5 Thermonuclear X-ray Bursts
225
with Sect. 5.1.2). At present this scenario remains highly speculative, and it is
computationally challenging to test with multi-dimensional models (Sect. 5.8.3).
Certainly, the balance in pressure resulting from the decrease in effective gravity at
the equator gives a larger effective accretion rate per unit area there, but not sufficient
to stabilise the burning [166]. Resolving this puzzle will likely require a substantial
improvement in our understanding of the global burst properties and perhaps also
the influence of the accretion flow.
5.3 The Burst Spectral Energy Distribution
Most of the energy from nuclear burning is generated at least several meters below
the neutron star surface, and processed before exiting the star from a cm-thick
photosphere. The burst radiation is, therefore, thermalized and leaves the burning
layer as a blackbody spectrum. Scattering off electrons and ions in the photosphere,
however, modifies the continuum spectrum and may introduce absorption lines and
edges. Those departures from a pure blackbody spectrum offer opportunities to
measure the neutron star mass and radius. In Sect. 5.4 we further discuss how some
fraction of the burst spectrum may be reprocessed by the accretion environment.
5.3.1 The Continuum Spectrum
The majority of observed burst spectra are well described by a Planck (blackbody)
distribution [180]. The rise of the burst flux, set by a combination of the time
scales of nuclear burning and flame spreading, can be as fast as a millisecond, or
as long as several seconds. The measured blackbody temperature increases to a
maximum value of 2–3 keV. In the burst tail, the decrease of the temperature is
determined by the time scales of cooling and of waiting points in the rp-process
(Sect. 5.9), and follows that of the burst luminosity, typically lasting ∼10–100 s.
This cooling of the neutron star atmosphere in the burst tail is often cited as prime
evidence of a thermonuclear flash (Type I burst) as opposed to an accretion flare
(Type II burst; [6, 115]). A small subset of thermonuclear events exhibit only weak
evidence of cooling. In an analysis of the remarkable burst source and 11 Hz pulsar
IGR J17480−2446, it has been shown that cooling may not be detectable when
the ratio of peak burst flux to persistent (accretion) flux is low, and the bursts
correspondingly have a low peak temperature [117].
Because of the evolution of the spectral parameters, time resolved spectroscopy
has to be performed, with each spectrum being collected in a short time interval.
The short duration over which each spectrum is accumulated in typical bursts
strongly limits the photon counts and the detail present in a single spectrum. Only
when considering exceptionally long bursts [172] or large numbers of bursts [203]
can deviations from a blackbody shape be detected. Some deviations are due to
225
with Sect. 5.1.2). At present this scenario remains highly speculative, and it is
computationally challenging to test with multi-dimensional models (Sect. 5.8.3).
Certainly, the balance in pressure resulting from the decrease in effective gravity at
the equator gives a larger effective accretion rate per unit area there, but not sufficient
to stabilise the burning [166]. Resolving this puzzle will likely require a substantial
improvement in our understanding of the global burst properties and perhaps also
the influence of the accretion flow.
5.3 The Burst Spectral Energy Distribution
Most of the energy from nuclear burning is generated at least several meters below
the neutron star surface, and processed before exiting the star from a cm-thick
photosphere. The burst radiation is, therefore, thermalized and leaves the burning
layer as a blackbody spectrum. Scattering off electrons and ions in the photosphere,
however, modifies the continuum spectrum and may introduce absorption lines and
edges. Those departures from a pure blackbody spectrum offer opportunities to
measure the neutron star mass and radius. In Sect. 5.4 we further discuss how some
fraction of the burst spectrum may be reprocessed by the accretion environment.
5.3.1 The Continuum Spectrum
The majority of observed burst spectra are well described by a Planck (blackbody)
distribution [180]. The rise of the burst flux, set by a combination of the time
scales of nuclear burning and flame spreading, can be as fast as a millisecond, or
as long as several seconds. The measured blackbody temperature increases to a
maximum value of 2–3 keV. In the burst tail, the decrease of the temperature is
determined by the time scales of cooling and of waiting points in the rp-process
(Sect. 5.9), and follows that of the burst luminosity, typically lasting ∼10–100 s.
This cooling of the neutron star atmosphere in the burst tail is often cited as prime
evidence of a thermonuclear flash (Type I burst) as opposed to an accretion flare
(Type II burst; [6, 115]). A small subset of thermonuclear events exhibit only weak
evidence of cooling. In an analysis of the remarkable burst source and 11 Hz pulsar
IGR J17480−2446, it has been shown that cooling may not be detectable when
the ratio of peak burst flux to persistent (accretion) flux is low, and the bursts
correspondingly have a low peak temperature [117].
Because of the evolution of the spectral parameters, time resolved spectroscopy
has to be performed, with each spectrum being collected in a short time interval.
The short duration over which each spectrum is accumulated in typical bursts
strongly limits the photon counts and the detail present in a single spectrum. Only
when considering exceptionally long bursts [172] or large numbers of bursts [203]
can deviations from a blackbody shape be detected. Some deviations are due to