Modeling chemical releases from building materials: The search for extended validity domain and parsimony
Résumé
This study aims to systematically evaluate and extend the validity domains of multiple mechanistic models of chemical emissions from building materials. We compare the validity domain of the analytical solution with four numerical solutions for a single layer material with one convective surface and a wide range of chemical properties, material thicknesses, and simulation time. We also develop a parsimonious simplified model, ensuring the widest possible validity domain with minimum simulation time. For diffusion coefficients lower than 10−15 m2/s, accuracy of the analytical solution requires at least 5000 positive roots. The numerical method using uneven discretization and finite difference approximation for the boundary conditions is the best numerical solution. The parsimonious combined D- and K-limited model achieves similar accuracy as the best numerical solution except slight overestimates at the interface between the D- and K-limited zones, while having simpler computations and much shorter simulation time. These models show good agreement against experimental data. This study demonstrates that the complex analytical solution can be well approximated by a simpler model with a wide validity domain, enabling the high-throughput screenings of a large number of chemical-product combinations.