arXiv subjects

Ben Davies

Publications and source records attributed to Ben Davies.

At least 19 recordsLinked to original sources

arXiv2026

JWST spectroscopy of SN 2010da/NGC 300 ULX-1: a surviving star hidden by dust

We present new James Webb Space Telescope ($JWST$) NIRSpec and MIRI integral-field spectroscopy of the remarkable system SN~2010da / NGC~300 ULX-1, the only known ultraluminous X-ray source powered by a neutron star with a supergiant donor. Our new data, taken in November 2024, reveal that the optical and near-infrared counterpart has dramatically faded since 2018 and no longer exhibits molecular absorption features characteristic of a red supergiant. Instead, the spectral energy distribution shows the donor has returned to its pre-outburst appearance, and is dominated by infrared continuum consistent with an optically thick warm ($\approx$900~K) dust shell. The bolometric luminosity indicates the presence of a surviving luminous source with $\log(L/L_{\odot})=4.11\pm0.02$. Radiative transfer modelling of the mid-infrared spectral energy distribution (SED) reveals a broad emission feature peaking near $\sim$11\,$\mu$m, best reproduced by silicon carbide (SiC) dust grains, a composition typically associated with carbon-rich evolved stars. We rule out a failed supernova scenario, which would predict a large drop in luminosity and continued fading. Given the association of SiC dust with AGB stars, we suggest the donor star may be an AGB star that has survived and is now heavily enshrouded - having returned to a dust-obscured state following a transient post-outburst phase in which the system appeared as a red supergiant. We propose a revised evolutionary timeline in which the 2010 outburst initiated sustained super-Eddington accretion and temporarily altered the circumstellar environment. These observations provide rare insight into eruptive mass loss, dust formation, and binary interaction in a unique system.

astro-ph.HE↗

arXiv2026

Atomic-resolution imaging of gold species at organic liquid-solid interfaces

Understanding solid-liquid interfaces at the atomic-scale is key to improved performance of heterogeneous catalysts, electrodes and membranes. Here we combine unique specimen design, record atomic resolution in situ electron microscopy, and artificial intelligence-enabled analysis to achieve a step change in quantitative understanding of interfacial atomic behaviour. We create the first graphene liquid cells with organic solvents and employ them to track over 106 gold adatoms and clusters at a graphene surface immersed in acetone and cyclohexanone. We reveal dynamic correlated behaviour of gold adatom monomers, dimers, trimers and clusters, strongly influenced by each other, the solvent properties, and the atomic lattice of the substrate, in good agreement with theoretical calculations. We use the results to interpret differences in catalytic activity towards the industrially important acetylene hydrochlorination reaction. This new capability for exploration of atomic scale chemistry could enable rational design of future catalysts, membranes and electrodes with improved functionality.

cond-mat.mtrl-sci↗

arXiv2025

SafePassage: High-Fidelity Information Extraction with Black Box LLMs

Black box large language models (LLMs) make information extraction (IE) easy to configure, but hard to trust. Unlike traditional information extraction pipelines, the information "extracted" is not guaranteed to be grounded in the document. To prevent this, this paper introduces the notion of a "safe passage": context generated by the LLM that is both grounded in the document and consistent with the extracted information. This is operationalized via a three-step pipeline, SafePassage, which consists of: (1) an LLM extractor that generates structured entities and their contexts from a document, (2) a string-based global aligner, and (3) a scoring model. Results show that using these three parts in conjunction reduces hallucinations by up to 85% on information extraction tasks with minimal risk of flagging non-hallucinations. High agreement between the SafePassage pipeline and human judgments of extraction quality mean that the pipeline can be dually used to evaluate LLMs. Surprisingly, results also show that using a transformer encoder fine-tuned on a small number of task-specific examples can outperform an LLM scoring model at flagging unsafe passages. These annotations can be collected in as little as 1-2 hours.

cs.CL↗

arXiv2025

Variable Polarization of WR~31a: Binary Companion or Co-Rotating Interaction Region?

