The local structures of layered covalent-organic frameworks (COFs) deviate from the average crystal structures assigned from X-ray diffraction experiments. For two prototype COFs of Tp-Azo and DAAQ-TFP, density functional theory calculations have shown that the eclipsed structure is not an energy minimum and that the internal energy is lowered for an inclined stacking arrangement. Here we explore the structural disorder of these frameworks at 300 K through molecular dynamics (MD) simulations using an on-the-fly machine learning force field (MLFF). We find that an initially eclipsed stacking mode spontaneously distorts to form a zigzag configuration that lowers the free energy of the crystal. The simulated diffraction patterns show good agreement with experimental observations. The dynamic disorder from the MLFF MD trajectories is found to persist in mesoscale MD simulations of 155 thousand atoms, giving further confidence in our conclusions. Our simulations show that the stacking behaviour of layered COFs is more complicated than previously understood.
Bibliographical noteFunding Information:
We thank Matthias Golomb for useful discussions on porous frameworks. J. H. acknowledges Imperial College London and the Chinese Scholarship Council (CSC) for providing a PhD scholarship. S.-J. S. is funded by the National Research Foundation of Korea (NRF-2018M3D1A1058793). K. T. is funded by the Independent Research Fund Denmark through the International Postdoctoral grant (0164-00015B). G. K. acknowledges the Faraday Institution for funding a PhD studentship (faraday.ac.uk; EP/S003053/1), grant number FIRG025. A. M. G. was supported by EPSRC Fellowship EP/T033231/1. We are also grateful to the UK Materials and Molecular Modelling Hub for computational resources, which is partially funded by EPSRC (EP/P020194/1 and EP/T022213/1). Via our membership of the UK's HEC Materials Chemistry Consortium, which is funded by EPSRC (EP/L000202), this work used the ARCHER2 UK National Supercomputing Service ( https://www.archer2.ac.uk ). Also, this work was supported by the Korea National Supercomputing Center with supercomputing resources including technical support (KSC-2022-CRE-0245).
© 2023 The Royal Society of Chemistry.