Monday, July 30, 2012

Introduction to Very Large Scale Integration (VLSI) Technology

Introduction to Very Large Scale Integration (VLSI) Technology

Structure of the transistor:

       The cross-section of an n-type MOS transistor as shown in fig.-1 an n-type transistor is embedded in a p-type substrate; it is formed by the intersection of an n-type wire and a polysilicon wire. The region at the intersection called the channel, is where the transistor action takes place. The channel connects to the two n-type wires, which form the source and drain, but is itself doped to be p-type. The insulating silicon dioxide at the channel (called the gate oxide) is much thinner than it is away from the channel (called the field oxide); having a thin oxide at the channel is critical to the successful operation of the transistor.

                            Fig.-1: Cross-section of an n-type transistor.
         The transistor works as a switch because the gate-to-source 
voltage modulates the amount of current that can flow between the source
and drain. When the gate voltage (Vgs) is zero, the p-type channel is full of 
holes, while the n-type source and drain contain electrons. The p-n 
junction at source terminal forms a diode, while the junction at the drain 
forms a second diode that conducts in the opposite direction. As a result,
no current can flow from the source to the drain.
         As  Vgs  rises above zero, the situation starts to change. While the 
channel region contains predominantly p-type carriers, it also has some 
n-type carriers. The positive voltage on the polysilicon which forms the 
gate attracts the electrons. Since they are stopped by the gate oxide, 
they collect at the top of the channel along the oxide boundary. At a 
critical voltage called the threshold voltage (Vt), enough electrons 
have collected at the channel boundary to form an inversion layer-a 
layer of electrons dense enough to conduct current between the 
source and the drain.
         The size of the channel region is labeled relative to the direction of 
current flow: the channel length (L) is along the direction of current
flow between source and drain, while the width (W) is perpendicular to 
current flow. The amount of current flow is a function of the W/L ratio, 
for the same reasons that bulk resistance changes with the object’s width 
and length: widening the channel gives a larger cross-section for 
conduction, while lengthening the channel increases the distance, current 
must flow through the channel. Since we can choose W and L when we 
draw the layout, we can very simply design the transistor current 

Drain current characteristics:

For an n-type transistor, we have:

   Linear region ( Vds < Vgs – Vt ):
        Id = k’(W/L)[(Vgs-Vt)(Vds-0.5Vds2)]
 Saturation region (Vds ≥ Vgs - Vt):
        Id =0.5k’(W/L)(Vgs- Vt)2(W/L) = the width-to-length ratio of the transistor.

Both  Vt  and k’ are measured, either directly or indirectly, for fabrication a process, W/L is determined by the layout of the transistor, but since it does not change during operation, it is a constant of the device equations.

0.5 mm transconductances:

From an MOSIS process:
k’n = 73 mA/V2           where,    Vt = 0.7 V

k’p = 21 mA/V2,         where,     Vt = -0.8 V

Current through a transistor:

Use 0.5 mm parameters. Let W/L = 3/2. Measure at the boundary between 
linear and saturation regions.

Vgs = 2V:     Id = 0.5k’(W/L)(Vgs-Vt)2= 93 mA

Vgs = 5V:     Id = 1 mA

Parallel plate capacitance:

1.     Formula for parallel plate capacitance:       
       Cox = eox / xox
2.     Permittivity of silicon:      
       eox = 3.46 x 10-13 F/cm2
3.     Gate capacitance helps determine charge in channel which forms 
       inversion region.

Threshold voltage

1.     Components of threshold voltage Vt:
2.     Vfb = flat band voltage; depends on difference in work function between 
       gate and the substrate and on fixed surface charge.
3.     fs = surface potential (about 2ff).
4.     The Voltage on parallel plate capacitor.
5.     Additional ion implantation.

Body effect

1.     Reorganize threshold voltage equation:   
      Vt = Vt0 + DVt
2.     The threshold voltage is a function of source/substrate voltage Vsb.
3.     Body effect g is the coefficient for the Vsb dependence factor.


Example: threshold voltage of a transistor

Vt0 = Vfb + fs + Qb/Cox + VII
  = -0.91 V + 0.58 V + (1.4E-8/1.73E-7) + 0.92 V
  = 0.68 V

Body effect gn = sqrt(2qeSiNA/Cox) = 0.1
DVt = gn[sqrt(fs + Vsb) - sqrt(Vs)]
        = 0.16 V


1.     Process transconductance k’ = mCox.
2.     Device transconductance b = k’W/L.


a.     Describes a small dependence of drain current on Vds in saturation.
b.    The Factor is measured empirically.
c.      New drain current equation:
      Id = 0.5k’ (W/L)(Vgs - Vt2(l l Vds)
d.     Equation has a discontinuity between linear and saturation 
      regions---small enough to be ignored.


 Basic transistor parasitics

a.     Gate to the substrate and Gate to source/drain.
b.     Source/drain capacitance, resistance.

c.     Gate capacitance Cg. Determined by active area.
d.     Source/drain overlap capacitances Cgs, Cgd
      Determined by source/gate and drain/gate overlaps. Independent of 
      transistor L.  Cgs = Col W
e.      Gate/bulk overlap capacitance.


a.     CMOS ICs have parasitic silicon-controlled rectifiers (SCRs).

b.     When powered up, SCRs can turn on, creating low-resistance path 
       from power to ground. Current can destroy the chip.

c.      Early CMOS problem. Can be solved with proper circuit/layout