1: % ms.tex
2: % Version for RESUBMISSION to ApJ
3: % March 18, 2009
4: %
5: %
6: %
7: %\documentclass[12pt, preprint]{aastex}
8: \documentclass{emulateapj}
9: %\usepackage[top=1in, bottom=1in, left=1in, right=1in]{geometry}
10: \usepackage{amssymb,amsmath,latexsym,natbib,wasysym}
11: %\usepackage[pdftex,usenames]{color}
12: \usepackage{pdfcolmk}
13: %\usepackage{multicol,subfigure}
14: %\usepackage{fullpage}%\usepackage{setspace}
15:
16: \pdfpagewidth 8.5in
17: \pdfpageheight 11in
18:
19:
20: %\usepackage{mcode}
21: %\numberwithin{equation}{section}
22: %Attempt to track changes
23:
24: \voffset -0.6in
25:
26: \DeclareMathAlphabet{\mathpzc}{OT1}{pzc}{m}{it}
27:
28: \newcommand\be{\begin{equation}}
29: \newcommand\ee{\end{equation}}
30:
31: %\newcommand{\annote}[2]{\textcolor{ForestGreen}{\bfseries \emph{ #1}
32: % \footnote{\textcolor{ForestGreen}{\bf Edit Note: #2}}}}
33: \newcommand{\annote}[2]{#1}
34:
35: \def\x{\times}
36: \def\e#1{10^{#1}}
37:
38: %\doublespacing
39:
40: \begin{document}
41:
42: \pagenumbering{arabic}
43:
44: \slugcomment{Accepted by the Astrophysical Journal}
45: \shorttitle{Probing Extra-solar Planet Interiors}
46: \shortauthors{Ragozzine \& Wolf}
47:
48: \title{Probing the Interiors of Very Hot Jupiters Using
49: Transit Light Curves}
50:
51: \author{Darin Ragozzine \& Aaron S. Wolf\footnote{Both authors
52: contributed equally to this work.}}
53: %\date{\today}
54: \affil{Division of Geological and Planetary Sciences, California Institute
55: of Technology, Pasadena, CA 91125}
56:
57: \email{darin@gps.caltech.edu, awolf@gps.caltech.edu}
58:
59: \begin{abstract}
60: Accurately understanding the interior structure of extra-solar planets
61: is critical for inferring their formation and evolution. The internal density
62: distribution of a planet has a direct effect on the star-planet
63: orbit through the gravitational quadrupole field created by the rotational and tidal
64: bulges. These quadrupoles induce apsidal precession that is proportional to
65: the planetary Love number ($k_{2p}$, twice the
66: apsidal motion constant), a bulk physical characteristic of the planet that
67: depends on the internal density distribution, including the presence or absence
68: of a massive solid core. We find that the quadrupole of the planetary
69: tidal bulge is the dominant source of apsidal precession for very
70: hot Jupiters ($a \lesssim 0.025$ AU), exceeding the effects of general relativity
71: and the stellar quadrupole by more than an order of magnitude.
72: For the shortest-period planets, the planetary interior
73: induces precession of a few degrees per year.
74: By investigating the full photometric signal of apsidal precession, we find that changes
75: in transit shapes are much more important than transit timing variations.
76: With its long baseline of ultra-precise photometry,
77: the space-based \emph{Kepler} mission can
78: realistically detect apsidal precession with the accuracy necessary to
79: infer the presence or absence of a massive core in very hot
80: Jupiters with orbital eccentricities as low as $e \simeq 0.003$. The signal
81: due to $k_{2p}$ creates unique transit light curve variations that are generally not
82: degenerate with other parameters or phenomena.
83: We discuss the plausibility of measuring $k_{2p}$ in an effort to
84: directly constrain the interior properties of extra-solar planets.
85: \end{abstract}
86:
87: \keywords{stars: planetary systems}
88:
89: \maketitle
90:
91:
92: \section{Introduction}
93: Whether studying planets within our solar system or planets orbiting other stars,
94: understanding planetary interiors represents our best
95: strategy for determining their bulk composition, internal dynamics,
96: and formation histories. For our closest neighbors, we have had the luxury of sending spacecraft to
97: accurately measure the higher-order gravity fields of these objects,
98: yielding invaluable constraints on their interior density distributions. Using
99: these observations, we have been able, for instance, to infer the presence of
100: large cores, providing support for the core-accretion theory of planet
101: formation \citep{2005AREPS..33..493G}. Study of planets outside our solar system, however, has
102: necessitated the development and usage of more indirect
103: techniques. Nevertheless, as the number of well-characterized
104: extra-solar planets grows, we gain more clues that help us answer the
105: most fundamental questions about how planets form and evolve.
106:
107: Guided by our current understanding of planetary physics, we have begun
108: to study the interiors
109: of extra-solar planets. This endeavor has been dominated by a
110: model-based approach, in which the mass and radius of a planet are
111: measured using radial velocity and transit photometry observations,
112: and the interior properties are inferred by finding the model most
113: consistent with those two observations. This strategy clearly requires
114: a set of assumptions, not the least of which is that the
115: physical processes at work in extra-solar planets are just like those
116: that we understand for our own giant planets. While it does seem that
117: this approach is adequate for explaining most of the known transiting
118: planets, there does exist a group of planets for which the usual set of
119: assumptions are not capable of reproducing the observations
120: \citep[e.g.,][]{2006A&A...453L..21G,2007ApJ...661..502B}.
121: These are the planets with so-called
122: positive ``radius
123: anomalies'', including the first-discovered transiting planet
124: HD 209458b \citep{2000ApJ...529L..45C}. Though most of these planets can be explained by
125: adjusting different pieces of the interior physics in the models (including
126: opacities, equations of state, and heat deposition), it is currently impossible to
127: discern which combination of these possible explanations is actually
128: responsible for their observed sizes \citep{2006A&A...453L..21G}.
129:
130: Additional uncertainties also exist for planets at the other end of
131: the size spectrum. For the group of under-sized extra-solar planets, such as HD 149026b,
132: the canonical approach is to give the planet a
133: massive highly condensed core of heavy elements in order to match the
134: observed radius. This approach also provides a first order estimate of
135: the planet's bulk composition, in terms of its fraction of heavy
136: elements. There is
137: also the added complication of how the assumed state of
138: differentiation affects the inferred composition and predicted
139: structure \citep{Baraffe2008}.
140:
141: Currently, the most promising approach to modeling the distinctive features of extra-solar planet interiors
142: is to study the known transiting planets as an ensemble. The group can be used to develop
143: either a single consistent model that reproduces all the observations \citep[e.g.,][]{2006A&A...453L..21G}
144: or to showcase the possible diversity in model parameters \citep[e.g., opacities, as in][]{2007ApJ...661..502B}.
145: Surely, a model-independent measure of interior structure would be valuable in order to begin disentangling
146: otherwise unconstrained physics.
147:
148: The idea of obtaining direct structural measurements for distant objects is
149: by no means a new one. For decades, the interiors of eclipsing binary
150: stars have been measured by observing ``apsidal motion,''
151: i.e. precession of the orbit due to the non-point-mass component of
152: the gravitational field \citep{1928MNRAS..88..641R,1938MNRAS..98..734C,1939MNRAS..99..451S,1939MNRAS..99..662S}.
153: The signal of the changing orbit is encoded in
154: the light curves of these systems by altering the timing of the
155: primary and secondary eclipses. From these eclipse times, it is
156: straightforward to determine the so-called apsidal motion
157: constant which then constrains the allowed interior density distributions.
158: Interior measurements inferred
159: from apsidal precession were among the first indications that stars
160: were highly centrally condensed. While it seems non-intuitive, we show
161: in this paper that we can use a similar technique to measure the
162: interior properties of very hot Jupiters. Most surprisingly, the
163: interior structure signal for very hot Jupiters actually
164: dominates over the signal from the star, yielding an unambiguous
165: determination of planetary interior properties.
166:
167: Our theoretical analysis is also extended to full simulated photometry in order to
168: explore the observability of apsidal precession.
169: We show that this precession is observable
170: by measuring the subtle variations in transit light curves. The photometric analysis is focused on the data expected from
171: NASA's \emph{Kepler} mission, which successfully launched on March 6, 2009
172: \citep{2003SPIE.4854..129B,2006Ap&SS.304..391K}.
173: \emph{Kepler} will obtain exquisite
174: photometry on $\sim$100,000 stars, of which about 30 are expected to
175: host hot Jupiters with periods less than 3 days \citep{2008arXiv0804.1150B}. \emph{Kepler} has the potential to
176: measure the gravitational quadrupoles of very hot Jupiters though the
177: technique described below. If successful, this will constitute a major
178: step towards an understanding of the diversity of planetary interiors.
179:
180: In Section 2, we describe the background theory that connects
181: interior structure and orbital dynamics and explore which effects are most
182: important. Section 3 applies this theory to the observable changes in
183: the transit photometry, including full \emph{Kepler} simulated
184: light curves. We show in Section 4 that the signal due to the planetary
185: interior has a unique signature. Other methods for
186: inferring planetary interior properties are discussed in Section
187: 5. The final section discusses the important conclusions of our work.
188:
189: \section{Background Theory}
190:
191: \subsection{Coordinate System and Notation}
192: The internal structure of very hot
193: Jupiters can be determined by observing changes in the planet's
194: orbit. These changes can be described in terms of two general types of
195: precession. Apsidal precession refers to rotation of the orbital
196: ellipse within the plane of the orbit. It is characterized by
197: circulation of the line of apsides, which lies along the major axis of the
198: orbit. Nodal precession, on
199: the other hand, occurs out of the plane of the orbit and refers to the
200: orbit normal precessing about the total angular momentum vector of the
201: system. For typical very hot Jupiter systems with no other planets, apsidal
202: precession has a much stronger observable signal than nodal
203: precession (see Section \ref{obliquitysec}), so we focus our discussion on
204: the simpler case of a fixed orbital plane.
205:
206: As is typical for non-Keplerian orbits, the star-planet
207: orbit is described using osculating orbital elements that change in time. We
208: identify the plane of the sky as the reference plane and orient the
209: coordinate axes in the usual way such that the sky lies in the x-z plane
210: with the y-axis pointing at Earth. The intersection of the
211: orbital plane and the reference plane is called the line of nodes, but
212: without directly resolving the system, there is no way to determine
213: the orientation of the line of nodes with respect to astronomical
214: North; thus, the longitude of the ascending node, $\Omega$,
215: cannot be determined. Given this degeneracy, we simplify the
216: description by orienting the z-axis to lie within the plane spanned by the orbit
217: normal and the line-of-sight. The angle between the line of sight and the orbit normal
218: is $i$, the inclination. The x-axis is in the plane of the sky and is the reference
219: line from which the argument of periapse ($\omega$) is measured (in
220: the standard counter-clockwise sense). For this choice of coordinates,
221: the argument of periapse and longitude of periapse ($\varpi$) are
222: equivalent. Given this coordinate system, transit centers occur when
223: the planet crosses
224: the y-z plane; this point lies $90^{\circ}$ past the reference x-axis,
225: and thus primary transits occur when the true anomaly, $f$, satisfies $f
226: _{tr}+\omega_{tr}\equiv90^{\circ}$, where the subscript $tr$ indicates the
227: value at transit center.\footnote{In elliptical orbits, if the inclination is not
228: 90$^{\circ}$, the photometric minima do not exactly coincide with the
229: planetary conjunctions. See \citet{1959cbs..book.....K}, p. 388 and section \ref{transitshapes}
230: below.}
231:
232: Throughout this paper, we refer to
233: parameters of the star (mass, radius, etc.) with
234: subscripts of ``$*$'' and parameters of the planet with
235: subscripts of ``$p$''.
236: For evaluation of various equations, we will take as
237: fiducial values the mass ratio $M_p/M_* = \e{-3}$,
238: the radius ratio
239: $R_p/R_* = 0.1$ (though some low density planets have radius ratios
240: greater than 1/6), and the semi-major axis in stellar radii
241: $a/R_* = 6$, typical for very hot Jupiters, which we define as planets
242: with semi-major axes $a \lesssim 0.025$ AU (see Table 1).\footnote{Throughout this work, we do
243: not distinguish between $M_{tot}$ and $M_*$, since $M_p \ll M_*$.} In this definition,
244: we deviate from \citet{2008arXiv0804.1150B}, who define very hot Jupiters as planets
245: with periods less than 3 days. These authors estimate that $\emph{Kepler}$ will find $\sim$30 such planets,
246: of which $\sim$16 will be brighter than V=14 (T. Beatty, pers. comm.).
247: Since our definition is more stringent, our
248: technique will be applicable to fewer \emph{Kepler} planets.
249:
250: \subsection{Rotational and Tidal Potentials}
251:
252: It is well known from classical
253: mechanics, that if stars and planets are considered to be purely spherical
254: masses, then they will obey a simple $r^{-2}$ force law and hence
255: execute closed elliptical orbits. Non-spherical mass effects are caused by the
256: application of external potential(s): the centrifugal potential of
257: spinning bodies causes rotational flattening and the tidal potential
258: of a nearby mass raises tidal bulges. Rotational and tidal bulges
259: create gravitational quadrupole fields ($r^{-3}$) that lead to
260: orbital precession.
261:
262: The complex subject of how planets\footnote{For clarity, in
263: these sections we focus on the planetary shape, though the derivations
264: are also valid for stars.} respond to
265: applied potentials is encapsulated in the so-called theory
266: of figures \citep{1978ppi..book.....Z}. As long as the distortions
267: are small, we can simplify
268: the problem by ignoring the small interaction terms between the tidal
269: and rotational potentials; in this paper, we thus restrict ourselves
270: to the first order theory, where the two planetary responses simply
271: add. Even in the linear case,
272: the way the fluid planet responds depends on the full radial density
273: structure of the planet. The planetary response is conveniently
274: captured in a single variable $k_{2p}$, using the definition
275: \be
276: \label{k2def}
277: V_2^{\rm ind}(R_p) \equiv k_{2p} V_2^{\rm app}(R_p)
278: \ee
279: where $k_{2p}$ is the Love number of the planet,
280: which is just a constant of proportionality between the applied second
281: degree potential field $V_2^{\rm app}$ and the resulting field that it
282: induces $V_2^{\rm ind}$ at the surface of the planet. Due to the
283: orthogonality of the Legendre polynomials used to express the gravity field,
284: if the planet is responding
285: to a second degree harmonic field, then only the second degree
286: harmonic of the planet's gravity field is altered, to first-order. Thus,
287: $k_{2p}$ is a measure of how the redistribution of mass caused by the applied
288: potential actually affects the external gravity field of the
289: planet. In the stellar literature, the symbol $k_2$ is used for the
290: apsidal motion constant, which is half of the secular/fluid Love
291: number that we use throughout this paper \citep{1939MNRAS..99..451S}.
292:
293: The Love number $k_2$ is an extremely useful parameterization, as it
294: hides the complex interactions of a planet and an
295: applied potential in just a single number. The process of calculating
296: $k_2$ of a fluid object (like stars and gas giants), from the interior density
297: distribution is fairly straightforward and outlined in several places
298: \citep[e.g.,][]{1939MNRAS..99..451S,1959cbs..book.....K}. Objects with most of their mass
299: near their cores, like stars, have very low $k_{2}$ values \citep[$\sim$0.03
300: for main sequence solar-like stars,][]{1995A&AS..109..441C} since the
301: distorted outer envelope has little mass and therefore little effect
302: on the gravity field. Planets have much flatter density
303: distributions, and thus distortions of their relatively more massive outer
304: envelopes
305: greatly affect the gravity field. At the upper extreme lies a uniform
306: density sphere, which has $k_2 = 3/2$. In this way, {\it $k_2$ can be
307: thought of as a measure of the level of central condensation of an
308: object}, with stronger central condensation corresponding to smaller
309: $k_2$.
310:
311: By examining the variations in $k_2$ for giant planets within our own
312: Solar
313: System, we can gain a feel for its expected values and how sensitive
314: it is to internal structure. The $n=1$ polytrope is commonly used to approximate
315: the density structure of (cold) gas giant planets; it has $k_2 \approx
316: 0.52$ \citep{1959cbs..book.....K}. This can be compared to the value
317: determined
318: from the gravity measurements of Jupiter, where $k_{2J} \simeq 0.49$. Even
319: though Jupiter may have a ~10 Earth mass core, it is small in
320: comparison to Jupiter's total mass, and thus it has minor effect
321: on the value of $k_2$. Saturn, on the other hand has a roughly 20
322: Earth mass core and is less than 1/3 of Jupiter's mass.
323: As a result, the presence of
324: Saturn's core is easily seen in the value of its Love number $k_{2S}
325: \approx 0.32$. From this, we can see that planets with and without
326: significant cores differ in $k_{2p}$ by about
327: $\sim 0.1$. This can also be inferred from \citet{2003ApJ...588..545B} by
328: using the Darwin-Radau relation to convert the moment of inertia
329: factor to $k_2$. Furthermore, \citet{2001ApJ...548..466B} list the moment
330: of inertia factors of various planet models of HD 209458 b and $\tau$
331: Bootis b, which correspond to a range of $k_{2p}$ values from $\sim$0.1 to $\sim$0.6.
332:
333: Current methods for inferring the internal structures of
334: extra-solar planets combine measurements of the mass and radius
335: with a model to obtain estimates of the planet's implied composition
336: and core size. Unfortunately, these models require one to make
337: assumptions about the degree of differentiation, among other things
338: \citep{Baraffe2008}. A
339: good measurement of $k_{2p}$, however, reveals important independent
340: structure information, which can break the degeneracies between
341: bulk composition and the state of differentiation. Given
342: such a wide range of potential $k_{2p}$ values, even an
343: imprecise measurement of $k_{2p}$ will be extremely valuable for
344: understanding extra-solar planets. By measuring the $k_{2p}$ values for
345: extra-solar planets, we can also uncover constraints on the
346: density structure that are independent of the measurement of the
347: planetary radius. This new information may
348: allow us to probe the unknown physics responsible for the currently
349: unexplained radius anomalies.
350:
351: \subsubsection{Induced External Gravity Field}
352: The internal structures of planets in our own solar system are most
353: readily characterized by the zonal harmonics of the planet's gravity
354: field, i.e. $J_2$, $J_4$, etc. It is these high-order harmonics that are
355: directly measured by spacecraft flybys. To better understand the
356: connection between the two, we can relate the $k_2$
357: formulation to $J_2$ by writing out the expression for the induced
358: potential at the surface of the planet in Equation \ref{k2def} in
359: terms of the definition of $J_2$, yielding:
360: $k_{2p} V_2^{\rm app}(R_p) = -J_2 \frac{GM_p}{R_p}P_2(\cos\theta)$, where
361: $P_2$ is the usual Legendre polynomial and $\theta$ is the planetary
362: co-latitude \citep{1999ssd..book.....M}.
363: We can use this equation to obtain expressions for the $J_2$ field
364: induced by both rotation and tides (discussed in more detail below).
365: The relation relies on dimensionless
366: constants which compare the strength of the acceleration due to
367: gravity with that of the rotational and tidal potentials:
368: \be
369: \label{qparams}
370: q_r = \frac{\nu_p^2 R_p^3}{G M_p} \qquad {\rm and} \qquad q_t =
371: -3\left( \frac{R_p}{r} \right)^3 \left(\frac{M_*}{M_p}\right)
372: \ee
373: where $\nu_p$ is the angular spin frequency of the planet. For the case
374: where the spin axis and tidal bulge axis are
375: perpendicular (i.e. zero obliquity), the relationship between $J_2$
376: and $k_2$ is, to first order:
377: \be
378: \label{J2perp}
379: J_{2} = \frac{k_2}{3}\left(q_r - \frac{q_t}{2} \right)
380: \ee
381: Note that $q_t$ is a function of the instantaneous orbital separation, $r$,
382: and is thus constantly changing in an eccentric orbit in response to the
383: changing tidal potential. Hence $J_2$ for eccentric extra-solar planets is a
384: complex function of time. This is why it is more sensible to analyze the orbital
385: precession in terms of $k_2$, which is a fixed intrinsic property of
386: the planet, rather than $J_2$.
387:
388: As very hot Jupiters are expected to be synchronously locked (denoted by $s$)
389: with small eccentricities, it can easily be shown that
390: $q_t^s\approx-3q_r$, which simplifies equation \ref{J2perp} yielding:
391: \be
392: J_{2p}^s \simeq \frac{5}{6}k_{2p} q_r \simeq \frac{5}{6} k_{2p} \left( \frac{M_*}{M_p}
393: \right) \left( \frac{R_p}{a} \right)^3
394: \ee
395: Using a moderate value of $k_{2p}=0.3$, the $J_2$ of very hot Jupiters
396: reaches as high as 5 $\x \e{-3}$, about half of the measured $J_2$ of
397: Jupiter and
398: Saturn.
399:
400:
401: \subsection{Apsidal Precession}
402: The quadrupole field created by rotational and tidal potentials
403: discussed above induces precession of the star-planet orbit.
404: Both Jupiter and Saturn have rather significant quadrupoles,
405: dominated entirely by their sizeable
406: rotational bulges resulting from rapid rotation periods
407: of less than 10 hours. In contrast, very hot Jupiters are expected to be synchronously
408: rotating, and thus their spin periods are longer by a factor of a
409: few. Since the rotational bulge size goes as the square of the spin
410: frequency, very hot Jupiters should have rotational bulges that are at
411: least an order of magnitude smaller than Jupiter and Saturn, inducing
412: only tiny quadrupole fields. These extra-solar planets are extremely
413: close to their parent stars, however, with semi-major axes of only
414: $\sim6$ stellar radii. Very hot Jupiters are thus expected to have large tidal bulges which are
415: shown below to dominate the quadrupole field and resulting apsidal
416: precession.
417:
418: \subsubsection{Precession Induced by Tidal Bulges}
419: The orbital effect of tidal bulges is complicated by their continuously changing size.
420: While tidal bulges always point directly\footnote{We can ignore the
421: lag due to dissipation, which has an angle of only $Q_p^{-1}
422: \lesssim 10^{-5}$ for giant planets \citep{1966Icar....5..375G,1999ssd..book.....M}.}
423: at the tide-raising object, their size is a function of orbital
424: distance. Since the height of the tidal bulge depends on the actual separation
425: between the objects, the second-order gravitational potential is
426: time-varying in eccentric orbits. Accounting for this dependence
427: (which cannot be captured by using a fixed $J_2$) is critical, as
428: illustrated by \citet{1939MNRAS..99..451S}. The dominant tidal
429: perturbation to the external gravity field of the planet, evaluated at
430: the position of the star, is a second-order potential:
431: \be
432: V_{tid}(r) = \frac{1}{2} k_2 GM_* R_p^5 r^{-6}
433: \ee
434:
435: The apsidal precession due to the tidal bulge, including the effect of
436: both the star and the planet is \citep{1939MNRAS..99..451S,2001ApJ...562.1012E}:
437: \begin{eqnarray}
438: \label{tidalprec}
439: \dot{\omega}_{\rm tidal} & = & \dot{\omega}_{\rm tidal,*} +
440: \dot{\omega}_{\rm tidal,p} \nonumber \\
441: & = &
442: \frac{15}{2} k_{2*} \left( \frac{R_*}{a}\right)^5 \frac{M_p}{M_*} f_2(e) n \nonumber \\
443: & + &
444: \frac{15}{2} k_{2p} \left( \frac{R_p}{a}\right)^5 \frac{M_*}{M_p} f_2(e) n
445: \end{eqnarray}
446: where $n$ is the mean motion and $f_2(e)$ is an eccentricity function:
447: \begin{eqnarray}
448: f_2(e) & = & (1-e^2)^{-5}(1 + \frac{3}{2} e^2 + \frac{1}{8} e^4) \nonumber \\
449: & \approx & 1 + \frac{13}{2} e^2 + \frac{181}{8} e^4 + ... \label{f2e}
450: \end{eqnarray}
451: Note that the factor of 15 does not appear for stationary
452: rotational bulges, as detailed below, and comes through Lagrange's
453: Planetary Equations from the higher dependence on radial separation
454: $(r^{-6})$ in the tidal potential. For this reason, tidal bulges
455: are much more important in producing apsidal precession.
456:
457: Furthermore, the main factor of
458: importance to extra-solar planets is the
459: mass ratio, which comes in because the height of the tide is
460: proportional to the mass of the tide-raising body. Consider the ratio
461: of the planetary and stellar effects:
462: \be
463: \frac{\dot{\omega}_{\rm tidal,p}}{\dot{\omega}_{\rm tidal,*}} =
464: \frac{k_{2p}}{k_{2*}} \left( \frac{R_p}{R_*} \right)^5 \left(
465: \frac{M_*}{M_p} \right)^2 \simeq 100
466: \ee
467: For tidal bulges, the apsidal motion due to the planet clearly dominates over
468: the contribution of the star. Even though the planet's radius is
469: smaller than the star's by a factor of ten, the star is so much more massive
470: than the planet that it raises a huge tidal bulge, which consequently
471: alters the star-planet orbit. The benefit provided by the inverse
472: square of the small mass ratio is compounded by the
473: order of magnitude increase in $k_2$ of the planet over the star.
474:
475:
476: \subsubsection{Precession Induced by Rotational Bulges}
477: The quadrupolar gravitational field due to the planetary rotational bulge, evaluated
478: at the star's position is:
479: \be
480: V_{\rm rot}(r) = \frac{1}{3} k_2 \nu_p^2 R_p^5 r^{-3} P_2(\cos \alpha_p)
481: \ee
482: where $\alpha_p$ is the planetary obliquity, the angle between the orbit normal
483: and the planetary spin axis. \citet{1939MNRAS..99..451S} assumes zero
484: obliquity and calculates the secular
485: effect of this perturbation on the
486: osculating Keplerian elements. This final result, including the effect of both the star and the planet
487: is\footnote{The full equation, including arbitrary obliquities, is
488: given in \citet{1978ASSL...68.....K}, Equation V.3.18
489: \citep[see also ][]{1939MNRAS..99..451S,2001ApJ...562.1012E}. Also recall that, unlike these authors, we
490: use the symbol $k_2$ to represent the Love number which is twice the
491: apsidal motion constant called $k_2$ in eclipsing binary literature.}:
492: \begin{eqnarray}
493: \label{dorot}
494: \dot{\omega}_{\rm rot} & = & \dot{\omega}_{\rm rot,*} +
495: \dot{\omega}_{\rm rot,p} \nonumber \\
496: & = &
497: \frac{k_{2*}}{2} \left( \frac{R_*}{a} \right)^5 \frac{\nu_*^2 a^3}{GM_*}
498: g_2(e) n \nonumber \\
499: & + &
500: \frac{k_{2p}}{2} \left( \frac{R_p}{a} \right)^5 \frac{\nu_p^2
501: a^3}{GM_p} g_2(e) n
502: \end{eqnarray}
503: where $g_2(e)$ is another eccentricity function:
504: \be
505: g_2(e) = (1-e^2)^{-2} \approx 1 + 2e^2 + 3e^4 + ... \label{g2e}
506: \ee
507:
508: Evaluating the importance of this effect requires an understanding
509: of the spin states of very hot Jupiters and their stars.
510: The rotation and spin pole orientation of very hot Jupiters should be
511: tidally damped on timescales $\lesssim 1$ MYr
512: \citep[e.g.,][]{2004ApJ...610..464D, Ferraz-Mello2008}. We therefore assume
513: that all planets have reached the psuedosynchronous rotation rate
514: derived by \citet{1981A&A....99..126H}. The rotation rate of the star is usually much slower
515: since the tidal stellar spin-up timescale is much longer than $\sim$1 GYr
516: \citep{2007ApJ...665..754F}.
517:
518: If both the star and the planet were spinning synchronously, the
519: stellar and planetary rotational bulges would have comparable contributions to apsidal precession.
520: However, since the tidal bulge of the planet is a much more
521: important effect, we find that even fast-spinning stars have a very weak contribution to apsidal
522: precession.
523:
524: \subsubsection{Total Apsidal Precession}
525:
526: The other major contributor to the apsidal precession in extra-solar
527: planetary systems is general relativity. The anomalous apsidal
528: advance of Mercury's orbit due to its motion near the massive Sun was
529: one of the first confirmations of general
530: relativity. This same apsidal advance is prevalent in
531: very hot Jupiter systems and has been shown to be possibly detectable
532: through long-term transit
533: timing \citep{2002ApJ...564.1019M,2007MNRAS.377.1511H,PK08,JB08}. The
534: relativistic advance is given (to lowest order) by:
535: \be
536: \dot{\omega}_{GR} = \frac{3 G M_* n}{ac^2(1-e^2)}
537: \ee
538:
539: One additional effect for non-synchronous planets is due to thermal tides
540: \citep{2009arXiv0901.0735A}, which create a bulge on the planet due to
541: temperature-dependent expansion of an unevenly-radiated upper atmosphere.
542: The thermal tidal bulge is very small in mass and is not expected to provide a
543: significant contribution to apsidal precession (P. Arras, pers. comm.)
544: and is thus neglected.
545:
546: Since we are considering only the lowest-order effects, all the
547: apsidal precession rates (rotational/tidal for the star/planet and
548: general relativity) simply add to give the total apsidal
549: precession (roughly in order of importance for very hot Jupiters):
550: \be
551: \dot{\omega}_{\rm tot} = \dot{\omega}_{\rm tid,p} + \dot{\omega}_{\rm GR}
552: + \dot{\omega}_{\rm rot,p} + \dot{\omega}_{\rm rot,*} + \dot{\omega}_{\rm tid,*}
553: \ee
554: We are ignoring the small cross-terms (geodetic
555: precession, quadrupole-quadrupole coupling, Lense-Thirring effect,
556: nutation, etc.) for the purposes of this paper as higher-order corrections.
557:
558: Calculating each of these contributions to the precession shows that
559: \emph{for very hot Jupiters, the dominant term in the total apsidal
560: precession is due to the
561: planetary tidal bulge}. For the known transiting planets, the
562: fraction of apsidal precession due to the planet is calculated and
563: illustrated in Figure 1. The precession due to the interiors of very
564: hot Jupiters towers over the other effects. General relativity, the
565: next largest effect
566: is $\sim$10 times slower than the precession caused by the planetary
567: tidal bulge.
568:
569: The apsidal precession rate of very hot Jupiters due solely to the
570: interior structure of the planet is:
571: \begin{eqnarray}
572: \dot{\omega}_{\rm p} & \approx & 3.26 \x \e{-10} \textrm{ rad/sec } \x \left(
573: \frac{k_{2p}}{0.3} \right)
574: \left( \frac{M_*}{M_{\astrosun}} \right)^{3/2} \x \nonumber \\
575: & & \left( \frac{M_p}{M_{\rm J}} \right)^{-1}
576: \left( \frac{R_p}{R_{\rm J}} \right)^{5}
577: \left( \frac{a}{0.025 \textrm{ AU}} \right)^{-13/2}
578: \end{eqnarray}
579: which explains why low density very close-in Jupiters are the prime targets for
580: measuring apsidal precession. For these planets, the precession rate can
581: reach a few degrees per year.
582:
583: The precession due to the planet has generally been neglected in extra-solar
584: planet transit timing work
585: to date \citep{2002ApJ...564.1019M,2007MNRAS.377.1511H}, which has
586: considered stellar oblateness or general relativity to be the dominant
587: effects (in the absence of other planets) though \citet{JB08} have
588: also pointed out that $\dot{\omega}_{\rm tidal,p}$ can be an
589: important source of apsidal precession. We find that the planetary
590: quadrupole is usually 1-2 orders of magnitude more important than effects
591: previously considered for single very hot Jupiters. Hence, measuring
592: apsidal precession essentially gives $\dot{\omega}_{\rm tid,p}$
593: which is directly proportional to
594: $k_{2p}$, implying that transit light curve variations due to apsidal
595: precession can
596: directly probe the interiors of extra-solar planets.
597:
598: \begin{figure}
599: \begin{center}
600: \includegraphics[width=3.5in]{f1}
601: \caption{\label{odotfig}
602: \textbf{Fraction of Apsidal Precession Due to the Planetary Quadrupole.}
603: The points show the planetary fraction of
604: the total apsidal precession calculated for the known transiting
605: extra-solar planets with properties taken from J. Schneider's Extra-Solar Planet Encyclopedia
606: (\texttt{http://www.exoplanet.eu}), assuming the planet has a typical Love number
607: of $k_{2p}=0.3$ (e.g. Saturn-like). The apsidal precession
608: induced by the tidal and
609: rotational bulges of the planet overcome precession due to general relativity
610: and the star, especially for short period planets.
611: The "error bars" show the range of planetary
612: contributions for a 5\% variation in stellar masses (and hence
613: $\dot{\omega}_{\rm GR}$) and the comparatively smaller effect of varying the stellar
614: Love number and rotation rate over all reasonable values.
615: The five cases where
616: the planetary contribution to apsidal precession is most
617: important (boxed) also have the shortest precession periods:
618: WASP-12b, CoRoT-1b, OGLE-TR-56b, WASP-4b, and TrES-3b would
619: fully precess
620: in about 18, 71, 116, 120, and 171 years, respectively. The planet in the lower left is CoRoT-7,
621: a super-Earth planet whose planetary contribution to precession is small because of its small radius. Transiting
622: planets with periods longer than 6 days all had planetary contributions less than 0.15.
623: In all cases,
624: the dominant signal in apsidal precession of very hot Jupiters is $k_{2p}$, which
625: is determined by their internal density distribution and is a powerful
626: probe into their interior structure.
627: }
628: \end{center}
629: \end{figure}
630:
631:
632: \subsection{Modification of the Mean Motion}
633:
634:
635: Non-Keplerian potentials also modify the
636: mean-motion, $n$, and cause a
637: small deviation from Kepler's Third Law. Including the effects
638: described above, the non-Keplerian mean motion, $n'$,
639: is (dropping second-order corrections):
640: \be
641: n' = n \left( 1 + \epsilon - \frac{3GM_*}{2ac^2} \right)
642: \ee
643: where $\epsilon$ is defined as
644: \begin{eqnarray}\label{epsilon}
645: \epsilon & = & \frac{k_{2*}}{2} q_{r,*} \left( \frac{R_*}{a}
646: \right)^2 + \frac{k_{2p}}{2} q_{r,p} \left( \frac{R_p}{a} \right)^2 \nonumber \\
647: & + & 3 k_{2*} \frac{M_p}{M_{tot}} \left( \frac{R_*}{a} \right)^5 +
648: 3 k_{2p} \frac{M_*}{M_p} \left( \frac{R_p}{a} \right)^5
649: \end{eqnarray}
650: and $n^2 \equiv \frac{GM_{tot}}{a^3}$. The general
651: relativistic correction to the mean motion is from \citet{1989racm.book.....S}. (Throughout
652: this paper, except where noted, the difference between $n'$ and $n$ is ignored as a higher-order
653: correction.)
654:
655: As with apsidal precession, the planetary quadrupole is more important than the stellar
656: quadrupole by about 2
657: orders of magnitude. At the largest, the correction to the mean motion
658: is a few times $\e{-5}$. \citet{2006NewA...11..490I} used the fact
659: that quadrupole moments cause deviations to Kepler's Third Law to
660: attempt to derive the $J_2$ of the star HD 209458 (the quadrupole of the
661: planet was incorrectly ignored).
662:
663: However, as \citet{2006NewA...11..490I} found, this method is only feasible if
664: you know the masses and semi-major axes of the orbit \emph{a priori}
665: or independently from Kepler's Law. Since the error in stellar masses
666: (from radial velocities and evolutionary codes) is usually 3-10 \%
667: \citep[e.g.][]{Torres2008}, the propagated error on
668: $k_{2p}$ would be a few times greater
669: than the highest $k_{2p}$ expected, making this method
670: impractical. It has been proposed that the stellar mass and semi-major axis can be
671: precisely and independently measured via the light-travel time effect described by
672: \citet{2005ApJ...623L..45L}. In practice, however, the light-travel
673: time effect is highly degenerate with the unknown transit epoch and/or the
674: orbital eccentricity. We find that a precise independent measurement of $M_*$
675: from light-travel time is
676: impractical even
677: with the excellent photometry of \emph{Kepler}.\footnote{We do note that
678: detailed observations of multiple-planet systems can
679: yield mass estimates of each of the bodies independently.
680: \emph{Kepler} asteroseismology can also provide independent information about
681: stellar mass and other properties \citep{2008arXiv0807.0508K}.}
682:
683: \subsection{Expectations for Planetary Eccentricities}
684:
685: Thus far, we have quantified how planetary interiors affect the orbit
686: through precession. The photometric observability of this apsidal
687: precession is highly dependent on the current orbital eccentricity
688: ($e$). Small eccentricities are the
689: largest limitation to using transit light curves to probe extra-solar planet
690: interiors. Indeed, if eccentricities are very low,
691: measuring apsidal precession from transit light curves may not be possible
692: for any of the \emph{Kepler} planets.
693:
694: Nearly all hot Jupiters have eccentricities consistent with zero,
695: though the radial velocity technique has difficulty putting 3-$\sigma$ upper
696: limits on eccentricities smaller than 0.05 \citep{2005ApJ...629L.121L}. So far, the
697: strongest constraints are placed by comparing the deviation of the
698: secondary transit time from half the orbital period, which are related
699: by \citep[e.g.,][]{2005ApJ...626..523C}:
700: \be \label{heqoffset}
701: e \cos \omega \simeq \frac{\pi}{2P_{\rm orb}} (t_{\rm sec}-t_{\rm
702: prim}-\frac{P_{\rm orb}}{2})
703: \ee
704:
705: Similarly, by measuring the primary and secondary transit durations
706: ($\Theta_I$ and $\Theta_{II}$), an additional constraint can be placed on
707: $e \sin \omega$. The equation commonly quoted in the extra-solar
708: planet literature \citep{1999ebs..conf.....K,2003ASPC..294..449C,
709: 2006ApJ...653L..69W}
710: has a sign error; the correct equation is derived by
711: \citet{1959cbs..book.....K}, p. 391 :
712: \be \label{keqdurratio}
713: e \sin \omega = \frac{\Theta_{II}-\Theta_I}{\Theta_{II}+\Theta_I}
714: \frac{\alpha^2 - \cos^2 i}{\alpha^2 - 2 \cos^2 i}
715: \ee
716: where $\alpha \equiv \frac{R_*+R_p}{a\sqrt{1-e^2}}$. The accuracy of this measurement is typically smaller than for
717: $e \cos \omega$, but we include this equation to note that there is information
718: about both the eccentricity and its orientation in the full transit light curve \citep[see also][]{2009arXiv0901.0282B}. \label{eccexpect}
719:
720: Combining secondary transit timing information with radial velocity and
721: Rossiter-McLaughlin measurements to help constrain $\omega$,
722: \citet{2005ApJ...631.1215W} found the best-fit eccentricity for HD 209458
723: was $\sim$0.015. Though \citet{2005ApJ...631.1215W} argue that the actual eccentricity is
724: probably less than 0.01, it is not necessarily 0 \citep{2007MNRAS.382.1768M}.
725: Recently, \citet{2008arXiv0806.1478J} revealed WASP-14b, a young massive hot
726: Jupiter with an eccentricity of 0.1; WASP-10b and WASP-12b also appear to be
727: eccentric \citep{2008arXiv0806.1482C,2009ApJ...693.1920H}, though these
728: eccentricities may be spurious or overestimated.
729:
730: The most accurate eccentricity constraint is
731: a detection by \citet{2007Natur.447..183K} for the very hot Jupiter
732: HD189733b. They observed continuously and at high cadence (0.4
733: seconds) with the
734: Spitzer space telescope and measured a secondary timing offset
735: corresponding to $e \cos \omega = 0.001 \pm 0.0002$, a 5-$\sigma$
736: result that they could not explain by any other means. (Preliminary analysis
737: of additional data for this planet by \citet{2009IAUS..253..209A}
738: indicates $e \cos \omega = 0.0002 \pm 0.0001$.) The constraint on $e \sin \omega$ is
739: much weaker. A non-zero
740: eccentricity of $e \simeq 0.003$ for hot Jupiters is therefore
741: consistent with every
742: measurement available in the literature, though the actual values of eccentricities
743: at the $\e{-3}$ level are essentially unconstrained.
744:
745: In the absence of excitation, the current eccentricities of
746: these planets depend on
747: the initial eccentricity and the rate of eccentricity
748: decay. Extrapolating from planets in our solar system \citep{1966Icar....5..375G}
749: implies short circularization
750: timescales of $\simeq$ 10 MYr, though recent studies have shown that using a fixed eccentricity
751: damping timescale is an inappropriate simplification of the full tidal evolution \citep[e.g.][]{Jackson2008,2009ApJ...692L...9L,2009arXiv0903.0763R}.
752: Even an analysis using the full tidal evolution equations cannot give a compelling case
753: for the present-day eccentricities of these planets, since there are essentially no
754: direct constraints on the tidal dissipation parameter for the planet, $Q_p$. Various estimates
755: show that $Q_p$ for exoplanets is not known and may be quite large \citep[e.g.,][]{2008ApJ...686L..29M},
756: implying that non-zero eccentricities are not impossible. Even so, we stress
757: that the best candidates for observing apsidal precession are also those
758: planets that have the fastest eccentricity damping, since the damping timescale
759: and apsidal precession rates are both proportional to $\frac{M_p}{M_*} \left( \frac{a}{R_p} \right)^5$.
760: Hence, those planets which have the fastest precession rates will also have the lowest
761: eccentricities. The first step in determining if this trade-off allows for apsidal precession to be measured by \emph{Kepler} data
762: is to apply the techniques described in this paper to the data themselves.
763: Furthermore, with the discovery and long-term characterization of more planets using
764: ground and space-based observations, the detectability of apsidal precession
765: will increase dramatically.
766:
767: We should note that there are several mechanisms that can excite eccentricities and
768: compete with or overwhelm tidal dissipation. The most prevalent is assumed to be
769: eccentricity pumping by an additional companion
770: \citep{1979Sci...203..892P,2001ApJ...548..466B,2006ApJ...649.1004A}.
771: Even very small (Earth-mass or less) companions
772: in certain orbits can provide significant eccentricity
773: excitation \citep{2007MNRAS.382.1768M}. (In this case, however, our
774: single-planet method for estimating $k_{2p}$ would need to be modified considerably.) Tidal dissipation
775: in rapidly rotating stars tends to increase the eccentricity, potentially prolonging circularization
776: in some systems \citep{Ferraz-Mello2008}. Very distant inclined
777: companions (e.g. a planet orbiting a star in a misaligned binary star system) can
778: induce Kozai oscillations that impart very large eccentricities on secular timescales
779: \citep[e.g.,][]{2007ApJ...669.1298F}. \citet{2009arXiv0901.0735A} proposed that thermal tides can significantly
780: affect the orbital and rotational properties of extra-solar planets, though their
781: conclusions appear to be overestimated \citep{2009arXiv0901.3279G,2009arXiv0901.3401G}.
782: Finally, recent (not necessarily primordial) dynamical instabilities in the planetary system can also be
783: responsible for generating eccentricity which simply hasn't damped away yet
784: \citep{2005Natur.434..873F,2005Natur.435..466G,2007astro.ph..3166C,2008ApJ...675.1538T}.
785: We, therefore, continue our analysis under the possibility that some very hot Jupiters
786: may have non-zero eccentricities.
787:
788: \section{Transit Light Curves of Apsidal Precession}
789:
790: Previous studies of transit light curve variability due to
791: non-Keplerian perturbations have focused almost
792: exclusively on transit timing. In contrast, we model
793: the full photometric light curve in order to estimate the detectability of $k_{2p}$.
794: This will automatically include the effect of changing transit
795: durations, which are very useful for detecting apsidal
796: precession \citep{PK08,JB08}. In addition, using full photometry can
797: provide a more direct and realistic estimate of the detectability of
798: $k_{2p}$. Of course, the drawback is additional computational cost,
799: though we found this to
800: be manageable, requiring less than 20 seconds to generate the $\sim2$ million
801: photometric measurements expected from \emph{Kepler}'s 1-minute cadence over
802: 3.5 years.
803:
804: \subsection{Our Transit Light Curve Model}
805:
806: Determining the photometric light curve of a transiting system requires knowing
807: the relative positions of the star and the planet at all times. These can be
808: calculated
809: by describing the motion of the planet with time-varying osculating orbital
810: elements.
811: When describing the motion of the planet using instantaneous
812: orbital elements, it is usually customary to ignore the periodic terms
813: by averaging, as in \citet{1939MNRAS..99..451S}, and calculate only the secular terms. These
814: small periodic terms describe how
815: the orbital elements change within a single orbit as a function of the true anomaly, $f$,
816: due to the
817: non-Keplerian potential. In precessing systems, the value of the true anomaly at central
818: transit, $f_{tr} \equiv 90^{\circ}-\omega_{tr}$, changes subtly from one transit to the
819: next, inducing slow variations in the osculating orbital elements at transit.
820: Therefore, we include in our model the dominant
821: periodic changes in orbital elements
822: as a function of orbital phase, using $M_{tr} \approx f_{tr}$ as an appropriate
823: approximation for low eccentricities. Using a direct integration
824: (described in Section \ref{obliquitysec}), we
825: verified that ignoring these periodic variations can cause non-negligible systematic
826: errors in determining transit times. The periodic changes are
827: derived from the same disturbing potentials used above. We follow the
828: method of \citet{1959AJ.....64..367K} for calculating osculating elements from mean elements,
829: and assume zero obliquity. The
830: correction is similar to the correction to the mean motion,
831: which is also applied in our model.
832: The correction to the semi-major axis, eccentricity, longitude of periapse, and mean
833: anomaly are $a_{\rm osc} = a_{\rm mean} + \frac{2ae}{1-e^2} \epsilon \cos
834: M \approx 2ae \epsilon \cos M$, $e_{\rm osc} = e_{\rm mean} + \epsilon (1 - \cos M)$,
835: $\omega_{\rm osc} = \omega_{\rm mean} + \frac{\epsilon}{e} \sin M$,
836: and $M_{\rm osc} = M_{\rm mean} - \frac{\epsilon}{e} \sin M$ where $\epsilon$
837: is defined in Equation \ref{epsilon}. General
838: relativistic periodic corrections are also added; these are taken
839: from \citet{1989racm.book.....S}, page 92 (with
840: $\alpha=0$, $\beta=\gamma=1$). Using our direct integrator (described below), we
841: verified that these corrections reproduced the actual orbit to
842: sufficient accuracy for this analysis as long as $e \gg \epsilon \sim
843: \e{-5}$. Other corrections
844: are higher order in small parameters and are ignored.
845:
846: Our model uses these corrected elements to generate astrocentric Cartesian
847: coordinates for a specific system inclination and, for completeness, also
848: includes the effect of
849: light-travel time \citep{2005ApJ...623L..45L} though we concur with
850: \citet{JB08} and \citet{PK08} that the
851: light-travel time change due to $\dot{\omega}$ is unimportant. The positions
852: are then translated to photometric light curves using the quadratic
853: limb-darkening code\footnote{Available at
854: \texttt{http://www.astro.washington.edu/
855: agol/transit.tar.gz}}
856: described in \citet{2002ApJ...580L.171M}. \emph{Kepler} data will have enough signal-to-noise to justify
857: using non-linear limb darkening laws
858: \citep{2007ApJ...655..564K}, but we do not expect that this
859: simplification will significantly alter our conclusions.
860:
861: In addition, we include the photometry of the secondary eclipse. As
862: suggested by \citet{2007ApJ...667L.191L}, very hot
863: Jupiters can reach temperatures exceeding 2000 K, where their blackbody
864: emission at optical wavelengths is detectable by \emph{Kepler}.
865: This thermal emission is added to the
866: reflected light of the planet, which appears to be small based on the
867: low upper limit of the albedo of HD 209458b and TrES-3 measured by
868: \citet{2007arXiv0711.4111R} and \citet{2008AJ....136..267W}, respectively.
869: We find that in \emph{Kepler}'s observing bandpass of 430-890 nm
870: \citep{2006Ap&SS.304..391K}, thermal emission of
871: very hot Jupiters can dominate over the weak reflected light.
872: We estimate the depth
873: of the secondary eclipse ($d_{\rm sec}$) in our simulated
874: \emph{Kepler} data by assuming that 1\% of the light is reflected and
875: the other 99\% absorbed and reemitted as processed thermal blackbody
876: emission from the entire planetary surface (day and night sides). To
877: be conservative and to account for unmodeled non-blackbody
878: effects, we divide the resulting planet/star flux
879: ratio by 2 \citep{2008arXiv0807.1561H}; the resulting depth of around
880: $2 \x \e{-4}$ is consistent with the lower values of
881: \citet{2008arXiv0803.2523B}, the tentative measurement of the
882: thermal emission from CoRoT-2b \citep{2009IAUS..253...91A}, and the detection of
883: secondary eclipse emission from OGLE-TR-56b \citep{2009A&A...493L..31S}.
884: We note that the best candidates for
885: detecting $k_{2p}$
886: are those with small semi-major axes and large radii; these same
887: planets have relatively large $d_{\rm sec}$ values (Table 1). Secondary
888: eclipses are very useful for determining $e$ and $\omega$. We will
889: also find that they can be important for observing apsidal precession.\label{secdepth}
890:
891: Our model generates accurate photometry for an extra-solar planet
892: undergoing apsidal precession. Several
893: other small photometric effects have been discussed in the literature,
894: which we do not include. Most of these effects are periodic (e.g. the
895: reflected light curve) and
896: therefore will not affect the long-term trend of precession. Care will need
897: to be taken to ensure that slow changes due to parallax and proper motion, which should
898: be quite small for relatively distant stars observed by \emph{Kepler}
899: \citep{2008arXiv0807.0008R,2007ApJ...661.1218S} or changes
900: in the stellar photosphere \citep{2008arXiv0807.0835L} are not significant.
901: Non-Gaussian astrophysical noise of
902: the star and other systematic
903: noise should degrade the accuracy with
904: which $k_{2p}$ can be measured compared to our ideal photometry. The long-term variability of
905: the star can be interpolated away or modeled \citep{2009A&A...493..193L}, though it is not
906: clear how short-term variability will affect transit light curves at \emph{Kepler}'s level
907: of precision. On the other hand, complimentary observations (e.g., warm Spitzer, HST,
908: radial velocities, JWST, etc.) should only enhance our understanding of the
909: systems studied.
910:
911: \subsection{Accuracy of $k_{2p}$ measurement}
912: With an accurate photometric model of apsidal precession, one could estimate the
913: measurement accuracy of $k_{2p}$
914: from \emph{Kepler} data by carrying out a full Monte Carlo study of the inversion
915: problem, going from realistic synthetic photometric data sets to a determination of all
916: system parameters. In this work, instead, we carry out a much
917: simpler calculation which cannot provide strict one-sigma error estimates
918: like the Monte Carlo analysis, but does give an indication of
919: how well $k_{2p}$ can be resolved given a large dataset.
920:
921: We obtain this accuracy estimate by comparing a realistic precessing
922: photometric model with $k_{2p} \ne 0$ to a base model with $k_{2p}= 0$. The
923: base model is still undergoing very slow apsidal precession, induced by
924: general relativity and $k_{2*}$. We
925: calculate the effect of a non-zero $k_{2p}$ value by subtracting the precessing
926: model from the base model. (See Figures \ref{candywrappermain} and
927: \ref{bowtiemain}.) Then, by calculating the root-sum-square of the residual
928: signal and comparing it to the photometric error on a single data
929: point, we obtain a numerical measure of the relative signal induced by
930: $k_{2p}$. The ``signal-to-noise'' ratio for the data set
931: is therefore given by:
932: \be
933: \frac{\rm S}{\rm N} \sim \frac{\sqrt{\sum_i (y_i -
934: y_i^0)^2}}{\sigma}
935: \ee
936: where $y_i$ and $y_i^0$ are the photometry model values for the $k_{2p}$
937: test model and the base model, respectively, and $\sigma$ is the
938: photometric error. We use $\sigma = 1000$ parts per million (ppm) flux per 1-minute integration,
939: corresponding to the expected noise of
940: \emph{Kepler} on a faint $V=14$ star
941: \citep{2006Ap&SS.304..391K}. Of the 30 planets with periods less than 3 days, 16 are expected
942: to be brighter than $V \simeq 14$ (T. Beatty, pers. comm.) and we can reasonably expect
943: some fraction of these to have
944: orbits comparable to the planets modeled here.
945:
946: Since our residual signal changes
947: as a function of time, this is not a true signal-to-noise calculation;
948: the distribution of values in time matters for a proper interpretation, but any
949: distribution would
950: yield the same effective $\frac{S}{N}$, and thus this construction is
951: not capturing all of the details. Even so, it does provide a
952: useful and reasonable rough estimate for detectability. In order to
953: identify the resolution on the $k_{2p}$ measurement, we search for the
954: value of $k_{2p}$ which yields a signal-to-noise of $\frac{S}{N} =
955: 1$. This is reasonable since it represents the threshold value of
956: $k_{2p}$, below which planetary induced precession cannot be distinguished
957: in the data with the given errors. The threshold $k_{2p}$ value can also be
958: loosely thought of as an estimate of the 1-$\sigma$ expected errors. \label{threshktwop}
959:
960: \begin{figure}
961: \begin{center}
962: \includegraphics[width=3.5in]{f2}
963: \caption{\label{candywrappermain}
964: \textbf{Photometric Difference Signal from $k_{2p}$.}
965: As described in the text, we use the difference between two
966: theoretical light curves in the
967: transit photometry to assess the observability of apsidal precession by
968: \emph{Kepler}. For WASP-4b at $\omega=0^{\circ}$, $e=0.003$, and a
969: central impact parameter, the difference between a model
970: with $k_{2p}=0$ and $k_{2p}=0.146$ would yield an effective
971: ``signal-to-noise'' of 1 on a moderately bright star
972: ($V=14$). Shown is this difference signal; the root sum of
973: squares of the signal is equal to 1000 ppm, the expected
974: photometric accuracy of \emph{Kepler} for a 1 minute
975: observation \citep{2006Ap&SS.304..391K}. The trends seen in the figure are
976: illustrated in Figure \ref{candywrapperpieces}
977: by considering excepts of single primary transits from
978: the regions labeled 1-5.
979: }
980: \end{center}
981: \end{figure}
982:
983: \begin{figure}
984: \begin{center}
985: \includegraphics[width=3.5in]{f3}
986: \caption{\label{candywrapperpieces}
987: \textbf{Excerpts of Photometric Difference Signal.}
988: Examining excerpts of the residual signal shown fully in Figure
989: \ref{candywrappermain}, the effects of transit timing and
990: ``transit shaping'' can both be seen. The five excerpts are offset
991: for clarity.
992: Transit timing has an
993: asymmetric signal (dotted lines), obtained when
994: subtracting two transit curves
995: slightly offset in time. Transit shaping, which is mostly due to changing
996: transit duration, creates a symmetric signal (dashed
997: lines). The total difference signal (solid lines) is
998: dominated by the effect of transit shaping, which has
999: $\sim$30 times more signal than transit timing alone. (See
1000: explanation in text.) Both effects are maximized at the
1001: beginning (1) and end (5), as
1002: expected for a signal that increases with longer
1003: baseline. The maximal signal occurs during ingress and
1004: egress, when the light curve changes the fastest. The transit
1005: shapes are equivalent at the center
1006: (3) by construction. The transit timing anomaly of precession is quadratic,
1007: which, when fitted with a best-fit straight line
1008: corresponding to a non-precessing signal, yields two
1009: intersections when
1010: transit timing is minimized (2,4). The transit timing offset
1011: at the beginning and end is only 0.085 seconds, while the
1012: center is offset by -0.042 seconds.
1013: }
1014: \end{center}
1015: \end{figure}
1016:
1017: \begin{figure}
1018: \begin{center}
1019: \includegraphics[width=3.5in]{f4}
1020: \caption{\label{bowtiemain}
1021: \textbf{Photometric Difference Signal from $k_{2p}$.}
1022: Similar to Figure \ref{candywrappermain}, but for
1023: $\omega=90^{\circ}$. This figure is dominated by the
1024: photometric difference between secondary transits slightly offset in
1025: time. At
1026: $\omega=90^{\circ}$ the changes in the primary
1027: transits due to precession are small, except far away from the central time.
1028: At this orientation, the primary-secondary timing offset
1029: (Equation \ref{heqoffset}) is
1030: maximized. This ``secondary transit timing'' signal is
1031: weaker than the signal from primary transit as the secondary
1032: transit depth is much shallower. Therefore, an unreasonably high $k_{2p}$ of 0.925 is required to detect the apsidal precession.
1033: Excerpts of single secondary
1034: transits taken from regions labeled 1-5 are shown in Figure
1035: \ref{bowtiepieces}.
1036: }
1037: \end{center}
1038: \end{figure}
1039:
1040: \begin{figure}
1041: \begin{center}
1042: \includegraphics[width=3.5in]{f5}
1043: \caption{\label{bowtiepieces}
1044: \textbf{Excerpts of Photometric Difference Signal.}
1045: Similar to Figure \ref{candywrapperpieces}, but for
1046: $\omega=90^{\circ}$. Single secondary transit differences
1047: are excised
1048: from the full difference signal shown in Figure
1049: \ref{bowtiemain}. The shape of the curves is due to the
1050: subtraction of two secondary transits slightly offset in
1051: time. Since the secondary transits are complete
1052: occultations, they are flat-bottomed and lack the additional
1053: structure due to limb-darkening seen in Figure
1054: \ref{candywrapperpieces}. By construction, the offset grows
1055: in time away from the
1056: center (3) of the signal and attains a maximum at the
1057: beginning (1) and end (5). Curves 2 and 4 are shown for
1058: comparison to Figure \ref{candywrapperpieces}.
1059: }
1060: \end{center}
1061: \end{figure}
1062:
1063: This is a realistic estimate only insofar as the residual signal
1064: ($y_i-y^0_i$) is due only to $k_{2p}$ and cannot be absorbed by any
1065: other parameters. Hence we seek to choose other parameters so as to
1066: minimize the residuals without changing $k_{2p}$. For most system
1067: parameters, this is accomplished by referencing the time to the center
1068: of the data set, and thus the difference between the signals grows
1069: similarly forward and backward in time as seen in Figures
1070: \ref{candywrappermain} - \ref{bowtiepieces}. The transit shapes in
1071: both models are equivalent at the center of the dataset as would be
1072: expected in an analysis of actual data.
1073:
1074: Additionally, a major \label{degeneracies}
1075: effect from changing the precession period is to alter the observed
1076: average period. When analyzing actual data, this would just be
1077: absorbed into a small adjustment to the (unknown) stellar mass, thereby
1078: adjusting the period to absorb much of the $k_{2p}$ signal. It is therefore
1079: important to correct for the average period change to avoid
1080: significantly overestimating the signal due to $k_{2p}$. Additionally,
1081: there is a similar, though less severe, effect for the epoch of the
1082: first transit, which is also adjusted to best absorb signal. This is
1083: achieved by using an analytic expression for the transit times (see
1084: Equation \ref{timeanombig} below) which match the transit times of the
1085: photometric model to very high accuracy. By fitting a line to these times, we
1086: can determine the average period and epoch that absorb the
1087: degenerate portions of the $k_{2p}$ signal, leaving behind the
1088: residual due only to $k_{2p}$. We have not explicitly accounted for
1089: degeneracies between the
1090: signal from $k_{2p}$ and the other parameters, like the radius, limb
1091: darkening, and system inclination, but since $k_{2p}$ induces a time varying
1092: signal while these other parameters are generally constant, there is
1093: little expected signal absorption from these parameters.
1094:
1095: The only
1096: major drawback of this approach is that it does not allow the
1097: eccentricity state of the system to change. With real data, the
1098: eccentricity and precession phase are not known in advance, and thus
1099: must be found by inversion. As detailed in Section \ref{eccexpect},
1100: eccentricity and orbital orientation are primarily constrained by
1101: comparing primary and secondary transit pairs, and thus proper inversion is
1102: greatly aided by accurate observations in
1103: wavelengths more favorable to secondary transit
1104: observations, obtained by Spitzer, HST, or from the ground
1105: \citep[e.g.][]{2007Natur.447..183K,2008Natur.452..329S,2008arXiv0806.4911G}.
1106: We also find that binned and folded \emph{Kepler} data has comparable sensitivity to
1107: a single Spitzer observation for characterizing the secondary eclipses of very hot Jupiters.
1108: In any case, our assessment of the threshold $k_{2p}$ assumes that the eccentricity
1109: of the system is very well known, which will likely require additional supporting observations.
1110:
1111: \subsection{Comparison to Expected Signal}
1112:
1113: The residual light curves calculated for each planet, Figures
1114: \ref{candywrappermain} - \ref{bowtiepieces}, match the theoretical
1115: expectations of the apsidal precession signal \citep{2002ApJ...564.1019M,2007MNRAS.377.1511H,PK08,JB08}.
1116: To interpret the results of our
1117: analysis, it will be useful to briefly review
1118: the major components of the apsidal precession signal: changes in the times of
1119: primary transits, changes in the shape of primary transits, and changes in the
1120: primary-secondary offset times \citep{2002ApJ...564.1019M,2007MNRAS.377.1511H,PK08,JB08}.
1121:
1122: The primary transit times, $T_N$, due to apsidal precession are
1123: well described by a sinusoid for very low eccentricities ($e \ll 0.1$):
1124: \be\label{timeanom}
1125: T_N = T_0 + NP_{\rm obs}+\frac{eP_{\rm obs}}{\pi}(\cos \omega_{tr,N}-\cos\omega_{tr,0})]
1126: \ee
1127: where $T_0$ is the epoch of the first transit, ${\omega_{tr,N} \equiv \dot{\omega}(T_N-T_0) + \omega_{tr,0}}$ is the
1128: argument of periapse for the $N^{\rm th}$ transit,
1129: and $P_{\rm obs}$ is the \emph{observed} period between successive transits, which deviates
1130: from the actual orbital period since the orbit has precessed a small
1131: amount between transits \citep{1973bmss.book.....B}. For small eccentricities,
1132: the amplitude of the transit timing variations due to $k_{2p}$ is:
1133: \be\label{timingamp}
1134: \frac{eP_{\rm obs}}{\pi} \simeq 119 \textrm{ sec} \x \left( \frac{e}{0.003} \right) \left(
1135: \frac{a}{0.025 \textrm{ AU}} \right)^{3/2} \left(
1136: \frac{M_*}{M_{\astrosun}} \right)^{-1/2}
1137: \ee
1138: Given that individual transit times can be measured with accuracies of only a few seconds,
1139: even tiny eccentricities $e \lesssim \e{-5}$ can induce detectable transit timing variations on precessional
1140: timescales ($\sim\dot{\omega}^{-1}$).
1141:
1142: For our analysis, we extended Equation \ref{timeanom} to fifth order in eccentricity allowing accurate determination of transit
1143: times for eccentricities up to of order 0.1. We also require a correction for the effect of a non-central impact
1144: parameter ($i < 90^{\circ}$, $e > 0$). For an inclined eccentric orbit, the apparent path of the planet across the
1145: stellar disk is curved. At orientations where the line
1146: of sight is not along the major axis of the ellipse, the curved path is also asymmetric. Therefore, the times of photometric minima, $T_N$,
1147: do not correspond exactly to the times of conjunction (when the planet crosses the $y-z$ plane and $f_{tr} \equiv 90^{\circ} -\omega_{tr}$).
1148: We follow the correction from Equation VI.9-21 of \citet{1959cbs..book.....K}, who find that at photometric
1149: minimum, $f_{tr} = 90^{\circ} - \omega'_{tr}$, where ${\omega'_{tr} \equiv
1150: \omega_{tr} + e \cos \omega_{tr} \cot^2(i) (1 - e \sin \omega_{tr} \csc^2(i))}$; in this corrective term, it is only required to keep terms up to second order in
1151: eccentricity. Assuming that $i$ and $\dot{\omega}$
1152: are constant, it can be shown that
1153: \begin{eqnarray}\label{timeanombig}
1154: T_N & = & T_0 + NP_{\rm obs} \nonumber \\
1155: & + & \frac{P_{\rm obs}}{\pi} \Bigg[ e(\cos \omega'_{tr,N} - \cos \omega'_{tr,0}) \nonumber \\
1156: & + & \frac{3}{8} e^2 (\sin 2\omega'_{tr,N} - \sin 2\omega'_{tr,0}) \nonumber \\
1157: & + & \frac{1}{6} e^3 (\cos 3\omega'_{tr,N} - \cos 3\omega'_{tr,0}) \nonumber \\
1158: & + & e^4 \Big( \frac{1}{16} (\sin 2\omega'_{tr,N} - \sin 2\omega'_{tr,0}) \nonumber \\
1159: & - & \frac{5}{64} (\sin 4\omega'_{tr,N} - \sin 4\omega'_{tr,0}) \Big) \nonumber \\
1160: & + & e^5 \Big( \frac{1}{16} (\cos 3\omega'_{tr,N} - \cos 3\omega'_{tr,0}) \nonumber \\
1161: & - & \frac{3}{80} (\cos 5\omega'_{tr,N} - \cos 5\omega'_{tr,0}) \Big) \Bigg] \nonumber \\
1162: \end{eqnarray}
1163: This transcendental equation is solved iteratively for ($T_N - T_0$) to obtain the transit times and has been tested thoroughly against the empirical
1164: determination of transit times calculated by our light curve model described above.
1165:
1166: The expected apsidal precession periods
1167: (including small contributions from GR and the star) for
1168: WASP-12b, CoRoT-1b, OGLE-TR-56b, WASP-4b, and TrES-3b are around
1169: 18, 71, 116, 120, and 171 years, respectively. In other words, they have precession rates induced
1170: by the planetary tidal bulge of a few degrees per
1171: year, compared to a few degrees per century as the fastest general
1172: relativistic precession \citep{JB08}. We caution that if $\frac{R_p}{a}$ for WASP-12b is
1173: overestimated due to imprecise data \citep[e.g.][]{2007AJ....134.1707W}, then
1174: the precession period would increase accordingly.
1175:
1176: Even with such fast precession rates, the duration of observations will generally be much shorter than the
1177: precession period. In addition, as discussed above, the linear timing anomalies
1178: will be absorbed into the effective period as a small change in the unknown stellar mass \citep{2007MNRAS.377.1511H,PK08,JB08}.
1179: Therefore, detection of apsidal precession from primary transit times alone will
1180: require a significant detection of the curvature over a small portion of a long-period
1181: sinusoid. Since the curvature in Equation \ref{timeanom} is maximal at
1182: $\omega \approx 0,180^{\circ}$, these orientations have the best primary
1183: transit timing signal. Even at these orientations, detecting $k_{2p}$ from primary transit
1184: times alone is difficult, since it can be shown that the signal strength is proportional to $e
1185: \dot{\omega}^2$, due to the need to detect curvature \citep{2007MNRAS.377.1511H}.
1186:
1187: When the observational baseline is much shorter
1188: than the decades-long precession period, utilizing the changing shape of the
1189: transits can significantly improve detectability of apsidal precession \citep{PK08,JB08}.
1190: Transit shapes are primarily determined by the orbital speed at transit $\dot{f}_{tr}$ and impact parameter
1191: $b$, both of which depend on the precession phase $\omega_{tr}$.
1192: For small eccentricities, the orbital angular speed at transit is given simply by
1193: $\dot{f}_{tr} \simeq n(1+2e \cos \omega_{tr})$.
1194: Changes in the impact parameter are somewhat more subtle, since $b$ is given by
1195: $r_{tr} \cos i / R_*$, where $r_{tr} \simeq a(1-e^2)/(1 + e \sin \omega_{tr})$ is the star-planet separation. Hence, the apparent
1196: impact parameter of the planet can change for non-central transits, even when the orbital plane
1197: remains fixed. The evolving transit shape of precessing orbits is determined by variations
1198: in both orbital speed and impact parameter. Simplifying the effect of transit shape by considering only the variations in transit duration as a
1199: function of $\omega_{tr}$, \citet{PK08} and \citet{JB08} find that these two effects are of comparable magnitude.
1200: These authors also show analytically that the two effects exactly cancel when $b=1/\sqrt{2}$. At this
1201: impact parameter, the transit duration stays constant throughout apsidal precession. The full photometric
1202: transit shape, however, still changes detectably in a precessing orbit, though the magnitude of
1203: signal is reduced (Figure \ref{k2refangfig}).
1204:
1205: The expected effect of changing transit shapes is fully consistent with
1206: the photometric difference signals calculated by our model (Figures \ref{candywrappermain} and \ref{candywrapperpieces}).
1207: Indeed, our model shows that transit shaping dominates the signal by a factor of
1208: $\gtrsim$30 (Figure \ref{candywrapperpieces}). We can also see that changes in the transit shape are maximized at orientations near
1209: $\omega \approx 0,180^{\circ}$ (as expected from Equation \ref{keqdurratio}). \label{transitshapes}
1210:
1211: For small eccentricities, the transit shaping
1212: signal strength is given by $\frac{S}{N} \propto e \dot{\omega} \propto e k_{2p}$.
1213: Therefore, when transit shaping dominates the observable signal, we should find that
1214: searching for the threshold $k_{2p}$ value that yields $\frac{S}{N} = 1$ results
1215: in a power law relationship between threshold $k_{2p}$ and $e$, such that $k_{2p} \propto e^{-1}$.
1216: By solving for threshold $k_{2p}$ for eccentricities from 0.001 to 0.1, we
1217: find, as expected, that threshold $k_{2p}$ very
1218: closely follows a power law in eccentricity with a slope of -1 for all planets. This power law
1219: relationship can be written as $ek_{2p} = C$, where $C$ is a constant calculated from our model that
1220: depends on the planetary, orbital, and stellar parameters of the system.
1221:
1222: At $\omega \approx 90,270^{\circ}$, transit timing and transit shaping effects are much weaker and
1223: are rather ineffective at constraining apsidal precession.
1224: At these orientations (when the Earth's line of sight is nearly aligned with the major axis of the orbit),
1225: another photometric signal emerges: variations in the difference between the times
1226: of primary and secondary transits. The changing
1227: orientation of the orbital ellipse causes a variation in the offset
1228: between primary and secondary transit times following
1229: Equation \ref{heqoffset} above \citep{2007MNRAS.377.1511H,JB08}. These authors show that
1230: the strength of this signal is also proportional to $e \dot{\omega}$ and we find that the variation
1231: in threshold $k_{2p}$ then also follows $k_{2p} \propto e^{-1}$.
1232:
1233: The photometric difference signal at $\omega=90^{\circ}$ is shown in Figures
1234: \ref{bowtiemain} and \ref{bowtiepieces}. Using the method described in Section \ref{degeneracies} to remove degeneracies
1235: almost eliminates the primary transit signal entirely, as expected, and the secondary
1236: transit offset becomes the more powerful signal. For WASP-12b, with an expected
1237: \emph{Kepler} secondary transit depth of $\sim$1830 ppm, the
1238: threshold $k_{2p}$ is actually \emph{lower} at $\omega=90^{\circ}$ (Figure
1239: \ref{ek2planet}). For the other planets, the secondaries are not as important.
1240:
1241: Our estimates of threshold $k_{2p}$ at $\omega=90^{\circ}$ are based on the unknown secondary transit depth ($d_{\rm sec}$) in the
1242: \emph{Kepler} bandpass (though our estimates of $d_{\rm sec}$ are consistent with all the measurements
1243: in the literature to date). Furthermore, we find that $\frac{S}{N} \propto d_{\rm sec}$, so that deeper secondary
1244: transits improve the accuracy with which $k_{2p}$ can be measured. It is important to note that combining \emph{Kepler}
1245: primary transit times with precise secondary transit times
1246: measured in the near-infrared (e.g. by warm Spitzer, HST, or JWST) is a very powerful way to constrain apsidal
1247: precession \citep{2007MNRAS.377.1511H} for any orientation. Even a few high-precision secondary eclipse observations
1248: are enough to lower the value of threshold $k_{2p}$ from our predictions, especially when $\omega \approx 90,270^{\circ}$.
1249:
1250: By construction, threshold $k_{2p}$ values vary linearly with the
1251: assumed photometric error $\sigma = 0.001 \x 10^{0.2(V-14)}$. In addition,
1252: re-performing our analysis using a 6-year long \emph{Kepler}
1253: mission improved threshold $k_{2p}$ values by a common factor of $\sim$2.2.
1254:
1255: \begin{figure*}
1256: \begin{center}
1257: \includegraphics[width=6in]{f6}
1258: \caption{\label{ek2planet}
1259: \textbf{Eccentricities Needed to Detect Interior Properties
1260: from Apsidal Precession.}
1261: The best-known planets for detecting $k_{2p}$
1262: precession are analogs to the hot Jupiters WASP-12b, WASP-4b, CoRoT-1b,
1263: OGLE-TR-56b, TrES-3b, HAT-P-7b, TrES-2b, and WASP-14b. Assuming that
1264: analogs to these planets exist in the \emph{Kepler} field around a
1265: V=14 magnitude star, the above graph shows the eccentricities
1266: required to detect $k_{2p}$. Black
1267: symbols correspond to calculations with $\omega=0^{\circ}$ and gray symbols
1268: correspond to $\omega=90^{\circ}$; in both cases, $b=0$. Apsidal precession is much easier to detect
1269: for larger eccentricities so increasing $e$ decreases the
1270: detectable $k_{2p}$. Using our transit light curve model,
1271: we found that threshold $k_{2p}$ values followed a power law
1272: $k_{2p} \propto e^{-1}$ (for low eccentricities), which is consistent
1273: with the analytical estimates that $\frac{S}{N} \propto e\dot{\omega} \propto
1274: ek_{2p}$ (see Section \ref{threshktwop}). Interpolating (and sometimes
1275: extrapolating) on this power law relationship, the graph identified the
1276: eccentricities required of these analog planets to detect precession
1277: due to a ``typical'' planetary interior of $k_{2p}=0.3$ (triangles).
1278: For example, when $e=0.00026$ and $\omega=0^{\circ}$,
1279: the apsidal precession due to an analog of WASP-12b should be just
1280: detectable by \emph{Kepler}. A higher eccentricity (shown in Table 1) would be
1281: needed to measure $k_{2p}$ with sufficient accuracy (0.1) to
1282: distinguish between a massive core and a core-less model
1283: (circles). Systematic errors are expected to become
1284: important once the measurement error on $k_{2p}$ reaches as
1285: low as 0.01 (squares). If any of the very hot
1286: Jupiters discovered by \emph{Kepler} have
1287: comparable eccentricities, the long-term high-precision
1288: photometry would allow for a powerful probe into their interior structure. HAT-P-7b and TrES-2b
1289: are known to lie in the \emph{Kepler} observing field, but the values
1290: above are not corrected for improved photometric accuracy obtainable on these bright stars. Note that
1291: the eccentricities shown above and in Table 1 are computed for $\frac{S}{N}=1$; 3-$\sigma$ measurements
1292: require eccentricities 3 times as high.
1293: }
1294: \end{center}
1295: \end{figure*}
1296:
1297: \begin{figure}
1298: \begin{center}
1299: \includegraphics[width=3.5in]{f7}
1300: \caption{\label{k2refangfig}
1301: \textbf{Effect of Impact Parameter on Precession Signal.}
1302: The detectability of apsidal precession depends on the
1303: impact parameter ($b$) of the orbital track across the
1304: star. For $\omega=0^{\circ}$ (solid), the
1305: signal of primary
1306: transits are most important, with transit shaping playing
1307: the largest role. (See Figure
1308: \ref{candywrapperpieces}.) However,
1309: the strength of transit shaping is a function of impact
1310: parameter with the minimum effect analytically estimated by \citet{JB08} and \citet{PK08} to be
1311: $b=1/\sqrt{2}$ (vertical solid line). Using a full photometric
1312: model, we see the expected decrease in the shaping signal
1313: (i.e. requiring a larger $k_{2p}$ to reach
1314: $\frac{S}{N}=1$). Note that the signal is nearly maximal, with small threshold $k_{2p}$ values, for a large
1315: range of impact parameters. When $\omega=90^{\circ}$ (dotted), the
1316: effect of primary transits are minimal and the offset in
1317: secondary transits become the determining factor. (See
1318: Figure \ref{bowtiemain}.) At high
1319: impact parameters secondary eclipses are grazing, reducing
1320: the observable signal. We also show the threshold $k_{2p}$ for an orientation of
1321: $\omega=45^{\circ}$, which lies, as expected, between the two extremes. The values
1322: of threshold $k_{2p}$ shown are for an V=14 CoRoT-1b analog in the \emph{Kepler} field
1323: with an eccentricity of 0.003.
1324: }
1325: \end{center}
1326: \end{figure}
1327:
1328:
1329: %\begin{tabular}{cccccccccccc}[8 pt]
1330:
1331: %\begin{deluxetable*}{lcccccccccccl}
1332: \begin{deluxetable*}{lrrrrrrrccccl}
1333: \label{paramstable}
1334: \tabletypesize{\scriptsize}
1335: %\tablewidth{0pt}
1336: \tablecaption{Extra-Solar System Parameters and Results}
1337:
1338: \tablehead{
1339: \colhead{Planet Analog} & \colhead{$M_*$} & \colhead{$R_*$} &
1340: \colhead{$M_p$} & \colhead{$R_p$} & \colhead{$a$} &
1341: \colhead{$d_{\rm sec}$\tablenotemark{b}} & \colhead{$\dot{\omega}_{tot}$} &
1342: \multicolumn{2}{c}{$e$ (Threshold $k_{2p}$=0.1)\tablenotemark{c} } &
1343: \colhead{Threshold $\dot{P}$\tablenotemark{d}} & \colhead{Threshold $Q_*$\tablenotemark{d}} & \colhead{Ref} \\
1344: \colhead{} & \colhead{$M_{\astrosun}$} & \colhead{$R_{\astrosun}$} &
1345: \colhead{$M_{J}$} & \colhead{$R_{J}$\tablenotemark{a}} &
1346: \colhead{AU} & \colhead{ppm} & \colhead{$^{\circ}$/yr} &
1347: \colhead{$\omega=0^{\circ}$} & \colhead{$\omega=90^{\circ}$} &
1348: \colhead{ms/yr} & & }
1349:
1350: \startdata
1351:
1352: WASP-12b & 1.35 & 1.57 & 1.41 & 1.79 & 0.0229 & 1830 & 19.9 & 0.0008 & 0.0004 & 0.95 & 92700 & 1 \\
1353: CoRoT-1b & 0.95 & 1.11 & 1.03 & 1.55 & 0.0245 & 314 & 4.96 & 0.0028 & 0.0085 & 0.93 & 12500 & 2,3 \\
1354: WASP-4b & 0.92 & 0.91 & 1.24 & 1.36 & 0.0234 & 109 & 2.91 & 0.0047 & 0.0394 & 0.68 & 9900 & 4 \\
1355: TrES-3b & 0.93 & 0.83 & 1.91 & 1.34 & 0.0228 & 106 & 2.04 & 0.0062 & 0.0614 & 0.53 & 13700 & 5 \\
1356: OGLE-TR-56b & 1.17 & 1.32 & 1.29 & 1.30 & 0.0236 & 451 & 3.00 & 0.0077 & 0.0096 & 1.36 & 24700 & 6 \\
1357: HAT-P-7 b & 1.47 & 1.84 & 1.77 & 1.36 & 0.0377 & 176 & 0.25 & 0.2085 & 0.3146 & 6.73 & 2800 & 7 \\
1358: TrES-2 b & 0.98 & 1.00 & 1.19 & 1.22 & 0.0367 & 18 & 0.13 & 0.2102 & \nodata\tablenotemark{e} & 2.94 & 350 & 8 \\
1359: WASP-14b & 1.21 & 1.31 & 7.34 & 1.28 & 0.0360 & 144 & 0.09 & 0.8352\tablenotemark{e} & \nodata\tablenotemark{e} & 3.92 & 5400 & 9 \\
1360: XO-3 b & 1.21 & 1.37 & 11.8 & 1.22 & 0.0454 & 46 & 0.04 & \nodata\tablenotemark{e} & \nodata\tablenotemark{e} & 8.00 & 1700 & 10 \\
1361: HAT-P-11b & 0.81 & 0.75 & 0.081 & 0.42 & 0.0530 & 0.2 & 0.01 & \nodata\tablenotemark{e} & \nodata\tablenotemark{e} & 29.2 & 0.1 & 11 \\
1362: CoRoT-7b & 0.91 & 1.02 & 0.028 & 0.16 & 0.0170 & 8 & 0.29 & \nodata\tablenotemark{e} & \nodata\tablenotemark{e} & 16.8 & 80 & 12 \\
1363:
1364: \enddata
1365:
1366: \tablerefs{
1367: (1) \citet{2009ApJ...693.1920H} \quad
1368: (2) \citet{2009arXiv0903.1845B} \quad
1369: (3) \citet{Barge2008} \quad
1370: (4) \citet{2009AJ....137.3826W} \quad
1371: (5) \citet{2009ApJ...691.1145S} \quad
1372: (6) \citet{2007AA...465.1069P} \quad
1373: (7) \citet{2009IAUS..253..428P} \quad
1374: (8) \citet{2007ApJ...664.1185H} \quad
1375: (9) \citet{2009MNRAS.392.1532J} \quad
1376: (10) \citet{2008ApJ...677..657J} \quad
1377: (11) \citet{2009arXiv0901.0282B} \quad
1378: (12) \texttt{www.exoplanet.eu}\tablenotemark{f}
1379: }
1380:
1381: \tablecomments{These system parameters were used to estimate the detectability of apsidal precession for these very hot Jupiter systems. The derivation
1382: of the values in the remaining columns is described in the text and in the footnotes below. For all systems, $k_{2*} = 0.03$ and quadratic limb
1383: darkening parameters $u_1=0.35$ and $u_2=0.4$ (appropriate for \emph{Kepler}'s bandpass) were used \citep{2002ApJ...580L.171M}. For reference, the
1384: measured eccentricity of WASP-12b, WASP-14b, HAT-P-11b, and XO-3b are 0.049 $\pm$ 0.015, 0.091 $\pm$ 0.003, 0.198 $\pm$ 0.046, and 0.2884 $\pm$ 0.0035 respectively. Other planets
1385: have unmeasured eccentricities or eccentricity upper limits of $\lesssim$0.05. A discussion of these results is provided in Section \ref{resultsdisc}.}
1386:
1387: \tablenotetext{a}{We use $R_{J}\equiv71492$ km, the equatorial radius at 1 bar.}
1388: \tablenotetext{b}{The estimated depth of the secondary transit in \emph{Kepler}'s bandpass (see Section \ref{secdepth}).}
1389: \tablenotetext{c}{The eccentricity required (at two different values
1390: of $\omega$) so that a $k_{2p}$ difference of 0.1 has an effective
1391: signal-to-noise of 1 in all of \emph{Kepler} data for a V=14 star, corresponding
1392: to a photometric accuracy of 1000 ppm/min. If analogs to these planets were
1393: found by \emph{Kepler} with the given eccentricities, the internal density distribution would
1394: be measured well enough to detect the presence of a large core (see Section \ref{threshktwop}). These values
1395: correspond to the circles in Figure \ref{ek2planet}. These results are for central transits (for $b > 0$, see Figure \ref{k2refangfig}). }
1396: \tablenotetext{d}{The value of the change in period, $\dot{P}$, that can be detected with a signal-to-noise of 1 in all of \emph{Kepler} data for a V=14 star (see Section \ref{threshpdot}). The value of threshold $Q_*$ is an estimate of the maximum value of the stellar tidal dissipation parameter, $Q_*$, assuming that the period decay is due entirely to tidal evolution
1397: of the planet. Lower values of $Q_*$ are detectable by \emph{Kepler}. Stars are thought to have time-averaged $Q_*$ values around 10000, though this value is highly uncertain and could be much higher for individual stars.}
1398: \tablenotetext{e}{Even with the precision of \emph{Kepler}, apsidal precession for these planets
1399: is undetectable. The extrapolation used to compute eccentricities at specific values of threshold $k_{2p}$ assumes the inverse relationship discussed in the text $k_{2p} \propto e^{-1}$, which is only true for low eccentricities.}
1400: \tablenotetext{f}{This ultra-short period low-mass planet was recently announced by the CoRoT team, but has not been published in a peer-reviewed
1401: journal. We take the parameters from J. Schneider's Extra-solar Planets Encyclopedia and use the mass-radius relation for terrestrial super-Earths of \citet{2007Icar..191..337S}
1402: to estimate the mass as $\sim$9 Earth masses (rather than using the quoted upper limit of 17 Earth masses).}
1403:
1404: \end{deluxetable*}
1405:
1406: \subsection{Results for Specific Planets} \label{resultsdisc}
1407: Using the method described above, we have determined the threshold
1408: $k_{2p}$ for the most favorable known transiting planets as analogs for the
1409: very hot Jupiters to be discovered by \emph{Kepler}. The threshold
1410: $k_{2p}$ for each planet was computed at a range of eccentricities
1411: from 0.001 to 0.1 and for $\omega=0^{\circ}$ and
1412: $\omega=90^{\circ}$. Using the relationship discussed above ($k_{2p} \propto e^{-1}$)
1413: we interpolated (and sometimes extrapolated) our calculations to determine
1414: the eccentricity required to reach threshold $k_{2p}$ values of 0.3, 0.1, and 0.01.
