Presentation and Excibition at 2025 Spring Meeting of E-MRS
PFCC will be exhibiting and presenting at the 2025 Spring Meeting of the European Materials Research Society (E-MRS) to be held from May 26 to 30, 2025, in Strasbourg, France. PFCC will deliver two oral presentations and one poster presentation at the conference. In addition, we invite you to visit Booth 37 to explore the latest advancements in Matlantis and discuss your research challenges with our experts. Please see below for exhibition details.
Exhibition and Booth Information
Expo hours: Tuesday, May 27 – Thursday, May 29, 2025
Location: Convention & Exhibition Centre of Strasbourg
PFCC Booth: 37
Presentation Information
Presenter 1: Junichi Ishida
Session: session S07
Session Date and Time: May 27, 2025, at 3:45 p.m.
Session Type: Oral Presentation
Title: Distilling Lightweight Machine Learning Potentials from a Universal Potential: Application to Micelle Formation
Abstract: To enable computations of large-scale systems that are difficult to handle with PFP, LightPFP—a service that creates efficient, system-specific models—proves highly useful. In this study, we trained a LightPFP model using PFP enhanced with data incorporating the r²SCAN functional, which exhibits high predictive accuracy for molecular systems. We successfully reproduced the aggregation process in aqueous solvent of the complex cationic surfactant 12-s-12, which possesses gemini-type hydrophobic-hydrophilic pairs.
Presenter 2: Lieven Bekaert
Session: session L09
Session Date and Time: May 28, 2025, at 9:45 a.m.
Session Type: Oral Presentation
Title: Suppressing chemical reactivity of Na metal with Na3PS4-derived solid-state electrolytes by heteroatom doping: a computational-experimental approach
Abstract: Using Li and Na metal electrodes in next-generation rechargeable batteries can significantly increase battery capacity. However, a high chemical reactivity and dendrite formation are hindering their implementation. It is therefore necessary to understand and engineer new electrolyte materials to be compatible at these electrode-electrolyte interfaces. Interfaces are challenging to characterize experimentally which is why computational methods are being widely used. A lack of classical compatible force fields in classical techniques and a limited size and time scales in Density Functional Theory (DFT) due to the high computational cost made computational analysis very limited. Recently, machine learning potentials have been developed which are significantly faster than ab-initio methods yet with a similar chemical accuracy as ab-initio methods. In this study, the chemical reactivity at the Na metal and Na3PS4 solid-state electrolyte interface is investigated using Preferred Potential (PFP) atomistic simulations combined with DFT ab-initio calculations and X-ray photoelectron spectroscopy. Insights from the decomposition mechanism were used to propose and screen a range of new analogues with heteroatom doping, upon which a material with significantly higher chemical stability was discovered and confirmed experimentally. This research shows that a combined machine-learning potentials, ab-initio, and experimental approach can provide an efficient way of accelerating material R&D.
Presenter 3: Taku Watanabe
Session: session SP03
Session Date and Time: May 28, 2025, at 13:45 p.m.
Session Type: Poster Presentation
Title: Computational Investigation of Pd-based Alloy Systems for Hydrogen Separation Membranes Using A Universal Neural Network Potential
Abstract: Pd-based alloy membranes for hydrogen separation offer high efficiency and important practical advantages over traditional methods like pressure swing adsorption or cryogenic distillation. However, the rarity and expense of Pd, along with its reactivity with certain gas mixture components, pose significant challenges. Our study uses the Preferred Potential (PFP) neural network to perform atomistic simulations that explore the impact of alloy configurations on hydrogen adsorption and diffusion at the surface. PFP allows near-DFT accuracy in simulating chemical reactions on larger systems and at higher speeds.
Our findings using Monte Carlo simulations indicate pronounced surface segregation behaviors in fcc Pd-based alloys, with specific patterns for Pd3Ag, Pd3Au, Pd3Cu, Pd3Ni, and Pd3Pt. These results align with known literature. Additionally, ternary phases like Au-Ag, Au-Cu, and Cu-Ag in Pd reveal similar segregation trends and interesting subsurface structures. The variation in surface composition significantly affects hydrogen binding energies and migration energy barriers, suggesting that the choice of alloy composition could optimize membrane performance for hydrogen separation.
For more information about the conference (External site) >> 2025 Spring Meeting of E-MRS
