Femtosecond light pulses
for single bunch injection
Currently SHR contains 1812
Stochastic cooling run requires
“single bunch injection”
South Hall Ring
Laser, linac RF and Ring RF must be phase-locked. Laser
frequency must be stable to 10-100 Hz (circumference tolerance)
Our bunches in the linac are 20 ps wide, why do we need
Fiber laser operate at λ=1030 nm or λ=1550 nm .
Neither wavelength is short enough for GaAs cathode.
Frequency doubling is required.
Efficiency of conversion is proportional to light intensity,
therefore we need very tightly focused and very short pulses
Ring laser T=(Lair+1.5*Lfiber)/c
diode L λ/2 M
λ/4 Grating pair
(~60 cm) λ/4
Laser frequency stabilization
f=37.6 MHz ± 30 Hz ⇒ T=26.6 ns ± 21 fs
21 fs corresponds to 7 µm of free space or 4.5 µm of fiber
The length of the cavity must be stable within this tolerance.
Silica has very low thermal expansion (~ 10 −6 per °C), yet 50 cm
of fiber gives .5 µm per °C
Thermal expansion of the breadboard may change the length of
the air fraction of the cavity.
Thermal drifts will be compensated with a remotely controlled
mirror, active feedback is possible if needed.
The necessity of fast feedback (piezo controlled mirror holder) to
compensate fast frequency changes (vibrations) will be evaluated.
Parasitic laser modes
Time 37.6 MHz f
Time 37.6 f
Parasitic modes greatly reduce the amplitude of the main mode.
Solution: run the laser close to lasing threshold and use amplifier
after the laser.
Laser compressor Modulator Amplifier
Amplifier works in saturation regime, amplitude of the output is
almost independent of the input amplitude
Very short, compressed pulses break down in the amplifier (very
high intensity, non-linear effects)
Drawback: high losses in the grating pair: 40-70%.
Note: output intensity from the frequency doubling crystal:
I blue = I i −red ⋅ ε conversion = α ⋅ I i2−red
Amplifier & frequency doubling
diodes PULSE COMPRESSOR
From λ/4 M
λ/2 L WDM M
T LBO crystal
Amplifier: new configuration
Pump Each diode (at 980 nm) ~ 350 mW
diodes After combiner ~ 650 mW
After WDM ~ 620 mW
Combiner Output unseeded (1030nm) ~ 340 mW
Output seeded ~ 425 mW
After grating pair ~ 230 mW
WDM After doubling (515 nm) ~ 40 mW
In 40 mW at f≈40 MHz gives about 1 nJ per
Out pulse. Significant increase is expected
with lower seeding frequency.
In a single bunch mode and cathode QE~1%, 1nJ gives
Q(C) ≈ λ(nm) ⋅ E (J ) ⋅ QE(%) ⋅ 8 ⋅ 10 −6 ≈ 4.1 ⋅ 10 −12 C = 2.6 ⋅ 10 −7 electrons