1415: These results are summarized in Figure \ref{ek2planet} and Table 1.
1416:
1417: WASP-12b is the best candidate for observing apsidal precession. With an eccentricity of
1418: $e \simeq 0.00026$ and $k_{2p}$=0.3, the apsidal precession would have
1419: an effective signal-to-noise of $\sim$1 for all of \emph{Kepler}
1420: data. If $e$ is $\sim$0.001, then $k_{2p}$ can be well
1421: characterized and not just detected. As the difference in $k_2$
1422: between Jupiter and Saturn of $\sim0.15$ is primarily due to the
1423: presence of a massive core, a resolution in $k_{2p}$ of 0.1 is enough
1424: to detect whether or not the planet has a core, at the $\sim$1-sigma level.
1425:
1426: Although WASP-12b does not lie in the \emph{Kepler} field, it
1427: clearly stands out as an excellent candidate for observing apsidal precession. Though the putative eccentricity of
1428: 0.049 \citep{2009ApJ...693.1920H} is probably
1429: an overestimate \citep{2005ApJ...629L.121L}, if it were real, it would cause sinusoidal
1430: transit timing deviations with an amplitude of $\sim$25 minutes (using Equation \ref{timeanom}) and
1431: a period of $\sim$18 years. Such a large deviation would be readily observed from the ground in
1432: either transit times or transit shapes. If apsidal precession is not observed, tight upper limits
1433: on the eccentricity can be established.
1434:
1435: Analogs to the very hot Jupiters WASP-4b, TrES-3b, CoRoT-1b, and OGLE-TR-56b are good candidates
1436: for observing apsidal precession if the eccentricities are above $\sim$0.003. (Note that CoRoT-1b has
1437: only $\sim$30 days of observations from the \emph{CoRoT} satellite \citep{Barge2008}, which is insufficient to observe
1438: any of the effects discussed in this paper.) These planets have precession periods of around 100 years so that the argument
1439: of periapse of these planets changes by $\sim$10$^{\circ}$ during the course of \emph{Kepler} observations. Though none
1440: of these planets lie in the \emph{Kepler} field, they are all good candidates for observing apsidal precession though precision
1441: photometry.
1442:
1443: WASP-14b is more massive and has a larger semi-major axis (0.035 instead of 0.025)
1444: which is enough to significantly reduce the detectability of apsidal precession which only proceeds at 0.1$^{\circ}$ per year.
1445: Unlike the previously mentioned planets, WASP-14b has a known non-zero eccentricity of 0.091 $\pm$ 0.003 \citep{2009MNRAS.392.1532J}.
1446: Thus, the amplitude of transit timing variations is known to be very large ($\sim$97 minutes), but with a $\sim$3400 year precession period.
1447:
1448: CoRoT-7b is a very hot super-Earth and has the shortest known orbital period (excepting the ultra-short period planets
1449: of \citealt{2006Natur.443..534S}). We included this planet in our analysis to get a feel for the plausibility of
1450: detecting the interior structure of terrestrial extra-solar planets. The small radius reduces the
1451: planetary contribution to apsidal precession (Figure 1) and significantly reduces the photometric signal. We note here
1452: that in bodies where material strength (rigidity) is more important than self-gravity, $k_{2p}$ is no longer directly related to
1453: internal density distribution. The correction factor is typically small for bodies larger than the Earth \citep{1999ssd..book.....M}.
1454:
1455: XO-3b is a super-massive eccentric planet that is not in the \emph{Kepler} field. Even so, it is interesting to note that, using the known
1456: eccentricity $e=0.2884 \pm 0.0035$ \citep{2009arXiv0902.3461W} and accounting for the brightness of the host star (V=9.8), the \emph{Kepler}
1457: threshold $k_{2p}$ is reduced to only 0.54. As pointed out by \citet{JB08} and \citet{PK08}, XO-3b is a good candidate for observing
1458: apsidal precession within the next decade or so. Furthermore, as discussed below, the non-zero obliquity of the stellar spin axis \citep{2009arXiv0902.3461W}
1459: may also result in an observable signal due to nodal precession.
1460:
1461: HAT-P-7b and HAT-P-11b are orbiting bright stars in the \emph{Kepler} field. The latter is an eccentric hot Neptune with a
1462: relatively large semi-major axis resulting in no eminently detectable apsidal precession. HAT-P-7b, on the other hand, is a good candidate
1463: for detecting apsidal precession. It is probably one of the brightest
1464: hot Jupiters in the \emph{Kepler} field, orbiting a V=10.5 star. The system brightness improves the expected photometric accuracy
1465: from 1000 ppm/min to 200 ppm/min, implying that an eccentricity of only 0.014 is needed to detect apsidal precession (threshold $k_{2p}$=0.3).
1466: \citet{2009IAUS..253..428P} report a best-fit eccentricity of 0.003 $\pm$ 0.012, indicating that the necessary eccentricity cannot be ruled out.
1467: Furthermore, this planet has transiting data extending back to 2004 and was observed by NASA's \emph{EPOXI} Mission in 2008
1468: (\citealp{2009IAUS..253..301C}; D. Deming, pers. comm.). This additional baseline, though sparsely sampled, may provide the additional
1469: leverage needed to detect apsidal precession if the eccentricity is non-zero.
1470: Note, however, that detecting changes in transit shapes is more difficult when the observations
1471: are made with a variety of telescopes because transit shapes depend on the observing filter used, due to wavelength-dependent
1472: limb darkening.
1473:
1474: TrES-2b is similar to HAT-P-7b in that it also lies in the \emph{Kepler} field, has observations dating to 2005, and was observed by NASA's \emph{EPOXI} Mission.
1475: TrES-2b is somewhat fainter than HAT-P-7b (V=11.4), and, correcting for the system brightness, an eccentricity of 0.021 would result in detectable apsidal
1476: precession (threshold $k_{2p}$=0.3). Observations of the secondary eclipse show no detectable deviations of the orbit from circularity \citep{2009IAUS..253..536O}. Even so,
1477: the light curve of this planet is quite sensitive to perturbations as it has a quite high impact parameter $b=0.854$. Accounting for this impact parameter
1478: does not significantly change the required eccentricity.
1479:
1480: We conclude that \emph{Kepler may detect the
1481: cores of very hot Jupiters} and probe their interior structure
1482: though their evolving transit light curve if eccentricities are above $\sim$0.003.
1483: As future observations provide longer baselines
1484: for these observations, the sensitivity to interior structure measurements
1485: will increase dramatically, significantly lowering the eccentricity needed
1486: to observe apsidal precession.
1487:
1488: In cases where apsidal precession is not
1489: observed, the data can set strong upper limits on
1490: planetary eccentricities. An upper limit on the eccentricity
1491: can be inferred by assuming that the planet
1492: has the minimal physically-plausible value of $k_{2p} \approx 0.1$.
1493: Null detections of apsidal motion should therefore provide
1494: upper limits on eccentricity comparable to the values shown in Table 1 (also shown by
1495: circles in Figure \ref{ek2planet}). Such strong eccentricity constraints are valuable
1496: for improving our understanding of these close-in planets.
1497:
1498: \section{Potential Confusion of the Apsidal Precession Signal}
1499:
1500: In the above, we have assumed that measuring $\dot{\omega}$ is
1501: tantamount to measuring $k_{2p}$. This is justified by noting that
1502: the conversion $\dot{\omega}$ to $k_{2p}$ involves only factors that
1503: are very well characterized. In Section 2 and Figure 1, we showed that
1504: $k_{2p}$ is usually the dominant source of apsidal precession.
1505: The effects of $k_{2*}$ and general relativity are well-understood and
1506: can typically be subtracted away without introducing serious uncertainty, even when they dominate the apsidal precession
1507: rate. From Equation \ref{tidalprec}, converting the remaining
1508: $\dot{\omega}_{\rm p}$ to $k_{2p}$ requires only knowing $\frac{M_p}{M_*}$, $e$, $\frac{R_p}{a}$,
1509: and $n$. The latter two are very accurately measured
1510: with even a few transit light curves \citep[e.g.,][]{Torres2008, Southworth2008}.
1511: The eccentricity only enters the equation through the $f_2(e)$ and $g_2(e)$
1512: eccentricity functions (Equations \ref{f2e} and \ref{g2e}), and \emph{Kepler} observations of secondary eclipse
1513: are sufficiently accurate to remove any systematic error due to these terms
1514: unless the eccentricity is large ($e\gtrsim0.3$).
1515: Determining the mass ratio requires well-sampled radial
1516: velocity observations. The systems detected by \emph{Kepler} are
1517: bright enough to get good mass measurements, especially
1518: since very hot Jupiters have large radial velocity amplitudes ($K \sim 200$
1519: m/s).\footnote{Other than determining the mass ratio and constraining the
1520: eccentricity, radial velocity
1521: information is thought to have a negligible contribution in constraining
1522: apsidal precession unless a serious observational campaign can
1523: measure the radial velocity period (independently of transits) to sub-second accuracies.
1524: \citep{2007MNRAS.377.1511H,JB08}.}
1525: The anticipated error in the mass
1526: ratio is a few percent \citep{Torres2008}. In all, we estimate that,
1527: converting from $\dot{\omega}$ to $k_{2p}$ leads to a typical systematic error on
1528: $k_{2p}$ of around $\sim$.01. This is a relatively small systematic effect in comparison to the
1529: potential range ($\sim$0.5) of $k_{2p}$ values. For reference, the eccentricity required to reach a
1530: threshold $k_{2p}$ of 0.01 is shown in Figure \ref{ek2planet} by squares.
1531:
1532: Another way to introduce systematic errors on the measurement of $k_{2p}$ is
1533: to misinterpret similar transit light curve variations. To ensure that the method
1534: outlined in this paper truly probes the interiors of extra-solar
1535: planets, we consider in this section whether the transit light curve
1536: resulting from apsidal
1537: precession can be confused with any other common circumstances. Although a
1538: very specific combination of parameters is required for any particular phenomenon
1539: to successfully mimic a signal due to $k_{2p}$, the below effects
1540: should be reconsidered when actual data is available.
1541:
1542: \subsection{Testing the Effect of Obliquity}
1543: \label{obliquitysec}
1544:
1545: If either the star or planet has a non-zero obliquity, the orbital plane will no longer
1546: be fixed as a result of nodal precession. The obliquities of
1547: very hot Jupiters rapidly ($\lesssim 1$ MYr) decay to a
1548: Cassini state, and recent work has shown that these planets are likely
1549: in Cassini state 1 \citep{2005ApJ...628L.159W,2007A&A...462L...5L,2007ApJ...665..754F}.
1550: Using a model based on the equations of \citet{2001ApJ...562.1012E}, we found that
1551: Cassini obliquities of very hot Jupiters are indeed negligible ($\alpha_p < 0.01^{\circ}$).
1552: Though tidal damping of the stellar obliquity occurs on far longer
1553: timescales,
1554: several measurements of the projected stellar obliquity through the
1555: Rossiter-McLaughlin effect indicates that planet-hosting stars generally have
1556: low obliquities $\lesssim 10^{\circ}$ like the Sun \citep{2009arXiv0902.0737F}.
1557: Hence, the general expectation is that both the star and planet will have rather
1558: low, but potentially non-zero obliquities.
1559:
1560: Understanding the specific orbital evolution resulting from non-zero
1561: obliquities is more complicated than the simple prescription for
1562: apsidal precession.
1563: To correctly account for non-Keplerian effects, we wrote a direct
1564: integrator, following \citet{2002ApJ...573..829M}, that
1565: calculates the Cartesian trajectory (and the direction of the spin axes) of a star-planet
1566: system including general
1567: relativity and the effects of quadrupolar distortion. This integrator
1568: reproduces the orbit-averaged analytic
1569: equations of \citet{2002ApJ...573..829M}, which are the same as those in
1570: \citet{2001ApJ...562.1012E}, \citet{1939MNRAS..99..451S}, and
1571: elsewhere.\footnote{This involved minor modifications to the "direct
1572: integrator" equations 3 and 5 in \citet{2002ApJ...573..829M}. In Equation 3, the coefficient 12 should
1573: be a 6 (R. Mardling, pers. comm.) and Equation 5 was replaced with the nearly equivalent
1574: equation from \citet{1989racm.book.....S}.} We did
1575: not include the effects of tidal forces or additional planets which
1576: are not relevant to our problem.
1577:
1578: Using this direct integrator, we investigated the effect of non-zero obliquities on the
1579: transit times, durations, and impact parameters. Integration of several cases with
1580: varying stellar and planetary obliquities showed that the largest effect on the
1581: photometry was due to changes in the impact parameter, as expected for an orbit with
1582: changing orientation \citep{2002ApJ...564.1019M}. However, even for large stellar
1583: obliquities ($\sim 45^{\circ}$) the transit light curve variations due to obliquity are
1584: generally small relative to the effects of purely apsidal precession, even with low
1585: eccentricities. One reason for this is that the tidal bulge, which does not contribute
1586: to nodal precession, is $\gtrsim15$ times
1587: more important than the rotational bulge.
1588: As with apsidal precession, the planetary contribution to orbital variations
1589: is much stronger than the stellar contribution (for equal obliquities). Unless the
1590: planetary obliquity is unexpectedly large ($\gtrsim 0.5^{\circ}$),
1591: the obliquity-induced nodal precession should have only a
1592: minor effect on the transit light curve.
1593:
1594: \subsection{Transit Timing due to Orbital Decay}
1595: \label{tidessec}
1596:
1597: Orbital decay generates a small secular trend in transit times.
1598: \citet{2003ApJ...596.1327S} proposed the detectability
1599: of the expected $\sim$1 ms/yr period change due to semi-major axis decay
1600: of
1601: OGLE-TR-56b. The transit timing anomaly due solely to orbital decay (or growth) is
1602: the result of constantly accumulating changes in the period:
1603: \be\label{ttta}
1604: T_N \simeq T_0 + NP_{\rm obs} + \frac{1}{2} N^2 \delta P
1605: \ee
1606: where $\delta P \equiv \dot{P}P$ is the change in the period during
1607: one orbit and $N$ is the number of transits after the initial
1608: transit. Equation \ref{ttta} can be derived by noting that the transit
1609: times are
1610: basically the integral of the instantaneous period. As before, the
1611: transit timing anomaly is composed of the quadratic deviation of $T_N$ from a straight line.
1612: The change in period can be due to magnetic stellar breaking \citep[e.g.,][]{2009AJ....137.3181L,2009arXiv0902.4554B},
1613: the Yarkovsky effect applied to planets \citep{2008ApJ...677L.117F}, and/or other effects.
1614:
1615: For planets orbiting an asynchronously rotating star, a major
1616: source of orbital decay is tidal evolution, which results in a slow change in
1617: semi-major axis, according to the formula \citep{1999ssd..book.....M}:
1618: \be
1619: \dot{a} = \textrm{sign}(\nu_* - n) \frac{3k_{2*}}{Q_*} \frac{M_p}{M_*} \left( \frac{R_*}{a} \right)^5 n a
1620: \ee
1621: where sign($x$) returns the sign of $x$ or 0 if $x=0$ and where $Q_*$ is the tidal
1622: quality parameter of the star, typically around $\e{4}$ \citep{2004ApJ...610..464D}.
1623: Though $\delta P$ due to
1624: tidal dissipation is only of order 3 micro-seconds, $N$ grows by $\sim 300$ each year,
1625: reaching $\sim$1000 during the duration of \emph{Kepler} for
1626: very hot Jupiters. This implies a transit timing signal of about a few seconds.
1627:
1628: Calculating the total ``signal-to-noise'' of tidal evolution, as was
1629: done for $k_{2p}$, we find that reasonable values of $Q_*$ can be
1630: measured even
1631: for faint stars ($V=14$; 1000 ppm/min noise). For a circular
1632: orbit with the
1633: parameters of OGLE-TR-56b, the effective $\frac{S}{N}$
1634: reaches 1 when $\dot{P}$ is 1.36 ms/yr (see Table 1), corresponding to $Q_* \approx 25000$. This implies the
1635: detectability
1636: of most of the empirically-motivated estimates of \citet{2003ApJ...596.1327S} for
1637: the tidal decay of OGLE-TR-56b, which
1638: are estimated to be within an order of magnitude of 1 ms/yr. On the other hand,
1639: \citet{2009arXiv0902.4563B} estimate that the tidal damping in F-stars like OGLE-TR-56
1640: and WASP-12 may be very low, which may explain the survival of these short period planets.
1641:
1642: The estimates of the threshold values of $\dot{P}$, shown in Table 1, include
1643: removing degeneracies in other parameters, except apsidal precession of
1644: eccentric orbits, and assume that everything but $\dot{P}$ is known. Note that the
1645: transit light curve signal due to orbital decay is due entirely to transit timing; the
1646: change in $a$ is far too small to observe in transit shaping. As the signal
1647: due to apsidal precession includes significant changes to the shapes of the
1648: transits, the signal due to $k_{2p}$ is qualitatively different than that of $Q_*$.
1649: The shifting of secondary transits from precession also help in this
1650: regard, as outlined above. However,
1651: the primary transit timing signals can be similar: quadratic transit timing
1652: anomalies with amplitudes of $\sim$1 second. \label{threshpdot}
1653:
1654: \emph{Kepler} analogs of very hot Jupiters WASP-12b, OGLE-TR-56b, CoRoT-1b, WASP-4b, and TrES-3b
1655: could have detectable transit timing anomalies due to tidal decay,
1656: implying a direct measurement of the current value of $Q_*$ for specific stars (Table 1). This is an
1657: exciting possibility, providing the first direct measurements (or constraints)
1658: of the currently unknown details of tidal dissipation in
1659: a variety of individual stars.\footnote{The vanishingly small effect of eccentricity decay
1660: is $\sim \frac{1}{Q_p}$ smaller than apsidal precession, so that
1661: direct measurements of $Q_p$ from eccentricity decay are not feasible.} We also note that interesting orbital decay of
1662: eclipsing binary systems seen by \emph{Kepler} could also be detectable.
1663:
1664: \subsection{Confusion Due to Other Planets}
1665:
1666: Could the signal due to $k_{2p}$ be confused with additional planets? In
1667: considering this issue, it should be noted that all known hot Jupiters (with $a
1668: \lesssim 0.05$ AU and $M_p \gtrsim 0.5 M_{Jup}$) have no currently known additional
1669: companions. The apparent single nature of these systems could very
1670: well be due to observational
1671: biases \citep{2008arXiv0806.4314F}. However, even for stars that
1672: have been observed for many years with radial velocity (e.g. 51 Peg, HD
1673: 209458), there appears
1674: to be a strong tendency towards hot Jupiters as the only close-in massive planets.
1675:
1676: Previous studies of transit timing variations focus on the effects of additional
1677: planetary perturbers \citep[e.g.,][]{2005Sci...307.1288H,2005MNRAS.359..567A,2007ApJ...664L..51F,2008ApJ...688..636N}.
1678: These authors find that nearby massive planets or even low-mass planets in mean-motion resonances
1679: would cause strong transit timing variations that are easily distinguishable from the comparatively
1680: long-period timing anomalies due to $k_{2p}$. Relatively distant
1681: companions or non-resonant low-mass planets, however, can induce a linear apsidal precession signal
1682: just like $k_{2p}$ \citep{2002ApJ...564.1019M,2007MNRAS.377.1511H,JB08}. The precession rate
1683: induced by a perturbing body is a function of its mass and semi-major axis. The interior structure of very hot Jupiters
1684: causes apsidal precession as fast as a few degrees per year. To match this precession rate
1685: would require, for example, another Jupiter-mass planet at $\lesssim 0.1$AU or a solar-mass star at $\sim$1 AU.
1686: Even perturbers an order of magnitude smaller than these would be readily detectable using radial velocity observations
1687: and/or high-frequency transit time variations. When restricted to planets that are undetectable by other means,
1688: adding the precession due to the unknown perturbing planet would lead to an
1689: insignificant overestimate of $k_{2p}$ for very hot Jupiters.\footnote{Conversely, as a consequence
1690: of the fast precession of very hot Jupiters due to their (unknown) interiors, it
1691: will be very difficult to detect the presence of additional perturbing planets in these systems
1692: from apsidal precession alone.} When observing transiting planets
1693: with larger semi-major axes ($a \gtrsim 0.05$ AU), the strength of planetary
1694: induced apsidal precession is reduced to a level comparable to apsidal precession
1695: from a low-mass perturbing planet \citep{JB08} and confusion may be possible in these cases.
1696:
1697: Since the transit timing signal for apsidal precession is similar to a
1698: sinusoid, another potential source of confusion would be light-travel
1699: time offsets due to a distant orbiting companion \citep[e.g.,][]{2008A&A...480..563D}. The transit timing signal due
1700: to stellar motion about the barycenter can be distinguished from
1701: $k_{2p}$ precession\footnote{Transit time anomalies due to $Q_*$ (Section
1702: \ref{tidessec}), however, can be confused with barycenter light-travel time
1703: shifts due to a distant planet that may be undetectable in radial velocities.}
1704: by considering the changes in transit shapes and
1705: primary-secondary transit time offsets, which are not affected by distant companions.
1706:
1707: We conclude that transit timing effects from other planets can be readily
1708: distinguished from the effects of apsidal precession. To address the issue of
1709: the transit shaping signal due to additional planets, we wrote a simple three-body
1710: integrator (similar to the integrator mentioned above) to investigate the kinds of
1711: transit light curve signals created by additional planets. For the vast majority of additional planet parameters,
1712: the transit timing deviations always carry far more signal than the minor deviations
1713: due to changes in the angular velocity\footnote{The angular velocity
1714: is directly related to the star-planet separation through conservation
1715: of angular momentum: $r\dot{f}^2$.} ($\dot{f}_{tr}$) or impact parameter ($b$), which together
1716: determine the transit shape as described in Section \ref{transitshapes} above. Generally,
1717: it is much easier to delay
1718: a transit by 5 seconds than it is to shift the apparent transit plane by an appreciable amount.
1719:
1720: However, when the perturbing planet is on a plane highly-inclined to the transiting planet, changes in the transit shape
1721: can become detectable, even while the transit timing variations are negligible. For example, a perturbing
1722: planet of mass $\e{-5}M_*$ at 0.1 AU with a mutual inclination of 45$^{\circ}$ caused very hot Jupiter transit
1723: durations to change by $\sim$1 second/year. This kind of signal is the result of nodal precession induced
1724: by the perturbing planet, as originally pointed out by \citet{2002ApJ...564.1019M}. In our investigation,
1725: we found that the three-body nodal precession alters the impact parameter ($b$) but does not significantly
1726: affect the orbital angular velocity ($\dot{f}_{tr}$). Conversely, the transit shaping signal due to $k_{2p}$ is
1727: generally produced by changes in both $b$ and $\dot{f}_{tr}$, but at near-central transits, the
1728: effect of changing orbital velocity is dominant (see Section \ref{transitshapes}). In high-precision
1729: transit light curves, both the angular velocity and the impact parameter can be
1730: independently measured and hence the signals of apsidal and nodal
1731: precession are usually distinct for all but the most grazing transits.
1732:
1733: Given the uniqueness of the apsidal precession signal induced by the planet's interior, it appears that if additional
1734: planets are not detectable in radial velocities, transit timing variations, or nodal precession, then
1735: they will not contribute to a misinterpretation of an inferred value of $k_{2p}$ for very hot Jupiters.
1736: Nevertheless, future measurements of $k_{2p}$ should check that these issues are unimportant
1737: within the context of the specific system being studied.
1738:
1739: Finally, we estimate that moons or rings with enough mass to bias an
1740: inferred $k_{2p}$ would cause other readily detectable photometric
1741: anomalies \citep[e.g., planet-moon barycentric motion][]{1999A&AS..134..553S}.
1742: In addition, extra-solar moons with any significant mass are tidally unstable,
1743: especially around very hot Jupiters \citep{2002ApJ...575.1087B}.
1744:
1745: \section{Other Methods for Determining $k_{2p}$}
1746:
1747: \subsection{Secular Evolution of a Two Planet System}
1748: Measuring $k_2$ for an extra-solar planet was suggested by \citet{2002ApJ...564.1024W}
1749: for the inner planet of HD 83443. Unfortunately, later analyses have
1750: indicated that the supposed second planet in this system was actually
1751: an artefact of the sparse radial velocity data \citep{2004A&A...415..391M}. Nevertheless,
1752: this technique could be applied to other eccentric planetary systems
1753: with similar conditions \citep{2007MNRAS.382.1768M}. \citet{2002ApJ...564.1024W}
1754: showed that in a
1755: regime of significant tidal circularization and
1756: excitation from an additional planet, the ratio of eccentricities
1757: depends on the precession rate which is dominated by
1758: $k_{2p}$ as shown above (see also \citealt{2006ApJ...649.1004A}, who do not
1759: include precession due to the planetary quadrupole). In theory, the
1760: current orbital state of such multi-planet systems gives an
1761: indirect measurement of the apsidal precession rate.
1762:
1763: \subsection{Direct Detection of Planetary Asphericity}
1764: Another method for determining interior properties of transiting planets
1765: would be to directly measure the asphericity due to the rotational or tidal
1766: bulge in primary transit photometry. The height of the
1767: rotational and tidal bulges are $q_r h_2 R_p$ and $q_t h_2 R_p$,
1768: respectively, where $q_r$ and $q_t$ are the dimensionless small
1769: parameters defined in Equation \ref{qparams} and $h_2$ is
1770: another Love number which, for fluid bodies, is simply $k_2 + 1$ \citep{1939MNRAS..99..451S}.
1771: These bulges cause the disk of the planet to be slightly
1772: elliptical, subtly modifying the photometric signal, as discussed for
1773: rotational bulges by \citet{2002ApJ...574.1004S} and \citet{2003ApJ...588..545B}. However, as
1774: discussed by \citet{2003ApJ...588..545B}, in real systems with actual
1775: observations, the size of the
1776: rotational bulge is very
1777: difficult to determine as it is highly correlated with
1778: stellar and orbital parameters that are not known
1779: \emph{a priori}, e.g. limb darkening coefficients.
1780:
1781: The tidal bulge, whose height is also set by $k_{2p}$,
1782: does not suffer from some of the difficulties involved
1783: with measuring the rotational bulge. It has a known orientation
1784: (pointing towards the star) so there is no degeneracy from an unknown
1785: obliquity \citep{2003ApJ...588..545B}. (Note, however, that for
1786: hot Jupiters, the obliquities must be tidally evolved to nearly zero, so this
1787: isn't really a problem with the rotational bulge.) In addition, the signal due to oblateness is only
1788: significant near ingress/egress, but the tidal bulge is continuously
1789: changing orientation throughout the entire
1790: transit. Though the tidal
1791: bulge is typically three times larger than the rotational bulge
1792: (Equation 2), the projection
1793: of the tidal bulge that is seen during a transit is small, proportional to
1794: $\sin \theta$ where $\theta$ is the angle between the planet position
1795: and the Earth's line of sight. For very hot Jupiters that have
1796: semi-major axes of only $\lesssim$6 stellar radii, $\sin \theta$ during transit
1797: ingress/egress
1798: reaches $\gtrsim \frac{1}{6}$ so that the projected tidal bulge is
1799: about half as large as the rotational bulge. The extra dimming due to the tidal bulges (and rotational
1800: bulges) is as high as $2 \x \e{-4}$ for some planets that are expected to
1801: have tides over
1802: 2000 km high (e.g. WASP-12b, WASP-4b,
1803: Corot-1b, OGLE-TR-56b); this compares very favorably with
1804: the photometric accuracy of binned \emph{Kepler} data at about
1805: 10 ppm per minute. However, we expect that, as with the
1806: rotational bulge alone, the combined signal from the rotational and tidal bulge
1807: will be highly
1808: degenerate with the unknown limb-darkening coefficients, as the
1809: size of the projection of the tidal bulge also varies as the distance to
1810: the center of the star.
1811:
1812: We note that using multi-color photometry should significantly improve
1813: the prospects of detecting non-spherical planetary transits since it
1814: breaks most of these degeneracies. For
1815: example, \citet{2007ApJ...655..564K} use HST to observe transits of HD 209458b in 10 wavelength
1816: bands and measure the planetary radius with a relative accuracy
1817: (between bands) of $0.003 R_J$, of the same level as the change in
1818: shape due to oblateness and the tidal bulge. \citet{2007A&A...476.1347P} made a
1819: similar measurement for HD 189733b and reached even higher relative
1820: accuracy. Combining such
1821: measurements with other data (e.g. primary transits in the infrared,
1822: where limb-darkening is much smaller) and a stellar photosphere model (to
1823: correctly correlate limb darkening parameters as in \citealt{2007MNRAS.374..941A})
1824: could yield detections
1825: of planetary asphericity, especially in very hot Jupiters which have
1826: the largest bulges.
1827:
1828: One possible source of confusion in interpreting planetary asphericity is the thermally-induced pressure
1829: effects of an unevenly radiated surface. In non-synchronous planets, the thermal tidal bulge \citep{2009arXiv0901.0735A} can
1830: shift the level of the photosphere by approximately an atmospheric scale height, about
1831: $\e{-2}$ or $\e{-3}$ planetary radii (P. Arras, pers. comm.). The orientation of the thermal bulge is significantly
1832: different from the tidal or rotational bulges and should be distinguishable.
1833: Furthermore, very hot Jupiters should orbit synchronously, reducing the importance of this effect. Nevertheless, the effect
1834: of atmospheric phenomena on measurements of planetary asphericity should be considered.
1835:
1836: Though difficult to disentangle from other small
1837: photometric effects, high-precision multi-color photometry may be another viable
1838: method for measuring $k_{2p}$. This technique is complimentary to detecting $k_{2p}$
1839: from apsidal precession since it does not require that the planet is eccentric,
1840: nor does it require a long time baseline. On some planets,
1841: the two methods could be used together as mutual confirmation
1842: of the planetary interior structure.
1843:
1844: \section{Conclusions}
1845:
1846: The planetary mass and radius are the only bulk physical characteristics measured
1847: for extra-solar planets to date. In this paper, we find that the planetary Love number ($k_{2p}$,
1848: equivalent to $J_2$)
1849: can also have an observationally detectable signal (quadrupole-induced apsidal precession) which can provide a new and
1850: unique probe into the interiors of very hot Jupiters. In particular, $k_{2p}$
1851: is influenced by the size of a solid core and other internal properties. Core sizes can be used to infer the formation
1852: and evolution of individual extra-solar planets \citep[e.g.,][]{2009arXiv0901.0582D,2009arXiv0903.1997H}.
1853:
1854: The presence of a nearby massive star creates a large tidal potential on these
1855: planets, raising significant tidal bulges which then induce non-Keplerian effects on
1856: the star-planet orbit itself. The resulting apsidal precession accounts for $\sim$95\% of
1857: the total apsidal precession in the best cases (Figure \ref{odotfig}). Hence, we
1858: find that the internal density distribution, characterized by $k_{2p}$, has a large
1859: and clear signal, not to be confused with any other parameters or phenomena. We urge
1860: those modeling the interior structures of extra-solar planets to tabulate the
1861: values of $k_{2p}$ for their various models.
1862:
1863: Encouraged by this result, we calculated full photometric light-curves like those
1864: expected from the \emph{Kepler} mission to determine the realistic observability
1865: of the interior signal. We estimate that \emph{Kepler} should be able to
1866: distinguish between interiors with and without massive cores ($\Delta k_{2p} \simeq
1867: 0.1$) for very hot Jupiters with eccentricities around $e \sim 0.003$ (Figure
1868: \ref{ek2planet}). Eccentricities this high may occur for some of the very
1869: hot Jupiters expected to be found by \emph{Kepler}, though
1870: these planets usually have highly damped eccentricities. Much stronger constraints
1871: on apsidal precession can be obtained by combining \emph{Kepler} photometry
1872: with precise secondary transits observed in the infrared. In cases where apsidal precession is not
1873: observed, the data can set strong upper limits on planetary eccentricities.
1874:
1875: In analyzing \emph{Kepler}'s photometric signal of apsidal precession, we find that transit
1876: timing variations are an almost negligible source of signal, though transit timing
1877: has been the focus of many observational and theoretical papers to date. The effect
1878: of ``transit shaping'' has $\sim$30 times the photometric signal of transit timing for
1879: apsidal precession \citep[see Figure \ref{candywrapperpieces},][]{PK08,JB08}). At
1880: orientations where transit timing and shaping are weakest, the changing offset
1881: between primary and secondary transit times can be used to measure $k_{2p}$ (Figure
1882: \ref{bowtiemain}). It may also be possible to measure $k_{2p}$ from high-precision
1883: multi-color photometry by directly detecting the planetary asphericity in transit.
1884: Such a measurement does not require a long baseline or an eccentric orbit.
1885:
1886: Very hot Jupiters are also excellent
1887: candidates for detecting tidal semi-major axis decay, where we find that relatively
1888: small period changes of $\dot{P} \simeq 1$ ms/yr should be detectable. This could
1889: constitute the first measurements (or constraints) on tidal $Q_*$ for a variety of
1890: individual stars. We note that \emph{Kepler} measurements of transit timing and shaping
1891: for eclipsing binaries should also provide powerful constraints on stellar interiors through
1892: apsidal motion and binary orbital decay (due to tides, if the components are asynchronous).
1893:
1894: Accurately measuring the interior structure of distant extra-solar planets seems
1895: too good to be true. Nevertheless, the exquisite precision, constant monitoring,
1896: and 3.5-year baseline of the \emph{Kepler} mission combined with the high
1897: sensitivity of transit light curves to small changes in the star-planet orbit make
1898: this measurement plausible.
1899:
1900: Our focus on \emph{Kepler} data should not be interpreted to mean that other observations will
1901: be incapable of measuring $k_{2p}$. In fact, the opposite is true since
1902: the size of the apsidal precession signal increases
1903: dramatically with a longer baseline. Combining \emph{Kepler} measurements with future ground
1904: and space based observations can create a powerful tool for measuring $k_{2p}$.
1905: In the far future, many planets will have
1906: measured apsidal precession rates (like eclipsing binary systems have now) and inferred
1907: $k_{2p}$ values. Incorporating these measurements into interior models holds
1908: promise for greater understanding of all extra-solar planets.
1909:
1910: \begin{acknowledgements}
1911: We thank Dave Stevenson, Mike Brown, Greg Laughlin, Oded Aharonson, Thomas Beatty, Phil Arras,
1912: Rosemary Mardling, Re'em Sari, Alejandro Soto, Ian McEwen
1913: and Chris Lee for help and useful discussions. We especially
1914: thank the referee, Dan Fabrycky, for helpful suggestions and discussions.
1915: DR is grateful for the
1916: support of the Moore Foundation. ASW is grateful for support from the National Science
1917: Foundation. This research has made use of NASA's Astrophysics Data System.
1918: \end{acknowledgements}
1919:
1920: \begin{thebibliography}{118}
1921: \expandafter\ifx\csname natexlab\endcsname\relax\def\natexlab#1{#1}\fi
1922:
1923: \bibitem[{{Adams} \& {Laughlin}(2006)}]{2006ApJ...649.1004A}
1924: {Adams}, F.~C., \& {Laughlin}, G. 2006, \apj, 649, 1004, arXiv:astro-ph/0606349
1925:
1926: \bibitem[{{Agol} {et~al.}(2009){Agol}, {Cowan}, {Bushong}, {Knutson},
1927: {Charbonneau}, {Deming}, \& {Steffen}}]{2009IAUS..253..209A}
1928: {Agol}, E., {Cowan}, N.~B., {Bushong}, J., {Knutson}, H., {Charbonneau}, D.,
1929: {Deming}, D., \& {Steffen}, J.~H. 2009, in IAU Symposium, Vol. 253, IAU
1930: Symposium, 209--215
1931:
1932: \bibitem[{{Agol} {et~al.}(2005){Agol}, {Steffen}, {Sari}, \&
1933: {Clarkson}}]{2005MNRAS.359..567A}
1934: {Agol}, E., {Steffen}, J., {Sari}, R., \& {Clarkson}, W. 2005, \mnras, 359,
1935: 567, arXiv:astro-ph/0412032
1936:
1937: \bibitem[{{Agol} \& {Steffen}(2007)}]{2007MNRAS.374..941A}
1938: {Agol}, E., \& {Steffen}, J.~H. 2007, \mnras, 374, 941, arXiv:astro-ph/0610159
1939:
1940: \bibitem[{{Alonso} {et~al.}(2009){Alonso}, {Aigrain}, {Pont}, {Mazeh}, \& {The
1941: CoRoT Exoplanet Science Team}}]{2009IAUS..253...91A}
1942: {Alonso}, R., {Aigrain}, S., {Pont}, F., {Mazeh}, T., \& {The CoRoT Exoplanet
1943: Science Team}. 2009, in IAU Symposium, Vol. 253, IAU Symposium, 91--96
1944:
1945: \bibitem[{{Arras} \& {Socrates}(2009)}]{2009arXiv0901.0735A}
1946: {Arras}, P., \& {Socrates}, A. 2009, ArXiv e-prints, 0901.0735
1947:
1948: \bibitem[{{Bakos} {et~al.}(2009){Bakos}, {Torres}, {P{\'a}l}, {Hartman},
1949: {Kov{\'a}cs}, {Noyes}, {Latham}, {Sasselov}, {Sip{\H o}cz}, {Esquerdo},
1950: {Fischer}, {Johnson}, {Marcy}, {Butler}, {Isaacson}, {Howard}, {Vogt},
1951: {Kov{\'a}cs}, {Fernandez}, {Mo{\'o}r}, {Stefanik}, {L{\'a}z{\'a}r}, {Papp},
1952: \& {S{\'a}ri}}]{2009arXiv0901.0282B}
1953: {Bakos}, G.~{\'A}. {et~al.} 2009, ArXiv e-prints, 0901.0282
1954:
1955: \bibitem[{Baraffe {et~al.}(2008)Baraffe, Chabrier, \& Barman}]{Baraffe2008}
1956: Baraffe, I., Chabrier, G., \& Barman, T. 2008, 0802.1810
1957:
1958: \bibitem[{Barge {et~al.}(2008)Barge, Baglin, Auvergne, Rauer, Leger, Schneider,
1959: Pont, Aigrain, Almenara, Alonso, Barbieri, Borde, Bouchy, Deeg, la~Reza,
1960: Deleuil, Dvorak, Erikson, Fridlund, Gillon, Gondoin, Guillot, Hatzes,
1961: Hebrard, Jorda, Kabath, Lammer, Llebaria, Loeillet, Magain, Mazeh, Moutou,
1962: Ollivier, Patzold, Queloz, Rouan, Shporer, \& Wuchterl}]{Barge2008}
1963: Barge, P. {et~al.} 2008, 0803.3202
1964:
1965: \bibitem[{{Barker} \& {Ogilvie}(2009{\natexlab{a}})}]{2009arXiv0902.4554B}
1966: {Barker}, A.~J., \& {Ogilvie}, G.~I. 2009{\natexlab{a}}, ArXiv e-prints,
1967: 0902.4554
1968:
1969: \bibitem[{{Barker} \& {Ogilvie}(2009{\natexlab{b}})}]{2009arXiv0902.4563B}
1970: ------. 2009{\natexlab{b}}, ArXiv e-prints, 0902.4563
1971:
1972: \bibitem[{{Barnes} \& {Fortney}(2003)}]{2003ApJ...588..545B}
1973: {Barnes}, J.~W., \& {Fortney}, J.~J. 2003, \apj, 588, 545,
1974: arXiv:astro-ph/0301156
1975:
1976: \bibitem[{{Barnes} \& {O'Brien}(2002)}]{2002ApJ...575.1087B}
1977: {Barnes}, J.~W., \& {O'Brien}, D.~P. 2002, \apj, 575, 1087,
1978: arXiv:astro-ph/0205035
1979:
1980: \bibitem[{{Batten}(1973)}]{1973bmss.book.....B}
1981: {Batten}, A.~H. 1973, {Binary and multiple systems of stars} (Binary and
1982: multiple systems of stars / by Alan H.~Batten.~Oxford ; New York : Pergamon
1983: Press, [1973] (International series of monographs in natural philosophy ;
1984: v.~51))
1985:
1986: \bibitem[{{Bean}(2009)}]{2009arXiv0903.1845B}
1987: {Bean}, J.~L. 2009, ArXiv e-prints, 0903.1845
1988:
1989: \bibitem[{{Beatty} \& {Gaudi}(2008)}]{2008arXiv0804.1150B}
1990: {Beatty}, T.~G., \& {Gaudi}, B.~S. 2008, ArXiv e-prints, 804, 0804.1150
1991:
1992: \bibitem[{{Bodenheimer} {et~al.}(2001){Bodenheimer}, {Lin}, \&
1993: {Mardling}}]{2001ApJ...548..466B}
1994: {Bodenheimer}, P., {Lin}, D.~N.~C., \& {Mardling}, R.~A. 2001, \apj, 548, 466
1995:
1996: \bibitem[{{Borucki} {et~al.}(2003){Borucki}, {Koch}, {Lissauer}, {Basri},
1997: {Caldwell}, {Cochran}, {Dunham}, {Geary}, {Latham}, {Gilliland}, {Caldwell},
1998: {Jenkins}, \& {Kondo}}]{2003SPIE.4854..129B}
1999: {Borucki}, W.~J. {et~al.} 2003, in Presented at the Society of Photo-Optical
2000: Instrumentation Engineers (SPIE) Conference, Vol. 4854, Future EUV/UV and
2001: Visible Space Astrophysics Missions and Instrumentation. Edited by J. Chris
2002: Blades, Oswald H. W. Siegmund. Proceedings of the SPIE, Volume 4854, pp.
2003: 129-140 (2003)., ed. J.~C. {Blades} \& O.~H.~W. {Siegmund}, 129--140
2004:
2005: \bibitem[{{Burrows} {et~al.}(2007){Burrows}, {Hubeny}, {Budaj}, \&
2006: {Hubbard}}]{2007ApJ...661..502B}
2007: {Burrows}, A., {Hubeny}, I., {Budaj}, J., \& {Hubbard}, W.~B. 2007, \apj, 661,
2008: 502, arXiv:astro-ph/0612703
2009:
2010: \bibitem[{{Burrows} {et~al.}(2008){Burrows}, {Ibgui}, \&
2011: {Hubeny}}]{2008arXiv0803.2523B}
2012: {Burrows}, A., {Ibgui}, L., \& {Hubeny}, I. 2008, ArXiv e-prints, 803,
2013: 0803.2523
2014:
2015: \bibitem[{{Charbonneau}(2003)}]{2003ASPC..294..449C}
2016: {Charbonneau}, D. 2003, in Astronomical Society of the Pacific Conference
2017: Series, Vol. 294, Scientific Frontiers in Research on Extrasolar Planets, ed.
2018: D.~{Deming} \& S.~{Seager}, 449--456
2019:
2020: \bibitem[{{Charbonneau} {et~al.}(2005){Charbonneau}, {Allen}, {Megeath},
2021: {Torres}, {Alonso}, {Brown}, {Gilliland}, {Latham}, {Mandushev}, {O'Donovan},
2022: \& {Sozzetti}}]{2005ApJ...626..523C}
2023: {Charbonneau}, D. {et~al.} 2005, \apj, 626, 523, arXiv:astro-ph/0503457
2024:
2025: \bibitem[{{Charbonneau} {et~al.}(2000){Charbonneau}, {Brown}, {Latham}, \&
2026: {Mayor}}]{2000ApJ...529L..45C}
2027: {Charbonneau}, D., {Brown}, T.~M., {Latham}, D.~W., \& {Mayor}, M. 2000, \apjl,
2028: 529, L45, arXiv:astro-ph/9911436
2029:
2030: \bibitem[{{Chatterjee} {et~al.}(2007){Chatterjee}, {Ford}, \&
2031: {Rasio}}]{2007astro.ph..3166C}
2032: {Chatterjee}, S., {Ford}, E.~B., \& {Rasio}, F.~A. 2007, ArXiv Astrophysics
2033: e-prints, astro-ph/0703166
2034:
2035: \bibitem[{{Christian} {et~al.}(2008){Christian}, {Gibson}, {Simpson}, {Street},
2036: {Skillen}, {Pollacco}, {Collier Cameron}, {Stempels}, {Haswell}, {Horne},
2037: {Joshi}, {Keenan}, {Anderson}, {Bentley}, {Bouchy}, {Clarkson}, {Enoch},
2038: {Hebb}, {H{\'e}brard}, {Hellier}, {Irwin}, {Kane}, {Lister}, {Loeillet},
2039: {Maxted}, {Mayor}, {McDonald}, {Moutou}, {Norton}, {Parley}, {Pont},
2040: {Queloz}, {Ryans}, {Smalley}, {Smith}, {Todd}, {Udry}, {West}, {Wheatley}, \&
2041: {Wilson}}]{2008arXiv0806.1482C}
2042: {Christian}, D.~J. {et~al.} 2008, ArXiv e-prints, 806, 0806.1482
2043:
2044: \bibitem[{{Christiansen} {et~al.}(2009){Christiansen}, {Charbonneau},
2045: {A'Hearn}, {Deming}, {Holman}, {Ballard}, {Weldrake}, {Barry}, {Kuchner},
2046: {Livengood}, {Pedelty}, {Schultz}, {Hewagama}, {Sunshine}, {Wellnitz},
2047: {Hampton}, {Lisse}, {Seager}, \& {Veverka}}]{2009IAUS..253..301C}
2048: {Christiansen}, J.~L. {et~al.} 2009, in IAU Symposium, Vol. 253, IAU Symposium,
2049: 301--307
2050:
2051: \bibitem[{{Claret}(1995)}]{1995A&AS..109..441C}
2052: {Claret}, A. 1995, \aaps, 109, 441
2053:
2054: \bibitem[{{Cowling}(1938)}]{1938MNRAS..98..734C}
2055: {Cowling}, T.~G. 1938, \mnras, 98, 734
2056:
2057: \bibitem[{{Deeg} {et~al.}(2008){Deeg}, {Oca{\~n}a}, {Kozhevnikov},
2058: {Charbonneau}, {O'Donovan}, \& {Doyle}}]{2008A&A...480..563D}
2059: {Deeg}, H.~J., {Oca{\~n}a}, B., {Kozhevnikov}, V.~P., {Charbonneau}, D.,
2060: {O'Donovan}, F.~T., \& {Doyle}, L.~R. 2008, \aap, 480, 563, arXiv:0801.2186
2061:
2062: \bibitem[{{Dobbs-Dixon} {et~al.}(2004){Dobbs-Dixon}, {Lin}, \&
2063: {Mardling}}]{2004ApJ...610..464D}
2064: {Dobbs-Dixon}, I., {Lin}, D.~N.~C., \& {Mardling}, R.~A. 2004, \apj, 610, 464,
2065: arXiv:astro-ph/0408191
2066:
2067: \bibitem[{{Dodson-Robinson} \& {Bodenheimer}(2009)}]{2009arXiv0901.0582D}
2068: {Dodson-Robinson}, S.~E., \& {Bodenheimer}, P. 2009, ArXiv e-prints, 0901.0582
2069:
2070: \bibitem[{{Eggleton} \& {Kiseleva-Eggleton}(2001)}]{2001ApJ...562.1012E}
2071: {Eggleton}, P.~P., \& {Kiseleva-Eggleton}, L. 2001, \apj, 562, 1012,
2072: arXiv:astro-ph/0104126
2073:
2074: \bibitem[{{Fabrycky}(2008{\natexlab{a}})}]{2008ApJ...677L.117F}
2075: {Fabrycky}, D. 2008{\natexlab{a}}, \apjl, 677, L117, 0803.1839
2076:
2077: \bibitem[{{Fabrycky} \& {Tremaine}(2007)}]{2007ApJ...669.1298F}
2078: {Fabrycky}, D., \& {Tremaine}, S. 2007, \apj, 669, 1298, arXiv:0705.4285
2079:
2080: \bibitem[{{Fabrycky}(2008{\natexlab{b}})}]{2008arXiv0806.4314F}
2081: {Fabrycky}, D.~C. 2008{\natexlab{b}}, ArXiv e-prints, 806, 0806.4314
2082:
2083: \bibitem[{{Fabrycky} {et~al.}(2007){Fabrycky}, {Johnson}, \&
2084: {Goodman}}]{2007ApJ...665..754F}
2085: {Fabrycky}, D.~C., {Johnson}, E.~T., \& {Goodman}, J. 2007, \apj, 665, 754,
2086: arXiv:astro-ph/0703418
2087:
2088: \bibitem[{{Fabrycky} \& {Winn}(2009)}]{2009arXiv0902.0737F}
2089: {Fabrycky}, D.~C., \& {Winn}, J.~N. 2009, ArXiv e-prints, 0902.0737
2090:
2091: \bibitem[{Ferraz-Mello {et~al.}(2008)Ferraz-Mello, Rodr{\'i}guez, \&
2092: Hussmann}]{Ferraz-Mello2008}
2093: Ferraz-Mello, S., Rodr{\'i}guez, A., \& Hussmann, H. 2008, 0712.1156
2094:
2095: \bibitem[{{Ford} \& {Holman}(2007)}]{2007ApJ...664L..51F}
2096: {Ford}, E.~B., \& {Holman}, M.~J. 2007, \apjl, 664, L51, arXiv:0705.0356
2097:
2098: \bibitem[{{Ford} {et~al.}(2005){Ford}, {Lystad}, \&
2099: {Rasio}}]{2005Natur.434..873F}
2100: {Ford}, E.~B., {Lystad}, V., \& {Rasio}, F.~A. 2005, \nat, 434, 873,
2101: arXiv:astro-ph/0502441
2102:
2103: \bibitem[{{Gillon} {et~al.}(2008){Gillon}, {Anderson}, {Demory}, {Wilson},
2104: {Hellier}, {Queloz}, \& {Waelkens}}]{2008arXiv0806.4911G}
2105: {Gillon}, M., {Anderson}, D.~R., {Demory}, B.~., {Wilson}, D.~M., {Hellier},
2106: C., {Queloz}, D., \& {Waelkens}, C. 2008, ArXiv e-prints, 806, 0806.4911
2107:
2108: \bibitem[{{Goldreich} \& {Soter}(1966)}]{1966Icar....5..375G}
2109: {Goldreich}, P., \& {Soter}, S. 1966, Icarus, 5, 375
2110:
2111: \bibitem[{{Gomes} {et~al.}(2005){Gomes}, {Levison}, {Tsiganis}, \&
2112: {Morbidelli}}]{2005Natur.435..466G}
2113: {Gomes}, R., {Levison}, H.~F., {Tsiganis}, K., \& {Morbidelli}, A. 2005, \nat,
2114: 435, 466
2115:
2116: \bibitem[{{Goodman}(2009)}]{2009arXiv0901.3279G}
2117: {Goodman}, J. 2009, ArXiv e-prints, 0901.3279
2118:
2119: \bibitem[{{Gu} \& {Ogilvie}(2009)}]{2009arXiv0901.3401G}
2120: {Gu}, P.-G., \& {Ogilvie}, G.~I. 2009, ArXiv e-prints, 0901.3401
2121:
2122: \bibitem[{{Guillot}(2005)}]{2005AREPS..33..493G}
2123: {Guillot}, T. 2005, Annual Review of Earth and Planetary Sciences, 33, 493,
2124: arXiv:astro-ph/0502068
2125:
2126: \bibitem[{{Guillot} {et~al.}(2006){Guillot}, {Santos}, {Pont}, {Iro}, {Melo},
2127: \& {Ribas}}]{2006A&A...453L..21G}
2128: {Guillot}, T., {Santos}, N.~C., {Pont}, F., {Iro}, N., {Melo}, C., \& {Ribas},
2129: I. 2006, \aap, 453, L21, arXiv:astro-ph/0605751
2130:
2131: \bibitem[{{Hebb} {et~al.}(2009){Hebb}, {Collier-Cameron}, {Loeillet},
2132: {Pollacco}, {H{\'e}brard}, {Street}, {Bouchy}, {Stempels}, {Moutou},
2133: {Simpson}, {Udry}, {Joshi}, {West}, {Skillen}, {Wilson}, {McDonald},
2134: {Gibson}, {Aigrain}, {Anderson}, {Benn}, {Christian}, {Enoch}, {Haswell},
2135: {Hellier}, {Horne}, {Irwin}, {Lister}, {Maxted}, {Mayor}, {Norton}, {Parley},
2136: {Pont}, {Queloz}, {Smalley}, \& {Wheatley}}]{2009ApJ...693.1920H}
2137: {Hebb}, L. {et~al.} 2009, \apj, 693, 1920, 0812.3240
2138:
2139: \bibitem[{{Helled} \& {Schubert}(2009)}]{2009arXiv0903.1997H}
2140: {Helled}, R., \& {Schubert}, G. 2009, ArXiv e-prints, 0903.1997
2141:
2142: \bibitem[{{Heyl} \& {Gladman}(2007)}]{2007MNRAS.377.1511H}
2143: {Heyl}, J.~S., \& {Gladman}, B.~J. 2007, \mnras, 377, 1511,
2144: arXiv:astro-ph/0610267
2145:
2146: \bibitem[{{Holman} \& {Murray}(2005)}]{2005Sci...307.1288H}
2147: {Holman}, M.~J., \& {Murray}, N.~W. 2005, Science, 307, 1288
2148:
2149: \bibitem[{{Holman} {et~al.}(2007){Holman}, {Winn}, {Latham}, {O'Donovan},
2150: {Charbonneau}, {Torres}, {Sozzetti}, {Fernandez}, \&
2151: {Everett}}]{2007ApJ...664.1185H}
2152: {Holman}, M.~J. {et~al.} 2007, \apj, 664, 1185, 0704.2907
2153:
2154: \bibitem[{{Hood} {et~al.}(2008){Hood}, {Wood}, {Seager}, \& {Collier
2155: Cameron}}]{2008arXiv0807.1561H}
2156: {Hood}, B., {Wood}, K., {Seager}, S., \& {Collier Cameron}, A. 2008, ArXiv
2157: e-prints, 807, 0807.1561
2158:
2159: \bibitem[{{Hut}(1981)}]{1981A&A....99..126H}
2160: {Hut}, P. 1981, \aap, 99, 126
2161:
2162: \bibitem[{{Iorio}(2006)}]{2006NewA...11..490I}
2163: {Iorio}, L. 2006, New Astronomy, 11, 490, arXiv:gr-qc/0505107
2164:
2165: \bibitem[{Jackson {et~al.}(2008)Jackson, Greenberg, \& Barnes}]{Jackson2008}
2166: Jackson, B., Greenberg, R., \& Barnes, R. 2008, 0801.0716
2167:
2168: \bibitem[{{Johns-Krull} {et~al.}(2008){Johns-Krull}, {McCullough}, {Burke},
2169: {Valenti}, {Janes}, {Heasley}, {Prato}, {Bissinger}, {Fleenor}, {Foote},
2170: {Garcia-Melendo}, {Gary}, {Howell}, {Mallia}, {Masi}, \&
2171: {Vanmunster}}]{2008ApJ...677..657J}
2172: {Johns-Krull}, C.~M. {et~al.} 2008, \apj, 677, 657, 0712.4283
2173:
2174: \bibitem[{{Jordan} \& {Bakos}(2008)}]{JB08}
2175: {Jordan}, A., \& {Bakos}, G.~A. 2008, ArXiv e-prints, 806, 0806.0630
2176:
2177: \bibitem[{{Joshi} {et~al.}(2009){Joshi}, {Pollacco}, {Cameron}, {Skillen},
2178: {Simpson}, {Steele}, {Street}, {Stempels}, {Christian}, {Hebb}, {Bouchy},
2179: {Gibson}, {H{\'e}brard}, {Keenan}, {Loeillet}, {Meaburn}, {Moutou},
2180: {Smalley}, {Todd}, {West}, {Anderson}, {Bentley}, {Enoch}, {Haswell},
2181: {Hellier}, {Horne}, {Irwin}, {Lister}, {McDonald}, {Maxted}, {Mayor},
2182: {Norton}, {Parley}, {Perrier}, {Pont}, {Queloz}, {Ryans}, {Smith}, {Udry},
2183: {Wheatley}, \& {Wilson}}]{2009MNRAS.392.1532J}
2184: {Joshi}, Y.~C. {et~al.} 2009, \mnras, 392, 1532, 0806.1478
2185:
2186: \bibitem[{{Joshi} {et~al.}(2008){Joshi}, {Pollacco}, {Collier Cameron},
2187: {Skillen}, {Simpson}, {Steele}, {Street}, {Stempels}, {Bouchy}, {Christian},
2188: {Gibson}, {Hebb}, {Hebrard}, {Keenan}, {Loeillet}, {Meaburn}, {Moutou},
2189: {Smalley}, {Todd}, {West}, {Anderson}, {Bentley}, {Enoch}, {Haswell},
2190: {Hellier}, {Horne}, {Irwin}, {Lister}, {McDonald}, {Maxted}, {Mayor},
2191: {Norton}, {Parley}, {Perrier}, {Pont}, {Queloz}, {Ryans}, {Smith}, {Udry},
2192: {Wheatley}, \& {Wilson}}]{2008arXiv0806.1478J}
2193: ------. 2008, ArXiv e-prints, 806, 0806.1478
2194:
2195: \bibitem[{{Kallrath} {et~al.}(1999){Kallrath}, {Milone}, {Kallrath}, \&
2196: {Milone}}]{1999ebs..conf.....K}
2197: {Kallrath}, J., {Milone}, E.~F., {Kallrath}, J., \& {Milone}, E.~F., eds. 1999,
2198: {Eclipsing binary stars : modeling and analysis}
2199:
2200: \bibitem[{{Kjeldsen} {et~al.}(2008){Kjeldsen}, {Bedding}, \&
2201: {Christensen-Dalsgaard}}]{2008arXiv0807.0508K}
2202: {Kjeldsen}, H., {Bedding}, T.~R., \& {Christensen-Dalsgaard}, J. 2008, ArXiv
2203: e-prints, 807, 0807.0508
2204:
2205: \bibitem[{{Knutson} {et~al.}(2007{\natexlab{a}}){Knutson}, {Charbonneau},
2206: {Allen}, {Fortney}, {Agol}, {Cowan}, {Showman}, {Cooper}, \&
2207: {Megeath}}]{2007Natur.447..183K}
2208: {Knutson}, H.~A. {et~al.} 2007{\natexlab{a}}, \nat, 447, 183, arXiv:0705.0993
2209:
2210: \bibitem[{{Knutson} {et~al.}(2007{\natexlab{b}}){Knutson}, {Charbonneau},
2211: {Noyes}, {Brown}, \& {Gilliland}}]{2007ApJ...655..564K}
2212: {Knutson}, H.~A., {Charbonneau}, D., {Noyes}, R.~W., {Brown}, T.~M., \&
2213: {Gilliland}, R.~L. 2007{\natexlab{b}}, \apj, 655, 564, arXiv:astro-ph/0603542
2214:
2215: \bibitem[{{Koch} {et~al.}(2006){Koch}, {Borucki}, {Basri}, {Brown}, {Caldwell},
2216: {Christensen-Dalsgaard}, {Cochran}, {Dunham}, {Gautier}, {Geary},
2217: {Gilliland}, {Jenkins}, {Kondo}, {Latham}, {Lissauer}, \&
2218: {Monet}}]{2006Ap&SS.304..391K}
2219: {Koch}, D. {et~al.} 2006, \apss, 304, 391
2220:
2221: \bibitem[{{Kopal}(1959)}]{1959cbs..book.....K}
2222: {Kopal}, Z. 1959, {Close binary systems} (The International Astrophysics
2223: Series, London: Chapman \& Hall, 1959)
2224:
2225: \bibitem[{{Kopal}(1978)}]{1978ASSL...68.....K}
2226: {Kopal}, Z., ed. 1978, Astrophysics and Space Science Library, Vol.~68,
2227: {Dynamics of Close Binary Systems}
2228:
2229: \bibitem[{{Kozai}(1959)}]{1959AJ.....64..367K}
2230: {Kozai}, Y. 1959, \aj, 64, 367
2231:
2232: \bibitem[{{Lanza} {et~al.}(2009){Lanza}, {Pagano}, {Leto}, {Messina},
2233: {Aigrain}, {Alonso}, {Auvergne}, {Baglin}, {Barge}, {Bonomo}, {Boumier},
2234: {Collier Cameron}, {Comparato}, {Cutispoto}, {de Medeiros}, {Foing},
2235: {Kaiser}, {Moutou}, {Parihar}, {Silva-Valio}, \&
2236: {Weiss}}]{2009A&A...493..193L}
2237: {Lanza}, A.~F. {et~al.} 2009, \aap, 493, 193, 0811.0461
2238:
2239: \bibitem[{{Laughlin} {et~al.}(2005){Laughlin}, {Marcy}, {Vogt}, {Fischer}, \&
2240: {Butler}}]{2005ApJ...629L.121L}
2241: {Laughlin}, G., {Marcy}, G.~W., {Vogt}, S.~S., {Fischer}, D.~A., \& {Butler},
2242: R.~P. 2005, \apjl, 629, L121
2243:
2244: \bibitem[{{Lee} {et~al.}(2009){Lee}, {Kim}, {Kim}, {Koch}, {Lee}, {Kim}, \&
2245: {Park}}]{2009AJ....137.3181L}
2246: {Lee}, J.~W., {Kim}, S.-L., {Kim}, C.-H., {Koch}, R.~H., {Lee}, C.-U., {Kim},
2247: H.-I., \& {Park}, J.-H. 2009, \aj, 137, 3181, 0811.3807
2248:
2249: \bibitem[{{Levrard} {et~al.}(2007){Levrard}, {Correia}, {Chabrier}, {Baraffe},
2250: {Selsis}, \& {Laskar}}]{2007A&A...462L...5L}
2251: {Levrard}, B., {Correia}, A.~C.~M., {Chabrier}, G., {Baraffe}, I., {Selsis},
2252: F., \& {Laskar}, J. 2007, \aap, 462, L5, arXiv:astro-ph/0612044
2253:
2254: \bibitem[{{Levrard} {et~al.}(2009){Levrard}, {Winisdoerffer}, \&
2255: {Chabrier}}]{2009ApJ...692L...9L}
2256: {Levrard}, B., {Winisdoerffer}, C., \& {Chabrier}, G. 2009, \apjl, 692, L9,
2257: 0901.2048
2258:
2259: \bibitem[{{Loeb}(2005)}]{2005ApJ...623L..45L}
2260: {Loeb}, A. 2005, \apjl, 623, L45, arXiv:astro-ph/0501548
2261:
2262: \bibitem[{{Loeb}(2008)}]{2008arXiv0807.0835L}
2263: ------. 2008, ArXiv e-prints, 807, 0807.0835
2264:
2265: \bibitem[{{L{\'o}pez-Morales} \& {Seager}(2007)}]{2007ApJ...667L.191L}
2266: {L{\'o}pez-Morales}, M., \& {Seager}, S. 2007, \apjl, 667, L191,
2267: arXiv:0708.0822
2268:
2269: \bibitem[{{Mandel} \& {Agol}(2002)}]{2002ApJ...580L.171M}
2270: {Mandel}, K., \& {Agol}, E. 2002, \apjl, 580, L171, arXiv:astro-ph/0210099
2271:
2272: \bibitem[{{Mardling}(2007)}]{2007MNRAS.382.1768M}
2273: {Mardling}, R.~A. 2007, \mnras, 382, 1768, arXiv:0706.0224
2274:
2275: \bibitem[{{Mardling} \& {Lin}(2002)}]{2002ApJ...573..829M}
2276: {Mardling}, R.~A., \& {Lin}, D.~N.~C. 2002, \apj, 573, 829
2277:
2278: \bibitem[{{Matsumura} {et~al.}(2008){Matsumura}, {Takeda}, \&
2279: {Rasio}}]{2008ApJ...686L..29M}
2280: {Matsumura}, S., {Takeda}, G., \& {Rasio}, F.~A. 2008, \apjl, 686, L29,
2281: 0808.3724
2282:
2283: \bibitem[{{Mayor} {et~al.}(2004){Mayor}, {Udry}, {Naef}, {Pepe}, {Queloz},
2284: {Santos}, \& {Burnet}}]{2004A&A...415..391M}
2285: {Mayor}, M., {Udry}, S., {Naef}, D., {Pepe}, F., {Queloz}, D., {Santos}, N.~C.,
2286: \& {Burnet}, M. 2004, \aap, 415, 391, arXiv:astro-ph/0310316
2287:
2288: \bibitem[{{Miralda-Escud{\'e}}(2002)}]{2002ApJ...564.1019M}
2289: {Miralda-Escud{\'e}}, J. 2002, \apj, 564, 1019, arXiv:astro-ph/0104034
2290:
2291: \bibitem[{{Murray} \& {Dermott}(1999)}]{1999ssd..book.....M}
2292: {Murray}, C.~D., \& {Dermott}, S.~F. 1999, {Solar system dynamics} (Solar
2293: system dynamics by Murray, C.~D., 1999)
2294:
2295: \bibitem[{{Nesvorn{\'y}} \& {Morbidelli}(2008)}]{2008ApJ...688..636N}
2296: {Nesvorn{\'y}}, D., \& {Morbidelli}, A. 2008, \apj, 688, 636
2297:
2298: \bibitem[{{O'Donovan} {et~al.}(2009){O'Donovan}, {Charbonneau}, {Harrington},
2299: {Seager}, {Deming}, \& {Knutson}}]{2009IAUS..253..536O}
2300: {O'Donovan}, F.~T., {Charbonneau}, D., {Harrington}, J., {Seager}, S.,
2301: {Deming}, D., \& {Knutson}, H.~A. 2009, in IAU Symposium, Vol. 253, IAU
2302: Symposium, 536--539
2303:
2304: \bibitem[{{P{\'a}l} {et~al.}(2009){P{\'a}l}, {Bakos}, {Noyes}, \&
2305: {Torres}}]{2009IAUS..253..428P}
2306: {P{\'a}l}, A., {Bakos}, G.~{\'A}., {Noyes}, R.~W., \& {Torres}, G. 2009, in IAU
2307: Symposium, Vol. 253, IAU Symposium, 428--431
2308:
2309: \bibitem[{{P{\'a}l} \& {Kocsis}(2008)}]{PK08}
2310: {P{\'a}l}, A., \& {Kocsis}, B. 2008, ArXiv e-prints, 806, 0806.0629
2311:
2312: \bibitem[{{Peale} {et~al.}(1979){Peale}, {Cassen}, \&
2313: {Reynolds}}]{1979Sci...203..892P}
2314: {Peale}, S.~J., {Cassen}, P., \& {Reynolds}, R.~T. 1979, Science, 203, 892
2315:
2316: \bibitem[{{Pont} {et~al.}(2007{\natexlab{a}}){Pont}, {Gilliland}, {Moutou},
2317: {Charbonneau}, {Bouchy}, {Brown}, {Mayor}, {Queloz}, {Santos}, \&
2318: {Udry}}]{2007A&A...476.1347P}
2319: {Pont}, F. {et~al.} 2007{\natexlab{a}}, \aap, 476, 1347, arXiv:0707.1940
2320:
2321: \bibitem[{{Pont} {et~al.}(2007{\natexlab{b}}){Pont}, {Moutou}, {Gillon},
2322: {Udalski}, {Bouchy}, {Fernandes}, {Gieren}, {Mayor}, {Mazeh}, {Minniti},
2323: {Melo}, {Naef}, {Pietrzynski}, {Queloz}, {Ruiz}, {Santos}, \&
2324: {Udry}}]{2007AA...465.1069P}
2325: ------. 2007{\natexlab{b}}, \aap, 465, 1069, arXiv:astro-ph/0610827
2326:
2327: \bibitem[{{Rafikov}(2008)}]{2008arXiv0807.0008R}
2328: {Rafikov}, R.~R. 2008, ArXiv e-prints, 807, 0807.0008
2329:
2330: \bibitem[{{Rodriguez} \& {Ferraz-Mello}(2009)}]{2009arXiv0903.0763R}
2331: {Rodriguez}, A., \& {Ferraz-Mello}, S. 2009, ArXiv e-prints, 0903.0763
2332:
2333: \bibitem[{{Rowe} {et~al.}(2007){Rowe}, {Matthews}, {Seager}, {Miller-Ricci},
2334: {Sasselov}, {Kuschnig}, {Guenther}, {Moffat}, {Rucinski}, {Walker}, \&
2335: {Weiss}}]{2007arXiv0711.4111R}
2336: {Rowe}, J.~F. {et~al.} 2007, ArXiv e-prints, 711, 0711.4111
2337:
2338: \bibitem[{{Russell}(1928)}]{1928MNRAS..88..641R}
2339: {Russell}, H.~N. 1928, \mnras, 88, 641
2340:
2341: \bibitem[{{Sahu} {et~al.}(2006){Sahu}, {Casertano}, {Bond}, {Valenti}, {Ed
2342: Smith}, {Minniti}, {Zoccali}, {Livio}, {Panagia}, {Piskunov}, {Brown},
2343: {Brown}, {Renzini}, {Rich}, {Clarkson}, \& {Lubow}}]{2006Natur.443..534S}
2344: {Sahu}, K.~C. {et~al.} 2006, \nat, 443, 534, arXiv:astro-ph/0610098
2345:
2346: \bibitem[{{Sartoretti} \& {Schneider}(1999)}]{1999A&AS..134..553S}
2347: {Sartoretti}, P., \& {Schneider}, J. 1999, \aaps, 134, 553
2348:
2349: \bibitem[{{Sasselov}(2003)}]{2003ApJ...596.1327S}
2350: {Sasselov}, D.~D. 2003, \apj, 596, 1327, arXiv:astro-ph/0303403
2351:
2352: \bibitem[{{Scharf}(2007)}]{2007ApJ...661.1218S}
2353: {Scharf}, C.~A. 2007, \apj, 661, 1218, arXiv:astro-ph/0702749
2354:
2355: \bibitem[{{Seager} \& {Hui}(2002)}]{2002ApJ...574.1004S}
2356: {Seager}, S., \& {Hui}, L. 2002, \apj, 574, 1004, arXiv:astro-ph/0204225
2357:
2358: \bibitem[{{Sing} \& {L{\'o}pez-Morales}(2009)}]{2009A&A...493L..31S}
2359: {Sing}, D.~K., \& {L{\'o}pez-Morales}, M. 2009, \aap, 493, L31, 0901.1876
2360:
2361: \bibitem[{{Soffel}(1989)}]{1989racm.book.....S}
2362: {Soffel}, M.~H. 1989, {Relativity in Astrometry, Celestial Mechanics and
2363: Geodesy} (Relativity in Astrometry, Celestial Mechanics and Geodesy, XIV, 208
2364: pp.~32 figs..~ Springer-Verlag Berlin Heidelberg New York.~ Also Astronomy
2365: and Astrophysics Library)
2366:
2367: \bibitem[{{Sotin} {et~al.}(2007){Sotin}, {Grasset}, \&
2368: {Mocquet}}]{2007Icar..191..337S}
2369: {Sotin}, C., {Grasset}, O., \& {Mocquet}, A. 2007, Icarus, 191, 337
2370:
2371: \bibitem[{Southworth(2008)}]{Southworth2008}
2372: Southworth, J. 2008, 0802.3764
2373:
2374: \bibitem[{{Sozzetti} {et~al.}(2009){Sozzetti}, {Torres}, {Charbonneau}, {Winn},
2375: {Korzennik}, {Holman}, {Latham}, {Laird}, {Fernandez}, {O'Donovan},
2376: {Mandushev}, {Dunham}, {Everett}, {Esquerdo}, {Rabus}, {Belmonte}, {Deeg},
2377: {Brown}, {Hidas}, \& {Baliber}}]{2009ApJ...691.1145S}
2378: {Sozzetti}, A. {et~al.} 2009, \apj, 691, 1145, 0809.4589
2379:
2380: \bibitem[{{Sterne}(1939{\natexlab{a}})}]{1939MNRAS..99..451S}
2381: {Sterne}, T.~E. 1939{\natexlab{a}}, \mnras, 99, 451
2382:
2383: \bibitem[{{Sterne}(1939{\natexlab{b}})}]{1939MNRAS..99..662S}
2384: ------. 1939{\natexlab{b}}, \mnras, 99, 662
2385:
2386: \bibitem[{{Swain} {et~al.}(2008){Swain}, {Vasisht}, \&
2387: {Tinetti}}]{2008Natur.452..329S}
2388: {Swain}, M.~R., {Vasisht}, G., \& {Tinetti}, G. 2008, \nat, 452, 329
2389:
2390: \bibitem[{{Thommes} {et~al.}(2008){Thommes}, {Bryden}, {Wu}, \&
2391: {Rasio}}]{2008ApJ...675.1538T}
2392: {Thommes}, E.~W., {Bryden}, G., {Wu}, Y., \& {Rasio}, F.~A. 2008, \apj, 675,
2393: 1538, arXiv:0706.1235
2394:
2395: \bibitem[{Torres {et~al.}(2008)Torres, Winn, \& Holman}]{Torres2008}
2396: Torres, G., Winn, J.~N., \& Holman, M.~J. 2008, 0801.1841
2397:
2398: \bibitem[{{Winn} \& {Holman}(2005)}]{2005ApJ...628L.159W}
2399: {Winn}, J.~N., \& {Holman}, M.~J. 2005, \apjl, 628, L159,
2400: arXiv:astro-ph/0506468
2401:
2402: \bibitem[{{Winn} {et~al.}(2007){Winn}, {Holman}, {Bakos}, {P{\'a}l}, {Johnson},
2403: {Williams}, {Shporer}, {Mazeh}, {Fernandez}, {Latham}, \&
2404: {Gillon}}]{2007AJ....134.1707W}
2405: {Winn}, J.~N. {et~al.} 2007, \aj, 134, 1707, arXiv:0707.1908
2406:
2407: \bibitem[{{Winn} {et~al.}(2009{\natexlab{a}}){Winn}, {Holman}, {Carter},
2408: {Torres}, {Osip}, \& {Beatty}}]{2009AJ....137.3826W}
2409: {Winn}, J.~N., {Holman}, M.~J., {Carter}, J.~A., {Torres}, G., {Osip}, D.~J.,
2410: \& {Beatty}, T. 2009{\natexlab{a}}, \aj, 137, 3826, 0901.4346
2411:
2412: \bibitem[{{Winn} {et~al.}(2008){Winn}, {Holman}, {Shporer}, {Fern{\'a}ndez},
2413: {Mazeh}, {Latham}, {Charbonneau}, \& {Everett}}]{2008AJ....136..267W}
2414: {Winn}, J.~N., {Holman}, M.~J., {Shporer}, A., {Fern{\'a}ndez}, J., {Mazeh},
2415: T., {Latham}, D.~W., {Charbonneau}, D., \& {Everett}, M.~E. 2008, \aj, 136,
2416: 267, arXiv:0804.2479
2417:
2418: \bibitem[{{Winn} {et~al.}(2009{\natexlab{b}}){Winn}, {Johnson}, {Fabrycky},
2419: {Howard}, {Marcy}, {Narita}, {Crossfield}, {Suto}, {Turner}, {Esquerdo}, \&
2420: {Holman}}]{2009arXiv0902.3461W}
2421: {Winn}, J.~N. {et~al.} 2009{\natexlab{b}}, ArXiv e-prints, 0902.3461
2422:
2423: \bibitem[{{Winn} {et~al.}(2006){Winn}, {Johnson}, {Marcy}, {Butler}, {Vogt},
2424: {Henry}, {Roussanova}, {Holman}, {Enya}, {Narita}, {Suto}, \&
2425: {Turner}}]{2006ApJ...653L..69W}
2426: ------. 2006, \apjl, 653, L69, arXiv:astro-ph/0609506
2427:
2428: \bibitem[{{Winn} {et~al.}(2005){Winn}, {Noyes}, {Holman}, {Charbonneau},
2429: {Ohta}, {Taruya}, {Suto}, {Narita}, {Turner}, {Johnson}, {Marcy}, {Butler},
2430: \& {Vogt}}]{2005ApJ...631.1215W}
2431: ------. 2005, \apj, 631, 1215, arXiv:astro-ph/0504555
2432:
2433: \bibitem[{{Wu} \& {Goldreich}(2002)}]{2002ApJ...564.1024W}
2434: {Wu}, Y., \& {Goldreich}, P. 2002, \apj, 564, 1024, arXiv:astro-ph/0108499
2435:
2436: \bibitem[{{Zharkov} \& {Trubitsyn}(1978)}]{1978ppi..book.....Z}
2437: {Zharkov}, V.~N., \& {Trubitsyn}, V.~P. 1978, {Physics of planetary interiors}
2438: (Astronomy and Astrophysics Series, Tucson: Pachart, 1978)
2439:
2440: \end{thebibliography}
2441:
2442: \end{document}
2443: