hide
Free keywords:
-
Abstract:
The synergy of experimental and theoretical investigations is highly powerful and is therefore utilized to gain a deeper understanding of (bio)chemical processes. Especially, molecular spectroscopy provides detailed information and thus, serves as a perfect meeting ground for the experimental and theoretical world. Therefore, quantum theoretical methods have to be refined towards more accurate computations of spectroscopic properties.
In this work, both the development of new quantum methods to calculate molecular properties of open shell species and a joint venture of sophisticated experiments with high level theory to investigate solvation effects on electron paramagnetic resonance (EPR) properties are presented.
EPR is a powerful spectroscopic method to study open shell molecules, with the electronic g-tensor being one of its key parameters. We developed an efficient implementation to compute the g-tensor at the level of second-order Møller-Plesset perturbation theory (MP2) using the resolution of identity (RI) approximation for an efficient treatment of the two-electron integrals. To circumvent the gauge problem, which is present for all magnetic property calculations, gauge including atomic orbitals (GIAOs) are applied. The implementation additionally enables calculations at double-hybrid density functional theory (DHDFT) level, of which we have tested the B2PLYP and DSD-PBEP86 functional next to pure RI-MP2. The computed g-values were compared to experimental values and published data from other methods, including coupled cluster singles doubles (CCSD). The results show a clear improvement of DHDFT upon RI-MP2 and good agreement with experimental values, however still being outperformed by the hybrid functional B3LYP for the tested set of small radicals. Evaluation of the computational performance for medium and large size radicals revealed that the RIJCOSX approximation for two-electron integrals distinctly reduces the time for large molecules which consist of more than 100 electrons.
In the joint study, we investigated a nitroxide spin label, HMI, in aqueous solution. Here, we computed the g-tensor and hyperfine couplings (HFCs) for a large set of “HMI in water” configurations (snapshots) to mimic the experimental sample. We conducted an elaborate calibration study to set-up a quantum mechanical/molecular mechanical (QM/MM) model. This facilitated calculations for many thousand snapshots. The established scheme includes HMI and all water molecules up to the second solvation shell in the QM regime, whereas the remaining waters were treated as point charges. Applying DLPNO-CCSD as the most efficient and accurate method for HFC calculations today, provided very good agreement with experiment. Furthermore, we investigated the g-strain effect, which is associated with the distribution of g-values within the measured sample, from a purely theoretical perspective for the first time. For this, we simulated “theoretical spectra” based on the set of computed EPR parameters at different spectrometer frequencies and compared those directly to the experimental EPR spectra. These investigations revealed that the g-strain mainly originates from the conformational flexibility of the molecule itself and is barely influenced by the explicit number of hydrogen bonds formed around the nitroxy group, but rather by solvation as such.
