北京十一选五开奖结果 www.frdg.net Not to appear in Nonlearned J., 45.
OBSERVATIONAL EVIDENCE FOR STRONG DISK COMPTONIZATION IN GRO J1655 ? 40
arXiv:astro-ph/0105426v1 24 May 2001
Aya Kubota1 and Kazuo Makishima2 [email protected] Department of Physics, University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-0033, Japan and Ken Ebisawa3 code 662, Laboratory of High Energy Astrophysics, NASA/Goddard Space Flight Center, Greenbelt, MD 20771, U.S.A ABSTRACT Analysis was made of the multiple RXTE/PCA data on the promised black hole candidate with superluminal jet, GRO J1655 ? 40, acquired during its 1996– 1997 outburst. The X-ray spectra can be adequately described by the sum of an optically thick disk spectrum and a power-law. When the estimated 1–100 keV power-law luminosity exceeds 1 × 1037 erg s?1 (assuming a distance of 3.2 kpc), the inner disk radius and the maximum color temperature derived from a simple accretion disk model (a multi-color disk model) vary signi?cantly with time. These results recon?rm the previous report by Sobczak et al. (1999). In this strong power-law state (once called “very high state”), the disk luminosity decreases with temperature, in contradiction to the prediction of the standard Shakura-Sunyaev model. In the same state, the intensity of the power-law component correlates negatively with that of the disk component, and positively
present address: Institute of Space and Astronautical Science, 3-1-1 Yoshinodai, Sagamihara, Kanagawa 229-8510, Japan
2 3 1
Also Cosmic Radiation Laboratory, Institute of Physical and Chemical Research Also Universities Space Research Association
–2– with the power-law photon index, suggesting that the strong power-law is simply the missing optically thick disk emission. One possible explanation for this behavior is inverse-Compton scattering in the disk. By re-?tting the same data incorporating a disk Comptonization, the inner radius and temperature of the underlying disk are found to become more constant. These results provide one of the ?rst observational con?rmations of the scenario of disk Comptonization in the strong power-law state. This strong power-law state seems to appear when color temperature of the disk exceeds the certain threshold, ? 1.2 – 1.3 keV. Subject headings: black hole physics
When the mass accretion rate is high, a black hole binary (BHB) resides in the soft state, which is characterized by a very soft spectrum, accompanied by a power-law tail. The soft spectral component is believed to be thermal emission from an optically thick accretion disk around a central BH (e.g., Makishima et al. 1986), because it can be reproduced by a multi-color disk model (MCD model, Mitsuda et al. 1984) which approximates emission from a standard accretion disk (Shakura & Sunyaev 1973). The model has two parameters; the maximum color temperature of the disk, Tin , and an apparent inner radius rin . Using a spectral hardening factor of κ ? 1.7–2.0 (Shimura & Takahara 1997), and a correction factor for the inner boundary condition, ξ = 0.41 (Kubota et al. 1998), rin can be related to the true inner radius Rin , as Rin = κ2 · ξ · rin . (1) X-ray observations indicate that Rin remains constant at 6Rg that is the last stable orbit for a non-spinning BH (Rg is the gravitational radius). Although this “standard picture” remained generally successful (e.g., Ebisawa et al. 1993, Tanaka & Lewin 1995), recently two deviations from its predictions have been noticed in some soft-state BHBs. One is quite small values of Rin , compared to 6Rg , and the other is signi?cant variations in Rin . The former is found in the superluminal jet sources, GRO J1655 ? 40 and GRS 1915 + 105 (e.g., Zhang et al. 1997). The latter includes GRO J1655 ? 40 (Sobczak et al. 1999; hereafter S99), LMC X-1 (Wilms et al. 2001) and XTE J1550 ? 564 (Kubota 2001). Although these anomalies are often attributed, e.g., to strong disk Comptonization and/or very high value of κ, no convincing evidence has been available. These issues may be related to the general theoretical belief that the standard-disk picture is valid only for a rather limited range of the mass accretion rate (e.g., Esin et al. 1997).
–3– In order to examine to what extent the standard picture is valid, we have analyzed X-ray spectra of GRO J1655 ? 40 obtained by multiple pointings with RXTE. This BHB has been reported to exhibit the two peculiarities mentioned above. Moreover, its BH mass, distance, and inclination angle are accurately estimated to be 7.02 ± 0.22 M⊙ , 3.2 ± 0.2 kpc, and 69.5 ± 0.08? , respectively (e.g., Orosz & Bailyn 1997). These make GRO J1655 ? 40 ideal for our purpose.
Observation and data reduction
We analyzed 72 archival data sets of GRO J1655 ? 40 obtained with the RXTE/PCA, covering the 16-months outburst in 1996–1997. This is the same dataset as utilized by S99 except for the last few observations when the source exhibited a signature of the hard state. We co-added the data from the individual proportional counter units, and produced one co-added spectrum for each pointing. We select good data in the basic manner for bright sources. We estimated the PCA background for each observation using pcabackest version 2.1e. In order to correct a possible < 10 % over/under-estimation of the background by the pcabackest, we compared the on-source spectra and the predicted model background spectra in the hardest energy band (60–100 keV), where the signal ?ux is usually negligible. When necessary, we adjusted the normalization factor of the background spectrum by ?10 ? +10%. We make the response matrix for each observation using pcarsp version 7.10, and add 1% systematic error to each energy bin of the PCA spectrum. Over the 20–35 keV range, we increased the systematic error to 10%, to take into account the response uncertainties associated with the Xe-K edge at ? 30 keV.
Data analysis and results 3.1. Standard modeling
According to the canonical spectral modeling, we ?t the obtained 3–30 keV PCA spectra of GRO J1655?40 with the MCD plus a power-law. After Yamaoka et al. (2001), we subject the MCD component to several absorption features (a line and edges). We do not discuss these absorption features any further. To the two constituent continuum components, we apply a common photoelectric absorption with the column ?xed at NH = 7 × 1021 cm?2 , by referring to the ASCA/GIS data of this source on 1997 Feb.26 (Kubota 2001). Moreover, as for the data obtained before 1996 May 21 (MJD 50224), we ?x the power-law photon index
–4– Γ 2.1, because the power-law component was too weak to constrain Γ. The ?ts have been acceptable for all the PCA spectra. In Fig.1, we show evolution of 2 4 the best-?t model parameters, including the disk bolometric luminosity, Ldisk (≡ 4πσrin Tin ), the 1–100 keV power-law luminosity, Lpow , calculated assuming an isotropic emission, and their sum, Ltot . Thus, the entire PCA data span of this source can be divided into three characteristic periods, referring mainly to Lpow . The 1st period (Period 1; before day 141) is characterized by a very low level of Lpow , while it is very high (> 1 × 1037 erg s?1 ) in Period 2, which was once called “very high state” by S99. In Period 3 (after the data gap), Lpow returns low. In Period 3, the values of Rin , which are obtained by utilizing eq.(1) with κ = 1.7 and ξ = 0.41, remain constant (26 km) against relatively large intensity variation, while in Period 2 they are observed to change signi?cantly between ? 6 km and ? 24 km. When we ?x Rin at 26 km and instead allow NH to vary, the ?ts for the Period 1 and 2 data become signi?cantly worse (Fig.1e), and NH changes violently. Thus, the variation of Rin is real as long as we utilize the canonical two components model with constant values of κ and ξ. Clearly, GRO J1655 ? 40 in Period 2 violates the standard picture. In order to highlight the anomalies of GRO J1655 ? 40, we plot Ldisk against Tin in Fig.2a. For comparison, we also plot the data points of a typical BHB, LMC X-3 (e.g., Kuiper et al. 1988), of which the BH mass (5–7.2 M⊙ ) and the inclination angle (65? –69? ) are both quite similar to those of GRO J1655 ? 40. The results on LMC X-3 were obtained through the spectral ?tting of the same PCA data as reported by Wilms et al. (2001). 4 Thus, the data points for LMC X-3 follow a simple relation of Ldisk ∝ Tin with a constant Rin as Ldisk varied by a factor of 10. Moreover, assuming its distance and inclination angle as 55 kpc and 66? respectively, the absolute value of Rin is calculated as ? 60 km, which coincide with 6Rg for a BH of 6M⊙ . In other words, the accretion disk in LMC X-3 perfectly satis?es the standard picture. In contrast, GRO J1655 ? 40 exhibits a distinct behavior on this Tin -Ldisk plane. The data points in Period 2 and Period 1 deviate from the standard 4 Ldisk ∝ Tin relation, while those of Period 3 satisfy the relation except that the value of Rin is much smaller than that of LMC X-3. Although the deviation from the standard picture has been found in both Period 1 and 2, it is much more signi?cant in Period 2 than in Period 1. In addition, the observed PCA spectra in Period 1 are relatively similar to those in Period 3 (see S99), while those in Period 2 are characterized by very strong hard emission component. Therefore in this letter, we mainly focussed on the anomalous behavior in Period 2. Hereafter, we call Period 2 anomalous regime, while call Period 3 standard regime.
–5– 3.2. Di?erences between anomalous and standard regimes
A prominent di?erence between the anomalous and standard regimes is found in behavior of the power-law component. In the anomalous regime, Lpow negatively correlates to Ldisk , in such a way that Ltot is kept approximately constant at a ceiling value of ? 1.7 × 1038 erg s?1 (Fig.1a). To our surprise, this ceiling corresponds roughly to ? 15 % of the Eddington luminosity, LE ? 1.1 × 1039 erg s?1 (assuming solar abundance) for a ? 7M⊙ BH in GRO J1655 ? 40, instead of LE itself. Furthermore, as shown in Fig.3, Γ gradually increases as Lpow gets higher in the anomalous regime, while it stays constant at ?2.1 in the standard regime. Therefore, the property of the hard component in the anomalous regime may be intrinsically di?erent from that in the standard regime. A simple interpretation of the source behavior in the anomalous regime is to presume that there emerges a third spectral component, which is harder than the MCD emission but softer than the hard component in the standard regime. Then, the strong anti-correlation between Lpow and Ldisk , seen in the anomalous regime, can be explained by assuming that this third component strongly and negatively correlates with the MCD component. It is therefore natural to assume that some fraction of the photons emitted from the optically thick accretion disk are converted into the third spectral component, instead of directly reaching us. The third component is most probably produced through inverse-Compton scattering of the MCD photons by high energy electrons that may reside around the disk.
Spectral ?tting incorporating a Comptonized component
We re-?t the same PCA spectra in the anomalous regime, with a three-component model, obtained by adding a Comptonized blackbody (“compbb”; Nishimura, Mitsuda, & Itoh 1986) component to the original two component model. The compbb model has four parameters; blackbody temperature Tbb , electron temperature Te , Compton optical depth τ , and radiative area of the blackbody for an isotropic emission. However, we cannot constrain all these additional parameters, since the previous two-component model has given acceptable ?ts. We accordingly tie Tbb to Tin , assuming the seed photons for the inverse-Compton to be supplied by the optically thick accretion disk. We ?x Γ of the original power-law component to 2.1, an average in the standard regime (Fig.3). Furthermore, to avoid a strong coupling between τ and Te , we ?x Te at a representative value of 10 keV, considering that strong Compton cooling by ample seed soft photons from the optically-thick disk will make Te signi?cantly lower than in the hard state (Te ?30–50 keV; e.g., Grove et al. 1998). In Fig.1c, we plot the re-estimated Tin with open circles. By considering the compbb
–6– component, the highly deviated data points in terms of Tin have thus settled back to a smooth long-term trend. We also re-estimate the luminosity in Fig.2b as Ldisk + Lcbb , where Lcbb is the estimated 0.1–100 keV compbb luminosity, assuming an isotropic emission. Thus, Ldisk + Lcbb plotted against the revised Tin approximately recovers the standard Ldisk ∝ Tin 4 relation for optically-thick accretion disks. Consequently, the value of Rin can also be considered to remain relatively stable, even when a signi?cant fraction of the MCD photons is Comptonized. We conclude that some part of the power-law seen in the anomalous regime has the origin in the MCD photons, modi?ed through the inverse-Compton process, and that the violent variations in the MCD parameters, on time scales of few days or shorter, is not real but apparent. Figure 4a shows the typical PCA spectrum of GRO J1655?40 in the anomalous regime, which corresponds to Observation A (1996 Aug. 6; day 218) presented in Fig.1a, ?tted with the three-component model. We also show the result from the previous two-component ?t in Fig.4b, where the best-?t model is obtained in the range of 3–30 keV. although the spectrum is shown in 3–50 keV.
In §3, we have shown that the scenario of the strong disk Comptonization successfully explains the anomalous regime of GRO J1655 ? 40, and that the underlying disk really satis?es the standard picture. A very similar phenomenon has been observed in another Galactic jet source, XTE J1550 ? 564 (Kubota 2001). Although the disk Comptonization has been discussed extensively in the literature, our results provide one of the ?rst unambiguous observational con?rmations of such a picture. Then, what causes such a strong Comptonization? As mentioned in §1, it was theoretically pointed out that the standard disk cannot be stable under high accretion rates. The anomalous regime can be hence considered to occur when the accretion rate reaches a certain upper critical level, as indicated by the location of anomalous regime (upper right) in Fig.2b. We however remember that the critical disk luminosity, at which the anomalous behavior of GRO J1655 ? 40 appears, is only ? 15% of LE , and that LMC X-3 exhibits no such deviation from the standard-disk even with Ldisk close to LE . Another important di?erence between these two BHBs is that the value of Rin for GRO J1655 ? 40 is only ? 2Rg for a 7M⊙ BH even in the standard regime , while that of LMC X-3 agrees with 6Rg . This fact makes the values of Tin for GRO J1655 ? 40 much higher than those for LMC X-3 when compared at the similar luminosity. Interestingly, the observed value of the temperature of GRO J1655 ? 40, ? 1.2 keV, which corresponds to the critical luminosity, is almost the same
–7– as the upper limit of Tin (? 1.3 keV) for LMC X-3. We hence suggest that the endpoint of the standard regime is not determined by the luminosity but instead by the temperature, which is thought to be ? 1.2–1.3 keV for both LMC X-3 and GRO J1655 ? 40. Such a consideration is consistent with a theoretical expectation that, at a high temperature, the opacity of the disk is given by electron scattering instead of photo-electric absorption that dominates at lower temperatures. Thus, we consider that the systematically higher Tin (or smaller Rin ) is responsible for the anomalous behavior of GRO J1655 ? 40. Because the innermost stable orbit becomes smaller for a prograde rotation around a spinning BH, down to ? Rg in the extreme case, the anomalous behavior of GRO J1655 ? 40 may be attributed ultimately to its BH spin, as already suggested by Zhang et al. (1997), and Makishima et al. (2000). We would like to thank Prof. H. Inoue and Prof. S. Mineshige for helpful discussions. We are also grateful to Dr. C. Done for valuable comments and discussions.
REFERENCES Ebisawa, K., Makino, F., Mitsuda, K., Belloni, T., Cowley, A., Schmidke, P., & Treves, A. 1993, ApJ, 403, 684 Esin, A., McClintock, J.E., & Narayan, R. 1997, ApJ, 489, 865 Grove et al. 1998, ApJ, 500, 899 Kubota, A., 2001, Ph.D. Thesis, University of Tokyo Kubota, A., Tanaka, Y., Makishima, K., Ueda, Y., Dotani, T., Inoue, H., & Yamaoka, K. 1998, PASJ, 50, 667 Kuiper, L., van Paradijs, J., & van der Klis, M. 1988, A & A, 203, 79 Makishima, K., et al. 1986, ApJ, 308, 635 Makishima, K., et al. 2000, ApJ, 535, 632 Mitsuda, K., et al. 1984, PASJ, 36, 741 Nishimura, J., Mitsuda, K., & Itoh, M. 1986, PASJ, 38, 819 Orosz, J.A. & Bailyn C.D. 1997, ApJ, 477, 876 Shakura, N.I., & Sunyaev, R.A. 1973, A&A, 24, 337 Shimura, T. & Takahara F. 1995, ApJ 445, 780 Sobczak, G.J., McClintock, J.E., Remillard, R.A., Bailyn, C.D. & Orosz J.A. 1999, ApJ 520, 776 (S99) Tanaka, Y., & Lewin, W.H.G. 1995, in X-ray Binaries, ed. W.H.G. Lewin, J. van paradijs, & W.P.J. van den Heuvel (Cambridge University Press, Cambridge), 126
–8– Wilms, J. et al. 2001, MNRAS, 320, 327 Yamaoka, K., Ueda, Y., Inoue, H., Nagase, F., Ebisawa, K., Kotani, T., Tanaka, Y., & Zhang, S.N. 2001, PASJ, 53, 179 Zhang, S.N., Cui, W., & Chen, W. 1997, ApJ, 482, L155
Fig. 1.— Long-term variation of the spectral parameters of GRO J1655 ? 40. (a) Histories of Ldisk (?lled star), Lpow (open star), and Ltot (open circle), The three characteristic periods are indicated in the top panel. (b)–(d) Time histories of rin , Tin , and Γ, respectively. Large circles in panel (c) represent Tin obtained by incorporating the disk-Comptonization. (e) That of χ2 /dof; ?lled circles represent χ2 /dof with the free-Rin and ?xed-NH ?ts, whereas crosses those when Rin is ?xed at 26 km and NH is left free.
A This preprint was prepared with the AAS L TEX macros v5.0.
Fig. 2.— (a) The calculated Ldisk plotted against the observed Tin . As for GRO J1655 ? 40, three kinds of symbols specify the data obtained during Period 1 (?lled circles), Period 2 (open circles), and Period 3 (crosses). The results of LMC X-3 (open squares) are also plotted 4 for comparison. The solid lines represent the Ldisk ∝ Tin relation. (b) Same as in panel (a), but the data points in Period 2 (anomalous regime) of GRO J1655 ? 40 are re-calculated considering the Comptonized component as Ldisk + Lcbb .
Fig. 3.— The observed Γ plotted against Lpow . Symbols are the same as in Fig.2.
– 10 –
Fig. 4.— The PCA spectrum of GRO J1655 ? 40 obtained in Observation A in Fig.1a. Predictions of the best-?t three-component model (panel a) and the canonical two-component model (panel b) are also shown, together with individual components.