Engineering topics
Publications and source records attributed to Van Paradijs, J..
Quasi-periodic oscillations (QPOs) have been detected in the X-ray flux of the bright Galactic bulge X-ray source GX 340+0. The QPO, the associated red noise, and their relation to the spectral state of the source are described. The QPO behavior is found to be similar to that seen in Cyg X-2 and GX 5-1.
During about 80 h of uninterrupted observations of the X-ray burst source 4 U/MXB 1735-44, made in August 1985 with Exosat, five bursts were observed. The bursts occur at irregular intervals, ranging between 30 min and 36 h. The burst sizes do not scale with the intervals preceding the bursts; in particular the energy in a burst which occurred after an interval of 50 min was only 30 percent less than that of a burst which occurred after more than 20 h. Due to the long uninterrupted observations possible with Exosat, this is the first time for 1735-44 that the burst intervals and the burst characteristics (e.g., fluence) can be compared. The observations show that the model recently proposed by Fujimoto et al. (1987) to explain irregular burst intervals, is not applicable to 1735-44. The amplitude of periodic X-ray flux variations at the 4.65 h orbital period is less than 4 percent.
The results of an Exosat observation of the 'Rapid Burster' in August 1985 during an active state are presented. Forty 1-10 minute long type 2 bursts were detected together with constant emission. Quasi-periodic oscillations (QPOs) were observed in 23 bursts with frequencies in the range of 2-5 Hz. Obscuration of the central X-ray source by an oscillating accretion disk rim is suggested as a possible scenario for the generation of the QPO from the Rapid Burster.
The burst-induced accretion-disk corona (BIADC) model proposed by Melia (1987) to explain the multiple-peak burst profiles of 4U/MXB 1636-53 in terms of direct and scattered components is examined critically. Published observational data (Sztajno et al., 1986, and Lewin et al., 1987) are presented in tables and graphs and analyzed. A number of possible BIADC scenarios are discussed, and it is argued that the observed characteristics of the 1636-53 bursts are not well accounted for by the BIADC model.
Results of Monte Carlo simulations of the Comptonization of photons in a finite spherical Comptonizing cloud are presented. The energy dependence of the arrival times of the Comptonized photons is very different for different input photon spectra. If the delays observed in the high-frequency QPO of Cyg X-2 and GX5-1 are the result of Comptonization, multichannel time-delay measurements may constrain the energy of the input photons and thus provide important information on the accretion process in low-mass X-ray binaries.
During Exosat observations of 4U/MXB 1636-53 in August 1985, three bursts were observed in 6 h and 24 bursts in 79 h of uninterrupted observation. The burst times, burst fluences, peak burst fluxes, burst rise times, decay times, burst intervals, mean persistent fluxes, and the product of the latter two are listed, analyzed, and discussed. The implications of the results for thermonuclear flash models are examined. A model which fits the data well involves bursts which suffer from nuclear energy losses due to stable hydrogen burning between bursts. Thus, the percentage of 'lost' energy increases with increasing burst intervals. For individual bursts, however, there are significant deviations from the approximate linear relationship.
The intensity variations of Cyg X-2 and GX 5-1 are analyzed with a Fourier cross-spectral technique making it possible to measure the frequency dependence of the time lags between the intensity variations in two X-ray spectral bands. In the 20-40 Hz frequency range of the quasi-periodic oscillations (QPO), the hard (about 5-18 keV) signal lags the soft (about 1-5 keV) one, while the reverse is true for the associated red noise. The observed time lags range from 0.4 to 8 ms. For similar QPO parameters, the observed lags are much smaller in GX 5-1 than they are in Cyg X-2. The 'hard lags' previously measured in the cross-correlation functions of these sources are now identified as due to lags in the QPO. It is argued that the 'soft lags' now detected in the red noise are not inconsistent with the interpretation of the hard lags as due to inverse Compton scattering.
The millisec-order delay between quasi-periodic oscillations (QPOs) observed from Cyg X-2 and GX 5-1, in a high-energy and a low-energy bandpass, may be due to the Compton scattering of photons in a cloud of hot electron gas, depending on optical depth, electron temperature, and cloud radius, as well as input photon spectra. Monte Carlo simulations are used to demonstrate how the observed delay and X-ray spectrum can be used to constrain the size of the scattering cloud, for an assumed input spectrum of photons. Both QPO in emergent photons and a time lag between high and low energy photons can be generated not only by the scattering of an oscillating signal in a constant cloud, but also by the scattering of a constant source of input photons through a cloud with oscillating optical depth.
Exosat observations of GX3+1 in 1985 showed the presence of low-frequency noise (LFN) roughly below 10 Hz, and the presence, during the September 4 observation, of quasi-periodic oscillations (QPOs) with a strength of about 3 percent. The LFN strength was seen to increase from 2-6 percent for photon energies increasing from 2-10 keV. No correlation is found between the occurrence of QPOs and source intensity. It is suggested that the QPOs detected in GX3+1 may be similar in nature to the 5.6-Hz QPOs observed in Cyg X-2 which are also not observed in the horizontal branch.
An analysis of 12 hours of continuous EXOSAT observations of 4U 1746-37 in NGC 6441, obtained between September 9 and 10, 1985, puts constraints on its mass-radius relation. The spectral shape of the persistent X-ray flux is well fitted by a thermal bremsstrahlung spectrum, with a kT temperature varying between 7.5 and 9.8 keV, and an interstellar column density of about 2.8 x 10 to the 21st/sq cm. Two X-ray bursts noted during the observation are found to be type I, with radius expansion of the neutron star photosphere during their peaks. If the neutron star has a canonical mass of 1.4 solar masses, the observed burst fluxes must be reduced as a result of anisotropy of the burst by a factor of at least 1.3-2.2. The present results are consistent with the very soft and intermediately stiff equations of state.
In a series of seven Exosat observations, the bimodal spectral behavior and the quasi-periodic oscillation (QPO)/red noise properties of GX 5-1 show a strict correlation. In one of the two spectral states (characterized by a 'horizontal branch' in the hardness-intensity diagram), strong 20-40 Hz QPO and red noise below about 60 Hz were always present. In the other ('normal branch'), no QPO between 6 and 60 Hz or red noise above 1 Hz were detected, but there was an indication for weak QPO near 5 Hz. In both states 'very low frequency noise' (VLFN) is detected below 0.1 Hz which has a power-law shape and and which extends down to the lowest observed frequencies (0.0001 Hz). The VLFN is probably not directly related to the QPO. The results are compared to those on Sco X-1 and Cyg X-2 and it is concluded that, although all three sources show bimodal spectral and QPO/red noise behavior, there is a qualitative difference between GX 5-1 and Cyg X-2 on one hand and Sco X-1 on the other.
From the spectral evolution of seven X-ray bursts observed from 4U/MXB 1820-30 (Haberl et al., 1986) during the phases of photospheric expansion and cooling, constraints are derived on the mass-radius relation of the neutron star in this system. Taking into account the uncertainty in the source distance and the possible anisotropy of the burst emission. If during bursts with radius expansion the outer layers of the neutron star are of a hydrogen-poor composition, none of the current equations of state of high-density matter is excluded. For a hydrogen-rich composition, only very soft equations of state are allowed.
A spectral analysis of the persistent X-ray flux from the X-ray burst source 4 U/MXB 1636-53, which was continuously observed with EXOSAT for about 25 h in May, 1984, is presented. It is found that no standard one-component spectral model provides an acceptable representation of all the data. However, almost all 'complex one-component' and 'two-component' models do accurately fit the data. A blackbody plus a 'multicolor' disk spectrum does not fit these data. In addition, some 'two-component' spectral models can be fitted to the data with more than one set of different parameter values. Based on these results, it is suggested that the various physical scenarios that have been proposed to interpret 'two-component' spectra of low-mass X-ray binaries may be based on mathematical parameterizations which may not provide unique representations of the data. If this is the case, the reported 'two-component' mathematical fits may not have any true physical significance.
The analysis of six X-ray bursts from 4U/MXB 1820-30 (one observed in May 1975, and five in March 1976) is presented. In all cases, photospheric contraction during burst decay was observed. The contraction manifests itself as a distinct hardening of the burst spectra, a phenomenon which was previously noted but misinterpreted. The photospheric expansion which must have preceded the contraction takes place in less than one second. The bolometric burst peak luminosities are all near the Eddington limit. This burst source has the best-known distance of all burst sources. These results are therefore of particular interest because of the many reports of super-Eddington burst peak luminosities from sources whose distances are not so well known.
During a one day Exosat observation eight X-ray bursts from 4U/MXB 1636-53 are observed. Four of these were very unusual. Their peak fluxes were relatively low, and they showed a distinct double peak in their bolometric flux profiles. These new double-peaked bursts are unexplained by presently available models of X-ray bursts. It is possible that the energy release in these bursts proceeds in two 'steps'. The burst profiles are not the result of an expansion and subsequent contraction of the photosphere of the neutron star. Thus, they are very different from previously observed bursts which do show a double peak in certain energy ranges but not in their bolometric flux profiles; these are satisfactorily explained in terms of photospheric radius expansion and contraction. The anticorrelation between the apparent blackbody radius and blackbody temperature is discussed in terms of the nonPlanckian character of burst spectra and it is concluded that the model calculations reported by London, Taam, and Howard in 1984 give a reasonable first-order description of the observed apparent radius changes in X-ray bursts.
The X-ray flux of the bright galactic bulge source GX5 - 1 shows intensity-dependent quasi-periodic oscillations between 20 and 40 Hz, appearing as a broad peak in the power spectrum whose centroid frequency, width, and integrated excess power strongly depend on the source intensity. The strength and steepness of low-frequency noise present in the power spectra below 15 Hz also depend on the source intensity. No evidence is found for coherent X-ray pulsations between 0.5 and 2000 Hz. Possible mechanisms to explain these new phenomena are discussed.
Since approximately 1979, it is known that galactic bulge X-ray sources are low-mass binary systems, and that the X-ray bursters are a subset of the bulge sources. Grindlay and Hertz (1983) reported on the close proximity (in four of twelve cases) of an X-ray burst source and a normal star. The two authors believe that these alignments are not coincidental. They suggest that the normal star is surviving giant in a disrupted globular cluster core, of which the X-ray source would also be a member. The present investigation has the objective to show that Grindlay and Hertz made an error in their statistical analysis of a factor of approximately 10. It is, thus, believed that there is insufficient evidence for the conclusion reported by Grindlay and Hertz. On the basis of the present investigation, it is concluded that there exists at present no statistically significant observational evidence that X-ray bursters, presently located outside globular clusters, were formed in globular clusters.
The results of observations of 15 and 26 optical bursts detected during June and July 1979 and June and July 1980, respectively, from the X-ray burst source 4U/MXB 1636-53 in a 'white light' passband are presented. The maximum burst fluxes above the persistent optical flux and integrated burst fluxes are correlated, and it appears that the maximum optical and X-ray burst fluxes are related according to a power law, consistent with optical emission from blackbody reprocessing of X-rays. The approximately linear relationship between the integrated optical and X-ray burst fluxes argues against this simple picture. The correlation of burst fluxes with waiting time since the previous burst suggests that large optical bursts come after a long waiting time. Short burst intervals observed both for optical and X-radiation suggest that not all available nuclear fuel is consumed in the thermonuclear flash which gives rise to the X-ray burst.