Damage-resistant single-pulse optics for x-ray free-electron lasers

X-ray free-electron lasers (XFELs) promise radiation of unprecedented peak brightness. This will enable unique research possibilities for condensed matter physics, materials science, and structural biology. For example, one eagerly anticipated experiment is the imaging of virtually any single biological molecule at the level of atoms.¹ The technology is, however, complex. A variety of optical components are required to use the XFEL beams. Moreover, radiation damage due to the high output fluence—i.e., energy per unit area—and short pulse duration pose significant challenges.

The damage threshold for single-pulse exposures is determined by thermal melting.² For multiple pulses, the threshold is potentially lower due to fatigue effects associated with thermomechanical stresses, chemical changes, and phase transitions, such as graphitization of diamond-like materials.³ We have been able to overcome these limits by taking advantage of the extremely short pulse duration of the FELs.^4–6 During the pulse, only a limited amount of damage occurs, and the optics still function. Material damage takes a finite amount of time to manifest itself, and only after the pulse has terminated are the optical elements destroyed.

Figure 1. Reflectivity as a function of the off-normal angle of incidence θ for varying fluence.⁴

We have demonstrated the concept of ‘disposable’ damage-resistant single-pulse optics in experiments at the first soft x-ray free-electron laser, called FLASH,⁷ located at the facility known as DESY in Hamburg, Germany. A variety of elements, such as multilayer mirrors,⁴ apertures,⁵ and nanolenses,⁵ remain intact for the duration of the 25fs FLASH pulse, even when exposed to fluences that exceed the melt damage threshold by 50 times or more. For example, multilayer mirrors still work during the pulse (see Figure 1), although they are completely destroyed afterward (see Figure 2). The multilayer consisted of alternating silicon and carbon layers and was irradiated with 32nm FLASH pulses.

Figure 2. Nomarski differential interference contrast (DIC) micrograph of the damaged multilayer surface. In this kind of microscopy, the reference beam is sheared to display the gradient of the optical paths present in the specimen.⁴

Figure 3. (Left) Scanning-electron micrograph image of a latex sphere on a silicon-nitride membrane after FLASH exposures. The dark rings correspond to depressions in the silicon-nitride. (Right) Calculated near-field diffraction pattern at the silicon-nitride surface.⁵

In another example, we used polystyrene spheres as nanolenses. At a wavelength of 32nm, the index of refraction is less than 1, and the sphere acts as a diverging lens. When deposited on a silicon-nitride membrane and exposed to the FLASH beam at low fluences, ring-shaped craters surrounding the sphere are formed, as shown in Figure 3 (left). The sphere is destroyed after the exposure to the FLASH beam. As shown in Figure 3 (right), we can match these rings with maxima in the near-field diffraction pattern of the sphere, demonstrating that the optical function is preserved during the pulse. We assume that these rings of high intensity also occur at larger FLASH fluences, but the evidence is not available owing to destruction of the entire silicon-nitride membrane. Based on these observations, we suggest that focusing, concave-shaped nanolenses can be constructed and used with FELs for single high-fluence pulses.

We have also developed predictive simulation tools that allow us to design damage-resistant single-pulse optics.⁴ We calculate the energy deposition in the sample considering the temperature and density dependency of the optical constants. We then calculate the damage dynamics of the optics using radiation hydrodynamics codes. These tools have allowed us to reproduce the experimental results obtained so far at FLASH and to design elements for the anticipated parameter space (shorter wavelengths and larger fluences).

Overcoming the damage barrier by applying short pulses has significant technological relevance. For hard x-rays, multilayer coatings can provide optics of higher numerical aperture (resulting in smaller focal spots) than uncoated mirrors. Extremely small focal spots will be required for the application of XFELs to biomolecular imaging, and to the creation and observation of extreme conditions in matter, such as exotic excited states of atoms and warm dense plasmas. To date, the smallest x-ray focal spots of 30nm diameter at 19.5keV photon energy have been achieved with thick multilayer structures, operating in a Laue geometry (i.e., the beam passes through the crystal volume).⁸ These multilayer Laue lenses should produce spot sizes below 5nm, and hence could be used to focus XFEL pulses to achieve x-ray power densities of 3×10²² W/cm², assuming anticipated XFEL light output. Since these lenses are only several hundred micrometers in diameter, they will be exposed to high incident power densities where damage will occur, but not before carrying out their function.

This work was performed under the auspices of the US Department of Energy by the University of California, Lawrence Livermore National Laboratory, under contract W-7405-Eng-48.