Marie Jacquet - POSIPOL 2007 Workshop - LAL 24/05/2007 Precise and fast measurement of the longitudinal polarization at HERA with a Fabry-Pérot cavity 1

laser HERA e ± polarised longitudinally spin rotators 2 e ± : 27.5 GeV / protons : 920 GeV since 1995 : LPOL and TPOL 2003: beginning of the Cavity LPOL project

Mean position of the diffused photons Principle of the polarization measurement: e   photon detect diffused  Compton diffusion e +   e +  P e transverse P e longitudinal SS Energy spectrum of the diffused photons degree of circular polarization of the laser beam 3

LPOL 33MW 100Hz pulsed laser 1000  / pulse 1% / 30 min~ 2% TPOL 10W 10 MHz CW laser 0.01  / bc ~ 3-4%1-2% / min P laser e-  crossing N  ΔP(stat) / bunchΔP(syst) Cavity 3KW 10 MHz CW laser 1  / bc per mille 1% / min 4 Technical solution to obtain Optical amplifier : Fabry-Pérot Cavity n   ~ LPOL/TPOL ratio vs day Enough large disagreements … Why a third polarimeter ?

e ± beam Polar. Lin. Polar. Circ. L ~ 2m mirror 5kW Beam analysis Fabry-Pérot cavity: principle Laser  = kHz ( precision ~ kHz ) P = 0.7 W Frequency adjustable 0.7W Fabry-Pérot Cavity Infrared Nd:YAG ( = 1064 nm ) monolithic cavity L=2m 2 spherical mirrors (solidly connected to the cavity body) e-  angle = 3.3 o 5 qwp /4

Cavity mechanical aspects Limit e ± energy losses because of wakefield along the beam pipe Keep the 2 cavity mirrors fix one to respect to the other Avoid ground and beam pipe vibrations ( ground motion, machine vibration … ) Keep mechanical deformation below 70  m ( 70  m ~ max value the feedback system is abble to correct) (monolithic cavity) (mirrors: SMA/IN2p3 Lyon) Thermic and radiation isolation of the cavity and of the optic system with a lead and aluminium housing Bellows Circular tube for laser beam, soldered to beam pipe 6

Ellipsometer to control the degree of circular polarisation inside the cavity 2 towers for the 4 flat 45 o mirrors 2 of them are motorised to align the laser beam with the cavity optical axis Entrance optical system 7 vaccum pump bellow mirror support laser entrance

8

beam pipe holes for vacuum conductance (for wake field) laser tube Inside cavity 9

Synchrotron isolation : 3 mm of lead Thermal isolation with aluminium HERMES ZEUS 10

Optics & Cavity performances 11

A 20 volts pic-pic triangular ramp is applied to the laser fast road ( 92 Mhz excursion for laser ). Start the locking procedure (Pound-Drever technic) 00 modes Merge incident and reflected beams (with CCD) and maximize the absorption reflected pics by moving motorised mirrors. time (sec) | I refl | 20 V ramp few Hertz odd 01 modes even 00 modes Alignment & locking 12 /4 cavity CCD motorized mirrors 2 glass plates laser pdiode

Very few delockings a.u. Good and stable cavity locking ~ 12 hours Cavity status 13

laser cavity laser Estimation of the cavity gain I transmitted Measurement of the cavity empty time Lock the cavity Switch off the laser and record the transmitted power Determine the extinction time of the laser : Switch on/off the laser and measure the laser power inside a p-diode 14

laser extinction time (sec) Fit ( Empty time of the cavity ) I transmitted after locking gain  ~ 9000 gain ~ ( c/ L)  unfolding   ~ 60  sec cavity decay time :  15

Power inside the cavity : P in  ~  00 X gain X P laser ~ 0.7 X 9000 X 0.7 W power reduction due to the spectral width of the laser beam cavity gain laser incident power = 4.4 kW ( 3 kW is necessary ) 16

Compton scattering process : P e at ‰ level dσ/dE  ~ σ 0 + σ 1 P e S  S   at ‰ level To control of the light polarisation inside cavity lepton beam polarisation degree of laser beam circular polarisation Why an ellipsometer at the cavity outer ? Outer of the cavity : ellipsometer 17

Ellipsometer principle SS Need a precise control of the optical elements (qwp,Wollaston, diodes) ( Determine the polarization (Ex versus Ey evolution) of the laser beam ) beam splitter φ p-diode 1 p-diode 2 p-diode 0 (reference) Wollaston laser beam transmitted by the cavity qwp 18 Measure I1/I0, I2/I0 for different value of φ allows to determine the docp S  ? extinction around Performance Wollaston : Diode response : σ  /mean < 1 ‰

e Gaussian laser beam uncoated quartz plate (anisotropic uniaxial medium) plate thickness e Quater wave plate characterzation Transmitted field depends on : e ~ 91  m ± few  m Induces a few per mille systematic on the laser circular polarisation known at the level Necessity to determine e precisely 2 optical indices n o and n e manufacturer value 19 Measure I1/I0 & I2/I0 as function of φ  qwp for different θ (qwp-laser incident angle) To determine e : Fit with a complete multiple reflexion model including also desalignment parameters

I2/I0 φφ Example of the fit : data fit I2/I0 vs φ for two different incidence angle θ Δe at the ‰ level Fitting ~ 360 X 2 X 6 points φ I1,I2 θ 20

Laser docp determination during data acq An acquisition of I1/I0 and I2/I0 vs φ thickness e previously precisely determined laser cavity to adjust the qwp ellipso docp of the  ‘s qwp docp’s determined at the per mill level 21 for a given qwp azymuthal position Then, fit with the complete model where e fixed

Synchrotron radiations First successful installation Then, problems, damages were occurred … cavity Laser controller and DAQ electronic board power suppliers damaged 2 times (every ~ 2 months) Holes in the lead sheet in front of the calorimeter electronicsacquisition summer 2003: Calorimeter crystals damage crystals NaBi( W0 4 ) 2 of one calo 22

HERMES Target Field Cavity LPOL Calorimeter 2.45 kW ! … because of enormous synchrotron radiation rate 23

Up to 3 cm lead sheet added limited by the lead weight (around the beam pipe, the cavity housing, the electronic rack and some optical devices) Radiation monitored (with dosimeters and PMs using the cavity DAQ) (  plot) Large expenses of time and manpower for reparations and actions against the radiations Construction and installation of a beam scraper between HERMES and our cavity area (  photos) Change the location of a large part of our electronic 24 Mirror quality has not been affected

qwp rotator controller laser controller mirror controller 25

PM activities, e-beam current and HERMES Target field correlations time Target field OFF no PM activity off PM activities e-beam current 26

Polarisation extraction 28

Cavity Laser 96 ns electron bunch Calorimeter about 60 m diffused  driver cable ~ 100 m Control room electron beam 1 hz alignment, locking, laser pola. 10 Mhz calo signal acqusition Electronic/DAQ system ACQ : acq / bunch ~ 10 sec Turn the laser ellipticity every 10 sec 220 histograms / 10 sec 29 Cavity control Acquisition system

From one doublet (right/left laser helicity histograms), in addition with the number per bunch of black body nR brem,nL brem, rads nR compton,nL compton pola Polarisation extraction Determine regularly (every ~100 files) the calo ( resolution + gain ) parameters using only the 3 “clean” bunches (preceded by 3 empty bunches) Using the determined calo parameters, determine the of the 180 non emtpy bunches, doublet by doublet. black body, nR brem, nL brem, rads, nR compton, nL compton, pola 30

Example of the fit quality : rads determined with precision ~ 10 Mev 31

… an other example (for one doublet right/left) : 32

Ratio cavity/tpol TPOL Cavity Higher statistical precision sec each measurement (<0.2% for all bunches error bars invisible) 33

Conclusion (designed gain and power reached) Cavity locked and stable in this difficult environment Acquisition system works well : 10 MHz (HERA bunch space time) Most of the systematic studies are under control (e beam position, laser polarization, pile up effect, detector response …) and will be finalized. Fitting procedure using the shape of the distribution allow to measure the polarization for each bunch every 20 sec with ΔP stat <5‰. 34 Ellispsometer controlled at the per mill level