I’itoi: Metal-poor Halo Substructure

Updated 2 July 2026

I’itoi is a chemo-dynamically defined, high-energy, metal-poor stellar substructure in the Milky Way halo, identified by strict dynamical (e.g., E > -1.25×10^5 km²/s², Lz criteria) and chemical ([Fe/H] < -2.0) selection.
Its chemical profile features a median [Fe/H] around -2.39 dex with a pronounced α-enhancement (≈+0.38 dex), indicating a rapid star-formation history with minimal Type Ia supernova contributions.
Orbital analyses reveal moderately eccentric trajectories (e ~ 0.47) with pericenters near 5 kpc and extended apocenters, supporting its origin as a disrupted low-mass dwarf galaxy accreted in the early Milky Way.

I’itoi is a chemo-dynamically defined, high-energy, metal-poor, retrograde substructure in the Milky Way’s stellar halo. Identified initially in the H3 Survey combined with Gaia astrometry, I’itoi is characterized by its distinctive locus in phase space, its extreme chemical primitivity, and its status as the most metal-poor accretion remnant among the disrupted dwarf galaxy progenitors that built the halo. The detection and characterization of I’itoi and its analogues provide critical constraints on the minimal-mass regime of galactic accretion, the chemical evolution of low-mass systems at high redshift, and the inventory of halo building blocks accessible to present-day chemodynamical surveys (2006.08625, Naidu et al., 2022, Sante et al., 12 Mar 2026).

1. Identification and Membership Criteria

I’itoi is defined by strict selection in both dynamical and chemical space. In kinematic terms, member stars must satisfy:

High orbital energy: $E_{\text{tot}} > -1.25\times10^5$ km $^2$ s $^{-2}$
Strongly retrograde angular momentum: $L_z > 0.7\times10^3$ kpc km s $^{-1}$ (in H3/Gaia sign convention; $L_z<0$ in other works due to different conventions)
Circularity $\eta > 0.15$
Low metallicity: $[\mathrm{Fe/H}] < -2.0$ These cuts isolate stars on moderately eccentric, retrograde orbits occupying a narrow region in energy– $L_z$ space, and exclude previously defined substructures (e.g., Sagittarius, Sausage/Enceladus, Sequoia, Arjuna). The canonical H3 sample identifies $N=65$ members, with other surveys (e.g., APOGEE+Gaia, LAMOST+Gaia) yielding similar populations under analogous selection (2006.08625, Horta et al., 2022, Ye et al., 2023, Liu et al., 2024).

2. Chemodynamical Properties

Chemical Abundances

I’itoi stars possess the most metal-poor metallicity distribution among disrupted dwarfs, with a median $^2$ 0 dex and interquartile range $^2$ 1 dex. The distribution is sharply peaked below $^2$ 2, with $^2$ 3 of members at $^2$ 4 (2006.08625, Naidu et al., 2022). Representative mean and spread from various samples:

Reference	Median $^2$ 5	Median $^2$ 6	$^2$ 7 (dex)
H3+Gaia (2006.08625)	$^2$ 8	$^2$ 9	$^{-2}$ 0
APOGEE+Gaia (Horta et al., 2022)	$^{-2}$ 1	$^{-2}$ 2	$^{-2}$ 3 (MDF)
t-SNE (Youakim et al., 30 Oct 2025)	$^{-2}$ 4	--	$^{-2}$ 5

The $^{-2}$ 6-element enhancement, $^{-2}$ 7 dex, is among the highest measured for any disrupted component, indicating an extremely short star-formation duration with minimal Type Ia SNe contribution. Enrichment in [Mg/Fe] and [Ca/Fe] is consistent with this origin (Horta et al., 2022, Shejeelammal et al., 2023, Naidu et al., 2022).

Orbital Parameters

The typical orbital configuration includes:

Orbital eccentricity $^{-2}$ 8 (range $^{-2}$ 9)
Pericenter $L_z > 0.7\times10^3$ 0 kpc
Apocenter $L_z > 0.7\times10^3$ 1 kpc (tail to $L_z > 0.7\times10^3$ 2 kpc in some catalogs)
Maximum vertical excursion $L_z > 0.7\times10^3$ 3 kpc
$L_z > 0.7\times10^3$ 4 kpc km s $L_z > 0.7\times10^3$ 5 ( $L_z > 0.7\times10^3$ 6 to $L_z > 0.7\times10^3$ 7 kpc km s $L_z > 0.7\times10^3$ 8 in other conventions)
Total energy median $L_z > 0.7\times10^3$ 9 km $^{-1}$ 0 s $^{-1}$ 1

Kinematic decomposition in the integrals-of-motion space ( $^{-1}$ 2, $^{-1}$ 3, $^{-1}$ 4) robustly distinguishes I’itoi from Sequoia, Arjuna, and in-situ halo populations (Liu et al., 2024, Ye et al., 2023, Malhan et al., 2022).

3. Astrophysical Origin and Halo Assembly

Stellar population synthesis and simulation-based inference methodologies estimate the progenitor stellar mass of I’itoi at $^{-1}$ 5 ( $^{-1}$ 6– $^{-1}$ 7), with a halo mass at infall of $^{-1}$ 8 (Naidu et al., 2022, Sante et al., 12 Mar 2026). The mass–metallicity relation for disrupted dwarfs places I’itoi $^{-1}$ 9 dex below the $L_z<0$ 0 mass–metallicity relation, corresponding to a quenching redshift $L_z<0$ 1. Simulation-based inference yields an infall lookback time $L_z<0$ 2 Gyr and halo mass ratio $L_z<0$ 3 relative to the Milky Way potential at infall (Sante et al., 12 Mar 2026).

The extremely $L_z<0$ 4-enhanced abundance pattern, very low metallicity, and moderately eccentric orbits point toward a progenitor subject to short ( $L_z<0$ 50.5 Gyr), intense early star formation, minimal SNe Ia enrichment, and rapid quenching upon accretion. No present-day classical dwarf galaxy is dynamically associated; most I’itoi field stars are now phase-mixed in the inner halo, though a handful of globular clusters and streams (e.g., GD-1, Kshir, NGC 3201, Phlegethon) share its chemo-dynamical properties (Malhan et al., 2022, Youakim et al., 30 Oct 2025).

4. Chemo-kinematical Associativity and Substructure Taxonomy

Multiple clustering algorithms—energy–action space SNN (Ye et al., 2023), HDBSCAN in $L_z<0$ 6– $L_z<0$ 7– $L_z<0$ 8 (Liu et al., 2024), t-SNE chemo-kinematical embedding (Youakim et al., 30 Oct 2025), and hierarchical ENLINK on actions (Malhan et al., 2022)—independently recover I’itoi as a coherent, highly retrograde, compact and metal-poor substructure. Its membership is robust to variation in selection methodology, with minor systematic shifts depending on the dynamical conventions adopted.

Among retrograde halo components, I’itoi is differentiated from Sequoia and Arjuna primarily by its lower metallicity and more extreme $L_z<0$ 9-enhancement. While some analyses (e.g., Ye et al.) cannot always cleanly separate I’itoi from its retrograde siblings and treat the ISA (I’itoi-Sequoia-Arjuna) complex as a phase-space continuum, MDFs and clustering in action–energy space persistently identify I’itoi as a distinct, metal-poor mode (Kim et al., 22 Aug 2025, Ye et al., 2023).

5. Detailed Chemical Evolution and Nucleosynthetic Record

High-resolution spectroscopy of I’itoi stars, including CEMP-no and r-process–enriched objects, indicates a chemically inhomogeneous enrichment history. Within I’itoi, the dispersion in [Eu/Fe] spans the full range of $\eta > 0.15$ 0-process enhancement (non-, r-I, r-II, actinide-boost). In the case of the actinide-boost star LAMOST J122216.85–063345.2 ([Eu/Fe] = +0.61, $\eta > 0.15$ 1), $\eta > 0.15$ 2-process yields are reproduced by neutron-star–merger and black hole–neutron-star merger models, with MR-SNe contributing only to lighter $\eta > 0.15$ 3-process elements (Jeong et al., 31 Mar 2026).

The variety of $\eta > 0.15$ 4-process enrichment levels requires multiple discrete nucleosynthetic events and inhomogeneous mixing in the low-mass progenitor ISM, with spatial scales for mixing estimated at $\eta > 0.15$ 5 pc. The uranium–thorium ratio in actinide-rich I’itoi members constrains an enrichment-to-star-formation timescale of $\eta > 0.15$ 6–11 Gyr, underscoring prolonged retention and incomplete mixing of merger ejecta in the progenitor (Jeong et al., 31 Mar 2026).

6. Associated Streams, Globular Clusters, and MDF Structure

I’itoi’s debris is represented among several dynamically cold stellar streams (GD-1, Kshir, Phlegethon, Gaia-9, Gjöll, Ylgr) and two globular clusters (NGC 3201, NGC 6101), all tightly clustered in $\eta > 0.15$ 7 phase space (Malhan et al., 2022). Mean orbital parameters for these systems span pericenters of 5–14 kpc, apocenters up to $\eta > 0.15$ 8 kpc, eccentricities $\eta > 0.15$ 9– $[\mathrm{Fe/H}] < -2.0$ 0, and inclinations $[\mathrm{Fe/H}] < -2.0$ 1– $[\mathrm{Fe/H}] < -2.0$ 2. The metallicity distribution of these groupings ranges from $[\mathrm{Fe/H}] < -2.0$ 3 to $[\mathrm{Fe/H}] < -2.0$ 4, with streams generally more metal-poor and the clusters peaking at higher $[\mathrm{Fe/H}] < -2.0$ 5 within I’itoi’s envelope.

Some clustering approaches conflate Arjuna, Sequoia, and I’itoi due to phase-mixing and similar kinematic footprints, but continuous MDF and $[\mathrm{Fe/H}] < -2.0$ 6-enhancement gradients—combined with orbital clustering—support a hierarchical, possibly time-dependent sequence of accretion for these systems (Liu et al., 2024, Malhan et al., 2022, Youakim et al., 30 Oct 2025).

7. Chronology and Assembly Context

Isochrone fitting to H3 main-sequence turn-off and subgiant members yields I’itoi ages $[\mathrm{Fe/H}] < -2.0$ 7–14 Gyr, with no discernible internal age-metallicity gradient due to small sample size (Woody et al., 2024). Bayesian SFH modeling gives a mean formation age $[\mathrm{Fe/H}] < -2.0$ 8 Gyr and $[\mathrm{Fe/H}] < -2.0$ 9 of stars in place by $L_z$ 0 Gyr, confirming I’itoi as among the first minor mergers. The chemically inferred quenching redshift exceeds $L_z$ 1, consistent with ultra-short star forming timescales and truncation upon accretion into the proto–Milky Way potential (Naidu et al., 2022, Woody et al., 2024).

I’itoi’s total stellar mass ( $L_z$ 2) contributes only $L_z$ 3 to the inner stellar halo mass within $L_z$ 4 kpc. Its survival as a distinct phase-space overdensity amidst dominant mergers (GSE, Sgr) highlights the sensitivity of modern chemo-dynamical surveys to even minute merger events and places a boundary on the minimum accretion mass detectable in the Galactic halo (2006.08625, Malhan et al., 2022, Sante et al., 12 Mar 2026).