Extended Main Sequence Turn-Off (eMSTO)

• eMSTO is a broadened or split feature in a cluster's CMD, marking deviations from a classical, narrow main sequence turn-off.
• Observations reveal that eMSTO morphology correlates with stellar rotation rates, internal mixing, and in some cases, genuine age spreads.
• Detailed CMD analyses, using metrics like MSTO-to-MS width ratios and pseudo-age distributions, offer insights into cluster formation and stellar evolution.

Searching arXiv for recent and foundational papers on extended main-sequence turn-offs to ground the article in the literature. arxiv.search({"^^^^^^^^3query3 main sequence turnoff\" OR ^^^^^^^^3all:\3 main-sequence turn-off\"","max_results":^^^^3all:\3query3^^^^,"sort_by":"submittedDate","sort_order":"descending"}) arXiv search results retrieved. I’ll use the provided corpus together with relevant arXiv records to ensure the article is grounded in the literature. Extended main-sequence turn-off (eMSTO) denotes a morphology of the color–magnitude diagram (CMD) in which the main-sequence turn-off is broadened or bifurcated far beyond the width expected for a classical simple stellar population, while the underlying population is often otherwise consistent with a single metallicity and a relatively narrow lower main sequence. Over the last decade, eMSTOs have been identified in intermediate-age clusters in the Magellanic Clouds and, with Gaia DR^^^^3 OR all:\3^^^^, in Galactic open clusters, making them a central diagnostic in the interpretation of stellar rotation, internal mixing, binarity, and possible age spreads in cluster populations (&&&^^^^3query3^^^^&&&, &&&^^^^3all:\3^^^^&&&).

^^^^3all:\3^^^^. Definition and observational morphology

In a canonical single-population cluster, the MSTO is narrow because stars of nearly identical age and composition leave the main sequence at nearly the same luminosity and color. An eMSTO appears instead as a broadened or bifurcated “fan” or wedge around the turn-off. In Galactic open clusters aged PRESERVED_PLACEHOLDER_^^^^3query3^^^^ Myr–PRESERVED_PLACEHOLDER_^^^^3all:\3^^^^ Gyr, the turn-off broadening typically spans PRESERVED_PLACEHOLDER_^^^^3 OR all:\3^^^^–$0.15$ mag or $\Delta G \simeq 0.2$–$0.4$ mag; in LMC and SMC systems, some CMDs show broader manifestations, such as $\Delta m_{\rm F475W}\approx 0.6$ mag and $\Delta c_{\rm F475W-F814W}\approx 0.25$ mag in NGC 4^^^^3all:\3all:\3^^^^ (&&&^^^^3query3^^^^&&&, Girardi et al., 2013).

A useful operational quantity is the MSTO width, $\Delta_{\rm MSTO}$, defined as the age difference between non-rotating isochrones that bracket the observed turn-off. In Galactic open clusters near $\sim 900$ Myr, PRESERVED_PLACEHOLDER_^^^^3all:\3query3^^^^–PRESERVED_PLACEHOLDER_^^^^3all:\3all:\3^^^^ Myr is typical, whereas the lower main sequence remains narrow, with PRESERVED_PLACEHOLDER_^^^^3all:\3 OR all:\3^^^^ Myr equivalent width. This localization of the broadening to the turn-off region is one of the basic empirical constraints on any physical explanation (&&&^^^^3query3^^^^&&&).

The morphology is age-dependent. In clusters younger than PRESERVED_PLACEHOLDER_^^^^3all:\33^^^^ Myr, broadened or split upper main sequences are commonly seen in addition to an eMSTO, while older systems more often display only the broadened turn-off. In very young clusters such as PRESERVED_PLACEHOLDER_^^^^3all:\3^^^^4 and PRESERVED_PLACEHOLDER_^^^^3all:\3^^^^5 Persei, stars below PRESERVED_PLACEHOLDER_^^^^3all:\3^^^^6 occupy narrow main sequences, while more massive stars define the eMSTO, again indicating that the phenomenon is not a uniform broadening of the full CMD (Cordoni et al., 2018, Li et al., 2019).

^^^^3 OR all:\3^^^^. Discovery across environments and demographic trends

The eMSTO was first established observationally in Magellanic Cloud clusters. In the LMC, intermediate-age clusters with PRESERVED_PLACEHOLDER_^^^^3all:\3^^^^7 and PRESERVED_PLACEHOLDER_^^^^3all:\3^^^^8–PRESERVED_PLACEHOLDER_^^^^3all:\3 were found to exhibit bifurcated or extended MSTOs in a large fraction of cases; Milone et al. (^^^^3 OR all:\3query3query3^^^^9), as summarized in the literature, examined ^^^^3all:\3^^^^6 clusters and found ^^^^3all:\3all:\3^^^^ with eMSTOs, with inferred age spreads up to PRESERVED_PLACEHOLDER_^^^^3 OR all:\3query3^^^^ Myr (&&&^^^^3all:\3^^^^&&&). In the SMC, NGC 4^^^^3all:\3all:\3^^^^ and NGC 4^^^^3all:\3^^^^9 showed strikingly similar eMSTO morphologies despite present masses differing by a factor of 4, reinforcing the view that the phenomenon is not confined to a narrow mass interval (Girardi et al., 2013).

Subsequent Gaia-based work established that eMSTOs are not peculiar to the Magellanic Clouds. NGC 67^^^^3query3^^^^5 (M^^^^3all:\3all:\3^^^^) provided the first clear Galactic open-cluster case, followed by NGC ^^^^3 OR all:\3query3^^^^99, NGC ^^^^3 OR all:\33submittedDate3query3^^^^, and NGC ^^^^3 OR all:\3descending3all:\3^^^^8; a broader Gaia DR^^^^3 OR all:\3^^^^ analysis then concluded that twelve Milky Way open clusters younger than PRESERVED_PLACEHOLDER_^^^^3 OR all:\3all:\3^^^^ Gyr show eMSTOs and/or broadened main sequences, and that nearly all clusters younger than PRESERVED_PLACEHOLDER_^^^^3 OR all:\3 OR all:\3^^^^ Gyr in both the Magellanic Clouds and the Milky Way exhibit the phenomenon (Marino et al., 2018, Cordoni et al., 2018).

The strongest large-sample trend is with age. For PRESERVED_PLACEHOLDER_^^^^3 OR all:\33^^^^ Myr PRESERVED_PLACEHOLDER_^^^^3 OR all:\3^^^^4 Gyr, one empirical fit gives PRESERVED_PLACEHOLDER_^^^^3 OR all:\3^^^^5, while another Gaia-based open-cluster study finds PRESERVED_PLACEHOLDER_^^^^3 OR all:\3^^^^6 when PRESERVED_PLACEHOLDER_^^^^3 OR all:\3^^^^7 is in Myr (&&&^^^^3all:\3all:\3^^^^&&&, Cordoni et al., 2018). In larger LMC samples extending to older ages, the width–age relation becomes non-monotonic, with a pronounced inverted-PRESERVED_PLACEHOLDER_^^^^3 OR all:\3^^^^8 or “PRESERVED_PLACEHOLDER_^^^^3 OR all:\3^^^^9” shape: MSTO widths are small in younger systems, peak near $0.15$^^^^3query3^^^^ ($0.15$^^^^3all:\3^^^^ Gyr), and decline again at older ages (&&&^^^^3all:\33^^^^&&&). This age dependence is central because it suggests that the dominant broadening mechanism is controlled by stellar evolution timescales rather than solely by cluster structure.

3. Quantification in CMD space

Modern eMSTO analyses are heavily CMD-based and, in Galactic work, depend on precise Gaia membership selection. One representative Gaia DR^^^^3 OR all:\3^^^^ workflow extracted all sources within $0.15$^^^^3 OR all:\3^^^^ arcmin of each cluster, selected members using proper-motion cuts $0.15$3, parallax selection within $0.15$4 of the parallax peak, and a 3D membership-probability filter following Kamann et al. (^^^^3 OR all:\3query3all:\3^^^^4). Distances were obtained from the median parallax of members with $0.15$5 mag, and extinctions were estimated by visually fitting non-rotating MIST isochrones to the unevolved main sequence (&&&^^^^3query3^^^^&&&).

One quantitative diagnostic compares the turn-off width to the lower-main-sequence width. Two fixed reference lines are defined in the CMD, one through the MSTO region and one along the lower main sequence. For each member, the perpendicular distance to each line is computed, and the ratio of the standard deviations of those distance distributions, $0.15$6, provides a width measure that is relatively insensitive to zero-point uncertainties and small age offsets. In well-behaved eMSTO clusters, $0.15$7–$0.15$8, whereas a narrow turn-off has $0.15$9 (&&&^^^^3query3^^^^&&&).

A second widely used formalism maps CMD position to pseudo-age. In this method, a parallelogram is drawn around the MSTO, with the long axis nearly perpendicular to non-rotating isochrones. Stars are projected onto that axis, converted to pseudo-ages via a calibrated relation, and the resulting distribution is characterized by a FWHM. The intrinsic width is then estimated after correcting for broadening caused by photometric scatter and unresolved binaries: $\Delta G \simeq 0.2$^^^^3query3^^^^ This framework underlies several of the large-sample Magellanic Cloud studies and makes it possible to compare MSTO widths in “age units” across clusters and photometric systems (&&&^^^^3all:\3^^^^6&&&, &&&^^^^3all:\3^^^^7&&&).

The same logic has recently been extended to low-mass clusters. In the VISCACHA analysis, the MSTO width was measured after field decontamination, differential-reddening correction, and comparison to synthetic SSP CMDs, and the resulting widths agreed between VISCACHA and HST photometry for NGC ^^^^3all:\3sort_by3 OR all:\3^^^^ to within $\Delta G \simeq 0.2$^^^^3all:\3^^^^, suggesting that the age-spread-style width measurement is robust across photometric systems (&&&^^^^3all:\3^^^^8&&&).

4. Age spreads, cluster potential, and the extended-star-formation interpretation

The earliest dominant interpretation treated eMSTOs as evidence for prolonged star formation. In that picture, the observed turn-off width reflects age spreads of order $\Delta G \simeq 0.2$^^^^3 OR all:\3^^^^–ΔG≃0.2\Delta G \simeq 0.23 Myr in intermediate-age LMC clusters, and the same extended star-formation histories were linked to the light-element abundance anomalies characteristic of ancient globular clusters. One proposed sequence is that a first generation forms, core-collapse SNe II expel natal gas, intermediate-mass AGB stars eject processed material over the next $\Delta G \simeq 0.2$4–$\Delta G \simeq 0.2$5 Myr, and a second generation then forms from retained ejecta mixed with accreted pristine gas (&&&^^^^3all:\3^^^^&&&).

A more dynamical version of this argument emphasized cluster escape velocity. In a mass-limited sample of ^^^^3all:\3^^^^8 intermediate-age Magellanic Cloud clusters, intrinsic MSTO widths of $\Delta G \simeq 0.2$6–$\Delta G \simeq 0.2$7 Myr were found to correlate with present and early central escape velocities, with eMSTO clusters satisfying $\Delta G \simeq 0.2$8 at early times, whereas two lower-mass clusters without eMSTOs remained at $\Delta G \simeq 0.2$9. That threshold was argued to be consistent with AGB-wind speeds, favoring in situ extended star formation regulated by cluster potential depth (&&&^^^^3all:\3^^^^6&&&).

This interpretation remains part of the literature, but it is not uncontested. A larger homogeneous LMC analysis found that eMSTO clusters span $0.4$^^^^3query3^^^^, $0.4$^^^^3all:\3^^^^, and $0.4$^^^^3 OR all:\3^^^^, with no statistically significant correlation between $0.4$3 and core radius, mass, or dynamical age. That study instead concluded that the MSTO width correlates most strongly with cluster age and argued that the phenomenon is not caused by actual age spreads (&&&^^^^3all:\33^^^^&&&).

A hybrid position also exists. Monte Carlo experiments with Geneva SYCLIST models showed that rotation-driven pseudo-age widths scale roughly with cluster age, but for $0.4$4 Gyr clusters such as NGC ^^^^3all:\3^^^^987 and NGC ^^^^3 OR all:\3 OR all:\3^^^^49, rotation-only models with realistic rotation distributions and random inclinations account for only $0.4$5 and $0.4$6 of the observed FWHM widths, respectively. In that analysis, the best fits required a combination of a rotation spread and a genuine age spread of $0.4$7–$0.4$8 yr (&&&^^^^3all:\3^^^^7&&&). The controversy is therefore not over whether CMD broadening exists, but over how much of it can be assigned to stellar rotation alone.

5. Stellar rotation as a unifying framework

The rotation paradigm attributes eMSTOs to two coupled effects. First, rotational mixing extends the main-sequence lifetime by transporting fresh H into the core. Second, gravity darkening alters the surface temperature distribution, with

$0.4$9

and $\Delta m_{\rm F475W}\approx 0.6$^^^^3query3^^^^ for radiative envelopes in the formulations summarized in the literature. In MIST models the relevant control parameter is

$\Delta m_{\rm F475W}\approx 0.6$^^^^3all:\3^^^^

the equatorial angular velocity relative to the ZAMS critical value. At fixed age, increasing $\Delta m_{\rm F475W}\approx 0.6$^^^^3 OR all:\3^^^^ shifts isochrones to redder colors and higher luminosities around the MSTO, so a population with a spread in $\Delta m_{\rm F475W}\approx 0.6$3 naturally develops an eMSTO (&&&^^^^3query3^^^^&&&).

Direct spectroscopic support for this interpretation is strong in several Galactic open clusters. In M^^^^3all:\3all:\3^^^^, high-resolution VLT/FLAMES spectroscopy showed that the blue main sequence is populated by slow rotators and the red main sequence by fast rotators, with $\Delta m_{\rm F475W}\approx 0.6$4 and $\Delta m_{\rm F475W}\approx 0.6$5 (Marino et al., 2018). In NGC ^^^^3 OR all:\3descending3all:\3^^^^8, stars on the red side of the eMSTO have $\Delta m_{\rm F475W}\approx 0.6$6 while stars on the blue side have $\Delta m_{\rm F475W}\approx 0.6$7, and the cluster follows the empirical age–eMSTO-extent relation previously established in Magellanic Cloud work (&&&^^^^3all:\3all:\3^^^^&&&). Comparable red-fast / blue-slow segregation is reported for NGC 58^^^^3 OR all:\3 OR all:\3^^^^, NGC ^^^^3 OR all:\33^^^^55, and NGC 6^^^^3query3^^^^67, where mean projected rotation velocities or direct correlations between $\Delta m_{\rm F475W}\approx 0.6$8 and pseudo-color again place faster rotators on the red side of the MSTO (&&&^^^^3 OR all:\3^^^^6&&&, &&&^^^^3 OR all:\3^^^^7&&&, &&&^^^^3 OR all:\3^^^^8&&&).

Within this framework, the common inference is that eMSTOs can mimic apparent age spreads even in coeval populations. The importance of this point is methodological as well as physical: if rotation is ignored, CMD fitting tends to recover spurious extended star-formation histories of order $\Delta m_{\rm F475W}\approx 0.6$9–$\Delta c_{\rm F475W-F814W}\approx 0.25$^^^^3query3^^^^ Myr in systems whose stars may be nearly coeval (Cordoni et al., 2018). This does not by itself eliminate every age-spread interpretation, but it does establish stellar rotation as an unavoidable ingredient in any realistic treatment of turn-off morphology.

6. Limits of the rotation picture and additional mechanisms

The strongest single-cluster argument for a rotation-dominated interpretation is the exceptional case of NGC ^^^^3 OR all:\3sort_by3query3^^^^9. Among Galactic open clusters around $\Delta c_{\rm F475W-F814W}\approx 0.25$^^^^3all:\3^^^^ Myr, it is unusual in showing an exceptionally narrow CMD, with an estimated age of $\Delta c_{\rm F475W-F814W}\approx 0.25$^^^^3 OR all:\3^^^^ Myr, $\Delta c_{\rm F475W-F814W}\approx 0.25$3 Myr from non-rotating fits, and $\Delta c_{\rm F475W-F814W}\approx 0.25$4. Differential extinction, stellar density, and binaries were explicitly tested and ruled out as unique explanations; rotating MIST models instead point to an unusually narrow stellar rotation distribution, $\Delta c_{\rm F475W-F814W}\approx 0.25$5, with best fits requiring essentially no age spread $\Delta c_{\rm F475W-F814W}\approx 0.25$6 and a very narrow $\Delta c_{\rm F475W-F814W}\approx 0.25$7 distribution peaked at $\Delta c_{\rm F475W-F814W}\approx 0.25$8 with width $\Delta c_{\rm F475W-F814W}\approx 0.25$9 (&&&^^^^3query3^^^^&&&). This functions as a control case: if a broad spread in rotation rates produces an eMSTO, a narrow spread can erase it.

Other studies delineate where rotation may be insufficient. In NGC ^^^^3all:\3^^^^83^^^^3all:\3 a main-sequence “kink” at an initial stellar mass of $\Delta_{\rm MSTO}$^^^^3query3^^^^ was interpreted as the threshold below which deep convection and magnetic braking render rotational photometric effects negligible. Since some $\Delta_{\rm MSTO}$^^^^3all:\3^^^^ Gyr clusters show wide eMSTOs already at masses well below that threshold, it was argued that stellar rotation cannot fully explain the wide extent of those intermediate-age Magellanic Cloud eMSTOs and that a distribution of stellar ages still seems necessary (&&&3^^^^3all:\3^^^^&&&). In the very young double cluster $\Delta_{\rm MSTO}$^^^^3 OR all:\3^^^^ and $\Delta_{\rm MSTO}$3 Persei, a single-age population spanning the full $\Delta_{\rm MSTO}$4 range cannot reproduce the observed $\Delta_{\rm MSTO}$5 mag spread; variable stars, binary interactions, Be-disc reddening, and possible star-to-star overshooting variations were proposed as additional contributors (Li et al., 2019).

Several alternative or complementary mechanisms have therefore been advanced. One is star-to-star variation in the overshooting of the convective core, parameterized by

$\Delta_{\rm MSTO}$6

Non-rotating YREC models with different $\Delta_{\rm MSTO}$7 values show that if $\Delta_{\rm MSTO}$8 varies within a coeval cluster, the resulting MSTO broadening can match observed equivalent age spreads in many young and intermediate-age systems; Yang and Tian further argue that OVCC can reproduce the eMSTO of NGC ^^^^3all:\3^^^^856, although it does not by itself produce a strongly split main sequence at $\Delta_{\rm MSTO}$9 (Yang et al., 2017). A related framework is the “isochrone-cloud” method, in which non-rotating MESA stellar models with varying convective-boundary and radiative-envelope mixing define a region rather than a line in CMD space; in NGC ^^^^3all:\3^^^^85^^^^3query3^^^^ and NGC 884, such isochrone-clouds explain a good fraction of the observed eMSTOs in terms of one coeval population, without invoking an age spread (Johnston et al., 2019).

Circumstellar dust has also been proposed as a modifier of eMSTO morphology. For NGC ^^^^3all:\3^^^^783, a dust-ring model around rapidly rotating stars in the Be stage was shown to reproduce a UV-dim tail on the red side of the eMSTO, with modest optical depths at $\sim 900$^^^^3query3^^^^ and color shifts that scale with inclination and dust column. In that picture, dusty stars could populate a substantial fraction of the eMSTO fan, especially in ultraviolet filters (D'Antona et al., 2023). By contrast, Swift UVOT analysis of NGC ^^^^3 OR all:\33^^^^55 found no statistically significant UV extinction excess among red-side eMSTO stars, effectively ruling out dust-like extinction from excretion discs as a major contributor in that cluster (&&&^^^^3 OR all:\3^^^^7&&&).

Finally, recent open-cluster studies increasingly focus on the origin of the rotation-rate spread itself. In NGC 6^^^^3query3^^^^67 and NGC ^^^^3 OR all:\33^^^^55, binary tidal locking explains some slow rotators, but the observed spatial distributions are inconsistent with tidal braking as the dominant driver of the red/blue split. Both studies instead propose pre-main-sequence star–disk interaction as the more plausible mechanism for establishing the spread in stellar spins that later manifests as an eMSTO (&&&^^^^3 OR all:\3^^^^8&&&, &&&^^^^3 OR all:\3^^^^7&&&).

The cumulative literature therefore supports a layered interpretation. Stellar rotation is strongly supported as a primary determinant of eMSTO morphology, especially in Galactic open clusters with direct $\sim 900$^^^^3all:\3^^^^ measurements. At the same time, the Magellanic Cloud literature preserves credible cases for hybrid models, true age spreads in some systems, and additional contributions from overshooting, interior mixing, binarity, pulsation, Be-related phenomena, and dust. A plausible implication is that “eMSTO” is best regarded as an observational phenotype rather than a single-cause category: the same CMD signature can emerge from different combinations of stellar spin physics, internal mixing, and cluster-specific evolutionary history (&&&^^^^3all:\3^^^^8&&&).