So a speaker is wholly
University of South Carolina.

Is a bachelor's degree in technology and

asters from Cornell University.

Joined the Department of Law
versus Carolina where he is

currently associate professor publish
the number of papers and a couple of.

One career warrior and
his research spanned the areas.

Some of the sensors and

that kind of thing of interest to I
think a lot of people here today.

So thank you.

Thanks David.

For the nice introduction and
thanks very much for inviting me here and.

Well like you mentioned about
I'm calling from the Department

of Electrical Engineering and
the University of South Carolina.

So today I was talking a little bit on

one of four major research
thrust areas which is.

Three five and I tried best micro And
then of course the talk

basically organized into
to kind of if you would.

In the first half of the module what
I'll do I'll go over some of the.

Al again Gann aluminum gallium nitride
telling one tried to best Pinfield

cantilevers and age for
sensing chemicals and

all that in the second phase of the taco
I'll talk about other aspects of

our research is more in the Nano domain

growing nano scale three five nine tried
materials and using them for anything.

So before I start Of course technology.

Some of the other group members might
value students because he could do it and.

Already graduated.

So part of their work has
been incorporated here.

So all of the sheet as if she
is one hundred registered

in Nicholas two year old are we also
collaborative very closely with Thomas and

that of Oak Ridge National Laboratory is
currently at the University of Alberta.

And he moved there last year sometime.

And of course all of our major fabrication
efforts are being carried out at

Georgia Tech in our city center right
here and like to thank David for

all the initiative and help that we
have received from Georgia Tech side.

Obviously we need funds
to carry out research.

And this research were supported by two
N.S.F. I was like technology that as well.

So here is a kind of formal
outline of my talk so

I'll give you a brief introduction
on different at the same story.

Or what interests are used for

that you guys don't know
which you probably already.

Which would be probably you know.

It's overkill.

Because I think most of you or

many of you would be working on
sensors to be quick to that and

then I would be introduced to a little bit
more specific sensing mechanisms using

micro cantilever So that part I don't
know how much they're familiar with and

we will specify our approach using micro
can't leave us the same thing which is

somewhat different from what is
normally pursued by researchers and

in that respect the micro cantilevered
themselves are also not made

of silicon in made up aluminum trade and
we talked a little bit about that.

And some of the results that we have
gotten on the micro can't leave Earth and.

Finally the right man.

And while growth and
their characterization and properties.

So obviously most of you
are aware of what the applications

of sensors might be they are.

Well varied all over.

From moment to call it turns to
Industrial Process control chemical war.

Medical applications and
so on and so forth.

So in my view without sensors
today's world will be.

Really pretty unpredictable
are pretty uncontrolled So

they're very important to do research on.

And our approach to would
be to quote use micro and

a scale devices to perform
the same things so.

So like I said before we are easing micro
cantilevers as our sensing elements so

what are micro cantilevers for
those of you who are not familiar.

They look something like a tiny diving
board and you can see one side of this is

attached to the main body the other
side is kind of flexible so

you can actually resonant them.

You can move them and
the movement motion can be

detected by using by shining
a throwing a laser and.

Detecting the laser reflected light.

So when the amplitude of the oscillation
of the cantilever changes

the detector actually produces
a signal of varying amplitude and

that can be detected to find out how them
how the microcontroller is behaving.

So this has been done for a while.

And have very good advantages
like High Fidelity they're

extremely sensitive to physical parameters
like temperature surface to air stress.

If you hold mass changes then what
we'll be focusing on is electrical.

So for surface work function
which is also a chemical concept.

Anyway.

So they're very sensitive to dolls and
can be used as a very very

sensitive transducer to perform
this kind of protection and

of course they can be
all microfiber get it.

That's another good thing there is
no guesswork here we can do exactly

as they can be done.

And so they're good for
me to radiation and integration and

if you distill it can best they can be
integrated with mainline silicon process

can be formed in from chips chips.

So.

Also micro cantilevered by themselves
consume very low power because it's more

of the oscillations that have they
tend to have high quality factor.

Very high quality factor much more
than electrical circuits anyway and

if you perform these are solutions
in vacuum the quality factors can

be as a few thousand maybe
six thousand dollars so

by themselves very low power and that
that's good and of the advantage as well.

The initial work on micro cantilever was
started the Oakridge National Laboratory

and that I had.

Had the pleasure to work with him for

a few years before he moved
to University of Alberta.

By the way.
Feel free to stop me any time that have

a question or if you want to discuss
something more be happy to go over

the point and discuss a bit more.

All right.

So in my country words are traditionally

have been used for detection
through two different principles.

One is the frequency change.

So I can see the frequencies can
be detected from looking at their

frequency vs amplitude characteristics and
when someone kills or

something at the micro cantilever
then the frequency shapes and

from that the attached mass can be
detected and you can find out what.

The point where the target and

light mass is there of course this kind
of cantilevers have to functionalized for

the mass to be attached and people have
done that and that's very nice and good.

You might notice though that we
need a very narrow pick here.

And a far better detection
better math resolution and

that is possible only when the quality
factor is very high which often means

that the micro can't leave or
has to be enclosed in vacuum.

