OSIRIS is a Day One instrument for the GTC of wide field of view, high efficiency, and cost competitiveness, for imaging and low resolution spectroscopy. It is easy upgradable and is multipurpose. Since it is optimized for line flux determination, OSIRIS can be designated as a Star Formation Machine.
Mechanics
The
instrument mechanics are driven by very stringent requeriments on image
stability in imaging and spectroscopic modes. In the imaging mode, the
movement of a point source
on the detector must be smaller that 1/5 of the smallest FWHM (GTC plus
seeing plus intrument) per hour and a spectral stability better than 10%
(with a goal of 5%) of the nominal resolution in one hour. The errors
due to slit fabrication and assembly, errors in spectral and focus directions
due to positioning, flexures, and temperature variations should not contribute
to long-slit flux calibration uncertainties by more than 3.7% during a
whole night.
The
requirements, coupled with the need to be able to work at the Nasmyth
and Cassegrain foci of the GTC, make the mechanical design quite challenging.
To fullfill them, detailed error budgets using finite element analysis
(FEA) and Zeemax software models for analysis have been developed to control
the contribution of each component to the image movement and considering
the collimator as compensator of flexure residuals in open loop.
Optics
The Camera barrel includes a barrel for the first -focusing- doublet, and second barrel for two singlet and two-doublet lenses. The last camera lens is mounted to the detector unit as a cryostat window. The Camera barrel is formed by two units: a focusing mechanism, using the first doublet as an image quality compensator, and a passive athermalization system, using some of the other lenses, for keeping the focal distance and image quality almost constan under temperature variations.
The all-refractive OSIRIS camera consists on 9 lenses defining 6 optical elements, all spherical surfaces. The last lens is the dewar window. The camera effective focal length of 181 mm provides the required detector scale (0.125arcsec/pixel) on a flat focal plane tilted 1.83 degrees. With the present element apertures, the camera does not vignette a telescope FOV of at least 11 arcmin in diameter. The first doublet is allowed to move to focus the camera for all combinations of pupil optics and for assembly and integration purposes.
Wavelength Selection
The wavelength selection consists on a structure formed by a hollow tubular shaft where four filter wheels are mounted. These wheels can rotate around the shaft and can be positioned independently by the user to select the appropiate filter to be used. Each wheel is driven by a motor via a geared transmission and a timing belt. An incremental encoder situated on the motor axis allows to identify the optical element positioned in the optical beam.the mechanical interface with the support structure.
Following the optical path from the telescope to the detectors conventional Filters are installed in wheels 1 and 2; Wheel 3 can accommodate both conventional filters and collimated beam masks, while the two tunable filters and grisms are located in Wheel 4 (the closest to the camera). Wheels 1 to 3 are aluminium disks 20 mm thick and of 15 kg mass each one. Wheel 4 is made of steel. Wheels external diameter is large, near one meter in diameter. In total, the system allows the simultaneous loading of the two tunable filters (blue and red), 6 grisms and 24 filters. Conventional filters include order sorters for the tunable filters and a set of the Sloan Digital Sky Survey (SDSS) broad band system.
Tunable Filters
An important aspect of OSIRIS is the use of tunable filters (TF) for narrow band imaging. The tunable filters are conventional Fabry-Perot etalons in that the cavity thickness, defined by high reflectance plates separation, can be adjusted in a wide thickness range with high accuracy. In this way, the filters allows selecting a wide range of resolving powers. Tunable imaging allows obtaining line fluxes: (i) simultaneously for all objects within OSIRIS FOV, (ii) avoiding sky emission lines, (iii) at an arbitrary selected wavelength compatible with the etalon configuration.
The OSIRIS TFs have been designed for covering a range of resolutions from R = 300 to R = 1000. In order to improve the coating reflectivity, two TFs will be used. A blue TF optimized from 365 to 670 nm with a reflectivity of 91%, and a red TF with a reflectivity of 94% covering from 620 to 1000 nm.
Grisms
For the spectroscopic modes grisms are used as dispersive elements. Grisms are a combination of transmission gratings and prims, manufactured in a way that the central wavelength of the first order spectrum is passed without deviation. The maximum guaranteed resolution will be R = 2500, and the minimun resolution will be R = 250. The higher dispersion grisms (R = 2500 and 5000) are based on Volume Phased Holographic Gratings (VPH). Peak efficiencies are higher than 80% for R = 2500 and 60% for R = 5000.
Collimator and Folder
The Collimator Subsystem consist of the Collimator Mirror, the Positioning mechanism to tilt the mirror and the Cell that supports the system and allows to attach it to the instrument main structure. The Collimator is an off-axis ellipsoid having an optical clear aperture of 533 x 431 mm. A lighweight substrate having an aspect ratio of 6:1 and a minimun mass reduction of 40% has been used to fullfil the mass requirements.
The folding mirror design was assumed by SESO (France). The mirror is made
of zerodur, characterized for its very low expansion thermal coefficient
and its long term stability.
The mirror fixed via 3 metallic flexures to a triangular base made of
steel, that insures the fixation interface onto the OSIRIS support frame.
The collimator unit design and manufacturing was shared between SESO and
CSEM (Switzerland). The collimator mirror is made of zerodur, and the
total weight of the collimator system is around 150 kg.
Mask Loader
Some observing modes (i.e. multi-object spectroscopy, or tunable filter imaging with change shuffling) will require that some parts of the field of view coming from the telescope must be blocked.
The masks (build in aluminium) are situated together in a mask cassette, and a mechanism select one mask and remove it from the cassette to place it on the GTC focal plane. A slit loader with capacity for up to 13 masks allows to insert and remove slits masks to and from the telescope focal plane. In addition to user-customized masks for multi-object spectroscopy, a number of fixed width long slit masks will be available. One extra mask for point-like fast photometry and another for charge-shuffled continuum subtraction will be available as well. A two degrees-of-freedom mechanism allows selecting one of the masks, removing it from the cassette and positioning it in the telescope focal plane with the required repeatibility, providing spectral stability better than 10% of nominal resolution per hour (including contributions from all instrument subsystems). The long slit flux calibration uncertainties will be less than 4\% during the night.
The slit subsystem can carry out three functions: mask storage and identification, mask selection and mask positioning in the focal plane. The changing time (currently 24 seconds) is smaller thant the detector readout time. Masks will be stored in the cassette. The masks can be exchanged through individual doors. The masks are automatically identified within the instrument by a bar-coded reader mounted beside the cassette. When a new mask is placed in the instrument, the contents of the cassette are read and recorded automatically. In the long slit mask the slit length is 8.5'. The multi-object spectroscopy masks cover a field of view of 8.5' x 5.2'.
Control and Detector
The detector mosaic is composed of two 2k x 4k CCDs abutted, with a plate scale of 0.125"/pixel (15 microns/pixel). The Day One arrays will be a set of two e2v CCD44-82, although these will be probably be upgraded to MIT/LL CCID-20 blue-enhanced ones afterwards. The detectors have been chosen to have a maximun quantum efficiency in the red but to be blue-sensitive (MIT ones) and with minimun fringing. Readout modes include chage shuffling up and down, continuous readout, and reading windows. Readout speeds will be at least the slowest possible, intermediate (nominal), and the fastest possible. Frame transfer is also contemplated with both devices for fast spectroscopic modes.
The detector controller will be a commercial SDSU-2 controller, using a timing board with parallel cable linked to a commercial digital PMC frame grabber. The CCD controller will allow to synchronized the TF switching frequency (up to over 100 Hz) with charge motion over large areas of the detector The software architecture allows to run any of the complex observing modes, that involves coordination and synchronization of critical operations (as charge shuffling and wavelength tuning) with real time constraints.
The strict requeriments for photometry decided the project to select a Bonn shutter, manufactured by Bonn University. The design is based in slit type shutter with two independent carbon fiber blades driven by stepper motor over a rectangular aperture. The shutter blades are made of carbor fiber material. The three main components (mechanics, control electronics hardware and control electronics software) play together to achieve a photometric accuracy of the order of 1% at an exposure time of 0.1sec (i.e. timing accuracy of 1msec) at any position across the aperture.
Software
The OSIRIS control software will be deployed on different machines executing both, conventional and Real Time Operating System (RTOS). These machines will be connected to the GTC Control System (GCS) through different networks, mainly ATM. An Ethernet (10/100 Mb/s) is also used for enginnering purposes, and are also available serial and CAN type field buses. The OSIRIS Control Software is being developed using a Use Case Method for requirements capture, and using a Distributed Object Oriented Approach which is integrated into a major framework, the above mentioned GTC Control System. The RUP (Rational Unified Process) has been used as software process framework.
Two specific software packages are being written for science operations, the OSIRIS Data Factory Pipeline (OSIRISDFP) and the OSIRIS Mask Designer. Both packages were developed by GMV S.A. (Spain). The OSIRISDFP controls an automatic set of procedures that will be available to process the data acquired from standard OSIRIS observing modes. At least (a) standard imaging, (b) long-slit spectroscopy, (c) charge shuffling imaging and (d) fast photometry observing modes will be processed using the pipeline. The OSIRIS Mask Designer is a utility prepared to define the exact positions and shapes of slits in a focal plane mask in order to perform multiobject spectroscopy. The Mask Designer can be run either by using a list of values presenting equatorial coordinates or by defining slit positions with a mouse and an input FITS format image.
Commisioning
The commissioning is the phase following construction during which the capabilities of an instrument are demonstrated in its final operational configuration. During commissioning, both verification and validation tests are performed on the complete system to ensure that the instrument meets all its science requirements and is ready for operation.
The basic operational parameters of the instrument will be measured during AIV. We aim during commissioning to characterize completely the behavior of the instrument in all the operational aspects, in a stand-alone mode and integrated with the GTC control system. During the basic commissioning the main focus is not the science operation.
See also: OSIRIS - Commisioning
Instrument Upgrades
Day One upgrades are already under development.
Solutions for achieving higher resolution spectroscopy, up to 5000 or more, have been studied at the IAC. An Acción Especial del Ministerio de Ciencia y Tecnología for implementation of higher spectral resolution in OSIRIS using VPHs has been approved. We plan to reach R = 5000, with expected peak efficiency of around 60%. We are studying to implement Fabry-Perot spectroscopy in a collaboration between the IA-UNAM and the IAC, to allow 2D spectroscopy at high resolution to be performed over the instrument FOV. This mode should open a completely new window in 8-10 m class telescopes. Finally, coronography is currently under study at the IA-UNAM. It will allow the observation of emission lines of host QSOs and galactic nebulae with bright stars embedded, among other applications.
A workshop for the future of the use of a Interferometric mode in OSIRIS was held on February 19, 2004 on the UNAN organized by Margarita Rosado. The participants discussed research programs that could be done using a high-resolution etalon in OSIRIS, in the areas of star formation, interstellar medium in our Galaxy and Extragalactic Astronomy.
The use of Fabry-Perot was compared with IFUs and MOS. IFUs also divide the
image of an extended object in an array of microlenses or optical fibers,
obtaining the spectra by grisms
or gratings. Usually the IFUs provide a relatively small field (~10")
and relatively low spectral resolution (R ~ 2000 to 5000). A Fabry-Perot
in OSIRIS should be able to work at resolutions up to R=20000 in all the
field (8').
MOS are ideal when the aim is to obtain spectra of relatively low dispersion (R ~ 2000 to 5000) from point sources in the same field (~ 10'). Their utility decreases when the sources are extended and show asymmetries, and the aim is to obtain spectra of all the object. Also, when it is required a higher spectral resolution. In that case, a FP has the advantage to supply high resolution spectra over all the spatial elements of the object in the field.
A High-Resolution Fabry-Perot can be installed in the same physical position of the Tunable Filters. They will require order-sorters, and use the same CS100 controller of the TF. Calibrations will be done using lamps of H, He, Ne, and Ar.
A candidate for the etalons would be ICOS ET100. It would be convenient to adquire two etalons, optimized one in the range 6300 to 7000 A (for galactic projects and Lyman alpha emitters redshifted), and 8000 to 9500 A (for example, for kinematics of objects in the OTELO survey). A High resolution etalon should have a R=20000, velocity resolution 15 km/s, FSR 600 km/s. The Low resolution etalon should have R=5000, velocity resolution 60 km/s, FSR 1000 km/s.
Last update July 18, 2005, by Héctor Castañeda