1: %\documentclass[12pt,preprint]{aastex}
2: \documentclass[apjl]{emulateapj}
3:
4: \def\kms{\ifmmode{\rm km\thinspace s^{-1}}\else km\thinspace s$^{-1}$\fi}
5: \def\ms{\ifmmode{\rm m\thinspace s^{-1}}\else m\thinspace s$^{-1}$\fi}
6: \def\gschat3{GSC~03466-00819}
7: \def\hatp3{HAT-P-3}
8: % Activity index R'_HK
9: \newcommand{\rhk}{\ensuremath{R^{\prime}_{\rm HK}}}
10: % log of R'_HK
11: \newcommand{\logrhk}{\ensuremath{\log\rhk}}
12:
13: \shortauthors{Torres}
14: \shorttitle{HAT-P-3b}
15:
16: \begin{document}
17:
18: \title{HAT-P-3\lowercase{b}: A heavy-element rich planet transiting a
19: K dwarf star}
20:
21: \author{
22: G.\ Torres\altaffilmark{1},
23: G.\ \'A.\ Bakos\altaffilmark{1,2},
24: G.\ Kov\'acs\altaffilmark{3},
25: D.\ W.\ Latham\altaffilmark{1},
26: J.\ M.\ Fern\'andez\altaffilmark{1},
27: R.\ W.\ Noyes\altaffilmark{1},
28: G.\ A.\ Esquerdo\altaffilmark{1},
29: A.\ Sozzetti\altaffilmark{1,4},
30: D.\ A.\ Fischer\altaffilmark{5},
31: R.\ P.\ Butler\altaffilmark{6},
32: G.\ W.\ Marcy\altaffilmark{7},
33: R.\ P.\ Stefanik\altaffilmark{1},
34: D.\ D.\ Sasselov\altaffilmark{1},
35: J.\ L\'az\'ar\altaffilmark{8},
36: I.\ Papp\altaffilmark{8},
37: P.\ S\'ari\altaffilmark{8}
38: }
39:
40: \altaffiltext{1}{Harvard-Smithsonian Center for Astrophysics,
41: Cambridge, MA, gtorres@cfa.harvard.edu}
42:
43: \altaffiltext{2}{Hubble Fellow}
44:
45: \altaffiltext{3}{Konkoly Observatory, Budapest, P.O.\ Box 67, H-1125,
46: Hungary}
47:
48: \altaffiltext{4}{INAF - Osservatorio Astronomico di Torino, 10025 Pino
49: Torinese, Italy}
50:
51: \altaffiltext{5}{Department of Physics and Astronomy, San Francisco
52: State University, San Francisco, CA}
53:
54: \altaffiltext{6}{Department of Terrestrial Magnetism, Carnegie
55: Institute of Washington, DC}
56:
57: \altaffiltext{7}{Department of Astronomy, University of California,
58: Berkeley, CA}
59:
60: \altaffiltext{8}{Hungarian Astronomical Association, Budapest,
61: Hungary}
62:
63: \begin{abstract}
64:
65: We report the discovery of a Jupiter-size planet transiting a
66: relatively bright ($V = 11.56$) and metal-rich early K dwarf star with
67: a period of $\sim2.9$ days. On the basis of follow-up photometry and
68: spectroscopy we determine the mass and radius of the planet, \hatp3b,
69: to be $M_p = 0.599 \pm 0.026$~M$_{\rm Jup}$ and $R_p = 0.890 \pm
70: 0.046$~R$_{\rm Jup}$. The relatively small size of the object for its
71: mass implies the presence of about 75~M$_{\earth}$ worth of heavy
72: elements ($\sim\frac{1}{3}$ of the total mass) based on current
73: theories of irradiated extrasolar giant planets, similar to the mass
74: of the core inferred for the transiting planet HD~149026b. The bulk
75: density of \hatp3b\ is found to be $\rho_p = 1.06 \pm 0.17$
76: g~cm$^{-3}$, and the planet orbits the star at a distance of 0.03894
77: AU. Ephemerides for the transit centers are $T_c = 2,\!454,\!218.7594
78: \pm 0.0029 + N\times (2.899703 \pm 0.000054)$ (HJD).
79:
80: \end{abstract}
81:
82: \keywords{
83: binaries: spectroscopic ---
84: planetary systems ---
85: stars: fundamental parameters ---
86: stars: individual (\hatp3) ---
87: techniques: spectroscopic
88: }
89:
90: \section{Introduction}
91: \label{sec:introduction}
92:
93: Ground-based surveys for extrasolar transiting planets have produced
94: 18 discoveries to date. This sample is still small enough that
95: individual discoveries often advance our understanding of these
96: objects significantly by pushing the limits of parameter space, either
97: in planet mass, radius, or some other property. Here we announce a new
98: transiting planet, the third to come out of the HATNet survey, around
99: a star previously known as \gschat3. It is the smallest yet discovered
100: photometrically, and appears to have a heavy-element content
101: representing a substantial fraction of its mass.
