AGN Feedback in the Hot Halo of NGC 4649
Using the deepest available Chandra observations of NGC 4649 we find strong evidences of cavities, ripples and ring like structures in the hot interstellar medium (ISM) that appear to be morphologically related with the central radio emission. These structures show no significant temperature variations in correspondence with higher pressure regions (). On the same spatial scale, a discrepancy between the mass profiles obtained from stellar dynamic and Chandra data represents the telltale evidence of a significant non-thermal pressure component in this hot gas, which is related to the radio jet and lobes. On larger scale we find agreement between the mass profile obtained form Chandra data and planetary nebulae and globular cluster dynamics. The nucleus of NGC 4649 appears to be extremely radiatively inefficient, with highly sub-Bondi accretion flow. Consistently with this finding, the jet power evaluated from the observed X-ray cavities implies that a small fraction of the accretion power calculated for the Bondi mass accretion rate emerges as kinetic energy. Comparing the jet power to radio and nuclear X-ray luminosity the observed cavities show similar behavior to those of other giant elliptical galaxies.
Evidence of the interaction of Active Galactic Nuclei (AGN) with the surrounding hot gas in nearby galaxies and clusters has been observed as morphological disturbances in the X-ray halos in the form of ripples and cavities (e.g. Fabian et al., 2000, 2003; Forman et al., 2005; Fabian et al., 2006). The AGN-induced disturbances have also been observed in the hot interstellar medium (ISM) in the halos of a number of normal elliptical galaxies (e.g. Diehl & Statler, 2007), and are interpreted as a consequence of the thermal X-ray emitting gas being displaced by the AGN jets.
NGC 4649, also known as M60, is a nearby
Finger-like structures in the inner kpc of the diffuse X-ray emission from NGC 4649 have been reported by Randall et al. (2004, 2006) in their study of Chandra and XMM-Newton data. These structures are compared by the authors with those predicted by hydrodynamical simulations of cooling flows in elliptical galaxies (Kritsuk et al., 1998), that is, brighter, cooler inflowing gas surrounded by fainter, hotter outflowing jets. However, the authors found no significant temperature variations across the observed structures. Shurkin et al. (2008), using the same Chandra observation of Randall et al. (2004) analysis, found morphological disturbances in the X-ray emitting gas, and interpreted them as the result of interaction with the central AGN. Instead a subsequent analysis of deeper Chandra observations by Humphrey et al. (2008) showed a generally undisturbed X-ray morphology, consistent with that expected from a hot ISM in hydrostatic equilibrium. Later, the analysis by Dunn et al. (2010) of Chandra observations shallower that those of Humphrey et al. - but deeper than Shurkin et al. - revealed (again) disturbances and cavities in the ISM connected with the radio emission.
Thanks to the relatively small distance and large supermassive black hole (SMBH) mass ( few ) of NGC 4649, Chandra resolves radii close to the Bondi accretion radius, , at which the gravitational binding energy of a gas element becomes larger than its thermal energy (Bondi, 1952). Thus, NGC 4649 represents an ideal case to investigate the following questions: what is the mass accretion rate? What fraction of the accretion power is in the observed nuclear luminosity, and what in the observed jet power? Is NGC 4649 consistent with the previously found correlations between the Bondi mass accretion rate , the power associated with the observed cavities , and radio luminosity (Cavagnolo et al., 2010; Russell et al., 2013)? Due to its low radio power, and its non-being a central dominant galaxy, NGC 4649 is also an ideal case to investigate whether these correlations, mostly found for radio-bright central dominant galaxies in groups or clusters, work equally well in more “normal” elliptical galaxies.
In this paper we revisit the properties of the hot ISM of NGC 4649, using much deeper Chandra data with respect to previous studies and updated atomic databases. We find strong evidence of cavities, ripples and ring like structures that appear to be morphologically related with the central radio emission. In addition, we find that the hot halo is subject to an additional non-thermal pressure term as already reported in previous studies (e.g. Brighenti et al., 2009; Das et al., 2010, 2011; Humphrey et al., 2013). We show that the non-thermal pressure is found on the same scale spatial as the disturbances of the halo and it is spatially correlated with the hot gas pressure and the minimum pressure derived from the radio data. The nucleus of NGC 4649 appears to be extremely radiatively inefficient, with highly sub-Bondi accretion flow, releasing a very small fraction of the accretion power in form of kinetic energy in the surrounding halo.
2 Data Reduction and Analysis
NGC 4649 has been observed by Chandra with the ACIS detector
(Garmire, 1997) six times, between April 2000 and August 2011.
Level 2 event data were retrieved from the Chandra Data
For each data set we generated a full resolution image in the energy band. To take advantage of the longer exposure time and identify fainter signatures we also produced a merged image of the six observations; to this end we used the wavdetect task to identify point sources in each observation with a 2 sequence of wavelet scales (i.e., 1, 2, 4, 8, 16 and 32 pixels) and a false-positive probability threshold of . Then we used the reproject_aspect task to modify the aspect solution minimizing position differences between the sources found, and finally merged the images with the merge_all script.
2.1 Image of halo
We examined the co-added image for evidence of morphological disturbances that may be connected with the nuclear radio source. To this end we then ran wavdetect on the merged image to detect fainter point sources. For each of these appropriate elliptical regions were generated with the roi CIAO task, both for the source and for the nearby background. We then processed the merged image with the dmfilth task to remove the detected point sources, replacing the counts by sampling the Poisson distribution of the pixel values in the concentric background region. The resulting data image is therefore expected to show the morphology of the ISM diffuse emission, although it will still include point sources below the detection threshold.
