London dispersion, which is a part of the famous van der Waals potential, may be an important element of structural stability, and thus affects chemical reactivity and selectivity^[1]. Recently we have shown that attractive, non-covalent interactions in the gas phase are largely, but not entirely, screened out in solution^[2]. Comparing experimental gas-phase bond dissociation energies and dispersion-corrected DFT or local coupled-cluster methods results, we showed that the weakest point in the computational models is the treatment of solvation. As an independent test, we have performed the experiment in the gas phase at temperatures of 4-50 K using an IR action spectrometer with a cryogenic Penning trap that was recently tested in our group (FT-ICR-MS)^[3] In particular, we have measured CIVP-spectra of pyridines, quinolones, and pyridinium dimers. The frequency of the N-H-N asymmetric stretching mode is strongly red-shifted relative to that for an unperturbed N-H stretch in a non-hydrogen-bonded pyridinium. The N-H stretch of a pyridinium cation is not involved in an ionic hydrogen bond is 3370 cm^-1. When it is hydrogen-bonded, our computational studies suggest the frequency drops to around 2200 cm^-1 (Figure 1). If the hydrogen bond is bent, the proton tends to localize and the frequency shifts back towards 3370 cm^-1, depending on how much the N-H-N angle is bent.  If a proton-bound dimer is built with side chains that can interact by dispersion, then the molecule serves as a molecular balance.  The stronger the non-covalent interaction is, the more the proton-bound dimer’s hydrogen bond is bent, and the more the frequency shifts to higher energy.  

Figure 1. Experimental design in which the N-H-N frequency of the ionic hydrogen bond is used as a probe for the strength of the noncovalent interactions between two substituents, whose interaction is balanced against the bending potential of the N-H-N linkage.
Therefore, the proton-bound dimer acts as a molecular balance where the attractive dispersion force is weighed against the bending potential of the N-H-N moiety. Thus the IRPD spectrum becomes a sensitive measure of gas-phase geometry and, in turn, a quantitative measure of the dispersion interactions between the substituents.
The conducted experiment will help to resolve the discrepancy between theory and experiment in the dispersion contribution to bond dissociation energies in large molecules as well as to fill the gap in the estimation of solvent effects in calculations.

[1] Philipp Wagner, Peter Schreiner, Angew. Chem. Int. Ed. Engl. 2015, 54, 12274-12296.
[2] Robert Pollice, Marek Bot, Ilia Kobylianskii, Ilya Shenderovich, Peter Chen, J. Am. Chem. Soc., 2017, 139,13126–13140.
[3] Lukas Fritsche, Andreas Bach, Larisa Miloglyadova, AlexandraTsybizova, Peter Chen, Rev. Sci Inst., 2018, 89, 063119-1−063119-10.