Machine learning made it possible to study the phase transitions of water that are inaccessible for experiments

(ORDO NEWS) — Using modeling and machine learning techniques, scientists have been able to investigate the phase transition of water, which has so far been impossible to recreate in experimental conditions.

The researchers confirmed the existence of a liquid-liquid transition occurring at temperatures below minus 100 degrees Celsius, in which water separates into two immiscible phases with different densities.

For the past 30 years, scientists have believed that when cooled to very low temperatures (less than minus 100 degrees Celsius), water can separate into two liquid phases with different densities.

These phases do not mix, and their presence may explain some of the other strange properties of water, such as why it becomes less dense as it cools.

However, this phenomenon is almost impossible to study in the laboratory, as water turns to ice very quickly at such low temperatures.

A new study by scientists from the Georgia Institute of Technology (USA) has overcome this limitation. The authors of the work used machine learning models to better understand the phase changes in water.

They were able to find compelling computational evidence in support of the liquid-to-liquid transition of water that can be applied to real systems.

In their calculations, scientists used quantum-chemical calculations that are as close as possible to real physics.

They ran molecular simulations on supercomputers that were compared to a virtual microscope that allows them to zoom in on individual molecules and observe their movement and interaction in real time.

So the researchers were able to characterize the structure of the liquid at various temperatures and pressures.

The scientists also used a machine learning algorithm that calculated the interaction energy of water molecules with each other.

This model performed calculations much faster than traditional methods, which made it possible to conduct a virtual experiment much faster.

The scientists carefully tested their predictions using a series of different algorithms.

One of the main problems of such studies is that the data obtained are almost impossible to compare with actually observed processes.

Some of the conditions from the virtual experiment are not possible on Earth at all, although they could potentially be present in various aquatic environments in the solar system, from the oceans of Jupiter’s moon Europa to water at the centers of comets.

However, the findings could help researchers better explain and predict the strange and complex properties of water, use it more efficiently in industrial processes, and develop better climate models.


Contact us: [email protected]

Our Standards, Terms of Use: Standard Terms And Conditions.