So delicious degrees of pleasure

to work on the whole of

there got his bachelor's and master's.

Bill you know so

I was starting with all the tools and


the you all have a very I

think you're acting out and


I am are you currently associate

professor and Captain here


with Chris you might also like

you're going to hear him.


Recently come to the record

of what he told.


Me More about that you know Also I'm

sure you know like you know you.


Get one another award notably the best

of your award and the reward for

the early career work right here on.
The.
Right thank you David.
As I don't say.

All right how many people just came

because they smell pizza in the hallway.

So it's.

So it's here in my Who's it was in physics

very few so I mean you were engineers.


Most were writing I mean the engineers or

scientists.

It's
today.
There are groups.
Like this where people are only right.
Here so only.
Without a doubt they do really want to be.
So we're.
Very.
Blessed to.
Do
my.
Best hundred.
Years.
Around.
The life.
Of the.
Course.
You know.
Work.
Here.
So it's.
A tremendous source of heat.

If you look at it in terms

of it into the earth.


You can really have a higher efficiency

of conversion if you can operate


conversion cycle with the cardinal cycle I

see you can get by ninety five percent and


we know that our technologies

here on earth we're


close to ninety five percent conversion

and there are really two basic reasons


one that we can't reach sunlight

temperatures with things that we use for


heat to electricity conversion or we have

fundamental limits in materials that


allow the printed materials to

operating at really high efficiency.


So it's not my one material so

solar cells people have have to.


Well you know anyway solar sail

they're becoming cheaper more


pervasive in our society if you look

at a single junction solar cell we've


known since the sixty's seventy's

that there's a fundamental limit


to how much efficiency we're going to

get out of a single junction solar cell


we're calling about thirty percent

now people clever engineers and


scientists who work to put together

ideas for most high junctions himself


he don't want the best of our ideas we're

still looking at only about fifty percent

of this energy that we can capture.
So.

The source that I was personally or

basically invented in the lab I don't know


any bill as it's a lot of fanfare but

if you dig into the story


it's really interesting when you learn how

the semiconductor being built was kind of


discovered but when the solar

cell was was was made public.


Around the same time there was researchers

said Raytheon and other places it produced


for pleasure but

he didn't succeed sixty three of


the researcher regular brown demonstrated

this concept called a retainer that was


kind of computing technology area but it

wasn't consider force already conversion


it was something I was working in my

own ways very kind of classic probably.


What originally it is basically

you have some sort of internal.


And you couple that is headed to a dive so

it's rectified by so so


you have already fired him you haven't

attended receives a signal the converse

is a signal to a D.C. He's worked as.

Back in sixty three already we're eighty

percent efficient and so grounded


a demonstration where he powered a small

helicopter you know remotely the wireless.

It's written in.

The seventy's there

straight single frequency.

One percent so pretty attractive.
To work around.

The instrument conductors and starting

to understand what the thermodynamic


limits of semiconductors were or

are inspire some researchers to start


thinking about red what happens you

could take this Rick tele concept and


you can scale it down to

really small scales so


light can be part of waves

typically with solar cells


we're looking at the particle version

two electrons but what if we can convert


the wave nature of light directly

to electricity so this is an idea.


Maybe a University of Florida squid

if it were first putting out one nine

hundred seventy two.

Back then really talking about you know

what if we can make in ten as they look


like the intent is on in sex they could be

the size of the wavelengths of light and


weak a couple of energy in this return

a concept to make a high proficiency


solar sail and so back in seventy

two before we kind of knew some of


things we know now he was thinking that

maybe this could even be a perfect


conversion mechanism one

hundred percent efficiency.


So let's look at it take a step back

just kind of look at the general car no


cycle plotted out versus temperature so

you see here kind of where the kernel goes


up to some terrestrial

temperature limits and see best


demonstrated photovoltaic fish and see

somewhere around here forty percent for


multiplex and so best generated

demonstrated return efficiency and

right way produces right here and

so inspired by Bailey's thoughts in

the seventy's there's been about


forty years of research looking into

Could an optical return to be possible.


Most of the work has been looking at first

trying to understand how materials operate


as intended in the nano scale there's

been a lot of good work there and we have


the ability to make those and then also

understanding how to make ultra fast dials


there's been some theoretical work and

so I you know I have here.


I have a you know not on one hundred

percent just because there's been people


who are proposing that maybe these written

that can actually be other two percent


efficient under monochromatic illumination

under some assumptions that it's not going


to be true I mean I'll do the spoiler

they're more dynamic to limit it but


this is just where the fields bit so forty

years people working on different parts


of this work tend to have people starting

to put out there is trying to push


our understanding of where

the physical bounds would be.


And last year there was a nice review

paper so if you wanted to catch up on


return to technology M.R.S. energy

sustainability and a nicer view Garber ten


a device theory to practice this work

is really surveying the work it's been


done in the past four years and

trying to identify some of the bottlenecks


from fabrication standpoint really the

take home is where that we don't really


have a good theory because theories

kind of push the extremes of where these


devices can operate from being not so

efficient to being highly efficient and


we haven't had a demonstration yet but

there's a lot interest so two thousand and


fourteen there was this review paper

before that in two thousand and


thirteen a guy named Gary of the I was

an Electrical Engineering University of


Colorado was probably

the leading researcher for


return to work in this published a lot of

theory did I serve a book looking at where


this technology is and

some of the challenges that is there that

we're still in the way to progressing it.

Are so so I think a step back and just

go over a couple of basics just a drop


in Paris and for you between how this

works and operates and how it a true.


Missional bandgap semiconductor

part of that sector operates So


you're mostly familiar with semiconductor

swear you have a solar photon


it comes and

it excites an electron out of the valence


man over this gap where inner G.

can't exist into the conduction bay and


there's some usually some sort

of lattice vibration involved so


there's heat generated or phone on

the cause of some inefficiency right so


it's like billiard balls lights coming in

like a you know one you know seven ball


hitting a ball in knocking about producing

electricity the way the return operates


is based on a classical way physics where

you have an antenna they can resonate


the wavelength and when it resonates

what happens is you get an A C.


electrical signal produced so

current and then some voltage


then when it hits a junction if the diode

is proper is designed correctly


then it can rectify the signal into a D.C.

signal So what does it mean for


it to be designed properly correctly

Well it needs to be almost ideal so


the dial needs to have a lot of

Melanie Eddy needs to be asymmetric so

that when the current goes one

way the preferred way that it travels that

way only so that's what we're looking for


relooking pork and then tell them they

can cut a hole in the dial that's ideal.


So let's take a little bit into

the return to operation so


we mention this rectification

from AC to D.C.


electrons hopping across there are some

basic parameters of the dictate whether or


not this is possible the most important

one being the cutoff frequency so


this made a barrier for high frequency

devices in general if you can


operate something in the optical spectrum

you're you're looking at one hundred fifty


to fifteen hundred Tera Hurd's operation

so that's really really fast so


that's how the speed

is going to eliminate.

Nial structures.

They rely on physical diffusion of

matter so what I mean by that so so


my conductors whenever you make a shocking

direction or somewhere there semiconductor


based junction you're going to have charge

diffusion in and out of your interface and


that physical charge of the fusion

causes the capacitance so


if it causes a fundamental limit or a lag

in that that device ability to respond to


an input signal so this capacitive so you

can see here in this expression you have


a cutoff frequency the says basically

when the frequency the output will


start to diminish when your frequency gets

a bug's value it depends or the resistance


of your antenna the resistance of

your diode in the capacitance and


so the intent in the diode or in parallel

in this circuit may not be apparent to you


lets you really get the circuits with

what you have to take my word for


that there are parallel which is why they

show up in this expression this way so


there are very few things that you

care to manipulate in this expression

to increase this cutoff frequency.
So you can you can reduce.

Capacitance that's kind of

a straightforward thing


where you look at capacitance from it's

just a parallel plate capacitor and


I'm showing this expression for the past

this because the type of diodes that


do operate at the speeds

are typically insulated


metal structures and the reason they

operate at that speed and semiconductor


diodes do not is because they operate

on quantum mechanical tunneling So


instead of having physical diffusion

electrons are actually tunneling across


a barrier gap so with these type of

devices the capacitors could be related


you know understood understood with

this parallel play pastor expression and


you see here you have material properties

permittivity and then you have area and


you have thickness these thickness so

if you want to to reduce this capacitance.


What happens so I let's say I

want to reduce the facets and

crease the thickness.

Right so I'm increasing my frequency

bill but what's the tradeoff there.


Right Chris the thickness of the insulator

in a metal insulated metal dial injunction


that going to decrease or

increase the resistance of the dial.

You know look like.

The active interact so

if you have the increase or decrease.


It's going to increase resistance right so

so yes the tradeoffs if I want to increase


the capacity or decrease it passes that

way then I'm going to increase my of


my resistance and that's going to be

bad for impedance matching later on so


really Ariel is about the only safe thing

to play with with your capacitance so


you that means that you need to make a

diode the has the smallest possible area.


We're not going to get smaller than atoms

but we want to get close to that in order


to have a really really

high cutoff reconsider.


So that was a sense is important

because you can dissipate power


across the junction without having

you know a low enough resistance


our impedance matching to the antenna so

the balance between this intent and


that resistance will come back to but that

resistance basically you know looks like


that so it's inverse of

the derivative of the I.V. curve and


then you have this response it simply

because I mentioned before yeah like.

Yes quantum
Don.

So you're looking you're thinking at least

I'm thinking of when you say that kind of


like a resident tunneling So it's not

a resident someone dials just it's just


you know those are for that you can play

to try to make it more pressure right now


this is you could think of this is

just a simple fact for our North or


or severance type theory from

the sixty's that you just.


I have this exponential increase in

current It's on it's going to keep


increasing until you reach a voltage where

you break your night your dielectric there


so because QUESTION So we want to

quantify it as we do we have to have this


ideal dial type behavior so we want

to quantify this with responsivity So


this is just looking at the linearity

of the I.V. characteristics and so so


these parameters give us

our basic equations for


understanding one how much voltage we

expect to be dissipated across the diode


I don't know if you in the back and see

this is just going to dial voltage equals


one half the responsivity times

optical voltage squared and


then you have the power output which

is a function of the responsivity that


the optical voltage in the resistance

in the thing is important there is if


the power produced is inversely

proportional to the dial resistance so.


Some of you that do narrow scale

electronics probably already have


a thought in your mind that that's going

to be a problem because contact resistance


in them a skill is usually pretty high but

we'll get back to that.


So like I said this technology has been

around for some some time at least some


concept even back in the seventy's

there was an excellent work of MIT in


the physics department

demonstrating these were ten or.


Ten microns in the I.R. so it works we

know that it works down to ten microns and


this work was done by just a very

simple experiment taking aluminum with


native oxide and a tungsten needle in

applying a controlled pressure and


using the tungsten as the antenna in

the oxide coated aluminum toxin as


you're in my AM dial structure and

so they've demonstrated this pretty


low efficiency but it mapped out some

theory to experimental in the spirit


of Pearson's that the were the foundation

actually for the work that we do.

In our group to build on that.

But one of the problems we start moving

to point out energies that are higher


than the ones at ten microns is that

the fulltime energies are comparable to


the energy of the optical voltage

fields and so you have this higher G.


pole times and you it forces you to

move into a semi classical picture


where you have to basically consider

a little bit of the particle nature


of the light we're going to consider

the particle later in the light for


producing R R voltage but we're going to

have to consider it because it's going to


skew the distribution of

electrons in the material.

So let me show you that So so

basically this is a paper looking at these

semi-classical effects theoretically and


Rick tell us where you basically

have electrons there from


the level that are contributing

primarily to the tunneling but


then when you add some H.B.O. or maybe

a full time energy into the system it's


going to move that electron distributions

higher energies on average and


so they're going to tunnel

with a higher probability so


that's something that you would

expect to see this all time a system


tunneling effect and any diode is going

to operate is not a correct term and so


what that does is it makes you have to

modify your basic diode responsivity and


your dial resistance expressions to

consider this response time energy and


house together so if you can imagine

if it's if it's making it easier for


electronics a tunnel in

the semi-classical picture


under illumination that

resistance will be created.


And then there will be some

changes in responsivity as well.


So about five years ago there

was a nice demonstration.

Of an optical rectification process and

rice where researchers

took a gold line that they


had with even the photography and

they broke it just from a low.


Gratian so high current through

broke it to create a vacuum again so


they took this experiment may

they took advance of the crowd

to temperatures eighty degrees Kelvin.
Illuminated it with a lot of energy but

ultimately they were able to produce

a signal that was very suggestive that we


should be able to produce

an optical recommendation effect.


But it's never that straightforward

because one of the things they found


in this system is that while you

might expect it to match up with


the semi-classical corrections that their

data actually looks somewhat classical


in the number linearity of your ID her and

the take home message from that was that.


If you look at this plot this is just

showing electron transmission across that


vacuum gap the fact that these lines are

pretty flat is just an indication that.


The transmission is happening in a voltage

range where the transmission probability.


You know it is about constant and because

of that that caviar you can actually


have data for a return or

a optical rift cation


system that looks a little

classic right so I'm so it


does get into the weeds a little bit but

it's important I tell you all this because


it's very difficult to know that you

are observing out a corrective action so


most of the papers that came out the past

decade that's why it took us about six


years to work in this project you never

could make any strong statements about


what you're seeing because you can never

be very difficult to make an experiment or


system or device that eliminates all

these other factors that could be causing


your signals so

that you can really trace it down and


say aha this is an optical

replication process.

So keep that in mind but so

the evidence was encouraging got

a nice high impact paper out of it.


At the end of the day there are three

things that are limiting our ability to


make a high efficiency return and

so we can make official.


Up to quit tennis has been a lot

of work METIRIA We can make


great dials people have been focusing on

that and we can make diode a race but


ultimately there's

a conflicting scaling with area


in both capacitance resistance low

area gives you local passiveness


lawyer gives you high resistance so

you got to deal with that somehow and


then also this intent of

dialed impedance matching.


Getting that to be it's close to

what is possible can also be tricky.


So I'm going to make a case for

carbon nanotubes.

You know it's kind of like this a.

Dance with the one they're running

here you know as I got here on carbon


able to say I've been trying to find

different ways to dance with her for


years so I always make the case of carbon

nanotubes but I like carbon able to for


a couple reasons for this application

One is that they have a really


fundamental resistance when they're

pristine some nice work here


by being here another in

collaborators back in ninety eight.

They can operate as optical intent.

And there's been a good deal of work that

has demonstrated multi Walker been able to


use this relatively efficient

obstacle in tandem.


And they can even be content and

the entire electromagnetic spectrum so


their research is now they have collected

data that is pretty much shown from radio

out to.

Two to U.V.

carbonyls who can act as an incentive so

it's pretty attractive on that side.

But we know and you know because we just

talked about it if we're going to make


a return on in there that it needs a diode

and I always have a small area and a high


enough cutoff frequency and so carbon

nanotubes can have very small diameters So


that's another attractive

property we want to leverage.

Our solutions are approached and so

this is rock that was

actually accepted yesterday.


They are now technology you want to

stay tune probably will be published in


October if you want to learn

more about what we're doing.


We are using vertical rays of

carbon able to use our antenna and


as our template for building dials

with nano scale persistent So


what we do is we grow particle rays in

that we use atomic layer deposition


to deposit can formally in oxide around

each individual nanotubes in the array.


And then we cap at the very end

with another lower function metal


to make our valve structure and

this is simple as that and


I say it's a simple as that

because in our class and


what motivated DARPA to fund this work

originally was that you know these


are these are Carbondale tubes that we

grow frequently just on the momentum for


oil that you can imagine it would have

processed a simple you have a conductor


substrate like alone without these carbon

nanotubes you do another step we deposit


in oxide and then you capture with metal

those that metal it will work function and


has a high transmissibility you

have a rectenna device it works so


in our system the junction is at the top

and so you know in this system we


studied extensively we have a thick layer

tie deposited on the silicon wafer and


we make electrical contact to the back

here and then off to the side


through this disc in aluminum feel this

this is actually the top middle contact is


calcium aluminum combination it's

about forty nanometer stick.


To allow transmissive eighty but the the

optical energy in the field is absorbed


in the nanotube and what happens is this

is the junction defined at the top tips


where this rectification happens

from an AC to D.C. so you.


Got it it's a basic device operation so

we really look at a little more


detail you can see here a side view

of our oxide cord it annulled.

Hoops just aluminum oxide or alumina.

So we have a process to get relatively

conformal coating this is a nail to that's


about eight hundred metres in diameter you

see the core here and it has six C.N.C.


Ross and

you see the smooth surface of the aluminum


you know coating already a way around

this now too so you have the four


basic features that are important so

you have eight millimeters of alumina So


you kind of have a pick and

that's that is the fine for a diode.


It's well established the N.T.'s are sharp

tips that will have field emission that's


something we're going to leverage to that

your electric field become stronger tip


of the C N T's so based off of this

ten the other meter approximate


junction diameter we expect to

have a capacitance of about


one adult fair to the minus eighteen

fair ads which is really really low and


it gets us into the space where we can

start talking about having a pet or hurts.

You know rectifier so

also important like I said before is this

calcium aluminum work function so we need


to have a lower function of calcium

actually some they don't that few minutes.

You know for this device to work.

So this is where the devices look like

we make we grow C and TS We make several


finger on the three pattern devices and

each C.N.C. broke so


we can probe them and get some some some

statistics on the things that we measure.


So here's a nice question

about the tumbling dial so


this is we have a basic tunneling dial and

our basic tunneling diode diary


a very simplistic way to look at this

is that we have a multi wall C.N.C.


that acts as a metal has a work on

should have about five Evie and


when we put it into contact

with oxide in the other metal


creates this trap as oil barrier and

so when it turns


on when it gets a little unaided you have

this field enhanced that I mentioned.


It helps the feeling that barrier even

further because of the C N T's geometry


and sharpness so that it can crease the

tumbling probability in this direction.


Because effectively the top metal is flat

you don't have that feeling here to make


going in a backwards direction so you

have a high reverse bias resistance and


as a result of that you get a dial

characteristic that looks pretty good


looks will start to look close

to what you want for a ten or so


these currents are on the order of

Eric's percentage meter square but


the most important thing the look at

is that when you have aluminum which


has a higher work function than calcium

meaning that the work force in difference


between the C. and C. is lower you don't

have a device that has as much a symmetry


and the linearity as you do with calcium

and the devices don't work at the moment


they only work with the council so this is

a bare important thing you can imagine.


That in addition to just having

the shape characteristics the turn


voltage matters so at the point where the

nonlinear restarts can be defined as when


the device is what voltages are required

to actually turn the device on and


what calcium may be low enough

into the middle range where


as you'll see later some of these optical

motions just can actually turn more and so


this is just a long so my log plot

of of the diode asymmetries just


give you a different view

of how more asymmetric

calcium contact is in the human contact.

So one of things that we want to

do here is quantify field here and


so we made a quibbling

device that's planar So


we took a device we use

gold to model C.N.C.


for this and we work function in this

planar device what we're able to do is


to look at what the current is and

to understand what


role the geometry of the carbon nanotubes

is playing and everybody situate.


As you know those of you anybody who

spent a lot of time working on the.

Tom dials.

So that's what people want to know that

it's understand that you can deposit


eight nanometers of physical oxide but

you're affected box I


often will be better in there are two

main reasons that can happen.


When you deposit your top metal you can

actually have physical metals I don't use


the fusing and Xerox are effectively in

that barrier and also there's some field.


Bending of pre-treating if you will

into the oxide so there's an effective


thickness of the oxide that often

this is better than the real so so


you want to compute that so we computed

that room for Mopar pointer control in for


our C N T's and so the effective barrier

for eight millimeters of the positive


is one point five B.N.

of meters for the C N T's and for


the player device it's it's about

three data meters so there's about.


A factor of two difference which we

are privy to a field management fan

in the car many analysts and.
So we measure the capacity so

we didn't devise crassness measurements

remember we don't have systems that can


measure passed this up to the frequencies

that we're interested in so


we just measure them up to the frequencies

that we can but from this data we can see


that it's a it's a pretty robust system

we look at both the capacitance data but


we also can extract a capacity that's

there is about to our ferrets so


we thought we would get something

on the order what our ferrets but


we know the number density of C N T's and

how the passengers add up in the network


of the circuit we get to measure

capacitance about two out of ferrets so


the vast mass of this is important because

there's actually a tricky trend that you


see in a device like this if

we blow it up and look at.

Our three be able to look connected.
You basically have these nailed to an.

You have been metal below has

relatively high resistance

that connects each nail to your step.

So if you look at the problem loss of

the electron So let's say we take.


A time period close to when

the rectification happens which will be on


the order from the seconds Well

most of this metal between the C.


and T.'s is going to look like an infinite

resistor because the electrons are only


traveling at maximum to the six meters

per second without traveling fast enough


at that time scale so this record without

saying another word in other words is


that the replication process is local so

you can think about this as each of these


negative junctions you have a lead

who charge rectification and


then the electrons queue up

in the queue up and wait for


longer times and the longer times than

they can be collected in the circuit.


And so because of that effect even though

you know the date of the we're seeing


is consistent with

the replication process.


The time it which you can

collect the signal in a circuit


depends on the capacitance

of the entire device here.


So there's two different time scales and

one scales for capacitance


you have a device level which determines

when you can collect the energy out and


then you have this local rectification

happening at the junctions which is


the phenomena that we're we're interested

in and want to study a little bit more.


So that the capacitance is

one place right so we can for


we have low capacity which which pushes

us into a place where we can have a high


frequency rectifier you know the things

people smash because if we have a really


really bad impedance matching becomes

difficult to measure in the signal to even


have a signal to the interpret So

one of the things that we have in favor of


the big this isn't a favorable

attribute of the carbon nanotubes is.


Is that when you make a way

if it's a proper density


there is a collective couple

laying effect that can lower


the resistance of the nanotubes is

an antenna and you can think about this


collective intent to attack as if you had

an isolated male to me try to put a light


home a column of light in absorb

it in that isolated mammal to.


The neighbors who will read

a lot of its energy but


if you take then the able to be put in a

dance or a day when it really radiates and


just radiates to neighbors so there's

like multiple times multiple chances for


to absorb the energy so this crisis

an effective lowering of the internal


resistance for each of these nanotubes in

the rate and that's that's the study by


by several people this is one

reference you can look at for that so


you have this collective intended effect

it lowers your internal resistance.


Right you still have a poor impedance

match in our devices we have.


Somewhere or just out here or

ohm of resistance per junction and so


we have about a tenth of the eleven

impedance mismatch at a single C.N.C.


So the only way that you actually

can get a measurable signal and


this is kind of where a big

advance happen is by being able to


reliably make billions of these dials

that are connected in parallel so

that's what the mercury allows you to do.

So let's get some action yet

another question.


Here what what why would why are we doing

what we're going to do it all that.


You gave a nice tie in

to the Action Man So


those are the rectification photo

response out saga that here.


Yes it's late and so

it's basically a photo detector or


or in an extreme scenario

where we improve efficiency.

It could be a solar sail.

Well it's not it's

the promise of a return of

is.

The promise of works and you remember me

on the slide telling you that carbon able


to it's going to soar electromagnetic

energy in the entire spectrum


is is to close all the gaps

in our visibility so


this right arm is interested armies

interested in this not Purcell or


sales interested in it because you can

now or whatever frequency or wavelength


are interested in this device should

be able to give you a signal based on.

That.

I mean it's really the only

similarity is in the use of broadly


that that term because the physics

is completely different.


So I mean it's similar in that

it's there it's the search for


a broadband energy conversion

process the search for


because you don't bang gaps in solar sails

you know or limited media you have to


have the energy over the band gap before

you can convert any significant energy and


Wes one of the big limitations of solar

cell silicon solar cells is a lot


of the IRA energy just dumped his heat so

people are trying to make I arst you know


vote solar cell detectors so

when something like a Rick ten a.


It's already been demonstrated up to ten

microns being that anything from ten


microns and below you basically can as

long as you can design a proper intent


you can coupled to the E.-M. and you

have a proper dial you should be able to

extract energy from that process or.
That.
It does.
Right or.
Wrong the way.
That.
It.

Was for you I mean you're asking

a question that is very aligned with.

Like finding a market niche for

right because ultimately it's got a fast

has a very very fast response just


not going to second that nothing is simple

second fast in the market so in terms of


what's in the market already it's probably

just as fast but the question that I get


from what you're saying is what's its

real value proposition as a detector well


semiconductors one of the big

problems with having a an I.R.


detector based on semiconductors

is the need to cool it.


Because with bandgap semiconductors you

know there is this efficiency drop with

every degree of temperature it increases.

I'll show you some data to show you

that that doesn't happen in this system


it is different physics and

that's why doesn't happen so so


it really repeller from this

being at something that


we're going to be going to John Goldman

and saying hey let's commercialize it


is basic science it's really just

demonstrating that this is possible and


really understanding where the bottlenecks

are to make it better I mean we still


don't even have up into this paper

the saber came out in two thousand and


thirteen we still don't even have a good

understanding of where the current carbon


nanotubes came from and so this paper

kind of concluded that the basis for


the community looking at metallic versus

semiconducting And basically the story is


that if you don't have a photo of her in

metallic carbon nanotubes typically it's


due to a thermal or back like a thermal

electric effect where you have a band that


in fact in semiconductor tools

we're using these houses so


what we're what we're going to show

as we look at our photo response


is we're looking to show we have

an effect it's not due to thermal


fax it doesn't fit in the signature

related thermal effect.


In that if that's the signature of

an article recommendation process so


this is the date I want to show you that

relates to kind of finding a a value


proposition for this is a detector

is that you look at this


data where we take an I.V. stand of our

structure from five to seventy seven see


is very weak very little temperature

change or dial change in temperature but


if you look at a similar

conducting see in T. base diode


just like any other subject of activity

here all this is a log scale here so


long to my log why you see a lot of change

in the ID characteristics with temperature


which ultimately relates to deleterious

effects in the dial structure.


So here's here's what we did so we have

a device we have a title semi-transparent


we eliminated it with green related

with red we eliminated with solar


we filtered solar we did

it several times and


this is a typical response this is just

looking at current density first voltage.


This plot can be misleading because

the voltage scale is really large


whereas because of the efficiency of our

system of the stage when I show you some


words doing optical replication is

really down here in this range and


another thing to note from

this plot is that the red and

green line to line up the grain is.

Three times lower intensity even the red

in this is just indicative of a full time


induced coma only this is this is this

just tells us nothing about the process


of energy conversion just says that when

you have a higher energy photon you have


you have a higher tunneling current

result from that which is expected.


So just to put this in

perspective because of this this.


This data the prior study I showed you

gives them a female technology looking at

optical reputation in a vacuum.

Class Matic Manal gap

a crowd in temperatures


had to use an illumination

of two thousand.


Times the illumination of regular solar

lights in order to see a signal so


and so

really their results than our actual solar


am you know one point five M. solar and

you know you know what level laser


light which is pretty significant

in that room temperature.


So you have to blow up the I.V.

curve like I told you.


And you need to look at where the energy

or power dissipation is happening and


so basically if you had a solar cell

you'd be looking in this fourth quadrant


people are used to trying to figure out

how to get a feel for actor where you get


a nice square in that shaded area will

be your energy conversion or your power.


The different physics of the rick

ten dictates that you have power


rectification in the second

quandary because your I.V.


curve this is the dark her ships so

that the more no linear your ID curve is


the larger your field prac Bill factor so

so this is looking at the ID scam and


these lines here are just showing you

where the measure open circuit voltage and


short circuit current match up with where

you would extract that from the ID scam so


these are separate measurements so you see

that for her red laser you see from green


laser that you have a smaller voltage then

you have a red laser which is different


because typically with higher full time

energies in a classical solar cell picture


you'd expect larger voltages and that's

consistent with a return to never solar


you see similar voltages that you see for

red light except you see much


lower currents which is also something

that's consistent with the idea that.


The reason that you're seeing the lower

currents when you have broadband versus


heading is you're mixing losses in your

intent to dial conversion process and


that's consistent with where people put

outlets in the theory so I just want to


relate this back to this paper here

this is two thousand and thirteen.


Very painful return which was kind of

predicting that you should see this energy


harvesting in the second quadrant and

basically it's the shifting of the I.V.

curve that causes the power dissipation.

So the full time assistant tunnelling this

read is the resistance the differential


resistance under illumination in

the black is when it's dark so

there's about one hundred fold decrease

in the resistance of the diode due

to this all time assistance only.


And we can predict that theory so

it's pretty consistent.


When you look at the measure voltage A So

these are the measure voltages for


a red solar grain the interesting

trend is that there are one or


three of these voltages are much

higher than a thermal letter


voltage that you can produce so

live a struggle just.


And you see that monochromatic red light

you get the highest voltage output

of a solar in the hollow by migraine.

So we stress in the Thousand Magnussen to

where some decide this is just looking at


us doing some thermal letter back

because I will spend much time here but


we spent probably eight months

in our lab just trying to


do a sanity check to make sure that

we were seeing thermal signals but


you can measure basically this crossing

of the zero of your voltage with position


from the negative electrode indicative

of being a normal after effect but


we collect this data at about

one hundred twelve times and

to be correct on a device.
So that responsivity if you remember.

Earlier science that was an important

brainer inter-meeting that the voltage and


power dissipated so this is the dial

responsivity under lumination for


the grain and for the red lasers and

what I want in your system


air this to what people see in

the literature in the take home is that


a lot of people most of the workers work

from two thousand from one thousand nine


hundred eight to two thousand thirteen the

receipt of this publication was focused on


optimizing dial structures and the type

responsivity as they were getting what


would be on parts what we're getting in

the argument always been that we can make


a dial that has a high responsivity

it should work isn't a return.


So it's a confirmation

a look back at that.


So another thing that you can do

once you have the voltage data is in


the responsivity is when you multiply

the voltage the measured over a certain


voltage by one over the responsivity

it gives you the optical voltage so


this normalization is actually quibbling

to extract the optical voltage in the car


may also be intent on and

as you expect you can draw a line through

data at the same illumination wavelength

a different powers and

these dots are from.


The open markers are from

the same device and


the red solar and green that you saw

before and close markers are for


a different devices that actually

tells you a bit something.

Something else this device that has closed

markers has a higher resistance

than the device has over markers


which means it has a lower impedance match

for the higher impedance mismatch which is


you know consistent with what you

would expect in these devices.


So when the light emitted bangle do some

polarization studies this is just kind of

a standard check you want to do to make.

Or that you see some classical

intended behavior in the structure.


So you're a messier engineer so see Carol

less about physics you want to know

how to use a device how fission is it.

You think that efficient yet but you know

what nobody's ever reported in the first


and see let alone measure an open circuit

vultures were happy about their progress


But we were happy Also if you want to

buy efficiencies for the first time so


you know this plot is table was just

showing Dark an illuminated resistance and


responsivity data showing a measure open

circuit voltage source circuit current


take home stories of the device gives you

about sense of the minus five officials


see right now and the major limit

on that is the dial resistance so


contact resistance even though we

have all these diodes in parallel


contact resistance is still something that

they were working on too to eliminate but


we're encouraged because you have

a favorable scaling because for


every order of magnitude you can

decrease the contact resistance you


increase the efficiency by

an order of magnitude so.


We know that we can make carbon nanotubes

our contacts to them a lot more

efficient than what we do right now.
So just to summarize.
The story here.
We made a device.

We made it by using Net all of our

repeat word nano scale fabrication


look fabulous using a all be made

of metal it's layer metal diode

in the tip of these carbon nanotubes.

We see how rectification in

the second quadrant we see


the best results matter

monochromatic red light.

Coherence matters and rates and.

There's this link that is

higher still trade off so.


Good questions about you know if

you want to use it as a potent.


Well one of the limitations is whatever

metals are connected to the circuit


matter in terms of your time response but

you want to use it as an energy conversion


device so we can make it more efficient

and that's up as a matter life is much.


Quicker than ten affected ten

effect is very important so

to emphasize that point the.
Two thousand and ten.

Malick gold medal gap optical

replication paper also estimated


the optical voltage under ten thousand

times higher lumination intensity to be


about thirty three million votes they were

producing in their their goal junction.


So a ten thousand times

less illumination intensity


we are producing about one hundred

millivolts as a maximum so


that tells you that the intent in

this race structure is at least ten


thousand times more efficient then an

isolated gold mammal junction than ten so


that's important because that allows us

to route here Tara because when you have.

Your antenna and

your diode in parallel when you want

to determine your cutoff frequency.


The lower resistance is one to

determine the resistance used in


the car frequency equation so

our coverage was used the term and


by the by one over the internal

resistance in the capacitance and


because both of them are low

allows us to achieve.

About a pair of herds of operation let me.

So the Power Point is there's more work

to be done because we still have core and

heat and matching.

In a single C.N.C. junction you know we're

using carbon able to that probably could


be a little bit more pure and there was

trucked or there there clearly was then so


that's something we're working on now is

opening the ends to try to make cars.


Untaxed some of the inner city and

sea walls.

I thank you for your country.
But.
You are still.

Joe Johns they are of course

you know that they have their.

Own really good one one that.

You think you'll have a similar

problem here how do you envision.

The problem or.
Are you and how.
We're like.
Well we saw it so we sold.

This this year what else were

good some modeling assuming


what happens we have mixing of

different wavelengths and so.


There they have the theoretical work and

I believe the most.


Under ideal dial characteristics their

prediction for energy conversion and


monochromatic is one hundred percent right

under ideal dial under ideal dial for

broadband solar it's fifty percent.

So you have same problems but we're hoping

that that upper limit of the higher.

We're hoping.
You're.
Here.
Also asking questions not so.

That's that that's the expected trend as

long as the antenna stays in the resistive

limit so are our C N T's or ten microns so

everything that's being absorbed is

seen a resistive it is no resonant into.

Nothing.

You can do some things where we we

actually try to match the scenes to


the light then we might see some resonance

next year will be interesting as well.


Right yes a typically though when you

have a disappointing resistive antenna


that's the trend you expect

you start going to longer I.R.


wavelengths in the voltage will start to

increase the current may drop a little bit

the motion.
Of.
The.
Temper said yes at what it.

Sold so that we can posit twenty

millimeters of calcium capped with


forty the enemies of the woman of it

in it's in this chorus it's porous

also.

It's hard so we need a better top contact

so the resistance that top contact is is


that just a sheet resistance

is pretty high possible so so


we have we have a dial resistance is

a problem as she resistance is a problem


when you saw from the the original

picture of the device that we were making


electrical contact here so electrons they

come across here have a novel the sheet


they have to go down this this wall and

then over here before they collect it so


the device in the circuit optimizations

That's a lot of work ahead of us.

So.
Let's

go.

To


that to reduce the concept resistor.

Here just because I guess if it will have


a carbon fiber lattice structure is but

the idea is that any fiber


carbon whatever if we can make

better contact to the actual atoms.

We should use the concept resistor.

Which we're trying to work on that right

now because we with our whole to our


structures going then we can try to do

that to address that is just to cut


the caps off and see if we can actually

make contact to to the end or was


I mean it's not it's not a trivial thing

to do but it's also not a trivial result


because people who have been successful

at cutting the caps and in making that


contact have have been able to drop the

resistance by eight orders of magnitude.