Subnanometer-Scale Measurements of the Interaction of Ultrafast Soft X-Ray Free-Electron-Laser Pulses with Matter

At the recently built FLASH x-ray free-electron laser, we studied the reflectivity of $\mathrm{Si}/\mathrm{C}$ multilayers with fluxes up to $3\times {10}^{14}\mathrm{W}/{\mathrm{cm}}^2$. Even though the nanostructures were ultimately completely destroyed, we found that they maintained their integrity and reflectance characteristics during the 25-fs-long pulse, with no evidence for any structural changes over lengths greater than 3 Å. This experiment demonstrates that with intense ultrafast pulses, structural damage does not occur during the pulse, giving credence to the concept of diffraction imaging of single macromolecules.

Figure 1

(a) A TEM image of a cross section of the multilayer nanostructure in an undamaged region, in which the 10 layers of 27.5 nm period can be seen. (b) Experimental geometry, showing the beam path from FEL undulator through a gas cell (to selectively attenuate the pulse energy by up to a factor of 100) and transparent gas ionization detector (to measure incident pulse energy) before focusing onto the sample with an ellipsoidal mirror. The energy of the pulse reflected from the sample is detected with a photodiode with 1.5° angular acceptance. (c) and (d) TEM image of cross sections of the nanostructure at different positions within the FEL-exposed spot. The positions are as indicated in (e), which is a NDIC micrograph of the multilayer surface. In (d), the multilayer structure and 400 nm of the substrate were ablated.

Figure 2

(a) Low-fluence reflectivity as a function of the off-normal angle of incidence $\theta$ measured at FLASH. Overlaid is the data measured at the synchrotron [11] at a very low flux of less than $1.5\mu \mathrm{W}/{\mathrm{cm}}^2$ and the calculated reflectivity (solid line). For both we assumed that the incident light consisted of 90% $p$-polarized light and 10% $s$-polarized light. (b) Reflectivity as a function of the off-normal angle of incidence $\theta$ at higher fluences. In this second set of experiments, we observed an angular shift of 1° with respect to the measurements shown in (a). This could be due to a difference in wavelength of about 0.6 nm, which is consistent wavelength fluctuations at FLASH. The error in the scattering angle is about 1°, and the error in the measured reflectivity varies between less than 0.5% at the lowest fluence range to less than 0.1% at the highest fluence range. (c) Multilayer reflectivity ranging from 44° to 47° [data taken from (b)] as a function of pulse energy. A linear fit through the data is shown as a thick solid line, and the 90% confidence level for this fit as thin lines. This fit suggests that the reflectivity drop at the high-fluence end lies between 7% and 41%. Overlaid as a dashed line are the simulation results.

Figure 3

Computational results for the spatial dependence of electron and ion temperatures at different times during the pulse, assuming a pulse energy of $20\mu \mathrm{J}$, a pulse length of 25 fs, and a beam diameter of $17\mu \mathrm{m}$.