Indonesian Journal of
Electrical
Engineer
ing and
Computer Science
V
o
l. 10
, No
. 3, Jun
e
20
18
, pp
. 93
4
~
94
2
ISSN: 2502-4752,
DOI: 10.
11591/ij
eecs.v10
.i3.pp934-942
9
34
Jo
urn
a
l
h
o
me
pa
ge
: http://iaescore.c
om/jo
urnals/index.php/ijeecs
Analysis
the E
ffect of Contro
l Factors Op
timization on the
Threshold Voltage of 18 nm
PMOS Using L27 Taguchi Method
Nor
a
ni At
an
1
,
Burhanud
din
Ye
op
Majlis
2
,
Ibrahin
Ahm
a
d
3
,
K.
H.
Ch
on
g
4
1,3,4
De
p
a
r
t
me
n
t
of
E
l
e
c
t
r
o
ni
c a
n
d
C
o
mmunication Engineering,
Un
ivers
iti Tenag
a
Nasional, Malaysia
2
In
stitu
te
o
f
Micro
e
ng
in
eeri
ng and
Nano
elect
ron
i
cs
(IMEN), Un
iv
ersiti Keb
a
ng
saan
Malaysia (UKM),
Malaysia
Article Info
A
B
STRAC
T
Article histo
r
y:
Received Ja
n
9, 2018
Rev
i
sed
Mar
14
, 20
18
Accepted
Mar 28, 2018
This resear
ch p
a
per is abou
t th
e inve
stig
at
ion of Halo Im
plant
a
tion
,
Halo
Im
plantat
i
on
Energ
y
,
Halo
Til
t
, Com
p
ensation Im
pla
n
tation
and
S
ource/Drain
I
m
plantation
.
Th
e
y
ar
e t
y
p
e
s
of
control
fa
ctors
that
us
ed in
achievement of
the thr
e
shold vo
ltag
e
value. To
support the successfully
of
the threshold v
o
ltag
e
(VTH) produci
ng, Tagu
chi method b
y
using L27
orthogonal arr
a
y
was used to optimize th
e co
ntrol factors var
i
ation
.
This
anal
ysis has inv
o
lved with 2 m
a
in fac
t
ors whic
h are break do
wn into five
control f
actors and two noise factors.
Th
e five
control f
actors
were varied
with thr
ee
lev
e
l
s
of ea
ch
and t
h
e two no
ise fa
ctors were
var
i
e
d
with tw
o
leve
ls of each in
27 experim
e
nts. In Taguchi m
e
t
hod, the sta
tisti
c
s
data of 18
nm
P
M
OS
trans
i
s
t
or ar
e from
th
e s
i
gna
l noise r
a
tio (SNR) with
nominal-the
best (NTB) and the analy
s
is of va
riance (ANOVA) are executed
to minimize
the variance of
threshold voltage.
Th
is exper
i
ment implanted b
y
using
Virtual Wafer
Fabricati
on SILVACO
software wh
ich is to design and
fabric
ate
th
e tr
ans
i
s
t
or dev
i
ce
.
Exper
i
m
e
ntal
res
u
lts
rev
eal
ed
tha
t
th
e
optimization method is achiev
ed
to perfo
rm th
e threshold voltage value with
least variance
and the per
cent,
which
is on
ly
2
.
16%. Th
e
thresh
old voltage
value from the
experiment show
s -0.308517
volts while th
e
target valu
e
that
is -0.302 volts from value o
f
Inte
rn
ation
a
l Technolog
y
Roadmap of
semiconductor,
ITRS 2012. Th
e threshol
d vo
ltage valu
e for
18
nm PMOS
transistor is well with
in the
range of -0.3
02 ± 12.7% volts that
i
s
recommendation
b
y
the Intern
ational
Roadmap for Semiconductor prediction
2012.
K
eyw
ords
:
18
nm
PM
OS
Thre
sh
ol
d
V
o
l
t
a
ge
SIL
VAC
O
C
ont
r
o
l
Fact
or
s
L2
7 Ta
guc
hi
M
e
t
h
o
d
Copyright ©
201
8 Institut
e
o
f
Ad
vanced
Engin
eer
ing and S
c
i
e
nce.
All rights re
se
rve
d
.
Co
rresp
ond
i
ng
Autho
r
:
No
ra
ni Ata
n
,
Depa
rt
m
e
nt
of
El
ect
roni
c
an
d
C
o
m
m
uni
cat
i
o
n E
n
gi
neeri
n
g,
Un
i
v
ersiti
Tenag
a
Nasion
al, Pu
traj
aya
Camp
u
s
,
Jalan IKR
A
M-UNITE
N
,
4
300
0 K
a
j
a
n
g
,
Selan
gor
, Malaysia.
Em
a
il: n
o
r
an
i@un
iten
.
edu
.
my
1.
INTRODUCTION
Fro
m
ti
m
e
to
ti
m
e
, th
e semico
n
d
u
c
tor i
n
du
stries t
h
rou
gho
u
t
t
h
e wo
rl
d
with co
ll
ab
oration
wit
h
research
es do
m
a
n
y
wo
rks to
u
p
g
r
ad
e th
e qu
ality an
d
p
e
rfo
rm
an
ce o
f
MOSFET. Th
is ach
iev
e
men
t
is
provide
d
the
better pe
rform
a
nce electroni
cs to user
s. Adva
ncem
ent
in
techno
log
y
allo
ws th
e size o
f
M
O
SFET
s
des
i
gn bec
o
m
e
sm
al
l
e
r and i
n
c
r
ease t
h
e swi
t
c
hi
ng spee
d. Ca
pacitance in
dicates switching spee
d
of t
h
e MOSFE
T
. Due to
our
necessa
ry
to c
o
m
p
act the Int
e
grate
d
Circui
t
as possible as
we can for
ge
tting
sm
al
l
el
ect
roni
cs devi
ces [
1
]
.
Scal
i
ng d
o
w
n
t
h
e si
ze of t
r
ansi
st
o
r
i
s
not
an easy
jo
b a
s
i
t
requi
res
h
i
ghe
r
transistor drive curren
t
,
hi
g
h
er i
n
t
e
grat
i
o
n
densi
t
y
an
d f
a
st
er swi
t
c
hi
n
g
spee
ds [
2
]
.
The m
a
i
n
aim behi
n
d
scalin
g
do
wn
MO
SFETs
g
e
ometr
y
is to
cr
eate a f
a
ster sw
i
t
ch
in
g sp
eed at a low
e
r pr
oductio
n
co
st.
Evaluation Warning : The document was created with Spire.PDF for Python.
In
d
onesi
a
n
J
E
l
ec En
g &
C
o
m
p
Sci
ISS
N
:
2
5
0
2
-
47
52
An
al
ysi
s
t
h
e
Ef
f
ect
of
C
o
nt
rol
Fact
or
s
Opt
i
m
i
z
at
i
o
n
on
t
h
e
T
h
res
hol
d…
(
N
o
r
ani
At
a
n
)
93
5
Th
e scaling
down
t
h
e tran
sisto
r
s
will b
r
i
n
g
g
r
eat eff
ects
on
th
e
reliab
ilit
y o
f
in
teg
r
ated circu
it and
man
u
f
act
u
r
i
n
g co
st. Howev
e
r, th
e p
r
ob
lems will arise to
th
e sh
ort ch
ann
e
l effects su
ch
as leak
ag
e cu
rren
t
whe
r
e the lea
k
age c
u
rre
nt is the m
a
in iss
u
es for sta
t
i
c
po
we
r di
ssi
pat
i
on i
n
st
an
dby
m
ode as t
h
e
si
ze o
f
transistor
been scale. T
h
ere
f
ore, t
h
e s
ubt
hre
s
hol
d lea
k
ag
e
cu
rren
t
rises
du
e to thresho
l
d vo
ltag
e
scaling
and
gate leaka
g
e c
u
rrent i
n
crease
s
due to
scale do
wn
o
f
ox
id
e th
ickn
ess
[
3
].
Th
e sub
t
h
r
eshold
sw
i
n
g is in
creasing
due t
o
t
u
nnel
i
n
g cu
rre
nt
an
d t
h
at
t
h
e pe
rf
or
m
a
nce of
nan
o
scal
ed M
O
SF
ETs i
s
de
gra
d
e
d
. It
i
s
sh
o
w
n t
h
at
t
h
e deg
r
adat
i
o
n of s
ubt
hres
h
o
l
d
swi
n
g i
n
cr
eases wi
t
h
bo
t
h
red
u
ct
i
o
n o
f
chan
nel
l
e
ngt
h an
d i
n
cre
a
se of
channel thic
kness [4].
While
the si
ze of M
O
SFE
Ts is re
duci
n
g, the at
om
s count insi
de the silicon
whic
h
affects th
e prod
u
c
tion
o
f
transisto
r
s pro
p
e
rties is red
u
c
i
n
g
as well. Th
is
will resu
lt in
co
n
s
isten
t
p
l
acemen
t
s
an
d
n
u
m
b
e
r
of co
n
t
ro
lled
dopan
t
[
5
]. Sem
i
c
o
ndu
ctor
p
r
o
c
ess is u
n
c
on
tro
lled
cau
si
n
g
th
e
p
r
o
cess var
i
ab
l
e
s to
have
st
at
i
s
t
i
cal va
ri
at
i
on.
I
n
o
r
de
r t
o
re
d
u
ce
t
h
e i
n
fl
ue
nce
o
f
vari
at
i
o
ns, t
h
e m
a
nufact
u
r
er
i
s
usi
n
g
t
h
e
bi
gge
r
scal
ed p
h
y
s
i
cal
gat
e
l
e
ngt
h f
o
r m
e
m
o
ry
appl
i
cat
i
ons p
u
r
p
os
es. The
d
o
w
n
s
cal
i
ng o
n
l
e
ngt
h gat
e
i
s
s
o
m
e
thi
n
g
th
at can
no
t b
e
av
o
i
d
e
d
an
d
it will b
e
ex
p
ected
to
con
tin
u
e
i
n
th
e co
m
i
n
g
y
ears. Th
is is du
e to
th
e in
ab
il
ity o
f
cont
rol factors to
trac
k
sca
ling of m
i
nim
u
m
feature sizes
.
In a
dva
nce
pr
ocess co
nt
r
o
l
,
ran
dom
vari
at
i
ons i
s
ex
pec
t
ed t
o
i
n
creas
e due t
o
m
i
ni
m
i
zat
i
on of
syste
m
at
ic sh
ifts in
th
e critical d
i
m
e
n
s
io
n. Th
is will le
ad
to
v
a
riation
in
i
m
p
act o
f
th
e o
v
e
rall p
o
wer
di
ssi
pat
i
o
n an
d
per
f
o
r
m
a
nces [6]
.
St
at
i
s
t
i
cal
desi
g
n
i
s
n
o
w
becom
i
ng m
o
re im
port
a
nt
be
cause
of t
h
e
ra
nd
om
vari
at
i
o
ns. T
h
e
st
at
i
s
t
i
c
al
desi
gn
w
h
i
c
h i
n
cl
u
d
es p
r
ocess
va
ri
at
i
on
param
e
t
e
rs has
becom
e
t
h
e gr
eat
eff
ect
i
n
circuit design
as it causing
huge e
ffect in the MOS tr
an
sistor. All
o
f
th
ese tran
si
st
ors nee
d
t
o
u
nde
r
g
o
o
p
tim
izat
io
n
pro
cess to
g
u
a
ran
t
ee a q
u
a
lity o
u
t
co
m
e
wh
ich
to
i
m
p
r
o
v
e th
e p
e
rfo
r
m
a
n
ce o
f
CMOS. Th
e
o
p
tim
izat
io
n
pr
o
cess is alw
a
ys b
e
ing
con
tin
u
e
d
t
h
rou
ghou
t th
ese d
ecade to
en
su
r
e
u
s
er
satisf
action
i
n
u
s
ing
el
ect
roni
c
devi
ces. The t
e
c
hni
que
of
o
p
t
i
m
i
zat
i
on ex
pl
ai
ne
d i
n
t
h
i
s
pape
r
i
s
foc
u
s o
n
t
h
e
cont
r
o
l
fact
or
whi
c
h
will g
i
v
e
effect in
th
e th
resho
l
d
v
o
ltag
e
. In th
is era, th
e Tag
u
c
h
i
m
e
th
o
d
h
a
s b
e
co
m
i
n
g
o
n
e
o
f
t
h
e powerfu
l
t
ool
s use
d
t
o
im
pro
v
e pr
o
d
u
c
t
i
v
i
t
y
duri
n
g t
h
e researc
h
and
devel
opm
ent
pr
ocess. T
h
us, i
t
i
s
possi
bl
e t
o
p
r
od
u
ce a
h
i
gher qu
ality o
f
pro
d
u
c
ts at a lo
wer co
st an
d
i
n
th
e shorter tim
e
[7
].
Wh
ile d
e
sig
n
i
n
g
t
h
e d
e
v
i
ces
b
y
u
s
ing
a d
e
ep
sub
-
m
i
cro
n
t
ech
no
log
y
and an
alyze th
e v
a
riab
ility, it h
a
d
g
r
o
w
n
to
b
ecome a v
e
ry i
m
p
o
rtan
t
to
o
l
. Th
is will allo
w p
r
ed
iction
o
f
th
e resp
onse
in
t
h
e v
e
ry
early
stag
e; j
u
st fro
m
th
e
co
n
t
ro
l
fact
o
r
itself.
The Tag
u
c
h
i
m
e
t
hod i
n
vol
v
i
ng anal
y
s
i
s
o
f
cont
rol
fact
or i
n
w
h
i
c
h
of t
h
e fact
ors
sho
u
l
d
be
m
a
ni
pul
at
ed a
n
d fi
nel
y
ad
ju
st
ed t
o
p
r
o
d
u
ce a
n
i
m
prov
em
ent of res
u
lts. T
h
e
optim
i
zation
for t
h
e c
ont
rol
factor
is the heart of
Taguchi m
e
thod as
the quality can be im
prove
d
while m
a
intaining its de
velopm
ent cost. The
reason
is th
e
pro
cesses are i
n
sen
s
itiv
e to
v
a
riatio
n
o
f
env
i
ron
m
en
tal co
nd
itio
n
s
and
o
t
her no
ise fact
o
r
. Th
is
m
e
t
hod s
o
l
v
es
t
h
e pr
o
b
l
e
m
wi
t
h
t
h
e s
p
eci
al
l
y
desi
gne
d
ort
h
o
g
onal
ar
r
a
y
s
t
h
at
are u
s
ed t
o
a
n
al
y
ze every
co
n
t
r
o
l
f
actor
i
n
sm
all ex
p
e
r
i
men
t
nu
m
b
er
.
By u
s
ing
an
orth
og
on
al ar
r
a
y, th
ese cou
l
d help
design
er
s to
f
i
nd
o
u
t
m
u
ltip
le man
i
pu
lated
facto
r
s on
each
characteristic a
nd faster v
a
riation
in a m
o
re eco
n
o
m
ical way [8
].
2.
R
E
SEARC
H M
ETHOD
2.
1.
Si
mul
a
ti
o
n
o
f
the Fa
bri
c
ati
o
n
At
he
na m
odul
e from
VWF was use
d
t
o
fa
bri
cat
e t
h
e 18
n
m
PM
OS nan
o
s
t
r
uct
u
re
d. Th
e fi
rst
st
ep i
n
fabricatio
n
is
a creatin
g
t
h
e
in
itial su
b
s
trat
e fro
m
a Silic
o
n
p
typ
e
(boro
n
do
p
e
d) wit
h
a dop
ing
7
x
10
14
atom
s/c
m
3
and
o
r
i
e
nt
at
i
o
n
<
1
00>
.
Next
p
r
o
cess i
s
t
o
gene
rat
e
ret
r
o
g
ra
de
N
-
wel
l
by
gr
owi
n
g
a
dry
o
x
y
g
e
n
20
0
Å
o
n
t
h
e t
o
p
o
f
t
h
e s
u
bst
r
at
e f
o
r
2
0
m
i
nut
es
wi
t
h
97
0
o
C
an
d do
pe
d w
i
t
h
Ph
os
p
h
o
r
o
u
s
.
T
h
e d
o
se
i
s
3.
75
x
10
12
atom
s /c
m
3
and ener
gy
im
pl
ant
a
t
i
on is 10
0Ke
V
. T
h
e next
st
ep i
s
t
o
fo
rm
t
h
e Shal
l
o
w Tre
n
ch
Is
ol
at
i
o
n
(ST
I) o
f
1
3
0
-
Å
t
h
i
c
kness i
n
anneal
i
n
g
pr
oce
ss wi
t
h
d
r
y
Ox
y
g
en i
n
2
5
m
i
nut
es at
90
0
o
C. In
th
is pro
cedure, th
e
Low
Pressure Ch
em
ica
l
Vapo
r Depo
sition
p
r
o
cess (L
PC
VD) and
Reactiv
e Ion
Etch
i
n
g
(RIE)
p
r
o
cess were
appl
i
e
d a
n
d i
n
vol
ved t
o
achi
e
ve t
h
e m
a
ki
n
g
o
f
t
h
e
desi
re
d de
pt
h
STI
.
T
h
en
, P
hos
p
h
o
r
Si
l
i
cat
e Gl
ass (PS
G
)
was
devel
o
ped
o
n
t
o
p
of
s
u
b
s
t
r
at
e aft
e
r
wa
f
e
r i
s
un
der
g
oi
n
g
t
h
e
an
neal
i
n
g
pr
ocess at
8
5
0
o
C fo
r 1
5
m
i
nutes
.
After co
m
p
lete th
e pro
cess
of grow
i
ng a
n
d
an
neal
ed
1.
1
nm
Gat
e
Oxi
d
e Thickness
(TOX), t
h
e B
o
ron
Di-
fl
u
o
ri
de
(B
F2
)
wi
t
h
1.
67
5
7
7
77
x 1
0
7
atoms /c
m
3
B
o
ro
n and t
h
e e
n
er
g
y
5KeV
with
a til
t an
g
l
e o
f
7
o
wa
s
i
m
p
l
an
ted
at t
h
e N-well activ
e. Fo
llowed b
y
d
e
po
sition
pro
cess o
f
in
su
lator th
at is Hafn
i
u
m
d
i
o
x
i
d
e
(d
ielectric p
e
rmit
tiv
ity
HfO2,
o
p
t
= 2
2
) on to
p
o
f
b
u
l
k
Silico
n
.
In
th
is
research
th
e leng
th
of HfO2
material
was
1
8
n
m
. Th
en
,
o
n
t
h
e top
o
f
t
h
e in
su
lator was t
h
e depositio
n
pro
c
ess
o
f
g
a
te m
a
teri
al, Titan
i
u
m
S
ilicid
e
(Ti
S
i
2
).
The
P
M
OS
de
vi
ce, P
hos
p
h
o
r
ous
wi
t
h
d
o
se
5
.
5
8
1
x1
0
13
at
om
s /c
m
3
at
30°
a
ngl
e an
d e
n
e
r
gy
,
29
0
K
e
V
was used
i
n
H
a
l
o
Im
pl
ant
a
t
i
on p
r
oce
ss.
T
h
e
chem
i
cal
vapor de
po
si
t
i
on (
C
VD
)
p
r
ocess was used
t
o
de
vel
o
p
sid
e
wall sp
acer with a 0.04
7
μ
m
Silico
n
Nitrid
e layer.
Ag
ai
n
,
B
o
ron with
d
e
nsity o
f
5
.
55
666
6 x 10
13
atom
s/c
m
3
and t
i
l
t
e
d at
7° wi
t
h
11
Ke
V i
m
plant
a
t
i
on e
n
er
g
y
was used i
n
So
urce/
Drai
n
I
m
pl
ant
a
t
i
on pr
ocess.
Next
pr
ocess i
s
t
o
de
vel
o
p a
0.
3
m
layer
o
f
B
o
ron
Ph
osp
hor
Silicate G
l
ass (
B
PSG) an
d fo
llo
w
e
d b
y
th
e
anneal
i
n
g
pr
oc
ess of st
r
u
ct
u
r
e at
850
o
C
.
T
h
e l
a
st
im
pl
ant
a
t
i
on pr
ocess
i
s
C
o
m
p
ensat
i
on
Im
pl
ant
a
t
i
on wi
t
h
Pho
s
pho
rou
s
do
se
o
f
2.5 x 10
13
atom
s /c
m
3
at 60
KeV en
erg
y
and
ang
l
e
tilted
at 7
°
.
Last b
u
t
no
t least, th
e
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
SN
:
2
502
-47
52
I
ndo
n
e
sian
J Elec Eng
& Com
p
Sci, V
o
l. 10
,
No
.
3
,
Jun
e
2
018
:
93
4 – 94
2
93
6
process of de
position the Al
uminum m
a
terial
on
top of the structure a
n
d etch
ing accordingly to form
m
e
ta
l
contacts for Source
and Drai
n
[9],
[
1
0]
. T
h
e
com
p
l
e
t
e
di
agr
a
m
of
1
8nm
str
u
ctur
e as
shown
in Figur
e
1
.
2.
2.
T
a
guc
h
i
L
2
7
Ort
h
o
g
o
n
a
l
Arr
a
y
Met
h
od
The
desi
g
n
i
n
g
of
1
8nm
PM
OS
devi
ce i
s
com
p
l
e
t
e
d by
usi
n
g S
I
L
V
A
C
O si
m
u
l
a
t
i
on. Ne
xt
, t
h
e
st
at
i
s
t
i
cal
of a
n
al
y
s
i
s
p
r
oce
s
s f
o
r
pa
ram
e
t
e
rs
devi
ce
by
usi
n
g
Tag
u
c
h
i
m
e
t
hod.
I
n
t
h
i
s
sect
i
o
n,
T
a
guc
hi
m
e
t
hod i
s
used
t
o
st
u
d
y
t
h
e co
nt
r
o
l
fact
o
r
ef
f
ect
s at
t
h
e t
h
re
shol
d
vol
t
a
ge
(
V
TH
).
The
ort
h
o
g
onal
a
rray
L2
7 i
n
Tag
u
chi
m
e
t
hod i
s
used t
o
opt
im
i
z
e t
h
e cont
r
o
l
fact
or
s i
n
t
e
n
d
ed t
o
c
onsi
d
er
t
h
e im
pact
of the i
n
t
e
ract
i
o
n.
The
i
m
p
o
r
tan
t
o
f
the in
teractio
n
st
u
d
y
is to
id
en
tify wh
ich
co
n
t
ro
l factor h
a
s i
n
teractio
n
with
th
e k
e
y fact
o
r
du
ri
ng
th
e op
ti
m
i
zatio
n
pro
cess is carried
ou
t.
T
h
e
key factor (fac
tor E
)
is sel
ect
ed f
r
om
opt
i
m
izat
i
on p
r
ocess
usi
n
g
ort
h
o
g
onal
ar
r
a
y
L9 Tag
u
chi
m
e
t
hod [
1
1]
. Thi
s
i
s
beca
us
e t
h
e expe
ri
m
e
nt
al
st
udy
o
f
T
a
guc
hi
m
e
t
hod
wi
t
h
ort
h
o
g
onal
ar
r
a
y
L9 onl
y
sh
ows t
h
e d
o
m
i
n
a
nt
fact
o
r
an
d
adj
u
st
m
e
nt
fact
ors wi
t
h
o
u
t
sh
owi
ng t
h
e im
pact
of
t
h
e i
n
t
e
ract
i
o
n.
The
key
fact
o
r
i
s
t
h
e
d
o
m
i
nant
fact
or
f
r
om
ar
ray
L9
Ta
guc
h
i
.
In t
h
i
s
researc
h
, t
h
e o
p
t
i
m
i
zat
ion
pr
ocess
by
usi
n
g o
r
t
h
o
g
o
n
a
l
array
L2
7 T
a
guc
hi
m
e
t
hod
needs
o
n
l
y
fi
ve i
m
port
a
nt
cont
rol
fact
o
r
s
t
o
be c
o
nsi
d
e
r
ed i
n
t
h
e
de
si
g
n
of
ex
pe
ri
m
e
nt
s
wi
t
h
i
n
t
e
ra
ct
i
on.
I
n
t
h
e
pr
evi
o
us
expe
ri
m
e
nt
, whi
c
h st
udi
es t
h
e p
r
ocess
by
usi
n
g a
n
o
r
t
h
o
g
onal
a
rray
L9 i
t
can
b
e
concl
u
de t
h
at
t
h
e
Source/
Drain Im
plantation is selected
as d
o
m
i
nant
fact
or
[1
1]
. S
o
t
h
i
s
st
udy
pr
ov
es t
h
at
t
h
e S
o
urce
/
D
ra
i
n
im
plantation dose is as
factor E
.
So in t
h
e ort
hog
on
al
arrays L27
stu
d
y
, it aim
e
d
to
inv
e
stig
at
e th
e
rel
a
t
i
ons
hi
p
of
i
n
t
e
ract
i
on fa
ct
or E wi
t
h
ot
her c
ont
r
o
l
fa
ct
ors. M
a
ny
expe
ri
m
e
nt
s were d
one t
o
sel
ect
t
h
e
sui
t
a
bl
e fi
ve c
ont
rol
fact
ors
and
n
o
i
s
e fact
ors
.
The st
udy
of i
n
t
e
ra
ct
i
o
n
s
i
n
v
o
l
v
i
n
g
do
se of S
o
urce/
Drai
n
Im
plantation (factor E
)
are c
onsi
d
ere
d
inte
racting with
dose o
f
Halo
Imp
l
an
tation
(facto
r
A),
do
se of Halo
En
erg
y
Im
p
l
antatio
n
(factor
B),
Halo
Tilt
(Factor C)
and d
o
se of C
o
m
p
en
sation
Im
p
l
an
tatio
n
(Fact
o
r
D) as
sh
own
in
Table 1
.
Wh
ile th
e no
is
e factors
are Phos
phor Silicate
Gl
ass (BPS) te
m
p
erature and Boron
Pho
s
pho
r Silicate Glass (BPSG) tem
p
erat
u
r
e at d
i
fferen
t
lev
e
ls ai
m
e
d
to
g
i
v
e
m
o
re sen
s
itiv
ity
to
ch
ang
e
s in
facto
r
s also listed
in Tab
l
e 2.
Sy
m
b
ol
Control
Factor
Unit Level
1
Level 2
Level 3
A Halo
I
m
plantation
ato
m
/c
m
3
5.
5950
00e1
3
[A
1]
5.
6000
00e1
3
[A
2]
5.
6050
00e1
3
[A
3]
B
Halo Energ
y
I
m
pla
n
tation
KeV
294
[B1]
295
[B2]
296
[B3]
C Halo
Tilt
°C
30
[C1]
32
[C2]
34
[C3]
D
Co
m
p
ensation
Im
plantation
ato
m
/c
m
3
14.
871
000e
13
[D
1]
14.
875
000e
13
[D
2]
14.
879
000e
13
[D
3]
E
Sour
ce-
Dr
ain
I
m
plantation
ato
m
/c
m
3
5.
3940
00e1
3
[E
1]
5.
4000
00e1
3
[E
2]
5.
4060
00e1
3
[E
3]
Tabl
e 1.
C
ont
rol Factors a
n
d
their ra
nges
Fi
gu
re
1.
A
d
o
p
i
n
g
pr
ofi
l
e
of
t
h
e 1
8
nm
gat
e
l
e
ngt
h
of
PM
OS t
r
ansi
st
or
Evaluation Warning : The document was created with Spire.PDF for Python.
In
d
onesi
a
n
J
E
l
ec En
g &
C
o
m
p
Sci
ISS
N
:
2
5
0
2
-
47
52
An
al
ysi
s
t
h
e
Ef
f
ect
of
C
o
nt
rol
Fact
or
s
Opt
i
m
i
z
at
i
o
n
on
t
h
e
T
h
res
hol
d…
(
N
o
r
ani
At
a
n
)
93
7
Table 2
. Noise Factor
and
its range
s
3.
R
E
SU
LTS AN
D ANA
LY
SIS
3.
1.
A
n
al
ysi
s
of
1
8
nm
P
M
O
S
De
vi
ce
Tagu
ch
i
L27
o
r
t
h
ogo
n
a
l ar
ray esti
m
a
tes t
w
en
ty sev
e
n
ex
p
e
r
i
m
e
n
t
s w
ith
on
e
hu
ndr
ed
an
d eigh
t
sim
u
l
a
t
i
ons ru
n bet
w
ee
n t
h
e
com
b
i
n
at
i
on fi
ve co
nt
r
o
l
fa
ct
ors an
d t
w
o
noi
se fact
ors
.
The desi
gn
va
l
u
e of
18
nm
PM
OS
t
h
res
h
ol
d v
o
l
t
a
ge
i
s
-0
.3
0
2
± 12
.7
% Vol
t
s
.
Th
is
v
a
lu
e is referred to
In
tern
atio
n
a
l
Ro
admap
of
Sem
i
cond
uct
o
r
(ITR
S
)
2
0
1
2
[1
2]
. T
h
at
m
e
an t
h
e t
h
res
h
ol
d
vol
t
a
ge
(
V
T
H
) i
s
i
n
N
o
m
i
nal
-
t
h
e
-
B
e
t
t
e
r
(NTB
)
q
u
a
lity Tagu
chi’s categ
ories.
Based
on
th
e co
n
t
ro
l
p
a
ram
e
t
e
rs listed
in
Tab
l
e 1
with
a co
m
b
in
atio
n
of
n
o
i
se
fact
or
s i
n
Ta
bl
e 2, t
h
e de
si
g
n
o
p
t
i
m
i
z
at
i
on pr
ocess i
s
ca
r
r
i
e
d
out
refe
rri
ng t
o
t
h
e
PM
OS t
r
a
n
si
st
o
r
array
s
co
m
b
in
atio
n
L2
7 Tag
u
c
h
i
m
e
th
od
and
th
e resu
lts of t
h
e
an
alysis VTH listed
i
n
Tab
l
e
3
.
Th
e
read
i
n
g for
VTH
is b
e
tween
-0.89
278
to -0
.12
7
7
1
9
Vo
lts.
It tak
e
s
8
ho
urs an
d 64
m
i
n
u
t
es to
co
m
p
lete th
e si
m
u
latio
n
.
Tabl
e 3.
18
n
m
PMOS Statisti
cal
Resu
lt-Tagu
ch
i L27
Ort
h
o
gon
al
Array
E
xp.
No
T
h
r
e
shold
voltage
X
1
Y
1
T
h
r
e
shold
voltage
X
1
Y
1
T
h
r
e
shold
voltage
X
1
Y
1
T
h
r
e
shold
voltage
X
1
Y
1
Mean
Variance
SNR
(No
m
in
al-th
e
Better),
dB
1
-
0
.
29829
-
0
.
27788
-
0
.
28759
-
0
.
27591
-
0
.
285
1.
06E
-
0
4
28.
86
2
-
0
.
24944
-
0
.
23777
-
0
.
24747
-
0
.
23581
-
0
.
243
4.
66E
-
0
5
31.
01
3
-
0
.
23932
-
0
.
22767
-
0
.
23735
-
0
.
22569
-
0
.
233
4.
66E
-
0
5
30.
64
4
-
0
.
33395
-
0
.
32700
-
0
.
33280
-
0
.
32560
-
0
.
330
1.
72E
-
0
5
38.
00
5
-
0
.
32859
-
0
.
31718
-
0
.
32696
-
0
.
31513
-
0
.
322
4.
61E
-
0
5
33.
52
6
-
0
.
31853
-
0
.
30690
-
0
.
31647
-
0
.
30489
-
0
.
312
4.
63E
-
0
5
33.
22
7
-
0
.
75252
-
0
.
89278
-
0
.
47339
-
0
.
82500
-
0
.
736
3.
39E
02
12.
03
8
-
0
.
89144
-
0
.
69478
-
0
.
82431
-
0
.
67680
-
0
.
772
1.
07E
-
0
2
17.
47
9
-
0
.
69536
-
0
.
61845
-
0
.
67749
-
0
.
60683
-
0
.
650
1.
89E
-
0
3
23.
49
10
-
0
.
29134
-
0
.
28005
-
0
.
28936
-
0
.
27809
-
0
.
285
4.
37E
-
0
5
32.
68
11
-
0
.
28134
-
0
.
27015
-
0
.
27938
-
0
.
26821
-
0
.
275
4.
30E
-
0
5
32.
45
12
-
0
.
27143
-
0
.
26031
-
0
.
26949
-
0
.
25838
-
0
.
265
4.
24E
-
0
5
32.
19
13
-
0
.
43722
-
0
.
41947
-
0
.
43357
-
0
.
41629
-
0
.
427
1.
06E
-
0
4
32.
34
14
-
0
.
42103
-
0
.
40225
-
0
.
41785
-
0
.
40058
-
0
.
410
1.
10E
-
0
4
31.
83
15
-
0
.
40694
-
0
.
39309
-
0
.
40414
-
0
.
39066
-
0
.
399
6.
45E
-
0
5
33.
91
16
-
0
.
36410
-
0
.
35455
-
0
.
36284
-
0
.
35235
-
0
.
358
3.
45E
-
0
5
35.
71
17
-
0
.
35631
-
0
.
34344
-
0
.
35411
-
0
.
34129
-
0
.
349
5.
66E
-
0
5
33.
32
18
-
0
.
34518
-
0
.
33257
-
0
.
34303
-
0
.
33048
-
0
.
338
5.
43E
-
0
5
33.
23
19
-
0
.
12772
-
0
.
13232
-
0
.
12919
-
0
.
13377
-
0
.
131
7.
74E
-
0
6
33.
44
20
-
0
.
13158
-
0
.
13609
-
0
.
13330
-
0
.
13748
-
0
.
135
7.
12E
-
0
6
34.
06
21
-
0
.
13530
-
0
.
14855
-
0
.
13998
-
0
.
15360
-
0
.
144
6.
81E
-
0
5
24.
86
22
-
0
.
29751
-
0
.
28559
-
0
.
29549
-
0
.
28360
-
0
.
291
4.
86E
-
0
5
32.
40
23
-
0
.
28721
-
0
.
27538
-
0
.
28521
-
0
.
27340
-
0
.
280
4.
79E
-
0
5
32.
15
24
-
0
.
27699
-
0
.
26218
-
0
.
27500
-
0
.
26324
-
0
.
269
5.
97E
-
0
5
30.
85
25
-
0
.
35178
-
0
.
34402
-
0
.
35048
-
0
.
34278
-
0
.
347
2.
05E
-
0
5
37.
70
26
-
0
.
34568
-
0
.
32382
-
0
.
34444
-
0
.
32037
-
0
.
334
1.
78E
-
0
4
27.
96
27
-
0
.
33987
-
0
.
33270
-
0
.
33777
-
0
.
33022
-
0
.
335
1.
98E
-
0
5
37.
53
Fro
m
th
e th
resh
o
l
d
vo
ltag
e
valu
es, th
e d
a
ta o
f
m
ean
(
μ
), varia
n
ce
(
and Si
gnal
-
t
o
-
N
oi
se (
S
NR
)
No
m
i
n
a
l-th
e-B
e
tter
NTB
) c
a
n
be cal
cul
a
t
e
d
by
usi
n
g t
h
e f
o
rm
ul
as bel
o
w
[
13]
:
SNR (NTB),
NTB
= 10 Log
10
2
2
(
1
)
W
h
ere:
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
SN
:
2
502
-47
52
I
ndo
n
e
sian
J Elec Eng
& Com
p
Sci, V
o
l. 10
,
No
.
3
,
Jun
e
2
018
:
93
4 – 94
2
93
8
M
ean,
μ
=
n
n
i
........
(
2
)
Variance,
1
1
2
n
n
i
i
B
a
sed o
n
t
h
e
equat
i
o
ns
, n i
s
t
h
e num
ber o
f
t
e
st
s, Y i
s
t
h
e ex
peri
m
e
nt
al
val
u
e of t
h
e VTH
.
The
equat
i
o
n (2
)
a
n
d
eq
uat
i
o
n (3
) are
t
h
e fo
rm
ulas to calculate
mean values
a
nd va
riance va
lues
res
p
ective
l
y.
By
appl
y
i
n
g
t
h
e b
o
t
h
f
o
rm
ul
as i
n
eq
uat
i
on
(3
),
t
h
e SNR
(N
T
B
),
NTB
fo
r th
e PM
OS de
vic
e
was calculated and
th
e resu
lts
were also
listed in
Tab
l
e
3
.
Based
o
n
t
h
e resu
lts ob
tain
ed in
Tab
l
e 3
,
n
e
x
t
p
a
rt is to
d
e
termin
e th
e con
t
ro
l p
a
ram
e
te
rs th
at affect
to
th
e ch
ang
i
ng
of th
e ch
aracteristics o
f
the d
e
v
i
ce
with
d
i
sp
lay th
e
h
i
gh
est SNR v
a
l
u
e o
f
each
lev
e
l. SNR
(NTB
)
fo
r eac
h level
o
f
c
ont
r
o
l pa
ram
e
ters is sh
o
w
n
in
Tab
l
e 4.
Tabl
e 4.
Resu
lt of
SN
R (N
T
B
)
,
C
ont
r
o
l
Fa
ct
ors
Thr
o
ug
h t
h
i
s
i
n
f
o
rm
at
i
on, t
h
e d
o
m
i
nat
i
ng fact
or ca
n
be d
e
term
in
ed
in
th
e
factor
A, fact
or B,
facto
r
C,
factor D and
factor E during
th
e
p
r
o
cess
v
a
riabilit
y is carried ou
t and
it can
b
e
selected
for the
veri
fi
cat
i
o
n
pr
ocess t
o
t
h
e e
n
d. R
e
fer t
o
Ta
bl
e 4
,
t
h
e
hi
g
h
e
st
val
u
e
of
ea
ch c
ont
r
o
l
fact
or i
n
t
h
e t
e
xt
i
ndi
cat
es
th
at th
e lev
e
l
o
f
th
is
p
a
ram
e
ter is th
e
h
i
ghest of SNR
(NTB), th
is m
e
an
s t
h
e
b
e
tter
q
u
a
lity of t
h
resh
o
l
d
vol
t
a
ge
, VT
H [1
4]
. So i
t
can be obse
r
v
e
d i
n
fact
o
r
A, s
h
o
w
ed
th
at th
e lev
e
l A2
of Halo
i
m
p
l
an
tatio
n
do
se
wi
t
h
a val
u
e
of
33
.0
7 dB
S
N
R
can be c
o
nsi
d
e
r
ed as
do
m
i
nat
i
ng as t
h
e hi
ghe
st
l
e
vel
of n
o
i
s
e ge
ne
rat
e
d
co
m
p
ared
with lev
e
l A1
and
l
e
v
e
l A3
. Factor B2 of
Ha
lo
En
erg
y
im
p
l
an
tatio
n
d
o
s
e and
facto
r
C
2
o
f
Halo
Tilt
w
h
ich
r
e
sp
ecti
v
ely show
th
e
h
i
gh
est
v
a
lu
es
o
f
SN
R
w
ith 33
.1
4 d
B
and
33
.9
2 d
B
.
W
h
il
e Factor
D
sh
ow
s t
h
e
l
e
vel
2 i
s
t
h
e h
i
ghest
val
u
e o
f
33.
2
6
dB
an
d
l
a
st
l
y
fact
or E i
ndi
cat
es t
h
e l
e
vel
3 i
s
t
h
e hi
g
h
est
val
u
e
of
3
1
.
10
dB
. T
h
e
ave
r
a
g
e
val
u
es
of
S
N
R
(
N
TB
)
dat
a
i
s
3
1
.
0
0
dB
an
d th
e ev
al
u
a
tio
n resu
lt
with
ou
t in
teracti
o
n
s
p
r
o
cess
is A2 B2 C2 D
2
E
1
.
3.
1.
An
al
ysi
s
of
A
N
O
V
A
R
e
sul
t
w
i
th
Inte
racti
o
n
This resea
r
ch
was stu
d
y
i
ng
the effects o
f
interactio
ns
(EX
A
)
,
(E
XB
), (E
XC) a
n
d
(EX
D
) i
n
d
e
term
in
atio
n
o
f
th
e op
tim
al
co
m
b
in
atio
n
of co
n
t
ro
l fact
ors. The aim is
to calculate
the avera
g
e value
of the
i
n
t
e
ract
i
o
n
of
Fact
or
E t
h
at
i
s
So
ur
ce/
D
r
a
i
n Im
pl
ant
a
t
i
o
n
(E
1,
E2
, E
3
)
wi
t
h
ot
he
r
f
act
ors
suc
h
as
Hal
o
Im
p
l
an
tatio
n
(A1,
A2,
A3
), Halo En
erg
y
Im
p
l
an
ta
tio
n (B1, B
2
, B3
),
Halo
Tilt (C1, C
2
, C3
), and
Co
m
p
en
satio
n Im
p
l
an
tatio
n
(D1
,
D2
,
D3). Th
en
, all th
e
d
a
ta were tran
sfer to
t
h
e
gr
aph
s
and
stud
y th
e
i
n
t
e
ract
i
o
n
.
T
h
e
gra
p
h
s
o
f
t
h
e rel
a
t
i
ons
hi
p
bet
w
ee
n c
ont
r
o
l
fact
o
r
s t
o
S
N
R
(
N
TB
)
can
be
pl
ot
t
e
d a
n
d
sh
ow
n
i
n
Fi
gu
re
6,
Fi
gu
re
7,
Fi
gu
re
8 a
n
d Fi
gu
re
9
.
T
h
e
pre
s
ence
o
f
i
n
t
e
ract
i
o
n
t
h
r
o
u
g
h
gra
p
hs
can
be
obse
r
v
e
d
by
t
h
e exi
s
t
e
nce
o
f
l
i
n
es t
h
at
i
n
t
e
rsect
an
d i
n
co
n
s
i
s
t
e
ncy
am
ong t
h
e
fact
ors i
n
vol
ved
.
Fi
gu
re
6,
Fi
g
u
re
7,
Fi
gu
re
8
and
Fi
g
u
r
e
9 re
spect
i
v
el
y
sh
o
w
t
h
e
rel
a
t
i
o
ns
hi
p
bet
w
ee
n i
n
t
e
ract
i
on
E (
E
XA
),
(E
XB
),
(
E
XC
)
an
d
(E
X
D
).
Fi
gu
re
6 s
h
o
w
s t
h
e S
N
R
(N
TB
) val
u
es
fo
r
gra
p
h i
n
t
e
ract
i
on
bet
w
ee
n
H
a
l
o
Im
pl
ant
a
t
i
o
n
(
f
act
or
A
)
and Source/Drain Im
plantation
(factor E
)
.
There
are
3 lines show
on
the graph, which ar
e A1
,
A2
an
d
A3
,
wh
ich
are in
tersect with
E lev
e
l, E1,
E2 a
n
d E3. From
that, it can be seen
as crossing the 2 lines
plot in
bet
w
ee
n
A2 a
n
d A
3
o
n
l
y
. B
u
t
A1
l
e
vel
cr
oss
i
ng
do
es n
o
t
e
x
i
s
t
.
S
o
det
e
rm
i
n
ed t
h
e hi
ghe
s
t
SNR
(
N
TB
)
l
e
vel
o
n
lin
es t
h
at intersect with
E1 p
a
ram
e
ter is A3
(34
.
51
d
B
).
Evaluation Warning : The document was created with Spire.PDF for Python.
In
d
onesi
a
n
J
E
l
ec En
g &
C
o
m
p
Sci
ISS
N
:
2
5
0
2
-
47
52
An
al
ysi
s
t
h
e
Ef
f
ect
of
C
o
nt
rol
Fact
or
s
Opt
i
m
i
z
at
i
o
n
on
t
h
e
T
h
res
hol
d…
(
N
o
r
ani
At
a
n
)
93
9
Fig
u
re
6
.
Th
e in
teraction
b
e
tween So
urce/Drain
Im
p
l
an
tati
o
n
and
Halo Im
p
l
an
tatio
n
Figure
7. T
h
e
interaction bet
w
een Source/Drai
n
Im
pl
ant
a
t
i
on
an
d
Hal
o
E
n
er
gy
Im
pl
ant
a
t
i
o
n
Fig
u
re
7 is the in
teraction
g
r
aph
for
SNR (NTB) b
e
t
w
een
Halo
Im
p
l
an
tatio
n
(facto
r
B
)
wit
h
So
urce/
Drai
n
I
m
pl
ant
a
t
i
on (f
act
or E)
.
The
r
e are 3 lines which are B1, B2
an
d B
3
are p
l
ot
t
i
ng o
n
t
h
e
gra
p
h,
wh
ere th
ey are in
tersect with
lev
e
l E1
, E2
an
d
E3
. Fro
m
that, 2 lines are
crossing to
ea
ch ot
he
r. Li
ne
B
1
i
s
cro
s
sed
with
lin
e B2
at E2
lev
e
l an
d
lin
e
B1
is cro
sse
d
with
lin
e B3
at E3
lev
e
l. Refer to
th
e graph
,
th
e
h
i
gh
est SNR
(NTB) lev
e
l th
at in
tersection
w
ith
E2
p
a
ram
e
ter is B2 (3
4.2
5
d
B
).
Fig
u
re
8
.
Th
e in
teraction
b
e
tween So
urce/
Drai
n Im
planta
tion a
n
d Hal
o
Tilt
Fig
u
re
8
sh
ows th
e in
teraction
resu
lt
b
e
tween
Ha
lo
Tilt (facto
r
C) and
So
urce/Drain
Im
p
l
an
tatio
n
(fact
o
r
E).
Gra
ph f
r
o
m
Fi
gure
8 i
s
sam
e
wi
t
h
grap
h f
r
om
Figu
re 6,
onl
y
2 l
i
n
es t
h
at
i
s
C
1
l
i
n
e i
s
crosse
d
wi
t
h
C2
lin
e.
Wh
ile C3
lin
e is crossin
g
do
es no
t
ex
ist. Re
su
lt
sh
ows
th
e h
i
gh
er v
a
lu
e of
SNR
(NTB)
of
Halo
Tilt
is in
teraction
with
Sou
r
ce /Drain
im
p
l
an
tati
o
n
is C2 with 36
.1
3 d
B
.
Fi
gu
re
9 s
h
ow
s t
h
e S
N
R
(N
TB
) i
n
t
e
ract
i
o
n
bet
w
ee
n C
o
m
p
ensat
i
on I
m
pl
ant
a
t
i
on (f
act
or
D)
a
n
d
Source/
Drain Im
plantation
(fact
or E). The
gra
p
h dis
p
lays the 2 lines
are
crossi
ng t
o
each ot
her. T
h
e
highest
v
a
lu
e
of
SN
R
(N
TB) fo
r Co
mp
ensatio
n I
m
pl
ant
a
t
i
on
i
s
D
2
wi
t
h
35
.7
2 dB
.
B
a
sed
o
n
i
n
f
o
r
m
at
i
on anal
y
s
i
s
pe
rf
orm
e
d, t
h
e
o
p
t
i
m
u
m
com
b
i
n
at
i
on
fo
r
PM
O
S
devi
ce
s VT
H
anal
y
s
i
s
t
h
at
takes int
o
acc
ount t
h
e e
ffect
of interacti
o
n is
A3, B2, C
2
,
D2,
E1.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
SN
:
2
502
-47
52
I
ndo
n
e
sian
J Elec Eng
& Com
p
Sci, V
o
l. 10
,
No
.
