1. Fabry-Perot cavity for the Compton polarimeter Goal: 10-100  J/pulse @ 5MHz repetition rate & small diameter ≈ 50  m (c.f. P. Schuler’s talks)

2. Fabry-Perot cavity: Principle (HERA cavity, cw laser) When Laser = 0  c/2L  r e sonance e beam Polar. Lin. Polar. Circ. L But :  / Laser = 10 -11 for Gain=10 4  laser/cavity feedback Done by changing the laser frequency Gain  10000

3. Some of the advantages of using a FP cavity Compact (& cheap) system compared to a laser of same power (500W in average) Laser power small outside the cavity: full power only at the electron-laser IP –no thermal effects producing parasitic birefringence & high quality frequency controlled beam  accurate control of the laser beam polarisation

4. Ti:sa oscillator 500 fs-1ps Pulse laser ≈ 5MHz / ≈10 nJ/pulse Fabry-Perot cavity with Super mirrors Electron beam Proposal: Cavity filled with a pulsed laser for a Compton polarimeter at FLC A priori impossible because the laser frequency width Dn ≈1/(1ps)=10 12 Hz for picosecond laser (c.f. 3kHz cavity banwidth) In fact possible with mode lock lasers Jones et al. Opt. Lett. 27 (2003) 1848, Jones at al. Phys. Rev. Lett. 15 (2001) 3288, Hood et al. Phys. Rev. A64 (2004)033804, Potma et al. Opt. Lett. 28 (2003)1835

5. t  t=1ps Fourier transform → superposition of N longitudinal laser mode – in phase  ~10 12 Hz=1/(1ps) ≈10 ns Mode lock laser If F.P. cavity length = laser cavity length  all modes are also resonant modes of the FP cavity Available laser pulse energy: 1-10nJ  cavity Gain ≈10 4

6. Pulse width limited by dispersion in the super-mirror coatings (Nb round trips=F/(2 p ) ≈ 5000 for F=30000  Gain ≈10000): circulating pulse gets broader and broader  power loss when overlapped to the incoming pulses (constructive interferences reduced) Cavity gain Width : 300fs-1ps for gain=10 4 R.J. Jones et al. Opt. Lett. 27 (2003) 1848

7. Reduction of the laser beam size at the IP To get a 50 m m laser beam size at the electron-laser beam IP –Use of a quasi-concentric cavity (mirror curvature radius ≈ half cavity length) –BUT, mechanical tolerance m m & m rad needed on relative mirror positions –Active feedback on relative mirror position needed (c.f. LIGO & VIRGO where nm tolerances are reached)

8. Present status of FP cavities filled with fs pulses Power amplification ≈ 120 and cavity Finesse ≈ 300 for pulse width 2-3ps ( Potma et al. Opt. Lett. 28 (2003)1835 ) Proposed R&D: –Reach a Finesse ≈ 30000 in a first step –And using a quasi-concentric FP cavity in a second step

9. Cavities in operation (for Compton polarimetry) CEBAF ( N. Falletto, NIM A459(2001)412 ): F≈24000 HERA ( upstream the HERMES experiment ): F≈30000 –Installation: 2003 summer –Laser & controllers dismounted after synch. rad. damages (huge, generated by 2 new dipoles in HERMES) –Presently: strong shielding and re-mounting –after 1 year of radiation, cavity finesse is still the same and locked again …

10. Optique input ligne ellipsometer HERA CAVITY 4 motorised miroirs bellow

11. 2003 installation shielding (3 mm pb) HERA CAVITY

12. Conclusion Proposal: a high finesse FP cavity filled with a pulse laser to produce 100 m J/pulse @5MHz –Will contribute to a high precision on the polarisation measurement This proposition make sense if the polarisation is to be measured bunch by bunch –If not, commercial laser with low rep. rate & high pulse energy do exist –But, this R&D may also be useful for other applications related to FLC (e.g. polarised positrons)

13. Laser/cavity feedback – similar to cw laser case ( Jones et al., Opt. Comm.175(2000)409 ) Stabilisation channels, e.g. MIRA (Coherent) Ti:sa oscillator –3 channels: 2 PZT mounted 2 mirrors & output coupler mounted on translation stage High frequency correction signal by an EOM if required Phase velocity & group velocity must be matched to the cavity ( both pulse-round-trip/pulse-repetition matching and frequency matching are required ) –A priori not a problem for 0.3-1ps pulse width but precise feedback techniques are known if needed

16. Aservissements