Performance and Cost of Machine Learning Interatomic Potentials

Our work on “Performance and Cost Assessment of Machine Learning Interatomic Potentials (ML-IAPs)” has been published in the Journal of Physical Chemistry A! Co-authored with the developers of four leading ML-IAPs, this work provides a rigorous assessment of ML-IAPs across several metrics – accuracy in energies and forces, materials properties and training and computing cost. This assessment was carried out using a diverse data set – bcc (Li, Mo) and fcc (Cu, Ni) metals and diamond group IV semiconductors (Si, Ge) – generated using high-throughput density functional theory (DFT) calculations. To facilitate the reuse and reproduction of our results, the code, data and optimized ML models in this work are published open-source on our mlearn Github repo. The code includes high-level Python interfaces for ML-IAPs development as well as LAMMPS material properties calculators. Check out the publication at this link.

Grain Boundary Database

GBDB

We are pleased to announce the release of the Grain Boundary Database (GBDB) together with the associated publication in Acta Materialia! The GBDB is the largest database of DFT-computed grain boundary properties to date, encompassing 327 GBs of 58 elemental metals. To construct the GBDB, we developed a novel scaled-structural template approach for GB calculations, which reduces the computational cost of converging GB structures by a factor of ~ 3–6. The grain boundary energies and work of separation have been rigorously validated against previous experimental and computational data. You can check out the GBDB at Crystalium@Materials Virtual Lab or the Materials Project.

Tuning Photoluminescence in K3YSi2O7

Our collaborative paper with the group of Zhiguo Xia in South China University of Technology on “Engineering of K3YSi2O7 To Tune Photoluminescence with Selected Activators and Site Occupancy” has been published in Chemistry of Materials. We have discovered the Eu- and Ce-activated K3YSi2O7 phosphors, which exhibit orange-red and green emission, respectively. Using DFT calculations, we show that Eu2+ occupies both K1 and Y2 crystallographic sites, while Ce3+ and Eu3+ only occupy the Y2 site. Hence, the broad-band red emission of Eu2+ are attributed to a small DFT band gap (3.69 eV) of K3YSi2O7 host and a selective occupancy of Eu2+ in a highly distorted K1 site and a high crystal field splitting around Y2 sites.

Congratulations Zhuoying and Chen Zheng!

The Materials Virtual Lab is pleased to announce two newly minted PhD graduates – Dr Zhuoying Zhu and Dr Chen Zheng! Congratulations on their successful PhD thesis defense! Both Zhuoying and Chen joined the group in 2014. They have taken radically different, but equally fruitful research paths over the course of their PhD career. Zhuoying’s thesis work is on superionic conductors solid electrolytes for all-solid-state rechargeable alkali-ion batteries. Using computational methods, she has successfully predicted completely novel lithium and sodium superionic conductors with superior ionic conductivity and electrochemical stability, several of which have already been realized experimentally. Her work showcases how first principles computations, combined with the application of thermodynamics and kinetics, can help accelerate materials optimization and discovery for technological applications. Chen’s thesis is focused on the relatively nascent field of machine learning (ML) in materials science. He has developed ML models that can interpret X-ray absorption spectra with accuracies exceeding that of humans, as well as applied graph-based deep learning techniques developed in our group to accelerate the exploration of vast compositional spaces for cathode materials.

Discovery of Novel Super-Broadband Phosphor

In a collaboration with the Xie group from Xiamen, our group is pleased to announce yet another novel phosphor discovered via first-principles computations. Sr2AlSi2O6N:Eu2+ has a superbroad emission with a bandwidth of 230 nm, the broadest emission bandwidth ever reported, and has excellent thermal quenching resistance (88% intensity at 150°C). A prototype white LED utilizing only this full-visible-spectrum phosphor exhibits superior color quality (Ra = 97, R9 = 91), outperforming commercial tricolor phosphor-converted LEDs. This work is published in Chemistry of Materials and is co-authored by Shuxing Li of the Xie group with Materials Virtual Lab alumnus Zhenbin Wang.

Water Contributes to Higher Energy Density and Cycling Stability of Prussian Blue Analogue Cathodes for Aqueous Sodium-Ion Batteries

Xingyu’s first paper titled “Water Contributes to Higher Energy Density and Cycling Stability of Prussian Blue Analogue Cathodes for Aqueous Sodium-Ion Batteries” is now published in Chemistry of Materials! In this work, we show that dry Prussian blue analogues (PBAs), one of the most promising cathode materials for aqueous sodium-ion batteries for large-scale energy-storage systems, generally undergo a phase transition from a rhombohedral Na2PR(CN)6 to a tetragonal/cubic PR(CN)6 during Na extraction. However, the presence of water fundamentally alters this phsae behavior, increasing an increase in the average voltage and a reduction in volume change during electrochemical cycling, resulting in both higher energy density and better cycling stability. We also identified four new promising PBA compositions, Na2CoMn(CN)6, Na2NiMn(CN)6, Na2CuMn(CN)6 and Na2ZnMn(CN)6 for further exploration.

Li3N eSNAP potential and other publications

Li3N arrhenius plot

Zhi Deng is the lead author in our recently published work in npj Computational Materials on a machine-learned (ML) electrostatic Spectral Neighbor Analysis Potential (eSNAP) for Li3N, a prototypical superionic conductor. By incorporating long-ranged electrostatics, we developed a highly accurate eSNAP model for Li3N that far outperforms traditional potentials in the prediction of energies, forces and properties such as lattice constants, elastic constants, and phonon dispersion curves. Most importantly, we demonstrate that the eSNAP enables long-time, large-scale Li diffusion studies in Li3N, computing the Haven ratio and simulating GB diffusion in Li3N for the first time to excellent agreement with experimental values. Our group members are also co-authors in several recently published works. Group alumnus Zhenbin Wang co-authored “Color Tunable Single-Phase Eu2+ and Ce3+ Co-Activated Sr2LiAlO4 Phosphors” published in Journal of Materials Chemistry C, a work that builds on the Sr2LiAlO4 phosphor previously discovered by our group using data mining and DFT computations to show that co-doping of Eu2+ and Ce3+ can be used to tune the color of the Sr2LiAlO4 phosphor. Zhuoying co-authored a work on “Elucidating the Limit of Li Insertion into the Spinel Li4Ti5O12” published in ACS Materials Letters. Our contribution is using DFT computations to identify […]

Database of Elemental Work Functions

We are pleased to announce that Richard’s follow-up work on the anisotropic work functions of the elements has been published in Surface Science. The work function is a fundamental electronic property of a solid that varies with the facets of a crystalline surface. It is a crucial parameter in spectroscopy as well as materials design, especially for technologies such as thermionic electron guns and Schottky barriers. In this work, we present the largest database of calculated work functions for elemental crystals to date. This well-validated database contains the anisotropic work functions of more than 100 polymorphs of about 72 elements. One significant advance is the development of an improved model for the work function of metals from atomic parameters such as the electronegativity and metallic radius based on Gauss’ law. The work function database can be accessed at the Crystalium website together with other surface properties.

MEGNet framework for Crystal and Molecules

Our paper on MatErials Graph Networks (MEGNet) for machine learning in crystals and molecules have been published in Chemistry of Materials. The article is available here. Key advances include the incorporation of state variables such as temperature, pressure and entropy, transfer learning from models with large data (e.g., formation energies) to models with smaller data (e.g., elastic constants) and extraction of chemical trends from learned elemental embeddings. These advances address key limitations in ML in materials science, such as data size limitations and physical interpretability. We have also released all our codes and data in our open Github repo at https://github.com/materialsvirtuallab/megnet to enable others to reproduce and improve on our models.

AI in Energy Materials

Prof Ong is the Feature Editor in Mar 2019’s MRS Energy Quarterly article on “Artificial intelligence is aiding the search for energy materials”. In this article written by Prachi Patel, we interview various leaders in the field on their perspectives on how AI is being applied in energy materials design, from discovering entirely novel materials to enabling large-scale complex simulations to providing insights into how to synthesize materials. Check it out at https://doi.org/10.1557/mrs.2019.51.