So how are we overcoming these limitations on telescopes? The biggest limitation historically was the size of the mirror. In 1948 the Palomar 200 inch telescope was commissioned in Southern California, the largest mirror ever built. And for decades, it will be the largest mirror in the United States until at the University of Arizona about ten years ago, we built a bigger mirror. When the Palomar 200 inch was built, nobody thought there would be a larger telescope. If you've ever been there, the dome is the size of a cathedral, has 10,000 tons of moving glass and steel. It's a huge and fantastic device. But our bigger telescopes are actually in smaller buildings because we have compact telescope designs. So the first progress has been made in building bigger mirrors than five meters or 200 inches. The mirror for the large synoptic survey telescope, or LSST, was the largest mirror ever made at the time we cast it under the stand of our football stadium at the University of Arizona. It's a phenomenally accurate as well as large mirror. If you imagine this mirror, which is larger than most people's living rooms, expanded to the size of the continental United States, the biggest imperfections will be bumps less than one inch high. It's the most accurate surface ever made. Telescopes with mirrors this large also have physically large detectors. They're still charged coupled devices like used in your phone and your camera, but they're much bigger. The CCD in your camera or cellphone might be the size of your little fingernail, but you can see the size of the detector in the large synoptic survey telescope. It's a huge data rate. Every time you push the button and take a picture, gigabytes pour in. And the data rate in one night will be 20 terabytes when this telescope is commissioned in a couple of years. That's a fire hose worth of data that astronomers will struggle to keep up with. As I mentioned, we've started making the biggest mirrors in the world under the stand of our football stadium. Our football team is kind of mediocre but the mirrors we make in the football stadium are exceptional. We use them for our own projects and we also sell them to other consortia for their own telescopes. And occasionally, we make mirrors for people like the Air Force, but that's so secret I can't even talk about it. The first time we eclipsed the Palomar 200 inch mirror was with a six and a half meter mirror for the multiple mirror telescope, with this one monolithic mirror replace the six 1.8 meter mirrors that had been inside previously. This telescope is located on Mount Hopkins in southern Arizona. Having cut our teeth on our first six and a half meter mirror, we started to make more, one about every 18 months. The guru of this work is a man called Roger Angel, a pioneer who first worked out how to make large and light mirrors. He's a professor in the department here. We made two 6.5-meter mirrors for the Twin Magellan Telescopes located in Chile at the Las Campanas Observatory. I've used both of them a number of times. They're magnificent telescopes on one of the darkest sites on earth. Then we stepped up to 8.4 meters which is the largest mirrors that can be made with the technique we're currently using. Two of these 8.4 meter mirrors were installed in the last couple of years in the large binocular telescope located on Mount Graham in southern Arizona. These two mirrors work together to make the effective aperture of over 11 meters, making this the largest telescope on earth. Seeing the twin mirrors in their cells inside the LVT dome is an awesome sight. These are enormous pieces of hardware, and they're also exquisitely precise. The object of our current dreams and aspirations is a telescope called the Giant Magellan Telescope. Also to be located in Chile on an adjacent mountain top to the twin 6.5 meter mirrors we already have. This one will take 7 of the 8.4 meter mirrors and put them together like flower petals around a central mirror, creating the equivalent of a 22 or 23 meter mirror telescope which would be by far the largest telescope in the world. We hope to finish this project within six years and get first light in about seven years. What's the trick by which we make such large mirrors when it seemed impossible? And for 30 or 40 years, no larger mirrors were made in the United States or anywhere in the world. These tricks all come from Roger Angel and the very clever scientists and engineers who work in the Mirror Lab at Stewart Observatory. The mirror is essentially a honeycomb, which means it's not a solid block of glass. The glass used is incidentally borosilicate, which is just a very uncommon sand, so a very simple ingredient. The honeycomb mirror has a smooth and continuous face plate of glass which is only one inch thick. But most of the volume of the mirror is a honeycomb where it's mostly air. The reason for a honeycomb mirror is two-fold. The first is to make the mirror light, much lighter than it would be if it was a solid block of glass that thick. The second is that a light mirror can have its temperature equalized with the surroundings. One of the things that degrades images from a large telescope is when the temperature of the mirror is different from the air surrounding it or the structure surrounding it. These light honeycomb mirrors can be kept in close harness with the temperature of the surrounding structure to within a tenth of a degree C. We can walk through the steps involved in making one of these huge mirrors. All of this taking place invisibly to the students and faculty of the University of Arizona under the stadium stand. A large tub has ceramic fiber boxes installed, each one in the shape of a hexagon. There are 1,681 of them. Each one is unique, and together they form the contour of a parabolic surface. So the mirror can be formed by liquid glass in more or less the shape it needs to be. Next, 18 tons of boro silicate glass are loaded on top of the preformed hexagon. Bora silicate is close to common sand but it needs to be of unbelievable purity to make a very accurate mirror surface. We get this glass from a small foundry on the Northern island of Hokkaido. In fact, it's a family run business and it's amazing that such an operation can make glass of the purity we need. We need so much of their glass that I think we have their order book filled for a decade. After carefully loading the top of the preformed shapes with the blocks of glass, each one about ten pounds the lid of the oven is closed and sealed. The oven starts to spin reaching a peak speed of about five revolutions per minute. Within it, a carefully controlled temperature cycle begins which will reach a peak temperature of 1160 degrees Celsius, hot enough to melt the glass. As the glass melts, it liquefies and falls between the shapes of the hexagons, filling the space between them and leaving a one inch faceplate that's in the perfect parabolic shape as the oven spins. Then the temperature is slowly reduced. Glass can fracture if it's cooled too quickly. So the mirrors go through a three-month slow cooling called annealing, which prevents bubbles from forming, and the glass from fracturing due to stresses. It's a nerve wracking time because you're not allowed to look at the mirror, or even crack the lid for a few weeks. And you really don't know until the end of the three months whether you've made a good mirror or a bad mirror. All this work is just the first step. By spinning the mirror, you've approximated a parabolic shape which is what we need for the final mirror. But it's not yet with the precision's needed. The mirror is then moved on to what's called a stressed-lap polishing machine where using something like Machines grind away the surface down to the millionth of a meter level to produce a perfect optical shape. This process also takes months, sometimes six or eight months, to produce the speck on the mirror. When the mirror is finished, if you could imagine the mirror doing this, you could read a newspaper at a distance of five miles. There are still challenges ahead. Once you've made your mirror and polished it to the required spec, you have to get it on top of a remote mountain top. These are huge pieces of glass that need to be encased and protected. It turns out that even the best military lift helicopter cannot get a mirror of this mass onto a high mountaintop where the air is thin. They must be taken up by road. Often the road has to be remade to get the huge mirror on its articulated vehicle around the hairpin turns. The vehicle itself has huge cantilever weights that swing in and out to stop the tractor trailer from toppling off the mountain side. It's a hair raising procedure and it takes a week to ten days to get one of these mirrors from Tuscon to one of our nearby mountain tops. Once the mirror is located in the telescope cell, it never moves. Older style telescopes like the Palomar 200 inch Used to have their mirrors taken out of their cells every year or so, moved across the dome floor, and re-aluminized in a special chamber. These modern telescopes are far too fragile to ever risk them by moving them out of the telescope, so they are re-aluminized every couple of years In situ, in place. Very thin coats of aluminum are sputtered onto the surface to make it clean. Remember the telescope sit in a natural environment that involves dust, moisture, wind, hopefully no rain, but they degrade over time. And so every couple of years all large telescopes need their surfaces recorded. We've seen that the historical impasse on the largest size of an optical mirror has been broken by making light wave mirrors. The University of California and Observatories pioneered the use of thin, light weight segments Which are mosaic to make a single large surface. Here at the University of Arizona, we make large monolithic mirrors, where they're made light by having a honeycomb structure where most of the volume is, in fact, air. Yet they're very stiff, and light, and can move easily around the sky. Mirrors 8.4 meters in size have been made this way. And we are planning to make six or more as a part of a huge telescope called the Giant Magellan Telescope.
