1:
2: \def\etal{{\it et al.\thinspace}}
3: \def\mearth{{\rm\,M_\oplus}}
4: \def\msun{{\rm\,M_\odot}}
5:
6: %\topmargin -0.5in
7:
8: %\documentclass[12pt,preprint]{aastex}
9: %\documentstyle[emulateapj,apjfonts]{article}
10: \documentclass[apjl]{emulateapj}
11:
12: %\received{}
13: %\accepted{}
14:
15: %\lefthead{}
16: %\righthead{}
17:
18: \begin{document}
19:
20: \shorttitle {Few habitable planets around low-mass stars}
21: \shortauthors{Raymond, Scalo \& Meadows}
22:
23: \title{A decreased probability of habitable planet formation around
24: low-mass stars}
25:
26: \author{Sean N. Raymond\altaffilmark{1,2,*}, John Scalo\altaffilmark{3,*},
27: \& Victoria S. Meadows\altaffilmark{4,*}}
28:
29: \altaffiltext{1}{Center for Astrophysics and Space Astronomy, University of
30: Colorado, 389 UCB, Boulder CO 80309-0389; raymond@lasp.colorado.edu}
31: \altaffiltext{2}{NASA Postdoctoral Program Fellow.}
32: \altaffiltext{3}{Department of Astronomy, University of Texas, Austin, TX; parrot@as.utexas.edu}
33: \altaffiltext{4}{Infrared Processing and Analysis Center, California
34: Institute of Technology, Pasadena, CA; vsm@ipac.caltech.edu}
35: \altaffiltext{*}{Member of NASA Astrobiology Institute}
36:
37: \begin{abstract}
38:
39: Smaller terrestrial planets ($\lesssim 0.3\mearth$) are less likely to retain
40: the substantial atmospheres and ongoing tectonic activity probably required to
41: support life. A key element in determining if suf{f}iciently massive
42: ``sustainably habitable'' planets can form is the availability of solid
43: planet-forming material. We use dynamical simulations of terrestrial planet
44: formation from planetary embryos and simple scaling arguments to explore the
45: implications of correlations between terrestrial planet mass, disk mass, and
46: the mass of the parent star. We assume that the protoplanetary disk mass
47: scales with stellar mass as $M_{disk}\propto f\,M_\star^h$, where $f$ measures
48: the relative disk mass, and $1/2 < h < 2$, so that disk mass decreases with
49: decreasing stellar mass. We consider systems without Jovian planets, based on
50: current models and observations for M stars. We assume the mass of a planet
51: formed in some annulus of a disk with given parameters is proportional to the
52: disk mass in that annulus, and show with a suite of simulations of late-stage
53: accretion that the adopted prescription is surprisingly accurate. Our results
54: suggest that the fraction of systems with suf{f}icient disk mass to form $>
55: 0.3\mearth$ habitable planets decreases for low-mass stars for every realistic
56: combination of parameters. This ``habitable fraction'' is small for stellar
57: masses below a mass in the interval 0.5 to 0.8 $\msun$, depending on disk
58: parameters, an interval that excludes most M stars. Radial mixing and
59: therefore water delivery are inef{f}icient in lower-mass disks commonly found
60: around low-mass stars, such that terrestrial planets in the habitable zones of
61: most low-mass stars are likely to be small and dry.
62: \end{abstract}
63:
64: \keywords{astrobiology --- planetary systems: formation ---
65: planetary systems: protoplanetary disks --- stars: low-mass, brown
66: dwarfs}
67:
68: \section{Introduction}
69:
70: The circumstellar habitable zone (HZ) is the distance annulus within which a
71: terrestrial planet can maintain liquid water on its surface. The average
72: distance of the HZ from the parent star depends primarily on the star's
73: luminosity, and is closer-in and narrower around lower-mass stars (Kasting
74: \etal 1993). For a planet to remain habitable, it must control its surface
75: temperature over long timescales, possibly via the CO$_2$-carbonate cycle
76: (Walker \etal 1981), which is enabled by a relatively thick atmosphere, a
77: hydrological cycle, and active plate tectonics. These planetary
78: characteristics are less likely to be sustained over long periods by less
79: massive terrestrial planets, or those with low water contents. Whether
80: potentially habitable planets of suf{f}icient mass will form depends on the
81: surface density of rocky material in their parent protoplanetary disks. Disks
82: with higher surface densities tend to form a smaller number of more massive
83: planets (Wetherill 1996; Raymond \etal 2004, Kokubo \etal 2006), and can
84: increase terrestrial planet water content by scattering more distant,
85: water-rich material into the HZ (Morbidelli \etal 2000; Raymond \etal 2004,
86: 2005a).
87:
88: Here we investigate the formation, masses and water contents of terrestrial
89: planets as a function of stellar mass by assuming a parameterized relationship
90: between disk and stellar mass. We adopt a simple model relating the HZ planet
91: mass to the disk mass, based on results in the literature. We use a set of
92: numerical simulations of terrestrial planet growth around stars between 0.2
93: and 1 $\msun$ to test the validity of the model. The simulations agree well
94: with the model, and justify an exploration of the consequences of such a
95: planet-disk-stellar mass relation, and in addition allow us to consider how
96: water delivery is af{f}ected as a function of stellar mass.
97:
98: Our main result is to derive probabilistic limits on the stellar mass below
99: which sustainably habitable planets cannot form because of insuf{f}icient disk
100: mass available for accretion to a critical planet mass of 0.3 $\mearth$. In
101: addition, we show how this stellar mass limit depends on the parameters of the
102: disk-star mass relationship, which is currently poorly constrained. In a
103: statistical framework, we show that the fraction of stars capable of forming
104: habitable planets {\it in situ} decreases quickly with decreasing stellar
105: mass, and that terrestrial planets in the habitable zones of M stars ($M_\star
106: \lesssim 0.5 \msun$; e.g., Reid \& Hawley 2000) are likely to be small and
107: dry. An interesting test of our models comes from the recent detection of a
108: $\sim 5 \mearth$ planet close to the habitable zone of the M dwarf Gliese 581
109: (Udry \etal 2007). The implications of this system for our formation models
110: are discussed in section 4.
111:
112: To simplify our calculations, we assume there are no gas giant (Jovian)
113: planets present, and we test the ef{f}ects of this assumption in section 3.3.
114: Systems without giant planets may represent the majority of planetary systems
115: with habitable planets. Available estimates of the fraction of solar-like
116: stars with massive planets (Tabachnick \& Tremaine 2002, Lineweaver \& Grether
117: 2004, Beer \etal 2004; Fischer \& Valenti 2005) allow for 70-90\% of stars to
118: have planetary systems without Jovians, and an even larger fraction for
119: lower-mass stars (Endl \etal 2006; Butler \etal 2006). In addition, Scholz
120: \etal (2006) point out that estimated submm and mm outer disk masses show that
121: only a small fraction of low-mass stars or substellar objects have suf{f}icient
122: disk mass to form a planet with the mass of Jupiter (but see Hartmann \etal
123: 2006). The lack of a correlation between the presence of debris disks and the
124: metallicity of the parent star (Greaves \etal 2006; Beichman \etal 2006;
125: Moro-Mart\'in \etal 2007), unlike the correlation known for exogiant planets
126: (Gonzalez 1997, Santos \etal, Fischer \& Valenti 2005), also supports the
127: likelihood of gas-giant-free planetary systems. There are observational (Endl
128: \etal 2006; Butler \etal 2006) and theoretical (Ward \& Hahn 1995; Adams \etal
129: 2004; Laughlin \etal 2004, Ida \& Lin 2005) reasons to think that the fraction
130: of M stars with Jovian planets is significantly smaller than for solar-mass
131: stars.
132:
133: \section{Methods}
134:
135: \subsection{Habitable Planet Mass Limit}
136: Following Williams \etal (1997), we assume a lower mass limit for planetary
137: habitability that supports active plate tectonics for several Gyr.
138: Significant mass is a necessary, but not suf{f}icient, requirement for
139: sustainable planetary habitability (e.g., Lissauer 1999). Using the
140: radioactive flux limit from Williams \etal (1997), the critical planetary mass
141: scales like $\rho^{-2} e^{3 \lambda t}$, where $\rho$ is the planet's bulk
142: density, $\lambda=1.5 \times 10^{10} yr^{-1}$ for $^{238}$U, and $t$ is the
143: duration of tectonic activity. For an active lifetime of 5 Gyr, and iterating
144: on the density to be consistent with the resulting critical mass, we get a
145: critical mass of 0.3 $\mearth$ for a density of 4.5 $g\, cm^{-3}$. This
146: result is sensitive to the plate tectonic activity timescale associated with
147: sustained habitability, which is fairly arbitrary. The average age of stars
148: in the Galactic disk is about 5 Gyr, so the choice of 5 Gyr allows about half
149: of such stars with sustainably habitable planets to be on the main sequence
150: today, for a roughly constant star formation rate. Throughout the paper we
151: consider 0.3 $\mearth$ as our critical mass limit for planetary habitability,
152: and we discuss the implications of changing this limit in section 4.
