Compton Polarimeter for Qweak

Evaluation of a Fiber Laserreference laserhigh-power fiber lasercomparison

S. Kowalski, M.I.T. (chair)D. Gaskell, Jefferson LabR.T. Jones, U. ConnecticutJeff Martin, Reginahopefully more…

Hall C Polarimetry WorkshopNewport News, June 9-10, 2003

Qweak Polarimetry Working Group:

2

laser l P Emax rate <A> t (1%)option (nm) (W) (MeV) (KHz) (%) (min)

Hall A 1064 1500 23.7 480 1.03 5

UV ArF 193 32 119.8 0.8 5.42 100

UV KrF 248 65 95.4 2.2 4.27 58

Ar-Ion (IC) 514 100 48.1 10.4 2.10 51

DPSS 532 100 46.5 10.8 2.03 54

Summary of reviewed options:

3

refererence design: 100W green pulsed High-power green laser (100 W @ 532 nm)

sold by Talis Laser industrial applications frequency-doubled solid state laser pulsed design, MW peak power

D. Gaskell: news as of 10/2005 product no longer being advertised Google search: “talis laser” “talis laser” findsfinds “laser tails” “laser tails” mispelledmispelled CoherentCoherent has a device with similar properties

4

New option: fiber laser with SHG Original suggestion by Matt Poelker (4/6/2006)

source group has good experience with fiber laser capable of very short pulses (40ps), high rate (500MHz) current design delivers 2W average power might be pushed up to 60W, duty factor around 50

Published result: Optics Letters v.30 no. 1 (2005) 67. high average power: 60W average power (520 nm). demonstrated high peak power: 2.4KW (d.f. = 30) almost ideal optical properties: M2 = 1.33 polarization extinction ratio better than 95%

5

laser diodesource: cw,broadband

Optics Letters v.30 no. 1 (2005) 67.

pulse startspulse startsherehere

polarizer modulator(chopper) pumped fiber

preamplifier

fiber laser(grating mirrors)

coupling to LMAamplifier laser

main pulseamplifier

(1080 nm)

main amplifierpump laser(976 nm)

non-lineardoublingcrystal

pulse comespulse comesout hereout here

6

Optics Letters v.30 no. 1 (2005) 67.

Is there anything exotic in this design? all optics elements are

coated for 1080 nm. FOPA pump coupling

mirror has dual coating. minimum pulse peak

power for efficienct SGH in non-linear crystal

minimum pulse width to avoid SRS in fiber.

LBO crystal has a narrow LBO crystal has a narrow temperature range.temperature range.

7

Optics Letters v.30 no. 1 (2005) 67. Performance: pictures tell the story!pictures tell the story!

8

Comparison Relevant features for a Compton laser:

1. high average power (in one polarization state)2. high instantaneous power (low duty factor)3. diffraction-limited optics (M2 of order unity)

Can one gain something by matching the laser pulse structure to the machine?

1. answer depends on crossing angle2. quantitative estimate follows…

9

Comparison

average power

minimum pulse width

pulse repetition rate

duty factor range

instantaneous power

M2 factor (emittance/HUP)

minimum crossing angle

reference laser option

100 W

100 ns

300 – 1000 Hz

(3 - 10) 10-5

1-3 MW

~30

3°

fiber laser option

60 W

< 40 ps

10 – 500 MHz

(0.05 – 2.5) 10-2

2.4 - ? KW

1.33

0.5°

10

Comparison How is “minimum crossing angle” derived?

crossing angle is important for stable alignment.

Raleigh range + crossing Raleigh range + crossing angle angle → eff. target length→ eff. target length.

larger M2 => shorter RR

might allow conversion of raw power into an “effective power factor”“effective power factor”

expected range

11

Comparison Near-ideal emittance feature of this device is

impossible to beat with diode-pumped SHG lasers.

To exploit this requires eithereither going to very small crossing angles (~ 1 mr) oror matching the laser pulse train to the electron pulse train, or some combinationor some combination.

Advantages of fiber laser design:Advantages of fiber laser design: in-house expertise at Jefferson Labin-house expertise at Jefferson Lab potential x10 effective power increase for same average powerpotential x10 effective power increase for same average power more flexible pulsing scheme (large range in duty factor)more flexible pulsing scheme (large range in duty factor)

12

Target heating limits maximum pulse duration and duty factor

Instantaneous rate limits maximum foil thickness

This can be achieved with a 1 m foil

Nreal/Nrandom≈10 at 200 A

Rather than moving continuously, beam will dwell at certain point on target for a few s

Status: tests with “half-target” foil

13

tests by Hall C team during December 2004

measurements measurements consistent at the consistent at the ~2% level~2% level

random coincidence rates were larger than expected

– reals/randoms 10:1 at reals/randoms 10:1 at 4040AA

– mabe due to distorted mabe due to distorted edge of foiledge of foil

– runs at 40A frequently interrupted by BLM trips

Status: tests with 1m “half-target” foil

14

Status: kicker + half-foil test summary Preliminary results look promising. Source polarization jumps under nominal run conditions make it

impossible to confirm ~1% stability. Running at very high currents may be difficult – problem may

have been exacerbated by foil edge distortion. Development is ongoing.

Dave Meekins is thinking about improved foil mounting design.

Future tests should be done when Moller already tuned and has been used for some period of time so that we are confident we understand the polarimeter and polarized source properties.

The next step is to make 1% polarization measurements at 80A during G0 backward angle run.

15

Configuration Kick width Precision Max. CurrentNominal - <1% 2 A

Prototype I 20 s few % 20 A

Prototype II 10 s few % 40 A

G0 Bkwd. (2006) 3.5-4 s Required: 2%

Goal: 1% 80 A

QWeak 2 s Required: 1% Goal: 1% 180 A

Plans: kicker + half-foil Moller R&D

16

1m foil with kicker should work fine at 1A average current (instantaneous current 180A)

1% measurement will take ~30 minutes

Conservative heating calculations indicate foil depolarization will be less than 1% in the worst case under these conditions – can be checked

Compton being shaken down during this phase

Plans: operation during Qweak phase I

17

To reach 1% combined systematic and statistical error, plans are to operate both Compton and Moller polarimeters during phase II.

Duration and frequency of Moller runs can be adjusted to reach the highest precision in average P-1

Can we estimate the systematic error associated with drifts of polarization between Moller samplings?

Plans: operation during Qweak phase II

Is there a worst-case model for polarization sampling errors?

18

Moller performance during G0 (2004)

19

Plans: estimation of Moller sampling systematicsWorst-case scenario for sampling

instantaneous jumps at unpredictable times model completely specified by just two parameters

maximum effective jump rate is set by duration of a sampling measurement (higher frequencies filtered out)

unpredictability of jumps uniquely specifies the model

1. average rate of jumps2. r.m.s. systematic fluctuations in P

y

sampling

20

Plans: estimation of Moller sampling systematics

model calculation

Monte Carlo simulation

Inputs:Inputs: Pave = 0.70

Prms = 0.15 fjump =

1/10min T = 2000hr fsamp =

variable

Rule of Rule of thumb: thumb: Adjust the sample frequency until the statistical errors per sample match P.

sampling systematics only

21

Short term plans (2006) Improve beamline for Moller and Moller kicker

operation

Long term plans (2008) Install Compton polarimeter

Longer term plans (12 GeV) Upgrade Moller for 12 GeV operation

Plans: time line for Hall C beamline

Jlab view:these arenotindependent

22

Overview: Compton design criteria

measure luminosity-weighted average polarization over period of ~1 hour with statistical error of 1% under Qweak running conditions

control systematic errors at 1% level

coexist with Moller on Hall C beamline

be capable of operation at energies 1-11 GeV

fomstat ~ E2 (for same laser and current)

23

Overview: the Compton chicane

10 m

2 m

D1

D2 D3

D4

Comptondetector

Comptonrecoildetector

D

4-dipole design accommodates both gamma and recoil electron

detection nonzero beam-laser crossing angle (~1 degree)

important for controlling alignment protects mirrors from direct synchrotron radiation implies some cost in luminosity

24

Alex Bogacz (CASA) has found a way to fit the chicane into the existing Hall C beamline. independent focusing at Compton and target last quad triplet moved 7.4 m downstream two new quads added, one upstream of Moller and one

between Moller arms fast raster moves closer to target, distance 12 m. beamline diagnostic elements also have to move

Focus with x y= 8m near center of chicane

Overview: the Compton chicane

25

Overview: the Compton chicane

26

Overview: the Compton chicane

27

3 configurations support energies up to 11 GeV

Beam energy bend B D xe (=520nm) (GeV) (deg) (T) (cm) (cm)

1.165 10 0.67 57 2.4 2.0 1.16 4.1 2.5 1.45 5.0 2.5 4.3 0.625 25 2.2 3.0 0.75 2.6 6.0 1.50 4.9 4.0 2.3 0.54 13 1.811.0 1.47 4.5

Overview: the Compton chicane

28

Plans: use of a crossing angle

assume a green laser = 514 nm

fix electron and laser foci at the same point

= 100 m emittance of laser scaled

by diffraction limit = M (/ 4

scales like 1/cross down to scale of beam divergence

29

Overview: Compton detectors Detect both gamma and recoil electron

two independent detectors different systematics – consistency checks

Gamma – electron coincidence– necessary for calibrating the response of gamma detector– marginally compatible with full-intensity running

Pulsed laser operation– backgrounds suppressed by duty factor of laser ( few 103 )– insensitive to essentially all types of beam background,

eg. bremsstrahlung in the chicane

30

Plans: example of pulsed-mode operation

detectorsignal

signal gate

background gate

laseroutput

* pulsed design used by Hermes, SLD

31

cannot count individual gammas because pulses overlap within a single shot

Q. How is the polarization extracted?A. By measuring the energy-weighted

asymmetry.

Consider the general weighted yield:

For a given polarization, the asymmetry in Y depends in general on the weights wi used.

i

iw Y

Plans: “counting” in pulsed mode

32

Problem can be solved analytically

wi = A(k) Solution is statistically

optimal, maybe not for systematics.

Standard counting is far from optimal

wi = 1 Energy weight is

better! wi = k

Plans: “counting” in pulsed mode

33

Define a figure-of-merit for a weighting scheme

f (ideal) f (wi=1)> f (wi=k)

514nm 2260 9070 3160

248 nm 550 2210 770

193 nm 340 1370 480

Nfp )ˆ(V

Plans: “counting” in pulsed mode

34

Systematics of energy-weighted counting measurement independent of gamma detector

gain no need for absolute calibration of gamma

detector no threshold method is now adopted by Hall-A Compton team

Electron counter can use the same technique rate per segment must be < 1/shot weighting used when combining results from

different segments

Plans: “counting” in pulsed mode

35

Status: Monte Carlo simulations

Needed to study systematics from detector misalignment detector nonlinearities beam-related backgrounds

Processes generated Compton scattering from laser synchrotron radiation in dipoles (with secondaries) bremsstrahlung from beam gas (with secondaries) standard Geant list of physical interactions

36

Monte Carlo simulationsCompton-geant: based on original Geant3 program by Pat Welch

dipole chicane

backscatter exit portgamma detector

37

Monte Carlo simulations

Example events (several events superimposed)

electron beam

Compton backscatter (and bremsstrahlung)

38

Monte Carlo simulations

39

Status: laser options

1. External locked cavity (cw) Hall A used as reference

2. High-power UV laser (pulsed) large analyzing power (10% at 180°) technology driven by industry (lithography) 65W unit now in tabletop size

3. High-power doubled solid-state laser (pulsed) 90W commercial units available

40

Status: laser configuration

two passes make up for losses in elements small crossing angle: 1° effective power from 2 passes: 100 W mirror reflectivity: >99% length of figure-8: 100 cm

laser

electron beam

monitor

41

Detector options

Photon detector Lead tungstate Lead glass BGO

Electron detector Silicon microstrip Quartz fibers

42

Summary Qweak collaboration should have two independent methods to

measure beam polarization. A Compton polarimeter would complement the Moller and

continuously monitor the average polarization. Using a pulsed laser system is feasible, and offers advantages

in terms of background rejection. Options now exist that satisfy to Qweak requirements with a

green pulsed laser, that use a simple two-pass setup. Monte Carlo studies are underway to determine tolerances on

detector performance and alignment required for 1% accuracy.

Space obtained at Jlab for a laser test area, together with Hall A. Specs of high-power laser to be submitted by 12/2005.

43

extra slides(do not show)

44

Addendum: recent progress

45

Addendum: recent progress

46

Addendum: laser choices

Properties of LPX 220i maximum power: 40 W (unstable resonator) maximum repetition rate: 200 Hz focal spot size: 100 x 300 m (unstable resonator) polarization: should be able to achieve ~90%

with a second stage “inverted unstable resonator” maximum power: 50 W repetition rate unchanged focal spot size: 100 x 150 m polarization above 90%

47

Addendum: laser choices

purchase cost for UV laser system LPX-220i (list): 175 k$ LPX-220 amplifier (list): 142 k$ control electronics: 15 k$ mirrors, ¼ wave plates, lenses: 10 k$

cost of operation (includes gas, maintenance) per hour @ full power: $35 (single)

$50 (with amplifier) continuous operation @ full power: 2000 hours

48

Initial tests with kicker and an iron wire target performed in Dec. 2003

Many useful lessons learned 25 mm wires too thick Large instantaneous rate

gave large rate of random coincidences

Duty factor too low – measurements would take too long

Status: tests with iron wire target