A Next-Generation Telescope for CFHT

David Crampton, C. L. Morbey

Dominion Astrophysical Observatory
HIA, NRC, Canada
Electronic-mail: David.Crampton@hia.nrc.ca

Abstract:

It is argued that Canadian astronomers, in particular, will require access to more large telescope time than currently being planned and that French astronomers will require access to Northern facilities. It is also argued that it is probable that the new 8m class telescopes will out-perform CFHT in delivered image quality and be more cost-effective in operation, so that CFHT will eventually have to be replaced. Given the timescale for acquiring funding and constructing large projects and also the number of 8-10m class telescopes that will soon be in operation, a widefield 25-30m class telescope is proposed. A novel design that is scalable to this size yet still will deliver superb images from 0.3 to 1$\micron$ over a one degree field is briefly described.

Introduction

Given that at least 16 telescopes of 8-10m aperture are either already completed, under construction or in planning stages, do we need yet another big telescope? At least for Canadian astronomers the answer must be yes or, at least, that access is required to additional telescopes to augment their 15% share of the Gemini telescopes.

Given that CFHT is already no longer competitive for most cutting-edge spectroscopic projects, will provision of widefield imaging capabilities in the 0.3 to 2.5$\micron$ region alone be sufficient to ensure a competitive future? We suspect not.

These issues are discussed in more detail in this brief summary, and suggestions for a next generation CFHT are explored.

Canadian Requirements

Canada has a very strong tradition in astronomy, dating from 1917 when the DAO 1.8m telescope became the world's largest telescope (for a very short time!). From then until 1950 when the 200-inch was completed, Canadian astronomers had access to about one-third of the world's large telescope time (including the DDO 1.8m which was completed in 1934). In 1979 when the CFHT was commissioned, Canada's 42.5% share represented $\sim$10% of the world's large telescope capabilities, but this has now declined to $\sim$5% of 4m class telescopes (i.e., not including the two Keck telescopes). When the two Gemini telescopes become operational in a few years, the Canadian share of the largest telescopes will have further declined to $\sim$2%.

In addition to CFHT, Canadian astronomers are frequent users of CTIO and other 4m telescopes. For example, among proposals for the CTIO 4m for the second semester of 1998, eight proposals were from Canadian PI's and 15 more involved Canadian co-I's. Hence, in addition to the $\sim$70 nights per semester that Canadians use on CFHT, they are obviously involved in at least as many nights again (assuming their proposals have average rate of success) on other 4m class telescopes. In other words, to support the current level of research in Canada that relies on optical telescopes, about 100 nights per semester on the world's largest telescopes are required (in general, users/programs simply must move up from the 4m to 8-10m class telescopes to remain competitive. Similarly, our students must be able to carry out projects on world-class facilities to be competitive).

In the past, Canadians have been fortunate in being granted competitive access to, particularly, telescopes like those at CTIO in order to use unique equipment or to observe southern hemisphere targets. With the new formal partnerships (e.g., Gemini), this will no longer be the case. This means that all that Canadian astronomers can expect is their 15% share of Gemini which will translate to about 36 nights per semester (total, on both telescopes) after engineering time and guaranteed host country nights are deducted. Hence, there will be severe competition for Canadian time, and Canadians may well not be able to undertake projects requiring substantial telescope time on their own. This conclusion is not new or surprising, because a decade ago Canadian astronomers argued very strongly that they needed at least a 25% share in Gemini, but then political and budgetary problems forced the share to be cut in 1991.

With the advent of the VLT, the situation for French astronomers is different, although they have a different problem: they will not be able to reach important northern targets. French astronomers currently have access to the equivalent of about one 4m (combining CFHT, ESO 3.6m and NTT time) or 180 nights per semester, and they will soon have roughly the same number of 8m nights. This is additional support for the arguments above, that Canadian astronomers will also need at least as many nights on 8m class telescopes as they currently have on 4m telescopes. Further support is evident in the recent IAU membership statistics published in IAU Bull. 82. Canada has roughly 1/3 as many members as France, but a higher percentage of Canadian astronomers depend on access to optical facilities than is the case in France (according to my french colleagues), once again suggesting that more like a 50% share of an 8m telescope is required for Canadians.

CFHT Status

The CFH community is used to thinking of CFHT as, arguably, the best 4m telescope - delivering the best images from the best site on earth. Of course, the Keck telescopes, especially, challenge that perception owing to their much larger light gathering power, but the forthcoming 8m telescopes will likely add a much bigger challenge - a combination of larger light gathering power and better image quality. The Gemini telescopes, for example, are being designed and engineered to deliver 0.25'' images, and all systems and instruments are being designed so that they do not degrade this image quality by more than 10%. The recent impressive results from the VLT UT1 demonstrating 0.27'' images during the early commissioning phase suggests that such image quality will be achieved. Furthermore, several instruments will be simultaneously mounted on Gemini and other 8m telescopes, allowing queue-scheduled observations to be executed such that full advantage is taken of the best seeing conditions. If these 8m telescopes deliver much better image quality than CFHT and can make optimal use of the site conditions in an effective way, then CFHT will no longer be really competitive in any way.

Operational costs may well also prove to be a significant problem for CFHT in the future, since one of the design criteria for telescopes like Gemini was one of operational efficiency. Unlike CFHT, Gemini will not have any ``upper end changes" and many fewer instrument changes since most instruments will be semi-permanently mounted on the Instrument Support Structure. Thus, either new, cost-effective operational models and/or major changes to the telescope and instruments will be required for CFHT to remain competitive in terms of ``photons per dollar".

CFHT has already met stiff (essentially unbeatable?) competition from HST for many imaging projects where excellent image quality is paramount, and the proposed NGST will completely dominate the near-IR wavelength range due to both its image quality and low sky emissivity. Despite the enormous success of AOB in delivering HST-like images, the realities of ground-based observing in the near-IR (with variable seeing, enhanced and variable sky emissivity, atmospheric absorption, moonlight, daylight, restricted isoplanatic angle, etc.) as compared with near-IR observations from space make the latter unbeatable. I, personally, have doubts about the efficiency and effectiveness of laser guide stars too, so sky coverage is an additional handicap for ground-based AO observations. If NGST is successful then most near-IR projects will be best carried out from space. Hence, towards the end of the timeframe under discussion, CFHT will face severe competition from space-based as well as large ground-based telescopes.

A Large Wide Field Telescope

Most of the large telescopes have relatively modest-sized fields of view. Gemini, for example, has a maximum field diameter of 10 and the fold mirror in the instrument support structure has a diameter of only 7. The VLT field is larger, 15 at the cassegrain focus and 30 at the Nasmyth focus. In addition, the f/ratios of most of these telescopes are such that the largest fields cannot be corrected by refractive optics due to sheer size limitations of the required optical elements. For example, the field lenses for the 7 field at the f/16 Gemini focus have diameters $\sim$350mm, near the maximum diameter for types of glasses that are necessary to provide panchromatic correction over the 0.35-1$\micron$ spectral region. Nevertheless, there are many projects where significant multiplex advantages could be realized if simultaneous spectroscopic studies of dozens of objects in a one degree field were possible. Hence, we decided to investigate designs for such fields on very large telescopes. Initially, this was in the context of exploring some of the options for replacing the 3.6m CFH telescope with a larger one within the same dome or, possibly, on the same pier but within a different enclosure. Some of these designs were included in a study for CFHT by Grundmann (1997).

Scientific and technical considerations

There were several considerations involved in the initial specifications:

• superb image quality is essential for scientific reasons as well as in order to capitalize on the advantages of the Mauna Kea site
• observations at wavelengths greater than 1$\micron$ will best be carried out from space, i.e., NGST
• there are many projects (e.g., QSO absorption lines, galaxies near the peak of their star formation near z = 2 [Ly $\alpha$ is beyond 0.3$\micron$ for z > 1.5]) which would gain substantially if good images were delivered in the ultraviolet, down to the atmospheric cutoff near 0.3$\micron$
• most current 8-10m telescopes and instruments are primarily being designed to deliver superb image quality over relatively small fields, e.g., with adaptive optics, hence wide fields are of interest.
• the basic large imaging and spectroscopic surveys required to study characteristics of populations are being carried out, or will be, with the current 4m and 8-10m telescopes, e.g., MegaCam, MOS, DEEP, VIRMOS, etc.
• the next step will be to understand the physical reasons behind observed phenomena so that detailed studies of individual objects will be important
• simultaneous Integral Field Unit (IFU) spectroscopy of many objects within a field is an excellent way of attaining significant multiplex advantages while obtaining 2-D spectra with relatively good spatial resolution.
• to sample fields or targets of reasonable size and to achieve a significant multiplex advantage, a field of $\sim1\arcdeg$ is required

General specifications

Hence, we have studied designs for large telescopes with:

• superb image quality (<0.1'') over the 0.3-1$\micron$ range
• a one degree field
• flat focal surfaces and relatively fast f/ratios (e.g., f/4, to reduce the scale). Such a design would be useful for direct imagers or slit mask spectroscopy.
• good telecentric properties and more moderate f/ratios (e.g., f/8) and curved fields that would be ideal for use with multiple IFUs.

 

Figure 1:   Design for a 13m f/10 telescope with a one degree field.

 

Figure 2:   RMS radius of spots as a function of the radius of the field for the telescope in Fig. 1

 

Figure 3:   Encircled energies as a function of image radius at various field locations ranging from the centre (0.0, 0.0) to the edge (00, 05).

Widefield four-mirror telescopes

Four-mirror (``Korsch") telescope designs with simple, conic, figures can meet such criteria. Our preferred option is to use an additional fifth mirror to fold the beam into a very compact design. For example, the f/10 telescope with a 13m primary shown in Figure 1 will just fit within the present CFHT dome (although the shutters won't open that wide!). The fold mirror in these designs could direct the beam in different directions if desired. Obviously, with the beam folded in this manner there is a central ``hole" at the centre of the field. This ``hole" is relatively small and would not pose a problem for, say, the multiple IFU mode of operation proposed above. In the design shown, the field is completely obscured at the very centre, suffers from >50% vignetting over 5% of the area of the total field, <10% over 88% of the area, and 84% of the field has no vignetting at all from the fold mirror. However, the gains are tremendous. As shown in Figure 2, the polychromatic images over the entire one degree field are superb - with an RMS diameter of $\sim22\micron$. Figure 3 shows that 85% of the encircled energy is within a diameter of 36$\micron$(006) for all images, even those at the edge.

Given the proliferation of 8-10m class telescopes (for which excellent performance has already been demonstrated) and the desire to get detailed information about, for example, very faint, very high redshift galaxies, a significant increase in aperture is warranted. Indeed, Gilmozzi et al. (1998) show that it is feasible to construct a 100m telescope with current technology. Matt Mountain (private communication) and others argue that a ground-based telescope must have an aperture >25-30m to be competitive with an 8m NGST. The design shown above is obviously scalable to larger sizes. We believe that its combination of aperture, field size and superb image quality throughout the UV-visible range make it extremely attractive.

Optimizing Scientific Output

If the field of such a telescope were populated with movable IFUs, each feeding spectrographs optimized for different spectral resolutions and wavelength regions, one could envisage performing several projects simultaneously at one telescope pointing. One example might be obtaining long exposures at relatively high spectral resolution of a quasar while simultaneously taking shorter, lower resolution exposures of all the nearby galaxies to identify the absorbers. One of my personal interests would be to obtain 2D spectroscopic observations of all the galaxy lenses and lensed objects in the field; current statistics indicate that there are $\sim$25 per square degree. In this way both the lensed and lensing objects could be studied, both the light and dark matter profiles of galaxies could be probed, and multiple observations of variable sources can be used to directly estimate cosmological constants. The characteristic sizes of these lenses are a few arcsec, so good seeing/image quality is required. In periods when the seeing is bad, however, the IFUs could simply be used as image slicers to observe point sources. Completely different projects could be carried out within one field, or a variety of ``parallel" type projects could be envisaged, in which spectra of various types of objects could be obtained during exposures on the primary target. Hence, as well as queue scheduling the telescope, the IFUs could also be queue scheduled to achieve scientific goals in a timely fashion.

Summary

There are many convincing arguments that Canadian astronomers will require more nights on large telescopes than provided by their share of the Gemini telescopes. Research by French astronomers will be hampered unless they can gain access to a large telescope in the northern hemisphere.

Excellent image quality can be achieved over wide fields throughout the 0.3-1$\micron$ range even with very large apertures if a fourth mirror is incorporated in the design. It is suggested that populating the field of such a telescope with up to 100 IFUs feeding spectrographs with a variety of resolutions is very attractive from a scientific viewpoint, and will provide significant multiplex advantages. We propose that Canadian and French astronomers and their partners should actively pursue such a facility, possibly as a replacement for CFHT.