XMM-Newton confirmation of a new intermediate polar: XMMU J185330.7-012815
aa r X i v : . [ a s t r o - ph . H E ] A ug Mon. Not. R. Astron. Soc. , 1–9 (2011) Printed 21 February 2018 (MN L A TEX style file v2.2)
XMM-Newton confirmation of a new intermediate polar:XMMU J185330.7-012815
C. Y. Hui ⋆ , K. Sriram and C.-S. Choi Department of Astronomy and Space Science, Chungnam National University, Daejeon 305-764, Republic of Korea International Center for Astrophysics, Korea Astronomy and Space Science Institute, 36-1 Hwaam, Yuseong, Daejeon 305-348,Republic of Korea
Received 2011 March 14
ABSTRACT
We report the results from a detailed spectro-imaging and temporal analysis of anarchival XMM-Newton observation of a new intermediate polar XMMU J185330.7-012815. Its X-ray spectrum can be well-described by a multi-temperature thermalplasma model with the K-lines of heavy elements clearly detected. Possible counter-parts of XMMU J185330.7-012815 have been identified in optical and UV bands. Thelow value of the inferred X-ray-to-UV and X-ray-to-optical flux ratios help to safelyrule out the possibility as an isolated neutron star. We confirm the X-ray period-icity of ∼
238 s but, different from the previous preliminary result, we do not findany convincing evidence of phase-shift in this observation. We further investigate itsproperties through an energy-resolved temporal analysis and find the pulsed fractionmonotonically increases with energy.
Key words: binaries: close — cataclysmic variables — stars: individual(XMMU J185330.7-012815, G31.9-1.1) — X-rays: stars
XMMU J185330.7-012815 is an X-ray object that hasits emission nature not yet completely confirmed. It hasbeen detected in the ROSAT All-Sky Survey (RASS).Based on its extent inferred from the RASS data (i.e.11 arcmin × ∼
14 arcmin. Al-though they found that XMMU J185330.7-012815 appearsto be elongated in an ASCA GIS image, the large off-axisangle precludes any constraining spatial analysis. Examin-ing the spectral data collected by ASCA, Schaudel (2003)found that the X-ray spectrum of XMMU J185330.7-012815is featureless and can be fitted with an absorbed power-law model with a photon index of Γ = 1 . +0 . − . (cf. Ta-ble 5.2 in Schaudel 2003). Together with the apparentlycentrally-brightened X-ray morphology, the author claimedthat the power-law spectral fit strongly supports the in-terpretation of a center-filled SNR or a Crab-like SNR.However, with only ∼ ⋆ E-mail: [email protected] data, one cannot unambiguously distinguish between thepower-law model and a single-temperature thermal plasmamodel with kT = 5 . +1 . − . keV (Schaudel 2003). Also,the non-detection of any radio emission from the positionof XMMU J185330.7-012815 makes the SNR interpretationquestionable.In an archival search for the Galactic magnetars, Munoet al. (2008) have made use of 506 archival Chandra data and441 archival XMM-Newton data. This search has included adedicated XMM-Newton observation of XMMU J185330.7-012815 with an off-axis angle of only ∼ . ∼
238 s from this observation. With this discovery, in-stead of being a magnetar candidate, Muno et al. (2008)have suggested this source is probably an accreting whitedwarf. The cataclysmic variable nature of this source is fur-ther supported by the optical spectroscopy performed by J.Halpern & E. Gotthelf (private communication reported inMuno et al. 2008).Although Muno et al. (2008) have identifiedXMMU J185330.7-012815 as a promising candidate ofcataclysmic variable, they have not further analysed anddiscussed the nature of this object as this is out of the scopeof their work. To confirm the X-ray emission properties ofXMMU J185330.7-012815, a detailed spectro-imaging andtemporal analysis of the aforementioned XMM-Newtonobservation is required and this provides the motivation of c (cid:13) Hui et al. this investigation. In §
2, we are going to describe the detailsof this XMM-Newton observation of XMMU J185330.7-012815 as well as the procedure of the data reduction. Themethod and the results of the data analysis are presented in §
3. Finally, we will discuss the implication and the possiblenature of XMMU J185330.7-012815 as an accreting whitedwarf.
XMMU J185330.7-012815 was observed by XMM-Newtonon 25-26 October 2004 (Observation ID: 0201500301).The X-ray data used in this investigation were obtainedwith the E uropean P hoton I maging C amera (EPIC) onboard XMM-Newton (Jansen et al. 2001). EPIC consists oftwo Metal Oxide Semiconductor (MOS1/2) CCD detectors(Turner et al. 2001) of which half of the beam from two of thethree X-ray telescopes is reflected to. The other two halves ofthe incoming photon beams are reflected to a grating spec-trometer (RGS) (den Herder et al. 2001). The third of thethree X-ray telescopes is dedicated to expose the EPIC-PNCCD detector solely (Str¨uder et al. 2001). The EPIC-PNCCD was operated in small-window mode with a mediumfilter to block optical stray light. This data provide imag-ing spectral and temporal information. All recorded eventsare time-tagged with a temporal resolution of 5.7 ms. TheMOS1/2 CCDs were setup to operate in full-window modewith a medium filter in each camera. The MOS1/2 camerasprovide imaging, spectral and timing information, thoughthe later with a temporal resolution of 2.6 s only.The aimpoint of the satellite in this observation isRA=18 h m . s and Dec= − ◦ ′ . ′′ (J2000). With themost updated instrumental calibration, we generate theevent lists from the raw data obtained from all EPIC in-struments with the tasks emproc and epproc of the XMMScience Analysis Software (XMMSAS version 9.0.0). Exam-ining the raw data from the EPIC-PN CCD, we did notfind any timing anomaly which was observed in many ofthe XMM-Newton data sets (cf. Hui & Becker 2006 and ref-erences therein). This provides us with opportunities for anaccurate timing analysis. We then created filtered event filesfor the energy range 0.2 keV to 12 keV for all EPIC instru-ments and selected only those events for which the patternwas between 0 −
12 for MOS cameras and 0 − omichain task. For imaging and source detection, the track historywas created, bad pixels were removed and the resulting im-age was subsequently used for source detection. Since in-dividual photons were centroid to one eighth of detectorpixel by onboard electronics which produces a noise pat-tern known as modulus 8 spatial fixed pattern noise andhence this noise was overcome using ommodmap task. The source detection was performed by using the task omdetect and counts were converted into magnitudes for the corre-sponding filters with the aid of the task ommag . The UVbandpass counts were converted into fluxes using the recipeprovided by Alice Breeveld. The composite MOS1/2 image of a 6 ′ × ′ fieldaround XMMU J185330.7-012815 is shown in Figure 1.XMMU J185330.7-012815 is observed as the brightestobject in this field. We determined its position and sig-nificance by means of a wavelet detection algorithm. TheX-ray position is found to be RA=18 h m . s ,Dec= − ◦ ′ . ′′ [J2000], where the numbersin the parentheses indicate the 1 σ statistical uncer-tainties of the last digit. The signal-to-noise ratio ofXMMU J185330.7-012815 is found to be 326 σ . ThisMOS1/2 image clearly rules out the claim of extendedsource as suggested by the RASS data (Schaudel 2003).There are two serendipitous X-ray sources found in thevicinity of XMMU J185330.7-012815, where are labeledas sources A and B in Figure 1. The wavelet detectionreports the locations and the significances of these sourcesto be (RA=18 h m . s , Dec= − ◦ ′ . ′′ [J2000]; S/N = 6 σ ) and (RA=18 h m . s ,Dec= − ◦ ′ . ′′ [J2000]; S/N = 17 σ ) for sources Aand B respectively. Since our focus is on characterising theemission nature of XMMU J185330.7-012815, we will notdiscuss the properties of these two sources further in thispaper. We estimated the effects of pileup in all EPIC data by usingthe XMMSAS task epatplot . Our results showed that all theEPIC data were not affected by CCD pileup. In order tomaximize the signal-to-noise ratio for XMMU J185330.7-012815, we extracted its source spectrum from circles with aradius of 50 ′′ and 30 ′′ in the MOS1/2 and EPIC-PN camerasrespectively. This choice of extraction regions corresponds tothe encircled energy fraction of ∼
90% in all cameras andat the same time it minimizes the contamination from thenearby X-ray sources. The background spectra were sampledfrom the nearby regions with circles of a radius of 60 ′′ and40 ′′ in MOS1/2 and EPIC-PN respectively. Response fileswere computed for all datasets by using the XMMSAS tasks rmfgen and arfgen . After the background subtraction, wehave 9130 ±
97 cts, 9314 ±
98 cts, and 19282 ±
140 cts collectedfor the spectral analysis from MOS1, MOS2 and EPIC-PNcameras respectively.In order to constrain the spectral parameters tightly, we http://xmm.esac.esa.int/sas/7.0.0/watchout/Evergreen tips and tricks/uvflux.shtml The source significances quoted in this paper are in units ofGehrels error: σ G = 1+ √ C B + 0 .
75 where C B is the backgroundcounts. http://xmm.esac.esa.int/external/xmm user sup-port/documentation/uhb/node17.htmlc (cid:13) , 1–9 MM observation of XMMU J185330.7-012815 Figure 1.
The raw X-ray image of the 6 ′ × ′ field-of-viewcentered at XMMU J185330.7-012815 generated by merging theMOS1 and MOS2 data in the energy range of 0 . −
12 keV. Twoserendipitous unidentified sources are also detected in this FOV. fitted the data obtained from three cameras simultaneously.For the spectrum extracted from each camera, we groupedthe data so as to have at least 50 counts per bin. We foundthere are fluctuations in the spectral data below 0.3 keVthat can be ascribed to the undesirable spectral response.Therefore, we limited all the spectral analysis in the energyrange 0 . −
12 keV. All the spectral fits are performed withXSPEC 12.5.1. The parameters of the best-fit models aresummarized in Table 1. All quoted errors are 1 σ for 2 pa-rameters of interest.The X-ray spectrum of XMMU J185330.7-012815 as ob-served by XMM-Newton is displayed in Figure 2 which hasshown the line features suggesting K-lines of heavy elements,in particular the features at ∼ . ∼ . n H = (4 . ± . × cm − and a plasma tem-perature of kT = 6 . ± . χ ν = 1 .
89 (583 D.O.F.). The large χ ν indicates that thismodel is unlikely to be the adequate description of the data.In examining the fitting residuals, we have identified thesystematic deviations at energies larger than ∼ ∼ n H = 4 . +0 . − . × cm − andthe plasma temperatures of kT = 1 . +0 . − . keV and kT = 8 . +0 . − . keV. In comparison with the single-temperaturemodel, the goodness-of-fit, χ ν = 1 .
27 (581 D.O.F.), is foundto be improved significantly. Statistically, the additionalcomponent is required at a confidence level > . > < ∼ ∼ . α and L lines, we take the abundance of iron as afitting parameter.By entangling the free parameters of Fe abundances inboth components, we found the residuals at ∼ ∼ . χ ν = 1 .
20, 580 D.O.F.). The model yields the column den-sity of n H = (5 . ± . × cm − , the plasma temper-atures of kT = 0 . ± .
03 keV and kT = 7 . +0 . − . keV,as well as an iron abundance of Fe=0 . ± .
08. Althoughthe residuals of the lines can be reduced with this model,with a more careful examination, we found that there arestill discrepancies between the observed data and this modelat energies larger than ∼ χ ν = 1 .
07, 579D.O.F.). Different from the two-temperature model, wedo not find any significant deviation of the Fe abundancefrom the solar value in this adopted model. Therefore, wefixed all the metal abundances at the solar values in three-temperature plasma model. It results in the column den-sity of n H = (5 . ± . × cm − , the plasma tem-peratures of kT = 0 . ± .
04 keV, kT = 1 . +0 . − . keVand kT = 10 . +1 . − . keV. The unabsorbed flux inferred bythis model is f X = (6 . ± . × − ergs cm − s − in0 . −
12 keV.We notice that there is a class of cataclysmic variablewhich contains a soft blackbody component (cf. Evans &Hellier 2007). Statistically the soft blackbody component isonly required at a confidence level of ∼
10% in this observedspectrum. Therefore, we do not consider this spectral con-tribution in this work.The ASCA and XMM-Newton spectra ofXMMU J185330.7-012815 appear to be qualitativelydifferent. ASCA observations show no line emission featuresand the spectrum with a single power-law model andobtained a reasonable goodness-of-fit with the ASCA data( χ ν = 0 .
86, 40 D.O.F.) (Schaudel 2003). However, withthe photon statistic improved by a factor of ∼
34, thespectral data obtained by XMM-Newton clearly show thepresence of the emission line features. Fitting a power-lawto the EPIC spectrum does not result in an acceptablegoodness-of-fit ( χ ν = 1 .
72, 583 D.O.F.). Similar to the c (cid:13) , 1–9 Hui et al. −3 no r m a li z ed c oun t s s − k e V − χ Energy (keV)
Figure 2.
Energy spectrum of XMMU J185330.7-012815 as ob-served with the EPIC-PN (upper spectrum) and MOS1/2 de-tectors (lower spectra) and simultaneously fitted to an absorbedthree-temperatures MEKAL model ( upper panel ) and contribu-tion to the χ fit statistic ( lower panel ). single-temperature plasma model, it cannot describe thedata below ∼ ∼ To confirm the X-ray period by Muno et al. (2008), we dida periodicity search by using an epoch folding method afterconverting the X-ray arrival times to the barycentric timesof the solar system. The best period, which was determinedby fitting a Gaussian function to the centroid of χ -peak,is P = 238 . ± . χ are included and the MOS1 and MOS2 data are com-bined). This period is consistent with the result reportedby Muno et al. (2008). If we take the possible irregularityof the pulses into account, then the error should increase to P = 238 . ± . σ of the Gaussian function fitted to the full χ -profile.Muno et al. (2008) have also pointed out that the phasesvary by ≈ . ≈ ≈ . +0 . − . , 1 . +0 . − . , 1 . +0 . − . , and 1 . +0 . − . in phases (where the errors are at the 90% confidence level).Taking the uncertainties into account, we are not able to un-ambiguously state whether there is a clear, systematic phaseshift. Furthermore, by running a χ -test among each pair ofthe time-sliced light curves in Figure 3, their distributionsare found to be consistent with each other at a confidencelevel > . Figure 3.
Pulse (or modulation) profiles for time-sliced PN data.The data ( plus signs ) were folded at the period 238.1 s from theepoch MJD 53303.98674. The pulse phase has been repeated overtwo cycles. A sine curve plus a constant ( dotted curve ) was fittedto the profiles. evidence for the incoherence can be found in our indepen-dent analysis.
A pulse or modulation profile folded at the period of 238.1 sshows a single broad peak over one cycle of the data. Itcan be approximated by a sinusoidal model plus a constantunpulsed level: A sin(2 π [ φ − φ ]) + C. (1)To see how the profiles vary with energy, we investigatethe energy-resolved profiles obtained in three different en-ergy bands, 0.2 − − −
10 keV. Forthis analysis, we focus on the PN dataset as it provides thesuperior photon statistic in each of the considered energybands. Figure 4 shows the profiles (plus signs), where thebackgrounds are not subtracted (the extracted backgroundsfrom a photon deficit region in the same CCD chip are verysmall and they are at most ∼
2% level compared to theirsource count rates). We fit the sinusoidal model to the pro-files with allowing all the parameters to be free and obtainthe following results from the best-fit parameters. Goodnessof the model-fit is given in each panel of the Figure 4. Theseprofiles show no significant phase shift in the chosen energybands. The pulse amplitude 2 A (cf. Equation 1) is found toincrease with increasing energy: 0 . ± .
03 (0.2 - 1.2 keV),0 . ± .
04 (1.2 - 3.0 keV), and 0 . ± .
06 (3.0 - 10 keV).In the other words, a pulsed (or modulation) fraction (i.e. c (cid:13) , 1–9 MM observation of XMMU J185330.7-012815 Figure 4.
Energy-resolved pulse profiles of XMMU J185330.7-012815 for PN data. The data were folded at the period 238.1 sfrom the observation start time. The intensity in each panelwas normalized by the average count rates of the same energyband. The pulse phase has been repeated over two cycles. A sinecurve plus a constant ( dotted curve ) was fitted to the profiles( plus signs ). A/C ) increases from 8% ±
2% to 23% ±
3% with increasingenergy. All the quoted errors for the temporal results are atthe 90% confidence level.
Within 10 ′′ from the X-ray position of XMMU J185330.7-012815, we have detected two UV sources in the OM data.One UV source (hereafter U1) has found to have the mag-nitudes of m UV W =15.99 ± UV M =16.49 ± σ and 53.46 σ respectively. On the other hand, the other source (here-after U2) has the magnitudes of m UV W =15.98 ± UV M =18.86 ± σ and 6.70 σ respectively. The corrected count rateof U1 and U2 for both filters are c =3.06 ± =0.82 ± =1.92 ± =0.06 ± × − erg cm − cts − ˚A − and 2.20 × − erg cm − cts − ˚A − , we obtain the fluxes of (4.36 ± × − erg cm − s − and (4.16 ± × − erg cm − s − for U1 in the corresponding filters. For U2, the fluxesare found to be (2.66 ± × − erg cm − s − and(0.30 ± × − erg cm − s − in UVW1 and UVM2 re-spectively. Of these two detected UV sources, U1 is found to haveits position coincided with the nominal X-ray position ofXMMU J185330.7-012815. Furthermore, while there is nooptical counterpart can be found for U2, we have identifiedan optical source in B − band at the position of U1 (see § In this paper, we present a detailed spectro-imaging andtemporal analysis of an archival XMM-Newton observationof XMMU J185330.7-012815, which has its emission naturenot yet been identified unambiguously.We found that the energy spectrum ofXMMU J185330.7-012815 obtained by the EPIC canbe well-described by an absorbed multi-temperature plasmamodel. The inferred column density is n H ∼ × cm − which is far lower than the total Galactic neutral hydrogenabsorption, n H ∼ cm − , in the direction towardsXMMU J185330.7-012815 (Kalberla et al. 2005; Dickey &Lockman 1990). This leads us to rule out the extragalacticorigin of XMMU J185330.7-012815.The unabsorbed X-ray flux inferred from the best-fitmodel is found to be f X ∼ × − ergs cm − s − . To-gether with the UV counterpart identified in OM, the in-ferred X-ray-to-UV flux ratio is f X /f UV ∼ .
5. To furtherconstrain the source nature, we also search for any opti-cal identification of XMMU J185330.7-012815 in the UnitedStates Naval Observatory (USNO)-B1.0 catalogue (Monetet al. 2003). Within the 20 σ X-ray positional uncertainty ofXMMU J185330.7-012815 (see § B = 17 .
11. The position of thissource is also found to be consistent with that of U1. By us-ing the n H inferred from the X-ray analysis to estimate theforeground extinction, we calculate the extinction-correctedoptical flux in B − band as f B = 1 . × − ergs cm − s − which implies f X /f B ∼ .
8. Such low value of X-ray-to-optical flux ratio can safely rule out the possibility as anisolated neutron star which typically has f X /f B > (cf.Haberl 2007).Through our independent search for the periodic X-raysignal from XMMU J185330.7-012815, we have identified apeak at ∼ . . This interpretationis further supported by the spectral properties. The X-ray http://asd.gsfc.nasa.gov/Koji.Mukai/iphome/catalog/alpha.htmlc (cid:13) , 1–9 Hui et al. spectrum of XMMU J185330.7-012815 can be described bya three-temperature plasma model and clearly shows thepresence of iron lines, which have often been seen in onetype of accreting white dwarf binaries, namely the inter-mediate polars (IPs) (e.g. Patterson 1994; Cropper et al.2002). The additional frequencies in the X-ray power spec-tra which corresponds to orbital or the beat period is absentin XMMU J185330.7-012815. In the context of an IP inter-pretation, this suggests that material accretes onto the polevia a disk.Although the spin period and the spectral propertiesstrongly favor the IP interpretation, the observed propertyof monotonically increasing modulation with increasing en-ergy in the light curves requires a further discussion. Incontrast to XMMU J185330.7-012815, many IPs show anopposite trend of decreasing modulation with energy (e.g.Norton & Watson 1989) which is generally explained by thevariation of photoelectric absorption in the accretion curtainacross the observer’s line of sight (Rosen, Mason, & C´ordova1988). The accretion rate is expected to have the maximumfrom the sector of the disc that is closest to the magneticpole and fall off with deviation from this sector. This givesrise to a continuous variation of column density along thecross-section of the curtain and hence results in the mod-ulation by photoelectric absorption. The minimum of thelight curve is thus at the phase when the accreting pole,where the absorption is the largest, is pointing toward theobserver. As the effect of photoelectric absorption is moreprominent in the soft band than in the hard band, a trendof decreasing pulsed fraction with increasing photon energyis not unexpected in this scenario (cf. Figure 9 in Rosen etal. 1988).Although the behaviour of decreasing amplitude withincreasing energy has been observed in many IPs, a numberof them do deviate from this general trend. For example, aclear increase in amplitude modulation with increasing en-ergy is also found in V2306 Cygni (WGA J1958.2+3232) bythe ASCA observation (Norton et al. 2002). Norton & Mukai(2007) found that an IP candidate XY Ari also shows anincreasing modulation with increasing energy in the XMM-Newton data, whereas it shows a decreasing behaviour withincreasing energy in the RXTE data (see Figure 1 and Figure5 in Norton & Mukai 2007). Another IP that shows a modu-lation different from the general trend is IGR J00234+6141in the RXTE observations (see Table 2 in Butters et al.2011). However, the XMM-Newton/EPIC-PN observationof this object shows no modulation above 2 keV (Anzolinet al. 2009). There is another IP, PQ Gem, which shows anincreasing modulation in sub-divided low energy bands anda decreasing modulation in high energy bands of ASCA andRXTE observations (James et al. 2002). On the other hand,a few other IPs show a constant modulation in RXTE ob-servations (e.g. Butters et al. 2007, 2008) and some show noamplitude modulation in the EPIC-PN energy band (e.g. deMartino et al. 2005). Since a diversity of temporal behaviourhas been observed, energy dependency of amplitude modu-lation is a weaker criterion to constrain the property of anIP. To further probe the temporal behaviour ofXMMU J185330.7-012815, we have also analysed thearchival RXTE observations for this source (ObsIDs: 90070-04-01-00, 90070-04-01-01, 90070-04-01-02, 90070-04-02-00, 90070-04-02-01 and 90070-04-02-02). However, no pulsedsignal from XMMU J185330.7-012815 was detected fromall these datasets. On the other hand, we found that theemission line features at ∼ . n H ∼ × cm − ) is much lower than typically observedin other IPs (cf. Ezuka & Ishida 1999). The absence of highvalues of absorption can possibly due to the low accretionrate. Such low absorption have also been seen in severalcases, such as HT Cam (de Martino et al. 2005), EX HYa(Mukai et al. 2003) and V1025 Cen (Hellier et al. 1998) alongwith a few other IP sources in which a strong soft compo-nent was always prominent (Evan & Hellier 2007). Never-theless, such soft component is not observed in the case ofXMMU J185330.7-012815. In the scenario of a low inclina-tion angle, the missing soft component may simply due tothe geometrical effect (Evan & Hellier 2007). Alternatively,this component might be much softer than the spectral cov-erage of XMM-Newton EPIC so that it is not revealed inthis observation.The spectral results of XMMU J185330.7-012815 sug-gest a relatively low shock temperature of ∼
11 keV (cf. Ta-ble 1). Very recently, Yuasa et al. (2010) have investigatedthe shock temperatures of 15 IPs and found two sourcesalso with low temperatures, FO Aqr ( ∼
14 keV) and EXHya ( ∼
12 keV). In another systematic study of 23 IPs, thelowest temperature is found to be ∼
12 keV for the sourceDO Dra (Brunschweiger et al. 2009). Furthermore, we noticethat the temperature of a new IP RX J0704 has exhibiteda change from a high value of >
44 keV to ∼
11 keV ineight months (Anzolin et al. 2008). There is an IP, AE Aqr,which shows exceptionally low temperatures plasma of 0.1 -7 keV (e.g. Choi et al. 1999; Choi & Dotani 2006; Mauche2009). However, for AE Aqr, it is widely believed that mostof the transferred material from the companion does notreach the magnetic poles due to a magnetic propeller effect(e.g., Wynn et al. 1997; Choi & Dotani 2006). Based onaforementioned temperature distribution in IPs, it is clearthat the observed temperature of XMMU J185330.7-012815is not significantly different from a few other IPs and alsothe observed low plasma temperature may not be persistent.A frequent monitoring of this source is encouraged.In summary, based on our temporal and spectral anal-ysis, we suggest that XMMU J185330.7-012815 belongs tothe IP sub-classification of cataclysmic variables. We wouldlike to stress that, including XY Ari, there are only 8 IPs/IPcandidates which show spin period less than 240 s. Owing to c (cid:13) , 1–9 MM observation of XMMU J185330.7-012815 this small population, the exact selection criteria to classifya source as an IP is not well-defined for this short-periodsub-class, which so far can only be constrained on the basisof the six observational criteria given by Patterson (1994).It is not certain if all six of these characteristics should beexhibited by an IP. For a further investigation of this newIP, a multi-wavelength observation campaign is required tounveil the physical and geometrical configuration of this sys-tem. In particular, the determination of the orbital periodof this source through a dedicated optical observation willdefinitely play a key role in determining the system param-eters such as orbital inclination. Furthermore, a dedicatedseries of X-ray observations is desirable to search for thepossible eclipses from XMMU J185330.7-012815. Throughthe timing of the eclipse ingress and egress, one is able todetermine the size of the X-ray emitting region and henceput an additional observational constraint on the emissionnature of this system. ACKNOWLEDGMENTS
The authors would like to thank the anonymous referee forthe useful comments. CYH is supported by the research fundof Chungnam National University in 2011.
REFERENCES
Anzolin G., de Martino D., Bonnet-Bidaud J. M., MouchetM., Gansicke B. T., Matt G., Mukai K., 2008, A&A 489,1243Anzolin G., de Martino D., Falanga M., Mukai K., Bonnet-Bidaud J.-M., Mouchet M., Terada Y., Ishida M., 2009,A&A, 501, 1047Brunschweiger J., Greiner J., Ajello M., Osborne M., 2009,A&A 496, 121Butters O. W., Barlow E. J., Norton A. J., Mukai K., 2007,A&A, 475, L29Butters O. W., Norton A. J., Hakala P., Mukai K., BarlowE. J., 2008, A&A, 487, 271Butters O. W., Norton A. J., Mukai K., Tomsick J. A.,2011, A&A 526 77Choi C. S., Dotani T., Agrawal P. C., 1999, ApJ, 525, 399Choi C. S., Dotani T., 2006, ApJ, 646, 1149Cropper M., Ramsay G., Hellier C., Mukai K., Mauche C.,Pandel D., 2002, Proc. R. Soc. London A, 360, 1951den Herder et al., 2001, A&A, 365, 7de Martino D., Matt G., Mukai K., Bonnet-Bidaud J.-M.,Gansicke B. T., Gonzalez Perez J. M., Haberl F., MouchetM., Solheim J.-E., 2005, A&A, 437, 935Dickey J. M., Lockman F. J., 1990, ARAA, 28, 215Evans P. A., Hellier C., 2007, ApJ, 663, 1277Ezuka H., Ishida M., 1999, ApJS, 120, 277Haberl F., 2007, Ap&SS, 308, 181Harrison F. A., et al., 2010, SPIE, 7732, 27Hellier C., Beardmore A. P., Buckley, D. A. H., 1998, MN-RAS, 29, 851Hui C. Y., Becker W., 2006, A&A, 454, 543James Cynthia H., Ramsay G., Cropper M., Branduardi-Raymont G., 2002, MNRAS, 336, 550Jansen F., et al., 2001, A&A, 365, L1 Kalberla P. M. W. et al., 2005, A&A, 440, 775Manchester R. N., Hobbs G. B., Teoh A., Hobbs M., 2005,AJ, 129, 1993Mason K. O., et al., 2001, A&A, 365, L36Mauche C. W., 2009, ApJ, 706, 130Mewe R., Gronenschild E. H. B. M., van den Oord G. H.J., 1985, A&AS, 62, 197Monet D., et al., 2003, AJ, 125, 984Mukai K., Kinkhabwala A., Peterson J. R., Kahn S. M.,Paerels F., 2003, ApJ, 586, L77Muno M. P., Gaensler B. M., Nechita A., Miller J. M.,Slane, P. O., 2008, ApJ, 680, 639Norton A. J., Watson M. G., 1989, MNRAS, 237, 853Norton A. J., Quaintrell H., Katajainen S., Lehto H. J.,Mukai K., Negueruela I., 2002, A&A, 384, 195Norton A. J., Mukai K., 2007, A&A, 472, 225Patterson J., 1994, PASP, 106, 209Ramsay G., Wheatley P. J., Norton A. J., Hakala P., BaskillD., 2008, MNRAS, 387, 1157Rosen S. R., Mason K. O., C´ordova F. A., 1988, MNRAS,231, 549Schaudel D., 2003, PhD thesis, Ludwig-Maximilians Uni-versit¨at M¨unchenStr¨uder L., et al., 2001, A&A, 365, L18Turner M. J. L., et al., 2001, A&A, 365, L27Wynn G. A., King A. R., Horne, K., 1997, MNRAS, 286,436Yuasa T., Nakazawa K., Makishima K., Saitou K., IshidaM., Ebisawa K., Mori H., Yamada S., 2010, A&A, 520, 25 c (cid:13) , 1–9 H u i e t a l . Table 1.
Best-fit spectral parameters of XMMU J185330.7-012815.
MEKAL MEKAL+MEKAL MEKAL+MEKAL MEKAL+MEKAL+MEKAL PL n H (10 cm − ) 4 . ± .
31 4 . +0 . − . . +0 . − . . +0 . − . . +0 . − . kT (keV) 6 . ± .
23 1 . +0 . − . . ± .
03 0 . ± .
04 - kT (keV) - 8 . +0 . − . . +0 . − . . +0 . − . - kT (keV) - - - 10 . +1 . − . -Fe a - 1.0 (fixed) 0 . ± .
08 1.0 (fixed) -Γ - - - - 1 . ± . (2 . ± . × − (1 . +0 . − . ) × − (1 . +0 . − . ) × − (8 . +1 . − . ) × − -Norm - (2 . +0 . − . ) × − (2 . ± . × − (3 . +1 . − . ) × − -Norm - - - (2 . +0 . − . ) × − -Norm PL c - - - - (1 . ± . × − χ a The abundance of iron relative to the solar photospheric values. b The normalization of MEKAL model is expressed as (10 − / πD ) R N e N H dV where D is the source distance in cm and N e and N H are the electron and hydrogen densities in cm − . c The normalization of the power-law model (PL) at 1 keV in units of photons keV − cm − s − . c (cid:13) R A S , M N R A S , MM observation of XMMU J185330.7-012815 This paper has been typeset from a TEX/ L A TEX file preparedby the author. c (cid:13)000