The highly variable X-ray spectrum of the luminous Seyfert 1 galaxy 1H 0419-577

K. L. Page 0 P K. A. Pounds 0 J. N. Reeves 0 P. T. O'Brien 0 0 X-Ray Astronomy Group, Department of Physics & Astronomy, University of Leicester , Leicester LE1 7RH An XMM-Newton observation of the luminous Seyfert 1 galaxy 1H 04192577 is presented. We find that the spectrum is well fitted by a power law of canonical slope G , 1:9 and three blackbody components (to model the strong soft excess). The XMM-Newton data are compared and contrasted with observations by ROSAT in 1992 and by ASCA and BeppoSAX in 1996. We find that the overall X-ray spectrum has changed substantially over the period, and suggest that the changes are driven by the soft X-ray component. When bright, as in our XMM-Newton observation, it appears that the enhanced soft flux cools the Comptonizing corona, causing the 2 - 10 keV power law to assume a 'typical' slope, in contrast to the unusually hard ('photon-starved') spectra observed by ASCA and BeppoSAX 4 years earlier. I N T R O D U C T I O N 1H 04192577 (also known as LB 1727, 1ES 04252573 and IRAS F0425025718) is a radio-quiet Seyfert galaxy, with a 60-mm flux of 0.18 Jy and an apparent magnitude of 14.1. It is a moderateredshift object z 0:104 and relatively bright X-ray source which has been observed over recent years by ASCA, ROSAT and BeppoSAX. 1H 04192577 was also one of the brightest Seyfert galaxies detected in the extreme-ultraviolet (EUV) by the ROSAT Wide Field Camera (Pye et al. 1995) and EUVE (Marshall, Fruscione & Carone 1995). Optical spectra taken over the same period in 1996 as the ASCA and BeppoSAX X-ray observations (Guainazzi et al. 1998) show 1H 04192577 to be a typical broadline Seyfert 1, in accordance with the classification by Brissenden (1989). Over the 2 10 keV band, Seyfert galaxies can usually be modelled by a power law, with photon index G , 1:8 2. Below about 1 keV a soft excess is often reported, although the limited bandwidth and resolution of previous missions have made it difficult to distinguish a soft emission component from the effects of absorption by ionized matter. In the case of 1H 04192577, Turner et al. (1999) did conclude that there is a soft emission component on the basis of simultaneous ROSAT High Resolution Imager (HRI) and ASCA observations. However, those authors found the 2 10 keV power law to be unusually flat, with G , 1:5 1:6, and to extend down to 0.7 keV. The BeppoSAX observation of 1H 04192577 in 1996 September (Guainazzi et al. 1998) also found an unusually flat power law (over 3 10 keV of G , 1:55, but provided no independent soft X-ray data, owing to technical problems with the Low Energy Concentrator Spectrometer (LECS). The link between the hard X-ray power law and a soft X-ray emission component is critical in the context of the accretion disc/ corona model for active galactic nuclei (AGN), where the hard X-ray emission is explained by Comptonization of optical/EUV photons from the disc by energetic electrons in an overlying corona (e.g. Haardt & Maraschi 1991). In this model the energy balance between the soft photon flux and the corona then determines the hardness of the spectrum in the 2 10 keV band. The soft excess may be the tail of the big blue bump, representing the thermal emission from an accretion disc surrounding the central black hole (Shields 1979; Czerny & Elvis 1987; Ross & Fabian 1993). In our XMM-Newton observation, reported here, we find a strong and broad soft emission component [consistent with an earlier ROSAT Position Sensitive Proportional Counter (PSPC) observation in 1992 (Guainazzi et al. 1998)] and a 2 10 keV power-law continuum slope typical of Seyfert 1 galaxies. We discuss our result in terms of the stronger soft photon flux cooling the coronal electrons, with a resulting steepening of the power law, and suggest that it represents direct observational support for the disc/corona model for the hard X-ray emission from radio-quiet AGN. In that general class of models, back-irradiation of the accretion disc by the hard X-ray flux can result in additional features being imprinted on the emerging X-ray spectrum. The most obvious of these reflection features (Pounds et al. 1990; Nandra & Pounds 1994) is often an emission line at 6.4 keV arising from fluorescence in near-neutral Fe. This line has emerged as a powerful diagnostic of the inner regions in AGN since ASCA observations found it to be broadened and redshifted (Tanaka et al. 1995; Nandra et al. 1997a,b). Early observations from XMM-Newton have shown a rather different situation to apply in several high-luminosity Seyferts (similar to 1H 04192577), with a weaker and higher energy (ionized) broad Fe K line, resolved from a narrow line at 6.4 keV (e.g. Reeves et al. 2001; Pounds et al. 2001). The latter component, interpreted as scattering from neutral matter distant from the hard X-ray source (e.g. in the molecular torus), is emerging as a common feature in AGN observations by XMMNewton and Chandra. In previous observations of 1H 04192577, Turner et al. (1999) found evidence for an emission line in the ASCA data at 6.39 keV, with equivalent width (EW) of 700 ^ 400 eV (data from 1996 August). A line is not detected in the BeppoSAX data (Guainazzi et al. 1998), but with a rather high upper limit of ,250 eV for the equivalent width. In Section 2 the data from XMM-Newton are summarized, followed by data analysis in Section 3. Comparisons with the ROSAT, ASCA and BeppoSAX observations are reviewed in Section 4. Note that all fit parameters are given for the rest frame of the AGN, with values of H0 50 km s21 Mpc21 and q0 0 assumed throughout. Errors are quoted at the 90 per cent confidence level Dx 2 2:7 for one interesting parameter). X M M - N E W T O N O B S E R VAT I O N S The XMM-Newton observation of 1H 04192577 took place on 2000 December 4 and lasted for just over 8 ks. Because of a coordinate error, X-ray data were only obtained from the European Photon Imaging Camera (EPIC) PN camera (Struder et al. 2001). The PN data were reduced with the XMM SAS (Science Analysis Software), using EPCHAIN to produce the event list. This was further filtered using XMMSELECT within SAS. Both single- and double-pixel events (patterns 0 4 in XMMSELECT) were selected, with the low-energy cut-off being set to 200 eV. The spectrum was extracted within a box-shaped region of size 20 by 10 pixels (corresponding to 87 43:5 arcsec2). The XSPEC v11.0 software package was used to analyse the background-subtracted spectrum, using the most recent response matrices. (The spectrum was first binned, using the FTOOL command GRPPHA, to provide a minimum of 20 counts per bin.) S P E C T R A L A N A LY S I S Since no significant changes occurred in the X-ray flux over the 8000-s observation, the summed data were used in modelling the spectrum. The spectral analysis was begun in the conventional way, comparing the data for 1H 04192577 with a range of parametric models until the best reduced x 2 value was obtained. (...truncated)