RESUMO

An energy-based structural indicator for water, [Formula: see text], has been recently introduced by our group. In turn, in this work we aim at: (1) demonstrating that [Formula: see text] is indeed able to correctly classify water molecules between locally structured tetrahedral (T) and locally distorted (D) ones, circumventing the usual problem of certain previous indicators of overestimating the distorted state; (2) correlating [Formula: see text] with dynamic propensity, a measure of the molecular mobility tendency, in order to seek for the existence of a connection between structure and dynamics within the supercooled regime. More specifically, in the first part of this work we will show that [Formula: see text] accurately discriminates between merely thermally deformed local molecular arrangements and truly distorted molecules (defects). This fact will be made evident not only from radial distribution function results but also from the dynamic propensity distributions of the different kinds of molecules. In turn, we shall devote the second part of this work to finding correlations between T and D molecules with low- and high-dynamic-propensity molecules, respectively, thus revealing the existence of a link between local structure and dynamics, while also making evident the dominant role of the D molecules (defects) in the structural relaxation. Moreover, the availability of a proper molecular classification technique will enable us to study the timescale of such influence of structure on dynamics by defining a modified dynamic propensity measure and by applying it to the structured and unstructured water molecular states.