University of Western Ontario
London, Ontario, CANADA N6A 3K7
Electronic-mail: jlandstr@uwo.ca
We present below an ongoing Franco-Canadian project for a new bright time instrument, a cross-dispersed échelle spectrograph/spectropolarimeter called ESPaDOnS, for general community use at the Canada-France-Hawaii Telescope (CFHT). The design specifications for this instrument are that it is to be able to obtain a complete optical spectrum (i.e. from 370 to 900 nm) in a single exposure, with a resolving power of about 50,000. If desired, ESPaDOnS will be able to measure all the polarisation components of the stellar light (circular and linear) with the same resolving power. The spectrograph will be bench-mounted and fed by a double optical fibre from a Cassegrain module containing all calibration facilities and the optional polarisation analyser, making it possible to have extremely good wavelength stability and minimal instrumental polarisation. ESPaDOnS should be a unique instrument worldwide in polarimetric mode, competitive with similar instruments on 8 m class telescopes in non-polarimetric mode. With this facility, French and Canadian astronomers should be able to address a broad range of important issues in stellar physics with unprecedented detail, from studies of stellar interiors, to investigations of stellar atmospheres, stellar surfaces, stellar magnetic fields, and to observations of circumstellar environments and extrasolar planets.
. The addition of polarisation optics, so that the wavelength
dependence of linear and/or circular polarisation can be measured with
the same resolving power and spectral coverage, greatly expands the
range of potential applications of the instrument. Quite a lot of
interesting scientific problems can be attacked with such an instrument,
both in spectropolarimetric and non-polarimetric mode.
Scuti, Be, O, roAp stars) pulsate in a variety of
non-radial modes which furnish invaluable information
about the structure of the stellar envelope. High
degree modes can be detected through small line
profile variations (e.g. Kennelly
et al. 1998) and will be very
useful additional constraints
for identifying pulsation modes on targets to be observed with the very first
asteroseismological spacecrafts COROT/MOST built by France/Canada. Addition of
information from many lines can make possible detection of
these tiny distortions
in spite of exposure times short enough to avoid phase smearing.
As often mentioned, many of the scientific programmes listed above can make use of multi-line (e.g. cross-correlation) techniques to boost their sensitivity by a considerable factor (e.g. a factor of 60 in signal-to-noise ratio in the particular case of Zeeman-Doppler or Doppler imaging of cool stars for instance). We estimate that such multi-line techniques could for instance allow us to:
,
,.
To carry out these programs, ESPaDOnS must comply with the following specifications:
To achieve these goals, we essentially plan to duplicate Semel's visitor
stellar polarimeter (Semel et al. 1993) and link it with
a double optical fibre to a bench-mounted spectrograph very similar to FEROS,
the Fibre-fed Extended Range Optical Spectrograph that ESO develops for the
1.5 m telescope on La Silla (Kaufer et al. 1997).
To obtain a polarimetric analysis with maximum efficiency, we will use a
quarter-wave and a half-wave Fresnel rhombus. These devices proved to be very
achromatic without producing spectral ripples such as those generated by
the super-achromatic Halle waveplates (Donati et al.
1998). We will use the low-OH
H-treated Ceram-Optec fibres (with 100/110 
m
core/cladding diameters) which provide
close to optimal transmission throughout
the whole spectral domain of interest.
Finally, we will copy most of FEROS optical
design (dual-pupil mounting, cross-dispersing prism,
79 gr/mm grating, fully dioptric
camera with 400 mm focal length,
2048
4096 15 m pxl CCD detector) except
for a larger pupil size (200 mm),
which should enable us to collect stellar light
through a 1.5 circular aperture
(with no increase in detector size nor
decrease in spectral domain and resolution).
The performances of ESPaDOnS should
therefore be very similar with those of FEROS,
i.e. full spectral coverage from
370 to 900 nm (orders #25 to #61) at
50,000 spectral resolution and with 20%
peak efficiency at 500 nm (and 6% throughput
on both spectral domain edges).
