Assessment of theoretical methods for the determination of the mechanochemical strength of covalent bonds

Abstract

The performance of some commonly used quantum-chemical methods in accurately and reliably describing the influence of applying an external mechanical force has been investigated for a set of small molecules. By applying coupled-cluster CCSD(T) theory in an extended basis set as benchmark, all methods tested provide a good qualitative description of the physical process, although the quantitative agreement varies considerably. Hartree–Fock (HF) theory overestimates both the values of the bond-breaking point and the rupture force, typically by 20–30%. The same applies to density-functional theory (DFT) based on the local density approximation (LDA). By introducing the generalized gradient approximation (GGA) in the form of the BLYP and PBE functionals, only a slight overestimation is observed. Moreover, these pure DFT methods perform better than the hybrid B3LYP and CAM-B3LYP methods. The excellent agreement observed between the CCSD(T) method and multiconfigurational methods for bond distances significantly beyond the bond-breaking point shows that the essence of mechanical bond breaking is captured by single-reference-based methods. Comparisons of accurate numerical bond-dissociation curves with simple analytical forms show that Morse-type curves provide useful approximate bond-breaking points and rupture forces, accurate to within 10%. By contrast, polynomial curves are much less useful. The outcome of kinetic calculations to estimate the dissociation probability as a function of the applied force depends strongly on the description of the potential-energy curve. The most probable rupture forces calculated by numerical integration appear to be significantly more accurate than those obtained from simple analytical expressions based on fitted Morse potentials.

Keywords:

• mechanochemistry
• atomic force microscopy
• bond-breaking point
• rupture force
• Morse potential

Acknowledgements

This work has received support from the Norwegian Research Council through a Centre of Excellence Grant (grant No. 179568/V30), as well as through a grant of computer time from the Norwegian Supercomputing Program.