Remarks on the provisional recommendation

DEFINITION MOF THE HYDROGEN BOND

These remarks originate from two chemists, one of whom (YM) is concerned with hydrogen bonding in liquid and supercritical water and solutions therein and the other (MR) with vibrational spectroscopy of hydrogen bonded systems.

In view of the insights that we have gained on the subject of hydrogen bonding, we now suggest a modification of the definition as follows:

The hydrogen bond is an attractive, directional, short range and saturated interaction between a hydrogen atom H from molecule or molecular fragment X–H in which X is more electro-negative than H, and an atom Y or a group of atoms in the same or a different molecule, in which there is evidence of bond formation X–H∙∙∙Y–Z.

The additions (in italics) sharpen the definition and make it more concrete, without detracting from its generality, and the deletion (struck through) removes rare cases of i) bifurcated hydrogen bonds and ii) hydrogen bonds to π-electron systems that ought to be dealt with in a footnote.

There are also some comments on the criteria. A major deficiency in the criteria is that according to the last phrase in the definition (“there is evidence of bond formation X–H∙∙∙Y–Z”), there should be recognition of the evidence, namely X–H stretching IR band shifts and enhancement and NMR proton deshielding. These must be satisfied, either experimentally or demonstrated theoretically for the existence of a hydrogen bond to be recognized.

The criteria suggested by persons dealing with computer simulations of (liquid) hydrogen bonded systems appear to have been disregarded. Xenides, Randolf, and Rode J. Mol. Liquids 123, 61 (2006) suggested as criteria for hydrogen bonds in water:

a) the angle of the O–H∙∙∙O configuration has to be θ ≥100°,

b) the distance between two oxygen atoms of neighboring water molecules has to

be 0.25 ≤ d_O–O/nm ≤ 0.35, and

c) the hydrogen bond distance has to be 0.15 ≤ d_O∙∙∙H/nm ≤ 0.25: “strong” hydrogen

bonds, have d_O∙∙∙H/nm ≤ 0.21 and “weak” ones have it up to d_O∙∙∙H/nm ≈ 0.25.

Kumar, Schmidt, and Skinner. J. Chem. Phys. 126, 204107 (2007) suggested somewhat more restrictive geometrical criteria:

a’) the angle of the O–H∙∙∙O configuration has to be θ ≥130°, and

b’) the distance between two neighboring oxygen atoms has to be d_O–O/nm ≤ 0.330, c’) the hydrogen bond distance has to be d_O∙∙∙H/nm ≤ 0.241,

but also specified the energetic criterion:

d’) the interaction energy between the hydrogen bonded water molecules should be more

negative than –12.9 kJ mol^–1.

e’) the occupancy of the σ_OH^* antibonding molecular orbitals of the acceptor atom has to

be ≥0.0085.

Furthermore, the hydrogen bond dynamics should also be considered. Voloshin and Naberukhin, J. Struct. Chem. 50, 78 (2009) among others suggested life times of the order of 1 ps as required for a viable hydrogen bond to exist, whereas shorter lived configurations conforming to the geometric criteria are only transients.

In view of these comments, E1 is not a criterion for the formation of a hydrogen bond, especially in view of footnote F2. It should be deleted. E3 should be modified to recognize more fully “bent” hydrogen bonds (in footnote F4 the example of Li₂F₂ is not relevant). Criterion E4 is not worded as a criterion but as a characteristic. Some limits on the d_{X–H∙∙∙Y} distance should be provided.

There are also comments on the characteristics. C1 is too sweeping, as the pK_a and pK_b are solvent dependent. When X–H and Y–Z are located in a Xe matrix, for example, no such correlation was found. Characteristic C3 should be expanded and criteria for the formation of hydrogen bond networks should be provided.

Yizhak Marcus

Mark Rozenberg