They Will Show What Is Hidden From the Eye: AI Algorithms Will Help Create Unique Devices for Inspecting Large Equipment

Russian scientists have invented a new way to turn X-rays into visible light. They used artificial intelligence to create unique materials that will be used in X-ray cameras to better see what is hidden inside ships, buses, subway cars, nuclear reactors, and submarines. The new technique will allow the creation of more powerful equipment for detailed study of the internal structure of such large objects.

Image source generated by the DALL•E 3 neural network

Visible Only to the "Armed" Eye

Researchers from the Siberian Federal University, the Kirensky Institute of Physics, and the South China University of Technology have created a new way to find substances that can convert X-ray radiation into visible light. They used artificial intelligence algorithms. The work was carried out under a joint Russian-Chinese grant from the Russian Science Foundation.

As the scientists explained, these substances, called scintillators, absorb ionizing radiation (X-rays, gamma rays, alpha particles) and emit light in the visible range. These properties make them useful, for example, for registering radiation passed through light-conducting objects.

Physics would not have developed without scintillators, because invisible radiation needs to be visualized somehow. Currently, there is a developed line of mid-range devices for these purposes. Special single-crystal scintillators are grown for them. However, there is a need to create even larger scanning equipment.

Maxim Molokeev, method developer, Associate Professor at the basic Department of Solid State Physics and Nanotechnology of SFU and Senior Researcher at the Laboratory of Crystal Physics of IF SB RAS

He explained that such equipment is necessary for detecting defects in the internal structure of airplanes, train cars, ships, submarines, nuclear reactors, and other large-sized technical devices. Growing crystals for devices of this size is impossible, so it is necessary to develop alternative approaches.

AI to the Rescue

Metal-organic compounds are a promising direction for the search for scintillators, noted Maxim Molokeev. These substances combine metal ions and organic molecules. Chemists have already created effective metal-organic compounds that are successfully used. However, their synthesis relies more on intuition than on accurate data.

We are the first group that, with the help of artificial intelligence, has shown methods for finding such compounds. Moreover, since there is not enough statistics, non-standard methods of data preparation for machine learning were used.

Maxim Molokeev, method developer, Associate Professor at the basic Department of Solid State Physics and Nanotechnology of SFU and Senior Researcher at the Laboratory of Crystal Physics of IF SB RAS

He said that a hypothesis was put forward that the high brightness of radiation depends on the distance between metal atoms. Then, the scientists collected a database of 296 open metal-organic compounds with known distances between metals and used it to train artificial intelligence.

The method is based on artificial radiation algorithms. A hypothesis was put forward that a certain distance between metal atoms contributes to the brightness of the radiation.

Maxim Molokeev, method developer, Associate Professor at the basic Department of Solid State Physics and Nanotechnology of SFU and Senior Researcher at the Laboratory of Crystal Physics of IF SB RAS

The program applied the identified dependencies to search for new compounds. The solution to the problem was based on a large amount of data, which increased the reliability of the forecast.

Earlier, Russian scientists created unique reagents known as"chemical shepherds". These substances are designed to clean the surface of seawater from fuel oil. These reagents are planned to be used for mechanical and sorption methods of eliminating accidental oil spills. The new substances have already proven their high efficiency.