Wavefront Sensing Concepts for Extremely Large Telescopes


The resolution of the ground-based telescope is limited by the optical effects of atmospheric turbulence, which distort the plane wavefronts arriving from point-like astronomical sources. Adaptive optics (AO) systems are used to correct for these wavefront errors in real time by means of a deformable mirror (DM). To achieve wavefront correction over wider field of view, a multi-conjugate AO system with multiple DMs conjugated to different altitudes in the atmosphere has been proposed. Current AO systems employ sodium laser guide stars (LGSs) to increase sky coverage and signal-to-noise ratio for the wavefront sensor (WFS) systems measuring the instantaneous wavefront errors. The proximity of LGSs to the telescope creates a severe problem of LGS re-imaging in the WFS system.


The virtual wavefront sensor concept, originally proposed at Lund Observatory, Sweden, attempts to resolve the LGS re-imaging problem by employing two dedicated WFSs and using additional test sources. As an alternative to conventional wavefront sensing performed at the final telescope focus, virtual wavefront sensing is carried out by two wavefronts sensors, namely, the primary WFS and test WFS. The primary WFS is located at the first available LGS focus in the first part of the telescope system, where images of the LGSs are not significantly aberrated despite their proximity to the telescope. To estimate the effect of the second part of the telescope on the LGS wavefronts, including contributions from relay optics and additional DMs, several artificial point sources are used at the intermediate telescope focus. The wavefronts from these test-sources are monitored by the test WFS working in the final science focus of the telescope. The wavefront sensing data obtained from both WFSs are converted into “virtual” wavefronts as if they were measured by conventional WFS in the final focus. Therefore, we call the combination of the primary and test WFSs a virtual wavefront sensor. Using, for example, modal atmospheric tomography, the required corrections to the DMs are derived from the estimated virtual wavefronts.



In order to experimentally verify the virtual WFS concept, we consider a downscaled model of a 10-m telescope with two DMs conjugated to the telescope pupil and a high altitude layer. The diameter of the equivalent FOV is 1 arcmin.  To simplify our experimental setup, we assume that synthetic NGS wavefronts (probing the atmospheric volume and the pupil DM) have been successfully generated using one of the techniques mentioned above. Ignoring the residual cone effect related to the formation of the synthetic NGSs from LGS wavefronts, we conjugate our three probe sources directly to infinity so that they appear as NGSs at the intermediate telescope focus. We also use three test sources to probe the second part of the telescope including the DM conjugated to an equivalent altitude of 10 km above the ground. The test sources are aligned with these NGSs to illuminate the high altitude DM along the same direction of propagation, as for the light from the NGSs. This is a mandatory condition for combining the corresponding wavefronts from the primary and test WFSs. The main objective of the experiment is to demonstrate the concept of combining the wavefronts from the two WFSs as proposed for the virtual WFS. The primary and test WFSs always work in quasi-open loop, whereas their combined wavefronts, which are the virtual NGS wavefronts, can be considered as closed loop measurements because the correction is achieved with all DMs. 


We have designed an optical system, which has four parts and represents a telescope operating in the dual-conjugate AO mode based on the virtual WFS concept. A module with three white light sources is followed by a beamsplitter, which forms a reference-source arm (containing an atmospheric module with three phase screens, first part of the telescope and the pupil DM) and the test-source arm (containing images of three test sources and the second DM). These arms are combined into one WFS system, which uses a single CCD and plays a role of the primary and test WFSs. To evaluate the correction achieved with both DMs, an additional pass is provided, which delivers corrected and uncorrected images of the three NGSs onto an acquisition camera. An optical layout of the experimental setup is presented below.


In order to make viewing of the optical layout easier we have numbered all optical components in the order of light travel through the primary WFS arm, that is, starting from the plane of the three reference sources, reflecting from 45-degree flat mirror M1, and passing through a collimator L1, aperture stop, which defines the pupil, beam splitter BSC1, pupil de-magnifying system of lenses L2, L3 and L4, atmospheric turbulence phase screens PS1, PS2 and PS3, telescope double objective L5 and L6, passing through a thin parallel plate BS1 (uncorrected images of the NGSs reflected from BS1 to the acquisition camera to the left as indicated by an arrow) and double collimator L7 and L8 to the pupil DM1. After reflecting from the DM1, the three light beams go through L8 and L7 and reflect from three flat mirrors attached to the back surface of the beamsplitter BS1, and then go through L9 and L10 and reflect on dichroic beamsplitter BS2, which transmits infrared (starting at 650 nm) and reflects in the visible. After reflecting on from BS2, the beams pass L11 and L12, and two flat mirrors M3 and M4 required for image inversion before entering a collimator array CA1, the beamsplitter BSC2 and the WFS. The CA1 consists of three off-axis doublets collimating the light from the three NGSs. In the test arm, we use the same three NGSs, but this time the light, after passing M1, L1, Pupil and BSC1, goes directly to the beamsplitter BS2. After reflecting from BS2, the beams travel via L13 and L14 and reflect (towards L15, L16, L17 and DM2) from three flat mirrors attached to the front surface of the reflective red dichroic filter BS3, which is needed to direct the return beams from DM2 and L17, L16 and L15 to the acquisition camera (passing L14, L13, BS2, L10, L9 and BS1). The green light on its return path from DM2 (conjugated to the highest phase screen PS1) goes through BS3 and enters a collimator array CA2, which is identical to CA1. After passing CA2, the three collimated beams enter BSC2 and join the other three beams (from the primary WFS arm) in the wavefront sensor. An afocal lens system L18 and L19 reduces the diameter of the beams on a lenslet array (LLA).