Физики разработали «панду» для вычисления углов смачивания в рамках численного эксперимента.

The results have been published in the journal Colloids and Surfaces A: Physicochemical and Engineering Aspects.

Contact Angles of Wetting characterize the interaction between molecules of a solid surface and a liquid. Understanding the contact angle between materials helps predict their behavior, assess surface cleanliness, and aids in the development of hydrophobic and hydrophilic coatings. Theoretical and numerical methods for calculating the contact angle require the input of many parameters and are dependent on defining the phase interface position. This poses a challenge, as the concept of "phase interface" is conditional, with numerous criteria for its determination.

“We have relieved researchers in the field of molecular modeling from the need to find the phase interface position — this has reduced the number of adjustable hyperparameters to two instead of 8–10 used in similar methods. There is no longer a need to adjust the angle calculation method for each liquid-solid pair; PANDA is universal in this regard. Moreover, there is no longer a need to apply a transport device to the system image,” said Ilya Kapanichuk, senior researcher at the MIPT Center for Computational Physics.

The method is based on analytical expressions that describe the shapes of one-dimensional density profiles for various types of surfaces formed between immiscible liquids, such as oil and water. The density profile shows how molecules are distributed across the gap space. The study examined all possible forms of the phase interface between two liquids in a flat gap.

The PANDA method calculates the wetting angle using a one-dimensional density profile. The expressions for the analytical density profile are defined by three parameters: the volume fraction of the non-wetting liquid, the cell size, and the contact angle. If the first two parameters are fixed, which is relatively easy at the start of the calculation, the angle value can be obtained through an optimization procedure that minimizes the deviation of the analytical density profile from the profile obtained in a numerical experiment.

Researchers tested the algorithm using synthetic data and data from molecular dynamics simulations. PANDA demonstrated high accuracy with an error of less than one percent for most contact angles in synthetic scenarios. For molecular dynamics data, the algorithm showed better agreement with real experiments than the experiments did among themselves.

PANDA simplifies density analysis at the nanoscale by eliminating the need to find the phase interface. It reduces the number of adjustable parameters to two, making it universal and user-friendly. This results in the algorithm being applicable to any system without prior parameter adjustments for specific compositions.

One limitation of PANDA is the requirement to visualize the droplet shape according to a developed nomenclature (doughnut, droplet, etc.). Researchers plan to remove this limitation by training a PointNet neural network. This will make the contact angle calculation using molecular dynamics fully automated.

The PANDA method still requires further refinement for use in actual physical experiments. One of the tasks set is to develop a universal method for estimating contact angles based on the interference pattern from liquid droplets. Solving this problem necessitates close collaboration with experimental physicists.

The PANDA method is already being used in the oil industry for optimizing hydrocarbon extraction and developing new technologies. In the future, PANDA could also be applied in other fields, including materials science and coating technology.

The study involved scientists from the MIPT Center for Computational Physics, the Institute of Applied Physics of the Russian Academy of Sciences, HSE, and AIRI.