Really imaginative new vistas must be fascinating also to non-specialists
in any particular field: perhaps one should even make it a criterion to
have such goals defined by persons not working in that field. Of
course, the limited time made this not possible during the recent workshop,
when rather each participant was invited to elaborate on the potential
in his/her own specific field. However, almost by definition, this leads
to a discussion of enhancements in those (already established) fields.
2. Topics where dramatic advances are in prospect
Although it may not be trivial to identify thresholds that arguably will be passed when going from 8-10 m to 30-100 m telescopes, one can think of some whose potential could be further explored than I felt was the case at the Workshop, such as:
2.1. Stellar surfaces
Many stars become surface objects and will no longer be only point sources. The "Catalogue of Apparent Diameters and Absolute Radii of Stars" (Fracassini et al., 1988) lists on the order of 100 stars with diameters of 10 milliarcsec or greater. An ELT with resolution 1 mas will thus produce at least 100-pixel-images of the *photospheres* of each of these. Most of these stars are K and M giants, many of which in addition have extended atmospheres, dust shells, and other circumstellar items that can be imaged. An integral-field spectrometer will give high-resolution spectra in different polarizations for each spatially resolved surface element, enabling studies such as the dynamic mapping of stellar oscillations, magnetic-field studies of starspots and flares; and the equator-pole differences in the acceleration of stellar winds.
2.2. Neutron stars
These have hitherto been studied mainly in radio and X-rays. However, to study the neutron star itself (i.e. not the accretion-flow emission in an X-ray binary), optical, highly time-resolved spectroscopy from an *isolated* neutron star may be the optimum. These objects are faint (likely candidates have V = 24 or 26), and a search for e.g. their nonradial oscillations (predicted periods in the range of milliseconds or 100's of microseconds) obviously require ELTs. Although, at first sight, it could appear that X-ray observations perhaps would be more suitable, their potential is limited by the limited number of X-ray photons that realistically can be collected over such short timescales by foreseen space instruments. A detailed probing of neutron-star interiors would enable a better understanding of baryonic matter and is, of course, of considerable interest also outside astronomy proper. Other neutron-star related observations include the optical counterpatrts of millisecond pulsars, of relevance for pulsar physics and branches of radio astronomy.
2.3. Quantum optics
All classical astronomical instruments (spectrometers, imagers, photometers, polarimeters, interferometers) measure quantities that can be deduced from the spatial and/or temporal first-order coherence of light. In a quantum description this corresponds to properties that can be ascribed to groups of individual photons. However, light also carries other properties (studied in laboratory quantum optics) such as the second- (and higher)-order coherence, properties that can be ascribed to groups of pairs (or a greater number) of photons. This "entropy" of light (equivalent to the ordering of photons in time) in principle carries information of how the photons have been created (spontaneous or stimulated emission), and how they have since been scattered in angle or frequency. It is possible (at least in principle) to segregate between two otherwise identical spectral line profiles, one where the Doppler broadening has come from those atoms that emitted the original photons, and another where the broadening has come from motions of intervening atoms that have since scattered already existing photons. While such phenomena are measured in the laboratory, their astrophysical observability is not yet known, and the potential thus somewhat speculative. However, it is clear that the signal-to-noise ratio for such measurements improves drastically for ELTs since the signal is proportional to the square of the light-collecting power, i.e. diameter**4 (for two-photon properties) or an even higher power for multi-photon correlations: this would be non-linear optics applied to astronomy.
2.4. Solar-system planets:
A spatial resolution of 1 mas resolves a few km on Jupiter or on Jovian
satellites such as Io, enabling systematic studies of the atmospheric and
surface conditions on all major and many minor planets, and planetary moons.
I believe this capability alone can easily justify an ELT - to obtain the
same with spacecraft would require a very large armada of them. Specific
challenges during the next few decades would include monitoring how the
atmopshere of Pluto is frozen out as Pluto's eccentric orbit carries it
further away from the Sun. Spatially resolved spectroscopy would reveal
how this freezing-out begins at the planet's poles (or does it?). Unfortunately,
it appears that there was really nobody really knowledgeable about planets
among the participants at the recent meeting: since this may be a field
where the ELTs may have their greatest impact, I believe such persons should
be invited to the later meetings.