153:
154: \subsection{Protoplanetary Disk Properties}
155:
156: We assume the surface density of protoplanetary disks, $\Sigma$, to scale
157: with heliocentric distance $r$ as:
158: \begin{equation}
159: \Sigma(r) = \Sigma_1\, f\,Z \left(\frac{r}{\rm 1 AU}\right)^{-\alpha} \left(\frac{M_{\star}}{M_{\odot}}\right)^h,
160: \end{equation}
161: where $\Sigma_1$ is the surface density of solids at 1 AU in the minimum-mass
162: solar nebula (MMSN) model ($\approx 6 g \, cm^{-2}$), $h$ adjusts the density
163: scaling with $M_{\star}$, $f$ is a scaling factor for the disk ($f$=1 for
164: MMSN), $Z$ is the stellar metallicity (1.0=solar), and $\alpha$ moderates the
165: steepness of the density profile ($\alpha = 3/2$ in the MMSN model; Hayashi
166: 1981, but more recent reconstructions using an accretion disk model as a
167: constraint gives $\alpha = 1/2$; Davis 2005). Most disk models that include
168: viscous and irradiation heating give $\alpha$ between about 1/2 and 1 for the
169: inner disk (e.g. D'Alessio \etal 1998; Garaud \& Lin 2007, Dullemond et
170: al. 2007), and high-resolution submillimeter observations suggest $\alpha$
171: between 1/2 and 1 in the outer disk (Mundy \etal 2000, Looney et al. 2003,
172: Andrews and Williams 2006). We take $\alpha = 1$ as a fiducial value but
173: consider values in the range 1/2 to 3/2.
174:
175: Present estimates of disk masses in nearby star-forming regions (see Andre \&
176: Montmerle 1994, Eisner \& Carpenter 2003, Andrews \& Williams 2005) give a
177: large spread, with median or average masses usually somewhat below the MMSN
178: value, but a significant fraction more massive, including a very massive disks
179: observed in both Orion and Taurus (e.g. Williams \etal 2005, Eisner \&
180: Carpenter 2006). At the other end of the mass spectrum, the estimates by
181: Williams \etal (2005) of mean masses for 18 Orion proplyds below the
182: $3-\sigma$ detection limit yielded $8 \times 10^{-4} \msun$, or only $3
183: \mearth$ in solids for a gas-to-dust ratio of 100. There also exists a factor
184: of 3-5 range in metallicity for thin-disk stars (Nordstrom \etal 2004), which
185: adjusts the ef{f}ective gas-to-dust ratio.
186:
187: There is considerable uncertainty in the scaling between stellar mass
188: $M_{\star}$ and disk mass $M_{disk}$. The most promising lead comes from
189: estimates of outer disk masses using submm and mm techniques (Beckwith et
190: al. 1990), which remain uncertain because of adopted dust temperatures,
191: opacities, distances, and radiative transfer model. Scholz \etal (2006)
192: estimated disk masses for brown dwarf disks and combined their results with
193: previous determinations for stellar and substellar objects from 0.02 to 3
194: $M_\odot$. Their Fig. 3 shows that the ratio of disk mass to stellar mass is
195: roughly constant with stellar mass, with a large amount of scatter. This
196: hints that $h \approx 1$, with a large range in $f$. Many of the data points
197: from Scholz \etal (2006) are only upper limits, but we think extreme cases
198: $h=0$ or $h>2$ would be apparent in this data.
199:
200: Additional clues come from the observed stellar accretion rate $\dot{M}$,
201: which scales as $\dot{M}\propto M_{\star}^2$, with a large dispersion
202: (Muzerolle \etal 2003, 2005, Natta \etal 2006). It might therefore seem
203: natural to assume that $M_{disk} \propto M_\star^2$, i.e. $h$=2, assuming that
204: the viscous timescale is independent of stellar mass (e.g., Ida \& Lin 2005).
205: On the contrary, nearly all accretion disk models predict $\dot{M}\propto
206: M_{\star}^a$ with $a \sim 1$. This matter is still unresolved; several
207: solutions have been proposed (see Padoan \etal 2006; Alexander and Armitage
208: 2006; Dullemond \etal 2006; Hartmann \etal 2006; Gregory \etal 2007) including
209: selection/detection limitations (Clarke \& Pringle 2006). Most of these
210: models are consistent with $h \approx 1$. Although $h$ is currently poorly
211: constrained, we emphasize its importance for understanding whether sustainable
212: HZ planets can form around low-mass stars. Given the results of Scholz \etal
213: (2006), we choose $h=1$ as our fiducial case, but explore the ef{f}ects of $h$
214: from 0.5 to 2.
215:
216: %proposed solutions include 1) a disk size - stellar mass relation (Alexander
217: %and Armitage 2006; but see Vicente \& Alves 2005); 2) Bondi-Hoyle accretion of
218: %ambient cloud gas (Padoan \etal 2006); 3) stellar mass trends in star
219: %formation initial conditions (Dullemond \etal 2006); or 4) selection/detection
220: %limitations (Clarke \& Pringle 2006).
221:
222: \subsection{Simulations and Models}
223:
224: We performed dynamical simulations to evaluate the scaling between terrestrial
225: planet mass and disk properties. For all runs, our starting conditions
226: reflect our fiducial case, with $fZ=1$, $h=1$, and $\alpha=1$. We vary the
227: mass of the central star from 0.2 to 1 $\msun$. Given the surface density
228: profile from Eq (1), we assume that planetary embryos have formed throughout
229: the inner disk of each star. We assume that the disk evolved following the
230: standard model of terrestrial planet growth (e.g. Chambers 2004): grains
231: coalesced to form km-sized planetesimals; and embryos formed from
232: planetesimals via runaway and oligarchic growth (e.g., Kokubo \& Ida 1998),
233: spaced by 5-10 mutual Hill radii ($R_{H,m} = 0.5 (a_1 + a_2) (M_1+M_2/3
234: M_\star)^{1/3}$; $a_1$ and $a_2$ are the orbital radii and $M_1$ and $M_2$ the
235: masses of two adjacent embryos). For the case of 1 $\msun$, we generated a
236: population of $\sim$ 75 embryos from 0.5 to 4 AU, totaling 4.95
237: $\mearth$.\footnote{Note that the outer edge of 4 AU for the embryo disk is
238: chosen based on previous simulations showing that the feeding zone of a
239: terrestrial planet in the habitable zone does not extend beyond 4 AU (Raymond
240: \etal 2007).} For other stellar masses, we assume the population of embryos
241: has the same temperature, as we are interested in HZ planets (defined to have
242: the same temperature), and the feeding zone boundaries should scale in the
243: same way as the HZ mean distance. We therefore scale the inner and outer
244: boundaries by the stellar flux (i.e., as the luminosity $L_\star^{1/2}$),
245: using a mass-luminosity relation from Scalo \etal (2007) that is a fit to data
246: from Hillenbrand \& White (2004): $y = 4.101 x^3 + 8.162 x^2 + 7.108 x +
247: 0.065$, where $y=log(L_\star)$ and $x=log(M_\star)$. Table 1 summarizes our
248: starting conditions for each stellar mass. Note the variations in the number
249: of embryos $N$ included in each simulation, from $\sim$ 75 for the 1 $\msun$
250: simulations to almost 200 for the 0.2 $\msun$ simulations. This variation in
251: $N$ is due to the Hill radius being very small at the small orbital distances
252: studied for lower-mass stars. In other words, models of embryo growth predict
253: that there really are more embryos in these inner disks than in the more
254: distant regions studied for higher-mass stars (e.g., Kokubo \& Ida 1998).
255: Kokubo \etal (2006) showed that the bulk properties of accreted planets are
256: not particularly sensitive to such variations in $N$, so we are not concerned
257: that our choice of embryo formation models will af{f}ect our results. Note also
258: that the $0.2 \msun$ star has less than 0.3 $\mearth$ in its entire disk, so
259: it serves primarily to tell us about the ef{f}iciency with which the disk mass
260: is used to make planets under these conditions, but cannot form a habitable
261: planet according to our criterion. Each numerical experiment was run
262: independently four times starting from dif{f}erent random initial embryo
263: positions for a total of twenty simulations. We give embryos random starting
264: eccentricities $<$0.02 and inclinations $<$0.1 degrees.
265:
266: Embryos are assigned water contents based on values from our Solar System,
267: where asteroids beyond $\sim$2.5 AU contain significant quantities of water
268: (Abe \etal 2000; see Fig. 2 in Raymond \etal 2004). We assumed that this
269: boundary between ``dry'' and ``wet'' embryos (called the ``water line'' in
270: Table 1) scales with the stellar flux.\footnote{We use the term ``water line''
271: instead of the more common ``snow line'' because we are not necessarily
272: assuming this to be the location where the temperature drops below $\sim 170$
273: K and water ice can condense (2.7 AU in the MMSN; Hayashi 1981). Rather, we
274: are simply assuming this boundary to divide dry and wet material. The water
275: line may actually be located somewhat interior to the snow line, as water-rich
276: bodies can be shifted inward via either gas drag or by eccentricity pumping
277: during embryo formation (e.g., Cyr \etal 1998; Kokubo \& Ida 1998).} Embryos
278: beyond this boundary contain 5\% water by mass, and those inside are dry
279: (Table 1). We integrated each simulation for 200 Myr using the hybrid
280: integrator {\it Mercury} (Chambers 1999). Timesteps were chosen to sample the
281: innermost body's orbit at least twenty times (e.g., Rauch \& Holman 1999;
282: Levison \& Duncan 2000; see Table 1), and so varied with each set of
283: simulations, from 6 days for 1 $\msun$ to 0.2 days for 0.2 $\msun$.
284: Collisions were treated as inelastic mergers conserving water.
285:
286: \section{Results}
287:
288: \subsection{Terrestrial Planet Mass vs Stellar Mass}
289:
290: Figure~\ref{fig:aet} shows the evolution of a simulation for a $0.6 \msun$
291: star: the disk is excited by gravitational perturbations among the embryos.
292: As eccentricities increase, orbits cross and collisions occur. In time, a few
293: planets grow and the number of bodies dwindles. Embryos may be scattered far
294: from their original locations, sometimes delivering water-rich material to
295: planets in the inner regions. Water delivery occurs relatively late, because
296: multiple scattering events are needed for significant radial movement (Raymond
297: \etal 2007). In this case some water delivery occurred: a 0.21 $\mearth$
298: planet formed at 0.41 AU, just beyond the outer boundary of the HZ, accreted
299: two water-rich embryos originating beyond the water line at 0.61 AU. However,
300: no water was delivered to any planets in the HZ in this case. This is a
301: typical outcome for the low-mass disks expected to be common around low-mass
302: stars (see below).
303:
304: Figure~\ref{fig:mall} shows the final outcome of ten simulations, with the
305: Solar System included for scale (the Earth's water content is $\sim 10^{-3} by
306: mass$; L\'ecuyer \etal 1998). It is clear that, for our assumptions of $h =
307: 1$, $\alpha = 1$ and $h =1$, terrestrial planets are much smaller around
308: low-mass stars. In addition, planets that form in the HZs of low-mass stars
309: tend to be dry, and more closely spaced. Note that, although we do not follow
310: their orbits for a full 5 Gyr, we assume water-rich planets that form in the
311: HZ to be potentially habitable. Very late-stage instabilities were not seen
312: in any of these simulations, nor in the $>$Gyr integrations from Raymond \etal
313: (2005a, 2006a), but the potential disruption of the system at times as late as
314: 5 Gyr cannot be ruled out. In addition, a late-stage instability could
315: potentially alter the orbit of a distant planet, causing it to collide with a
316: planet in the HZ. Such an event could even deliver water at a very late
317: stage. Although this is certainly possible, we have found this type of event
318: to be rare in previous long-term simulations. In addition, the source of such
319: a rogue planet would have to be quite far out, because accretion tends to
320: occur faster closer to the star, and crossing orbits are needed to cause a
321: strong scattering event.
322:
323: Figure~\ref{fig:mpl} shows the mean mass of simulated planets that formed in
324: the HZ as a function of stellar mass, with error bars representing the range
325: of outcomes from the four simulations for a given stellar mass $M_\star$. The
326: solid curve shows the prediction of a simple scenario in which the mass of a
327: planet in the HZ is proportional to the mass contained within the HZ annulus,
328: such that
329: \begin{equation}
330: M_{pl} \propto \frac{\Sigma_1 f Z
331: M_\star^h}{2-\alpha}\left(r_{HZ,out}^{2-\alpha} - r_{HZ,in}^{2-\alpha}\right),
332: \end{equation}
333: \noindent where $r_{HZ,in}$ and $r_{HZ,out}$ are the inner and outer
334: boundaries of the HZ (see Table 1). This model is a very simple, but not
335: unreasonable, approximation for planet mass, and the quality of the fit is
336: remarkable -- note that this model was calibrated such that a disk with $fZ=1$
337: will form a 1 $\mearth$ planet in the HZ around a solar-mass star. Kokubo
338: \etal (2006) showed that the planet mass scales roughly linearly with the
339: available mass, albeit for a fixed stellar mass, and we independently found
340: the same result for $M_\star = 0.4 \msun$. In fact, Kokubo \etal (2006) found
341: a slightly stronger than linear correlation, $M_{pl} \propto
342: M_{disk}^{0.97-1.1}$. Fig~\ref{fig:mall} shows that the mean interplanetary
343: spacing decreases somewhat for lower-mass stars, and that there are slightly
344: more planets for the lower-mass stars. We know that the total mass in the
345: habitable zone, $M_{HZ}$, is equal to the number of planets, $N$, times the
346: average planet mass, $M_{pl}$. Our simulations suggest that $N \propto
347: M_{HZ}^{-0.1}$ or so. Indeed, a model with $M_{pl} \propto M_{HZ}^{1.1}$
348: provides a fit that is comparable to the one in Fig.~\ref{fig:mall}. However,
349: for the remainder of the paper we assume that $M_{pl} \propto M_{HZ}$ (Eqn 2).
350: The reason for this assumption has to do with the goals of the paper. We are
351: attempting to constrain the locations in $M_\star - h - fZ$ parameter space
352: that might harbor potentially habitable planets with $M_{pl} \geq 0.3
353: \mearth$. To be conservative in our evaluations, we prefer to slightly
354: overestimate, rather than underestimate, $M_{pl}$. The dif{f}erence between
355: the two estimates is negligible for larger stellar masses, but is as much as
356: $\sim$ 40\% below 0.1 $\msun$.
357:
358: The planet mass decreases monotonically with stellar mass for all reasonable
359: parameter values (Eq. 2; Fig~\ref{fig:mpl}), with a scatter in the details of
360: a given system based on the stochastic nature of the accretion process (e.g.,
361: Wetherill 1996). Only for very steep density profiles ($\alpha > 2$) or
362: reversed disk mass scalings ($h<0$) can the planet mass increase at lower
363: stellar masses. These ef{f}ects are the result of the strong dependence of
364: the HZ's location's on stellar luminosity, and therefore on stellar mass. For
365: a given value of $f$, $Z$, $\alpha$, and $h$, there exists a stellar mass
366: limit below which the formation of a $>0.3 \mearth$ planet in the HZ is
367: unlikely. For $\alpha = 1$ and $h=1$, this limit ranges from 1 $\msun$ for
368: $fZ < 0.3$ to 0.74 $M_\odot$ for $fZ=1$ to 0.43 $M_\odot$ for $fZ=5$. These
369: limits clearly depend on the critical mass for habitability; for instance, the
370: limit is 0.53 $\msun$ for the $fZ=1$ case if the critical habitable mass is
371: 0.1 $\mearth$. Recall that $fZ$ represents a scaling of the disk mass, i.e.,
372: the disk's relative mass $f$ times the relative abundance of solids, assumed
373: to scale with the stellar metallicity $Z$.
374:
375: Figure~\ref{fig:mstar-h} shows the location in $M_\star-fZ$ space where
376: planets $>0.3 \mearth$ can form in the HZ, assuming $\alpha=1$. Each curve
377: corresponds to a given value of $h$; planets $> 0.3 \mearth$ form above and to
378: the right of each curve. More massive or metal-rich disks can form habitable
379: planets around lower-mass stars. In addition, it is easier to form $> 0.3
380: \mearth$ planets in the HZ for more centrally-condensed disks, i.e., for
381: larger values of $\alpha$ (not shown in Fig.~\ref{fig:mstar-h}). Given the
382: large amount of variation in $fZ$ and other uncertainties, we do not consider
383: these limits to be firm. However, given the large uncertainties and expected
384: variation in $f$ and other parameters, we do not consider these limits to be
385: meaningful except in a statistical sense. For an ensemble of disks, the
386: fraction of $>0.3 \mearth$ planets that form decreases significantly for
387: low-mass stars. A probabilistic version of the mass limit estimate is
388: discussed further in \S 4.
389:
390: %For $\alpha=3/2$ and $h$=1, this limit ranges from 1 $M_\odot$ for
391: %$fZ < 0.3$ to 0.66 $M_\odot$ for $fZ=1$ to 0.26 $M_\odot$ for $fZ=5$.
392:
393: \subsection{Formation Timescales and Planetary Water Contents}
394:
395: Figure~\ref{fig:tf} shows the mean formation timescales for HZ planets in our
396: simulations. Around 0.2 $\msun$ stars, terrestrial planets in the HZ form in
397: a few Myr. This increases to 20-50 Myr for Sun-like stars, consistent with
398: estimates from Hf/W isotopic measurements (e.g., Jacobsen 2005). The reason
399: for the speedup in accretion times in the HZs of low-mass stars is due to a
400: combination of the faster orbital speeds in the HZs of low-mass stars and the
401: higher surface densities (albeit much lower total HZ masses). For instance,
402: if we assume $L_\star \propto M_\star^4$ (a rough fit to Hillenbrand \& White
403: 2004), then the location of the HZ, $r_{HZ}$, scales with the stellar mass as
404: $r_{HZ} \propto L_\star^{1/2} \propto M_\star^2$. The orbital speed in the
405: HZ, $v_{HZ}$, scales as $v_{HZ} \propto (M_\star/r_{HZ})^{1/2} \propto
406: M_\star^{-1/2}$. The surface density in the HZ, $\Sigma_{HZ}$, scales as
407: $\Sigma_{HZ} \propto r_{HZ}^{-\alpha} M_\star^h \propto M_\star^{-2\alpha+h}
408: \propto M_\star^{-1}$ for our case of $\alpha = h = 1$. So, using a very
409: rough approximation that the growth time $t_G$ scales inversely with the
410: product of the orbital frequency ($\sim v_{HZ}/r_{HZ}$) and the local surface
411: density (Safronov 1969) yields $t_G \propto M_\star^{7/2}$ for our case. This
412: scaling is a very poor match to our simulations, yielding accretion timescales
413: that are far too short for stellar masses below $0.6 \msun$ (dotted line in
414: Fig.~\ref{fig:tf}). This is because the final stage of planetary growth is
415: dominated by isolated scattering events between embryos rather than accretion
416: from a continuous medium of planetesimals. In fact, we notice that a slightly
417: dif{f}erent scaling, with $t_G \propto (\Sigma_{HZ} v_{HZ})^{-1} \propto
418: M_\star^{3/2}$, provides a good empirical fit to our simulated formation times
419: (dashed line in Fig.~\ref{fig:tf}). We suspect that this is simply because of
420: the dynamics of this accretion regime, in which the relevant bodies are not on
421: initially crossing orbits, and secular perturbations are required to excite
422: eccentricities and thus to cause orbits to cross.
423:
424: Lissauer (2007) argued that the formation time of HZ planets around low-mass
425: stars is very short, a few hundred thousand years for a 1/3 $\msun$ star. Our
426: simulations confirm the trend that HZ planets form faster around low-mass
427: stars. However, Fig.~\ref{fig:tf} shows that these formation times do not
428: scale nearly as strongly with stellar mass as indicated by Lissauer (he
429: calculated $t_G \sim M_\star^{6.2}$). The reason for this discrepancy is that
430: Lissauer required an Earth-mass planet to form in the HZ, so to accomplish
431: this he increased the disk mass by a large factor. Our models suggest that a
432: factor of 20-30 increase is needed to form a 1 $\mearth$ planet in the HZ of a
433: $1/3 \msun$ star; as expected, that factor corresponds to the approximate
434: dif{f}erence in accretion times between our simulations and Lissauer's.
435: However, the fraction of 1/3 $\msun$ stars with disks massive enough to form a
436: 1 $\mearth$ planet in the HZ is less than 5\%, although that fraction
437: increases to about 15\% if $\alpha=3/2$ (see Fig.~\ref{fig:frac} below). In
438: addition, recall that the much shallower empirical scaling ($t_G \propto
439: M_\star^{3/2}$ for our case) provides a far better fit to our simulations than
440: the simple derivation using the orbital frequency ($t_G \propto M_\star^{7/2}$
441: for our case; based on Safronov 1969 and calculated in similar fashion to
442: Lissauer 2007).
443:
444: The mean water content of our HZ planets decreases drastically around low-mass
445: stars. In fact, only one out of 19 HZ planets formed from a star with
446: $M_\star < 0.6 M_\odot$ contained a significant amount of water, compared with
447: two out of six HZ planets for $M_\star = 0.8 \msun$ and two out of four for
448: $M_\star = 1 \msun$ (see Fig.~\ref{fig:mall}). The reason for this trend --
449: drier planets in the HZs of lower-mass stars -- is inef{f}icient dynamical
450: stirring in the low-mass disks found around low-mass stars, because more
451: massive embryos are needed to increase eccentricities enough for significant
452: radial mixing and therefore water delivery (Morbidelli \etal 2000; Raymond
453: \etal 2004, 2007). Lissauer (2007) argued that planets in the HZs of M stars
454: would be deficient in volatiles for three reasons: 1) collision velocities are
455: higher; 2) formation timescales are faster; and 3) pre-main sequence evolution
456: of the lowest-mass M stars is slow, such that the snow line and HZ move inward
457: on a Myr to Gyr timescale (see Kennedy \etal 2006). We agree with point 1:
458: collision velocities are proportional to the orbital speed, so the velocity of
459: impactors at infinity (neglecting the escape speed) scales with stellar mass
460: roughly as $M_\star^{-1/2}$ (see above). We also agree qualitatively with
461: point 2: formation timescales are indeed likely to be shorter around low-mass
462: stars by perhaps an order of magnitude compared with solar-mass stars
463: (Fig.~\ref{fig:tf}). Point 3 is more uncertain: the inward movement of the
464: snow line could af{f}ect the water contents of HZ planets if the formation
465: timescale is shorter than the snow line's drift timescale. In such a case,
466: material in a given region may be too hot to contain water at early times,
467: during accretion. After several Myr, that zone may drop below the threshold
468: for water to condense. If, however, the planets are fully formed by this
469: time, then water delivery is impossible. Note that this pre-main sequence
470: scenario applies only to very low-mass stars. Our results suggest that this
471: is a moot point, as HZ planets around such very low-mass stars are unlikely to
472: have wide enough feeding zones to accrete water-bearing material, even at late
473: times.
474:
475: Thus, we argue that terrestrial planets in the HZs of low-mass stars are
476: likely to be dry, but for a dif{f}erent reason than Lissauer. Simply put, very
477: little water-rich material is likely to impact such planets at all, because
478: radial mixing is inef{f}icient in low-mass disks preferentially found around
479: low-mass stars. The influence of additional gas or ice giant planets in the
480: system is not expected to increase the water contents of HZ planets, at least
481: not from the asteroidal source of water considered here (S. Raymond 2007, in
482: preparation). However, migrating giant planets (not modeled here) can stir up
483: eccentricities and induce the formation of very water-rich planets in their
484: wake (Raymond \etal 2006b; Fogg \& Nelson 2007; Mandell \etal 2007). In
485: addition, subsequent delivery of water from a cometary source is possible.
486:
487: \subsection{Ef{f}ects of Giant Planets}
488:
489: An obvious criticism of this work is the lack of giant planets. Simulations
490: have shown that giant planets play an important role in terrestrial planet
491: formation (e.g., Wetherill 1996; Levison \& Agnor 2003). Giant planets on
492: orbits exterior to the terrestrial region, such as Jupiter and Saturn, stir up
493: the eccentricities of embryos from the outside in via secular and resonant
494: perturbations, while mutual scattering of embryos excites eccentricities from
495: the inside out (e.g., Raymond \etal 2005b). Accretion is suppressed in the
496: vicinity of the giant planet, because perturbations between embryos during
497: their growth can scatter them into unstable regions such as strong giant
498: planet resonances (Wetherill 1994). Indeed, the combined ef{f}ects of embryo
499: scattering and perturbations from Jupiter and Saturn are thought to be the
500: cause of the depletion of the Solar System's asteroid belt (Wetherill 1992;
501: Chambers \& Wetherill 2001; Petit \etal 2001).
502:
503: The net ef{f}ect of giant planets is to increase the eccentricities of embryos
504: during accretion. This can cause the width of the feeding zone of terrestrial
505: planets to be increased, causing planets to be somewhat more massive and less
506: numerous than in the absence of giant planets (Levison \& Agnor 2003). This
507: increase in planet mass can be as high as $\sim 30\%$ (or planet masses can be
508: decreased), but occurs primarily for relatively weak perturbations, i.e., for
509: less massive giant planets on relatively circular orbits (S. Raymond 2007, in
510: preparation). Even on circular orbits, external giant planets rarely enhance
511: water delivery. Rather, their perturbations clear out external, water-rich
512: material leaving behind relatively dry terrestrial planets (S. Raymond 2007,
513: in preparation). Indeed, terrestrial planets in systems with giant planets on
514: eccentric orbits are likely to be dry (Chambers \& Cassen 2002; Raymond \etal
515: 2004; Raymond 2006). Note that systems with close-in giant planets may also
516: contain Earth-like planets, which should be very water-rich (Raymond \etal
517: 2006b; Mandell \etal 2007).
518:
519: To test the ef{f}ects of giant planets, we performed eight additional
520: simulations including one giant planet on an exterior orbit. We chose the 0.6
521: $\msun$ case for these simulations, to see if giant planet perturbations could
522: increase the mean terrestrial planet mass above $0.3 \mearth$. We used the
523: same four sets of embryos described in Table 1 and included a giant planet on
524: a circular orbit at 1.3 AU, corresponding roughly to the orbital distance with
525: the same temperature (and therefore at a comparable dynamical separation from
526: the HZ) as Jupiter for a 0.6 $\msun$ star. We ran two sets of simulations,
527: one each for a Neptune- and a Jupiter-mass giant planet.
528:
529: Planets that formed in the HZ of our 0.6 $\msun$ simulations without giant
530: planets had masses between 0.06 and 0.18 $\mearth$, with a mean of 0.10
531: $\mearth$. Systems with a Jupiter- [or Neptune-] mass giant planet formed,
532: respectively, HZ planets with masses between 0.05 [0.06] and 0.22 [0.10]
533: $\mearth$, with a mean of 0.11 [0.09] $\mearth$, very close to the no giant
534: planets cases. None of the seven HZ planets in systems with no giant planets
535: contained any water-rich material. Two out of eleven HZ planets that formed
536: in the eight giant planet systems contained a substantial amount of water.
537: While this is only a small fraction of outcomes, it does show that giant
538: planet stirring can, in some cases, help in water delivery. However, in the
539: vast majority of situations, giant planets either hinder or have no effect on
540: water delivery (S. Raymond 2007, in preparation). The formation timescales of
541: HZ planets in giant planet simulations were comparable to, or even slightly
542: longer than, those with no giant planets. In almost all cases, planets with
543: or without giant planets reached 75\% of their final masses within 10-20 Myr
544: (see Fig.~\ref{fig:tf}). We suspect that the reason giant planets do not
545: accelerate accretion in this case is because more radial mixing is happening,
546: such that accretion is somewhat less confined to a given annular zone. This
547: ef{f}ect appears to be small, as less than 20\% of planets in the giant planet
548: simulations contain water.
549:
550: \section{Discussion and Conclusions}
551:
552: Our analysis is based on three key assumptions: 1) terrestrial planets
553: below a given mass are unlikely to sustain life on Gyr timescales -- following
554: Williams \etal (1998), we derive a limiting planet mass of roughly $0.3
555: \mearth$; 2) a protoplanetary disk can be well described by just a few
556: parameters (see Eq. 1): the radial surface density slope $\alpha$, the
557: relative disk mass in solids $fZ$, and the exponent characterizing the disk
558: mass-stellar mass relationship, $h$; and 3) the typical planet mass in the
559: habitable zone is proportional to the disk mass in that zone (see Eq. 2 and
560: Fig.~\ref{fig:mpl}). We performed a suite of twenty dynamical simulations of
561: the late-stage growth of terrestrial planets that confirmed that this scaling
562: does hold for a wide range of stellar masses. We assumed in our calculations
563: that no giant planets were present. We tested that assumption for the 0.6
564: $\msun$ case and found that the ef{f}ects of exterior giant planets do not
565: increase the masses or decrease the formation times of terrestrial planets, at
566: least for the cases considered here.
567:
568: Starting from these assumptions, we explored the range of parameters that
569: allow planets more massive than $0.3 \mearth$ to form in the habitable zones
570: (HZs) of their host stars. We found that any realistic combination of
571: parameters led to a stellar mass - planet mass correlation such that HZ planet
572: masses decrease with decreasing stellar mass (see Fig.~\ref{fig:mpl}). For a
573: given set of parameters ($\alpha$, $h$,$fZ$), one can derive a stellar mass
574: below which the probability of a habitable planet forming in a disk with those
575: parameters decreases to zero (see Fig.~\ref{fig:mstar-h}). For instance, for
576: our fiducial parameters ($h=1$, $\alpha=1$,$f=1.2$, chosen so that the
577: simulations produce a mean planet mass of 1 $\mearth$ for a 1 $\msun$ star),
578: this critical stellar mass is $0.7 \msun$. This would mean that no M stars
579: and few K stars should have sustainably habitable planets. However, there are
580: some caveats to this approach. First, for main sequence stars, it is
581: impossible to go back in time to learn the properties of their protoplanetary
582: disks, most importantly $f$ and $\alpha$. Second, for a given set of
583: parameters, there still exists a range of potential outcomes because of the
584: stochastic nature of the final stages of planetary growth (e.g., Wetherill
585: 1996, Raymond \etal 2004; see Fig.~\ref{fig:mall}). Third and most
586: importantly, there is no universal set of parameters that applies for all
587: disks. A given set of parameter values, corresponding to a specific disk or
588: set of disks, has a corresponding planet mass distribution. For instance, the
589: above-mentioned 0.7 $\msun$ limit decreases to $0.31 \msun$ for a more massive
590: disk with $fZ = 10$, and would decrease farther still for $h<1$ or $\alpha
591: >1$. Thus, our approach is only viable in a statistical sense, by combining
592: likely outcomes from a distribution of disks with varying properties.
593:
594: What are the values of the crucial parameters? We expect $h$ and $\alpha$ to
595: be universal values characterizing the distributions of all disks, but $f$ and
596: $Z$ to vary from disk to disk. As discussed above, there is considerable
597: uncertainty in the value of $h$, which controls the disk mass-stellar mass
598: scaling. If $h$ were larger than 2 or smaller than 0.5, we expect that the
599: trend would be apparent in Fig. 3 in Scholz \etal (2006). Thus, a roughly
600: linear disk mass-stellar mass relationship ($h=1$) seems likely. A range of
601: roughly 2 orders of magnitude in disk mass is observed around a given stellar
602: mass (e.g., Eisner \& Carpenter 2003; Williams \etal 2005; Scholz \etal 2006),
603: implying a corresponding range in $f$ values. Note, however, that there
604: exists a large uncertainty in tying a given $f$ value to an absolute disk mass
605: or surface density, because of both observational and theoretical
606: uncertainties (e.g., Carpenter \etal 2005). There exists a range of 3-5 in
607: metallicity $Z$ for thin-disk stars (Nordstrom \etal 2004), although the range
608: for currently-forming stars may be much smaller, only $\sim$20\% (see
609: Cartledge et al. 2006). Nearly all disk models and observations suggest that
610: the surface density exponent $\alpha = 0.5-1.5$ (e.g., Dullemond \etal 2007;
611: Andrews \& Williams 2006). Our choice of $\alpha =1$ as a fiducial case was a
612: compromise between these varying estimates.
613:
614: Given the large amount of intrinsic variation in the disk mass and accretion
615: outcome, as well as the uncertainty in $h$ and $\alpha$, we do not consider
616: any stellar mass limits from Fig.~\ref{fig:mstar-h} to be firm. Rather, our
617: primary result is a statistical observation that, in our framework, the
618: fraction of disks capable of forming planets of a given mass in the habitable
619: zone decreases around low-mass stars. We investigate a model with $fZ$
620: distributed as a Gaussian in log space, with a mean of $fZ=1$ and a standard
621: deviation of 0.5 dex. For simplicity, all other parameters are held fixed,
622: with $\alpha=1$, and $h$=1. Figure~\ref{fig:frac} shows that the fraction of
623: systems that can form a $>0.3 \mearth$ HZ planet decreases from 80\% around
624: solar-type stars to 42\% around 0.7 $M_\odot$ stars to 6\% around 0.4
625: $M_\odot$ stars to only 0.3\% around 0.2 $M_\odot$ stars. Disks with more
626: centrally concentrated mass distributions (larger $\alpha$ values) have less
627: dif{f}iculty forming massive planets in HZs very close to low-mass stars (e.g.,
628: Raymond \etal 2005b). For our chosen fiducial parameters, we estimate that
629: only 25\% of systems with $M_\star = 0.6 \msun$ can form planets $> 0.3
630: \mearth$, and only about 5\% for $M_\star =0.4 \msun$. Disks with radial
631: surface density profiles as flat as $r^{-1/2}$ have near zero probability for
632: stars with $M_\star < 0.5 \msun$, while probabilities are 2-4 times larger for
633: disks with $r^{-3/2}$ profiles. Reducing the required planetary mass for
634: habitability to $0.1 \msun$ increases the probability by a factor of 2-4 (see
635: below).
636:
637: Our estimate of the minimum planet mass that can sustain enough tectonic
638: activity for a period of time that we took as 5 Gyr is also very uncertain.
639: It is simple to show that if the mass luminosity relation is of the form
640: $L_\star \propto M_\star^p$, then the derived critical stellar mass
641: $M_{\star,cr}$ below which there should be few planets with masses above a
642: critical mass $M_{pl,cr}$ is given by $M_{\star,cr} \sim (M_{pl,cr}/f)^B$,
643: with $B=2/[p(2-\alpha) + 2h]$. Thus, a reduction in the adopted critical mass
644: for sustained habitability from 0.3 to 0.2 $\mearth$ reduces the required
645: value of disk mass normalization by the same factor, in order to obtain the
646: same critical stellar mass. Alternatively, for a given chosen disk mass
647: normalization, the derived $M_{\star,cr}$ varies with the assumed planet mass
648: to a power that is between about 0.3 to 0.7 for most of the range of
649: parameters described above. The inset of Fig.~\ref{fig:frac} shows the ef{f}ect
650: of changing our $0.3 \mearth$ lower limit, retaining fixed values of $h=1$ and
651: $\alpha = 1$.
652:
653: A system of three planets with minimum masses between 5 and 15 $\mearth$ was
654: recently discovered orbiting the 0.31 $\msun$ star Gliese 581 (Bonfils \etal
655: 2005; Udry \etal 2007). Figures~\ref{fig:mstar-h} and~\ref{fig:frac} suggest
656: that such a low-mass star is unlikely to form massive terrestrial planets.
657: How can we reconcile this? We see two possible explanations: 1) the planets
658: of Gl 581 formed {\it in situ} from a very massive disk, with $f Z \sim
659: 30-50$; or 2) our assumption of {\it in situ} formation failed in this case:
660: the planets of Gl 581 formed farther from the star (probably in the icy
661: regions of a relatively massive disk, with $f Z$ of at least 5-10) and
662: migrated inward (Goldreich \& Tremaine 1980). If any of the Gl 581 planets
663: were to transit the parent star, then it would be possible to derive a rough
664: composition (rocky vs. icy: Valencia \etal 2007; Fortney \etal 2007; Sotin
665: \etal 2007), and to distinguish between these two formation scenarios (Gaidos
666: \etal 2007; S. Raymond \etal 2007, in preparation). Note that, even if the Gl
667: 581 planets formed {\it in situ}, we expect that they would contain a
668: substantial amount of water. Following the arguments made above, massive
669: disks promote eccentricity growth and radial mixing such that planets in the
670: HZ should accrete a large amount of water-rich material.
671:
672: Even if HZ planets of suf{f}icient mass can form around M stars, they face
673: other challenges, including loss of atmosphere by intense stellar activity
674: coupled with their small distance to their parent stars, depending on the
675: strength of the planetary magnetic field (Lammer \etal 2007), a snow line that
676: takes perhaps a Gyr to get close enough to the HZ to give suf{f}icient water
677: (Kennedy \etal 2006), or loss of volatiles due to larger impact speeds and
678: faster formation times (Lissauer 2007; but see \S 3.2 and
679: Fig.~\ref{fig:tf}). However these latter two points should not af{f}ect
680: decisions concerning M star planet searches, since they only apply to the very
681: lowest-mass stars, $\sim 0.1-0.2 \msun$, whose apparent brightness and HZ
682: angular separation are far too small for any planned search (Scalo \etal
683: 2007). Unless the masses of disks are larger than is currently thought by a
684: significant factor and/or the critical planet mass for habitability is
685: considerably smaller, only a small fraction of accessible M star systems (with
686: $M_\star \gtrsim 0.3-0.4 \msun$) should have habitable planets that remain
687: habitable for billions of years.
688:
689: \section{Acknowledgments}
690:
691: This paper benefited from the referee's thoughtful comments, and from
692: discussions with Jim Kasting. This work was performed by the NASA
693: Astrobiology Institute's Virtual Planetary Laboratory Lead Team, supported via
694: NASA CAN-00-OSS-01. J.M.S. was supported by NASA Exobiology grant
695: NNG04GK43G. S.N.R. was supported by an appointment to the NASA Postdoctoral
696: Program at the University of Colorado's Center for Astrobiology, administered
697: by Oak Ridge Associated Universities through a contract with NASA.
698: Simulations were performed at Weber State University and JPL using CONDOR
699: (www.cs.wisc.edu/condor).
700:
701:
702: \newpage
703: \begin{thebibliography}{999}
704:
705:
706: \bibitem[]{394} Abe, Y., Ohtani, E., Okuchi, T., Righter, K., and Drake,
707: M., 2000. In: Righter, K. \& Canup, R. (Eds.), Origin of the Earth and the
708: Moon. University of Arizona Press, Tucson, 413-433.
709:
710: \bibitem[]{} Adams, F.C., Hollenbach, D., Laughlin, G., and Gorti, U. 2004. \apj, 611, 360.
711:
712: \bibitem[]{398} Alexander, R., \& Armitage, P. 2006, \apjl, 639, L83
713:
714: \bibitem[]{401} Andre, P. and Montmerle, T. 1994, ApJ, 420, 837.
715:
716: \bibitem[]{403} Andrews, S. M. and Williams, J. P. 2005, ApJ, 631, 1134.
717:
718: \bibitem[]{} Andrews, S. M. and Williams, J. P. 2006, ApJ, submitted (astro-ph/0610813).
719:
720: \bibitem[]{} Beckwith, S.~V.~W., Sargent, A.~I., Chini, R.~S., \& Guesten, R.\ 1990, \aj,
721: 99, 924
722:
723: \bibitem[]{405} Beer, M. E., King, A. R., Livio, M., Pringle, J. E. 2004 MNRAS, 354, 763
724:
725: \bibitem[]{} Beichman, C.~A., et al.\ 2006, \apj, 652, 1674
726:
727: \bibitem{} Bonfils, X., Forveille, T., Delfosse, X., Udry, S., Mayor, M.,
728: Perrier, C., Bouchy, F., Pepe, F., Queloz, D., Bertaux, J.-L.\ 2005. \aap,
729: 443, L15-L18.
730:
731: \bibitem[]{} Butler, R.~P., Johnson, J.~A., Marcy, G.~W., Wright, J.~T., Vogt, S.~S., \& Fischer, D.~A.\ 2006, \pasp, 118, 1685
732:
733: \bibitem[]{} Carpenter, J.~M., Wolf, S., Schreyer, K., Launhardt, R., \& Henning, T.\ 2005, \aj, 129, 1049
734:
735: \bibitem[]{} Cartledge, S. I. B., Lauroesch, J. T., Meyer, D. M., and Sofia, U. J. 2006 ApJ, 641, 327.
736:
737: %\bibitem[]{407} Chambers, J.~E., Wetherill, G.~W., \& Boss, A.~P.\ 1996, Icarus, 119, 261
738:
739: \bibitem[]{409} Chambers, J. E., 1999. MNRAS, 304, 793.
740:
741: \bibitem[Chambers \& Cassen(2002)]{2002M&PS...37.1523C} Chambers, J.~E., \&
742: Cassen, P.\ 2002, Meteoritics and Planetary Science, 37, 1523
743:
744: \bibitem[Chambers \& Wetherill(2001)]{2001M&PS...36..381C} Chambers, J.~E.,
745: \& Wetherill, G.~W.\ 2001, Meteoritics and Planetary Science, 36, 381
746:
747: \bibitem[]{411} Chambers, J.~E.\ 2004, Earth and Planetary Science Letters, 223, 241
748:
749: \bibitem[]{} Clarke, C.~J., \& Pringle, J.~E.\ 2006, \mnras, 370, L10
750:
751: \bibitem[]{} Cyr, K.~E., Sears, W.~D., \& Lunine, J.~I.\ 1998, Icarus, 135, 537
752:
753: \bibitem[]{}D'Alessio, P., Canto, J., Calvet, N., \& Lizano, S.\ 1998, \apj, 500, 411
754:
755: \bibitem[Davis(2005)]{2005ApJ...627L.153D} Davis, S.~S.\ 2005, \apjl, 627,
756: L153
757:
758: \bibitem[Dullemond et al.(2006)]{2006ApJ...645L..69D} Dullemond, C.~P.,
759: Natta, A., \& Testi, L.\ 2006, \apjl, 645, L69
760:
761: \bibitem[]{}Dullemond, C.~P., Hollenbach, D., Kamp, I., \& D'Alessio, P.\ 2007, Protostars and Planets V,
762: 555
763:
764: \bibitem[]{412} Eisner, J. A. and Carpenter, J. M. 2003, ApJ, 598, 1341.
765:
766: \bibitem[]{414} Eisner, J. A. and Carpenter, J. M. 2006, \apj, 641, 1162.
767:
768: \bibitem[]{417} Endl, M., Cochran, W.~D.,
769: K{\"u}rster, M., Paulson, D.~B., Wittenmyer, R.~A., MacQueen, P.~J., \&
770: Tull, R.~G.\ 2006, \apj, 649, 436
771:
772: \bibitem[]{421} Fischer, D. A. and Valenti, J. 2005. ApJ 622, 1102
773:
774: \bibitem[]{} Fogg, M.~J., \& Nelson, R.~P.\ 2007, \aap, 461, 1195
775:
776: \bibitem{} Fortney, J.~J., Marley, M.~S., Barnes, J.~W.\ 2007, \apj, 659, 1661.
777:
778: \bibitem[]{} Gaidos, E., Haghighipour, N., Agol, E., Latham, D., Raymond,
779: S.~N., Rayner, J. 2007, Science, submitted.
780:
781:
782: \bibitem[]{} Garaud, P., \& Lin, D.~N.~C.\ 2007, \apj, 654, 606
783:
784: \bibitem[Gonzalez(1997)]{1997MNRAS.285..403G} Gonzalez, G.\ 1997, \mnras,
785: 285, 403
786:
787: \bibitem[]{423} Greaves, J.~S., Fischer, D.~A., \& Wyatt, M.~C.\ 2006, \mnras,
788: 366, 283
789:
790: \bibitem[]{} Gregory, S. G., Jardine, M., Simpson, I., and Donati, J.-F. 2007 MNRAS 371, 999.
791:
792: \bibitem[Goldreich \& Tremaine(1980)]{1980ApJ...241..425G} Goldreich, P.,
793: \& Tremaine, S.\ 1980, \apj, 241, 425
794:
795: \bibitem[Hartmann et al.(2006)]{2006astro.ph..5294H} Hartmann, L.,
796: D'Alessio, P., Calvet, N., \& Muzerolle, J.\ 2006, \apj, 648, 484
797:
798: \bibitem[]{} Hayashi, C.\ 1981, Progress of Theoretical Physics Supplement, 70, 35
799:
800: \bibitem[]{425} Hillenbrand, L.~A., \& White, R.~J.\ 2004, \apj, 604, 741
801:
802: \bibitem[]{427} Ida, S., \& Lin, D.~N.~C.\ 2005, \apj, 626, 1045
803:
804: \bibitem[]{429} Kasting, J. F., Whitmire, D. P., and Reynolds, R. T.,
805: 1993. Icarus 101, 108.
806:
807: \bibitem[]{}Kennedy, G.~M., Kenyon, S.~J., \& Bromley, B.~C.\ 2006, \apjl, 650, L139
808:
809: \bibitem[]{432} Kokubo, E., \& Ida, S.\ 1998, Icarus, 131, 171
810:
811: \bibitem[]{434} Kokubo, E., Kominami, J., \& Ida, S., 2006. \apj, 642, 1131
812:
813: \bibitem[]{} Kornet, K., Wolf, S. and Rozyczka, M. 2006. \aap 458, 661
814:
815: \bibitem[]{436} Laughlin, G., Bodenheimer, P., \& Adams, F.~C.\ 2004, \apjl, 612, L73
816:
817: \bibitem[]{} L\'ecuyer, C., Gillet, P., \& Robert, F. 1998, Chemical Geology
818: 145, 249.
819:
820: \bibitem[]{439} L{\'e}ger, A., \etal\ 2004, Icarus, 169, 499
821:
822: \bibitem[]{} Levison, H.~F., \& Duncan, M.~J.\ 2000, \aj, 120, 2117
823:
824: \bibitem[Levison \& Agnor(2003)]{2003AJ....125.2692L} Levison, H.~F., \&
825: Agnor, C.\ 2003, \aj, 125, 2692
826:
827: \bibitem[]{441} Lineweaver, C. H. and Grether, D. 2003, ApJ 598, 1350
828:
829: \bibitem[]{443} Lissauer, J. J. 1999 Nature 402, C11.
830:
831: \bibitem[Lissauer(2007)]{2007ApJ...660L.149L} Lissauer, J.~J.\ 2007, \apjl,
832: 660, L149
833:
834: \bibitem[]{} Looney, L. W., Mundy, L. G., and Welch, W. J. 2003 ApJ 5592, L255.
835:
836: \bibitem[Mandell et al.(2007)]{2007ApJ...660..823M} Mandell, A.~M.,
837: Raymond, S.~N., \& Sigurdsson, S.\ 2007, \apj, 660, 823
838:
839: \bibitem[]{445} Morbidelli, A., Chambers, J., Lunine, J. I., Petit, J. M.,
840: Robert, F., Valsecchi, G. B., and Cyr, K. E., 2000. Meteoritics and
841: Planetary Science 35, 1309.
842:
843: \bibitem[]{}Moro-Mart{\'{\i}}n, A., et al.\ 2007, \apj, 658, 1312
844:
845: \bibitem[]{} Mundy, L. G., Looney, L. W., and Welch, W. J. 2000, in Protostars and Planets IV, ed. V. Manning, A. P. Boss, S. S. Russell, p. 355.
846:
847: \bibitem[]{449} Muzerolle, J., Hillenbrand, L., Calvet, N., Brice{\~n}o, C., \& Hartmann, L.\ 2003, \apj, 592, 266
848:
849: \bibitem[]{452} Muzerollle, J., Luhman, K. L., Briceno, C., Hartmann, L., and Calvet, N. 2005, ApJ, 625, 906.
850:
851: \bibitem[]{} Natta, A., Testi, L., \& Randich, S.\ 2006, \aap, 452, 245
852:
853: \bibitem[]{} Nordstrom, B. \etal 2004, \aap, 418, 989.
854:
855: \bibitem[]{454} Padoan, P., Kritsuk, A., Norman, M. L., and Nordlund, A. 2005, ApJL, 622, L61.
856:
857: \bibitem[Petit et al.(2001)]{2001Icar..153..338P} Petit, J.-M., Morbidelli,
858: A., \& Chambers, J.\ 2001, Icarus, 153, 338
859:
860: \bibitem[]{456} Quinn, T.~R., Tremaine, S., \& Duncan, M., 1991. AJ 101, 2287.
861:
862: \bibitem[Rauch \& Holman(1999)]{1999AJ....117.1087R} Rauch, K.~P., \& Holman, M.\ 1999, \aj, 117, 1087
863:
864: \bibitem[]{458} Raymond, S.~N., Quinn, T., \& Lunine, J. I., 2004. Icarus, 168, 1.
865:
866: \bibitem[]{461}Raymond, S.~N., Quinn,~T., \& Lunine,~J.~I. 2005a. Icarus, 177, 256.
867:
868: \bibitem[]{464}Raymond, S.~N., Quinn,~T., \& Lunine,~J.~I. 2005b. \apj, 632, 670.
869:
870: \bibitem[]{466}Raymond, S.~N., Quinn,~T., \& Lunine,~J.~I. 2006a. Icarus, 183, 265.
871:
872: \bibitem[]{}Raymond, S.~N., Mandell, A.~M., \& Sigurdsson, S. 2006b. Science, 313, 1413.
873:
874: \bibitem[]{469}Raymond, S.~N., Quinn,~T., \& Lunine,~J.~I. 2007. Astrobiology,
875: 7, 66.
876:
877: \bibitem[]{} Raymond, S.~N. 2006, ApJ, 643, L131.
878:
879: \bibitem[]{474} Reid, N., \& Hawley, S.~L.\ 2000, New light on dark stars :
880: red dwarfs, low mass stars, brown dwarfs / Neill Reid and Suzanne
881: L.~Hawley.~New York : Springer, 2000.~(Springer-Praxis series in
882: astronomy and astrophysics),
883:
884: \bibitem[]{} Safronov, V.~S. 1969, Evolution of the Protoplanetary Could and
885: Formation of the Earth and Planets (Moscow, Nauka Press), English translation:
886: NASA TTF-677, 1972.
887:
888: \bibitem[Santos et al.(2001)]{2001A&A...373.1019S} Santos, N.~C.,
889: Israelian, G., \& Mayor, M.\ 2001, \aap, 373, 1019
890:
891: \bibitem[]{}Scalo, J., et al. 2007, Astrobiology, 7, 85.
892:
893: \bibitem[Scholz et al.(2006)]{} Scholz, A., Jayawardhana, R., \& Wood, K.\ 2006, \apj, 645, 1498
894:
895: \bibitem{} Sotin, C., Grasset, O. \& Mocquet, A. 2007, Icarus, in press.
896:
897: \bibitem[]{479} Tabachnik, S. and Tremaine, S. 2002, MNRAS 335, 151.
898:
899: \bibitem{2007arXiv0704.3841U} Udry, S., and 10 colleagues 2007. \aap, submitted.
900:
901: \bibitem{} Valencia, D., Sasselov, D.~D., O'Connell, R.~J.\ 2007. \apj, in
902: press, astro-ph/07043454.
903:
904: %\bibitem[]{481} Vicente, S.~M., \& Alves, J.\ 2005, \aap, 441, 195
905:
906: \bibitem[]{483} Walker, J.C.G., Hays, P. B., and Kasting, J. F. 1981, J. Geophys. Res. 86, 9776.
907:
908: \bibitem[]{} Ward, W.~R., \& Hahn, J.~M.\ 1995, \apjl, 440, L25
909:
910: \bibitem[]{485} Weidenschilling, S.~J.\ 1977. Ap\&SS, 51, 153
911:
912: \bibitem[Wetherill(1992)]{1992Icar..100..307W} Wetherill, G.~W.\ 1992, Icarus, 100, 307
913:
914: \bibitem[]{487} Wetherill, G. W., 1996. Icarus, 119, 219.
915:
916: \bibitem[]{489} Williams, D.~M., Kasting, J.~F., \& Wade, R.~A.\ 1997, Nature,
917: 385, 234.
918:
919: \bibitem[]{492} Williams, J. P., Andrews, S. M., and Wilner, D. J. 2005, ApJ, 634, 495.
920:
921: \end{thebibliography}
922:
923:
924: %\newpage
925:
926: \scriptsize
927: \begin{deluxetable}{ccccccc}
928: %\begin{deluxetable}{p{2cm} p{2cm} p{2cm} p{2cm} p{2cm} p{2cm} p{2cm}}
929: \tablewidth{0pt}
930: \tablecaption{Initial Conditions for Simulations\tablenotemark{1}}
931: \renewcommand{\arraystretch}{.6}
932: \tablehead{
933: \colhead{$M_\star$} &
934: \colhead{Range (AU)} &
935: \colhead{Mass ($\mearth$)} &
936: \colhead{N} &
937: \colhead{Timestep (d)} &
938: \colhead{HZ (AU)} &
939: \colhead{Water line (AU)}}
940: \startdata
941: 1.0 & 0.5-4 & 4.95 & 75 & 6 & 0.8-1.5 & 2.5 \\
942: 0.8 & 0.25-2 & 1.98 & 75 & 2.5 & 0.39-0.74 & 1.23 \\
943: 0.6 & 0.12-1 & 0.75 & 100 & 1.0 & 0.20-0.37 & 0.61 \\
944: 0.4 & 0.06-1 & 0.53 & 130 & 0.4 & 0.10-0.19 & 0.32 \\
945: 0.2 & 0.03-0.5 & 0.13 & 190 & 0.2 & 0.05-0.1 & 0.16 \\
946: \enddata
947: \tablenotetext{1}{Columns are labeled as follows: `$M_\star$' is the stellar
948: mass in Solar masses; `Range' is the initial radial distribution of embryos;
949: `Mass' represents the total mass in embryos; `N' is the number of embryos;
950: `Timestep' is the timestep used for the integration in days; `HZ' is the
951: extent of the habitable zone for each case; and `Water line' indicates the
952: boundary beyond which embryos are assumed to contain 5\% water by mass.}
953: \end{deluxetable}
954: \normalsize
955:
956: %\scriptsize
957: %\begin{deluxetable}{cccccccc}
958: %\tablewidth{0pt}
959: %\tablecaption{Initial Conditions for Simulations}
960: %\renewcommand{\arraystretch}{.6}
961: %\tablehead{
962: %\colhead{$M_\star$} &
963: %\colhead{embryo range (AU)} &
964: %\colhead{Mass ($\mearth$)} &
965: %\colhead{Embryo separation ($\Delta$)} &
966: %\colhead{N} &
967: %\colhead{Timestep (days)} &
968: %\colhead{HZ (AU)} &
969: %\colhead{Water line (AU)}}
970: %\startdata
971: %1.0 & 0.5-4 & 4.95 & 4-8 & 75 & 6 & 0.8-1.5 & 2.5 \\
972: %0.8 & 0.25-2 & 1.98 & 5-10 & 75 & 2.5 & 0.39-0.74 & 1.23 \\
973: %0.6 & 0.12-1 & 0.75 & 5-10 & 100 & 1.0 & 0.20-0.37 & 0.61 \\
974: %0.4 & 0.06-1 & 0.53 & 7-10 & 130 & 0.4 & 0.10-0.19 & 0.32 \\
975: %0.2 & 0.03-0.5 & 0.13 & 6-10 & 190 & 0.2 & 0.05-0.1 & 0.16 \\
976: %\enddata
977: %\end{deluxetable}
978: %\normalsize
979:
980: \begin{figure}
981: \centerline{\epsscale{0.7}\plotone{f1.eps}}
982: \caption{Evolution of a simulation around a 0.6 $\msun$ star. Particle size
983: scales with the planetary size as the planet mass$^{1/3}$, but is not to scale
984: on the x axis. Colors correspond to water contents, from red (dry) to blue
985: (5\% water by mass -- see color bar). For scale, the planet that formed at
986: 0.29 AU in the HZ is 0.06 $\mearth$ and the planet at 0.42 AU is 0.21
987: $\mearth$. The HZ is shaded in the final panel of the simulation. Note that,
988: although the HZ planet is dry, some water delivery did occur, as two
989: water-rich embryos were accreted by the planet at 0.41 AU.}
990: \label{fig:aet}
991: \end{figure}
992:
993: \begin{figure}
994: \centerline{\epsscale{0.7}\plotone{f2.eps}}
995: \caption{Final outcomes of ten simulations, two for each stellar mass chosen,
996: and with the Solar System for scale. As in Fig~\ref{fig:aet}, color represents
997: water content, and particle size scales with the planet size, i.e., the planet
998: mass$^{1/3}$. The lines under each planet represent the radial excursion of a
999: planet over its orbit, i.e. its orbital eccentricity -- these values are
1000: averaged over the last 10 Myr of each simulation. Solar System water contents
1001: are from Lodders \& Fegley (1998) and L\'ecuyer \etal (1998), and
1002: eccentricities are 3 Myr averages from Quinn \etal (1991). The HZ is shaded
1003: for each stellar mass. }
1004: \label{fig:mall}
1005: \end{figure}
1006:
1007: \begin{figure}
1008: \centerline{\epsscale{0.9}\plotone{f3.eps}}
1009: \caption{Mass of planets formed in the HZ as a function of stellar
1010: mass, for a model with $h=1$, $\alpha=1$, and $fZ = 1.2$ (so that the mean
1011: planet mass for a 1 $\msun$ star is 1 $\mearth$). Error bars represent the
1012: range of values for HZ planets. The solid curve represents a model in which
1013: the HZ planet mass scales linearly with the total annular mass in the HZ. The
1014: shaded region represents reasonable estimates of the limiting planet mass for
1015: habitability (0.1-0.5 $\mearth$); our chosen value of 0.3 $\mearth$ is
1016: indicated with the dashed line.}
1017: \label{fig:mpl}
1018: \end{figure}
1019:
1020: \begin{figure}
1021: \centerline{\epsscale{0.9}\plotone{f4.eps}}
1022: %\centerline{\epsscale{1.0}\plotone{Mst-Npl_emb2.eps}}
1023: \caption{Regions of $M_\star-fZ$ space in which habitable planets more
1024: massive that 0.3 $\mearth$ can form, assuming $\alpha=1$ and for three
1025: dif{f}erent values of $h$. Planets larger than 0.3 $\mearth$ can form above
1026: and to the right of each curve. }
1027: \label{fig:mstar-h}
1028: \end{figure}
1029:
1030: \begin{figure}
1031: \centerline{\epsscale{0.9}\plotone{f5.eps}}
1032: \caption{Formation times of HZ planets in our simulations. Dif{f}erent symbols
1033: correspond to the time for a planet to reach a fraction (50\%,75\%,90\%) of
1034: its final mass. Shaded are estimates for the formation time of the Earth,
1035: derived from Hf/W isotopic measurements (e.g., Jacobsen 2005). The dotted
1036: line corresponds to a simple estimate from Safronov (1969), assuming the
1037: formation time scales inversely with the product of the orbital frequency and
1038: the local surface density. The dashed line represents a dif{f}erent, simple
1039: model in which the formation time scales inversely as the product of the
1040: orbital {\it velocity} and the local surface density. Both estimates are
1041: referenced to 50 Myr for 1 $\msun$.}
1042: \label{fig:tf}
1043: \end{figure}
1044:
1045: \begin{figure}
1046: \centerline{\epsscale{1.0}\plotone{f6.eps}}
1047: \caption{Fraction of stars able to form a $>0.3\mearth$ HZ planet as a
1048: function of stellar mass. We assume an $fZ$ distribution that is gaussian
1049: with a standard deviation of 0.5 dex, and a fixed value of $h=1$. The inset
1050: shows the ef{f}ect of varying the limiting mass for habitability from 0.1 to
1051: 0.5 $\mearth$, for $\alpha=h=1$.}
1052: \label{fig:frac}
1053: \end{figure}
1054:
1055:
1056:
1057: \end{document}
1058: