Interview series about ENCI, part 2: X-ray physics‘The Challenge is to Combine Lightweight Construction and Radiation Protection’

25 January 2024

ENCI is a globally unique mobile CT scanner for sealed cuneiform tablets. In this interview, Christian Schroer explains the main challenges during the construction of ENCI and talks about the current fine-tuning before it is used in the Louvre.

Lesen Sie hier die deutsche Version dieses Interviews

Christian Schroer, it’s not exactly common for X-ray physicists to work together with Assyriologists. How did this come about?

During a retreat, Cécile Michel and I once had a long chat over dinner. She told me about the sealed cuneiform tablets that she and her colleagues were unable to read, and we quickly realised that X-ray tomography was ideally suited to solving this problem. However, you need a mobile device, because archives and museums don’t let their collections travel. My colleagues and I had previously built a technology demonstrator for the European Space Agency (ESA), which also had to be very compact. We therefore already had some expertise in this area that our project built on.

The result is the world’s first mobile device that can read sealed cuneiform tablets. What technical challenges had to be overcome along the way?

In terms of X-ray physics, what we have built has already been tried and tested before. Thus, we never had any doubts that the technology itself would work. The difficulty was to combine lightweight construction and radiation protection, which is first and foremost an engineering challenge. We operate with very strong radiation: we literally have to go through a brick, after all. At the same time, of course, we don’t want the radiation to come out. That sounds contradictory: we want to look through walls, but the radiation must remain inside. This requires a very effective radiation shield made of tungsten, a very heavy metal. Comparable devices weigh several tonnes. By making the radiation chamber as compact as possible and thus minimising the surface area to be shielded, we were able to reduce the weight of ENCI to just over 400 kilograms.

How does ENCI reveal what’s inside the clay envelopes then?

Tomography first provides us with a huge data set of around 12 billion volume pixels, known as voxels. Because the cover and the cuneiform tablet are made of the same material, namely clay, we can simply assign each voxel to one of two categories based on its grey value: either ‘there is clay’ or ‘there is no clay’. In this way, we can determine the boundary areas of the clay tablet. Since the cuneiform writing was carved into the clay, this allows us to reconstruct the characters on the tablet’s surface.

Of course, this isn’t done by hand but by an algorithm. Optimising this process is a challenge for our colleagues in computer science, with whom we work together on this project. This is already working very well, but in a few cases, we still have to intervene manually to make some corrections.

You initially tested ENCI on replicas before trying it on real ancient cuneiform tablets in December 2023. Were there any important differences?

Yes, the first tests with real objects provided some new insights. The replicas we used were made of clay from the craft shop, which is perfectly homogenous. There are a lot more irregularities in the real tablets, such as small holes, organic remains of plants, and so on. We also discovered that small parts of cuneiform tablets have broken off and are now lying in the hollow space between the tablet and the envelope. This makes the overall picture much less clear and thus more difficult for the algorithm. There were cases in which it ‘discovered’ writing floating freely in the air; we then have to clean up these areas manually. In the long term, however, we want to automate the whole process.

How severe are these deviations? Do they prevent Assyriologists from reading the texts on the tablets?

No, they can already read it, it just doesn’t look as nice. If we want to make a 3D print of the tablet, for example, some parts would not be true to the original. Our aim is to reproduce all the details correctly and in high resolution. We are still working on this.

Can you use ENCI to analyse objects other than clay tablets as well?

In principle, the device is suitable for experiments with various materials. However, the objects must not be too large or absorb too much of the radiation. We can’t use ENCI to look at very thick metallic objects, for example a bronze vase. But organic materials of all kinds, such as paper or palm-leaf manuscripts, are no problem. So, yes, there are also other types of written artefacts that we can examine with ENCI.