TELKOM
NIKA
, Vol.13, No
.2, June 20
15
, pp. 451 ~ 4
5
9
ISSN: 1693-6
930,
accredited
A
by DIKTI, De
cree No: 58/DIK
T
I/Kep/2013
DOI
:
10.12928/TELKOMNIKA.v13i2.1298
451
Re
cei
v
ed
De
cem
ber 7, 20
14; Re
vised
Ma
rch 6, 201
5; Acce
pted
March 27, 20
15
Effect of Underlap and Its Soft Error Performance in
30 nm Junctionless Based 6T- SRAM Cell
P Chitra*, V N Ram
a
krish
n
an
Schoo
l of Elelc
t
ronics Eng
i
ne
erin
g, VIT
Univ
ersit
y
, Vel
l
or
e, 632
01
4,
T
a
milnad
u, India
*Corres
p
o
ndi
n
g
author, e-ma
i
l
: chitra.p@vit.ac.in
A
b
st
r
a
ct
As CMOS devi
c
e is scali
ng d
o
w
n
signific
antl
y
, the s
ensitivit
y of Integrated
Cir
cuits (ICs) to Single
Event Upset (
SEU) radi
ation
increas
es.
T
he Sensitiv
ity of ICs to soft errors emerg
e
a
s
relia
bil
i
ty threat
w
h
ich motivate
s signific
ant i
n
terest in the
de
velo
p
m
ent of v
a
rio
u
s techn
i
q
u
e
s both
at the
devic
e an
d circ
uit
level for SE
U
hardn
ess in
SRAM me
mori
es. To fac
ilitat
e the scal
i
n
g
the conc
ept of
und
erla
p Gate-
Source/Dr
a
in
(
G
-S/D) w
a
s su
ggeste
d
in th
e
liter
ature.
L
un
is o
ne
of the
s
ensitiv
e g
e
o
m
etrical
par
a
m
et
er
consi
dere
d
to
differ fro
m
1
n
m
to
5 n
m
in
a
SEU ra
diati
n
g
envir
on
me
nt. T
he effect of
Gate-Source/D
rai
n
und
erla
p (L
un
) on soft error p
e
rformanc
e in
30 n
m
Ju
nctio
n
less T
r
ans
istor (JLT
) base
d
on 6T
-SRAM
cell
has be
en exa
m
i
n
e
d
throug
h extensiv
e mix
e
d mo
de-
devic
e
and circuit si
mulati
ons usi
ng
T
C
AD. T
he critica
l
dose
obs
erve
d
in J
L
T
bas
ed
6T
-SRAM w
i
th
L
un
rangi
ng
fro
m
1 n
m
t
o
5
n
m
to
flip
the c
e
l
l
is
give
n by
Li
near
Energy T
r
ansf
e
r (LET
) between 0.0
5
to 0.06 pC/
µ
m.
T
h
e simulati
on re
sult ana
ly
z
e
s
electric
al an
d SEU
radi
ation p
a
ra
meters to stud
y its impact o
n
JLT
based 6T
-
S
RAM cell.
Ke
y
w
ords
: Ju
nction
less 6T
-
S
RAM, Under
l
ap, SEU Rad
i
a
t
ion, LET
,
Hea
vyIon, T
C
AD Simulati
on
1. Introduc
tion
Scaling of CMOS device
s
are mainly sufferi
ng from
Short Ch
ann
el Effect (SCEs) an
d
Drai
n
Indu
ce
d
Barrier Lo
wering (DIBL). The sen
s
itiv
ity of ICs to S
E
U ra
diation,
comin
g
fro
m
the
natural
sp
ace or p
r
e
s
ent
in the terre
s
trial e
n
viron
m
ent is g
o
in
g up [1] i.e. the amou
nt of
minimum cha
r
ge (kn
o
wn a
s
criti
c
al charge) requi
red t
o
flip the memory cell (S
RAM) decre
ases
with
scaling
[2]. This
e
v
ent is
kno
w
n
as soft
error. In
crea
sing
de
sign
and
fabri
c
a
t
ion
compl
e
xities, smalle
r feature sizes, lo
we
r volt
age and
higher
cu
rre
nt levels, intrinsi
c paramet
e
r
fluctuation
s
,
highe
r o
perating fre
que
nci
e
s
are
al
so
proje
c
ted
to
cau
s
e
an i
n
crea
se i
n
the
soft
error failure rate in s
ub-90 nm ICs
[3].
