New magnetic materials for energy and information-processing applications are of paramount importance in view of significant global challenges in environmental and information security. The discovery and design of materials requires efficient computational and experimental approaches for high throughput and efficiency. When increasingly powerful computational techniques are combined with special non-equilibrium fabrication methods, the search can uncover metastable compounds with desired magnetic properties. Here we review recent results on novel Fe-, Co- and Mn-rich magnetic compounds with high magnetocrystalline anisotropy, saturation magnetization, and Curie temperature created by combining experiments, adaptive genetic algorithm searches, and advanced electronic-structure computational methods. We discuss structural and magnetic properties of such materials including Co– and/or Fe–X compounds (X = N, Si, Sn, Zr, Hf, Y, C, S, Ti, or Mn), and their prospects for practical applications.
more »
« less
Evolutive screening of candidates for new materials using genetic algorithms and deep learning
ifferent mechanisms are used for the discovery of materials. These include creating a material by trial-and-error process without knowing its properties. Other methods are based on computational simulations or mathematical and statistical approaches, such as Density Functional Theory (DFT). A well-known strategy combines elements to predict their properties and selects a set of those with the properties of interest. Carrying out exhaustive calculations to predict the properties of these found compounds may require a high computational cost. Therefore, there is a need to create methods for identifying materials with a desired set of properties while reducing the search space and, consequently, the computational cost. In this work, we present a genetic algorithm that can find a higher percentage of compounds with specific properties than state-of-the-art methods, such as those based on combinatorial screening. Both methods are compared in the search for ternary compounds in an unconstrained space, using a Deep Neural Network (DNN) to predict properties such as formation enthalpy, band gap, and stability; we will focus on formation enthalpy. As a result, we provide a genetic algorithm capable of finding up to 60% more compounds with atypical values of properties, using DNNs for their prediction.
more »
« less
- Award ID(s):
- 1750970
- PAR ID:
- 10513258
- Publisher / Repository:
- Latin American and Caribbean Consortium of Engineering Institutions
- Date Published:
- Journal Name:
- Proceedings of the LACCEI international multiconference for engineering education and technology
- ISSN:
- 2414-6390
- ISBN:
- 9786289520743
- Format(s):
- Medium: X
- Location:
- Buenos Aires
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
The search for new magnetic materials with high magnetization and magnetocrystalline anisotropy is important for a wide range of applications including information and energy processing. There is only a limited number of naturally occurring magnetic compounds that are suitable. This situation stimulates an exploration of new phases that occur far from thermal-equilibrium conditions, but their stabilization is generally inhibited due to high positive formation energies. Here a nanocluster-deposition method has enabled the discovery of a set of new non-equilibrium Co–N intermetallic compounds. The experimental search was assisted by computational methods including adaptive-genetic-algorithm and electronic-structure calculations. Conventional wisdom is that the interstitial or substitutional solubility of N in Co is much lower than that in Fe and that N in Co in equilibrium alloys does not produce materials with significant magnetization and anisotropy. By contrast, our experiments identify new Co–N compounds with favorable magnetic properties including hexagonal Co 3 N nanoparticles with a high saturation magnetic polarization ( J s = 1.28 T or 12.8 kG) and an appreciable uniaxial magnetocrystalline anisotropy ( K 1 = 1.01 MJ m −3 or 10.1 Mergs per cm 3 ). This research provides a pathway for uncovering new magnetic compounds with computational efficiency beyond the existing materials database, which is significant for future technologies.more » « less
-
We develop an open-access database that provides a large array of datasets specialized for magnetic compounds as well as magnetic clusters. Our focus is on rare-earth-free magnets. Available datasets include (i) crystallography, (ii) thermodynamic properties, such as the formation energy, and (iii) magnetic properties that are essential for magnetic-material design. Our database features a large number of stable and metastable structures discovered through our adaptive genetic algorithm (AGA) searches. Many of these AGA structures have better magnetic properties when compared to those of the existing rare-earth-free magnets and the theoretical structures in other databases. Our database places particular emphasis on site-specific magnetic data, which are obtained by high-throughput first-principles calculations. Such site-resolved data are indispensable for machine-learning modeling. We illustrate how our data-intensive methods promote efficiency of the experimental discovery of new magnetic materials. Our database provides massive datasets that will facilitate an efficient computational screening, machine-learning-assisted design, and the experimental fabrication of new promising magnets.more » « less
-
Domain‐based local pair natural orbital methods within the correlation consistent composite approachmore » « less
Ab initio composite approaches have been utilized to model and predict main group thermochemistry within 1 kcal mol−1, on average, from well‐established reliable experiments, primarily for molecules with less than 30 atoms. For molecules of increasing size and complexity, such as biomolecular complexes, composite methodologies have been limited in their application. Therefore, the domain‐based local pair natural orbital (DLPNO) methods have been implemented within the correlation consistent composite approach (ccCA) framework, namely DLPNO‐ccCA, to reduce the computational cost (disk space, CPU (central processing unit) time, memory) and predict energetic properties such as enthalpies of formation, noncovalent interactions, and conformation energies for organic biomolecular complexes including one of the largest molecules examined via composite strategies, within 1 kcal mol−1, after calibration with 119 molecules and a set of linear alkanes. © 2019 Wiley Periodicals, Inc.
-
null (Ed.)Abstract Accurate theoretical predictions of desired properties of materials play an important role in materials research and development. Machine learning (ML) can accelerate the materials design by building a model from input data. For complex datasets, such as those of crystalline compounds, a vital issue is how to construct low-dimensional representations for input crystal structures with chemical insights. In this work, we introduce an algebraic topology-based method, called atom-specific persistent homology (ASPH), as a unique representation of crystal structures. The ASPH can capture both pairwise and many-body interactions and reveal the topology-property relationship of a group of atoms at various scales. Combined with composition-based attributes, ASPH-based ML model provides a highly accurate prediction of the formation energy calculated by density functional theory (DFT). After training with more than 30,000 different structure types and compositions, our model achieves a mean absolute error of 61 meV/atom in cross-validation, which outperforms previous work such as Voronoi tessellations and Coulomb matrix method using the same ML algorithm and datasets. Our results indicate that the proposed topology-based method provides a powerful computational tool for predicting materials properties compared to previous works.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Conference Paper:
- https://doi.org/10.18687/LACCEI2023.1.1.651
