3 Magnetars: A Short Review and Some Sparse Considerations
107
to an X-ray telescope that provides ∼arcsec localisation), the short bursts have
become the primary channel for the discovery of new magnetars and of the onset
of magnetic activity from known sources. The bursting activity is unpredictable:
Magnetars usually go through long stretches of quiescence (even decades) that can
be interrupted by sparse bursts or by paroxysmal activity, during which hundreds or
thousands of events are clustered in days. Or by everything in between.
The impulsive emission of X/gamma ray photons lasts from milliseconds to a few
seconds and the peak luminosity is typically in the range ≈10 36 –10 43 erg s −1 . Their
profile is usually single-peaked and asymmetric, with a faster rise than decay. Twoor multi-peak bursts are not rare, but since series of bursts can be very rapid, the
distinction between single events and multi-peak bursts may be tenuous. Figure 3.4
shows the light curves of 21 bursts collected from SGR 1806–20 with INTEGRAL
in 2003 (from Götz et al. [86]).
Some bursts, in general the brightest (‘intermediate flares’), can be followed by
X-ray tails surviving from minutes to hours. These events resemble the overall shape
of the giant flares and their afterglows, except that in some instances the energy in
the tail exceed that emitted in the spike (see e.g. Pintore et al. [173]). Sometimes
these tails show flux modulation at the neutron star spin period, linking directly the
afterglow to the neutron star (e.g. to a region of the star surface that was heated
by the burst), but a significant fraction of them may be due to (or receive a large
contribution from) dust-scattering of the burst emission, in which a fraction of the
photons of the burst is re-emitted after reprocessing by clouds/layers of interstellar
dust between us and the source [61, 95, 137, 173]. In general, it is possible to
disentangle the contribution of the delayed dust scattering from that intrinsic to
the neutron star only when data sets with a large number of photons and good
spatial resolution are available, but sometimes the phenomenon is truly spectacular:
see Tiengo et al. [211] for the study of a scattering halo surrounding the magnetar
1E 1547.0–5408 that took the shape of at least three expanding symmetric rings.
A number of models have been used to model the spectral emission of short
magnetar bursts, but a single-blackbody model, or a double-blackbody model when
broad band data are available (e.g. 1–200 keV), are usually a safe bet [115, 119]. The
blackbody temperature kT is typically in the range from ∼2 to 12 keV and when
the double-blackbody decomposition is viable, a bimodal distribution of radii end
temperatures takes shape, with a soft blackbody and a hotter (∼12 keV) blackbody
with smaller surface.
The fluence (S) distribution for number (N) of bursts usually follows a powerlaw function N(> S) ∝ S −α over several orders of magnitude, with α ∼ 0.6–0.9,
depending on the sources and the instruments [5, 87, 93, 94]. It has been pointed
out many times that this behaviour is similar to what observed for earthquakes. The
truth is that such distribution is rather ubiquitous in nature and also in artificial
and human-influenced systems: classic examples are the sizes of landslides and
avalanches, solar flares, forest fires, lunar craters and cities, or the frequency of use
of words in human languages (e.g. Newman [162]). As observed by Paizis and Sidoli
[166] for the distribution of hard-X luminosity of supergiant fast X-ray transient,
the power-law distribution is characteristic (but not exclusive) of the self-organized
107
to an X-ray telescope that provides ∼arcsec localisation), the short bursts have
become the primary channel for the discovery of new magnetars and of the onset
of magnetic activity from known sources. The bursting activity is unpredictable:
Magnetars usually go through long stretches of quiescence (even decades) that can
be interrupted by sparse bursts or by paroxysmal activity, during which hundreds or
thousands of events are clustered in days. Or by everything in between.
The impulsive emission of X/gamma ray photons lasts from milliseconds to a few
seconds and the peak luminosity is typically in the range ≈10 36 –10 43 erg s −1 . Their
profile is usually single-peaked and asymmetric, with a faster rise than decay. Twoor multi-peak bursts are not rare, but since series of bursts can be very rapid, the
distinction between single events and multi-peak bursts may be tenuous. Figure 3.4
shows the light curves of 21 bursts collected from SGR 1806–20 with INTEGRAL
in 2003 (from Götz et al. [86]).
Some bursts, in general the brightest (‘intermediate flares’), can be followed by
X-ray tails surviving from minutes to hours. These events resemble the overall shape
of the giant flares and their afterglows, except that in some instances the energy in
the tail exceed that emitted in the spike (see e.g. Pintore et al. [173]). Sometimes
these tails show flux modulation at the neutron star spin period, linking directly the
afterglow to the neutron star (e.g. to a region of the star surface that was heated
by the burst), but a significant fraction of them may be due to (or receive a large
contribution from) dust-scattering of the burst emission, in which a fraction of the
photons of the burst is re-emitted after reprocessing by clouds/layers of interstellar
dust between us and the source [61, 95, 137, 173]. In general, it is possible to
disentangle the contribution of the delayed dust scattering from that intrinsic to
the neutron star only when data sets with a large number of photons and good
spatial resolution are available, but sometimes the phenomenon is truly spectacular:
see Tiengo et al. [211] for the study of a scattering halo surrounding the magnetar
1E 1547.0–5408 that took the shape of at least three expanding symmetric rings.
A number of models have been used to model the spectral emission of short
magnetar bursts, but a single-blackbody model, or a double-blackbody model when
broad band data are available (e.g. 1–200 keV), are usually a safe bet [115, 119]. The
blackbody temperature kT is typically in the range from ∼2 to 12 keV and when
the double-blackbody decomposition is viable, a bimodal distribution of radii end
temperatures takes shape, with a soft blackbody and a hotter (∼12 keV) blackbody
with smaller surface.
The fluence (S) distribution for number (N) of bursts usually follows a powerlaw function N(> S) ∝ S −α over several orders of magnitude, with α ∼ 0.6–0.9,
depending on the sources and the instruments [5, 87, 93, 94]. It has been pointed
out many times that this behaviour is similar to what observed for earthquakes. The
truth is that such distribution is rather ubiquitous in nature and also in artificial
and human-influenced systems: classic examples are the sizes of landslides and
avalanches, solar flares, forest fires, lunar craters and cities, or the frequency of use
of words in human languages (e.g. Newman [162]). As observed by Paizis and Sidoli
[166] for the distribution of hard-X luminosity of supergiant fast X-ray transient,
the power-law distribution is characteristic (but not exclusive) of the self-organized