You are in: Instruments > GMOS > Performance and Use > GMOS Components > MOS    
[GMOS logo]

Multi-Object Spectroscopy

Introduction

The multi-object mode of GMOS offers the possibility of obtaining spectra of several hundred objects simultaneously. The GMOS MOS design is based upon precisely fabricating and locating a plate containing many small slits within the spectrograph's entrance aperture. With a 5.5 arcmin field of view, 30-60 slits can typically be located in a single mask, with a maximum of several hundred slits when narrow-band filters are used. A total of 18 masks (plus the IFU) can be loaded into GMOS at any given time. Actual mask production is done with a laser milling machine located in Hilo. Currently this laser milling machine produces masks for both GMOSs at Gemini North and Gemini South.

The GMOS MOS masks can be designed either from images acquired with GMOS, or from object catalogs with very accurate relative astrometry. In the first case observers use the GMOS imaging mode to record an image of the sky from which the spectroscopic target pixel positions are measured. The required imaging is obtained in queue mode prior to either classical or queue mode MOS spectroscopic observations. In the second case, accurate relative coordinates (RA and Dec) are required for the spectroscopic targets, and known transformations are used to calculate the GMOS pixel positions from these coordinates.

The MOS mode requires that the PI designs the masks and submits the relevant information to Gemini prior to the planned observations.

This page gives a brief overview of the mask design process. Users should also read the:

Currently offered capabilities

GMOS pre-imaging

GMOS imaging of the field is executed in queue mode only (even for Classical programs), and must be obtained before the MOS mask can be designed. Images which are primarily obtained for the purpose of mask design are referred to as pre-imaging in order to distinguish them from regular imaging for queue planning purposes. Normally, this imaging is executed as early in the semester as possible once the field becomes available. MOS masks are cut approximately once a month, there are monthly deadlines for submission of mask designs in order to have them included in the batch of masks cut at the next cutting date. These deadlines are available through the "special instructions" page linked from the Phase II Overview webpage. PIs can use these deadlines to estimate when they may need to be available to design their masks.

Observing time to obtain the required direct imaging must be requested during Phase I and the observations defined as part of the Phase II process. The observations executed will be charged against the program's total allocated time.

Imaging obtained with one GMOS in a prior semester may be used for the mask design for the same instrument if the observations were taken with exactly the same pointing and orientation on the sky as is planned for the MOS observations. For a given pointing and orientation, the location of the OIWFS patrol field on the sky depends on whether GMOS is on the up-looking port or one of the side-looking ports on the telescope. Imaging data obtained with GMOS on the up-looking port may be used for mask design for a side-looking port only if a suitable guide star is available without changing the pointing and orientation of the field on the sky. The same constraint applies to imaging data obtained with GMOS on a side-looking port and used for mask design for the up-looking port. GMOS North was on the up-looking port in 2001B, and has since been on the side-looking port. GMOS South is on the side-looking port. If you are in doubt about whether existing pre-imaging can be used for your mask design, please contact the Gemini HelpDesk.

If a PI has prior imaging from other telescopes, in most cases it is possible to supplement the existing images with additional GMOS pre-imaging that would then allow the PI to design a GMOS mask using the existing imaging. The time required for this additional GMOS pre-imaging would be quite short and could be taken in worse conditions than required for the spectroscopy. If the GMOS images contain a sufficient number of objects (> 50) with good signal-to-noise then these objects may be used to "boot-strap" the pixel positions of the science targets. This would be useful if, for example, the relative astrometry provided by the WCS is not accurate enough to construct a reliable object catalog. This might be the best procedure to use, for example, when designing a GMOS-S MOS mask based on imaging obtained previously with GMOS-N. Further details are given in the detailed mask design instructions.

PIs may want to access carefully the observing conditions needed for the imaging of the field. Requesting very good observing conditions for the imaging may lower the chance of getting the imaging done early during a semester, and therefore negatively impact the overall chance of getting the MOS observations completed within a given semester.

Using Object Catalogs

Beginning in October 2007 Gemini is now offering the capability to design masks from object catalogs. The PI must have an object catalog with accurate relative coordinates for spectroscopic targets and alignment stars. It is crucial that all coordinates in the object catalog be given in the same astrometric system, as systematic offsets between science targets and acquisition targets for example will produce a fatally flawed MOS mask. In most cases these coordinates should be known with a relative accuracy < 0.1 arcsec. More details on how to assess whether your program is suitable for mask design without pre-imaging are given on the object catalog instructions page. Gemini recommends that masks with slitwidths narrower than 1 arcsec be designed from GMOS direct imaging rather than solely from object catalogs.

The PI must also create a "pseudo-GMOS" image of the field, using the gmskcreate task available as part of the Gemini IRAF package. The pseudo-GMOS image may be helpful for designing the mask if one is using the Gemini MOS Mask Preparation Software (GMMPS), and must be submitted via the OT along with the mask design as it is required during the mask checking process. This image is created from a PI-supplied image taken with another telescope and the gmskcreate task transforms that image onto GMOS pixels when you specify fl_getim=yes.

As with MOS masks designed from GMOS direct imaging, the PI designs the mask for a specific field center and position angle on the sky. Once the mask design is submitted these cannot be changed, therefore the PI must verify that a suitable guide star exists for that field center and position angle. The OT can be used for this. The MOS observations defined in the OT must use the same coordinates and position angle as those used for the mask design.

Mask Design

GMOS MOS masks are normally designed using the Gemini MOS Mask Preparation Software (GMMPS). The software is available pre-compiled for Sun/Solaris, Linux/Redhat/Fedora Core and Mac OS X. The software takes the following input:

The mask design software assumes that the input GMOS images of the field have been reduced with the Gemini IRAF Package for GMOS. Of special importance is the mosaicing of the images from the three CCDs. The Gemini IRAF Package for GMOS has a task for handling this taking into account the gaps between the CCDs and the misalignment of the CCDs relative to each other. Users who use their own reduction software should ensure that the correct transformations are used for the mosaicing.

Pseudo-GMOS images produced by gmskcreate are already properly formatted for use with GMMPS.

The software allows automatic design of the mask with the possibility of interactive editing of the result. The output from the software are the Object Definition Files for each of the desired masks.

The mask design allows slit-lets of different widths and lengths to be used in the same mask. It is also possible to place slit-lets in several "banks" in the spatial direction. This is useful if combined with a choice of grating and filter that ensures that the spectra do not overlap on the detectors.

The field of view within which the laser cutter used for manufacturing the masks can cut good quality slits is approximately 305 arcsec by 305 arcsec, which is slightly smaller than the GMOS imaging field of view.

Submission of the Object Definition Files

The PI is required to submit to Gemini the Object Definition Files for each of the desired masks. The detailed mask design instructions include information about the procedure for submission.

Mask Cutting

The mask designs are checked by NGO staff and when they have been accepted they are forwarded to Gemini. The cutting of the masks is done by Gemini Staff using the laser cutting machine located at the Gemini North (Hilo) base facility. Plans to procure and commission a second laser cutting machine at Gemini South (La Serena) are underway, which would allow for faster response times between the mask submission and the MOS observations execution with GMOS-S. Gemini staff will take care of cutting the masks and shipping these to the telescope to be used for the observations.

Further information

The original description of the user requirements to the Gemini MOS Mask Preparation Software is also available. The current version of the software does not yet have all the features described in this document.

[Science Operations home][GMOS home]

Last update October 24, 2007; Kathy Roth, Ilona Soechting
Previous version October 20, 2006; Kathy Roth