It has been observed that, when a Free Electron Laser oscillates at high power, a set of sidebands appears located near the laser carrier wavelength. These sidebands result from an instability, and have been examined theoretically by analytic theory as well as numerical simulations in the past. The cause of the instability has to do with the slippage of the faster light wave in front of the more slowly moving bunched electrons in the FEL, which provides a feedback mechanism for the growth of parasitic waves on the carrier. The appearance of a complicated spectrum suggests that narrow temporal pulses might be produced, and indeed numerical work suggests that the FEL micropulse (~ 5 ps long) can break up into a series of regularly spaced or chaotic sha rp, narrow pulses ("spikes") at high intensity. We have proposed to study sideband and spiking radiation as they may occur in the visible FEL facility at the ATF in Brookhaven.

The motivation to study such spiking is as follows. One can show that the location of the sideband is about 1.5% away from the carrier; thus the spectrum is about 10E13 Hz wide, and pulses as narrow as 100 fs could result by Fourier transform. Such high intensity narrow pulses could be useful if they are regularly spaced. On the other hand, other applications might require the suppression of spiking. Therefore, it is useful to determine if such spiking appears in the ATF FEL, and what mechanisms or conditions affect the phenomenon.

The FEL should produce a high intensity carrier at about 500 nm wavelength at a level > 100 MW/cm2 inside the optical resonator. At thus level, simple estimates from analytic theory show that sidebands -- spaced about 1.5% from the carrier -- should grow at a rate somewhat smaller than the carrier itself. Thus, a few passes after the carrier reaches saturation level, the sidebands should appear. The first part of the experimental program consists in observing sideband radiation with a spectrometer. This instrument can also be used to study how the sideband develops throughout the FEL macropulse. In addition, we have a 1D numerical program which has computed the sideband growth spectrum under conditions appropriate to the ATF FEL.

If sidebands are found, there is motivation to study the time-dependent structure of the radiation pulses, especially the spikes. At first, a streak camera, with a resolution of 1 ps, will be used to measure the temporal width. We also plan to combine streak camera with spectrometer to study the frequency content across the optical micropulse. For the observation of the optical spikes which can be as short as 100 fs, our streak camera will not be sensitive enough to observe it. We plan to build an optical autocorrelator to observe the spikes. The study of the correlator construction is currently underway. Once spiking is detected, operating parameters of the FEL can be changed in order to understand what physics favors regular spiking. For information please contact : Thomas Marshall (tcm2@columbia.edu)