A meeting was held at STScI on Feb, 26 and 27 to discuss the science case for a 20-30m telescope in space. It became clear during the discussions that a 30 meter can do qualitatively different science than can a telescope with 20m or less aperture, so the following summary focuses on that case. The speakers and talks are listed on the STScI webpage.
Four Science Drivers
There are four big programs that demand a 30 meter telescope in order to make significant scientific progress.
A) Making a volume/redshift census of black hole masses. This study (analogous to luminosity-function determinations for galaxies) sets the resolution demand for the UV telescope by asking to study the sphere of influence of the central objects in AGNs. Four milliarcsecond angular resolution meets this need, for the largest black holes, anywhere in the Universe. The study demands full aperture, not interferometry, because of the large number of objects that need to be measured. Lower mass objects have smaller spheres of influence; the telescope scale adequate to do the more distant, high mass objects, is adequate for the low mass black holes, at low z. A larger volume is needed, containing distant AGNs, to obtain a large sample of massive objects, but for low masses, lower volumes (redshift) are sufficient, as the low mass holes have a higher number per cubic megaparsec.
B) Producing an unbiased study of the spheres of influence of 1 solar mass stars. For planets in orbit around solar-like stars, 4 mas is needed to probe an adequate volume of one million cubic parsecs (0.5 AU orbits at 100 parsec distance). Large numbers of planets around low mass stars, with higher space densities, can be done at higher spatial resolution at shorter distances from Earth; the higher mass stars, with lower space density, would have to be done at greater distances to get a large sample (as for the AGNs). The filled aperture is needed because, once again, a systematic attack on the issue of spheres of influence is needed to encompass not particular, known cases of disks, jets and orbiting rocks, but all of the above in a fair sample of all stars.
C) Mapping the cosmic web. The story of the assembly of galaxies is believed to be found in the connection between the cosmic web of gas and the known galaxies, at a variety of redshifts. The gaseous web is generally invisible except by probing the gas using distant QSOs as background sources for the study of the UV absorption lines of H I, O VI and C II, for instance. The map of the cosmic web does not require high angular resolution, as do the first two problems, but does a require a very large aperture because the number of points to sample (distant QSOs) is so large. The Sloan Digital Sky Survey provides 1 million QSOs over 10,000 square degrees, virtually all the sources needed for this study. They are largely too faint for HST and at any rate it is not twos and threes but two hundreds and three hundreds of sources that are needed, even for a "small " study, never mind a key project. Hundreds of each of several types of objects (galaxies, groups of galaxies, clusters of galaxies) need to be probed in tens of angularly proximate, background QSOs for each foreground galaxy, group, or cluster, to map the web and its connections to galaxies, in detail.
D) Constructing HR diagrams for star clusters out to Virgo. Population synthesis studies promise an independent story of time evolution if we can reach the main sequence in Virgo. The suggested volume also includes enough examples to provide a "fair sample" of stellar populations. A 30 meter aperture is needed to do a complete study in a finite time (still years, however). High angular resolution is necessary to obtain accurate photometry and uncontaminated spectra of individual stars in a large sample of clusters in a large sample of galaxies, to study the history of element build up in the Universe and the constancy (or not) of the initial mass function.
A good case was made for studies reaching below the Lyman alpha line for Project C. Working longward of 1200A is often adequate, but lower wavelengths, as low as 1000A, are sometimes demanded, for a number of reasons (for example, for the cosmic web, lines of O VI, at 1031A and 1037A, and H I, 1026A are essential; for other programs, line of molecular hydrogen, 1048-1108A, are essential). Warren Moos made the point that big spectrographs need lots of reflections, which makes the region below 1000A challenging. In talking to attendees outside the meeting, it seems a reasonable approach to allow the spectrographs to include the 900-1000A region (it is not much more detector real estate than is needed for the main program) but to just take the efficiency hit for S VI (933, 944A) and C III (977A) and the higher lines of D, at low z, consistent with maximizing the telescope throughput from 1000A to longer wavelengths. Therefore, these lines will only be explored in special cases. We think the case is strong for the UV for Project B as well, though this was not stated in the meeting by the scientists interested in extra-solar planets. The angular resolution of the 30 meter telescope at UV wavelengths greatly relieves the issues of scattered light rejection that plague Terrestrial Planet Finder. These problems are reduced further by working with emission lines in the UV, where the parent stars are not so bright and the contrast is higher. The 30 meter makes all the difference in terms of detecting the resonance scattering from planetary atmospheres. However, the case, cast thusly, needs to be developed in detail.
Project A can be done in the UV or the optical with the 30 meter telescope in space: the critical issue is the angular resolution. It is just a case of numbers of objects that can be studied. By going to the UV, one gets margin in the needed angular resolution. The UV will also make separation of the AGN phenomenon from star formation regions possibly associated with the AGN, easier. The optimal wavelength is a trade between angular resolution and a bright enough galaxy flux level to measure velocity dispersions. It was felt that there are good lines at 2600A, not yet used, for the latter measurement.
The UV was also favored by another current during the presentations: the need to know the local (low redshift) swimming hole well for making comparisons with the distant (high redshift) Universe. This was especially true in the areas of population synthesis, of star forming galaxies and of the cosmic web (O VI at low z is essential).
A number of talks addressed problems that required extension to 1.6 microns. Roger Angel kept repeating, and we agree, that in the time frame of the 30 meter UV telescope, the near IR studies are likely to be doable from the ground and should not be done with high-value space dollars.
Spectral Resolving Power
There was considerable discussion of spectral resolving power. For Project A, two separate talks advocated R=3,000 and 30,000, for different purposes. For planet detection and disks (Project B), R=30,000 is needed up close to the host star, but as one observes objects further out, the Keplerian drop in motion makes 300,000 a desideratum. For the cosmic web, 30,000 is adequate, at first glance. But, for the comparison with the low z Universe, R>1,000,000 is needed. This last requirement is tied to abundances and their rise in the Universe, the determination of which requires facing the same problems as we face in the Local Interstellar Medium with blended, saturated lines: we just have not faced the issue because we do not have the equipment to do it correctly. HST studies of the ISM of the Magellanic Clouds tell us what to expect and the very highest resolving power is essential.
Field of View
So, what is needed is a 30 meter aperture, diffraction-limited telescope to work from 1000A to 10,000A. Imaging fields as large as 20 arc minutes would be desirable for stellar population studies (Project D) and for spectroscopy at R=30,000. In the latter case, the typical separations of common objects (QSOs) is 1-6 arcminutes and a multislit capability with movable slits to do 10-20 objects over that field will make the telescope much more efficient. For high angular resolution (Projects A, B, and D), 20 arcseconds seems adequate. The very highest resolving power can be done on- axis. In this last case, the finest images are needed because of the need for a very small slit, to control the size of the spectrographs.
The technologies needed were discussed a bit. The telescope should be serviceable and have a long life. Even with the large aperture of 30 meters, all subsystems will need to be designed for maximum throughput to carry out the very large programs suggested above. For instance, large, photocathode-based devices with very high quantum efficiency may be feasible that are more efficient that current devices by up to a factor of 5, but studies are needed. These studies do not need to wait for any particular space opportunity to move forward. Spectrographs, likewise, can be developed and studied. For these items, only money is needed, not a space opportunity, in the near term.
The real show-stoppers for the 30-meter, UV telescope are access to an efficient, servicable observing location (one of the Lagrangian points) and the management of the surface of the mirror. These last two involve other aspects of NASA and are very long lead items. Special funding may be needed, and special opportunities.
Prioritizing The Science -- A Robust Mission
Warren led a lively discussion on making the extra solar-system planetary science a prime project. This is a good idea, right now, because the requirements for the planetary work will satisfy all of the above needs and because the study of planets has been announced as the focus of the NASA exploration program. The nice outcome of this meeting, which we all sensed, is that the main scientific drivers are clear for each project and that all drive the mission requirements into the same territory, for angular resolution, aperture, wavelength coverage and instrument modes. The mission is thus robust to changing agency goals and to changing general science priorities of the community, an essential component of a mission that is over a decade away.
The above summary was prepared by Donald York, with minor modifications by Warren Moos. We thank the Space Telescope Institute for inviting us to facilitate discussions and prepare this summary.