Well I mean one water could
use a stop over here.

Well you know.

What to get through our speakers today so
we're pleased to have Dr that is a.

Georgia Tech the love
doctors here yesterday.

With just heard worked in department
a real science there is a thirty.

Started here ten fold last fall
straight from fifty five that.

Industry.

Is here today to talk to
us as they do a little.

Engineering.

And he's here to talk to us today
about how we do performance.

Materials.

All right.

Thank you it's a pleasure to be here I
appreciate the invitation to speak at

the nano tech seminar everybody here
me pretty well Mike's on all right so

today I'll be talking about some of
my previous work related to dynamic

process is an electric chemical systems
and understanding these processes and

engineering materials for
improved performance.

And at the end I'll briefly mention some
of the things that my current group here

at Georgia Tech is working on
now since I started in the fall.

So in general kind of an introduction
to electro chemical systems.

Electro chemical processes and devices are
very important for many types of energy

conversion and storage so you guys have
probably seen the Tesla power wall right

this is kind of a bed for
personalized home energy storage to get

off the grid there's fuel cell vehicles
there's electric vehicles are run on

batteries as well this is a toilet a fuel
cell vehicle they've been pushing the.

For a long time there's also devices
that can make electric chemical fuels so

basically splitting water is one way
to make a fuel you do this electric

chemically to produce hydrogen right and
so all these different.

Applications are enabled by different
types of litter chemical devices for

the quite important for for
energy storage and conversion.

So if we take a look at what
a letter chemical device generically

looks like this is kind of a schematic
here so this is maybe a battery or

a fuel cell you have two electrodes and
anode in the cathode

you have some sort of electrolyte in
between electrolyte passes ions and

not electrons you then have some external
circuit where the electrons pass right and

so during operation of this electro
chemical system you have ion motion in

the electrolyte an electron motion
through an external circuit so

on the device scale these are the types of
dynamic process we're concerned about but

there's different lengths scales that are
important in these devices so if we zoom

in a little bit look at the surface of one
of these electrodes we may have a bunch of

particles that are stacked together in
a certain way that are important for

device operation and this is this is
we've zoomed into the Misa scale now so

we're on a smaller scale bar but
on this scale

there's other things that are important as
particle particle interactions that become

important they have to be controlled for
good performance there's also electron

an ion transport in this architecture now
between particles and among particles

if we take another level of zoom go to the
nano scale now there's different aspects

of device performance that we have
to care about so in a battery for

instance we may have phase transformations
that occur in particles This can cause

mechanical degradation fracture lots of
different dynamic things that happen.

We could also have a material so
this is process such as an electro

deposition right when you're growing a new
crystal on a surface from the liquid

solution this is another dynamic process
finally you can have corrosion and

dissolution of a material in the nano
scale which usually you want to avoid.

In order to have good and durable
performance so the summary of the slide is

that across all these like scales we have
these dynamic things that are happening in

system and these electrochemical systems
and really to engineer better systems we

have to understand these dynamic processes
and control these dynamic processes.

So I'll talk about two different research
areas today these are things that I

focused on in the last few years as
a student in a post-doc and continuing

to focus on in my faculty position
the first is high capacity materials for

lithium ion batteries and the second is
developing stable photo electrodes for

electrochemical fuel production these are
simply conductors that absorb light and

produce fuels like hydrogen.

And the overall overarching goal is to
understand it control the dynamic behavior

and interfaces in these systems and

hopefully the last slide was a good
introduction about why that's important.

All right so we'll start with batteries so

batteries of course everybody knows
lithium ion batteries have enabled.

Portable electronics they will enable
the revolution of portable electronic in

the last fifteen twenty years with the I
am batteries have a very high energy Misty

compared to other types of battery
systems like nickel cadmium or nickel.

But even though they've been very good for
phones and laptops moving to

new applications like low costs and
long lasting electric vehicles which maybe

not is this one woods a nice picture
we need different types of batteries so

we need batteries that are lighter
smaller and cheaper and so

this is a big research area of course I'm
sure many people are familiar with it.

But if we take a look at what
a battery looks like and

try to understand how we can make them
smaller lighter cheaper basically that

presents itself pretty quickly so
this is a schematic of a battery

we have two different materials that
are used as the anode and cathode

then there's an electrolyte in between we
talked about this before with the ions

going through the electrolyte but the
basic the energy limits on this battery

are given by this simple formula down
here where the lithium storage capacity.

Times the voltage of the total system
is your is your energy content and so

to increase the total energy content
with the same size system we need

to either increase the lithium
storage capacity in your materials or

we need to increase the voltage between
the materials and so the materials I'll

talk about today are materials with
significantly higher lithium storage

capacity which could lead to higher
energy density in a full battery system.

All right so one of the materials and one
of the primary materials that has received

research attention is silicon from lithium
ion batteries so silicon is a replacement

for graphite as the end of material
graphite holds lithium and the capacity of

graphite is this number here three hundred
seventy two million power's program that's

basically like homes William Powers is
another way to say that silicon You

can see is almost an order of magnitude or
more than an order of magnitude.

Higher right and so very high capacity so
we know it has I theoretical capacity so

why don't we use it the problem is that
there's a different reaction mechanism

when lithium react with silicon compared
to when lithium reacts with carbon or

graphite So that's shown here
schematically So lithium actually alloys

with silicon so you take fifteen lithiums
atoms and stick them with four silicon so

you form a new phase a new ally
phase lithium fifteen silicon for

very let the enrich has a very large
volume compared to the initial silicon

You can see here if you start with silicon
particles they react with lithium to form

larger particles but they break up into
smaller pieces and fracture is mechanical

degradation that occurs and this is bad
for your battery performance basically

your particles some lose electrical
contact in your battery electrode and

they become dead weight basically and
you lose capacity over time so

this is in contrast to graphite
where we're lithium can actually

be inserted directly between the graphene
layers in the graphite and this is a very

versatile process very small volume
change only about five percent and so

you can insert lithium and remove it
over hundreds and hundreds of cycles and

that's why your batteries last
a long time and your phones.

So these different reaction mechanisms are
fundamentally the reason why even though

we know this has a high capacity we we
have not been able to use it successfully

because of the capacity to K.
with cycling so.

The goal here basically
in the research arena

on these materials is we know that
there's this large capacity so

maybe we can try to understand exactly
how the reaction mechanism occurs and

if we can really understand the volume
changes the reaction mechanism and

also the mechanical degradation the
fracture that occurs in these particles

maybe we can engineer them specifically to
accommodate this expansion reversible e

expansion contraction if we could do that
maybe we could make a good long lasting

battery with these materials but
understanding these reaction mechanisms

are kind of difficult because conventional
electro chemical experiments don't really

give you any insight into these mechanisms
so that's that's what I focused on in

the past understanding these reaction
mechanisms with new techniques.

All right so the main technique or
one of the main techniques that I've used

is called in situ transmission
electron microscopy or T.M. so T.M.

is a is a very high resolution
electron microscopy technique

where you can image materials
at the atomic scale.

But in situ means that we're actually
doing some sort of reaction inside

the microscope we can observe
it while it happens so

to do that we use a special sample holder.

So the T.M.C.
or Georgia Tech are downstairs and

they're pretty large the sample Holder
is pretty small this is only about

a centimeter wide here this image and but
you can see here this is zoomed in image

of this an optical image
there's two probes one here and

one here so on the sample holder we have
these two probes we can actually put

battery materials on these probes so
for instance we put lithium Coble oxide

which is a common lithium containing
material in one probe and

we put silicon or silicon nano wire and
animate serial on the other probe and

then inside the microscope we can actually
move these probes together actually with

nano scale precision almost an animator
level precision and we can contact them

there's an electrolyte in between so this
is a liquid electrolyte it's an ionic.

Wood has a very low pressure so in T.M.
It actually doesn't evaporate You can also

use solid electrolytes which I've done and
it's very useful but the electrolyte is

key because when you bring these two
things together you've now contacted them

but there's no legible connection only an
ionic connection so you apply a bias and

you can cause lithium ion to flow from
this electrode through the electrolyte and

to react with the silicon nano wire and
you can image the reaction and

real time to it's just it's very
similar to a real battery reaction but

you basically build a nano scale
battery cell that you're operating.

All right so here's a video of
this process a little hard to see

here because it's dark but
hopefully you get the picture.

So this is a silicon nanowires
sorry this isn't a particle

this nanoparticle is attached to
a nano wire kind of down here and

you'll see this process lithium is going
to be reacting with this particle and

you'll see it expand and
something interesting happens at the end.

So this is fifteen times actual speed.

And you'll see this darker core region
see the darker region the shrinking

that's the silicon Crystal and
silicon core that shrinking

as a lithium reaction to this lighter
region around the edge is actually

an amorphous litigated silicon very highly
lithium it has a larger specific volume so

that's why you see the volume expansion
what happens is one at some point

towards the end of the reaction is you
see this this crack initiate grows across

the particle the particle breaks into two
pieces so it's pretty cool this is a this

is kind of a real time demonstration of
what might have to be happening in a real

battery electrode that could be leading
to capacity decay with cycling but

what you saw there was this process where
you have this highly with the added

silicon phase it's a morph a sexually
we know that from the imaging and

there's this sharp reaction front between
this phase and the crystalline silicon and

the reaction front moves in concentric
lead towards the middle of the particle

and eventually we see the fracture occur
and the particle break into two so

some questions arise from
these types of videos So

first of all why does fracture
occur at the surface.

We saw the crack initiated the surface
we need to understand stress evolution

to understand how that happens and then
second of all with these types of images

and videos we can directly correlate the
reaction kinetics to the driving force and

things like that so we want to
understand the reaction kinetics.

So first I'll just briefly mention
these this stress evolution and

why fracture occurs at the surface so
everyone may not be familiar with

mechanical stress analysis this is
actually a calculated plot that shows

the hoop stress mechanical stress inside
this particle in the midst of litigation

so imagine the particle looks something
like this where there's a silicon core and

the lithium silicon shells were
about half way with the added here

if we look at the stress across this
particle starting at the middle

going to the surface we could see that
it's fairly large First of all so

we have Pascal level stresses this is
much larger than you'd normally find

in a chemical reaction and you can see in
the core we have compressive stress so

we have pressure in the core to move out
to the surface we have this big drop here

which we'll talk about later but then
we actually move towards a service and

we get tension so
there's tension at the surface so

we have these large stresses that we have
tension of the surface and this tension of

the surface that arises actually causes
the fracture to initiate the surface.

So an interesting thing that I observed
along with one of my colleagues So clearly

this is at Stanford when I do this work is
that there's a strong size dependence to

fractures of particles and pillars
nanostructures of different sizes do or

do not fracture depending on their size so
that here's just an example plot showing

this so this shows this is basically we
imaged a bunch of these particles and

pillars and determine whether they
fractured after that the nation

you can see the particles above about
three hundred sixty nine meters in

diameter fractured ninety five percent of
the time or more below that it drops off

significantly almost zero percent so
there's this critical size for fracture

it's about three hundred meters so below
that you can let the these particles you.

You have this large volume expansion and
there's no fracture that occurs

this is a really important number for for
manufacture real battery electrodes so

nanostructure smaller than this three
hundred meter value should be used in real

batteries so the reason why there is this
strong size dependence has to do with

this strain energy that's inside these
particles so the strain if you're a member

of some of you may have taken a fracture
mechanics classes or something like that.

The strain energy is what drives
the formation of a new crack right so

the strategy is just the energy this
contained within the whole structure due

to the strain that's there
there's lots of strain there but

with small particles there's just
not enough material to allow for

strain the strain or G. to build up
to the point to drive the fracture so

in larger particles This is the case so
we have this strain energy effect to be

that that is seen when we use
particles of different sizes.

All right so
on to the kinetics measurement so

this is another interesting thing and
you'll find out that after we can analyze

the kinetics actually the stress evolution
of these large stresses that exist in

these particles actually influence
the kinetic strongly as well.

But this is an individual particle here
we can see and it's going to become with

the added I'm going to show some images
here and what you'll see is we can map out

as a function of time how big the particle
is and how big the crystal in core is so

we started about eighteen and meters in
diameter down here in this particle or

this image here from this side.

After a little bit of time the particle
has begun to be lifted and

you can see this little darker
region that's the crystal in core so

we can map the outer diameter and
the crystal and core as a function of time

after more let the ation you can
see that this initial fast with

the Asian trajectory slows down quite
a bit so we have fast of the ation and

then the particle kind of vilification
process stalls over here and so

you can see that in a single
particle here as we let the eight so

the reaction slows down so
initially this wasn't too surprising but

upon further analysis turns out
that it was very surprising so.

In terms of the detail kinetics I asked
the question what kind of kinetic

should be expect so we have these
particles the reactor with lithium

it's some inner facial type reaction we
think because we had this this reaction

front that's moving the sharp reaction
front so there's basically two different.

Very simple types of of limiting cases
that we could expect in terms of

the kinetics of this process the first is
if the if the reaction at the interface or

at the silicon lithium silicon
interface was the rate limiting step

this is the really an extent we have short
range interactions that are controlling

the kinetics and the interface itself
should move proportionally with time so

it should be a linear
a linear shrinking with time.

Actually a little bit faster in
this because we have a sphere but

it's very close to linear.

For the fusion limited kinetics which
is the other possible limiting case or

one of the possible any cases we expect
that the the the lithium to thickness

should grow with the square root of time
and this is commonly known square root

time kinetics for diffusion limited cases
but in fact we have neither of these

the data that we observe from lots of
different particles slows greater and

this the slowing is more dramatic than
either one of these is more dramatic than

linear of course because we have
this non-linear effect but it's also

a greater slowing than we'd expect from
just a squirt of time based kinetics.

And I measured many many particles
sixty to seventy of these particles and

they all showed the slowing effect so
so what's happening here so what

we think is happening sorry is we have
this strong effect from the hydrostatic

stress that evolves in these particles so
the hydrostatic stress is the pressure so

as we lift it this particles are already
told you that there are these high

stresses that evolve near this reaction
front there is actually a very high

pressure that evolves and it changes as a
function of the extent of let the ation So

this is just this is
a little bit complicated but

this is the hydrostatic stress or
pressure inside this part of.

All right at this reaction front as we're
lifting idiots over as we lift the eight

we move this way and the stress gets
more and more negative so as higher and

higher pressures and these pressures
actually influence the driving force for

the reaction so this is the good free
energy of a generic reaction to get free

energy of course in the letter chemical
system has a chemical component and

an electrical component and this is how we
usually write it in electrochemical system

we usually don't write the mechanical
component to give free energy but

is there and the reason we don't usually
write it is because it's usually extremely

small compared to the chemical driving
force of the electrical driving force but

in the case of this extreme volume change
reaction which results in these very large

get past level pressures actually
the mechanical driving force for this for

a component of this gives free
energy becomes the same or

close to the chemical driving force and

so what we think is happening is basically
the particle starts let the aging and

eventually the pressure reaches the point
inside the particle that it just

stops the process from occurring because
it cancels out the chemical driving force

of the reaction it was a very
interesting process and

it wouldn't happen unless we had
these extreme volume changes and

these particles so I should mention
there's a teach you here at Georgia Tech.

Very soon after I publish this
initial paper reported similar

observations in nano wires silicon
nanowires underlip the ation and they

undergo a very similar let the process and
similar stress evolution to particles so

it seems like there's good agreement
across different types of geometries.

And this is this is
an important observation.

Mainly because you saw how this we
had this reaction slowing process so

in real batteries this this this
fundamental process this fundamental

limitation of the kinetics could
influence the actual maximum charging

rate that you could have in a battery
system using these materials so

that's kind of the link
towards a real battery system.

OK So so
far I've talked about Crystal and so.

And so we have this nice crystal these
are all single crystal particles and

they react with this sharp reaction
front but what about a more facility and

so so many people may know that amorphous
silicon is a relatively common material

silicon is is a little bit unique in that
way in that you can easily prepare it

in both the amorphous Crystal in states.

Receive for the crystal and say phase of
silicon there's a two phase interface

during the reaction fracture occurs when
the particles are about three hundred

meters and we also see this slowing in
the reaction kinetics so what about for

a more facility in an amorphous silicon
Of course we have a disordered.

Matrix of atoms there's
no atomic lattice But

how does this influence
the reaction process.

So I have a video here this is
going to be really hard to see but

we will see a little bit of expansion but
the interior will be hard to see

this is a this is a morph a silicon
sphere it's about four or

five hundred meters in diameter
it's attached to a silicon rod here

this is only this is acting as the lithium
highway so this really has no influence on

what we're observing but the lithium
travels through this other Silicon rod

to react with the silicon sphere so we're
concerned with the silicon sphere here and

again this is fifteen times actual speed.

So if you look closely there's this
darker region that shrinks again

it's kind of hard to tell here but

there's this darker region you can
see the expansion maybe a side.

Not that important to see I have an image
here you see a little bit more clearly

here this is some spheres in the midst
of the ation you can see there's this

sharp front here between this darker
region and this lighter region so

there's less of a contrast here
because materials inside the T.M.

don't scatter as strongly
as crystal materials but

we can still say that there's a pretty
sharp reaction front between this.

Silicon region and the amorphous pristine
silicon region and in fact there's been

some other work that shown this is
actually anatomically flat reaction from.

So this is.

The result of this is is basically this
reactor looks very similar to the crystal

and case and one of the reasons why I
bring this up is this was very surprising

because from previous electro chemical
experiments actually the reaction front

was not expected to be atomically flat or
even flat at all this supposed to be very

diffuse but we've observed this this very
sharp reaction from these experiments.

So the one thing I want to say about
amorphous silicon beyond that is

I want to contrast the fracture behavior
to the crystalline silicon case so

before we saw actually that crystalline
silicon particles fracture and

there's a critical factor size actually
that the particle fracture size

is about two hundred nanometers.

The pillar fracture size or nano pillar
fracture size critical fracture size

is about three hundred meters so
you can see here is an image of a particle

fracturing upon with the Asian So how
does this compare to more facility and so

I tested I tested amorphous silicon
particles up to eight hundred meters

in diameter and none of them fractured and
actually since this work there's

been additional experiments showing up
to two or three microns in diameter for

amorphous structures there's no
fracture upon that the Asian so

amorphous silicon appears to be much
more resilient against fracture than

crystalline silicon and so it appears that
using amorphous silicon in batteries is

probably a good idea number
one Fundamentally though.

The reason why this is because.

A more of a silicon expands isotropically
During with the ation So this

is something I don't really talk about but
crystalline silicon undergoes an isotropic

expansion you can kind of see that here so
this is the crystalline silicon core and

this is a half way with the added particle
you can see this core is highly faceted

even though the original particle was
a sphere so this is this facet of core

results because lithium actually reacts
with different crystal plains of silicon

at different rates and because of that you
have expansion in different directions at

different rates and this causes actually
stress concentrations at certain places

and that's one of the main reasons that
fracture occurs in these crucial in part.

So that was a very brief explanation
of this there's been a lot of work on

understand the an isotropic expansion
of silicon but it's rather obvious now

that a more facility and has no underlying
lattice rides just an amorphous structure

there's isotropic expansion because of
that there's no stress concentrations and

so it's more robust and
will not fracture upon the.

To very promising these experiments have
shown there is a very promising material.

All right so that's what I'm going
to talk about in terms of silicon

understanding silicon
reaction with lithium

there's one more slide here this is just
a kind of highlight some of the parallel.

Efforts that have gone into engineering
in silicon and structures and

testing them in and battery systems and
trying to improve their performance so

there's been a lot of work
developing engineer nanostructure So

this is actually work from
group at Georgia Tech

this is some work from a previous group
at Stanford for some citations but

basically the these studies are dedicated
to engineering these nanostructures and

testing them in batteries and
trying to improve their cycle life so

here's an example of capacity over
a number of cycles you can see that this

this material here this engineer now
structure is actually a hollow carbon

material with silicon struck nano
particles inside results and

very high capacity over a thousand cycles
so this is much much better than need

than you would get ten years ago
where you have maybe a capacity to

cycle life of about ten cycles so
really improve cycle life and the main

reason is because of these engineer Nash
treachers they're specifically designed to

accommodate this volume expansion and
contraction and I you know I like

to think that a lot of this work across
the world really has it's gone hand in

hand with some of the fundamental
understanding that has also been pretty

widespread some of these studies it just
showed in terms of understanding these

reaction mechanisms and I think that the
fundamental understanding has been very

important for for showing you know
the pathways forward in terms of

engineering better in structures
that can accommodate these changes.

So that's been a really nice area of
research I think the last seven or

eight years even.

All right so a little bit more
about these battery reactions so

I've been talking about silicon primarily
or only until now but lithium actually

reacts with many many materials and
there's lots of different options for

materials and battery systems so this
is just an example of another material

that actually undergoes a different type
of reaction with lithium So this is a T.M.

image of copper sulfide nano particles so

this is a fraction pattern these
nanoparticles are actually disks and

they're excitedly shaped so we're looking
down on the face of the disk here

well this is actually more like a circle
but they're supposed to be exactly shapes.

And it has exceptional symmetry.

But these copper sulfide materials undergo
what's called a conversion reaction so

instead of an alloy reaction which
we saw before where lithium simply

forms a new phase a conversion reaction
is a little bit more complicated so

there's the chemical reaction or the
electric reaction I should say of lithium

with Congress will fight so you start
with copper sulfide you have lithium ion

electrons to reduce the material and you
end up with three different phases you end

up with lithium sulfide which is a lithium
reactor with the sulfur and then you and

you also have copper metal So this is more
complicated and the question is you know

we have these particles we know what
should happen electric chemically but

we don't know what how the morphology
should change we don't know how the volume

should change because this final mixture
of two phases could take any number of

forms so studying this type of
reaction is also quite important.

So what this looks like
actually is the this is

these are two different particles and
they're sitting on a carbon

carbon support here looking along
the edge of one of these particles so

you can see actually the lattice fringes
here very high resolution image.

This is the for you transform this image
to these spots here correspond to these

lattice planes and these lads planes
correspond of these spots here so

we can determine that this is.

And has that correct a lot of spacing and
all that So this is before reaction so

we can in this process we can image these
particles right before reaction and

right after reaction so
after reaction they look like this so

before we have these
nice lattice French's And

right after it's changed quite a bit
we still have this particle down

here you can see we still have these
lattice French's see before and after but

now we have this little hat right this
guy's wearing a hat now and what's this so

this is actually the copper this
darker region is the copper metal

this lighter region down here is
actually lithium sulfide So this is kind

of subtly very interesting before we had
this copper sulfide particle after we've

ejected the copper metal to form a new
copper particle and then a particle and

now we have a lithium sulfite particle
that is in the exact same location

with almost the exact same lot of spacing
as the initial copper sulfide So we have

we retain the morphology even though we
undergo this rather dramatic transfer.

So that was interesting when we
discovered this so why is this happen so

if we look at the atomic level
structures of these two materials copper

sulfide lithium sulfide we can get
a good idea of why this this retained

morphology occurs so here's copper
sulfide high resolution image and

this is a Fast Forward transform same
thing for lithium sulfide you can't

really tell much for these images they
look pretty good though I must say but

let's look at a schematic to help us
understand what's going on here so this is

the you take this image actually you can
basically map it out on the schematic so

this shows the sulfur atoms and
the atomic structure and the yellow and

then the copper atoms which
which are dispersed rather and

are rather complex manner and this copper
sulphate material they put in part here is

if you look at Lithium sulfide the sulfur
atoms are in almost exactly the same space

space lattice lattice points.

The spacing between the sulphur layers is
almost the same only about two percent

different and so basically
the lattice of these two materials is

almost exactly the same so
what happens is when we insert.

If you mean the copper sulfide the lithium
comes in the copper can actually diffuse

very quickly this is a material known for
that and the copper actually diffuses

out of the material and forms a new
copper metal particle on the surface and

lithium basically just
replaces it in a lattice and

so this is kind of like a cat
ion exchange process almost

where this usually happens and a liquid
environment here is in solid state.

And so we have this interesting and
we basically retain

the particle morphology
during this exchange process.

And just this is just very brief here but

we can also image this process while it
happens in slightly larger particles so

this is a copper sulfide
particle here before the reaction

this is looking at the same particle
lithium sulfide after the reaction but

we can color code the two different
phases as the reactions occur and

so in red we have copper sulfide to start
with as we start the reaction we can see

that some blue phase occur forms at the
left side that's the lithium sulfide phase

and as we go across the lithium sulfide
phase grows across the material and

eventually converts completely to lithium
sulfide So this is this is different

than the images before this is actually a
high resolution atomic level image showing

a reaction this occurring across
the single particles so this is

this is actually kind of difficult to get
this data so I'm pretty proud of that.

All right so one more one more
slide here this is pretty cool so

I showed you what happens when a single
copper sulfide particle reaction lithium

we had this transformation in the lattice
the copper gets kicked out to form

this new particle but something really
interesting is what happens when you

have a bunch of these five particles
all touching each other so

that's what's happening here there's
probably one hundred or two hundred

copper sulfide particles in this in this
region here and it's kind of thick That's

why you see it's dark you can't
really see through it with the T E M.

But what's happening is they're all
reacting with lithium and the copper

I mentioned before is a very fast diffuser
in this system just like lithium actually.

Well what's happening is it's more
energetically favorable for the copper to

diffuse together from all these particles
and diffuse through the particles and

then form a copper metal denge right that
grows that's more energetically favorable

than simply forming a bunch of little
copper particles so what we see actually

is this large copper didn't write that
grows during lift the Asian process so

this is actually you know this is
obviously dynamic we're seeing this

this copper metal then dry to grow
out of this solid material so

this is really weird to see the first time
I saw it I mean unexpected definitely but

it's only enabled by this
interesting structural change and

also the very high diffuse cities
of copper in this material.

So it's and it'll be interesting
to compare this to the reaction of

of copper sulphate with other
alkali metals like sodium or

potassium to see if the same the same
kinetics and the same morphology changes

hold.

All right OK so
in the last fifteen minutes or

so I'm going to switch gears a little bit
no longer talking about batteries but

talking about a different letter chemical
system and this is materials for

solar fuels.

So.

So solar fuels fundamentally means
we're going to absorb sunlight and

create storable fuels either hydrogen or

hydrocarbons one way to do this is
actually to use semiconductors which we

know can absorb sunlight like in a solar
cell but instead of using them in a solar

cell use the semiconductors to do electric
chemistry so basically used to make

it actors that absorb light and
split water so the idea here

is that theoretically Well theoretically
you could make a simple and cheap

solar fuels device using semiconductors
that are doing electric chemistry.

To split water this is the simple reaction
of course you all know this you need

gives free energy of one point two three
electron volts to split water so this is

rather high especially for a semiconductor
system I mean most most solar cells.

Single solar cells will produce between
five to seven hundred millivolts So

here we have one point two
three volts that's required so

this this complicates
the situation a little bit.

And I'll just mention hydrogen
is useful fuel because it can

be used regularly in fuel cells
as many of you know of course but

it can also be used to further
process carbon monoxide and to and

to store will fuels liquid fuels and
so this could be a nice feedstock for

renewable fuels that's why
hydrogen is of interest.

But this is a schematic of a solar fuels
device so this is an original design

that was invented at Cal Tech
probably about a decade ago now but

this basically contains two different
semiconductor materials and you

take this whole device and you actually
dunk it in water that's how it operates.

So the signatures are labeled photo
anode photo cathode the photo and

it is a larger bandgap semiconductors so
it's a it can absorb U.V.

light or near U.V. light the photo cathode
is a smaller band gets I mean Dr you

can have this membrane separator that's
quite important actually in between so

it separates the products and
it also allows for ion transport.

But the important reactions that
are occurring as you're absorbing

you're absorbing light in
the photo anode semiconductor and

so you have photo generated holes so
you have

excited holes charge carriers that move
to the surface of the semiconductor and

they react with water actually the holes
react with water to oxidize water and

they form oxygen so you have the oxygen
evolution reaction happening here and

the bottom semiconductor you have
the hydrogen evolution reaction so

the photo generated electrons
move to the surface and

they react with water to form hydrogen and
so this is a totally contained for

water splitting process where on the top
electrode you're forming water or

sorry you're forming oxygen you're of all
the oxygen in the bottom bottom electrode

you're of all the hydrogen and so there's
a lot of details here will go into all

the details but that's the important
things that you need to know.

Regarding the whole device.

The the really big problem
with these devices and

with this whole field actually is that
basically all samee actors are or

many many similar actors are unstable and
eight weeks environments and

this is actually exacerbated when you when
you shine a light on it because often

the photo generated carriers like the
holes are have a very high energy and they

can actually corrode your material instead
of doing the electrical history you want

them to do and so photo anodes where that
where the water oxidation happens they

operate at the high potential and they are
especially susceptible to photo corrosion

because these put it holes can corrode
the material itself so this is an example

of what photographing looks like this
is a photo current stability test so

this is this is a semiconductor material
it's dunked and it's actually a slightly

basic or pretty basic solution and
light is shown on and we're measuring

the current that's evolved during water
oxidation as a function of time and

you can see first we start with about one
point four million some similar squared

and over the course of fifteen minutes
we drop off to a very low value and

so our performance is decreasing This is
because the surface of the material is

corroding away and we can verify
that by looking at before and

after images of this thin film this is
a businessman a date actually it's a type

of oxide semiconductor this is before and
this is after we can see actually that

we've lost the material here so this is
an example of the photo corrosion process.

So the you know the first thought would be
OK So we have this we have these doctors

who want to protect them underwater we
want them to still you know retain their

functionality how do we do that well the
first thought would be I guess to try to

put something else on top you're seeing a
doctor right to protect it from corrosion

that's rather an obvious thought actually
I mean people been thinking about that for

a long time maybe forty years but
the problem is it's

actually kind of difficult because
there's a lot of requirements for

these protection layers so they have to
be resistant to corrosion first of all

they have to be transparent
they have to let light through.

They also have to be conductive
defo to generate a carry.

And this can be a hard one because many
many materials that are resistant to

corrosion and transparent are are wideband
oxides that are not conductive.

They also have to be catalytic for
the oxygen evolution reaction or

the hydrogen evolution reaction depending
on which one you're doing and so

these are a lot of different fireman's And
so there are some tradeoffs rights of

the insulating oxides can be corrosion
resistance and transparent but

they're generally not conductive or
electric catalytic So that's a problem.

So basically we need to find the right
material and the right fabrication process

this is actually something that I
started working on as a postdoc

in conjunction with a number of other
colleagues this is a Cal Tech and

the Joint Center for
artificial photosynthesis there.

And so we basically developed a number of
different ways to deposit very thin films

and thicker films designed to protect
the semiconductors from photo corrosion

this is one example so this is using ultra
thin ailed the protection layers this is

a schematic so this is the green here is
the semiconductor business fan today and

I developed this dual ultra
thin coating dual layer

coating of titanium oxide which is
deposited with atomic layer deposition and

also Necco which is catalytic So
we kind of.

Taking apart the two functionalities.

This was pretty promising this this
coating allowed for an imp and increase in

stability time from about you
know ten minutes to two hours.

But the electronic conduction process
through these very thin layers

is most likely a tunneling process so
that's why we're able to conduct electrons

to oxidize the water.

But the problem with these very thin
materials is then coatings that is there's

always defects in pin holes and it's very
difficult to get sustain stability over

ten twenty one hundred thousand hours
with these very thin coatings so

moving towards thicker coatings
that are very carefully processed.

Is necessary.

And so just briefly here.

There are two different types of coatings
that I've worked on and in conjunction

with some other colleagues as well
to different oxides both dioxide and

nickel oxide that have shown very
good performance and as as protective

layers for photo electrodes for these
water splitting devices basically have

improved the stability of these electrodes
now to unprecedented levels and so

this is been really an important.

Breakthrough basically in this field as
not just me this done it's been probably

ten to fifteen grad students and postdocs
at Caltech in the working on this.

So first talk about titanium dioxide.

So basically what we're going to use
silicon now is are semiconductor in type

silicon this is a very easy material to
work with it's probably many of you know

and we deposit one hundred meters
of amorphous such a name dioxide

on the surface with atomic Lare deposition
and then a very thin nickel layer as

the catalyst layer this is kind of similar
way just shows that this is much thicker

than in dioxide and so we don't expect
that tunneling will have any kind of.

Effect here right we have to have
some sort of conduct for that to for

the photo generate characters
other than tunneling

what we see is when we test these devices
so this is a letter chemical test of this

electrode that's built up with
this multilayer stack here.

The details here on important but in the
dark you see this is a standard catalytic

electrode you see a catalytic turn
on a current turn on at some point.

This is the silicon electrode over here
and we can see that we do have photo

current that flows if anybody is
familiar with photovoltaic curves I.V.

curves we analyze these and
basically the same way these foetal intros

will get in the details there though
what this plot shows is we do have

a catalytic current that's due to
the flowing we have this flowing and this

is surprising initially because we expect
we actually expected that this amorphous.

I would completely.

Be a blocking layer to the holes that
are trying to flow through it but

in fact we do get flow of current and we
do a pretty good performance actually so

another aspect of this is that
the stability is very long so

this is the stability of these photo
electrodes that are oxidizing water and

ph fourteen K. O. H. and

you can see here that we're showing the
photo career over one hundred hours and

it is fairly stable these have been
further tested over now to many hundreds

of hours without any kind of degradation
so as promising as a protective layer and

it still allows for this current to
flow which was initially quite puzzling.

So to kind of just briefly mention some
of the properties of this material so

this is a T.M. image showing the amorphous
T O two layer on top of silicon

within a thin nickel layer on top which
again is the the catalytic layer.

And this is a high resolution image
of the interface here the silicon

interface there's always some
silicon oxides interface

which we actually think plays a role and
developing the photo voltage but

the interesting thing with this
material is that we're actually well.

I was going to want to show this
because it's very complicated this

is actually the band diagram
this been mapped out using X.

P.S. by my colleague here Hsu who.

Basically has set of experiments allowed
us to really understand how current was

flowing through these materials by mapping
out the band diagram but I'm going to show

a little simplified version right because
the other one was way too complicated so

here we're looking at the end type silicon
this is the material that's actually

absorbing the light and the photo
generated holes are being generated there.

To the studio to layer sits right
across the interface from silicon and

the point is we want to get
holes from the valence band

of silicon through the T O two to
the nickel so they can oxidize water so

the question is how is this happening
this question is not completely resolved

unfortunately but
there's a few options the first option is.

Basically the holes can
recombine with electrons and

the conduction man that are coming
from the water oxidation process.

That's one option the second option is
actually that the holes move through these

mid gap states in the two This is a this
is a controversial option I would say and

there are ongoing experiments to discern
whether this is actually happening if it

is happening it's quite exciting because
it could be a way that we could modify

the bands structure of two to allow for
you know different types of different

energy energies of holes electrons to pass
through but this is an ongoing process but

fundamentally we do have
conduction through the T O two and

we think it's either through this
conduction band mechanism or

through a midcap States mechanism
we have measured the big gap States

with X P S of the techniques they
are there and they're at this energy and

this is significant they're about ten of
the twenty percent per cent meter cubed

is a pretty high density
of admin get States but

the actual could actually make it
a mechanism as yet to be resolved.

All right so.

Nickel oxide is another protection layer.

That's yet to was has been studied for

a few years now and they're still we're
still trying to understand exactly how

the connections happening Nicol
oxide happens to be actually better

than T O two in terms of performance
this is only discovered recently but

one of the reasons why is because its
properties are exactly what you want for

these protection materials first of all as
transparent as a high band gap it's also

very stable alkaline environments even
more stable than to tame oxide and

some cases it's also P.
type and so there's no question

about how the holes are being
transported through the material.

And finally is catalytic for all we are so

all these things together make
it a very good protection layer.

So here's the stability of a of a silicon
electro that's been protected with nickel

oxide and you can see here that we're
going to more than a thousand hours.

With very little current decay and
so that's.

Very promising performance.

People at Cal Tech are now making full
water splitting devices from these

utilize these protective
layers basically for for

a performance that's basically never been
seen before so it's pretty exciting.

All right so I think I had this slide of
that one more slide just to end up and

sure enough time you have just a few more
minutes so just in summary I talked about

two different systems right we had battery
systems and I talked a lot about how

fundamental understanding of these
reaction mechanisms of mechanical

degradation has really actually greeted
direct links between engineering.

Nanostructures for
better performance and in a lot of ways

we wouldn't have been able to do that
without these fundamental studies.

And then also talked about water splitting
and solar driven water splitting and

how we can tailor interfaces and
these materials in these devices for

dramatically improve stability of
performance so examples of two different

systems so for the last slide I'll
just talk a little bit about what my

group here is working on so actually meant
to graduate students are here in the room.

Are my group here at Georgia Tech
is focused on bridging material so

this in-situ characterization and
device performance and the letter chemical

systems and as you can probably tell this
follows directly from my previous work

we're interested in a few different
projects right now we've started a project

on on tailoring interfaces in solid
state electric chemical systems a lot of

a lot of times interfaces in solid state
systems like batteries are reactive and

have to be tailored controlled for
greater stability.

I'm also quite interested in
working on the scale in situ probes

to understand Ledger chemical reactions in
chemo mechanics and battery material so

this goes beyond in situ D.M. which we're
still doing but trying to correlate.

Face transformations in structural
changes across the length scales and.

Systems is quite important I think
it's really the next frontier.

And finally we're working on a project
related to ionic materials for

resistive switching and this is kind of
an esoteric application when it comes

to electric chemistry but it's actually
for nonvolatile switching and memories and

resistive switching All right so.

I like to thank everybody for your tension
these are some of my previous coworkers

and advisors my current group
here at Georgia Tech and

previous funding sources and
also a little bit of current funding so

that you very much for attention I'll
take any questions there are any Thank

you.

Thank you.

Yeah largest a rhesus Yeah probably
about one point five zero History sis.

But interestingly it has
a lower history assist then

than other conversion reactions and is
thought to be due to the fact that copper

has a very high diffusive eighty and
the material.

Has to do with me that's
right that's right.

So.

There should be here.

Next Tuesday.

You little.

Use.

Like some sort of oxidized metal.

Yeah yeah yeah yeah the question
of the comment is that perhaps it

could be more reversible to go to an
oxidized metal in terms of the conversion

reaction Yeah I think you're right
about that manganese could be.

Actually the useful I'm sure people have
studied manganese oxide are making useful

fires a conversion reaction Yeah.

The problem with that is you'd
have less capacity available

because the capacity is driven by the
change in the oxidation state more lithium

you can put in right and so you'd be
trading off maybe more reversibility for

less capacity.

Mercury and I yeah maybe so
I don't know if I want that in my battery

that.

Is.

Just.

So at the interface.

So.

Yeah yeah that's a good comment
Yeah definitely actually

my colleagues have done some X. P.S.

on that interface careful work to show
that there definitely is a mixed silicon

tame oxide I don't know about them on
oxide but if that's been shown previously.

Maybe you showed it.

OK OK OK great thank

you.

What is the pressure size.

One factor so it's driven by the
fundamental nature of the of the volume

expansion the reaction itself so
in this in the crystalline silicon case.

You had this sharp reaction front and you
actually have an isotropic expansion so

you have group more expansion in certain
directions this driven by the crystal

energy of the lattice.

And because of that you have stress
concentrations that are much higher and so

you have a small critical fracture size
in the amorphous case the critical factor

size much larger because you have
basically have lower stresses that evolve

and that's just because of
the way that the reaction works.

Yeah I mean I'm officer can should
be more popular so I think.

In terms of.

In terms of previous research that this
been done most people have looked at

Crystal and silicon just because
it's easier to make usually and

then a structure form but in terms
of actual applications I think that

manufacturers that are looking
at Silicon in their

batteries are definitely considering
a more facility and because of its greater

versatility.

Like sodium or something yeah yeah that's
definitely a large research area each

different alkali I on that you use
has its own serious challenges

beyond just fracture right so so in this
lithium silicon case we've gone from

many many choices of materials to silicon
and we've narrowed in on this last one

of these last few problems which is the
fracture mechanical the British in problem

bigger Titian problem but
if you go to if you go to sodium or

magnesium ions as your battery as
your active ion in your battery These

are very young fields actually They've
only started in the last few years and

there are many many issues associated with
that but there's a large research area

many people are looking into that
lots of work to be done in that area

so.

Yeah it's a great question so I didn't
talk about that all at all I guess so

it turns out that if you deal with the
eight silicon no matter if it starts as

a more facility in a crystalline silicon
when you deal with the it it remains

a more FAS Yeah so it always through
cycling it remains amorphous so

starting with crystalline silicon
is not really a good idea.

As it turns amorphous anyway and it stays
amorphous but it turns and the way that

people have studied this material in the
past is that they've they've pretty much

always used crystalline silicon in
the last few years this is changed

you know as this research has come out and
as people have seen this but definitely

amorphous is more reversible and also
it retains that structure with cycling

so.

Yeah yeah so the electric field

are you kind of asking about the specifics
of the in situ experiment or just in

general.

Yeah so so in a normal battery
the electric field is is generally pretty

uniform because you have these parallel
plates and you have a large surface area

in my in situ experiments the electric
field is highly concentrated because you

have maybe a single man a wire and
a counter electrode and

you know a very small gap and so
you can have very eyelet your fields.

But in general in general one of
the important things about these in situ

experiments is that you always have to do
some exit you experiments that maybe not

will tell you the same things but
they will give you some idea

that what you're seeing in situ is
actually relatable to a real battery so

in all these cases we were we did exit
you experiments to look at before and

after transformations to ensure that what
we were seeing was what would happen or

what is similar to what happens
in a real battery system so.

Yeah yeah so it is a bit different
than a real battery system but

it's it's the only way that you
can really do it and incited so.

Well.

I mean.

How do you nationalize a yes so you apply
a voltage Yeah yeah so just like a real

battery you have an electric car a letter
of material active electromechanical you

bring them together there's an electrolyte
of a tween and then inside the T.M.

you apply a voltage the special sample
holder allows you to apply a voltage.

That.

The beam Yeah yeah yeah so there are beam
effects but they don't cause let the ation

the beam effects could be controlled if
you if you're under high magnification or

these experiments you can actually see the
material changing the main beam effects

are that the lithium containing materials
inside the T.M. are kind of unstable so

if you zoom in they actually
the lithium actually evaporates but

basically all the images I showed with
the silicon it was a man if occasion and

if you did the experiment
with the beam blanked

you get the exact same transformations so
that's a good way to test that so.