On 21 September, a joint FLNP seminar “High-temperature superconducting hydrides: current status” was held. Artem Oganov (Skolkovo Institute of Science and Technology, MIPT, Northwestern Polytechnical University) and Dmitrii Semenok (Skolkovo Institute of Science and Technology) presented their reports. The seminar, held also online, brought together many participants in the Laboratory’s Conference Hall. As FLNP Director V. N. Shvetsov noted, it was the first full-fledged seminar mostly attended in-person in the Conference Hall rather that in front of computers. He reminded the audience that the creation of the state programme on the research of high-temperature superconductivity had been initiated by V. L. Aksenov. Results of it are already tangible.

Artem Oganov, a crystallographer by education, together with his team developed the USPEX software allowing predicting and acquiring stable chemical compounds with specified properties, including those “forbidden” by traditional chemistry. Now, thousands of researchers from all over the world use this software. Together with his team, A. R. Oganov studies high-temperature superconductors.

– Thanks to the breakthrough in theory allowing us to predict and identify new substances,” A. R. Oganov began his speech, “We are half step away from room-temperature superconductivity. We predict new materials using artificial intelligence that uses machine learning and big data. Previously, there were two possible ways to predict new substances: machine learning and global optimization. And they diverged in different directions. The first method did not allow specialists to predict brand-new compounds, and the second one was very expensive. They think that the pioneer of the chemical databases was D. I. Mendeleev. And I strongly support this point of view. He was a brave man, he had the courage to say that empty spaces in the periodic table did not violate the correlation but predicted undiscovered elements. I talked today to Yu. Ts. Oganessian, the successor of Mendeleev. We face the following
question today: where does the periodic table end?

Our method for determining new compounds is based on neural networks that allow predicting the properties for which there are no theories yet. Machine learning gives an opportunity to predict stable structures and their properties without using supercomputers and without losing accuracy. 118 chemical elements known today may form the 7021 binary system, while there may be more than a million triple systems. It is just impossible to consider all of them, and the evolution algorithm used in our software manages to do it: it “understands” what compounds are more promising, and at each next step of the search, it focuses on more promising compounds.”

The speaker also reported on new chemistry they were discovering. Under the pressure of about 10 million atmospheres, the periodicity is broken, and under the pressure of 1 million atmospheres, the periodicity still works but unexpected compounds are produced with unexpected properties, including superconductivity. For example, oxygen under the pressure of 10 GPa becomes a superconductor, and sodium becomes a transparent nonmetal. The speaker gave examples of new compounds with unexpected, interesting properties, and told the audience how they checked other people’s non-obvious results using their software. We heard very similar verdicts: the theory is wrong but very interesting; work is incorrect but very interesting, etc.

“Who stole calcium carbide from construction sites and threw it in a puddle!”, RAS Professor and a Full Member of the Royal Society of Chemistry said to the audience, and several hands rose. Artem mentioned research of new explosives but the military theme does not interest him a lot. Then, we got acquainted with the latest history of high-temperature superconductors. In 2014, scientists from China predicted a new compound using the USPEX programme: H3S with a record high superconductivity temperature of 203 K, the cuprate record of 135 K was broken. The triumph of the union of theoreticians and experimentalists contributed to hundreds of scientific papers on the topic.

A. R. Oganov presented research conducted by his post-graduate student and colleague Dmitrii Semenok who proposed to visualize the results of high-temperature superconductivity using the periodic table. “Dmitrii has acquired breakthrough results in the fields of high-temperature superconductivity, his scientific papers have put us at the forefront of the race for room-temperature superconductivity. It turned out that metal hydrids produced under the pressure of 100-200 GPa have superconductivity under record temperatures, and the lanthanum hybrid LaH10 under the pressure of 190 GPa has the official record (260 K is almost room temperature). “Where’s hope? In complicating chemical composition.” Why? I do not know but the temperature of superconductivity increases with the increase in chemical complexity.” Moreover, the Oganov’s team learned to predict stable chemical compounds not only in the binary but also in the molecular state. “We have understood the principle for distinguishing stable molecules from unstable ones.

The palette of stable substances is way much wider than the known one. When I was four, I was dreaming to work with Flerov and Oganessian, to synthesize new elements,” Artem Romaevich said and demonstrated the SinOm silicon oxide’s stability ridge very similar to the famous stability island of isotopes.

The speech by A. R. Oganov evoked numerous questions among participants of the event, and the moderator had to limit them as far as the report by Dmitrii Semenok was planned then. He reported on the state-of-the-art of the superconducting metal hydrid field, peculiarities of their synthesis at diamond anvils, i.e. the most expensive part of the experiment. Finding out the structure of synthesized hydrids is possible only by using synchrotron radiation, and there is hope for a source in Kurchatov Institute not to go to Grenoble, Hamburg or China. Various hydrids of uranium have been synthesized. Many of them are stable at 1 atmosphere, they may be studied by neutron diffraction. Dmitrii distributed superconducting properties of hydrids using neural networks basing on the periodic table. One of the most promising room-temperature superconductors is actinium, and according to the speaker, it is also one of the very interesting materials for studying superhydrids.

Scientific Leader of the Laboratory V. L. Aksenov expressed gratitude to the speakers and made final comments. And V. N. Shvetsov said the following concluding the seminar, “I am impressed by the results and the general scientific approach covering the theory, predictions and synthesis. I am sure that it is not our last meeting.” After the seminar, A. R. Oganov answered the questions of our journalist.

– You have already known Yu. Ts. Oganessian before and have met him this time as well. Do these meetings give you something in addition to up-to-date scientific information?

Of course, in addition to scientific information – and Yuri Tsolakovich just has lots of ideas that inspire, in which one always finds something close to their research area. So, he is also a very warm and kind man, and his childish sparkle in eyes, his interest in life, science are, of course, inspiring. I always strive to take my cue from such people, one always wish to see a man they take the cue of more often. There is no way to visit Dubna without meeting Yuri Tsolakovich.

– I hope your group will establish long-standing cooperation with FLNP. Neutrons are a useful tool for your studies, and Dubna is closer than Grenoble…

I really hope so. Firstly, it provides additional information we have not had before. It is possible to obtain information about our objects using neutron scattering that cannot be acquired by any other methods, it is crucially important. On the other hand… you know I believe in collective intelligence, that we will advance not only in terms of experimental implementation of our plans but also in terms of new ideas cooperating with scientists from such a laboratory as FLNP JINR.

– How did you come with the idea to search for non-existent compounds?

In general, new chemical compounds regularly appear. Someone synthesises something that has not existed before. There is a peculiarity: compounds that we study have been recently considered impossible. For example, new sodium chlorides, such as NaCl₃ or Na₃Cl, or current record-breaking superconductors, such as LaH₁₀ or YH₆. All this was achieved thanks to prediction methods I have invented, methods for predicting stable chemical compounds. And this was born in many ways from the questions of my childhood. When I studied at school, I had a question: why is the NaCl compound possible and the NaCl₂ or NaCl₃ are not? What does impossible mean? I can create a model of the crystal structure with such a composition, I can calculate the properties, the energy of a substance. Okay, it will be unstable for some finite values but if we change the conditions of the experiment or calculations to higher pressures, strong electrical or magnet field, maybe it will be stable? When I invented the method, it became possible to answer this question. And the answer was actually positive. Such substances that contradict the chemical intuition and do not really appear in normal conditions can, in fact, emerge under high pressure. By the way, other interesting things are possible in strong magnet or electrical fields but it is still understudied. We do not delve into this field, some other groups have already made predictions, and they are interesting: they also predict things that do not fit into classical rules.

– Are there great prospects?

Yes, of course, because most of the matter in stars and planets is under high pressure, and some stars, neutron stars, for example, have very strong magnetic fields. I believe that it is not so interesting to apply this to neutron stars as far as there is no chemistry, only a soup of neutrons and a small number of protons, and the gas around a neutron star has the atomic state. So, its behaviour in the strong magnetic field may be a very curious question. I still do not know why it should be studied other than to check that we know it. However, there definitely will be interesting applications, new interesting chemistry, new physics. It turned out that helium atoms in a strong magnetic field do not repel each other but, on the contrary, form very strong chemical bonds. It is absolutely unexpected. Our chemical intuition is based on the fact that we know the behaviour of chemical elements under normal conditions but this behaviour changes with the change in conditions, and things that seem strange appear. And it is important because there are places in the Universe where the matter is under such conditions. Most of the matter on our planet is in the state of very high pressure, the pressure in the centre of the Earth is almost 4 million atmospheres, and, of course, chemistry is totally different there.

Olga Tarantina, JINR Weekly Newspaper
photos by Elena Puzynina