Mixtures involving water or carbon dioxide

Carbon dioxide capture, transport and storage are vital to reducing the greenhouse gas content of the atmosphere. We are planning to use new theoretical and computational techniques to discover important properties such as the phase behaviour, density, critical point and maximum water content of a range of carbon dioxide mixtures at different pressures and temperatures.

Our research on water mixtures involves the calculation of thermodynamic properties for mixtures of atmospheric gases and combustion gases. The thermodynamic properties of low-pressure gases can be related to the virial coefficients, which themselves can be deduced from a knowledge of the forces between the molecules. We also calculate high-order virial coefficients, equations of state and correlation functions for simple model interactions, such as hard spheres and hard discs, steeply repulsive potentials, square-well potentials, and ellipsoidal molecules.

Van der Waals forces

Weak Van der Waals forces between molecules arise from a combination of electrostatic, induction, dispersion and electron exchange effects. For non-polar, closed-shell molecules, the dispersion attraction and exchange repulsion dominate. For hydrogen-bonded complexes, the electrostatic and induction forces are also important.

The goal is the calculation of an intermolecular potential energy surface, which describes the interaction between the molecules of interest. The intermolecular potential is a multidimensional function which must be calculated at hundreds or thousands of different points corresponding to different relative positions of the molecules.

Direct calculations using quantum chemistry computer programs are not usually suitable for this, because they take too long, and are inaccurate, especially when calculating the dispersion energy. My group uses a variety of methods to overcome this.


Suppose that we are interested in the interaction between two small molecules A and B. It will usually be possible to calculate the properties of A and B accurately using quantum mechanics (monomer calculations). However the calculations of the A-B intermolecular potential (dimer calculations) will be less accurate, since the calculations are harder, and there are many more of them to do. The table illustrates the situation.

Low-level monomer calculations Low-level dimer calculations
High-level monomer calculations High-level dimer calculations

Using the Systematic InterMolecular Potential Extrapolation Routine (SIMPER), we shall be able to estimate the results of high-level dimer calculations by extrapolating the low-level dimer results. This will be made possible by new findings connecting the dimer results to the corresponding monomer results.

Computational studies of supercritical fluids.

Supercritical carbon dioxide is one of the chemical solvents of the future. It is environmentally friendly, cheap and easy to produce, and its physical properties can be changed substantially with small changes in temperature and pressure, leading to control over chemical reaction products.

Non-critical and supercritical conditions

We are interested in the unexpectedly large solubility of fluorinated hydrocarbons in supercritical carbon dioxide. These molecules dissolve in carbon dioxide much better than water does, so understanding this phenomenon may lead to the design of new surfactants for stabilising water-containing micelles within the carbon dioxide solvent.

Our investigation is a joint theoretical study (by Dr Wheatley and Prof Hirst) and experimental investigation (by Dr Ke and Prof Poliakoff) into the thermophysical properties of carbon dioxide / fluorohydrocarbon mixtures. We are currently investigating the effect of non-additive forces and non-rigidity of carbon dioxide molecules on the vapour-liquid curve of pure carbon dioxide.

Ab initio spectroscopy

With the intermolecular potential energy surfaces calculated by SIMPER, we are able to predict the rotational and vibrational energy levels of the Van der Waals molecules, and therefore predict the spectrum in the microwave and far-infrared region. We have recently shown that SIMPER is better than much more expensive ab initio methods for predicting the spectra of a range of different interactions, including Ar-He and HF-Ne.