Work at Daresbury is leading to new, stronger, alloys.
Pure metals are now well understood. The positions of the atoms and the electronic structure - the way in which the energy is shared between the electrons - are known to a very high precision. Experimental measurements are closely matched by theoretical predictions, and fully understanding the properties of metals has been one of the triumphs of modern solid-state physics.
The next challenge is alloys, in which two or more metals are mixed together. Alloying can drastically change physical and chemical properties; most metals in everyday use are alloys. For instance, pure aluminium and pure magnesium are very soft, yet an alloy of the two can be hard enough to be used to make ships; dental amalgam is an alloy of mercury and other metals, pure mercury being a toxic liquid. These examples are typical of alloy science, where their material or chemical properties are important. Now, it is important to understand the electronic properties, partly for their own sake, and partly because some alloys are being used as catalysts, or to improve the output from thermionic cathodes.
The difficulty with modelling the electronic properties of alloys is that different atoms are in different places in the material, and this lack of translational symmetry precludes the use of the normal methods in the theory of electrons in solids. Also, from the experimental point of view, X-ray crystallography can give the overall arrangement of the atoms in the alloy, but does not tell us about the relative arrangement of the different types of atom. This information can come from EXAFS, and - using the tunability of the synchrotron radiation - the local environment around each of the different types of atom in the alloy can be determined.
Recent work at the SRS, by a team led by Peter Weightman and Carol Jones from the University of Liverpool, in collaboration with Bob Bilsborrow and David Norman at Daresbury, has used EXAFS to shed new light on the arrangement of the atoms in copper-palladium alloys. Each of the pure elements, copper (Cu) and palladium (Pd), has a very simple crystal structure, known as face-centred cubic, in which each atom has 12 nearest-neighbours in a perfect cubic arrangement. When mixed together, they still make a crystal, but two types of alloy can be formed: disordered alloys, where each point in the crystal has a random chance of being occupied by a Cu or a Pd atom, and ordered alloys, where the Cu and Pd atoms have a regular, periodic, arrangement.
The nub of the problem is that Pd atoms are a bit bigger than Cu atoms. Work a few years ago by this group showed that the distance between neighbouring Pd and Cu atoms increased by about 2% in dilute alloys compared to the usual Cu-Cu distance. In other words, the Pd atoms fit themselves into a lattice of Cu atoms by pushing their neighbours out a bit. This helped to explain the disagreement between theory and experiment for the electronic structure of the alloys: theory, assuming a uniform, average, distance, overestimated the contribution from electrons on Pd sites. On the other hand, smaller atoms can fit into a lattice of bigger atoms without much trouble, as happens with dilute alloys of Pd in silver, and the theory works well there.
All of this is fine for dilute alloys, where there are not many Pd atoms. With just 1% Pd in Cu, on average, all of the nearest neighbours and the next-nearest neighbours are copper, and it is quite easy for the palladium to put the squeeze on its neighbours. At higher concentrations, some of the neighbours will also be palladium atoms, pushing back. What happens then? The recent experiments studied the fine details of the arrangement of Cu and Pd atoms in ordered and disordered Cu3 Pd, where, on average, three-quarters of the atoms are copper and one-quarter palladium (figure 25). EXAFS showed the nearest-neighbour distances to be identical in the ordered and disordered phases, and the same as in the dilute alloy as well, suggesting that the Pd-Cu bond length is independent of composition. This means that the palladium atoms keep on pushing the copper atoms, and the only way in which higher concentrations of Pd can be accommodated is by distorting the lattice from the perfect cubic structure.
Figure 25: Ordered Cu3 Pd alloy
Some other interesting effects were seen. Usually, the EXAFS intensity drops off quickly with distance away from the central atom (figure 26), where the spectrum is mainly dominated by the nearest neighbours, in this case the Cu atoms surrounding the Pd, at a distance of 2.61 Å. When looking at the neighbour environment around a Cu atom (figure 27), an enormous peak is seen at a distance of about 5 Å. This is due to the fourth shell of atoms, which are directly in line with those in the first shell, the nearest neighbours (figure 25). What is happening here is that the electron scattering, which causes the EXAFS effect, is focused by the intervening atoms so that the fourth shell has a much larger effect than would normally be expected. Pd exhibits a much stronger focusing effect than Cu, so giving a much enhanced peak in the EXAFS.
Figure 26: Pd EXAFS spectrum in ordered Cu3 Pd.
Figure 27: Cu EXAFS spectrum in ordered Cu3 Pd.
What are the implications of this work? It is interesting to understand the electron scattering properties of Cu and Pd, but - more importantly - the link between physical and electronic structure should help us to explain why alloys work better than pure metals as catalysts, and, perhaps, how to design stronger materials in the future.
