Submillimeter Camera Is Under Development

Hank Hogan

There's more to the night sky than meets the naked eye. Some of the most scientifically fascinating objects and processes aren't visible at all, prompting an interest in submillimeter astronomy, in which imaging is performed at wavelengths of 200 µm to 1 mm. This is the spectral domain of cold objects such as early galaxies, stars and planets, and is the purview of a new camera project dubbed Submillimeter Common User Bolometer Array 2.

With the successes of the original array, which is used at the 15-m-diameter James Clerk Maxwell Telescope atop Mauna Kea on the Big Island of Hawaii and which has had an impact second only to the Hubble Space Telescope, astronomers have desired a better tool to image through the atmospheric spectral windows at 450 and 850 µm. "The telescope already exists," said Wayne S. Holland, a researcher at the UK Astronomy Technology Centre in Edinburgh and a principal scientist on the project. "We are building an innovative new camera for it."

A completed 40 × 32-pixel subarray for the Submillimeter Common User Bolometer Array 2 is ready for laser dicing. Courtesy of Eric F. Schulte.

The prototype detectors have been constructed for the second version. The current schedule calls for the instrument to be delivered to the Maxwell Telescope in early 2006, but the work hasn't been easy.

"This project is a real challenge for everyone," said Eric F. Schulte. He was a senior principal engineer at Raytheon Vision Systems in Goleta, Calif., and worked on this project until recently. He's now an independent consultant.

The challenges arose because Array 2 will be the first instrument to use superconducting transition-edge sensors in a large CCD-like focal plane array. These sensors should boost the rate at which astronomers can map and image the submillimeter universe by a factor of 1000 compared with that of the original setup.

In operation, the detector hovers at the threshold temperature of approximately 200 mK, at which the material making up the sensors, a molybdenum/copper bilayer, ceases to be a superconductor. When a submillimeter photon strikes an element of the detector, it produces a change in resistance. This signal is amplified and multiplexed with those from the other pixels to generate an image. The final detector comprises four close-butted 40 × 32-pixel subarrays that are end-butted into an 80 × 64 large-area focal plane. The optics of the telescope direct the photons onto the sensor array.

The 1.135-mm pixels are constructed on 75-mm-diameter wafers fabricated at the National Institute of Standards and Technology in Boulder, Colo., and at the University of Edinburgh. Each wafer is bonded to a mechanical waffle that provides support and that acts as an optical channel. The superconducting amplifier and other circuitry are built on another wafer, which is then bonded to the other two by Raytheon using equipment from Suss MicroTec SAS in Saint-Jeoire, France.

When the bonding takes place, the alignment of the structures on one wafer to its corresponding component on the other must be better than a micron, and this must be maintained as pressure is applied. Holding that alignment is a challenge, particularly given the size of the wafers.

"Parallelism of the bonded die is very critical when you're trying to align large components covered by very tiny bumps," explained Gilbert Lecarpentier, international technical product manager for device bonders at Suss MicroTec.

The sensors will be used to look at the universe in a cold light. The sensor technology, however, may have other uses in astronomy, such as imaging extrasolar planets, as well as in earthly applications such as drug discovery and spotting semiconductor contamination.