Hydrogen bond (HB) is by its importance a unique chemical phenomenon in nature and has been widely studied from almost any possible viewpoint. HBs have an enormous role in the structure (proteins, nucleic acids, cellulose fibers, etc) and functionality (enzyme catalysis, ligand-receptor complexes, etc) of all forms of life as well as a countless number of supramolecular systems. Computational prediction of HB is of high interest both for rationalizing existing and developing new chemical and biochemical systems and processes. A comprehensive study of modeling HB with the COSMO-RS computational method has been recently carried out at UT Chair of analytical chemistry (ChemPhysChem, 2013, 14). The method displayed mixed behavior, being quite successful with some systems but failing with others.

Q: What is the connection here with measurements?
A: A direct one. The level of success of the computations was assessed by comparison with measurement results of hydrogen bond formation equilibrium constants. And, surprisingly, it turned out that a large part of data on HB formation constants in the literature was unusable because of inconsistrencies and lack of any information characterizing the uncertainty of the values. The values from different groups sometimes varied by up to an order of magnitude (!). Such doubtful datasets of course were left aside when carrying out the evaluation.

This work nicely proves the importance of good measurement science also for theoretical chemistry: it is not possible to develop and improve theoretical computation methods if teh obtained results cannot be compared to accurate measurement results.

_END_