1: \documentclass[12pt,preprint]{aastex}

2: %\documentclass[12pt,manuscript,doublespace]{aastex}

3: \doublespace

4: \slugcomment{73 pages, including 12 figures and 7 tables}

5: \shorttitle{solid surface density of solar nebula}

6: \shortauthors{Dodson-Robinson et al.}

7: \begin{document}

8: 

9: \title{Ice Lines, Planetesimal Composition and Solid Surface Density in

10: the Solar Nebula}

11: 

12: \author{Sarah E. Dodson-Robinson}

13: \affil{NASA Exoplanet Science Institute, California Institute of

14: Technology \\

15: MC 100-22 Pasadena, CA 91125}

16: \email{sdr@ipac.caltech.edu}

17: 

18: \author{Karen Willacy}

19: \affil{Jet Propulsion Laboratory,  California Institute of Technology \\

20: 4800 Oak Grove Dr, Pasadena, CA 91109}

21: 

22: \author{Peter Bodenheimer}

23: \affil{ Department of Astronomy and Astrophysics,  University of

24: California at Santa Cruz \\

25: 1156 High St, Santa Cruz, CA 95064}

26: 

27: \author{Neal J. Turner}

28: \affil{Jet Propulsion Laboratory, California Institute of Technology \\

29: 4800 Oak Grove Dr, Pasadena, CA 91109}

30: 

31: \author{Charles A. Beichman}

32: 

33: \affil{NASA Exoplanet Science Institute/Jet Propulsion Laboratory \\

34: California Institute of Technology \\

35: MC 100-22, Pasadena, CA 91125}

36: 

37: %Proposed running head: Solid Surface Density in the Solar Nebula

38: 

39: %\bigskip

40: 

41: %Please direct correspondence to:

42: 

43: %\begin{center}

44: %\begin{tabular}{l}

45: %Sarah Dodson-Robinson \\

46: %NASA Exoplanet Science Institute \\

47: %California Institute of Technology \\

48: %Mail Code 100-22 \\

49: %Pasadena, CA 91125 \\

50: %sdr@ipac.caltech.edu \\

51: %\end{tabular}

52: %\end{center}

53: 

54: %\clearpage

55: 

56: %\begin{center}

57: %{\bf \large ABSTRACT}

58: %\end{center}

59: 

60: \begin{abstract}

61: To date, there is no core accretion simulation that can successfully

62: account for the formation of Uranus or Neptune within the observed

63: 2-3~Myr lifetimes of protoplanetary disks.  Since solid accretion rate

64: is directly proportional to the available planetesimal surface density,

65: one way to speed up planet formation is to take a full accounting of all

66: the planetesimal-forming solids present in the solar nebula.  By

67: combining a viscously evolving protostellar disk with a kinetic model of

68: ice formation, which includes not just water but methane, ammonia, CO

69: and 54 minor ices, we calculate the solid surface density of a possible

70: giant planet-forming solar nebula as a function of heliocentric distance

71: and time.  Our results can be used to provide the starting planetesimal

72: surface density and evolving solar nebula conditions for core accretion

73: simulations, or to predict the composition of planetesimals as a

74: function of radius.

75: 

76: We find three effects that favor giant planet formation by the core

77: accretion mechanism: (1) a decretion flow that brings mass from the

78: inner solar nebula to the giant planet-forming region, (2) the fact that

79: the ammonia and water ice lines should coincide, according to recent lab

80: results from \cite{collings04}, and (3) the presence of a substantial

81: amount of methane ice in the trans-Saturnian region.  Our results show

82: higher solid surface densities than assumed in the core accretion models

83: of Pollack et al. (1996) by a factor of 3--4 throughout the

84: trans-Saturnian region.  We also discuss the location of ice lines and

85: their movement through the solar nebula, and provide new constraints on

86: the possible initial disk configurations from gravitational stability

87: arguments.

88: \end{abstract}

89: 

90: {\bf Keywords:} solar nebula; planetary formation; ices

91: 

92: \clearpage

93: 

94: \section{Introduction}

95: \label{introduction}

96: 

97: Previous core accretion simulations of the formation of Uranus and

98: Neptune require the solar nebula lifetime to be at least 10~Myr longer

99: than the observed median protostellar disk lifetime (2--3 Myr; Haisch et

100: al. 2001).  Similarly, the simulation of Saturn's formation by Alibert

101: et al. (2001) allows the planet to form in 2.5~Myr, but only if it

102: migrates from 11.9~AU to 9.5~AU during formation.

103: %and incorporates 100\% of the available

104: %planetesimals into its core.

105: This scenario is incompatible with the

106: outward migration of Saturn predicted by the Nice model of planetary

107: dynamics (Tsiganis et al. 2005)

108: %and the presence of $\sim 40 M_{\oplus}$

109: %of comets scattered into the Oort cloud by Jupiter and Saturn (Weissman

110: %1996).

111: This research was motivated by the need to create a model of the

112: solar nebula with enough solid mass to form Jupiter, Saturn, Uranus and

113: Neptune via core accretion at heliocentric distances of $5 < R < 20$~AU.

114: 

115: For core accretion simulations, planetesimal mass available for building

116: giant planet cores is usually calculated by invoking the scaled

117: minimum-mass solar nebula with surface density $\Sigma \propto

118: R^{-3/2}$, obtained from spreading out the current mass of the planets

119: (Weidenschilling 1977), and assuming a gas/solid ratio of $\sim 70$.

120: However, a single planet core embedded in a disk of planetesimals grows

121: at the rate of

122: \begin{equation}

123: {dM_{\rm core} \over dt} = \pi a_{\rm core}^2 \Sigma_{\rm solid}

124: \Omega F_g ,

125: \label{planetgrowth}

126: \end{equation}

127: where $F_g$ is a factor taking into account gravitational focusing by

128: the core (Safronov 1969), $\Sigma_{\rm solid}$ is the solid surface

129: density of planetesimals, $\Omega$ is the orbital angular frequency of

130: the core, and $a_{\rm core}$ is the core's effective capture radius.

131: For Keplerian orbits, we have

132: \begin{equation}

133: \Omega = \left ( {GM_* \over R^3} \right )^{1/2} .

134: \label{kepler}

135: \end{equation}

136: The $R^{-3/2}$ dependence of $\Omega$ means that even planets forming

137: in a disk with uniform surface density should, in the end, reach masses

138: that suggest $\Sigma \propto R^{-3/2}$ if growth rate is not taken

139: into account.  A near-uniform surface density distribution, in which

140: mass is not concentrated in the inner disk, gives higher surface

141: densities in the outer disk and may speed up the formation

142: of Uranus and Neptune.

143: 

144: Furthermore, the canonical gas/solid ratio of 70, which is based on the

145: composition of Comet Halley (Jessberger et al. 1989), includes only one

146: ice in the solid inventory---H$_2$O.  However, there were certainly

147: other ices present in the solar nebula: the Deep Impact ejecta from

148: Tempel 1 contained $\sim 1000$~tons of excavated CO, CO$_2$ and CH$_3$OH

149: (A'Hearn 2008).  Taking a full accounting of the solids in the solar

150: nebula may reveal other volatile mass sources besides H$_2$O that

151: condense in the outer solar nebula and improve the formation prospects

152: of Uranus and Neptune.

153: 

154: In this work, we calculate the time-evolving solid surface density

155: available for giant planet formation.  We make two updates to previous

156: solar nebula simulations: (1) we relax the assumption of a steady-state

157: disk and allow for viscous redistribution of mass, which tends to drive

158: the surface density profile toward uniformity, and (2) we add a chemical

159: reaction network to trace the formation and freezeout of volatiles.

160: 

161: This paper is organized as follows.  In \S \ref{overview}, we outline

162: our method for calculating solid surface density in the solar nebula:

163: which codes we use, how they are linked, and how they differ from

164: previous solar nebula models.  In \S \ref{alphadisk}, we describe our

165: protostellar disk model and use it to calculate the total surface

166: density (gas+solid) and temperature evolution of the disk.  \S

167: \ref{chemmodel} details the physics behind our chemical reaction

168: network, the abundances of the major ice species, and the location of

169: the ice lines.  We calculate the solid surface density available for

170: planet formation in \S \ref{planetesimal} and present our conclusions in

171: \S \ref{conclusions}.

172: 

173: \section{Simulation Overview}

174: \label{overview}

175: 

176: This simulation consists of two independent, non-interacting codes used

177: in sequence.  The first code is a non-steady-state, 1+1-dimensional

178: $\alpha$-disk model that simulates the dynamical evolution of the solar

179: nebula.  Code 1 gives the surface density distribution $\Sigma(R,t)$,

180: along with the vertical and radial temperature and volume density

181: distributions $T(R,z,t)$ and $\rho(R,z,t)$.

182: 

183: The results from Code 1 fill an important niche in the body of work on

184: protostellar disk evolution.  Currently available are 2-d, static disk

185: models that include stellar insolation and radiogenic heating (e.g.

186: D'Alessio et al. 2006); quasi-steady-state 1+1-d models where

187: $\dot{M}(t)$ follows an imposed prescription (e.g. Hersant et al. 2001);

188: and non-steady-state models with only one zone in the vertical direction

189: (Ruden and Pollack 1991).  The evolving protostellar disks of Boss

190: (2007) include 3-d hydrodynamics and radiative diffusion, but cover a

191: timespan of $< 4000$ years.

192: 

193: In this work, we couple full vertical structure models with the

194: time-dependent, non-steady-state surface density evolution of the disk

195: and present the first long-term solar nebula simulation with viscous

196: angular momentum transport determined self-consistently with vertical

197: structure.  For a description of how the resulting dynamical evolution

198: of our disk differs from previous simulations, see \S \ref{massflux}.

199: 

200: Code 2 is a non-equilibrium chemical model, including gas reactions,

201: grain-surface reactions, freezeout and desorption.  This code traces the

202: abundance of 211 species, including 58 ices, as the disk cools and the

203: ice lines move inward.  Chemical reaction, freezeout and desorption

204: rates are governed by the temperature and density of the solar nebula

205: gas and grains.  We use the midplane temperatures and densities

206: calculated by Code 1 to generate reaction rate coefficients for Code 2.

207: 

208: %Even though the energy release during freezeout and

209: %exothermic reactions can heat the disk and affect its structure and

210: %evolution, we do not yet have the capacity to integrate a complex

211: %chemical-reaction network into a long-term dynamical simulation.  

212: 

213: Previous chemical models of the solar nebula include those of Aikawa et

214: al. (1996), Willacy and Langer (2000), and Ilgner and Nelson (2006).

215: \cite{aikawa96} extended the minimum-mass solar nebula (MMSN; Hayashi

216: 1981) to 800~AU to predict the observed depletion of gaseous CO in the

217: outer regions of T-Tauri disks.  Their calculation included gas-phase

218: reactions of CO and other C, H, O, N and S-based molecules, freezeout of

219: CO and its reaction products, formation of H$_2$, and electron-ion

220: recombination on grain surfaces.  The density and temperature of the

221: underlying disk were assumed to stay constant with time, while the

222: time-evolving midplane CO abundance was calculated with an independent

223: sequence of radial one-zone models.

