1: \documentclass[apjl]{emulateapj}

2: 

3: \def\kms{\ifmmode{\rm km\thinspace s^{-1}}\else km\thinspace s$^{-1}$\fi}

4: \def\ms{\ifmmode{\rm m\thinspace s^{-1}}\else m\thinspace s$^{-1}$\fi}

5: \def\hip{{\it Hipparcos\/}}

6: 

7: \shortauthors{Torres}

8: \shorttitle{GJ 436}

9: 

10: \begin{document}

11: 

12: \journalinfo{Accepted for publication in The Astrophysical Journal Letters}

13: 

14: \title{The transiting exoplanet host star GJ 436: a test of stellar

15: evolution models in the lower main sequence, and revised planetary

16: parameters}

17: 

18: \author{Guillermo Torres}

19: 

20: \affil{Harvard-Smithsonian Center for Astrophysics, 60 Garden St.,

21: Cambridge, MA 02138}

22: 

23: \email{gtorres@cfa.harvard.edu}

24: 

25: \begin{abstract} 

26: 

27: Knowledge of the stellar parameters for the parent stars of transiting

28: exoplanets is pre-requisite for establishing the planet properties

29: themselves, and often relies on stellar evolution models. GJ~436,

30: which is orbited by a transiting Neptune-mass object, presents a

31: difficult case because it is an M dwarf. Stellar models in this mass

32: regime are not as reliable as for higher mass stars, and tend to

33: underestimate the radius. Here we use constraints from published

34: transit light curve solutions for GJ~436 along with other

35: spectroscopic quantities to show how the models can still be used to

36: infer the mass and radius accurately, and at the same time allow the

37: radius discrepancy to be estimated. Similar systems should be found

38: during the upcoming \emph{Kepler\/} mission, and could provide in this

39: way valuable constraints to stellar evolution models in the lower main

40: sequence. The stellar mass and radius of GJ~436 are $M_{\star} =

41: 0.452_{-0.012}^{+0.014}$~M$_{\sun}$ and $R_{\star} =

42: 0.464_{-0.011}^{+0.009}$~R$_{\sun}$, and the radius is 10\% larger

43: than predicted by the standard models, in agreement with previous

44: results from well studied double-lined eclipsing binaries.  We obtain

45: an improved planet mass and radius of $M_p = 23.17 \pm

46: 0.79$~M$_{\earth}$ and $R_p = 4.22_{-0.10}^{+0.09}$~R$_{\earth}$, a

47: density of $\rho_p = 1.69_{-0.12}^{+0.14}$ g~cm$^{-3}$, and an orbital

48: semimajor axis of $a = 0.02872 \pm 0.00027$ AU.

49: 		

50: \end{abstract}

51: 

52: \keywords{

53: planetary systems ---

54: stars: evolution ---

55: stars: fundamental parameters ---

56: stars: individual (GJ~436) ---

57: stars: low-mass, brown dwarfs

58: }

59: 

60: \section{Introduction}

61: \label{sec:introduction}

62: 

63: The applications of stellar evolution theory to astrophysics are so

64: widespread, and its validity so often taken for granted, that it is

65: easy to forget that it took decades to develop, and significant effort

66: to validate by comparison with careful measurements, a process that

67: still continues. It is usually only when those theoretical predictions

68: fail that the classical discipline of stellar evolution ``makes the

69: headlines'', and even then it draws the attention of relatively

70: few. One such instance has occurred for low mass main-sequence

71: stars. Over the last 10 years or so it has become clear that our

72: understanding of the structure and evolution of these objects is still

73: incomplete.  Discrepancies between theory and observation in the radii

74: of stars under 1~M$_{\sun}$, first mentioned by \cite{Hoxie:73},

75: \cite{Lacy:77}, and others, are now well documented for several

76: low-mass eclipsing binaries \citep[see, e.g.,][]{Popper:97,

77: Clausen:99, Torres:02, Ribas:03, Lopez-Morales:05}. Differences in the

78: effective temperatures have been observed as well. The direction of

79: these disagreements is such that model radii are underestimated by

80: roughly 10\%, while effective temperatures are overestimated.

81: 

82: In recent years stellar evolution has had important applications in

83: the field of transiting extrasolar planets. This is because the

84: planetary parameters of interest (mass $M_p$, radius $R_p$) depend

85: rather directly on those of the star ($M_{\star}$, $R_{\star}$), and

86: in most cases models provide the only means of determining the

87: latter. The subject of this paper is GJ~436, a late-type star found by

88: \cite{Butler:04} to be orbited by a Neptune-mass planet with a period

89: of 2.644 days. This object was later discovered by \cite{Gillon:07a}

90: to undergo transits, enabling its size to be determined

91: ($\sim$4~R$_{\earth}$).  As the only M dwarf among the 22 currently

92: known transiting planet host stars, GJ~436 (M2.5V) presents a special

93: challenge for establishing the stellar parameters, because of the

94: disagreements noted above. Little mention seems to have been made of

95: this, and for the most part past studies have relied instead on

96: empirical mass-luminosity ($M\!-\!L$) relations to set the mass of

97: GJ~436. Radius estimates have often rested on the assumption of

98: numerical equality between $M_{\star}$ and $R_{\star}$ for M

99: stars. Despite being the closest transiting planet system (only 10 pc

100: away), it is rather surprising that the mass of the star is only known

101: to about 10\% \citep{Maness:07, Gillon:07b}. This is currently

102: limiting the precision of the planetary mass, and some of that

103: uncertainty translates also to the radius. These two properties are

104: critical for studying the structure of the object.

