OCCAM Survey: Mapping Open Cluster Chemistry

• OCCAM Survey is a Galactic-archaeology program that uses homogeneous APOGEE and Gaia data from open clusters to trace the Milky Way’s chemical structure and evolution.
• It has evolved from small-scale APOGEE studies to a large, multi-dimensional resource integrating infrared spectroscopy and high-resolution optical follow-ups to refine cluster memberships and abundance measurements.
• Key results include robust radial metallicity gradients, empirical [C/N]-age calibrations, and detailed neutron-capture element analyses that advance our understanding of Galactic chemical evolution.

OCCAM, the Open Cluster Chemical Abundances and Mapping Survey, is a Galactic-archaeology program built around open clusters and designed to produce “a comprehensive, uniform, infrared-based spectroscopic dataset for hundreds of open clusters” and to “constrain key Galactic dynamical and chemical parameters from this sample.” Its scientific premise is that open clusters are age-datable, approximately coeval, and chemically homogeneous stellar populations, so they can serve as unusually clean tracers of the Milky Way disk’s chemical structure, its temporal evolution, and the calibration of age-dating methods such as chemical clocks, gyrochronology, and asteroseismology (Frinchaboy et al., 2013, 2206.13650, Spoo et al., 2022). Over time, OCCAM has expanded from an APOGEE-based infrared abundance survey into a broader platform that combines Gaia membership vetting, cluster-scale chemical mapping, empirical age calibration, and high-resolution optical follow-up of neutron-capture elements (Otto et al., 9 Jul 2025, Myers et al., 14 Oct 2025).

1. Survey scope and historical development

The first OCCAM science paper established the survey’s basic design in SDSS DR10, presenting analysis of 141 member stars in 28 open clusters. That sample included “the first high-resolution metallicity measurements for 22 open clusters,” and it was used to examine the Galactic disk gradients of both $[M/H]$ and $[\alpha/M]$ from a single homogeneous APOGEE dataset (Frinchaboy et al., 2013). The second contribution, based on SDSS/DR14, moved from an early proof-of-concept to “precision cluster abundances,” with 259 member stars in 19 open clusters, and it extended the survey from bulk metallicity to multi-element radial trends in $$[{\rm O/Fe}], [{\rm Mg/Fe}], [{\rm Si/Fe}], [{\rm Mn/Fe}],$$ and $[{\rm Ni/Fe}]$ (Donor et al., 2018).

A major expansion arrived with the DR16 analysis. OCCAM IV assembled a raw sample of 128 open clusters and 914 member stars, of which 83 clusters passed the CMD-based “high quality” screen and 71 high-quality clusters remained in the main abundance-gradient analysis after excluding 12 very young clusters. This release was important because it converted OCCAM from a moderate-size APOGEE cluster project into a large, homogeneous Galactic disk resource with gradients for 16 elements and age-binned trend analysis (Donor et al., 2020). OCCAM VI, based on APOGEE DR17, then presented the full APOGEE-2-era abundance-gradient analysis with 150 open clusters and 2061 member stars in the full sample, 94 “high quality” clusters, and 85 clusters in the final abundance-gradient sample (2206.13650).

The survey continued into the SDSS-V/MWM DR19 era in OCCAM VIII. That release established a sample of 164 high quality open clusters and 1083 member stars, explicitly emphasizing that the cluster set had become large enough not only to refine radial abundance gradients but also, “for the first time using the OCCAM sample,” to investigate Galactic azimuthal variations (Otto et al., 9 Jul 2025). In parallel, OCCAM broadened scientifically beyond the original APOGEE-accessible light, alpha, odd-$Z$, and iron-peak species. The cerium study derived homogeneous Ce abundances for 218 stars in 42 open clusters (Sales-Silva et al., 2021), OCCAM VII produced an empirical $[{\rm C/N}]$-age calibration (Spoo et al., 2022), and later optical follow-up papers used OCCAM as the membership and target-selection backbone for neutron-capture work inaccessible or unreliable in APOGEE (Myers et al., 14 Oct 2025, Myers et al., 1 Jul 2026).

2. Membership architecture and sample construction

OCCAM’s membership methodology evolved substantially across releases. In the earliest DR10 analysis, non-calibration clusters were pre-selected with an infrared strategy built around spatial location, color–magnitude behavior, and especially extinction filtering via the Rayleigh-Jeans Color Excess (RJCE) method. Candidate members then had to satisfy APOGEE photometric cuts, primarily

$(J-K_s)_0 \ge 0.5$

and

$7 \le H \le 12.2$

for a standard 3-hour field, with targeting possible down to $H=13.8$ in deeper fields. Final membership relied mainly on radial velocity, with stars having radial-velocity membership probabilities $>50\%$ accepted, followed by an iterative $[\alpha/M]$0 metallicity cut (Frinchaboy et al., 2013).

By DR14, Gaia astrometry had become central. OCCAM II combined APOGEE radial velocities and metallicities with Gaia DR2 proper motions, and a star was considered a likely member if it lay within $[\alpha/M]$1 of the cluster mean in RV, [Fe/H], and proper-motion space. The abundance analysis was then restricted to likely giant members, approximately

$[\alpha/M]$2

and a cluster entered the “high reliability” sample only if it had at least 4 likely member giant stars (Donor et al., 2018).

The DR16 and DR17 analyses preserved this multi-dimensional logic while intensifying quality control. OCCAM IV combined APOGEE spectroscopy with Gaia DR2 astrometry and photometry, started from stars within $[\alpha/M]$3 the cluster radius, and required $[\alpha/M]$4 consistency in RV, $[\alpha/M]$5, and proper motion. It then applied a visual screen based on proper-motion-cleaned color–magnitude diagrams, retaining only clusters in which the APOGEE-selected members formed a plausible main sequence, subgiant branch, giant branch, or red clump (Donor et al., 2020). OCCAM VI updated this structure with Gaia EDR3-era information, starting from about 26,700 stars near known open clusters and retaining stars whose reported probabilities were $[\alpha/M]$6 in RV, [Fe/H], and PM space within the $[\alpha/M]$7 acceptance region (2206.13650).

The DR19 membership pipeline introduced a slightly different balance between Gaia and spectroscopy. For each cluster, OCCAM VIII selected stars within

$[\alpha/M]$8

and cross-matched them to Cantat-Gaudin et al. (2020) members with membership probability $[\alpha/M]$9. It then refined membership in RV and [Fe/H] space using Gaussian-kernel smoothing and retained stars with probability $$[{\rm O/Fe}], [{\rm Mg/Fe}], [{\rm Si/Fe}], [{\rm Mn/Fe}],$$0 in both, which the paper describes as a $$[{\rm O/Fe}], [{\rm Mg/Fe}], [{\rm Si/Fe}], [{\rm Mn/Fe}],$$1 cut for deriving bulk cluster parameters. Visual quality classes based on Gaia CMDs, Kiel diagrams, and PARSEC isochrones then separated calibration clusters, high-quality 5+ member clusters, high-quality 2–4 member clusters, and good one-star clusters from rejected systems (Otto et al., 9 Jul 2025). This suggests a steady methodological trajectory from infrared field cleaning toward a fully multi-dimensional Gaia–APOGEE membership architecture.

3. Spectroscopic framework and abundance products

The original OCCAM survey is fundamentally an APOGEE survey. APOGEE provides high-resolution near-infrared spectroscopy in the $$[{\rm O/Fe}], [{\rm Mg/Fe}], [{\rm Si/Fe}], [{\rm Mn/Fe}],$$2-band at $$[{\rm O/Fe}], [{\rm Mg/Fe}], [{\rm Si/Fe}], [{\rm Mn/Fe}],$$3, which is especially valuable for open clusters in the dust-obscured Galactic plane because the infrared reduces extinction and allows observation of distant or reddened giant stars (Frinchaboy et al., 2013). In OCCAM II and later papers, cluster abundances were not generally rederived from raw spectra by the OCCAM authors; instead, the survey relied on ASPCAP pipeline products, then converted member-star abundances into cluster means, using the observed scatter among vetted members rather than propagating individual-star uncertainties (Donor et al., 2018).

OCCAM IV standardized the use of DR16 “named tag” abundances and explicitly analyzed 16 chemical species: Fe plus O, Na, Mg, Al, Si, S, K, Ca, Ti, V, Cr, Mn, Co, Ni, Cu. Cluster abundances were produced by determining reliable members, adopting APOGEE DR16 calibrated abundances for those stars, and computing cluster-level means and uncertainties from the member set (Donor et al., 2020). OCCAM VI then updated this framework to DR17 and analyzed “16 reliable chemical species available in APOGEE DR17”: Fe, the $$[{\rm O/Fe}], [{\rm Mg/Fe}], [{\rm Si/Fe}], [{\rm Mn/Fe}],$$4-elements O, Mg, Si, S, Ca, Ti, the iron-peak elements V, Cr, Mn, Co, Ni, the odd-$$[{\rm O/Fe}], [{\rm Mg/Fe}], [{\rm Si/Fe}], [{\rm Mn/Fe}],$$5 species Na, Al, K, and the neutron-capture element Ce (2206.13650).

The DR19 analysis extended the catalog structure further. OCCAM VIII used SDSS-V / Milky Way Mapper DR19 abundances from ASPCAP within the astra framework and investigated O, Mg, Si, S, Ca, Ti, Cr, Mn, Fe, Co, Ni, Na, Al, K, Ce, Nd. It also released two value-added catalogs, occam_member-DR19.fits and occam_cluster-DR19.fits, containing member-level probabilities, bulk cluster chemistry, motions, and orbital parameters including $$[{\rm O/Fe}], [{\rm Mg/Fe}], [{\rm Si/Fe}], [{\rm Mn/Fe}],$$6 and $$[{\rm O/Fe}], [{\rm Mg/Fe}], [{\rm Si/Fe}], [{\rm Mn/Fe}],$$7 (Otto et al., 9 Jul 2025). Across releases, the survey repeatedly emphasized that its principal advantage lies in internal homogeneity rather than the mere size of the sample.

4. Galactic abundance gradients and disk structure

OCCAM’s most widely cited results concern the Milky Way’s radial abundance gradients. The first DR10 paper already found that a single global linear metallicity gradient was an incomplete description. Over

$$[{\rm O/Fe}], [{\rm Mg/Fe}], [{\rm Si/Fe}], [{\rm Mn/Fe}],$$8

a single-line fit yielded

$$[{\rm O/Fe}], [{\rm Mg/Fe}], [{\rm Si/Fe}], [{\rm Mn/Fe}],$$9

but splitting at

$[{\rm Ni/Fe}]$0

produced

$[{\rm Ni/Fe}]$1

for $[{\rm Ni/Fe}]$2 and

$[{\rm Ni/Fe}]$3

for $[{\rm Ni/Fe}]$4, while $[{\rm Ni/Fe}]$5 showed

$[{\rm Ni/Fe}]$6

(Frinchaboy et al., 2013).

OCCAM II remeasured the metallicity gradient with Gaia-refined membership and found the preferred

$[{\rm Ni/Fe}]$7

after excluding NGC 6791, with significant positive gradients in $[{\rm Ni/Fe}]$8 and significant negative gradients in $[{\rm Ni/Fe}]$9 and $Z$0 (Donor et al., 2018). OCCAM IV then used a larger DR16 sample and fit the break radius as a free parameter for the first time, obtaining an inner-disk metallicity gradient of

$Z$1

over $Z$2 kpc, an outer gradient of

$Z$3

and a fitted break at

$Z$4

The same paper also found that older open-cluster populations show steeper $Z$5 gradients, while younger populations are flatter (Donor et al., 2020).

OCCAM VI, using the final APOGEE DR17 sample, produced the survey’s definitive APOGEE-2 metallicity-gradient statement. For present-day Galactocentric radius, it found an inner-disk slope

$Z$6

over

$Z$7

with a knee at

$Z$8

and an outer slope

$Z$9

Using guiding-center radius,

$[{\rm C/N}]$0

it obtained a similar inner slope,

$[{\rm C/N}]$1

with a knee at

$[{\rm C/N}]$2

and outer slope

$[{\rm C/N}]$3

The same DR17 paper reported significant $[{\rm C/N}]$4 gradients for O, Mg, S, Ca, Mn, Na, Al, K, and Ce (2206.13650).

OCCAM VIII revised the gradient picture in DR19. For the 164 high-quality open clusters, it found an overall linear radial metallicity gradient

$[{\rm C/N}]$5

and

$[{\rm C/N}]$6

Although bilinear fits could still be obtained, the paper concluded from an Akaike Information Criterion comparison that the DR19 metallicity gradient was better described by a single linear trend than by a broken-linear model. For the first time in the OCCAM series, the sample size also allowed an azimuthal analysis, and the paper reported evidence of azimuthal variations in the measured radial abundance gradient in the Galactic disk (Otto et al., 9 Jul 2025). A common misconception is therefore that OCCAM has converged on one immutable “true” break radius; the series instead shows that the inferred shape depends on release, radial coverage, coordinate choice, and model-comparison criterion.

5. Chemical clocks and neutron-capture extensions

OCCAM’s role in chemical-clock calibration is most explicit in OCCAM VII. Using APOGEE DR17 giant-star abundances in a final calibration sample of 49 clusters and 530 stars, the survey derived the empirical relation

$[{\rm C/N}]$7

usable for

$[{\rm C/N}]$8

and described as applicable “primarily to metal-rich, thin and thick disk giant stars.” The paper also found no evidence for a significant difference between RGB-only and RC-only relations within the current uncertainties (Spoo et al., 2022). This gave OCCAM a direct role in field-star age inference rather than only cluster abundance cartography.

The survey’s neutron-capture extension began inside APOGEE space with the cerium study. Using BACCHUS, MARCS model atmospheres, and Turbospectrum, that paper derived Ce abundances from seven Ce II lines in 218 stars belonging to 42 open clusters. Its central results were that, for Ages $[{\rm C/N}]$9 Gyr, younger open clusters have higher $(J-K_s)_0 \ge 0.5$0 and $(J-K_s)_0 \ge 0.5$1 than older clusters, that metallicity segregates clusters in the $(J-K_s)_0 \ge 0.5$2-Age plane, and that these relations are therefore not universal clocks. It also derived, for the first time, age-binned radial gradients of $(J-K_s)_0 \ge 0.5$3 and $(J-K_s)_0 \ge 0.5$4, with

$(J-K_s)_0 \ge 0.5$5

and

$(J-K_s)_0 \ge 0.5$6

and estimated that the $(J-K_s)_0 \ge 0.5$7 gradient becomes slightly steeper, changing by $(J-K_s)_0 \ge 0.5$8 (Sales-Silva et al., 2021).

The next step moved beyond APOGEE entirely. The conference proceeding “Beyond OCCAM: Measuring Optical Neutron Capture Abundances of Open Cluster Stars” explicitly states that it “utilize[s] the SDSS-IV/APOGEE-based OCCAM survey as the foundation for our optical follow-up observations.” It obtained Keck I/HIRES spectra at

$(J-K_s)_0 \ge 0.5$9

with signal-to-noise ratio at least 75 at 5500 Å for targets with

$7 \le H \le 12.2$0

The observed sample comprised 29 stars across 8 clusters, while the preliminary abundance analysis used 16 stars and required at least three high-quality members per cluster, primarily from OCCAM and supplemented by Cantat-Gaudin et al. (2020) when needed. Abundances were derived with BACCHUS, Turbospectrum, and MARCS model atmospheres, starting from APOGEE stellar parameters, and the preliminary neutron-capture results focused on Ce, Ba, Y, Mo, La, and Zr. The main reported pattern was “a non-linear trend of increasing abundance with cluster metallicity $7 \le H \le 12.2$1,” visible across all measured species but “shallower for molybdenum and zirconium” (Myers et al., 14 Oct 2025).

OCCAM X then formalized the optical neutron-capture program on a larger scale. It used OCCAM as the foundation for new optical observations of 56 stars in 18 open clusters, combining Keck/HIRES and Magellan/MIKE spectroscopy at $7 \le H \le 12.2$2 and high-S/N ($7 \le H \le 12.2$3 at 5500Å). With BACCHUS, it derived abundances for 23 elements, including 7 neutron capture abundances not measurable by APOGEE, and characterized their radial distribution in the Milky Way. The paper reports that the second-peak $7 \le H \le 12.2$4-process and $7 \le H \le 12.2$5-process abundances exhibit relatively flat gradients, while the first-peak $7 \le H \le 12.2$6-process abundances also have slopes which are shallower than the alpha and iron-peak elements. It further states that “a metallicity dependence of the AGB stars responsible for producing the heaviest s-process abundances may be necessary to consider in Galactic evolution models” (Myers et al., 1 Jul 2026). In this phase, OCCAM functions both as a membership catalog and target-selection engine and as a contextual abundance/age database for high-resolution optical work.

6. Survey significance, boundaries, and related distinctions

OCCAM’s long-term significance lies in the fact that it turned open clusters into a reusable, internally consistent framework for Milky Way chemo-dynamical inference. By combining homogeneous APOGEE abundances, Gaia astrometry, cluster ages, and later optical follow-up, it created a bridge between classical open-cluster gradient work and survey-scale Galactic archaeology. The survey’s scientific outputs span radial metallicity structure, $7 \le H \le 12.2$7 gradients, age-binned trend analysis, empirical chemical-clock calibration, and neutron-capture abundances across multiple nucleosynthetic channels (Donor et al., 2020, 2206.13650, Spoo et al., 2022).

At the same time, the series is explicit about its limitations. OCCAM IV noted that the measured $7 \le H \le 12.2$8 slope can vary by as much as $7 \le H \le 12.2$9 with the adopted distance catalog, that the outer disk remained sparsely sampled, and that pipeline unreliability for very young stars required excluding clusters younger than 50 Myr from abundance analysis (Donor et al., 2020). OCCAM VI emphasized age uncertainties, possible abundance systematics for elements such as V, Ti, S, Co, and Ce, selection effects in cluster survival, and the unresolved discrepancy between DR17 ASPCAP cerium and BACCHUS-based or optical cerium abundances (2206.13650). OCCAM VIII further showed that the inferred DR19 metallicity gradients depend on the membership catalog and that several element-by-element trends—especially O, Ca, Ti, Co, Ni, Ce, and Nd—can deviate considerably from other large spectroscopic surveys (Otto et al., 9 Jul 2025). This suggests that OCCAM’s chief strength is not immunity to systematics, but the explicit attempt to control them within a uniform cluster framework.

A recurrent point of confusion in the literature concerns the distinction between OCCAM and OCCASO. They are not the same survey. OCCASO is the Open Cluster Chemical Abundances from Spanish Observatories survey, a distinct northern-hemisphere optical program using FIES, HERMES, and CAFE, designed to obtain homogeneous radial velocities, physical parameters, and chemical abundances for Milky Way open clusters (Casamiquela et al., 2016). OCCAM, by contrast, is the APOGEE- and Gaia-centered Open Cluster Chemical Abundances and Mapping Survey that later expanded into optical neutron-capture follow-up.

In cumulative terms, the survey’s trajectory is clear. Early OCCAM established that homogeneous cluster spectroscopy could sharpen the Milky Way’s metallicity-gradient problem. Mid-series OCCAM converted that insight into a large APOGEE resource for multi-element Galactic chemical evolution. Later OCCAM papers turned the survey into a calibration framework for $H=13.8$0-based ages and a platform for multi-wavelength neutron-capture studies. A plausible implication is that OCCAM is no longer only an infrared cluster-abundance survey; it is a general cluster-based infrastructure for connecting ages, kinematics, and chemistry across multiple spectroscopic regimes.