Quantum tunnelling in water clusters

Many areas of computational chemistry and biology require accurate and computationally efficient potential energy surfaces to describe the interactions between water molecules. A great deal of progress has been made in developing and modelling such potentials, but much remains to be understood. The contemporary importance of this field is evident from new activity generated by recent experiments, and the opportunity afforded by novel instanton theory for quantum dynamics calculations suggests that rapid progress will now be possible.

The study of water clusters is in principle a very powerful technique for developing and refining water potentials. Although the dynamics of such clusters may be far from that of water in the bulk, the interactions between the water molecules are of course the same, and the advantage of studying water clusters is that they are prepared at very low temperatures in a molecular beam, thus allowing precise and detailed spectroscopic measurements to be made, which respond sensitively to the properties of the water potential. If one can develop a method for computing these spectral lines from the potentials, then one has established a powerful, direct, link between the water potential and experiment.

Developing such a method, and applying it to clusters containing from 4 to at least 20 water molecules is the primary goal of this proposed research. The particular transitions that we will study are those that involve quantum tunnelling between different permutational isomers of the water clusters. This analysis will allow us to use a novel 'instanton' method, which is a systematic way of obtaining a good approximation to the dominant tunnelling paths. This method has already been tested on water dimer and trimer, and shown to give excellent results that reproduce experiment. The proposed research will augment and develop further these techniques, permitting them to be applied to clusters containing up to around 20 water molecules. This work will result in the first predictions of the tunnelling splitting patterns for these clusters, which will then be compared with experimental measurements made in the group of project partner Rich Saykally (Berkeley, USA). These comparisons will then allow us to improve and refine the water potential energy surface, which will be conducted in collaboration with project partner Joel Bowman (Emory, USA).

In addition to water clusters, we will also study complexes of water with hydrocarbons. This work will result in better potential energy surfaces for describing the interactions in gas hydrates, which will lead to more reliable simulations of these systems and new results that will be relevant to studies of global warming and exploitation of alternative energy reserves. An oil consultancy software company (InfoChem) is very interested in possible developments resulting from this work, and is named as one of our project partners. High resolution spectra for hydrocarbon complexes such as water-methane have already been obtained in the Saykally group, and our calculations will be carried forward with ongoing feedback from experiment.

The improvements to the water potentials that result from this work are likely to lead to more reliable simulations of water in all its phases, and thus to lead to better representations and understanding of the vast range of important chemical and biological systems that contain water.

The results of this project will lead to an improved understanding of the interactions between water molecules. A better understanding of water is key to many areas of science, technology and medicine. For example, water potentials are used to simulate and predict the efficacy of drugs and to model chemical reactions in an appropriate biochemical environment. Any improvements in such water potential energy surfaces resulting from this work will therefore have a far-ranging impact. A major improvement, allowing such potentials to yield predictions of near-quantitative accuracy, would of course have a massive impact. This is indeed our aim in this project. The capability to calculate realistic tunnelling splittings for larger water clusters and compare directly with experiment is now provided through a combination of state-of-the-art instanton theory and optimisation techniques. We will therefore be able to treat new systems that have hitherto been out of reach, and the insight obtained will be used to produce new intermolecular water potentials.

These results will be of great interest to the extensive community engaged in simulations involving explicit water molecules, which includes some of the most active areas of contemporary computer modelling. To widen the impact as far as possible we will include potentials with alternative balances of accuracy and complexity, which can then be applied depending on the problem of interest. In fact, it may be desirable to employ different descriptions within the same simulation, for example, treating water molecules near active sites or functional groups at a higher level of theory than more distant regions.

Another aspect of the project that will have wide-ranging societal implications are the proposed applications of the framework we develop to water-hydrocarbon clusters. Just as our studies of homogeneous water clusters will provide new potential energy surfaces for pure water, the water-hydrocarbon work will determine improved potential energy surfaces for modelling gas-hydrates. Gas hydrates are relevant for gas storage and transportation, natural gas recovery, and safety in oil and gas pipelines. While the formation of hydrates can block pipelines, causing practical issues for natural gas transport, naturally occurring deposits represent a huge reserve that is largely unexploited. More accurate yet efficient intermolecular potentials for the corresponding molecules could form the basis for new modelling efforts aiming to understand and prevent hydrate formation where it is an impediment. This work will have an immediate impact on our industrial project partner (InfoChem). We have consulted InfoChem directly to better understand the most important potential impacts of our research for the petrochemical industry. In fact, there are a number of applications where accurate water-hydrocarbon potentials would be immediately useful, ranging from structure prediction in gas-hydrate lattices (especially host-guest systems), to simulation of phase behaviour. Large-scale calculations of phase changes, such as crystallisation, using accurate potentials cab inform the more coarse-grained descriptions that are used for complex, multi-component system.

These initial impacts in the field of biomolecular simulation, and computer modelling in the petrochemical industry can serve as a platform for future work, where the framework we develop is applied to an even more diverse range of systems. For example, the ability to calculate tunnelling splittings as a basis for developing new interatomic and intermolecular potentials could be employed in a variety of diverse areas of materials science.