Soft error stu
d
ies of dee
p sub
-
mi
cro
n
MOSFET-b
ased
6T-SRA
Ms a
r
e dealt in det
ail [4],[5]. IT
RS roa
d
map i
ndicates that
semi
con
d
u
c
tor memo
rie
s
will
occupy m
a
jo
r portio
n
of
chip a
r
ea
in t
he n
ear future technol
ogie
s
[6]. SEUs
have al
so
be
en
analyzed a
s
an imp
o
rtant
reliability fa
ctor in
the
dev
elopme
n
t of
semi
con
d
u
c
tor m
e
mo
rie
s
[7].
Semico
ndu
ct
or memo
rie
s
such as SRAM is by far the dominan
t form of embedd
ed mem
o
ry
found in tod
a
y
’s ICs
occu
p
y
ing as m
u
ch
as 6
0
-7
0%
o
f
the total chi
p
are
a
an
d a
bout 75%
-85
%
o
f
t
he t
r
an
si
st
or
cou
n
t
in
so
me I
C
pr
o
duc
ts
[8
]. T
h
e
mo
s
t
c
o
mmon
ly u
s
ed
6
T
-SR
A
M me
mor
y
c
e
ll
desi
gn uses six
tran
si
stors
to sto
r
e a bit.
T
he m
e
th
od to
pro
d
u
c
e lo
w p
o
wer
6T-SRA
M de
sign
was
disc
uss
e
d in [9].
Device level cha
r
a
c
teri
stics of Doubl
e Gate
Vertical (DGV) MOS
F
ET with and
without
SOI stru
ctures a
r
e exa
m
ined [10].
Since, t
he
CMOS
scali
ng technolo
g
y
bring
s
in
many
unconventio
n
a
l device
s
especi
a
lly FinFET, it is nec
essary to stud
y the radiatio
n effect on these
device
s
and circuits. The
transi
ent re
spon
se of
pla
nar Do
uble
-
G
a
te
an
d
Fin
F
ET
to
he
avy
ion
irra
diation h
a
s
be
en
studi
ed by 3-D n
u
meri
cal
sim
u
lation in [1
1]. FinFET-b
ase
d
6T
-SRAM
circuits are studied
using
TCAD simulati
ons
in
[12],[13]. When these devi
c
es are scaled down to
extreme dim
ensi
o
n
s
, the formation
of ultra-sh
arp j
u
nction
s b
e
tween
sou
r
ce/d
rain a
nd
cha
nnel
become
s
co
mplex si
nce t
he d
oping
co
nce
n
tration
h
a
s to
vary by
seve
ral
ord
e
r
s
of ma
gnitu
des
over a di
stan
ce
of a few n
anomete
r
s. Theref
o
r
e, hig
h
ly accurate
dopin
g
tech
ni
que
s and
ultrafast
dopa
nt activation processe
s are
re
qu
ired to avoid
the lateral
diffusion of
source a
nd d
r
ain
impuritie
s int
o
the cha
nnel
region.
Based
on Lili
enfeld’
s first
transi
s
to
r architec
ture a solution to thi
s
problem
(L
ilienfeld
1930
) ve
ry rece
ntly, a Ju
nctionl
ess Fi
nFET intr
odu
ced
by
Colin
ge et. al. [1
4] ha
s attracted
device
co
mm
unity. The SEU radiation
a
nalysi
s
of
a j
unctio
n
le
ss
d
e
vice i
s
studi
ed by Mu
nte
anu
et. al [15]. So
ft error
study of Junctionless Fi
nF
ET based SRAM i
s
studied in [16]. To facilitate
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 13, No. 2, June 20
15 : 451 – 45
9
452
the scalin
g
phen
omen
a
the con
c
ept
of Gate
-So
u
rce/Drain
u
nderl
ap
(L
un
) with lo
w-do
ped
cha
nnel wa
s sug
g
e
s
ted
[1
7].
