A Chandra observation of the H₂O megamaser IC 2560

Abstract

A short Chandra ACIS-S observation of the Seyfert 2 galaxy IC 2560, which hosts a luminous nuclear water megamaser, shows (1) that the X-ray emission is extended; (2) that the X-ray spectrum displays emission features in the soft (E < 2 keV) X-ray band (this is the major component of the extended emission); and (3) a very strong (EW ∼ 3.6 keV) iron Kα line at 6.4 keV on a flat continuum. This last feature clearly indicates that the X-ray source is hidden behind Compton-thick obscuration, so that the intrinsic hard X-ray luminosity must be much higher than that observed, probably close to ∼3 × 10⁴² erg s⁻¹. We briefly discuss the implications for powering of the maser emission and the central source.

masers, galaxies: individual: IC 2560, X-rays: galaxies

1 Introduction

IC 2560 is a relatively nearby (D= 26 Mpc) barred spiral galaxy, classified as a Seyfert 2 (Fairall 1986; Kewley et al. 2001). It is notable for exhibiting luminous H₂O maser emission from its nucleus (Braatz, Wilson & Henkel 1996). This maser emission resembles that from the archetypal water megamaser source NGC 4258 in that high-velocity emission is seen up to ΔV≈ 400 km s⁻¹ away from the systemic velocity, the systemic emission is much stronger than the high-velocity emission, and centripetal acceleration of systemic velocity features has been reported (Ishihara et al. 2001). The high-velocity emission has not yet been imaged; if the data are interpreted in the framework of a Keplerian disc, as in NGC 4258, the implied central mass is M_c≈ 2.8 ×^{10^6}M_⊙. Ishihara et al. also analysed an ASCA observation of IC 2560 and concluded that it possesses a fairly heavily obscured (N_H∼ 3 × 10²³ cm⁻²) but low-luminosity (L_{2−10 keV}∼ 10⁴¹ erg s⁻¹) X-ray source.

In this paper we report on a short Chandra observation of IC 2560, which reveals spatially extended soft X-ray emission and a substantially different nature to the hard X-ray source from that inferred by Ishihara et al. (2001) from the ASCA data. Throughout we assume the distance to the galaxy to be 26 Mpc, giving an angular scale of 120 pc arcsec⁻¹.

2 Observation and data reduction

IC 2560 was observed with the Chandra X-ray Observatory (hereafter Chandra: Weisskopf et al. 2000) on 2000 October 29–30, using the Advanced CCD Imaging Spectrometer (ACIS). The galaxy was positioned on the back-illuminated CCD, ACIS-S3 detector. The focal-plane temperature was −120^○ during this observation. The data reduction was carried out using the Chandra Interactive Analysis of Observation (_ciao) version 2.2 package and calibration files in the calibration data base (CALDB) version 2.10. The S3 detector has been inspected for background flares using a source-free region in the 2.5–7 keV band. The detector background appears to be stable during this observation: the background light curve shows that all the data points are within 30 per cent of the mean value. Since the X-ray source of interest is barely extended beyond the point spread function (see Section 3.1), little impact from such moderate background flares is expected on the spectral data of the source. Therefore no background flare rejection has been applied and the resulting good exposure time is 9.8 ks. The mean count rate of the source, corrected for the background, is 3.2 × 10⁻² count s⁻¹, sufficiently low that the data are unaffected by pile-up. An aspect offset, which was present in the original event file, has been corrected using the latest 2002-May-02 alignment file; the observed X-ray emission peaks at the position , about 0.2 arcsec away from the nuclear position in the NASA/IPAC Extragalactic Data base (NED) and well within the absolute astrometric uncertainty (∼0.6 arcsec) of Chandra pointings. The spectral analysis presented below was performed with the spectral analysis software _xspec version 11.2.

3 results

3.1 Extended X-ray emission

X-ray emission from IC 2560 is found to be extended in the Chandra image. Fig. 1 shows the 0.4–7 keV band ACIS-S3 image of IC 2560. The image suggests that there is a faint extension to the south-west up to 6 arcsec (∼720 pc), and possibly to the north-east up to 5 arcsec (∼600 pc), although the significance of these features is low (∼2σ). The azimuthally averaged radial surface brightness profile of the image in Fig. 1 is shown in Fig. 2. To investigate the extension of the X-ray core, a point spread function (PSF) is computed and plotted in Fig. 2 for a comparison. The PSF was computed for the same position on the detector as that of IC 2560 in this observation, and the energy, at which the PSF was constructed, was assumed to be 1 keV, as much of the detected counts are distributed around that energy (see the energy spectrum in Fig. 3). A simple comparison with the PSF indicates that the core region of IC 2560 is extended with a significance of larger than 3σ and at least by 80 pc in radius. The hard band (3–7 keV) image shows a possible elongation in the south-east–north-west direction, but, with only 58 counts detected in the energy range, the reality of this feature is highly uncertain.

1

The ChandraACIS-S3 image of IC 2560 in the 0.4–7 keV band. The contours are drawn at seven logarithmic intervals from 2 to 80 per cent of the X-ray peak at the nucleus position. Low surface brightness extension appears to be present in the north-east–south-west direction. The sky coordinates are of J2000. The angular scale is ≈120 pc arcsec⁻¹ .

2

The radial surface brightness profiles of the 0.4–7 keV emission from IC 2560 (filled circles with solid-line histogram) and of a PSF computed at 1 keV (dotted-line histogram). The data for IC 2560 have been corrected for background and the two profiles are normalized by the peak brightnesses, respectively. One pixel corresponds to ≈0.5 arcsec or 60 pc.

3

The ACIS-S spectrum of IC 2560. The solid histogram shows the best-fitting model consisting of a broken power law and a Gaussian line for Fe K, modified only by Galactic absorption. Note the residuals in the soft X-ray range (see also Fig. 4 ).

3.2 Spectrum

Fig. 3 shows the Chandra ACIS-S spectrum extracted from a circular region with radius of 2.5 arcsec centred on the X-ray peak. A very strong Fe Kα line at 6.4 keV is clearly seen on a faint, flat continuum. A steep rise of soft X-ray emission is seen below 2 keV, which appears to be the major component of the extended emission. When the 0.4–7 keV data are fitted by a broken power law modified only by Galactic absorption1 ( N_H = 6.5 × 10 ²⁰cm ⁻²: Dickey & Lockman 1990) plus a Gaussian for the Fe K line (Fig. 3), a spectral break is found atE_br = 2.0 ^+0.5_−0.4keV with photon indices of Γ= 2.8^+0.2_−0.2 and 0.5^+0.3_−0.7 below and above the break energy, respectively (the quoted errors throughout are 90 per cent confidence limits for one parameter of interest). The line centroid of the Fe K emission is 6.41 ± 0.03 keV when corrected for the redshift of the galaxy. The line is not spectrally resolved: the upper limit to the dispersion of a Gaussian is 150 eV. Besides the Fe K line, the residuals of the above fit suggest the presence of soft X-ray emission features, the origin of which is discussed later. The fluxes as observed are 5.7 × 10⁻¹⁴ erg cm⁻² s⁻¹ in the 0.5–2 keV band, and 3.6 × 10⁻¹³ erg cm⁻² s⁻¹ in the 2–10 keV band. The 0.5–2 keV luminosity corrected for Galactic absorption is 5.0 × 10³⁹ erg s⁻¹. The observed 2–10 keV luminosity is 2.7 × 10⁴⁰ erg s⁻¹ , of which half is contributed by the iron K line.

The flat hard continuum is consistent with a cold reflection spectrum (e.g. Iwasawa, Fabian & Matt 1997, for an example seen in NGC 1068). A spectrum of reflection alone from an optically thick cold slab [e.g. computed using _pexrav by Magdziarz & Zdziarski (1995), in _xspec, assuming the incident source to have a power-law spectrum with Γ= 2] is in good agreement with the continuum shape above 3 keV. The Fe K line flux is (1.32 ± 0.55) × 10⁻⁵ photon s⁻¹ cm⁻², and its equivalent width with respect to the reflection continuum is extremely large, 3.6 ± 1.5 keV.

Based on the possible emission-line features, the soft excess emission could be thermal emission (i.e. collisionally ionized plasma) arising from a nuclear starburst. A fit with a thermal emission spectrum (computed by _mekal in _xspec based on the model calculations of R. Mewe and J. Kaastra with updated Fe L calculations by D.A. Liedahl; Mewe, Gronenschild & van den Oord 1985; Kaastra 1992; Liedahl, Osterheld & Goldstein 1995) gives a temperature of kT= 0.63^+0.11_−0.11 keV and 0.03^+0.03_−0.02 solar metallicity. Since the implied metallicity from the fit is much too low to be realistic, this single-temperature thermal emission model may not be appropriate. It is unclear whether there is a nuclear starburst in IC 2560 capable of producing the thermal emission. Although a number of H _ii regions distributed along spiral arms have been imaged with Hα+[N _ii] (Tsvetanov & Petrosian 1995), these H _ii regions are located 20 arcsec or more away from the nucleus where the X-ray emission is observed, and none of them is detected in the Chandra image. The optical light from the nucleus is dominated by Seyfert 2 emission. The 1.6-μm luminosity of the unresolved nucleus, 2.3 × 10⁴⁰ erg s⁻¹, as measured with the Near-Infrared Camera and Multiobject Spectrometer (NICMOS) on the Hubble Space Telescope (Quillen et al. 2001), is as low as typically seen in the non-Seyfert control sample, suggesting that the near-infrared emission from the nucleus of IC 2560 is largely due to a stellar cluster. Since the near-infrared to soft X-ray luminosity ratio is more than an order of magnitude too low to match those of starburst galaxies, a nuclear starburst does not appear to be powerful enough to produce the observed soft X-rays. An alternative source for the soft X-ray emission is extended photoionized gas. A few suggestive features in the energy range between 0.5 and 1.4 keV (Fig. 4) could be due to recombination lines and radiative recombination continua from highly ionized N, O, Ne, Fe (L-shell emission) and Mg. The steep rise of soft X-ray emission towards lower energies could be dominated by these and other emission features, which are not resolved at the spectral resolution of a CCD, and a scattered nuclear continuum may not necessarily be present. Higher resolution spectroscopy is required to test this hypothesis.

4

The low-energy part of the spectrum of IC 2560. The data have been corrected for the detector efficiency and the energy scale has been corrected for the redshift of the galaxy (z= 0.00975) .

We analysed the ASCA data of IC 2560, available from the public archive, and confirmed the presence of a strong iron K line with equivalent width EW = 2.2 ± 0.8 keV at an energy of 6.4 keV in the spectrum. Prior to our Chandra observation, two contrary claims have been reported for the ASCA data: Risaliti, Maiolino & Salvati (1999) found an enormous iron line with EW = 6.3^+2.6_−3.0 keV at 6.56^+0.25_−0.15 keV, suggesting a Compton-thick source, while Ishihara et al. (2001) interpreted the hard X-ray emission as an absorbed power law with N_H∼ 3 × 10²³ cm⁻² without noting the iron line. Our Chandra data support the interpretation of Risaliti et al. (1999), although our ASCA results do not agree exactly.

4 Discussion

4.1 The true luminosity of the hidden active nucleus

The hard X-ray spectrum dominated by an iron K line indicates the absence of direct continuum emission from a central source in the Chandra band, meaning that we are seeing only reflected light from a hidden nucleus. The absorption column density must be larger than N_H∼ 1 × 10²⁴ cm⁻², i.e. the X-ray source is Compton-thick. How much of the intrinsic luminosity emitted by the hidden nucleus is seen in reflection depends on the obscuration/reflection geometry. Reflection from cold material, as inferred from the iron line energy and the flat hard X-ray continuum, is not an efficient process, since photoelectric absorption within the reflector suppresses reflected light significantly, in particular at low energies. With the observed continuum luminosity of 1.3 × 10⁴⁰ erg s⁻¹ in the 2–10 keV band, the minimum incident luminosity to yield the reflection is about 2.6 × 10⁴¹ erg s⁻¹. However, in a realistic toroidal geometry in which the incident source is hidden from our direct view, the true value will be much larger.

We estimate the upper limit to the 2–10 keV luminosity to be ∼3 × 10⁴² erg s⁻¹, 10 per cent of the infrared (≈bolometric) luminosity, L_{8−1000 μm}≃ 3 × 10⁴³ erg s⁻¹[obtained from the Infra-Red Astronomical Satellite (IRAS) measurements using the formula given by Sanders & Mirabel (1996)], assuming the typical 2–10 keV X-ray to bolometric luminosity ratio for Seyfert galaxies (e.g. Mushotzky, Done & Pounds 1993). The active galactic nucleus (AGN) luminosity is likely to be close to the above upper limit, given the warm IRAS colour (S_{60 μm}/S_{25 μm}= 3.4) and the AGN-dominated nuclear optical spectrum, suggesting that the infrared emission is predominantly powered by a hidden active nucleus.

4.2 Large EW of Fe K line

The EW for the iron Kα line, in excess of 3 keV, is one of the largest measured among the reflection-dominated Seyfert 2 galaxies (see Matt et al. (2000) and Levenson et al. (2002) for recent compilations). Iron overabundance (2–3 solar) could be a possible reason (Ballantyne, Fabian & Ross 2002), but, even with solar metallicity, an optically thick torus can produce a very large EW. In models assuming obscuring matter in a toroidal form, the Fe K line EW depends on the optical depth and geometry of the torus (Leahy & Creighton 1993; Ghisellini, Haardt & Matt 1994; Krolik, Madau & Życki 1994; Levenson et al. 2002). All these models indicate that, in order to produce an EW as large as 3.6 keV, a torus needs to have a column density N_H≥ 3 × 10²⁴ cm⁻² (or Thomson optical depth τ_T≥ 2) and a small half-opening angle, θ≤ 20°, and to be viewed nearly edge-on.

The required column density is consistent with the lack of transmitted primary source emission in the Chandra spectrum, from which the lower limit of the line-of-sight absorption column density, N_H≥ 1 × 10²⁴ cm⁻², has been derived (Section 4.1). The requirement of a small opening angle means that the X-ray absorbing matter should be located close to the central source. The inner radius of the water maser disc was inferred to be 0.07 pc (Ishihara et al. 2001), but the accretion disc could extend inwards, with the inner part of the disc too cold to induce masing (Neufeld & Maloney 1995, hereafter NM95). Perhaps this dense (n>10¹⁰ cm⁻³) inner part of the accretion disc may be the region where the X-ray absorption primarily occurs. However, in this case, the water maser emitting part of the disc must be warped in order to see the X-ray flux directly, otherwise the X-rays from the central source will not impinge on the disc.

4.3 Water maser disc and accretion flow

If the geometry of the maser emission in IC 2560 is similar to that in NGC 4258, which has not yet been determined by imaging, then we can use the X-ray emission and the kinematic data to place constraints on fuelling of the central massive black hole. The observed velocities of maser emission in IC 2560 imply that the outer disc radius is R_out≈ 0.26 pc. Assuming that the maser emission is powered by the X-ray flux from the central source (Neufeld et al. 1994; NM95), as appears likely for the majority of H₂O megamaser sources (see Maloney 2002, for a recent review), we can identify the outer radius of the disc with the location of the molecular to atomic phase transition [see equation (4) of NM95]. For a hard X-ray luminosity of L₂₋₁₀= 10⁴²L₄₂ erg s⁻¹ and black hole mass M_c= 10⁶M₆ M_⊙, the ratio of the mass accretion rate to the viscosity parameter α is then  

1

which, for the inferred mass of the central black hole M_c≈ 2.8 × 10⁶ M_⊙, is . Assuming that the bolometric luminosity is approximately 10 times the 2–10 keV X-ray luminosity, the product of α and the radiative efficiency factor ε is then αε∼ 0.06, similar to the value inferred for NGC 4258 (NM95). (In making this estimate we have assumed that flow through the disc is steady over the accretion time-scale from the disc outer edge, and that the disc sees the true X-ray luminosity.) Although the fractional Eddington luminosity (∼0.03) and mass accretion rate are much higher than in NGC 4258, the efficiency of radiation appears to be similar. We also note that the Compton-thick obscuration of the hard X-ray source is consistent with the accretion disc itself acting as the obscurer, as in NGC 4258, given the larger derived value of .

Acknowledgments

ACF and KI thank Royal Society and PPARC, respectively, for support. PRM is supported by the National Science Foundation under grant AST 99-00871 and by NASA through Chandra X-Ray Observatory grant GO2-3112X. The NASA/IPAC Extragalactic Data base (NED) is operated by the Jet Propulsion Laboratory, California Institute of Technology, under contract with the National Aeronautics and Space Administration.

Footnotes

1

This value is obtained by averaging the data of the H_imap over a 1-deg ² field at the position of the galaxy, and may not be an exact column density towards the position of IC 2560. However, any fluctuations within the field are not expected to be large enough to affect our results.