Electronics Nanoelectronics Devices and Materials mod01lec01

welcome to this lecture series on nanoelectronics
ah which will include a lot of discussion on devices as well as materials this course
will be taught jointly by three faculty members i am professor navakanta bhat i will probably
cover about one third of these lecture series and subsequently two other colleagues of mine
professor shivashankar and professor k n bhat will also cover rest of the course we are
from centre for nano science and engineering which is a brand new department at the indian
institute of science bangalore and you can go to this website which gives a lot more
details about the centre and the capabilities of the centre and the vision of the centre
and so on and so forth and the building that you see out here is the new building which
is home to the centre for nano science and engineering it houses state of the art nanofabrication
facility ah and also ah very extensive nano characterization facility ok
so lets get started with the course ah just a couple of logistics first ah some of the
ah there is no text book really for this course however there are quite a few reference books
some of those are listed here and you know these consist of two things ah two broad areas
ah that is ah devices as i mentioned ah you know you can go through following series of
books ah depending on which topics that you are interested in fundamentals of modern vlsi
devices by yaun tour and ning ah cambridge university press solid state electron devices
by streetman and banerjee ah which is a very classic text book on semiconductor devices
especially for the basics of devices fundamentals of electron devices which is written by achutan
and bhat k n bhat who is also ah you know co instructor on this course ah that we have
ah then also there is one ah very important book on ah m o s technology which is metal
oxide semiconductor physics and technology by nicollian and brews which is also considered
as the bible for m o s technology and then you know there is there other set of books
on process technologies and materials ok ah which include the silicon vlsi technology
by plummer deal and griffin u l s i technology edited by s m sze mc graw hill publication
and also a very extensive ah you know ah topics on material characterization covered in encyclopedia
of material characterization published by elsevier right so as i mentioned ah in addition
to the reference book books that i have listed during the course ah we will also point to
ah certain publications which have appeared in the recent conferences and journals that
are also very relevant to understand the material that we are going to cover in this course
ok so ah just put things in perspective when
we talk of nanoelectronics right in our nano science of course is the most fundamental
aspect ah just as the macro counterpart at nano scale ah the nano science essentially
talks about you know studying phenomenon whether it is physical phenomenon chemical phenomenon
or biological phenomenon but most importantly at length scale which are less than hundred
nanometer right that is what we broadly classify as nanoscience now when we talk of nanotechnology
what we are saying again just analogous to the macro counterpart based on the understanding
that we gather ah after studying nanoscience we would like to make use of that understanding
and create useful products right that is what technologies all about the distinction here
is that the products are created with at least one building block right products will have
multiple building blocks in them any product for that matter at least one building block
should be at lengths scale which is less than hundred nanometer if that is the case
then you know we call that product as a nanotechnology product or nanotechnology enabled product
as ah most of you know there has been a very tremendous ah activity in nanoscience and
nanotechnology in the last few years all over the world and we can certainly say that nanoelectronics
which happens to be one of the branches of nanotechnology you see in nanotechnology we
can talk of nanomaterials nanobiotechnology nano electronics nanomechanics and so on and
so forth out of this ah i can argue that nanoelectronics is probably the most mature of nanotechnology
and in fact nanoelectronics is the most successful commercial manifestation of nanoscience and
nanotechnology right so when we are talking of nanoelectronics we use nanoelectronics
enabled products in our day to day life although we may not really know about that right and
the most important thing ah to ah look ahead is that nanoelectronics growth in future relies
on innovations in new materials and new device structures only then we can create new nanoelectronics
products in the future right so this course is essentially trying to address that aspect
right ah a very specific ah aspect of nanotechnology as i mentioned which is nanoelectronics and
to be able to go along the nanoelectronics path in future what do we need to do in terms
of materials technology and device technology ok and that is what we ah you know hope to
cover ah during this course just to get a feel for ah perspectives on length scale ah
what i have tried to do here is to sort of just suppose to different ah ah areas you
know one is of course biology that we are all made of right and electronics
that we create right as you know biology is of course based on carbon where as electronics
is entirely based on silicon right you know that is ah underlined distinction out here
this is carbon atom which is sort of ah of course you will have other atoms as well but
carbon is the most fundamental ah aspect out here whereas here in addition to silicon you
will also have various other things but know silicon is essentially the enabler right now
when we are talking of atomic scale right we are really talking of of the order of hundred
picometers right ah you know on the screen for some reasons these numbers are not coming
out properly but what you have here in terms of these horizontal lines out here this is
hundred meters this is hundred millimeter hundred micro meter hundred nanometer and
hundred picometer right three orders of magnitude decrease in length scale ok
now in this area out here which is less than hundred nanometer right this is the bar that
we have you know in biology we have of course in addition to carbon atom then the building
block of life which is n a and cell nucleus all this essentially fall in this domain right
and similarly in the electronics counterpart ah you know in addition to the silicon atom
you have the gate oxide which falls in this regime and the gate lengths that we talk about
today are also in this regime right using this fundamental constructs that we have you
know we then construct ah higher level abstractions right ah in electronics of course we do that
but in biology nature does that right we don’t still don’t know how to do that right but
you know the the idea here is that then we talk of red blood cells human arteries human
hair you know this is a very classic example that
was given when we were talking about micro electronics right just to give ah the length
perspective here as you know less than hundred nanometer i already told you is nanotechnology
right if you are less than hundred micro meter but more than hundred nanometer then you are
really in the regime of micro technology in general and if that micro technology applies
to electronics then you know we are doing micro electronics right so when we started
doing micro electronics way back in ah fifties and sixties the classic example that was given
was that we build transistors which are of the size less than diameter of the human hair
right diameter of the human hair is about hundred micro meter right then you know between
hundred micrometer to hundred millimeter you have mesos scale technologies right ah
you know if you talk about electronics you know if you look at cache memory size or a
micro processor dye or even a package chip right that will probably fall in this ah kind
of length scale that we are talking about right ah then of course you know using this
we build a a variety of electronics product right ah you know so this is the big distinction
right you know i have also sort of indicated here biology has really figured out how to
build the systems which are also you know ah to the environment meaning that you know
once ah the living organism ceases to live right you know its bio degradable right whereas
you know disposing electronics waste has become one of the biggest challenge today right now
these are some of the things ah you know we have still not being able figure how although
we have being saying that today we can create a computer which can be as as powerful as
a as a human brain which of course is a very degradable point right but none the less ah
we have not being able to do it with the efficiency that ah biology has been doing right so in
other words the point to also understand here is that there are two two important messages
that i would like to drive home here right in future especially in the era of the nanotechnology
right we there is a there is going to be a very strong interaction between biology and
electronics on one hand we would like to build electronics which we would call as bio inspired
electronics really trying to mimic how the brain does computation right i mean ah the
way we do digital computation is quite different from ah how biological systems do computation
right so that is one aspect right related to that
aspect is also the the point of ah you know self feeling architectures you know redundancy
in electronics and so on and so forth which is to sort of learn from biology right on
other hand i think one of the most important ah you know application area for electronics
is really to apply electronics in biology in the next ah couple of decades right in
terms of really building ah senses for biological application diagnostic tools ah you know so
called lab on chip devices and so on so forth right so although you know when we started
the discussion on this slide we said that you know these two system seem to be distinct
right but there is going to a lot of synergy going forward among these two so ah seemingly
which appear quite different right I think they will have to co exist and in
fact that is why there is so much interest in nanoelectronics and nanotechnology because
toady we have the capability of building transistors or electronic components which are at the
same scale as that of the building blocks of life right and that is where we can actually
start interfacing the electronics and biology at their fundamental ah length scale right
so ah you know you need to keep this in mind although in the course we are not going to
cover anything on application of electronics in biology or bio inspired electronics and
so on and so forth these are very good research topics to look forward to we will only focus
on the silicon cmos technology and also some other materials which are going to compliment
the silicon technology right so that is going to be the emphasis of ah this course right
ok so lets move on then so what is a big picture of electronics right
ah you know electronics is very distinct right i mean you compare electronics industry with
any other industry for example automobile right ah what is distinct about electronics
as i already pointed out this is the only industry where you can create intelligence
you see we create intelligence on silicon right and that is why electronics is is really
an enabling ah manufacturing field right in fact it is predicted that if you look at a
high and automobile today most of the cost i would say easily more than fifty percent
of the cost is really in electronics because its this electronics which brings intelligence
to automobile right i mean same is true with any other industrial segment in other words
electronics industry has a far reaching impact on various other industrial segments that
we have right and there is a very interesting survey conducted
by ah ah body in european union a while ago which said that the electronics industry growth
in the last few years was twice as much as world g d p right just imagine that right
you know that is the kind of growth that we have in electronics right so how is it enabled
the big picture of electronics is that we construct these chips which will do either
ah you know computation as is indicated here or storage right you know any field that you
would like to look at there is lot of information that you gather
so you need to store information and you need to process information right the electronic
chips enable us to do that when we talk of computation it could be a microprocessor or
a micro controller or a d s p chip or a network processors and so on and so forth right all
these essentially take in digital inputs and digit bits and manipulate this bits right
do lot of computation on this bits having done that you need to store the result right
i mean then you need storage elements the storage elements could be a static random
access memory or a dynamic random access memory flash memory and so on and so forth right
so when we talk of a device the device essentially stores and manipulates the computational state
variable today most of the time we are essentially talking of digital electronics right so ah
essentially it’s a binary ah digital logic that we built till today so you know computational
variable has two states ah and we manipulate this computational state variable ok computational
state variable can take one of several forms but today it essentially takes the form of
voltage which comes about because we store charge on the circuit nodes that we have in
cmos circuits right and ah you know this is how we build electronics right so any electronics
chip that you have no matter what chip it is fundamentally it either does computation
or it does storage ok now if you go an put an electronic chip under
a microscope under an optical microscope this is what you will see ok ah you will see a
regular patterns on an electron i mean on optical microscope ok as is indicated here
if it is a chip in this example i have taken intel pentium chip you will have you may not
know that unless you are a designer but the one person who has designed it would know
that ok there is a data cashier you know there is a floating point unit here and so on and
so forth right so this is what you will see when you look at the plan view or a top view
and it doesn’t matter which chip you look at you know
this is another example of one of the chip that we have designed at i s e which was a
neuralnet chip again you see some regular arrangement of some building blocks right
and in other words integrated circuit chip is really silicon floor planning just as when
you build a house you do floor planning right ah to begin with right similarly you need
to decide how many blocks should be there on your silicon chip where should a blocks
should be placed and so on and so forth this is the task that is done by a chip designer
as far as the subject matter of this course is concerned we are not going to talk about
chip design ok instead what we will do is really to ask the question what is inside
the chip ok I mean you got that dye dye is essentially
a piece of silicon on which you have made the chip right you have put it under a op
optical microscope but if you take a cross section now you cut this silicon piece and
take a side view what will you see right what you will see is essentially a very large ensemble
of transistors which are connected appropriately right this is a key depending on how you connect
them how you interconnect them you get a particular function right you may get a memory function
you may get a microprocessor function and so on and so forth right and in order to do
that you will also have large number of metal interconnection lines ok
so when you do the cross section at a very top right ah you probably will see a multilevel
metal interconnects today state of the art chip can easily have more than ten layers
of metal lines right for example this is the top metal line and underneath there will be
you know another metal line and so on and so forth and eventually you will reach silicon
on silicon you have a transistor and this is what is your transistor right and today
we are only talking about m o s transistor when we talk of cmos its complementary metal
oxide semi conductor transistor if you look at any electronics today more than ninety
percent of the electronics that is built today is based on cmos technology a very very small
fraction ok may be less than five percent is by bipolar junction technology either on
silicon or mesprits on gallium arsenide and so on so forth right
so majority of the silicon electronics that you see is cmos and hence the entire focus
of this course is going to be on cmos ok and this is what you will see at the very bottom
on silicon right you will have constructed a silicon m o s transistor which will have
you know a source terminal inside silicon a drain terminal inside silicon and a gate
oxide we will see a blown up version a little later and a gate electrode right it is essentially
a three terminal device well strictly speaking it’s a four terminal device there is also
a body terminal but we will ignore that for the time being right
so todays modern chip will have close to billion such transistors on a very small silicon area
silicon area could be one centimeter by one centimeter in other words you will have you
know a square silicon dye about one centimeter by one centimeter and inside this you would
have put close to billion transistors just amazing population density of the transistors
right on silicon ok if you have billion transistor on such a tiny area they need to be interconnected
right and that is why we need multilevel interconnects right otherwise i cant pack them in one centimeter
by one centimeter if i am given only three layers of metal interconnect maybe this chip
will have to be five centimeter by five centimeter or whatever that is right
so that is why over the generation of the silicon technology we have also increased
the number of metal interconnects and you know this is what you see in fact you know
if you have an artistic ah view you know in fact somebody had remarked that today we are
building cities on silicon right these look a flyovers [laughter] that you see in modern
cities today ok so this is a kind of ah things that you will see inside a silicon right and
the entire focus of this course is only on transistor we will not even talk about interconnects
right so that is how focused this course will be right the next ah forty five lectures or
so ah maybe we will you know have a brief ah reference to interconnects in one of the
lecture but that is not going to be a theme right the reason simply is that you see in
silicon chips transistor is the active element interconnects are passive elements right it
is active element which is doing all the functions right it doesn’t mean that the passive elements
are not important but you know it is very important to design the active element right
ok and that is why you know we will only focus on the transistor right ok so as i mentioned
metal oxide semi conductor field effect transistor right
this is what we are going to look at right a very simple model for a transistor is you
know you can abstract it as a switch right so there is a controller input that you have
in a switch right and a signal will pass from input to the output if the control input has
the right condition then the switch could be closed and then you have input signal propagating
to the output otherwise switch is open you know it’s a very very abstraction ok when
we talk of doing circuit we come up with what is called a schematic abstraction of a transistor
right so this is how you will represent a schematic
of a n channel m o s transistor it will have a gate terminal it will have a drain terminal
which will drain the carriers it will have a source terminal which will source the carriers
right the carriers could be electrons and holes and accordingly you will have two flavors
of transistors we will see in a minute look at the top view as i mentioned already this
is what you see you know some regular arrangement of rectangles or you know some structures
right this is what you see but what we are really interested is cross section ok so as
i mentioned you will have n channel transistor or n mosfet and p channel transistor and if
you build chips combining n and p channel transistor and that is your complementary
metal oxide semi conductor technology right and
you know your n channel transistor will be built in a p well which will have n plus drain
and plus source and a gate insulator which is what we call oxide and a gate terminal
and similarly you have a complement of this and that is why the name complementary metal
oxide semi conductor wherever you have you p n you have replaced that with p you have
p plus in fact these will typically be represented as n plus n plus which means they are very
heavily doped n region ok n plus doesn’t mean that you know n is of course electrons are
majority carriers electrons do not have positive charge right electrons have negative charge
plus here indicates the doping concentration ok very heavily doped ah n plus ah regions
out here and here you have very heavily doped p regions right a source and drain region
and insulator is same n well they a built on n well and a gate insulator and
you know they have complementary characteristics right with respect to the voltage that you
apply to the gate this is what we call gate voltage right v g you will have the transistor
conducting very very little current ideally zero current when it is off but it is never
the case as becomes as it becomes clear later on in the course it will have very small non
zero current and if you apply the appropriate voltage you know n channel transistor will
turn on right so this is off state for n channel transistor and this is on state for an n channel
transistor but p is exactly complement of this right when n is off p will be on and
that is why it is conducting large current out there and when n is on p will be off correct
so this is how we build ah you know a circuits based on the ah ah transistors right so in
other words ah if you look at the integrated circuit hierarchy right we can classify broadly
that there is a technology domain which enables the construction of the building blocks of
chips transistors and interconnects that is where you need to have very good knowledge
of device physics only if you know device physics you can come up with very innovative
designs of transistors and having gotten that design you need to build that that is where
you need appropriate process technology you need to use appropriate materials and sequence
of processes to build that transistor this is what we call a technology domain
if you look at any foundries silicon foundries that is not mechanical foundries they essentially
do this they do silicon chip fabrication they focus on device physics and process technology
ok whereas when we talk of application domain this is where the designers system designers
and chip designers are working on they talk about what kind of system architecture they
want to build for example if it’s a micro processor you know you know what kind of architecture
you know should it be cisc architecture or a risc architecture things like that if you
are building an a d c what kind of an a d c analog to digital convertor that you want
to build successive approximation or flash a to d convertor
so this kind of system architecting happens here and once you do the system architecting
you do circuit design right at the fundamental building block which is transistor you need
to architect the circuit design the circuit and do the lay out only when you do the layout
you complete your design then pass the design onto the foundry then a foundry can the design
run through all the processes and build your chip right so this is really an interdisciplinary
area you see you need lot of understanding of ah materials chemistry out here good understanding
of physics good understanding of circuits computer science or whatever that application
domain is and of course mathematics to develop models computated design is key here right
we do lot of cad both in the technology domain as well as ah you know circuit ah domain right
so today as i mentioned our process technology is based on silicon our devices are cmos complementary
metal oxide semi conductor and our computational variable as i said charge and hence we are
not going to talk anything about for example using spin as ah logic variable right so called
the area of spintronics right that is not at all going to be discussed in this course
right and of course we build circuits today based on boolean logic digital architecture
this is where the point i mentioned earlier that bio inspired circuits could be quite
different from you know digital architecture and boolean logic right ah but again we will
not really discuss ah too much about that we need to really do optimization across all
these domains only then we can get a very high performance integrated circuit right
that is the key hm otherwise we will not be able to ah do that
any discussion on ah you know scaling and nanoelectronics is incomplete in my opinion
without reference to moores law ok you you might have heard about this moores law but
let me just walk you through that gordon moore ah published a paper way back in nineteen
seventy five nineteen sixty five i am sorry this was also published in a trade journal
called electronics ok the paper was titled cramming more components onto integrated circuit
that was the title of the paper it’s a very interesting paper easy to read you know you
just google it you will certainly get this paper on the internet right ah what he did
there are lot of things that he discusses but there is one very interesting the thing
that he did right he constructed a graph right which is shown out here hm what he did in
this graph on x axis he put the year time line is on x axis on y axis he put the number
of components per integrated ah function ok ah essentially what it meant means is that
by the way this graph first of all is plotted on a semi log sheet ok
so this is log scale here ah y axis is a log scale ok and x axis is a linear scale all
he did was he took integrated circuit chip introduced in any given year and counted how
many transistors where integrated on that chip on that given year right so number of
transistor is plotted on y axis ok and ah year on the x axis and he didn’t have too
many data points you see the first integrated circuit was invented way back in nineteen
fifty eight if you may recall the first i c was invented by jack kilby at texas instruments
that was in nineteen fifty eight right the first semi conductor transistor itself was
way back in nineteen forty seven right which was a bipolar junction transistor not a m
o s transistor although in this course we only talk about
m o s transistor that’s an irony right the first semi conductor transistor was indeed
bipolar junction transistor although you know the scientists at bell labs were really trying
to make a m o s transistor but making a m o s transistor was so difficult they ended
up doing a bipolar junction transistor right but that apart the data points are only from
nineteen fifty eight till the year nineteen sixty five when wrote this particular paper
right so you know he has about five data points here he notice that it fits a straight line
right in other words because it’s a semi log sheet it indicated that the number of transistors
that were put on the chips were increasing exponentially over the time in particular
at least during that period the number of components or transistor was doubling every
year ok but the lot of credit is given to moore because of this dotted line that he
has in this paper you see ah you know this is not ah based on any data
point you see this was written in nineteen sixty five obviously no chip was done in subsequent
years at that time the point he made here was that this will continue to happen in the
years to come hm that was a very very bold prediction ok ah at least what he said was
that he doesn’t see any fundamental ah impediments from science ok its only the limitations of
technology if any we are not able to construct that but otherwise we should be able to increase
the number of transistors exponentially over the years right and indeed subsequently there
was a paper published in i triple e spectrum in nineteen seventy right ah again it’s a
very interesting paper you can go back and read that and
what is done here is that now there are few more data points on x axis because this was
ah you know published the subsequently right and what you see is that these are the original
data points which were published in the moores paper and you see that all subsequent data
points are also on the same line right it has continued to happen hm in fact moore wrote
in his handwriting out here on you know sometime in nineteen ninety six you know a very interesting
ah he says the ah the definition of the moores law has come to refer to almost anything related
to the semi conductor industry make a note of it almost anything related to semi conductor
industry right that when plotted in a semi log paper is a straight line right [laughter]
right so you see moores law is not a fundamental science law its not a law in physics for example
you know it’s a it’s a very very bold prediction that was made by moore you take feature size
or the dimension of the transistor plot it as a function of year meaning this is year
and this is the transistors size lets say length
of the transistor if this is log scale we see that over the years the length of the
transistor has decreased exponentially any anything that in semi conductor industry you
know goes according to this law right but you see the key here is really this because
we are able to miniaturize the transistors because we are able to miniaturize this this
will happen if you had not done this there is no way we can cram more components on to
integrated circuit right this key here the message here is scaling we need to scale technology
in fact we will have more discussion in one of the future lecture why do we need to scale
in the first place apart from putting more transistor in a given real estate of silicon
there are various other advantages of scaling as well we will discuss a lot more of that
in one of the future lectures ok of course moore not only did this prediction but he
also proved his vision you nee you see moore co founded intel on of the technological giants
in electronics right gordon moore and ah robert noyce co founded intel way back and of course
you know ah because micro processor happened then of course you know there was an explosive
growth of electronics right before that we had only memories memory chips were there
of course right but memory chips were only for storage the real intelligence comes that
is information processing comes because of micro processor or any of these logic chips
ok and very interesting point i always tell my students is that you know gordon moore
actually had a degree in chemistry ok although he went on to lead a technological giant in
electronics industry right so there is really no barriers in ah science
and technology right if you have you know ah real ah liking for any field you know pick
up that at point in time of course you need to have perseverance and interest and you
you should really want it right only then you can get it ok ok so this is a just a same
data ah just to sort of ah highlight in from what is called international technology road
map for semi conductor ah this is ah an authoritative body i t r s as is indicated here again you
know you can google ah i t r s you get lot of information right ah at least i would recommend
everybody who is interested in nanoelectronics to really to take a look at i t r s website
it has lot of interesting technical information out there what i t r s does is to look at
the history and make perdition for future make perdition
for future based on not only historical trends but also based on a fairly good understanding
of current capabilities and where is science heading in future right its not like kite
flying future prediction right right the future prediction are done ah very consistently but
more importantly i t r s revises its technology road map every year right you go to i t r
s website and you can take a look at all the previous ah technology road maps right only
by having that yearly revision and upgrade you know you can have these road maps correct
right but what is plotted here again on x axis is year of production ok going form nineteen
ninety five the prediction is of course all the way to twenty twenty ok today we are somewhere
around ah twenty twelve hm what is plotted on y axis are multiple things ok a product
functions per chip is what it broadly indicates but what is plotted is flash memory number
of bits that you can make on a chip and ah if it is a single level cell in flash memory
you can also build what is called a multilevel cell if it is a multilevel cell chip how many
bits can you put in a chip if it is a microprocessor unit right then how many transistors that
you can put on a chip ok and again there are different kinds of microprocessor called high
performance microprocessor low power microprocessor different flavors ok this is also a different
flavor of a microprocessor if it is a dram how many bits can you put can i make sixteen
g b dram thirty two g b dram and so on and so forth so this line right essentially says
that all this products their capabilities are increasing exponentially over the years
ok and this is happening because if miniaturization right that is the key otherwise you know we
wouldn’t be able to do this yes this is the point right now the previous
comes about because of this again on the x axis i have the years and on the y axis ah
you know what we call as a feature size ok there are multiple things that are plotted
here called product half pitch gate length and so on and so forth again for multiple
products for dram for microprocessor for flash and so on and so forth all of them have the
same trend the feature size is decreasing over the calendar year of course the the graph
that you have at the at the very bottom is essentially the gate length you see this is
the gate length ah the gate length of the transistor that is the lowest dimension using
this you build other constructs right pitch will be half pitch will be little more than
the gate length and so on and so forth ok we will talk more about that later now
this point here is an important point i want you to look at and in fact that is where it
says nanotechnology era begins in electronics when did it begin it began in nineteen ninety
nine not today ok it began in nineteen ninety nine which is almost thirteen years ago today
we are talking of two thousand twelve right although you know there was not much fan fare
right because electronics for electronics it was just a evolution you know you are doing
little more than a hundred nano meter you cross that hundred nano meter barrier right
so whats the big deal kind of right so ah there was not much ah fan fare about saying
oh i have a nanotechnology product right but on the fact that now we have so much hype
on nanotechnology its not just hype you see we have been using nanotechnology product
at least nao the building nanometer have been used in chips since nineteen ninety nine ok
and of course we are continuing to scale this and very soon of course we will also cross
the ten nanometer barrier ok that is how small we will be building the
transistor ok ok so in other words what have it done we have taken a m o s transistor which
is the building block remember that source drain a well or a substrate a gate insulator
and a gate electrode ok and made it smaller and smaller and smaller over the years hm
this is all silicon cmos right as i mentioned the first transistor was not cmos b j t and
first few years right in fifties ah it we were essentially building chips based on bipolar
junction transistor the m o s transistor was actually invented although concept was very
well known way back in thirties and twenties it was actually made possible only in nineteen
sixty four ok and then of course in nineteen sixties once we made m o s transistors the
technology was just silicon cmos it just displaced bipolar junction technology but we took the
fundamental building block and just made it smaller and smaller that is how we have ah
been doing the business in the last ah a few years
but today we are facing some very fundamental challenges in making this transistor smaller
in future because we are really hitting some fundamental limit there is lot of gate tunneling
current w use to say that the insulated gate m o s transistor has an insulated gate if
i apply voltage on the gate no current flows through the gate terminal because there is
an insulator that’s no longer true now ok this causes lot of problems we need to overcome
that problem there is lot of issues with channel quantization ok the channel here between source
and drain is so small ok that the quantization effects cannot be neglected and the transport
itself is no longer drift diffusion based transport right it is becoming quasi ballistic
transport right so if we have to scale in the future miniaturize
the transistor in the future we can do that only if you have new materials and new structures
ok otherwise you know there is no way we can go down to fifteen nanometer kind of dimensions
right now ah there is ah one important point ah that i want to bring up here is something
called technology life cycle right this also gives you a feel ah the place of electronics
right you know at least ah from the point of view of ah using electronics products right
you see the best way to best time to buy electronics product is always tomorrow right if you want
to buy a cell phone you wait till tomorrow because you get a better cell phone with better
features for a lower cost right you will buy a cell phone today and your friend will buy
one six months down the line and you will curse yourself why did i buy this right six
months ago right you know technology is becoming obsolete so fast right
if you look at the history the way things are happening is that a new technology cycles
start emerging at time line this lets say reference zero is reference x axis is month
it takes about two years develop a new technology during this period you know you are working
on your equipments you are working on your processes you are working on your device structures
and so on and so forth but there is no product in the market really however you will be seeing
lot of papers published in the conference leading conferences and journals right and
very quickly you figure out how to make a product using this because making a single
transistor is one thing making billion transistor all of which are working properly and interconnect
them appropriately is a marvelous feet right that is where you need to you know have lot
of ah learning right that comes about and you actually start making the products right
you start selling chips what is plotted here is volume parts per month ok and here parts
here is chips per month and here wafers per month right you know twenty wafers can give
you ten thousand chips right this is ten k here this is two twenty two hundred and so
on and so forth the scales are different ok because each wafer can support a large number
of chips you see ok so then of course the production ramps up
and you really have large number of chips porting to a new technology at the same time
in the r and d there will be a new wave that will start right you start working on a new
generation of the technology you know its its sort of ongoing right and that is why
you have ah continuous ah miniaturization introduction of new technology nodes and so
on and so forth right and that is a very unique thing about electronics ah unparalleled compared
to any other industrial segments ah this is the only industrial segment where
the cost of your product is decreasing every year with more functionality right so what
is the projections for the next few years right this is taken from i t r s road map
ah as i mentioned you should go and look at the i t r s road map ok you know today we
have ah something like a forty five and thirty two nanometer technology in volume production
ok ah you know in future we will of course go down to something like eleven nanometer
ah technology node ok there are multiple matrix here technology node is quite different there
is also something called transistor gate length right when i am building a chips on so called
forty five nanometer technology the transistor dimension may actually be much less than that
you know i may be making transistor whose gate length is as small as twenty nanometer
right there is a big difference between the so called technology label ah which is derived
based on a definition of half pitch as the actual very much smaller right when you have
the gate dimension and spacing equivalent then the gate length becomes equal to half
pitch ok whereas if the gate length is much smaller and di spacing is much larger than
half pitch is more than gate length you see there is a distinction between the technology
node definition which is derived from the half pitch that is you have one transistor
here you have another transistor here so this is the pitch right hm how closely you can
pack the transistors ok and if the spacing and dimensions were to be identical half pitch
and gate length would be identical but that is not the case right we miniaturize the gate
length more aggressively we can pack the transistors together ok
so that is why there is a difference out here wafer diameters today we are building silicon
chips on twelve inch wafer diameter right these are like you know big dinner place that
you have twelve inch wafer diameter reason being that if you have twelve inch wafer you
can populate large number of chips right that is how you know you can reduce the cost of
chip you will have so many chips from a single wafer ok and of course in order to do that
you have to go through a very complex process steps you will probably go through at least
ah thirty five photolithography steps in a nanofab facility ok the number of transistor
that you are talking about is about ah two billion transistors that you would integrate
ok and ah you know number of interconnects as i mentioned you know something like more
than ten that i already mentioned because you have so many transistor you need to be
able to interconnect them right that is why you need large number of interconnects right
and ah of course toady we are talking of ultra low voltage operation right about one volt
is what the supply voltage for all these chips they are no longer you know your five volt
chips you know at if you apply five volt the transistor will dry instantaneously ok
so the supply voltage has to be very very low of one volt ok and of course you know
very large number of ah i packaged a pins right you can never remember the pin out of
the chip right fourteen fifty package pins right very interestingly out of this fourteen
fifty at least for microprocessor kind of chip at least two thirds of the these pins
are only v d d and ground pins right that is required ah to bring in supply into the
chip from multiple ah a regions so that you can minimize the noise because your chip is
working only at one volt right you need to make sure that you don’t have much noise margin
here you need to make sure that you minimize the noise right and
that how that is why you need to be able to distribute your so called power network very
efficiently ok and of course because we do that we get very high performance we talk
of frequencies upwards of ten gigahertz and of course ah we also consume very large power
today you know state of the art microprocessor is consuming more than hundred watt this is
also becoming a serious issue its as if you have a hundred watt light bulb sitting in
your ah c p u box right you know that is the kind of ah power dissipation that we have
in this chips another important point i wanted to bring about is that just because ah we
have a newer and newer technology introduced almost every two years it doesn’t mean that
all older generation technologies are obsolete ok each technology has its relevance size
the fit all all it means is that the highest performance the so called leading and bleeding
edge chip such as microprocessor very high performance flash dram and what have you that
will immediately go on to the new technology node ok what
this is a very interesting plot again x axis is year time line ok but y axis is a very
interesting plot out here you know this essentially is a worldwide wafer production across technologies
if you look at nineteen ninety seven along this nineteen ninety seven you have actually
created another x y graph ok on out here w p c refers to wafer production ok and here
this refers to the feature size way back in nineteen ninety seven we had only technologies
going up to point five micron that is why along this horizontal line you don’t have
any data points coming down here ok the technology the best technology that you had stopped there
ok you had lot of products out here but at that time one micron technology which was
state of the art maybe in nineteen ninety continued to stay there for niche application
ok so there was wafer production to make one
micron chips as well and as you go along the years you see along this line there are new
rectangles that are being added you see right all that means is that today you know if i
look at two thousand seven it’s not today of course five years ago you see in addition
to this there are new rectangles added which corresponds to different generations of the
technology right going all the way to ninety nanometer technology back then if i construct
this plot today it will go all the way down to forty five nanometer but the point to note
here is that of course the highest performance chip will immediately go to the latest current
technology but even then you still have worldwide wafer production happening on point five micron
technology point seven one micron and so on and so forth
just to give an example if you want to make a very high power chip ok you want to drive
several watts ah you know into you know your audio amplifier or what have you right then
i cant do that with one volt right i need to maybe use five volt or ten volt then the
only way i can construct is using one micron bigger transistors because if i go to you
know smaller transistors the transistors will just die as i said they will break down ok
so each technology has its relevance but the point is that the leading edge technology
will occupy most of the business because most of the business really comes from the very
high end chips like microprocessors and ah you know a very high capacity memories and
so on and so forth in addition to this there is something interesting happening for last
few years ah and that is ah something called ah building sensors on silicon ok so called
cmos sensors ah presumably these will become next technology drivers ok in other words
if you see last few decades in every decade there was a technology driver right way back
ah you know in nineteen seventies until microprocessor happened we had only memory chips ok well
of course you had microprocessor chips that microprocessor was invented and nineteen nineties
was really internet era right internet was invented and
you had lot of new chips the so called network chips chips which enable networking came about
it doesn’t mean that we didn’t fabricate processor we didn’t fabricate memory we continued to
do that but there is a vector that has emerged ok in the last decade you know its really
an explosion of hand held devices right and accordingly we had large number of chips which
enabled ah you know your cellphone p d as digital cameras and so on and so forth i think
next couple of decades really is going to be you in terms of systems we will really
be building so called wireless sensor network based system right or
so called ambient intelligence right we need to be able to have very low cost sensors spread
all around which will ah collect lot of information in real time and ah you know makes sense out
of it right in order to do that we should really enable a large number of sensor chips
right so this is also that is happening although in this course we will not really talk about
sensor right but when we talk of nanoelectronics this is also a very very important ah aspects
in other words we talk of what is called more moore and more than moore what we mean by
more moore is essentially this is conventional electronics we are building computational
units such as c p u logic and storage unit such as memories right all that we are doing
is miniaturization ok and this is what is happening as per the moores law we are miniaturizing
scaling and so on and so forth more than moore is essentially along with electronics trying
to embed something different right to begin with along with digital you may say i want
to embed analog and r f function then i want to embed passives such as inductors capacitors
on the same chip then also i want embed high voltage high power transistors on the same
chip which has low voltage low power transistors then taking it further sensors and actuators
and biochips and so on and so forth right this is the that we will not really talk about
ah ok so we will restrict ourselves only to what
we call traditional scaling right what i call traditional scaling is essentially shrinking
feature size we take the building block which is transistor we are only looking at electrical
domain make the transistor smaller and smaller the technology drivers have been memory and
logic the technology drivers will continue to be memory and logic along that traditional
scaling part the material and devices are silicon and its derivative classical cmos
is what we had going forward silicon will coexist with a variety of materials that will
be the focus of this course we will talk quite a bit not only on silicon
on germanium on compound semi conductors and so on and so forth right in addition to that
device architecture not only cmos we will also talk about non classical cmos ok that
will also form a significant ah chunk of ah the course ah that we have ok whereas this
equivalent scaling path is really staying at the same dimension not necessarily miniaturizing
but doing functional integration add more functions to the chip not just compute and
storage but add you know sensing and actuation this is what we call hybrid systems on chip
or s o cs silicon will coexist with variety of functional materials polymer plezo ferroelectric
magnetic so on and so forth in terms of devices not only cmos but mems chemical sensor biological
sensors everything will be sort of integrated on a single chip ok
so we would need ah new materials also to be able to do that ok this as i said is going
to be the ah emphasis of this course new materials and new device structure right so you know
what i will do at this time ah you know is just flash this periodic table for you to
go back and think about ah you know of course you know silicon is out here in column four
which a semi conductor and our entire electronics is based on silicon right in addition to silicon
there are lot of other materials ah we will see in the next lecture that ah what it used
to be until ten years ago we will recognize that we had only handful of material if you
were to ask what is the ingredient of the chip you will see only about five or six material
that you could list but today its quite different we will see why that is the case and we will
also talk about the device structure right so we will save it for the next lecture and
ah you know with that ah we will conclude the lecture for the day

One comment on “Electronics Nanoelectronics Devices and Materials mod01lec01”

  1. samia abdelhaleem says:

    where can i find the slides?

Leave a Reply

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