Making Metals: Designing A New Alloy

Dave - Metal alloys are mixtures of different chemical elements. Adding certain elements can make a metal harder, others can alter melting point or help the metal to resist corrosion. But there are tens of metals to choose from, so how can we work out which to include and in what proportions? Professor Roger Reed, Director of Research at the School of Metallurgy and Materials at Birmingham University, joins us to explore this. So Roger, what's the traditional way of working out how to make an alloy?

Roger - Well, traditionally of course, a lot of so-called "bucket chemistry" has been used to find out which elements to use for the alloys. So you would typically mix up elements such as nickel and chromium, and aluminium and titanium, and so on for the nickel based superalloys, and then you would test them in mechanical testing apparatus, and work out whether one has adequate properties for a given application. So, for example, in the jet engine, you might be interested in making sure that the alloys have got sufficient strength or toughness, or even something as complicated as creep resistance. So a lot of empirical studies have been done by 'make test' iterations. You make the alloys, you test them and perhaps up to maybe a few hundred alloys, you have one which you think you can use.

Dave - That sounds like a hideously time consuming and expensive process.

Roger - Well, that's right. Many people spend lots of time and of course, lots of money. For a typical turbine disk for a jet engine - it may be up to a million pounds or perhaps even ten times that to have the alloy ready for use in the engine with all of the necessary test data measured carefully.

Dave - So, what are you doing which is different?

Roger - We've been working on computer modelling programmes to try to use theoretical methods to come up with what we think are the optimum compositions for the sorts of components for jet engines. So the idea is that you would do the necessary theory in calculations on the computer and then to choose the alloy composition which you think is best matched to the properties that you need for these applications.

Dave - So, are you starting off at the atomic scale with quantummechanics or doing something slightly different?

Roger - There are different ways to do it. So yes, certainly we make use of quantum mechanical methods; so-called density functional theory where a number of assumptions solves Schrödinger's equation to work out the bond energies for the different elements and in particular for the energy associated with the different elements swapping positions on the crystal lattice. So that's certainly something we do. We also use high level thermodynamical methods based on solution theory which have got now parameters in there which describe how the bonding behaves.

Dave - So essentially, you're using quantum mechanics to get figures which you then feed into another larger scale model which you then feed into another one?

Roger - We come up with what we call merit indices and those merit indices are figures of merit which describe how any given trial of synthetic composition would behave in creep, or in fatigue, or in strength. And we do that by making estimates for suitable defect energies which describe how easy it is for that alloy to deform. It comes down to a number of composition dependent energies which we can estimate using the theory that you just described.

Dave - I guess you're not just dealing with two metals in an alloy. You're dealing with many. Does this add more difficulty?

Roger - Yes, that's quite right. The kind of alloys that I look at in my research are the nickel-based superalloys. Those are the alloys which are used in the very hot parts of the turbine engine, of the type that you would use for a jet propulsion or for electricity generation in a power generation circuit. Those alloys are quite complicated. They typically have maybe 7, 8, 9 - maybe as many as 10 different alloying elements. We add chromium, for example, and cobalt. The chromium is added for much the same reason that you would add chromium to a stainless steel - to make it oxidation and corrosion resistant. We add also elements such as aluminium and titanium to promote precipitation strengthening. That makes the alloy start to behave a bit like a composite, so you have small particles of a second phase present which adds strength to the materials. Then we add other exotic elements, much rarer and more expensive refractory elements like tungsten, candidum, and even rhenium. And even some precious metals like ruthenium have been shown to confer very best properties particularly in creep, where one is interested in making sure that the alloys don't deform and stretch continuously over time under the stress that they experience in the engine.

So yes, there's quite of range of elements which are added, and this does provide the complexity which makes the computer modelling methods useful because if you think about it, each element needs to have a particular concentration associated with it. The concentration ranges are quite wide and with so many elements there, perhaps 8 or 9, or 10, the number of combinations that you can come up with synthetically is quite large. That's the power of computer modelling methods; we are a适合于do that sorting and ranking, and prediction on the computer without the need to go to the laboratory to make the alloys and do the testing. That does two things. It reduces the cost, but it also allows us to identify the compositions quickly. In a market where you need to get new alloys into the engine quickly, so that new products can be flown within perhaps the period of 1 or 2 years, rather than the traditional 8 or 9 years, that's very, very useful.

Dave - Are you fine tuning principles which people have worked out by the rough and ready system or are you working out new things which can go into the mixture?

Roger - New theories have been needed, yes. Of course, we are building on the shoulders of the giants as they say. Lots of people have worked on these sorts of alloys over the last 20, 30 years and of course, they can traced back to Whittle's very first engine which flew round about the time of the second World War. But theory has been needed to explain and to put in place composition dependent theory for the important properties such as strength, creep and toughness. I think in the last 5 or so years, that theory has matured and that's really what's allowed us to propose these new computer modelling methods to do the alloy design using theoretical methods.

Dave - So once you've got one of these, once you've run your models, what do you pass on to the guys that actually got to make the things?

Roger - Typically, we would suggest alloy compositions which we believe to be useful. That would mean the exact concentration in either weight percentage or maybe atomic percentage - percentage of the different elements there. So for example, for a turbine disk alloy, we may specify perhaps 15 or 16 weight percent of chromium in the latest alloys. You may then have smaller amounts of aluminium and titanium, tungsten and so on. We specify the compositions which we believe are required to lead to the very best properties.