Superstrong nature of covalently bonded glass-forming liquids at select compositions

Abstract

Variation of fragility (m) of specially homogenized Ge x Se 100−x melts are established from complex specific heat measurements, and show m(x) has a global minimum at an extremely low value (m=14.8(0.5)) in the 21.5% < x < 23% range of Ge. Outside of that compositional range, m(x) then increases at first rapidly then slowly to about m=25-30. By directly mapping melt stoichiometry as a function of reaction time at a fixed temperature T>T g , we observe a slowdown of melt-homogenization by the super-strong melt compositions, 21.5% < x < 23%. This range furthermore appears to be correlated to the one observed between the flexible and stressed rigid phase in network glasses. These spectacular features underscore the crucial role played by topology and rigidity in the properties of network-forming liquids and glasses which are highlighted when fragility is represented as a function of variables tracking the effect of rigidity.Finally, we investigate the fragility-glass transition temperature relationship, and find that reported scaling laws do not apply in the flexible phase, while being valid for intermediate and stressed rigid compositions.

Topological Ordering and Viscosity in the Glass-Forming Ge–Se System: The Search for a Structural or Dynamical Signature of the Intermediate Phase

“…There may also be an issue in interpreting the imaginary part of the heat capacity signal C ′′ P from TMDSC experiments, which is used to extract m DSC , when it cannot be represented by a single Gaussian function, e.g., when there are two relaxation channels that originate from different structural motifs (Yang et al, 2012;Gulbiten, 2014). A shift in the Gunasekera et al (2013) m DSC values to better match the fragility index of glassy Ge 0.10 Se 0.90 found in the work by Svoboda and Málek (2015) leads to results that are more consistent with the m DSC values of 23(2)-27(2) measured for Ge 0.22 Se 0.78 by Li et al (2017), and better match the measured composition dependence of m visc (Figure 12). In comparison, the m DSC values of Zhao et al (2013) FIGURE 6 | The composition dependence of the measured total structure factors S(q) ≃ S NN (q) for the GexSe 1− x system.…”

“…For example, a least-squares parabolic fit to the m visc values of Senapati and Varshneya (1996) leads to a minimum at x = 0.196(2), whereas a similar fit to all of the m visc data points leads to a minimum at x = 0.223(2), consistent with the value x = 0.225 previously reported by Stølen et al (2002). The m DSC values of Gunasekera et al (2013) are smaller than other values of the fragility index and, for several intermediate phase compositions, are even smaller than the fragility index of silica m visc ≃ 21, where the latter was obtained by applying the MYEGA model to the viscosity data listed by Doremus (2002). A large disparity between m visc and m DSC is, however, unexpected for strong glassforming systems: the approximation m visc ≃ m DSC is expected to become less accurate with increasing fragility because of the use of an Arrhenius approximation in DSC work, where the m DSC values are often smaller than their m visc counterparts (Zheng et al, 2017).…”

“…The pairs of vertical dashed (black) or chained (red) lines, and associated horizonal arrows, mark compositions for which ∆Hnr ≃ 0 as found in the present work or in the work of Boolchand et al (2001a), respectively. Gunasekera et al (2013), but the calculated viscosities are several orders of magnitude larger than expected from viscosity measurements (see also Figure 11). A maximum in log 10 η(T L ) occurs at (iv) m = 0.20(1) if m visc is estimated from the m DSC values of Zhao et al (2013), but in this case the calculated viscosities are significantly smaller than expected from viscosity measurements.…”

“…A large disparity between m visc and m DSC is, however, unexpected for strong glassforming systems: the approximation m visc ≃ m DSC is expected to become less accurate with increasing fragility because of the use of an Arrhenius approximation in DSC work, where the m DSC values are often smaller than their m visc counterparts (Zheng et al, 2017). As discussed by Svoboda and Málek (2015), the small m DSC values of Gunasekera et al (2013) may originate from the exploration of a narrow range of relaxation times in their TMDSC experiments. There may also be an issue in interpreting the imaginary part of the heat capacity signal C ′′ P from TMDSC experiments, which is used to extract m DSC , when it cannot be represented by a single Gaussian function, e.g., when there are two relaxation channels that originate from different structural motifs (Yang et al, 2012;Gulbiten, 2014).…”

“…However, Ge x Se 1−x glasses made by using an almost identical rocking-furnace procedure show no evidence of sample heterogeneity (Section 3.1). In the work of Gunasekera et al (2013), a small fragility index m DSC = 14.8(5) for Ge 0.22 Se 0.78 (Figure 12) might be attributed to the preparation of homogeneous glass after a long reaction time of 144-216 h for samples of mass 2 g. However, systematically larger values of m DSC = 23(2) and m DSC = 27(2) are reported for Ge 0.22 Se 0.78 samples of similar mass (∼1.5 g) prepared using short versus long reaction times of 34 and 192 h, respectively (Li et al, 2017). The neutron diffraction results of the present work do not show any obvious structural signature of the intermediate phase.…”