MOS Capacitor
MOS Capacitor
Capacitor
Şenol Mutlu
Invention of the Field-Effect
Transistor. The first patent for
the field-effect transistor
principle was filed in Canada
by Austrian-Hungarian
physicist Julius Edgar
Lilienfeld on October 22,
1925. In 1934 German
physicist Dr. Oskar Heil
patented another field-effect
transistor. A working MOSFET
was not demonstrated until
1955. It started to dominate
integrated circuits after 1970s.
E0
W, depletion region
n+ poly-Si 3.1 eV
VG SiO EC
2
+ EG
_ qfF E
Ei
EC=EFM FS
EV
p-type Si
EV
VG = 0
4.8 eV
EV
Semiconductors and Electronic Devices 9
Flat-Band Condition
EC
EC=EFM
n+ poly-Si EC
VG SiO 𝑞 𝑉𝐹𝐵 0.56 eV
2 qfF EEi
FS
+ EV EV
_
p-type Si
VG = VFB ≈ -(0.56 + ϕF) V
EV
𝑞 f𝐹 ≡ 𝐸𝑖 (𝑏𝑢𝑙𝑘) − 𝐸𝐹
• p-type Si:
Ec
𝑘𝑇
f𝐹 = ln( 𝑁𝐴 /𝑛𝑖 ) > 0 qfF
Ei
𝑞 EF
Ev
Ec
• n-type Si: EF
𝑘𝑇 |qfF|
f𝐹 = − ln( 𝑁𝐷 /𝑛𝑖 ) < 0 Ei
𝑞 Ev
GATE Ev
|qVG |
|qfS| is small, 0
- - - - - -
+ + + + + +
VG + Ec
_
p-type Si 4.8 eV EFS
Ev
𝑉𝐺 ≅ 𝑉𝐹𝐵 + 𝑉𝑜𝑥
Mobile carriers (holes) accumulate at Si surface
EE130, UC Berkeley
Semiconductors and Electronic Devices 15
Accumulation Layer Charge Density
GATE
- - - - - -
Ε𝑜𝑥 = −𝑄𝑎𝑐𝑐 /ε𝑆𝑖𝑂2
xo
+ + + + + + 𝑉𝑜𝑥 = Ε𝑜𝑥 𝑥𝑜 = −𝑄𝑎𝑐𝑐 /𝐶𝑜𝑥
VG +
_ Qacc (C/cm2) where 𝐶𝑜𝑥 ≡ ε𝑆𝑖𝑂2 /𝑥𝑜
p-type Si F/cm2 (units: F/cm2)
𝜌 ≅ −𝑞𝑁𝐴 (0 ≤ 𝑥 ≤ 𝑊)
𝑑Ε ρ 𝑞𝑁𝐴
• Poisson’s equation: = ≅− (0 ≤ 𝑥 ≤ 𝑊)
𝑑𝑥 ε𝑆𝑖 ε𝑆𝑖
• Integrate twice, to obtain fS:
2𝑞𝑁𝐴 𝜀𝑠𝑖 f𝑆
𝑉𝐺 = 𝑉𝐹𝐵 + f𝑆 + 𝑉𝑜𝑥 = 𝑉𝐹𝐵 + f𝑆 +
𝐶𝑜𝑥
2𝑞𝑁𝐴 𝜀𝑠𝑖 f𝑆
𝑉𝐺 = 𝑉𝐹𝐵 + f𝑆 +
𝐶𝑜𝑥
2
𝑞𝑁𝐴 𝜀𝑠𝑖 2𝐶𝑜𝑥 2 (𝑉𝐺 − 𝑉𝐹𝐵 )
f𝑆 = 1+ −1
2𝐶𝑜𝑥 2 𝑞𝑁𝐴 𝜀𝑠𝑖
⇒ f𝑆 = 2f𝐹
𝐸𝑖 (𝑏𝑢𝑙𝑘) − 𝐸𝑖 (𝑠𝑢𝑟𝑓𝑎𝑐𝑒) = 2 𝐸𝑖 (𝑏𝑢𝑙𝑘) − 𝐸𝐹
𝐸𝑖 (𝑠𝑢𝑟𝑓𝑎𝑐𝑒) − 𝐸𝐹 = − 𝐸𝑖 (𝑏𝑢𝑙𝑘) − 𝐸𝐹
⇒ 𝑛𝑠𝑢𝑟𝑓𝑎𝑐𝑒 = 𝑁𝐴
qfF EFS
2𝜀𝑆𝑖 (2f𝐹 ) qfs Ev
qVG
𝑊 = 𝑊𝑇 =
𝑞𝑁𝐴 Ec= EFM
Ev
EE130, UC Berkeley
Semiconductors and Electronic Devices 22
Threshold Voltage
WT r(x)
GATE M O S
+ + + + + +
- - - - - -
VG + x
_
p-type Si
f𝑆 ≅ 2f𝐹
Significant density of mobile electrons at
surface 2𝜀𝑠𝑖 (2f𝐹 )
(surface is n-type)
𝑊 ≅ 𝑊T =
𝑞𝑁𝐴
Semiconductors and Electronic Devices 25
Inversion Layer Charge Density (p-type Si)
𝑉𝐺 = 𝑉𝐹𝐵 + f𝑆 + 𝑉𝑜𝑥
(𝑄𝑑𝑒𝑝 + 𝑄𝑖𝑛𝑣 )
= 𝑉𝐹𝐵 + 2f𝐹 −
𝐶𝑜𝑥
2𝑞𝑁𝐴 𝜀𝑠 (2f𝐹 ) 𝑄𝑖𝑛𝑣
= 𝑉𝐹𝐵 + 2f𝐹 + −
𝐶𝑜𝑥 𝐶𝑜𝑥
𝑄𝑖𝑛𝑣
= 𝑉𝑇 −
𝐶𝑜𝑥
2fF
f S: 𝑞𝑁𝐴 𝜀𝑠𝑖 2
2𝐶𝑜𝑥 (𝑉𝐺 − 𝑉𝐹𝐵 )
2
2εSi (2fF )
WT =
qNA
W: 2𝜀𝑆𝑖 f𝑆 𝜀𝑆𝑖 2𝐶𝑜𝑥 2 (𝑉𝐺 − 𝑉𝐹𝐵 )
𝑊= = 1+ −1 (for 𝑉𝐹𝐵 < 𝑉𝐺 < 𝑉𝑇 )
𝑞𝑁𝐴 𝐶𝑜𝑥 𝑞𝑁𝐴 𝜀𝑠𝑖
0 VG
accumulation depletion inversion
VFB VT
EE130, UC Berkeley
Semiconductors and Electronic Devices 27
Total Charge Density in Si, Qs
Q acc = −𝐶𝑜𝑥 (𝑉𝐺 − 𝑉𝐹𝐵 )
depletion inversion
0
accumulation
VG 𝑄𝑠 = 𝑄𝑎𝑐𝑐 + 𝑄𝑑𝑒𝑝 + 𝑄𝑖𝑛𝑣
VFB VT
VFB VT
Q dep = −𝑞𝑁𝐴 𝑊 VG
0 VFB VT
accumulation depletion inversion
0 VG
Qinv
VFB VT
slope = -Cox
VG
𝑑𝑄𝑠
𝐶=
VFB VT 𝑑𝑉𝐺
Qinv C
slope = -Cox
Cox
VG
VFB VT
accumulation depletion inversion
+Q
𝑑𝑄𝑎𝑐𝑐
𝐶= = 𝐶𝑜𝑥
-Q 𝑑𝑉𝐺
−DQ
Cox
+ + + + + + Cdep W
p-type Si p-type Si
Case 1 Case 2
Inversion: gate gate
Cox Cox
N+ - - - - - - DC - - - - - -
Cdep,min
-
AC
DC and AC WT WT
p-type Si p-type Si
p-type Si: C
Cmax=Cox
CFB
Cmin
VG
accumulation depletion inversion
VFB VT
The quasi-static C-V characteristic is obtained by slowly ramping the
gate voltage (< 0.1 V/s), while measuring the gate current IG with a
very sensitive DC ammeter. C is calculated from IG = C·(dVG/dt)
Cmin
VG
VFB VT
VG
VFB VT
VG
VFB VT
• At the interface:
– Fixed charge QF
• Excess Si (?)
– Trapped charge QIT
• Dangling bonds
M O S
qQF / Cox
3.1 eV
Ec= EFM
|qVFB | QF
VFB = f MS
Ev Ec
EFS
−
Ev
Cox
4.8 eV
𝑥𝑜
–0.15V
𝑉𝐹𝐵 = f𝑀𝑆 − 𝑄𝐹
𝜀𝑆𝑖𝑂2
–0.3V
EE130, UC Berkeley
Semiconductors and Electronic Devices 46
Mobile Ions
• Odd shifts in C-V characteristics were once a mystery:
QM
DVFB =−
Cox