Amid uncertainty over the fate of the Hubble Space Telescope and with a key instrument not working, NASA Administrator Sean O'Keefe gave the go-ahead Monday for planning a robotic servicing mission.

"Let's go save the Hubble," O'Keefe said. He did not say whether the failed instrument would be repaired.

O'Keefe told about 200 employees at NASA's Goddard Space Flight Center to develop a firm proposal within a year, at which time a decision would be made whether to proceed.

Costs not known

Al Diaz, NASA associate administrator for the agency's Science Mission Directorate, would not say how much NASA expects the mission to cost.

NASA officials have stated previously that a full-fledged robotic servicing mission would likely cost in excess of $1 billion. A recent internal NASA study, according to government and industry sources, estimated the cost at $1.6 billion to $2.3 billion -- several times more than what NASA has spent in the past mounting space shuttle missions to the telescope.

It is not clear where the money would come from.

"That would have to be worked," NASA spokesperson Don Savage told SPACE.com Tuesday. The funding might involve a supplemental request to Congress or could come from shifting funds in NASA's existing budget, Savage said.

However, other sources, who did not wish to be identified, told Space News that NASA plans to ask Congress for supplemental funding when the lawmakers return to Washington in September.

Details of the robotic mission are still to be laid out. But in an encouraging sign for astronomers, O'Keefe asked engineers to consider a mission that would meet all of the objectives for a manned servicing mission that had originally been planned but was since canceled. That mission was designed to replace crucial parts and add two new instruments.

"He's thinking it's looking more and more like we ought to do full robotic servicing," Savage said.

Meanwhile, in a major blow to the Hubble program, prospects look dim for reviving an important instrument that stopped working last week. Its loss would leave a huge hole in astronomers' ability to study certain aspects of the universe, adding urgency to unsettled plans to service the observatory.

'Big loss'

The Space Telescope Imaging Spectrograph (STIS) conked out Aug. 3 when a power converter malfunctioned. People associated with the 14-year-old observatory said engineers were still trying to figure out whether the problem could be fixed remotely, but that early indications were a solution did not appear likely.

A spectrograph does not produce traditional pictures but instead analyzes light to discern a source's chemical compositions and temperature. STIS is the only spectrograph in space that records ultraviolet light, covering a range of the spectrum that can't be observed from Earth. The device made the only discovery of an atmosphere around another star. It has also studied black holes and provided confirmation of the age of the universe.

"It's a big loss," said Paul Scowen, a researcher at Arizona State University. "Without that a lot of science simply cannot be done now, until a new mission flies."

There are no firm plans for any such mission.

With colleagues, Scowen has proposed to NASA a future space-based observatory that would include a UV spectrograph, effectively replacing the STIS. It is one of nine competing proposals under the agency's Astronomical Search for Origins Program.

Scowen called the STIS second only to Hubble's Advanced Camera for Surveys (ACS) in importance to astronomers. The ACS and two other science instruments on Hubble continue to operate normally.

Partial replacement ready

The STIS was added to Hubble by shuttle astronauts in 1997. It was designed to last five years.

A partial replacement and complementary instrument, called the Cosmic Origins Spectrograph (COS), is built and was supposed to be installed during another servicing mission originally slated for last year. The mission was delayed by the Space Shuttle Columbia disaster on Feb. 1, 2003, and it was canceled by O'Keefe, NASA's chief, this January.

After a public, political and scientific outcry and accusations of back-room decision making, O'Keefe decided to consider a robotic mission to service Hubble. Its primary goals would be to replace the telescope's batteries and gyroscopes and to install a module that would eventually be used to safely de-orbit the observatory. If not serviced, Hubble would likely stop working altogether by 2008.

It has not been determined whether a robotic servicing mission could fix the vital STIS instrument or take on the daunting task of hooking up the phone-booth-sized COS.

Scowen is familiar with engineering discussions over the possible robotic servicing mission.

"Installing COS is a very tall pole right now," he told SPACE.com. "It is a proposal fraught with risk."

Astronaut John Grunsfeld, who has twice serviced Hubble and is now NASA's Chief Scientist, did much of the lobbying that led to O'Keefe considering a robotic mission. In an interview in May, Grunsfeld said swapping equipment robotically poses challenges. Astronauts can make adjustments if things get stuck, he said, but on the other hand "robots can do really pure motions."

Leaning toward robots

On July 13, a 20-member National Academy of Sciences commission suggested NASA Administrator Sean O'Keefe reconsider his ban on a crewed servicing mission.

The group's interim report on the matter cited limitations of robotics given the industry's infant stage. It pointed out that four previous crewed trips to Hubble were successful and said safety concerns do not preclude sending astronauts again. The panel urged NASA to commit to the original servicing mission, including installation of COS and another already-built camera called the Wide Field Camera 3 (WFC3), which would replace Hubble's Wide Field and Planetary Camera 2.

(Another of the nine proposals in the Astronomical Search for Origins Program would fly COS and WFC3 on a new observatory.)

O'Keefe praised the National Academies interim report but sidestepped any serious rethinking of his opposition to using astronauts. He has called such a mission too dangerous to ponder in the wake of the Space Shuttle Columbia disaster and subsequent investigations and recommendations made by the Columbia Accident Investigation Board.

While he hasn't ruled out a shuttle mission, a spokesperson said last month, O'Keefe is focusing on the robotic option.

"Everybody says, 'We want to save the Hubble' -- well, let's go save the Hubble," O'Keefe told The Orlando Sentinel after Monday's meeting with employees.

O'Keefe said it would take nine months to a year to formulate a solid mission plan from various proposals that NASA has been studying in recent weeks.

Fix STIS?

Engineers are probably a couple of weeks away from determining whether or not the power supply to the defunct STIS can be restarted from the ground. Importantly, the power unit was not designed to be serviceable, so any mission to the telescope -- robotic or human -- would face an as-yet undetermined project if officials decide to resurrect the STIS.

Savage, the NASA spokesperson, said it's too early to decide if a robotic mission could fix the STIS, as "engineers haven't been able to determine what caused it to fail." The possibility of repairing it would have to compete with other aspects of the robotic mission.

When asked if a manned shuttle mission to Hubble was still being considered, Savage said, "We're not doing anything to preclude doing that."

"Although it is too early to tell if STIS can be serviced, the astronauts have been remarkably versatile on previous missions for fixing instruments not designed to be serviced," Steven Beckwith, outgoing director of the Space Telescope Science Institute, which operates Hubble for NASA, said over the weekend.

In one example of astronaut capability, a shuttle mission in 2002 revived Hubble's infrared camera by installing a new cooling system.

Engineers have for months been testing robotic equipment for a possible automated effort, controlled from the ground, to repair or upgrade Hubble.

O'Keefe told employees Monday the robotic mission would cost up to $1.6 billion, according to press reports, but that a firm estimate could not be calculated until a final plan was developed.

It is not clear, however, if NASA will even get the basic funds it has asked for next year. Despite a White House-backed request for more money than last year, to support President Bush's vision of sending astronauts to the Moon and Mars, NASA's budget stands to be trimmed $229 million in the latest version in Congress.

Urgent need

If STIS is not fixed or replaced, one key loss will be the ability to examine atmospheres of extrasolar planets.

The spectrograph was used to discover and then characterize the atmosphere of HD 209458b, a planet orbiting a distant star. Astronomers termed it a shrinking lava-like world.

The compelling observations were possible only because HD 209458 transits directly in front of its host star, as seen from our vantagepoint. Astronomers expect as many as 10 similarly configured planets to found over the next five years, Beckwith said. "STIS will be the only way to analyze the atmospheres."

David Charbonneau, who worked on the team that probed HD 209458b, agrees.

"If the decision not to carry out the final servicing mission for Hubble is reconsidered, and if it is possible to repair the failed power supply, then of course this would be just fantastic," said Charbonneau, formerly at Caltech and now with the Harvard-Smithsonian Center for Astrophysics.

Charbonneau said the case for a Hubble repair mission was already compelling. The loss of the STIS has "made that discussion all the more urgent," he said via e-mail.

"It is a big task to weigh the unparalleled science that the Hubble Space Telescope can achieve with the risk and cost of a servicing mission, Charbonneau said, "but I certainly hope that the decision can be considered in much greater depth than has been presented to date."

Space News reporter Brian Berger and the Associated Press contributed to this report