MSA: Programmable Slit Mask for JWST

Updated 10 June 2026

Micro-Shutter Array (MSA) is a MEMS-based configurable slit mask comprising nearly 250,000 individually addressable shutters for JWST's NIRSpec, enabling simultaneous spectral capture.
It employs a precise electrostatic actuation mechanism with independent quadrant control, achieving rapid reconfiguration (<2 minutes) and >96% operational success.
MSA supports high-multiplex spectroscopic surveys with advanced algorithmic mask planning and dynamic failure mitigation, setting the stage for next-generation space spectrographs.

The Micro-Shutter Array (MSA) is a micro-electromechanical system (MEMS) device composed of nearly 250,000 individually addressable shutters that serves as the configurable slit mask for the Near-Infrared Spectrograph (NIRSpec) instrument on the James Webb Space Telescope (JWST). It enables programmable multi-object spectroscopy (MOS) over a wide field of view, allowing simultaneous spectral acquisition of hundreds of astronomical targets. As the first space-based deployable slit device of its kind, the MSA defines the technical landscape for high-multiplex near-infrared spectroscopy in space and is core to the scientific productivity of JWST and future observatories.

1. Structural Design and Actuation Mechanism

Each NIRSpec MSA comprises four independent quadrants, each with a 365 × 171 grid of shutters, yielding a total of 249,660 micro-shutters. A single shutter measures approximately 78 μm × 178 μm, corresponding to a 0.20″ × 0.46″ aperture on the sky. Each cell consists of a magnetically latched, bi-stable silicon-nitride door integrated with electrostatic electrodes. The addressing scheme uses orthogonal sets of row (V₋) and column (V₊) lines, allowing individual control via voltage differentials:

Opening: A magnet arm initial sweep unlocks all shutters, and a voltage ΔV = V₊–V₋ latches open desired elements.
Programming: On reverse sweep, only the selected shutters maintain V_open, while others drop to V_close and close under torsion spring force.
Hold: After patterning, V_hold is applied to retain the shutter configuration during exposures (Bechtold et al., 18 Aug 2025, Bechtold et al., 2024, Rawle et al., 2022).

This design enables rapid (<2 min) reconfiguration of hundreds of thousands of apertures, with >96% success rate in typical science configurations (Rawle et al., 2022).

2. Operational Metrics, Failure Modes, and Yield

At commissioning, ≈82.5% of non-vignetted shutters were available for science; the remainder being disabled due to electronic shorts (“short masks”) or failed-closed (FC) states. Shutters may also fail open (FO), though this occurs rarely (≈20 FO at any epoch). Inoperability has evolved gradually, reaching ≈18.5% after two years on orbit:

Short-masked: 10.5 → 10.9% (from commissioning to 2024)
Failed-closed: 7.0 → 7.6%
Failed-open: ~constant at ~20 cells (Bechtold et al., 2024)

Shorts predominantly result from particulate-induced conductive paths between electrodes and manifest as elevated quadrant currents or thermal infrared glow. The MSA sustains a gradual exponential decrease in operable units, with a λ ≈ 1.46×10⁻² yr⁻¹ per shutter, leading to a 3.1% reduction in viable 3-shutter slitlets over 16 months (Jakobsen, 2024).

Epoch	Failed Open	Failed Closed (%)	Short Masked (%)	Total Inoperable (%)
Commissioning	22	7.0	10.5	17.5
Feb 2024	20	7.6	10.9	18.5

3. Detection, Localization, and Mitigation of Electrical Shorts

Shorts are isolated through two primary pathways:

Electrical Short Detection (ESD): Automated scripts under the Observation Plan recursively bisect rows and columns, measuring segmental quadrant current against a threshold. A key enhancement to the ESD script in July 2023, involving even/odd subset partitioning, increased nearest-neighbor short detection from ~50% to ~100% (Bechtold et al., 18 Aug 2025).
Optical Short Detection (OSD): For low-level shorts below electrical detection thresholds, diagnostic exposures acquired with internal calibration lamps are used to generate and uplink candidate line masks that iteratively minimize observed glow. Validated masks become the new operational ZPM (Zero-Potential Mask).

Mitigation actions require masking entire affected rows or columns, with a typical impact of 171 or 365 lost shutters per event. The decision logic seeks to minimize masked area while reliably suppressing glow. Regular re-checking has shown that many shorts are transient; successful recovery of a single line typically restores ≈288 shutters.

4. Multiplexing, Survey Strategies, and Algorithmic Mask Planning

The principal scientific leverage of the MSA is massive multiplexing—enabling 150–200 simultaneous low-overlap spectra in standard MOS, and up to ≳850 in high-density Dense-Shutter Spectroscopy (DSS) mode (D'Eugenio et al., 13 Oct 2025). The actual multiplex is a function of:

The viable slitlet fraction, strongly impacted by inoperable and masked shutters.
Target field density, geometrical slitlet packing, vignetting, exclusion constraints (e.g., avoidance of failed-open/colliding spectra).
Slitlet length and background subtraction scheme (typically three-shutter slitlets with positional nodding).

Planning mask configurations is computationally nontrivial. The eMPT software formalizes the planning as a discrete optimization problem. Its Initial Pointing Algorithm (IPA) and the Matrix (Arribas) algorithm maximize non-overlapping target allocations subject to inoperable shutters, field-dependent optical distortions, and spectral collisions. At high densities, these algorithms improve mask yield by up to ≈17% over simple greedy approaches, and dynamic exclusion of spectra projected onto detector gaps raises completeness by ~10% for PRISM (Bonaventura et al., 2023).

5. Target Acquisition, Astrometric Registration, and Impact of Shutter Degradation

Precise alignment (<20 mas error) of the science mask on the sky is essential. NIRSpec's autonomous target acquisition (MSATA) uses pre-loaded catalogs of reference stars, observed in a broadband undispersed “TA image” with all shutters open. The system requires each Reference Star to fall within a fully operable 5×3 shutter window. Degradation impacts TA more acutely than science multiplexing: the fraction of valid TA windows dropped from 45.8% to 43.6% between March 2022 and June 2023, increasing the probability of insufficient VRS to 6.3% (from 4.9%) in deep GOODS-S pointings (Jakobsen, 2024, Holwerda et al., 2016).

Mitigation includes pre-screening, dither-sequence reordering, modest catalog "fudging," and advocating for filter-magnitude limit revision. Projections suggest a linear increase in TA rejection rate, reaching ~10% by the end of 2026 if no significant improvements in shutter operability occur.

6. Observing Modes: Dense-Shutter Spectroscopy and Background Subtraction

The DSS mode increases target allocation by relaxing overlap constraints, deliberately permitting 5–10 overlapping spectra per detector row. While this elevates background noise—imposing a 1.7× exposure penalty per source—the net survey speed gain is ~5×, and line sensitivity at λ≈4 μm exceeds NIRCam WFSS by a factor ≳5 (i.e., requiring ~25× less exposure for the same sensitivity) (D'Eugenio et al., 13 Oct 2025).

In crowded fields (e.g., 30 Doradus), the canonical MSA configuration is a three-shutter mini-slit with positionally nodded exposures, optimizing background sampling. Quantitative simulations confirm that the subtraction of nebular contamination introduces a mean equivalent width error of +0.8% with σ=13% for S/N>4 (Rogers et al., 2023). The configuration strongly affects the ability to recover backgrounds in regimes of spatially varying nebular emission.

7. Lessons Learned and Outlook for Future MSA-Based Instruments

Operational experience has refined several key strategies:

Transient shorts: Confirmed by repeat observation, supporting dynamic mask re-checking and potential recovery of previously disabled lines.
Autonomy and scriptability: Onboard scripts for subset testing, mask application, and current-threshold logic are essential for managing hardware failures without constant ground intervention.
Design evolution: Full electrostatic actuation is favored for future arrays (e.g., Habitable Worlds Observatory), to mitigate particle-induced shorts.
Real-time capacity tracking: Maintaining simple, on-the-fly multiplexing metrics enables risk assessment for both science and target acquisition (Bechtold et al., 18 Aug 2025).

The MSA's field performance, despite modest degradation, has validated the MEMS-based, programmable slit paradigm for ultra-multiplexed near-IR spectroscopy in space. The operational, algorithmic, and engineering toolkit developed for the JWST/NIRSpec MSA sets a strong precedent for next-generation space-based MOS implementations (Bechtold et al., 18 Aug 2025, Bechtold et al., 2024).