Gravitational wave detectors such as LIGO pick up the trembling of space-time from collisions of black holes and neutron stars. To do this, they measure changes in distance that are a thousand times smaller than the width of a proton. To make such precision possible, 40-kilogram mirrors separated by four kilometers must be held in strictly specified positions using laser light.
Before beginning to "listen" to the Universe, the mirrors are gently calmed and fixed by a special arm-length stabilization system called the locking system. The next LIGO upgrade should use new crystalline mirror coatings to increase sensitivity. However, these coatings absorb the green laser light on which the current locking system operates.
A team of researchers from the Australian nodes of the ARC Centre of Excellence for Gravitational Wave Discovery (OzGrav) at the University of Western Australia, the Australian National University and the University of Adelaide developed and tested an alternative scheme. Instead of green light, they used an infrared laser with a wavelength of 1596 nanometers, which the new coating does not absorb. Special crystals convert the frequency so that both the infrared locking light and the detector's main laser light (1064 nm) are turned into green. This allows control to be transferred smoothly from the locking laser to the main one.
Lead author Dr Jian Liu of the UWA School of Physics, Mathematics and Computing said: "We demonstrated stable handover in about 60 seconds - much faster than the current system - and successfully repeated it over several days." According to him, judging by the results, the measurement error between the lasers is extremely small - about one tenth of the width of an atom, even when the measurement is made along LIGO's four-kilometer arm.
The work is published in the journal Physical Review D. The authors conclude that a practical locking method has been found for gravitational wave detectors with new low-noise crystalline mirror coatings. This does not mean that the upgraded detector is already ready to catch signals - this is a test of a method that may be useful in a future upgrade.