Modern Hydrogen Bonding Theory

The H-bond was discovered in 1920 by W.M. Latimer and W.H. Rodebush [1] with the collaboration of M.L Huggins [2], three young men working in the laboratory of G.N. Lewis who gave of it a definition based on the Lewis electron-dot formalism which appears to be quite lucid and accurate even in modern terms. By the time that L. Pauling wrote his famous book “The Nature of the Chemical Bond” (1939-1940) [3], the H-bond had received complete systematization within the scheme of the newly developing VB theory, including the distinction between weak electrostatic and strong covalent H-bonds which was successively given VB theoretical dignity by Coulson and Danielsson (1954) [4]. This line of thought was accepted during the 1957 Ljubljana Conference [5] (the first H-bond meeting) and in “The Hydrogen Bond” by Pimentel and McClellan (1960) [6] (the first H-bond book). This unified approach did not survive the division of sciences in more specialized branches occurred in the post-war period. The accumulation of ever new thermodynamic, spectroscopic and structural data, together with the underlying battle between VB and MO methods, lead to a period of general confusion, summarized in the Hopfinger’s (1973) statement “The only one definite fact about H-bonds is that there does not appear to be any definite rules which govern their geometry” [7]. It became clear, however, that the main point of the discussion was centered on the H-bond nature itself, that is on whether the H-bond was electrostatic, covalent, or both, a subject on which the most imaginative positions became allowed. In 1991, Jeffrey and Saenger published “Hydrogen Bonding in Biological Structures” [8] where, for the first time, the most reliance is placed on the restricted number of accurate neutron structures and, in their absence, on carefully selected X-rays ones. This marks a turning point in H-bond studies: we accept the idea that our previous theories may be in error because based on insufficiently accurate experimental data, suspend temporarily any judgment on them, and start again to collect the widest and most reliable set of H-bond data from which to infer the true nature of the H-bond and then to lay sound foundations for any further theoretical advance. In the last 15 years, this novel data-oriented method of dealing with the H-bond problem has involved many researchers worldwide who, taking advantage of the existing crystallographic (CSD) [9] and thermodynamic (NIST) databases, have produced substantial changes in our way of considering the H-bond phenomenon. These changes will be the object of the present lecture.