105: 

106: Given the importance of GJ~436 as the parent star of the only

107: Neptune-mass transiting exoplanet found so far, and hence the closest

108: analog to our Earth with a mass and radius determination, one of the

109: motivations of this paper is to improve the precision of the stellar

110: and planetary parameters by making use of additional observational

111: constraints not used before. Specifically, we incorporate the

112: information on the stellar \emph{density} directly available from the

113: transit light curve \citep{Sozzetti:07}, which provides a strong

114: handle on the size of the star.

115: 

116: The nature of the discrepancies between evolutionary models and

117: observations for low-mass stars has been examined recently from both

118: the observational and theoretical points of view by

119: \cite{Lopez-Morales:07} and \cite{Chabrier:07}. Further progress

120: depends on gathering more evidence to supplement the few available

121: highly accurate mass and radius measurements based on double-lined

122: eclipsing binaries.  Thus, a second motivation for this work, despite

123: the fact that GJ~436 is not a double-lined eclipsing binary, is to

124: present a way of using all observational constraints simultaneously to

125: show that the star presents the same radius anomaly found for the

126: other systems, or more generally, to test the models.  Because similar

127: constraints may become available in the future for other M dwarfs

128: given the keen interest in finding smaller and smaller transiting

129: planets, we anticipate that the indirect technique described here may

130: yield valuable information on this problem and eventually help improve

131: our understanding of low-mass stars.

132: 

133: \section{Constraints and methodology}

134: \label{sec:method}

135: 

136: The basic procedure for establishing the mass and radius of a star

137: that is not in a double-lined eclipsing binary is to place it on an

138: H-R diagram using observational constraints, and compare it with

139: stellar evolution models. We adopt here the models by

140: \cite{Baraffe:98} for a mixing-length parameter of $\alpha_{\rm ML} =

141: 1.0$, which are widely used for low-mass stars.  The constraints

142: available for GJ~436 are several. The spectroscopic study by

143: \cite{Maness:07} established the effective temperature to be $T_{\rm

144: eff} = 3350 \pm 300$~K, and the metallicity was estimated

145: photometrically by \cite{Bonfils:05} to be [Fe/H] $= -0.03 \pm

146: 0.20$. The well-determined {\it Hipparcos} parallax is $\pi_{\rm HIP}

147: = 97.73 \pm 2.27$ mas. With available visual and near infrared

148: photometry from \cite{Leggett:88} and 2MASS, the absolute magnitudes

149: become $M_V = 10.610 \pm 0.051$ and $M_K = 6.048 \pm

150: 0.052$. Additionally, in view of the difficulties in determining

151: effective temperatures for M dwarfs \citep[see, e.g.,][]{Maness:07},

152: we consider as well the infrared color $J\!-\!K = 0.802 \pm

153: 0.024$. Following \cite{Carpenter:01} the 2MASS magnitudes have been

154: transformed to the CIT system of \cite{Elias:82} adopted in the

155: \cite{Baraffe:98} models, and averaged with those of

156: \cite{Leggett:88}, already on that system. The agreement between the

157: two sources is excellent.

158: 

159: Transit light curves for GJ~436 have been obtained from the ground by

160: \cite{Gillon:07a} in the $V$ band, and also at 8$\mu$m by

161: \cite{Deming:07} and \cite{Gillon:07b} using the Spitzer Space

162: Telescope. Aside from minor corrections due to limb-darkening, transit

163: light curves can in general be described using three parameters: the

164: radius ratio between the planet and the star ($R_p/R_{\star}$), the

165: normalized planet-star separation ($a/R_{\star}$), and the impact

166: parameter ($b \equiv a\cos i/R_{\star}$), where $i$ is the inclination

167: angle of the orbit. \cite{Seager:03} have shown that $a/R_{\star}$ is

168: directly related to the density of the star, and thus contains

169: valuable information on its size. \cite{Sozzetti:07} have described

170: how $a/R_{\star}$ can be used together with $T_{\rm eff}$ and stellar

171: evolution models to infer $M_{\star}$ and $R_{\star}$. Briefly, the

172: measured values of $a/R_{\star}$ and $T_{\rm eff}$ are compared with a

173: fine grid of model isochrones for a wide range of ages and

174: metallicities. Theoretical stellar properties are interpolated along

175: each isochrone using a small step in mass, and all points in the H-R

176: diagram matching the observations within their uncertainties are

177: recorded. The best-fitting mass and corresponding radius are assigned

178: errors based on the full range of model values that are consistent

179: with the observations. Other stellar properties can then be read off

180: the best fitting model. This is the procedure we apply below.  For

181: GJ~436 we have restricted the comparison to solar metallicity, given

182: the [Fe/H] estimate by \cite{Bonfils:05}.  The values of the light

183: curve parameters we adopt are weighted averages of the results from

184: the ground-based and Spitzer photometry: $a/R_{\star} = 13.34 \pm

185: 0.58$, $R_p/R_{\star} = 0.0834 \pm 0.0007$, and $b = 0.848 \pm 0.010$.

186: 

187: \section{Mass and radius determinations}

188: \label{sec:mass}

189: 

190: In previous studies the mass and radius of GJ~436 have been derived in

191: three ways: either 1) the mass has been obtained from the

192: near-infrared ($JHK$) mass-luminosity relations of \cite{Delfosse:00}

193: \citep[giving $M_{\star} = 0.44 \pm 0.04$~M$_{\sun}$;][]{Maness:07}

194: and the radius has been assumed to be numerically equal to the mass

195: \citep{Gillon:07a}, or 2) the mass has been held fixed at the above

196: value and the radius constrained directly from the light curve

197: \citep{Gillon:07a, Gillon:07b}, or 3) both $M_{\star}$ and $R_{\star}$

198: have been solved for simultaneously subject to the constraint that

199: they be numerically equal and making use of the implicit sensitivity

200: $R \propto M^{1/3}$ between mass and radius in the light curve fitting

201: procedure \citep{Deming:07, Gillon:07b}. In the second case the radius

202: obtained is near 0.46~R$_{\sun}$ for $M_{\star} = 0.44$~M$_{\sun}$,

203: and in the third case $M_{\star} = R_{\star} \approx 0.47$ or 0.48 in

204: solar units. Thus, differences of order 0.02 to 0.04 remain between

205: these determinations, depending on the procedure used.

206: 

207: \begin{figure}

208: \vskip -0.3in 

209: \epsscale{1.15} 

210: \plotone{f1.eps}

211: \vskip -0.1in 

212: 

213: \figcaption[]{Observational constraints for GJ~436 provided by

214: $a/R_{\star}$, $T_{\rm eff}$, $M_K$, and $J\!-\!K$, shown in four

215: different combinations against \cite{Baraffe:98} isochrones for solar

216: composition. Ages between 1 and 10 Gyr are represented, although the

217: curves are hardly distinguishable because stars in this mass range

218: evolve very slowly.  The raggedness of the isochrones in the right

219: panels is due to the limited precision with which magnitudes are

220: tabulated in the published models. The agreement between the latter

221: and the observations appears good in all panels, but the implied

222: masses and radii show large systematic differences indicative of

223: internal inconsistencies in the models.

224: \label{fig:baraffe}}

225: 

226: \end{figure}

227: 

228: The methodology in the present paper is completely different, and can

229: yield improved precision and also give a better understanding of

230: possible systematics. We initially applied the $a/R_{\star}$ and

231: $T_{\rm eff}$ constraints as described in \S\ref{sec:method} using the

232: \cite{Baraffe:98} models, and obtained a mass value near $M_{\star} =

233: 0.50$~M$_{\sun}$, which is considerably larger than previous

234: estimates. The predicted absolute visual magnitude ($M_V = 9.86$) is

235: also much brighter than that computed directly from the {\it

236: Hipparcos\/} parallax. This inconsistency strongly suggests a problem

237: with the models, which is not entirely unexpected for a star of this

238: type.  As an alternative to $a/R_{\star}$, we then experimented using

239: the absolute $K$ magnitude as a proxy for luminosity, as well as

240: replacing $T_{\rm eff}$ with the color index $J\!-\!K$.\footnote{We

241: have refrained from directly applying any constraint based on the $V$

242: magnitude because of suspected deficiencies in the models for optical

243: passbands, related to missing molecular opacity sources shortward of

244: 1~$\mu$m \citep[see, e.g.,][]{Baraffe:98, Delfosse:00}. Experiments

245: using the $I\!-\!K$ color as an alternative indicate that the $I$ band

246: is also affected at some level.}  Figure~\ref{fig:baraffe} displays

247: the four constraints in different combinations against

248: solar-metallicity isochrones from 1 to 10 Gyr.  The seemingly good fit

249: in all planes belies the serious discrepancies present in other

250: derived quantities that are not shown explicitly. Those results are

251: listed in Table~\ref{tab:baraffe}. Masses inferred from $a/R_{\star}$

252: are systematically larger than those from $M_K$, and so are the

253: radii. The masses from $M_K$ come close to the estimates from the

254: empirical $M\!-\!L$ relations, but the corresponding radii are

255: considerably smaller than expected. This would seem to go in the

256: direction of the results from eclipsing binaries (see

257: \S\ref{sec:introduction}).

