A GNN-Based QSPR Model for Surfactant Properties DOI Creative Commons
Seokgyun Ham, Xin Wang, Hongwei Zhang

et al.

Colloids and Interfaces, Journal Year: 2024, Volume and Issue: 8(6), P. 63 - 63

Published: Nov. 19, 2024

Surfactants are among the most versatile molecules in chemical industry because they can self-assemble bulk solutions and at interfaces. Predicting properties of surfactant solutions, such as their critical micelle concentration (CMC), limiting surface tension (γcmc), maximal packing density (Γmax) water–air interfaces, is essential to rational design. However, relationship between structure these complex difficult predict theoretically. Here, we develop a graph neural network (GNN)-based quantitative structure–property (QSPR) model CMC, γcmc, Γmax. Ninety-two data points, encompassing all types surfactants—anionic, cationic, zwitterionic, nonionic—are fed into model, covering temperature range [20–30 °C], which contributes its generalization across types. We show that our models have high accuracy (R2 = 0.87 on average tests) predicting three parameters surfactants. The effectiveness QSPR capturing variation Γmax with molecular design carefully assessed. curated dataset, developed assessment will contribute development improved surfactants facilitate for diverse applications.

Language: Английский

Experimentally Determined Aqueous Diffusion Coefficients of PFAS Using 19F NMR Diffusion-Ordered Spectroscopy DOI
Jeremy R. Gauthier, Scott A. Mabury

ACS ES&T Water, Journal Year: 2024, Volume and Issue: unknown

Published: Sept. 6, 2024

Language: Английский

Citations

4
Oxidation Stability of Hydrocarbons: A Machine-Learning-Based Study DOI
Adrian Venegas-Reynoso, Benoît Creton, Lucia Giarracca

et al.

Energy & Fuels, Journal Year: 2025, Volume and Issue: 39(9), P. 4361 - 4373

Published: Feb. 24, 2025

Having fluids that are stable over time is important for many applications, particularly sustainable aviation fuels (SAFs) derived from various renewable sources. Being able to understand this characteristic as early possible during the development of SAFs would facilitate blending sources with or without fossil fuels. Oxidation stability, defined a hydrocarbon's resistance reacting oxygen at near-ambient temperatures, one most hydrocarbon-stability-related properties. Indeed, accumulation byproducts oxidation reactions may result in system failures. Assessing property experimentally remains time-consuming; thus developing fast and accurate predictive models becomes relevant approaches based on machine learning appear valuable alternatives. The quantitative structure–property relationships (QSPRs) subject availability reference data, unfortunately, these currently lacking literature. In study, we built database containing consistent experimental results accelerated tests conducted diverse pure hydrocarbons─within carbon atom number range SAFs─using PetroOxy/RapidOxy test method, second, applied two machine-learning-based techniques (SVM XGBoost) generated data set derive QSPR-based models. contribution such augmentation our was also investigated compared more classical approaches. best model (RMSEP = 2.7 h) obtained after log-transforming Induction Period, performing Smart Data Augmentation enrich content, using XGBoost linear learners. While model's accuracy not adequate predictions, it allows semiquantitative predictions.

Language: Английский

Citations

0
A Coarse-Grained Model Describing the Critical Micelle Concentration of Perfluoroalkyl Surfactants in Ionic Aqueous Phase DOI
Eddy Barraud, Christine Dalmazzone, Aurélie Mouret

et al.

Langmuir, Journal Year: 2025, Volume and Issue: unknown

Published: March 14, 2025

In this study, dissipative particle dynamics (DPD) simulations were employed to determine the critical micelle concentration (CMC) of perfluoroalkyl and polyfluoroalkyl substances (PFAS) in ionic aqueous solutions. This approach provides precise CMC data for PFAS surfactants presence various species, thereby addressing a gap current literature. Additionally, study contributes development open-source molecular force fields charged perfluorinated compounds, which are currently limited. These models incorporate hydration free energy values obtained from density functional theory (DFT) account interactions through well-established linear relationship. Hydrophobic between surfactant tail water fine-tuned match chosen surfactants. Then, DPD successfully predicted diverse range surfactants, including those based on hydrocarbons PFAS, demonstrating ability represent realistic salinities encountered natural waters. Experimental validation methodology was conducted using sodium n-nonyl sulfate (SNS) n-dodecyl (SDS) via interfacial tension measurements, confirming accurate representation changes with salinity. enhances our understanding behavior solutions valuable tool predicting complex environmental systems.

Language: Английский

Citations

0
PCL-PEtOx-based Crystalline-core Micelles for the Targeted Delivery of Paclitaxel and Trabectedin in Ovarian Cancer Therapy DOI
Zixiu Du, Wei Wei, Sen Lu

et al.

Acta Biomaterialia, Journal Year: 2025, Volume and Issue: unknown

Published: April 1, 2025

Language: Английский

Citations

0
PFAS adsorption and desorption on functionalized surfaces: A QCM and kinetic modeling approach DOI Creative Commons
Olanrewaju Eunice Beyioku,

Gilboa Arye,

Avner Ronen

et al.

Separation and Purification Technology, Journal Year: 2025, Volume and Issue: unknown, P. 133457 - 133457

Published: May 1, 2025

Language: Английский

Citations

0
Numerical Approaches to Determine Cetane Number of Hydrocarbons and Oxygenated Compounds, Mixtures, and their Blends DOI
Benoît Creton,

Nathalie Brassart,

Amandine Herbaut

et al.

Energy & Fuels, Journal Year: 2024, Volume and Issue: 38(16), P. 15652 - 15661

Published: Aug. 5, 2024

In the present work, we report development and use of models to predict cetane number hydrocarbons oxygenated compounds, mixtures, their blends. The study is divided in three steps: (i) prediction pure compounds' CN using ML-based approaches, (ii) application mixing rules, (iii) external validation on a set real fuels. Experimental values for 658 compounds are collected from literature merged obtain consistent comprehensive database. then trained A second database built collection 572 experimental mixtures. Existing proposed rules powered by either or predicted assessed basis new rule involving activity coefficients mixtures' components shows best performance. Finally, our predictive numerical approach 27 fuels demonstrates its accuracy relevance, that it could be further used testing large numbers samples.

Language: Английский

Citations

3
Current Challenges in Monitoring Low Contaminant Levels of Per- and Polyfluoroalkyl Substances in Water Matrices in the Field DOI Creative Commons
Hector Medina, Carson Farmer

Toxics, Journal Year: 2024, Volume and Issue: 12(8), P. 610 - 610

Published: Aug. 20, 2024

The Environmental Protection Agency (EPA) of the United States recently released first-ever federal regulation on per- and polyfluoroalkyl substances (PFASs) for drinking water. While this represents an important landmark, it also brings about compliance challenges to stakeholders in water industry as well concerns general public. In work, we address some most associated with measuring low concentrations PFASs field real matrices. First, review "continuous monitoring compliance" process laid out by EPA hurdles. requires measuring, frequency, (e.g., below 2 ppt or ng/L) targeted PFASs, presence many other co-contaminants various conditions. Currently, task can only (and is expected to) be accomplished using specific protocols that rely expensive, specialized, laboratory-scale instrumentation, which adds time increases cost. To potentially reduce burden, portable, high-fidelity, low-cost, real-time PFAS sensors are desirable; however, path commercialization promising technologies confronted challenges, well, they still at infant stages. Here, provide insights related those based results from

Language: Английский

Citations

3
Analysis of oral and inhalation toxicity of per- and polyfluoroalkylated organic compounds in rats and mice using multivariate QSAR DOI
Nuno Silva, Eduardo Borges de Melo

SAR and QSAR in environmental research, Journal Year: 2024, Volume and Issue: 35(10), P. 877 - 897

Published: Oct. 2, 2024

Per- and polyfluoroalkylated organic compounds (PFAs) are versatile extensively used in global industries. However, they also persistent pollutants (POPs). This study aimed to develop new models for assessing oral inhalation toxicity rat mice models. A set of 407 PFAs from the literature was divided into four groups based on endpoints interest. The were constructed using only 2D structure descriptors derived SMILES strings. resulting showed a strong statistical quality all endpoints. They present an applicability domain (AD) that ensures good reliability, provided meaningful interpretation, which partially supported by existing literature. Consequently, these valuable understanding how exert their toxic effect mammals predicting risk associated with significant industrial chemical agents.

Language: Английский

Citations

0
A GNN-Based QSPR Model for Surfactant Properties DOI Creative Commons
Seokgyun Ham, Xin Wang, Hongwei Zhang

et al.

Colloids and Interfaces, Journal Year: 2024, Volume and Issue: 8(6), P. 63 - 63

Published: Nov. 19, 2024

Surfactants are among the most versatile molecules in chemical industry because they can self-assemble bulk solutions and at interfaces. Predicting properties of surfactant solutions, such as their critical micelle concentration (CMC), limiting surface tension (γcmc), maximal packing density (Γmax) water–air interfaces, is essential to rational design. However, relationship between structure these complex difficult predict theoretically. Here, we develop a graph neural network (GNN)-based quantitative structure–property (QSPR) model CMC, γcmc, Γmax. Ninety-two data points, encompassing all types surfactants—anionic, cationic, zwitterionic, nonionic—are fed into model, covering temperature range [20–30 °C], which contributes its generalization across types. We show that our models have high accuracy (R2 = 0.87 on average tests) predicting three parameters surfactants. The effectiveness QSPR capturing variation Γmax with molecular design carefully assessed. curated dataset, developed assessment will contribute development improved surfactants facilitate for diverse applications.

Language: Английский

Citations

0