PASJ:Publ.Astron.Soc.Japan,1–??, c 2013.AstronomicalSocietyofJapan. (cid:13) Spectral Evolution of a New X-ray Transient MAXI J0556−332 Observed by MAXI, Swift, and RXTE Mutsumi Sugizaki1, Kazutaka Yamaoka2, Masaru Matsuoka1,3, Jamie A. Kennea4, Tatehiro Mihara1, Kazuo Hiroi5, Masaki Ishikawa6, Naoki Isobe7, Nobuyuki Kawai8, Masashi Kimura9, Hiroki Kitayama9, Mitsuhiro Kohama3, Takanori Matsumura10, Mikio Morii8, Yujin E. Nakagawa11, Satoshi Nakahira1, Motoki Nakajima12, Hitoshi Negoro13, Motoko Serino1, Megumi Shidatsu5, Tetsuya Sootome1, Kousuke Sugimori8, Fumitoshi Suwa13, Takahiro Toizumi8, Hiroshi Tomida3, Yoko Tsuboi10, Hiroshi Tsunemi9, Yoshihiro Ueda5, 3 Shiro Ueno5, Ryuichi Usui8, Takayuki Yamamoto1, Makoto Yamauchi14, Kyohei Yamazaki10, Atsumasa 1 Yoshida2, and the MAXI team 0 2 1MAXI team, Institute of Physical and Chemical Research (RIKEN), 2-1 Hirosawa, Wako, Saitama 351-0198 [email protected] n 2Department of Physics and Mathematics, Aoyama Gakuin University, 5-10-1 Fuchinobe, Sagamihara, Kanagawa 229-8558 a J 3ISS Science Project Office, Institute of Space and Astronautical Science (ISAS), Japan Aerospace Exploration Agency 0 (JAXA), 2-1-1 Sengen, Tsukuba, Ibaraki 305-8505 1 4 Department of Astronomy and Astrophysics, 0525 Davey Laboratory, Pennsylvania State University, University Park, PA 16802, USA ] 5Department of Astronomy, Kyoto University, Oiwake-cho, Sakyo-ku, Kyoto 606-8502 E 6School of Physical Science, Space and Astronautical Science, The Graduate University for Advanced Studies (Sokendai), 3-1-1 H Yoshinodai, Chuo-ku, Sagamihara, Kanagawa 252-5210 h. 7Institute of Space and Astronautical Science(ISAS), Japan Aerospace Exploration Agency(JAXA) ,3-1-1 Yoshino-dai, p Chuo-ku, Sagamihara, Kanagawa 252-5210 - 8Department of Physics, Tokyo Institute of Technology, 2-12-1 Ookayama, Meguro-ku, Tokyo 152-8551 o 9Department of Earth and Space Science, Osaka University, 1-1 Machikaneyama, Toyonaka, Osaka 560-0043 r t 10Department of Physics, Chuo University, 1-13-27 Kasuga, Bunkyo-ku, Tokyo 112-8551 s a 11Research Institute for Science and Engineering, Waseda University, 17 Kikui-cho, Shinjuku-ku, Tokyo 162-0044 [ 12School of Dentistry at Matsudo, Nihon University, 2-870-1 Sakaecho-nishi, Matsudo, Chiba 101-8308 13Department of Physics, Nihon University, 1-8-14, Kanda-Surugadai, Chiyoda-ku, Tokyo 101-8308 1 14Department of Applied Physics, University of Miyazaki, 1-1 Gakuen Kibanadai-nishi, Miyazaki, Miyazaki 889-2192 v 8 9 (Received ;accepted ) 0 Abstract 2 1. We report on the spectral evolution of a new X-ray transient, MAXI J0556 332,observed by MAXI, − 0 Swift, and RXTE. The source was discovered on 2011 January 11 (MJD=55572) by MAXI Gas Slit 3 Camera all-sky survey at (l,b)=(238 .9, 25 .2), relatively away from the Galactic plane. Swift/XRT ◦ ◦ 1 follow-up observations identified it with a−previously uncatalogued bright X-ray source and led to optical : v identification. For more than one year since its appearance, MAXI J0556 332 has been X-ray active, − i with a 2-10 keV intensity above 30 mCrab. The MAXI/GSC data revealed rapid X-ray brightening in X the first five days, and a hard-to-soft transition in the meantime. For the following 70days, the 0.5–30 r keV spectra, obtained by the Swift/XRT and the RXTE/PCA on an almost daily b∼asis, show a gradual a hardening,withlargefluxvariability. Thesespectraareapproximatedbyacutoffpower-lawwithaphoton index of 0.4–1 and a high-energy exponential cutoff at 1.5–5keV, throughout the initial 10 months where the spectral evolution is mainly represented by a change of the cutoff energy. To be more physical, the spectraareconsistentlyexplainedbythermalemissionfromanaccretiondiskplusaComptonizedemission from a boundary layer around a neutron star. This supports the source identification as a neutron-star X-raybinary. Theobtainedspectralparametersagreewiththoseofneutron-starX-raybinariesinthe soft state, whose luminosity is higher than 1.8 1037 erg s 1. This suggests a source distance of >17 kpc. − × Key words: X-rays: stars — X-rays: individual (MAXI J0556 332)— stars: neutron − 1. Introduction thought to occur when the gas from a companion (often late type) star accretes onto a compact star, a neutron Galactic X-ray binaries distributed near the Galactic star (NS) or a black hole (BH). They often show a large plane,exhibiting someofthebrightestX-rayfluxesinthe degree of X-ray flux variability and long periods of quies- sky, have been well studied since the beginning of X-ray cence, only appearing in single (or sometimes recurring) astronomy(e.g. Hayakawa1981). TheirX-rayemissionis periods of transient X-ray activity. Many attempts have 2 Sugizaki et al. [Vol. , been made to understand the behavior of these objects around the compact accretor (Halpern 2011). It con- inaunifiedscheme,particularlyemployingaccretion-disk tinued to brighten in the R-band until January 17, but theory (Shakura & Sunyaev 1973). slightlyfadedonJanuary18(Russelletal.2011). Optical So far, the standard-disk picture has successfully ex- spectroscopyrevealednarrowemissionlines inthe Bowen plained the X-ray emission from NS binaries with weak blend, and a period search in radial velocities of these magnetic fields in their bright phase (Mitsuda et al. lines provided two candidate orbital periods, 16.43 0.12 ± 1984; Makishima et al. 1986). Our next task is to un- hrsand9.754 0.048hrs(Cornelisseetal.2011). Aradio ± derstandtheir variations,particluarlyspectralstatetran- observation on January 19 with the Australia Telescope sitionsthatareoftenseenwhenthesesourcesexhibittran- Compact Array detected a faint radio source (Coriat et sientoutbursts. AlthoughRXTE,INTEGRAL,andSwift al. 2011). Both the optical-to-X-ray and the radio-to-X- surveyobservationswithwideskycoveragesprovideduse- ray flux ratios suggest that the X-ray source is probably ful information, few works have been done on the initial a NS binary rather than a BH binary, if it belongs to our transitions. Inparticular,thesestudiesonNSX-raytran- Galaxy with a distance less than 20 kpc (Russell et al. sients are limited because they are much fainter than BH 2011;Coriatet al.2011). An XMM-Newtona observation X-ray novae. with the reflection grating spectrometer (RGS) detected Theunbiasedall-skymonitoringwithMonitorofAllX- a strong emission line near 24.8 ˚Awhose center energy is ray Image (MAXI; Matsuoka et al. 2009)allows us to de- consistent with the Lyα transition of N VII in the rest tectX-raynovaeandtransients,andtofollowtheirinten- frame. From an extremely high N/O line ratio revealed sity evolution from the beginning to the end. The MAXI by this observation, the donor star is suggested to be a mission started in 2009 August and has already detected peculiar exotic star such as a hot subdwarf (sdB, sdO) or several X-ray transients and novae in their initial phase a white dwarf (Maitra et al. 2011). (e.g. Nakahira et al. 2010; Yamaoka et al. 2011; MAXI These observational results suggest that MAXI web site1). Asai et al. (2012) studied the initial rising J0556 332 is an X-ray binary, involving a collapsed ob- − behavior of outbursts from two transientNS low-mass X- ject, located in the Galactic halo. The companion is not raybinaries(LMXBs),AqlX-1and4U1608-52,andthen a regular late-type star, so that the source distance has derived a relation between the initial hard-state duration not been determined better than the X-ray estimate of and the hard-to-soft transition luminosity. 20–35 kpc by Homan et al. (2011). Most of these results In the constellation Columba, a new X-ray transient, favor the collapsed component being a NS, although the MAXI J0556 332,was discovered by the MAXI Gas Slit BH scenario has not yet been completely ruled out. − Camera (GSC; Mihara et al. 2011) at 9:21 (UT) on This paper presents the X-ray behavior of MAXI 2011 January 11 (Matsumura et al. 2011). Its position in J0556 332,including aninitialtransitionandcontinuous − Galactic coordinates, (l,b)=(238 .9, -25 .1), is relatively long-term variations, observed by the MAXI GSC. We ◦ ◦ away from the Galactic plane. A Swift (Gehrels et al. also analyzed spectral variations using data taken by the 2004) follow-up observation confirmed a bright uncata- Swift/X-Ray Telescope (XRT; Burrows et al. 2005) and logued X-ray source within the MAXI error circle and lo- the RXTE/ProportionalCounterArray(PCA;Jahodaet calized the source position at J2000 coordidates of (α, δ) al.2006)onanapproximatelydailycadence. Wedescribe = (89 .19300, -33 .17451) = (5h56m46.s32, 33 1028.2) the observations and the data reduction in section 2, and ◦ ◦ ◦ ′ ′′ − with the positional uncertainty of 1.7 (Kennea et al. present the analysis results in section 3. In section 4, ′′ 2011). The X-ray source agrees in position with an op- the originof the X-ray emission and its evolution are dis- tical star with a B-magnitude of 19.4. RXTE Target- cussed. All the quoted errors are hereafter given at the of-Opportunity (ToO) observations were also performed. 90% confidence limit, unless otherwise specified. The results revealed complex time variability, together with energy spectra that can be represented by a sum 2. Observation and Data Reduction of a multi-color disk blackbody (diskBB in Xspec ter- 2.1. MAXI/GSC minology) and a blackbody (BB) (Strohmayer & Smith 2011; Strohmayer 2011; Belloni et al. 2011). Based on Every 92-minute orbital revolution, MAXI on the color-color (CD) and hardness-intensity (HID) diagrams International Space Station (ISS) scans almost the whole extractedfromthe RXTE data,Homan et al.(2011)sug- sky with two kinds of X-ray cameras: the GSC work- gestedthatthesourceisatransientneutron-starZsource. ing in the 2–30 keV energy band, and the Solid-state Slit Theyestimatedthesourcedistancetobe20–35kpc,from Camera (SSC; Tsunemi et al. 2010; Tomida et al. 2011) achangeofthe CD/HIDtrackswhichisthoughttooccur in the 0.5–10keV band. The new X-ray transient, MAXI at the same luminosity as in another transient Z source, J0556 332, first detected by the GSC on 2011 January − XTE J1701 462(e.g. Homan et al. 2010). 11, brightened to 80 mCrab in the 4–10 keV band within − Follow-upobservationsoftheopticalcounterpartfound a day (Matsumura et al. 2011). The upper limit on the that the star had brightened to R 17.8 from its USNO- average4–10keVflux priorto the detection is 1.2mCrab B1.0 magnitude of R=19.9, and∼the spectrum revealed =1.5 10 11ergcm 2s 1(Hiroietal.2011). Thesource − − − × emissionlines indicating the presenceofanaccretiondisk was also detected by the SSC when it was discovered by the GSC. However,the time coverageof the SSC was too 1 http://maxi.riken.jp/top/ limited to study spectral and flux changes. We therefore No. ] Spectral Evolution of MAXI J0556 332 3 − via NASA/GSFC. Since the WT data are 1-dimensional, 0.3 2-4 keV only spatial information in the CCD detector X (DETX) 0.25 direction is available. The source events were collected -1]s 0.2 from a 40 pixel wide region centered on the target posi- -2 m0.15 tion, and the backgrounds were collected from a region s c with the same width as the source region and 40-pixel h p 0.1 awayfromthe targetalong the DETX direction. In spec- [ 0.05 tral model fitting, we used an XRT response matrix file, swxwt0to2s6 20010101v014.rmf,and ancillary response 0 files built by xrtmkarf. A systematic error of 2% was 0.12 4-10 keV implemented (Godet et al. 2009). 0.1 2.3. RXTE/PCA -2-1] scm00..0068 J0F55ro6m323021w1aJsaonbusaervyed13wittoh2R0X11TED/ePceCmAbeorn2a9n, aMlmAoXstI s − ph0.04 daily basis with a typical exposure of 1–2 ks. The ob- [ tained data provides useful information in the energy 0.02 range from 3 to 30 keV. We performed reduction of the 0 PCAdatawiththestandardRXTEanalysistoolsreleased MJD 55600 55700 55800 55900 56000 as a part of HEASOFT 6.11 and the CALDB files of ver- 2010/12/30 2012/06/02 sion 20111102 provided via NASA/GSFC. We used the PCA standard-2 data with a time resolution of 16-s for Fig. 1. MAXI/GSClightcurvesofMAXIJ0556 332in2-4 the spectral analysis, and the Good-Xenon data with a − keV(top)and4-10keV(bottom). Eachdatapointrepresents timeresolutionof1-µsforthelight-curveanalysis. Allthe adailyaverage. datawasscreenedwiththestandardselectioncriteria: the spacecraftpointingoffsetis smallerthan0 .02,the earth- concentrate on the GSC data. ◦ limbelevationangleislargerthan10 , andthe time since Reduction and analysis of the GSC data were car- ◦ the last South Atlantic Anomaly passage is longer than ried out following the standard procedure described by 30 minutes. We used event data detected only on the Sugizaki et al. (2011). The source event data were ex- top layer of the Proportional Counter Unit (PCU) #2, tractedfroma rectangularareaof3 .6 alongthe detector ◦ whichis the best calibratedamongall counterunits. The anodewiresand3 .0inthe scandirectioncenteredatthe ◦ backgroundwasestimatedusingthearchivedbackground sourceposition;thelattercorrespondstothepointspread model provided by the instrument team2. The response function for eachsingle scan transit of 40 seconds. The ∼ matrix files were built with pcarsp for each pointing ob- backgroundwascollectedfromdatatakeninthesamean- servation. A systematic error of 0.5% was applied. odeareajustbeforeandaftereachscantransit. Withthe same analysis procedure, we processed the light curve of 3. Analysis and Results the Crab nebula for the same period, and then confirmed that calibrationuncertaintiesin the standardlightcurves 3.1. Long-Term Light Curve and Color Variations in the 2–4 keV, 4–10 keV, and 10–20 keV energy bands Inordertoinvestigatelong-termfluxandhardnessvari- are at most 5% at the 1-σ level. ations with a good statistic accuracy, we first extracted Figure 1 shows the obtained GSC light curve of MAXI five-band X-ray light curves from the Swift/XRT (0.3–6 J0556 332 in the 2–4 keV and 4–10 keV bands for the − keV) and RXTE/PCA (2–20 keV) data. Figure 2 shows entire 1.5-year active period. These data represent the the results from 2011 January 13 (MJD=55574) to 2011 unfoldedphotonfluxper1-daytimebin. Inthe10–20keV December29(MJD=55924),togetherwithsoft-color(3.6– band,theGSCdidnotdetectanysignificantfluxabove3- 5.6 keV / 2.0-3.6 keV) and hard-color (8.5–18.4 keV / σ confidence limit, which is typically 0.015 photons cm 2 − 5.6–8.5keV)variations. Inordertoenabledirectcompar- s 1 (45 mCrab). − ison among the different instruments, the observed count 2.2. Swift/XRT ratesper256-stimebinineachinstrumentwereconverted to the unfolded photon flux, assuming that the spectrum Swift/XRT pointing observations of MAXI J0556 332 − has a power-law with a photon index of 1 and a high- were performed over 112 epochs until 2011 November energy exponential cutoff at E = 3 keV, with an in- 22 utilizing the Windowed Timing (WT) mode and and cut terstellar absorption of 0.29 1021 cm 2; these are based a typical exposure of 0.5–1 ks. Using the archival × − on the spectral analysis in section 3.5. The Swift/XRT Swift/XRTdata,weinvestigatedtheevolutionofthe0.3– and RXTE/PCA data in the same energy band, taken 10keVenergyspectrum. Thedatareductionandanalysis within about a day, thus mostly agree with each other, were performed using the Swift analysis software version which confirms general consistency between the two in- 3.8, released as a part of HEASOFT 6.11 and CALDB (calibration database) files of version 20110915, provided 2 http://heasarc.gsfc.nasa.gov/docs/xte/pca news.html 4 Sugizaki et al. [Vol. , struments. The largeflux discrepancyseenfor afew data emit type-I X-ray bursts, characterized by a fast rise in points close in time is indicative of flux changes between a few seconds and an exponential decay in a few tens of the pair of observations, which were not exactly simulta- seconds (e.g. Lewin et al. 1993). Therefore, we searched neous. theentireRXTE/PCAlightcurvesfortype-IX-raybursts Theselightcurvesshowcomplexenergy-dependentvari- using 1-s time bins. However, we did not find any burst- ations, particularly during the initial 50 days. In the like event with a count-rate increase higher than 50 c s 1 − lowestenergyband(0.3–2keV),theflux∼reachedthemax- PCU 1 in the 2–20 keV band. − imum within a few days after the source emergence, and Even if the source is located at the distance of 35 then decayed in about 50 days. In the highest energy kpc, which is the farthest limit estimated by Homan et band (8.5–18.4 keV), the flux in contrast increased grad- al. (2011), we would have detected 98 c s 1 with the − ∼ ually, reaching the maximum with large short-time vari- RXTE/PCAfromatypicalX-rayburst,assumingablack- ability on about the 50th day. The contrast between the body spectrum with a temperature of 2.1 keV and a lu- soft and hard flux evolutions are clearly reflected in the minosity as high as the Eddington luminosity of 2 1038 soft-color and hard-color variations, because both colors erg s 1 as seen in the typical type-I burst. Such a×burst − largely increased during the initial 50 days. should have been observed with the RXTE/PCA. As shown in figure 2, we divided the entire observa- 3.4. Initial Transition Behavior tion period into six intervals, A, B, C, D, E, and F; they are characterized as follows. Int-A: initial (30 days) The initial emergent phase for about three days of phasewhentheintensitybelow2keVwashighestandthe MAXI J0556 332 was covered only by the MAXI/GSC − hardest-bandintensity increasedgradually. Int-B:bright- all-sky survey. To clarify the source behavior meanwhile, est phase in the 5.6-8.5keV and harder bands, with large we extracted 2–4 keV and 4–10 keV lightcurves in a 12-h variabilityinthewholeband. Int-C:decayingphasewhere time bin, and present them in figure 5. Like in figure 2, the large variability in the 5.6-8.5 keV and harder bands the GSC count rates were converted to the incident pho- stillremains. Int-D:intermissionphasewheretheaverage ton fluxes, to make the results approximately free from flux is lower than in the intervals just before and after. the instrumental responses. There, the Swift/XRT and Int-E: re-brightening phase in the whole band with mod- the RXTE/PCA data are superposed, together with the erate variability. Int-F: low activity phase with a similar hardness ratio and the 2 10 keV flux. The Swift/XRT − average flux to Int-D. andRXTE/PCAdatapoints,eachcoveringtypically0.5– To better grasp the spectral changes associated with 2 ks (plotted with a bin width of 256 s), are significantly the intensity variations, we plot several CDs in figure 3, morescatteredthanthe12-haveragedGSCdata,presum- representing the correlation between the two colors ex- ably due to short-term (e.g. <1 h) variations. However, tracted from the RXTE-PCAdata. There, data from the the hardnessratiosarealwaysconsistentamongthe three six intervals defined in figure 2 are shown separately. All missions. This implies that the spectral shape did not the six CDs apparently exhibit Z-like shapes as often ob- changelargelyonatime scaleofadayorshorter,atleast servedinbrightNSLMXBs(e.g. vanderKlis2006). This in the initial phase. strongly supports the source identification as a NS X-ray The MAXI/GSC hardness ratio clearly shows a step- binary. These CDs are also classified into two groups. likedecreaseatMJD=55574.0beforetheRXTEandSwift The Int-C behavior resembles that of Int-E, while Int-D observations started. Therefore, the source is considered resemblesInt-F.ThisagreeswiththereportbyHomanet tohavemadeahard-to-softtransitionatthisepoch,when al. (2011) on this source, that the CD track changed be- the flux was still rising rapidly. The flux then peaked at tween 2011 February (MJD 55600) and 2011 September MJD=55575.5 in the presumable soft state. ∼ (MJD 55800). According to two subgroups of the Z-like ∼ 3.5. X-ray Spectral Evolution CDs, represented by Cyg X-2 and Sco X-1 (e.g. Homan et al. 2010), the CDs of Int-A, B, C, and E are classified We carried out X-ray spectral analysis over the 0.5– into the Cyg X-2 group, while those of Int-D and F are 30 keV broad band, by combining the Swift/XRT and into the Sco X-1 group. RXTE/PCA data which are averaged over each observa- In figure4, we plot the twocolorsagainstthe intensity, tion(0.5–2kstypically). AlthoughtheseSwiftandRXTE to create two HIDs. There, the six intervals are specified observations were not exactly simultaneous, any spectral with the same color scheme as in figure 3. As Homan et changewithinadayisconsideredsmallasalreadynoticed al. (2011) reported, The change of the HID tracks among in section3.1. We thus selected pairs of Swift and RXTE the six intervals is also clearly seen. observations carried out within 12 hours, and performed their simultaneous fits. To avoid possible inconsistency 3.2. Color-Color and Hardness-Intensity Diagrams betweenthetwoinstruments,weleftthePCAversusXRT normalization, f , free, and discarded those pairs 3.3. Search for X-ray Burst Activity PCA/XRT ofwhichthe value off isdeviatedfromthe aver- PCA/XRT From multi-wavelength observations reported so far age (=1.12) by more than 30%. The logs of the selected (section 1) and the results of data analysis presented observations, totaling 72 pairs, are summarized in table above,MAXIJ0556 332isconsideredmostlikelytobea 1. Allspectralfitting wascarriedoutusingXspec version − NS X-ray binary. Transient NS X-ray binaries sometimes 12.7.0, released as a part of HEASOFT 6.11. No. ] Spectral Evolution of MAXI J0556 332 5 − 1.2 0.3-2.0 keV ]s-1 1 -2 m 0.8 Swift/XRT c 0.6 s ph 0.4 [ 0.2 0.3 A B C D E F 2.0-3.6 keV 70 -2] s[phs cm-1 00..0012..5521 RXTE/PCA 2345600000 -1]PCA Rate [c s 0.05 10 0.12 3.6-5.6 keV 100 ]s-1 0.1 80 -1]c s -2 hs cm 00..0068 4600 A Rate [ [p 00..0024 20 PC 0.05 5.6-8.5 keV 50 ]s-1 0.04 40 -1]c s -2 hs cm 00..0032 2300 A Rate [ [p 0.01 10 PC 0.022 18 0.02 8.5-18.4 keV -2] s[phs cm-1 0000000.......00000000.0001111046824681 46811110246 -1]PCA Rate [c s 0.002 2 0.8 3.6-5.6 keV/2.0-3.6 keV 3 Soft flux ratio 0000....4567 122..55 Soft Color 0.3 1 0.9 1 8.5-18.4 keV/5.6-8.5 keV 0.8 0.9 Hard flux ratio 000000......345678 00000.....34567 Hard Color 0.2 0.2 0.1 0.1 MJD 55600 55650 55700 55750 55800 55850 55900 2010/12/30 2012/01/14 Fig. 2. Light curves in 0.3–2.0 keV, 2.0–3.6 keV, 3.6–5.6 keV, 5.6–8.5 keV, and 8.5-18.4 keV energy bands obtained by the Swift/XRT(gray)andtheRXTE/PCA.Black,red,green,blue,purpleandcyanspecifyintervalsA,B,C,D,E,andFrespectively, aslabeledinthesecondpanel. Softcolor(3.5–5.6keV/2.0–3.6keV)andhardcolor(8.5–18.4keV/5.6–8.5keV)oftheRXTE/PCA data areshown inthe bottom two panels. Allpanels utilize256-s timebin. Allthe vertical errorbars representthe 1-σ statistical uncertainty. 6 Sugizaki et al. [Vol. , 0.8 A 0.8 B ) A B C D E F V 0.7 0.7 e 2.2 k 0.6 0.6 6 3. 0.5 0.5 0- 2 2. 0.4 0.4 V)/( 1.8 0.3 0.3 e V)0.2 0.2 6 k e 3. 1.6 8.5 k0.11 1.2 1.4 1.6 1.8 2 2.2 0.11 1.2 1.4 1.6 1.8 2 2.2 (5.6- 1.4 5.6-0.8 C 0.8 D olor eV)/(00..67 00..67 Soft C 1.2 k 4 0.5 0.5 1 8. 0 20 40 60 80 100 120 140 160 180 200 5-10.4 0.4 V) 0.8 8.0.3 0.3 ke or (0.2 0.2 8.5 0.7 ol0.1 0.1 6- C 5. rd 1 1.2 1.4 1.6 1.8 2 2.2 1 1.2 1.4 1.6 1.8 2 2.2 V)/( 0.6 Ha0.8 E 0.8 F ke 0.5 0.7 0.7 4 8. 0.6 0.6 1 - 0.4 5 0.5 0.5 8. ( 0.4 0.4 or 0.3 ol 0.3 0.3 C d 0.2 0.2 0.2 ar H 0.1 0.1 0 20 40 60 80 100 120 140 160 180 200 1 1.2 1.4 1.6 1.8 2 2.2 1 1.2 1.4 1.6 1.8 2 2.2 2.0-18.4 keV Count Rate [c s-1] Soft Color (3.6-5.6 keV)/(2.0-3.6 keV) Fig. 4. Soft-colorversusintensity(top)andhard– Fig. 3. Color-color diagrams (CDs) extracted from the color versus intensity (bottom) diagrams extracted RXTE-PCAdata,forindividualintervalsofA,B,C,D,E,andF fromtheRXTE-PCAdata. Data intervals of A,B, definedinfigure2. C, D, E, and F are colored similarly as in figure 2 andfigure3. Figure 6 shows the XRT and PCA spectra in their un- the cutoff energy E changed from 1.5 keV to 5 cut folded νfν form, from six typical observations, as indi- keV. The flux decreased from 4 10 9∼to 5 10 10∼erg − − cated with arrowsin the top panelof figure 7a. Including cm 2 s 1 overthe 10 months in a×greementwi×th the light − − thesixrepresentatives,alltheobtainedspectrawerefound curve (figure 2). The best-fit parameters for the spectra to show a featureless continuum, without any signifi- in figure 6 are summarized in table 2. cant emission or absorption features. We first attempted The spectrum, represented by a cutoff power-law with to fit them with a simple continuum model of either a Γ = 0.4–1 and E = 1.5–5 keV, roughly agrees with cut power law (PL), a blackbody (BB), a broken power-law those of typical bright NS-LMXBs. The cutoff energy (bknpower), or a power law with a high-energy exponen- could represent the temperature of blackbody radiation tial cutoff (cutoffpl), all multiplied with photoelectric from a neutron-star surface (e.g. Mitsuda et al. 1984, absorption (wabs, Morrison & McCammon 1983) with Mitsuda et al. 1989). We thus attempted a canonical free column density N . Then, 50 out of the 72 spec- two-component model for bright NS-LMXBs, consisting H tral pairs were reproduced by the wabs*cutoffpl model of a multi-color-disk blackbody (diskBB in Xspec termi- with reduced chi-squared(χ2) within the 99% confidence nology) and a blackbody (BB) (Mitsuda et al. 1984). ν limits, while none of the other models were as success- However, as shown in figure 7a (fifth panel) and table ful on any data set. Figure 7a plots time evolution of 2, the best-fit χ2 values of the wabs*(diskBB+BB) fits ν best-fit wabs*cutoffplmodel parameters, together with were no better than those with the wabs*cutoffpl. The absorption-corrected 0.1–100 keV fluxes. Thus, the pho- residuals revealed excess features at both low and high ton index Γ was in the range of 0.4–1 throughout, while energies, which can be considered as a signature of a No. ] Spectral Evolution of MAXI J0556 332 7 − 1 9 1 2 1 Mx/GSC 2-4 keV Γ= + (1) -2-1] sm 00..34 SRwx//PXCRAT (Sunyae"v4& TmkiTetcae2rτch(1uk+1τ39)8#0).−2 s c 0.2 The nthcomp model alone, with seed photons of nei- h p ther BB nor diskBB, was able to fit the observed spectra, [ 0.1 evenif the N wasleft free. The residuals showeda large H 0 discrepancy in the soft X-ray band below 2 keV. Adding another soft thermal component, of which the spectral 0.16 4-10 keV 0.14 shapeisthe sameasthatoftheseedphotons,didnotim- -1]s 0.12 prove the fit. However, when the temperature of the soft -2 m 0.1 component was allowed to be free, and hence is different c 0.08 hs 0.06 from the seed-photon temparature, the Comptonization [p 0.04 plus soft-component model became acceptale. Therefore, 0.02 the seed photons for Comptonization and the additional 0 soft component must be of different origin. 4-10 keV/2-4 keV When the additional soft component is represented by a BB model, the fit alwasys made the absorption col- o ess rati 1 uismsnigndiefincsaitnytlyNHlow<e0r.1th×an102t1hecmG−a2la.ctTichiHsIupdpenesritlyimoitf ardn 0.29×1021 cm−2 in the source direction (Kalberla et al. H 2005). This would contradict the high Galactic latitude (b =25 2) and the suggested large distance, which place ◦ | | MAXI J0556 332well outside the Galactic disk. On the 2-10 keV flux other hand, a−n alternative model employing a diskBBfor 2.5 -1-2] sm 2 tlahregesroftthcaonmtphoenGenatlaacltwicayHsIg.aTvehetrheefobrees,tw-fiethNerHeawftheirchusies g c 1.5 the diskBB model for the soft component. er -9 0 1 Then, how about the seed-photonspectrum? Ifthe op- [1 0.5 tical depth of the Comptonizing medium is high enough (τ 1), whether the seed photon spectrum follows a BB 0 ≫ or a diskBB model causes little difference in the emer- MJD 55570 55575 55580 55585 55590 55595 gent spectrum, and therefore are difficult to distinguish. 2011/01/06 2011/02/06 Hence, we assume that the seed photons have a BB spec- trum. Insection4.2,Wediscussthevalidityofthismodel Fig. 5. Two-band MAXI/GSC light curves (black) during from the obtained best-fit parameters. the initial phase, plotted since five days before the first de- tection. Each data point is an average over 12 hours. The Using the wabs*(diskBB+nthcomp) model with a BB Swift/XRTandRXTE/PCAdatain256-stimebinareover- seed-photon source, we fitted all the spectrum pairs, and laid in blue and red, respectively. Hardness ratio (4–10 keV obtained successful fits as exemplified in figure 6. Then, /2–4keV)andthe2-10keVfluxareshowninthethirdand assuming that the Comptonization process conserves the thefourthpanels,respectively. photon number in the originalBB radiation,we estimated the radius of the BB seed-photon sphere R according Comptonization process. Following Lin et al. (2007) and seed to the formula of Lin et al. (2009), we hence tried to add a broken power- law function approximating the Comptonized component Fnthcomp(Tseed,Rseed,d) photonscm−2s−1 to the canonical two-component model, but the model wabs*(diskBB+BB+bknpower)didnotimprovethefitsig- = fBB(E,Tseed,Rseed,d)dE(cid:0) (cid:1) nificantly. Z 3 2 2 WethenexaminedmorerigorouslytheComptonization kTseed Rseed d − =20.1 , (2) emission process, employing a thermally Comptonized 1keV 1km 10kpc (cid:18) (cid:19) (cid:18) (cid:19) (cid:18) (cid:19) continuum model, nthcomp in Xspec terminology where F (T ,R ,d) is the incident photon flux (Zdziarski et al. 1996; Z˙ycki et al. 1999). It describes nthcomp seed seed calculated from the best-fit model parameters in Xspec, Comptonizedemissionarisingwhenthermalseedphotons T is the seed-photon BB temperature, d is the source seed with a BB or a diskBB spectrum is up-scattered by a hot distance, and f (E,T ,R ,d) is the photon-flux BB seed seed thermal electrons with a temperature kT . This model e spectrum of a BB emission. Including the values of R , seed uses an asymptotic power-law photon index Γ, which is the best-fit model parameters from the six representative related with kT and the scattering optical depth τ as e spectra are given in table 2. Figures 7b show temporal variations of the best-fit pa- rameters, including τ and R derived using equations seed 8 Sugizaki et al. [Vol. , Table 1. ListofSwift/XRTandRXTE/PCAobservationpairsusedinthesimultaneousspectal fits MJD Swift/XRT RXTE/PCA ObsID Date† Start End Exp.[s] ObsID Date† Start End Exp.[s] 55574* 00031914001 01/13 10:51 11:14 1413 96371-01-02-00 01/13 12:32 13:22 2912 55575 00031914002 01/14 06:10 06:31 1277 96371-01-01-00 01/14 16:44 19:06 5936 55578 00031914005 01/17 14:23 14:42 1125 96371-01-03-00 01/17 12:13 19:24 15872 55580 00031914007 01/19 16:09 16:24 891 96371-01-04-01 01/19 19:35 20:00 864 55582 00031914009 01/21 10:10 10:33 1375 96371-01-05-00 01/21 08:52 09:17 1456 55583* 00031914010 01/22 18:15 18:39 1466 96371-01-05-01 01/22 13:02 13:51 2800 55584 00031914011 01/23 05:30 05:54 1480 96371-01-05-02 01/23 14:27 14:57 1616 55587 00031914014 01/26 15:13 15:37 1442 96371-01-05-05 01/26 08:06 08:48 2352 55588 00031914015 01/27 04:03 04:19 955 96371-01-05-06 01/27 13:56 14:36 2224 55589 00031914016 01/28 04:08 04:27 1167 96414-01-01-00 01/28 08:29 09:25 3200 55590 00031914017 01/29 04:13 04:28 938 96414-01-01-01 01/29 06:33 07:22 2752 55591 00031914018 01/30 01:21 01:38 1006 96414-01-01-02 01/30 07:30 08:27 3200 55593 00031914020 02/01 04:28 04:44 981 96414-01-01-07 02/01 04:05 04:21 800 55594 00031914021 02/02 08:01 08:16 900 96414-01-01-05 02/02 11:09 11:43 1872 55596 00031914023 02/04 04:48 05:07 1180 96414-01-02-00 02/04 02:33 02:54 1152 55597 00031914024 02/05 16:04 16:17 760 96414-01-02-07 02/05 23:26 23:39 768 55598 00031914025 02/06 19:30 19:34 252 96414-01-02-09 02/06 21:57 22:21 416 55600* 00031914027 02/08 10:10 10:30 1165 96414-01-02-04 02/08 11:10 11:59 2368 55606 00031914033 02/14 13:55 14:09 873 96414-01-03-03 02/14 13:03 15:00 2848 55607 00031914034 02/15 13:49 14:08 1147 96414-01-03-04 02/15 10:54 12:53 3408 55608 00031914035 02/16 13:52 14:08 959 96414-01-03-05 02/16 10:23 11:24 2896 55609 00031914036 02/17 12:32 12:52 1208 96414-01-03-06 02/17 16:23 16:43 336 55611 00031914038 02/19 07:43 08:05 1275 96414-01-04-01 02/19 05:48 09:52 8912 55614 00031914039 02/22 01:35 01:54 1144 96414-01-04-04 02/22 07:34 10:51 6144 55616 00031914040 02/24 04:58 05:18 1243 96414-01-04-06 02/24 11:19 11:40 880 55618 00031914041 02/26 02:06 02:13 441 96414-01-05-01 02/26 11:57 13:26 3040 55620* 00031914042 02/28 18:08 18:29 1260 96414-01-05-03 02/28 11:25 11:59 1872 55624 00031914044 03/04 17:07 17:30 1373 96414-01-06-00 03/04 12:14 12:48 2000 55626 00031914045 03/06 04:18 04:40 1309 96414-01-06-02 03/06 05:06 05:37 1872 55628 00031914046 03/08 02:46 03:09 1398 96414-01-06-04 03/08 13:25 14:22 3184 55630 00031914047 03/10 09:43 09:59 994 96414-01-06-06 03/10 09:28 09:49 1232 55632 00031914048 03/12 04:40 04:49 538 96414-01-07-01 03/12 11:31 12:21 2944 55634 00031914049 03/14 01:42 02:01 1148 96414-01-07-02 03/14 02:41 03:13 1888 55638 00031914051 03/18 13:21 13:40 1165 96414-01-08-00 03/18 12:02 12:35 1952 55640* 00031914052 03/20 15:12 15:32 1216 96414-01-08-02 03/20 10:59 11:20 1216 55642 00031914053 03/22 02:30 02:46 968 96414-01-08-04 03/22 01:09 01:19 608 55646 00031914055 03/26 10:45 11:04 1145 96414-01-09-00 03/26 07:55 08:47 2416 55648 00031914056 03/28 12:30 12:46 972 96414-01-09-03 03/28 08:36 09:19 2128 55650 00031914057 03/30 17:37 17:53 978 96414-01-09-05 03/30 10:52 11:10 592 55652 00031914058 04/01 19:13 19:34 1277 96414-01-10-01 04/02 06:04 08:33 5216 55654 00031914059 04/03 14:39 14:57 1092 96414-01-10-02 04/03 07:11 08:05 2592 55656 00031914060 04/05 21:14 21:33 1169 96414-01-10-05 04/06 02:53 03:26 2016 55658 00031914061 04/07 03:44 04:03 1156 96414-01-10-06 04/07 05:08 05:23 464 55662* 00031914063 04/11 05:50 06:11 1258 96414-01-11-02 04/11 14:21 15:14 2272 55664 00031914064 04/13 01:06 01:25 1143 96414-01-11-04 04/13 00:46 01:26 2320 55666 00031914065 04/15 01:19 01:42 1392 96414-01-12-00 04/15 12:21 12:43 1072 55670 00031914067 04/19 03:25 03:47 1340 96414-01-12-02 04/19 02:50 03:15 1488 55684 00031914071 05/04 02:48 03:03 874 96414-01-14-02 05/03 20:49 21:11 1328 55688 00031914072 05/07 11:16 11:33 1052 96414-01-15-01 05/07 07:51 08:48 3168 55694 00031914074 05/13 03:29 03:48 1183 96414-01-16-00 05/13 06:42 07:25 1904 55697 00031914075 05/16 10:20 10:39 1110 96414-01-16-03 05/16 06:50 07:24 1696 55700 00031914076 05/19 10:34 10:58 1479 96414-01-16-06 05/19 05:15 05:36 832 55706 00031914078 05/25 12:36 12:57 1245 96414-01-17-03 05/25 03:46 04:42 2688 55709 00031914079 05/28 01:35 01:56 1220 96414-01-18-01 05/28 02:12 03:13 3152 55712 00031914080 05/31 22:48 23:05 1055 96414-01-18-05 05/31 21:33 21:54 1280 55715 00031914081 06/03 19:57 20:15 1101 96414-01-19-00 06/04 03:57 04:33 2160 55718 00031914082 06/06 00:50 01:07 1059 96414-01-19-01 06/06 02:37 03:30 3088 55721 00031914083 06/09 01:05 01:27 1322 96414-01-19-03 06/09 02:43 04:15 3120 55724 00031914084 06/12 01:21 01:44 1377 96414-01-20-01 06/12 02:48 03:42 3184 55746 00031914091 07/04 16:06 16:29 1330 96414-01-23-01 07/05 02:32 02:56 1024 55749 00031914092 07/07 19:28 19:43 922 96414-01-24-00 07/08 00:57 02:56 3120 55752 00031914093 07/10 00:35 00:36 66 96414-01-24-02 07/10 03:12 03:34 864 55758 00031914095 07/16 12:02 12:18 974 96414-01-25-00 07/16 21:54 00:30 6288 55774 00031914099 08/01 10:35 10:51 997 96414-01-27-04 08/01 21:57 00:26 6032 55778 00031914100 08/05 18:34 18:49 948 96414-01-28-00 08/05 13:46 13:59 800 55782 00031914101 08/09 22:04 22:21 974 96414-01-28-04 08/09 18:12 18:33 1280 55785 00031914102 08/13 06:38 06:55 1021 96414-01-29-00 08/12 23:24 23:39 720 55790 00031914103 08/17 00:13 00:29 943 96414-01-29-03 08/17 03:22 03:43 784 55805 00031914107 09/02 03:13 05:01 6463 96414-01-31-05 09/01 20:35 21:16 2096 55809 00031914108 09/06 00:20 08:36 29714 96414-01-32-01 09/05 19:03 19:27 1472 55826 00031914112 09/22 06:40 06:57 974 96414-01-34-02 09/22 15:05 15:22 1024 55832 00031914113 09/29 05:46 05:57 678 96414-01-35-02 09/28 23:07 23:40 1792 ∗: Energy spectra are shown in figure 6 and thebest-fit parameters are summarized table 2. †: Date and time are in UT (UniversalTime). No. ] Spectral Evolution of MAXI J0556 332 9 − MJD=55574 UT:2011/01/13 MJD=55583 UT:2011/01/22 MJD=55600 UT:2011/02/08 1 Obs#1 1 Obs#2 1 Obs#3 m s)−2−1 0.1 m s)−2−1 0.1 m s)−2−1 0.1 V c V c V c E*F (keE0.01 E*F (keE0.01 E*F (keE0.01 4 2 2 2 χ 0 χ 0 χ 0 −2 −2 −2 0.5 1 2 5 10 20 0.5 1 2 5 10 20 0.5 1 2 5 10 20 Energy (keV) Energy (keV) Energy (keV) MJD=55620 UT:2011/02/28 MJD=55640 UT:2011/03/20 MJD=55662 UT:2011/04/11 1 Obs#4 1 Obs#5 1 Obs#6 m s)−2−1 0.1 m s)−2−1 0.1 m s)−2−1 0.1 V c V c V c E*F (keE0.01 E*F (keE0.01 E*F (keE0.01 2 2 2 1 0 χ 0 χ χ 0 −2 −1 −2 −2 0.5 1 2 5 10 20 0.5 1 2 5 10 20 0.5 1 2 5 10 20 Energy (keV) Energy (keV) Energy (keV) Fig. 6. Examples of unfolded νfν spectra of MAXI J0556 332 obtained by the Swift/XRT (black) and the RXTE/PCA (red). − Thebest-fitwabs*(diskBB+nthcomp) modelanditsresidualsareshowntogether. Themodelparametersaresummarizedintable2. Epochsofthesesixobservations areindicatedinthetoppaneloffigures7a. (1) and (2) respectively,and the unabsorbedfluxes of the ten appear as transients or reccurents. The luminosity of diskBB and nthcomp components. The values of R atollsourcesvariesoveralargerangefrom 10 3to 0.2 seed − ∼ ∼ and the diskBB-model parameter, R √cosi, related to times L , accompaniedby spectral state changes. The in EDD the disk inner radius R and the inclination i, are calcu- higher-energy cutoff of 1.5–5 keV of MAXI J0556 332, in − lated assuming d=10 kpc. The actual radii are propor- represented by E in the wabs*cutoffpl model or kT cut e tional to the source distance. in the wabs*(diskBB+nthcomp) model, agrees well with those of the Z sources such as Cyg X-2 (Di Salvo et 4. Discussion al. 2002), Sco X-1 (D’A´ı et al. 2007), GX 5 1 (Sriram − et al. 2011), GX 17+2 (Farinelli et al. 2005), and GX MAXI J0556 332 is a new X-ray transient which ap- 349+2 (Di Salvo et al. 2001), whose luminosities are al- − peared on 2011 January 11. Using the MAXI/GSC, ways>0.5L . TheE valuesofatollsourcesarealso EDD cut Swift/XRT, and RXTE/PCAdata, we studied the inten- below 5 keV in the soft state with the luminosity higher sity and spectal evolution of this source over the entire than 0.1L , while they tend to increase up to > 15 EDD active period for more than one year. keV in the hard state with the luminosity of the order of 0.01L ;thelatterexpamplesincludeAqlX-1(Linetal. 4.1. Source Identification from X-ray Spectrum EDD 2007;Sakuraietal.2012),4U0614+09(Singh&Apparao The wide-band(0.5–30keV)X-rayspectraobtainedby 1994), 4U 1608 52 (Gierlin´ski & Done 2002b; Lin et al. − the Swift/XRT and XTE/PCA are featureless, and can 2007;Takahashi et al. 2011),4U 1705-44(Barret & Olive be approximated by a cutoff power-law with Γ 0.4 1 2002; Lin et al. 2010), and 4U 1728 34 (Tarana et al. ≈ − − andE 1.5 5keV.Theyarebetter fitted withatwo- 2011). Therefore, if the source is a NS LMXB, the X-ray cut componen≈tmod−elthatconsistsofamulti-color-diskblack- luminosity should be higher than 0.1L =1.8 1037 EDD × bodyandathermallyComptonizedblackbody. Thisspec- erg s 1 in the observed period. The absorption-corrected − tralmodelhasoftenbeenusedtodescribeX-rayemission model flux, whichchangedfrom4 10 9 to 5 10 10 erg − − × × from LMXBs containing a weakly-magnetized NS (e.g. cm 2 s 1,constrainsthesourcedistancetobed>17kpc, − − Gierlin´ski & Done 2002b; Sakurai et al. 2012) or a BH inagreementwiththeestimateofd=20–35kpcbyHoman candidate (e.g. Done et al. 2004). et al. (2011) deduced from the luminisity at a CD-track NS LMXBs have been largely classified into “Z-type” transition. and ”atoll-type” sources according to their CD-track Galactic BH X-ray binaries have been observed mostly shapes(Hasinger&vanderKlis1989;vanderKlis2006). in either of the two major spectral states, the low/hard The former objects are persistently as bright as L , or the high/soft state. These spectra are represented EDD and sometimes show rapid flux variations up to a factor by a combination of thermal emission from the accretion of 5. The latter are fainter than the former, and of- disk anda Comptonizedharder component (althoughde- ∼ 10 Sugizaki et al. [Vol. , (a) Model: wabs*cutoffpl (b) Model: wabs*(diskBB+nthcomp) A B C D E F A B C D E F 12 3 4 5 6 ]m-2 ]m-2 Hc 1 Hc 1 N21 0 N 21 0 1 1 [ [ Galactic N H 10-1 10-1 10 10 :kT :kT :kT in seed e e T k Ecut[keV] , kTseed[keV] 1 1 , Tin k 1.8 :Rin :Rseed 11..46 Rseed 102 Γ 1.21 , is km] 0.8 o [ c 00..46 Rin 10 0.2 Unabs. flux-9] s erg cm0-1-2 1 τ 10 1 [ cutoffpl :diskBB :nthcomp :Sum χ2ν diskBB+BB bs. flux ] sg cm-1-2 1 1 Una-9 er0 1 [10-1 F DO 102 χ2ν 1 55600 55650 55700 55750 55800 55600 55650 55700 55750 55800 2011/01/04 2011/10/11 11/01/04 11/10/11 Fig. 7. Evolutionofthe best-fitspectralparameters, jointlydeterminedwiththeSwift/XRTandtheRXTE/PCAdata. (a) The resultsfromtheempiricalwabs*cutoffpl fits. Fromtoptobottom,theabsorbingcolumndensity, thecutoffenergy, thepower-law index, the absorption-corrected model flux, the χ2 values, and the degree of freedom. In the fifth panel, the fit goodness with an ν alternativewabs*(diskBB+BB)isshownin(red)cross. (b)Thecasewiththemorephysicalwabs*(diskBB+nthcomp) fits. Inthethird panel,Rseed andRin√cosiarecalculatedintheassumedsourcedistanceof10kpc.