258: 

259: On the other hand, there is good evidence from various sources that

260: the bolometric luminosities from these models are not seriously in

261: error \citep{Delfosse:00, Torres:02, Ribas:06, Torres:06}.  This

262: suggests that adjustments to the model radii and temperatures might

263: resolve the discrepancies in Table~\ref{tab:baraffe}, and allow us to

264: obtain a meaningful result for GJ~436. We explored this by introducing

265: a correction factor $\beta$ to the radii, and at the same time

266: applying a factor $\beta^{-1/2}$ to the temperatures in order to

267: preserve the bolometric luminosity. We repeated the comparison between

268: the adjusted models and each of the four sets of constraints for a

269: range of $\beta$ factors centered on the value indicated by the

270: eclipsing binary studies.  Figure~\ref{fig:beta} shows the result for

271: several of the key stellar properties. The lines corresponding to the

272: four sets of constraints seem to converge for a value of $\beta$ near

273: 1.1 (representing a 10\% correction to the model radii), which happens

274: to be the typical factor found by the eclipsing binary studies

275: mentioned earlier. For this value of $\beta$ the models yield

276: essentially the same mass, radius, and luminosity for GJ~436,

277: independently of which set of observational constraints is used, as

278: one would expect from a realistic model. Thus, a self-consistent

279: solution is achieved. To arrive at the best possible values of

280: $M_{\star}$ and $R_{\star}$ we next applied all the constraints

281: simultaneously, and varied $\beta$ as before, seeking the best

282: agreement with the measurements. The result is illustrated in the

283: bottom right panel of Figure~\ref{fig:beta}, where the quality of the

284: match as represented by $\chi^2$ is shown as a function of the

285: correction factor over the restricted range in which the models agree

286: with all four observables within their errors. The best match is again

287: near $\beta = 1.1$. The resulting mass and radius are $M_{\star} =

288: 0.452_{-0.012}^{+0.014}$~M$_{\sun}$ and $R_{\star} =

289: 0.464_{-0.011}^{+0.009}$~R$_{\sun}$. These and other inferred stellar

290: properties for GJ~436 are listed in the top section of

291: Table~\ref{tab:properties}. We emphasize that these quantities are the

292: result of the simultaneous application of the four constraints, and

293: the agreement with some of the values in Table~\ref{tab:baraffe} is

294: accidental.

295: 

296: \begin{figure}

297: \vskip -0.3in 

298: \epsscale{1.15} 

299: \plotone{f2.eps}

300: \vskip 0.1in 

301: 

302: \figcaption[]{Stellar properties for GJ~436 derived from the

303: comparison with stellar evolution models by \cite{Baraffe:98}, as a

304: function of the adjustment factor applied to the model radii

305: ($\beta$). The observational constraints given by $a/R_{\star}$,

306: $T_{\rm eff}$, $M_K$, and $J\!-\!K$ are applied and shown in pairwise

307: combinations, as labeled. Kinks in the curves are a reflection of the

308: discreteness of some of the quantities tabulated in the models. The

309: simultaneous application of all four constraints results in the

310: $\chi^2$ curve shown at the bottom right, indicating a best fit for a

311: $\beta$ value near 1.1 (see text). This value is represented with the

312: vertical dotted line running through this and the other panels.

313: \label{fig:beta}}

314: 

315: \end{figure}

316: 

317: \section{The radius disagreement with the models}

318: \label{sec:disagreement}

319: 

320: Our mass and radius estimates are compared with measurements for

321: late-type double-lined eclipsing binaries in Figure~\ref{fig:mlr}.

322: Only systems with the most accurate determinations are shown (relative

323: errors below 3\%), which are taken from the summary by

324: \cite{Ribas:06}. For comparison we include two isochrones from

325: \cite{Baraffe:98} corresponding to ages of 300 Myr (as estimated for

326: two of these binaries) and 3 Gyr (more representative of the

327: field). All eclipsing binary systems are seen to have larger sizes

328: than predicted for their mass.  We note also that GJ~436 lies in the

329: gap between masses of 0.43 M$_{\sun}$ (for CU~Cnc~A) and 0.60

330: M$_{\sun}$ (GU~Boo~B), and thus provides valuable additional

331: information on the radius discrepancies for low mass stars.

