Two thousand years ago, Mount Vesuvius carbonized an entire library at Herculaneum — hundreds of papyrus scrolls turned to brittle charcoal, too fragile to unroll and, until recently, too uniform to read. Under X-rays, the carbon ink vanishes against the carbon page. As NIST physicist Jacob LaManna put it, “You’re basically looking for carbon letters on a carbon scroll, and there’s no contrast between the two.”

A study published Sept. 16 in PLOS One describes a promising way around that problem — and it arrived by an unlikely route: lithium-ion batteries.

Four years ago, NIST researcher Cyrus Daugherty built software that turned X-ray images of a battery, taken at different angles, into a CT scan: a 3D portrait of its tightly wound internal layers. Separately, research volunteer Douglas Seiler had been making replica scrolls from modern Egyptian papyrus, inscribing them with a stylus dipped in carbon ink mixed with varying concentrations of lead sulfate, a blend believed to resemble Roman ink. With David Kreimer at UC Berkeley, he sealed each replica in a partially sealed steel container and carbonized it in a low-oxygen furnace to mimic the eruption’s conditions.

Daugherty spent about a week adapting the battery algorithm to read the squashed, partly fused scroll layers — no clean spiral, unlike a battery cell. The software deciphered the hidden text in days. Because Seiler had written the inscriptions himself but kept them secret, the team could check its results against ground truth — something impossible with the real papyri.

The sharper contribution is chemical. Lead interacts far more strongly with X-rays than carbon does, so a scroll written in lead-based ink should be markedly easier to read. LaManna found that an ordinary X-ray spectrometer — a standard lab instrument that reads each element’s fluorescent “fingerprint” — can identify lead quickly, cheaply and without damaging the scroll. That could let scholars sort the readable papyri from the hopeless ones in minutes rather than months.

The 2024 Vesuvius Challenge, led by University of Kentucky computer scientist Brent Seales, has already used CT scans and machine learning to recover passages from one scroll and fragments of another. LaManna said the challenge’s organizers are aware of the NIST work but have not decided whether to adopt it.

Daugherty, for one, is keeping his focus elsewhere. “Although the scrolls are definitely interesting in their own ‘Indiana Jones’ kind of way, I’m an electrochemist, and my primary research interest is batteries,” he said.

Sources: NIST | Live Science | PLOS One | Gizmodo