XMM-Newton discovery of a possible cyclotron emission feature from the SFXT IGR J18483-0311
aa r X i v : . [ a s t r o - ph . H E ] J un XMM-Newton discovery of a possible cyclotronemission feature from the SFXT IGR J18483 − V. Sguera ∗ † INAF, Istituto di Astrofisica Spaziale e Fisica Cosmica, Via Gobetti 101, I-40129 Bologna, ItalyE-mail: [email protected]
L. Sidoli
INAF, Istituto di Astrofisica Spaziale e Fisica Cosmica, Via E. Bassini 15, I-20133 Milano, Italy
A. Bazzano
INAF, Istituto di Astrofisica Spaziale e Fisica Cosmica, Via Fosso del Cavaliere 100, I-00133Rome, Italy
L. Bassani
INAF, Istituto di Astrofisica Spaziale e Fisica Cosmica, Via Gobetti 101, I-40129 Bologna, Italy
M. Orlandini
INAF, Istituto di Astrofisica Spaziale e Fisica Cosmica, Via Gobetti 101, I-40129 Bologna, Italy
We report the results from an archival
XMM-Newton observation of the Supergiant Fast X-rayTransient (SFXT) IGR J18483 − × erg s − ) is the lowest ever reported in the literature, it is best fitted byan absorbed black body model yielding parameters consistent with previous measurements. Inaddition, we find evidence of an emission line feature at ∼ ∼ × G. A possible hint of thefirst harmonic is also found. If firmly confirmed by future longer X-ray observations, this wouldbe the first detection ever of a cyclotron feature in the X-ray spectrum of a SFXT, with importantimplications on theoretical models. ∗ Speaker. † The authors acknowledge the ASI financial support via grant ASI-INAF I/033/10/0 and I/009/10/0. c (cid:13) Copyright owned by the author(s) under the terms of the Creative Commons Attribution-NonCommercial-ShareAlikeLicence. http://pos.sissa.it/ ossible cyclotron emission feature from the SFXT IGR J18483 − V. Sguera
1. Introduction
During the last few years, the INTEGRAL satellite has played a key role in discovering apreviously unrecognized class of high-mass X-ray binaries which display a very unusual and in-triguing fast X-ray transient behaviour and host a massive OB supergiant star as companion donor:the so-called Supergiant Fast X-ray Transients (SFXTs; Sguera et al. 2005, 2006; Negueruela etal. 2006). To date ∼
10 firm SFXTs are known (plus the same number of candidates), they arecharacterized by long periods of low X-ray activity level (with luminosities values or upper limitsin the range 10 –10 erg s − ), interspersed with short, strong flares lasting from a few hours tono more than a few days and reaching peak luminosities of 10 –10 erg s − . The broad bandX-ray spectra of SFXTs are very similar to those of accreting X-ray pulsars (i.e. flat power lawbelow 10 keV and cut-off at 10–30 keV) strongly suggesting that they host a neutron star as well.Indeed X-ray pulsations have been detected in four SFXTs with spin periods ranging from 5 to 230seconds.The physical reasons behind the intriguing SFXTs behaviour are still not explained and severaltheoretical mechanisms have been proposed in the literature (see Sidoli 2009 for a review). All suchmodels are based on the general consensus that the supergiant wind is highly inhomogeneus andstructured (clumpy), however they can be mainly divided into two major groups i) models invokingspherically symmetric/anisotropic clumpy winds ii) models invoking gated mechanisms able tostop/allow the accretion onto a highly magnetized (B ∼ G, i.e. magnetar) and slow (P ≥ − ∼ ∼ × erg s − and ∼ ∼ XMM-Newton observation of IGR J18483 − XMM-Newton observation and results An XMM-Newton observation of the source, performed on 12 October 2006 for a total ex-posure of ∼
20 ks, was fortunately timed catching it right during the apastron passage at f =0.52.However, after rejecting time intervals affected by high background the net good exposure timereduced to 14.4 ks. The source net count rates (1–10 keV) are 0.097 ± .
003 counts s − (pn),0.035 ± .
002 counts s − (MOS1), and 0.027 ± .
001 counts s − (MOS2). Since the spectroscopywith the MOS cameras does not provide any improvement with respect to the analysis of the pnspectrum alone, in the following we report only on the source EPIC pn results.We first fit the 0.5–10 keV EPIC-pn spectrum of IGR J18483 − H = 7.8 + . − . × cm − , G =2.4 ± c n =1.46 for 52 d.o.f.This motivates us to investigate alternative spectral models. The best fit was achieved by using an2 ossible cyclotron emission feature from the SFXT IGR J18483 − V. Sguera − . r m a li z ed c oun t s / s e c / k e V − . . . r e s i dua l s channel energy (keV) Figure 1:
Data-to-model (absorbed black body) and corresponding residuals of the 0.5–10 keV EPIC-pnspectrum of IGR J18483 − − × − × − no r m a li z ed c oun t s / s e c / k e V − − − r e s i dua l s channel energy (keV) Figure 2:
Data-to-model (power law) and corresponding residuals of the 0.5–10 keV EPIC-pn backgroundspectrum of IGR J18483 − absorbed thermal black body ( c n =1.17, 52 d.o.f.) yielding spectral parameters (N H =3.4 + . − . × cm − , kT=1.35 ± × − erg cm − s − which translates into a X-ray luminosity of 1.3 × erg s − by assuming a distanceof the optical counterpart of 3 kpc. This is the lowest X-ray state of the source ever reportedin the literature, compatible with the 3 s upper limit measured by Swift/XRT in the same energyband (Romano et al. 2010). The detection of X-ray pulsations during this state (Giunta et al.2009) strongly suggest that it is very likely due to accretion onto the neutron star even if at muchlower rate than that during the outbursts activity. Above 20 keV, the lowest hard X-ray state of thesource has been measured by IBIS/ISGRI and it is about one order of magnitude higher (Sgueraet al. 2010). Although the absorbed black body model is a reasonable statistical description of3 ossible cyclotron emission feature from the SFXT IGR J18483 − V. Sguera
Figure 3:
EPIC-PN image (0.5-10 keV) of the field of IGR J18483 − the continuum, the ratio of data to model clearly show an excess around ∼ ∼ s ) according to a F-test ( c n =0.97, 49 d.o.f), it has an energycentroid of ∼ + . − . × − photon cm − s − (uncertainties givenat 68% confidence level) for a significance detection of ∼ s . We are aware that the F-test is not agood and proper measure of the actual significance of such spectral line feature (see Protassov et al.2002), however it could give an indication and to this aim we point out that the obtained low F-testprobability value should make the detection of the line stable against mistakes in the calculationof its significance. In addition, we also performed a Run Test which can provide useful additionalinformation. In fact, the Run Test (Barlow 1989) can be used to check a randomness hypothesisfor a two-valued data sequence or whether a function fits well to a data set or not. We performeda Run test to the residuals in Fig. 1 to test the randomness of their distribution and determine ifthere are any patterns or trends. As result, we obtained a chance probability equal to about 1% thatthe pattern under analysis has been generated by a random process. Such reasonably low valueprovides a statistical evidence that the pattern of residuals in Fig. 1 was not generated randomly.In order to rule out an artificial nature of the spectral emission line due to instrumental effects,we checked that the response matrix did not introduce any strong feature around 3.3 keV and notethat no uncalibrated instrumental features or edges are expected or known close to the same energy.In addition the inspection of the background spectrum, whose spectral shape is best fitted by a hardpower law ( c n =0.6, 7 d.o.f), did not reveal any spectral feature around 3.3 keV as clearly visible inFig. 2. We also note that the background spectrum was extracted from a region of 40 arcsecondsradius in the same CCD as IGR J18483 − ) we found that two lines from highly ionized Argon are expectedclose to ∼ http://cxc.harvard.edu/atomdb/WebGUIDE/index.html ossible cyclotron emission feature from the SFXT IGR J18483 − V. Sguera − − − P ho t on s c m − s − k e V − Energy (keV)
Figure 4:
Unfolded EPIC-pn spectrum (0.2–10 keV) best fit with an absorbed black body model plus thefundamental cyclotron line at 3.28 keV and its first harmonic. require an unexplainable high overabundance of Ar which is not obvious in the neutron star atmo-sphere or in its surrounding ambient, ii) it would not explain why only the highly ionized Ar lineis observed and no other lines are seen from more abundant elements, iii) it would not justify thelack of a similar spectral feature in any of the many known accreting X-ray pulsar in HMXBs forwhich much higher signal to noise X-ray spectra are available. Therefore, we explored an alterna-tive physical explanation in terms of electron cyclotron emission from an X-ray pulsar accretingat a low rate, as predicted by Nelson et al. (1995,1993). Specifically, these authors predicted thepossible detection of electron cyclotron emission lines in the X-ray spectra of magnetized and tran-sient X-ray pulsars during their low luminosity quiescent state (i.e L x ≤ erg s − ). The energyline center is expected to peak at energies in the range ∼ − ∼ × G (if the form-ing region is situated far above the neutron star polar cap) or alternatively ∼ × G (if it issituated close to the neutron star surface, i.e. at the base of the accretion column, and so affectedby a gravitational redshift of z=0.3)In addition, the possible detection of eventual harmonics has been also investigated. The firstharmonic line is expected around ∼ ∼ ∼ ∼
94% significance level of confidence ( ∼ s )according to a simple F-test, its energy centroid and intensity are ∼ + . − . × − photon cm − s − (uncertainties given at 68% confidence level) for a significance detection of ∼ s .5 ossible cyclotron emission feature from the SFXT IGR J18483 − V. Sguera
Since no atomic transition lines are expected close to ∼ − XMM-Newton could have possibly favoured the detection of the putative cyclotronemission line. Unfortunately, we can go no further on the above issues because the availableexposure time and statistics of the data prevent us from a more detailed investigation. Longer X-ray observations of IGR J18483 − XMM-Newton , for example, are strongly needed inorder to achieve a higher signal to noise ratio. This would allow us a much deeper investigation, inorder to support or reject our proposed interpretation in terms of electron cyclotron emission line. Iffirmly confirmed by a future longer
XMM-Newton observation, this would be the first detection everof a cyclotron feature in the X-ray spectrum of a SFXT. Its implications are very important: it willhelp in discriminating between different theoretical models proposed in the literature to physicallyexplain the outbursts mechanisms at work in SFXTs, i.e. involving highly magnetized (B ≥ G)or lower magnetized neutron stars (B ∼ G).
References [1] Barlow, R.J., 1989, The Manchester Physics Series, New York, Wiley[2] Giunta, A. et al., 2009, MNRAS, 399, 744[3] Negueruela, I. et al., 2005, ESA SP-604, 165[4] Nelson, R. W. et al., 1995, ApJ, 438L, 99[5] Nelson, R. W. et al., 1993, ApJ, 418, 874[6] Protassov R. et al., 2002, ApJ, 571, 545[7] Rahoui, F., Chaty, S., 2008, A&A, 492, 63[8] Romano, P. et al., 2010, MNRAS, 401, 1564[9] Sguera, V. et al., 2005, A&A, 444, 221[10] Sguera, V. et al., 2006, ApJ, 646, 452[11] Sguera, V. et al., 2007, A&A, 467, 249[12] Sguera, V. et al., 2010, MNRAS, 402L, 49S[13] Sidoli, L. 2009, AdSpR, 43, 1464[1] Barlow, R.J., 1989, The Manchester Physics Series, New York, Wiley[2] Giunta, A. et al., 2009, MNRAS, 399, 744[3] Negueruela, I. et al., 2005, ESA SP-604, 165[4] Nelson, R. W. et al., 1995, ApJ, 438L, 99[5] Nelson, R. W. et al., 1993, ApJ, 418, 874[6] Protassov R. et al., 2002, ApJ, 571, 545[7] Rahoui, F., Chaty, S., 2008, A&A, 492, 63[8] Romano, P. et al., 2010, MNRAS, 401, 1564[9] Sguera, V. et al., 2005, A&A, 444, 221[10] Sguera, V. et al., 2006, ApJ, 646, 452[11] Sguera, V. et al., 2007, A&A, 467, 249[12] Sguera, V. et al., 2010, MNRAS, 402L, 49S[13] Sidoli, L. 2009, AdSpR, 43, 1464