Articles, Blog

nanoHUB-U Organic Electronic Devices L5.4: Photovoltaic & Emerging Devices: Polymer Thermoelectrics

September 13, 2019


>>[Slide 1] Hello. And welcome back to the course Organic
Electronic Devices. In our last lecture, we discussed polymer
thermoelectric devices, and we introduced you to the motivation
behind why we would want to use thermoelectric devices, what some of the inorganic
materials used today are, why we might want to use polymer
thermoelectric materials, and really kind of motivated how
polymer thermoelectric materials could make a significant impact
in thermoelectric devices in a grand scheme of things. Today what I’d like to do is
move beyond what we’ve seen before and go into the state
of the art with respect to polymer thermoelectric
devices and polymer thermoelectric
materials,, and hopefully understanding what the state of
the art in these materials are, will help us design better
materials with respect to polymer thermoelectrics and really understand how we
can play and tune conditions such that we have
this high performance and higher figures
of merits. [Slide 2] So what we’ll do is we’ll
discuss what initial polymers were used in terms of
thermoelectric materials and discuss why some of
the strategies worked well and what some of the
strategies needed to be improved to reach the state of
the art we have today. We’ll then move on to
talk about PEDOT:PSS, so material that we’ve
seen numerous times, both on OLEDs and OPVs’
the whole conducting layer and how PEDOT:PSS is really
pushing the front in terms of the p-type leg of polymer
thermoelectric devices as well. And then we’ll touch on how
we can use both polymers and inorganic nanoparticles,
and the composite of those two materials to make
high performance thermoelectric devices as well. So we’ll go into hybrid systems. So by the end of the lecture
today, what we’ll hope to see, is that we’ll be able to
name at least three materials that have been of
high utility we’ll say in polymer thermoelectric
devices. And really what we hope
we get out of this is that we can explain why
doping matters so much in polymer thermoelectrics. We haven’t talked a whole a lot about doping throughout the
other device applications because we’re going to
really keen on looking at the semiconducting
properties. Now we want to push those
semiconducting properties to the limit and figure
out how we can maximize that electrical conductivity
so that we can maximize ZT and really doping I’m using
molecular dopants is a very powerful way to do that. And then the last that we like
to be able to do is justify why one might move to inorganic,
organic, so inorganic polymer, hybrid thermoelectric systems. And there’re some very good
reasons why won’t we want to do that, and we’ll touch on
those with two example cases. So let’s get started. [Slide 3] And before we move
into polymers, I’d first like to give the
feel for the order of magnitude of materials we’re
working with here. So here we have a plot
of ZT versus temperature, for a whole suite of
inorganic systems. And this goes anywhere from
very, very low temperature, so 100K all the way
up to 1400 Kelvin. And really I’ll remind
that for our polymers, we’re going to really
want to work in this very low
temperature array so almost absolute
zero, up to about 500K. At that point our polymers if
they haven’t started to flow or melt at that temperature,
we’re definitely going to start to think
about the grading between the carbon-carbon
bonds. So we’re going to be operating
in this low temperature regime. So we can see in this
low temperature regime, really the figure of merit ZT for these materials
is somewhere, and these are the best
inorganic materials we can find. For these materials is
somewhere between 0.8 and 1, is what ZT maxes out to be. So with that in mind,
when we look at our polymer thermoelectric
materials and we look at the ZT values
that we can calculate for our polymer materials, and we measure our ZT
values experimentally for our polymer thermoelectric
devices, we really need to keep the benchmark of the
best materials in the world, is between 0.8 and 1. [Slide 4] And when polymer thermoelectric
material started out, this is roughly in
the early ’90s. The first material
that was really looked at in depth was polyacetylene,
or PA. And here we have
polyacetylene, and we can see that we haven’t doped
it, but if you look at the electrical conductivity
of polyacetylene, is somewhere on the order of 6,000 inverse
ohms inverse centimeters. Its Seebeck coefficient is
around 21 microvolts per Kelvin, and that give it a power factor
of somewhere on the order of 10 to the negative 4th watts
per meter Kelvin squared. If you look at polyaniline, you can see that it has
lower performance value so it’s not studied
quite as much, but those are two materials
that kind of were at the outset of this polymer thermoelectrics
idea. Now, one way you can increase
the conductivity when we talk about inorganic system
is to dope it, right? And polymer systems have
a similar terminology and a somewhat similar
mechanism and the fact that they can also be
doped but they’re doped with molecules instead
of individual atoms. And if we doped polyacetylene,
and the first case here was to dope with a FeCl4, we can
see the conductivity goes up, the Seebeck coefficient
remains roughly the same. And that gives us a
higher power factor. And you can see depending
on how you dope it in, who is performing the task that
it varies a little bit but it’s on the order again of 10 to the negative 4th watts
per meter Kelvin squared. So FeCl4 is a decent dopant and increase the
conductivity a little bit. It does have some problems with
respect to long-term stability so what people move to is
doping with iodine. And when dope polyacetylene
with iodine you can see that the conductivity shoots way
up, all the way up to 50,000, so almost an order of
magnitude higher. The Seebeck coefficient
between the very first entry in the table and the one in the bottom is
exactly the same at 20. And that puts it a
power factor of 10 to the negative 3rd watts
per meter Kelvin squared. So now we have this
relatively high power factor. Remember, if we multiply
that by 298 because all of these operate roughly around
room temperature, and divide it by the thermal conductivity,
and the thermal conductivity for polymers are somewhere
in the order of 0.1 or 0.2, then we can back out ZT. But let’s just look at this
power factor right now. We are in the order
of 10 to the minus 2. That’s extremely high. This looks great. So the big question becomes,
why don’t we just have a bunch of polyacetylene
doped with iodine and thermoelectric
modules in place today? These are very high values. And the answer is because
they’re not stable. The iodine comes in vapor form
and it can easily be pulled out of the polymer and will
over time and become undoped, and I’ll reduce back down to a
very low power factor value. So while we need to dope
and when you think about how to dope, we also need to
be careful and make sure that our dopants aren’t
transient with respect to time, that is both the
device won’t break down and our dopant won’t
leave the device. [Slide 5] And the best way to dope
the people found to date without having a
transient materials to have the dopant be
polymeric as well. So now we’re going to
have our PEDOT:PSS. So remember PEDOT:PSS is this
polyethylenedioxythiophene, right? And it’s doped with
polystyrene sulfonate. And people have used
PEDOT:PSS in a quite number of applications including
thermoelectrics, may be used pristine
PEDOT:PSS as well. It performs fairly well. It has a nice figure
of merit. But in recent result, what people have done is
actually oxidize it further. As they do that, they’ve used
an organic compound that’s based with an iron atom in the
center, this iron oxalate. And they’ve used this to oxidize
the PEDOT:PSS beyond what it is straight out of the bottle. And when you do this,
you can actually look at how the device performance
changes as a function of oxidation level. And if we look here in
the close triangles, we can see that initially
the Seebeck coefficient is relatively high and it begins to
decrease, begins to decrease with the increasing
oxidation level. It goes somewhere from 800
down to roughly 100. The conductivity on the
other hand shoots up from 10 to negative 3 centimeter
or siemens per centimeter, all the way up to over 100. And because of these
two competing effect, just like we see a lot of times
in the inorganic systems, we see a maximum in
S squared sigma. And here it occurs at the
oxidation level of around 22 and a half percent. So now the question becomes,
how does this impact ZT? Well, it turns out that this
oxidation doesn’t necessarily affect the thermal conductivity. So then the plots of ZT
versus oxidation percent, looks very similar to the
numerator of the ZT values, so the power factor. And what this tells us
that our temperature, we see that the ZT value for this doped PEDOT:PSS
is roughly 0.25. And that’s not bad, right? Because remember, the best
performing inorganic materials at room temperature are
somewhere in the order of 0.8 to 1, so we’re a
quarter of the way there through a simple
doping strategy. And because of this, now we can
think that if we’re a quarter to the way there, we can really
go back to what we talked about in the last
lecture and think about how much do all
the materials cost and which way is the
more efficient way to produce these materials
in a large scale. So that’s great. [Slide 6] However, we can take a
one step further. And if we take it
one step further, we can even increase the
thermoelectric performance of PEDOT:PSS even more. And the way we do that is
by removing some of the PSS. Now, you remember PSS is that
polystyrene sulfonate. It’s electrically insulating,
the reason we used it is because we polymerize EDOT
to make PEDOT in the presence of polystyrene sulfonate. But once we’ve polymerized the
EDOT, have it soluble solution to make our thin film,
the PSS is just kind of endothelin film not doing a
whole lot for charge transport. So what people have
been able to do is actually dip the thin film
in ethylene glycol or EG, and that’s what EG treatment
means, for various amounts of times, to remove
some of that PSS. And you can see when
they do that, that the Seebeck coefficient
of PEDOT:PSS increases. The conductivity
increases substantially. The thermal conductivity
doesn’t really change, it stays at roughly 0.28. So that tells us that our
ZT will actually increase in the same manner as the
power factor increases. And when that happens
we see that we could get up to 0.42 ZT at
room temperature. So now we’ve almost doubled
what we did last time and we’re basically
halfway to being where the best inorganic
systems, which has been studied for much longer than
polymer systems are. So now we’re starting to really
be able to play an optimization of PEDOT:PSS, could be
an amazing way forward for polymer thermoelectric
materials. Now you recall, our thermoelectric devices
have two legs, they have both of p-type leg and an
n-type leg. [Slide 7] And our p-type leg might
be very well be PEDOT:PSS but you’ll recall from
our organic photovoltaics and our organic light-emitting
diodes, the PEDOT:PSS is whole
selective. It really likes to p-type, but we need an n-type
leg as well. So, discovering new
materials, new polymers that are preferentially electron
transporters is also very important in terms of thermoelectric
materials and devices. And that work is just
starting to get on going over the last two
to three years. And one of the initial
high performance materials that have been found, is this bulky conjugated
polymer here where we have this
electron donor, electron acceptor structure
that we might remember from our low band gap
polymer discussions. And it turns out that if
you have this material, and you add in dopants,
either DMBI, the upper one or one or DPBI, the
lower one. And then if it’s between
these it’s either methyl group or the phenethyl group on these
materials, we actually see that with adding dopants we
can increase the conductivity of this material. And what we see is that relatively low molecular
doping concentrations between 5 and 25 percent, we actually see
a 3 order of magnitude increase in the conductivity of
these materials. You’ll note the conductivity
is 10 to the minus 3 here at the maximum, siemen
per centimeter. So this is roughly 6
orders of magnitude less than what we were seeing in
the PEDOT:PSS, but you have to remember this is a much
earlier generation of material, much less well studied
than PEDOT:PSS. So hopefully we’ll
be able to take it up to a higher level soon,
but the interesting thing is that the Seebeck coefficient
is very, very high. It’s roughly 850. And you’ll note that when
we have an n-type material, the Seebeck coefficient
is negative. And when we have
the p-type material, the Seebeck coefficient
is positive. And that’s by definition,
that’s by nomenclature, but remember S is always
squared in the power of factor in the ZT terms, so the
sign doesn’t matter in terms of our device performance it’s
just let’s us know whether it’s p-type or n-type. But here we see this
is roughly 850 so then we get power factors
now that are on the order of microwatts per meter
kelvin squared. So we’re at, for n-type
materials at that level that we were for p-type
materials, roughly ten years ago. So we’re getting there, it’ll
just take more research, more material design on how
to get to the higher level. [Slide 8] And in fact, an exciting thing
to do is to take the material and instead of doping it
with the external dopants, is to have internal dopants. So here we have our n-type
polymer but now we’re going to dope it with a material
that’s actually covalently bound to it. So now we have an
intramolecular doping. And it turns out how you dope
these things and what the length of this N the space or group,
how many CH2 groups I put between the main chain, and the side chain,
it matters in terms of its thermoelectric response. And what we see is that we
have a space or length of two, four, six the Seebeck coefficient
stays relatively flat around negative 200 but the
electrical conductivity shoots up, and it approaches
one siemen per centimeter. And if that’s the case, then we
can see there are power factor, is roughly one microwatt
per meter Kelvin squared, and really that’s one of the
highest performing devices for n-type materials
that has ever been seen for organic thermoelectric
materials. So we’re getting there, n-type
is not as well developed as p-type materials,
but it’s coming closer and closer each day and really
the design and synthesis and implementation of these
new materials will help drive thermoelectric devices and
polymer thermoelectric devices in particular to the next level. Well, last thing
we’ll discuss, is the idea of hybrid
inorganic polymer systems. [Slide 9] And we saw on the first slide that that bismuth
telluride was one of the highest performing
p-type inorganic materials. So maybe rational strategy would
be to combine bismuth telluride which performs well inorganic
material, PEDOT:PSS, one of our highest performing
polymer thermoelectric materials for p-types, and combine
them to see if we can get to some kind of synergy,. And the answer is yes. And people have done this
and they can make nice sheets of bismuth telluride,
right here, that are roughly 500
nanometers in width. And when they do
this, what they see is as they load more bismuth
telluride nano sheets in there, they see an
increase and then a leveling of the electrical conductivity. You can see this is very high, roughly 1200 siemens
per centimeter. And we also see that the
Seebeck coefficient goes ahead and it increases as well but it
stays within this modest range of ten to eighteen, what we’re used to
see for polymer semiconductors. So what happens is, is that the
power factor goes all the way high as 35 microwatts per
meter kelvin squared. So now we’re getting
the benefits of both the inorganic material
and the polymer material, while retaining, while retaining the thermal
conductivity of polymer which is very, very low,
which is great for ZT, because we want to keep
that thermogradient high between T hot and T cold. [Slide 10] We can also think
about doing this with PEDOT:PSS and tellurium. So what happen here is actually
tellurium nanowire formed in the presence of PEDOT:PSS, to form these nice tellurium
nanowires that have a sheet of PEDOT:PSS around them. And now you have inorganic
nanorod with a layer of polymer wrapped
around it. And the question is, does this
perform the same as the polymer, the same as the nanorod, or
better or worse than either one? And the answer is better. So here we have our PEDOT:PSS
its conductivity is roughly one, the T nanowires
are roughly in the order of magnitude to– in
order of magnitude lower, but in the PEDOT:PSS
with tellurium, jumps up by a factor of 20. We’ll see that the Seebeck
coefficient increases drastically, what leads us
to is ZT of around 0.1. So this isn’t quite as
high as pristine PEDOT, at least modified PEDOT:PSS. But this is pristine
PEDOT wrapped with tellurium around it. So the idea then is, is can
we now take our PEDOT:PSS with tellurium nanowires
around it and do those same chemical
modifications we saw earlier to get the ZT of 0.42. Can we do those same
chemical treatments to increase this
nanorod system. Because this material right here
is just a solution process just as cheap as PEDOT:PSS
itself. And really moving forward, inorganic polymer
hybrids might end up being the best
of both worlds. [Slide 11] So with that, what I’ve hoped
to show you today is the fact that polymer thermoelectric
materials can really play in a low temperature
regime with respect to both device performance
and always with this underlying thought of
cost of fabrication and the cost of materials that go into the
thermoelectric devices. And then not only
can we use polymers but if we design
inorganic nanoparticles or nanorods the appropriate
way polymer inorganic composites
could play a huge role in order to have a synergy that goes
beyond the thermoelectric performance of either one of
the individual components. So the prospects for
thermoelectric materials and polymer-based thermoelectric
devices is very bright, we need to discover new
materials, we need to work on n-type materials,
but the ground works there, the theories there, and
the limitations associated with inorganic materials
aren’t quite as obvious, so that could be a great
way to move forward with these new materials
and these new devices. In our next lecture, we’ll
review what we’ve talked about throughout the course of
this series, and we’ll talk about how all of these
concepts pull back together and really lets us
go from molecules to modules inorganic
electronic devices. With that, I thank you for your
attention, and I look forward to seeing you in
the next installment of organic electronic devices.

You Might Also Like

No Comments

Leave a Reply