Highlights

  • ARPES gives first observation of dispersive excitons in a low-dimensional metallic system

    Already used in some solar cells, excitons, with their charge neutrality and expected mobility, have also been proposed as potential transmitters of quantum information. In either use, it’s essential to understand how and why these quasiparticles move. Investigations into exciton mobility have been inaccessible to traditional optical experiments though because they only create and detect excitons with negligible momentum. Now, using angle-resolved photoemission spectroscopy (ARPES), scientists led by NCCR MARVEL’s Professor Ming Shi at the Paul Scherrer Institute have detected several types of mobile excitons in the quasi-one-dimensional metallic trichalcogenide, TaSe3. They’ve also shown that certain exciton properties can be tuned by surface doping. The paper was recently published in Nature Materials. 

  • Manifolds in commonly used atomic fingerprints lead to failure in machine-learning four-body interactions

    The existence of manifolds in two atomic environment fingerprints commonly used to characterize the local environments of atoms in machine learning and other contexts causes a failure to machine learn four-body interactions such as torsional energies, which are an important part of standard force fields. No such manifolds can be found for the Overlap Matrix (OM) fingerprint due to its intrinsic many-body character, making it an appealing alternative, NCCR MARVEL researchers at the University of Basel found in a paper recently published in the Journal of Chemical Physics.

  • Equivariant representations for molecular Hamiltonians and N-center atomic-scale properties

    Considerations of symmetry underpin the mathematical representations of the atomic configurations that are used by machine learning models to predict properties of various molecular structures. Though these models generally rely on a description of atom-centered environments, many of the quantities that are relevant for quantum mechanical calculations – notably the single-particle Hamiltonian Ĥ matrix when written in an atomic-orbital basis – aren’t associated with a single center, but rather with two or more atoms in the structure. In the paper “Equivariant representations for molecular Hamiltonians and N-center atomic-scale properties,” recently published in the Journal of Chemical Physics, researchers discuss a family of structural descriptors that generalize atom-centered density correlation features to the N-centers case and show how it can be applied to efficiently learn the matrix elements of Ĥ. These N-centers features are fully equivariant in terms of both translations and rotations as well as in terms of permutations of the indices associated with the atoms and are therefore suitable for use in constructing symmetry-adapted machine-learning models of new classes of properties of molecules and materials.

  • Quantum physics across dimensions: unidirectional Kondo scattering

    NCCR MARVEL researchers Ana Akrap and Oleg Yazyev formed part of an international team led by EPFL scientists who unveiled a unique quantum-mechanical interaction between electrons and topological defects in layered materials. Only observed in engineered atomic thin layers, the phenomenon can be reproduced by the native defects of lab-grown large crystals, making future investigation of Kondo systems and quantum electronic devices more accessible. EPFL journalists report on the phenomenon below.

  • Ab initio model of Ca2RuO4 perovskite in remarkably good agreement with available experimental data

    NCCR MARVEL researchers at the University of Fribourg have tested the GW + EDMFT ab initio approach for correlated materials modelling. Using the insulator-metal transition in the perovskite Ca2RuO4 as a benchmark, they found that their parameter-free simulation was in close agreement with the available experimental data and that the extension to the nonlocal polarization and self-energy provided by GW are essential to attaining such accuracy. The calculations represent an important test and an encouraging result for the further application and development of the GW + EDMFT framework, the authors said.

  • Machine learning solves the who’s who problem in NMR spectra of organic crystals

    A team of EPFL researchers has combined a large database of 3D structures with a machine learning model of chemical shifts and topological representations of molecular environments to allow for the probabilistic assignment of NMR spectra of organic crystals directly from their 2D chemical structures. They demonstrated the approach on seven molecular solids with experimental shifts and benchmarked it on 100 crystals using predicted shifts. The correct assignment was found among the two most probable assignments in more than 80% of cases. The paper, Bayesian Probabilistic Assignment of Chemical Shifts in Organic Solids, was published today in Science Advances.

  • First-ever rare earth nickelate single crystals lead to first experimental evidence supporting predicted multiferroicity

    Perovskites—materials with crystal structures similar to that of calcium titanium oxide—have unique properties. Rare earth nickelates such as RENiO3, for instance, are metallic at high temperatures, but insulating and magnetically ordered at low temperatures. Moreover, it has been theoretically predicted that these materials might be multiferroic, that is, featuring simultaneously occurring ferroelectric and magnetic order in the low temperature phase. While the materials have drawn much attention for potential applications in fields of research ranging from optoelectronics, to battery engineering and neuromorphic computation, crucial experimental data needed to validate theoretical predictions has been lacking because the materials are very difficult to synthesize—to date, it has not been possible to grow sizable bulk single crystals based on rare earths other than La, Pr and Nd. Now, MARVEL researchers at PSI and colleagues have successfully grown bulk single crystals of the full nickelate family while another team including former MARVEL members from the University of Geneva has used them to provide experimental evidence supporting the existence of multiferroicity in these materials.

  • New approach to ab initio modelling of electron-phonon interactions in correlated materials

    Correlated materials, which feature highly localized electrons and strong coupling between electrons, their spin, and atomic vibrations, are among the most mysterious and exotic of all solids. They can host states of matter ranging from high-temperature superconductivity to metal-nonmetal transitions, colossal magnetoresistance and multiferroicity. Much-needed theoretical research into these materials is nonetheless hampered by a lack of quantitative methods capable of accurately describing the electron-phonon interactions that play a critical role in determining their unique properties. In a letter just published in Physical Review Letters, a team of researchers from EPFL, Caltech and colleagues introduce an ab initio approach that allows for the quantitative calculations of such interactions. The method can be broadly applied to various families of strongly correlated materials, capturing the strong coupling of electron, spin and lattice degrees of freedom and their combined effect on electron-phonon interactions, paving the way for quantitative studies of their rich physics.

  • Newly identified R-2 2D material may show promise in development of spin-layer-locking spinFETs

    Today’s electronic devices rely on the electron’s negative charge to manipulate electron motion or store information. So-called spintronic devices, which would also exploit the spin of electrons for information processing and storage, may ultimately allow us to reduce energy consumption while increasing information processing capabilities, giving us multi-functional, high-speed, low-energy electronic technologies. An essential first step, however, is finding appropriate high-performance materials and integrating them into devices that allow us to control their properties well. In the paper “Gate Control of Spin-Layer-Locking FETs and Application to Monolayer LuIO,” recently published in Nano Letters, NCCR MARVEL researchers and colleagues identified lutetium oxide iodide (LuIO) as just such a material. They studied how to control its properties with electric gates—simulating for the first time ever the effect of electronic doping—and provided practical guidelines for building and operating associated devices from such material.

  • Common workflows for computing material properties with various quantum engines

    Using electronic-structure simulations based on density-functional theory to predict material properties has become routine, thanks at least in part to an ever-widening choice of increasingly robust simulation packages. This wide selection of codes and methods allows for cross-verification, useful in ascertaining accuracy and reliability. But the wide range of methods, algorithms and paradigms available make it difficult for non-experts to select or efficiently use any one for a given task. In the paper “Common workflows for computing material properties using different quantum engines,” published today in npj Computational Materials, a team led by researchers in NCCR MARVEL and in the MaX CoE  shows how the development of common interfaces for workflows that automatically compute material properties can address these challenges and demonstrate the approach with an implementation involving 11 different simulation codes. Also thanks to the use of the AiiDA workflow engine, they guarantee reproducibility of the simulations, simplify interoperability and cross-verification, and open up the use of quantum engines to a wider range of researchers.   

  • OPTIMADE API enables seamless access and interoperability across materials databases

    More than 30 research institutions including NCCR MARVEL have come together to form the Open Databases Integration for Materials Design consortium and develop an API specification enabling seamless access and interoperability among materials databases. The paper “OPTIMADE, an API for exchanging materials data,” published today in Nature Research’s journal Scientific Data, presents the OPTIMADE API specification, illustrates its use and discusses future prospects and ongoing development. 

  • Machine learning cracks the oxidation states of crystal structures

    Chemical engineers at EPFL have developed a machine-learning model that can predict a compound’s oxidation state, a property that is so essential that many chemists argue it must be included in the periodic table.