Luminous Fast Blue Optical Transients as very massive star core-collapse events

TL;DR

The paper addresses the origin of LFBOTs by testing very massive star core-collapse as a progenitor channel. It couples BPASS-based population synthesis with a metallicity-dependent CSFH to derive BH formation rates and compares them with the LFBOT rate and host metallicities. Key findings show BHs with $M_{ m BH} \gtrsim 38$–$41\,M_\odot$ form at a rate compatible with the LFBOT rate, preferentially from progenitors with $Z<0.3\,Z_\odot$, and capable of producing dense CSM and long-lived accretion-disc emission. The scenario also points to connections with super-kilonovae and possible contributions to the $r$-process in galaxies, while noting uncertainties in late-stage mass loss and local metallicity conditions.

Abstract

Luminous Fast Blue Optical Transients (LFBOTs) are rare extragalactic events of unknown origin. Tidal disruptions of white dwarfs by intermediate mass black holes, mergers of black holes and Wolf-Rayet stars, and failed supernovae are among the suggestions. In this paper, we explore the viability of very massive star core-collapse events as the origin of LFBOTs. The appeal of such a model is that the formation of massive black holes via core collapse may yield observational signatures that can match the disparate lines of evidence that point towards both core-collapse and tidal disruption origins for LFBOTs. We explore the formation rate of massive black holes in population synthesis models, and compare the metallicities of their progenitors with the observed metallicities of LFBOT host galaxies. We further examine the composition, mass loss rates and fallback masses of these stars, placing them in the context of LFBOT observations. The formation rate of black holes with mass greater than ~30-40Msol is similar to the observed LFBOT rate. The stars producing these black holes are biased to low metallicity (Z<0.3Zsol), are H and He-poor and have dense circumstellar media. However, some LFBOTs have host galaxies with higher metallicities than predicted, and others have denser environments (plausibly due to late mass loss not captured in the models). We find that long-lived emission from an accretion disc (as implicated in the prototypical LFBOT AT2018cow) can plausibly be produced in these events. We conclude that (very) massive star core-collapse is a plausible explanation for LFBOTs. The preferred progenitors for LFBOTs in this scenario overlap with those predicted to produce super-kilonovae. We therefore suggest that LFBOTs are promising targets to search for super-kilonovae, and that they may contribute non-negligibly to the r-process enrichment of galaxies.

Figures (6)

• Figure 1: The mean ratio of the rate of black hole formation above a minimum mass M$_{\rm BH,min}$ to the rate of core-collapse supernovae at z$<0.4$. The black line shows this ratio as a function of M$_{\rm BH,min}$ assuming the metallicity-dependence cosmic star-formation history (CSFH) of 2006ApJ...638L..63L. The red and blue lines, bounding the grey shaded region, are the result of adopting a CSFH shifted up/down by 0.2 dex in 12$+$log(O/H) respectively. The dashed horizontal line is the estimated LFBOT volumetric rate 2023ApJ...949..120H. We therefore find that the predicted formation rate of BHs with masses greater than (38$^{+3}_{-5}$) M$_{\odot}$ is consistent with the LFBOT event rate.
• Figure 2: The cumulative distribution of LFBOT host galaxy metallicities is shown in orange (see Table \ref{['tab:host_galaxy']}). We sample the Z values, assuming Gaussian uncertainties, 100 times, producing the many realisations of the cumulative distribution shown. Cumulative distributions of our selected progenitor metallicities, weighted by the CSFH(Z) at z$<$0.4, are also shown. The black line is the $z<0.4$ mean metallicity distribution of star-forming galaxies 2006ApJ...638L..63L. The grey shaded region bounded by dashed lines is defined by shifting the distribution of 2006ApJ...638L..63L by $\pm$0.2 in 12+log$_{10}$(OH), covering the high and low Z extremes defined by 2019MNRAS.488.5300C.
• Figure 3: The leftover (i.e. ejected and/or accreted) mass fractions of hydrogen (X$_{\rm surf}$) and helium (Y$_{\rm surf}$) in the last time-step of the selected bpass models, with contributions from different metallicities determined by the mean metallicity spread at $z<0.4$2006ApJ...638L..63L. The colourbar indicates the number of events per 10$^{6}$ M$_{\odot}$ of star formation, with contributions from different metallicities following the fiducial case as shown in Figure \ref{['fig:Z']}, and model weightings as defined in Section \ref{['sec:rates']}.
• Figure 4: Remnant black hole mass versus the 'leftover' mass for the models selected when the fiducial CSFH is used (black lines in Figures \ref{['fig:Rbh']} and \ref{['fig:Z']}). These are the remnant and leftover (ejecta) masses when an core-collapse energy injection of 10$^{51}$ erg is assumed 2004MNRAS.353...87E, taking into account the mass gap from PISNe and PPISNe 2019ApJ...887...53F2023MNRAS.520.5724B. The histograms show the number events expected per 10$^{6}$ M$_{\odot}$ of star formation, as described in Figure \ref{['fig:XY']}.
• Figure 5: The wind density parameter for the selected models, normalised by a typical Wolf-Rayet mass-loss rate and wind speed. The circumstellar densities resulting from these values are lower than typical Wolf-Rayet stars and lower than observed in some LFBOTs, likely due to weaker winds as a result of the strong low-metallicity bias introduced by requiring such massive black holes. Typical LFBOT values are in the range 0.1--10. The $y$-axis indicates the number events per 10$^{6}$ M$_{\odot}$ of star formation, as described in Figure \ref{['fig:XY']}.