224: 

225: Willacy and Langer (2000) calculated the vertical distribution and

226: column density of common interstellar molecules as a function of time in

227: the flared protoplanetary disk of Chiang and Goldreich (1997).  They

228: used a similar species set to Aikawa et al. (1996) but added

229: grain-surface reactions and photodesorption to the chemistry.  The

230: simulation consisted of a 2-d grid of noninteracting one-zone models in

231: the $R-z$ plane.  The underlying temperature and density distribution of

232: the protostellar disk remained constant with time.

233: 

234: Finally, Ilgner and Nelson (2006) embedded a five-species chemical

235: reaction network in an evolving $\alpha$-disk model, fully coupling the

236: chemical and dynamical evolution of the disk.  They used the chemical

237: reaction network to trace the ionization fraction, which is affected by

238: charge-transfer reactions, electron-ion recombination, and viscous

239: diffusion.  The simulation was computationally tractable due to the

240: small reaction network (5 species, 4 reactions), and the limited extent

241: of the protostellar disk (0.1--10~AU).

242: 

243: Code 2 of this work is the first to combine an extensive chemical

244: reaction network, including ice formation and grain-surface reactions,

245: with a time-evolving protostellar disk model (as used by Ilgner and

246: Nelson 2006).  Our chemical model consists of a radial sequence of

247: noninteracting one-zone models, as in Aikawa et al. (1996).  The disk

248: model in Code 1 therefore functions as a stand-alone input to the

249: chemical model of Code 2.  Our models differ from those of Aikawa et al.

250: (1996) and Willacy and Langer (2000) in that we allow the disk

251: temperature and density to change with time, as prescribed by the

252: results of Code 1.  Although Code 1 simulates the full vertical

253: structure of the protostellar disk, we apply Code 2 only to the

254: midplane, which is where solids settle and build planetesimals.

255: 

256: Code 2 calculates the abundance of each ice and gas species as a

257: function of heliocentric radius and time.  Combining these results with

258: information from the literature about the composition of refractory

259: grains, we calculate the gas/solid mass ratio in the solar nebula.  This

260: ratio tells us what fraction of the total surface density calculated by

261: Code 1 is solid and can build giant planet cores.  The end result of

262: both codes is the solid surface density distribution $\Sigma_{\rm

263: solid}(R,t)$.

264: 

265: Since most molecules in the solar nebula formed in a molecular cloud, we

266: first run Code 2 for molecular cloud conditions ($T = 10$~K, hydrogen

267: number density $2 \times 10^4 \; {\rm cm}^{-3}$, grain radius $0.258 \mu

268: m$).  The molecular cloud model begins with atomic gas of solar

269: composition (except for hydrogen, 99\% of which is in H$_2$) and bare

270: refractory grains at the standard interstellar abundance $n_g = 10^{-12}

271: (n_{\rm H} + 2 n_{\rm H2})$, where $n_g$ is the grain abundance and

272: $n_{\rm H}$ and $n_{\rm H2}$ are the abundances of H and H$_2$,

273: respectively.  We run the molecular cloud model for $10^6$~yr and use

274: the resulting ice-gas mixture as the starting composition of the solar

275: nebula.

276: 

277: \section{Disk Model}

278: \label{alphadisk}

279: 

280: \subsection{Overall Model Formulation}

281: \label{framework}

282: 

283: The goals of this simulation are to calculate the time-evolving

284: temperature and density in the solar nebula, which determine the

285: reaction rates for our kinetic model of ice formation, and to follow the

286: viscous evolution of the surface density profile.  Although from a

287: planet formation perspective we are primarily concerned with the ice

288: inventory at the solar nebula midplane, an accurate dynamical

289: description of the midplane requires understanding the vertical energy

290: balance in the disk.  To simplify the problem of long-term protostellar

291: disk evolution, we use the following assumptions:

292: 

293: \begin{enumerate}

294: 

295: \item The disk is axisymmetric and symmetric about the midplane.

296: 

297: \item The disk is geometrically thin: $H / R << 1$.

298: 

299: \end{enumerate}

300: 

301: The disk is represented in cylindrical polar coordinates $R$

302: (heliocentric distance) and $z$ (height above midplane), where the

303: $z$-axis is the rotation axis of the nebula.  Assumption 1 reduces the

304: physical 3-dimensional disk to a 2-dimensional quadrant with zero flux

305: at the midplane, allowing us to suppress the azimuthal coordinate

306: $\theta$.  Assumption 2 allows the vertical and radial dimensions of the

307: disk to be decoupled to form a 1+1-dimensional framework: at each radial

308: gridpoint resides an independent vertical structure model.  At each

309: timestep, energy balance between viscous heating and radiative cooling

310: is solved within in a single annulus.

311: 

312: \subsection{Radial Diffusion}

313: \label{raddiff}

314: 

315: We calculate the radial motion of mass using the surface density

316: diffusion equation of Lynden-Bell and Pringle (1974):

317: \begin{equation}

318: {\partial \Sigma \over \partial t} = {3 \over R} {\partial \over

319: \partial R} \left [ R^{1/2} {\partial \over \partial R} \left ( \Sigma

320: \nu R^{1/2} \right ) \right ] .

321: \label{diffusion}

322: \end{equation}

323: Eq. \ref{diffusion} is valid when the protostar's accretion rate is

324: small, so the sun is almost fully assembled at the beginning of the

325: simulation.  This assumption is consistent with the T-Tauri phase of

326: disk evolution.  We take the viscosity $\nu$ governing radial diffusion

327: to be the midplane value.

328: 

329: To calculate $\nu_{\rm midplane}$, we make the following 

330: assumptions about angular momentum transport in the disk:

331: 

332: \begin{enumerate}

333: 

334: \item {\it Viscous stresses follow the standard $\alpha$-viscosity

335: prescription} (Shakura and Syunyaev 1973).  We assume the

336: magnetorotational instability (Balbus and Hawley 1991; hereafter MRI)

337: provides the requisite turbulent viscosity.

338: 

339: \item {\it The grains and gas are well mixed.} This assumption allows us

340: to track the disk's solid surface density evolution by modeling only a

341: single fluid.  

342: 

343: \end{enumerate}

344: 

345: The turbulent viscosity is

346: \begin{equation}

347: \nu = \frac{2}{3} \alpha c_s H , 

348: \label{visc}

349: \end{equation}

350: where $c_s$ is the isothermal sound speed and $H$ is the modified

351: pressure scale height, softened into a nonsingular form (Milsom et al.

352: 1994):

353: \begin{equation}

354: H = { c_s / \Omega \over \sqrt{ 1 + \left ( {2

355: z^2 \Omega^2 / c_s^2} \right ) }} .

356: \label{scale}

357: \end{equation}

358: The angular velocity, neglecting disk self-gravity, is

359: \begin{equation}

360: \Omega = \left [ {G M_* \over \left ( R^2 + z^2

361: \right )^{3/2} } \right ]^{1/2} ,

362: \label{omega}

363: \end{equation}

364: and the isothermal sound speed is

365: \begin{equation}

366: c_s^2 = {R_g \over \mu} T .

367: \label{soundspeed}

368: \end{equation}

369: In parts of the disk where $H > z_{\rm surf}$, the disk surface height,

370: we substitute $z_{\rm surf}$ for $H$ in Eq. \ref{visc}.  We take a

371: mean molecular weight of $\mu = 2.33$~g~mol$^{-1}$ (Ruden and Pollack

372: 1991).  Based on recent global-disk MHD simulations (Lyra et al. 2008),

373: we choose $\alpha = 2 \times 10^{-3}$.  We further discuss the interplay

374: between $\alpha$, disk mass and surface density profile, along with our

375: constraints on these parameters, in \S \ref{initial}.  All free

376: parameters in the disk model are listed in Table \ref{freedisk}.

377: 

378: %If MRI is the source of viscous turbulence, the grains must be small

379: %enough to be coupled to the gas, but not so small as to remove charge

380: %from the gas and lower the ionization fraction.  We assume a mean grain

381: %radius of $0.258 \; \mu m$, slightly above the $0.1 \; \mu m$

382: %minimum required to keep enough free charge in the gas to sustain MRI

383: %turbulence (Turner et al.  2007).

384: 

385: Although we assume turbulence is caused by MRI, we do not include the

386: radial variation of the viscosity coefficient $\alpha$ caused by the

387: dead zone (e.g. Kretke and Lin 2007, Reyes-Ruiz et al. 2003).

388: %We will

389: %investigate the effect of the dead zone on disk evolution in a

390: %forthcoming paper.

391: However, we do account for the cessation of MRI

392: turbulence in the tenuous outer regions of the disk, where ions and

393: neutrals may be only weakly coupled.  \cite{hawley98} find that neutrals

394: participate fully in MRI turbulence only when the ion-neutral collision

395: timescale $T_{\rm coll}$ is much less than the orbital timescale;

396: $T_{\rm coll} \leq 0.01 / \Omega$.  Ions and neutrals decouple and MRI

397: turbulence ceases when $T_{\rm coll} \geq 100 / \Omega$.  In the

398: transition region between coupled and decoupled neutral-ion interaction,

399: where $0.01 \leq T_{\rm coll} \Omega \leq 100$, we decrease $\log

400: \alpha$ linearly from its fiducial value.

401: 

402: To calculate $T_{\rm coll}$, we find the mean free time of a neutral

403: particle traveling through a gas of moving ions.  The neutrals and ions

404: collide at an average angle of $90^{\circ}$ and thus a mean impact

405: velocity of

406: \begin{equation}

407: v_{\rm coll} = \sqrt{v_n^2 + v_i^2} = v_n \left (1 + {\mu_n \over \mu_i}

408: \right )^{1/2} ,

409: \label{vcoll}

410: \end{equation}

411: where $\mu_n$ and $\mu_i$ are the molecular weights of neutrals and

412: ions, respectively.  The mean collision time is then

413: \begin{equation}

414: T_{\rm coll} = {1 \over \pi \left ( a_n + a_i \right )^2 n_i v_n

415: \left (1 + \mu_n / \mu_i \right )^{1/2}} ,

416: \label{tcoll}

417: \end{equation}

418: where $a_n$ and $a_i$ are the radii of neutrals and ions, $v_n =

419: \sqrt{3} c_s$ is the mean velocity of a neutral and $n_i$ is the number

420: density of ions:

421: \begin{equation}

422: n_i = {f_i \rho \over m_i}

423: \label{ni}

424: \end{equation}

425: We assume an ionization fraction of $f_i = 10^{-9}$ and an ion mass of

426: $m_i = 30$~AMU.  $f_i$ and $m_i$ are listed along with other free

427: parameters in the simulation in Table \ref{freedisk}.

428: 

429: \subsection{Vertical Structure}

430: 

431: Since the disk midplane viscosity is determined by the balance between

432: viscous energy generation and radiative energy losses at the

433: photosphere, we calculate full vertical structure models at each

434: radial grid point.  We use the following assumptions about the mass and

435: energy balance in the vertical direction:

436: 

437: \begin{enumerate}

438: 

439: \item {\it The disk stays in hydrostatic equilibrium}.  Even though the

440: mass distribution evolves, the vertical sound crossing timescale is much

441: less than the radial diffusion timescale.

