Lanthanide organic frameworks geometry prediction accuracies of quantum chemical calculations

Semiempirical methods emerge as good candidates for solid-state calculation of molecules with hundreds up to a few thousand atoms, as the use of methods considered more accurate such as DFT demands high computational cost even for molecules with a hundred atoms. In this work, we evaluated how accurately different semiempirical quantum chemical models predict the solid-state geometries of Lanthanide Organic Frameworks (LOFs), comparing the performance of Sparkle/AM1, Sparkle/PM3, Sparkle/PM6, Sparkle/PM7, Sparkle/RM1 and the RM1 methods in the calculation of 83 crystal structures. The geometric parameters chosen were bond lengths and angles, and the accuracy of the models was evaluated based on root mean square deviation (RMSD). Results show that Sparkle/PM3 and Sparkle/RM1 were the most accurate models although all of models evaluated had acceptable accuracies. (C) 2019 Published by Elsevier B.V.