194
A. Patruno and A. L. Watts
accretion-powered pulsars have played a key role. Their primary contribution, of
course, has been to highlight the very close relationship between the burst oscillation
frequency and the spin frequency of the star. This fact underpins the two main
classes of model: hot spot models and global mode models. Small temperature
variations in the surface layers mean that ignition is expected to begin at a point, with
a flame front then spreading out across the star [308]. This flame may then either
stall, confining the burning to a small region (the hot spot model) or excite largescale waves in the surface layers (global mode models). The resulting temperature
asymmetry gives rise to the burst oscillations. In hot spot models, the relationship
with the spin rate is straightforward since the hot spot should be near stationary
in the rotating frame of the star [322]. The main open theoretical question for hot
spot models is what might cause the flame to stall, with both magnetic and Coriolis
forces under consideration [33, 46, 310]. For global mode models, the requirement
that burst oscillation frequency be very close to spin frequency puts very stringent
restrictions on the types of surface modes that might be responsible [133, 137, 270].
Efforts are now ongoing to develop both classes of model to determine whether they
could match the rest of the observed properties [349]. The differences between the
burst oscillations from the pulsars and the non-pulsars are informative, and suggest
a role for the magnetic field. What is still unclear, however, is whether we are seeing
one mechanism with a continuum of properties set by stellar parameters such as
the magnetic field—or whether two different mechanisms may be required. The
apparent gradual changes in burst oscillation properties from the non-pulsars and
intermittent pulsars (where magnetic field is likely to be weak) to the persistent
pulsars (with stronger fields) had until recently favoured the former. However new
results from the pulsars point increasingly towards the possibility that we are
indeed seeing two different mechanisms. The development of burst oscillations
in the slowly rotating mildly recycled pulsar IGR J17480-2466 is particularly
hard to understand in the context of global modes or Coriolis force induced hot
spots [46]. Magnetic confinement of the flame front is at present the only plausible
explanation for this source, and would also explain the extraordinary phase-locking
between accretion-powered pulsations and burst oscillations. This mechanism might
also operate in the other strong magnetic field source with phase-locked burst
oscillations, the AMXP XTE J1814-338 [46, 352]. However for the other sources,
with weaker magnetic fields, magnetic confinement is unlikely to be effective. For
these sources global mode models remain a good possibility, with the differences
in the pulsar burst oscillations being perhaps due to magnetic modifications to the
mode structure.
4.8 Aperiodic Variability and kHz QPOs
The study of aperiodic signals in power spectra of AMXPs reveals a rich phenomenology which has helped us to understand the physics of accretion disks and
compact objects. All AMXPs are “atoll sources”, so-called because of their pattern
A. Patruno and A. L. Watts
accretion-powered pulsars have played a key role. Their primary contribution, of
course, has been to highlight the very close relationship between the burst oscillation
frequency and the spin frequency of the star. This fact underpins the two main
classes of model: hot spot models and global mode models. Small temperature
variations in the surface layers mean that ignition is expected to begin at a point, with
a flame front then spreading out across the star [308]. This flame may then either
stall, confining the burning to a small region (the hot spot model) or excite largescale waves in the surface layers (global mode models). The resulting temperature
asymmetry gives rise to the burst oscillations. In hot spot models, the relationship
with the spin rate is straightforward since the hot spot should be near stationary
in the rotating frame of the star [322]. The main open theoretical question for hot
spot models is what might cause the flame to stall, with both magnetic and Coriolis
forces under consideration [33, 46, 310]. For global mode models, the requirement
that burst oscillation frequency be very close to spin frequency puts very stringent
restrictions on the types of surface modes that might be responsible [133, 137, 270].
Efforts are now ongoing to develop both classes of model to determine whether they
could match the rest of the observed properties [349]. The differences between the
burst oscillations from the pulsars and the non-pulsars are informative, and suggest
a role for the magnetic field. What is still unclear, however, is whether we are seeing
one mechanism with a continuum of properties set by stellar parameters such as
the magnetic field—or whether two different mechanisms may be required. The
apparent gradual changes in burst oscillation properties from the non-pulsars and
intermittent pulsars (where magnetic field is likely to be weak) to the persistent
pulsars (with stronger fields) had until recently favoured the former. However new
results from the pulsars point increasingly towards the possibility that we are
indeed seeing two different mechanisms. The development of burst oscillations
in the slowly rotating mildly recycled pulsar IGR J17480-2466 is particularly
hard to understand in the context of global modes or Coriolis force induced hot
spots [46]. Magnetic confinement of the flame front is at present the only plausible
explanation for this source, and would also explain the extraordinary phase-locking
between accretion-powered pulsations and burst oscillations. This mechanism might
also operate in the other strong magnetic field source with phase-locked burst
oscillations, the AMXP XTE J1814-338 [46, 352]. However for the other sources,
with weaker magnetic fields, magnetic confinement is unlikely to be effective. For
these sources global mode models remain a good possibility, with the differences
in the pulsar burst oscillations being perhaps due to magnetic modifications to the
mode structure.
4.8 Aperiodic Variability and kHz QPOs
The study of aperiodic signals in power spectra of AMXPs reveals a rich phenomenology which has helped us to understand the physics of accretion disks and
compact objects. All AMXPs are “atoll sources”, so-called because of their pattern