1: \documentclass{emulateapj}
2: \usepackage{apjfonts}
3:
4: %\usepackage{amsmath}
5: \usepackage{amssymb}
6: \usepackage{natbib}
7:
8: \usepackage{xspace}
9: \usepackage{graphicx}
10:
11: \newcommand{\err}[2]{\ensuremath{^{+#1}_{-#2}}\xspace}
12: \newcommand{\Rsun}{\ensuremath{\mbox{R}_\odot}\xspace}
13: \newcommand{\Msun}{\ensuremath{M_\odot}\xspace}
14: \newcommand{\Lsun}{\ensuremath{\mbox{L}_\odot}\xspace}
15:
16: \slugcomment{Received 2005 June 24; accepted 2005 October 10}
17:
18: \shorttitle{Cyclotron lines in V\,0332$+$53}
19: \shortauthors{Pottschmidt et al.}
20:
21: \begin{document}
22:
23: \title{\textsl{RXTE} Discovery of Multiple Cyclotron Lines during the 2004
24: December Outburst of V\,0332$+$53}
25:
26: \author{Katja~Pottschmidt\altaffilmark{1},
27: Ingo~Kreykenbohm\altaffilmark{2,3},
28: J\"orn~Wilms\altaffilmark{4},
29: Wayne~Coburn\altaffilmark{5},
30: Richard~E.~Rothschild\altaffilmark{1},
31: Peter~Kretschmar\altaffilmark{6},
32: Vanessa~McBride\altaffilmark{7},
33: Slawomir~Suchy\altaffilmark{1},
34: R\"udiger~Staubert\altaffilmark{2}}
35:
36: \altaffiltext{1}{University of California, San Diego, Center for
37: Astrophysics and Space Sciences, 9500 Gilman Dr., La Jolla, CA
38: 92093-0424, USA, \{kpottschmidt,rrothschild,ssuchy\}@ucsd.edu}
39: \altaffiltext{2}{Institut f\"ur Astronomie und Astrophysik,
40: Astronomy Section, Sand 1, 72076 T\"ubingen, Germany,
41: staubert@astro.uni-tuebingen.de} \altaffiltext{3}{\textsl{INTEGRAL}
42: Science Data Centre, Chemin d'\'Ecogia 16, 1290 Versoix,
43: Switzerland, ingo.kreykenbohm@obs.unige.ch}
44: \altaffiltext{4}{Department of Physics, University of Warwick,
45: Coventry CV4 7AL, United Kingdom, j.wilms@warwick.ac.uk}
46: \altaffiltext{5}{Space Sciences Laboratory, University of
47: California, Berkeley, Berkeley, CA 94702-7450, USA,
48: wcoburn@ssl.berkeley.edu} \altaffiltext{6}{European Space Agency,
49: European Space Astronomy Centre, Villafranca del Castillo, 28080
50: Madrid, Spain, peter.kretschmar@esa.int} \altaffiltext{7}{School of
51: Physics and Astronomy, Southampton University, Southampton SO17 1BJ,
52: United Kingdom, vanessa@astro.soton.ac.uk}
53:
54: \email{kpottschmidt@ucsd.edu}
55:
56: \begin{abstract}
57: We present an analysis of the 2--150\,keV spectrum of the transient
58: X-ray pulsar V\,0332$+$53 taken with the Rossi X-Ray Timing Explorer
59: (\textsl{RXTE}) in 2004 December. We report on the detection of
60: three cyclotron resonance features at 27, 51, and 74\,keV in the
61: phase-averaged data, corresponding to a polar magnetic field of
62: $2.7\times 10^{12}$\,G. After 4U\,0115+63, this makes V\,0332+53 the
63: second accreting neutron star in which more than two cyclotron lines
64: have been detected; this has now also been confirmed by
65: \textsl{INTEGRAL}. Pulse-phase spectroscopy reveals remarkably
66: little variability of the cyclotron line through the 4.4\,s X-ray
67: pulse.
68: \end{abstract}
69:
70: \keywords{pulsars: individual (V\,0332+53) --- stars: flare --- stars:
71: magnetic fields--- X-rays: binaries --- X-rays: stars}
72:
73:
74:
75: \section{Introduction}
76:
77: The recurring transient X-ray pulsar V\,0332+53 was discovered in 1983
78: in \textsl{Tenma} observations made from 1983 November to 1984 January
79: \citep{tanaka83a,makishima90a}. Subsequently, an outburst was revealed
80: in \textsl{Vela 5B} data to have occurred in 1973 summer
81: \citep{terrell83a,terrell84a}. In rapid follow-up observations to the
82: \textsl{Tenma} detection by \textsl{EXOSAT} during the 1983 outburst,
83: 4.4\,s pulsations were discovered, an accurate position was
84: determined, and indications were found for an orbital period of
85: 34.25\,days \citep{stella85a}. The position and subsequent
86: optical/X-ray variability secured the identification of the
87: counterpart with the heavily reddened B star BQ~Cam \citep{Argyle83a}.
88: Later investigations refined the spectral type to O8--9
89: \citep{negueruela99a}.
90:
91: Analysis of the \textsl{Tenma} data revealed a shape similar to that
92: seen in other accreting X-ray pulsars with a flat power law and
93: exponential cutoff, and also showed evidence of a cyclotron line at
94: $\sim$28\,keV \citep{makishima90a}. In 1989 September the source
95: experienced another outburst, this time observed by \textsl{Ginga}
96: \citep{makino89a}. With the energy range of the Large Area Counters
97: adjusted to cover the 2--60\,keV range, cyclotron resonant scattering
98: features (CRSF) were detected at 28.5 and 53\,keV
99: \citep{makishima90b}. At the time, this was the fourth accreting X-ray
100: pulsar system to exhibit cyclotron features. Since then a significant
101: number of this class of objects have been shown to have CRSFs, with
102: 4U~0115$+$63 setting the record for the number of harmonics at five
103: \citep[see][for a review]{heindl04a}.
104:
105: V\,0332+53 went into outburst again in 2004 November, reaching a
106: 1.5--12\,keV intensity of $\sim$1\,Crab \citep{remillard04a}. A long
107: series of observations by the Proportional Counter Array
108: \citep[PCA;][]{jahoda:05a}\footnote{See also
109: \url{http://lheawww.gsfc.nasa.gov/docs/xray/xte/pca/}.} and High
110: Energy X-Ray Timing Experiment \citep[HEXTE;][]{rothschild:98a} aboard
111: \textsl{RXTE} were made throughout the outburst. Following up on our
112: initial announcement \citep{coburn05a}, here we report the discovery
113: of a third cyclotron line at $\sim$74\,keV, along with clear evidence
114: for a non-Gaussian shape of the fundamental. In \S2 we present the
115: observations and the data extraction. The phase-averaged and the
116: phase-resolved spectra are analyzed and discussed in \S3 and \S4,
117: respectively. In \S5 we summarize the results and present our
118: conclusions.
119:
120: \section{Observations and Data Extraction}
121:
122: Since the luminosity of the source is strongly variable during the
123: outburst, we decided to concentrate on the three longest observations
124: during the peak of the outburst as observed in the soft X-ray band by
125: the \textsl{RXTE}-All Sky Monitor (ASM; Fig.~\ref{fig:asm}). Results
126: from the full monitoring data set will be presented elsewhere. We
127: combined the PCA and HEXTE data taken 2004 December 24--26
128: (\textsl{RXTE} ObsIDs 90089-11-05-08G, 90089-22-01-00G, and
129: 90089-22-01-01G), giving a total of $\sim$58\,ks of exposure in the
130: PCA and 14.4\,ks of dead-time-corrected live time in HEXTE cluster B.
131: At the time of the selected observations, HEXTE's cluster A was not
132: rocking due to a single event upset, and we discarded data from this
133: cluster\footnote{Cluster~A was subsequently rebooted and continues to
134: work normally.}. The data were reduced with HEASOFT version 5.3.1, and
135: spectral modeling was performed with XSPEC version 11.3.1w
136: \citep{arnaud:96a}. For modeling the phase-averaged spectra
137: (Fig.~\ref{fig:cycl}a), PCA data from 3--20\,keV and HEXTE data from
138: 18--150\,keV were used, and the HEXTE data were rebinned by factors of
139: 3 and 5 between 70--100 and 100--150\,keV, respectively. A systematic
140: error of 0.5\% was assumed for the PCA.
141:
142: \begin{figure}
143: \plotone{f1.eps}
144: \caption{\textsl{RXTE} ASM 1.5--12\,keV light curve of the 2004/2005
145: outburst of V\,0332+53, binned to 0.5\,day resolution. Tick marks
146: denote the times of the observations analyzed in this Letter.}
147: \label{fig:asm}
148: \end{figure}
149:
150: \section{Phase-averaged Spectra}
151:
152: The X-ray spectrum observed from accreting X-ray pulsars is the
153: superposition of X-rays originating in the accretion column and the
154: hot spots on the surface of the neutron star. The neutron star's
155: X-rays are partially intercepted by the material present in the
156: system, such as the stellar wind, giving rise to significant
157: absorption which can exceed $10^{23}\,\rm{H-atoms}\,\rm{cm}^{-2}$
158: (see, e.g., \citealt{becker05a} for a discussion of the formation of
159: the X-ray continuum in the accretion column).
160:
161: The ${\cal O}(10^{12}\,\rm{G})$ magnetic field close to the magnetic
162: poles of the neutron star leads to the formation of CRSFs in the X-ray
163: spectrum. The observer's frame energy of the CRSFs is given by
164: \citep{meszaros92a}
165:
166: %\begin{align}
167: \begin{eqnarray}
168: \label{eq:cyce}
169: E_{\rm{C},n} & = \frac{2nE_{\rm F}}{1+z} \left(1 + \sqrt{1 + 2n \frac{E_{\rm F}}{m_{\rm{e}} c^2}\sin^2\theta} \right)^{-1} \\
170: \label{eq:nonrel} & \sim \frac{nE_{\rm F}}{1+z},
171: \end{eqnarray}
172: %\end{align}
173: where $n=1$,2,\ldots, is the harmonic number, $E_{\rm F} = 11.6\,{\rm
174: keV} \cdot (B/10^{12}\,{\rm G})$ is the fundamental nonrelativistic
175: cyclotron energy ($n=1$), $\theta$ is the angle between the wave
176: vector of the incoming photon and the magnetic field $B$, and
177: $z\sim0.3$ is the gravitational redshift (assuming a $1.4\,M_\odot$
178: neutron star with a 10\,km radius). The great importance of CRSF
179: observations is that they provide the only direct way to measure
180: neutron star $B$-field strengths.
181:
182: The lack of adequate theoretical continuum models for accreting
183: neutron stars necessitates the use of empirical models to describe the
184: observations \citep{kreykenbohm99a}. We use a power law modified at
185: higher energies by a ``Ferm- Dirac cutoff'' \citep{tanaka86a} to fit
186: the PCA and HEXTE data simultaneously, including a multiplicative
187: constant as a fitting parameter to allow for the slight difference
188: ($\lesssim$5\%) in the flux calibration of both
189: instruments. Photoelectric absorption and a strong Fe K$\alpha$
190: fluorescence line at 6.35\,keV from neutral iron, with a width of
191: $\sigma_{\rm{Fe K\alpha}}=0.43$\,keV and an equivalent width of
192: 123\,eV are also taken into account. The soft continuum parameters
193: found (see below) are in general agreement with earlier results
194: \citep[e.g.,][]{unger92a}.
195:
196: While this model fits the overall shape of the continuum reasonably
197: well, strong absorption-like features are present above 20\,keV
198: (Fig.~\ref{fig:cycl}b), which we model as lines with Gaussian optical
199: depth profiles of the form
200: \begin{equation}\label{gabs}
201: \tau(E) = \tau_{\rm{C}}
202: \cdot \exp\left(
203: -\frac{1}{2}\left(\frac{E-E_{\rm{C}}}{\sigma_{\rm{C}}}\right)^2
204: \right),
205: \end{equation}
206: where $E_{\rm{C}}$ is the line energy, $\sigma_{\rm{C}}$ is the line
207: width, and $\tau_{\rm{C}}$ is the optical depth at the line
208: center. The spectral shape is then given by multiplying the continuum
209: spectrum by $\exp(-\tau(E))$ \citep[][and references
210: therein]{kreykenbohm04a}.
211:
212: \begin{figure}
213: \plotone{f2.eps}
214: \caption{Cyclotron line modeling of the phase averaged PCA/HEXTE
215: spectrum of V\,0332+53. (a) Combined PCA/HEXTE spectrum (crosses),
216: best fit model (histogram) and unfolded spectrum (dotted and solid
217: lines, right $y$-axis label), illustrating the cumulative effect of
218: the CRSFs on the continuum (see text for a description of the
219: continuum model). Note that the lines are already visible in the raw
220: count spectrum. The inset shows the best-fit spectra with one
221: (dotted line) and two (dashed line) line components for the
222: fundamental line relative to the best-fit continuum with no lines
223: (solid line). Panels (b)--(f) show residuals for models taking
224: different numbers of CRSFs into account (see Table~1 for the best
225: fit values). (b) No line; (c) one line at 25.5\,keV; (d) two lines,
226: one at 26.6\,keV and one at 50\,keV; (e) two lines, a fundamental
227: line that is best modeled with two Gaussian components at 25.2 and
228: 26.6\,keV and its harmonic at 49.9\,keV; (f) three lines, the
229: fundamental modeled by two components at $\sim$25 and $\sim$27\,keV,
230: and two lines at $\sim$51 and $\sim$74\,keV. [\textsl{See the
231: elctronic edition of the Journal for a color version of this
232: figure.}]}\label{fig:cycl}
233: \end{figure}
234:
235: We start by modeling the strongest feature between 20 and 30\,keV with
236: one of these Gaussian components (see Table~1 for the relevant best
237: fit parameters). This feature corresponds to the CRSF first reported
238: by \citet{makishima90b}. While the 20--30\,keV residuals improve
239: significantly compared to the pure continuum fit when this component
240: is included, another strong and remarkably well defined
241: absorption-like feature emerges at $\sim$50\,keV
242: (Fig.~\ref{fig:cycl}c). This feature is the first harmonic of the
243: cyclotron line, found at $\sim 2$ times the energy of the fundamental
244: and already tentatively reported by \citet{makishima90b}. Taking these
245: two lines into account, weaker residuals remain at $\sim$25\,keV
246: (Fig.~\ref{fig:cycl}d). Similar to 4U\,0115+63 \citep{heindl99a} and
247: other sources, these residuals are caused by the fundamental CRSF
248: having a non-trivial shape, which cannot be modeled well with any
249: simple line shape (see \S\ref{sec:summary}). To describe the complex
250: shape we add a shallow ($\tau_{\rm{cyc,1b}}\sim0.3$) and narrow
251: ($\sigma_{\rm{cyc,1b}}\sim 1.5$\,keV) line component at an energy
252: similar to that of the fundamental,
253: $E_{\rm{cyc,1b}}=25.2\err{0.2}{0.2}$\,keV; the shape of this more
254: complex line profile is shown in the inset of Fig.~\ref{fig:cycl}a.
255: Although lowering the $\chi^2$ considerably, the fit is still not
256: acceptable, and significant residuals remain (Fig.~\ref{fig:cycl}e).
257: Adding another line at $73.7\err{1.5}{1.3}$\,keV improves the fit
258: significantly to $\chi^2_{\rm{red}} = 1.04$ (Fig.~\ref{fig:cycl}f, the
259: unfolded spectrum and the consecutive effect of the individual CRSFs
260: are shown in Fig.~\ref{fig:cycl}a). The $F$-test probability that
261: this improvement is due to chance is only $3\times 10^{-6}$ (however,
262: see \citealt{protassov02a} concerning the applicability of the
263: $F$-test in this case).
264:
265: \begin{deluxetable}{lrrrr}
266: \tablecaption{Spectral parameters for the phase-averaged
267: spectrum. \label{tab:values}} \tablecolumns{5} \tablehead{Parameter &
268: 2c & 2d & 2e & 2f} \startdata $N_{\rm{H}}$ ($10^{22}\,\rm{cm}^{-2}$) &
269: $0.0^{+0.2}_{-0.0}$ & $1.3^{+0.5}_{-0.3}$ & $1.5^{+0.5}_{-0.7}$ &
270: $1.6^{+0.6}_{-0.6}$ \\[1mm] $\Gamma$ & $-0.16^{+0.02}_{-0.01}$ &
271: $0.04^{+0.04}_{-0.03}$ & $0.37^{+0.07}_{-0.07}$ &
272: $0.42^{+0.05}_{-0.05}$ \\[1mm] $E_{\rm{cut}}$ (keV) &
273: $0.0^{+0.1}_{-0.0}$ & $0.0^{+0.7}_{-0.0}$ & $13.0^{+5.9}_{-4.2}$ &
274: $17.2^{+3.2}_{-4.6}$ \\[1mm] $E_{\rm{fold}}$ (keV) &
275: $6.1^{+0.0}_{-0.0}$ & $7.1^{+0.2}_{-0.1}$ & $7.0^{+0.2}_{-0.2}$ &
276: $8.0^{+0.7}_{-0.5}$ \\[1mm] $E_{\rm{cyc,1a}}$ (keV) &
277: $25.5^{+0.0}_{-0.0}$ & $26.3^{+0.1}_{-0.1}$ & $26.6^{+0.1}_{-0.1}$ &
278: $27.1^{+0.2}_{-0.1}$ \\[1mm] $\sigma_{\rm{cyc,1a}}$ (keV) &
279: $4.7^{+0.0}_{-0.0}$ & $5.5^{+0.1}_{-0.1}$ & $7.0^{+0.5}_{-0.3}$ &
280: $7.6^{+0.2}_{-0.2}$ \\[1mm] $\tau_{\rm{cyc,1a}}$ &
281: $1.02^{+0.01}_{-0.01}$ & $1.22^{+0.03}_{-0.04}$ &
282: $1.38^{+0.20}_{-0.11}$ & $1.81^{+0.08}_{-0.06}$ \\[1mm]
283: $E_{\rm{cyc,1b}}$ (keV) & \nodata & \nodata & $25.2^{+0.2}_{-0.2}$ &
284: $25.2^{+0.2}_{-0.2}$ \\[1mm] $\sigma_{\rm{cyc,1b}}$ (keV) & \nodata &
285: \nodata & $1.3^{+0.5}_{-0.5}$ & $1.5^{+0.3}_{-0.5}$ \\[1mm]
286: $\tau_{\rm{cyc,1b}}$ & \nodata & \nodata & $0.35^{+0.13}_{-0.04}$ &
287: $0.34^{+0.09}_{-0.03}$ \\[1mm] $E_{\rm{cyc,2}}$ (keV) & \nodata &
288: $50.0^{+0.4}_{-0.4}$ & $49.9^{+0.4}_{-0.3}$ & $51.3^{+0.4}_{-0.4}$
289: \\[1mm] $\sigma_{\rm{cyc,2}}$ (keV) & \nodata & $6.5^{+0.5}_{-0.3}$ &
290: $6.5^{+0.5}_{-0.5}$ & $8.9^{+0.6}_{-0.6}$ \\[1mm] $\tau_{\rm{cyc,2}}$
291: & \nodata & $0.94^{+0.09}_{-0.09}$ & $1.19^{+0.14}_{-0.11}$ &
292: $2.21^{+0.31}_{-0.26}$ \\[1mm] $E_{\rm{cyc,3}}$ (keV) & \nodata &
293: \nodata & \nodata & $73.7^{+1.5}_{-1.3}$ \\[1mm] $\sigma_{\rm{cyc,3}}$
294: (keV) & \nodata & \nodata & \nodata & $4.5^{+2.5}_{-2.7}$ \\[1mm]
295: $\tau_{\rm{cyc,3}}$ & \nodata & \nodata & \nodata &
296: $3.3^{+2.6}_{-1.1}$ \\[1mm] $\chi^2_{\rm{red}}/\rm{dof}$ & $16.24/104$
297: & $4.05/101$ & $1.37/98$ & $1.04/95$ \enddata \tablecomments{ The
298: columns refer to residuals shown in Fig.~2. Symbols denote the
299: equivalent hydrogen column, $N_{\rm{H}}$, the photon index, $\Gamma$,
300: the cutoff and folding energies of the Fermi-Dirac continuum,
301: $E_{\rm{cut}}$ and $E_{\rm{fold}}$, and the cyclotron line parameters
