1: \documentclass[12pt,preprint]{aastex}
2: %\documentclass[manuscript]{aastex}
3:
4: \shorttitle{Magnetic fields in M dwarfs}
5: \shortauthors{Phan-Bao et al.}
6:
7: \begin{document}
8:
9: \title{Magnetic fields in M dwarfs: rapid magnetic field variability in EV Lac}
10:
11: \author{Ngoc Phan-Bao,\altaffilmark{1,2}
12: Eduardo L. Mart\'{\i}n,\altaffilmark{3,2}
13: Jean-Fran\c{c}ois Donati,\altaffilmark{4}
14: and Jeremy Lim\altaffilmark{1}}
15:
16: \altaffiltext{1}{Institute of Astronomy and Astrophysics, Academia Sinica, P.O. Box 23-141, Taipei 106, Taiwan, ROC; pbngoc@asiaa.sinica.edu.tw; jlim@asiaa.sinica.edu.tw.}
17: \altaffiltext{2}{present address (PBN):University of Central Florida, Dept. of Physics, PO Box 162385, Orlando, FL 32816-2385, USA; pngoc@physics.ucf.edu.}
18: \altaffiltext{3}{Instituto de Astrof\'{\i}sica de Canarias, C/ V\'{\i}a L\'actea s/n, E-38200 La Laguna (Tenerife), Spain; ege@iac.es.}
19: \altaffiltext{4}{Laboratoire d'Astrophysique, Observatoire Midi-Pyr\'en\'ees, F-31400 Toulouse, France; donati@ast.obs-mip.fr.}
20:
21: \begin{abstract}
22: We report here our spectropolarimetric observations obtained
23: using the Espadons/CFHT high resolution spectrograph of two M dwarf
24: stars which standard models suggest are fully convective: EV~Lac (M3.5) and HH~And (M5.5).
25: The difference in their rotational velocity makes them good targets to study
26: the dependence of the magnetic field topology in M dwarfs on rotation.
27: Our results reveal some aspects of the field topology in EV~Lac and HH~And.
28: We measured mean longitudinal magnetic field strengths ($B_{\rm z}$) in EV~Lac
29: ranging from 18$\pm$3~G to $-$40$\pm$3~G. The $B_{\rm z}$ variations are seen to
30: occur in a timescale of less than 50 minutes, significantly shorter than the rotation period,
31: and are not due to a flaring event.
32: We discuss some formation scenarios of the Zeeman signatures found in EV~Lac.
33: For HH~And we could not detect circular polarization and thus we place an upper limit
34: to $B_{\rm z}$ of 5~G.
35:
36: \end{abstract}
37:
38: \keywords{stars: low mass --- stars: magnetic fields --- techniques: spectroscopic --- techniques: polarimetric
39: --- stars: individual (EV~Lac, HH~And)}
40:
41: \section{Introduction}
42:
43: A large number of M dwarfs have been detected in the last few
44: decades, and our knowlegde of these faint stars has kept improving, but there are still many
45: challenges for research on their magnetic activity.
46: Main-sequence stars are expected to be fully convective if their mass lies below a certain
47: value, about 0.3-0.4 $M_{\odot}$ (M3-M4 spectral types) as suggested by the stellar evolution theory;
48: this probably shifts toward lower masses due to the influence of the magnetic field \citep{cox,mullan}.
49: In fully convective stars, their magnetic field may be produced by dynamo processes \citep{durney}
50: that differ from the shell-dynamo at work in partly-convective Sun-like stars.
51: However the nature of these dynamos is not clear \citep{liebert}.
52: Additionally, recent models \citep{dobler,chabrier} disagree with the observations
53: as pointed out in \citet{donati06}.
54:
55: One way to understand the nature of magnetic activity in M dwarfs, especially in fully-convective
56: stars, is to directly measure magnetic fields.
57: In the two last decades, magnetic fields in late type main sequence stars have been measured for:
58: G, K dwarf stars \citep{robinson80,marcy84,saar86,basri88,valenti95,ruedi}
59: or M dwarfs \citep[and references therein]{saar85,johns-krull96,saar01,reiners}.
60: The basic idea of the measurement method employed by these authors is a comparison of differential
61: Zeeman broadening between magnetically sensitive
62: and insensitive spectral lines, through unpolarised high-resolution spectra.
63: These line broadening measurements have indicated that in some active M dwarfs the
64: stellar surface is covered with magnetic fields of 3-4~kG and the filling
65: factor of about 50\%; however they are poorly informative on the field topology.
66: One should note that an early attempt to search for circular polarisation in M dwarfs
67: was carried out by \citet{vogt} but there was no clear detection.
68: In this paper, we report our analysis of polarised high-resolution spectra of 2 M-dwarf stars:
69: a rapidly rotating M3.5 (EV~Lac) and a slowly rotating M5.5 (HH~And).
70: Based on their spectropolarimetry presented in Sec.~2, we discuss some interesting aspects on
71: the complex field topology in EV~Lac and HH~And in Sec.~3.
72: %
73: \section{Targets, spectropolarimetric observation and data reduction}
74: %
75: \subsection{Targets}
76: \subsubsection{EV Lac (Gl 873)}
77: %
78: The star is a fast rotator for an M dwarf: $vsini = 4.5$~km/s \citep{johns-krull96},
79: with this rotational velocity EV~Lac is expected to have a strong magnetic field.
80: This is therefore a good target to study the dependence of magnetic
81: dynamo on rapidly rotating M stars.
82: EV~Lac is also well-known as an M3.5 flare star \citep[and references therein]{osten};
83: the strong $H_{\alpha}$ emission observed \citep{stauffer} indicates a high chromospheric
84: activity level in EV~Lac. The rotational period of the star is about 4 days, derived from its rotational velocity
85: and a corresponding radius of an M3.5 dwarf ($R \sim 0.36R_{\odot}$, \citealt{chabrier97, favata}).
86: \citet{pettersen92} have also well determined a photometric period of 4.4~days for EV~Lac.
87: The mean field measurements
88: using synthetic spectrum fitting have previously been reported:
89: $|B|$ = 4.3 kG, $f = 85$\%
90: in \citet{saar94b}; and
91: $|B|$ = 3.8 kG, $f=50$\% in \citet{johns-krull96}.
92: However, no measurements of its longitudinal magnetic field $B_{\rm z}$
93: have been reported so far.
94: %
95: \subsubsection{HH And (Gl 905)}
96: The M5.5 dwarf star in contrast with EV~Lac is a slow rotator: $vsini < 1.2$~km/s \citep{delfosse98}.
97: This is therefore a good target to study the dependence of magnetic dynamo on slowly rotating M stars.
98: No $H_{\alpha}$ emission detected \citep{stauffer} indicates a very low chromospheric activity level
99: in HH~And.
100: Its rotational period is longer than 7 days, derived from an upper limit for $vsini$ as given above
101: and an M5.5 dwarf radius ($R \sim 0.17R_{\odot}$, \citealt{chabrier97}).
102: There are no measurements of its magnetic field up to now.
103: The X-ray observations \citep{schmitt} of nearby stars have implied
104: that the coronal activity level in EV~Lac is from 10 to 100 times higher than in HH~And.
105: %
106: \subsection{Spectropolarimetric observation and data reduction}
107: %
108: We observed EV~Lac and HH~And with the Espadons/CFHT high resolution
109: spectrograph ($R = 65,000$; \citealt{donati03}) providing a wavelength coverage of 370-1,000~nm,
110: in spectropolarimetric mode to measure Stokes $I$ and $V$ parameters.
111: For each star we took three successive exposures, 50 minutes of each
112: exposure for EV~Lac and 40 minutes for HH~And.
113: We obtained high signal-to-noise spectra
114: of about 270:1 and 150:1 for EV~Lac and HH~And, respectively.
115: Wavelength calibrated unpolarised and polarised spectra corresponding to each
116: observing sequences are extracted with the dedicated software package Libre-ESpRIT
117: (\citealt{donati97}; Donati et al., in preparation)
118: following the principles of optimal extraction \citet{horne}.
119:
120: To compute the mean longitudinal magnetic field of the stars from the photospheric lines, we use
121: the Least-Squares Deconvolution (LSD) multi-line analysis procedure\footnote{
122: For this work, we used a line list \citep{kurucz} corresponding to M spectral types
123: matching that of EV~Lac and HH~And.
124: About 5,000 intermediate to strong atomic spectral lines with the average Land\'e factor $g_{\rm eff}$
125: of about 1.2 are used simultaneously to retrieve the average polarisation information in line profiles, with typical
126: noise levels of $\approx$0.06\% (relative to the unpolarised continuum level) per
127: 1.8~km/s velocity bin and per individual polarisation spectrum, corresponding
128: to a multiplex gain in S/N of about 10 with respect to a single average line analysis.
129: }
130: given in \citet{donati97}, producing mean Stokes $I$ (unpolarised) and $V$ (circularly polarised)
131: profiles for all collected spectra.
132: Figures 1 and 2 show our mean Stokes $V$ profiles of EV~Lac and HH~And.
133: In the case of EV~Lac,
134: mean longitudinal field strengths corresponding to three
135: successive exposures are:
136: $B_{\rm z}$ = 18$\pm$3~G, $-$40$\pm$3~G and
137: $-$37$\pm$3~G, respectively.
138: It is very interesting to note that using the formula given in \citet{wade}
139: the mean longitudinal field strengths estimated from the
140: chromospheric lines are much stronger than from the photospheric lines;
141: and they are different between the Balmer lines ($B_{\rm z} = 260\pm10$~G) and the Ca~II infra-red triplet (IRT)
142: ($B_{\rm z} = 150\pm20$~G, an average over the three triplet lines). We defer a discussion of this result to Sec.~3.
143:
144: For HH~And, its Stokes $V$ profiles from our observations have not revealed a significant magnetic field,
145: $B_{\rm z} = -5\pm2$~G. We have therefore set up an upper limit of 5~G on $B_{\rm z}$ for the star.
146: %
147: \section{Discussion}
148: %
149: In this paper, we could not detect the circularly polarised signature in HH~And
150: but we did detect it in EV~Lac.
151: The explanation might be a difference in the kind of magnetic dynamo that
152: dominates in each star.
153: Since EV~Lac (M3.5) and HH~And (M5.5) have low enough masses,
154: $\sim 0.35 M_{\odot}$ for EV~Lac and $\sim 0.15 M_{\odot}$ for HH~And \citep{delfosse00},
155: and they are therefore expected to be
156: fully convective \citep{chabrier97}. If this is the case, a turbulent
157: dynamo \citep{durney} may dominate in HH~And and generates a small-scale structure
158: that is not accessible to us.
159: In the case of EV~Lac, the magnetic field is probably generated from
160: both a turbulent dynamo and another dynamo kind.
161: The latter may only work in rapidly rotating stars
162: and produces a large-scale structure
163: whose magnetic fields are able to be detected,
164: e.g. EV~Lac (this paper) or V374~Peg \citep{donati06}.
165:
166: Figures 3-6 show the Stokes $V$ and $I$ profiles of the H$_{\alpha}$ at 656.28~nm, and the Ca~II~IRT
167: (at 849.802, 854.209, and 866.214~nm) respectively for the last exposure.
168: The difference in the Stokes $V$ profile between the H$_{\alpha}$
169: and the Ca~II~IRT reflects: (1) an imhomogeneous and complex field topology in EV~Lac;
170: (2) the different shapes of the emission cores ($V \sim \delta I/\delta \lambda$);
171: (3) the difference in the effective Land\'e factor and the $\sigma$ component distribution
172: of the lines; (4) the difference in line formation heights of the H$_{\alpha}$ and the Ca~II~IRT,
173: and the field strengths at those heights \citep{vernazza}.
174:
175: During our observation, the Balmer lines as well as the emission cores
176: of the Ca~II~IRT in EV~Lac indicate chromospheric (and therefore magnetic) activity.
177: This supports the idea that the Zeeman signatures of the Balmer lines and the Ca~II~IRT have
178: a chromospheric origin.
179: On the other hand, the longitudinal magnetic field strength computed
180: from the Balmer lines ($B_{\rm z} = 260$~G) is much stronger
181: than that from the Ca~II~IRT ($B_{\rm z} = 150$~G),
182: implying that the Zeeman signatures of the Balmer lines and the Ca~II~IRT might be
183: formed at different heights \citep{vernazza}.
184:
185: Our observations have also indicated a significant variability of the field in EV~Lac:
186: an opposite sign of $B_{\rm z}$ observed in the first exposure compared with the second and the third one.
187: The time scale of this variability is very short, i.e. less than 50 minutes.
188: We consider several possibilities that could explain the variability.
189: First, the possibility of star rotation changing the field vector is ruled out
190: since in the case of EV~Lac its rotational period of 4 days could not significantly change
191: the field vector
192: in 50 minutes or $\sim$0.8\% of a whole phase.
193: %
194: Second, the possibility of a strong flare during our observations,
195: modifying the Zeeman signature obtained in the first exposure,
196: is also precluded since we do not find
197: any significant change in the H$_{\alpha}$ profile whose equivalent width is of about
198: 3.3~\AA~for all three exposures.
199: %
200: Finally, with high probability the magnetic field in EV~Lac is intrinsically variable.
201: This again indicates the imhomogeneous and complex field in the star
202: that probably leads observational results as discussed above.
203:
204: More spectropolarimetric observations at different rotational phases
205: will allow us to reconstruct the EV~Lac field topology
206: not only on the stellar surface but also in three dimensions by using different lines coming
207: from different layers of the star.
208: This will lead to a strong improvement of our understanding on magnetic field
209: in EV~Lac and hence in low mass stars.
210:
211: \acknowledgments
212: We thank the referee for many helpful comments that clarified and significantly improved our paper.
213: P.B.N. acknowledges useful discussions with J. Valenti and G. Chabrier.
214: This paper is based on observations obtained at the Canada-France-Hawaii Telescope (CFHT)
215: which is operated by the National Research Council of Canada, the Institute National des Sciences
216: de l'Univers of the Centre National de la Recherche Scientifique of France, and the University of Hawaii.
217:
218: \begin{thebibliography}{}
219:
220: \bibitem[Basri \& Marcy(1988)]{basri88}
221: Basri, G., \& Marcy, G. W. 1988, \apj, 330, 274
222:
223: \bibitem[Chabrier \& Baraffe(1997)]{chabrier97}
224: Chabrier, G., \& Baraffe, I. 1997, \aap, 327, 1039
225:
226: \bibitem[Chabrier \& K\"uker(2006)]{chabrier}
227: Chabrier, G., \& K\"uker, M. 2006, \aap, 446, 1027
228:
229: \bibitem[Cox et al.(1981)]{cox}
230: Cox, A. N., Shaviv, G., \& Hodson, S. W. 1981, \apj, 245, L37
231:
232: \bibitem[Delfosse et al.(1998)]{delfosse98}
233: Delfosse, X., Forveille, T., Perrier, C., \& Mayor, M. 1998, \aap, 331, 581
234:
235: \bibitem[Delfosse et al.(2000)]{delfosse00}
236: Delfosse, X., Forveille, T., S\'egransan, D., et al. 2000, \aap, 364, 217
237:
238: \bibitem[Dobler et al.(2006)]{dobler}
239: Dobler, W., Stix, M., \& Brandenburg, A. 2006, \apj, 638, 336
240:
241: \bibitem[Donati et al.(1997)]{donati97}
242: Donati, J.-F., Semel. M., Carter, B. D., et al. 1997, \mnras, 291, 658
243:
244: \bibitem[Donati(2003)]{donati03}
245: Donati, J.-F. 2003, ASP Conference Proceedings, 307, 41
246:
247: \bibitem[Donati et al.(2006)]{donati06}
248: Donati, J.-F., Forveille, T., Cameron, A. C., et al. 2006, Science, 311, 633
249:
250: \bibitem[Durney et al.(1993)]{durney}
251: Durney, B. R., De Young, D. S., \& Roxburgh, I. W. 1993, Sol. Phys., 145, 207
252:
253: \bibitem[Favata et al.(2000)]{favata}
254: Favata, F., Reale, F., Micela, G., Sciortino, S., et al. 2000, \aap, 353, 987
255:
256: \bibitem[Horne(1986)]{horne}
257: Horne, K. 1986, PASP, 98, 609
258:
259: \bibitem[Johns-Krull \& Valenti(1996)]{johns-krull96}
260: Johns-Krull, C. M., \& Valenti, J. A. 1996, \apj, 459, L95
261:
262: \bibitem[Kurucz(1993)]{kurucz}
263: Kurucz, R. L. 1993, CDROM \# 13 (ATLAS9 atmospheric models) and \# 18 (ATLAS9 and SYNTHE routines,
264: spectral line database).
265:
266: \bibitem[Liebert et al.(2003)]{liebert}
267: Liebert, J., Kirkpatrick, J. D., Cruz, K. L., et al. 2003, \aj, 125, 343
268:
269: \bibitem[Marcy(1984)]{marcy84}
270: Marcy, G. W. 1984, \apj, 276, 286
271:
272: \bibitem[Mullan \& MacDonald(2001)]{mullan}
273: Mullan, D. J., \& MacDonald, J. 2001, \apj, 559, 353
274:
275: \bibitem[Osten et al.(2005)]{osten}
276: Osten, R. A., Hawley, S. L., Allred, J. C., et al. 2005, \apj, 621, 398
277:
278: \bibitem[Pettersen et al.(1992)]{pettersen92}
279: Pettersen, B. R., Olah, K., \& Sandmann, W. H. 1992, \aaps, 96, 497
280:
281: \bibitem[Reiners \& Basri(2006)]{reiners}
282: Reiners, A., \& Basri, G. 2006, arXiv: astro-ph/0602221
283:
284: \bibitem[Robinson et al.(1980)]{robinson80}
285: Robinson, R. D., Worden, S. P., \& Harvey, J. W. 1980, \apj, 236, L155
286:
287: \bibitem[R\"uedi et al.(1997)]{ruedi}
288: R\"uedi, I., Solanki, S. K., Mathys, G., \& Saar, S. H. 1997, \aap, 318, 429
289:
290: \bibitem[Saar \& Linsky(1985)]{saar85}
291: Saar, S. H., \& Linsky, J. L. 1985, \apj, 299, L47
292:
293: \bibitem[Saar et al.(1986)]{saar86}
294: Saar, S. H., Linsky, J. L., \& Beckers, J. M. 1986, \apj, 302, 777
295:
296: \bibitem[Saar(1994)]{saar94b}
297: Saar, S. H. 1994, IAU Symp. Infrared Solar Physics, 154, 493
298:
299: \bibitem[Saar(2001)]{saar01}
300: Saar, S. H. 2001, ASP Conference Proceedings, 223, 292
301:
302: \bibitem[Schmitt \& Liefke(2004)]{schmitt}
303: Schmitt, J.H.M.M, \& Liefke, C. 2004, \aap, 417, 651
304:
305: \bibitem[Stauffer \& Hartmann(1986)]{stauffer}
306: Stauffer, J. R., \& Hartmann, L. W. 1986, \apjs, 61, 531
307:
308: \bibitem[Valenti et al.(1995)]{valenti95}
309: Valenti, J. A., Marcy, G. W., \& Basri, G. 1995, \apj, 439, 939
310:
311: \bibitem[Vernazza et al.(1981)]{vernazza}
312: Vernazza, J. E., Avrett, E. H., \& Loeser, R. 1981, \apjs, 45, 635
313:
314: \bibitem[Vogt(1980)]{vogt}
315: Vogt, S. S. 1980, \apj, 240, 567
316:
317: \bibitem[Wade et al.(2000)]{wade}
318: Wade, G. A., Donati, J.-F., Landstreet, J. D., \& Shorlin, S. L. S. 2000, MNRAS, 313, 851
319:
320: \end{thebibliography}
321:
322: \clearpage
323:
324: \begin{figure}
325: \vskip 1in
326: \hskip -0.25in
327: \centerline{\includegraphics[width=4.5in,angle=-90]{f1.ps}}
328: \caption{Stokes $V$ profiles of EV Lac (Gl 873, M4.5); solid line: first
329: exposure, dashed line: second exposure, and dash-dotted line: third exposure; 50 minutes
330: for each exposure were taken. A positive blueward component (solid line) implies
331: a mixed polarity in the first exposure.
332: \label{StokesV_EVLac}}
333: \end{figure}
334:
335: \clearpage
336:
337: \begin{figure}
338: \vskip 1in
339: \centerline{\includegraphics[width=4.5in,angle=-90]{f2.ps}}
340: \caption{Stokes $V$ profiles of HH And (Gl 905, M5); three exposures were also
341: taken, 40 minutes for each one.
342: No longitudinal magnetic field has been detected, we set up an upper limit of 5~G on $B_{\rm z}$.
343: \label{StokesV_EVLac}}
344: \end{figure}
345:
346: \clearpage
347:
348: \begin{figure}
349: \centerline{\includegraphics[width=4.5in,angle=-90]{f3.ps}}
350: \caption{Stokes $V$ (top) and $I$ (bottom) profile (exposure \#3) of the H$_{\alpha}$ line at 656.28~nm.
351: Additional components in the Stokes $V$ profile (a positive component
352: at 656.19~nm and a negative one at 656.38~nm) implies a mixed polarity.
353: \label{Halpha}}
354: \end{figure}
355:
356: \clearpage
357:
358: \begin{figure}
359: \centerline{\includegraphics[width=4.5in,angle=-90]{f4.ps}}
360: \caption{Stokes $V$ (top) and $I$ (bottom) profile (exposure \#3) of the
361: Ca~II IRT at 849.802~nm.
362: \label{CaIIIRT1}}
363: \end{figure}
364:
365: \clearpage
366:
367: \begin{figure}
368: \centerline{\includegraphics[width=4.5in,angle=-90]{f5.ps}}
369: \caption{Stokes $V$ (top) and $I$ (bottom) profile (exposure \#3) of the Ca~II IRT at
370: 854.209~nm.
371: \label{CaIIIRT2}}
372: \end{figure}
373:
374: \clearpage
375:
376: \begin{figure}
377: \centerline{\includegraphics[width=4.5in,angle=-90]{f6.ps}}
378: \caption{Stokes $V$ (top) and $I$ (bottom) profile (exposure \#3) of the Ca~II IRT at
379: 866.214~nm.
380: \label{CaIIIRT3}}
381: \end{figure}
382:
383:
384:
385: \end{document}
386:
387:
388: