General description

As of a few years ago, the MOS masks have been fabricated in Copenhagen. As the masks have to be shipped to the NOT, the design of the masks has to be started timely, e.g. not later than 1 month before the run. This implies that any required pre-imaging has to be planned about 2 months before the run, at the latest.

When and How to use Multi Object Spectroscopy

• You can use the MOS option if you have 3 or more objects within the "central section" (y=750-1250) of your field. If objects are situated outside this field, the spectra will run out of the top or bottom of the CCD. This means of course, that the "central section" depends on the length (dispersion) of your spectra. If you use low dispersion grisms like 10,11, or 12, the central section will be larger; if you use high dispersion grisms like 6, 7, or 8, the central section will be smaller. You will get information on this when using the "mosplate" programme.

• Before the run:
  A necessary pre-requisite is that you obtain a WELL-CENTERED image of your field - for most practial applications the centering accuracy should be around 10" in alpha and delta. The pre-imaging needs to be done well in advance of the run, as the design and fabrication of the masks typically takes several weeks. We will try to obtain the necesary pre-imaging during technical or service nights.

  You may want to optimise the field rotation to best allow for atmospheric dispersion. E.g. if you have 3 masks a night, then it may be good to optimise the field angle of the first and last according to the parallactic angle.
  On the other hand, you may want to optimise the field rotation to fit most of your targets in the central region of the chip in y-direction (see above), in order to obtain the most objects with full-wave-range spectra.

  The pre-imaging should be done without a filter in the ALFOSC filter wheel. However, if the spectra are to be taken with a filter in the beam, then that filter should also be in the beam when doing the pre-imaging. This is because filters inside ALFOSC tend to change the transformation function discussed below. Filters in either of the FASU wheels do not affect this.

  Note that in order to manufacture the MOS-masks, a non-trival transformation is made from CCD pixel coordinates to mm-coordinates on the MOS-mask (to be manufactured in a CNC-controlled milling machine). Therefore it is not possible to artificially shift your field on the CCD, nor to use absolute alpha and delta coordinates.

  You will need to identify 3-5 stars in your field (here full field can - and should - be used) for aligning purposes. Inevitable spectra of these stars will also be obtained, so you must make sure that the spectra of the stars do not interfere with the spectra of you science targets. Therefore you should not use too many stars for aligning. Usually 3 well distributed stars will do. The apertures for the alignment stars should be square 4"x4". The aligning, or fiducial, stars should not be too bright in order to limit scattered light.

  For the sizes of the apertures for your science targets you should consider the following:

  1. width: The width will as usual determine your spectral resolution. However, it is difficult to make slits smaller than 1" (corresponding to 140 microns) and the aligning procedure will not allow you to center better than say 0.2" - 0.3". So it will generally be more safe to use a slit width larger than 1".
  2. length: Usually it is important to obtain a good subtraction of the sky background. Therefore be sure to make slits long enough for this to be possible. Usually one will be tempted to make too short slits in order to eventually obtain more science spectra.

• During the run:
  When you have aligned your MOS-mask following the procedure described elsewhere you will need to adress the following points carefully:
  1. a) arc-line spectra for wavelength calibration
  2. b) flat-field spectra (internal halogen lamp)
  For MOS spectra it is very important to obtain these calibration spectra at exactly the same pointing of the telescope as for the science targets. If you want to determine the slit function (for each MOS slit) you should obtain spectroscopic sky flats.

  If you need standards (stars or galaxies) for e.g. radial velocities you should obtain these through one (or more) of the MOS slits, to ensure that the spectra will have the same resolution as for the science targets.

  Absolute calibration (using spectrophotometric standard stars) will be difficult and is probably not to be recommended unless you are an expert user. Ideally the standards should be observed through each individual slit. This is of course not feasible in practice. The second best is probably to take the standard through (one or) two of the slits and to obtain spectroscopic sky flats, which then can be used for determining large scale spatial variations of the sensitivity of the whole system.

Practical hints for preparation of a MOS observing run

• The use of more than 3-4 MOS masks per night is not advisable. The time used to offset the telescope is around 10 min for an expert user and probably on the order of 20 min for well prepared users.

• It is not practical to have more than 20 and typically 8 - 12 objects per MOS mask. In practice the 3 aquisition stars will fill some of the useful space on the mask.

• For mask design it is advisable to have:
  1. A hardcopy of the ALFOSC field(s) in question with the objects for which slitlets should be made marked in a clear way. The printouts should be numbered.
  2. A numbered list of ALFOSC ccd image pixel X and Y coordinates for these objects e.g. Field 1 (grism 4, slitwidth: 1 asec, slitlenght: 10 asec) obj 1 x: 1267 y: 739 obj 2 x: 1898 y: 866 . . . obj n x: 1634 y: 802 The lists should be numbered so the accopanying field printouts can be identified.
  The reason for both the printout and the list is that if the object are faint, they cannot be identified in the image display of the layout program unless the pixel coordinates are available.

  The mask design is done using the MOSPLATE program. This program takes as input an ALFOSC image, and the ascii lists of objects. Using the program the slitlets of the masks are laid out superimposed on the ALFOSC image, and then the program can generate a CNC machine file used for machining the MOS mask.

  Note that the output of the MOSPLATE program comes in 3 types

  1. The CNC files
  2. ASCII files with star X,Y list
  3. PS files

  The first type is used for mask fabrication. But also note that the ASCII lists and PS files are crucial for target acquisition at the telescope. Please make sure these files are present at the telescope before the run starts.