Neutron star atmosphere composition: the quiescent, low-mass X-ray binary in the globular cluster M28

Abstract

Using deep Chandra observations of the globular cluster M28, we study the quiescent X‐ray emission of a neutron star in a low‐mass X‐ray binary in order to constrain the chemical composition of the neutron star atmosphere and the equation of state of dense matter. We fit the spectrum with different neutron star atmosphere models composed of hydrogen, helium or carbon. The parameter values obtained with the carbon model are unphysical and such a model can be ruled out. Hydrogen and helium models give realistic parameter values for a neutron star, and the derived mass and radius are clearly distinct depending on the composition of the atmosphere. The hydrogen model gives masses/radii consistent with the canonical values of 1.4 M⊙ and 10 km, and would allow for the presence of exotic matter inside neutron stars. On the other hand, the helium model provides solutions with higher masses/radii, consistent with the stiffest equations of state. Measurements of neutron star masses/radii by spectral fitting should consider the possibility of heavier element atmospheres, which produce larger masses/radii for the same data, unless the composition of the accretor is known independently.

Limits on thermal variations in a dozen quiescent neutron stars over a decade

“…The X-ray spectra of helium atmospheres are similar to hydrogen atmospheres, but slightly harder (Romani 1987;Ho & Heinke 2009). An observed spectrum will thus give different parameter values for He vs. H atmospheres, and several papers have considered the importance of atmosphere composition for attempts to constrain the NS mass and radius (Servillat et al 2012;Catuneanu et al 2013;Lattimer & Steiner 2014;Heinke et al 2014). For relatively low-count spectra such as CX9, which parameter is different from other NS qLMXBs-e.g., atmospheric composition (H or He), vs. NH -may not be possible to differentiate.…”

“…However, low-level accretion at sufficiently high rates (corresponding to L ∼ 10 33 erg/s) could maintain metals in the atmosphere at abundances sufficient to soften the spectrum, leading to an overestimate of R (Rutledge et al 2002b). The question of the origin of the thermal component in qLMXBs is of intense interest, since fitting the thermal spectra of qLMXBs in globular clusters is often used to measure the radius of NSs and place constraints on the dense matter equation of state (Rutledge et al 2002a;Heinke et al 2006a;Webb & Barret 2007;Steiner, Lattimer & Brown 2010, 2013Guillot, Rutledge & Brown 2011;Servillat et al 2012;Guillot et al 2013;Guillot & Rutledge 2014;Lattimer & Steiner 2014;Heinke et al 2014;Ozel et al 2015; see also Miller et al 2013). Confirmation that thermal spectra of qLMXBs (especially those without powerlaw components) are not produced by accretion would eliminate the possibility that metals remain in the atmosphere, and thus eliminate a systematic uncertainty in this method of constraining the NS radius.…”

“…Nowak, Heinz & Begelman 2002;Heinke et al 2003c;Wijnands et al 2003). A few qLMXBs without detectable powerlaw components have repeated deep observations with Chandra, permitting sensitive measurements for variability; these include (with 90% confidence variability limits) 47 Tuc X7, ∆T /T <1.0% over 2 years (Heinke et al 2006a); NGC 6397 U18, ∆T /T <1.4% over 10 years (Guillot, Rutledge & Brown 2011;Heinke et al 2014); M28 source 26 (no variation found, Servillat et al 2012) and ω Cen, ∆T /T <2.1% over 12 years .…”

Atmospheres and radiating surfaces of neutron stars

“…In most cases it is stipulated by the fact that a joint evaluation of R and M (let alone R, M , and D) from the currently available thermal spectra leaves too large uncertainties and almost does not constrain M and D. A comparison of the results obtained for the same objects with different assumptions on D values readily shows that a choice of D can drastically affect the R estimate. In addition, the estimate of R is strongly affected by assumptions on the photosphere composition, as one can see, for example, from a comparison of the results obtained by assuming hydrogen and helium photospheres for the qLMXBs in globular clusters M28 [243] and M13 [244].…”

“…However, the estimates that were obtained in [242] by an analogous analysis for five qLMXBs in globular clusters, although widely scattered, generally better agree with soft EOSs. In [243,244], thermal spectra of two qLMXBs were analyzed using hydrogen and helium atmosphere models. It turned out that the former model leads to low estimates of M and R, compatible with the soft EOSs, while the latter yields high values, which require a stiff EOS of superdense matter.…”