The XMM-Newton survey of the Small Magellanic Cloud: A new X-ray view of the symbiotic binary SMC3
R. Sturm, F. Haberl, J. Greiner, W. Pietsch, N. La Palombara, M. Ehle, M. Gilfanov, A. Udalski, S. Mereghetti, M. Filipović
aa r X i v : . [ a s t r o - ph . H E ] O c t Astronomy&Astrophysicsmanuscript no. aa˙xmm˙SMC3 c (cid:13)
ESO 2018November 21, 2018
The XMM-Newton survey of the Small Magellanic Cloud:A new X-ray view of the symbiotic binary SMC 3
R. Sturm , F. Haberl , J. Greiner , W. Pietsch , N. La Palombara , M. Ehle , M. Gilfanov , A. Udalski , S.Mereghetti , and M. Filipovi´c Max-Planck-Institut f¨ur extraterrestrische Physik, Giessenbachstraße, 85748 Garching, Germany INAF, Istituto di Astrofisica Spaziale e Fisica Cosmica Milano, via E. Bassini 15, 20133 Milano, Italy XMM-Newton Science Operations Centre, ESAC, ESA, PO Box 50727, 28080 Madrid, Spain Max-Planck-Institut f¨ur Astrophysik, Karl-Schwarzschild-Str.1, 85741 Garching, Germany;Space Research Institute, Russian Academy of Sciences, Profsoyuznaya 84 /
32, 117997 Moscow, Russia Warsaw University Observatory, Aleje Ujazdowskie 4, 00-478 Warsaw, Poland University of Western Sydney, Locked Bag 1797, Penrith South DC, NSW1797, AustraliaReceived 21 October 2010
ABSTRACT
Context.
The XMM-Newton survey of the Small Magellanic Cloud (SMC) was performed to study the population of X-ray sourcesin this neighbouring galaxy. During one of the observations, the symbiotic binary SMC 3 was found at its highest X-ray luminosityobserved until now.
Aims.
In SMC 3 wind accretion from a giant donor star onto a white dwarf is believed to cause steady hydrogen burning on the whitedwarf surface, making such systems candidates for supernova type Ia progenitors. It was suggested that the X-ray source is eclipsedevery ∼ Methods.
We present the ∼
20 year X-ray light curve of SMC 3 and study the spectral evolution as seen with XMM-Newton / EPIC-pnto investigate possible scenarios which can reproduce the high X-ray variability.
Results.
We did not find significant variations in the photo-electric absorption, as it would be expected during eclipse ingress andegress. Instead, the X-ray spectra from di ff erent intensity levels, when modelled by black-body emission, can be better explained byvariations either in normalisation (by a factor of ∼
50) or in temperature (kT between 24 eV and 34 eV). The light curve shows maximaand minima with slow transitions between them.
Conclusions.
To explain the gradual variations in the X-ray light curve and to avoid changes in absorption by neutral gas, a predom-inant part of the stellar wind must be ionised by the X-ray source. Compton scattering with variable electron column density (of theorder of 5 × cm − ) along the line of sight could then be responsible for the intensity changes. The X-ray variability of SMC 3 couldalso be caused by temperature changes in the hydrogen burning envelope of the white dwarf, an e ff ect which could even dominate ifthe stellar wind density is not su ffi ciently high. Key words. stars: individual: SMC3 – (stars:) binaries: symbiotic – X-rays: binaries – (stars:) white dwarfs – galaxies: individual:Small Magellanic Cloud
1. Introduction
The symbiotic star SMC 3 (Morgan 1992) in the SmallMagellanic Cloud (SMC) was discovered as super-soft source(SSS, Hasinger 1994) in X-rays during the ROSAT all sky sur-vey (Kahabka & Pietsch 1993). It is thought to be an interactingbinary system, consisting of a cool M0 giant and a hot whitedwarf (WD) in a wide orbit. In this model accretion from thestellar wind of the giant donor onto the WD leads to steady hy-drogen burning on the WD surface which powers the high X-rayluminosity (Kahabka & van den Heuvel 1997).A series of ROSAT observations covering ∼ ≥
