Highlights

Edited by NCCR MARVEL group leaders Michele Ceriotti and Anatole von Lilienfeld as well as colleague Cecilia Clementi of Rice University, the JCP Special Topic on “Machine Learning Meets Chemical Physics” features papers from the labs of Clemence Corminboeuf and Ceriotti at EPFL and von Lilienfeld at the University of Basel. The issue focuses on research that puts an emphasis on the interplay between machine learning and chemical physics, whether by incorporating physical principles and chemical intuition into model construction or by using machine learning to recognize new laws or general design principles. In the article below, we look at highlights of the research contributed from the three MARVEL labs. 

Researchers led by Mathieu Luisier from the Institute for Integrated Systems (IIS) at ETH Zurich and NCCR MARVEL's Director Nicola Marzari at EPFL have set out to develop a comprehensive atlas of 2-D materials that might be capable of challenging currently  manufactured silicon-based transistors, so-called Si FinFETs. The new simulations are based on earlier results from Marzari and his team, who used complex simulations on the Swiss National Supercomuting Centre's supercomputer "Piz Daint" to sift through a pool of more than 100,000 materials to identify 1,825 from which it might be possible to obtain 2-D layers of material. The paper 2‑D Materials for Ultrascaled Field-Effect Transistors: One Hundred Candidates under the Ab Initio Microscope, recently published in the journal ACS Nano, identifies 13 particularly promising candidates. 

High-throughput computational materials design is an emerging field that looks set to accelerate reliable, cost-effective design and optimisation of new materials that feature specific desirable properties. Maximally-localized Wannier functions (MLWFs)—a means of representing the Bloch eigenstates of a periodic system—are tools that have been regularly used to compute certain advanced materials properties from first principles. Bringing the two approaches together has been complicated by the fact that generating MLWFs automatically and robustly without user intervention and for arbitrary materials is difficult. Researchers at Nicola Marzari’s THEOS lab and colleagues have now addressed this problem by proposing a procedure for automatically generating MLWFs for use in high throughput frameworks. In the interest of Open Science, they have also developed a virtual machine that allows researchers to perform their own simulations, either with different parameters or on new materials using this new protocol. The research was a collaboration between the EU H2020 E-CAM and MaX Centres of Excellence. A Success Story on the project can be found here.

Dr. Carlo Pignedoli, deputy group leader of the Atomistic Simulations Group at the nanotech@surfaces Laboratory of Empa, has been given special recognition by his laboratory head for two years in row. His feat? Following a record number of projects from experimental colleagues and coordinating simulations in the field of carbon-based nanomaterials, one of the lab’s core activities, also associated with MARVEL Design & Discovery Project 3 – Low Dimensional Materials. For Pignedoli, the achievement was made possible by excellent researchers, certainly, but also by the exceptional efficiency in performing computational tasks thanks to AiiDA, open-source infrastructure for managing and storing the ever-growing amount and complexity of workflows and data in computational science. “High computational efficiency is what made the difference in the last two years compared to the past and it's all about AiiDA and AiiDA lab,” Pignedoli said. “With this "success story" I would like to acknowledge the AiiDA & AiiDA lab teams for all their work and for continuously inspiring new solutions for boosting computational materials science.”

Recent experiments have shown that adding palladium (Pd) to the high-entropy alloy CoCrFeNi produces a material that is significantly stronger. In the paper Origin of high strength in the CoCrFeNiPd high-entropy alloy, NCCR MARVEL researchers led by Prof. Bill Curtin, head of EPFL’s Laboratory for Multiscale Materials Modelling, show how a recent parameter-free theory for initial yield strength in random alloys predicts a strength of CoCrFeNiPd that is in good agreement with experimental results. The strengthening is mainly due to the large misfit volume of Pd in CoCrFeNi.

Researchers including Professor Oleg Yazyev, chair of Chair of Computational Condensed Matter Physics at EPFL, and scientist QuanSheng Wu have shown that flat bands are a fundamental feature twisted double bilayer graphene (TDBG) rather than the result of any external fields and have identified a so-called magic angle of 1.3 degrees, at which both electron and hole gaps are maximized. A separate paper, published together with experimental colleagues in China, shows that TDBG is an easily tunable platform for exploring quantum many-body states thanks to vertical displacement fields. The corresponding papers were recently published in Nano Letters and Nature Physics, respectively. 

New collaborative work involving NCCR MARVEL researchers has provided additional insight into the nature of jacutingaite (Pt2HgSe3), a species of platinum-group mineral first discovered in a Brazilian mine in 2008. The new studies show that the material is one of only a few known dual-topological insulators, featuring different surface states that are linked to crystalline symmetries rather than to the topological properties of the 2D monolayer, which is a quantum spin Hall insulator (QSHI). The work has been published in Physical Review Letters and Physical Review Research. 

Heterogeneous catalysts are often preferred in industrial settings because of their robustness and lower operating costs, but homogenous catalysts still dominate when high selectivity is needed—finding superior heterogeneous catalysts has been a challenge. A recent collaboration between the Paul Scherrer Institute's experimental Syncat Group, led by Marco Ranocchiari, and EPFL’s Laboratory of Molecular Simulation, a computational group led by Berend Smit, has shown how micropores in metal-organic frameworks (MOFs) can enhance selectivity to levels that cannot be achieved with existing catalysts. Though the findings have significant potential in the production of aldehydes, the easy experimental protocol and the chemical and structural flexibility of MOFs means that the approach represents a powerful tool for designing selective catalytic heterogeneous processes in the fine chemical industry overall. A paper on the research has just been published in Nature Communications.

NCCR MARVEL researchers at EPFL’s Chair of Computational Condensed Matter Physics (C3MP) and colleagues have developed an artificial neural network (ANN) model that may serve as a basis for solving inverse problems. Their approach reaches the same level of accuracy as the now commonly used maximum entropy (MaxEnt) method for low-noise data, performs significantly better than this standard technique when the noise strength increases, and features a reduction in computational cost by orders of magnitude. The research has just been published in Physical Review Letters.

Michele Simoncelli, a PhD student here at EPFL, together with Andrea Cepellotti, a former EPFL student now at Harvard, and Nicola Marzari, head of EPFL's Theory and Simulation of Materials laboratory as well as director of NCCR MARVEL, have developed a novel set of equations for heat propagation that goes beyond Fourier’s law and explains why and under which conditions heat propagation can become fluid-like, rather than diffusive. These "viscous heat equations'' show how heat conduction is not only governed by thermal conductivity, which was introduced by Fourier in his well-known macroscopic law of heat conduction, but also by another quantity, thermal viscosity. The theory is in striking agreement with pioneering experimental results in graphite published last year and may pave the way for the design of the next generation of more efficient electronic devices. The paper, Generalization of Fourier's law into viscous heat equations, has been published in Physical Review X.  

Wannier90—a computer program for generating maximally-localized Wannier functions and using them in the computation of advanced electronic properties of materials—has become a community code with a wide base of contributors over the last few years. This has resulted in a major new release with novel features described in the paper Wannier90 as a community code: new features and applications, published in Journal of Physics: Condensed Matter.

Though researchers have been looking for solid-state electrolytes that could enhance both the safety and performance of lithium-ion batteries for decades, no thoroughly suitable candidate has yet been found. Computational screening may offer better chances of success than previous methods of investigation, largely led by chemical intuition and experiment, but such methods must also meet certain criteria. In a recent paper published in the journal Energy & Environmental Science, NCCR MARVEL researchers Leonid Kahle, Aris Marcolongo and Nicola Marzari present a suitable computational framework for predicting the diffusion of Li-ions in solid-state materials, show how to employ it in large-scale computational screening and use it to identify new ceramic compounds for further experimental investigation.