CHANDRA REVEALS HEAVY OBSCURATION AND CIRCUMNUCLEAR STAR FORMATION IN SEYFERT 2 GALAXY NGC 4968

Active galactic nuclei (AGNs) are powered by accreting supermassive black holes in galactic centers. A significant portion of this black hole growth is obscured by gas and dust (e.g., Fabian 1999; Risaliti et al. 1999; Treister et al. 2004, 2009), which according to the AGN unification model, resides on ∼1–10 pc scales in a toroidal geometry (Antonucci 1993; Urry & Padovani 1995). According to this model, Type 2 AGNs lack broad lines in their optical spectra due to the observer line of sight intersecting the obscuring torus, while Type 1 AGNs are unobscured, allowing both broad emission lines and the AGN continuum to be visible.

Compton-thick AGNs, in which the column density (N_H) of the obscuring material becomes optically thick to Compton scattering (≥1.25 × 10²⁴ cm⁻²),⁸ appear to represent a significant fraction of cosmic black hole growth, though the percentage of this population is a matter of some debate (e.g., Treister & Urry 2005; Gilli et al. 2007; Treister et al. 2009; Ueda et al. 2014; Aird et al. 2015; Buchner et al. 2015; Ricci et al. 2015; Akylas et al. 2016). Hence, to constrain population synthesis models and accurately determine the accretion history of cosmic black hole growth, it is important to have a proper census of Compton-thick AGNs. These heavily obscured sources can be identified by signatures in their X-ray spectra, with the fluorescent Fe Kα line emission at 6.4 keV being perhaps the most notable feature (Krolik & Kallman 1987; Ghisellini et al. 1994; Matt et al. 1996). Detailed study of individual Compton-thick AGN then provides an opportunity to probe AGN physics. Furthermore, the suppression of optical AGN emission allows us to study the host galaxy in greater detail than is possible with unobscured AGNs since galaxy features are not swamped by AGN light. Circumnuclear star formation may even play a role in obscuring some AGNs (e.g., Ballantyne 2008).

NGC 4968 is a nearby (z = 0.00986) Type 2 Seyfert (Sy2) identified in the IRAS 12 μm sample (Spinoglio & Malkan 1989), hosted in an Sa galaxy (Malkan et al. 1998) with a WISE measured 12 μm luminosity of 1.7 × 10⁴³ erg s⁻¹ (Wright et al. 2010). Our X-ray analysis of this source with XMM-Newton revealed a strong Fe Kα line with an equivalent width (EW) of ${3.10}_{-0.8}^{+1.0}\,\mathrm{keV}$ (LaMassa et al. 2011; see also Guainazzi et al. 2005). This extreme value indicates that the AGN in NGC 4968 is embedded within Compton-thick material: the total continuum from the transmitted and reflected components is suppressed relative to the fluorescent line photons, which can be created anywhere throughout the obscuring medium. However, using a phenomenological model of an absorbed power law plus a Gaussian component to fit the AGN continuum and Fe Kα line, respectively, we did not find an absorbing column larger than that from our Galaxy. Excess emission at soft energies (0.5–2 keV) prevented the column density from being accurately measured with this overly simplistic model. Is this soft X-ray emission due to the host galaxy, scattered AGN emission, or both? Levenson et al. (2002) noted that Sy2s with large Fe Kα EW values (>2 keV), like NGC 4968, reside in galaxies with on-going star formation, so it is likely that star formation contributes to the soft X-ray emission in NGC 4968.

With the quality and resolution of the XMM-Newton data, we were unable to fit self-consistent, physically motivated models (e.g., MYTorus, Murphy & Yaqoob 2009; spherical and toroidal absorption models of Brightman & Nandra 2011) to the spectra, nor could we attempt to disentangle nuclear from unresolved galaxy light. Therefore, we obtained a deep (50 ks) Chandra observation in Cycle 16 in order to physically diagnose the conditions at the center of this galaxy. We present our results in this paper as follows. In Section 2, we detail the reduction of the Chandra data. We discuss the imaging and spectral analysis of NGC 4968 in Sections 3 and 4, respectively. From these results, we comment on the nature of the X-ray obscuring medium, estimate the star-formation rate (SFR) of the host galaxy, and propose that the gas responsible for circumnuclear star formation increases the covering fraction of the obscuration toward unity, such that the AGN is almost completely enshrouded within Compton-thick material (Section 5). Throughout, we use a cosmology where H₀ = 67.8 km s⁻¹ Mpc⁻¹, and Ω_M = 0.31, Ω_Λ = 0.69 (Planck Collaboration et al. 2015).

