Time To Cast Another Mirror!

Polishing mirrors, particularly the behemoths we make at the Steward Observatory Mirror Lab is a slow, exacting process. While we work long hours, in fact, we're up to 3 shifts a day now, it still takes years to finish a mirror surface to the accuracy of nearly a glass molecule! But the casting of the mirror substrates is more a matter of engineering and science. We've got a fantastic casting crew that can push out the nearly 30 foot diameter substrates nearly once a year. Our last casting was a year ago January, and we're about to start another in a few days. This one is the 3rd mirror substrate for the Giant Magellan Telescope (GMT), and yesterday was the loading of the glass into the telescope mold.


Few things are more photogenic that watching the loading of 20 tons of gleaming glass chunks into the mold.  The E6 glass is a Pyrex equivalent, made for us from Ohara Optical in Japan.  They make it in small batches of a ton at a time, and cleave it into the chunks seen so that they will melt together in the substrate without trapping air bubbles.  And while these days we could obtain glass with nearly zero coefficient of thermal expansion(CTE), we use borosilicate glasses which become fluid enough to melt and run into the mold.  Ultra-low CTE glasses do not become fluid enough to do that.

The process started a couple weeks ago when the glass was brought in from our off-site storage.  Each block was inspected under crossed Polaroid filters to look for internal stresses.  While Ohara goes to considerable effort to maintain precise uniformity from batch to batch, occasional impurities result in internal stresses that do not get relieved by re-melting.  In the crossed-Polaroid test, stress birefringence rotates the plane of polarization allowing it to become visible.  In the photo at left, a seed, a speck of impurity introduces a high-stress point (circled at lower right).  If these are close enough to the block's surface, the glass can be re-cleaved to remove it and the rest of the block can be used.  Besides pulling out the small percentage that is rejected for high stress as above, it is also graded for quality.  The best glass is layered on top of the mold last so that the mirror faceplate will contain the highest quality.

The actual loading is a day-long Herculean effort by the crew.  They usually recruit a couple volunteers from other areas of the lab to lend a hand.  The shot at left shows a panorama of the operation.  A forklift brings up pallets of the inspected glass blocks at left, and a crew loads them into the inclined rollers to hoist them to the 3-man crew placing the individual blocks into the mold.  Another worker removes and collapses the empty boxes for recycling.  Peeking into the oven, you can see some of the resistive heater elements lining the oven interior, and the Inconel bands that encase the tub walls of the mold.  The crew had started loading about 5am and these pictures were taken about 6+ hours later.  It makes for a long, strenuous day, followed by the removal of the loading scaffolding.  Today after a final look around, they installed the top of the oven and make final preparations for the 3-month casting process.

Round about this weekend they will flick the switch and it will start the temperature ramp up towards the melting point of the glass.  It takes about a week to reach the high temperature of 1180C when the mold will fill with the molten glass as the oven spins at the correct speed to form the right curve on the molten surface.  This is scheduled to happen during the weekend of the 24th, after which it will cool slowly to prevent stress in the blank.  If all goes well, we'll get to open up the oven and take our first look about Thanksgiving.  If only we could pump out polished mirrors as fast!