Assessing collision cross section calculations using MobCal-MPI with a variety of commonly used computational methods

Structural studies with ion mobility require an accurate methodology to bridge theoretical modelling of chemical structure with experimental determination of an ion's collision cross section (CCS). The parallelized MobCal-MPI package enables rapid and accurate evaluation of CCSs that are applicable to several chemical classes, but was only assessed for accuracy using a single model chemistry: B3LYP-D3/6-31++G(d,p). In this work, the performance of MobCal-MPI was validated across 25 different model chemistries, which encompassed PM7, HartreeFock, and three common DFT functionals (B3LYP-D3, omega B97X-D, and M06-2X-D3) using six different basis sets (6-31 G, 6-31 G(d,p), 6-31++G(d,p), def2-SVP, def2-TZVP, and def2-TZVPP). Performance assessment was accomplished using geometries generated from a set of 50 structurally diverse molecules at each level of theory. MobCal-MPI calculates CCSs that correlate well with experimental values for all model chemistries explored (< 2.5% RMSD) with the exception of PM7 (3.0 % RMSD) and methods that employ basis sets lacking polarization functions (e.g., 6-31G; < 4% RMSD). While any of the 25 model chemistries can be used with MobCal-MPI with reasonable accuracy, caution should be exercised when coupling CCS calculations with PM7 or basis sets that lack polarization functions. Following benchmarking, MobCal-MPI was used to calculate the CCS of a macromolecular construct consisting of atropine and beta-cyclodextrin. The CCSs calculated for the beta-cyclodextrin complex using either the PM7 or B3LYP-D3 model chemistries agree with experimental values within the expected error of the method (< 2.5 %).

Automated summary

This study validated the performance of MobCal-MPI across 25 different model chemistries and found that the calculated collision cross sections (CCSs) correlated well with experimental values for almost all model chemistries, with only a few exceptions where the deviation was slightly larger.