Free Access
Issue
A&A
Volume 527, March 2011
Article Number A69
Number of page(s) 8
Section Extragalactic astronomy
DOI https://doi.org/10.1051/0004-6361/201015878
Published online 26 January 2011

© ESO, 2011

1. Introduction

Molecular line emission is an important tool for probing the highly obscured inner regions of starburst galaxies and buried AGNs. Line ratios within and between species help determine physical conditions and chemistry of the gas, which provide essential clues to the type and evolutionary stage of the nuclear activity. Important extragalactic probes of cloud properties include molecules such as HCN, HCO+, HNC, CN, and HC3N (e.g. Costagliola & Aalto 2010; Krips et al. 2008; Graciá-Carpio et al. 2008; Loenen et al. 2008; Imanishi et al. 2004; Gao & Solomon 2004; Aalto et al. 2002) that trace the dense (n ≳ 104 cm-3) star forming phase of the molecular gas. HCN, HNC, HCO+ and CN are all species that can be associated both with photon dominated regions (PDRs) (e.g. Tielens & Hollenbach 1985) in starbursts and X-ray dominated regions (XDRs) (e.g. Maloney et al. 1996; Meijerink & Spaans 2005) surrounding active galactic nuclei (AGN). In contrast, HC3N requires shielded dense gas to survive in significant abundance since it is destroyed by UV and particle radiation (e.g. reactions with the ions C+ and He+) (Prasad & Huntress 1980; Rodriguez-Franco et al. 1998, e.g.). These ions are expected to be abundant in for example XDRs. Thus HC3N line emission may identify galaxies where the starburst is in the early, embedded, stage of its evolution. There is still, however, substantial dichotomy in the interpretation of the line emission from the above molecules and therefore new tracer species to combine with existing information are important.

The hydronium ion H3O+ is a key species in the oxygen chemistry of dense molecular clouds, and useful as a measure of the ionization degree of the gas (Phillips et al. 1992). Since H3O+ formation requires H2O to exist in the gas-phase, the H3O+ molecule acts as a natural filter to select hot (T > 100 K) molecular gas, and therefore traces more specific regions than molecules usually surveyed towards other galaxies. The chemistry of this filtering is the evaporation of icy grain mantles at T ≈ 100 K (van der Tak et al. 2006b), or that H2O forms in the gas phase at high (T > 300 K) temperatures.

Observations of H3O+ emission will therefore help probe the location of dense, warm and active gas in galactic nuclei. Combined with information on the H2O abundance, these observations may further be used to trace the ionization rate by cosmic rays (produced in supernovae, i.e., starbursts) and/or X-rays (from an AGN). Studies by van der Tak et al. (2006a) have demonstrated this use of H3O+ for Sgr B2 in the Galactic centre. Furthermore, recent Herschel observations of Galactic sources show 984 GHz 0 H3O+ absorption in the diffuse gas towards G10.6-0.4 (W31C) (Gerin et al. 2010) and 1.03–1.63 THz H3O+ emission from the high mass star forming region W3IRS5 (Benz et al. 2010). In general H3O+ abundances agree well with expectations from PDR models and – together with H2O+ and OH+ – provide important further insight into gas phase oxygen chemistry.

Recently we have detected 32+ – 22+ H3O+ in the nearby starburst M 82 and the ultraluminous galaxy Arp 220 (van der Tak et al. 2008). Derived column densities, abundances, and H3O+/H2O ratios indicate ionization rates similar to or even exceeding that in the Galactic centre. In M 82 the extended evolved starburst (PDR) is a likely source of this ionization rate while, for the ULIRG Arp 220, an AGN-origin (XDR) is suggested.

In the XDR and PDR models grain-processing is not taken into account since the chemistry is taking place in the gas phase. However, a formation route for H3O+ involving the evaporation (or removal by shocks) of H2O from grain surfaces need also to be considered. For example, evaporating H2O reacting with HCO+ may provide an important source of H3O+ (e.g. Phillips et al. 1992; van der Tak & van Dishoeck 2000).

We have used the James Clerk Maxwell Telescope (JCMT) in Hawaii to observe the 364 H3O+ line (upper level energy Eu = 139 K) in seven starburst and active galaxies which cover a range of environments. Our goal is to use H3O+ as a tracer of gas properties in galactic nuclei and to see if H3O+ can serve as a diagnostic tool to distinguish AGN from starburst activity. In Sects. 24 the sample, observations and their results are presented. H3O+ line parameters are presented in Sect. 4.1 and in Sect. 4.2 H3O+ column densities and fractional abundances are calculated. In Sects. 5.1 and 5.2 H3O+ abundances in the context of X-ray and UV irradiated models are discussed and in 5.3 the potential importance of grain chemistry. In the last Sect. 5.5, we present a brief future outlook.

2. The sample

Table 1

Sample galaxiesa.

We have selected a sample consisting of seven nearby luminous starburst and AGNs – and one distant ULIRG. The galaxies are selected from their bright HCN line emission. From our previous experience with extragalactic H3O+ we knew that the line is weaker than the standard high density gas tracers such as HCN and HCO+ so we restricted ourselves to a relatively small sample of seven objects (coordinates, FIR luminosities and distances are presented in Table 1):

IC 342 is a nearby barred Scd galaxy of moderate luminosity (central 400 pc has LFIR of 6 × 108   L) with a central starburst. Within its central 300 pc (30″) two molecular arms end in a clumpy central ring of dense gas (e.g. Downes et al. 1992) which surrounds a young star cluster. The ring is suggested to outline the X2 orbits in a larger-scale bar. The chemistry of IC 342 has been investigated in detail in a high-resolution study by Meier & Turner (2005). They find that the chemistry in the ring is a mixture of PDR-dominated regions and regions of younger star-forming clouds. The chemistry in the bar/arms is dominated by shocks as shown by CH3OH (Meier & Turner 2005) and SiO (Usero et al. 2006). Five (A–E) giant molecular clouds (GMCs) are found in the ring and arms. The 13″ JCMT beam of our H3O+ observations is pointed towards the region of GMC B – but also includes GMCs A and E. GMCs B and C are the sites of young (a few Myr) star formation and are also the regions where incoming molecular clouds meet the ring. GMC A has more PDR-like characteristics. The dust temperature of IC 342 is estimated to 42 K (e.g. Becklin et al. 1980).

NGC 253 is also a nearby barred galaxy located in the Sculptor group with a compact nuclear starburst and a IR luminosity that appears to originate in regions of intense massive star formation within its central few hundred parsecs (Strickland et al. 2004). From their 2 mm spectral scan Martín et al. (2006) suggest that the chemistry of NGC 253 shows strong similarities to that of the Galactic centre molecular region, which is thought to be dominated by low-velocity shocks. High resolution SiO observations show bright emission resulting from large scale shocks as well as gas entrained in a nuclear outflow (García-Burillo et al. 2000). Note also that it is suggested that NGC 253 is a galaxy in which a strong starburst and a weak AGN coexist (e.g. Weaver et al. 2002; Müller-Sánchez et al. 2010). High resolution observations of HCN and HCO+ 1–0 (Knudsen et al. 2007) show strongly centrally concentrated emission. The 13″ JCMT beam covers the bulk of the nuclear emission from HCN and HCO+. The central dust temperature of NGC 253 is estimated to 50 K (Melo et al. 2002).

NGC 1068 is the nearest example of a type 2 Seyfert galaxy luminous in the infrared. Surrounding the AGN there is a 4″ circumnuclear molecular ring or -disk (CND) and on a larger scale there is a NIR stellar bar 2.3 kpc long. This bar is connected to a large-scale, molecular starburst ring that contributes about half the bolometric luminosity of the galaxy (e.g. Telesco et al. 1984; Scoville et al. 1988; Tacconi et al. 1994; Helfer & Blitz 1995; Tacconi et al. 1997). Bright HCN 1–0 line emission is observed towards the CND while the HCO+ 1–0 emission is relatively fainter by a factor of 1.5 (e.g. Kohno et al. 2001). Only the CND and a fraction of the bar is covered by our JCMT beam and we adopt a source size of 2″ for the CND. The dust of the inner 4″ appears to show a strong temperature gradient – from about 800 K in the very inner region to 150–275 K at a distance of 0''̣8 or greater (Tomono et al. 2006). Alloin et al. (2000) find temperatures of about 150 K 200 pc from the nucleus. There is a also radio jet from the nucleus which falls into our JCMT beam (e.g. Wilson & Ulvestad 1983).

NGC 4418 is a luminous, edge-on, Sa-type galaxy with a deeply dust-enshrouded nucleus (Spoon et al. 2001). NGC 4418 is a FIR-excess galaxy with a logarithmic IR-to-radio continuum ratio (q) of 3 (Roussel et al. 2003). This excess may be caused by either a young pre-supernova starburst or a buried AGN (Aalto et al. 2007; Roussel et al. 2003; Imanishi et al. 2004). Unusually luminous HC3N line emission (Aalto et al. 2002, 2007; Costagliola & Aalto 2010) has been interpreted as a signature of young starburst activity. Mid-IR intensities are indicative of dust temperatures of 85 K (Evans et al. 2003) inside a radius of 50 pc (0''̣5). The IR luminosity-to-molecular gas mass ratio is high for a non-ULIRG galaxy indicating that intense, compact activity is hidden behind the dust. Recent high resolution imaging of CO 2–1 (Costagliola et al., in prep.) indicate a molecular source size of 0''̣5.

NGC 6240 is an infrared luminous merger of two massive spiral galaxies. Two AGN/LINER nuclei are separated by 1″ and the bulk of the molecular gas has piled up between the two nuclei (Iono et al. 2007; Tacconi et al. 1999). Luminous HCO+ 4–3 emission is also emerging from in between the two nuclei where the H2 emission is also located. The mid-IR sources (24.5 μm) are associated with the two X-ray nuclei – the brightest by far being the southern nucleus (T = 55–60 K) (Egami et al. 2006). It has been suggested that the medium between the two nuclei is dominated by starburst superwinds from the southern nucleus (Ohyama et al. 2000) and the molecular gas is indeed highly turbulent (Iono et al. 2007; Tacconi et al. 1999). The JCMT beam encompasses the entire molecular structure of this galaxy and we adopt a source size of 1″ for the H3O+ emission.

IRAS 15250+3609 is a relatively distant ultraluminous galaxy – probably a major merger with a dominant bright nucleus. A tidal feature appears to emerge from the southwest side of the nucleus and turns around on the eastern side to create an enormous closed ring, 27 kpc in diameter. The optical spectrum is a composite of H II and LINER features (Veilleux et al. 1999). IRAS 15250 exhibits deep silicate absorption features (Spoon et al. 2006) indicating a deeply enshrouded nucleus. Our JCMT beam covers the whole optical galaxy.

Arp 299 is an IR-luminous merging system of two galaxies, IC 694 and NGC 3690. Strong 12CO emission has been detected from the nuclei of IC 694 and NGC 3690 and from the interface between the two galaxies (e.g. Sargent & Scoville 1991; Aalto et al. 1997). The two nuclei, as well as the western overlap region, currently undergo intense star formation activity (e.g. Gehrz et al. 1983). Bright HCN 1–0 emission is emerging from both nuclei as well as from the overlap region (e.g. Aalto et al. 1997). Water emission from NGC 3690 suggests that it also harbours an AGN (Tarchi et al. 2007). To cover all three regions a map was carried out.

3. Observations

The 364.7974 GHz line of H3O+ was observed towards the sample galaxies in February and August 2008, with the James Clerk Maxwell Telescope (JCMT1) on Mauna Kea, Hawaii. We used the 16 elements heterodyne array HARP. Each of the receptors of the array has a beam size of 14″ at 345 GHz and the sources were placed in the receptor labeled H052. The back end was the Auto-Correlation Spectrometer and Imaging System (ACSIS), providing 1.0 GHz bandwidth in 2048 channels and a maximum resolution of 0.4 km s-1. In order to optimize baseline stability, double beam switching at a rate of 1 Hz was used with an offset of 120″. Weather conditions were optimal, with about 1 mm of precipitable water vapor, leading to a typical noise rms of 2 mK at 10 km s-1 resolution. Telescope pointing was checked every hour on the CO line emission of nearby AGB stars and always found to be within 2″. Data were extracted into fits files with the starlink software and analyzed in CLASS. Linear baselines were subtracted, line parameters were calculated by fitting Gaussian profiles to the spectra. The resulting spectra are shown in Fig. 1. Throughout this paper, velocities are in the heliocentric frame and redshift is computed using the radio convention.

thumbnail Fig. 1

Observed spectra of H3O+. The intensity scale is in , not corrected for the JCMT beam efficiency, that at this frequencies is 0.7. The parameters of the detected lines were obtained by Gaussian fits, shown on the plots as dashed lines.

Open with DEXTER

4. Results

4.1. Line intensities and line widths

Integrated line intensities, line widths and fitted velocities (heliocentric) for the seven observed galaxies can be found in Table 2. p-H3O+ line emission was detected towards IC 342, NGC 4418, NGC 253, NGC 1068 and NGC 6240, while Arp 299 and IRAS 15250 were not detected.

In general it is found that line widths and shapes (Fig. 1) agree well with those found for HCN and HCO+ in the central regions of the systems (Nguyen et al. 1992; Aalto et al. 2007; Knudsen et al. 2007). The most noteworthy deviation from this is NGC 6240 where the linewidth of HCN and HCO+ (Greve et al. 2009) is more than a factor of two greater than what we find for H3O+. A possible explanation for this could be the weak H3O+ signal resulting in a moderate signal-to-noise profile. For NGC 1068 the line width agrees well with that of HCO+ 4–3 but is somewhat narrower than what is observed for HCN 4–3 (202 km   s-1) (Pérez-Beaupuits et al. 2009).

For IC 342 an unidentified (U) line was detected in February 2008 – but it did not appear again in August the same year. The line is somewhat narrower than the H3O+ line and appear at a velocity of 263 km   s-1 (rest frequency 364.477 GHz). It is possible that it is an atmospheric O3 line.

Table 2

H3O+ line resultsa.

4.2. Column densities, excitation and abundances

Since we have no direct information on the excitation of the H3O+ molecule we make some assumptions that will have to be tested in future observations. The critical density of the H3O+ transition is high (about ncrit = 106 cm-3, Phillips et al. 1992) and the low reduced mass of the H3O+ molecule makes its excitation very sensitive to radiative pumping by dust. We therefore assume that the coupling between the colour temperature of the IR emission and the excitation temperature of H3O+ holds as long as the dust colour temperature is above 30 K (see discussion in van der Tak et al. 2008).

For each galaxy we ran a RADEX model (van der Tak et al. 2007) with temperatures ranging from T = 10–500 K. In these calculations, the volume density is set to an arbitrary large value, so that the excitation of H3O+ is thermalized. The adopted excitation temperature is essentially the radiation temperature and a radiative excitation with one excitation temperature is simulated. We therefore do not use the full non-LTE capacity of RADEX.

The results are shown in Fig. 2. Each plot has temperature (kinetic or radiation) on the x axis and H3O+ column on the y axis. The contour line corresponds to the observed brightness temperature. This depends on our assumption for the source size.

H3O+ abundances relative to H2 (X(H3O+)) are listed in Table 3. We have used CO 2–1 spectra from the literature and used a Galactic conversion factor from CO luminosity to H2 mass (2.5 × 1020 cm-2 K-1 km-1 s) to estimate N(H2). Note that for Galactic nuclei (and starbursts) it is argued that this conversion factor overestimates the H2 column density by factors of 5–10 (see e.g. Martín et al. (2010)). Since we wish to reduce the risk of overestimating the H3O+ abundances we have however still adopted the standard conversion factor. For IC 342 and NGC 253 we have estimated and average N(H2) for the 13″ beam size. For NGC 1068, NGC 4418 and NGC 6240 we have used the same adopted source size as for H3O+.

thumbnail Fig. 2

On the x axis we find the temperature of the exciting background radiation and on the y axis the column density of H3O+. For each galaxy, the brightness temperature of the H3O+ transition is shown in a gray scale map. The thick black contour marks the observed value, with one sigma errors marked by dashed lines. Background temperatures estimated by IR observations are also reported and used to derive a column density consistent with the observed brightness temperature. Note that the column density fits include only the para column.

Open with DEXTER

4.2.1. Errors in H3O+ column density and abundances

Our assumptions on the excitation and spatial extent of H3O+ introduce errors in the column density estimates. For IC 342 in particular the source size error can be significant. We assume that the source fills the beam which gives us an average column density for GMCs A, B and E, but the real column density towards a particular cloud should be higher. For NGC 253 the double peaked line shape suggests that the H3O+ emission is emerging from both dense-gas peaks in the centre and is filling the JCMT beam. For NGC 1068 the H3O+ source size in the CND could be anything from  <1″ to 4″. We have assumed a source size of 2″. For NGC 4418 and NGC 6240 the source sizes are  ≲ 1″ which means that the beam dilution is significant, but no H3O+ emission is missed.

In Fig. 2 the impact of the assumption of the excitation temperature on the resulting H3O+ column density is illustrated. Note the strong dependence of N(H3O+) for Tex ≲ 60 K. For NGC 253 (for example) N(p-H3O+) changes from 9 × 1014 cm-2 when Tex = 50 K to 7 × 1015 for Tex = 25 K. When Tex exceeds 60 K, however, the temperature dependence on N(H3O+) is almost gone. Thus for all galaxies, except IC 342, we are not likely to overestimate N(H3O+) by more than a factor of 2 since the adopted Tex ≳ 50 K. It is possible that H3O+ could be collisionally excited. In this case the excitation temperature is likely significantly lower than the ones we assume here (beacuse of the high critical density). In Fig. 2 it is evident that the N(H3O+) would go up considerably with decreasing temperature. With the assumed Tex the resulting N(H3O+) is already quite large for all galaxies – even larger H3O+ columns would be an interesting result indeed, but very difficult to explain with current models. The excitation can be constrained through observing multiple transitions and Requena-Torres et al. (in prep.) are currently studying the 307 and 364 GHz line of H3O+ in a sample of starburst galaxies.

Table 3

Parametersa for the RADEX modeling of the H3O+ Line.

We conclude that the source size uncertainties dominate the errors in the column density calculations. To improve future N(H3O+) calculations it is necessary to determine source sizes through high resolution observations and to obtain information on Tex(H3O+) through multi-transition observations.

In addition it should be noted that the estimates of the H3O+ relative abundances are dependent on the reliability of the conversion factor from CO luminosity to H2 mass.

5. Discussion

5.1. H3O+ abundances and PDR/XDR models

We compare the column densities of H3O+ and H2 in Table 3 to chemical models of clouds irradiated by either far-UV, photon dominated regions (PDRs) or X-ray photons (XDRs). In general, we expect the XDRs to dominate in molecular ISMs surrounding an AGN. The XDR and PDR models we use are presented in Meijerink & Spaans (2005) and Meijerink et al. (2007). In general, the thermal and chemical structure of XDRs and PDRs are different. Larger parts of an XDR can be maintained at high temperaure and the ionization fraction of an XDR can be up to two orders of magnitude higher (xe = 10-2 − 10-1) than in a PDR. In an XDR H2O may form in the gas phase at high temperatures (200–300 K) and then react with either H or HCO+ to form H3O+.

Source averaged HCO+ abundances are typically  ≈ 10-8 for all detected galaxies except for IC 342. A general result of the models presented in van der Tak et al. (2008) is that the X(H3O+) does not exceed 3 × 10-9 in PDRs – even in the models with an extremely high cosmic-ray ionization rate of ζ = 5.0 × 10-15 s-1. H3O+ abundances approaching 10-8 and beyond are more likely to occur in X-ray dominated regions and we conclude that the XDR scenario fits the X(H3O+) values of NGC 253, NGC 1068, NGC 4418 and NGC 6240. For IC 342, the results are consistent with a PDR, but a model with a high ζ of 5.0 × 10-15 s-1.

5.1.1. Water abundances

The H3O+/H2O ratio is an even better probe of the ionization rate, and potential nature of the emitting source. It is straight forward to obtain H3O+/H2O abundance ratios as large as 10-2 in XDR models, while for PDR models ratios are generally 10-3 or less. Herschel has already measured water abundances in M 82 (Weiß et al. 2010) (see also Sect. 5.4) and results for other galaxies will follow soon.

5.2. Relative H3O+ and HCO+ abundances and XDR/PDR models

Kohno et al. (2001) and Imanishi et al. (2004) find HCN/HCO+ 1–0 line intensity ratios greater than unity in several Seyfert nuclei – where also the HCN/CO 1–0 line ratio is high. Furthermore, Graciá-Carpio et al. (2006) find elevated HCN/HCO+ 1–0 line ratios in ULIRGs. This is often interpreted as a sign of an underabundance of HCO+ compared to HCN due to X-ray chemistry. (Although it is important to remember that an abundance difference cannot be unambiguously deduced from a single-transition line ratio). Underabundant HCO+ in XDRs has been proposed in theoretical work by Maloney et al. (1996). However, more recent models by Meijerink & Spaans (2005) suggest that HCO+ may be underabundant in moderate column density (NH < 1022.5 cm-2) XDRs – but for larger columns the reverse is true and X(HCO+) generally exceeds X(HCN). Here we compare our derived N(H3O+) with N(HCO+) for the sample galaxies to see if large H3O+ abundances are paired with a particularly low HCO+ abundance, and to compare N(H3O+)/N(HCO+) with the expectations from current models.

In the XDR/PDR models, the formation of both H3O+ as well as HCO+ is mainly driven by reactions with H2, H, and H, and destruction by electrons, unless the electron fractional abundance is very low xe ≲ 10-7 − 10-8. However, the interpretation of an N(H3O+)/N(HCO+) abundance ratio in an XDR/PDR scenario is not entirely straightforward, but the models provide some useful limits. For example, in the PDR models H3O+/HCO+ ratios are unlikely to become larger than 3 (Meijerink et al. 2010), while for the XDR models it is quite straightforward to obtain N(H3O+)  > N(HCO+), but the obtained ratio is very column density dependent. When the column densities are small (e.g N(H)  ≈ 1022 cm-2) ratios as large as 20 can be obtained, but abundances (and associated brightness temperatures) are small, and in order to obtain significant HCO+ and H3O+ column densities larger clouds are needed. However, when increasing the column, the H3O+/HCO+ column density ratio slowly decreases, but moderate ratios around 4–5 are easy to reproduce with the XDR model.

5.2.1. N(H3)/N(HCO+) in the sample galaxies

Table 4

HCO+/H3O+ratios.

In Table 4 we list N(HCO+) and N(H3O+)/N(HCO+) for the sample galaxies. The largest N(H3O+)/N(HCO+) value is found in NGC 1068 which has a Seyfert nucleus and its inner few hundred pc has been suggested to be an XDR (e.g. Usero et al. 2004; Pérez-Beaupuits et al. 2009; García-Burillo et al. 2010). The large abundance ratio of  ≈ 24 suggested for NGC 1068 is consistent with (a low column density) XDR.

For NGC 4418 and NGC 253 XDR models can explain the H3O+/HCO+ ratios (but we note that the values with errors are also within the range of PDR models). For NGC 253 there is evidence that both an AGN and a young starburst is present while for NGC 4418 the nuclear activity is so obscured that the nature of the activity cannot be discerned (see Sect. 2). The H3O+/HCO+ ratio for IC 342 is consistent with a PDR model as is the relative H3O+ abundance.

For NGC 6240 the H3O+/HCO+ ratio is consistent with both XDR and PDR models while the H3O+ relative abundance favours an XDR. It is interesting to note that even if the molecular gas of NGC 6240 is collected in-between the two nuclei, their radiation could still impact the gas. The two AGNs are separated by 1''̣5 which means that the molecular gas is irradiated by intense X-ray emission from two directions on only 370 pc distance. This scenario could possibly explain both the relatively high H3O+ and HCO+ abundances.

5.3. Grain chemistry and the formation of H3O+

Above we have found that the XDR/PDR models can potentially explain the H3O+ and HCO+ column densities and abundances we found for the observed galaxies. However, for NGC 4418 the problem with an XDR explanation is the large column densities of gas and dust observed towards its centre (see footnote of Table 3) where Costagliola & Aalto (2010) find global HC3N abundances similar to those found towards hot cores in Sgr B2. Intense vibrational line emission suggest that the HC3N indeed exists in warm-to-hot environments in the very centre of the galaxy. The HC3N abundances are not consistent with either XDRs or PDRs.

Thus, we have searched for alternative explanations to the H3O+ line emission outside of the XDR and PDR models. Elevated HC3N line emission is consistent with the conditions in warm, dense shielded gas associated with embedded star formation. Is it then possible to obtain relative H3O+ abundances of  ≈ 10-8 without the X-ray chemistry?

In a scenario where grains are important, the H2O can evaporate off the grains at temperatures  > 100 K and in Phillips et al. (1992) the destruction of HCO+ through the reaction of H2O is found to be an important formation process for H3O+ when the gas is dense and warm. For a water abundance X(H2O) of 10-5 and a density n(H2) of 105 cm-3Phillips et al. (1992) find an H3O+ abundance, X(H3O+), of  ≈ 10-8 and X(HCO+) a factor 2–3 lower. It is unclear however, if the N(H3O+)/N(H2O) can be greater than 10-3 within the context of their model. van der Tak & van Dishoeck (2000) find the same effect in their study of the of Galactic protostar GL 2136. In their Fig. 1 the temperature and density structure is presented including calculated concentrations of HCO+ both with and without destruction by H2O.

We suggest that for NGC 4418 it is possible that the H3O+ abundance and the N(H3O+)/N(HCO+) ratio can be explained by a model where the gas is warm and dense and water is coming off the grains to react with HCO+ to form H3O+. This process could contribute to a reduction of the HCO+ abundance in the gas phase which may result in a lower HCO+ line intensity in any transition. Note that shocks may be responsible for removing the water from the grains making it available for further reactions including H3O+ formation. Flower & Pineau Des Forêts (2010) presents models where water abundances are related to the type and strength of shocks and there are a multitide of Galactic observational studies showing enhancement of water emission in shocks (Nisini et al. 2010; Lefloch et al. 2010; Wampfler et al. 2010; Melnick et al. 2008). Apart from causing H2O to come off the grains, shocks can result in efficient gas-phase formation of H2O from OH in the shocked high-temperature regime.

5.4. H2O+, H3O+, and H2O

Weiß et al. (2010) find a (potentially) surprisingly low N(H2O)/N(H2O+) ratio of only a few which they attribute to H2O evaporating off the grains through shocks, the H2O then becomes photodissociated into O and OH. Ion-molecule reactions then make H2O+ while H2O abundances remain low. It is furthermore noteworthy that X(H2O + ) > X(H3O+) in M 82 (van der Tak et al. 2008; Weiß et al. 2010). This is consistent with the low density (n = 103 cm-3), high cosmic ray rate (>5 × 10-15 s-1) model (Model 2) of Meijerink et al. (2010). This model does however not include any grain processing.

In van der Tak (2010a) a photoevaporation-ionization hypothesis for the H3O+/H2O+/H2O ratio is discussed for M 82 noting that testing this theory requires calculation of the photodissociation cross-section of H3O+. Whether this hypothesis holds for other galaxies, however, remains to be seen: the H2O/H2O+ ratio in M 82 may well be unusually low. In the ULIRG Mrk 231 N(H2O) ≫ N(H2O+) and the H2O+ line emission feature is consistent with an XDR interpretation (van der Werf et al. 2010b).

5.5. Future studies

With this study we have confirmed that the 364 GHz line of p-H3O+ is feasible to detect in starburst and active galaxies and that H3O+ fractional abundances are substantial. The HIFI heterodyne spectrometer onboard ESA’s Herschel space observatory is currently obtaining H2O and H3O+ data around 1 THz and more accurate estimates of the H2O/H3O+ abundance ratios can be determined for many sources improving the understanding of the ionization rates of nuclear molecular regions. Determining the cource size and excitation of H3O+ is essential to further improve the understanding of the impact of the nuclear activity on its surrounding interstellar medium and to facilitate a more accurate comparison with models. In the near future, grain chemistry and impact of shocks should be added to the interpretation of the data. In particular the results for NGC 4418 – and to some degree also NGC 253 – emphasizes this conclusion.


1

The JCMT is operated by the Joint Astronomy Centre on behalf of the Science and Technology Facilities Council of the United Kingdom, the Netherlands Organization for Scientific Research, and the National Research Council of Canada.

Acknowledgments

We are very grateful to the staff of the JCMT for their help and support and we thank the referee, Santiago Garcia-Burillo, for a thorough and very useful report which improved the paper. This research has made use of the NASA/IPAC Extragalactic Database (NED) which is operated by the Jet Propulsion Laboratory, California Institute of Technology, under contract with the National Aeronautics and Space Administration.

References

All Tables

Table 1

Sample galaxiesa.

Table 2

H3O+ line resultsa.

Table 3

Parametersa for the RADEX modeling of the H3O+ Line.

Table 4

HCO+/H3O+ratios.

All Figures

thumbnail Fig. 1

Observed spectra of H3O+. The intensity scale is in , not corrected for the JCMT beam efficiency, that at this frequencies is 0.7. The parameters of the detected lines were obtained by Gaussian fits, shown on the plots as dashed lines.

Open with DEXTER
In the text
thumbnail Fig. 2

On the x axis we find the temperature of the exciting background radiation and on the y axis the column density of H3O+. For each galaxy, the brightness temperature of the H3O+ transition is shown in a gray scale map. The thick black contour marks the observed value, with one sigma errors marked by dashed lines. Background temperatures estimated by IR observations are also reported and used to derive a column density consistent with the observed brightness temperature. Note that the column density fits include only the para column.

Open with DEXTER
In the text

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.