H3 Spectroscopic Survey: Outer Halo Chemodynamics

Updated 2 July 2026

H3 Spectroscopic Survey is a ground-based, high-resolution stellar spectroscopy program designed to map the Milky Way halo and outer disk through detailed kinematics and chemical abundances.
It utilizes the Hectochelle spectrograph on the 6.5 m MMT to deliver robust radial velocities and precise stellar parameters, supporting studies of Galactic substructure and mass estimates.
The survey's analysis pipeline combines Bayesian spectral fitting with Gaia astrometry to yield accurate distances, chemical abundances, and kinematic measurements for a vast sample of stars.

The H3 Spectroscopic Survey is a ground-based, medium- to high-resolution stellar spectroscopic program expressly designed to probe the chemodynamical structure of the Milky Way halo and outer disk. It delivers radial velocities, detailed chemical abundances, and precise spectrophotometric distances for an unprecedented volume- and depth-limited sample of stars. By leveraging its unique design and high-resolution capabilities, the survey underpins a range of scientific studies—from the mapping of Galactic substructure and accretion signatures to fundamental Milky Way mass constraints and the chronology of ancient star formation.

1. Scientific Objectives and Survey Rationale

The H3 Spectroscopic Survey was designed to address key open questions in Galactic archaeology, particularly the hierarchical assembly of the Galactic halo through the accretion and disruption of dwarf galaxy progenitors. ΛCDM-based galaxy formation simulations predict that the outer halo contains numerous remnants of these accretion events, preserved in phase-space and chemical-abundance substructure. Full exploitation of such predictions requires six-dimensional (position and velocity) phase-space information as well as multi-element chemical tagging.

Gaia DR2 astrometry supplies precise positions and proper motions for >1 billion stars to G ≈ 20, but radial velocities and metallicities are confined primarily to G ≲ 15, limiting the reach to about 10 kpc. H3 targets the gap by delivering high-resolution (R~23,000) spectra for stars to r = 18, pushing precision chemical and kinematic mapping to distances of ≳100 kpc in the outer halo. The key survey goals are:

Identification and characterization of kinematic and chemical substructure in the halo, essential for reconstructing the accretion history of the Milky Way.
Dynamical constraints on the total mass and mass profile of the Milky Way out to large radii.
Age-dating and orbital analysis of metal-poor stars to test models of early disk and halo assembly.
Direct chemodynamical mapping of rare populations, such as very metal-poor stars, accreted halo subcomponents, and distant giants (Conroy et al., 2019, Carter et al., 2020, Zaritsky et al., 2019).

2. Instrumentation, Footprint, and Target Selection

H3 utilizes the Hectochelle multi-object echelle spectrograph on the 6.5 m MMT. Hectochelle provides:

Up to 240 fibers over a 1° field of view.
Resolution R ≈ 23,000, covering 5150–5300 Å, encompassing strong metallic (Mg b) and Fe I lines.
Sky coverage: ≈15,000 deg² (Dec > –20°, |b| > 30°), sparsely sampled in ≈3° tiles for efficient mapping.
Target selection is conducted with Gaia DR2 astrometry. The main criteria are 15 < r < 18 and π – 2σ_π < 0.5 mas (or π < 0.4 mas in the main selection), efficiently isolating distant (d ≳ 2 kpc) halo candidates with high completeness.
High-value science populations (e.g., photometric K giants, blue horizontal-branch stars, RR Lyrae) are included via photometric and parallax selection; rare objects reach ≈100 kpc.

This design yields kinematically and chemically clean outer-halo samples with a halo fraction (|V–V_LSR| > 220 km/s) of ≈20%, rising to higher Galactic latitude, and supports robust statistical coverage for rare distant populations (Conroy et al., 2019).

3. Data Acquisition, Reduction, and Parameter Inference

Each field undergoes nominal 30-minute exposures (3 × 10 min), with typical S/N per pixel ≈6 (r = 15–18). Hectochelle data are reduced via:

Bias subtraction, flat-fielding, wavelength calibration (ThAr arcs), fiber extraction, sky subtraction (from dedicated sky fibers), and optimal profile extraction.
Spectra are processed through MINESweeper, a Bayesian fitting framework that performs joint analysis of high-resolution spectra, broadband SEDs (Pan-STARRS, SDSS, Gaia, 2MASS, WISE), and parallax. Fitting is performed using ATLAS12/SYNTHE LTE synthetic spectra and MIST isochrones.
Key free parameters: stellar mass, age, [Fe/H], [α/Fe], distance, A_V, radial velocity, rotational/macroturbulent broadening, instrumental resolution, and continuum coefficients.
Typical formal precisions: σTeff ≈ 20–50 K, σ_logg ≈ 0.05–0.15 dex (dwarfs/giants), σ[Fe/H] ≈ 0.05–0.10 dex, σ_[α/Fe] ≈ 0.05–0.10 dex, σ_RV ≈ 0.5–1 km/s, and σ_D/D ≈ 5–8%.

The analysis pipeline directly yields full posterior samples for all parameters, facilitating forward modeling of the Milky Way’s chemodynamical structure and robust uncertainty propagation into downstream applications (Conroy et al., 2019).

4. Early Science and Legacy Results

The H3 Survey has delivered results across several domains:

Halo Substructure: The survey resolves phase-space clumps in energy–angular momentum and [Fe/H]/[α/Fe], allowing direct identification of accretion features and comparison to Gaia-Enceladus, Helmi streams, and additional structures beyond Gaia's radial-velocity reach.
Very Metal-Poor Stars: H3 isolates 482 very metal-poor ([Fe/H] < –2) dwarfs near the disk (1 ≲ |Z| ≲ 3 kpc). These are uniformly old (≈12 Gyr) regardless of [α/Fe] or metallicity, with >70% on prograde orbits—a signature inconsistent with purely isotropic accreted populations and implying both in situ and aligned accreted components to the ancient disk (Carter et al., 2020).
Outer Halo Kinematics and the Milky Way Mass: Applying the classical timing argument to 32 stars at R > 60 kpc yields a 90% confidence lower limit on the Galaxy's virial mass: M_{200} > 0.91 × 10¹² M_⊙. Simulation-based calibration (Auriga) reinforces that previous low-mass solutions are disfavored, and the most likely value is ≈1.4 × 10¹² M_⊙. The scale and selection of H3 make such bounds robust against substructure and distance bias (Zaritsky et al., 2019).
Chemo-dynamical mapping: Joint construction of the 6D phase-space plus chemistry dataset enables reconstruction of the accretion history, estimation of the halo density profile, binary statistics, and quantification of outer halo perturbations (e.g., from the LMC).

5. Survey Comparison, Technical Distinctions, and Data Products

H3 occupies a unique niche relative to contemporary Milky Way spectroscopic surveys:

Survey	Resolution (R)	Depth (r mag)	Halo Fraction	Main Strength
H3	~23,000	18	~20%	Outer halo chemodynamics
APOGEE	~22,500	16.5 (H)	~2–4%	Inner MW, disk, bulge
GALAH	~28,000	14	~2–4%	Disk, chemistry
Gaia-ESO	~20,000	19 (VLT)	<5%	Bulge, inner disk

H3's combination of moderate-to-high resolution and relatively deep limiting magnitude, matched to a broad sky footprint, ensures an unrivaled sample for detailed chemical and dynamical study of the outer halo (10–100 kpc). The pipeline's Bayesian distance inference yields spectrophotometric distances with ≲8% uncertainty even for low-S/N data and severely limited Gaia parallaxes (Conroy et al., 2019).

Public data releases include radial velocities, T_eff, log g, [Fe/H], [α/Fe], extinction, distances, and full posterior samples, with the catalog size projected to reach ≈200,000 stars (Conroy et al., 2019).

6. Broader Scientific Implications and Future Prospects

The H3 Survey's design directly supports key programmatic advances in near-field cosmology:

Provides vital radial-velocity and chemical input for leveraging Gaia's proper motions/parallaxes in mapping hierarchical assembly.
Supplies statistically clean, outer-halo tracer samples critical for robust Milky Way mass estimates, constraining models of satellite accretion and LMC dynamical perturbations.
Enables detailed investigations of rare populations—carbon-enhanced, fast rotators, binary fractions, and metal-poor giants at large distances.
The unique combination of depth, resolution, and wide-area selection maximizes the halo science return of upcoming large-scale photometric and astrometric datasets.
Forthcoming expansions (e.g., through complementary surveys: e.g., DESI, WEAVE, 4MOST) will further augment the census of distant halo stars, driving down statistical and systematic errors in both mass and accretion history constraints (Zaritsky et al., 2019).

A salient result of the initial H3 phase is the empirical demonstration that prograde ancient very metal-poor dwarfs—linked to both in situ and aligned-accreted origins—coexist near the disk, and that the angular momentum vector of the Galaxy has remained stably oriented for more than 12 Gyr (Carter et al., 2020). These findings challenge prior assumptions of fully isotropic accreted halos and underpin new paradigms for disk formation and the persistence of kinematical alignment over cosmic time.

7. Summary Table: H3 Survey at a Glance

Attribute	Value/Implementation
Telescope	MMT 6.5 m
Instrument	Hectochelle
Spectral Resolution	R ≈ 23,000 (σ_instr ≈ 5.5 km/s)
Wavelength Window	5150–5300 Å
Field Size	1° diameter, 240 fibers/field
Footprint	Dec > –20°,
Mag. Range	15 < r < 18 (main), to 18.5 (fill)
Main Selection	Gaia DR2 parallax cuts
Sample Size	≈200,000 (final)
Median S/N/pixel	~6 (r = 15–18)
[Fe/H] Precision	0.05–0.10 dex
[α/Fe] Precision	0.05–0.10 dex
Distance Precision	5–8%
Halo Fraction	≈20%

The H3 Spectroscopic Survey sets the current benchmark for outer-halo chemodynamical studies and will continue to serve as a reference dataset for Galactic structure and evolution research (Conroy et al., 2019, Carter et al., 2020, Zaritsky et al., 2019).