On the sixth day of the final space shuttle visit to the Hubble Space Telescope, astronauts John Grunsfeld and Andrew Feustel stepped into void to begin another long day of Hubble Space Telescope repairs. At the top of their agenda on May 16, 2009 was installing James Green’s Cosmic Origins Spectrograph, or COS.

Green, the University of Colorado astrophysicist who had designed the spectrograph, was the $70-million instrument’s principal investigator. Shuttle-related delays, most profoundly the Columbia tragedy in early 2003, had kept COS grounded for five long years past its planned date with Hubble.

Green had first considered building a Hubble instrument in 1996, when he was still busy with an ultraviolet spectrograph of his design for NASA’s Far Ultraviolet Spectroscopic Explorer spacecraft. The space agency issued an announcement of opportunity for Hubble instruments. He sat down with a colleague and decided he had just the thing for the Great Observatory.

Embodying lessons learned from FUSE, Green figured he could build a spectrograph 20 times more sensitive in the far ultraviolet than the next best thing out there – or not-quite out there – the Ball Aerospace-built Space Shuttle Imaging Spectrograph, or STIS, slated for installation in 1997. There would be a tradeoff: COS would have lower spectral resolution than STIS. But given the science data COS was to collect, sensitivity would be paramount.

COS would tease out the composition of enormous clouds of gas wafting about the interstellar medium – thought to hold perhaps 60 percent of the universe’s periodic-table matter but still little understood. The instrument would do this through absorption spectroscopy, using quasars far beyond the Milky Way like cosmic flashlights to pierce the fog. Not knowing how far away in space and time a given cloud might be, COS would capture a wide expanse of light with each exposure to account for different ages and distances of the intervening clouds, as manifested in varying redshifts.

Green wanted to build three diffraction gratings mounted on what looked like an 18-inch-diameter paddlewheel. When rotated into position, each would be capable of capturing thick slabs of neighboring UV spectrum (from 1,150-2,050 angstroms, all told) with a hyperbolic concavity squeezing spectral lines into the sharp points preferred by the intended detector, a University of California Berkeley model nearly identical to the one to fly on FUSE. Each grating would also correct for the spherical aberration in Hubble’s primary mirror. The combination allowed COS to massage light coming from telescope in a single bounce, avoiding losses inherent each time optics steer UV light (in STIS, built before the technologies to develop such an all-in-one grating existed, light takes four bounces).

With a plan in hand, it was time to look for an industrial partner. It was an easy decision.

“With this one, it was pretty obvious that Ball was the only way to go,” Green said.

The proposal demanded Hubble-specific expertise ranging from connectivity to compatibility, and Ball had either built or was slated to build five of the seven instruments that had been on Hubble to date – including the COSTAR corrective optics that helped restore the telescope’s blurred vision in 1993. Plus the company was right up the street.

“Clearly their credibility as an industrial partner would be unquestioned, and they also have a history in UV,” Green said.

Seven of the eight PIs proposing Hubble instrument had the same idea, it turned out, with Ball supporting them all. Green’s Cosmic Origins Spectrograph prevailed.

David Lekrone, NASA’s senior project scientist for Hubble, had said COS has the potential to acquire UV information “farther out across the universe than we’ve ever been able to do before,” thereby helping answer questions about “the global cosmic process of how you form the large-scale structure of how material is distributed in the universe and what role that played in forming new galaxies.”

NASA threw the COS team an immediate curveball, asking for a near-ultraviolet channel (1,700-3,200 angstroms) to complement the far-ultraviolet channel proposed, which meant adding a second grating-laden paddlewheel mechanism as well as a second detector – as it turned out, one originally built for STIS. NASA also upped the instrument’s expected lifetime from three years to seven years, triggering a wholesale switch to more radiation-hardened electronics that added $10 million to the mission budget.

As they began work, Green saw that he was, in his words, “a bit of an odd bird” to Ball Aerospace.
“They were used to a PI coming in with limited hardware experience but big on science,” Green said. “They might say, ‘I want an instrument. It has to have this much resolution, it has to have this much field of view, it has to have this much sensitivity. Can you build it?” In short, the PI’s job was to sell the mission and Ball’s was to build the hardware to make it happen. Not so with Green.

“I said, ‘I’m doing the optical design. I’m going to procure the gratings. I’m going to procure the detector – Berkeley’s going to build the thing. Berkeley’s going to talk to me, Berkeley’s not going to talk to you,’” Green recalled.

To meet the instrument’s exceedingly stringent needs, Green went so far as asking Ball to change clean-room procedures the company had honed over years of instrument building.

“There were several times when we went in and said, ‘This is the better way to do things,’” Green said. Each time, Green said, Ball engineers considered the request and changed their approach to suit the mission.
But other times, Green relied on Ball Aerospace’s experience and expertise. Ball built the instrument’s mechanisms and most of its electronics, bonded the glass, and wrestled with formidable thermal issues that cropped up due to a formidable recycling effort. In the proposal, the team had suggested saving money by reusing the optical bench – a telephone-booth-size house for optics – from the Goddard High Resolution Spectrograph, a Ball-built Hubble instrument brought back in 1997.

COS would need a different configuration, which meant removing stabilizer bars and shear plates and adding new ones elsewhere. But the flavor of graphite epoxy in the old spectrograph housing could no longer be bought. The new materials had different coefficients of thermal expansion and moisture expansion, which complicated Ball’s thermal modeling and testing, Green said.

As Green managed the project and monitored the efforts of optics subcontractors in France and the United States, Ball was also working behind the scenes to navigate COS through the enormous enterprise of a space-shuttle launch. Green said he first understood how much his industrial partner had been doing behind the scenes during a ground-operations meeting at the Kennedy Space Center.

“I find the conference room, I walk in, and there must be 200 people there. I know five of them, and all of them know about COS. They’ve been writing documents about COS,” Green recalled. “And Ball was managing that whole effort and all those interfaces for me just as part of the routine job. They were doing a whole bunch of things I didn’t even know had to be done.”

Years – and countless meetings – later, astronauts Grunsfeld and Feustel removed the vestigial COSTAR (vision-correction had been built into newer Hubble instruments) and slid its replacement, COS, into place, where its scientific work will begin in the coming months. With the installation of COS and the new Wide Field Camera 3, Ball had built every instrument on the great space telescope.

“Certainly from my perspective, and I hope they would say the same, I though it was really a collaboration more than a partnership – and a highly successful one at that,” Green said.


Fact Sheet

Corporate Executives


Employee Profiles-Diversity

Employee Resource Groups

STEM Programs

Educational Resources


Contact Information

Science at Ball Aerospace

PI Perspectives

Employee Profiles- Science

Professional Papers


Small Business Opportunities

Supplier Survey

Supplier Diversity

Small Business Innovation Research

Supplier Terms & Conditions

Ethics Handbook

Investor Relations

Ball Corporation

A-Z List of
Programs & Products