Weike’s paper on “Deep Neural Networks for Accurate Predictions of Crystal Stability” is now out in Nature Communications. Predicting the stability of crystals is one of the central problems in materials science. Here, we show that deep neural networks, i.e., algorithms that mimic the animal brain, utilizing just the electronegativity and ionic radii as inputs can predict formation energies of crystals with extremely high accuracy. We also demonstrate how these models can be generalized for mixed crystals using a binary encoding scheme, and use it to identify thousands of potentially stable new compositions. We have published a web app (http://crystals.ai) that enables anyone to use these models. News: UCSD News: Scientists use artificial neural networks to predict new stable materials Paper: W. Ye, C. Chen, Z. Wang, I.-H. Chu, S.P. Ong, Deep neural networks for accurate predictions of crystal stability, Nat. Commun. 9 (2018) 3800. doi:10.1038/s41467-018-06322-x.
Our paper on “Mining Unexplored Chemistries for Phosphors for High-Color-Quality White-Light-Emitting Diodes” has been published in Joule. Using supercomputers and data mining, we identified Sr2LiAlO4, the first known Sr-Li-Al-O quaternary crystal, as a highly promising phosphor material in low-cost, high-color-quality white LEDs. Eu2+ and Ce3+-activated Sr2LiAlO4 phosphors exhibit broad green-yellow and blue emissions, respectively, with excellent thermal quenching resistance of > 88% intensity at 150oC. A prototype phosphor-converted white LED utilizing Sr2LiAlO4-based phosphors yields an excellent color rendering index exceeding 90. This work is a collaboration between the Materials Virtual Lab (UCSD), McKittrick group (UCSD) and Im group (Chonnam University). The lead authors are Zhenbin Wang, Jungmin Ha and Yoon Hwa Kim. More information about this work can be found in the Jacobs School of Engineering News as well as Science Daily, Phys.org, etc.
Hui Zheng’s first paper on “Role of Zr in strengthening MoSi2 from density functional theory calculations” has just been published in Acta Materialia. MoSi2 is an important intermetallic with excellent oxidation resistance at high temperatures. However, “pesting” by oxygen limits its application at intermediate temperatures. Using DFT calculations, we show that Zr nanoparticles act as a getter for oxygen, and in the process, significantly strengthens MoSi2 interfaces and grain boundaries. We also use the Materials Project to efficiently screen for other potential getter elements using simple thermodynamic descriptors, a general approach that can be extended to other alloy systems of interest.
Hanmei’s first paper on “Probing Solid-Solid Interfacial Reactions in All-Solid-State Sodium-ion Batteries with First Principles Calculations” has just been published in Chemistry of Materials. In this comprehensive work, we show how explicit AIMD models can lead to different predictions of interfacial reaction products from simple thermodynamic approximations. Specifically, SO4 formation is predicted to be favored over PO4 formation at the interface between NaCoO2 and Na3PS4. We also identified several promising new candidates for buffer materials that potentially show lower reactivity with common electrodes and solid electrolytes.
Chi’s paper on “Accurate force ﬁeld for molybdenum by machine learning large materials data” has just been published in Physical Review Materials. This work addresses a crucial gap in the available force field for Mo. We will show that by ﬁtting to the energies, forces, and stress tensors of a large DFT dataset on a diverse set of Mo structures, a Mo Spectral Neighbor Analysis Potential (SNAP) model can be developed that achieves close to DFT accuracy in the prediction of a broad range of properties, including elastic constants, melting point, phonon spectra, surface energies, grain boundary energies, etc. Examples and parameters of the new potential can be obtained at our Github page.
Paul’s paper titled “Comparison of the polymorphs of VOPO4 as multi-electron cathodes for rechargeable alkali-ion batteries” has just been published in Journal of Materials Chemistry A. In collaboration with the Whittingham group, we performed a systematic first principles investigation, supported by careful electrochemical characterization and published experimental data, of the relative thermodynamic stability, voltage, band gap, and diffusion kinetics for alkali intercalation into the β, ε and αI polymorphs of VOPO4, a highly promising family of multi-electron cathodes. We identify the β polymorph as the most promising for Li insertion, and the αI polymorph as the most promising for Na insertion. We show that differences in the voltage, kinetics and rate capability of these different polymorphs for Li and Na insertion can be traced back to their fundamentally different VO6/VO5–PO4 frameworks.
Hanmei Tang and Iek-Heng Chu are co-authors on “Atomate: A High-Level Interface to Generate, Execute, and Analyze Computational Materials Science Workflows” just published in Computational Materials Science. This paper describes atomate, an open-source Python framework for computational materials science simulation, analysis, and design with an emphasis on automation and extensibility, that is built on top of pymatgen, FireWorks, and custodian. The Materials Virtual Lab are proud contributors to this great open science initiative! Check out the atomate package here.
Our work on “Effects of Transition-Metal Mixing on Na Ordering and Kinetics in Layered P2 Oxides” has just been published in Physical Review Applied. In this work by Chen Zheng and other co-authors, we systematically investigate the effects of transition-metal (TM) mixing on Na ordering and kinetics in the NaxCo1−yMnyO2 model system using DFT calculations. We show that the TM composition at the Na(1) (face-sharing) site has a strong influence on the Na site energies, which in turn impacts the kinetics of Na diffusion towards the end of the charge. By employing a site-percolation model, we establish theoretical upper and lower bounds for TM concentrations based on their effect on Na(1) site energies, providing a framework to rationally tune mixed-TM compositions for optimal Na diffusion.
Our work on “Elucidating Structure–Composition–Property Relationships of the β-SiAlON:Eu2+ Phosphor” has been published in Chemistry of Materials. Using ﬁrst-principles calculations, we identiﬁed and conﬁrmed various chemical rules for Si−Al, O−N, and Eu activator ordering in β-SiAlON, one of the most promising narrow-band green phosphors for high-power light-emitting diodes and liquid crystal display backlighting with wide color gamut. Through the construction of energetically favorable models based on these chemical rules, we studied the eﬀect of oxygen content and Eu2+ activator concentrations on the local EuN9 activator environment, and its impact on important photoluminescence properties such as emission peak position (using the band gap as a proxy), bandwidth, and thermal quenching resistance. Based on these insights, we discuss potential strategies for further composition optimization of β-SiAlON.
We have published the world’s largest database of surface energies and Wulff shapes, dubbed Crystalium. A collaborative effort between the Materials Virtual Lab and the Materials Project, this new open-source database can help researchers design new materials for technologies in which surfaces and interfaces play an important role, such as fuel cells, catalytic converters in cars, computer microchips, nanomaterials and solid-state batteries. You can read more about it in the UCSD press release on EurekAlert! This work is published in Scientific Data as an open-access article, and Richard Tran, one of our undergraduate volunteers, is the first author.