MOFs for Hydrogen Storage - University of Pretoria

This dataset consists of 58 Fe-based MOFs samples created by University of Pretoria is also included in the F.A.I.R. database. As in the case of AC, this dataset is in tabular form, with each row representing one sample, and the columns provide the necessary information, labelled as input or output. Input columns provide the elemental composition of the samples, and the experimental conditions such as treatment temperature and duration. Output columns include porosity estimates (pore area and volume) and H uptake. The first 32 samples are mixtures of up to 2 elements among Al, Ca, Zn, Mg, and Mn, added to the Fe base, all processed at 110 °C for 20 hours. The remaining 26 samples are only made of Fe, using varying treatment temperatures and durations. This methodological split in the dataset was found to lower the quality of machine-learning models significantly, because the dataset does not extensively cover the space of compositions and experimental conditions simultaneously. For this reason, it is also available split into separate subsets.