Cavity losses and extraction efficiencies in a short-pulse infrared free-electron laser

Recent data taken at the FELIX facility are presented. Losses in the optical cavity have been artificially increased by the introduction of fine wires, and measurements of the resulting optical and electron energy spectra have been made over a range of cavity losses. The experimental data show that the extraction efficiency varies as the inverse square root of the cavity losses, consistent with the observation of superradiance in the free-electron laser oscillator. * Corresponding author. Tel. +44 1382 308242, Fax +44 1382 308261, email [email protected]. The electron-to-photon conversion efficiency, η, is an important parameter in a free-electron laser (FEL). Measurements of the intrinsically low efficiency of Compton FELs have been performed at a number of laboratories and have been reported previously. [1,2]. Efficiency measurements performed at the FELIX facility [3] used an electron energy spectrometer equipped with a transition radiation detector [4] and demonstrated efficiencies as high as 2.3%. During these measurements electron energy spreads regularly exceeded the acceptance of the electron spectrometer and caused background problems, and therefore a technique was developed to calculate the efficiency from a combination of the electron and optical data. The optical power envelope may be used to determine the extraction efficiency as a function of time within the macropulse, up to an arbitrary scaling factor [3]. During the onset of lasing, the energy spreads observed fall easily within the 8 % energy bite of the detector. Fitting this part of the data to the efficiency derived from the optical power allows the scaling factor to be determined. In this way the optical data may be used to determine the efficiency even in those cases where the energy spread is too large for the spectrometer to handle. It should be noted that when the energy spread falls within the acceptance of the spectrometer, consistent results are obtained for the efficiency values calculated from the optical data and those measured directly from the electron energy. In this paper we discuss the effects of cavity losses on the extraction efficiency in a FEL operating in the strong slippage regime with short electron bunches, short optical pulse lengths and relatively long radiation wavelengths. The experimental data were taken on FEL–2 at the FELIX facility [5] operating at a wavelength of 15.2 μm and with round-trip cavity losses which were varied from 4 % to 16 %. The results obtained are consistent with the recent work of Jaroszynski et al. [6,7] which describes superradiance in short-pulse free-electron lasers and is relevant to situations where the optical pulse length is considerably shorter that the electron bunch length. In our case the optical pulse length is of the order of 600 fs [8] and the electron bunch length is of the order of 3 ps. Efficiencies considerably in excess of the conventional 1 2N u estimate are produced at small cavity detunings and arise because the ultra-short optical pulses interact with electrons at a given position in a bunch for only part of the journey down the undulator before contact is lost due to slippage. Our measurements show that the efficiency η depends on the cavity losses α as η α ∝ − 1 2 and that there is also good numerical agreement between our results and the effective and measurable efficiencies as defined in [6] The electron diagnostics comprise a high-resolution time-resolved electron spectrometer employing a curved OTR radiator, and variable imaging optics set to give a dispersion of 0.22 % per channel [4]. The time resolution of the spectrometer is 50 ns. The optical power envelope is measured by a pyroelectric detector with a 100 ns response time. Cavity losses are increased by introducing fine wires into the optical cavity and measured using the ring-down time of the cavity at the end of the macropulse. Following [6] we use the expression η ρ α = l l b