Several of the most compelling open questions in astrophysics, such as those regarding the nature of the ‘dark energy’ that is thought responsible for the universe's accelerating expansion or the processes governing galaxy formation, rely heavily on the possibility of taking a large number of spectra of distant objects simultaneously. This desired multiplexing capability represents a significant technical challenge, however. Since celestial objects are scattered randomly across the sky, any viable solution must allow for a high degree of versatility. Simple focal-plane masks made by punching small apertures into metal plates are an option, but their preparation requires sophisticated machining capabilities that are rarely available at the remote sites hosting state-of-the-art astronomical observatories. Another strategy uses multiple optical fibers, positioned either manually or using robotic arms. Micro-electromechanical systems (MEMS) offer an alternative approach. Randomly addressable optical switches that can selectively transmit or reflect light to the spectrograph (see Figure 1) could allow the generation of slit patterns in real time, which could lead to enormous gains in terms of versatility and cost.1
Multi-object spectrograph concept2
based on a micromirror array positioned in the telescope's focal plane (with a field of view of 3×3arcmin2
Employing MEMS, either in transmission (micro-shutter arrays: MSAs) or reflection (digital micromirror devices: DMDs), was first considered for astrophysical applications by NASA in the late 1990s. NASA developed the MSA for the Near-IR Spectrograph (NIRSpec) on board the James Webb Space Telescope (JWST). MSAs are arrays of small (100×200μm2) shutter blades connected to a frame by narrow torsion bars that can be opened through magnetic actuation and latched open electrostatically.3 The most recent MSAs qualified for flight on JWST/NIRspec contain 171×365 pixels and operate in cryogenic conditions.
An alternative approach, invented by L. Hornbeck and W. E. Nelson at Texas Instruments (TI) in 1987 and also originally considered by NASA, involves employing DMDs. These are arrays of randomly addressable microscopic mirrors that can switch along their diagonal direction thousands of times per second as a result of electrostatic attraction between the mirror structure and the underlying electrodes. The latest generations of DMDs exhibit a ±12° tilt angle, with a center-to-center pitch of either 10.8 or 13.68μm. To date, TI has produced more than 18 million DMDs for projection applications. Even if commercial devices are not suitable for the extreme cryogenic conditions encountered by NIRSpec, NASA's early interest has triggered construction of at least two multi-object spectrographs, the IR Multi-Object Spectrograph (IRMOS) and the Rochester Institute of Technology Multi-Object Spectrometer (RITMOS).
IRMOS was conceived by NASA to explore the design and performance of a DMD-based instrument while simultaneously providing a high-performance scientific instrument for astrophysical research.4 It is based on an early 848×600-element DMD, cooled to approximately −45°C to enable IR observations from 0.85 to 2.5μm. The all-reflective optics of IRMOS currently deliver a field of view of 170×210 arcseconds2 at the Kitt Peak National Observatory's 4m telescope. RITMOS5 also uses a 848×600-pixel TI DMD, but at room temperature. Its main characteristic is that it exploits both the ‘on’ and ‘off’ beams reflected by the DMD, feeding spectrographic and imaging channels in parallel. RITMOS, built for the Mees Observatory 24 inch telescope, owes its scientific potential to the clever exploitation of the unique DMD principles.
In 2007, we proposed a novel space mission, the Spectroscopic All-Sky Cosmic Explorer (SPACE),6 in response to a call for proposals issued by the European Space Agency (ESA). Its main goal is to produce the largest ever 3D evolutionary map of the universe, covering the past ten billion years, by taking spectra of hundreds of millions of galaxies. To perform this deep all-sky spectroscopic survey, SPACE relies on the latest-generation DMD, the 2048×1080-element TI Cinema chip. The SPACE mission concept, envisioned as a joint ESA/NASA collaboration, has passed the first ESA selection round and is now part of the Euclid mission under consideration for a 2017 launch date.
With the design study of SPACE/Euclid, we are currently exploring several solutions for DMD-based spectrographs.7 However, the range of new instrument concepts enabled by employing DMDs as optical switches remains to be fully explored.
Space Telescope Science Institute
Massimo Robberto is an instrument scientist for the near-IR camera under construction for the James Webb Space Telescope. He worked previously on the Hubble Space Telescope's Wide Field Camera 3 and on IR instrumentation for large ground-based telescopes.