NGC 4968 was observed for 50 ks on 2015 March 9 with ACIS-S (ObsID: 17126; PI: LaMassa). The data were reduced with CIAO v4.8, with CALDB v. 4.7.1. We used CIAO task chandra_repro to produce a filtered level 2 events file with the latest calibration files.

From this events file, we extracted images of NGC 4968 in the soft (0.5–2 keV) and hard (2–10 keV) bands, as shown in Figure 1 (top), where the location of the narrow line region (Schmitt et al. 2003), and slightly elongated radio emission (Schmitt et al. 2001) are noted by the solid blue and dashed magenta lines, respectively.⁹ As can be seen via visual inspection of the images, the source is more spatially extended in the soft band compared with the hard band (see Section 3 for a full treatment of the spatial analysis). For reference, we show the Hubble Space Telescope WFPC2 image of NGC 4968 (Malkan et al. 1998) in the F606W filter in the bottom panel of Figure 1, overplotted with X-ray contours from the soft (red) and hard (cyan) bands. While the emission at softer energies has a larger spatial extent, it is confined to the bulge of the host galaxy (at least in projection).

Zoom In Zoom Out Reset image size

We used CIAO routine specextract to extract the spectrum from the full X-ray emitting region as well as the nuclear and extended components separately. For the global modeling, the spectrum was extracted from a circular aperture with radius 6'' to ensure all extended emission was captured. To isolate the nuclear emission, we ran wavdetect on the hard energy band, extracting the spectrum from the output region file centered on the AGN (an ellipse with major axis 186, minor axis 16 and angle 1374, measured counter-clockwise from West to East in WCS coordinates). This nuclear region was excised from the global region to produce a spectrum of the extended-only emission (as noted in Section 3 the estimated PSF at the source location is smaller than the nuclear region). In all cases, the background was extracted from an annulus centered on the source with an inner radius of 10'' and outer radius of 30''. The spectra were grouped by 5 counts per bin; visual inspection of the spectra demonstrates that the chosen binning does not over sample the resolution and thus we lose no information with this binning.

From Figure 1, the X-ray emission appears more extended in the soft band than in the hard band, and lies in a direction not coincident with the narrow line region (Schmitt et al. 2003) nor radio structure (Schmitt et al. 2001). To test whether the X-ray emission is extended and to quantify the size of this region, we simulated a point-spread function (PSF) at the position where the source was detected. Since the Chandra PSF varies as a function of energy as well as position, we fit the Chandra spectrum with an ad hoc model that reproduces the shape of the observed spectrum so that the energy dependence of the source can be accounted for by ChaRT, the Chandra ray tracing software (see Section 4 for a full treatment of the spectral modeling). The output ray-traced data from the simulated PSF are then input into MARX, which projects the photons onto the ACIS-S detector-plane to create a pseudo-events file. We simulated pseudo-events files for a point source separately for the soft and hard band.

CIAO tool srcextent compares the PSF with the actual data to measure the size of the PSF at the source location on the detector and to quanity whether the source is extended at the 90% confidence level. From this routine, we find that in both the soft and hard bands, the source is extended at the 90% confidence level. The observed size of the PSF at the location of the source is 069, where the PSF is measured as $\sqrt{(1/2)({a}^{2}+{b}^{2})}$ with a and b as the semimajor and semiminor axes of an elliptical Gaussian.

To measure the extent of the X-ray emission, we compared the surface brightness of the source with what would be expected from a point source: if the surface brightness of the source exceeds that of a point source over a range of distances, then we can confirm that the source is extended and measure this spatial extent. We perform this exercise separately for the soft and hard bands. In each band, we extracted radial surface brightness profiles from the soft and hard bands by calculating the background-subtracted net counts per pixel squared within a series of annuli around the nucleus. The radius of the inner annulus was set to 1'' and the subsequent annuli were incremented in steps of 05.

In Figure 2, we compare the radial surface brightness profiles from the data with that of the simulated point source. Figure 2 demonstrates that the soft X-ray emission is significantly extended up to ∼4'' while the hard X-ray emission is more compact, yet extends up to ∼225. These sizes correspond to physical scales of 900 pc and 500 pc, respectively. As discussed in Section 5.3, the size scale, and direction of the extended hard emission is consistent with the narrow line region measured via the [O iii] ionization cone (Schmitt et al. 2003), while the soft emission appears to be independent of the narrow line region.

Zoom In Zoom Out Reset image size

The Chandra spectrum was fitted with XSpec version 12.9.0 (Arnaud 1996). We used the Cash-statistic for fitting, which is more appropriate in the low count regime than χ² since it only requires that the count distribution be Poissonian (Cash 1979). The energy range was restricted to be between 0.5 and 8 keV. Errors on the fit parameters represent the 90% confidence interval for one interesting parameter.

4.1. Global Fit

4.1.1. Phenomenological Fit

We begin by fitting the Chandra spectrum with a phenomenological model to measure the EW of the Fe Kα feature because this provides insight into the geometry of the obscurer. A simple absorbed power law fails to fit the spectrum, requiring a more complex model. Here, we adopt a double absorbed power-law model, which describes a partial covering geometry where a portion of the intrinsic AGN continuum is scattered into our line of sight or leaks through the circumnuclear obscuration, while the rest is absorbed. To this model, we add a Gaussian component, at the redshift of the source, to accomodate the Fe Kα emission and a thermal emission model (apec) to best fit the soft emission (see below). One absorption component (N_H) is frozen to the Galactic value of 9 × 10²⁰ cm⁻² and an additional component is left free to attenuate the the transmitted AGN emission. This fit to the Chandra spectrum is shown in Figure 3, where we find N_H = 1.8${}_{-0.9}^{+5.0}\times {10}^{23}$ cm⁻² and a power-law slope (Γ) of ${1.69}_{-0.51}^{+0.47}$.

Zoom In Zoom Out Reset image size

In Figure 4 (top left), we show a close-up of the Fe K region and note that this model cannot completely fit the line, despite the line being spectrally unresolved (i.e., the best-fit σ pegged at our imposed lower limit of 0.01 keV, which is several times below the approximate Chandra resolution of ∼0.06 keV); this could be due to poor photon statistics in the line region, causing the residuals to be more narrow than the spectral line resolution. We measured the Fe Kα EW against the total continum, finding a value of 2.9 keV in the rest-frame.

Zoom In Zoom Out Reset image size

We also perform a local fit (3.5–7.8 keV) for a more detailed investigation of the Fe Kα EW. Here, the spectrum is modeled with a power law and Gaussian component. Taking into account errors on the local power-law slope and normalization, and width and normalization of the Gaussian model (the line center was well measured at 6.38 keV and thus was frozen), we calculated Δχ² as a function of Fe Kα EW in Figure 5; the horizontal line denotes the 90% confidence level (Δχ² = 7.779 for four free parameters). From this exercise, we find the (rest-frame) Fe Kα $\mathrm{EW}={2.5}_{-1.0}^{+2.6}\,\mathrm{keV}$.

Zoom In Zoom Out Reset image size

This large EW is quite telling. To obtain such large values, the transmitted emission must be attenuated more than the fluorescent line emission produced within the medium, which is possible when the column density reaches Compton-thick levels of obscuration. The same effect could be induced by variability, where the AGN continuum weakens and the travel time to the Fe Kα emitting region causes a delay in the response to the EW. As we discuss further in Section 5.1, archival X-ray observations of this source indicate that the flux and EW have remained constant over time (20 and 15 years, respectively), such that this scenario seems unlikely. The measured EW thus implies a higher N_H than what is measured, indicating that another continuum component must dominate at energies below the Fe Kα line. Additionally, as discussed in Murphy & Yaqoob (2009), the Fe Kα EW is sensitive to the inclination angle of the obscuring medium when the line of sight intersects the torus, with the most extreme EW values (i.e., >2 keV) obtained when the torus is completely edge-on at 90° for the geometry assumed in the MYTorus model, i.e., a torus with a fixed 60° opening angle.

4.1.2. Toroidal Obscuration

To obtain a physically motivated fit to the NGC 4968 spectrum, we use the MYTorus model, which self-consistently treats the attenuated transmitted continuum (modeled as a power law), reflection spectrum, and Fe Kα and Kβ fluorescence emission (Murphy & Yaqoob 2009). To preserve the self-consistency of the model, the normalization and Γ of the power law, the inclination angle of the torus, and the column density (N_H) are tied together among the MYTorus model components. Motivated by our phenomenological modeling of the spectrum and resulting Fe Kα EW measurement, we freeze the inclination angle to 90°.

The data require an additional power-law component to properly model the soft emission (i.e., ΔC-Stat = 703 for a change of 1 in degrees of freedom (dof) between the model lacking this component and the one where it is included). We attribute this component to be due to a portion of the intrinsic AGN continuum that is scattered into or directly enters our line of sight after leaking through the torus: the power-law normalization and Γ are tied to the MYTorus components, with a constant factor left free to account for the scattered/leaked fraction of AGN light. In Figure 6 (top left), we show this fit to the Chandra spectrum, with the model parameters summarized in the first column of Table 1. Residuals at softer energies are apparent.

Zoom In Zoom Out Reset image size

Table 1.  Spectral Fit to Global Emission

Parameter	MYTorus	MYTorus + apec	Sphere + apec
Γ	${1.57}_{-0.12}^{+0.16}$	<1.43	<1.27
NH (1024 cm−2)	${1.36}_{-0.34}^{+0.31}$	${1.54}_{-0.26}^{+0.23}$	${7.64}_{-2.37}^{+8.57}$
Power-law normalization (10−3)a	${1.81}_{-0.63}^{+0.79}$	${1.83}_{-0.63}^{+0.83}$	${2.85}_{-1.96}^{+0.23}$
fscatt (10−2)b	${1.3}_{-0.4}^{+0.8}$	${0.9}_{-0.3}^{+0.5}$	${0.45}_{-0.34}^{+0.05}$
kT (keV)	...	${0.72}_{-0.11}^{+0.12}$	${0.72}_{-0.11}^{+0.10}$
apec normalization (10−6)c	...	${6.77}_{-1.59}^{+1.68}$	${8.71}_{-1.42}^{+1.95}$
C-Stat (dof)	217.2 (106)	156.9 (104)	128.5 (104)

Notes.

^aPower-law normalization is units of photons keV⁻¹ cm⁻² s⁻¹ at 1 keV. ^bFraction of intrinsic AGN light that leaks through the circumnuclear obscuration or is scattered into the observer line of sight. ^cThe apec normalization is given as $\tfrac{{10}^{-14}}{4\pi {[{D}_{A}(1+z)]}^{2}}$ $\int {n}_{e}{n}_{h}{dV}$, where D_A is the angular diameter distance to the source in cm and n_e and n_H are the electron and hydrogen densities, respectively, in cm⁻³.

Download table as:  ASCIITypeset image

We therefore added an apec component to this model that accounts for thermal emission, putatively associated with star formation; the abundances are kept frozen at solar. The lower energy spectrum is much better modeled (ΔC-Stat = 60.3 for Δdof = 2, corresponding to P < 0.005 that the statistical improvement to the fit is due to chance). However, the Fe Kα complex is worse fit when compared with the phenomenological model, with larger fit residuals around the emission feature (Figure 4). We note that in this modeling, we used MYTorus tables, which have a power-law termination energy of 200 keV. However, when we fit the spectra with termination energies of 100 and 400 keV, which span the range of the highest and lowest values available with MYTorus, we obtain essentially identical results.

4.1.3. Spherical Absorption

To produce such an extreme Fe Kα EW, the central engine may be completely enshrouded, because differential extinction of continuum and fluorescent line photons is greater for a closed geometry such as a uniform spherical distribution, compared to an open geometry such as a torus (e.g., compare left panel of Figure 2 from Brightman & Nandra 2011 with Figure 8 from Murphy & Yaqoob 2009). We therefore fit the spectra with the spherical absorption model of Brightman & Nandra (2011).

Similar to the MYTorus model, the spherical absorption model treats the transmitted, reflected, and fluorescent line emission self-consistently. In this model, the central continuum is fully covered over 4π steradian, though we do include a scattered power-law component to account for light leaking through this obscuration, similar to the toroidal model; the percentage of escaping light is small enough to not significantly affect the self-consistency of the model. Again, we include an apec component to best fit the soft emission (ΔC-Stat = 48.2 for Δdof = 2, corresponding to P < 0.005). The bottom panel of Figure 6 shows the spherical absorption fit to the X-ray spectra of NGC 4968, with a close-up of the Fe K complex shown in Figure 4. This model provides a better fit to the data than the toroidal geometry; we note that thawing the Fe abundance does not improve the fit and leads to an abundance that is 10 times solar, which is unphysical. The fit parameters for the spherical absorption model are listed in Table 1.

4.1.4. Summary of the Chandra Spectral Modeling to the Global Emission

The spectral modeling points to the AGN in NGC 4968 being heavily obscured to Compton-thick. The implied column density of the obscurer is inconsistent between the MYTorus and spherical absorption models, ranging from ${N}_{{\rm{H}}}={1.54}_{-0.26}^{+0.23}\times {10}^{24}$ cm⁻² (marginally Compton-thick) to ${7.64}_{-2.37}^{+8.57}\times {10}^{24}$ cm⁻² (extending to extreme Compton-thick levels). Furthermore, circumnuclear star formation is on-going in the host galaxy, evidenced by the thermal nature of the soft X-ray spectrum.

We note that both model fits return a shallow spectral slope, inconsistent with the canonical Γ = 1.8–1.9 for AGNs (e.g., Kim et al. 2004; Tozzi et al. 2006). However, if we freeze the photon index to 1.9, we obtain statistically worse fits (ΔC-Stat = 53.5, 40.7 for Δdof = 1, for the MYTorus and spherical absorption models, respectively), with larger residuals around the Fe K complex than when the photon index is a free parameter. Furthermore, a hard spectrum facilitates the production of extreme Fe Kα EW values: with more photons in the continuum above the Fe K edge, more Fe Kα photons can be produced (see Murphy & Yaqoob 2009).

