Multiwavelength Analysis of a Nearby Heavily Obscured AGN in NGC 449

A log of the XMM-Newton and NuSTAR observations analyzed herein is presented in Table 1, and the individual data sets are described in this section.

We search the archives for its mid-IR data. There is good mid-IR spectral data, covering wavelengths from 5 to 37 µm, observed by Spitzer’s IrsStare on February 5, 2009 (PI: Levenson, Nancy, ObsID: 25408000). The mid-IR spectrum was processed using the Spitzer data analysis software irs_merge v2.1 with the pipeline script (version: S18.18.0) on December 18, 2011. For more detailed processing of the mid-IR spectrum of this source, please refer to Section 2 of Lebouteiller et al. (2011).

To construct the SED for NGC 449, we search for available multi-band photometric data from different surveys or the archives of telescopes, i.e., the Sloan Digital Sky Survey (SDSS) Data Release 7 (Abazajian et al., 2009), the Two Micron All-Sky Survey (2MASS, Skrutskie et al., 2006), the Wide-field Infrared Survey Explorer (WISE, Wright et al., 2010), the AKARI and the Infrared Astronomical Satellite (IRAS). In total, we collect the photometric data for 17 filters. Among them, only the filter SDSS-u is in the UV band. There are three filters of the optical band, including SDSS-g, SDSS-r, and SDSS-i. The other filters are in the IR band, including four near-IR filters (SDSS-z, 2MASS-J, 2MASS-H, 2MASS-K), seven mid-IR filters (W1, W2, W3, W4, IRAS-PSC 25, AKARI-PSC 09, AKARI-PSC 18) and two far-IR filters (IRAS-PSC 60, IRAS-PSC 100). The details for photometric data and filters are listed in Table 2.

In this section, we analyze the X-ray spectra of NGC 449 in Xspec v.12.11.0 (Arnaud, 1996). First, we use phenomenological modeling to determine whether there is variation in the spectra for the same energy range of XMM-Newton and NuSTAR. We then analyzed which components contributed to the X-ray spectra of this source. Based on the analysis of the X-ray components, we finally use the self-consistent and physical models to fit the X-ray spectra. All errors represent the 68.0% confidence intervals unless otherwise stated.

In this section, we use DeblendIRS²²2For more details, refer to http://www.denebola.org/ahc/deblendIRS/ (Hernán-Caballero et al., 2015) to analyze the mid-IR spectrum of NGC 449. DeblendIRS is an IDL package that fits the mid-IR spectra with a linear combination of three spectral templates, i.e., a “pure” AGN template, a “pure” stellar template, and a “pure” Polycyclic Aromatic Hydrocarbon (PAH, which accounts for the interstellar emission) template. The mid-IR spectra of the sources can be well decomposed into the contributions of different components using this package. Thus we obtain some physical properties of the AGNs and their host galaxies, such as the silicate strength³³3The silicate strength is defined as $$\rm{S}_{\rm{Si}}=\ln\frac{F(\lambda_{p})}{F_{C}(\lambda_{p})},$$ where ${\rm F}(\lambda_{p})$and ${\rm F}_{C}(\lambda_{p})$ stand for the maximum flux density of the silicate line profile near 9.7 µm and the corresponding flux density of the underlying continuum profile,respectively. $\rm{S}_{\rm{Si}}$ is a negative value, which means that the central radiation of the AGN is absorbed by the torus.The optical depth of the silicate absorption $\tau_{9.7}=-\rm{S}_{\rm{Si}}$. ($\rm{S}_{\rm{Si}}$), spectral index ($\alpha$), AGN emission at rest-frame 6 µm, and 12 µm.

We use this package to fit the mid-IR spectrum (5 – 16 µm) and provide the best-fitting results. Figure 4 shows the best-fitting of the mid-IR spectrum of NGC 449. From this fitting, we derive a silicate strength of $-0.71$, a luminosity at 6 µm of $7.65\times 10^{42}\rm{erg\ s^{-1}}$, and a luminosity at 12 µm of $1.84\times 10^{43}\rm{erg\ s^{-1}}$.

In this section, we analyze the Multi-band SED of this source using Code Investigating GALaxy Emission (CIGALE 2020.0, Boquien et al., 2019). CIGALE 2020.0 is an open python code designed to estimate the physical properties (i.e., SFR, stellar mass, AGN luminosity) of galaxies and AGNs. We fit the SED of NGC 449 with the same modules and parameters as Guo et al. (2020). For details, please refer to Section 3 of Guo et al. (2020).

