EDP Sciences
Free Access
Issue
A&A
Volume 599, March 2017
Article Number A22
Number of page(s) 24
Section Interstellar and circumstellar matter
DOI https://doi.org/10.1051/0004-6361/201628916
Published online 22 February 2017

© ESO, 2017

1. Introduction

Massive stars have a strong impact on their environment and play an important role in influencing the physical and chemical structure of the interstellar medium. One of the most important feedback effects is radiation feedback by the far-ultraviolet (FUV) radiation of massive stars, which dissociates molecules in their surroundings and creates different chemical layers as a function of depth into the molecular clouds. These regions between the fully ionized and fully molecular clouds are photon dominated regions (PDRs), which can be studied well at sub-mm and far-infrared wavelengths. See Hollenbach & Tielens (1997) for a review.

In this paper we study the physical and chemical structure of the well-known Orion Bar PDR based on data from an unbiased spectral line survey. The Orion Bar is part of the Orion molecular cloud (OMC) 1, located at a distance of ~420 pc (Menten et al. 2007; Hirota et al. 2007). Parts of the OMC 1 region are irradiated by the Trapezium cluster which consists of four massive stars including the O6 star Θ1 Ori C. The Hii region around the Trapezium is surrounded by the Orion Bar on its south-eastern side and the Orion Ridge on its western side (Fig. 1). In addition to the PDR interfaces around the Trapezium, there is ongoing star formation in two parts of the OMC 1 region: the Orion BN/KL and Orion S regions. While Orion BN/KL (e.g. Tercero et al. 2010) is part of the OMC 1, Orion S is located within the ionized nebula in front of the OMC 1 cloud (O’Dell et al. 2009). Thanks to its nearly edge-on orientation, the Orion Bar is an ideal source to study PDR physics and chemistry. With its high FUV field (1−4×104χ0), in Draine (1978) units and average kinetic temperature of 85 K (Hogerheijde et al. 1995), it is a prototypical high UV-illumination, warm PDR. Part of the molecular line emission measured toward the Orion Bar corresponds to an “interclump medium” with densities between a few 104 and 2 × 105 cm-3 (Simon et al. 1997). It has been suggested that other molecular lines originate in clumps with densities in the range between 1.5 × 106 and 6 × 106 cm-3 (Lis & Schilke 2003). Evidence for densities higher than the average gas density of 105 cm-3 (Hogerheijde et al. 1995) has also been found based on OH transitions, tracing a warm (Tkin~ 160220 K) and dense (nH~ 106–107 cm-3) gas component with a small filling factor (Goicoechea et al. 2011). The clumpiness of the Orion Bar was found to be consistent with the intensity of [Cii] and different 12/13CO and HCO+ transitions using a three-dimensional PDR model representing the Orion Bar PDR by a clumpy edge-on cavity wall (Andree-Labsch et al. 2017).

Previous spectral line surveys of the Orion Bar were carried out at mm- and sub-mm wavelengths including the 279–308 GHz (Leurini et al. 2006) and 330–360 GHz (van der Wiel et al. 2009) ranges, and in different bands of the IRAM 30 m telescope in the 80–359 GHz frequency range (Cuadrado et al. 2015). In addition to the line surveys performed with ground-based telescopes, a line survey in the 4471545 GHz range was carried out at low spectral resolution using Herschel/SPIRE (Habart et al. 2010).

In this paper, we present a Herschel/HIFI line survey of spectrally resolved data covering the 480–1250 GHz and 14101910 GHz frequency ranges. Previous papers based on this line survey include the first detection of HF in emission toward a Galactic source (van der Tak et al. 2012). This line survey also resulted in the analysis of the properties of reactive ions such as CH+ and SH+ (Nagy et al. 2013), and OH+ (van der Tak et al. 2013). Two HO transitions detected as part of this line survey allowed the ortho-to-para ratio of water to be measured toward the Orion Bar (Choi et al. 2014). We also analysed six high-N transitions (N = 6–5, ..., 10–9) of C2H as a part of this line survey and found them to trace warm and dense gas including a component with densities of n(H2) ~ 5 × 106 cm-3 and temperatures of Tkin~ 400 K (Nagy et al. 2015b). The already studied molecules detected in this line survey suggest a complex structure with disparate gas components traced by different species.

In this paper we give a general description of the line survey and aim to provide a complete picture of the physical and chemical structure of the Orion Bar based on our Herschel/HIFI line survey.

thumbnail Fig. 1

13CO 3–2 integrated intensity (in K km s-1 units) of the Orion Bar observed with the JCMT. The circles at the position of the CO+ peak (αJ2000 = 05h35m20.6s, δJ2000 = −05°25′14′′) indicate the beam sizes corresponding to the smallest (band 7b, ~11′′) and largest (band 1a, ~44′′) beam sizes of the data presented in this line survey. The white symbols show the position of the Trapezium stars. The white contours show the distribution of the integrated intensity (in K km s-1 units) of HCN 43 measured with the JCMT (van der Wiel et al. 2009). The contours are 25%, 40%, 60%, and 80% of the peak HCN intensity of ~24 K km s-1.

Open with DEXTER

2. Observations and data reduction

The CO+ peak (αJ2000 = 05h35m20.6s, δJ2000 = −05°25′14′′) of the Orion Bar (Störzer et al. 1995) was observed using the Heterodyne Instrument for the Far-Infrared (HIFI, De Graauw et al. 2010) on board the Herschel Space Observatory Pilbratt et al. (2010) as part of the HEXOS1 guaranteed-time key program (Bergin et al. 2010). The observations were carried out as spectral scans, and cover the full frequency coverage of HIFI (480–1250 GHz and 1410–1910 GHz). The total integration times of the observations are 2.4 h (band 1a), 2.2 h (band 1b), 3.4 h (band 2a), 2.6 h (band 2b), 1.6 h (band 3a), 2.8 h (band 3b), 4.3 h (band 4a), 2.6 h (band 4b), 5.7 h (band 5a), 4.5 h (band 5b), 7.5 h (band 6a), 5.6 h (band 6b), 4.4 h (band 7a), and 5.9 h (band 7b). The average beam sizes of the scans are in the range between ~39′′ (band 1) and 11′′ (band 7) corresponding to 0.08 pc and 0.02 pc, respectively. The Wide-Band Spectrometer (WBS) backend was used which covers the 4 GHz bandwidth in four 1140 MHz sub-bands at 1.1 MHz resolution, equivalent to 0.66 km s-1 at 500 GHz and 0.18 km s-1 at 1850 GHz.

thumbnail Fig. 2

Full baseline-subtracted spectrum toward the CO+ peak of the Orion Bar. Bands 5, 6, and 7 were smoothed from the original velocity resolutions of ~0.12 km s-1, ~0.09 km s-1, and ~0.08 km s-1 to the ~1 km s-1, ~1.5 km s-1, and ~1.4 km s-1 channels, respectively.

Open with DEXTER

The data were reduced using the Herschel interactive processing environment (HIPE, Ott et al. 2010) pipeline versions 9.0 and 10.0 and are calibrated to TA scale. First, the double sideband (DSB) scans were deconvolved using the doDeconvolution task in HIPE with the strongest lines (TA> 10 K) removed, reducing the probability of ghost lines in the resultant single sideband (SSB) spectrum. Second, we performed another deconvolution with the strongest lines present to recover the data at strong line frequencies. Finally, the strong lines were incorporated into the weak line SSB spectrum. Before the deconvolution the baseline subtraction was done using the fitBaseline task of HIPE. Spurs and other bad data such as spectra with very high rms noise levels (compared to the typical rms noise levels of the scans) were flagged using the flagTool in HIPE. The main error of the deconvolution step is the uncertainty from the sideband gain divided by two, i.e. 5% for most frequencies and about a factor of two higher at a few frequencies (around 630, 870, 1260 GHz) based on Higgins et al. (2014). The velocity calibration of HIFI data is accurate to ~0.5 km s-1 or better. Based on fitting second-order polynomials, the continuum at the observed frequencies is negligible as it is similar to the measured rms noise levels. The rms noise levels at the frequency of the detected lines is shown in Table A.1. The average rms noise levels of the spectral scans are 0.02–0.04 K for band 1, 0.04–0.09 K for band 2, 0.09–0.15 K for band 3, 0.08–0.27 K for band 4, 0.30–0.42 K for band 5, 0.35–0.46 K for band 6, and 0.30–0.48 K for band 7. High rms noise levels (compared to the average values) were measured in the three frequency ranges that were affected by problems in the HIFI gain stability, namely for LO frequencies around 863 GHz and 1046 GHz, and between 1540 and 1560 GHz. Apart from these frequencies, the baseline quality is good, as shown in Appendix B. The resulting spectrum toward the Orion Bar CO+ peak is shown in Fig. 2.

Finally, we derived correction factors for the integrated intensities measured with different beam sizes, which is important when comparing the line intensities of such data, as is done in Sects. 4.2 and 4.3. The largest and smallest average beam sizes for the HIFI bands (~39′′ and 11′′) corresponding to the line survey data are shown in Fig. 1 on a 13CO 3–2 map observed with the JCMT2 (JCMT archive3). To derive correction factors for the integrated intensities measured with different beam sizes corresponding to the different bands, we need a model or proxy for the spatial extent of the emission (a simple scaling by the beam size is only possible for point sources). To this end, we used JCMT images at a spatial resolution of ~15′′, as we did in Nagy et al. (2013). In Nagy et al. (2013) we used an HCN J = 4−3 map. In addition to HCN 4–3, we also used 13CO 3–2 (Fig. 1) in this paper. The beam size of 15′′ is slightly above but comparable to the smallest beam size of the scans corresponding to band 7 (11′′). For this method we assume that the spatial distribution of molecular species observed in this line survey is similar to those of HCN or 13CO. The two species represent limiting cases: HCN predominantly traces gas near the PDR surface, while 13CO represents the bulk of the emission from the molecular material. We use the 13CO integrated intensity distribution as a reference to derive correction factors, and use the HCN integrated intensity distribution to estimate error bars of the integrated intensities converted to the different beam sizes. We convolved the 13CO 3–2 and HCN 4–3 images to spatial resolutions equivalent to the values corresponding to the HIFI bands, i.e. 39′′ for band 1, 30′′ for band 2, 25′′ for band 3, 21′′ for band 4, and 19′′ for band 5. To trace the spatial distribution of the gas at the resolutions of band 6 (15′′) and 7 (11′′) we used the original (15′′ beam) JCMT images. We compared the integrated intensities corresponding to the different spatial resolutions at a position close to the CO+ peak to derive correction factors between the different beam sizes. Based on the derived correction factors, we converted the observed integrated intensities to a ~39′′ beam size, equivalent to the beam size of band 1 (Roelfsema et al. 2012) when using LTE and non-LTE methods to estimate physical parameters in Sects. 4.2 and 4.3. The derived correction factors based on HCN between the integrated intensities of the different bands and band 1 are 0.84 for band 2, 0.77 for band 3, 0.70 for band 4, 0.67 for band 5, and 0.59 for bands 6 and 7. The derived correction factors based on 13CO between the integrated intensities of the different bands and band 1 are 0.94 for band 2, 0.91 for band 3, 0.88 for band 4, 0.87 for band 5, and 0.84 for bands 6 and 7. In this paper we used the conversion factors based on 13CO, and used the values based on HCN to give an estimate of the errors of the integrated intensities resulting from applying conversion factors to correct for the changing beam size over the observed frequency range. The errors of the integrated intensities which have been corrected using the conversion factors between each band and band 1 derived from 13CO: 15% for band 1, 22% for band 2, 26% for band 3, 30% for band 4, 31% for band 5, and 36% for bands 6 and 7. The error for band 1 includes the calibration error and the error from the Gaussian fitting used to obtain the line parameters. The errors for the other bands also include a contribution from the scaling factors used to convert intensities to the same beam size. Table A.1. contains the original intensity values, which have not been converted to the spatial resolution of band 1, unlike the values used in Sects. 4.2 and 4.3. As the HIFI beam efficiencies were re-calculated compared to those listed in Roelfsema et al. (2012), we used the updated beam efficiencies4 in this paper. In earlier works based on this line survey the Roelfsema et al. (2012) beam efficiency was used (van der Tak et al. 2012, 2013; Nagy et al. 2013, 2015b; Choi et al. 2014). The reduced data are available for download from the Herschel Science Archive at the user provided products5.

3. Detected species

The line identification is based on the Cologne Database for Molecular Spectroscopy (CDMS, Müller et al. 2005)6 and the Jet Propulsion Laboratory (JPL, Pickett et al. 1998)7 molecular databases. We have identified about 120 lines corresponding to 29 molecules. There are two unidentified lines with a line detection threshold of about 3σ rms noise level. Several molecules show multiple transitions detected with different beam sizes. To correct for the effect of the changing beam size on the integrated intensities obtained with different beams we apply the method explained above (Sect. 2). The effect of the changing beam size is most important for CO and its isotopologues, for which transitions were detected across a wide frequency range. For many species, the detected transitions fall in the frequency range of band 1, and for those cases the changing beam size does not need to be taken into account. These species include SO, H2CO, SH+, and three of the four detected CS transitions. The parameters of the lines obtained by a Gaussian fit are shown in Table A.1.

We detected 11 transitions of CO (from J = 5–4 up to J = 16–15), 7 transitions of 13CO, 6 transitions of C18O, and 5 transitions of C17O. The CO line emission observed as a part of this line survey will be analysed in detail using PDR models by Joblin et al. (in prep.). Six transitions of H2O have been detected toward the Orion Bar CO+ peak (Choi et al., in prep.), and two transitions of HO, which were used by Choi et al. (2014) to measure the ortho-to-para ratio of water, have also been detected toward the Orion Bar. Ten transitions of formaldehyde (H2CO) have been detected in the line survey, nine of them being ortho and one of them a para transition. Several transitions of sulphur-bearing species have been detected including four transitions of H2S and three transitions of SO. Five doublets of C2H have been detected in the line survey from N = 6–5 to N = 10–9, and have been found to trace warm and dense gas (Nagy et al. 2015b), most of it at temperatures of Tkin ~ 100−150 K and densities of n(H2) ~ 105−106 cm-3, and the highest-N transitions tracing densities of n(H2) ~ 5 × 106 cm-3 and temperatures of Tkin~ 400 K. Two molecules which are expected to originate in the outer PDR layers, HCO+ and CH, have also been detected with six and five lines, respectively. The H2Cl+ line detected in this line survey was previously reported by Melnick et al. (2012) and was analysed in more detail by Neufeld et al. (2012). Two other species of the chlorine chemistry have also been detected, namely the H35Cl and the H37Cl isotopologues of hydrogen chloride. The rotational ground state line of ammonia, o-NH3 (J, K) = (1, 0) (0, 0) at 572.5 GHz detected with Odin (Larsson et al. 2003) is confirmed by our observations with a higher spatial resolution compared to the large Odin beam (about 2 arcminutes). The observed NH3 line parameters are consistent with the VLSR = 10.5 km s-1 and full width at half maximum (FWHM) of 3.3 km s-1 observed by the Odin telescope, which are also consistent with the CO J = 4–3 line parameters (Wilson et al. 2001). Other N-bearing molecules detected in the HIFI line survey include NO, CN, HCN, and HNC. The first detection of HF emission in a Galactic source has been reported by van der Tak et al. (2012) based on data from the line survey summarized in this paper, who found that HF toward the Orion Bar is most likely excited by collisions with electrons. The CH+, SH+, and CF+ ions from this line survey toward the Orion Bar have been analysed in Nagy et al. (2013). Emission of OH+ and upper limits for H2O+ and H3O+ based on data from this line survey were presented in van der Tak et al. (2013). HIFI data of C+ and 13C+ toward the Orion Bar have been presented by Ossenkopf et al. (2013). Both fine structure transitions of atomic carbon and four transitions of CS are also covered by this line survey. The strongest hyperfine transition (F = 11/2–9/2 at 589 872.2 MHz) of the CO+N = 5–4 line is tentatively detected in the line survey. This transition has also been detected with Herschel/HIFI toward another position of the OMC-1 cloud, the Orion S molecular core (Nagy 2013), which is located in the Hii region cavity around the Trapezium cluster (O’Dell et al. 2009).

Most previous line surveys of the Orion Bar were carried out at lower frequencies, such as the line survey presented in Hogerheijde et al. (1995) using different mm/sub-mm instruments between 90 and 492 GHz. This line survey was carried out toward five positions perpendicular to the Orion Bar, one of which near the ionization front close to the CO+ peak (ΔRA = 6.0′′, ΔDec = 6.8′′). The line survey of Leurini et al. (2006) was carried out between 279 and 361.5 GHz using APEX toward the most massive clump identified using H13CN by Lis & Schilke (2003) at offsets of ΔRA = 70.0′′, ΔDec = 40.0′′ from the CO+ peak. The line survey of Cuadrado et al. (2015) with the IRAM 30 m was carried out in several frequency ranges between 80 and 359 GHz toward a position near the CO+ peak (ΔRA = 3.0′′, ΔDec = 3.0′′), and mostly includes transitions of hydrocarbons. The JCMT line survey of van der Wiel et al. (2009) includes transitions in the range between 330 and 350 GHz and two more transitions around 230 GHz for a 2 × 2 arcminute region. The CO+ peak is at the edge of the region covered by the JCMT line survey, which is centred on ΔRA = 58.0′′, ΔDec = 14.0′′ relative to the CO+ peak. The line survey of Habart et al. (2010) used the SPIRE instrument of Herschel in a frequency range similar to our line survey, between 447 and 1545 GHz, but with a very low spectral resolution (0.04 cm-1 or 1.2 GHz), so that they identified only 43 lines. Their observed position is ΔRA = 33.3′′, ΔDec = 16.3′′ relative to the CO+ peak position. The spectral line survey analysed in this paper represents the dataset with the largest frequency coverage at the highest frequencies and spectral resolution (1.1 MHz). While the line survey presented in this paper contains the largest number of molecular lines detected toward the Orion Bar, a few molecules which were detected in earlier line surveys at lower frequencies, such as CH3OH, are not detected in this line survey (Leurini et al. 2006). CH3OH was detected toward the nearby Orion S region, which was demonstrated to have a strong UV-irradiation and similar abundances to the Orion Bar (Tahani et al. 2016; see Sect. 5).

thumbnail Fig. 3

Average line width versus average velocity (weighted by the signal-to-noise ratio of the lines) of all of the detected species in the HEXOS Orion Bar line survey. The error bars represent the ranges of line widths and LSR velocities for the species with multiple detected transitions including an error which is based on the errors of the parameters for the individual transitions from the Gaussian fitting. For the species with only one detected transition, the error bars are from the Gaussian fitting of the line profiles.

Open with DEXTER

In this paper we present the data corresponding to the Herschel/HIFI line survey with conclusions on the physical and chemical structure of the Orion Bar. We do not go into details about the molecules which have already been analysed in earlier works based on this line survey.

4. Results

4.1. Line properties

Molecules with similar kinematical properties may originate in the same gas component, i.e. species tracing diffuse or dense, cold or hot gas may have different kinematical properties (peak velocities and line widths). Owing to the high spectral resolution, the fitted line properties (line width, VLSR) can be used to obtain information on the kinematics/structure of the source. Even though the spectral line survey presented here corresponds to one position, the kinematical properties of the observed species may provide information on the structure of the PDR. In Fig. 3 the fitted peak LSR velocities of the observed lines are plotted against their observed line width (FWHM). In the case of molecules with multiple transitions, the average VLSR and FWHM were weighted by the signal-to-noise ratio of the lines. The error bars of these species represent the range in the observed line parameters including an error which is based on the errors of the parameters for the individual transitions from the Gaussian fitting. The data points without (upper or lower) error bars are due to the weighted average of the observed width or velocity being closer to the maximum or the minimum value of the width or velocity. For the species with only one detected transition the error bars are from the Gaussian fitting of the line profiles. For OH+ we only used its transition at 1 033 118.6 MHz as the other observed transition (at 971 805.3 MHz) is an unresolved blend of multiple hyperfine components. For most detected molecules there is a decreasing trend for the FWHM values with increasing velocities between 9.5 and 11 km s-1, as can be seen in Fig. 3.

Molecules which do not follow this trend include CH+. This was expected considering that the line-broadening of this ion is strongly related to its chemistry rather than the physical properties of its emitting region, as discussed in Nagy et al. (2013). Other species with FWHM values above the typical 23 km s-1 for the Orion Bar include OH+, HF, and H2O. The width of the lines is expected to be affected by a combination of thermal line broadening, opacity, and chemical pumping such as for CH+.

In Appendix C we show that the spread of the velocity widths is partially due to a low signal-to-noise ratio for the weak lines and partially due to opacity broadening for CO. The variation in the line position is mainly produced by the change in the pickup of signal from the Orion Ridge for different beam widths.

In summary, the VLSR-line width relation is likely affected by the combination of the changing beam size, opacity broadening, and the species originating in different temperature and density components. Lines stemming from dense molecular material have a typical velocity of 10.6 km s-1 and a line width between 1.5 and 3 km s-1. The higher line velocities shown by HF and CF+ may stem from some diffuse gas; to some degree this may also apply to CN. OH+, CO (and 13CO), and the HCl isotopologues show significant velocity contributions from the Orion Ridge. Broader line profiles are also observed for optically thick lines.

4.2. Physical parameters: Single excitation temperature

In this section, we estimate column densities for the observed molecules in LTE approximation. As not every molecule detected in this line survey has collisional excitation rates for a non-LTE analysis, we use this as a common method to obtain a column density estimate for each observed molecule.

For some of the species detected in this line survey, multiple transitions also provide an estimate on the excitation temperature by applying the rotational diagram method. We have created rotational diagrams for species with at least three observed transitions, covering a sufficient range in upper level energy. In the case of SO the transitions cover an upper level energy range between ~166 and 174 K, which is too low to result in an accurate estimate of the SO rotational temperature and column density. This method assumes optically thin lines, uniform beam filling, and a single excitation temperature to describe the observed transitions. For most observed transitions the optically thin assumption is reasonable. The single excitation temperature is a simplification as we expect the gas covered by the Herschel beam to originate in a range of temperatures. In some cases, a non-uniform beam filling may also affect the excitation of the observed molecules and the different contributions from the Orion Ridge. In the rotational diagram method the measured integrated main-beam temperatures of lines ( K km s-1) can be converted to the column densities of the molecules in the upper level (Nu) using (1)and (2)with gu the statistical weight of level u, Ntot the total column density, Q(Trot) the partition function for Trot, Eu the upper level energy, ν the frequency, and Aul the spontaneous decay rate. A linear fit to ln(Nu/gu)Eu gives Trot as the inverse of the slope, and Ntot can be derived. The rotational temperature is expected to be equal to the kinetic temperature if all levels are thermalized. We use the rotational diagram method for molecules which have more than three detected transitions. For the rest of the detected species, we derive column densities in the LTE approximation (see below) using the integrated intensity of their strongest detected transition. Equation (2) represents the Rayleigh-Jeans limit without background, which is very accurate for frequencies below 1 THz, where most observed lines in this line survey are. Figure D.1 shows the rotational diagrams of molecules observed toward the Orion Bar CO+ peak with at least three observed transitions. The derived excitation temperatures and column densities are shown in Table 2.

thumbnail Fig. 4

Derived rotational temperatures as a function of the dipole moment.

Open with DEXTER

Figure 4 shows a summary of the derived excitation temperatures as a function of the dipole moment. The highest values are consistent with or slightly above the 85 ± 30 K average kinetic temperature derived by Hogerheijde et al. (1995). The value found for CO is close to the ~150 K kinetic temperature measured near the ionization front of the Orion Bar by Batrla & Wilson (2003) and Goicoechea et al. (2011). The lower values most likely correspond to non-LTE conditions. In the non-LTE case, the radiative decay dominates the collisional excitation, resulting in excitation temperatures below the kinetic temperature. As seen in Fig. 4, the rotational temperatures found for HCO+, HCN, and CS are below the gas temperature in the regions where these species originate, as indicated by their high dipole moments. CF+, o-H2S, and C2H have lower dipole moments than HCO+, HCN, and CS, and are therefore expected to be closer to thermalized.

For molecules with no information available on their excitation (less than three observed transitions), we derived a column density in the LTE approximation for a range of excitation temperatures using Eq. (2). Table 2 includes the column densities derived in LTE approximation, assuming excitation temperatures in the range between ~20 and 150 K.

4.3. Physical parameters: Non-LTE approximation

For some of the detected molecules, collisional excitation rates are available, and therefore, non-LTE calculations with the RADEX (van der Tak et al. 2007) code can be performed. We adopt a background radiation field which was used for the non-LTE analysis of OH+ in van der Tak et al. (2013) and which is based on the results of Arab et al. (2012), who describe the continuum toward the Orion Bar with a modified blackbody distribution with a dust temperature of 50 K and a dust emissivity index of β = 1.6. For molecules with at least three detected transitions we aim to derive a best fit kinetic temperature and H2 volume density. In the case of species with at least three transitions, we compared the observed line intensities to the values calculated with RADEX. In the case of species with only two transitions such as [Ci] we compared the observed line ratio to those predicted by the model. When fitting the line intensities we assumed uniform beam filling. Most species observed toward the Orion Bar are expected to originate in both “clumps” with a lower filling factor and in the “interclump” medium with a filling factor close to one. Probing the contribution of the components with different filling factors is beyond the scope of this work, as many of the species modelled with RADEX have a low number (34) of transitions, which makes it difficult to constrain a beam filling factor8. We ran a 30 × 30 grid of RADEX models with kinetic temperatures in the range between 50 K and 300 K and H2 volume densities between 104 and 106 cm-3 to cover conditions that are expected in the clump and interclump gas. In a few cases we extended the grid to densities up to 107 cm-3. Based on previous observations most temperatures are expected to be in the range between 50 K and 300 K. As the CO+ peak is near the ionization front of the Orion Bar, some of the molecules originate in regions with temperatures of a few 100 K, such as those measured for OH (160220 K, Goicoechea et al. 2011). However, some of the detected species, such as H2O which may trace temperatures as low as 60 K, may trace larger depths into the PDR (Choi et al. 2014). Most species detected toward the Orion Bar are expected to trace the average density of 105 cm-3, but some of the species may trace densities of a few times 104 cm-3 (Simon et al. 1997) or densities around 106–107 cm-3 (e.g. Goicoechea et al. 2011). The logarithmic temperature and density steps in the grid are 0.026 dex and 0.067 dex, respectively. We also adopt a column density based on the LTE calculations, the best fit value for the molecules with rotational diagram results, and the average value for the rest. We apply a χ2 minimization to the results of the grid. When the minimum χ2 is larger than a few ×1, we change the column density and repeat the method. The error bars of the best fit H2 volume densities and kinetic temperatures given in Table 2 are determined as the range of densities and temperatures within χ2 ≤ min(χ2) + 1. Aiming to derive best fit kinetic temperatures and H2 volume densities traced by the observed molecules and transitions is a simplification as a range of physical conditions is expected to be sampled in the HIFI beam. Therefore, only beam-averaged densities and temperatures can be derived. In Sect. E we discuss each molecule in Table 2. The applied inelastic collision rates are listed in Table 1. All our tracers are expected to show their maximum abundance at visual extinctions well above AV = 0.1 where most of the hydrogen is in molecular form, which means that collisions with atomic hydrogen and electrons can be neglected.

Table 1

Inelastic collision data used for the RADEX modelling.

Figure E.1 lists the results from RADEX described above, showing the comparison of the observed and modelled line intensities. Table 2 lists the parameters used for the non-LTE analysis and the results. Previous results based on this line survey are also shown in Table 2 for C2H, HF, H2O, CH+, SH+, and OH+.

Table 2

Results derived from the single excitation temperature and non-LTE assumptions, when applicable.

The summary of the RADEX results is shown in Fig. 5, including the earlier results based on this line survey for CH+ (Nagy et al. 2013) and C2H (Nagy et al. 2015b). As seen in Fig. 5, most species trace temperatures in the range between 100–150 K and densities in the range between 105 and 106 cm-3. The best fit kinetic temperature close to 150 K for CO (138 K) is similar to those found by Goicoechea et al. (2011) and Batrla & Wilson (2003) near the surface of the Orion Bar. The observed CO+ peak position is indeed close to the PDR surface; therefore, a kinetic temperature above the average value of 85 ± 30 K (Hogerheijde et al. 1995) was expected.

thumbnail Fig. 5

Summary of the kinetic temperatures and H2 volume densities derived from the RADEX models for the molecules with at least three detected transitions. The results for CH+ from Nagy et al. (2013) and the results for C2H from Nagy et al. (2015b) are also included. The results for C2H include two components, one with temperatures of 100150 K and densities of 105–106 cm-3 representing most C2H column density, and a small fraction with a temperature of 400 K and density of 5 × 106 cm-3 required to fit the highest N transitions of C2H.

Open with DEXTER

A comparison between the rotational temperatures from the LTE and the kinetic temperatures from the non-LTE analysis is shown in Fig. 6. The warm surface of the PDR is mainly traced in the optically thick CO; the inner shielded regions produce most of the observed C17O and C18O emission. As expected for partially subthermal excitation, the kinetic temperatures are always higher than the excitation temperatures. For species with a small dipole moment, the gas densities are sufficient for a good coupling so that we find small differences between kinetic and rotation temperature. For species with higher dipole moment, the two temperatures deviate much more. HCO+ with the highest dipole moment shows the greatest difference; CS with its high dipole moment is a similar case.

We found a range of temperatures and densities, indicating a morphology of the PDR that is more complex than the clump/interclump model.

5. Discussion

In the previous sections we summarized observations of the Herschel/HIFI spectral line survey carried out toward the CO+ peak of the Orion Bar in the frequency range between 480–1250 GHz and 1410–1910 GHz. About 120 lines corresponding to 29 molecules were detected, including isotopologues. We derived excitation temperatures and column densities for molecules with multiple detected rotational transitions. The derived excitation temperatures are in the range between ~22 K and ~146 K. For several molecules we also derived physical parameters using a non-LTE analysis, making use of the available collision rates. For species with at least three detected transitions, we derived the best fit kinetic temperatures and H2 volume densities of their emitting region. As mentioned above, the modelling using the RADEX code is a simplification as it assumes a single temperature and density, while the observed line intensities are expected to originate in gas with a range of temperatures and densities.

thumbnail Fig. 6

Summary of the rotational temperatures derived in Sect. 4.2 and the kinetic temperatures derived in Sect. 4.3. For C2H we show the range of kinetic temperatures which corresponds to the highest C2H column density for the observed transitions (Nagy et al. 2015b).

Open with DEXTER

Even though the observations presented in this paper correspond to a single position, the temperatures and densities given by the RADEX fits for various molecules – especially those with the largest number of transitions – give an indication of what kind of gas component these molecules originate in. In Sect. 5.1 we discuss a few groups of species based on the best fit kinetic temperatures and H2 volume densities derived from RADEX. In Sect. 5.2 we discuss what the kinematical properties of the molecules suggest about their origin in the PDR.

Table 3

Origin of the molecules discussed in Sect. 5 based on their transitions detected in this line survey.

5.1. Temperature and density differentiation of the Orion Bar

  • Tracers of hot gas at average Orion Bar density. Very hot (500–1000 K) CH+ gas is found to trace the average Orion Bar density of 105 cm-3 (Nagy et al. 2013). Two more species detected in this line survey trace similar densities based on the RADEX models: HCO+ with a best fit kinetic temperature of 210 K and CN with a best fit kinetic temperature of 283 K.

  • Tracers of hot and dense gas. Two of the observed molecules correspond (at least partly) to very high-density hot gas. As investigated in Nagy et al. (2015b), a small fraction of C2H traces unresolved hot (400 K) and dense (5 × 106 cm-3) structures, and therefore provides additional evidence for clumpiness of the Orion Bar. Similar to the hot component of C2H, the o-H2CO transitions observed in this line survey are most consistent with hot (315 K) and dense (1.4 × 106 cm-3) gas. Though the H2CO transitions observed by Leurini et al. (2006) toward a different position are also consistent with a high kinetic temperature and density (150 K and 5 × 105 cm-3), these values are below the best fit parameters which the transitions observed in this line survey suggest. Some of the transitions covered in this line survey likely trace gas close to the ionization front, which includes the hot and dense structures revealed by C2H (Nagy et al. 2015b) and OH (Goicoechea et al. 2011) emission.

  • Species tracing warm and dense gas. Most molecules covered in this line survey, such as CO, 13CO, C18O, o-H2S, CS, and the bulk of C2H, trace temperatures of 100–150 K and densities of 105–106 cm-3. These temperatures are below what can be expected from species tracing the PDR surface, but are higher than what can be expected deep in the cloud. These species represent most of the gas detected toward the CO+ peak and are very likely associated with the pattern of high-density substructures recently revealed by ALMA in the HCO+J = 43 line (Goicoechea et al. 2016).

  • Species tracing lower temperature (<100 K) gas. For C17O, SO, and HCN the best fit kinetic temperatures are below the 100–150 K range which most species observed in this line survey trace. The lowest kinetic temperature among the species modelled in this paper (56 K) was found for SO. This may be interpreted as tracing gas which is not as directly exposed to UV irradiation as the species tracing higher temperature gas.

Table 3 includes a summary of the species discussed above with their possible origin toward the Orion Bar based on the transitions observed in this line survey.

5.2. Structure of the Orion Bar as suggested by the observed kinematical properties

Apart from the kinetic temperatures estimated with RADEX, the line properties of the species discussed in Sect. 4.1 can also be used to trace the origin of the species detected in this line survey. As mentioned in Sect. 4.1, for some species we expect a contribution of emission from the Orion Ridge region. The LSR velocity of the Orion Ridge is around 9 km s-1 (van der Tak et al. 2013); therefore, the species detected at the lowest velocities are expected to have a contribution from gas toward the Orion Ridge. These species are OH+ (as also seen in maps analysed in van der Tak et al. 2013), HCl, part of the emission in the lowest J CO, and 13CO transitions covered by this line survey, and possibly part of the SO line emission. Apart from the groups related to the dense PDR and the Orion Ridge material, some of the species are likely to trace lower density/diffuse gas, such as CF+, HF, and CN. Based on their physical properties and their expected origin close to the surface of the PDR, part of CH+ and HCO+ may also trace low density gas.

In addition to the velocity of the species, their line width may also give an indication to some detected molecules tracing the same gas component. The molecules tracing hot gas at average density include CH+; however, as discussed in Nagy et al. (2013), its width is related to its chemistry rather than the conditions of its emitting region. The other two species tracing hot gas at average density are CN and HCO+ with average FWHM widths (weighted by the signal-to-noise ratio of the lines) of 2.5 km s-1 and 2.7 km s-1, respectively. The other two species tracing hot gas, probably the surfaces of high-density substructures (Goicoechea et al. 2016), are C2H (partly) and H2CO with average line widths of 2.5 km s-1 and 2.3 km s-1, respectively. Most species detected toward the CO+ peak correspond to warm (100150 K) and dense (105–106 cm-3) gas, such as CO, 13CO, C18O, CS, and C2H. The large CO average line width of 3.9 km s-1 is mostly due to the optical depth of the CO transitions. The difference in the line width of the 13CO and C18O lines (2.5 km s-1 and 2.0 km s-1, respectively) may be due to the difference in the kinetic temperature of their emitting region, and therefore means that C18O is less exposed to UV irradiation than is 13CO. The 13CO lines are also more optically thick than the C18O lines. Despite the best fit kinetic temperature of 123 K, the CS lines have a very low line width of 1.7 km s-1. The species with the lowest kinetic temperatures include C17O, with a 2.0 km s-1 line width as measured for C18O. As expected, the SO lines are narrow (FWHM~ 1.8 km s-1); however, the HCN lines are broader with an average line width of 2.5 km s-1. Based on the above, there is no correlation between the line width of the species and the best fit kinetic temperatures found using RADEX (Fig. C.4). This may be related to the fact that physical conditions in the different regions of the PDR change rapidly, and the species trace temperature gradients and different density components around the best fit values found with RADEX.

Based on the results of the RADEX models we conclude that the species detected in this line survey allow us to trace multiple temperature and density components of a PDR with a complex morphology. The large HIFI beam (as shown in Fig. 1) captures gas from several components of the PDR, and the transitions covered in this line survey correspond to temperature and density gradients including the values found as best fits in the RADEX models. The presence of dense substructures is also required to explain the existence of some of the observed line intensities; the high-density and hot component of C2H and the high-density and high-temperature gas traced by the H2CO transitions covered in this line survey may trace the same dense, unresolved structures close to the ionization front of the PDR. Such small, dense structures were detected in HCO+ 43 at a high angular resolution with ALMA (Goicoechea et al. 2016).

5.3. Comparison with Orion South

Figure 7 shows the abundances of molecular species in the Orion Bar, derived by dividing the column densities in Table 2 by the reference N(H2) value of 2.9 × 1022 cm-2 as derived from our C17O data (Appendix D), which is basically the same as the value from Cuadrado et al. (2015). Although the column densities in Table 2 should be accurate to ~20%, we adopt an uncertainty of a factor of 2 for the abundances to allow for the uncertainty in N(H2). No values are plotted for DCN, H2CS, and SO2, which have been detected in APEX line surveys of the Orion Bar (Leurini et al. 2006; Parise et al. 2009), but not in our HIFI survey. Since the APEX line surveys were carried out at another position in the Bar and at higher angular resolution, including their results in Fig. 7 would add an extra uncertainty.

thumbnail Fig. 7

Comparison of molecular abundances in the Orion Bar (blue squares) and those in Orion South (red triangles).

Open with DEXTER

Overplotted in Fig. 7 are the molecular abundances derived from the HIFI line survey of Orion South (Tahani et al. 2016). This object is a massive protostellar envelope under strong external irradiation by the Trapezium stars, also affected by internal irradiation of outflow cavity walls. The Orion Bar and Orion South are compared to determine the role of the different excitation mechanisms since both objects are at the same distance and since both surveys were carried out with the same instrument. Further comparison of abundances in Orion South with those for the various components of the Orion-KL (Crockett et al. 2014) massive star-forming region is presented in Tahani et al. (2016).

Many species are seen to have similar abundances in the two sources (Fig. 7), including classic PDR tracers such as C2H (Nagy et al. 2015b; Cuadrado et al. 2015). This chemical uniformity of Orion, observed previously on larger scales (Ungerechts et al. 1997), is somewhat surprising as PDR chemistry is expected to favour small radicals and ions as a result of UV irradiation, whereas in protostellar envelopes, grain surface chemistry is expected to lead to higher abundances of more highly saturated species. The higher abundances of H2CO and isotopic CS in Orion South compared with the Bar seem to follow this expectation, but the higher NH3 abundance is in disagreement with this picture. We conclude that the chemistry of Orion South is actually more PDR-like and not like the chemistry of other protostellar envelopes such as AFGL 2591 (Kaźmierczak-Barthel et al. 2015), as also suggested by the lack of “hot core” signatures in Orion South (Tahani et al. 2016).

6. Summary

We have carried out a line survey with Herschel/HIFI toward the CO+ peak of the Orion Bar in the frequency range between 4801250 GHz and 14101910 GHz. About 120 lines corresponding to 29 molecules were detected, including isotopologues.

The average FWHM line width for each detected species as a function of their average LSR velocity shows a decreasing trend for the line widths with velocity. This trend may be due to a combination of the changing beam size, the effect of opacity, and different species tracing different temperature and density components toward the observed position.

For species with at least three transitions we have created rotational diagrams to get a column density and an excitation temperature estimate. The excitation temperatures from the rotational diagrams are in the range between ~22 and ~146 K and the column densities are in the range between 1.8 × 1012 cm-2 and 4.5 × 1017 cm-2.

For species with at least three transitions and available collisional excitation rates we also performed a non-LTE analysis with RADEX and derived best fit kinetic temperatures and H2 volume densities. Most detected species have temperatures in the range between 100 and 150 K and densities in the range between 105 and 106 cm-3. The exceptions include SO, which has the lowest kinetic temperature among the species detected in the line survey, 56 K, probably tracing gas deeper into the PDR. Another exception is H2CO, which – based on the transitions observed in this line survey – involves a high-density (~1.4 × 106 cm-3) component but also a higher temperature (315 K) than the other studied species.

The physical conditions in the Orion Bar encompass a range of temperatures and densities, indicating that the morphology of the PDR is more complex than the clump/interclump model, which is consistent with the complex morphology revealed by the recent ALMA observations (Goicoechea et al. 2016). Furthermore, we warn against classifying a source and its chemistry based on the mere detection of one or a few species. The comparison made in Sect. 5.3 with Orion S shows that such simple comparisons may be rather misleading.


1

Herschel observations of EXtra-Ordinary Sources.

2

The James Clerk Maxwell Telescope is operated by The Joint Astronomy Centre on behalf of the Science and Technology Facilities Council of the United Kingdom, the Netherlands Organisation for Scientific Research, and the National Research Council of Canada.

4

The updated beam efficiencies are listed in the technical note “Measured beam efficiencies on Mars (revision v1.1, 1 October 2014)” http://herschel.esac.esa.int/twiki/bin/view/Public/HifiCalibrationWeb.

8

For the species with the highest number of transitions (CO and its isotopologues) the changing beam size introduces an uncertainty from the varying Orion ridge pickup that would make a two-component fit questionable.

9

The velocity difference between the three hyperfine components of the 13C+ line may indicate some problem with the accuracy of the published rest frequencies.

Acknowledgments

We thank the referee for the useful comments which helped to improve the paper. HIFI has been designed and built by a consortium of institutes and university departments from across Europe, Canada, and the US under the leadership of SRON Netherlands Institute for Space Research, Groningen, The Netherlands, with major contributions from Germany, France, and the US. Consortium members are: Canada: CSA, U.Waterloo; France: CESR, LAB, LERMA, IRAM; Germany: KOSMA, MPIfR, MPS; Ireland, NUIMaynooth; Italy: ASI, IFSI-INAF, Arcetri-INAF; Netherlands: SRON, TUD; Poland: CAMK, CBK; Spain: Observatorio Astronomico Nacional (IGN), Centro de Astrobiología (CSIC-INTA); Sweden: Chalmers University of Technology – MC2, RSS and GARD, Onsala Space Observatory, Swedish National Space Board, Stockholm University – Stockholm Observatory; Switzerland: ETH Zürich, FHNW; USA: Caltech, JPL, NHSC. HIPE is a joint development by the Herschel Science Ground Segment Consortium, consisting of ESA, the NASA Herschel Science Center, and the HIFI, PACS, and SPIRE consortia. V.O. was supported by Deutsche Forschungsgemeinschaft (DFG) via the collaborative research grant SFB 956, project C1. C.J. and M.G. acknowledge funding from the French space agency CNES. Y.C. was supported by the Basic Science Research Program through the National Research Foundation of Korea (NRF) (grant No. NRF-2015R1A2A2A01004769) and the Korea Astronomy and Space Science Institute under the R&D program (Project No. 2015-1-320-18) supervised by the Ministry of Science, ICT and Future Planning.

References

Appendix A: Observed lines and their parameters

Table A.1

Gaussian fit and spectroscopic parameters of the lines identified in the spectral line survey. The Gaussian fit parameters without errors have been fixed in the fit, for example in the case of blended lines.

Appendix B: The observed data

thumbnail Fig. B.1

From top to bottom: bands 1a, 1b, 2a, and 2b.

Open with DEXTER

thumbnail Fig. B.2

From top to bottom: bands 3a, 3b, 4a, and 4b.

Open with DEXTER

thumbnail Fig. B.3

From top to bottom: bands 5a, 5b, 6a, and 6b. Bands 5 and 6 were smoothed from the original velocity resolutions of ~0.12 km s-1 and ~0.09 km s-1 to the ~1 km s-1 and ~1.5 km s-1 channels, respectively.

Open with DEXTER

thumbnail Fig. B.4

From top to bottom: bands 7a, 7b. The spectra were smoothed from the original velocity resolution of ~0.08 km s-1 to the ~1.4 km s-1 channel.

Open with DEXTER

Appendix C: Line properties

To test whether some of the fitted line parameters are the result of low signal-to-noise, we plotted the fitted velocities as a function of the peak intensities in Fig. C.1. Some of the parameters of lower intensity lines may indeed be affected by the low signal-to-noise as there is no trend showing that the lowest intensity lines peak at the lowest or highest velocities of the observed range, the effect of line parameters due to low signal-to-noise on the VLSR-line width relation is small9.

Figure C.2 shows the line width–velocity plot for the species with the highest number of detected transitions in the line survey. The general trend that the line widths decrease with the velocity as seen in Fig. 3 is also seen in the case of individual transitions. The CS line at a velocity of ~10.95 km s-1 is a marginal (3σ) detection, and is also close to the edge of bands 1b and 2a. Some of the variations in the observed peak velocities may be due to

the changing beam size. The lower frequency transitions capture more emission of the Orion Ridge region, which is dominated by lower velocity gas with VLSR< 9 km s-1 (van der Tak et al. 2013) compared to that toward the Orion Bar. To test whether the changing beam size has a dominant effect on the trend seen in the VLSR–line width relation, we plot the observed LSR velocities of the molecules with the most detected transitions as a function of the beam size (Fig. C.3). For some of the plotted molecules there indeed seems to be a decrease in the velocity with increasing beam size as a result of capturing more emission from the Orion Ridge region. These molecules are CO, 13CO, C18O, and CS. The low J CO transitions are dominated by emission from the Orion Ridge. The low J transitions of 13CO also show a significant contribution from the Orion Ridge. For C17O and HCO+ the velocity does not depend on the beam-size based on the observed transitions as these molecules are not excited in the Orion Ridge.

thumbnail Fig. C.1

Peak intensities of the species shown in Fig. 3 as a function VLSR. For some of the transitions in this figure the error bars are smaller than the symbol sizes.

Open with DEXTER

thumbnail Fig. C.2

Line width versus the LSR velocity of the species with most detected transitions. For most transitions in this figure the error bars are smaller than the symbol sizes.

Open with DEXTER

thumbnail Fig. C.3

LSR velocity versus the beam size of the species with most detected transitions. For most transitions in this figure the error bars are smaller than the symbol sizes.

Open with DEXTER

thumbnail Fig. C.4

Average line widths (weighted by the signal-to-noise ratio of the transitions) of species with kinetic temperature estimates. For C2H the lower kinetic temperature is plotted as it represents most C2H emission for the observed C2H transitions.

Open with DEXTER

Appendix D: Rotational diagrams

For the five C17O transitions the rotational diagram gives an excitation temperature estimate of 65 ± 3 K. The six detected C18O transitions suggest a slightly higher rotational temperature: 71 ± 2 K. The C18O column density of (6.2 ± 0.2) × 1015 cm-2 is consistent with the value derived from the C17O column density ((1.8 ± 0.1) × 1015 cm-2) and an abundance ratio of 18O/17O = 3.2 (Wilson & Rood 1994).

The seven 13CO transitions correspond to a rotational temperature higher than that of the C17O and C18O lines: 92 ± 3 K. Its column density is (4.7 ± 0.2) × 1016 cm-2. It is expected that at least some of the 11 CO transitions are optically thick; therefore, we include a correction for the opacity for CO, as explained in Nagy et al. (2015a), and evaluated Eq. (1) in Nagy et al. (2015a) for a set of CO column densities, excitation temperatures, and opacities. We assumed a uniform beam filling, as for the rotational diagrams without opacity correction. This method results in an excitation temperature of 137 K, and a best fit CO column density of 4.5 × 1017 cm-2. This fit suggests that the CO 5–4, ..., 11–10 transitions are optically thick and have opacities in the range between 1.2 and 2.4. The CO column density of 4.5 × 1017 cm-2 is about a factor of 10 below the 3.2 × 1018 cm-2 value suggested by the C17O column density and the abundance ratios of 16O/18O = 560 and 18O/17O = 3.2 (Wilson & Rood 1994). The difference between the CO column density derived from the C17O column density and the value given by the CO rotation diagram is consistent with the expected high optical depth of 12CO emission. Non-uniform beam filling with different emitting region sizes for CO and C17O could also contribute to the differences.

A CO column density of 3.2 × 1018 cm-2 (based on the C17O column density and the abundance ratios quoted above) and a CO abundance of 1.1 × 10-4 (Johnstone et al. 2003) suggest an H2 column density of 2.9 × 1022 cm-2. The difference between this value and the 2.1 × 1022 cm-2 derived in Nagy et al. (2015b) is due to re-calculating the C17O column density using the line intensities scaled by the updated HIFI main beam efficiencies (see Sect. 2). An H2 column density of 2.9 × 1022 cm-2 is consistent with the 3 × 1022 cm-2 value found by Cuadrado et al. (2015) toward a position close to the CO+ peak. van der Wiel et al. (2009) measured an H2 column density of 1023 cm-2 toward a position further away from the CO+ peak.

The nine detected o-H2CO transitions are consistent with a rotational temperature of 146 ± 48 K which is consistent with the rotational temperature derived from the CO transitions. The corresponding column density is (1.5 ± 0.5) × 1012 cm-2.

For HCO+ the rotational diagram suggests an excitation temperature of 46 ± 3 K, and a column density of (5.4 ± 0.3) × 1012 cm-2. This column density is consistent with that derived by Jansen et al. (1995) toward some of their observed positions.

The rotational diagram of the four detected CS transitions results in an excitation temperature of 39 ± 2 K and a column density of (1.0 ± 0.1) × 1013 cm-2.

An even lower value, 27 ± 10 K is estimated for o-H2S. A similarly low excitation temperature, 22 ± 4 K, is suggested for CF+, when creating a rotational diagram using the 5–4 transition observed in this line survey and the 3–2, 2–1, and 1–0 transitions previously observed by Neufeld et al. (2006). The corresponding column density is (1.8 ± 0.3) × 1012 cm-2. The

difference between the value quoted in Nagy et al. (2013) and in this paper is due to the re-reduction of the data with a more recent HIPE version, which affects the intensity of the low signal-to-noise CF+ 5–4 transition. The rotation temperature value quoted in Nagy et al. (2013) was 32 K, and the column density 2.1 × 1012 cm-2.

The three detected HCN transitions are also consistent with a low excitation temperature: 24 ± 2 K, and a column density of (1.1 ± 0.1) × 1013 cm-2. The low column density suggested by the rotational diagram compared to that estimated using RADEX (see Sect. 4.3) suggests non-LTE conditions, which is expected as a result of the high critical densities of the HCN transitions, for example 108 cm-3 for the 6–5 transition.

The seven detected CN transitions correspond to three rotational transitions. These transitions suggest a rotational temperature of 27 ± 8 K and a CN column density of (4.7 ± 1.3) × 1013 cm-2. The [CN]/[HCN] column density ratio based on the rotational diagrams is ~4 which is similar to previous measurements (e.g. Jansen et al. 1995).

Six H2O transitions were detected in this line survey, with three ortho and three para transitions. The excitation temperatures have large errors in both cases, especially for p-H2O, but the o/p ratio based on the two column densities is ~0.14, which is consistent with the non-LTE estimate of Choi et al. (2014). The large scatter seen in the H2O rotational diagrams confirms that non-LTE effects are important in the excitation of H2O.

thumbnail Fig. D.1

Rotational diagrams of molecules with multiple detected transitions. The red triangles for CO have been corrected for the effect of opacity.

Open with DEXTER

Appendix E: RADEX results

To model the CO line intensities, we adopt a CO column density based on the C17O rotational diagram result and isotopic ratios of C18O/C17O of 3.2 and CO/C18O of 560 (Wilson & Rood 1994). With the adopted column density, the CO line intensities are consistent with an H2 volume density of 1.6× 105 cm-3 and a kinetic temperature of 138 K. This temperature is similar to the excitation temperature derived by the rotational diagram method within the given error bars. This temperature is also close to the value of ~150 K found by Batrla & Wilson (2003) near the ionization front of the Orion Bar.

The H2 volume density of 3.4 × 105 cm-3 corresponding to the 13CO transitions is similar to the density derived from the CO transitions. The best fit kinetic temperature of 115 K is lower than the kinetic temperature implied by the CO transitions, but higher than the 13CO excitation temperature of 93 ± 3 K resulted from the rotational diagram. The 13CO column density in the fit was slightly higher than the value given by the rotational diagram, in order to better fit the intensity of the lowest-J13CO transition covered by this line survey.

The observed C18O transitions suggest a similar density to that of CO and 13CO, with a best fit value of 3.4 × 105 cm-3. The best fit kinetic temperature of 91 K is below the value given for the 13CO fit, but above the excitation temperature of 71 ± 2 K given by the rotational diagram fit.

The five detected C17O transitions are consistent with a kinetic temperature of 72 K which is close to the average kinetic temperature of 85 K (Hogerheijde et al. 1995). The corresponding H2 volume density is 8.6 × 105 cm-3. The difference between the best fit H2 volume density of C17O and those given by the other CO isotopologues is likely due to the lower number of C17O transitions. The H2 volume densities derived from the other CO isotopologues fall in the 1σ limit range derived for the C17O fit.

The six detected HCO+ transitions are most consistent with a kinetic temperature of 210 K, which is much higher than the excitation temperature of 46 ± 3 K given by the rotational diagram method. The best fit H2 volume density given by the HCO+ transitions is the average H2 volume density toward the Orion Bar, 1.0× 105 cm-3.

The four detected CS transitions suggest a kinetic temperature of 123 K, much above the 39 ± 2 K excitation temperature given by the rotational diagram method. The best fit H2 volume density is 8.6 × 105 cm-3.

To get a better fit for the observed o-H2CO transitions we extended the search range of temperatures and densities to 50–400 K and 104–107 cm-3, respectively. The nine detected o-H2CO transitions are most consistent with a very high-temperature and high-density component, 315 K and 1.4 × 106 cm-3. The relative intensities of the transitions are not well fitted by the RADEX model.

We use the three detected o–H2S transitions to constrain the physical parameters of the H2S emitting region. The best fit H2 volume density of 1.4 × 105 cm-3 and kinetic temperature of 123 K are similar to those derived from the RADEX models for CO and 13CO.

The seven detected CN lines which correspond to three rotational transitions suggest a best fit kinetic temperature of 283 K, which is similar to the value obtained for o-H2CO. The best fit H2 volume density of 1.8 × 105 cm-3 is close to the values derived from the CO isotopologues and HCO+. The chemically related molecule HCN requires a component with a high density ( cm-3) but a moderate temperature ( K). The H2 volume density above 106 cm-3 is due to the fact that we extended the upper limit of the search range in the case of HCN to get an idea of the error bars of its best fit density. The [CN]/[HCN] column density ratio in these models is ~3, which is close to what was measured earlier toward the Orion Bar (e.g. Jansen et al. 1995).

Among the molecules discussed in this paper, SO is consistent with the lowest kinetic temperature, K. This is similar to the 60 K used by Choi et al. (2014) when deriving parameters for the observed HO transitions. The best fit H2 volume density is cm-3, similar to the values derived from the 13CO and C18O transitions. As no rotational diagram result is available as a first guess for the SO column density, we ran models with several different column densities in the range given by the excitation temperatures used in the LTE calculation. The temperature and density shown above correspond to the best fit.

The four observed NO transitions which correspond to two rotational transitions suggest an H2 volume density of 2.9 × 105 cm-3, close to that derived for the CO isotopologues. The kinetic temperature is not well constrained by the fit, and is in the range between 150 and 300 K.

The o-NH3 (J, K) = (1, 0) (0, 0) transition was observed with a 2 beam using Odin by Larsson et al. (2003), who conclude an NH3 abundance of NH3/H2 = 5 × 10-9. This abundance and an N(H2) = 2.2 × 1022 cm-2 is equivalent to N(NH3) = 1.1 × 1014 cm-2. Assuming a kinetic temperature of ~145 K (Wilson et al. 2001), the NH3 line emission observed with HIFI originates in a ~ 2 × 105 cm-3 density gas component.

Assuming typical Orion Bar conditions (Tkin = 100 K and n(H2) = 105 cm-3) for HNC, the intensity of the observed (J = 6–5) HNC transition can be reproduced by a column density of 5 × 1012 cm-2, which is slightly above the LTE estimate.

Two transitions of [Ci] are covered by the HIFI line survey. Based on the comparison of the observed ratio of the transitions to those calculated with RADEX the best fit kinetic temperature is 86 K and the best fit H2 volume density is 8.6 × 104 cm-3 when assuming a [Ci] column density of 1018 cm-2. These parameters are not well constrained, as is shown by the large error bars in Table 2.

Five transitions of CH, which correspond to two rotational transitions, were detected in this line survey. The available collision rates (Marinakis et al. 2015) did not take into account the hyperfine structure of CH; therefore, the temperature and density of the gas traced by CH can only be constrained by two rotational transitions. We searched for the temperatures and densities which fit the observed line intensity ratio best. The obtained temperature and density cover a broad range of parameters with temperatures of K and densities of cm-3. Given that these parameters are not well constrained by the two observed transitions, we do not consider this molecule in Sect. 5.

We used the CF+ 5–4 line intensity measured as a part of this line survey and the CF+ 1–0, 2–1, and 3–2 line intensities observed by Neufeld et al. (2006) toward a position (05h35m22.8s, −5°25′01′′) close to the CO+ peak. Using the column density obtained from the rotational diagram the H2 volume density given by the fit is consistent with cm-3, like the values traced by the CO isotopologues and HCO+. The kinetic temperature of K is not well constrained by the fit, but provides a lower limit on the kinetic temperature of the gas traced by CF+.

thumbnail Fig. E.1

RADEX fits of molecules with multiple detected transitions. The blue circles are the observed line intensities and the red diamonds are the line intensities predicted using RADEX.

Open with DEXTER

All Tables

Table 1

Inelastic collision data used for the RADEX modelling.

Table 2

Results derived from the single excitation temperature and non-LTE assumptions, when applicable.

Table 3

Origin of the molecules discussed in Sect. 5 based on their transitions detected in this line survey.

Table A.1

Gaussian fit and spectroscopic parameters of the lines identified in the spectral line survey. The Gaussian fit parameters without errors have been fixed in the fit, for example in the case of blended lines.

All Figures

thumbnail Fig. 1

13CO 3–2 integrated intensity (in K km s-1 units) of the Orion Bar observed with the JCMT. The circles at the position of the CO+ peak (αJ2000 = 05h35m20.6s, δJ2000 = −05°25′14′′) indicate the beam sizes corresponding to the smallest (band 7b, ~11′′) and largest (band 1a, ~44′′) beam sizes of the data presented in this line survey. The white symbols show the position of the Trapezium stars. The white contours show the distribution of the integrated intensity (in K km s-1 units) of HCN 43 measured with the JCMT (van der Wiel et al. 2009). The contours are 25%, 40%, 60%, and 80% of the peak HCN intensity of ~24 K km s-1.

Open with DEXTER
In the text
thumbnail Fig. 2

Full baseline-subtracted spectrum toward the CO+ peak of the Orion Bar. Bands 5, 6, and 7 were smoothed from the original velocity resolutions of ~0.12 km s-1, ~0.09 km s-1, and ~0.08 km s-1 to the ~1 km s-1, ~1.5 km s-1, and ~1.4 km s-1 channels, respectively.

Open with DEXTER
In the text
thumbnail Fig. 3

Average line width versus average velocity (weighted by the signal-to-noise ratio of the lines) of all of the detected species in the HEXOS Orion Bar line survey. The error bars represent the ranges of line widths and LSR velocities for the species with multiple detected transitions including an error which is based on the errors of the parameters for the individual transitions from the Gaussian fitting. For the species with only one detected transition, the error bars are from the Gaussian fitting of the line profiles.

Open with DEXTER
In the text
thumbnail Fig. 4

Derived rotational temperatures as a function of the dipole moment.

Open with DEXTER
In the text
thumbnail Fig. 5

Summary of the kinetic temperatures and H2 volume densities derived from the RADEX models for the molecules with at least three detected transitions. The results for CH+ from Nagy et al. (2013) and the results for C2H from Nagy et al. (2015b) are also included. The results for C2H include two components, one with temperatures of 100150 K and densities of 105–106 cm-3 representing most C2H column density, and a small fraction with a temperature of 400 K and density of 5 × 106 cm-3 required to fit the highest N transitions of C2H.

Open with DEXTER
In the text
thumbnail Fig. 6

Summary of the rotational temperatures derived in Sect. 4.2 and the kinetic temperatures derived in Sect. 4.3. For C2H we show the range of kinetic temperatures which corresponds to the highest C2H column density for the observed transitions (Nagy et al. 2015b).

Open with DEXTER
In the text
thumbnail Fig. 7

Comparison of molecular abundances in the Orion Bar (blue squares) and those in Orion South (red triangles).

Open with DEXTER
In the text
thumbnail Fig. B.1

From top to bottom: bands 1a, 1b, 2a, and 2b.

Open with DEXTER
In the text
thumbnail Fig. B.2

From top to bottom: bands 3a, 3b, 4a, and 4b.

Open with DEXTER
In the text
thumbnail Fig. B.3

From top to bottom: bands 5a, 5b, 6a, and 6b. Bands 5 and 6 were smoothed from the original velocity resolutions of ~0.12 km s-1 and ~0.09 km s-1 to the ~1 km s-1 and ~1.5 km s-1 channels, respectively.

Open with DEXTER
In the text
thumbnail Fig. B.4

From top to bottom: bands 7a, 7b. The spectra were smoothed from the original velocity resolution of ~0.08 km s-1 to the ~1.4 km s-1 channel.

Open with DEXTER
In the text
thumbnail Fig. C.1

Peak intensities of the species shown in Fig. 3 as a function VLSR. For some of the transitions in this figure the error bars are smaller than the symbol sizes.

Open with DEXTER
In the text
thumbnail Fig. C.2

Line width versus the LSR velocity of the species with most detected transitions. For most transitions in this figure the error bars are smaller than the symbol sizes.

Open with DEXTER
In the text
thumbnail Fig. C.3

LSR velocity versus the beam size of the species with most detected transitions. For most transitions in this figure the error bars are smaller than the symbol sizes.

Open with DEXTER
In the text
thumbnail Fig. C.4

Average line widths (weighted by the signal-to-noise ratio of the transitions) of species with kinetic temperature estimates. For C2H the lower kinetic temperature is plotted as it represents most C2H emission for the observed C2H transitions.

Open with DEXTER
In the text
thumbnail Fig. D.1

Rotational diagrams of molecules with multiple detected transitions. The red triangles for CO have been corrected for the effect of opacity.

Open with DEXTER
In the text
thumbnail Fig. E.1

RADEX fits of molecules with multiple detected transitions. The blue circles are the observed line intensities and the red diamonds are the line intensities predicted using RADEX.

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.