1: \documentclass[12pt,preprint]{aastex}
2: %\documentclass[manuscript]{aastex}
3: %\usepackage{emulateapj5}
4: %\slugcomment{submitted to The Astrophysical Journal}
5: \shorttitle{Stellar Oscillations in $\beta$ Gem}
6: \shortauthors{Hatzes and Zechmeister}
7: \bibliographystyle{apj}
8: \begin{document}
9:
10: \title{The Discovery of Stellar Oscillations in the Planet Hosting Giant Star
11: $\beta$~Geminorum}
12:
13: \author{Artie P. Hatzes}
14: \and
15: \author{Mathias Zechmeister}
16: \affil{Th\"uringer Landessternwarte, D - 07778 Tautenburg, Germany}
17: \email{artie@tls-tautenburg.de}
18: \email{zechmeister@tls-tautenburg.de}
19:
20:
21: \begin{abstract}
22: We present the results of a long time series of precise stellar radial velocity measurements of
23: the planet hosting K giant star $\beta$~Geminorum. A total of 20 hours
24: of observations
25: spanning three nights were obtained and the radial velocity variations show the presence of solar-like
26: stellar oscillations. Our period analysis yields six
27: significant pulsation modes that have frequencies in the
28: range of 30 -- 150 $\mu$Hz. The dominant mode is at a frequency of 86.9 $\mu$Hz
29: and has an amplitude of 5.3 m\,s$^{-1}$. These values are consistent with
30: stellar oscillations for a giant star with a stellar mass of
31: $\approx$ 2 $M_\odot$.
32: This stellar mass implies a companion minimum mass of 2.6 $M_{Jupiter}$.
33: $\beta$~Gem is the first planet hosting giant star
34: in which multi-periodic stellar oscillations have been detected.
35: The study of stellar oscillations in planet hosting giant stars may provide
36: an independent, and more accurate determination of the stellar mass.
37: \end{abstract}
38:
39: \keywords{planetary systems --- techniques: radial velocities}
40:
41: \section{Introduction}
42: \label{intro}
43:
44: Radial Velocity (RV) surveys have discovered a number of giant planets in orbit
45: around giant stars (Frink et al. 2002; Sato et al. 2003, Setiawan et al. 2003,
46: Hatzes et al. 2005). These discoveries are important because the
47: evolved host stars often have masses in the range 1--3 $M_\odot$. Precise stellar
48: radial velocity measurements of
49: main sequence stars in this mass range are difficult due to the high effective
50: temperatures of the photosphere
51: and the paucity of stellar lines. Furthermore, these lines are often broadened by
52: high rates of stellar rotation. Consequently, most
53: RV surveys have focused primarily on stars later than about spectral type F6.
54: The study of planets around giant stars
55: can thus give us valuable clues as to the process or planet formation around
56: stars more massive than the sun, {\it if} one can determine an accurate stellar mass.
57: For giant stars this can be difficult.
58: Evolutionary tracks
59: of main sequence stars spanning a
60: wide range of masses all converge to the giant branch in
61: the color magnitude diagram. One has to rely on stellar
62: evolutionary tracks
63: which are model dependent, and these in turn rely on accurate determinations
64: of such stellar parameters as effective temperature, surface gravity,
65: abundance, and absolute luminosity.
66:
67: The mass determination of Arcturus offers us a good example.
68: Using spectral analysis M\"ackle et al. (1975) determined a stellar mass
69: of 0.1 -- 0.6 $M_\odot$. The analysis of Martin (1977) yielded a mass in the
70: range 0.6 -- 1.3 $M_\odot$. This is consistent with a later mass determination
71: of 0.95 $M_\odot$ by Bell, Edvardsson, and Gustafson (1985). These authors
72: noted that due to uncertainties in the surface gravity the
73: stellar mass could be as low as
74: 0.7 $M_\odot$. An accurate stellar mass is not only important for comparing
75: results to planet formation theories, but it is also required to calculate
76: the companion mass. Planets around intermediate mass
77: stars have masses in the range 3--10 $M_{Jupiter}$ and many lie on the deuterium
78: burning border which separates brown dwarfs from planets (approximately 13 $M_{Jupiter}$).
79: More accurate determinations of the stellar mass may establish if a
80: companion has a mass that is still consistent with a bona fide planet
81: or is certainly a brown dwarf, even if the orbital inclination were to be nearly
82: 90$^\circ$ (i.e sin $i$ = 90$^\circ$).
