GENERAL DISCUSSION 42 1 Closing Remarks BY A. D. BUCKINGHAM We have had a splendid Discussion on a topic of considerable general interest. Chemists have often been tempted to postulate " complex formation " to cover ignorance or inability to reconcile experiment and simple theory. The study of the weakly bound species that we call Van der Waals molecules clarifies this notion of a " complex." At this meeting we have seen and heard that many different techniques can provide information on the structure and properties of these " complexes " in the gas phase. What is a Van der Waals molecule? Between sessions I have been asking experts for their views, but I have to report that my consultations have not brought out an agreed definition. There is a feeling that the concept may be temperature-dependent, and that at high enough temperatures molecules that we normally regard as " good ", such as H, and N,, might qualify.There is general agreement that Van der Waals molecules are weakly bound and have large-amplitude vibrations. But what do we mean by " weakly bound " ? Do we mean that the depth of the potential well is small compared with that in H,, or in (H,O),, or relative to kT? My own suggestion is that a Van der Waals molecule is one whose dissociation energy is less than ca. 40 kJ mol-' (10 kcal mol-1 or 3000 cm-l). So some hydrogen-bonded complexes are on the borderline between Van der Waals molecules and " good '' molecules. An im- portant feature of a Van der Waals molecule is the large-amplitude vibrational motion, and this means that the structural and electronic properties (" bond " lengths and422 GENERAL DISCUSSION angles, dipole moments, polarizabilities, etc.) may differ widely in different vibrational states, and perhaps even in different rotational states.And in a gas of Van der Waals molecules at equilibrium at room temperature, many different quantum states will be populated. This definition of a Van der Waals molecule is not temperature-dependent, but those who think that H2 would be a Van der Waals molecule in an environment at T E 5 x lo4 K may be able to reconcile their view with the definition by noting that the dissociation energies of highly excited vibrational states of H, are much less than 40 kJ mol-l. If we accept this definition then we should have to rule that Dr. J.Tennyson’s poster entitled “Ab initio structure and dynamics of a Van der Waals molecule: potassium cyanide (KCN) ” does not qualify! It is not for me to attempt to summarise the many contributions to this Discussion -the authors have done that for themselves, and the varied and lengthy discussions have brought out numerous points of interest. 1 was particularly impressed by the beautiful results reported by the spectroscopists; their progress will obviously en- hance our knowledge of intermolecular potential-energy surfaces. Atld the study of molecular clusters is very interesting, particularly in bridging the gap between the gaseous and condensed phases. We heard much on predissociation and on calcula- tions of rates of dissociation, and those contributions are specially topical because of their relevance to unimolecular reaction rate theory, To conclude, 1 should like to suggest that, in spite of what we have heard in this Discussion, there is still hope that we can achieve a qualitative understanding of the structures of Van der Waals molecules through the simple notions of long-range forces and molecular shape.Prof. Klemperer reported that his observation of a structure of the type shown in (I) for the formaldehyde-hydrogen-fluoride Van der Waals molecule meant that dipolar forces could not be responsible. I agree that dipole-dipole forces do not describe the attraction at these short distances, but can we rule out the electro- static interaction? It would be interesting to examine the actual electrostic energy in this case. But whether or not it favours the observed structure, I believe that useful predictions can be made on the basis of electrostatic forces coupled with the shape of the interacting molecules, as represented, for example, by atomic Van der Waals radii.After Klemperer et al. had reported some years ago that the benzene and hexa- fluorobenzene dirners (C6H6), and (C6F& were dipolar, I suggested that the dimer of I ,3,5-C6F3H, might be non-polar, with a face-to-face structure staggered by 60” so that the H atoms are near the F atoms of the neighbour. Such a structure is favoured by the electrostatic and dispersion forces, because of the closeness of the approach. This particular dimer has, I understand, been found by Klemperer and his group to be non-dipolar. It will be interesting to discover if the crystal structure of 1,3,5-tri- fluorobenzene differs radically from those of benzene and hexafluorobenzene. Similarly, the dimer of s-triazine, (s-C,N,H,),, might be non-dipolar. So, following Prof. Ewing, I suggest that there are ‘‘ propensity rules ” for under- standing and predicting the structures of Van der Waals molecules. The ‘‘ rules ” are based only on considerations of the long-range forces associated with the properties of the isolated molecules, including their shapes, and they will be valid if the effects ofGENERAL DISCUSSION 423 electron exchange can be neglected except in so far as they determine the Van der Waals radii of the atoms. Finally, it is abundantly clear that there is no shortage of interesting research to be done by both experimentalists and theorists on Van der Waals molecules. W. Klemperer et al., J. Chem. Phys., 1975, 63, 1419; 1979, 70, 4940. A. D. Buckingham, personal communication to Prof. Klemperer, March 1978.