Monday, August 30, 2004
Water changes phase (from liquid to gas) at 100 degrees C (assuming atmospheric pressure). What's interesting about that phase change (and even more so for phase changes of other substances - water is unique for all sorts of reasons not worth getting into at the moment) is that at these temperatures and pressures small changes in the environment can cause large changes in the macroscopic properties of the system. For example, small changes in the pressure in a vat of nearly boiling water can cause most of the water to boil off (more or less the idea behind a pressure cooker).
So phase changes are interesting in general, then, because they're times when small changes in the environment can lead to big changes in materials properties. If you want to design, e.g. concrete which cracks less readily in cold, plastics which are easier to shape or, as it turns out, steel which better resists rust, you need to understand what's happening at the relevant phase change and how to design your system to either inhibit or encourage change (depending on the specifics).
But back to corrosion. Prior work has demonstrated that viritually all metals form a protective oxide film when exposed to atmospheric conditions (because of reactions with water). Corrosion happens when holes are made through that film which rapidly grow and extend, downwards, into the underlying metal. Prior work has also demonstrated that under 'normal' conditions this protective film has small metastable pits. These pits typically appear and vanish on fairly small time scales, usually showing up at defects in the underlying crystal structure of the metal. Punkt et al. image metal surfaces under small changes in environmental conditions (small changes in temperature/salt content in the overlying water) and find that corrosion occurs when a critical density of the metastable pits occur: when the ms pits are close enough to feel each other. The observed sensitivity of corrosion to environmental conditions, and the pattern which corrosion follows once it starts can all be predicted (at least qualitatively) by a simple model of the process.
Why is this important you might now ask? If you're interested in a car which doesn't rust as easily or any other sort of metal part which works better when exposed to the environment, this knowledge is key. Now you know that one of the ways to beat corrosion is work really, really, hard at making your product defect free (so the ms pits occurs much further apart in the first place).
Unfortunately accessint the article requires a subscription to Science. For those so inclined the full reference is...
Punckt et al., Sudden Onset of Pitting Corrosion on Stainless Steel as a Critical Phenomenon, Science 2004 305: 1133-1136
Two of the groups involved in the research do have easily accessible web pages at which you can find out more about the sort of stuff they study:
-> Hudson Group at UVA
-> Mikhailov Group at the Fritz-Haber Institute