L
un
is one
of the sen
s
it
ive geometri
cal paramete
r
con
s
ide
r
e
d
to
differ fro
m
1
nm to
5 nm
i
n
a SE
U
radi
ating e
n
viron
m
ent. In thi
s
pape
r, we h
a
v
e studi
ed th
e
Effec
t
of L
un
on SEU/soft error
perfo
rm
ance of tri-ga
te Jnle
ss 6T
-SRAM cell. L
un
has b
een t
a
ke
n
as a
pa
ramet
e
r u
nde
r ou
r
control. Mini
mum radiat
io
n do
se
req
u
ired to flip the
cell i
s
foun
d
out
by doing the tran
sient si
mu
lations u
s
in
g TCAD d
e
vice
simulato
r.
This p
ape
r i
s
organi
ze
d as follo
ws: S
e
ction II discusse
s
abo
ut Jun
c
tionle
ss device
construction and cali
bration.
Se
ction III talks about
Junctionless
6T-SRAM
structure
simulati
on
and S
R
AM o
peratio
n. Se
ction IV explai
ns SE
U
radia
t
ion ph
enom
e
na o
n
Jun
c
tio
n
less
6T-S
RAM
and provide
s
the simulatio
n
results an
d discu
ssi
on
s. Section V pro
v
ides the con
c
lu
sion.
2. Dev
i
ce Constru
c
tion a
nd Calibra
tion
The vario
u
s
device p
a
ra
meters are sho
w
n
in th
e schem
atic diagra
m
of Figure 1
.
Sentaurus T
C
AD
simulat
o
r from Syn
opsy
s
is u
s
e
d
in this
wo
rk. Th
e sim
u
lator ha
s ma
ny
feature
s
a
nd
module
s
use
d
in o
u
r
simu
lation a
r
e:
Se
ntauru
s
Structure Edito
r
(S
DE) i
s
u
s
e
d
t
o
cre
a
te individ
ual devi
c
e
structure. Com
m
on
T
r
iple
-Gate-NMOS
and PMOS
JLT is
created
as
sho
w
n in Fig
u
re 2 an
d Fig
u
re 3 p
r
ior to
SRAM simul
a
tion. Durin
g
d
e
vice sim
u
lati
on, the mobili
ty
model in
clu
d
e
s d
oping
de
pend
en
cy, high-field
satu
ration and t
r
a
n
sverse fiel
d
(i.e. gate fie
l
d)
depe
nden
cy etc.
Figure 1. Schematic vie
w
of 2-D
Jun
c
tionle
ss FET
Figure 2. Co
mmon Tri
-
gat
e JLT
NMOS
Device
Struct
ure (Gate Oxi
de is re
move
d)
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
1693-6
930
Effec
t
of Underlap and its
Soft Error Performanc
e
in 30 nm J
unc
tionless
bas
e
d ....
(P Chitra)
453
Figure 3. Co
mmon Tri
-
gat
e JLT PMOS
Device
Struct
ure (Gate Oxi
de is re
move
d)
Sentaurus Device (SDEVI
CE) Simulato
r is use
d
to s
i
mulate trans
i
ent c
u
rves
; Ins
p
ec
t is
use
d
to view the re
sults [
18]. An I
D
-V
G
cha
r
a
c
teri
sti
c
of NM
OS
device i
s
sho
w
n in Fi
gure
4
.
Supply Volta
ge (V
dd
) u
s
e
d
in thi
s
stud
y is 1
V. Tab
l
e 1
gives va
riou
s d
e
vice
dimen
s
ion
s
a
nd
dopin
g
value
s
.
Figure 4. Co
mmon Tri
-
gat
e JLT
NMOS
Device I
D
-V
G
Simulation
Table 1. Devi
ce Dim
e
n
s
ion
s
Parameter Value
Gate Le
ngth(L
g
)
30
nm
Gate-
o
xide thickness (T
ox
)
2
nm
Fin
w
i
d
t
h (W )
5 nm
Fin height (H
)
5 nm
Underlap (
Lu
n
)
1 to 5 nm
Channel doping
8× 10
19
/cm
3
(N transistor)
2.5× 10
19
/cm
3
(P
transistor)
Source-dr
a
in dop
ing
8 × 10
19
/cm
3
( N t
r
ans
i
s
tor)
2.5 × 10
19
/cm
3
( P
trans
i
s
tor)
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 13, No. 2, June 20
15 : 451 – 45
9
454
3. Junctionle
ss 6T SRAM
Structure
JLT i
s
a multigate FET wi
th neither N
+
P nor P
+
N ju
nction
s. Unli
ke a MOSFET, for a
JLFET
ch
ann
el the do
ping
con
c
e
n
tration
and type
are
equal to
(N
+
N
+
N
+
) i
n
the
source
- chan
n
e
l-
drain
regi
on.
Since this d
e
vice ha
s no
con
c
ent
ra
tio
n
gradi
ent in
the lateral
dire
ction of the
con
d
u
c
tion la
yer, it does n
o
t have any p–n jun
c
tion.
The dopi
ng
con
c
e
n
tration
used in
JLT
is
typically in th
e ra
nge
of 1
0
19
cm
-3
, unif
o
rm, an
d h
o
m
ogen
ou
s a
c
ross th
e
sou
r
ce
(S),
chan
nel,
and drain (D)
regio
n
.
A JLT
6T-S
RAM cell i
s
desi
gne
d by
repl
aci
ng th
e co
nvention
a
l MOSFET
s or SOI
(sili
con
-
O
n
-In
s
ulato
r) Fi
nF
ET based mu
lti-gate tra
n
si
st
or
s
wit
h
jnle
ss t
r
an
sisto
r
s. The gen
erated
6T-SRA
M structure fro
m
S
D
E is
sh
own i
n
Figu
re
5. M
e
shi
ng i
s
sho
w
n o
n
on
e-si
de of the
devi
c
e
namely (N2, ACC2, P2 transi
s
tors) in Figure 5.
In this ne
w de
sign app
roa
c
h
,
trigate jnless
device provides gate controllability
on three
sides. Using a trigat
e device
architecture i
t
is
possibl
e to turn the device on and off to obtai
n MOSF
ET- like ele
c
trical characte
ristics.
Figure 5. Co
mmon Tri
-
gat
e
JLT 6T
-SR
A
M Structur
e
In this wo
rk,
JLT ba
se
d 6T-SRA
M is
subj
ecte
d to heavy ion ba
sed SEU
rad
i
ation.
Minimum
radi
ation do
se
re
quire
d to flip
the cell
is fou
nd out
by doi
ng the t
r
an
si
ent sim
u
latio
n
s
with ra
diation
model
s turne
d
-on i
n
TCA
D
devic
e si
mul
a
tor. Figure 6 sho
w
s 6T
-SRAM ope
rat
i
on
simulatio
n
cu
rve befo
r
e S
E
U ra
diation.
Mixed
mod
e
simul
a
tion
approa
ch i
s
use
d
in SRA
M
simulatio
n
. In mixed mode simulatio
n
so
me portio
n
of the circuit ca
n be simul
a
te
d at the device
level and so
me part of the ci
rcuit ca
n use
com
p
a
c
t model
s. In this study, interconn
ect
s
are
assume
d to
be p
e
rfe
c
t int
e
rconn
ect
s
.
The
details o
f
rise a
nd fall
time, pul
se
width
of data
and
acce
ss p
u
lse
s
used in SRAM simulatio
n
are given in
Table 2.
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
1693-6
930
Effec
t
of Underlap and its
Soft Error Performanc
e
in 30 nm J
unc
tionless
bas
e
d ....
