An official website of the United States government
Abstract Interatomic potentials derived with Machine Learning algorithms such as Deep-Neural Networks (DNNs), achieve the accuracy of high-fidelity quantum mechanical (QM) methods in areas traditionally dominated by empirical force fields and allow performing massive simulations. Most DNN potentials were parametrized for neutral molecules or closed-shell ions due to architectural limitations. In this work, we propose an improved machine learning framework for simulating open-shell anions and cations. We introduce the AIMNet-NSE (Neural Spin Equilibration) architecture, which can predict molecular energies for an arbitrary combination of molecular charge and spin multiplicity with errors of about 2–3 kcal/mol and spin-charges with error errors ~0.01e for small and medium-sized organic molecules, compared to the reference QM simulations. The AIMNet-NSE model allows to fully bypass QM calculations and derive the ionization potential, electron affinity, and conceptual Density Functional Theory quantities like electronegativity, hardness, and condensed Fukui functions. We show that these descriptors, along with learned atomic representations, could be used to model chemical reactivity through an example of regioselectivity in electrophilic aromatic substitution reactions.
more »
« less
Machine-learning-assisted modeling of alloy ordering phenomena at the electronic scale through electronegativity
Many studies attribute the excellent properties of high-entropy alloys to the ordering-phenomena. It can be known from density functional theory that the macroscopic properties of the system can be described by the electron density. Electronegativity is related to electron density, and models describing ordering can be established based on electronegativity scales through machine learning. In this study, a large dataset was established and predicted the ordered state corresponding to the alloy composition. The accuracy of the model on the test set was 94%. Furthermore, this study used different methods to explain the machine learning model and learned more model information.
more »
« less
- Award ID(s):
- 1809640
- PAR ID:
- 10553136
- Publisher / Repository:
- AIP
- Date Published:
- Journal Name:
- Applied Physics Letters
- Volume:
- 124
- Issue:
- 11
- ISSN:
- 0003-6951
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
This talk focuses on the principles of 4D-STEM based electron nanodiffraction techniques for defect, strain and short-range ordering analysis using electron diffuse scattering [8, 9]. We review recent progress made in scanning electron nanodiffraction (SEND) data collection, new algorithms based on cepstral analysis, and machine learning based electron DP analysis. These progresses will be highlighted using defect detection, and short-range ordering analysis as application examples. The materials of the study are the medium entropy alloy, CrCoNi, which has exceptional low-temperature mechanical strength and ductility. We will show how SEND helps our understanding of non-random chemical mixing in a CrCoNi alloy, resulting from short-range ordering, behind the mechanical strength in CrCoNi and how these developments provide general opportunities for an atomistic-structure study in advanced alloys.more » « less
-
Abstract The theorems of density functional theory (DFT) establish bijective maps between the local external potential of a many-body system and its electron density, wavefunction and, therefore, one-particle reduced density matrix. Building on this foundation, we show that machine learning models based on the one-electron reduced density matrix can be used to generate surrogate electronic structure methods. We generate surrogates of local and hybrid DFT, Hartree-Fock and full configuration interaction theories for systems ranging from small molecules such as water to more complex compounds like benzene and propanol. The surrogate models use the one-electron reduced density matrix as the central quantity to be learned. From the predicted density matrices, we show that either standard quantum chemistry or a second machine-learning model can be used to compute molecular observables, energies, and atomic forces. The surrogate models can generate essentially anything that a standard electronic structure method can, ranging from band gaps and Kohn-Sham orbitals to energy-conserving ab-initio molecular dynamics simulations and infrared spectra, which account for anharmonicity and thermal effects, without the need to employ computationally expensive algorithms such as self-consistent field theory. The algorithms are packaged in an efficient and easy to use Python code, QMLearn, accessible on popular platforms.more » « less
-
Abstract The construction of a better exchange-correlation potential in time-dependent density functional theory (TDDFT) can improve the accuracy of TDDFT calculations and provide more accurate predictions of the properties of many-electron systems. Here, we propose a machine learning method to develop the energy functional and the Kohn–Sham potential of a time-dependent Kohn–Sham (TDKS) system is proposed. The method is based on the dynamics of the Kohn–Sham system and does not require any data on the exact Kohn–Sham potential for training the model. We demonstrate the results of our method with a 1D harmonic oscillator example and a 1D two-electron example. We show that the machine-learned Kohn–Sham potential matches the exact Kohn–Sham potential in the absence of memory effect. Our method can still capture the dynamics of the Kohn–Sham system in the presence of memory effects. The machine learning method developed in this article provides insight into making better approximations of the energy functional and the Kohn–Sham potential in the TDKS system.more » « less
-
The electronic structures of three highly mismatched alloys (HMAs)—GeC(Sn), Ga(In)NAs, and BGa(In)As—were studied using density functional theory with HSE06 hybrid functionals, with an emphasis on the local environment near the mismatched, highly electronegative atom (B, C, and N). These alloys are known for their counterintuitive reduction in the bandgap when adding the smaller atom, due to a band anticrossing (BAC) or splitting of the conduction band. Surprisingly, the existence of band splitting was found to be completely unrelated to the local displacement of the lattice ions near the mismatched atom. Furthermore, in BGaAs, the reduction in the bandgap due to BAC was weaker than the increase due to the lattice constant, which has not been observed among other HMAs but may explain differences among experimental reports. While local distortion in GeC and GaNAs was not the cause for BAC, it was found to enhance the bandgap reduction due to BAC. This work also found that mere contrast in electronegativity between neighboring atoms does not induce BAC. In fact, surrounding the electronegative atom with elements of even smaller electronegativity than the host (e.g., Sn or In) consistently decreased or even eliminated BAC. For a fixed composition, moving Sn toward C and In toward either N or B was always energetically favorable and increased the bandgap, consistent with experimental annealing results. Such rearrangement also delocalized the conduction band wavefunctions near the mismatched atom to resemble the original host states in unperturbed Ge or GaAs, causing the BAC to progressively weaken. These collective results were consistent whether the mismatched atom was a cation (N), anion (B), or fully covalent (C), varying only with the magnitude of its electronegativity, with B having the least effect. The effects can be explained by charge screening of the mismatched atom's deep electrostatic potential. Together, these results help explain differences in the bandgap and other properties reported for HMAs from different groups and provide insight into the creation of materials with designer properties.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Journal Article:
- https://doi.org/10.1063/5.0188516
