On the trail of gravitational waves
After decades of intensive preparations, the researchers achieved their goal; on 14 September 2015 two detectors known as Advanced LIGO finally managed to ensnare gravitational waves. The installation in the USA uses technologies developed primarily at the Max Planck Institute for Gravitational Physics, because many of these technologies were developed for the GEO600 detector, whose home since the 1990s has been a field in Ruthe near Hanover. So it is worth taking a look at the details of the gravitational wave trap which serves as a global test laboratory for all other instruments with a similar design.
Text: Helmut Hornung
How to measure “nothing”
Inconceivable, but true: if you want to detect gravitational waves, you must measure the mutual shift of two light waves by a hundred billionth of a degree (10-11) – a “nothing” which an interferometer converts into a tiny difference in brightness. Although the interferometer was invented more than 130 years ago and the first laser built in 1960, the technologies available did not meet the high demands of gravitational wave astronomers.
The detector requires a light source as monochromatic (single coloured) as possible with extremely constant brightness: the perturbing signal caused by intensity and frequency fluctuations must be kept as low as possible. “Diode pumped” solid-state lasers are ideal for constructing such a highly stable light source.
The scientists selected an Nd:YAG laser for GEO600 owing to its high efficiency and output power, its long service life and the fact that it needs no maintenance. At its heart is a neodymium-doped yttrium-aluminium garnet crystal whose end surfaces are cut in a special way to give it the shape of a ring resonator.
Two laser diodes, as are found in a commercial CD player, shine in light. With each cycle, this light is converted into infrared laser light with a wavelength of 1064 nanometres. Small piezoelectric crystals and Peltier elements are located on the laser crystal. The former react to mechanical stresses and always keep the crystal in the correct shape, the latter keep its temperature constant.
The corrections are made electronically via a reference system that continuously compares actual and target data concerning frequency and intensity with each other, and sends the necessary commands to the sensors on the crystal.
The recycling trick
The output power of the laser is around one watt – too low for proper measurements. This is because the sensitivity of the detector depends on the circulating light power. From a purely theoretical point of view, the interferometer could only operate at its optimum with an output power of one million watts. In order to come slightly closer to this target, the light is sent into a second ring laser with more power, which takes over the good properties of the first laser.
Two further ring resonators filter the laser beam geometrically and lay bare the stable core of the beam. The system thus generates an output power of around ten watts. To further amplify the light power circulating in the interferometer, the method of “power recycling” is used: the exit of the interferometer is dark in the normal case (no gravitational wave). A tiny variation in brightness caused by a gravitational wave is then easier to observe than against a very bright background.
In this so-called destructive interference, light is not destroyed, it is merely redistributed: it is returned to the entrance and then reused. This is done by placing a mirror between laser and interferometer which reflects the light and transmits further laser light. The light circulating in the interferometer is thus amplified. This trick is used in the Advanced LIGO to get what “feels” like almost a million watts (megawatts) from the laser.
The researchers also place high demands on the optical components such as mirrors and beam splitters. Let’s take a look at the beam splitter. It is permeated by the light circulating in the installation, the material absorbing a fraction of the light power. The result: the beam splitter becomes warm like a piece of glass in the sun and expands in the process.
Unfortunately, this thermal expansion is not uniform, meaning the surface of the beam splitter bends. This change in shape is minimal, but it nevertheless causes the beam splitter to act like an optical lens that focuses the laser beam. In the least favourable case this can lead to the destruction of another optical component in the beam path.
A silica glass was thus developed for GEO600 which absorbs only one hundredth of the light absorbed by all the types of glass used before; the absorption of this new material amounts to less than one millionth part per centimetre. In addition, the designers placed great value on there being as little stray light as possible. This meant that the surface of the finished silica glass body had to be as highly polished as possible in order to prevent irregular scattering of the light.
“As highly polished as possible” means: the surface should be smooth on the atomic scale. This extreme requirement was also actually fulfilled. The average roughness of the mirror surface is a mere 10-10 metres over a distance of 26 centimetres – corresponding to the diameter of an atom.
Light which exerts pressure
The extreme sensitivity of the whole installation to vibrations means effects occur which are normally not important. The effect in space that creates the dust tail of a comet in the sky – namely the radiation pressure of light – causes problems for GEO600. The laser beam can also be imagined as an irregular sequence of light particles (photons). These transfer momentum so that the mirror is shaken in an irregular way which thus feigns a signal.
Although this effect is used to calibrate the detector by sending a laser pulse of a defined power onto the end mirror and then observing the signal produced, the effect is unwelcome when the detector is in operation. In order to keep it as small as possible, the mirrors are solid silica blocks with a mass of around ten kilograms. Their form is also determined by internal perturbations that originate from the thermal motion of the atoms.
This thermal excitement causes the surface of the mirror to move and produces an interfering signal which is far larger than the actual signal expected. The simplest solution would be to cool the optics – which is practically impossible to realize, as the system would have to be cooled to close to absolute zero, to temperatures below one Kelvin (minus 272 degrees Celsius). The next generation of detectors is to use this technique, and put the beam splitter and end mirrors into cryostats (a kind of thermos flask).
For the time being, the designers make do with a trick and ensure that the undesired surface oscillations lie in a frequency range which cannot be used anyway. To this effect, they give the mirror a very special shape – its thickness is around half the dimension of its diameter. The general rule is: the larger the mass of the mirror and the better the mechanical quality of the material, the smaller the perturbations that remain.