News Coatings Technologies
Cheminformatics approach to predict VOC release from coatings
Miles Brockbank and Erik Sapper (Cal Poly) are developing structure-property models that link molecular structure to the volatility and transport of small molecules in coatings systems. The approach combines thermogravimetric analysis with quantitative structure-property relationship (QSAR/QSPR) modelling and aims to predict how volatile additives behave in evolving polymer matrices, supporting the development of lower-emission coatings.
Volatile organic compounds (VOCs) are released from solid or liquid mixtures due to their low vapour pressure and contribute to photochemical ozone formation, smog and respiratory health concerns. Their content in industrial products is therefore strictly regulated. Current standard methods in the United States, such as ASTM D2369 (a weigh-by-difference oven test) and ASTM D6886 (gas chromatography), either lack mechanistic detail or require considerable analytical complexity. Thermogravimetric analysis (TGA), by contrast, allows volatilisation to be monitored continuously as a function of temperature and time, offering a more mechanistic view.
Building on this analytical basis, Miles Brockbank and Erik Sapper at California Polytechnic State University (Cal Poly) have been developing structure-property models that connect molecular structure to volatility and transport behaviour in coatings systems. TGA data collected under isothermal conditions at 110 °C are analysed with custom Python scripts to determine the slope of volatilisation and the time required to reach 0 % mass. Molecular geometries are optimised in Avogadro using a Universal Force Field, exported as .mol2 files and processed with Dragon7 to generate thousands of molecular descriptors per compound. Correlation analysis and initial QSAR/QSPR modelling using train-test splits have yielded promising predictive performance for approximately 100 individual VOCs.
Event tip:
The EC Conference Digitalisation in Coatings Formulation is the ideal platform to learn about the latest advances and trends in the field of digitally assisted coatings production processes. Join experts, researchers and industry leaders as they share their insights and expertise in automation, big data, AI, and predictive modelling to enhance efficiency, accuracy, and speed in formulation, testing, and quality control.
From pure compounds to complex polymer matrices
The current model has clear limitations: it was developed for a relatively small number of individual VOCs rather than for additives embedded in complex, interacting coating mixtures, and its performance is sensitive to the choice of training data. The proposed next step is to expand the diversity of compounds analysed, refine descriptor generation using new force fields, and identify which molecular descriptors most reliably predict volatility parameters across a chemically diverse dataset. A central question is which structural features consistently govern volatilisation behaviour under standardised TGA conditions.
A further, more ambitious aim of Brockbank and Sapper is to extend descriptor-based transport modelling from pure compounds to mixtures containing both polymer resins and volatile additives. This raises the challenge of predicting volatility in dynamic polymer matrices that undergo drying, latex formation, shrinking, swelling or curing. By linking molecular-scale structure to time-dependent transport in real formulations, the research could reduce reliance on trial-and-error formulation, support the development of environmentally safer coatings and provide a framework potentially transferable to other fields, including controlled release in medical applications and functional materials.
Source: Brockbank, M. & Sapper, E., Structure-Property Models for Small Molecule Volatility and Transport in Evolving Polymer Matrices. California Polytechnic State University, Sapper Group (2024).
