Predictive Modeling of Liquid Density and Surface Tension for Sustainable Aviation Fuels Using Nuclear Magnetic Resonance Atom Types

Prescreening of sustainable aviation fuels (SAFs) is
crucial for
early stage development and ASTM D4054 evaluation. This study develops
models to predict two key properties: temperature-dependent liquid
density and surface tension of complex hydrocarbon mixtures. 1H 13C heteronuclear single quantum coherence nuclear
magnetic resonance spectroscopy is used to determine atom type compositions.
Multiple linear regression models, trained on 1241 liquid density
and 1260 surface tension experimental data points, identified seven
key atom types and a temperature-dependent term as predictors. Applied
to fossil-derived and synthetic fuels, density predictions had an
error range of 0.00–5.35%, and surface tension predictions
ranged from 0.29–4.41%. The prescreening method proved to be
effective for predicting critical fuel properties in early stage SAF
development.