442: 

443: \item {\it We neglect stellar irradiation as an energy source, as it has

444: little effect on the midplane temperature}.  In section \ref{initial} we

445: show that the disk is self-shadowed ($H/R$ decreases with $R$), making

446: this assumption self-consistent with our simulation results.

447: 

448: \end{enumerate}

449: 

450: A system of three coupled differential equations specifies the vertical

451: structure of the disk.  First, the disk is in hydrostatic equilibrium:

452: \begin{equation}

453: {\partial P \over \partial z} = -\rho \Omega^2 z .

454: \label{hydrostatic}

455: \end{equation}

456: Second, the viscous energy generation rate per unit volume (Pringle

457: 1981) is

458: \begin{equation}

459: {\partial F \over \partial z} = \frac{9}{4} \nu \Omega^2 \rho .

460: \label{flux}

461: \end{equation}

462: Third, where vertical energy transport is regulated by radiative

463: diffusion and energy transport in the radial direction is negligible

464: (thin disk assumption), the energy equation is

465: \begin{equation}

466: {\partial T \over \partial z} = {\partial P \over \partial z} \nabla {T

467: \over P} .

468: \label{tempgrad}

469: \end{equation}

470: 

471: To calculate the thermodynamic gradient $\nabla = d \ln T / d \ln P$, we

472: use the Schwarzschild criterion for stability against convection

473: (Kippenhahn and Weigert 1994):

474: \begin{equation}

475: \nabla = \left \{ \begin{array}{l} \nabla_{\rm rad},  

476: \nabla_{\rm rad} \le \nabla_{\rm ad} \\ \nabla_{\rm conv},  

477: \nabla_{\rm rad} > \nabla_{\rm ad} \end{array} \right. .

478: \label{nabla}

479: \end{equation}

480: In Eq. \ref{nabla}, $\nabla_{\rm ad}$ is the adiabatic gradient,

481: which is $2/7$ for a diatomic gas; $\nabla_{\rm conv}$ is the convective

482: gradient; and $\nabla_{\rm rad}$ is the radiative gradient: 

483: \begin{equation}

484: \nabla_{\rm rad} = \frac{3}{4} {\kappa P F \over a c \Omega^2 z T^4} ,

485: \label{nablarad}

486: \end{equation}

487: where $a$ is the radiation density constant and $\kappa$ is the local

488: Rosseland mean opacity (for a description of the opacity tables used in

489: this calculation, see \S \ref{opsec}).

490: 

491: We use mixing-length theory to calculate $\nabla_{\rm conv}$.  We begin

492: the convection treatment by defining two dimensionless quantities:

493: \begin{equation}

494: W \equiv \nabla_{\rm rad} - \nabla_{\rm ad}

495: \label{convw}

496: \end{equation}

497: and

498: \begin{equation}

499: U \equiv {3 a c T^3 \over c_P \rho^2 \kappa \ell_m^2}

500: \sqrt{8 \varepsilon} .

501: \label{convu}

502: \end{equation}

503: In Equation \ref{convu}, $c_P$ is the specific heat at constant pressure

504: and $\ell_m$ is the mixing length.  As in \cite{milsom94}, we define

505: $\ell_m$ as the minimum of the convection zone top or the pressure scale

506: height $H$ (Equation \ref{scale}).  $\varepsilon$ is an analytical

507: softening of the ratio $H / \Omega^2 z$, eliminating the singularity at

508: $z = 0$:

509: \begin{equation}

510: \varepsilon = \sqrt{1 \over \Omega^4 (1 + z^2 / H^2)} .

511: \label{varparm}

512: \end{equation}

513: Defining a new variable $\xi$ as

514: \begin{equation}

515: \xi \equiv \nabla_{\rm conv} - \nabla_{\rm ad} + U^2 = \sqrt{\nabla_{\rm

516: conv} - \nabla_e} + U

517: \label{xieqn}

518: \end{equation}

519: (where $\nabla_e$ is the thermodynamic gradient of an individual

520: convective element), we can then solve the cubic equation

521: \begin{equation}

522: \left ( \xi - U \right )^3 + \frac{8U}{9} \left ( \xi^2 - U^2 - W \right

523: ) = 0

524: \label{cubic}

525: \end{equation}

526: to find $\nabla_{\rm conv}$.  For the derivations of Equations

527: \ref{xieqn} and \ref{cubic} see Chapter 7 of \cite{kippenhahn94}.

528: 

529: Finally, we close the system of equations with the ideal gas equation of

530: state,

531: \begin{equation}

532: P = \left ({R_g \over \mu} \right ) \rho T .

533: \label{idealgas}

534: \end{equation}

535: 

536: \subsection{Computational Methods and Boundary Conditions}

537: 

538: We begin the simulation with a specified surface density profile for the

539: disk, $\Sigma \propto r^{-3/2}$ (Hayashi 1981).  The total disk mass

540: used to normalize $\Sigma(r)$ and the viscosity coefficient $\alpha$ are

541: determined by constraints of gravitational stability

542: % presence of

543: %crystalline silicates in comets

544: and the need for enough mass to form

545: giant planets.  For full details of how the $t = 0$ disk is constructed,

546: see \S \ref{initial}.  We also assume the disk is embedded in a medium

547: with an ambient temperature $T_{\rm amb}$ (for example, a molecular

548: cloud core).  This places a firm lower limit on the nebula temperature

549: and forces inactive regions of the disk to assume an isothermal vertical

550: structure.

551: 

552: As in \cite{bell97}, we place the disk surface at

553: optical depth $\tau = 0.03$.  We begin the vertical structure

554: calculation with guesses for the free parameters $T_{\rm surf}$ and

555: $\rho_{\rm surf}$, the density and temperature at the $\tau = 0.03$

556: surface.

557: 

558: The net flux leaving the disk surface is determined by the accretion

559: temperature $T_{\rm acc}$, generated solely by viscous stresses:

560: \begin{equation}

561: F = \sigma T_{\rm acc}^4 .

562: \label{teff}

563: \end{equation}

564: We require a value of $T_{\rm acc}$ at the disk surface to to start

565: integrating Equations \ref{flux} and \ref{tempgrad}.  The free parameter

566: $T_{\rm surf}$ is the flux sum of the accretion temperature $T_{\rm

567: acc}$ and the ambient temperature $T_{\rm amb}$, invoking the Eddington

568: approximation:

569: \begin{equation}

570: T_{\rm surf}^4 = T_{\rm amb}^4 + \frac{3}{4} \left [ \tau + f \left (

571: \tau \right ) \right ] T_{\rm acc}^4 .

572: \label{fluxsum}

573: \end{equation}

574: In equation \ref{fluxsum}, $f(\tau)$ is the Hopf function, with a value

575: of 0.601242 for $\tau = 0.03$ (Bell et al. 1997; Mihalas 1978).

576: 

577: After calculating $P$ from Eq. \ref{idealgas}, the variables of

578: integration $T_{\rm acc}$, $P$ and $F$ are all specified.  We find the

579: height of the $\tau = 0.03$ surface by solving the following equation

580: for z:

581: \begin{equation}

582: { 1 \over \Omega^2 \, z } = {\tau \over \kappa(\rho, T) \, P} .

583: \label{zsurfeqn}

584: \end{equation}

585: Integration proceeds inward from the disk surface to the midplane using

586: the Runge-Kutta method with adaptive stepsize control (Press et al.

587: 1992).  We repeat the integration, adjusting $T_{\rm surf}$ and

588: $\rho_{\rm surf}$ with the Newton-Raphson algorithm (Press et al. 1992),

589: until a solution is found that satisfies

590: \begin{equation}

591: F_{\rm midplane} = 0

592: \label{fluxconv}

593: \end{equation}

594: (following Assumption 1, \S \ref{framework}) and

595: \begin{equation}

596: 2 \int_{z=0}^{z_{\rm surf}} \rho \: dz = \Sigma .

597: \label{sdconv}

598: \end{equation}

599: 

600: After the vertical structure is solved for each radius, we use the

601: resulting $\nu_{\rm midplane}$ to update $\Sigma(r)$.  We implement

602: implicit finite differencing to solve Eq. \ref{diffusion},

603: adjusting the timestep $\Delta t$ so that surface density varies by a

604: maximum of 1.5\% during a single timestep (e.g. Ruden and Pollack 1991).

605: 

606: The radial grid is equispaced in $r^{1/2}$ and contains 600 cells

607: between the inner boundary at $0.3$~AU and the initial outer boundary at

608: $30$~AU.  We implement a zero viscous-stress inner boundary condition,

609: so matter falls directly onto the star from the innermost gridpoint.

610: Although the actual accretion shock is at a much smaller radius than

611: $0.3$~AU, \cite{ruden86} find that the location of the inner boundary

612: $R_{\rm in}$ is unimportant as long as $R_{\rm out} >> R_{\rm in}$.  We

613: allow the disk to expand freely from the original outer boundary, so new

614: radial grid cells are added to the disk as the simulation progresses.

615: 

616: \subsection{Opacities}

617: \label{opsec}

618: 

619: Solving Eq. \ref{nablarad} requires knowing local mean opacity.

620: Semenov et al. (2003) calculated Rosseland mean opacities due to icy

621: grains for temperatures down to 5K.  The authors modeled several

622: different grain types, including spheres, aggregates, and porous grains.

623: We use the 5-layered sphere topology, where each grain consists of a

624: silicate/iron nucleus covered with successive layers of volatiles as the

625: temperature decreases.  Following Semenov et al., we assume the grains

626: condensed from gas of solar composition, with the abundances given by

627: Helling et al. (2000).  These abundances are primarily derived from the

628: catalog of Anders and Grevesse (1989).  We list the abundances of the

629: most common elements in Table \ref{solarcomp}, along with the types of

630: solids they can form: rock, metal, ice, or refractory CHON (e.g.

631: graphite or kerogen; see Jessberger et al. 1988).

632: 

633: At temperatures above 1200K, iron and silicates sublimate and the main

634: opacity source switches from dust to molecular gas.  Semenov et al.

635: (2003) caution that the gas opacities in their model are only

636: approximate in the temperature range where dust and gas opacity are

637: comparable.  For temperatures above 1000K, we use the opacities of

638: Ferguson et al. (2005), updated for our chosen set of solar abundances

639: (J. Ferguson, private communication).  It should be noted that the

640: number of atomic and molecular lines included in the calculation of an

641: opacity table has a greater impact on the resulting opacities than minor

642: changes in assumed solar composition.

643: 

644: Regions where opacity is a strong function of temperature can cause

645: convergence problems for the Newton-Raphson algorithm, which will

646: oscillate between two solutions with similar temperatures but quite

647: different opacities.  To mitigate this problem, we smoothed the raw

648: opacity tables with an averaging filter.  Figure \ref{opplot} shows

649: opacity as a function of temperature and density in our model.

650: 

651: 

652: \subsection{Initial Conditions}

653: \label{initial}

654: 

655: Constraints on the early stages of the solar nebula come from

656: observations of Class 1 protostars and T-Tauri stars and the presence

657: and orbital configuration of the planets.  We require our $t = 0$ disk

658: to meet two criteria:

659: 

660: %analysis of the

661: %chemical composition of comets and other pristine bodies,

662: 

663: \begin{enumerate}

664: 

665: \item The disk is gravitationally stable to axisymmetric perturbations

666: at all radii.

667: 

668: %\item The temperature in the inner nebula, $r \leq 2$~AU, is hot enough

669: %to produce crystalline silicates ($> 800$~K; Gail 1998).

670: 

671: \item The disk is massive enough to produce Jupiter and Saturn by core

672: accretion within the observed mean protostellar disk lifetime of

673: 2--3~Myr (Haisch et al. 2001).

674: 

675: \end{enumerate}

676: 

677: To test whether a particular disk meets Criterion 1, we calculate the

678: Toomre Q parameter at every radial gridpoint.  Gravitational stability

679: requires that

680: \begin{equation}

681: Q = {c_s \Omega \over \pi G \Sigma} > 1 .

682: \label{toomreq}

683: \end{equation}

684: Gravitationally unstable disks will fragment and form stellar or

685: substellar clumps on a $10^3$ year timescale.  The lack of a binary

686: companion for the sun and the near-circular, non-chaotic orbits

687: displayed by all the planets makes gravitational fragmentation anytime

688: during solar nebula evolution unlikely.

689: 

690: Our $Q > 1$ criterion determines the combinations of viscosity parameter

691: $\alpha$, mass distribution and total disk mass that create viable

692: accretion disks.  High values of $\alpha$ act to stabilize the disk by

693: increasing energy generation and therefore sound speed.  Large disk

694: masses destabilize the disk by increasing $\Sigma$.  Surface density

695: profiles that decline steeply with radius damp gravitational instability

696: by loading the mass preferentially toward the inner disk, where

697: dissipative shear stresses are strongest.

698: 

699: Hersant et al. (2001; hereafter HGH) used an evolving $\alpha$-disk

700: model and measured solar-system D/H ratios to constrain the structure of

701: the solar nebula.  They chose the surface density profile $\Sigma

702: \propto R^{-1}$.  Even with an extremely high turbulent energy

703: generation rate, $\alpha \geq 0.01$, disks with this surface density

704: distribution are gravitationally stable only if $M_{\rm disk} \leq 0.08

705: M_{\odot}$.  However, Criterion 3 requires that there be sufficient

706: solid mass to form Jupiter and Saturn within $\sim 3$~Myr.  Assuming a

707: gas/solid mass ratio of 70 (Pollack et al. 1996), a disk with $\Sigma

708: \propto R^{-1}$ out to 30~AU and $M_{\rm disk} = 0.08 M_{\odot}$ has a

709: solid surface density of only 6~g~cm$^{-2}$ in Saturn's feeding zone at

710: 9.5~AU.  \cite{hubickyj05} found that a planet forming from

711: 6~g~cm$^{-2}$ of solids at 5~AU takes more than 13~Myr to initiate rapid

712: gas accretion, necessary for forming a massive atmosphere.  Since planet

713: formation proceeds more slowly at larger heliocentric distances (Eq.

714: \ref{planetgrowth}), this disk is unlikely to be able to form Saturn.

715: We thus require a more massive disk, with a steeper surface density

716: profile to keep the outer regions gravitationally stable.

717: 

718: We choose a disk mass $M_{\rm disk} = 0.12 M_{\odot}$ and a surface

719: density $\Sigma \propto R^{-3/2}$ (Hayashi 1981).  We take $\alpha = 2

720: \times 10^{-3}$, the approximate azimuthally and vertically averaged

721: value at Jupiter's heliocentric distance in the global MHD turbulence

722: simulations of Lyra et al. (2008).  Our solar nebula has inner and outer

723: radii of $R_{\rm in} = 0.3$ AU and $R_{\rm out} = 30$~AU, covering the

724: range of heliocentric distances where planets exist today.  It is worth

725: noting that studies of the relative motions of dust and gas predict a

726: much larger outer radius: the gas drag on cm-sized bodies can cause them

727: to be quickly lost to the Sun unless they have a very large distance to

728: migrate (Stepinski and Valageas [1997]; Ciesla and Cuzzi [2006]).

729: Mitigating the deleterious effect of gas drag on large grains may

730: require that the nebula has an additional low-density component outside

731: $R = 30$~AU, which we do not model in this work.

732: 

733: We assume our $t = 0$ disk coincides with the beginning of the sun's

734: T-Tauri phase.  Most of the sun's mass is already in place: $M_* = 0.95

735: M_{\odot}$ at the start of our simulation.  We assume an ambient

736: temperature of 20~K, approximating a disk embedded in the remnants of a

737: molecular cloud.  All free parameters in the disk model are listed in

738: Table \ref{freedisk}.

739: 

740: Although our model neglects solar irradiation, we find that the giant

741: planet-forming region of our t=0 disk is in fact shadowed.  The maximum

742: aspect ratio $H / R$ of the $\tau = 2/3$ photosphere occurs at $R = 2$

743: AU.  Beyond this radius, no part of the disk receives direct sunlight.

744: The shadow persists until $t \sim 1$~Myr, when regions with $R > 20$~AU

745: attain an aspect ratio greater than the local maximum at 2~AU.  In this

746: way, our disk evolution is broadly consistent with model T-Tauri disk

747: SEDs, which require flaring to explain observed spectra (e.g. Chiang et

748: al. 2001), yet the giant planet-forming region ($\sim 5$--20~AU) still

749: receives no solar heating during the icy planetesimal formation epoch.

750: 

751: \subsection{Disk Model Results}

752: \label{viscous}

753: 

754: \subsubsection{Surface Density}

755: \label{surfacedensity}

756: 

757: We run the solar nebula model for 2~Myr, about the median age of

758: observed protostellar disks (Haisch et al. 2001).  Figure \ref{sdplanet}

759: shows the evolution of the disk surface density profile in the giant

760: planet-forming region, 5-20~AU (Tsiganis et al. 2005).  The surface

761: density profile at $t = 0$ is shown on the upper left.  Subsequent

762: panels show snapshots of $\Sigma(R)$ as the simulation progresses

763: plotted in black.  For reference, surface density profiles from

764: preceding plot panels are retained in gray.  Figure \ref{sdwholedisk}

765: shows the surface density evolution of the entire disk as it expands

766: from 30~AU at $t = 0$ to $> 80$~AU after 2~Myr.

767: 

768: From examining Fig. \ref{sdplanet}, we can divide the evolution of the

769: solar nebula's giant planet-forming region into two epochs.  In Epoch 1,

770: the surface density profile flattens as mass is redistributed from the

771: inner disk to the outer disk.  At $t = 0$, the net mass flow is outward

772: everywhere except at $R < 1$~AU, which is accreting inward toward the

773: star.  A wave of local surface density enhancement consequently

774: propagates outward, reaching $\sim 10$~AU at $t = 3 \times 10^4$ yr and

775: dissipating near $\sim 17$~AU at $t = 3 \times 10^5$ yr.  At the end of

776: Epoch 1, at $t \sim 5 \times 10^5$~yr, the solar nebula has assumed a

777: surface density profile flatter than $\Sigma \propto R^{-1}$ (equal mass

778: in each annulus) in all but the inner $0.5$~AU of the disk.  The effect

779: of this mass redistribution epoch, therefore, is to drive the surface

780: density distribution toward uniformity.

781: 

782: In Epoch 2, $t > 5 \times 10^5$~yr, surface density decreases with time

783: everywhere in the disk.  In Fig. \ref{sdwholedisk}, we see the solar

784: nebula grow more tenuous as it expands: after 2 Myr, the total disk mass

785: is 0.067~$M_{\odot}$ spread over 80~AU, as opposed to 0.12~$M_{\odot}$

786: within 30 AU at the beginning of the simulation.  The protosun reaches a

787: final mass of 1.003~$M_{\odot}$.

788: 

789: Figure \ref{sdann} shows surface density as a function of time at 5~AU,

790: 10~AU, 15~AU and 20~AU.  At each heliocentric distance, we see the

791: temporary mass growth as the density wave passes through.  The largest

792: relative surface density increase $\Delta \Sigma / \Sigma_0$ occurs at

793: 11.5~AU, where the initial surface density of 442~g~cm$^{-2}$ grows by

794: 28\% to 558~g~cm$^{-2}$ in 85,000~yr.  {\it This surface density

795: enhancement moving outward through the giant planet-forming region

796: strongly favors solid planet-core formation.}

797: 

798: At the beginning of the sun's T-Tauri phase, before substantial grain

799: growth occurs, gas and grains are well mixed: accreting gas drags grains

800: with it.  If gas moves inward with a constant mass flux, surface density

801: decreases with time everywhere in the disk and solids are lost into the

802: sun.  Our net outward gas flow deposits small grains in the 5--15~AU

803: region within the planetesimal formation timescale of $\sim 5 \times

804: 10^4$~yr (Hubbard and Blackman 2006).  {\it These grains provide surface

805: area for ice deposition and add solid mass to the giant planet feeding

806: zones, speeding up solid core formation.}

807: 

808: The existence of Epoch 1 and consequent driving of the surface density

809: distribution toward uniformity is a robust result that we see even using

810: $t = 0$ surface density profiles with power-law indices of 0.7 and 1.

811: Our chosen fiducial disk remains stable to axisymmetric perturbations

812: (c.f. Eq. \ref{toomreq}) throughout its evolution even as mass

813: flows outward.

814: 

815: \subsubsection{Mass Flux}

816: \label{massflux}

817: 

818: In their study of the similarity solutions of viscously evolving disks,

819: \cite{lyndenbell74} found that the natural behavior of an accretion disk

820: is to diffuse and spread out, creating a net inward flow in the inner

821: disk, a turnaround radius, and an outward decretion flow in the outer

822: disk.  \cite{ruden86} uncovered the same behavior in numerical

823: simulations of the evolution of the minimum-mass solar nebula.  Why

824: haven't many other $\alpha$-disk models predicted the outward mass flow

825: into the trans-Saturnian region that could so improve the prospects for

826: giant planet formation?  Most previous evolving disk models such as HGH

827: and the pseudoevolutionary sequence of \cite{bell97} have required the

828: disk to be in steady state.  In steady-state models, each annulus

829: receives an equal inward-driven mass flux with the sun serving as the

830: sink particle.  There is therefore no chance for any part of the disk

831: interior to $R_{\rm out}$ at $t = 0$ to gain mass.  Note that the

832: assumption of steady-state, net inward accretion is still consistent

833: with viscous expansion of the disk from its original outer boundary, as

834: occurs in the HGH models and this work.

835: 

836: The non-steady-state evolving disk model of \cite{ruden91} assumed

837: turbulent stresses were caused solely by convection, and set the

838: viscosity to zero in optically thin regions where radiative transport is

839: efficient.  Far from the star, the shallow potential well does not allow

840: the disk to attain high volume densities, so reaching an optical depth

841: of $\tau > 1$ requires more mass than in the inner disk.  At $R =

842: 30$~AU, the outer radius at the beginning of the Ruden and Pollack

843: calculation, convection ceases almost immediately and the disk becomes

844: quiescent.  The inward accretion flow then removes enough mass from

845: progressively interior annuli to make them optically thin, leaving

846: behind a remnant surface density profile where $\Sigma$ {\it increases}

847: with $R$.

848: 

849: In the Ruden and Pollack simulation, mass is not redistributed from the

850: inner disk to the outer disk---the slope of $\Sigma(R)$ changes from

851: negative to positive because the inner parts of the disk remain

852: optically thick for the longest, participate in convection-driven

853: accretion the longest, and therefore lose the most mass.  However, in

854: their analysis of the coupling between the kinetic stress tensor

855: governing hydrodynamic stability, the magnetic stress tensor, and the

856: shear field, \cite{hawley99} find that MRI is uniquely capable of

857: providing turbulent angular momentum transport in disks.  By relaxing

858: the steady-state condition, but still permitting angular momentum

859: transport by MRI in optically thin regions (though MRI ceases if ions

860: and neutrals decouple; see \S \ref{raddiff}), we allow outward flows to

861: develop and viscous stress to redistribute mass from the inner disk to

862: the giant planet-forming region.

863: 

864: %Indeed, the immediate cessation of viscous

865: %evolution at $R \sim 30$~AU found by Ruden and Pollack is inconsistent

866: %with the work of HRH, who compare their disk model results with

867: %observed D/H ratios in planet atmospheres,

868: %satellites and comets to argue that the solar nebula

869: 

870: Figure \ref{mdot} shows time snapshots of the mass accretion rate as a

871: function of radius.  The location with $\dot{M} = 0$ is the turnaround

872: between inward accretion and outward ``decretion.'' This turnaround

873: radius moves outward over time, so that more and more of the disk is

874: accreting toward the star.  This behavior is required to open a hole in

875: the inner disk and make the transition from classical T-Tauri to

876: non-accreting, weak-lined T-Tauri star.  Instead of accreting onto the

877: star, the outwardly decreting material ($R > 15$~AU) will eventually be

878: lost to photoevaporation (Alexander et al. 2006).

879: 

880: %Our prediction of a decretion disk is consistent with the analysis of Desch (2007), who 

881: %combined the masses and starting locations of the giant planets in the Nice model

882: %(Tsiganis et al. 2005) with an a net outward mass flow combined with photoevaporative

883: %truncation of the disk.  

884: 

885: \subsubsection{Midplane Temperature}

886: 

887: Figure \ref{midplanetemp} shows shapshots of the temperature of the

888: protostellar disk midplane at $R < 30$~AU as the disk evolves.  Outside

889: of 30~AU, the entire disk (midplane and above) is isothermal at the

890: ambient temperature throughout the simulation (20~K, plotted with a

891: dash-dotted line).  The first three panels of Fig. \ref{midplanetemp}

892: show that during Epoch 1, areas of the disk that increase in mass are

893: correspondingly heated: the temperature at 10~AU increases from 40~K at

894: $t = 0$ to 49~K at $3 \times 10^4$~yr.

895: 

896: We first evaluate the suitability of our model disk for delivering

897: crystalline silicates to comets, as detected by \cite{keller06} in comet

898: 81P/Wild 2 dust captured by the Stardust spacecraft.  The dashed line in

899: Fig. \ref{midplanetemp} shows the minimum temperature for silicate

900: crystallization of 800~K (Gail 1998).  (Note that forsterite formation

901: requires $T > 1067$~K; Hallenbeck et al. [2000].)  The inner nebula

902: retains temperatures above 800~K for $10^5$~yr.  \cite{dullemond06}

903: calculated the formation and transport rates of crystalline silicates in

904: a 1-d evolving $\alpha$-disk model with infall from a collapsing cloud.

905: Relaxing the steady-state assumption, they found that viscous spreading

906: of the disk, rather than turbulent radial mixing, could relocate

907: crystalline silicates produced early on in the inner, hot ($T \ga

908: 800$~K) parts of the disk to distances of $R > 15$~AU, where comets

909: formed.  A gradual outward push of crystalline silicates is consistent

910: with the nebula flow pattern observed in this work, and may explain the

911: anticorrelation between disk age and crystallinity in the inner disk ($R

912: < 10$~AU) discussed by \cite{vanboekel04}.  However, the danger that

913: crystalline silicates transported outward through the midplane would be

914: swept up by protoplanets before ever reaching the comet-forming region

915: led \cite{ciesla07} to favor more rapid transport than the decretion

916: flow in this model.

917: 

918: %One danger when crystalline silicates are formed in the inner nebula and

919: %transported outward through the disk midplane is that they will be swept

920: %up by forming protoplanets and never incorporated into comets.

921: %Fortunately, there are other possibilities for both the formation and

922: %transport of silicates.

923: %Scott and Krot (2005) inferred high cooling rates of chondritic

924: %meteorites, on the order of 10--1000 K~hr$^{-1}$, favoring localized,

925: %brief heating (as from a shock) as the source of crystalline silicates

926: %in asteroids.  In a study of melilite mantle formation on type B CAIs,

927: %Richter et al. (2006) find that temperatures greater than $\sim 1300$~K

928: %were sustained for only $\sim 10^4$~yr---not long enough to erase

929: %melilite-\aa kermanite zoning.

930: 

931: %Grains annealed in the quiescent disk could be transported to

932: %protocometary material at 5--20~AU by the decretion flow described in \S

933: %\ref{massflux} that occurred in the crystalline silicate-forming region

934: %during the first $2 \times 10^4$~yr.

935: 

936: %Turbulent mixing with $\alpha \ge 4

937: %\times 10^{-4}$ could then deliver crystalline silicates to the

938: %asteroid- and comet-forming regions within a few Myr (Desch 2007).

939: 

940: %The need for a hot, long-lived inner

941: %solar nebula with efficient radial mixing is far from clear.

942: 

943: The dotted line on Fig. \ref{midplanetemp} shows the H$_2$O ice line

944: at 160~K (Collings et al. 2004).  Although we require a chemical kinetic

945: model to determine the exact position and width of the ice line (since

946: ice deposition is not instantaneous but proceeds at a finite rate

947: limited by the gas velocity of H$_2$O and the available grain surface

948: area), we can nevertheless estimate its location and compare with

949: previous disk models.  In our simulation, the H$_2$O ice line is always

950: inside the formation zones of Saturn, Uranus and Neptune, and crosses

951: Jupiter's feeding zone (5.2~AU) at $5 \times 10^4$~yr.  All giant

952: planet cores, therefore, should have formed from icy grains.

953: 

954: %In the HGH fiducial model, the H$_2$O ice line crosses Jupiter's feeding

955: %zone at $\sim 5 \times 10^5$~yr, with a total surface density of $\sim

956: %300$~g~cm$^{-2}$.  Assuming a gas/solid mass ratio of 70 for icy regions

957: %(Pollack et al. 1996), the solid surface density available for forming

958: %Jupiter would be only $\sim 4.2$~g~cm$^{-2}$ in the HGH model.

959: %\cite{hubickyj05} find that forming Jupiter from 6~g~cm$^{-2}$ of solids

960: %takes 13~Myr.  The HGH disk, therefore, must be extraordinarily

961: %long-lived in order to form Jupiter.

962: 

963: After 2~Myr, the midplane outside 15~AU has become isothermal at the

964: ambient temperature.  This temperature, therefore, has a profound effect

965: on the ice inventory of the solar nebula.  The most volatile ices (CO,

966: CH$_4$ and N$_2$) have sublimation temperatures below 45~K.  Simply

967: raising the ambient temperature above this value would ensure that none

968: of these ices could freeze on grains.  However, C/H enrichments of

969: 20--40 over the solar value measured in Uranus and Neptune's atmospheres

970: indicate that direct gas accretion from the nebula cannot explain these

971: planets' composition (Baines et al. 1995).  We therefore postulate that

972: the outer solar nebula midplane contained CO and CH$_4$ ice, making our

973: choice of ambient temperature appropriate.  Further evidence for a cold

974: solar nebula comes from the observed noble gas enrichment in Jupiter's

975: atmosphere---trapping argon in icy planetesimals requires temperatures

976: of 30~K (Owen et al. 1999).

977: 

978: %Late-stage accretion of

979: %refractory CHON grains similar to those observed in Halley's comet dust

980: %by \cite{jessberger88} cannot provide such a large carbon overabundance:

981: %a mechanism of tapping into the massive CO and CH$_4$ reservoirs is

982: %required.  

983: 

984: In the next section, we describe our chemical model of the solar nebula

985: midplane.  We will use these results, along with the surface density

986: curves shown in Fig. \ref{sdplanet}, to predict the solar nebula's

987: solid mass distribution in \S \ref{planetesimal}.

988: 

989: \section{Chemical Model}

990: \label{chemmodel}

991: 

992: \subsection{Reaction Set}

993: 

994: The foundation of our chemical model is the UMIST database RATE95

995: (Millar et al. 1997).  We follow the chemistry of 211 species, 153

996: gas-phase and 58 ices.  Our reaction network contains 2479 reactions,

997: including thermal desorption, gas-grain reactions and grain surface

998: reactions.  No non-thermal desorption processes (e.g. cosmic ray

999: heating) are included.

1000: 

1001: We make the simplifying assumption that radial and vertical motions of

1002: gas and dust are slow compared to chemical reaction timescales, and do

1003: not track the motion of gas and dust through the nebula.  This

1004: assumption is true for grain surface reactions and freezeout/desorption,

1005: which have short timescales, but gas-phase chemistry proceeds more

1006: slowly.  Furthermore, inward migration of cm-size particles can deplete

1007: the outer solar nebula of ice and concentrate each species at its

1008: condensation front (e.g. Ciesla and Cuzzi 2006).  In this paper, we

1009: ignore relative motions of solid and gas.

1010: 

1011: Since protostellar disks have a high optical depth to UV radiation, we

1012: neglect photoionization in the disk midplane model.  However, we include

1013: cosmic rays at the standard interstellar flux in parts of the disk with

1014: a half-plane surface density lower than the penetration depth of

1015: 150~g~cm$^{-2}$ (Umebayashi and Nakano 1981).  The cosmic-ray

1016: penetration depth is listed in Table \ref{freechem} with the other free

1017: parameters in the chemical model.

1018: 

1019: The following sections describe our treatment of gas-grain and grain

1020: surface reactions and the binding energies we use to calculate those

1021: reaction rates.

1022: 

1023: \subsubsection{Grain Surface Reactions}

1024: 

1025: Following \cite{willacy06} and \cite{garrod06}, we use rate equations to

1026: calculate grain surface reaction rates.  Rate equations work in the

1027: mean-field approximation, neglecting the stochastic variation of

1028: abundances on different grains.  The accuracy of rate equations is

1029: compromised when a species reacts on the grain surface faster than it

1030: can adsorb or desorb: this drives the reactant abundance per grain to

1031: less than unity and the mean-field approximation is no longer valid

1032: (Caselli et al. 1998).  However, unlike stochastic methods that account

1033: for the discrete nature of the grains, solving rate equations is

1034: computationally tractable even for large-scale simulations.  For

1035: example, a model that uses direct integration of the master equation

1036: (the probability distribution governing reaction rates) has the number

1037: of equations in the simulation increasing exponentially with the number

1038: of species (Barzel and Biham 2007).

1039: 

1040: The species most likely to violate the mean-field approximation is

1041: atomic hydrogen, which can scan grain surfaces and find reaction

1042: partners quickly due to its low atomic weight.  While heavier atoms move

1043: across grain surfaces by thermal hopping, Hornekaer et al. (2003) find

1044: that H atoms often tunnel between adjacent vacancies in the grain

1045: surface lattice.  This result conflicts with previous work by Katz et

1046: al. (1999), who found that the quantum tunneling mechanism is

1047: improbable.  Although the question of whether quantum tunneling or

1048: thermal hopping is primarily responsible for hydrogen mobility on grain

1049: surfaces is not resolved, Ruffle and Herbst (2000) show that rate

1050: equation simulations limiting H motion to thermal hopping only---and

1051: thus slowing down reactions involving atomic H---show good agreement

1052: with stochastic methods.  We therefore assume thermal hopping regulates

1053: all grain surface reaction rates.

1054: 

1055: The thermal hopping rate to an adjacent vacancy on the grain surface

1056: lattice is

1057: \begin{equation}

1058: t_{\rm hop}^{-1} = \nu_0 \exp \left ( -0.3 E_D / k T_g \right ) ,

1059: \label{hop}

1060: \end{equation}

1061: where $E_D$ is the binding energy, $\nu_0$ is the oscillation frequency

1062: between the ice and the grain surface, and $T_g$ is the grain

1063: surface temperature (assumed to be in equilibrium with the gas).  The

1064: oscillation frequency is

1065: \begin{equation}

1066: \nu_0 = \sqrt{2 n_s E_D / \pi^2 m} ,

1067: \label{osc}

1068: \end{equation}

1069: where $n_s \sim 1.5 \times 10^{15} {\rm cm}^{-2}$ is the surface density

1070: of lattice vacancies (Hasegawa et al. 1992; Table \ref{freechem}) and

1071: $m$ is the reactant mass.  Typical oscillation frequencies are

1072: 1--3$\times 10^{12}$~s$^{-1}$ (Herbst et al. 2005).  Following

1073: \cite{willacy06}, we assume that only atoms can move across grain

1074: surfaces: molecules, which have higher binding energies, are stationary.

1075: All grain surface reactions thus involve at least one atom.

1076: 

1077: The grain-surface reaction rate ($s^{-1}$) is given by

1078: \begin{equation}

1079: t_{\rm reac}^{-1} = { t_{\rm hop,1}^{-1} + t_{\rm hop,2}^{-1}

1080: \over 4 \pi a^2 n_s n_g } ,

1081: \label{surfrate}

1082: \end{equation}

1083: where $t_{\rm hop,1}$ and $t_{\rm hop,2}$ are the thermal hopping rates

1084: of reactants 1 and 2, $a$ is the grain radius and $n_g$ is the number

1085: density of grains.  As in the molecular cloud model, we assume a mean

1086: grain radius of $0.258 \; \mu m$ (Table \ref{freechem}).  We take a

1087: grain number density of $n_g / (n_{\rm H} + 2 n_{\rm H2}) = 10^{-12}$

1088: (Table \ref{freechem}), consistent with our assumption of well-mixed

1089: grains and gas.

1090: 

1091: \subsubsection{Gas-Grain Reactions}

1092: 

1093: We assume that all species can freeze out on grains except He, which has

1094: a very low binding energy, $E_D / k = 100$~K (Tielens and Hagen 1982).

1095: The freeze-out rate is

1096: \begin{equation}

1097: k_f = S_x \left \langle \pi a^2 n_g \right \rangle C_i v_x n_x ,

1098: \label{freezerate}

1099: \end{equation}

1100: where $S_x$ is the sticking coefficient (assumed to be 0.3 for H atoms

1101: and 1 for all other species; Table \ref{freechem}), $v_x$ is the

1102: gas-phase velocity of species $x$, and $n_x$ is the gas-phase number

1103: density of $x$.  The factor $C_i$ takes into account the attraction of

1104: positively charged ions to grains which, on average, are negatively

1105: charged (Umebayashi and Nakano 1980):

1106: \begin{equation}

1107: \left \{ \begin{array}{l} C_i = 1 \; {\rm for \; neutrals} \\ C_i = 1 +

1108: e/(akT) \; {\rm for \; single-charged \; positive \; ions.} \end{array}

1109: \right.

1110: \label{ci}

1111: \end{equation}

1112: In Eq. \ref{ci}, $k$ is the Boltzmann constant and $e$ is the

1113: electron charge.  Once an ion hits a grain, we assume it is neutralized

1114: by an adsorbed free electron.  The thermal desorption rate ($s^{-1}$) at

1115: which ice mantle species are returned to the gas is

1116: \begin{equation}

1117: k_t = \nu_0 \exp \left ( E_D / k T_g \right ) .

1118: \label{therm}

1119: \end{equation}

1120: 

1121: %Although formation of $> 10 \mu m$ grains proceeds rapidly

1122: %in microgravity experiments (Blum et al. 2000), we assume the continued

1123: %presence of small grains and keep the mean grain radius fixed throughout

1124: %our simulation.  

1125: 

1126: \subsubsection{Binding Energies}

1127: 

1128: The binding energies $E_D$ for each species depend on the surface onto

1129: which the molecules adsorb and the mixture of ices on that surface.  If

1130: each ice were a pure solid, $E_D$ would simply be the sublimation energy

1131: and desorption physics could be calculated nonkinetically using vapor

1132: pressure.  However, molecules in mixed ices may have substantially

1133: different binding energies than those in pure ices, so the sublimation

1134: energy approximation is inadequate.

1135: 

1136: The most abundant astrophysical ice is H$_2$O, so the binding energy of

1137: each species under protostellar disk conditions is primarily determined

1138: by how it behaves on/in an amorphous water-ice matrix.

1139: \cite{collings04} group binding energies into three categories based on

1140: their behavior when co-deposited with H$_2$O: CO-like, water-like, and

1141: intermediate.  The extremely volatile CO-like species include CO, N$_2$,

1142: O$_2$, and CH$_4$.  Molecules of these species do not bond efficiently

1143: either with themselves or with water: they quickly diffuse through an

1144: amorphous H$_2$O matrix and desorb mainly at their sublimation

1145: temperatures, which are quite low ($<40$~K), with minor amounts

1146: remaining trapped in the H$_2$O ice.

1147: 

1148: Molecules of water-like species, including NH$_3$ and CH$_3$OH, bind

1149: more strongly to water than to each other or to a metal substrate.

1150: These polar molecules form hydrogen bonds with water that keep them in

1151: the solid phase at higher temperatures than possible in their pure form.

1152: The data of \cite{collings04} show that ammonia-water mixtures are

1153: substantially less volatile than pure ammonia, which has a binding

1154: energy of only 1110~K (Hasegawa and Herbst 1993).  NH$_3$ therefore takes

1155: on the binding energy of H$_2$O, 5700~K.

1156: 

1157: Finally, intermediate species are trapped by H$_2$O to some extent, but

1158: do not have the same mobility as CO inside a water-ice matrix.  Some

1159: co-desorption with H$_2$O occurs, but most desorption takes place at the

1160: sublimation energy.  These species include H$_2$S, OCS and CH$_3$CN.  We

1161: have not included the partial co-desorption with water and treat each

1162: molecule as having only one specific desorption tmeperature.  Table

1163: \ref{binding} gives the binding energy data used in our model.

1164: 

1165: \subsection{Preprocessing in Molecular Cloud}

1166: 

1167: The ices in protostellar disks have undergone significant chemical

1168: processing before giant planet formation begins.  To simulate this

1169: preprocessing, we model $10^6$ years of molecular cloud evolution to

1170: derive input abundances for the protostellar disk.  The molecular cloud

1171: model uses the same species set and reactions as our disk model, with

1172: the addition of photoionization due to the interstellar UV field.

1173: 

1174: The atoms that form the species in our simulation are H, He, C, N, O,

1175: and S.  We also include Si adsorption, desorption and ionization as a

1176: tracer of the metal population.  Except for hydrogen, we assume all

1177: material enters the molecular cloud in atomic form, as if coming from a

1178: recent supernova.  We take $n_{\rm H} / (n_{\rm H} + 2 n_{\rm H2}) =

1179: 10^{-2}$ (Table \ref{freechem}), so that 99\% of hydrogen atoms are in

1180: H$_2$.  We assume a density of $2 \times 10^4$~cm$^{-3}$, a temperature

1181: of 10~K and an extinction $A_V = 10$~mag.

1182: 

1183: Not all of the atomic CHONS material is available to form ices: some

1184: will help form the solid grains where ice mantles deposit.  To figure

1185: out the fraction of C, N, O, and S locked away in grains, we use the

1186: atomic abundance ratios in Halley's comet dust grains reported by

1187: \cite{jessberger88}: C:N:O:Si:S = 814:42:890:185:72.  Assuming only

1188: trace amounts of silicon remain in the gas phase (gas/grain $< \:

1189: 0.001$), we calculate grain fractions for C, N, O and S as

1190: \begin{equation}

1191: n_{\rm grain} / n_{\rm gas} = \left ( {N_{\rm Si} \over N_x} \right

1192: )_{\odot} \left ( {N_x \over N_{\rm Si}} \right )_{\rm comet} .

1193: \label{dustfrac}

1194: \end{equation}

1195: 

1196: Using the solar composition of \cite{helling00}, as in the dynamical

1197: disk model, we find $F_{\rm grain}$ = 0.4, 0.1, 0.2 and 0.85 for C, N, O

1198: and S respectively (Table \ref{freechem}).  This portion of each atom's

1199: inventory does not participate in chemistry and is excluded from the

1200: simulation.  The gas and ice abundances of major species after 1~Myr of

1201: molecular cloud evolution are listed in Table \ref{cloudoutput}.  These

1202: abundances form the chemical initial conditions for our protostellar

1203: disk model.

1204: 

1205: The major ice species formed in the molecular cloud are H$_2$O, NH$_3$,

1206: CO, and CH$_4$.  More than 90\% of the C, N and O atoms that aren't part

1207: of refractory grains are in these compounds.  Sulfur is sequestered in

1208: solid H$_2$S after the molecular cloud phase, but this composition is

1209: subsequently modified during disk evolution.  Minor CNO carriers are

1210: HCN, N$_2$ and NO.  Figure \ref{pies} shows the division of solid C, N

1211: and O atoms among the species in our model.

1212: 

1213: Our molecular cloud model predicts total ${\rm CO / CH_4} = 0.14$.  Both

1214: of these species reside in ice mantles.  Abundances of major ices

1215: relative to H$_2$O are listed in Table \ref{iceratios}.  Our CO/CH$_4$

1216: ice ratio is consistent with the \cite{aikawa08} models of prestellar

1217: cores, which give ${\rm CO / CH_4} = 0.43$ in an embedded core and ${\rm

1218: CO / CH_4} = 0.1$ in an isolated core.  Likewise, Lodders (2003)

1219: concluded that the reactions

1220: \begin{equation}

1221: \begin{array}{ll}

1222: 3 {\rm H_2 + CO} \rightarrow {\rm CH_4 + H_2O} \\

1223: {\rm CO} \rightarrow {\rm C} \; ({\rm graphite}) + {\rm CO_2}

1224: \end{array}

1225: \label{noco}

1226: \end{equation}

1227: would deplete CO from presolar gas (note, however, that in our model

1228: CH$_4$ forms on grains and not in the gas).  Lodders predicts ${\rm CH_4

1229: / H_2O} = 0.58$, whereas we find a methane-to-water ratio of 0.38 (Table

1230: \ref{iceratios}).  This discrepancy reflects the fact that we have a

1231: higher input oxygen abundance (to match the opacity tables calculated by

1232: Semenov [2003]).

1233: 

1234: We note that the dominant components of interstellar ice mantles are

1235: typically H$_2$O, NH$_3$, CO, CO$_2$, and CH$_3$OH (Charnley and Rodgers

1236: 2008), which suggests carbon oxide formation in the ISM should be

1237: favored and hydrocarbon formation kinetically inhibited.  One possible

1238: way to reduce the ${\rm CH_4 / CO}$ ratio is to lower the assumed atomic

1239: hydrogen abundance, which would reduce the efficiency of hydrogenated

1240: molecule formation on the grains at early times when atomic carbon is

1241: abundant.  However, the purpose of this work is to determine the ice

1242: mass available for giant planet formation. Since CH$_4$ and CO have such

1243: similar condensation temperatures (41 and 35 K, respectively) and we

1244: assume ${\rm O/C} > 1$, so CO cannot incorporate all available oxygen,

1245: ${\rm CO/CH_4}$ is not critical to our final result.

1246: 

1247: We find that the most abundant nitrogen-carrying molecule is ammonia,

1248: with ${\rm N_2 / NH_3} = 2.5 \times 10^{-3}$ (both species reside in ice

1249: mantles).  We calculate ${\rm NH_3 / H_2O} = 0.14$, which matches the

1250: input N/O ratio.

1251: %This result is consistent

1252: %with observed abundances in the multiple protostellar system IRAS 16293

1253: %(Charnley and Rodgers 2008).

1254: Lodders (2003) also predicted that hydrated ammonia was the major

1255: nitrogen reservoir in the solar nebula (see \S \ref{ammonia} for a

1256: discussion of hydrated ammonia in Jupiter and Saturn's feeding zones).

1257: However, \cite{womack92} infer ${\rm N_2 / NH_3} = 4$ in the

1258: comet-forming region of the solar nebula from models of elemental

1259: nitrogen depletion in Comet Halley.

1260: 

1261: Overall, our model molecular cloud favors the formation of hydrogenated

1262: C, N, and O compounds over oxides or diatomic molecules (except for

1263: H$_2$).  We next discuss the subsequent chemical evolution of the solar

1264: nebula.

1265: 

1266: \subsection{Chemical Model Results: Ice Lines} 

1267: \label{icelinesec}

1268: 

1269: In this section, we report the locations of important condensation

1270: fronts and the relative abundances of the most common ices.  We use the

1271: gas and ice mixture resulting from the molecular cloud model as the

1272: starting point for a radial series of midplane chemical models that span

1273: 2~Myr of disk evolution.  These models track the ice inventory of

1274: the disk midplane as a function of heliocentric distance and time.

1275: 

1276: The ices in the disk can be subdivided into two categories: (1) those

1277: that form in the molecular cloud and freeze out from the disk gas, and

1278: (2) those that form or are destroyed in the warm disk midplane.  All

1279: Category 1 molecules are highly stable both as gases and solids: in the

1280: radiation-shielded disk midplane, their reaction probabilities are very

1281: low.  The stability and ease of formation of these molecules makes them

1282: highly abundant, $> 99$\% of the solar nebula ice mass.  In Category 1

1283: are water, methane, CO, ammonia, and HCN. In Category 2 are all

1284: sulfur-carrying molecules, simple hydrocarbons, N$_2$, and NO.

1285: 

1286: \subsubsection{Category 1: Stable and Abundant}

1287: 

1288: Figure \ref{freezeabun} shows time snapshots of the abundance of

1289: Category 1 ice abundances (solid phase only) as a function of radius.

1290: Note that the ammonia and water condensation fronts coincide; we will

1291: discuss the implications of this further in \S \ref{ammonia}.  The most

1292: abundant ice is H$_2$O, followed by methane, ammonia, CO and HCN.  These

1293: five ices account for 98\% of the ice mass in regions where $T \leq

1294: 35$~K, the sublimation temperature of CO.

1295: 

1296: In Fig. \ref{freezeabun}, we see the H$_2$O condensation front sweeping

1297: from 6~AU at $t = 0$ to 1.5~AU after 2~Myr.  Observations of the

1298: asteroid belt place the H$_2$O ice line at 2.7~AU (Abe et al. 2000;

1299: Rivkin et al. 2002), a position it reaches at $t = 5 \times 10^5$~yr in

1300: our model.  The slow motion of the snow line through the inner solar

1301: system could perhaps account for the H$_2$O abundance gradient observed

1302: in the asteroid belt (Barucci et al. 1996): we find that icy grains are

1303: present for $\sim 10^5$ years longer at 3~AU than at 2~AU.  The snow

1304: line crosses Jupiter's heliocentric distance, 5.2~AU, after only $5

1305: \times 10^4$ yr, leaving ample time for the formation of Jupiter's core

1306: from icy planetesimals.

1307: 

1308: If we assume Saturn formed {\it in situ} at 9.5~AU, primordial methane

1309: accretion onto both the giant planet and its satellites is possible.  We

1310: find that grain mantles in Saturn's feeding zone should have ${\rm CH_4

1311: / H_2O} \ge 0.01$ after $2.3 \times 10^5$~yr.  CH$_4$ is 90\% solid at

1312: 9.5~AU after $5.4 \times 10^5$~yr.  We note that the C/H ratio in

1313: Saturn's atmosphere has recently been revised upward by a factor of 2

1314: (Flasar et al. 2005).

1315: 

1316: We find that if Uranus and Neptune formed outside 12~AU, their

1317: core-forming planetesimals contained both CH$_4$ and CO.  The

1318: composition of both planets strongly suggests that they accreted

1319: carbonaceous ices: ${\rm (C/H) / (C/H)}_{\odot}$ is 41 for Neptune and

1320: $\leq 260$ for Uranus (Lodders and Fegley 1994).  Furthermore, Grundy et

1321: al.  (2002) find that methane ice is widely distributed across the

1322: surface of Triton.

1323: 

1324: \subsubsection{Category 2: Chemically Active}

1325: 

1326: In Fig. \ref{freezeabun} we see that the typical behavior of an ice

1327: line is to move inward as the nebula cools.  This can happen only if the

1328: molecule is relatively inert.  Sulfur species are missing from Fig.

1329: \ref{freezeabun} because H$_2$S, the dominant sulfuric ice at $t = 0$

1330: with 99.6\% of the sulfur atoms, does not survive in the gas phase.

1331: Likewise, while C$_3$H$_n$ chains are stable and freezeout-dominated in

1332: our model, acetylene ice partially hydrogenates to form ethylene and

1333: ethane.  Gaseous NO and N$_2$ participate in ion exchange reactions in

1334: the gas, reducing the abundance available for ice formation.  In this

1335: section, we consider the sulfuric, nitric and hydrocarbon ice systems.

1336: 

1337: As soon as our midplane chemical model begins, the reaction

1338: \begin{equation}

1339: {\rm H_2S + H \rightarrow HS + H_2}

1340: \label{h2s}

1341: \end{equation}

1342: starts destroying gaseous H$_2$S where temperature is above $\sim 50$~K.

1343: The product HS then initiates a network of reactions that form OCS and

1344: H$_2$CS.  The reactions linking the main sulfur-carrying molecules are

1345: shown in Fig. \ref{sulfurnetwork}.  We do not find any depletion of

1346: H$_2$CS once formed on grains.  Figure \ref{sulfuric} shows time

1347: snapshots of the abundances (solid phase only) of the main

1348: sulfur-carrying molecules H$_2$S, OCS and H$_2$CS.  After 2~Myr, we see

1349: a banded structure where H$_2$S dominates at $R > 8$~AU, OCS is most

1350: abundant between 5 and 8~AU and H$_2$CS dominates at $R < 5$~AU.

1351: 

1352: H$_2$S remains in the outer solar nebula because the gas at $R > 9$~AU

1353: never gets hotter than the H$_2$S sublimation temperature in our model.

1354: This scenario may be physically unrealistic: infalling molecular cloud

1355: material is likely heated in an accretion shock during disk formation.

1356: However, if we examine the total abundance of icy sulfur atoms, $N({\rm

1357: H_2S}) + N({\rm OCS}) + N({\rm H_2CS})$, we see that it is conserved

1358: after OCS formation finishes at $3 \times 10^5$ yr.  From Fig.

1359: \ref{sulfuric}, we can imagine a sulfurous ice line sweeping through the

1360: solar nebula, moving from 8.5~AU at $t = 0$ to 2.5~AU after 2~Myr.

1361: 

1362: Figure \ref{sulfuric} also shows the abundances of solid NO and N$_2$.

1363: NO is chemically active during the late stages of the solar nebula, $T

1364: \geq 10^6$~yr.  Even when the solar nebula temperature drops below the

1365: nominal sublimation temperature of NO, some thermal desorption still

1366: takes place, albeit at a much-reduced rate (see Eq. \ref{therm}).

1367: For most species, re-freezing after desorption occurs faster than

1368: reactions in the gas and no ice molecules are lost.  However, NO can be

1369: destroyed in the gas by reactions with ${\rm H^+}$.  Once the half-plane

1370: surface density drops below the cosmic ray penetration depth at late

1371: times, the production rate of ${\rm H^+}$ increases by 33\% due to the

1372: reaction of ${\rm H_2}$ with cosmic ray-produced photons:

1373: \begin{equation}

1374: {\rm H_2} + \gamma_{CR} \rightarrow {\rm H^+ + H + e^-} .

