# Imaging the molecular outflows of the prototypical ULIRG NGC 6240 with ALMA

Imaging the molecular outflows of the prototypical ULIRG NGC 6240 with ALMA Abstract We present 0.97 × 0.53 arcsec2 (470 pc × 250 pc) resolution CO (J = 2–1) observations towards the nearby luminous merging galaxy NGC 6240 with the Atacama Large Millimeter/submillimeter Array. We confirmed a strong CO concentration within the central 700 pc, which peaks between the double nuclei, surrounded by extended CO features along the optical dust lanes (∼11 kpc). We found that the CO emission around the central, a few kpc, has extremely broad velocity wings with full width at zero intensity ∼ 2000 km s−1, suggesting a possible signature of molecular outflow(s). In order to extract and visualize the high-velocity components in NGC 6240, we performed a multiple Gaussian fit to the CO data cube. The distribution of the broad CO components shows four extremely large line width regions (∼1000 km s−1) located 1–2 kpc away from both nuclei. Spatial coincidence of the large line width regions with H α, near-IR H2, and X-ray suggests that the broad CO (2–1) components are associated with nuclear outflows launched from the double nuclei. galaxies: active, galaxies: evolution, galaxies: individual: NGC 6240, galaxies: interactions 1 INTRODUCTION Gas-rich galaxy mergers play a major role in the formation and evolution of galaxies by triggering intense star formation and changing their morphology as suggested by numerical simulations (e.g. Barnes & Hernquist 1991). During the process of a galaxy merger, radial streaming can feed gas to the central supermassive black hole (SMBH; e.g. Hopkins et al. 2008), possibly reaching at the quasar-phase after dispersing the obscuring gas and dust (Urrutia, Lacy & Becker 2008). Powerful winds or outflows, driven by the active galactic nucleus (AGN) or the surrounding starbursts [(ultra-)luminous infrared galaxy, (U)LIRG], are predicted to suppress gas feeding to the central SMBH and its host galaxy's spheroidal component in order to explain the empirical correlation between the stellar velocity dispersion of a galaxy bulge and the mass of the central SMBH (Costa, Sijacki & Haehnelt 2014). Observational studies of such galactic winds/outflows in (U)LIRGs have been frequently done by spectral analyses of hot molecular gas, atomic gas, and ionized gas emission/absorption lines (e.g. Bellocchi et al. 2013; Veilleux et al. 2013). Those tracers are bright enough to probe generally faint, broad outflowing gas profiles, although they are not likely to be major constituents of galactic outflows. Cold molecular gas is likely to dominate the outflow in terms of mass, and thus it is the key to understanding fundamental feedback effects of galactic outflows on the surrounding interstellar medium (ISM; Feruglio et al. 2010; Cicone et al. 2014; Fiore et al. 2017). NGC 6240 is a well-studied close galaxy pair in the local Universe (z = 0.024480, 1.0 arcsec = 480 pc), which is often regarded as a prototype of ULIRGs, even though its infrared (IR) luminosity is slightly lower than 1012 L⊙ (LIR = 1011.93 L⊙; Armus et al. 2009). This system has a double nucleus detected in a variety of wavelengths from X-ray to radio (e.g. Komossa et al. 2003; Hagiwara, Baan & Klöckner 2011; Stierwalt et al. 2014; Ilha, Bianchin & Riffel 2016), indicating the presence of two separated AGNs. The cold molecular ISM in NGC 6240 has been observed through many rotational transitions of CO, HCN, and HCO+, showing a strong gas concentration between the nuclei (e.g. Nakanishi et al. 2005; Iono et al. 2007; Papadopoulos et al. 2014; Scoville et al. 2015; Tunnard et al. 2015; Sliwa & Downes 2017). Although the properties shown above are indeed suitable for representing nearby ULIRGs, past observations have revealed a different side of NGC 6240, which makes it rather unique. Lu et al. (2015) reported that NGC 6240 shows an order of magnitude higher LCO(7–6)/LIR ratio than other local (U)LIRGs. Such an extreme CO line-to-continuum ratio can be explained by a simple C-shock model (Meijerink et al. 2013). Furthermore, deep and wide Subaru observations revealed a vastly extended, filamentary H α nebulae (∼90 kpc; Yoshida et al. 2016) coinciding with soft X-ray emission. Contrary to those extended features, the H α and X-ray observations have also detected a compact ‘butterfly nebula (∼4.5 kpc) surrounding the nuclei, showing the signatures of ionized gas outflows (Komossa et al. 2003). Recent interferometric observations have revealed the presence of a broad molecular line profile (≤1400 km s−1) towards the centre (Tacconi et al. 1999; Ohyama, Yoshida & Takata 2003; Iono et al. 2007; Feruglio et al. 2013a,b). However, the spatial and velocity structures of the central kpc of NGC 6240 are still unclear mainly because of the limited sensitivity, angular resolution, and/or bandwidth of previous observations. Also, the complex velocity structures due to the ongoing violent merging event prevent us from modelling and revealing the underlying gaseous structures. In this Letter, we present Atacama Large Millimeter/submillimeter Array (ALMA) Band 6 observations of NGC 6240 in the CO (2–1) line emission and its underlying continuum with sub-arcsecond resolution. The goal of this Letter is to reveal the complete spatial and velocity distribution of the molecular outflow, which has been previously only partly mapped, in NGC 6240. We assumed H0 = 70 km s−1 Mpc−1, ΩM = 0.3, ΩΛ = 0.7 throughout this Letter. 2 ALMA BAND 6 OBSERVATION NGC 6240 was observed on 2016 June 26 for the ALMA Cycle 3 program ID = 2015.1.00003.S using 42 12-m antennas. The Band 6 receiver was tuned to cover CO (2–1) (νobs = 225.02928 GHz). The field of view, on-source time, and baseline lengths are ∼28 arcsec, ∼4 min, and 15.1–783.5 m, respectively. One of the spectral windows (spw) containing CO has a resolution of 1.953 MHz (∼2.6 km s−1) with a bandwidth of 1.875 GHz. The single sideband system temperature after flagging is 50–100 K with two peaks of ∼190 K at atmospheric absorptions. Precipitable water vapour during the observations towards NGC 6240 and all calibrators are 0.68–0.82 mm. Titan, J1550+0527, and J1651+0129 were observed as the amplitude, bandpass, and phase calibrators, respectively. We used the delivered calibrated uv data. The flux of J1550+0527 at each spw is in agreement with the ALMA calibrator source catalogue1 (a flux measurement obtained on the same date). Thus, we regard the accuracy of the absolute flux calibration as 5 per cent throughout this Letter. The data reduction was carried out using casa version 4.5.3 (McMullin et al. 2007). All maps are reconstructed with natural weighting. We used the casa task clean in multiscale mode to make use of the multiscale clean deconvolution algorithm (ms-clean; Cornwell 2008). ms-clean can reduce extended low surface brightness structures in the residual image, and thus recover extended emission significantly larger than the synthesized beam size. For the deconvolution process using clean, we performed an iterative auto-masking procedure as described in the casa guide.2 In order to improve the image fidelity, we carried out a two-rounded phase self-calibration after the normal ALMA calibration process. We chose the bright compact source corresponding to the peak of the CO spectrum (∼7050 km s−1) for the model used in the self-calibration. Continuum emission was subtracted in the uv-plane by fitting the line-free channels in both lower and upper sidebands with a first order polynomial function, and then we made a CO (2–1) data cube with 5 km s−1 resolution. The synthesized beam size and sensitivity are 0.97 × 0.53 arcsec2 (PA = 64$$_{.}^{\circ}$$5) and 2.5 mJy beam−1 (at 5 km s−1 bin), respectively. 3 RESULTS The CO (2–1) integrated intensity contours overlaid on a velocity field map are shown in Fig. 1(a). It shows a bright nuclear concentration (∼700 pc, a rough size of the contour that has 50 per cent of the peak value by eye) and extended filamentary structures (length ∼ 11 kpc), which coincide with the dust lanes in the optical image (Yoshida et al. 2016). Both features match the CO (1–0) emission as reported by Feruglio et al. (2013a). The extended CO emission is compact compared to the H α distribution (∼90 kpc) and the remnant of the main stellar disc (∼20 kpc) (Yoshida et al. 2016). There are extremely broad, asymmetric velocity wings, ranging from ∼6200 to ∼8200 km s−1, in the radio local standard of rest (kinematic) (LSRK) velocity frame [full width at zero intensity (FWZI) ∼ 2000 km s−1], within the central 4-arcsec aperture (Fig. 1a and 1b). This line width is broader than that previously reported for CO lines (e.g. Feruglio et al. 2013a), because of the higher sensitivity and wider bandwidth of our CO data. The observed total integrated intensity inside the 3σ contour is 1283 ± 64 Jy km s−1, which is 86 ± 15  per cent of the single-dish value (Papadopoulos et al. 2014). We also detected CN (25/2–13/2) and CS (3–2) lines, which will be presented in a future paper. Figure 1. View largeDownload slide (a) CO (2–1) velocity field (moment 1) image of NGC 6240. The velocity field in colour scale ranges from 6600 to 7600 km s−1. Contours show the CO (2–1) integrated intensity (moment 0) map. The contours are 208 × (0.125, 0.25, 0.5, 1.0, 2.0, 4.0, 8.0, 16.0, 32.0 64.0, and 96.0) Jy beam−1 km s−1. The crosses indicate the peak positions of the Band 6 continuum emission. (b) CO (2–1) spectrum towards the central 4 arcsec aperture, which is shown in Fig. 1a, with a zoomed panel to emphasize the wings. Figure 1. View largeDownload slide (a) CO (2–1) velocity field (moment 1) image of NGC 6240. The velocity field in colour scale ranges from 6600 to 7600 km s−1. Contours show the CO (2–1) integrated intensity (moment 0) map. The contours are 208 × (0.125, 0.25, 0.5, 1.0, 2.0, 4.0, 8.0, 16.0, 32.0 64.0, and 96.0) Jy beam−1 km s−1. The crosses indicate the peak positions of the Band 6 continuum emission. (b) CO (2–1) spectrum towards the central 4 arcsec aperture, which is shown in Fig. 1a, with a zoomed panel to emphasize the wings. We implemented one and two Gaussian fits to disentangle outflow signatures from the chaotic velocity field of NGC 6240 (Fig. 1a). A hexagonal Nyquist sampling with 2.0-arcsec aperture (step = 1.0 arcsec) was performed on the CO (2–1) cube to decrease the number of spectra (from 500 × 500 pixels to 40 × 40 apertures) for the fit. We fit the data as follows: (1) perform a one-component fit for spectra with a peak S/N > 5, and then (2) if the residual of the one-component fit has a peak S/N > 5, perform a two-component fit to the data (not the residual). Thus, the resultant best-fitting model could have one or two Gaussian component(s) at each aperture. For the one-component fit, the three initial parameters [i.e. peak, velocity, and full width at half-maximum (FWHM)] were chosen from the peak flux, peak velocity, and the number of channels, which satisfy peak S/N > 5. For the two-component fit, initial parameters of the first Gaussian are obtained from the one-component fit, although those of the second Gaussian were chosen from the residual of the one-component fit following the procedure used for estimating the initial guess of the one-component fit. Examples of multiple Gaussian fits that require two Gaussian components are shown in Fig. 2. The peaks seen in the residual of the two-component fit are statistically insignificant (peak S/N < 5). Thus, we did not fit any third Gaussian component. Figure 2. View largeDownload slide CO (2–1) spectrum (blue), residual (green), and two-component Gaussian model (red = narrow, purple = broad, and black = narrow + broad) within 2.0 arcsec aperture at the four highest velocity dispersion positions of the CO (2–1) map. These four positions are shown in Fig. 3c. Figure 2. View largeDownload slide CO (2–1) spectrum (blue), residual (green), and two-component Gaussian model (red = narrow, purple = broad, and black = narrow + broad) within 2.0 arcsec aperture at the four highest velocity dispersion positions of the CO (2–1) map. These four positions are shown in Fig. 3c. Fig. 3(a) shows the histogram of the CO (2–1) line width in FWHM from every Gaussian component from our fitting procedure. The small line width components (≤240 km s−1) are mainly found along the optical dust lanes, which is similar to that of giant molecular associations found in the nearby merging galaxy the antennas (≤210 km s−1; Espada et al. 2012; Ueda et al. 2012). On the other hand, the large line width components reach unusually large values of ∼1000 km s−1. We performed a Gaussian fit to this plot. The best-fitting Gaussian (green line) shows a peak at 49 km s−1 and the dispersion of 160 km s−1 (i.e. +1σ ≃ 210 km s−1). There are 12 per cent outliers larger than the +3σ limit (at FWHM ∼ 530 km s−1) of the Gaussian population, and thus, we simply define data whose line widths are larger than 500 km s−1 as ‘broad components’. The broad components account for ≲10 per cent of the total integrated intensity. The line width image of the broad components is shown in Fig. 3(b) and (c). The image area of Fig. 3(b) roughly corresponds to that of Fig. 1(a). Figure 3. View largeDownload slide (a) Distribution of the CO (2–1) line width in FWHM (km s−1) from every Gaussian component fit. The green line shows the best-fitting Gaussian to the histogram, showing that there are outliers higher than 500 km s−1 (i.e. broad component). The black dashed line marks where FWHM = 500 km s−1. (b) CO (2–1) line width map of the broad component in FHWM (km s−1). The green contour shows the outline of the H α (‘butterfly nebula’). Black dots show the centre position of each aperture. (c) Zoomed-in CO (2–1) line width map of the broad components with the colour bar showing the line width range. Figure 3. View largeDownload slide (a) Distribution of the CO (2–1) line width in FWHM (km s−1) from every Gaussian component fit. The green line shows the best-fitting Gaussian to the histogram, showing that there are outliers higher than 500 km s−1 (i.e. broad component). The black dashed line marks where FWHM = 500 km s−1. (b) CO (2–1) line width map of the broad component in FHWM (km s−1). The green contour shows the outline of the H α (‘butterfly nebula’). Black dots show the centre position of each aperture. (c) Zoomed-in CO (2–1) line width map of the broad components with the colour bar showing the line width range. 4 DISCUSSION The origin of the CO velocity wings in the central region of NGC 6240 has been discussed for years, and recent studies based on interferometric observations suggested the presence of massive molecular outflow(s) (see Section 1). However, the spatial and velocity distributions of the molecular outflows are unclear yet mainly because of the complicated CO velocity structures and the instrumental limits. The broad component is located around the central 3.4 kpc, although large line width regions are not located at the nuclear positions. The line width of the broad components peaks at four positions around the nuclei, whose spectra are shown in Fig. 2. The broad components show a clearly different spatial distribution with respect to the dense gas tracer, HCN (4–3), which likely traces nuclear rotating discs (Scoville et al. 2015). The four peak positions of the broad components roughly coincide with peaks of H α (Gerssen et al. 2004), near-IR H2 (Max et al. 2005), and 0.5-1.5 keV X-ray (Komossa et al. 2003), indicating that the CO broad components arise from a multiphase ISM. This is also supported by the maximum velocity of the CO (2–1) emission (FWZI/2 ∼ 1000 km s−1; Fig. 1b), which is consistent with the maximum velocity of a warm molecular outflow traced by blueshifted OH absorption (∼1200 km s−1; Veilleux et al. 2013) and H α emission (≥1000 km s−1; Heckman, Armus & Miley 1990). As seen in other LIRGs (e.g. NGC 1068; García-Burillo et al. 2014, NGC 3256; Sakamoto et al. 2014, and NGC 1614; Saito et al. 2017), molecular outflows are often spatially extended (a few kpc) and concomitant with (or entrained by) ionized gas outflow. Considering all pieces of evidence described above and some similarities with other LIRGs, we suggest that the four peaks of the broad CO (2–1) components trace molecular outflows in NGC 6240. We derive physical parameters of the molecular outflows found in NGC 6240. The spatially resolved map (Fig. 3c) allows us to estimate the mass and projected distance of the line width peaks from the launching points. We assume the launching points are the double nuclei, that is, the northern and eastern (southern and western) components of the molecular outflow come from the northern (southern) nucleus (see grey arrows in Fig. 4). This spatial configuration is supported by radio continuum observations with very long baseline interferometers (Gallimore & Beswick 2004; Hagiwara, Baan & Klöckner 2011), showing that the northern nucleus has an east–west bipolar structure. The direction of those bipolar structures is similar to our assumed morphology of the molecular outflows. We also assume that the inclination (α), CO (2–1)/CO (1–0) line intensity ratio, and CO (1–0) luminosity to H2 mass conversion factor of the molecular outflows in NGC 6240 are 45°, unity, and 0.8 M⊙ (K km s−1 pc2)−1, respectively, for the sake of simplicity. Figure 4. View largeDownload slide Peak velocity map of the broad component. The velocity field in colour scale ranges from 6850 to 7350 km s−1. Data points around the four peaks in Fig. 3c are shown. Figure 4. View largeDownload slide Peak velocity map of the broad component. The velocity field in colour scale ranges from 6850 to 7350 km s−1. Data points around the four peaks in Fig. 3c are shown. To estimate the mass outflow rate ($$\dot{M}_{\rm H_2, out}$$), we employ the expression,   $$\dot{M}_{\rm H_2, out} = \frac{3v_{\rm out, proj}M_{\rm H_2, out}}{R_{\rm out,\, proj}}\tan {\alpha },$$ (1)where vout, proj is the projected velocity of the outflow, $$M_{\rm H_2, out}$$ is the gas mass of the outflow, Rout, proj is the distance from the launching point to the outflowing gas, and α is the inclination of the outflow. This equation assumes a uniformly filled cone geometry (Maiolino et al. 2012). All derived parameters related to $$\dot{M}_{\rm H_2, out}$$ are listed in Table 1. vout, proj of the northern and eastern (southern and western) outflows is derived by using the projected velocity of the outflowing gas (Fig. 4) and the systemic velocity of the northern (southern) nucleus of 7122 (7164) km s−1 in the radio LSRK velocity frame (Hagiwara 2010). Table 1. NGC 6240 molecular outflow properties. Property  East  South  West  North  Total  Rout, proj (pc)  1960  1240  1370  830  –  vout, proj (km s−1)  −70  +150  −270  +180  –  age (Myr)  27  8  5  5  –  $$\log M_{\rm H_2, out}$$ (M⊙)  7.9  8.1  8.2  8.2  8.7  $$\dot{M}_{\rm H_2, out}$$ (M⊙ yr−1)  8  43  86  94  231  Property  East  South  West  North  Total  Rout, proj (pc)  1960  1240  1370  830  –  vout, proj (km s−1)  −70  +150  −270  +180  –  age (Myr)  27  8  5  5  –  $$\log M_{\rm H_2, out}$$ (M⊙)  7.9  8.1  8.2  8.2  8.7  $$\dot{M}_{\rm H_2, out}$$ (M⊙ yr−1)  8  43  86  94  231  View Large The derived mass of each broad component is ∼107.9–8.2 M⊙ (∼108.7 M⊙ in total, which is consistent with the value derived from unresolved CO data; Cicone et al. 2014). Since the ionized gas mass of the butterfly nebula is <1.4 × 108 M⊙ (Yoshida et al. 2016), >70  per cent of the total gas mass in the outflow consists of molecules. The total $$\dot{M}_{\rm H_2, out}$$ of ∼230 M⊙ yr−1 is 3.5 times lower than that estimated by Cicone et al. (2014). This is due to the difference of the assumed geometry. Cicone et al. (2014) assumed a spherical geometry because their data did not spatially resolve the broad CO component, and we assumed a conical geometry with a certain inclination. Using the nuclear SFR of ∼61 M⊙ yr−1 derived from radio-to-FIR SED fitting (Yun & Carilli 2002), the mass loading factor ($$\dot{M}_{\rm H_2, out}$$ divided by SFR) is ∼4, indicating that either the AGN or the starburst, or both are the main drivers of the outflows. We also estimated the age (= Rout, proj/vout, proj assuming the inclination of 45°) of each broad component, and found that it ranges from 5 to 27 Myr. This is consistent with the age of the central H α structures from the butterfly nebula (8.4 Myr) to the ‘hourglass (24 Myr) (Yoshida et al. 2016), indicating that, again, the H α and CO (2–1) outflows are co-located. Note that, although the spatial characteristics of the broad CO components in NGC 6240 could also be explained by inflow motions (e.g. Gaspari, Temi & Brighenti 2017) rather than outflow, we favour the latter because the absorption profile of the OH doublet towards the central 9 arcsec shows that a high velocity component faster than 1000 km s−1 is only detected on the blueshifted side (Veilleux et al. 2013). 5 CONCLUSION We present the ALMA high-resolution observations of CO (2–1) line towards the prototypical ULIRG NGC 6240, which is known to have broad CO profiles around the centre, although no one has succeeded in extracting their spatial distributions. Our high sensitivity and wide-band ALMA data revealed the presence of extremely broad CO wings (FWZI ∼ 2000 km s−1), which are as fast as the OH and H α outflows. We performed multiple Gaussian fitting for the CO data cube to visualize the morphology of the wings. We found that for the first time the broad component presents four peaks 1–2 kpc away from the double nuclei. We also found a spatial connection between nuclear bipolar structures in radio, CO, and the H α and X-ray emission, and thus we suggest that the CO wings are associated with twin east–west bipolar molecular outflows launched from the nuclei. ACKNOWLEDGEMENTS The authors thanks an anonymous referee for comments that improved the contents of this Letter. TS thanks T. H. Saitoh and R. Maiolino for useful discussion. TS and the other authors thank the ALMA staff for their kind support. NL acknowledges support by National Key R&D Program of China #2017YFA0402704 and NSFC #11673028. This Letter makes use of the following ALMA data: ADS/JAO.ALMA#2015.1.00003.S. ALMA is a partnership of ESO (representing its member states), NSF (USA), and NINS (Japan), together with NRC (Canada), NSC and ASIAA (Taiwan), and KASI (Republic of Korea), in cooperation with the Republic of Chile. The Joint ALMA Observatory is operated by ESO, AUI/NRAO, and NAOJ. Footnotes 1 https://almascience.nrao.edu/sc/ 2 https://casaguides.nrao.edu/index.php/M100_Band3_Combine_4.3#Image_Using_An_Automasking_Technique REFERENCES Armus L. et al.  , 2009, PASP , 121, 559 CrossRef Search ADS   Barnes J. E., Hernquist L. E., 1991, ApJ , 370, L65 CrossRef Search ADS   Bellocchi E., Arribas S., Colina L., Miralles-Caballero D., 2013, A&A , 557, A59 CrossRef Search ADS   Cicone C. et al.  , 2014, A&A , 562, A21 CrossRef Search ADS   Cornwell T. J., 2008, IEEE J. Sel. Top. Signal Process. , 2, 793 CrossRef Search ADS   Costa T., Sijacki D., Haehnelt M. G., 2014, MNRAS , 444, 2355 CrossRef Search ADS   Espada D. et al.  , 2012, ApJ , 760, L25 CrossRef Search ADS   Feruglio C. et al.  , 2010, A&A , 518, L155 CrossRef Search ADS   Feruglio C. et al.  , 2013, A&A , 549, A51 CrossRef Search ADS   Feruglio C. et al.  , 2013, A&A , 558, A87 CrossRef Search ADS   Fiore F. et al.  , 2017, A&A , 601, A143 CrossRef Search ADS   Gallimore J. F., Beswick R., 2004, AJ , 127, 239 CrossRef Search ADS   García-Burillo S. et al.  , 2014, A&A , 567, A125 CrossRef Search ADS   Gaspari M., Temi P., Brighenti F., 2017, MNRAS , 466, 677 CrossRef Search ADS   Gerssen J., van der Marel R. P., Axon D., Mihos J. C., Hernquist L., Barnes J. E., 2004, AJ , 127, 75 CrossRef Search ADS   Hagiwara Y., 2010, AJ , 140, 1905 CrossRef Search ADS   Hagiwara Y., Baan W. A., Klöckner H.-R., 2011, AJ , 142, 17 CrossRef Search ADS   Heckman T. M., Armus L., Miley G. K., 1990, ApJS , 74, 833 CrossRef Search ADS   Hopkins P. F., Hernquist L., Cox T. J., Kereš D., 2008, ApJS , 175, 356 CrossRef Search ADS   Ilha G. d. S., Bianchin M., Riffel R. A., 2016, Ap&SS , 361, 178 CrossRef Search ADS   Iono D. et al.  , 2007, ApJ , 659, 283 CrossRef Search ADS   Komossa S., Burwitz V., Hasinger G., Predehl P., Kaastra J. S., Ikebe Y., 2003, ApJ , 582, L15 CrossRef Search ADS   Lu N. et al.  , 2015, ApJ , 802, L11 CrossRef Search ADS   Maiolino R. et al.  , 2012, MNRAS , 425, L66 CrossRef Search ADS   Max C. E., Canalizo G., Macintosh B. A., Raschke L., Whysong D., Antonucci R., Schneider G., 2005, ApJ , 621, 738 CrossRef Search ADS   McMullin J. P., Waters B., Schiebel D., Young W., Golap K., 2007, in Shaw R. A., Hill F., Bell D. J., eds, ASP Conf. Ser. Vol. 376, Astronomical Data Analysis Software and Systems XVI . Astron. Soc. Pac., San Francisco, p. 127 Meijerink R. et al.  , 2013, ApJ , 762, L16 CrossRef Search ADS   Nakanishi K., Okumura S. K., Kohno K., Kawabe R., Nakagawa T., 2005, PASJ , 57, 575 CrossRef Search ADS   Ohyama Y., Yoshida M., Takata T., 2003, AJ , 126, 2291 CrossRef Search ADS   Papadopoulos P. P. et al.  , 2014, ApJ , 788, 153 CrossRef Search ADS   Saito T. et al.  , 2017, ApJ , 835, 174 CrossRef Search ADS   Sakamoto K., Aalto S., Combes F., Evans A., Peck A., 2014, ApJ , 797, 90 CrossRef Search ADS   Scoville N. et al.  , 2015, ApJ , 800, 70 CrossRef Search ADS   Sliwa K., Downes D., 2017, A&A , 604, A2 CrossRef Search ADS   Stierwalt S. et al.  , 2014, ApJ , 790, 124 CrossRef Search ADS   Tacconi L. J., Genzel R., Tecza M., Gallimore J. F., Downes D., Scoville N. Z., 1999, ApJ , 524, 732 CrossRef Search ADS   Tunnard R., Greve T. R., Garcia-Burillo S., Graci Carpio J., Fuente A., Tacconi L., Neri R., Usero A., 2015, ApJ , 815, 114 CrossRef Search ADS   Ueda J. et al.  , 2012, ApJ , 745, 65 CrossRef Search ADS   Urrutia T., Lacy M., Becker R. H., 2008, ApJ , 674, 80 CrossRef Search ADS   Veilleux S. et al.  , 2013, ApJ , 776, 27 CrossRef Search ADS   Yoshida M., Yagi M., Ohyama Y., Komiyama Y., Kashikawa N., Tanaka H., Okamura S., 2016, ApJ , 820, 48 CrossRef Search ADS   Yun M. S., Carilli C. L., 2002, ApJ , 568, 88 CrossRef Search ADS   © 2017 The Author(s) Published by Oxford University Press on behalf of the Royal Astronomical Society http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Monthly Notices of the Royal Astronomical Society: Letters Oxford University Press

# Imaging the molecular outflows of the prototypical ULIRG NGC 6240 with ALMA

, Volume 475 (1) – Mar 1, 2018
5 pages

/lp/ou_press/imaging-the-molecular-outflows-of-the-prototypical-ulirg-ngc-6240-with-fsvfZ7zLA8
Publisher
Oxford University Press
ISSN
1745-3925
eISSN
1745-3933
D.O.I.
10.1093/mnrasl/slx207
Publisher site
See Article on Publisher Site

### Abstract

Abstract We present 0.97 × 0.53 arcsec2 (470 pc × 250 pc) resolution CO (J = 2–1) observations towards the nearby luminous merging galaxy NGC 6240 with the Atacama Large Millimeter/submillimeter Array. We confirmed a strong CO concentration within the central 700 pc, which peaks between the double nuclei, surrounded by extended CO features along the optical dust lanes (∼11 kpc). We found that the CO emission around the central, a few kpc, has extremely broad velocity wings with full width at zero intensity ∼ 2000 km s−1, suggesting a possible signature of molecular outflow(s). In order to extract and visualize the high-velocity components in NGC 6240, we performed a multiple Gaussian fit to the CO data cube. The distribution of the broad CO components shows four extremely large line width regions (∼1000 km s−1) located 1–2 kpc away from both nuclei. Spatial coincidence of the large line width regions with H α, near-IR H2, and X-ray suggests that the broad CO (2–1) components are associated with nuclear outflows launched from the double nuclei. galaxies: active, galaxies: evolution, galaxies: individual: NGC 6240, galaxies: interactions 1 INTRODUCTION Gas-rich galaxy mergers play a major role in the formation and evolution of galaxies by triggering intense star formation and changing their morphology as suggested by numerical simulations (e.g. Barnes & Hernquist 1991). During the process of a galaxy merger, radial streaming can feed gas to the central supermassive black hole (SMBH; e.g. Hopkins et al. 2008), possibly reaching at the quasar-phase after dispersing the obscuring gas and dust (Urrutia, Lacy & Becker 2008). Powerful winds or outflows, driven by the active galactic nucleus (AGN) or the surrounding starbursts [(ultra-)luminous infrared galaxy, (U)LIRG], are predicted to suppress gas feeding to the central SMBH and its host galaxy's spheroidal component in order to explain the empirical correlation between the stellar velocity dispersion of a galaxy bulge and the mass of the central SMBH (Costa, Sijacki & Haehnelt 2014). Observational studies of such galactic winds/outflows in (U)LIRGs have been frequently done by spectral analyses of hot molecular gas, atomic gas, and ionized gas emission/absorption lines (e.g. Bellocchi et al. 2013; Veilleux et al. 2013). Those tracers are bright enough to probe generally faint, broad outflowing gas profiles, although they are not likely to be major constituents of galactic outflows. Cold molecular gas is likely to dominate the outflow in terms of mass, and thus it is the key to understanding fundamental feedback effects of galactic outflows on the surrounding interstellar medium (ISM; Feruglio et al. 2010; Cicone et al. 2014; Fiore et al. 2017). NGC 6240 is a well-studied close galaxy pair in the local Universe (z = 0.024480, 1.0 arcsec = 480 pc), which is often regarded as a prototype of ULIRGs, even though its infrared (IR) luminosity is slightly lower than 1012 L⊙ (LIR = 1011.93 L⊙; Armus et al. 2009). This system has a double nucleus detected in a variety of wavelengths from X-ray to radio (e.g. Komossa et al. 2003; Hagiwara, Baan & Klöckner 2011; Stierwalt et al. 2014; Ilha, Bianchin & Riffel 2016), indicating the presence of two separated AGNs. The cold molecular ISM in NGC 6240 has been observed through many rotational transitions of CO, HCN, and HCO+, showing a strong gas concentration between the nuclei (e.g. Nakanishi et al. 2005; Iono et al. 2007; Papadopoulos et al. 2014; Scoville et al. 2015; Tunnard et al. 2015; Sliwa & Downes 2017). Although the properties shown above are indeed suitable for representing nearby ULIRGs, past observations have revealed a different side of NGC 6240, which makes it rather unique. Lu et al. (2015) reported that NGC 6240 shows an order of magnitude higher LCO(7–6)/LIR ratio than other local (U)LIRGs. Such an extreme CO line-to-continuum ratio can be explained by a simple C-shock model (Meijerink et al. 2013). Furthermore, deep and wide Subaru observations revealed a vastly extended, filamentary H α nebulae (∼90 kpc; Yoshida et al. 2016) coinciding with soft X-ray emission. Contrary to those extended features, the H α and X-ray observations have also detected a compact ‘butterfly nebula (∼4.5 kpc) surrounding the nuclei, showing the signatures of ionized gas outflows (Komossa et al. 2003). Recent interferometric observations have revealed the presence of a broad molecular line profile (≤1400 km s−1) towards the centre (Tacconi et al. 1999; Ohyama, Yoshida & Takata 2003; Iono et al. 2007; Feruglio et al. 2013a,b). However, the spatial and velocity structures of the central kpc of NGC 6240 are still unclear mainly because of the limited sensitivity, angular resolution, and/or bandwidth of previous observations. Also, the complex velocity structures due to the ongoing violent merging event prevent us from modelling and revealing the underlying gaseous structures. In this Letter, we present Atacama Large Millimeter/submillimeter Array (ALMA) Band 6 observations of NGC 6240 in the CO (2–1) line emission and its underlying continuum with sub-arcsecond resolution. The goal of this Letter is to reveal the complete spatial and velocity distribution of the molecular outflow, which has been previously only partly mapped, in NGC 6240. We assumed H0 = 70 km s−1 Mpc−1, ΩM = 0.3, ΩΛ = 0.7 throughout this Letter. 2 ALMA BAND 6 OBSERVATION NGC 6240 was observed on 2016 June 26 for the ALMA Cycle 3 program ID = 2015.1.00003.S using 42 12-m antennas. The Band 6 receiver was tuned to cover CO (2–1) (νobs = 225.02928 GHz). The field of view, on-source time, and baseline lengths are ∼28 arcsec, ∼4 min, and 15.1–783.5 m, respectively. One of the spectral windows (spw) containing CO has a resolution of 1.953 MHz (∼2.6 km s−1) with a bandwidth of 1.875 GHz. The single sideband system temperature after flagging is 50–100 K with two peaks of ∼190 K at atmospheric absorptions. Precipitable water vapour during the observations towards NGC 6240 and all calibrators are 0.68–0.82 mm. Titan, J1550+0527, and J1651+0129 were observed as the amplitude, bandpass, and phase calibrators, respectively. We used the delivered calibrated uv data. The flux of J1550+0527 at each spw is in agreement with the ALMA calibrator source catalogue1 (a flux measurement obtained on the same date). Thus, we regard the accuracy of the absolute flux calibration as 5 per cent throughout this Letter. The data reduction was carried out using casa version 4.5.3 (McMullin et al. 2007). All maps are reconstructed with natural weighting. We used the casa task clean in multiscale mode to make use of the multiscale clean deconvolution algorithm (ms-clean; Cornwell 2008). ms-clean can reduce extended low surface brightness structures in the residual image, and thus recover extended emission significantly larger than the synthesized beam size. For the deconvolution process using clean, we performed an iterative auto-masking procedure as described in the casa guide.2 In order to improve the image fidelity, we carried out a two-rounded phase self-calibration after the normal ALMA calibration process. We chose the bright compact source corresponding to the peak of the CO spectrum (∼7050 km s−1) for the model used in the self-calibration. Continuum emission was subtracted in the uv-plane by fitting the line-free channels in both lower and upper sidebands with a first order polynomial function, and then we made a CO (2–1) data cube with 5 km s−1 resolution. The synthesized beam size and sensitivity are 0.97 × 0.53 arcsec2 (PA = 64$$_{.}^{\circ}$$5) and 2.5 mJy beam−1 (at 5 km s−1 bin), respectively. 3 RESULTS The CO (2–1) integrated intensity contours overlaid on a velocity field map are shown in Fig. 1(a). It shows a bright nuclear concentration (∼700 pc, a rough size of the contour that has 50 per cent of the peak value by eye) and extended filamentary structures (length ∼ 11 kpc), which coincide with the dust lanes in the optical image (Yoshida et al. 2016). Both features match the CO (1–0) emission as reported by Feruglio et al. (2013a). The extended CO emission is compact compared to the H α distribution (∼90 kpc) and the remnant of the main stellar disc (∼20 kpc) (Yoshida et al. 2016). There are extremely broad, asymmetric velocity wings, ranging from ∼6200 to ∼8200 km s−1, in the radio local standard of rest (kinematic) (LSRK) velocity frame [full width at zero intensity (FWZI) ∼ 2000 km s−1], within the central 4-arcsec aperture (Fig. 1a and 1b). This line width is broader than that previously reported for CO lines (e.g. Feruglio et al. 2013a), because of the higher sensitivity and wider bandwidth of our CO data. The observed total integrated intensity inside the 3σ contour is 1283 ± 64 Jy km s−1, which is 86 ± 15  per cent of the single-dish value (Papadopoulos et al. 2014). We also detected CN (25/2–13/2) and CS (3–2) lines, which will be presented in a future paper. Figure 1. View largeDownload slide (a) CO (2–1) velocity field (moment 1) image of NGC 6240. The velocity field in colour scale ranges from 6600 to 7600 km s−1. Contours show the CO (2–1) integrated intensity (moment 0) map. The contours are 208 × (0.125, 0.25, 0.5, 1.0, 2.0, 4.0, 8.0, 16.0, 32.0 64.0, and 96.0) Jy beam−1 km s−1. The crosses indicate the peak positions of the Band 6 continuum emission. (b) CO (2–1) spectrum towards the central 4 arcsec aperture, which is shown in Fig. 1a, with a zoomed panel to emphasize the wings. Figure 1. View largeDownload slide (a) CO (2–1) velocity field (moment 1) image of NGC 6240. The velocity field in colour scale ranges from 6600 to 7600 km s−1. Contours show the CO (2–1) integrated intensity (moment 0) map. The contours are 208 × (0.125, 0.25, 0.5, 1.0, 2.0, 4.0, 8.0, 16.0, 32.0 64.0, and 96.0) Jy beam−1 km s−1. The crosses indicate the peak positions of the Band 6 continuum emission. (b) CO (2–1) spectrum towards the central 4 arcsec aperture, which is shown in Fig. 1a, with a zoomed panel to emphasize the wings. We implemented one and two Gaussian fits to disentangle outflow signatures from the chaotic velocity field of NGC 6240 (Fig. 1a). A hexagonal Nyquist sampling with 2.0-arcsec aperture (step = 1.0 arcsec) was performed on the CO (2–1) cube to decrease the number of spectra (from 500 × 500 pixels to 40 × 40 apertures) for the fit. We fit the data as follows: (1) perform a one-component fit for spectra with a peak S/N > 5, and then (2) if the residual of the one-component fit has a peak S/N > 5, perform a two-component fit to the data (not the residual). Thus, the resultant best-fitting model could have one or two Gaussian component(s) at each aperture. For the one-component fit, the three initial parameters [i.e. peak, velocity, and full width at half-maximum (FWHM)] were chosen from the peak flux, peak velocity, and the number of channels, which satisfy peak S/N > 5. For the two-component fit, initial parameters of the first Gaussian are obtained from the one-component fit, although those of the second Gaussian were chosen from the residual of the one-component fit following the procedure used for estimating the initial guess of the one-component fit. Examples of multiple Gaussian fits that require two Gaussian components are shown in Fig. 2. The peaks seen in the residual of the two-component fit are statistically insignificant (peak S/N < 5). Thus, we did not fit any third Gaussian component. Figure 2. View largeDownload slide CO (2–1) spectrum (blue), residual (green), and two-component Gaussian model (red = narrow, purple = broad, and black = narrow + broad) within 2.0 arcsec aperture at the four highest velocity dispersion positions of the CO (2–1) map. These four positions are shown in Fig. 3c. Figure 2. View largeDownload slide CO (2–1) spectrum (blue), residual (green), and two-component Gaussian model (red = narrow, purple = broad, and black = narrow + broad) within 2.0 arcsec aperture at the four highest velocity dispersion positions of the CO (2–1) map. These four positions are shown in Fig. 3c. Fig. 3(a) shows the histogram of the CO (2–1) line width in FWHM from every Gaussian component from our fitting procedure. The small line width components (≤240 km s−1) are mainly found along the optical dust lanes, which is similar to that of giant molecular associations found in the nearby merging galaxy the antennas (≤210 km s−1; Espada et al. 2012; Ueda et al. 2012). On the other hand, the large line width components reach unusually large values of ∼1000 km s−1. We performed a Gaussian fit to this plot. The best-fitting Gaussian (green line) shows a peak at 49 km s−1 and the dispersion of 160 km s−1 (i.e. +1σ ≃ 210 km s−1). There are 12 per cent outliers larger than the +3σ limit (at FWHM ∼ 530 km s−1) of the Gaussian population, and thus, we simply define data whose line widths are larger than 500 km s−1 as ‘broad components’. The broad components account for ≲10 per cent of the total integrated intensity. The line width image of the broad components is shown in Fig. 3(b) and (c). The image area of Fig. 3(b) roughly corresponds to that of Fig. 1(a). Figure 3. View largeDownload slide (a) Distribution of the CO (2–1) line width in FWHM (km s−1) from every Gaussian component fit. The green line shows the best-fitting Gaussian to the histogram, showing that there are outliers higher than 500 km s−1 (i.e. broad component). The black dashed line marks where FWHM = 500 km s−1. (b) CO (2–1) line width map of the broad component in FHWM (km s−1). The green contour shows the outline of the H α (‘butterfly nebula’). Black dots show the centre position of each aperture. (c) Zoomed-in CO (2–1) line width map of the broad components with the colour bar showing the line width range. Figure 3. View largeDownload slide (a) Distribution of the CO (2–1) line width in FWHM (km s−1) from every Gaussian component fit. The green line shows the best-fitting Gaussian to the histogram, showing that there are outliers higher than 500 km s−1 (i.e. broad component). The black dashed line marks where FWHM = 500 km s−1. (b) CO (2–1) line width map of the broad component in FHWM (km s−1). The green contour shows the outline of the H α (‘butterfly nebula’). Black dots show the centre position of each aperture. (c) Zoomed-in CO (2–1) line width map of the broad components with the colour bar showing the line width range. 4 DISCUSSION The origin of the CO velocity wings in the central region of NGC 6240 has been discussed for years, and recent studies based on interferometric observations suggested the presence of massive molecular outflow(s) (see Section 1). However, the spatial and velocity distributions of the molecular outflows are unclear yet mainly because of the complicated CO velocity structures and the instrumental limits. The broad component is located around the central 3.4 kpc, although large line width regions are not located at the nuclear positions. The line width of the broad components peaks at four positions around the nuclei, whose spectra are shown in Fig. 2. The broad components show a clearly different spatial distribution with respect to the dense gas tracer, HCN (4–3), which likely traces nuclear rotating discs (Scoville et al. 2015). The four peak positions of the broad components roughly coincide with peaks of H α (Gerssen et al. 2004), near-IR H2 (Max et al. 2005), and 0.5-1.5 keV X-ray (Komossa et al. 2003), indicating that the CO broad components arise from a multiphase ISM. This is also supported by the maximum velocity of the CO (2–1) emission (FWZI/2 ∼ 1000 km s−1; Fig. 1b), which is consistent with the maximum velocity of a warm molecular outflow traced by blueshifted OH absorption (∼1200 km s−1; Veilleux et al. 2013) and H α emission (≥1000 km s−1; Heckman, Armus & Miley 1990). As seen in other LIRGs (e.g. NGC 1068; García-Burillo et al. 2014, NGC 3256; Sakamoto et al. 2014, and NGC 1614; Saito et al. 2017), molecular outflows are often spatially extended (a few kpc) and concomitant with (or entrained by) ionized gas outflow. Considering all pieces of evidence described above and some similarities with other LIRGs, we suggest that the four peaks of the broad CO (2–1) components trace molecular outflows in NGC 6240. We derive physical parameters of the molecular outflows found in NGC 6240. The spatially resolved map (Fig. 3c) allows us to estimate the mass and projected distance of the line width peaks from the launching points. We assume the launching points are the double nuclei, that is, the northern and eastern (southern and western) components of the molecular outflow come from the northern (southern) nucleus (see grey arrows in Fig. 4). This spatial configuration is supported by radio continuum observations with very long baseline interferometers (Gallimore & Beswick 2004; Hagiwara, Baan & Klöckner 2011), showing that the northern nucleus has an east–west bipolar structure. The direction of those bipolar structures is similar to our assumed morphology of the molecular outflows. We also assume that the inclination (α), CO (2–1)/CO (1–0) line intensity ratio, and CO (1–0) luminosity to H2 mass conversion factor of the molecular outflows in NGC 6240 are 45°, unity, and 0.8 M⊙ (K km s−1 pc2)−1, respectively, for the sake of simplicity. Figure 4. View largeDownload slide Peak velocity map of the broad component. The velocity field in colour scale ranges from 6850 to 7350 km s−1. Data points around the four peaks in Fig. 3c are shown. Figure 4. View largeDownload slide Peak velocity map of the broad component. The velocity field in colour scale ranges from 6850 to 7350 km s−1. Data points around the four peaks in Fig. 3c are shown. To estimate the mass outflow rate ($$\dot{M}_{\rm H_2, out}$$), we employ the expression,   $$\dot{M}_{\rm H_2, out} = \frac{3v_{\rm out, proj}M_{\rm H_2, out}}{R_{\rm out,\, proj}}\tan {\alpha },$$ (1)where vout, proj is the projected velocity of the outflow, $$M_{\rm H_2, out}$$ is the gas mass of the outflow, Rout, proj is the distance from the launching point to the outflowing gas, and α is the inclination of the outflow. This equation assumes a uniformly filled cone geometry (Maiolino et al. 2012). All derived parameters related to $$\dot{M}_{\rm H_2, out}$$ are listed in Table 1. vout, proj of the northern and eastern (southern and western) outflows is derived by using the projected velocity of the outflowing gas (Fig. 4) and the systemic velocity of the northern (southern) nucleus of 7122 (7164) km s−1 in the radio LSRK velocity frame (Hagiwara 2010). Table 1. NGC 6240 molecular outflow properties. Property  East  South  West  North  Total  Rout, proj (pc)  1960  1240  1370  830  –  vout, proj (km s−1)  −70  +150  −270  +180  –  age (Myr)  27  8  5  5  –  $$\log M_{\rm H_2, out}$$ (M⊙)  7.9  8.1  8.2  8.2  8.7  $$\dot{M}_{\rm H_2, out}$$ (M⊙ yr−1)  8  43  86  94  231  Property  East  South  West  North  Total  Rout, proj (pc)  1960  1240  1370  830  –  vout, proj (km s−1)  −70  +150  −270  +180  –  age (Myr)  27  8  5  5  –  $$\log M_{\rm H_2, out}$$ (M⊙)  7.9  8.1  8.2  8.2  8.7  $$\dot{M}_{\rm H_2, out}$$ (M⊙ yr−1)  8  43  86  94  231  View Large The derived mass of each broad component is ∼107.9–8.2 M⊙ (∼108.7 M⊙ in total, which is consistent with the value derived from unresolved CO data; Cicone et al. 2014). Since the ionized gas mass of the butterfly nebula is <1.4 × 108 M⊙ (Yoshida et al. 2016), >70  per cent of the total gas mass in the outflow consists of molecules. The total $$\dot{M}_{\rm H_2, out}$$ of ∼230 M⊙ yr−1 is 3.5 times lower than that estimated by Cicone et al. (2014). This is due to the difference of the assumed geometry. Cicone et al. (2014) assumed a spherical geometry because their data did not spatially resolve the broad CO component, and we assumed a conical geometry with a certain inclination. Using the nuclear SFR of ∼61 M⊙ yr−1 derived from radio-to-FIR SED fitting (Yun & Carilli 2002), the mass loading factor ($$\dot{M}_{\rm H_2, out}$$ divided by SFR) is ∼4, indicating that either the AGN or the starburst, or both are the main drivers of the outflows. We also estimated the age (= Rout, proj/vout, proj assuming the inclination of 45°) of each broad component, and found that it ranges from 5 to 27 Myr. This is consistent with the age of the central H α structures from the butterfly nebula (8.4 Myr) to the ‘hourglass (24 Myr) (Yoshida et al. 2016), indicating that, again, the H α and CO (2–1) outflows are co-located. Note that, although the spatial characteristics of the broad CO components in NGC 6240 could also be explained by inflow motions (e.g. Gaspari, Temi & Brighenti 2017) rather than outflow, we favour the latter because the absorption profile of the OH doublet towards the central 9 arcsec shows that a high velocity component faster than 1000 km s−1 is only detected on the blueshifted side (Veilleux et al. 2013). 5 CONCLUSION We present the ALMA high-resolution observations of CO (2–1) line towards the prototypical ULIRG NGC 6240, which is known to have broad CO profiles around the centre, although no one has succeeded in extracting their spatial distributions. Our high sensitivity and wide-band ALMA data revealed the presence of extremely broad CO wings (FWZI ∼ 2000 km s−1), which are as fast as the OH and H α outflows. We performed multiple Gaussian fitting for the CO data cube to visualize the morphology of the wings. We found that for the first time the broad component presents four peaks 1–2 kpc away from the double nuclei. We also found a spatial connection between nuclear bipolar structures in radio, CO, and the H α and X-ray emission, and thus we suggest that the CO wings are associated with twin east–west bipolar molecular outflows launched from the nuclei. ACKNOWLEDGEMENTS The authors thanks an anonymous referee for comments that improved the contents of this Letter. TS thanks T. H. Saitoh and R. Maiolino for useful discussion. TS and the other authors thank the ALMA staff for their kind support. NL acknowledges support by National Key R&D Program of China #2017YFA0402704 and NSFC #11673028. This Letter makes use of the following ALMA data: ADS/JAO.ALMA#2015.1.00003.S. ALMA is a partnership of ESO (representing its member states), NSF (USA), and NINS (Japan), together with NRC (Canada), NSC and ASIAA (Taiwan), and KASI (Republic of Korea), in cooperation with the Republic of Chile. The Joint ALMA Observatory is operated by ESO, AUI/NRAO, and NAOJ. Footnotes 1 https://almascience.nrao.edu/sc/ 2 https://casaguides.nrao.edu/index.php/M100_Band3_Combine_4.3#Image_Using_An_Automasking_Technique REFERENCES Armus L. et al.  , 2009, PASP , 121, 559 CrossRef Search ADS   Barnes J. E., Hernquist L. E., 1991, ApJ , 370, L65 CrossRef Search ADS   Bellocchi E., Arribas S., Colina L., Miralles-Caballero D., 2013, A&A , 557, A59 CrossRef Search ADS   Cicone C. et al.  , 2014, A&A , 562, A21 CrossRef Search ADS   Cornwell T. J., 2008, IEEE J. Sel. Top. Signal Process. , 2, 793 CrossRef Search ADS   Costa T., Sijacki D., Haehnelt M. G., 2014, MNRAS , 444, 2355 CrossRef Search ADS   Espada D. et al.  , 2012, ApJ , 760, L25 CrossRef Search ADS   Feruglio C. et al.  , 2010, A&A , 518, L155 CrossRef Search ADS   Feruglio C. et al.  , 2013, A&A , 549, A51 CrossRef Search ADS   Feruglio C. et al.  , 2013, A&A , 558, A87 CrossRef Search ADS   Fiore F. et al.  , 2017, A&A , 601, A143 CrossRef Search ADS   Gallimore J. F., Beswick R., 2004, AJ , 127, 239 CrossRef Search ADS   García-Burillo S. et al.  , 2014, A&A , 567, A125 CrossRef Search ADS   Gaspari M., Temi P., Brighenti F., 2017, MNRAS , 466, 677 CrossRef Search ADS   Gerssen J., van der Marel R. P., Axon D., Mihos J. C., Hernquist L., Barnes J. E., 2004, AJ , 127, 75 CrossRef Search ADS   Hagiwara Y., 2010, AJ , 140, 1905 CrossRef Search ADS   Hagiwara Y., Baan W. A., Klöckner H.-R., 2011, AJ , 142, 17 CrossRef Search ADS   Heckman T. M., Armus L., Miley G. K., 1990, ApJS , 74, 833 CrossRef Search ADS   Hopkins P. F., Hernquist L., Cox T. J., Kereš D., 2008, ApJS , 175, 356 CrossRef Search ADS   Ilha G. d. S., Bianchin M., Riffel R. A., 2016, Ap&SS , 361, 178 CrossRef Search ADS   Iono D. et al.  , 2007, ApJ , 659, 283 CrossRef Search ADS   Komossa S., Burwitz V., Hasinger G., Predehl P., Kaastra J. S., Ikebe Y., 2003, ApJ , 582, L15 CrossRef Search ADS   Lu N. et al.  , 2015, ApJ , 802, L11 CrossRef Search ADS   Maiolino R. et al.  , 2012, MNRAS , 425, L66 CrossRef Search ADS   Max C. E., Canalizo G., Macintosh B. A., Raschke L., Whysong D., Antonucci R., Schneider G., 2005, ApJ , 621, 738 CrossRef Search ADS   McMullin J. P., Waters B., Schiebel D., Young W., Golap K., 2007, in Shaw R. A., Hill F., Bell D. J., eds, ASP Conf. Ser. Vol. 376, Astronomical Data Analysis Software and Systems XVI . Astron. Soc. Pac., San Francisco, p. 127 Meijerink R. et al.  , 2013, ApJ , 762, L16 CrossRef Search ADS   Nakanishi K., Okumura S. K., Kohno K., Kawabe R., Nakagawa T., 2005, PASJ , 57, 575 CrossRef Search ADS   Ohyama Y., Yoshida M., Takata T., 2003, AJ , 126, 2291 CrossRef Search ADS   Papadopoulos P. P. et al.  , 2014, ApJ , 788, 153 CrossRef Search ADS   Saito T. et al.  , 2017, ApJ , 835, 174 CrossRef Search ADS   Sakamoto K., Aalto S., Combes F., Evans A., Peck A., 2014, ApJ , 797, 90 CrossRef Search ADS   Scoville N. et al.  , 2015, ApJ , 800, 70 CrossRef Search ADS   Sliwa K., Downes D., 2017, A&A , 604, A2 CrossRef Search ADS   Stierwalt S. et al.  , 2014, ApJ , 790, 124 CrossRef Search ADS   Tacconi L. J., Genzel R., Tecza M., Gallimore J. F., Downes D., Scoville N. Z., 1999, ApJ , 524, 732 CrossRef Search ADS   Tunnard R., Greve T. R., Garcia-Burillo S., Graci Carpio J., Fuente A., Tacconi L., Neri R., Usero A., 2015, ApJ , 815, 114 CrossRef Search ADS   Ueda J. et al.  , 2012, ApJ , 745, 65 CrossRef Search ADS   Urrutia T., Lacy M., Becker R. H., 2008, ApJ , 674, 80 CrossRef Search ADS   Veilleux S. et al.  , 2013, ApJ , 776, 27 CrossRef Search ADS   Yoshida M., Yagi M., Ohyama Y., Komiyama Y., Kashikawa N., Tanaka H., Okamura S., 2016, ApJ , 820, 48 CrossRef Search ADS   Yun M. S., Carilli C. L., 2002, ApJ , 568, 88 CrossRef Search ADS   © 2017 The Author(s) Published by Oxford University Press on behalf of the Royal Astronomical Society

### Journal

Monthly Notices of the Royal Astronomical Society: LettersOxford University Press

Published: Mar 1, 2018

## You’re reading a free preview. Subscribe to read the entire article.

### DeepDyve is your personal research library

It’s your single place to instantly
that matters to you.

over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month ### Explore the DeepDyve Library ### Search Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly ### Organize Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place. ### Access Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals. ### Your journals are on DeepDyve Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more. All the latest content is available, no embargo periods. DeepDyve ### Freelancer DeepDyve ### Pro Price FREE$49/month
\$360/year

Save searches from
PubMed

Create lists to

Export lists, citations