ALMA IMAGING OF THE CO (6-5) LINE EMISSION IN NGC 7130^*

Figures 1(a)–(d) show the integrated CO(6-5) emission, the 434 μm continuum, the first and second moment maps, respectively. All images are overlaid by the same contours of the integrated CO(6-5) line emission. In Figure 1(a) we marked the three distinct regions: NE, central, and SW. The central region appears clumpy, and several peaks can be identified, whereas the other two regions look more diffuse. For example, three clumps were found by the clump-finding algorithm of Williams et al. (1994)¹⁹ , which performs a blind search for clumps by contouring the map at different levels to identify peaks. The information of these clumps, such as integrated flux, size and separation, is given in Table 2.

The velocity field (Figure 1(c)) indicates that the gas in all 3 regions participates in a common rotation that covers radii up to ∼1 kpc. The mid-point between the two separate CO concentrations in the nucleus has a systemic velocity of ∼4790 km s⁻¹. From this mid-point, the underlying (projected) rotational velocity increases radially out, to a value of a few tens of km s⁻¹ seen at the outer edge of the nuclear gas. At the radii of the NE or SW regions, the underlying rotational velocity is reduced to ∼10 km s⁻¹. This overall kinematic pattern can also been seen in the channel maps displayed in Figure 2, where channels are overlaid on the integrated line emission image. The CO(6-5) emission in the NE and SW regions coincide well with the dust lane associated with the inner spiral arms as seen in the Hubble Space Telescope (HST) optical image of González Delgado et al. (1998), suggesting that the observed rotational velocity field likely reflects the galactic rotation.

Zoom In Zoom Out Reset image size

Figure 1(d) shows that the line of sight velocity dispersion is from ∼30 to 50 km s⁻¹ in the central region, around 20 km s⁻¹ in the NE region, and between 10 and 20 km s⁻¹ in the SW region. However, the core of the SW region has a large velocity dispersion, as high as ∼50 km s⁻¹.

Figures 3(a)–(d) present the integrated CO(6-5) line profiles for the 3 regions as well as for the total ALMA detection, together with the central velocity and FWHM from a Gaussian fit. The central velocities are consistent with a systematic rotation that has a projected velocity of 12–15 km s⁻¹ at the radii of NE or SW. Among the three regions, the SW one has the broadest line width, consistent with its large velocity dispersion seen in Figure 1(d). The FWHM of the total line emission in Figure 3(d) is similar to that of CO(1-0) and CO(2-1) (Albrecht et al. 2007).

Zoom In Zoom Out Reset image size

The total flux of CO(6-5) measured from the integrated image is 1230 ± 74 Jy km s⁻¹ (within an ellipse of a = 39 and b = 17, P.A = 126), which agrees well with our SPIRE/FTS-measured flux of 1223 ± 82 Jy km s⁻¹ (N. Lu et al. 2016, in preparation), obtained with a much larger aperture of ∼33''. The central (∼15 in diameter; or ∼530 pc) region contains ∼50% of the total ALMA-detected emission, and the NE and SW regions about 20% and 30% respectively. These results suggest that our ALMA observation has recovered all of the CO(6-5) emission in NGC 7130.

As shown in Figure 1(b), the 434 μm continuum emission generally correlates spatially with the CO(6-5) emission in NGC 7130, but with two notable differences over smaller scales: (1) The brightest peaks of the nuclear continuum and CO(6-5) line emission have a small offset of ∼02 (or ∼70 pc). (2) The average line-to-continuum ratio (R_CO/cont) is particularly high in the SW region compared to those in the central or NE regions. The corresponding values of R_CO/cont are 2100 (SW), 1700 (NE) and 1300 (central), suggesting an intrinsically faint 434 μm flux density in the SW region. In star-forming regions, the dust emission is always dominated by far-UV photon heating in the photo-dissociation regions (PDRs) near young massive stars. In contrast, the dominant heating mechanism for the CO(6-5) emission is still controversial despite the fact that its heating source should still be associated with the ongoing SF on a global scale (Lu et al. 2014). The observed differences between the dust continuum and CO(6-5) emission at small scales may lend direct support to the notion that shocks (or other mechanical energy input) are mainly responsible for the warm CO emission as often suggested by many authors who modeled the observed CO spectral line energy distributions in galaxies (e.g., Nikola et al. 2011; Rangwala et al. 2011; Kamenetzky et al. 2012; Meijerink et al. 2013; Rosenberg et al. 2014).

