The Herschel-HIFI Spectral Line Catalogue

 

Example of a line list content for obsid 1342191559. Detected and identified lines are indicated with coloured ticks. Specific line names are only given for the most common and/or bright species

 

Introduction

The HIFI spectral line catalogues provide a set of spectral features present in the HIFI Spectral Scans, and offer for each of them a best-guess of which species and transition those correspond to. Out of the about 500 observations HIFI performed in this mode, the catalogue generation was attempted for about 2/3 of them, which correspond to the spectral scan observations for which accurate baseline-corrected deconvolved spectra were produced, and which are provided as a separate Highly-Processed Data Product data-set.

The catalogues have been build based on a fully automatic line detection engine, coupled to a semi-automatic line assignment method making use of the JPL and CDMS spectroscopic databases. In some cases, spectral smoothing of the data was necessary in order to detect lines otherwise missed by the detection algorithm. The whole automated process was performed under HIPE, while manual checks and adjustments of the catalogue content was done with the CASSIS software.  

Catalogue content

The delivered Spectral Line Catalogue corresponds to the successful line detection and line identification achieved on a total of 278 individual Spectral Scan observations. Those are a combination of catalogues obtained from data at the native spectral resolution and/or at a smoothed spectral resolution. Their respective break-down is as follows:

  • There are 270 observations that lead to a Spectral Line Catalogue of at least one entry when using the native resolution Spectral Scan data. In those, 13 observations correspond to Spectral Scan data where only sub-optimal baseline correction could be achieved, and for which no flux or line width is provided (see also Table 8 from the Explanatory Supplement)
  • There are 76 observations for which complementary catalogues are provided based on smoothed Spectral Scan data. Only 8 of those observations have no counterpart in the list of the above 270 observations because they lead to at least one line detection only when the input data were smoothed
  • The above line lists are also provided as full concatenated catalogues for each and can be searched in the Herschel Science Archive (HSA)

All details about the generation of those line lists, as well as their validation against published catalogues, are given in the Explanatory Supplement. This document also provides further information about the content of the products that can be downloaded from the HSA. Individual files and postcards can also be inspected in the legacy product repository, together with the full concatenated catalogues at each resolution.

Cautionary notes

The completeness of the catalogue is limited by the data noise (a signal-to-noise ratio threshold of 5 was applied to the selection of spectral line feature), as well as the complexity of the line profiles (intrinsic to the source, or altered by possible OFF position contamination). On top of that, not all spectral lines could be assigned to a specific transition (so-called U-lines), and some could actually correspond to multiple solutions (so-called blended lines). The U-lines are not provided as output of this study.

Although all integrity checks show a high level of reliability in both line detection and transition assignment (see Section 7 of the Explanatory Supplement), special care should be taken when using lines with low intensity (typically below some hundreds of mK) and/or low signal-to-noise ratio, in particular when those lines are stamped as blends. Finally, the fitted line intensities should be used with precaution when the line profiles deviate from a simple Gaussian, or when the spectral line density is particularly high (which will manifest also in the line blends). See Section 5.2 of the Explanatory Supplement for additional warning flags provided with the catalogues.