Hydrogen bonding dynamics in liquids

Abstract

Liquids may appear disordered compared to the structured nature of crystals, but this is only true when considering long-range order. Unlike ideal gases, most liquids exhibit local order due to molecular interactions, ranging from dipole-dipole to electrostatic interactions. For example, water forms an extensive hydrogen bond network or molecular liquids such as alcohols can form aggregates. Understanding intermolecular forces, especially hydrogen bonding, is crucial for insights into phenomena like water’s density anomalies. This thesis addresses three topics within the framework of hydrogen bonding in liquids, with a particular focus on the dynamics of these bonds. We employ linear and two-dimensional infrared spectroscopy (2DIR) alongside ab initio molecular dynamics (AIMD) simulations to explore bonding in liquids.

In the first project, we employ linear and two dimensional infrared (2DIR) spectroscopy to investigate the correlation of the donating hydrogen bonds of water. Water has previously been proposed to form asymmetric hydrogen bonds based on X-ray absorption studies. This supposed asymmetry has been controversially discussed. Our results provide experimental evidence for dynamic anti correlations instead of static asymmetric hydrogen bonds. We use 2DIR spectroscopy to isolate the inhomogeneous contributions to the lineshape for pure and isotopically dilute (HOD in H2O) D2O in dimethylformamide (DMF). Through the dilution in DMF each water molecule only forms two donating hydrogen bonds to the solvent. This allows us to distinguish between symmetric and asymmetric stretching modes for D2O. Comparison with density functional theory calculations (DFT) show, that the about twice broader inhomogeneous lineshape of HOD can be explained by an anti correlated distribution of hydrogen bonds. Furthermore analysis of the crosspeaks for D2O give direct experimental evidence for the anti correlation. We find that this anti correlation quickly decays (< 500 fs). Furthermore similar experiments on urea also show this anti correlation, albeit less pronounced. This confirms that these short lived anti correlations are not special for water but rather expected for all XH2 containing molecules.[1]

In the second project we investigate the solvent properties of the fluorinated mono alcohol hexafluoroisopropanol (HFIP). HFIP is a solvent that is widely used in synthetic chemistry as well as chemical biology. It has been shown to open up novel reaction pathways and to increase the reaction speed of several reactions. This increase in reaction speed has been coined a "booster effect". We investigate the dynamic origins of the booster effect of HFIP, via vibrational and dielectric spectroscopy of HFIP and its non fluorinated counterpart isopropanol (IP). With polarization controlled time resolved IR spectroscopy (TRIR) we found that individual HFIP molecules show a slower reorientation than IP, even though linear IR spectroscopy suggests weaker hydrogen bonds between HFIP molecules compared to IP. 2DIR spectroscopy shows slower hydrogen bonding dynamics for HFIP. Using dielectric relaxation spectroscopy (DRS), we find faster collective reorientation of HFIP, suggesting a smaller average cluster size. Titration experiments with a hydrogen bond accepting substrate (diethylether) reveal that HFIP forms much stronger hydrogen bonds with the substrate than IP. Together our findings suggest that HFIP forms smaller clusters than IP, and thus has more terminal OH groups as potential hydrogen bond donors. The slower hydrogen bond dynamics and reorientation of HFIP can stabilize the hydrogen bonded clusters, and thus make solvent substrate interactions more likely.[2]

The last projects is about specific ion effects on peptides. The strucure of peptides depends on the solvation environment. Ions can affect peptides directly through electrostatic interactions as well as indirectly through disruption of the solvent environment. We investigate the effects of Ca2+ on the dipeptide L-alanyl-L-alanine (2Ala). Calcium is located on the protein denaturation side of the Hofmeister series. We in particular focus on the competition between the two potential binding sites with the carboxylate group and the amide carbonyl group. With linear and 2DIR spectroscopy we find a blueshifted shoulder for the carboxylate peak as well as a small redshift for the amide carbonyl. Experiments with 13C labeled 2ALa reveal that the shoulder is connected with Ca2+ interaction at the carboxylate group. Ab initio MD simulations show that Ca2+ binds to the carboxylate and to a lesser degree at the amide CO. This is consistent with an increasing lifetime of the amide CO peak found from 2DIR spectroscopy with increasing salt concentration. At the carboxylate we only find Ca2+ only monodentate bonded. Vibrational density of states (VDOS) show that this is consistent with a spectral blueshift. Additionally we find configurations of two Ca2+ bound mono dentally to the two carboxylate oxygens. This configuration leads to an additional blueshift. Our results highlight the importance of taking competition between different binding sites into account, when designing model systems for peptide interactions.[3]