development of a dedicated hard x-ray...

Development of a DedicatedHard X-Ray Polarimeter

Development of a DedicatedDevelopment of a DedicatedHard X-Ray Hard X-Ray PolarimeterPolarimeter

Mark L. McConnell, James R. Ledoux,John R. Macri, and James M. Ryan

Space Science CenterUniversity of New Hampshire

Durham, NH

AAS-HEAD Mt. Tremblant, Quebec 23-26 March 2003

2

Hard X-Ray Hard X-Ray PolarimetryPolarimetry

Gamma-Ray Bursts

ÿ In the context of fireball models of classical GRBs, polarimetry providesuseful constraints on the extent to which the flow is beamed.

ÿ Energy-dependent polarization measurements can also distinguishbetween synchrotron and inverse-Compton emissions.

ÿ Polarimetry can test the importance of photon splitting in SGRs.

Solar Flares

ÿ Polarimetry provides a means to determine the extent to which theenergetic electrons are beamed.

3

Basic Principles of Compton Basic Principles of Compton PolarimetryPolarimetry

Polarimetry relies on the fact that…

photons tend to Compton scatter at right anglesto the incident polarization vector

q

h

X

Y

Z

Eok

k'

†

ds =ro

2

2dW

¢ E Eo

Ê

Ë Á

ˆ

¯ ˜

2Eo

¢ E +

¢ E Eo

- 2sin2 q cos2 hÊ

Ë Á

ˆ

¯ ˜

¢ E =Eo

1 +Eo

mc 2 1- cosq( )

q is the Compton Scatter Angle, h is the Azimuthal Scatter Angle

incident photon

scattered photon

4

The Polarization SignatureThe Polarization SignatureFor a fixed Compton scatter angle (q), the azimuthal distribution

of scattered photons contains the polarization signature.

The amplitude of the modulation defines the level of polarization.

The scattering angle corresponding to the minimum of thedistribution defines the plane of polarization.

†

Q =Cmax - CminCmax + Cmin

=AB

2000

1500

1000

500

0

Coun

ts

350300250200150100500Azimuthal Scatter Angle (degs)

C-max

C-min

pola

rizat

ion

angl

e

†

C h( ) = A cos 2 (h-j) + B

B

AModulation Factor for a100% polarized beamrepresents a figure-of-

merit for the polarimeter

5

Minimum Detectable Polarization (MDP)Minimum Detectable Polarization (MDP)

S = source counting rate

B = background counting rate

T = observation time

Q100 = modulation factor for 100% polarization

†

MDP =n

s

Q100 S2 S + B( )

T

Sensitivity can be improved by :

1) Increasing S (efficiency or geometric area)

2) Decreasing B

3) Increasing T

4) Increasing Q100 (optimizing geometry)

6

Generating a Polarized BeamGenerating a Polarized Beam

Polarized X-rays can be generated in the lab by Compton scattering photonsfrom a g-ray calibration source within a block of plastic scintillator.

A signal from the scintillator provides an electronic tag for each scatteredphoton, which can be used as a coincidence signal with the polarimeter.

0.8

0.6

0.4

0.2

0.0

Pola

rizat

ion

Frac

tion

18016014012010080604020Scatter Angle (degs)

60 keV 662 keV 1200 keV

This graph shows the level of polarization that can be achieved for various input photon energies

(corresponding to 241Am, 137Cs and 60Co) and various scatter angles.

A 662 keV photon beam, scattered at 90°, is ~60% polarized;

the scattered energy is 288 keV.

7

The GRAPE Science ModelThe GRAPE Science Model

Gamma-RAy Polarimeter Experiment

Position-Sensitive PMTHamamatsu R3292

Plastic ElementsBC-404

CsI(Tl) Array

Plastic Scattering Elements :• 280 element array of BC-404• Each element optically isolated• Wrapped in Tyvek® and Kapton® tape • Each 5 mm square by 5 cm long • Readout provided a 5-inch PSPMT• Hamamatsu R3292 PSMT

Calorimeter Elements :• 4 element array of CsI(Tl)• Each 1cm square by 5 cm long• Readout provided by a MAPMT • Hamamatsu R5900 MAPMT

PSPMT / Plastic Array Housing :• 1mm thick aluminum• Optically isolated from CsI/MAPMT

8

GRAPE GRAPE Polarimetric Polarimetric ResponseResponse

Simulated Results for Monoenergetic On-Axis Beam at 150 keV

The simulated data for anunpolarized beam are used

to correct for geometric effectsand to extract the true modulation

pattern from the data.

1200

800

400

0

Cou

nts

350300250200150100500

Azimuthal Scatter Angle (degs)

Polarized Data1200

800

400

0

Cou

nts

350300250200150100500


Unpolarized Data

1200

800

400

0

Cou

nts

350300250200150100500Azimuthal Scatter Angle (degs)

Corrected Data

9

Simulated Performance: On-AxisSimulated Performance: On-Axis

1.0

0.8

0.6

0.4

0.2

0.0

Mod

ulat

ion

Fact

or

3 4 5 6 7 8 9100

2 3 4 5

Energy (keV)

101.6 mm 88.9 mm 76.2 mm 63.5 mm 50.8 mm 25.4 mm

Modulation Factor vs. Depth4

3

2

1

0

Effe

ctiv

e A

rea

(cm

2 )

3 4 5 6 7 8 9100

2 3 4 5

Energy (keV)

101.6 mm 88.9 mm 76.2 mm 63.5 mm 50.8 mm 25.4 mm

Effective Area vs. Depth

7

6

5

4

3

2

1

0

Figu

re o

f M

erit

(rel

ativ

e)

3 4 5 6 7 8 9100

2 3 4 5

Energy (keV)

76.2 mm 50.8 mm 25.4 mm 63.5 mm 88.9 mm 101.6 mm

Figure-of-Merit vs. Depth The figure-of-merit incorporates anestimate of the increased background

as a function of detector volume.

Optimum Depth for thisdesign is ~3 inches

†

FoM =m100 eA

Vdet

10

Simulated Performance: Off-AxisSimulated Performance: Off-Axis

1.0

0.8

0.6

0.4

0.2

0.0

Mod

ulat

ion

Fac

tor

706050403020100Angle (Degs)

3.0

2.5

2.0

1.5

1.0

0.5

0.0

Effective A

rea (cm2)

Modulation Factor Effective Area

7.62 cm depth @ 200 keV

The GRAPE design offers significant response even atlarge off-axis angles.

This makes GRAPE an attractive design for polarizationstudies of transient sources (Gamma-Ray Bursts).

11

Science Model ComponentsScience Model Components

This photo shows the 5-inch PSPMT, the array ofplastic scattering elements and the MAPMT/CsI assembly.

12

GRAPE Science Model Lab Set-UpGRAPE Science Model Lab Set-Up(drawing not to scale)

Photons are partially polarized by scatteringoff a plastic scintillator detector.

This also provides electronic tagging of eachscattered photon.

13

PSPMT Response to Plastic ArrayPSPMT Response to Plastic Array

The spatial resolution in the PSPMT is sufficient to resolve the individual 5 mm plastic scattering elements.

Distribution of 137Cs Events

This figure shows the spatialdistribution of events in the plastic

scintillator array.

These are coincident events,representing photons that havescattered from one of the plastic

elements into the central CsI array.

The individual plastic elements areclearly resolved, as is the square

well for the central CsI array.

Distortions near the edge of thePSPMT are also evident in these

raw data.

14

Results from Science ModelResults from Science Model

The incident energy spectrumis reconstructed by combiningthe energy loss in both the CsI

and in the plastic array.

This energy spectrum is that ofa 662 keV beam that has been

scattered once.

The variable scatteringgeometry results in an incidentbeam that covers a broad rangeof energies, thus accounting for

the broad nature of thisspectrum.

2500

2000

1500

1000

500

0

Coun

ts

8007006005004003002001000Energy (keV)

Polarized Source Spectrum(Nominal Energy = 288 keV)

Eo = 284.2 (±0.9) keV

∆E/E = 39%

15

Lab Results from Science ModelLab Results from Science Model

1000

800

600

400

200

0

Corre

cted

Cou

nts

360300240180120600


150100

500

-50-100

GRAPE Science Model0° Incidence

∆f = 0°

600

500

400

300

200

100

0

Corre

cted

Cou

nts

360300240180120600


-100-50

050

100

GRAPE Science Model0° Incidence

∆f = 90°

Runs taken with the incident beam having apolarization angle that differs by ~90°.

measured polarization = 61(±3)%polarization angle = 49(±1)°

measured polarization = 73(±7)%polarization angle = 135(±3)°

16

GRB SensitivityGRB Sensitivity

GRB Polarization Sensitivity50-300 keV energy band

Fluence(ergs cm-2)

MDPT = 10 sec

MDPT= 100 sec

Observed Rate(N > Fluence)

1 ¥ 10-4 2.8% 2.9% 1 every 320 days

5 ¥ 10-5 3.9% 4.4% 1 every 80 days

1 ¥ 10-5 9.3% 13.7% 1 every 10 days

5 ¥ 10-6 14.0% 24.4% 1 every 6 days

3 ¥ 10-6 19.4% 38.6% 1 every 4 days

1 ¥ 10-6 43.2% ––– 1 every 2 days

We have estimated the polarization sensitivity of a long-duration balloonpayload consisting of a planar array of 36 GRAPE modules. Such an

array would have a FoV of ~2! steradian and cover less than 1 m2 of area.

The table below gives the polarization sensitivity over the full energyrange (50-300 keV) and the expected frequency of observed events

(based on the fluence distribution of the 4B catalog).

17

Solar Flare SensitivitySolar Flare Sensitivity

The solar flare polarization sensitivity for various energy bands.These data assume a balloon-borne array of 16 GRAPE modules.

30

25

20

15

10

5

0

Min

imum

Det

ecta

ble

Pola

rizat

ion

(%)

6005004003002001000Energy (keV)

X1-Class Flare

X10-Class Flare

18

A New DesignA New Design

We are currently preparing to conduct labtests with a more compact design based on

Hamamatsu’s new flat-panel MAPMT.

This new PMT design (Hamamatsu H8500)provides an 8 x 8 array of anodes.

The anode size of 5.6 mm (on a 6 mm pitch)is an excellent match to the 5 mm plastic

elements that we are currently using.

The MAPMT can be used to read out boththe plastic scattering elements and the CsI

calorimeter elements.

This also eliminates a significant amount ofmass in front of the detector array.

19

Polarimeter Polarimeter Design Based on MAPMTDesign Based on MAPMT

Single module withMAPMT and FEE

electronics.

Array of modules suitable for use as acoded aperture imaging plane.

MAPMT

FEE

plasticplasticelementselements

CsICsI

20

Development StatusDevelopment Status

ÿ We have developed a design for a compact polarimetermodule that could be adapted to a variety of configurations.

ÿ Testing of a science model based on the design has so faryielded encouraging results.

ÿ Further testing of the science model will be completed in thecoming months.

ÿ We will soon be testing a new design based on the use a flatpanel multi-anode PMT.

ÿ Long-term goal will be to develop a dedicated balloon payload.

ÿ An array of GRAPE modules (a “bunch of GRAPEs”) willprovide sufficient sensitivity to constrain models GRBs andsolar flares.

21

Results from Science ModelResults from Science Model

1000

800

600

400

200

0

Coun

ts


Polarized Data

4000

3000

2000

1000

0

Coun

ts


Unpolarized Data

1000

800

600

400

200

0

Coun

ts


Corrected Data

Data from the science model are shown below. These are energy selecteddata from both a polarized beam (left) and an unpolarized beam (middle).

The corrected data is on the right.

Measured Polarization = 53(±3)%

Polarization Angle = 48 (±1)°

288 keV 356 keV 288 keV

22

PSPMT CalibrationPSPMT Calibration

300

250

200

150

100

Y-Lo

catio

n (a

rbitr

ary

units

)

300250200150100 X-Location (arbitrary units)

Measured LED Locations

-40

-20

0

20

40

Y-Lo

catio

n (m

m)

-40 -20 0 20 40X-Location (mm)

LED Calibration Grid

The PSPMT response was calibrated during the summer of 2001.

These data show both the calibration grid (2.5 mm spacing) and thereproduced locations. Distortions near the edge of the PSPMT are evident.

23

Compton ScatterCompton Scatter Polarimetry Polarimetry

A Compton scatter polarimeter measuresthe angular distribution of the scattered

photons in a plane which is perpendicularto the incident photon direction.

The asymmetry of this azimuthal scatterangle distribution can be exploited tomeasure the linear polarization of the

incident flux.1

10

100

Asy

mm

etry

Rat

io

1201008060Scattering Angle (degrees)

50 keV

100 keV

250 keV

500 keV

1 MeV

A Compton scatter polarimeter consists of two basic components:

1) low-Z scattering detector(s) to provide Compton scattering medium2) high-Z calorimeter detector(s) to absorb the scattered photon

development of a dedicated hard x-ray...

Documents