Transverse displacement of the electron beam at input to the wiggler induces betatron oscillation, imparting a sinusoidal transverse displacement to the gain medium seen by the optical beam. As the light propagates down this ``rippled'' medium it is guided by the dispersion, thereby acquiring some transverse sinusoidal displacement. At output from the wiggler the optical beam in general has acquired some transverse component of velocity, that is, it propagates off axis. As the e-beam displacement fluctuates statistically with every pulse, the optical output direction jitters around the nominal axis. The jitter is estimated with a model that includes off-axis optical propagation through a rippled gain medium but neglects feedback of the off-axis light on the e beam. The model is solved in perturbation theory. The jitter is calculated at the final aperture of the optical system beyond the wiggler. The result is (normalized to the final aperture diffraction) (rms focused jitter)/(diffraction width)${\mathit{\ensuremath{\simeq}}\mathit{\ensuremath{\Vert}}\mathit{n}}_{\mathrm{i}}$\ensuremath{\Vert}(w${\ensuremath{\lambda}}_{\mathrm{\ensuremath{\beta}}}$a${\ensuremath{\lambda}}_{\mathrm{bb}}$\ensuremath{\Delta} where ${r}_{b}$ is the e-beam radius, w is the optical beam radius (1/${e}^{2}$ intensity), \ensuremath{\lambda} is the optical wavelength, ${n}_{i}$ is the imaginary optical index on axis in e beam, ${\ensuremath{\lambda}}_{\ensuremath{\beta}}$ is the betatron wavelength, a is the rms e-beam transverse displacement, L is the wiggler length, and \ensuremath{\Delta}(L) is a damped oscillating function of L. For the proposed Paladin free-electron laser at \ensuremath{\lambda}=10 \ensuremath{\mu}m on the Advanced Test Accelerator (ATA), this implies (jitter)/(diffraction)0.25a/${r}_{b}$.
Read full abstract