Si melt density has been found to have the maximum value around 2.54 g/cm(3
) at 24 degreesC above the melting point, like H2O. The anomalous variation
near the melting point reported by Sasaki et al. [1] has been found to be
due to the melt climbing to the bob when just touching the Si melt surface.
It is needed to wait more than 2h to get constant values. We used fresh bo
bs to measure one point temperature in order to eliminate the bob thickness
variations. The density variations are discussed in terms of weight signal
variations. Tentative density values are proposed. (C) 2001 Elsevier Scien
ce Ltd. All rights reserved.