Online material

Appendix A: Spectral survey and line catalog

The full spectrum, from 30 to 50 GHz, is available as an ASCII file at the Strasbourg Astronomical Data Center (CDS).

Fig. A.1 Spectrum of the ATCA 7 mm survey toward PKS 1830 − 211. The grey envelope indicates the 3σ noise level. The LTE synthetic model, assuming Trot = 5.14 K for all molecules, is overlaid in red. The spectrum in magenta, offset at a level of 0.995, gives the relative differences between observations taken in September 2009 and March 2010, normalized to the average spectrum. The corresponding scale is given on the right of each box (±5%). Some non-detections are indicated in light blue.

Open with DEXTER

	Fig. A.1 Continued.
Open with DEXTER

	Fig. A.1 Continued.
Open with DEXTER

	Fig. A.1 Continued.
Open with DEXTER

	Fig. A.1 Continued.
Open with DEXTER

	Fig. A.1 Continued.
Open with DEXTER

Appendix B: Notes on individual molecules

In this section, we give short reviews of the different species and discuss constraints or new results obtained from this survey.

B.1. C₂H

The highly reactive ethynyl radical was first discovered and identified toward the Orion nebula by Tucker et al. (1974) based on its hyperfine structure, while it had still not been observed in the laboratory. C₂H has turned out to be both an ubiquitous and abundant species in the interstellar medium, including in diffuse clouds (Lucas & Liszt 2000b) where its abundance is of the order of a few 10^-8 with respect to H₂.

The ground state transition N = 1–0 at ν₀ = 87 GHz is decomposed into six components by hyperfine splitting. Toward PKS 1830−211, the strongest hyperfine components were previously observed by Menten et al. (1999) with the VLA, although at low spectral resolution. We detect here all six N = 1–0 hyperfine components arising from the SW absorption, as well as the two strongest hyperfine components (the J = 3/2–1/2, F = 2–1 and F = 1–0) related to the NE absorption at  −147 km s^-1 (see Fig. B.1). The absorption depths reflect the LTE relative intensities of the hyperfine components, indicating that the lines are optically thin.

B.2. HCN, HNC and HCO⁺

Molecular absorption was first reported toward PKS 1830 − 211 by Wiklind & Combes (1996b) after the detection of the HCN, HNC, and HCO⁺ 2–1 transitions, allowing the redshift of the intervening galaxy to be derived. Figure B.2 shows the full spectra of the HCO⁺, HCN, and HNC 1–0 transitions observed in our survey. The first two lines are likely saturated and their opacities could be underestimated. In particular two findings seem to indicate that the HCO⁺ column density might be underestimated by a factor of 2–3. First, the HCO⁺/H¹³CO⁺ ratio toward FG0.89SW is a factor of  ~2 lower than the corresponding HCN/H¹³CN and HNC/HN¹³C ratios. Second, the ratio of the column densities along the two lines of sight, SW/NE is also a factor 2–3 lower for HCO⁺ than for other species (see Table 4). In any case, the uncertainty in the opacity introduced by the uncertainty in the continuum background illumination is nevertheless large for both HCO⁺ and HCN (see Sect. 3.3)

Time variations and the discovery of additional velocity components are discussed in Sect. 3.2 and Sect. 3.5, respectively.

B.3. N₂H⁺

N₂H⁺ is thought to be one of the most abundant interstellar ions, and plays a key role in nitrogen chemistry. It is detected in a large number of Galactic clouds. However, it remains elusive in diffuse clouds (Liszt & Lucas 2001), with an abundance upper limit more than two orders of magnitude lower than its abundance in FG0.89SW. We note that N₂H⁺ is also rare in translucent clouds, e.g., Turner (1995a) having detected it in only 2 out of 16 such clouds.

B.4. HOC⁺

The hydroxymethylidynium ion, HOC⁺, isomer of the formyl ion, was first tentatively identified in Sgr B2 by Woods et al. (1983). Its presence in several dense molecular clouds was then confirmed by Ziurys & Apponi (1995) and Apponi & Ziurys (1997), with a relative abundance HCO⁺/HOC⁺ ~ 360–6000. This abundance ratio is however about one order of magnitude lower (~70–120) in diffuse clouds (Liszt et al. 2004), and seemingly has an in-between value in PDR: Apponi et al. (1999) derived HCO⁺/HOC ⁺  ~ 270 toward the Orion bar. Martín et al. (2009) find a ratio  ~70 toward NGC 253, similar to that in diffuse clouds.

In FG0.89SW, we derive a ratio HCO⁺/HOC⁺ of  ~60, although this value might be slightly underestimated because of the large opacity of HCO⁺. The H¹³CO⁺/HOC⁺ ratio is indeed 2.6 ± 0.1. Assuming a ¹²C/¹³C ratio of 35, the HCO⁺/HOC⁺ ratio in FG0.89SW is comparable to that in diffuse clouds.

B.5. HCO

The formyl radical, HCO, was discovered in the interstellar medium by Snyder et al. (1976) and has been identified as a tracer of the cloud-illuminated interface in PDRs (de Jong et al. 1980; Schilke et al. 2001). García-Burillo et al. (2002) revealed widespread HCO emission in the nuclear starburst of M 82, with HCO, CO, and HII emission in nested rings, HCO extending farther out in the disk. Gérin et al. (2009) demonstrated that the HCO emission in the Horsehead nebula traces the PDR front, coincident with PAH and hydrocarbon emission. The abundance of HCO in the PDR is comparable to that of HCO⁺, whereas it becomes about one order of magnitude less in the dense regions shielded from UV radiation. To the best of our knowledge, HCO has not been detected in diffuse or translucent clouds to date, although it is located close in frequency to the H¹³CO⁺ (1–0) and SiO (2–1) lines.

We clearly detect three (out of the four) hyperfine components of the ground state transition 1₀₁−0₀₀, the remaining low intensity component being buried in the noise level. The relative optical depths are consistent with LTE intensities, suggesting that the lines are optically thin. We derive an abundance ratio HCO⁺/HCO of  ~13, similar to the values observed in dense shielded regions. This value is even larger if we consider that the HCO⁺ opacity might be underestimated by a factor 2–3.

B.6. CH₂NH

We identify the absorption feature near 33.93 GHz with the 1₀₁−0₀₀ ground state transition of methanimine, as it is the simplest interstellar molecule with lowest energy transition close to that frequency. The hyperfine structure is not resolved. Only one line of methanimine is located within the limits of our frequency coverage. The strongest transitions, with predicted τ > 0.1 based on the 1₀₁−0₀₀ absorption depth, have rest frequencies of 127.9, 166.9, 225.6 and 284.3 GHz. Of these, only the second and last fall in an ALMA band for a redshift of z = 0.89.

Methanimine was first detected in Sgr B2 by Godfrey et al. (1973) and has been further observed in several star-forming regions and translucent clouds (Turner et al. 1999), but is not detected in cold dark clouds (Turner 1991; Dickens et al. 1997). It has been tentatively detected in NGC 253 (Martín et al. 2006) and has been observed in Arp 220 (Salter et al. 2008). Methanimine is a prebiotic molecule, precursor to more complex molecules such as glycine.

B.7. CH₃NH₂

Methylamine is the saturated and terminal product of the hydrogenation series based on the cyanide radical: CN-HCN-CH₂NH-CH₃NH₂.

Fig. B.1 Spectrum of the C2H N = 1–0 line. The different hyperfine components are indicated with their relative intensities for the SW (top) and NE (bottom) absorptions. Note the clear detection of the two strongest hyperfine components arising from FG0.89NE. The red curve shows the LTE synthetic spectrum, taking column densities of 1.3 × 1015 and 4 × 1013 cm-2 toward the SW and NE images respectively, and a rotation temperature of 5.14 K.

Open with DEXTER

	Fig. B.2 Full spectra of the HCO+, HCN, and HNC 1–0 transitions, showing the absorption toward the SW (0 km s-1) and NE ( − 147 km s-1) images of the quasar. A zoom in with a narrower intensity scale, shown in Fig. 6, reveals additional velocity components.
Open with DEXTER

Methylamine has a very complex spectrum resulting from the combination of internal rotation of the CH₃ group, the asymmetry of the molecule and the inversion of the NH₂ group (Takagi & Kojima 1973; Ilyushin & Lovas 2007). It was first detected in Sgr B2 by Kaifu et al. (1974) (see also Turner 1991). No extragalactic detection has been reported to date.

B.8. CH₃OH

Interstellar methanol was first detected toward Sgr B2 and Sgr A by Ball et al. (1970) and is widely observed toward star forming regions, dark clouds, and also translucent clouds (Turner 1998). It has been detected in various extragalactic sources (e.g. Henkel et al. 1987). Methanol is about one order of magnitude less abundant in cold/translucent clouds than in star forming regions, where its abundance can reach a few 10^-8 relative to H₂. It is not detected in diffuse clouds, down to upper limits of a few lower than its abundance in cold clouds (Liszt et al. 2008). Methanol is the terminal product of hydrogenation of CO.

The methanol spectrum has a multitude of transitions, of which only one is strong enough to be detected within our frequency coverage. Therefore, no excitation analysis can be deduced. Assuming LTE equilibrium with the CMB radiation field, we derive a relatively high methanol abundance of a few 10^-8 relative to H₂ in FG0.89SW. Several strong lines (with predicted τ ~ 0.02–0.4) near rest frequency of 300 GHz and redshifted in the 2 mm band are (simultaneously) observable (e.g. with ALMA or PdBI) and could provide a stringent measurement of the methanol rotation temperature.

The calculations of Jansen et al. (2011) have shown that methanol could be a sensitive probe of spatial and temporal variations in the proton-to-electron mass ratio.

B.9. l-C₃H

We identify the feature around 40.4 GHz with the J = 7/2–5/2 transition of the rotational ladder of the propynylidyne radical (ν₀ = 76.2 GHz). The transition is split into two Λ-doublets (see Fig. B.3), approximately equal in strength and spaced by about 5 MHz in the rest frame, which secures our identification. The hyperfine structure is unresolved. No other transitions of l-C₃H are detected in our frequency coverage.

Linear propynylidyne was first identified by Thaddeus et al. (1985a) toward IRC+10216 and TMC 1. It is observed in translucent clouds (Turner et al. 2000) but remains undetected in diffuse clouds (Lucas & Liszt 2000b, although this might just be a sensitivity issue, as their upper limits of l-C₃H/HCO⁺ is still higher than the value we derive). No extragalactic detections have been reported to date that we are aware of.

The cyclic isomer form c-C₃H is detected in dense (Yamamoto et al. 1987; Fossé et al. 2001) and translucent (Turner et al. 2000) clouds, with c-C₃H/l-C₃H ratio  ~ 5. No transitions of c-C₃H are detected in our survey. In particular, the non-detection of the N_KaKc = 2₁₂-1₁₁ transition at a rest-frame frequency of 91.5 GHz, which is connected to the ground state, yields an upper limit of  ~3 × 10¹² cm^-2, and a c-C₃H/l-C₃H ratio  < 0.4. We note that a low ratio c-C₃H/l-C₃H  <  1 was found by Turner et al. (2000) toward one translucent cloud (CB228).

	Fig. B.3 Spectrum of the l-C3H J = 7/2–5/2 transition. The locations of the two Λ-doublets with hyperfine structure are indicated.
Open with DEXTER

	Fig. B.4 c-C3H2 transitions. The dashed curve on the wing of the 212 − 101 line corresponds to the LTE prediction for the HCS+J = 2–1 line with a HCS+ column density of 5 × 1012 cm-2.
Open with DEXTER

B.10. c-C₃H₂

The ring molecule cyclo-propenylidene c-C₃H₂ is a fairly ubiquituous interstellar species (e.g., Thaddeus et al. 1985b; Matthews & Irvine 1985; Cox et al. 1988; Madden et al. 1989).

We detect 2 ortho and 2 para lines, yielding an ortho/para ratio of 2.8 ± 0.1, comparable to the statistical weight 3:1.

A tight correlation exists between the abundances of C₂H and c-C₃H₂ in diffuse clouds (Lucas & Liszt 2000b; Gérin et al. 2011), with N(C₂H) = (28 ± 1.4)N(c-C₃H₂). We derive a comparable abundance ratio C₂H/c-C₃H₂ = 23.6 ± 0.2.

The excess absorption in the blue wing of the c-C₃H₂ 2₁₂–1₀₁ line (see Fig. B.4) could be due to the known velocity component at  ~−20  km s^-1 toward FG0.89SW (see e.g. Fig. 5 from Muller et al. 2006). Alternatively, we note that this feature could be due the HCS⁺ 2–1 line, which would fall at  ~−32 km s^-1 with respect to the center of the c-C₃H₂ 2₁₂–1₀₁ line. HCS⁺ is indeed detected in diffuse clouds with an abundance of  ~3 × 10^-10 relative to H₂ (Lucas & Liszt 2002). If its abundance were similar in FG0.89SW, then the LTE predicted absorption depth would be of the order of τ = 0.015, i.e. compatible with the observations. There are no other lines of HCS⁺ within our frequency coverage. Observations of the J = 1–0, 3–2, or 4–3 transitions (redshifted to 22.6, 67.9, and 90.5 GHz, and with predicted opacity of τ = 0.007, 0.013, and 0.006, respectively) could confirm the presence of HCS⁺.

B.11. l-C₃H₂

Propadienylidene, l-C₃H₂, is the linear isomer of c-C₃H₂ that was first discovered by Cernicharo et al. (1991) in the carbon-chain rich dark cloud TMC 1. Noticeably, Maier et al. (2011) identified l-C₃H₂ as the probable carrier of two conspicuous diffuse interstellar bands (at 4881 and 5450 Å).

The ortho/para ratio is , in agreement with the statistical weight of 3:1.

The linear form is usually less abundant than the cyclic one, with a cyclic/linear ratio of 3–5 in diffuse clouds, 70 in TMC 1, and  >150 in Sgr B2 (Cernicharo et al. 1991, 1999, see also Turner et al. 2000). We derive a cyclic/linear ratio of 25 ± 3, which is in-between the ratio observed toward diffuse clouds and TMC 1.

B.12. H₂CCN

The non-linear cyanomethyl radical has a very complicated spectrum owing to the unpaired electron and hyperfine splitting induced by the H and N atoms. We detect two absorption features, corresponding to N_Ka,Kb = 3₀₃–2₀₂ and 4₀₄–3₀₃. Both are from the ortho species (ortho states are characterized by K_a even and para states by K_a odd).

Irvine et al. (1988) first identified H₂CCN toward Sgr B2 and TMC 1. It was also detected in the circumstellar envelope of IRC+10216 (Agúndez et al. 2008), but, to the best of our knowledge, it has not been observed toward other sources. In particular, this is the first extragalactic detection of this molecule.

B.13. CH₃CCH

Methyl acetylene was first detected in Sgr B2 and Orion A by Snyder & Buhl (1973). It is observed in dark clouds (e.g., Irvine et al. 1981) and has been detected in the nearby starburst galaxies NGC253, NGC4945 and M82 (e.g. Mauersberger et al. 1991; Wang et al. 2004).

We detect the J_K = 5₀–4₀ transition, but not the J_K = 4₀–3₀, suggesting that the rotation temperature is higher than 5.14 K. The kinetic temperature was fixed to 50 K (see B.14).

We note that CH₃CCH is the molecule in our survey which shows the miminum scatter in relative abundances between the different sources.

	Fig. B.5 Spectra of the 4K − 3K and 5K − 4K transitions of CH3CN. Our LTE prediction assuming a kinetic temperature of 50 K is overlaid.
Open with DEXTER

B.14. CH₃CN

Methyl cyanide was first detected in space toward the Galactic center by Solomon et al. (1971). It is commonly detected in hot cores (e.g. Olmi et al. 1993), where the heating from massive stars is thought to evaporate the products of grain surface chemistry. The abundance of CH₃CN in these regions of active star formation exceeds 10^-9 relative to H₂ (e.g. Kalenskii et al. 2000). We note that CH₃CN is also observed in dark clouds (Matthews & Sears 1983) and has been detected in NGC253 (Mauersberger et al. 1991). On the other hand, CH₃CN is detected in neither diffuse (Liszt & Lucas 2001) nor translucent clouds (Turner 2000), with upper limits clearly lower than abundances in dark clouds.

Symmetric-top molecule are excellent temperature probes, because radiative transitions between the different K-ladders are forbidden, so that only collisional excitation is effective. Consequently, the rotation temperature of the molecule can be derived within the same K-ladder on the one hand, and the kinetic temperature is reflected in the population of the K levels on the other hand (e.g. Churchwell & Hollis 1983).

In our LTE synthetic spectrum approach, the term in Eq. (2) was changed to , where E_KK is the lowest energy level of each K-ladder, to evaluate the opacity of each transition (see e.g., Olmi et al. 1993). The partition function was calculated accordingly by taking into account both the rotation and kinetic temperatures (B.1)where g_J = (2J+1) and the product g_Ig_K = 2 for K = 0 and K ≠ 3n (n > 0), and g_Ig_K = 4 for K = 3n.

Two groups of transitions of methyl cyanide fall in the frequency coverage of our survey, the J_K = 4_K–3_K redshifted near 39 GHz, and the J_K = 5_K–4_K, redshifted to  ~ 48.8 GHz. Both are clearly detected. Remarkably, the K-ladder transitions are detected up to K = 3 (see Fig. B.5), with lower energy state up to 70 K above the ground level. We note that CH₃CN is the only species in our survey showing absorption from such high energy levels. This suggests that the kinetic temperature is high, in agreement with the observations of inversion lines of ammonia up to (J,K) = 10,10 (which lies at about 1000 K above the ground state) by Henkel et al. (2008). On the basis of their estimates, the kinetic temperature T_kin is about 80 K for 80–90% of the ammonia gas, while the remaining fraction could even reach T_kin > 600 K.

A robust fitting of the CH₃CN spectra is difficult as the different K-ladder transitions are close in frequency relative to the width of the lines. Nevertheless, a good match of the observations could be found using our LTE model by assuming a CH₃CN column density of 1.3 × 10¹³ cm^-2 and a kinetic temperature of 50 K (see Fig. B.5). This latter parameter value is comparable to, although slightly lower than that obtained by Henkel et al. (2008). We neglect the broadening of the line due to hyperfine splitting, as it is at most of 1.5 MHz for the principal hyperfine components of the J_K = 5₄–4₄ transition.

B.15. H₂CCO

Ketene was first detected in Sgr B2 by Turner (1977). It was detected in line surveys of regions of low-mass and massive star formation (e.g. Turner 1991; van Dishoeck et al. 1995; Ruiterkamp et al. 2007), as well as in both translucent and dark clouds (Turner et al. 1999; Kaifu et al. 2004), suggesting that it is a common interstellar species. However, only few specific studies of the molecule have been conducted, and its formation mechanism(s) is subject to controversy: H₂CCO could be either produced by gas-phase reactions (Turner et al. 1999) or formed by means of surface chemistry, followed by grain desorption (Ruiterkamp et al. 2007). Its typical abundance is of the order of 10^-10–10^-9 relative to H₂.

We detect four different (ortho) transitions of H₂CCO in our survey, yielding a rotation temperature of  K, significantly lower than T_CMB = 5.14 K at z = 0.89.

B.16. HNCO

Isocyanic acid was first discovered toward Sgr B2 (Snyder & Buhl 1972). We detect two lines of HNCO, the 3₀₃–2₀₂ and the 4₀₄–3₀₃, the latter in a low-noise region of the survey. They do not appear to be at LTE equilibrium, and we could not derive a positive rotation temperature. The centroid velocity of HNCO is also significantly offset with respect to the average velocity of all molecules. We could not identify lines from other molecules close in frequency, and therefore discard the possibility of a false identification.

B.17. SiO

Silicon monoxide is another commonly observed interstellar molecule, considered as a good shock tracer. Its abundance is low in cold dark clouds (of the order of or less than 10^-12 relative to H₂), where it is presumably heavily depleted onto dust grains, but can reach values several orders of magnitude higher in outflows and shocked gas. It is found to have an intermediate abundance (SiO/H₂ ~ 10^-10) in diffuse and translucent clouds (Lucas & Liszt 2000a; Turner 2000). Toward FG0.89SW, we find a SiO abundance of nearly 4 × 10^-10 that is comparable to, though slightly higher than, that in diffuse/translucent clouds.

Interestingly, Ziurys et al. (1989) found a correlation between ln[SiO/HCN] and the inverse of the kinetic temperature in Galactic molecular clouds: were this correlation to hold for the gas in FG0.89SW, the SiO/HCN ratio would then point to a kinetic temperature of  ~ 30 K, a value not so far from the kinetic temperature estimated from CH₃CN.

The ²⁹SiO isotopologue is also detected and we derive a ²⁸Si/²⁹Si ratio of , which is roughly half the Solar System value.

	Fig. B.6 Spectrum of NS (2–1). The hyperfine structure with relative intensities is indicated.
Open with DEXTER

B.18. CH₃CHO

Acetaldehyde is an asymmetric rotor, with two symmetry states (A and E) resulting from the rotation of the methyl group. Electric dipole transitions are not allowed between these states, which have the same weight statistic. Each transition is dedoubled by the symmetry, and the resulting spectrum consists of a series of closeby A/E doublets.

The detection of acetaldehyde was first reported toward Sgr B2 (Ball et al. 1971) and the molecule has been observed in cold dark clouds (Matthews et al. 1985) and translucent clouds (Turner et al. 1999). Interestingly, Chengalur & Kanekar (2003) observed that CH₃CHO is more widespread toward the Galactic center than most other complex organic molecules. They argue that the molecule could be produced by shock-induced disruption on dust mantles.

No extragalactic detection has yet been reported for this species.

B.19. NS

Nitrogen sulfide was first detected in Sgr B2 by Gottlieb et al. (1975). It is detected in quiescent clouds (McGonagle et al. 1994) and in massive-star forming regions (McGonagle & Irvine 1997). There are no reports of NS detection in diffuse or translucent clouds to date.

The rotational levels of the molecule are divided into two levels (e/f) with opposite parity by Λ-doubling, each of them subdivided into hyperfine structure corresponding to the nitrogen nucleus. We clearly detect the hyperfine components with highest relative intensities (see Fig. B.6). We note that the frequency interval between  ~36 and 40 GHz is the region observed with the lowest noise level in our survey, allowing these weak lines of NS, H₂CS, C₂S, and SO⁺ to be detected.

Taking the CS column density estimated by Muller et al. (2006), we find a ratio NS/CS  ~  0.02, comparable to the values 0.02–0.05 measured by Hatchell & Viti (2002) in a sample of hot cores. Martín et al. (2003) estimated a ratio NS/CS = 0.4 in NGC 253, which they interpreted as the signature of NS chemical enhancement by shocks (see also Viti et al. 2001).

B.20. H₂CS

Thioformaldehyde was first detected in absorption toward Sgr B2 by Sinclair et al. (1973). It is observed in star forming regions and cold dark clouds (e.g. Irvine et al. 1989; Minh et al. 1991) as well as in translucent clouds (Turner 1996). Tentative extragalactic detections are reported by Heikkilä et al. (1999) in the LMC and Martín et al. (2006) in NGC253.

Two lines of H₂CS are detected, one of each ortho and para form, from which we derive an ortho/para ratio of . However, we note that another ortho line, the 2₁₂–1₁₁ at a rest frequency of 67.653820 GHz, should have been detected at the same intensity as the 2₁₁–1₁₀ line, according to the LTE prediction. It is fitted with an opacity twice lower, although with a significance of just 3σ on its integrated opacity, which is the reason we discarded it from our estimate of the ortho-H₂CS column density.

The H₂CO/H₂CS ratio, derived from their para forms, is 32 ± 6, fully consistent with the present oxygen-to-sulphur cosmic abundance ratio (33:1, Lodders 2003).

B.21. SO

Sulfur monoxide was first detected by Gottlieb & Ball (1973) toward several star forming regions and later observed in dark clouds by Rydbeck et al. (1980).

From the J_K = 2₁–1₀ line detected in our survey, we derive an abundance of  ~10^-9 relative to H₂, similar to that observed in Galactic diffuse clouds (Lucas & Liszt 2002), and  ~25 lower than that in translucent clouds (Turner 1995b).

The non-detection of the SO₂ 1₁₁–0₀₀ ground state transition yields a ratio SO/SO₂ > 8. Both species are observed in quiescent translucent clouds with an abundance ratio SO/SO₂ ~ 1–15 (Turner 1995b). It has been found that SO₂ is not detected in diffuse clouds, down to a fractional abundance  < 10^-9 (Lucas & Liszt 2002).

The non-detection of ³⁴SO points to an upper limit of SO/³⁴SO  >  13.7 (3σ). Muller et al. (2006) derived a ratio ³²S/³⁴S of 10.5 ± 0.6 from measurements of CS and H₂S, and their isotopologues.

B.22. SO⁺

The reactive radical SO⁺ is characterized by a ²Π_1/2 ground state. Its spin-doubled rotational transitions are split by Λ-type doubling.

The SO⁺J = 3/2–1/2 e line is partly blended with the F = 5/2–3/2 f hyperfine transition of NS. The width of the spectral feature at ν_sky ~ 36.807 GHz, as obtained by the fit of a single Gaussian component, is 37 ± 5 km s^-1, i.e. much larger than the width of other lines. Nevertheless, after taking into account the SO +  e/f doublet together with the NS line, the LTE synthetic spectrum reproduces fairly well the observations.

It has been found that SO⁺ is a rather widespread and quite abundant interstellar species (Turner 1994), and is seen, although weakly, in translucent clouds (Turner 1996). There, the abundance ratio SO/SO⁺ is of the order of 30. We estimate a ratio SO/SO ⁺  = 15 ± 3 of comparable order of magnitude toward FG0.89SW.

B.23. C₄H

The butadiynyl radical was first reported by Guélin et al. (1978) toward IRC+10216 and has been observed toward dark clouds by Irvine et al. (1981). It is present in translucent clouds with a quite high abundance of a few 10^-8 relative to H₂ (Turner 2000). No extragalactic detection has been reported so far.

The interaction of the unpaired electron with the molecular rotation produces fine structure, which is divided further into (here unresolved) hyperfine structure by the proton spin. The resulting spectrum of C₄H consists of a series of closeby doublets, of which we detect four (N = 6–5 to N = 9–8) within the limits of our frequency coverage.

B.24. HC₃N

The highly unsaturated cyanoacetylene molecule was first detected in Sgr B2 (Turner 1971) and is observed in various sources, such as translucent clouds, dense dark clouds, and star forming regions, with abundances of  ~ 10^-10–10^-9 relative to H₂. HC₃N emission has also been detected in several extragalactic objects (see Lindberg et al. 2011 for a recent review).

We detect all four transitions of HC₃N within the frequency range of our survey, from J = 7–6 to J = 10–9. The J = 3–2 and J = 5–4 lines of HC₃N were previously detected with the VLA by Menten et al. (1999). More recently, Henkel et al. (2009) reported the detection of all transitions from J = 3–2 to J = 7–6 with the Effelsberg 100 m telescope. The J = 10–9 line was not detected. The results of their observations are reported in the rotation diagram together with our data points (Fig. 4). They appear to be inconsistent with one another. The integrated opacities of the J = 7–6 line, observed in common with both works, differ by a factor of  ~ 2, as for their fitted linewidths. Time variablity and/or errors in the spectral baseline could explain the observed differences. We note that the higher S/N J = 7–6 line appears slightly asymmetrical, with a shoulder on the blue-shifted side, reminescent of that of the HCO⁺ 2–1 profile observed at higher spectral resolution (Muller et al. 2006; Muller & Guélin 2008). This transition was therefore not included in the Gaussian fit to derive the velocity and linewidth.

The points corresponding to the four lines observed by us appear remarkably well-aligned in the rotation diagram, suggesting that the lines are at LTE. They yield a rotation temperature of T_rot = 6.3 ± 1.3 K, just slightly higher than the expected CMB temperature. On the other hand, the points corresponding to the lines observed by Henkel et al. (2009) at lower frequencies are not aligned in the rotation diagram. While we cannot discard possible time variations, another explanation could be a change in the source covering factor with frequency, as discussed by Henkel et al. (2009). Observations of all these transitions within a short period of time would help clarify the interpretation of these data.

B.25. C₂S

The thioxoethenylidene radical was first identified by Saito et al. (1987) in previous spectra of Sgr B2 and TMC 1, based on laboratory spectroscopy studies. It has been found that C₂S is present in translucent clouds (Turner et al. 1998) with a typical abundance of  ~10^-9 and is detected in the 2 mm survey of NGC 253 by Martín et al. (2006). It is the heaviest species detected in our survey. We measure an abundance of  ~2 × 10^-10 relative to H₂.

Few lines (10 ≥ J_u ≥ 16) of C₃S fall within the frequency coverage of our survey. Their non-detection gives a (not tightly constraining) ratio C₂S/C₃S  >  0.7. Search for lower J transitions of C₃S would be more sensitive. Fuente et al. (1990) measured a typical abundance ratio C₂S/C₃S = 5 in dark clouds.

© ESO, 2011