EDP Sciences
Free Access
Issue
A&A
Volume 606, October 2017
Article Number A109
Number of page(s) 10
Section Extragalactic astronomy
DOI https://doi.org/10.1051/0004-6361/201731405
Published online 20 October 2017

© ESO, 2017

1. Introduction

Hydrides are formed by the first chemical reactions starting from atomic gas and are therefore at the root of interstellar chemistry. Due to the relative simplicity of their chemical network, they provide excellent diagnostics of the physico-chemical properties of the interstellar gas, especially of the low-density diffuse component (Gérin et al. 2016).

Methylidynium, CH+, was one of the first interstellar molecules to be identified, in absorption in the optical spectrum toward bright stars (Douglas & Herzberg 1941). This discovery established a long-lasting problem for astrochemistry, the formation and survival of this reactive species in the interstellar medium (ISM). Its large observed abundances are at odds with predictions from steady-state quiescent gas-phase chemical models by several orders of magnitude. The formation path of the molecule, C+ + H2 CH+ + H, has a high endothermicity of ΔE/k ~ 4300 K, and mechanisms such as shocks (e.g., Elitzur & Watson 1980), dissipation of turbulence (e.g., Godard et al. 2009, 2014), intense FUV or X-ray radiation (e.g., Morris et al. 2016), or formation from vibrationally excited H2 (e.g., Agúndez et al. 2010; Nagy et al. 2013; Faure et al. 2017) are advocated to overcome this problem.

The J = 1–0 transition of CH+ (and its 13C isotopologue), at 835 GHz (830 GHz, respectively), has been observed in absorption in Galactic sightlines toward massive star-forming regions with the Herschel Space Observatory (Falgarone et al. 2010; Godard et al. 2012). In addition, optical absorption lines of CH+ have been observed toward stars in the Magellanic Clouds (Welty et al. 2006) and even bright supernovae outside the Local Group (SN1986G in NGC 5128, D’Odorico et al. 1989; SN2006X in M 100, Cox & Patat 2008; SN2008fp in ESO 428G14, Cox & Patat 2014; SN2014J in M 82, Ritchey et al. 2015), allowing us to probe extragalactic ISM lines of sight.

The sulfanylium ion SH+ has an even higher endothermicity of ΔE/k ~ 9900 K in its formation via the reaction of S+ with ground-state H2. SH+ can also be formed by reactions of S+ with vibrationally excited H2(v> 1) (Zanchet et al. 2013) and in X-ray dominated regions by reactions of S++ with H2 (Abel et al. 2008). SH+ was first detected in space in emission toward the massive star forming region W3 IRS5 (Benz et al. 2010) with the Herschel Space Observatory and in absorption toward the star-forming region complex Sgr B2 near the Galactic Center (Menten et al. 2011) with the Atacama Pathfinder EXperiment telescope (APEX). Unlike CH+, it has never been detected in an extragalactic source before.

Because the formation of CH+ and SH+ have different endothermicity, their abundance ratio could reflect the physical properties of the region where these molecules form. Godard et al. (2012) did a comparative study of the CH+ and SH+ absorptions along multiple Galactic sightlines, and found that the CH+/SH+ column density ratios can vary by more than two orders of magnitude, from one to more than 100. They find, however, a relatively good correlation between N(CH+)/N(SH+) and N(CH+)/N(H), except toward the Central Molecular Zone in the Milky Way. There, they argue that the stronger X-ray radiation field could trigger reactions of C++ and S++ ions with H2 to enhance the abundances of CH+ and SH+.

Here, we report the detection of CH+, SH+, and their 13C- and 34S-isotopologues in the z = 0.89 absorber located in front of the quasar PKS 1830211. The quasar, at z = 2.5 (Lidman et al. 1999), is gravitationally lensed by the foreground absorber, a nearly face-on spiral galaxy (Winn et al. 2002). The directions to the two bright and compact lensed images of the quasar (separated by 1) form two independent lines of sight through the disk of the intervening galaxy, with absorption detected for many molecular species (e.g., Wiklind & Combes 1996; Muller et al. 2011, 2014a).

2. Observations

The observations were carried out with the Atacama Large Millimeter/submillimeter Array (ALMA) during its early cycles (1 and 2).

CH+ tuning: the CH+ and 13CH+J = 1–0 transitions, redshifted to ~440 GHz in ALMA band 8, were observed simultaneously on 2015 May 20. The weather conditions were excellent, with a precipitable water vapor content of ~0.3 mm. The total on-source time was approximately 17 min. The array was composed of 35 antennas, resulting in a synthesized beam of ~0.3, full-width at half maximum (FWHM). Hence, the two lensed images of PKS 1830211 were well resolved. The correlator was configured with 1.875 GHz wide spectral windows and a spectral resolution of 1.1 MHz, corresponding to a velocity resolution, after Hanning smoothing, of ~0.8 km s-1.

SH+ tuning: the SH+ and 34SH+NJ = 1201 transitions, redshifted to ~280 GHz in ALMA band 7, were observed on 2014 May 5 (two executions) and July 18 (one execution). The HO (110101) line (rest frequency 547.676 GHz, redshifted to 290.4 GHz) was also observed with the same tuning. The weather conditions were good to moderate, with a precipitable water vapor content between 0.5–2.5 mm. The total on-source time comprised about one hour. In each execution, the array was composed of 30 antennas, resulting in a final synthesized beam smaller than 0.5FWHM. The correlator was also configured with 1.875 GHz wide spectral windows and a spectral resolution of 1.1 MHz, corresponding to a velocity resolution, after Hanning smoothing, of ~1.2 km s-1.

For both the CH+ and the SH+ tunings, the data calibration was done within the CASA1 package, following a standard procedure. The bandpass response of the antennas was calibrated from observations of the bright quasar J 1924292. The gain solutions were self-calibrated on the continuum of PKS 1830211. The final spectra were extracted toward both lensed images of PKS 1830211 using the CASA-python task UVMULTIFIT (Martí-Vidal et al. 2014) and fitting a model of two point sources to the interferometric visibilities.

3. Laboratory spectroscopic data

The spectroscopic parameters for the CH+ and SH+ lines discussed in this paper are given in Table 1. The CH+ and 13CH+ rest frequencies were taken from the Cologne Database for Molecular Spectroscopy (CDMS)2 (Müller et al. 2001, 2005), based on Müller (2010). Both J = 1–0 transition frequencies were determined by Amano (2010). The rest frequencies for the SH+NJ = 1201 transitions were taken from laboratory measurements by Halfen & Ziurys (2015). The CDMS SH+ entry is based on the present work as detailed in Appendix A.

In the Born-Oppenheimer approximation, the SH+ spectroscopic data can be taken directly to derive rest frequencies of 34SH+ (Brown et al. 1986), see also Müller et al. (2015b) for the determination of NO spectroscopic parameters from data of several isotopic variants and references therein for further examples. The resulting 34SH+N = 1–0 transition frequencies are given in Table 2. Deviations from the Born-Oppenheimer approximation are not known accurately but are small in the case of a substitution of 32S by 34S. Trial fits with plausible values suggest that the most important correction, the one to the rotational constant B, may shift all N = 1–0 transition frequencies by around 1 MHz, usually, but not always, to higher frequencies. The correction to the spin-spin coupling parameter λ may cause shifts of around one to a few megahertz also, but affects mostly the upper and lower frequency FS components in opposite directions, whereas the shift of the J = 2−1 line is much smaller, possibly several 100 kHz. Other corrections as well as uncertainties from the values and uncertainties of the higher spectroscopic parameters are most likely smaller, but may add up to several 100 kHz.

We have adopted the electric dipole moment μCH+ = 1.68 Debye (Cheng et al. 2007) and note that a nearly identical value Debye was reported for the heavier isotopologue (Follmeg et al. 1987). For the sulfanylium ion we have adopted the same value μSH+ = 1.28 Debye for both isotopologues (Senekowitsch et al. 1985; Cheng et al. 2007).

Table 1

Line parameters.

Table 2

Quantum numbers, frequencies, Einstein A coefficients, upper gu and lower gl state degeneracies, and upper Eu and lower El state energies of the N = 1–0 ground state rotational transition of 34SH+.

4. Results

In this section, we present the absorption spectra of CH+, 13CH+, SH+, and 34SH+, obtained with ALMA toward the two lensed images of PKS 1830211. The spectra were first normalized to the continuum level of each image and are described as: (1)where fc is the continuum source covering factor (i.e., the fraction of the continuum emission actually covered by the absorbing clouds) which can vary for different species, and τi(v), the optical depths of different velocity components, which we assume to have Gaussian profiles. In view of the following sections and Figs. 15, this choice of Gaussian profiles is a good approximation.

The size of the continuum emission at submillimeter wavelengths is as yet unknown, but Very Long Baseline Interferometry (VLBI) measurements at 7 mm indicate a size of ~0.2 mas for the southwest image, scaling linearly with λ (Jin et al. 2003). Projected in the plane of the absorber, the apparent sizes of the submm continuum images are thus ~0.1 pc in diameter. Since ALMA cannot spatially resolve this scale, the degeneracy between fc and optical depth can only be broken for a heavily saturated line (as for CH+), or possibly if spectrally resolved multiple fine and/or hyperfine components of different strengths allow us to constrain their opacity. In the absence of such constraints, we assumed fc = 1 to derive apparent optical depths, and thus obtain lower limits to the column densities. From previous ALMA observations, Muller et al. (2014a) observed fc(SW) between 91–95%, with a trend of increasing fc with increasing frequency (i.e., when the size of the continuum emission becomes smaller). However, chemical segregation and time variations of the continuum morphology complicate the picture. The covering factor is unknown for the NE image, the OH+ absorption observed by Muller et al. (2016) implying fc(NE) > 50% in June 2015.

Next, we derived column densities by assuming that the excitation of CH+ and SH+ is strongly coupled to the cosmic microwave background (CMB, TCMB  = 5.14 K at z = 0.89, Muller et al. 2013) and that other contributions to the excitation are small. Although rotationally inelastic collisions with H, H2, and e might need to be considered, reactive collisions of CH+ with the same collision partners are even faster, so that the formation and destruction processes must be considered in proper analysis of the excitation (cf. Godard & Cernicharo 2013). Departures from the CMB can also be caused by radiative excitation in the local background radiation of the absorbing galaxy itself. We have no information about the local visible, infrared, and submm-wave continuum inside the absorbing galaxy. A local continuum comparable to that of the average background in the Milky Way would dominate the excess excitation of CH+ in competition with collisions at the densities and temperatures of a Galactic diffuse molecular cloud. Even so, the excitation temperature in CH+J = 1–0 might be raised to ~6 K, which would still have a negligible effect on the following analysis. Local excitation effects are unlikely to be any larger in SH+ under diffuse-cloud conditions. Accordingly, we calculate the column densities as: (2)where the αij coefficients (Table 1) are calculated for a given transition ij for Trot  = 5.14 K (see e.g., Muller et al. 2014a).

4.1. PKS 1830211 south-west line of sight

4.1.1. CH+ and 13CH+

thumbnail Fig. 1

Absorption spectra toward the southwest image of PKS 1830211: top: of the CH+J = 1–0, OH+N = 1–0 (deconvolved from its hyperfine structure and shown for its strongest hyperfine component), HF J = 1–0, ortho-H2O 110101, ArH+J = 1–0, and para-H2S 111000 lines; bottom: of the 13CH+J = 1–0, SH+NJ = 1201 (deconvolved from its hyperfine structure and shown for its strongest hyperfine component), ArH+J = 1–0 (opacity scaled down by a factor of three), para-H2S 111000 (opacity scaled down by a factor of three), and ortho-HO 110101 lines. All spectra are normalized to the continuum level and are referenced to the heliocentric frame taking z = 0.88582.

Open with DEXTER

Toward the SW image, the CH+ spectrum shows deep and broad absorption (Fig. 1), the broadest of all molecular species observed so far at millimeter wavelengths toward this source (e.g., Muller et al. 2014a). The full width at zero power exceeds 160 km s-1. The line appears flat-bottomed over a velocity range of ~50 km s-1 (i.e., between 30 km s-1 and +20 km s-1), although the absorption does not exactly reach the zero level, but implies a continuum source covering factor, fc ~ 97%. The covering factor is slightly higher than for saturated species observed before (Muller et al. 2014a). Most likely, this is due to the smaller continuum size at higher frequencies and/or to the intrinsic larger filling factor of CH+, although we cannot exclude time variations (Muller & Guélin 2008).

The heavy saturation prevents us to derive directly the peak opacity of CH+, its total column density, and chemical correlation with other species near v = 0 km s-1 velocities. By cutting the optical depths at a threshold of τ = 3, we estimate a lower limit of 6.4 × 1014 cm-2. The weak v = + 170 km s-1 component (Muller et al. 2011, 2014a) is also detected in CH+, as shown in Fig. 2, with an apparent peak opacity of ~0.1 and a FWHM ~ 10 km s-1, leading to a column density ~4 × 1012 cm-2.

The absorption from 13CH+ is detected in the same tuning as CH+ and barely reaches a depth of 10% of the continuum level near v = 0 km s-1. The hyperfine splitting of the 13CH+J = 1−0 line is smaller than 2 MHz (rest frame), that is, spread over a velocity interval ~0.7 km s-1. This is much smaller than the FWHM of the line, and is thus neglected here. The total column density is about 9 × 1012 cm-2.

We further discuss the CH+ absorption in comparison with other species and the [CH+]/[13CH+] ratio in Sect. 5.

thumbnail Fig. 2

Same as Fig. 1, upper panel, zoomed on the v = + 170 km s-1 velocity component toward the southwest image. The dotted curves show the best fits with one Gaussian component.

Open with DEXTER

4.1.2. SH+ and 34SH+

The SH+ absorption is detected near v = 0 km s-1 with an optical depth ~0.15. The identification of SH+ is corroborated by the presence of 34SH+ at the expected frequency (see below and Sect. 3). Spectra for both isotopologues, converted into opacity scale, are shown in Fig. 3. The absorption profile of SH+, deconvolved from its hyperfine structure and in velocity scale, is shown in Fig. 1. This profile was obtained by the fit of Gaussian velocity components convolved with the hyperfine structure. Only three Gaussian components are required to fit the profile, leaving residuals at the level of the noise, see Fig. 3. In particular, two Gaussian components are necessary to reproduce the large and asymmetric blue and red wings.

The companion absorption from 34SH+ can also be included in the same fit. As an exercise to compare with spectroscopic calculations (see Sect. 3), we assume for 34SH+ the same hyperfine structure as for SH+ (i.e., same splitting and relative line intensities) but shifted in frequency by a constant value δ, and the same intrinsic velocity profile. We find δ = 952.5 ± 1.9 MHz, consistent with but not providing more accurate values than the frequencies listed in Table 2. In this fit, we also find a ratio SH+/34SH+ of 16.2 ± 1.3, slightly higher than the 32S/34S ratios obtained previously from CS () and H2S (10.6 ± 0.9) isotopologues by Muller et al. (2006), but significantly smaller than the solar system value of 22 (Lodders 2003).

thumbnail Fig. 3

Opacity spectra of the SH+ and 34SH+ lines toward the SW image of PKS 1830211, shown with their best fit model (three Gaussian components convolved with the hyperfine structure, in red) and the fit residuals (in blue, offset by 0.03). The hyperfine structure is indicated for SH+. We assumed the same hyperfine structure, shifted in frequency by a constant value for 34SH+ (see Sect. 4.1.2).

Open with DEXTER

thumbnail Fig. 4

Absorption spectra toward the northeast image of PKS 1830211: top: of the CH+J = 1–0, OH+N = 1–0 (deconvolved from its hyperfine structure and shown for its strongest hyperfine component), HF J = 1–0, ortho-H2O 110101, and ArH+J = 1–0 lines; bottom: of the 13CH+J = 1–0, SH+NJ = 1201 (deconvolved from its hyperfine structure and shown for its strongest hyperfine component), and CH+J = 1–0 (opacity scaled down by a factor of 146). All spectra are normalized to the continuum level and are referenced to the heliocentric frame taking z = 0.88582. The 13CH+ spectrum was smoothed to 3.3 km s-1 for better signal-to-noise ratio (S/N) in individual channels.

Open with DEXTER

4.2. PKS 1830211 north-east line of sight

4.2.1. CH+ and 13CH+

The CH+ NE absorption profile consists of a series of individual velocity components, with widths as small as a few km s-1, spanning the large continuous interval between 300 km s-1 and 100 km s-1 (Fig. 4). A fit of 12 Gaussian velocity components reproduces well the overall line opacity profile (see Fig. 5 and the list of velocity components in Table 3). The absorption reaches a depth of about 80% of the continuum level at v ~ −156 km s-1, implying that the source covering factor fc is >80% for this peak absorption. Without stronger constraints, we have assumed fc = 1 toward this image. The total CH+ column density is about 1.9 × 1014 cm-2.

The 13CH+ absorption is weakly detected. It is the first such detection for a 13C isotopologue toward the NE image of PKS 1830211. It reaches a depth of ~1% of the continuum level. A simultaneous fit of the CH+ and 13CH+ spectra (with the same 12-Gaussian velocity components profile obtained above) yields a [CH+]/[13CH+] ratio of 146 ± 43. The total 13CH+ column density is about 1.3 × 1012 cm-2along the NE line of sight.

4.2.2. SH+

Despite a sensitivity better than 0.2% of the continuum level, SH+ is not detected toward the NE image. We have estimated an upper limit of the SH+ integrated opacity as km s-1, where we take the FWHM = 42 km s-1 determined from the fit of the CH+ absorption with only one Gaussian velocity component, and where σ is the standard deviation of the spectrum and δV the velocity resolution. This corresponds to an upper limit of 3.2 × 1011 cm-2for the column density. The abundance ratio between the SW and NE lines of sight, provided excitation conditions are comparable, is then larger than 120, which is much higher than for other hydrides like ArH+, OH+, H2O+, H2Cl+, and CH+ (see Table 4 and Sect. 5.2).

5. Discussion

5.1. Comparison with other species

5.1.1. Time variations

As previously shown (e.g., Muller & Guélin 2008; Schulz et al. 2015), we have to account for the effect of time variations for the comparison of spectra taken at different epochs toward PKS 1830211, with an observed timescale of the order of one month to several years. The lensing geometry of the system is such that morphological changes in the background quasar, for example due to the emission of new plasmons in the (possibly helical) jet (Garrett et al. 1997; Jin et al. 2003; Nair et al. 2005), can cause a slightly varying illumination of clouds in the absorber and thereby changes in the absorption line profiles toward both lensed images.

thumbnail Fig. 5

Fit of the CH+J = 1–0 line opacity profile toward the NE image of PKS 1830211. The individual Gaussian components (listed in Table 3) are marked in red and the global resulting profile is shown in green. The fit residuals are shown in blue, offset by 0.2 in opacity.

Open with DEXTER

Table 3

Gaussian velocity components used to fit the CH+ opacity profile toward the NE image of PKS 1830211, as shown in Fig. 5.

We do not expect micro-lensing, for example, by a stellar-mass object in the lens plane, to cause effective variability, because the continuum emitting size of the quasar at mm wavelengths is still several orders of magnitude larger than the corresponding Einstein radius, θE,micro ~ 2 μas. A transverse velocity ~1000 km s-1 in the lens plane would convert into an apparent drift of ~0.5 mpc within one year. The timescale associated with milli-lensing by an object of 1 M would thus be of the order of weeks or longer.

ArH+ (Müller et al. 2015a) and HF (Kawaguchi et al. 2016) were observed by ALMA the day before CH+ in May 2015, while OH+ was observed about two weeks later (Muller et al. 2016). In May and July 2014, the ground-state transitions of ortho-HO were observed in the same tuning as SH+, and another sulfur-bearing molecule, para-H2S, was observed within one month. Within such short time intervals, we do not expect significant variations of the absorption profiles.

Muller et al. (2016) discuss the temporal variations of H2O, CH, and H2O+ profiles between the different ALMA observing sessions in 2014, 2015, and 2016. The major changes occur in the blue and red wings of the saturated line of H2O toward the SW image, with an increase of ~50% of the integrated opacities in the velocity ranges −60 to −20 km s-1 and +20 to +80 km s-1 between 2014 and 2016. For CH, the total integrated opacity varies by less than 6%. Toward the NE image, the total integrated opacity of the water line varies by less than 15% in this period. Different fine structure transitions of H2O+, observed between 2014 and 2015, do not show significant differences. These checks have allowed us to link the profiles of the different species observed with ALMA between 2014 and 2015, and suggest little evolution of the absorption line profiles toward the two images of PKS 1830211 during this period.

Allison et al. (2017) recently published a study of the long-term variability of the 21 cm H I line toward PKS 1830211. They compare the H I profiles observed in 1996 by Chengalur et al. (1999) and 1999 by Koopmans & de Bruyn (2005) with new data obtained in 2014–2015. In stark contrast with the molecular variability at millimeter wavelengths, they find only marginal variations of at most few percent within this long time span. This is consistent with the fact that the continuum at cm wavelengths is much more extended, smoothing away all variations. In the absence of better estimates of H I column densities, we used those derived from the source kinematical model by Koopmans & de Bruyn (2005) and given in Table 4 (see also Muller et al. 2016).

Table 4

Total column densities of various species along the SW and NE lines of sight toward PKS 1830211.

5.1.2. Column density ratios

Among the species used for comparison in this paper, only OH+, and ArH+ have wings sufficiently broad toward the SW image to avoid the saturated region in the CH+ spectrum and are strong enough toward the NE image, for determining column density ratios with CH+. These ratios are shown in Fig. 6 for both lines of sight. The [OH+]/[CH+] ratio is rather constant, = 1.8 ± 0.2, in the wings of the SW absorption (v = −60 to 40 km s-1 and +20 to +80 km s-1), but varies significantly between ~1 and 18 for all other velocities of the NE absorption (v = −300 to 100 km s-1). It is ~9 for the v = + 170 km s-1 component toward the SW image. On the other hand, the [CH+]/[ArH+] ratio is varying abruptly between ~2 and 60, and never stabilizes to a constant value. We note a clear trend that when [OH+]/[CH+] is low, [CH+]/[ArH+] is high, and vice versa, one possible explanation being that CH+ traces gas with higher fH2 than ArH+ and OH+.

Assuming that the [OH+]/[CH+] ratio is constant over the SW absorption, that is, also in the saturated region of the CH+ absorption, we can perform a simultaneous fit of the CH+, 13CH+, and OH+ spectra, using Eq. (1) with a single intrinsic velocity profile for all species. The free parameters of the fit are thus the centroid velocity, FWHM, and integrated opacity for each Gaussian velocity component, the source covering factor, fc (we assumed the same for each species), and the scaling factors between species. The best fit with three Gaussian components yields fc = 97.8 ± 0.3%, [CH+]/[13CH+ ] = 97 ± 6, and [OH+]/[CH+ ] = 1.7 ± 0.1. It reproduces the spectra well and leaves residuals close to the noise level. For simplicity, we assumed Gaussian statistics on the errors in the fit. From this, we derive a new “saturation-corrected” total column density for CH+ of 9.7 × 1014 cm-2. Using the H I and H2 column densities from Table 4, we finally obtained the CH+ abundance relative to total hydrogen: N(CH+)/N(H) = N(CH+)/(N(H I) + 2×N(H2)) = 2 × 10-8 and 4 × 10-8 along the SW and NE lines of sight, respectively, comparable to typical values found in the Milky Way (e.g., Falgarone et al. 2010; Menten et al. 2011; Godard et al. 2012).

thumbnail Fig. 6

Column density ratios of [OH+]/[CH+] (top) and [CH+]/[ArH+] (bottom) toward the NE (v< −100 km s-1) and SW (v> −100 km s-1) lines of sight. The ratios are shown in black curves. The green and red curves are optical depths and translate into column densities with factors 2.85 × 1012, 20.0 × 1012, and 1.70 × 1012 cm-2 km-1 s for CH+, OH+, and ArH+, respectively. Channels with S/N< 3 or opacity larger than 2 were flagged prior to calculate the ratios.

Open with DEXTER

In Fig. 7, we show the column density ratios of SH+ relative to other species discussed in this work, for the SW line of sight. These ratios show a clear pattern with different species and velocity ranges. At | v | < 10 km s-1, the ratios [X]/[SH+] go to a minimum for X = OH+, H2O+, 13CH+, and ArH+, all hydrides preferentially tracing gas with low molecular fraction. The ratios increase by a factor two or more in the wings, for | v | > 10 km s-1. For X = ortho-HO and para-H2S, the ratios show a reverse trend in the velocity interval | v | < 10 km s-1, increasing toward the line center and dropping by a factor two or more in the wings. The | v | < 10 km s-1 velocity interval is dominated by translucent gas with a relatively high density of a few 103 cm-3, as determined from the excitation of NH3 (Henkel et al. 2008) and several other species (Henkel et al. 2009; Muller et al. 2013). In the line wings, where the low-fH2 tracers show enhanced absorption, the density is likely lower and the gas more diffuse. Hence, we can conclude that SH+ behaves intermediate between tracers of low- and high-molecular fraction, in agreement with the analysis by Neufeld et al. (2015) along Galactic sightlines.

Taking the total column densities, we find a large difference in the [CH+]/[SH+] ratios between the two lines of sight toward PKS 1830211: ~25 toward the SW image and >600 toward the NE. This agrees with the correlation derived by Godard et al. (2012) in the Milky Way, that shows a higher N(CH+)/N(SH+) ratio in regions with high N(CH+)/N(H). However, the difference of N(CH+)/N(H) between the two PKS 1830211 sightlines is only a factor of two, when the [CH+]/[SH+] varies by a factor > 24. The [CH+]/[SH+] ratios collected by Godard et al. (2012) show a large scatter of more than two orders of magnitude. They interpret these wild variations as linked to the ion-drift velocity in the turbulent dissipation regions model, therefore the amount of suprathermal energy injected in the system, and to the difference in formation endothermicities between the two species.

thumbnail Fig. 7

Column density ratios (in black) of several species (in red) relative to SH+ (in green) toward the SW image of PKS 1830211. Opacity spectra were smoothed by five channels to improve the S/N in individual channels. The column density ratios are calculated only where both species have opacity S/N larger than five.

Open with DEXTER

5.2. Comparison of the two lines of sight toward PKS 1830211

CH+, SH+, and their 13C- and 34S-isotopologues are detected toward the SW image of PKS 1830211. This line of sight is known to contain nearly 50 molecular species, not counting isotopologues, with an H2 column density of about 2 × 1022 cm-2 (see, e.g., Muller et al. 2011, 2014a). At v ~ 0 km s-1, it mostly consists of translucent clouds at moderate density of the order of 103 cm-3 and with a kinetic temperature of ~80 K (Henkel et al. 2009; Muller et al. 2013). In contrast, the NE line of sight consists of more diffuse gas (see e.g., Muller & Guélin 2008; Muller et al. 2014b, 2016), although the physical conditions are not as well constrained as toward the SW image.

Previously, four species have been found to have enhanced abundances in the NE line of sight, with respect to the SW line of sight: H2Cl+ (Muller et al. 2014b), ArH+ (Müller et al. 2015a), and OH+ and H2O+ (Muller et al. 2016). These species are also known to trace gas with low molecular fraction (Gérin et al. 2016). Muller et al. (2016) noticed that the total column density ratio between the SW and NE images (hereafter γSW/NE) for a given species is a good indicator of what kind of gas it preferentially traces. We list in Table 4 the column densities of several species observed toward both lines of sight, as well as their γSW/NE ratios. When ordered with increasing value of γSW/NE, there is a clear trend with tracers of increasing fH2 .

In this picture, CH+, with a γSW/NE = 5, is intermediate between the low fH2tracers, for example OH+ and H2O+ (tracing fH2of a few percent) and high fH2tracers, such as CH and HF, for which γSW/NE ≳ 20. On the other hand, SH+, with its non detection in the NE line of sight, yields a lower limit γSW/NE> 120, a rather extreme ratio among the species listed, either suggesting that SH+ is found in gas with relatively high-fH2 or that the physical conditions in the NE line of sight are not favorable to the formation of SH+. By comparison, Godard et al. (2012) find molecular fractions with a large scatter 0.04 <fH2< 1 for the diffuse gas seen in CH+ and SH+ absorption along Galactic sightlines, with an average value of 10%.

In their recent chemical predictions, Neufeld & Wolfire (2016) find a sequence ArH+–HCl+–H2Cl+–OH+–H2O+ where the five ions reach their peak abundance at increasing molecular fraction (or visual extinction). ArH+ is confirmed to be a unique tracer of almost purely atomic gas (i.e., even better than H I), peaking at molecular fraction 10-5−10-2 (see also Schilke et al. 2014). OH+ and H2O+ are found to reside primarily in gas with fH2 ~ 0.01−0.1, with the chlorine-bearing ions tracing intermediate fH2 ≲ 0.1. Although not perfect, the agreement between the γSW/NE-classification from the two PKS 1830211 sightlines and the chemical predictions is remarkable.

Finally, we note that the [CH+]/[CH] ratio in the NE line of sight is higher (~5) than in the SW (~1). In the Milky Way, an elevated [CH+]/[CH] ratio suggests a larger contribution of the CH+ chemistry to the formation of CH, in contrast to regions where relatively more CH arises from quiescent chemistry, as discussed by Federman et al. (1997) and Porras et al. (2014).

5.3. 12C/13C ratio

Because of its relatively large value (~60 in the local ISM, Lucas & Liszt 1998), the interstellar 12C/13C ratio can be difficult to measure, either due to saturation of the main 12C-species or sensitivity issues with the detection of the 13C-isotopologues. In addition, the interpretation of the [12CX]/[13CX] abundance ratio can be complicated by fractionation and/or selective photodissociation. From optical-line absorption studies in the Milky Way, Ritchey et al. (2011) find that the [12CH+]/[13CH+] ratio does not deviate from the 12C/13C isotopic ratio, as expected if the molecule forms via energetic processes.

Our measurement of [12CH+]/[13CH+] toward the SW image is indirect, using the [OH+]/[CH+] ratio determined from the line wings to link CH+ in the saturated velocity interval near v = 0 km s-1 to 13CH+. We derive [12CH+]/[13CH+] = 97 ± 6 in a combined fit, which would imply a large optical depth of between approximately five and ten for the peak of the CH+ absorption. This [12CX]/[13CX] ratio is much higher than the previous measurements ~30–40 using HCO+, HCN, and HNC (Muller et al. 2006, 2011). Non detection of the 13C-variants for H2CO and C2H sets limits to the ratio significantly higher than 40 (Muller et al. 2011), and may suggest fractionation issues (see e.g., Roueff et al. 2015). Alternatively, it could be possible that the isotopic ratio is different due to incomplete mixing between different gas components: low- vs. high-fH2 gas or molecular gas that has been more enriched by recent stellar formation vs. a more highly disturbed component of the ISM.

For the first time toward the NE image, we are able to measure a [12CX]/[13CX] ratio. From a combined fit of the CH+ and 13CH+ spectra, we obtain [12CH+]/[13CH+] = 146 ± 43. In this fit, we set the ratio as a free parameter and assume Gaussian statistics on the uncertainties. As for the 36Ar/38Ar ratio (Müller et al. 2015a), we find a slight difference between the two lines of sight, suggesting that the NE lines of sight, intercepting the absorber at a galactocentric radius of ~4 kpc compared to ~2 kpc for the SW line of sight, might be composed of less processed material.

6. Summary and conclusions

We report ALMA observations of CH+, SH+, and their 13C- and 34S-isotopologues along two independent lines of sight, with different physico-chemical properties, across the z = 0.89 absorber toward PKS 1830211. CH+ shows deep absorption spanning a velocity range of ~200 km s-1 along both sightlines, with 13CH+ also detected, albeit weakly, along both. In contrast, SH+ is only detected toward the SW line of sight, characterized by an higher average molecular fraction fH2 . We report the first interstellar detection of 34SH+ in the same line of sight.

The [CH+]/[SH+] column density ratios differ widely between the two sightlines, ~25 in the SW, and >600 in the NE. This suggests, in agreement with previous observations, that SH+ resides in gas with high fH2(10%). Alternatively, the difference of column density ratios might be due to the difference of formation endothermicity between the two species, with physical conditions not as favorable to the formation of SH+ in the NE line of sight.

We suggest that the total column density ratios between the two lines of sight for a given species is a good indicator of the molecular fraction where the species primarily resides. In this picture, CH+ is placed in gas with fH2 ≳ 10%, that is, with similar or possibly slightly higher fH2than for OH+ and H2O+. The CH+J = 1–0 line is the most heavily saturated line among all species previously observed in this source, and as such, appears as the best tracer of diffuse gas in terms of detectability, in particular at high redshift.

The detection of 13CH+ allows us to estimate [12CH+]/[13CH+ ] ~ 100 and ~150, toward the SW and NE sightlines, respectively. The SW value is larger than any previous [12CX]/[13CX] ratios determined from HCO+, HCN, and HNC. The even larger (although somewhat uncertain) value derived in the NE line of sight, intercepting the disk of the absorber at a larger galactocentric radius, suggests that the material might be less processed by stellar nucleosynthesis. This work shows that with the unprecedented sensitivity and frequency coverage of ALMA, we now have the opportunity to follow up the recent breakthroughs in our understandings of the physics and chemistry in the ISM of our Milky Way, with observations of hydrides in distant galaxies, and of the z = 0.89 absorber toward PKS 1830211, in particular.


Acknowledgments

We thank the referee for useful comments and corrections. This paper makes use of the following ALMA data: ADS/JAO.ALMA#2012.1.00056.S and #2013.1.00020.S. ALMA is a partnership of ESO (representing its member states), NSF (USA) and NINS (Japan), together with NRC (Canada) and NSC and ASIAA (Taiwan) and KASI (Republic of Korea), in cooperation with the Republic of Chile. The Joint ALMA Observatory is operated by ESO, AUI/NRAO and NAOJ. This research has made use of NASA’s Astrophysics Data System.

References

Appendix A: Complementary laboratory data

We have evaluated the SH+ spectroscopic parameters in the present work. Transition frequencies with microwave accuracy exist only for the N = 1–0 rotational transition; all the following data, including multiple determinations, were used in the present fit. Savage et al. (2004) reported data for the J = 0−1 and J = 2−1 fine structure (FS) components. Müller et al. (2014) analyzed ALMA data of the Orion Bar region and showed that the J = 0−1 datum of Savage et al. (2004) was in error by several megahertz and reported both hyperfine structure (HFS) components. Halfen & Ziurys (2015) obtained improved transition frequencies for all three FS components and confirmed the findings from ALMA observations (Müller et al. 2014). Even though Halfen & Ziurys (2015) only fit their own data, their fit displayed large residuals (up to 204 kHz) between measured transition frequencies and those calculated from their spectroscopic parameters much larger than the experimental 50 kHz. Exchanging the transition frequency of the J = 1−1, F = 0.5–0.5 HFS component at 683 359.227 MHz with that of a nearby line at 683 360.577 MHz yielded residuals of less than 20 kHz on average. Consequently, we employed the latter frequency in our fits. Brown & Müller (2009) derived extrapolated zero-field frequencies from laser magnetic resonance data (Hovde & Saykally 1987); in addition, rovibrational data (Brown et al. 1986; Civiš et al. 1989) were also used in the fit. The resulting parameters are given in Table A.1. We summarize for convenience the N = 1–0 transition frequencies determined with microwave accuracy in Table A.2. The re-evaluated SH+ spectroscopic parameters will be used to create an updated entry for the CDMS catalog (Müller et al. 2001, 2005).

Table A.1

Spectroscopic parametersa (MHz, cm-1) of sulfanylium, SH+.

Table A.2

Quantum numbers J and F, frequencies (MHz), uncertainties (Unc., kHz), and residuals oc (kHz) between observed rest frequencies obtained with microwave accuracy used in the present fit and those calculated from the present set of spectroscopic parameters of the N = 1–0 transition of sulfanylium, SH+, and notes on the source.

All Tables

Table 1

Line parameters.

Table 2

Quantum numbers, frequencies, Einstein A coefficients, upper gu and lower gl state degeneracies, and upper Eu and lower El state energies of the N = 1–0 ground state rotational transition of 34SH+.

Table 3

Gaussian velocity components used to fit the CH+ opacity profile toward the NE image of PKS 1830211, as shown in Fig. 5.

Table 4

Total column densities of various species along the SW and NE lines of sight toward PKS 1830211.

Table A.1

Spectroscopic parametersa (MHz, cm-1) of sulfanylium, SH+.

Table A.2

Quantum numbers J and F, frequencies (MHz), uncertainties (Unc., kHz), and residuals oc (kHz) between observed rest frequencies obtained with microwave accuracy used in the present fit and those calculated from the present set of spectroscopic parameters of the N = 1–0 transition of sulfanylium, SH+, and notes on the source.

All Figures

thumbnail Fig. 1

Absorption spectra toward the southwest image of PKS 1830211: top: of the CH+J = 1–0, OH+N = 1–0 (deconvolved from its hyperfine structure and shown for its strongest hyperfine component), HF J = 1–0, ortho-H2O 110101, ArH+J = 1–0, and para-H2S 111000 lines; bottom: of the 13CH+J = 1–0, SH+NJ = 1201 (deconvolved from its hyperfine structure and shown for its strongest hyperfine component), ArH+J = 1–0 (opacity scaled down by a factor of three), para-H2S 111000 (opacity scaled down by a factor of three), and ortho-HO 110101 lines. All spectra are normalized to the continuum level and are referenced to the heliocentric frame taking z = 0.88582.

Open with DEXTER
In the text
thumbnail Fig. 2

Same as Fig. 1, upper panel, zoomed on the v = + 170 km s-1 velocity component toward the southwest image. The dotted curves show the best fits with one Gaussian component.

Open with DEXTER
In the text
thumbnail Fig. 3

Opacity spectra of the SH+ and 34SH+ lines toward the SW image of PKS 1830211, shown with their best fit model (three Gaussian components convolved with the hyperfine structure, in red) and the fit residuals (in blue, offset by 0.03). The hyperfine structure is indicated for SH+. We assumed the same hyperfine structure, shifted in frequency by a constant value for 34SH+ (see Sect. 4.1.2).

Open with DEXTER
In the text
thumbnail Fig. 4

Absorption spectra toward the northeast image of PKS 1830211: top: of the CH+J = 1–0, OH+N = 1–0 (deconvolved from its hyperfine structure and shown for its strongest hyperfine component), HF J = 1–0, ortho-H2O 110101, and ArH+J = 1–0 lines; bottom: of the 13CH+J = 1–0, SH+NJ = 1201 (deconvolved from its hyperfine structure and shown for its strongest hyperfine component), and CH+J = 1–0 (opacity scaled down by a factor of 146). All spectra are normalized to the continuum level and are referenced to the heliocentric frame taking z = 0.88582. The 13CH+ spectrum was smoothed to 3.3 km s-1 for better signal-to-noise ratio (S/N) in individual channels.

Open with DEXTER
In the text
thumbnail Fig. 5

Fit of the CH+J = 1–0 line opacity profile toward the NE image of PKS 1830211. The individual Gaussian components (listed in Table 3) are marked in red and the global resulting profile is shown in green. The fit residuals are shown in blue, offset by 0.2 in opacity.

Open with DEXTER
In the text
thumbnail Fig. 6

Column density ratios of [OH+]/[CH+] (top) and [CH+]/[ArH+] (bottom) toward the NE (v< −100 km s-1) and SW (v> −100 km s-1) lines of sight. The ratios are shown in black curves. The green and red curves are optical depths and translate into column densities with factors 2.85 × 1012, 20.0 × 1012, and 1.70 × 1012 cm-2 km-1 s for CH+, OH+, and ArH+, respectively. Channels with S/N< 3 or opacity larger than 2 were flagged prior to calculate the ratios.

Open with DEXTER
In the text
thumbnail Fig. 7

Column density ratios (in black) of several species (in red) relative to SH+ (in green) toward the SW image of PKS 1830211. Opacity spectra were smoothed by five channels to improve the S/N in individual channels. The column density ratios are calculated only where both species have opacity S/N larger than five.

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.