We use cookies to ensure that we give you the best experience on our website. You can change your cookie settings at any time. Otherwise, we'll assume you're OK to continue.

Centre for Advanced Instrumentation



KMOS is a second-generation instrument for the European Southern Observatory's (ESO) Very Large Telescope (VLT). KMOS is a unique design of near-infrared multi-object spectrograph which uses deployable integral field units (d-IFUs) to obtain spatially-resolved spectra for multiple target objects selected from within an extended field of view. d-IFUs have a significant advantage over multi-slit spectrographs because of the reduced slit contention in crowded fields and their insensitivity to slit losses due to extended galaxy morphology and orientation. KMOS will mount onto the Nasmyth platform of ESO's UT1 and will use the Nasmyth A&G facilities. The top-level scientific requirements are: (i) to support spatially-resolved (3-D) spectroscopy; (ii) to allow multiplexed spectroscopic observations; (iii) to allow observations across the IZ, YJ, H, and K infrared atmospheric windows from 0.8 to 2.5 μm. These have been flowed down during a series of ever more detailed design reviews to arrive at the final design specification. KMOS completed its Final Design Review in April 2008 and completed commissioning at Paranal observatory in Chile in April 2013. The instrument is now in regular use for scientific observations at ESO.

The final design employs 24 robotic arms that position fold mirrors at user-specified locations in a 7.2 arcmin diameter field of view. Each arm selects a sub-field on the sky of 2.8x2.8 arcseconds. The size of the sub-fields is tailored specifically to the compact sizes of high redshift galaxies, with a spatial sampling (0ʼʼ.2 per pixel) designed to sample the excellent infrared seeing at Paranal (FWHMmedian=0”.52 in the K-band). The sub-fields are then anamorphically magnified onto 24 advanced image slicer IFUs that partition each sub-field into 14 slices, with 14 spatial pixels along each slice. Light from the IFUs is dispersed by three cryogenic grating spectrometers which generate 14x14 spectra, each with ~1000 Nyquist-sampled spectral resolution elements, for all of the 24 independent sub-fields. The spectrometers each employ a single 2kx2k Hawaii-2RG HgCdTe detector. The optical layout for the whole system has a threefold symmetry about the Nasmyth optical axis allowing a staged modular approach to assembly, integration and test.

Figure 1 shows a cutaway drawing of the main KMOS cryostat design showing the entrance window followed by the pickoff arm module layer, the integral field unit module layer and the spectrograph module layer. Each 1/3 sector contains 8 robotic arms, 8 IFUs and 1 spectrograph to allow a sequential approach to assembly, integration & test and easier maintenance.

Figure 1: KMOS Cryostat

Design Specifications



Final Design
Instrument Total Throughput (mean) IZ>20% , YJ>20%, H>30%, K>30%
Wavelength coverage 0.8 to 2.5 µm
Spectral Resolution R=3300,3400,3800,3800 (IZ,YJ,H,K)
Number of IFUs 24
Extent of each IFU 2.8 x 2.8 arcseconds
Spatial Sampling 0.2 x 0.2 arcseconds
Patrol field 7.2 arcmin diameter circle
Close packing of IFUs >3 within 1 sq. arcmin
Closest approach of IFUs edge-to-edge separation of >6 arcsec


Cryostat and CACOR

The KMOS cryostat is the main support structure for the optomechanical components which make up the instrument and mounts directly onto the Nasmyth rotator. It uses three low-vibration Leybold 10MD cryocoolers to maintain an internal optical bench at a temperature below T=140K, in order to minimise the thermal background radiation, and to keep the detector at a temperature below T=80K to minimise dark current and persistence. A major design change after PDR has been to remove the electronics racks and services from the outside of the cryostat and to mount these on a co-rotating structure (CACOR) located separately on the Nasmyth platform (Figure 2). This provides an improved mass margin and also improved the expected flexure performance. The cryostat is a hybrid aluminium-steel vessel with a diameter of 2 metres. 

Figure 2: KMOS and CACOR

Figure 3 shows the KMOS cryostat mounted on a Bode positioner at UKATC. The closed-cycle coolers are visible around the circumference of the cryostat. Also visible is the ESO NGC detector controller used to read out the Hawaii-2RG detectors. The red CACOR to the right contains the three instrument electronics racks containing all of the mechanism and housekeeping electronics, all of which need to co-rotate with the instrument.

