Unveiling soft gamma-ray repeaters with INTEGRAL

Thanks to INTEGRAL’s long exposures of the Galactic Plane, the two brightest Soft Gamma-Ray Repeaters, SGR 1806-20 and SGR 1900+14, have been monitored and studied in detail for the first time at hard-X/soft gamma rays. This has produced a wealth of new scientific results, which we will review here. Since SGR 1806-20 was particularly active during the last two years, more than 300 short bursts have been observed with INTEGRAL and their characteristics have been studied with unprecedented sensitivity in the 15–200 keV range. A hardness-intensity anticorrelation within the bursts has been discovered and the overall Number-Intensity distribution of the bursts has been determined. In addition, a particularly active state, during which 100 bursts were emitted in 10 minutes, has been observed on October 5 2004, indicating that the source activity was rapidly increasing. This eventually led to the Giant Flare of December 27th 2004, for which a possible soft gamma-ray (>80 keV) early afterglow has been detected. The deep observations allowed us to discover the persistent emission in hard X-rays (20–150 keV) from 1806-20 and 1900+14, the latter being in a quiescent state, and to directly compare the spectral characteristics of all Magnetars (two SGRs and three Anomalous X-ray Pulsars) detected with INTEGRAL. D.G. acknowledges the French Space Agency (CNES) for financial support. Based on observations with INTEGRAL, an ESA project with instruments and the science data centre funded by ESA member states (especially the PI countries: Denmark, France, Germany, Italy, Switzerland, Spain), Czech Republic and Poland, and with the participation of Russia and the USA. ISGRI has been realized and maintained in flight by CEA-Saclay/DAPNIA with the support of CNES. K.H. is grateful for support under NASA’s INTEGRAL U.S. Guest Investigator program, Grants NAG5-13738 and NNG05GG35G.     

On the iron interpretation of the 6.4 keV emission line from SGR 1900+14

A 6.4 keV emission line was discovered in an unusual burst from the soft gamma repeater SGR 1900+14 with the Rossi X-ray Timing Explorer (RXTE). The line was detected in part of a complex multipeak precursor that preceded the unusual burst of 1998 August 29, i.e. two days after the giant flare of August 27 from the source. The origin of the line was not firmly identified and two possible interpretations were equally plausible including (a) Kα fluorescence from a small iron rich material that was ejected to the magnetosphere during the August 27 flare, and (b) proton or α-particle cyclotron resonance. If the iron scenario was correct, we expect to find evidence for the line during the intervening interval between the flare and the August 29 burst, i.e. on August 28. Here we present the results of the August 28 burst observation, taken with RXTE. We detect a total of seven bursts whose individual and joint spectra do not show evidence for spectral lines. We also investigated a sample of nine bursts before and after the August 29 burst (from 1998 June to December) that do not reveal evidence for a spectral line near 6.4 keV or elsewhere. These results disfavor the iron scenario and make the proton/α-particle cyclotron resonance interpretation more plausible. The appearance of the emission line in part of a complex burst and its absence from the studied sample indicate that the line is likely due to a transient phenomenon that may depend on the burst morphology, energetics and the properties of the emission region.      

Hard Burst Emission from the Soft Gamma Repeater SGR 1900+14

We present evidence for burst emission from SGR 1900+14 with a power-law high-energy spectrum extending beyond 500 keV. Unlike previous detections of high-energy photons during bursts from soft gamma repeaters (SGRs), these emissions are not associated with extraordinarily bright flares. Not only is the emission hard, but the spectra are better fitted by D. Band's gamma-ray burst (GRB) function rather than by the traditional optically thin thermal bremsstrahlung model. We find that the spectral evolution within these hard events obeys a hardness/intensity anticorrelation. Temporally, these events are distinct from typical SGR burst emissions in that they are longer ( approximately 1 s) and have relatively smooth profiles. Despite a difference in peak luminosity of greater, similar1011 between these bursts from SGR 1900+14 and cosmological GRBs, there are striking temporal and spectral similarities between the two kinds of bursts, aside from spectral evolution. We outline an interpretation of these events in the context of the magnetar model.     

The continuum and line spectra of SGR 1806-20 bursts

The defining property of Soft Gamma Repeaters is the emission of short, bright bursts of X-rays and soft γ-rays. Here we present the continuum and line spectral properties of a large sample of bursts from SGR 1806-20, observed with the Proportional Counter Array (PCA) onboard the Rossi X-ray Timing Explorer (RXTE). Using 10 trail spectral models (5 single and 5 two component models), we find that the burst continua are best fitted by the single component models: cutoff power-law, optically thin bremsstrahlung, and simple power-law. Time resolved spectroscopy show that there are two absorption lines at ∼5 keV and 20 keV in some bursts. The lines are relatively narrow with 90% upper limit on the line widths of 0.5–1.5 keV for the 5 keV feature and 1–3 keV for the 20 keV feature. Both lines have considerable equivalent width of 330–850 eV for the 5 keV feature and 780–2590 eV for the 20 keV feature. We examined whether theses spectral lines are dependent upon the choice of a particular continuum model and find no such dependence. Besides, we find that the 5 keV feature is pronounced with high confidence in the cumulative joint spectrum of the entire burst sample, both in the individual detectors of the PCA and in the co-added detectors spectrum. We confront the features against possible instrumental effects and find that none can account for the observed line properties. The two features do not seem to be connected to the same physical mechanism because (1) they do not always occur simultaneously, (2) while the 5 keV feature occurs at about the same energy, the 20 keV line centroid varies significantly from burst to burst over the range 18–22 keV, and (3) the centroid of the lines shows anti-correlated red/blue shifts. The transient appearance of the features in the individual bursts and in portions of the same burst, together with the spectral evolution seen in some bursts point to a complex emission mechanism that requires further investigation.      

Ignition latitude and the shape of Type I X-ray bursts

The shape of the light curve during the rising phase of Type I X-ray bursts is determined by many factors including the ignition latitude, the accretion rate, and the rotation rate of the star. We develop a phenomenological model of the burst rise process and show that simple measures of the burst morphology can be robust diagnostics of ignition latitude and burning regime. We apply our results to the large sample of bursts from the low-mass X-ray binary 4U 1636−536, and find evidence for off-equatorial ignition for many of the bursts. We argue that such behaviour may be associated with the transition from hydrogen to helium ignition at accretion rates a few per cent of Eddington. We show that this model can also explain variations in the detectability of burst oscillations, and discuss the implications for other burst sources.     

Discovery of type I X-ray bursts from the low-mass X-ray binary 4U 1708 – 40

We report the discovery of type I X-ray bursts from the low-mass X-ray binary  4U 1708 − 40  during the 100-ks observation performed by BeppoSAX on 1999 August 15–16. Six X-ray bursts have been observed. The unabsorbed 2–10 keV fluxes of the bursts range from ∼3 to  9 × 10⁻¹⁰ erg cm⁻² s⁻¹  . A correlation between peak flux and fluence of the bursts is found, in agreement with the behaviour observed in other similar sources. There is a trend of the burst flux to increase with the time interval from the previous burst. From the value of the persistent flux we infer a mass accretion rate     , which may correspond to the mixed hydrogen/helium burning regime triggered by thermally unstable hydrogen. We have also analysed a BeppoSAX observation performed on 2001 August 22 and previous RXTE observations of  4U 1708 − 40  , where no bursts have been observed; we find persistent fluxes of more than a factor of 7 higher than the persistent flux observed during the BeppoSAX observation showing X-ray bursts.     

Systematic variation in the apparent burning area of thermonuclear bursts and its implication for neutron star radius measurement

Precision measurements of neutron star radii can provide a powerful probe of the properties of cold matter beyond nuclear density. Beginning in the late 1970s, it was proposed that the radius could be obtained from the apparent or inferred emitting area during the decay portions of thermonuclear (type I) X-ray bursts. However, this apparent area is generally not constant, preventing a reliable measurement of the source radius. Here, we report for the first time a correlation between the variation of the inferred area and the burst properties, measured in a sample of almost 900 bursts from 43 sources. We found that the rate of change of the inferred area during decay is anticorrelated with the burst decay duration. A Spearman rank correlation test shows that this relation is significant at the  <10⁻⁴⁵  level for our entire sample, and at the  7 × 10⁻³⁷  level for the 625 bursts without photospheric radius expansion. This anticorrelation is also highly significant for individual sources exhibiting a wide range of burst durations, such as 4U 1636–536 and Aql X-1. We suggest that variations in the colour factor, which relates the colour temperature resulted from the scattering in the neutron star atmosphere to the effective temperature of the burning layer, may explain the correlation. This in turn implies significant variations in the composition of the atmosphere between bursts with long and short durations.     

Short gamma-ray bursts from SGR giant flares and neutron star mergers: two populations are better than one

There is increasing evidence of a local population of short duration gamma-ray bursts (sGRB), but it remains to be seen whether this is a separate population to higher redshift bursts. Here we choose plausible luminosity functions (LFs) for both neutron star binary mergers and giant flares from soft gamma repeaters (SGR), and combined with theoretical and observed Galactic intrinsic rates we examine whether a single progenitor model can reproduce both the overall Burst and Transient Source Experiment (BATSE) sGRB number counts and a local population, or whether a dual progenitor population is required. Though there are large uncertainties in the intrinsic rates, we find that at least a bimodal LF consisting of lower and higher luminosity populations is required to reproduce both the overall BATSE sGRB number counts and a local burst distribution. Furthermore, the best-fitting parameters of the lower luminosity population agree well with the known properties of SGR giant flares, and the predicted numbers are sufficient to account for previous estimates of the local sGRB population.     

Time-dependent photoionization opacities in dense gamma-ray burst environments

The recent detection of a transient absorption feature in the X-ray prompt emission of GRB 990705 showed the importance of such observations in the understanding of gamma-ray bursts and their progenitors. We investigate the time dependence of photoionization edges during the prompt emission of bursts in different environments. We show that their variability can be used to infer the density and geometry of the surrounding medium, giving important clues to unveil the nature of the burst progenitor.     

The steady spin-down rate of 4U 1907+09

Using X-ray data from the Rossi X-ray Timing Explorer , we report the pulse timing results of the accretion-powered, high-mass X-ray binary pulsar 4U 1907+09, covering a time-span of almost two years. We measured three new pulse periods in addition to the previously measured four pulse periods. We are able to connect pulse arrival times in phase for more than a year. The source has been spinning down almost at a constant rate, with a spin-down rate of     for more than 15 yr. Residuals of pulse arrival times yield a very low level of random-walk noise, with a strength of ∼     on a time-scale of 383 d, which is 40 times lower than that of the high-mass X-ray binary pulsar Vela X-1. The noise strength is only a factor of 5 greater than that of the low-mass X-ray binary pulsar 4U 1626−67. The low level of the timing noise and the very stable spin-down rate of 4U 1907+09 make this source unique among the high-mass X-ray binary pulsars, providing another example, in addition to 4U 1626−67, of long-term quiet spin down from an accreting source. These examples show that the extended quiet spin-down episodes observed in the anomalous X-ray pulsars 1RXS J170849.0−400910 and 1E 2259+586 do not necessarily imply that these sources are not accreting pulsars.     

Giant flare in SGR 1806-20 and its Compton reflection from the Moon

We analyze the data obtained when the Konus-Wind gamma-ray spectrometer detected a giant flare in SGR 1806-20 on December 27, 2004. The flare is similar in appearance to the two known flares in SGR 0526-66 and SGR 1900+14 while exceeding them significantly in intensity. The enormous X-ray and gamma-ray flux in the narrow initial pulse of the flare leads to almost instantaneous deep saturation of the gamma-ray detectors, ruling out the possibility of directly measuring the intensity, time profile, and energy spectrum of the initial pulse. In this situation, the detection of an attenuated signal of inverse Compton scattering of the initial pulse emission by the Moon with the Helicon gamma-ray spectrometer onboard the Coronas-F satellite was an extremely favorable circumstance. Analysis of this signal has yielded the most reliable temporal, energy, and spectral characteristics of the pulse. The temporal and spectral characteristics of the pulsating flare tail have been determined from Konus-Wind data. Its soft spectra have been found to contain also a hard power-law component extending to 10 MeV. A weak afterglow of SGR 1806-20 decaying over several hours is traceable up to 1 MeV. We also consider the overall picture of activity of SGR 1806-20 in the emission of recurrent bursts before and after the giant flare.     

Biases for neutron star mass, radius and distance measurements from Eddington-limited X-ray bursts

Eddington-limited X-ray bursts from neutron stars can be used in conjunction with other spectroscopic observations to measure neutron star masses, radii and distances. In order to quantify some of the uncertainties in the determination of the Eddington limit, we analysed a large sample of photospheric radius-expansion thermonuclear bursts observed with the Rossi X-ray Timing Explorer . We identified the instant at which the expanded photosphere 'touches down' back on to the surface of the neutron star and compared the corresponding touchdown flux to the peak flux of each burst. We found that for the majority of sources, the ratio of these fluxes is smaller than ≃1.6, which is the maximum value expected from the changing gravitational redshift during the radius expansion episodes (for a  2 M_⊙  neutron star). The only sources for which this ratio is larger than ≃1.6 are high-inclination sources that include dippers and Cyg X-2. We discuss two possible geometric interpretations of this effect and show that the inferred masses and radii of neutron stars are not affected by this bias. On the other hand, systematic uncertainties as large as ∼50 per cent may be introduced to the distance determination.     

A very rare triple-peaked type-I X-ray burst in the low-mass X-ray binary 4U 1636−53

We have discovered a triple-peaked X-ray burst from the low-mass X-ray binary (LMXB) 4U 1636−53 with the Rossi X-ray Timing Explorer ( RXTE ). This is the first triple-peaked burst reported from any LMXB using RXTE , and it is only the second burst of this kind observed from any source. (The previous one was also from 4U 1636−53, and was observed with EXOSAT .) From fits to time-resolved spectra, we find that this is not a radius-expansion burst, and the same triple-peaked pattern seen in the X-ray light curve is also present in the bolometric light curve of the burst. Similar to what was previously observed in double-peaked bursts from this source, the radius of the emitting area increases steadily during the burst, with short periods in between during which the radius remains more or less constant. The temperature first increases steeply, and then decreases across the burst also showing three peaks. The first and last peak in the temperature profile occur, respectively, significantly before and after the first and last peaks in the X-ray and bolometric light curves. We found no significant oscillations during this burst. This triple-peaked burst, as well as the one observed with EXOSAT and the double-peak bursts in this source, all took place when 4U 1636−53 occupied a relatively narrow region in the colour–colour diagram, corresponding to a relatively high (inferred) mass-accretion rate. No model presently available is able to explain the multiple-peaked bursts.     

On the role of extinction in failed gamma-ray burst optical/infrared afterglows

While all but one of the gamma-ray bursts observed in the X-ray band showed an X-ray afterglow, about 60 per cent of them have not been detected in the optical band. We demonstrate that in many cases this is not as a result of adverse observing conditions, or delay in performing the observations. We also show that the optically non-detected afterglows are not affected by particularly large Galactic absorbing columns, since its distribution is similar for both the detected and non-detected burst subclasses. We then investigate the hypothesis that the failure of detecting the optical afterglow is due to absorption at the source location. We find that this is a marginally viable interpretation, but only if the X-ray burst and afterglow emission and the possible optical/UV flash do not destroy the dust responsible for absorption in the optical band. If dust is efficiently destroyed, we are led to conclude that bursts with no detected optical afterglow are intrinsically different. Prompt infrared observations are the key to solving this issue.     

Temporal properties of short gamma-ray bursts

We analyse a sample of bright short bursts from the BATSE 4B-catalog and find that many short bursts are highly variable  ( δt _min/ T ≪1  , where δt _min is the shortest pulse duration and T is the burst duration). This indicates that it is unlikely that short bursts are produced by external shocks. We also analyse the available (first  1–2 s)  high-resolution Time Tagged Events (TTE) data of some of the long bursts. We find that variability on a 10-ms time-scale is common in long bursts. This result shows that some long bursts are even more variable than it was thought before  ( δt _min/ T ≈10^-4–10^-3)  .     

Radiation characteristics of quasi-periodic radio bursts in the Jovian high-latitude region

Ulysses had a “distant encounter” with Jupiter in February 2004. The spacecraft passed from north to south, and it observed Jovian radio waves from high to low latitudes (from +80° to +10°) for few months during its encounter. In this study, we present a statistical investigation of the occurrence characteristics of Jovian quasi-periodic bursts, using spectral data from the unified radio and plasma wave experiment (URAP) onboard Ulysses. The latitudinal distribution of quasi-periodic bursts is derived for the first time. The analysis suggested that the bursts can be roughly categorized into two types: one having periods shorter than 30 min and one with periods longer than 30 min, which is consistent with the results of the previous analysis of data from Ulysses’ first Jovian flyby [MacDowall, R.J., Kaiser, M.L., Desch, M.D., Farrell, W.M., Hess, R.A., Stone, R.G., 1993. Quasi-periodic Jovian radio bursts: observations from the Ulysses radio and plasma wave. Experiment. Planet. Space Sci. 41, 1059-1072]. It is also suggested that the groups of quasi-periodic bursts showed a dependence on the Jovian longitude of the sub-solar point, which means that these burst groups are triggered during a particular rotational phase of the planet. Maps of the occurrence probability of these quasi-periodic bursts also showed a unique CML/MLAT dependence. We performed a 3D ray tracing analysis of the quasi-periodic burst emission to learn more about the source distribution. The results suggest that the longitudinal distribution of the occurrence probability depends on the rotational phase. The source region of quasi-periodic bursts seems to be located at an altitude between 0.4 and 1.4 Rj above the polar cap region (L>30).     

The broad-band noise characteristics of selected anomalous X-ray pulsars and soft gamma-ray repeaters

We present the broad-band noise structure of selected anomalous X-ray pulsars (AXPs) and soft gamma-ray repeaters (SGRs) in the 2–60 keV energy band. We have analysed Rossi X-Ray Timing Explorer Proportional Counter Array archival light curves for four AXPs and one SGR. We detect that the persistent emission of these sources shows band-limited noise at low frequencies in the range 0.005–0.05 Hz varying from 2.5 to 70 per cent integrated rms in times of prolonged quiescence and following outbursts. We discovered band-limited red noise in 1E 2259+586 only for ∼2 yr after its major 2002 outburst. The system shows no broad-band noise otherwise. Although this rise in noise in 1E 2259+586 occurred following an outburst which included a rotational glitch, the other glitching AXPs showed no obvious change in broad-band noise, thus it does not seem that this noise is correlated with glitches. The only source that showed significant variation in broad-band noise was 1E   1048.1−5937  , where the noise gradually rose for 1.95 yr at a rate of ∼3.6 per cent per year. For this source the increases in broad-band noise was not correlated with the large increases in persistent and pulsed flux, or its two short SGR-like bursts. This rise in noise did commence after a long burst, however, given the sparsity of this event, and the possibility that similar bursts went unnoticed the trigger for the rise is noise in 1E   1048.1−5937  is not as clear as for 1E 2259+586. The other three sources indicate a persistent band-limited noise at low levels in comparison.     

Gamma-ray burst efficiency and possible physical processes shaping the early afterglow

The discovery by Swift that a good fraction of gamma-ray bursts (GRBs) have a slowly decaying X-ray afterglow phase led to the suggestion that energy injection into the blast wave takes place several hundred seconds after the burst. This implies that right after the burst the kinetic energy of the blast wave was very low and in turn the efficiency of production of γ-rays during the burst was extremely high, rendering the internal shocks model unlikely. We re-examine the estimates of kinetic energy in GRB afterglows and show that the efficiency of converting the kinetic energy into γ-rays is moderate and does not challenge the standard internal shock model. We also examine several models, including in particular energy injection, suggested to interpret this slow decay phase. We show that with proper parameters, all these models give rise to a slow decline lasting several hours. However, even those models that fit all X-ray observations, and in particular the energy injection model, cannot account self-consistently for both the X-ray and the optical afterglows of well-monitored GRBs such as GRB 050319 and GRB 050401. We speculate about a possible alternative resolution of this puzzle.     

Analysis of COMPTEL gamma-ray burst locations and spectra

In its first three years of operation, the COMPTEL instrument on theCompton Gamma-Ray Observatory has measured the locations (mean accuracy 1°) and spectra (0.75-30 MeV) of 18 gamma-ray bursts and continues to observe new events at a rate of 1/month. With good angular resolution and sensitivity at MeV energies, the growing COMPTEL burst catalog is an important new piece of evidence in the on-going GRB mystery. The COMPTEL burst locations are consistent with an isotropic distribution of sources, yet the spatial coincidence of two of the bursts indicates the possibility of repetition. The COMPTEL burst spectra are in most cases consistent with a single power law model with spectral index in the range 2–3. However, two bursts show evidence of a spectral break in the MeV range. Measurement of rapid variability at MeV energies in the stronger bursts provides evidence that either the sources are nearby (within the Galaxy) or the gamma-ray emission is relativistically beamed. We present an overview of analysis results obtained from the COMPTEL burst catalog concentrating on the search for burst repetition and the implications of highly variable MeV emission.     

The energy dependence of burst oscillations from the accreting millisecond pulsar XTE J1814−338

The nature of the asymmetry that gives rise to Type I X-ray burst oscillations on accreting neutron stars remains a matter of debate. Of particular interest is whether the burst oscillation mechanism differs between the bursting millisecond pulsars and the non-pulsing systems. One means to diagnose this is to study the energy dependence of the burst oscillations: here we present an analysis of oscillations from 28 bursts observed during the 2003 outburst of the accreting millisecond pulsar XTE J1814−338. We find that the fractional amplitude of the burst oscillations falls with energy, in contrast to the behaviour found by Muno et al. in the burst oscillations from a set of non-pulsing systems. The drop with energy mirrors that seen in the accretion-powered pulsations; in this respect XTE J1814−338 behaves like the other accreting millisecond pulsars. The burst oscillations show no evidence for either hard or soft lags, in contrast to the persistent pulsations, which show soft lags of up to 50 μs. The fall in amplitude with energy is inconsistent with current surface-mode and simple hotspot models of burst oscillations. We discuss improvements to the models and uncertainties in the physics that might resolve these issues.