Tiny mirrors for huge telescopes

Observing the formation of early galaxies helps us understand our origins. The light coming from these faraway objects is very faint and shifted to the infrared. Multi-object spectroscopy (MOS) is the central method for studying many isolated objects simultaneously, using a slit mask in the focal plane of the telescope for blocking spoiling sources and background light. Today these masks are static perforated sheets or complex fiber-optic- based systems. In the future, microelectromechanical systems (MEMS) could provide a remote controllable, reconfigurable slit mask, increasing the scientific efficiency of MOS.

In the framework of studies on the future European Extremely Large Telescope (E-ELT), we are developing a micromirror array (MMA)-based reflective slit mask. Another MEMS solution, a microshutter-based slit mask, is being developed for the James Webb Space Telescope by Moseley and colleagues at NASA.¹ However, we believe that MMAs are an appropriate solution for extremely large telescopes (ELTs), as they have a large focal plane (in the meter range). Because we work with reflected light, the electric wiring and mounting of the MEMS can be realized under the MMA. These integrated MMA building blocks can be used as ‘paving stones’ to cover large surfaces nearly seamlessly.

Figure 1 shows the basic concept of a MOS using an MMA as a slit mask. The mirrors at the locations of the objects of interest are tilted (ON state), and the light is sent toward the spectrograph. All the other mirrors remain unactuated (OFF state) and send the light of the spoiling sources back to the telescope. The ratio of wanted to unwanted light must be 3000:1 or better. Thus a mechanical tilt angle of at least 20° is needed.² This tilt angle must be uniform across the whole array, so that the light of all selected objects goes through the entrance pupil of the spectrograph. In order not to disturb the wavefront of the reflected light, the mirror must remain optically flat (λ/20) during operation. For infrared applications, it must be in a cryogenic environment.

Figure 1. In the multi-object spectroscopy microelectromechanical system (MOS-MEMS) setup, the micromirror array in the focal plane of the telescope selects the light from the objects of interest and sends it to the spectrograph.

The micromirrors are actuated electrostatically. Figure 2 shows the basic concept. The device consists of two assembled chips. The first chip comprises the micromirrors, and the second chip the electrodes and integrated spacers. The micromirrors are made out of thick single-crystal silicon to ensure the required flatness even in operation. The cantilever suspension is hidden by the mirror, reducing unwanted stray light and increasing the fill factor. A locking system, based on stoppers on the mirror and the frame, provides precise control of the tilt angle. Together with uniform spacing, this system ensures regular tilt angle across the array. ³

Figure 2. Here we see the MOS micromirror concept in operation. The two stopper beams provide precise tilt-angle control (d).

We fabricated arrays of 5×5 micromirrors successfully (see Figure 3). Mirrors and electrode chips are fabricated separately, both using a silicon-on-insulator (SOI) substrate, ensuring compatibility in terms of the thermal expansion coefficient (for cryogenic operation). The micromirror chip is fabricated using a combination of bulk and surface micromachining. The micromirrors are etched by deep reactive ion etching (DRIE) into the single-crystal bulk silicon, whereas the cantilever suspension is formed by a deposited and patterned polysilicon layer on the back of the mirror. The electrode chip is fabricated using a delay-mask DRIE process, creating electrodes and an integrated spacer at the same time. A constant height for the spacers is guaranteed by the uniform layer thickness of the SOI wafer.

Figure 3. On the left is a front side close-up view of a 5×5 micromirror array. On the right, the suspension side of the mirror shows the cantilever suspension and the stopper beams. One micromirror measures 100×200μm.

Figure 4. In this 3D optical profiler image of a 3×3 subset of a 5×5 micromirror array, one row is actuated, implementing a long slit. The fill factor is 97% along the slit.

The fabricated micromirrors show an excellent surface quality. The peak-to-valley deformation of a 100×200μm large mirror is only 7nm, and the surface RMS roughness is about 1nm. The mechanical tilt angle obtained is 20° at 90V actuation. The micromirrors equipped with the stopper beams remain stable in the ON position within 1arcmin over a voltage range of 15V. This proves the functionality of the locking mechanism, providing uniform tilt angle over the array.

We demonstrated object selection in an optical setup simulating an astronomical field of view. Using one of our 5×5 MMAs, objects could be ‘switched’ on or off by tilting the corresponding mirrors (see Figure 4). To complete the demonstration of the suitability of our MMA for MOS, cryogenic characterization and contrast measurements will be carried out in the near future. Currently we fabricate large arrays of 100×200 micromirrors. These MMAs could be used as building blocks to cover large surfaces, such as the focal plane of the future E-ELT.