CARMENES: Dual-Channel Exoplanet Spectrograph

• CARMENES Spectrograph is a dual-channel, high-resolution echelle instrument that precisely measures stellar radial velocities to detect low-mass exoplanets.
• Its design features fiber-fed visible and near-infrared arms housed in vacuum chambers with advanced calibration, achieving sub-m/s stability.
• The instrument supports extensive M-dwarf surveys and atmospheric studies through optimized scheduling and data reduction techniques, enhancing planet recovery.

CARMENES is an ultra-stabilised, dual-channel, high-resolution echelle spectrograph installed at the 3.5 m Calar Alto telescope in southern Spain. Designed by a German–Spanish consortium, the instrument operates continuously over the 0.52–1.71 μm range (simultaneously covering visible and near-infrared) at resolving powers R ≈ 80,000–100,000. The project’s core objective is the comprehensive radial velocity (RV) survey and characterization of exoplanets orbiting nearby, bright M-dwarf stars, with additional applications in transiting planet follow-up and atmospheric studies. CARMENES employs a state-of-the-art, fiber-fed, environmentally controlled design, pioneering stability and precision suitable for detecting terrestrial/low-mass planets—including those in habitable zones—and probing the properties of their host stars and planetary systems.

1. Instrument Design and Technical Architecture

CARMENES consists of two spectrographs: a visible (VIS) arm spanning 0.52–0.96 μm and a near-infrared (NIR) arm spanning 0.96–1.71 μm, both achieving R ≈ 80,000–100,000 spectral resolution (Amado et al., 2012, Caballero et al., 7 Mar 2025). Fiber feeds link the spectrographs to the telescope’s Cassegrain focus, providing stable light injection and minimizing mechanical flexure. The cross-dispersed echelle design minimizes spectral gaps and ensures broad, contiguous spectral coverage.

Thermal and mechanical stability are essential for sub-m/s RV precision. Both spectrographs are housed in vacuum chambers: the VIS channel is passively temperature-stabilized, while the NIR channel originally relied on cryogenic nitrogen cooling. As part of the CARMENES-PLUS upgrade, the NIR cooling system was re-engineered to use a continuous flow of cryogenic N₂ managed by a PID-controlled proportional valve, coupled with a pressure regulation unit and dedicated vacuum system for the transfer lines. These modifications reduced thermal fluctuations to ΔT ≈ 0.002 K, improved intrinsic NIR RV precision from 2.62 m/s to 0.67 m/s, and brought NIR stability closer to that of the VIS channel (Varas et al., 22 Sep 2025).

Calibration is achieved via hollow-cathode lamps (U–Ne, U–Ar, Th–Ne) and stabilized Fabry–Pérot etalons that monitor instrumental drifts. Dedicated algorithms ensure that temperature, pressure, and drift corrections systematically maintain the wavelength solution at the milliKelvin and sub-m/s level (Bauer et al., 2020, Varas et al., 22 Sep 2025).

2. Survey Strategy, Scheduling, and Operational Modes

The CARMENES consortium designed a five-year RV survey of ~300 M dwarfs employing 600+ guaranteed nights at Calar Alto (Amado et al., 2012). Survey design leverages a curated all-sky input catalogue ("CARMENCITA"), emphasizing the selection of single, bright, spectroscopically characterized targets (Alonso-Floriano et al., 2015). Target characterization employs low-resolution indices (spectral types, activity, metallicity, gravity) and high-resolution model-atmosphere fits (PHOENIX-ACES, BT-Settl) to derive accurate stellar parameters and optimize planet detectability (Passegger et al., 2016, Passegger et al., 2018, Tabernero et al., 29 Jul 2024).

Survey scheduling maximizes detection efficiency and observing time utilization. The CAST scheduler employs a hierarchical, multi-objective strategy based on the NSGA-II genetic algorithm, integrating observational constraints (visibility, elevation, moon phase), instrument overheads, and science priorities. CAST achieves >99% use of weather-approved telescope time, equitable coverage of all targets (SD ≈ 3 on the number of observations per target), and an anticipated planet recovery rate of ≈65% for signals with K > 1 m/s (photon noise limit) (Garcia-Piquer et al., 2017).

3. Radial Velocity Measurement and Precision

CARMENES uniquely provides simultaneous VIS and NIR RV measurements, allowing multi-wavelength diagnostics of both planetary and stellar signals. For most early- and mid-type M dwarfs (M0–M6), the 700–900 nm region yields optimal RV precision due to a high density of molecular and atomic lines and relatively high stellar photon flux (Reiners et al., 2017). The photon-limited RV uncertainty is expressed as

$$\delta v_\mathrm{rms} = \frac{c}{Q \cdot \mathrm{S/N}},$$

where $Q$ quantifies RV information content (Reiners et al., 2017).

Stellar spectra are extracted, wavelength-calibrated, and corrected for instrumental drifts using SERVAL and CARACAL pipelines (Kaminski et al., 2018, Caballero et al., 7 Mar 2025). The order selection process further refines the measurement by excluding orders contaminated by tellurics or with excessive noise, especially for the NIR where detectors and telluric absorption dominate (Bauer et al., 2020). Following the CARMENES-PLUS upgrade, the NIR channel achieves an intrinsic calibration precision of 0.67 m/s and nightly zero-point scatter of 3.9 m/s, approaching VIS performance (NZP scatter ≈ 2.5 m/s) (Varas et al., 22 Sep 2025).

Multi-wavelength RV measurements allow for the separation of wavelength-independent planetary signals from stellar activity-related RV variations, often chromatic. Gaussian process (GP) regression kernels are frequently used to model and remove activity-induced correlated noise in the RV time series (González-Álvarez et al., 2020, Stauffenberg et al., 16 Jul 2024).

4. Scientific Contributions: Planet Detection and Stellar Characterization

CARMENES has yielded a wealth of discoveries and major contributions to exoplanet and stellar astrophysics:

• M-dwarf exoplanet yield: The survey has detected Neptune-mass and super-Earth planets at low RV amplitudes (K ≈ 2–3 m/s), including habitable zone planets and systems in binaries (Kaminski et al., 2018, González-Álvarez et al., 2020).
• Stellar characterization: Accurate determinations of T_eff, log g, and [Fe/H] for M dwarfs using high-resolution spectra and spectral synthesis (PHOENIX-ACES, BT-Settl, Turbospectrum) improve the mass and radius determinations for planet hosts (Passegger et al., 2018, Tabernero et al., 29 Jul 2024). Line-by-line abundance analyses provide critical information on Mg, Si, and iron-to-silicate fractions relevant to rocky planet composition (Tabernero et al., 29 Jul 2024).
• Activity monitoring: The large spectral coverage enables the monitoring of chromospheric activity (e.g., Hα, Ca II IRT) and the analysis of activity–rotation relations. While CARMENES cannot cover the blue Ca II H&K lines directly, a time-resolved Ca II H&K catalog was assembled from complementary archival data for rotational and activity cycle studies, providing context for RV "jitter" and its correction (Perdelwitz et al., 2021).
• Binary systems and calibration: Double-line SB2 detection and characterization refine mass–luminosity relations for low-mass stars (Baroch et al., 2018).

5. Atmospheric and Ancillary Science

CARMENES's spectral range and precision have supported atmospheric characterization of transiting exoplanets and benchmark ultrahot Jupiters:

• Detection of atomic/ionic species (e.g., Ca II, Fe, Ti) in exoplanet atmospheres from high-resolution emission and transmission spectroscopy, confirming thermal inversions and retrieving chemical abundances (Casasayas-Barris et al., 2021, Guo et al., 29 Apr 2024).
• Cross-correlation techniques between high-resolution models and observed emission spectra (with post-processing of telluric and instrumental features via SYSREM and template division) enable the identification of species such as Fe I and Ti I in hot Jupiter dayside atmospheres, quantification of inversion layers, and measurement of planetary rotation (Guo et al., 29 Apr 2024).
• Telluric removal via template division telluric modeling (TDTM) exploits the Earth’s barycentric motion over time: high S/N stellar templates are constructed, divided into each observed spectrum, and fitted with synthetic transmission models, yielding a telluric-free, high-resolution spectrum for each target. This approach is optimized for late-type stars and enhances not only RV precision but also atmospheric retrievals (Nagel et al., 2023).
• Studies of wing asymmetries in chromospheric lines of M dwarfs reveal complex atmospheric dynamics (flares, coronal rain, chromospheric condensations) and inform corrections for RV noise (Fuhrmeister et al., 2018).

6. Consortium, Upgrades, and Scientific Legacy

CARMENES is an exemplar of a major German–Spanish collaborative effort, leveraging balanced technical and scientific expertise across instrument design, operations, and data analysis (Amado et al., 2012, Caballero et al., 7 Mar 2025). Rigorous environmental controls and systematic hardware upgrades—most notably the CARMENES-PLUS NIR cooling system—have strongly mitigated instrumental noise (Varas et al., 22 Sep 2025).

Continued and future upgrades focus on calibration (e.g., joint VIS+NIR Fabry–Pérot etalon cryostat), further stabilization, and next-generation data reduction pipelines. These advances will improve detection sensitivity to longer-period and lower-mass planets and maximize the synergy with ongoing and upcoming space missions such as TESS and PLATO.

The establishment of comprehensive, public high S/N, high-resolution template libraries for ~400 M dwarfs, and the homogeneously calibrated abundance and activity catalogs, provide fundamental resources for both current and future studies in stellar and exoplanet astrophysics (Nagel et al., 2023, Tabernero et al., 29 Jul 2024).

7. Impact, Limitations, and Future Prospects

CARMENES has substantially advanced the detection and characterization of small exoplanets around the most common stars in the solar neighborhood, with particular sensitivity in the parameter space previously inaccessible to optical-only surveys. Its dual-channel approach uniquely enables wavelength-dependent scrutiny of stellar activity, crucial for robust and precise exoplanet mass determinations and understanding planetary system architectures.

Some limitations remain intrinsic to ground-based RV work, including telluric contamination and stellar "jitter," but tailored techniques such as TDTM, order selection, and GP regression, as well as ongoing instrument upgrades, continue to mitigate these challenges (Nagel et al., 2023, Varas et al., 22 Sep 2025). The integration of CARMENES capabilities with atmospheric spectroscopy and planetary interior modeling is expected to become increasingly important, especially in the detailed paper of rocky planet formation and exoplanet habitability around M dwarfs.

As a legacy project, the data products, technical solutions, and survey design principles of CARMENES serve as a benchmark for the exoplanet community and inform the conceptualization of future instruments targeting the low-mass regime at both ground and space facilities.