Mongolian Buddhist Shrine Scrolls Virtually Unrolled Using X-Ray Tomography

Researchers at the Helmholtz-Zentrum Berlin (HZB) have digitally reconstructed one of three miniature scrolls found in a Mongolian Buddhist shrine, currently housed at the Ethnological Museum in Berlin.

Using high-resolution 3D X-ray tomography at the BESSY II synchrotron facility, the team produced a virtual unrolling of the scroll, revealing a dharani prayer inscribed in Sanskrit script, but rendered in Tibetan characters.

Preserving cultural heritage

The Gungervaa shrine, part of the Ethnological Museum’s Mongolian collection, is a uniquely interesting artifact. Buddhism in Mongolia developed its own traditions linked to nomadic culture, with many families possessing portable shrines that featured various statutes and images. These shrines also sometimes contained small, tightly rolled scrolls inscribed with prayers, known as ‘dharanis’.

This cultural practice was nearly eradicated during Mongolia’s revolutionary period (1921–1930), when many shrines were destroyed.

One surviving shrine eventually found its way to Berlin, where it was stored in the museum archives for decades, with little documentation about its origin. Museum restorer Birgit Kantzenbach recently initiated research on the shrine, which revealed a jumble of disordered fabric flowers, relics and three silk-wrapped scrolls. 

Non-destructive scanning at BESSY II

Eager to study these scrolls, Kantzenbach brought the artifacts to the attention of HZB physicist, Tobias Arlt.

Conventional methods of examining such artifacts typically require physical unrolling, which runs the risk of harming the fragile scroll materials.

To avoid this, Arlt brought the scrolls to the tomography station of the Federal Institute for Materials Research and Testing (BAM) at BESSY II. Using 3D x-ray tomography, Arlt was able to produce high-resolution images revealing the inner contents of the scrolls.

"The high-resolution 3D images show that there are around 50 windings in each scroll, with strips measuring over 80 centimetres that are wound tightly and carefully," Arlt said.

Scrolls virtually unrolled using AI

To access the text without disturbing the original artifact, Arltz used a mathematical method developed at the Konrad Zuse Institute and the corresponding Amira software to digitally unroll the scroll from the 3D imaging data. This method, initially time-consuming, has been significantly accelerated through the use of artificial intelligence. 

"We are continuing to optimise this complex process of virtual unrolling," said Arlt. "We also use this method in our own research, for example to analyse changes in tightly wound or folded batteries."

Upon virtually unrolling one scroll, Arlt and Kantzenbach were able to identify traces of ink, though the ink's composition was different to what might ordinarily be expected.

"This is interesting because Chinese ink traditionally consists of a mixture of soot and animal glue, but in this case, ink containing metal particles was apparently used," noted Kantzenbach.

The ink characters spelled out a well-known Buddhist mantra, "Om mani padme hum", which is associated with the concept of "great compassion" in Tibetan and Chinese Buddhism. Intriguingly, the mantra was written in Sanskrit language but using Tibetan script.

The Gungervaa shrine is now on display at the Humboldt Forum in Berlin as part of the exhibition “Restoration in Dialogue,” which runs until June 1, 2026. There are additional plans to display the shrine in Mongolia in the future.

Reference: Arlt T, Kantzenbach B. Revealing the history of a Mongolian shrine by virtually unrolling Buddhist Dharanis. J Cult Herit. 2025;74:190-194. doi: 10.1016/j.culher.2025.06.009

This article has been republished from the following materials. Note: material may have been edited for length and content. For further information, please contact the cited source. Our press release publishing policy can be accessed here.

This content includes text that has been generated with the assistance of AI. Technology Networks' AI policy can be found here.