There is growing interest in applying machine learning techniques in the research of materials science. However, although it is recognized that materials datasets are typically smaller and sometimes more diverse compared to other fields, the influence of availability of materials data on training machine learning models has not yet been studied, which prevents the possibility to establish accurate predictive rules using small materials datasets. Here we analyzed the fundamental interplay between the availability of materials data and the predictive capability of machine learning models. Instead of affecting the model precision directly, the effect of data size is mediated by the degree of freedom (DoF) of model, resulting in the phenomenon of association between precision and DoF. The appearance of precision-DoF association signals the issue of underfitting and is characterized by large bias of prediction, which consequently restricts the accurate prediction in unknown domains. We proposed to incorporate the crude estimation of property in the feature space to establish ML models using small sized materials data, which increases the accuracy of prediction without the cost of higher DoF. In three case studies of predicting the band gap of binary semiconductors, lattice thermal conductivity, and elastic properties of zeolites, the integration of crude estimation effectively boosted the predictive capability of machine learning models to state-of-art levels, demonstrating the generality of the proposed strategy to construct accurate machine learning models using small materials dataset.

A high energy-density Sn anode capable of displaying superior operating voltages and capacity, for rechargeable Mg-ion batteries, is highlighted. The intended application and performance of the anode is confirmed by coupling it with a Mo(6)S(8) cathode in a conventional battery electrolyte.

No abstract

Here we report the study of Li and Li oxides insertion in αMnO 2 with firstprinciples density functional theory calculations. For Li insertion, the redox reaction is characterized with a particular order of the reduction of Mn ions. It induces asymmetric changes of lattice parameters through Jahn−Teller distortion of MnO 6 units. At composition of LiMn 2 O 4 , the ratio between lattice parameter a and b extends to the highest value, 1.19. The severe structural deformation is concluded to be the major cause of the irreversible capacity when αMnO 2 is used as Li-ion battery cathodes. On the other hand, the structure of αMnO 2 is stable to reversibly insert and remove Li oxides. Li oxides inserted αMnO 2 is half metallic instead of insulating. On the basis of our results, a mechanism is proposed to explain the role of αMnO 2 in the catalyzed decomposition of Li 2 O 2 . The key step in this mechanism is the insertion of Li oxides in αMnO 2 . The insertion transfers electrons from Li oxides to MnO 2 and partially oxidized Li oxides, making the subsequent release of O 2 easier.

The Na-ion battery has recently gained a lot of interest as a low-cost alternative to the current Li-ion battery technology. Its feasibility strongly depends on the development of suitable electrode materials. In the present work we propose a novel anode candidate, boron-doped graphene, for the Na-ion battery. Our first-principles calculations demonstrate that the sodiation of boron-doped graphene well preserves its structural integrity. The 2D-BC3 anode has the average sodiation voltage of 0.44 V in an appropriate range to avoid the safety concerns caused by the formation of dendritic deposits. The capacity of the 2D-BC3 anode reaches ∼2.04 times that of the graphite anode in a Li-ion battery and ∼2.52 times that of hard carbon in a Na-ion battery. The high electronic mobility and Na mobility on boron-doped graphene indicates that it has a high potential to reach good rate performance. These suggest the promising potential of boron-doped graphene to serve as an anode for a rechargeable Na-ion battery.