102:
103: \section{Photometric detection}
104: \label{sec:detection}
105:
106: % Alternate names
107: %
108: % TYC 3466-819-1
109: % 2MASS 13442258+4801432
110: % GSC 03466-00819
111: % SDSS J134422.59+480143.1
112:
113: \gschat3\ is located in our internally labeled field ``G145'' centered
114: at $\alpha = 13^{\rm h} 48^{\rm m}$, $\delta = +45\arcdeg 00\arcmin$,
115: which was observed using the 11-cm aperture HAT-5 instrument of the
116: HAT network \citep[HATNet;][]{Bakos:02, Bakos:04}, located at the F.\
117: L.\ Whipple Observatory (FLWO) on Mt.\ Hopkins (Arizona). The field
118: was observed in single-station mode, i.e., without networked
119: operations between Arizona and HATNet's other station in
120: Hawaii. Observations over an extended interval of six months (from
121: 2006 January 28 through 2006 July 22) yielded some 3200 photometric
122: measurements per star for approximately $10,\!000$ stars in the field
123: down to $I\sim14$. Following standard frame calibration procedures
124: astrometry was performed as described in \citet{Pal:06}, and aperture
125: photometry results were subjected to the de-trending algorithms we
126: refer to as External Parameter Decorrelation \citep[EPD, described
127: briefly in][]{Bakos:07}, and Trend Filtering Algorithm
128: \citep[TFA;][]{Kovacs:05}. We searched the light curves for box-shaped
129: transit signals using the BLS algorithm of \citet{Kovacs:02}. A
130: prominent 2.9-day signal was detected in the $I\approx 10.52$
131: magnitude star \gschat3 (also known as 2MASS 13442258+4801432; $\alpha
132: = 13^{\rm h} 44^{\rm m} 22\fs58$, $\delta = +48\arcdeg 01\arcmin
133: 43\farcs2$; J2000) with a transit depth of $0.014$ mag and a
134: dip-significance parameter (DSP) of $\sim11$ \citep{Kovacs:05b}. This
135: discovery light curve is shown in Figure~\ref{fig:lc}a. The standard
136: deviation of the measurements after the EPD procedure was 0.009
137: mag. Altogether 5 partial and 2 full transits were observed. As is
138: shown in the following sections the transiting companion is a Jovian
139: planet. Hereafter we refer to the star as HAT-P-3, and to the
140: planetary companion as HAT-P-3b.
141: %
142: \begin{figure}
143: %\vskip -1in
144: \epsscale{1.0}
145: \plotone{f1.ps}
146: \vskip -0.8in
147: \figcaption[]{(a) Unbinned instrumental $I$-band
148: discovery light curve of \hatp3\ obtained with HATNet, folded with the
149: period of $P = 2.899703$ days. (b) Unbinned instrumental Sloan
150: $i$-band photometry collected on UT 2007 April 27 with the 1.2-m
151: telescope at FLWO, along with our best-fit transit model curve (see
152: text).\label{fig:lc}}
153: \end{figure}
154: %
155: The near infrared color of the star ($J\!-\!K_s = 0.49$) corresponds
156: to an early K dwarf, and the annual proper motion listed in the
157: Tycho-2 catalog \citep[$\mu_{\alpha}\cos\delta = -23.3 \pm 2.9$
158: mas~yr$^{-1}$, $\mu_{\delta} = -24.0 \pm 2.6$
159: mas~yr$^{-1}$;][]{Hog:00} is consistent with this. The photometric
160: ephemeris was determined by combining all recorded HATNet transit
161: events with a full transit observed more recently at higher precision
162: with a larger telescope, as described below. The period we obtain is
163: $P = 2.899703 \pm 0.000054$ days, and the reference epoch of transit
164: center is $T_c = 2,\!454,\!218.7594 \pm 0.0029$ (HJD). We use this
165: ephemeris for the remainder of the paper.
166:
167: We note that HAT-P-3 has a faint ($I\approx15$) and red ($J\!-\!K_s =
168: 0.86$) companion at an angular separation of $\sim$10\arcsec\ (2MASS
169: 13442345+4801387), which is entirely within the aperture used for the
170: HATNet photometry. If this happened to be an eclipsing binary, we
171: estimate, based on the relative fluxes, that 1.6-mag deep eclipses of
172: this star could mimic the observed shallow transits seen in the merged
173: light curve. However, our follow-up photometry described in
174: \S\ref{sec:followup} with much better spatial resolution shows that
175: the shallow dips in brightness are indeed present in the light curve
176: of HAT-P-3, whereas the faint companion is constant at the 0.05 mag
177: level.
178:
179: \section{Follow-up observations}
180: \label{sec:followup}
181:
182: \hatp3\ was observed spectroscopically with the FLWO 1.5-m Tillinghast
183: reflector in order to rule out the possibility that the observed drop
184: in brightness is caused by a stellar companion in an eclipsing
185: binary. Three spectra were obtained with the CfA Digital Speedometer
186: \citep{Latham:92} over an interval of 25 days. These observations
187: cover 45~\AA\ in a single echelle order centered at 5187~\AA, and have
188: a resolving power of $\lambda/\Delta\lambda \approx 35,\!000$. Radial
189: velocities were obtained by cross-correlation and have a typical
190: precision of 0.4~\kms. They showed no variation within the
191: uncertainties, ruling out a companion of stellar mass. The mean
192: velocity is $-23.8 \pm 0.1$~\kms.
193:
194: Higher resolution spectroscopy was conducted with the HIRES instrument
195: \citep{Vogt:94} on the Keck~I telescope. With a spectrometer slit of
196: $0\farcs86$ the resolving power is $\lambda/\Delta\lambda \approx
197: 55,\!000$, and the wavelength coverage is $\sim$3800--8000~\AA. An
198: iodine gas absorption cell was used to superimpose a dense forest of
199: $I_2$ lines on the stellar spectrum and establish a highly accurate
200: wavelength fiducial \citep[see][]{Marcy:92}. A total of nine 12-minute
201: exposures were obtained between 2007 March and June with the iodine
202: cell, along with one without $I_2$ for use as a template. Relative
203: radial velocities in the Solar System barycentric frame were derived
204: as described by \cite{Butler:96}, including full modeling of the
205: spatial and temporal variations of the instrumental profile. The
206: results, listed in Table~\ref{tab:rvs}, have internal errors
207: $\lesssim 3~\ms$.
208: %
209: \begin{deluxetable}{lccc}
210: \tablewidth{0pc}
211: \tablecaption{Relative radial velocity measurements of \hatp3.\label{tab:rvs}}
212: \tablehead{\colhead{BJD} & \colhead{RV} & \colhead{$\sigma_{\rm RV}$} & \colhead{} \\
213: \colhead{\hbox{~~~~(2,400,000$+$)~~~~}} & \colhead{(\ms)} & \colhead{(\ms)} & \colhead{Phase}}
214: \startdata
215: 54187.01101\dotfill & $-$38.9 & 1.5 & 0.051 \\
216: 54187.12931\dotfill & $-$67.4 & 1.6 & 0.092 \\
217: 54187.97909\dotfill & $-$78.9 & 1.8 & 0.385 \\
218: 54188.05795\dotfill & $-$59.3 & 1.3 & 0.412 \\
219: 54188.98084\dotfill & $+$75.9 & 1.4 & 0.730 \\
220: 54189.06253\dotfill & $+$75.9 & 1.4 & 0.759 \\
221: 54189.11203\dotfill & $+$73.9 & 1.5 & 0.776 \\
222: 54250.85697\dotfill & $-$48.1 & 2.1 & 0.069 \\
223: 54278.86111\dotfill & $+$65.2 & 2.9 & 0.727 \\
224: \noalign{\vskip -6pt}
225: \enddata
226: \end{deluxetable}
227:
228: The iodine-free template spectrum from Keck was used to infer the
229: atmospheric parameters of the star. A spectral synthesis modeling was
230: carried out using the SME software \citep{Valenti:96}, with wavelength
231: ranges and atomic line data as described by \cite{Valenti:05}. The
232: effective temperature ($T_{\rm eff}$), surface gravity ($\log
233: g_\star$), iron abundance ([Fe/H]), and projected rotational velocity
234: ($v \sin i$) we obtain are listed in Table~\ref{tab:stellar}, where
235: the uncertainties represent internal errors. The temperature and
236: surface gravity correspond to an early K dwarf.
237: %
238: \vskip -10pt
239: \begin{deluxetable}{lcl}
240: \tablewidth{0pc}
241: \tablecaption{Stellar parameters for \hatp3.\label{tab:stellar}}
242: \tablehead{\colhead{~~~~~~Parameter~~~~~~} & \colhead{Value} & \colhead{Source}}
243: \startdata
244: $T_{\rm eff}$ (K)\dotfill & 5185~$\pm$~46\phn\phn & SME\tablenotemark{a} \\
245: $[$Fe/H$]$\dotfill & +0.27~$\pm$~0.04\phs & SME \\
246: $\log g_\star$ (cgs)\dotfill & 4.61~$\pm$~0.05 & SME \\
247: $v \sin i$ (\kms)\dotfill & 0.5~$\pm$~0.5 & SME \\
248: $M_\star$ (M$_{\sun}$)\dotfill & $0.936_{-0.062}^{+0.036}$ & Y$^2$+LC+SME\tablenotemark{b} \\
249: $R_\star$ (R$_{\sun}$)\dotfill & $0.824_{-0.035}^{+0.043}$ & Y$^2$+LC+SME \\
250: $L_\star$ (L$_{\sun}$)\dotfill & $0.442_{-0.057}^{+0.078}$ & Y$^2$+LC+SME \\
251: $M_V$ (mag)\dotfill & 5.86~$\pm$~0.20 & Y$^2$+LC+SME \\
252: Age (Gyr)\dotfill & $0.4_{-0.3}^{+6.5}$ & Y$^2$+LC+SME \\
253: Distance (pc)\dotfill & 140~$\pm$~13\phn & Y$^2$+LC+SME \\
254: \noalign{\vskip -5pt}
255: \enddata
256: \tablenotetext{a}{SME = `Spectroscopy Made Easy' package for analysis
257: of high-resolution spectra \cite{Valenti:96}. See text.}
258: \tablenotetext{b}{Y$^2$+LC+SME = Yale-Yonsei isochrones \citep{Yi:01},
259: light curve parameters, and SME results.\vskip 3pt}
260: \end{deluxetable}
261: \vskip 0.1in
262: %
263: Photometric follow-up of \hatp3\ was carried out in the Sloan $i$ band
