1: \documentclass[12pt,aasms4]{aastex}
2: %\documentclass[12pt,preprint]{aastex}
3: %\usepackage{emulateapj5}
4: \def\ergsec{\hbox{ergs s$^{-1}$}}
5: \def\ergcmsec{\hbox{ergs cm$^{-2}$ s$^{-1}$}}
6: \def\countsec{\hbox{counts s$^{-1}$}}
7:
8: \slugcomment{Accepted for publication in the {\it Astrophysical Journal Letters}}
9: \shorttitle{LMXRBs in M15}
10: \shortauthors{White \& Angelini}
11:
12: \begin{document}
13:
14: \title{The Discovery of a Second Luminous Low Mass X-ray Binary in
15: the Globular Cluster M15}
16: \author{Nicholas E. White, and Lorella Angelini\altaffilmark{1}}
17:
18: \affil{Laboratory for High Energy Astrophysics, NASA's Goddard Space Flight Center,
19: Greenbelt, MD 20771}
20: \altaffiltext{1}{Also Universities Space Research Association}
21: \email{nwhite@lheapop.gsfc.nasa.gov}
22: \email{angelini@davide.gsfc.nasa.gov}
23:
24: %==============================================================================
25: \begin{abstract}
26:
27: We report an observation by the {\it Chandra} X-ray Observatory of 4U2127+119, the X-ray source identified with the globular cluster M15. The {\it Chandra} observation reveals that 4U2127+119 is in fact two bright sources, separated by $2.7''$. One source is associated with AC211, the previously identified optical counterpart to
28: 4U2127+119, a low mass X-ray binary (LMXB). The second source, M15-X2, is coincident with a $19^{th}$ U magnitude blue star that is $3.3"$ from the cluster core. The {\it Chandra} count rate of M15-X2 is 2.5 times higher than that of AC211. Prior
29: to the $0.5''$ imaging capability of {\it Chandra} the presence of two so closely separated bright sources would
30: not have been resolved. The optical counterpart, X-ray luminosity and
31: spectrum of M15-X2 are consistent with it also being an LMXB system. This is the first time that two
32: LMXBs have been seen to be simultaneously active in a globular cluster.
33: The discovery of a second active LMXB in M15 solves a long standing puzzle where the properties of AC211 appear consistent with it
34: being dominated by an extended accretion disk corona, and yet 4U2127+119 also shows
35: luminous X-ray bursts requiring that the neutron star be directly visible.
36: The resolution of 4U2127+119 into two sources suggests that the X-ray bursts did not come from AC211, but rather from M15-X2. We discuss
37: the implications of this discovery for understanding the origin and evolution of LMXBs in GCs as well as X-ray observations of globular clusters in nearby galaxies.
38:
39: \end{abstract}
40:
41: \keywords{ globular clusters: individual (M15) -- Stars: individual
42: (AC211) -- X-rays:binaries -- X-rays:individual (4U2127+119)}
43:
44: %=============================================================================
45:
46: \section{Introduction}
47:
48:
49: The X-ray source 4U2127+119 associated with the globular cluster (GC) M15
50: was the first to be identified with an individual star within a
51: GC. The identification with the $V\sim 15$ star AC211 was made
52: by Auriere, le Fevre and Terzan
53: (1984) using the Einstein high resolution imager (HRI) position
54: (Hertz and Grindlay 1983; HG83). The
55: identification of AC211 with 4U2127+119 was further strengthened from a spectroscopic study
56: by Charles, Jones and Naylor (1986), which showed characteristic
57: signatures of a LMXB. AC211 was found to have a modulation of 8.5
58: hr in the optical (Ilovaisky {\it et al.} 1987), which was then subsequently
59: seen in the X-ray flux of 4U2127+119 (Hertz 1987; H87). Further observations and a more detailed analysis by Ilovaisky {\it et al.} (1993; I93) revealed the true orbital period of AC211 to be 17.1 hr, twice the originally proposed value.
60:
61: AC211 is optically one of the brightest known LMXBs and yet has
62: a relatively low X-ray luminosity of $\sim 10^{36}$ \ergsec. The high
63: optical to X-ray luminosity ratio suggests that a very luminous
64: central X-ray source is hidden behind the accretion disk, with
65: X-ray emission scattered into our line of sight via an
66: accretion disk corona, ADC (Auriere {\it et al.} 1984, H87,
67: %Auriere, le Fevre and Terzan 1984, Hertz 1987,
68: Naylor {\it et al.} 1988).
69: An ADC is also required to explain
70: the X-ray orbital modulation (H87, I93).
71: This neat picture was put into doubt when a luminous X-ray
72: burst from 4U2127+119 was recorded by the Ginga satellite in 1988 (Dotani
73: {\it et al.} 1990). This burst was long lived ($>$ 150 s) with a precursor
74: event $\sim 6$~s before the longer event. The peak luminosity of the
75: burst was above $10^{38}$ \ergsec, with an expansion of the neutron
76: star photosphere (Dotani {\it et al.} 1990; van Paradijs {\it et al.} 1990),
77: meaning that the neutron star surface had been directly
78: observed. A dip in the continuum flux between the precursor and the
79: main burst seemed to tie this event to the X-ray source in
80: M15. A second X-ray burst from 4U2127+119 in September 2000, with similar properties to the first, has been reported by Smale (2001) using RXTE. The observation of X-ray bursts from 4U2127+119 has been hard to reconcile with the idea that the
81: central source is hidden behind an ADC. In the {\it letter} we report the results of a {\it Chandra} High Energy Transmission Grating (HETG) observation that solves the puzzling behavior of 4U2127+119.
82:
83: %=====================================================================
84:
85: \section{Results}
86:
87: 4U 2127+119 was observed with the {\it Chandra} HETG grating in
88: conjunction with the ACIS-S CCD array on 2000 August 24, for a total
89: exposure of order 21,000 s. The event file was gain corrected, using
90: the calibration released on June 7 2001 (Caldb 2.6), screened for
91: bad pixels and good
92: time intervals. Streaks caused by flaws in the serial readout of
93: the CCDs were also removed. A CCD grade histogram shows
94: a large fraction of grade seven events, indicating
95: considerable pile up. The zero order image from the cleaned event
96: file using all grades is shown in Figure 1a -- two bright sources separated by $2.7''$ can be seen. To determine the best
97: source positions we used data from all grades and determined the peak of the
98: distribution in a box of $3 \times 3$ pixels. The
99: positions of the two
100: sources are M15-X1: R.A. (J2000) = $21^{h} 29^{m} 58{^s}.25$,
101: Dec (J2000) = $+12^{o} 10' 02''.9$
102: and M15-X2: R.A. (J2000) = $21^{h} 29^{m} 58{^s}.06$,
103: Dec (J2000) = $+12^{o} 10' 02''.6$.
104:
105: Accurate positions for AC211 were reported by Kulkarni {\it et al.} (1990)
106: based on radio measurements and by Geffert {\it et al.} (1994)
107: based on meridian circle measurements and the PPM catalogue. The
108: separation between the radio-meridian and radio-PPM positions range
109: between $0.2''-0.27''$. M15-X1 is within $ 0.9''$ of the Kulkarni position of
110: R.A.(J2000)=$21^{h} 29^{m} 58{^s}.31$, Dec(J2000)=$+12^{o} 10' 02''.9$ for AC211.
111: This is within the
112: uncertainty of the {\it Chandra} attitude solution. To obtain the best
113: coordinates for the second source we then corrected the image reference
114: coordinates so as to give the Kulkarni position for AC211. This corresponded
115: to a shift of $0.94 ''$ in RA and $0.03''$ in Dec, and gives a revised
116: position for M15-X2 of R.A. (J2000)=$21^{h} 29^{m} 58{^s}.13$ and Dec
117: (J2000)=$+12^{o} 10' 02''.6$. Since the M15-X2 source shows pile-up, which could effect the centroiding, we adopted a conservative positional uncertainty
118: of $0.5 ''$ diameter, based on the size of 1 pixel in the {\it Chandra} CCD.
119: We designate
120: the new source {\it CXO J212958.1+121002}, although continue
121: to refer to it here by the more concise name M15-X2.
122:
123: To search for an optical counterpart we used the {\it Hubble Space Telescope}
124: (HST) images published in Guhathakurta {\it et al.} (1996). In Figure 1b we over the $0.5''$ diameter circle centered on the M15-X2 position (Figure 1b). Near the center of the error circle is a faint blue
125: star. De Marchi and Paresce (1994) have cataloged the M15 stars seen in earlier HST images and in their list the faint blue star is number 590, a star with an equivalent U magnitude of 18.6. Star 590 is $0.13 ''$ from the {\it Chandra} position centroid for M15-X2. In Figure 1c we indicate on the original color (U+B+V) image from Guhathakurta {\it et al.} (1996) the positions of AC211 and
126: the blue optical counterpart to M15-X2. The new source M15-X2 is $3.4"$ from the center of M15 (the green cross in Figure 1c).
127: A faint blue optical counterpart is the classic signature of a LMXB (e.g. van
128: Paradijs and McClintock 1996) and it seems very likely that this is the
129: counterpart to the X-ray source.
130:
131: The HEG and MEG dispersion directions and the two source positions relative to the {\it Chandra} roll angle are such that the dispersed
132: grating spectra of the two sources overlap (Figure 1a). Due to the
133: 10 degree difference in dispersion angle between the two gratings,
134: the overlap is more severe in the HEG compared to the MEG.
135: To extract the spectra of the two sources, histograms for the MEG and HEG
136: along the cross-dispersion dimension were fitted with Lorenzian models. The Lorenzian width was fixed
137: at the value obtained from a similar histogram of a point source.
138: The width of the HEG histogram is consistent with a single
139: source, while the MEG requires a double Lorenzian fit (see Fig 2).
140: We concentrated on the MEG grating spectrum because it is the best separated.
141: The relative normalization
142: between the two Lorenzian models for the MEG histogram is $\sim 2.5$, with M15-X2 the
143: brighter source. This is consistent with the fact that most of the zero order
144: grade seven events, caused by pile-up, are coincident with M15-X2.
145: MEG spectra and light curves were accumulated by selecting regions across the
146: cross-dispersed direction, where the contamination was a minimum (Fig 2). The percentage of counts from each source
147: included in the extracted spectra are $\sim$ 50\% for AC211 and $\sim$ 20\%
148: for M15-X2.
149:
150: The lightcurves of the two sources exhibit quite different variability, confirming that the cross contamination is minimal (Figure 3).
151: The X-ray flux of AC211 is highly variable across the observation, typical of that
152: reported in the past (see e.g. Ilovaisky {\it et al.} 1987). Using the
153: ephemeris given in I93 the observation began at
154: orbital phase 0.6 and ended at 0.9, just prior to the predicted time
155: of the partial X-ray eclipse. In contrast M15-X2 shows little variability,
156: with just a very slight decline of a few percent across the
157: observation. A power spectrum analysis of each source for a frequency range $4.8\times10^{-5}$ Hz to 0.196 Hz (twice the CCD read out) did not reveal any periodic signal with a peak to mean amplitude $>$10\%, neither from AC211, nor from M15-X2. There was some quasi-periodic activity on a timescale of $\sim 2.7$ hr from AC211 (which can be seen in the lightcurve) and a 700s modulation in both sources caused by a deliberate wobble in the spacecraft pointing to reduce pileup in the CCD.
158:
159: The spectra of the two sources, separated by orders,
160: were extracted using the {\it tgextract} routine included in Ciao 2.1.
161: For each source the first order positive and negative MEG spectra were added together and grouped with a minimum of 20 counts per bin.
162: The MEG spectra of the two sources are shown in Figure 4. The X-ray spectrum
163: of AC211 is harder than that of M15-X2. The new source, M15-X2, can
164: be fit with a single power law with an energy index of $1.72 \pm 0.06$
165: and an absorption of $< 3.4 \times 10^{20}$ H cm$^{-2}$.
166: Fixing the overall absorption at the expected value to M15 equivalent
167: to a hydrogen column density of
168: $6.7 \times 10^{20}$ cm$^{-2}$ still gives an acceptable fit
169: (reduced $\chi^2$ of 1.14), with a power law
170: index of $1.89\pm 0.05$. In contrast, AC211 is not well fit by a single
171: component power law or thermal bremsstrahlung model. A power law
172: model gives a relatively hard photon index of 1.2 with an absorption
173: of $6\times 10^{21}$ cm$^{-2}$, but with a reduced $\chi^2$ of 1.75. A
174: partial covering model, one of several components used by Sidoli, Parmar and Oosterbroek (2000; SPO)
175: to fit the broader band BeppoSAX integrated spectrum of both sources,
176: does provide an acceptable fit. This gives a power law index of 2.1, a
177: covering fraction of 0.87 with an absorption of $2\times 10^{22}$
178: cm$^{-2}$ and an interstellar absorption of $1.6 \times 10^{21}$
179: cm$^{-2}$. Fixing the absorption at the expected interstellar value for M15 gives
180: a power law index of $2.0\pm 0.1$, a covering fraction of $0.92\pm 0.01$ and an intrinsic
181: absorption of $2.05\pm 0.15 \times 10^{22}$ cm$^{-2}$. This is higher than the covering fraction of 0.64 reported by SPO for the combined spectrum of both sources.
182: %The integrated spectrum of both sources fits a partial covering model with
183: %a power law of index $\sim 2$, but with a lower covering fraction.
184: %Since the individual spectra have a similar
185: %slope, but being M15-X2 brighter the net result is having less fraction
186: %of a putative single source system covered.
187: Using the expected point spread function for a single source we
188: corrected for the partial extraction of the two sources. The 0.5--7.0
189: keV flux from AC211 is $7 \times 10^{-11}$ \ergcmsec, corresponding to
190: $9 \times 10^{35}$ \ergsec\ using a
191: distance of 10.3 kpc to M15 (Harris 1996). For the new source, M15-X2
192: the 0.5-7.0 flux is $1.1 \times 10^{-10}$ \ergcmsec, corresponding
193: to $1.4 \times 10^{36}$ \ergsec.
194:
195:
196: The second burst from 4U2127+119 reported by Smale (2001), occurred one month after the {\it Chandra} observation -- indicating the burst source to still be active. The RXTE all sky monitor
197: (ASM) light curve of M15 does not show any transient outburst or unusual behavior around
198: the time of the {\it Chandra} and RXTE observations.
199: The 5yr long ASM light curve is relatively steady, suggesting that the source is long lived and not highly variable. There are short outbursts in the ASM every
200: $\sim 365$ days, but they occur when the source passes close
201: to the sun and are probably caused by sun glint on the detector
202: collimator upsetting the solutions. It is interesting
203: to note that the original $\pm 1''$ position from the Einstein HRI (HG83)
204: lies closer to that of M15-X2, than AC211, and that the search
205: for the optical counterpart by Auriere {\it at al.} (1984) used a circle
206: of $3.3''$, the standard HRI 90\% confidence error circle.
207: The ROSAT HRI archival images from the mid-1990s appear somewhat
208: elongated in the East-West direction --
209: suggesting the M15-X2 source may have been present.
210:
211:
212:
213: %=============================================================================
214: \section{Discussion}
215:
216: This is the first time that two LMXBs have been seen to be
217: simultaneously active in a globular cluster associated with our Galaxy. The
218: separation of $ \sim 2.7''$ is less than the resolution of previous X-ray
219: telescopes. It is only with the superb $0.5''$ quality imaging
220: of the {\it Chandra} X-ray Observatory that 4U2127+119 can be resolved into two sources.
221: The new
222: X-ray source M15-X2 is 2.5 times brighter than AC211 and associated with a faint
223: $19^{th}$ U magnitude blue star, characteristic of a LMXB. It is 3.4 arc from the center of M15 (Figure 1c). The discovery of a second active LMXB in M15
224: provides a simple explanation for the apparently schizophrenic
225: properties of 4U2127+117 -- M15-X2 is
226: the source of the X-ray bursts, not the ADC dominated AC211. The
227: presence of a second source also resolves why the spectrum of
228: 4U2127+119 is unusually
229: complex (SPO). The X-ray continuum spectrum of M15-X2 is a power law with a photon index of $\sim 1.9$ and a luminosity of $\sim 10^{36}$ \ergsec,
230: both typical of LMXBs that show X-ray bursts. RXTE and archival data suggests that the 2nd source may have been present over the entire past 20 yr. This confirms earlier suspicions that two sources could explain the conflicting properties of 4U2127+117 (Grindlay 1992, 1993; Charles, van Zyl \& Clarkson 2001). The ADC model to explain the X-ray and optical properties of AC211 is now self consistent. The X-ray spectrum of AC211 alone is harder and resembles that of the classic ADC source 4U1822-371 (White {\it et al.} 1981; White, Kallman and Angelini
231: 1997). The average X-ray luminosity of AC211 is really one third lower than previously
232: thought. This both increases the amplitude of the orbital modulation
233: and reduces the luminosity -- strengthening the analogy with 4U1822-371 (White and Holt 1982).
234: %A 0.6 s X-ray pulsar with a modulation of a few percent has recently been discovered in %4U1822-371 by Jonker and van der Klis (2001) -- this {\it Chandra} observation of %4U2127+119 would not have been sensitive to such a short pulse period, and the presence of %the second brighter source in M15 would make such a search by RXTE very challenging.
235:
236: The ratio of LMXBs to stellar mass is more
237: than two orders of magnitudes higher for globular clusters than
238: it is for the rest of the Galaxy (Clark 1975). This overabundance of
239: LMXBs led Fabian, Pringle \& Rees (1975) to propose that the LMXBs in
240: globular clusters are formed via tidal capture of neutron stars
241: in close encounters with main-sequence or giant stars, a
242: mechanism that operates efficiently in the high stellar density
243: found in globular clusters. Hut, Murphy and Verbunt (1991; HMV) discuss the
244: probability of finding one or more LMXBs in any particular globular cluster.
245: This depends strongly on how many neutron stars stay in the
246: cluster after they are born and the lifetime of the LMXBs. In general
247: these calculations suggest that more than one LMXB should be found
248: in a globular cluster. The fact that so few LMXBs in globular clusters are observed has required either
249: short lifetimes (HMV) or larger fractions of neutron
250: stars ejected from the globular cluster (Verbunt and Hut 1987). The large number of millisecond radio pulsars found in globular clusters, thought to be the remains of a LMXB
251: systems, also points to many LMXBs having been active in the past
252: (see HMV). {\it Chandra} observations of other GCs have revealed
253: faint sources some of which may be a transient LMXB in quiescence, suggesting that there are more LMXB in a single GC ({\it e.g.} Grindlay {\it at al.} 2001). This discovery of two active LMXBs in M15 moves the observations in the right direction with respect to the theory and
254: number of radio pulsars, especially given the small number of globular cluster systems in our Galaxy with active LMXBs.
255:
256: {\it Chandra} X-ray observations of nearby galaxies have identified many
257: point X-ray sources with globular clusters (Sarazin, Irwin, \& Bregman 2001,
258: Angelini, Loewenstein, \& Mushotzky 1991).
259: Many of these GC sources have a luminosity above the Eddington
260: limit for accretion onto a neutron star. Angelini {\it et al.} (1991) have suggested that some of these high luminosity systems may be due to multiple LMXBs being active in some
261: GC systems. The discovery of two LMXBs active at the
262: same time in M15 adds weight to that argument.
263:
264:
265: Acknowledgements. We recognize the critical contribution of Cynthia
266: Hess, the original principle investigator for this observation. We
267: thank Ian George and Roy Kilgard for help with the grating analysis, Brian Yanny
268: for providing an HST image of M15, Phil Charles for
269: discussions on the optical counterpart, Robin Corbet for advice on the RXTE ASM and Josh Grindlay, Peter Edmonds and Sergio Ilovaisky for comments on the manuscript. This research made used of data extracted from the HEASARC, MAST and CDS.
270:
271:
272:
273: %==============================================================================
274:
275:
276:
277: \begin{thebibliography}{}
278:
279:
280: \bibitem[Angelini et al. 2001]{A01}
281: Angelini, L., Loewenstein, M., \& Mushotzky, R. F., 2001, \apj, 557, L35
282:
283: \bibitem[Auriere 1984]{A84}
284: Auriere, M., le Fevre, O. , \& Terzan, A. 1984, A\&A, 138, 415
285:
286: \bibitem[Charles 1986]{C86}
287: Charles, P.A., Jones,D.C., \& Naylor, T., 1986, Nature , 323, 417
288:
289: \bibitem[Charles 1991]{C91}
290: Charles, P.A., Clarkson, W.I., \& van Zyl, L, 1991, New Astr. submitted
291:
292: \bibitem[Clark 1975]{C75}
293: Clark , G.W., 1975, \apj, 199, L143
294:
295: \bibitem[De Marchi \& Paresce 1994]{DMP94}
296: De Marchi, G., \& Paresce, F., 1994, AJ, 422, 597
297:
298: \bibitem[Dotani 1990]{D90}
299: Dotani,T., et al., 1990, Nature, 347,534
300:
301: \bibitem[Fabian, Pringle, and Rees 1975]{FPR75}
302: Fabian,A. C., Pringle,F.E., \& Rees, M.J. , 1975, \mnras, 172, 15p
303:
304: \bibitem[Geffert 1992]{G94}
305: Geffert, M., Auriere, M., Requieme,Y.,\& Mazurier, J-M., 1994, A\&A, 282, 989
306:
307: \bibitem[Grindlay 1992]{Gr92}
308: Grindlay, J.E. 1992, in "X-ray Binaries and Recycled Pulsars" (Ed. E. van den Heuvel \& S. Rappaport), Kluwer, NATO ASI Series, 377, 365
309:
310: \bibitem[Grindlay 1993]{Gr93}
311: Grindlay, J.E. 1993, ASP Conf Ser., 48, 1993
312:
313: \bibitem[Grindlay 2001]{Gr01}
314: Grindlay, J.E., Heinke, C., Edmonds, P.D. and Murray, S.S. 2001, Science, 292, 2290
315:
316: \bibitem[Guhathakurta 1996]{GU96}
317: Guhathakurta, P., Yanny, B., Schneider, D.P., \& Bahcall, J.N., 1996, AJ, 111,267
318:
319: \bibitem[Harris 1996]{Ha96}
320: Harris, W.E., 1996, AJ, 112,1487
321:
322: \bibitem[Hertz 1987]{He87}
323: Hertz, P., 1987, \apj, 315,119 [H87]
324:
325: \bibitem[Hertz \& Grindlay 1983]{HG83}
326: Hertz, P., \& Grindlay,J.E, 1983, \apj, 275, 105 [HG83]
327:
328:
329: \bibitem[Hut, Murphy \& Verbunt 1992]{HMF91}
330: Hut, P., Murphy,B.W., \& Verbunt, F.,1991, A\&A, 241,137 [HMV]
331:
332: \bibitem[Ilovaisky 1987]{I87}
333: Ilovaisky, S.A., Auriere, M., Kock-Miramond, L., Chevalier, C., \&
334: Cordoni, J.P., 1987, A\&A, 179, L1
335:
336: \bibitem[Ilovaisky 1993]{I93}
337: Ilovaisky, S.A., Auriere, M., Kock-Miramond, L., Chevalier, C.,
338: Cordoni, J.P., \& Crowe, R.A., 1993, A\&A, 270, 139 [I93]
339:
340: %\bibitem[Jonker \& vdklis 2001]{J01}
341: %Jonker, P.G. \& van der Klis, M., 2001, \apjl, 553, L43
342: %
343: \bibitem[Kulkarni 1990]{K90}
344: Kulkarni, S.R., Goss,W.M., Wolszczan, A., \& Middleditch, J.M., 1990,
345: \apj, 363,L5
346:
347: \bibitem[Nylon et al. 1988]{N88}
348: Nylon, T., Charles, P.A., Drew, J.E., \& Hassall, B.J.M., 1988, MNRAS, 233,285
349:
350: \bibitem[Sarazin, Irwin, \& Bregman 2001]{SIB01}
351: Sarazin, C. L., Irwin, J. A., \& Bregman, J.N. 2001, \apj, in press
352:
353: \bibitem[Sidoli, Parmar, Oosterbroek 1997]{S97}
354: Sidoli, L., Parmar, A. N., \& Oosterbroek, T., 2000, A\&A, 360, 520 [SPO]
355:
356: \bibitem[Smale 2001]{S01}
357: Smale, A.P. 2001, \apj, in press
358:
359:
360: \bibitem[Van Paradijs et al. 1990]{VA99}
361: Van Paradijs, J., Dotani, T., Tanaka, Y., \& Tsuru, T., 1990, PASJ, 42, 633
362:
363: \bibitem[Van Paradijs et al. 1996]{VA98}
364: Van Paradijs, J., \& McClintock, J.E., 1996,{\it X-ray binaries},
365: ed. W.H.G Lewin, J. van Paradijs \& E.P.J. van den Heuvel,
366: Cambridge University Press, 126
367:
368:
369: \bibitem[Verbunt, van Paradijs, J. \& Elson 1984]{VPE84}
370: Verbunt, F., \& Hut, P., 1987, {\it The Origin and Evolution of Neutron Stars},
371: Proceedings of the IAU Symposium, Nanjing, People's Republic of China,
372: May 26-30, 1986, eds D. J. Helfand and J.-H. Huang, Dordrecht, D. Reidel
373: Publishing Co. (IAU Symposium, No. 125), 187
374:
375: \bibitem[White et al. 1981]{W81}
376: White, N. E.,Becker, R. H.,Boldt, E. A., Holt, S. S., Serlemitsos, P. J.,
377: \& Swank, J. H., 1981, \apj, 247,994
378:
379: \bibitem[White,\& Holt 1982]{WH}
380: White, N. E., \& Holt, S.S., 1982, \apj, 257, 318
381:
382: \bibitem[White et al 1997]{W97}
383: White, N. E., Kallman, T.R., \& Angelini, L., 1997, {\it X-ray Imaging and
384: Spectroscopy of Cosmic Hot Plasmas}, ed.
385: F.Makino \& K. Mitsuda, Universal Academy Press, 411
386:
387:
388: \end{thebibliography}
389:
390:
391:
392:
393: %\end{document}
394: \small
395:
396: \begin{figure}
397: \figurenum{1a \& 1b}
398: \centerline{
399: \includegraphics[scale=0.6,angle=-90]{fig1a_ximg.eps}\includegraphics
400: [scale=0.6,angle=-90]{fig1b_oimg.eps}}
401: \caption{{\it a) Left :} The ACIS-S/HETG {\it Chandra} zero order image
402: of M15 showing the two sources. The tick lines across
403: the image mark the dispersion directions for the MEG and the HEG, which are
404: separated by 10 degrees. The dispersion direction
405: for the MEG and HEG with respect to M15-X2 is offset by $\sim 5$~pixels
406: in the X direction and $\sim 1$~pixel in the Y direction.
407: The perpendicular line marked with TG\_D (shown only for the MEG)
408: indicates the cross-dispersion direction. The MEG count histogram
409: (shown in Figure 2) is accumulated with respect to that dimension.
410: {\it b) Right :} An enlargement of the
411: U+B+V HST image taken from Guhathakurta et al. (1996) centered on the
412: {\it Chandra} position of M15-X2.
413: The {\it blue star} corresponds to star number 590 in
414: De Marchi \& Paresce (1994).}
415: \end{figure}
416:
417: \begin{figure}
418: \figurenum{1c}
419: \centerline{\includegraphics[scale=1.0,angle=0]{fig1c_colorimage.eps}}
420: \caption{A true color (U+B+V) Hubble Space Telescope image of the central 9"x9" region of M15 reproduced from Guhathakurta et al. (1996). The positions of AC211 and the $\sim 19^{th}$ U magnitude blue star identified with M15-X2 are indicated. The green cross shows the center of M15 (Guhathakurta et al. 1996). Note the orientation is different from Fig 1b.}
421: \end{figure}
422:
423:
424:
425: \begin{figure}
426: \figurenum{2}
427: \centerline{\includegraphics[scale=0.6, angle=0]{fig2_xhisto.eps}}
428: \caption{Histogram of the MEG cross-dispersion direction (TG\_D)
429: with respect to AC211, shown with the best Lorenzian fits to model the
430: two sources. The dash and dash-dotted lines mark the regions selected
431: to extract the spectra for AC211 and M15-X2. These are taken from
432: the tail of the distribution to minimize contamination.
433: The X-axis boundary of the M15-X2 region are shifted relative to the
434: AC211 position to give an overall view from where the spectra where
435: taken with respect to TG\_D.}
436: \end{figure}
437:
438:
439:
440: \begin{figure}
441: \figurenum{3}
442: \centerline{\includegraphics[scale=0.5,angle=-90]{fig3_xcurve.eps}}
443: \caption{The 0.3-10 keV light curves of the two sources using the same extraction
444: regions used for the spectra. These are not corrected for the extraction region used and
445: the count rates must be scaled up by $\sim 2$ and $\sim 5.5$ for AC211 and M15-X2, respectively. The {\it Chandra} observation covers from phase 0.6 to 0.9
446: of the AC211 orbit, using the ephemeris given by I93.}
447: \end{figure}
448:
449:
450: \begin{figure}
451:
452: \figurenum{4}
453: \centerline{\includegraphics[scale=0.5,angle=-90]{fig4_xspec.eps}}
454: \caption{The MEG spectra for M15-X2 (dashed) and AC211 (solid) with
455: the best fit models shown as histograms. The overall normalization is not corrected
456: for the extraction region used. The correction factors for AC211 and M15-X2 are $\sim 2$ and $\sim 5.5$, respectively}
457: \end{figure}
458:
459:
460:
461: \end{document}
462: