The technical "first light" KMOS spectrum with the instrument cold, was obtained using the instrument calibration unit on 20 January 2010. This significant milestone, which tests all the integrated subsystems, is the first step to obtain an end-to-end qualification check of the complete cryogenic performance. It was carried out using a part-populated front segment with two complete channels comprising pickoff arms, IFUs, spectrograph and detector system.
KMOS is a new cryogenic near infrared instrument which will combine the advantages of multiplexing with the power of integral field spectroscopy. A consortium of German and British institutes together with ESO develop KMOS as one of the so called second generation VLT instruments. It is planned to be operational in late 2011.
The baseline concept is to have 24 integral field units (IFUs), each of which has 14x14 spatial elements providing a 2.8arcsec field of view and can be positioned anywhere in the 7.2arcmin diameter unvignetted field of the VLT Nasmyth focus.
There will be 3 identical spectrographs with a resolving power of about 3500 in the J, H, and K bands, each fed by 8 IFUs and equipped with a Hawaii 2RG array with 2K x 2K pixels. The cryostat is 2m in diameter and 1.5m high, and will be cooled to 77K by 3 closed cycle coolers.
The IFUs are located at the tips of 24 robotic arms that patrol the field of view. The arms are arranged in two planes 20mm above and below focal plane to maximise the freedom of movement. The arms weigh about 4.5kg each and have a size of 30 cm. Positioning within 0.1” (<60μm) is feasible.
KMOS will generate its own internal flatfields and wavelength calibration. For these it uses 2 halogen lamps, a neon lamp and an argon lamp mounted in an integrating sphere outside the instrument. The light is directed through a sealed tube to another integration sphere in the centre of the cryostat, and thence to each arm. In order to detect light from the flatfield lamps, the arms must be positioned correctly outside the patrol field.
KMOS is designed so that 8 IFU arms fed into a single spectrograph and have their light dispersed onto a single detector. Thus, in total there are 3 spectrographs and 3 detectors. In the picture on the right the optical path for 1/3 of the instrument is shown, the two other segments are identical. Hence, the format of the data on each detector is, modulo optical alignment and manufacturing tolerances, identical.
The optical path of each arm has 45 optical surfaces. In total the instrument comprises 1080 optical surfaces and 60 cryogenic motors.
At the right one can see a raw image (2k x 6k pixels) as it will be produced by KMOS. The three images of the three detectors are attached one to the other with gaps in between. The vertical axis is the wavelength, the horizontal axis the spatial position. The bright, horizontal lines are OH-lines.
In the magnified area one can distinguish vertical slitlets. Each slilet is 14 pixels wide and 14 slitlets set up an IFU. All IFUs are joined horizontally. When the 14 slitlets are stacked, a datacube is obtained with a depth of 2000 pixels which equals the wavelength resolution.
Before reconstructing datacubes the calibration data obtained before each observation run is processed together with the raw image
Ahead of each observation run several calibration measurements (Dark Frame, Flatfield, Spectrum Alignment, Wavelength Alignment, Telluric Correction) have to be done. Before reconstructing datacubes the calibration data will be processed together with the raw image.
Since the targets to be observed are very faint and long integration times (> 1h) are needed to get data, a Real Time Display (RTD) is needed. Therefore some of the IFUs point to bright reference objects which are visible with short integration times (~10min). For the observation itself the calculated offsets can then be applied to the instrument.
For the Real time Display (RTD) the datacubes of all IFUs are flattened and mosaiced together.
Additionally a Patrol View of the whole field of view can be displayed, where the images of the IFUs are pasted at their position at the sky.
In general, multi-IFU allows one to select 20 to 50 sub-fields from the telescope field by positioning a pick-off device appropriately in order to relay the light form the sources to the IFUs. The spatial elements of these sub-fields (100 to 1000 each) are rearranged along the entrance slit of a spectrograph. In this way, spectra from all source locations within the white squares (left column in the figure) are recorded; and from these source images at all wavelengths can be reconstructed. In contrast to standard multi object slit spectroscopy, the spectrograph slit position is independent of the source position and one does not need to worry about the orientation of the slit.