Showing that a rotating liquid forms a parabola is an elementary exercise in equations. In optics, one learns that a parabola perfectly focuses a collimated beam of light, which means that the images on axis should be diffraction limited. The derived focal length f of a rotating liquid turns out to be f = g/2(ω)2, where g is the acceleration of gravity and ω is the angular velocity (radians per second) of the liquid. Mercury is a highly reflecting liquid. Put the concepts together and, in principle, you have a perfect telescope of focal length f. Robert W. Wood tried it nearly 100 years ago, when he rotated a 50-cm-diameter pan containing mercury:
The instrument resolved stars three seconds of arc apart, showed the small craterlets on the moon, and yielded wonderfully bright images of nebulae when running with a short focus. It was, however, merely a scientific curiosity. 1
Due to vibrations from the bearing, the room, and the drive belt, Wood’s rotating mercury surface had standing waves that affected his images. Much later, rotating curing dishes of epoxies have shown similar problems. 2 The researchers at Laval University were the first to fabricate a container with a preformed parabolic bottom; the container was rotated on a precision air bearing and had only a 1-mm-thick mercury coating that could not support long-wavelength surface waves. Laval’s liquid mirror telescopes (LMTs), 2.5 meters in diameter, have been found to be diffraction limited. 3
The November 2003 Physics Today story Physics Today 0031-9228 56 11 2003 24 https://doi.org/10.1063/1.1634524 (page 24) presents the state of the art for LMTs. Compare, though, the $1 million 6-meter Hickson LMT to the proposed 30-meter multimirror telescope, which has a projected cost of $700 million. The apertures of these two telescopes are in the ratio of 1/25, while the costs are 1/700! LMTs have many applications, such as the proposed Large Aperture Mirror Array (LAMA) with a 50-meter effective diameter; that is 3.33 times more aperture than the proposed 30-meter, at 1/14 to 1/7 its cost.
I was involved with the choice and installation of one of the 2.7-meter LMTs mentioned in the story—a lidar collector with a 4.5-meter focal length, located at UCLA’s HIPAS ionospheric research facility near Fairbanks, Alaska. 4 First light was 7 May 1995, with parts costing $45 000. That LMT is in its own two-story structure with an overhead glass skylight, to protect it from local outside winter temperatures of around −4°C.
The HIPAS LMT has been operating reliably since first light, and has even run for three months continuously on occasion. When the mercury container is first rotated, mercury vapor levels are high; however, with time, the surface oxidizes and the levels are well below the US government safety threshold of 50 µg/m3 per 5 hours’ exposure. After two weeks of operation, the vapor level is typically 20 µg/m3.
Not mentioned in the article is the ease of cleaning mercury, particularly since all common objects float on it. The container is stopped, and a lead-weighted rubber tube is used to drag debris and mercury oxide to an edge of the puddle, from which the debris is aspirated away. The mirror is then restarted. Because the parabolic mercury reflecting surface is only due to equilibrium between gravity and centrifugal pressures, such easily cleaned mirrors can be used to focus kilojoule laser pulses into the ionosphere for the creation of plasma columns at 100-m altitudes without fear of permanently damaging the focusing surface.