83:
84: One of the best ways of determining stellar mass, outside of dynamical methods, is
85: via asteroseismology. This is also the most accurate method
86: for determining the masses of isolated stars.
87: The stellar oscillations can be used to
88: derive such
89: fundamental parameters as the stellar mass, radius, age, and,
90: depending on the number of modes detected, the internal structure.
91: Asteroseismology has been used with spectacular success on white dwarf stars
92: using multi-site photometric campaigns
93: (e.g. Castanheira et al. 2004).
94: More recently, thanks to an increase in the precision of stellar RV
95: measurements, asteroseismology has been applied with some success
96: to solar-like stars (e.g. Bedding et al. 2006; Bazot et al. 2005).
97:
98: It is well established that many cool giants exhibit short period
99: RV or photometric
100: variations with periods ranging from hours (e.g. Hatzes \& Cochran 1994b;
101: Frandsen et al. 2002; de Ridder et al. 2006) to days
102: (e.g. Hatzes \& Cochran 1994a, Retter et al. 2003).
103: These periods are consistent with p-mode oscillations in giant stars.
104: Although many of these modes have not been identified with certainty, the observed
105: periods seem to be consistent with radial fundamental or overtone modes
106: (Hatzes \& Cochran 1994a),
107: although nonradial modes can still not be excluded.
108: The fact that extrasolar planets have been discovered around a class of
109: stars known to exhibit stellar oscillations opens up the exciting
110: possibility of using these stellar oscillations as an independent means
111: of deriving important properties of the planet host star.
112:
113: Long period variations with a period of 545 days were discovered in
114: $\beta$~Gem by Hatzes \& Cochran (1993). One proposed explanation was that
115: these were due to a planet with a minimum mass of 2.9 $M_{Jupiter}$,
116: assuming a stellar mass of 2.8 $M_\odot$.
117: Over a decade later Hatzes et al. (2006) and Reffert et al. (2006)
118: confirmed that these variations were in fact due to a planet in orbit with
119: a revised period of 590 days. Hatzes et al. (2006) used a more recent
120: stellar mass determination of 1.7 $M_\odot$ to derive a companion
121: mass of 2.3 $M_{Jupiter}$. However, given the difficulties of deriving
122: stellar masses from evolutionary tracks of giants, $\beta$~Gem
123: could easily have a much lower or higher mass. If one is interested in understanding the stellar mass
124: dependence of planet formation it is important to know if indeed $\beta$~Gem has a mass
125: in the intermediate ($\approx$ 2 $M_\odot$) range.
126:
127:
128: Here we present time series RV measurements for $\beta$~Gem taken on three
129: nights. The star shows RV variability consistent with stellar
130: oscillations. We then use these oscillation frequencies to estimate the stellar
131: mass.
132:
133: \section{Observations}
134:
135: A long time series of spectral observations were made of
136: $\beta$~Gem using the coude echelle spectrograph of the 2m Alfred Jensch
137: Telescope of the Th\"uringer Landessternwarte (Thuringia State Observatory).
138: Precise stellar radial velocity measurements were achieved by using an
139: iodine absorption cell placed in the optical light path to provide the
140: wavelength reference. A detailed description of the instrumental setup
141: and data reduction and analysis process can be found in Hatzes et al. (2005).
142:
143: Exposure times were 90 secs and with a CCD readout time of
144: 70 secs. The signal-to-noise ratio ($S/N$) of the observations depended on atmospheric
145: transparency and seeing conditions. The $S/N$ for our observations
146: ranged from about 100 to 300 per pixel.
147: A total of over 20 hours of observations were made
148: $\beta$~Gem spanning 3 different nights.
149: Table 1 gives the journal of observations which includes the
150: Julian day at the start of the time series, the length of the time
151: series, and the number of spectra that were obtained. The last column
152: is the nightly rms scatter of the
153: RV measurements about a mult-sine fit to the data (see below).
154:
155: Figure~\ref{f1} shows the time series of the RV measurements on
156: the 3 nights. It is evident from the figure that $\beta$~Gem shows
157: low amplitude, periodic variability on short time scales. It is also clear
158: that a single period cannot reproduce the observed variations.
