The Project Manager's
Corner: TMT Builds Its First Mirror
Gary Sanders, TMT Project Manager
April 2006
The Thirty Meter Telescope project has finished building its
first mirror—a tiny flexible mirror that fits in the palm of
your hand. Building and testing this small mirror means that
the giant telescope project has passed a major technology hurdle.
This tiny mirror may be the most challenging mirror in TMT.
A large primary mirror is the heart of any telescope. That
mirror defines the optical quality of the telescope and the
resulting image quality. The diameter of the primary mirror
defines the maximum amount of light collected. This determines
how well the telescope can take images of faint and far objects.
Its diameter also defines how sharp a point the telescope can
image, and how close two objects can be in the sky and still
be seen to be separate by the telescope. This is the resolution
of the telescope and that limit is set by the optical theory
of diffraction. For the 30 meter-diameter mirror of TMT, the
limit in imaging objects sharply will be nearly 5,000 times
better than the human eye!
It should not be a surprise that the building of the mirror
is a great event in the building of a telescope. Read The
Perfect Machine: Building the Palomar Telescope, by Ronald
Florence, for an account of the building of the epic Palomar
200-inch mirror and telescope. For a segmented mirror telescope
like TMT, however, building the first segment is a smaller milestone.
Since TMT is founded on the technology base of the Keck 10-meter
telescopes, each with 36 mirror segments very similar to TMT,
building the first TMT segment may not be such a great step.
Together with spares, the Keck team built 84 segments. This
same team is working with TMT toward building our 860 segments,
including spares. The challenge is not building the segments,
but building 10 times more of them at a very low price and very
fast pace. So building our first segment is a minor step (though
we will drink some champagne in celebration this year when we
see our first segment).
Achieving resolution nearly 5,000 times better than a human
eye depends upon the big primary mirror. But it depends even
more critically on another mirror. A tiny mirror, about 30 centimeters
(one foot) across. A deformable mirror.
This mirror has a high-quality optical surface. The thin glassy
surface is mounted on about 4,200 tiny actuators that can push
or pull at a point on the back of the mirror, thus distorting
the surface up or down locally. This is the mirror that can
compensate for the shimmering distortion that Earth’s atmosphere
imposes on light waves arriving from a distant star or galaxy.
This deformable mirror is the heart of adaptive optics technology.
And adaptive optics is really the very heart of TMT. Yes, our
big mirror is important. But TMT makes a great leap forward
because it can sweep away the disturbances made by our atmosphere,
and allow TMT to achieve diffraction limited performance at
infrared wavelengths.
To understand how this works, you need a primer on adaptive
optics. We have one on this website that is very well done,
courtesy of the Gemini Observatory.
Click here to view the adaptive
optics animation and watch the film several times. You
can use your media player to interrupt it at each step and
slow it down after you have watched it once or twice. This
animation shows you step-by-step how adaptive optics works.
And it shows you the crucial role of a deformable mirror.
When you are done, come back to this column and we will continue
with the story of our deformable mirror.
Now that you are back and expert in adaptive optics, you know
that the surface of the deformable mirror has to be very agile.
How agile? Our agility goal is set by our larger goal: to correct
the arriving starlight after it is disturbed by the atmosphere
well enough to achieve the resolution limit of the 30-meter
telescope as if there were no atmosphere at all!
Light leaves a distant star heading out in all directions.
The wavefront of one wave in the light forms a spherical surface
leaving the star, like the spherical surface of an expanding
balloon. By the time that surface travels a very great distance
to reach Earth, the sphere is very large and any individual
patch of area on the wavefront is quite flat. In other words,
that wavefront arrives as a very flat plane with all of the
rays of light in step (in phase), having traveled essentially
the same distance. All of this is true at the top of the Earth’s
atmosphere.
If there were no atmosphere, the wavefront would arrive at
TMT and TMT would act like a 30-meter diameter cookie cutter
and slice out a 30-meter diameter disk in the wavefront and
focus it with the primary mirror, and eventually our instruments
would see an image that is very sharp indeed, defined by this
perfect wavefront in a disc.
There is an atmosphere. High in this atmosphere there is turbulence.
There are regions of the atmosphere that are warmer and colder.
The warmer regions are less dense and so light sees a different
index of refraction. That means that it travels at a slightly
slower speed for more dense regions of the atmosphere, and faster
for the lower density, warmer regions. Turbulence mixes the
denser and more rarified patches of sky so that any ray of light
may experience a different mixture of transit speeds compared
to other rays of light.
Due to this effect, the very flat plane wavefront arrives at
the surface of the Earth with some portions a bit ahead and
some a bit behind. Different rays of light have experienced
different optical path lengths, and so they arrive out of step.
"…Twinkle twinkle, little star. How I wonder where you
are…"
The image is now blurred and squirming around in the focal
plane of the telescope. Look back at the animation discussed
earlier. Look at the star image on the computer monitor before
the AO system is turned on and the deformable mirror performs
its bit of wizardry. Squirming around.
The distorted wavefront arrives like a jiggling potato chip
and the deformable mirror in the path jiggles in the opposite
way by just the right amount to restore near flatness to the
wavefront. Voila, a sharp image approaching the limit set by
diffraction theory. The telescope behaves as if the atmosphere
had been swept away.
TMT needs such a mirror with about 4,200 actuators. The actuators
are about 5 millimeters apart in a regular array, and they have
to be able to push and pull the mirror surface up and down by
about 8 to 10 microns (a human hair is about 70 microns thick).
Remember that the surface of the mirror is a glass-like material.
When one actuator pushes up, the nearby actuator spots must
not move much. The surface must be flexible. And the entire
device must operate at -30 degrees Celsius, a very cold temperature
where materials stiffen and actuators are sluggish. If such
a mirror could be built, TMT can correct the 30-meter diameter
plane wavefront to about one-thousandth of the thickness of
a human hair.
Mirrors like this, with over 4,000 actuators packed closely
together that can reliably push and pull this far at very cold
temperatures, do not yet exist. The best deformable mirrors
fall short of these requirements. But TMT cannot operate well
without such a deformable mirror.
We have built the first mirror of this type. More correctly,
our industrial partner, CILAS (a
French company whose name stands for, in English, "Industrial
Laser Company") designed the full 4,200-actuator version
shown in Figure 1. They then built a portion of this mirror
sufficient in size to demonstrate the performance (with 57 actuators)
but with lower cost than the full mirror, and tested it. The
mirror passed all of the required tests, even at the cold temperatures.
A photograph of the mirror is shown in Figure 2.
This is a very important development. We know how to build
our primary mirror segments. And now we know that we can build
the crucial adaptive optics mirror to use at first light for
the telescope.
— Gary Sanders |