(P Chitra)
455
Figure 6. 6T-SRAM simul
a
tion
Table 2. Data
and Access
Pulse Timin
g
s
nbit(data)
n(
w
o
r
d
)
Rise Time(pS)
Fall Time(pS)
Pulse Width(pS)
Rise Time(pS)
Fall Time(pS)
Pulse Width(pS)
10
10 500 10
10
250
4. SEU Radi
ation Effec
t
s
and Res
u
lts
The
simulatio
n
of SEU
ca
use
d
by a
he
avy ion impa
ct is
activate
d by u
s
ing th
e prop
e
r
keyword
(‘
He
avyIon’) in th
e phy
sics
se
ction
of
SDE
V
ICE. The
chara
c
te
risti
c
s of the h
eavy ion
like di
re
ction,
cha
r
a
c
teri
sti
c
radiu
s
, do
se val
ue o
r
Li
near Energy
Tran
sfe
r
(LE
T
in p
C
/µm)
and
stri
ke lo
cation
can
be
spe
c
i
f
ied. The he
a
vy ion stri
kes
at locatio
n
wit
h
dire
ction
of motion of ion
s
from top to b
o
ttom along v
e
rtical
axis. L
ength of the i
on tra
ck i
s
0.
035
μ
m, cha
r
acteri
stic
ra
di
us
w
t
=
0.
0
1
μ
m.
He
avy ion
b
o
mba
r
dme
n
t
is initiate
d in
the sim
u
latio
n
du
ring
no
n-acce
ss pe
rio
d
at
the d
r
ain
re
gi
on
whi
c
h
ha
s
logic value
‘1’
.
Figure
7
sho
w
s no
de volta
ges after he
a
vy ion st
rike a
t
t=275
ps for
LET =
0.04 p
C
/ µm. By properly
cho
o
si
ng LET valu
e
the state
of the cell
ca
n
be
flipped a
s
sh
own in Fig
u
re
8. The minimum req
u
ire
d
LET value to flip the cell is of our inte
rest.
In Table
3, th
e L
un
Vs
LET
is
sho
w
n fo
r
1 to 5
nm. Th
e LET valu
e
differen
c
e
bet
wee
n
L
un
1 to
5
nm is only on
e–fold mag
n
itude differe
nce due to less
variation in th
e cha
nnel resistan
ce [19]
.
Figure 7. Nod
e
Voltages aft
e
r He
avy Ion
Strike at 275
ps for LET
=
0
.
04 pC/µm
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 13, No. 2, June 20
15 : 451 – 45
9
456
Figure 8.
Nod
e
Voltages aft
e
r He
avy Ion Strike at 275
ps (LE
T
= 0.0
5pC/µm
)
Figure 9
an
d
10
sh
ows th
e he
avy ion
g
eneration, h
e
a
vy ion
cha
r
g
e
de
nsity a
n
d
ele
c
tro
-
static pote
n
tial distrib
u
tion
acro
ss the tri-g
a
te
J
n
less 6T-
S
R
A
M s
t
r
u
c
t
ure w
h
ic
h sw
itches
its
state at dose/LET value of 0.05 pC/µm f
o
r L
un
= 1, 2 nm and 0.06
pC/µm for L
un
= 3, 4, 5 n
m
.
There are variou
s p
a
ra
meters like
electron a
n
d
hole de
nsiti
e
s, ele
c
tro
n
and hol
e current
den
sities,
ele
c
tro
-
stati
c
p
o
t
ential, SRH
recombi
natio
n rate
etc.
can b
e
an
alyzed to
study t
h
e
SEU ra
diatio
n effect
s. Fig
u
re
11
( L
un
=1,2
nm) a
n
d
Figu
re
12
(L
un
=
3,4 & 5
nm)
sho
w
s
SRH
recombi
natio
n
rate acro
ss the
structu
r
e at
three
different time in
sta
n
ts, i.e. @2
5
0
pS, @275
pS,
and
@33
0
p
S
. These tim
e
instant
s co
rre
sp
ond to
pre, pe
ak a
n
d
post radiat
ions. It can
be
observed
fro
m
Figu
re
11
and
12
that
SRH
re
com
b
ination
rate
is very lo
w
for p
r
e
and
post
radiatio
ns.
Figure 9. 2D
profile of He
a
vyIonGeneration, Hea
vyIon
Cha
r
ge
De
nsit
y, Electrostati
c Potential of
Tri-g
a
te JL
T 6T-SRA
M structure at
275
pS (LET= 0.05 pC/µm for
L
UN
= 1 and 2
nm)
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
1693-6
930
Effec
t
of Underlap and its
Soft Error Performanc
e
in 30 nm J
unc
tionless
bas
e
d ....
(P Chitra)
457
Figure 10. 2D profile of He
avyIonGene
ration, H
eavyIonCharge
De
nsity, Electro
s
taticPotentia
l of
Tri-g
a
te JL
T 6T-SRA
M structure at
275
pS (LET= 0.0
6
pC/µm for L
UN
= 3, 4 & 5
nm)
Table 3. Lun
Vs LET value
L
un
(nm
)
LET ( pC/µm
)
1 0.05
2 0.05
3 0.06
4 0.06
5 0.06
As we in
crea
se the L
un
fro
m
1 to 5 nm the HeavyIon
Gene
ration in
cre
a
ses from
9.8e31
to 2.2e32 to f
lip the cell. T
h
is indi
cate
s
highe
r LET is requi
red fo
r i
n
crea
sing
L
un
. In the s
i
milar
way, the
HeavyIonCh
a
rgeDen
sity in
cre
a
ses a
s
L
un
in
cre
a
s
e
s
. It is
also
note
d
that
HeavyIon
Cha
r
ge
Den
s
ity increa
se
s fro
m
9.9e
20
to 1.2e
21
cm
-1
s
-1
to flip the cell. Thi
s
also
indicates hi
g
her LET i
s
re
quire
d. Electrostatic
potent
ial after the strike g
e
ts mo
dified dep
end
ing
on the ge
ne
rated he
avyion ch
arg
e
de
n
s
ity. It is obse
r
ved that it flips the
node
whe
n
it is cl
o
s
e to
1V, otherwi
se
critical
cha
r
g
e
is not suffici
ent to flip the
node.
Figure 11 an
d 12 sho
w
s 2
D
sp
atial distribution
of SRH re
com
b
inat
ion rate for L
un 1 to 5
nm. It can b
e
observed f
r
om figu
re 1
1
the re
com
b
ination rate
is lower for
Lun
=1 an
d 2
nm
comp
ared to f
i
gure
12 fo
r L
un =3,4 & 5
n
m
. At the stri
ke location, S
RH
re
com
b
in
ation value
s
are
given in
Ta
bl
e 4. T
h
is ma
ke
s n
ode
volt
age
distu
r
b
a
n
c
e i
s
lo
wer in
Lun
=3, 4
&
5 nm
comp
ared
to Lun = 1, 2 nm. Hen
c
e it requi
re
s high
er LET to flip the cell.
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 13, No. 2, June 20
15 : 451 – 45
9
458
Figure 11. Tri
-
gate JLT 6T
-SRAM SRH
Re
combi
natio
n pre
-
pe
ak-p
ost radi
ation
simulatio
n
(L
ET=
0.05 pC/µm L
UN
=1 and 2 n
m
@25
0
p
s
2
75 ps 3
30 p
s
)
Table 4. Pre
-
Peak-Post ti
me SRH
Re
combinatio
n for different Lu
n
values
Pre-peak-
post
Ti
me (ps
)
SRH Recombina
t
ion value for Lun
=
1,2 nm
SRH Recombina
t
ion value for Lun
= 3,4
& 5 nm
250 4.8e+26
5.2e+26
275 1.8e+28
4.1e+28
330 4.4e+27
4.8e+27
Figure 12. Tri
-
gate JLT 6T
-SRAM SRH
Re
combi
natio
n pre
-
pe
ak-p
ost radi
ation
simulatio
n
(L
ET=
0.06 pC/µm L
UN
=3, 4 & 5 n
m
@25
0
p
s
2
75 ps 3
30 p
s
)
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
1693-6
930
Effec
t
of Underlap and its
Soft Error Performanc
e
in 30 nm J
unc
tionless
bas
e
d ....
(P Chitra)
459
5. Conclusio
n
Tri-gate Jnless
6T-S
RAM
is sim
u
la
ted to s
t
udy the effec
t
of L
un
and inve
stig
ated their
SEU/soft erro
r perfo
rma
n
ce (the minimu
m LET value
requi
re
d to flip the cell) in
TCAD. Th
e LET
requi
re
d to flip the Jnless
6T-SRA
M is
observ
ed
and
the value is f
ound to b
e
b
e
twee
n 0.05
pC/
µm for L
UN
= 1, 2 nm an
d 0.06 p
C
/µm for L
UN
=
3, 4, 5 nm. The si
mulatio
n
re
sult anal
yzes
electri
c
al
and
SEU radi
ation pa
ram
e
ters to
stu
d
y its impact o
n
JLT ba
sed
6T
-SRAM me
m
o
ry
c
i
rc
uit.
Referen
ces
[1]
Bauma
nn
RC.
Rad
i
ati
on-Ind
u
ced
Soft Err
o
rs in
Adv
anc
ed S
e
mico
nd
u
c
tor T
e
chnolo
g
ies.
IEEE
T
r
ansactio
n
s o
n
Devic
e
Mater
i
al Re
li
abi
lity
. 2005; 5(3): 3
05-
316.
[2]
Petersen E
L
, Shapiro P, Adams JH, Burke
EA.
Calcul
atio
n of Cos
m
ic-R
ay Induc
ed Sof
t
Upsets an
d
Scali
ng in V
L
SI Device.
19
82; 29(6):
20
55
–20
63.
[3]
Kartik
M.
Simu
latio
n
of transi
ents caus
ed b
y
singl
e-ev
ent
upsets i
n
co
mbin
ation
a
l l
o
g
i
c
.
International
T
e
st Conferen
ce, 2005; 3
7
: 3.
[4]
Lia
n
li
an Z
,
Be
ckett P
.
Soft
Error Rate Estim
a
ti
on i
n
D
e
ep Su
b-Micro
n
CMOS
. 13
th
Pacific Rim
Internatio
na
l Symp
osi
u
m on
Dep
end
ab
le C
o
mputi
n
g
.
2
007
:
210-2
16.
[5]
Prasha
nt KS,
Navak
anta B. SEU Reli
ab
ilit
y Impr
ovement
Due to Sourc
e
-Side C
harg
e
Coll
ectio
n
i
n
the D
eep-S
u
b
m
icron S
R
AM
Cell.
IEEE
T
r
a
n
sactio
ns o
n
D
e
vice
an
d M
a
terials
R
e
li
abi
lit
y.
200
3; 3(
1):
14-1
7
.
[6]
http://public.itrs.net.
[7]
Palkes
h J, Juzer V, Rakes
h
K. Lal. SE
U
Rel
i
ab
ilit
y
Anal
ys
is of Advanc
ed D
e
e
p
-Subm
icro
n
T
r
ansistors – Introducti
on.
IE
EE Transactio
n
s on
Devic
e
and M
a
teria
l
s
Reli
ab
ility.
2
005
;
5(2): 28
9-
295.
[8]
Hen C, Er E, S
oh PN, L
oh SK
, W
ang Q
, T
eong
J.
F
a
il
ure A
nalysis
of 65
n
m
T
e
ch
no
logy
Nod
e
SRAM.
Internatio
na
l Symp
osi
u
m on
Ph
y
s
ica
l
and F
a
ilur
e
Ana
l
ysis of
Integrated
C
i
rcuits. 200
8: 1-4.
[9]
Lab
on
nah
F
R
, Moh
a
mmad
F
BA, Mamun
BIR, M
ohd. M
a
r
u
fuzzama
n, H
a
fizah
H. Adv
a
nces
on
Lo
w
Po
w
e
r
Designs for SRAM Cell.
T
E
LKOMN
I
KA Indon
esia
n Jour
nal
of
E
l
ectrical
Eng
i
n
eeri
ng.
2
014
;
12(8): 60
63-
60
82.
[10]
Jatmiko ES, R
a
zali
I. Desi
gn
of Do
ubl
e Ga
te
Vertical
MOSF
ET
using S
i
licon
On Ins
u
l
a
tor (SOI)
T
e
chnolog
y.
In
ternatio
nal J
o
u
r
nal of Na
no D
e
vices, Sens
or
s and Syste
m
s
(IJ- Nano).
1(1
)
: 34-38.
[1
1
]
Mu
n
t
e
a
n
u
D
,
Au
tra
n
JL
.
3
-
D
si
mul
a
tio
n
an
a
l
y
s
i
s
o
f
b
i
p
o
l
a
r
a
m
p
l
i
f
i
c
a
t
i
on i
n
p
l
a
n
a
r
d
o
u
b
l
e
-
g
a
t
e
and
F
i
nF
ET
w
i
th in
dep
en
dent ga
tes.
IEEE Trans. Nucl. Sci.
2009; 56(4): 2
083
–20
90.
[12]
Ramakris
hn
an
VN, Sriniv
asan
R. Soft Error
St
ud
y in
Do
ubl
e gate
d
F
i
nF
E
T
-Based SRAM Cells
w
i
t
h
Simulta
neo
us and Ind
e
p
e
n
d
e
n
t Driven Gate.
Microelectr
oni
cs Journa
l.
201
2;
43(11):8
88-
893.
[13]
Ramakris
hn
an
VN, Srinivasa
n
R. Effect of
und
erla
p an
d soft error performanc
e in 3
0
nm F
i
nF
ET
-
base
d
6T
-SRAM cells
w
i
th
si
multan
eous
a
n
d
in
de
pe
nde
nt
drive
n
gates.
Journ
a
l of
Co
mp
utatio
na
l
Electron
ics
. 20
13; 12(3): 4
69-
475.
[14]
Lili
enfe
l
d JE.
Method a
nd a
p
p
a
ratus for contr
o
lli
ng e
l
ectric c
u
rrent,
193
0, U
.
S Patent 1 74
5 175.
[15]
Muntea
nu D, Autran JL.
3D
Numerica
l Si
mu
lati
on of Bi
pol
ar Amplific
a
t
ion in Ju
nctio
n
less D
oub
le-
Gate MOSF
ETs und
er h
eavy
–
io
n irra
di
ation
.
12th Eur
o
p
e
a
n
confer
enc
e o
n
rad
i
ati
on a
n
d
its effect
s
on comp
on
ent
s and s
y
stems. 2011: 7
3
-76.
[16]
Ramakris
hn
an
VN, Srinivas
an
R.
3D Si
mu
lat
i
on stu
d
y of S
o
ft Error on Ju
ntion
l
ess 6T
S
R
AM.
IEEE
Internatio
na
l C
onfere
n
ce o
n
Innov
atio
ns
in
Engi
neer
in
g an
d T
e
chnolo
g
y
,
201
4.
[17]
T
r
ivedi V, F
o
ssum JG, Cho
w
dhur
y MM. Na
no
sca
le F
i
nF
E
T
s
w
i
th g
a
te-s
ource/dr
a
in
Un
derl
ap.
IEEE
T
r
ans. Electron
Devices
. 20
05
; 52(1): 56–
62.
[18]
S
y
n
ops
ys Se
nt
aurus D
e
vice
User Guid
e Ve
rsion-A 2
008.
0
9
[19]
Chi WL, Isa
b
e
lle F, Ar
ya
n A
,
Ran
Y, Nim
a
DA
, Pe
dram
R, Jea
n
PC.
Performanc
e e
s
timation
o
f
juncti
onl
ess multig
ate transist
o
rs.
Solid-Stat
e
Electronics.
201
0; 54: 97–
1
03.
Evaluation Warning : The document was created with Spire.PDF for Python.