159:
160: \section{Period Analysis}
161:
162: A period analysis was performed on the full data set using the program {\it Period04}
163: (Lenz \& Breger 2004).
164: This program offers convenient means of searching for multiple periods
165: in data via a pre-whitening procedure. A
166: sine wave fit is made to the data using the
167: dominant period found by Fourier analysis. This is subtracted and
168: additional periods are found in the residuals by further Fourier analysis.
169: Finally, {\it Period04} can be used to improve the solution by performing
170: a simultaneous least squares fit using a sum
171: of sine functions with the initial guess periods found in the
172: pre-whitening procedure.
173:
174: The Fourier analysis revealed a long period component corresponding to a period of
175: 3 days. Our data string is too short to establish if this is real or possibly an alias
176: of a shorter period. Since we are searching for periods comparable to, or shorter than our
177: nightly coverage this long-period component was subtracted at the start of the
178: period analysis.
179:
180:
181: Figure~\ref{f2} shows the power spectra derived using the discrete
182: Fourier transform at
183: each step of the pre-whitening procedure.
184: The top panel is for the
185: ``raw'' time series, and each subsequent lower panel is the power spectrum
186: after removal of the dominant frequency.
187:
188: The statistical significance of the periods were assessed using a
189: ``bootstrap randomization technique''.
190: The RV values was randomly shuffled
191: keeping the times fixed and a Scargle periodogram calculated
192: (Scargle 1982). A Scargle periodogram was used since the power is a measure
193: of the statistical significance of a signal. After a large
194: number of shuffles (200,000) the fraction of random periodograms
195: having Scargle power greater than the data periodogram gave an
196: estimate of the false alarm probability (FAP)
197: that a signal is due purely to noise.
198: Our bootstrap analysis indicates that a Scargle power, $P$ = 15 corresponds
199: to a FAP = 10$^{-5}$.
200: The pre-whitening procedure continued until
201: the Fourier analysis
202: found a period which we deemed not to be significant. For this analysis we
203: considered any signal having a FAP $>$ 10$^{-4}$ as not significant.
204: The dominant peak in each panel of
205: Figure~\ref{f2} had Scargle power greater than 15, with the exception of the
206: last ($f_7$) mode.
207:
208: Table 2 lists all the periods, frequencies, and corresponding amplitudes
209: for all the statistically significant periods found by our analysis. The modes are listed in the
210: order they were found by the pre-whitening procedure. The first six frequencies are highly significant
211: having FAP $<$ 5 $\times$ 10$^{-6}$. In other words after 2 $\times$ 10$^{5}$
212: shuffles of the bootstrap there was no instance when the power in the random periodograms
213: had higher power than the data periodograms. The least significant frequency is the one
214: at $\nu$ = 193 $\mu$Hz. The false alarm probability of this signal is 1.6 $\times$ 10$^{-3}$ which is
215: higher than our adopted threshold for significance.
216: The amplitude of this mode is significantly less than the mean measurement error
217: (1.5 -- 2 m\,s$^{-1})$. We thus regard this mode as uncertain and in need of confirmation
218: with data spanning a longer time.
219: The line in Fig.~\ref{f1} shows the multi-component sine fit to the full data set. The rms scatter about the
220: fit for the individual nights are 1.2, 1.5, and 1.7 m\,s$^{-1}$, respectively.
221:
222: \section{Stellar Mass Determination}
223:
224: To estimate the stellar mass we used
225: the scaling relations of Kjeldsen \& Bedding (1995) which they showed to
226: be valid
227: for stars covering a wide range of masses and luminosity classes.
228: In particular
229: we will use their expression for the
230: frequency of the maximum power:
231:
232: $$\nu_{max} = { {M/M_\odot} \over {(R/R_\odot)^2 \sqrt{T_{eff}/5777 K} }}
233: \hskip 5pt 3.05 \hskip 5pt \rm{mHz} $$
234:
235: We chose to use this expression rather than the one for the
236: frequency splitting because of the short time span of our observations.
237: We are not confident that we have detected all possible modes in $\beta$~Gem
238: which are necessary for determining a frequency spacing for high order
239: p-modes.
240: We take the 87.9 $\mu$Hz frequency as the dominant mode in the
241: data. This has the highest amplitude and is the most obvious peak in the periodogram
242: of the un-whitened data.
243:
244: The stellar radius of $\beta$~Gem has been measured with long baseline
245: interferometry. Nordgren, Sudol, \& Mozurkewich (2001) determined an
246: angular diameter of 7.96 $\pm$ 0.09 mas which corresponds to a radius
247: of 8.8 $\pm$ 0.1 $R_\odot$ using the Hipparcos distance of 96.74 $\pm$ 0.94
248: mas. McWilliam (1990) derived an effective temperature of
249: 4850 K. Using a value of $\nu_{max}$ = 0.0868 mHz results in an stellar mass, $M$ = 2.04 $\pm$ 0.04 $M_\odot$.
250: The orbital solution for the companion has a mass function, $f(m)$ =
251: (4.21 $\pm$ 0.48) $\times 10^{-9}$ $M_\odot$ (Hatzes et al. 2006). Using our
252: nominal stellar mass
253: results in a minimum mass
254: for the companion, $m$~sin~$i$ = 2.56 $\pm$ 0.5 $M_{Jupiter}$.
255:
256:
257: A caveat is in order regarding the uncertainty of our mass determination.
258: The error is based on the uncertainties
259: in the stellar parameters. However, the Kjeldsen \& Bedding scaling
260: relations are based on assumptions which may not be valid for a giant
261: star like $\beta$ Gem. Furthermore, due to the short time span of our
262: measurements the frequency of maximum power may actually be in
263: an adjacent mode. If the frequency of maximum power is actually in
264: the modes at $\nu$ = 86.9 $\mu$Hz or $\nu$ = 104.4 $\mu$Hz then
265: the stellar mass can be as low as 1.85 $M_\odot$ or as high as
266: 2.42 $M_\odot$. We thus adopt a value of $\pm$0.3 $M_\odot$ as the error in our
267: mass determination.
268:
269:
270: \section{Discussion}
271:
272: We have detected, for the first time, stellar oscillations in
273: a K giant star known to host an extrasolar giant planet. Our analysis
274: reveals six significant periods, with the dominant mode at a frequency
275: of 86.8 $\mu$Hz. Using the scaling relations of Kjeldsen \& Bedding (1995)
276: as well as the interferometric stellar radius determination results in
277: a stellar mass of 2.04 $M_\odot$. This is in reasonable agreement with the stellar
278: mass of 1.7 $M_\odot$ used by Hatzes et al. (2006) in deriving a companion
279: mass of 2.3 $M_{Jupiter}$. This provides independent confirmation
280: that $\beta$~Gem is indeed an intermediate mass star.
281:
282: The scaling relations of Kjeldson \& Bedding (1995) can also be used
283: to predict the velocity amplitude:
284:
285: $$v_{osc} = {{L/L_\odot} \over {M/M_\odot}} (23.4 \pm 1.4) \hskip 5pt
286: {\rm cm\,s^{-1}}$$
287:
288: The absolute $V$-mag of $\beta$~Gem is 1.08 mag. The effective temperature
289: corresponds to a bolometric correction for a giant star of $BC$ = $-$0.42
290: (Cox 2000) which gives a luminosity, $L$ = 42.8 $L_\odot$. The predicted
291: pulsational velocity amplitude is thus $v_{osc}$ = 5.0 $\pm$ 0.3 m\,s$^{-1}$,
292: in excellent agreement with the 5.3 $\pm$ 0.38 m\,s$^{-1}$ of the dominant mode.
293:
294: The predicted order of the mode, $n$, can be estimated using
295: Equation 10 of Kjeldson \& Bedding (1995). For the stellar parameters of $\beta$
296: Gem this corresponds to $n$ $\approx$
297: 10. The detected modes are most likely high order radial or non-radial modes.
298:
299: The mean frequency spacing of the modes in Table 2 is
300: approximately $\approx$ 20 $\mu$Hz.
301: This is considerably
302: higher than the 7.3 $\mu$Hz large spacing expected given the mass and radius of
303: $\beta$~Gem. We believe that we have not detected all possible modes
304: due to the short time span of our measurements.
305: Deeming (1975) showed that for modes of unequal amplitudes the modes must be separated by
306: a frequency of 2.5/T, where $T$ is the time span of the measurements. Our measurements
307: have $T$ $\approx$ 3 days which corresponds to
308: $\delta\nu$ = 8.8 $\mu$Hz. Furthermore, we do not know what the mode lifetimes
309: are. Continuous measurements spanning
310: a week or more may be required to derive the full oscillation spectrum of $\beta$~Gem.
311:
312:
313: Our investigation of $\beta$~Gem shows the potential of using stellar
314: oscillations to determine the stellar mass of giant stars.
315: We suspect that stellar oscillations in K giant stars are ubiquitous. If
316: that
317: is the case, then an investigation of the stellar oscillations in planet hosting
318: giant star can be used to determine a more accurate stellar mass.
319: We are aware that our mass estimate is only an approximation. More modes
320: need to be detected in
321: $\beta$~Gem and a proper theoretical modeling
322: is required to derive a more accurate stellar mass. We are currently analyzing additional
323: RV measurements for this star to search for additional modes.
324:
325:
326: As is the case with ground-based asteroseismic studies it is very difficult
327: to get the requisite telescope time needed detect all possible modes.
328: Observations using multi-site campaigns are needed due
329: to the one day alias at single sites.
330: This could be difficult
331: because not all astronomical facilities are equipped for making precise
332: stellar RV measurements. Furthermore, even with multiple observing sites,
333: poor weather conditions at one or more sites can still produce data gaps.
334: In this respect the CoRoT space telescope (Baglin et al. 2001)
335: will provide a major breakthrough in the study of K giant oscillations.
336: CoRoT is a 27cm telescope with a dual mission -- asteroseismology on bright
337: stars and a search for transiting exoplanets in a field of
338: up to 12,000 stars in the visual magnitude range $V$ = 11--16.
339: CoRoT can achieve a photometric precision of $\approx$ 2 $\times$
340: 10$^{-4}$ in one hour for a star in the exofield with $V$ =
341: 15.4. CoRoT was launched on 27 December 2007 and as of this writing the
342: light curves from the first observed field had not yet been released.
343:
344:
345:
346:
347: The Kjeldsen \& Bedding (1995) scaling relations yield a predicted
348: photometric amplitude of 100 ppm (10$^{-4}$) for a giant star exhibiting
349: stellar oscillations like $\beta$~Gem, a precision that can be
350: reached for most stars in the CoRoT exofield, and
351: many of these will be giant stars.
352: The 150-day, uninterrupted observations of a given long run CoRoT
353: field should yield the full oscillation spectrum for stars throughout the
354: giant branch.
355: Given that sub-stellar companions may be common around
356: giant stars a search for companion around pulsating K giants
357: found by CoRoT should prove
358: fruitful for planet formation studies.
359:
360: \hskip 10pt
361: \centerline{Acknowledgments}
362:
363: This work was based on observations made on the 2m Alfred Jensch Telescope
364: of the Th\"uringer Landessternwarte Tautenburg.
365: We acknowledge the the support of grant 50OW0204
366: from the Deutsches Zentrum f\"ur Luft- und Raumfahrt e.V. (DLR).
367: This research has made use of the SIMBAD database,
368: operated at CDS, Strasbourg, France.
369:
370:
371: \begin{thebibliography}{11}
372: \expandafter\ifx\csname natexlab\endcsname\relax\def\natexlab#1{#1}\fi
373:
374:
375: \bibitem[]{}
376: Baglin, A. et al. in Proceedings of the SOHO 10/GONG 2000 Workshop: Helio-and
377: Asteroseismology at the Dawn of the Millennium, ed. A. Wilson, ESP SP-464, 2001,
378: p395.
379:
380: \bibitem[]{341}
381: Bazot, M., Vauclair, S., Bouchy, F., \& Santos, N.C. 2005, A\&A, 440, 615
382:
383: \bibitem[]{344}
384: Bedding, T.R, and 11 other authors, 2006, ApJ, 647, 558.
385:
386:
387: \bibitem[]{348}
388: Bell, R.A., Edvardsson, B., \& Gustafsson, B. 1985, MNRAS, 212, 497.
389:
390: \bibitem[]{351}
391: Castanheira, B.G. and 34 co-authors, A\&A, 413, 623
392:
393: \bibitem[]{354}
394: Cox, A. 2000, Allen's Astrophysical Quanties, (Fourth Edition, New York,
395: Springer-Verlag)
396:
397: \bibitem[]{358}
398: Deeming, T.J. 1975, Ap\&SS, 36, 137.
399:
400: \bibitem[]{435}
401: de Ridder, J., Barban, C., Carrier, F., Mazumdar, A., Eggenberger, P., Aerts, C.,
402: Deruyter, S., Vanautgaerden, J. 2006, A\&A, 448, 689
403:
404:
405: %\bibitem[]{361}
406: %D\"ollinger, M., Hatzes, A., Pasquini, L., Guenther, E., Hartmann, M.,
407: %Girardi, L., \& Esposito, M. 2007, A\&A, in press (arXiv:astro-ph/0703672).
408:
409:
410:
411: %\bibitem[]{368}
412: %ESA 2006, {\it The CoRoT Mission, Pre-Launch Status, Stellar Seismology and
413: %Planet Finding}, Eds.: M. Fridlund, A. Baglin, J. Lochard, \& L. Conroy,
414: %ESA SP-1306, ESA Publications Division, Noordwijk, The Netherlands.
415:
416: \bibitem[]{}
417: Frandsen, S., Carrier, F., Aerts, C., Stello, D., Maas, T., Burnet,
418: M., Bruntt, H., Teixeira, T.C., de Medeiros, J.R., Bouchy, F., Kjeldsen,
419: H.,Pijpers, F., \& Christensen-Dalsgaard, J. 2002, A\&A, 394, L5.
420:
421:
422: \bibitem[2002] {frink02}
423: Frink, S., Mitchell, D.S.,
424: Quirrenbach, A., Fischer, D.A.,
425: Marcy, G.W., \& Butler, R.P. 2002, ApJ 576, 478
426:
427: \bibitem[]{390}
428: Hatzes, A.P. \& Cochran, W.D. 1993, ApJ, 413, 339
429:
430: \bibitem[]{393}
431: Hatzes, A.P. \& Cochran, W.D. 1994a, ApJ, 422, 366
432:
433: \bibitem[]{393}
434: Hatzes, A.P. \& Cochran, W.D. 1994b, ApJ, 432, 763
435:
436: \bibitem[]{378}
437: Hatzes, A. P., Guenther, E.W., Endl, M., Cochran, W.D., D\"ollinger, M.P.,
438: Bedalov, A. 2005, A\&A, 437, 743.
439:
440: \bibitem[]{382}
441: Hatzes, A. P., Cochran, W.D., Endl, M., Guenther, E.W.,
442: Saar, S.H., Walker, G.A.H., Yang, S., Hartmann, M.,
443: Esposito, M., Paulson, D.B., D\"ollinger, M.P.,
444: 2006, A\&A, 457, 335.
445:
446:
447:
448:
449: %\bibitem[]{399}
450: %Horne, J.H. \& Baliunas, S.L. 1986, ApJ, 302, 757
451:
452: \bibitem[]{402}
453: Kjeldsen, H. \& Bedding, T.R. 1995, A\&A, 293, 87
454:
455: \bibitem[]{405}
456: Lenz, P. \& Breger, M. 2004, in ``The A-Star Puzzle'', eds. J. Zverko,
457: J. Ziznovsky, S.J. Adelman, \& W. Weiss, IAU Symposium No. 224,
458: Cambridge, UK, Cambridge University Press, 2004, p786.
459:
460: %\bibitem[1976]{Lomb76} Lomb, N.R. 1976, Ap\&SS, 39, 477
461:
462:
463:
464: \bibitem[]{414}
465: M\"ackle, R., Holweger, H., Griffin, R., \& Griffin, R. 1975, A\&AS, 19,
466: 303
467:
468: \bibitem[]{418}
469: Martin, D. 1977, A\&A, 61, 591
470:
471: \bibitem[2002]{mcw90}
472: McWilliam, A. 1990, ApJS, 74, 1075
473:
474: \bibitem[]{424}
475: Nordgren, T.E., Sudol, J.J., Mozurkewich, D. 2001, AJ, 122, 2707
476:
477:
478: \bibitem[]{428}
479: Reffert, S., Quirrenbach, A., Mitchell, S., ALbrecht, S.,
480: Hekker, S., Fischer, D.A., Marcy, G.W., \& Butler, R.P. 2006, ApJ, 652, 661.
481:
482: \bibitem[]{432}
483: Retter, A., Bedding, T.R., Buzasi, D.L., Kjeldsen, H., \& Kiss, L.L. 2003, ApJ, 591, L151
484:
485: \bibitem[2003]{sato03}
486: Sato, B., Ando, H., Kambe, E., Takeda, Y., Izumiura, H.,
487: Masuda, S., Watanabe, E., Noguchi, K., Wada, S., et al. 2003,
488: ApJ 597, L157
489:
490: \bibitem[1982]{Sca82} Scargle, J.D. 1982, ApJ, 263, 835
491:
492:
493: \bibitem[2003]{setiawan03}
494: Setiawan, J., Hatzes, A. P., von der L\"uhe, O., Pasquini, L.,
495: Naef, D., da Silva, L., Udry, S., Queloz, D., Girardi, L.
496: 2003, A\&A 398, L19
497:
498: %\bibitem[]{450}
499: %Setiawan, J., Pasquini, L., da Silva, L., Hatzes, A. P., von der L\"uhe, O.,
500: %Girardi, L., de Medeiros, J.R. \& Guenther, E. 2004, A\&A, 421, 241
501:
502:
503:
504: \end{thebibliography}
505: \clearpage
506:
507: \begin{deluxetable}{cccc}
508: %\footnotesize
509: \tablecaption{Journal of Observations}
510: \tablewidth{0pt}
511: \tablehead{
512: \colhead{Start} & \colhead{Time Coverage} & \colhead{$N_{Obs}$} & \colhead{$\sigma$} \\
513: \colhead{(Julian Day)} & \colhead{(hours)} & & \colhead{(m\,s$^{-1}$)} }
514: \startdata
515: 2450168.308 & 6.88 & 115 & 1.2 \\
516: 2450170.266 & 5.04 & 74 & 1.5 \\
517: 2450171.259 & 8.64 & 152 & 1.7 \\
518: \enddata
519: \end{deluxetable}
520:
521: \begin{deluxetable}{ccrl}
522: %\footnotesize
523: \tablecaption{Oscillation modes for $\beta$~Gem }
524: \tablewidth{0pt}
525: \tablehead{
526: \colhead{Mode} & \colhead{Period} & \colhead{Frequency} & \colhead{Amplitude
527: } \\
528: & \colhead{(hours)} & \colhead{$\mu$Hz} & \colhead{m\,s$^{-1}$
529: } }
530: \startdata
531: $f_1$ & 3.20 & 86.91 $\pm$ 0.37 & 5.29 $\pm$ 0.38 \\
532: $f_2$ & 2.66 & 104.40 $\pm$ 0.49 & 4.09 $\pm$ 0.28 \\
533: $f_3$ & 9.34 & 29.75 $\pm$ 0.54 & 3.61 $\pm$ 0.22 \\
534: $f_4$ & 6.23 & 48.41 $\pm$ 1.06 & 1.85 $\pm$ 0.18 \\
535: $f_5$ & 3.85 & 79.47 $\pm$ 0.76 & 2.58 $\pm$ 0.16 \\
536: $f_6$ & 2.02 & 149.25 $\pm$ 0.63 & 1.25 $\pm$ 0.15 \\
537: $f_7$ & 1.56 & 193.63 $\pm$ 3.07 & 0.64 $\pm$ 0.13 \\
538: \enddata
539: \end{deluxetable}
540:
541: \clearpage
542:
543: \begin{figure}
544: \plotone{f1.eps}
545: \figcaption[ftall.ps]{Time series of the RV measurements
546: for $\beta$~Gem for three nights. The line represents a
547: multi-component sine fit using the six frequencies of Table 1.
548: The standard deviations of the RV values about this fit are
549: 1.2, 1.5, and 1.7 m\,s$^{-1}$ on the three nights, respectively.
550: \label{f1}}
551: \end{figure}
552: \clearpage
553:
554: \begin{figure}
555: \plotone{f2.eps}
556: \figcaption[]{Power spectra of the RV data at each step of the
557: pre-whitening procedure. The top left panel is for the raw RV data.
558: Each successive lower panel shows the power spectrum
559: after subtracting the contribution of the dominant mode
560: from the previous (upper adjacent) panel. The lower right panel is for the final
561: RV residuals after removal of the possible
562: seventh mode ($f_7$). Note the change in y-axis scale for each panel.
563: \label{f2}}
564: \end{figure}
565:
566:
567: \end{document}
568:
569:
570: