astro-ph0604046/ms.tex
1: \documentclass{emulateapj}
2: \usepackage{amstext,amsmath,natbib}
3: 
4: \shorttitle{Ionized Lines in 4U~1916$-$05}
5: \shortauthors{JUETT AND CHAKRABARTY}
6: \slugcomment{Accepted for publication in the Astrophysical Journal}
7: 
8: \begin{document}
9: 
10: \newcommand{\source}{4U~1916$-$05}
11: 
12: \title{Detection of Highly Ionized Metal Absorption Lines in the
13: Ultracompact X-ray Dipper 4U 1916$-$05}
14: 
15: \author{Adrienne~M.~Juett\altaffilmark{1} and
16: Deepto~Chakrabarty\altaffilmark{2}}
17: 
18: \altaffiltext{1}{Department of Astronomy, University
19: of Virginia, Charlottesville, VA 22903; ajuett@virginia.edu} 
20: 
21: \altaffiltext{2}{Department of Physics and Kavli Institute for
22: Astrophysics and Space Research, Massachusetts Institute of
23: Technology, Cambridge, MA 02139; deepto@space.mit.edu}
24: 
25: 
26: \begin{abstract}
27: We present the high-resolution {\em Chandra X-ray Observatory}
28: persistent (non-dip) spectrum of \source\/ which revealed narrow
29: absorption lines from hydrogenic neon, magnesium, silicon, and sulfur,
30: in addition to the previous identified hydrogenic and helium-like iron
31: absorption lines.  This makes \source\/ only the second of the
32: classical X-ray dipper systems to show narrow absorption lines from
33: elements other than iron.  We propose two possible explanations for
34: the small measured line widths ($\lesssim500$--2000~km~s$^{-1}$),
35: compared to the expected Keplerian velocities ($> 1000$~km~s$^{-1}$)
36: of the accretion disk in this 50-min orbital period system, and lack
37: of wavelength shifts ($\lesssim 250$~km~s$^{-1}$). First, the ionized
38: absorber may be stationary.  Alternatively, the line properties may
39: measure the relative size of the emission region.  From this
40: hypothesis, we find that the emission region is constrained to be
41: $\lesssim$0.25 times the radial extent of the absorber.  Our results
42: also imply that the ionized absorber spans a range of ionization
43: parameters.
44: \end{abstract}
45: 
46: \keywords{binaries: close ---
47:   stars: individual (\source) ---
48:   X-rays: binaries}
49: 
50: \section{Introduction}
51: It was anticipated that high-resolution spectra of low-mass X-ray
52: binaries (LMXBs) would reveal a wealth of line features which could be
53: used to better understand accretion disk structure, but such features
54: have been largely absent in most {\em Chandra X-ray Observatory} and
55: {\em XMM-Newton} spectra of LMXBs.  One exception is the class of
56: X-ray dippers.  X-ray dippers show intensity dips, although not
57: necessarily eclipses, on the orbital period and have inclination
58: angles in the range $60^{\circ}$$<$$i$$<$$80^{\circ}$.  The dipping
59: behavior is caused by absorption of the continuum by the accretion
60: disk or its atmosphere.
61: 
62: The best studied of the X-ray dippers, the neutron star (NS) binary,
63: EXO~0748$-$676 ($P_{\rm orb}$$=$3.8~hr), showed emission lines from
64: the hydrogenic and helium-like species of nitrogen, oxygen, neon,
65: magnesium, and silicon in its {\em XMM} and {\em Chandra} spectra
66: \citep{ckb+01,jsm03}.  The {\em XMM} spectrum of the X-ray dipper
67: MXB~1659$-$298 ($P_{\rm orb}$$=$7.1~hr) showed narrow absorption lines
68: from highly ionized oxygen, neon, and iron, and a broad Fe-$K$
69: emission line \citep{sop+01}.  {\em XMM} data from four other dippers,
70: X1624$-$490 ($P_{\rm orb}$$=$21.0~hr), 4U~1254$-$69 ($P_{\rm
71: orb}$$=$3.9~hr), 4U~1916$-$05 ($P_{\rm orb}$$=$0.83~hr), and
72: 4U~1323$-$62 ($P_{\rm orb}$$=$2.94~hr) revealed similar narrow iron
73: absorption lines and some evidence for broad iron emission lines
74: \citep{pob+02,bp03,bpb+04,bmt+05,crd+05}.
75: 
76: The presence of these absorption features in X-ray dippers, but not
77: in other LMXBs, has led to the suggestion that the absorbing material
78: has a cylindrical distribution \citep[e.g.,][]{bpb+04}.  The lack of
79: variation in the properties of the absorption lines as a function of
80: orbital phase, excluding the dipping region, implies an azimuthal
81: symmetry \citep[see, e.g.,][and references therein]{bpb+04}.  This in
82: turn points to some relationship between the accretion disk and the
83: absorbing material.
84: 
85: In this paper, we present the high-resolution {\em Chandra} spectrum
86: of \source, the shortest orbital period X-ray dipper.  The 50-min
87: orbital period of 4U~1916$-$05 was discovered independently by
88: \citet{ws82} and \citet{wbm+82} from dips seen in the X-ray emission.
89: The stability of the dip period over many years led the authors to
90: associate it with the orbital period of the system \citep[but also
91: see,][]{cgb01,hch+01,rcb+02}.  The dipping behavior of \source\/ is
92: variable on the order of $\sim$4 days, varying from short
93: ($\approx$10\% of orbital phase), shallow dips to long ($\approx$40\%
94: of orbital phase), deep dips \citep[e.g.,][]{cgb01}.  Unique among the
95: X-ray dippers, the 50-min orbital period of \source\/ places it in the
96: class of ultracompact LMXBs which require hydrogen-deficient and/or
97: degenerate donors \citep[e.g.,][]{jar78}.  It has been proposed that
98: the companion is a hydrogen-deficient but not yet degenerate star
99: \citep{nrj86}.  Observational evidence points to a helium-rich donor.
100: The X-ray burst properties of \source\/ suggest that the accreting
101: material is helium-rich and that the distance to \source\/ is
102: 8.9$\pm$1.3~kpc (Galloway et al. 2005, in prep.).  A recent optical
103: spectrum of \source\/ shows a feature at 4540~\AA\/ which may be due
104: to \ion{He}{2} \citep{nj04}.
105: 
106: The non-dip spectrum can be fit with a power-law or power-law $+$
107: blackbody model in the range 0.5--10 keV.  Variations in the best-fit
108: spectral parameters are correlated with the broad-band luminosity and
109: the position of \source\/ in its color-color diagram \citep{bgb+00}.
110: In addition, \citet{adn+00} detected a broad (FWHM~$=$~0.7~keV)
111: emission feature in the {\em ASCA} spectrum at 5.9 keV with an
112: equivalent width of 87 eV attributed to Fe-$K$ emission.  Highly
113: ionized iron absorption lines were found in the {\em XMM} spectrum of
114: 4U~1916$-$05 with marginal detections of other highly ionized features
115: \citep[e.g., \ion{Mg}{12}, \ion{S}{16};][]{bpb+04}.
116: 
117: 
118: \section{Observations and Data Reduction}
119: We observed \source\/ on 2004 August 07 for 50 ks with {\em Chandra}
120: using the High Energy Transmission Grating Spectrometer (HETGS) and
121: the Advanced CCD Imaging Spectrometer \citep[ACIS;][]{cdd+05}.  The
122: HETGS carries two transmission gratings: the Medium Energy Gratings
123: (MEGs) with a range of 2.5--31~\AA\/ (0.4--5.0~keV) and the High
124: Energy Gratings (HEGs) with a range of 1.2--15~\AA\/ (0.8--10.0~keV).
125: The HETGS spectra are imaged by ACIS, an array of six CCD detectors.
126: The HETGS/ACIS combination provides both an undispersed (zeroth order)
127: image and dispersed spectra from the gratings.  The various orders
128: overlap and are sorted using the intrinsic energy resolution of the
129: ACIS CCDs.  The first-order MEG (HEG) spectrum has a spectral
130: resolution of $\Delta\lambda=$ 0.023~\AA\/ (0.012~\AA).  To reduce
131: pileup, the observation used a sub-array of 512 rows, yielding a
132: frametime of 1.7~s.  The ``level 1'' event files were processed using
133: the CIAO v3.2 data analysis
134: package\footnote{http://asc.harvard.edu/ciao/}.  The standard CIAO
135: spectral reduction procedure was performed.
136: 
137: The combined MEG $+$ HEG first order dispersed spectrum of \source\/
138: has an average count rate of 10.6$\pm$1.3~cts~s$^{-1}$.  During our
139: observation, \source\/ showed only mild dipping behavior as well as
140: two X-ray bursts (see Figure~\ref{fig:lc}).  In the analysis presented
141: here, we focus only on the persistent (non-dip) emission from \source.
142: We inspected the lightcurve at times of expected dipping, using the
143: dip ephemeris of \citet{cgb01}.  Of the 16 orbital cycles covered by
144: our observation, only six showed statistically significant dipping
145: behavior.  These dips were centered on a phase of $\approx0.15$ and
146: had width of $\approx0.15$ in phase.  To produce our persistent
147: (non-dip) spectrum, we excluded data from phases 0.0--0.3 for all
148: orbital cycles.  In addition, we excluded data during the two X-ray
149: bursts.  We filtered the level 2 event file, yielding a persistent
150: (non-dip, non-burst) exposure time of 31~ks.  We then extracted source
151: and background spectra for both the plus and minus first order MEG and
152: HEG.  Detector response files (RMFs and ARFs) were created for the
153: four spectra using the standard CIAO tools.
154: 
155: For bright sources, pileup can be a problem for CCD detectors
156: \citep[see, e.g.,][]{d03}.  We checked the dispersed spectra of
157: \source\/ and found signs of pileup in the first order MEG spectrum.
158: In dispersed spectra, pileup can affect only a limited wavelength
159: range, particularly where the effective area of the instrument is the
160: highest.  In our observation, pileup was present between 2--10 \AA\/
161: in the MEG plus and minus first orders.  No pileup was found in the
162: HEG spectra.
163: 
164: The spectral analysis was performed using ISIS \citep{hd00}.  The $+1$
165: and $-1$ order spectra were combined for the HEG and MEG respectively
166: using the ISIS function {\tt combine\_datasets}.  This function is
167: similar to summing the datasets, but more accurately takes into
168: account the different responses for each dataset.  We fit jointly the
169: HEG spectra over the range 1.6--11.5~\AA\/ and MEG spectra over the
170: range 10.0--14.0~\AA.
171: 
172: 
173: \section{Analysis and Results}
174: An initial inspection of the spectrum of \source\/ revealed narrow
175: absorption features at 12.1, 8.4, 6.2, and 4.7~\AA\/ (see
176: Figure~\ref{fig:flux}), attributable to hydrogenic neon, magnesium,
177: silicon, and sulfur.  To determine the continuum model, we fit the
178: spectrum ignoring regions around these lines as well as the Fe-$K$
179: region at 1.8~\AA.  We tested three different continuum models: power
180: law, power law $+$ blackbody, and power law $+$ disk blackbody, all
181: with interstellar absorption as described by the {\tt tbabs} model
182: using the abundances of \citet{wam00}.  Both the power law $+$
183: blackbody and power law $+$ disk blackbody models give better fits to
184: the data than the power law alone (see Table~\ref{tab:cont}).  We take
185: the power law $+$ disk blackbody model as the continuum model,
186: although we note that the choice of the power law $+$ blackbody model
187: would make little difference to our results.  The absorbed 0.5--10~keV
188: flux from \source\/ was $8.0\times 10^{-10}$~erg~cm$^{-2}$~s$^{-1}$.
189: 
190: Our best-fit model is significantly different from that found by
191: \citet{bpb+04}.  But as noted by \citet{bpb+04}, the limited energy
192: range covered by {\em XMM} and {\em Chandra} make it difficult to
193: determine a unique solution to the continuum fit.  One interesting
194: difference between our {\em Chandra} data and the {\em XMM} data, is
195: the presence of an edge at 0.98~keV.  \citet{bpb+04} found that both
196: the EPIC and RGS data were best-fit when an edge, with depth $\tau =
197: 0.11\pm0.03$, was included.  In contrast, the addition of an edge at
198: 0.98~keV does not improve the chi-squared value of the continuum fit
199: for the {\em Chandra} data.  We find an upper limit of $\tau < 0.08$
200: for the edge, barely consistent with the {\em XMM} result.  The
201: 0.5--10~keV luminosity of \source\/ was $9.0 \times
202: 10^{36}$~erg~s$^{-1}$, using a distance of 8.9 kpc, during the {\em
203: Chandra} observation.  This is $\approx$2 times greater than found
204: during the {\em XMM} observation \citep{bpb+04}.
205: 
206: With the best-fit continuum model fixed, we fit the lines from
207: \ion{Ne}{10}, \ion{Mg}{12}, \ion{Si}{14}, \ion{S}{16}, \ion{Fe}{25},
208: and \ion{Fe}{26} with Gaussian models to determine the line positions,
209: widths, and fluxes.  In addition, we determined the upper limits to
210: the flux from the helium-like ions of neon, magnesium, silicon, and
211: sulfur.  For the upper limit measurements, we fixed the wavelength to
212: those given in \citet{bn02} and we fixed the line width $\sigma =
213: 0.005$~\AA.  We marginally detected the helium-like ion of sulfur.
214: The neon lines were covered by the MEG spectrum, while the rest of the
215: lines were located in the HEG spectral region.  The best-fit
216: parameters are given in Table~\ref{tab:line} and the spectra and
217: best-fit models are shown in Figures~\ref{fig:lowz} and \ref{fig:fe}.
218: 
219: The line positions are consistent with the rest frame wavelengths of
220: hydrogenic neon, magnesium, silicon, and sulfur, and hydrogenic and
221: helium-like iron.  This suggests that there is no bulk motion along
222: our line of sight associated with the absorbing material.  In
223: addition, the measured line widths are comparable to, or less than,
224: the instrument resolution (0.023~\AA\/ for MEG, 0.012~\AA\/ for HEG).
225: Therefore, direct measurements of the velocity widths, or temperature,
226: of the absorbing material are not possible with these data.  The
227: derived temperature limits range from $<4 \times 10^{8}$~K for
228: \ion{Mg}{12}, to $1.2\pm1.1 \times 10^{10}$~K for \ion{Fe}{26}.
229: 
230: From the best fit equivalent widths (EW), we can estimate the column
231: density ($N_{\rm Z}$) of the ions using the relationship between EW,
232: $N_{\rm Z}$, and the transition oscillator strength $f_{ij}$
233: \begin{equation}
234: \frac{W_{\lambda}}{\lambda} = \frac{\pi e^2}{m_e c^2} N_{\rm Z}
235: \lambda f_{ij} = 8.85 \times 10^{-13} \, N_{\rm Z} \lambda f_{ij},
236: \end{equation}
237: where $W_{\lambda}$ is the EW in wavelength units and $\lambda$, the
238: line wavelength, is given in cm units \citep{s78}.  The equation above
239: is true only when a line is unsaturated, on the linear part of the
240: curve of growth.  For saturated lines, the exact relationship between
241: EW and $N_{\rm Z}$ becomes a more complicated function, and includes a
242: dependence on the velocity of the ions.  We used the oscillator
243: strengths given by \citet{vvf96} for the hydrogenic lines, and those
244: of \citet{bn02} for the helium-like lines.  Our results are given in
245: Table~\ref{tab:line}.  If the lines are saturated, the values
246: represent a lower limit to the column densities for each ion.
247: 
248: Comparing our results with those of \citet{bpb+04}, we find that the
249: iron line EWs are barely consistent within the errors.  Our
250: \ion{Fe}{26} EW is at the upper limit of the {\em XMM} result, while
251: our \ion{Fe}{25} EW is at the lower limit.  It is possible that the
252: greater luminosity of \source\/ during the {\em Chandra} observation
253: has caused a real difference in the \ion{Fe}{25} and \ion{Fe}{26}
254: column densities, compared to the lower luminosity {\em XMM} data, but
255: the difference is not significant enough to confirm this.  The iron
256: line widths found in the {\em Chandra} data are lower than the upper
257: limits found in the {\em XMM} data owing to the higher resolution of
258: the HETG compared to the EPIC pn.  Our EW measurements for \ion{S}{16}
259: and \ion{Ne}{10} are consistent with the values reported for the {\em
260: XMM} data, but our \ion{Mg}{12} EW is significantly below the result
261: of \citet{bpb+04}.  We note however that \citet{bpb+04} concluded that
262: their \ion{Mg}{12} detection was only marginal, therefore the
263: difference between the {\em Chandra} and {\em XMM} results may not
264: indicate a real change between the two observations.  As in the other
265: X-ray dipper observations, we found no variation in the line
266: parameters as a function of orbital phase (see Figure~\ref{fig:comp}).
267: 
268: 
269: \section{Discussion}
270: Our {\em Chandra}/HETGS observation of \source\/ revealed narrow,
271: unresolved absorption lines in the persistent emission, attributable
272: to hydrogenic neon, magnesium, silicon, and sulfur, in addition to the
273: previous identified hydrogenic and helium-like iron absorption lines.
274: This makes \source\/ only the second of the classical X-ray dipper
275: systems to show narrow absorption lines from mid-$Z$ elements
276: \citep[and references therein]{bpb+04}.
277: 
278: The properties of these lines is of particular interest for
279: understanding the emission and absorption processes.  With a 50~min
280: orbital period, \source\/ is the shortest period X-ray dipper, and as
281: such should have the smallest accretion disk.  If the lines are
282: associated with the accretion disk, they should have widths comparable
283: to the expected Keplerian velocities in the disk.  Assuming a 1.4--2.0
284: $M_{\odot}$ primary and a 0.1--0.15 $M_{\odot}$ secondary \citep[see,
285: e.g.][]{nrj86}, we can estimate the outer disk velocity assuming that
286: the disk fills 70\% of the primary Roche lobe.  We find that the outer
287: disk velocity should be in the range of 1000--1300~km~s$^{-1}$, while
288: the companion velocity should be 740--840~km~s$^{-1}$.  In all cases,
289: the measured line widths are consistent with, or below, the instrument
290: resolution.  For \ion{Mg}{12} and \ion{Fe}{25}, the upper limits on
291: the line widths are less than the estimated outer disk velocity.
292: Additionally, we measure no shift in the wavelength of the lines,
293: ruling out a strong outflow.  The \ion{S}{16} line shows a slight
294: shift ($+$0.004~\AA) but given that this is the least significant of
295: our line detections and the possibility of systematic errors, we do not
296: feel it is a significant result.
297: 
298: We suggest two possible explanations for these results.  First, the
299: absorber could be static, producing the narrow lines and lack of
300: wavelengths shifts.  This however seems unlikely given the large
301: quantities of angular momentum present in the system.  Instead, we
302: propose that the line properties are a measure of the emission
303: properties, particularly the extent of the emission.  In this case, we
304: assume that the ionized absorber is associated with either the
305: accretion disk or its atmosphere, as suggested by the implied
306: cylindrical geometry \cite[see e.g.,][]{bpb+04}.  The rotation of the
307: absorbing material is only measurable when it has a component along
308: the line of sight to the emitter.  If the bulk velocity is
309: perpendicular to the line of sight, no velocity effects will be found
310: in the absorption lines (excluding thermal or turbulent motions which
311: we take as small compared to the Keplerian velocities in the disk).
312: From the limits on the linewidths, we can calculate the maximum
313: allowable radial extent of the emitter.  We find that the X-ray
314: emission region is $\lesssim 32000$~km, assuming that the absorber is
315: located at the outer edge of the accretion disk.  This represents the
316: maximum upper limit on the extent.  We note that if the absorber is
317: located at a smaller disk radius, the upper limit will be smaller,
318: roughly 1/4 of the absorber radius.
319: 
320: Assuming that photoionization is the dominant ionization process for
321: the absorbing material, we can estimate the ionization parameter, $\xi
322: = L/n_{e} r^{2}$ \citep{tts69}, from the ratio of the helium-like to
323: hydrogenic column densities for each atomic species.  We assume that
324: the helium-like and hydrogenic ions are present at the same,
325: single-valued ionization parameter.  This is a simple approximation to
326: the expected true state of the system in which the ions exist in
327: different relative proportions over a range of ionization parameters.
328: 
329: We used
330: XSTAR\footnote{http://heasarc.gsfc.nasa.gov/docs/software/xstar/xstar.html}
331: to estimate the ion abundances as a function of $\xi$ for an optically
332: thin plasma, with constant number density and solar abundances.  We
333: note that while the companion in \source\/ is expected to be a
334: hydrogen-deficient star, and solar abundances would therefore not
335: apply, we found that varying the abundances made no significant
336: difference in the helium-like to hydrogenic column density ratios for
337: any element.  The only XSTAR input that made a substantial difference
338: in the inferred ionization parameters was the shape of the input
339: spectrum.  We tried three models, a power law with $\alpha = -1$, a
340: 10~keV bremsstrahlung, and a user defined model based on the continuum
341: emission of \source.
342: 
343: We compared the XSTAR results with the column density ratios for neon,
344: magnesium, silicon, sulfur, and iron as given in Table~\ref{tab:line}.
345: We note that these values are only valid if the lines are unsaturated.
346: The limits for neon and magnesium are too high to provide any
347: constraint on the ionization parameter.  For silicon, we find $\log
348: \xi \gtrsim$1.7--2.0 and for sulfur, $\log \xi = 1.5$--3.3, given the
349: error in the measurement and the variation in $\xi$ with spectral
350: model (highest values for the power law model, lower for the
351: bremsstrahlung and user defined models).  The most constraining result
352: comes of course from the iron ratio, where we find $\log \xi =$
353: 3.50$\pm$0.13 for the bremsstrahlung model, 3.73$\pm$0.13 for the
354: power law model, and 3.02$\pm$0.12 for the user defined model.  This
355: rough estimate would suggest that the iron absorption is found in a
356: material with higher ionization than the lower $Z$ elements.  Our
357: result is consistent with modeling of the ionized absorber in
358: MXB~1658$-$298 which required a distribution of ionization parameters
359: \citep{tpb+05}.  This is of course expected since in a material
360: ionized enough to contain helium-like and hydrogenic iron, the lower
361: $Z$ elements would be completely stripped.
362: 
363: 
364: 
365: \acknowledgements{We thank Jonathan Gelbord, Kazunori Ishibashi, and
366: Michael Nowak for useful discussions.  We would also like to thank Tim
367: Kallman for his help with XSTAR.  We also thank the anonymous referee
368: whose comments improved the paper.  Partial support for this work was
369: provided by NASA through Chandra award number GO4-5054X.}
370: 
371: 
372: \begin{thebibliography}{25}
373: \expandafter\ifx\csname natexlab\endcsname\relax\def\natexlab#1{#1}\fi
374: 
375: \bibitem[{{Asai} {et~al.}(2000)}]{adn+00}
376: {Asai}, K., {Dotani}, T., {Nagase}, F., \& {Mitsuda}, K. 2000, \apjs, 131, 571
377: 
378: \bibitem[{{Behar} \& {Netzer}(2002)}]{bn02}
379: {Behar}, E., \& {Netzer}, H. 2002, \apj, 570, 165
380: 
381: \bibitem[{{Bloser} {et~al.}(2000)}]{bgb+00}
382: {Bloser}, P.~F., {Grindlay}, J.~E., {Barret}, D., \& {Boirin}, L. 2000, \apj,
383:   542, 989
384: 
385: \bibitem[{{Boirin} {et~al.}(2005)}]{bmt+05}
386: {Boirin}, L., {M{\'e}ndez}, M., {D{\'{\i}}az Trigo}, M., {Parmar}, A.~N., \&
387:   {Kaastra}, J.~S. 2005, \aap, 436, 195
388: 
389: \bibitem[{{Boirin} \& {Parmar}(2003)}]{bp03}
390: {Boirin}, L., \& {Parmar}, A.~N. 2003, \aap, 407, 1079
391: 
392: \bibitem[{{Boirin} {et~al.}(2004)}]{bpb+04}
393: {Boirin}, L., {Parmar}, A.~N., {Barret}, D., {Paltani}, S., \& {Grindlay},
394:   J.~E. 2004, \aap, 418, 1061
395: 
396: \bibitem[{{Canizares} {et~al.}(2005)}]{cdd+05}
397: {Canizares}, C.~R., et al. 2005, \pasp, 117, 1144
398: 
399: \bibitem[{{Chou} {et~al.}(2001)}]{cgb01}
400: {Chou}, Y., {Grindlay}, J.~E., \& {Bloser}, P.~F. 2001, \apj, 549, 1135
401: 
402: \bibitem[{{Church} {et~al.}(2005)}]{crd+05}
403: {Church}, M.~J., {Reed}, D., {Dotani}, T., {Ba{\l}uci{\'n}ska-Church}, M., \&
404:   {Smale}, A.~P. 2005, \mnras, 359, 1336
405: 
406: \bibitem[{{Cottam} {et~al.}(2001)}]{ckb+01}
407: {Cottam}, J., {Kahn}, S.~M., {Brinkman}, A.~C., {den Herder}, J.~W., \& {Erd},
408:   C. 2001, \aap, 365, L277
409: 
410: \bibitem[{{Davis}(2003)}]{d03}
411: {Davis}, J.~E. 2003, in Proceedings of the SPIE, Vol. 4851, X-Ray and Gamma-Ray
412:   Telescopes and Instruments for Astronomy, ed. J.~E. {Truemper} \& H.~D.
413:   {Tananbaum}, 101--111
414: 
415: \bibitem[{{D{\'{\i}}az Trigo} {et~al.}(2006)}]{tpb+05}
416: {D{\'{\i}}az Trigo}, M., {Parmar}, A.~N., {Boirin}, L., {M{\'e}ndez}, M., \&
417:   {Kaastra}, J.~S. 2006, \aap, 445, 179
418: 
419: \bibitem[{{Homer} {et~al.}(2001)}]{hch+01}
420: {Homer}, L., {Charles}, P.~A., {Hakala}, P., {Muhli}, P., {Shih}, I.-C.,
421:   {Smale}, A.~P., \& {Ramsay}, G. 2001, \mnras, 322, 827
422: 
423: \bibitem[{{Houck} \& {Denicola}(2000)}]{hd00}
424: {Houck}, J.~C., \& {Denicola}, L.~A. 2000, in ASP Conf. Ser. 216: Astronomical
425:   Data Analysis Software and Systems IX, Vol.~9, 591--594
426: 
427: \bibitem[{{Jimenez-Garate} {et~al.}(2003){Jimenez-Garate}, {Schulz}, \&
428:   {Marshall}}]{jsm03}
429: {Jimenez-Garate}, M.~A., {Schulz}, N.~S., \& {Marshall}, H.~L. 2003, \apj, 590,
430:   432
431: 
432: \bibitem[{{Joss} {et~al.}(1978){Joss}, {Avni}, \& {Rappaport}}]{jar78}
433: {Joss}, P.~C., {Avni}, Y., \& {Rappaport}, S. 1978, \apj, 221, 645
434: 
435: \bibitem[{Nelemans \& Jonker(2005)}]{nj04} 
436: 
437: Nelemans, G., \& Jonker, P.~G. 2005, in AIP Conf. Proc. 797:
438: Interacting Binaries: Accretion, Evolution and Outcomes,
439: ed. L.~A. {Antonelli} {et al.}, 396--401
440: 
441: \bibitem[{{Nelson} {et~al.}(1986){Nelson}, {Rappaport}, \& {Joss}}]{nrj86}
442: {Nelson}, L.~A., {Rappaport}, S.~A., \& {Joss}, P.~C. 1986, \apj, 304, 231
443: 
444: \bibitem[{{Parmar} {et~al.}(2002)}]{pob+02}
445: {Parmar}, A.~N., {Oosterbroek}, T., {Boirin}, L., \& {Lumb}, D. 2002, \aap,
446:   386, 910
447: 
448: \bibitem[{{Retter} {et~al.}(2002)}]{rcb+02}
449: {Retter}, A., {Chou}, Y., {Bedding}, T.~R., \& {Naylor}, T. 2002, \mnras, 330,
450:   L37
451: 
452: \bibitem[{{Sidoli} {et~al.}(2001)}]{sop+01}
453: {Sidoli}, L., {Oosterbroek}, T., {Parmar}, A.~N., {Lumb}, D., \& {Erd}, C.
454:   2001, \aap, 379, 540
455: 
456: \bibitem[{{Spitzer}(1978)}]{s78}
457: {Spitzer}, L. 1978, {Physical processes in the interstellar medium} (New York:
458:   Wiley-Interscience)
459: 
460: \bibitem[{{Tarter} {et~al.}(1969){Tarter}, {Tucker}, \& {Salpeter}}]{tts69}
461: {Tarter}, C.~B., {Tucker}, W.~H., \& {Salpeter}, E.~E. 1969, \apj, 156, 943
462: 
463: \bibitem[{{Verner} {et~al.}(1996){Verner}, {Verner}, \& {Ferland}}]{vvf96}
464: {Verner}, D.~A., {Verner}, E.~M., \& {Ferland}, G.~J. 1996, Atomic Data and
465:   Nuclear Data Tables, 64, 1
466: 
467: \bibitem[{{Walter} {et~al.}(1982)}]{wbm+82}
468: {Walter}, F.~M., {Bowyer}, S., {Mason}, K.~O., {Clarke}, J.~T., {Henry}, J.~P.,
469:   {Halpern}, J., \& {Grindlay}, J.~E. 1982, \apjl, 253, L67
470: 
471: \bibitem[{{White} \& {Swank}(1982)}]{ws82}
472: {White}, N.~E., \& {Swank}, J.~H. 1982, \apjl, 253, L61
473: 
474: \bibitem[{{Wilms} {et~al.}(2000){Wilms}, {Allen}, \& {McCray}}]{wam00}
475: {Wilms}, J., {Allen}, A., \& {McCray}, R. 2000, \apj, 542, 914
476: 
477: \end{thebibliography}
478: 
479: \clearpage
480: \begin{deluxetable}{lcccccc}
481: \tabletypesize{\footnotesize} 
482: \tablewidth{0pt} 
483: \setlength{\tabcolsep}{0.1in}
484: \tablecaption{Best-Fit Continuum Spectral Parameters\tablenotemark{a}}
485: \tablehead{\colhead{Model} & \colhead{$N_{\rm H}$ ($10^{21}$ cm$^{-2}$)} 
486:   & \colhead{$\Gamma$} & \colhead{$A_{1}$\tablenotemark{b}} & 
487:   \colhead{$kT$ (keV)} & \colhead{Norm\tablenotemark{c}} & 
488:   \colhead{$\chi^2_{\nu}/ \nu$} } 
489: \startdata
490: PL          &  6.9$\pm$0.2 &  1.519$\pm$0.019 &  12.2$\pm$0.3 & 
491:      \nodata       & \nodata     &  1.123/1187 \\
492: PL$+$BB     &  5.6$\pm$0.7 &  1.31$\pm$0.11   &  8.5$\pm$1.7  & 
493:      0.54$\pm$0.04 &  70$\pm$30  &  1.114/1185 \\
494: PL$+$DISKBB &  5.4$\pm$0.7 &  1.0$\pm$0.4     &  5$\pm$3      & 
495:      0.91$\pm$0.11 &  15$\pm$5   &  1.113/1185 \\
496: \enddata
497: \tablenotetext{a}{All errors quoted at the 90\%-confidence level.}
498: \tablenotetext{b}{Power-law normalization in units of $10^{-2}$
499: photons cm$^{-2}$ s$^{-1}$ keV$^{-1}$.}  
500: \tablenotetext{c}{Thermal component normalization.  For the blackbody
501: model, norm $= R^2_{\rm km}/D^2_{\rm 10 \, kpc}$; for the disk
502: blackbody model, norm $= (R_{\rm in,km}/D_{\rm 10 \, kpc})^2
503: \cos\theta$, where $\theta$ is the inclination angle of the disk.}
504: \label{tab:cont}
505: \end{deluxetable}
506: 
507: 
508: \begin{deluxetable}{lccccc}
509: \tabletypesize{\footnotesize} 
510: \tablewidth{0pt} 
511: \setlength{\tabcolsep}{0.1in}
512: \tablecaption{Best-Fit Line Parameters\tablenotemark{a}}
513: \tablehead{\colhead{Ion} & \colhead{Wavelength (\AA)} & 
514:   \colhead{FWHM (\AA)} & \colhead{FWHM (km~s$^{-1}$)} & \colhead{EW (m\AA)} & 
515:   \colhead{$N_{\rm Z}$\tablenotemark{b} ($10^{16}$~cm$^{-2}$)}} 
516: \startdata
517: \ion{Ne}{10} & 12.140$\pm$0.006  & 0.035$\pm$0.019 & 870$\pm$470  & 
518:      $-25\pm$8    & 4.6$\pm$1.5 \\
519: \ion{Mg}{12} & 8.425$\pm$0.002   & $<$0.014        & $<$500       & 
520:      $-9\pm$3     & 3.4$\pm$1.1 \\
521: \ion{Si}{14} & 6.185$\pm$0.002   & 0.016$\pm$0.009 & 800$\pm$460  & 
522:      $-11\pm$3    & 8$\pm$2 \\
523: \ion{S}{16}  & 4.736$\pm$0.003   & 0.012$\pm$0.009 & 750$\pm$600  & 
524:      $-7\pm$3     & 8$\pm$4 \\
525: \ion{Fe}{25} & 1.8519$\pm$0.0017 & $<$0.005        & $<$950       & 
526:      $-4.9\pm$1.1 & 20$\pm$5 \\
527: \ion{Fe}{26} & 1.7811$\pm$0.0015 & 0.012$\pm$0.005 & 1900$\pm$900 & 
528:      $-11\pm$3    & 90$\pm$30 \\ \tableline
529: \multicolumn{6}{c}{Upper Limits on Helium-like Ions} \\
530: \tableline
531: \ion{Ne}{9}  & 13.448 (fixed)    & 0.012 (fixed)   & 260 (fixed)  & 
532:      $> -9$       & $< 0.9$ \\
533: \ion{Mg}{11} & 9.170 (fixed)     & 0.012 (fixed)   & 380 (fixed)  & 
534:      $> -6$       & $< 1.2$ \\
535: \ion{Si}{13} & 6.648 (fixed)     & 0.012 (fixed)   & 530 (fixed)  & 
536:      $> -3$       & $< 1.1$ \\
537: \ion{S}{14}  & 5.039 (fixed)     & 0.012 (fixed)   & 700 (fixed)  & 
538:      $-4\pm$3     & 2.5$\pm$1.9 \\
539: \enddata
540: \tablenotetext{a}{All errors quoted at the 90\%-confidence level.}
541: \tablenotetext{b}{Assumes lines are unsaturated.  Values represent a
542: lower limit to the column density if lines are saturated.}
543: \label{tab:line}
544: \end{deluxetable}
545: 
546: 
547: \begin{figure}
548: \epsscale{0.7}
549: \plotone{f1.ps}
550: \caption{{\em Left panel, top:} First order HEG$+$MEG 0.5--8.0~keV
551: count rate over the {\em Chandra} observation of \source.  The solid
552: marks at the top of the figure indicate the expected dip times using
553: the dip ephemeris of \citet{cgb01}.  {\em Left panel, bottom:} The
554: hardness ratio ([4--8~keV count rate]/[0.5--1.5~keV count rate])
555: during the observation.  As expected the hardness ratio increases
556: during the X-ray dips (due to low energy absorption) and X-ray bursts.
557: {\em Right Panel:} Normalized phase folded lightcurve of the first
558: order HEG$+$MEG 0.5--8.0~keV counts from \source.  The times of X-ray
559: bursts have been removed and the lightcurve has been corrected for
560: exposure effects.  The solid line from phases 0.3--1.0 indicates the
561: data used to produce the persistent (non-dip) spectrum.}
562: \label{fig:lc}
563: \end{figure}
564: 
565: \begin{figure}
566: \epsscale{0.8}
567: \plotone{f2.ps}
568: \caption{Flux spectra and best-fit model for the combined first order
569: MEG$+$HEG {\em Chandra} spectrum of \source.  The data and model are
570: binned to 0.010~\AA.  The solid line is the full best-fit model
571: including both continuum and line components.  The dotted and dashed
572: lines show the best-fit power law and disk blackbody continuum models,
573: respectively.  The absorption features from neon, magnesium, silicon,
574: sulfur, and iron are labeled.}
575: \label{fig:flux}
576: \end{figure}
577: 
578: \begin{figure}
579: \epsscale{1.0}
580: \plotone{f3.ps}
581: \caption{Flux spectra and best-fit models for the \ion{Ne}{10},
582: \ion{Mg}{12}, \ion{Si}{14}, and \ion{S}{16} line regions.  Note that
583: the \ion{Ne}{10} line is covered by the lower resolution MEG, while
584: the other lines are covered by the HEG.  The line-like feature at
585: $\approx$11.97~\AA\/ in the \ion{Ne}{10} region is more narrow than
586: the instrument resolution and is likely an instrumental artifact.}
587: \label{fig:lowz}
588: \end{figure}
589: 
590: \begin{figure}
591: \epsscale{0.7}
592: \plotone{f4.ps}
593: \caption{Flux spectrum and best-fit model for the iron line region.
594: The \ion{Fe}{25} line is found at 1.85~\AA\/ and the \ion{Fe}{26} line
595: is found at 1.78~\AA.}
596: \label{fig:fe}
597: \end{figure}
598: 
599: 
600: \begin{figure}
601: \epsscale{0.7}
602: \plotone{f5.ps}
603: \caption{Comparison of the absorption line equivalent widths as a
604: function of orbital phase.  We plot the EWs for the full dataset
605: (diamond), in addition to three phase groups: group A covering phases
606: 0.30--0.53 (triangle), group B covering phases 0.53--0.76 (square),
607: and group C covering phases 0.76--1.00 (x).  No significant variation
608: of the line properties are found.}
609: \label{fig:comp}
610: \end{figure}
611: 
612: 
613: \end{document}
614: