[MUSIC]
Hi.
Welcome to lecture four of week two of
nanotechnology the basics.
In this lecture we're going to talk about
some nano materials
that may be useful eventually to replace
wires in, integrated circuits.
The standard architecture in integrated
circuit today
uses voltages to control currents in
transistors.
So those voltages and currents need to be
delivered to the transistors by wires, a
standard interconnect.
The basic interconnect these days is
simple copper wire.
It's an evaporated film of copper
patterned onto a semi-conductor substrate.
There are two problems that arise when you
try to shrink this down to nano
dimensions.
The first problem has to do not merely
with how small the wire
is, but how fast you're trying to switch
the voltages on that wire.
We all know that our computers have gotten
faster as time goes on, so the cutting
edge laptop computers these days have a
clock speed of
several gigahertz, which means that there
is a wire in
there where the voltage is being switched
from a high
value, maybe 5 volts, to a low value, say
zero.
Several billion times a second.
So the faster you try to switch, a voltage
on a
wire, the more energy loss there is as the
wire gets smaller.
So smaller wires, it's harder to switch
the
voltage then larger wires, and one, one
reason for
that is, that the wires tend to radiate
more efficiently.
And so the energy is radiated rather than
transported along the wire.
A second issue has to do, again, not so
much with
the size of the wire, but the proximity to
it's neighbor.
So when we're shrinking the whole circuit
down, we're moving wires closer to each
other.
And if there's a voltage switching on one
wire,
that can lead to switching on the wire
next door.
So that parasitic coupling
between wires also becomes a big challenge
in
designing circuits to operate on the nano
scale.
There are a couple of potential solutions
that
nano materials can offer, and that is to
through the route of different kinds of
materials,
other than just conventional copper
evaporated on silicon.
One candidate, is a carbon nanotube.
So if you take a single sheet of graphite,
which makes this hexagonal array of carbon
atoms, and you roll it
up, and make a tube, then what you got is
a carbon nanotube.
It's got nothing but carbon, every atom is
a carbon atom, it has a diameter
Of perhaps a nano meter, and it has a
length which can be well really long.
It could be only ten nano meters or it
could
be ten microns long or it could be
millimeters long.
These things could be made with
ridiculously large aspect ratios,
so they look like wires, and in fact
certain types of
carbon nano tubes can conduct electricity
extremely well, better than copper even.
So they could, in principal, be
a replacement for lithographically
deposited metal wires.
one of the challenges, well there's
several challenges, associated with these
things.
One of them is how do you, how do you grow
them?
How do you make them?
So if you want a characterize a, an
electronic circuit
that contains the carbon nanotube.
You need to be able to put down a nanotube
between two electrodes.
So they are really three ways to do that,
one is
simply put down the tube and then put
electrodes on top.
The second way is to put electrodes down
on the surface and then put a tube on top.
Andf the third is to put down electrodes
on the surface and that actually grow the
tube.
In situ between the electrodes.
All three of those procedures have been
done and these pictures show you a couple
of examples.
This is really a single carbon nanotube
bridging a gap between two metal
electrodes.
This is a great way to measure the
properties
of a single tube and to confirm that they
really do have these remarkable properties
that are associated
with the nano scale size and the molecular
perfection.
The challenge here is, how do you scale
this up?
What if I want to grow not just one tube,
but a billion of them.
and is it really reproducable?
And those are really big challenges,
because those are necessary steps
in order to be able to use these things in
real circuits.
So, there's another challenge, which is
the wiring problem.
Right?
We want to take advantage of the fact that
these structures are nanoscale in size.
So if we build a single this stands for
single-walled nanotube.
If we build a single wall nanotube
transistor.
Which is a nanotube bridging the gap
between two metal contacts.
Which we can call the source and the drain
of a transistor.
And then the nanotube acts as the channel
between the source and the drain.
That's great, right?
We have a a transistor with an active
region that's only one nanometer thick.
The thickness of the tube, right.
The diameter of the tube.
The problem is that it's connected to
these metal contacts, which are giant, by
comparison.
And which extend out to electrodes which
are even more giant.
Which are really enormous.
So are we really building a nanoscale
circuit if we have all of this other stuff
surrounding our nanoscale element?
Right, this is known as the wiring
problem, and it's really
an issue not just for carbon nanotubes,
but for any nanocircuit element.
If we have a single nano object.
Connected to an extended large structure,
we're not doing nano science any more.
Or to put it another way, we're doing nano
science, but we're
not taking advantage of the fact that the
object is nano sized.
So there are of course reproduce-ability
issues here, there are
three of them that I'll mention associated
with nano tubes.
Number one is the tube type, if you take
a, a sheet of graphite and
roll it up there are many, many ways to do
that depending on the roll angle.
Right?
The chirality of the, of the tube.
And some of those are metals and some of
those are not.
There are some of them are semiconducting.
So if you want to make a wire between two
points using a semiconducting tube
wouldn't be so useful.
so, you need to be able to know what kind
of tube you have.
And right now sorting tubes is difficult.
It's a hard thing to do.
Number two is how do you get the tube to
sit in the right place?
Right?
It's very difficult to place a one
nanometer object exactly
where you want it, repeatedly over and
over and over.
so that's a real challenge.
And number three is the issue of contacts.
Ultimately these tubes have to be
connected to some metal object.
And so that means you're going to have a
connection between a carbon and
a metal.
And those connections are, are at this
point, not well understood.
Theoretically not well characterized
experimentally and have large variability.
Meaning if you do it a hundred times you
don't always get the same answer, right?
And so that turns out, those turn out to
be
real big challenges in the fabrication of
nano scale circuitry.
There are also some surprising or
unintended effects in carbon nanotubes as
wires.
These are really one dimensional
conductors.
The electrons flow in just one direction
along these things.
There's no way for them to flow
in any other direction except the
longitude, right?
And so if there's a defect in your tube,
if one atom is missing from your, from
your perfect molecular lattice there, then
that's a
defect and every electron is going to see
that defect.
Now compare that to the case of a bulk
piece of silicon which might
be a 100 nanometer cube, right, if I
remove one atom from the middle
and, and leave a, a vacancy, a defect,
then
most of the electrons flowing through this
cube won't
see that defect at all, and so that defect
will have a relatively small impact on the
conduction.
But in a nanotube where the conduction is
one dimensional every electron
sees every defect and so this really has
a, a, a big challenge.
That means we have to be able to control
or eliminate defects and eliminating
them is really hard so we have to at least
be able to control them.
And the last thing I'll
mention is this issue of adsorbates.
What I'm talking about here is if you have
a tube lying
in air then any molecule that falls on it
can stick to it.
Water molecule floating around in the air,
or
some organic, and when something sticks to
the outside.
Then, it can impact the conduction.
It can influence the electrons that are
flowing through
the tube, and change the way that they
flow.
So, these absorbates can strongly affect
the electronic properties, and that means
either
you have to package this thing really
well, to keep
the absorbates off, to keep off all
molecules, basically have to
put it in vacuum or you have to regard
this thing,
not as a conductor of electricity, but as
a sensor, right?
It can sense when molecules fall on it.
So one person's packaging nightmare is
another person's sensor.
That's the joke that people like to use.
Okay.
So, if you don't like carbon, then there
are other alternatives.
We can also make nanowires out of
semi-conductors.
So let's talk of it, a bit about
semi-conducting nanowires.
There are methods for growing single
crystal semi-conductor wires, that are
maybe.
A few nanometers or 10 nanometers in
diameter, and again, many microns long.
So they have very very large aspect ratio,
they look like wires, okay.
And these structures have been
demonstrated
in many different kinds of semi-conductor,
including Silicon and lots of other things
too, okay.
And i-, in addition to being able to grow
these things in Silicon or Gallium
Arsenide or other materials.
You can also dope them, so you can do
controlled doping distribution,
either along the length of the wire, so
you could have a p-type
length, and an n-type length, a p-type
length, and intrinsic length and so
on, or from the inside out, so you can
make core shell structures,
where the interior is, is, conducts
electrons and the exterior
conducts holes, or something like that,
and so these complicated structures.
which can be grown in controlled ways
offer
a lot of very exciting possibilities for
example you
might imagine building an, a pn junction
into the
wire so your wire then becomes a
transistor itself.
So there are, these are exciting
possibilities.
there are a number of advantages, number
one, because these things are grown in
a more controlled environment, the
reproducability is
much improved relative to carbon nano
tubes.
the they're grown as crystalline,
crystalline materials with facets, and
so you can really know which facet you're
growing on.
And so on.
They're similar in their benefits, as far
as electron conduction to carbon
nano tubes, it's still a 1 d conductor,
and so you still
have a lot of the same benefits.
there's a known chemistry for the contact,
so when these nano wires are contacting
the metal, that junction is much more
well characterized than a carbon metal
junction.
this issue of surface states, is a little
bit more well controlled, because you can
passivate
the surface, and in principle, it should
be
easier to interface these materials with
other materials.
So a silicon nano wire contacting a
germanium nano wire is going to be
a very well characterized contact.
But of course there are also challenges,
right?
You have the same wiring problem as you do
with nanotubes, if
you have a single nanowire connected to a
macroscopic metal set of
leads, then you really, are you really
doing nano science, are you
really taking advantage of the nano scale
nature of your active area?
many of these materials will oxidize in
air, they will undergo oxidation chemistry
and
that will change their properties, so one
has to be careful about that.
You have the same susceptibility to single
defects, as you
do in nano tubes, because these are still
1 d conductors.
So, you'll have that same issues, and
you'll have a similar vulnerability to
surface modifications.
So, once again is it a sensor, or is it a
packaging nightmare?
and if you are trying to build electronic
circuits,
then packaging is going to be the answer
to your
question, and nightmare is a word you'd
rather not have have to use.
So, let's talk about what we've discussed
in this lecture.
First of all, really, there's a key
challenge when shrinking down circuits.
It's not simply the active components,
it's also the passive components.
That is to say, just the wires that
connect things up.
These can really be a big challenge.
Number two, there are other materials
other than the conventional materials used
by the semiconductor industry that offer
possible nano-solutions.
That are very small.
And can be really, really good conductors.
Even better than copper.
Small does mean highly sensitive to the
environment.
And so packaging becomes a real challenge
on the nanoscale.
Probably a bigger challenge the smaller
you go.
And lastly we still have not yet figured
out how to wire up a billion nano-devices.
We can wire up one nano-device, but then
again, are you really doing nanoscience
if you're not taking advantage of the fact
that your active area is small?
Because you're sort of killing the
advantage by connecting it
to a big thing, so we've got to figure out
how to place a million or a billion nano
things
and wire those up, and that's still a big
challenge.
So, we'll talk more about nano
materials as potential solutions for the
semi-conductor
industry in the next lecture.
Thanks for listening.