Discovery of a 0.02 Hz QPO feature in the Transient X-ray Pulsar KS 1947+300
Marykutty James, Biswajit Paul, Jincy Devasia, Kavila Indulekha
aa r X i v : . [ a s t r o - ph . H E ] M a y Mon. Not. R. Astron. Soc. , 1– ?? (2009) Printed 22 October 2018 (MN L A TEX style file v2.2)
Discovery of a 0.02 Hz QPO feature in the Transient X-rayPulsar KS 1947+300
Marykutty James , ⋆ , Biswajit Paul , Jincy Devasia , and Kavila Indulekha School of Pure and Applied Physics, Mahatma Gandhi University, Priyadarshini Hills P.O, Kottayam-686560, Kerala, India Raman Research Institute, Sadashivnagar, C. V. Raman Avenue, Bangalore 560080, India
22 October 2018
ABSTRACT
We report the discovery of Quasi Periodic Oscillations (QPO) at 0.02 Hz in a transienthigh mass X-ray binary pulsar KS 1947+300 using
RXTE -PCA. The QPOs weredetected during May-June 2001, at the end of a long outburst. This is the 9th transientaccretion powered high magnetic field X-ray pulsar in which QPOs have been detectedand the QPO frequency of this source is lowest in this class of sources. The unusualfeature of this source is that though the outburst lasted for more than 100 days, theQPOs were detected only during the last few days of the outburst when the X-rayintensity had decayed to 1.6% of the peak intensity. The rms value of the QPO islarge, ∼ . ± .
0% with a slight positive correlation with energy. The detection ofQPOs and strong pulsations at a low luminosity level suggests that the magnetic fieldstrength of the neutron star is not as high as was predicted earlier on the basis of acorrelation between the spin-up torque and the X-ray luminosity.
Key words:
X-ray: Neutron Stars - X-ray Binaries: individual (KS 1947+300)
KS 1947+300 is an accretion powered X-ray pulsar whichwas first detected in June 1989 in observations made withthe TTM coded-mask imaging spectrometer aboard the
Kvant module of the
Mir Space Station (Skinner 1989).The 2-27 keV flux was 70 ±
10 mCrab. In two monthsthe flux from the source faded by a factor of seven. ItsX-ray spectrum could be described by a power law withphoton index Γ = 1 . ± .
31 and a hydrogen columndensity N H = (3 . ± . × (Borozdin et al. 1990).The coordinates of the source were determined to be :R.A.= 19 h m s .
2, Dec.= +30 ′ ′′ (Eq. 1950.0). InApril 1994, the Burst and Burst and Transient SourceExperiment (BATSE) aboard the
Compton Gamma RayObservatory (CGRO) detected 18.7 s pulsations from anX-ray source within a few degrees of KS 1947+300. Thenewly detected source GRO J1948+32 was later found tobe same as KS 1947+300 (Swank and Morgan 2000). Theoptical counterpart is a V=14 . Rossi X-ray TimingExplorer (RXTE) -ASM. The first outburst of this source ⋆ E-mail: [email protected] was detected by
RXTE -ASM in October 2000 (Levineand Corbet 2000). Following this the intensity declined,but the source became highly active again in November2000. The outburst reached its peak in February 2001 andslowly declined till June 2001. Based on data acquiredwith
RXTE-PCA during the 2000-2001 outburst, Gallowayet al. (2004) determined the orbital parameters of thebinary: the orbital period P orb = 40 . ± .
010 d, theprojected semi major axis a x sini = 137 ± e = 0 . ± . RXTE analysis of this source revealed an increase inpulse frequency at an unusually high rate giving evidencefor the first time for a glitch in an accretion powered pulsar(Galloway 2004). A broad band (0.1-100.0 keV) study ofthis source was first carried out with
BeppoSAX . Thisrevealed that the energy spectrum has three components - aComptonized component, a 0.6 keV blackbody component,and a narrow and weak iron emission line at 6.7 keV(Naik et al. 2006). The absorption column density mea-sured towards this source is low (4.0-5.0 × atoms cm − ).Quasi Periodic Oscillations (QPOs) in X-ray binariesare generally thought to be related to the rotation of the in-ner accretion disk (Paul & Rao 1998). Any inhomogeneousmatter distribution or blobs of material in the inner diskmay result in QPOs in the power spectrum. In the case ofaccretion powered X-ray pulsars this gives useful informa- c (cid:13) Marykutty James, Biswajit Paul, Jincy Devasia and Kavila Indulekha
Table 1.
List of PCA ObservationsYear Obs Ids No. of Total DurationsPointings (ks)2000 P50425 19 111.832001 P50068 30 280.1P50425 18 73.7P60402 53 162.82002 P70404 15 46.3 tion about the interaction between the accretion disk andthe neutron star magnetosphere. Black hole X-ray binariesand low magnetic field neutron stars show QPOs over a widerange of frequency from a few Hz to a few hundred Hz. Highmagnetic field neutron star systems show only low frequencyQPOs, in the range 10 mHz upto about 1 Hz. We have inves-tigated the timing properties of the transient X-ray pulsarKS 1947+300 using observations made with the RXTE-PCAand report here the discovery of a transient QPO feature inthis source.
KS 1947+300 was observed extensively by
RXTE during2000-2002.
RXTE carries three X-ray astronomy instru-ments. The All Sky Monitor (ASM) is sensitive to X-rayphotons between 1.5 and 12 keV. The Proportional CounterArray (PCA) consists of five xenon proportional counter de-tectors, sensitive in the energy range of 2-60 keV with aneffective area of 6500 cm at 6 keV. The High Energy Tim-ing Experiment (HEXTE) operates in the energy band of15-250 keV.In Figure 1 we show the one day averaged ASM lightcurve of this source during 2000-2002 in 1.5-12 keV energyband. In this period the source showed one large outburstthat lasted for about four months, and several smalleroutbursts. The first PCA observation of KS 1947+300 wasperformed in 2000 November 21. Subsequently, observationswere made every 2 to 3 days until June 18 around whenthe main outburst ended. Some more PCA observationswere carried out again in 2002, around a smaller outburst.A background subtracted lightcurve of the source madewith data from one of the RXTE PCA detectors, PCU2 isshown in Figure 2. We have used all PCA observations ofthis source available in the archive.Any contamination from other sources are negligible be-cause there is no known hard X-ray source within 2 degreefrom KS 1947+300. A total of 674 ks of useful data was ob-tained with the PCA from 135 observations. The details ofthe observations are given in Table 1. We extracted 2-60 keV light curves from the PCA obser-vations using the Standard-1 mode data which has a timeresolution of 0.125 s. To search for QPOs we created PowerDensity Spectra (PDS) using the FTOOL powspec for small
Figure 1.
One day averaged
RXTE -ASM light curve of KS1947+300 during 2000-2002 in the 1.5-12 keV energy band. Thevertical arrows indicate the times of the
RXTE /PCA observa-tions and the ’+’ signs indicate the period when the QPOs aredetected.
Figure 2.
Background subtracted light curve of KS 1947+300obtained with
RXTE -PCU2 during the 2001 outburst with a binsize of 800 s. The vertical arrows indicate the times of the QPOdetection. data segments of duration 2048 s. PDS from 5-10 consecu-tive segments were averaged to improve the detectability ofany QPO like feature. The PDS were normalized such thattheir integral gives the squared rms fractional variabilityand the white noise level was subtracted. A narrow peakat around 0 .
055 Hz corresponding to the spin frequency ofthe pulsar and several harmonics are seen in all the PDSgenerated from the data of all the PCA observations in2000-2002. In addition to the peaks due to the pulsations, aQPO feature is seen at 0 . ± . χ of 178 for110 degrees of freedom. The fit clearly showed the need for c (cid:13) , 1– ?? iscovery of a 0.02 Hz QPO feature in the Transient X-ray Pulsar KS 1947+300 Figure 3. T he power spectrum of KS1947+300 obtained duringthe peak of the outburst (top curve) and during the QPO detec-tion period are shown here. The top curve has been multiplied byan arbitrary factor. a third component around 0.2 Hz. With the addition of athird component, a gaussian, the χ was reduced to 119for 107 degrees of freedom. This improvement in χ by 59for the addition of one component is very significant. Theratio of the amplitude and the uncertainty of the gaussiancomponent indicates a 6 σ detection of the QPO feature.The peaks corresponding to the pulsations are shown inFigure 3, but these were not included while fitting thePDS continuum. From the fitted power spectrum shownin Figure 3, and after correcting for the background countrate, we calculated the rms value of the QPO to be quitelarge, 15 . ± . ν/F W HM ofthe 0.02 Hz QPO is 3.6, comparable to the quality factorof 4-10 in other HMXB pulsars. In PDS created with datasegments of shorter duration the QPO feature has poorsignal to noise ratio. However, by looking at the PDSaveraged over smaller segments we have verified that thebroad nature of the QPO seen in Figure 3 is intrinsic andit is not due to averaging of a narrow QPO feature with avariable frequency.We have also calculated the energy dependence of theQPO feature. We created power spectra in different energybands of 2.6-6.6, 6.6-11.0, 11.0-15.4 and 15.4-19.8 keV usingevent data from the PCA. There is marginal evidence for anincrease in the rms value with energy and is shown in Fig-ure 4. The PCA detectors have small effective area at higherenergy and we did not detect the QPO feature above 20 keV.To measure the X-ray flux during the period whenQPOs are detected, we have generated an X-ray spec-trum in 129 binned channels of the PCA observations. Amodel for the background spectrum was generated using pcabackest with appropriate background model providedby XTE guest Observer facility (GOF). We employed thecaldb version 1.0.2 and the appropriate response matrixwas generated for this observation using pcarsp version 11.7.The spectrum is fitted with an absorbed power lawand an iron emission line. The best fitted spectrum has Figure 4.
RMS fluctuation in the 0.02 Hz QPO is shown here asa function of energy. C oun t s s − c m − k e V −
105 20−202 ∆ S χ Energy (keV)
Figure 5.
The X-ray spectrum of KS 1947+300 is fitted withpower law, absorption and a narrow emission line. a photon index of 1.44 ± ± χ r value forthe spectrum is 1.7 for 44 degrees of freedom. The fluxin the 2-20 keV band during the time of the QPO detec-tion is 5.0 × − erg cm s − . The spectrum is shown inFigure 5 along with the best fit model and the residuals.The parameters of the best fitted model are given in Table 2. Table 2.
Best fit Spectral Parameters of KS 1947+300Parameter Value N H (10 cm − ) 2.43 ± ± ± ± ± ± (cid:13) , 1– ?? Marykutty James, Biswajit Paul, Jincy Devasia and Kavila Indulekha
We reported the detection of a low frequency QPO featurein the HMXB pulsar KS 1947+300 at ν QPO = 0 .
02 Hz.The power density spectra of accretion powered pulsarsconsist of narrow peaks corresponding to the spin fre-quency of the pulsar and its harmonics, accompaniedby aperiodic variabilities like broad bumps. QPOs havebeen detected in 16 accretion powered high magnetic fieldpulsars. These include many HMXBs and a few LMXBs;both transient and persistent. The list of transient andpersistent sources with QPO features are given in Table 3.There are only three LMXB sources in which QPOs havebeen detected. Table 3 summarizes the QPO detectionsin high magnetic field accreting pulsars, giving spin fre-quency ν s , the observed QPO frequency ν QPO , and ratioof the two. The frequency of the QPO feature, detectedin KS 1947+300 is lowest among the transient X-ray pulsars.Various models have been proposed to explain thesub-Hz QPO features in high mass X-ray binaries. Amongthe most popular models are the Magnetospheric BeatFrequency Model (MBFM; Alpar and Shaham 1985) andthe Keplerian Frequency Model (KFM; van der Klis et al.1987). In the KFM the QPOs arise from the modulationof the X-rays by inhomogeneities in the accretion disk,at the Keplerian frequency. In the MBFM, the QPOfrequency is the difference between the spin frequencyand the Keplerian frequency of the inner edge of theaccretion disk ν QPO = ν k − ν s (Shibazaki & Lamb 1987).Mass flow to the neutron star is expected to be stoppedat magnetospheric boundary by centrifugal inhibition ofaccretion if the Keplerian frequency at the magnetosphericboundary is less than the neutron star spin frequency(Stella et al. 1986). Among the sources listed in Table 3,the KFM is not applicable for some of the sources includingKS 1947+300 as the observed QPO frequency is smallerthan the pulse frequency. At larger mass accretion rate, theaccretion disk is expected to extend closer to the neutronstar, and therefore, in either model, a positive correlationis expected between the QPO centroid frequency and theX-ray intensity (Finger 1998). In some sources, the QPOfrequency is found to be quite constant with more than afactor of 10 variation in X-ray luminosity. In some of thesesources, it is believed that the inner disk origin may not beapplicable while in some other sources there are a varietyof other reasons for a lack of correlation (Raichur & Paul2008a,b).A positive energy dependence of the QPO rms, similarto that seen in KS 1947+300 is found in some sources (XTEJ1858+034, Paul & Rao 1998). This favours the MBFM overthe KFM, while the QPO rms is independent of the energyin the persistent X-ray pulsar Cen X-3 (Raichur & Paul2008a).With some exceptions (4U 1626-67; Kaur et al. 2008but see also Jain et al. 2009), the QPOs are transientphenomena in all types of pulsars. The persistent X-raypulsars do not show persistent QPOs (Raichur & Paul2008a). In the transient X-ray pulsars that show QPOs, thefeature is not detected in all the outbursts. For examplethe recent outbursts of the source A 0535+262 does not show the flux dependent QPO features that was observedwith BATSE during a large outburst in 1994 (Finger et al.1996). However, if QPOs are present during an outburst,it is usually present throughout the outburst (Finger et al.1996, Mukherjee et al. 2006). In the case of KS 1947+300,the QPO feature was not seen in 2000 and 2002 data. Itappeared very near the end of the 2001 outburst duringMJD 52052-52062. We also analysed the 2-20 keV spectrumduring the peak of the outburst and the flux measured is3 × − erg cm − s − . The QPO feature is seen to be presentwhen the X-ray flux had dropped to 1.6% of this peak value.Assuming MBFM for KS 1947+300, the Keplerian fre-quency of the inner disk corresponding to a 0.02 Hz oscil-lation is ν k = 0 .
02 + 0 .
053 = 0 . R BF M = (cid:18) GM π ( ν QPO + ν s ) (cid:19) / (1)The R BF M obtained for a mass of 1 . M ⊙ is 9 . × km.The X-ray flux of KS 1947+300 in the 2-20 keV band is 5 . × − erg cm − s − during the time of QPO detection. Thisflux amounts to an X-ray luminosity of 0 . × erg s − for a source distance of 10 kpc. Using a correlation betweenthe spin-up torque and the broad band X-ray luminosity ofthis source a very strong magnetic field of B = 2 . × Gwas inferred for the neutron star (Tsygankov & Lutovinov2005). Using the above, and canonical values for the stellarradius and mass of 10 km and 1.4 M ⊙ respectively, the sizeof the magnetosphere for a dipole field configuration can beestimated as (Ghosh & Lamb 1991) R m = 2 . × ( M/ . M ⊙ ) / × ( B/ . × G ) / × ( R/ cm ) / × ( L x / . × ergs s − ) − / km (2)The magnetospheric radius is obtained as 2 . × km.Although the magnetic field value is rather uncertain,such a strong magnetic field implies a magnetospheric ra-dius larger by a factor of 2.8 compared to the inner diskradius at which the QPOs are likely to be produced in theMBFM and by a factor of 2 compared to the corotationradius. This is true even if we assume a 50% bolometriccorrection, 12% larger distance (Kiziloglu et al. 2007) andthe neutron star mass and radius larger by 20% (Lattimer& Prakash 2007). If the magnetic field is so strong, the X-ray flux value during the QPO detection indicates that thesource should be in propeller regime. The presence of QPOsand the strong pulsations at the low flux value thus suggeststhat the magnetic field may not be as high as was suggestedby Tsygankov & Lutovinov (2005). On the other hand, foran X-ray luminosity of 0 . × erg s − d kpc if the ob-served QPOs are produced due to inhomogeneities near themagnetospheric radius, it implies a magnetic field strengthof about B = 4 × d kpc G for the neutron star. TheX-ray spectrum of KS 1947+300, however, is devoid of anycyclotron absorption line corresponding to such a magneticfield strength (Naik et al. 2006). c (cid:13) , 1–, 1–
053 = 0 . R BF M = (cid:18) GM π ( ν QPO + ν s ) (cid:19) / (1)The R BF M obtained for a mass of 1 . M ⊙ is 9 . × km.The X-ray flux of KS 1947+300 in the 2-20 keV band is 5 . × − erg cm − s − during the time of QPO detection. Thisflux amounts to an X-ray luminosity of 0 . × erg s − for a source distance of 10 kpc. Using a correlation betweenthe spin-up torque and the broad band X-ray luminosity ofthis source a very strong magnetic field of B = 2 . × Gwas inferred for the neutron star (Tsygankov & Lutovinov2005). Using the above, and canonical values for the stellarradius and mass of 10 km and 1.4 M ⊙ respectively, the sizeof the magnetosphere for a dipole field configuration can beestimated as (Ghosh & Lamb 1991) R m = 2 . × ( M/ . M ⊙ ) / × ( B/ . × G ) / × ( R/ cm ) / × ( L x / . × ergs s − ) − / km (2)The magnetospheric radius is obtained as 2 . × km.Although the magnetic field value is rather uncertain,such a strong magnetic field implies a magnetospheric ra-dius larger by a factor of 2.8 compared to the inner diskradius at which the QPOs are likely to be produced in theMBFM and by a factor of 2 compared to the corotationradius. This is true even if we assume a 50% bolometriccorrection, 12% larger distance (Kiziloglu et al. 2007) andthe neutron star mass and radius larger by 20% (Lattimer& Prakash 2007). If the magnetic field is so strong, the X-ray flux value during the QPO detection indicates that thesource should be in propeller regime. The presence of QPOsand the strong pulsations at the low flux value thus suggeststhat the magnetic field may not be as high as was suggestedby Tsygankov & Lutovinov (2005). On the other hand, foran X-ray luminosity of 0 . × erg s − d kpc if the ob-served QPOs are produced due to inhomogeneities near themagnetospheric radius, it implies a magnetic field strengthof about B = 4 × d kpc G for the neutron star. TheX-ray spectrum of KS 1947+300, however, is devoid of anycyclotron absorption line corresponding to such a magneticfield strength (Naik et al. 2006). c (cid:13) , 1–, 1– ?? iscovery of a 0.02 Hz QPO feature in the Transient X-ray Pulsar KS 1947+300 We thank an anonymous referee for many suggestions thathelped us to improve the paper. This research has madeuse of data obtained through the High Energy AstrophysicsScience Archive Research Center Online Service, providedby the NASA/Goddard Space Flight Center.
REFERENCES
Alpar, M. A., & Shaham, J., 1985, Nature, 316, 239Angelini, L., Stella, L., Parmar, A.N., 1989, ApJ, 346, 906Angelini, L., White, N.E., Stella, L., Parmar, A.N., 1991, ApJ,371, 332Borozdin, K., Gil’fanov, M., et al. 1990, SvAL, 16, 345BFinger, M. H., Wilson, R. B., & Harmon, B. A., 1996, ApJ, 459,288Finger, M. H., 1998, AdSpR, 22, 1007Galloway, D. K., Morgan, E. H., Levine, A.M., 2004, ApJ, 613,1164Ghosh, P., Lamb, F.K. 1991, in Neutron Stars Theory and Ob-servation, eds. J. Ventura and D. Pines, NATO ASI Series 43,363Inam, S. C., Baykal, A., Swank, J., Stark, M. J., 2004, ApJ, 616,463In’t Zand, J. J. M., Baykal, A., & Strohmayer, T.E., 1998, ApJ,496, 386Jain, Chetana., Paul, Biswajit., & Dutt, Anjan., 2009,arXiv:0906.4169[astro-ph.HE]Kaur, R., Paul, B., Raichur, H. & Sagar, R., 2007, ApJ, 660,1409Kaur, R., Paul, B., Kumar, B., & Sagar, R., 2008, ApJ, 676, 1184Kiziloglu, U., Baykal, A., & Kiziloglu, N., 2007, AN, 328, 142Lattimer, James M., & Prakash, Madappa 2007, PhR, 442, 109Levine, A., & Corbet, R., 2000, IAU, Circ., No.7523Mukherjee, U., Bapna, S., Raichur, H., Paul, B., & Jaaffrey, S.N. A., 2006, JApA, 27, 25Moon, D., Eikenberry, Stephen. S., 2001a, ApJ, 549, L225Moon, D., Eikenberry, Stephen. S., 2001b, ApJ, 552, L135Naik, S., Callanan, P.J., Paul, B., & Dotani, T., 2006, ApJ, 647,1293Negueruela, I., Israel, G. L., et al., 2003, A&A, 397, 739Paul, B., & Rao, A. R., 1998, A&A, 337, 815Raichur, Harsha, Paul, Biswajit, 2008a, ApJ, 685, 1109Raichur, Harsha, Paul, Biswajit, 2008b, MNRAS, 387, 439Shibazaki, N., & Lamb, F. K., 1987, ApJ, 318, 767Shinoda, K., Kii, T., et al., 1990, PASJ, 42, L27Skinner, G. K., 1989, IAUC, 4850Soong, Y.,& Swank, J.H., 1989, ESASP, 296, 617Stella, L., N.E. White, and R. Rosner, 1986, ApJ, 308, 669Swank, J.,& Morgan, E., 2000, IAUC, 7531Takeshima T., Dotani T., Mitsuda K., Nagase F., 1991, PASJ,43, L43Takeshima T., Dotani T., Mitsuda K., Nagase F., 1994, ApJ, 436,871Takeshima T., 1997, BAAS, 191, 111.04Tsygankov, S. S.; Lutovinov, A. A., 2005, Astronomy Letters, 31,88van der Klis, M., Stella, L., White, N., Jansen, F., & Parmar,A.N., 1987, ApJ, 316, 411Zhang, W., Morgan, E.H., et al. 1996, ApJ, 469, L29c (cid:13) , 1– ?? Marykutty James, Biswajit Paul, Jincy Devasia and Kavila Indulekha
Table 3.
List of QPO sourcesSource Type ν s ν QPO ν QPO / ν s Reference (mHz) (mHz) (mHz)Transient pulsarsKS 1947+300 HMXB 53 20 0.38 This workSAX J2103.5+4545 HMXB 2.79 44 15.77 1A0535+26 HMXB 9.7 50 5.15 2V0332+53 HMXB 229 51 0.223 34U 0115+63 HMXB 277 62 0.224 4XTE J1858+034 HMXB 4.53 110 24.3 5EXO 2030+375 HMXB 24 200 8.33 6XTE J0111.2-7317 HMXB 32 1270 39.68 7GRO J1744-28 LMXB 2100 20000 9.52 8Persistent pulsarsSMC X-1 HMXB 1410 10 0.0071 9Her X-1 LMXB 806 13 0.016 10LMC X-4 HMXB 74 0.65-20 0.0087-0.27 11Cen X-3 HMXB 207 35 0.17 124U 1626-67 LMXB 130 48 0.37 13X Per HMXB 1.2 54 45 144U 1907+09 HMXB 2.27 69 30.4 15References: (1) Inam et al. 2004; (2) Finger et al. 1996; (3) Takeshima et al. 1994; (4) Soong & Swank 1989; (5) Paul &Rao, 1998; (6) Angelini et al. 1989; (7) Kaur et al. 2007; (8) Zhang et al. 1996. (9) Angelini et al. 1991; (10) Moon et al.2001b; (11) Moon et al. 2001a; (12) Takeshima et al. 1991; (13) Shinoda et al. 1990; (14) Takeshima 1997; (15) In’t Zandet al. 1998; c (cid:13) , 1–, 1–