264: with KeplerCam \citep[see][]{Holman:07} on the 1.2-m telescope at
265: FLWO. After several attempts thwarted by weather we obtained a
266: reasonably complete and high-precision light curve of \hatp3\ on UT
267: 2007 April 27. An astrometric solution between the individual frames
268: and the 2MASS catalog was carried out using first order polynomials
269: based on $\sim$100 stars per frame. Aperture photometry was performed
270: using a series of apertures on fixed positions around the 2MASS-based
271: [$X$,$Y$] pixel coordinates. We selected a frame taken near the
272: meridian and used $\sim$70 stars and their magnitudes as measured on
273: this reference frame to transform all other frames to a common
274: instrumental magnitude system (weighting by the estimated photometric
275: errors of each star). The aperture yielding the lowest scatter outside
276: of transit was used in the subsequent analysis. The light curve of
277: HAT-P-3 (Figure~\ref{fig:lc}b) was then de-correlated against trends
278: using the out-of-transit sections and a dependence on hour angle and
279: zenith distance. The photometry is somewhat affected by occasional
280: thin cirrus during the night, and has a typical precision of
281: $\sim$2.5~mmag per individual measurement at 20-second cadence.
282:
283: \section{Analysis}
284: \label{sec:analysis}
285:
286: Modeling of the $i$-band light curve was carried out using the
287: formalism of \cite{Mandel:02}. Quadratic limb-darkening coefficients
288: ($u_1 = 0.3832$ and $u_2 = 0.2683$) were taken from the tables of
289: \cite{Claret:04} by interpolation to the stellar properties given in
290: Table~\ref{tab:stellar}. The period was held fixed, and the orbit was
291: assumed to be circular, since the eccentricity is not sufficiently
292: well constrained by the presently available velocity measurements.
293: The four adjusted parameters are the planet-to-star ratio of the radii
294: ($R_p/R_\star$), the normalized separation ($a/R_\star$) where $a$ is
295: the semimajor axis of the relative orbit, the impact parameter ($b
296: \equiv a \cos i/R_\star$), and the time of the center of the transit
297: ($T_c$). A grid search was performed at a fixed $T_c$ to locate the
298: $\chi^2$ minimum in the parameter space of the other three variables,
299: and these were subsequently held fixed to optimize $T_c$. The process
300: was iterated until convergence. Uncertainties were estimated by
301: alternately fixing each parameter at a range of values near the
302: minimum and allowing the others to float, until the best-fit solution
303: produced an increase in the $\chi^2$ corresponding to a 1$\sigma$
304: change. The results are shown in Table~\ref{tab:parameters}.
305:
306: \begin{deluxetable}{lc}
307: \tablewidth{0pc}
308: \tablecaption{Spectroscopic and light curve solutions for \hatp3, and
309: inferred planet parameters.\label{tab:parameters}}
310: \tablehead{\colhead{~~~~~~~~~~~~~~Parameter~~~~~~~~~~~~~~} & \colhead{Value}}
311: \startdata
312: \noalign{\vskip -4pt}
313: \sidehead{Spectroscopic parameters}
314: ~~~$P$ (days)\tablenotemark{a}\dotfill & 2.899703~$\pm$~0.000054 \\
315: ~~~$T_c$ (HJD$-2,\!400,\!000$)\tablenotemark{a}\dotfill & $54,\!218.7594$~$\pm$~0.0029\phm{,2222} \\
316: ~~~$K$ (\ms)\dotfill & 89.1~$\pm$~2.0\phn \\
317: ~~~$\gamma$ (\ms)\dotfill & $-$14.8~$\pm$~1.5\phn\phs \\
318: ~~~$e$\dotfill & 0 (adopted) \\
319: \sidehead{Light curve parameters}
320: ~~~$a/R_\star$\dotfill & $10.59_{-0.84}^{+0.66}$ \\
321: ~~~$R_p/R_\star$\dotfill & $0.1109_{-0.0022}^{+0.0025}$ \\
322: ~~~$b \equiv a \cos i/R_\star$\dotfill & $0.51_{-0.13}^{+0.11}$ \\
323: ~~~$i$ (deg)\dotfill & 87.24~$\pm$~0.69\phn \\
324: ~~~Transit duration (days)\dotfill & 0.0858~$\pm$~0.0020 \\
325: %~~~$u_1$\tablenotemark{b}\dotfill & 0.3832 (adopted) \\
326: %~~~$u_2$\tablenotemark{b}\dotfill & 0.2683 (adopted) \\
327: \sidehead{Planet parameters}
328: ~~~$M_p$ (M$_{\rm Jup}$)\dotfill & 0.599~$\pm$~0.026 \\
329: ~~~$R_p$ (R$_{\rm Jup}$)\dotfill & 0.890~$\pm$~0.046 \\
330: ~~~$\rho_p$ (g~cm$^{-3}$)\dotfill & 1.06~$\pm$~0.17 \\
331: ~~~$a$ (AU)\dotfill & 0.03894~$\pm$~0.00070 \\
332: ~~~$\log g_p$ (cgs)\dotfill & 3.310~$\pm$~0.066 \\
333: \noalign{\vskip -3pt}
334: \enddata
335: \tablenotetext{a}{Held fixed from the photometric determination (\S\ref{sec:detection}).}
336: \end{deluxetable}
337:
338: The relative radial velocities of \hatp3\ were fitted with a Keplerian
339: model constrained to have zero eccentricity. The ephemeris was held
340: fixed at the value for $P$ determined in \S\ref{sec:detection} and
341: $T_c$ from the light curve analysis. The solution fits the data well
342: (see Figure~\ref{fig:rvbis}a), but has an rms scatter of 4.7~\ms\
343: (accounting for the number of degrees of freedom) that is larger than
344: the typical internal errors (1.3--2.9~\ms). No obvious long-term
345: trends are seen in the residuals. Inspection of the spectra reveals
346: that the \ion{Ca}{2} H and K lines display modest emission cores,
347: indicative of a small degree of chromospheric activity. This is known
348: to cause excess scatter in the velocities \citep[e.g.,][]{Saar:97},
349: and is often referred to as ``jitter''. While we cannot rule out
350: other systematics that may be causing our internal errors to be
351: underestimated, we attribute most of the excess scatter to
352: activity. For our final fit we have added 4.2~\ms\ in quadrature to
353: our formal errors, which yields a reduced $\chi^2$ near unity. The
354: parameters of the fit are not significantly changed, and are listed in
355: Table~\ref{tab:parameters}. The minimum companion mass is $M_p \sin i
356: = 0.625 \pm 0.015 \times [(M_\star + M_p)/M_{\sun}]^{2/3}$~M$_{\rm Jup}$,
357: where $M_\star$ represents the mass of the star.
358:
359: \begin{figure}
360: %\vskip -1in
361: \epsscale{1.05}
362: \plotone{f2.ps}
363: %\vskip -0.1in
364: \figcaption[]{(a) Radial-velocity measurements for
365: \hatp3\ along with our orbital fit, shown as a function of orbital
366: phase. The center-of-mass velocity has been subtracted. (b) Bisector
367: spans (BS) for our 10 Keck spectra (template + 9 iodine exposures),
368: computed as described in the text. The mean value has been subtracted,
369: and the vertical scale is the same as in the top
370: panel.\vskip 5pt\label{fig:rvbis}}
371: \end{figure}
372:
373: We investigated the possibility that the radial velocities we measured
374: are the result of distortions in the line profiles due to
375: contamination from an unresolved eclipsing binary \citep{Santos:02,
376: Torres:05}, instead of being due to true Doppler motion in response to
377: a planetary companion. We cross-correlated each Keck spectrum against
378: a synthetic template matching the properties of the star, and averaged
379: the correlation functions over all orders blueward of the region
380: affected by the iodine lines. From this representation of the average
381: spectral line profile we computed the mean bisectors, and as a measure
382: of the line asymmetry we calculated the ``bisector spans'' as the
383: velocity difference between points selected near the top and bottom of
384: the mean bisectors \citep{Torres:05}. If the velocities were the
385: result of a blend with an eclipsing binary, we would expect the line
386: bisectors to vary in phase with the photometric period with an
387: amplitude similar to that of the velocities \citep{Queloz:01,
388: Mandushev:05}. Instead, we detect no variation in excess of the
389: measurement uncertainties (see Figure~\ref{fig:rvbis}, bottom panel).
390: We conclude that the velocity variations are real and that the star is
391: orbited by a Jovian planet.
392:
393: \section{Star and planet parameters}
394: \label{sec:parameters}
395:
396: The mass and radius of the planetary companion scale with those of the
397: parent star. We have estimated the stellar properties by comparison
398: with stellar evolution models from the Yonsei-Yale (Y$^2$) series by
399: \cite{Yi:01}. However, because a direct estimate of the distance is
400: unavailable (\hatp3\ was not observed during the {\it Hipparcos}
401: mission), we have used as a proxy for luminosity ($L_\star$) the
402: quantity $a/R_\star$ from the light curve fit, which is closely
403: related to the stellar density \citep[see][]{Sozzetti:07}. As
404: described by those authors, $a/R_\star$ is typically a better
405: constraint on $L_\star$ than the spectroscopic value of $\log
406: g_\star$, which has a relatively subtle effect on the line profiles
407: and whose determination is therefore more susceptible to systematic
408: errors. Following \cite{Sozzetti:07} we have determined the range of
409: stellar masses and radii that are consistent with the measured values
410: of $T_{\rm eff}$, [Fe/H], and $a/R_\star$ within their
411: uncertainties. Based on previous experience, for this application we
412: have adopted more conservative errors for the spectroscopic quantities
413: of 80~K for $T_{\rm eff}$ and 0.08~dex for [Fe/H]. We obtain $M_\star
414: = 0.936_{-0.062}^{+0.036}$~M$_{\sun}$ and $R_\star =
415: 0.824_{-0.035}^{+0.043}$~R$_{\sun}$. The predicted surface gravity,
416: $\log g_\star = 4.577_{-0.045}^{+0.019}$, agrees well with the
417: spectroscopic estimate (see Table~\ref{tab:stellar}). The nominal age
418: we infer for \hatp3\ is only 0.4~Gyr, although the uncertainty is very
419: large due to the unevolved nature of this K dwarf. While this nominal
420: age would make \hatp3\ the youngest known transiting extrasolar planet
421: host star, we note that there is no sign of the
422: \ion{Li}{1}~$\lambda$6708 line in the spectra of \hatp3\ down to a
423: factor of $\sim$4 lower than the Li level in stars of similar spectral
424: type in the Hyades cluster, which have an age of $\sim$0.7 Gyr. This
425: would suggest an age for \hatp3\ greater than 0.7 Gyr. Further
426: evidence for an older age is provided by the measured chromospheric
427: activity index of the star based on the \ion{Ca}{2} H and K lines in
428: the Keck spectra: $\logrhk$ is $-4.59\pm0.06$, indicating an age of
429: $\sim$1.3 Gyr based on the calibration by \cite{noyes84}.
430:
431: Combined with the results from the light curve and radial velocity
432: modeling, the above stellar parameters yield a planet mass of $M_p =
433: 0.599 \pm 0.026$~M$_{\rm Jup}$ and a radius of $R_p = 0.890 \pm
434: 0.046$~R$_{\rm Jup}$. The surface gravity of the planet, which is
435: independent of the stellar parameters \citep[see,
436: e.g.,][]{Southworth:07}, is $\log g_p = 3.310 \pm 0.066$ (cgs), and is
437: consistent with the $P$/$\log g_p$ relation shown by those authors.
438: These and other planet properties are listed in
439: Table~\ref{tab:parameters}.
440:
441: The absolute visual magnitude we estimate for the star using the
442: stellar evolution models ($M_V = 5.86 \pm 0.20$) combined with the
443: apparent visual magnitude \citep[$V = 11.561 \pm 0.067$;][]{Droege:06}
444: yields a distance of $140 \pm 13$~pc, ignoring extinction. With this
445: distance estimate we investigated the possibility that the faint
446: 10\arcsec\ companion to \hatp3\ mentioned in \S\ref{sec:detection} is
447: physically associated. We find that it is not: its $K_s$-band
448: brightness as measured by 2MASS is approximately 2 magnitudes fainter
449: than predicted for a star with its $J\!-\!K_s$ color, using the
450: best-fit isochrone to \hatp3. It is therefore a background star.
451:
452: \section{Discussion and concluding remarks}
453:
454: The properties of \hatp3b\ place it among the smaller transiting
455: planets discovered so far, and are remarkably similar to those of the
456: recently announced XO-2b \citep{Burke:07}, not only in mass and radius
457: (reported as 0.57~M$_{\rm Jup}$ and 0.97~R$_{\rm Jup}$, respectively),
458: but also in some of the characteristics of the parent star such as
459: $T_{\rm eff}$ and $M_\star$. In the case of \hatp3b, with a mass of
460: $M_p = 0.599$~M$_{\rm Jup}$, the radius measured here appears too
461: small for a pure hydrogen-helium planet according to current theory.
462: The models for irradiated giant planets by \cite{Fortney:07} indicate
463: the radius could be explained by the presence of a heavy-element core
464: of some 75~M$_{\earth}$ ($\sim\frac{1}{3}$ of the total $M_p$) for the
465: nominal mass, orbital semimajor axis, and age we infer, although
466: uncertainties in these properties and in the planetary radius allow
467: the range to be from about 50 to 100~M$_{\earth}$. \hatp3b\ thus joins
468: HD~149026b \citep{Sato:05} as a hot Jupiter for which theory predicts
469: a substantial fraction of heavy elements. Using the same models and
470: the published properties of XO-2b, we find that XO-2b is also expected
471: to have a 40~M$_{\earth}$ core representing about $\frac{1}{4}$ of its
472: total mass. These three heavy-element rich planets, along with others
473: suggested by \cite{Burrows:07}, provide important support for the
474: core-accretion scenario of planet formation. \hatp3b\ appears
475: consistent with the correlation proposed by \cite{Guillot:06} between
476: the host star metallicity and the amount of heavy elements in the
477: planet interior, although there seems to be at least one exception to
478: this relation \citep[WASP-1b;][]{Stempels:07}. Compared to HD~149026b,
479: \hatp3b\ has a similar mass of heavy elements but considerably more
480: hydrogen and helium, a feature that will need to be explained by
481: planet formation models.
482:
483: As of this writing \hatp3b\ is the smallest extrasolar transiting
484: planet discovered photometrically. While HD~149026b and GJ~436b
485: \citep{Gillon:07} do have smaller radii, they were both originally
486: discovered in Doppler surveys before transits were noticed.
487:
488: \cite{Gaudi:05} has argued that selection effects in
489: signal-to-noise-limited transit surveys tend to bias the size
490: distribution of close-in giant extrasolar planets toward larger radii,
491: since those objects have deeper transits and are therefore easier to
492: detect. The fact that the recent discoveries of \hatp3b\ and XO-2b
493: are on the small side of the radius distribution would seem to
494: contradict the argument. However, the transit depth in both of these
495: cases is not particularly small ($\sim$1.4\%), a result of the fact
496: that the parent stars are K dwarfs rather than earlier type
497: stars. Thus, there may still be a bias, although relatively small. We
498: note also that the properties of \hatp3b\ are entirely consistent with
499: the Monte Carlo simulations of the population of transiting planets by
500: \cite{Fressin:07} (see, e.g., their Fig.\ 17), as are other recently
501: announced transiting planets. This supports the idea that these
502: objects are quite representative of the general population of giant
503: planets, and are not significantly different from the Doppler planets.
504:
505: \acknowledgements
506:
507: We are grateful to the referee, T.\ Guillot, for helpful comments on
508: the original manuscript. Operation of the HATNet project is funded in
509: part by NASA grant NNG04GN74G. We acknowledge partial support also
510: from the Kepler Mission under NASA Cooperative Agreement NCC2-1390
511: (D.W.L., PI). G.T.\ acknowledges partial support from NASA under
512: grant NNG04LG89G, and work by G.\'A.B.\ was supported by NASA through
513: Hubble Fellowship Grant HST-HF-01170.01-A. G.\'A.B.\ also wishes to
514: acknowledge the use of the photometry software that is under
515: development by A.\ P\'al. G.K.\ thanks the Hungarian Scientific
516: Research Foundation (OTKA) for support through grant K-60750. This
517: research has made use of Keck telescope time granted through NASA, of
518: the VizieR service \citep{Ochsenbein:00} operated at CDS, Strasbourg,
519: France, of NASA's Astrophysics Data System Abstract Service, and of
520: the 2MASS Catalog.
521:
522: \begin{thebibliography}{}
523:
524: \bibitem[Bakos et al.(2002)]{Bakos:02}
525: Bakos, G.\ \'A., L\'az\'ar, J., Papp, I., S\'ari, P., \& Green,
526: E.\ M.\ 2002, \pasp, 114, 974
527:
528: \bibitem[Bakos et al.(2004)]{Bakos:04}
529: Bakos, G.\ \'A., Noyes, R.\ W., Kov\'acs, G., Stanek, K.\ Z.,
530: Sasselov, D.\ D., \& Domsa, I. 2004, \pasp, 116, 266
531:
532: \bibitem[Bakos et al.(2007)]{Bakos:07}
533: Bakos, G.\ \'A., et al. 2007, \apj, submitted (arXiv:0705.0126)
534:
535: \bibitem[Burrows et al.(2007)]{Burrows:07}
536: Burrows, A., Hubeny, I., Budaj, J., \& Hubbard, W.\ B. 2007, \apj,
537: 661, 502
538:
539: \bibitem[Butler et al.(1996)]{Butler:96}
540: Butler, R.\ P., Marcy, G.\ W., Williams, E.\ et al. 1996, \pasp, 108,
541: 500
542:
543: \bibitem[Burke et al.(2007)]{Burke:07}
544: Burke, C.\ J.\ et al. 2007, ApJ, submitted (arXiv:0705.0003v1)
545:
546: \bibitem[Claret(2004)]{Claret:04}
547: Claret, A. 2004, \aap, 428, 1001
548:
549: \bibitem[Droege et al.(2006)]{Droege:06}
550: Droege, T.\ F., Richmond, M.\ W., \& Sallman, M. 2006, \pasp, 118,
551: 1666
552:
553: \bibitem[Fortney et al.(2007)]{Fortney:07}
554: Fortney, J.\ J., Marley, M.\ S., \& Barnes, W. 2007, \apj, 659, 1661
555:
556: \bibitem[Fressin et al.(2007)]{Fressin:07}
