
functionality VLSI chips. The driving forces behind these developments are increasing the demand for portable
systems requiring high throughput and high integration capacity. Effects of scaling on the performance
characteristics of a MOS device are analyzed in this paper.
Keywords: Scaling, constant field scaling, constant voltage scaling, threshold voltage, scaling factor.
I. Introduction
Scaling of a MOS transistor is concerned with reducing the dimension of the device. The size of a
MOS transistor is reduced by a factor of 0.7 in each technology generation [1]. Scaling results in the decrease of
the dimensions of a MOS device, and thus increases the device density and functional capacity of the chip. More
number of smaller MOS transistors is packed into a smaller chip area, and thus increases the device density and
functional capacity of the chip. The cost of a VLSI chip depends upon the number of chips that can be produced
per wafer. Smaller MOS transistors reduce the chip area, and thus allow more chips per wafer. This reduces the
overall fabrication cost of the chip. MOSFET capacitances are also reduced, which in turn reduces the output
switching time and switching power dissipation [1]. Despite some of the advantages, scaling of MOS device
presents a series of challenges to device design. The electrical characteristics of a MOS transistor change with
the reduction in the device dimensions. Further reducing the size of MOS transistors is restricted due to
difficulties in device fabrication process and need for low voltages. Lowering the supply voltage reduces the
switching speed of a MOS transistor. So, in order to maintain the same switching speed, its threshold voltage
should be reduced at the same rate as the supply voltage is reduced. However, low threshold voltage MOS
transistor results in increasing the subthreshold leakage power dissipation [2-3]. Thus, the relationship between
power dissipation and switching speed is critical in obtaining optimal scaled MOS device. The internal MOS
transistor characteristic is changed with the reduction in the dimension of MOS device, and thus the current flow
characteristics also changes [4-5]. MOSFET characteristics degrade with the reduction in its physical dimension.
Main characteristics are threshold voltage and subthreshold swing. Threshold voltage decreases and
subthreshold swing increases because of two-dimensional (2-D) electrostatic charge sharing between the gate
and the source-drain regions [6-7]. Consequently, the on-to-off current ratio is reduced substantially, which
results in a significant increase in standby power and compromised overall performance [7]. Weaker electrical
performance due to scaling of MOS transistor further necessitate for new initiatives in improving the
performance of the device.
II. Mosfet Scaling
Fig. 1 shows the scaling of a MOS transistor by a scaling factor (S >1). Before scaling, the channel
length and the channel width of the MOS transistor are W and L respectively. After scaling by a factor S, the
dimensions of the scaled MOS device are W’=W/S and L’=L/S. The channel area of the MOS transistor before
scaling is A=WL, whereas the channel area of the MOS transistor after scaling is A’= W’L’=A/S2
. Hence after
scaling, the channel area of a MOS device reduces significantly. Two different scaling options are employed for
scaling the MOS device. These different options are: constant field scaling and constant voltage scaling [8].
Both types of scaling methods have different effects on the performance characteristics of the MOS device.
Fig.1 MOSFET Scaling

2.1 Constant Field Scaling
Fig. 2 shows constant field scaling in an nMOS transistor. In this scaling method [8], physical
dimensions of a MOS transistor are scaled down by a scaling factor (S >1). Table 1 shows the MOS physical
dimensions, potentials, and doping densities in constant field scaling. All terminal potentials and the threshold
voltage of the MOS transistor are also scaled down by the same scaling factor (S). The doping densities (NA and
ND) are increased by a factor S to maintain the charge densities and electric field relations.
Fig. 2 Constant field scaling in a MOSFET
Table 1: MOS physical dimensions, potentials, and doping densities in constant field scaling [8]
Quantity Symbol Before Scaling After Scaling
Channel length L L L’= L/S
Channel width W W W’= W/S
Channel area A A A’= A/S
Gate oxide thickness tox tox t’ox= tox/S
Junction depth xj xj x’j = xj/S
Supply voltage VDD VDD V’DD = VDD/S
Gate to source voltage VGS VGS V’GS = VGS/S
Drain to source voltage VDS VDS V’DS = VDS/S
Substrate to source voltage VBS VBS V’BS = VBS/S
Threshold voltage VTH VTH V’TH = VTH/S
Doping densities NA
ND
NA
ND
N’A = S. NA
N’D = S. ND
2.1.1 Effects Of Constant Field Scaling On Mos Device Performance
The effects of constant field scaling on MOS device performance such as gate oxide capacitance per
unit area, transconductance, drain current, power dissipation, and power dissipation density are shown from
equations (1) – (6).
Gate oxide capacitance per unit area, C’ox = εox/t’ox = S. εox/tox = S. Cox (1)
Transconductance, k’n = µn.C’ox.W’/L’ = S. kn (2)
Drain current, I’D (lin) = k’n/2. [2.(V’GS-V’TH).V’DS – V’2
DS]
= S. kn/2. 1/S2
.[2.(VGS-VTH).VDS – V2
DS]
Hence, I’D (lin) = ID (lin)/S (3)
I’D (sat) = k’n/2. (V’GS-V’TH)2
= S. kn/2. 1/S2
. (VGS-VTH)2
Hence, I’D (sat) = ID (Sat)/S (4)
Power dissipation, P’= I’D . V’DS
= 1/S2
. ID . VDS
Hence, P’= P/S2
(5)
Power dissipation density, P’d = P’/ (W’. L’) = Pd (6)
2.2 Constant Voltage Scaling
Fig. 3 shows constant voltage scaling in an nMOS transistor. In this scaling method [8], physical
dimensions of a MOS transistor are reduced by a scaling factor (S>1). Table 2 shows the MOS physical
dimensions, potentials, and doping densities in constant voltage scaling. The supply voltage, terminal voltages,

and the threshold voltage in a MOS transistor remain unchanged in this scaling method. The doping densities
(NA and ND) are increased by a factor of S2
in order to maintain the charge densities and electric field relations.
Fig. 3 Constant voltage scaling in a MOSFET
Table 2: MOS physical dimensions, potentials, and doping densities in constant voltage scaling [8]
Quantity Symbol Before Scaling After Scaling
Channel length L L L’= L/S
Channel width W W W’= W/S
Channel area A A A’= A/S
Gate oxide thickness tox tox t’ox= tox/S
Junction depth xj xj x’j = xj/S
Supply voltage VDD VDD V’DD = VDD
Gate to source voltage VGS VGS V’GS = VGS
Drain to source voltage VDS VDS V’DS = VDS
Substrate to source voltage VBS VBS V’BS = VBS
Threshold voltage VTH VTH V’TH = VTH
Doping densities NA
ND
NA
ND
N’A = S2
.NA
N’D = S2
.ND
2.2.1 Effects Of Constant Voltage Scaling On Mos Device Performance
The effects of constant voltage scaling on MOS device performance such as gate oxide capacitance per
unit area, transconductance, drain current, power dissipation, and power dissipation density are shown from
equations (7) – (12).
Gate oxide capacitance per unit area, C’ox = εox/t’ox = S. εox/tox = S. Cox (7)
Transconductance, k’n = µn.C’ox.W’/L’ = S. kn (8)
Drain current, I’D (lin) = k’n/2. [2.(V’GS-V’TH).V’DS – V’2
DS]
= S. kn/2. [2.(VGS-VTH).VDS – V2
DS]
Hence, I’D (lin) = S. ID (lin) (9)
I’D (sat) = k’n/2. (V’GS-V’TH)2
= S. kn/2. (VGS-VTH)2
Hence, I’D (sat) = S. ID (Sat) (10)
Power dissipation, P’= I’D . V’DS
= S. ID . VDS
Hence, P’= S. P (11)
Power dissipation density, P’d = P’/ (W’. L’) = S3
. Pd (12)
III. Performance Analysis
Constant voltage scaling may be preferred over constant field scaling in many cases because the scaling
of voltages may not be very practical in many cases. The peripheral and interface circuitry may require certain
voltage levels for all input and output voltages, which in turn would necessitate multiple power supply voltages
and complicated level shifter arrangements [9]. However, it is observed that by using constant voltage scaling in
a MOS transistor, the drain current and power dissipation increases by a factor of S, while power dissipation
density increases by a factor of S3
. This large increase in power dissipation density may cause serious reliability
problems for the scaled transistor, such as electromigration, hot carrier degradation, oxide breakdown, and
electrical over stress [9]. In constant voltage scaling, the performance enhancement results primarily from the
improvement of the drain current. In constant field scaling, the drain current and power dissipation decrease by

S and S2
respectively, while power dissipation density remains unchanged. In this scaling, a much higher
off-state leakage is observed due to the scaling down of the threshold voltage.
IV. Conclusion
Reducing the device dimensions allows higher density and higher logic integration. As the device
dimensions are systematically reduced through constant field scaling or constant voltage scaling, various
physical limitations become increasingly more prominent, and ultimately restrict the amount of feasible scaling
for some device dimensions. In constant field scaling, power dissipation decreases by S2
, while in constant
voltage scaling, drain current increases by S. Hence, constant field scaling can be used in low power
applications, while constant voltage scaling can be used in high switching speed applications. The appropriate
scaling approach in a MOSFET may be used depending upon the applications.
References
[1]. G. E. Moore, Cramming More Components onto Integrated Circuits, Proceedings of the IEEE, 86, 1998, 82-85.
[2]. M. Kumar, M. A. Hussain, and S. K. Paul, Performance of a Two Input Nand Gate Using Subthreshold Leakage Control
Techniques, Journal of Electron Devices, 14, 2012, 1161-1169.
[3]. M. Kumar, M. A. Hussain, and L. K. Singh, Design of a Low Power High Speed ALU in 45nm Using GDI Technique and its
Performance Comparison, Communications in Computer and Information Science, Springer Berlin Heidelberg, 142, 2011, 458-
463.
[4]. T. Yuan, D. A. Buchanan, C. Wei, D. J. Frank, K. E. Ismail, L. Shih-Hsien, G. A. Sai-Halasz, R. G. Viswanathan, H. J. C. Wann, S.
J. Wind, and W. Hon-Sum, CMOS Scaling into the Nanometer Regime, Proceedings of the IEEE, 85, 1997, 486-504.
[5]. S.M.Sze, Physics of Semiconductor Devices, New York: Wiley, 1981.
[6]. T. Ghani, K. Mistry, P. Packan, S. Thompson, M. Stettler, S. Tyagi, and M. Bohr, Scaling Challenges and Device Design
Requirements for High Performance Sub-50 nm Gate Length Planar CMOS Transistors, Symposium on VLSI Technology, 2000,
174-175.
[7]. J. D. Meindl, Low Power Microelectronics: Retrospect and Prospect, Proceedings of the IEEE, 83, 1995, 619-635.
[8]. D. L. Critchlow, MOSFET Scaling: the Driver of VLSI Technology, Proceedings of the IEEE, 87, 1999, 659-667.
[9]. S. Kang, and Y. Leblebici, CMOS Digital Integrated Circuits, New York: McGraw-Hill, 2003.

Effects of Scaling on MOS Device Performance

Recommended

Recommended

More Related Content

What's hot

What's hot (20)

Similar to Effects of Scaling on MOS Device Performance

Similar to Effects of Scaling on MOS Device Performance (20)

More from iosrjce

More from iosrjce (20)

Recently uploaded

Recently uploaded (20)

Effects of Scaling on MOS Device Performance