The existence of ultra-fast rotating neutron stars (spin period P ∼ < 1 ms) is expected on the basis of current models for the secular evolution of interacting binaries, though they have not been detected yet. Their formation depends on the quantity of matter accreted by the neutron star which, in turn, is limited by the mechanism of mass ejection from the binary. An efficient mass ejection can avoid the formation of ultra-fast pulsars or their accretion induced collapse to a black hole. We propose that significant reductions of the mass-transfer rate may cause the switch-on of a radio pulsar phase, whose radiation pressure may be capable of ejecting out of the system most of the matter transferred by the companion. This can prevent, for long orbital periods and if a sufficiently fast spin has been reached, any further accretion, even if the original transfer rate is restored, thus limiting the minimum spin period attainable by the neutron star. We show that close systems (orbital periods P orb ∼ 1 hr) are the only possible hosts for ultra-fast spinning neutron stars. This could explain why ultra-fast radio pulsars have not been detected so far, as the detection of pulsars with very short spin periods in close systems is hampered, in current radio surveys, by strong Doppler modulation and computational limitations.

We report on a 63-ks long XMM-Newton observation of the accreting millisecond pulsar SAX J1808.4-3658 during the latest X-ray outburst which started on September 21st 2008. The pn spectrum shows a highly significant emission line in the energy band where the iron K-α line is expected, and which we identify as emission from neutral (or mildly ionized) iron. The line profile appears to be quite broad (more than 1 keV FWHM) and asymmetric; the most probable explanation for this profile is Doppler and relativistic broadening from the inner accretion disc. From a fit with a diskline profile we find an inner radius of the disc of 8.7+3.7 −2.7 Rg, corresponding to 18.0 +7.6 −5.6 km for a 1.4 M⊙ neutron star. The disc therefore appears truncated inside the corotation radius (31 km for SAX J1808.4-3658) in agreement with the fact that the source was still showing pulsations during the XMM-Newton observation.

We report on the results of a broad band (0.1-100 keV) spectral analysis of the bursting atoll source MXB 1728MXB -34 (4U 1728 observed by the BeppoSAX satellite. Three bursts were present during this observation. The spectrum during the bursts can be fitted by a blackbody with a temperature of ∼ 2 keV. The radius of the blackbody emitting region is compatible with the radius of the neutron star if we correct for the difference between the observed color temperature and the effective temperature. From the bursts we also estimate a distance to the source of ∼ 5.1 kpc. MXB 1728-34 was in a rather soft state during the BeppoSAX observation. The persistent spectrum is well fitted by a continuum consisting of a soft blackbody emission and a comptonized spectrum. We interpreted the soft component as the emission from the accretion disk. Taking into account a spectral hardening factor of ∼ 1.7 (because of electron scattering which modifies the blackbody spectrum emitted by the disk), we estimated that the inner disk radius is R in √ cos i ∼ 20 km, where i is the inclination angle. The comptonized component could originate in a spherical corona, with temperature ∼ 10 keV and optical depth ∼ 5, surrounding the neutron star. A broad gaussian emission line at ∼ 6.7 keV is observed in the spectrum, probably emitted in the ionized corona or in the inner part of the disk. Another emission line is present at ∼ 1.66 keV. No reflection component is detected with high statistical significance, probably because of the low temperature of the corona in this state of the source. If the iron emission line is due to reflection of the comptonized spectrum by the accretion disk, it requires a ionized disk (ξ ∼ 280) and a solid angle of ∼ 0.2 (in units of 2π) subtended by the reflector as seen from the corona.

We present the results of a BeppoSAX observation of the Z source GX 349+2 covering the energy range 0.1-200 keV. The presence of flares in the light curve indicates that the source was in the flaring branch during the BeppoSAX observation. We accumulated energy spectra separately for the non-flaring intervals and the flares. In both cases the continuum is well described by a soft blackbody (kT BB ∼ 0.5 keV) and a Comptonized spectrum corresponding to an electron temperature of kT e ∼ 2.7 keV, optical depth τ ∼ 10 (for a spherical geometry), and seed photon temperature of kT W ∼ 1 keV. All temperatures tend to increase during the flares. In the non-flaring emission a hard tail dominates the spectrum above 30 keV. This can be fit by a power law with photon index ∼ 2, contributing ∼ 2% of the total source luminosity over the BeppoSAX energy range. A comparison with hard tails detected in some soft states of black hole binaries suggests that a similar mechanism could originate these components in black hole and neutron star systems.

We report on the spectral analysis of the peculiar source Cir X-1 observed by the BeppoSAX satellite when the X-ray source was near the periastron. A Ñare lasting D6 ] 103 s is present at the beginning of the observation. The luminosity during the persistent emission is 1 ] 1038 ergs s~1, while during the Ñare it is 2 ] 1038 ergs s~1. We produced broadband (0.1È100 keV) energy spectra during the Ñare and the persistent emission. At low energies the continuum is well Ðtted by a model consisting of Comptonization of soft photons, with a temperature of D0.4 keV, by electrons at a temperature of D1 keV. After the Ñare, a power-law component with photon index D3 is dominant at energies higher than 10 keV. This component contributes D4% of the total luminosity. During the Ñare its addition is not statistically signiÐcant. An absorption edge at D8.4 keV, with optical depth D1, corresponding to the K edge of Fe XXIIIÈFe XXV, and an iron emission line at 6.7 keV are also present. The iron-line energy is in agreement with the ionization level inferred from the absorption edge. The hydrogen column deduced from the absorption edge is D1024 cm~2, 2 orders of magnitude larger than the low-energy absorption measured in this source. We calculated the radius of the region originating the Comptonized seed photons, km. We propose a scenario where (the Wien radius) is the inner disk radius, a R W D 150 R W highly ionized torus surrounds the accretion disk, and a magnetosphere is present up to The R W . absorption edge and the emission line could originate in the photoionized torus, while the Comptonized component originates in an inner region of the disk.