Yes.

That's a

serious.

Air strikes.

Right.

There.

OK.

Thanks.

Very.

Much.

For.

A lot.

Was

elected.

OK Thank you Bernard so
I don't have a microphone today so

if people find me hard to hear maybe you
could have a closer they think is the best

solution so just before I start with
the talk I should say a little bit to buy.

Imperial College So this is actually
the fifth time that I've visited

Georgia Tech or
Georgia Tech organized meeting and

this sort of relationship
between Imperial and

the organic electronics community at
Georgia Tech dates back about ten years or

maybe more maybe more than that and
we have as you have this center for

Gammick Protonix and electronics so
we have an imperial center for

plastic electronics and the director of
it is Donor Bradley who's shown here and

he's the guy who brought.

He was one of the guys who discovered
electro luminescence in conjugated

polymers and he was a guy who brought the
organic electronics to Imperial College

and so we now have a research community
that spans across the chemistry physics

material science and a bit of device
engineering of organic semiconductors and

I should say printed semiconductors
generally because not all of them organic

and various applications and
as Bernard mentioned so

I've been active in the field of organic
photovoltaic for quite some time over

over ten years and also a lot of other
people in our community at Imperial and

the community in the center for
plastic electronics there is able to work

together on you know sort of some
application area if you like but

bringing together the different
skills of of the people involved so

for example in the area of O P V We have
a number of different problems that need

to be solved and
then we also have a number of experts and

some of you may may know some of
the the individual shown but but

the good thing is that under
one roof we have experts

in the chemistry physics
material science and

engineering of materials who can work.

I gather on common problems so
it's very nice for

me to be able to represent the C.P.A.
at the C O P OK so

I'm going to talk about what I decided to
talk about some work we've been doing.

Really over a number of years but most
of it is quite recent where we've been

working on the problem of trying
to relate the molecular packing

inorganic semiconductors to their
own to electronic properties and so

when we should maybe maybe say that the
main people involved in this work will be

Gilbert who support stock and was
a student in my group Dora to need to lead

who is also a post-doc in my group and Flo
Steiner and Beth Rice who are two Ph D.

students and another key contributor
is Jarvis Frost who was a Ph D.

student and post-doc with me and
now is in the University of Bath.

So when we speak of structure inorganic
semiconductors then this this is important

this can be defined on a lot of
different scales we can think about

the chemical structure of
the compound concerned and

the molecule may exist in a different
configuration different this or

twice for example says so
Trans configurations or

you may have different isomers with
the same chemical structure and

it may also exist in it in a large
number of different conformations So

the molecule itself varies in its
structure but then in any solid

film we're concerned with the molecular
packing and then that's something that

will concern not only the characteristics
of the individual molecules concerned but

how their backbones and side chains
interact and then when we move to a multi

component system for example this is
a kind of like an artist's impression so

Jarvis made this image a number of
years ago as an example of how things.

Might look at the interface between
P three H.T. and peace B.M.

inside a solar cell if we want to really
build up a picture of an interface

like that there's an awful lot that we
need to do that we need to know before

we would have any chance of
doing that with some accuracy.

So just to.

Prove.

To you just to to do to give some
pride some motivation really for

the work the point.

Here I wanted to give a couple of examples
of the way in which the structure

adopted by a conjugated polymer matters so
the first example

data shown after there is showing how the
whole mobility of a poly flurrying polymer

can be changed by more than two orders of
magnitude by the processing conditions

the change the confirmation
that the polymer adopts.

We can then think of examples where
the optical properties particularly in

the solid state are influenced by
the structure that the polymer adopts or

the semiconductor adopts controlled for
example by choice of solvent or

processing technique and as well as the up
to electronic properties there is also

plenty of evidence that the structure for
example the molecular weight

of a polymer can influence its mechanical
its thermal and mechanical properties and

as an example they're showing
are the tensile strength the elasticities

of polymer films depends upon the
molecular weight these are all motivations

to understand it and then to give
the context where I come from so

as well as being somebody who works on
organic semiconductors I'm also I've also

had a longstanding interest in trying
to simulate the electronic properties

of different organic
semiconductor materials and.

And in that sort of framework that we and
of course other

groups have had similar approaches
have been working on we need to know

the electronic structure of the component
molecules we need to have some models for

how the molecules would be brought
together into a solid film.

And having done that we can
then calculate in principle

the coupling between different
components different individuals.

Transport sites or sites it could
accommodate charge within the film and

then we can use that
simulation with one or

another realisation of the of
the of a part of the solid film and

then simulate the dynamics of charges or
in principle of exit on Zor or

other species through that film and
in this big picture the really difficult

part of course is a part where you
simulate the molecular packing and

the reason it's difficult is because for
all the reasons we just looked at it's

rather complex that makes it
a complication really intensive to

explore the many different
configurations that may exist.

And as well as that it's rather hard
to know when you have a description

which is actually correct of course
you can simulate a structure.

But how do you know when
that structure is correct

how do you know when the the ingredients
the potentials the formulae

that you use to build that
structure are valid or not.

And that has remained really a challenge
throughout the time I've been

working in this area and

of course this really matters because
the organization of the components is

going to have a very strong influence
on the optical or the electronic

properties that you would then simulate
from the structures that you build.

So what I want.

Good to do then today was to say
a bit about methods that we used to

simulate structure and then go on to
a number of case studies where we had been

using largely using molecular
dynamics to try to understand

the behavior of some different material
systems so to start I just wanted to

mention something here so back in I just
remembered when I was writing the talk

at back in two thousand and seven I pray
I was at the same meeting as Bernard and

and Seth in in in an island off
the state of of Washington and

I remember showing this
simulation where we tried to

reproduce the an experimental observation
so that the data on the right were from

Joe Klein who was there who had been
a student in Stanford and he showed

that you get a different structural
organization and a different mobility for

short and long molecular way police fire
feelings and then we wanted to try to

see if we could reproduce we produce this
appearance of enhanced Crystal in a T.

for the low molecular weight using a kind
of I'm going to say a toy model and

the results of the toy
model are on the left and

this is something I
won't say more about it.

But simply to say that in a model like
this so we used reputation to kind of and

introduced an interaction between
different units in our structure in order

to to to to then build up some
sort of typical structures

that would result from the molecular
weight of the species that were used and

the size of the interactions between
the different layers in the structure so

that sort of thing it's quite good for
asking a kind of what if question you

know could different molecular weight give
us something with different structures but

it's not very good for

predicting it won't tell you how any
new chemical structure will behave.

It doesn't contain the detail that you
need so to get the detail you need to go

further and the sort of method of.

Sort of choice to study
how different chemical

structures would perform
we would behave in in

some situation in solution or in a solid
film is atomistic molecular dynamics and

here you would consider
individually the atoms or

possibly groups of atoms inside your
molecule and and calculate the strength

of all the different types of interactions
that they can have both the bonded

interactions between the atoms in them in
within the same molecular structure and

the non bonded interactions between atoms
in different molecular structures and

a lot of work needs to do is to be done to
develop those force fields when you have

a set of force fields you can then
do some kind of simulation and

generate a structure so you could
simulate the structure that a load of or

legal birds might form in the solid state
or as you condense it into the solid state

and then allow it to reach what you
would then judge to be an equilibrium

configuration and then you could
look at typical dynamics and

typical sort of structural.

Sort of characteristics of that system so
this is this is generated assembly and

this is sort of some information that we
were interested in this is the sort of

frequency of flips in the torsional
angles within those illegal Mares OK.

Now.

If you try atomistic molecular
dynamics to a relatively

simple system then it
can together with this

multi scale sort of simulation of
transport it can give you some predictive.

Our This is an example it also
actually dates from two thousand and

seven this is another study we
were we were collaborating on but

this is the work of Dennis and
drank all in the Max Planck Institute for

Polymer research in mind and
his group and they were trying to explain

by just looking at the differences in
the side chains attached to a hex abends

of Coron in discussing liquid crystal they
were trying to explain why different side

chains lead to different sort of degrees
of order in the discotheque pattern

packing that those molecules adopted and
also the different electronic

transport properties that had been
measured and they were able to do that so

they they they characterized
the force fields between the.

The disk otic liquid crystals with
different sign chains attached and

then they were able to predict
the conditions in which you would get

the different phases that had
been observed experimentally and

then we collaborated with
them to predict the home

abilities that were measured in the
different systems and we could see that

yes if you have these long alcool side
change you get a more ordered packing and

you get a higher home ability and
if you have some Both the side chains

then you get a more disrupted packing and
a lower mobility so that's fine but

that was a very simple system and this is
another simple system that we worked on

where we we were growing
films of C sixty and

then varying in the simulation you can
vary the substrate temperature and

that has an influence on the size of
the grains that can be produced and

experimentally it had been seen that
films with different sized brains

were expressing different
electron mobility.

And.

We build the structures and
then we could simulate F.

E. T. electron transport and
the power the trend was well there were

two very different trends in the
literature in our trend was in between so

at least we could say we agreed with that
we read as well with the literature is

the literature did with itself but
again it's a very simple system.

For applications we're often interested
in using conjugated polymers for

a large number of reasons and
conjugated polymers of course it's not so

simple because you have many more
degrees of freedom and particularly.

There are degrees of freedom
that affect the pie system for

example the torsion within the backbone
of a conjugated polymer but

also the degrees of freedom that would
affect side chains where they come in to

the conjugated backbone and whilst there
are a number of packages out there

that are used to do molecular dynamics
they are generally speaking not that

well characterized by conjugating systems
they tend to have been developed for

proteins maybe for biological systems so
there's work to do and

when we come to apply atomistic M.D.T.

conjugated polymers then there's
you know a number of issues and

I just identified some want some
that I want to talk about day so

one is the richness of the number
of different structures

that you might predict and how you can
choose between them another one is

how do you know when you've got
force fields that are any good.

And then the other issue is of course
the sort of the computational expense

of exploring a larger very
large number of different.

Sort of.

Configurations of your system.

Within a sort of reasonable computational.

Time and just to sort of comment on
the last I mean this is an image on the.

Right which was the result of M D on.

A series of legal birth of Polly siphoning
and we could look at the structures

there but
these are only hexadecimal urse and

anybody who works with police in they know
that you you have molecular weights that

are maybe an order of magnitude higher
than that as you increase the molecular

weight the time that you need to sort
of try and simulator a reasonable solid

state structure is going to increase super
linearly So there really is a constraint.

OK so I'm gone now and the first thing
I'm just going to give an example and

this is work that we've already done and
and published but

it's it's just a nice example of
what you can do with atomistic M.D.

to try to explain some observations so
if we go back a number of years the.

Slide to pan to die fire fiend.

BT Co polymers they were very interesting
low bandgap polymers particularly for

solar cells but also you know
potentially for other applications and

first of all the one with the carbon
bridge was was published and

then sometime later an analog with
the silicon bridge was published and

between the two if you don't
call process the carbon bridge

done along with something else
then it performs rather worse in

generating four to current in a solar cell
compared to the silicon bridge to analog

which seems to work quite nicely by
itself the silicon bridge analog is also

it also has a higher degree of crystal
limiting it has a higher mobility and

also when you go in and inspect the X.
ray diffraction patterns

you can see there's also differences in
the type of structure which is adopted.

So in both cases you have a limited
structure but in the case of the silicon

bridge the pie striking distance seems
to be rather shorter and that is at

the expense of the miller striking
distance which is rather larger so.

This sort of body of experimental data on
these polymers then gave us quite a nice

sort of system to look at to try and
explore using molecular dynamics so

this was the work of an shown of her and

she studied then the different
degrees of freedom by doing so

she'll do quantum chemical calculations
on the potential energy of your molecule

when it's constrained to different
positions and then you can constrain it so

that you vary let's say the torsion
angle between the units in the chain or

to vary the angle of flexion between
the two sides chains etc and

so when she explored the different degrees
of freedom in the system she found that

the place where silicon substituting
the carbon bridge or the silicon bridge

a place where that makes a difference is
in the flexibility of the side chains so

the potential for flexing of the side
chains becomes softer when you use silicon

and also the potential for rotation
of the side chain around the first

carbon in the chain again it becomes much
softer when you use silicon so to say to

put that simply the silicon bridge makes
society change much more flexible.

So then we wanted to understand
this polymorphism and

to try to understand why it is that in the
case of silicon you get this closer price

that crystal structure that in case of
carbon and so what what she did was she

then built this kind of model sort of mini
crystal of with the mellow stacking and

varied the sort of principle distances
the pie striking distance and

the distance and then would.

Do molecular dynamics on whichever
set of parameters we've got and

calculate the average sort of energy
of this configuration and then build

up a potential energy surface we wanted
to then look for the minima because those

who'd be the places where the most stable
crystal structure might be found and

the interesting thing was we find actually
exactly the same potential energy surface

for or you know the same to within the you
know the area of the fluctuations for

the silicon and the carbon bridged they
both seem to show more than one minimum so

they're both expressing polymorphism.

But what actually had happened was that
in the samples we looked at with X.

ray diffraction it seemed that
the silicon one had managed to find

this polymorph with the PI
striking distances shorter and

the Lamell are stacking distances larger
whilst the carbon bridge one had not

managed to find it and that we
attributed that to sort of an increased

this sort of the to the effect of the side
chain flexibility on the kinetics

of structure formation during
solidification of the film.

So the.

Bridging at him is important but
it was important not because it changed

some torsional potential
in the backbone but

because it changed the sort
of the the dynamics

the ability of the polymer to explore
the different configurations.

So then another one to go on to another
family of polymers this is something

that's occupy our attention rather
Dorothy's attention in particular recently

and this is the family thought there are
almost related to the last series that we

looked at but now these are the Indus seen
no Die thought if in coal bones a fire

die is all polymer So my colleague and
McCullough did a lot of work on this.

And there they played a bit with
silicon and carbon bridging in

analogous way to the die five
fins that we just looked at and

the question I want that we decided
that we wanted to address is

what happens when you fly or
in this BT unit.

So.

In.

His student Bob Schroeder made a very
nice series of polymers where they

included the BT unit either with two
hydrogens or two flurries and then

also looked at other structures with
with fire fins flanking the BT and

then looked at their structural and
electronic properties and

the rationale was that as I understood it
was that introducing the floor in a flurry

is should play in your eyes the backbone
because you have an interaction between

the sulphur on the fire fin in the floor
in that should help the plane rise it and

so the understanding of the belief at the
time was that with a more planar backbone

you want to then get.

Better and sort of facial packing
of the polymer chains and

that would enable higher home ability.

But the the outcome was
that yes one student did

you did get a very well
defined crystal structure.

Very well high degree
of crystal in a city.

Where when the polymer is related and
the mobility what was not better in fact

a direct comparison showed the mobility
of the fluorinated compound was worse for

this particular version of
the I degenerated compound and

quite a lot worse than a number of
other hydrogenated compounds and

in addition so we also looked at them for
solar cells and then in a recent study and

we compared the fluoridated one
with another Nowshera native id T..

BT and the photo current
is very disappointing and

in this particular case it turned out
that charges were then leading to

triplets in this material system so the.

Fluorinated id t b t is not
performing as we would like so

we'd like to try to understand that so

we investigated the molecular packing of
these materials used using and doing.

So the sort of method
that we used was this so

we wanted to try to identify the ways in
which these chains could pack together and

then having packed them together to see
how they're what their electronic copying

would be like so if we think about
the confirmations that could lead

to a crystallization then it
makes sense to select the linear

conformation where the orientation of
the BT unity is flipping every monomer and

then we have to state or consider in what
in what ways could two chains like that.

Join together and so we can consider
whether they are if you like.

Or sits sitting on top of each
other with the bts in the same or

the opposite orientations we call
the system trans and we can also consider

the relative shift in the molecule for
one relative to the other and

so there's then a number of different
structures that could in principle exist

some of them can immediately be
ruled out because of the staring

hindrance like this transition here
this wouldn't happen and then we also

need to consider the interactions between
the chains that would lead to a greater or

lower stability for
one structure relative to the other and

then having selected some plausible
call facial parking structures we

then look at what parking
structures could be possible so.

So we explore a lot of different
possible rangelands of the molecules and

then we calculate the relative
stability in order to try and

identify the ones that we think
are more likely to form and then for

we also calculate the transfer
integral for whole transfer between

neighboring labors in
the pipe stacking direction.

So what we found.

In the case of the this is
the hydro generated component

we found the most likely.

Packing structure is one where the chains
are on top of each other in this

sort of issue like Trans configuration.

And the side chains are into did
you take in pairs in this quarter

pen like mechanism and an important
feature here is that the BT unit without

the flaring has a rather strong dipole
moment and that's going to encourage

the BT units to sit on top of
each other and to interact.

With each other but with opposite
orientation and this particular structure

led to a transfer into a which was
the order of between ten to minus two and

ten to the minus one electron volt whereas
in the case of the flurry native compound

then you no longer have this
dipole dipole interaction.

And the most stable of the structures
that we found was one where one

chain is shifted by.

A unit relative to the one below and
they sit on top of each other in an assist

configuration and
that has a rather weak and

in fact a very weak electronic coupling
and there are other structures which could

be formed but none of them has
a substantially higher coupling

they don't have the same.

They are there they don't have
the same tendency to park.

In this.

Sort of well aligned structure
that's brought about by the BT unit.

We also were able to look at
the dynamics of the chain assembly and

what we could see was
it in the case of the.

Of sorry that's right in the case of
the destroyed in case of the fluorinated

compound the chain assembly happens.

Almost similar Taney Asli so
the price starting happens first but

the Lamell are stacking happens very
quickly after it it all happens very fast

it kind of these these chains sort of zip
up and we can see that in a in a movie and

in a minute if it works whereas in
the case of the Hydra generated one

the price psyching still the price
are still happens but it happens first and

it takes rather more time for
the stacking to occur and so

this could have some I suppose
impact on the structures that you

get when you use these and
they're in a blend with another component.

So we could say was that
in this case the large J.

Results from really from a call facial
packing structure that are driven

by the interactions of the I
dredge elated BT unit.

So also is a kind of in
defense if you like of

the structures that we found we simulated
the X. ray diffraction pattern and

could compare that with the patterns that
have been reported experimentally and

we could identify features in the specter
that are due to the site change due to

the packing and
due to the side chain packing as well.

So this family of materials is also
quite interesting because this

si sixteen analog Now this
is of an id t b t would you.

Not silicon bridged its carbon bridged and

that was published in a paper
by sending searing house.

Had made the polymer and the group
in Mom's where were co-authors and

they published this last
year in nature arguing that

the polymer is showing
trap free transport.

And.

In their measurements they had a mobility
a very nice if you will billet you over

two to three centimeters squared Popov
all second now this structure from

our simulations is actually not
the one that we would expect

to be the most stable to form and
neither is the one that gives the highest

electronic coupling However
the structure that we do expect

would expect a form for this compound
sorry this sorry that this is the wrong

structure that should be a carbon and that
should be sixteen I meant to change it but

I didn't the structure that would form.

The we would expect a form for that
chemical structure will be the one that

gives a very high electronic copying
a whole coupling compared to the a number

of other alternatives so what we could say
is it was a very good choice of polymer

to get a high mobility transport
because the limitation of

interchange hoping is going to be
relatively low because the chains can and

do assemble in such a way that
their orbitals are aligned

to give you a vertically strong electronic
coupling it doesn't have to make a crystal

that just needs to happen some of
the time whereas with the change

to the structure like fluorinated in
the BT unit you would no longer have

the likelihood of forming
the strong interactions but we do

have a question about whether that is
actually the published structure is there.

When we think it's unlikely and another
thing to maybe comment is that you don't

need to have an extremely close
stacking in the experiment and

in our calculations it's around about
four and a half and strums which is not

particularly close but nevertheless
the coupling seems to be strong.

So the conclusion from our
part is that the dipole driven

by stacking can in some cases interact
in a dominate the interchange coupling.

And then with just.

I want to turn.

This Earth it's just a movie showing the
the this is this is the fluoridated analog

showing how you you start
off you've already.

Said the chains are assembled in the pie
stuck in direction really quite fast and

once they've done that.

These these pairs of chains
can then interact and

interlock in this market mechanism.

OK so I raised the question
a little bit earlier about.

What how do we know that force is the
force field that we're using are any good

and so I wanted to mention and partly
cause the neutron scattering expert here

some of the work that we've been
we been doing to explore that.

Using data from quasi
elastic neutron scattering.

So.

We do all the simulations we do all the
simulations with force fields either that

somebody else has calculated and published
or if that hasn't been done we'll

calculate them and then we introduce
those force fields into a molecular

dynamics program and then you grow some
structures and then you calculate some

properties of the structures that you grow
when you compare them with experiment and

you say that's great or that's not so
great but how do you know.

Force fields are any good and so

we had an opportunity to look
at this by analyzing and

modeling the results of a set of
course the elastic neutron scattering

measurements that were done on films
of probably five Faine in one case

police three Haxo five and
in the other case police three octal five.

And these are the neutron
scattering the elastic incoherent

structure factor data against temperature
for the two different polymers OK so

you can't read very much from that but
what you can see is that there is a lot

of data here which will reflect
the dynamics of the hydrogens

inside chains as a function of temperature
for the two different compounds

now how can we use that how can we use
molecular dynamics to look at that well so

what we did was we took our our structures
the police and he three H T M P three O.T.

and in this case force fields had already
been published for police I mean by

the group in in Pali me in Milan and so
we used and adapted their force fields and

then carried out molecular dynamics of the
P. three H.T. in the piece we oversee and

fill a sort of a little
mirth long leg of mirth.

As a function of temperature.

And then we could analyze
the dynamics are simulated to.

Look at a few things one is to establish
which dynamics we would expect to see or

not see within the time window of
a neutron scattering experiment and

then having established which ones we
think we could look at we could then

compare the measured elastic
incoherent structure factor

with one that can be simulated
as from the molecular dynamics

simulation and then finally having done
that we can go right in and find out.

The activation energy for different
degrees of freedom in the film and

I'm not going to go into a lot of detail
about how we did this because there isn't

enough time but
just to give some of the results so.

We built one built up these
sort of relaxation maps and

this is a map of the time constant for
a number of different

motions in the side chain as a function
of the inverse temperature so for

example you could define motion as
being the relaxation of the dihedral

angle between two carbons in
the side chain back to its kind of

equilibrium position and you can look at
the results of molecular dynamics and

and find the time constant so we find the.

The the.

Time constant by calculating
the correlation function for that dihedral

relaxation and then if it has relaxed
within the time scale of the of our M.

D. simulation go out to about
thirty nine zero seconds.

Then we would include it in if it
hasn't relaxed then we just say we

don't know anything about it so then
the thing the different degrees of freedom

is these are different twists different
contrasts between different carbons in

the side chain we find the time
constant from the M.D. and then we

compare that with the times that different
instruments are able to look at so

we did some of the some of the
measurements were done with the or Cyrus.

Beam line in Oxfordshire and some of
the others were done at a tire level and

you can see that of all the possible
motions only a certain subset of them

are actually likely to be visible within
the time window of the neutron scattering

experiment so
then we were able to sort of simulate

the structure of the structure
factor from the molecular.

Dynamics and we do that by basically
building up the sort of distribution

function of the of the hydrogens
as a function of time and

then taking a furrier transform of
that probability distribution and

we take it in the long time limit and
that should give us the equivalent of

the incoherent struct
incoherent structure factor and

so here the experimental measurements
are represented by symbols and

the the simulation are represented
by these and full lines and

we would say that we were quite pleased
with the agreement so it's not.

It's it's it's not perfect agreement but

the agreement is reasonably good and
certainly more or

less within the scatter of
the of the experimental data so

this we would say gives us some confidence
in the force fields that are used and

now we can go back and
look at some of the.

Some particular features of the of of
obviously dynamics that we simulated And

here we've tabulated the activation energy
so for different degrees of freedom and

this is all the degrees of freedom
along with find chain for the heck so

case for the octal case.

And point to make about this is that for
all of these sort of dihedral

rotations in the alcohol chain when you're
far from the backbone like out here or

out here it has a typical activation
energy that you would expect

for a chain in an isolated environment but

when you get close to the backbone the
dynamics of those signed chains are more

restricted fact they're more like what
you would see for I saw Pentium than for

the typical Transco she reorientation so.

Here the M.D.
study is telling us that the sort

of the activation energy that you would
sort of calculate from D.F.T. or from

a sort of a prior study for a molecule
in a vacuum is not representative for

the molecule in a packed environment
the molecules the side chains are actually

stiffer close to the backbone than
you would expect from a study on

isolated molecules and why is that useful
Well that is likely to be useful for

the next thing that we want to do and also
the next thing I'll talk about the not for

polymers which is how you would
parameterize the interactions if you want

to then of course Crane the structure and
build it with a more efficient with a more

of a lower resolution and with this
sort of greater computational scope.

So.

The conclusion from one part is that if
you like the force field are not too bad

at least for
these probably five in polymers but

we have seen that the molecular packing.

Has an influence on the motions close
to the bank bomb that the molecular

packing in the solid film sort of
restricts the motion of the carbon atoms

close to the backbone and
that can be quantified OK.

So a comment a bit earlier that one
of the limitations of atomistic M.D.

is size of the structures
that you can build.

And.

We can't build a conjugated polymer which
is got one hundred repeat units and

have a solid film of them and really
hope to get any representative structure

we would like to do is to
simplify our system and

pick out the key degrees of freedom and
then make some simulation of the structure

based on those now that's called
course craning It's a technique which

is well known it has been
applied to a certain extent for.

Gated polymers But
what I want to talk about here is

an application where we've sort of
been able to use course craning.

With abandon and that was to
the case of the elec this sort of

molecular packing and then the electronic
transport properties are full of rings and

I use it to give an example of of
hard this approach can work so

we were interested in these different
adductor one two and three adult.

They were introduced anybody who's worked
you know P.V. will know them quite well so

the hieratic who are introduced they
have weaker except for strength and

they were expected to and indeed they
did give rise to a higher voltage

in solar cells but then they were
limited by their ability to pack and

their low transport properties and all of
those problems were many of those problems

arose from the fact that when you make
base for the trees out of one of these for

a range you don't just get one
isomer you get several and

they're not easy to separate So
what's nice about

these Is it gave us a system where
we've plenty of experimental data but

something which could
then be coarse grained.

And and the results of the Course
graining could be studied and

so we course grained our this is
a business and we represented it by

saying saw the core screening of C sixty
is already done it's a spherical potential

It's got a certain weight a certain decay
with distance and then we introduce to

a comparable potential for the side
chains we find chains is a ball and

then we needed to fit two
parameters two to sort of disc to

set the position and
the center of the center of force and

the strength of the of the force
fields do to the ball and

that was done by doing sort of flipped.

No Completely atomistic
simulation of P.C.B. and

then doing course grained
simulations of P.C.B..

Where we represent the P.C.M. by the two
balls and then we do this over and

over again until we get the parameters for
the for

this course craning that give us a
reasonable agreement so we're not going to

get a perfect agreement with the radial
distribution function from the atomistic

simulation but it's you know it could
be fitted so that it's not too bad so

then having to find the force field what
can we then do so then we can use the same

force field to make the best
films of the base and the Triss.

And build up structures and then we could
build up structures containing these only

contain about a thousand but we also have
structures that contain up to one hundred

thousand molecules and then we can look
at their properties so in terms of

their structural properties we could look
at the radial distribution function and

you can see that as you put on more
side chains you are effectively reducing

the coordination number so the probability
of finding another fullerene

right next just at one kind of
wonder radius away is reducing

as you increase the number assigned
chains enough because it becomes more and

more likely that the BE A side
chain stuck in between.

So however in spite of that the effect
of these additional side chains.

Wasn't very strong I mean the.

First coordination number actually
hasn't been changed the coordination

number hasn't been changed that much
by the addition of side chains so

we have our structures and
we can then simulate electron transport in

the structures and we did that in
a number of different cases so

these are the experimental if you team
abilities over here this is a log scale.

So we then simulate electron transport
in the case where the we have

only we have no energy disorder every
site every fullerene is identical for

the transport of electrons and and

the only effect on transport is going
to be the different distances between

the fullerenes that results from their
additional side chains and in that case

we see that the additional side chains
reduce the mobility but not that much.

And then if we also are in energetic
disorder that results from

there being a number of
different isomers then you get

a trend in the simulation which is much
more like the trend in the experiment and

the thing that we concluded
from this was the for

something like a full arena which has
a very isotropic electronic coupling and

that packing disorder and is the effect
of packing disorder on transport

is relatively weak if you're molecules
are going to couple no matter how they're

oriented relative to each other then their
electronic properties will be much more

resistant to disorder disruption in
the molecular packing and also the fact

that they're approximately spherical
also allows them to part quite well.

So in the case when our coupling is
isotropic It's a site energy disorder

which dominates.

Now another advantage of
being able to treat your

course grain your organic electronic
system is it then offers you

a sort of rather nice tool to look
at the electronic structure so

we can go bang into very basic
quantum mechanics and we could treat

the electronic structure of a coarse
grained structure using a tight binding.

Model So what do we need for
that well with fullerenes I just

mentioned the coupling is isotropic so
we can we can already calculate

the electron transfer integral as
a function of the separation of course and

it doesn't depend too
strongly on which add a Q.

and we can then use these couplings we
can take our positions from a simulated

structure take the couplings from
the separations in the simulated structure

we can build up a tight binding Tony and
and we can solve it and

that can then give us two things one so
it give us the the set of eigenvalues for

this system and it will also source so
this set of eigenvectors of this system

and also the eigenvalues the eigenvalues
to give us the density of states and

then the eigenvectors which I'm
not able to illustrated the moment

give us some idea of the spread of these
states throughout the system my P.C.P.

is quite relevant system for
that because it's been argued Can you or

can you not treat.

P.C.B. the transport in P.C.B.
am in the weak coupling limit and

it's been argued that you that you can't
and if we look at the distribution of the.

Or the extent of the way functions.

That would support that idea but
then there's a whole other side of it

which is what happens when you populate
one of those states with a charge and then

you form a polar on and that would then
tend to localize that's another part of

the story it's not particularly relevant
here but my point here is that the.

The sort of course craning kind of
will will will will quite nicely

carry you through to a treatment
of the electronic structure of

the approximated compounds and
one thing just to let you know what

we're doing with the with the full
rings so we were rather almost.

Shagreen to find that this the was
almost there was almost no effect on

the density of states over replacing your
one with two with three side chains.

And so we're trying to force
a disorder transition in Philippines

by putting more side chains on and it
does something interesting maybe somebody

will make these who know us anyway.

So this is just to say the core screening
of polymer is obviously is an objective

and there are a number of
of of challenges here and

I mean I think particularly amongst them
when we think about the things that we

know polymers do one is that with
coarse graining It's not that easy to.

To to to bring about
the crystallization of side chains and

we've seen that the organization of
slide chains is quite important in tort

terminating the structures that
the conjugated polymers will form and

another is that when you represent
any group of atom by and by a single

unit that single unit then represents
a lot of different microstates and

make different microstates would then
embody a lot of different electronic

interactions because the electronic
interactions are strongly dependent

on how close the closest atoms get and
their relative position so

you can't you can build kind of the maze
of phase but the woods there's a number of

questions that still have to be worked
through regarding how you would.

Get a reliable course graining approach
for proper polymers that's going to

reproduce the electronic
interactions correctly so I have.

One more topic which is about
the grain boundaries but

I think it's I should
perhaps keep that and.

Simply to say we've been simulating

brain by injuries in maybe
I'll just talk about the.

We've been simulating the effect
of quote buying degrees between

grains This is actually in
the case of tip spent to see.

Where we start off with some
kind of stylized crystals and

then we work out the force field for this
material we allow the crystals to relax

and you see some merging of
the bar Andries into each other.

But then we could calculate the this is
work that we did together with Dennis and

corncob poking and we could we could
calculate the electrostatic interactions.

And also the effect of
induction the effect of

electrostatic interactions on the energy.

Of the individual molecules and find that
there's a kind of mountain range appears

in terms of energy between
the two grains and that

has the effect of impeding transport it
gets worse as the grains become more and

more out of orientation with each other
and that's compatible with with what's

been seen in experiments on a similar
material by Lynn live in Princeton so

let brings me to the end I think.

And just to make some
some some comments then.

So clearly understanding the structure of
the molecules that organic semiconductors

can form is kind of critical to
understanding their electronic conduct

when it Tonic properties and if we
were to try and extract some sort of.

You know.

Recommendations from from from the
different case studies we've looked at I

mean one is that the effect of
the enabling side chain flexibility is

important in allowing different polymorphs
to be discovered another is that

we've observed a rather sort
of important effect of.

Dipole driven of dipole interactions in
controlling the electronic interaction.

That can result for
different structures and

then finally that when we have a system
where the electronic coupling is

relatively isotropic then
those systems remain.

Person remain robust in their electronic
properties with respect to disorder or

I didn't really talk about it but
grain boundaries.

In their effect on
the charge of transport.

And.

I'd like to thank you for your attention
and I'll just jump back to that but

before I finish then the particular people
who were involved in this work and.

Florian Steiner he's writing up his Ph D.
right now and Beth lice and

Jarvis two is not here so he's not in
the picture but also Jarvis Frost So thank

you.

For.

Sure.

Have to tell me when food is.

A company that.

I don't know where so
what was it on Seth can you tell me.

OK

fine.

That one.

OK.

You're like it.

Way.

They do they do they
don't they don't they do

this.

To reconfigure the force field
in principle yes in principle.

I mean has got an idea why they do that.

It is not something that we're thought.

I think I think this can be done.

We're actually so
yeah I mean this isn't perfect for

sure it's not perfect yet and.

But.

You know we have.

One thing that I haven't mentioned here
when I was going through it is you may see

that these actually don't fall to zero and
what we in fair about that and

maybe it will relate to what's happening
at Nokia as well I don't know but

what we infer from the fact that these
sort of curves are not enough for

them to zero is that there is a certain
fraction of the volume that's actually

not involved so for example you could
have side change locked in crystals which

are then not moving in the same way as the
rest of it's we consider this is probably

a more spartan.

Sure.

Yes I think so I think so.

But I can't answer.

I mean all I can say is that we know this
is happening and you know comedian and

I think it's quite hard to.

I mean we know what the limitations are of
the simulations so you've got you know

a box which contains
a number of chains and.

You know they interact but
it's a limited volume and

the chains have got limited length.

You know because you can't really do them
day with something which means more than

about twenty units long as we know
what the limitations are and I come to

the moment think whether which of those
limitations is likely to influence but

we know where we have to look if you like
and the other thing is of course and

the simulation doesn't contain crystals
the experiment will contain crystals

just because a simulation
isn't big enough you know.

Because that's not can't
be done can't be done by us

anyway.

Sorry.

I think partly because of
the finite size I mean why is it

not.

That.

Well it's not that's not the transfer
into growth but it's the energies.

I believe it's find in
finite size effects of are.

There.

Well.

I mean you've got because of
the finite I mean you I mean.

I mean I hear what you say but
in the system that we've we've gone out so

you've gone you know you
go you're one thousand or

ten thousand the these are done for
the one thousand so

let's say you got something which
is basically ten by ten by ten.

Your Lowest eigenstates will be ones
which you know whose extent is limited

by the by the finite size of
the volume you're working in.

And then they highest energy state

or.

You.

I mean it's not it's
not it's not periodic.

It doesn't a period boundary

conditions.

OK we can talk we can

talk.

You know it.

Was like.

Yeah.

OK I mean you know so I don't know but
I can speculate I mean the mobility

of our so first of all the mobility values
were two measurements from from the pay.

Right they weren't the best.

Yeah so I mean that this is this is what
you know this is what Bob got when he

measured those systems and.

The.

Field their field effect mobility
is less going to be and so

you know I mean I think I commented when I
was saying this that that wasn't the best

mobility that had ever been observed for
a polymer without structure

that's just what they got when they
compared four similar ones together.

But another side of that.

Is that the.

The.

The in this particular structure
this this is this this was the most

this parking was the most.

Sort of the lowest the lowest that
the most stable one that came out of

the exploration of structures.

There were a number of others and some of
them slightly higher not not as high as

this but slightly higher coupling also for
many think what you could say is this

this there would be a number of different
couple Ling's and the there would also

be is also a likelihood that there's
going to be some chain interactions so

this is about as valid data that can
trigger a kid but that is it can get OK

and it may be the most stable but it won't
be what will always always happen in some

places where the chains would be in
a configuration that's not structurally

the most stable.

And so and so those would be
the defect of those if you've got

if the chain if the interchange
coupling is what limits transport and

then you have a few kind of defect
where it's better than that

then it would mean the mobility would
be likely to be less different.

The coupling would be yeah yeah yeah yeah
yeah sure it's simply I mean I'm not

trying to say we predict or anything
like that it was just that you know

we're interested in the chemical structure
and it seems as though it makes sense

here.

That's a good question actually because.

The answer was yes we did but
we don't anymore.

And the reason is that
we had a plan to do and

it's exactly what I wanted to do so

we wanted to repeat the neutrons guttering
measurements on those materials.

And then we discovered first of all you
need about four grams per sample Secondly

you need somebody to do treated for you.

And the silicon bridge cycle
painted It's not that easy to get

nevermind deuterated So in the end so

we have a proposal going on to
look at a need to look at blends.

Both in the end we decided
to look at P three H.T.

simply because of the difficulty
of getting hold of the polymer so

that I would we would like
to look at that question

so.

Yeah OK well.

One more time we have a.

Write off.

With a little bit small here for
you to bring back to infill colleagues.

I'm going to go thank you.