{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,9,30]],"date-time":"2026-09-30T13:56:11Z","timestamp":1790776571904,"version":"4.1.0"},"reference-count":31,"publisher":"MDPI AG","issue":"19","license":[{"start":{"date-parts":[[2026,9,30]],"date-time":"2026-09-30T00:00:00Z","timestamp":1790726400000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["Int. J. Mol. Sci.","IJMS"],"abstract":"<jats:p>This fully synthetic study tests a structure-preserving surrogate on NRTL-generated phase equilibrium systems; it contains no experimental molecular validation. A symmetric neural potential represents GE\/(nRT), from which activity coefficients, thermal derivatives, binodals, flashes, and tangent-plane-distance diagnostics are derived. Across ten binary training seeds, its median ln\u03b3 MAE was 1.16\u00d710\u22123, better than matched direct regression (2.40\u00d710\u22123) but not Redlich\u2013Kister (9.65\u00d710\u22124) or correctly specified refitted NRTL (2.36\u00d710\u22124). Gibbs\u2013Duhem residuals decayed as O(h2) under grid refinement, confirming diagnostic truncation around an analytic identity. A 41,600-point binary hull gave endpoint MAE 2.31\u00d710\u22124. On 100 ternary type-I feeds, phase-count agreement with numerical reference labels was 98.0% (exact 95% interval, 93.0\u201399.8%), with two false negatives after recovery of one infeasible-floor solver failure. A 20-member ensemble showed pooled error ranking (\u03c1=0.894) but weak in-domain ranking (\u03c1=0.164); a temperature\u2013distance heuristic was stronger (\u03c1=0.914), and nominal 90% conformal coverage fell from 88.0% in-domain to 0.0% under shift. Controlled flashes were 2.77\u00d7 slower than NRTL. A hard thermodynamic structure is therefore valuable within neural modeling, but classical dominance, shift sensitivity, and absent molecular data bound the claim.<\/jats:p>","DOI":"10.3390\/ijms27198767","type":"journal-article","created":{"date-parts":[[2026,9,30]],"date-time":"2026-09-30T13:01:19Z","timestamp":1790773279000},"page":"8767","update-policy":"https:\/\/doi.org\/10.3390\/mdpi_crossmark_policy","source":"Crossref","is-referenced-by-count":0,"title":["Structure-Preserving Excess Gibbs Learning for Multicomponent Phase Equilibria: A Falsifiable and Uncertainty-Aware Synthetic Study"],"prefix":"10.3390","volume":"27","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-8312-7554","authenticated-orcid":false,"given":"Pedro","family":"Robles","sequence":"first","affiliation":[{"name":"Escuela de Ingenier\u00eda Qu\u00edmica, Pontificia Universidad Cat\u00f3lica de Valpara\u00edso, Valpara\u00edso 2340000, Chile"}],"role":[{"vocabulary":"crossref","role":"author"},{"role":"conceptualization","vocabulary":"credit"},{"role":"methodology","vocabulary":"credit"},{"role":"validation","vocabulary":"credit"},{"role":"investigation","vocabulary":"credit"},{"role":"resources","vocabulary":"credit"},{"role":"project-administration","vocabulary":"credit"},{"role":"funding-acquisition","vocabulary":"credit"},{"role":"writing-review-editing","vocabulary":"credit"}]},{"ORCID":"https:\/\/orcid.org\/0009-0007-3408-9233","authenticated-orcid":false,"given":"Luis","family":"Rios-Colque","sequence":"additional","affiliation":[{"name":"Facultad de Ingenier\u00eda, Pontificia Universidad Cat\u00f3lica de Valpara\u00edso, Valpara\u00edso 2340000, Chile"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-5766-6004","authenticated-orcid":false,"given":"Vanesa","family":"Bazan","sequence":"additional","affiliation":[{"name":"Instituto de Investigaciones Mineras (IIM), Facultad de Ingenier\u00eda, Universidad Nacional de San Juan, San Juan 5400, Argentina"},{"name":"Consejo Nacional de Investigaciones Cient\u00edficas y T\u00e9cnicas (CONICET), San Juan 5400, Argentina"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0005-8935-7595","authenticated-orcid":false,"given":"Luis","family":"Rojas-Valdivia","sequence":"additional","affiliation":[{"name":"Facultad de Ingenier\u00eda, Pontificia Universidad Cat\u00f3lica de Valpara\u00edso, Valpara\u00edso 2340000, Chile"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1968","published-online":{"date-parts":[[2026,9,30]]},"reference":[{"key":"ref_1","doi-asserted-by":"crossref","first-page":"345","DOI":"10.1021\/ie50458a036","article-title":"Algebraic Representation of Thermodynamic Properties and the Classification of Solutions","volume":"40","author":"Redlich","year":"1948","journal-title":"Ind. Eng. Chem."},{"key":"ref_2","doi-asserted-by":"crossref","first-page":"135","DOI":"10.1002\/aic.690140124","article-title":"Local Compositions in Thermodynamic Excess Functions for Liquid Mixtures","volume":"14","author":"Renon","year":"1968","journal-title":"AIChE J."},{"key":"ref_3","doi-asserted-by":"crossref","first-page":"116","DOI":"10.1002\/aic.690210115","article-title":"Statistical Thermodynamics of Liquid Mixtures: A New Expression for the Excess Gibbs Energy of Partly or Completely Miscible Systems","volume":"21","author":"Abrams","year":"1975","journal-title":"AIChE J."},{"key":"ref_4","doi-asserted-by":"crossref","first-page":"1086","DOI":"10.1002\/aic.690210607","article-title":"Group-Contribution Estimation of Activity Coefficients in Nonideal Liquid Mixtures","volume":"21","author":"Fredenslund","year":"1975","journal-title":"AIChE J."},{"key":"ref_5","doi-asserted-by":"crossref","first-page":"899","DOI":"10.1021\/ie001047w","article-title":"A Priori Phase Equilibrium Prediction from a Segment Contribution Solvation Model","volume":"41","author":"Lin","year":"2002","journal-title":"Ind. Eng. Chem. Res."},{"key":"ref_6","doi-asserted-by":"crossref","first-page":"1244","DOI":"10.1021\/ie0003887","article-title":"Perturbed-Chain SAFT: An Equation of State Based on a Perturbation Theory for Chain Molecules","volume":"40","author":"Gross","year":"2001","journal-title":"Ind. Eng. Chem. Res."},{"key":"ref_7","doi-asserted-by":"crossref","first-page":"9","DOI":"10.1016\/j.fluid.2010.03.030","article-title":"Modeling of High-Pressure Vapor\u2013Liquid Equilibrium in Ionic Liquids + Gas Systems Using the PRSV Equation of State","volume":"295","author":"Arce","year":"2010","journal-title":"Fluid Phase Equilibria"},{"key":"ref_8","doi-asserted-by":"crossref","first-page":"119","DOI":"10.1016\/j.fluid.2016.08.031","article-title":"Modeling of Liquid\u2013Liquid Equilibrium for Binary and Ternary Systems Containing Ionic Liquids with the Hexafluorophosphate Anion Using the ASOG Method","volume":"429","author":"Robles","year":"2016","journal-title":"Fluid Phase Equilibria"},{"key":"ref_9","doi-asserted-by":"crossref","first-page":"981","DOI":"10.1021\/acs.jpclett.9b03657","article-title":"Machine Learning in Thermodynamics: Prediction of Activity Coefficients by Matrix Completion","volume":"11","author":"Jirasek","year":"2020","journal-title":"J. Phys. Chem. Lett."},{"key":"ref_10","doi-asserted-by":"crossref","first-page":"4854","DOI":"10.1039\/D1SC07210B","article-title":"Making Thermodynamic Models of Mixtures Predictive by Machine Learning: Matrix Completion of Pair Interactions","volume":"13","author":"Jirasek","year":"2022","journal-title":"Chem. Sci."},{"key":"ref_11","doi-asserted-by":"crossref","first-page":"1054","DOI":"10.1039\/D2CP04478A","article-title":"Prediction of Parameters of Group Contribution Models of Mixtures by Matrix Completion","volume":"25","author":"Jirasek","year":"2023","journal-title":"Phys. Chem. Chem. Phys."},{"key":"ref_12","doi-asserted-by":"crossref","first-page":"3141","DOI":"10.1021\/acs.jpca.4c08360","article-title":"Similarity-Informed Matrix Completion Method for Predicting Activity Coefficients","volume":"129","author":"Hayer","year":"2025","journal-title":"J. Phys. Chem. A"},{"key":"ref_13","doi-asserted-by":"crossref","first-page":"859","DOI":"10.1039\/D2DD00058J","article-title":"A Smile Is All You Need: Predicting Limiting Activity Coefficients from SMILES with Natural Language Processing","volume":"1","author":"Winter","year":"2022","journal-title":"Digit. Discov."},{"key":"ref_14","doi-asserted-by":"crossref","first-page":"138","DOI":"10.1039\/D2DD00045H","article-title":"Capturing Molecular Interactions in Graph Neural Networks: A Case Study in Multi-Component Phase Equilibrium","volume":"2","author":"Qin","year":"2023","journal-title":"Digit. Discov."},{"key":"ref_15","doi-asserted-by":"crossref","first-page":"108153","DOI":"10.1016\/j.compchemeng.2023.108153","article-title":"Graph Neural Networks for Temperature-Dependent Activity Coefficient Prediction of Solutes in Ionic Liquids","volume":"171","author":"Rittig","year":"2023","journal-title":"Comput. Chem. Eng."},{"key":"ref_16","doi-asserted-by":"crossref","first-page":"1752","DOI":"10.1039\/D3DD00103B","article-title":"Gibbs\u2013Duhem-Informed Neural Networks for Binary Activity Coefficient Prediction","volume":"2","author":"Rittig","year":"2023","journal-title":"Digit. Discov."},{"key":"ref_17","doi-asserted-by":"crossref","first-page":"18504","DOI":"10.1039\/D4SC04554H","article-title":"Thermodynamics-Consistent Graph Neural Networks","volume":"15","author":"Rittig","year":"2024","journal-title":"Chem. Sci."},{"key":"ref_18","doi-asserted-by":"crossref","first-page":"19777","DOI":"10.1039\/D4SC05115G","article-title":"HANNA: Hard-Constraint Neural Network for Consistent Activity Coefficient Prediction","volume":"15","author":"Specht","year":"2024","journal-title":"Chem. Sci."},{"key":"ref_19","doi-asserted-by":"crossref","first-page":"045301","DOI":"10.1103\/PhysRevE.105.045301","article-title":"Machine Learning of Consistent Thermodynamic Models Using Automatic Differentiation","volume":"105","author":"Rosenberger","year":"2022","journal-title":"Phys. Rev. E"},{"key":"ref_20","doi-asserted-by":"crossref","first-page":"406","DOI":"10.1038\/s42005-024-01892-3","article-title":"Development of a Helmholtz Free Energy Equation of State for Fluid and Solid Phases via Artificial Neural Networks","volume":"7","author":"Chaparro","year":"2024","journal-title":"Commun. Phys."},{"key":"ref_21","doi-asserted-by":"crossref","first-page":"422","DOI":"10.1038\/s42254-021-00314-5","article-title":"Physics-Informed Machine Learning","volume":"3","author":"Karniadakis","year":"2021","journal-title":"Nat. Rev. Phys."},{"key":"ref_22","doi-asserted-by":"crossref","first-page":"2491","DOI":"10.1039\/D5DD00489F","article-title":"Physics-Informed Machine Learning for Predicting Temperature-Dependent Chemical Properties","volume":"5","author":"Rezaei","year":"2026","journal-title":"Digit. Discov."},{"key":"ref_23","doi-asserted-by":"crossref","first-page":"89","DOI":"10.1039\/D4ME00168K","article-title":"Accelerating Multicomponent Phase-Coexistence Calculations with Physics-Informed Neural Networks","volume":"10","author":"Dhamankar","year":"2025","journal-title":"Mol. Syst. Des. Eng."},{"key":"ref_24","doi-asserted-by":"crossref","first-page":"10487","DOI":"10.1021\/acs.jcim.6c00728","article-title":"Differentiable Thermodynamic Phase-Equilibria for Machine Learning","volume":"66","author":"Ascani","year":"2026","journal-title":"J. Chem. Inf. Model."},{"key":"ref_25","doi-asserted-by":"crossref","first-page":"1798","DOI":"10.1002\/aic.690410715","article-title":"Global Optimization for the Phase Stability Problem","volume":"41","author":"McDonald","year":"1995","journal-title":"AIChE J."},{"key":"ref_26","doi-asserted-by":"crossref","first-page":"1159","DOI":"10.1016\/S0098-1354(98)00218-X","article-title":"A Global Optimization Method, \u03b1BB, for General Twice-Differentiable Constrained NLPs\u2014II. Implementation and Computational Results","volume":"22","author":"Adjiman","year":"1998","journal-title":"Comput. Chem. Eng."},{"key":"ref_27","doi-asserted-by":"crossref","first-page":"494","DOI":"10.1561\/2200000101","article-title":"Conformal Prediction: A Gentle Introduction","volume":"16","author":"Angelopoulos","year":"2023","journal-title":"Found. Trends Mach. Learn."},{"key":"ref_28","doi-asserted-by":"crossref","first-page":"5510","DOI":"10.1021\/ie010954d","article-title":"Application of the Perturbed-Chain SAFT Equation of State to Associating Systems","volume":"41","author":"Gross","year":"2002","journal-title":"Ind. Eng. Chem. Res."},{"key":"ref_29","doi-asserted-by":"crossref","first-page":"268","DOI":"10.1021\/j100621a026","article-title":"Thermodynamics of Electrolytes. I. Theoretical Basis and General Equations","volume":"77","author":"Pitzer","year":"1973","journal-title":"J. Phys. Chem."},{"key":"ref_30","doi-asserted-by":"crossref","unstructured":"Rasmussen, C.E., and Williams, C.K.I. (2006). Gaussian Processes for Machine Learning, MIT Press.","DOI":"10.7551\/mitpress\/3206.001.0001"},{"key":"ref_31","doi-asserted-by":"crossref","first-page":"404","DOI":"10.1093\/biomet\/26.4.404","article-title":"The Use of Confidence or Fiducial Limits Illustrated in the Case of the Binomial","volume":"26","author":"Clopper","year":"1934","journal-title":"Biometrika"}],"container-title":["International Journal of Molecular Sciences"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.mdpi.com\/1422-0067\/27\/19\/8767\/pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,9,30]],"date-time":"2026-09-30T13:03:55Z","timestamp":1790773435000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.mdpi.com\/1422-0067\/27\/19\/8767"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,9,30]]},"references-count":31,"journal-issue":{"issue":"19","published-online":{"date-parts":[[2026,10]]}},"alternative-id":["ijms27198767"],"URL":"https:\/\/doi.org\/10.3390\/ijms27198767","relation":{},"ISSN":["1422-0067"],"issn-type":[{"value":"1422-0067","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,9,30]]}}}