EDP Sciences
Free Access
Issue
A&A
Volume 584, December 2015
Article Number A28
Number of page(s) 84
Section Interstellar and circumstellar matter
DOI https://doi.org/10.1051/0004-6361/201526222
Published online 18 November 2015

Online material

Appendix A: Baseline subtraction

As noted in Sect. 2, the spectra observed by APEX exhibit unstable quasi-sinusoidal baselines (see Vassilev et al. 2008). When investigating the spectra, we found that polynomial and/or sinusoidal baselines were not giving desirable results. To account for this, we developed a relatively advanced baseline-fitting algorithm, computing running-mean baselines, which is described here.

In principle, the fitted baseline is a boxcar smoothing of the spectrum (where line channels have been removed) with weights increasing towards the centre of the box: for each channel in the spectrum, the algorithm calculates the weighted mean of the line-free channels in a box around the channel. The weights have a Gaussian distribution around the central channel of the box. The width of this Gaussian must be adjusted so that all irregularities in the baseline are removed but no real spectral features are removed. In addition to the quasi-sinusoidal baselines, strong atmospheric lines arising from the different elevations of the position-switching on- and off-positions made it necessary to introduce discontinuities in the fitted baseline. By this method, a piecewise smooth but in general non-analytical baseline can be fitted to the data (see examples in Fig. A.1). Some artefacts of the strongest of these atmospheric lines still remain after the baseline subtraction and appear as broad spectral line features in the resulting spectra. In this dataset, they are easy to distinguish from the source lines due to the large difference in line widths.

thumbnail Fig. A.1

Examples of spectral data before they were treated by the running-mean baseline-fitting routine (blue) and the fitted baseline that was subsequently subtracted (red). The right spectrum has a discontinuity in the fitted baseline that is due to an atmospheric line.

Open with DEXTER

Appendix B: Observed spectral line parameters

The integrated intensities and other parameters of the detected spectral lines are listed in the tables of this appendix. The values

of vLSR, Δv, and Tpeak are results of Gaussian fits. In the cases where no values are given, Gaussian fits could not be performed to that line because of irregular line shapes.

Table B.1

Observed line strengths towards IRS7B, both for the 1.3 mm unbiased line survey and for the IRS7B part of the source survey, which also covered the 0.9 mm and 0.8 mm windows.

Table B.2

Observed line strengths towards CrA-46.

Table B.3

Observed line strengths towards CrA-3.

Table B.4

Observed line strengths towards CrA-5.

Table B.5

Observed line strengths towards LS-RCrA1.

Table B.6

Observed line strengths towards Haas4.

Table B.7

Observed line strengths towards IRS2.

Table B.8

Observed line strengths towards IRS5A.

Table B.9

Observed line strengths towards IRS5N.

Table B.10

Observed line strengths towards IRS1.

Table B.11

Observed line strengths towards IRS7A.

Table B.12

Observed line strengths towards CrA-24, total flux (both source and outflow components).

Table B.13

Observed line strengths towards CrA-24, on-source velocity component.

Table B.14

Observed line strengths towards CrA-24, outflow velocity component.

Table B.15

Observed line strengths towards SMM 2, total flux (both source and outflow components).

Table B.16

Observed line strengths towards SMM 2, on-source velocity component.

Table B.17

Observed line strengths towards SMM 2, outflow velocity component.

Table B.18

Observed line strengths towards CXO42.

Table B.19

Observed line strengths towards CrA-44.

Table B.20

Observed line strengths towards CrA-33.

Table B.21

Observed line strengths towards VV CrA.

Table B.22

Observed line strengths towards CrA-37.

Appendix C: Complete IRS7B spectrum at 217–245.5 GHz

All spectra in this and the next section are cut off at 0.5 K to also show the weaker spectral lines.

thumbnail Fig. C.1

Full spectrum of IRS7B, smoothed by a factor 8, corresponding to a channel width of 1.2–1.4 km s-1.

Open with DEXTER

thumbnail Fig. D.1

0.9 mm spectrum of IRS7B, smoothed by a factor 8, corresponding to a channel width of 0.6 km s-1.

