Skip to content.

GMT

Metrology Rules!

Scientists at the University of Arizona's Steward Observatory mirror lab continue to develop and implement exacting tests to make the GMT a success. Recent photos taken at the mirror lab feature the newly aluminized 3.8-meter diameter "fold sphere" mirror that will be used for testing the large GMT primary mirror segments. Preparation of the Test Tower also continues.
Tucson, AZ.  As grinding and polishing is performed on the first mirror segment for the Giant Magellan Telescope, work continues on systems designed to measure and test the mirror's accuracy. This is the field of optical metrology, and it is one of the most challenging aspects of making the GMT.

The first of seven GMT mirror segments is being ground and polished with a computer controlled deformable lap, called a Stressed Lap. The tool precisely changes its shape as it moves across the mirror's surface. Photo by SOML.

Due to the mirror arrangement in the GMT, each mirror has to be highly aspheric – that is, the mirror’s face has a steeply curved slope. One side of the mirror is thicker than the other, differing by as much as 14 mm.

The difficulty of shaping each mirror segment is compounded by the fact that these large segments must have the exact same curvature in order to perform together.


To achieve the stringent demand for accuracy, each mirror is periodically taken off of the polishing machine and carefully measured. The results of these measurements in turn guide the polishing program as it progresses.

To measure the mirror’s surface and confirm that it is polished into the correct shape, several tests are used: a principal test and several verification tests.

Using a laser interferometer, the principal test must first compensate for the unprecedented asymmetry of the mirror's surface.  Instead of measuring the mirror directly, a beam is first bounced off of two mirrors at oblique angles, and then passed through a computer generated hologram.  With the aspheric departure removed, interferometric surface measurements may be taken.

In theory this is complicated, and in practice it is even more difficult: one of the bounce-off mirrors, called fold spheres, must be quite large: 3.8 meters in diameter. And it must be positioned a considerable distance away from the primary mirror segment. The Test Tower protruding above the mirror lab actually had to be rebuilt to support this large fold sphere and accommodate the width of the interferometer beam.


This 3.75m (147.6 inches) diameter mirror is called the Large Fold Sphere, shown here having just been aluminized at the Kitt Peak Observatory facility. Mounted near the top of a 25m (82 feet) tall testing tower, this mirror is designed to reflect (or, fold) a beam of laser light down onto the primary mirror segment. Photo by SOML.

The fold sphere will be mounted near the top of the test tower. Before risking lifting the mirror itself, a trial run is done lifting the mirror’s mounting cell only. (left) a crane used to lift the mirror cell into the tower is seen outside the mirror lab adjacent to the University of Arizona’s football stadium. (right) the mirror cell is being hoisted into position. Photo by SOML.

Looking up the tower, the mirror’s mounting cell is seen arriving near its final position near the top of the tower.
Photo by SOML.

Three other methods are used to verify the mirrors shape and direct the polishing activity:

  • The scanning pentaprism test is an independent measurement of low-order aberrations that guards against the possibility of a mistake in the implementation of the principal test.
  • The laser tracker measurement supports surface generation and loose-abrasive grinding processes by providing an independent measurement of radius of curvature and astigmatism.
  • The shear test, performed by displacing the segment by about ±0.5 m around the optical axis, will confirm the accuracy of the principal test on small scales.

For information regarding the science and implementation of mirror metrology, refer to the following publications:

printer-friendly pdf version of this article