It seemed like the kind of discovery amateur archaeologists dream of making. In 1998 two of them declared that they had found objects made from gold sheet, apparently from a crown, a belt and other ornaments, in the soil under uprooted trees at the hamlet of Bernstorf in Bavaria, Germany. The site had been known since the early twentieth century as the location of a Bronze Age settlement, but much of it had been destroyed by gravel mining. Here, though, were precious artefacts disclosed by that very ground clearance.

Or were they? An analysis in 1999 using X-ray fluorescence (XRF), led by Rupert Gebhard of the State Archaeological Collection in Munich, showed that the gold was unusually pure. Natural gold typically contains significant amounts — up to 40% — of silver, as well as other trace elements such as copper. But the XRF measurements revealed less than 0.2% silver1. Subsequent, more precise studies by Ernst Pernicka at Mannheim using mass spectrometry2 showed that the gold was at least 99.99% pure — an astonishing degree of refinement hitherto known only in modern samples.

So are the Bernstorf artefacts fakes? Pernicka thinks so. But it's hard to see how amateur archaeologists could achieve so sophisticated an act of deception, and there are no other features of the objects that make them suspicious in the context of Middle Bronze Age culture. Then how could they be so pure?

Maybe the measurements were wrong. After all, the mass-spectrometry results showed some trace-element discrepancies with other earlier studies3. But now a new XRF analysis using intense synchrotron radiation from the BESSY source in Berlin, led by Martin Radtke in collaboration with Gebhard, has confirmed that the Bernstorf gold indeed has a purity of more than 99.9%, comparable to a modern sample produced by the Degussa chemicals company4.

Bronze Age gold purer than that in nature is not in itself a complete mystery. A technique for stripping silver impurities from gold, which involves heating foils of the metal with salt at around 800 °C for several hours, was known at least in the sixth century BC, when the earliest archaeological evidence of it appears. Known as cementation, it is described by Pliny in his Natural History from the first century AD. It's not impossible that it was known in the Bronze Age.

Credit: PHILIP BALL

But would it do the job this well? Roman gold coins purified by cementation — it was the introduction of gold currency that made the process necessary — are no more than 99.8% pure. It seems possible that greater purity could have been attained by several repeated rounds of cementation. But there is no obvious motivation for that. Most silver is removed in the first go (it is volatilized as silver chloride), after which there is no discernible further change in the colour of the gold — the general means of assessing its purity in the ancient world. So why would Bronze Age metalworkers have bothered?

At this stage, then, the Bernstorf gold is still a puzzle — and a reminder that sometimes not even state-of-the-art science can definitively answer questions about the past. However, the onus is now on advocates of the artefacts' authenticity to explain how this could be possible.