Ever since Galileo first looked through an eyepiece at Saturn, astronomers have puzzled over the night sky. With each successive increase in lens or mirror size our horizon of knowledge has widened, while Earth's relative primacy within the universe has astronomically shrunken. In the near future, as Extremely Large Telescopes (dubbed ELTs) become operational, we may finally be able to answer some of the most perplexing questions of cosmology, although it's also likely there will be new surprises in store, initiating questions no one has yet thought to ask.


To discover what's driving the development of ELTs, just ask any astronomer what they wish they could do: Image rocky exo-earths to study their spectra for bio-indications of water and oxygen. . . detect the earliest supernovae as a means of gauging when and how the first stars formed. . . develop testable theories on the nature of mysterious dark energy. . . study individual stars in other galaxies--like low metal stars in ellipticals--to shed light on how galaxies assemble. . . map the kinetics of high redshift galaxies within dark matter haloes to determine their structure and possible nature. . . discover if distant dwarf galaxies show early indication of massive black hole development.
   The list goes on. With so many new questions, it's no mystery that observational astronomers need ELTs in much the same way that physicists need larger particle accelerators. Primarily, they hope to overcome the "confusion limit" inherent in poor angular resolution. This is true because, while the sensitivity of light-analyzing detectors is nearing 100% efficiency, an image's clarity or optical resolution depends chiefly upon the number of photons that can be gathered to a common focal point. Meaning advancements in knowledge can now only come from an increase in the light gathering ability of larger mirrors.  
Of course orbiting telescopes like Hubble, Spitzer, and Chandra will continue to provide stunning images, due to their lack of atmospheric aberration. But even future STs, (like the James Webb Space Telescope, slated for launch in 2015), will be more useful in wavelengths other than the visible, and won't be able to match the resolution to be achieved by the more massive ground based mirrors coming, once adaptive optics (AO) achieves the same technical advances that spectrographic instruments already possess. 


Get ready, the monsters are coming, although not from outer space. Here's a peek at the three largest telescope projects now on the drawing boards.  
   * The 24.5 meter Giant Magellan Telescope (GMT) will consist of seven massive 8.4 meter mirrors, just like the two currently employed on the Large Binocular Telescope. These mirrors, to be cast from a design by Roger Angel of the University of Arizona, will utilize a honeycomb design to provide the lightest possible weight while maintaining stability against image degradation when tilted. The result? When phased together, GMT will produce 11 times the light gathering power of its predecessors--(the two 6.5m Magellan scopes)--and will be 130 times optically faster in detecting faint objects.  
   Once deployed in Chile during the projected 2010-2016 construction and commissioning phase of the project, both spherical aberration and atmospheric diffraction will have been nullified by corrective elements, including an adaptive secondary mirror. With a mount rigid enough to withstand wind-caused tracking deterioration, and with new computer algorithms to facilitate total telescope control, steering of the behemoth will be facilitated by both radically new technologies, and a revolutionary scaling up of many previous mechanical systems. The result? At completion, the GMT (together with two other ELTs pursuing similar timelines) will usher astronomy into a new era of observational performance and discovery.  
   * The Thirty Meter Telescope. When TMT opens its huge 492 segmented eye on the universe sometime after the middle of the next decade, it will have nine times the light collection capacity of a single Keck scope, and three times the spacial resolution. Some of the same Keck designers are at work here, too, in a project that consolidates three independent concepts (the California ELT or CELT, the NOAO's Giant Segmented Mirror Telescope or GSMT, and the Canadian Very Large Optical Telescope or VLOT). Among the primary goals is to penetrate, with unprecedented precision, the most distant horizons in near infrared wavelengths using thinner and smaller mirrors of shorter focal length, (similar to Keck), yet scaled up into such a mammoth matrix that TMT will be able to image objects nine times fainter than Keck in the same time frame, and with ten times the spacial resolution of Hubble.
   TMT's fast f/1 segmented primary mirror will be complemented by a concave Gregorian deformable "active" secondary, able to correct for subtle wavefront and mirror stress aberrations. An articulated tertiary mirror will then direct the telescope's corrected image to a suite of instruments positioned on stable Nasmyth platforms surrounding the giant azimuth structure. These instruments--needed for specialized study--will also benefit from adaptive optics systems designed to cancel the effects of atmospheric turbulence, and will include an infrared multi-object wide-field spectrograph (IRMOS); a PFI or Planet Formation Imager, (for high contrast coronagraphic photography); and NIRES, a sensitive near infrared Echelle spectrograph, fed by a high tech NFIRAOS AO system that accurately gauges and compensates for light wavefront instability. While some of the technologies to be employed are still under development, the hopeful result will be the world's first 30m Earth-bound telescope capable of eliminating the twinkle blurring the stars from the ground.  
   * Finally, there is the E-ELT, short for European Extremely Large Telescope. Its gargantuan 42 meter mirror will consist of 906 hexagonal segments, making it 100 times more sensitive than Keck. Its concept combines the contributions of more than a hundred astronomers in the ESO (European Organization for Astronomical Research in the Southern Hemisphere), adapting earlier designs like the OWL (OverWhelming Large) and Euro-50, after cost parameters were delineated. Site selection will be determined in late 2008 (as in TMT's case), with estimated "first light" being 2017+. "A telescope of this size could not be built without a complete rethinking of the way we make telescopes," says Catherine Cesarsky of the ESO Project Office.  
   E-ELT's 6m secondary mirror is as large as many of today's primary mirrors, and will feed a 4.2m tertiary, relaying the beam to an AO system that includes a 2.5m active mirror (possessing over 5000 actuators able to distort its shape 1000 times a second), and another 2.7m mirror configured for final image correction. Once deployed at a suitable high and dry elevation, the E-ELT will open its gaping maw to the sky, and eat more photons than any sci-fi Godzilla ever could gnats.


Under current developmental models, next generation telescopes become most scientifically justified at twice or more their previous size. This naturally strains the economic rationale behind their construction, making funding the #1 obstacle to proceeding beyond the planning stage. While advancing science is overcoming many technical challenges, getting the money to actually build huge observatories requires dizzying collaborations between multiple institutions, both public and private.  In the case of GMT, the main project partners include The Carnegie Institution, Harvard, MIT, Australian National University, Smithsonian Astrophysical Observatory, Texas A&M, University of Arizona, University of Michigan, and the University of Texas at Austin. For TMT, support comes from the Moore Foundation, the National Science Foundation, and several Canadian agencies, along with its partnership institutions, Caltech and the University of California. Still, the myriad tasks required just to complete the instrument design phase are daunting, and are "being paced by technological readiness and financial resources," according to their Science Advisory Committee. Then there are the "trade off studies" in progress within the projected billion dollar E-ELT project budget that pit cost to performance in the areas of optical layout and instrument platforms as they relate to the five major spectrometers and imagers to be utilized in just the first phase of operations.
   Beyond funding, and among the biggest technological challenges to overcome, is the immense computer power required to control the extremely complex AO systems that represent the forefront of R&D for ELTs. Whether this problem will be solved at the software level (in addition to faster supercomputers) remains to be seen. As for the AO instruments themselves, they are marvels of engineering. Particularly MCAO, or Multi-Conjugate Adaptive Optics, a system that utilizes multiple deformable mirrors phased to laser beams that project artificial "guide stars" into different layers of the mesosphere (90-160 km high) via fluorescence of the sodium atoms present there. These stars, as with natural stars (typically useful for only 5% of the sky), are then tracked by wavefront sensors to calculate and compensate for phase aberrations resulting from the varying turbulence throughout the three dimensional volume of air above the telescope. Little wonder that a mere laptop hasn't a prayer of crunching those numbers!  
   MCAO, currently being employed on the Gemini scopes in Cerro Pachon, Chile, possess the ability to overcome the cone effect and anisoplanatism (wavefront errors) characteristic of previous systems, and provide diffraction-limited quality images over a one arc minute field of view. But although there is a tenfold gain in optical resolution from this, MCAO systems cannot solve all problems, and have problems of their own, too. According to Francois Rigaut, a senior scientist at the Gemini Observatory, "MCAO is not suitable for all needs.  While it provides a good Stehl ratio (related to image quality) over a moderate field of view, things like planet detection will require using bright natural guide stars in addition to manmade laser guide stars.  Or for integral field spectroscopy over larger fields, you'll need Multiple Object Adaptive Optics, or MOAO, where only a few arc seconds are corrected over a wider field than MCAO." Adds Jerry Nelson, (Center for Adaptive Optics, UC Santa Cruz), "The adaptive optics challenge grows with the size of the telescope. Building sodium lasers for ELTs, and collecting the light, and processing data to make a tomographic map of the atmosphere, then correcting the wavefront errors with deformable mirrors. . . all are very serious challenges."  
   Such challenges extend across the spectrum of the institutions involved in ELTs, from science programs that utilize GLAO, or Ground Layer Adaptive Optics, (optically conjugated so that only the lower level of the atmosphere is corrected), to ever more complex systems now under design to solve other "air quality" obstacles, as when astrophysical constraints require hypersensitive photometry to find rocky earth-like planets that are currently undetectable.


In science fiction you can scale up a monster by using visual effects, or with a zoom lens. In the real world of science, scaling anything up requires redesign of the very tools of manufacture, with a sharp eye on tolerances, power mechanisms, metal fatigue, and a host of ancillary sticklers. Simply enlarging the mirrors of monster ELTs by adding more and bigger shiny "scales" is not all that's required. Nor is merely aligning a new generation of instruments to harvest their reflection. Radical new technologies must come to play that drive and monitor telescope and facility control. Only by keeping the beast on a short leash will it comply to commands that aim to answer questions that require 10 milli-arc second resolutions, or better.
   Of course this dilemma is not new. In scaling up engineering systems and software programs for just the current generation of new telescopes, Norm Cushing, chief software engineer on the Large Binocular Telescope, required that his team write low level embedded software in assembly language, including some higher level software modules running in Linux that talk to each other through reflective memory. "New algorithms for right ascension and declination had to be created just to steer the telescope and make it move to position and track objects," Cushing explained. "The lower level commands that control the motion of the building to follow the telescope were all new, too, written from scratch for this system."
   What drives teams similar to Cushing's, now, is both scientific cooperation and a competitive team spirit, knowing the timing is critical for making monumental discoveries with instruments of monumental size. For instance, ELTs will be crucial in doing follow-up studies of discoveries made by the future James Webb Space Telescope, when their uniquely high angular resolution will enable a finer order of spectroscopy than can be obtained from space. Explains Francois Rigaut, "JWST will go extremely deep, but deep doesn't necessarily mean you have enough photons to get good signal-to-noise spectra. So the scenario is that JWST discovers objects, thanks to its large field and low background, then ELTs follow up and make spectra of these objects, to investigate their physical properties. This high angular resolution is important in two respects. First, because it will allow astronomers to make spectra at a resolution similar to the JWST discovery images, so you can compare oranges with oranges. Then, because most of these objects will probably be clumpy, an ELT's higher angular resolution will concentrate the light in fewer pixels and boost the signal-to-noise ratio, allowing it to go fainter. Otherwise, because high redshift galaxies are about 1 arc-second in size, you would only see a blob with seeing-limited telescopes, and be unable to make any accurate study of the galaxy's dynamic or structural composition."
   In short, although bigger may not always be better in other realms, in ELT astronomy it had better be better, to justify the astronomical costs involved. This goes not just for instrumentation, mounts, platforms, computers, cranes, enclosures, or mirrors, but even mirror coatings. As another example of the complexity of support operations, the re-coating of mirrors is daunting, considering the size and number of segments involved, and their delicate handling. Huge vacuum tanks must be employed to seal atomized aluminum onto glass, once the old coating has been stripped away by a weak acid solution. To get certain desired effects, even the coatings themselves will come in several flavors, too. While aluminum has been the standard element, specific scientific goals now require an alternate material that possesses higher reflectance, particularly in the optical range of 0.3-30 microns (part of the visible/infrared spectrum).  Since aluminum doesn't perform well in this wavelength range, one team of scientists at TMT is working to develop a multi-layer coating with higher reflectance and lower emissivity (tendency to give off its own infrared radiation).  With costs linked to longevity and resulting down time, coatings made at the Pacific Northwest National Laboratory will be tested at Lick Observatory with a spectro-photometer before being placed in telescope dome environments to study long term quality degradation.


So how much bigger and better can ELTs get? Sandra Dawson of TMT's public relations office predicts, "The next level of telescope size may be an extension of the finely segmented mirror technology pioneered by Keck, employed by HET, GTC and SALT, and expanded by TMT. This technique naturally scales to a telescope perhaps three times the diameter of TMT, as already studied carefully in the European OWL study." However, as Jerry Nelson postulates, "Eventually the largest telescopes will confront impossible difficulties from gravity, and while it may be possible to make a useful telescope on the ground at the scale of 100m, much beyond that the largest telescopes must be in space, or not at all." 
If ELTs do move beyond Earth, it's anyone's guess as to their size or composition. Perhaps a slowly rotating liquid mirror the size of the Superbowl will be needed to finally answer the questions to be raised by GMT, TMT, and E-ELT. Questions, perhaps, involving the nature and influence of cosmic strings left over from collisions between multiverse "branes," of which our own universe is postulated to be just one. Only the future knows the exciting discoveries to be made, once we have opened our eyes wide enough to penetrate the darkness that now awaits. -0-  


The article on this page was originally published as the cover article in Sky & Telescope, April 2008. 
An article on the Large Binocular Telescope in Tucson, also in Sky & Telescope
Seeing Limited VS. Diffraction Limited
If the goal of ELTs is to achieve diffraction-limited performance, what exactly does that mean? "ELT instruments come in two basic types," Jerry Nelson explains. "Seeing-limited instruments observe the sky without making any adaptive optics corrections. Diffraction-limited instruments are designed to accept image sizes limited by diffraction. A diffraction-limited image is therefore as small as it can be, limited by interference effects due to the wavelength of light. Since adaptive optics is difficult--and the difficulty grows rapidly with shorter wavelengths--usually seeing-limited instruments work in the region 0.3-1.0µm. At longer wavelengths we expect that adaptive optics will be able to significantly shrink the image size relative to the seeing-limited size. Then, with residual wavefront errors below about 10% of the wavelength of light, we can achieve diffraction-limited performance, as with instruments that work in the infrared." (1-30µm)

Novel featuring an optics engineer who witnesses the first naked eye supernova in 400 years.
Astronomy books
Google Play