Developing 3D spectroscopy in Europe

One of the inherent preoccupations of astronomy is to obtain a three-dimensional view of the Universe and its components. Except for Solar System objects, we are always presented with a two-dimensional view of celestial objects. Spiral galaxies for example could be considered only as flat structures if it were not for the rotational velocity which shows them to be spinning threedimensional entities. The distance scale is another fundamental aspect of this question – placing astronomical sources in the third dimension. Once their distance is determined, physical parameters can follow such as luminosity, radius and mass. In order to determine this information one could ideally imagine a “maximal spectrograph’’ which produced the spectrum of the whole sky at some desired spectral resolution and spatial sampling on the sky. The complexity of such an instrument is obviously beyond current technological means and the sheer size of the resulting data set would be prohibitively large. Nevertheless, a small, but significant, step towards this goal is to obtain the spectrum of an area of sky and this is what 3D spectroscopy achieves. With advances in technology the sampled area is becoming bigger. 3D spectroscopy is called integral field (IFS) or area spectroscopy and the principle is summarized in Figure 1. The resulting data have three dimensions – two spatial and one spectral. The spectrum at a spatial pixel (dubbed a “spaxel’’), or in an aperture of any desired shape over a substructure of interest, can be extracted or an image over a spectral range can be formed by summing in the spectral direction. A long-slit spectrum can be formed by slicing the 3D data in one spatial and the spectral direction. Such data have very powerful advantages over aperture or long-slit spectra which sample pre-defined spatial regions. With 3D spectroscopy not only can spectra of a whole extended object be obtained, such as a nearby elliptical galaxy to investigate its velocity field, but areas of the sky can be searched for objects which are difficult to detect in wideband imaging, such as emission line sources with a few, even only one, visible line over the wavelength range of the instrument (such as a search for Lyman-alpha emission from very highredshift galaxies). Since there is no slit in the conventional sense, there are neither slit losses nor the problems associated with atmospheric refraction which plague conventional spectroscopy. Most importantly, there is, by definition, no prejudice as to the selection of the region of interest: thus a particular slit pointing and orientation may easily miss a physically important structure which appears to be unassuming on a broadband image. Integral-field observation can unveil the complete spectroscopic signature within the 2-dimensional field of view of the wavelength coverage of the 3D instrument. A good example is the 3D observations of the stellar kinematics of the nuclear region of the nearby galaxy M31. Whilst the distribution of light shows a double nucleus, the velocity map reveals a structure that is not aligned with the two nuclei and the peak in the velocity dispersion, assumed to mark the location of the supermassive black hole, is also offset from one of the peaks (Bacon et al. 2001a). Orientation of a long slit over the obviously visible features would have missed the essentials of the velocity structure.