Comment on ``Quasisaddles as relevant points of the potential energy surface in the dynamics of supercooled liquids'' [J. Chem. Phys. 116, 10297 (2002); cond-mat/0203301]

Abstract

Recently, the properties of supercooled liquids have been studied by mapping instaneous configurations onto minima of the gradient squared. It was originally suggested that this mapping would probe higher-order saddle points of the potential energy surface. However, it was subsequently shown that the majority of the minima of this function are only local minima and so do not correspond to saddles. In this comment, we provide a critique of the suggestion made by Angelani et al. [J. Chem. Phys. 116, 10297 (2002); cond-mat/0203301] that although these minima are not true saddles, they are almost saddles (hence the term quasisaddles). This issue has important implications for the interpretation of the results obtained by this approach.