A $Herschel$ mapping of [C ii], [O i] and [O iii] lines from the circumnuclear region of M31

Galactic circumnuclear regions, in which the multiphase interstellar medium (ISM) and various stellar populations are actively interacting with the supermassive black holes (SMBHs), are crucial to our understanding of the accretion and feedback of SMBHs, and in turn the global evolution of the host galaxies. As the closest external spiral galaxy, the Andromeda galaxy (M31) gives us a unique opportunity to study the circumnuclear environment in a close-up view, thanks to its proximity (D $\approx$ 785 kpc, where 1$\arcsec$ = 3.8 pc, McConnachie et al., 2005), appropriate inclination ($\sim 77^{\circ}$) and low line-of-sight extinction ($A\rm_{V}\lesssim$ 1, Dong et al., 2016). The most conspicuous structure of this environment is the so-called nuclear spiral, which consists of spiral-like filaments across the central few hundred parsecs of M31. Both optical emission (Jacoby, Ford & Ciardullo, 1985) and extinction (Dong et al., 2016), and infrared (IR) emission (Li et al., 2009) from the nuclear spiral indicate that it is composed of neutral, dusty gas clouds with ionized outer layers. The investigation of dust (Draine et al., 2014) and molecular gas (e.g. Melchior & Combes, 2013; Li et al., 2019, 2020) has revealed that the kinetic temperature is higher in the nucleus than in the disk. However, the ionization/excitation mechanism of the nuclear spiral remains a long-standing puzzle, given the weak nuclear activity (Li et al., 2011) and the absence of on-going star formation (Brown et al., 1998; Groves et al., 2012; Dong et al., 2015) in this region.

Far infrared (FIR) atomic/ionic fine-structure lines, such as [C ii] 158$\,\micron$, [O i] 63$\,\micron$ and [O iii] 88$\,\micron$, arise from multiple ISM phases, and thus can be used as diagnostics of the physical conditions of the circumnuclear environment. Specifically, due to the moderate ionization potential of carbon (11.26 eV), [C ii] 158 $\micron$ line is present in low-density diffuse ionized gas, neutral photodissociation regions (PDRs) and also diffuse molecular clouds (Petuchowski et al., 1993; Heiles, 1994; Abel, 2006). On the other hand, [O i] 63 $\micron$ line is only found in neutral gas (PDRs and molecular clouds) since its ionization potential is 13.62 eV, very close to that of hydrogen; while [O iii] 88 $\micron$ is exclusively found in ionized gas because the production of O⁺⁺ requires energetic photons (35.12 eV). These lines are also widely used as star formation rate (SFR) tracers (e.g. De Looze et al., 2014) since they are from the ISM that fuels and closely interacts with star formation. Among them, [C ii] 158$\,\micron$ is most often investigated since it is typically the most luminous line in galaxies (e.g., Stacey et al., 1991). The relationship of SFR surface density and [C ii] intensity has been extensively investigated in nearby galaxies (Herrera-Camus et al., 2015), such as M51 (Pineda et al., 2018), and also in M31 disk (Kapala et al., 2015). These studies all found tight correlations between the two quantities. However, since the circumnuclear region of M31 is lack of current star formation (e.g. Dong et al., 2015), we will not further discuss this correlation in this paper.

Moreover, [C ii] 158$\,\micron$ usually acts as the main coolant for low density neutral gas (the bulk of neutral gas) due to its lower excitation energies and critical densities ($\Delta E/k$ $\sim$91 K and $n\rm_{crit}\sim$ 3$\times 10^{3}$ cm^-3), while [O i] 63 $\micron$ is dominant in high density and temperature regime ($\Delta E/k$ $\sim$228 K; $n\rm_{crit}\sim$8.5$\times$10⁵ cm^-3). The neutral ISM is mainly heated by photoelectrons ejected by dust grains absorbing far-ultraviolet (FUV) photons within the range 6-13.6 eV (Draine et al., 1978). The energy of the heated ISM is later taken away by collisional excited FIR emission lines, and the gas is thus cooled. On the other hand, the dust grains can be heated by photons of nearly all energies (mostly 0.01-13.6 eV) and then re-radiated in total infrared (TIR, 8-1000 $\micron$; Sanders & Mirabel, 1996) band. Thus, the [C ii]/TIR ratio can be an approximation of the theoretical photoelectric heating efficiency, $\epsilon\rm_{PE}$, defined as the photoelectric heating rate divided by the dust grain photon absorption rate (Tielens & Hollenbach, 1985). Investigations of these lines could give us insights about the physical properties of the ISM, in particular the M31 center.

The first [C ii] 158$\micron$ observations of M31 (Mochizuki, 2000) was made with the Infrared Space Observatory ($ISO$; Kessler et al., 1996), resulting in a strip map of 31 positions with a spatial resolution of $\sim$70$\arcsec$. Mochizuki (2000) detected [C ii] flux in the central region at the level of 1$\times 10^{-5}$ erg s^-1 cm^-2 sr^-1 and found a [C ii]/100 $\micron$ flux ratio of 6$\times 10^{-3}$, higher than that found in the Milky Way, which the author interpreted as due to the higher gas column density of the Galactic center. More recent observations (Kapala et al., 2015) carried out with Herschel Space Observatory (Pilbratt et al., 2010) targeting five regions in M31 disk have mapped the [C ii] lines with much higher resolution (12$\arcsec$) and found the [C ii]/TIR ratio increases with radius. Using the same set of data, Kapala et al. (2017) later claimed that since TIR emission involves the contribution from all photons that can heat dust, not just the FUV photons capable of ejecting photoelectrons, the [C ii]/FUV_att ratio could be a better proxy of $\epsilon\rm_{PE}$, where FUV_att stands for the attenuated FUV emission.

The circumnuclear region of M31 is a relatively small reservoir of neutral gas. To date, there is no unambiguous detection of circumnuclear atomic hydrogen reported. An upper limit of $10^{6}$ M_⊙ was placed on the HI mass in the central 500 pc (Brinks, 1984), and a more recent survey by Braun et al. (2009) yielded no substantial detection in the central kpc. Only until recently, molecular gas has been mapped in this region with sensitive IRAM 30m CO(2–1) (Melchior & Combes, 2013) and JCMT CO(3–2) observations (Li et al., 2019), which suggest a small amount of molecular gas ($\sim 3.7\times 10^{5}$ M_⊙). In this work we present $Herschel$ spectroscopic observations of the [C ii] 158 $\micron$, [O i] 63 $\micron$ and [O iii] 88 $\micron$ lines in the central 2 arcmin of M31, which is the first mapping of atomic phase gas in this region with a resolution comparable to molecular gas and dust observations. We make maps and analyze the flux ratios of these lines to study the ISM properties of the nuclear spiral and compare the results with that of the disk. The paper is organized as follows. The observations and data preparation are described in Section 2. The results are presented in Section 3, including the morphology and kinematics of the lines, and the analysis of their ratios. The discussion and implication of the results are presented in Section 4. A brief summary is given in Section 5.

The observations of [C ii] 158$\,\micron$, [O i] 63$\,\micron$ and [O iii] 88$\,\micron$ lines were carried out during 2012-6-19 and 2012-6-20 for a total time of 9.9 hours (obsid: 1342247148, 1342247149; PI: Z. Li) with $Herschel~{}Space~{}Observatory$. The data were obtained with Photodetector Array Camera and Spectrometer (PACS; Poglitsch et al., 2010) using the 3 $\times$ 3 raster mode covering a 120$\arcsec\times 120\arcsec$ ($\sim 500\times 500$ pc) area of M31 nuclear spiral. This field is centering on ([RA,DEC]=[10.691149, 41.275203]), $\sim$100 pc northeast of the M31 center ([RA,DEC]=[10.684793, 41.269065]). We adopted the unchopped grating scan mode, with an off-source field chosen at an inter-arm region $\sim 500\arcsec$ northwest of the M31 nucleus. The wavelength, angular and velocity resolution of these lines are given in Table 2.

The data were reduced using the Herschel Interactive Processing Environment (HIPE, Ott, 2010) version 14.0 standard pipeline (SPG 14). We made use of the level 2 data from the Herschel Science Archive (HSA), which are fully reduced and flux calibrated. We then applied the script provided by $Herschel$ named “FittingPacsCubes_mappingI.py”, which is suitable for PACS spectroscopy data cube analysis, to process the data and fit the spectra. The main steps are summarized below.

We started with the interpolated cube with a pixel size of 3$\arcsec$ since it not only satisfies the Nyquist sampling requirement for all three lines but also is smoother than the recommended projected cube. First, we cropped the noisy end of the spectra to avoid contamination from the spectral edges. Next, we used a first-order polynomial (second-order for the [O i] spectrum, since the signal is relatively weak and the baseline is not as stable as the other two lines) plus a Gaussian to fit each spectrum, and obtained the peak, center and width of the line. We then eliminated the spaxels with peak signal-to-noise ratio (SNR) smaller than 3 and obtained the integrated intensity and velocity maps of all three lines using the masked cube. For comparison of CO(3–2) and [C ii] emission, we smoothed them to a common spatial resolution and regridded them to a pixel size of 15$\arcsec$. The flux is converted to erg s^-1 cm^-2 sr^-1, and the velocity reference frame is Local Standard of Rest (LSR). Throughout this paper, we adopt a systemic velocity of M31 of -300 km s^-1 (van den Bergh, 1969), a position angle of 38^∘ and an inclination angle of 77^∘ for the M31 disk (McConnachie et al., 2005), and an inclination angle of 45^∘ for the central 1 kpc (Ciardullo et al., 1988).

We retrieved several ancillary data sets to assist our analysis: (1) The CO(3–2) map of the central kpc of M31 obtained from the James Clerk Maxwell Telescope (JCMT) with HARP raster mode (Li et al., 2019), taken with a resolution of 15$\arcsec$ and a typical RMS of 3.5 mK in a 13 km s^-1 channel. (2) $Herschel$ PACS imaging in [C ii] 158 $\micron$ and optical integral-field spectroscopy of five fields in M31 disk from the Survey of Lines In M31 (SLIM, Kapala et al., 2015). These fields are reduced in the same way as described above. (3) The CO(3–2) observations of the five SLIM fields from a JCMT large program HARP and SCUBA-2 High-Resolution Terahertz Andromeda Galaxy Survey (HASHTAG; Li et al., 2020, M. Smith et al. in preparation). (4) Galaxy Evolution Explorer (GALEX) FUV (1350-1750 Å) map of M31 (Thilker et al., 2005). (5) The dust mass surface density map of the whole galaxy (Draine et al., 2014), obtained with spectral energy distribution (SED) fitting of infrared emission. (6) Total far infrared (FIR) luminosity map from Ford et al. (2013), which was produced by integrating over six frequency bands, including $Spitzer$ MIPS 70 $\micron$, $Herschel$ PACS and SPIRE 100, 160, 250, 350, 500 $\micron$. The footprint of the five SLIM fields as well as the Nucleus field are shown in Figure 1.