188
A. Patruno and A. L. Watts
where y = ρdz is the column density, z is the direction orthogonal to the surface
in the plane parallel geometry approximation, ˙
m is the local mass accretion rate
per unit area (in units of g cm −2 s −1 ), H =
dlnP
dz is the pressure (defined as
P ) scale-height, and η is the magnetic diffusivity coefficient. This model predicts
that LMXBs with very low mass accretion rates should be AMXPs, and that
the most luminous LMXBs should not pulsate. This seems to be the case, since
persistently pulsating AMXPs are indeed all faint LMXBs. However, the discovery
of intermittent pulsations as short as ∼100 s in Aql X-1 poses severe problems
for the model since it is unclear how a buried B field could emerge for this brief
episode. Another interesting observational result comes from HETE J1900.1-2455,
which, with a mean bolometric luminosity of 4.4 × 10 36 erg s −1 , has a mean mass
accretion rate ˙
M ≈ 8 × 10 −10 M yr −1 for an assumed accretion efficiency of
20%. This accretion rate is about twice that of most AMXPs, but is smaller than the
bolometric luminosity of SAX J1808.4-3658 during observations taken close to the
peak of the 2008 outburst [159]. This might indicate that the screening timescale
is long compared to the outburst duration of most AMXPs. In HETE J1900.1-2455
observational evidence for magnetic field screening has been found [250], with the
magnetic field possibly reducing by orders of magnitude on a timescale of about 100
days during which channeled accretion becomes less and less efficient until it stops
working. If confirmed this indicates that indeed most LMXBs might be not pulsating
because of the lack of an extended magnetosphere. Recent 2-D and 3-D numerical
simulations have started to address this problem and results seems to suggest that
submergence of the magnetic field is possible at least when the accretion rate is high
([20] but see also [222, 268, 269] for different results). Despite still open problems,
the screening model remains therefore one of the most promising, although further
efforts are necessary to formulate new sophisticated predictions and tests.
Smearing from an Optically Thick Corona This model suggests that X-ray emission from AMXPs is formed primarily via Comptonization processes in a relatively
optically thin medium. If the Compton cloud is thick, then pulsations can be washed
out. This scenario has been criticized on the grounds that spectral analyses of most
LMXBs show small optical depths, well below the limit required to smear out
pulsations [116], which should therefore be visible (but see [332]). It is not clear
why in this model the intermittent AMXP SAX J1748.9-2021 shows pulsations that
appear and disappear on timescales of a few hundred seconds [2, 253], uncorrelated
with spectral changes. However it provides a reasonable explanation for the time
lags and pulse energy dependence, especially for the pulses of IGR J00291+5934
which have a turnover at 7 keV that is otherwise difficult to explain [86, 88].
Gravitational Light Bending For large stellar compactness (M/R), gravitational
light bending can distort the path of X-ray photons up to the point where pulsations
are strongly reduced [19, 234, 364]. Persistent sources may therefore be less likely
to show pulsations because they accrete more mass than transient systems. However,
the total amount of mass that a NS has accreted does not depend on the current mass
accretion rate and on the persistence of the X-ray flux, but on the evolutionary stage
of the donor star and on the average mass transfer rate from the donor towards the
A. Patruno and A. L. Watts
where y = ρdz is the column density, z is the direction orthogonal to the surface
in the plane parallel geometry approximation, ˙
m is the local mass accretion rate
per unit area (in units of g cm −2 s −1 ), H =
dlnP
dz is the pressure (defined as
P ) scale-height, and η is the magnetic diffusivity coefficient. This model predicts
that LMXBs with very low mass accretion rates should be AMXPs, and that
the most luminous LMXBs should not pulsate. This seems to be the case, since
persistently pulsating AMXPs are indeed all faint LMXBs. However, the discovery
of intermittent pulsations as short as ∼100 s in Aql X-1 poses severe problems
for the model since it is unclear how a buried B field could emerge for this brief
episode. Another interesting observational result comes from HETE J1900.1-2455,
which, with a mean bolometric luminosity of 4.4 × 10 36 erg s −1 , has a mean mass
accretion rate ˙
M ≈ 8 × 10 −10 M yr −1 for an assumed accretion efficiency of
20%. This accretion rate is about twice that of most AMXPs, but is smaller than the
bolometric luminosity of SAX J1808.4-3658 during observations taken close to the
peak of the 2008 outburst [159]. This might indicate that the screening timescale
is long compared to the outburst duration of most AMXPs. In HETE J1900.1-2455
observational evidence for magnetic field screening has been found [250], with the
magnetic field possibly reducing by orders of magnitude on a timescale of about 100
days during which channeled accretion becomes less and less efficient until it stops
working. If confirmed this indicates that indeed most LMXBs might be not pulsating
because of the lack of an extended magnetosphere. Recent 2-D and 3-D numerical
simulations have started to address this problem and results seems to suggest that
submergence of the magnetic field is possible at least when the accretion rate is high
([20] but see also [222, 268, 269] for different results). Despite still open problems,
the screening model remains therefore one of the most promising, although further
efforts are necessary to formulate new sophisticated predictions and tests.
Smearing from an Optically Thick Corona This model suggests that X-ray emission from AMXPs is formed primarily via Comptonization processes in a relatively
optically thin medium. If the Compton cloud is thick, then pulsations can be washed
out. This scenario has been criticized on the grounds that spectral analyses of most
LMXBs show small optical depths, well below the limit required to smear out
pulsations [116], which should therefore be visible (but see [332]). It is not clear
why in this model the intermittent AMXP SAX J1748.9-2021 shows pulsations that
appear and disappear on timescales of a few hundred seconds [2, 253], uncorrelated
with spectral changes. However it provides a reasonable explanation for the time
lags and pulse energy dependence, especially for the pulses of IGR J00291+5934
which have a turnover at 7 keV that is otherwise difficult to explain [86, 88].
Gravitational Light Bending For large stellar compactness (M/R), gravitational
light bending can distort the path of X-ray photons up to the point where pulsations
are strongly reduced [19, 234, 364]. Persistent sources may therefore be less likely
to show pulsations because they accrete more mass than transient systems. However,
the total amount of mass that a NS has accreted does not depend on the current mass
accretion rate and on the persistence of the X-ray flux, but on the evolutionary stage
of the donor star and on the average mass transfer rate from the donor towards the