Direct Observation of Melting in Shock-Compressed Bismuth With Femtosecond X-ray Diffraction

The melting of bismuth in response to shock compression has been studied using in situ femtosecond x-ray diffraction at an x-ray free electron laser. Both solid-solid and solid-liquid phase transitions are documented using changes in discrete diffraction peaks and the emergence of broad, liquid scattering upon release from shock pressures up to 14 GPa. The transformation from the solid state to the liquid is found to occur in less than 3 ns, very much faster than previously believed. These results are the first quantitative measurements of a liquid material obtained on shock release using x-ray diffraction, and provide an upper limit for the time scale of melting of bismuth under shock loading.

Figure 1

The equilibrium phase diagram of Bi to 20 GPa [13, 14]. The calculated Principal Hugoniot, the locus of states accessible with a single shock, is shown in the Bi-V phase as a dashed line [15], and the release isentropes from the Hugoniot states, adapted from Ye et al. [16], are shown using different dotted-dashed lines. The uncertainty in the position reached on the Hugoniot due to shot-to-shot energy jitter from the drive laser is taken into account and is highlighted by shaded regions around the release isentropes.

Figure 2

Experimental configuration for laser-shock experiments using femtosecond x-ray diffraction at the MEC beam line at LCLS. The VISAR beam is collected perpendicular to the target rear surface (green arrow). 2D diffraction images (top right) are collected on CSPAD detectors in transmission and reflection (not shown) and are then integrated into 1D diffraction profiles (bottom right).

Figure 3

Simulated pressure profiles in the Bi targets superimposed on the initial target dimensions. These dimensions change with time and the initial position of the polyimide-epoxy–bismuth interface, and its position at $t=0\mathrm{ns}$ and $t=+3\mathrm{ns}$ are shown by the black, green, and blue dashed lines, respectively, at the bottom of the plot. After rear surface breakout, the sample pressure is reduced by both the weak release wave originating at the ablation surface, and by the centered rarefaction wave originating at the rear surface, the propagation directions of which are indicated by arrows.

Figure 4

Raw and integrated diffraction profiles collected at different times relative to rear surface breakout ($t=0$). At $t=-2.0\mathrm{ns}$ [(a) and (c)(i)] both the high-pressure Bi-V (110) peak and the Bi-I (012) peak from the as-yet uncompressed sample are observed. At $t=+3.0\mathrm{ns}$ [(b) and (c)(ii)], the profiles are dominated by a broad liquid diffraction peak from the melted region of the sample. The liquid diffraction signal is still observed at $t=+6.5\mathrm{ns}$ [profile (c)(iii)] at which time the Bi-V (110) peak has disappeared. The origin of the peak marked with an asterisk is discussed in Ref. [28]. A diffraction profile from liquid Bi obtained in a diamond anvil cell study at 2 GPa and 500 K [29, 30] is shown in profile (c)(iv) for comparison.