14 - Chapter 7
14 - Chapter 7
14 - Chapter 7
7.1 Conclusion
The continuous miniaturization of the CMOS technology resulted in ultra fast
processors and will also degrade the device performance due to Short Channel Effects
(SCEs). In order to overcome these problems, various device engineering techniques
were proposed like source drain engineering (LDD MOSFET) lateral channel
engineering (i.e. graded channel MOSFET, halo and Dielectric Pocket MOSFET),
gate dielectric engineering (i.e. gate stack MOSFET), substrate engineering (i.e.
Silicon On Insulator and Silicon On Nothing MOSFET) and gate electrode
engineering (i.e. double gate MOSFET). The work presented in this dissertation was
mainly focused on Dielectric Pocket MOSFET from lateral channel engineering,
Silicon On Nothing MOSFET from substrate engineering and Double Gate MOSFET
from gate electrode engineering techniques that can overcome the demerits associated
with bulk MOSFET and also offers improve device performance.
In addition, CMOS circuit design demands the accurate modeling of non-classical
devices for describing the behavior of various electrical parameters prior to the device
fabrication. This requires exact solution of the basic semiconductor equation i.e.
Poisson’s equation, continuity equation, current transport equation and other related
equations. The solution of these equations invariably involved numerical analysis.
The situation becomes even more complex for nano-scale devices where 2-
dimensional or 3- dimensional effects are to be accounted in modeling. Thus, a two
dimensional analytical model which can give approximately same results as obtained
from the device simulation within acceptable tolerance would be very useful.
Furthermore, in nano-scale regime, the process variation has also major impact on the
device behavior.
Thus the conclusions derived from the present dissertation are divided into two
sections, namely, (a) Impact of dielectric pockets (or insulating layers) on the Single
Gate geometry and (b) on Double Gate geometry. These dielectric pockets (or
insulating layers) are either present at the side wall of the channel (i.e. inside the
Source/Drain region) or inside the channel region. Physics based analytical model of
the different non-classical architectures are used to study the electrical characteristics
of the device and the same is verified with the numerical simulations obtained from
propagation delay that boost the speed of the device in comparison to SON and ISE
MOSFET.
The results also show that ISESON MOSFET has better Voltage Transfer
Characteristics and higher noise margin than conventional SON and ISE MOSFET.
The analysis predicts that ISESON MOSFET is more reliable in terms of variation in
supply voltage and channel length and is also more resistant towards temperature
variation as the device exhibits better circuit performance as compared to other
devices. The impact of process variation in the position of dielectric pockets (i.e.
either it was present inside the S/D region or inside the channel region) on the circuit
performance is also marginal in ISESON MOSFET resulting in immunity against
process variation.
Moreover the improved performance of ISESON based various logic gates (NAND,
NOR and XOR) in terms of propagation delay is also observed in chapter 4. It was
also found that performance of sequential circuits (i.e. R-S Flip Flop and D Flip Flop)
based on ISESON MOSFET has minimum propagation delay in comparison to SON
and ISE based circuits. In addition, ISESON MOSFET also shows relatively lower
parasitic capacitances and improved device reliability in comparison to SON, and ISE
MOSFETs for channel length down to 22 nm.
The impact of insulating layers (or dielectric pockets) on the performance of double
gate MOSFET has been discussed by considering two cases: 1) presence of dielectric
pocket at the side wall of the source/drain regions i.e. Dielectric Pocket Double Gate
MOSFET and 2) presence of insulating layer inside the channel region i.e. Empty
Space in Double Gate MOSFET.
Chapter 5 explores the impact of Dielectric Pocket on the performance of Double
Gate i.e. DP-DG MOSFET through an efficient 2-D analytical model solved using
superposition technique and consequentially comparing the results with the
conventional double gate and dielectric pocket MOSFETs. The discussed model is
also useful for estimating the drain current from sub-threshold to saturation region of
operation. The proposed model includes evaluation of surface potential, electric field,
threshold voltage (Vth), Drain Induced Barrier Lowering (DIBL), sub-threshold slope
(S), drain current (Ids), trans-conductance (gm), device efficiency (gm/Ids) and Voltage
Transfer Characteristics (VTCs) of NMOS inverter that greatly simplifies the device
characteristics assessment. Analytical results of temperature dependent drain current
model have also been verified with the simulated results. With increasing
temperature, leakage current of the device also increases because of two additional
current components i.e. 1) current due to thermal generation and 2) impact ionization
current.
The results reflect that, DP-DG MOSFET offers better sub-threshold performance
along with improved device reliability in terms of lower electron temperature and
lateral electric field and better linearity distortion metrics like VIP2, VIP3 and IIP3 as
compared to bulk, DP and DG MOSFETs. The higher Ion/Ioff ratio in DP-DG
MOSFET indicates faster switching response (i.e. lower propagation and intrinsic
delay) as compared DG, DP and bulk MOSFETs. The results also highlight the lower
sensitivity of DP-DG MOSFET against temperature variation as compared to other
devices i.e. DP and DG MOSFETs. This is so because of the 1) presence of dielectric
pocket at the side walls that can increase the breakdown voltage of the device, 2)
double gate architecture which can enhance the gate controllability over the channel
resulting in lower leakage current.
DP-DG MOSFET also exhibits potential advantages in terms of lower parasitic
capacitances (Cgs and Cgd), higher noise margin, and lower static and dynamic power
dissipation as compared to other devices for high speed digital applications. In
addition to this, the impact of structural modification (i.e. either dielectric pocket is
present at the S/D junctions i.e. partially inside the channel region or partially inside
the S/D region or inside the drain region) on the performance of DP-DG MOSFET is
also studied. It was found that the presence of DP inside the S/D region shows
enhanced linearity and digital performance. The chapter, thus, presents DP-DG
MOSFET as a viable candidate for analog and high speed digital circuits at 60-nm
technology node.
Chapter 6 explores the impact of Empty Space layer (present inside the channel
region) on the performance of Double Gate MOSFET known as Empty Space in
Double Gate (ESDG) MOSFET through physics based analytical drain current model.
The accuracy of the proposed model calculated using EMA technique is verified using
ATLAS 3D device simulation results. Lower leakage current, higher drive current and
lower threshold voltage roll-off with device scaling pertained by the ESDG MOSFET,
extends its applicability at shorter channel length in comparison to DG, bulk ESS
MOSFET and conventional bulk MOSFET. The impact of varying the length of
empty space layer on the sub-threshold performance like sub-threshold slope (S),
threshold voltage (Vth), DIBL and leakage current has also been examined to
determine optimum length of the empty space layer for efficient device performance.
It was also found that, ESDG MOSFET has improved device performance such as
higher device gain (gm/gd), device efficiency (gm/Ids), output resistance (Rout), early
voltage (Vea), Ion/Ioff ratio, and noise margin and lower delay time ( d ) and on
Scaling CMOS towards sub-100 nm channel length generation necessitates the need
for innovative alternative MOSFET architectures, as proposed in this dissertation, to
circumvent the major problems of scaled MOSFET i.e. undesirable SCEs and leakage
current. The work presented in this dissertation mainly focused on analog and digital
performances of proposed non-classical MOSFETs. Various imperative issues and
applications of proposed devices still need exploration at nano-scale regime. Thus the
suggestions for future work related to this dissertation are highlighted below.
The device reliability issues of ISESON MOSFET can be further suppressed
by using lower work function of the metal gate near the drain side i.e. DMGISESON
MOSFET. Thus the impact of DMG on the performance of ISESON MOSFET will be