A short note on the possible stabilization of the local water structure in the liquid due to an oxygen or a nitrogen molecule

Abstract

In a previous paper [l] we investigated the possibility of an oxygen or a nitrogen molecule having a stabilizing effect on the local structure in liquid water. Surface tension experiments [2] had indicated that such a stabilization actually occurs. To study this problem from a theoretical point of view we used a CNDO method [3]. As a model for the local water structure we chose a cage of ten water molecules connected by twelve hydrogen bonds. The structure chosen occurs in ice Ic. Furthermore the water molecules were oriented in such a way that the cage did not have any resulting dipole moment. In that way the choice of position for the interstitial oxygen or nitrogen molecule was simplified. Two comments on this model are that first, the hydrogen bonded structure used in this work has many similarities with several of the models proposed for water in the literature (see [4] for a rCsumC), and second, our model takes only into account the interaction between the foreign molecule and the first layer of water molecules. This last limitation in our approach is not supposed to qualitatively change the conclusions. The stabilization effect calculated in [l] may perhaps be smaller when the interaction between the first layer of water molecules around the foreign molecule and the rest of the liquid is included. This is because the surrounding water may partly inhibit the charge redistribution in the first layer as calculated in [l]. The result in [l] was that the interstitial molecule could stabilize the cage by 4 to 7 kcal/mole or around 0.5 kcal/mole per hydrogen bond, depending on its orientation.We have made some further investigations of this problem. It might be argued that the cage chosen as a model for the local water structure could be too favorable for the interaction with the oxygen or the nitrogen. For that reason we have calculated the interaction between some monomer, dimer, and trimer fragments of the cage and a nitrogen molecule. When the geometries were those of cage plus interstitial molecule, the monomers were stabilized between 0.0 and 0.6 kcal/mole, the dimers between 1.1 and 1.5 kcal/mole, and the trimers from 1.1 to 2.6 kcal/mole. If one assumed the hydrogen bonds in the cage to be independent, these results would predict a stabilization of the cage due to the interstitial molecule of around 14 to 18 kcal/mole instead of the calculated 5 to 7 kcal/mole. What happens is, of course, that in the cage each hydrogen bond is not free to readjust to the interstitial molecule completely, due to the rest of the cage. By varying the.position of the nitrogen relative to the 177 01974 by John Wiley & Sons, Inc.