SLAC, Menlo Park, California, USA
Paul Scherrer Institut, Villigen, Switzerland
Coherent diffraction imaging of complex molecules, like proteins, requires a large number of hard X-ray photons, ~10+13/pulse, within a time ~10 fs or less. This is equivalent to a peak power of about one TW, much larger than that currently generated by LCLS or other proposed X-ray FELs. We study the feasibility of producing such pulses from LCLS and the proposed LCLS-II, employing a configuration beginning with a SASE amplifier, followed by a "self-seeding" crystal monochromator [1], and finishing with a long tapered undulator. Results suggest that TW-level output power at 8 keV is possible, with a total undulator length below 200 m. We use a 40 pC electron bunch charge, normalized transverse emittance of 0.2-mm-mrad, peak current of 4 kA, and electron energy about 14 GeV. We present a tapering strategy that extends the original "resonant particle" formalism by optimizing the transport lattice to maximize optical guiding and enhance net energy extraction. We also discuss the transverse and longitudinal coherence properties of the output radiation pulse. Fluctuation of such a tapered FEL is studied with realistic jitter measured at LCLS and with start-to-end simulation.
