8/13/2019 VLSI lec 3
1/24
MOS Transistor Theor
VLSI DesignVLSI Design
By Dr. Yaseer A. DurraniDept. of Electronics Engineering
University of Engineering & Technology, Taxila
Outline Semiconductor IC Chip pn-Junction Transistor I-V Characteristics RC Delay Model Transistor Families
2
8/13/2019 VLSI lec 3
2/24
Silicon Chip A pattern of interconnected switches & gates on surface of crystal of
semiconductor (typically Si) Silicon is a Group IV semiconducting material
Crystal lattice: covalent bonds hold each atom to four neighbours These switches & gates are made of:
Areas of n-type silicon Areas of p-type silicon Areas of insulator Lines of conductor (interconnects) joining areas together
, , , , , Geometry of these areas is known as layout of the chip Connections from chip to outside world are made around the edge of the chip to
facilitate connections to other devices
3
Semiconductor is a solid material that has electrical conductivity in betweenconductor & insulator Similar to insulators Less Larger energy band gaps
Doping is the process of intentionally introducing impurities into extremelypure/intrinsic semiconductor in order to change its electrical properties Number of dopant atoms needed to create a difference in ability of a
semiconductor to conduct is very small Small number of dopant atoms are added (order of 1 every 100,000,000
Semiconductor
, More dopant atoms are added (order of 1 in 10,000) then doping is said to
be heavy, or high. This is often shown as n+ for n-type dopant or p+ for p-type doping
4
8/13/2019 VLSI lec 3
3/24
8/13/2019 VLSI lec 3
4/24
p-n Junction A p-n junction is a boundary or interface b/w two types of semiconductor
material, p-type & n-type, inside a single crystal of semiconductor
pn junctions are elementary "building blocks" of most semiconductorelectronic devices such as diodes, transistors, solar cells, LEDs, ICs Common type of transistor, bipolar junction transistor, consists of two pn
junctions in series, in the form npn or pnp Its main electrical property is that it rectifies (allow current to flow easily in one
direction only)
7
p-type n-type
anode cathode
Type of junctions Homojunction: is a semiconductor interface that occurs b/w layers of similar
semiconductor material, these materials have equal band gaps but typicallyhave different doping
Heterojunction: is interface that occurs b/w two layers or regions of dissimilarcrystalline semiconductors. These semiconducting materials have unequal bandgaps as opposed to a homojunction
8
Image of nanoscale heterojunctionb/w iron oxide & cadmium sulfide
8/13/2019 VLSI lec 3
5/24
8/13/2019 VLSI lec 3
6/24
Holes diffused to metalurgical junction & combine with electrons. They leavebehind -vely charged acceptor, while electrons diffused & leave behind +velycharged donor
Diffusion process can not go on forever, because, increasing amount of fixedcharge wants to electrostatically attract the carriers that are trying to diffuseaway & equlibrium is reached
Fixed charge produce an electric field which slows down the diffusion process Fixed charge region is called depletion/space charge region
Idealized p-n Junction
11
Contact Potential When p & n-type materials are joined, electrons & holes diffuse due to their
large carrier concentration gradients at junction where holes diffuse from p-side into n-side, & electron diffuse from n to p. Thus, concentration gradientscreates a diffusion component of current from p to n region
+++-
-- -+
12
++
- -- - -
++++
8/13/2019 VLSI lec 3
7/24
Transistors as ideal switches An ON transistor passes a finite amount of current
Depends on terminal voltages Derive current-voltage (I-V) relationships
Transistor gate, source, drain all have capacitance I = C (V/t) ->t = (C/I)V Capacitance and current determine speed
13
NMOS Capacitor Gate & body form MOS capacitor Operating modes
Accumulation Depletion Inversion
polysilicon gate
(a)
silicon dioxide insulator
p-type body+-
Vg< 0
+-
0 < Vg< V
t
depletion region
(b)
(c)
+-
Vg> Vt
depletion region
inversion region
14
8/13/2019 VLSI lec 3
8/24
Terminal Voltages Mode of operation depends on Vg, Vd, Vs
Vgs = Vg Vs Vgd = Vg Vd
Vds = Vd Vs = Vgs - Vgd Source & drain are symmetric diffusion terminals
However, Vds 0 Three- regions of operation
Cutoff, Linear, & Saturation
Vg
V s V d
Vgd
Vgs
Vds
+-
+
-
+
-
NMOS Saturation RegionChannel pinches off if Vds > Vgs Vt.I inde endent of V i.e. current saturates
+-
Vgs
> Vt
n+ n+
+-
Vgd
< Vt
Vds> V
gs-V
t
p-type body
b
g
s d Ids
+-
Vgs
= 0
n+ n+
+-
Vgd
p-type bodyb
g
s d
NMOS Cutoff RegionAssume Vs = VbNo channel, if Vgs = 0Ids = 0
+-
Vgs
> Vt
n+ n+
+-
Vgd
= Vgs
+-
Vgs
> Vt
n+ n+
+-
Vgs
> Vgd
> Vt
Vds
= 0
0 < Vds
< Vgs
-Vt
p-type body
p-type bodyb
g
s d
b
g
s dIds
NMOS Linear RegionChannel forms if Vgs > Vt
No Currernt if Vds = 0
Linear Region:If Vds > 0, Current flowsfrom d to s ( e- from s to d)Ids increases linearlywith Vds if Vds > Vgs Vt.Similar to linear resistor
Similar to current source
15
I-V Characteristics In Linear region, Ids depends on
How much charge is in the channel How fast is the charge moving
Channel Charge: MOS structure looks like parallel plate capacitor whileoperating in inversion Gate oxide (dielectric) channel
Qchannel = CV C = C
g
= ox
WL/tox
= Cox
WL V = Vgc Vt = (Vgs Vds/2) Vt Cox = ox/ tox
n+ n+
p-type body
+
Vgd
gate
+ +
source
-
Vgs
-drain
Vds
channel-
Vg
Vs
Vd
Cg
n+ n+
p-type body
W
L
tox
SiO2gate oxide(good insulator, ox= 3.9)
polysilicongate
16
8/13/2019 VLSI lec 3
9/24
Carrier Velocity Carrier Velocity
Charge is carried by e- Carrier velocity vproportional to lateral E-field b/w source & drain
v= E, is mobility, E = Vds/L Time for carrier to cross channel: t=L/v NMOS Linear I-V
How much charge Qchannel is in the channel How much time teach carrier takes to cross
NMOS Saturation I-V If V < V channel inches off near drain
channel
ox 2
2
ds
dsgs t ds
dsgs t ds
QI
t
W VC V V V
L
VV V V
=
=
=
ox=
WC
L
When Vds > Vdsat = Vgs Vt
Now drain voltage no longer increases current
NMOS IV Summary
Shockley1st
order transistor models (valid for Large channel devices only
( )2
2
2
dsatds gs t dsat
gs t
VI V V V
V V
=
=
( )2
cutoff
linear
saturatio
0
2
2n
gs t
dsds gs t ds ds dsat
gs t ds dsat
V V
VI V V V V V
V V V V
<
= 17
Example For a 0.6m process
From AMI Semiconductor tox = 1 0 0 = 350 cm2/V*s Vt = 0 . 7 V
Plot Ids vs. Vds Vgs = 0 , 1 , 2 , 3 , 4 , 5 Use W/L = 4/2
1423.9 8.85 10W W W
0 1 2 3 4 50
0.5
1
1.5
2
2.5
Ids
(mA)
Vgs
= 5
Vgs
= 4
Vgs
= 3
Vgs
= 2
Vgs
= 1
8100 10
oxL L L
= = =
ds
18
8/13/2019 VLSI lec 3
10/24
PMOS I-V All dopings & voltages are inverted for PMOS Mobility p is determined by holes
Typically 2-3x lower than that of electrons n
120 cm2/V*s in AMI 0.6 m process Thus PMOS must be wider to provide same current
In this class, assume n/ p = 2
19
Gate Capacitance Any two conductors separated by an insulator have capacitance Gate to channel capacitor is very important
Creates channel charge necessary for operation Source & drain have capacitance to body
Across reverse-biased diodes Called diffusion capacitance because it is associated with source/drain
diffusion Approximate channel as connected to source Cgs = oxWL/tox = CoxWL = CpermicronW
polysilicongate
Cpermicron is typically about 2 fF/m
n+ n+
p-type body
W
L
tox
SiO2gate oxide
(good insulator, ox= 3.90)
20
8/13/2019 VLSI lec 3
11/24
Dynamic Behavior of the Transistor Propagation Delay, Tp: defines how quickly output is affected by input
Measured between 50% transition from input to output tpLH defines delay for output going from low to high
tpHL defines delay for output going from high to low Overall delay, tp, defined as the averageof tpLH and tpHL Rise and fall time, Tr & Tf: Defines slope of the signal
Defined b/w 10% and 90% of the signal swing Propagation delay and rise and fall times affected by the fan-out due to
larger capacitance loads
Standard method to measure gate delay is based on ring oscillator: 2Ntp >> tf + tr for proper operation
tpH L
tpL H
t
t
Vin
Vou t
50 %
50 %
tr
10 %
90 %
tf
tpH L
tpL H
t
t
Vin
Vou t
50 %
50 %
tr
10 %
90 %
tf
v 0 v 1 v 2 v 3 v 4 v 5
v0 v 1 v5
T = 2 tp N
Ring Oscillator21
Switch Model of Dynamic Behavior
VDD
Rp
Vout
CL
VDD
R
VoutCL
Gate response time is determined by the time to charge CL through Rp(discharge CL through Rn)
VDD
Vout
CLIav
tpH L = C LVswing /2
Iav
CL
kn VDD~
Technique-1
Vin = 0 Vin = V DD
in= DD
VDD
Vout
Vin= V DD
Ron
CL
tpHL = f(Ron.CL)
= 0.69 RonCL
t
Vou t
VDD
RonCL
1
0.5
ln(0.5)
0.36
)/( LonCRt
OHout eVV
=
Technique-2 22
8/13/2019 VLSI lec 3
12/24
How can the designer build a fast gate? tpHL = f(Ron*CL) Keep output capacitance, CL, small
low fan-out
Keep interconnections short (floor-plan your layout!) Decrease on-resistance of transiston Increase W/L ratio Make good contacts (slight effect)
3
23
0 0 .5 1 1 .5 2 2 .5
x 1 0-1 0
-0.5
0
0 .5
1
1 .5
2
2 .5
t ( sec)
Vout(
V) tp = 0.69 CL (Reqn+Reqp)/2tpHL tpLH
Transient Response
Inverter Transient Response
Vin
tf trtpHL tpLH
VDD=2.5V0.25mW/Ln = 1.5W/Lp = 4.5Reqn= 13 k ( 1.5)Reqp= 31 k ( 4.5)
tpHL = 36 psec
tpLH = 29 psec
so
tp = 32.5 psec
t (sec) x 10-10
From simulation: tpHL = 39.9 psec & tpLH = 31.7 psec
24
8/13/2019 VLSI lec 3
13/24
Propagation Delay Analysis - Switch Model
V
Propagation delay depends on input patterns Example: Two input NAND gate
Two PMOS transistors are ON
Delay: 0.69x(Rp/2)xCL Only one PMOS transistor ON then Delay: 0.69xRpxCL
Large number of transistors (2N) increases overall capacitance of gate Series connection of transistors in PUN/PDN of gate causes additional slow down
25
VDD
CL
FCL
CL
F
F
Rp
Rp Rp Rp
Rp
Rn
Rn
Rn Rn Rn
AA A
A
A
A
B B
B
B
Inverter 2 input NAND 2-input NOR
t p = 0.69 Ron C L
(assuming that C Ldominates!)
= RON
Analysis of Propagation DelayVDD
CL
F
Rp Rp
Rn
Rn
A
A B
B
1. Assume Rn=Rp= resistance of minimumsized NMOS inverter
2. Determine Worst Case Input transition(Delay depends on input values)
3. Example: tpLH for 2input NAND- Worst case when only ONE PMOS Pulls
up the output node
- For 2 PMOS devices in parallel, theresistance is lower
VDD
CL
F
A B
B
2
1 1
VDD
A
B
C
D
A 2
2
24
4
F
26
2-input NAND
4. Example: tpHL for 2input NAND- Worst case : TWO NMOS in series
tpLH = 0.69RpCL
tpHL = 0.69(2Rn)CL
AD
B C
122
Here it is assumed that Rp=Rn
Design for worse case
8/13/2019 VLSI lec 3
14/24
Elmore Delay
R1
C1
R2
C2
Ri-1
Ci-1
Ri
Ci
RN
CN
Vin
N
1 2 i-1i
Elmore delay is equivalent to first-order time constant of n/w Time constant represents simple approx. of actual delay b/w source
node and node i
27Delay Optimization
Approx RC model of digital cir circuits
= = =
==
N
i
i
j
N
i
iiijiDN RCRC1 1 1
RC Delay Model Use equivalent circuits for MOS transistors
Ideal switch + capacitance and ON resistance NMOS has resistance R, capacitance C PMOS has resistance 2R, capacitance C
Capacitance proportional to width Resistance inversely proportional to width
kg
d
g
d
kCR/k
d
s
kC
kC
2R/k
s
s
kCkC s
d
kC
k1 Capacitance
C = Cg = Cs = Cd = 2 fF/m of gate width Values similar across many processes
Resistance R 6 K*m in 0.6um process Improves with shorter channel lengths
Unit transistors
May refer to minimum contacted device (4/2 ) Or maybe 1 m wide device Doesnt matter as long as you are consistent 28
8/13/2019 VLSI lec 3
15/24
Inverter Delay Estimate Estimate the delay of a fanout-of-1 inverter
C
2C
2C
R
2
1A
Y
C
2C
C
2C
C
2C
RY
2
1
Cd = 6RC
29
Transistor Families CMOS: Displace bipolar technology in digital applications are as follows:
CMOS logic circuits dissipate much less power than bipolar logic circuits High input impedance of MOS transistor allows designer to use charge
storage for temporary storage of information in both logic & memory circuits Feature size (i.e., minimum channel length) of MOS transistor has
decreased which permits very tight circuit packing & integration Bipolar: Two logic-circuit families based on bipolar junction transistor are in
some use at present: TTL & ECL
BiCMOS: Combines the high operating speeds possible with BJTs with lowpower dissipation & other excellent characteristics of CMOS Gallium Arsenide (GaAs): High carrier mobility results in very high speeds
30
8/13/2019 VLSI lec 3
16/24
Transistor Families Resistortransistor logic (RTL)
Direct-coupled transistor logic (DCTL) Resistortransistor logic (RCTL)
Diodetransistor logic (DTL) Complemented transistor diode logic (CTDL) High-threshold logic (HTL)
Emitter-coupled logic (ECL) Positive emitter-coupled logic (PECL) Low-voltage positive emitter-coupled logic(LVPECL)
31
Transistortransistor logic (TTL) P-type metaloxidesemiconductor logic (PMOS) N-type metaloxidesemiconductor logic (NMOS)
Depletion-load NMOS logic
Complementary metaloxidesemiconductor logic (CMOS) Bipolar complementary metaloxidesemiconductor logic (BiCMOS) Integrated injection logic (I2L)
Transistor Families
Family DescriptionPropagation
delay (ns)
Togglespeed(MHz)
Power/gate@1MHz (mW)
Typical supplyvoltage V (range) Year Remarks
RTL Resistortransistor logic
__ 4 10 3.3 1963 First CPU built f rom used RTL.
DTL Diodetransistor logic
__ __ 10 5 1962 Introduced by Signetics, Fairchild 930line became industry standard in 1964
CMOS AC/ACT 3 125 0.5 3.3 or 5 (2-6 or 4.5-5.5)
1985 ACT has TTL Compatible levels
CMOS HC/HCT 9 30 0.5 5 (2-6 or 4.5-5.5) 1982 HCT has TTL compatible levels
CMOS 4000B/74C 30 5 1.2 10V (3-18) 1970 Approximately half speed & power at 5V
TTL Original series 10 25 10 5)4.75-5.25( 1964 Several manufacturers
32
TTL L 33 3 1 5)4.75-5.25( 1964 Low power
TTL H 6 43 22 5)4.75-5.25( 1964 High speed
TTL S 3 110 19 5)4.75-5.25( 1969 Schottky high speed
TTL LS 10 33 2 5)4.75-5.25( 1976 Low power Schottky high speed
TTL ALS 4 34 1.3 5)4.5-5.5( 1976 Advanced Low power Schottky
TTL F 3.5 100 5.4 5)4.75-5.25( 1979 Fast
TTL AS 2 105 8 5)4.5-5.5( 1980 Advanced Schottky
TTL G 1.5 1125(1.125GHz)
1.65-3.6 2004 First GHz 7400 series logic
ECL ECL III 1 500 60 -5.2)-5.19--5.21( 1968 Improved ECL
ECL MECL I 8 31 -5.2 1962 first IC commercially produced
ECL ECL 10K 2 125 25 -5.2)-5.19--5.21( 1971 Motorola
ECL ECL 100K .75 350 40 -4.5)-4.2--5.2( 1981ECL ECL 100KH 1 250 25 -5.2)-4.9--5.5( 1981
8/13/2019 VLSI lec 3
17/24
Field Effect Transistors (FET) Field effect devices are voltage controlled by action of electric field, rather than
carrier injection FETs have weak electrical signal coming in through one electrode creates an
electrical field through the rest of transistor
FET transistor uses an electric field to control the shape & conductivity of achannel of one type of charge carrier in a semiconductor material
FETs are unipolar, as they involve single-carrier-type operation FET behaves like bipolar transistor with important difference that gate has a very
33
FET Types N Channel (as NPN transistor) P Channel (as PNP transistor) N & P channel each come as:
Enhancement mode (just IGFET) Depletion mode (IGFET or JFET)
34
8/13/2019 VLSI lec 3
18/24
Basic Operation of JFET JFET operation can be compared to a water spigot: Source of water pressure: Accumulated electrons at ve pole of applied voltage
from Drain to Source Drain of water: Electron deficiency (or holes) at +ve pole of applied voltage from
Drain to Source Control of flow of water: Gate voltage that controls the width of n-channel, which
in turn controls the flow of electrons in n-channel from Source to Drain
35
JFET If channel is doped with donor impurity, n-type material is formed & channel
current will consist of electrons If channel is doped with acceptor impurity, p-type material will be formed &
channel current will consist of holes N-channel devices have greater conductivity than p-channel types, since
electrons have higher mobility than do holes; thus n-channel JFETs areapproximately twice as efficient conductors compared to their p-channel
counterparts Magnitude of current is controlled by voltage applied to gate, which is reverse-
36
biased
8/13/2019 VLSI lec 3
19/24
An Insulated Gate FET
Source Gate Drain
SubstrateNType
P Type
Insulator
Metal
Source Gate Drain
SubstrateNType
P Type
Insulator
Metal
Insulated Gate FETn-Channel Enhancement mode
Source Gate+ Drain
SubstrateNType
P Type
Insulator
Metal
Channel
n-Channel enhancement
JFET
Source Gate+ Drain+
SubstrateNType
P Type
Insulator
Metal
Channel
N enhance FET atPinchoff
N enhance FETbeyond Pinchoff
Source Gate+ Drain++
SubstrateNType
P Type
Insulator
Metal
Channel
n-Channel Depletion modeFET ON No gate bias
Source Gate Drain
SubstrateNType
P Type
Insulator
Metal
Channel
n-Channel Depletionmode FET off
Source Gate- Drain
SubstrateNType
P Type
Insulator
Metal
NoChannel
37
MOSFET MOSFETMetalOxideSemiconductor Field-Effect Transistor is a transistor
used for amplifying or switching electronic signals Source & Drain terminals are specified by operation voltage MOSFET has four-terminals: Source (S), Gate (G), Drain (D), Body (B) Body (or substrate) of MOSFET often is connected to source terminal, making it
three-terminal MOSFET is by far the most common transistor in both digital and analog circuits,
though the bipolar junction transistor was at one time much more common
38
8/13/2019 VLSI lec 3
20/24
BJTBipolar Junction Transistors is current controlled device consists of two pn-junctions
Three region: Base, Collector , Emitter
Operational modes are Base-Emitter & Base-Collector voltages When there is no base current, almost no collector current flows When base current flow, collector current can flow BJT consists of N-type material with P-type on either side, or visa-versa
PNP P-type N-type P-typeNPN N-type P-type N-type
BJTs
39
VMOS VMOS Vertical MOSFET increases the surface area of the device Advantage:
This allows the device to handle higher currents by providing it moresurface area to dissipate the heat
VMOSs also have faster switching times
40
8/13/2019 VLSI lec 3
21/24
CMOS CMOS Complementary Metal Oxide Semiconductor employs both PMOS &
NMOS devices CMOS Complementary MOSFET p & n-channel MOSFET on same substrate
If substrate is p-type, PMOS transistors are formed in n-well (n-type body need) If substrate is n-type, NMOS transistors are formed in p-well (p-type body need) Substrate & well are connected to voltages which reverse bias the junctions for
device isolation Advantage:
Lower power consumption Higher input impedance Faster switching speeds
41
BJT Vs. CMOS Advantages of Bipolar over CMOS
Switching speed Currents drive per unit area Noise performance Analog capability Input/output speed
Advantages of CMOS over Bipolar
Power dissipation Noise margin Packing density Ability to integrate large complex functions with high yields
42
8/13/2019 VLSI lec 3
22/24
BiCMOS BiCMOS technology combines Bipolar & CMOS transistors onto a single IC
where the advantages of both can be utilized BiCMOS technologies have tended to evolve from CMOS processes in order to
obtain the highest CMOS performance possible Bipolar processing steps have been added to core CMOS flow to realize the
desired device characteristics BiCMOS gates can be an efficient way of speeding up VLSI circuits
Advantages:
Lower power dissipation than Bipolar Flexible input/outputs High performance analog Latch up immunity
43
BiCMOS Structure NMOS device, built on top of P+substrate PMOS transistor, built in an N-well approximately 5um deep P+substrate is used to reduce latch up susceptibility by providing low impedance
patch through a vertical pnp device Polysilicon gates are used for both the PMOS & NMOS transistors
Simplest way to add npn bipolar transistor to CMOS structure is by using PMOSN-well as collector of Bipolar device & introducing an additional mask level for P-
base region P-base is approx 1-um deep with doping level of about 1e17 atoms/cm^3 N+source/drain ion implantation step is used for emitter & collector contact of
bipolar structure P+source/drain ion implantation step is used to create a P+ base contact to
minimize the base series resistance
44
8/13/2019 VLSI lec 3
23/24
CMOS Nanoelectronics Key of success of CMOS nanoelectronics: additive technology features
Ultra-thin body FET Double- (or Multi-) gate FET
Strained Si (bulk, on insulator) Ge (bulk, on insulator) High-k gate dielectrics Metal gates Crystal orientation
45
Emerging Devices New technologies based on something other than electronic charge
thatextend the scaling of information processing technologies throughmultiple generations beyond 2019
Provide a path to scale CMOS to the end of Roadmap using new transistorstructural designs & new materials Transport-Enhanced FETs: Enhance the velocity of carriers
Ultra-thin Body SOI FETs: Thin transistor body is employed Source/Drain Engineered FETs: Engineering the source/drain Multiple Gate FETs:
N-Gate (N>2) FETs: Use more than two gates to improve electrostatic control Double-gate FETs: Use two isolated gates for low-power & mixed-signal
processing
46
8/13/2019 VLSI lec 3
24/24
Emerging Devices
Resonant tunnel devices
Molecular devices
ng e-e ec ron rans s ors
Spin transistors 47