AUSTRALIAN FRONTIERS OF SCIENCE, 2003
Canberra, 31 July to 1 August 2003
Advanced materials
characterisation
by Associate Professor Dougal McCulloch
 |
Dougal McCulloch was recently promoted to Associate Professor – Microscopy and Microanalysis in the Department of Applied Physics at RMIT University. In 1999 he established, and is now a director of, the RMIT Microscopy and Microanalysis Facility, which is utilised by researchers from many disciplines within RMIT and beyond. He is expert in applying a wide range of materials analysis techniques to the investigation of carbonaceous and thin film coatings including electron microscopy, surface and ion beam analysis and Raman and infrared spectroscopy. He also established the Car-Parrinello ab-initio molecular dynamics code in Australia and performed an extensive range of simulations on carbon and light element binary and ternary alloy systems. His achievements in this area were recently recognised when he was invited to give a keynote lecture at a symposium on Amorphous, Disordered and Incommensurate Materials at the 15th International Congress on Electron Microscopy (ICEM-15) held in Durban, South Africa in September 2002. |
In this part of the session we are going to talk more specifically about advanced materials characterisation.
I am first going to talk about what we mean by limits of resolution, that is, how small we can 'see' or analyse. Then I will talk about what is currently possible in terms of analysing materials, then a little bit about atomistic modelling, which is used to help understand what you might be seeing in a particular method, as well as to predict structures and properties – Marcela has already talked a bit about that – and then some words on the future as I see it.
Click on image for a larger version of figure 1
How small can we see? What I mean here is that if you have an imaging device and you want to actually image something, Rayleigh's criterion tells us that resolution depends basically on the wavelengths of the radiation that we use – as well as some other things, of course, but that is the main consideration. I have a diagram here of the scale of things, to put things into perspective. When we want to see smaller things, it is really a very powerful technique to be able to actually image what you are trying to understand. And of course when we cannot see it any more with our eyes, we use microscopes. The light microscope is able to resolve down to approximately the wavelength that you are using in that light microscope, about 200nm to 500nm. If we want to see things smaller than that, then we need to use a more sophisticated instrument, like an electron microscope or something else. A lot of the things that we are interested in occur down at these scales, so it is really important to be able to analyse them.
I have put there [in diagram at right] some things that you might want to look at. Most of the new nanomaterials that are being constructed are of the order of 10nm or so, and if you want to understand how they work you need to be able to look at interfaces and defects around the 1nm or less level.
So what can be done now? Well, you can think about three things you might want to know about a material or a system. You want to be able to actually image it, to understand how it is built or what structure it has; you might want to know what elements are present and where they are, and how many of them there are; and another important area is maybe how all those elements are bonded, because that can be very important in determining properties.
You can categorise these techniques into two broad areas – surface ones that look at just the surface of the structure, or ones that look at the internal structure. I am not going to be able, obviously, to go through all the techniques that are around, in this short 20-minute talk, but I thought I would show a few.
Click on image for a larger version of figure 2
This is a scanning electron microscope, which is quite familiar to a lot of people. This is a brand new one that has just gone in at UTS, in Sydney. It is all computer-controlled these days, and it has what is called a field emission gun, a very bright gun which allows you to image at very low voltages, which really brings out details in the surface of samples. We have a couple of samples here from that lab, of sapphire plates and small pores in inactivated charcoal.
Click on image for a larger version of figure 3
The other big advance recently has been the development of what they call environmental electron microscopes. The problem with electron microscopes is that normally you need to put the sample in a vacuum to look at it, and it needs to be electrically conducting as well, which rules out many sorts of systems. But now the engineers have become clever enough to build a vacuum system so that you can keep the gun where the electrons are generated under a vacuum and you can put your sample into a higher-pressure environment so that you can actually look at wet samples. These are some examples of what you can do. One is a cross-section of a blade of grass, uncoated or untreated. You can look at live things, actually. People have actually done movies of poor things that have been looked at in an electron microscope. These are wet bacteria. The other exciting part about this area is that you can actually do in situ things: you can cure concrete and watch it cure in an electron microscope, which means that you can look at it in much more detail than you could with an optical microscope.
Click on image for a larger version of figure 4
Another big area these days which is really taking off is scanning probe microscopes. Marcela has already shown you this image. There are two sorts: scanning tunnelling microscopy and atomic force microscopy, basically. Both of them give you a very detailed view of the surface structure of a sample, down to the atomic scale. These days, people are becoming cleverer with these instruments and they are doing things like measuring adhesion of biological molecules onto surfaces, by designing the experiment with these devices.
Click on image for a larger version of figure 5
So they are three areas where you are looking at the surface of samples. If you want to look at the interior structure of a sample, it is quite a lot more difficult. The main technique around that I am familiar with is transmission electron microscopy. The main problem here is that you have got to prepare a thin sample so that the electrons can pass through that sample and then you can image what is going on inside that thin sample. These sorts of devices have been around for a long time. As the electrons pass through the sample they interact and we can form an image. This [at right] is an image of a GaAs/AlGaAs interface, which is what you would call an atomic level image, a lattice image. These blobs are supposedly the columns of atoms, or pairs of atoms in this case.
Click on image for a larger version of figure 6
As the electron beam passes through the sample, it interacts in many ways and we can get X-rays generated which are characteristic of the elements, and these can be used to do elemental mapping.
This [first example on slide] is a quick example of a titanium oxide catalyst, which was treated with silicon oxide and aluminium oxide. The scientist wanted to know the distribution of these elements on the surface, which is very important in determining its properties. With this sort of analysis you can use a probe these days of about 1nm, which is quite small, and you can, for instance, do a trace along this edge and find out that the aluminium is actually coating the silicon a bit more than the other way around.
 Figure 7 |
 Figure 8 |
| Click on images for larger versions | |
Another area that has become more advanced recently is the attachment of energy filters to electron microscopes. This [at left] is a typical transmission electron microscope. You put your sample in the middle – for those who are not familiar with this – and you have got your electrons generated at the top. You normally look through the screen or you look at an image on a TV camera. But these days you can put energy filters at the bottom, which actually get you more information out of your normal image. What do I mean by that? Well, as the electrons pass through the sample they interact differently in different parts of the sample, depending on both what elements are there and how they are bonded.
If you filter the image in this spectrometer you can do some interesting things. This is a titanium nitride film on silicon. It has been prepared in cross-section. That is, you take the sample and you have to fiddle around and thin it so you are looking at it side-on. This is the surface along here and this is the substrate. It is about 150nm thick. If you put it through the filter you can actually find out where the different elements are. This can be done at quite high spatial resolutions because of the way the interactions occur. People can do this sort of mapping at about of the order of 0.2nm.
Click on image for a larger version of figure 9
You can do other things too. To help humans identify what is going on it is a good idea sometimes to combine images and see their coloured arrangements, or you can do correlation mapping. This is a cross-sectional image again of an oxide, the surface of an aerospace alloy, which happened to have quite a high concentration of magnesium in it.
Click on image for a larger version of figure 10
The other thing about this technique is that you do not only get elemental information, you can also get bonding information, which is quite important. In this case here we have an example of doing that, using a CVD diamond film, which is just a diamond film that has been deposited onto a silicon substrate – quite common these days. These researchers were interested in knowing what was at the very interface between where the diamond starts to grow and the silicon. So what they did is that they set up their microscope so they were looking at it side-on again. This is the bottom of the diamond crystal and this is the silicon. These are actually not elemental maps; you can think of them as bonding maps. This bottom one here is white where there is diamond bonding, and this little one is something called a diamond-like map which is picking up also what we call diamond-like amorphous carbon. Then at the top we have a graphite-like bonding map. As you probably are all aware, two main forms of carbon are diamond and graphite. You can see here clearly that on top of the silicon is a graphite-like layer which is only about 5nm thick, between the silicon substrate and the CVD diamond.
Click on image for a larger version of figure 11
Marcella has already mentioned atomistic modelling, which is used not only to predict new structures but also to help us interpret what we might be trying to image or characterise. Also, it allows us to do experiments in the computer without having to try and do them in the lab, which can sometimes be a lot harder. The other thing about this sort of modelling is that it can really give you insights which you may not have imagined, and then you can go and look for that in your experiments. These methods are now being done on the quantum scale, so you are really trying to do them as accurately as possible, to take into account what the electrons are doing.
This [at bottom left] is an example of a study which was looking at stability of diamond nanostructures: how small can you make a diamond nanoparticle for it still to be diamond? There was a view that maybe it relaxes when it gets too small. In this case, this morphology actually does, in this computer model. It changed, or relaxed, from a diamond-like structure to a fullerene-like structure. As Marcela showed you, we can actually identify bonding in these models by looking at what the electrons are doing, and these sorts of maps.
The other thing you can do is that once you know what the electrons are doing you can also calculate some electronic properties in the computer and compare them with what you might measure in an experiment.
Click on image for a larger version of figure 12
To come to the future, this diagram here [at left] shows you the resolution of the light microscope and electron microscope as a function of years, and various people who have made a major advance in improving resolution. The light microscope, starting in here [at about 1820] has obviously improved with time as we improved the lenses and then asymptote towards its theoretical limit of resolution – which as I said before is limited by the wavelength. In fact, our light microscopes are made these days with hardly any problems in terms of aberrations or defects in the lenses.
The electron microscope has got the same sort of curve. However, the problem with electron microscopes is that the lenses are not defect-free. They are very poor, in fact, compared with what you are used to in light microscopes. So, whereas in a light microscope your resolution is of the order of the wavelength, in your electron microscope where you have a wavelength of, say, 0.02nm, your resolution is 100 times worse. That is because of the lenses and other problems. So there is now quite a lot of work being done in trying to correct for these aberrations or defects in the lenses.
These are called aberration correctors, and there are a couple of ways of doing it. You can use software, taking images at different conditions and putting them together, or you can actually put a physical aberration corrector in front of your lens that you are trying to correct. It is a bit like thinking about an electron microscope with glasses. So, up to now, our poor electron microscope has had to have glasses to be able to see better.
Click on image for a larger version of figure 13
Now we are starting to see some of this work come into effect. This [image at left] is some work from last year, some work from Phil Batson and others. We have seen gold nanoparticles for a long time, but this was done with an aberration correction such that the resolution was 0.1nm, or 1 Ångstrom. Straight away they saw much more detail around this gold nanoparticle. They actually started to see individual atoms and clusters of atoms which were identified using these diffraction patterns.
So, just by improving the resolution from a typical conventional electron microscope, which might have a resolution of, say, 0.2nm, to 0.1nm, we start to see a lot more detail than we ever saw before. Every bit of resolution you can get helps in trying to understand what you are looking at. For instance, they suggested that 0.1nm should be sufficient to start to detect single dopant atoms in semiconductors, which is quite a challenge.
Click on image for a larger version of figure 14
The other area that is becoming more popular, or more developed, is to go from 2D imaging like this to 3D imaging or representations. Here there is actually a platinum nanoparticle on an alumina rod or something. There are two ways of doing this. You can do holography [as in top image] or you can do tomography, where you take images at different angles and then construct what you are looking at in the computer. I must admit that I find it much easier to interpret something like that than just a 2D image, to understand what I am looking at.
To come to the future for characterisation: I have not had time to go through every possible type of method, but basically and probably similar to other areas, it is all about trying to improve spatial resolution, being able to resolve smaller things. There are quite a lot of challenges to do that. As you start to improve the resolution, then you have got to worry about your environment more and more, because if you are trying to image at 1x10-10 m, which is 1 Ångstrom, then you cannot have any outside influences, otherwise your image or whatever you are trying to do will be disturbed. So there is a lot of work to be done there by engineers to design these things. Another idea is to start to do remote operation, where you might put your instrument in a place that is isolated from humans, to remove these potential problems.
I think we are going more and more to the analysis of individual atoms, and, as already mentioned by Marcela, a lot of these methods that are used for analysis can also be used for manipulation. So you can do both at once. And you need to be able to do that on the nanoscale, because you cannot see what you are doing most of the time if you want to try and construct something that way.
[Some words not recorded during audiotape changeover – probably beginning to address the fifth dot point: Understanding the nano-world is complex!]
...then we cannot think in our normal way. We have to think in the land of quantum mechanics, and it is very hard sometimes to imagine what you might see. So you need these accurate, computerised models to predict what you might see, or to help understand what you are looking at.
I have not talked a lot about other microscopies because there is not enough time, but obviously there are opportunities for other microscopies to make major impacts. The X-ray microscope at the moment is again limited by being able to make a lens which is good enough to be able to do a lot of imaging. And there are other techniques like atom probe field ion microscopy, which is an atomic scale type analysis technique.
So what would we see, for instance, if we could make our aberration corrected electron microscope which had a resolution better than 0.1nm? Well, we might be able to see the actual atomic structure of an atom. What that would look like, I am not sure, because it is at the quantum level.
