Fig. 3: Temperature-dependence of the structure factor first-peak position, q_1(T), obtained from wide-angle X-ray scattering (WAXS) and MD simulations.

A Experimental data for the glycerol-water solutions (χ_g = 3.2% glycerol mole fraction) obtained from WAXS measurements using table-top X-ray diffraction (XRD) at moderately supercooled temperatures (T ≥ 250–295 K, open black squares), and by femtosecond X-ray scattering of evaporatively cooled microdroplets (T = 229.3–250 K, open black circles). For comparison, we include the experimental q₁(T) of pure liquid water (open red circles and triangles) from refs. ^7,53. B q₁(T) obtained from MD simulations of glycerol-water solutions (χ_g = 3.2% glycerol mole fraction, open black circles) and pure liquid water from ref. ⁴⁹ (open red triangles). Solid lines are smoothing spline fits. The inset in (B) shows the temperature-derivative dq₁/dT for the glycerol-water solution (open black circles) and pure water (open red triangles). The maximum in dq₁/dT occurs at T = 228 K for the glycerol-water solution and at T = 233 K for pure water, respectively. In both figures, error bars smaller than the size of the corresponding data points and the red arrows are used to indicate the q₁ of low-density amorphous (LDA) ice, according to experiments⁸⁴ at T = 80 K, and simulations⁸⁵ at T = 80 K. Error bars indicate the standard error obtained from the gaussian fit to the first S(q) peak.