AI Unlocks the Mystery of Supercooled Water

Breaking Down Supercooled Water

According to НВ — Техно: On July 9, at 18:30, researchers at Osaka University turned to artificial intelligence to investigate the structure of supercooled water. Their findings were published in the journal Communications Chemistry. Supercooled water is a state in which water remains liquid even when cooled below its normal freezing point. Under typical conditions, for a liquid to turn into ice, its molecules must arrange into an ordered crystalline lattice, and the crystallization process begins at nucleation sites.

Water’s Unique Traits

The study sheds light on the unusual properties of water in its supercooled state, which arise from competition between two liquid forms: high-density liquid (HDL) and low-density liquid (LDL). As temperature increases, the compact structures of high-density liquid begin to dominate over those of low-density liquid. Researchers had previously proposed various parameters to describe the local arrangement of molecules, such as local density and the order of tetrahedral bonds.

To train the AI system, the researchers fed it data sets obtained from molecular dynamics simulations. The system analyzed 16 different descriptors, evaluating how well each could distinguish between the high-density and low-density forms under varying temperature conditions. Because these descriptors were based on different scales and types of data, direct comparisons between them were not possible.

“Machine learning has already proven effective in classifying structural data.”

Kan Kim, lead author of the study

Co-author Nobuyuki Matubayasi noted that 'the system analyzed 16 different descriptors, assessing their ability to differentiate between high-density and low-density forms under various temperature conditions.' This work has the potential to deepen our understanding of water’s behavior in different states and to drive further technological advances in the field.

The study of supercooled water and its properties could have a major impact on fields such as physics, chemistry, and materials science. By understanding how water behaves under non-standard conditions, researchers may develop new technologies that harness the unique characteristics of liquids. Additionally, this research opens new avenues for scientific exploration in molecular dynamics and artificial intelligence, simplifying the analysis of complex systems. In short, these findings could carry far-reaching implications for science and technology as a whole.

In addition to unlocking the secrets of supercooled water, advancements in artificial intelligence are also making significant strides in other scientific fields. For instance, researchers recently utilized AI to enhance our understanding of underwater geological formations, leading to a remarkable increase in the number of identified submarine calderas. This innovative approach showcases the transformative power of AI in various scientific inquiries, highlighting its potential impact across disciplines. To learn more about these exciting developments, check out how AI is reshaping our understanding of geological features in the oceanic landscape.