And that is not very
practical in many cases.

OK in our approach we have found
a way to go around that problem.

We can still enclose the micro cantilever
but perform detection explained how.

The other one way to do it is to stress
change this works in the same way

the microcontroller is functionalized
when molecules come in and

estimate the surface work function
as surface free energy changes and

the surface stress changes and
the microcontroller bends up or down and

from that you can actually find out that
some and a light molecules are present.

OK So traditionally these are the two
ways of doing things one is static or

other is dynamic.

But both in both cases are to
functionalize the micro cantilever itself.

And when you functionalize micro
cantilevered they tend to go bad.

Not just that uniformity would be
an issue as well from batch to batch so.

And once the micro can't stop
you have to throw it away and

that's kind of expensive process.

So what.

So what we started to look at is

to avoid from these challenges and
use potential metric detection.

So what we do we do not functional is a
cantilevered we functionalize a substrate

which is kind of a counter electrode and
based on the Kelvin probe technique which.

Works on the basis of what function
difference between the microscope

cantilever and the counter electrode.

Michael country was going
to make to also lead and.

Did that change in work function.

There are some mathematical formulas which
I just go for pretty quickly as to give

you some kind of flavor of how the math
looks like and then you'll understand.

Hopefully it'll be better how
the difference in what function within

the counter electrode and
the cantilever can be measured and

can be used as a parameter to detect
presence of I like molecules or

attachment fairly certain but one of
the other advantages one advantage is

obviously can't we were does not mean to
be coated Gentiva can human pristine your

transducer elements pristine and it is
possible to enclose that in vacuum and

preserve its high quality Factor U.K.
and I'll explain how.

The other good thing is obviously if you
don't have a country called the cantilever

then it does not have
a chance to get buried or

degraded over time so that you don't
have it throughout the country which is

which can be expensive and
the third thing is by changing and

you can see how I can't live with
my as we put ideas important here.

And if you voltage together and
this one substrate is grounded.

So basically capacitor is forming
in the cantilever and the function.

Hopefully that would be connecting.

And then we basically apply with an AC and
make it awful it.

So any change in what function will
change the amplitude of oscillation for

the cantilever.

And this is how the math worked out.

Not sure if you would enjoy
the big formulas or not but

at least you can see from here that.

Total capacity forth is made
of one constant term one

hundred thirty four in term and others.

Certainly varying term
with to ice the frequency.

So this is basically dependent on it.

B.C. bias and Delta file Delta flight
presents the difference in walk function

between the cantilever surface and
the counter electrode surface and

this is our focus basically So
if you lock in to this magnitude.

It's easy to lock in this one because it's
sinusoid So a lock in world basically

looking at prefer would kind of reject all
of the frequencies and just to zero in

on that particular frequency and so if you
look at the manga of that then there is

proportional to this
change in what function or

difference in walk function we can
the cantilever on the surface so

effectively we are getting out of
the idea of functionalism the cantilever.

But we are putting our functional
usually are on the counter electrode and

looking at the change in what
function of the counter electrode

when molecules adsorbed on it and this
is how the work function because of this

change in work function the amplitude of
oscillation of the cantilever changes and

we can get our detection done.

So this kind of going more
detail into the band diagram.

I would avoid that.

But once again.

This is the displacement amplitude and
it amplified by the quality factor.

So if you put it in a vacuum
where they can't even vacuum.

It will become more sensitive.

And this is what function
change that appears here.

OK So
just to summarize the entire process.

We have molecule or at the top sion.

Stuff is what function of the counter
electrode that causes a capacitor force

change that causes the solution
a picture change which can be transfused

into some kind of readout so ultimately.

If there are large concentration things
will start to change much faster and

so on and hopefully I can give you
a couple of examples of how we detect and

you will understand more better.

So what do you.

In the initial stages now we're moving
away from it but just to give you an idea.

We basically modified
an airfare into action system

we have made some kind of fixtures for
throwing gases and helping the fam pull.

People to be here and
we are basically looking at

using this can't leave our best detection
and we can't leave it still at us.

Today.
If we have what they're doing here in

Georgia Tech is improved the process
take it a step farther and make harsh

environment sensors based on elegant again
cantilever the initial process of course

using fill you can with micro cantilevers
which are commercial cantilevers and

putting them in the air fan in a modified
air from the power from the detection.

So.

Some of our sensing results.

What we did is using this technique.

We took a function like this
one layer of platinum platinum

coated glass light to be more specific and
then that was a counter electrode and

we passed gas on top of it while
measuring the surface work function.

And suddenly the work function changes.

And what he thinks is as soon
as they start to get one hundred

percent pure hydrogen it of
course goes very fast very fast.

And be twenty thousand ppm hydrogen and

there you are basically feeding the P.O.
nitrogen hydrogen into

an IF YOU CAN hydrogen the highly unfair
producing record of our best nitrogen.

Now the question is why does
nitrogen produce response.

Well the answer here is what do you
use for the pure and dry nitrogen and

that when you flowing on the sample that
also show us the work function because

the sample is exposed then Hahnemann
it has all sorts of IT oxygen in it and

it has water vapor in it.

So when you floor nitrogen
that's kind of pure and

doesn't have any water vapor so
that for the system changes as well.

So we analyzed is there.

In detail and we found that it's
definitely possible to produce a P.P.M.

hydrogen in air using and
pretty selectively using platinum coating

and which is one of the best results so
far I think the lowest one was about four

pm but
it has about point five ppm hydrogen.

So that's that's pretty close to that.

Another example of the one that you
published in two thousand and seven

is using based on what we call is
nano structured graphite and this

is kind of a poor man's way of producing
graphite rather than using a fondness and

a silicon carbide substrate what he
did is just which people other people

also do like mechanical exploit
graphite or kind of rub it.

So we took up the substrate and it could
be any substrate it doesn't matter but

it's a metal insulator so I can record and
if you put a piece of graphite or

even your pencil on the substrate

it leaves enough amount to flex if it's
the pencil it'll be mixed with clay.

But if it is just pure graphite by itself
it will leave pretty nice flexed and

here is what that pyramid looks like for
this flex and this all indeed is just

rubbed this high purity graphite on the on
the substrate and you can see that.

Pretty nice at a big
layer that is already.

And even here in the lower the Lucian
more you can still see differently and

graphite layers and
because these layers that kind of

broken we are we're not shooting for
any electronic electronic device so

we didn't really care whether
our multi layered graphite or

graphene whatever you want to call was
largely or not it was just some flex

all over but that probably enhanced
the sensing a lot and what we could do

is we could go into don't appear to
be level and we did sixty people be.

Detection for
two is in this kind of flux so

it's probably well known by now that
graphene really gets affected by.

And that's in use the what
which would be detected.

Yes place.

We did all of these experiments in here
just to be consistent with what the normal

operation would be of this material.

So it was the ambient will probably
have about ten people B. and

it took so we passed like
sixty people be in zero two.

Mixed with nitrogen.

So that's what the responses.

That's right.

Yes.

Yes.

How does it reflect well we can
do two three things actually.

We can look at the rise of it and
from that I see that

you can actually get some conservation
idea if you look at here.

So at the concentration
of an open increases.

Basically the rise that will change.

So you can calibrate in principle you can
do that but until we question calibration

effort that's a very good question.

Yes and that's exactly the reason
why you moved from it.

But this was to prove the initial
concept that it works OK So what you're

saying is that OK you are using a very
cheap and dirty method of using this

graphene to our graphite to put it on
the substrate right every time you do it.

It could be different.

That's what your point is that this is.

Yeah yeah exactly.

So that's actually true.

And that's why we moved away from that and
we're using graphene

the largely Agrafena silicon carbide but
this is just to show the concept and

show how easily you can actually
detect to easing this technique and

those things like dirt cheap I mean
if you take any soft and even paper.

So what we tried is used paper and

put wrap pencil on it or
graphite it would take the same way.

So it has some advantages but

obviously the state standardized
be difficult unless you do.

We don't know we didn't really look at.

More into that.

OK So we the initial proof of the concept
for potential metrics tensing done.

What he focused is actually
everyone will talk about great so

you have a great sense there but
what about selectivity and.

So the way we've looked at it is again.

We tried to combine two techniques.

So we did the work function change
measurement and we also did connect and

change measurement and so on one side we
put the connectors change of the thin

film of structured graphite and
one thirty put what function changed.

And basically you can see the red one
is here is what function change and

the blue one is going to can change

what function saturates because
it's a surface best technique.

Connectors.

Moving up because molecules
go through the ports and

change the overall
connectivity of the film.

That fine.

Everyone expected it to be fine.

What is interesting is twenty
plot take these two things and

remove the time dependence and
plot it to owning a starter.

What do you see for that ice.

It follows a nice straight line and
for the fall part.

It also follows a nice straight line and
the slope of this rise and

fall are pretty much constant so.

This for the first time really
got interested he thought.

Well if we do this together the work
function and the connectivity measurement.

Maybe maybe we could get signatures for
each of these and I like gases so.

So that would be one more step
toward any identifying one and

a light of course is still run into
trouble if you have a mixture.

If you're mix or two or three mixtures I
think we would still be able to do it with

a binary mixture by looking at three
dimensional signatures capacitance

change what function change and
connectivity change it would be getting

messy from that point on one hundred
mixtures to be very difficult hours but

still it illustrates the concept that
by doing multimodal measurements.

You can actually identify molecules and
this one.

What we are showing is.

Different concentrations of
energy to produce a similar slope

in the first quadrant.

Whereas other things like acetone ammonia
methanol water vapor they are producing

signatures in different quadrants and so
this was the initial first proof that.

Well this concept might work
into five hundred volatiles.

Or X. So now what we did is so far is kind
of give you a flavor of our initial work I

guess I'll have to go a little bit faster
and have a lot of material to cover.

So we then started to design my Cantley
verse which do not need a laser

which basically has to have proper based
on the to get it be as a resistor and

we tried to make this
cantilevered of elegant again.

So here is our design.

We have a gate train and key bias which
basically will have its counter electrode

which will functionalized and this will
make the cantilever too awful it or

the folds and the drain will basically
pass the current and the gate.

We modulate it didn't create its
sensitivity or kill its range or.

If you would.

So why did you choose gallium nitride
because of several different advantages.

So one of the advantages of
gallium nitride is that it has

very strong correlation properties.

So if you have a multi layered structure.

So as to stay off aluminum gallium nitride
and gallium nitride because they are so

strongly polar and
the pollution is different.

You do not have to go up anything you
will basically don't have drop any

of this material but you'll get a two
dimensional electron gas or a fairly high

two dimensional density of electrons
that the interface of this material and

that that has been very well studied for
the last ten fifteen years forms the basis

for aluminum gallium nitride gallium
nitride based microwave transistors.

OK So we are just boring that concept from
that domain and trying to use the same

technology for our purposes and because
this material wideband up white man

get meaning the bank of gallium nitride
that were three point five electron volts

aluminum gallium nitride for about thirty
percent concentration would be about four

point one electron gold aluminum nitrate
has about six point one electron fold.

So these are whiter man get material.

Don't easily get damaged by radiation.

Don't get too much affected
by changing temperature

because their intrinsic carriers
don't change too much temperature.

And so on and
more importantly chemical you know art and

mechanical is very strong.

It's hard to break this kind of material.

So and because it's central symmetric
non-central symmetric crystals.

So that's why the two D. information is
very easy I'm not going to go into details

into non-central symmetry and
all that but because the polarization and

North-Central symmetric we're having these
electric gas forming at the interface.

And we can calculate all
of that using formulas.

We're probably not going to that.

Anyway.

So our approach to this.

Getting the micro cantilevered involved
starting with substrate that you bought

from Micro next and going through a state
of the fun process starting from message

outline and so on and so forth and
finally the back silicon pocket so

they can deliver a structure is
all grown on a silicon substrate.

And this is way out of jail.

To show how the proportional
basically Allegan is very thin

about twenty five nanometer gallium
nitride is about two micro meter and

fill again substrate is a regular silicon
substrate on which we are growing these

aluminum gallium nitride gallium nitride.

So it's way out of proportion.

Some of these you can see from
here actually much better.

So and so you follow all the steps and
let me go over this step by step.

So we first start with the gallium
nitride at and defining this al again.

And then after that we basically.

Put our cantilever outline on
the top with as you can see and

this is how it looks like the picture and.

Following that we do the only contact
people vision for source and.

Drain right here.

Regular process flow for

gallium nitride which my students
developed right here in Georgia Tech.

And after that we do that get
made a live edition which is

to be Shakti the other would be omic to
allow easy flow of corn to be shocking

to control it and
then get the polish on and

then obviously you have to put the borning
parents to contact the outside world and

this is how it looks like on
after all this process is done.

And finally this is a massive
part of basically have to

release the cantilever by.

Going through this pocket back pocket
as you can see the kind of pocket

things where it's kind of.

Carved a region that's where
the cantilevered start and

we actually developed a process where
you can do it in form of chip as well.

So this is makes life much more easier
when you want to take a chip out and

and wear mourning.

So these are the cantilevers
that it looks like but

you can see the country where it's not
Trek like silicon cantilever the reason is

because there is some stress in built so
when the silicon is removed

the stress takes effect and
they can't lever bends down ward

in these cases there are some advantages
to that in terms of switches and

we are investigating those aspects
of bending and this is the much

more magnified image distorted and
in that context for the cantilever.

So that's fine.

So we went through something basic
electrical measurements and the process is

working well for the last one and a half
years we developed on these processes and

the process is working all
the parameters that we get for

the contact resistance between short and
thin and all of the parameters are have

been finalized now so so
that that's great and

we did the mission of gas density
that comes out to be about three

point five times in the twelve which is
not as high as either expected it should

be somewhere close to ten to thirteen
eight times in the twelve or something.

So there is still some process
some trapping effect going on

the process has not been optimized yet so
we're losing some of the charge channel

charge interest charge
which is very important but

it still can function of
the sensor as you will see.

So these are the transistor
characteristics for the resistor by

the way it appears as this very thing to
get at the other peers so there's this.

Each Fed is integrated at
the base of the cantilever

just like silicon cantilever has
a silicon was fed embedded or

appears a resistor network embedded
we are using these elegant Gand H.

fed at the best to stick
to detect the strange

change that occurs when the cantilever
bands and that's how so.

So we are right now focusing on
the transistor aspect of things and

this is the I think a Christic
pretty decent I'd be characteristic.

And we're looking at the ideal B.G.
characteristic as well.

So all this seems to work and.

It obviously needs to be changed to Bart.

And modified and
we find it seems to work at a basic level.

We also measure the quality factor and

quite a factor is pretty decent actually
having a quality factor in here is

not too bad matches decent
way with stimulation as well.

These are all done with
console simulation.

Predicts that what thirty to
thirty one hundred frequencies

from the console we get about twenty seven
or so there's still some mismatch and

probably has to do with that I mentioned
the can't lever not being completely

accurate and this is the same
station that you have developed

to perform the sensitivity
analysis of the cantilever.

And I'll go over that in a moment.

So here is how the cantilever
works based on.

As you bend the cantilever as you can
see the characteristics change due to

different types of bending and
we get some change in train current horses

not bending and for
no good bias we find a gauge

factor of about forty and we've gained
by probably eighty there's a typo here.

And so they should be about
eighty which is pretty decent

gauge factors compared to
be as it is to flick and.

Can't you it.

Here is like you want to change the
dynamic the sensitivity the now instead

of putting it to measure the changes
we are putting in is the bias and

so we basically can monitor the AC changes

if the resistance changes and that gives
a gauge factor of about twenty as well.

So finally we did some step changes and
we calculated the gauge factor once again.

Basically all the gauge factory
coming everybody in twenty and

eighty in these cases and
that that's pretty

pretty reasonable has room to improve but
we're very happy about it.

And the First Level One of the interesting
aspects that we saw fairly recently and

you may not have enough time to discuss
that is if we look at trends and changes

for example instead of slowly bending and
holding and bending again and holding

if we do bend and don't do anything about
it don't make any step or anything.

What happens is that as
soon as the banded town.

There is a small transients I should
say small it's pretty large actually

much larger than here and that can give us
a gauge factor of more than thousand and

this is something that we
are still investigating.

So because this country was to
be used as off the letters.

We don't really have to measure
the steady state and gauge factor.

We can take advantage of the trends and
gauge factor and

that would be really nice to
have such a gauge factor.

So this kind of the summary of our work so
far and

we are basically demonstrably
militiaman technique for

non-contact potential metric gas and
think platform and

demonstrated it different gases used for
money clarity creation and

also used this concept to apply

it in form of elegant and
best micro cantilevers which can be.

It can be extremely sensitive and like
you mentioned the gauge factor operating

hundred has been observed intransient
cases and this is really very good and

at least an order of magnitude higher than
the silicon ph there is the can't reverse

and maybe even carbon energy
based ones that are you know.

So we're still investigating
this though this is kind of for

one only one data point.

So I'm not going to spend too much time on
that it feel encouraging where it can go.

So so far so good.

One more point before I go to
the nano scale aspect of things

I want to make this point clear because if
non-contact electrostatic best interaction

which is used for
detection we can in principle enclose

the cantilever in a chamber or
if in a small pocket

have its sapphire we know
being high dielectric

constant material will effectively have if
you have twenty five micron sapphire have

a thickness of you can read an article
as of two micron So if you have a twenty

five micron sapphire window it can't
leave it on one side enclosed in vacuum.

You can still perform detection by using a
counter electrode which is functionalized

and that concept is extremely important
has not been demonstrated ever before

that the transducer is in enclosed and
pristine whereas the outside

can be probed from that using on
contactless direct measurement.

Or so so far so good.

We have understood some part of these
scans liver best sensing but can it be

done in a scale that would be really
interesting because not only will it.

If we can control and make this.

Micro cantilever in form
of nano cantilever.

Maybe someday we don't even have
to go through some big process.

Maybe we can be stealth organization.

And get these nano cantilever and
get them memes.

So this was another effort of our
research where we started with

synthesizing in one thousand and
I am come to

explain why you want your one thousand and
works why we wanted to study these things.

In a few slides later.

But anyway so
we started in your mind to it.

Man and wife entity starting from
a catalyst particle like everyone else

is doing synthesis of nano words we used
gold particle which seem to work best and

using the growth mechanism
which is very famous by now for

liquid solid So we have the Nano why a
growing with a catalyst particle and top.

And what you observed from
pretty Are you on and

this is again a catalyst covered
particle nano welcoming outside.

What he observed very early on is that

the Nano where growth can be controlled
by putting catalyst particular

only selected areas I'll
come to that a little later.

But let me just explain for

a picture of our basic rudimentary
Stevie fornace So here is this before.

Our controllers and so
on gases are coming in reacting

on the silicon dioxide coated feeling and
stuff that and

boron nitride hold up and you have
the time a couple in the palm and so on.

So generally very very similar
to what people normally use and

this is a little bit primitive photograph
of our system it has changed much more

since the early days apologize for
not having a more picture here.

Anyway.
So

the growth proceeds at about
seven hundred fifty seven or

degree centigrade that is a fraction
of an atmospheric pressure and we use.

Two nanometer called Catalyst pads.

Which seems to work best any thicker you
get a jumble of matter where anything

are you basically have much
narrower certainty and

the growth rate is pretty decent
about thirty Micron per hour.

So.

If so what happened is that when
you started growing this man

where we basically had the Ollie
stead started to put gold everywhere.

And that resulted in this
jungle of man hours and

one day my student actually accidentally
scratched the surface of gold and

he found that nano welds and the surfaces
are actually not becoming entangled so

what they are doing is they're growing
alum the surface and on the surface and

showing some strange behavior
they're reflecting from one another.

And basically sometimes going along and
worse but always growing

on the surface and this was something
of interest to us because this actually

gives to a lot of different possibilities
here is an example of where from growing

from the age of at least BACK OUR know it
by now we realized that OK we cannot cover

the entire surface that catalyst and that
only gives rise to a chunk of land away.

So we covered only part of the areas and

then let the animals go out
from those areas into the open.

And this is what resulted.

So this in and worth as you can see is
going back and nicely deflecting back.

So several properties to no here are at
least absolve here one is that the animals

grow outward from the catalyst
particles that all his people.

Then and worth are actually making
a reflection from other networks and

that is very interesting because that
gives rise to lots of many places and

techniques for dismantling and
this must be core planners right.

If this is not on the plan then this could
easily mislead the thickness of the Nano

where would be about twenty
nanometer forty nanometer So

unless that core plan or I'm going only.

The same plan they wouldn't reflect
from one another they would just go

on different directions so always caught
our attention and this was absurd for

anyone tried.

Unfortunately we didn't see the full
gallop nitride Holly started it.

So at the initial stages we just about
at that plan in any way gallium nitride

very well studied and so at that time we
thought OK we're not studying at night X.

United also has other good properties
like such as a completion of electrons

that we'll talk about later as well
which also makes it very attractive for

sensing purposes and

because surface accumulation of electron
means they do for extremely sensitive to

impoundment because the electron on
the carrier charges right at the surface.

So that's very important as well.

Anyway so this was a structural
property that really made us

investigate things in more detail.

And so we started playing with
Catalyst part design we never

went back to covering entire
surface with Catalyst anymore.

We just only patterned things and
we looked at the different effects of

different sized catalysts image of
different catalysts I support and.

What we found is that if we do.

I think one my coming to earth part we get
nano a growth from that but that's kind

of half likely to go to call this three D.
growth because it's uncontrolled and

these would be two dimensional growth
which is on the plane of the substrate.

And here is an example
of how that changes so

five micrometer gives a jungle
of nano where fit and

control treaty as well as some two D. and
then two hundred millimeter gives you only

part of it and one hundred nanometers
gives you can't give a single anywhere.

So by controlling the spot size we can
actually control the number of man

a website coming out of that so
that if that's very good and

we have come to you with regularly

in terms of the diameters of
the Nano WEST Well they're right and Spot.

The diameter distribution and
one time I asked my students to count the.

Go through individual nanowires you had
to go through one thousand different and

what's obviously not happy about it.

But we did get some important
message out of it and

if you can see here the media and
meter and

the maximum frequency of occurrence
is around twenty nanometer and

that is very very very nice for
making names type devices.

So if we want to make transistors and
names twenty nanometer is pretty decent

size for semiconductor nano
wants to make connecting and

what we are even getting some in the range
but mostly in the twenty thirty.

That's what we're getting.

So here is that here we made showing
the catalyst spot at the age of

the aware and sometimes the gold
catalyst part get dislodged or

it's hard to say whether it's really
in the top of the surface or not.

And basically.

These are the effects spectra and

you can see a number one it's called
a lot of gold here but not that much.

Here we have three and
nitrogen and gold as expected.

The kind of confirms the composition.

Here is the height of
the Lucian which gives you

the keys to the graphic direction which
goes along the one one zero direction.

And here conference
deflection confronts the.

Structure of what it's like structure and
works as well.

OK So one other property that we got.

Into or noticing this property is that
the man was there bending that fine

bark is there a specific
pattern increased and

I know it's the bending Well if you
look carefully what we found is that.

Now it was indeed brain but
brain at multiples of thirty.

Sixty degrees.

So what happens is that if you look at
that image when they're going in one

one zero direction and they still
can maintain one one zero reaction

if their brain that
multiples of sixty degrees.

If their brain that if they want multiples
of thirty like thirty ninety one fifty

then they have to bend into one one zero
direction and we looked a lot and I'm not

very much into molecule or simulation so
I don't know exactly where the free energy

is would be the same in this particular
direction one one zero one one one zero.

We are still looking for people who could
look into periodical aspects of things and

predict that OK well this free energy or
formation of the same in this direction

and therefore they should bend
in those particular direction.

Nevertheless.

From an engineering standpoint it
does give us some of the leverage

in making a name space devices.

Imagine if we could control
the direction of the UN and

where it's the way it deflects back and
forth.

We could easily make a triangular
cantilever members can't leave it

out of that if we remove the oxide
material underneath by Actually

that was the whole motivation
into playing this game.

And they can bend at ninety degrees
as well like you mentioned before so

this is what also capture attention and
work and form different type of pattern as

they can form loops and so on and this is
very nice almost looks like a miniature

radio tower and
two which are acting as guides and

one animal going in between them and
it keeps going doing that maintain a given

mind it has to be coplanar otherwise
these will not deflect back and forth.

OK so it's highly precise mechanism and

it's doing that so that that tells
us that it is possible to control

the growth of these nano words in
a very highly precise manner so

we're looking at all the junctions
formation and so on.

Sometimes they deflect back and
they grow and.

West.

So there are many aspects of
that which we started in detail.

You probably don't have time to
going to too much into them.

If you look at the junctions though
they seem to be perfectly well and

they're very nicely stop and
you're using F.M. connected

to establish that these are indeed
electrically connecting these are not.

Detached are insulated from each other.

So.

Another important aspect.