From these fits, we calculated the total observed soft and hard X-ray luminosities (${L}_{0.5\mbox{--}2\mathrm{keV}}$ and ${L}_{2\mbox{--}10\mathrm{keV}}$, respectively), the apec contribution to the soft X-ray luminosity (L_thermal) from which we estimate the SFR in Section 5.4, and the estimated instrinsic 2–10 keV X-ray luminosity (${L}_{2\mbox{--}10\mathrm{keV},\mathrm{int}}$) from the power-law fit parameters. These luminosities, and associated errors (which reflect the statistical errors on the power law and/or apec normalizations) are listed in Table 2. Both models return similar observed X-ray luminosities and predict intrinsic 2–10 keV luminosities that are over two orders of magnitude larger than that observed.

Table 2.  X-Ray Luminosities^a

Luminosity	MYTorus	Sphere
${L}_{0.5\mbox{--}2\mathrm{keV}}$ (observed)	${40.11}_{-0.23}^{+0.19}$	${40.10}_{-0.53}^{+0.10}$
${L}_{2\mbox{--}10\mathrm{keV}}$ (observed)	${40.91}_{-0.18}^{+0.17}$	${40.77}_{-0.51}^{+0.04}$
${L}_{\mathrm{thermal}}$	${39.67}_{-0.12}^{+0.10}$	${39.78}_{-0.08}^{+0.09}$
${L}_{2\mbox{--}10\mathrm{keV},\mathrm{int}}$	${42.45}_{-0.18}^{+0.17}$	${42.89}_{-0.51}^{+0.04}$

Note.

^aLuminosities are in log space and units of erg s⁻¹ and are rest-frame values. The quoted errors on the luminosities refer to the statistical error of the fit.

Download table as:  ASCIITypeset image

4.1.5. Clues from the Mid-infrared

The scenario where the source is completely obscured geometrically, with a small amount of leakage permitted, appears to be the most plausible for explaining the X-ray spectrum of NGC 4968: the residuals around the Fe K complex are the lowest among the self-consistent models, as would be expected if complete coverage facilitates conditions to produce an extreme Fe Kα EW. This model also predicts the highest column density.

Further clues of extreme obscuration are evidenced by the mid-infrared (MIR) emission of this source. The 12 μm luminosity of NGC 4968 as measured by WISE (Wright et al. 2010) is 1.7 × 10⁴³ erg s⁻¹. Based on the correlation found between the nuclear 12 μm luminosity and intrinsic 2–10 keV X-ray luminosity ($\mathrm{log}\left(\tfrac{{L}_{12\mu {\rm{m}}}}{{10}^{43}\,\mathrm{erg}\,{{\rm{s}}}^{-1}}\right)$ $=\,(0.33\pm 0.04)\;+(0.97\pm 0.03)$ $\mathrm{log}\left(\tfrac{{L}_{2\mbox{--}10\mathrm{keV},\mathrm{int}}}{{10}^{43}\,\mathrm{erg}\,{{\rm{s}}}^{-1}}\right);$ Gandhi et al. 2009; Asmus et al. 2015), the predicted X-ray luminosity is ∼8 × 10⁴² erg s⁻¹, fully consistent with the value we derived via fitting the X-ray spectrum with the spherical absorption model. Though the MIR–X-ray correlation can also be driven by the X-ray spectral slope and covering factor of the obscurer (Yaqoob & Murphy 2011), the MIR W1 − W2 color is 0.89 (Vega),¹⁰ consistent with the MIR emission being dominated by AGN-heated dust (W1 − W2 > 0.8 for W2 ≤ 15; Stern et al. 2012). A high quality X-ray spectrum above 10 keV from e.g., NuSTAR, would determine whether a spherical or toroidal geometry is correct, since the predicted spectral shapes above 10 keV differ between these models.

4.2. Nuclear Region

Here we capitalize on Chandra's arcsecond resolution to hone in on nuclear emission at scales of ∼17 (<400 pc). As discussed above, a model where the AGN continuum is completely enshrouded is the most likely description of this source. In addition to fitting the nuclear spectrum with the spherical absorption model, we also applied MYTorus and summarize those results in Table 3 for completeness.

Table 3.  Spectral Fit to Nuclear Emission

Parameter	MYTorus	MYTorus + apec	Sphere	Sphere + apec
Γ	<1.59	<1.44	1.55 ${}_{-0.14}^{+0.14}$	<1.18
NH (1024 cm−2)	${1.36}_{-0.33}^{+0.30}$	${1.36}_{-0.26}^{+0.26}$	${2.00}_{-1.00}^{+1.12}$	${2.8}_{-1.2}^{+13.6}$
Power-law normalization (10−3)	${1.47}_{-0.54}^{+0.74}$	${1.42}_{-0.55}^{+0.72}$	${3.62}_{-2.65}^{+6.38}$	${1.25}_{-0.75}^{+2.15}$
${f}_{\mathrm{scatt}}$ (10−2)	${1.4}_{-0.5}^{+0.9}$	${1.2}_{-0.4}^{+0.8}$	${0.64}_{-0.4}^{+1.8}$	${1.0}_{-0.6}^{+1.5}$
kT (keV)	...	${0.76}_{-0.14}^{+0.14}$	...	${0.76}_{-0.12}^{+0.12}$
apec normalization (10−6)	...	${4.63}_{-1.46}^{+1.58}$	...	${6.52}_{-1.74}^{+1.55}$
C-Stat (dof)	173.3 (101)	141.0 (99)	146.9 (101)	112.7 (99)

Download table as:  ASCIITypeset image

For both models, where we include a scattered power-law model to accommodate leaked AGN emission, we again find residuals at softer energies, such that adding in an apec component significantly improves the fit (the change in C-Stat is significant at the P < 0.005 level; Table 3). Thus it appears that thermal X-ray emission is not confined to the extended emission, but also arises at nuclear scales. The spherical absorption plus apec fit is shown in Figure 7.

Zoom In Zoom Out Reset image size

4.3. Extended Emission

The spectrum from the extended-only emission, extracted from a 6'' (∼1.3 kpc) region with the central point source excised, is shown in Figure 8. As expected from the imaging analysis, the spectrum is dominated by emission at soft energies. It is well accomodated by an apec plus power-law model (see Table 4), where both components are required to fit the spectrum. The power-law emission can arise from unresolved X-ray binaries in the host galaxy and/or AGN continuum photons that scatter into our line of sight.

Zoom In Zoom Out Reset image size

Table 4.  Spectral Fit to Extended Emission

Parameter	apec + powerlaw
kT (keV)	${0.33}_{-0.08}^{+0.19}$
apec normalization (10−6)	${4.40}_{-2.00}^{+2.58}$
Γ	${0.69}_{-0.75}^{+0.69}$
Power-law normalization (10−7)	${9.39}_{-5.70}^{+9.62}$
C-Stat (dof)	19.7 (14)

Download table as:  ASCIITypeset image

In this extended zone, the temperature of the gas is significantly lower than that observed on nuclear scales with $\mathrm{kT}={0.33}_{-0.08}^{+0.19}\,\mathrm{keV}$ versus ${0.76}_{-0.12}^{+0.12}\,\mathrm{keV}$. In fact, from the global fit to the X-ray emission, we find a gas temperature (kT ∼ 0.7) consistent with the nuclear value. Additionaly, the thermal luminosity is higher on nuclear than on extended scales (${4.33}_{-1.15}^{+1.03}\times {10}^{39}$ erg s⁻¹ and ${1.77}_{-0.80}^{+1.04}\times {10}^{39}$ erg s⁻¹, respectively), suggesting that circumnuclear star formation dominates the total thermal emission in this galaxy.

5.1. X-Ray History of NGC 4968

5.2. Comparison of Torus Properties with Past X-Ray and Infrared Studies

5.3. Extended X-Ray Emission Is Not Coincident with Features Discovered at Other Wavelengths

5.4. The X-Ray View of Star Formation

We use the thermal emission from the apec component as a starting point to estimate the SFR. As pointed out in LaMassa et al. (2012), using solely the apec luminosity neglects the contribution from X-ray binaries, which emit as power laws. However, the unresolved X-ray binary population cannot be easily disentangled from the AGN power-law emission. In LaMassa et al. (2012), we presented a template to adjust L_thermal to recover X-ray binary emission and more accurately estimate the SFR. Using a sample of 22 star-forming galaxies from the XSINGS sample (a Chandra survey of the SINGS sample, Kennicutt et al. 2003), we calculated the mean ratio of the X-ray emission from non-nuclear point sources to the total 0.5–2 keV emission (R). The 0.5–2 keV luminosity attributable to star formation (${L}_{0.5\mbox{--}2\mathrm{keV},\mathrm{SF}}$) is then

$$\begin{eqnarray}&&{L}_{0.5\mbox{--}2\mathrm{keV},\mathrm{SF}}=\displaystyle \frac{{L}_{\mathrm{thermal}}}{1-R},\end{eqnarray} \tag{ 1 }$$

where R = 0.51 ± 0.26.

Using the above expression and L_thermal = 6.0 × 10³⁹ erg s⁻¹ (Table 2), we estimate that the X-ray luminosity associated with star formation is 1.2 ± 0.6 × 10⁴⁰ erg s⁻¹ (taking into account errors on the fitted apec luminosity and R value). Using the X-ray luminosity to SFR conversion presented in Pereira-Santaella et al. (2011), where they combine ultraviolet and infrared data to account for both the unobscured and obscured star formation in a local sample of luminous infrared galaxies:

$$\begin{eqnarray}&&\mathrm{SFR}\ ({M}_{\odot }\,{\mathrm{yr}}^{-1})=3.4\times {10}^{-40}\,{L}_{0.5\mbox{--}2\mathrm{keV},\mathrm{SF}}\ (\mathrm{erg}\,{{\rm{s}}}^{-1}),\end{eqnarray} \tag{ 2 }$$

we find an SFR of 4.1 ± 2.0 M_⊙ yr⁻¹. We note that when using the Ranalli et al. (2003) ${L}_{0.5\mbox{--}2\mathrm{keV}}$/SFR relation, which is calibrated on infrared data only and thus neglects unobscured star formation and also assumes a different initial mass function than Pereira-Santaella et al. (2011), we find an SFR of 2.6 ± 1.3 M_⊙ yr⁻¹.

Additionally, we estimate the far-infrared SFR from the IRAS flux densities at 60 and 100 μm (2.35 Jy and 3.75 Jy, respectively; Spinoglio & Malkan 1989). Using the far-infrared derived SFR calibration from Kennicutt (1998):

$$\begin{eqnarray}&&\mathrm{SFR}\ ({M}_{\odot }\,{\mathrm{yr}}^{-1})=4.5\times {10}^{-44}{L}_{\mathrm{FIR}}\ (\mathrm{erg}\,{{\rm{s}}}^{-1}),\end{eqnarray} \tag{ 3 }$$

we obtain an SFR of 2.4 M_⊙ yr,⁻¹ which is essentially identical to our X-ray derived SFR using the Ranalli et al. (2003) relation; both works assume the same initial mass function. We note that the Kennicutt (1998) SFR calibration includes mid-infrared wavelengths, which we did not include here to minimize contamination from AGN-heated dust to our star-formation estimates.

The mid-infrared spectroscopy provides an independent probe of the AGN/star formation connection in NGC 4968. In LaMassa et al. (2010), we reported the polycyclic aromatic hydrocarbons (PAHs) EW measurements at 11.3 μm and 17 μm for the 12 μm Sy2 sample using Spitzer data. PAHs are associated with star formation though the presence of AGN-heated dust dilutes the strength of the PAH EW: the stronger the AGN contribution to the MIR continuum, the weaker the PAH EW. We found the PAH EWs to be weak. We can infer that circumnuclear star formation is on-going by the detection of the PAH features and presence of thermal X-ray emission, though AGN-heated dust dominates the mid-infrared emission, evidenced by both the MIR spectroscopy and WISE photometry.

5.5. Extreme Fe Kα EW: An Indication of More Than Just Obscuration?

We presented the Chandra imaging and spectral analysis of the Seyfert 2 galaxy NGC 4968. Capitalizing on Chandra's arcsecond resolution, we investigated the spatially resolved X-ray emission, finding the following.

• 1.  
  The hard (2–10 keV) X-ray emission is extended on scales of ∼500 pc, coincident with the narrow line region as mapped by [O iii] 5007 Å emission detected by the Hubble Space Telescope (Schmitt et al. 2003).
• 2.  
  The soft (0.5–2 keV) X-ray emission is significantly extended to scales of about a kiloparsec, and is more extended than, and spatially distinct from, both the narrow line region and a previously detected elongated radio structure (Schmitt et al. 2001).

The most striking feature of NGC 4968 comes from its X-ray spectrum, which reveals a prominent Fe Kα EW (>2 keV) as reported in Guainazzi et al. (2005) and LaMassa et al. (2011). From the Chandra spectrum, we measure an EW of ${2.5}_{-1.0}^{+2.6}\,\mathrm{keV}$. This extreme value can only be achieved if the continuum is suppressed by Compton-thick levels of obscuration, and is more likely to occur in a geometry where the central engine is completely enshrouded. In that case, the differential extinction of the total continuum (emanating from the center of the obscuring medium, and including the reflected and scattered continua) compared with the fluorescent line emission (produced throughout the medium) is the greatest, boosting the EW.

We use the MYTorus (Murphy & Yaqoob 2009) and the spherical absorption models of Brightman & Nandra (2011), which are physically motivated models that self-consistently treat the transmitted, reflected, and fluorescent line emission, to fit the X-ray spectra of NGC 4968. From the spectral analysis, we learn the following.

• 1.  
  The obscuration is at least near Compton-thick levels and is likely more extreme.

  • (i)  
    The MYTorus fit to the spectrum returns N_H = ${1.35}_{-0.22}^{+0.22}\times {10}^{24}$ cm⁻², while the spherical absorption model measures an N_H of ${7.64}_{-2.37}^{+8.57}\times {10}^{24}$ cm⁻².
  • (ii)  
    The instrinsic X-ray luminosity from the spherical absorption fit (${L}_{2-10\mathrm{keV},\mathrm{int}}$ = ${7.8}_{-5.4}^{+0.7}\times {10}^{42}$ erg s⁻¹), about two orders of magnitude higher than that which is observed, is consistent with what would be predicted based on the MIR–X-ray correlation (Gandhi et al. 2009; Asmus et al. 2015).

  A future significant high energy (i.e., >10 keV) observation from, e.g., NuSTAR should differentiate between a toroidal and spherical absorption model. These data would provide the most accurate constraints on the column density and intrinsic X-ray luminosity.
• 2.  
  A thermal model (apec) is needed to adequately fit the soft X-ray spectrum, indicating the presence of hot gas with temperature kT ∼ 0.7 keV around the AGN. Such emission is likely linked with on-going star formation.
• 3.  
  We separately investigated the X-ray spectrum from the nuclear (<400 pc) and extended (400 pc–1.3 kpc) regions.

  • (i)  
    Nuclear spectrum: thermal X-ray emission is present at the smallest scales we are able to resolve, with a temperature of kT ∼ 0.75 keV.
  • (ii)  
    Extended emission spectrum: within the extended region, both thermal and non-thermal emission are present and are ascribed to star formation and unresolved X-ray binaries and/or scattered AGN photons, respectively. The temperature of the gas (kT ∼ 0.33 keV) is lower than that within the nuclear region.

We derived the SFR based on the thermal X-ray luminosity (L_thermal ∼ 6 × 10³⁹ erg s⁻¹) and the template presented in LaMassa et al. (2012), finding an SFR of ∼2.6–4 M_⊙ yr⁻¹, consistent with the far-infrared SFR (based on the 60 μm and 100 μm fluxes) of 2.4 M_⊙ yr⁻¹. Though this rate is not as prodigious as most starburst galaxies, we can conclude that star formation is on-going in the central ∼kiloparsec of this galaxy.

It is interesting that AGNs with extreme Fe Kα EWs appear to reside in galaxies with circumnuclear star formation (e.g., Levenson et al. 2002; Boorman et al. 2016). This association may be a natural consequence of nuclear star formation, where gas completely envelops the AGN (perhaps due to stellar feedback inflating the obscuring medium and increasing its covering factor to unity; Wada & Norman 2002; Schartmann et al. 2009; Wada et al. 2009), providing the conditions necessary to produce large EW values. Though a global trend between star-formation activity and X-ray obscuration is not always apparent (see, e.g., LaMassa et al. 2011), gas from star formation may contribute to the column density in these systems (Thompson et al. 2005; Ballantyne 2008). We emphasize that the converse is not necessarily true: not all starburst/AGN composities have prominent Fe Kα EWs. However, AGNs with large Fe Kα EW values are promising candidates to look for clues of on-going, circumnuclear star formation. Such a search may be especially relevant at higher redshift where the SFRs and central gas masses are much higher than in the present day universe.

We close by noting that though the Fe Kα emission line is the most prominent feature in the Chandra spectrum, the line is spectrally unresolved. Follow-up high-resolution grating spectroscopy, or microcalorimeter observations, would spectrally resolve this feature and determine whether it originates within the standard broad line region or putative torus (e.g., Yaqoob et al. 2001, 2003; Shu et al. 2011).

We thank the referee for a thorough reading of this manuscript and for providing constructive comments. Support for this work was provided by the National Aeronautics and Space Administration through Chandra Award Number GO5-16112X issued by the Smithsonian Astrophysical Observatory for and on behalf of the National Aeronautics Space Administration under contract NAS8-03060. S.M.L. is supported by an appointment to the NASA Postdoctoral Program at the NASA Goddard Space Flight Center, administered by Universities Space Research Association under contract with NASA. N.A.L. is supported by the Gemini Observatory, which is operated by the Association of Universities for Research in Astronomy, Inc., on behalf of the international Gemini partnership of Argentina, Brazil, Canada, Chile, and the United States of America. P.B. and P.G. thank STFC for support (trant reference ST/J003697/2).

The scientific results reported in this article are based to a significant degree on observations made by the Chandra X-ray Observatory. This research has made use of software provided by the Chandra X-ray Center (CXC) in the application packages CIAO, ChIPS, and Sherpa.

Chandra ObsId 17126.

Facility: CXO - Chandra X-ray Observatory satellite.