Relaxation effect on crystal nucleation in a glass unveiled by experimental, numerical, and analytical approaches

Understanding the mechanisms and dynamics of crystal nucleation and growth in glass-forming materials is fundamental to avoid or control crystallization - the critical steps for developing and producing novel glasses and glass-ceramics. Below the temperature Tmax, corresponding to the maximum of experimentally measured nucleation rates, the reported nucleation rates are frequently smaller than the theoretical predictions by the Classical Nucleation Theory (CNT). This discrepancy dramatically increases with decreasing temperature, and has been denoted as the “breakdown” of the CNT. For understanding this alleged breakdown, it is important to stress that the experimental Tmax is located somewhat above or at the glass transition temperature, Tg. Here, using a 5BaO·8SiO2 (B5S8) glass, first, we showed a significant change in the characteristic relaxation time with temperature in the Tg region: it is four times shorter for the supercooled liquid than for the glass. Then a combination of detailed experiments and theoretical (analytical and numerical) analyses for the B5S8 glass confirmed our recent findings for Li2O·2SiO2 and 2Na2O·CaO·3SiO2 glasses. The “breakdown” is just an artefact, since the experimental nucleation times used are often not long enough to allow the complete structural relaxation of the glass and hence reach the ultimate steady-state nucleation rate. Finally, for the first time, we also show that due to the almost complete crystallization of the B5S8 glass, and likely of other glasses that exhibit sufficiently high nucleation and growth rates, the structural relaxation process cannot be finished and, correspondingly, the ultimate steady-state nucleation regime at T