AUSTRALIAN FRONTIERS OF SCIENCE, 2005

Walter and Eliza Hall Institute of Medical Research, Melbourne, 12-13 April

Session 1: Discussion

Question – I have a question about biological imaging. There are two basic ways that people do it: you prepare a slide and you do your imaging, or you prepare tissue. Can you explain, with each of those methods, what is the limit of resolution in depth that you would get using your new technology, as compared with what we have now?

Min Gu – The theoretical limit of the axial resolution, the depth resolution for this multiphoton technology is about 200 to 300 nanometre resolution. But that is only based on using a conventional lens, if you just buy the lens from Thiess or another commercial company. But we are doing nanophotonics, so if we can squeeze the light even smaller, break that formula I talked about, then we can do much better. In recent technology in the laboratory at the Max Planck Institute they can do 30 nanometres axial resolution with this three-dimensional technology.

Question – How does that compare with the 4π microscope?

Min Gu – That is exactly what I am talking about, the 4π. But it is very expensive, about €1 million for the system.

Question – Very innovative work you are doing here. Could you just highlight for me the timescale of this research? How long will it take you to achieve your ultimate objectives? In addition to that, you showed some great lithography, I think, of Australian icons and what not. Can we make things with movable parts? There is a fair bit of press about nanomachines, I think, in the UK.

Ben Eggleton – I will comment on the first part. Min Gu can comment on the movable parts, because that is Min’s, I believe.

Our program is a five-year centre, so we set out with a pretty ambitious goal to create, at the end of the five years, working prototypes of chip devices that would have some of the functionalities that we believe are going to be part of the next generation system. We are about six months away from having a demonstration of one of the key functionalities, regeneration.

One of the bottlenecks in a very high speed communications system is that you need to regenerate periodically through a network, and right now that is being done by these very clumsy – we say – systems, electronic solutions. We are about six months away from having a very impressive proof of concept of that, and that is the first demonstration. That will operate at about 500 gigabits per second, so it is 10 times faster than anything in the ground now.

That device will take a number of years to develop into something that would be commercially viable. I think most of our research that is chip-based is 5 to 10 years away from being ready for applications.

At the same time, we have a big program in nonlinear signal processing using off-the-shelf components that will be ready for commercialisation as soon as the industry and the demand picks up, which is a few years away from now.

So I think the program is in two phases. We have the chip-based devices, for which what we have been setting out to achieve has taken a few years to build a momentum. The microfabrication is now under control, we have got the materials, we are making the devices, we are validating right now in labs in Sydney and at ANU some of these basic functionalities. The final phase of our Centre is starting to bring together different functions onto a chip, and, depending if we go beyond the five years, that is when we will be really starting to demonstrate chip-based devices and focus on technology.

Was that the first part of your question? Centres of excellence are about ambitious, very exciting, great-impact problems. We look at this as the grand challenge of our generation, if you like. It is making sure that there is the infrastructure, the technology that will enable systems, not three years, not five years, but 10 years beyond that. And as to the bottleneck, we are talking about a factor of 100,000, potentially, in the speeds at which these little routers can process data.

So it is a grand challenge and it will take some time. But we have got a pretty exciting result in the next six months, which is the first building block for that chip.

Professor Barry Luther-Davies (Chair) – I think the second part of the question was about whether you can make dynamic reconfigurable microscale systems.

Min Gu – Maybe the answer is to talk about the 3D structure we made. Eventually if we want to use this concept using photons to replace electrons, we have to have a material like a semiconductor, so it must be a three-dimensional structure. With the technology we can develop we know we can make such a 3D structure in any arbitrary shape. So what we are doing now is to combine this technology with the technology Ben is developing in Sydney, trying to put it on this chip device. I think we need some communication between those two-dimensional structures and the three-dimensional structure interacting with each other, and that will make this functionality. The technology we are developing is ready to do that.

Ben Eggleton – Let me have a go. Can one reconfigure a microscale system? There are ways of doing that. One way is that one can put active materials into a system like liquid crystals, which can be modified electronically to modify the properties of the system dynamically, at a speed of kilohertz. I didn’t talk about this, but the nonlinearity that we rely on, which in effect is a reconfigurable property of the material, is femtosecond speed. That is an example of how the system has been reconfigured.

There is one other example that might intersect with people here. I have a small project in microfluidics, which may or may not be an emerging paradigm any more but in the context of lab on a chip it is particularly hot in the US. There is a big program funded by DARPA in optical fluidics, in a sense the biology equivalent of the photonic chip – constructing a wafer that has microscale channels for moving fluids around and making new fluids. And one of the themes that we are exploring is how we can integrate microfluidics into a microphotonic system to actually achieve dynamic functions on a chip. A very simple embodiment of that is if you have an optical waveguide that is a few microns wide and you are sending light, and you have a microfluidic channel that intersects with that optical waveguide. Well, if I move a fluid back and forth through that beam of light I am going to scatter or transmit, attenuate/transmit. That is an example of modifying, reconfiguring the microscale system.

Spin that around, and you can use your optics to do spectroscopy on the fluids on a very, very small scale. So that is a spin-off application of microphotonic systems.

Chair – I guess I could also comment on this a little bit, in the sense that there is an organisation in Melbourne called MiniFAB which is involved much more in the lab on a chip type of work. I guess the general structures that you use for those kinds of approaches are rather bigger than the ones we deal with here, probably by an order of magnitude. In that sense, the job becomes a little easier, but then of course if you are trying to make a lab on a chip type functionality really effective, first of all it has to cost virtually nothing, because the biologists want to plug a thing in and then throw it away, and secondly you have got to figure out ways of getting incorporated the kinds of functionality that will actually provide you with some measurement off the system.

While some of the techniques that have been talked about here are relevant to lab on a chip concepts, I think that, as I say, the main emphasis of what we do is probably at a smaller scale than is viable for a lot of those concept channels. It is hard to push fluids through submicron gaps. You can do it; it is just that the forces required are somewhat large.

Question – We use lasers now to collect or destroy single cells. I was just wondering where you illustrate that you can register a signal from a particular class of molecule, maybe a single molecule, in part of a subcellular structure. Do you see then that that could be subsequently accompanied by peta-oblation or whatever it is going to be, that you could destroy or modify that particular molecule if there is a signal, for example, that is a molecule that is likely to lead to abnormality? Are we going to see surgery at that level?

Min Gu – The answer is yes. As I said, I was in Germany to celebrate Abbe, who died 100 years ago, and one of the topics was nanosurgery, including at both the cell level and the tissue level. There are two ways to do it. A lot of people want to do it inside tissue, and there are some problems but that presents a challenge for optics people to solve, and others want to do it at a cellular level. There the challenge is to try and reduce the spot to even smaller than Abbe’s limitations, say down to 100 nanometres. If we can do that, then it is possible to do the cutting, the surgery, at a cellular level, using these multiphotons.

That brings me to one technology I didn’t talk about. As I said, I was just using two photons to illustrate using this sort of nanotechnology. You can do this by what is called CARS, coherent anti-Stokes Raman scattering, and that is exactly what people are developing for this molecular imaging.

So the answer is yes.

Question – I presume that there are clones of your Centre all round the world – maybe not clones but similar laboratories. Do we stand a chance to protect intellectual property that you guys are perhaps generating?

Ben Eggleton – In terms of clones, there aren’t clones. We debated this to death at the beginning, and there really is not an example of a centre that has everything together like we have. That is what makes us very strong, because we have Barry Luther-Davies in Canberra, world class materials, lithography; we have Yuri Kivshar, we have the de Sterke-McPhedran theorists, we have devices. We cannot find an analogous centre. We can find centres that represent a subset of what we are, but nothing that brings together microfabrication, the lasers, the theory of electromagnetics, the devices. From that point of view we are out there in the lead and we should, in the next few years, really start to have a lot of impact.

In terms of intellectual property, I certainly hope that we can capture the intellectual property, in that I hope that one of the outcomes of this Centre, going forward, will be a real impact on the Australian industrial sector, commercialisation, proved infrastructure and so forth. That will come in time. Right now it is very early days, we are building up momentum. But there is no reason we can’t capture that.

There is no reason we have to manufacture here. I don’t confuse that with manufacturing. Israel is an example of a very strong knowledge-based economy that is doing very well but doesn’t do all its manufacturing in Israel. But we can benefit enormously if we can capture intellectual property, develop some of the technology here, partner, outsource and so forth. That is certainly the goal, but it is very early on at this point.

The IP is being captured at the universities. It is being consolidated at some level between the different universities, so we have an IP vision, if you like, going forward.

Question – But we have three-layer CDs or DVDs now out there. Have we missed that?

Min Gu – Well, for the 3DCD we have just received patent approval from the United States, so Singapore, the United States and, I think, Taiwan are all approved, and we are waiting for final approval from Japan.

The multidimensional concept – I am talking about the quantum dots – is a new one. Probably we are the first group to propose this idea, and we presented some preliminary results last year in a conference on optical memory, and immediately there were a number of companies, including Simpson, LG, Philips, Sony, who wanted to join with us to develop it. I think that is probably a good strategy, because in Australia there is no large optical memory industry. If we insist that we want to develop everything here, we probably won’t reach too far. So we changed the strategy, when we developed the multidimensional optical storages, to joint development.

Question – With your switching devices you basically mentioned that smaller is better. Is there a theoretical limitation going from, say, the micron scale, the periodicity of microns, to, say, the nanoscale or even smaller?

Ben Eggleton – Is your question about whether there is a limit to the speed of the switching, or to the size of the switch?

Question (continued) – Just the theory of light. You are talking about Maxwell’s equations and things like that. If you go from a periodicity of, say, micron level down to even smaller than submicron or even nanometre scale, do you get different theories occurring – or can you go to the nanoscale? Why are you at the micron scale for your switching devices?

Ben Eggleton – It is all classical. Quantum effects are just not showing up on this regime. You are dealing with the wavelength of light, which is about a micron – 1.5 microns is the region of interest – and it becomes more and more difficult to confine light at very small and even smaller scales. So if you think of an optical fibre or a piece of glass, if you make the glass strand smaller than the wavelength of light then the light is no longer confined or trapped by that glass strand. It becomes more and more difficult, then, to actually control the light. You are limited to operating on a scale of the wavelength of light.

Now, it is important to note that electronics can go smaller than that, obviously. If you visit Yorktown Heights you will find that they are talking a roadmap that will take them out to 30-nanometre electronic transistors. So this is not replacing the hardware that goes in a computer for doing basic logic functions; it is really enabling you to do, in a sense, logic functions at a much, much higher speed – thousands of times faster than electronics. It will be larger-scale, because we are limited by the wavelength of light, really. The wires that we are talking about are about half the size of the wavelength of light.

So everything is classical here. There are no quantum effects that show up in most of our systems. You talked about putting quantum dots, and obviously there are some quantum effects and there are some interesting aspects.

Min Gu – At a device level, probably this is not a problem, but when you talk about the fabrication there are some quantum effects. For example, one of the things that we had to make the photonic crystal is nonlinear. So in that case we had to dope the quantum dots into the bulk medium, and in that interaction, when we fabricated such a material, there were some quantum effects. But that is only in the fabrication process.

Ben Eggleton – That being said, even Maxwell’s equations become pretty tricky. And so we have got some of Australia’s top theorists – I mean, these are real heavyweight theorists – who are doing very heavy, sophisticated calculations. They are running these calculations on clusters and state-of-the-art computers and so forth. So it becomes very difficult down in that regime even to model Maxwell’s equations, let alone add quantum effects.

Question – There seems to be a dichotomy between the sorts of wonderful discoveries that you blokes make and the development work for which the Centres of Excellence are funded. The development work is all in terms of milestones and what products you can produce and so forth, whereas in fact progress depends on the sorts of discoveries for which you get the IP and so on. Is there a philosophical problem at the political level of what you are funded for –the concept of Centres of Excellence versus, really, discovering new things, which is what progress will depend on?

Ben Eggleton – We are about discovery. It is a science centre, and it is about discovery. And the research we are doing is very much blue-sky research. There also is some focus. So it is a biased random walk. We have come together to work as a team and we all have bought in to this vision of creating microscale photonic signal processing devices, but within that context there is an enormous amount of flexibility in the program. So we don’t talk about product. We have milestones that are agreed upon on a yearly basis, but I think even then there is a lot of flexibility. We all are working towards a common goal and there is a very strong sense of collaboration between the groups in the Centre, and so it is scale and focus.

We do have debates – and all the people are here, I can talk openly about it. We had our Advisory Board meeting last week, and one of the issues that come up in our Advisory Board is end users and applications, and making sure that we are positioned to have impact in the Australian context. You have to reconcile that with also being a centre that is based on discovery, research and innovation. And it is not a CRC, it is not a corporation. But I think, looking at my colleagues, we have a good philosophy at the moment with things. It is the excitement of science, the excitement of discovery, but we have a vision of creating chip-based microscale devices and there is a great opportunity there.

Question – Forgive me for my ignorance; I am no physicist. But I was just wondering, when you were talking about biological imaging, how this nanophotonic technology either compares with or perhaps complements synchrotron technology.

Min Gu – I think it is different. First of all, the wavelengths are different. The synchrotron is in the X-ray/UV range, and as far as I understand it this type of light produces a lot of side effects for biological tissues. What we are talking about here is in the optical wavelengths region, and in particular the multiphoton technology is using infrared lasers. With the infrared laser there are a lot of advantages. I didn’t talk about that.

As one example, I have emphasised that you can do imaging inside the bulk medium. When you go inside biological tissue there is a lot of multiple scattering. If you are using infrared wavelengths, compared with the UV light that most people are using, the scattering reduces dramatically because it is in inverse proportion to the wavelength’s force power. So there are a lot of advantages in using these optical wavelengths and then you can have rich colour information. So, for example, with the confocal microscope you are using, you can now do multiple channels. That is the advantage, compared with the synchrotron technology, though synchrotron can certainly provide high resolution because of the short wavelengths.

Question – I have a question crossing the boundaries, as a biologist. Are you guys looking to nature for examples of structures for transmission of light? In our area of coral research there have been some recent discoveries that coral structures are some of the most highly scattering substrates, and there are patents out for that. I saw a lecture just recently on butterfly wings and the creation of fluorescent signals from monolayers within butterfly wings. There are micro-algae, there is a whole bunch of natural substrates out there that I am not sure if you are tapping into.

Ben Eggleton – Not tapping into, but certainly well aware of it. Many of my colleagues are involved in the butterfly research. You know, the students love the butterflies. It really gets the students excited, and they are observing photonic crystals on the back of butterfly wings and they are calculating – using the same tools that we are – band structure and density of states and emission properties of butterflies.

One of the chief investigators in the Centre is trying to incorporate dye into a sea-mouse, an animal like a slug that lives on the bottom of the ocean, which has a photonic crystal. It looks like one of these holey fibres, in a sense; it is a channel with a periodic structure around it. And it is wonderful that nature has provided these.

A number of times I have had people in the audience say, ‘Get rid of this lithography. Grow photonic crystals.’ I guess we should give that serious thought, but I can’t see any practical way of taking advantage of that at this point. But we can learn from it, and it is fascinating stuff.

Chair – This is very much an example where the physicists are trying to catch up on nature. If you look at the structures of butterfly wings, for example, you see that they are much more complex than the structures that we currently make. But fundamentally, from the optics point of view, they work the same.

There is another interesting area that again the physicists are just cottoning on to. A lot of what Ben has been talking about, and what our program relies on, is so-called nonlinear optical materials. There is a growing amount of interest now in the use of DNA as a nonlinear optical material, or at least to template other materials that you essentially can insert in the helix to obtain the kind of lack of centrosymmetry that is very important for some nonlinear optical processes. Looking at natural materials is quite fundamental to a lot of what optics does, and of course optics provides a lot of the tools for looking at those materials themselves. So there is, I think, a quite nice synergy there.