A solution to the tension of burning on neutron stars and nuclear physics

Abstract

When neutron stars accrete matter from a companion star, this matter forms a disc around them and eventually falls on their surface. Here, the fuel can ignite into bright flashes called Type I bursts. Theoretical calculations based on state-of-the-art nuclear reactions are able to explain many features of the bursts. However, models predict that the bursts should cease at high accretion rates, whereas in many sources they disappear at much lower rates. Moreover, their recurrence times also show strong discrepancies with predictions. Although various solutions have been proposed, none can account for all the observational constraints. Here we describe a new model that explains all the contradictory behaviours within a single picture. We are able to reconstruct the conditions on the star surface that determine the burst properties by comparing data to new simulations. We find strong evidence that the physical mechanism driving the burst behaviour is the structure of the accretion disc in the regions closest to the star. This connection reconciles the puzzling burst phenomenology with nuclear physics and also opens a new window on the study of accretion processes around compact objects.

Figures (6)

• Figure 1: Burst recurrence time $t_{\rm{rec}}$ (left) and $\alpha$ (right) for the database models. The results are displayed as a function of weight $w$ and accretion rate $\dot{m}_{\rm{loc}} / \dot{m}_{\rm{Edd}}$. The missing points are due to burst quenching.
• Figure 2: Results of the fit of the burst database to the observations for two of the five sources considered in this paper. Top two rows: Each quantity is plotted as a function of $\left<\dot{m}\right>_{\rm{obs}}$ in units of $\dot{m}_{\rm{Edd}}$. Bottom two rows: Each function is plotted as a function of $S_z$, which tracks the position on the spectral color-color diagram for each source. The vertical (black, dashed) lines in each panel indicate the approximate position of the turnover in $\left<\dot{m}\right>_{\rm{obs}}$ and $S_z$. Left column: fit local $\dot{m}_{\rm{loc}}$ in units of $\dot{m}_{\rm{Edd}}$. The black dotted line shows the values expected if $\dot{m}_{\rm{loc}}$ and $\left<\dot{m}\right>_{\rm{obs}}$ coincided, while the orange line assumes $\dot{m}_{\rm{loc}} = \bar{f} \left<\dot{m}\right>_{\rm{obs}}$, with $\bar{f}$ set by the average value of the ratio during the regime R1. Central column:$f = \dot{m}_{\rm{loc}} / \left<\dot{m}\right>_{\rm{obs}}$. The horizontal orange line indicates the average value during regime R1. The black and orange lines in the left and middle columns correspond to each other. Right column: the purity $w$ as a function of $\left<\dot{m}\right>_{\rm{obs}}$. The horizontal orange line indicates the average value during regime R1. The values of the turnover $\dot{m}_{\rm{to}}$ and $S_{z,\, \rm{to}}$ and the averages are reported in Tab. \ref{['tab:res']}.
• Figure 3: Schematic of the solution proposed in this paper. Top panel: evolution of the burst recurrence time vs. mass accretion rate. The different background colors correspond to the spectral states: hard state (blue), intermediate transition (orange), and soft state (yellow). The lower three rows correspond to each of these states. Left column: position of the source in the spectral color-color diagram. Central column: accretion configuration. We indicate the neutron star, the thin disk, the hot inner flow, and the spreading layer. Right column: corresponding distribution of the matter on the surface, through the parameter $f(\theta)$. First row: R1, the thin disk is truncated, and an inner flow reaches the star. The spreading layer (indicated in light gray) does not develop. The burst behavior is determined by a band around the equator set by the hot inner flow (in red, up to the latitude of the hot inner flow $\theta_{\rm{H}}$). The recurrence time decreases with $\dot{M}$ as expected. The efficiency (not shown) is high, indicating that no pollution is present. Second row: the inner flow has almost disappeared, and the spreading layer emerges (up to the latitude $\theta_*$, now on the order of $\theta_{\rm{H}}$). The burning around the equator is transitioning to stable burning (red and hatched). After the turnover the burning is determined by the regions outside the stable zone (brown). The burst recurrence time turns over and begins to increase, and the efficiency decreases. Third row: R2, the accretion is now dominated by the spreading layer which is growing closer to the pole as a function of accretion rate. The lower latitudes are stable, and the burning is only determined by the unstable region closer to the pole. The burst recurrence time keeps increasing and the efficiency decreasing due to the large amount of ashes present. Angles and sizes are not to scale. The rectangles on the color-color diagrams have dashed boundaries to indicate that the regions are indicative and the precise limits vary from source to source.
• Figure 21: Results of the fit of the burst database to the observations of the 5 sources considered in this paper. Data are displayed as a function of $\left<\dot{m}\right>_{\rm{obs}}/\dot{m}_{\rm{Edd}}$. See Fig. \ref{['fig:res']} for details and Tabs. \ref{['tab:fitres']} and \ref{['tab:res']}.
• Figure 22: Same as Fig. \ref{['fig:resM']}, but displayed as a function of $S_z$.