Avaliação da anomalia da teoria de nucleação clássica em vidros silicatos

Abstract

Under some conditions, the Classical Nucleation Theory (CNT) provides a good description of the temperature dependence of the experimental homogeneous nucleation rates in supercooled liquids. Allegedly, however, it strongly overestimates the experimental nucleation rates below the temperature of maximum nucleation (𝑇_max ~ 𝑇_𝑔, glass transition temperature), configuring the so-called nucleation anomaly. This apparent discrepancy has been associated with some hypothesis related to thermodynamic and diffusional parameters. However, a definitive explanation about the origin of this issue is still open, since some of the hypotheses raised cannot explain it, while others are limited to theoretical analysis without experimental verification. In this thesis, we evaluate the alternative hypothesis that the CNT anomaly can be due to an experimental artifact, which leads to the underestimation of the experimental nucleation rates below 𝑇_𝑔. This hypothesis is based on the fact that the CNT is only valid when the analyzed system reaches a stationary regime. To validate or refute this hypothesis, Li2Si2O5, and BaSi2O5 glasses were chosen as model systems. The study was conducted through the implementation of heat treatments above and below the 𝑇_𝑔 of these glasses. In particular, the heat treatments below 𝑇_𝑔 were considerable extended and examined by a steady-state criterion. Also, viscosity and growth rates were used as alternative proxies for the nucleation diffusion coefficient. Additionally, the possible influence on the nucleation rates of factors as the chemical homogeneity, water content, and thermal history of the glasses were evaluated. The results not only validate the proposed hypothesis, but also offer new experimental evidence supporting recent studies that suggest the interdependence between the structural relaxation and the nucleation process, as the fundamental origin of the alleged CNT anomaly.