1: %\documentclass[12pt,preprint2]{aastex}
2: %\documentclass[12pt,preprint]{aastex}
3: %\documentclass[manuscript]{aastex}
4:
5: \documentclass{emulateapj}
6: \usepackage{apjfonts}
7:
8: \usepackage{graphicx}
9:
10:
11: % Lines & atoms
12: \newcommand{\La}{\mbox{${\rm Ly\alpha}$}}
13: \newcommand{\Line}[3]{\Ion{#1}{#2}\,$\lambda$\,#3}
14: \newcommand{\Lines}[3]{\Ion{#1}{#2}\,$\lambda\lambda$\,#3}
15: \newcommand{\Ion}[2]{#1{\,\scriptsize #2}}
16: %
17: % Stellar properties
18: \newcommand{\Mwd}{\mbox{$M_{\rm wd}$}}
19: \newcommand{\Rsec}{\mbox{$R_{\rm sec}$}}
20: \newcommand{\Msec}{\mbox{$M_{\rm sec}$}}
21: \newcommand{\Teff}{\mbox{$T_{\rm eff}$}}
22: \newcommand{\Porb}{\mbox{$P_{\rm orb}$}}
23: %
24:
25:
26: \begin{document}
27:
28: \title{A 150\,MG magnetic white dwarf in the cataclysmic variable
29: RX\,J1554.2+2721\altaffilmark{1}}
30:
31: \author{
32: B.T. G\"ansicke\altaffilmark{2},
33: S. Jordan\altaffilmark{3},
34: K. Beuermann\altaffilmark{4},
35: D. de Martino\altaffilmark{5}, \\
36: P. Szkody\altaffilmark{6},
37: T. Marsh\altaffilmark{2},
38: J. Thorstensen\altaffilmark{7}}
39:
40: \altaffiltext{1}{Based on observations made with the NASA/ESA Hubble Space
41: Telescope, obtained at the Space Telescope Science Institute, which is
42: operated by the Association of Universities for Research in Astronomy,
43: Inc., under NASA contract NAS 5-26555.}
44: %
45: \altaffiltext{2}{Department of Physics,University of Warwick, Coventry
46: CV4 7AL, UK, boris.gaensicke@warwick.ac.uk}
47: %
48: \altaffiltext{3}{Institut f\"ur Astronomie und Astrophysik, Universit\"at
49: T\"ubingen, Sand 1, 72076 T\"ubingen, Germany, jordan@ari.uni-heidelberg.de}
50: %
51: \altaffiltext{4}{Universit\"ats-Sternwarte, Geismarlandstr. 11, 37083
52: G\"ottingen, Germany, beuermann@uni-sw.gwdg.de}
53: %
54: \altaffiltext{6}{Osservatorio di Capodimonte, Via Moiariello 16,
55: I-80131 Napoli, Italy, demartin@na.astro.it}
56: %
57: \altaffiltext{6}{Astronomy Department, University of Washington,
58: Seattle, WA 98195, USA, szkody@astro.washington.edu}
59: %
60: \altaffiltext{7}{Department of Physics and Astronomy, Dartmouth
61: College, 6127 Wilder Laboratory, Hanover, NH 03755-3528, USA,
62: j.thorstensen@dartmouth.edu}
63:
64: \begin{abstract}
65: We report the detection of Zeeman-split \La\ absorption $\pi$ and
66: $\sigma^+$ lines in the far-ultraviolet \textit{Hubble Space
67: Telescope}/Space Telescope Imaging Spectrograph spectrum of the
68: magnetic cataclysmic variable RX\,J1554.2+2721. Fitting the STIS data
69: with magnetic white dwarf model spectra, we derive a field strength of
70: $B\simeq144$\,MG and an effective temperature of
71: $17\,000\,\mathrm{K}\la\Teff\la23\,000$\,K. This measurement makes
72: RX\,J1554.2+2721 only the third cataclysmic variable containing a
73: white dwarf with a field exceeding 100\,MG. Similar to the other
74: high-field polar AR\,UMa, RX\,J1554.2+2721 is often found in a state
75: of feeble mass transfer, which suggests that a considerable number
76: of high-field polars may still remain undiscovered.
77: \end{abstract}
78:
79:
80: \keywords{stars: individual (RX\,J1554.2+2721) --
81: novae, cataclysmic variables
82: line: formation --
83: stars: magnetic fields --
84: white dwarfs
85: }
86:
87: \section{Introduction}
88: In polars, a sub-class of cataclysmic variables (CV), the strong
89: magnetic field of the white dwarf suppresses the formation of an
90: accretion disk, and synchronizes the white dwarf spin with the orbital
91: period of the binary. Whereas single magnetic white dwarfs show a field
92: distribution ranging from a few tens kG to $\simeq1000$\,MG
93: \citep{wickramasinghe+ferrario00-1, jordan01-1, gaensickeetal02-5,
94: schmidtetal03-1} there has been (and still is) a discomforting lack of
95: CVs containing white dwarfs with magnetic fields in excess of 100\,MG.
96: The first high-field polar to be identified was AR\,UMa, where a field
97: strength of $B\simeq230$\,MG was estimated from the Zeeman-splitting
98: of the \La\ $\sigma^+$ component detected in a far-ultraviolet (FUV)
99: spectrum obtained with the \textit{IUE}
100: \citep{schmidtetal96-1}. Using detailed \textit{HST}/STIS data,
101: \citet{gaensickeetal01-1} refined the field strength measurement to
102: $B\simeq200$\,MG. \citet{schmidtetal01-1} reported a second high-field
103: polar, V884\,Her, and determined $B\simeq150$\,MG from matching a
104: number of narrow circularly polarized absorption features in the
105: optical spectrum with Zeeman-split Balmer line components.
106:
107: Here, we report the identification of a third high-field polar,
108: RX\,J1554.2+2721, henceforth RX\,J1554. RX\,J1554 was identified as a
109: CV on the basis of its emission line spectrum in the Hamburg Quasar
110: Survey \citep{jiangetal00-1}. Its magnetic nature was suggested by
111: \citet{tovmassianetal01-2}, who also provided a first estimate of the
112: orbital period, placing the system in the $2-3$\,h period
113: gap. \citet{thorstensen+fenton02-1} improved the period determination
114: to $\Porb=151.865\pm0.009$\,min. Based on the identification of broad
115: bumps in their optical spectra as cyclotron harmonics,
116: \citet{tovmassianetal01-2} tentatively suggested a field strength of
117: $B=31$\,MG for RX\,J1554.
118:
119: \section{FUV spectroscopy}
120: Medium-resolution FUV spectroscopy of RX\,J1554 was obtained with
121: \textit{HST}/STIS on February 27, 2003, as part of our ongoing
122: snapshot survey of cataclysmic variables \citep{gaensickeetal03-1}. We
123: used the G140L grating in conjunction with the
124: $52\arcsec\times0.2\arcsec$ aperture, providing a spectral resolution
125: of $\simeq1.2$\,\AA\ and a spectral coverage of $1150-1710$\,\AA. The
126: STIS spectrum of RX\,J1554 (Fig.\,\ref{f-stis}) contains a blue
127: continuum superimposed with narrow emission lines of
128: \Line{C}{III}{1176}, \Line{Si}{III}{1206}, \Line{N}{V}{1240},
129: \Line{C}{II}{1335}, \Lines{Si}{IV}{1393,1402}, \Line{C}{IV}{1550}, and
130: \Line{He}{II}{1640}. A weak broad emission bump centered on
131: $\sim1300$\,\AA\ may be related to the \Lines{Si}{III}{1294--1309}
132: multiplet, the \Lines{Si}{II}{1260,65} doublet is not detected. A most
133: unusual feature is the broad ($\sim75$\,\AA) absorption line centered
134: on $\sim1280$\,\AA. It falls clearly blue-ward of the 1400\,\AA\
135: quasi-molecular $H_2^+$ absorption observed in cool white
136: dwarfs. Considering the magnetic nature of RX\,J1554, a tantalising
137: possible identification for this feature is the $\sigma^+$ component
138: of \La\ Zeeman-split in a field $B>100$\,MG, analogous to that
139: observed in the FUV spectrum of AR\,UMa \citep{schmidtetal96-1,
140: gaensickeetal01-1}. We measured a F28$\times$50LP magnitude of
141: $\simeq16.8$ (roughly equivalent to the $R$ band) from the STIS CCD
142: acquisition image taken prior to the FUV spectroscopy. The system was
143: hence in a state of relatively low accretion activity
144: \citep{tovmassianetal01-2, thorstensen+fenton02-1}.
145:
146: \begin{figure}
147: \includegraphics[angle=-90,width=8cm]{f1.eps}
148: \caption{\label{f-stis}\textit{HST}/STIS G140L snapshot spectrum of
149: RX\,J1554.2+2721, along with emission line identifications.}
150: \end{figure}
151:
152: \section{Spectral Analysis}
153: We have analysed the STIS spectrum of RX\,J1554 using magnetic white
154: dwarf model spectra computed with the code of \citet{jordan92-1} and a
155: fitting routine very similar to the one described by
156: \citet{euchneretal02-1}. In brief, our fit involves three free
157: parameters: the magnetic field strength $B$ (where we restricted our
158: models to centered dipole fields), the angle between the line-of-sight
159: and the magnetic axis $\Psi$, and the effective temperature of the
160: white dwarf \Teff. Flux and circular polarization spectra are
161: calculated for a large number of surface elements representing the
162: surface of the white dwarf, and the actual white dwarf model spectrum
163: is obtained by adding up the contributions of all visible elements. In
164: contrast to \citet{euchneretal02-1}, we used 12 viewing angles
165: (instead of 9), and interpolated in a grid of 12 propagation angles
166: within the atmosphere to account for the effect of limb darkening
167: (instead of a linear approximation). The optimization is initially
168: done using a genetic algorithm \citep{carroll01-1} and subsequently
169: refined by a downhill simplex algorithm. Once the best-fit parameters
170: are determined, a final model is calculated by consistently solving
171: the radiative transfer equations \citep{jordan92-1} in order to avoid
172: the interpolations used in the fitting code.
173:
174: Whereas the position of the \La\ Zeeman components as well as their
175: profiles allow a relatively accurate measurement of $B$ and $\Psi$,
176: determining the effective temperature of high-field magnetic white
177: dwarfs is notoriously difficult. In non-magnetic white dwarfs, the
178: analysis of the Stark-broadened Balmer and/or Lyman lines is most
179: useful in establishing the effective temperature. However, the
180: combination of Stark broadening and Zeeman splitting at high field
181: strengths is still too poorly understood to use the hydrogen line
182: profiles as exact temperature indicators in strongly magnetic white
183: dwarfs. Similarly, also the slope of the continuum is subject to
184: uncertainties in the atomic physics. A general trend at high fields
185: appears to be that the optical continuum implies higher temperatures
186: than the FUV continuum, as observed e.g. in PG\,1031+234
187: \citep{schmidtetal86-1} and in AR\,UMa \citep{gaensickeetal01-1}.
188:
189: Our best-fit to the STIS spectrum of RX\,J1554 results in
190: $B=144\pm15$\,MG, $\Teff\sim17500$\,K and
191: $\Psi=38\pm15^{\circ}$. Thus, RX\,J1554 is the third CV containing a
192: magnetic white dwarf with a field exceeding 100\,MG. The field
193: strength determined here from the unambigous detection of the \La\
194: Zeeman $\sigma^+$ component exceeds the estimated $B=31$\,MG of
195: \citet{tovmassianetal01-2} by a large factor. The most likely cause
196: for this disagreement is an erroneous assignment of cyclotron harmonic
197: numbers to the weak emission bumps observed in the blue part of the
198: optical spectrum \citep{tovmassianetal01-2}. For a field strength of
199: $\sim150$\,MG, the emission of the cyclotron fundamental is centered
200: at $\sim7200$\,\AA, the first harmonic at $\sim3600$\,\AA. Neither the
201: spectra of \citet{tovmassianetal01-2} nor those of
202: \citet{thorstensen+fenton02-1} display any feature readily
203: identifiable as the cyclotron fundamental or first
204: harmonic. Phase-resolved spectropolarimetry would be desirable to
205: probe the nature of the optical continuum emission.
206:
207: \begin{figure}
208: \includegraphics[angle=-90,width=8cm]{f2.eps}
209: \caption{\label{f-stisfit}The best-fit magnetic white dwarf model
210: spectrum ($B=144$\,MG, $\Teff=17\,500$\,K) reproduces well the
211: position and profile of the observe \La\ $\pi$ and $\sigma^+$
212: components. }
213: \end{figure}
214:
215: Figure\,\ref{f-stisfit} shows the STIS spectrum along with
216: our best-fit model extended blue-wards up to the Lyman limit. It is
217: apparent that future phase-resolved \textit{FUSE} and \textit{HST}
218: spectroscopy of RX\,J1554 would be extremely useful to confine in more
219: detail the topology of the magnetic field in RX\,J1554, which could
220: deviate from a simple dipole geometry (as shown in the case of AR\,UMa
221: by \citealt{gaensickeetal01-1} and
222: \citealt{ferrarioetal03-1}). \textit{FUSE} observations would also
223: provide an additional constraint on the effective temperature of the
224: white dwarf.
225:
226: In Fig.\,\ref{f-sed} we show an average low-state spectrum of
227: RX\,J1554 from \citet{thorstensen+fenton02-1} along with our STIS
228: data. The red part of the optical spectrum clearly reveals the
229: secondary star, the blue continuum is most likely associated with
230: photospheric emission from the white dwarf. Following
231: \citet{thorstensen+fenton02-1}, the spectral contribution of the
232: secondary star is well described by a M4 template scaled for a
233: distance of $d=210$\,pc and a Roche-lobe equivalent radius of the
234: secondary of $R_{L_2}=1.76\times10^{10}$\,cm (corresponding to a mass
235: ratio $\Mwd/\Msec=3$). Whereas the relatively flat spectral slope of
236: the STIS spectrum is best described with a temperature of
237: $\simeq17\,500$\,K, the combined optical-FUV spectral energy
238: distribution suggests a higher temperature of $\sim23\,000$\,K
239: (reddening is negligible along the line of sight to RX\,J1554). At a
240: distance of $d=210$\,pc, the flux scaling factors implied by the
241: models shown in Fig.\,\ref{f-sed} give white dwarf radii of
242: $9.8\times10^8$\,cm, $7.9\times10^8$\,cm, and $6.6\times10^8$\,cm for
243: $\Teff=15\,000$\,K, 20\,000\,K, and 25\,000\,K~--~which are all
244: plausible, but do not help to constrain the temperature of the white
245: dwarf. We conclude that the temperature of the white dwarf in
246: RX\,J1554 is likely to be $17\,000\,\mathrm{K}\la\Teff\la23\,000$\,K.
247:
248: \section{Magnetic fields in white dwarfs}
249: The statistics of magnetic fields in single white dwarfs has been
250: substantially improved by the first results from the Sloan Digital Sky
251: survey \citep{gaensickeetal02-5, schmidtetal03-1}, even though
252: \citet{liebertetal03-1} argue that the fraction of magnetic white
253: dwarfs may still be significantly underestimated. Combining the SDSS
254: results with those summarized in \citet{wickramasinghe+ferrario00-1},
255: the number of single white dwarfs with known field strength is 108. Our
256: current census of polars is 76, of which some measure of the surface
257: field is available for a subset of 53. Figure\,\ref{f-fielddist} shows
258: the distribution of the measured (mean or typical) field strengths in
259: the single white dwarfs and in polars. For the latter, the quoted
260: field strength is preferentially that of the main accreting pole as
261: determined from the spacing of cyclotron lines, replaced by Zeeman
262: detections where no cyclotron measurement is available. It is likely
263: that this distribution is biased against the detection of both high
264: and low field strengths, because the high-field systems have weak
265: spectral features in the optical range and the low-field systems often
266: display a cyclotron continuum with all lines washed out. We suspect
267: that most of the 23 polars with no field measurement and all of the 40
268: confirmed or suspected intermediate polars in the 2003 edition of the
269: \citeauthor{ritter+kolb03-1} catalogue have field strengths typically in
270: the 1--25\,MG range, given that the minimum field strength which
271: allows a magnetic CV to synchronize is of the order of 3\,--\,10\,MG.
272:
273:
274: \section{Discussion}
275: The lack of high-field polars has been a long-standing puzzle.
276: \citet{hameuryetal89-3} suggested that for strong fields ($B>40$\,MG)
277: the wind of the secondary may couple to the magnetic field lines of
278: the white dwarf rather to those of the secondary, and, as a
279: consequence of the large magnetospheric radius of the white dwarf,
280: magnetic braking would be much more efficient than in non-magnetic
281: CVs~--~in fact so efficient that the mass transfer would be thermally
282: unstable, with a correspondingly very short life time of such systems.
283:
284: \begin{figure}
285: \includegraphics[angle=-90,width=8cm]{f3.eps}
286: \caption{\label{f-sed} Combined ultraviolet/optical spectrum of
287: RX\,J1554, using our STIS data and the mean low-state spectrum of
288: \citet{thorstensen+fenton02-1}. Overplotted as dashed lines are
289: (from top to bottom) magnetic white dwarf models ($B=144$\,MG,
290: $\Psi=28^{\circ}$ with \Teff\,=\,30\,000\,K, 25\,000\,K, 20\,000\,K,
291: and 15\,000\,K. The models have been scaled to the $V$-band flux of
292: RX\,J1554, after subtracting the contribution from the companion
293: star.}
294: \end{figure}
295:
296: \begin{figure}
297: \includegraphics[width=8cm]{f4.eps}
298: \caption{\label{f-fielddist} Distribution of magnetic field strength
299: for single white dwarfs and for white dwarfs in CVs.}
300: \end{figure}
301:
302: \citet{wickramasinghe+wu94-1} and \citet{lietal94-1} alternatively
303: argued that in high-field polars the magnetic field lines of the
304: secondary will form closed loops or connect to the field lines of the
305: white dwarf, reducing the magnetic flux in open field
306: lines. Therefore, their prediction is that the angular momentum loss
307: and mass transfer rates in strongly magnetic CVs should be lower
308: compared to non-magnetic ones, and they predict a maximum field
309: strength of $\sim70-100$\,MG, above which mass transfer will be
310: suppressed completely (see, however, the discussion by
311: \citealt{kingetal94-1}).
312:
313: Our determination of a white dwarf magnetic field strength of
314: $B\simeq150$\,MG in RX\,J1554 raises the number of high-field polars
315: ($B>100$\,MG) to three~--~which makes these objects a rare but no
316: longer exceptional species. AR\,UMa ($B\simeq200$\,MG,
317: $\Porb=116$\,min) and V884\,Her ($B\simeq150$\,MG, $\Porb=113$\,min)
318: are both below the period gap, and RX\,J1554 ($B\simeq145$\,MG,
319: $\Porb=152$\,min) is right in the middle of the gap~--~i.e. all known
320: high-field polars have periods below 3\,h. It seems, however,
321: premature to draw any conclusion based on three systems only. One
322: property that at least AR\,UMa and RX\,J1554 share is that they are
323: frequently encountered in states of low mass transfer
324: \citep{remillardetal94-2, schmidtetal96-1, schmidtetal99-1,
325: tovmassianetal01-2, thorstensen+fenton02-1}, while the long-term
326: variability of V884\,Her is less well documented. Assuming that
327: frequent low states are a general property of high-field polars it is
328: likely that a considerable fraction of these systems have evaded
329: discovery so far. The exact causes of low states in polars are
330: not yet fully understood, possible mechanisms that could decrease the
331: mass transfer rate on time scales of years include starspots moving
332: across the $L_1$ point \citep{king+cannizzo98-1, hessmanetal00-1} and
333: the swinging dipole model by \citet{andronov87-1}. Detailed long-term
334: monitoring of the accretion activity of polars in general, and
335: high-field polars in particular, would be desirable to establish a
336: measure of their accretion rates following the approach of
337: \citet{hessmanetal00-1}.
338:
339: \citet{schmidtetal03-1} believe that the maximum in the distribution
340: of single white dwarfs near 20\,MG is real. Taking into account the
341: selection effects for identifying magnetic CVs in the first place, and
342: measuring the magnetic fields of their white dwarfs in the second
343: place, it seems possible that the intrinsic distribution of magnetic
344: field strengths in CV white dwarfs is not dissimilar from the one of
345: single white dwarfs. RXJ\,1554 is one example of the systems not
346: readily recognisable as high-field polars by optical means, and
347: systems with white dwarfs of still higher field strength possibly may
348: exist but appear no different than ordinary polar.
349:
350: BTG was supported by a PPARC Advanced Fellowship. Additional support
351: was provided through NASA grant GO-9357 from the Space Telescope
352: Science Institute, which is operated by the Association of
353: Universities for Research in Astronomy, Inc., under NASA contract NAS
354: 5-26555.
355:
356: \bibliographystyle{apj}
357: \bibliography{aamnem99,aabib}
358:
359:
360:
361: \end{document}
362:
363: