Unexpected late-time temperature increase observed in the two neutron star crust-cooling sources XTE J1701−462 and EXO 0748−676

Abstract

Transient low-mass X-ray binaries (LMXBs) that host neutron stars (NSs) provide excellent laboratories for probing the dense matter physics present in NS crusts. During accretion outbursts in LMXBs, exothermic reactions may heat the NS crust, disrupting the crust-core equilibrium. When the outburst ceases, the crust cools to restore thermal equilibrium with the core. Monitoring this cooling evolution allows us to probe the dense matter physics in the crust. Properties of the deeper crustal layers can be probed at later times after the end of the outburst. We report on the unexpected late-time temperature evolution (≳2000 days after the end of their outbursts) of two NSs in LMXBs, XTE J1701−462 and EXO 0748−676. Although both these sources exhibited very different outbursts (in terms of duration and the average accretion rate), they exhibit an unusually steep decay of ∼7 eV in the observed effective temperature (occurring in a time span of ∼700 days) around ∼2000 days after the end of their outbursts. Furthermore, they both showed an even more unexpected rise of ∼3 eV in temperature (over a time period of ∼500–2000 days) after this steep decay. This rise was significant at the 2.4σ and 8.5σ level for XTE J1701−462 and EXO 0748−676, respectively. The physical explanation for such behaviour is unknown and cannot be straightforwardly be explained within the cooling hypothesis. In addition, this observed evolution cannot be well explained by low-level accretion either without invoking many assumptions. We investigate the potential pathways in the theoretical heating and cooling models that could reproduce this unusual behaviour, which so far has been observed in two crust-cooling sources. Such a temperature increase has not been observed in the other NS crust-cooling sources at similarly late times, although it cannot be excluded that this might be a result of the inadequate sampling obtained at such late times.

Crust structure and thermal evolution of neutron stars in soft X-ray transients

“…Unlike the case of MXB 1659−29, these temperature variations are difficult to explain by an increased hydrogen column density on the line of sight. Among different tentative explanations discussed by Parikh et al (2020), the most viable one appears to be convection, driven by chemical separation at the ocean-crust boundary, which was previously studied by Medin & Cumming (2014. The latter authors predicted dips of the effective temperature at ∼ 5 -6 years of crust cooling, similar to those observed for XTE J1701−462 and EXO 0748−676.…”

“…Rapid late-time variations of soft X-ray flux during the postoutburst neutron star crust cooling are not unique to MXB 1659−29. Recently, Parikh et al (2020) reported an unusually steep decay of ∼ 7 eV followed by a rise of ∼ 3 eV in the observed effective temperature during the crust cooling of two other SXTs, XTE J1701−462 and EXO 0748−676, around ∼ 5.5 years after the end of their outbursts. Unlike the case of MXB 1659−29, these temperature variations are difficult to explain by an increased hydrogen column density on the line of sight.…”

UV and X-ray observations of the neutron star LMXB EXO 0748-676 in its quiescent state

“…Observations of the source in quiescence indicate that the crust of the NS was heated up during outburst and exhibited cooling in quiescence (e.g., Degenaar et al 2009aDegenaar et al , 2011aDegenaar et al , 2014Díaz Trigo et al 2011;Cheng et al 2017). However, more recent quiescent observations, obtained >2000 d after the end of its accretion outburst, indicate a rise in the observed effective NS temperature (Parikh et al 2020). This rise is highly surprising for a crust-cooling source as it is expected to exhibit a continuous decay in the temperature rather than a rise.…”

“…Since EXO 0748-676 hosts a cooling NS crust (e.g., Parikh et al 2020, see the references therein for more details), we fitted the XRT spectrum with (a NS atmosphere model; Heinke et al 2006). The source distance was set to 𝐷 = 7.4 kpc (e.g., Galloway et al 2008) and the NS mass and radius to 𝑀 = 1.6 M and 𝑅 = 12 km, respectively.…”

“…The source distance was set to 𝐷 = 7.4 kpc (e.g., Galloway et al 2008) and the NS mass and radius to 𝑀 = 1.6 M and 𝑅 = 12 km, respectively. Given the limited statistics of the XRT data, we fixed several of the parameters to the values obtained from fitting high-quality quiescent Chandra and XMM-Newton spectra of EXO 0748-676 (see Parikh et al 2020, for details, where the same parameters have been fixed). An additional power-law component was needed to fit some high-quality quiescent spectra of this source (Parikh et al 2020).…”