Generate negative resistance for AC signal
so in IMPATT diode for a typical let us say P plus I N plus it is kept under reverse biased
condition first of all so positive side will be connected to N plus region and electrons
they will be accelerated and move towards the N plus region and this N plus region for
this typical P plus I N plus device it is used for electron holes pair generation or
for that avalanche impact effect So once electron hole pairs is generated then holes they will
drift to I region and they will be collected by the P plus region Now there is a typical
transit time to collect these holes since these holes it takes finite time to travel
this I region So this transit times it depends on the length of I region Now this picture
it shows a typical housing of IMPATT diode usually IMPATT diode itself it would provide
capacitive reactance and to nullify that we need some inductive reactance It is usually obtained by using a bonding
wire as shown by this magenta line here and altogether they sit inside a cavity So you
can see in right hand side picture how it looks from outside we have arrangement for
external biasing and we have the RF output from this end now how this chain reaction
it continues let us consider we have biased the IMPATT diode by using a DC source and
it is just near to the breakdown voltage Now we are sending some RF signal through the
IMPATT diode So in positive half of this RF signal it process that V th required for the
breakdown so in the positive half cycle of this super imposed AC breakdown will start
but it is delayed by a finite time So if I look at this picture it shows the applied
AC voltage versus Omega t and in the 2nd picture it shows the electron hole pairs generation
and the current due to that So when the AC voltage it crosses this 0 point
it is positive and the diode it is in breakdown mode so slowly electron hole pair generation
it starts and it continues And when the applied AC voltage it crosses the point Pie just after
this it is this total superimposed AC plus DC it is just below (requi) just below the
voltage that required for the breakdown So then the electrons hole pair generation it
continues throughout this positive half cycle and it reaches a maximum point when this applied
AC voltage it crosses this high point And now after that it does not instantaneously
stop it even after that it continues to some extent and due to that we have some more electron
holes pair generation So at the negative most point of this AC it
stops completely And now once these electron holes pairs are generated these holes they
drift through this I region and it will take finite time to reach this P plus region So
if I look at the external current component it would be exactly in opposite phase and
usually the length of the I region is so made that it will provide us another Pie by 2 phase
shift or the current generated due to this avalanche break down it is exactly in opposite
phase to the applied RF signal So the RF signal then it can absorb energy from the avalanche
breakdown and that is how we can use IMPATT diode as an amplifier So this right hand side
picture it shows the equivalent circuit of that IMPATT diode in addition to the housing
so we have the capacitive diode effect and the inductive (wa) wire bonding and also the
effect of packaging it is modelled by a parasitic resistor and inductor So diode terminal resistance it depends on
the series resistance passive series resistance of the in active region or I region we can
say V d the carrier drift velocity through I L length of the drift space charge region
so I region it plays an important role its length its resistance and the carrier drift
velocity through it Next A this is the diode cross section e s this semiconductor dielectric
permittivity and Theta is the transit (an) angle So transit angle it has one important
relationship with R R depends on transit angle So here we have a plot of negative resistance
versus transit angle transit angle is simply Omega t where t can be given by L by V d L
is the length of the intrinsic region So we see that negative resistance it is a function
of transit angle and it is maximum when transit angle is equal to Pie so it has almost periodic
nature with decreasing amplitude at transit angle equal to 0 and twice Pie it has minimum
value These are photographs of IMPATT diode typically used at 90 GHz band Next is field effect transistor MOSFET typically
so in low frequency applications we already know the equivalent circuit of MOSFET it is
a unipolar device and the device characteristics is controlled by the applied electric field
But the problem is due to the capacitance high capacitance value used as the oxide layer
for the gate So we have to consider this capacitance effect at the high frequency and the high
frequency operation is mainly limited due to the capacitor and not only that inside
MOSFET we have N type material and P type material so it is associated with P N junction Now carrier mobility is highest in intrinsic
material Now if we introduce some doping so that means this foreign atoms they behave
as a scattered while carriers like electrons and holes they move through this doped material
they will be scattered by this foreign agent So if we increase doping concentration obviously
carrier velocity will decrease So that is why inside P type or N type material carrier
velocity will be lower than that inside intrinsic layer So we will see there is another modification
of conventional MOSFET where we avoid this P type or N type material and we use simply
intrinsic material for realisation of the channel and that is called the high electron
mobility transistor or HEMT Let us 1st discuss the basic operation of MOSFET and why its
high frequency operation is very much limited So if I look at the characteristics transfer
characteristics and output characteristics you will see D s drain to source current it
is a function of gate to source voltage so we can control the drain current by controlling
gate to source voltage and if we look at the output characteristics V ds it is if I increase
V ds there is almost no change in I ds so after this pinch off effect or in saturation
region we can represent MOSFET by a current source and this current value it can be controlled
by gate to source voltage And below this region it behaves as a linear device or we can replace
the drain to source by a resistance So this is a typical MOSFET device you can
see below is the body and we have source drain over that we have a thin oxide layer and on
top of this oxide layer we have gate layer made of metal or poly silicon and if I now
look at the frequency versus the (F) F t and F max realised F t and F max with years so
in recent years there are conventional MOSFETs they have been developed which for which the
F t value is as high as 500 or 600 gigahertz So for that what we have to do to increase
F t we have to decrease the transit time just like BJT so we have to reduce the gate length
now even 10 nanometre technology is possible and we have to increase the carrier mobility
so that transit time it decreases So for normal operation so I will just a few
basic points so that we can go to HEMT and we can discuss the problems of basic MOSFET
So when we do not apply any external biasing in that case if we consider a N MOS it starts
with P type substrate and we have N plus source region and N plus drain region and since it
involves PN 2 back to back PN junctions at normal condition we do not have any current
flow from source to drain Now let us consider a situation we have applied some drain to
source voltage but it is less than V gs minus V th V th is the threshold voltage that needed
to realise the channel in between source and drain So in this case one small inversion
layer is created and we have a thin contacting layer here and we can represent the device
by a resistance or we have this linear region of this curve Now if I keep on increasing the drain to source
voltage for a given gate to source voltage then what will happen we have this pinch off
condition so we have this pinch off condition So one once this pinch off condition is achieved
when V ds is equal to V gs minus V th this device it behaves like a current source So
after that if we keep on increasing the drain to source voltage the device current I d it
becomes almost constant So this figure shows the cut off frequency
in terahertz versus Gate length this is the theoretical plot for double gate MOSFET and
single gate MOSFET in different technology So in 10 nanometre double gate or dual gate
MOSFET technology you see the typical value is 3 terahertz cut off frequency whereas in
20 nanometre single gate MOSFET technology gate length is 20 nanometre so typical cut
off frequency it is 1 terahertz using conventional MOSFET Now when we go for uhh amplifier application
in that case the thumb rule is if Ft is given we can easily use the device to Ft by 10 frequency
points Now you look at the electric field plot between the gate source and through oxide
layer this electric field it is associated with capacitance So we have a capacitor between gate and source
which can be represented by C gs and we have another capacitor between gate and drain which
is represented by C gd and in addition to that we have body effect So in the equivalent
circuit then for high frequency operation we have to add all these capacitors one between
gate and source one between gate and drain so this capacitors resistors inside the rectangular
box it shows the intrinsic device property In addition to that due to packaging and some
other effects we have some additional capacitors and that is sometimes called the extrinsic
model of the device So because of these capacitors the gain of the device it becomes a function
of (proper) function of frequency And the frequency at which this gain becomes unity
current gain typically we call Ft of the device Now for high frequency operation one thing
can be done that we can decrease the treat capacitance So instead of oxide layer there
is one type of MOSFET where Schottky junction is used simply metal semiconductor we call
it MESFET So you can see the picture here so just below gate we have N plus gallium
arsenide layer so there is no oxide layer and it will give lower capacitance and we
can increase F t of the device but it is not very popular it has some other problems So
next we will move to a modified version of MOSFET we call it high electron mobility transistor
So as I discussed that if we keep on increasing the doping concentration then the carrier
velocity it will decrease due to the scattering effect of this region which are being used
for doping So in a HEMT device typically it is a hetero junction device two different
types of semiconductors are used to realise this type of device Here we use one intrinsic layer and where
channel it will be completely inside the intrinsic layer so that the carriers typically electrons
since electrons mobility is higher will be will prefer electrons as the carrier it is
much higher inside intrinsic layer compared to that doping layer and that is how we can
improve the mobility of the carriers and that is why the name is high electron mobility
transistor So some characteristics this is a hetero structure field effect transistor
it incorporates a junction between 2 materials with different band gaps so that is why it
is called a hetero junction as the channel Commonly used materials are gallium arsenide
with aluminium gallium arsenide indium it can give high better high frequency performance
and gallium nitride it can give high power performance So in the right hand side picture you can
see a schematic diagram of a HEMT So below gate we have thin oxide layer and below that
we have N type aluminium gallium nitride and below that we have simply gallium nitride
based carrier layer (Ga) gallium nitride based carrier electron layer So now how it operates
the electron concentration inside N type aluminium gallium nitride is much higher compared to
the second layer gallium nitride so because of that we have band bending effect and electronic
accumulation effect in the second layer So look at this band diagram aluminium gallium
nitride it is a wide band gap semiconductor whereas gallium nitride its band gap is lower
At room temperature the Fermi lever it aligns itself and now at this transition point this
band conduction band it bends itself for continuity and it bends deep inside the Fermi level So I am plotting the energy of conduction
band versus X so X it started from the top layer and as X increases we are going inside
gallium nitride so just at this transition inside gallium nitride the conduction band
bends below Fermi level and electrons in the conduction band of aluminium gallium nitride
it sees a lower energy state unoccupied inside gallium nitride So it will move to gallium
nitride and it will be occupied by electrons So you can see inside the device then we have
electrons diffusion initially from this N type material to gallium nitride and because
of that we have a very thin layer of electrons and it will form a small depletion region
here and this will uhh stop any further electron depletion due to the built in field of this
depletion region The thickness of this electron layer is typically very small just maybe 10
of Armstrong and sometimes because of that we call it is a 2 dimensional electron gas Now when some electric field is applied so
we have (betwe) electric field between source and drain then mostly carrier movement it
is inside the gallium nitride layer not inside the N type layer so that is how we can increase
the velocity because of the increased carrier mobility inside un-doped gallium nitride and
that is why it is called high electron mobility transistor or HEMT
So it is being widely used at uhh for different millimetre wave products for satellite television
receivers radar equipments et cetera Now when we use two different types of materials they
simply cannot be placed one after another one there is some problem due to their lattice
different lattice structure When we use 2 different weighted semiconductor materials
their lattice constant should be very close to each other otherwise it will provide some
loading effect and because of that electron mobility again it will decrease So there are 2 techniques to avoid that usually
a very thin layer of 1 semiconductor material is used on another semiconductor material
so that that thin layer of semiconductor material it close it stretches itself and closely follow
the lattice formation of the second material that is how we can obtain lattice matching
and the HEMT due to that we call pseudo morphic HEMT or PHEMT There is another technique where
another material is used a third material is used as a buffer layer between this N type
material and the intrinsic material This buffer layer is used to match the lattice constant
between these 2 layers and this is called uhh metamorphic HEMT or MHEMT so we have two
different types of HEMT devices PHEMT and MHEMT So this picture it shows HEMT DC characteristics
gate current we still have some leakage current gate current but typically it is very small
but with the increasing gate to source voltage it increases and of the order of fraction
of some micro ampere and look at the variation of I d versus gate to source voltage you can
compare with the conventional MOSFET At very high gate to source voltage it is almost constant
this is the measured characteristics and after that it follows nearly what we see for conventional
MOSFET And looking at the output characteristics it almost looks like a conventional MOSFET
device it has the linear region and the saturation region so in saturation region we can replace
the device as the current source Comparison of different solid state devices
available in market we are comparing breakdown voltage versus average cut off frequency in
gigahertz so they are available commercially and this yellow curve it shows devices based
on silicon and silicon germanium technology and this blue one it shows based on indium
phosphide so if I look at the different types of the the this is showing uhh high frequency
and high power high breakdown voltage mean So in most of the cases HBT is the winner
Indium phosphide based HBT is the winner So after this we will see one application of
active devices that is electronic switch but before that we will take a short break thank

Leave a Reply

Your email address will not be published. Required fields are marked *