I should say interesting aspect but
we did not really have a chance

to pursue this aspect for that but
he did feel look formation.

And to this day that we don't really know
what is causing the looks to form and

we suspect that there are some
electrostatic force which are responsible.

So whenever it comes to deflect back there
must be some charges forming because

it's a pull out material a lot of static
charges would be impossible to form and

when it starts to go deflect back maybe
the charges make it fall back and

again started deflected from almost
like a ball dropping on the floor only

it is getting translated
as well in the process.

So there can be many explanations the
charge explanation is one which you could

imagine about some aspect that we're still
doing it and it's nice though here and

it gets a nice picture we have multiple
occurrences of this looks formation and

double up.

Looks like Caterpillar.

So it's interesting if we could
manage all of these things

maybe who knows when we can form
a microfluidic channel through this or

something or trap a molecule or
some something inside.

Anyway so

we need some calculation just to make sure
that we're not really relying only on sci.

Which is kind of a two D. So

we did get the similar structures
in a similar deflections and

all that confirmed that also
similar thirty nine a meter.

Looked at from the other Kate.

Actions which are called A good because.

Reflect back and forth.

One thing and confirming that
these are very well very nicely

solid structures that
are being formed here.

And so.

All this is nice and good but how can
we use them for practical purposes so

that that's where we started looking
a little bit more closely into how

the Nano West Bank and every thought
before the Nano were being fun tenuously

basic angles or
when they meet obstacles like here.

They basically start to bend.

Interestingly they can
form multiple deflections

between two parallel manner works and
so on so.

So if in your mind and
words can themselves act as guide for

other in your mind tonight or
is it possible to also

to create little graphic patterns and then
cried then in the night and an emotion

that would be interesting because that's
what would be the first step toward

realizing some of these nanostructures
in a more controlled fashion.

So for example you think.

Changes or is able to barriers.

Can he do it.

He started getting some changes using
our I.E. and O.E. and you can see

that this is not as well developed this
is slightly better but still not so good.

These are our initial efforts and
some of the user and

some not very well trained at the time.

And nevertheless we did good on an aware
and he saw that when and where for

kind of deep as you can see here the Nano
word for deflect back from the tiger well.

Whereas the the naturally shallow.

It's kind of nano world can
go down go up along this.

So both of these aspects
can actually have effects.

If you can control that.

Then we can find different
types of structures just

by defining the trench in different ways.

So there are lot of course for
pursuing research in these aspects and

I still don't understand everything but

it is an interesting field then
what is a thought that well for

the train just might not be perfect
way to realize and aware of it might.

To use barriers because in
that case all the things

that plan are and we put the barriers and
then maybe can control them and

what direction that will put all the
barriers and this was your in Georgia Tech

to bury earth on a plane or structures and
then put catalyst particles in between.

And a story that they called Catalyst
particles which were deposited by electron

beam little graphic later after putting
optical optical deposited optical

it'll have the best depletion of barriers
and here is what the results are so this

is not a very clean result but you can see
the gold particle was somewhere here and

you can see the Nano work is going
back and forth between the various So

even using a little
graphically patterned barrier.

We were actually able to deflect the Nano
words and if you look at the junction.

It seems to be reflection from
the junction from the barrier here.

And the goal would be of course to realize
something like this in a much more

controlled fashion.

This was one of the random more random
effects but if we have a catalyst but

somewhere here we deflect it and
go back and

then contact the students electronically
and remove the bottom surface then you can

make a nano cantilever which can
also let which can measure stuff for

us and can use the appropriate for

scanning probe microscopy and
lots of different applications can.

Result So this is still the holy grail and

we are going after days
trying to you know.

Get it more repeatedly.

As much as we can and
also put them in an area.

So that's what the goal is.

And of course one of the things that
you have to control is why on earth.

Sometimes the wall or along the barrier
along the barrier and why sometimes

it gets deflected So those are the
questions that you have cancer as well.

So.

Now.

After fabricating this never works.

We of course started to look at
the electronic property because without

electronic property you wouldn't be
able to use these anonymous much so

fabricated devices and so

had a central Goldspot which gave rise to
an awareness at the source and try and

contact court after that in this
fashion and one would believe that.

Then a little characteristic of the Nano
were also came very nicely thought.

Train Carnforth to train voltage like
normal transistors for the first

time he showed a very good saturation
behavior drunken situation as well as very

good get control in the United and watch
the not receive difficult to actually give

rise to a device out of because they
have extremely high concentration.

So people always tend to get only
contact but not very good contact.

So this was one of the first results that
is short for you could get contacts.

Anywhere so there are formulas to
calculate the electron mobility and

carrier concentration and what we
found is that electric mobility and

get a concentrations are pretty
respectable so no worries about nine fifty

mobility of thin films about eight
hundred or so this very close.

Not very close but pretty comparable.

And we would like to see these Pinfield
mobility which would really make our

devices much much more sensitive.

And one more important aspect that we
found energy is really important for

sensors is that if the Nano were to reduce
this the mobility actually goes up.

And sort of the connectivity and
the character's intuition.

So it's really kind of amazing and
it's kind of animal as there are reports

on this as well but nobody could study
it very systematically we did for

the first time and this is this
result is really important because.

If you go.

To lower and what I mentioned slower
smaller thinner and thinner and

worse they can still function.

They can still get high mobility they
can have good electrical properties and

can be useful for sense of appreciation.

So these are the bandwagon still
not talk too much about it.

But we did get a chance to
model these variations and

fit our model fits very well
based on our first in charge.

I can relation which is again
a good news for sensors.

So anyway so some of these not aware if we
could actually many political manage to

get a quick oxide coating around them and
the central part of the in and

nitrate out there to be a mock side and
if we do that then they

actually could devise out of them then and

expose it to add on to what we found
that actually enter to can produce

a lot of change in electrical properties
and from minus thirty five is good for

shifts to minus fifteen volts voltage for

about ninety two and
point five point nine ppm produces about.

Minus twenty.

So in order to because any
M oxide formed around this

interim nitride gets affected by N.
or to its connectivity so

arrow talk sensing is possible
using these networks as well.

And here is some other result which
shows that down to forty five people

be still produces some effect on the.

On the Nano where if you see
although not as pronounced.

So so basically the question structure

has this oxide and I tried structure
interface at the charges and it

becomes then does this charge amount and
basically that so the sensing is trivial.

So anyway so
basically we are at the end of the talk.

What I would like to show you and
mention one more time if.

We have these nano where it can be
deflected from a barrier goes back any

patterns on top of them.

We can actually make probes and

then a scale probes like F.M.
and would be this would be very useful for

molecule or sensing probing
in the nano scale and so on.

So that's that's the whole motivation.

So all this work is leading toward that.

So here is the summary of my
talk we have demonstrated

all the good aspects of the Nano
where growth on the thing will play.

And we have found that they

can the growth can be one of us
can bend spontaneously or when.

Obstructed by a barrier and
which can be utilized for

making a structure then devices.

Electronic properties of the Nano
watch have been excellent and

really better than any other
group at present I believe we

also found that the Nano where properties
get even better if we go to smaller and

smaller diameter which
is really interesting.

In fact anomalous to some
extent in the nitride

sheller can detect two
to very low level and

taken all together and
I tried best nanostructures devices and

names pretty promising and
we'll continue our work in that area.

The future.

So thank you very much for your attention.

If you have any questions
please feel free to ask.

Please.

For in one thread and then.

Yes we would use them as
name structures our first

start of goal would be to use the of the
separate structures like you saw before

to make like a fish or paired name so
there's an interest and we would go.

We do similar way like we use
the cantilevers to detect molecules and

chemicals that were the first or
a thing we could also pull teeth on

a regular silicon cantilevered
use them as probes for air from.

Measurements because the Nano words are
also very little time of mass they live

very good for mapping thermal maps or

electrostatic potential things like that
applications let's say used in animals to

probe the new dawns at different locations
very high using F.M. on F.M. tapes.

So those are sound applications that
it would someone in the future.

More immediate application
would be to use this ship and

then when I'm kind of oscillate
it a very high quality factor and

do chemical sensing and if you have
an area of this we can do very highly

specially resolved and
sensitive detection.

Here's a resistance change
of the Nano where sue me or

you're talking with
the elegant cantilever.

The cantilever strike.

So the peers the resistance
change is due to bending right.

Want to bend it there was some strain
developing at the best and the P.

as a resistor and that would
change because it's a pull out of

material correlation will change and
will have some charge change so

that would be the Pirs a resistive
change what is the other thing you said.

The contact resistance.

The contact resistance chain if you're
making a very good only contact

that shouldn't change
too much with bending.

Good question.

I wouldn't.

Well we haven't come to this problem yet
I would say what you could

probably think of is if it is
contact resistance related change.

Probably the change will be
in the same direction right.

So if you bend it up or bend it down.

It should still.

I hoped it would change
in the same direction but

because the wrist and will probably
change in the opposite direction.

I don't know if that would be
really the case I really didn't

To be honest with you.

I didn't think about this problem but

good that you pointed out
probably take a look at it.

Sure.

I think that's where the fear and

she is are minimum three initially
the starting two one one one zero or

one man's erection and then they can
change as they probably meet some kind of

obstacles on the surface or
maybe they find some growth or

some concentration gradient and
and to rain so on but

I think the way the crystal a graphic
depictions of the nanowires are fixed.

They are determined from which direction
they get the minimum growth phrenology.

OK so if the molecules are stable in that
configuration in that direction then

that minimizes the surface of the energy
that's how it is determined America.

Probably cannot give you a more
detailed answer than that but

that's my understanding that
minimises the growth theorem of

the Greens gives free energy and
all that it's all common elements of it so

yes yes.

So it's a hexagonal interpenetrated
hexagonal structure.

Yes.

It's actually not.

I think most people most to
researchers have reported one one by

one one zero direction as the preferred
growth direction but that's non-polar one

one zero is non-polar but one one zero one
zero zero one in the pull out election.

So it can be done in both ways.

I guess but.

There are other conditions
perhaps not just

the free energy which are not
over because the site.

Yeah I think that people
should want to thank you.