Abstract

The formaldehyde megamaser emission has been mapped for the three host galaxies IC 860, IRAS 15107+0724 and Arp 220. Elongated emission components are found at the nuclear centres of all galaxies with an extent ranging between 30 and 100 pc. These components are superposed on the peaks of the nuclear continuum. Additional isolated emission components are found superposed in the outskirts of the radio continuum structure. The brightness temperatures of the detected features ranges from 0.6 to 13.4 × 104 K, which confirms their masering nature. The masering scenario is interpreted as amplification of the radio continuum by foreground molecular gas that is pumped by far-infrared radiation fields in these starburst environments of the host galaxies.

1 INTRODUCTION

High-luminosity megamaser (MM) emission has been detected in external galaxies for four molecular species. Hydroxyl (OH) masers are found in the nuclear regions of ultra-luminous infrared galaxies ((U)LIRGs; Baan, Wood & Haschick 1982), water vapour (H2O) masers in accretion discs of active galactic nuclei (AGNs; Braatz et al. 2010; Kuo et al. 2015) and in interaction regions of jets and circumnuclear clouds (e.g. Claussen, Heiligman & Lo 1984; Haschick & Baan 1985; Middelberg et al. 2007), formaldehyde (H2CO) masers in the nuclear regions of ULIRGs (Baan, Güsten & Haschick 1986; Araya, Baan & Hofner 2004), and recently methanol (CH3OH) and silicon oxide (SiO) maser emission has been found in the starburst regions of (U)LIRGs as well as in nuclear feedback regions (Ellingsen et al. 2014; Wang et al. 2014; Chen et al. 2015).

A significant contribution to these observed molecular emissions may result from the amplification of the embedded or background radio continuum emission by foreground pumped molecular gas, as already proposed for the prototype OH MM source Arp 220 (IC 4553 – IRAS 15234+2354) (Baan et al. 1982). Rather than relying on the conventional high-gain maser scenario with a tunnel-like column of inverted molecules that amplifies some spontaneous seed emission, a large fraction of these OH MM emissions is produced by low-gain unsaturated amplification of the background radio continuum (Baan 1985). This scenario naturally accounts for the superposition of maser emission originating in molecular cloud structures with variable local pumping conditions and the extended radio continuum in the source (Baan 1989; Parra et al. 2005), resulting in both low- and high-brightness maser components.

Superposition of maser and continuum sources has already clearly been shown for the powerful OH MM sources such as Arp 220 (e.g. Baan et al. 1987), Mrk 273 (e.g. Klöckner & Baan 2004) and IRAS 17207−0054 (e.g. Momjian et al. 2006), as well as for the strong central components of the H2O MM source NGC 4258 (Miyoshi et al. 1995; Herrnstein et al. 1997) and other powerful sources such as NGC 3079 having an extended nuclear continuum structure (e.g. Haschick et al. 1990; Kondratko, Greenhill & Moran 2005). Both the Class I and II methanol MM emissions observed in nearby starburst galaxies also suggest that they are confined within the extended continuum emission (Ellingsen et al. 2014; Chen et al. 2015). A first map of the formaldehyde emission in Arp 220 also shows a superposition of line emission and continuum (Baan & Haschick 1995), and this study should confirm this composite maser amplification scenario. Similarly, Galactic OH and H2O maser sources lose much of their flux at long terrestrial and space interferometer baselines, indicating the contribution of extended low-brightness maser emission (Slysh et al. 2001).

Formaldehyde masering activity remains very rare in the Galaxy and in extragalactic sources, and the observed masers are relatively weak and difficult to find. Although formaldehyde absorption is widespread in the Galaxy, formaldehyde emission in the 4.829 GHz Ka= 101 − 111 transition is currently known to occur in only eight Galactic sources|$^{\vphantom{1}}$| (e.g. Araya et al. 2008; Ginsburg et al. 2015). The H2CO MM emission in the extragalactic sources IC 860 and IRAS 15107+0724, and Arp 220 have been confirmed (Baan, Haschick & Uglesich 1993; Araya, Baan & Hofner 2004; Mangum et al. 2008), while other sources require interferometric confirmation. In addition, the source NGC 6240 displays two emission components that probably coincide with lower brightness continuum structures and that cover the velocity range of the single-dish spectrum (Baan et al. 1993; Wang et al. 2014). While most extragalactic sources show absorption in both the 4.83 GHz Ka = 110 − 111 and the 14.5 GHz Ka= 211 − 212 transitions (Mangum et al. 2013), the few sources with ground-state emission also exhibit dominant absorption in the 14.5 GHz transition except for possible partial infilling of the line by emission. This study suggests that the localized emission in the ground state may also be accompanied by more distributed absorption. The detection of such weak localized emission components embedded within more extended absorption requires higher resolution observations.

The masering activity in extragalactic sources strongly depends on the availability of an appropriate pumping agent for the molecular species and favourable environmental conditions. The known OH MM emissions generally occur in (U)LIRGs where radiative pumping of the OH results from star formation induced dust infrared emissions (Baan 1989; Henkel & Wilson 1990; Darling & Giovanelli 2002). Since all three H2CO MM sources are also known as OH MMs, it may be assumed that the infrared radiation fields in the nuclei of these galaxies also contribute to the pumping of the formaldehyde.

Issues to be raised in this study of IC 860, IRAS 15107+0724 and Arp 220 are the spatial structure of the masering formaldehyde emission components, their superposition on the nuclear continuum emission and their velocity characteristics. The observed brightness temperature of these components and the radio continuum will be used confirm the masering nature of the line emission. This paper will also address the ability of the infrared radiation field to pump the masering activity of formaldehyde and to provide sufficient gains. A final issue to be considered is to seek evidence of more extended absorption in these sources.

2 MERLIN OBSERVATIONS

The 4.83 GHz H2CO Ka = 110 − 111 transition data for the sources IC 860 (IRAS 13126+2452), IRAS 15107+0724 and Arp 220 (IRAS 15234+2354) have been obtained with the MERLIN array using the antennas at Defford, Cambridge, Knockin, Darnhall, Jodrell Mk2, Lovell and Tabley.

IC 860 was observed from 2009 November for a total of approximately 35 h including the flux density calibrator, 3C 286 (7.57 Jy) and the phase reference source, 1318+225 (0.26 Jy). The total target on-source time was 23.4 h, part of which was affected by weather. IRAS 15107+0724 was also observed during 2009 November for a total of 36.7 h including the flux calibrator, 3C 286, baseline calibrator, OQ 208 (0.99 Jy) and phase reference source, 1509+054 (0.84 Jy). The target on-source time was 27.1 h, part of which was also affected by weather. The observations of both IRAS 15107+0724 and IC 860 have an observing bandwidth of 16 MHz centred at a sky frequency of 4768.4 MHz, which corresponds to a redshift at band centre for both sources of 3802.8 km s−1.

The heliocentric radial velocity of 3347 km s−1 for IC 860 indicates a distance of 46.0 Mpc and a spatial scale of 223 pc arcsec−1. The emission from IC 860 has been found to be close to its systemic velocity of 3911 km s−1 and not at the heliocentric velocity of 3347 km s−1 found in the literature. The heliocentric radial velocity of 3897 km s−1 for IRAS 15107+0724 indicates a distance 53.8 Mpc and a spatial scale of 261 pc arcsec−1.

The observations for Arp 220 were obtained in 2004 May and lasted for a total of 32 h with a target on-source time of 24.6 h. The flux density calibrator used was 3C 286 (7.60 Jy,) the baseline calibrator was OQ 208 (0.99 Jy) and the phase reference source were 1511+238 (0.80 Jy) and 1530+239 (0.28 Jy). The observations for Arp 220 had an observing bandwidth of 16 MHz centred at a frequency of 4744.4 MHz giving a centre velocity of 5292.6 km s−1 close to the peak velocity in the single-dish spectrum of 5430 km s−1. For Arp 220 at a radial velocity 5375 km s−1 and distance of 73 Mpc, the conversion from angular size to projected linear size is 356 pc arcsec−1. The projected separation between the two continuum peaks in Arp 220 is 0.9 arcsec or 321 pc (Sakamoto et al. 2008).

The data reduction procedure followed the standard AIPS routines for flagging, calibration, bandpass calibration, self-calibration of phase calibrator and imaging. The incremental phase and amplitude corrections obtained from the nearby phase calibrators were applied to the target source before imaging. The 31 channels in the observing bands had a channel width of 0.5 MHz, which corresponds to a velocity resolution of approximately 31.04 km s−1. The synthesized beam for IC 860 is 0.081 arcsec × 0.058 arcsec (18 pc × 13 pc) and 0.101 arcsec × 0.071 arcsec (26 pc × 18 pc) for IRAS 15107+0724. For Arp 220 the corresponding channel velocity resolution is 31.06 km s−1 and the synthesized beam size of Arp 220 is 0.079 arcsec × 0.068 arcsec (25 pc × 23 pc).

The continuum structure of the sources has initially been obtained from the bandpass-corrected data cube by averaging two or three channels at the edge of the cube that show no evidence of the remaining bandpass structure and no discernible line emission. However, in order to obtain more accurate outer contours of the continuum structure, the centre 75 per cent of the channel maps have been used to represent the continuum structure for IC 860 and IRAS 15107+0724, which then includes the weak line emission at the centre locations. The peak of the line emission for Arp 220 and also IC 860 and IRAS 15107+0724 is found to be close to the centre of the band, and the line emission in the centre channels has been obtained by subtraction of a flat continuum structure based on the edge channels across the whole cube. A flat baseline subtraction does not remove any existing channel structure in the band and only brings the continuum level close to zero. Only for the final results for IC 860, the continuum subtraction was based on only the low-frequency edge channels in order to take into account the possible line emission near the high-frequency edge. For Arp 220, the emission extends beyond the edge of the observing band, which may result in errors in the continuum subtraction. A larger observing band of these MERLIN observations would have helped to take away any uncertainty continuum subtractions.

The formaldehyde line emission features in all three sources are known to be very weak from single-dish observations (e.g. Araya et al. 2004) and special attention has been paid during the data reduction process in order to arrive at reliable results. Some of the data have also been reduced using Miriad, which produced the same results. Simple tests such as determining the emission structures by stacking the emission line channels also produced the same results as using moments. There is no evidence that any of the observed weaker features away from the nuclear region result from imperfect data reduction. All emission features in the maps are indeed found to be superposed on the radio continuum contours of the sources.

The results of these studies show some differences between line profiles obtained from the interferometric data and those obtained with single-dish experiments. Some differences may result from residual errors in calibration and continuum subtraction using a limited number of line-free edge channels. In addition, the presence of diffuse absorption against the radio continuum may account for differences when sampled with different beam sizes.

The analysis of the three target sources requires an evaluation of the observed brightness temperatures of the formaldehyde spectral components and the continuum structures. The brightness temperatures of a source component may be determined as (Kellermann & Owen 1988)
(1)
where the observing frequency ν is in unit of GHz, the component flux density Sν has a unit of Jy, the source sizes θmaj and θmin are in milliarcsec.

In this paper, distances were determined using a cosmological model with H0 = 73 km s−1 Mpc−1, ΩM = 0.27 and |$\Omega _\Lambda = 0.73$| (Spergel et al. 2007).

3 FORMALDEHYDE IN IC 860 – IRAS 13126+2452

3.1 IC 860 – formaldehyde emission

The 4.83 GHz formaldehyde Ka = 1 emission in IC 860 – IRAS 13126+2452 has first been observed with the Arecibo Observatory with a peak flux density of 2.0–2.2 mJy (Baan et al. 1993). A representative spectrum shows an asymmetric profile centred at 3860 km s−1 with a total width of 160 km s−1 as presented in Fig. 1 (Araya et al. 2004). IC 860 also exhibits HI absorption centred at 3866 km s−1 as well as a combined OH spectrum with absorption at 3850 km s−1 and emission at 4000 km s−1 (Schmelz et al. 1986). These spectra of IC 860 cover a total velocity range from 3700 to 4100 km s−1. Similarly, the formaldehyde spectrum of IC 860 may also extend from 3550 to 4100 km s−1with a weak emission feature at 3580 km s−1 and weak absorption features at 3700 and 4050 km s−1. This signature may also be recognized in the spectrum obtained with the Green Bank Telescope Mangum et al. (2013). These weak features may be further verified using single-dish or interferometric data.

A single-dish spectrum of the formaldehyde emission in IC 860 obtained at 4.83 GHz with the Arecibo radio telescope in 1993 (Araya et al. 2004).
Figure 1.

A single-dish spectrum of the formaldehyde emission in IC 860 obtained at 4.83 GHz with the Arecibo radio telescope in 1993 (Araya et al. 2004).

The formaldehyde line emission structure of IC 860 is presented in Fig. 2 as a zeroth-moment colour map overlaid on the radio continuum map. The moment map incorporates line emission features above a threshold of 0.8 mJy beam−1, while the rms in the line channel maps (away from the location of the source) is 0.31 mJy beam−1. The central line emission region extends 31 pc at PA = 130° with an additional extension to the east at PA = 270°. The region is centred on the nuclear continuum peak and its orientation agrees well with that of the two micron 2Mass image of the galaxy of PA = 150° (Skrutskie et al. 2006) and lies perpendicular to the north-west continuum extension. One weaker and significant emission feature, south-east of the nucleus, is found in the zeroth-moment map inside the continuum contours.

The H2CO emission structure superposed on the nuclear continuum structure of IC 860. The contoured continuum emission structure has a peak flux density of 6.42 mJy beam−1, which includes the weak line emission. The actual continuum peak intensity is 4.68 mJy beam−1. The continuum contour levels are 0.18 mJy beam−1 × (−1, 1, 2, 3, 4, 8, 16, 32) with an rms noise in the map of 0.077 mJy beam−1. The moment 0 map of the H2CO emission structure is presented as a colour scale map with a range of 120–840 mJy beam−1 km s−1. Two emissions components are found in the map as the prominent centre region and the weaker south-east region. The integrated peak of the centre line emission converts to 1.74 mJy beam−1.
Figure 2.

The H2CO emission structure superposed on the nuclear continuum structure of IC 860. The contoured continuum emission structure has a peak flux density of 6.42 mJy beam−1, which includes the weak line emission. The actual continuum peak intensity is 4.68 mJy beam−1. The continuum contour levels are 0.18 mJy beam−1 × (−1, 1, 2, 3, 4, 8, 16, 32) with an rms noise in the map of 0.077 mJy beam−1. The moment 0 map of the H2CO emission structure is presented as a colour scale map with a range of 120–840 mJy beam−1 km s−1. Two emissions components are found in the map as the prominent centre region and the weaker south-east region. The integrated peak of the centre line emission converts to 1.74 mJy beam−1.

The underlying continuum structure of the nuclear region of IC 860 consists of a compact component and extended components towards the north-west, the southwest and to the east. With an integrated flux density 15.31 mJy, the compact component has a peak flux density of 4.68 mJy beam−1. The brightness temperatures, Tb, of the central component (5.13 × 104 K) and the other components (in the range of 0.45 to 0.78 × 104) K are consistent with the occurrence of star formation in all continuum components (Condon 1992). In order to better define the outer continuum structure, the (weak) line emission close to the nuclear centre has been included in the continuum map in Fig. 2, which results in an increased peak flux density of 6.42 mJy beam−1. The parameters for the nuclear and circumnuclear components are presented in Table 1.

Table 1.

The continuum components.

SourceRAaDec.aContinuumBrightness
componentfluxtemperature
(s)( arcsec)(mJy b−1)(104 K)
IC 860
C03.50507.7914.685.13
E03.51607.7550.610.45
S03.50007.5750.780.63
W03.49807.8201.750.78
NW03.49707.9460.750.12
IRAS 15107+0724
C13.09431.8727.929.40
NW13.09132.0300.720.53
W13.08331.8470.470.34
SW13.08831.6160.480.35
S13.09531.5870.560.41
SE13.10331.8230.370.27
Arp 220
W57.09431.87630.4230.2
E57.09132.03013.3813.3
SourceRAaDec.aContinuumBrightness
componentfluxtemperature
(s)( arcsec)(mJy b−1)(104 K)
IC 860
C03.50507.7914.685.13
E03.51607.7550.610.45
S03.50007.5750.780.63
W03.49807.8201.750.78
NW03.49707.9460.750.12
IRAS 15107+0724
C13.09431.8727.929.40
NW13.09132.0300.720.53
W13.08331.8470.470.34
SW13.08831.6160.480.35
S13.09531.5870.560.41
SE13.10331.8230.370.27
Arp 220
W57.09431.87630.4230.2
E57.09132.03013.3813.3

Notes. (a) Positions of IC 860 are relative to RA = 13h15m and Dec. = +24°37΄. Positions of IRAS 15107+0724 are relative to RA = 15h13m and Dec. = +07°13΄. Positions for Arp 220 are relative to RA = 15h34m and Dec. = +23°30΄.

Table 1.

The continuum components.

SourceRAaDec.aContinuumBrightness
componentfluxtemperature
(s)( arcsec)(mJy b−1)(104 K)
IC 860
C03.50507.7914.685.13
E03.51607.7550.610.45
S03.50007.5750.780.63
W03.49807.8201.750.78
NW03.49707.9460.750.12
IRAS 15107+0724
C13.09431.8727.929.40
NW13.09132.0300.720.53
W13.08331.8470.470.34
SW13.08831.6160.480.35
S13.09531.5870.560.41
SE13.10331.8230.370.27
Arp 220
W57.09431.87630.4230.2
E57.09132.03013.3813.3
SourceRAaDec.aContinuumBrightness
componentfluxtemperature
(s)( arcsec)(mJy b−1)(104 K)
IC 860
C03.50507.7914.685.13
E03.51607.7550.610.45
S03.50007.5750.780.63
W03.49807.8201.750.78
NW03.49707.9460.750.12
IRAS 15107+0724
C13.09431.8727.929.40
NW13.09132.0300.720.53
W13.08331.8470.470.34
SW13.08831.6160.480.35
S13.09531.5870.560.41
SE13.10331.8230.370.27
Arp 220
W57.09431.87630.4230.2
E57.09132.03013.3813.3

Notes. (a) Positions of IC 860 are relative to RA = 13h15m and Dec. = +24°37΄. Positions of IRAS 15107+0724 are relative to RA = 15h13m and Dec. = +07°13΄. Positions for Arp 220 are relative to RA = 15h34m and Dec. = +23°30΄.

3.2 IC 860 – formaldehyde emission structure

The channel maps of the line emission for IC 860 have been presented in Fig. 3. The negative (single solid line) contours indicate the general noise level of each of the maps as well as the location of (extended) structures possibly representing weak absorption against the radio continuum of the source. The features with multiple positive contours identify the emission features located within the contours of the radio continuum. These feature have the approximate size of the resolving beam and represent very localized emission regions with single or multiple maser sources that often fill a single spectral channel in the maps. Features above the second positive contour have been added to the Moment 0 map in Fig. 2.

Channel maps of IC 860. The contour levels for the 27 central channels emphasize the negative (single solid line) contours as well as the strongly positive (multiple concentric) contours of the maser emission features. The contour levels are 0.5 mJy beam−1 × (−1, 1, 2, 3) such that the extended single contours are negative.
Figure 3.

Channel maps of IC 860. The contour levels for the 27 central channels emphasize the negative (single solid line) contours as well as the strongly positive (multiple concentric) contours of the maser emission features. The contour levels are 0.5 mJy beam−1 × (−1, 1, 2, 3) such that the extended single contours are negative.

The first moment map of the central emission components in IC 860 does not show a distinct velocity gradient (Fig. 4). Instead the colour contours indicate a dominant emission component with a velocity in the range of 3730–3820 km s−1 superposed on some larger scale background velocity structure with 3500 km s−1 at the east and west edges and 3850 km s−1 in the north and south.

The velocity structure of the central H2CO emission structures of the nuclear region of IC 860. The first moment map shows colour contours between 3500 and 3900 km s−1. The velocity structure may reveal a superposition of point sources but does show an overall velocity gradient.
Figure 4.

The velocity structure of the central H2CO emission structures of the nuclear region of IC 860. The first moment map shows colour contours between 3500 and 3900 km s−1. The velocity structure may reveal a superposition of point sources but does show an overall velocity gradient.

3.3 IC 860 – spectral characteristics

The integrated spectral profile of the H2CO line emission in the nuclear centre region of IC 860 is presented in Fig. 5(a) and shows three distinct emission components. The features centred at 3830 and 3990 km s−1 (ranging from 3720 to 4050 km s−1) are in rough agreement with the single-dish emission components peaking at 3890 km s−1 in Fig. 1, although the high-velocity edge is less prominent in the single-dish spectrum. The broad feature at 3590 km s−1 with a mean flux of 0.95 mJy beam−1 (line width of 230 km s−1) only has a weak counterpart in the single-dish spectrum. The integrated spectrum of the south-east region is displayed in Fig. 5(b). This spectrum shows not only emission peaks at 3700 and 3880 km s−1 but also the smaller feature at 3580 km s−1 that is also found in the centre region.

The integrated spectra of the formaldehyde emission in IC 860 integrated of the centre emission region and the south-east region. Top: the centre region exhibits three distinct velocity components at V = 3590, 3830 and 3990 km s−1. The 3590 km s−1 component has a weak counterpart in the single-dish spectrum of Fig. 1, while the other components find their counterparts in the main feature of the single-dish spectrum and a possible high velocity wing. Bottom: the spectrum at the south-east region shows three distinct components at shifted velocities of V = 3520, 3720 and 3870 km s−1.
Figure 5.

The integrated spectra of the formaldehyde emission in IC 860 integrated of the centre emission region and the south-east region. Top: the centre region exhibits three distinct velocity components at V = 3590, 3830 and 3990 km s−1. The 3590 km s−1 component has a weak counterpart in the single-dish spectrum of Fig. 1, while the other components find their counterparts in the main feature of the single-dish spectrum and a possible high velocity wing. Bottom: the spectrum at the south-east region shows three distinct components at shifted velocities of V = 3520, 3720 and 3870 km s−1.

In addition to the main emission features seen in the single-dish spectrum of Fig. 1, the emission component close to 3580 km s−1 does not have a strong counterpart in the single-dish spectrum and occurs precariously close to the edge of the observing band. Careful inspection of the data reduction procedure shows that this emission is not the result the bandpass calibration or the flat continuum subtraction using the edge channels. (A great effort has been done to make this feature disappear but it would not.) While the strength of this emission component remains unexpected, we suggest that the feature really exists and is compensated in the single-dish spectrum by absorption in the nuclear region in certain channels.

The systemic velocity of IC 860 is nominally at 3347 km s−1(Haynes et al. 1997), but all known OH and HI emission and absorption features are found close to a higher velocity of 3911 km s−1. While the 2Mass image does not show clear signs of a galaxy interaction (Skrutskie et al. 2006), these two velocities may indicate the interactive nature of the IC 860 system consistent with its large FIR luminosity. The observed velocity component at 3580 km s−1 would be consistent with the maser emissions occurring closer to the systemic velocity.

Close to the peak of the continuum emission in Fig. 2, the estimated amplifying optical depths for the spectral features vary as τ = 0.34–0.46 and the brightness temperatures range from 3.0 to 5.2 × 104 K. The spectral components in the weaker south-east region show an optical depth ranging from 1.11 to 1.90 with brightness temperatures ranging from 2.2 to 4.2 × 104 K (Table 2). The regions with the highest flux densities and the lowest gains are located towards the peak of the continuum distribution, which is consistent with a stronger background requiring a lower amplifying gain.

Table 2.

The line emission components.

SourceRAaDec.aContinuumVelocityLineLineBrightness
componentfluxbfluxoptical depthctemperature Tb
(s)( arcsec)(mJy b−1)(km s−1)(mJy b−1)(104 K)
IC 860
Centre03.50407.794.043490, 3640, 3830, 3990(3.30), 1.90, 2.80, 1.90(0.62), 0.34, 0.46, 0.343.01 – 5.23
South-east03.51607.670.203520, 3720, 38700.71, 0.79, 1.341.26, 1.37, 1.902.23 – 4.16
IRAS 15107+0724
Centre13.09731.926.613760, 39101.45, 1,180.20, 0.164.22, 3.43
South13.09331.580.503460, 3710, 3900,40500.80, 0.95, 0.66, 0.280.95, 1.06, 0.84, 0.440.56 – 1.92
North-west13.08232.030.223710, 40300.80, 1.001.53, 1.713.23, 4.04
Arp 220 West
W-centre57.19111.6620.352302.200.105.28
W-south57.19111.585.125230, 53601.03, 0.760.18, 0.143.37, 2.48
W-west57.18111.765.195140, 5230, 53401.30, 1.38, 1.750.22, 0.24, 0.2910.06, 13.43
W-north57.19211.804.625090, 52801.30, 1.550.28, 0.293.89, 4.64
Arp 220 East
E-east57.26511.536.1854301.050.163.14
E-west57.25811.457.845160, 5320, 5400, 55402.20, 1.80, 1.75, 1.200.14 – 0.242.88 – 5.28
SourceRAaDec.aContinuumVelocityLineLineBrightness
componentfluxbfluxoptical depthctemperature Tb
(s)( arcsec)(mJy b−1)(km s−1)(mJy b−1)(104 K)
IC 860
Centre03.50407.794.043490, 3640, 3830, 3990(3.30), 1.90, 2.80, 1.90(0.62), 0.34, 0.46, 0.343.01 – 5.23
South-east03.51607.670.203520, 3720, 38700.71, 0.79, 1.341.26, 1.37, 1.902.23 – 4.16
IRAS 15107+0724
Centre13.09731.926.613760, 39101.45, 1,180.20, 0.164.22, 3.43
South13.09331.580.503460, 3710, 3900,40500.80, 0.95, 0.66, 0.280.95, 1.06, 0.84, 0.440.56 – 1.92
North-west13.08232.030.223710, 40300.80, 1.001.53, 1.713.23, 4.04
Arp 220 West
W-centre57.19111.6620.352302.200.105.28
W-south57.19111.585.125230, 53601.03, 0.760.18, 0.143.37, 2.48
W-west57.18111.765.195140, 5230, 53401.30, 1.38, 1.750.22, 0.24, 0.2910.06, 13.43
W-north57.19211.804.625090, 52801.30, 1.550.28, 0.293.89, 4.64
Arp 220 East
E-east57.26511.536.1854301.050.163.14
E-west57.25811.457.845160, 5320, 5400, 55402.20, 1.80, 1.75, 1.200.14 – 0.242.88 – 5.28

Notes. (a) Positions of IC 860 relative to RA = 13h15m and Dec. = 24°37΄. Positions of IRAS 15107+0724 relative to RA = 15h13m and Dec. = 07°13΄. Positions for Arp 220 relative to RA = 15h34m and Dec. = 23°30΄. (b) The continuum flux density represents a mean value across the emission region. (c) The line optical depth assumes exponential amplification of the background radio continuum.

Table 2.

The line emission components.

SourceRAaDec.aContinuumVelocityLineLineBrightness
componentfluxbfluxoptical depthctemperature Tb
(s)( arcsec)(mJy b−1)(km s−1)(mJy b−1)(104 K)
IC 860
Centre03.50407.794.043490, 3640, 3830, 3990(3.30), 1.90, 2.80, 1.90(0.62), 0.34, 0.46, 0.343.01 – 5.23
South-east03.51607.670.203520, 3720, 38700.71, 0.79, 1.341.26, 1.37, 1.902.23 – 4.16
IRAS 15107+0724
Centre13.09731.926.613760, 39101.45, 1,180.20, 0.164.22, 3.43
South13.09331.580.503460, 3710, 3900,40500.80, 0.95, 0.66, 0.280.95, 1.06, 0.84, 0.440.56 – 1.92
North-west13.08232.030.223710, 40300.80, 1.001.53, 1.713.23, 4.04
Arp 220 West
W-centre57.19111.6620.352302.200.105.28
W-south57.19111.585.125230, 53601.03, 0.760.18, 0.143.37, 2.48
W-west57.18111.765.195140, 5230, 53401.30, 1.38, 1.750.22, 0.24, 0.2910.06, 13.43
W-north57.19211.804.625090, 52801.30, 1.550.28, 0.293.89, 4.64
Arp 220 East
E-east57.26511.536.1854301.050.163.14
E-west57.25811.457.845160, 5320, 5400, 55402.20, 1.80, 1.75, 1.200.14 – 0.242.88 – 5.28
SourceRAaDec.aContinuumVelocityLineLineBrightness
componentfluxbfluxoptical depthctemperature Tb
(s)( arcsec)(mJy b−1)(km s−1)(mJy b−1)(104 K)
IC 860
Centre03.50407.794.043490, 3640, 3830, 3990(3.30), 1.90, 2.80, 1.90(0.62), 0.34, 0.46, 0.343.01 – 5.23
South-east03.51607.670.203520, 3720, 38700.71, 0.79, 1.341.26, 1.37, 1.902.23 – 4.16
IRAS 15107+0724
Centre13.09731.926.613760, 39101.45, 1,180.20, 0.164.22, 3.43
South13.09331.580.503460, 3710, 3900,40500.80, 0.95, 0.66, 0.280.95, 1.06, 0.84, 0.440.56 – 1.92
North-west13.08232.030.223710, 40300.80, 1.001.53, 1.713.23, 4.04
Arp 220 West
W-centre57.19111.6620.352302.200.105.28
W-south57.19111.585.125230, 53601.03, 0.760.18, 0.143.37, 2.48
W-west57.18111.765.195140, 5230, 53401.30, 1.38, 1.750.22, 0.24, 0.2910.06, 13.43
W-north57.19211.804.625090, 52801.30, 1.550.28, 0.293.89, 4.64
Arp 220 East
E-east57.26511.536.1854301.050.163.14
E-west57.25811.457.845160, 5320, 5400, 55402.20, 1.80, 1.75, 1.200.14 – 0.242.88 – 5.28

Notes. (a) Positions of IC 860 relative to RA = 13h15m and Dec. = 24°37΄. Positions of IRAS 15107+0724 relative to RA = 15h13m and Dec. = 07°13΄. Positions for Arp 220 relative to RA = 15h34m and Dec. = 23°30΄. (b) The continuum flux density represents a mean value across the emission region. (c) The line optical depth assumes exponential amplification of the background radio continuum.

The distinct components at the centre and south-east locations between 3650 and 4000 km s−1 contribute to emission in the integrated single-dish spectrum (Fig. 1). A superposition of the emission (and possible absorption) across the face of the nuclear continuum could account for the differences between the single-dish spectrum and the current MERLIN spectra. However, the reality of the weak emission feature at 3580 km s−1 in the single-dish spectrum and the feature found in the current data does require further verification with more sensitive observations having a larger observing bandwidth.

4 FORMALDEHYDE IN IRAS 15107+0724

4.1 IRAS 15107 – formaldehyde emission

IRAS 15107+0724 has been first detected at Arecibo Observatory (Baan et al. 1993) with a peak flux of 1.91 mJy. A representative spectrum, presented in Fig. 6 (Araya et al. 2004), shows a spectral line with a mean velocity of 3880 km s−1 and a total width of 320 km s−1, but clearly consisting of two components. No other tentative features can be seen in the spectrum. An OH spectrum also shows two emission components centred at 3782 and 3900 km s−1, whereas the HI spectrum shows a broad absorption feature centred at 3902 km s−1 (Baan et al. 1987). The galaxy itself has a systemic velocity of 3897 km s−1.

A single-dish spectrum of IRAS 15107+0724 obtained with the Arecibo radio telescope. Taken from Araya et al. (2004).
Figure 6.

A single-dish spectrum of IRAS 15107+0724 obtained with the Arecibo radio telescope. Taken from Araya et al. (2004).

The formaldehyde line emission in IRAS 15107+0724 is presented as a zeroth-moment colour overlay on the continuum contours in Fig. 7. The line channels in the data cube have an rms noise of 0.17 mJy beam−1 and the moment maps incorporate signals above a flux threshold of 0.4 mJy beam−1. The main emission feature straddles the central continuum peak and has an extent of 52 pc at PA = 140°. This orientation agrees well with the orientation of the 2Mass image of the galaxy at PA = 155° (Skrutskie et al. 2006). Two other compact emission regions, south and north-west, are superposed on the extended radio continuum structure.

The formaldehyde line emission and radio continuum structure of the IR 15107+0724. A continuum source shows a central component with a peak flux of 7.92 mJy beam−1 surrounded by distributed emission regions resembling an ring or inner spiral arms. The contour levels are 0.11 mJy beam−1 × (−1, 1, 2, 3, 5, 8, 24, 48, 72). The rms in the continuum map is 0.051 mJy beam−1. The zeroth moment of the formaldehyde emission is presented as a superposed colour scale map. The peak integrated flux is 308.8 mJy km s−1 and the colour scale runs from 60 to 320 mJy km s−1. Three distinct emission regions are found within the continuum contours at the centre position and at the south and north-west positions.
Figure 7.

The formaldehyde line emission and radio continuum structure of the IR 15107+0724. A continuum source shows a central component with a peak flux of 7.92 mJy beam−1 surrounded by distributed emission regions resembling an ring or inner spiral arms. The contour levels are 0.11 mJy beam−1 × (−1, 1, 2, 3, 5, 8, 24, 48, 72). The rms in the continuum map is 0.051 mJy beam−1. The zeroth moment of the formaldehyde emission is presented as a superposed colour scale map. The peak integrated flux is 308.8 mJy km s−1 and the colour scale runs from 60 to 320 mJy km s−1. Three distinct emission regions are found within the continuum contours at the centre position and at the south and north-west positions.

The underlying continuum structure of IRAS 15107+0724 shows a central nuclear component surrounded by a ring of emission tracing two inner spiral arms and resembling an S-shaped structure (Fig. 7). The total flux density of the source is 29.72 mJy with a peak of 12.57 mJy beam−1. The brightness temperature at the nuclear source is 9.4 × 104 K, while the peaks in the surrounding structure have brightness temperatures in the range of (0.27–0.53) × 104 K. These temperatures and the emission structure support the hypothesis that this entire S-shaped continuum structure traces star formation activity, although the presence of an embedded radio AGN at the nucleus cannot be excluded. In order to better define the outer continuum structure, the contour map presented in Fig. 7 also includes the central line emission. The positions, the continuum flux densities and the brightness temperatures of these identifiable components are presented in Table 1.

4.2 IRAS 15107 – formaldehyde emission structure

The channel maps of the line emission for IR15107+0724 have been presented in Fig. 8. The negative (single solid line) contours indicate the general noise level of each of the maps as well as the location of possible absorption features against the radio continuum of the source. The features with multiple positive contours identify the localized emission features in the maps lying mostly within the contours of the radio continuum. The features above the first positive contour have been presented in the Moment 0 map in Fig. 7. The emission structure of formaldehyde in IR 15107+0724 is also made up of multiple (single) features that often fill only one spectral channel and are found across the centre of the velocity range. In addition, the presence of extended negative contours may be consistent with the concept that certain emission features are compensated by regions of absorption.

Channel maps of IR 15107+0724. The contour levels for the 27 central channels emphasize the negative (single solid line) contours as well as the strongly positive (multiple concentric) contours of the maser emission features. The contour levels are 0.2 mJy beam−1 × (−2, 4, 6, 8).
Figure 8.

Channel maps of IR 15107+0724. The contour levels for the 27 central channels emphasize the negative (single solid line) contours as well as the strongly positive (multiple concentric) contours of the maser emission features. The contour levels are 0.2 mJy beam−1 × (−2, 4, 6, 8).

4.3 IRAS 15107 – spectral characteristics

The velocity field in the first moment map shows a semi-organized northeast–southwest velocity gradient (PA = 145°) across the nuclear emission region (Fig. 9) with a central dominant component at 3850 km s−1 superposed close to the centre.

The velocity field of the formaldehyde emission region at the centre of IRAS 15107+0724. The velocity colours in the first moment map range between 3750 and 4050 km s−1. No clear organized motion can be seen across the centre emission region except for a (possible) velocity gradient covering 250 km s−1 from north-east to south-west. A dominant central component exists close to the systemic velocity of 3850 km s−1 located close to the central continuum peak of the source. The emission and the associated velocity field seem dominated by a superposition of strong emission components.
Figure 9.

The velocity field of the formaldehyde emission region at the centre of IRAS 15107+0724. The velocity colours in the first moment map range between 3750 and 4050 km s−1. No clear organized motion can be seen across the centre emission region except for a (possible) velocity gradient covering 250 km s−1 from north-east to south-west. A dominant central component exists close to the systemic velocity of 3850 km s−1 located close to the central continuum peak of the source. The emission and the associated velocity field seem dominated by a superposition of strong emission components.

The spectral signature of the central emission component in the zeroth-moment map in Fig. 7 has been presented in Fig.  10. The central component contributes most to the integrated emission close to the systemic velocity with a peak at 3750 km s−1, which is below the observed 3880 km s−1 velocity of the emission peak and the 3897 km s−1 systemic velocity. The spectra for the two weaker emission components, south and north-west, are presented in Fig. 11. Both the south and north-west emission regions show additional velocity components below (3710 km s−1) and above (4015 km s−1) the systemic velocity of 3820 km s−1. These coarse velocity designations are quite consistent with the velocities found at the centre region, although the multiple velocity systems are reminiscent of a merger scenario.

Integrated formaldehyde emission spectrum at the centre position of IRAS 15107+0724.
Figure 10.

Integrated formaldehyde emission spectrum at the centre position of IRAS 15107+0724.

Integrated formaldehyde spectra at the two additional emission components of IRAS 15107+0724 at the southern component (top), and the north-west component (bottom). The spectra suggest that there is also absorption in the source and that individual emission components exist within the observed range of 3650–4050 km s−1 in the single-dish spectra (see Fig. 6).
Figure 11.

Integrated formaldehyde spectra at the two additional emission components of IRAS 15107+0724 at the southern component (top), and the north-west component (bottom). The spectra suggest that there is also absorption in the source and that individual emission components exist within the observed range of 3650–4050 km s−1 in the single-dish spectra (see Fig. 6).

The characteristics of all emission regions within the continuum confines of IRAS 15107+0724 have been presented in Table 2. The peak line flux density at the nucleus is 1.45 mJy beam−1 and the integrated fluxes at the three other regions are in the range 0.45–1.0 mJy beam−1. The central line emission regions in IRAS 15107+0724 superposed on the continuum peak have a brightness temperature in the range of 3.4–4.2 × 104 K and apparent optical depths of only 0.16–0.20 (see Table 2). The line emission features in the two other features have brightness temperatures higher than those of the radio continuum ranging from 0.6 to 4.0 × 104 K and they have higher optical depths ranging from 0.44 to 1.71. Although some larger optical depths are found, both the brightness temperatures and the deduced optical depths are consistent with modelling results with radiative pumping schemes (details in Section 6.3).

The spectral features emission found in IRAS 15107+7024 (Figs 10 and 11) cover the velocity range of the observed single-dish spectrum extending from 3650 to 4050 km s−1 with some of the weaker features covering the shoulders. The dropouts at higher velocities at the southern location and at lower velocities at the north-west location appear unusual, but they are found to be a local characteristic. They are not the result of continuum subtraction. These differences are consistent with the presence of weak absorption structures against the continuum structure that do not stand out clearly in the single-dish spectrum.

5 FORMALDEHYDE IN ARP 220 (IC 4553)

5.1 Arp 220 – formaldehyde emission

The extragalactic formaldehyde emission from Arp 220 was first detected using the Effelsberg telescope in 1984 together with the absorption lines in NGC 3628 and NGC 3079 (Baan et al. 1986). A representative Arecibo spectrum of the line has been presented in Fig. 12 (from Baan et al. 1993). With a single-dish line strength of only ∼4.0 mJy and an isotropic luminosity |$L_{\rm H_{2}CO} = 61\, {\rm L}_{\odot }$|⁠, it is the strongest and most luminous among the known H2CO MMs (Araya et al. 2004). The formaldehyde emission is found to extend across the central molecular zones of each of the nuclei, while the peak of the formaldehyde spectrum at 5340 km s−1 is dominated by emission at the systemic velocity of the western nucleus (Baan & Haschick 1995). Based on the OH MM emission at 1667 MHz and a number of high-density tracer molecular emissions (Baan 2007), the systemic velocities of the two nuclei are 5683 km s−1 for the eastern nucleus and 5365 km s−1 for the western nucleus. As a result the emissions at the systemic velocity of the eastern nucleus do not fall within the frequency window of these observations.

A single-dish spectrum of Arp 220 obtained with the Arecibo radio telescope. The two arrows indicate the systemic velocities of the western (left) and eastern nuclei. The horizontal bars below the spectrum indicate the observed velocity range of the OH 1612 MHz absorption, the OH 1667/1665 MHz MM emissions and the thermal CO(1-0) and HCN(1-0) emissions in the nuclear region. The dashed line indicates the range of the observed blueshifted OH outflow emission.
Figure 12.

A single-dish spectrum of Arp 220 obtained with the Arecibo radio telescope. The two arrows indicate the systemic velocities of the western (left) and eastern nuclei. The horizontal bars below the spectrum indicate the observed velocity range of the OH 1612 MHz absorption, the OH 1667/1665 MHz MM emissions and the thermal CO(1-0) and HCN(1-0) emissions in the nuclear region. The dashed line indicates the range of the observed blueshifted OH outflow emission.

The formaldehyde emission structures at the nuclei of Arp 220 are presented in Fig. 13 as a zeroth-moment formaldehyde emission colour map superposed on the continuum contours. These emission regions exhibit both discrete and extended components of different strengths superposed on the extended continuum structure of the west and the east nuclear regions. The dominating central emission structure at the west nucleus is elongated by 110 pc in a north–south direction and is similar to the elongated structures found at the nuclei of IC 860 and IR 15107+0724. However, the western nucleus shows a more complex structure than just a north–south nuclear disc structure. Possibly the east–west structure with a distinct (east–west) velocity gradient represents a superposed structural component resulting from the ongoing merger. Alternatively, these are star formation regions away from the disc. The emission region at the eastern nucleus covers approximately 53 pc and consists of two regions in northeast–southeast direction. It should be noted that this emission occurs at the systemic velocity of the western nucleus, while emission at the systemic velocity of the eastern nucleus falls at the upper edge (or outside) of the observing band.

A composite map of the continuum emission and the formaldehyde line emission in Arp 220. The continuum structure is depicted with contours at levels 0.30 × (−1, 1, 2, 4, 8, 16, 32, 64, 80) mJy beam−1 and a peak flux of 30.42 and 13.38 mJy beam−1 for the western and eastern nucleus, respectively. The colour scale zeroth-moment image depicts a range of 0.1–0.56 mJy beam−1 km s−1. The line emission image shows four emission regions at the western nucleus: W-centre, W-north, W-west and W-south. The eastern nucleus displays two line emission regions: E-west and E-east.
Figure 13.

A composite map of the continuum emission and the formaldehyde line emission in Arp 220. The continuum structure is depicted with contours at levels 0.30 × (−1, 1, 2, 4, 8, 16, 32, 64, 80) mJy beam−1 and a peak flux of 30.42 and 13.38 mJy beam−1 for the western and eastern nucleus, respectively. The colour scale zeroth-moment image depicts a range of 0.1–0.56 mJy beam−1 km s−1. The line emission image shows four emission regions at the western nucleus: W-centre, W-north, W-west and W-south. The eastern nucleus displays two line emission regions: E-west and E-east.

The underlying continuum structure of the merger system Arp 220 in Fig. 13 shows two isolated nuclei with peak flux densities of the east and west nuclei of 13.38 and 30.42 mJy beam−1). The peak brightness temperatures at 4.83 GHz for the eastern and western nuclei are 1.3 and 3.0 × 105 K, respectively (Table 1), which is consistent with intense star formation. The (loop-like) structure that extends southward from the western nucleus to the eastern nucleus has been resolved out at this higher resolution (Baan & Haschick 1995; Rovilos et al. 2003). The extended structures at various position angles would be evidence that the star formation extends into pre-existing spiral arms or larger scale structures within the merging nuclei. No evidence has been found yet for the presence of an AGN in either one of the nuclei (Genzel et al. 1998; Smith et al. 1998).

5.2 Arp 220 – formaldehyde emission structure

The channel maps of the line emission for Arp 220 have been presented in Fig. 14. The negative (single solid line) contours indicate the general noise level of each of the maps as well as the location of possible absorption features against the radio continuum of the source. The features with multiple positive contours identify the localized emission features in the maps lying mostly within the contours of the radio continuum of the two nuclei. The features from the dominant emission regions in both nuclei above the first positive contour have been presented in the Moment 0 map in Fig. 13. The emission structure of formaldehyde in Arp 220 is also made up of multiple (single) features that often fill only one spectral channel. They are also found across a large fraction of the velocity range, which does not fully cover the range of 5000–5850 km s−1 of the observed emission from the two nuclei (see Fig. 12).

Channel maps of Arp 220. The contour levels for the 27 central channels indicate the negative (single solid line) contours as well as the strongly positive (multiple concentric) contours of the maser emission features. The contour levels are 0.3 mJy beam−1 × (−2, 4, 6, 8).
Figure 14.

Channel maps of Arp 220. The contour levels for the 27 central channels indicate the negative (single solid line) contours as well as the strongly positive (multiple concentric) contours of the maser emission features. The contour levels are 0.3 mJy beam−1 × (−2, 4, 6, 8).

5.3 Arp 220 – spectral characteristics

The velocity field of the emissions at the two nuclei of Arp 220 has been presented in the first moment maps of Fig. 15. Because of the 300 km s−1 velocity difference between the two nuclei, the observed emission mostly covers the velocity range of the west nucleus. The weaker emission at the systemic velocity of the eastern nucleus enters the band only above 5600 km s−1 . Both emission regions exhibit very complex velocity fields that suggest that the observed emission represents a superposition of compact emission regions at different radial velocities. The emission regions at the west nucleus may show a possible north–south gradient at PA = 6° and centred at Vwest = 5365 km s−1 across the W-north, W-centre and W-south regions. On the other hand, the W-west region shows higher velocities and the weaker W-east region shows lower velocities. The emission at the east nucleus has a velocity field also centred at the Vwest velocity with a (possible) southwest–northeast gradient at PA = 315° across the regions. While the orientation of these velocity gradients are in rough agreement with those of the OH MM (1667 MHz) data, the velocity gradient at the west nucleus appears opposite to the local HI and OH gradient (Rovilos et al. 2003; Baan 2007). The whole nuclear region shows a north–south gradient based on CO and HCN molecular data (Downes & Solomon 1998; Zhao & An 2008; Sakamoto et al. 2008). Considering that the formaldehyde material seen at the east nucleus mostly belongs to the west galaxy and that the observed emission has a very complex structure, no detailed information can be derived from the velocity structure in the moment maps.

The velocity fields at the western nucleus (top) and eastern nucleus (bottom) of Arp 220. The emission regions in the western nucleus possibly display a weak north-south velocity gradient and also the regions in the eastern nucleus indicate a southwest–northeast gradient.
Figure 15.

The velocity fields at the western nucleus (top) and eastern nucleus (bottom) of Arp 220. The emission regions in the western nucleus possibly display a weak north-south velocity gradient and also the regions in the eastern nucleus indicate a southwest–northeast gradient.

The line emission spectra of the four prominent components in the west nucleus (Figs 16 and 17) and of two prominent components in the east nucleus (Fig. 18) show a complementary picture. The central emission regions W-centre and W-south at the west nucleus at 5230 km s−1 provide the dominant peak in the single-dish spectrum while the emission from the western regions W-west contributes to the higher velocity shoulder of the profile. Region W-north in the west nucleus also contributes to the lower velocity shoulder. The east nucleus in E-west shows multiple peaks covering a larger velocity range that contributes to both the low-velocity wing down to 5100 km s−1 and the extended high-velocity wing up to 5540 km s−1. This higher velocity component at the east nucleus and also the 5600 km s−1 structure in the W-south location are consistent with emission close to the systemic velocity of the east nucleus as has also been seen in the OH 1667 MHz MM data (Baan & Haschick 1984; Rovilos et al. 2003). The spectra at both nuclei as presented in Figs 1618 again show the presence of localized (narrow and broad) absorption-like features.

Integrated spectra resulting from the central emission regions in the western nucleus of Arp 220. The spectra have been taken at northern part W-centre (top) and at the southern part W-south (bottom). Together these emission components account for much the single-dish emission profile as presented in Fig. 12. The structure at 5650 km s−1 .
Figure 16.

Integrated spectra resulting from the central emission regions in the western nucleus of Arp 220. The spectra have been taken at northern part W-centre (top) and at the southern part W-south (bottom). Together these emission components account for much the single-dish emission profile as presented in Fig. 12. The structure at 5650 km s−1 .

Integrated spectra resulting from the other emission regions in the western nucleus of Arp 220. The spectra are those taken across the western nuclear component W-west (top) and the northern nuclear component W-north (bottom).
Figure 17.

Integrated spectra resulting from the other emission regions in the western nucleus of Arp 220. The spectra are those taken across the western nuclear component W-west (top) and the northern nuclear component W-north (bottom).

Integrated spectra resulting from emission regions at the eastern nucleus of Arp 220. The spectra are those from the E-east region (top) and the E-west region (bottom).
Figure 18.

Integrated spectra resulting from emission regions at the eastern nucleus of Arp 220. The spectra are those from the E-east region (top) and the E-west region (bottom).

The amplifying optical depth and the brightness temperature of the emission line features in Arp 220 are presented in Table 2. The optical depth of the features at both nuclei varies between 0.10 and 0.29, with smaller values at locations with higher continuum fluxes, consistent with amplification of the continuum. The brightness temperatures of the features vary between 2.5 and 13.4 × 104 K. The high brightness temperatures of all emission components confirm masering nature of these emissions.

6 INTERPRETING THE EMISSION

6.1 Extragalactic emission structures

The formaldehyde Ka = 101 − 111 data at 4.829 GHz in IC 860, IRAS 15107+0724 and Arp 220 confirms that the emission originates in regions centred on the nuclei of the galaxies and that the emission results from masering amplification of the radio continuum by excited foreground molecular gas. The dominant H2CO emission arises from a central (elongated) molecular structure of size 30–100 pc in a line of sight close to the peak of the continuum source, and also from some isolated star formation regions, whose projected positions are also within a region defined by the radio contours. The central emission regions are consistent with a clumped masering medium in the central section of an edge-on molecular disc causing a hierarchy of emission structures from compact high-brightness to more extended lower brightness components. Any emission from the outer regions of the disc is undetected because of the decreasing continuum flux. The central regions do not show a prominent velocity gradient because of the superposition of dominant point sources at different velocities superposed on a more extended structure. An elongated emission structure is also found in Sgr B2 as a string of six single maser features covering a linear distance of nearly 5 pc (Mehringer, Goss & Palmer 1994; Qin et al. 2008), although there is no evidence yet of a more extended and diffuse component.

The formaldehyde emission at the west nucleus of Arp 220 coincides quite well with the presence of a compact circumnuclear disc as deduced from the OH MM and the CO and H i emission. The emission from its east nucleus may not be representative of the observed disc rotation, but it is still well aligned with the larger scale rotation structure (Sakamoto et al. 1999; Mundell, Ferruit & Pedlar 2001; Baan 2007). In the cases of IC 860 and IR 15107+0724, no detailed molecular studies have yet been done. However, the observed formaldehyde emission structure is well aligned with the optical image of both galaxies, and in the case of IC 860, it also lies perpendicular to the extended relic continuum structure.

The weaker emission components found in IC 860 and IRAS 15107+0724 represent discrete star formation regions in the circumnuclear environment. The emission also results from amplification of the lower background continuum, but they require higher amplifying optical depths than those in the nuclear region (Table 2). The estimated brightness temperatures of the emissions in the circumnuclear regions are slightly lower than those in the central regions, which gives a range for the observed temperatures between 6 × 103 K and 1.3 × 105 K. Such brightness temperatures confirm that the observed line emission results from masering activity and that these sources are indeed MMs.

Except for the three galaxies considered here, single-dish spectra of nearby galaxies show absorption for both the 4.83 GHz Ka= 1 and the 14.5 GHz Ka= 2 transitions (Baan et al. 1993; Araya et al. 2004; Mangum et al. 2013). This would indicate that the ground-state inversions are weak and possibly rare and that there is competition between emission and absorption in the 4.83 GHz line of most galaxies. As a result the integrated profiles may show evidence of absorption from extended regions, of localized emission and possibly of localized re-absorption depending on the physical conditions of the intervening molecular material. The spectral results presented in this paper show some evidence of both narrow and broad absorption components, which could account for the differences when compared with the single-dish spectra (e.g. the feature at 3580 km s−1 in the spectrum of IC 860). This spectral superposition is further supported by some recent single-dish 4.83 GHz H2CO spectra of extragalactic sources that show clear evidence of emission components superposed on dominant absorption (see Mangum et al. 2013).

The occurrence of formaldehyde maser action remains rare in the Galaxy and in extragalactic sources. Although there are many luminous infrared sources in the Galaxy and in extragalactic sources, only a small fraction of these are associated with H2CO and also OH maser activity. Therefore, masering activity only happens during certain evolutionary stages when both the right pumping agent and the right molecular environment exist along a certain line of sight. As a result, the special conditions under which H2CO masering action happens hold a key for understanding the sequence of events during star formation and starburst processes.

6.2 The H2CO pumping environment

The establishment of maser action in four extragalactic sources again raises the issue of the responsible pumping mechanism for formaldehyde in Galactic and extragalactic sources. Early suggestions for pumping the formaldehyde molecules included radiative pumping by FIR radiation fields (e.g. Litvak 1969) and a pumping model using the free–free radio continuum (e.g. Boland & de Jong 1981). Free–free pumping can produce small-level inversions, but it does not explain the maser components discussed in this paper (Baan et al. 1986), and the formaldehyde masers in Sgr B2 (Mehringer et al. 1994).

Collisional pumping models have been suggested based on H2 and electron collisions (Araya, Hofner & Goss 2007) and shocks (Mehringer, Goss & Palmer 1994; Araya, Baan & Hofner 2004; Hoffman, Goss & Palmer 2007). Collisional excitation with H2 and radiative excitation by the free–free radio continuum radiation from a nearby ultra- or hyper-compact H ii region can indeed invert the H2CO Ka = 101 − 111 transition (van der Walt 2014). However, collisional pumping in shocks does not explain the simultaneous flaring of the H2CO and 6.7 GHz Class II CH3OH maser components of the Galactic source IRAS 18566+0408 (Araya et al. 2010).

The FIR radiation fields in the host galaxies of formaldehyde MM remain as the most viable pumping agent for extragalactic formaldehyde emitters and for Galactic sources. The FIR radiation fields are already found to be responsible for pumping the OH in Galactic sources and OH MM sources and they can also account for pumping many Class II maser transitions for Galactic methanol (Cragg, Sobolev & Godfrey 2005) and naturally explains the flaring in IRAS 18566+0408. A early link between formaldehyde MM activity and FIR luminosity and spectral colour temperature has been established based on these galaxies (Baan et al. 1993; Araya et al. 2004), in analogy with similar relations for OH MM sources.

6.3 FIR radiative excitation of H2CO

A radiative FIR pumping scenario for the formaldehyde 4.829 GHz Ka 110 − 111 transition requires a certain temperature range for the prevailing FIR radiation field in order to achieve population inversions. The estimated FIR radiation blackbody temperatures are in the range of 40–65 K for the three galaxies (see Gao & Solomon 2004), which is also the range that results in pumping of the OH molecules. However, inversion of the formaldehyde Ka = 1 level cannot be achieved when the kinetic temperature (Tk) of the local environment is larger or equal to the blackbody temperature (Td) of the dust FIR radiation (van der Walt 2014). Therefore, extragalactic environments with coupled dust and kinetic temperatures will only show absorption in both transitions combined with possible thermal emission in the Ka = 1 transition but will not show any maser activity (as seen by Mangum et al. 2013).

Simulations with the Radex facility (van der Tak et al. 2007) indicate that for local kinetic temperatures lower than the radiative temperature, inversions in both the Ka = 1 and Ka = 2 transitions can be achieved by FIR radiative pumping. This describes molecular environments with densities up to 105 cm−3, where the kinetic and dust temperatures are yet not coupled. The results of the simulations depicted in Fig. 19 show that the attainable (negative) optical depth varies strongly with density and local kinetic temperature Tk and column density N(H2). Assuming a representative dust temperature of Td = 50 K giving the blackbody FIR radiation field, the optimal densities are in the range of n = 104–105 cm−3. Weak or no inversions are found at both higher densities (n = 106 cm−3), where temperature coupling would result, and also at lower (n = 103 cm−3) densities. No inversion occurs when Tk becomes equal or higher than Td. A maximum in the optical depth of τ4.8 = −2.2 is found for a density of n = 104 cm−3 and temperature difference (Td − Tk) = 40 K; the gains decrease with a decreasing temperature (Td − Tk) difference.

Inversion of the formaldehyde molecules by FIR radiative pumping. A representative dust temperature Td = 50 K has been assumed to represent the FIR radiation field. The optical depth curves for the 4.83 GHz Ka = 1 transition are in solid lines and the curves for the 14.5 GHz Ka = 2 transition are in dashed lines. The two groups of curves are for two gas densities (blue for 105 cm−3 and red for 104 cm−3) for three kinetic temperatures. The gains increase for lower gas densities and decrease for higher values of the kinetic temperature Tk.
Figure 19.

Inversion of the formaldehyde molecules by FIR radiative pumping. A representative dust temperature Td = 50 K has been assumed to represent the FIR radiation field. The optical depth curves for the 4.83 GHz Ka = 1 transition are in solid lines and the curves for the 14.5 GHz Ka = 2 transition are in dashed lines. The two groups of curves are for two gas densities (blue for 105 cm−3 and red for 104 cm−3) for three kinetic temperatures. The gains increase for lower gas densities and decrease for higher values of the kinetic temperature Tk.

Similar optical depth curves are found for the optical depth in the 14.5 GHz Ka = 211 − 212 transition except that the peak values typically scale as τ14.5 ≈ 0.45 τ4.8 (Fig. 19). However, no discernible maser emission has yet been detected in the 14.5 GHz transition in single-dish spectra of extragalactic sources (Mangum et al. 2013). This may be explained by the much lower gains found for the Ka = 2 transition and the disappearance of the inversion for smaller values of the (Tbb − Tk) temperature difference. In the parameter range where both transitions could amplify the background radio continuum, a radio spectral index of α = −1 and a (peak) assumed gain τ4.8 = −1.3 will generate an expected emission line flux in the 14.5 GHz line that is only 9 per cent of the 4.8 GHz emission line strength. Such 14.5 GHz line emission features may indeed go unidentified in the presence of a more dominant absorption component.

The evaluation of the FIR pumping of the formaldehyde in a non-coupled environment show that there is sufficient parameter space to provide the maser action observed in the four extragalactic sources. The optical depths deduced for the three sources (Table 2) lie within the range of attainable optical depths found with these pumping simulations. In addition, the non-coupled (dust-kinetic) temperature environment required for the pumping represents a plausible condition in the star-forming environments. The requirement of a non-coupled temperature environment may also define the transient environment that provides the window of opportunity for maser action and may explain the small numbers of observed formaldehyde masers.

7 CONCLUSION

The emissions in the formaldehyde ground-state transition in the three galaxies, IC 860, IRAS 15107+0724 and Arp 220, are extended and show a range of structural scale sizes up to 100 pc. The brightness temperatures of the emission features (Tb = 6 × 103–1.3 × 105 K) falling within the contours of the nuclear radio continuum indicate that they are maser features resulting from amplification of the continuum background. A clumped foreground medium pumped by the FIR radiation fields will produce a hierarchy in the emission structure, giving a combination of extended lower brightness and compact higher brightness maser components at a range of radial velocities.

The observed central emission regions are close to the peak of the nuclear continuum and are consistent with them being the front section of a larger scale disc structure in the galaxies. The central emission regions do not exhibit clear velocity gradients resulting from disc rotation, because dominant high-brightness components that contribute to the emission and the complex structure of the molecular ISM following a nuclear merger. The outer parts of the disc have not yet been detected because of the outward decrease of the continuum and possibly changing pumping conditions. In the case of Arp 220, it appears that the OH emission from the disc in the west nucleus actually straddles the observed formaldehyde emission (see Rovilos et al. 2003). In addition to the central emission regions, a number of less prominent and isolated regions have been found in all three sources. Most likely these are star formation regions with a higher amplifying gain superposed on weaker continuum components.

The pumping of the formaldehyde emission has been associated with the FIR radiation fields in the (U)LIRGs, which is analogous to the pumping scenarios for OH MMs. The FIR radiation field can invert the formaldehyde molecular population for a density range of n = 104–105 cm−3 when the local kinetic temperature Tk is lower than the blackbody temperature Td of the radiation field. The observed optical depths of the amplifying medium fall within the range predicted by the simulations of the pumping environment. In general, the observed optical depths are found to be relatively low, which would be consistent with the concept of low-gain amplification by an extended (more diffuse) and clumpy interstellar medium.

The current data for these extragalactic sources show some evidence for localized absorption in addition to the emission regions. The presence of absorption, localized emissions and possible re-absorption in the ISM of the nuclear region may account for some of the spectral differences between the single-dish data and the interferometric results. The detection of localized emission will thus depend strongly on the beam size of the instrument and the other (unconfirmed) candidate formaldehyde emitters should also be studied at higher resolution.

The more extended emission found in formaldehyde in these galaxies provide a new view of the medium in these sources. In the past, maser studies have mostly concentrated on only the highest brightness components. However, it is becoming increasingly clear that a large fraction of the Galactic and also extragalactic OH, H2CO and likely H2O maser emission is extended and will be at least partially resolved in the highest resolution experiments. These extended maser regions may thus provide tools to study the structure and nature of the ISM in these sources in combination with other diagnostic tools.

The requirement for pumping of the formaldehyde by FIR radiation fields implies that cold material must still be present in the molecular ISM of the starbursting nucleus. Considering that the ongoing star formation process continues to heat any regions in the ISM with a non-coupled dust and kinetic temperature regime, such regions will exist during a relatively small evolutionary time window during the relatively early stages of evolution of Galactic star formation and extragalactic starbursts. This may indicate that formaldehyde MM activity occurs at an earlier evolutionary stage than OH MM activity and this may also explain why only a few H2CO MM have been found among the more extended OH MM sample.

For Galactic maser environments, an empirical time sequence has been suggested for maser activity with different flavours (Ellingsen et al. 2007; Breen et al. 2010), which did not (yet) include formaldehyde masers. A similar sequence may now also exist for the maser activity during the evolution of starbursting nuclei including OH, H2CO and also Class II CH3OH masers.

Acknowledgments

This paper is written in memory of our friend James R. Cohen (1948–2006) with whom this project was started. WAB has been supported as a Visiting Professor of the Chinese Academy of Sciences (KJZD-EW-T01). WAB thanks the Shanghai Astronomical Observatory staff for their hospitality during the tenure. TA has been supported by the NSTC 973 Program (2013CB837900) and Shanghai Rising Star programme. The e-MERLIN is a National Facility operated by the University of Manchester at Jodrell Bank Observatory on behalf of the UK Science and Technology Facilities Council (STFC). This research has made use of the NASA/IPAC Extragalactic Database (NED) that is operated by the Jet Propulsion Laboratory, California Institute of Technology, under contract with the National Aeronautics and Space Administration.

REFERENCES

Araya
E.
,
Baan
W. A.
,
Hofner
P.
,
2004
,
ApJS
,
154
,
541

Araya
E.
,
Hofner
P.
,
Goss
W. M.
,
2007
,
Proc. IAU Symp.
,
242
,
110

Araya
E. D.
et al. ,
2008
,
ApJS
,
178
,
330

Araya
E. D.
et al. ,
2010
,
ApJ
,
717
,
L133

Baan
W. A.
,
1985
,
Nature
,
315
,
26

Baan
W. A.
,
1989
,
ApJ
,
338
,
804

Baan
W. A.
,
2007
,
Proc. IAU Symp.
,
242
,
437

Baan
W. A.
,
Haschick
A. D.
,
1984
,
ApJ
,
279
,
541

Baan
W. A.
,
Haschick
A. D.
,
1995
,
ApJ
,
454
,
745

Baan
W. A.
,
Wood
P. A. D.
,
Haschick
A. D.
,
1982
,
ApJ
,
260
,
L49

Baan
W. A.
,
Güsten
R.
,
Haschick
A. D.
,
1986
,
ApJ
,
305
,
830

Baan
W. A.
,
van Gorkom
J. H.
,
Schmelz
J. T.
,
Mirabel
I. F.
,
1987
,
ApJ
,
313
,
102

Baan
W. A.
,
Haschick
A. D.
,
Uglesich
R.
,
1993
,
ApJ
,
415
,
140

Boland
W.
,
de Jong
T.
,
1981
,
A&A
,
98
,
149

Braatz
J. A.
et al. ,
2010
,
ApJ
,
718
,
657

Breen
S. L.
,
Ellingsen
S. P.
,
Caswell
J. L.
,
Lewis
B. E.
,
2010
,
MNRAS
,
401
,
2219

Chen
X.
et al. ,
2015
,
ApJ
,
800
,
L2

Claussen
M. J.
,
Heiligman
G. M.
,
Lo
K. Y.
,
1984
,
Nature
,
310
,
298

Condon
J. J.
,
1992
,
ARA&A
,
30
,
575

Cragg
D. M.
,
Sobolev
A. M.
,
Godfrey
P. D.
,
2005
,
MNRAS
,
360
,
533

Darling
J.
,
Giovanelli
R.
,
2002
,
AJ
,
124
,
100

Downes
D.
,
Solomon
P. M.
,
1998
,
ApJ
,
507
,
615

Ellingsen
S. P.
et al. ,
2007
,
Proc. IAU Symp.
,
242
,
213

Ellingsen
S. P.
et al. ,
2014
,
ApJ
,
790
,
L28

Gao
Y.
,
Solomon
P. M.
,
2004
,
ApJ
,
606
,
271

Genzel
R.
et al. ,
1998
,
ApJ
,
498
,
579

Ginsburg
A.
et al. ,
2015
,
A&A
,
584
,
L7

Haschick
A. D.
,
Baan
W. A.
,
1985
,
Nature
,
314
,
144

Haschick
A. D.
et al. ,
1990
,
ApJ
,
356
,
149

Haynes
M. P.
et al. ,
1997
,
AJ
,
113
,
1197

Henkel
C.
,
Wilson
T. L.
,
1990
,
A&A
,
229
,
431

Herrnstein
J. R.
et al. ,
1997
,
ApJ
,
475
,
L17

Hoffman
I. M.
,
Goss
W. M.
,
Palmer
P.
,
2007
,
ApJ
,
654
,
971

Kellermann
K. I.
Owen
F. N.
,
1988
,
Verschuur
G. L.
Kellermann
K. I.
,
Galactic and Extragalactic Astronomy
.
Springer-Verlag
,
Berlin
, .
577

Klöckner
H.-R.
,
Baan
W. A.
,
2004
,
A&A
,
419
,
887

Kondratko
P. T.
,
Greenhill
L. J.
,
Moran
J. M.
,
2005
,
ApJ
,
618
,
618

Kuo
C. Y.
et al. ,
2015
,
ApJ
,
800
,
26

Litvak
M. M.
,
1969
,
ApJ
,
156
,
471

Mangum
J. G.
,
Darling
J.
,
Menten
K. M.
,
Henkel
C.
,
2008
,
ApJ
,
673
,
832

Mangum
J. G.
,
Darling
J.
,
Henkel
C.
,
Menten
K. M.
,
2013
,
ApJ
,
766
,
108

Mehringer
D. M.
,
Goss
W. M.
,
Palmer
P.
,
1994
,
ApJ
,
434
,
237

Middelberg
E.
,
Agudo
I.
,
Roy
A. L.
,
Krichbaum
T. P.
,
2007
,
MNRAS
,
377
,
731

Miyoshi
M.
et al. ,
1995
,
Nature
,
373
,
127

Momjian
E.
,
Romney
J. D.
,
Carilli
C. L.
,
Troland
T. H.
,
2006
,
ApJ
,
653
,
1172

Mundell
C. G.
,
Ferruit
P.
,
Pedlar
A.
,
2001
,
ApJ
,
560
,
168

Parra
R.
,
Conway
J. E.
,
Elitzur
M.
,
Pihlström
Y. M.
,
2005
,
A&A
,
443
,
383

Qin
S.-L.
et al. ,
2008
,
ApJ
,
677
,
353

Rovilos
E.
et al. ,
2003
,
MNRAS
,
342
,
373

Sakamoto
K.
et al. ,
1999
,
ApJ
,
514
,
68

Sakamoto
K.
et al. ,
2008
,
ApJ
,
684
,
957

Schmelz
J. T.
,
Baan
W. A.
,
Haschick
A. D.
,
Eder
J.
,
1986
,
AJ
,
92
,
1291

Skrutskie
M. F.
et al. ,
2006
,
AJ
,
131
,
1163

Slysh
V. I.
et al. ,
2001
,
MNRAS
,
320
,
217

Smith
H. E.
,
Lonsdale
C. J.
,
Lonsdale
C. J.
,
Diamond
P. J.
,
1998
,
ApJ
,
493
,
L17

Spergel
D. N.
et al. ,
2007
,
ApJS
,
170
,
377

van der Tak
F. F. S.
et al. ,
2007
,
A&A
,
468
,
627

van der Walt
D. J.
,
2014
,
A&A
,
562
,
A68

Wang
J.-Z.
et al. ,
2014
,
Nat. Commun.
,
5
6449

Zhao
J.-H.
,
An
T.
,
2008
,
Relativ. Astrophys.
,
966
,
77