A self-consistent explanation of the MeV line in GRB 221009A unveils a dense circum-stellar medium

TL;DR

This work presents a self-consistent interpretation of a narrow ~10 MeV emission line in GRB 221009A as high-latitude emission from a thin shell where $e^+e^-$ pairs annihilate. The model links the line evolution to the shell radius $r$, bulk Lorentz factor $\Gamma$, and pair multiplicity $N_\pm$ via $L_\text{line}(t_obs)$ and $h\nu_\text{line}(t_obs)$, deriving stringent constraints that point to radiative-pair loading of a precursor-driven external shock at $r\sim10^{15}$–$4\times10^{16}$ cm with $\Gamma$ in the hundreds. Achieving the required pair abundance demands an extremely dense external medium ( $n_\text{ext} \sim 10^8$–$10^9$ cm$^{-3}$ or $A_\star\sim10^3$–$10^4$ cm$^{-1}$ ), consistent with a highly mass‑loaded circum-stellar medium around a massive-star progenitor, possibly LBV-like or in a common-envelope phase. The proposed illumination phase also naturally explains the abrupt LHAASO afterglow rise and has implications for non-thermal emission by the created pairs and the occurrence of similar lines in other GRBs with dense environments and long quiescent intervals.

Abstract

GRB~221009A has been the brightest gamma-ray burst (GRB) observed to date, and its afterglow has been characterised with unprecedented detail at TeV energies by LHAASO. Quite puzzlingly, it is also the most energetic GRB known. Among the riddles posed by this mysterious source, however, the sheer energetics are hardly the most intriguing: an unprecedented, narrow, luminous emission line at around 10 MeV has been uncovered by a detailed spectral analysis of \textit{Fermi}/GBM data immediately following the brightest peak in the GRB prompt emission and the peak of the TeV afterglow. As noted in the discovery article, the temporal evolution of the line properties can be explained as being due to high-latitude emission from a geometrically thin, relativistically expanding shell where annihilation of a large number of electron-positron pairs took place. We show that this interpretation yields stringent constraints on the properties of such shell, that point to a process that happens at radii typical of external shocks. We then demonstrate that the shell could have been the blastwave associated with the GRB precursor, with the line arising after pair loading of such blastwave as it was illuminated by the bright and hard radiation of the GRB main event. The scenario, which also explains the abrupt initial rise of the LHAASO afterglow, requires the progenitor of the GRB to have been surrounded by a circum-stellar medium (CSM) extending out to a few $10^{15}\,\mathrm{cm}$, with a density $n_\mathrm{ext}\sim 10^{8}-10^{9}\,\mathrm{cm^{-3}}$ reminiscent of those found from studies of Type IIn supernovae. This provides a precious clue to the nature of the progenitor of this peculiar GRB, which could also be present in other bursts that feature a long quiescence followed by a bright emission episode with a hard spectrum.

Figures (8)

• Figure 1: GRB 221009A light curves and narrow line properties. Panel (a): background-subtracted count rate light curve of the fourth sodium-iodide (NaI) detector of Fermi/GBM (red histogram, sensitive to emission in the 8 keV -- 1 MeV band). The pink shading shows the one-sigma-equivalent uncertainty, assuming Poisson-distributed background counts. The gray-shaded area shows the 'Bad Time Interval' (BTI) during which the GBM detectors were affected by saturation. Panel (b): Gaussian line luminosity evolution. Error bars show the estimated Gaussian line luminosity according to the analysis of GBM data reported in Ravasio2024 (red stars), Axelsson2024 (Axelsson2024, orange circles), Zhang2024_observation, and that of the tentatively detected line in Konus-WIND data Frederiks2023. The BTI time range (only relevant to GBM) is reported with a lighter gray shading. Panel (c): same as panel (b), but for the line central photon energy. Panel (d): LHAASO WCDA count rate light curve LHAASO2023. The cyan shading shows the one-sigma-equivalent uncertainty, assuming Poisson-distributed background counts.
• Figure 2: HLE model fit results. From top to bottom: in the first (second) panel, purple lines show the evolution of $L_\mathrm{line}$ ($h\nu_\mathrm{line}$) from one hundred posterior samples of the HLE model fitted to the evolution inferred from the observations (shown by the error bars -- see text), including data in the BTI (grey shaded area). Cyan lines show the corresponding result when such data is excluded. The third panel shows the Fermi/GBM and LHAASO WCDA light curves as temporal references. In the last panel, the purple and cyan histograms show the posterior probability of the HLE start time $t_0$ inferred from the same two model fitting procedures.
• Figure 3: Sketch of the proposed scenario (not to scale). After emitting their gamma-rays (a), the precursor ejecta expand into the external medium and drive a blastwave (b). The main event ejecta emit their gamma-rays (c) and initially expand in the cavity left by the precursor. The main event gamma-ray front illuminates the precursor blastwave (d) leading to copious pair creation and radiative acceleration. Soon after, the main event ejecta collide with the pair-enriched blastwave (e). The mass-annihilation of the pairs ('pair bubble bursting') happens somewhere between the end of (d) and the end of (e). When the main-event-ejecta-driven shock reaches the external medium, it leads to a much more energetic blastwave (f), whose emission (including the component seen by LHAASO) rises steeply before settling to the usual, self-similar evolution. Panel (g) shows the evolution of the radii of the components described in the other panels as a function of the 'lab-frame' time $t$ in the progenitor rest frame. The inset shows the Fermi/GBM light curve, highlighting the connection between $t$ and the observer-frame time $t_\mathrm{obs}$.
• Figure 4: Fundamental block of the process that leads to $e^+e^-$ pair enrichment and acceleration within the blastwave. A packet of incident photons with energy $\varepsilon$ (orange) is inverse-Compton scattered at time $t_\mathrm{s}$ by an electron or positron (green) into a packet of photons with energy $\varepsilon_\mathrm{sc}$ (blue) that propagates with a cosine angle $\cos\theta=\mu$ with respect to the incident photon direction. Subsequent interactions of the scattered photons with incident photons of typical energy $\varepsilon\sim\varepsilon_\mathrm{th}=2/(1-\mu)\varepsilon_\mathrm{sc}$ (red) lead to pair production (two example such interactions happening at times $t_1$ and $t_2$ are shown in the sketch), gradually suppressing the scattered photon packet intensity by a factor $\exp\left[-c(t-t_\mathrm{s})/\lambda_{\gamma\gamma}(\varepsilon_\mathrm{sc},\mu)\right]$, where $\lambda_{\gamma\gamma}$ is the mean free path of the interaction.
• Figure 5: Example illumination phase evolution of pair enrichment and radiative acceleration. The three panels show the evolution of the pair multiplicity ($Z_\pm$, top panel), number of annihilated pairs ($N_\mathrm{\pm,ann}$, middle panel) and bulk Lorentz factor ($\Gamma_1$, bottom panel) as a function of time or equivalently of $E_\mathrm{\gamma f}$ (i.e. the total energy in the incident radiation until a given time) for a blastwave in an external medium with $k=2$, with density parameter $A_\star=4\times 10^3\,\mathrm{cm^{-3}}$ and with a precursor blastwave energy $E_\mathrm{pre}=10^{54}\,\mathrm{erg}$, corresponding to an illumination radius $r_0\approx 2.1\times 10^{15}\,\mathrm{cm}$. The grey shaded area corresponds to times after the main event ejecta collided with the blastwave. The vertical red band shows the range of $E_\mathrm{\gamma f}$ for which the number of annihilated pairs satisfy the constraints imposed by the HLE interpretation. The horizontal bands show the pair multiplicity (top panel), number of annihilated pairs (middle panel) and bulk Lorentz factor (bottom panel), respectively, required by the HLE interpretation if the annihilation takes place at $r_0$.