In this video, we will discuss some advanced operation of MOSFET.

Before we do
that, I want to rewrite the long-channel MOSFET equation, that we derived in the
previous video in this form. You can see that this equation has a typical standard
drift equation form in that, the charge density, this force to C oxide times W
times the quantity in the parenthesis, represents the charge density and then mu
times the VD over L is an electric field inside the channel. So, this is a very
standard drift current equation and the charge density term represents the charge
density at the middle of the channel or the average charge density. Then, VD over L
is the electric field produced along the y direction, the channel direction,
produced by the drain voltage. So, the long-channel MOSFET equation really
represents the current that would flow in the channel containing on average charge
density of this guy is here, drift along a constant electric field of VD over L.
Now, in reality, of course the inversion layer charge density changes and it
actually decreases, charge density actually decreases as it approaches the drain.
All the while, your ID should remain constant as a function of y because Kirchhoff
law says that as long as there is no branch, your current should be constant
everywhere, right? So, our simple model predicts that in saturation, your channel
current is determined by the rate at which the carrier arrives at the pinch-off
point. Once, the carrier reaches the pinch-off point then it just gets drifted by
the large built-in electric field inside the depletion region. So, that was what we
assumed. Now, in the long-channel MOSFET, the small movement of the pinch-off point
inward is not important. However, as the channel length becomes reduced in modern
IC technology, the channel lengths reduction has been one of the main drives in the
technological development. So, this pinch-off region with L minus L prime in this
figure here, that region width now may not be completely negligible anymore. In
that case, we have to somehow take that into account. How do we do that? The best
way to do it, is to consider this as a reduced channel length. So, we still assume
that the carriers arrived at the pinch-off point gets promptly swept away by the
larger built-in electric field. So, the net effect of this pinch-off region is to
reduce the channel length and this effect is called the channel length modulation.
So, your current actually increases due to the reduction of your effective channel
length and if your delta L, which is defined as L minus L prime. If that remains
small compared to your total channel length, you can approximate this as this. So,
now you can show that your ID, your drain current, increases linearly with the
reduction of your channel length. You can calculate the reduced channel length
using the Poisson's equation and you can, in much the same way that as we did in
the p-n junction and the Schottky contact case and that will give you this equation
here. However, this equation is not entirely accurate because we have ignored the
free carrier density that's still floating around in the region and also, we have
adopted this gradual channel approximation and uses a simply 1D case. So, in order
to be more accurate you have to use the full 2D Poisson's equation. Nevertheless,
the approach of describing the saturation drain current with the reduced channel
length is effective and valid. So, saturation drain current is written in a similar
way to the only fact in Bipolar Junction Transistor and is written by this. So, the
effect of channel length modulation is represented by this extra term here, which
is linearly proportional to the drain voltage. So, this leads to a finite
resistance in the saturation region. Then so, if you plot the IV characteristic
once again, the saturation region which we use to describe it with a flap constant
saturation current, now is linearly increasing with the drain voltage and the slope
is determined. Slope is characterized by this parameter V sub A and if you
extrapolate this this saturation region, then it will intersect the x-axis at this
value negative V sub A, this voltage is called the early voltage. The effect of
this, it describes the effect of the channel length modulation leading to a
linearly increasing saturation drain current relative to the drain voltage. The
next effect I want to discuss is the Body Bias Effect. So far, we assumed the
source voltage and the body voltage discharge are the same, but it may not always
be the case. What is the effect of the body bias? Well if you look at the source
and the body, this is a p-n junction. So, when you apply a positive voltage on the
body, then you are applying a reverse bias voltage on this p-n junction. When you
apply a reverse bias on the p-n junction, you increase the band bending. So, what
it does is, you're effectively changing the potential of your bulk region. So,
because of this energy band bending, you are affecting the voltage, gate voltage,
that is required to create the inversion layer. So, if you look at the Module one,
video four, in this course. You can find that applying a body voltage creates a
shift in threshold voltage and the threshold voltage change is related to your VSP,
the source to body voltage as this. So, the primary gamma is called the body effect
parameter and has a unit of square root of volt and this equation is valid for a
long-channel MOSFET but it will lose its validity for a short-channel device and in
order to rigorously described this body bias effect, you have to use the full 2D
model that describes the actual charge distribution in the short-channel device.
But in any case, what is the effect of body bias? Body bias changes the threshold
voltage. How does that work? Pictorially, this is the best illustration. So, here
is the MOSFET on its side. So, the gate is on the front, substrate is on the back,
source on the left and drain on the right. If you draw the three-dimensional energy
band diagram, then for zero body bias case, it looks like this. So, this figure B,
shows the case of the flat band and then this is the case, this figure C is the
case of onset of strong inversion. So, you can see that by applying a voltage on
the gate, you create an energy band bending along x direction here. Okay, if the
band bending is large enough then you can create strong inversion. If you have a
body bias however, there is a larger energy band bending between source and body.
So, you start out with a larger band bending here and here. Therefore, if you apply
the same voltage on the gate, the energy band bending is not large enough to create
the inversion layer you have to apply a larger voltage on the gate to create the
inversion layer. That is described by this delta VT shift change in the threshold
voltage.