fccjxxw.com
非常超级学习网 学习超级帮手
当前位置:首页 >> >>

The Low Quiescent X-Ray Luminosity of the Neutron Star Transient XTE J2123-058


D RAFT VERSION F EBRUARY 2, 2008
A Preprint typeset using LTEX style emulateapj v. 7/15/03

THE LOW QUIESCENT X-RAY LUMINOSITY OF THE NEUTRON STAR TRANSIENT XTE J2123–058
J OHN A. T OMSICK 1 , DAWN M. G ELINO 1 , J ULES P. H ALPERN 2 , P HILIP K AARET 3
Draft version February 2, 2008

arXiv:astro-ph/0404287v1 14 Apr 2004

ABSTRACT We report on the ?rst X-ray observations of the neutron star soft X-ray transient (SXT) XTE J2123–058 in quiescence, made by the Chandra X-ray Observatory and BeppoSAX, as well as contemporaneous optical observations. In 2002, the Chandra spectrum of XTE J2123–058 is consistent with a power-law model, or the combination of a blackbody plus a power-law, but it is not well-described by a pure blackbody. Using the interstellar value of NH , the power-law ?t gives Γ = 3.1+0.7 and indicates a 0.3–8 keV unabsorbed luminosity ?0.6 of (9+4 ) × 1031 (d/8.5 kpc)2 ergs s?1 (90% con?dence errors). Fits with models consisting of thermal plus ?3 power-law components indicate that the upper limit on the temperature of a 1.4 M⊙ , 10 km radius neutron star with a hydrogen atmosphere is kTeff < 66 eV, and the upper limit on the unabsorbed, bolometric luminosity is L∞ < 1.4 × 1032 ergs s?1 , assuming d = 8.5 kpc. Of the neutron star SXTs that exhibit short (<1 year) outbursts, including Aql X-1, 4U 1608–522, Cen X-4, and SAX J1810.8–2609, the lowest temperatures and luminosities are found for XTE J2123–058 and SAX J1810.8–2609. From the BeppoSAX observation of XTE J2123–058 in 2000, we obtained an upper limit on the 1–10 keV unabsorbed luminosity of 9 × 1032 ergs s?1 . Although this upper limit allows that the X-ray luminosity may have decreased between 2000 and 2002, that possibility is not supported by our contemporaneous R-band observations, which indicate that the optical ?ux increased signi?cantly. Motivated by the theory of deep crustal heating by Brown and co-workers, we characterize the outburst histories of the ?ve SXTs. The low quiescent luminosity for XTE J2123–058 is consistent with the theory of deep crustal heating without requiring enhanced neutron star cooling if the outburst recurrence time is > 70 years. ? Subject headings: accretion, accretion disks — stars: individual (XTE J2123–058, SAX J1810.8–2609) — stars: neutron — X-rays: stars
1. INTRODUCTION

Accreting neutron stars can be found in high-mass (HMXB) or low-mass (LMXB) X-ray binary systems. The majority of HMXBs have transient X-ray emission. Their outburst spectra are relatively hard and X-ray pulsations from these highly magnetized (B ? 1012 G) neutron stars are typically detected. A wide variety of X-ray behaviors are seen for neutron star LMXBs, but, in general, the lack of X-ray pulsations from most (but not all) of these systems, and the emission of type I X-ray bursts from some, suggest that they harbor neutron stars with relatively low magnetic ?eld strengths (B ? 108?9 G). During ≈ 33 years of X-ray observations, some sources (e.g., Sco X-1, Cyg X-2) have maintained luminosities approaching the Eddington limit of ≈ 1038 ergs s?1 , while others are able to maintain persistent luminosities several orders of magnitude lower (Wilson et al. 2003). In addition, there is a class of transient neutron star LMXBs for which the luminosity varies from a substantial fraction of Eddington to quiescent levels typically near 1032?33 ergs s?1 . In outburst, these systems have relatively soft spectra compared to the HMXBs, and are commonly grouped with black hole transients as soft X-ray transients (SXTs). In quiescence, most neutron star SXTs exhibit X-ray energy spectra with a component that is typically ?tted well by a blackbody, suggesting that the origin of this compo1 Center for Astrophysics and Space Sciences, Code 0424, University of California at San Diego, La Jolla, CA, 92093, USA (e-mail: jtomsick@ucsd.edu) 2 Columbia Astrophysics Laboratory, Columbia University, 550 West 120th Street, New York, NY 10027 3 Harvard-Smithsonian Center for Astrophysics, 60 Garden Street, Cambridge, MA, 02138, USA

nent is thermal emission from the surface of a cooling neutron star. Although a pure blackbody often provides a good ?t to the spectrum, unphysical neutron star radii near 1 km are inferred unless an atmosphere is modeled (Rutledge et al. 1999). In addition to the thermal component, the energy spectra often contain a second component that has a powerlaw shape. The brightest and best studied systems in this class, Cen X-4 and Aql X-1, usually display both components (Rutledge et al. 2001, 2002a; Campana & Stella 2003; Campana et al. 2004). However, other systems may be dominated by the thermal component, such as MXB 1659–29 (Wijnands et al. 2003b) and sources X-5 and X-7 in the globular cluster 47 Tucanae (Heinke et al. 2003), or by the powerlaw component, such as SAX J1808.4–3658 (Campana et al. 2002) and EXO 1745–248 (Wijnands et al. 2003a). Although theories for the thermal component, such as the deep crustal heating model of Brown, Bildsten & Rutledge (1998), are relatively well-developed and are being tested with observations, the origin of the power-law component is not understood beyond suggestions that it may be related to accretion onto the neutron star magnetosphere (Campana et al. 1998) or a putative pulsar wind colliding with infalling matter from the companion star (Tavani 1991). In addition to our lack of understanding of the power-law component, questions remain about the mass accretion rate in quiescence, the origin of rapid (100–10,000 s) variability (Rutledge et al. 2002a; Campana et al. 2004), and the origin of variability in the thermal component on longer time scales (Rutledge et al. 2002a). Another important question is if quiescent observational properties correlate with other known differences between neutron star SXTs, such as whether they are millisecond X-ray pulsars (during outbursts) or not, whether the systems are in the ?eld

2

J.A. Tomsick et al. ? 1′ ), and this source, which we call SAX J2122.8–0575, is also present in the LECS image. It is clear that SAX J2122.8– 0575 is not XTE J2123–058 as they are separated by 6′.6. We conclude that XTE J2123–058 was not detected during the BeppoSAX observation, and we derive an upper limit on its X-ray ?ux below. Finally, we obtained R-band images on three occasions close to the times of the X-ray observations. As shown in Table 1, we observed XTE J2123–058 using the 2.4 m Hiltner telescope of the MDM Observatory on 2000 July 24, about 2 months after the BeppoSAX observation. We also observed XTE J2123–058 with the Shane 3 m telescope of Lick Observatory about 2 months before the Chandra observation and again at MDM about 2 weeks after the Chandra observation. For both MDM runs, we used the same SITe 2048 × 2048 pixel, thinned, back-illuminated CCD with a spatial scale of 0′′.275 per 24 ?m pixel, and an R ?lter that is very close to Harris R. 4 At Lick, we used the Prime Focus Camera, with a SITe 2048 × 2048 pixel, thinned CCD with a spatial scale of 0′′.296 per 24 ?m pixel, and a Kron-Cousins R ?lter. In all, we obtained thirteen 600–700 s exposures (see Table 1), and we reduced the images using standard IRAF 5 routines. For XTE J2123–058, we carried out the photometry with the IRAF package phot and used two calibrated reference stars with R magnitudes of 19.47 and 19.51. In 2002, the conditions at Lick were photometric, and we observed Landolt (1992) standards to obtain the calibration. We note that this calibration is about 0.1 magnitudes brighter than the calibration previously obtained using MDM observations from 1998 reported in Tomsick et al. (1999).
3. ENERGY SPECTRUM AND SOURCE LUMINOSITY

or in globular clusters, and whether their X-ray outbursts are long (years to decades) or short (weeks to months). Here, we report on X-ray and optical observations of the ?eld neutron star SXT XTE J2123–058 taken during quiescence. XTE J2123–058 had its only detected X-ray outburst in 1998 June-August (Levine, Swank & Smith 1998; Tomsick et al. 1999), and we focus on observations made with Chandra, BeppoSAX, and optical telescopes 2–4 years after the outburst. During the outburst, the Rossi X-ray Timing Explorer (RXTE) detected type I X-ray bursts and a pair of kHz quasi-periodic oscillations (Homan et al. 1999; Tomsick et al. 1999), indicating that the system contains a rapidly rotating neutron star. However, coherent X-ray pulsations were not found. The 6 hr binary orbital period and the fact that the binary inclination of the system is relatively high were established from optical modulation and the presence of partial eclipses in the optical light curve (Tomsick et al. 1999; Soria, Wu & Galloway 1999; Zurita et al. 2000; Shahbaz et al. 2003). The high Galactic latitude (b = ?36? ) and low extinction have allowed for detailed optical studies of XTE J2123–058 in quiescence even though the source is rather faint at its relatively large distance (8.5 ± 2.5 kpc). The optical observations show that XTE J2123–058 consists of a K7 V star on or close to the main sequence and a neutron star for which mass determinations of 1.5 ± 0.3 M⊙ (Tomsick et al. 2001; Casares et al. 2002; Tomsick et al. 2002) and 1.04–1.56 M⊙ (Shahbaz et al. 2003) have been obtained. The focus of this paper is the ?rst X-ray study of XTE J2123–058 in quiescence.
2. OBSERVATIONS AND ANALYSIS

We observed XTE J2123–058 with Chandra on UT 2002 November 13 (ObsID 2709), using the Advanced CCD Imaging Spectrometer (ACIS) with the target placed on one of the back-illuminated ACIS chips (ACIS-S3). For our analysis, we used the “level 2” event list produced by the standard data processing with ASCDS version 6.9.2 using Calibration Data Base (CALDB) version 2.17. Light curves using counts from the full ?eld-of-view do not show any background ?ares, allowing us to use the data from the full 17,706 s integration. Using the Chandra Interactive Analysis of Observations (CIAO) version 3.0 software routine wavdetect (Freeman et al. 2002), we searched for sources on the S3 chip in the 0.3–8 keV energy band. We detected 22 sources with counts between 5 and 108 per source, using a detection threshold of 10?6 , including a 24 count source at R.A. = 21h 23m 14s.54, decl. = ?05? 47′ 53′′.2 (equinox J2000, uncertainty 0′′.6). This position is consistent with the target’s optical position (Tomsick et al. 1999), and we conclude that this source is the quiescent X-ray counterpart of XTE J2123– 058. We also report on a BeppoSAX observation of the XTE J2123–058 ?eld made on UT 2000 May 12–13. We produced 1–10 keV images using the data from the two Medium Energy Concentrator/Spectrometer (MECS) units that were operational during the observation (units 2 and 3). We also produced a 0.1–10 keV image using data from the Low Energy Concentrator/Spectrometer (LECS). We obtained a MECS exposure time of 46,340 s and a LECS exposure time of 17,080 s. To search for sources, we convolved each of the three images with a two-dimensional Gaussian with a width (σ) of 5 pixels (40′′ ) in both directions. Only one source was clearly detected in both of the MECS images at R.A. = 21h 22m 50s.6, decl. = ?05? 45′ 09′′ (equinox J2000, uncertainty

We used the CIAO software routine psextract to produce the ACIS energy spectrum for XTE J2123–058 and to create the appropriate instrument response matrix for the spectrum. The software used CALDB 2.23 to create the response matrix, and we included a correction for the time-dependence of the ACIS response. The source spectrum included counts from a circular region with a 5 pixel (about 2′′.5) radius, and we estimated the background level using counts from a source-free annulus around the target position. The source spectrum consists of 24 counts, and we estimate a background level of 0.7 counts in the extraction region. We produced a “light curve” of the source in six time bins of ≈ 3000 s. Each bin contains between 2 and 6 counts, which is consistent with a constant ?ux; however, the low count rate does not allow us to place tight constraints on the possible amplitude of variability. The small number of counts also indicates that χ2 statistics are not appropriate for spectral analysis, as the assumption of a Gaussian probability distribution in each spectral bin is not met. Thus, we carried out our spectral analysis by minimizing the Cash statistic (Cash 1979), which is appropriate in cases where the assumption of a Poisson probability distribution in each spectral bin is valid. We ?tted the spectrum using XSPEC 11.2. The spectra of other neutron star SXTs are typically welldescribed by a blackbody, a power-law, or both components with interstellar absorption. We began by ?tting the spec4 See http://www.astro.lsa.umich.edu/obs/mdm/technical/?lters for the exact transmission curve. 5 IRAF (Image Reduction and Analysis Facility) is distributed by the National Optical Astronomy Observatories, which are operated by the Association of Universities for Research in Astronomy, Inc., under cooperative agreement with the National Science Foundation.

XTE J2123–058 in Quiescence

3

10-2 Counts s-1 keV-1 10-3 10-4 10-5

a

E FE (keV cm-2 s-1)

b
10-6

10-7 1 Energy (keV) 10

F IG . 1.— Chandra/ACIS energy spectrum of XTE J2123–058. Top: folded through the detector response. Bottom: unfolded. In each panel, the solid line is a power-law ?t to the data (using Cash statistics as described in the text) with NH ?xed at the interstellar value. The power-law photon index is 3.1+0.7 , and the unabsorbed 0.3–8 keV luminosity is (9+4 ) × 1031 ergs s?1 . A ?3 ?0.6 blackbody model (dashed line), by itself, does not provide a good description of the spectrum.

trum for XTE J2123–058 with an absorbed power-law, and the results are given in Table 2. Once we found the best-?tted parameters by minimizing the Cash statistic, we determined the quality of the ?t and the 90% con?dence errors on the parameters by producing and ?tting 10,000 simulated spectra. We used the best-?tted parameters from the ?ts to the actual data as input to the simulations. Using simulations to determine the parameter errors is necessary only because of the small number of counts in the spectrum. We tested our method by producing spectra with 2–3 times as many counts, and we found that the errors produced by calculating changes in the Cash statistic (the standard technique) match the values we obtain using simulations. In addition to the power-law model, we used simulations to determine the errors on parameter for all other ?ts presented in this work. For the power-law model, we obtain NH = (7.0+40.0 ) × 1020 cm?2 for the column ?7.0 density and Γ = 3.1+2.8 for the photon index. The ?t using ?0.8 a blackbody model is signi?cantly worse than the power-law model as indicated by the fact that for 97% of the simulated spectra we obtained a better Cash statistic than the one obtained when ?tting the actual spectrum (compared to 69% for the power-law model). We re?tted the spectrum with the same spectral models but with NH ?xed to the interstellar column density. The interstellar NH comes from the AV measurement of Hynes et al. (2001), which gives NH = (6.6 ± 2.7) × 1020 cm?2 , and the total Galactic H I value of 5.7 × 1020 cm?2 from Dickey & Lockman (1990). We adopt a value of 6 × 1020 cm?2 . For the power-law alone, we obtain Γ = 3.1+0.7 and an ?0.6 unabsorbed 0.3–8 keV ?ux of (1.1+0.5 ) × 10?14 ergs cm?2 s?1 , ?0.4

which corresponds to a luminosity of (9+4 ) × 1031 ergs s?1 ?3 at a distance of 8.5 kpc. The best ?tted blackbody temperature is 240+60 eV, but the quality of the blackbody ?t is even ?50 worse (98.7%) with NH ?xed. Figure 1 shows the spectrum (rebinned for clarity, although the Cash ?ts are performed without rebinning) along with the best ?tted power-law and blackbody models with NH ?xed. It is clear that the curvature of the blackbody model is too large, underpredicting the observed spectrum below 0.7 keV and above 2 keV. This, along with the measurements of the ?t quality from the simulations, indicates that the spectrum is well-described by a power-law model or the combination of a blackbody and a power-law model, but it is not well-described by a blackbody alone. Although a power-law alone provides an adequate description of the ACIS spectrum, we performed additional ?ts to obtain limits on the temperature and luminosity of the putative thermal component that is expected to be emitted from the surface of the neutron star. We ?tted the spectrum with a model consisting of a power-law component along with thermal emission from an atmosphere composed of hydrogen. The latter component was modeled using the “Neutron Star Atmosphere” (NSA) model of Zavlin, Pavlov & Shibanov (1996). We assumed a neutron star mass of 1.4 M⊙ , which is consistent with the measured mass, and a radius of 10 km. With these parameters ?xed, the remaining free parameters in the NSA model are the temperature (kTeff ) and the distance. For XTE J2123–058, there are several arguments that lead to distance estimates in the range 5–15 kpc (Tomsick et al. 1999; Homan et al. 1999; Zurita et al. 2000), but the most reliable estimates come from optical observations of the source in quiescence. Previous estimates include 8 ± 3 kpc (Zurita et al. 2000), 8.5 ± 2.5 kpc (Tomsick et al. 2001), and 9.6 ± 1.3 kpc (Casares et al. 2002). Here, we adopt a range of 6–11 kpc. We performed the NSA plus power-law ?ts with the distance ?xed to three values spanning this range (6, 8.5, and 11 kpc). As shown in Table 3, the 90% con?dence upper limits on kTeff are 57, 66, and 73 eV, respectively, for these three distances. Us4 ing L∞ = 4πR2 σTeff (1 ? 2GM/Rc2 ), where R = 10 km and M = 1.4 M⊙ , upper limits on the unabsorbed luminosity from the NSA component as seen by a distant observer are 8.0 × 1031, 1.4 × 1032 and 2.1 × 1032 ergs s?1 for distances of 6, 8.5, and 11 kpc, respectively. While these values are bolometric luminosities, it should be noted that we cannot rule out the possibility that there is thermal emission at energies below the Chandra bandpass. Although XTE J2123–058 was not detected by BeppoSAX in 2000 May, we can calculate upper limits on its luminosity if we assume that the energy spectrum was similar to that measured by Chandra. To calculate the upper limits, we assume the simplest model of an absorbed power-law with a photon index of 3.1 and the interstellar column density. For the MECS (units 2 and 3 combined), we measure a 1–10 keV count rate of 1.684 ks?1 in a circle of radius 2′ centered on the XTE J2123–058 position. This is not signi?cantly higher than the expected background rate (using a blank-sky pointing) of 1.675 ks?1 . These values imply a 3-σ upper limit on the count rate from XTE J2123–058 of 0.676 ks?1 , which corresponds to an unabsorbed ?ux < 1.02 × 10?13 ergs cm?2 s?1 , and a 1– 10 keV luminosity < 9 × 1032 ergs s?1 for d = 8.5 kpc. For the LECS, the 0.1–1 keV count rate in a 2′ radius circle centered on XTE J2123–058 is 1.230 ks?1 , which is actually somewhat lower than the expected background rate of 1.466 ks?1 . Using the expected background rate, the 3-σ upper limit on the unabsorbed 0.1–1 keV ?ux is 1.75 × 10?12 ergs cm?2 s?1 , corre-

4

J.A. Tomsick et al.

sponding to a luminosity < 1.5 × 1034 ergs s?1 for d = 8.5 kpc.
4. OPTICAL RESULTS

Figure 2 shows the R magnitudes for XTE J2123–058 from three exposures taken in 2000 July at MDM, two exposures taken in 2002 September at Lick, and eight exposures taken in 2002 November at MDM. Previously, Shahbaz et al. (2003) reported on extensive R-band photometry of XTE J2123–058 taken between 1999 June and 2000 August. The R-band light curve was relatively stable at that time, showing orbital ellipsoidal modulations with a peak-to-peak amplitude of about 0.25 magnitudes. The dashed lines in Figure 2 indicate the range of the modulation in 1999–2000, which was 21.70 < R < 21.95. While our 2000 May measurements of 21.63 ± 0.07, 21.75 ± 0.08, and 21.57 ± 0.07 are consistent with the brighter end of this range, our later observations indicate that XTE J2123–058 brightened between 2000 and 2002. The range of R between 21.41 ± 0.07 and 21.01 ± 0.03 over 25% of the 6 hr binary orbital period during the 2002 November run indicates that the source continues to vary, but with a peak level that is at least 0.6–0.7 magnitudes brighter than the peak level of R = 21.7 measured by Shahbaz et al. (2003). As our observations do not cover the full binary orbit, it is not clear whether this change is due to the addition of a constant component to the light curve or if the shape of the modulations has changed. Although it would be useful to know the orbital phases corresponding to the 2002 data, the binary ephemeris is too uncertain to extrapolate to 2002 as discussed in Tomsick et al. (2002). In the future, it may be worthwhile to obtain an X-ray observation while measuring the optical or IR light curve for a full 6 hr binary orbit. As we are reporting the ?rst sensitive X-ray observations of XTE J2123–058 in quiescence, comparison can be made with a prior prediction for the quiescent X-ray luminosity of this source that was made by modeling quiescent R-band light curves in 1999 and 2000. Shahbaz et al. (2003) predicted an X-ray luminosity of ?1033 ergs s?1 , which is only slightly higher than our upper limit from the BeppoSAX observation made in 2000, but an order of magnitude higher than we observed with Chandra in 2002. While the X-ray observations would allow for the possibility that the quiescent X-ray luminosity decreased signi?cantly between 2000 and 2002, our optical observations made in 2000 and 2002 indicate that the R-band ?ux actually increased over this time. If the 2000 Rband light curve showed a signi?cant contribution from Xray heating, it is dif?cult to see how a drop in X-ray ?ux could lead to an increase in the optical, thus, it is possible that Shahbaz et al. (2003) over-estimated the contribution from Xray heating.
5. INTERPRETATION

21.0 21.2 21.4 21.6 21.8 22.0 749.860 749.865 749.870 749.875 HJD-2451000 (days)

R-magnitude

a

749.880

21.0 21.2 21.4 21.6 21.8 22.0 1529.72 1529.74 1529.76 1529.78 HJD-2451000 (days)

R-magnitude

b

21.0 21.2 21.4 21.6 21.8 22.0 1603.56 1603.58 1603.60 1603.62 HJD-2451000 (days)

R-magnitude

c

1603.64

F IG . 2.— R-band measurements of XTE J2123–058. (a) 2000 July at MDM Observatory. (b) 2002 September at Lick Observatory. (c) 2002 November at MDM Observatory. The dashed lines delimit the levels measured by Shahbaz et al. (2003) during 1999–2000. An increase in the R-band ?ux between 2000 and 2002 is apparent.

X-ray observations of neutron star SXTs in quiescence provide tests of theoretical models for the thermal component. According to the theory of deep crustal heating by Brown, Bildsten & Rutledge (1998), the temperature of the neutron star core is maintained by nuclear reactions in the deep crust that occur when the mass accretion rate is high during outburst. As the thermal time scale for the core is ?10,000 years (Colpi et al. 2001), the level of quiescent thermal emission is set by the average mass accretion rate over this time span according to Lq = 9 × 1032 ˙ <M>?11 ergs s?1 (see Equation 1 of Rutledge et al. 2002b), assuming 1.45 MeV of heat deposited in the crust per ac˙ creted nucleon (Haensel & Zdunik 1990). Here <M>?11 is

the mass accretion rate averaged over the thermal time scale of the core in units of 10?11 M⊙ yr?1 , and Lq is the quiescent bolometric luminosity. While the time-averaged mass accretion rate is only a predictor of Lq if the neutron star core has reached thermal equilibrium, the lifetimes of LMXB systems are much longer than the thermal time scale of the core, and it is expected that thermal equilibrium has been established for all or nearly all of the LMXBs. While the long thermal time scale for the core precludes appreciable changes in the core temperature during a single SXT outburst, the neutron star crust can be heated signi?cantly for the systems with longer outbursts (years to decades) so that the quiescent emission is determined by the evolution of the physical conditions in the crust (Rutledge et al. 2002b). However, for the systems with shorter outbursts (weeks to months), the properties of the quiescent thermal emission are primarily set by the conditions in the neutron star core (Brown, Bildsten & Rutledge 1998). For a system like XTE J2123–058 that has undergone one ≈ 40 day outburst, the deep crustal heating theory implies that quiescent thermal emission provides information about the neutron star core. Here, we compare the quiescent properties of XTE J2123–058 to similar systems. The comparison group includes ?eld neutron star SXTs with outbursts lasting less than 1 year for which quiescent X-ray observations have been reported. We only consider LMXB systems

XTE J2123–058 in Quiescence from which X-ray bursts have been detected, proving that the accreting object is a neutron star. We restrict the comparison group to ?eld systems because the outburst histories for most transients in globular clusters are uncertain due to source confusion in instruments with angular resolution worse than Chandra. Thus, our comparison group consists of the neutron star SXTs Aql X-1, 4U 1608–522, Cen X-4, and SAX J1810.8–2609. We also compare the quiescent properties of XTE J2123–058 to those of the millisecond X-ray pulsar SAX J1808.8–3658 as the quiescent X-ray luminosities of the two sources are comparable. Table 4 compares the parameters of the thermal component for the ?ve systems, including kTeff , R, L∞ , and the gravitational redshift parameter, g = 1 ? 2GM/Rc2, assuming that all the neutron star masses are 1.4 M⊙ . For the spectral ?ts to XTE J2123–058, we set R = 10 km, and Jonker, Wijnands & van der Klis (2004) make the same assumption for SAX J1810.8–2609. However, for the other sources, we use the best ?tted radii, and the parameters are taken from Rutledge et al. (1999, 2001, 2002a). The thermal component is detected for Cen X-4, Aql X-1, and 4U 1608– 522, and a power-law component is also required for Cen X4 and for some of the observations of Aql X-1. For XTE J2123–058 and SAX J1810.8–2609, the thermal component is not statistically required, and we report upper limits on the thermal component parameters for these two systems. Also, for all the sources we assume the best current distance measurement. The temperature upper limits for XTE J2123–058 (66 eV) and SAX J1810.8–2609 (72 eV) are about a factor of 2.5 lower than the 4U 1608–522 measurement of 170 eV, and the luminosity upper limits for XTE J2123–058 and SAX J1810.8–2609 are about a factor of 50 lower than the maximum quiescent luminosity measured for Aql X-1. The temperature and luminosity of Cen X-4 are intermediate between the hottest and most luminous sources (Aql X-1 and 4U 1608– 522) and the coolest and least luminous sources (XTE J2123– 058 and SAX J1810.8–2609). Here, we characterize the outburst histories of these ?ve sources to determine if they are related to the quiescent thermal properties as is expected in the Brown, Bildsten & Rutledge (1998) theory. From a review of SXT light curves covering the time period between 1969 and 1996 (Chen, Shrader & Livio 1997) and the RXTE All-Sky Monitor (ASM) covering 1996 to 2003, 4U 1608– 522 had 16 outbursts with a mean peak X-ray luminosity, ? Lpeak , of 2.5 × 1037 (d/3.6 kpc)2 ergs s?1 . Aql X-1 had 21 ? outbursts with Lpeak = 3.6 × 1037 (d/5 kpc)2 ergs s?1 . Cen X-4 ? had two outbursts (in 1969 and 1979) with Lpeak = 5.6 × 1037 2 ?1 (d/1.2 kpc) ergs s , and XTE J2123–058 had one outburst ? (in 1998) with Lpeak = 1.7 × 1037 (d/8.5 kpc)2 ergs s?1 . Like XTE J2123–058, SAX J1810.8–2609 had only one outburst, also in 1998. SAX J1810.8–2609 was discovered by Natalucci et al. (2000); and the only reported X-ray detections of this source in outburst were from BeppoSAX and ROSAT in 1998 March (Natalucci et al. 2000; Greiner et al. 1999). However, the ASM light curve shown in Figure 3c indicates that SAX J1810.8–2609 was in outburst by 1998 January 4 (MJD 50,817) and probably as early as 1997 December 18 (MJD 50,800). The dashed lines in Figure 3c mark the times of the BeppoSAX discovery at MJD 50,882–50,885 and the ROSAT observation at MJD 50,896, while it is clear that the source was considerably brighter prior to this. From

5

ASM s-1 (1.5-12 keV)

80 60 40 20 0

a. Aql X-1

2600

2650

2700 2750 2800 MJD-50000 (days)

2850

ASM s-1 (1.5-12 keV)

80 60 40 20 0

b. 4U 1608-522

2400

2450 2500 2550 MJD-50000 (days)

2600

ASM s-1 (1.5-12 keV)

10 5 0 -5 700

c. SAX J1810.8-2609

750

800 850 MJD-50000 (days)

900

950

ASM s-1 (1.5-12 keV)

10 5 0 -5

d. XTE J2123-058

900

950

1000 1050 MJD-50000 (days)

1100

F IG . 3.— All-Sky Monitor (ASM) light curves for the four neutron star SXTs in our comparison group that had outbursts during the RXTE lifetime. Each point represents the average ASM count rate over 1 day. One Crab ?ux equals 75 ASM counts s?1 . For SAX J1810.8–2609 (c), the vertical dashed lines mark the times of previously reported BeppoSAX and ROSAT observations. Aql X-1 and 4U 1608–522 have had many outbursts during the lifetime of RXTE, and the light curves in (a) and (b) are representative.

? the 1.5–12 keV ASM light curve, Lpeak = 1.0 × 1037 (d/4.9 2 ?1 kpc) ergs s . Although it has been suggested that SAX J1810.8–2609 had an unusually low outburst luminosity (Jonker, Wijnands & van der Klis 2004), the value we derive from the ASM measurements is comparable to the other neutron star SXTs. Figure 3 also shows the ASM light curve for the XTE J2123–058 outburst as well as sample outburst light curves for Aql X-1 and 4U 1608–522 (Cen X-4 has not

6

J.A. Tomsick et al. mechanism of enhanced cooling of the core would be necessary. The most often mentioned mechanism is neutrino cooling of the core due to the direct Urca process, and this mechanism requires a neutron star with a mass higher than 1.7– 1.8 M⊙(Colpi et al. 2001). The mass of the neutron star in XTE J2123–058 has been measured at 1.5 ± 0.3 M⊙ or 1.04– 1.56 M⊙, so that the mass would have to be at the upper end of the error range for the direct Urca process to operate. While we have been focusing primarily on the implications of the upper limit on the temperature and luminosity of the thermal component, it is notable that the power-law ?t to the spectrum of XTE J2123–058 also indicates that its X-ray luminosity, at (9+4 ) × 1031 ergs s?1 in the unabsorbed ?3 0.3–8 keV band, is among the lowest for neutron star SXTs. The quiescent luminosity of SAX J1810.8–2609, ≈ 1 × 1032 ergs s?1 (Jonker, Wijnands & van der Klis 2004), is essentially the same as for XTE J2123–058, and the only other system in this class known to have a lower quiescent X-ray luminosity, SAX J1808.4–3658, is also the only millisecond X-ray pulsar system for which there is a measurement of its quiescent X-ray spectrum. The XMM-Newton spectrum of SAX J1808.4–3658 is well described by a power-law with Γ = 1.5 and a 0.5–10 keV unabsorbed luminosity of 5 × 1031 ergs s?1 , and the upper limit on the contribution from a thermal component is 10% of the total X-ray luminosity (Campana et al. 2002). If the spectrum of XTE J2123–058 consists of only +0.7 a power-law component, then its Γ = 3.1?0.6 is quite a bit softer than that of SAX J1808.4–3658, and the difference would have to be explained. On the other hand, if the spectrum for XTE J2123–058 consists of a thermal component and a power-law, then its power-law could have the same slope as SAX J1808.4–3658. Concerning the SAX J1808.4–3658 thermal component, Campana et al. (2002) conclude that enhanced neutron star cooling is required to obtain a quiescent luminosity as low as the XMM-Newton upper limit implies. However, as mentioned above, such conclusions depend on the assumption that the recent outburst behavior is indicative of the average mass accretion rate over the past 10,000 years. In summary, our X-ray measurements of XTE J2123–058 in quiescence show that this system is one of the faintest and least luminous of the neutron star SXTs. A comparison of the XTE J2123–058 X-ray properties in outburst and quiescence to that of four other neutron star SXTs that exhibit short outbursts shows that the systems with a higher degree of outburst activity tend to have more luminous thermal components in quiescence, and this is in-line with predictions of the theory of deep crustal heating. For XTE J2123–058, the upper limit on the thermal luminosity is consistent with this theory without requiring enhanced neutron star cooling if the outburst recurrence time is >67 years, which is consistent with the known outburst history of this source. JAT would especially like to thank Keith Arnaud for his help with statistical techniques and XSPEC. JAT acknowledges useful conversations with Tariq Shahbaz and Rudy Wijnands. JAT and PK acknowledge partial support from Chandra awards GO2-3051X and GO3-4043X issued by the Chandra X-ray Observatory Center, which is operated by the Smithsonian Astrophysical Observatory for and on behalf of NASA under contract NAS8-39073. DMG acknowledges support from a CASS postdoctoral fellowship.

had an outburst during the RXTE lifetime). These light curves demonstrate that individual outbursts from different sources are similar in duration (typically 40–80 days) and overall shape, although it should be noted that exceptional outbursts do occur (Chen, Shrader & Livio 1997). For each of the neutron star SXTs, the time-averaged mass accretion rate over the past 33 years can be expressed as ? Lpeak Ntoutburst f ˙ ? < M >= = sLpeak N ?c2 (33yr) , (1)

where N is the number of outbursts, toutburst is the typical duration of an outburst from the source, f is a factor with a value less than 1.0 that accounts for the shape of the outburst light curve, ? is the fraction of the rest mass energy released from accreted matter, and s = toutburst f /?c2 (33yr). As the durations and light curve shapes are similar for the sources in our comparison group, s is approximately the same for these sources, and the time-averaged mass accretion rate over the ? past 33 years is proportional to Lpeak N. We estimate s using toutburst = 60 days and ? = 0.2, which is the value of ? that is typically assumed for accretion onto a neutron star (Rutledge et al. 2002b). Although the precise value of f depends on the exact shape of the outburst light curve, it can be approximated as the mean outburst ?ux divided by the peak ?ux. For the XTE J2123–058 ASM light curve, the mean ?ux over a 60 day period that includes the 1998 outburst is 41% of the peak level, and we use f = 0.4, giving s = 1.1 × 10?23 s2 cm?2 . Using this value of s for all ?ve sources and the values ? for Lpeak and the number of outbursts (N) given above, we calculate estimates of the time-averaged mass accretion rate over ˙ the past 33 years. The values for <M> are given in Table 4, and they show that this quantity is a relatively good predictor of L∞ . Aql X-1 and 4U 1608–522 have the the highest values ˙ of both <M> and L∞ , XTE J2123–058 and SAX J1810.8– 2609 have the lowest, and Cen X-4 is intermediate. The ratio ˙ of the value of <M> for Aql X-1 to that for XTE J2123– 058 is 45, which is consistent with their ratio of L∞ , which is > 38. Similarly, the Aql X-1 to SAX J1810.8–2609 ratio ˙ of <M> is 76, and this is consistent with their ratio of L∞ , which is > 27. While these measurements are in line with the Brown, Bildsten & Rutledge (1998) theory, a caveat is that we cannot be certain that X-ray ?ux history over the past 33 years re?ects the behavior over the last 10,000 years. We can use the Brown, Bildsten & Rutledge (1998) theory to predict the recurrence time for outbursts of XTE J2123– 058, assuming that the outbursts are similar. Using L = ˙ ?Mc2 and the expression relating Lq to the time-averaged ˙ mass accretion rate, Lq = 9 × 1032 <M>?11 ergs s?1 , one obtains (see also Wijnands et al. 2001) Lq = [toutburst /(toutburst + tq )](<Loutburst >/130). Here, tq is the average time the source spends in quiescence between outbursts, <Loutburst > is the average luminosity during an outburst, and toutburst is de?ned above. For XTE J2123–058, we used the RXTE/ASM light curve shown in Figure 3d to determine that the mean 1.5–12 keV count rate during the 40 days (toutburst ) between MJD 50,990 and MJD 51,030 was 3.2 s?1 , corresponding to <Loutburst >= 1.1 × 1037 ergs s?1 for d = 8.5 kpc. From the luminosity upper limit of Lq < 1.4 × 1032 ergs s?1 , we obtain a lower limit on the recurrence time, toutburst + tq , of 67 years. The known outburst history of XTE J2123–058, one outburst in 33 years, assuming that no outbursts were missed, is consistent with these long predicted recurrence times. However, if the recurrence time is in fact shorter than 67 years, then a

XTE J2123–058 in Quiescence

7

REFERENCES Brown, E. F., Bildsten, L., & Rutledge, R. E., 1998, ApJ, 504, L95 Campana, S., Colpi, M., Mereghetti, S., Stella, L., & Tavani, M., 1998, A&A Rev., 8, 279 Campana, S., Israel, G. L., Stella, L., Gastaldello, F., & Mereghetti, S., 2004, ApJ, 601, 474 Campana, S., & Stella, L., 2003, ApJ, 597, 474 Campana, S., et al., 2002, ApJ, 575, L15 Casares, J., Dubus, G., Shahbaz, T., Zurita, C., & Charles, P. A., 2002, MNRAS, 329, 29 Cash, W., 1979, ApJ, 228, 939 Chen, W., Shrader, C. R., & Livio, M., 1997, ApJ, 491, 312 Colpi, M., Geppert, U., Page, D., & Possenti, A., 2001, ApJ, 548, L175 Dickey, J. M., & Lockman, F. J., 1990, ARA&A, 28, 215 Freeman, P. E., Kashyap, V., Rosner, R., & Lamb, D. Q., 2002, ApJS, 138, 185 Greiner, J., Castro-Tirado, A. J., Boller, T., Duerbeck, H. W., Covino, S., Israel, G. L., Linden-V?rnle, M. J. D., & Otazu-Porter, X., 1999, MNRAS, 308, L17 Haensel, P., & Zdunik, J. L., 1990, A&A, 227, 431 Heinke, C. O., Grindlay, J. E., Lloyd, D. A., & Edmonds, P. D., 2003, ApJ, 588, 452 Homan, J., Méndez, M., Wijnands, R., van der Klis, M., & van Paradijs, J., 1999, ApJ, 513, L119 Hynes, R. I., Charles, P. A., Haswell, C. A., Casares, J., Zurita, C., & SerraRicart, M., 2001, MNRAS, 324, 180 Jonker, P. G., Wijnands, R., & van der Klis, M., 2004, MNRAS, 349, 94 Landolt, A. U., 1992, AJ, 104, 340 Levine, A., Swank, J., & Smith, E., 1998, IAU Circular, 6955 Natalucci, L., Bazzano, A., Cocchi, M., Ubertini, P., Heise, J., Kuulkers, E., in ’t Zand, J. J. M., & Smith, M. J. S., 2000, ApJ, 536, 891 Rutledge, R. E., Bildsten, L., Brown, E. F., Pavlov, G. G., & Zavlin, V. E., 1999, ApJ, 514, 945 Rutledge, R. E., Bildsten, L., Brown, E. F., Pavlov, G. G., & Zavlin, V. E., 2001, ApJ, 551, 921 Rutledge, R. E., Bildsten, L., Brown, E. F., Pavlov, G. G., & Zavlin, V. E., 2002a, ApJ, 577, 346 Rutledge, R. E., Bildsten, L., Brown, E. F., Pavlov, G. G., Zavlin, V. E., & Ushomirsky, G., 2002b, ApJ, 580, 413 Shahbaz, T., Zurita, C., Casares, J., Dubus, G., Charles, P. A., Wagner, R. M., & Ryan, E., 2003, ApJ, 585, 443 Soria, R., Wu, K., & Galloway, D. K., 1999, MNRAS, 309, 528 Tavani, M., 1991, ApJ, 379, L69 Tomsick, J. A., Halpern, J. P., Kemp, J., & Kaaret, P., 1999, ApJ, 521, 341 Tomsick, J. A., Heindl, W. A., Chakrabarty, D., Halpern, J. P., & Kaaret, P., 2001, ApJ, 559, L123 Tomsick, J. A., Heindl, W. A., Chakrabarty, D., & Kaaret, P., 2002, ApJ, 581, 570 Wijnands, R., Heinke, C. O., Pooley, D., Edmonds, P. D., Lewin, W. H. G., Grindlay, J. E., Jonker, P. G., & Miller, J. M., 2003a, astro-ph/0310144 Wijnands, R., Homan, J., Miller, J. M., & Lewin, W. H. G., 2003b, astro-ph/0310612 Wijnands, R., Miller, J. M., Markwardt, C., Lewin, W. H. G., & van der Klis, M., 2001, ApJ, 560, L159 Wilson, C. A., Patel, S. K., Kouveliotou, C., Jonker, P. G., van der Klis, M., Lewin, W. H. G., Belloni, T., & Méndez, M., 2003, ApJ, 596, 1220 Zavlin, V. E., Pavlov, G. G., & Shibanov, Y. A., 1996, A&A, 315, 141 Zurita, C., et al., 2000, MNRAS, 316, 137

8

J.A. Tomsick et al.
TABLE 1 XTE J2123–058 X-R AY AND O PTICAL O BSERVATIONS

Observatory

UT Date

Energy Band

Exposure Time (s)

BeppoSAX 2000 May 12–13 0.1 ? 10 keV 46,340a MDM 2000 July 24 Harris R 3 × 600 Lick 2002 September 11 Kron-Cousins R 2 × 700 Chandra 2002 November 13 0.3 ? 8 keV 17,706 MDM 2002 November 25 Harris R 8 × 600 a This is the MECS exposure time. The LECS exposure time is 17,080 s.

TABLE 2 S PECTRAL F ITS
Model (1020 NH a cm?2 ) Γ (10?14 Fpl b ergs cm?2 s?1 ) kTbb (eV) (1032
2 L/d8.5 c Fit ergs s?1 ) Qualityd

— — 0.69 1.2+31.9 pl 7.0+40.0 3.1+2.8 ?0.5 ?0.8 ?7.0 0.97 0.49+2.76 — — 260+50 bb 0.0+40.0 ?0.12 ?110 ?0.0 +3.67 0.66 0.9+160 100+130 0.70?0.48 bb+pl 8.0+38.0 2.5+2.9 ?0.7 ?60 ?2.3 ?8.0 — — 0.69 1.1+0.5 pl 6.0 3.1+0.7 ?0.4 ?0.6 0.987 0.63+0.22 bb 6.0 — — 240+60 ?0.20 ?50 +180 +2.1 +0.5 0.64 0.6+36.9 110?90 bb+pl 6.0 2.5?2.5 0.7?0.5 ?0.5 a Errors are 90% con?dence for all parameters. b 0.3–8 keV unabsorbed ?ux. At d = 8.5 kpc, F = 10?14 ergs cm?2 s?1 corresponds to 8.6 × 1031 ergs s?1 . pl c Bolometric luminosity at d = 8.5 kpc. d Fraction of 10,000 simulated spectra for which the C-statistic was better than the C-statistic obtained for the actual spectrum.

TABLE 3 NSA PLUS P OWER - LAW F ITS
d kTeff FNSA a L∞ b (kpc) (eV) (10?14 ergs cm?2 s?1 ) (1032 ergs s?1 ) Γ Fpl c (10?14 ergs cm?2 s?1 )

6 < 57 < 1.1 < 0.80 2.3+1.4 0.56+0.82 ?2.5 ?0.42 8.5 < 66 < 1.1 < 1.4 2.5+1.4 0.63+0.82 ?3.1 ?0.60 0.77+0.65 11 < 73 < 1.0 < 2.1 2.8+1.9 ?0.77 ?3.4 a 0.3–8 keV unabsorbed ?ux. b Bolometric luminosity as seen by a distant observer using 4πR2 σT 4 (1 ? 2GM/Rc2 ), where R = 10 km and M = 1.4 M . ⊙ eff c 0.3–8 keV unabsorbed ?ux. At d = 8.5 kpc, F = 10?14 ergs cm?2 s?1 corresponds to 8.6 × 1031 ergs s?1 . pl

TABLE 4 NSA PARAMETERS FOR S HORT O UTBURST F IELD NS SXT S AND X-R AY ACTIVITY
Sourcea d (kpc) kTeff (eV) R (km) L∞ b (ergs s?1 ) gc ˙ <M>d ( M⊙ yr?1 )

Aql X-1 5 94 ? 108e 13.2 (5.3 ? 9.4) × 1033 0.828 1.3 × 10?10 4U 1608–522 3.6 170 ± 30 9.4 5.3 × 1033 0.748 6.9 × 10?11 Cen X-4 1.2 76 ± 7 12.9 4.8 × 1032 0.824 1.9 × 10?11 SAX J1810.8–2609 4.9 < 72 10 < 2.0 × 1032 0.765 1.7 × 10?12 XTE J2123–058 8.5 < 66 10 < 1.4 × 1032 0.765 2.9 × 10?12 a Parameters for sources other than XTE J2123–058 are from Rutledge et al. (1999, 2001, 2002a) and Jonker, Wijnands & van der Klis (2004). b Bolometric luminosity as seen by a distant observer calculated using 4πR2 σT 4 g2 . eff cg = 1 ? 2GM/Rc2 assuming M = 1.4 M⊙ . d The time-averaged mass accretion rate over the past 33 years, estimated according to <M>= sL ˙ ?peak N, where s = 1.1 × 10?23 s2 cm?2 (see Equation 1). e The range of values found in three separate Chandra observations.


更多相关文章:

非常超级学习网 fccjxxw.com

copyright ©right 2010-2021。
非常超级学习网内容来自网络,如有侵犯请联系客服。zhit325@126.com|网站地图