Water is a particularly important substance; it is the essential medium in biology, a major component in terrestrial geology, and a key reagent in the chemistry of the atmosphere.
Biological systems are mostly water, the physiological liquid of the human body contains about 0.8mol of salt, ie one ion pair per 100 water molecules.
The ocean is a concentrated salt solution, sea-salt aerosols serve as the global source of Cl- and Br- in atmosphere.
Many geological processes involve the chemistry of water, such as the dissolution of minerals in leaching and weathering processes.
Clouds are formed by water that condenses around around aerosols, and control the radiative balance of the planet.
At a molecular level water is a particularly challenging system to study. The picture above shows a snap-shot of water molecules in a super-cell, the red atoms are oxygen and they white atoms are hydrogen, imagine this repeated infinitely many times in every direction.
Water is very difficult to model computationally because:
- it is extended, ie it has no boundaries the way a discrete molecule does
- it is disordered, ie there is no atomic symmetry or ordering as in a crystalline or metallic material.
- an extended network of hydrogen bonds which breaks and reforms over very short periods of time (femto seconds!).
Previous studies in this area have been carried out using purely classical methods, and have been unable to represent any chemistry that is occurring. The Car-Parrinello method has uses density functional theory to model the electronic structure, while retaining classical equations for the nuclear motions.
Over the last few years it has been shown that the structure, dynamics and transport properties of liquid water can be successfully simulated using Car-Parrinello molecular dynamics.