We trained two models for saturation vapor pressure estimation on experimental data. The adGC2NN-broad is a general model with broad scope that is suitable for both organic and inorganic molecules and achieves a mean absolute error (MAE) of 0.67 log-units (R2 = 0.86) on the training data.
The adGC2NN-confined model is specialized on organic compounds with functional groups often encountered in atmospheric SOA, achieving an even stronger correlation with independent test data (MAE = 0.36 log-units, R2 = 0.97).
The models use molecular descriptors like molar mass alongside molecular graphs containing atom and bond features as representations of molecular structure. In adaptive-depth GC2NN, the number of evaluated graph layers depends on molecular size. The most suitable model is automatically selected for each compound.
Note to Users: if you use results from the adGC2NN model, please cite the paper listed below and this website giving the URL (http://www.multiphasekinetics.org/pvap).
Krüger, M., Galeazzo, T., Eremets, I., Schmidt, B., Pöschl, U., Shiraiwa, M., and Berkemeier, T.: Improved vapor pressure predictions using group contribution-assisted graph convolutional neural networks (GC2NN), Geosci. Model Dev. 18, 7357–7371, (2025).