Table 5.6: Ratio of XIS count rate of each region to that of the whole rem-nant. The spatial distribution of X-ray emission obtained withASCA GIS is assumed.
Region ID ratio
1 0.0674
2 0.0784
3 0.0278
4 0.0131
5 0.0279
6 0.0224
7 0.0148
8 0.0321
9 0.0235
10 0.0259
X [arcmin]
-30 -20 -10 0 10 20 30
Y [arcmin]
-30 -20 -10 0 10 20 30
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Figure 5.19: Angular response of HXD-PIN. The color scale indicates the transmission of fine collimators at each position. The energy dependence of the response is negligible below∼80 keV.
0 0.2 0.4 0.6 0.8 1 1.2 1.4
(0)
FoV (FWHM) FoV (Full)
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4
(1)
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6
(2)
0 0.2 0.4 0.6 0.8 1 1.2
(3) 1.4
0 0.2 0.4 0.6 0.8 1 1.2
(4) 1.4
0 0.2 0.4 0.6 0.8 1
(5) 1.2
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
(6) 0.8
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
(7) 0.8
0 0.1 0.2 0.3 0.4 0.5 0.6
(8)
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
(9)
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
(10)
Table 5.7: Ratio of HXD count rate of each pointing to that of the whole remnant when the spatial distribution of X-ray emission obtained withASCA GIS is assumed.
Pointing ID ARF·ratio
0 0.334
1 0.313
2 0.391
3 0.254
4 0.239
5 0.155
6 0.351
7 0.226
8 0.285
9 0.117
10 0.194
power-law function, independently. Table 5.8 shows the summary of the fitting results.
The fit to the XIS spectrum is not acceptable at the 99% confidence level, and the residuals begin to become large from ∼ 6 keV. This fact is clearly seen in Figure 5.21, where the spectra are plotted together with a power-law function which represents the XIS spectrum. The spectral steepening begins in the XIS band and continues smoothly to the HXD band. This plot strongly suggests there exists spectral steepening between 5 keV and 20 keV. Such a spectral feature has been clearly shown for the first time with this work.
Table 5.8: Summary of power-law fitting to the XIS and HXD spectrum of the whole remnant of RX J1713.7−3946a
Γ NH Fluxb χν2 (ν)
(1022 cm−2) (erg s−1 cm−2)
XIS (0.5–10 keV) 2.36±0.01 0.77±0.01 (7.71±0.03)×10−10 1.18 (2158) HXD (12–40 keV) 3.0±0.1 — (1.32±0.05)×10−10 1.20 (72)
aErrors represent 90% confidence.
bCorrected for the absorption. The calculated energy band is 1–10 keV for XIS and 10–40 keV for HXD.
Figure 5.21: The Suzaku (XIS+HXD-PIN) spectrum of RX J1713.7−3946 normalized to the whole remnant. The solid lines are absorbed power-law model obtained by fitting the XIS spectrum in the energy range of 0.5–10 keV.
The lower panel shows the ration of the data to the model.
Next, we try to deduce physical parameters by fitting models with a high-energy cutoff to the combined spectrum. The observed X-ray spectrum is interpreted as synchrotron emission from high-energy electrons (Koyama et al. 1997). The electrons responsible for the X-ray emission are believed to be the highest energy populations of them. These electrons could suffer from cutoff due to synchrotron radiative losses. Although we can-not predict the cutoff shape of the electron spectrum with current diffusive acceleration theory, we assume here a conventional form of electron spectrum, a power law with an exponential cutoff,
Ne(E)∝E−pexp (
−E E0
)
(5.5) Then, we can calculate the synchrotron spectrum using equation (2.42). By using δ-functional approximation, we obtain the synchrotron spectrum in the form of
dN
dǫ ∝ ǫ−Γexp [
− ( ǫ
ζǫ0
)1/2]
, (5.6)
Γ = p+ 1
2 (5.7)
where
ǫ0 =~ωc ≃5.3 ( B
10 µG
) ( E0
100 TeV )2
keV (5.8)
is the characteristic synchrotron energy for an electron of energyE0, andζ is a parameter introduced to adapt this formula to accurate numerical calculations, which show that the best broad-band fit, with accuracy better than 25% in the entire region up to ǫ ∼20ǫ0, is provided by ζ = 1 (Aharonian 2000). Therefore, we takeζ = 1 here.
The typical cooling time of highest energy electrons which emit synchrotron X-rays is shorter than the source age, even for young, 1000 yrs-old SNRs. This causes spectral steepening, hence, p = 3 rather than p = 2 is the most probable case (Uchiyama et al.
2003). We, then, fitted the Suzaku spectrum with the model shown as equation (5.6) with fixed value of Γ = 2. The cutoff energy is obtained as 6.8+0.4−0.3 keV in this case. We also fitted with two other values of Γ = 1.5 and Γ = 1.75 in order to see the dependence of the parameters on the photon index. The resultant best-fit parameters are shown in Table 5.9, and the spectrum is plotted together with the model of Γ = 2 in Figure 5.22.
In the fitting procedures, constant factors which represent the relative normalization of HXD-PIN above XIS are introduced, considering some uncertainties in calculating the average spectra. The fit is marginally unacceptable at the 99% confidence level. However, the statistic of the data (in the energy band around 2 keV) is comparable to that of the observational data of the Crab nebula and the fit statistic has a similar value to that obtained when we fit the Crab spectrum with an absorbed power-law model. Thus, we consider that this fitting is essentially acceptable under the current level of instrumental
Figure 5.22: Same as Figure 5.21 but with the best-fit model with a high-energy cutoff described as equation (5.6)
.
calibrations. The fit becomes acceptable at the 99% confidence level if we consider 2%
of the count in each bin as the systematic error.
