1: %barb dop comments di scott 15/4/08
2: %frf e barb 10/4/08
3: %ap 12/3/08
4:
5: %frf 10/3/2008-12/3/08
6:
7: %barb 26/2/08
8:
9: %ap 26/2/08
10:
11: %frf 30/1/08-19/2/08
12:
13: %bfrf 22/1/08
14:
15: %barb 18/1/08
16:
17: %\documentclass[preprint]{aastex}
18:
19: %\documentclass[12pt,referee]{../apj/aastex502/aastex}
20:
21: \documentclass[12pt,preprint]{aastex}
22:
23: %\documentstyle[emulateapj5,epsf]{aastex}
24:
25:
26:
27: \def\alwaysmath#1{\ifmmode{#1}\else{$#1$}\fi}
28: \def\msun{\alwaysmath{{M}_{\odot}}}
29: \def\teff{\alwaysmath{T_{\rm eff}}}
30: \def\arcsec{\hbox{$^{\prime\prime}$}}
31: \def\etal{{et al.~}}
32: \def\kelvin{{\rm \, K}}
33: \def\pulb{PSR~J1701$-$3006B~}
34: \def\compulb{COM~J1701$-$3006B~}
35: \def\kms{\alwaysmath{{{\rm \,km\,sec^{-1}}}}}
36: \def\ersa{\alwaysmath{{\rm \, erg sec^{-1}/\AA}}}
37: \def\ers{\alwaysmath{{\rm \, erg\,sec^{-1}}}}
38: \newcommand\HST{{\sl HST}}
39: \newcommand\hst{{\sl HST}}
40: %\newcommand\ROSAT{{\sl ROSAT}}
41:
42: \newcommand\CHANDRA{{\it Chandra }}
43: \newcommand\SYNPHOT{{\sl SYNPHOT~}}
44: \newcommand\PC{{\sl PC~}}
45: %\newcommand\WFPC2{{\sl WFPC2}}
46:
47: %\newcommand\FOC{{\sl FOC}}
48:
49: %\newcommand\STIS{{\sl STIS}}
50:
51: %\newcommand\ROMAFOT{{\sl ROMAFOT}}
52:
53: %\newcommand\METRIC{{\sl METRIC}}
54:
55: %\newcommand\IRAF{{\sl IRAF}}
56:
57: %\newcommand\ngc[1]{NGC\,#1}
58:
59: %\setlength{\voffset}{-1.5truecm}
60:
61: \newcommand{\gras}[1]{\mbox{\boldmath $#1$}}
62:
63: \def\ltsima{$\; \buildrel < \over \sim \;$}
64:
65: \def\gtsima{$\; \buildrel > \over \sim \;$}
66:
67: \def\lsim{\lower.5ex\hbox{\ltsima}}
68: \def\gsim{\lower.5ex\hbox{\gtsima}}
69: \def\lapp{\ifmmode\stackrel{<}{_{\sim}}\else$\stackrel{<}{_{\sim}}$\fi}
70: \def\gapp{\ifmmode\stackrel{>}{_{\sim}}\else$\stackrel{<}{_{\sim}}$\fi}
71:
72: \slugcomment{Submitted to ApJ Letters}
73:
74: \shorttitle{The optical companion to PSRJ1701-3006B in NGC\,6266}
75: \shortauthors{Cocozza et al.}
76:
77: \begin{document}
78:
79: \title{A puzzling millisecond pulsar companion in
80: NGC\,6266\altaffilmark{1}}
81:
82: \author{
83: G. Cocozza\altaffilmark{2,3},
84: F. R. Ferraro\altaffilmark{2},
85: A. Possenti\altaffilmark{4},
86: G. Beccari\altaffilmark{2},
87: B. Lanzoni\altaffilmark{2},
88: S. Ranson\altaffilmark{5},
89: R. T. Rood\altaffilmark{6},
90: N. D'Amico\altaffilmark{4,7}}
91: %\medskip
92:
93: \altaffiltext{1}{Based on observations with the NASA/ESA HST,
94: obtained at the Space Telescope Science Institute, which is operated
95: by AURA, Inc., under NASA contract NAS5-26555.}
96: \altaffiltext{2}{Dipartimento di Astronomia Universit\`a
97: di Bologna, via Ranzani 1, I--40127 Bologna, Italy,}
98: \altaffiltext{3}{INAF, Osservatorio Astronomico di Bologna,
99: via Ranzani 1, I--40127 Bologna, Italy}
100: \altaffiltext{4}{INAF, Osservatorio Astronomico di Cagliari,
101: Loc. Poggio dei Pini, Strada 54, I--09012 Capoterra, Italy}
102: \altaffiltext{5}{National Astronomy Observatory, 520 Edgemont Road,
103: Charlottesville, VA, 22903, USA}
104: \altaffiltext{6}{Astronomy Department, University of Virginia,
105: P.O. Box 400325, Charlottesville, VA, 22904, USA}
106: \altaffiltext{7}{Dipartimento di Fisica Universit\`a di Cagliari,
107: Cittadella Universitaria, I-09042 Monserrato, Italy}
108:
109: %\date{2008, March 12}
110:
111: \begin{abstract}
112: We report on the optical identification of the companion to the
113: eclipsing millisecond pulsar PSR\,J1701$-$3006B in the globular
114: cluster NGC\,6266. A relatively bright star with an anomalous red
115: colour and an optical variability ($\sim$ 0.2 mag) that nicely
116: correlates with the orbital period of the pulsar ($\sim$ 0.144 days)
117: has been found nearly coincident with the pulsar nominal
118: position. This star is also found to lie within the error box position
119: of an X-ray source detected by Chandra observations, thus supporting
120: the hypothesis that some interaction is occurring between the pulsar
121: wind and the gas streaming off the companion. Although the shape of
122: the optical light curve is suggestive of a tidally deformed star which
123: has nearly completely filled its Roche lobe, the luminosity ($\sim
124: 1.9\,L_\odot$) and the surface temperature ($\sim 6000$\,K) of the
125: star, deduced from the observed magnitude and colours, would imply a
126: stellar radius significantly larger than the Roche lobe radius.
127: Possible explanations for this apparent inconsistency are discussed.
128: \end{abstract}
129:
130: \keywords{Globular clusters: individual (NGC\,6266); stars: evolution
131: -- pulsars: individual (PSR\,J1701$-$3006B) -- binaries: close.}
132:
133: \section{Introduction}
134: \label{sec:intro}
135:
136: NGC\,6266 (M62) is one of the most massive and brightest ($M_V=-9.19$;
137: Harris 1996\footnote{Everywhere in this paper, we refer to the
138: data at {\tt http://www.physics.mcmaster.ca/Globular.html}})
139: Galactic Globular Clusters (GCs), also is characterized by high values of
140: the central density (log $\rho_0 \sim 5.47$, with $\rho_0$ in units of
141: $M_\odot\,$pc$^{-3}$; Beccari et al. 2006). It displays a King-model
142: density profile with an extended core ($\sim 19\arcsec$) and a modest
143: value of the concentration parameter ($c=1.5$; Beccari et al. 2006).
144:
145: Six binary millisecond pulsars (MSPs) have been discovered in M62
146: (D'Amico et al. 2001\nocite{damico01}; Jacoby et
147: al. 2002\nocite{jacoby02}; Possenti et al. 2003\nocite{P03}, hereafter
148: P03) and it ranks fifth of the GCs in wealth of MSPs, after Terzan 5,
149: 47 Tucanae, M15 and M28\footnote{See {\tt
150: http://www.naic.edu/{~}pfreire/GCpsr.html} for a catalog of MSPs
151: in GCs with relative references}. P03 presented phase--connected
152: timing solutions, yielding precise celestial coordinates, for three of
153: the MSPs in NGC\,6266. One of them, PSR\,J1701$-$3006B (hereafter
154: PSR\,6266B) displays partial or total eclipses of the radio signal at
155: 1.4 GHz near its superior conjunction (in the convection adopted
156: throughout this paper, this corresponds to orbital phase $\phi
157: =0.25$), clearly due to gas streaming off the companion. The pulsar
158: orbit is circular, with a projected semi-major axis of only $\sim
159: 0.11\,R_\odot$. P03 suggested two options for explaining the
160: behaviour of the PSR\,6266B system. The first option is that the
161: pulsar companion is a non-degenerate bloated star, whose mass loss is
162: sustained by ablation of its loosely bound surface layers by the
163: relativistic wind emitted by the pulsar. In this case, PSR\,6266B may
164: resemble PSRs B1957+20 (Fruchter et al. 1990\nocite{fbb+90}) and
165: J2051$-$0827 (Stappers et al. 2001\nocite{sbl+01}): the optical light
166: curve of their companion star presents a maximum when the side of the
167: companion facing the pulsar points toward the observer (i.e., at the
168: pulsar inferior conjunction: $\phi=0.75$). This is a clear signature
169: of the irradiation of the companion surface by the pulsar flux. The
170: second option has the pulsar companion as a tidally deformed star
171: overflowing its Roche lobe due to the internal nuclear evolution. In
172: this case, the system may be more like PSR\,J1740$-$5340 (D'Amico et
173: al. 2001\nocite{damico01}; Ferraro et al. 2001\nocite{ferraro01}),
174: where irradiation effects are negligible (Orosz \& van Kerkwijk
175: 2003\nocite{ov03}) and the optical light curve of the companion is
176: dominated by ellipsoidal variations (Ferraro et
177: al. 2001\nocite{ferraro01}), i.e. it shows two maxima at quadratures
178: ($\phi=0.0$ and $0.5$).
179:
180: We present the optical identification of the companion to PSR\,6266B,
181: based on high-quality phase-resolved photometry obtained with the {\it
182: Hubble Space Telescope} (HST), and X-ray emission detected with
183: \CHANDRA.
184:
185: \section{Observations and data analysis}
186: {\sc HST observations.} The photometric data presented here consist
187: of a set of high-resolution images obtained on August 1 2004 by using
188: the Advanced Camera for Survey (ACS), on-board the HST, and retrieved
189: from the {\it ESO/ST-ECF Science Archive}. The data comprise three
190: $R$-band exposures ($2\times340\,$s and $1\times30\,$s exposures), two
191: $B$-band exposures (of 120 and 340\,s) and four $H_\alpha$-band
192: exposures (of 340, 1050, 1125, 1095\,s). The three longest exposures
193: in $H_\alpha$ are the combination of three sub-exposures. An
194: additional set of Wide Field Planetary Camera 2 (WFPC2, Prop. 10845,
195: PI:Ferraro) data has been secured through the $H_\alpha$ filter with
196: the specific aim of testing the possible variability of the companion:
197: ten\footnote{Only 9 exposures have been used in the analysis since one
198: was made useless by a cosmic ray passage.} 1200\,s exposures were
199: taken between May 2 and May 5 2007. In order to best resolve stars in
200: the most crowded regions, the Planetary Camera (with a resolution of
201: $0.046\arcsec/$pixel) of the WFPC2 was pointed at the cluster center.
202: The photometric analysis of the ACS dataset has been carried out using
203: the ACS module of the Dolphot\footnote{The DOLPHOT packages, including
204: the ACS module and all the documentation, can be found at {\tt
205: http://purcell.as.arizona.edu/dolphot/}.} package \cite{dolphin},
206: using the FLT.fits datasets for the photometry, and the drizzled
207: images as references for the geometric distortion correction. We set
208: the photometry parameters as recommended in the Dolphot manual,
209: obtaining a final catalog of $B$, $R$ and $H_\alpha$ magnitudes,
210: calibrated according to Sirianni et al. (2005). The final catalog is
211: reported to an absolute astrometric system by cross-correlating the
212: stars in common with the data set of Beccari et al. (2006), which has
213: been astrometrized by using the standard stars from the new {\it Guide
214: Star Catalog II} ($GSCII$) lying in the considered field of view
215: (FoV). At the end of the procedure, the rms residuals were of the
216: order of $\sim 0.5\arcsec$ both in RA and Dec; we assume this value as
217: representative of the astrometric accuracy.
218:
219: In order to carefully search for any variability from the MSP
220: companion, we reanalyzed the $H_\alpha$ images in a small area around
221: the nominal position of PSR\,6266B by using ROMAFOT (Buonanno et
222: al. 1983). This package has been specifically developed to perform
223: accurate photometry in crowded fields, and it also allows a visual
224: inspection of the quality of the PSF-fitting procedure. In
225: particular, we extracted $50\times 50$-pixels ($\sim 2.5\arcsec \times
226: 2.5\arcsec$) sub-images from the ten original $H_\alpha$ ACS frames,
227: each centered on the nominal position of PSR\,6266B. In order to
228: optimize the detection of faint objects, we performed the search
229: procedure on the median image. Then, we adopted the resulting mask
230: with the star positions to the individual sub-images, and performed
231: the PSF-fitting procedure separately on each of them. We adopted the
232: same strategy for the analysis of the nine WFPC2 $H_\alpha$ images,
233: after the careful alignement with the previously astrometrized ACS
234: frames. The resulting instrumental magnitudes were transformed to a
235: common system, and we compiled a final catalog with coordinates and
236: magnitudes for all the stars identified in the considered 19
237: sub-images. The photometric calibration of the instrumental magnitudes
238: and the absolute celestial coordinates were determined by using all
239: the stars in common with the overall ACS catalog, calibrated and
240: astrometrized as discussed above.
241:
242: {\sc Chandra Observations.} NGC\,6266 was observed for 63 ks on May
243: 2002 with the \CHANDRA Advanced CCD Imaging Spectrometer (ACIS). The
244: ACIS-S3 chip was pointed at the cluster center, its FoV ($8\arcmin
245: \times 8 \arcmin$) entirely covering the half-mass radius
246: ($1.23\arcmin$; Harris 1996). The data were reduced using the CIAO
247: software (v. 3.3.0), reprocessing the level 1 event files; the task
248: {\tt acis\_run\_hotpix} was used to generate a new badpixel file. Then
249: we applied the corrections for pixels randomization, good time
250: intervals and grades. Our analysis includes only photons with ASCA
251: grades of 0, 2, 3, 4 and 6. About 1.5 ks of observations were not
252: considered in the analysis because of the high background level. We
253: searched for discrete sources in our level 2 event file using {\tt
254: Wavdetect} (Freeman et al. 2002), performing the source detection
255: separetely on the 0.2-1, 1-2, 2-8, and 0.2-8 keV bands. The detection
256: threshold was set to $10^{-5}$, and the source detection was performed
257: increasing the sequence of wavelets from 1 to 16 by factors of
258: $\sqrt{2}$. More than 110 sources were detected in the 0.2-8 keV band
259: in the entire chip. The astrometry of all the detected sources was
260: corrected by using the Aspect Calculator provided by the \CHANDRA
261: X-Ray Center. Then we searched for possible optical counterparts both
262: in our HST images and catalog. About 5 bright stars and 2 faint and
263: extended objects (likely background galaxies) were found to coincide
264: with the X-ray sources well outside the half-mass radius, where the
265: stellar density is relatively low. We therefore used these optical
266: counterparts for extending the astrometric solution of HST to the
267: \CHANDRA sources, thus obtaining the same accuracy in the absolute
268: astrometry for both the optical and the X-ray sources.
269:
270: \section{Results}
271: Figure \ref{fig:chart} shows the $3\arcsec \times 3\arcsec$ WFPC2
272: $H_\alpha$ map centered on the PSR\,6266B nominal position (marked
273: with a {\it cross}), as derived from timing (see P03). An accurate
274: photometric analysis has been carried out for all the $\sim30$ stars
275: found in our catalog within a distance of $1.5\arcsec$ from the
276: PSR\,6266B position. This correspond to 3 times the $1\sigma$
277: uncertainty ($0.5\arcsec$) in the absolute astrometry of the HST data.
278: In particular, we have extracted 19 individual images from our dataset
279: (10 exposures with ACS and 9 with WFPC2) and compared the resulting
280: magnitudes for each object in order to search for variability and
281: estimate the typical photometric uncertainties at different magnitude
282: levels. Only one source in the catalog (hereafter named COM\,6266B,
283: marked with a small circle in Fig.~\ref{fig:chart}) showed $H_\alpha$
284: variability ($\Delta H_\alpha \sim 0.2$ mag) significantly larger than
285: the typical rms magnitude fluctuations of stars of similar luminosity
286: ($\delta H_\alpha \sim 0.02$ mag in ACS data, and $\delta H_\alpha
287: \sim 0.04$ mag in WFPC2 data). Its celestial coordinates are
288: $\alpha_{\rm J2000}=17^{\rm h}01^{\rm m} 12^{\rm s}\!.690$ and
289: $\delta_{\rm J2000}=-30\arcdeg 06\arcmin 48\farcs61$, whereas the mean
290: apparent magnitudes are $B=20.58$, $V=19.48$, $R=19.02$,
291: $H_\alpha=18.65$ ($V$ is from Beccari et
292: al. 2006\nocite{Beccari06}). It is located at $0.5''$ from PSR\,6266B
293: just at the edge of its positional error circle.
294:
295: Figure \ref{fig:CMDs} reports the $(R,\,H\alpha-R)$ and $(R,\,B-R)$
296: Color Magnitude Diagrams (CMDs) for all the stars detected in the
297: $20\arcsec\times20\arcsec$ box centered on the PSR\,6266B nominal
298: position. COM\,6266B is marked with a {\it large filled triangle} in
299: both the CMDs. As apparent from the figure, the star has almost the
300: same luminosity of the cluster Turn Off, but it shows an anomalous red
301: color which locates it out of the Main Sequence. The photometric
302: properties imply that it is not a degenerate star and that it has a
303: moderate $H_\alpha$ excess, indicating the presence of ionized gas
304: surrounding the system. These findings are consistent with the
305: irregulaties seen in the radio signal of PSR\,6266B and supports the
306: scenario that COM\,6266B is a mass losing star.
307:
308: In order to firmly assess the physical connection between COM\,6266B
309: and the binary pulsar PSR\,6266B, we have accurately investigated the
310: time scale of the optical variability\footnote{Note that COM\,6266B
311: has a close companion located at only 4 pixels ($\sim 0\farcs19$)
312: West in the ACS map (see Fig.\ref{fig:chart}). This object turns out
313: to be a normal Sub-Giant Branch star, slightly brighter ($\sim0.5$
314: mag in $B$, $R$ and $H_\alpha$) than COM\,6266B and not displaying
315: any evidence of variability. Although very close to each other, the
316: two stars are clearly separable in the high resolution ACS and WFPC2
317: images and we have verified that the photometric analysis of
318: COM\,6266B is not affected by the presence of this close star.} (in
319: the 0-1 day range). The $H_\alpha$ time series (19 points) was
320: processed using the GRATIS $\chi^2$ fitting routine, a code developed
321: at the Bologna Astronomical Observatory in order to study the
322: periodicity of variable stars (see, e.g., Clementini et
323: al. 2000\nocite{clem+00}). Following the procedure fully described in
324: Ferraro et al. (2001), the most significant (99\% confidence
325: interval)\footnote{The significance level has been obtained from
326: the reduced $\chi^2$ of the Fourier time series fit (at different
327: modulation periods) to the $H_\alpha$ data (see for example Figure
328: 2 of Ferraro et al. 2001).} periodicity was found at a period of
329: $0.1446\pm 0.0015$ day, consistent, within the uncertainties, with the
330: orbital period of PSR\,6266B derived from timing
331: ($P_b=0.1445454303(6)$ day, where the figure in parenthesis is the
332: uncertainty, at 99\% confidence level, on the last quoted digit;
333: P03). Given that, we have folded the 19 magnitude values, by using
334: $P_b$ and the reference epoch of the PSR\,6266B radio ephemeris
335: ($T_0=52047.258199(2)$; P03). The results are shown in
336: Fig.\ref{fig:lc}. As apparent, the ACS and the WFPC2 data (obtained
337: about 3 years later) agree with each other, drawing a well defined
338: light curve with two distinct maxima at about $\phi=0.0$ and
339: $\phi=0.5.$ This fact, as well as the phases and the amplitudes of the
340: minima, confirms that the optical modulation is associated with the
341: pulsar binary motion and strongly suggests that COM\,6266B is a
342: deformed star overflowing its Roche lobe (see \S 1).
343:
344: Additional properties of this system can be derived from the analysis
345: of the \CHANDRA data. We found an X--ray source located at only $\sim
346: 0\farcs4$ from COM\,6266B and at $0\farcs3$ from the nominal radio
347: position of PSR\,6266B. The derived coordinates of the X-ray source
348: are $\alpha_{\rm J2000}=17^{\rm h}01^{\rm m}12^{\rm s}\!.700$ and
349: $\delta_{\rm J2000}=-30\arcdeg 06\arcmin 49\farcs08,$ and the
350: $1\sigma$ uncertainty area on its position is encircled in
351: Fig.\ref{fig:chart} (dashed line). Given the number of sources (50)
352: detected within the half-mass radius of NGC\,6266, the probability of
353: a chance superposition of a X--ray source within $0\farcs3$ from the
354: radio position of PSR\,6266B is $\sim 7\times10^{-4},$ strongly
355: supporting the association of the detected source with the PSR\,6266B
356: system. The (background substracted) photons counts are
357: $18.3_{-4.2}^{+5.3}$ (soft: 0.2--1 keV), $23.7_{-4.9}^{+6.0}$ (medium:
358: 1--2 keV) and $9.8_{-3.1}^{+4.3}$ (hard: 2--8 keV), implying
359: HR1$\simeq 1.3$ and HR2$\simeq 0.4$ (where HR1=medium/soft counts and
360: HR2=hard/medium counts). Hence the source is harder than the typical
361: MSP population observed in GCs (e.g. Grindlay et
362: al. 2002\nocite{grind02}) and the HR1, HR2 values resemble those of a
363: source in which the pulsar wind shocks the material released from the
364: companion (see, e.g., the cases of PSR\,J1740$-$5340 in NGC\,6397, and
365: PSR\,J0024$-$7204W in 47Tuc; Grindlay et al. 2002\nocite{grind02}, and
366: Bogdanov et al. 2005\nocite{bogd05}, respectively). Even if the small
367: number of photons prevents a detailed spectral analysis, for an
368: absorption column density $N_H=2.2\times 10^{21}\,{\rm cm^{-2}}$
369: (Pooley et al. 2003\nocite{Pooley03}), the counts in each band and the
370: HR1 and HR2 values are consistent\footnote{The PIMMS tool has been
371: used for these estimates: {\tt
372: http://heasarc.gsfc.nasa.gov/Tools/w3pimms.html}} with a power law
373: model having a photon index $\simeq 2.5$, and a total unabsorbed flux
374: $F_X\sim 10^{-14}\,{\rm erg\ cm^{-2} s^{-1}}$ in the 0.2-8 keV band,
375: translating (for a distance of 6.6\,kpc; Beccari et
376: al. 2006\nocite{Beccari06}) into a X--ray luminosity of $L_X\sim
377: 6\times 10^{31}\,{\rm erg\ s^{-1}}$.
378:
379: \section{Discussion}
380:
381: Many observed features of COM\,6266B (namely, its anomalous red
382: colour, the $H_\alpha$ excess, the shape of the light curve, the X-ray
383: emission from the binary) are suggestive of a tidally deformed star
384: which is experiencing heavy mass loss, similar to the system found in
385: NGC\,6397 by Ferraro et al. (2001\nocite{ferraro01}). On the other
386: hand, other photometric properties of the source do not fit this
387: picture, as described below.
388:
389: By assuming the reddening $E(B-V)\sim 0.47$ and the distance quoted by
390: Beccari et al. (2006), and adopting $[Fe/H]=-1.1$ and an age of 12 Gyr
391: for NGC\,6266, we have determined the physical parameters of
392: COM\,6266B from the comparison of its position on the CMDs
393: (Fig.\,\ref{fig:CMDs}) with the isochrones of Pietrinferni et
394: al. (2004\nocite{Pietr04}). The resulting effective temperature,
395: bolometric luminosity and radius of this star are $T_{\rm eff}\sim
396: 6000\pm 500\,$K, $L_{\rm bol}\sim 1.9\pm0.2\,L_\odot,$ and $R\sim
397: 1.2\pm 0.2\,R_\odot$, respectively, where the quoted uncertainties are
398: conservative estimates, essentially due to the uncertanties in the
399: distance modulus ($\pm 0.15$ mag) and in the reddening ($\pm 0.05$
400: mag). If the luminosity variations shown in Fig.~\ref{fig:lc} are
401: mainly due to tidal deformations of the companion star, the stellar
402: radius $R$ is expected to be of the order of the Roche lobe radius
403: $R_L$. The latter can be estimated from the mass function and the
404: projected semi-major axis of PSR\,6266B (derived from pulsar timing;
405: P03), provided that a pulsar mass $M_{\rm psr}$ and an orbital
406: inclination $i$ are assumed. As already discussed in P03, the presence
407: of radio eclipses indicates that $i$ is not small.
408: %\footnote{This is also an implicit
409: %consequence of the hypothesis that the shape of the light curve is due
410: %to ellipsoidal variations, the modulation of which tends to vanish for
411: %small values of $i$.}.
412: Conservatively taking $i\gsim20\arcdeg$ and assuming $M_{\rm
413: psr}=1.4\,M_\odot$, the companion star is expected to span a mass
414: interval ranging from $0.125\,M_\odot$ ($i=90\arcdeg$) to
415: $0.41\,M_\odot$ ($i=20\arcdeg$)\footnote{Note that in both cases the
416: orbital separation of the system turns out to be $a\sim
417: 1.4\,R_\odot$.}, corresponding to $R_L\sim 0.26\,R_\odot$ and
418: $0.40\, R_\odot$, respectively. The value of $R_L$ is only slightly
419: affected by the choice of $M_{\rm psr},$ being, for instance,
420: $R_L\lsim 0.45\, R_\odot$ for $0.7\, M_\odot\le M_{\rm psr}\le 3.0\,
421: M_\odot$ (a safely large neutron star mass range; see, e.g., Freire et
422: al. 2008\nocite{freire08}). Note that in order to match the estimated
423: $R_L$ with the observed $R$ it should be $i\sim 3\arcdeg$ (for $M_{\rm
424: psr}=1.4\, M_\odot$), thus implying an unreliable companion mass of
425: $\sim 6\, M_\odot$. In summary, for all plausible binary parameters,
426: it turns out that the observed stellar radius of COM\,6266B is
427: significantly larger ($R\gsim 3\,R_L$) than the expected Roche lobe
428: radius of the companion to PSR\,6266B, i.e. COM\,6266B appears $\gsim
429: 10$ times brighter than the expected luminosity of any pulsar
430: companion having an effective temperature $T_e=6000~K.$
431:
432: Under the hypothesis that COM\,6266B is the companion to PSR\,6266B,
433: what is the origin of this remarkable discrepancy? A number of
434: possibilities are examined below.
435:
436:
437: {\it (i)} The binary system could not belong to the cluster. In this
438: case the adopted distance would be inappropriate and the value of
439: the radius deduced from the luminosity would be largely overestimated.
440: On the other hand, the celestial position of PSR\,6266B is at less
441: than $2\arcsec$ from the cluster center, and the pulsar shows a
442: negative value of the spin period derivative, as expected if
443: PSR\,6266B lies in the cluster potential well.
444:
445: {\it (ii)} The optical luminosity of COM\,6266B could be dominated by
446: strong non-thermal processes, possibly triggered by the pulsar spin
447: down power $L_{\rm sd}\lsim 10 \,L_\odot$ (see P03). Also in this
448: case the value of the stellar radius deduced from the observed
449: luminosity could be significantly overestimated. However, At the
450: orbital separation of $1.4 R_\odot$, the energy captured by a Roche
451: lobe filling companion would be of the order of $L\sim 0.1-0.2
452: L_\odot$ which is too low to explain the observed value. Moreover, it
453: is hard to explain how a large non-thermal component may generate a
454: light curve with maxima that fall at orbital phases nicely consistent
455: with those expected for light curves modulated by ellipsoidal
456: variations.
457:
458: {\it (iii)} COM\,6266B could be the blend of two low-luminosity stars:
459: one star would be the optically variable companion to PSR\,6266B and
460: the other one a non-variable star in the foreground
461: (background). While stellar blends are not uncommon in the highly
462: crowded GC cores, this possibility only partially alleviates the
463: problem. In fact, by assuming that COM\,6266B is the blend of two
464: stars of equal mean $H_\alpha$ magnitude (i.e., $H_\alpha=19.35$), the
465: observed ellipsoidal modulation could only be produced if the pulsar
466: companion varies by $\Delta{H_\alpha}\sim 0.35$ mag, which is a rather
467: extreme value for variabilities induced by tidal
468: distortions. Moreover, even in this case, the star would have a radius
469: $R \gsim 1.7 R_L$. Matching $R$ with $R_L$ would require that
470: COM\,6266B is e.g. the blend of a non-variable star with
471: $H_\alpha=18.7$ and a star with mean magnitude $H_\alpha \sim 21$
472: varying by $\Delta{H\alpha}>3 $ mag. We note that similar
473: considerations would apply also if PSR\,6266B were a triple system.
474:
475: In conclusion, while option {\it (iii)} seems to be a quite unlike
476: {\it ad hoc} scenario, options {\it (i)} and {\it (ii)} will be
477: finally addressed as soon as spectroscopic data are available for the
478: system. This will allow us to investigate the nature (thermal or not
479: thermal) of the optical emission and the association of the source to
480: the cluster (via the determination of the radial velocity), hence
481: shedding some light on the nature of this puzzling object.
482:
483: \acknowledgements This research was supported by the Agenzia Spaziale
484: Italiana (ASI I/088/06/0), {\it PRIN-INAF2006} and by the Ministero
485: dell'Istruzione, dell'Universit\`a e della Ricerca, and it is part of
486: the {\it Progetti Strategici di Ateneo 2006} granted by University of
487: Bologna. AP and NDA also acknowledge financial support from {\it
488: PRIN-MIUR2005}. We thank the Referee for the careful reading of the
489: manuscript.
490:
491: \begin{thebibliography}{}
492:
493: \bibitem[Beccari et al. 2006]{Beccari06}
494: Beccari, G., Ferraro, F.~R., Possenti, A., Valenti, E., Origlia, L.,
495: \& Rood, R.~T., 2006, AJ, 131, 2551
496:
497: \bibitem[Bogdanov et al. 2005]{bogd05}
498: Bogdanov, S., Grindlay, J.~E. and van den Berg, M., 2005, ApJ, 630, 1029
499:
500: \bibitem[Buonanno et al. 1983]{buonanno83}
501: Buonanno, R., Buscema, G., Corsi, C.~E, Ferraro, I., \& Iannicola, G., 1983,
502: A\&A, 126, 278
503:
504: \bibitem[Clementini et al., 2000]{clem+00}
505: Clementini, et al., 2000, AJ, 120, 2054
506:
507: \bibitem[D'Amico et al. 2001]{damico01}
508: D'Amico, N., Lyne, A.~G.,Manchester, R.~N., Possenti, A., \& Camilo, F.,
509: 2001, ApJ, 561, L171
510:
511: \bibitem[Dolphin 2000]{dolphin}
512: Dolphin, A.~E., 2002, PASP, 112, 1383
513:
514: \bibitem[Ferraro et al. 2001b]{ferraro01}
515: Ferraro, F.~R., Possenti, A., D'Amico, N., \& Sabbi, E., 2001, ApJ, 561, L93
516:
517: \bibitem[Freeman et al. 2002]{freeman02}
518: Freeman, P. E., Kashyap, V., Rosner, R., \& Lamb, D. Q., 2002, \apjs, 138, 185
519:
520: \bibitem[Freire et al. 2008]{freire08}
521: Freire, P.C.C., Wolszczan, A., van den Berg, M., \& Hessels, J. W. T.,
522: 2008, ApJ, in press
523:
524: \bibitem[Fruchter et al. 1990]{fbb+90}
525: Fruchter, A.~S. et al., 1990, ApJ, 351, 642
526:
527: \bibitem[Grindlay et al. 2002]{grind02}
528: Grindlay, J. E., Camilo, F., Heinke, C. O., Edmonds, P. D.,
529: Cohn, H. \& Lugger, P., 2002, ApJ, 581, 470
530:
531: \bibitem[Harris 1996]{harris96}
532: Harris, W.~E., 1996, AJ, 112, 1487
533:
534: \bibitem[Jacoby et al. 2002]{jacoby02}
535: Jacoby, B.~A., Chandler, A.~M., Backer, D.~C. Anderson, S.~B.,
536: \& Kulkarni, S.~R., IAUC 7783, 2002
537:
538: \bibitem[Orosz \& van Kerkwijk 2003]{ov03} Orosz, J.~A.,
539: \& van Kerkwijk, M.~H., 2003, A\&A, 397, 237
540:
541: \bibitem[Pietrinferni et al. 2004]{Pietr04} Pietrinferni, A., Cassisi, S.,
542: Salaris, M., \& Castelli, F., 2004, ApJ, 612, 168
543:
544: \bibitem[Pooley at al. 2003]{Pooley03}
545: Pooley, D., et al., 2003, ApJ, 591, L131
546:
547: \bibitem[Possenti et al. 2003]{P03}
548: Possenti, A., D'Amico, N., Manchester, R.~N., Camilo, F.,Lyne, A.~G.,
549: Sarkissian, J., Corongiu, A., 2003, ApJ, 599, 475 (P03)
550:
551: \bibitem[Sirianni et al. 2005]{Sirianni05}
552: Sirianni, et al., 2005, PASP, 117, 1049
553:
554: \bibitem[Stappers et al. 2001]{sbl+01}
555: Stappers, B.~W., van Kerkwijk, M.~H., Bell, J.~F., \& Kulkarni, S.~R.,
556: 2001, ApJ, 548, L183
557:
558: \end{thebibliography}
559:
560: \clearpage
561:
562: \begin{figure}
563: \plotone{f1.ps}
564: \caption{Finding chart for \compulb showing the median-combined WFPC2
565: $H_\alpha$ image. The region covers $3\arcsec \times 3\arcsec$; the
566: cross marks the nominal position of PSR\,6266B, with the $1\sigma$
567: error circle ($0.5\arcsec$, solid line). The smaller circle shows
568: the optical counterpart, whereas the dashed line represents the
569: $1\sigma$ circle ($0.7\arcsec$) in the \CHANDRA position. North is
570: up and East is to the left.
571: \label{fig:chart}}
572: \end{figure}
573:
574: %\clearpage
575:
576:
577: \begin{figure}
578: \plotone{f2.ps}
579: \caption{ $(R,\,H_\alpha-R)$ and $(R,\,B-R)$ CMDs for stars identified
580: in a region of $20\arcsec \times20\arcsec$ centered at the nominal
581: position of PSR\,6266B. The optical counterpart to the pulsar
582: companion COM\,6266B is marked with a {\it large filled triangle}.
583: \label{fig:CMDs}}
584: \end{figure}
585:
586: \clearpage
587:
588: \begin{figure}
589: \plotone{f3.ps}
590: \caption{ $H_\alpha$ light curve of COM\,6266B obtained by using the
591: period and the reference epoch of the radio ephemeris of
592: PSR\,6266B. Asterisks represent ACS archive observations performed
593: in August 2004; large empty triangles are the WFPC2 data collected
594: in May 2007. The phases of quadrature and conjunction of COM\,6266B
595: are reported for clarity. The {\it solid line} represents the
596: Fourier time series (trucated to the second harmonic) that best fits
597: the data.
598: \label{fig:lc}}
599: \end{figure}
600:
601: \end{document}
602: