One of the Herculaneum scrolls scientists are hoping to read with x-ray technology.
(Credit: Diamond Light Source/University of Kentucky)
Two thousand years ago the heat of a volcano burned the contents of a library in the doomed Roman city of Herculaneum. Today, a team of researchers from the University of Kentucky, working at the UK and STFC’s x-ray synchrotron Diamond Light Source research facility, are using the light of a billion Suns to hopefully reveal the writing locked away within the ancient charred papyri.
In October 79 AD, one of the most luxurious villas to be found in the wealthy Roman coastal town of Herculaneum on the coast of southern Italy lay abandoned. Among the villa’s buildings, which stood atop four highly-manicured and well-appointed terraces that cascaded towards the sea, there was one of the finest libraries of its day. Within the library were thousands of papyrus scrolls that had been hastily packed into cases – ready to be carried to safety from the chaos that had been unfolding outside.
For weeks, the citizens of Herculaneum (and the nearby town of Pompeii) had ignored the earth-shaking tremors that had portended the approaching natural disaster. This was a time in human history when the Roman Empire stretched across most of Europe and North Africa. However, that meant nothing when a little after 1pm, the mighty Mount Vesuvius unleashed its pent up fury and erupted – hurling millions of tonnes of volcanic material thousands of metres into the air. As the ash began to fall, the inhabitants of Herculaneum started to evacuate their homes. In the villa, it’s occupants had scrambled to save household treasures and the wealth of knowledge contained within the library’s papyri but, as the situation deteriorated, the would-be rescuers abandoned their tasks and fled for their lives.
It is just as well they did because, at about 1am, the column of volcanic material, which had risen high into the stratosphere, suddenly collapsed down onto Vesuvius’ flanks and a series of pyroclastic flows of super-heated gas and ash surged towards Herculaneum at more than 100 miles per hour.
A fragment of one of the the Herculaneum scroll that will be used to help computer programmes to read the text hidden within the scrolls.
(Credit: Diamond Light Source)
When the surges hit, buildings like the villa became buried in layers of ash and the hot gases (still at temperatures of 250ºC) turned the library into an oven in which the abandoned scrolls were flash-heated and transformed into fragile rolls of carbonised papyrus. By the time it was all over, the library lay buried beneath more than 20 metres of ash that entombed the building and slowly became compacted into a type of volcanic rock called tuff – preserving the scrolls within.
Nearly seventeen centuries later, in 1709, long-forgotten Herculaneum was rediscovered and excavations began. In 1752, the opulent villa (dubbed the ‘Villa of the Papyri’ and now thought to have belonged to Julius Caesar’s father-in-law) was also discovered, along with the library, which is the only one to survive intact from antiquity. Unfortunately, after some initial digging (during which many papyri scrolls were found and some even then thrown away), the site was then lost again for another two centuries until it was rediscovered in 1980.
In all, some 1,800 scrolls were eventually excavated from the site and housed in the Biblioteca Nazionale in nearby Naples (where the majority are kept), along with the Institut de France in Paris, the British Library in London, and Oxford’s Bodleian Library.
For more than two hundred years, scholars have been attempting to read what remains of the heavily-charred scrolls, which resemble lumps of coal or burned logs more than they do rolls of papyri. But the carbonised papyrus is so delicate and brittle, any attempt to unroll the scrolls causes irreparable damage and, once exposed to the air, the ink can quickly break down, fade and become unreadable.
And then there is the problem of differentiating between the ink on the papyrus and the papyrus itself. In the past, scientists have achieved some success pulling words from within rolled, or damaged, artefacts, by looking for presence of metals, such as iron, in the ink. Unfortunately, the ink used on the Herculaneum papyri is made up of ash (carbon) mixed with water, which is a problem because the carbon-based ink is almost indistinguishable from the carbonised papyri.
So how do you even attempt to read something that is illegible and is locked away in a complex matrix of tightly rolled, blackened and compacted material that is too fragile to touch? To answer this question, two intact scrolls and four smaller fragments have found their way to Diamond Light Source in Oxfordshire where a team of scientists are hoping to be able to ‘virtually unwrap’ the scrolls to reveal their secrets.
The team, led by ancient artefacts expert Professor Brent Seales from the University of Kentucky, have devised a two-stage method involving artificial intelligence and x-rays to achieve this seemingly impossible task.
The team ready a papyrus fragment for examination. (Front row – left to right) Jens Dopke, Brent Seales, Francoise Berard, and Christy Chapman. (Back row – left to right) Robert Atwood and Thomas Connolley.
(Credit: Diamond Light Source)
The first stage requires the use of the UK’s research facility Diamond Light Source – a state-of-the-art synchrotron that can probe objects with beams of x-ray light, 100 billion times brighter than the Sun. These super-bright x-rays can penetrate the scrolls and map their internal structure in a completely non-invasive and non-destructive way. The x-rays pass though the scrolls and are detected on the other side to create a series of two-dimensional images that can then be stitched together to create a three-dimensional model of the scroll.
STFC’s senior detector scientist Dr. Jens Dopke provided the Kentucky team with the technical expertise that allows them to use Diamond’s beamline to gather data from the scroll samples.
Dr Dopke said: “With Diamond Light Source, we get such a high resolution within the object that we can then detect changes in the microscopic structure of the papyrus it was written on and therefore may be able to reconstruct where the writing happened on that scroll.”
The second stage involves the fragments of papyrus that are the result of previous botched efforts to physically unroll the papyri. These fragments, which have writing on them that is visible to the naked eye, are photographed, scanned and the results fed into computers. Complex computer algorithms (machine learning) are used to teach the computer to recognise the subtle structural differences between the inked and blackened areas (such as differences in the structure of papyrus fibres). After the computer has been trained on these fragments, the idea is to apply it to the data from the intact scrolls collected at Diamond Light Source and, hopefully, reveal the text hidden within.
‘It's ironic, and somewhat poetic, that the scrolls sacrificed during the past era of disastrous physical methods will serve as the key to retrieving the text from those [that] survive but are unreadable,’ says Professor Seales.
"By digitally restoring and reading these texts, which are arguably the most challenging and prestigious to decipher, we will forge a pathway for revealing any type of ink on any type of substrate in any type of damaged cultural artefact.’
The team has finished collecting the x-ray data and is now in the process of training the computer algorithms, which will then be applied to the scrolls in the coming months. In 2015, Seales and his team proved that the technique works on metallic-based inks when they were able to read the text hidden within an ancient Hebrew text, the first time this had been achieved with an ancient scroll without physically opening it.
If the same can be done for carbon-based inks, an entire library of lost prose, poetry, philosophy and history might suddenly be opened up to the modern world… and the knowledge those ancient librarians in doomed Herculaneum once tried to preserve might be finally be readable once more.
Last updated: 29 October 2019