Climate Challenge Workshop 2021 - Prof. Sir Colin Humphreys
Well, thank you very much, and it's a very great pleasure to be here, and I'll now try
to put the most difficult part of the talk on, that's uploading my slides, so let me
do that.
Fine.
So can everybody hear me?
Yes?
Good.
Okay, fine.
Thank you.
So it's a great pleasure to be here and to talk at this climate change workshop on delivering
a battery revolution.
So I'd like to start by talking about predicting battery failure, and let me start by telling
you something you already know, and that is batteries are critically important.
So this research institute in Indonesia is just so important.
It's important today, but it's also important for the foreseeable future, because we're
going to need batteries and battery improvements for many, many years to come.
So as you all know, if the sun's not shining, solar panels are useless, right?
They're only useful if the sun's not shining, if you want batteries to store the energy
they generate when the sun is shining.
And of course, with electric vehicles, then if you want to replace diesel and petroelectric
vehicles, then batteries, again, are critically important.
So I'd like just to start by saying the area that you're all working in is a really critically
important area, not only for Indonesia, but for the whole world.
So congratulations on setting up this National Battery Institute in Indonesia.
I'm going to talk about battery failure, because it's important to be able to predict battery
failure in, for example, buses and trains.
So for example, imagine you're traveling along in the countryside of Indonesia in a battery-powered
bus and the battery fails.
And you're miles from anywhere, and it's a catastrophe for the passengers.
And so what we need to have is a system which gives you advance warning of failure.
We need a system in our batteries which will tell you in the next hundred miles the battery
may fail, but it gives you a warning so that before you set a bus out from the bus station
in the morning, the battery's been changed to a battery which isn't going to fail.
So there's a real need for in-situ, real-time monitoring of battery performance.
And I know we have battery maintenance systems that are present in batteries, but there's
a real need for a more sensitive detection system than we have at present.
So we need to detect the first signs of failure in the individual cells in batteries.
And also, for battery development work, we need to pinpoint the place where that battery
starts to fail.
So there is a need for more sensitive battery sensors.
So how do we detect the first signs of battery failure?
Well, as you all know, when we're using a battery, current flows through it.
And when we're charging a battery, current flows through it.
And if we've got a defect in the cell, for example, in a solid-state electrolyte, for
example, then if there's a micro-crack in there, which will develop maybe into a full
crack, if there's an inclusion in the material, if there's precipitates in the material, if
there are dislocations, as there will be in the material, then this changes locally the
current which flows at that point.
So, for example, if you've got a micro-crack, this locally stops the current flowing through.
And when we have a current in a wire, there's a magnetic field surrounding that.
And so whenever we have a current, there's a magnetic field.
And if the current locally changes, there's a local change in the magnetic field.
And so the idea which we have is to use an array of small magnetic sensors on the cell
to detect the position of defects in the cell by detecting tiny changes in the magnetic
field in certain points of the cell.
And sometimes you get hot spots in a cell where the current increases because of defects
at that point, and you can get local heating and the cell can catch fire.
So this is also a way of eliminating fires in batteries.
So what we need is a super-sensitive magnetic field or current sensor.
And so we asked ourselves, can we make a super-sensitive graphene sensor?
So let me now say a few words about graphene.
Let me just introduce graphene.
So as many of you will know, in 2010, two Russian scientists working at the University
of Manchester in England won the Nobel Prize.
And the citation was, a groundbreaking experiment regarding the two-dimensional battery material,
sorry, regarding the two-dimensional material graphene.
And it's quite unusual for Nobel Prizes to be awarded to a material.
But this is one of these rare cases.
And graphene is a two-dimensional network of carbon atoms, as I'm sure you all know.
It's often called carbon chicken wire.
I'll give an illustration in the next slide saying why it's called carbon chicken wire.
It is the first material in which all the atoms are in a single plane.
And because the carbon atom is very small, it is the thinnest material in the world.
So it's a remarkable material.
So this is a diagram on the left-hand side of what graphene looks like.
And each dot there represents a carbon atom.
And you see it looks a bit like chicken wire.
And if you pile up layer upon layer of graphene, you get graphite.
And graphite's a mineral.
You can dig it up from the earth.
And in England, we have graphite mines where you can dig up the graphite.
And I'm sure you do in Indonesia as well.
So it's quite a common mineral.
And how do you get graphene from graphite?
Well, in 2004, in Manchester, these two people got Nobel Prizes.
They isolated graphene for the first time.
And what they did, they dug up a block of graphite.
And they put sticky tape on both sides of the block.
And then they pulled the sticky tape apart.
And they got the graphite divided into half, roughly.
And then they put another piece of sticky tape on.
And they pulled this sticky tape and put some more on, pulled it in half,
stuck it on, pulled it in half, until they got a thinner and thinner and thinner
piece of graphite, until they were down to a single layer of graphite,
which is called graphene.
And they got the Nobel Prize for that.
So in a sense, they got the Nobel Prize for using sticky tape.
And what I'd like to say is, most of the simple science experiments
in the world have already been done.
But not all of them.
And this is a really simple experiment which anyone could have conceived.
A schoolchild could have conceived of this and gone ahead and got a Nobel Prize.
So don't ignore simple ideas.
It's really important.
And you might get the Nobel Prize for that.
So it turned out that this single flake of graphene had remarkable properties.
If it had not had remarkable properties, no Nobel Prize.
Everyone would have forgotten about this experiment.
And so it had remarkable properties, which I'll show in the next slide.
But graphene plates and graphene powders are used for adding to car tires,
paint concrete, running shoes, and so on.
And it's useful for these lower tech applications.
But if we want electronic devices like transistors, if you want sensors,
you need larger area graphene that you can get by the sticky tape method.
And so that's what's needed.
We need to be able to scale up this method.
So these are the properties of these exfoliated graphene plates,
which you get by sticky tape method.
And now what you can do instead of a sticky tape method, of course,
you can, for example, ball mill the graphite and get graphene plates and powders from that.
But a single flake of graphene has remarkable properties,
very high electrical conductivity, a thousand times better than copper.
And I think it's the best, the highest in the world.
And that promises very high speed electronics.
And because it has high electrical conductivity, it has very low resistivity.
And that promises very low power consumption from your devices.
It's also flexible, a single flake of graphene.
That promises flexible electronics.
It has very high thermal conductivity.
And that's useful for improved device performance.
So if you look at high power semiconductor devices,
you can't actually operate them at the moment at their maximum capability
because you can't get rid of the heat generated quickly enough.
So for example, gallium nitride power electronics,
which I've been involved in in the past,
the problem, the limiting factor there is the heat dissipation.
Because it is a poor conductor.
So that's a limiting.
Whereas if you can have graphene electronic devices,
because of the conductor thermally,
you can dissipate the heat and get much better performance
than you have at the moment, in theory at least.
And also, we have high optical transparency from a single layer of graphene.
Over 97% transparent.
And this is useful because it promises to replace transparent conductors
in, for example, our mobile phones.
Because indium tin oxide is used there,
and indium is on the world's list of critical materials.
Now, what is remarkable is all these world-beating properties
are in a single material.
And that's what's remarkable about graphene.
And that's why graphene has been called the wonder material.
So here's graphene, the wonder material.
And at the top of the picture,
you kind of see a block of graphite and then a single layer of graphene,
more conductive copper, more transparent indium tin oxide,
all these wonderful properties.
And so that's attracted, of course, attention throughout the world.
And companies like Intel have invested huge sums of money
in trying to scale up graphene and make useful devices like transistors.
So Intel have invested $750 million US dollars,
IBM $3 billion US dollars,
Samsung $2.2 billion US dollars.
But where are all these predicted wonder electronic products?
And so far, they have not existed.
The problem has been in trying to scale up
from a small phase to something you can manufacture.
And this is just some newspaper headlines on this slide.
So here you see, this is a newspaper from the World Economic Forum at Davos.
And they said, this is some years ago,
graphene could soon make your computer a thousand times faster.
Do we have graphene computers a thousand times faster?
No, we don't.
So this prediction has not happened.
Bosch, another newspaper,
Bosch finds graphene magnetic sensors a hundred times more sensitive than silicon.
I'll come back to that in the next slide.
This third newspaper,
graphene filters change the economics of water.
Has it happened?
Do we have graphene filters in Indonesia to get pure water?
I don't think so.
So this hasn't happened.
So there's been all these predictions made
based on graphene results in the laboratory.
But the challenge is to scale up these results to commercial scale,
where you can actually have computers a thousand times faster.
So that is the graphene challenge.
And this shows a transistor made by IBM.
And it's assembled by hand.
So you can't manufacture like this.
It's assembled by hand using graphene plates.
And it was 10 times faster than silicon.
Well, this is a transparent contact made by Samsung.
You can't actually scale it up.
And this is a graphene hall sensor made by Bosch.
And they found it was a hundred times more sensitive than silicon.
Just made by hand from tiny graphene plates.
And Bosch issued a press release about this in 2015.
And the press release says,
Bosch creates graphene sensor 100 times as sensitive to silicon,
but unlikely to compete for 5 to 10 years
due to the current lack of large-scale,
wafer-based and transfer-free synthesis techniques.
And at the time this press release was issued,
there was a larger scale method of making graphene.
Called CVD, chemical vapor deposition.
But it was not suitable for electronic devices
because you deposit gases like methane.
And they're decomposed on a copper substrate.
And the copper acts as a catalyst to decompose the gases.
And the copper gets embedded some impurities into the graphene.
And you don't want copper impurities in your graphene for electronic devices.
So it's difficult to use that existing method of large-scale graphene
for using in electronic devices because of contamination.
Then you also have to transfer the graphene from the copper.
So this is where I got involved.
So this is where a personal story of mine starts.
So I had two very good postdocs when I was working at Cambridge.
I've now moved from Cambridge to Queen Mary University of London.
But this work initiated in Cambridge.
Now we continue it at Queen Mary.
Two postdocs.
And they solved some key problems.
And what they used was a new technique for making graphene called MOCVD.
Metal Organic Chemical Vapor Deposition.
This technique has been well known for other materials.
But no one has thought of doing it for graphene.
And so this was the first time.
And they got it to work.
And it was patented in 2015.
And there's no metallic contamination because you're going to use a metal catalyst.
There's no transfer process because we deposit directly on the substrate we want to use.
And it is very promising.
The next generation of transistors and so on.
So we set up a company to exploit this research.
Raised SEAS funding in October 27.
Moved into premises in 2018.
And we now have over 60 employees.
So it's a rapidly expanding company.
And these are the premises we move into.
And I'd just like to tell you about this.
Because a number of you who are listening may want to start up companies.
So let me tell you something about how we started up this company.
Because it may be very relevant.
So the obvious place, if we're in Cambridge, was to go to the Cambridge Science Park.
But it's very popular and very expensive.
And we thought we're not going to pay these large sums of money for going to Cambridge Science Park.
So we looked around.
And there was a farmer in Cambridge.
An innovative farmer.
And the farmer put up large sheds on his fields.
And he called it an industrial park.
And these large sheds were totally empty.
So this is one of the large sheds that we moved into.
And it didn't have windows at all.
There were no windows.
It only had a door.
Absolutely empty.
No electricity.
No water.
It had a hole in the roof, which we discovered when we moved in in February 2018.
Because it snowed.
So snow came into this empty shed.
And we got this extremely cheaply.
We still have it extremely cheaply.
And we brought in electricity from a substation outside.
And we brought in water as well.
So we just set up.
And very, very cheaply.
So you can set up a company in the UK and also in Indonesia really cheaply.
And then you can use a shed.
You can put up a really nice building inside it.
Actually, you can put in clean rooms, which we've done in this shed and so on.
So this is a shed on a farmer's field.
And it works very well for us.
And we've installed.
We've got these windows installed.
We've put in a floor.
So it's a ground floor, an upper floor, and so on.
So that's some advice if you want to set up a company.
It doesn't need to cost too much.
And graphene has no metallic contamination.
No chemical problems.
Large area.
100% substrate coverage.
We can dope it when we grow.
Suitable high-volume device production.
And Tuta will be home for making this for devices and for sensors.
Most of you, I'm sure, will know what a Hall Effect is.
So I don't dwell on this.
But basically, if you have any material, in fact, in a magnetic field coming up vertically
through this material.
And if you apply a voltage at right angles, this magnetic field is voltage at right angles,
then a current flows through this material.
And the magnetic field deflects the current carriers, whether they're electrons or they're
diodes, it deflects them by something called a Lorentz force.
And then that sets up a voltage in this third perpendicular direction.
And if you know what the applied voltage is, as you do, and you measure this voltage in
this direction, called the Hall Voltage, then you can determine what the magnetic field
is.
So that's how it works with a sensor.
And I'm sure you all know that already.
And you can also, if you know, you can also use this to measure current, because this
is a current carrying wire, there's a magnetic field around it, and you measure a magnetic
field with this Hall Effect sensor.
And let's just show you some sensitivities.
Let's show you just how good graphene is.
So the standard Hall sensor is made from silicon.
It has a sensitivity of 70.
The units are volts per amp tesla.
Compound semiconductors like gallium arsenide, quantum wells, they're expensive, but they
have a sensitivity much better, which is 850.
Bosch's graphene flake, sensitivity, 1250.
Paragraph graphene, 2000.
So 30 times better than silicon.
Minimum operating temperature for our Hall sensor, less than 100 millikelvin.
For silicon, the minimum operating temperature is 230 kelvin.
So graphene has remarkable properties.
The only Hall sensor which operates at cryogenic temperatures.
And this was a surprise to us.
We didn't expect this, but when we got it, we thought, well, of course, it will work
at these very low temperatures.
Maximum operating temperature is limited by the packaging being used at the moment.
Radiation resistant.
Works in very high magnetic fields.
It works from room temperature, works from zero field, up to 30 tesla, huge field.
And typical Hall sensors just work up to one tesla.
Very low power consumption.
Operates in harsh environments.
And this shows you schematically what graphene Hall sensor is like compared with other Hall
sensors.
So this is a standard Hall sensor in this column.
This is a specialized Hall sensor like made from gallium arsenide.
This is not a Hall sensor, a magnetoresistive sensor for detecting magnetic fields.
They're also flux based.
And NMR, nuclear magnetic resonance, can detect Hall or can detect magnetic fields.
If we look at the sensitivity, then the paragraph Hall sensor is more sensitive than a normal
Hall sensor, a specialized one, a magnetoresistive one.
It's not quite as sensitive as flux based, not quite as sensitive as NMR, but the sensitivity
is very high.
For looking at the magnetic field range, it surpasses anything else that exists.
Nothing else can go from zero to 30 tesla.
There's no planar Hall effect.
It operates at very low temperatures.
Lots of ticks.
Operates at high frequency.
Lots of ticks.
Low power.
Lots of ticks.
So if you add up these ticks, the sort of figure of merit, then overall, this paragraph
graphene Hall sensor is the best magnetic detector that there is.
And let me just speed up now.
I'm always at the end.
We're working with CERN to see if we can use in their upgrade with the Large Hadron Collider.
We're looking to see if we can use in satellites.
But here we come to the sensors and batteries.
So we're looking with a major manufacturer of batteries to see if it can be used as a
sensor to predict and detect failure at an early stage.
So the idea, as I said, right at the start is you have a battery, which has a number
of cells, and you have an array of tiny graphene magnetic sensors.
And you put these on the cells and you can detect tiny changes in the magnetic current.
And this talk, I know it's really up to date talk.
So we're working with this company.
We do not know yet if it's going to work.
We believe it's going to work.
We believe it will give you early warning of a failure of batteries.
But if it does, we think it's really important.
So this is very advanced information of something which we hope is going to happen.
So let me now go to the final slide.
Conclusions.
So we've got a new way to make graphene.
It's the best performing full sensor in the world.
It raises expectations that graphene could be used for biosensors and transistors.
And it raises expectations that it may be able to detect the first signs of failure
in batteries.
So thank you very much for listening.
It's a pleasure to talk to you.
Thank you.
OK, this is a slide I wasn't intending to use.
I'll just tell you what this is, though.
This was a slide after the talk, and it's gone automatically onto this.
These are high-resolution electron microscope images, really staggering ones.
And on the left-hand side, you can see the individual atoms in graphene.
And it's doped with silicon.
And you can see where the silicon dopant atom has gone, a larger atom.
And it actually goes in two different places in the graphene lattice.
It can go there, and it can go there.
So this is, I wasn't going to talk about this.
But this shows the power of electron microscopy.
And you need these sorts of characterization techniques to look at the atomic structure
of the material in detail.
So let me now stop showing my screen.
OK, thank you, Prof. Colin, Sir Colin.
And now I will give the floor for those who have a question.
But first of all, Sir Colin, it is indeed an honor for us to have you in our workshop.
So thank you so much.
Well, thank you.
And it's an honor for me to be here as well and to be part of this workshop.
And your vision for the future, which your vision is absolutely correct,
batteries are just so important.
So it's a great pleasure for me to be here as well.
So thank you very much.
So if you don't mind, I have a question, Colin.
Although I could ask you this.
What sort of sensitivities do you expect to need in order for you to be able to do this
future prediction of batteries?
And how does that relate to the sensitivity of your Hall probes for measuring current?
Right, so I'm bad at remembering figures.
But what we have, we have, what we know is that we have a Hall sensor, which is
more sensitive than other Hall sensors.
It's not as sensitive as nuclear magnetic resonance.
But there's no way you'll fit an NMR machine in a battery because they're huge, right?
It's not as sensitive as flux state measurements.
And those are huge bits of kit.
So we believe we've got, apart from those two, which are big pieces of kit, expensive
piece of kit, we have the most sensitive, we believe, sensor, Hall sensor there is.
And so we believe that it gives us the best opportunity, the best chance to measure these
tiny changes in magnetic field you get when coming close to a battery.
There's an imperfection in the battery material, which can then lead to failure in the future.
And we think we can then detect micro cracks before they develop into major cracks.
We can develop inclusions in material, impurities in material, which shouldn't be there.
And so enabling the battery manufacturer to develop better materials.
But the idea is to give you warning that in the next 100 miles, your battery may fail.
I think I can envision and visit a proposal here where you do in situ, in operando tomography
and have your Hall sensor running at the same time
so that we can actually purposefully break a cell slowly to try and simulate a cell that's
dying, you know, and then you can do three dimensional imaging of it at the same time as
looking at your Hall sensor at the same time as knowing what it is that you're doing
to the cells because you're purposefully breaking it.
That might be quite interesting.
That's a great idea. This is how science works, right? I give a talk,
Alan comes along, has a great idea, and then we set up a grant proposal to do this.
So that's a great idea, Alan. So it's brilliant. Thank you.
We have to think about this. I think a conversation when we're both able to meet
up in Queen Mary, we have to finish another grant first. But that's another story.
Right.
I'll put it over to anybody else that might have a...
May I have a question, Alan?
Please, please.
So thank you for calling. Thank you for inspiring lecture from you about the graphene.
Question is related to several study has shown that the graphene can increase the specific
capacity of the battery up to 23%. So how do you think about this and then how it works?
Because it will be very, very few chairs for the graphene to be added to the cathode material,
since they will not increase the volume, but the specific capacity.
Okay. Thank you.
No, no, thank you. That's a very good question. So I think these are the flakes and powders of
graphene which are added. And so I'm not an expert in that area. So I've been working on
large area graphene. I've also only moved into graphene relatively recently. So before I was
doing a lot of work on gallium nitride. But I think when the graphene is added,
a very good question is the mechanism of how it works. So let me give a different example as well.
So for example, when graphene is added to car tires, it makes them last longer. And you say,
how does that work? And the belief is with car tires, that because graphene is such a good
thermal conductor, it makes them cooler when they run. So when you drive your car, if you've driven,
you know, 100 miles or kilometers, I should say 100 kilometers, Britain still uses miles often,
the 100 kilometers, then the tires are very, very hot. And they wear faster when they're very,
very hot. And so to have the graphene there, it conducts electricity away. So it reduces
the temperature. And that might be one possible effect when you're operating a battery that it
keeps the electrodes cooler. But there are other possible things that can happen.
And graphene sounds a very simple material, it sounds the simplest possible material.
In practice, it's very complicated. One complication is it interacts with the atmosphere
a lot. And so it absorbs water, it reacts with water, it reacts with oxygen, and those actually
don't the graphene. So when you add the graphene plates and powders to an electrode material,
then those materials themselves will be doping the graphene and this may be having an effect
on the electrical performance of the device. So it really is quite complicated. And you know,
we don't understand these things very well. So although you think graphene is very simple,
it is not so simple in practice. And it's a fascinating material to you. So it's a great
thing for you to do research on, you know, just the mechanism of why this is improving performance.
Thank you. Thank you, Prof. Molin. Yes, for the explanation.
Okay, we have another...
Just one question. Prof. Molin, I think you have a nice presentation. It's quite inspiring.
I want to ask about how to get the single-layer graphene. Because it's quite difficult to have
single-layer graphene. Because I've got the graphene, it's more agglomerate. We couldn't
get the single-layer graphene. Is there any method to get the single-layer of graphene?
Right, so what you see what... The single-layer graphene is very, very tiny.
So we have agglomerate of graphene, but we need the single-layer graphene. But it's quite
difficult to have the single-layer graphene. What's the method to have the single-layer graphene?
Right, so that's a very, very good question. So what we thought of, we used this new technique
which people haven't used before for graphene at all, called MOCVD. MOCVD, Metal Organic Chemical
Vapor Deposition, is used for manufacturing semiconductors. And in semiconductors, you grow
them layer by layer. So you actually grow single layer at a time. So when you grow a gallium
nitride transistor or an LED, which has a quantum well in it, which is just a few single layers,
they're actually growing layer by layer. And so we thought, can we grow graphene in a single
layer? And this is what happens. So you have, say we're growing... Let me just describe what
happens. You have graphene grown on a substrate, say sapphire we use quite a lot, and the chemical
gases come down, and the substrate is at high temperature, about a thousand degrees we're using.
And the gases come down, and when they hit the substrate, the gases decompose. And things like
hydrogen come off and go away. And the carbon is on the surface, and it scoots around on the surface
until one carbon atom finds another carbon atom, then they find another carbon atom. And the
bonding between them is very strong. It's called sp2 bonding. And they bond up and they form a
single layer. And in fact, that growth method, they like to form a single layer because this
bonding is very, very strong. So you've got a strong, strong single layer. And if you keep on
growing, they'll form another layer on top. But that other layer on top is van der Waals bonded
to the layer underneath, so it's a weak bonding. So it actually likes to form a single layer. So
that's how we get a single layer. The growth method we use, which is important. Yes, it's
very important. Because I couldn't get any single layer. It's difficult to have a single layer.
It is difficult, but you have to think about it. It's just difficult. I agree. Yes.
Thank you, Professor Colin. It's very interesting lecture this afternoon.
Oh, thank you very much. Pleasure. Thank you. Professor Colin, we have another question
that's sent through the checkbox. The question from Mrs. Rika Iudianti. She said,
thank you, Professor Colin, for your presentation. It's very inspiring and interesting.
You talk about graphene. Could you have an opinion on CNT as a sensor? Because CNT also
has a sensitive surface for sensors. What are the advantages and disadvantages between CNT
and graphene as the sensor? Right. So CNT carbon nanotubes.
So a carbon nanotube, that's right, is really a rolled up piece of graphene.
And if you want to, I think you can use carbon nanotubes as sensors. I think graphene is easier
to use as a sensor because it's a planar material. So if you want to have, well, there's a number of
points, right? So to start with, if you want to make a hall sensor, then the easiest way to make
a hall sensor is to have a piece of graphene, which may be 50 millimeters in diameter or more.
And then you can use a laser beam to etch out on the surface these pieces of graphene in a sort of
cross shape. And you can just easily manufacture like that. Now, it's very hard to manufacture
using carbon nanotubes, right? And that's why carbon nanotubes have not been that widely used
so far. You can add them. You can add carbon nanotubes to composite materials and improve
the performance of composite materials, but to actually manufacture sensors from carbon nanotubes
is very, very difficult. It's much easier to use graphene to manufacture sensors. I hope that helps.

