Recent results

The results shown on this page can still be preliminary.

N/O-enhanced chemical enrichment by massive stars: the interplay of optically-thick winds and rotation (arXiv:2609.26998)

Recent JWST observations have revealed a sample of high-z galaxies with highly-enhanced N/O abundance ratios, log(N/O)≳−1.1. Strong optically-thick winds of rotating massive stars going through the Wolf-Rayet (WR) phase provide a viable explanation for such N/O enhancements, supported by the direct spectroscopic WR signatures discovered in a few galaxies. Such strong winds are also needed to reproduce the single WR stars in the Small Magellanic Cloud (SMC). Here, we compute the metal yields of massive (∼20−100 M⊙), metal-poor stars using two sets of MESA single stellar evolution models (Boco et al. 2025) with and without optically-thick winds in a dense grid of metallicities (Z∼0.001−0.0045) and initial spins (ω≡Ω/Ωcrit∼0−0.7). Considering an initial spin distribution based on observations of O-type stars in the SMC, we find that both sets produce strong N/O enhancements up to log(N/O)∼0.1 within a few Myr after a starburst. The models with optically-thick winds undergo a rapid transition from high N/O to the "normal" state of log(N/O)∼−1.5 at t∼3−4 Myr driven by C/O-rich winds, while the transition is slower and weaker without optically-thick winds. The optically-thick-wind scenario can reproduce the nebular C, N, O, and He abundance patterns of three representative high-z galaxies showing direct WR signatures: RXCJ2248-ID, Sunburst Arc, and MARTA 4327, with stellar metallicities and ages similar to those inferred from SED-fitting. This can be interpreted as a coherent evolution sequence captured before (t≲3 Myr), during (t∼3−4 Myr), and after (t≳5 Myr) the transition. In contrast, the optically-thin-wind-only scenario ejects much less C and O via winds and cannot reproduce Sunburst Arc and MARTA 4327. Our yields are published on Zenodo.

Abundance patterns of population-averaged cumulative stellar ejecta that include the contributions of stars with different initial spins according to the observed spin distribution of O stars in the SMC for the models without (left) and with (right) optically-thick winds. The shaded regions denote the abundance-ratio space generally occupied by observed N/O-enhanced (NOE) galaxies, for which we show the sample compiled by Ji et al. (2026) with the small squares. The big circle, square, hexagon, and star show the abundance patterns of three representative high-z galaxies with WR signatures, RXCJ2248-ID, Sunburst Arc, and MARTA 4327 (Berg et al. 2026; Curti et al. 2026; Moreschini et al. 2026), and the GAll stacked spectrum of z ∼ 4.5 − 10.1 star-forming galaxies in lensing-cluster fields (Umeda et al. 2026). The solid and dashed curves show the [N/O]-[O/H] and [C/O]-[O/H] scaling relations from local observations (Nicholls et al. 2017; Cataldi et al. 2026).

Metallity-stellar mass relation (MZR) at z~4-10. Simulated galaxies from our best-fit model are shown as dots/circles, color-coded by SFR. A linear fit in log-log space is derived from the simulated galaxies with log(SFR [Msun/yr])>-0.5 that are detectable in current JWST spectroscopy surveys as the long-dashed line, i.e., the predicted MZR. The other lines show the MZRs obtained from observations, which are corrected for the bias of observations towards high-SFR galaxies. The diamonds with errorbars show the original average values (without corrections) and 1 sigma scatters of the stellar mass, metallicity, and SFR of the galaxies observed by JWST from Nakajima et al. (2023). The small dots with errorbars denote the extremely metal-poor galaxy (EMPG, [O/H]<-1.7) candidates observed by JWST.

Impact of initial mass function on the chemical evolution of high-redshift galaxies (arXiv:2506.06139)

Star formation and metal enrichment in galaxies are regulated by supernova (SN) explosions and metal yields from massive stars, which are sensitive to the high-mass end of the initial mass function (IMF). Recent JWST observations have reached a consensus on an invariant relation between stellar mass, metallicity, and star formation rate up to z~8 and its breakdown at higher redshifts. It is crucial to understand the underlying physics, especially the role played by the IMF. We explore the impact of IMF on the chemical evolution of high-redshift galaxies and the interplay between IMF and galactic outflow parameters. The ultimate goal is to constrain the high-mass end of the IMF by the cosmic star formation history and stellar mass-metallicity-star formation rate relation (MZSFR) inferred from observations at z~4-10. Using the semi-analytical galaxy evolution code A-SLOTH, we follow galactic baryon cycles along merger trees built from a high-resolution cosmological simulation. Stellar feedback is modeled with up-to-date stellar evolution tracks covering the full metallicity range (Z~10^-11-0.03) and a broad stellar mass range (2-600 Msun) including the metal yields from stellar winds, core-collapse SNe, (pulsational) pair-instability SNe, and Type Ia SNe. Assuming that the IMF follows a Kroupa-like shape with a varying upper mass limit m_max, we find m_max≳200 Msun is required to reproduce the observed MZSFR. Observational data at z≳6 favor a galactic outflow model where the outflow rate is proportional to the supernova energy injection rate divided by the halo binding energy. We conclude that very massive (≳200 Msun) stars can play important roles in the star formation and chemical enrichment histories of high-z galaxies. Besides, while reproducing current observations biased by luminous objects, our best-fit model self-consistently produces a large population of (unseen) faint, low-mass, metal-poor (Z ≲ 0.1 Z⊙) galaxies that dominate cosmic star formation and metal enrichment at z ≳ 10. We also discuss the implications of such model predictions for reionization and transient sources of both electromagnetic waves and gravitational waves.

Signatures of rapidly rotating stars with chemically homogeneous evolution in the first galaxies (2025, ApJ, 980, L30)

The James Webb Space Telescope (JWST) has revealed an unexpectedly high abundance of UV luminous galaxies at redshifts z ≳ 10, challenging ‘standard’ galaxy formation models. We investigate the role of rapidly rotating (massive) stars undergoing chemically homogeneous evolution (CHE) in reconciling this potential tension. These stars are more compact, hotter, and exhibit enhanced UV emission. We find that the rest-frame UV luminosity of star-forming galaxies can be significantly enhanced by a factor of ∼ 3 − 6 when CHE stars above a minimum initial mass of ∼ 2 − 10 M⊙ account for more than half of the total stellar mass following a Salpeter initial mass function (IMF). As a result, the UV luminosity functions observed at z ∼ 12 − 16 can be reproduced with less extreme values of star formation efficiency and UV luminosity stochastic variability. Moreover, in the CHE scenario, the feedback of supernova explosions is less enhanced compared with the alternative scenario of top-heavy IMFs. The UV spectra also become harder under CHE, which can possibly explain the bluer galaxies at higher redshifts in observations.

Our results highlight the potential of CHE in explaining the UV-bright galaxy populations detected by JWST and call for future work to explore the broader astrophysical implications of CHE and its associated phenomena in the early universe, such as gamma-ray bursts, compact object binaries, and metal enrichment.

Rest-frame UV luminosity per unit star formation rate η_UV for a top-hat star formation history within a timescale of 100 Myr as a function of the minimum initial mass of CHE stars given the Salpeter IMF slope α = 2.35. For the solid curves, the minimum mass of non-rotating stars is fixed to 1.7 M⊙, and the mass fraction of CHE stars f_CHE increases for lighter colors. The dotted curve with errorbars shows the maximum η_UV in an alternative population model with a fixed overall IMF. We also show the reference value ηUV ≃ 8.70 × 10^27 from Madau & Dickinson (2014) with the dashed line, which is very close to our result for f_CHE = 0. The shaded regions show the results for Pop III stars with extremely top-heavy IMFs from Inayoshi et al. (2022).

Characteristic masses for primordial star formation as functions of initial gas inflow rate (thin curves of different colors), which are combined to predict the final mass of Pop III stars (thick solid curve), assuming that only one star forms per cloud. Simulation results are shown for comparison (data points and a fit as the long-dashed line).

Initial mass functions of Pop III stars formed in the progenitors of typical luminous quasar host haloes (Li et at. 2021) under different degrees of fragmentation (denoted by the number of stars per cloud N), compared with the IMF predicted by Hirano et al. (2015) using 2D (single-star) simulations for Pop III star formation in unbiased regions (thick dashed).

Towards a universal analytical model for Population III star formation: interplay between feedback and fragmentation (arXiv:2407.14294)

JWST has brought us new insights into Cosmic Dawn with tentative detection of the unique signatures of metal-free Population III (Pop III) stars, such as strong HeII emission, extremely blue UV spectrum, and enhanced nitrogen abundance. Self-consistent theoretical predictions of the formation rates, sites, and masses of Pop III stars are crucial for interpreting the observations, but are challenging due to complex physical processes operating over the large range of length scales involved. One solution is to combine analytical models for small-scale star formation processes with cosmological simulations that capture the large-scale physics, such as structure formation, radiation backgrounds, and baryon-dark matter streaming motion that regulate the conditions of Pop III star formation.

We build an analytical model to predict the final masses of Pop III stars/clusters from the properties of star-forming clouds, based on the key results of small-scale star formation simulations and stellar evolution models. Our model for the first time considers the interplay between feedback and fragmentation and covers different modes of Pop III star formation ranging from ordinary small (∼10−2000 M⊙) clusters in molecular-cooling clouds to massive (≳10^4 M⊙) clusters containing supermassive (∼10^4−3×10^5 M⊙) stars under violent collapse of atomic-cooling clouds.

As an example, the model is applied to the Pop III star-forming clouds in the progenitors of typical haloes hosting high-z luminous quasars, which shows that formation of Pop III massive clusters is common (∼20−70%) in such biased (∼4σ) regions, and the resulting heavy black hole seeds from supermassive stars can account for a significant fraction of observed luminous (≳10^46 erg s−1) quasars at z∼6.

Fig. 1. Co-moving merger rate density of Pop III BBHs as a function of redshift in our fiducial model (using the LOG1 initial condition model for Pop III binary stars, see Sec. 2.3 in Costa et al., 2023). The results for mergers in galaxy fields, NSCs, and combined are shown with the thin solid, dotted, and thick solid curves, respectively. The rate density of NSC infall is also shown with the dash-dotted curve. For comparison, the other curves show the results for other populations of BBH mergers from Franciolini et al. (2022). The local rate inferred from observations is shown with the triangle (Abbott et al., 2021).

Fig. 2. Detection rates of Pop III BBH mergers in galaxy fields (diamonds) and NSCs (circles, triangles, and squares) by ET with SNR > 8 for 18 models combining 3 IMFs (delineated by the power-law slope 𝛼), 2 models for initial binary statistics (close/left versus wide/right), and 3 models for NSCs (obs, low, and full).

Fig. 3. Intrinsic primary mass distributions of Pop III BBH mergers in galaxy fields (solid) and NSCs (dotted), initial BBHs (dashed), BBHs that fall into NSCs (dash-dotted), and progenitor stars (shaded) in the fiducial model. The primary mass distribution inferred from the 69 confident BBH mergers in GWTC-3 (Edelman et al. 2023) is shown with the thick solid curve.

Binary black hole mergers from the first stars: roles played by two evolution channels (arXiv:2406.17397)

The gravitational wave (GW) signal from binary black hole (BBH) mergers is a promising probe of population III (Pop III) stars, which are more efficient at producing massive black holes (BHs) than their population I/II (Pop I/II) counterparts. To fully unleash the power of the GW probe, one important step is to understand the relative importance and unique features of different evolution channels. We implement two channels, i.e., isolated binary stellar evolution (IBSE) and nuclear star cluster-dynamical hardening (NSC-DH), in the semi-analytical model A-SLOTH (Hartwig et al., 2022) for early star/galaxy formation. This framework is combined with binary population results from the code SEVN (Costa et al., 2023) to predict the BBH merger properties under various assumptions on Pop III initial mass function (IMF), binary statistics and high-z NSCs.

We find that the NSC-DH channel can be as efficient as the IBSE channel in producing Pop~III BBH mergers at z<9 (Fig. 1), which contributes 8-95% of Pop III BBH mergers across cosmic history, with higher contributions achieved by initially wider binary stars, more top-heavy IMFs, and more abundant high-z NSCs. The stochastic GW background (SGWB) produced by Pop III BBH mergers has a peak value of 1-8*10^-11 around observer-frame frequencies 10-100 Hz, which can be a non-negligible (~2-32%) component in the total SGWB below 10 Hz. The estimated detection rates of Pop~III BBH mergers by the Einstein Telescope (ET) are ~6-230 and ~30-1230 events per year for the NSC-DH and IBSE channels, respectively (Fig. 2). BBH mergers in NSCs are more massive than those from IBSE (Fig. 3), so they dominate the Pop III SGWB below ~20 Hz in most cases (Fig. 4). Besides, the detection rate of Pop III BBH mergers involving at least one intermediate-mass BH (IMBH) above 100 Msun by the Einstein Telescope is 0.5-200/yr in NSCs (Fig. 5) but remains below 0.1/yr for IBSE.

Fig. 4. SGWB energy density spectrum contributed by Pop III BBH mergers in the fiducial model. The annotations are identical to those in Fig. 1. The contributions of other populations of BBH mergers are taken from Bvera et al. (2022). The shaded regions show the sensitivities of two gravitational wave detectors. The triangle shows the total SGWB energy density estimated by population analysis through GWTC-3 (Abbott et al. 2023).

Fig. 5. Same as Fig. 2 but for the detection rates of Pop III BBH mergers including at least one IMBH (>100 Msun) in NSCs. The detection rates of such mergers in galaxy fields remain below 0.1/yr and are therefore not shown here.

Face-on view of an isolated dwarf galaxy simulated by RIGEL for 1000 Myr. Gas distribution is shown in red and yellow. Stars are color-coded by their ages.

Fig. 1. Production rates of photons in select energy bands considered by AREPO-RT as functions of metallicity and initial stellar mass. The curves of Opt., FUV, LW, and EUV0 bands are shifted by adding constants of 4, 3, 2, and 1, respectively, to reduce overlapping.

RIGEL: Simulating dwarf galaxies at solar mass resolution with radiative transfer and feedback from individual massive stars (arXiv:2405.08869)

Based on the SMUGGLE model (Marinacci et al., 2019), we design a novel framework RIGEL (Realise Interstellar medium in Galactic Ecosystem and Lifecycle) of star formation and stellar feedback for the radiative hydrodynamic simulation code AREPO-RT (Kannan et al., 2019) that self-consistently models the radiative feedback, stellar winds, and supernovae from individual massive stars, taking into account the metallicity dependence of stellar evolution (Fig. 1). The model has just been applied to simulations of isolated metal-poor dwarf galaxies in the local Universe. An example is shown in the video. We found that early radiative feedback from massive stars has significant effects on star cluster properties.

For instance, compared with the run without radiative feedback, the star formation rate (SFR) and cluster formation efficiency (CFE) are lower when radiative feedback is turned on in our fiducial simulation (Fig. 2). The scatter in CFE is also larger. In particular, radiative feedback suppresses the formation of massive star clusters (Fig. 3). On the other hand, increasing metallicity enhances star formation but reduces both the CFE and its scatter. Besides, the number of massive star clusters is slightly increased with higher metallicity.

Fig. 2. CFE (ratio of the formation rate of stars in bound clusters over the total star formation rate) as a function of SFR surface density. The results for 3 high-resolution (with gas mass resolution ~ 1 Msun) simulations are shown, including two cases with full feedback for different metallicities, and a model without radiative feedback for the lower metallicity case. The crosses show the results from observations of nearby dwarf galaxies (Cook et al., 2012, 2023).

Fig. 3. Mass distribution of star clusters in the 3 sets of simulations described in Fig. 2. The dotted curves show the results from low-resolution runs (with gas mass resolution ~ 10 Msun).

Effects of chemically homogeneous evolution of the first stars on the 21-cm signal and reionization (arXiv:2504.00535)

The UV radiation from massive stars is sensitive to stellar evolution and regulates the thermal and ionization histories at Cosmic Dawn, which are captured by the 21-cm signal. It is shown in our earlier work Sibony et al. (2022) that the UV emission from Pop III stars can be significantly boosted by chemically homogeneous evolution (CHE) due to fast rotation, which may facilitate early cosmic reionization. In the hope of constraining Pop III stellar evolution with observations of Cosmic Dawn, we investigate the effects of Pop III CHE by combining spectra of Pop III stars (Fig. 1) derived by the atmosphere model TLUSTY (Hubeny et al., 2017) with semi-numerical simulations of Cosmic Dawn (Gessey-Jones et al., 2023). Our results show that CHE boosts the ionizing power of Pop III stars by a factor of 2, which reduces the Pop III stellar mass density required to reproduce the observed Thomson scattering optical depth proportionally (Fig. 2). Meanwhile, the maximum 21-cm global absorption signal is shallower by up to 15 mK (11%), partly due to the reduced Lyman-band emission from CHE, and the large-scale (k ~ 0.2/cMpc) power drops by a factor of a few at z>25 (Fig. 3). In general, the effects of CHE can be comparable to those of Pop~III star formation parameters. Our results highlight the importance of metal-free/poor stellar evolution in understanding the early Universe and suggest that future studies should consider joint constraints on the physics of star/galaxy formation and stellar evolution.

Fig. 1. Main sequence lifetime-integrated spectra of chemically-homogeneous (CH, solid) and non-rotating (NR, dotted) Pop III stars with initial masses of 9, 40, 120 and 300 Msun. The vertical lines label the characteristic energies of 5 types of UV photons.

Fig. 2. Impact of Pop III CHE on the optical depth of Thomson scattering (left) and cosmic ionization history (right) with varying Pop III star formation efficiency (SFE). The dotted curves show the `standard' case for non-rotating (NR) Pop III stars, while the solid curves show the CHE case. The curves are color-coded by the Pop III SFE. The special models with SFE values required to produce the best-fitting value and the 3-sigma upper limit of the optical depth for Thomson scattering measured by Planck are shown by the lower and upper thick curves, respectively.

Fig. 3. Impact of Pop III CHE on the global 21-cm signal (left) and the power spectrum of the 21-cm signal (right) with varying Pop III SFE. The annotations are identical to those in Fig. 2.

Page updated

Google Sites

Report abuse