Refurbishing the CFHT MOCAM camera for our own
purpose (by replacing the existing
chips by e.g. one 20484096 pxl
spare chip of the CFH12k) and turning it into a
dedicated camera should minimise tuning
procedures and make ESPaDOnS a truly `point
and shoot' instrument.
Along with a versatile telescope mounting (either on the bare
Cassegrain bonnette or behind MOS or PUEO), ESPaDOnS should altogether be very
economic in terms of CFHT staff support.
By halving the slit width, we may bring the resolution up to about 80,000 (assuming a instrument point-spread-function with a full width at half maximum of about 1 pxl) at the expense of losing about half the light and being severely under sampled (with a resolved element of 1.3 pxl). However, by tilting the slit with respect to CCD lines, one can ensure that the different spectral columns throughout each order are sampled on a different pixel grid; using micro-scanning techniques as part of the extraction routines should then allow the recovery of the observed spectrum with at least twice as many sampling points, bringing the resolution element up to a comfortable size of more than 2.6 subpixels. The total throughput in this mode (peaking at about 10%) will still be very attractive in today's spectrograph standards.
Finally, we will make available modern optimal extraction and cross-correlation tools (e.g. the ESpRIT and `Least-Squares Deconvolution' packages, Donati et al. 1997) to enable real-time spectrum processing and preliminary data analysis as soon as photons are collected.
Located at an altitude of 4200 m on the summit of Mauna Kea, Hawaii, CFHT has been the largest and best optical telescope available to Canadian and French astronomers since it was commissioned in 1979. It occupies one of the best telescope sites in the world, and is extremely well equipped for direct imaging and low-resolution, multi-object spectroscopy for studies of faint stars and extra-galactic objects. It also has a powerful, but quite specialized, stellar spectrograph with a resolving power of about 120,000, which however can only obtain a spectrum about 3 nm long at once. This spectrograph, although certainly one of the finest instruments for problems for which it is optimized, is nonetheless quite inefficient for many of the research programmes described above.
Canadian and French astronomers have quite limited resources for working on the programmes described above apart from CFHT. Other telescopes, equipped with adequate instrumentation for these programmes and accessible to our communities, have substantially smaller apertures (2 m or less), and are in much worse sites, resulting in extremely severe limitations of the observable star samples (sometimes the observable sample is exactly one, or even zero, stars). Thus most of the topics described above can only be studied by obtaining observing time from a foreign observatory (at the Anglo-Australian or William Hershell Telescopes for instance).
ESPaDOnS will not only fill this gap, but should also be by far the best spectropolarimeter worldwide. Comparing in particular with the performances of the three only similar instruments we know of, it will provide a twice wider spectral coverage and 10 times higher throughput than Semel's visitor instrument on the UCL échelle spectrograph of the 3.9 m Anglo-Australian Telescope, a four times wider spectral coverage, a four times higher throughput and a much more reliable polarisation modulator than the CASPEC spectrograph on the 3.6 m ESO telescope (which simply fails at measuring the tiny Zeeman signatures in the line profile of cool active stars), and a 2.5 times wider coverage and 100 times higher throughput than the MuSiCoS spectropolarimeter on the 2 m Télescope Bernard Lyot atop Pic du Midi (including the difference in telescope collecting power). As far as we know, there is no plan for a high resolution échelle spectropolarimeter on any of the 8 to 10 m class telescopes being built at the moment (except may be on the HROS spectrograph of the Gemini South telescope). It is therefore quite clear that ESPaDOnS should be a unique instrument worldwide (or at least in the northern hemisphere if the hypothetic polarisation module of HROS/Gemini is funded).
Moreover, because of the very large wavelength coverage and high total throughput for which our spectrograph is designed, we expect that, in non-polarimetric mode, ESPaDOnS on CFHT will be even competitive in overall information acquisition (within a factor of order two) with the high resolution cross-dispersed spectrographs being built for the 8 m telescopes (whose spectral coverage is at least twice smaller, except for HROS/Gemini). Furthermore, most of the programmes described above do not require access to instruments with such large collecting powers, because of the relatively large number of stars that will be already observable with ESPaDOnS at CFHT. In fact, because the above programmes usually require relatively extended observing runs, to study periodically or erratically variable phenomena, they are not well suited to being programmed on 8 m class telescope where demand will be extremely high and time allocations short.
In short, CFHT is particularly well suited to the kinds of programmes for which ESPaDOnS will be optimized. Furthermore, there is a real need on CFHT for another instrument to make the best use of available bright time in the coming Gemini/VLT era which will lead to greater emphasis at CFHT on using a smaller number of instruments to do a limited number of kinds of observations (particularly large-field direct imaging, but only during the dark period) extremely well. The proposed instrument is probably the most versatile and generally useful instrument that could be developed for stellar observations with CFHT.
This project is the result of the increased level of interaction that has gradually built up, principally because of the joint ownership of CFHT, between French and Canadian astronomers. This proposal was first presented two years ago to French, Canadian and Hawaiian astronomers. The proposal was discussed by the Scientific Advisory Council of CFHT, who concluded in their November 1996 report:
Recommendation #3 on Spectropolarimeter:Considering that this instrument will be unique in the world; considering that this instrument provides more opportunities for research than the current Coudé spectrograph; considering that this instrument will provide an excellent use of CFHT bright time; considering that many members of all three communities (about 30) have expressed support for its construction; The SAC recognizes that the spectropolarimeter proposed by Donati and Catala has high scientific merit. The SAC therefore supports the development of this instrument and its inclusion in the instrumentation plan if money is available beyond that committed to the other projects.
The main obstacle to having CFHT build ESPaDOnS is shortage of personnel and funds. The new fund for wide-field imaging, created with special allocations from both Canada's NRC and France's INSU as well as from savings in CFHT operation, appears to be totally engaged for imaging projects, primarily for direct imaging in visible light with a field of one square degree (the MEGACAM project). CFHT does not have enough money to fund this project alone, however valuable it would be. Furthermore, the personnel needed for specification, design, construction, and assembly of ESPaDOnS is simply not available within CFHT.
The solution appears to be a joint funding and construction. A total cost of about US$ 400 k was estimated for this instrument, spreading as follows (prices in US$):
| Polarimeter & fibre link | 70 k |
| Diffraction grating | 60 k |
| Collimators & flat mirror | 60 k |
| Cross dispersing prism | 30 k |
| Fully dioptric f/2 camera | 150 k |
| Mechanics & electronics | 30 k |
This cost can be shared by France, Canada, and CFHT budget. The proposal for funding submitted to INSU in 1997 (at a level of US$ 70 k) was ranked at the very top, making ESPaDOnS the only new ``intermediate size'' project funded by INSU in 1998. The Director of CFHT has informed us that he will most likely be able to make a contribution of about US$ 100 k within CFHT budget between 1998 and 2000, bringing our secured funds to a level of 40% the requested amount. Another proposal (worth US$ 80 k, bringing the overall French contribution to a total level of US$ 150 k, salaries excluded) has been been submitted in 1998 to the French Ministère de l'Enseignement Supérieur et de la Recherche. After a first unsuccessful attempt at NSERC in 1998, we intend to resubmit a new proposal (at a level of US$ 150 k) in Sept. 1998.
The instrument will be designed, constructed, assembled and tested at Observatoire Midi-Pyrénées. The project team includes in particular J.-F. Donati as PI, C. Catala, T. Böhm and J. Arnaud as CoIs, J.-L. Prieur as CoI/Project Scientist and J.-P. Dupin as Project Manager. The optical design will be carried out by P. Rabou (Grenoble) and L. Parès, the mechanical design by G. Gallou, the electronics and control hardware by G. Delaigue, and the control software by P. Tilloles and H. Valentin. The phase of detailed optical and mechanical design has already started and should be completed by mid 1999. It should be followed by a construction phase in late 1999, then an integration and testing phase in 2000. The instrument should be installed and tested at CFHT in late 2000 for a commissioning early 2001. Up-to-date information on ESPaDOnS will be available on-line here .
