1: \documentclass[12pt,preprint]{aastex}
2: %\documentclass[manuscript]{aastex}
3: %\documentclass[preprint2]{aastex}
4: %\documentclass{aastex}
5: %\usepackage{emulateapj5}
6:
7: \def\xr {X--ray}
8: \newcommand{\gcc}{g~cm$^{-3}\ $}
9: \newcommand{\sfun}[2]{$#1(#2)\ $}
10: \newcommand{\rhonot}{$\rho_{\circ}\ $}
11: \newcommand{\msun}{$M_{\odot}\ $}
12: \newcommand{\greq}{$\stackrel{>}{ _{\sim}}$}
13: \newcommand{\lteq}{$\stackrel{<}{ _{\sim}}$}
14: \newcommand{\etal}{{\it et al.}}
15: \newcommand{\lsim}{\raisebox{-0.3ex}{\mbox{$\stackrel{<}{_\sim} \,$}}}
16: \newcommand{\gsim}{\raisebox{-0.3ex}{\mbox{$\stackrel{>}{_\sim} \,$}}}
17: \def\gta{\ifmmode {\mathbin{\lower 3pt\hbox %> or of order
18: {$\,\rlap{\raise 5pt\hbox{$\char'076$}}\mathchar"7218\,$}}}
19: \else {${\mathbin{\lower 3pt\hbox
20: {$\rlap{\raise 5pt\hbox{$\char'076$}}\mathchar"7218\,$}}}
21: $}\fi}
22: \def\lta{\ifmmode {\,\mathbin{\lower 3pt\hbox %< or of order
23: {$\,\rlap{\raise 5pt\hbox{$\char'074$}}\mathchar"7218\,$}}}
24: \else {${\mathbin{\lower 3pt\hbox
25: {$\rlap{\raise 5pt\hbox{$\char'074$}}\mathchar"7218\,$}}}
26: $}\fi}
27:
28: \shorttitle {Broad Relativistic Iron Line from Ser X-1}
29: \shortauthors {Bhattacharyya and Strohmayer}
30:
31: \begin{document}
32:
33: \title {Evidence of a Broad Relativistic Iron Line from the Neutron
34: Star Low-Mass X-ray Binary Serpens X-1}
35:
36: \author {Sudip Bhattacharyya\altaffilmark{1,2}, and Tod
37: E. Strohmayer\altaffilmark{3}}
38:
39: \altaffiltext{1}{CRESST and X-ray Astrophysics Lab, Astrophysics
40: Science Division, NASA's Goddard Space Flight Center, Greenbelt, MD
41: 20771; sudip@milkyway.gsfc.nasa.gov} \altaffiltext{2}{Department of
42: Astronomy, University of Maryland, College Park, MD 20742}
43: \altaffiltext{3}{X-ray Astrophysics Lab, Astrophysics Science
44: Division, NASA's Goddard Space Flight Center, Greenbelt, MD 20771;
45: stroh@clarence.gsfc.nasa.gov}
46:
47: \begin{abstract}
48:
49: We report on an analysis of {\it XMM-Newton} data from the neutron
50: star low mass X-ray binary (LMXB) Serpens X-1 (Ser X-1). Spectral
51: analysis of EPIC PN data indicates that the previously known broad
52: iron K$\alpha$ emission line from this source has a significantly
53: skewed structure with a moderately extended red wing. The asymmetric
54: shape of the line is well described with the {\tt laor} and {\tt
55: diskline} models in XSPEC and strongly supports an inner accretion
56: disk origin of the line. To our knowledge this is the first strong
57: evidence of a relativistic line in a neutron star LMXB. This finding
58: suggests that the broad lines seen in other neutron star LMXBs likely
59: originate from the inner disk as well. Detailed study of such lines
60: opens up a new way to probe neutron star parameters and their strong
61: gravitational fields. The red wing of the iron line from Ser X-1 is
62: not as broad as that observed from some black hole systems. This is
63: not unreasonable for a neutron star system, as the accretion disk has
64: to terminate at or before the hard stellar surface. Finally, the
65: inferred source inclination angle in the approximate range $40^{\rm
66: o}$-$60^{\rm o}$ is consistent with the lack of dips and eclipses from
67: Ser X-1.
68:
69: \end{abstract}
70:
71: \keywords{accretion, accretion disks --- line: profiles --- relativity
72: --- stars: neutron --- X-rays: binaries --- X-rays: individual (Serpens
73: X-1)}
74:
75: \section {Introduction} \label{sec: 1}
76:
77: The bright persistent LMXB Ser X-1 was discovered in 1965 (Bowyer et
78: al. 1965). The detection of type I X-ray bursts from
79: this source established that it harbors a neutron star (Swank et
80: al. 1976; Li et al. 1977), as such bursts originate from the
81: thermonuclear burning of matter on the stellar surfaces (Woosley
82: \& Taam 1976; Lamb \& Lamb 1978; Strohmayer \& Bildsten 2006). The
83: lack of energy dependent dips and eclipses in the X-ray light curve of
84: Ser X-1 suggests that its inclination angle may be less than $60^{\rm
85: o}$ (Frank et al 1987; White \& Swank 1982). The analysis of {\it
86: RXTE} and {\it BeppoSAX} data revealed that the continuum spectra
87: could be well fitted with a combination of absorbed Comptonization and
88: disk blackbody models (Oosterbroek et al. 2001). Moreover, these
89: authors reported the existence of a broad iron emission line near 6
90: keV, which they could adequately fit with a Gaussian model. Broad
91: iron emission lines (near 6 keV) have been observed from many LMXBs
92: (Asai et al. 2000; Bhattacharyya et al. 2006; Miller et al. 2002a),
93: and active galactic nuclei (AGN; Reynolds \& Nowak 2003). These lines
94: could be broadened by either Doppler and relativistic effects (due to
95: Keplerian motion in the inner accretion disk; Fabian et al. 1989), or
96: Compton scattering (in a disk corona; Misra \& Kembhavi 1998). It has
97: also been suggested recently that complex absorption from perhaps
98: several ionized components may partially account for some of the broad
99: line component (e.g., Turner et al. 2005). The {\it ASCA} data from
100: the Seyfert-1 galaxy MCG-6-30-15 gave the first strong clues that such
101: lines originate in the inner accretion disk, as the high
102: signal-to-noise ratio data revealed a skewed and double-horned
103: profile, consistent with Doppler and
104: relativistic effects in the inner accretion disk, where the speed of
105: matter is a substantial fraction of the speed of light (Tanaka et
106: al. 1995). Analysis of more recent data has strongly supported this
107: inner disk origin of the line for MCG-6-30-15 and other AGN (Reynolds
108: \& Nowak 2003; Wilms et al. 2001; Reynolds et al. 2004; Reeves et
109: al. 2007; Miniutti et al. 2007). Such lines, therefore, are extremely
110: useful probes of strong gravity, and black hole properties, including
111: spin (Brenneman \& Reynolds 2006).
112:
113: An inner disk origin has also been suggested for broad lines in some
114: stellar mass black hole binaries, such as Cyg X-1 and GRS 1650--500
115: (Miller et al. 2002a; 2002b; Fabian et al. 1989). But, the origin of
116: the broad iron lines from neutron star LMXBs has not been as well
117: understood. This is largely because the available data have a modest
118: signal-to-noise ratio, which was not precise enough to rule out simple
119: symmetric profiles (such as a Gaussian profile). If the inner
120: accretion disk origin can be established for broad iron lines from
121: neutron star LMXBs, then the shape and width of the line may be used
122: to constrain the radius of the inner edge of the disk, as well as the
123: Keplerian speed at that radius. The former can give an upper limit to
124: the neutron star radius (as the disk inner edge radius must be greater
125: than or equal to the stellar radius), while both quantities may be
126: utilized to constrain the stellar mass. Moreover, high frequency
127: quasi-periodic oscillations (kHz QPOs) are observed from many neutron
128: star LMXBs (van der Klis 2006). Although the actual physical
129: mechanism responsible for these timing features is still debated, the
130: leading models involve strong gravity, with the accretion disk close
131: to the neutron star surface. Therefore, if the inner disk origin can
132: be established, the broad iron lines can be useful to constrain kHz
133: QPO models, and together these two spectral and timing features will
134: be extremely important as probes of strong gravity, and to constrain
135: neutron star parameters. We note that the constraints on neutron star
136: mass and radius can be useful to determine the equation of state (EoS)
137: of high density cold matter at the stellar core, which is a fundamental
138: problem of physics (see Bhattacharyya et al. 2005). In this Letter,
139: we present evidence for an inner accretion disk origin of the broad
140: iron line from the neutron star LMXB Ser X-1. In \S~2, we
141: describe our analysis of {\it XMM-Newton} data, and in \S~3 we discuss
142: the implications of our results.
143:
144: \section {Spectral Analysis} \label{sec: 2}
145:
146: The neutron star LMXB Ser X-1 was observed three times with {\it
147: XMM-Newton} in March, 2004 (obsIds: 0084020401, 0084020501,
148: 0084020601). These observations were each separated in time by two
149: days, and each consists of about 22 ks of data. Due to the brightness
150: of the source, there are no European Photon Imaging Camera (EPIC) MOS
151: detector data available for these observations. EPIC PN data were
152: obtained in timing mode, and the source photons were also registered
153: by the Reflection Grating Spectrometer (RGS) instruments. However, as
154: no significant source spectral lines were detected in the RGS energy
155: range, here we report only the analysis of the EPIC PN spectra.
156: We have excluded the portions of the EPIC PN data with high and
157: variable background (due to soft proton flares), and extracted source
158: spectra, background spectra, and response matrices using the {\it
159: XMM-Newton} Science Analysis Software (SAS; version 3.0). The SAS
160: task `epatplot' has indicated that this set of spectra (set 1) is
161: modestly piled up. We have, therefore, extracted another set of
162: spectra (set 2) after excluding the contributions from the central
163: (and hence the brightest) pixels in order to minimize the pile-up
164: effect. Results from the two spectral sets are consistent with each
165: other, and the broad double-peaked line, which is the focus of this
166: Letter, is present in both sets in the same energy range. This is
167: expected, as pile-up cannot generate a double-peaked line. Based on
168: these findings we do not believe pile-up is a significant concern for
169: our study. Therefore, we have used spectral set 1 because of its
170: higher signal-to-noise ratio.
171:
172: We have rebinned the spectra (see, for example, Ibarra et al. 2007),
173: fitted them with various models (within XSPEC), and found that the
174: best model to describe the observed continuum spectra consists of an
175: absorbed Comptonization ({\tt compTT}) component plus a disk blackbody
176: ({\tt diskbb}). The addition of {\tt diskbb} to the {\tt compTT}
177: component is essential, as can be seen from Table 1. This is in
178: accordance with observations of Ser X-1 obtained with other X-ray
179: missions (see Oosterbroek et al. 2001). Oosterbroek et al. noted that
180: the choice between a disk blackbody and a single temperature blackbody
181: was arbitrary, however, our spectral fitting shows that a disk
182: blackbody is preferred to a simple blackbody. For example, the choice
183: of the former reduces $\chi^2$ by about 47 compared to that for the
184: latter (obs. 1, i.e., obsId 0084020401). Such a decrease in $\chi^2$
185: is significant, and a similar decrease is found for obs. 2 (obsId
186: 0084020501) and obs. 3 (obsId 0084020601). Table 1 shows that the
187: addition of a Gaussian emission line near 0.54 keV improves the fit
188: very significantly. As we have not found any source spectral lines at
189: this energy in the RGS data, we suspect that this feature may be of
190: instrumental origin, so we have fixed the line centroid of this
191: feature at $\approx 0.54$ keV in our subsequent model fitting.
192: Finally, we have found highly significant excess emission near 6
193: keV. As a broad iron emission line is detected near this energy from
194: many LMXBs (see \S~1), we have fitted this excess emission with a
195: Gaussian. This extra component is very significant (see Table 1),
196: consistent with the results reported in Oosterbroek et
197: al. (2001). However, as we have mentioned in \S~1, it is not yet known
198: with any certainty what broadens this line for neutron star LMXBs. To
199: explore this question in more detail we have replaced the 6 keV
200: Gaussian component with the {\tt diskline} model component of XSPEC.
201: This component represents the spectral line emission from the inner
202: portion of an accretion disk in the Schwarzschild spacetime (Fabian et
203: al. 1989). The corresponding fits show that the {\tt diskline} model
204: describes the line profile significantly better than the simple
205: Gaussian profile (see Table 1). This result supports the notion that
206: the line is produced in the inner accretion disk. In order to further
207: substantiate this result, we have also modeled the spectrum using the
208: {\tt laor} model. This model is similar to the {\tt diskline} model,
209: but it includes the effects of the spin of the central star (Laor
210: 1991). The fact that the {\tt laor} model also fits the line profile
211: significantly better than the simple Gaussian (see Table 1) provides
212: strong evidence of the inner disk origin of this line. In fact, {\tt
213: laor} describes the line profile slightly better than {\tt diskline}
214: (primarily for obs. 1; see table 1). However, while there is some
215: indication of a smaller disk inner radius in the {\tt laor} fit using
216: obs. 1, a value of $6r_{\rm g}$ is not yet excluded with sufficient
217: significance to argue that the spin of the neutron star is affecting
218: the line profile. In principle, deeper observations could test for
219: such an effect.
220:
221: %This suggests that
222: %the spin of the central star may have some effect on the line profile,
223: %by decreasing the disk inner edge radius (Laor 1991). Indeed, while
224: %the best-fit parameter values (including the continuum parameters) for
225: %the {\tt diskline} component and the {\tt laor} component are
226: %generally very consistent with each other, the best-fit inner disk
227: %radius ($R_{\rm in}$) for the latter model is smaller than that for
228: %the former (again, primarily for obs. 1).
229: %Here, we note that $R_{\rm in} \ge 6r_{\rm g}$ for {\tt
230: %diskline} ($r_{\rm g} = GM/c^2$; $M=$ central star mass).
231:
232: In Table 2, we give the best-fit parameter values for the XSPEC model
233: {\tt wabs*(compTT+ diskbb+gauss+laor)} for each observation. These
234: values are consistent across the three observations. Here we note
235: that we have used the disk geometry for the Comptonizing plasma ({\tt
236: compTT} model). However, even the spherical
237: geometry gives very similar parameter values, except for the {\tt
238: compTT} optical depth $\tau_{\rm C}$ (the latter geometry gives
239: $\tau_{\rm C} \approx 15$). We have constrained the rest frame line
240: energy ($E_{\rm L}$; {\tt laor} component) in the range $6.4-6.97$ keV
241: while fitting. This is because, K$\alpha$ spectral lines from neutral
242: or ionized iron are expected in this range. However, although we have
243: always found $6.4$ keV as the best-fit value of $E_{\rm L}$, relaxing
244: this constraint does not lower the best-fit $E_{\rm L}$ much. We have
245: found the best-fit source inclination angle, $i_{\rm L}$, in the range
246: $\approx 40^{\rm o}-60^{\rm o}$, which is consistent with the expected
247: value ($< 60^{\rm o}$) for Ser X-1 (due to the lack of eclipses and
248: dips; see \S~1). Here we note that, since there are many model
249: parameters, we have fixed the line parameters (except the
250: normalizations) to their best-fit values to estimate the errors on the
251: parameters of the {\tt wabs}, {\tt compTT}, and {\tt diskbb}
252: components. Similarly, we have frozen the parameters (except the
253: normalizations) of the {\tt compTT} and {\tt diskbb} models in order
254: to estimate the errors on the spectral line parameters. Therefore, the
255: quoted error values are likely somewhat underestimated.
256: Fig. 1 shows the data, model components, and data-to-model ratio for
257: observation 2. The {\tt laor} model component for the broad iron line
258: is clearly shown by the double-peaked dotted line. The moderately
259: structured shape of the data-to-model ratio results in a relatively
260: high overall reduced $\chi^2$ for Ser X-1 (Table 2). Note that similar
261: residuals, and a high reduced $\chi^2$ were also reported by
262: Oosterbroek et al. (2001; see their Figs. 5 \& 6). These authors
263: argued that these narrow structures could not be caused by an
264: incorrect continuum modeling, since such incorrectness would probably
265: give more smoothly varying data-to-model ratios. Fig. 2 exhibits the
266: structure of the broad relativistic iron line. The two panels show
267: this line from two observations. The data points of the figure clearly
268: show the line with an extended red wing. This figure also explicitly
269: shows that both the {\tt laor} model and the {\tt diskline} model
270: (dotted line in each panel) fit the spectral line profile well.
271:
272: \section {Discussion and Conclusions} \label{sec: 3}
273:
274: In this Letter, we report the results of the spectral fitting of the {\it
275: XMM-Newton} EPIC PN data from the neutron star LMXB Ser X-1. The
276: best-fit continuum parameter values (see Table 2) are generally
277: consistent with those of Oosterbroek et al. (2001). Our slightly
278: higher Comptonization plasma temperature ($T_{\rm C}$) compared to
279: that of Oosterbroek et al. can be explained in terms of the lower
280: $2-10$ keV flux that we have found. This is because, as the intensity
281: of an LMXB decreases, its energy spectrum generally becomes harder.
282:
283: It was previously known that Ser X-1 exhibits a broad iron line (see
284: \S~1), but earlier data did not have the statistical quality to rule
285: out simple symmetric profiles (such as a Gaussian). As a result, its
286: origin was not strongly constrained. Knowledge of the line's origin is
287: very important, because if it is produced in the inner accretion disk,
288: it can be used (1) to probe the strong gravitational field of the
289: neutron star, (2) to constrain the stellar parameters, and (3) to
290: constrain models of the kHz QPOs, which may, in turn, be useful to
291: achieve the first two goals (see \S~1). By analyzing {\it XMM-Newton}
292: EPIC PN data we have demonstrated that the line is significantly
293: asymmetric, and can be accurately modeled with physical models of line
294: formation in the inner disk (using the {\tt diskline} and {\tt laor}
295: models in XSPEC; see Fig. 2). This strongly supports the idea that
296: this spectral line originates from the inner accretion disk of Ser X-1
297: (see \S~1). The high signal-to-noise ratio data also revealed the
298: detailed shape of the broad and relativistically skewed line with an
299: extended red wing (Fig. 2). Such lines have so far been observed from
300: AGNs and a few Galactic black hole X-ray binaries (Tanaka et al. 1995;
301: Miniutti et al. 2007; Miller et al. 2002a; 2002b), but this is the
302: first strong evidence of a relativistic line in a neutron star
303: system. As we have mentioned in \S~1, other neutron star LMXBs
304: exhibit broad iron emission lines, which so far have been adequately
305: modeled with simple Gaussian profiles. Our finding for Ser X-1
306: suggests that the lines seen in other LMXBs may also originate from
307: the inner accretion disk, and its relativistically skewed nature could
308: be confirmed with sufficiently long observations. If future
309: observations bear this out it will open up an exciting new opportunity
310: to probe neutron stars and strong gravity with deep spectroscopic
311: observations.
312:
313: %The best-fit line energies ($E_{\rm L}$) shown in Table 2 should be
314: %the result of the gravitational redshift of one of the K$\alpha$
315: %spectral lines (in the rest energy range $6.40-6.97$ keV) from neutral
316: %or ionized iron. The required redshift values are consistent with the
317: %radial range defined by the parameters $R_{\rm in}$ and $R_{\rm out}$
318: %of the {\tt laor} model component (see Table 2).
319:
320: We have found that the equivalent width ($EW_{\rm L}$) of the iron
321: line from Ser X-1 is lower than that reported in Oosterbroek et
322: al. (2001), which may suggest that the strength of this line decreases
323: with the source intensity. The source inclination angle inferred from
324: both the {\tt diskline} and {\tt laor} fits is consistent with $\lsim
325: 60^{\rm o}$, which is expected for the non-dipping nature of Ser X-1
326: (see \S~1). The best-fit values of $R_{\rm in}$ for obs. 1 \& 3 (see
327: Table 2) suggest that the accretion disk extends almost to the neutron
328: star surface. Although this is not the case for obs. 2, it is to be
329: noted that both the source count rate and the iron line strength are
330: relatively low for this observation (Table 2), and hence the spectral
331: line statistics are not as good as for the other two obsIDs. However,
332: we cannot make a definite conclusion regarding the disk extension with
333: the current data. But we note that the accretion disk can plausibly
334: approach the neutron star surface because the stellar magnetic field
335: in LMXBs is relatively low ($\approx 10^{8 - 9}$ G), and the implied
336: accretion rate for Ser X-1 is high ($\approx 0.28 -
337: 0.32\dot M_{\rm Edd}$; based on the observed flux, and the Eddington
338: luminosity $\approx 2.0-3.8\times10^{38}$ ergs s$^{-1}$, for a
339: distance of $\approx 9.5-12.7$ kpc; Jonker \& Nelemans 2004). An
340: important difference between neutron star and black hole systems in
341: the context of disk line formation is the presence of the neutron star
342: surface. This sets a firm limit on the inner edge of the disk in a
343: neutron star system, even a rapidly spinning one, although the exact
344: limit will depend on the EoS of neutron star matter. For black holes,
345: significant spin always implies a smaller disk inner edge radius (for
346: a corotating disk). Therefore, for very fast spinning black holes, the
347: red wing of the relativistic iron line is expected to be very
348: broad. So the relatively modest extended red wing (compared to some
349: black hole systems; Miniutti et al. 2007) of the iron line from
350: Ser X-1 appears consistent with the fact that this source harbors a
351: neutron star. This relativistic iron line is expected to be
352: accompanied by a continuum reflection spectral component (Reynolds \&
353: Nowak 2003). A distinct feature of this component is a broad hump near
354: 30 keV. As the signal-to-noise ratio of the EPIC PN data is small at
355: higher energies ($\gsim 10$ keV), we could not determine if this
356: component is present in the Ser X-1 spectra.
357:
358: \acknowledgments
359:
360: The authors thank Tim Kallman and Jean Cottam for useful discussions,
361: and John Miller and an anonymous referee for helpful comments.
362:
363: %\clearpage
364: \begin{thebibliography}{}
365:
366: \bibitem[]{342} Asai, K., Dotani, T., Nagase, F., \& Mitsuda, K. 2000,
367: ApJSS, 131, 571
368:
369: %\bibitem[]{345} Bhattacharyya, S., Strohmayer, T. E., Markwardt, C. B., \&
370: %Swank, J. H. 2006a,
371: %\apj, 639, L31
372:
373: \bibitem[]{349} Bhattacharyya, S., Strohmayer, T. E., Swank, J. H., \&
374: Markwardt, C. B. 2006, \apj, 652, 603
375:
376: \bibitem[]{352} Bhattacharyya, S., Strohmayer, T. E., Miller, M. C., \&
377: Markwardt, C. B. 2005, \apj, 619, 483
378:
379: %\bibitem[]{355} Bhattacharyya, S., Thampan, A. V., Misra, R., \& Datta,
380: %B. 2000, \apj, 542, 473
381:
382: \bibitem[]{358} Bowyer, S., Byram, E. T., Chubb, T. A., \& Friedman,
383: H. 1965, Science, 147, 394
384:
385: \bibitem[]{361} Brenneman, L. W., \& Reynolds, C. S. 2006, \apj, 652, 1028
386:
387: \bibitem[]{363} Fabian, A. C., Rees, M. J., Stella, L., \& White,
388: N. E. 1989, \mnras, 238, 729
389:
390: \bibitem[]{366} Frank, J., King, A. R., \& Lasota, J.-P. 1987, A\&A, 178, 137
391:
392: \bibitem[]{413} Ibarra, A. et al. 2007, A\&A, 465, 501
393:
394: \bibitem[]{368} Jonker, P. G., \& Nelemans, G. 2004, \mnras, 354, 355
395:
396: \bibitem[]{370} Lamb, D. Q., \& Lamb, F. K. 1978, \apj, 220, 291
397:
398: \bibitem[]{372} Laor, A. 1991, \apj, 376, 90
399:
400: \bibitem[]{374} Li, F. K. et al. 1977, \mnras, 179, 21
401:
402: \bibitem[]{376} Miller, J. M. et al. 2002a, \apj, 578, 348
403:
404: \bibitem[]{378} Miller, J. M. et al. 2002b, \apj, 570, L69
405:
406: \bibitem[]{380} Miniutti, G. et al. 2007, PASJ, 59, S315
407:
408: \bibitem[]{382} Misra, R., \& Kembhavi, A. K. 1998, \apj, 499, 205
409:
410: \bibitem[]{384} Oosterbroek, T., Barret, D., Guainazzi, M., \& Ford,
411: E. C. 2001, A\&A, 366, 138
412:
413: \bibitem[]{387} Reeves, J. N. et al. 2007, PASJ, 59, 301
414:
415: \bibitem[]{389} Reynolds, C. S., \& Nowak, M. A. 2003, Physics Reports, 377, 389
416:
417: \bibitem[]{391} Reynolds, C. S., Wilms, J., Begelman, M. C., Staubert,
418: R., \& Kendziorra, E. 2004, \mnras, 349, 1153
419:
420: \bibitem[]{394} Strohmayer, T. E., \& Bildsten, L. 2006, in {\it Compact
421: Stellar X-ray Sources}, Eds. W.H.G. Lewin and M. van der Klis,
422: (Cambridge University Press: Cambridge), 113
423:
424: \bibitem[]{398} Swank, J. H., Becker, R. H., Pravdo, S. H., \&
425: Serlemitsos, P. J. 1976, IAU Circ., 2963
426:
427: \bibitem[]{401} Tanaka, Y. et al. 1995, Nature, 375, 659
428:
429: \bibitem[]{403} Turner, T. J., Kraemer, S. B., George, I. M., Reeves,
430: J. N., \& Bottorff, M. C. 2005, \apj, 618, 155
431:
432: \bibitem[]{406} van der Klis, M. 2006, in {\it Compact
433: Stellar X-ray Sources}, Eds. W.H.G. Lewin and M. van der Klis,
434: (Cambridge University Press: Cambridge), 39
435:
436: \bibitem[]{410} White, N. E., \& Swank, J. H. 1982, ApJ, 253, L61
437:
438: \bibitem[]{412} Wilms, J. et al. 2001, \mnras, 328, L27
439:
440: \bibitem[]{414} Woosley, S. E., \& Taam, R. E. 1976, \nat, 263, 101
441:
442: \end{thebibliography}{}
443:
444: \clearpage
445:
446: \begin{deluxetable}{cllll}
447: \tablecolumns{5}
448: \tablewidth{0pc}
449: \tablecaption{Fitting of the {\it XMM-Newton} EPIC PN energy spectra from Ser X-1
450: with various XSPEC models.}
451: \tablehead{No. & XSPEC Model & $\chi^2/{\rm dof}$\tablenotemark{a} & $\chi^2/{\rm
452: dof}$\tablenotemark{a} & $\chi^2/{\rm dof}$\tablenotemark{a}\\
453: & & (Obs. 1\tablenotemark{b}) & (Obs. 2\tablenotemark{b}) & (Obs. 3\tablenotemark{b})}
454: \startdata
455: 1 & {\tt wabs*compTT} & $\frac{2689.1}{706}$ & $\frac{2517.2}{706}$ & $\frac{2464.3}{706}$ \\ \\
456: 2 & {\tt wabs*(compTT+diskbb)} & $\frac{1708.9}{704}$ (0) & $\frac{1198.0}{704}$ (0) &
457: $\frac{1432.5}{704}$ (0) \\ \\
458: 3 & {\tt wabs*(compTT+diskbb+gauss)} & $\frac{1517.6}{702}$ (8.0E-19) &
459: $\frac{1103.8}{702}$ (3.3E-13) & $\frac{1338.2}{702}$ (4.2E-11) \\ \\
460: 4 & {\tt wabs*(compTT+diskbb+gauss+gauss)} & $\frac{1192.7}{699}$ (2.7E-36) &
461: $\frac{949.1}{699}$ (9.7E-23) & $\frac{1098.0}{699}$ (8.4E-30) \\ \\
462: 5 & {\tt wabs*(compTT+diskbb+gauss+diskline)} & $\frac{1165.6}{696}$ (1.1E-3) &
463: $\frac{917.4}{696}$ (3.0E-5) & $\frac{1068.9}{696}$ (3.1E-4)
464: \\ \\
465: 6 & {\tt wabs*(compTT+diskbb+gauss+laor)} & $\frac{1153.6}{696}$ (3.6E-5) &
466: $\frac{917.8}{696}$ (3.4E-5) & $\frac{1060.5}{696}$ (2.3E-5)
467: \enddata
468: \tablenotetext{a}{The number in the parentheses is the probability (calculated from F-test using
469: XSPEC) of the decrease of $\chi^2/{\rm dof}$ by chance from the value of the previous row to
470: that of the current row (for models $2-5$). For model 6, the comparison is with model 4.
471: A very small value is replaced with a zero.}
472: \tablenotetext{b}{Obs. 1: ObsId 0084020401; Obs. 2: ObsId 0084020501; Obs. 3: ObsId 0084020601.}
473: \end{deluxetable}
474:
475: \clearpage
476: \begin{deluxetable}{lccc}
477: \tablecolumns{4}
478: \tablewidth{0pc}
479: \tablecaption{Best fit parameters (with 90\% confidence) for the {\it XMM-Newton} EPIC PN
480: energy spectra from Ser X-1.}
481: \tablehead{Parameter\tablenotemark{a} & Obs. 1\tablenotemark{b} & Obs. 2\tablenotemark{b} & Obs.
482: 3\tablenotemark{b}}
483: \startdata
484: $N_{\rm H}$\tablenotemark{c} ($10^{22}$ atoms cm$^{-2}$) & $0.48_{-0.02}^{+0.02}$ &
485: $0.44_{-0.01}^{+0.01}$ & $0.49_{-0.02}^{+0.02}$ \\
486: $T_{\rm 0}$\tablenotemark{d} (keV) & $0.17_{-0.01}^{+0.01}$ & $0.19_{-0.01}^{+0.01}$ &
487: $0.17_{-0.01}^{+0.02}$ \\
488: $T_{\rm C}$\tablenotemark{e} (keV) & $3.18_{-0.36}^{+0.54}$ & $2.63_{-0.11}^{+0.14}$ &
489: $2.81_{-0.18}^{+0.35}$ \\
490: $\tau_{\rm C}$\tablenotemark{f} & $6.19_{-0.74}^{+0.68}$ & $7.25_{-0.32}^{+0.30}$ &
491: $6.71_{-0.52}^{+0.48}$ \\
492: $T_{\rm in}$\tablenotemark{g} (keV) & $1.71_{-0.12}^{+0.10}$ & $1.21_{-0.04}^{+0.04}$ &
493: $1.40_{-0.08}^{+0.08}$ \\
494: %$EW_{\rm G}$\tablenotemark{h} (eV) & $46.3_{-7.7}^{+6.9}$ & $29.6_{-6.6}^{+3.5}$ & $45.6_{-9.1}^{+11.4}$ \\
495: $E_{\rm L}$\tablenotemark{h} (keV) & $6.40_{-0.0}^{+0.08}$ & $6.40_{-0.0}^{+0.03}$ &
496: $6.40_{-0.0}^{+0.04}$ \\
497: $\beta_{\rm L}$\tablenotemark{i} & $2.17_{-0.13}^{+0.30}$ & $2.38_{-0.56}^{+0.31}$ &
498: $2.55_{-0.30}^{+0.39}$ \\
499: $R_{\rm in}$\tablenotemark{j} & $4.04_{-0.68}^{+2.14}$ & $16.19_{-3.19}^{+8.77}$ &
500: $5.59_{-1.19}^{+1.48}$ \\
501: $R_{\rm out}$\tablenotemark{k} & $271.5_{-93.7}^{+115.2}$ & $400.0_{-70.9}^{+0.0}$ &
502: $313.5_{-166.9}^{+86.5}$ \\
503: $i_{\rm L}$\tablenotemark{l} (degree) & $44.7_{-3.2}^{+1.7}$ & $50.2_{-5.4}^{+8.8}$ &
504: $39.7_{-1.5}^{+1.4}$ \\
505: $EW_{\rm L}$\tablenotemark{m} (eV) & $105.1_{-8.0}^{+6.8}$ & $85.9_{-10.8}^{+8.9}$ &
506: $95.2_{-8.9}^{+11.6}$ \\ \hline
507: ${\chi^2}/{{\rm dof}}$ & ${1153.6}/{696}$ & ${917.8}/{696}$ & ${1060.5}/{696}$ \\
508: Flux\tablenotemark{n} ($0.5-2$ keV) & $0.93$ & $0.87$ & $0.91$ \\
509: Flux\tablenotemark{n} ($2-10$ keV) & $4.22$ & $3.34$ & $3.79$ \\
510: \enddata
511: \tablenotetext{a}{Parameters of the XSPEC model {\tt wabs*(compTT+diskbb+gauss+laor)}.}
512: \tablenotetext{b}{Obs. 1: ObsId 0084020401; Obs. 2: ObsId 0084020501; Obs. 3: ObsId 0084020601.}
513: \tablenotetext{c}{Hydrogen column density from the {\tt wabs} model component.}
514: \tablenotetext{d}{Input soft photon (Wien) temperature of the {\tt compTT} model component.}
515: \tablenotetext{e}{Temperature of the Comptonizing plasma ({\tt compTT} model
516: component; disk geometry).}
517: \tablenotetext{f}{Optical depth of the Comptonizing plasma ({\tt compTT} model component).}
518: \tablenotetext{g}{Temperature at inner disk radius from {\tt diskbb} model component.}
519: %\tablenotetext{h}{Equivalent width of the low energy emission line ({\tt gauss} model component). The central energy of the line is frozen at 0.538 keV (23.05 \AA).}
520: \tablenotetext{h}{Rest frame energy of the broad relativistic iron emission line ({\tt laor} model component).}
521: \tablenotetext{i}{Power law index of emissivity ({\tt laor} model component).}
522: \tablenotetext{j}{Inner radius (in the unit of $GM/c^2$; $M$ is the mass of the neutron star)
523: from the {\tt laor} model component.}
524: \tablenotetext{k}{Outer radius (in the unit of $GM/c^2$) from the {\tt laor} model component.}
525: \tablenotetext{l}{Source inclination angle from the {\tt laor} model component.}
526: \tablenotetext{m}{Equivalent width of the broad relativistic iron emission line ({\tt laor} model component).}
527: \tablenotetext{n}{Observed flux in $10^{-9}$ ergs cm$^{-2}$ s$^{-1}$.}
528: \end{deluxetable}
529:
530: \clearpage
531:
532: \begin{figure}
533: \hspace{-1.9 cm}
534: \epsscale{0.7}
535: \plotone{f1.ps}
536: \vspace{-0 cm}
537: \caption {{\it XMM-Newton} EPIC PN energy spectrum from Ser X-1 for
538: obs. 2 (obsId 0084020501). The upper panel shows the data (error
539: bars), model (solid line), and individual additive model components
540: (dotted lines). Here we have used the best-fit XSPEC model {\tt
541: wabs*(compTT+diskbb+gauss+laor)} (see Table 2). Among the two
542: continuum additive model components, the upper dotted line shows {\tt
543: compTT}, and the lower dotted line shows {\tt diskbb}. The low energy
544: emission line is the {\tt gauss} component, while the high energy
545: broad iron emission line is the {\tt laor} component. The lower panel
546: shows the data to model ratio.}
547: \end{figure}
548:
549: \clearpage
550: \begin{figure}
551: \hspace{-0.3 cm}
552: \epsscale{0.4}
553: \plotone{f2.ps}
554: \vspace{-0 cm}
555: \caption {{\it Upper panel}: {\it XMM-Newton} EPIC PN energy spectrum
556: from Ser X-1 for obs. 1 (obsId 0084020401). With the error bars we
557: show the data intensity in excess of the model intensity, and
558: normalized by the model intensity (i.e., (Data$-$Model)/Model). Here
559: the model is the best-fit model (of Table 2) minus the {\tt laor}
560: component. The {\tt laor} component is separately shown with the
561: dotted line. {\it Lower panel}: Similar to the upper panel, but for
562: obs. 3 (obsId 0084020601), and for the {\tt diskline} component (model
563: 5 of Table 1) instead of the {\tt laor} component. For each panel, the
564: data points clearly show a broad relativistic iron emission line with
565: an extended red wing. }
566: \end{figure}
567:
568: \end{document}
569: