205 Acta Polytechnica CTU Proceedings 1(1): 205–209, 2014 205 doi: 10.14311/APP.2014.01.0205 Expected Hard X-Ray and Soft Gamma-Ray from Supernovae Keiichi Maeda1,2, Yukikatsu Terada3, Aya Bamba4 1Department of Astronomy, Kyoto University, Japan 2Kavli Institute for the Physics and Mathematics of the Universe (WPI), University of Tokyo, Japan 3Department of Physics, Saitama University, Japan 4Department of Physics and Mathematics, College of Science and Engineering, Aoyama Gakuin University, Japan Corresponding author: keiichi.maeda@kusastro.kyoto-u.ac.jp Abstract High energy emissions from supernovae (SNe), originated from newly formed radioactive species, provide direct evidence of nucleosynthesis at SN explosions. However, observational difficulties in the MeV range have so far allowed the signal detected only from the extremely nearby core-collapse SN 1987A. No solid detection has been reported for thermonuclear SNe Ia, despite the importance of the direct confirmation of the formation of 56Ni, which is believed to be a key ingredient in their nature as distance indicators. In this paper, we show that the new generation hard X-ray and soft γ-ray instruments, on board Astro-H and NuStar, are capable of detecting the signal, at least at a pace of once in a few years, opening up this new window for studying SN explosion and nucleosynthesis. Keywords: nuclear reaction - nucleosynthesis - abundances - supernovae: general. 1 Introduction Supernova (SN) explosions trigger (or are triggered by) explosive nucleosynthesis, and they are believed to be main production sites of heavy elements in the Uni- verse. The resulting yields are sensitive to explosion mechanism(s), and thus studying nucleosynthesis prod- ucts is important to uncover the still-debated explosion mechanism. Especially important is the production of 56Ni – this is the origin of Fe (as a result of the ra- dioactive decay chain 56Ni → 56Co → 56Fe), and the decay is believed to provide a source of emissions from (many classes of) SNe through thermalization of emit- ted γ-rays and positrons. In type Ia supernovae (SNe Ia), about half of an exploding white dwarf in mass is processed into 56Ni, supporting their huge luminosities as distance indicators. However, the most direct evi- dence in this scenario is still missing – there has been no solid detection of the decay γ-rays from SNe except for SN 1987A (e.g., Dotani et al., 1987; Sunyaev et al., 1987). Especially, no solid detection has been reported for SNe Ia (see Milne et al., 2004 for a review). From a theoretical point of view, studying this high energy emission has been restricted to one-dimensional models (see Milne et al., 2004, for a review) despite the importance of multi-dimensional structures of the ex- plosion both in theory and observation (e.g., Kasen et al., 2009; Maeda et al., 2010a). Most previous studies also focused on the emission in the MeV range. In this paper, we present our radiation transfer simulations of the high energy emission based on the state-of-the-art SN Ia explosion models. We extend our analysis to hard X-ray and soft γ-ray regimes, for which dramatic improvement is expected in the observational sensitivi- ties thanks to new generation observatories like NuStar (Koglin et al., 2005) or Astro-H (Takahashi et al., 2010). We predict that these telescopes are capable of detect- ing the radioactive decay signals from SNe Ia, at a rate of once in a year or at least once in a few years. We also briefly comment on perspectives for core-collapse SNe. 2 Expected Characteristics We performed radiation transfer simulations (Maeda et al. 2012) based on a series of two-dimensional delayed detonation models by Kasen et al. (2009). The de- layed detonation model is among the most popular sce- narios for SNe Ia, resulting from a near-center ignition of thermonuclear sparks within a Chandrasekhar-mass white dwarf (Khokhlov, 1991). Conditions for the ini- tial triggers have not been clarified