<title>Recent developments in VUV femtosecond pulse generation: proposal for a higher efficiency F2 (157-nm) laser</title>

Abstract

The efficiency of the four-wave mixing (2w1 -(1)2), VUV ferntosecond source described by us last year (SPIE Vol. 1861) inherently relies UOfl a two-photon resonance enhancement occurring (in Xe gas) when 2 248.5 nrn. Ilowever, it also depends upon the spatial and temporal coherence properties of the input pulses. Development of a new source for producing beams at w with greatly improved coherence properties is here reported. Performance of the four-wave mixiig source with these 'i pulses is described. lerntosecond VUV P1lsCS with energies as high as 60 1J are generated. As a separate topic, we propose the use of a pulsed (nanosecond) infrared laser to deplete the terminal state of an ( 57-nm) laser in order to greatly enhance its performance.(I) 60-12J Femtosecond VUV Pulse (;elerafi.)n A practical method of generating femtosecond-tiine-scale pulses in the spectral range 6 -8.5 cY was discussed in Ref. I. This method is based upon four-wave, difference-frequency (w3 2 (1)1 -(1)2) mixing in Xe. It was shown in [1] that one can utilize a two-photon resonance in Xe to increase the VUV output energy by almost two orders of magnitude (over the output generated in a non-resonant four-wave mixing scheme, such as the OflC employed in Ref. 2) by choosing the beam at O)j to be a Kr1-excimer-amplified (5.0 eV) femtosecond pulse. The Xe energy levels responsihie for the resonant. enhancement in this scheme are shown in Fig. 1. Even with the use of 300-fs, 5.0-eV pulses having less-than-optimal coherence properties (as briefly discussed in Ref. 1, and as more fully described in Ref. 3), this resonantly-enhanced mixing process was utilized to generate 4-1i..J, femtosccond-time-scale, pulses at 1 56 urn, with w2 -f (1)3 energy conversion efliciencics 0.03. It was our original intent to use these seed pulses to test the ultrafast amplificatK)fl characteristics of an P2-discharge [I]. however, the attempts we subsequently made to measure single-pass gain in an l72-discharge (one capable of producing 14 mJ output when operated as a standard nanosecond-time-scale laser) gave only ambiguous results leading us to propose the