Accelerated mapping of electronic density of states patterns of metallic nanoparticles via machine-learning

Abstract

Within first-principles density functional theory (DFT) frameworks, it is challenging to predict the electronic structures of nanoparticles (NPs) accurately but fast. Herein, a machine-learning architecture is proposed to rapidly but reasonably predict electronic density of states (DOS) patterns of metallic NPs via a combination of principal component analysis (PCA) and the crystal graph convolutional neural network (CGCNN). With the PCA, a mathematically high-dimensional DOS image can be converted to a low-dimensional vector. The CGCNN plays a key role in reflecting the effects of local atomic structures on the DOS patterns of NPs with only a few of material features that are easily extracted from a periodic table. The PCA-CGCNN model is applicable for all pure and bimetallic NPs, in which a handful DOS training sets that are easily obtained with the typical DFT method are considered. The PCA-CGCNN model predicts the R2 value to be 0.85 or higher for Au pure NPs and 0.77 or higher for Au@Pt core@shell bimetallic NPs, respectively, in which the values are for the test sets. Although the PCA-CGCNN method showed a small loss of accuracy when compared with DFT calculations, the prediction time takes just ~ 160 s irrespective of the NP size in contrast to DFT method, for example, 13,000 times faster than the DFT method for Pt147. Our approach not only can be immediately applied to predict electronic structures of actual nanometer scaled NPs to be experimentally synthesized, but also be used to explore correlations between atomic structures and other spectrum image data of the materials (e.g., X-ray diffraction, X-ray photoelectron spectroscopy, and Raman spectroscopy).

Physically Informed Machine Learning Prediction of Electronic Density of States

“…With regards to model design, we note that there have been significant recent advances in graph neural network designs for materials chemistry applications. − These approaches can be incorporated in our framework with only minor changes to our overall workflow. Additional avenues for improvement can also include better output representations for the DOS, such as using a principal component basis ,, or using autoencoders, , which would help by reducing the output dimensions, as long as the reconstruction errors are sufficiently low. Alternatively, coupling the ML predictions to a physical model such as a tight-binding model or a lower level of DFT theory in a Δ-ML approach could be used to improve accuracy and reduce data requirements.…”

Intelligent control of nanoparticle synthesis through machine learning

“…Generally, they are in the ratio of 7 : 3. 65 There are many kinds of predictable machine learning models being selected, but we have to match the selection by weighing between its predictive accuracy, computational cost and feasibility. 66,67 The predictive power is usually affected by the amount of data in the test set, the choice of kernel function and the definition of loss function in the data set.…”

Big Data in a Nano World: A Review on Computational, Data-Driven Design of Nanomaterials Structures, Properties, and Synthesis

“…Emphasizing the importance of physically meaningful descriptors, various studies − have shown excellent mapping of bulk band gaps onto compositional representations for the materials. Single-value scalars such as band gap, ionization energy, and electron affinity are commonly used descriptors, and even the densities of states have been used as a complex input feature for mapping the relationship between materials and their electronic structure. − …”

“…Single-value scalars such as band gap, ionization energy, and electron affinity are commonly used descriptors, 101 and even the densities of states have been used as a complex input feature for mapping the relationship between materials and their electronic structure. 102 − 104 …”