Can Machine Learning Predict the Phase Behavior of Surfactants?

Thacker, Joseph C. R.; Bray, David J.; Warren, Patrick B.; Anderson, R. L.

doi:10.1021/acs.jpcb.2c08232

Cited by 7 publications

(3 citation statements)

References 57 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…(Some closely related, but distinct problems in machine-learning for aqueous phase thermodynamics include predicting partition coefficient of biomolecules in the two different phases of an ATPS [23][24][25] and predicting the phase behavior of one surfactant in water as a function of temperature and surfactant concentration. 26 )…”

Section: Toc Graphicmentioning

confidence: 99%

Data-driven imputation of miscibility of aqueous solutions via graph-regularized logistic matrix factorization

Behnoudfar

Simon

Schrier

2023

Preprint

View full text Add to dashboard Cite

Aqueous, two-phase systems (ATPSs) may form upon mixing two solutions of independently water-soluble compounds. Many separation, purification, and extraction processes rely on ATPSs. Predicting the miscibility of solutions can accelerate and reduce the cost of the discovery of new ATPSs for these applications. Whereas previous machine learning approaches to ATPS prediction used physicochemical properties of each solute as a descriptor, in this work, we show how we can impute missing miscibility outcomes directly from an incomplete collection of pairwise miscibility experiments. We use graph-regularized logistic matrix factorization to learn a latent vector of each solution from (i) the observed entries in the pairwise miscibility matrix and (ii) a graph (nodes: solutes, edges: shared relationships) indicating the general category of the solute (i.e., polymer, surfactant, salt, protein). Using an experimental dataset of the pairwise miscibility of 68 solutions from Peacock et al. [ACS Appl. Mater. Interfaces 2021, 13, 9], we show that graph-regularized logistic matrix factorization more accurately predicts missing (im)miscibility outcomes of pairs of solutions than ordinary logistic matrix factorization and random forest classifiers using physicochemical features of the compounds.

show abstract

Section: Toc Graphicmentioning

confidence: 99%

Data-driven imputation of miscibility of aqueous solutions via graph-regularized logistic matrix factorization

Behnoudfar

Simon

Schrier

2023

Preprint

View full text Add to dashboard Cite

show abstract

“…The random forest achieved an ∼74% accuracy. (Some closely related but distinct problems in machine learning for aqueous phase thermodynamics include predicting the partition coefficient of biomolecules in the two different phases of an ATPS − and predicting the phase behavior of one surfactant in water as a function of temperature and surfactant concentration …”

Section: Introductionmentioning

confidence: 99%

“…(Some closely related but distinct problems in machine learning for aqueous phase thermodynamics include predicting the partition coefficient of biomolecules in the two different phases of an ATPS 23−25 and predicting the phase behavior of one surfactant in water as a function of temperature and surfactant concentration. 26 ) Matrix Factorization. The pairwise structure of ATPSs suggests the possibility of imputing missing or unobserved mixture miscibilities using matrix factorization.…”

Section: ■ Introductionmentioning

confidence: 99%

Data-Driven Imputation of Miscibility of Aqueous Solutions via Graph-Regularized Logistic Matrix Factorization

Behnoudfar,

Simon,

Schrier

2023

J. Phys. Chem. B

View full text Add to dashboard Cite

Aqueous, two-phase systems (ATPSs) may form upon mixing two solutions of independently water-soluble compounds. Many separation, purification, and extraction processes rely on ATPSs. Predicting the miscibility of solutions can accelerate and reduce the cost of the discovery of new ATPSs for these applications. Whereas previous machine learning approaches to ATPS prediction used physicochemical properties of each solute as a descriptor, in this work, we show how to impute missing miscibility outcomes directly from an incomplete collection of pairwise miscibility experiments. We use graph-regularized logistic matrix factorization (GR-LMF) to learn a latent vector of each solution from (i) the observed entries in the pairwise miscibility matrix and (ii) a graph where each node is a solution and edges are relationships indicating the general category of the solute (i.e., polymer, surfactant, salt, protein). For an experimental data set of the pairwise miscibility of 68 solutions from Peacock et al. [ACS Appl. Mater. Interfaces 2021, 13, 11449–11460], we find that GR-LMF more accurately predicts missing (im)miscibility outcomes of pairs of solutions than ordinary logistic matrix factorization and random forest classifiers that use physicochemical features of the solutes. GR-LMF obviates the need for features of the solutions and solutions to impute missing miscibility outcomes, but it cannot predict the miscibility of a new solution without some observations of its miscibility with other solutions in the training data set.

show abstract

Surfactants produced from carbohydrate derivatives: Part 2. A review on the value chain, synthesis, and the potential role of artificial intelligence within the biorefinery concept

Marquez,

Ortiz,

Barrios

et al. 2024

J Surfact & Detergents

View full text Add to dashboard Cite

This comprehensive and critical review explores the synthesis and applications of carbohydrate‐based surfactants within the biorefinery concept, focusing on biobased sugar‐head molecules suitable for use across several manufacturing sectors, including cosmetics, pharmaceuticals, household products, detergents, and foods. The main focus relies on sustainable alternatives to conventional surfactants, which could reduce the final manufacturing carbon footprint of several industrial feedstocks and products. A thorough analysis of raw materials, highlighting the significance of feedstock sources, and the current biobased surfactants and rhamnolipid biosurfactants production trends, is presented. Key organic reactions for the production of sorbitan esters, sucrose esters, alkyl polyglycosides, and fatty acid glucamines, such as glycosidation, acylation, and etherification, as well as the production of rhamnolipids through fermentation are described. Given the scarce literature on the characterization of these surfactant types within the hydrophilic–lipophilic deviation (HLD) framework, the surfactant contribution parameter (SCP) in the HLD equation for sugar‐head surfactants is critically assessed. The economic landscape is also discussed, noting the significant growth in the biobased surfactants and biosurfactant market, driven by environmental awareness and regulatory changes, with projections indicating a substantial market increase in the forthcoming years. Finally, the promising potential of generative artificial intelligence (AI) in developing customized surfactant molecules, with optimized properties for targeted applications, is emphasized as a promising avenue for future research.

show abstract

Can Machine Learning Predict the Phase Behavior of Surfactants?

Cited by 7 publications

References 57 publications

Data-driven imputation of miscibility of aqueous solutions via graph-regularized logistic matrix factorization

Data-driven imputation of miscibility of aqueous solutions via graph-regularized logistic matrix factorization

Data-Driven Imputation of Miscibility of Aqueous Solutions via Graph-Regularized Logistic Matrix Factorization

Surfactants produced from carbohydrate derivatives: Part 2. A review on the value chain, synthesis, and the potential role of artificial intelligence within the biorefinery concept

Contact Info

Product

Resources

About