Polarimetry at high energies

Transcription

1 34 Polarimetry at high energies Wojtek Hajdas I and Estela Suarez-Garcia II Abstract Using polarimetry as a tool to study high-energy photons was until now rather uncommon. Recent satellite measurements proved the high power of polarization observables and initiated a rapid development of new instruments. We describe the physical mechanisms leading to photon polarization, and discuss the processes used for its detection at energies from few kiloelectronvolts X-rays up to megaelectronvolts γ-rays. Past attempts and instrumentation applied for polarization measurements are summarized and reviewed with respect to experimental difficulties. We also present current and planned polarimeters and the utilization of novel, energy specific, detection technologies. Finally, we discuss the importance of proper modelling and calibration of the polarimeter, and describe several methods used to characterize it. Introduction The polarization data at X- and γ-ray photon energies remain until now very scarce. Several past attempts to measure polarization concentrated mainly on observations of the Sun, Crab nebula, and a few X-ray binaries. They resulted in a single detection of a relative polarization of 19 % from the Crab (Weisskopf et al 1978), and hardly conclusive results from solar flares (Suarez-Garcia et al 2006; Boggs et al 2006). On the other hand, the lower-energy regimes of the electromagnetic spectrum routinely benefit from polarization observations. Such disparity is largely caused by great difficulties related with both construction of the high-energy polarimeters and conducting of the measurements in space. Polarized high-energy photons can be produced by the following physical processes: Compton and inverse Compton scattering, magneto- and electrostatic bremsstrahlung, and photon splitting in extreme magnetic fields (Kallman 2004). As polarization signatures are clearly distinctive, studying them can provide essential information on both the emission mechanism and the source geometry with its size, uniformity or even structure of the magnetic field. In addition, polarization can sample the extreme physics, basic symmetries and invariants, as well as the nature of gravity (Fan 2007). Therefore it has a great potential in answering the fundamental questions I PSI Paul Scherrer Institut, Villigen, Switzerland II DPNC & ISDC, University of Geneva, Switzerland 559

2 Polarimetry at high energies on the physics of the Universe. The list of targets for polarization observation includes black holes, pulsars, X-ray binaries and flare emitters and exotic objects, such as γ-ray bursts (GRBs), soft gamma repeaters and magnetars (Lazzati 2006). Polarization measurements aim to determine two variables: the polarization degree and the polarization direction of the incoming photon flux (Lei et al 1997). Measuring principles The selection of detection techniques applied to determine the level of the linear polarization of photons is strongly related to their energies. Thomson scattering, photo-electric effect and Bragg reflection are utilized at energies of about a few kiloelectronvolts to a few tens of kiloelectronvolts. For the Bragg-reflection technique, the incoming photons undergo a constructive interference during reflection off the crystal at the glancing angle (θ B ) (see Equation 2.50 in Chapter 2, Wilhelm and Fröhlich 2010). The maximum reflectivity occurs for photons having their electric vectors parallel to the crystal planes, while it equals zero if the direction of electric vectors is normal. The Bragg-crystal method, although highly efficient as a polarization analyzer, can suffer from systematic effects and a narrow energy range strictly selected by the Bragg law. Another technique in the same energy regime is based on the photo-electric effect. It analyzes the angular distribution of electrons emitted from the atom shell after absorption of the photon. The cross section depends on the azimuth angle ξ between the photon electric vector and the direction of the electron emission. The formula for electrons ejected from the K-shell (Ramsey et al 1994) is: dσ dω sin2 θ e cosξ (1 β cosθ e ) 4, (34.1) where θ e is the electron ejection angle and β is its velocity divided by the speed of light. In order to maximize their efficiency, polarimeters based on the photoeffect utilize high-z materials as detectors and work in the energy range where the photo-absorption cross section is the highest. The Thomson-scattering technique is also utilized at energies below 30 kev. The process is equivalent to the Compton effect in the classic limit in which photon energy is much smaller than the mass of the electron. The azimuth distribution of the Compton-scattered photon depends on the initial direction of its polarization vector and is given by the Klein-Nishina cross section (Heitler 1954): dσ dω = r2 0 2 ) 2 ( ) (E E + E E E E 2 sin2 θ cos 2 η, (34.2) where r 0 is the classical electron radius, θ is the scattering angle and η is the azimuth angle between the scattering plane and the direction of the photon polarization vector, and E and E are the initial and final photon energies, respectively. The cross section in the classic limit is obtained for E = E, i.e., photon energy conservation.

3 B+A 5? = J J A H H - D G, A J A? J H - arbitrary units B-A Polarization angle azimuthal scattering angle /deg Figure 34.1: Left: Measurement of the linear polarization of γ-rays. E is the photon polarization (electric) vector. Right: Ideal modulation curve pattern from the Compton scattering of polarized photons. The most widely-used technique to measure the photon polarization at energies from tens of kiloelectronvolts up to several megaelectronvolts is the Compton scattering. The measurement principle is described in Figure Polarimeters based on the Compton effect usually select low-z materials as scatterers in order to maximize the probability of Compton scattering over the photo-electric absorption. At even higher energies of tens of megaelectronvolts the electron-positron pair production becomes a dominating process in interactions of photons with matter. The cross section for pair production with a polarized photon can be written as (Mattox et al 1990): σ(δ) = σ 0 (1 + Π R cos2δ), (34.3) 2π where σ 0 is the total cross section for pair production, δ is the angle between the electron-positron plane and the incident direction of the photon electric vector. Π is the initial polarization degree of the photons and R specifies the asymmetry ratio for the pair-creation process. To date, there are no working polarimeters based on the pair production. Their typical construction makes use of layers of heavyelement materials maximizing the probability of pair production. They are followed by position-sensitive detectors that measure the electron and positron tracks. For each of the above processes, the angular distribution of the measured observable, either photon or electron, can be described by a cosine function of the azimuth angle, C(η), (McConnell et al 2002). The modulation amplitude A and the offset B of the cosine function are linked to the polarization level of the incident photon flux. The phase of the modulation curve, η 0, corresponds to the direction of the photon polarization vector (see Figure 34.1). C(η) = Acos[2(η η 0 ) + π/2] + B. (34.4) The main experimental challenge is to measure this distribution with the highest possible precision. Polarimetric capabilities of any instrument based on the measurement techniques described above are determined by a single parameter called

4 Polarimetry at high energies modulation factor. The modulation factor is extracted from the parameters A and B from Equation 34.4 as: µ Π = A B. (34.5) It describes the response of the polarimeter to a photon flux with a polarization degree of Π. Given an instrumental modulation factor for fully polarized photons, µ 100, the degree of polarization Π of the incoming photons is equal to: Π = µ Π µ 100 (34.6) The polarization degree is defined as a positive quantity with non-gaussian statistics spanned between 0 and 1. For the source flux S f and background rate B r, the minimum detectable polarization Π mind at the significance level n σ is expressed as (Novick 1975): n σ Sf ε A p + B r Π mind =, (34.7) µ 100 S f ε A p T where A p is the detection area, ε the detection efficiency, and T the observing time. The product εa p is also known as the effective area A eff of the detector. Another quantity used to characterize an operational performance of the polarimeter is its figure of merit defined as a product of A eff and µ 100 divided by a square root of the instrument background rate. Existing polarimeters The choice of the polarimeter detection technique, the optimization of its parameters and the observation methodology depend both on the objectives of the project and on the constraints imposed by the satellite. This section illustrates the topic with several representative examples of existing polarimeters. SXRP and SPN-R The Stellar X-Ray Polarimeter SXRP (Ferocci et al 1994) was designed for the SRG mission that has not been flown yet. It is a hybrid consisting of the Bragg graphite crystal and the low-z lithium scatterer surrounded by the imaging proportional counters (Figure 34.2). The instrument covers the energy range from 2 kev to 20 kev. To reduce systematic effects the polarimeter rotates around its optical axis. An extra anticoincidence shield minimizes the background counting rate. The SXRP has very high values of µ 100 and A eff : 99 % and 10 cm 2 for the Bragg and 71 % and 65 cm 2 for the Thomson processes, respectively. The polarimeter is well understood with a thorough laboratory calibration and modelling. Its flight model is still waiting for the next flight opportunity. A similar Thomson polarimeter called SPN-R flew aboard the Coronas-F satellite. Its scatterer is made of beryllium and is surrounded by three pairs of the CsI(Na) detectors. The active area of SPN-R A eff is 1 cm 2 only. Over five years, between 2001 and 2005, the instrument observed

5 563 Figure 34.2: A diagram of the SXRP rotating polarimeter with the Bragg crystal and lithium scatterer. Figure 34.3: The RHESSI spectrometer array with nine Ge-detectors and beryllium scatterer. tens of solar flares. For only eight of them it was possible to provide just the upper polarization limits: Π 3σ < 8 % to 40 %. COMPTEL and BATSE The large area imaging Compton Telescope COMPTEL (Schönfelder et al 1993) and the Burst And Transient Source Experiment BATSE (Fishman et al 1992) were flown on board the NASA CGRO mission. COMPTEL operated at energies from 0.75 MeV to 30 MeV and consisted of seven low-z scintillators as scatterers (liquid NE213, A = 4188 cm 2 ) followed by fourteen high-z detectors (NaI(Tl), A = 8620 cm 2 ) placed 1.5 m below the scatterer plane (see also Chapter 11, Schönfelder and Kanbach 2010, for a more detailed description of the instrument). Its field of view (FOV) was equal to 1 sr but its A eff was smaller than 20 cm 2 and µ 100 below 8 %. The instrument was not optimized for polarimetry and the analysis of several GRBs revealed very low statistics and serious systematic effects. Even polarimetric observations of the Crab nebula turned out to be impossible. Slightly more successful was the other instrument BATSE. It had eight large area NaI(Tl) detectors (A = 2000 cm 2 each) placed at the spacecraft corners. It operated between 40 kev and 600 kev and had a full 4π FOV. BATSE was not a polarimeter but had good directional capabilities to localize GRBs. Polarization detection was possible using the Earth albedo scattering. After tedious analysis supported by complex Monte Carlo modelling, only two GRBs out of the about 3000 in the BATSE catalogue were analyzed. For both bursts the results gave the lower limit of the polarization: Π > 30 % and 50 % (Willis et al 2005).

6 Polarimetry at high energies RHESSI RHESSI (Lin et al 2002) is a NASA-SMEX mission for imaging the Sun at energies between 3 kev and 20 MeV. Its spectrometer, made of nine germanium detectors, is located in the plane perpendicular to the satellite rotation axis. Although not primarily designed as polarimeter, RHESSI can be used as such in two different modes: passive and active. In the passive mode (30 kev to 80 kev) one measures polarization studying the photons that scatter in the Be-block (see Figure 34.3) and stop in the Ge-detectors surrounding it (McConnell et al 2002, 2004). In the active mode (100 kev to 2 MeV), the polarization information is extracted from the photons scattered between two neighbouring Ge-detectors. As a polarimeter, RHESSI has a wide FOV, but a moderate µ 100 and a small A eff (1 cm 2 ). In 2003, a result obtained using the active mode was published claiming a very high polarization (Π = 80 %) from GRB (Coburn and Boggs 2003). Although later analysis found this result rather controversial (Rutledge and Fox 2004; Wigger et al 2004), it stimulated the theoretical and experimental astrophysics communities, and it accelerated the development of hard X-ray polarimetry. To date, RHESSI has been used to study the polarization of eight solar flares (2 % < Π < 54 % with a relative standard uncertainty from 5 % to 26 %) (Boggs et al 2006; Suarez- Garcia et al 2006). RHESSI polarimetric capabilities are limited by the high levels of Earth-scattered photons and accidental background. SPI and IBIS on INTEGRAL INTEGRAL is an ESA satellite launched in Its two main instruments, the SPI spectrometer and the IBIS imager, were not designed for polarization detection, but have such capabilities (Lei et al 1997). The spectrometer has 19 large hexagonal Ge-detectors covering energies from 20 kev to 8 MeV. Its µ 100 is moderate (20 %) but its large A eff reaches 50 cm 2. To date the SPI instrument could measure polarization of only one GRB (GRB041219a). Two groups performed independent analysis detecting very high polarization degrees of Π = 60 % (McGlynn et al 2007) and Π = 98 % (Kalemci et al 2007) with relative uncertainties of 31 % and 33 %, respectively. One should note that although the results are mutually consistent, the level of uncertainty is very large. With respect to observations of Crab (routinely used for the energy and flux calibration), calculations predict the Π mind to be on the level of a few percent for 10 6 s long measurements. The first polarization data from several weeks of Crab observations resulted, as in case of the GRB, in very high values of polarization. In the energy range from 100 kev to 1 MeV the polarization level was found to be Π = 46 % with a relative uncertainty of 10 % (Dean et al 2008). The IBIS imager has two detection layers with CdTl and CsI detector arrays. These detectors can be combined to work as a Compton telescope with A eff = 100 cm 2. Despite the IBIS large µ 100 (30 %), no polarization data have yet been published. The major problem of the INTEGRAL polarimetry is its lack of proper calibration of its modulation factors and a large uncertainty of the systematic and instrumental effects.

