Long-term X-ray variation of the colliding wind Wolf-Rayet binary WR 125
MMNRAS , 1–5 (2018) Preprint 31 December 2018 Compiled using MNRAS L A TEX style file v3.0
Long-term X-ray variation of the colliding windWolf-Rayet binary WR 125
Takuya Midooka , , (cid:63) Yasuharu Sugawara , Ken Ebisawa , Institute of Space and Astronautical Science (ISAS), Japan Aerospace Exploration Agency (JAXA), 3-1-1 Yoshinodai, Chuo-ku,Sagamihara, Kanagawa 252-5210, Japan Department of Astronomy, Graduate School of Science, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-0033, Japan
Accepted XXX. Received YYY; in original form ZZZ
ABSTRACT
WR 125 is considered as a
Colliding Wind Wolf-rayet Binary (CWWB), from whichthe most recent infrared flux increase was reported between 1990 and 1993. We ob-served the object four times from November 2016 to May 2017 with
Swift and
XMM-Newton , and carried out a precise X-ray spectral study for the first time. There werehardly any changes of the fluxes and spectral shapes for half a year, and the absorption-corrected luminosity was 3.0 × erg s − in the 0.5–10.0 keV range at a distance of4.1 kpc. The hydrogen column density was higher than that expected from the inter-stellar absorption, thus the X-ray spectra were probably absorbed by the WR wind.The energy spectrum was successfully modeled by a collisional equilibrium plasmaemission, where both the plasma and the absorbing wind have unusual elementalabundances particular to the WR stars. In 1981, the Einstein satellite clearly detectedX-rays from WR 125, whereas the
ROSAT satellite hardly detected X-rays in 1991,when the binary was probably around the periastron passage. We discuss possiblecauses for the unexpectedly low soft X-ray flux near the periastron.
Key words: stars: Wolf-Rayet — X-rays: individual: WR 125 — binaries: spectro-scopic
Most of the Wolf-Rayet (WR) stars, massive stars with sig-nificant mass-loss, are known to be binaries (Rosslowe &Crowther 2015). In particular, those WR binaries whichproduce hot plasma from their stellar wind collision arecalled
Colliding Wind Wolf-rayet Binaries (CWWBs). Theshocked plasma has a temperature of 10 –10 K and high ab-sorption columns of 10 –10 H cm − (Schild et al. 2004).The X-ray luminosity is highly dependent on binary sepa-rations, mass-loss rates, and wind velocities (Stevens et al.1992; Usov 1992). The X-ray energy spectra significantlyvary with the binary orbital phase, which enable us to studyorbital dependence of the plasma parameters and amountof the circumstellar absorption through spectral analysis.In this manner, we are able to constrain the wind acceler-ation and the mass-loss rate from the WR star. We havealready applied this methodology to the CWWBs WR 140(Sugawara et al. 2015) and WR 19 (Sugawara et al. 2017),which are relatively bright with known orbital parameters.We measured variations of the circumstellar absorptions on (cid:63) E-mail: [email protected] the orbital phases, and successfully constrained the mass-loss rates from these WR stars (Sugawara et al. 2015, 2017).WR 125 is considered as a CWWB, consisting of a WC7type WR star and an O9 III companion star (Williams et al.1994). The orbital period is unknown, while it is reportedthat the infrared flux started to increase in 1990 July andreached the maximum during 1992 and 1993 (Williams et al.1994). In general, infrared brightening in the long-periodCWWBs is thought to be caused by dust formation near theperiastron passage in their eccentric binary orbits (Williamset al. 1987, 2012).In 1981, the X-ray observatory
Einstein detected X-rays from WR 125 for the first time (Pollock 1987). Theabsorption-corrected luminosity was ( . ± . ) × erg s − in the 0.2–4.0 keV band at an assumed distance of 1.9 kpc.As discussed in Pollock et al. (1981), the log-likelihood de-tection statistic λ gives a scale of the significant detection.In the case of Einstein
IPC, λ being greater than 3 is con-sidered to be a significant detection. Since Pollock (1987)showed that λ of WR 125 was 39.1, the detection was signif-icant. Later, Pollock et al. (1995) claimed a marginal detec-tion with ROSAT in 1991, where λ was 5.8; this detection © a r X i v : . [ a s t r o - ph . H E ] D ec Midooka T. et al. was not significant because
ROSAT usually takes λ >
10 asthe detection threshold.In this paper, we present new
Swift and
XMM-Newton monitoring observations of WR 125, and investigate thelong-term X-ray variation. In section 2 we introduce the ob-servation and data reduction, and in section 3 we presentdata analysis and results. We discuss long-term X-ray vari-ation of WR 125 using all the available X-ray observationalresults in section 4.
Table 1 gives the observation log and the observed countrates. We proposed a Target of Opportunity (ToO) ob-servation of WR 125 with
Neil Gehrels Swift Observatory (Gehrels et al. 2004), and three pointings were made from2016 November 28 to 2017 March 16 for a total exposure ofabout 12 ksec. The X-ray Telescope (XRT; Burrows et al.2005) was operated in the Photon-Counting mode. We pro-cessed the XRT data through the
Swift -XRT data productgenerator (Evans et al. 2007, 2009). We produced the XRTlight curves, images and spectra by using the Swift -XRTdata product generator (Evans et al. 2007, 2009).Having confirmed significant detection by
Swift , weproposed a more detailed observation with
XMM-Newton (Jansen et al. 2001), and the observation was carried outon 2017 May 11. The European Photon Imaging Camera(EPIC) is sensitive in the 0.2 to 12.0 keV energy range(Turner et al. 2001; Str¨uder et al. 2001). The data werereduced with SAS version 15.0.0 to obtain the filtered eventfiles for EPIC-MOS1, 2 and pn in 0.3–10.0 keV.Good time intervals were selected by removing the in-tervals dominated by flaring particle background when thesingle event (PATTERN =
0) count rate in the >
10 keV bandwas larger than 0.35 counts s − and that in the 10–12 keVband larger than 0.4 counts s − for EPIC-MOS and EPIC-pndata, respectively. We used a circular regions of 22 (cid:48)(cid:48) radiusfrom the same CCD for extracting source and backgroundevents . Following the SAS Data Analysis Threads , we ob-tained light curves and spectra. In the following analysis, weused HEASOFT version 6.22.1 and XSPEC version 12.9.1p. In the
Swift of ToO observations, we detected a source at(19h 28m 15.6s, +19 ◦ (cid:48) (cid:48)(cid:48) ) with a 90% radial error of2.7 (cid:48)(cid:48) . The most precise coordinate of WR 125 is (19h 28m15.61s, +19 ◦ (cid:48) (cid:48)(cid:48) ) by Gaia Collaboration et al. (2018), thus the detected object is certainly WR 125. Count ratesof three Swift observations were almost the same (Table 1).We used XSPEC to analyze X-ray spectra. We madethree energy spectra corresponding to three
Swift observa-tions. For
XMM-Newton , we made MOS1, MOS2 and pnenergy spectra, separately. We grouped three
Swift spectraevery 10 counts per bin and three
XMM-Newton spectra(MOS1, MOS2, pn) every 15 counts per bin.We set the solar abundance by Wilms et al. (2000), andfitted the spectra using a simple model (
TBabs*apec ), wherean emission spectrum from collisionally-ionized diffuse gasis affected by the interstellar absorption (Smith et al. 2001).First, we fitted the six spectra separately, and found thatthere were hardly spectral variations. Consequently, we fit-ted the six spectra simultaneously. The left-hand side of Ta-ble 2 shows the best-fit parameters using the simple model(
TBabs*apec ).Now we estimate N H of the interstellar absorption fromoptical extinction. According to a catalogue of Galactic WRstars (van der Hucht 2001), A v is 6.68 mag for WR 125.Consequently, N H was estimated as . × cm − usingthe following equation (Vuong et al. 2003), N H = . × A v × cm − . Meanwhile, our best-fit column density was . × cm − .Therefore, we suppose that X-rays were further absorbed bythe WR wind.Next, we introduced another absorption model( varabs ), in which elemental abundance is variable, in or-der to take the additional circumstellar absorption into ac-count, fixing N H of TBabs at the expected interstellar value( . × cm − ). We also changed apec to vvapec inorder to specify abundances of the collisional equilibriumplasma and set H abundance to zero. We took the C, O andNe abundances of WR 90, which is another WC7-type WRsingle star and fixed other chemical abundances to the un-known Fe abundance. Since hydrogen is depleted, we spec-ified the C, O and Ne abundances relative to He, as (i.e.(C/He) ∗ /(C/He) (cid:12) = 101.7, (O/He) ∗ /(O/He) (cid:12) = 5.98 and(Ne/He) ∗ /(Ne/He) (cid:12) = 3.81; Dessart et al. 2000). Abun-dances of H and N were set to zero, which is expected forWR 125 being a WC-type WR star, and the abundances ofthe emission ( vvapec ) and absorption ( varabs ) componentswere made equal. We fitted the six spectra simultaneously,allowing only the He abundance of varabs and the Fe abun-dance and kT of vvapec to be free parameters.The right-hand side of Table 2 shows the best-fit parameters using the more sophisticated model( TBabs*varabs*vvapec ). χ / dof was 187/134(= 1.40),slightly better than that of the simple model. According toGagn´e et al. (2012), absorption-corrected luminosities andtemperatures of CWWBs range from 10 to 10 erg s − and from 1 to 4 keV. The best-fit luminosity and plasmatemperature of WR 125 were found within these ranges. Theenergy spectra and the best-fit models are shown in Figure1. We detected persistent X-ray emission from the CollidingWind Wolf-rayet Binary WR 125 with
Swift and
XMM-
MNRAS , 1–5 (2018) ong-term X-ray variation of the colliding wind Wolf-Rayet binary WR 125 Table 1.
Observation logs and the count rates with
Swift and
XMM-Newton
Satellite/Detector Obs. mode Obs. ID Start time [UT] Exposure time (ks) (cid:219) C @0.3–1.5keV a (10 − counts s − ) (cid:219) C @1.5–10.0keV a (10 − counts s − ) Swift /XRT Photon-Counting 00034826001 2016-11-28T01:50 4.8 0.9 ± ± Swift /XRT Photon-Counting 00034826002 2016-12-17T13:27 4.7 0.6 ± ± Swift /XRT Photon-Counting 00034826003 2017-03-16T06:19 2.3 0.6 ± ± XMM /EPIC Full frame 0794581101 2017-05-11T09:06 b ± ± a Observed count rates of each energy band. The rate of
XMM-Newton is that by EPIC-pn detector. b Starting time of EPIC-pn observation. − − no r m a li z ed c oun t s s − k e V − − ( da t a − m ode l ) / e rr o r Energy (keV)
Swift/XRT/2016-11-28Swift/XRT/2016-12-17Swift/XRT/2017-03-16XMM/MOS1/2017-05-11XMM/MOS2/2017-05-11XMM/PN/2017-05-11
Figure 1.
Spectra of WR 125 and the best-fitting models. The six spectra are fitted simultaneously. In the upper-panel, the solid linesshow the best fitting models, which is
TBabs*varabs*vvapec . Newton in a series of four observations carried out in 2016-2017, following a clear detection with
Einstein in 1981 anda marginal detection with
ROSAT in 1991. No significantflux/spectral changes were found throughout the first ob-servation in 2016 November to the last one in 2017 May.We suppose that the orbital period may be longer than 24years, considering that the last reported periastron passage(expected from the near infrared flux increase) was in 1993(Williams et al. 1994), and there was no flux increase re-ported since then.We carried out X-ray spectral analysis in 0.3–10 keVfrom WR 125 for the first time. From the spectra analysis,we found that the column density was probably increasedby WR 125’s stellar wind component, and the plasma pa-rameters (luminosity and temperature) were not so extremevalues among WC-type WR binaries (Gagn´e et al. 2012).We carefully looked into the archival data of
Ein-stein and
ROSAT . Einstein data by Imaging Proportional Counter (IPC) instrument was sensitive in the 0.4 to 4.0 keVenergy range. It was obtained in 1981 April 9 (sequence No.8680), and the count rate was 0.0122 (± . ) counts s − ,which is considered as a significant detection (Pollock 1987;Harris et al. 1996). On 1991 October 28, ROSAT data wastaken by Position Sensitive Proportional Counters (PSPC)instrument in 0.1-2.0 energy range for an exposure of 2105seconds (sequence ID RP500042N00). As a result of scru-tinizing the
ROSAT data, we conclude that there was nomeaningful X-ray detection from WR 125; the WR 125 countrate was less than that of the dimmest point source signifi-cantly detected in the field-of-view ( . × − counts s − ).With WebPIMMS (ver. 4.9) , we converted the Ein-stein count rate into the flux in the 0.5–10.0 keV en- https://heasarc.gsfc.nasa.gov/cgi-bin/Tools/w3pimms/w3pimms.pl MNRAS000
ROSAT data, we conclude that there was nomeaningful X-ray detection from WR 125; the WR 125 countrate was less than that of the dimmest point source signifi-cantly detected in the field-of-view ( . × − counts s − ).With WebPIMMS (ver. 4.9) , we converted the Ein-stein count rate into the flux in the 0.5–10.0 keV en- https://heasarc.gsfc.nasa.gov/cgi-bin/Tools/w3pimms/w3pimms.pl MNRAS000 , 1–5 (2018)
Midooka T. et al.
Table 2.
The best-fitting parameter of spectral fitting
Parameter
TBabs*apec TBabs*varabs*vvapec
Interstellar absorption N H ( cm − ) ± . N He ( cm − ) ——- 0.16 ± . Thin thermal plasma kT (keV) a ± . + . − . (Fe/He)/(Fe/He) (cid:12) ——- 0.29 ± E . M . ( cm − ) b, c + . − . + . − . F x ( − erg cm − s − ) d + . − . ± . L x ( erg s − ) c, e + . − . + . − . χ / dof a Plasma temperature. b The emission measure in cm − . c Assuming a distance of 4.1 + . − . kpc (Gaia Collaboration et al. 2018). d The absorbed flux (0.5–10.0 keV). e The luminosity which is dereddened by the interstellar andcircumstellar absorptions (0.5–10.0 keV).The error of the luminosity is computed using themaximum/minimum unabsorbed flux and distance.
Table 3.
Observed flux with
Einstein , ROSAT , Swift and
XMM-Newton
Obs. date Satellite/Detector Observed flux a (yyyy.mm) (10 − erg s − cm − )1981.04 Einstein /IPC 7.3 ± . ROSAT /PSPC < Swift /XRT&
XMM /EPIC 7.9 + . − . a We converted each count rate to the flux in the 0.5–10.0 keVrange using model
APEC with WebPIMMS. ergy range assuming N H = 1.59 × cm − , plasma tem-perature = 2.16 keV and 1.0 solar abundance in APEC model based on
Swift and
XMM-Newton (see above). Theconverted flux was 7.3 (± . ) × − erg s − cm − .We also converted the ROSAT upper-limit count rate( . × − counts s − ) into the flux in the 0.5–10.0 keVrange; it was 4.2 × − erg s − cm − .Table 3 shows the observed flux with Einstein , ROSAT and
XMM-Newton ( Swift ) in the 0.5–10.0 keV energy range,where the flux observed with
ROSAT in 1991 was obvi-ously the lowest. Meanwhile, the infrared flux was increas-ing then, thus WR 125 was probably in the periastron pas-sage (Williams et al. 1994). Namely, contrary to the expec-tation that X-ray luminosity of the internal shock layer isinversely proportional to the binary separation (Usov 1992),the
ROSAT observation suggested that the soft X-ray lumi-nosity decreased at the periastron. Coincidentally, a similarsoft X-ray decrease at the periastron was also observed fromWR 22 (Gosset et al. 2009), Eta Carinae (Corcoran et al.2010), WR 140 (Sugawara et al. 2015) and WR 21a (Gosset& Naz´e 2016).In order to understand the observed low luminosity nearthe periastron, we examined three possibilities: First, theremay be a chance that the WR star or the companion star co-incidentally fully occulted the colliding wind region in 1991exactly when
ROSAT observed the source. Since the orbitalinclination angle is not restricted at all from IR or opticalobservations, it is difficult to estimate the eclipse possibility. In any case, even though the orbit of WR 125 has a highinclination, eclipse may not be expected in X-ray band. Ac-tually, a total X-ray eclipse was never reported up to nowin any CWWBs; for example, WR 20a does not show anyeclipses (Naz´e et al. 2008) and V444 Cyg shows only a partialone (e.g. Lomax et al. 2015) despite of their high inclinationangles. Therefore, the first possibility may be low.Second, soft X-ray from the colliding wind region mayhave been heavily absorbed by the WR star wind, whileintrinsic X-ray luminosity of WR 125 is not significantlyvariable. While
ROSAT /PSPC was sensitive to X-rays onlybetween 0.1 and 2.0 keV,
Swift and
XMM-Newton are re-spectively sensitive in the 0.3 to 10.0 keV and 0.3 to 12.0 keVenergy ranges. Therefore, it might be possible that
ROSAT was not able to detect the soft X-rays if significantly ab-sorbed by the WR star wind. When we increased N H fromthe best-fit . × cm − to . × cm − assumingthe same intrinsic luminosity and the spectra determined by Swift/XMM , we found it impossible to detect WR 125 using
ROSAT /PSPC. However, in fact, there are few observationsthat CWWB has such a high column density (Rauw et al.2000; Schild et al. 2004; Sugawara et al. 2015). Then, weexamined requirements that WR 125 column density wouldreach . × cm − . According to Pollock et al. (2005),column density of a spherically symmetric WR wind at adistance R from the WR star surface along the line of sightcan be written as N H ( R , φ, i ) ∼ . × (cid:219) M − µ − v − ( R ∗ / R (cid:12) ) − × ( γ / sin γ ) ∫ ∞ R / R ∗ x − ( − / x ) − dx cm − , where cos γ = cos φ sin i , mass-loss rate (cid:219) M − = − M (cid:12) yr − ,wind velocity v = km s − , φ and i are the orbitalazimuthal and inclination angle, and R ∗ is the WR stellarradius. In the case of WR 125, we used typical physical con-ditions in the WC type WR wind (cid:219) M = × − M (cid:12) yr − , v ∞ = 2000 km s − (Dessart et al. 2000), R ∗ = . R (cid:12) (Koesterke& Hamann 1995), and mean atomic weight for nucleons µ =6, which was estimated by the He, C, O and Ne abundances.Since we cannot constrain other orbital parameters, we as-sumed φ = ◦ , which gives the maximum column density.We examined two cases for different location of the X-rayemitting plasma, (1) R = 0.5AU and (2) R = 1.0AU. As aresult, we found that only when inclination i is more than78 ◦ in situation (1) or i is more than 84 ◦ in situation (2),the column density reaches . × cm − . Therefore, it ispossible to attain N H = . × cm − only under the verylimited circumstances with particular binary separation, or-bital inclination angle and azimuthal angle.Third, size of the X-ray emitting (colliding wind) mightbe reduced near the periastron under some circumstantialconditions. For example, one possibility is lack of the enoughacceleration in the O-star wind. In general, wind momen-tum of the WR star overwhelms that of the O-star, so thatthe colliding region almost reaches the O-star surface. Con-sequently, near the periastron, O-star wind may not havesufficient space to reach its terminal velocity before enter-ing the shock region, and collides with the WR star windbefore reaching the terminal velocity; this will lead to re-duction in the wind momentum fluxes (e.g. Luo et al. 1990;Stevens et al. 1992; Myasnikov & Zhekov 1993). Other possi- MNRAS , 1–5 (2018) ong-term X-ray variation of the colliding wind Wolf-Rayet binary WR 125 bilities are radiative inhibition and radiative braking, whichcan be obstacles of wind-acceleration (e.g. Stevens & Pol-lock 1994, Gayley et al. 1997). The radiative inhibition isa process where the acceleration of each wind is reduced bythe radiation from its companion star. The radiative brakingdescribes a scenario in which the WR wind is slowed afterreaching a large velocity. These mechanisms require smallbinary separations; for example, the smallest separation inV444 Cyg is 35.97 R (cid:12) (0.33AU) (Eri¸s & Ekmek¸ci 2011).If binary separation of WR 125 is sufficiently small, theseprocesses can slow down the wind velocity significantly, andreduce size of the colliding wind region, decreasing the X-ray flux. With a hydrodynamical simulation, it is suggestedthat X-ray flux could even disappear due to a full disruptionof the colliding wind region (e.g. Parkin & Gosset 2011). Inconclusion, we suppose that the significant low soft X-rayflux in 1991 was likely to be a consequence of mixture of thesecond and third possibilities.In summary, we have confirmed a long-term X-ray vari-ation from WR 125 over 36 years for the first time usingfour X-ray satellites. Still, WR 125 has many unknown as-pects, even its orbital period. If we can determine the or-bital parameters precisely in future, we may understand rea-son of the significantly low luminosity in 1991. According toWilliams et al. (1992), extinction of the non-thermal radioemission is expected to increase by the dense WR wind ma-terial just before the dust formation. Thus, we suppose thatsignificant change of the radio flux may become a sign of theperiastron passage. We propose multi-band monitoring ob-servations of WR 125 including radio, in order to determinethe orbital parameters and clarify the wind parameters. ACKNOWLEDGEMENTS
We are grateful to the anonymous referee for the compre-hensive review and many constructive comments. This re-search has made use of data and software provided by theHigh Energy Astrophysics Science Archive Research Cen-ter (HEASARC), which is a service of the Astrophysics Sci-ence Division at NASA/GSFC and the High Energy Astro-physics Division of the Smithsonian Astrophysical Observa-tory. We acknowledge the use of public data from the
Swift data archive and the UK Swift Science Data Center at theUniversity of Leicester. This study was based on observa-tions obtained with
XMM-Newton , an ESA science missionwith instruments and contributions directly funded by ESAMember States and NASA. This research was partially sup-ported by JSPS KAKENHI Grant Numbers JP16K17667(Y.S.), JP16K05309 (K.E.).
REFERENCES
Burrows D. N., et al., 2005, Space Sci. Rev., 120, 165Corcoran M. F., Hamaguchi K., Pittard J. M., Russell C. M. P.,Owocki S. P., Parkin E. R., Okazaki A., 2010, ApJ, 725, 1528Dessart L., Crowther P. A., Hillier D. J., Willis A. J., MorrisP. W., van der Hucht K. A., 2000, MNRAS, 315, 407Eri¸s F. Z., Ekmek¸ci F., 2011, Astronomische Nachrichten, 332,616Evans P. A., et al., 2007, A&A, 469, 379Evans P. A., et al., 2009, MNRAS, 397, 1177 Gagn´e M., Fehon G., Savoy M. R., Cartagena C. A., Cohen D. H.,Owocki S. P., 2012, in Drissen L., Robert C., St-Louis N.,Moffat A. F. J., eds, Astronomical Society of the Pacific Con-ference Series Vol. 465, Proceedings of a Scientific Meeting inHonor of Anthony F. J. Moffat. p. 301 ( arXiv:1205.3510 )Gaia Collaboration et al., 2018, A&A, 616, A1Gayley K. G., Owocki S. P., Cranmer S. R., 1997, ApJ, 475, 786Gehrels N., et al., 2004, ApJ, 611, 1005Gosset E., Naz´e Y., 2016, A&A, 590, A113Gosset E., Naz´e Y., Sana H., Rauw G., Vreux J.-M., 2009, A&A,508, 805Harris D. E., et al., 1996, VizieR Online Data Catalog, 9013Jansen F., et al., 2001, A&A, 365, L1Koesterke L., Hamann W.-R., 1995, A&A, 299, 503Lomax J. R., et al., 2015, A&A, 573, A43Luo D., McCray R., Mac Low M.-M., 1990, ApJ, 362, 267Myasnikov A. V., Zhekov S. A., 1993, MNRAS, 260, 221Naz´e Y., Rauw G., Manfroid J., 2008, A&A, 483, 171Parkin E. R., Gosset E., 2011, A&A, 530, A119Pollock A. M. T., 1987, ApJ, 320, 283Pollock A. M. T., Bignami G. F., Hermsen W., Kanbach G., LichtiG. G., Masnou J. L., Swanenburg B. N., Wills R. D., 1981,A&A, 94, 116Pollock A. M. T., Haberl F., Corcoran M. F., 1995, in van derHucht K. A., Williams P. M., eds, IAU Symposium Vol.163, Wolf-Rayet Stars: Binaries; Colliding Winds; Evolution.p. 512Pollock A. M. T., Corcoran M. F., Stevens I. R., Williams P. M.,2005, ApJ, 629, 482Rauw G., Stevens I. R., Pittard J. M., Corcoran M. F., 2000,MNRAS, 316, 129Rosslowe C. K., Crowther P. A., 2015, MNRAS, 447, 2322Schild H., et al., 2004, A&A, 422, 177Smith R. K., Brickhouse N. S., Liedahl D. A., Raymond J. C.,2001, ApJ, 556, L91Stevens I. R., Pollock A. M. T., 1994, MNRAS, 269, 226Stevens I. R., Blondin J. M., Pollock A. M. T., 1992, ApJ, 386,265Str¨uder L., et al., 2001, A&A, 365, L18Sugawara Y., et al., 2015, PASJ, 67, 121Sugawara Y., Maeda Y., Tsuboi Y., 2017, in Ness J.-U., MigliariS., eds, The X-ray Universe 2017. p. 215Turner M. J. L., et al., 2001, A&A, 365, L27Usov V. V., 1992, ApJ, 389, 635Vuong M. H., Montmerle T., Grosso N., Feigelson E. D., Ver-straete L., Ozawa H., 2003, A&A, 408, 581Williams P. M., van der Hucht K. A., van der Woerd H.,Wamsteker W. M., Geballe T. R., Garmany C. D., PollockA. M. T., 1987, in Instabilities in Luminous Early Type Stars.p. 221, doi:10.1007/978-94-009-3901-1 22Williams P. M., van der Hucht K. A., Bouchet P., SpoelstraT. A. T., Eenens P. R. J., Geballe T. R., Kidger M. R.,Churchwell E., 1992, MNRAS, 258, 461Williams P. M., van der Hucht K. A., Kidger M. R., Geballe T. R.,Bouchet P., 1994, MNRAS, 266, 247Williams P. M., van der Hucht K. A., van Wyk F., Marang F.,Whitelock P. A., Bouchet P., Setia Gunawan D. Y. A., 2012,MNRAS, 420, 2526Wilms J., Allen A., McCray R., 2000, ApJ, 542, 914van der Hucht K. A., 2001, New Astronomy Reviews, 45, 135This paper has been typeset from a TEX/L A TEX file prepared bythe author.MNRAS000