SPIRou: nIR Spectropolarimetry & Exoplanet Studies

• SPIRou is a near-infrared high-resolution spectropolarimeter and velocimeter that integrates broad spectral coverage and advanced calibration techniques to study exoplanets and stellar magnetic fields.
• It features a stabilized cryogenic spectrograph, achromatic polarimeter, and robust data pipelines to achieve sub-meter-per-second radial velocity precision.
• Its legacy surveys, including programs targeting M dwarfs and young stellar objects, exemplify its role in revealing exoplanet demographics and magnetic field dynamics.

SPIRou is a near-infrared (nIR) high-resolution spectropolarimeter and high-precision velocimeter designed and installed at the 3.6-m Canada–France–Hawaii Telescope (CFHT) on Maunakea. Its broad spectral coverage (0.95–2.50 μm; resolving power R ≈ 70,000–75,000) and simultaneous spectroscopic and polarimetric capabilities make it uniquely suited to advance two primary science goals: detecting and characterizing low-mass exoplanets (particularly Earth-like and habitable-zone planets) around M dwarfs via precise radial-velocity (RV) measurements, and probing the role of magnetic fields in star and planet formation. SPIRou's instrument architecture—comprising a stabilized cryogenic spectrograph, an achromatic polarimeter, a sophisticated calibration unit, and robust data reduction and analysis pipelines—has set new standards in high-precision nIR velocimetry and stellar spectropolarimetry, enabling a series of legacy surveys and science programs probing exoplanet demographics, dynamo processes, and exoplanet atmospheres.

1. Instrument Architecture and Capabilities

SPIRou is constructed around several integrated subsystems designed to maximize RV precision and polarimetric sensitivity in the nIR regime:

• Spectral Coverage and Resolution: Simultaneous YJHK-band coverage from 0.95 to 2.50 μm, with a spectral resolving power R ≈ 70,000–75,000, is achieved with a fiber-fed, bench-mounted, cross-dispersed echelle spectrograph. The instrument design ensures high throughput (up to 15%) and stable illumination, leveraging a unique combination of fluoride fibers, an advanced pupil slicer, and a stabilized cryostat operating at 80 K (Santerne et al., 2013, Artigau et al., 2014, Donati et al., 2018, Donati et al., 2020).
• Achromatic Polarimeter: The Cassegrain-mounted front-end unit incorporates Fresnel rhombs and a Wollaston prism to separate and analyze orthogonal polarization states prior to fiber injection. This allows acquisition of both intensity (Stokes I) and circular/linear polarization (Stokes V, Q/U) spectra, enabling simultaneous spectropolarimetry and high-precision velocimetry (Delfosse et al., 2013, Moutou et al., 2015, Moutou et al., 2020).
• Thermal and Mechanical Stability: The spectrograph bench is suspended within a multi-layered cryostat achieving temperature stability at the milli-Kelvin level. The instrument's optical bench, collimators, and imaging components are isolated with a hexapod mount and thermal shields, ensuring internal RV drifts remain below 0.7 m/s overnight, and sub-meter-per-second correction is possible using a dedicated calibration module (Artigau et al., 2014, Boisse et al., 2016, Donati et al., 2018).
• Calibration Unit: The wavelength calibration system employs a uranium-neon (UNe) hollow-cathode lamp for absolute calibration, a stabilized Fabry–Pérot (FP) etalon for relative wavelength referencing, and flat-field lamps. Simultaneous calibration light may be injected during science exposures to monitor and correct for instrumental drifts (Boisse et al., 2016, Cersullo et al., 2017, Hobson et al., 2021).
• Detector System: The spectrograph uses a mosaic H4RG-15 HgCdTe infrared array (4096×4096 15 μm pixels), with careful pre-processing in the reduction pipeline to address detector-specific noise (1/f noise, amplifier drifts, hot pixels, cosmic-ray events) (Artigau et al., 2014, Cook et al., 2022).

2. Data Reduction, Calibration, and Precision RV Techniques

SPIRou's data acquisition is supported by the APERO pipeline, which orchestrates a modular, multi-phase reduction process:

• Preprocessing and Extraction: APERO corrects for detector artifacts and cosmic rays, applies pixel grid registration, and aligns curved echelle orders using affine transformations derived from FP and HC exposures. Orders are optimally extracted and barycentric corrections applied (Cook et al., 2022).
• Wavelength Calibration: A combined solution using UNe lamp transitions and the dense, evenly-spaced FP etalon lines achieves a pixel-wavelength mapping internal error of ∼0.15 m/s, well within the sub-meter-per-second requirements for exoplanet detection. The FP cavity width d(λ) is modeled using a high-order polynomial fit, correcting for chromatic penetration in the mirror coatings (Hobson et al., 2021).
• Telluric Correction: Rigorous correction employs TAPAS synthetic transmission models, empirical hot-star templates, airmass detrending, and principal component analysis to mitigate time-variable atmospheric transmission and correlated noise—particularly crucial in the nIR (Moutou et al., 2023, Cook et al., 2022, Debras et al., 2023).
• Radial Velocity Extraction: While cross-correlation techniques using weighted masks are available, science-grade RVs are primarily derived using a line-by-line (LBL) approach. The LBL method extracts RVs from thousands of spectral lines individually and combines them using robust statistical weighting, with further dimensionality reduction and systematics correction (via Weighted PCA in the “Wapiti” method) to eliminate non-astrophysical trends (Moutou et al., 2023, Ould-Elhkim et al., 10 Feb 2025).
• Polarimetric Product Generation: Full polarimetric analysis is performed by combining multiple exposures in different retarder plate states. Both “Ratio” and “Difference” methods are used to calculate Stokes parameters, with quality control provided by null diagnostic spectra (Cook et al., 2022).

APERO utilizes scalable, multi-core architectures and SQL-based databases to process large data volumes spanning several years, facilitating fast turnaround for survey and time-domain applications (Cook et al., 2022).

3. Scientific Programs and Legacy Surveys

SPIRou serves as the core instrument for the SPIRou Legacy Survey (SLS), a CFHT Large Program allocated hundreds of nights, designed to address several major science cases:

• Exoplanet Discovery via SLS-PS: The SLS-Planet Search (SLS-PS) targets ~100–330 nearby M dwarfs, conducting dense RV monitoring to detect low-mass planets with RV amplitudes down to 1 m/s. Monte Carlo simulations predict a detection yield of $85.3^{+29.3}_{-12.4}$ planets (0.5–200 d period), including $20.0^{+16.8}_{-7.2}$ habitable-zone planets and $8.1^{+7.6}_{-3.2}$ Earth-like worlds, for occurrence rate constraints $\eta_\oplus$ at a ∼45% precision (Cloutier et al., 2017).
• Magnetism and Star Formation with SLS-MP: The SLS-Magnetic Protostar/Planet (SLS-MP) survey maps large-scale surface magnetic fields (via Stokes V and Zeeman–Doppler Imaging, ZDI) of young stellar objects (protostars through weak-line T Tauri stars) and M dwarfs, investigating dynamo regimes, magnetic topology evolution, and the role of magnetism in disk accretion and angular momentum regulation (Moutou et al., 2017, Donati et al., 2023).
• Atmospheres and Transmission Spectroscopy: SPIRou enables high-resolution transmission spectroscopy of transiting exoplanets, facilitating atmospheric retrievals (including abundance, C/O ratio, and limb-to-limb dynamics) from the NIR molecular bands of H₂O, CO, CH₄, NH₃, and atomic tracers. Dedicated pipelines (e.g., ATMOSPHERIX) employ Bayesian retrievals and cross-correlation techniques to extract weak planetary absorption or emission in the presence of correlated systematics and tellurics (Debras et al., 2023, Hood et al., 28 Mar 2024).
• Multi-Instrument Synergy: SPIRou data are systematically combined with optical RV datasets (from HARPS, CARMENES, etc.), photometric transits (TESS, PLATO, CHEOPS), and mm/sub-mm imaging (ALMA), optimizing the discrimination of planetary signals from stellar activity, and maximizing detection and characterization efficiency across wavelength domains (Ould-Elhkim et al., 10 Feb 2025, Santerne et al., 2013).

4. Methodological Advancements: Activity Filtering and Magnetic Topology

Mitigating stellar activity in M dwarfs is critical for reliable planet detection:

• Activity Indicators and Diagnostics: Multiple proxies—chromospheric emission indices, Zeeman broadening (via FeH lines), rotation periods, and longitudinal field B_\ell—are monitored and modeled. An “activity merit function” (AMF) is constructed to prioritize “quiet” targets for planet search campaigns (Moutou et al., 2017).
• Gaussian Process (GP) Regression: Surveys employ quasi-periodic GP models, with hyperparameters constrained by ancillary activity indicators (dLW, dET, Hα, Ca II IRT, etc.), to model and remove correlated noise in the RV time series. The importance of GP prior selection is evidenced by the sensitivity of candidate planet detection to the hyperparameter values, especially near the detection threshold for low-amplitude signals (Ould-Elhkim et al., 10 Feb 2025, Cloutier et al., 2017).
• Direct Magnetic Mapping: ZDI from time-series spectropolarimetry maps the poloidal/toroidal components and axisymmetry of large-scale magnetic fields, enabling studies of dynamo operation across the fully convective boundary. Bayesian frameworks with quasi-periodic GP kernels provide reliable inferences of stellar rotation periods even in weakly magnetic, slowly rotating M dwarfs (Donati et al., 2023, Moutou et al., 2017).

5. Exoplanetary and Stellar Results

SPIRou has delivered several key science results:

• Confirmed Planets: Detection of a $9.5537 \pm 0.0005$ d planet with $M_p \sin i = 8.8 \pm 0.7\,M_\oplus$ orbiting Gl 480, and non-detections with strong upper limits in cases dominated by stellar activity (Gl 382), illustrating the strong utility of nIR RVs coupled with advanced activity filtering (Ould-Elhkim et al., 10 Feb 2025).
• Multiplanet System Characterization: Revisiting systems like GJ 876 and GJ 1148 with long-baseline SPIRou data, including N-body dynamical modeling (Wisdom–Holman integrator, MCMC fitting of $e$, $\omega$, $K$, $\lambda$), and explicit tracking of mean-motion and Laplace resonant angles ($\phi_{12,k} = 2\lambda_2 - \lambda_1 - \omega_k$ and $$\phi_{\text{Lap}} = \lambda_c - 3\lambda_b + 2\lambda_e$$) demonstrating resonance and chaotic orbital evolution (Moutou et al., 2023).
• Atmospheric Characterization: Recovery of H$_2$O and CO in WASP‑76 b, estimation of atmospheric abundances ($$\log(\mathrm{H}_2\mathrm{O})_\mathrm{MMR} = -4.52 \pm 0.77$$, $\log(\mathrm{CO})_\mathrm{MMR} = -3.09 \pm 1.05$) and a C/O ratio of $0.94 \pm 0.39$ ($\sim$1.7$\times$ solar), along with evidence for day/night limb asymmetry in line Doppler shifts—signifying 3D atmospheric structure and dynamics (Hood et al., 28 Mar 2024).
• Stellar Magnetic Trends: For M dwarfs with rotation periods from ∼15 to >400 d, the large-scale field strength does not decrease with Rossby number as in more massive stars, indicating possible differences in dynamo regime below the fully convective threshold. Temporal evolution and polarity switching are observed, suggesting long magnetic cycles (Donati et al., 2023).

6. Limitations, Challenges, and Future Directions

• Instrumental Systematics: Detector persistence, residual instrumental systematics (modal noise, fiber illumination), and imperfect modal scrambling remain challenges at the few m/s level. Planned upgrades (e.g., improved fiber scramblers, adoption of a laser frequency comb) will further stabilize wavelength calibration and throughput (Donati et al., 2020, Cook et al., 2022).
• Telluric and Activity Residuals: Although sophisticated correction algorithms exist, strong telluric absorption in bands (especially in water-dominated regions) and the complexity of stellar activity require on-going improvement in statistical modeling and data selection for both RV and atmospheric analyses (Debras et al., 2023, Cook et al., 2022, Moutou et al., 2023).
• Spectral Modeling Uncertainties: Systematic offsets exist between stellar parameter estimates from PHOENIX and MARCS synthetic spectral grids, with typical discrepancies up to 30–50 K in $T_{\rm eff}$ or 0.05–0.4 dex in $\log g$ and [M/H], especially problematic for atmospheric retrievals and planet property inference (Cristofari et al., 2021).
• Degeneracies in Atmospheric Retrievals: Retrieval of chemical abundances is subject to significant degeneracies, especially between mass, radius, and mean molecular weight (as the transmission spectrum amplitude is sensitive to $H\,R_p/R_\star^2$; $H = \mathcal{R}\,T/\mu g$). Interpreting multi-species signals and separating overlapping absorbers remains challenging at current SNRs (Debras et al., 2023, Hood et al., 28 Mar 2024).

SPIRou remains at the forefront of nIR high-precision velocimetry, spectropolarimetry, and exoplanet characterization. Its legacy surveys and ongoing upgrades will continue to yield high-fidelity constraints on both exoplanet demographics and the physics of low-mass stars and their planetary systems as complementary facilities (JWST, ELT, PLATO, TESS) come online.