Dating by Water Clock

Meera -   This is the Naked Scientists with Meera and Diana.  Now, we're talking of things thermal.  Here's a story on fired clay.  Now normally, ancient bits of pot and brick are dated using carbon samples or thermoluminescense.  But these techniques require lots of sampling and calibrating which takes time.  So, it was quite a surprise when construction scientist, Moira Wilson, came up with a really simple method of dating ceramics...

Moira -   I think the problem is, that most people think that carbon-14 dating can be used to date anything.  And in fact, it can't.  It can only be used to date things that have been living like, you know - plant material, bone, or whatever that's been exchanging carbon with the atmosphere throughout its lifetime.  So you can't carbon date everything.  That's the first thing and you certainly can't carbon date pottery.  Thermoluminescense - that can be used to date pottery, this is actually quite a complicated thing.  You need to excavate quite a large volume of materials from around the fragment of pot during the archaeological dig because you need this material to get a background count for calibration and thermoluminescense is calibrated against the carbon-14 dating curve.  Now, our method is self-calibrating and doesn't require any material from around the excavated site.  So, we can actually use it retrospectively on a lump of clay or a lump of pot or brick or whatever that's been sitting in a museum.

Diana -   And how did you go about discovering this technique?  What gave you the inspiration in the first place?

Moira -   Right.  Well it sounds a bit of a weird thing to do, but they say that chance favours the prepared mind.  What we were actually working on at the time was moisture expansion in clay masonry and this is an engineering problem really in masonry construction that bricks expand.  This is why you have expansion joints in walls, for example.  So, as soon as the bricks come out of the kiln, they start expanding and carry on expanding forever.  And this is why in general practice, you know, bricks would never be used straight from the brickyard.  They'd be left for a few days.  So, we started off, looking at the engineering problems.

Diana -   So basically, you were trying to find a way of being able to predict how these bricks would behave after they came out of the kiln?

Moira -   Yeah, but what we were looking at was moisture expansion in clay brick as in walls.  In the process of doing so, we discovered a new law and we found that these materials expand as the rate that's proportional to the fourth root of time.  For example, if you have a kilometre of wall that's about 200 years old, it will, in that course of that 200 years, expanded by about a metre.

Diana -   Wow!

Moira -   Yeah.  But that doesn't mean the wall gets longer by a metre because what happens is that their expansion of each individual brick is accommodated by the mortar joints getting squashed up.

Diana -   All right then, so what are the actual particulars of the technique that other archaeologists could do themselves eventually?

Moira -   Right, so, how it works?  As I say, the dating method relies on this new law.  But fundamentally, what it relies on is that, as soon as any fired clay product is removed from the kiln, it starts gaining weight and expanding caused by the reaction between the ceramic and atmospheric moisture.  And this, as we say, carries on forever.  So, if we take the example of a whole brick, we would weigh the brick first of all and the weight now, is its original weight as it came out of the kiln plus the weight of the chemically combined water that's reacted with it over its lifetime.  Then we heat the material to several hundred degrees to drive-off this chemically combined water.  And then we wait again, so we can establish from that how much water it has gained over its lifetime.  And in the case of a Roman brick, say, that would be about 40 grams.  So, once we've driven all these chemically combined water off, it starts to recombine again.  So, it starts all over again.  So, we monitor this for a few days and it obeys our t^1/4 (t-to-the-quarter) law so the data come out as a nice straight line.  And from this, we can establish the rate at which the material is reacting with water.  And if we know how much it's lost, we can work out how long it's going to take to gain that much water again.

Diana -   So, each measurement is specific to a different type of ceramic?

Moira -   It is and it's self-calibrating.  So, we don't need any other external calibration data for it at all.  It doesn't matter what the -the clay is made of anything and it very simply you see.

Diana -   Can sort of ambient humidity change the result at all?

Moira -   Nope! That is the other nice thing about it. This chemical reaction is incredibly slow.  It's proceeding in increments of 1, 16, 81, 256, minutes, days, decades, whatever.  So, the reaction rate, very fast to start and it slows down very, very quickly.  And 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 it's sitting at 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.

Diana -   Okay, and how accurate are the dates do you think that you could get from say, a bit of pottery, that's maybe, something around 2000 years-old?  I mean, how accurate if it was that old, what would your date be?

Moira -   Well, I think the biggest inaccuracy that we've got in our results is the supposed known age of the sample in the first place.  I mean, we don't know that the brick was fired on the 5th of July, half past two in the afternoon or something like that, you know.  We just know that it's Roman, for example.  So, if we can date a Roman brick to 2001 years, that's great and it's certainly in the right ballpark.  But as far as the magnitude of it goes, say we're looking at a Roman brick, we don't measure whole bricks, we measure 4-gram pieces.  So, in the course of 2000 years say, we'd expect the 4-gram piece to have gained about 68mg in mass.  Now, if we are measuring to a tenth of a millionth of a gram, that's actually a lot of units of measurement that are available to us.  So, really, I suppose what you're saying is, how accurate are all the components of this. The mass measurements is virtually error-free, we would say.

Diana -   Fantastic!  You mentioned that you used a 4-gram sample, just like a starting point but is that the smallest you can go to or could you take a really small scraping and work from that?

Moira -   Well, actually, given that we started off measuring the mass gain in whole bricks, which is about 2.5 kilograms, and then we move down to four grams, which is about half a size of a sugar cube, and now, we're down to milligram-size pieces.  So, we're getting the sample size down all the time, which is an advantage.

Diana -   Wonderful.  And apart from archaeology, how else could this technique be used?

Moira -   Well, because the mass going in these materials follows this precisely defined kinetic law, and because it's also temperature-dependent, we could, if we knew the precise dates of which something that's fired.  And there are instances of this like Pompeii for example, was immersed in pyro-clastic ash in AD 79.  So, we know the precise dates at which all the ceramics in Pompeii were heated up.  We could then turn the method on its head and because we know the exact firing date, we could work out the average temperature that the material had been exposed to, over this precisely known lifetime.  So, there might be a bit of climate change potential there as well.

Diana -   So, what's the next stage in research in this technique?  Are you going to apply it to objects you know the dates of first or are you just going to sort of have a play to see if you can debunk any other previous dates that have been taken on pottery?

Moira -   It's going to be having a play really.  We have this fantastic overwhelming international response from the archaeological community to this work.  I mean, we're really shocked actually.  So, the next stage of the work really is to validate the method.  I mean, suppose you could certainly have already validated it.  We proved the concept.  What we need to do now is to work out a standard methodology for carrying out these measurements so that other people can do it.  We also need to extend the time range of the samples that we've measured.  We've only measured things to 2000 years old so far.  So, we really like to look at some older samples and we've heard in our correspondence with archaeologists that pyro-technology apparently goes back to 16,000 BC.  So, if we'd be able to get our hands on a piece of pottery that was 18,000 years old, we'd be really happy.

Diana -   It seems too good to be true and the team were actually thrown to start with as they were given this medieval brick to test their theories on.  And they just couldn't work out why the date came out as around 1939, but it turned out that the house in which the bricks had been stored was bombed during the second world war.  So, the brick itself was actually re-fired, giving a much more recent date.  So, you might not want to use this method on the bricks at Pompeii!