Observing the image of a faint object that lies close to a star is a demanding task as the object is generally hidden in the glare of the star. Characterising this object, by taking spectra, is an even harder challenge. Still, thanks to ingenious scientists and a new ESO imaging spectrograph, this is now feasible, paving the way to an eldorado of many new thrilling discoveries.
These very high contrast observations are fundamental for directly imaging unknown extra-solar planets (i.e. planets orbiting a star other than the Sun), as well as low-mass stars and brown dwarfs, those failed stars that are too small to start burning hydrogen into helium. Astronomer Niranjan Thatte and his colleagues developed a new method for exactly this purpose. The basis of the concept is relatively simple: while the positions of most of the features associated with the host star and artefacts produced by the telescope and the instrument scale with the wavelength, the location of a faint companion does not. So if the image has an internal reflection of the star masquerading as a planet, this phantom planet will be in one location in the image when looking in red light, and another when looking in blue; a real planet will stay at the same place no matter what colour of light one examines. Therefore, with the combined detection of spectra and position, one can see what is scaling, subtract it, and be left with what is fixed, that is the target dim object. Such observations can be done with specific instruments, called 'integral field spectrographs', such as the SINFONI instrument on ESO's VLT. This technique, termed Spectral Deconvolution (SD), although first proposed in 2002 for space-based applications, has never been applied to obtain spectra of a real object until now.