We produced separate images in different energy bands (, and ) in order to obtain a three-color image that may pinpoint evidences of structures in the ISM. We then adaptively-smoothed each band image using the csmooth tool (Ebeling et al., 2006), with minimum and maximum significance S/N levels of and , respectively to enhance fainter, extended features of the diffuse emission. The merged three-color image (Figure 1) suggests hints of structures and cavities in the soft emission in the inner region.
To gain a more quantitative insight on morphological disturbances, we performed a one-dimensional fit of the surface brightness profile - evaluated in concentric annuli in the source-free merged image - with a standard model using Sherpa. The fit was performed in the inner 100” () in order to avoid contamination from the emission of the companion galaxy NGC 4647. The results (Figure 2) show that, although the model reproduces the large-scale behavior of the surface brightness profile, the fit is poor, with a reduced . Adding a second model did not increase the goodness of the fit, while yielding unphysical model parameters. Most of the contribution comes from well defined radii at ” and ” (corresponding to and , respectively).
To investigate the origin of these residuals, we performed a 2-D fit of the full-band source-free merged image with a two-dimensional model. Again, adding a second model in our fit did not increase the fit goodness significantly. The residual distribution in the inner region is presented in the left panel of Figure 3, where we applied a 3 pixel FWHM Gaussian smoothing to highlight the presence of structures. In the same panel we superimpose to the residual contours (shown in green) the VLA emission contours (Stanger & Warwick, 1986, shown in red). We notice a striking correlation between the outer radio lobes and the regions of negative residuals; regions of positive residuals appear to lie on both sides of the radio emission and are responsible for the features at a galactocentric radius of seen in Figure 2. On a larger scale (), the residual map shows a ring-like structure - recalling those observed in NGC 1275 (e.g., Fabian et al., 2006) - as shown in 4, where on the left panel we present the inner region, where we applied a 10 pixel FWHM Gaussian smoothing. Radio emission contours are superimposed as in Figure 3. The significance of these structures is shown in the right panels of Figures 3 and 4. Here we show the residual S/N map, evaluated as the ratio between the residuals and the X-ray counts error, binned to a pixel size 4 and 14 times the size of the native ACIS-S pixel, respectively. At this pixel size the residual S/N map closely follows that of the residual map with the gaussian smoothing, as shown by the superimposed contours of the residual map (shown in green). In particular, we notice that the S/N of the binned pixel in the observed structures is of the order of 3 or higher, pointing to an higher significance of the structures as a whole.
In order to evaluate the reliability of the observed structures and their dependence on the particular model used, we performed the 2-D fitting procedure described above excluding data between 5” and 70” (so avoiding the structures shown in Figure 4) and then subtracted the best-fit model so obtained from the complete dataset. In addition, we repeated the same procedure in the four 90 degrees quadrants to investigate possible azimuth anisotropies that could yield the observed structures. All the residual maps so obtained are very similar and show the same structures presented in Figures 3 and 4. To investigate the effect of the point-source removal procedure described above on the observed residual structure, we also used the source catalog produced by Luo et al. (2013) (which used a finer sequence of wavelet scales), obtaining again very similar results. We therefore conclude that the observed structures are real and do not strongly depend on the details of our analysis.
2.2 Spatially resolved spectral analysis
To estimate the properties of the hot gas in the regions of enhanced and suppressed X-ray emission, we performed spectral analysis. Figure 5 shows the region used for spectral extraction with the CIAO specextract task. For each extraction region, background spectra were extracted in the same region from appropriate “blank-sky” fields and normalized equating the count rates of the observed and background data, since essentially all those hard X-rays are due to particle background (Hickox & Markevitch, 2006). Additionally, we extracted spectra in a series of concentric, contiguous annuli, with widths chosen so as to contain approximately the same number of background-subtracted photons (). We placed a lower limit of 2.5” on the annulus width to ensure that the instrumental spatial resolution does not lead to strong mixing between the spectra in adjacent annuli. To account for projection effects, we used the projct model implemented in Xspec (ver. 12.8.0, Arnaud, 1996). We excluded data in the vicinity of any detected point source, as well as in the central part of the interloper galaxy NGC 4647 to prevent possible contamination. We extracted data individually for each observation and combined the source and background spectra, generating spectral response matrices weighted by the count distribution within the aperture (as appropriate for extended sources). To make use of the fit statistic we binned the spectra to obtain a minimum of counts per bin using the specextract task; in the following, errors correspond to the - confidence level for one interesting parameter (). In all the spectral fits we included photo-electric absorption by the Galactic column density along the line of sight (Kalberla et al., 2005).
Spectral fitting were performed in the energy range using a model comprising a vapec thermal component, plus a thermal bremsstrahlung component to account for undetected point sources (Irwin et al., 2003). We adopted solar abundances from Anders & Grevesse (1989), and we allowed the global ratios of O, Ne, Mg and Si with respect to Fe to fit freely (see e.g. Kim et al., 2012), and fixed the remaining ratios at the solar value. In addition, as NGC 4649 is within the Virgo cluster we included an additional hot gas component, with kT fixed at 2.5 keV (e.g. Gastaldello & Molendi, 2002) to account for possible interloper cluster emission. The results are presented in Table 2. The details of the element abundances obtained in the fits will be discussed in a forthcoming paper (Paggi et al., 2014). Here we concentrate on the physical parameters (temperature and gas density). In Figure 6 we show gas temperature and projected pressure (where is the vapec component normalization per arcsec square) in the extraction regions (shown in the corresponding colors) in comparison with the average values obtained in the annulus at the same radius (shown in grey). There is no significant temperature variation between overdense and underdense regions. Instead, the projected pressure is higher in brighter regions.
Figure 7 shows temperature and gas density profiles out to a radius in comparison with that obtained with similar techniques by Humphrey et al. (2008) using shallower Chandra data. The temperature we derive is systematically higher than that of Humphrey et al. (2008), while our gas density is systematically lower. This effect is manly due to the use of the updated ATOMDB 2.0.2 with respect to the 1.3.2 version available before 2010 (see Loewenstein & Davis, 2012). In particular with respect to this previous analysis, while the gas density profiles have similar shapes, we see that our temperature profile, rather than a decrease up to followed by a monotonic increase up to the outer regions, is characterized by a decrease up to followed by a “plateau” of almost constant temperature up to (the same scales of the structures shown in Figure 4). The two profiles are properly fitted by a double broken power-law model (shown in Figure 7 with the dashed lines), with breaks at and .
In the local universe, most central SMBHs in elliptical galaxies are radiatively quiescent (Pellegrini, 2005, 2010; Soria et al., 2006; Ho, 2008; Gallo et al., 2010), and often seen in a “radio-mode”, where they are able to drive jets (e.g., Ho, 2002; Merloni & Heinz, 2007). Mechanical feedback of these jets has a significant impact on the surrounding ISM, with implications for galaxy and SMBH coevolution (e.g. Di Matteo et al., 2005; Croton et al., 2006; Sijacki et al., 2007; Ostriker et al., 2010; King, 2013), and regulation of the hot gas cooling (Dunn et al., 2010). Many aspects of this feedback process, though, are not well understood, including the nature of the material feeding the SMBHs, the details of how accretion originates a mechanical energy output, in the form of a wind or a jet, and then how this couples to the surrounding ISM heating it. Chandra observations of hot gas rich elliptical galaxies often show direct evidence of mechanical feedback in the form of radio jets and inflated bubbles displacing the hot plasma (e.g. McNamara & Nulsen, 2007). From a study of nine X-ray luminous ellipticals harboring cavities, sufficiently nearby to measure the hot gas density and temperature reasonably close to the SMBH, a tight coupling was found between the Bondi mass accretion rate () and the jet power estimated from the energy associated with the observed cavities (Allen et al. 2006; Merloni & Heinz 2007, see also Russell et al. 2013). correlates also with the (extended) radio luminosity, albeit in relationships showing a large scatter (Bîrzan et al., 2008; Cavagnolo et al., 2010; O’Sullivan et al., 2011).
Thanks to the long cumulative Chandra pointing on NGC 4649 we have determined the hot gas density and temperature in a wide range of radii, discovering cavities and a ring-like higher-pressure ripple. We also determine these parameters in the nuclear region with unprecedented accuracy.
3.1 Temperature profile
The residual map of the X-ray emission with respect to a standard model (Figure 4) shows significant cavities and bright spots on sub-kpc scale, as well as more extended ring like structures at scales . These previously unreported features appear to be morphologically related with the weak nuclear radio source (Condon et al., 2002; Shurkin et al., 2008). We find no significant temperature variations between under-dense and over-dense regions (see Figure 6). Moreover, looking at the average temperature profile presented in Figure 7 (left panel), we see an almost constant temperature in the range. This resembles the case of the ripples and cavities around NGC 1275 in the Perseus cluster, discovered by Fabian et al. (2006) with Chandra, which also shows no sign of increased temperature in higher pressure regions. Fabian et al. concluded that the NGC 1275 structures may be isothermal waves whose energy is dissipated by viscosity, with thermal conduction and sound waves effectively distributing the energy from the radio source.
The radial kT profile of NGC 4649 shows a sudden increase up to in the inner . Chandra observations revealed central temperatures higher than the surroundings in other early type galaxies with low or very low power radio sources: NGC 4594 (Pellegrini et al., 2003), NGC 4552 (Machacek et al., 2006), NGC 3115 (Wong et al., 2011) and NGC 4278 (Pellegrini et al., 2012). This hotter core has been interpreted as a result of heating from gravitational effects due to the central SMBH, a recent AGN outburst, or interaction with confined nuclear jets. The clear evidence for AGN feedback on the large scale in NGC 4649 (i.e., the series of cavities, likely more numerous than the N and S ones that will be investigated in detail in Sect. 3.5) provides support to the idea of repeated AGN heating occurring at the center, possibly producing the temperature peak, for example through shock activity. Note that the radiative cooling time in the innermost radial bin is just yr (using the cooling curve from Sazonov et al., 2005), then if such cooling is to be offset a heating mechanism is required. Following the calculation by Fabian et al. (2006), we can interpret the observed structures as isothermal waves and show that they can provide an effective way to heat the ISM. In fact, the thermal pressure deviations from the average value (shown in Figure 6) are about 5-10% (comparable with those observed in NGC 1275, although with less significance). Then, such waves can balance radiative cooling if the cooling time is times the crossing time evaluated from the local sound speed (Bîrzan et al., 2004). This condition is met, since the cross time for the inner region is , while the cooling time at this radius turns out to be . So, even if constraints on isothermal waves are not as tight as in the case of NGC 1275 due to larger uncertainties, it is possible for them to be the source of the ISM disturbances we observe in NGC 4649.
3.2 Non-thermal pressure component
To further investigate the origin of the observed ISM disturbances, we performed an analysis similar to that proposed by Humphrey et al. (2013), confronting the mass profiles obtained from the hot gas profiles presented in Figure 7 through the hydrostatic equilibrium equation and the ones obtained from independent diagnostics such as stellar kinematics, globular clusters (GCs) and planetary nebulae (PNe) velocities (Shen & Gebhardt, 2010; Deason et al., 2012). The profiles are presented in Figure 8 (left panel), the hot gas-derived profile is in light blue and that from stellar and GC kinematics in red (Shen & Gebhardt, 2010), with the strips width representing the relative uncertainties. In the outer regions the mass profiles from Deason et al. (2012) are plotted in green (GCs) and dark blue (PNe).
