Fig. 6: Structural origin of the m-MRD correlation revealed by in-situ neutron total-scattering analysis.

a FSDP regions of F(Q) of three CAS glasses measured from room temperature up to 1.1 T_g and NS20 glass from RT to 1.25 T_g. The RT−F(Q) is plotted in black, then the color changes following the rainbow color spectrum, until the F(Q) of 1.1 T_g is plotted in red. The three F(Q)s with T higher than 1.1 T_g for NS20 glass are plotted by gray dashed lines. Both CAS40 and CAS50 glasses were measured at ten temperatures, as labeled in the figure legend; due to time limitations, CAS70 was only measured at seven temperatures with T = 0.4, 0.6 and 0.8 T_g skipped. b An analogous OLUA plot of scaled-MRD_T/MRD_Tg as a function of scaled temperature, T/T_g for the four glasses. The slopes for the MRD change in both supercooled-liquid and glass ranges are determined by linear fits, as shown by the dotted lines. Note the significant discontinuous increase of MRD_T/MRD_Tg on passing through T_g. The inverse linear correlation between the through-T_g slope ratio (liquid/glass) of MRD as a function of T/T_g and RT-MRD (c), along with the positive correlation between the through-T_g slope ratio and fragility-m (d), provide a semi-quantitative justification for the inverse linear correlation between m and RT-MRD (e), as illustrated in Fig. 5 for 89 glasses. The error of the slope ratio is propagated from the linear fitting error in (b).