This can be better understood in a simplified analytical way: The CO(6-5) to continuum ratio, I_CO(6−5)/(ν I_ν,434μm), equals (I_CO(6−5)/I_CO(1−0)) × (I_CO(1−0)/(ν I_434μm)). In the inner galaxy where most of the gas is molecular (e.g., Tanaka et al. 2014), I_CO(1−0) scales with the total gas mass (M_gas); I_ν,434μm is likely in or near the Rayleigh–Jeans limit and thus scales approximately with the dust mass (M_dust) times the effective dust temperature (T_dust). Therefore, I_CO(6−5)/(ν I_{ν,434 μm}) ∝ (I_CO(6−5)/I_CO(1−0))(M_gas/M_dust)(1/T_dust). Since M_gas/M_dust is more or less fixed, a higher I_CO(6−5)/(ν I_ν,434μm) ratio can result from a higher CO excitation or/and a lower T_dust. If both CO(6-5) and dust emissions are tied to the same PDR mechanism, their emission peaks should coincide with each other spatially. This is not consistent with our data at scales ≤70 pc. Furthermore, if supernova-driven shocks are responsible for the CO(6-5) emission, the high I_CO(6−5)/(ν I_ν,434μm) ratio in the SW region could be associated with the epoch when the O stars (with lifetimes of several times 10⁶ years) from the last starburst episode have burned off, leading to a lower T_dust. Since supernova progenitors are less massive (down to ∼5 M_⊙) and have longer lifetimes (up to ∼10⁸ years), the I_CO(6−5)/I_CO(1−0) ratio could still remain high, leading to a higher I_CO(6−5)/(ν I_ν,434μm) ratio. This picture is overall consistent with the observations that the gas in the SW region displays a complex velocity field and little Pa-α emission (see Figure 4 below). However, there exists another possibility that the NE/SW gases were excited by bar-induced shocks, which can inhibit star formation when the shock and shear is too strong (e.g., Athanassoula 1992; Reynaud & Downes 1998).

Zoom In Zoom Out Reset image size

Our ALMA-detected flux density of the continuum at 434 μm is f_434μm,ALMA = 814 ± 52 mJy. Following a similar procedure (i.e., interpolating the SPIRE measurements at 350 and 500 μm to obtain the total flux) in Xu et al. (2014, 2015), we estimated that about 0.51 ± 0.05 of the total dust continuum is detected by our ALMA observation, which is lower by a factor of about 2 than the interferometer-to-single-dish flux ratio of the CO(6-5) line emission. This result indicates that the dust distribution is substantially more extended or more diffuse overall than that of the warm dense gas.

Figures 4(a)–(d) display the integrated CO(6-5) line emission contours overlaid on the HST F210M (UV) (González Delgado et al. 1998), F160W continuum and Pa-α (Alonso-Herrero et al. 2006; Díaz-Santos et al. 2008), and the VLA 8.4 GHz continuum images (Thean et al. 2000), respectively. The extinction-corrected Pa-α map was obtained from a set of HST near-IR (NIR) narrow- and broad-band images (Díaz-Santos et al. 2008). These NIR images have a resolution of ∼02, which nearly matches our ALMA observations. The radio image was obtained by reducing the raw data downloaded from the NRAO VLA science archive.²⁰ The elongated beam (due to the target being in the southern sky), using a uniform weighting, has a size of 060 × 019. The maximum recoverable scale for this snapshot observation is about 32, very close to our ALMA observation.

The astrometry of these HST observations has been improved by utilizing the 2MASS point source catalog and the ring-like feature identified in González Delgado et al. (1998). Furthermore, we have also assumed that the nucleus seen in the F160W peak coincides with the radio peak, since an AGN usually peaks at the nucleus that is the bottom of the gravitational potential well. Indeed González Delgado et al. (1998) identified the optical peak (which correlates with the F160W-band) as the nucleus. This way, we can match ALMA and NIR maps to an accuracy of <01, dominated by the ALMA/radio offset.

Figure 4(a) shows that the clumps in the UV image have offsets relative to the dense gas emission, which may be caused by extinction since the UV data are heavily obscured (González Delgado et al. 1998), and/or by the break in the local SF law (Xu et al. 2015). The western peak in the CO(6-5) image seems to be complementary to the ring-like structure shown in the UV image. Figure 4(b) reveals that the two dust lanes, which run in the north–south direction and are also seen in the optical image (González Delgado et al. 1998), coincide well with our CO(6-5) emission (SW and NE regions). This indicates that our method for improving the astrometry of the HST images is reasonable. Furthermore, the dust lanes and NE/SW gas offset toward the leading sides of the inner bar, which can be identified in the K-band image and is oriented at P.A. = 0° (Mulchaey et al. 1997). This result is consistent with numerical simulations (e.g., Athanassoula 1992) that in the presence of a barred potential, gas will hit the bar from behind and shock, forming a dust-lane, and then the gas will fall inward along the bar. The shocks can heat gas much more efficiently than dust. Therefore, both the observed gas morphology and high CO(6-5)-to-continuum ratio indicate that the NE/SW gas might have been accumulated through bar-induced gas inflow.

From Figure 4(c) it is evident that the Pa-α emission varies smoothly with radius, and and has no clumps corresponding to the ones seen in the central region of the CO(6-5) image. Unless the extinction correction is severely uncertain, our result indicates a breakdown of the correlation between SF and warm dense gas at sub-100 pc scales, which has been found in NGC 1614 (Xu et al. 2015; but see Wu et al. 2010 and Chen et al. 2015 for opposite results). Nevertheless, a quantitative analysis, using a higher-quality radio image, is needed to reach a solid conclusion since the extinction correction for the Pa-α image may suffer from some uncertainties.

The only peak seen in the radio continuum image (Figure 4(d)) is close to the eastern clumps of the CO(6-5) emission (central region). There is no detectable radio emission for the NE and SW regions ∼2'' away from the central region seen in CO(6-5), whereas Liu et al. (2015b) found that there is no difference in the global FIR and radio correlations with dense gas and CO emission. The majority of and the peak emission of the radio map may correspond to the AGN, as a compact radio core has been detected in NGC 7130 at 2.3 GHz (Corbett et al. 2002). This is also supported by the fact that, for star-forming regions in NGC 1614, the average CO(6-5)-to-radio flux ratio (R_CO/radio) is 5.4 × 10⁴ km s⁻¹, and for NGC 34, whose AGN contribution to the radio continuum is negligible (Corbett et al. 2002), R_CO/radio = 6.1 × 10⁴ km s⁻¹. For the western knot (i.e., the western peak of the CO emission in the central region; here we have smoothed our ALMA data to match the radio resolution) in the radio map, R_CO/radio is 5.3 × 10⁴ km s⁻¹. Whereas for the region centered at the radio peak (measured within an ellipse of a = 06 and b = 02) in NGC 7130, it is 1.1 × 10⁴ km s⁻¹, ∼5 times lower than the forementioned values, indicating a significant enhancement of the radio emission in this region.

Assuming that the X-ray dominated region associated with the AGN is unresolved by ALMA, we measured the CO(6-5) flux within the region (r = 02; hereafter A_XDR) centered at the radio peak and found that it only accounts for ∼15% of our ALMA-detected flux within the central region, or ∼7% of the total flux of the entire galaxy. However, there might be a contribution from massive stars to the heating of CO(6-5) in A_XDR. So it is an upper limit of the AGN contribution to the CO(6-5) heating. Our results are also consistent with the finding in Lu et al. (2014), i.e., SF rather than the AGN dominates the excitation of mid-J CO emission. The spectrum extracted within A_XDR has a central velocity of 4799 km s⁻¹ and FWHM of 86.4 km s⁻¹, which are ∼7 km s⁻¹ and 12 km s⁻¹ larger and smaller, respectively, than those of the spectrum from the entire central region. However, unlike the optical lines from the nucleus (e.g., Davies et al. 2014), no extended wing was detected in all of our CO(6-5) spectra (see Figure 3).