At larger radii we notice a significant discrepancy between the mass profile derived by Shen & Gebhardt (2010) using stellar (Gebhardt et al., 2003) and GC (Hwang et al., 2008) kinematics, and the mass profiles obtained by Deason et al. (2012) using PNe (Teodorescu et al., 2011) and GC (same dataset as Shen & Gebhardt) kinematics assuming spherical symmetry. We note that Deason et al. report that PNe are more centrally concentrated and are on more radial orbits with respect to the total GC population, while the red GC subpopulation is similar to that of both the stars and PNe. These discrepancies have been discussed by Teodorescu et al. (2011), Das et al. (2011) and Coccato et al. (2013), who suggested that GC an PNe may represent dynamically distinct systems. In addition, the analysis of stellar kinematics observation from the SLUGSS survey by Arnold et al. (2013) shows strong evidence for NGC 4649 of a fast rotating embedded disk structure, pointing to this galaxy as a candidate for a major merger remnant. Finally, a recent analysis by D’Abrusco et al. (2013) revealed significant anisotropies in the two-dimensional distribution of GCs in NGC 4649, in particular in red and blue GC populations, that may affect these mass profiles. However, we note that the GC anisotropies revealed by D’Abrusco et al. (2013) are on a bigger scale than the ISM disturbances studied here.
In the inner radii, the mass profile from X-ray emitting hot gas and that from stellar and GC kinematics differ significantly between 0.5 and 3 kpc, again the same scales of the structures shown in Figure 4. In particular, the blue profile has been obtained using the equation
where is the radius, is the gas temperature, is the gas densities, the proton mass and is the average molecular weight factor. A non-thermal pressure term can, however, be added to Eq. 1, that is
Assuming that the stellar kinematics profile accounts for the total mass expressed by Eq. 2, we can estimate the non-thermal pressure term writing
and can be obtained by numerically integrating Eq. 3. In Figure 8 (right panel) we show the profile of the ratio between the non-thermal pressure and the average gas pressure obtained in the annular fit, which is consistent with that obtained by Humphrey et al. (2013). For comparison, in the same panel, we plot the ratio of gas pressure (i.e., excess and deficit in Figure 6b) in the same regions shown in Figure 5. Adopting the radio lobe parameters reported by Shurkin et al. (2008) we also computed the minimum radio pressure assuming energy equipartition between particles and magnetic field in the jet (see, e. g., Hardcastle et al., 2004), and plotted it for N and S lobes in Figure 8 (right panel) as black squares. Although there are large uncertainties in these estimates, the radio jet pressure appears comparable with the thermal gas pressure at similar radii.
We note that the non thermal pressure can account on average for of the gas pressure in the regions of significant ISM disturbance, and that there is a striking correlation between peaks and dips in the two pressure trends. A simple cross-correlation test between the two profiles yields that without any lag the correlation is significant at the , indicating a correlation between the two profiles. This links the non-thermal pressure to the nuclear radio source and its jets. Humphrey et al. (2013) advanced several possibilities for the non-thermal pressure component observed in NGC 4649, including gas rotation, random turbulence, magnetic field and cosmic ray pressure. Our analysis strongly correlates the non-thermal pressure with the radio jet structure, and therefore we conclude that cosmic ray injection into the ISM from the weak radio jets is the most likely origin of this pressure component.
3.3 Black Hole Mass determination
To investigate the effects of the non-thermal pressure on the SMBH mass estimate we fitted the mass profile derived with Eq. 1 with the four standard galaxy mass components, that is, the gas component shown in Figure 7, the dark matter described by a standard NFW profile , the stellar mass obtained from the observed V-band luminosity (Kormendy et al., 2009) (with a free to vary ), and the central SMBH. The result of the fit procedure is presented in Figure 9. The best fit parameters are , and . These yield a SMBH mass estimate , compatible with - but nominally larger than - the SMBH mass estimate of by Shen & Gebhardt (2010). Repeating the fitting procedure using the dark matter logarithmic profile proposed by Shen & Gebhardt (2010), we obtained a similar value of but an even larger . Besides the details of the DM halo (more relevant in the outer regions), comparing with the mass profile form (Shen & Gebhardt, 2010) we note the main effect of the non-thermal pressure (significant in the range) on this fitting procedure is to yield higher values of , that is, a higher contribution of the stellar component to the total mass, and as a consequence a lower SMBH mass. However, since the central SMBH mass is mainly driven by the mass profile in innermost radii, its estimate does not appear to be strongly affected by the presence of the non-thermal pressure component. We note, however, that our SMBH mass estimate is higher than the value obtained by Humphrey et al. (2008) using shallower Chandra X-ray data. This discrepancy is explained with the higher gas temperature we observe with respect to Humphrey et al. (see Figure 7 and Eq. 1).
3.4 Nuclear luminosity and Bondi accretion
To evaluate the nuclear luminosity we extracted the Chandra-ACIS spectrum of the inner , adding to the model described in Sect. 2.2 a power-law component with photon index fixed to 2 in order to model the AGN emission (e.g., Gallo et al., 2010). The nuclear luminosity within is therefore , , and . Given the mass of the SMBH, this X-ray nuclear emission corresponds to a very sub-Eddington bolometric emission. In fact, for a large sample of nearby low luminosity AGNs, the median bolometric correction is (Ho, 2008); Merloni & Heinz (2007) adopt for their study of 15 low luminosity AGNs; Vasudevan & Fabian (2007) find for low-luminosity sources. The latter choice gives erg s, and then . From an estimate of the mass accretion rate , we can next discuss the accretion modalities.
The simplest assumption for gas accretion is that it is steady, spherically symmetric, with negligible angular momentum, as in the theory developed for gas accreting onto a point mass (Bondi, 1952). The accretion rate then comes from the gas density, temperature, and the SMBH mass, for which an accurate estimate is available for NGC 4649 (, Shen & Gebhardt 2010). Ideally, one should insert in the calculation the density and temperature at , where the dynamics of the gas start to be dominated by the potential of the SMBH ( is the ambient sound speed; Frank et al. 2002). In practice, one uses fiducial temperature and density for the circumnuclear region, determined as close as possible to the SMBH. In our case, the gas properties are derived reasonably close to the black hole, thus giving a estimate far better than usual (e.g. Pellegrini, 2005; Russell et al., 2013). Inserting in the temperature ( keV) at the innermost bin, that extends out to 200 pc, gives pc (for the polytropic index of the adiabatic case), or pc (, isothermal case). For the gas density of the innermost bin, g cm, the Bondi mass accretion rate is yr for (as adopted hereafter; would be 4.5 times larger if ; Frank et al. 2002). Taking into account the uncertainties on gas density, temperature and SMBH mass, and maximizing their effect on the computation of , the resulting range is yr. The accretion power is then erg s.
The nucleus of NGC 4649 is extremely radiatively inefficient, with (as already noticed Pellegrini, 2005); our estimate of the accretion rate allows to establish the accretion modalities. In fact , a very low value well within the radiatively inefficient accretion (RIAF) regime, that can take place when (Narayan & Yi, 1995). RIAF models are the viscous rotating analog of the spherical Bondi accretion, with an efficiency for producing radiation of . The expected is then erg s, that is times larger than erg s. Adopting instead a bolometric correction factor specific for the spectral energy distribution of a RIAF, i.e., (Mahadevan, 1997), then erg s, which is times lower than the predicted . Reductions of the mass accretion rate with respect to the mass available at large radii (i.e., ) on the way to the SMBH are possible, since RIAF solutions include cases of outflows or convective motions (Blandford & Begelman, 1999); another source of reduction is given by the possibility that the gas has non-negligible angular momentum at the Bondi radius (Proga & Begelman, 2003; Narayan & Fabian, 2011). The latter authors calculated the rate at which mass accretes onto a SMBH from rotating gas, for plausible RIAF solutions for galactic nuclei, and found that . Indeed, the stellar component of NGC 4649 is known to possess a significant rotation (Pinkney et al., 2003). If , the predicted goes down to erg s, which is now larger than by a factor between one hundred and six hundreds, considering the range for .
In conclusion, the accretion flow seems to emit less than predicted from the fuel observed to be available, even when allowing for a RIAF with angular momentum at the outer radius of the accretion flow (the various powers and luminosities considered here are summarized in Table 3); possible solutions could be that the bolometric correction is larger, so that becomes larger (this is not expected to fix entirely the problem, though), or that the accretion flow should be modeled differently (i.e., with an even lower radiative efficiency), or that outflows/convective motions are important. We note that a very low radiation efficiency was also found in the deep Ms Chandra observations of NGC 3115 (Wong et al., 2014; Shcherbakov et al., 2014) indicating for this source either a remarkably inefficient flow or a very low accretion rate suppressed by outflow, rotational support or stellar feedback.
However, the total (observed) power output from the black hole is the sum of the radiating (e.g., ) and mechanical powers, where the latter by far dominates usually in the low luminosity nuclei of the local universe. Therefore, a further important constraint on how much mass must be accreting comes from , and is examined below.
3.5 Jet power and total accretion output
Pointed VLA observations by Shurkin et al. (2008) showed the radio properties at 1.4 GHz (see Fig. 3), and their connection with the cavities in the hot ISM (from a shallower 37 ksec Chandra exposure). Shurkin et al. derived a total flux density of Jy at 1.4 GHz; for a distance of 16 Mpc, this gives a radio power of erg s Hz, that is erg s. Thus NGC 4649 is a low-power radio source, one of the least radio emitting giant ellipticals with respect to those in the samples that produced the correlations between , , and radio luminosity mentioned above; it is useful then to establish whether extrapolation to lower luminosity sources is applicable.
Following Allen et al. (2006), Shurkin et al. (2008) used the
1.4 GHz image to determine the edges of the cavities, and then calculate the
kinetic energy of the jet, as the sum of their internal energy and the PdV
work done to inflate them (as ). Since the majority of the energy
carried off by these jets is mechanical, not radiative
(e.g. Merloni & Heinz, 2007), the resulting energy , when
divided by an approximate age for the cavities, gives an estimate of (a
lower limit to
We calculated the cavity power from our deeper image adopting the same method of Shurkin et al., but rather than using the radio contours as a guide we used instead the X-ray residual map to delineate the cavities (see Fig. 3). These are approximated as ellipsoids centered at distance from the SMBH and with semi-axes along the radio jet direction ( east of north) and across it (the semi axis along the line of sight is taken equal to ). The regions shown in Fig. 3 have , , and , , for the N and S cavities, respectively, with a uncertainty on these sizes. We evaluated an energetic content and for the north and south cavity, respectively. The sound-speed expansion time has been evaluated as , where (Bîrzan et al., 2004), yielding and for the N and S cavity, respectively. Buoyancy rise times are evaluated as , where . Using from Eq. 1 we obtain and for the N and S cavity, respectively.
We then obtained for the two N and S cavities, for the sound speed expansion time erg s and erg s, and for the buoyancy rise time erg s and erg s. The two estimates are compatible between the errors, and give a total erg s (for the buoyancy rise time), and erg s (for the sound speed expansion time). Note that, as mentioned above, this may not be the whole of the mechanical power injected in the hot ISM; shocks are not seen here, though, possibly sound waves. Moreover, in addition to the larger cavities close to the ends of the jets, there are other smaller cavities seen in the X-ray gas (as well as other smaller radio structures, see Fig. 2 of Shurkin et al.). The total of is then such that .
In the correlation found by Allen et al. (2006) between and for a sample of elliptical galaxies, on average , a much larger fraction than found here. For NGC 4649, that has equal to the largest one in the Allen et al.’s sample, a larger is predicted from our by their correlation: erg s (with erg s the lowest value allowed by all the uncertainties on and on the correlation coefficients), about 100 times larger the jet power we estimate above. The Allen et al.’s relation implies that a significant fraction of emerges as kinetic energy, which requires that a significant fraction of the mass entering flows all the way down to the SMBH, with little mass loss (in outflows) along the way. This seems not to be the case here, consistent with the results of the previous section, where a significant reduction of with respect to was needed. A similar result is obtained if one were to infer the accretion rate from the “observed” erg s. For the RIAF low radiative efficiency, one would derive yr, that is . Such a value for gives erg s, which can still account for the measured (and is much closer to it, ). It seems so that the same reduction in required by the (far smaller) nuclear luminosity could explain also the (far larger) accretion output observed in mechanical form. In addition we note that in principle is an average over the yr age of the cavities, while is a measurement of the instantaneous accretion rate (since the flow time from to the SMBH is very small, of the order of yr). Thus when linking to we assume that the accretion flow has been the same over the past few yr. We note that Russell et al. (2013) were not able to reproduce Allen et al.’s relation. This is mainly due to a different method for delineating the cavities adopted in the two papers, that is, radio maps for Allen et al. and X-ray maps for Russell et al. (the same method adopted in the present paper). As a result, the trend between cavity power and Bondi power appears significantly weakened in Russell et al. analysis (see their Figure 11) and with a larger scatter due to the different profiles adopted by these authors to extrapolate the gas density profiles to the accretion radius.
Note that NGC 4649 does not display extended optical emission line regions and is depleted of molecular and atomic gas, as shown by Young (2002), that put an upper limit of to the mass content of this galaxy. If we assume that the AGN mechanical power we evaluate in the cavities is generated by accretion at , the total accreted mass to power the jet would be , well below the upper limit evaluated by Young. However, in NGC 4649 there is (as in general for hot gas rich galaxies) a large hot gas reservoir , much larger than the cold gas upper limit. Then, on a secular scale it is more likely that accretion is taking and took place thanks to the hot gas. Even if the primary source of accretion is likely to be the hot gas, however, the low accretion rate we evaluate does not exclude molecular gas as a source of fuel for the SMBH, so cold accretion of molecular and atomic gas represents a possible alternate (or additional) scenario to Bondi accretion.
3.6 Jet power, radio and nuclear luminosity
Cavagnolo et al. (2010) re-examined the scaling relationship between jet
power, , and synchrotron luminosity at 1.4 GHz and 200-400 MHz,
incorporating measurements for 21 giant elliptical galaxies, thus expanding
the sample of Bîrzan et al. (2008), dominated by bubbles in clusters,
to lower radio power sources. The was estimated as , the
power required to inflate the cavities. The
Cavagnolo et al.’s relationship predicts erg s from of NGC 4649, and between and erg s considering the large scatter
around the correlation
The nuclear 2-10 keV luminosity of NGC 4649 also fits on the correlation between nuclear 2-10 keV luminosity and , derived for 27 detected nuclei of central dominant galaxies (Russell et al. 2013; see also Merloni & Heinz 2007). The majority of the sample is in a RIAF mode, and the mechanical cavity power dominates the radiative output. NGC 4649 sits at the lowest values of the correlation, with , from which we can predict erg s (see Figure 10, right panel). Even though also in this correlation the scatter is large, it was taken as evidence that the radiative efficiency of the X-ray nucleus increases with increasing , until the quasar regime is reached (where the nuclear luminosity becomes comparable to ).
4 Summary and Conclusions
We investigated the presence of AGN feedback in the ISM of the giant elliptical NGC 4649 by using a total of 280 ks Chandra observations This source has been studied several times in different wavelength, and in particular in the X-rays (e.g. Churazov et al., 2010; Das et al., 2010; Loewenstein & Davis, 2012). Shurkin et al. (2008) and Humphrey et al. (2008), making use of Chandra observations, studied the properties of the ISM in NGC 4649 using the unsharp-mask technique looking for morphological disturbances pointing to deviations from the hydrostatic equilibrium condition suggested by the generally relaxed X-ray morphology. Interestingly, while the former authors using a shallow observation found evidences of structures and cavities in the ISM that they interpreted as connected with the central, faint radio source, the latter authors using deeper data did not found any evidence of such disturbances. A subsequent analysis by Dunn et al. (2010) showed disturbances and cavities in the ISM as residuals of the X-ray surface brightness from a spherical model, connected with the radio emission.
Using much deeper Chandra data with a total exposure , we used the latter approach to investigate the morphological distribution of the ISM in NGC 4649. We studied the deviation of the X-ray surface brightness from an elliptical model, which is expected to describe the hot gas distribution in relaxed galaxies. The residuals of this fitting procedure, presented in Figure 4, show significant cavities, ripples and ring like structures on the inner scale. This is at variance with the significance of the cavities reported by Shurkin et al. (2008) as evaluated by Humphrey et al. (2008). The deeper data considered here revealed these structures with high significance; moreover the structures appear to be morphologically related with the central radio emission, with cavities lying in correspondence with the extended radio lobes and regions of enhanced emission situated on the side of them and, on larger scale, taking the form of ring like ripples which seems reminiscent of the structures observed in NGC 1275 (Fabian et al., 2006). In common with this source, we found no significant temperature variations in correspondence with higher pressure regions. So, if radio ejecta driven shocks are responsible for the observed ISM morphology, the observed structures may be isothermal waves whose energy is dissipated by viscosity, with thermal conduction and sound waves effectively distributing the energy from the radio source. Evidences of deviations from the hydrostatic equilibrium are also provided by the mass profiles presented in Figure 8 (left panel). A significant non-thermal pressure is found on the same scale of the residual structures, where it reaches of the observed gas pressure. In addition, the excess gas pressure and non-thermal pressure profiles appear to be strongly correlated, indicating the radio ejecta as the likely origin for this additional pressure component.
At smaller scales, similarly to a few other early type galaxies harboring low power radio sources, NGC 4649 shows increased temperatures in the inner region. The nucleus of NGC 4649 appears to be extremely sub-Eddington, with the accretion flow emitting less than predicted from the fuel observed to be available, even when allowing for a RIAF with angular momentum at the outer radius of the accretion flow. Also the jet power evaluated from the observed X-ray cavities appears to be much smaller than that predicted for elliptical galaxies from the Bondi accretion power . If the mass accretion rate accounting for the observed nuclear X-ray luminosity is adopted - which requires, in addition to a low radiative efficiency, a significant reduction of the accretion rate with respect to the Bondi value, due, e.g., to outflows/convective motions - then the corresponding accretion power is times larger than the observed kinetic power. When comparing the jet power to radio and nuclear X-ray luminosity, on the other hand, the observed cavities show similar behavior to those of other giant elliptical galaxies.
|Obs ID||Net Exposure (ksec)||Date||PI Name|
|Region||Net Counts (error)||Residuals sign.||(d.o.f.)|
|A||7659 (88)||9||1.06 (90)|
|C||3865 (65)||6||1.00 (79)|
|D||5840 (77)||6||1.08 (76)|
|E||14565 (121)||7||1.08 (111)|
|F||8125 (93)||11||0.96 (102)|
|G||11657 (117)||10||1.07 (161)|
|Bondi accretion power|
|RIAF luminosity + ang. mom.||-|
|RIAF bolometric luminosity|
|Cavity power||buoyancy rise|
|”||sound speed expansion|
- slugcomment: version August 22, 2019
- We assume a distance to NGC 4649 of (). At this distance corresponds to .
- http://cda.harvard.edu /chaser
- Shocks and sound waves have also been found to contribute significantly to the power output of the central AGN (e.g. Fabian et al., 2006; Forman et al., 2007; Randall et al., 2011).
- The relation shows a large scatter, possibly due to the fact that the radio emission from lobes depends on their composition, and processes such as gas entrainment, shocks and aging. In the case of NGC 4649 we do not see evidence of shocks. Shurkin et al. investigated the particle content of the cavities by determining k/f, where k is the ratio of the total particle energy contained in the cavity to the energy accounted for by electrons emitting synchrotron radiation in the range of 10 MHz to 10 GHz, and f is the volume filling factor of the relativistic plasma in the cavity. In general, k/f is in the range of to for cavities that are active, while the values for NGC 4649 are very large (). This could be explained by entrained particles, or electrons that have aged, which is likely given the small radio flux; NGC 4649 could soon become a ghost cavity system.
- Allen, S. W., Dunn, R. J. H., Fabian, A. C., Taylor, G. B., & Reynolds, C. S. 2006, MNRAS, 372, 21
- Anders, E., & Grevesse, N. 1989, Geochim. Cosmochim. Acta, 53, 197
- Arnaud K. A., 1996, ASPC, 101, 17
- Arnold, J. A., Romanowsky, A. J., Brodie, J. P., et al. 2013, arXiv:1310.2607
- Bîrzan, L., Rafferty, D. A., McNamara, B. R., Wise, M. W., & Nulsen, P. E. J. 2004, ApJ, 607, 800
- Bîrzan, L., McNamara, B. R., Nulsen, P. E. J., Carilli, C. L., & Wise, M. W. 2008, ApJ, 686, 859
- Blandford, R. D., & Begelman, M. C. 1999, MNRAS, 303, L1
- Bondi, H. 1952, MNRAS, 112, 195
- Brighenti, F., Mathews, W. G., Humphrey, P. J., & Buote, D. A. 2009, ApJ, 705, 1672
- Buote, D. A., & Tsai, J. C. 1995, ApJ, 452, 522
- Cavagnolo, K. W., McNamara, B. R., Nulsen, P. E. J., et al. 2010, ApJ, 720, 1066
- Churazov, E., Tremaine, S., Forman, W., et al. 2010, MNRAS, 404, 1165
- Coccato, L., Arnaboldi, M., & Gerhard, O. 2013, MNRAS, 436, 1322
- Condon, J. J., Cotton, W. D., & Broderick, J. J. 2002, AJ, 124, 675
- Croton, D. J., Springel, V., White, S. D. M., et al. 2006, MNRAS, 365, 11
- D’Abrusco, R.. 2013, ApJ, in press
- Das, P., Gerhard, O., Churazov, E., & Zhuravleva, I. 2010, MNRAS, 409, 1362
- Das, P., Gerhard, O., Mendez, R. H., Teodorescu, A. M., & de Lorenzi, F. 2011, MNRAS, 415, 1244
- Deason, A. J., Belokurov, V., Evans, N. W., & McCarthy, I. G. 2012, ApJ, 748, 2
- Diehl, S., & Statler, T. S. 2007, ApJ, 668, 150
- Di Matteo, T., Springel, V., & Hernquist, L. 2005, Nature, 433, 604
- Dunn, R. J. H., Allen, S. W., Taylor, G. B., et al. 2010, MNRAS, 404, 180
- Ebeling, H., White, D. A., & Rangarajan, F. V. N. 2006, MNRAS, 368, 65
- Fabian, A. C., Sanders, J. S., Ettori, S., et al. 2000, MNRAS, 318, L65
- Fabian, A. C., Sanders, J. S., Allen, S. W., et al. 2003, MNRAS, 344, L43
- Fabian, A. C., Sanders, J. S., Taylor, G. B., et al. 2006, MNRAS, 366, 417
- Forman, W., Nulsen, P., Heinz, S., et al. 2005, ApJ, 635, 894
- Forman, W., Jones, C., Churazov, E., et al. 2007, ApJ, 665, 1057
- Frank, J., King, A., & Raine, D. J. 2002, Accretion Power in Astrophysics, by Juhan Frank and Andrew King and Derek Raine, pp. 398. ISBN 0521620538. Cambridge, UK: Cambridge University Press, February 2002.,
- Fruscione, A., McDowell, J. C., Allen, G. E., et al. 2006, Proc. SPIE, 6270
- Gallo, E., Treu, T., Marshall, P. J., et al. 2010, ApJ, 714, 25
- Garmire, G. P. 1997, Bulletin of the American Astronomical Society, 29, 823
- Gastaldello, F., & Molendi, S. 2002, ApJ, 572, 160
- Gebhardt, K., Richstone, D., Tremaine, S., et al. 2003, ApJ, 583, 92
- Hardcastle, M. J., Harris, D. E., Worrall, D. M., & Birkinshaw, M. 2004, ApJ, 612, 729
- Ho, L. C. 2002, ApJ, 564, 120
- Ho, L. C. 2008, ARA&A, 46, 475
- Hickox, R. C., & Markevitch, M. 2006, ApJ, 645, 95
- Humphrey, P. J., Buote, D. A., Brighenti, F., Gebhardt, K., & Mathews, W. G. 2008, ApJ, 683, 161
- Humphrey, P. J., Buote, D. A., Brighenti, F., Gebhardt, K., & Mathews, W. G. 2013, MNRAS, 430, 1516
- Hwang, H. S., Lee, M. G., Park, H. S., et al. 2008, ApJ, 674, 869
- Irwin, J. A., Athey, A. E., & Bregman, J. N. 2003, ApJ, 587, 356
- Kalberla P. M. W., Burton W. B., Hartmann D., Arnal E. M., Bajaja E., Morras R., Pöppel W. G. L., 2005, A&A, 440, 775
- Kim, D.-W., Fabbiano, G., & Pipino, A. 2012, ApJ, 751, 38
- King, A. 2013, Space Sci. Rev., 85
- Kormendy, J., Fisher, D. B., Cornell, M. E., & Bender, R. 2009, ApJS, 182, 216
- Kritsuk, A., Bohringer, H., & Muller, E. 1998, MNRAS, 301, 343
- Loewenstein, M., & Davis, D. S. 2012, ApJ, 757, 121
- Luo, B., Fabbiano, G., Strader, J., et al. 2013, ApJS, 204, 14
- Machacek, M., Nulsen, P. E. J., Jones, C., & Forman, W. R. 2006, ApJ, 648, 947
- Mahadevan, R. 1997, ApJ, 477, 585
- McNamara, B. R., & Nulsen, P. E. J. 2007, ARA&A, 45, 117
- Merloni, A., & Heinz, S. 2007, MNRAS, 381, 589
- Narayan, R., & Yi, I. 1995, ApJ, 452, 710
- Narayan, R., & Fabian, A. C. 2011, MNRAS, 415, 3721
- Ostriker, J. P., Choi, E., Ciotti, L., Novak, G. S., & Proga, D. 2010, ApJ, 722, 642
- O’Sullivan, E., Giacintucci, S., David, L. P., et al. 2011, ApJ, 735, 11
- Paggi, A., et al. 2014, in prep.
- Pellegrini, S., Baldi, A., Fabbiano, G., & Kim, D.-W. 2003, ApJ, 597, 175
- Pellegrini, S. 2005, ApJ, 624, 155
- Pellegrini, S. 2010, ApJ, 717, 640
- Pellegrini, S., Wang, J., Fabbiano, G., et al. 2012, ApJ, 758, 94
- Pinkney, J., Gebhardt, K., Bender, R., et al. 2003, ApJ, 596, 903
- Proga, D., & Begelman, M. C. 2003, ApJ, 592, 767
- Randall, S. W., Sarazin, C. L., & Irwin, J. A. 2004, ApJ, 600, 729
- Randall, S. W., Sarazin, C. L., & Irwin, J. A. 2006, ApJ, 636, 200
- Randall, S. W., Forman, W. R., Giacintucci, S., et al. 2011, ApJ, 726, 86
- Russell, H. R., McNamara, B. R., Edge, A. C., et al. 2013, MNRAS, 432, 530
- Sazonov, S. Y., Ostriker, J. P., Ciotti, L., & Sunyaev, R. A. 2005, MNRAS, 358, 168
- Shcherbakov, R. V., Wong, K.-W., Irwin, J. A., & Reynolds, C. S. 2014, ApJ, 782, 103
- Shen, J., & Gebhardt, K. 2010, ApJ, 711, 484
- Shurkin, K., Dunn, R. J. H., Gentile, G., Taylor, G. B., & Allen, S. W. 2008, MNRAS, 383, 923
- Sijacki, D., Springel, V., Di Matteo, T., & Hernquist, L. 2007, MNRAS, 380, 877
- Soria, R., Fabbiano, G., Graham, A. W., et al. 2006, ApJ, 640, 126
- Stanger, V. J., & Warwick, R. S. 1986, MNRAS, 220, 363
- Teodorescu, A. M., Méndez, R. H., Bernardi, F., et al. 2011, ApJ, 736, 65
- Vasudevan, R. V., & Fabian, A. C. 2007, MNRAS, 381, 1235
- Wong, K.-W., Irwin, J. A., Yukita, M., et al. 2011, ApJ, 736, L23
- Wong, K.-W., Irwin, J. A., Shcherbakov, R. V., et al. 2014, ApJ, 780, 9
- Young, L. M. 2002, AJ, 124, 788