Once light reflects off the actively controlled mirrors, it is further enhanced at an adaptive optics filter, which also employs a deformable mirror and works to counter errors, such as high-altitude thermal changes and atmospheric turbulence, that occur more rapidly than those the active optics can correct. These conditions are monitored by focusing on a single guide star; in response, the computer orders changes in the mirror's shape 100 times per second. Though several observatories around the world, most notably the Subaru Telescope atop Mauna Kea, Hawaii, are applying or developing active and adaptive optics, the VLT is by far its largest application, and the ESO was one of the first organizations to develop the technology. "When we were developing the active optics system in the 1980s, nobody was sure that this would work at this scale, and many people actually opposed the idea," Tarenghi recalls.
The other remarkable achievement taking shape on Paranal is the VLT interferometer. This setup, common in radio astronomy, capitalizes on the fact that if a telescope has a larger aperture, it can collect more of the light from objects in space. Interferometry enables the four telescopes to focus on the same object and gather its light as if the group were a single telescope as big as the combined distances between the individual telescopes-in the VLT's case, a distance of 426 feet.
Light beams collected by the four telescopes are deflected by mirrors into underground tunnels, where they are gathered at a single sensor. The sensor generates an image that is a cumulative product of the four beams. The trick is getting the light waves to meet at the sensor at the same time: As objects are tracked, the telescopes' relative positions change. Consequently, in the tunnel, the light is bounced off several retro-reflectors sitting on small, precisely positioned rail carts that move on 200-foot tracks to compensate for these changes. In addition, three small auxiliary telescopes, also moveable, on the surface will fill in the spaces between the four UTs to further punch up the resolution.
Final processing will prove that the VLT is far greater than the sum of its parts, with an angular resolution of 0.001 arcsecond. (The "celestial sphere" around Earth is 360 degrees; the full moon has an apparent size of 0.5 degree; a degree has 60 arcminutes; an arcminute has 60 arcseconds.) The Hubble, which sits above the atmosphere but has only a 2.4-meter mirror, can resolve to 0.1 arcsecond. The VLT's resolution is fine enough to capture detailed images of distant galaxies, clues about the chemical and biological composition of extrasolar planets-and snapshots of lunar rovers.
Putting all this to work means long nights on the mountain. In the evening, the base camp is pitch dark (no exterior lights are permitted, with the exception of a few dim safety lights) and increasingly silent as the lively chatter and music coming from the dorms of the ESO's 150-plus Chilean and European engineers and administrative and support staff gradually taper off. But atop Paranal, in the control building-already decorated with posters of some of the more spectacular images captured through UT1-a steady buzz of activity lasts until dawn. On average, the observing conditions (the "seeing") are considered excellent 350 days a year, a number envied by most observatories. On those evenings, the telescopes and their attached instruments work feverishly to flush every photon of light-borne information out of the sky.
The scheduling of science operations at the VLT is controlled at the ESO's headquarters in Garching, Germany. Astronomers compete for observing time, and if their proposal is accepted, they can travel to Chile to supervise the session themselves-as might be necessary for complicated or variable-dependent projects-or request that the VLT's staff astronomers, such as Marconi, conduct the program on their behalf. Marconi also operates the VLT for visiting astronomers, so that they don't lose time struggling with the technology.
The scientific programs Marconi helps execute are challenging, chosen to push the VLT's capabilities as far as possible. The results are sometimes breathtaking. "We made an observation three weeks ago which is particularly alive in my mind," Marconi recalls. "The target was a jet of material ejected from the famous active galaxy 3C273. While looking at the details obtained in the images, I forgot for a moment that I was on the ground."
The VLT will undertake a variety of scientific projects, including measuring the universe, studying galaxy structure and formation, and observing star birth and planetary system formation. Director Gilmozzi is confident that the VLT will push astronomy even further. "There's a final question that astronomers are asking, and that's whether we are alone in the universe," he says. "We will soon begin to ask how to detect biospheres and ecospheres on extrasolar planets, and the VLT will make those searches possible. I'm also sure that over the next several decades we will discover an enormous class of objects of which we have no hint today."
Along the way, ESO hopes to develop new strategies for studying astronomical phenomena. Instrumentation plays a key role in this. The VLT will initially be outfitted with 11 instruments-some as big as a room-capable of wide-range spectroscopy and other types of imaging. Some of these instruments have already paid big dividends for astronomers using UT1. Munich University's Rolf-Peter Kudritzki spent four nights at Paranal late last year using the VLT's FORS-focal reducer/low-dispersion spectrograph-to examine a newly discovered population of stars floating in the space between the galaxies of certain clusters. Though Kudritzki's team came well prepared, the members encountered some surprising results and consequently had to alter their program on the fly. "Many of our objects turned out to be completely different from what we expected," Kudritzki says."We detected a new class of extreme-emission-line galaxies at very high redshift." High redshifts indicate that the galaxies are moving away from our own very quickly and are at the edge of an expanding universe; because they are so far away, their light took billions of years to reach Earth, so we are seeing them as they were when the universe began. "I knew on the spot that I had something new and I was very excited about that," says Kudritzki.
Peter Barthel, an astronomer at the Kapteyn Institute in the Netherlands, spent several days at Paranal in late January. Collaborating with Willem de Vries of Lawrence Livermore National Laboratory in Berkeley, California, and Chris O'Dea of Baltimore's Space Telescope Science Institute, Barthel investigated distant galaxies that appear to be harboring small, young radio sources-potentially other galaxies. "The VLT data will tell if our ideas about these young radio galaxies make sense, and fortunately the data we gathered were excellent," Barthel says. "We got a number of new identifications with very faint galaxies. Though we still have substantial data processing and analysis left, we can, from the raw images, already see the faint host galaxies and study their spectra."