The obtained cutoff energy should depend on the constant factor introduced in the fitting. As described above, we have already included the relative difference in the ab-solute flux of XIS and HXD-PIN which was obtained with fitting to the Crab spectra.
Therefore, the constant factor should be unity if there is no systematic uncertainties.
However, we expect two major components as the systematic uncertainties. One is sys-tematic errors in the HXD flux due to the misestimation of the background and the other is those in the procedure of obtaining the averaged spectra using the ASCA. As already described in§5.4.3, the former is estimated as 20%. The latter can be estimated as∼10%, considering the photon index distribution of 2.8–3.5 above 10 keV. Therefore, we take 20% as the systematic error in the constant factor.
We evaluated the cutoff energy by fitting the spectrum with some fixed values of the constant factor in the range of∼ ±20%. We plot the obtained cutoff energy against the value of constant factor in Figure 5.23. As shown in this figure, the cutoff energy changes almost linearly depending on the constant factor. When the constant factor changes
±20%, the cutoff energy varies ±2 keV, from which we conclude that the systematic
Table 5.9: Fitting results with a model defined by equation (5.6)a
Γ ǫ0 Norm.b NH Const. χν2 (ν)
(fixed) [keV] [photons s−1 keV−1 cm−2] [1022 cm−2]
1.5 1.24±0.03 0.57±0.01 0.64±0.01 1.09+0.04−0.03 1.11 (2232) 1.75 2.45+0.08−0.07 0.481±0.009 0.68±0.01 0.94±0.03 1.10 (2232) 2.0 6.8+0.4−0.3 0.405+0.008−0.007 0.72±0.01 0.80±0.03 1.11 (2232)
aErrors represent 90% confidence.
bNormalization at 1 keV.
Figure 5.23: The relation between the cutoff energy and the constant factor when Γ = 2.0 (black). The red point shows the best-fit parameter when the Suzakuspectrum is fitted with the constant factor free. The error bars indicate the statistical errors of 90% confidence level.
Chapter 6
Observation & Results of RX J0852.0 − 4622
6.1 Overview of RX J0852.0 − 4622
RX J0852.0−4622 (also called G226.2−1.2 or Vela Jr.) is the second shell-type SNR which was spatially resolved in TeV gamma-rays (Aharonian et al. 2005). Similarly with RX J1713.7−3946, RX J0852.0−4622 was discovered in the data of ROSAT All-Sky Survey in the line of sight to the Vela SNR (Aschenbach et al. 1998). Figure 6.1 (left) shows the X-ray image of the Vela region byROSAT. The SNR, RX J0852.0−4622, which can be seen in the lower left of the image, extends over a circular region with a radius of ∼ 1◦. Based on the ROSAT data, whose bandpass was limited to soft X-rays below 2 keV, Aschenbach et al. (1998) argued that the emission from the SNR can be explained either by a hot thermal model with a temperature of∼2.5 keV or by a power-law model with a photon index of∼2.6.
The ASCA observations, with an imaging capability up to ∼ 10 keV, confrimed the X-ray emission from the bright rims is dominated by non-thermal components (Slane et al. 2001). This is the second shell-type SNR in which non-thermal emission dominates in X-ray, following RX J1713.7−3946 (Koyama et al. 1997; Slane et al. 1999). The photon index of the non-thermal components was obtained∼2.6. Due to the bright and spatially varying emission from the Vela SNR, it is not confirmed whether the thermal component comes only from the Vela SNR or it also comes from RX J0852.0−4622 itself. The following Chandra observations of the northwest (NW) rim revealed sharp filamentary structures similar to those discovered in SN1006 and RX J1713.7−3946 (Bamba et al.
2005).
The CANGAROO collaboration detected TeV gamma-rays from the direction consis-tent with the peak of the X-ray emission in the NW rim (Katagiri et al. 2005). Then, with highly sensitive stereoscopic observations, the H.E.S.S. collaboration detected the TeV
Figure 6.1: (left) The X-ray image of the Vela SNR region obtained with ROSAT taken from Aschenbach et al. (1998). X-rays with energy
> 1.3 keV are accumulated. The circular structure in the left side is SNR RX J0852.0−4622. (right) TeV gamma-ray image of RX J0852.0−4622 obtained with the H.E.S.S. Cherenkov telescope (Aharonian et al. 2005).
of X-rays detected with ROSAT (Aharonian et al. 2005). Figure 6.1 (right) shows the image of RX J0852.0−4622 obtained with the H.E.S.S. telescope. The H.E.S.S. spectrum of RX J0852.0−4622 can be well fitted with a power law of Γ = 2.1.
The age and distance of RX J0852.0−4622 are not decisive yet. Iyudin et al. (1998) estimated the age of∼680 yr and the distance of∼200 pc based on the flux of44Ti lines detected with COMPTEL onboard CGRO. Tsunemi et al. (2000) estimated a more or less similar age between 630 and 970 yr based on the observations of Ca lines withASCA. These estimates suggest that RX J0852.0−4622 is one of the closest supernovae to the earth. However, Slane et al. (2001) argued that the column density for the spectrum of the SNR is larger than that for the Vela SNR and that the distance to RX J0852.0−4622 should be much larger, 1–2 kpc.