A perspective on the fragility of glass-forming liquids

We discuss possible extraneous effects entering in the conventional measures of "fragility" at atmospheric pressure that may obscure a characterization of the genuine super-Arrhenius slowdown of relaxation. We first consider the role of density, which increases with decreasing temperature at constant pressure, and then the potential influence of the high-temperature dynamical behavior and of the associated activation energy scale. These two effects involve thermodynamic parameters and the strength of the "bare" activation energy reflecting the specific bonding between neighboring molecules. They vary from system to system with, most likely, little connection with any putative collective behavior associated with glass formation. We show how to scale these effects out by refining the definition of fragility and modifying the celebrated Angell plot. We dedicate this note to our great and so inspiring friend, Austen Angell, and associate in this tribute another dear colleague who died too soon, Daniel Kivelson.