557: Fressin, F., Guillot, T., Morello, V., \& Pont, F. 2007, \aap, in
558: press (arXiv:0704.1919)
559:
560: \bibitem[Gaudi(2005)]{Gaudi:05}
561: Gaudi, B.\ S. 2005, \apj, 628, L73
562:
563: \bibitem[Gillon et al.(2007)]{Gillon:07}
564: Gillon, M. et al. 2007, \aap, in press (arXiv:0705.2219v2)
565:
566: \bibitem[Guillot et al.(2006)]{Guillot:06}
567: Guillot, T., Santos, N.\ C., Pont, F., Iro, N., Melo, C., \& Ribas,
568: I. 2006, \aap, 453, L21
569:
570: \bibitem[H\o g et al.(2000)]{Hog:00}
571: H\o g, E. et al. 2000, \aap, 355, L27
572:
573: \bibitem[Holman et al.(2007)]{Holman:07}
574: Holman, M.\ J. et al. 2007, \apj, in press (arXiv:0704.2907)
575:
576: \bibitem[Kov\'acs et al.(2002)]{Kovacs:02}
577: Kov\'acs, G., Zucker, S., \& Mazeh, T. 2002, \aap, 391, 369
578:
579: \bibitem[Kov\'acs et al.(2005)]{Kovacs:05}
580: Kov\'acs, G., Bakos, G.\ \'A., \& Noyes, R.\ W. 2005, \mnras, 356,
581: 557
582:
583: \bibitem[Kov{\'a}cs \& Bakos(2005)]{Kovacs:05b}
584: Kov{\'a}cs, G., \& Bakos, G.~\'A.~2005, astro-ph/0508081
585:
586: \bibitem[Latham(1992)]{Latham:92}
587: Latham, D.\ W. 1992, in IAU Coll.\ 135, Complementary Approaches to
588: Double and Multiple Star Research, ASP Conf.\ Ser.\ 32, eds.\ H.\ A.\
589: McAlister \& W.\ I.\ Hartkopf (San Francisco: ASP), 110
590:
591: \bibitem[Mandel \& Agol(2002)]{Mandel:02}
592: Mandel, K., \& Agol, E. 2002, \apjl, 580, L171
593:
594: \bibitem[Mandushev et al.(2005)]{Mandushev:05}
595: Mandushev, G., et al. 2005, \apj, 621, 1061
596:
597: \bibitem[Marcy \& Butler(1992)]{Marcy:92}
598: Marcy, G.\ W., \& Butler, R.\ P. 1992, \pasp, 104, 270
599:
600: \bibitem[Noyes et al.(1984)]{noyes84}
601: Noyes, R.~W., Weiss, N.~O., \& Vaughan, A.~H.~1984, \apj, 287, 769
602:
603: \bibitem[Ochsenbein et al.(2000)]{Ochsenbein:00}
604: Ochsenbein, F., Bauer, P., \& Marcout, J. 2000, \aaps, 143, 23
605:
606: \bibitem[P\'al \& Bakos(2006)]{Pal:06}
607: P\'al, A., \& Bakos, G.\ \'A. 2006, \pasp, 118, 1474
608:
609: \bibitem[Queloz et al.(2001)]{Queloz:01}
610: Queloz, D., et al. 2001, \aap, 379, 279
611:
612: \bibitem[Saar \& Donahue(1997)]{Saar:97}
613: Saar, S.\ H., \& Donahue, R.\ A. 1997, \apj, 485, 319
614:
615: \bibitem[Santos et al.(2002)]{Santos:02}
616: Santos, N.\ C. et al. 2002, \aap, 392, 215
617:
618: \bibitem[Sato et al.(2005)]{Sato:05}
619: Sato, B. et al. 2005, \apj, 633, 465
620:
621: \bibitem[Southworth et al.(2007)]{Southworth:07}
622: Southworth, J., Wheatley, P.\ J., \& Sams, G. 2007, \mnras, in press
623: (arXiv:0704.1570)
624:
625: \bibitem[Sozzetti et al.(2007)]{Sozzetti:07}
626: Sozzetti, A., Torres, G., Charbonneau, D., Latham, D.\ W., Holman,
627: M.\ J., Winn, J.\ N., Laird, J.\ B., \& O'Donovan, F.\ T. 2007, \apj,
628: in press (arXiv:0704.2938)
629:
630: \bibitem[Stempels et al.(2007)]{Stempels:07}
631: Stempels, H.\ C., Collier Cameron, A., Hebb, L., Smalley, B., \&
632: Frandsen, S. 2007, \mnras, in press (arXiv:0705.1677v1)
633:
634: \bibitem[Torres et al.(2005)]{Torres:05}
635: Torres, G., Konacki, M., Sasselov, D.\ D., \& Jha, S. 2005, \apj,
636: 619, 558
637:
638: \bibitem[Valenti \& Fischer(2005)]{Valenti:05}
639: Valenti, J.\ A., \& Fischer, D.\ A. 2005, \apjs, 159, 141
640:
641: \bibitem[Valenti \& Piskunov(1996)]{Valenti:96}
642: Valenti, J.\ A., \& Piskunov, N. 1996, \aaps, 118, 595
643:
644: \bibitem[Vogt et al.(1994)]{Vogt:94}
645: Vogt, S.\ S.\ et al. 1994, Proc.\ SPIE, 2198, 362
646:
647: \bibitem[Yi et al.(2001)]{Yi:01}
648: Yi, S.\ K., Demarque, P., Kim, Y.-C., Lee, Y.-W., Ree, C.\ H.,
649: Lejeune, T., \& Barnes, S. 2001, \apjs, 136, 417
650:
651: \end{thebibliography}
652:
653: %\clearpage
654:
655: %\clearpage
656:
657: %\clearpage
658:
659: \end{document}
660:
661: