The Monte Carlo simulation framework of the ASTRO-H X-ray observatory

Transcription

1 The Monte Carlo simulation framework of the ASTRO-H X-ray observatory Masanobu Ozaki a, Masanori Ohno a, Yukikatsu Terada b, Shin Watanabe a, Tsunefumi Mizuno c, Tadayuki Takahashi a, Motohide Kokubun a, Masahiro Tsujimoto a, Noriko Y. Yamasaki a, Hirokazu Odaka a, Yoh Takei a, Takayuki Yuasa d, Akihiro Furuzawa e, Hideyuki Mori e, Hironori Matsumoto e, Takashi Okajima f, Caroline A. Kilbourne f, Hiroyasu Tajima g and Yoshitaka Ishisaki h a Institute of Space and Astronautical Science/JAXA, Yoshinodai, Chuou-ku, Sagamihara, Kanagawa , Japan; b Department of Physics, Saitama University, 255 Shimo-Ohkubo, Saitama , Japan; c Department of Physical Science, Hiroshima University, Kagamiyama, Higashi-Hiroshima, Hiroshima , Japan; d Department of Physics, University of Tokyo, Hongo, Bunkyo-ku, Tokyo , Japan; e Department of Astrophysics, Nagoya University, Furo-cho, Chikusa-ku, Nagoya , Japan; f Astrophysics Science Division, NASA/Goddard Space Flight Center, Greenbelt, MD 2771, USA, ; g Kavli Institute for Particle Astrophysics and Cosmology, Stanford University, CA 9439, USA; h Department of Physics, Tokyo Metropolitan University, Minami-Osawa 1-1, Hachiouji, Tokyo , Japan ABSTRACT We are developing an ASTRO-H data analysis framework with the Geant4-based Monte Carlo simulation core, and numerical models of the on-orbit environmental radiation and full-satellite mass structure. In addition, the framework also includes a mechanism to connect and control data processing modules that are developed independently and data communication channels among them, which has been technically proven by simulations and analysis of the Suzaku HXD, many other detectors and astrophysical issues. Keywords: ASTRO-H, Monte Carlo, Geant4, Framework 1. INTRODUCTION The ASTRO-H 1 is the 6th Japanese X-ray astronomy satellite scheduled for launch in 214, which will be operated as an observatory open to the global astronomy community. We are developing the data analysis framework with the Geant4-based Monte Carlo simulation core and numerical models of the on-orbit environmental radiation and full-satellite mass structure, to construct high SNR detectors, their high precision response matrices and background databases. The satellite will carry 4 kinds of detectors (Figure 1), an X-ray calorimeter (SXS), 2 X-ray CCDs (SXI), 3 Si strip + CdTe detectors (HXI) 4 and Compton telescopes (SGD), 5 each of which uses different photon detection technology and has different sensitivity to background radiation. In addition, the X-ray mirrors and SGD have Further author information: (Send correspondence to M. Ozaki) M. Ozaki: ozaki@astro.isas.jaxa.jp, Telephone:

2 SGD SXS SXI L = 14 m HXI Figure 1. A CG image (left) and a schematic frame drawing (right) of the ASTRO-H satellite. As the detectors and heavy components are placed all over the space craft body, secondary radiative backgrounds from them must also be taken into account when the radiation environment inside the satellite is studied. complicated focusing feature and FOV response, respectively, for their extremely light weight nature and compactness compared to their effective area and energy range. To simulate and evaluate the particle transportation and energy deposition in such satellite and sensors, as well as the simulator core and the proper models, other software components such as filtering, data fitting and downstream evaluation processes are also essential. The data analysis framework thus includes a mechanism to connect and control those software components developed independently and the data communication channel among them, which has been technically proven by the simulations and analysis of the Suzaku HXD, many other detectors and astrophysical issues. 2. PURPOSES AND REQUIREMENTS The radiation environment simulator occupies important roles in fields such as the following: (1) estimating the detector outputs due to the background radiation field, (2) generating the detector response functions and (3) evaluating the preliminary geometry using a ray-tracing method in the designing phase. The reason why the first one is necessary is that the orbital environment is filled with radiation fields, and many of the radiation particles penetrates the spacecraft structure and reach the detector devices. The particles activate the detectors, generate electrical noise and sometimes choke the instrument short-term. For a precise scientific measurement, precise estimation of their effects is essential. As the X- and gamma-ray photons are also kinds of radiation particles, the simulator can also be effectively used for the second purpose. In particular, it has been demonstrated that not only the collimator or shielding geometry but also grazing-incident X-ray mirrors can be simulated quantitatively by introducing proper physical processes and parameters to the simulator. The third field is important when we design a spacecraft consisting of multiple detectors with complex configuration. An example will be shown in section 4.1.

3 Geometry description CAD base Geant4 code Incident particle generation (based on Mizuno et al. 24) ANL++ Computing Visualization (Terada et al. 25) Physics processes, including activation (Mizuno et al. 21) Geant4 core (Ozaki et al. 26) BGD simulation Gamma-ray FOV visibility map Detector response... Figure 2. A schematic diagram of the framework structure and examples of the products. The object-oriented analysis framework ANL++ connects all the software modules and controls the execution flow. Most of the components required for the simulator have already been developed. Products are obtained according to the selected modules. For the simulation method, Monte Carlo (MC) is usually the only choice. This is because the parameter space that has to be numerically integrated for a non-mc simulation generally becomes excessive for reasonable computing resources. For the simulation, the following software components are required: Monte Carlo engine, calculation engine of the orbital propagation of a spacecraft, calculation engine of the coordinate conversion between celestial and local spacecraft variables, on-orbit radiation environment generator, spacecraft geometry generator, physical processes including secondary particle generation by radio activation, data mining or visualization and the analysis framework that controls all of the above and other possible components. A schematic relationship among them is shown in Figure 2. While the framework is listed above, it is not essential to carry out the analysis: users can aggregate the necessary libraries and data sets and build their own software by themselves, as is done for most software running on modern computers. The reasons, or merits, of introducing a framework are as follows: (1) all the software components can be designed with a common interface that is specified by the framework, and generally the components become simple, ordered and easy to understand, and (2) it becomes easy to write reusable software that would be a part of the heritage, thanks to the previous features. As the framework, we developed ANL++,6 which was used for the MC simulations with the Geant4 engine for the Suzaku instruments, especially HXD.7

4 3. HERITAGES AND NEW FEATURES In addition to the Geant4 module for ANL++, we have already developed most of the components listed in section 2. As the basic development strategy, we use the heritages of previous missions, such as Suzaku, to reduce the development cost and make the outputs as reliable as possible: a few exceptions being modules for universal geometry description and secondary particle generation, both of which, however, are supported by the Geant4 toolkit. Nevertheless, we chose to develop software to be treated as a supplement or alternative to Geant4, to facilitate development by collaborators familiar with non-geant4 tools. The major heritages are as follows: Analysis framework : we use the object-oriented analysis framework ANL++, which is sometimes pronounced or written as ANL plus. 7 This is the successor to the official Suzaku software framework ASTE ANL, which is also embedded in HEAsoft maintained by NASA/GSFC, and the usability and reliability of which has been proven by its adoption in certain experiments and astrophysical data analysis. Orbital radiation environment generator : the generator based on Mizuno et al is working, while some component involves considerable uncertainty, by a factor of 2 3. Monte Carlo engine : the core Monte Carlo engine is Geant4. Downstream analysis modules such as energy spectrum accumulation and column density map generation. In addition to these heritages and original ASTRO-H man power, we also collaborate with nearby projects as much as possible to develop tools that are commonly used among them. To ensure these collaboration progress effectively, we will introduce geometry and radiation computing tools external to (or maybe redundant to) the Geant4 functions, such as the CAD geometry and external radioactivation library. 3.1 Geometry (mass model) description As the ASTRO-H is a large scale and complex satellite, its parts shall be supplied from many constructors. Some are scientific instruments, whose supplier might be able to supply geometric information suitable for Geant4 simulation; others are, however, unfamiliar with such simulations and the geometries are expected to be supplied in CAD format such as STEP. With this in mind, we began developing a geometry importing method by converting the STEP file to GDML (Geometry Description Markup Language) supported by the Geant4 toolkit. On the other hand, traditional Geant4 geometry can be converted to GDML by Geant4 itself, and we can merge it into a CAD-based version. GDML can supply Geant4 geometry from outside the executable image: this means that the geometry update no longer requires a software update and the geometries can be maintained as independent data sets. 3.2 Radioactivation Mizuno et al. 21 developed a radioactivation module 9 that employs MGGPOD software suite. 1 They carried out a detailed validation of the simulator including this module through the comparison with literature and the CdTe sensor, one of the key components of the SGD, activation beam test data, and successfully reproduced the results of the beam test. 4.1 A demonstration of geometry tool 4. DISCUSSION As the first step of the geometry generator, we built simplified mass models of the ASTRO-H satellite body with GDML converted from STEP and an SGD with traditional Geant4 code, and merged them into a single GDML. Figure 3 shows the result. While the ASTRO-H body is very simplified, we can see that two components are seamlessly attached to each other. Note that not only the shape but also materials of the ASTRO-H body are different from the real configuration: for example, all the structure is assumed as Al while CFRP-Al honeycomb will be used in real.

5 Figure 3. An example of geometry generation from both STEP and Geant4 code. Absorption probability Zenith angle (deg) shadow by side panels 1 2 Azimuth angle (deg) 3 Figure 4. 4π-Sr. absorption probability map from the SGD center toward the outside computed for the Figure 3 geometry. The photon energy is 1 MeV. The shadow by the side panels outside the SGD is clearly seen. With this geometry, we computed some absorption probability map for incident photons at the center of the SGD part, as was done for the Suzaku HXD by Terada et al Figure 4 shows the 4π-Sr. map for 1 MeV photons. This kind of information can be used to estimate the backgrounds generated by photons coming from the side or the bottom of the detectors. Another usage of this geometry is the optimization of the detector configuration. Figure 5 shows the map of the absorption probability for 1 MeV photons toward the SGD FOV. We can see clear asymmetry in the Y direction due to the absorption by the satellite structure. The effect of the satellite structure should be much smaller in the real case because the model assumes that all the structure is made of Al. 4.2 To-do s There remain several items to be done to use the framework for the real analysis and the satellite configuration as follows: (1) a sophisticated and automated STEP to GDML conversion method should be implemented. A

6 2 15 Azimuth Absorption probability Zenith angle (deg) Figure 5. The map of the absorption probability for 1 MeV photons toward the SGD FOV computed for the Figure 3 geometry. The satellite structure obscures Y > 3 region in case of this simple geometry. STEP geometry sometimes includes very precise information such as screw threads that consume much computing resources but can be ignored in the simulation. It is preferable that they will be removed automatically. A possible solution is to embed that kind of information in the original CAD data maintained by each manufacturer. (2) A geometry selector that picks up necessary geometry components from prepared GDML files and synthesize the final geometry should be implemented for user friendly simulation. (3) A mechanism that manages the geometry component ID through all the GDML file is required. This mechanism is essential to optimize physics processes geometrical components by components or trace a simulation results through whole the geometry. 5. CONCLUSION We are developing the Monte Carlo simulation framework of the ASTRO-H by using many heritages and newly developed modules. While the CAD geometry import is a key and it is not fully implemented yet, we succeeded in synthesizing a GDML geometry from a STEP geometry and traditional Geant4 code and demonstrated that this works as a base of necessary simulations. REFERENCES [1] Takahashi T. et al., The ASTRO-H Mission, in [Space Telescopes and Instrumentation 21: Ultraviolet to Gamma Ray], Arnaud, M., Murray, S. S. and Takahashi, T., eds., Proc. SPIE 7732 (in press). [2] Mitsuda, K. et al., The high-resolution x-ray microcalorimeter spectrometer system for the SXS on ASTRO- H, in [Space Telescopes and Instrumentation 21: Ultraviolet to Gamma Ray], Arnaud, M., Murray, S. S. and Takahashi, T., eds., Proc. SPIE 7732 (in press). [3] Tsunemi, H. et al., Soft x-ray imager (SXI) onboard ASTRO-H, in [Space Telescopes and Instrumentation 21: Ultraviolet to Gamma Ray], Arnaud, M., Murray, S. S. and Takahashi, T., eds., Proc. SPIE 7732 (in press). [4] Kokubun, M. et al., Hard x-ray imager for the ASTRO-H Mission, in [Space Telescopes and Instrumentation 21: Ultraviolet to Gamma Ray], Arnaud, M., Murray, S. S. and Takahashi, T., eds., Proc. SPIE 7732 (in press). [5] Tajima, H. et al., Soft gamma-ray detector for the ASTRO-H Mission, in [Space Telescopes and Instrumentation 21: Ultraviolet to Gamma Ray], Arnaud, M., Murray, S. S. and Takahashi, T., eds., Proc. SPIE 7732 (in press).

7 [6] Ozaki, M. et al., Framework for a Geant4-Based Simulator of the Radiation Background and Detector Responses of the Space X-Ray Observatory Suzaku (Astro-E2), IEEE Transactions on Nuclear Science 53, (26). [7] Terada, Y. et al., Development of a Monte Carlo Simulator for the Astro-E2 Hard X-ray Detector (HXD- II), IEEE Transactions on Nuclear Science 52, (25). [8] Mizuno, T. et al., Cosmic-Ray Background Flux Model Based on a Gamma-Ray Large Area Space Telescope Balloon Flight Engineering Model, Astrophysical Journal, 614, (Oct. 24). [9] Mizuno, T. et al., Monte Carlo simulation study of in-orbit background for the soft gamma-ray detector onboard ASTRO-H, in [Space Telescopes and Instrumentation 21: Ultraviolet to Gamma Ray], Arnaud, M., Murray, S. S. and Takahashi, T., eds., Proc. SPIE 7732 (in press). [1] Weidenspointner, G. et al., MGGPOD: a Monte Carlo Suite for Modeling Instrumental Line and Continuum Backgrounds in Gamma-Ray Astronomy, Astrophysical Journal Supplement Series 156, (25).