XMM-Newton Science Analysis System: User Guide

6.4.1 Grism specific tasks

Some of the tasks used in the processing of grism data are the same ones used in omichain and they perform the same functions and produce similar output. These tasks are omcomb, omprep and ommodmap (steps 0-1-2).

Other tasks are tools for spectroscopic data processing:

• omgprep (step 3) Grism images, as any other OM image, are affected by geometrical distortion which is corrected in the case of grisms by omgprep so as to obtain a linearised image. The grisms dispersion direction is inclined with respect to the detector axes. To facilitate the extraction of the spectrum, the image, after correcting it for distortion, is rotated so that the dispersion coincides with the Y axis. The rotation is done also by omgprep. The output from this task is a file (RIMAGE) from which spectral extraction can be done with any processing system. As it shall be described later, this file is the input for the interactive omgsource.

• omdetect (step 4). The same detection algorithm used in image mode data is applied to grism images coming from omgprep (RIMAGE) to search for spectral signatures, namely zero and first orders, present in the image. Each source is labelled as zero or first order depending on its shape. The results of the detection are written in a source list (SWSRLI) which contains the detected source positions in detector pixels in the rotated image. A ds9 region file type (REGION) contains all performed detections.

• omatt (step 4.1). OM grism astrometry is performed using omatt. It is applied to the source list (SWSRLI) produced by omdetect to compute the astronomical coordinates of the detected spectra. New columns with RA/Dec positions are added to the source list for all detected zero and first orders. A north aligned grism image (SIMAGE) in sky coordinates is also produced.

• omgrism (step 5). Spectral extraction and calibration of the spectra is performed by omgrism. The source list generated by omdetect is analysed to search for the spectrum of the target. If it cannot be found at the proper position, then a default extraction is performed. The positions of the zero and first orders and a flag indicating the nature of the spectrum are written in another source list (SPECLI). A ds9 region file type (SPCREG) depicts the zero and first orders.

  If the parameter extractfieldspectra=yes was set, then omgrism is used to establish correlations between the detected zero and first orders. The correlation is declared when the X coordinates of both orders are within a certain range and also their Y coordinates. After finding all the possible relations, the corresponding spectra are extracted. The proper identification of the zero order is essential since its position fixes the wavelength scale. In case a first order cannot be associated with a zero one, or Vice Versa, a default extraction is used. All detected zero and first orders are included in the SWSRLI source file, as well as the relations found among them. Only the successful ones giving rise to an extracted spectrum and the default extractions are written into the SPECLI source file and the SPCREG region file.

  The performed extraction is by default a box car, with predefined width and positions around the first order (or the Y coordinate of the zero order, if defaulted) for the spectrum and background regions. Signal weighted extraction is optional although its use is not currently recommended.

  The extracted net spectrum and the corresponding background are calibrated in wavelength and they are written to the output fits file (SPECTR) as count rate per angstrom in the whole spectral range contained in the image. Then the flux calibration, defined in a restricted wavelength range for each grism, is applied and the fluxed spectrum is written to the SPECTR fits file. The corresponding errors are also written to this file.

• omgrismplot (step 6). The produced spectra are plotted by omgrismplot in PDF and/or Postscript format. Net spectrum, background and fluxed spectrum are plotted. A picture showing the grism rotated image (RIMAGE) with the superimposed extraction regions (SPCREG) is also produced by omgrismplot.