If you are an archeologist looking at a new site, one of the first things you want to know is how old it is. Radiocarbon dating can answer this question for organic objects that contain carbon, but carbon can be quite rare as organic material gets eaten.
One thing that is very common in almost all archeological sites is pottery. It is easy to make and cheap, it breaks easily and you can't recycle it, so there was a lot thrown away and it lasts thousands if not millions of years. So it gets everywhere, unfortunately there is no carbon in it and it is very difficult to date.
However Moira Wilson and colleagues may have come up with a solution. When you make pottery, you fire it. You heat clay up to a temperature between 1000 and 1400 Celcius which sinters it, causing the particles of the clay to stick together and crucially drives the water out of some of the minerals which make up the clay.
Then as soon as the clay cools down a very slow reaction between these minerals and water starts, and the team is using this reaction to date the pottery.
It is easy to measure the amount of water in the minerals of the pot, you just dry the pot out normally to get rid of the water in amongst the grains, then cook it at 600C for a few hours and measure the difference in weight.
Different pottery takes up water at different rates, but the rate at which it starts taking up water for a couple of days after it is dried out predicts how it will take up water over the next few hundred years very accurately, and crucially as Moira told us:
The reaction is sustained by an incredibly small quantity of water so there's actually sufficient moisture in the atmosphere to keep the reaction going. So it doesn't actually matter whether your brick is sitting on the table, or sitting on the bottom of a lake. As long as there's enough water there to sustain the reaction, any excess water, for example if the material is saturated it doesn't contribute to the reaction it just sits there doing nothing.
In fact the only thing that does affect the rate of uptake of water seems to be the temperature, which is going to be reasonably constant over a fairly large area, so if you can take this into account, by measuring a few pots you know the ages of, you can make remarkably accurate predictions.
They Dated brick from a Charles the second buildig in greenwich which was built between 1664-1669 and altered 1690s as dating from 1691 ± 22 years.
A roman brick was dated as 2000 years old, and is known to be 2001 years old.
They had more problems with a medieval brick from canterbury which they repeatedly dated to be 60 years old. but it turned out that during the blitz there was a major fire in this area, essentially refiring the brick and resetting the timer
If this system turns out to be as good as it promises to be it should be able to date thousands of sites which are so far undated giving us a much more accurate view of the past..