Douglas Seiler painstakingly sourced papyrus and reed pens from Egypt and traditional lampblack ink from Japan, then paid some high school students to inscribe the papyrus with lines from "Star Wars" and the Bible and a quote from the '60s sci-fi TV show "The Outer Limits."
Seiler, an affiliate of Berkeley SETI , then rolled up the papyrus, placed it in a container inside a high-temperature furnace and burned it to a crisp.
The procedure replicated what happened to more than a thousand papyrus scrolls in a library in the Italian city of Herculaneum, when ash from Mount Vesuvius inundated the city in 79 C.E. and turned ancient learning — some likely unknown today — into brittle, carbonized husks. The eruption, which also buried nearby Pompeii, preserved for posterity the only known intact library from antiquity.
Seiler's goal was to determine whether it's possible to read the text on carbonized scrolls without unrolling them and reducing them to black dust. The key, he hypothesized, was that inks that contain lead would stand out when a scroll is scanned by an X-ray beam, and that computer algorithms could virtually unroll the scroll to reveal the hidden text.
In a paper published today (Sept. 16) in the journal PLOS ONE, Seiler and his collaborators, including a range of Berkeley experts, report that even minuscule amounts of lead in the ink make the letters readable by X-ray tomography, or X-ray CT. The lead absorbs up to 25 times the X-rays absorbed by the charred papyrus, making the inked letters stand out.
"With lead in the ink, you would get a huge friggin' signature, so you really need to be looking for scrolls with lead in them," Seiler said. "They're having problems reading a lot of them because of the low contrast of carbon ink on carbon paper. We're relatively certain that if they start searching for lead, or they let us search for lead, it will help this whole process. This is the Holy Grail."
In the paper, the researchers also showed that an inexpensive handheld X-ray fluorescence scanner can easily detect leaded ink in a burned scroll, providing a simple way to identify those scrolls that would be more easily deciphered using X-ray CT.
To date, only a few charred scrolls from the Villa of the Papyri at Herculaneum have been virtually unrolled and read and none of them were tested for leaded ink. Though no one has systematically looked for leaded ink in the scrolls, at least one researcher found lead in the ink of a fragment that remained from earlier efforts to physically unroll some of them. Most of those attempts followed the scrolls' discovery in 1752, but when unrolling mostly produced heaps of ashes, the Italian authorities called a halt to such efforts.
Of those physically unrolled, all have been texts unknown to scholars, many of them written by the Epicurean philosopher Philodemus, who lived in Herculaneum a century before the Vesuvius eruption. Currently, what remains of the library's 1,800 or so scrolls and scroll fragments are housed in Italy, France and England.
"Almost everything from antiquity has been destroyed," Seiler said. "There's very little left except papyrus that was in a dry climate or some works that were copied by scribes and transferred down. But some of the Herculaneum scrolls aren't copies. These are books directly from the ancient world frozen in time. They're actually from Roman intellectual circles."
"The value in this project is practically unquantifiable," said Berkeley archaeology graduate student Leah Packard-Grams, who consulted on the project as a papyrologist. "The stakes are the largest in the history of Greek literature."
As someone enamored of the ancient Greek and Roman world, Packard-Grams mourns the many lost works of philosophy, science and literature that disintegrated before they could be copied for posterity. She is now translating Greek and Egyptian writing on ancient papyri housed at UC Berkeley in the Center for the Tebtunis Papyri in the Bancroft Library. Unlike the Herculaneum papyri, these were preserved in the dry sands of Egypt. One set of fragments was written by a man named Dionysios, a school teacher and scribe-for-hire, who recorded his mundane daily activities, down to the costs of food he bought for his children along the canal.
"This is what drew me to papyrology — that I could read the words of ancient people not through a library book or through a Renaissance copy but actually sit down and know that someone wrote this letter to whomever, his father or something," she said. "It's autograph history — not just from a book, but from what they were actually writing."
Lucky for Seiler, Packard-Grams has studied the ink chemistry of Egyptian papyri in Berkeley's collection. She has hopes that the composition of ink — traditionally made from soot, water and a binder like gum Arabic — can provide a signature allowing her to identify the author of a specific text. Any metals in the ink, accidental or purposely added, would help in this identification.
The Tebtunis papyri are fragments of scrolls dating from between 300 B.C.E. and 300 C.E. and were excavated for UC Berkeley 126 years ago at the site of the ancient city of Tebtunis. Using a hand-held X-ray fluorescence scanner, Packard-Grams confirmed that some of these scrolls — generally those inscribed after the first century C.E. — are written with ink containing lead and/or copper. That bodes well for finding lead in some of the Herculaneum scrolls.
"If we know that the Herculaneum papyri have lead in them — some of them do, some of them don't — then when they take the CT scan, they can scan it with a sensitivity for the element lead and we can see where the ink is," she said. "As you unroll it, you can map the letters a lot more clearly."
Seiler also tapped the expertise of retired Berkeley chemists David Kreimer and Elena Kreimer, a retired manager of the College of Chemistry's Microanalytical Facility. With their help, he added calibrated amounts of lead — in the form of lead nitrate — to lampblack ink to get samples with different lead concentrations. Seiler asked the children of friends to write passages on a new papyrus scroll using traditional reed pens and the various lead-spiked inks. Packard-Grams and Jesse Obert of the Archaeological Research Facility scanned these newly created scrolls to confirm the lead concentration of the inks used to inscribe the papyrus.
Seiler then carbonized the scrolls in his home lab by placing them inside a semi-sealed steel container containing very little oxygen and heating it in a high temperature furnace. Using his connections at the Space Sciences Laboratory, Seiler was able to find and enlist the help of Jake LaManna, a physicist at the Center for Neutron Research at the National Institute of Standards and Technology in Gaithersburg, Maryland. LaManna gladly created a 3D X-ray CT scan of the charred scroll with the center's laboratory X-ray source.
"It's the first scroll I've ever done," he said. "I don't tend to say no to interesting projects."
LaManna jury-rigged a stand for the charred scroll and rotated it in an X-ray beam, recording thousands of slices that were turned by computer into a 3D rendering. Because lead absorbs more X-rays, the leaded ink stood out as bright spots against the darker paper, like an underexposed portion of a black-and-white negative. Ink with a lead concentration as low as 25 micrograms per square centimeter was easily detected both by X-ray CT and X-ray fluorescence.
"If you look at the images, the letters lit up like a Christmas tree," Seiler said.
Coincidentally, a postdoctoral fellow at NIST, Michael Cyrus Daugherty, had developed a program to unroll CT scans of the "jelly rolls" inside lithium-ion batteries, and LaManna asked if that algorithm would work with the scrolls.
"He got excited and ran off with the data and within a couple of days and after just a couple little tweaks to his program, he came back with examples of the scroll unrolled," LaManna said. "He just had to tweak his code to follow the uneven and changing thickness along the length of the scroll."
The Vesuvius Challenge
In 2009, Brent Seales at the University of Kentucky was the first to take micro-CT scans of a Herculaneum scroll — one owned by the Institut de France — and in 2015 he revealed a successful digital unrolling of a charred scroll from a synagogue that burned in about 600 C.E.
With that success and the consent of Italy, Seales and some venture capitalists launched the Vesuvius Challenge in 2023, promising a $700,000 Grand Prize to the first person to decipher four passages of at least 140 characters each from two Herculaneum scrolls that had been scanned by X-ray CT. It took only five months for two people working independently to decipher the first word: πορφύραc, Greek for purple. They used AI.
In 2024, the same team employed AI to decipher 15 full columns — less than one-tenth of a Herculaneum scroll that belongs to the Institut de France. It turned out to be an Epicurean philosophical treatise concerning perception and pleasure.
AI again came to the rescue in 2026, when Seales' team decoded the remains of a Herculaneum scroll, PHerc. 1667, that people had tried to unroll and had instead nearly destroyed. It appears to be a commentary on the philosophy of the ancient Greek Stoics and may date from the second or third century B.C.E., making it perhaps one of the oldest of the Herculaneum scrolls.
All of these texts were detected by virtually unrolling the scrolls and then using machine learning to look for subtle texture or morphological differences associated with the ink — an extremely faint signal.
Finding lead in the ink would be a game changer. LaManna estimates that leaded ink is up to 25 times brighter than the papyrus on X-ray CT scans.
Seiler holds out hope that X-ray fluorescence scanning of the scrolls will turn up many with leaded ink, which could then be scanned by X-ray CT to reveal many more texts, advancing the work done by Seales and his collaborators.
"What I would like to see and what I'd like to do is a much more rigorous study into the ink concentrations or what's in the ink, not just lead," LaManna said. "And see how far we could push this into making training data sets to actually make more robust algorithms for reading the real ones."
Model scrolls provide a low-risk method to develop algorithms without endangering the priceless artifacts, the researchers said.
"Doug had a really great idea scientifically with creating a model," Packard-Grams said. "Why? Because if you hurt a model, it's fine. If you hurt a 2,300 year old ancient artifact, you'll have a bunch of archeologists ready to jump you. Not to mention me jump you!"
Seiler's not sure where his research will go from here. He took up the project on a whim, having heard about the scrolls while working with Berkeley SETI on a new telescope called Panoseti , which will search for laser signals from intelligent life in the galaxy. Now retired, he previously made a living in real estate and banking and as an inventor.
"Honestly, getting here is, for me, just as unique as our research," he said. "I mean, inorganic chemistry, papyrus, X-ray tomography, AI — it's really quite an eclectic group of scientists and methodology to get to the point that yes, if there's lead in those scrolls, you guys will be able to read the images much better. I'll give you 10 to 1 odds. We'd like it to be our team, but if some other team is going to take this idea — which is okay — we don't care."
"Following Douglas' progress on the scroll project has been fascinating," said Karl van Bibber, a Berkeley professor of nuclear engineering who served as Seiler's informal advisor on everything from X-ray imaging to the ins and outs of scientific publishing. "One might be tempted to imagine the 'gentleman scientist' as a bygone era, but it's alive and well here in Berkeley."
Seiler, who funded the research, is first author of the paper. His co-authors are LaManna, Daugherty, David Kreimer, Michael McOsker of University College London and Jens Dopke of the Rutherford Appleton Laboratory in the United Kingdom.
