---
title: Extended Main Sequence Turn-Off (eMSTO)
url: https://www.emergentmind.com/topics/extended-main-sequence-turn-off-emsto-f0fa7621-8ef9-450c-99c8-f470e5f304d0
type: topic
---

# Extended Main Sequence Turn-Off (eMSTO)

Searching arXiv for recent and foundational papers on extended main-sequence turn-offs to ground the article in the literature.
arxiv.search({"query":"all:\"extended main sequence turnoff\" OR all:\"extended main-sequence turn-off\"","max_results":10,"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 DR2, 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 [1911.02075, 1102.1723].

## 1. 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 \(200\) Myr–\(2\) Gyr, the turn-off broadening typically spans \(\Delta(G_{\rm BP}-G_{\rm RP}) \simeq 0.05\)–\(0.15\) 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 411 [1911.02075, 1303.1361].

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, \(\Delta_{\rm MSTO}\sim 200\)–\(400\) Myr is typical, whereas the lower main sequence remains narrow, with \(\Delta_{\rm MS}<50\) Myr equivalent width. This localization of the broadening to the turn-off region is one of the basic empirical constraints on any physical explanation [1911.02075].

The morphology is age-dependent. In clusters younger than \(\sim 700\) 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 \(h\) and \(\chi\) Persei, stars below \(\sim 1.3\,M_\odot\) 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 [1811.01192, 1904.02005].

## 2. Discovery across environments and demographic trends

The eMSTO was first established observationally in Magellanic Cloud clusters. In the LMC, intermediate-age clusters with \(1.0\,{\rm Gyr}<\tau<3.0\,{\rm Gyr}\) and \(M_{\rm cluster}\approx 10^{4.8}\)–\(10^{5.3}\,M_\odot\) were found to exhibit bifurcated or extended MSTOs in a large fraction of cases; Milone et al. (2009), as summarized in the literature, examined 16 clusters and found 11 with eMSTOs, with inferred age spreads up to \(300\) Myr [1102.1723]. In the SMC, NGC 411 and NGC 419 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 [1303.1361].

Subsequent Gaia-based work established that eMSTOs are not peculiar to the Magellanic Clouds. NGC 6705 (M11) provided the first clear Galactic open-cluster case, followed by NGC 2099, NGC 2360, and NGC 2818; a broader Gaia DR2 analysis then concluded that twelve Milky Way open clusters younger than \(\sim 1.5\) Gyr show eMSTOs and/or broadened main sequences, and that nearly all clusters younger than \(\sim 2.5\) Gyr in both the Magellanic Clouds and the Milky Way exhibit the phenomenon [1807.05888, 1811.01192].

The strongest large-sample trend is with age. For \(100\) Myr \(\lesssim t \lesssim 1\) Gyr, one empirical fit gives \(\Delta t_{\rm FWHM}\simeq (0.33\pm0.03)\,t\), while another Gaia-based open-cluster study finds \({\rm FWHM}\simeq (0.25\pm0.03)\,t_{\rm cluster}\) when \(t\) is in Myr [1807.10779, 1811.01192]. In larger LMC samples extending to older ages, the width–age relation becomes non-monotonic, with a pronounced inverted-\(V\) or “\(\Lambda\)” shape: MSTO widths are small in younger systems, peak near \(\log(t/{\rm yr})\simeq 9.2\) (\(\sim 1.6\) Gyr), and decline again at older ages [1608.04554]. 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 DR2 workflow extracted all sources within \(10\) arcmin of each cluster, selected members using proper-motion cuts \((\sigma_\mu<0.5\ {\rm mas\ yr}^{-1})\), parallax selection within \(2\sigma\) of the parallax peak, and a 3D membership-probability filter following Kamann et al. (2014). Distances were obtained from the median parallax of members with \(G<17\) mag, and extinctions were estimated by visually fitting non-rotating MIST isochrones to the unevolved main sequence [1911.02075].

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, \(\sigma_{\rm MSTO}/\sigma_{\rm MS}\), provides a width measure that is relatively insensitive to zero-point uncertainties and small age offsets. In well-behaved eMSTO clusters, \(\sigma_{\rm MSTO}/\sigma_{\rm MS}\approx 2.4\)–\(4.2\), whereas a narrow turn-off has \(\sigma_{\rm MSTO}/\sigma_{\rm MS}\sim 1.4\) [1911.02075].

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:
\[
{\rm FWHM}_{\rm MSTO}=\sqrt{{\rm FWHM}_{\rm obs}^2-{\rm FWHM}_{\rm SSP}^2}\, .
\]
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 [1410.3840, 1707.03847].

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 152 to within \(1.3\,\sigma\), suggesting that the age-spread-style width measurement is robust across photometric systems [2507.15921].

## 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 \(\Delta\tau \approx 50\)–\(300\) 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 \(50\)–\(300\) Myr, and a second generation then forms from retained ejecta mixed with accreted pristine gas [1102.1723].

A more dynamical version of this argument emphasized cluster escape velocity. In a mass-limited sample of 18 intermediate-age Magellanic Cloud clusters, intrinsic MSTO widths of \(200\)–\(550\) Myr were found to correlate with present and early central escape velocities, with eMSTO clusters satisfying \(v_{\rm esc}\gtrsim 15\ {\rm km\ s^{-1}}\) at early times, whereas two lower-mass clusters without eMSTOs remained at \(v_{\rm esc}\lesssim 12\ {\rm km\ s^{-1}}\). That threshold was argued to be consistent with AGB-wind speeds, favoring in situ extended star formation regulated by cluster potential depth [1410.3840].

This interpretation remains part of the literature, but it is not uncontested. A larger homogeneous LMC analysis found that eMSTO clusters span \(2\,{\rm pc}\lesssim r_c\lesssim 6\,{\rm pc}\), \(3.35\lesssim \log(M_{\rm cls}/M_\odot)\lesssim 5.50\), and \(0.2\lesssim \tau_{\rm age}/\tau_{\rm relax}\lesssim 8.0\), with no statistically significant correlation between \( {\rm FWHM}_{\rm MSTO}\) 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 [1608.04554].

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 \(\sim 1\) Gyr clusters such as NGC 1987 and NGC 2249, rotation-only models with realistic rotation distributions and random inclinations account for only \(\sim 60\%\) and \(\sim 40\%\) 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 \(\sim 2\)–\(3\times 10^8\) yr [1707.03847]. 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
\[
T_{\rm eff}\propto g_{\rm eff}^{\beta},
\]
and \(\beta\approx 0.25\) for radiative envelopes in the formulations summarized in the literature. In MIST models the relevant control parameter is
\[
f_{\rm rot}\equiv \Omega/\Omega_{\rm crit,ZAMS},
\]
the equatorial angular velocity relative to the ZAMS critical value. At fixed age, increasing \(f_{\rm rot}\) shifts isochrones to redder colors and higher luminosities around the MSTO, so a population with a spread in \(f_{\rm rot}\) naturally develops an eMSTO [1911.02075].

Direct spectroscopic support for this interpretation is strong in several Galactic open clusters. In M11, 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 \(\langle v\sin i\rangle_{\rm bMS}=83\pm14\ {\rm km\ s^{-1}}\) and \(\langle v\sin i\rangle_{\rm rMS}=194\pm15\ {\rm km\ s^{-1}}\) [1807.05888]. In NGC 2818, stars on the red side of the eMSTO have \(V\sin i>160\ {\rm km\ s^{-1}}\) while stars on the blue side have \(V\sin i<160\ {\rm km\ s^{-1}}\), and the cluster follows the empirical age–eMSTO-extent relation previously established in Magellanic Cloud work [1807.10779]. Comparable red-fast / blue-slow segregation is reported for NGC 5822, NGC 2355, and NGC 6067, where mean projected rotation velocities or direct correlations between \(v\sin i\) and pseudo-color again place faster rotators on the red side of the MSTO [1904.03547, 2406.19132, 2507.14433].

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 \(100\)–\(300\) Myr in systems whose stars may be nearly coeval [1811.01192]. 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 2509. Among Galactic open clusters around \(\sim 900\) Myr, it is unusual in showing an exceptionally narrow CMD, with an estimated age of \(860\) Myr, \(\Delta_{\rm MSTO}\approx 60\) Myr from non-rotating fits, and \(\sigma_{\rm MSTO}/\sigma_{\rm MS}=1.38\). 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, \(\Omega/\Omega_{\rm crit,ZAMS}=[0.4,0.6]\), with best fits requiring essentially no age spread \((\sigma_t\lesssim 15\ {\rm Myr})\) and a very narrow \(f_{\rm rot}\) distribution peaked at \(f_{\rm rot}=0.5\) with width \(\Delta f_{\rm rot}\lesssim 0.1\) [1911.02075]. 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 1831, a main-sequence “kink” at an initial stellar mass of \(1.45\pm0.02\,M_\odot\) was interpreted as the threshold below which deep convection and magnetic braking render rotational photometric effects negligible. Since some \(\sim 1.7\) 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 [1807.04737]. In the very young double cluster \(h\) and \(\chi\) Persei, a single-age population spanning the full \(0\leq \omega\leq 0.95\) range cannot reproduce the observed \(\Delta(G_{\rm BP}-G_{\rm RP})\simeq 0.25\) mag spread; variable stars, binary interactions, Be-disc reddening, and possible star-to-star overshooting variations were proposed as additional contributors [1904.02005].

Several alternative or complementary mechanisms have therefore been advanced. One is star-to-star variation in the overshooting of the convective core, parameterized by
\[
\Delta_{\rm ov}=\delta_{\rm ov} H_P .
\]
Non-rotating YREC models with different \(\delta_{\rm ov}\) values show that if \(\delta_{\rm ov}\) 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 1856, although it does not by itself produce a strongly split main sequence at \(m_U\lesssim 21\) [1701.05963]. 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 1850 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 [1910.00591].

Circumstellar dust has also been proposed as a modifier of eMSTO morphology. For NGC 1783, 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 \(10\,\mu{\rm m}\) 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 [2303.16049]. By contrast, Swift UVOT analysis of NGC 2355 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 [2406.19132].

Finally, recent open-cluster studies increasingly focus on the origin of the rotation-rate spread itself. In NGC 6067 and NGC 2355, 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 [2507.14433, 2406.19132].

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 \(v\sin i\) 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 [2507.15921].

Source: https://www.emergentmind.com/topics/extended-main-sequence-turn-off-emsto-f0fa7621-8ef9-450c-99c8-f470e5f304d0