Construction of a 3.7 meters liquid mirror
The construction of the 3.7 meters liquid mirror started in April of 1996. Here is, in pictures, a short chronology of the operation.
First, the foam...
The construction of the mirror's container begins with its bottom side.
Blocks of styrofoam are glued together with X-40 superglue.
A heated wire is then used to carve the foam into a cone-like shape.
The hot wire has done its job. After removing the excess of foam cut by
the wire, we start to work on the central part of the container.
The slope of the cone was computed by Robert Content (University of Durham) using finite element software.
The foam is sanded to correct irregularities left by the carving. It is now ready to receive the
first layer of Kevlar.
Then the Kevlar...
In this photo, we prepare the application of the first layers of kevlar on the central part of the
container. We begin with a small layer in the center and each successive layer is slightly larger, ending with
a complete surface coverage. This distribution was calculated by
Robert Content. The first twenty layers are finished.
We use an epoxy glue to laminate the kevlar. The glue takes about 24 hours to dry.
After this step, we are done with the central part of the container.
Expanding the structure...
Additional styrofoam blocks are added to the container
to increase the mirror's diameter to 3.7 m, its final size.
A scaffolding is assembled around the container. Care must be taken to avoid stepping on the
high density foam.
The hot wire at work again! The bottom part of the container is carved. Now we will prepare this surface
to receive the final layers of Kevlar. On the wall, behind us, you can see the leftover foam carved with the hot wire.
The sanding of the surface is complete and we begin to cover
the bottom of the "mirror" with the first full-surface layer of Kevlar.
Additional layers of Kevlar are deposited on the bottom surface. The epoxy glue
used to laminate the kevlar will be dry in 24 hours, after which the bottom
side of the container will be finished. We are ready to turn it over.
The underside of the mirror is finished. We have to turn the container over to work the top side.
This picture, taken half way through the operation, gives a good idea of the mirror's scale.
The three-point mount...
The three points mount is made of 600 kg of steel tubes and plates. It is designed
to conteract efficiently any unbalance of the liquid mercury in the container.
It was made following a design by Éric Masson.
A liquid mirror has to be precisely levelled.
To obtain a good precision, one can either use a large base, or a fine height ajustement
on each of the platform's feet. We chose someting in between. We use a
commercial adjustable wedge to level our instrument.
The base was complete before the container. You can see us still working in the
back of the laboratory. Meenwhile, Gilberto, the photographer, was resting on the side of the pool.
With a touch of paint, our base starts to look like a professional instrument. We install a mechanical bearing
and it is ready to receive the container. So far, we have about half of the job done.
The container was built near an optical test tower. We only had to move it
a few meters to put it on its base. For that purpose, we used a system of chains and pulleys fixed at
the center of the container, making the move with relative ease.
Carving the sphere...
For a 3.7 m diameter F/1.2 mirror, the radius of curvature is 8.64 m long. The difference (Peak-to-Valley) between a parabola and a sphere of the same specifications is 0.6 mm. We therefore decided to carve the surface into a spherical shape, technically easier than carving a parabola.
For that purpose, we used an 8,74 m steel pipe. The pipe was fixed at the center of curvature of the mirror.
To carve the foam, we fixed a milling machine at the end of the pipe. The tool rotates at 20000 rpm
and gives excellent results on foam.
The container is slowly spun on its bearing while we move the milling machine towards the center.
We had to proceed in two steps to make sure we were carving a perfect sphere. We first made an
approximate carving and then, once sure the operation was going well, we made the final cut.
The sphere is carved. One can see the foam blocks used to build the heart of the container.
Laminating the surface...
The container is removed from the three point mount. It will not come back until the top side is laminated.
The procedure we used to laminate the top part of the container is quite similar to the one we used for the bottom.
We began with 20 smaller layers in the center.
It is easier to apply a kevlar layer when the inferior layers are still wet with resin.
We therefore applied the final 8 layers in a single day.
Once the kevlar is dry, we put the container back on its three point mount and trim the edge to the
correct diameter using the milling machine.
To complete the container, we only need to install the rim.
We already have a good idea of the final look of the setup.
Installing the rim...
We begin by turning the container upside down. It will be easier for us to work on the rim this way.
With this picture, one can easily comprehend the size of our future mirror.
A foam ring is installed all around the container. It will be used to hold the edges in place.
We than laminate the foam ring with kevlar. It is now part of the setup.
The side of the container is now thick enough to support the rim.
The rim , made of thin sheet of kevlar, is glued to the container using a FLOX mixture.
We used bolts and washers to hold the sheet in place while the glue set.
And here is the result. Soon, we will spin-cast urethane resin on the container.
This will define our parabola.
Spin-casting...
Spin-casting is used to give to the surface of our container a parabolic shape.
We used a urethane resin to perform the operation. For optimal results, the resin must
be mixed following very precise procedures.
The container rotates at the future mirror's optimal velocity so that the
resin surface will match the distribution mercury as closely as possible. This will allow us to
use only thin layers of mercury. The resin layer is not too thick, so as to minimize
the total weight of the container.
The operation is over. It took about 15 minutes. Once the resin has polymerised (and after a
few final adjustments) the container will be ready.
Mirror, mirror...
Here it is at last. A 3.7 meter liquid mirror. The pictures were taken on April 7 1997, 3 days after the mirror was first activated. A thin layer of oxide had formed on the mercury surface 4 to 5 hours after startup, virtually eliminating mercury evaporation. After 3 days there are practically no traces of mercury vapor in the air. The surface of the mirror is 10.75 square meters. The mercury layer is 1.8-mm thick for a total volume of 22 liters of mercury.