192
A. Patruno and A. L. Watts
helium, for a source where the accreted fuel is a mix of hydrogen and helium [98].
The only other sources that are thought to have pure helium bursts are ultra-compact
sources where the donor star is too small for there to be significant hydrogen in the
accreted material. Pure helium bursting is thought to occur in mixed fuel sources
only in a very narrow range of accretion rates where hydrogen can burn stably to
form helium before the helium ignites unstably [96].
Bursting and Intermittent Accretion-Powered Pulsations Two of the intermittent
AMXPs, HETE J1900.1-2455 and SAX J1748.9-2021, show intermittent pulsations
with variable amplitudes at times when the source is also bursting [2, 101]. There
does not appear to be any clear-cut causal relationship between the two phenomena
(and the third intermittent AMXP, Aql X-1, does not show bursts at the time of its
one intermittent pulsation episode, [44]). Nonetheless there has been speculation in
the literature as to whether the two might be related. Radiation from the burst might
perhaps lead to a temporary disruption of the inner edge of the disk, perturbing the
flow of material onto the magnetic poles and moving the accretion footprint to a
location more favorable for the formation of pulsations [179].
Burst Oscillations Where the AMXPs have had the biggest impact is in our
understanding of burst oscillations (see [349] for a more comprehensive review
of this topic). Burst oscillations were first observed in 1996, from the non-pulsar
4U 1728-34 [322]. The oscillations manifested as a coherent signal at about
363 Hz, with an upwards drift in frequency of ∼1 Hz as the bursts progressed,
towards an asymptotic maximum in the tails. The fact that the same frequency was
seen in multiple bursts from the same source, and the stability of the asymptotic
maximum frequency, suggested a link to a stable clock such as the stellar spin
rate [323]. Firm identification of burst oscillation frequency with the spin frequency
however had to wait until the first robust detection of burst oscillations from an
AMXP [51]. Although the two frequencies appear to be very close (within a
few Hz), however, the separation does differ markedly for different AMXPs. The
AMXP XTE J1814-338 and the mildly recycled pulsar IGR J17480-2466 have burst
oscillation frequencies that agree with the spin frequency very closely (within the
error bars, ∼10 −8 Hz for XTE J1814-338 [324, 351, 352], ∼10 −4 Hz for IGR
J17480-2466 [46]). SAX J1808.4-3658 and IGR J17511-3057, by contrast, have
burst oscillations that show a rapid increase in frequency in the burst rise, slightly
overshooting the spin frequency, and then stabilizing to within ∼10 mHz of the spin
frequency in the tails [4, 51]. The intermittent pulsars Aql X-1 and HETE J1900.12455 show burst oscillations with slow drifts throughout the bursts, more similar
to those of the non-pulsars, and in these cases the asymptotic maximum frequency
is ∼1 Hz below the spin frequency [44, 224, 354, 368]. Any model for the burst
oscillation mechanism must be able to explain this diversity in frequency drift and
the small variations in offset from the spin frequency.
There are interesting differences in the properties of the burst oscillations from
the pulsars as compared to those of the non-pulsars. For the non-pulsing bursters,
burst oscillations tend only to be seen when the source is in the soft (high accretion
rate) state [103]. The two intermittent pulsars conform to this rule: indeed HETE
Précédent

- 202/344

Suivant