JWST/MIRI Medium-Resolution Spectrometer

• JWST/MIRI MRS is a cryogenic, four-channel integral field spectrometer offering moderate spectral resolution (R≈1,300–3,700) over 4.9–28.5 µm with simultaneous spatial and spectral coverage.
• It employs image-slicer IFUs across four channels and a multi-stage calibration pipeline to achieve high astrometric and photometric precision, with on-sky uncertainties as low as 50 mas.
• The instrument enables groundbreaking studies of the ISM, exoplanet atmospheres, and star formation by providing uniform, high-resolution spectral cubes and robust flux calibration.

The Medium-Resolution Spectrometer (MRS) of the JWST Mid-Infrared Instrument (MIRI) is a cryogenic, four-channel, integral field spectrometer providing moderate resolving power (R ≈ 1,300–3,700) from 4.9 to 28.5 µm, with simultaneous spatial and spectral coverage. Each channel operates via an image-slicer IFU, subdivided into three spectral sub-bands (A=SHORT, B=MEDIUM, C=LONG), all processed through multi-stage calibration pipelines. The MRS’s architecture, calibration, data reduction, and its astrophysical applications are summarized below.

1. Instrument Architecture and Optical Design

MIRI/MRS comprises four coaxial IFUs, each covering distinct wavelength intervals via dichroic/grating wheel assemblies. The spectrograph accepts the JWST f/12 output, with pre-optical elements collimating the beam and sequential dichroics splitting the passband into four channels:

• Channel 1: 4.9–7.6 μm; 21 slices; FOV ≈ 3.7″×3.7″; slice width ≈ 0.17–0.18″
• Channel 2: 7.5–11.87 μm; 17 slices; FOV ≈ 4.7″×4.5″; slice width ≈ 0.28″
• Channel 3: 11.6–18.24 μm; 16 slices; FOV ≈ 6.2″×6.2″; slice width ≈ 0.39″
• Channel 4: 17.7–28.3 μm; 12 slices; FOV ≈ 7.7″×7.7″; slice width ≈ 0.64″

Each sub-band is covered through grating-dichroic wheel settings, for a total of twelve sub-bands spanning 4.9–28.3 μm. Sliced light is dispersed and imaged onto two 1024×1024 Si:As BIB detector arrays (SW: Ch1+2, LW: Ch3+4) (Wells et al., 2015, Argyriou et al., 2023, Patapis et al., 2023).

2. Spectral Resolution, Field, and Calibration Performance

Spectral Resolving Power and Wavelength Coverage

The resolving power varies as R(λ)=λ/Δλ, with flight and ground measurements yielding:

• R ≈ 3,700 at 5 μm (channel 1A; best),
• R ≈ 1,300–1,500 at 28 μm (channel 4C),
• R(λ) = 4603 − 128 λ [μm] provides an in-flight fit (Argyriou et al., 2023, Jones et al., 2023).

Analysis of Fabry–Pérot etalon exposures and celestial calibrators yields absolute wavelength accuracy better than 0.1 pixel, or Δv < 10–30 km s⁻¹, across all bands, and internal repeatability at the 1% level (Argyriou et al., 2023, Jones et al., 2023, Labiano et al., 2021).

Geometric and Astrometric Calibration

A slice-by-slice polynomial distortion solution achieves relative astrometric accuracy of 8 mas (5 μm) to 23 mas (28 μm), and total on-sky uncertainty of ≈50 mas when including target acquisition and grating wheel repeatability (Patapis et al., 2023). Transformations map detector (x, y) to IFU (α, β), then to observatory (V2, V3) coordinates with <10% of a spatial resolution element per band.

Flux Calibration and Sensitivity

Calibration is linked to high S/N observations of spectrophotometric standards (O/A/G dwarfs) and, longward of 18 μm, bootstrapped with asteroid and circumstellar standards (Law et al., 2024). Photometric repeatability is <1% (5–18 μm) and a few percent thereafter. Absolute calibration agrees with MIRI imager and Spitzer IRAC/IRS to ≲1–2%. The 5σ line sensitivity for point sources in 10,000 s reaches ≲10⁻²¹ W m⁻² (Ch1–2), ≃3×10⁻²¹ W m⁻² (Ch4) (Argyriou et al., 2023, Vermot et al., 12 Nov 2025).

3. Detector Effects and Pipeline Corrections

Fringing and Flatfielding

Si:As BIB arrays form two dominant low-finesse Fabry–Pérot cavities; their interference fringes (peak-to-peak up to 15% pre-correction) are modeled as multiplicative Airy functions in wavenumber (Crouzet et al., 15 Apr 2025). Static fringe flats—derived from extended sources (e.g., NGC 7027)—reduce amplitudes to <1% in extended sources, with further reduction via residual fringe fitting. Point sources after full correction show 1–2% residuals per spaxel (Crouzet et al., 15 Apr 2025).

Spectral Stitching and Data Cube Construction

Twelve sub-band cubes require spectral stitching to correct for flux mismatches in overlap regions. Non-negative matrix factorization approaches (e.g., Haute Couture algorithm) achieve flux consistency via global scaling and matrix completion, delivering uniform, high-resolution cubes and preserving spatial fidelity over spatially varying PSF and band edge discontinuities (Canin et al., 17 Nov 2025).

3D Drizzle

Spectral cubes are assembled via a 3D drizzle algorithm that weights detector data onto (α, β, λ) cube voxels. Covariance between neighboring spaxels—arising from shared detector pixels—necessitates correction factors (1.5–3×) for error propagation in extracted spectra (Law et al., 2023). Four-point dither strategies are essential for suppressing residual undersampling systematics below 1%.

4. Calibration Pipeline Structure

The MRS calibration pipeline consists of detector-level corrections, spectro-photometric calibration, astrometric transformations, and construction of science-ready cubes (Labiano et al., 2016):

1. CALSPEC2: WCS assignment, pixel flat-fielding, stray-light removal, fringe correction, flux calibration.
2. CALSPEC3: Background matching/subtraction, residual latency and fringe correction, cube building, aperture extraction, leakage correction, multi-band cube stitching.

Two modes—Baseline (typical science) and Optimal (edge cases, e.g., faint extended regions, stray-light diagnostics)—invoke different sets of calibration products and algorithms. Standard reference files include pixel flats, fringe flats, WCS transforms, PSF-based aperture corrections, and spectral leak models.

5. On-Sky Performance and Scientific Applications

MRS consistently achieves or exceeds its design goals in spectral, astrometric, and photometric domains (Argyriou et al., 2023, Jones et al., 2023). Applications include:

• Warm H₂ and dust in metal-poor ISM: Direct detection of weak mid-IR H₂ rotational lines, analysis of excitation under LTE and PDR models, and measurement of non-equilibrium ortho-to-para ratios (OPR > 3) inaccessible before JWST/MIRI (Hunt et al., 2 Sep 2025).
• Exoplanet characterization: Time-series and integral-field datasets yield photon-limited precision for bright transiting systems and direct imaging of cool planets. The stability of spectral response (≲1.4×10⁻⁴ over weeks) enables cross-correlation analyses for molecular mapping (H₂O, CO, CH₄, NH₃, CO₂) (Deming et al., 2024, Patapis et al., 2021, Mâlin et al., 2023).
• Studies of ISM phases, supernova-enriched environments, and circumstellar material: Simultaneous detection of ionic, atomic, and molecular features enables phase decomposition and abundance measurements in regions such as the Galactic center (Vermot et al., 12 Nov 2025) and planetary nebulae (Jones et al., 2023).

6. Limitations, Systematics, and Best Practices

• Fringe correction: For point sources, residuals are limited by uncorrected high-frequency dichroic fringes and intra-band systematics; optimal pipeline modules and proper dither strategies are essential (Crouzet et al., 15 Apr 2025).
• Long-term throughput loss: A time-dependent decrease in throughput at λ ≳ 20 μm is tracked and corrected using regular internal calibrations (Law et al., 2024).
• Spatial undersampling: Channels 1–2 are undersampled; a four-point dither is mandatory to recover full spatial fidelity and suppress reconstruction artifacts (Law et al., 2023).
• Spectral leak correction: A small fraction (∼2.5%) of 6.1 μm light leaks into the 12.2 μm band and is subtracted in the pipeline (Law et al., 2024, Labiano et al., 2016).
• Aperture corrections: Extraction radii scale with λ to maintain a constant PSF fraction, with well-characterized correction curves for photometry (Law et al., 2024).

7. Impact and Scientific Legacy

The JWST/MIRI MRS delivers integral-field, moderate-resolution spectroscopy across the entire mid-infrared atmospheric window with photometric stability, astrometric precision, and systematic control surpassing prelaunch goals. It enables direct astrophysical inferences on ISM, star formation, planetary atmospheres, and dust mineralogy under conditions previously unattainable—establishing it as the reference instrument for mid-IR spectroscopy for the coming decade (Argyriou et al., 2023, Hunt et al., 2 Sep 2025, Deming et al., 2024).