Role of graphite crystal structure on the shock-induced formation of cubic and hexagonal diamond

Since cubic diamond was first recovered from explosively shocked graphite samples in 1961, the shock-induced graphite to diamond phase transformation has been of great scientific and technological interest. Recent real-time x-ray diffraction results on different types of pyrolytic graphite under shock compression have reported hexagonal diamond and cubic diamond formation at comparable stresses. To resolve and understand these differences, synchrotron x-ray diffraction measurements were used to examine, in real time, the plate impact shock response of two grades of highly oriented pyrolytic graphite and as-deposited pyrolytic graphite—at stresses below and above their respective phase transformation stresses. The present results show that at their respective transformation stresses, crystallites in as-deposited pyrolytic graphite are compressed $\sim 30\%$ more along the $c$ axis than crystallites in both highly oriented pyrolytic graphite types. This work establishes that the high-pressure phase of even ZYH-grade highly oriented pyrolytic graphite (a less oriented variety with mosaic spread $3.5^{\circ }\pm 1.5^{\circ }$), at $\sim 50$ GPa, is hexagonal diamond. In contrast, the high-pressure phase of as-deposited pyrolytic graphite (mosaic spread $\sim {45}^{\circ }$) in the present work, at $\sim 60$ GPa, is cubic diamond. Analysis of ambient x-ray diffraction data demonstrates that the crystallites in the highly oriented pyrolytic graphite samples have the hexagonal graphite crystal structure with three-dimensional long-range order. In contrast, the crystallites in the as-deposited pyrolytic graphite samples have a turbostratic carbon crystal structure which lacks rotational/translational order between parallel adjacent graphene layers. The ambient results suggest that the observed high-pressure crystal structure of shocked graphite depends strongly on the initial crystal structure—shock compression along the $c$ axis of hexagonal graphite (in highly oriented pyrolytic graphite) results in highly textured hexagonal diamond and shock compression of turbostratic carbon (in as-deposited pyrolytic graphite) results in nanograined cubic diamond. The present results reconcile previous disparate findings, establish the definitive role of the initial crystal structure, and provide a benchmark for theoretical simulations.

Figure 1

Experimental configuration for in situ plate impact XRD experiments. The snapshot shown is after the flyer plate impacts the graphite sample. Boundaries between shocked and unshocked materials are indicated with black lines and the graphite/impactor interface is indicated with a blue line. The x-ray detector is perpendicular to the direct x-ray beam.

Figure 2

Comparison of graphite crystal structures and microstructures. HOPG (left) and PG (right) are depicted from the macroscopic scale (top) down to the atomic scale (bottom). The many crystallites comprising the bulk sample are shown with the resulting mosaic spread in the top of the figure. Closeup side views of the crystal structure interlayer spacings are shown in the middle. Top views of the hexagonal graphite and turbostratic carbon crystal structures looking down at the average $c$ axes are shown at the bottom. Note that the orientations of crystallites transverse to the average $c$ axis are randomly distributed giving rise to a fiber texture.

Figure 3

Representative XRD patterns for ambient and shock-compressed HOPG and PG. Due to the fiber texture of the samples, all diffraction patterns are symmetric about the horizontal center line of each image. Bright spots which are not symmetric relative to the horizontal axis or are not present in each similar experiment are from LiF and are not analyzed. Some peaks, discussed in the main text, are labeled $hk$ or hkl denoting planes in the 2D {$hk$} or 3D {hkl} family of planes, respectively. Diffraction images show different crystal structures: hexagonal graphite (HG), turbostratic carbon (TC), hexagonal diamond (HD), and cubic diamond (CD). (a) Ambient ZYB HOPG. Asterisk denotes the integrated peak shown in Fig. 4. (b) Shocked ZYB HOPG (17 GPa, below transition). (c) Ambient PG. (d) Shocked PG (15 GPa, below transition). (e) Ambient ZYH HOPG. (f) Shocked ZYH HOPG (52 GPa, above transition). This frame, obtained ∼5 ns after the shock reaches the rear surface, is shown to highlight the new XRD peaks, though the weaker peaks observed in the previous frame were analyzed to determine the lattice spacings. Due to the finite x-ray phosphor decay time, 7% of the ambient diffraction image was subtracted from the shocked image to eliminate XRD patterns from the ambient material. (g) Ambient PG. (h) Shocked PG minus 13% of ambient diffraction image (59 GPa, above transition).

Figure 4

Representative XRD line profile and fit for a single XRD peak. The azimuthally integrated intensity vs scattering angle line profile (black) is shown along with the background line profile (dot-dashed green) and the resulting background subtracted line profile (blue) is shown below. The best fit simulated XRD line profile is shown in red. The measured and simulated line profiles correspond to the hexagonal graphite (104) peak indicated with an asterisk in Fig. 3.

Figure 5

Relative $c$-axis lattice parameters for shock-compressed HOPG and PG. The larger lattice parameter uncertainties for HOPG are due to the fact that none of the observed HOPG peaks directly provide the $c$-lattice parameter, whereas the (002) and (004) peaks directly provide the $c$-lattice parameter for PG.