Volumetric study consistent with a glass-to-liquid transition in amorphous ices under pressure

Dilatometry experiments on low- and high-density amorphous ices up to 0.30 GPa are presented together with powder x-ray diffraction data. Repeated isobaric heating and cooling cycles reveal three competing processes: irreversible (micro)structural relaxation, reversible relaxation, and (irreversible) crystallization. The third and subsequent heating runs produce identical curves, i.e., irreversible relaxation is absent. We interpret the deviation from linear expansivity in these curves as the onset temperature of the volumetric glass-to-liquid transition ($T_{g,\hspace{0.5em}\mathrm{onset}}$) and report its dependence on pressure.

Figure 1

Dilatometry curves obtained upon isobarically heating 300 mg of (a) LDA and (b) HDA. Both were prepared by isothermal decompression of VHDA (Refs. 28 and 29). All samples were heated to $T⩾170$ K at heating rates of 2 K/min in order to obtain fully crystalline samples. Traces labeled “2${}^{\mathrm{nd}}$” were obtained after cooling the crystallized samples and reheating at 2 K/min.

Figure 2

Powder XRDs for samples recovered to 77 K and ambient pressure. Data were recorded on a diffractometer in θ-θ geometry (Siemens, model D 5000, Cu$K$α), equipped with a low-temperature camera from Paar at ∼80 K. Curves are offset for clarity. Temperature labels indicate the maximum temperature experienced in isobaric heating runs for (a) LDA at 0.006 GPa and (b) HDA at 0.20 GPa; the labels “1${}^{\mathrm{st}}$,” “2${}^{\mathrm{nd}}$,” “3${}^{\mathrm{rd}}$,” and “6${}^{\mathrm{th}}$” indicate that the samples have been heated one, two, three, and six times to the labeled temperature, respectively.

Figure 3

Test for repeatability of dilatometry curves. (a) LDA was heated six times at 2 K/min and 0.006 GPa to ∼150 K and (b) HDA was heated six times at 2 K/min and 0.20 GPa to 144 K in order to avoid crystallization. Cooling rates were also 2 K/min. Vertical lines mark where the curves start to deviate from linearity, which we interpret to be the volumetric $T_{g,\hspace{0.5em}\mathrm{onset}}$. For comparison, the dilatometry curves of crystalline ices are shown (labeled “cryst.”). Top panel: Original data, offset for clarity. Bottom panel: Magnified view of the fifth and sixth heating runs of (a) LDA and (b) HDA are shown together with the data obtained on crystalline ices. The linear functions (black line) shown in the top panel have been subtracted from the original data, i.e., no volume change corresponds to perfectly linear expansion.

Figure 4

Volumetric $T_{g,\hspace{0.5em}\mathrm{onset}}$ for LDA (open circle) and HDA (open triangles) at heating rates of 2 K/min (not corrected for thermal lag). $T_{g,\hspace{0.5em}\mathrm{onset}}$ is determined by the method shown in Fig. 3. Error bars of ±2 K reflect the ambiguity in placing the vertical dashed line in Fig. 3. For comparison, the calorimetric $T_g$ of LDA (Ref. 8) (filled circle) and HDA (Ref. 11) (filled triangle) at 1 bar are also shown.