302: $E_{\rm{cyc}}$, $\sigma_{\rm{cyc}}$, and $\tau_{\rm{cyc}}$ (defined in
303: Eq.~\ref{gabs}) for fits with 1, 2, and 3 CRSFs. CRSF parameters
304: labeled 1a and 1b correspond to the complex profile of the fundamental
305: line. See text for a description of the continuum
306: model. $\chi^2_{\rm{red}}$ is the reduced $\chi^2$ and dof is the
307: number of degrees of freedom. All uncertainties are 90\% confidence
308: levels for one interesting parameter ($\Delta\chi^2=2.71$). }
309: \end{deluxetable}
310:
311: To verify the detection of the CRSFs and to confirm their independence
312: from the choice of the continuum, we modeled the data with other
313: continua commonly used to describe the spectra of accreting pulsars.
314: The negative positive exponential continuum
315: \citep[NPEX,][]{mihara95a}, for example, also requires the presence of
316: three CRSFs at 27, 51.7, and 74.8\,keV, again with a non-Gaussian
317: fundamental. Other continua gave similar results, as do other
318: prescriptions for the shape of the CRSF. The actual choice of the
319: continuum or line shape is thus irrelevant for the detection and
320: characterization of the CRSFs at the level of the \textsl{RXTE} energy
321: resolution. We therefore report the discovery of the second harmonic
322: of the cyclotron line in V\,0332+53 with \textsl{RXTE}, confirming our
323: preliminary initial report for this data set \citep{coburn05a}. This
324: discovery has been independently confirmed by subsequent
325: \textsl{INTEGRAL} observations \citep{kreykenbohm05a}.
326:
327: \section{Phase-resolved Spectra}
328:
329: \begin{figure}
330: \plotone{f3.eps}
331: \caption{Pulse profile and CRSF variability of V\,0332+53. (a) HEXTE
332: (full band, upper profile) and PCA (20--50\,keV; lower profile)
333: pulse profile. Also shown are the phase evolution of (b) the energy
334: and (c) the depth of the fundamental and (d, e) the first harmonic
335: cyclotron line. Error bars are at the 90\% level for one interesting
336: parameter; dotted lines indicate the best-fit values from the two
337: CRSF fit to the phase averaged spectrum. [\textsl{See the elctronic
338: edition of the Journal for a color version of this figure.}]}
339: \label{fig:profile}
340: \end{figure}
341:
342: The neutron star's magnetic axis is offset from its spin axis, not
343: only giving rise to X-ray pulsations but also confronting the observer
344: with a complex situation. Due to the rotation of the neutron star, the
345: accretion column and the hot spots are seen under a constantly
346: changing viewing angle. Since the physical conditions are likely to be
347: variable within the emission region and since the CRSF line shape
348: depends on the viewing angle, we expect the X-ray continuum and line
349: parameters to depend on the pulse phase
350: \citep{araya00a,araya99a,isenberg98a}. This has been confirmed by
351: pulse-phase-resolved spectroscopy in a number of sources, such as
352: GX~301$-$2 \citep[][finding $\sim$20\% change in the line
353: energy]{kreykenbohm04a}, Vela~X-1 \citep{kreykenbohm02a,barbera03a},
354: or 4U~0115+63 \citep{mihara04a}.
355:
356: To obtain phase-resolved spectra, we determined the pulse period for
357: each of the three observations using standard techniques. This
358: approach is necessary, as the accumulated uncertainty of the only
359: available orbital ephemeris of the neutron star \citep{stella85a} is
360: too large to allow any meaningful orbit correction. The average
361: non-orbit-corrected X-ray period is $P=4.3745$\,s. After
362: phase-aligning the resulting pulse profiles, we determined
363: phase-resolved spectra. Fig.~\ref{fig:profile}a shows the resulting
364: HEXTE and PCA pulse profile of V\,0332+53.
365:
366: Since the data modes of the observation do not contain a PCA data mode
367: with energy resolution below 20\,keV and a sufficiently high time
368: resolution for phase-resolved spectroscopy, we performed phase
369: resolved spectroscopy using the HEXTE data only. Signal-to-noise ratio
370: considerations allow a maximum of six phase bins
371: (Fig.~\ref{fig:profile}a). In the resulting phase-resolved spectra,
372: the source is detected to $\sim$100\,keV, such that we only model the
373: lower two CRSFs and do not include the structure of the fundamental
374: line. The absence of soft spectral data forces us to hold $\Gamma$
375: and $E_{\rm{cut}}$ at the values found from the appropriate
376: phase-averaged spectrum. This is justified by the relative constancy
377: of these parameters as seen with earlier instruments
378: \citep{unger92a}. Both CRSFs are detected with high significance in
379: all pulse phases. Figs.~\ref{fig:profile}b--e show the evolution of
380: the CRSF line energy and depth over the X-ray pulse. In contrast to
381: the sources mentioned above, there is little to no variation in these
382: parameters, no other spectral parameters show any significant
383: variation either.
384:
385: \section{Summary and Conclusions}\label{sec:summary}
386:
387: We have presented the discovery of a third cyclotron line in
388: V\,0332+53, making this X-ray pulsar the second, after 4U\,0115+63,
389: with more than two cyclotron lines, and confirming the identification
390: of the 26\,keV line as the fundamental CRSF. Similar to sources such
391: as 4U\,0115+63 or GX~301$-$2, the HEXTE spectrum shows that the
392: profile of the fundamental line has an asymmetric profile that is
393: shallower toward lower energies. Such a profile is consistent with
394: expectations from the asymmetric shape of the relativistic cross
395: section for resonant electron scattering, but could also be influenced
396: by the emission wings predicted in Monte Carlo simulations
397: \citep{araya00a,isenberg98a} or by a variation of the $B$-field in the
398: emission region.
399:
400: Consistent with the trend seen in most sources with multiple CRSFs,
401: the energies of the fundamental and its harmonics do not scale like
402: 1:2:3, as predicted by the non-relativistic Eq.~\ref{eq:nonrel}.
403: Rather, the line ratios are slightly smaller, in agreement with the
404: prediction of relativistic quantum mechanics (Eq.~\ref{eq:cyce}).
405: Using all three observed line energies (taking $E=26.3$\,keV for the
406: fundamental line) and assuming $z=0.3$, for V\,0332$+$53 a most
407: probable polar magnetic field of $2.7\times 10^{12}$\,G is found.
408: Consistency between all three CRSF energies can only be reached for
409: $\theta\gtrsim 60^\circ$; i.e., photons scattered out of the line have
410: an initial direction that is perpendicular to the $B$-field, as is
411: expected from the significant increase of the electron scattering
412: cross section with $\theta\rightarrow 90^\circ$. This result provides
413: a much stronger confirmation of the interpretation of the line
414: features as CRSFs than is possible in sources where just the
415: fundamental is observed.
416:
417: Pulse-phase-resolved spectroscopy revealed that V\,0332+53 shows far
418: less variability in the CRSF parameters than many other sources. The
419: origin of the CRSF variability in these sources is unclear, although
420: it has been pointed out that their complex pulse profiles make it
421: likely that different regions of the accretion column or even
422: different magnetic poles are seen during the X-ray pulse. The
423: remarkably simple hard X-ray pulse profile of V\,0332+53 suggests a
424: less complex accretion geometry than in other sources, which might
425: also explain the very weak CRSF variability.
426:
427: In conclusion, our observation of multiple lines strongly confirms the
428: interpretation of these features as cyclotron lines. For a further
429: interpretation of our results, however, more realistic models for the
430: neutron star continuum and especially for the line shape are needed,
431: to include the radiative transfer in the line, multiple photon
432: scattering, etc., and thus also predict emission wings and strongly
433: asymmetric profiles. Work is currently ongoing in our collaboration
434: to compute such models for a variety of parameters to make such
435: progress possible in the near future.
436:
437: \begin{acknowledgments}
438: We thank both first referees for their remarks, which greatly
439: improved this \textsl{Letter}. We acknowledge the support of NASA
440: contract NAS5-30720, NASA grant NAG5-10691, NSF grants INT-9815741
441: and INT-0003773, and DLR grants 50OG9601, 50OG0501, and
442: 50OR0302. V.~McBride is funded jointly by the NRF (South Africa),
443: British Council, and Southampton University.
444: \end{acknowledgments}
445:
446: \begin{thebibliography}{}
447:
448: \bibitem[\protect\astroncite{Araya \& Harding}{1999}]{araya99a}
449: Araya, R.~A., \& Harding, A.~K., 1999, ApJ, 517, 334
450:
451: \bibitem[\protect\astroncite{Araya-G{\'o}chez \& Harding}{2000}]{araya00a}
452: Araya-G{\'o}chez, R.~A., \& Harding, A.~K., 2000, ApJ, 544, 1067
453:
454: \bibitem[\protect\astroncite{{Argyle} et~al.}{1983}]{Argyle83a}
455: {Argyle}, R.~W., {Kodaira}, K., {Bernacca}, P.~L., {Iijima}, T., \& {Stagni},
456: R., 1983,
457: \newblock {V0332+53},
458: \newblock {IAU Circ.\ 3897}
459:
460: \bibitem[\protect\astroncite{Arnaud}{1996}]{arnaud:96a}
461: Arnaud, K.~A., 1996,
462: \newblock in Astronomical Data Analysis Software and Systems {V}, ed. J.~H.
463: Jacoby, J. Barnes, (San Francisco: Astron.\ Soc.\ Pacific), ~17
464:
465: \bibitem[\protect\astroncite{{Becker} \& {Wolff}}{2005}]{becker05a}
466: {Becker}, P.~A., \& {Wolff}, M.~T., 2005, ApJ, 621, L45
467:
468: \bibitem[\protect\astroncite{{Coburn} et~al.}{2005}]{coburn05a}
469: {Coburn}, W., {Kretschmar}, P., {Kreykenbohm}, I., {McBride}, V.~A.,
470: {Rothschild}, R.~E., \& {Wilms}, J., 2005,
471: \newblock {The Astronomer's Telegram 381}
472:
473: \bibitem[\protect\astroncite{{Heindl} et~al.}{1999}]{heindl99a}
474: {Heindl}, W.~A., {Coburn}, W., {Gruber}, D.~E., {Pelling}, M.~R., {Rothschild},
475: R.~E., {Wilms}, J., {Pottschmidt}, K., \& {Staubert}, R., 1999, ApJ, 521,
476: L49
477:
478: \bibitem[\protect\astroncite{{Heindl} et~al.}{2004}]{heindl04a}
479: {Heindl}, W.~A., {Rothschild}, R.~E., {Coburn}, W., {Staubert}, R., {Wilms},
480: J., {Kreykenbohm}, I., \& {Kretschmar}, P., 2004,
481: \newblock in AIP Conf. Proc. 714: X-ray Timing 2003: Rossi and Beyond, 323
482:
483: \bibitem[\protect\astroncite{Isenberg, Lamb \& Wang}{1998}]{isenberg98a}
484: Isenberg, M., Lamb, D.~Q., \& Wang, J. C.~L., 1998, ApJ, 493, 154
485:
486: \bibitem[\protect\astroncite{Jahoda et~al.}{2005}]{jahoda:05a}
487: Jahoda, K., Markwardt, C.~B., Radeva, Y., Rots, A.~H., Stark, M.~J., Swank,
488: J.~H., Strohmayer, T.~E., \& Zhang, W., 2005, ApJS,
489: \newblock submitted
490:
491: \bibitem[\protect\astroncite{{Kreykenbohm} et~al.}{2002}]{kreykenbohm02a}
492: {Kreykenbohm}, I., {Coburn}, W., {Wilms}, J., {Kretschmar}, P., {Staubert}, R.,
493: {Heindl}, W.~A., \& {Rothschild}, R.~E., 2002, A\&A, 395, 129
494:
495: \bibitem[\protect\astroncite{Kreykenbohm et~al.}{1999}]{kreykenbohm99a}
496: Kreykenbohm, I., Kretschmar, P., Wilms, J., Staubert, R., Kendziorra, E.,
497: Gruber, D.~E., Heindl, W.~A., \& Rothschild, R.~E., 1999, A\&A, 341, 141
498:
499: \bibitem[\protect\astroncite{{Kreykenbohm} et~al.}{2005}]{kreykenbohm05a}
500: {Kreykenbohm}, I., et~al., 2005, A\&A, 433, L45
501:
502: \bibitem[\protect\astroncite{{Kreykenbohm} et~al.}{2004}]{kreykenbohm04a}
503: {Kreykenbohm}, I., {Wilms}, J., {Coburn}, W., {Kuster}, M., {Rothschild},
504: R.~E., {Heindl}, W.~A., {Kretschmar}, P., \& {Staubert}, R., 2004, A\&A,
505: 427, 975
506:
507: \bibitem[\protect\astroncite{{La Barbera} et~al.}{2003}]{barbera03a}
508: {La Barbera}, A., {Santangelo}, A., {Orlandini}, M., \& {Segreto}, A., 2003,
509: A\&A, 400, 993
510:
511: \bibitem[\protect\astroncite{{Makino}}{1989}]{makino89a}
512: {Makino}, F., 1989,
513: \newblock {IAU Circ.\ 4872}
514:
515: \bibitem[\protect\astroncite{Makishima et~al.}{1990a}]{makishima90b}
516: Makishima, K., et~al., 1990a, ApJ, 365, L59
517:
518: \bibitem[\protect\astroncite{{Makishima} et~al.}{1990b}]{makishima90a}
519: {Makishima}, K., et~al., 1990b, PASJ, 42, 295
520:
521: \bibitem[\protect\astroncite{M\'esz\'aros}{1992}]{meszaros92a}
522: M\'esz\'aros, P., 1992, High-Energy Radiation from Magnetized Neutron
523: Stars, Chicago \& London: Univ.\ Chicago Press
524:
525: \bibitem[\protect\astroncite{Mihara}{1995}]{mihara95a}
526: Mihara, T., 1995,
527: \newblock Ph.D. thesis, RIKEN, Tokio
528:
529: \bibitem[\protect\astroncite{{Mihara}, {Makishima} \&
530: {Nagase}}{2004}]{mihara04a}
531: {Mihara}, T., {Makishima}, K., \& {Nagase}, F., 2004, ApJ, 610, 390
532:
533: \bibitem[\protect\astroncite{{Negueruela} et~al.}{1999}]{negueruela99a}
534: {Negueruela}, I., {Roche}, P., {Fabregat}, J., \& {Coe}, M.~J., 1999, MNRAS,
535: 307, 695
536:
537: \bibitem[\protect\astroncite{{Protassov} et~al.}{2002}]{protassov02a}
538: {Protassov}, R., {van Dyk}, D.~A., {Connors}, A., {Kashyap}, V.~L., \&
539: {Siemiginowska}, A., 2002, ApJ, 571, 545
540:
541: \bibitem[\protect\astroncite{{Remillard}}{2004}]{remillard04a}
542: {Remillard}, R., 2004,
543: \newblock The Astronomer's Telegram 371
544:
545: \bibitem[\protect\astroncite{Rothschild et~al.}{1998}]{rothschild:98a}
546: Rothschild, R.~E., et~al., 1998, ApJ, 496, 538
547:
548: \bibitem[\protect\astroncite{{Stella} et~al.}{1985}]{stella85a}
549: {Stella}, L., {White}, N.~E., {Davelaar}, J., {Parmar}, A.~N., {Blissett},
550: R.~J., \& {van der Klis}, M., 1985, ApJ, 288, L45
551:
552: \bibitem[\protect\astroncite{{Tanaka}}{1983}]{tanaka83a}
553: {Tanaka}, Y., 1983,
554: \newblock {IAU Circ.\ 3891}
555:
556: \bibitem[\protect\astroncite{Tanaka}{1986}]{tanaka86a}
557: Tanaka, Y., 1986,
558: \newblock in Radiation Hydrodynamics in Stars and Compact Objects, ed. D.
559: Mihalas, K.-H.~A. Winkler, (Heidelberg: Springer), 198
560:
561: \bibitem[\protect\astroncite{{Terrell} \& {Priedhorsky}}{1984}]{terrell84a}
562: {Terrell}, J., \& {Priedhorsky}, W.~C., 1984, ApJ, 285, L15
563:
564: \bibitem[\protect\astroncite{{Terrell} et~al.}{1983}]{terrell83a}
565: {Terrell}, J., et~al., 1983,
566: \newblock {IAU Circ.\ 3893}
567:
568: \bibitem[\protect\astroncite{{Unger} et~al.}{1992}]{unger92a}
569: {Unger}, S.~J., {Norton}, A.~J., {Coe}, M.~J., \& {Lehto}, H.~J., 1992, MNRAS,
570: 256, 725
571:
572: \end{thebibliography}
573:
574:
575: \end{document}
576:
577:
578:
579:
580:
581:
582: