What are the honeycomb mirrors made of?

The mirror substrates are made of glass. Glass is used because it is a very stable material. As long as you don't drop it, a piece of glass will maintain its shape for thousands of years. Glass can also be polished to a very smooth and precise surface without any granular structure. The spun-cast mirror blanks produced at the Mirror Lab are made from E6 low expansion borosilicate glass produced by the Ohara Corporation in Japan. The coefficient of thermal expansion (CTE) of E6 glass is 2.8 x 10-6/C. Ohara makes this glass in large clay pots, each holding more than a ton. Ohara supplies shipments of glass to the Mirror Lab in pieces typically weighing ~ 4-5 kg.

What makes the glass reflective?

After the glass surface is polished to a very precise figure, it is turned into a front-surface mirror by applying a very thin metallic coating. The function of the glass substrate is to hold the shape of this thin metal layer which reflects the light. Aluminum is most frequently used, although silver or gold is used in some applications. Aluminum is the preferred coating since it has a high reflectivity over the whole visible spectrum whereas the reflectivity of silver decreases significantly shortward of a wavelength of ~ 400 nm. The thickness is typically ~ 100 nm ( 4.0 x 10-6 inch) thick and weighs only a few grams. Aluminum is applied in a vacuum chamber by evaporating a small amount of metal and allowing it to bond to the clean glass surface. For astronomical telescopes, the mirror coating (aka aluminization) is normally done at the mountaintop observatory site.

Why is borosilicate glass used for the castings?

Borosilicate glass is chosen for the casting of the honeycomb mirrors because:

  • Borosilicate glass has a relatively low coefficient of thermal expansion. Ohara E6 glass has a coefficient of thermal expansion of 2.8 x10-6/C. This expansion rate keeps the mirror figure accurate as the temperature varies.
  • The working point of the borosilicate glass is low enough that we can mold it into the complex honeycomb structures at temperatures which are easy to obtain. At 1180 C (2156 F), the viscosity of the glass is about the same as cold honey at ~ 15 C.
  • The borosilicate glass is relatively inexpensive compared to other types of glass and glass-ceramics used in making that mirror blanks. While the Mirror Lab orders special production E6 borosilicate glass from Ohara, mirror blanks could be, and have been, cast from common borosilicate glass, like that used in cookware, i.e., Pyrex (Corning 7740), with a coefficient of thermal expansion of 3.3 x 10-6/C.
  • Borosilicate glass has good resistance to attack by chemicals. This property is important because the polished glass surface is recoated with a fresh coating of aluminum every one to three years. Strong acids and/or bases are used to dissolve the previous coating and contaminants. The mirrors must survive this cleaning process dozens of times without requiring the precise surface to be re-polished.

Why do the mirrors have a honeycomb structure?

The mirrors are made with a honeycomb structure to have low weight and high stiffness. Traditionally, mirror blanks have been made of solid slabs of glass with a thickness about 1/6th of their diameter. If this technology is scaled to a diameter of 6.5 meters (256 inches), a solid mirror blank will weigh about 60 tons. A mirror of this mass is difficult to handle and to support in the telescope. If made of zero-expansion materials, the cost would be significantly more expensive. A honeycomb structure (much like the structure of a beehive) can reduce the weight by a factor of five to seven times.

A second important advantage of the honeycomb structure is that it can be controlled to the nighttime air temperature. A thick mirror will tend to lag behind the night air temperature by many hours as the mountaintop cools during the night. If the mirror is warmer than the surround air, convection currents will form above the surface. These pockets of warmer air have a different refractive index, and when mixed in the wind cause a blurring of the images in the telescope. The effect is similar to the mirage seen above the highway on a warm summer day. The honeycomb mirror has less mass of glass to cool, and so can remain in closer equilibrium with the night air. We can also circulate ambient cool night air through the interior of the honeycomb structure to further improve the temperature equilibrium.

How is the honeycomb structure formed?

The honeycomb sandwich structure is formed by melting the glass into a mold which is the negative of the desired honeycomb shape. The glass fills the voids in this mold and becomes a honeycomb structure. The main components of the mold are a cylindrical tub which sets the outside shape and alumina-silica hexagonal columns or cores which form the voids in the honeycomb. Each of the hollow core boxes is held down against the floatation force with a special silicon carbide bolt. The cylindrical tub is held together by bands of Inconel 601 which wrap around the outside of the mold in a manner similar to hoops of steel around the staves of a barrel. After the honeycomb structure has been cast and cooled, the mold material is removed by breaking it apart with a high pressure water spray.

Why are the mirrors figured as paraboloids?

Mirrors for large reflecting telescopes are polished to precise paraboloidal (figure of revolution like a parabola) or nearly paraboloidal shapes. This shape can perfectly focus starlight from far away into an image just above the mirror. In this manner, a concave mirror acts the same as a camera lens - forming an image for the film or CCD detector to record. The f/number, or focal ratio, is the ratio of the focal length to diameter of the mirror. Our large spun-spun cast mirrors typically have focal ratios of f/1.25 (for 6.5-m mirrors) or f/1.14 (for 8.4-m mirrors). Frequently, a second smaller mirror is used to redirect the image from the prime focus to a more convenient location or image scale.

How is the paraboloidal shape formed?

The rough paraboloidal shape is formed when the mirror blank is cast in the spinning furnace. By spinning the furnace at the proper speed while the glass is molten, the surface takes on a paraboloidal shape. By the time the cooling process is complete, this surface is accurate to a small fraction of an inch. By utilizing the spin-casting technique, we avoid the need to grind tons of glass in the sagitta (or center) of the mirror. This method saves many tons of glass and years of grinding.

A precise paraboloidal shape is generated (ground with a spinning tool impregnated with diamond particles) with a numerically-controlled milling machine. This procedure improves the surface accuracy to about 50 microns (0.002 inch).  

The final shape of the surface is produced by polishing with a lap using a very fine polishing compound. This shape is carefully polished to an accuracy of better than 25 nanometers  (1.0 x 10-6 inch).

Why does the mirror surface have to be polished so accurately?

The surface of a telescope mirror must be polished to its precise paraboloidal shape within approximately 1/25 of the wavelength of light. For typical blue light, that means a surface accuracy of order 15-20 nanometers (less than 1.0 x 10-6 inch).  Any small scale roughness (the lack of a good polish) will cause the light to be scattered and result in reduced contrast. Inaccuracies on larger scales, such as bending of the entire mirror, can result in an inability to focus the light into sharp images. The value of these large mirrors in telescopes is to collect lots of light from very faint astronomical objects and to focus it into very sharp images.

How are these very large and fragile pieces of glass moved?

The castings are first lifted out of the oven with a lifting fixture that has 36 steel pads which are glued to the front surface of the mirror with RTV silicone rubber. An overhead crane in the casting lab sets the lifting fixture into the turning ring so the mirror can be rotated from a horizontal position to a vertical one.

The first phase of optical finishing involves grinding on the back and edges of the mirror blank. Once the back of the mirror is finished, load spreaders, to distribute the weight of the glass, are glued in place and become a permanent interface to the glass. The mirror blank remains attached to the lifting fixture until it is placed on the polishing cell. At that point, the pads glued to the front of the mirror are removed.

After the front surface of the mirror has been polished, the glued pads are replaced with a set of 36 vacuum pads which lift the mirror by suction. This vacuum lifting fixture is used to lift the mirror in and out of the telescope cell and the shipping box.

While both the casting lab and the integration lab have overhead cranes, the optical test tower in the polishing lab would interfere with a crane. Mirrors and mirror cells are moved between the three areas of the Mirror Lab with an "air cart". The air cart is a type of hovercraft that can move heavy loads over flat surfaces.