A "Multi-Argus" Facility at CFHT

Geneviève Soucail

UMR 5572, Observatoire Midi-Pyrénées
Electronic-mail: soucail@obs-mip.fr

and

Jacques Baudrand

DAEC, Observatoire de Paris-Meudon
Electronic-mail: baudrand@obspm.fr

and

Paul Felenbok

DAEC, Observatoire de Paris-Meudon
Electronic-mail: felenbok@daec.obspm.fr

and

Jean Guérin

DAEC, Observatoire de Paris-Meudon
Electronic-mail: Jean.Guerin@daec.obspm.fr

and

Jean-Pierre Picat

UMR 5572, Observatoire Midi-Pyrénées
Electronic-mail: picat@obs-mip.fr

Abstract:

We propose to implement a ``multi-argus'' facility at CFHT. The main scientific driver of this project is the study of the evolution of the Tully-Fisher relation at intermediate redshift, thanks to the use of the optical rotation curves of spiral galaxies. The multiplex gain makes this project quite attractive, even on a 4m-size telescope such as CFHT. From the instrumental side, the idea is to modify the multi-fiber positionner FIFI (a copy of the MEFOS instrument in use on the ESO 3.6m telescope) and to include small bundles of fibers, then fix it at the Cassegrain focus and link the fibers to the entrance of the MOS. We present the way it would work in the CFHT environment with minor changes and demonstrate the scientific feasability and the interest of such configuration for several astrophysical programmes.

Scientific Case

  Multiobject spectroscopy with fibers is a versatile mode for which the field of view depends on the telescope only. It allows to reach higher spectral resolutions than slit spectroscopy with a focal reducer, although the quality of the sky subtraction for faint objects is generally poorer and requires very careful calibrations of the fibers. Another powerful advantage of the fibres is the opportunity of doing 2D spectroscopy as they can be organised to cover a compact bundle at the focus of the telescope and then be re-arranged to mimic a slit at the entrance of a spectrograph. Such systems are under construction on most of the 8-10m very large telescopes, and some are already in use on 4m telescopes like the MOS/ARGUS instrument at CFHT. The scientific programmes covered by this kind of instrument are related to the detailed 2D description of objects: kinematic maps, emission line distributions, etc ... But several programmes would benefit from a multiplex gain while still preserving the 2D approach, requiring a ``multi-argus'' spectroscopic mode. This is what we propose to implement on the CFHT, in order to prepare observations on larger telescopes where similar facilities are already planned (Gemini, VLT, Grantecan ...). We first describe a scientific testcase that would be well suited to this instrumental configuration. It concerns the dynamics of galaxies and their evolution at intermediate redshift (z < 0.3). We know from the Tully-Fisher (TF) relation that the rotation velocity of a galaxy (or its total mass) is strongly correlated to its absolute luminosity. In the nearby Universe, this scaling relation is used to measure the Hubble constant H₀ and to map the large scale velocity fields. For more distant galaxies, we want to study galaxy evolution and the increase in luminosity and/or the decrease in mass, as suggested by current models of galaxy evolution. Indeed any significant shift of the canonical TF relation versus redshift will characterize both the space-time curvature (q₀) and the evolution of the M/L ratio of galaxies. This degeneracy can be raised by coupling this study with other photometric parameters such as the color or surface brightness of galaxies, less sensitive to the cosmological effects.

Rotation curves of spiral galaxies or line widths have already been measured up to z=0.3 on 3-4m telescopes (Vogt et al. 1993, Bershady 1995) and up to z = 1 with the Keck telescope (Vogt et al. 1996). Present results are not significant, no clear evolution of the M/L ratio with redshift is detected and some controversy is found among the authors. Indeed, the selected samples are quite small and generally driven by observational constraints (large surface brightness, very blue galaxies, morphological appearance of the galaxies ...). In addition these observations are done with long slits and additional uncertainties are due to the inclination corrections. With a ``multi-argus'' instrument, we would benefit from both an integral field spectrographic mode to reconstruct the 2D velocity field of the galaxies, and an increase in the efficiency of the observations by a multiplex factor of 10. A sample of more than 100 galaxies would then be observable in a reasonable telescope time. This will allow a study of the TF relation versus morphological types at different redshifts in a more refined way and with a more homogeneous sample. At redshifts lower than 0.3, evolutionary effects are small, at least in terms of luminosity evolution. For the mass evolution of galaxies, the expected result is less clear and the constraints on the determination of q₀ are also prospective. Similar studies can be proposed in dense environment such as clusters of galaxies. The mass evolution of spiral galaxies in clusters can also be tested with the proposed instrument, and again will strongly benefit from the multiplex gain.

The MEFOS/FIFI Instrument

 The positionner necessary to place the fiber bundles already exists and is called FIFI. It is a copy of the MEFOS system developped for the prime focus of the ESO 3.6m telescope (Felenbok et al. 1997). We remind shortly its main features, useful for the implementation of a ``multi-argus'' instrument at CFHT. Thirty independant arms can point within a 20 cm field. One arm is used for guiding, while the other 29 arms are dedicated to the astronomical objects. All the arms are distributed around the field as "fishermen-around-the-pond" and can be moved radially and in rotation. Hence, each arm can access objects at the center of the field, whereas only one specific arm can reach an object at the field periphery. This situation changes gradually from the center to the edge of the field. An arm-object assignation software is provided to optimize the configuration, taking care of constraints of size and moving capacity of the arms. The results can also be modified interactively. Each arm is driven by an individual electronic slave board. The whole instrument is controlled by a PC, independently of the Telescope Control System (TCS). Each arm tip is carrying two fibers separated by 1 arcmin: one is used for the object and the other one is for sky subtraction. Each arm is also supporting an imaging fiber bundle covering a $35'' \times 35''$ area on the sky. The image bundles are projected on a single $1024 \times 1024$ CCD and an image of the selected objects can be obtained, for field recognition at the object coordinates (Figure 1).

 

Figure 1:   Schematic view of the MEFOS arm, as it worked on the ESO 3.6m telescope.

With this procedure, the objects on which the spectral fiber will be positioned can be seen in advance. By analysing the position of the object on the image and knowing the fixed displacement between the image bundle center and the spectral fiber within an arm, a precise offset can be given to the arm to reach a perfect positionning of the scientific fiber on the object. The final positioning has been checked on stellar sources at a 0.2'' rms accuracy.

Guideline for a CFHT Multi-Argus Project

The instrument described above is well performing and the operation mode has been fully tested and improved for several years at ESO. We propose to install this system at CFHT, taking advantage of the knowledge acquired at ESO, and replacing the single scientific fibers (object + sky) by a bundle of fibers allowing 2D observations of at least 7 to 10 objects at the same time (Figure 2). This is a technical change which seems easy to do (if not for free!). The instrument should be considered as a visitor one with low load (and possibly no load) on the CFHT staff. This is the main reason why we have looked at a possibility of using the instrument at Cassegrain in place of Prime Focus.

 

Figure 2:   Details of the FIFI configuration and main caracteristics of teh ``mini-argus'' bundles.

The configuration we propose will use fibers with an overall diameter of 115 (0.9'' on the sky at the Cassegrain focus, behind the field corrector) and a core of 75 (0.6''). The equivalent slit at the MOS/SIS spectrograph entrance is then 120 which gives a sampling of 2.8 pixels on the CCD (with 15 pixels). In these conditions we can accommodate 400 fibers for a single slit or twice this number for a double slit, with a reduced spectral coverage. The solution shown in Figure 2 consists in implementing 10 -11 ``mini-argus'' of 37 fibers to cover 5.7'' in diameter for each bundle. The stroke of the arms allows to cover a 20 cm field, which means about 20' at the Cassegrain.

The setting of the instrument on the telescope is shown on Figure 3. It is important to note that the only interface between FIFI and the rest of the CFHT instrumentation is the bundle of scientific fibers which feeds the MOS/SIS spectrograph the same way as ARGUS does presently. The guiding will be performed by the telescope and the CCD exposures are acquired within the CFHT environment. FIFI is fully and independentlty controlled by its own controller.

 

Figure 3:   Global view of the instrumental mounting of the multi-argus system at CFHT.

Operating Mode at the Telescope

To prepare an observation, the positions of objects on the sky are required, with an accuracy of a few arcseconds only. On the telescope the first night implies set-up operations for 2 to 4 hours to know accurately the scale, the alignment of the instrument with the North-South direction, the conjugation of the telescope and FIFI centers. To do that, a set of star fields must be observed with the Image fibers at the expected positions. All these operations are driven by automatic procedures through the observing software.

In normal operation, the telescope is pointed towards the selected field and the guiding is locked. The arms are then sent to assigned positions to have the image fiber bundles centered on the objects. During the arm positioning the PC continously displays the movements of the arms. It typically takes about 4 minutes for the arms to reach their final positions. When in position, an image is taken of all image fiber bundles and analysed to derive the offsets to be applied to each arm. This operation lasts for a few minutes. Final repositioning of the arms on the scientific fibers is about 1 minute.

Calibrations are standard ones with continuum and spectral lamps, with all the fibers in a central position using the calibration Gumball at the CFHT Cassegrain. The overhead between two scientific exposures on different fields is typically 20 to 30 minutes, including calibrations.

Expected Performances

We have made a flux estimate with reference to some real data observed at CFHT with MOS/ARGUS. For the FIFI instrument and setup (with the O600 grism in MOS), we expect to reach S/N = 5 in 1 hour on a resolved element (3 pixels) for an emission line with a line equivalent width of $W_\lambda = 30$ Å above a background with a surface brigthness $\mu_R = 22.8$ after sky subtraction. In these conditions, the line centroid can be measured up to FWHM/5, or 40 km/s. These values are quite representative of what is expected for a spiral galaxy at z < 0.3 at the edges of the disk, where the Vterm used for the TF relation is measured. We estimate than 2 to 3 hours are necessary for objects up to a magnitude R = 20. At this magnitude level, about 100 galaxies are available in the proposed field of view of 20' diameter. Note that the half-light radius of a typical spiral galaxy is about 4 kpc and the average radius of a disc is 6 kpc, or equivalently 3'' at a redshift of 0.2. A bundle size of 6'' will correctly cover the galaxy field.

A comparison of the FIFI setup and the Keck/LRIS spectrograph in terms of performances for the measurement of galaxy rotation curves shows an advantage of a factor of about 5.7 for Keck on a single object. With 10 objects per exposure CFHT+FIFI could then be equivalent to Keck (assuming 2 objects on the slit). Note also that presently no such instruments exit on the telescopes, and it is necessary to wait for more than 3 years before any multiplex 2D facility is available on the 8-10m telescopes.

Conclusion

FIFI is an existing instrument wich has been fully evaluated at ESO. The instrument and its software procedures and control are ready to be mounted on the telescope. The main modification is to replace the scientific fibers by the ``mini-argus''. The flux evaluation has shown that with this configuration CFHT could be as efficient than Keck with LRIS. A goal would be to have the instrument used at CFHT in 1999 for about a 2-3 years use waiting for the same kind of instruments on 8 meter telescopes. This would allow a ``première'' at CFHT for this kind of observation with a very prospective scientific programme. There is a short window of opportunity before similar instruments become available on 8-10m telescopes.