The Be Star HD 215227: A Candidate Gamma-ray Binary
S.J. Williams, D.R. Gies, R.A. Matson, Y. Touhami, E.D. Grundstrom, W. Huang, M.V. McSwain
aa r X i v : . [ a s t r o - ph . S R ] S e p V2. 09/03/2010 – Submitted to ApJL
The Be Star HD 215227: A Candidate Gamma-ray Binary
S. J. Williams , D. R. Gies, R. A. Matson, Y. Touhami Center for High Angular Resolution Astronomy and Department of Physics andAstronomy, Georgia State University, P. O. Box 4106, Atlanta, GA 30302-4106;[email protected], [email protected], [email protected],[email protected]
E. D. Grundstrom
Physics and Astronomy Department, Vanderbilt University, 6301 Stevens Center,Nashville, TN 37235; [email protected]
W. Huang
Department of Astronomy, University of Washington, Box 351580, Seattle, WA98195-1580; [email protected]
M. V. McSwain
Department of Physics, Lehigh University, 16 Memorial Drive E., Bethlehem, PA 18015;[email protected]
ABSTRACT
The emission-line Be star HD 215227 lies within the positional error circle ofthe newly identified gamma-ray source AGL J2241+4454. We present new bluespectra of the star, and we point out the morphological and variability similaritiesto other Be binaries. An analysis of the available optical photometry indicatesa variation with a period of 60 . ± .
04 d, which may correspond to an orbitalmodulation of the flux from the disk surrounding the Be star. The distance to thestar of 2.6 kpc and its relatively large Galactic latitude suggest that the binarywas ejected from the plane by a supernova explosion that created the neutronstar or black hole companion. The binary and runaway properties of HD 215227make it an attractive candidate as the optical counterpart of AGL J2241+4454and as a new member of the small class of gamma-ray emitting binaries. Guest investigator, Dominion Astrophysical Observatory, Herzberg Institute of Astrophysics, NationalResearch Council of Canada.
Subject headings: stars: emission-line, Be — stars: early-type — stars: evolution— gamma rays: stars — stars: individual (HD 215227; AGL J2241+4454)
1. Introduction
Gamma-ray binaries are a class of high energy and very-high energy emitting sourcesthat consist of a massive star and compact companion (Mirabel 2007; Dubus et al. 2010;McSwain 2010). Six such objects are known sources of TeV emission: LS 5039, Cygnus X-1, Cygnus X-3, LS I +61 303, PSR B1259 −
63, and HESS J0632+057. The massive starcomponent is a luminous O-star in the first two, a probable WR star in the third, and aBe star in the last three cases. All these massive stars have winds and the Be stars alsoeject mass into an outflowing circumstellar disk. The interaction of this mass loss witha degenerate companion can lead to gamma-ray emission in several ways (Parades 2008).First, if the companion is a pulsar, then a high speed wind from the mass donor can collidewith the pulsar wind in a shock region, and inverse Compton scattering of stellar photonswith relativistic electrons in the shock can create gamma-rays (Dubus et al. 2010). Second,if the companion is a black hole, then gas accretion can lead to the formation of relativisticjets, and gamma-ray emission may occur by inverse Compton scattering from jet electronsand/or by the decay of neutral pions that originate in inelastic proton - proton collisions(Romero et al. 2007).Recently, Lucarelli et al. (2010) announced the detection with the AGILE satellite (Ta-vani et al. 2009) of gamma-ray emission above 100 MeV from a new unidentified source,AGL J2241+4454. The source has Galactic coordinates of ( l, b ) = (100 . ◦ , − . ◦
2) with anerror circle radius of approximately 0 . ◦
6. The source has not yet been detected by the NASAFermi Gamma-Ray Observatory . The AGILE point sources found to date include pulsars,blazars, supernova remnants, and high mass X-ray binaries (Pittori et al. 2009), but there areno cataloged examples of any of these in the region close to AGL J2241+4454 (Lucarelli etal. 2010). We point out two possibilities that deserve further attention. First, Brinkmann etal. (1997) found an X-ray source and probable quasar in this region, RXJ2243 . V ≈
16 (based upon the observed X-ray flux and an assumed frequency power law with α = 1 . Fermi LAT Report, 2010 July 30; http://fermisky.blogspot.com/ ◦ l, b ) = (100 . ◦ , − . ◦ γ (Petrie & Lee 1965), and back to B0 (Hern´andez et al. 2005). The classifica-tion is difficult because the spectral lines are very broad and weak (the projected rotationalvelocity is V sin i = 262 ±
26 km s − ; Yudin 2001) and often blended with emission features.Here we present new blue spectra of the target, and we argue that HD 215227 is a poten-tial optical counterpart of AGL J2241+4454 based upon its probable runaway and binarycharacter.
2. Spectroscopic Observations
We obtained blue spectra of HD 215227 with the HIA Dominion Astrophysical Obser-vatory 1.8 m telescope on 2010 July 28 and 29. These observations were made with theCassegrain spectrograph with grating 1200B (1200 grooves mm − ) in first order, and thespectra cover the range 4260 – 4669 ˚A. The detector was the SITe-2 CCD (a 1752 × × µ m pixels), and the resulting spectra have a resolving power of R = λ/ ∆ λ = 4290 as measured from the Fe Ar comparison lines. Exposures were 300 sin duration, leading to spectra with a S/N = 100 per pixel in the continuum. The spectrawere extracted and calibrated using standard routines in IRAF , and then each continuum-rectified spectrum was transformed to a common heliocentric wavelength grid in log λ incre-ments.The two spectra are illustrated in Figure 1. There is clear evidence of double-peakedemission from the circumstellar disk that is seen in the core of H γ and in numerous Fe II emis-sion lines. The only clear photospheric lines are those of C II λ I λλ , γ λ IRAF is distributed by the National Optical Astronomy Observatory, which is operated by the Associa-tion of Universities for Research in Astronomy, Inc., under cooperative agreement with the National ScienceFoundation. V sin i was estimated from the shape of the He I λ II λ T eff derivation was based primarilyon the He I λ II λ T eff , although the ratio is almost independent of T eff near T eff = 21 kK, which correspondsto the peak of the He I strength. However, we can estimate T eff even in this vicinity by therelative strength of the O II λ T eff . The gravity log g was determined by fitting the emission-free parts of the Stark broadened, H γ line wings. Wefound that the model photospheric spectrum had line depths that were consistently deeperthan the observed ones, and this is probably due to disk continuum flux in this spectralregion. Consequently, we renormalized the model spectrum by adding a pure continuumcomponent with a disk-to-star, monochromatic flux ratio F d /F ⋆ . This renormalized modelspectrum is shown as the lower plot in Figure 1. Finally, we estimated the radial velocity V r for both spectra by cross-correlating the observed and model spectra over the wavelengthrange including the emission-free H γ wings and He I λ γ . Note that a similar shell feature may be present in He I λ V sin i = 30 km s − , compared to our result. We suspectthat their spectrum was obtained at a time when the shell lines dominated and that theirmeasurement corresponds to the shell line width. 5 –
3. Discussion
The spectral properties of HD 215227 bear some resemblance to those of the gamma-raybinary LS I +61 303, which is a Be star in a 26.4960 d orbit (Aragona et al. 2009). TheH α profile in LS I +61 303 displays systematic variations in the ratio of the violet-to-red( V /R ) peak emission around the time of periastron in this eccentric orbit ( e = 0 .
54) binary(Grundstrom et al. 2007; McSwain et al. 2010). We find that the
V /R ratio of the H γ emission changed significantly in just one day (Fig. 1), which is unusually fast for most Bestars (Grundstrom 2007). We suggest that such rapid variability might be associated withthe changing tidal effects of a companion on the disk (especially strong near periastron).If the disk in HD 215227 is modulated with a binary orbit, then the continuum flux fromthe disk may also show an orbital modulation. We found that the disk contributes ≈ Hipparcos light curve (258 measurements). There are two other setsof photometric measurements from all sky survey experiments. The first set of 101 Cousins I C measurements were made between 2003 and 2007 by The Amateur Sky Survey (TASS ;Droege et al. 2006). The second set of 52 points were made from 1999 to 2000 with theNorthern Sky Variability Survey (NSVS ; Wo´zniak et al. 2004). Since these three sets weremade with different broad-band filters, we assumed that color variations are minimal andthen simply subtracted the mean magnitude of each set to form a combined photometrictime series. These residual magnitudes are based upon differences from the average valuesof < H p > = 8 . < I C > = 8 .
52, and < m
V ,
ROTSE > = 9 .
12 for the
Hipparcos , TASS, andNSVS photometry sets, respectively. A discrete Fourier transform period search revealedone significant signal with a period of 60 . ± .
04 d, an epoch of maximum brightness atHJD 2,453,243.3 ± .
8, and a semiamplitude of 0 . ± .
002 mag. The resulting light curve(Fig. 2) displays a low-amplitude, quasi-sinusoidal variation that probably corresponds tothe binary orbital period. Note that the small radial velocity changes we observed over a oneday interval are consistent with the small variations that are expected for an orbital periodthis long.We note that the Galactic latitude of HD 215227, b = − ◦ , suggests that the star isquite far from the Galactic plane, and hence it may be a runaway star formed by a supernova(SN) explosion in a binary system (Gies & Bolton 1986; Hoogerwerf et al. 2000). In this http://sallman.tass-survey.org/servlet/markiv/ http://skydot.lanl.gov/nsvs/nsvs.php T eff and log g parameters (Table 1). We constructed the SED using UV fluxes from the TD1satellite (Thompson et al. 1978), Johnson magnitudes from Mermilliod (1991; transformedto flux using the calibration of Colina et al. 1996), 2MASS magnitudes (Skrutskie et al.2006; transformed to flux according to the calibration of Cohen et al. 2003), and an AKARI9 µ m measurement (Ishihara et al. 2010). The SED (Fig. 3) indicates that there is a strongIR-excess from the Be star’s disk that extends into the optical range. In order to fit thephotosphere of the star alone, we restricted the range to the UV fluxes and B -band revisedflux without the disk contribution (the lower point plotted at 4443 ˚A in Fig. 3). We adopteda theoretical flux spectrum from the models of R. L. Kurucz for solar metallicity, T eff andlog g from Table 1, and a microturbulent velocity of 2 km s − . This flux spectrum was fit tothe restricted set of observations with two parameters: the limb-darkened, angular diameter θ LD and the reddening E ( B − V ) (assuming a ratio of total-to-selective extinction R V = 3 . f λ (obs) = θ LD F λ (mod)10 − . A λ where A λ is the wavelength dependent extinction (Fitzpatrick 1999). The first parameter θ LD acts as a normalizing factor while E ( B − V ) defines how the shape of the SED is alteredby the extinction A λ . The results for these two fitting parameters are given in Table 1, andthe derived f λ spectrum is plotted as a solid line in Figure 3. We find a very low reddeningalong this line of sight, E ( B − V ) = 0 .
02 mag. Earlier estimates consistently arrive at ahigher reddening of E ( B − V ) ≈ . B − V colorand ignore the disk contribution in the optical that makes the star appear too red (Fig. 3). http://kurucz.harvard.edu/grids.html θ LD with the evolutionary radius R ⋆ yields alarge distance, d = 2 . ± . z = − . ± .
20 kpc. This is an extreme distance for a normal OB star. For example, inthe Be star kinematical survey by Berger & Gies (2001), the mean distance is < | z | > = 69 pcand only one other star (HD 20340 at z = − .
71 kpc) out of a sample of 344 has a distancefrom the plane as large as that of HD 215227. This suggests that HD 215227 is a runawaystar that probably obtained its initial high velocity and current position by a supernovaexplosion in a binary. We adopted the proper motions from van Leeuwen (2007) and theaverage radial velocity and distance from Table 1 to estimate the star’s current peculiarvalues of tangential V T p , radial V R p , and space velocity V S p using the method described byBerger & Gies (2001). These peculiar velocities are measured relative to the star’s localstandard of rest by accounting for the Sun’s motion in the Galaxy and differential Galacticrotation. Our estimates of the peculiar velocities (Table 1) are not unusually large, but theerrors are significant and the star may have decelerated in the gravitational potential of theGalaxy.The similarity of this Be binary to other gamma-ray binaries, its probable runawaystatus, and its close proximity to the gamma-ray source location all indicate that HD 215227may be the optical counterpart of AGL J2241+4454. We encourage additional spectroscopicobservations to search for orbital motion and to document the emission line variability aroundthe orbit. New X-ray and radio observations with good angular resolution will be essentialto secure the identification of the target (see the case of HESS J0632+057; Hinton et al.2009; Skilton et al. 2009). Furthermore, a search for periodic gamma-ray brightening on the60 d cycle would offer conclusive evidence of the connection with the Be star. If the orbitaleccentricity is high, the gamma-ray emission may be restricted to a limited part of the orbit.We thank Dr. Dmitry Monin and the other staff of the Dominion Astrophysical Obser-vatory, Herzberg Institute of Astrophysics, National Research Council of Canada, for theirassistance in making these observations possible. This material is based on work supportedby the National Science Foundation under Grant AST-0606861. Institutional support hasbeen provided from the GSU College of Arts and Sciences and from the Research ProgramEnhancement fund of the Board of Regents of the University System of Georgia, adminis-tered through the GSU Office of the Vice President for Research. We gratefully acknowledgeall this support. Facility:
DAO:1.85m 8 –
REFERENCES
Aragona, C., et al. 2009, ApJ, 698, 514Berger, D. H., & Gies, D. R. 2001, ApJ, 555, 364B¨ohm-Vitense, E. 1981, ARA&A, 19, 295Brinkmann, W., et al. 1997, A&A, 323, 739Cohen, M., Wheaton, W. A., & Megeath, S. T. 2003, AJ, 126, 1090Colina, L., Bohlin, R., & Castelli, F. 1996, HST Instrument Science Report CAL/SCS-008(Baltimore: STScI)Droege, T. F., Richmond, M. W., & Sallman, M. 2006, PASP, 118, 1666Dubus, G., Cerutti, B., & Henri, G. 2010, A&A, 516, A18Fitzpatrick, E. L. 1999, PASP, 111, 63Gies, D. R., & Bolton, C. T. 1986, ApJS, 61, 419Grundstrom, E. D. 2007, Ph.D. dissertation, Georgia State Univ.Grundstrom, E. D., et al. 2007, ApJ, 656, 437Hanuschik, R. W. 1995, A&A, 295, 423Harris, D. L., III 1955, ApJ, 121, 554Hern´andez, J., Calvet, N., Hartmann, L., Brice˜no, C., Sicilia-Aguilar, A., & Berlind, P. 2005,AJ, 129, 856Hinton, J. A., et al. 2009, ApJ, 690, L101Hoogerwerf, R., de Bruijne, J. H. J., & de Zeeuw, P. T. 2000, ApJ, 544, L133Huang, W., & Gies, D. R. 2008, ApJ, 683, 1045Ishihara, D., et al. 2010, A&A, 514, A1Koen, C., & Eyer, L. 2002, MNRAS, 331, 45Lanz, T., & Hubeny, I. 2007, ApJS, 169, 83Lucarelli, F., et al. 2010, ATel, 2761 9 –McSwain, M. V. 2010, Be Star Newsletter, 40, in pressMcSwain, M. V., Grundstrom, E. D., Gies, D. R., & Ray, P. S. 2010, ApJ, submittedMermilliod, J. C. 1991, Catalogue of Homogeneous Means in the UBV System (Lusanne:Inst. d’Astronomie, Univ. de Lausanne)Merrill, P. W., & Burwell, C. G. 1943, ApJ, 98, 153Mirabel, I. F. 2007, Ap&SS, 309, 267Neckel, Th., & Klare, G. 1980, A&AS, 42, 251Oktariani, F., & Okazaki, A. T. 2009, PASJ, 61, 57Paredes, J. M. 2008, in High Energy Gamma-Ray Astronomy: Proc. 4th International Meet-ing on High Energy Gamma-Ray Astronomy, AIP Conf. Proc. 1085, ed. F. A. Aha-ronian, W. Hofmann, & F. Rieger (Melville, NY: AIP), 157Peters, G. J. 1983, PASP, 95, 311Petrie, R. M., & Lee, E. K. 1965, Publ. Dominion Astrophys. Obs., 12, 435Pittori, C., et al. 2009, A&A, 506, 1563Romero, G. E., Okazaki, A. T., Orellana, M., & Owocki, S. P. 2007, A&A, 474, 15Schaller, G., Schaerer, D., Meynet, G., & Maeder, A. 1992, A&AS, 96, 269Skilton, J. L., et al. 2009, MNRAS, 399, 317Skrutskie, M. F., et al. 2006, AJ, 131, 1163Snow, T. P., Jr., York, D. G., & Welty, D. E. 1977, AJ, 82, 113Tavani, M., et al. 2009, A&A, 502, 995Thompson, G. I., Nandy, K., Jamar, C., Monfils, A., Houziaux, L., Carnochan, D. J.,& Wilson R. 1978, Catalogue of stellar ultraviolet fluxes (TD1): A compilation ofabsolute stellar fluxes measured by the Sky Survey Telescope (S2/68) aboard theESRO satellite TD-1 (London: The Science Research Council, U.K.)van Leeuwen, F. 2007, A&A, 474, 653Vercellone, S., et al. 2008, ApJ, 676, L13 10 –Wolff, S. C., Strom, S. E., Dror, D., & Venn, K. 2007, AJ, 133, 1092Wo´zniak, P. R., et al. 2004, AJ, 127, 2436Yudin, R. V. 2001, A&A, 368, 912Zhang, P., Chen P. S., & Yang, H. T. 2005, NewA, 10,325
This preprint was prepared with the AAS L A TEX macros v5.2.
11 –Table 1. Stellar PropertiesParameter Value T eff (kK) . . . . . . . . . . . . . . . . . . . . . . . . 19 ± g (cm s − ) . . . . . . . . . . . . . . . . . . . 3 . ± . V sin i (km s − ) . . . . . . . . . . . . . . . . . 300 ± V r (HJD 2,455,405.9461) (km s − ) 0 . ± . V r (HJD 2,455,406.9124) (km s − ) 2 . ± . F d /F ⋆ . . . . . . . . . . . . . . . . . . . . . . . . . . . 0 . ± . E ( B − V ) (mag) . . . . . . . . . . . . . . . . 0 . ± . θ LD (10 − arcsec) . . . . . . . . . . . . . . . 24 ± M ⋆ ( M ⊙ ) . . . . . . . . . . . . . . . . . . . . . . . 7 . ± . R ⋆ ( R ⊙ ) . . . . . . . . . . . . . . . . . . . . . . . . 6 . ± . d (kpc) . . . . . . . . . . . . . . . . . . . . . . . . . 2 . ± . z (kpc) . . . . . . . . . . . . . . . . . . . . . . . . . − . ± . V T p (km s − ) . . . . . . . . . . . . . . . . . . . . 19 ± V R p (km s − ) . . . . . . . . . . . . . . . . . . . . 21 ± V S p (km s − ) . . . . . . . . . . . . . . . . . . . . 28 ±
24 12 – λ (Å)0.40.60.81.01.2 R E L A T I VE F L U X HR 2142HJD 2,455,405.9HJD 2,455,406.9MODEL
Fig. 1.— The continuum normalized spectrum of HD 215227 from the first and second nights(the latter is offset by − .
20 for clarity). Above is a similar spectrum of the Be binaryHR 2142 (Grundstrom 2007) and below is a TLUSTY/SYNSPEC synthetic spectrum forthe parameters listed in Table 1 (offset by +0 .
25 and − .