🔎Discover Alex, Patricia P., Sebastian, and Christoph's new paper, "Automatic Process Exploration through #machinelearning Assisted Transition State Searches," now online at Physical Review Letters.
📣 Press release: https://www.fhi.mpg.de/1732161/2025-03-04_APE_MLIP?c=321765
📰 Paper: https://journals.aps.org/prl/abstract/10.1103/PhysRevLett.134.096201
👋 Today we welcome back Julius Lonnes, who is joining our department again, this time as #PhD student in Nico Hörmann´s group.
During the next three years Julius will be working on "Electrochemical Barriers and Kinetics at applied Potential from First Principles".
Before that, Julius obtained his M.Sc. in Chemistry from the @tu_muenchen, with a thesis entitled "Electrochemical Adsorption of Hydrogen on Platinum: A Theoretical Perspective".
Looking forward to be working with you Julius!
👋 Today we welcome Leon Müller, who will be starting as a new #PhD student in Matthias Kick´s group, working on light storage in solar batteries. 👋
Leon obtained his M. Sc. in Electrical Engineering from the @tu_muenchen. He carried out his master project on "Accelerated Perovskite Solar Cell Development Through Machine Learning" in the PV Lab at the @epfl.
We are looking forward to be working with you Leon!
👋 Today we welcome Aditya Kumar, who will be starting a Postdoc in Sebastian Matera´s group, working on developing an automatic procedure to translate transition-state search results into a kinetic Monte Carlo model.👋
Aditya obtained his PhD in Chemical Engineering from the Indian Institute of Technology (IIT) Bombay with a thesis entitled "Molecular Scale Studies on Kinetic Aspects of Heterogeneous Catalysis".
We are looking forward to be working with you Aditya!
2️⃣ What is the key focus?
The review highlights the coupled proton-electron transfer process at electrified metal-water interfaces and contrasts it with classical understandings.
It also discusses recent insights into the energetics of protons along their adsorption path and underscores the advantages of GC descriptions, such as avoiding unclear concepts and being applicable to various reactions and systems.
1️⃣ What is the review article about?
In his review article, Nico gives exciting insights into modern grand canonical (GC) descriptions of electrified interfaces.
This framework offers a thermodynamically consistent way to assess energetics and kinetics based on electrode potential.
👏 Congratulations to Nico, whose review article "Grand canonical view on electrochemical energetics at applied potential in a nutshell" is now out on Current Opinion in Electrochemistry! 👏
If you want to know more about the review article, have a look at it online https://www.sciencedirect.com/science/article/pii/S2451910325000158 or read the paper digest below 👇
3️⃣ What are the implications for the future?
This approach could enhance the derivation of microkinetic equations from experimental data and serve as a kernel method for designing experiments. It offers a systematic way to discover optimal models, potentially improving the robustness and accuracy of data-driven modeling in various scientific fields.
2️⃣ What is new?
The paper introduces an alternative approach that optimizes the library of functions while imposing sparsity. It employs a multi-objective genetic algorithm (NSGA-II) to generate a subset of optimal models, evaluated by both the quality of fit and the statistical distribution of residuals.
1️⃣ What is the paper about?
The paper discusses sparse data-driven approaches for approximating the governing laws of physical processes using parsimonious equations. It highlights the challenge of selecting an optimal set of basis functions for promoting sparsity in these models.
👏 Congratulations to Gianmarco, Maryke and @christophscheurer, whose paper "Pareto-based Optimization of Sparse Dynamical Systems" made it into The Journal of Chemical Physics! 👏
While we wait for their paper to be published online, you can have a look at the #preprint on #chemrxiv or read the paper digest below! 👇
https://chemrxiv.org/engage/chemrxiv/article-details/673da9a47be152b1d0d0d388
👋 Today we welcome Zhe-Yun Kee, who will spend one year in the group of @christophscheurer supported by a fellowship from the National Science and Technology Council of Taiwan.👋
During this time, he will focus on the modeling of single-atom electrocatalysts consisting of metalloporphyrin complexes on graphite surfaces as a model for multi-walled nanotubes.
When not in the office, Zhe-Yun likes hiking, jogging, and playing badminton.
We are looking forward to be working with you Zhe-Yun!
3️⃣ What are the implications for the future?
The discovery of these low-barrier processes significantly increases the kMC-determined island diffusivity, suggesting that APE could enhance the accuracy and efficiency of simulations in surface science and potentially other fields, leading to better understanding and innovation in material design and analysis.
2️⃣ What is new?
The APE framework utilizes a fuzzy machine-learning classification algorithm to minimize redundancy in transition-state searches, focusing on previously unexplored local atomic environments.
This approach uncovers numerous low-barrier collective processes that were previously overlooked.
1️⃣ What is the paper about?
The paper introduces an Automatic Process Explorer (APE) framework designed to reduce the dependence on human intuition in identifying relevant elementary processes in systems, such as those used in kinetic Monte Carlo (kMC) simulations.
👏 Congratulations to Alex L., Patricia P., Sebastian, and @christophscheurer, whose work "Automatic Process Exploration Driven by Diversity in Local Atomic Environments: Beyond Look-Up Table Kinetic Monte Carlo" made it into Physical Review Letters!
While we wait for their paper to be published online, you can have a look at the #preprint on #chemrxiv or read the paper digest below! 👇
https://chemrxiv.org/engage/chemrxiv/article-details/6710a9cecec5d6c14263fc9e
🥳 Happy international Day of Women and Girls in Science! 🥳
Let's take a moment to celebrate all the amazing talents in our department, whether they are just beginning their careers, participating in conferences or receiving awards.
Let´s celebrate them all, regardless their achievements or various position within the department!
🥂 Here's to ensuring that no one ever takes away your imagination, creativity, or scientific curiosity!
3️⃣ What are the implications for the future?
The findings suggest that using atom-pairwise methods like XDM could enhance the efficiency and accuracy of BEEF-vdW in various systems, potentially leading to more effective and computationally feasible approaches in surface science and catalysis research.
2️⃣ What's new?
The study evaluates the performance of commonly used atom-pairwise corrections, such as the Tkatchenko-Scheffler (TS) and exchange-hole dipole moment (XDM) approaches, as well as many-body dispersion (MBD) treatments.
It finds that atom-pairwise methods, particularly XDM, offer a good balance between computational cost and accuracy across molecular, surface, and solid-state systems.
1️⃣ What is the paper about?
The paper investigates alternatives to the nonlocal van-der-Waals density functional (vdW-DF2) used in the Bayesian error estimation functional (BEEF-vdW) for surface science and catalysis.
It focuses on exploring atom-pairwise and many-body dispersion treatments to improve computational efficiency and accuracy for molecular systems.