Tag Archives: Roman glass

The Jewels Of Transition

Chemical analysis of Iron Age glass beads, found in North-East Scotland, shows that the jewellery was made from recycled Roman glassware

Related: a great post from Grrl Scientist about the colours of cobalt

The soils around Aberdeenshire and Moray, in northern Scotland, are resplendent with Iron Age glass jewellery – deeply colourful,
intricately patterned beads which stand amongst the ancient world’s finest
examples of craftsmanship. The beads come in many translucent colours – shades of blue, green, purple, red and yellow. And on their surface, they carry swirls of opaque colour.

But they’re something of a mystery – no-one really knows who
made them. Southern Britons of the time did make glass beads, but a preliminary analysis of the Scottish beads showed that they were not southern British.  So, where did they come from?

Now, a team of UK scientists have used laser ablation technology to unravel the history of those jewels. It’s a story that takes us to the sands of the Levant and the city-states of ancient Rome, before finally depositing the beads in the soils of the eastern Scotland, to be uncovered millennia later.

The UK team, led by Martina Bertini of the University of Aberdeen, decided to explore the trace element composition of the beads. They hoped that by exposing the underlying chemistry of the glass, they could find out who made the beads – and how such intense, striking colours were produced.

So, they subjected nineteen of the beads to a technique called Laser Ablation Inductively Coupled Plasma-Mass Spectrometry. It’s a gobful, yes, which is why scientists – ever fond of their acronyms – call it LA-ICP-MS.

Now, we’ve met ICP-MS before – it’s a way for scientists to identify chemicals in a sample by measuring their molecular mass. And Laser Ablation is simply a way to get some material off the hard glass beads and into the ICP-MS. As the name suggests, it uses a laser to bash off a bit of the bead material. Those fine glassy particles are then swept up in a stream of argon gas, and into the ICP-MS – which will take a look, and tell us about the beads’ chemistry. Specifically, it will tell us about the beads’ trace metal content.

So what did Bertini and co. find when they analysed their results?

Well, the sand used to make the glass, it turned out, bore all the chemical signatures of a Middle Eastern origin – the chemical element fingerprint of the glass (the relative amounts of elements like rubidium, zirconium and gallium) matched that of other glasses which are known to come from the Levant region, near modern Lebanon. And, because of the glass’s strontium content, which reflects a high level of marine shell in the sand, Bertini’s’s team think that the sand came from a coastal area, not an inland desert.

So, the sand used to make the glass beads came, most likely,
from the banks of a Levantine river – and not from the cold beaches of
Scotland. But there was, plainly, more than just basic glass in those
beads – the translucent base colours, and the opaque surface decorations, were clearly created using some kind of pigment. And the LA-ICP-MS analysis showed, quite definitively, that the colours of the glass had been produced using Roman chemical recipes.

Roman glassworkers used transition metal ion complexes – yes, another gobful, but it’s a simple concept – to make the colours. Transition metals are a special class of metal, and their defining feature is, in a way, incompleteness. Electrons around a nucleus can inhabit several levels, or orbitals: a bit like rows of seats around an amphitheatre. In transition metals, one of these levels – the d-orbital – is incomplete. It does not have a full set of electrons. That incompleteness leads to all kinds of strange chemical properties –their striking colours, for example.

So Romans coloured their glass by adding these transition metal ions to
the sand mixture, before heating the whole concoction in a furnace. And the eventual colour of the glass could be adjusted, using a little bit of chemical knowledge. Let’s take iron, as an example. If you oxidise iron so that it loses three electrons – the trivalent state, in chemical terms – you get Fe3+.
This gives a nice, subtle amber colour to your glass. If the iron is reduced to
the Fe2+, or divalent, state, then you get the blue-green shade which is more typical of Roman glass.

Not that the Romans knew about electrons, of course – it was
simply trial and error. But the results could be spectacular. To make an emerald
green bead, Roman glassmakers mixed a bit of oxidised copper in with their iron.
Oxidised copper on its own, though, would give a nice shade of medium blue. Reduce your copper a bit, and you’ll get a shade of red. Manganese, when it has been oxidised to lose three electrons, gives a purple colour.

The team recognised these chemical recipes, uncovered by their analysis, as being classically Roman.

So what can we say for sure about how these beads ended up in Scotland? Well, the story is still incomplete. Even though we now know that the ingredients came from the Middle East, and the beads were made to Roman specifications,
we’re still not exactly sure who made them. Archaeological evidence, though,
points to an enterprising local Scot. It looks as if the beads were, in fact, recycled
– made from Roman glassware that had reached Scotland and then, perhaps, been broken or was otherwise unwanted. So someone melted it down, and used it to make these iridescent beads.

In a way, each bead is a microcosm of the Roman empire of the time – throughout all those years of lying in the damp Scottish earth, waiting
to be discovered, they were holding inside them a chemical memory of the hot
winds of the Levant, and of the chatter of ancient Rome. To find out more,
check out Martina Bertini’s article about her research, here: http://www.leopardmag.co.uk/feats/305/puzzle-of-our-iron-age-beads


Bertini, M. et al (2011). Investigation of Iron Age north-eastern Scottish glass beads using element analysis with LA-ICP-MS. Journal of Archaeological Science 38 (10)  pp 2750 – 2766 doi:10.1016/j.jas.2011.06.019