Fig. 1. Schematic of experimental setup. He_α emission was generated from a titanium backlighter foil by irradiating it with the PHELIX beam. A gold pinhole aperture was installed to collimate the X rays, which then illuminated the monocrystalline diamond sample. Three X-ray diagnostic instruments were fielded to investigate the material during and after the heating by the lead ion beam: an X-ray Thomson spectrometer (XRTS), an imaging plate recording the diffraction pattern (XRD), and a monitor spectrometer recording the spectrum of the X-ray source.

Fig. 2. Ion–ion structure factor S_ii(k) for a diamond lattice at different temperatures calculated by DFT-MD. The gray lines indicate Bragg reflections. Our XRD diagnostic measured diffraction from the first peak. The black histogram shows the range of scattering vector lengths k covered in our XRTS setup. The inset on the right visualizes the predicted ratio x_el/x_inel of elastic to inelastic scattering, calculated with Eq. (3) for this k distribution. The simulation suggests a linear increase in elastic scattering with a ratio ∼10 times higher for 3000 K compared with ambient conditions.

Fig. 3. Posterior predictions modeling the spectrally resolved XRTS signal (black) as the sum of an elastic (orange) and inelastic (yellow) contribution. The latter two curves were obtained by simulating the experiment using our ray tracing code with input from DFT-MD in the case of the incoherent curve. The full model (blue) additionally assumed a Gaussian noise with constant unknown magnitude. To guide the eye, smoothed curves of the noisy data and full predictions are plotted as solid lines.

Fig. 4. Comparison of the temperatures obtained from the ratio of elastic to inelastic scattering in the XRTS spectrum (blue) with those from SRIM²⁹ simulations of energy deposited within the sample corrected for slight shot-to-shot variations in the ion number (orange). While both methods suggest comparable temperatures, XRTS measures a slightly colder diamond, on average. However, later delays coincide with a change in the diffraction pattern, described in Sec. IV. The orange points neglect effects of cooling and heat transfer within the diamond. This simplification was justified by thermal simulations including conductivity with the ANSYS software package for the maximal ion numbers achieved in our experiments, suggesting that these effects have minor implications on the timescales examined. The ion flux as measured by a fast current transformer for a representative event is shown by the gray area, where the axis is given by the solid horizontal line.

Fig. 5. (a) XRD signals recorded on IPs at different X-ray probing times Δt. Each diffraction pattern shows three rings, caused by the He_α main, intercombination, and lithium-like lines from the Ti backlighter. Because the source was significantly extended on the sample, the Laue spots of the monocrystalline material are smeared into circular segments. The measured two-dimensional data have been offset horizontally to align on these circles (denoted by dotted lines to guide the eye). The intensity of individual plots has been corrected for the source intensity and the decay of the IP signal. For the longest delay of Δt = 20 000 ns, only a faint signal was detected, and the intensity has been scaled up by a factor of four for better readability. For Δt = 1500 and 5000 ns, the signal changes qualitatively: the pattern no longer shows three concentric circles, but the formerly continuous rings seem to break apart. We can obtain similar images with our ray tracing code when splitting our sample in two parts along the ion beam axis and rotating the upper half by 5°: see the orange plot in (b).

Fig. 6. Sketch visualizing the selection of crystallite normals when simulating mosaic crystals in our ray tracing code. All vectors that enclose the angle α with the surface normal n form the mantle of the blue cone. From this set of potential crystallite normals, we identify those vectors n_c that also satisfy the Bragg condition, i.e., we require that the angle between the incoming ray r and n_c be equal to π/2 − θ_B (the pink cone depicts all vectors for which this is the case). To account for the probability of both conditions being realized for a specific ray and given α, we reduce the weight of the reflected ray by a factor proportional to sin α.