Research Interests

Fast proton transport (PT) between hydrogen-bond donor and acceptor atoms is of paramount importance in many aspects of chemistry and biology. Applications are diverse and include many technological developments such as hydrogen fuel cells and storage materials. In nature all reactions that convert energy from one form into another are mediated by PT, which also serves as a vital route to achieve cell pH stabilisation. Proton migration (PM) is a much more subtle effect, where the time-averaged position of the H atom in a short, strong hydrogen bond changes as a function of temperature and pressure. Our aim is to understand the mechanisms of PT and PM using simulation. This work is challenging for three main reasons. First, bond formation and breaking events are not described by conventional molecular mechanics force fields, necessitating the use of a quantum-mechanical technique. Secondly, because of the small mass of the hydrogen atom, quantum effects such as tunnelling and zero-point energy contributions can radically alter the reaction landscape. Thirdly, PT is a rare event, meaning that it will only be observed a couple of times during the molecular dynamics trajectory, with obvious consequences for statistical sampling. We have presented a reaction mechanism for the double PT reaction that occurs spontaneously in carboxylic acid dimers, considering atoms both as classical and quantum particles. We have begun complex simulations on a flexible model-system of transmembrane proteins, where excess protons are transported by chains of water molecules embedded in an alpha-helix domain. We have also performed extensive molecular dynamics calculations that go a considerable way towards offering an explanation for the PM phenomenon.

This research area has the ultimate aim of enhancing the way crystallographic data are analysed. Currently, the vibrational motions of atoms are restricted to rectilinear motions (i.e. standard thermal ellipsoids). There are many cases, however, where that approximation breaks down, e.g. groups with nearly-free rotation, hydrogen-bonded materials, and super-conducting materials close to the critical temperature etc. This weakness impacts strongly on the quality of the structure obtained, with geometric parameters (bond lengths, angles etc.) in error. Our approach makes use of molecular dynamics simulations to obtain more realistic PDFs in the solid-state. A recent proof-of-concept study reported the calculation of a PDF for a deuterium atom in the crystal structure of CD₃NO₂ (see diagram). The distribution is not a standard ellipsoid: it is curved and the distribution is asymmetric. We are now working to fit generic functions of this type to analytical equations that will require only a couple more parameters than that needed to describe a standard ellipsoid. The new functions will ultimately be incorporated into public-domain crystallographic software.

The use of pressure as a thermodynamics variable to access new polymorphs is a very active area of research. High pressure single crystal X-ray diffraction (XRD) experiments can be readily performed on standard diffractometers via the use of diamond anvil compression cells (DAC). There is, however, one considerable drawback with the method - the chunky steel casing of the DAC restricts the amount of reciprocal diffraction space that can be harvested to ca. 30%, compared to a standard ambient pressure data collection. This results in structures with much larger standard deviations, the inability to distinguish between atoms that scatter X-rays to the similar degrees (e.g. N and O in the diagram given), but perhaps more worrying for molecular systems, the absence of information for the positioning of hydrogen atoms, which is of particular importance in verifying the hydrogen bonding network. One route to a obtaining complete crystal structure at high pressure is a neutron diffraction (ND) experiment, but this involves making an external application for beam time and introduces a time delay in obtaining results. We therefore set ourselves the challenge of applying QM modelling techniques (plane-wave DFT geometry optimisations and molecular dynamics simulations) to complete the partially determined experimental structures before the ND experiments were run. The outcome from these advanced experiments would therefore act to validate our modelling approach.