WR 31a (Hen 3-519) is likely a post-luminous blue variable (LBV) star that is evolving to become a classical Wolf-Rayet star. Multicolor (UBVR) photopolarimetric observations of WR 31a were obtained over nine nights in early 2007. The linear polarization data of WR 31a trace a "loop" structure in a Stokes Q-U diagram, which is similar in all four passbands. After mean subtraction, the four loops align to form a single overall pattern. Such loops can be expected to arise from binary systems. We test the binary hypothesis with two models. The data are fit for a strictly circular orbit to derive an orbital period of 16.7 d, requiring a high inclination perspective of $i\sim 80^\circ$. We also consider an elliptical orbit under simplifying assumptions, yielding a match for $i\sim 75^\circ$ with eccentricity $e\sim 0.5$ and a longer orbital period of about 70 d. The prevalence of binarity among massive stars is well-known; the prospect of detecting a binary companion during the post-LBV stage of WR 31a would add to an emerging narrative of diverse interactions between massive multiple components as a function of evolutionary stage. However, if the loop originates because of a co-rotating interaction region (CIR), then the rotation period could be 8.5 d or 17 d. This would give an estimated equatorial rotation speed of 95 or 190 km/s. Either of these is a significant fraction of the estimated critical speed of rotational break-up at 320 km/s (for an Eddington factor of $\Gamma=0$).

astro-ph.SR↗

arXiv2024

JWST reveals a luminous infrared source at the position of the failed supernova candidate N6946-BH1

N6946-BH1 is the first plausible candidate for a failed supernova (SN), a peculiar event in which a massive star disappears without the expected bright SN, accompanied by collapse into a black hole (BH). Following a luminous outburst in 2009, the source experienced a significant decline in optical brightness, while maintaining a persistent infrared (IR) presence. While it was proposed to be a potential failed SN, such behavior has been observed in SN impostor events in nearby galaxies. Here, we present late-time observations of BH1, taken 14 years after disappearance, using JWST's NIRCam and MIRI instruments to probe a never-before-observed region of the object's spectral energy distribution. We show for the first time that all previous observations of BH1 (pre- and post-disappearance) are actually a blend of at least 3 sources. In the near-IR, BH1 is notably fainter than the progenitor but retains similar brightness to its state in 2017. In the mid-IR, the flux appears to have brightened compared to the inferred fluxes from the best-fitting progenitor model. The total luminosity of the source is between 13 - 25% that of the progenitor. We also show that the IR SED appears consistent with PAH features that arise when dust is illuminated by near-ultraviolet radiation. At present, the interpretation of N6946-BH1 remains uncertain. The observations match expectations for a stellar merger, but theoretical ambiguity in the failed SN hypothesis makes it hard to dismiss.

astro-ph.SR↗

arXiv2022

The effect of winds in red supergiants: modeling for interferometry

Red supergiants (RSGs) are evolved massive stars in a stage preceding core-collapse supernova. Understanding evolved-phases of these cool stars is key to understanding the cosmic matter cycle of our Universe, since they enrich the cosmos with newly formed elements. However, the physical processes that trigger mass loss in their atmospheres are still not fully understood, and remain one of the key questions in stellar astrophysics. We use a new method to study the extended atmospheres of these cold stars, exploring the effect of a stellar wind for both a simple radiative equilibrium model and a semi-empirical model that accounts for a chromospheric temperature structure. We then can compute the intensities, fluxes and visibilities matching the observations for the different instruments at the Very Large Telescope Interferometer (VLTI). Specifically, when comparing with the atmospheric structure of HD 95687 based on published VLTI/AMBER data, we find that our model can accurately match these observations in the Kband, showing the enormous potential of this methodology to reproduce extended atmospheres of RSGs.

astro-ph.SR↗

arXiv2022

Explosion Imminent: the appearance of Red Supergiants at the point of core-collapse

