That's.

Right a chemical engineer and you know.

The other.

He

already.

Number more.

Right.

Here.

Thank you very much Martha can you all
hear me the microphone on very good

I've been told I need to
wear two microphones so.

I want to thank Martha and all of you for
the invitation to come here today it's

a great opportunity for me to see some old
friends and meet some new people it's been

a day packed with great conversation so
thank you for the invitation to come here.

And like to tell you a little bit about
what we're doing in Minnesota in the area

of flexible Tronics and one of the things
I like so much about this area as Martha

mentioned is it's a really nice blend
of chemical materials engineering.

And there are some very significant
challenges that I'll try to.

Motivate the talk will be less analytical
and more about synthesis right so

I think you'll find it accessible.

Let me start by acknowledging my
collaborators in particular I in this work

I work a lot with Maureen Francis who runs
our coding process fundamentals program in

Minnesota and
also Timothy watch polymer scientists and

Chris Kim helps us with
some circuit design and

Jennifer Lewis at Harvard to provide
some electronics that we use.

Most of the work I'm going to tell you
about has been done by a graduate student

and John is about to join three M.
and a post-doc.

So let me give you my outline here.

I'd like to start by motivating why
we want to do printed electronics or

large area flexible like Tronics and

then I'll tell you about some new
materials some gel electrolytes that we

find are very useful in designing
flexible a trunk circuits I'll tell you

briefly about a kind of device a squishing
device that we build in my group and

which we also use in principle it tracks
then I'll show you some results for

printed circuits and then I'm going to
spend actually the last third of the talk

talking about a self aligned manufacturing
process because I think this is really

from my perspective anyway the great
challenge in printed electronics is how do

you how do you do registration
of multiple layers of material.

So first a little motivation I think of
course on this campus there's lots of

people interested in thin film electronics
organic electronics one of the reasons

we'd like to be able to do thin film
electronics is to take advantage

on flexible substrates is to take
advantage of all the roll processing and

in particular role the role processing
of solar cells seems potentially

an attractive application but
I think also if you think about displays.

You know of course in our you know all
of our televisions at home we have

a backplane behind the display full of
transistors that turn the pixels on and

off that backplane is built on
glass if you wanted a more robust

display one that was basically unbreakable
you might want to put it on plastic and

if you put it in plastic.

You then have a limited thermal budget for
processing materials you can only go to

maybe up to two hundred degrees C.
and so that changes the material

sets that you have available to make
that backplane but moving to plastic

also allows you to think about moving to
roll the roll continuous manufacture.

In by electronic So lots of people
are interested in flexible circuits and

here I think for a lot of applications
it may be completely sufficient just to

make thin silicon chips and put them on
a plastic substrates you may not need to

do a whole new kind of
manufacturing methodology to make

bioelectric circuits wearable
delivery patches and so

on the key point I want to make is when
you think about flexible Alectryon and

printed electronics we need to target
applications where we're thinking about

large area wherever you need large area
you're not going to do you not going

to use conventional single crystal silicon
if you can use silicon you will but for

a large area kinds of applications that's
that's not likely to be practical so

this idea is emerging of using
the real processing on flexible

webs to put circuits on to flexible
substrates and this is also

driven by the fact that there have been
some great advancements in electronic inks

in the last ten years you can buy metallic
semiconducting and insulating inks and

you can imagine and you can pattern
them using in Jet or aerosol jet or

screen printing or give your printing and
so on and so you could imagine then that

there you could build design some process
by which you would pattern these materials

precisely to build up functional circuits.

But I think there's a couple of key
challenges that we have to keep in mind

with this with this idea and the first
is this issue of aligning layers of

electronic ink if you think of a typical
thin film transistor structure and

I'll describe this in more detail later
but it's a multi layer stack of materials

and you might like to separate this
let's say these are two metals so

be a metal let's say an insulator a
semiconductor and two metals and you might

like this dimension here to be microns
maybe a micron So now you've got four.

Layers of material that you have to
deposit with precise thickness and

then register materials with micron level
precision on top of one another and

you want to do this in a continuous
fashion on a deformable substrate that's

moving right so
that's that's a huge challenge and

we don't have a way right now of doing
that there isn't a printed technology

to roll that's out there so
I think that's challenge number one and

that's really what I'm going to talk about
in the in the last third of the talk but

the other challenge is having
sufficient circuit performance.

You're going to deposit electronic
materials from solution and

when when the material dries when the ink
dries you want a good semiconductor

left behind something that's going
to give you good performance and so

that's a that's a major challenge here's
another way of looking at the performance

issue this is a little bit dated now but
the trends are still the same if you look

at the switching time of a transistor
made by printed electronics

versus supply voltage which you can find
is there's always a trade off and then and

this is true for silicon too there's
always a trade off if you want fast

witching times you get apply more voltage
The problem is is that because printing

process these are imprecise the device
dimensions are big thicknesses are not

well controlled you tend to have to
apply rather large voltages to get let's

say microsecond switching times and so
where you want to be down here you want to

be down here low voltages
comparable to battery supplied.

Voltages and
you want to have short switching times and

I'll say this if you could develop
a technology that would allow

you to print circuits over large areas and
be clocked at let's

say megahertz which is not fast by today's
standards but if you could do that I think

there's a lot of applications that would
open up so the speeds we need to achieve

are not terribly high sort of
microsecond type type switching speeds.

So those are the two challenges one
is how do you align the materials

the second is do the materials perform
the way you need them to perform to make

functional circuits.

So let me start to get into
the technical aspects of this.

And think about the basic building
block the basic building block for

thin film electronics is a T.F.T.

or thin film transistor and that's shown
here you have a semiconductor layer

deposited on an insulator there's
an underlying gate electrode and

there are two metal contacts of
the semiconductor called Source and drain.

And the idea behind this device is that
when you put a voltage on the gate

you induce carriers in
the semiconductor because the gate is

capacitively coupled to the semiconductor.

And so when you do that essentially
increase the conduct of any of

the semiconductor and now you can flow
current from the source to the drain so

the device is on devices on when
I apply a voltage to the gate but

if I don't apply voltage to the gate
I have no carriers here and

I don't get any current So it's a switch.

Fundamentally the current that flows
from source to drain is just a function

of the charge concentration in the
semiconductor channel times the average

speed which is given by a material
parameter called mobility so we need.

We want current to be large so we want
large charge concentrations at low voltage

and we want high mobility even Tiriel.

But a big problem as I showed you
in the previous slide is generally

you have to apply large voltages to get
fast witching times right that's like

saying you have to apply large voltages
to get large current all right and

the speed of the carriers depends on the
voltage that you apply between the source

and the drain so we want to dial down
the voltage We don't want to have to apply

that much voltage and
the way you can do that is by using

a high capacitance dielectric here so that
for every volts you apply at the gate.

You get a large amount of
charge in the semiconductor So

this makes you think about OK well what's
the capacitance of the dielectric it goes

is the dielectric constant
over the thickness so

we could make the dielectric very
very thin fifty nanometers thick and

then maybe we could use low voltages
this is the voltage the the carrier

concentration in the semiconductor is
given by the capacitance times they gave

the so if I give a if I have a big C.
I can use a smaller.

Which But the problem with going very thin
is that in a printing process it's hard to

control that it's hard to make very
thin layers in a printing process so

the other way you can go is boost
the dielectric constant you know use high

dielectric constant materials that will
give you the high capacity you need here

to get a lot of charge if I have a lot
of charge provide that means I can lower

the gate voltage right so
the strategy that we've been pursuing for

a while is to use electrolytes in
this kind of an interesting choice of

dielectric material so here I show you
basically a gel electrolyte and I'll

describe it in more detail in a minute but
basically I've got polymer matrix and

I've got dissolved ons and here's
a semiconductor very typical one poly.

And here's a source electrode
to drain here's the gate so

now if I wire this up and I apply
a voltage to the gate these ions move and

they induce carriers in
the in the semiconductor and

I get a current all right and the
switching speed of this device is actually

in the end fundamentally going to be
limited now by how fast I can polarize

the Zions how fast I form double
layers if I can form a double as fast

I'll charge up the semiconductor fast and
I'll get fast switching.

Hopefully that hopefully
that makes sense right so

this is this is what we call
an electrolyte gated transistor and.

Now the ions can also penetrate
into the semiconductor and

when that happens this is basically.

What's also called an electric chemical
transistor you have an electric chemical

oxidation that's going on.

All right so let me just show you what
you can do so you can you can print

these devices pretty easily take a plastic
substrate like kept on and I'll show you

the printing method we use in a minute
you print a gold nanoparticle ink

print a semiconductor in this case this
is called P Q T twelve it's another poly

we're going to print a gel electrolyte
that I'll describe in a minute and

then I print a conducting polymer as
the gate this stack of materials I print.

And this is an early
version number of years ago

this is an all printed device we printed
the source we print to drain the printing

technique was aerosol jet printing
which I'll describe in a minute and

then here's the reddish semiconductor
material the clear material is

the gel the gel electrolyte and
here is the gate.

And what I want you to see here on this
slide is that these are low voltage

devices so this is the current that flows
from the source to the drain through this

semiconductor as a function of the voltage
I put on the gate right here and

what I want you to see is there's five
orders of magnitude change in the current

from the off state of the on state and
it happens in the space of a Volt right so

the device is fully on the ball.

So you can print these E.G. T. devices and

they operate at low voltage and
they give you good output currents.

Let me show you that you
can also watch them switch.

This is a little video clip
which I hope will work.

So what you can see here is
the semiconductor changing color as we put

charge into and out of the semiconductor
basically the clear box here is the gel

electrolyte and this is an undersized
gauge electrode allowing you to look

through and see the Semiconductor which
is electrochromic it's a pike conjugated.

Polymer semiconductor and
when you charge it it changes color.

So.

We can make these switches.

And we can make them operate at low
voltage Now let me talk a little bit about

the printing so for this work that I'm
telling you in the first part of this talk

we use aerosol jet printing
an aerosol jet is basically

a high tech spray can you
make an aerosol out of any so

typically by Sonic cation you would train
nearest all in an inert gas stream and

that brings the you bring the aerosol over
the printhead where second sheath gas

accelerates and focuses this
aerosol on to the substrate and so

you can draw lines and basically if you
can make an aerosol you can print which is

why we like the technique very much you
can print carbon nanotubes you can print

silver nanoparticles you can print
zinc oxide precursor materials

you can print graphene you can print as
I showed you probably five things and

conducting polymers and gel electrolytes
So it's a very versatile printing

technique if you can make an aerosol you
can print and you can get sort of ten

micron wide lines.

I'll tell you about a different
strategy later OK so we can make

these electrolyte gated devices we can
print them they work at low voltage Let's

talk about the electrolyte the electrolyte
we use we develop with him logic Minnesota

is based on an ionic liquid and a
structuring block polymer so I am a quick

words you're probably familiar with these
are big organic salts they have a large

organic cat ion and a large organic and
ion because the ions are large and

there is symmetric the melting point
is depressed below room temperature and

so at room temperature you
have a liquid salt and

it has very high on
an ionic conduct of eighty.

But not electronic on the T.V.
as high on a convict of eighty and

it also therefore has high capacitance
because if I put that ionic.

Between two metal plates and I polarized
the metal plates I get electrical

double layers and
I just get huge huge capacitance so

we make a gel out of this by
adding a tri block polymer

that has sticky polystyrene ends so
it's polystyrene block.

BLOCK polystyrene and when you put this
in the eye on a quick witted the A blocks

the polystyrene blocks associate to
form glassy nodes cross-link nodes and

the polar P M M A chains spin
the gap between the nodes and

that that gives you your
basically elastic net so

when you stretch the Joe you're stretching
those Pima may change in the ionic liquid.

All right so that's a very nice weight of
it to make a gel Here's another look at

the structure chemical structure of
an ionic liquid that we use a lot it's

in a midazolam sulfa made salt.

I on a quick woods are often clear you can
see here perhaps a miniscule between this

ionic liquid and distilled water so
they are electric chemically inert

they have wide electro chemical Windows
they have essentially no vapor pressure

very nice materials you can put them
in a vacuum system and pump on them for

a month and they don't evaporate.

So very very interesting electrolyte
materials so we have the polymer and

that that gives us a gel and we can tune
the elastic modulus from a killer Pascal

to a mega Pascal depending on
which try block polymer we use and

the molecular weights and
how much polymer we add to the ironically.

So let me tell you a little bit then
more about these materials because

I think they're interested in there
in their own right if you take this

material maybe make a slab and
you sandwich it between two metal plates

you could measure the capacitance
by impedance spectroscopy

this is the capacitance versus thickness
and you get a very large number ten

micro ferrets per square centimeter
that's a that's a very large number.

If you measure the resistance
of course these gels have.

I have an electrical
resistance due to ion motion.

The resistance will scale with thickness
the thicker I make the gel the higher

the resistance right and

it's just what you expect from sort of an
ohms law picture so capacitance specific

capacitance does not scale with thickness
because the capacitance it has to do with

the formation of a double layers
at the metal in jail interfaces so

that's thickness independent It doesn't
matter whether it's this thick or

this thick which is different in
a conventional capacitor if I double

the capacity of the thickness I reduce the
bastards in half here that doesn't happen

so I have high capacitance that
is independent of thickness and

I have resistance that depends on
thickness if I multiply the two together

I get an approximate estimate of the
polarization time the R.C. time constant

of that gel if I just model the gel as
a resistor and capacitor and series.

Can be microseconds so if I make
a gel that's two microns thick I can

polarize it in a few microseconds and two
microns this is still really relatively

thick That's a thick film certainly
something we can code for print and so

what we have electrolytes that
are polarized the ball in microseconds and

chemically inert transparent very nice for

the kinds of applications
we are we're interested in.

So we have a pretty good idea of
how to engineer these materials how

the choice of mid block and
block actually influences the properties

we can make shells that have melting
points above one hundred C. simply by

making the polystyrene blocks longer than
about four thousand molecular weight and

then the mid lock the key thing we need to
think about here is the glass transition

temperature of that mid block because it
turns out that the mid block which is

bathed in the ionic liquid provides
resistance to ions right so

if I have a high T. G.
The chains are sort of stiff and I have I.

Higher ionic resistance to that shell than
if I pick a mid block with a low T.G.

low T. G. provides less impedance to
the eye in motion so we can actually

go to the properties of these gels
Here's an example the blue data points.

Are for a gel where the mid block
molecular weight is one hundred fifty

thousand and
the try block here is polystyrene block

poly Ethel accolade that's Ethel accolade
this is this is Ethel accolade here and

then this is polystyrene So
this we call this as a ass.

And here the mid block the F.L.
accolade is varied in weight percentage.

And we've measured modulus of
the gel versus ionic conduct and

you have this tradeoff right if you
want high modulus you add more polymer

that raises the modulus but
it reduces the comic to Vittie So

generally we want higher
ionic activity and

in the here it gets up to seven times ten
to the minus three percent a meter of good

value at ten weight percent but
if we switch the mid block con

molecular way we reduce it from one fifty
to twenty three thousand we get a big

bump in the modulus even while keeping
the conduct of any the same and

the reason is that now all these polymer
chains are shorter and I sensually have

a tighter mesh if you think about these
these the cross-linking as a two D.

grid right when I have big molecular
weight I have a big loose mesh and so

I have a low modulus when I reduce the mid
block weight I have a tighter mesh more

sort of patches per volume and
I have a larger modules so

that's a that's a nice result.

We can also.

Compare for
example gels made with blocks like effort

with a low two G.
to gels made with blocks with a.

T.G. So this is the S.M.S.
polystyrene polymath from a factor

late polystyrene this has a T.G. of one
hundred this F L accolade has a T.G.

of minus twenty four and
what you can see is the S.M.S.

gels are here in blue the S.E.'s gels are
here in red and what you can see is for

a given weight percentage
you get a higher.

Modulus but also.

Let's let's look at it this way if I take
this weight percentage and I add and

I switch to from S.M.S. to S.C.S.

actually my conduct of it goes up at the
same weight percent that's what I want to

see the connectivity goes up by a factor
of five maybe even a factor of ten in

some cases simply because I've
lowered the T.G. of the mid block.

So we know how to tune the properties
of these of these gel electrolytes and

in general General Electric
lights are interesting for

lots of applications various kinds of
sensors electrochromic devices and

these gel electrolytes that we make a very
high on a comic to ease up to ten to

the minus two seaman's percent a meter
only acidic electrolytes or higher.

So.

Now we return back to.

Devices and printing.

So the device I showed you before this
printed E G T was what we call a P.

type device that means it switches on for
negative voltages on the gate.

But we need to actually have N.
type devices and so

we need in type materials one thing we
can do is make an ink based on an equal

sync oxide precursor we printed and
then a needle it and

we get as ink oxide film and we can make.

Good printed zinc oxide E.G.
TS on plastic This shows you

now a device turning on for positive
gate voltage very low histories this

orders of magnitude change
from the off to the on state.

And these ID.

Curves look like what you would expect so
we can make any type devices and we can

make P. type devices and again these are
low voltage devices working about of all.

So we've got you can start to
build circuits you can make

complementary circuits where you have
a P.T. device and then type device and

these are the kinds of inks that we load
into the aerosol jet printer and we could

make complementary circuits so this is
an inverter this is a piece which and

then types which this is the output
voltage so this is the voltage at this

node as a function of the input voltage
that's on the X. axis and what an inverter

does is it flips the voltage here where at
the high voltage on the input low voltage

on the output and when we sweep lower
here we trip high on the output and

I just want you to see that you get the
behavior that you expect when the supply

voltage is two volts this trips at
one Volt which is what you want.

So we can make good inverters and
we can switch them and

in this business switching it
a few killer hurts is good so

here we are switching this
inverter at five kilohertz.

As I said it's kind of a target as it is
a megahertz So we have some work to do but

but that's a nice initial result.

All right.

The other thing you can do is start to
build circuits like a ring oscillator this

is a circuit centrally when you energize
it the voltage output at this node

oscillates in time and from this kind
of oscillation in time this is milli

seconds here you can back up the delay
time first stage for this signal and

you know you get something that's
fifty fifty microseconds or so

so we can actually start
to make circuits that.

That have reasonable for formants
reasonable speed and reasonable stability

this is an inverter switch for
twenty two hours continuously and

there is some change in the and the output
voltage in red but it's actually.

A reasonably good start towards stable
operation I'll show you quickly.

You that you can make these
devices with carbon nanotubes

if you if you actually make these devices
where you print nets of carbon nanotubes

you can make it inverters and
I'll skip this slide and

I'll just go you can make ring oscillators
the ring at twenty killer hertz or so and

in this case that the late time for
stages is five microseconds so

you can actually start now to get down to
switching times that are about as fast as

you can polarize the gel right.

OK so other groups have different
strategies to print electronics and

similar results are seen by by other
groups the field is really I would say

progressed in terms of the quality of the
inks that are available and the kinds of

device performances that group various
groups can get so that's nice to see.

So now you can find printed circuits
that switch in sort of the microsecond

to tens of microsecond range and they work
at supply voltage is less than three volts

and as I argued before you want to be down
here because ultimately you want to use

low voltage power sources like printed
batteries so you need to be below two or

three volts.

OK so now let me switch to this
I think more challenging issue

which is how do we make a system it's
one thing to make a transistor but

the fields not going to advance unless
we actually build systems so for

example it would be wonderful if we
could print an amplifier like a whole

operational amplifier would be wonderful
if we could print in the analog to

digital converter this is
a very simple circuit we built

that its function was to basically
switch the selector chromic pixel on and

off that's all the circuit was to do for
a given input voltage and this particular

circuit had a couple of dozen transistors
in a dozen capacitors or so and.

Sisters and
we made it by aerosol jet printing and

it took my student two days of basically
continuous work to to make this and

she made two of them at the same time so
this is just the simple circuit layout for

this and the precise operation
isn't what's important here that

the point is that you know we need
to go far far beyond this to make

printed circuits on on
flexible substrates.

Something real.

It occurs to me while I'm talking that
maybe I should motivate the printing

aspect a little bit.

What's nice about printing is that it's
an additive manufacturing process so

you only put the inquiry you will want to
go where it is generally in a clean room

you're doing subtractive processing
you put metals down everywhere and

then you at your off metal in certain
areas and you just leave little traces of

metal behind so if you thinking about
large area electronics where you might

have to use lots of material and
subtracted away that's that's you know

it's not very attractive from a cost
perspective it's not very attractive from.

A sustainability perspective so you'd like
to have additive manufacturing processes

and printing does that but the problem
with printing is that it's low resolution

I showed you sort of ten twenty
Micron features here by aerosol jet

it's hard to do any better than that
with any other printing technique.

So how are you going to
to pattern materials

with precision high resolution and how
are you then going to overlay that that's

the key the key question that I
want I want to think about here so

here's here's just
a picture out of a proposal

where we said OK look at the very least we
need to be able to make a NAND gate to P.

type two N. type transistors we need
to be able to make an inverter and

we need to be able to to
connect them right I mean.

To be able to take this output and connect
it let's say to this output so how do I

how do I make the interconnects and
how do I overlay all these materials and

we thought well we could always line up
the P. type materials in one channel and

then type in another and we could
have routing channels and tracks but

still at the end of the day this doesn't
give you a manufacturing approach so

what I want to do know is think about
a self aligned process a process where we

do all the registration in the first step
and then we think about how we deliver

the sequence fully afterwards and
what I'd also like to argue here is

you know from a chemical engineering
perspective working with liquidating.

Is very interesting we can take
advantage of differential wedding we can

take advantage of capillarity and working
processing electronic materials from

liquid inks this is kind of a paradigm
change it's a shift in the way you think

about making making circuits so this
idea here we call scale which stands for

self aligned capillarity assisted
lithography for electronics and

what we intend to do is a boss
a substrate take a plastic substrate

slot coat let's say a thermal
plastic polymer or maybe a U.

the liquid U.V. curable resist we now
imprint into that thermoplastic or

imprint into that you'd be
curable resists and expose so

that we make a pattern that
consists of reservoirs channels and

as I'll show you I'll show you device
cavities so we do that all in one step.

And the device cavities as I've showed you
the device is like a thin film transistor

has four layers in it so that imprint step
in that device cavities actually going

to have complex topography it's going to
have the possibility for four layers but

we're going to make that all in one
shot and we're going to bring the the.

Feeder channels in.

At different elevations so

that if we drop the ink in the right
sequence we first put in metal then we put

in semiconductor then we put in dielectric
then we put in metal that's the idea.

And we think it's additive it is that
if we think this is a nice approach

to do a self aligned process we think
it's scalable in the following sense

you can imprint over large areas and
people have shown that you can imprint.

Down one hundred nanometers features so
you're scalable to small sizes but

you're also scalable to a large area and
so

it's scalable in two senses of the word
and if we can if we can implement it

implemented role the role it has
the opportunity to be high throughput.

OK so here's here's just a very
basic demonstration if we take

these are imprinted channels in
a in a poxy a U.V. curable poxy and

we use this sophisticated eyedropper here
basically to drop some ink into this

reservoir you can see a disk here that's
an indentation and because of refractive

index matching when the ink starts to
flow you lose sight of the channels but

when the ink starts to dry some seconds
later you get the channels back and so

that's the basic the basic idea we're
going to use capillarity to deliver

the into the channels So
let's think about how this might work and.

We'll start with just metal lines and
see what we can do with metal eyes.

In this process we're going to end up
spending a lot of time initially in

a clean room to make a silicon master so
we use conventional micro fabrication

to make a silicon master which we then
copy with P.D. mess and in a way that's

quite standard now so we make a rubber
copy we peel off this copy and

now in principle then this is wrapped on
a roll but I'm just going to show you.

A discrete process so we now.

Imprint in the a poxy U.V.
cure and detach so now we have a solid.

Cross-link film with the imprint.

We then take in jet.

And drop it a precise amount of ink a
precise volume of ink into the capillaries

the ink flows and initially it
looks like this in the channels but

of course it dries because it has some
certain solids content some solvent

content and so you get a skin
of silver in this capillary and

what you want of course is a solid line
you want to solid conducting line so

now we're going to electro loosely plate
so we play copper into that line and

we get we get these these copper lines.

Your S.C.M.
images of a capillary with silver.

Lining and you can see it
here this is all a poxy this

is the capillary channel
with silver in it.

We then plate this is
the plating chemistry.

Very simple to do you could imagine
this role the role is substrate for

certain amount of time and in a trough.

We play copper and
this is a copper filled channel so

this is now a copper wire nominally two or
three microns wide we overdeveloped it so

it is grown out and mushroom over the
Channel and this is a fib cut focused and

being so you can see the nice
cross-section so we can make you know

micron level conductors
that are solid this way.

We could track this is a function
of time that's not so important but

if you measure the resistance a person's
length of course you should be able to

back out the resistivity in that shown
in blue here and the point is it's

the conduct of it is fifty
percent that a ball copper.

So we can make good conductive lines in
this very simple and this very simple way.

So how would I make
a resistor that's got to be.

The most basic.

Element So what we do is we imprint.

This three reservoir structure.

We drop metal.

Into these two reservoirs it flows
to the ends of the line here and

then we put in resistive material like
carbon black to make the connection

between the two metal lines and
this is what it looks like and

we can measure the resistance so
I should show you I.V. curves but

they're very nice very very linear and we
can in this case this is a sixty four kilo

resistance we can tune the resistance by
which material we use here whether we

use a conducting polymer we use carbon
black we can tune it by the thickness

of the material we can also tune it by the
separation distance between the electrodes

so we have a way to make
make a resistor by.

Additive printing techniques.

How do we make a capacitor Well
here's a here's a structure

this is again in a poxy So
what you're looking at is an S.C.M.

image of this imprinted
a poxy layer after.

That was cured on the stamp and so we
have been introduced to the Dellec trode

structure and we have this other reservoir
this raised reservoir we deliver metal to

these two reservoirs we drill
the liver are gel electrolyte

to this third reservoir and we can measure
capacitance as a function of frequency to

make we can make a pastor's.

All right.

So now.

Should you we can make resistors
capacitors we have recent work on

diodes what I want to show you is
that we can make transistors and

I'm going to use a different architecture
what I've been showing you before

is transistors with a gate aligned
over the semiconductor channel so

this is the source and semiconductor and
normally the gate would be here but

when you use an electrolyte and
that's what this green.

Patch is it's the electrolyte I can
actually move the gate off to the side and

the device will still work that's
a trick it's a trick that we can

do this because we're using
an electrolyte is the gate dielectric but

what's nice about it is that you can
produce these source training gate layers

in one shot on the same level.

That's not exactly how we're going to make
this device but that shows you the idea so

on the right here is the in
red is the embossed the poxy

pattern we're going to
emboss this pattern.

Into a poxy and then deliver metals and

semiconductor and dielectric and
another and another conductor.

So this is this is the side gated A.G.T.
structure made by this scale process again

here's the here's the profile of
the imprint this is an S.C.M. image of

this U.V. cured a poxy on a plastic
substrate like polyester.

This is the silicon master so
we have two reservoirs nine microns

below the surface of the silicon
another area that's minus six

minus six and then these are raised areas
of all of the above the silicon surface so

this is our master we make
a pedia mess copy we then stamp.

That pattern into this U.V. curable
poxy and this is the S.C.M. image so

you can make very nice features this way.

And now what we're going to do is start
delivering ink Here's a link delivered to

the source and drain reservoirs and
you can see the metal here so

this is optical micrograph.

Then these are fed cuts in an S.C.M.

image through those metal lines so you
can see we have nice silver coatings here

through this through this section
we then deliver semiconductor here.

So we now have semiconductor in
the Channel we can get force microscope

images of the semiconductor
it's reasonably smooth.

And then we can deliver conducting
polymer and what we've done here

is we've we've made a pinning line
right here so we deliver the ink and

it there's a trench cut here which
diverts the flow sideways so

that we get this side gate architecture
looks like this and profile.

There's a shot of it this is
the conducting polymer gate you can

kind of see it right here.

So then we flood with electrolyte
that's the last step.

So these devices work this is the source
string current versus gatefold the G.

again turning on the negative gate bolt
is fully on it of bolt and it's made

essentially by this self aligning process
why is this process self aligning itself

aligning because in the first step I made
all the high resolution features and

all the low resolution features and
all I had to do was deliver ink to these

reservoirs which are pretty easy targets
so if you imagine the roll process

those reservoirs could be lined up in
tracks tracks for semiconductor tracks for

metal tracks for dielectric and
you can have a bank of heads that deliver

the.

So of course yield matters and
reproducibility and

we're doing this all out in open
air no clean room environment so

we have about eighty percent
yield on the devices that work.

Eighty percent yield
period on offer a shows so

this is a log of the on current to the off
off current of about five good carrier

mobility point eight some of
your square provoked second and

nice small negative
threshold will to just.

We can also.

Just print metal.

Gate get electrodes these are E.G. tees
with a copper gate works just fine and

we can start to think about.

Making circuits this way so this is
an inverter where we have a transistor and

now a resistor all again made with this
imprint process and delivery of these

reservoirs and this is the output input
curve looks very nice a nice Triple H.

So again as I mentioned before showing
that you can make discrete devices is

a long way from making systems and
and we don't have a system made yet

this way but what we do have is that
at least a process more than just

on paper that allows you to deliver
in that high resolution and

get and get registration.

So let me wrap up by talking about
how you do this crossover issue so

we have to be able to connect
devices to each other and

the only way that you can really
do this is if you come up

out of the plane of the substrate
you go into the third dimension so

you have to have ways of running
two metal lines over each other and

having them be insulated so this is one
version of how we do a crossover we

imprint a structure that looks like this
this is a cross-section here right so

it's a multi-level structure we then
deliver ink to these reservoirs so

this connects to here
these are not connected.

We then deliver a poxy into these
reservoirs which then coats this

this under layer here we then splash in

more metallic ink here which connects
these two and that's the crossover.

So this is what it looks like.

And in fact.

You have you can get very large
resistances gomers distances between

those two lines so
at this point to show you how to.

Crossovers and transistors and
capacitors and resistors and

now we have diodes working I'm not
going to share that with you today but

those are all the building elements
that you need to be able to make and so

the next step for us is is two steps one
is to continue to improve the performance

of these devices but then the next step is
actually build systems and it's my view

that this area of flexible Atrox is not
really going to advance unless we actually

build systems discrete devices is is is
a start but we actually need to go to.

Integrated circuits so.

We have designed to roll the roll lines so

that we can implement this this
process for all the role and

the first one actually gets delivered
this month to our labs at Minnesota and

it's sort of a multipurpose
role the role line.

Will be able to take up to six inch wide
web and we have four tension controlled

zones because each printing process needs
slightly different attention control but

we have slot coding groove your
coding aerosol jet we will have

jet on here we have
a photonic centering unit and

this system allows you to go forward and

reverse so you know obviously.

In production you might have all these
operations strong out down the line you

have a fifty foot long line but we don't
have space for that this whole machine is

fourteen feet so what we can do is
we can print a layer center it and

then rewind print the next layer and
through it and rewind so that's.

This is our deposition machine the other
machine is an imprint nano imprint.

Machine where we will make we will cope
this curable poxy and then imprint and

then we will take that role off that
machine to this machine so this we have.

One line dedicated to nano imprint it's
worked out best to do it that way and

then this I think will will reconfigure.

And lots of different ways says
as has our needs require it.

So let me just end here and to be happy to
take questions I have tried to tell you

a little bit about my perspective on what
the challenges are imprinted electronics.

And and in particular I want to
emphasize that I think you know

in order to really do role
the role printed electronics

you're going to need a self aligning
process and I'll finish with.

Having you reflect on your smartphone
which is an amazing device and

I was going to show you mine but
it's over my jacket there but

your smartphone is it is an amazing device
it does lots of things but it is rigid and

there are many examples of things that
you could think about doing that be much

better if you had soft squishy stretchable
electronics a nice example is electronic

skin for robotics you'd like distributed
pressure and temperature sensors for

example I already gave you
the the idea of a more robust

flat panel display one that you
can bend and it won't shatter.

People are interested in distributed
sensing on the wings of aircraft so

you can detect the onset of turbulence by
slight pressure changes across the surface

of a wing that's a large area to
cover you're not likely to solve that

large area sensing problem by picking and
placing lots of silicon chips so I think

there are real good reasons to do large
area electronics and certainly when you do

it on plastic the idea of continuous roll
the roll processing is very attractive.

But we have some great chemical
materials engineering challenges to

solve before we get there thank you for
your attention to be happy to ask.

And.

The fundamental limit to
how small a device can be.

Right so the question is how small can
the devices be and what limits that So

what I like about imprint lithography.

People number of groups have shown that
you can imprint one hundred nanometer or

even smaller features writes one
hundred nanometers is actually

quote unquote easy to do
by nano imprint right so.

Right now for a lot of these applications
I don't think there is a need

to even go that small so
having sort of one thousand and

seventy design rules
you know where devices

are microns in size is just fine for
a lot of applications so

I actually don't think right now the issue
is so much about size it's more about

reproducibility How well do
the do the devices perform and

can you make them in a manufacturable
process those are the big issues I think.

With the ratio.

You need to do the rest
of the great question.

So actually you can make you can get
inkjet printers that deliver drops that

have ten micron diameters So that means
that I would probably want to target

of twenty maybe thirty Micron reservoir
size but if you if you do some layouts and

we've done some of this you can easily
imagine a thousand devices per square

centimeter even with those relatively
large reservoirs and a thousand there's a.

There's no process that we have now
that gets anywhere close to that so

a thousand devices per square centimeter
is orders of magnitude away from what

Intel is doing but
in this space it would be

great.

But that's a great question I mean I
guess one way you can think about three

dimensional circuit designs is simply
layering four oils on top of one another

and then you would have to have
some sort of interconnect strategy.

But but certainly there's nothing about
that that seems completely undoable to me

so I guess the first crack at going three
D. from my mind is to layer foils on

top of one another.

It's all.

Right I mean rightly Yeah I think
in the system the the lines that

we're designing are going to be enclosed
in hepa filter so they're not going to be

in a clean room they're going to be in
a room like this but there have their own.

They'll be self-contained inside.

Inside a box basically.

But you know certainly being in
a clean room would would help solve.

Some of our ranks are equally spaced
which is nice but not all of them so

many of the polymer semiconductors
you have a coronated solvent

in that case though what's nice
is you're not slot coating a six

inch wide web with with ink you're just
you're delivering from an inkjet head so

the amount of material that you're
delivering is actually quite small but

still you'll have to you'll have
to remove corn and solve them.

And that.

Led to.

OK So the question goes back to
the ion gel electrolytes and

the molecular weight influence
on the on the gel properties in

particular the molecular
weight distribution.

I don't know the answer but I would kind
of expect it not to be very sensitive and

the reason is that these
are swollen networks so

that it's not like this is a close packed
you know dye block microstructure.

Where.

Some of the molecular weight distributions
can help relieve interface full tension

this is a swollen network where
the styrene cores are relatively

far away from each other and
I don't think you're going to see much

measurable impact the molecular
weight distribution.

And.

Me.

We're not know is that you know.

That.

Yeah really good good question so when you
can you can make the gel by dissolving

the ionic liquid and the block polymer
in a CO solvent like Ethel acetate and

that's the ink so then we print
that the F L acetate evaporates and

the gel spontaneously forms.

It's a physical cross-linking not
a chemical cross-linking and in principle

it's thermally reversible we tend
to make gels with large enough and

block molecular weights that they're
stable up to one hundred C..

So.

OK So the question is what's the
performance of these soft circuits like

compared to silicon circuits our circuits
are terrible compared to silicon service.

You know in my lab clocking devices
that are killer it's is great.

In your laptop your clock at
A couple of guards right.

So there's no comparison in
terms of raw performance so

there's a long way to
go in that aspect but

I think there's reason to be hopeful I
mean there are there are some very nice

improvements in printed devices
in even the last five years and

I think that the inks
continue to get better.

And so I think I think there's
reason to be optimistic but

the other thing to really keep in mind and
I said this a couple times but

I'll say it again is anywhere
in flux more trucks you can use

silicon by say gluing a silicon
chip on your end of the circuit on

to the piece of plastic that's what you're
going to do anywhere that makes sense

economically that's what you're going to
do because you have so much power and so

we can ship what we're trying to do is
develop a technology that would apply

to situations that are large area
where you would never get away with

picking and placing all that many silicon
chips and the best to every day example

is the backplane to your laptop or
the backplane to your flat screen T.V.

We use amorphous silicon on glass there
because it doesn't make sense to pick and

place all silicon chips so a large area
applications where you need flexibility

stretch ability that's that's the kind of
application and you might be able to then

get away with less performance.

Thank you.

In the case of these rules real we've got
these reservoirs is there are the three

of them we could be standing.

So I guess it will.

Come out you.

No, that can't happen.

That's that's an interesting question.

I'm not aware that we've seen it.

Certainly when you think about a moving
web in the vibrations on the web

you would maybe wonder about that.

So.

And so and so sizing sizing will be
important how much do we deliver

right how much ink do we deliver to the
reservoir how big how big is the reservoir

it's an interesting question.