LIGO Collaboration Prepares Cryogenic Update

by Prof. Rana X Adhikari, California Institute of Technology; Dr. Brett N. Shapiro, Stanford University

A sketch of a cryogenic LIGO mirror inside a vacuum enclosure. Image: LIGO Collaboration

Figure 1. A sketch of a cryogenic LIGO mirror inside a vacuum enclosure. Image: LIGO Collaboration

In the past year, the Laser Interferometer Gravitational-wave Observatory (LIGO) has twice observed gravitational waves from the merger of black holes in deep space. To make this observation, researchers must be able to measure phase shifts in the light equivalent to motions of ~10-19 meters over the 4 km arm length of the interferometers.

In order to make such a sensitive measurement, the laser power impinging on the interferometers’ mirrors exceeds 100 kW (continuous). In addition to filtering out the vibrations from the environment, the microscopic thermal vibrations of the mirror surface must also be tamed.

The LIGO interferometers operate at room temperature and so each eigenmode of the mirror is vibrating with 1/2 kB (300 K) of thermal energy. The trick to making sensitive displacement measurements in the presence of thermal noise is to not look where the noise is: the mirror materials are engineered to have such a high mechanical Q (low internal friction) that the energy is well contained within the narrow frequency band of the high Q eigenfrequencies. Still, the residual thermal noise far from the resonance does set a limit to how small a motion may be measured interferometrically.

In order to take the next big step in gravitational-wave astronomy, this thermal noise limit must be surpassed. We can either find higher Q materials or operate the system cryogenically. Happily, nature has conspired to make these properties go hand in hand and the LIGO Scientific Collaboration is now making designs and measurements to enable a near future upgrade of this sort. The LIGO Collaboration is exploring the optical absorption and noise characteristics of single-crystal silicon, how to synthesize a 200 W laser with a two micron wavelength and testing of various UHV compatible, high-emissivity coatings to enable efficient radiative cooling. The aim is to have a room-size prototype demonstration within three years.

Many crystalline materials have a Q inversely proportional to temperature. The ~150 kg silicon mirrors that will be needed for the LIGO upgrade should have a Q 109. The remaining challenge is to build a thin film Bragg coating for the silicon substrate. This coating must be able to have a high Q (> 105) and absorption below 1 ppm.

One of the toughest engineering challenges in a cryogenic laser interferometer is to extract several watts of heat from the mirror without disturbing the mirror motion at the 10-19 meters level. Thermal straps of OFHC copper could remove the heat but would connect the mirror to the noisy cold head. An exchange gas could also be used for the initial cooldown but would produce an acoustic short between the vacuum chamber and the mirror. Radiative cooling is ordinarily quite weak, but a mirror this large (45 cm dia., 50 cm thick) should be able to radiate ~10 W while operating at 123 K.

Figure 1 illustrates the chosen design for the cooling of the mirrors inside LIGO’s updated vacuum enclosure. The primary source of the heat absorbed into the mirror will come from the interferometer’s laser, depositing ~10 W into the mirror’s surface.

The design features a dual shielding system employed to maintain the mirror at 123 K, the desired temperature in the presence of this heat load. An inner shield at about 80 K surrounds the mirror. This shield will collect the ~10 W radiated away by the 123 K mirror. This inner shield is suspended from wires and springs in order to give it vibration isolation down to the 10-10 meters level. This isolation is important because a small amount (~30 ppm) of light will scatter off the mirror, bounce off the inner shield and find its way back to the interferometer. The phase of this returning light will be contaminated with the velocity of the inner shield, and can potentially mask the phase shift induced by the gravitational radiation.

The inner shield will be cooled through flexible copper straps due to its vibration isolation requirement. These straps limit the amount of heat that can be extracted from it. Thus, to minimize the heat load on this shield, an outer shield surrounds the inner one. The outer shield’s vibration is much less critical since it is largely inaccessible to the laser beam. Consequently, it will be mounted rigidly to the ground and cooled to 77 K via liquid nitrogen pipes. The inner shield’s flexible copper straps will be mounted between this outer shield and the inner for conductive heat transfer between the two.

While radiative cooling works well in the steady state, it would take weeks to cool down from room temperature. Either a UHV compatible heat switch or an exchange gas will thus be used during each initial cooldown to achieve the desired state in about one day, thus allowing for frequent incursions into the vacuum envelope to make repairs and upgrades.