Empirical measurement of MEMS stroke saturation, with implications for woofer-tweeter architectures

The Gemini Planet Imager (GPI) is currently in production for the Gemini Telescope in Chile. This instrument will directly image young jovian exoplanets, aided by a micro-electrical mechanical systems (MEMS) deformable mirror (DM). Boston Micromachines MEMS mirrors operate thousands of actuators to provide a well-sampled correction at high spatial frequencies. However, because MEMS stroke alone is insufficient to fully correct the atmosphere in the near-IR on an 8-meter telescope, a dual-mirror system is planned for GPI: The MEMS is used as a "tweeter" to correct the higher spatial frequencies while a separate "woofer" DM will be used to correct the lower frequencies. During operation at GPI, any saturated actuators would scatter starlight into the dark hole instead of allowing it to be removed coronagraphically; thus, stroke saturation on the MEMS is tolerated only at the 5-sigma level. In the Laboratory for Adaptive Optics, we test the ability of the MEMS to counter atmospheric turbulence. The MEMS shape is set to random iterations of woofer-corrected Kolmogorov phase screens with varying woofer sizes. We find that, for r₀ = 10 cm, saturation decreases from several percent to a few tenths of a percent (∼-sigma) when using a 100cm-pitch woofer. The MEMS we tested has 0.2 μu; m inter-actuator stroke for a 200V-range. Nonetheless, saturation (when it occurs) appears to be due to low-order peak-to-valley stroke even in the woofer-corrected case. Gemini characteristically has r₀ = 15 cm, so future work includes extrapolating to find where the 5-sigma saturation level occurs.