Tremaine 2002 0494
Tremaine 2002 0494
Tremaine 2002 0494
Bragg
Diffraction
have been pursued in recent years using time-resolved x- Detector
they probe structure on the atomic length scale, their time Fig. 1 Schematic of laser-x-ray pump-probe experiments
resolution (~ 100 ps) is not well-matched to the natural electrons at the LLNL 100 MeV electron linac. The
dynamics of elemental processes in solids, such as the scattered laser photons are relativistically up shifted in
time scale for atomic motion (~10-50 fs). Laser melting in frequency into the hard x-ray range, and are emitted in a
semiconductors may involve excitation, electron-phonon narrow cone about the electron beam direction.
coupling, melt front motion and shock waves, all with < As viewed in the frame of the moving electrons, the
30 ps relevant timescales. X-rays produced from hot incident laser pulse train appears as an electromagnetic
plasma sources can achieve very short pulse lengths, but undulator of wavelength λu=λL/γ(1–cosΨ), where
the low peak brightness forces multiple shot
γ=E/moc2 and Ψ is the incident angle between electron
measurements and may lead to target destruction by the
and laser beams as shown in Figure 2. The electrons
pump excitation.
radiate photons, which are up shifted back into the
We have undertaken the development of an x-ray
laboratory frame by a second factor of 2γ. The x-ray
source based on laser scattering off a relativistic electron
wavelength is therefore related to the initial laser
bunch. This Thomson scattering approach offers the
wavelength by
potential to produce x-rays in a unique regime. The LLNL
source will exceed the brilliances currently offered at third λx=λL/2γ2(1-cosΨ) (1)
generation synchrotron light sources while delivering
638
Proceedings of LINAC2002, Gyeongju, Korea
where λL is the laser wavelength. In the laboratory frame, photoinjector, shown schematically in Figure 3 below, has
the up shifted x-rays are confined to a narrow cone with been designed, constructed, and characterized.
opening angle ~1/γ, and their energy varies with A pulse of S-band (2.8545 GHz) RF input with 7 MW
observation angle in the laboratory due to the kinematics peak power and 3 µs pulse length produces a peak
of the Lorentz transformation. The laboratory x-ray standing wave electric field of 100 MV/m that accelerates
energy (for scattering of the fundamental) is the electrons to 5 MeV in less than a 10 cm distance. A
laser pulse, split from the Falcon laser oscillator is
Ex=EL 2γ2 (1-cosΨ)/(1 + γ2θ2 + ao2) (2) amplified and frequency tripled before striking a copper
where EL is the laser photon energy, θ is the observation photocathode near the peak RF field. The electrons
angle and ao is the usual normalized vector potential of the bunches produced have 1-10 nC of charge with pulse
laser field, which is analogous to the K parameter of a lengths of about 10 ps. Focusing solenoids are employed
static field undulator. to preserve the transverse emittance [11] of the electron
The LBNL Advanced Light Source injector linac beam immediately off the cathode and to help match the
previously demonstrated generation of sub-ps pulses of electron beam into 4 slac type linac sections, which
hard x-rays by Thomson scattering [4,5]. The LLNL increase the beam energy to 40-100 MeV.
Thomson source presently under development is expected Initial production of Thomson x-rays at LLNL utilized
to achieve an x-ray beam flux some four to five orders of 20 mJ laser pulses and 5 MeV electron bunches from the
magnitude larger, enabling the accumulation of sufficient photoinjector at 10 Hz. Successful overlap of the two
data for a high quality Bragg diffraction spectrum on each beam demonstrated 0.6 keV x-ray production [12]. This
shot. result established the ability to focus the laser and electron
We expect to produce up to 107 x-ray photons in a ~100 beams to small spots and synchronize both beams. For
fs pulse, and up to 109 x-rays in a 1-10 ps pulse, by high flux x-rays, the electron beam, with an energy of 100
scattering the laser respectively either across the electron MeV, and Falcon laser will both need to be focused to
beam (Ψ= 90o) which minimizes the temporal overlap, or under 40 microns with synchronization <3 ps
in a head-on (Ψ=180o) geometry [6,7], which maximizes At high energy (100 MeV) the minimum focal spot
the interaction of the electrons and Falcon laser. To obtainable for the electron beam is emittance dominated
achieve enhancements over other Thomson sources, we for bunch charges in the nC range. Simulations
are making two straightforward but critical improvements: Microwave Power Input
a low emittance electron beam derived from a (10 MW, 3 µs)
photoinjector and a high power, 10 TW Ti:sapphire laser
with a future upgrade to 100 TW. UV Laser Input
synchronized to FALCON
Ψ
Electron Beam Output
Laser Accelerating fields:
100 MV/m peak 5MeV, 1-10 nC, 1-10 ps
pulse (to 100 MeV Linac)
639
Proceedings of LINAC2002, Gyeongju, Korea
4 10
-7 m2 1.5195e-07 8.0324e-09 [3] C. W. Siders, et al., "Detection of nonthermal melting by
m3
Chisq
4.0605e-08
1.7518e-15
4.459e-09
NA
ultrafast X-ray diffraction", Science 286, 1340-1342
R 0.99496 NA (1999).
3 10
-7
[4] R. W. Schoenlein, et al., "Femtosecond X-ray pulses at 0.4
Spot size
640