NIRSpec: JWST Infrared Spectrograph

• NIRSpec is a cryogenic multi-mode near-infrared spectrograph on JWST, offering high-sensitivity spectroscopy from 0.6 to 5.3 μm with versatile observing modes.
• Its advanced micro-shutter array enables simultaneous spectroscopy of up to 200 targets over a 3.6′×3.4′ field, optimizing multi-object observations.
• The instrument delivers impressive spectral resolution and sensitivity through state-of-the-art calibration and dispersion systems, supporting extragalactic, Galactic, and exoplanet research.

The Near-Infrared Spectrograph (NIRSpec) is a cryogenic, multi-mode near-IR spectrograph onboard the James Webb Space Telescope (JWST), designed with a unique set of functionalities for high-multiplex, high-sensitivity spectroscopy from 0.6 to 5.3 μm. Its defining features include a pioneering micro-shutter array (MSA) for multi-object spectroscopy (MOS), multiple fixed slits, an integral-field unit (IFU), and a broad suite of dispersers delivering spectral resolving powers from R ~ 30 to R ~ 3700. NIRSpec's architecture, operational history, technical performance, calibration methodology, and data reduction frameworks collectively underpin its role as a cornerstone facility for extragalactic, Galactic, and exoplanet research (Bechtold et al., 2024, Ferruit et al., 2022, Böker et al., 2023).

1. Instrument Architecture and Observing Modes

NIRSpec integrates several subsystems enabling versatile observing strategies:

• Micro-Shutter Array (MSA): Four quadrants each with 365 × 171 addressable MEMS shutters (total 249,660), each ~0.20″ × 0.46″ on-sky, enabling MOS of up to 200 targets per configuration over a 3.6′ × 3.4′ field. The MSA is implemented as a cruciform mosaic; selective opening and latching of shutters realizes slit patterns for simultaneous acquisition of spatially separated objects (Bechtold et al., 2024, Ferruit et al., 2022, Maseda et al., 2024).
• Fixed Slits (FS): Five monolithic slits, including three S200 (~0.2″ × 3.3″), S400 (0.4″ × 3.8″), and S1600 (1.6″ × 1.6″), optimized for high-contrast, bright-object, and time-series exoplanet work.
• Integral Field Unit (IFU): Image-slicing relay dissecting a contiguous 3.1″ × 3.2″ region into 30 slices of 0.103″ width, feeding the disperser via eight additional reflections to enable spatially resolved, contiguous 3D spectroscopy (Böker et al., 2022).
• Dispersers: Six medium- and high-resolution gratings (G140M/H, G235M/H, G395M/H) and a double-pass CaF₂ prism provide R≃100 (0.6–5.3 μm), R≃1000 (0.7–5.2 μm), R≃2700 (same bands), with corresponding band-limiting filters (Jakobsen et al., 2022).
• Focal Plane Assembly (FPA): Two HAWAII-2RG HgCdTe arrays (2040–2048²), with specialized readout ASICs and support for both full-frame and subarray (including high-gain, fast-read) operation (Birkmann et al., 2022).
• Observing Modes: NIRSpec supports: MOS (via MSA), FS, IFU/IFS, and BOTS (Bright Object Time Series, using S1600A1 or equivalent subarrays) (Böker et al., 2023, Birkmann et al., 2022).

2. The Micro-Shutter Array: Design and Operations

The MSA is central to NIRSpec’s MOS capability:

• Device Architecture: Each quadrant consists of a 171 × 365 grid of microshutters (62,415 per quadrant), each ~78 × 178 μm. Shutters are magnetically opened and electrostatically latched for user-defined patterns (Bechtold et al., 2024).
• Electrical and Mechanical Control: Two redundant Micro-shutter Control Electronics (MCE) units address the grid. A motorized magnet arm opens all mechanically operable shutters; specific patterns are latched via controlled voltages, and a hold mask secures the configuration during exposures.
• Operability Metrics:
  • Operable Fraction: $f_{op} = N_{operable}/N_{total}$, currently ≈81% (41,812 masked + 17,144 failed-closed + 20 failed-open, out of 249,660) (Bechtold et al., 2024).
  • Failure Modes: Mechanically failed-closed (FC, currently 7.6%, up from 6.6% at commissioning), failed-open (FO, stable at ~20), and masked (shorted) rows/columns (10.9% masked).
• Short Mitigation: Electrical shorts (particulate/circuit bridging) are mitigated via ESD (current-sensed) and OSD (stray-light imaged) algorithms, with dynamic masking and infrastructure for recovery of transient events (Bechtold et al., 2024).
• Longevity: Magnet-arm usage remains well below the lifetime budget; after ~2.3 cycles/day, only ~8.4% of the mechanism’s life is expended, supporting long-term operation.

3. Spectral Performance, Throughput, and Sensitivity

• Spectral Resolution: Empirical in-flight characterizations show R(λ) in all modes meets or exceeds pre-launch predictions by 5–45%, with the largest gains at long wavelengths (Shajib et al., 4 Jul 2025). For example, MOS/high-R: $R(2\,\mu\mathrm{m})\sim4000$, $R(4\,\mu\mathrm{m})\sim9300$.
• Point-Source Sensitivity: For $S/N=10$ per resolution element:
  • PRISM: $f_\nu\lesssim100$ nJy (AB 26.4; 10×1000 s)
  • $R\sim1000$: $f_\nu\lesssim1$ μJy (AB 23.9)
  • $R\sim2700$: $f_\nu\lesssim3$ μJy (AB 22.7)
  • Assumes optimal centering; off-center degrades S/N up to ≈30% (Ferruit et al., 2022, Böker et al., 2023, Giardino et al., 2022).
• Photon Conversion Efficiency (PCE): Peaks at ≈0.50 around 2.5 μm (PRISM); measured in-flight PCE generally exceeds or matches models across 0.8–4 μm, with lower efficiency (<20%) below 1 μm and some reduction (>10%) above 4 μm due to PSF broadening/slit losses (Giardino et al., 2022).
• Detector Noise: Read noise (IRS² mode): 4–6 e– per pixel; dark current: 0.005–0.01 e– s⁻¹ pixel⁻¹. Total ramp noise for 1000 s: <10 e– (Birkmann et al., 2022).

Mode	R (at 2 μm)	PCE (peak)	Sensitivity (10 ks, S/N=10)
PRISM	330	0.50	$f_\nu\sim100$ nJy (AB 26.4)
G140M/H	1000–3700	0.45–0.55	$f_\nu\sim1–3$ μJy ($R\sim1000–2700$ modes)
IFU	see above	up to 0.45	Comparable to MOS/FS, especially at λ<2μm

4. Calibration, Astrometric, and Wavelength Solutions

• Instrument Geometric Model: NIRSpec calibration is underpinned by a parametric instrument model incorporating:
  • Coordinate transforms between optical planes, including distortion polynomials and misalignment parameters for MSA quadrants, IFU slices, and detectors (Lützgendorf et al., 2022, Oliveira et al., 2018).
  • Real-time ingestion of Grating Wheel Assembly (GWA) tip/tilt sensor voltage readings, converting to per-exposure angular offsets for any disperser position using linearized calibration fits (Oliveira et al., 2022).
• Calibration Pipeline: Stages include:
  1. Ramp fitting, cosmic ray/saturation flagging.
  2. Bias/subtraction, non-linearity and dark correction.
  3. Wavelength and spatial assignment using the physical model and GWA state.
  4. Pixel- and field-dependent flatfielding (D-, S-, F-flat); path-loss and slit- or bar-shadow correction via Fourier optics models, with empirical dither validation.
  5. Flux calibration using standard stars.
  6. Final product in absolute flux units (e.g., erg s⁻¹ cm⁻² Å⁻¹) (Oliveira et al., 2018, Lützgendorf et al., 2022).

• Achieved Precision: Residuals after model-based extraction and calibration are $\lesssim0.1$ pix in wavelength (1/10 resel), with target centroids in TA placed to $\lesssim0.05$″ ($<$1/4 shutter) (Oliveira et al., 2022, Böker et al., 2023).

5. Operational Workflow, MOS Planning, and Software Infrastructure

• MSA Planning: MOS observation preparation is supported by the ESA-provided eMPT (extended MSA Planning Tool), supplementing the baseline MSA Planning Tool (MPT) in APT. eMPT algorithmic sequence:
  1. Initial Pointing Algorithm: search for pointing/roll maximizing highest-priority targets.
  2. Candidate/viable slitlet identification and contamination rejection.
  3. Matrix (conflict-minimization) slit-assignment algorithm for optimal non-overlapping spectral assignment.
  4. Output: pointing, slit-mask, and target files for ingestion into APT (Bonaventura et al., 2023, Maseda et al., 2024).

• Slitlet, Dither, and Background Schemes: Standard MOS uses $1\times3$ or $1\times2$ adjacent shutter "slitlets" and nodding for background subtraction. Extended sources/faint backgrounds require tailored dither and masking strategies (Ferruit et al., 2022, Maseda et al., 2024).

• TA Procedures: MOS utilizes multi-reference-star MSA-TA for $\sim$25 mas target placement; FS/IFS use wide-aperture TA. Catalog astrometry must be aligned to Gaia for optimal TA performance (Böker et al., 2023).

6. Key Scientific Applications and Survey Results

• High-Redshift Extraterrestrial Surveys: The NIRSpec Wide GTO Survey demonstrates large-scale MOS efficiency: 3200 $z>1$ galaxies over 320 arcmin² in 105 hr, exploiting flexible MSA configurations for complex, prioritized science cases. PRISM enables census-level redshift surveys; high-R gratings enable kinematic and nebular diagnostics (Maseda et al., 2024).
• Exoplanet and Time Series: S1600A1 slit/subarray modes with low (~200 ppm) systematic noise floors allow high-precision transit and eclipse spectroscopy; simultaneous coverage 0.6–5.3 μm provides access to H$_2$O, CO$_2$, CH$_4$, and other atmospheric features (Birkmann et al., 2022).
• Integral-Field Spectroscopy: IFU mode enables spatially resolved spectroscopy of compact galaxies, star-forming regions, AGN, and Solar System targets. Path-loss corrections, sub-spaxel dithering, and leakage subtraction are essential for robust 3D data products (Böker et al., 2022).
• Galactic and Stellar Science: Stellar metallicity, brown dwarf, and Galactic cluster studies leverage the high S/N, multiplex, and wide spectral access.

7. Engineering Trends, Reliability, and Future Prospects

• Operational Stability: Over two years in orbit, NIRSpec MSA and mechanisms exhibit no significant degradation beyond predictable trends. Masked (shorted) shutter growth slowed after initial stabilization; FC and FO shutters remain within operational margins.
• Remaining Risks: Dominant residual risk is contamination-induced shorts, transient in nature and mitigated by ESD/OSD tools. Future systematic re-checking of short-masked lines will enable recovery of lost shutters, potentially raising $f_{op}>85\%$.
• Improvements: Science-like operability assessments and probabilistic per-shutter operability maps are proposed for further multiplex optimization (Bechtold et al., 2024).
• Science Pipeline Evolution: Calibration and pipeline updates are continuous. Instrument geometric model and flat-fields are refined post-commissioning for long-term fidelity.

NIRSpec’s combination of multiplexed, high-sensitivity, broad-wavelength infrared capability with sophisticated calibration and operational machinery marks it as an essential tool for contemporary astrophysical spectroscopy (Böker et al., 2023, Ferruit et al., 2022, Bechtold et al., 2024).