Professional Documents
Culture Documents
VLSI Unit-1
VLSI Unit-1
VLSI Unit-1
Introduction to VLSI
Advantages of VLSI:
Physically smaller
Fabrication process.
Introduction to IC Technology:
Reliability
Low cost
vacuum tube,
single transistor,
IC (Integrated Circuits)
ICs:
Integration improves
size
speed
power
Integration improves
size
speed
power
IC Evolution:
5th Generation:
• Size of chip.
• Smaller chip.
• Faster.
• More reliable.
• Bipolar technology
MOS (Metal-oxide-silicon)
Moore’s Law:
Moore’s Plot:
Variation of Gate length & supply voltg w.r.t Year:
• cost effective
These devices have feature size (such as film thickness, line width
etc) measured in nanometres or 10-9 metres
Recurring Costs:
VLSI Applications:
Microprocessors
personal computers
microcontrollers
Leading Ic technology:
• nMOS
• CMOS
3. Easy transition.
Make sure that the inside of the machine is very clean too and that
the gas flow - the gas you introduce but also the SiO coming from
the molten Si because parts of the crucible dissolve - does not
interfere with the growing crystal.
Dissolve the Si in the crucible and keep its temperature close to the
melting point. Since you cannot avoid temperature gradients in the
crucible, there will be some convection in the liquid Si. You may
want to suppress this by big magnetic fields.
Insert your seed crystal, adjust the temperature to "just right", and
start withdrawing the seed crystal. For homogeneity, rotate the
seed crystal and the crucible. Rotation directions and speeds and
their development during growth, are closely guarded secrets!
First pull rather fast - the diameter of the growing crystal will
decrease to a few mm. This is the "Dash process" ensuring that
the crystal will be dislocation free even though the seed crystal may
contain dislocations.
Now decrease the growth rate - the crystal diameter will increase -
until you have the desired diameter and commence to grow the
commercial part of your crystal at a few mm/second.
Fabrication Cycle:
Processing Wafers:
Metal - wires
Polysilicon - gates
Oxide
Mos transistor:
P Transistor Operation:
Opposite of N-Transistor
The fabrication processes used for nMOS are relevant to CMOS and
Bi-CMOS which may be viewed as involving additional fabrication
steps.
Such wafers are typically 4" to 8" in diameter and 0.4 mm thick and
are doped with, say, boron to impurity concentrations of 10 15/cm3 to
1016/cm3, giving resistivity in the approximate range 25 Ohm/cm to 2
Ohm/cm.
5 . These areas are subsequently readily etched away together with the
underlying silicon dioxide so that the wafer surface is exposed in the
window defined by the mask.
is grown over the entire chip surface and then polysilicon is deposited on
top of this to form the gate structure.
(as shown in Step 6), and then the thin oxide is removed to expose areas
into which n-type impurities are to be diffused to form the source and
drain as shown.
8. Thick oxide (SiO2) is grown over all again and is then masked with
photoresist and etched to expose selected areas of the polysilicon gate
and the drain and source areas where connections (i.e. contact cuts) are
to be made.
9. The whole chip then has metal (aluminium) deposited over its
surface to a thickness typically of 1 mm. This metal layer is then masked
and etched to form the required interconnection pattern.
The process revolves around the formation or deposition and patterning of
three layers, separated by silicon dioxide insulation.
layer of SiO2.
• Mask 1 - Pattern SiO2 to expose the silicon surface in areas where paths
in the diffusion layer or source, drain or gate areas of transistors are
required. Deposit thin oxide over all. For this reason, this mask is often
known as the ‘thinox’ mask but some texts refer to it as the diffusion
mask.
• Mask 2 - Pattern the ion implantation within the thinox region where
depletion mode
Using the same mask, remove thin oxide layer where it is not covered by
polysilicon.
• Diffuse n+ regions into areas where thin oxide has been removed.
Transistor drains and sources are thus self-aligning with respect to the
gate structure.
• Mask 4 - Grow thick oxide over all and then etch for contact cuts.
• Mask 6 - This mask is required for the overglassing process step, and is
called the
passivation mask.
The p-well and n-well processes are widely used in practice (the n-
well process was an easy retrofit to existing nMOS lines).
– N-Well
– Twin Tub
• p-well process:
The basic processing steps are of the same nature as those used for
nMOS.
This diffusion must be carried out with special care since the p-well
doping concentration and depth will affect the threshold voltages as well
as the breakdown voltages of the ntransistors.
P-well on N-substrate :
Steps :
• N-type substrate
• P-well doping
The two areas are electrically isolated using thick field oxide (and often
isolation implants [not shown here])
Steps :
• Grow thin layer of SiO2 (~0.1mm) gate oxide, over the entire chip
surface
Advantages:
The n-channel devices are used to form logic elements that provides :
2. High Density.
Disadvantages:
In conventional p & n-well CMOS process, the doping density of the well
region is typically higher than the substrate , which results in un-balanced
drain Parasitics.
Since two independent doping steps are performed for the creation
of the well regions, the dopant concentrations can be carefully optimized
to produce the desired device characteristics. The Twin-Tub process is
shown below.
1. speed and
2. latch-up susceptibility.
Advantages:
Bi-CMOS Technology:
Low output drive current (issue when driving large capacitive loads)
essentially unidirectional
• Design uses CMOS gates along with bipolar stage where driving of
high capacitance loads is required
Advantages of Bi-CMOS :
A known deficiency of MOS technology is its limited load
driving capabilities (due to limited current sourcing and sinking
abilities of pMOS and nMOS transistors.
higher gain
Example Applications
CMOS - Logic
2. The n+ sub-collector .
Length of channel ( L)
Tsd
Velocity (V )
but Velocity
V = µEds
Where
mVds
V
L
L2
Tsd 2
mVds
Vg = VGS - Vt
Vt = Threshold Voltage
The charge gets enduced into the channel due to gate voltage and if E g is
average electric field. From gate to channel.
Qc=EgtinstoWL
[(Vgs Vt ) Vds ]
Eg
Eg= D
Where,
D is oxide thickness.
Eins EoWL V
Qc (Vgs Vt ) ds 3
D 2
E ins Eo m W V
I ds = (Vgs Vt ) ds Vds
D L 2
W Vds
2
I ds = K (V
gs Vt )Vds 4 or resistive region where
In the non-saturated
L 2
Vds <Vvs – Vt and
Let Eins Eo m
K
D
W
K
L
So that
V
2
I ds = (Vgs Vt )Vds ds
2
5
C gm Vds
2
6
I ds = (V
gs Vt )Vds
L2 2
Cg=CoWL
W Vds
2
I ds = Co m (Vgs Vt )Vds 7
L 2
Saturation begins when Vds=Vvs-V, at this point the IR drop in the channel
equals the effective gate to channel voltage. The current through the
channel remainsI fairlyW (V V ) 2
constant for any further increase in V ds.
gs t
ds = K
L 2
W
(Vgs Vt ) 2 K
I ds = N
2
Eins toWL
I ds =
C gm
2L2
V gs Vt
2
C g
D
QB=The Charge per unit area in the depletion layer beneath the oxide.
fN
Fermi level potential between invested surface and bulk si
KT n
fN n
Q ni voth
Where VSB = substrate bias voltage (-Ve wrt source for nµos and +Ve wrt
PMOS)
Q=1.6×10-19 coulomb
The Variation with Vt with VSB (Substrate bias) is called body effect.
Body effect Co-efficient = 0.223 (Volts)1/2
Vt interms of VSB is
D
2 EO ESi qN VSB
1/ 2
Vt = Vt(O) ±
E O Eins
Empirical formula for V(t)
D=Oxide thickness
EO=Absolute permittivity
Q=Electron charge
VSB=Substrate bias.
Vt Vt o VSB
1/ 2
Where
0.233 (Volts)1/2 is body effect Co-efficient
L2
ds
mVds
SQcVds m
SI ds
L2
But change in change
SQc C g SV gs
So that
C g SVgs mVds
SI ds
L2
Now
SI ds C g mVds
gm
SVgs L2
In saturation
Vds V gs Vt
Cg m
gm V gs Vt
L2
Eins EoWL
Cg
And substituting for D
mEins Eo W
gm Vgs V
D L
g m Vgs V
Vds
Rds Vgs 1
I ds Constant
We have
W
Vgs Vt Vds Vds
2
I ds K 2
L 2
I ds K
W
L
Vgs Vt Vds 3
By Substituting in we get
3 1
Vds
Rds
K
W
L
Vgs Vt Vds
1
Rds
K
W
L
Vgs Vt
1
W
K Vgs Vt
L
g ds K
W
Vgs Vt
L
m 1
2 Vgs Vt
gm
Wo 1
cg L sd
This shows that switching speed depends on gate goltage above threshold
and on carries mobility and inversely as the square of channel length. A
fast circuit require that gm be as high as possible.
Body Effect:
Vt
2qN A E si
Co
2 s Vsb 2 s
Vt n 2 s Vsb 2 s
Where
n in body effect factor depends on gate oxide thickness and substrate
doping.
Pass Transistor:
When the transistor is operating in the linear region, the device act as
a linear resistance under gate voltage control.
PMOS has opposite behavior passing strong ‘I’ s and degraded ‘O’.
NMOS Inverter:
With no current drawn from the output, the currents I ds for both
transistor4s must be equal.
For the depletion made transistor, the gate is connected to the source
so it is always on and only the characteristic curve Vgs=O is relevant.
In this configuration the depletion made device is called the pull-up
(P.u.) and the enhancement made device the pull-down, (P.d)
transistor.
Note that as Vin(=Vgs P.d Transistor) exceeds the P.d threshold voltage
current begins to flow. The output voltage V out will cause the P.d
transistor to come out of saturation and become resistive. Note that
the P.u transistor is initially resistive as the p.d turns on.
The point at which Vout=Vin is denoted as Vinv and it will be noted that
the transfer characteristics and Vinv can be shifted by variation of the
ration of pull-up to pull – down resistances.
Vin Vout
Consider the depletion mode transistor for which V gs=O under all
conditions, and further assume that in order to cascade inverters without
degradation of levels we are aiming to meet the requirements.
Vin=Vout=Vinv
w p.u rtd 2
I ds k 2
l p .u 2
Since Vgs=O
w p.d vinv vt 2
I ds k 3
l p.d 2 Since V =V
gs inv
By equating 2
& 3
w p.d vinv vt 2 w p.d
vtd 2
l p .d 2 l p.u
Where Wp.d, Lp.d, Wp.u and Lp.u are the widths and length of pull-down and
pull-up transistor respectively.
Now write
L p.d L p.u
Z p.d ; z p.u
W p.d W p.u
We have
1
vinv vt 2 1 vtd 2
Z p.d Z p .d
Hence
Vtd
Vinv Vt
Z p .u Z p.d 4
4 in
By substituting the typical values of the following
Vinv=0.5VDD
0.6
0.5 0.2
Z p.u Z p .d
Z p .u Z p . d
Hence =2
Z p .u 4
Z p .d 1
With all pass transistors gates connected to V DD, these is a loss of V tp,
however many are connected in series, since no static current flows
through them and there can be no voltage drop in the channels.
Vin2=VDD-Vtp
Vdsl l L p .dl 1
R1
I ds k Wp.dl V
VDD V dsl t
2
vds1
Note that Vdsi is small and 2 is ignored
1 1
R1 Z P.d 1
K VDD Vt
2
Z p .d 1 1 1 V td
I1 R1 Vout 1
Z p.u1 VDD Vt VDD Vt 2
Z p.d 2 1 V td 2
Vout 2 I 2 R2
Z p.u 2 VDD Vtp Vt 2
If the inverter2 is to have the same output voltage under these conditions
then.
Vout1=Vout2
Vtp 0.3VDD
Z P.u 2 Z 0 .8
P.u1
Z p.d 2 Z p . d 1 0 .5
Z P .u 2 Z 8
2 P .u1
Z p .d 2 Z p .d 1 5
CMOS Inverter:
The current / voltage relationship for the MOS transistor may be written.
2
Vds
w
I ds k v gs vt Vds
L 2
In the resistive region
2
w v gs vt
I ds k
L 2
Eins Eo m p W p
P
D Lp
Region1:
Region2:
The input voltage increased to a level which just exceeds threshold
voltage of the n-transistor. The n-transistor conducts and has a large
voltage between source and drain; so it is in saturation. The p-transistor
is also conducting but with only a small voltage across it, it operates
unsaturated resistive region.
Region3:
Is the region in which the inverter en habits gain and in which both
transistors one in saturation.
The current in each device must be the same since the transistors are in
series, so we may write.
I dsp I dsn
Where
I dsp
2
And
n
I dsn Vin Vtn 2
2
Hence we can express Vin in terms of the β ratio and the other circuit
voltages and currents.
1
2
VDD Vtp Vtn n
Vin p
1
2
1 n
p
Region4:
Region5:
Vin=logic1, the n-transistor is fully on while the p-transistor is fully off.
Again, no current flows and a good logic 0 appear at the output.
BiCMOS Inverter:
Operation:
When the input voltage Vin of the BiCMOS inverter in figure is low, the
transistor QN and Q2 are cut – off. The transistor Q p is turned on and
provides base current ot Q11 so that Q1 turned on and supplies current to
the load capacitance CL. The load capacitance charges and the output
voltage goes high.
When the input voltage Vin goes high, Qp turns OFF eliminating any bias
current to Q1, so Q1 is also OFF. The two transistors Q N and Q2 turn ON
and provide voltage goes low. The output voltage will discharge to a
minimum value of Vo (min) =VBE (on).
Q1 and Q2 will present low impedances when turned ON into saturation the
load CL will charge and discharged rapidly.
The inverter has high input impedance and low output impedance.
An added advantage of this circuit is, when V in goes high and output
goes low, the very small output current through Q N and R2 means the
output voltage is pulled to ground potential.
Also, as Vin goes low and output goes high, the very small load current
means that the output is pulled up to essentially v DD through the
resistor R1.
The below figure shows the way in which R 1 are implemented nMOS
devices QR1 and QR2 are used to realize R1 and R2.
And these two transistors are made to conduct only when needed. The
QR1 conducts only, when Vin rises. At which time its drain current
constitutes a reverse base current for Q11 speeding up its turn off.
Similarly RQ2 will conduct only when Vin falls and Qp conducts, pulling
the gate of QR2 high.