1. Could you explain the significance of your article to the non-specialist?

The continued advancement of (widely applicable) high technologies is critical both for continual national economic prosperity and the movement towards a more eco-friendly social model. The ability to control material formation on an atomistic level (effectively "nano-engineering" is crucial if the potentially useful material properties are to be fully exploited. Computer simulations offer the possibility of both constructing these novel materials and understanding the factors which control their formation without initial recourse to potentially dangerous and expensive experimental investigation. The nanotubular structures observed in the present work may, for example, have unique mechanical and electronic properties which may be usefully exploited. These properties may even be "tunable'' by controlling the nanotube composition, diameter and morphology.

2. What has motivated you to conduct this work?

A significant motivation for the present work involves advancing the role of computer simulation modelling from a traditional responsive mode (in which simulations are applied to aid post hoc interpretation of experimental investigations) to a predictive mode in which models act to drive on-going experimental investigation. The advantages of such a move are clear. For example, simulation models can effectively drive experimental investigation then such investigations may be more focussed. Furthermore, model investigation is significantly more cost-effective. However, in order to achieve this change truly reliable and transferable potential models are required in the sense that one must have confidence that the results predicted by a given model (away from the state-points on which it is parameterised) are valid.

3. Where do you see this work developing in the future?

The work is presently developing along many directions. In one, additional nanotubular structures are being observed by filling the carbon nanotubes directly from different molten salts. The aim here, therefore, is to develop an understanding of the general classes on nanotube structures formed both as a function of the carbon nanotube itself (the diameter and morphology as well as the presence of any inherent defects) and the chemical identity of the filling salt. In another research programme, the behaviour of mixtures of salts in filling the carbon nanotubes is being investigated. One possibility is that these tubes may behave as ``nano-sieves'' and allow the effective separation of materials which are, at present, difficult to separate. In a third research programme, the simulation models themselves are continually being improved. The mooted move from a responsive to predictive simulation role relies on the construction of simulation models which reproduce experimental properties over a wide range of conditions. The lack of reliable models is, at present, a major draw back.

4. Are there any particular challenges facing future research in this area?

As ever, there are many potential challenges ahead. One major challenge in this area will be in translating the atomistic control attainable in the simulations to the experimental world. Again, however, (realistic) simulation models may act in a predictive mode in order to highlight possible controllable formation pathways. What is required, therefore, is the development of a truly symbiotic relationship between theory and experiment.