Monte Carlo simulation of ELT-scale multi-object adaptive optics deformable mirror requirements and tolerances

Abstract

Multi-object adaptive optics (MOAO) has been demonstrated by the CANARY instrument on the William Herschel Telescope. However, for proposed MOAO systems on the next generation Extremely Large Telescopes, such as EAGLE, many challenges remain. Here we investigate requirements that MOAO operation places on deformable mirrors (DMs) using a full end-to-end Monte-Carlo AO simulation code. By taking into consideration a prior global ground-layer (GL) correction, we show that actuator density for the MOAO DMs can be reduced with little performance loss. We note that this reduction is only possible with the addition of a GL DM, whose order is greater than or equal to that of the original MOAO mirrors. The addition of a GL DM of lesser order does not affect system performance (if tip/tilt star sharpening is ignored). We also quantify the maximum mechanical DM stroke requirements (3.5 µm desired) and provide tolerances for the DM alignment accuracy, both lateral (to within an eighth of a sub-aperture) and rotational (to within 0.2 • ). By presenting results over a range of laser guide star asterism diameters, we ensure that these results are equally applicable for laser tomographic AO systems. We provide the opportunity for significant cost savings to be made in the implementation of MOAO systems, resulting from the lower requirement for DM actuator density.

Monte Carlo modelling of multi-object adaptive optics performance on the European Extremely Large Telescope

“…However, a study of performance trends is possible, and we find that several of the performance trends that we present here (e.g. performance as a function of asterism diameter, with scaling of NGS flux, number of LGS, and LGS pixel scale) are similar to those in previous studies using Monte-Carlo end-to-end AO simulation (Basden 2015b;Le Louarn et al 2012;Tallon et al 2011;Foppiani et al 2010;Basden, Evans & Morris 2014;Basden et al 2013). We note that Monte-Carlo models are pessimistic when compared to analytical model results (Neichel, Fusco & Conan 2008).…”

“…However, individual instrumental science requirements may mean that performance goals can still be met with a lower performance, and so using lower order DMs should not be ruled out simply because performance is reduced. We note that these results are similar to those reported by Basden et al (2013). We also consider the use of a hexagonal actuator pattern for M4, (the global ground layer correction), with the DM having 75 × 65 actuators.…”

“…Although the performance of LTAO and MOAO as a function of LGS asterism diameter has been well studied elsewhere (Le Louarn et al 2012;Basden et al 2013), we include here such a study, to provide a reference, and for comparisons with other results. Fig.…”

“…The impact of sub-aperture rotation has been studied previously (Basden et al 2013). Here, we consider only the effect of LGS motion (tracking the sky) resulting in a change in tomography, i.e.…”

Wide-field adaptive optics performance in cosmological deep fields for multi-object spectroscopy with the European Extremely Large Telescope

“…number and final specification of science channels), our objective here was to investigate the broader aspects of a MOAOcorrected MOS, so we investigate a wide range of NGS magnitudes (Section 3), detector performances (Section 3.1), tip-tilt correction (Section 3.2) and spatial element size on the sky (Section 3.3). We note that the MOAO performance across the telescope field of view has been investigated previously (Basden et al 2013), so here we provide on-axis performance estimates (at the centre of the LGS asterism) to simplify our results. In summary, we use six sodium LGSs (589 nm wavelength) equally spaced around the edge of a 7.3 arcmin diameter circle (the widest LGS asterism that can be transported to the E-ELT focal plane), each with 74 × 74 subapertures and 16 × 16 pixels per subaperture.…”