Open with DEXTER

thumbnail Fig. D.2

0.8 mm spectrum of IRS7B, smoothed by a factor 8, corresponding to a channel width of 0.5 km s-1. Strong atmospheric lines are visible at 364.1 GHz, 364.4 GHz, and 365.3 GHz.

Open with DEXTER

thumbnail Fig. D.3

1.4 mm spectrum of CrA-46, smoothed by a factor 8, corresponding to a channel width of 0.8 km s-1.

Open with DEXTER

thumbnail Fig. D.4

1.4 mm spectrum of CrA-3, smoothed by a factor 8, corresponding to a channel width of 0.8 km s-1.

Open with DEXTER

thumbnail Fig. D.5

0.9 mm spectrum of CrA-3, smoothed by a factor 8, corresponding to a channel width of 0.6 km s-1.

Open with DEXTER

thumbnail Fig. D.6

1.4 mm spectrum of CrA-5, smoothed by a factor 8, corresponding to a channel width of 0.8 km s-1.

Open with DEXTER

thumbnail Fig. D.7

0.9 mm spectrum of CrA-5, smoothed by a factor 8, corresponding to a channel width of 0.6 km s-1.

Open with DEXTER

thumbnail Fig. D.8

0.8 mm spectrum of CrA-5, smoothed by a factor 8, corresponding to a channel width of 0.5 km s-1.

Open with DEXTER

thumbnail Fig. D.9

1.4 mm spectrum of LS-RCrA1, smoothed by a factor 8, corresponding to a channel width of 0.8 km s-1.

Open with DEXTER

thumbnail Fig. D.10

0.9 mm spectrum of LS-RCrA1, smoothed by a factor 8, corresponding to a channel width of 0.6 km s-1.

Open with DEXTER

thumbnail Fig. D.11

1.4 mm spectrum of Haas4, smoothed by a factor 8, corresponding to a channel width of 0.8 km s-1.

Open with DEXTER

thumbnail Fig. D.12

0.9 mm spectrum of Haas4, smoothed by a factor 8, corresponding to a channel width of 0.6 km s-1.

Open with DEXTER

thumbnail Fig. D.13

1.4 mm spectrum of IRS2, smoothed by a factor 8, corresponding to a channel width of 0.8 km s-1.

Open with DEXTER

thumbnail Fig. D.14

0.9 mm spectrum of IRS2, smoothed by a factor 8, corresponding to a channel width of 0.6 km s-1.

Open with DEXTER

thumbnail Fig. D.15

1.4 mm spectrum of IRS5A, smoothed by a factor 8, corresponding to a channel width of 0.8 km s-1.

Open with DEXTER

thumbnail Fig. D.16

0.9 mm spectrum of IRS5A, smoothed by a factor 8, corresponding to a channel width of 0.6 km s-1.

Open with DEXTER

thumbnail Fig. D.17

0.8 mm spectrum of IRS5A, smoothed by a factor 8, corresponding to a channel width of 0.5 km s-1.

Open with DEXTER

thumbnail Fig. D.18

1.4 mm spectrum of IRS5N, smoothed by a factor 8, corresponding to a channel width of 0.8 km s-1.

Open with DEXTER

thumbnail Fig. D.19

0.9 mm spectrum of IRS5N, smoothed by a factor 8, corresponding to a channel width of 0.6 km s-1.

Open with DEXTER

thumbnail Fig. D.20

0.8 mm spectrum of IRS5N, smoothed by a factor 8, corresponding to a channel width of 0.5 km s-1.

Open with DEXTER

thumbnail Fig. D.21

1.4 mm spectrum of IRS1, smoothed by a factor 8, corresponding to a channel width of 0.8 km s-1.

Open with DEXTER

thumbnail Fig. D.22

0.9 mm spectrum of IRS1, smoothed by a factor 8, corresponding to a channel width of 0.6 km s-1.

Open with DEXTER

thumbnail Fig. D.23

0.8 mm spectrum of IRS1, smoothed by a factor 8, corresponding to a channel width of 0.5 km s-1.

Open with DEXTER

thumbnail Fig. D.24

0.9 mm spectrum of IRS7A, smoothed by a factor 8, corresponding to a channel width of 0.6 km s-1.

Open with DEXTER

thumbnail Fig. D.25

0.8 mm spectrum of IRS7A, smoothed by a factor 8, corresponding to a channel width of 0.5 km s-1. Strong atmospheric lines are visible at 364.1 GHz, 364.4 GHz, and 365.3 GHz.

Open with DEXTER

thumbnail Fig. D.26

1.4 mm spectrum of CrA-24, smoothed by a factor 8, corresponding to a channel width of 0.8 km s-1. Atmospheric line at 217.7 GHz.

Open with DEXTER

thumbnail Fig. D.27

0.9 mm spectrum of CrA-24, smoothed by a factor 8, corresponding to a channel width of 0.6 km s-1.

Open with DEXTER

thumbnail Fig. D.28

0.8 mm spectrum of CrA-24, smoothed by a factor 8, corresponding to a channel width of 0.5 km s-1.

Open with DEXTER

thumbnail Fig. D.29

1.4 mm spectrum of SMM 2, smoothed by a factor 8, corresponding to a channel width of 0.8 km s-1.

Open with DEXTER

thumbnail Fig. D.30

0.9 mm spectrum of SMM 2, smoothed by a factor 8, corresponding to a channel width of 0.6 km s-1.

Open with DEXTER

thumbnail Fig. D.31

0.8 mm spectrum of SMM 2, smoothed by a factor 8, corresponding to a channel width of 0.5 km s-1.

Open with DEXTER

thumbnail Fig. D.32

1.4 mm spectrum of CXO42, smoothed by a factor 8, corresponding to a channel width of 0.8 km s-1.

Open with DEXTER

thumbnail Fig. D.33

0.9 mm spectrum of CXO42, smoothed by a factor 8, corresponding to a channel width of 0.6 km s-1.

Open with DEXTER

thumbnail Fig. D.34

1.4 mm spectrum of CrA-44, smoothed by a factor 8, corresponding to a channel width of 0.8 km s-1.

Open with DEXTER

thumbnail Fig. D.35

0.9 mm spectrum of CrA-44, smoothed by a factor 8, corresponding to a channel width of 0.6 km s-1.

Open with DEXTER

thumbnail Fig. D.36

1.4 mm spectrum of CrA-33, smoothed by a factor 8, corresponding to a channel width of 0.8 km s-1.

Open with DEXTER

thumbnail Fig. D.37

0.9 mm spectrum of CrA-33, smoothed by a factor 8, corresponding to a channel width of 0.6 km s-1.

Open with DEXTER

thumbnail Fig. D.38

1.4 mm spectrum of VV CrA, smoothed by a factor 8, corresponding to a channel width of 0.8 km s-1.

Open with DEXTER

thumbnail Fig. D.39

0.9 mm spectrum of VV CrA, smoothed by a factor 8, corresponding to a channel width of 0.6 km s-1.

Open with DEXTER

thumbnail Fig. D.40

1.4 mm spectrum of CrA-37, smoothed by a factor 8, corresponding to a channel width of 0.8 km s-1.

Open with DEXTER

thumbnail Fig. D.41

0.9 mm spectrum of CrA-37, smoothed by a factor 8, corresponding to a channel width of 0.6 km s-1.

Open with DEXTER

Appendix E: Unidentified spectral lines

We here present our attempts to identify all U lines detected towards IRS7B: The faint line at 229.7760 GHz (assuming vLSR = 5.7 km s-1) could be the CH3CHO line at 229.7750 GHz, but that would give this line a vLSR = 4.5 km s-1, which is 1.2 km s-1 lower than the median LSR velocity and 0.5 km s-1 lower than any other line, including the more certain CH3CHO lines (also if taking the JPL uncertainty of 50 kHz on the measured laboratory line frequency into account). The line at 225.160 GHz coincides with a CH3SH line, but this is not the strongest line expected from this species in the studied frequency range. The 237.998 GHz line is consistent with a c-H13CCCH line, but if this is a true detection, the c-H13CCCH abundance would be of the same order as the c-C3H2 abundance, which is very unlikely. The remaining U lines do not align with any spectral lines found in Splatalogue of species expected to be present at this rms level. Two of the detected lines agree with known U lines at 223.756 GHz and 233.456 GHz. As a result of incomplete image-band rejection, a few so-called ghost lines appear at the image frequencies of the strongest spectral lines (CO isotopologue lines and strong H2CO, CH3OH, and CN lines). We did not tabulate these ghost lines, but they are labelled in the spectra in Appendix C. For all identified and unidentified lines, the image frequencies were investigated to exclude the possibility that any reported line would be such a ghost line.

Appendix F: Non-LTE models of H2CO emission

Mangum & Wootten (1993) showed that ratios of H2CO transitions involving the same Ju-level but from different K-ladders (such as the 303 → 202/ 322 → 221 and 505 → 404/ 524 → 423 ratios) are excellent tracers of the kinetic temperature of the gas because they only operate through collisional excitation. However, at densities n ≲ 108 cm-3, the different Ju-levels are not fully thermalised, which means that ratios of transitions involving different Ju-levels (such as 303 → 202/ 505 → 404) are sensitive not only to the temperature, but also to the molecular density n(H2).

Mangum & Wootten (1993) used LVG models to derive a method for extracting temperature, density, and column density from H2CO line observations. Jansen (1995) performed RADEX modelling of H2CO line ratios to show which H2CO line ratios can be used to trace the kinetic temperature at rather low column densities, investigating certain line ratios at a p-H2CO column density of N = 1012 cm-2. In Fig. F.1 we show such plots at a p-H2CO column density N = 1014 cm-2, which agrees better with the properties of the sources in this study. The 303 → 202/ 505 → 404 ratio clearly is a particularly poor temperature probe, while the 303 → 202/ 322 → 221, 303 → 202/ 321 → 220, and 505 → 404/ 523 → 422 ratios probe the temperature well at T ≲ 50 K and n ≳ 105 cm-2, also at this relatively high column density. We used the p-H2CO molecular data file from the LAMDA database (Schöier et al. 2005), which uses the collisional rates for p-H2CO-H2 from Wiesenfeld & Faure (2013). If we instead use the older Green (1991) p-H2CO-He collisional rates corrected to collisions with H2 by correctional factors for pressure broadening and relative collision velocities the estimated n(H2) values become ~50% higher.

To acquire the H2CO temperature T, column density N, and the H2 density n towards the sources where we have measurements of both Ju = 3 and Ju = 5 transitions, we used the following recipe developed from the method of Mangum & Wootten (1993):

  • 1.

    Measure the rotational temperatures of the Ju = 3 and Ju = 5 transitions separately and assume that the kinetic temperature is the weighted average of these two temperatures. If only the Ju = 3 transitions can be measured, we use this as the kinetic temperature (and conversely, we use only the Ju = 5 temperature in IRS7A, since the Ju = 3 transitions were not covered).

  • 2.

    Use the non-LTE radiative transfer code RADEX (van der Tak et al. 2007) to calculate model line strengths of the 303 → 202 and 505 → 404 lines for the kinetic temperature and a grid of n and N (where N is the total H2CO column density assuming an ortho-to-para ratio of 1.6).

  • 3.

    On this (n,N) model grid, plot lines corresponding to the measured 303 → 202/ 505 → 404 line ratio and the measured 303 → 202 line strength. The values of n and N are determined from the intersection of these lines (see example in Fig. F.2).

thumbnail Fig. F.1

Ratios of H2CO line intensities as modelled by RADEX (solid) and assuming LTE (dashed) with a column density N = 1014 cm-2 and line widths of 2 km s-1.

Open with DEXTER

thumbnail Fig. F.2

RADEX fit for H2CO in IRS7B, given T = 40 K (from the rotational diagram fit of the full IRS7B survey). The solid line shows the measured 303 → 202 line intensity, and the dashed line is the measured 303 → 202/ 505 → 404 line ratio. 3σ errors are shown as grey lines, but can barely be distinguished from the measured values. When we estimate uncertainties, the errors on the fitted temperatures are also taken into account.

Open with DEXTER


© ESO, 2015

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.