Comment on: Melting behavior of SiO 2 up to 120 GPa (Andrault et al. 2020)

The additional work we have done using our new laser heating in the diamond anvil cell system since the publication of Andrault et al. (Phys Chem Mineral 47(2), 2020) leads us to the conclusion that there was a systematic bias in the determination of temperature. First, the temperature of the W-lamp used for the calibration of the optical system was overestimated by ~ 22 K at 2273 K. Then, we made the assumption that hot SiO 2 was a grey-body (constant emissivity ε ( λ )), while the available measurements suggest instead that ε ( λ ) of SiO 2 is similar to that of tungsten. Applying these two corrections lowers the SiO 2 melting temperatures significantly. In LMV, we performed a new experimental determination of the SiO 2 melting temperature, at 5000 (200) K and ~ 70 (4) GPa, which is well compatible with the amplitude of the correction proposed. The reevaluation of the melting temperature profile does not affect largely the interpretations or the main conclusions presented in Andrault et al. (Phys Chem Mineral 47(2), 2020). Within the stability field of stishovite, the melting curve still presents a relatively sharp change of slope at P–T recalculated as ~ 40 GPa and ~ 4800 K. It is related to a change of the melt structure. At higher pressures, the melting curve is almost flat up to the subsolidus transition from stishovite to the CaCl 2 -form around 85 GPa, where the slope of the melting curve increases again up to ~ 120 GPa. We present corrected figures and tables of the original publication.