7 565 New polarimeters As shown, until now X- and γ-ray polarimetry has been largely neglected and attempts to measure polarization have suffered from various obstacles. The difficulties in the instrument construction and in carrying out the observations came in parallel with beliefs that the expected polarization levels should be quite low. These subjects have been recently challenged by the latest results from several instruments on satellites and by the new theoretical models. It caused a rapid development of novel polarimeters and the perfection of measurement techniques making use of the latest advances in the detection technology (McConnell and Bloser 2006). Modern gas micropattern detectors are used at low energies by PIXIE (Costa et al 2001), Mu-PIC or 1DPSPC (Nakajima et al 2004). Also classic types like Thomson and crystal polarimeters are further developed and improved like XPE and PLEXAS (Marshall et al 1998). The large scintillator arrays are core to the design in the classic Compton energy range for instruments like GRAPE (McConnell et al 1999), POGO (Larsson and Pearce 2004) and POLAR (Produit et al 2005). For polarimetry at even higher energies, either Si or Ge microstrip detectors are applied as for MEGA (Kanbach et al 2001), TIGRE (O Neill et al 1996) and NCT (Boggs et al 2001). Novel semiconductor pixel detectors (CdTl, CdZnTl) are planned for the SGD and CIPHER polarimeters (Curado da Silva et al 2003) and a time projection chamber filled with liquid xenon is adopted for the high-energy polarimetry in the LXeGRIT detector (Aprile et al 2002). For all of the above instruments, big emphasis is given to maximizing the values of A eff and µ 100 and optimizing the signal-to-noise ratio. Current developments and modern detection techniques are described below for three energy regions. For this purpose a few novel polarimeters are selected as typical examples. Energies around 10 kev PIXIE (IXO) New polarimetric techniques based on the photo-effect have recently been developed to study polarization at lower energies. An example of such an instrument is the Pixel Imaging Experiment (PIXIE) under design for the ESA IXO (formerly XEUS) mission (Costa et al 2001). PIXIE measures the tracks of electrons ejected from the atom after absorption of the photon. From the angular distributions of the tracks one can deduce photon polarization. At energies around 10 kev the length of the electron track in matter is very short. Thus, its precise visualization requires micropattern gas detectors as a photon absorption medium. The image of the ionization pattern is made using gas electron multipliers (GEMs) directly coupled to the readout electronics (ASIC, application specific integrated circuit, see Figure 34.4). In the typical operational energy range between 2 kev and 12 kev, the µ 100 reaches values of about 50 %. The total area of the GEM is slightly below 2 cm 2 while its pixel size is of the order 50 ñm. The thickness of the gas chamber is about 10 mm only. This implies that for efficient operation the detector must be mounted behind a large photon concentrator, e.g., an IXO mirror with 3 m 2 effective area. Further challenges are related with a precise modelling and advanced reconstruction techniques of the complex 3D electron tracks. Despite such obstacles

8 Polarimetry at high energies Figure 34.4: PIXIE polarimeter concept with a gas micropattern detector GEM directly connected to the readout ASIC. Figure 34.5: Concept of the POGO pointing instrument made of a well-type array of phoswich detectors. the prototype was constructed and its polarimetric potential has been successfully demonstrated. Energies around 100 kev GRAPE, POLAR and POGO Not only the energies but also the field of view that a polarimeter has to cover depends on the sources that one would like to study. For point-like objects like the Crab nebula a pointing instrument with a narrow field of view is preferable as its background levels are smaller. In the case of transient sources, like γ-ray bursts, that can appear in any position of the sky the field of view must be as large as allowed by external constraints, e.g., by the satellite or Earth shadow. When a large FOV is required at energies around 100 kev, the classical design consists of an array of low-z plastic scatterers where photons experience Compton scattering, and high-z absorbers where they undergo photopeak reaction. This is the concept of the Gamma Ray Polarimeter Experiment GRAPE (McConnell et al 1999), constituted of a array of plastic and BGO scintillators (see Figure 34.6). An alternative design is being developed for POLAR (Produit et al 2005), where all elements are low-z plastic scintillators and Compton scattering is the main effect. In this case a photon usually produces several energy depositions and the two largest ones are selected to draw the modulation curve. Despite the difference in their designs, both GRAPE and POLAR detect photons in the range 50 kev to 500 kev with high µ % to 60 % and large A eff 100 cm 2 to 400 cm 2. Both instruments have also very good off-axis performance that makes them perfect for GRB polarimetry. The demo model of GRAPE has already been tested in a balloon flight and the proposal for a new space mission dedicated to γ-ray polarimetry is under evaluation by NASA. The Polarized Gamma-ray Observer POGO (Larsson and Pearce 2004) is a polarimeter designed to study point-like sources. It is constituted of ca. 400 welltype phoswich counters with fast and slow scintillators placed on top of BGOs (see Figure 34.5). Its energy range is 25 to 200 kev and it has a small FOV Large A tot 1 m 2 and µ %, and a good signal-to-noise ratio maximize POGO s figure of merit. The instrument is suitable for long-duration

9 567 Figure 34.6: A diagram of the GRAPE polarimeter with high-z absorbing scintillators surrounded by an array of the low-z scattering ones. Figure 34.7: Si-microstrip tracker and arrays of CsI(Tl) absorbers build the MEGA instrument. The 4π anticoincidence shield reduces the chargedparticle background. balloon flights with pointing capabilities. Within 6 h it can reach a 6 % level of Π mind for 100 mcrab strong sources. A small version of POGO has been flown in a balloon, and a larger version will fly in the near future. Energy around and above 1000 kev MEGA Astrophysical γ-ray polarimetry at energies where electron-positron pair production starts to dominate is to date fully untouched. Negative analysis of data from both the COS-B and the EGRET on CGRO instruments in the past as well as a limited sensitivity of the present Fermi (GLAST) mission indicate certain difficulties. In order to succeed, the future instruments have to assure very high detection efficiency and superior tracking precision. Both parameters are intensely optimized for the Medium Energy Gamma-Ray Astronomy Experiment MEGA (Kanbach et al 2001). This instrument consists of a tracker made of 32 layers of large area (36 cm 36 cm), double-sided Si-microstrip detectors and a calorimeter made of the three-dimensional CsI(Tl)/PIN diode arrays as shown in Figure MEGA operates between 0.4 MeV and 50 MeV although measurements of polarization (at least with the Compton scattering) are feasible only up to 5 MeV. The predicted µ 100 values for the Compton mode span between 19 % at 0.7 MeV and 3 % at 5 MeV. They were confirmed with the prototype version of MEGA (scaled down to about 10 % of the full instrument volume) during calibration runs at the synchrotron. The instrument has a large field of view of 130 (full cone) and is equipped on the sides and on top with an anticoincidence shielding. Its moderate overall dimensions of 1.2 m width and 1.3 m height make it an ideal polarimeter for both a satellite mission and a long-duration balloon flight.

10 Polarimetry at high energies Modelling and calibration The modulation factor derived from calculations with the Compton cross-section formula overestimates in most cases the instrumental one. The reason for this lies in the extreme difficulty of including all polarimeter aspects into the analytical calculation. Therefore, the determination of the polarimeter modulation factor (µ 100 ) and its effective area (A eff ) must rely on both Monte Carlo simulations and laboratory calibrations. Polarization-sensitive codes of photon tracking, e.g., GEANT4 (CERN 2007), EGSnrc (NRC/SLAC 2007) or MCNP (LANL 2007), are necessary for construction of the instrument mass-model. Together with an accurate description of the geometry and materials that constitute the instrument and its surroundings (spacecraft, balloon,...), it is also important to introduce into the simulation the numerous sources of background that will affect the polarimeter in flight. The cosmic ray particles, the cosmic X-ray background, and the photons that arrive at the polarimeter either after being scattered in the materials of the spacecraft or in the Earth s atmosphere are only some of the most significant examples of background that have to be considered. More advanced studies of the signal conversion in the detector can also be performed using recent software developments. For example, with the optical photon package of GEANT4 one can generate and track optical photons through the scintillator down to the photomultiplier including their absorption as well as reflection from the walls and the coating. Further implementation of the sensitive detector volumes and particle or γ-ray hits is also possible. Independently of the simulation packages chosen, it is always necessary to verify the Monte Carlo results with proper laboratory tests and calibration procedures. These tests have to be done in a wide range of energies paying special attention to the extremes of the spectral range designed for the polarimeter. The laboratory calibration of µ 100, its angular variability as well as the effective area, detection thresholds and counting rate dependence, etc., are essential for the polarimeter s inflight success. Studies must be done with both polarized and unpolarized beams, to properly verify the zero polarization response of the instrument and discover possible systematic effects causing false polarization patterns or responses. Synchrotron radiation is commonly used to obtain fully polarized photon fluxes. For somewhat lower levels of polarization a simple laboratory option is to build a setup where γ-rays from a radioactive source (e.g., Cs 137 ) are applied to the polarimeter after being Compton scattered in some material. The Compton process will partially polarize the reflected photons, with the maximum polarization level for the scattering angles around 90. In addition, it is recommended to test the polarimeter in particle beams to study the influence of the background particles (e.g., protons from radiation belts and cosmic rays or neutrons from the nuclear reactions). Both simulations and laboratory verification help in the elimination and correction of systematic effects and spurious signals. For example sensitivity variations between individual channels of the photomultiplier can be diminished after adjustment of their discriminator thresholds. For those effects which are inherent characteristics of the detector, a specific data analysis technique must be applied. An example is the pixellation effect, inherent to any pixellated geometry of the po-

11 569 larimeter. It can be eliminated by randomizing the position of the hit inside each pixel, or by applying the Decoupled Ring Technique (DRT) from Lei et al (1997), where further descriptions of data analysis techniques useful in different situations can be found. Summary During recent years X- and γ-ray polarimetry experienced considerable progress. Innovative detection technologies, like gas tracking detectors, pixellated semiconductor or large scintillator arrays, greatly influence the design of novel polarimeters. Lessons learnt from the existing instruments brought further improvements especially in reduction of various background sources and systematic effects. Modern simulation tools and accelerator-based calibration methods are now routinely applied to enhance the performance. In addition one also observes big theoretical advances with polarization used as a probe of various physical processes and mechanisms in the Universe. A number of original polarimeters are currently under development worldwide. They will cover the whole energy range from hard X-rays up to medium-energy γ-rays. Some instruments are already approved for future missions while others are in the application phase. Long awaited, high scientific return provided by X- and γ-ray polarization measurements can be expected within the next decade. Bibliography Aprile E, Curioni A, Giboni KL (plus seven authors) (2002) The LXeGRIT Compton telescope prototype: current status and future prospects. Proc SPIE 4851: Boggs SE, Jean P, Lin RP (plus nine authors) (2001) The nuclear Compton telescope: a balloon-borne soft γ-ray spectrometer polarimeter and imager. AIP Conf Proc 587: Boggs SE, Coburn W, Kalemci E (2006) Gamma-ray polarimetry of two X-class solar flares. Astrophys J 638: Coburn W, Boggs SE (2003) Polarization of the prompt γ-ray emission from the γ-ray burst of 6 December Nature 423: Costa E, Soffitta P, Bellazzini R (plus three authors) (2001) An efficient photoelectric X-ray polarimeter for the study of black holes and neutron stars. Nature 411: Curado da Silva RM, Caroli E, Stephen JB, Siffert P (2003) Polarimeter telescope concept for hard X-ray astronomy. Experim Astron 15:45 65 Dean AJ, Clark DJ, Stephen JB (plus seven authors) (2008) Polarized gamma-ray emission from the Crab. Science 321: EGSnrc NRC webpage (2007): Fan Y-Z, Wei D-M, Xu D (2007) γ-ray burst ultraviolet/optical afterglow polarimetry as a probe of quantum gravity. MNRAS 376: Ferocci M, Costa E, Matt G (plus two authors) (1994) X-ray scattering polarimetry with scintillating fibers of different materials. Proc SPIE 2283:

12 Polarimetry at high energies Fishman GJ, Meegan CA, Wilson RB (plus two authors) (1992) The BATSE experiment on the Compton Gamma Ray Observatory: Status and some early results. NASA CP-3137:26 34 Geant4 CERN webpage (2007): Heitler W (1954) Quantum theory of radiation. 3rd edition, Oxford University Press, Oxford Kalemci T, Boggs SE, Kouveliotou C (plus two authors) (2007) Search for polarization from the prompt gamma-ray emission of GRB a with SPI on INTEGRAL. Astrophys J Suppl 169:75 82 Kallman T (2004) Astrophysical motivation for X-ray polarimetry. Adv Space Res 34: Kanbach G and the MEGA collaboration (2001) MEGA-A next generation mission in medium energy gamma-ray astronomy. AIP Conf Proc 587:887 MCNP webpage: (2007) LANL Lazzati D (2006) Polarization in the prompt emission of gamma-ray bursts and their afterglows. New J Phys 8:131 (8 pages) Larsson S, Pearce M (2004) POGO: The polarised gamma-ray observer. NIMP A 525: Lei F, Dean AJ, Hills GL (1997) Compton polarimetry in gamma-ray astronomy. Space Sci Rev 82: Lin, RP, Dennis BR, Hurford GJ (plus 63 authors) (2002) The Reuven Ramaty High-Energy Solar Spectroscopic Imager (RHESSI). Sol Phys 210:3 32 Mattox JR, Mayer-Hasselwander HA, Strong AW (1990) Analysis of the COS B data for evidence of linear polarization of VELA pulsar gamma rays. Astrophys J 363: Marshall HL, Corbet R, Takeshima T (plus six authors) (1998) RXTE observations of gamma-ray burst afterglows. Bull Am Astr Soc 30:762 McConnell ML, Ryan JM, Smith DM (plus two authors) (2002) RHESSI as a hard X-ray polarimeter. Sol Phys 210: McConnell ML, Macri JR, McClish M, Ryan JM (1999) Recent laboratory tests of a hard X-ray solar flare polarimeter. Proc SPIE 3764:70 78 McConnell ML, Smith DM, Emslie AG (plus three authors) (2004) Hard X-ray solar flare polarimetry with RHESSI. Adv Space Res 34: McConnell ML, Bloser PF (2006) Status and future prospects for γ-ray polarimetry. Chinese J Astron Astrophys Suppl 6: McGlynn S, Clark DJ, Dean AJ (plus six authors) (2007) Polarisation studies of the prompt gamma-ray emission from GRB a using the spectrometer aboard INTEGRAL. Astron Astrophys 466: Nakajima M, Mihara T, Tamagawa T (plus two authors) (2004) Proportional counter with 7 ñm carbon fiber and detection of X-ray polarization. X-ray Polarimetry Workshop, SLAC, Abstract under xray polar/reg/absdetail.asp?absid=48 Novick R (1975) Stellar and solar X-ray polarimetry. Space Sci Rev 18: O Neill TJ, Akyuez A, Bhattacharya D (plus six authors) (1996) Tracking, imaging and polarimeter properties of the TIGRE instrument. Astron Astrophys 120:

13 571 Produit N, Barao F, Deluit S (plus seven authors) (2005) POLAR, a compact detector for gamma-ray bursts photon polarization measurements. NIMP A 550: Ramsey BD, Austin RA, Decher R (1994) Instrumentation for X-ray astronomy. Space Sci Rev 69: Rutledge RE, Fox DB (2004) Re-analysis of polarization in the γ-ray flux of GRB MNRAS 350: Schönfelder V, Aarts H, Bennett K (plus 27 authors) (1993) Instrument description and performance of the Imaging Gamma-Ray Telescope COMPTEL aboard the Compton Gamma-Ray Observatory. Astrophys J Suppl 86: Schönfelder V, Kanbach G (2010) Imaging through Compton scattering and pair creation. ISSI SR-009: Suarez-Garcia E, Hajdas W, Wigger C (plus four authors) (2006) X-ray polarization of solar flares measured with RHESSI. Sol Phys 239: Weisskopf MC, Silver EH, Kestenbaum HL (plus two authors) (1978) A precision measurement of the X-ray polarization of the Crab Nebula without pulsar contamination. Astrophys J 220:L117 L121 Wigger C, Hajdas W, Arzner K (plus two authors) (2004) Gamma-ray burst polarization: Limits from RHESSI measurements. Astrophys J 613: Wilhelm K, Fröhlich C (2010) Photons from source to detector. ISSI SR-009:23 54 Willis DR, Barlow EJ, Bird AJ (plus six authors) (2005) Evidence of polarisation in the prompt gamma-ray emission from GRB and GRB Astron Astrophys 439: