X-ray laser related investigations

Stig Borgström^†, Ulf Litzén^*, Anders Persson, Tomas Starczewski, Jürgen Steingruber, Sune Svanberg and Claes-Göran Wahlström

^†deceased

^*Atomic Spectroscopy Division

We have continued to investigate different routes to compact, table-top, X-ray lasers. During the past two years, we have concentrated on a number of optical-field ionisation schemes in gas targets. We have also pursued a number of issues of more basic interest to X-ray laser research.

Fig. B1. Axial oxygen spectrum. Geometry of the nozzle as seen from below (inset).

In one study, we have investigated the possibility of obtaining population inversion with respect to the ground states in O⁺ and O²⁺, in an effort to reproduce results published by Chichkov et al. [Phys. Rev. A 52, 1629 (1995)]. The terawatt laser was focused into a pulsed oxygen jet. The pulsed gas nozzle consisted of an arrangement of tubes, as illustrated in Fig. B1. By moving the nozzle perpendicularly to the laser beam, the length of the laser-produced plasma could be varied.

Fig. B2. Axial and transverse oxygen spectra at different plasma lengths.

Fig. B1 shows a spectrum from 29 to 66 nm. The gain line candidate (O²⁺ line at 37.4 nm) is clearly seen to dominate the spectrum. To investigate the question of gain, we observed the plasma simultaneously on-axis and perpendicular to that direction (transversely) for different plasma lengths and gas pressures. The results show the same nonlinear enhancement with plasma length of the lines in both directions of observation (Fig. B2). This indicates that we are observing a volume effect rather than gain [26].

Another scheme, demonstrated by Lemoff et al. [Phys. Rev. Lett. 74, 1574 (1995)], where electrons and ions are produced by tunnelling ionisation and accelerated in a circularly polarised laser field was investigated using xenon as the target gas. Comparisons were made using linearly and circularly polarised laser radiation. However, with the laser-pulse characteristics available at the time of our experiment (pulse duration 150 fs, pulse energy onto the target 100 mJ), gain could not be observed [26].

Fig. B3. Experimental setup for prepulse dependence experiments.

In order to gain a better understanding of the importance of laser prepulses in optimizing the plasma production needed to obtain lasing, we systematically studied the influence of femtosecond laser prepulses on the soft X­ray emission from solid target plasmas [B3].

Fig. B4. Spectrally and spatially integrated X­ray yields from laser-irradiated aluminum: a) versus the prepulse-to-main-pulse ratio (for different delays); b) versus the time delay between the prepulse and the main pulse (for two different prepulse intensities).

The laser beam was directed, at normal incidence, onto a target consisting of polished aluminum or vanadium disks. The laser system itself produces prepulses which are difficult to eliminate completely. Under normal laser operation at the time of the experiments, the prepulse-to-main-pulse ratio was of the order of 1 × 10^-4. The exact number and relative amplitudes of the prepulses in the train depended on the actual laser adjustment, but the most common configuration comprised three distinguishable prepulses with energies of the same order of magnitude. These prepulses were separated by 11 ns in time (the round-trip time of the regenerative amplifier) and the last of the prepulses arrived 11 ns ahead of the main pulse. These prepulses could be eliminated by inserting a saturable-absorber cell in the laser beam after pulse compression. A small fraction of the beam could also be split off at the edge and allowed to travel a shorter distance through a system of smaller mirrors, thus arriving before the main pulse, creating a so-called artificial prepulse (Fig. B3). Comparisons were made between spectra obtained with an inherent laser prepulse, with a clean laser pulse without any prepulses, and with a clean laser pulse with an artificially added prepulse.

The X-ray emission was also studied as a function of the delay time between the prepulse and the main pulse. In order to increase the reproducibility, the emission from the laser-produced plasma was recorded spectrally integrated, but spatially resolved by use of a pinhole and a back-side illuminated, X-ray-sensitive CCD chip (Fig. B4).