Lecture 5 Introduction of MOS (3 Terminal)

IIT Roorkee July 201
61 followers
Jan 10, 2025
Hello everybody and welcome to the NPTEL online certification course on Semiconductor Devices
for the Next Generation FET, a Physics Perspective.
This is lecture number 5, where we will start from where we left in lecture number 4 and
primarily we will be looking in this lecture MOS as a 3 terminal device, look into a formal
verification of the threshold voltage and then we will next subsequently go to 2 terminal
device as a capacitor right and how the capacitance will behave when you operate the device in
accumulation, inversion and depletion.
So with this short introduction let me start today's lecture and let's see.
So let me introduce to you the 3 terminal MOS structure.
As you can see here till now we were all discussing it to be as a 2 terminal where you put basically
the metal gate and the body.
The body here on the right, on the left hand side if you see on the figure, the gate is
basically the metal gate which I was talking about, this is the metal gate right, this
is the oxide layer right, this is the semiconductor and at the bottom most layer where you have
this light green colour is basically your body or substrate.
And till now we were actually applying a gate to body bias and trying to figure out how
the charges near the silicon-silicon dioxide interface would change with application of
a reverse bias, a 0 bias, a small forward bias and a large forward bias.
We saw that if it is a reversed biased then you will have an accumulation, at 0 it will
be flat band, as you make your gate voltage slightly higher you enter into the depletion
region and very large value of gate voltage you enter into inversion region.
So I will talk about inversion in my subsequent slides but let us see what we have added here.
We have added an extra N plus region here right, this is highly doped N plus region
which you see on the left hand side blue coloured and this let me name it as C terminal, terminal
number C and it is heavily doped.
And as you can see this is basically a self-aligned structure which means that wherever the oxide
stops your this layer, let me as usual this is source actually, so wherever the oxide
layer stops that is where your source terminal starts right and therefore these are aligned
in the nature.
This is not overlap otherwise in reality what will happen is this is the oxide layer generally
the N plus layer will start like this, this is my overlap region right, this is my overlap
region.
Similarly the drain will start from here N plus region, this is my overlap, this overlap
gives you C overlap or overlap capacities.
But at this stage let us assume that there is no overlap and they are perfectly aligned
with respect to each other and we have added an extra terminal there known as C.
With this knowledge if you can see these inversion layer is already formed, so you have these
negative charges being formed near the silicon-silicon dioxide interface right, you are into the
inversion layer and there is a depletion layer here.
But you see what interesting thing has happened now is if you concentrate at this particular
part this is nothing but N plus P layer right, so there will be always a depletion layer
available within N plus P layer right, this will be always there because it is a p-n junction
diode and the thickness of the depletion layer will be more towards the P side right.
So when you actually do not give, even if you do not give any external bias this N plus
P junction will be actually giving you a depletion region which will be more towards the P side
and that is what I have drawn here if you see, this is the depletion region I am talking
about.
This depletion region is coming by virtue of the gate voltage which we have already
discussed in our previous discussions.
So you see what thing we learn when you introduce a third terminal here known as a source terminal
is that you have been able to actually formulate a depletion region even below the source region
which is depleted of any free charge carriers, there are only immobile charge carriers available
there.
Now what I do look at the, now let us look and throw back on the right hand side which
is this schematic.
In this schematic I have applied a bias on the source side which is with respect to of
course the body which is more than equal to the 0 right.
So when you apply a bias which is more than equal to 0 you actually do what, you actually
reverse bias this N plus P junction because on the N side you are putting a positive bias
and the P side you are putting a negative bias.
This implies that the depletion thickness on this side, I do not want to name it as
source and drain at this stage, on this terminal at least just below this terminal the depletion
thickness will be growing larger and larger right.
So higher the value of V CB which is this one more will be the depletion thickness because
it is a reversed biased PN junction sort of a element which you can see here.
And there will be very very small current between N plus and P again reversed bias because
of the minority current carriers at this stage we will assume that this is infinitesimally
small right.
So it is infinitesimally small whatever contribution is coming is just because you have a depletion
region available there and there will be also an inversion layer which is available here.
So this I have still not introduced the next terminal which is on this side, on the right
hand side this side but this will give you an idea that as you therefore increase the
bias you will obviously the depletion thickness will grow.
But please understand once you increase the bias on the C with respect to B you are also
and let us suppose this side is grounded let us suppose then you are also introduce an
electric field in the longitudinal direction right like this and you can pull the charge
carriers which is available here right.
These are all N type charge carriers so they will be forming here they will go easily if
you apply a positive gate bias and that is basically the current which you see.
So this is the basic introduction of a 3 terminal MOS device right.
Let me come to the concept here remember in my previous all discussions I told you what
is the concept of inversion was that when the number of electrons.
So if you start with the P type substrate and say there were 10 number of holes available
per unit volume near the silicon-silicon dioxide interface right and if you go on giving
a gate voltage such that you are able to form the depletion and then inversion we say that
the onset of inversion will take place when you have as many number of electrons as there
were initially the number of holes that is when you say that your channel is actually
being inverted right.
So you will form a depletion and then you will form an inversion and that can happen
when your surface charges are more than the intrinsic carrier concentration near the silicon-silicon
dioxide interface right.
N surface is basically the electron concentration at the surface which surface silicon-silicon
dioxide surface.
All these discussions are with the fact that typically there are no traps available in
the silicon-silicon dioxide there are no impurities available there the surface is
perfectly clean at this stage and we just say that therefore the electrons which are
getting accumulated near the silicon-silicon dioxide interface are primarily because of
the electric field in the transverse direction which allows it the electrons to flow there.
Therefore we write down N surface to be this formula here where N A is the doping concentration
of the p-type by which you dope it psi s is the surface potential remember 2 phi f why
is it 2 phi f now remember 2 phi f was that this was the minimum amount of bending of
the intrinsic Fermi level so that your inversion is there I will discuss this point in the
subsequent blank slides.
But understand this point that as you apply go on applying the gate bias the intrinsic
Fermi level will start going down right and it becomes lower than the Fermi level by as
much it was higher will entail that a inversion layer is formed and that is the reason you
see we have added a term 2 phi f here what is VCB it is between C and bulk right between
C and bulk you apply the voltage why are you adding this 2 phi f and VCB the reason being
that the reason you are adding VCB in at least in this case is that once your VCB becomes
larger and larger right you are actually doing what you are actually increasing the depletion
width I discussed with you in the previous slide once you do that your inversion layer
goes down and that is the reason you see it is psi s minus it is psi s minus 2 phi f plus
VCB.
So when my this term VCB goes up this quantity goes up when this quantity goes up this quantity
goes down right this quantity goes down and therefore exponential to that quantity which
I have written here will actually go down and this is true also from first principles
of physics this is true also that as the VCB value goes on increasing in a positive sense
you do have N surface going down. So therefore as my VCB goes up my N surface goes down
the number of charge carriers inversion charge carrier goes down drastically right. So this
is sort of a way but let me frame a question for you think about it if VCB would have been
negative then what would have been the case the case would have been that you are actually
forward biasing this N plus P layer this one N plus P layer and when you forward biasing
there will be no depletion region when there is no depletion region it will not eat away
into the inversion layer here you got the point and that is the problem.
So if you look if you let me rub this ink and show it to you so if this is N so this
is P right this is P this is N plus so the depletion region is actually also hitting
this part this part also and that is where it is eating away into the inversion layer.
So your surface charge density is falling down so more you try to increase the value
of C with respect to B your surface charge density goes down inversion surface charge
density goes down and therefore you have to apply a larger gate voltage in order to sustain
a channel you are able to get the picture and therefore the threshold voltage goes up.
Therefore with this basic sort of principle available to you we can look at this fact
that VGB gate to bulk voltage or gate to bulk will be VFB plus SiS we have already seen
plus this term which you see this term right this term which you see will come into picture
and there will be a gamma this is known as body factor we will not go into details at
this stage but this is a body factor right this is a body factor and obviously very straightforward
and simple if you increase the VGB your VGB will increase if you increase your SiS because
of some problem VGB will increase if this whole quantity increases obviously your VGB
will increase you got the idea and that is the reason why your if you see for example
the surface potential surface potentials comes from the fact that how much amount of potential
is available at silicon-silicon dioxide interface.
If you go on increasing the gate bias obviously your SiS will go on increasing so if you increase
this quantity here obviously this quantity will this is almost constant this quantity
will increase right this quantity when this quantity increases this quantity also goes
up SiS also goes up right and so on and so forth.
So this is a basic fundamental principle of the formation of an inversion layer basic
inversion fundamental let me come to now a detailed study of the threshold voltage from
the band diagram as perspective of it and then I will try to give you the value of the
threshold voltage in terms of total charges.
So let us suppose I have a MOS layer right I have a MOS layer here this is my metal and
oxide layer this is my oxide this is my metal and this is a semiconductor which is basically
p-type initially so let us say I have this valence band here EV right I have also a Fermi
level which is close to EV let us suppose so it is a p-type semiconductor and then I
have an intrinsic Fermi level right this is EFI EFI and then I have a conduction band
here this is my conduction band this is the conduction band this is a flat band condition
so there is no bending of the bands taking place on the semiconductor side.
This difference between EFI and EFEF is basically equals to Q times Phi of Fp right what does
it tell me this Q Phi of Fp tells me how much the device is basically p-type that is
also one of the reasons EF is very close to EV.
So if you see if EF is very close to EV EFI minus EF will be typically a large quantity
and which means that it is more of a p-type right so that is the reason this Q times Phi
Fp gives you a sort of a qualitative as well as quantitative analysis of how far the substrate
which you have taken is p-type right and this is no I have not applied any bias here.
So now if I apply a bias here which is positive in nature what will happen is this thing will
happen so let me draw it in inversion this so let me draw a line here and then what will
happen this conduction band edge will start going down this will go down and this will
also go down.
So all these three EC, EFI and EV they will all go down by how much by the amount of potential
of applied here into Q right so if you have applied V then you multiply this with Q all
will fall by that factor in the reverse direction because it is minus Q into V which you will
see.
Now we define and the potential at this particular point at the silicon-silicon dioxide interface
is referred to as SiS the surface potential.
Now we define that we will have threshold voltage right we will have threshold voltage
when your SiS becomes equals to twice Phi of Fp means so this is one Fp and this is
another Fp so this is your Phi Fp and this is also your Phi Fp you got the idea.
So it is dropped by the same factor as you increase the voltage so I get 2 Phi Fp right
when you if this is for a P type semiconductor if you have an N type semiconductor P type
if you have an N type semiconductor then SiS must be equals to 2 Phi Fn for an N type semiconductor
with this knowledge or with this basic idea let us see how the charges will behave once
you apply a positive bias right on the metal side.
So when you apply on the metal side a positive bias let me plot for you the metal oxide semiconductor
charge distribution and show it to you.
So this is my metal layer this is my oxide layer so this is the interface silicon-silicon
dioxide interface this whole thing is oxide the whole part is oxide.
So when you apply a metal you will get on the metal side a value known as QMT prime
which is the charge per unit area in the metal right QMT prime is the charge per unit area
in the metal this will entail a charge on the oxide side which is QSS prime charge per
unit area right.
This consists of many factors I have already discussed in the class for example it is consists
of fixed charges, mobile charges, immobile charges, impurities and so on as traps and
so on and so forth.
So that is QSS in order to balance these two what happens is that on the P side which is
on the P substrate on the semiconductor P type substrate you have this charge this is
XDT, XDT is basically the length of the depletion layer you will have a charge which is equals
to QSD right depletion prime this is this will give you the maximum value and this will
be equals to Q times you will have NA right into XDT this is the total charge which you
will see and this charge has to be balanced across the network right why because it is
Q time why it is Q times NA XDT is the thickness of the depletion region I will show you what
does it mean basically this region this is XDT.
So if you multiply that with NA into Q you get the total charge in that particular region
and you can see it is much large so it is relatively much larger right as compared to
among the metal on the semiconductor side it is in fact if not larger it is actually
extending more towards the semiconductor whereas in the metal and oxide they are very much
very near to the metal layer and very near to the oxide layer.
With this fundamental principle with this let me write down the expression now so this
will be Q of MT prime plus QSS prime must be equals to Q of SD max maximum which you
can get prime mod of it I am not worried at this stage about the sign of this so I get
the metal charge plus the oxide charge must be equals to the QSD charge right.
So if you take it and find out we know that VG right will be equals to change in VG will
be given by so change in VG will be change of V oxide plus change of phi S psi S. Now
you can write down this to be as VOX right voltage across the oxide layer plus surface
potential plus phi MS we have already discussed this point in the initial stages that it is
the voltage drop across the oxide layer surface potential at silicon-silicon dioxide interface
and the metal semiconductor work function difference which you see comes out to be equals
to VGS here.
Now at threshold I will let me write down what will happen at threshold so at threshold
let me write down for you what happens is that I write down VGS or VGS to be equals
to VTN the threshold voltage will be equals to VOX at threshold plus 2 times phi FP by
the condition this condition which condition this condition psi S should be equals to 2
phi FP plus phi MS and if you look at VOX at threshold will be given as Q of MT prime
MT prime divided by COX prime fine. So VOXT is equals to QMT prime by COX prime.
Now this can be written as this QMT prime by COX prime can be written as 1 by COX prime
right into QMT can be written as this could be written as this quantity minus QSS prime.
So I can write down this to be as QSD prime max right minus QSS prime. So if you divide
by C oxide on both the sides I can get therefore I can write down let me write down the value
of VTN now right let me divide into 3 parts so my VTN value which I have taken from here
will be VOXT VOXT can be written as QSD prime max right prime max divided by COX prime right
minus QSS prime divided by COX prime plus phi MS plus 2 phi FP and therefore this can
be now COX can be taken together and COX is equals to COX by epsilon ox remember the
value of COX prime which is charge per unit area is basically capacitance per unit area
is you have to divide it by area so what you get from here is you get VTN to be equals
to QSD prime max QSD prime minus QSS prime blank right into TOX by COX TOX by epsilon
ox why because it is divided by COX prime COX is equals to epsilon ox by TOX so when
you when you divide it becomes TOX by epsilon ox this plus phi MS plus 2 phi FP the same
thing will again come into picture.
Now we already know we already know that VFB we already know that VFB is equals to phi
MS right minus Q prime by COX prime we have already studied this particular problem and
therefore the final expression for VT let me write down for you that is very important
expression is VTN will be equals to QSD max prime divided by COX prime plus VFB plus 2
phi FP this happens to be your final threshold voltage value what does it tell me it that
the threshold voltage depends upon 2 phi FP it depends upon VFB it also depends upon the
maximum depletion charge which you are storing at the MOS device and also on the capacitor
right and so you can see very easily here not doing anything that if you actually reduce
the if you increase the value of COX right your VTN goes down that means if your COX
actually goes on increasing right COX prime goes on increasing your VTN actually starts
to go down.
Now how can you how can you make your COX go high and why is it so physically true because
if COX prime goes high you are actually storing larger amount of charge for the same amount
of gate voltage right so I need to give a lesser amount of gate voltage to invert got
the idea and that is the reason I say COX prime increase but let us correlate this with
the physical parameter COX prime increases COX prime is epsilon ox by T ox right.
Now this can this will only increase when T ox prime actually reduces because epsilon
ox is almost constant so this reduces which means that when T ox actually reduces or the
oxide thickness goes on reducing you are trying to bring the gate closer to the channel and
therefore you have a better control over the channel and that is the reason your threshold
voltage actually shows a drop or shows a decrement in the value.
So this is one very fundamental first principle sort of way to design a low threshold voltage
devices with this knowledge if I therefore plot on the x axis and y axis on the x axis
I do a doping right and this is an A doping acceptor and doping and on the y axis we are
trying to plot the value of ETH the typical way of looking at it is something like this
it goes like this right it is something like this and it suddenly increases and it suddenly
increases why because for low values of acceptor ion concentration right your depletion region
charge is relatively very very small and therefore threshold voltage is small but as the doping
goes on increasing that factor on the left hand side becomes higher and higher and because
of which the threshold voltage suddenly shoots up to a very large value when your doping
basically increases from a low value to a high value this is the doping of the substrate
here right not of the not of the channel or channel region or the any other region.
This is I have told you a brief idea about the threshold voltage how to model threshold
voltage and what are the principles of modeling the threshold voltage let me end this discussion
with a one small numerical example to showcase this say I have a two terminal device where
I have N substrate right so it is basically a PMOS, PMOS I have an M substrate available
there whose TOX value is 120 angstrom right and QSS which is the oxide charge density
is 2 into 10 to the power 10 per centimeter square and my threshold voltage the now my
the question which I am giving it to you is that I want to design a threshold voltage
such that the value is equal to 0.03 volts approximately how will you design it that
is the question.
So I have been given a P type MOSFET with obviously an N type substrate TOX is already
given to you I know the value of QSS prime you are told to get the value of VT threshold
voltage of the device.
So let us assume that I take a P plus poly gate at the MOS right now in the P plus poly
gate MOS let us consider ND is also to be equal to 10 to the power 17 per centimeter
cube and phi MS let us assume that for a P plus poly gate my phi MS value is 0.1 this
all will be given to you 1.1 volts approximately.
So what I get from here is that the phi FP is nothing but KT by Q Ln of Na by Ni right
and this comes out to be 0.0407 volts right this is phi FP which is the difference between
what EFI and EF the difference between them is 0.047.
So this has to be so when this goes to 0.814 which is twice the value of phi FP then only
I get an inversion right so this is you get EFI minus EFP as the value here right.
So now I need to find out the value of XDT, XDT is the thickness of the depletion layer
and this comes from the PN junction diode theory that this will be phi Fn divided by
Q times ND to the power half and if you put the values I will get 1.026 into 10 to the
power minus 8 centimeter.
Once you know the value of XDT you can now extract the value of QSD prime max right and
this comes out to be what QXDT right into Na or ND in this case and if you solve it
I will get 1.642 into 10 to the power minus 7 coulomb per centimeter square.
Now what I am supposed to do take these values put into the expression of VTP threshold voltage.
So let me put it in the threshold voltage expression and I will get VTP to be equals
to minus of QSD prime max please understand I have put a minus sign and you know understand
the reason why is it minus because it is not a P type it is a P type MOSFET.
So it will be always negative in nature right for this to be true minus QSS prime right
minus into T ox by epsilon ox right plus phi MS plus 2 phi FP if you put all these values
have been calculated and if you put all these values I will get VTP to be approximately
equals to minus 0.296 equal to 0.296 which can be approximated as minus 0.3 volts right.
So what I have shown you in this small numerical here is how can you design specific threshold
voltage device right from the notations or from the input values as far as phi MS is
concerned and ND is concerned it will be given to you by the designer.
So you need to find out the value of VTP in this manner right.
So with this let me end today's lecture this lecture and then we will start the next
lecture subsequently, thank you.
