A unique instrument for the biggest telescope

MOSAIC is the proposed multiple-object spectrograph for the ELT that will utilise the widest possible field of view provided by the telescope. In terms of adaptive optics, there are two distinct operating modes required to meet the top-level science requirements. The MOSAIC High Multiplex Mode (HMM) requires either seeing-limited or GLAO correction within a 0.6 (NIR) and 0.9 (VIS) arcsecond sub-fields over the widest possible field for a few hundred objects. To achieve seeing limited operation whilst maintaining the maximum unvignetted field of view for scientific observation will require recreating some of the functionality present in the Pre-Focal Station relating to control of the ELT active optics. MOSAIC High Definition Mode Control (HDM) requires a 25% Ensquared Energy (EE) within 150mas in the H-band element for approximately 10 targets distributed across the full ELT field, implying the use of Multiple Object AO (MOAO). Initial studies have shown that to meet the EE requirements whilst maintaining high-sky coverage will require the combination of wavefront signals from both high-order NGS and LGS to provide a tomographic estimate for the correction to be applied to the open-loop MOAO DMs. In this paper we present the current MOSAIC AO design and provide the first performance estimates for the baseline instrument design. We then report on the various trade-offs that will be investigated throughout the course of the Phase A study, such as the requirement to mix NGS and LGS signals tomographically. Finally, we discuss how these will impact the AO architecture, the MOSAIC design and ultimately the scientific performance of this wide-field workhorse instrument at the ELT.

More information here (SAO/NASA ADS Astronomy Abstract Service)

Positioner System Architecture

Baseline for the full architecture is a stepped tiled focal plane. This is the only configuration which can compensate adequately for the non-telecentricity of the ELT and provide support for all required modes. With 7.4’, and a tile size of 0.5’ (~100mm), we can achieve the 200 multiplex including LGS vignetting (TBD)


REM : this is just a concept representative of the type of tile we envision for MOSAIC.

  • Total system bandwidth is 0.37 to 1.8μm
  • Cross over between Vis optimised system and NIR optimised system is 0.8μm
  • Coverage at R5000 is achieved with 3 VPH in Vis and 3 VPH in NIR


  • Initial optical concept for the visible spectrograph, with the final optical architecture, indicates that the beam size will be around 200-230 mm
  • Camera speed is ~ F/1.6
  • ie a feasible design
  • [The concept is actually quite close in terms of parameters to the WEAVE spectrograph that Oxford, GEPI and NOVA are involved in]
  • More detailed concept is currently under investigation...

Both HMM and HDM modes use microlenses and fibres to transmit light to the spectrographs, the principle being to make a pupil image on the fibre cores. Microlenses and fibres are directly on the tile in the focal plate for HMM, while they are offset from the focal plate, after an optical relay including a deformable mirror for HDM. In the case of HDM, the microlens surface is conjugated to the image of the sky, and acts as a high-definition sampler.

Who are we? Infos on the MOSAIC consortium.


Scientific goals and milestones: why MOSAIC?


How do we get there? All the technology behind MOSAIC.


What performance can we expect from MOSAIC?


How will MOSAIC fit in the instrumental landscape?