Enhancing crystal structure prediction by combining computational and experimental data via graph networks DOI Creative Commons
Chenglong Qin, Jinde Liu, Shiyin Ma

et al.

arXiv (Cornell University), Journal Year: 2023, Volume and Issue: unknown

Published: Jan. 1, 2023

Crystal structure prediction (CSP) stands as a powerful tool in materials science, driving the discovery and design of innovative materials. However, existing CSP methods heavily rely on formation enthalpies derived from density functional theory (DFT) calculations, often overlooking differences between DFT experimental values. Moreover, material synthesis is intricately influenced by factors such kinetics conditions. To overcome these limitations, novel collaborative approach was proposed for that combines with data, utilizing advanced deep learning models optimization algorithms. We illustrate capability to predict closely align actual observations through transfer data. By incorporating synthesizable information crystals, our model capable reverse engineering crystal structures can be synthesized experiments. Applying 17 representative compounds, results indicate accurately identify experimentally high precision. obtained lattice constants values, underscoring model's effectiveness. The synergistic theoretical data bridges longstanding disparities predictions results, thereby alleviating demand extensive costly trials.

Language: Английский

Applications of Artificial Intelligence and Machine Learning Algorithms to Crystallization DOI
Christos Xiouras, Fabio Cameli, Gustavo Lunardon Quilló

et al.

Chemical Reviews, Journal Year: 2022, Volume and Issue: 122(15), P. 13006 - 13042

Published: June 27, 2022

Artificial intelligence and specifically machine learning applications are nowadays used in a variety of scientific cutting-edge technologies, where they have transformative impact. Such an assembly statistical linear algebra methods making use large data sets is becoming more integrated into chemistry crystallization research workflows. This review aims to present, for the first time, holistic overview cheminformatics as novel, powerful means accelerate discovery new crystal structures, predict key properties organic crystalline materials, simulate, understand, control dynamics complex process systems, well contribute high throughput automation chemical development involving materials. We critically advances these new, rapidly emerging areas, raising awareness issues such bridging models with first-principles mechanistic models, set size, structure, quality, selection appropriate descriptors. At same we propose future at interface applied mathematics, chemistry, crystallography. Overall, this increase adoption tools by chemists scientists across industry academia.

Language: Английский

Citations

76
Optimality guarantees for crystal structure prediction DOI
Vladimir V. Gusev, Duncan Adamson, Argyrios Deligkas

et al.

Nature, Journal Year: 2023, Volume and Issue: 619(7968), P. 68 - 72

Published: July 5, 2023

Language: Английский

Citations

31
Towards quantitative evaluation of crystal structure prediction performance DOI Creative Commons
Lai Wei, Q. Li, Sadman Sadeed Omee

et al.

Computational Materials Science, Journal Year: 2024, Volume and Issue: 235, P. 112802 - 112802

Published: Jan. 23, 2024

Language: Английский

Citations

7
Recent progress in generative adversarial networks applied to inversely designing inorganic materials: A brief review DOI
Rahma Jabbar, Rateb Jabbar, S. Kamoun

et al.

Computational Materials Science, Journal Year: 2022, Volume and Issue: 213, P. 111612 - 111612

Published: July 1, 2022

Language: Английский

Citations

24
Accelerating Materials Discovery through Machine Learning: Predicting Crystallographic Symmetry Groups DOI Creative Commons

Yousef A. Alghofaili,

Mohammed Alghadeer, Abdulmohsen Alsaui

et al.

The Journal of Physical Chemistry C, Journal Year: 2023, Volume and Issue: 127(33), P. 16645 - 16653

Published: Aug. 11, 2023

Predicting crystal structure from the chemical composition is one of most challenging and long-standing problems in condensed matter physics. This problem resides at interface between materials sciences With reliable data proper physics-guided modeling, machine learning (ML) can provide an alternative venue to undertake reduce problem's complexity. In this work, very robust ML classifiers for crystallographic symmetry groups were developed applied ternary (AlBmCn) binary (AlBm) starting only formula. first essential step toward predicting full geometry. Such a highly multi-label multi-class perspective requires careful preprocessing due size imbalance data. The resulting predictive models are accurate all groups, including systems, point Bravais lattices, space with weighted balanced accuracies exceeding 95%. small set ionic compositional features, namely, stoichiometry, radii, ionization energies, oxidation states each element compounds. Considering such minimal feature space, obtained high ascertain that physics well captured. even further confirmed as we demonstrate accuracy our approach limited by comparing models. presented work could effectively contribute accelerating new discovery development.

Language: Английский

Citations

15
TCSP: a Template-Based Crystal Structure Prediction Algorithm for Materials Discovery DOI
Lai Wei, Nihang Fu, Edirisuriya M. Dilanga Siriwardane

et al.

Inorganic Chemistry, Journal Year: 2022, Volume and Issue: 61(22), P. 8431 - 8439

Published: April 14, 2022

Fast and accurate crystal structure prediction (CSP) algorithms web servers are highly desirable for the exploration discovery of new materials out infinite chemical design space. However, currently, computationally expensive first-principles calculation-based CSP applicable to relatively small systems reach most researchers. Several teams have used an element substitution approach generating or predicting structures, but usually in ad hoc way. Here we develop a template-based (TCSP) algorithm its companion server, which makes this tool accessible all Our uses elemental/chemical similarity oxidation states guide selection template structures then rank them based on compatibility can return multiple predictions with ranking scores few minutes. A benchmark study 98290 formulas Materials Project database using leave-one-out evaluation shows that our achieve high accuracy (for 13145 target TCSP predicted their root-mean-square deviation < 0.1) large portion formulas. We also discover Ga-B-N system, showing potential high-throughput discovery. user-friendly app be accessed freely at www.materialsatlas.org/crystalstructure MaterialsAtlas.org platform.

Language: Английский

Citations

21
Machine Learning-Driven Prediction of Composite Materials Properties Based on Experimental Testing Data DOI Open Access
Kristina Berladir, Katarzyna Antosz, Vitalii Ivanov

et al.

Polymers, Journal Year: 2025, Volume and Issue: 17(5), P. 694 - 694

Published: March 5, 2025

The growing demand for high-performance and cost-effective composite materials necessitates advanced computational approaches optimizing their composition properties. This study aimed at the application of machine learning prediction optimization functional properties composites based on a thermoplastic matrix with various fillers (two types fibrous, four dispersed, two nano-dispersed fillers). experimental methods involved material production through powder metallurgy, further microstructural analysis, mechanical tribological testing. analysis revealed distinct structural modifications interfacial interactions influencing key findings indicate that optimal filler selection can significantly enhance wear resistance while maintaining adequate strength. Carbon fibers 20 wt. % improved (by 17–25 times) reducing tensile strength elongation. Basalt 10 provided an effective balance between reinforcement 11–16 times). Kaolin 2 greatly enhanced 45–57 moderate reduction. Coke maximized 9−15 acceptable Graphite ensured wear, as higher concentrations drastically decreased Sodium chloride 5 offered improvement 3–4 minimal impact Titanium dioxide 3 11–12.5 slightly Ultra-dispersed PTFE 1 optimized both work analyzed in detail effect content learning-driven prediction. Regression models demonstrated high R-squared values (0.74 density, 0.67 strength, 0.80 relative elongation, 0.79 intensity), explaining up to 80% variability Despite its efficiency, limitations include potential multicollinearity, lack consideration external factors, need validation under real-world conditions. Thus, approach reduces extensive testing, minimizing waste costs, contributing SDG 9. highlights use polymer design, offering data-driven framework rational choice fillers, thereby sustainable industrial practices.

Language: Английский

Citations

0
Machine Learning Supported Annealing for Prediction of Grand Canonical Crystal Structures DOI
Yannick Couzinié, Yuya Seki, Yusuke Nishiya

et al.

Journal of the Physical Society of Japan, Journal Year: 2025, Volume and Issue: 94(4)

Published: March 24, 2025

Language: Английский

Citations

0
Inverse design of experimentally synthesizable crystal structures by leveraging computational and experimental data DOI
Chenglong Qin, Jinde Liu, Shiyin Ma

et al.

Journal of Materials Chemistry A, Journal Year: 2024, Volume and Issue: 12(23), P. 13713 - 13723

Published: Jan. 1, 2024

A novel collaborative approach was proposed for crystal structure prediction that utilizes advanced deep learning models and optimization algorithms combined with experimental data.

Language: Английский

Citations

3
Crystal Structure Prediction DOI
Marta K. Dudek

Royal Society of Chemistry eBooks, Journal Year: 2025, Volume and Issue: unknown, P. 199 - 223

Published: March 31, 2025

In most NMR crystallography applications experimental techniques are used to build an appropriate structural model, which can be later refined using quantum-chemical calculations. some cases, this viewed as obstacle, in particular when constraints extracted from the data ambiguous or not abundant enough. One of promising solutions problem is crystal structure prediction (CSP). On other hand, for complicated, flexible and/or multicomponent systems number degrees freedom (DOF) need accounted CSP starts overwhelming, thus limiting applicability computational method. such instances, solid-state spectra help reduce vast a perfectly manageable DOFs, making combination and calculations very powerful approach. This chapter focuses on context crystallography, including brief overview modern approaches, together with their advantages limitations.

Language: Английский

Citations

0