Computing vibrational free energies (F vib ) and entropies (S vib ) has posed a long standing challenge to the high-throughput ab initio investigation of finite temperature properties of solids. Here we use machine-learning techniques to efficiently predict F vib and S vib of crystalline compounds in the Inorganic Crystal Structure Database. By employing descriptors based simply on the chemical formula and using a training set of only 300 compounds, mean absolute errors of less than 0.04 meV/K/atom (15 meV/atom) are achieved for S vib (F vib ), whose values are distributed within a range of 0.9 meV/K/atom (300 meV/atom.) In addition, for training sets containing fewer than 2,000 compounds the chemical formula alone is shown to perform as well as, if not better than, four other more complex descriptors previously used in the literature. The accuracy and simplicity of the approach mean that it can be advantageously used for the fast screening of phase diagrams or chemical reactions at finite temperatures.

No abstract

Machine learning (ML) is increasingly becoming a helpful tool in the search for novel functional compounds. Here we use classification via random forests to predict the stability of half-Heusler (HH) compounds, using only experimentally reported compounds as a training set. Cross-validation yields an excellent agreement between the fraction of compounds classified as stable and the actual fraction of truly stable compounds in the ICSD. The ML model is then employed to screen 71,178 different 1:1:1 compositions, yielding 481 likely stable candidates. The predicted stability of HH compounds from three previous high throughput ab initio studies is critically analyzed from the perspective of the alternative ML approach. The incomplete consistency among the three separate ab initio studies and between them and the ML predictions suggests that additional factors beyond those considered by ab initio phase stability calculations might be determinant to the stability of the compounds. Such factors can include configurational entropies and quasiharmonic contributions.

To assess the potential of amorphous Si (a-Si) as an anode for Li, Na, and Mg-ion batteries, the energetics of Li, Na, and Mg atoms in a-Si are computed from first-principles and compared to those in crystalline Si (c-Si). It is shown that Si preamorphization increases the average anode voltage and reduces the volume expansion of the anode during the insertion of the metal atoms.Analysis of computed formation energies of Li, Na, and Mg defects in a-Si and c-Si suggests that the energetics of the single atoms into a-Si are thermodynamically more favorable. For instance, defect formation energies of Li, Na, and Mg defects in a-Si are respectively 0.71, 1.72, and 1.82 eV lower compared to those in c-Si. Moreover, the defect formation energies of Li, Na, and Mg defects (vs. vacuum reference states) in a-Si are comparable with the metal cohesive energies and consequently the insertion of the metal atoms might be possible with appropriate control of charging process. This is in contrast to c-Si, where the storage of Na and Mg atoms is limited due to high energy cost of Na and Mg insertion into c-Si.

We present a comparative, combined ab initio and experimental study of sodium and lithium storage in amorphous (glassy) carbon (a-C) vs. graphite. Amorphous structures are obtained by fitting stochastically generated structures to a reference radial distribution function.Li insertion is thermodynamically favored in both graphite and a-C. While sodium insertion is thermodynamically unfavored in graphite, a-C possesses multiple insertion sites with binding energies stronger than Na cohesive energy, making it usable as anode material for Na-ion batteries. Binding energy of Na is predicted to be stronger than the Na cohesive energy for Na concentrations corresponding to a capacity of about 200 mAh/g. These results are confirmed by experimental measurements using highly amorphous carbon, in which a specific capacity of 173 mAh/g for Na is obtained after 100 cycles.

While Si is an effective insertion type anode for Li-ion batteries, crystalline Si has been shown to be unsuitable for Na and Mg storage due, in particular, to insufficient binding strength. It has recently been reported that Si nanowires could be synthesized with high-concentration (several atomic %) and dispersed Al doping. Here we show based on density functional theory calculations that Al doping significantly improves the energetics for Na and Mg insertion, specifically, making it thermodynamically favored versus vacuum reference states. For high Al concentrations, the energy of Mg in Al-doped Si approaches the cohesive energy of Mg. However, the migration barriers for the diffusion of Li (0.57-0.70 eV), Na (1.07-1.19 eV) and Mg (0.97-1.18 eV) in Al-doped Si are found to remain about as high as in pure Si, likely preventing effective electrochemical sodiation and magnesiation.

No abstract

Machine learning approaches, enabled by the emergence of comprehensive databases of materials properties, are becoming a fruitful direction for materials analysis. As a result, a plethora of models have been constructed and trained on existing data to predict properties of new systems. These powerful methods allow researchers to target studies only at interesting materialsneglecting the non-synthesizable systems and those without the desired properties -thus reducing the amount of resources spent on expensive computations and/or time-consuming experimental synthesis. However, using these predictive models is not always straightforward. Often, they require a panoply of technical expertise, creating barriers for general users. AFLOW-ML (AFLOW Machine Learning) overcomes the problem by streamlining the use of the machine learning methods developed within the AFLOW consortium. The framework provides an open RESTful API to directly access the continuously updated algorithms, which can be transparently integrated into any workflow to retrieve predictions of electronic, thermal and mechanical properties. These types of interconnected cloud-based applications are envisioned to be capable of further accelerating the adoption of machine learning methods into materials development.