332: 

333: \begin{figure}

334: \vskip -0.4in 

335: \epsscale{1.25} 

336: \plotone{f3.eps}

337: \vskip -0.3in 

338: 

339: \figcaption[]{Mass-radius relation for all double-lined eclipsing

340: binaries with relative mass and radius errors under 3\% \citep[data

341: from][]{Ribas:06}. GJ~436 is shown at the values determined from our

342: modeling (open circle), and is seen to display the same radius

343: discrepancy as the other systems. Two solar-metallicity model

344: isochrones by \cite{Baraffe:98} are shown for reference, corresponding

345: to ages as labeled.\label{fig:mlr}}

346: 

347: \end{figure}

348: 

349: \cite{Lopez-Morales:07} has investigated how these discrepancies

350: ($\Delta R_{\star}/R_{\star}$) depend on metallicity and the strength

351: of the chromospheric activity, quantified in terms of the X-ray

352: luminosity (specifically, $L_X/L_{\rm bol}$), for the rather limited

353: sample available so far. Both of these factors have been suggested to

354: play a role. The above study examined single M dwarfs as well as M

355: dwarfs in binary systems.  GJ~436 is a rather inactive star for its

356: type \citep{Endl:03, Butler:04}, but was detected nonetheless by ROSAT

357: as an X-ray source because of its proximity. The X-ray luminosity was

358: reported by \cite{Hunsch:99} to be $L_X = 0.7 \times 10^{27}$

359: erg~s$^{-1}$. When combined with the bolometric luminosity in

360: Table~\ref{tab:properties}, we obtain $L_X/L_{\rm bol} = 7.0 \times

361: 10^{-6}$. The metallicity was estimated by \cite{Bonfils:05} to be

362: near solar: [Fe/H] $= -0.03 \pm 0.20$.  Considering GJ~436 as a single

363: star, the value $\Delta R_{\star}/R_{\star} \sim$ 10\% we find is

364: consistent with the overall conclusions of \cite{Lopez-Morales:07} in

365: the sense that it is similar to the offsets for other systems

366: regardless of $L_X/L_{\rm bol}$, and at the same time it seems to

367: follow the trend with [Fe/H] exhibited by other single M dwarfs.

368: 

369: \section{Planet parameters and final remarks}

370: \label{sec:discussion}

371: 

372: With the host star properties known, the planet parameters we infer

373: are given in the bottom section of Table~\ref{tab:properties}.  The

374: required orbital period and velocity semi-amplitude $K_{\star}$ are

375: adopted from \cite{Maness:07}, along with the eccentricity from

376: \cite{Demory:07}, who obtain a similar value for $K_{\star}$. The

377: improved precision of these derived parameters is a reflection of the

378: better stellar parameters. The slightly larger planet radius than in

379: previous studies confirms with even greater statistical significance

380: the conclusions of earlier authors regarding the presence of a

381: hydrogen/helium envelope \citep{Gillon:07a, Deming:07, Gillon:07b},

382: and agrees very well with the models by \cite{Fortney:07} for a 10\%

383: fraction of those elements.

384: 

385: In this paper we have shown that GJ~436 provides a valuable test of

386: stellar evolution theory near the bottom of the main sequence, made

387: possible by the fact that it has a transiting planet.  Traditional

388: studies in the area of low-mass stars have made the comparison with

389: models by measuring the mass and radius directly for double-lined

390: eclipsing systems containing M dwarfs. In a few other cases angular

391: diameters have been measured interferometrically for single stars, and

392: the mass has been inferred from empirical $M\!-\!L$ relations

393: \citep[e.g.,][]{Lane:01, Segransan:03}. More recently, a variety of

394: constraints and assumptions have been used to infer the mass and

395: radius of the late-type secondaries in F+M systems observed as part of

396: transiting planet surveys \citep[e.g.,][]{Bouchy:05, Pont:05,

397: Beatty:07}. Though perhaps not as compelling as having actual

398: model-independent mass and radius measurements, the approach in the

399: present work is able to make use of available information for GJ~436

400: and compare the models directly with the observational constraints

401: \emph{without} requiring a direct measurement of the mass and radius.

402: The discrepancy in $R_{\star}$ is derived by parameterizing it in

403: terms of a single adjustment factor to the model radii ($\beta$),

404: assuming the luminosity from theory is accurate, as other observations

405: seem to indicate.

406: 

407: NASA's upcoming \emph{Kepler\/} mission, currently slated to launch in

408: early 2009, will emphasize the search for transiting Earth-size

409: planets.  These should be easier to detect around late-type

410: stars. Therefore, we anticipate that many systems similar to GJ~436

411: could be found and become a significant source of information on radii

412: for low-mass stars, since they will have all the observational

413: constraints needed (including trigonometric parallaxes) to test models

414: of stellar evolution in the way we have done here.

415: 

416: \acknowledgements 

417: 

418: The anonymous referee is thanked for a prompt and helpful report. The

419: author acknowledges partial support for this work from NASA grant

420: NNG04LG89G and NSF grant AST-0708229. This research has made use of

421: the SIMBAD database, operated at CDS, Strasbourg, France, of NASA's

422: Astrophysics Data System Abstract Service, and of data products from

423: the Two Micron All Sky Survey, which is a joint project of the

424: University of Massachusetts and the Infrared Processing and Analysis

425: Center/California Institute of Technology, funded by NASA and the NSF.

426: 

427: \begin{thebibliography}{}

428: 

429: \bibitem[Baraffe et al.(1998)]{Baraffe:98}

430:  Baraffe, I., Chabrier, G., Allard, F., \& Hauschildt, P.\ H. 1998,

431:  \aap, 337, 403

432: 

433: \bibitem[Beatty et al.(2007)]{Beatty:07}

434:  Beatty, T.\ G.\ et al.\ 2007, \apj, 663, 573

435: 

436: \bibitem[Bonfils et al.(2005)]{Bonfils:05}

437:  Bonfils, X., Delfosse, X., Udry, S., Santos, N.\ C., Forveille, T.,

438:  \& S\'egransan, D. 2005, \aap, 442, 635

439: 

440: \bibitem[Bouchy et al.(2005)]{Bouchy:05}

441:  Bouchy, F., Pont, F., Melo, C., Santos, N.\ C., Mayor, M., Queloz,

442:  D., \& Udry, S. 2005, \aap, 431, 1105

443: 

444: \bibitem[Butler et al.(2004)]{Butler:04}

445:  Butler, R.\ P., Vogt, S.\ S., Marcy, G.\ W., Fischer, D., Wright, J.\

446:  T., Henry, G.\ W., Laughlin, G., \& Lissauer, J.\ J. 2004, \apj, 617,

447:  580

448: 

449: \bibitem[Carpenter(2001)]{Carpenter:01}

450:  Carpenter, J.\ M. 2001, \aj, 121, 2851

451: 

452: %\bibitem[Chabrier \& Baraffe(1997)]{Chabrier:97}

453: % Chabrier, G., \& Baraffe, I. 1997, \aap, 327, 1039

454: 

455: \bibitem[Chabrier et al.(2007)]{Chabrier:07}

456:  Chabrier, G., Gallardo, J., \& Baraffe, I. 2007, \aap, in press

457:  (arXiv:0707.1792v1)

458: 

459: \bibitem[Clausen et al.(1999)]{Clausen:99}

460:  Clausen, J.\ V., Baraffe, I., Claret, A., \& VandenBerg, D.\ A. 1999,

461: in Theory and Tests of Convection in Stellar Structure, eds.\ A.\

462: Gim\'enez, E.\ F.\ Guinan, \& B.\ Montesinos, ASP Conf.\ Ser. 173 (San

463: Francisco: ASP), 265

464: 

465: \bibitem[Delfosse et al.(2000)]{Delfosse:00}

466: Delfosse, X., Forveille, T., S\'egransan, D., Beuzit, J.-L., Udry, S.,

467: Perrier, C., \& Mayor, M. 2000, \aap, 364, 217

468: 

469: \bibitem[Deming et al.(2007)]{Deming:07}

470: Deming, D., Harrington, J., Laughlin, G., Seager, S., Navarro, S.\ B.,

471: Bowman, W.\ C., \& Horning, K. 2007, \apj, in press

472: (arXiv:0707.2778v1)

473: 

474: \bibitem[Demory et al.(2007)]{Demory:07}

475:  Demory, B.-O.\ et al.\ 2007, \aap, submitted (arXiv:0707.3809v2)

476: 

477: \bibitem[Elias et al.(1982)]{Elias:82}

478:  Elias, J.\ H., Frogel, J.\ A., Matthews, K., \& Neugebauer, G. 1982,

479:  \aj, 87, 1029

480: 

481: \bibitem[Endl et al.(2003)]{Endl:03}

482:  Endl, M., Cochran, W.\ D., Tull, R.\ G., \& MacQueen, P.\ J. 2003,

483:  \aj, 126, 3099

484: 

485: \bibitem[Fortney et al.(2007)]{Fortney:07}

486:  Fortney, J.\ J., Marley, M.\ S., Barnes, J.\ W. 2007, \apj, 659, 1661

487: 

488: % Secondary Spitzer transit

489: %\bibitem[Gillon et al.(2007c)]{Gillon:07c}

490: % Gillon, M.\ et al.\ 2007c, \aap, 466, 743

491: 

492: % Transit discovery paper

493: \bibitem[Gillon et al.(2007a)]{Gillon:07a}

494:  Gillon, M.\ et al.\ 2007a, \aap, in press, (arXiv:0705.2219v2)

495: 

496: % Primary Spitzer transit

497: \bibitem[Gillon et al.(2007b)]{Gillon:07b}

498:  Gillon, M.\ et al.\ 2007b, \aap, 471, L51

499: 

500: %\bibitem[Henry \& McCarthy(1993)]{Henry:93}

501: % Henry, T.\ J., \& McCarthy, D.\ W.\ Jr. 1993, \aj, 106, 773

502: 

503: \bibitem[Hoxie(1973)]{Hoxie:73}

504:  Hoxie, D.\ T. 1973, \aap, 26, 437

505: 

506: \bibitem[H\"unsch et al.(1999)]{Hunsch:99}

507:  H\"unsch, M., Schmidtt, J.\ H.\ M.\ M., Sterzik, M.\ F., \& Voges,

508:  W. 1999, \aaps, 135, 319

509: 

510: %\bibitem[Kervella et al.(2004)]{Kervella:04}

511: % Kervella, P., Th\'evenin, F., Di Folco, E., \& S\'egransan, D. 2004,

512: % \aap, 426, 297

513: 

514: \bibitem[Lacy(1977)]{Lacy:77}

515:  Lacy, C.\ H. 1977, \apjs, 34, 479

516: 

517: \bibitem[Lane et al.(2001)]{Lane:01}

518:  Lane, B.\ F., Boden, A.\ F., \& Kulkarni, S.\ R. 2001, \apj, 551, L81

519: 

520: \bibitem[Leggett \& Hawkins(1988)]{Leggett:88}

521:  Leggett, S.\ K., \& Hawkins, M.\ R.\ S. 1988, \mnras, 234, 1065

522: 

523: \bibitem[L\'opez-Morales \& Ribas(2005)]{Lopez-Morales:05}

524:  L\'opez-Morales, M., \& Ribas, I. 2005, \apj, 631, 1120

525: 

526: \bibitem[L\'opez-Morales(2007)]{Lopez-Morales:07}

527:  L\'opez-Morales, M. 2007, \apj, 660, 732

528: 

529: \bibitem[Maness et al.(2007)]{Maness:07}

530:  Maness, H.\ L., Marcy, G.\ W., Ford, E.\ B., Hauschildt, P.\ H.,

531:  Shreve, A .\ T., Basri, G.\ B., Butler, R.\ P., \& Vogt, S.\ S. 2007,

532:  \pasp, 119, 90

533: 

534: %\bibitem[Perryman et al.(1997)]{Perryman:97}

535: % Perryman, M.\ A.\ C., et al. 1997, The \hip\ and Tycho Catalogues

536: % (ESA SP-1200; Noordwjik: ESA)

537: 

538: \bibitem[Popper(1997)]{Popper:97}

539:  Popper, D.\ M. 1997, \aj, 114, 1195

540: 

541: \bibitem[Pont et al.(2005)]{Pont:05}

542:  Pont, F., Bouchy, F., Melo, C., Santos, N.\ C., Mayor, M., Queloz,

543:  D., \& Udry, S. 2005, \aap, 438, 1123

544: 

545: \bibitem[Ribas(2003)]{Ribas:03}

546:  Ribas, I. 2003, \aap, 398, 239

547: 

548: \bibitem[Ribas(2006)]{Ribas:06}

549:  Ribas, I. 2006, \apss, 304, 89

550: 

551: \bibitem[Seager \& Mall\'en-Ornelas(2003)]{Seager:03}

552:  Seager, S., \& Mall\'en-Ornelas, G. 2003, \apj, 585, 1038

553: 

554: \bibitem[S\'egransan et al.(2003)]{Segransan:03}

555:  S\'egransan, D., Kervella, P., Forveille, T., \& Queloz, D. 2003,

556:  \aap, 397, L5

557: 

558: \bibitem[Sozzetti et al.(2007)]{Sozzetti:07}

559:  Sozzetti, A., Torres, G., Charbonneau, D., Latham, D.\ W., Holman,

560: M.\ J., Winn, J.\ N., Laird, J.\ B., \& O'Donovan, F.\ T. 2007, \apj,

561: 664, 1190

562: 

563: \bibitem[Torres et al.(2006)]{Torres:06}

564:  Torres, G., Lacy, C.\ H.\ S., Marschall, L.\ A., Sheets, H.\ A., \&

565:  Mader, J.\ A. 2006, \apj, 640, 1018

566: 

567: \bibitem[Torres \& Ribas(2002)]{Torres:02}

568:  Torres, G., \& Ribas, I. 2002, \apj, 567, 1140

569: 

570: \end{thebibliography}

571: 

572: \begin{deluxetable}{lcccc}

573: %\tabletypesize{\scriptsize}

574: \tablewidth{0pc}

575: %\rotate

576: \tablecaption{Stellar parameters for GJ~436 based on different sets of

577: constraints, using the models by

578: \cite{Baraffe:98}.\label{tab:baraffe}}

579: \tablehead{  & \multicolumn{4}{c}{Observational constraints} \\ [+1.0ex]

580: \cline{2-5} \\ [-1.5ex]

581: \colhead {~~~~~~Parameter~~~~~~} & 

582: \colhead{$T_{\rm eff}$ and $a/R_{\star}$} &

583: \colhead{$J\!-\!K$ and $a/R_{\star}$} & 

584: \colhead{$T_{\rm eff}$ and $M_K$} &

585: \colhead{$J\!-\!K$ and $M_K$}}

586: \startdata

587: $M_{\star}$ (M$_{\sun}$)\dotfill & $0.497_{-0.015}^{+0.003}$    & $0.515_{-0.033}^{+0.042}$    & $0.448_{-0.008}^{+0.018}$    & $0.452_{-0.012}^{+0.014}$    \\ [+1.0ex]

588: $R_{\star}$ (R$_{\sun}$)\dotfill & $0.469_{-0.015}^{+0.000}$    & $0.484_{-0.030}^{+0.035}$    & $0.421_{-0.009}^{+0.009}$    & $0.421_{-0.010}^{+0.009}$    \\ [+1.0ex]

589: $\log g_{\star}$ (cgs)\dotfill   & $4.792_{-0.000}^{+0.020}$    & $4.780_{-0.032}^{+0.032}$    & $4.841_{-0.009}^{+0.020}$    & $4.843_{-0.011}^{+0.018}$    \\ [+1.0ex]

590: $a/R_{\star}$\dotfill            & \nodata                      & \nodata                      & $14.63_{-0.19}^{+0.34}$      & $14.65_{-0.21}^{+0.31}$      \\ [+1.0ex]

591: $T_{\rm eff}$ (K)\dotfill        & \nodata                      & $3684_{-55}^{+87}$           & \nodata                      & $3585_{-13}^{+19}$           \\ [+1.0ex]

592: $J\!-\!K$ (mag)\tablenotemark{a}\dotfill         & $0.801_{-0.001}^{+0.008}$    & \nodata                      & $0.810_{-0.007}^{+0.002}$    & \nodata                      \\ [+1.0ex]

593: $M_V$ (mag)\dotfill              & $9.86_{-0.00}^{+0.12}$    & $9.71_{-0.36}^{+0.27}$    & $10.256_{-0.094}^{+0.070}$   & $10.244_{-0.082}^{+0.082}$   \\ [+1.0ex]

594: $M_K$ (mag)\tablenotemark{a}\dotfill              & $5.758_{-0.001}^{+0.090}$    & $5.66_{-0.22}^{+0.18}$    & \nodata                      & \nodata                      \\ [+1.0ex]

595: $L_{\star}$ (L$_{\sun}$)\dotfill & $0.0348_{-0.0035}^{+0.0000}$ & $0.0383_{-0.0069}^{+0.0099}$ & $0.0260_{-0.0016}^{+0.0015}$ & $0.0260_{-0.0017}^{+0.0014}$ \\ [-1.0ex]

596: \enddata

597: \tablenotetext{a}{Magnitudes are in the CIT photometric system of \cite{Elias:82}.}

598: \end{deluxetable}

599: 

600: \begin{deluxetable}{lc}

601: \tablewidth{0pc}

602: \tablecaption{Stellar and planetary parameters for the GJ~436 system.\label{tab:properties}}

603: \tablehead{\colhead{~~~~~~~~~~~~~~Parameter~~~~~~~~~~~~~~} & \colhead{Value}}

604: \startdata

605: \noalign{\vskip -5pt}

606: \sidehead{Stellar parameters}

607: ~~~$M_{\star}$ (M$_{\sun}$)\dotfill & $0.452_{-0.012}^{+0.014}$ \\ [+1.0ex]

608: ~~~$R_{\star}$ (R$_{\sun}$)\dotfill & $0.464_{-0.011}^{+0.009}$ \\ [+1.0ex]

609: ~~~$L_{\star}$ (L$_{\sun}$)\dotfill & $0.0260_{-0.0017}^{+0.0014}$ \\ [+1.0ex]

610: ~~~$\log g_{\star}$ (cgs)\dotfill   & $4.843_{-0.011}^{+0.018}$ \\ [+1.0ex]

611: ~~~Age (Gyr)\tablenotemark{a}\dotfill                & $6_{-5}^{+4}$ \\

612: \noalign{\vskip -2pt}

613: \sidehead{Planetary parameters}

614: ~~~$M_p$ (M$_{\earth}$)\dotfill & 23.17~$\pm$~0.79\phn \\ [+0.7ex]

615: ~~~$R_p$ (R$_{\earth}$)\dotfill & $4.22_{-0.10}^{+0.09}$ \\ [+1.0ex]

616: ~~~$\rho_p$ (g~cm$^{-3}$)\dotfill  & $1.69_{-0.12}^{+0.14}$ \\ [+0.7ex]

617: ~~~$\log g_p$ (cgs)\dotfill        & 3.107~$\pm$~0.040 \\ [+0.7ex]

618: ~~~$a$ (AU)\dotfill              & 0.02872~$\pm$~0.00027 \\ [-1.5ex]

619: \enddata

620: \tablenotetext{a}{Due to the unevolved nature of GJ~436 the age is

621: essentially unconstrained by the observations. We list this value

622: only for completeness.}

623: \end{deluxetable}

624: 

625: \end{document}

626: 

627: