#270      30 min 46 sec
Useful imperfections: Nanodiamonds for quantum sensors in living cells

Physicist Prof Lloyd Hollenberg explains how quantum technologies are leading to the development of sensors of only a few atoms’ size -- small enough to be placed inside living cells to enable monitoring of biological processes. Presented by Dr Shane Huntington.

"Standard sensors are actually quite large. In our case, our sensor is literally a single atom and we exploit its quantum properties in new ways to sense its surroundings." -- Prof Lloyd Hollenberg




Prof Lloyd Hollenberg
Prof Lloyd Hollenberg

Lloyd Hollenberg is the Deputy Director of the Centre for Quantum Computation and Communication Technology. He completed his PhD in theoretical particle physics the University of Melbourne in 1989 after which he was awarded a JSPS Fellowship at the KEK accelerator laboratory in Tsukuba, Japan. After his postdoctoral period he returned to the University of Melbourne where he is now a Professor in the School of Physics.  He is a well-known proponent of quantum technology, with broad interests including quantum computers, quantum communication and the development of quantum sensing techniques crossing over to the nano-bio realm.

Publications

Credits

Host: Dr Shane Huntington
Producers: Eric van Bemmel, Dr Dyani Lewis
Audio Engineer: Gavin Nebauer
Voiceover: Nerissa Hannink
Series Creators: Kelvin Param & Eric van Bemmel

View Tags  click a tag to find other episodes associated with it.

 Download mp3 (28.1 MB)

VOICEOVER
Welcome to Up Close, the research talk show from the University of Melbourne, Australia.

SHANE HUNTINGTON
I’m Dr Shane Huntington. Thanks for joining us. Sensors are all around us. We find them in our cars, on the front of our television screens and even in the humble household kettle. Each sensor is designed to determine a particular state of its external environment. Sensors detect engine temperature in our cars, infrared signals from our television remote controls and when the water is boiled so that our kettles shut off safely. All these common sensors are physically quite large, certainly large enough to be seen and touched. But the promise of quantum technologies suggest we'll soon be able to design and build sensors to the size of only a few atoms, small enough to fit inside one of our cells. Will these tiny sensors work in the same way as their larger cousins? What new insights can they give us about the inner workings of our bodies, and what influence would the strange effects of quantum mechanics have on the way these sensors do their job? To answer these questions and help us explore the world of quantum sensors we are joined by physicist Professor Lloyd Hollenberg, Deputy Director of the Australian Research Council Centre of Excellence for Quantum Computation and Communication Technology in the School of Physics at the University of Melbourne. Welcome to Up Close, Lloyd.

LLOYD HOLLENBERG
Hi Shane.

SHANE HUNTINGTON
Now, let's start with the common view of a sensor. I think most people have an idea of what a sensor is. How does this compare to what you're talking about in the quantum sense, so those standard electrical devices, how do they compare to what you're talking about?

LLOYD HOLENBERG
Well, it's really what you said in the introduction. Standard sensors are actually quite large. You might typically measure currents through a circuit as a result of the sensing action. In our case, our sensor is literally a single atom and we exploit its quantum properties in new ways to sense its surroundings.

SHANE HUNTINGTON
When I ask what one of these looks like, when we're talking individual atoms, obviously this doesn't quite work.

LLOYD HOLENBERG
Yes.

SHANE HUNTINGTON
But can you describe one of these sensors?

LLOYD HOLENBERG
Okay, the sensor is contained in a nanodiamond so a nanodiamond is an extremely small particle of diamond, and diamond of course is primarily carbon atoms and occasionally you can get a defect in the lattice. Often these defects which are single atoms essentially, give rise to the various colour hues of diamond. In our case, our defect is a nitrogen atom that's replaced the carbon atom. It just so happens that there's also a little vacancy in the lattice next door to it, so it's essentially replaced two carbon atoms. This little nitrogen vacancy centre, that we call it, an NV centre, is our little sensor. 

SHANE HUNTINGTON
So Lloyd, we take this carbon lattice that we have for diamond, we yank out a carbon atom and we yank out a second carbon atom, which we would rather than leave a hole replace with a single nitrogen atom. What is it about that that allows us to do any sort of sensing? It seems like, just as you say, something would change colour or even perhaps the properties of the material?

LLOYD HOLENBERG
Well, it's a remarkable system. As I said, essentially a single nitrogen atom. So when a nitrogen atom is replacing a single carbon atom in the lattice there's a spare electron. When you have a vacancy next to it there's a few more electrons. It just so happens this system has also captured a further electron so the whole thing is electrically neutral so we call it a nitrogen vacancy minus centre, and it's the atomic properties of that centre that are really unique. So it has quantised energy levels, it has a ground state and excited state. Within that ground state there are magnetic levels as well. So what we can do first of all, if we shine a green laser on this system then we can literally see the light coming back at us as red. So the nitrogen vacancy centre absorbs the green light and then re-emits it as a reddish light, and we can literally see that in a confocal microscope. We would have a substrate with our nanodiamond sitting there and we can just look at these little pinpoints of red light, and that's literally the light coming from the single nitrogen vacancy centre. That light tells us what state it's in, so that's very useful information; not only where it is but what quantum state it's actually in. Then with a microwave antenna we can actually put that atom into quantum superposition, we can control its quantum state and then read it out. 

SHANE HUNTINGTON
Well, you talk about the NV centre sitting within a lattice of carbon atoms, the diamond; how big does that lattice have to be for this to work? Can you just have a few atoms of carbon, and physically what does that mean in terms of the sensor’s size?

LLOYD HOLENBERG
Well, the studies at present show that you need a certain critical mass of lattice around the NV centre to make it stable. So the smallest nanodiamonds that people have found stable NV centres in are about five nanometres across, and that means that if the nitrogen vacancy centre was in the middle you would have a few nanometres either side of it of diamond. That also means then that your stand-off from whatever you're seeing outside the diamond is just a few nanometres. 

SHANE HUNTINGTON
So you've got effectively an atom where, unlike the other atoms necessarily in the diamond structure, you can actually deliberately make this atom sit in a certain energy state. Can you tell us a bit more about what that means, because these energy states are quantised, meaning they can only have specific values, is that right?

LLOYD HOLENBERG
That's right. So when electrons are in atoms they don't exist in any old energy level, they can only exist in certain energies and so it's the same situation here, except we have more electrons. But overall the energy of those electrons can only be in certain values so the lowest energy state is called the ground state, and then there is an excited state and so forth. Now it just so happens that these energy levels also are sensitive in different ways to magnetic fields so the ground state is essentially insensitive to a magnetic field but the immediate excited state is sensitive. So we can use that property as our basic sensing mechanism, but that's where it differs from a normal classical object, that we have quantised energy levels and we can actually put that atom in any of those energy levels that we want. But also, according to quantum mechanics, we can put it into a superposition of energy levels so we can put it into two energy levels at the same time. 

SHANE HUNTINGTON
So this where things get a bit funky, isn’t it?

LLOYD HOLENBERG
Absolutely.

SHANE HUNTINGTON
Because unlike a normal sense of whether it's on or off, or one or zero, you have the potential here for the sensor to be all of those things at the same time and everything in between. What does that mean for the external world to the object? You mentioned you can shine in green light and get out red; that seems fairly binary. But what does it mean when you have this superposition of states to the external world, to the sensor?

LLOYD HOLENBERG
So imagine now that we have our two energy levels and we put the atom into a superposition where it's both at the same time. Now when we shine the green light on our atom we are basically forcing that quantum system to make a decision because it has to tell us one answer. It can't tell us that it's in both at the same time because that's essentially an answer that we can't really process. We can only process whether it's been forced into the lower state or the upper state. We can understand and tell that decision - actually measure it - by looking at the light that comes off the atom. Essentially if the red light is emitted immediately it was forced into the lower state. If there was a slight delay it was forced into the upper state. But before you made the measurement it was actually in both at the same time.

SHANE HUNTINGTON
This sounds very familiar. Is this the famous Schrödinger's cat-type experiment? Is it on the same principle?

LLOYD HOLENBERG
It is. It's related to that.

SHANE HUNTINGTON
Can you explain that Schrödinger's cat experiment? I think it's a good one conceptually for people to try and get their head in quantum physics?

LLOYD HOLENBERG
It's a great thought experiment. Schrödinger proposed the following idea as a consequence of quantum mechanics. He imagined you have a cat in a box and in that box is a little radio isotope source and the radio isotope source is made up of atoms that either have decayed or not decayed. But some of them will be in a quantum superposition perhaps of decayed and not decayed. So if quantum superpositions exist then those atoms will be in those two states at the same time. Now the device is organised so that if a Geiger counter measures a decay then it will trigger a mechanism to release poison gas into the box, killing the cat. But what Schrödinger reasoned was the hierarchy from a quantum superposition would mean that the cat would eventually be in a superposition of alive or dead, because the atom would be in a superposition of decayed and not decayed, which means the Geiger counter would now be in a superposition of click and no click. Then the poison would be in a superposition of released and not released and then the cat ultimately would be in a super position of alive and dead. So that's Schrödinger's cat.

SHANE HUNTINGTON
Ultimately, we kill or bring the cat to life by essentially opening the box…

LLOYD HOLENBERG
Exactly.

SHANE HUNTINGTON
…and forcing one of these states; is that how it works?

LLOYD HOLENBERG
Exactly. So the act of measurement in that system would be opening the box and then because we don't observe cats in two states at the same time of alive and dead, we force that system into one of those possibilities, and that's really what's happening at the quantum level. For a system as large as a cat that probably won't happen, and the reason for that is what's called decoherence. Every time you have a quantum system in a superposition then it's still interacting with the environment, and the environment is essentially composed of many, many atoms and the larger your quantum system is the more interaction you have with the environment. So the environment is essentially acting like a measuring device and will force the system into one state or the other. So something as large as a cat will decohere incredibly quickly. But this is the actually the mechanism that underpins our sensing, that our NV centre, our nitrogen vacancy centre, when we put it into a quantum superposition, that's when it's really vulnerable to fluctuating magnetic fields in its environment. So we can measure how fast it actually loses its quantum superposition over many, many tries at this and that tells us information about the environment so it's very much related to Schrödinger's cat.

SHANE HUNTINGTON
Now this is very interesting because if you look at that nitrogen vacancy centre, so two specific things within the carbon lattice of diamond, it seems when you think about that, that its closest neighbours are just other carbon atoms. So its immediate environment is just other atoms of carbon but you're suggesting there are other fundamental forces - magnetic, perhaps gravitational - others that affect this as well?

LLOYD HOLENBERG
Well, gravitation won't be a factor here because it's so much weaker than anything else around. But the interesting thing is, yes, we are in a solid and this is essentially a single quantum atom in a solid made up of many carbon atoms. But all of the electrons involved in the lattice are busy binding and so their properties are essentially locked down and they have very little effect on our nitrogen vacancy centre. Of course, electrons have a property called spin which is its intrinsic magnetic moment, and you could imagine that with 1023-odd electrons around our little nitrogen vacancy system that there'd be a lot of magnetic noise. But it's not the case because when electrons bind in a diamond lattice they bind as spin up and spin down, and that cancels that effect out totally. So the lattice is actually a very quiet place for our sensor. But there are things that do affect it in the lattice itself. The first one is the fact that we don't always have carbon-12. We might have carbon-13 atoms. So carbon-13 has a nuclear spin whereas carbon-12 doesn't. So in ultra-purified diamond where you only have carbon-12 there are no nuclear spin background magnetic fields. But if you have even one per cent of carbon-13 that presents a decohering source. 

SHANE HUNTINGTON
So each one of these little carbon-13s is effectively like a mini-magnet…

LLOYD HOLENBERG
Exactly.

SHANE HUNTINGTON
…that's sitting nearby.

LLOYD HOLENBERG
Exactly, and it's fluctuating up and down. Because typically we do this at room temperature. The other interesting thing about diamond is that we can measure the quantum coherence of our NV centre at room temperature, whereas in a material like silicon if you were to measure a similar quantum coherence of a single donor atom, in that case phosphorous, you need to cool it down to milliKelvin temperatures. The reason for that is that in silicon you get a lot of lattice vibrations involved and they ultimately tend to swamp the quantum coherence once you've cleaned up all the isotopic impurities there, which would be silicon-29. But in diamond, because it's very hard you don’t tend to get a lot of lattice vibrations, even at room temperature. So this is the other remarkable property of these diamond quantum systems, our NV system in diamond, is that the quantum coherence actually is measurable at room temperature. 

SHANE HUNTINGTON
Just coming back to the Schrödinger's cat idea for a moment, this was simply what was called a Gedankenexperiment or a thought experiment.

LLOYD HOLENBERG
Yes.

SHANE HUNTINGTON
So this was something that was just on paper but what you're talking about is a literal connection between the quantum world and the classical world, an actual experiment, isn't it?

LLOYD HOLENBERG
That's right, and these sorts of experiments have been done for quite a few years now. In the march to develop a quantum computer you need to have such a quantum system that you can put into one or two states or both simultaneously, and that would be your zero and one that is encoded in your system. This has been in atoms, in atom traps, super conducting systems, even phosphorous donors in silicon now. That basic measurement of measuring the quantum superposition and its response to its environment, its decoherence rate, has been done many times now. So it's Schrödinger's cat in miniature.

SHANE HUNTINGTON
You're listening to Up Close. Today we're talking about quantum sensors with Professor Lloyd Hollenberg. I'm Shane Huntington. Lloyd, let's talk a bit now about quantum computing. This is an area that can also use the nitrogen vacancy centres in diamond as what we call a qubit. I'll ask you to explain what that is and most people will have heard of a bit in normal computers. What is a qubit and how is it different?

LLOYD HOLENBERG
Well, the word qubit, it's spelt with a Q not a C, derives simply from quantum bit. So no matter what your quantum system is, if you have two energy levels that are separated you can call one of them a zero and label the other one a one so that you're connecting with the bit representation. Now classically, if this was a transistor and a computer or a processor, then you could only represent zero or one at any one time. But in a quantum bit you can represent that system as a zero and a one at the same time. If you imagine you replicate that representation over many atoms, then you're essentially representing binary numbers but all possible binary numbers over that many bits at the same time. So you have in principle a quantum superposition of all binary numbers existing at the same time. Then if you can organise interactions between your quantum systems, you're essentially performing mathematical operations between those numbers. But you're doing it in one step over all of those numbers at the same time, and that's essentially the paradigm that powers the quantum computer.

SHANE HUNTINGTON
One of the things we've seen with the drive towards building a quantum computer is the incredible requirements on the devices themselves. You know, low temperature, this decoherence problem being particularly problematic. Some of them do try and use the NV centre. Are those problems also an issue for the construction of these sensors or are they in a different field entirely?

LLOYD HOLENBERG
It's interesting. The two fields are related but in the case of quantum computing decoherence is your enemy. So you need to guard your quantum computer against the environment because the environment will try to collapse that large super position over your binary numbers. The only way a quantum computer can work actually is through a lot of redundancy and error correction, which is a huge field in itself. It means that ultimately a quantum computer would be made up of many millions of qubits, and all of that is because eventually the environment tries to interfere with your quantum coherence. Because these quantum systems are so vulnerable and sensitive to quantum decoherence we thought that we could turn that inside out and use that property as the sensing mechanism. So literally knowing how a system decoheres in a particular environment means that when we measure that qubit system in the environment we can infer information about the environment, just by understanding our quantum system very well.

SHANE HUNTINGTON
You've gone to the point now where you're actually starting to test some of these sensors in particular environments. Tell us about what environments you’ve chosen and why.

LLOYD HOLENBERG
Well essentially, we chose what I thought would be the most horrible environments for a qubit to be in and the most interesting, and for me that's room temperature biology. For a physicist, normally working on very simple systems, an atom is a very simple system although conceptionally very complex but physically quite simple. Biology on the other hand is vastly complicated and full of tremendous self-organisation systems and a lot of order out of chaos. So for a physicist these are fascinating systems. It turns out that even when you expose our qubits to that sort of system, you can still get some meaningful information out. So we've been through a process of understanding the sensing ability of our little nanodiamonds and we're at the stage now where we're thinking of and actually putting these into biological systems to see what we can sense. 

SHANE HUNTINGTON
Now, correct me if I'm wrong but diamond is biologically a compatible material. It's quite inert as well.

LLOYD HOLENBERG
That's right. That's one of the nice properties of diamond. So there are many remarkable properties that have come to intersection in this system. So as a qubit it's just wonderful because we can measure it optically through that stimulation of laser light and reading red light. In a lot of other systems the detection of the quantum properties is electrical. So that's one really nice property. We can control it just through microwaves, which is very straightforward. But also for biology, the diamond itself is very low toxicity material. The other property that's very nice too, just a tracking object in biology, is that the fluorescence is extremely stable and tends not to blink or bleach, which can happen with other fluorophores and quantum dot-type systems.

SHANE HUNTINGTON
We should clear up one another property, and that is this misconception that diamond's an expensive material. 

LLOYD HOLENBERG
Oh yes.

SHANE HUNTINGTON
It's expensive if you're buying gem quality, literally mined diamonds, but it's not hard to make, is it?

LLOYD HOLENBERG
No, it's not that hard to make these diamonds. We tend not to do that. There are places where you can buy nanodiamonds and I think it comes out to be about a picocent per nanodiamond, which is pretty cheap.

SHANE HUNTINGTON
I have to get my head around that number. That's an extremely small amount of diamond but a very small amount of money.

LLOYD HOLENBERG
But when you consider you have billions of them in a sample that adds up a little bit too. You know, maybe a few hundred dollars. The expensive material is bulk diamond and there are applications where instead of using a single nitrogen vacancy centre in a nanodiamond you could enhance the sensitivity by having many nitrogen vacancy centres in a single crystal that's relatively large. It might be millimetres by millimetres. 

SHANE HUNTINGTON
I want us to start getting our head around the resolution we're talking about here. So most imaging devices, like an optical microscope for example is limited in resolution to a little better than half of the wavelength of the light being used. MRI machines are limited in their wavelength, not much better than that. But here you're talking about just a few atoms as the sensing device. Does it track that the resolution is commensurate with the size?

LLOYD HOLENBERG
It does. To give you an example, when we put nanodiamonds into a lipid bilayer - well more correctly, we formed the lipid bilayer around our nanodiamonds, and our nanodiamonds are about 15 nanometres in size in that case. The sensing volume is commensurate with the size of the nanodiamond itself. In that case we were able to detect Gadolinium labels - Gadolinium is an atom that has a fairly high magnetic moment. We were able to label some of the lipids in the bilayer with these Gadolinium atoms and then detect their presence through just measuring our NV centres’ properties and deduce that we were detecting down to about four Gadolinium atoms in the neighbourhood of our sensor. So to give you a comparison, an MRI machine which gives you beautiful images based on detecting small regions of atoms in the body, needs about several billion of these atoms to get any sort of a signal. That corresponds to the resolution limit which is around about a millimetre or below. So in this case we can detect down to about four atoms and if we use a smaller nanodiamond, we could detect down to a single atom in that case.

SHANE HUNTINGTON
You're listening to Up Close and today we're talking to Professor Lloyd Hollenberg about quantum sensors based on diamond. I'm Shane Huntington. Lloyd, what sort of biological structures or processes do you hope to be able to measure with this kind of sensor down the track?

LLOYD HOLENBERG
That's a great question. We're still finding that out. We've investigated a few possibilities. Because the sensor is steeped in quantum physics, to understand that question, to really answer that question, you need to understand the interaction of our sensor with these sorts of environments. So it's not as simple as saying, oh, we just want to measure a certain magnetic field. We have to understand how certain magnetic fields in biological processes generated at the atomic and molecular level will affect the decoherence of our quantum sensor. Some of the initial work was to look at iconic problems to see how the quantum sensor might detect in those sorts of situations. So one of them was the flow of ions through an ion channel. So ion channels are, of course, incredibly important. They're the little openings in cell membranes that effectively allow ions to pass through. Normally you would detect those by electrical means through a patch clamp method, which is terrific. In our case the quantum sensor would see that completely differently. Our quantum sensor is not really sensitive to electric fields, which is actually really good because the micro scale at that level is full of electric fields. It's very noisy in that sense and this is a challenge for electric sensors at that level. In terms of magnetic fields, an ion channel process would correspond to about 10,000 ions passing through the channel with their nuclear spins orientated at random. Little bar magnets at random orientations, essentially. That presents a certain fluctuating magnetic field to our sensor. So we did some extensive calculations where we looked at bringing a nanodiamond into proximity to the ion channel and looked at what the signal would be in terms of measuring the quantum decoherence of our sensor and we found that it would be able to detect that ion channel event in near real time. 

SHANE HUNTINGTON
Lloyd, given these sensors are very much based around the measurement of magnetic fields, I suppose most people wouldn't think of their bodies as creators of magnetic fields. So what are we measuring in terms of magnetic fields within the body?

LLOYD HOLENBERG
Well, it turns out your bodies are replete with magnetic fields. All the electrons and many of the nuclei in the atoms in your bodies have these very small magnetic moments and generate magnetic fields. More than that, fluctuating magnetic fields. The timescale of those fluctuations are characteristic of what those atoms and indeed molecules are doing. So a really good example are free radicals. So free radicals have a free electron, essentially, and that electron has its own magnetic moment and as it moves around it's essentially a magnetic field that's changing in time. If it's flipping backwards and forwards or indeed spinning around, it's also fluctuating at a certain rate. It's those sorts of processes that we believe now that we can measure, even down to the single free radical scale.

SHANE HUNTINGTON
Obviously, one of the things that the MRI machines are used for most is imaging the brain where it is difficult to get to the brain in a non-invasive fashion and the MRI and its predecessor, the CT scanners and so forth, have given us incredible views of that. When you consider things like neurons and so forth, where there is a lot of electrical activity - and as you say, these sensors are not affected by that, what can that tell us about those particular objects and have you looked into things like neurons?

LLOYD HOLENBERG
Yes, we have. It's a very interesting problem in that sense, that usually again the measurement of a neuron firing is through an electrical probe. Of course, we know as physicists that when you have a current travelling down a structure and a neuron firing is essentially generating that, there's always a magnetic field associated with that. So we simulated a neuron firing through fairly accurate physiological models, and we did this with Steve Petrou's group. We then converted that to magnetic field strengths and we looked at what would happen if we had our neurons growing on a diamond surface, so a bulk crystal surface not nanodiamonds, wherein we have a layer of nitrogen vacancy centres in the lattice just below the surface. Then measuring the light over that area would be a CCD device that would basically take an image of the fluorescence from the NV centres. We found that that collection of NV centres, by having many of them you multiply the sensitivity because the magnetic fields from the neurons are very small. They're only about a few nanotesla, so that's a very small field. We found that the image would actually capture those magnetic field changes, corresponding to the neural activity. Then the interesting question was, well then, what would be the resolution, both spatial and temporal? We found that the spatial resolution could be down to just a few microns or less, and the temporal resolution would be down to about a millisecond or less. So that's getting into a unique space because normally if you try to measure neurons you need to actually probe them with little wires and you're quite limited in your resolution. You get very good temporal resolution but the spatial resolution wouldn't be that good. This seems to be in the middle ground where we could have potentially very good spatial resolution and pretty good temporal resolution. So this is one of the detection problems that we're looking at right now.

SHANE HUNTINGTON
So Lloyd let's say for example you get all this resolved, you have a system you want to test in the body, what would you be involved for a patient? Do they need to drink a cup of nanodiamond? Is it something you'd have to specifically place? How would that work?

LLOYD HOLENBERG
Well unfortunately, in vivo is challenging for this technology. We imagine a lot of this would be in vitro, where you're testing the effect of drugs and so forth. So our initial systems that we're building at the Centre for Neural Engineering will be for that purpose. Initially of course, we want to show that we can actually carry out this detection. In vivo is a challenge because not only do you need to get green light through tissue but you need to get the red light out through tissue as well. So you're often very limited in how you can do that. But I think watch this space. I think a lot of people are looking at this. Obviously, we're not the only people looking at quantum sensing using this diamond technology around the world. It's actually quite competitive now. That's one of the key challenges that people will be looking at in terms of applications in biology.

SHANE HUNTINGTON
Lloyd, I think you've given us all a lot to think about during this episode and we thank you very much for being our guest in Up Close today and talking about diamond based quantum sensors. 

LLOYD HOLENBERG
Thank you, it's been a pleasure.

SHANE HUNTINGTON
That was Professor Lloyd Hollenberg, the Deputy Director of the Australian Research Council Centre of Excellence for Quantum Computation and Communication Technology in the School of Physics at the University of Melbourne. If you'd like more information on this episode visit the Up Close website where you'll also find a full transcript. Up Close is a production of the University of Melbourne Australia. This episode was recorded on 9 October 2013. Producers were Eric van Bemmel, Peter Clarke and Dr Dyani Lewis. Audio engineering by Gavin Nebauer. Up Close is created by Eric van Bemmel and Kelvin Param. I am Dr Shane Huntington. Until next time, goodbye.

VOICEOVER
You've been listening to Up Close. We're also on Twitter and Facebook. For more info visit upclose.unimelb.edu.au. Copyright 2013, the University of Melbourne.


show transcript | print transcript | download pdf