VOLUME 74, NUMBER 8 PH YSICAL REVIEW LETTERS 20 FEBRUARY 1995 Demonstration of High Gain in a Recombination XUV Laser at 18. 2 nm Driven by a 20 J, 2 ps Glass Laser J. Zhang, ' M. H. Key i, 2 P. A. Norreys, G. J. Tallents, A. Behjat, C. Danson, A. Demir, L. Dwivedi, M. Holden, P. B. Holden, C. L. S. Lewis, A. G. MacPhee, D. Neely, G. J. Pert, 4 S. A. Ramsden, S. J. Rose, Y. F. Shao, * O. Thomas, F. Walsh, and Y. L. You~ 'Department of Atomic and Laser Physics, University of Oxford, Oxford, OX1 3PU, United Kingdom 'Central Laser Facility, Rutherford Appleton Laboratory, Chilton, OX12 OQX, United Kingdom 'Department of Physics, University of Essex, Colchester, CO4 3SQ, United Kingdom Department of Physics, University of York, York, YO1 5DD, United Kingdom 'Department of Pure and Applied Physics, Queen's University, Belfast, BT7 12VN, United Kingdom Department of Physics and Space Science, University of Birmingham, Birmingham, B15 2TT, United Kingdom (Received 13 June 1994) An exceptionally high gain coefficient has been obtained at 18. 2 nm in a C VI recombination laser driven by a 2 ps, 20 J, Nd-glass laser operating by chirped pulse amplification. Carbon fiber targets of 7 p, m diameter and up to 5 mm length were irradiated at 6 X 10" Wcm . The time and space integrated gain coefficient on the 18. 2 nm Balmer a transition was measured to be 12. 5 ~ 1. 5 cm Comparison with numerical modeling suggests that saturated laser action would be obtained with less than a factor of 2 increase in length. PACS numbers: 42. 60. Lh, 32. 70. — n, 42. 55.Vc, 52.50. Jm A major objective in the development of extreme ultraviolet (XUV) lasers is to reduce the driver energy significantly below the kilojoule level currently needed for saturated laser action in collisionally excited lasers [1, 2]. Adiabatically cooled recombination lasers have long been promising in this respect [3 — 7] but have so far suffered from a reduction of apparent gain coefficient with increase in plasma length and a consequent limit on the gain-length product to a value much lower than is required to achieve saturated laser action. One possible solution to this problem is to produce saturated output in a smaller length of plasma by increas- ing the gain coefficient. We have shown by numerical modeling that this could be achieved for the C vI Balmer n transition at 18. 2 nm in less than 1 cm of plasma us- ing a driver laser of picosecond pulse duration, which is much shorter than in earlier work [8]. We report here ex- perimental results of the first observation of high gain in a C VI recombination XUV laser driven by 2 ps, 20 J laser pulses at A = 1. 053 p, m generated by chirped pulse am- plification in a Nd-glass laser [9]. The experimental system had its final compression grating located inside the target chamber under vacuum and an f/4 off-axis parabolic mirror rellected the beam to a spot focus which was imaged by an f /4 off- axis spherical mirror to produce a line focus 7 mm long and 20 p, m wide. The pulse length monitored by a single-shot second-harmonic autocorrelator was 2 ~ 0. 6 ps. The pulse to background contrast ratio, measured by a single-shot third-order autocorrelation technique, was between 10 and 10 [9]. The spatially averaged incident irradiance in the line focus was (6 ~ 1) X 10'5 Wcm for the data discussed here. There was a fall in focused intensity along the line focus toward the output end of the plasma due to the geometry and beam intensity profile with an intensity ratio of approximately 1:0. 7. About twice the incident energy calculated to be optimum for uniform irradiation was used in the experiment to compensate for the mismatched space-time gain windows with nonuniform irradiation [10]. The targets used in the experiment were carbon fibers of 1 cm length and 7 p, m diameter supported at one end. They were positioned with better than ~ 2 p, m spatial accuracy and ~l mrad angular accuracy using a split-field microscope system [5, 7]. The free end of the target was placed well within the line focus to avoid creating a cold output end in the plasma, and the irradiated length was varied by moving the line focus axially along the fiber which was always at the same location. The primary diagnostic along the fiber axis was a Oat-field grazing incidence XUV spectrometer with a 1200 line/mm aperiodically ruled grating. It recorded the spectral range from 5. 0 nm to 30. 0 nm on Kodak 101-04 film. Spatial resolution in the spectra was pro- vided by two cylindrical mirrors which imaged the fiber end onto the detector plane of the spectrometer giving -40 p, rn resolution along the direction of the incident laser beam at 2X magnification [11]. Results presented here are effectively spatially integrated, however, as the images produced had a width of the zone of emission no larger than 40 p, m. The spectrometer was replaced for some measurements by a normal-incidence concave x- ray multilayer mirror rejecting at 18. 2 nm in a bandwidth of 2.0 nm. It was of 250 mm focal length and located 320 mm from the output end of the lasing plasma produc- ing an image at a magnification of 3. 6x. The detector was an XUV sensitive phosphor coupled to a charged-coupled 0031-9007/95/74(8)/1335(4)$06. 00 1995 The American Physical Society 1335