A method that uses laser frequency combs to calibrate astronomical spectrographs to unprecedented accuracies has been developed and successfully tested by researchers in Europe. The method could be used to find Earth-sized exoplanets by detecting their tiny influence on the motions of their companion stars. The comb was tested on the European Southern Observatory's High Accuracy Radial velocity Planet Searcher (HARPS) spectrograph at the La Silla Observatory in Chile.

Astronomical spectrographs separate light according to wavelength and the spectra that they produce play important roles in many aspects of astronomy. As a result, astronomers are constantly looking at ways to make their spectrographs more accurate, stable and precisely calibrated. Currently, the best spectrographs, such as HARPS, use thorium-argon lamps or iodine cells for calibration – however, these do not deliver the precision to detect the tiny shifts in the wavelength of starlight caused by the presence of an exoplanet.

These shifts in wavelength correspond to changes in the radial velocity of a star – which, in turn, could be caused by the gravitational influence of any exoplanets that may be orbiting the star. Radial-velocity changes are derived from shifts in the parent star's spectral lines caused by the Doppler effect. While this method works very well for enormous planets that orbit very close to their parent stars, the accuracy needed to measure the tiny shifts caused by an Earth-sized planet orbiting within the habitable zone of a Sun-like star cannot be achieved today.

The idea of using a frequency comb to calibrate a spectrograph has been discussed for several years, but this is the first time that the technique has been tested and verified, thanks to the efforts of Tobias Wilken of the Max-Planck-Institute for Quantum Optics in Munich and colleagues in Germany and Spain, in collaboration with scientists at the European Southern Observatory (ESO). Wilken is part of a research group headed by Theodor Hänsch, who shared the 2005 Nobel Prize for Physics for his development of frequency combs.

Frequency ruler

The team installed a laser frequency comb to calibrate HARPS over two test runs carried out in November 2010 and January 2011. The comb delivers a series of equally spaced spectral lines that act as a "frequency ruler" against which the light emitted from distant stars can be measured. Wilken told physicsworld.com that complex adaptions had to be made to the comb before it could be used. Spectrographs that can detect exoplanets do so by looking at light in the 400–700 nm range. But Wilken uses a fibre laser as his comb because it is "very robust against thermal variations or acoustic vibrations that are bound to occur when the comb is permanently attached to a working spectrograph". The problem is that the comb operates in the 1000–1500 nm range.

Also, he explains that with fibre lasers, the length of the laser cavity is directly proportional to the line spacing – that is, with a longer time between each laser pulse, the pulse frequency is lowered. The pulse frequency also determines the line spacing, so for a low frequency there is low spacing. But to accurately resolve a spectral line, very high spacing is required. So the team had to increase the laser line spacing. This was done using Fabry–Pérot cavities – highly reflecting parallel mirrors. "They transmit only a small portion of the light, increasing the spacing," explains Wilken.

For the entire system to work, two channels (fibres) send light to the spectrograph – one channel guides the calibration light; the other channel, starlight – through the grating and prism set-up, from where the light is detected by a CCD camera, showing a very high-resolution spectrum. "The comb actually has 10,000 lines, which appear as several rows of dots in the CCD image, and you see the calibration lines directly beneath the starlight spectra, so you can constantly compare them and check for any inaccuracies," says Wilken. For example, there might be a slight expansion of the spectrograph caused by heat, which would be detectable thanks to the calibration lines, and thus the instrumental error can be measured and removed.

Calibration light

For its initial tests, the team used calibration light in both channels, so that the researchers could ensure that the calibration was working accurately. "This way, if you have the same source, any changes should be seen in both spectra. So then, with different sources, all shifts are down to instrumental effects and so can be corrected for," Wilken explains. "And finally, after that any shifts you see are caused by astronomical stellar effects, like a planet," he says.

The team also used the calibrated spectrograph to make eight measurements of the radial velocity of the star HD 75289, known to have an exoplanet, to ensure that it could be detected. They observed the system for five nights and "this was done mainly to show that the comb can do night-by-night observations, as this was previously doubted by the astronomical community, who felt that the method was complex and too futuristic for now," says Wilken.

Despite this success, the comb is not quite ready to be installed permanently. The company that produces the frequency combs, Menlo, set up by Hänsch and colleagues and a spin-off from the Max-Plank-Institute, is now perfecting the combs for permanent long-time use. More durable combs should be ready for use in a year's time.

Unexpected drifting

During testing on HARPS, the team also saw some unexpected drifting of the spectral lines. This allowed astronomers to identify a previously unknown problem with the HARPS instrumentation. "We need to identify where these irregularities are coming from at HARPS, so that the comb can identify the shifts, characterize them and then correct for them, so improving the stability of HARPS," says Wilken.

Wilken also points out that the combs could help in designing more accurate and stable instruments in the future. These include the ESPRESSO spectrograph that will look for rocky exoplanets and is due to be installed in 2015 on ESO's Very Large Telescope.

Zoe Leinhardt, from Bristol University in the UK, who heads a group studying planet formation and evolution, is "very impressed" by the work and believes it could prove to be very useful in detecting Earth-mass exoplanets. "If the method is perfected, and can stably reproduce previous data, it could be hugely beneficial for radial-velocity measurements," she says. As well as pushing radial-velocity measurements closer to the detection of Earth-sized planets, she also believes that the detection process could be faster than conventional techniques.

The research is published in Nature.