Figure 5 presents the best-fit SED for this source. In addition, we also obtain some physical parameters of NGC 449 by the SED fitting, such as SFR, stellar mass, and AGN luminosity. Table 6 lists several parameters that may be used in this work. The SFR estimate by the SED fitting is more dependent on the star formation history model, so we also use the calibration from Bell et al. (2005) to estimate SFR by the UV and IR luminosities, scaled to Chabrier (2003) initial mass function:

where $\rm{L_{UV}}=\nu L_{\nu}$ is an estimation of the integrated 1216 – 3000Å rest-frame UV luminosity, and $\rm{L_{IR}}$ is the 8 – 1000 µm rest-frame IR luminosity. Both $\rm{L_{UV}}$ and $\rm{L_{IR}}$ (listed in Table 6) are in units of L_⊙. The SFR of NGC 449 is 3.12 M_⊙/yr which is estimated by the UV and IR luminosities.

The AGN at the center of NGC 449 is an optical type-2 AGN, so we cannot directly observe its central emission (e.g., accretion disk, corona). It is difficultly to understand the coronal (intrinsic) emission only by the X-ray spectrum fitting. For heavily obscured AGNs, the estimation of intrinsic luminosities (flux densities) is very dependent on the X-ray model employed. Although Table 5 presents the intrinsic 2–10 keV luminosities (flux densities) obtained by the Pexmon and Borus models, they are unreliable. To derive a reliable intrinsic luminosity, we had better estimate it in other ways. Fortunately, there is a very close correlation between 2–10 keV and mid-IR 12 µm luminosities of AGNs (e.g., Asmus et al., 2015). In Section 3.2, we have obtained the mid-IR 12 µm luminosity ($\nu\rm{L}_{\nu}(12~{}\micron)=1.84\times 10^{43}\rm{erg\ s^{-1}}$) of central AGN. We use Equation 2 of Asmus et al. (2015) to estimate intrinsic 2–10 keV luminosity of $(8.54\pm 0.75)\times 10^{42}\ \rm{erg\ s^{-1}}$. Moreover, the fraction of the scattering component is also not well constrained. However, the Pexmon and Borus models have well-constrained the scattering luminosity, whose average value is $(2.18\pm 1.11)\times 10^{40}\ \rm{erg\ s^{-1}}$. The relationship between the intrinsic luminosity and scattering luminosity is $f_{\rm{scat}}={\rm L_{scat}}/{\rm L_{Int}}$. So the scattering fraction of the X-ray emission of central AGN is about $(0.26\pm 0.13)\%$.

Since the UV-to-optical continuum of this source is mainly contributed by the host galaxy, we cannot obtain the bolometric emission of the accretion disk by SED fitting. However, the coronal and torus emissions can well track the thermal radiation of the accretion disk. So they can be used to estimate the bolometric luminosity of type-2 AGNs. Using the intrinsic 2–10 keV luminosity, we solve the Equation 21 of Marconi et al. (2004) to estimate the bolometric luminosity $\rm{L_{bol}}=1.44\times 10^{44}\ \rm{erg\ s^{-1}}$ (bolometric correction $k_{bol}\simeq 16.85$). This bolometric luminosity is consistent with that provided by Woo & Urry (2002).

The mass accretion rate is a measurement of SMBH mass growth and is related to bolometric luminosity. The relationship is written as $\dot{m}_{\rm{acc}}=\rm{L_{bol}}/\eta c^{2}$, where $\eta$ is the efficiency that converts the rest-mass energy of accreted material into radiation. Assuming $\eta=0.1$ (Frank et al., 2002), we estimate the mass accretion rate of NGC449, which is $2.54\times 10^{-2}\rm{M}_{\odot}\ \rm{yr}^{-1}$. The Eddington ratio is a measurement of the SMBH accretion efficiency and is defined as $\lambda_{\rm{Edd}}=\rm{L_{bol}/L_{Edd}}$, where $\rm{L_{bol}}$ is the bolometric luminosity of the AGN, $\rm{L_{Edd}}$ is the Eddington luminosity of the AGN ($1.3\times 10^{38}\ \rm{M_{BH}/M_{\odot}\ erg\ s^{-1}}$). Using the SMBH mass ($1.32\times 10^{7}\ \rm{M}_{\odot}$) given in Woo & Urry (2002), we estimate the Eddington Luminosity to be $\rm{L_{Edd}}\simeq 1.72\times 10^{45}\ \rm{erg\ s^{-1}}$. The Eddington ratio is about $\lambda_{\rm{Edd}}=8.39\times 10^{-2}$. The Eddington ratio of most CT-AGNs in the local universe is similar to our result (e.g., Tanimoto et al., 2022).

Previous studies have attempted to constrain the column density of NGC449 using XMM-Newton data or the ratio of X-ray to high-ionized emission lines ([O III]5007, [Ne V]3426) luminosities (Guainazzi et al., 2005; Heckman et al., 2005; Gilli et al., 2010). Those results suggested that the center of NGC449 hosted a CT-AGN. We use the latest X-ray data above 10 keV to construct a broadband X-ray spectrum. Compared with the previous studies, we can better constrain the parameters of the torus in Section 3.1. In the Pexmon model, the column density of the torus is obtained as $(1.27\pm 0.26)\times 10^{24}\ \rm{cm}^{-2}$, indicating that it may be a CT-AGN. In Section 3.1.4, we re-fit the broadband X-ray spectrum using the Borus model and obtain a series of parameters about torus. Among these parameters, the column density of LOS is $(0.97\pm 0.16)\times 10^{24}\ \rm{cm}^{-2}$, and the average column density of the torus is about $1.23\times 10^{24}\ \rm{cm}^{-2}$, which is close to Compton-thick. In conclusion, the column density of this source we obtain is similar to that estimated by previous works (Guainazzi et al., 2005; Heckman et al., 2005; Gilli et al., 2010), but it is well-constrained.

In Section 3.2, we decomposed the contribution of AGN by the mid-IR spectrum fitting, and derived the silicate strength of the AGN component. Using the silicate strength, we can derive the optical depth of silicate absorption as $\tau_{9.7}=0.71$. The optical depth of silicate absorption is a good way to track the dust in the torus, i.e., the V-band extinction is $A_{\rm V}=19\times A_{9.7}=19\times 1.086\tau_{9.7}=14.65\ \rm{mag}$ (Roche & Aitken, 1985). Therefore, the gas-to-dust ratio of the torus is $N_{\rm H}/A_{\rm V}=8.40\times 10^{22}\ \rm{cm}^{-2}\ \rm{mag}^{-1}$ ($N_{\rm H}$ is the average column density of torus). This gas-to-dust ratio is about 45 times that of the Galaxy ($N_{\rm H}/A_{\rm V}=1.87\times 10^{21}\ \rm{cm}^{-2}\ \rm{mag}^{-1}$, Draine, 2003). Even this gas-to-dust ratio is higher than that of nearby CT-AGN in the Circinus galaxy (Tanimoto et al., 2019), which is the most obscured AGN in the nearby universe. Such a high gas-to-dust ratio of this source means that the radiation of the central AGN may have destroyed dust in the torus.

Many studies have suggested a co-evolution scheme between host galaxies and AGNs (e.g., Kormendy & Ho, 2013), such as $\rm{M_{BH}}-\sigma_{*}$ relation (e.g., Ferrarese & Merritt, 2000), AGN feedback (e.g., Bower et al., 2006). Some studies have suggested that the host galaxies of heavily obscured AGNs appear to be at the stage of intense star formation (e.g., Hopkins et al., 2006). These intense star-forming activities may be associated with AGNs, i.e. AGNs trigger star formation of host galaxies.

NGC 449 is a late-type galaxy, meaning it is in an early stage of galaxy evolution. Its center hosts a heavily obscured AGN, which is considered to be at the early stage of AGN evolution. These suggest that the host galaxy and central AGN of NGC 449 are at the early stage of co-evolution. In Section 3.3, we derive some parameters about the host galaxy, including SFR and stellar mass. The specific star formation rate ($\rm{sSFR=SFR}/\rm{M}_{*}$) of this source is approximately $0.59\ \rm{Gyr}^{-1}$. This value is consistent with other star formation galaxies at similar redshifts (Salim et al., 2007). This result suggests that the central AGN does not appear to trigger intense star formation in its host galaxy.