From the early radiation of type II-P supernovae (SNe), it has been claimed that the majority of their red supergiant (RSG) progenitors are enshrouded by large amounts of circumstellar material (CSM) at the point of explosion. The inferred density of this CSM is orders of magnitude above that seen around RSGs in the field, and is therefore indicative of a short phase of elevated mass-loss prior to explosion. It is not known over what timescale this material gets there: is it formed over several decades by a `superwind' with mass-loss rate $\dot{M} \sim10^{-3}\,{\rm M_\odot\,yr^{-1}}$; or is it formed in less than a year by a brief `outburst' with $\dot{M}\sim10^{-1}\,{\rm M_\odot\,yr^{-1}}$? In this paper, we simulate spectra for RSGs undergoing such mass-loss events, and demonstrate that in either scenario the CSM suppresses the optical flux by over a factor of 100, and that of the near-IR by a factor of 10. We argue that the `superwind' model can be excluded as it causes the progenitor to be heavily obscured for decades before explosion, and is strongly at odds with observations of II-P progenitors taken within 10 years of core-collapse. Instead, our results favour abrupt outbursts $<$1 year before explosion as the explanation for the early optical radiation of II-P SNe. We therefore predict that RSGs will undergo dramatic photometric variability in the optical and infrared in the weeks-to-months before core-collapse.

astro-ph.SR↗

arXiv2021

Red Supergiants in M31: The Humphreys-Davidson limit at high metallicity

The empirical upper limit to Red Supergiant (RSG) luminosity, known as the Humphreys-Davidson (HD) limit, has been commonly explained as being caused by the stripping of stellar envelopes by metallicity-dependent, line-driven winds. As such, the theoretical expectation is that the HD limit should be higher at lower metallicity, where weaker mass-loss rates mean that higher initial masses are required for an envelope to be stripped. In this paper, we test this prediction by measuring the luminosity function of RSGs in M31 and comparing to those in the LMC and SMC. We find that $\log (L_{\rm max}/L_{\odot}) = 5.53 \pm 0.03$ in M31 (Z $\gtrsim$ Z$_{\odot}$), consistent with the limit found for both the LMC (Z $\sim$ 0.5 Z$_{\odot}$) and SMC (Z $\sim$ 0.25 Z$_{\odot}$), while the RSG luminosity distributions in these 3 galaxies are consistent to within 1$\sigma$. We therefore find no evidence for a metallicity dependence on both the HD limit and the RSG luminosity function, and conclude that line-driven winds on the main sequence are not the cause of the HD limit.

astro-ph.GA↗

arXiv2021

The impact of winds on the spectral appearance of Red Supergiants

The rate at which mass is lost during the Red Supergiant evolutionary stage may strongly influence how the star appears. Though there have been many studies discussing how RSGs appear in the mid and far infrared (IR) as a function of their mass-loss rate, to date there have been no such investigations at optical and near-IR wavelengths. In a preliminary study we construct model atmospheres for RSGs which include a wind, and use these models to compute synthetic spectra from the optical to the mid-infrared. The inclusion of a wind has two important effects. Firstly, higher mass-loss rates result in stronger absorption in the TiO bands, causing the star to appear as a later spectral type despite its effective temperature remaining constant. This explains the observed relation between spectral type, evolutionary stage and mid-IR excess, as well as the mismatch between temperatures derived from the optical and infrared. Secondly, the wind mimics many observed characteristics of a `MOLsphere', potentially providing an explanation for the extended molecular zone inferred to exist around nearby RSGs. Thirdly, we show that wind fluctuations can explain the spectral variability of Betelgeuse during its recent dimming, without the need for dust.

astro-ph.SR↗

arXiv2021

The impact of realistic red supergiant mass-loss on stellar evolution

Accurate mass-loss rates are essential for meaningful stellar evolutionary models. For massive single stars with initial masses between 8 - 30\msun the implementation of cool supergiant mass loss in stellar models strongly affects the resulting evolution, and the most commonly used prescription for these cool-star phases is that of de Jager. Recently, we published a new \mdot\ prescription calibrated to RSGs with initial masses between 10 - 25\msun, which unlike previous prescriptions does not over estimate \mdot\ for the most massive stars. Here, we carry out a comparative study to the MESA-MIST models, in which we test the effect of altering mass-loss by recomputing the evolution of stars with masses 12-27\msun\ with the new \mdot-prescription implemented. We show that while the evolutionary tracks in the HR diagram of the stars do not change appreciably, the mass of the H-rich envelope at core-collapse is drastically increased compared to models using the de Jager prescription. This increased envelope mass would have a strong impact on the Type II-P SN lightcurve, and would not allow stars under 30\msun\ to evolve back to the blue and explode as H-poor SN. We also predict that the amount of H-envelope around single stars at explosion should be correlated with initial mass, and we discuss the prospects of using this as a method of determining progenitor masses from supernova light curves.

astro-ph.SR↗

arXiv2021

The temperatures of red supergiants in low metallicity environments

The temperatures of red supergiants (RSGs) are expected to depend on metallicity (Z) in such a way that lower-Z RSGs are warmer. In this work, we investigate the Z-dependence of the Hayashi limit by analysing RSGs in the low-Z galaxy Wolf-Lundmark-Mellote (WLM), and compare with the RSGs in the higher-Z environments of the Small Magellanic Cloud (SMC) and Large Magellanic Cloud (LMC). We determine the effective temperature ($T_{\textrm{eff}}$) of each star by fitting their spectral energy distributions, as observed by VLT+SHOOTER, with MARCS model atmospheres. We find average temperatures of $T_{\textrm{eff}_{\textrm{WLM}}}=4400\pm202$ K, $T_{\textrm{eff}_{\textrm{SMC}}}=4130\pm103$ K, and $T_{\textrm{eff}_{\textrm{LMC}}}=4140\pm148$ K. From population synthesis analysis, we find that although the Geneva evolutionary models reproduce this trend qualitatively, the RSGs in these models are systematically too cool. We speculate that our results can be explained by the inapplicability of the standard solar mixing length to RSGs.

astro-ph.GA↗

arXiv2021

The Age of Westerlund 1 Revisited

The cluster Westerlund~1 (Wd1) is host to a large variety of post main-sequence (MS) massive stars. The simultaneous presence of these stars can only be explained by stellar models if the cluster has a finely-tuned age of 4-5Myr, with several published studies independently claiming ages within this range. At this age, stellar models predict that the cool supergiants (CSGs) should have luminosities of $\log(L/L_\odot) \approx 5.5$, close to the empirical luminosity limit. Here, we test that prediction using archival data and new photometry from SOFIA to estimate bolometric luminosities for the CSGs. We find that these stars are on average 0.4dex too faint to be 5Myr old, regardless of which stellar evolution model is used, and instead are indicative of a much older age of $10.4^{+1.3}_{-1.2}$Myr. We argue that neither systematic uncertainties in the extinction law nor stellar variability can explain this discrepancy. In reviewing various independent age estimates of Wd1 in the literature, we firstly show that those based on stellar diversity are unreliable. Secondly, we re-analyse Wd1's pre-MS stars employing the Damineli extinction law, finding an age of $7.2^{+1.1}_{-2.3}$Myr; older than that of previous studies, but which is vulnerable to systematic errors that could push the age close to 10Myr. However, there remains significant tension between the CSG age and that inferred from the eclipsing binary W13. We conclude that stellar evolution models cannot explain Wd1 under the single age paradigm. Instead, we propose that the stars in the Wd1 region formed over a period of several Myr.

astro-ph.SR↗

arXiv2020

On `On the Red Supergiant Problem': a rebuttal, and a consensus on the upper mass cutoff for II-P progenitors

The `Red Supergiant Problem' describes the claim that the brightest Red Supergiant (RSG) progenitors to type II-P supernovae are significantly fainter than RSGs in the field. This mismatch has been interpreted by several authors as being a manifestation of the mass threshold for the production of black holes (BHs), such that stars with initial masses above a cutoff of $M_{\rm hi}=17$M$_\odot$ and below 25$M_\odot$ will die as RSGs, but with no visible SN explosion as the BH is formed. However, we have previously cautioned that this cutoff is more likely to be higher and has large uncertainties ($M_{\rm hi}=19^{+4}_{-2}M_{\odot}$), meaning that the statistical significance of the RSG Problem is less than $2σ$. Recently, Kochanek (2020) has claimed that our work is statistically flawed, and with their analysis has argued that the upper mass cutoff is as low as $M_{\rm hi} = 15.7 \pm 0.8M_\odot$, giving the RSG Problem a significance of $>10σ$. In this letter, we show that Kochanek's low cutoff is caused by a statistical misinterpretation, and the associated fit to the progenitor mass spectrum can be ruled out at the 99.6\% confidence level. Once this problem is remedied, Kochanek's best fit becomes $M_{\rm hi} =19^{+4}_{-2}M_{\odot}$, in excellent agreement with our work. Finally, we argue that, in the search for a RSG `vanishing' as it collapses directly to a BH, any such survey would have to operate for decades before the absence of any such detection became statistically significant.

astro-ph.SR↗

arXiv2020

A new mass-loss rate prescription for red supergiants

Evolutionary models have shown the substantial effect that strong mass-loss rates ($\dot{M}$) can have on the fate of massive stars. Red supergiant (RSG) mass-loss is poorly understood theoretically, and so stellar models rely on purely empirical \mdot-luminosity relations to calculate evolution. Empirical prescriptions usually scale with luminosity and effective temperature, but $\dot{M}$ should also depend on the current mass and hence the surface gravity of the star, yielding more than one possible $\dot{M}$ for the same position on the Hertzsprung-Russell diagram. One can solve this degeneracy by measuring $\dot{M}$ for RSGs that reside in clusters, where age and initial mass ($M_{\rm init}$) are known. In this paper we derive $\dot{M}$ values and luminosities for RSGs in two clusters, NGC 2004 and RSGC1. Using newly derived $M_{\rm init}$ measurements, we combine the results with those of clusters with a range of ages and derive an $M_{\rm init}$-dependent $\dot{M}$-prescription. When comparing this new prescription to the treatment of mass-loss currently implemented in evolutionary models, we find models drastically over-predict the total mass-loss, by up to a factor of 20. Importantly, the most massive RSGs experience the largest downward revision in their mass-loss rates, drastically changing the impact of wind mass-loss on their evolution. Our results suggest that for most initial masses of RSG progenitors, quiescent mass-loss during the RSG phase is not effective at removing a significant fraction of the H-envelope prior to core-collapse, and we discuss the implications of this for stellar evolution and observations of SNe and SN progenitors.

astro-ph.SR↗

arXiv2020

The `Red Supergiant Problem': the upper luminosity boundary of type-II supernova progenitors

By comparing the properties of Red Supergiant (RSG) supernova progenitors to those of field RSGs, it has been claimed that there is an absence of progenitors with luminosities $L$ above $\log(L/L_\odot) > 5.2$. This is in tension with the empirical upper luminosity limit of RSGs at $\log(L/L_\odot) = 5.5$, a result known as the `Red Supergiant Problem'. This has been interpreted as evidence for an upper mass threshold for the formation of black-holes. In this paper, we compare the observed luminosities of RSG SN progenitors with the observed RSG $L$-distribution in the Magellanic Clouds. Our results indicate that the absence of bright SN II-P/L progenitors in the current sample can be explained at least in part by the steepness of the $L$-distribution and a small sample size, and that the statistical significance of the Red Supergiant Problem is between 1-2$σ$ . Secondly, we model the luminosity distribution of II-P/L progenitors as a simple power-law with an upper and lower cutoff, and find an upper luminosity limit of $\log(L_{\rm hi}/L_\odot) = 5.20^{+0.17}_{-0.11}$ (68\% confidence), though this increases to $\sim$5.3 if one fixes the power-law slope to be that expected from theoretical arguments. Again, the results point to the significance of the RSG Problem being within $\sim 2 σ$. Under the assumption that all progenitors are the result of single-star evolution, this corresponds to an upper mass limit for the parent distribution of $M_{\rm hi} = 19.2{\rm M_\odot}$, $\pm1.3 {\rm M_\odot (systematic)}$, $^{+4.5}_{-2.3} {\rm M_\odot}$ (random) (68\% confidence limits).

astro-ph.SR↗

arXiv2019

The distances to star-clusters hosting Red Supergiants: $χ$ Per, NGC 7419, and Westerlund 1

Galactic, young massive star clusters are approximately coeval aggregates of stars, close enough to resolve the individual stars, massive enough to have produced large numbers of massive stars, and young enough for these stars to be in a pre-supernova state. As such these objects represent powerful natural laboratories in which to study the evolution of massive stars. To be used in this way, it is crucial that accurate and precise distances are known, since this affects both the inferred luminosities of the cluster members and the age estimate for the cluster itself. Here we present distance estimates for three star clusters rich in Red Supergiants ($χ$ Per, NGC 7419 and Westerlund 1) based on their average astrometric parallaxes $\barπ$ in Gaia Data Release 2, where the measurement of $\barπ$ is obtained from a proper-motion screened sample of spectroscopically-confirmed cluster members. We determine distances of $d=2.25^{+0.16}_{-0.14}$kpc, $d=3.00^{+0.35}_{-0.29}$kpc, and $d=3.87^{+0.95}_{-0.64}$kpc for the three clusters respectively. We find that the dominant source of error is that in Gaia's zero-point parallax offset $π_{\rm ZP}$, and we argue that more precise distances cannot be determined without an improved characterization of this quantity.

astro-ph.SR↗

arXiv2019

Discrepancies in the ages of young star clusters; evidence for mergers?

There is growing evidence that star clusters can no longer be considered simple stellar populations (SSPs). Intermediate and old age clusters are often found to have extended main sequence turn-offs (eMSTOs) which are difficult to explain with single age isochrones, an effect attributed to rotation. In this paper, we provide the first characterisation of this effect in young (<20Myr) clusters. We determine ages for 4 young massive clusters (2 LMC, 2 Galactic) by three different methods: using the brightest single turn-off (TO) star; using the luminosity function (LF) of the TO; and by using the lowest $L_{\rm bol}$ red supergiant (RSG). The age found using the cluster TO is consistently younger than the age found using the lowest RSG $L_{\rm bol}$. Under the assumption that the lowest luminosity RSG age is the `true' age, we argue that the eMSTOs of these clusters cannot be explained solely by rotation or unresolved binaries. We speculate that the most luminous stars above the TO are massive blue straggler stars formed via binary interaction, either as mass gainers or merger products. Therefore, using the cluster TO method to infer ages and initial masses of post-main sequence stars such as Wolf-Rayet stars, luminous blue variables and RSGs, will result in ages inferred being too young and masses too high.

astro-ph.SR↗

arXiv2018

The surface abundances of Red Supergiants at core-collapse

In the first weeks-to-months of a type II-P supernova (SN), the spectrum formation region is within the hydrogen-rich envelope of the exploding star. Optical spectra taken within a few days of the SN explosion, when the photosphere is hot, show features of ionised carbon, nitrogen and oxygen, as well as hydrogen and helium. Quantitative analysis of this very early phase may therefore constrain the chemical abundances of the stellar envelope at the point of core-collapse. Using existing and new evolutionary calculations for Red Supergiants (RSGs), we show that the predictions for the terminal surface [C/N] ratio is correlated with the initial mass of the progenitor star. Specifically, a star with an initial mass above 20M$_{\odot}$ exploding in the RSG phase should have an unequivocal signal of a low [C/N] abundance. Furthermore, we show that the model predictions are relatively insensitive to uncertainties in the treatment of convective mixing. Although there is a dependence on initial rotation, this can be dealt with in a probabilistic sense by convolving the model predictions with the observed distribution of stellar rotation rates. Using numerical experiments, we present a strategy for using very early-time spectroscopy to determine the upper limit to the progenitor mass distribution for type II-P SNe.

astro-ph.SR↗