Density functional theory and material databases in the era of machine learning DOI
Arti Kashyap

Applied Physics Letters, Journal Year: 2024, Volume and Issue: 125(22)

Published: Nov. 25, 2024

This perspective article presents the density functional theory and traces its evolution. With advancement in theory-based computations efforts to collate data generated through theory, field now has a good repository/database of materials their properties. repository, though not as substantial generally used for machine learning, nonetheless made it possible combine learning. highlights current research challenges an optimistic outlook future “Density Functional Theory with Machine Learning” by discussing some specific examples.

Language: Английский

Hybrid Semi-mechanistic and Machine Learning Solubility Regression Modeling for Crystallization Process Development DOI
Gustavo Lunardon Quilló, Satyajeet Bhonsale, A. Collas

et al.

Crystal Growth & Design, Journal Year: 2025, Volume and Issue: unknown

Published: Feb. 10, 2025

Solubility regression modeling is foundational for several chemical engineering applications, particularly crystallization process development. Traditionally, these models rely on parametric semimechanistic approaches such as the Van't Hoff Jouyban-Acree (VH-JA) cosolvency model. Although generally provide narrow prediction intervals, they can exhibit increased bias when dealing with significant solute heat capacities or complex mixture effects. This study explores machine learning, including Random Forests, Support Vector Machines, Gaussian Process Regression, and Neural Networks, potential alternatives. While most learning offered a lower training error, it was observed that their predictive quality quickly deteriorates further from data. Hence, hybrid approach explored to leverage low of variance VH-JA model through heterogeneous locally weighted bagging ensembles. Key methodology quantifying, tracking, minimizing uncertainty using ensemble. illustrated case solubility ketoconazole in binary mixtures 2-propanol water. The optimal ensemble, comprising 58% stepwise 42% models, reduced root-mean-squared error maximum absolute percentage by ≈30% compared full VH-JA, while preserving comparable interval.

Language: Английский

Citations

0
Flux‐Regulated Crystallization of Perovskites Using Machine Learning‐Predicted Solvent Evaporation Rates for X‐Ray Detectors DOI Creative Commons
Tatiane Pretto, Sergey Dayneko, Antoine Pavesic

et al.

Advanced Functional Materials, Journal Year: 2025, Volume and Issue: unknown

Published: April 17, 2025

Abstract Flux‐regulated crystallization (FRC), a method that dynamically monitors and adjusts crystal growth from solutions in real time using computer vision feedback control, has been recently introduced. Using FRC, centimeter‐scale perovskite single crystals at linear rate of 0.2 mm h −1 with standard deviation ( σ ) 0.061 is synthesized. Here, machine learning integrated into FRC to predict solvent evaporation rates during time, thus leading an over threefold decrease 0.018 . This also results improved reproducibility crystallinity, as evidenced by average full width half maximum 22 ± 5 arcsec X‐ray rocking curve measurements; detectors, sensitivity 4500 500 µC Gy air cm −2 electric field 100 V across 13 devices.

Language: Английский

Citations

0
Prioritizing Computational Cocrystal Prediction Methods for Experimental Researchers: A Review to Find Efficient, Cost-Effective, and User-Friendly Approaches DOI Open Access
Beáta Lemli, Szilárd Pál, Ala’ Salem

et al.

International Journal of Molecular Sciences, Journal Year: 2024, Volume and Issue: 25(22), P. 12045 - 12045

Published: Nov. 9, 2024

Pharmaceutical cocrystals offer a versatile approach to enhancing the properties of drug compounds, making them an important tool in formulation and development by improving therapeutic performance patient experience pharmaceutical products. The prediction involves using computational theoretical methods identify potential cocrystal formers understand interactions between active ingredient coformers. This process aims predict whether two or more molecules can form stable structure before performing experimental synthesis, thus saving time resources. In this review, commonly used are first overviewed then evaluated based on three criteria: efficiency, cost-effectiveness, user-friendliness. Based these considerations, we suggest researchers without strong experiences which tools should be tested as step workflow rational design cocrystals. However, optimal choice depends specific needs resources, combining from different categories powerful approach.

Language: Английский

Citations

2
A new coamorphous ethionamide with enhanced solubility: Preparation, characterization, in silico pharmacokinetics, and controlled release by encapsulation DOI
João G. de Oliveira Neto, Raychimam D. S. Bezerra,

F. Domingos

et al.

International Journal of Pharmaceutics, Journal Year: 2024, Volume and Issue: unknown, P. 125159 - 125159

Published: Dec. 1, 2024

Language: Английский

Citations

1
Applications of machine learning for modeling and advanced control of crystallization processes: Developments and perspectives DOI Creative Commons
Fernando Arrais Romero Dias Lima, Marcellus G.F. de Moraes, Evaristo C. Biscaia

et al.

Digital Chemical Engineering, Journal Year: 2024, Volume and Issue: unknown, P. 100208 - 100208

Published: Dec. 1, 2024

Language: Английский

Citations

1
On-line image analysis for evaporative crystallization of xylose DOI

Qihang Zhu,

Guangzheng Zhou, Gary G. Hou

et al.

Powder Technology, Journal Year: 2024, Volume and Issue: unknown, P. 120446 - 120446

Published: Nov. 1, 2024

Language: Английский

Citations

0
In Situ Microscopy with Real-Time Image Analysis Enables Online Monitoring of Technical Protein Crystallization Kinetics in Stirred Crystallizers DOI Creative Commons

Julian Mentges,

Daniel Bischoff, Brigitte Walla

et al.

Crystals, Journal Year: 2024, Volume and Issue: 14(12), P. 1009 - 1009

Published: Nov. 21, 2024

Controlling protein crystallization processes is essential for improving downstream processing in biotechnology. This study investigates the combination of machine learning-based image analysis and situ microscopy real-time monitoring kinetics. The experimental research focused on batch an alcohol dehydrogenase from Lactobacillus brevis (LbADH) two selected rational crystal contact mutants. Technical experiments were performed a 1 L stirred crystallizer by adding polyethyleneglycol 550 monomethyl ether (PEG MME). estimated volumes online correlated well with offline measured concentrations solution. In addition, was superior to data if amorphous precipitation occurred. Real-time provides basis estimation important performance indicators like yield, kinetics, size distributions, number crystals. Surprisingly, one LbADH mutants, which should theoretically crystallize more slowly than wild type based molecular dynamics (MD) simulations, showed better except yield. Thus, scalable improves precision studies industrial settings providing comprehensive data, reducing limitations traditional analytical techniques, enabling new insights into process dynamics.

Language: Английский

Citations

0
Density functional theory and material databases in the era of machine learning DOI
Arti Kashyap

Applied Physics Letters, Journal Year: 2024, Volume and Issue: 125(22)

Published: Nov. 25, 2024

This perspective article presents the density functional theory and traces its evolution. With advancement in theory-based computations efforts to collate data generated through theory, field now has a good repository/database of materials their properties. repository, though not as substantial generally used for machine learning, nonetheless made it possible combine learning. highlights current research challenges an optimistic outlook future “Density Functional Theory with Machine Learning” by discussing some specific examples.

Language: Английский

Citations

0