The Strike in the Glass

The Rothschild Mystery

Mara: Sheffield, October 2026. Lisa Hollands, a glass technical specialist at the University of Sheffield, is talking about a glass she's been trying to make from scratch. It copies something Roman. “People have tried it before with limited success,” she said. She called the project “a labour of love.”

Eli: A listener asked us: how did the Romans make glass that changes colour, and why did it take 1,600 years to copy it? Today's angle is a modern furnace trial testing an ancient empirical secret.

Mara: So we follow Hollands. From a cup in London, through a recipe nobody kept, into her furnace. And along the way there's a fight that isn't settled, over whether the Romans meant to make this glass at all.

Eli: The cup's in London. One visitor called it a bit difficult to find among all the other treasures in its gallery.

Mara: Worth finding, though. It's the Lycurgus Cup, at the British Museum. Roman, generally dated to the fourth century, so about 1,600 years old. It's a cage cup, the glass cut into an openwork shell around the bowl. And it's dichroic. Lit from the front, it looks jade green. Lit from behind, it glows ruby red.

Eli: Two colours from one glass, then. One bounced back, one let through.

Mara: Exactly. And it's the only complete Roman vessel that does it. Other cage cups survive from the period, but this is the only one with the effect. The rest of ancient dichroic glass survives as a few small fragments.

Eli: And for a long time nobody was even sure it was glass. In 1950 its owner asked the British Museum to investigate. The first full studies settled what had been argued without resolution: glass, not a hard stone.

Mara: A gemstone could have explained the colours. Plain glass shouldn't.

Eli: Except it isn't quite plain. The body is soda-lime glass, the everyday kind made from sand, soda and lime. But in 1990 two researchers looked at it with an analytical electron microscope, an instrument that uses a beam of electrons to detect microscopic particles. They found colloidal gold and silver: metal particles spread through the glass, thousands of times smaller than a grain of sand.

Mara: So the colour lives in the metal, not the glass.

Eli: In how the metal handles light. Particles that small absorb and scatter it selectively. Light bounced off the surface comes back green. Light pushed through the wall loses other colours and comes out red. The modern name is surface plasmon resonance: electrons in the metal moving together, soaking up some colours and throwing others back.

Mara: And knowing what's inside never told anyone how to melt it. That's the problem Hollands took on.

The Lost Secret

Eli: Somebody once knew how. The knowledge of making this glass was lost after the fourth century, apparently after about a century of use. Specialized glassmaking knowledge often passed from master to pupil, kept as the equivalent of a trade secret. That makes old techniques hard to trace at all.

Mara: And the 1990 microscope showed the ingredients, not the method.

Eli: Right. A glass of similar colour was made in the 1960s, according to Ian Freestone, an emeritus professor of archaeological materials at University College London. Similar, not the same. Sheffield's own account says many tried to reproduce the effect with varying success, and a consistent, repeatable recipe long proved elusive.

Mara: And in 2020 a Wageningen team matched it, in 3D-printed plastic, not molten glass.

Eli: Which is the corner Hollands worked from. She'd studied and worked with glass for over twenty-five years, and paired that with a modern scientific understanding. Then she added: “if the Romans could have seen me experimenting, they’d probably laugh and ask: ‘Why does she not know how to make this?’”

Mara: Unless some of them didn't know either, and simply got lucky.

Eli: That's the live argument, Mara. Recipe or luck, and it decides what Hollands was even trying to recover.

Fluke or Formula?

Mara: Hollands took this on knowing it might not work. And here's what nobody knows: whether Roman glassmakers understood that tiny particles made the colour. There's no evidence they knew the physics. RSC Education reads the cup as deliberate all the same: Roman makers with a precise recipe, worked out by experience, for adding traces of gold and silver. Ian Freestone reads it the other way. He says the colour was quite dependent on conditions in the furnace. Put the glass in a different spot and it came out differently.

Eli: In his words: “It makes me think that, with the Lycurgus glass, they hit the jackpot.” The International Institute for Nanotechnology says the cup's properties were most likely created by accident.

Mara: A jackpot means long odds. How long?

Eli: A 2026 study, physics plus machine learning, modelled the gold-silver mixes. Only about 0.05 percent of the ones tested matched both colours at once. That's roughly one in two thousand. Call it the composition window: the thin slice of metal recipes that gives green one way and red the other. On the accident reading, that's a measure of how narrow the target was.

Mara: And the cup wasn't the only time someone hit it. It's the only one that survives whole, but fragments turn up. Pieces dug up at Dülük Baba Tepesi, in Turkey, show the effect wasn't confined to this one cup. And the very few surviving fragments vary considerably in their two colours. Same trick, different results.

Eli: Which fits Freestone's view: furnace position, not a fixed recipe, set the shade.

Mara: On RSC Education's reading, those same fragments come from makers working to a recipe. On Freestone's, they're the odds showing through. What would move the argument is a furnace that lands the cup's colours on purpose.

Eli: That's the test Hollands set herself in Sheffield.

The Furnace Breakthrough

Mara: She started with a Roman glass expert, researching what was known about glass of the era. Then she built a base glass from modern materials, mimicking what the Romans might have used. Then the hard part: getting small amounts of gold and silver into the glass in the right form. She tried different furnace types. Different starting materials. Different oxygen levels. She even tried adding sugar.

Eli: Sugar. And not one ready-made nanoparticle on the list.

Mara: That was the point. She'd said it was “not just a case of adding nanoparticles to glass.” And then a recipe worked: both of the cup's colours, in her glass. It did better in a gas-fired furnace than in electric equipment.

Eli: So the particles that 1990 microscope found weren't put in. They formed.

Mara: They form as the glass cools. Hollands calls it self-striking: the colour switches itself on. “The amazing thing about this formulation is that the glass self-strikes,” she said. No complex heat treatment to grow them. They're produced “simply by annealing to relieve stress,” the slow cool that keeps glass from cracking. Her verdict: “To me, it’s magic.”

Eli: The Roman cup may have had its own helper. Its glass contains antimony. On the deliberate reading RSC Education holds, that was probably added on purpose, to clear the glass and act as a reducing agent, something that coaxes dissolved metal back out as metal.

Mara: And her answer came from the furnace and the air inside it, the same lever seen in those uneven fragments. Earlier tries had limited success. Freestone says Sheffield's glass is much closer to the cup's real green and red than the 1960s version.

Eli: Still a slab of glass, though. The Romans made a cup.

Mara: So the team brought in a glass artist, a winner of Netflix's Blown Away, to make a sculpture from Hollands's recipe in the Sheffield melting shop. He wanted it hot-worked, blown and shaped while soft from the furnace. In his words, “we needed a blown element within this to really show the dichroic glass.” It's due to be unveiled at Sheffield's Millennium Gallery in November 2026, then kept for good at the university's Turner Museum of Glass.

Eli: Showing it off was a Roman concern too. Behind the cup's figures, the glass is hollowed out deep, so it's about as thick as the wall and the colour stays even with light behind it. That care is part of RSC Education's case.

Mara: Hollands joked the Romans would laugh at her. Now a modern maker is working their glass around the same problem: getting the light through.

Eli: So, the listener's question. The Sheffield team believes the Romans added tiny amounts of gold and silver, and under the right conditions those metals formed layered nanoparticles inside the glass. Whatever the Romans knew was workshop knowledge, worked out at the furnace. Sheffield paired furnace trials with a modern scientific understanding of the particles. That's the gap between ancient and modern glass chemistry.

Mara: And the 1,600 years?

Eli: There's no record of sixteen centuries of people trying and failing. The craft vanished after antiquity, a lost recipe that lived only in the hands of the people making it, and then went unstudied until modern times. So the gap reflects a lost tradition, and how hard it was to get a modern furnace to do the same thing.

Mara: So the labour of love Hollands described is real glass now, headed for a gallery. That one-in-two-thousand window, Sheffield hit on purpose. Whether a Roman workshop did is still the argument.

Eli: Two things to watch. As of October 2026, there's no peer-reviewed paper setting out the Sheffield recipe, particle sizes or colour measurements, so outside chemists can't yet check how close the match is. And a fourth-century workshop, its records or its furnace, would be the find that says whether the Romans meant it.

Mara: One topic for further reading: Roman dichroic glass.

Eli: Listeners can follow science with us: what can modern researchers still learn from ancient crafts?

Mara: That's our story for today. I'm Mara.

Eli: And I'm Eli. Thanks for listening to Angle, and until next time.