1: %\documentclass[letterpaper]{article}
2: \documentclass{aastex}
3: %\usepackage{emulateapj}
4: \usepackage{emulateapj5}
5: %\usepackage{onecolfloat}
6: \usepackage{apjfonts}
7: \usepackage{natbib}
8: \usepackage{graphicx}
9: \usepackage{amsmath}
10:
11:
12:
13: %\include{macro}
14: %jw \newcommand{\gtsim} {\lower.5ex\hbox{$\; \buildrel > \over \sim \;$}}
15: %jw\newcommand{\ltsim} {\lower.5ex\hbox{$\; \buildrel < \over \sim\;$}}
16:
17: %jw the following ones insert the correct characters (i.e., the ones
18: %that apJ will use anyway...)
19: \newcommand{\lumin}{\mbox{$\rm\,ergs\,s^{-1}$}}
20: \newcommand{\gtsim}{\ensuremath{\gtrsim}}
21: \newcommand{\ltsim}{\ensuremath{\lesssim}}
22:
23:
24: \newcommand{\chisq} {\mbox{$\rm\,\chi^2$}}
25: \newcommand{\amin} {\mbox{$^{\prime}$}}
26: \newcommand{\aprx} {\mbox{$\sim$}}
27: \newcommand{\rxte} {{\em RXTE}}
28: %jw extra dash removed here (ApJ style rules)
29: \newcommand{\xray} {\mbox{X-ray}}
30: \newcommand{\xrays} {\mbox{X-rays}}
31: \newcommand{\hxt}{HEXTE}
32: %jw changed {\em } to \textit{ } (latex2e style, that way the kerning
33: %will be correct)
34: \newcommand{\sax}{{\textit{Beppo}SAX}}
35: \newcommand{\eflux} {\mbox{$\rm\,ergs~cm^{-2}~s^{-1}$}}
36: \def\heaoone{\textit{HEAO-1}}
37: \def\ginga{\textit{Ginga}}
38: \def\batse{{BATSE}}
39: \def\degree{\hbox{$^\circ$}}
40: \def\cmsq{{$\rm cm^2$}}
41:
42: \def\cgro{\textit{CGRO}}
43:
44:
45: \newcommand{\srcnm} {XTE~J1946+274}
46: %
47:
48: \shorttitle{Cyclotron Line in \srcnm}
49: \shortauthors{Heindl, et al.}
50:
51: \begin{document}
52:
53: %\twocolumn[
54: \title{Discovery of a Cyclotron Resonance Scattering Feature in the X-ray Spectrum of \srcnm}
55: \author{W.A. Heindl, W. Coburn, D.E. Gruber, R.E. Rothschild,}
56: \affil{Center for Astrophysics and Space Sciences, Code 0424, University of
57: California, San Diego, La Jolla, CA 92093}
58: \author{I. Kreykenbohm, J. Wilms, R. Staubert}
59: \affil{Institut f\"ur Astronomie und Astrophysik -- Astronomie,
60: University of T\"ubingen, Waldh\"auser Strasse 64, D-72076
61: T\"ubingen, Germany}
62:
63:
64: %\authoremail{wheindl@ucsd.edu}
65: \email{wheindl@ucsd.edu}
66:
67: \begin{abstract}
68: Observations of the transient accreting pulsar \srcnm\ made with the
69: \emph{Rossi X-ray Timing Explorer} during the course of the 1998
70: September--November outburst, reveal a cyclotron resonance scattering
71: feature (or ``cyclotron line'') in the hard \xray\ spectrum near
72: 35\,keV. We determine a centroid energy of $36.2_{-0.7}^{+0.5}$\,keV,
73: which implies a magnetic field strength of $3.1(1+z) \times
74: 10^{12}$\,G, where $z$ is the gravitational redshift of the scattering
75: region. The optical depth, $\tau = 0.33_{-0.06}^{+0.07}$, and width,
76: $\sigma = 3.37_{-0.75}^{+0.92}$\,keV, are typical of known cyclotron
77: lines in other pulsars. This discovery makes \srcnm\ one of thirteen
78: pulsars with securely detected cyclotron lines resulting in
79: direct magnetic field measurements.
80: \end{abstract}
81:
82: \keywords{stars: individual (\srcnm) --- stars: neutron --- stars:
83: magnetic fields --- \xrays: binaries --- \xrays: stars}
84: %]
85:
86: \section{Introduction}
87:
88: The transient accreting \xray\ pulsar \srcnm\ was discovered during an
89: outburst in 1998 September with the All Sky Monitor (ASM) on board the
90: \emph{Rossi X-ray Timing Explorer} (\rxte) \citep{Smi98}. At the same
91: time, 15.8 second pulsations were observed with the \cgro/BATSE, with
92: the pulsating source designated GRO~J1944+26 \citep{Wil98}. The best
93: \xray\ position \citep[\sax/MECS: 1\amin\ radius, 90\% confidence, ][]{Cam98}
94: falls within the {\textit{Ariel V}} error region for the
95: transient source 3A~1946+274, making it likely that the two objects
96: are identical \citep{Cam99}.
97:
98: The initial outburst (see Fig.~\ref{f:asmlc}), which reached
99: \aprx6\,cps in the \rxte/ASM, lasted about 100\,days and was followed
100: by a series of smaller flares. \citet{Cam99} discussed the extended
101: \rxte/ASM lightcurve, finding that the flaring was nearly periodic
102: with a repetition period of \aprx80\,d. They interpret this as either
103: the half or full binary orbital period. Either case is consistent
104: with the known orbital period range of Be/\xray\ binary pulsars.
105: Given its probable orbital period and the outburst characteristics,
106: \srcnm\ is most likely another example of a Be star/\xray\ binary
107: pulsar transient. This class of binaries accounts for over half of
108: the known accreting pulsars \citep{Liu00}.
109:
110: Several \rxte\ observations (see Figure~\ref{f:asmlc}) were made
111: spanning the peak of the initial outburst.
112: In this \emph{Letter}, we report on the spectral analysis of these
113: observations and, in particular, the discovery of a cyclotron
114: resonance scattering feature, or ``cyclotron line'', at
115: \aprx35\,keV. \citet{San01} report their independent discovery of the
116: \srcnm\ cyclotron line with \sax\ in this volume.
117:
118: Cyclotron lines result from the scattering of \xrays\ by electrons in
119: quantized Landau orbits in the \aprx10$^{12}$\,G fields near the
120: magnetic poles of accretion powered pulsars. The characteristic energy
121: of the Landau transition scales with the magnetic field as $E_{cyc} =
122: (11.6\text{\,keV})(1+z)^{-1}B_{12}$, where $E_{cyc}$ is the cyclotron
123: line energy in keV, $z$ is the gravitational redshift in the
124: scattering region, and $B_{12}$ is the magnetic field in units of
125: $10^{12}$\,G. Because of this proportionality, cyclotron lines give
126: us the only direct measure of the neutron star magnetic field. In
127: general, harmonically spaced lines (corresponding to higher order
128: Landau transitions) may exist \citep[see for
129: example][]{Hei99,Hei00,San99}. Depending on the temperature and geometry of
130: the emitting and scattering material and the viewing angle with
131: respect to the magnetic field, the line profiles may be broad and
132: complex \citep[e.g.][]{Ara00,Kre00}. To date, about a dozen pulsars
133: have well-established cyclotron lines, for the most part discovered
134: with \ginga, \rxte, and \sax\ \citep[see for
135: example][]{Mak99,Hei00,Dal00}.
136:
137: %jw no, that is plenty for a letter. the above is a good summary.
138: %XXX Do I need more on cyclotron lines in the introduction?XXX
139:
140:
141: \section{Observations}
142:
143: During the \srcnm\ outburst, 12 pointed observations were made with
144: the Proportional Counter Array (PCA) and the High Energy X-ray Timing
145: Experiment (HEXTE) on-board the \rxte\ (see Tab.~\ref{t:obs} and
146: Fig.~\ref{f:asmlc}). The PCA \citep{Jah96} is a set of five Xenon
147: proportional counter units (PCUs) sensitive in the energy range
148: 2--60\,keV with a total effective area of \aprx 7000\,$\rm cm^2$. The
149: \hxt\ \citep{Rot98} consists of two arrays (``clusters A and B'') of 4
150: NaI(Tl)/CsI(Na) phoswich scintillation counters (15--250\,keV)
151: totaling \aprx1600\,\cmsq. Early in the mission, the pulse height
152: analyzer of a single cluster B phoswich failed, making the effective
153: area of cluster B approximately 3/4 of cluster A. The \hxt\ clusters
154: alternate pointing between target and nearby blank fields in order to
155: measure the background. The PCA and \hxt\ fields of view are
156: collimated to the same 1\degree\ full width half maximum (FWHM)
157: region. In order to extend the life of the PCA detectors,
158: most observations are currently made with one or more PCUs turned off.
159: All the observations here were performed with PCUs 0, 1, and 2
160: on. Because PCUs 3 and 4 were sometimes off, they have been excluded
161: from this analysis.
162: %jw I think it would make more sense to give the count rate PER PCU
163: %instead of PER 3 PCU, or at least indicate that in the table
164: % wah -- done
165:
166:
167:
168: \includegraphics[width=3.25in]{fig1.ps}
169: \figcaption{The \rxte/ASM lightcurve for the 1998
170: September outburst of \srcnm. Bars at the top indicate the PCA/HEXTE
171: observations (Table~\ref{t:obs}).}\label{f:asmlc}
172:
173: \section{Analysis}
174: For each observation, we accumulated source and background spectra
175: from the PCA standard 2 data and the \hxt\ event mode data. We used
176: the FTOOLS 5.0.1 and the procedures detailed by \citet{Wil99}.
177: In order to take advantage of the best PCA calibrations, we limited
178: the data to detector layer 1 only. We summed the spectra for the 3
179: active PCUs to form a single pulse height
180: spectrum. We used the bright source background model (appropriate for
181: the counting rate of \srcnm) to estimate the PCA background.
182: Because of the high PCA counting rate of \srcnm\ and the duration
183: of the observations, the statistical errors in the energy spectra ($\ll
184: 1\%$) were small compared to the uncertainties in the instrument
185: calibration. It was therefore necessary to include systematic
186: uncertainties on the PCA data in order to achieve a reasonable fit to
187: the data. We determined the size of these errors through the
188: procedures recommended by the PCA team \citep{Jah00} and explained in
189: \citet{Cob01}. In short, we fit contemporaneous observations of the
190: Crab Nebula and Pulsar to a two power law model, using the \srcnm\
191: response matrices. This model allows for different power law indices
192: for the pulsar and nebular components \citep{Kni82}. We
193: adjusted the systematic errors until a reduced chi-squared of one was
194: achieved -- at a level of 0.3\% per channel below 20\,keV. Above
195: 20\,keV, the systematic errors became unacceptably large, \gtsim5\%,
196: so we chose to ignore these PCA data, rather than apply large systematic
197: errors.
198:
199: \label{s:anal}
200: \begin{table*}[b]
201: \caption{\label{t:obs} Observations}
202: \begin{minipage}{\linewidth}
203: \renewcommand{\thefootnote}{\thempfootnote}
204: \begin{center}
205: \begin{tabular}{clccccc} \hline \hline
206: & & \multicolumn{2}{c}{PCA} & & \multicolumn{2}{c}{HEXTE}\\
207: \cline{3-4} \cline{6-7}
208: & Date\footnote{1998} & Time \footnote{exposure (ks)} &
209: Rate\footnote{cts s$^{-1}$ per PCU, 3--20\,keV}
210: & & Time \footnote{exposure (ks)} &
211: Rate \footnote{cts s$^{-1}$, 16--100\,keV} \\ \hline
212: 1 & Sep 16 & 1.42 & $195.0\pm0.2$ & & 0.84 & $30.1\pm0.5$ \\
213: 2 & Sep 17 & 2.35 & $208.2\pm0.2$ & & 1.50 & $33.6\pm0.4$ \\
214: 3 & Sep 18 & 2.51 & $217.5\pm0.2$ & & 1.64 & $34.6\pm0.3$ \\
215: 4 & Sep 19 & 2.94 & $216.3\pm0.2$ & & 1.85 & $33.3\pm0.3$ \\
216: 5 & Sep 21 & 2.69 & $250.0\pm0.2$ & & 1.76 & $37.9\pm0.4$ \\
217: 6 & Sep 22 & 3.82 & $254.5\pm0.2$ & & 2.42 & $38.5\pm0.3$ \\
218: 7 & Sep 23 & 2.70 & $243.0\pm0.2$ & & 1.75 & $37.6\pm0.4$ \\
219: 8 & Sep 24 & 1.54 & $258.4\pm0.2$ & & 1.09 & $38.0\pm0.4$ \\
220: 9 & Sep 29 & 4.93 & $244.9\pm0.4$ & & 3.26 & $36.4\pm0.2$ \\
221: 10 & Oct 4 & 2.93 & $219.7\pm0.2$ & & 1.81 & $33.1\pm0.3$ \\
222: 11 & Oct 8 & 2.77 & $200.6\pm0.2$ & & 1.79 & $30.7\pm0.4$ \\
223: 12 & Oct 14 & 1.08 & $167.9\pm0.2$ & & 0.71 & $25.4\pm0.5$ \\
224: & Total \footnote{Observations 1 -- 11. Observation 12 was excluded from the
225: analysis (see \S\ref{s:anal}).}
226: & 30.6 & $230.7\pm0.3$ & & 19.7 & $35.3\pm0.1$\\ \hline \hline
227: \end{tabular}
228: \end{center}
229: \end{minipage}
230: \end{table*}
231:
232: We accumulated the \hxt\ cluster A and B spectra separately then
233: summed them to form joint source and background spectra for each
234: observation. For the analysis of these summed spectra, we added the
235: cluster A and B response matrices, weighted by the cluster effective
236: %jw here, give the detailed weighing that is used in the data analysis
237: %instead of an approximate value...
238: % wah -- done
239: areas (4:3, respectively). Fits with this matrix to a single power law
240: model, appropriate in the \hxt\ energy band, to the Crab have
241: residuals of $\pm$1\%. This is smaller than the statistical
242: uncertainties, so we applied no additional systematic errors to the
243: \hxt\ data.
244:
245: Because the observations were spread over \aprx1\,month and a range of
246: \aprx40\% in flux, we were concerned that changes in the source
247: spectrum could affect our analysis. To search for spectral changes,
248: we calculated for each observation the ratio of the net PCA and HEXTE counts
249: spectra to the total spectrum of all observations (see
250: Figure~\ref{f:ratio}). In all but observation 12, we
251: \includegraphics[width=3.25in]{fig2.ps}
252: \figcaption{Ratios of the net PCA counts spectra from the
253: twelve observations (Table~\ref{t:obs}) to the total spectrum,
254: normalized to the 8--20\,keV counting rate. While a small residual
255: slope is apparent in observation 9, deviations above 8\,keV were less
256: than 2\%, so this spectrum was included in the analysis. However,
257: owing to the overall steep slope of its ratio and large deviations from one,
258: observation 12 was eliminated.}\label{f:ratio}
259:
260: \vspace{1ex}
261: \noindent found that above8\,keV, the ratio was nearly flat,
262: indicating that only the flux, and not the
263: spectral shape, had changed. In observation 12 (the last and dimmest
264: observation -- see Table~\ref{t:obs}), significant spectral changes
265: were observed, and this pointing was eliminated from our analysis.
266: Observations 1 to 11 showed no changes in spectral shape at a level of
267: \aprx1--2\% between 8--20\,keV. Below 8\,keV, systematic changes of
268: up to \aprx5\% were
269: \includegraphics[width=3.25in]{fig3.ps}
270: \figcaption{The joint PCA/HEXTE spectrum of \srcnm. Top
271: panel shows the data (crosses), the best fit NPEX model with a
272: cyclotron line (histograms), and the inferred incident photon spectrum
273: (smooth curve). See Table~\ref{t:fit} for fit parameters. The middle
274: and bottom panels show the residuals for the best fit model and the
275: best fit NPEX model with no cyclotron line.}\label{f:spec}
276:
277: \vspace{1ex}
278: \noindent seen, particularly in the lowest energy channels.
279: These changes are consistent with \ltsim10\% variations in the
280: absorbing column and the iron line emission. Such variations are not
281: unexpected during a Be/\xray\ binary outburst, and they are likely not
282: indicative of changes in the underlying continuum emission. The HEXTE
283: ratios showed no evidence for spectral variability at a limiting level of
284: \aprx5--10\% below 35\,keV. Above 35\,keV the statistics did not allow
285: for a meaningful comparison. For these
286: reasons, we concluded that by excluding data below 8\,keV we could
287: safely combine observations 1 to 11.
288:
289: After accumulating observations 1 through 11 as described above, we
290: attempted to fit the PCA and HEXTE spectra jointly with a set of
291: continuum models typically used for accreting pulsars. While there is
292: a notable lack of adequate theoretical models, pulsar spectra are
293: heuristically well-described by a power-law at low energies which
294: breaks to an exponentially cut-off power-law at high energies
295: \citep{Whi83}.
296: %jw rewrote the following
297: %\citet{Kre99} review several models -- PLCUT \citep{Whi83}, FDCO
298: %\citep{Tan86}, and NPEX \citep{Mih95} -- with this asymptotic behavior
299: %that appear in the literature.
300: \citet{Kre99} review several models with this asymptotic behavior
301: -- the exponentially cut-off power law \citep[PLCUT,][]{Whi83}, the
302: Fermi-Dirac cut-off \citep[FDCUT,][]{Tan86}, and a combination of two
303: power-laws with a positive and a negative exponent
304: \citep[NPEX,][]{Mih95}. None of these models provided an acceptable
305: fit to the \srcnm\ spectrum. All models left significant, cyclotron
306: line-like residuals
307: %jw replaced ``then'' by ``therefore''
308: near 35\,keV (see Figure~\ref{f:spec}). We therefore added to each model a
309: cyclotron absorption term with a Gaussian optical depth profile:
310: \begin{equation}
311: \tau(E) = \tau_{\text{cyc}} \times e^{-(E - E_{\text{cyc}})^2/{2\sigma_{\text{cyc}}^2}}.
312: \end{equation}
313: %jw changed the style a little bit
314: Adding the Gaussian absorption line greatly improved the fits. The
315: best fit was achieved with the NPEX continuum, given by:
316: \begin{equation}
317: \text{NPEX}(E) \propto (E^{-\Gamma_1} + \alpha E^{+\Gamma_2}) \times
318: e^{-(E/E_{\text{fold}})}
319: \end{equation}
320: where $\text{NPEX}(E)$ is the photon flux at energy $E$; $\Gamma_1$
321: and $\Gamma_2$ are the indices of the falling and rising power law
322: components, with $\alpha$ their relative normalizations; and
323: $E_{\text{fold}}$ is the exponential folding energy. Our best-fit
324: model was of the form $ \text{Flux} \propto \text{NPEX}(E) \cdot
325: \text{e}^{-\tau (E)}$. In this case, the reduced $\chisq$ changed
326: (with the addition of the line) from 3.61 to 0.97 for 52 and 49
327: degrees of freedom respectively. The resulting $F$-Test probability
328: for this to be a chance improvement is $ 1.2 \times 10^{-14}
329: $. Table~\ref{t:fit} gives the best fit parameters for this model.
330: \newpage
331:
332: \section{Results and Discussion}
333:
334: Figure~\ref{f:spec} shows the best fit model and residuals together
335: with the inferred incident photon spectrum. It also shows residuals
336: to the best NPEX model without a cyclotron line, which is then
337: apparent in the residuals around 35\,keV. From the cyclotron line
338: energy, we deduce a magnetic field strength in the scattering region
339: of $3.1(1+z) \times 10^{12}$\,G. The line is weakly resolved with the
340: HEXTE energy resolution of \aprx8\,keV (FWHM, at 35\,keV), and the
341: fitted width of the line is $\rm 9.3^{+2.5}_{-2.1}$\% of the line
342: energy. This is within the range of typical values determined with
343: the Gaussian absorption profile model \citep{Cob01b}. No harmonic
344: line near 70\,keV was found, but as is evident in Figure~\ref{f:spec},
345: the falling high energy continuum provides inadequate statistics to
346: make a sensitive search for such a feature.
347: \vspace{2ex}
348:
349: \includegraphics[width=3.25in]{fig4.ps}
350: \figcaption{Spectral cut-off energy as a function of
351: cyclotron line energy. A possible saturation of the correlation is
352: seen at energies above 30\,keV. All points are from \rxte\ except
353: A0535+26 which is from HEXE/TTM \citep{Ken94}. The line is a power law
354: with $E_{cut} \propto E_{cyc}^{0.7}$ \citep{Mak99}. See text for
355: parameter definitions.}\label{f:ecutvsecyc}
356: \vspace{1ex}
357:
358: Figure~\ref{f:ecutvsecyc} \citep[after][]{Cob01b} shows the spectral
359: cut-off energy plotted against the cyclotron line energy for 8 pulsars
360: measured with \rxte. Also shown is the (somewhat controversial) HEXE/TTM
361: result for A0535+26 \citep{Ken94}. For uniformity, the pulse-phase
362: average spectra of all sources were fit with the PLCUT continuum:
363: \vspace{0.5ex}
364: \begin{equation}
365: \text{PLCUT}(E) \propto E^{-\Gamma} \times \left\{
366: \begin{array}{ll}
367: 1 & \text{if~} E \leq E_{\text{cut}}\text{;} \\
368: e^{-(E-E_{\text{cut}})/E_{\text{fold}}} & \text{otherwise}
369: \end{array}
370: \right.
371: \end{equation}
372: smoothed at $E_{cut}$ \citep[see][~for details]{Cob01b}. Two sources
373: are excluded from the plot: 4U~1626-67, whose cut-off energy varies by
374: a factor of four with pulse phase \citep{Cob01b}, and the low
375: luminosity (\aprx$4\times 10^{34}$\,\lumin) object 4U~0352+309 (X~Per)
376: whose spectrum is not well fit by the usual pulsar models. Due to the
377: complex shape of the 4U~0115+63 fundamental line \citep{Hei00}, its
378: plotted energy is half of the first harmonic value.
379:
380:
381:
382: \begin{table*}[b]
383: \caption{Best fit spectral parameters for an
384: NPEX times Gaussian absorption profile model.}\label{t:fit}
385: \begin{minipage}{\linewidth}
386: \renewcommand{\thefootnote}{\thempfootnote}
387: \begin{tabular}{ll} \hline \hline
388: Parameter & value \\ \hline
389: Continuum & \\ \cline{1-1}
390: $\Gamma_1$ & $2.80_{-0.27}^{+0.14}$ \\
391: $\Gamma_2$ & $-1.38_{-0.10}^{+0.27}$ \\
392: $\alpha$ & $(3.1_{-1.8}^{+3.2}) \times 10^{-4}$ \\
393: $E_{fold}$ (keV) & $5.54_{-0.13}^{+0.16}$ \\
394: Flux\footnote{\eflux, 2--10\,keV} & $5.5 \times 10^{-9}$ \\ \hline
395:
396: Cyclotron line & \\ \cline{1-1}
397: $\rm \tau_{cyc} $ & $0.33_{-0.06}^{+0.07}$ \\
398: $\rm E_{cyc} $ (keV) & $36.2_{-0.7}^{+0.5}$ \\
399: $\sigma_{cyc}$ (keV) & $3.37_{-0.75}^{+0.92}$ \\ \hline
400:
401: $\chi^2_r$/degrees of freedom & 0.97/49 \\ \hline \hline
402:
403: \end{tabular}
404: \end{minipage}
405: \end{table*}
406:
407: A clear correlation exists between the line energy and spectral
408: cut-off indicating that the cut-off is related to the magnetic field.
409: \citet{Mak92} and \citet{Mak99} first noted this relationship by
410: plotting cyclotron line energies from \ginga\ and other instruments,
411: derived with a variety of continuum models, against the PLCUT cut-off
412: energy from sometimes non-contemporaneous \ginga\ observations. They
413: found that the relationship was consistent with a power law, $E_{cyc}
414: \propto E_{cut}^{1.4}$ (or, $E_{cut} \propto E_{cyc}^{0.7}$),
415: indicating a saturation as compared to a linear
416: correlation. Figure~\ref{f:ecutvsecyc} has the advantage that each
417: point is derived from a uniform model fit to a single spectrum. And,
418: with the exception of A0535+26, all points are from the same set of
419: instruments. \srcnm\ fits nicely on the correlation, and below
420: 30\,keV, the slope is consistent with $E_{cut} \propto
421: E_{cyc}^{0.7}$. However, there appears to be a break in the
422: relationship above \aprx30\,keV. This flattening
423: is more abrupt than the smooth turnover of the
424: $E_{cyc}^{0.7}$ power law suggesting that the processes that form
425: the continuum saturate at higher magnetic fields.
426:
427: With \srcnm, we have added a 13$^{th}$ accreting pulsar to the list of
428: objects with secure cyclotron line detections. Nearly all of these
429: have been confirmed, and several were discovered, with \rxte\ and \sax.
430: We have now begun detailed studies
431: of these objects as a class. In particular, since ten of the thirteen
432: sources have been observed with \rxte, we are applying uniform
433: analyses to all these objects to further understand the line forming
434: regions. First results of such studies, including correlations
435: between other line parameters, are given in \citet{Cob01b}.
436:
437:
438: \acknowledgements
439:
440: ASM data are provided by the \rxte/ASM teams at
441: MIT and at the \rxte\ SOF and GOF at NASA's GSFC. This work was
442: supported by NASA grant NAS5-30720, NSF travel grant INT-9815741,
443: DFG grant Sta~173/25-1, and a travel grant from the DAAD.
444:
445: %\bibliographystyle{jwapjbib}
446: %\bibliography{paper}
447: \newpage
448: \begin{thebibliography}{}
449:
450: \bibitem[\protect\astroncite{{Araya-G{\'o}chez} \& {Harding}}{2000}]{Ara00}
451: {Araya-G{\'o}chez}, R.~A., \& {Harding}, A.~K., 2000, ApJ, 544, 1067
452:
453:
454: \bibitem[\protect\astroncite{Campana, Israel \& Stella}{1999}]{Cam99}
455: Campana, S., Israel, G., \& Stella, L., 1999, A\&A, 352, L91
456:
457: \bibitem[\protect\astroncite{{Campana} et~al.}{1998}]{Cam98}
458: {Campana}, S., {Israel}, G.~L., {Stella}, L., \& {Santangelo}, A., 1998, IAU
459: Circ., 7039, 2
460:
461: \bibitem[\protect\astroncite{Coburn}{2001}]{Cob01b}
462: Coburn, W., 2001,
463: \newblock {\em Ph.D. thesis\/}, University of California, San Diego,
464: \newblock in preparation
465:
466: \bibitem[\protect\astroncite{Coburn et~al.}{2001}]{Cob01}
467: Coburn, W., Heindl, W., Gruber, D., Rothschild, R., Staubert, R., Wilms, J., \&
468: Kreykenbohm, I., 2001, ApJ, 552, 738
469:
470: \bibitem[\protect\astroncite{{Dal Fiume} et~al.}{2000}]{Dal00}
471: {Dal Fiume}, D., et~al., 2000,
472: \newblock in The $\rm 5^{th}$ Compton Symposium, ed. M. McConnell, J. Ryan,
473: (Melville, New York: AIP), 183
474:
475: \bibitem[\protect\astroncite{Heindl et~al.}{2000}]{Hei00}
476: Heindl, W., et~al., 2000,
477: \newblock in The $\rm 5^{th}$ Compton Symposium, ed. M. McConnell, J. Ryan,
478: (Melville, New York: AIP), 173
479:
480: \bibitem[\protect\astroncite{{Heindl} et~al.}{1999}]{Hei99}
481: {Heindl}, W.~A., {Coburn}, W., {Gruber}, D.~E., {Pelling}, M.~R., {Rothschild},
482: R.~E., {Wilms}, J., {Pottschmidt}, K., \& {Staubert}, R., 1999, ApJ, 521,
483: L49
484:
485: \bibitem[\protect\astroncite{Jahoda}{2000}]{Jah00}
486: Jahoda, K., 2000, in \emph{Rossi2000}: Astrophysics with the Rossi X-ray
487: Timing Explorer
488:
489: \bibitem[\protect\astroncite{Jahoda et~al.}{1996}]{Jah96}
490: Jahoda, K., Swank, J.~H., Giles, A.~B., Stark, M.~J., Strohmayer, T., \& Zhang,
491: W., 1996, SPIE, 2808, 59
492:
493: \bibitem[\protect\astroncite{{Kendziorra} et~al.}{1994}]{Ken94}
494: {Kendziorra}, E., et~al., 1994, \aap, 291, L31
495:
496: \bibitem[\protect\astroncite{Knight}{1982}]{Kni82}
497: Knight, F., 1982, ApJ, 260, 538
498:
499: \bibitem[\protect\astroncite{Kretschmar et~al.}{2000}]{Kre00}
500: Kretschmar, P., Araya-G{\'o}chez, R.~A., Kreykenbohm, I., Wilms, J., Staubert, R.,
501: Heindl, W.~A., Rothschild, R.~E., \& Gruber, D.~E., 2000,
502: \newblock in Proc. 4th INTEGRAL Symposium, ESA-SP),
503: \newblock submitted
504:
505: \bibitem[\protect\astroncite{{Kreykenbohm} et~al.}{1999}]{Kre99}
506: {Kreykenbohm}, I., {Kretschmar}, P., {Wilms}, J., {Staubert}, R., {Kendziorra},
507: E., {Gruber}, D.~E., {Heindl}, W.~A., \& {Rothschild}, R.~E., 1999, A\&A,
508: 341, 141
509:
510: \bibitem[\protect\astroncite{Liu, {van Paradijs} \& {van den Heuvel}}{2000}]{Liu00}
511: {Liu}, Q.~Z., {van Paradijs}, J., \& {van den Heuvel}, E.~P.~J., 2000,
512: A\&AS, 147, 25
513:
514: \bibitem[\protect\astroncite{Makishima \& Mihara}{1992}]{Mak92}
515: Makishima, K., \& Mihara, T., 1992,
516: \newblock in Frontiers of X-Ray Astronomy (Proc. of the 28$\rm^{th}$ Yamada
517: Conf.), (Tokyo: Uni. Acad. Press), ~23
518:
519: \bibitem[\protect\astroncite{{Makishima} et~al.}{1999}]{Mak99}
520: {Makishima}, K., {Mihara}, T., {Nagase}, F., \& {Tanaka}, Y., 1999, ApJ, 525,
521: 978
522:
523: \bibitem[\protect\astroncite{Mihara}{1995}]{Mih95}
524: Mihara, T., 1995,
525: \newblock {\em Ph.D. thesis\/}, University of Tokyo
526:
527: \bibitem[\protect\astroncite{Rothschild et~al.}{1998}]{Rot98}
528: Rothschild, R., et~al., 1998, ApJ, 496, 538
529:
530: \bibitem[\protect\astroncite{Santangelo et~al.}{2001}]{San01}
531: Santangelo, A., et~al., 2001, ApJ, submitted
532:
533: \bibitem[\protect\astroncite{{Santangelo} et~al.}{1999}]{San99}
534: {Santangelo}, A., et~al., 1999, ApJ, 523, L85
535:
536: \bibitem[\protect\astroncite{Smith \& Takeshima}{1998}]{Smi98}
537: Smith, D., \& Takeshima, T., 1998, IAU Circ., No. 7014
538:
539: \bibitem[\protect\astroncite{Tanaka}{1986}]{Tan86}
540: Tanaka, Y., 1986,
541: \newblock in Radiations hydrodynamics in stars and compact objects, ed. D.
542: Mihalas, K. Winkler, (New York, Heidelberg: Springer), 198
543:
544: \bibitem[\protect\astroncite{White, Swank \& Holt}{1983}]{Whi83}
545: White, N., Swank, J., \& Holt, S., 1983, ApJ, 270, 711
546:
547: \bibitem[\protect\astroncite{Wilms et~al.}{1999}]{Wil99}
548: Wilms, J., Nowak, M., Dove, J., Fender, R., \& {di Matteo}, T., 1999, ApJ, 522
549:
550: \bibitem[\protect\astroncite{Wilson et~al.}{1998}]{Wil98}
551: Wilson, C.~A., Finger, M.~H., Wilson, R.~B., \& Scott, D.~M., 1998, IAU Circ.,
552: No. 7014
553:
554: \end{thebibliography}
555:
556:
557:
558: \end{document}
559:
560: }
561:
562: