Terzan 2: Inner Bulge Globular Cluster

• Terzan 2 is a highly obscured globular cluster in the inner Milky Way bulge, characterized by an old age, moderate metallicity, and severe dust extinction.
• High-resolution NIR spectroscopy reveals clear multiple population signatures, with uniform iron levels and distinct C-N anti-correlation indicating standard globular cluster chemistry.
• Dynamic studies show Terzan 2’s chaotic, high-eccentricity orbit influenced by the Galactic bar, with frequent high-speed encounters enhancing its mass loss and morphological evolution.

Terzan 2 (Ter 2) is a heavily reddened globular cluster in the inner Milky Way bulge, located at Galactic coordinates $l = 356.3^\circ,\ b = 2.30^\circ$. Contemporary work characterizes it as an old bulge globular cluster with a mean metallicity of $[\mathrm{Fe/H}] = -0.84 \pm 0.04$, no evidence for an intrinsic iron spread, clear evidence for multiple populations through C and N variations, and a chaotic inner-bulge orbit influenced by the Galactic bar but not trapped by it. Direct $N$-body modelling further places Terzan 2 in a dynamically crowded central environment, where repeated high-speed encounters with Terzan 4 and Terzan 5 can enhance mass loss and drive transient triaxial or prolate distortions (Uribe et al., 2 Sep 2025, Ishchenko et al., 8 May 2026).

1. Classification, location, and basic parameters

Terzan 2 is observed deep in the inner Galaxy and is treated in current work as a bona fide bulge globular cluster rather than a halo intruder. At $b = 2.30^\circ$ and distances of order $8$–$10$ kpc, it lies only a few hundred parsecs from the Galactic plane. The cluster is strongly obscured by dust, with a new extinction analysis giving $R_V = 2.48 \pm 0.10$, $A_V = 5.16 \pm 0.05$, and a corresponding foreground reddening $E(B-V) = 2.08$; the Harris value quoted for comparison is $E(B-V) = 1.87$. An age of $[\mathrm{Fe/H}] = -0.84 \pm 0.04$0 Gyr is adopted for isochrone fitting, consistent with an old bulge globular cluster (Uribe et al., 2 Sep 2025).

Quantity	Value	Context
Galactic coordinates	$[\mathrm{Fe/H}] = -0.84 \pm 0.04$1	inner bulge
Mean metallicity	$[\mathrm{Fe/H}] = -0.84 \pm 0.04$2	four BACCHUS members
Extinction law	$[\mathrm{Fe/H}] = -0.84 \pm 0.04$3	fitted for Terzan 2
Visual extinction	$[\mathrm{Fe/H}] = -0.84 \pm 0.04$4	severe obscuration
Foreground reddening	$[\mathrm{Fe/H}] = -0.84 \pm 0.04$5	fitted value
Photometric distance	$[\mathrm{Fe/H}] = -0.84 \pm 0.04$6	from isochrone fitting
Orbit-integration distance	$[\mathrm{Fe/H}] = -0.84 \pm 0.04$7	adopted for dynamics

Two distance scales are discussed. Isochrone fitting gives a true distance modulus $[\mathrm{Fe/H}] = -0.84 \pm 0.04$8, corresponding to $[\mathrm{Fe/H}] = -0.84 \pm 0.04$9, whereas the orbital analysis adopts $N$0 from Baumgardt & Vasiliev for consistency with previous dynamical work. The same source is quoted as giving $N$1 kpc and a mass of $N$2. A plausible implication is that basic structural inference for Terzan 2 remains somewhat sensitive to the adopted distance scale, even though its inner-bulge classification is robust.

2. Observational basis and abundance methodology

The first detailed high-resolution spectroscopic abundance analysis of Terzan 2 was carried out within CAPOS, the bulge Cluster APOgee Survey, using APOGEE-2S near-infrared spectra and BACCHUS line-by-line abundance analysis (Uribe et al., 2 Sep 2025). Near-infrared spectroscopy is central to the modern study of Terzan 2 because the cluster is so heavily extincted that optical spectroscopy is difficult; APOGEE-2S provides $N$3 spectroscopy across most of the $N$4 band, $N$5, with small gaps near $N$6–$N$7 and $N$8–$N$9.

Membership determination combines positional, kinematic, radial-velocity, metallicity, and color-magnitude information. Candidate stars were selected on the red giant branch and bright enough to achieve adequate APOGEE signal-to-noise, and refined membership later used Gaia proper motions. Four stars survive the full set of cuts: they lie within the tidal radius, cluster in proper-motion space around $b = 2.30^\circ$0 and $b = 2.30^\circ$1, have very similar radial velocities around $b = 2.30^\circ$2, cluster near $b = 2.30^\circ$3, and lie cleanly on the differentially de-reddened red giant branch.

The atmospheric-parameter pipeline is explicitly photometric rather than purely spectroscopic. A Gaia+2MASS color-magnitude diagram is first corrected for strong differential reddening. PARSEC isochrones with $b = 2.30^\circ$4, fixed age $b = 2.30^\circ$5 Gyr, and initial global metallicity $b = 2.30^\circ$6 are then used to optimize $b = 2.30^\circ$7, $b = 2.30^\circ$8, distance, and $b = 2.30^\circ$9. For each target, $8$0 and $8$1 are obtained by projecting the star horizontally onto the best-fit red giant branch isochrone at the same $8$2 magnitude, with a temperature-dependent extinction correction applied. Microturbulence is then assigned from an empirical relation for FGK stars.

BACCHUS computes synthetic spectra with Turbospectrum and 1D LTE MARCS atmospheres and applies four line diagnostics, with the final adopted abundance taken from the $8$3 profile fit. Carbon, nitrogen, and oxygen are derived iteratively and self-consistently from CO, OH, and CN molecular features. The analysis reports reliable abundances for 11 elements. Sodium is not measured because APOGEE $8$4-band Na lines are weak and blended, so the classical O-Na anti-correlation is not tested in this data set.

3. Chemical composition and multiple populations

Terzan 2 is chemically homogeneous in iron but not in the light elements that trace multiple populations. The BACCHUS analysis gives $8$5 with dispersion $8$6, consistent with a total iron error of $8$7 dex, so no intrinsic $8$8 spread is detected. Nickel is essentially solar relative to iron, with $8$9 and negligible internal scatter. This is the abundance pattern expected of a normal globular cluster rather than an iron-complex system (Uribe et al., 2 Sep 2025).

The $10$0-elements are uniformly enhanced: $10$1, $10$2, $10$3, $10$4, and $10$5. No intrinsic spreads are detected in these species. The reported interpretation is rapid early enrichment by core-collapse supernovae, consistent with the general trend of Galactic globular clusters and bulge field stars at similar metallicity.

The clearest multiple-population signal appears in carbon and nitrogen. The mean abundances are $10$6 and $10$7, with a large $10$8 spread of about $10$9 dex, well above the quoted typical error of about $R_V = 2.48 \pm 0.10$0 dex. The stars define a C-N anti-correlation and a modest N-O anti-correlation. One star, 2M17273540−3047308, is described as having relatively low $R_V = 2.48 \pm 0.10$1, closer to first-generation composition. These patterns are identified as the standard chemical hallmark of multiple populations in globular clusters.

Aluminum is only mildly enhanced, $R_V = 2.48 \pm 0.10$2, and no statistically secure internal Al spread is claimed. The Mg-Al plane shows only a hint of a relation, treated as tentative because the sample contains only four stars. Cerium is mildly super-solar, $R_V = 2.48 \pm 0.10$3, with very small internal scatter and no Ce-N correlation. The absence of a Ce-N relation differentiates Terzan 2 from some other bulge clusters in which Ce-N or Ce-Al correlations have been reported. This suggests that Terzan 2 follows the “normal” Galactic chemical evolution of the bulge in Ce rather than showing a cluster-specific internal $R_V = 2.48 \pm 0.10$4-process signature.

4. Orbit in the Galactic bar region

Terzan 2’s present-day orbital characterization is based on Gaia EDR3 proper motions, radial velocity, and a bulge potential that includes a rotating boxy/peanut bar. The adopted dynamical inputs are $R_V = 2.48 \pm 0.10$5, $R_V = 2.48 \pm 0.10$6, $R_V = 2.48 \pm 0.10$7, and $R_V = 2.48 \pm 0.10$8, with orbit integration performed in the GravPot16 Milky Way potential for bar pattern speeds of $R_V = 2.48 \pm 0.10$9, $A_V = 5.16 \pm 0.05$0, and $A_V = 5.16 \pm 0.05$1. The study computes $A_V = 5.16 \pm 0.05$2 million backward orbit realizations over $A_V = 5.16 \pm 0.05$3 Gyr in the non-inertial frame corotating with the bar (Uribe et al., 2 Sep 2025).

Across the tested bar pattern speeds, the orbital morphology changes little. Terzan 2 remains on a highly eccentric orbit with $A_V = 5.16 \pm 0.05$4, apogalactocentric distances $A_V = 5.16 \pm 0.05$5 kpc, and vertical excursions $A_V = 5.16 \pm 0.05$6 kpc. The cluster is therefore dynamically confined to the inner bulge and close to the Galactic plane.

The orbit is described as chaotic, influenced by the bar but not trapped by it. In the terminology of the analysis, Terzan 2 crosses the bar region but does not follow a simple bar-aligned trapped family. Instead, it can display both prograde and retrograde motion within the inner bulge region. This distinction matters: bar influence is strong, but the cluster is not treated as a bar-supporting orbit in the same sense as a trapped $A_V = 5.16 \pm 0.05$7-type trajectory. The chemical and dynamical evidence are therefore jointly interpreted as favoring an in-situ bulge origin.

5. Long-term dynamical evolution and mutual interactions

A separate dynamical study follows Terzan 2, Terzan 4, and Terzan 5 for $A_V = 5.16 \pm 0.05$8 Gyr with high-resolution direct $A_V = 5.16 \pm 0.05$9-body simulations in a time-variable Milky Way-like potential extracted from IllustrisTNG-100. In that framework, Terzan 2 is represented at $E(B-V) = 2.08$0 Gyr by an initial mass $E(B-V) = 2.08$1, $E(B-V) = 2.08$2 simulation particles, an initial half-mass radius $E(B-V) = 2.08$3 pc, a King-model concentration $E(B-V) = 2.08$4, and metallicity parameter $E(B-V) = 2.08$5. These are model initial conditions calibrated to approximate present-day mass and size after long-term evolution, not direct present-day observables (Ishchenko et al., 8 May 2026).

The simulations compare isolated reference runs with a combined three-cluster run that includes full mutual gravity. They compute the instantaneous tidal radius via

$E(B-V) = 2.08$6

and quantify morphology through axis lengths $E(B-V) = 2.08$7 and the triaxiality parameter

$E(B-V) = 2.08$8

In the combined run, Terzan 2 undergoes multiple close encounters with Terzan 4 and Terzan 5. The most extreme Terzan 2–Terzan 4 event occurs at $E(B-V) = 2.08$9 Gyr, when the clusters approach within $E(B-V) = 1.87$0 pc at a relative velocity of $E(B-V) = 1.87$1, with $E(B-V) = 1.87$2 pc and $E(B-V) = 1.87$3 pc; the tidal volumes overlap strongly. A notable Terzan 2–Terzan 5 encounter occurs at $E(B-V) = 1.87$4 Gyr, with $E(B-V) = 1.87$5 pc, $E(B-V) = 1.87$6, $E(B-V) = 1.87$7 pc, and $E(B-V) = 1.87$8 pc.

These flybys are interpreted as impulsive rather than merger-producing because the relative speeds are very high. Even so, the combined simulations show that mutual interactions enhance Terzan 2’s early mass-loss rate relative to the isolated case. Over the full $E(B-V) = 1.87$9 Gyr interval, Terzan 2 loses almost $[\mathrm{Fe/H}] = -0.84 \pm 0.04$00 of its initial mass in both isolated and combined runs, but remains bound at the end. The common run also produces distinct morphological evolution: in isolation, Terzan 2 remains almost perfectly spherical with $[\mathrm{Fe/H}] = -0.84 \pm 0.04$01, whereas in the combined run it evolves at some epochs to a distinctly prolate shape with $[\mathrm{Fe/H}] = -0.84 \pm 0.04$02 and $[\mathrm{Fe/H}] = -0.84 \pm 0.04$03–$[\mathrm{Fe/H}] = -0.84 \pm 0.04$04. The shape oscillations track encounter timing. Extended tidal tails are also formed, with characteristic densities of $[\mathrm{Fe/H}] = -0.84 \pm 0.04$05, implying that any observational detection would be difficult in the dense inner Galaxy.

6. Relation to the broader Terzan family and astrophysical significance

Terzan 2 is often discussed alongside other inner-bulge “Terzan” systems, but it is not equivalent in phenomenology to Terzan 5. Terzan 5 has been described as a stellar system with three sub-populations, iron content varying by more than one order of magnitude, and two distinct main-sequence turn-off ages of $[\mathrm{Fe/H}] = -0.84 \pm 0.04$06 Gyr and $[\mathrm{Fe/H}] = -0.84 \pm 0.04$07 Gyr for its dominant sub-solar and super-solar components. Terzan 2, by contrast, shows no intrinsic $[\mathrm{Fe/H}] = -0.84 \pm 0.04$08 spread in the four-star BACCHUS sample and is characterized as a typical bulge globular cluster with standard multiple-population chemistry rather than as a Terzan 5-like iron-complex bulge building-block remnant (Ferraro et al., 2016, Uribe et al., 2 Sep 2025).

That distinction is astrophysically useful. Terzan 2 combines severe extinction, old age, bulge-like $[\mathrm{Fe/H}] = -0.84 \pm 0.04$09-enhancement, and normal globular-cluster light-element anomalies, making it a compact tracer of early bulge star formation under extreme observational conditions. Its N-rich stars also fit naturally into the broader discussion of N-rich bulge field stars as possible debris from dissolved clusters. The long-term dynamical calculations add a complementary point: the present-day state of inner-bulge clusters need not be explained by Galactic tides alone. A plausible implication is that Terzan 2 exemplifies a class of low-mass bulge clusters whose chemistry is ordinary for globular clusters, but whose structural evolution is strongly conditioned by the collective gravitational environment of the central Milky Way.

In that combined observational and dynamical picture, Terzan 2 is best understood as an old, in-situ bulge globular cluster that is chemically typical, dynamically extreme, and embedded in a bar-dominated and interaction-rich central Galactic ecosystem.