from the first prin- ciple (e.g., Seitenzahl et al., 2013), thus Kasen et al. (2009) adopted various conditions (i.e., distribution of the sparks) and produced a range of the ejecta models. In this scenario, different initial conditions can be asso- ciated with observed diversities. Figure 1 shows exam- ples of the ejecta structure. The model DD2D asym 04 is for bright SNe Ia (resulting in ∼ 1M� of 56Ni), while 205 http://dx.doi.org/10.14311/APP.2014.01.0205 Keiichi Maeda, Yukikatsu Terada, Aya Bamba Figure 1: Examples of the delayed-detonation models (Kasen et al., 2009; see also Maeda et al., 2010b). The mass fractions of Si (left) and 56Ni (right) are shown, on a logarithmic scale. The axes are in 10, 000 km s−1. DD2D iso 04 is for fainter ones (∼ 0.4M� of 56Ni). Generally, this model sequence predicts more asym- metric structure for fainter SNe Ia (note that the ini- tial condition of model DD2D asym 04 is indeed more asymmetric than DD2D iso 04, but the post-explosion ejecta are less asymmetric). Examples of the synthetic spectra are shown in Fig- ure 2. The spectra are characterized by the decay lines, Compton scattering continuum, and the low energy cut off by the photoelectric absorption. At 20 days, the decay lines from 56Ni → 56Co (e-folding time of ∼ 8.8 days) are more important than those from 56Co → 56Fe (∼ 113 days) as characterized by strong lines in the soft γ-ray range (e.g., the 158 keV line followed in strength by the 270 and 480 keV lines). Later on, the strong lines are mostly in the MeV range (i.e., the 847 keV line as the strongest) except for the annihilation line. The cut off energy becomes higher as time goes by due to the increasing contribution to the emission from the deeper part where the mean atomic number and pho- toelectric cross sections are larger. Thus, overall the spectra evolve from soft to hard as time goes by. This indicates that follow-up of SNe at relatively early phases is important in hard X-ray and soft γ-ray range. The classical 1D model W7 has lower energy cut off than the delayed detonation model sequence, since the W7 model has less extended explosive nucleosynthesis in the surface layer than the delayed detonation model. Due to increasing transparency, the emission is sen- sitive to model variants (including the viewing direc- tion) in the earlier phase, while the mass of 56Ni plays a dominant role in the later phase. Thus, observation at relatively early phases in (relatively) soft bands can pro- vide unique diagnostics in clarifying the explosion mech- anism(s). In the multi-D delayed detonation model se- quence, a unique prediction is that faint SNe Ia should show larger variations in the high energy emission aris- ing from various viewing directions than brighter ones. Such prediction can be tested once there are at least a few samples of high energy emission detected for SNe Ia. Another diagnostics using the ‘soft’ bands includes the surface composition analysis from the photoelectric absorption, e.g., the different behavior shown for the W7 model and delayed detonation models. 3 Observational Perspectives Table 1 summarizes expected detectability of extra- galactic SNe Ia at various band passes by a few current and future instruments. While detecting the MeV lines from the 56Co decay has been challenged in the past, even with SPI on board INTEGRAL and 106 s exposure (Roques et al., 2003; see also Isern et al., 2013), this is limited to extremely nearby SNe Ia up to 5 - 6 Mpc (or 8 Mpc for extremely bright SNe Ia). Such nearby events are expected only once in a decade. This frus- trating situation in the MeV range will be improved only when the sensitivity is improved by an order of magnitude, hopefully by proposed new generation ob- servatories like GRIPS (Greiner et al., 2012; see also Summa et al., 2013). We propose that new generation hard-X and/or soft γ-ray instruments can change the situation. In hard X- rays, NuStar has been already launched. Astro-H is scheduled for launch in 2015, which will be attached with HXI (hard X-ray) and SGD (soft γ-ray). These instruments are expected to have 106 s exposure sen- sitivities sufficient to reach SNe Ia at 15 (conservative estimate) or 25 Mpc (optimistic estimate) (Tab. 1). The ‘line detection’ is more challenging, and we esti- mate the distance for the 5σ detection of the 158 keV line is ∼ 10 − 15 Mpc for Model DD2D asym 04 and ∼ 3 − 5 Mpc for the other two models shown in Table 1. Figure 3 shows simulations for expected signals from SNe Ia at 15 Mpc, by convolving the synthetic spectra 206 Expected Hard X-Ray and Soft Gamma-Ray from Supernovae 10 10020 50 200 1 0 − 8 1 0 − 7 1 0 − 6 1 0 − 5 1 0 − 4 n o rm a liz e d c o u n ts s k e V Energy (keV) DD2D_asym_04_dc2 15Mpc 10 10020 50 200 1 0 − 8 1 0 − 7 1 0 − 6 1 0 − 5 1 0 − 4 n o rm a liz e d c o u n ts s k e V Energy (keV) w7 15Mpc 10 10020 50 200 1 0 − 8 1 0 − 7 1 0 − 6 1 0 − 5 1 0 − 4 n o rm a liz e d c o u n ts s k e V Energy (keV) DD2D_iso_04_dc3 15Mpc Figure 2: Detector response simulations for an exposure of 106 seconds for selected models (Tab. 1), for HXI (black) and SGD (red) on board Astro-H. The model spectra at 20 days after the explosion are used as input, placed at distances of 15 Mpc. The sensitivity curves are adopted from Kokubun et al. (2010), Tajima et al. (2010), and Takahashi et al. (2010). Note that the photon count is very low in the HXI band for all the models at this distance, thus an apparent detection by HXI (left panel) just comes from the statistical fluctuation. Table 1: Expected Detectability (for an exposure of 106 s centered at the peak date in each band pass). Shown here are limiting distance and the expected number of SNe Ia within the distance (shown in parenthesis). ‘cons’ and ‘opt’ are conservative and optimistic estimates, respectively. See Maeda et al. (2012) for details. DD2D asym 04 W7 DD2D iso 04 M(56Ni)/M� 1.02 0.64 0.42 Band (keV) Instrument Mpc (SNe year−1) 60–80 HXI 13.9 (0.43) 17.7 (0.96) 10.5 (0.09) NuStar (cons.) 13.0 (0.43) 16.5 (0.70) 9.7 (0.09) NuStar (opt.) 18.4 (1.13) 23.3 (2.52) 13.8 (0.43) 158 SPI 4.6 (<0.09) 2.9 (<0.09) 2.3 (<0.09) SGD (cons.) 22.2 (2.09) 14.2 (0.43) 11.4 (0.09) SGD (opt.) 38.5 (6.70) 24.6 (2.96) 19.7 (1.57) 200–460 SPI 3.7 (<0.09) 2.7 (<0.09) 2.3 (<0.09) SGD (cons.) 11.6 (0.09) 8.6 (0.09) 7.1 (0.09) SGD (opt.) 20.2 (1.74) 14.8 (0.43) 12.3 (0.26) 812 SPI 4.3 (<0.09) 2.6 (<0.09) 2.0 (<0.09) GRIPS 16.8 (0.87) 10.0 (0.09) 7.6 (0.09) 847 SPI 7.7 (0.09) 5.4 (<0.09) 4.6 (<0.09) GRIPS 29.8 (4.52) 21.0 (2.00) 18.0 (1.04) and designed sensitivity curves of HXI and SGD (here the adopted sensitivity curve corresponds to our ‘opti- mistic’ case). In 2011-2012, 6 SNe Ia were discovered within ∼ 20 Mpc, 3 of which were within ∼ 15 Mpc (from the Asiago SN Catalog; Barbon et al., 1999). Most of these were discovered soon after the explosion, and especially the nearest ones were all discovered within a week after the explosion – thus, ToO follow-up at the hard X and soft γ-ray peak (2 - 3 weeks after the explosion) is feasible. With 106 s exposure, we predict that a few (optimistic) or one (conservative) SNe Ia per year are reachable by Astro-H. 207 Keiichi Maeda, Yukikatsu Terada, Aya Bamba 100 1000 1E-8 1E-7 1E-6 SGD HXI (b) 60 day Fl ux [c m 2 s 1 k eV 1 ] Energy (keV) 100 1000 1E-8 1E-7 1E-6 SGD HXI (a) 20 day Fl ux [c m 2 s 1 k eV 1 Figure 3: Examples of synthetic spectra at (a) 20 days and (b) 60 days after the explosion. Shown here are angle- averaged spectra for models DD2D asym 04 (red line), W7 (gray; Nomoto et al., 1984), and DD2D iso 04 (dark blue) (at 10 Mpc). The angle-dependent spectra seen from two opposite directions are shown for DD2D iso 04 (green and cyan) at 20 days. At 60 days the angle dependence is small. The angle dependence is small for DD2D asym 04 at both epochs. Sensitivity curves for an exposure with 106 seconds of HXI and SGD on board Astro-H (Tajima et al., 2010; Takahashi et al., 2010) are shown by black lines. 4 Discussion and Conclusions According to our simulations of radioactive decay sig- nals from SNe Ia, the new generation hard X-ray and soft γ-ray observatories (either NuStar or Astro-H) are expected to be capable of detecting these signals from SNe Ia up to ∼ 20 Mpc with 106 s exposure. This will hopefully lead to nearly annual detections, dramatically changing the field. We thus propose follow-up of nearby SNe Ia by these telescopes in a ToO mode. With a stan- dard set up with a few 105 s exposure, the detection will be limited to extremely nearby objects (i.e., up to ∼8 - 12 Mpc for SNe Ia with average brightness), but still there is a good chance of first solid detection of the sig- nal from SNe Ia. Once detected, it will provide various diagnostics on explosive nucleosynthesis and explosion mechanisms, and here a combination of hard X-ray and soft γ-ray will be essential. A similar argument applies for core-collapse SNe. By combining results from similar simulations for core- collapse SNe (Maeda, 2006) and those obtained for SNe Ia (Maeda et al., 2012), we find the following. SNe IIp (an explosion of a red supergiant) is not a promising tar- get in the soft bands, since the thick H envelope is still opaque in the early phase where the 56Ni decay can pro- vide the strong emission in these band passes. Among different types of core-collapse SNe, SNe IIb/Ib/Ic (an explosion of a He or C+O star) are most promising. We estimate that the peak date in the high energy emission will be similar to (or a bit delayed than) that of SNe Ia (i.e., 2 - 3 weeks since the explosion). The smaller amount of 56Ni here (typically ∼ 0.1M�), taking into account the delayed peak date, results in the expected horizon as 5 − 8 Mpc (conservative and optimistic). Such nearby objects are rare, but expected to be dis- covered at a rate of a few in a decade according to the past statistics. Acknowledgement KM thanks Franco Giovannelli and the organizers of Frascati Workshop 2013 for creating the friendly and stimulating atmosphere. The work by KM has been supported by WPI initiative, MEXT, Japan. The au- thors acknowledge financial support by Grant-in-Aid for Scientific Research from MEXT (22684012, 23340055, 23740141). References [1] Barbon, R., et al.: 1999, A&AS, 139, 531 [2] Dotani, T., et al.: 1987, Nature, 330, 230 doi:10.1038/330230a0 208 http://dx.doi.org/10.1038/330230a0 Expected Hard X-Ray and Soft Gamma-Ray from Supernovae [3] Greiner, J., et al.: 2012, Exp. Astron., 34, 551 doi:10.1007/s10686-011-9255-0 [4] Isern, J., et al.: 2013, A&A, 552, 97 [5] Kasen, D., Röpke, F.K., Woosley, S.E.: 2009, Na- ture, 460, 869 doi:10.1038/nature08256 [6] Koglin, J.E., et al.: 2005, Proc. SPIE, 5900, 266 [7] Kokubun, M., et al.: 2010, Proc. SPIE, 7732, 33 doi:10.1117/12.857933 [8] Khokhlov, A.: 1991, A&A, 245, 114 [9] Maeda, K.: 2006, ApJ, 644, 385 [10] Maeda, K., et al.: 2010a, Nature, 466, 82 doi:10.1038/nature09122 [11] Maeda, K., et al.: 2010b, ApJ, 712, 624 doi:10.1088/0004-637X/712/1/624 [12] Maeda, K., et al.: 2012, ApJ, 760, 54 doi:10.1088/0004-637X/760/1/54 [13] Milne, P.A., et al.: 2004, ApJ, 613, 1101 doi:10.1086/423235 [14] Nomoto, K., Thielemann, F.-K., Yokoi, K.: 1984, ApJ, 286, 644 [15] Roques, J.P., et al.: 2003, A&A, 411, L91 [16] Seitenzahl, I.R., et al.: 2013, MNRAS, 429, 1156 [17] Summa, A., et al.: 2013, A&A, 554, 67 [18] Sunyaev, R.A., et al.: 1987, Nature, 330, 227 doi:10.1038/330227a0 [19] Tajima, T., et al.: 2010, Proc. SPIE, 7732, 34 [20] Takahashi, T., et al.: 2010, Proc. SPIE., 7732, 27 209 http://dx.doi.org/10.1007/s10686-011-9255-0 http://dx.doi.org/10.1038/nature08256 http://dx.doi.org/10.1117/12.857933 http://dx.doi.org/10.1038/nature09122 http://dx.doi.org/10.1088/0004-637X/712/1/624 http://dx.doi.org/10.1088/0004-637X/760/1/54 http://dx.doi.org/10.1086/423235 http://dx.doi.org/10.1038/330227a0 Introduction Expected Characteristics Observational Perspectives Discussion and Conclusions