Figure 3: KMOS and CACOR at UKATC

Pick-Off Arms

One of the more unusual elements in KMOS is the pickoff module which relays the light from the 24 selected regions distributed within the patrol field, to an intermediate focus position at the entrance to the integral field unit module. The robotic pickoff arms are of an (r,θ) design (Figure 4) with pivot points located in a circle around the periphery of the patrol field and are driven in radial and angular motions by two cryogenically prepared stepper motors.

The arms are arranged into two layers on either side of the Nasmyth focal plane to improve the access to target objects in crowded fields whilst avoiding interference between neighbouring arms. This focal plane is flat and telecentric thanks to a pair of all-silica field lenses, one of which forms the entrance window to the cryostat. The pickoff arm design has been through a number of refinements based on repeatability, flexure and lifetime tests conducted in a relevant environment (i.e. in vacuum at T~140K). Whilst the positioning of the arms is open-loop via step-counting from datum switches, there is a linear variable differential transformer (LVDT) encoder on each arm which is used to check for successful movement. In addition a hardware collision-detection system is also implemented as a third level of protection which can sense if any two arms have come into contact and stop the movement of arms within 10μm. The absolute positioning accuracy of the arm when cold has been measured using an automated laser tracker system and is < 50μm (0.1 arcsec).

The pickoff module also contains the instrument calibration unit (Tungsten, Argon and Neon sources) and an order-sorting filter wheels which provides focus compensation between the different bands. The cold-stop for each channel is at the base of the arm, after which an intermediate image is formed by a fixed K-mirror assembly which orientates the pickoff fields so that their edges are parallel on the sky. This enables a sparse matrix configuration for the arms where the KMOS IFUs can be used to map a contiguous region of sky covering 65x43 (33x16) arcseconds in 16(9) dither pointings.

Figure 4 shows one of the fully assembled KMOS pickoff arms. The linear motion is a stepper motor drive in the top of the arm, and the angular motion is a stepper drive in the base. The total length of the mechanism is ~50cm.

Figure 4: Pick-Off Arm

Integral Field Unit

The IFU subsystem contains optics that collects the output beams from each of the 24 pickoffs and reimages them with appropriate anamorphic magnification onto the 24 image slicers. The anamorphic magnification is required in order that the spatial sampling pixels (ʻspaxelsʼ) on the sky are square whilst maintaining Nyquist sampling of the spectra on the detector in the spectral dimension. The slices from groups of 8 sub-fields are aligned and reformatted into a single 254mm long slit at the entrance to the three spectrometers. The optics in a single IFU comprises: two off-axis aspheric re-imaging mirrors, a third re-imaging mirror defined with a more complex geometry using Zernike polynomials, one monolithic slicing mirror array containing 14 slices with spherical surfaces in different orientations, two monolithic pupil mirror arrays containing 7 facets each with spherical surfaces, and one monolithic slit mirror array containing 14 facets with toroidal surface form. All of the micro-optics in the IFUs are produced by diamond-machining using a combination of diamondturning and raster fly-cutting techniques. This technique allows arrays of multi-faceted components to be manufactured with in-built mounting surfaces, all to sub-micron accuracy. Particular attention was paid to minimising the micro roughness on the optical surfaces which was in the range 5-10nm rms for many components. Detailed metrology measurements are being performed on every element to ensure that the stringent tolerances in form error and alignment for the IFUs have been achieved. Figure 5 shows a complete set (one of three) of IFU components produced by CfAI's Precision Optics Laboratory. Each set comprises 72 separate mirrors with 384 individual optical surfaces. Figure 6 shows a view of the 'front' of KMOS with the 24 pick-off arms visible.

Figure 5: One of three complete sets of IFU mirrors

Figure 6: KMOS with all three pick-off and IFU segments installed

Current Status

KMOS is currently in regular scientific use at ESO's Very Large Telescope at the Paranal observatory in Chile. Figure 7 shows a view of the instrument installed on VLT UT1. The first light KMOS spectrum was collected on 22nd Nov 2012. The formal press release can be found on the ESO news pages. Some recent science results from KMOS can be found here.

Figure 7: KMOS installed on ESO VLT UT1 (Image credit: ESO / G. Lombardi)



KMOS was built by a consortium of UK and German institutes working in partnership with ESO:

CfAI, Department of Physics, Durham University, Durham, UK
Universitätssternwarte München, München, Germany
UK Astronomy Technology Centre, Royal Observatory, Edinburgh, UK
Max-Planck-Institut für Extraterrestrische Physik, Garching, Germany
Sub-Department of Astrophysics, University of Oxford, Oxford, UK
European Southern Observatory, Garching, Germany

The KMOS PI was Ray Sharples:

The Durham Work Package Manager was Paul Clark: