Comment on 'Structure of supercooled liquid silicon' by Ansell et al

We correct a statement made in a recent paper (Angell S et al 1998 J. Phys.: Condens. Matter 10 L73-8) that computer simulations on supercooled liquid silicon were not available. We further point out that the simulations that have been made provide an important confirmation of the findings of Angell et al, from the x-ray diffraction studies, that the Si coordination number rapidly decreases in the supercooling range. Of particular interest is the observation that, at the limit of the 340 K supercooling reported in Angell et al, the laboratory liquid has the same coordination number that the simulated system reaches at the temperature of its first order liquid-liquid transition. This implies that the crystallization is promoted by the liquid-liquid transition, as is also seen in the simulation. We point out other systems in which such provocative behaviour should be found.