80 in ROSAT PSPC count rate), which was explained byan eclipse of the WD by the donor star (Kahabka 2004). Thisscenario needs obscuration of the X-ray emission region by thedense stellar wind close to the giant to account for the shape andthe long duration of several months of the eclipse ingress andegress. An optical outburst between December 1980 and November1981 of up to 3 mag in the U band was reported by Morgan(1992). During this outburst no changes were detected in the Iband and therefore, its origin was assigned to the hot stellar com-ponent. It is not clear, if the non-detection of SMC 3 with theEinstein satellite was due to X-ray inactivity before the opticaloutburst or to insu ffi cient sensitivity (Kahabka 2004). The en-richment of nitrogen also suggests evidence for a thermonuclearevent (Vogel & Morgan 1994). Results from modelling multi-wavelength data of SMC 3 with non-LTE models under the as-sumption of a constant X-ray source were presented in Orio et al.(2007) and Jordan et al. (1996). Orio et al. (2007) found that thevariability cannot be caused by photo-electric absorption andsuggested a ”real” eclipse by the red giant.SMC 3 was in a very luminous state during observation offield number 13 of the XMM-Newton (Jansen et al. 2001) largeprogram SMC survey (Haberl & Pietsch 2008). This enablesspectral analysis of the EPIC-pn (Str¨uder et al. 2001) data withunprecedented statistical quality. Four spectra from di ff erent in- Sturm et al.: A new X-ray view of the symbiotic binary SMC 3 tensity states allow us now to study the spectral evolution of thehot component of SMC 3. We present the light curve startingwith the first ROSAT detection in 1990, to investigate the natureof the variability seen from this system.
2. Observations and data reduction
SMC 3 was serendipitously observed four times with XMM-Newton at o ff -axis angles between 8 ′ and 14 ′ . Table 1 lists somedetails of the observations with the EPIC instruments operatedin full-frame mode. In addition to the observation in Oct. 2009from the SMC large survey program, we analysed three obser-vations from 2006 and 2007 available in the archive. The firstobservation in March 2006 revealed SMC 3 in a high intensitystate, but su ff ered from very high background. These data wereused in the study of Orio et al. (2007). The two observationsin April and October 2007 showed the source at low intensity.The detection of the source in the later observation was noted byZezas & Orio (2008).To process the data, we used XMM-Newton SAS 10.0.0 with calibration files available until 17 June 2010, including thelatest refinement of the EPIC-pn energy redistribution. As a stan-dard, we selected good time intervals (GTIs) with an EPIC-pnbackground rate below 8 cts ks − arcmin − (single- and double-pixel events, 7 −
15 keV). However, for the observation in 2006the background was between 500 and 2500 cts ks − arcmin − and in order to retain any data, no background screening wasapplied. Since soft proton flares usually show a rather hard spec-trum, the contribution to the super-soft spectrum of SMC 3 wasstill acceptable (cf. Table 1). The SAS task eregionanalyse was used to determine circular extraction regions by optimisingthe signal to noise ratio, as shown in Fig. 1 and listed in Table 1.We ensured that the source extraction region had a distance of > ′′ to other detected sources. For the background extractionregion, we chose a circle on a point source free area on the sameCCD as the source. Since the MOS-spectra have lower statisti-cal quality by a factor of 10 for such soft spectra and to avoidcross calibration e ff ects between the EPIC instruments , we areconcentrating on the EPIC-pn spectra in this study. For the ex-traction of EPIC-pn spectra, we selected single-pixel events with FLAG = 0 . We binned the spectra to a minimum signal-to-noiseratio of 5 for each bin using the task specgroup .
3. Spectral analysis of the EPIC-pn data
We used xspec (Arnaud 1996) version 12.5.0x for spectral fit-ting. For all models, the Galactic photo-electric absorption wasfixed at a column density of N H , gal = × cm − and elementalabundances according to Wilms et al. (2000), whereas the SMCcolumn density was a free parameter with abundances at 0.2 forelements heavier than Helium. At first, we investigated the recentEPIC-pn spectrum of 2009, which has unprecedented statisticscompared to previous X-ray observations. For black-body emis-sion, we obtain a best-fit with χ / dof = /
74 with the best-fitparameters: N H , SMC = . + . − . × cm − , kT = . ± . L bol = . + . − . × ergs − . This luminosity is super-Eddington (and leading to a WDradius ∼ −
10 times larger than expected for a 1 M ⊙ WD),which is often caused by the black-body approximation (see Science Analysis Software (SAS), http: // xmm.esac.esa.int / sas / EPIC Calibration Status Document,http: // xmm2.esac.esa.int / external / xmm sw cal / calib / index.shtml D ec li n a t i on D ec li n a t i on Fig. 1.
Combined EPIC colour images of SMC 3 from the fourXMM-Newton observations. Red, green, and blue colours de-note X-ray intensities in the 0.2 − − − χ / dof = /
72 by includ-ing two Gaussian lines with fixed energy at 431 eV (N vi ) and500 eV (N vii ) and line widths fixed to 0. The equivalent widthsof 19 eV and 32 eV for the N vi and N vii lines, respectively,are physically plausible, but the residuals could at least partiallybe also caused by calibration uncertainties. Alternatively, allow-ing the oxygen abundance in the SMC absorption componentas free parameter also improved the fit ( χ / dof = / . + . − . times solar with anN H , SMC = . + . − . × cm − . A hard spectral component,which could possibly be caused by the wind-nebula, is not seenin the spectrum. Adding an apec plasma emission component tothe black-body emission, with fixed temperature of kT = EM = . × cm − .The derived values for the absorption N H , SMC are well belowthe total SMC absorption in the direction of SMC 3 ( ∼ × cm − ; see Stanimirovic et al. 1999). This suggests that the sym-biotic system is located on the near side of the large amount ofH I present in the SMC Bar.We also tested non-local thermal equilibrium models pro-vided by Thomas Rauch (Rauch & Werner 2010). We found thebest-fit for a pure He atmosphere with WD surface gravity logg = χ / dof = / = χ / dof = /
74. If the spec-trum in fact contains emission lines, this might be the reasonof the inferior fit of the non-LTE models which produce spectrawhich are dominated by absorption lines from the white dwarf http: // astro.uni-tuebingen.de / ∼ rauch / turm et al.: A new X-ray view of the symbiotic binary SMC 3 3 Table 1.
XMM-Newton EPIC-pn observations of SMC 3
ObsID Satellite Date Time Filter Net Exp Net cts. Bg ( a ) Net cts. Bg ( a ) R ( b )sc R ( b )bg Revolution (UT) [s] (0.2 − − ′′ ] [ ′′ ]0301170501 1149 2006 Mar 19 14:45-20:17 medium 10446 ( c ) ( a ) ratio of background count rate to source count rate in the same energy band. ( b ) radius of the source and background extraction region. ( c ) no GTI screening was applied. atmosphere. Since the black-body spectrum resulted in a betterfit, we decided to use this model in our further investigations.To study the spectral evolution of SMC 3, we fitted the EPIC-pn spectra of all four epochs simultaneously with a set of modelsbased on absorbed black-body emission. In each model only oneindividual parameter for each spectrum and two common param-eters for all spectra were allowed to vary (see below). Since thestatistical quality of the high-flux spectrum is far better than forthe other spectra, it also dominates the resulting χ . This leadsto relatively bad fits for the simple black-body models. Addingemission lines would improve the fits (see above). However,given the limited spectral resolution of the EPIC-pn instrument itis not clear if the lines have any physical meaning. Because wewere mainly interested in the evolution of spectral parameters,we decided to use the simplest model.Our first model (model 1 in Table 2) assumes temperatureand luminosity not varying with time, while the absorbing col-umn density can change with time. This corresponds to theeclipse model with varying absorption by the dense donor wind,as suggested by Kahabka (2004). This model gives an insu ffi -cient fit to the data (see Table 2). Although, the fit is statisticallydominated by the two high-flux spectra, the spectral shape of thelow-flux spectra cannot be reproduced by a high column densitywhich predicts much less flux at lowest energies.For model 2, we fixed the spectral shape (same temperatureand absorption) and allowed the luminosity to change (i.e. fittingindividual normalisations which corresponds to a variable size ofthe emission area). This fits the data much better (Table 2).An even better fit was achieved by our third model withvarying source temperature. The corresponding black-body lu-minosities were related to the temperature ( L bol ∝ T ). The re-sults for this model are again described in Table 2 and the indi-vidual spectra with the model fit are plotted in Fig. 2.
4. The X-ray light curve of SMC 3
To analyse the temporal behaviour of the system, we reconciledthe X-ray light curve of SMC 3 starting from the first detectionby ROSAT in 1990. To convert the ROSAT count rates, providedby Kahabka (2004), into fluxes, we simulated a ROSAT PSPCspectrum based on the spectral model derived from the simul-taneous fit with variable normalisation and the PSPC detectorresponse. We obtained a conversion factor of 1 . × − ergcm − cts − . All fluxes are computed for the 0.2 − C oun t s s − k e V − χ Channel energy (keV)
Fig. 2.
EPIC-pn spectra of SMC 3 together with the best-fitblack-body model 3 with variable temperature. No significantemission is seen above 0.7 keV.of .
2. Analogous, for the ROSAT HRI, this simulation yields aconversion factor of 4 . × − erg cm − cts − .We searched the Swift archive for observations coveringSMC 3. The source was detected in observation 00037787001with ∼ xselect and the e ff ectivearea file was created using xrtmkarf . The resulting spectrumcontains 135 net counts, which is insu ffi cient to distinguish be-tween the above described models. Thus we fitted only the nor-malisation and assumed the spectral shape according to the si-multaneous fit to the EPIC-pn spectra with variable normalisa-tion. This fit yields a flux of 2 . + . − . × − erg cm − s − .To derive XMM-Newton fluxes, we integrated the best-fitmodel with variable temperature as described above. The sourceflux during a Chandra observation in February 2003 was de-duced from the parameters of the best-fit black-body model re-ported by Orio et al. (2007).Figure 3 shows the X-ray light curve of SMC 3 since the firstdetection by ROSAT in 1990. By modelling the light curve withseveral eclipses, we realised, that (i) the transition from high tolow intensity occurs over a long time period and that (ii) the timescales of the duration of the high and low intensity intervals areof the same order. Thus, instead of eclipses the light curve mayalso be interpreted by several periodic outbursts. For demonstra-tion, the dashed line in Fig. 3 shows a fit with a sine function.To account for uncertainties in the flux conversion and cross cal- Sturm et al.: A new X-ray view of the symbiotic binary SMC 3
Table 2.
Results from the simultaneous black-body fit to the EPIC-pn spectra.
Model 1 Model 2 Model 3 T and L bol constant with time N H and T constant with time N H constant with timeN H variable L bol variable T variable, L bol , i = L bol , ( T i / T ) kT = (32 . ± .
5) eV N H = (0 . ± . × cm − N H = (0 . ± . × cm − L bol = (8 . ± . × erg s − kT = (33 . ± .
5) eV L bol , = (5 . ± . × erg s − N H , = (1 . ± . × cm − L bol , = (4 . ± . × erg s − kT = (32 . ± .
2) eVN H , = (6 . ± . × cm − L bol , = (0 . ± . × erg s − kT = (24 . ± .
3) eVN H , = (6 . ± . × cm − L bol , = (0 . ± . × erg s − kT = (25 . ± .
2) eVN H , = (0 . ± . × cm − L bol , = (6 . ± . × erg s − kT = (33 . ± .
4) eV χ / dof = / = . χ / dof = / = . χ / dof = / = . ibration between the di ff erent instruments we included a 20%systematic error on the flux. As best-fit ephemeris for the X-rayminimum we then obtain (90% confidence errors):MJD min , x = (49382 ± + N × (1634 ±
7) days . The relatively high flux measured in the last XMM-Newton ob-servation might suggest possible changes in the amplitude of themodulation.
5. MACHO and OGLE data
The OGLE-II (Udalski et al. 1997) and OGLE-III (Udalski et al.2008) I-band as well as the MACHO B-band light curves areshown in the lower two panels of Fig. 3. The calibrated MACHOlight curve was shifted in magnitude to match its average B mag-nitude with that measured by Zaritsky et al. (2002). In the I-bandof OGLE-II and the B-band of MACHO, Kahabka (2004) foundcorrelating quasi-periodic oscillations with periods around 110days which might be related to pulsations of the red giant star.These short variations are also present in the OGLE-III data.With the OGLE-III data, which cover a much longer time in-terval, we can now rule out a significant variation with the 1630day cycle as suggested by the X-ray light curve. As noted byKahabka (2004), the MACHO light curve shows a quasi si-nusoidal modulation with a period of ∼ min , B = (49242 ± + N × (1647 ±
24) days . To account for the short-term variations we added a systematicerror of 0.05 mag to the B-band magnitudes. Formally, the fits toX-ray and MACHO light curves indicate a phase shift of (140 ±
14) days (optical preceding the X-rays) while the periods agreewithin the errors. However, it should be noted that the MACHOlight curve covers only ∼
6. Discussion
SMC 3 was observed with XMM-Newton at four epochs, cover-ing the super-soft X-ray source twice at high and twice at lowintensity. A strong variation in the X-ray flux by a factor of morethan ∼
50 between minimum and maximum intensity is foundin the 0.2 − F . − . k e V ( − e r g s c m − s − ) I m ag ~ B m ag MJDROSAT PSPCROSAT HRI Chandra XMM−Newton SwiftOGLE II + IIIMACHO
Fig. 3.
The 0.2 − ∼
20 year coverage indicates ahigh regularity of the period with similar duration of high andlow intensity intervals. The regularity of the X-ray light curve,with meanwhile four observed minima, strongly supports the in-terpretation of the 4.5 year period as the orbital period of thebinary system. Assuming masses of 15 M ⊙ and 1 M ⊙ for the M-giant and the white dwarf, respectively, the orbital period impliesa semi-major axis of the binary system of 6.83 AU.We analysed the spectral evolution and found, that the vari-ability of the X-ray flux cannot be explained by photo-electricabsorption by neutral gas with varying column density. Toavoid the strongly energy-dependent attenuation of soft X-rays,Kahabka (2004) discussed absorption due to highly ionised gas.In this picture the strong X-ray source ionises the stellar windaround it. Compton scattering on free electrons would then re-duce the X-ray flux along the line of sight most e ffi cient whenlooking through the dense innermost regions near the M-star.This mimics variable intensity with little energy dependence (nosignificant change of spectral shape). Using a Compton scatter-ing model ( cabs in xspec ), instead of variable normalisation,would require a column density of > × cm − (completely turm et al.: A new X-ray view of the symbiotic binary SMC 3 5 ionised absorber) to reduce the X-ray intensity from maximumto minimum. In this picture scattering of X-rays into the line ofsight is neglected or at least assumed not to change significantlybetween the two states. Using the estimated mass, size and den-sity of the ionised wind region as given by Orio et al. (2007)yields a column density to its centre of 5.8 × cm − . This isa factor of ∼ and its stellar wind. An eclipse by thestar only would be short ( ∼
3% of the orbital period) with sharpingress and egress while the dense inner wind regions cause along gradual eclipse ingress (and egress) by increasing (decreas-ing) Compton scattering along the line of sight. The shape of thelight curve should then depend on the geometry of the binarysystem and the distribution of free electrons in the stellar wind.Our spectral analysis shows, that the X-ray variability canalternatively be dominated by temperature changes, varying be-tween 24 and 34 eV. Assuming a constant size of the emittingarea, this corresponds to a variation in L bol (for spherically sym-metric emission) by a factor of 4.3. The larger variation in ob-served instrumental count rates would then be caused by shiftingthe spectra with lower temperature out of the sensitive energyband of EPIC. A possible scenario might be an elliptical orbitof the white dwarf around the M-giant (or equatorial mass ejec-tion with inclined WD orbit), causing accretion at di ff erent rates.Variable accretion, even at low level, can lead to large temper-ature changes in the burning layer (Paczynski & Rudak 1980).Similar scenarios were used to explain X-ray variability in otherSSS (e.g. AG Dra, Greiner et al. 1996). However, we note, thatin those cases usually an anti-correlation of X-ray and opticalluminosity is observed, whereas in the case of SMC 3 these twoare clearly correlated.Assuming the same temperature and absorption (and nochange in Compton scattering) for the low and high intensityspectra, the inferred radii would be di ff erent by a factor of ∼ ff erence in L bol . In general forstable shell burning on the WD surface, an increase of the hydro-gen burning envelope (e.g. due to a higher accretion rate) leadsto an increase of both, temperature and radius (Fujimoto 1982).Increasing temperature and declining Compton scattering bothlead to an increasing X-ray luminosity. It depends on the orien-tation of the orbit with respect to the observer, how much thetwo e ff ects act in phase. Additional temperature variations maytherefore reduce or increase the amount of Compton scatteringrequired to explain the X-ray luminosity variations.Superimposed on the general long term variation in the X-ray light curve, we probably see e ff ects imposed by the donorstar. The ∼
110 days brightness variations seen in the I-band sug-gests changes in the stellar wind which can lead to variationsin the mass accretion rate onto the white dwarf. Since the 1630day X-ray period is not visible in the I-band, it is unlikely thatthis period is caused by the cool stellar component, which seemsto remain rather una ff ected by the process producing this varia-tion. In the B-band, both modulations are seen, the 1630 day pe-riod derived from the X-rays and the ∼
110 day variations whichcorrelate with the I mag (as already pointed out by Kahabka2004). Therefore, the cool companion star and the region wherethe X-ray emission is produced most likely both contribute to the B-band. If viewing e ff ects produce the variation in B in asimilar way as in the X-rays (by changing extinction) or if heat-ing of the cool star by the X-ray source is causing this varia-tion remains unclear: While Kahabka (2004) finds X-ray heat-ing insu ffi cient to account for the observed B-band modulation,Orio et al. (2007) discuss irradiation e ff ects influencing the massoutflow as very important.
7. Conclusions
The new X-ray data of SMC 3 show that two scenarios can qual-itatively explain the spectral evolution and the shape of the lightcurve. The evolution of the X-ray spectra is incompatible withchanging photo-electric absorption by neutral gas, but is consis-tent with energy-independent intensity and / or with temperaturevariations. As suggested before by Kahabka (2004), Comptonscattering in a predominantly ionised stellar wind could lead tothe observed intensity variations if the stellar wind density (massloss rate) is high enough. Additional temperature changes in theburning layer of the WD which are caused by variable accretion,can reduce the required wind densities. Acknowledgements.
This publication is partly based on observations withXMM-Newton, an ESA Science Mission with instruments and contributionsdirectly funded by ESA Member states and the USA (NASA). The XMM-Newton project is supported by the Bundesministerium f¨ur Wirtschaft undTechnologie / Deutsches Zentrum f¨ur Luft- und Raumfahrt (BMWI / DLR, FKZ50 OX 0001) and the Max-Planck Society. R.S. acknowledges support from theBMWI / DLR grant FKZ 50 OR 0907. This paper utilises public domain data ob-tained by the MACHO Project, jointly funded by the US Department of Energythrough the University of California, Lawrence Livermore National Laboratoryunder contract No. W-7405-Eng-48, by the National Science Foundation throughthe Center for Particle Astrophysics of the University of California under coop-erative agreement AST-8809616, and by the Mount Stromlo and Siding SpringObservatory, part of the Australian National University.
References