Hydrogen bond strength as a function of ring size.

One frequently has to confront the question: will a hydrogen bond form between a suitable donor (lone pair or π) and an acceptor? One of the factors to be taken into consideration for hydrogen bonds which are part of a cycle is the ring size. Here I explore one way of quantifying the effect for the series below, n=1-5 (4-8 membered rings).

I will use the NBO approach. To remind, this reduces the wavefunction for a molecule to a set of localised orbitals, referred to as natural bond orbitals. The perturbation interaction energy E(2) between any (doubly occupied, i.e. donor) orbital and an (unoccupied) acceptor orbital establishes the strength of that interaction. For a hydrogen bond, this can be expressed as the NBO corresponding to the (in this case oxygen) lone pair (shown in orange and purple below) and the corresponding H-O σ* empty orbital (shown as red and blue below). E(2) is a function both of how close in energy this pair of orbitals is (the smaller the energy gap the better) and how well they overlap (the relevant overlap in this case is the positive one between purple and blue). This latter attribute is shown below for the series n=2,3,4,5 (n=1 does not form any discernible hydrogen bond), at the ωB97XD/6-311G(d,p) computational level.

NBO interaction for 5-ring H-bond.

NBO interaction for 6-ring H-bond.

NBO interaction for 7-ring H-bond.

NBO interaction for 8-ring H-bond.

The interaction energies E(2) are collected below, together with the computed lengths. To put E(2) into context, it is around 16 kcal/mol for a strong anomeric interaction, and about 6 kcal/mol for the stereoelectronic influence in di-fluoroethane. One can see that by the time the angle subtended at the hydrogen has increased to ~150°, the interaction energy has reached a respectable value.

E(2), kcal/mol  O…H length, Å  Angle, °
1 ~0.0  –  81.8
2 0.75 2.294  109.7
3 3.56 1.984  139.6
4 6.24 2.017  146.5
5 8.35 1.957  153.4

So the simple trick of looking at the donor-acceptor NBO interaction in a cyclic hydrogen bond can give us a straightforward way of quantifying how the size of the ring and hence the orbital overlap (one presumes that the Lp/C-O σ* energy gap is similar for all the systems) affects the strength of the interaction. One might also explore this by looking at structures in the Cambridge crystal database. But note from the above that whilst the  E(2) energies follow ring size, this does not appear to happen for the H…O lengths! The analysis reveals that the maximum number of structures for the span 5 to 8-rings occurs at ~2.15, 1.85, 1.65 and 1.85Å respectively. 

Crystal data for 5-rings

Crystal data for 6-rings.

Crystal data for 7-rings

Crystal data for 8-rings

Henry Rzepa

Henry Rzepa is Emeritus Professor of Computational Chemistry at Imperial College London.

View Comments

Recent Posts

Internet Archeology: reviving a 2001 article published in the Internet Journal of Chemistry.

In the mid to late 1990s as the Web developed, it was becoming more obvious…

1 month ago

Detecting anomeric effects in tetrahedral carbon bearing four oxygen substituents.

I have written a few times about the so-called "anomeric effect", which relates to stereoelectronic…

1 month ago

Data Citation – a snapshot of the chemical landscape.

The recent release of the DataCite Data Citation corpus, which has the stated aim of…

2 months ago

Mechanistic templates computed for the Grubbs alkene-metathesis reaction.

Following on from my template exploration of the Wilkinson hydrogenation catalyst, I now repeat this…

2 months ago

3D Molecular model visualisation: 3 Million atoms +

In the late 1980s, as I recollected here the equipment needed for real time molecular…

3 months ago

The Macintosh computer at 40.

On 24th January 1984, the Macintosh computer was released, as all the media are informing…

3 months ago