<title>X-ray autocorrelator</title>

Abstract

The preliminary measurement of a 100 eV ultrashort soft x-ray pulsewidth using cross correlation is presented based on the principle that ponderomotive potentials of high intensity visible light shift x-ray absorption lines at femtosecond speeds. We are developing a high resolution technique for the direct measurement of pulse durations of soft x-ray pulses that are emitted from laser produced plasmas driven by subpicosecond lasers. This technique relies on the interaction of a high intensity ultrashort pulse visible laser with the bound levels of an atom through the high intensity A.C. Stark effect. In the presence of the visible laser, all atomic levels undergo some shift in energy1 . Specifically, the final state will shift significantly for a bound-bound inner shell transition in a noble gas in the soft x-ry regime (in which an inner shell electron is excited to an autoionizing Rydberg state). Thus, for a narrow spectral region centered on the unperturbed inner shell transition energy, the absorption of radiation may be 'switched off' when the visible laser is present. This switching of the absorption characteristics forms the basis for our pulsewidth measurement. If the timing between the x-ray pulse and the visible laser pulse is varied, the switch will map out the temporal evolution of the x-ray pulse with a resolution equal to the visible laser pulsewidth. Our experiment uses a high density pulsed krypton gas source as the abosrbing medium, and montior the 3d-5p inner shell transition at 91.2 eV with a 1.5 m grazing incidence monochromator. The plasma is produced on a solid gold plated copper rod by 3 mJ of 308 nm light with a pulsewidth of about 300 fs. The high intensity visible laser has a peak intensity of about iO' W/cm2 in a 100 fs pulse at 616 nm. By delaying the visible laser with respect to the x-ray pulse, we have made a very preliminary measurement of about 1.5 ps for the duration of the x-ray burst. The time to accumulate the data in this experiment was on the order of 2 hours with poor signal to noise. Although the data indicate a change in the absorption due to the visible laser, it is not good enough to make a conclusive measurement. We are currently in the process of redesigning the experiment to incorporate a multilayer coated focussing optic to enhance the x-ray flux by a factor of 1000. This should enable us to obtain good signal to noise and establish this method as a valuable diagnostic for laser produced plasmas.