3
,
Jun
e
2
018
:
93
4 – 94
2
94
0
Fig
u
re
9
.
Th
e in
teraction
b
e
tween So
urce/Drain
Im
p
l
an
tati
o
n
and
C
o
m
p
en
sation
Im
p
l
antatio
n
3.
2.
C
o
n
f
i
r
m
a
ti
on T
e
s
t
The e
nd
of t
h
e
sim
u
l
a
t
i
on wi
t
h
t
h
e
noi
se
fac
t
or
of t
h
e
fi
nal
t
e
st
veri
fi
cat
i
o
n sh
o
u
l
d
be ca
rri
ed
o
u
t
t
o
verify the acc
uracy of the forecast Ta
guc
hi
m
e
thod. T
h
e com
b
ination of optim
u
m
cont
rol factors
(A3, B2,
C2
, D2,
E1)
listed
in
Tab
l
e 5
.
Table 5
. C
o
m
b
i
n
at
i
on
o
f
opt
i
m
u
m
L27 anal
y
s
i
s
fo
r
18
nm
PM
OS
Wh
ile th
e fin
a
l v
e
rificatio
n
decisio
n
with
no
ise f
actor and com
b
ination of
op
tim
u
m
co
n
t
ro
l facto
r
s
for
th
e
an
alysis
of
arrays
L27
Tagu
ch
i
m
e
th
od
s for
t
h
resho
l
d
v
o
ltag
e
, VTH with
v
a
lu
e -0
.3
085
17
vo
lts.
Tabl
e 6 s
h
o
w
s
t
h
e l
i
s
t
s
of t
h
e perce
n
t
a
ge
di
f
f
ere
n
ce bet
w
ee
n the res
u
lts of the com
b
ination
without
th
e in
teraction an
d
co
m
b
in
atio
n
with
in
teractio
n
during
dev
i
ce op
timiza
ti
o
n
pro
cess is carried ou
t,
wh
ich
it
refe
rs t
o
t
h
e
no
m
i
nal
val
u
e
pr
oject
e
d
by
ITR
S
2
0
1
2
.
T
h
e fi
nal
deci
si
on a
n
al
y
s
i
s
m
e
t
hods
i
n
de
si
g
n
i
n
g
m
odel
L2
7 Tag
u
c
h
i
18
nm
PM
OS devi
ce sh
o
w
s
t
h
at
t
h
e t
w
o expe
ri
m
e
nt
s i
n
whi
c
h t
h
e i
n
t
e
ract
i
on
or
wi
t
h
o
u
t
in
teractio
n
take
effect of VTH v
a
lu
e with
in
th
e
r
a
ng
e
o
f
the no
m
i
n
a
l v
a
lue (-
0.302
12
.7%
vol
t
s
)
.
Howe
ver, the perce
n
tage
of
the expe
rim
e
nt that ta
kes into account the
effect
of the interaction is
bet
t
e
r t
h
a
n
n
o
m
i
n
al
val
u
es
w
i
t
hout
t
a
ki
n
g
i
n
t
e
ract
i
on
wi
t
h
2.
16
%.
Anal
y
s
i
s
sh
ows
t
h
at
t
h
e
key
fact
or
o
r
fact
or
E i
n
t
h
i
s
st
u
d
y
,
So
ur
ce/
Drai
n
Im
pl
ant
a
t
i
on d
o
sa
ge fact
or
ha
d i
n
t
e
ract
i
o
n
w
i
t
h
ot
he
r c
o
nt
r
o
l
fact
ors
.
Table 6
. A
n
al
y
s
i
s
of
Pe
rcent
a
ge
T
h
re
sh
ol
d
Vol
t
a
ge
, VT
H 18
nm
PM
OS
Evaluation Warning : The document was created with Spire.PDF for Python.
In
d
onesi
a
n
J
E
l
ec En
g &
C
o
m
p
Sci
ISS
N
:
2
5
0
2
-
47
52
An
al
ysi
s
t
h
e
Ef
f
ect
of
C
o
nt
rol
Fact
or
s
Opt
i
m
i
z
at
i
o
n
on
t
h
e
T
h
res
hol
d…
(
N
o
r
ani
At
a
n
)
94
1
4.
CO
NCL
USI
O
N
In the
prese
n
t
study proves t
h
at
, the control factors effect
s the
t
h
res
hol
d
vol
t
a
ge
, VT
H
of 1
8
n
m
PMOS t
r
ansist
or was
success
f
ul
found together with th
e optim
a
l
factors level
pred
icted by Ta
guchi m
e
thod.
So
urce/
Drai
n
I
m
pl
ant
a
t
i
on d
o
s
age
fact
o
r
has
bee
n
i
d
e
n
t
i
f
i
e
d as
key
fact
or
ha
d i
n
t
e
ract
i
o
n
wi
t
h
ot
he
r c
ont
rol
facto
r
s su
ch as Hal
o
Im
p
l
antatio
n
,
Halo
Im
p
l
an
tatio
n
En
erg
y
,
Halo Tilt an
d
C
o
m
p
en
sation
Im
p
l
an
tatio
n
.
There
f
ore,
i
t
h
a
s bee
n
p
r
ove
n t
h
at
18
nm
t
r
ansi
st
o
r
can
be
achi
e
ve
d
pr
od
uced t
h
e
VTH
val
u
e i
s
wel
l
wi
t
h
i
n
th
e I
T
R
S
2
012
r
e
qu
ir
em
en
ts of
-0
.30
2
± 12
.7% vo
lts.
ACKNOWLE
DGE
M
ENTS
Th
e au
t
h
or wou
l
d
lik
e to
th
ank
to
Min
i
stry
o
f
High
er Ed
u
c
atio
n
(MOE), In
stitu
te
of
Micro
e
ng
in
eerin
g
an
d
Nanoelectro
n
i
cs (IMEN) Un
i
v
er
siti Keb
a
ng
saan
Malaysia (UKM
), and
C
e
n
t
re of
Micro
and
Nan
o
Eng
i
n
e
ering
(CeM
NE) Un
iv
ersiti
Ten
a
ga
Nasion
al (UNITEN) fo
r
fi
n
a
n
c
ial,
faciliti
es and
m
o
ral
t
h
ro
ug
h
out
t
h
e
pr
o
j
ect
.
REFERE
NC
ES
[1]
Md.
Ala
m
gir H
o
ssa
in,
et
al
., “
C
apac
itan
ce-Vol
t
age Ch
ara
c
t
e
ris
tics
of Nanowi
r
e Trig
at
e M
O
S
F
ET Cons
iderin
g
Wave Function
Penetration,”
International
Journal of Electrical
and Computer Engineering (
I
JECE)
,
Vol.
2,
No.6,
pp. 785~791, December 2012
.
[2]
R. Arghavani
, et
al.
, "High-k/Metal Gates and Prepar
e for
High-Volume Manufactur
ing
,
"
Semiconducto
r
International, Reed
Elsevier
In
c.,
2007
.
[3]
W
oo W
e
i Kai,
et al
., “Variable Bod
y
Biasing (
V
BB) based VLSI
Design Ap
proach to Redu
ce Static Power,”
International Jo
urnal of
Electrical and Computer
Engineering (
I
JECE)
, Vol. 7,
No. 6, pp
. 3010
~3019, December
2017
[4]
Hak Kee Jung
, e
t
al
., “The Impact of Tunneling o
n
the Subthresho
ld Swing in Sub-20 nm Asy
mmetric Double G
a
te
MOSF
ETs,”
International Journ
a
l of Electrica
l
and Computer Engineering (
I
JECE)
, Vol.6, No.6, pp. 2730~273
4,
Decem
ber 2016
.
[5]
H. Ramakrishnan,
et
al.
,
"
A
na
lys
i
ng th
e effect of
process varia
tio
n to redu
ce parametric
yield
loss,"
IEEE ICICDT
2008,
pp.171-1
75, 2008
.
[6]
K. Kuhn,
et al
., "Managing process variation
in Intel’s 45nm C
M
OS technolog
y
,
"
Intel Techno
logy Journal, Vo
l,
12(
2
)
,
pp. 93-10
9, 2008
.
[7]
F.Salahuddin
,
et
al
., "
Analyze
and Optimize th
e Silicid
e
Thickne
ss in 45nm CMOS Technolog
y Using Taguchi
Method,”
ICSE
Proc. 2010
.
[8]
Elgomati,
et al.
,
"Application o
f
Taguchi Meth
od in The Optim
ization of Process Variation f
o
r 32nm C
M
OS
Techno
log
y
,"
Australian Journa
l of Basic and
Applied
Sciences,
5(
7
)
, pp. 346-3
t
5
5
, 2011
.
[9]
Husam
Ah
m
a
d
Elgom
ati
,
”Char
act
erizing Cobal
t
Silicid
e
and G
a
te Dielectri
c
Th
ickness in 65nm
NMOS Device,”
Master of
Scien
c
e Th
esis, Univ
er
siti Keb
a
ngsaan
Mala
y
s
ia
,
2007
.
[10]
Belal Ahmed Hameda, “Opt
imization of p
M
OS 65n
m Using Tagu
chi
M
e
thod,” Master
of Science Th
esis,
Universiti Kebangs
aan Malay
s
ia, 2007
.
[11]
N.
Atan,
et al
.,
"
I
nf
luen
ce
of Optimizationof Process
Parame
ter ob Threshold Voltag
e
for
Developmen
t
of
HfO2/TiSi2 18nm PMOS,”
MATEC Web Conf & 2
nd
Internatio
nal Conferen
ce
on Green Design and Manufacturer
2016 (IConGDM 2016),
Vol. 7
8
, 2016
.
[12]
ITRS. 2012, “ITRS Repor
t. www.ITRS2012.net”.
[13]
Douglas C
et
al
., ”Design
and An
aly
s
is of
Exper
iments,”
John Wiley
&
Son
s Inc.,
6
th
Edition
2005.
[14]
Ugur Esme, “Application o
f
Tag
u
chi of
Tagu
chi
Method for
th
e Optimizatio
n
of resistance
Spot Welding
Process
,
”
The Arabian
Jou
r
nal for science
and Engin
eering
,
Vo
l. 34,
No.2B, 2009.
BIOGRAP
HI
ES
OF AUTH
ORS
Noran
i
Bte Atan
receiv
e
d h
e
r B.Eng
.
(Hons
) in El
ectr
i
c
a
l
Engine
ering f
r
om
Univers
i
t
y
Teknolog
y
Mar
a
UITM) and
MSc in Electrical
and Electron
i
c
Eng
i
neer
ing from
Universiti
Tenag
a
Nasional (UNITEN) in 1995 and 200
8 respectiv
ely
.
She has 7 y
e
ar
s of working
experience in the industries as
a R&D design se
nior engineer at Matsushita Telev
i
sion and
Network (M) Sd
n. Bhd. Shah Alam, Selangor, Mala
y
s
ia. She has involves in
electrical design and
radio fr
equen
c
y tuning
for
co
lour telev
i
si
on
production fo
r intern
ational global
exports.
Responsible for
the qu
ality
and
cost redu
ction
d
e
sign for Pan
a
sonic
TVs brand.
She is curr
ently
working at College of
Engineerin
g, Universiti
Tenaga Nasion
al
(
U
niten),
M
a
l
a
ya
s
i
a as
a l
e
c
t
urer
.
Her rese
arch
int
e
rests in
clude
n
a
no dev
i
c
e
in
te
gration
and
dev
i
ce
m
odelling
.
She is curr
ent
l
y
doing Ph.D.
deg
r
ee wi
th Univ
ersiti K
e
bangsaa
n
Mala
y
s
ia
(UKM), Mal
a
y
s
ia
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
SN
:
2
502
-47
52
I
ndo
n
e
sian
J Elec Eng
& Com
p
Sci, V
o
l. 10
,
No
.
3
,
Jun
e
2
018
:
93
4 – 94
2
94
2
Burhanuddin Yeop Majlis is a pr
ofessor of
m
i
croelect
roni
cs and now the Directo
r
of Institute of
Microengin
eerin
g and Nanoelectronics(IMEN).
He
rec
e
ived
the
P
h
.D degree fro
m
Univers
i
t
y
o
f
Durham, UK in
1988. He r
e
ceiv
ed the M.Sc. degr
ee from University
of Wales,
UK in 1980, an
d
B.Sc.(Hons) fro
m
UKM in 1979. He
is senior m
e
m
b
er
of th
e Ins
titut
i
on of
Electr
ical,
Electron
ics
Engine
ers (S
MIEEE)
and fel
l
o
w
of Mala
y
s
i
a
n
S
o
lid S
t
ate S
c
ienc
e and T
ech
nolog
y
S
o
c
e
i
t
y
(FMASS).
Ibrahim Ahma
d
receiv
e
d the
B.S
c
. degr
ee
in
P
h
y
s
ics in 198
0 from
Universiti Keba
ngsaan
M
a
la
y
s
ia
(UKM
). He r
e
c
e
ived
t
h
e M
.
S
c
.
degr
ee
in Nucl
ear S
c
ie
nce
and Anal
yt
i
cal P
h
ys
ics
from
UKM and University
of Wales
respectively
,
in
y
e
ar of 1991
an
d 1992. He r
e
ceived th
e Ph.D.
degree in El
ec
tri
cal
, Ele
c
tron
ic a
nd Sy
st
em
E
ngineering from
Universiti Keb
a
ngsaan Mala
ysi
a
in
2007. He was a Nuclear Science
Officer at Nuclear
Energ
y
Unit (
M
INT) in charg
e
of Radio scope
production
for
medical and
in
dustr
y
from198
7 to 1992
. Fro
m
1993 to 199
6, he worked o
n
Sem
i
conductor
Techno
log
y
Div
i
sion at Malay
s
i
a
n in
stitu
te for
Microel
ectron
i
cs
Research Cen
t
er
& S
y
s
t
em
(M
IM
OS
), Kuala
L
u
m
pur. He jo
in
ed th
e Dep
a
rtm
e
nt of
El
ec
tric
al
, E
l
ec
troni
c an
d
S
y
stem Engineering, University
Kebangsaan Malay
s
ia (UKM) as a lecturer in 19
97 to 2002,and
as Associate Pr
ofessor from 20
02 to 2007. He i
nvolved in several mana
g
e
ment and technical
positions with
MINT, MIMOS, Em
isis Sm
artkom
S
dn. Bhd. K.Lumpur, Bumi
Hibiy
a
Sdn. Bh
d.
K.Lumpur and
UKM. He is currently
a Prof
esso
r with th
e
Department of
Electronics and
Com
m
unication
Engine
ering
,
Universiti Tena
ga
Nasional
,
M
a
la
ysi
a
. He pub
lished over 15
0
res
earch
pap
e
rs
in J
ournals
and
conferen
ces
.
He
is
a
s
e
nior
m
e
m
b
er of th
e Ins
t
itute
of
El
ectr
i
ca
l
and Ele
c
troni
cs Engine
ers (S
enior MIEEE); Mem
b
er of Institute of P
h
y
s
i
c
s Mala
ysi
a
(MIP
M)
a
nd Me
mbe
r
of
Ma
lay
s
ia
n Assoc
i
ati
on of Solid
Sta
t
e
s
Scie
nce
(MAS
S).
K. H. Chong
, gr
aduated with
B.
Eng (Hons) in
Electron
i
cs and
Electr
i
ca
l, M. Sc and PhD in
Electronic from
University
Putr
a Malay
s
ia
in
y
ear 2000, 2002
and
2008. His curr
ent research
inter
e
sts inc
l
ude
Artific
ial
Int
e
ll
ig
ent,
Evolu
tion
a
r
y
El
ectron
i
c
,
Ind
u
strial
Process C
ontrol
and
Autom
a
tic Con
t
r
o
l S
y
st
em
.
Evaluation Warning : The document was created with Spire.PDF for Python.