1375: \label{makehplus}

1376: \end{equation}

1377: Thermally desorbed NO atoms are then efficiently removed from the gas

1378: before refreezing by the reaction sequence

1379: \begin{equation}

1380: \begin{array}{c}

1381: {\rm NO + H^+ \rightarrow NO^+ + H} \\

1382: {\rm then} \\

1383: {\rm NO^+} \stackrel{\rm freeze}{\longrightarrow} {\rm N + O} \\

1384: {\rm or} \\

1385: {\rm NO^+ + e^- \rightarrow N + O} . \\

1386: \end{array}

1387: \end{equation}

1388: 

1389: Laboratory work by \"{O}berg et al. (2005) indicates that N$_2$ is

1390: extremely volatile, with $E_D / k = 790$~K.  This puts its 

1391: sublimation temperature near the assumed ambient temperature of our

1392: protostellar disk simulation, 20~K.  The smooth shape of the N$_2$

1393: abundance curve at $t = 1$~yr reflects the midplane temperature's

1394: asymptotic approach to isothermality at 20~K.  As the disk evolves,

1395: the N$_2$ ice line moves inward, but the overall N$_2$ abundance

1396: declines due to the gas-phase reaction

1397: \begin{equation}

1398: {\rm He^+ + N_2 \rightarrow N^+ + N + He} .

1399: \label{dinitrogen}

1400: \end{equation}

1401: The ${\rm N^+}$ ion then begins a sequence of reactions with H$_2$ that

1402: produce ${\rm NH^+}$, ${\rm N_2H^+}$, ${\rm N_3H^+}$ and finally the

1403: ammonium ion ${\rm N_4H^+}$.  Ammonium recombines with an electron to

1404: form NH$_3$ + H, so the net effect is to convert gaseous N$_2$ into

1405: NH$_3$.

1406: 

1407: Our model includes the small aliphatic hydrocarbons C$_2$H$_2$,

1408: C$_2$H$_4$, C$_2$H$_6$, C$_3$H$_2$ and C$_3$H$_4$.  We do not include

1409: longer chains in our model---we merely use these five species to find

1410: the location of the ``oil line'' of icy hydrocarbons.  Lodders (2004)

1411: suggested that tarry compounds, rather than water, provided the bulk of

1412: the solid mass for Jupiter's core.  This work does not truly test that

1413: hypothesis, since Lodders' scenario requires a substantial amount of

1414: hydrocarbons that are more refractory than water (our light oils have

1415: sublimation temperatures of 55-70~K).  PAHs would be suitable tar

1416: candidates, but we do not include them in our model for lack of accurate

1417: binding energy measurements.  We plan on incorporating PAH freezeout

1418: into future work, but we note that Jupiter falls inside the water-ice

1419: line for all but the first 50,000~yr of disk evolution.

1420: 

1421: %Broadly speaking, our results imply that

1422: %unsaturated hydrocarbons are preferred over saturated ones, but

1423: %concatenation of small hydrocarbons is likely.

1424: 

1425: Figure \ref{hydrocarbon} shows the evolution of the hydrocarbon ice

1426: lines.  The most abundant hydrocarbons in our simulation are C$_3$H$_n$

1427: chains, which are at the end of our reaction sequence.  This suggests

1428: that concatenation of small carbon chains is efficient.  The small

1429: hydrocarbons in our model have similar sublimation temperatures,

1430: creating an ``oil line'' that moves from 8~AU at $t = 0$ to 3~AU after

1431: 2~Myr.

1432: 

1433: For the C$_2$H$_n$ hydrocarbons, acetylene dominates over ethylene and

1434: ethane in the trans-Saturnian region, $R > 10$~AU.  The activation

1435: barrier $E/k = 1210$~K for the reaction ${\rm C_2H_2 + H \rightarrow

1436: C_2H_3}$ (Hasegawa et al. 1992), which has to break the strong C-C

1437: triple bond, means that at the cold temperatures required for acetylene

1438: to deposit on grain surfaces, hydrogenation proceeds slowly.  Only when

1439: the disk is near the acetylene sublimation temperature ($\sim 55$~K)

1440: does a significant amount of ethane form.  Throughout most of the disk,

1441: ethylene is more abundant than ethane.  However, an ``ethylene gap''

1442: forms between 7 and 10~AU after $3 \times 10^4$~yr.  We explain the

1443: ethylene gap as follows: between $10^4$ and $10^5$~yr, 7--10~AU region

1444: takes on mass due to an outward-moving density wave (recall the

1445: discussion in \S \ref{surfacedensity}).  This surface density increase

1446: causes a temperature increase of $\sim 15$~K, increasing the rate of

1447: ethylene hydrogenation.

1448: 

1449: %A secondary acetylene gap and ethane peak forms at late times ($t >

1450: %1$~Myr) near 4~AU.  The temperature at the start of the secondary

1451: %acetylene gap formation is also $\sim 60$~K.  

1452: 

1453: %C$_2$H$_n$ hydrocarbons move toward

1454: %saturation everywhere in the disk that

1455: 

1456: \subsubsection{The Snow-Ammonia Line}

1457: \label{ammonia}

1458: 

1459: Since we have adapted lab results showing co-desorption of water and

1460: ammonia (Collings et al. 2004) for our model, we find that the ammonia

1461: and water condensation fronts coincide.  This statement may seem

1462: controversial, and indeed, further lab experiments are needed to verify

1463: that hydrogen bonding between the species is strong enough to govern

1464: ammonia desorption.  However, the composition of Jupiter, Saturn and

1465: their satellites provides strong evidence for primordial ammonia-rich

1466: planetesimals between 5 and 10 AU.

1467: 

1468: Lopes et al. (2007) examined the rheological properties of a

1469: cryovolcanic flow on Titan with Synthetic Aperture Radar imaging and

1470: found a slurry composition most consistent with an

1471: ammonia-water-methanol mixture (recall that methanol is another ice

1472: found to co-desorb with water in lab experiments).  The presence of

1473: ammonia in cryovolcanic outflows strongly suggests primordial ammonia

1474: accretion: late-stage volatile delivery by comets would provide a small

1475: surface reservoir of ammonia but not embed it in the geologically active

1476: subsurface.  Based on the low temperatures ($T < 75$~K) necessary for

1477: trapping N$_2$ in water ice, \cite{owen00} argued that Titan's massive

1478: nitrogen atmosphere originally supplied as NH$_3$, {\it unmodified from

1479: its relative abundance in the outer solar nebula.}

1480: 

1481: %Similarly, Tobie et al. (2005) examined tidal

1482: %dissipation on Titan and found that a primordial liquid water shell {\it

1483: %with a few percent of ammonia} is required to maintain Titan's high

1484: %orbital eccentricity.

1485: 

1486: Other Saturnian moons also show evidence of primordial ammonia

1487: accretion: Prentice (2007) calculates that Iapetus is 27\% ammonia by

1488: mass, assuming homologous contraction of the Saturnian subnebula and

1489: placing Enceladus just at the stability point of liquid water.  The

1490: source material for the rock/ice Iapetus cannot be solar nebula gas, nor

1491: could N$_2$ have provided the initial nitrogen budget: with 34\% rock

1492: and 34\% H$_2$O, there would have been no free hydrogen reservoir for

1493: ammonia formation.  Freeman et al. (2007) suggest that an ammonia-water

1494: ice mantle is necessary to explain the subsurface ocean driving

1495: Enceladus' south pole geyser.  Finally, Prentice (2006) finds that

1496: the lack of compressional features on the surface of Rhea, which are

1497: expected to result from phase II of H$_2$O crystalline ice, favors an

1498: ammonia-rich satellite composition (25\% NH$_3$ by mass), since ammonia

1499: inhibits phase II H$_2$O ice formation.

1500: 

1501: In the Jovian system, conducting subsurface oceans are required to

1502: explain the magnetic fields of Europa, Ganymede, and Callisto.  One way

1503: to lower water's melting point and maintain liquid oceans is to mix in

1504: ammonia (Hussmann et al. 2006).  Spohn and Schubert (2003) use

1505: equilibrium heat transfer models to confirm the likelihood of subsurface

1506: ammonia-water oceans on all Galilean satellites except Io.

1507: 

1508: Perhaps the best piece of evidence for ammonia-containing planetesimals

1509: between 5 and 10 AU is the nitrogen enrichment in Jupiter and Saturn's

1510: atmospheres.  ${\rm (N/H) / (N/H)}_{\odot} = 3.3$ for Jupiter and 2--4

1511: for Saturn (Owen and Encrenaz 2003).  Such a large enrichment suggests a

1512: non-gaseous origin for nitrogen in both planets.  Trapping N$_2$ or

1513: NH$_3$ in clathrate hydrates is possible, but \cite{hersant04} find that

1514: the ammonia clathrate cannot form until 0.9~Myr after solar nebula

1515: formation: the ammonia clathrate is more volatile than a co-deposited

1516: ammonia-water ice mixture.  Based on the upward revision of Saturn's C/H

1517: ratio, Hersant et al. (2008) suggest that ammonia was present as a

1518: hydrate, rather than a clathrate, in Saturn's feeding zone.  {\it We

1519: recommend that the canonical ``snow line'' be reconceived as a water

1520: ice-ammonia line, and that NH$_3$ be added to the volatile inventory of

1521: Jupiter- and Saturn-forming planetesimals.}

1522: 

1523: %We note, however, that self-shielding of $^16$CO and preferential

1524: %photodestruction of heavier isotopes altered the isotopic composition of

1525: %refractory material throughout the solar system (Lyons and Young 2005).

1526: %Even though the Category 1 ices can be considered inert for the purpose

1527: %of predicting bulk planetesimal composition, they do react in the upper

1528: %layers of the disk and reaction products subsequently diffuse to the

1529: %midplane.

1530: 

1531: \section{Solid Surface Density}

1532: \label{planetesimal}

1533: 

1534: Knowing the solid-phase abundance of each ice species and the location

1535: of its condensation front, we can now calculate the solid surface

1536: density available for planet formation everywhere in the disk.  We first

1537: calculate the gas/solid mass ratio

1538: \begin{equation}

1539: G / S = {\sum_g \mu_g n_g \over \sum_s \mu_s n_s} ,

1540: \label{gsr}

1541: \end{equation}

1542: where $\mu$ is the mean molecular weight of a species, $n$ is its

1543: abundance, and the subscripts $g$ and $s$ denote gas and solid species,

1544: respectively.  We include Na, Mg, Si, Fe, Ni, Al and Ca in the solid

1545: inventory in solar proportions and assume these atoms form refractory

1546: compounds.  Finally, we calculate the solid surface density as

1547: \begin{equation}

1548: \Sigma_{\rm solid} = {\Sigma_{\rm total} \over 1 + G / S} .

1549: \label{sigmasolid}

1550: \end{equation}

1551: 

1552: Figure \ref{solids} shows the evolution of the solid surface density

1553: distribution in the solar nebula.  Note that as the disk evolves, the

1554: solids should grow and decouple from the gas, whereas these solid

1555: surface density calculations assume grains and gas are well mixed.  We

1556: see the large bump caused by the snow-ammonia line beginning at 6~AU and

1557: moving inward to 1.5~AU after 2~Myr.  In addition, methane adsorption

1558: creates a second local maximum in the solid surface density profile at

1559: 11~AU, which moves inward to 6.5~AU by the end of the simulation.

1560: 

1561: Panels 1 and 2 in Fig. \ref{solids}, $t = 1$ and $3 \times 10^4$~yr,

1562: show the effect of the decretion flow on the solid surface density

1563: profile.  The solid mass between 6 and 14~AU actually increases during

1564: the early stages of disk evolution, at the expense of the inner 4~AU.

1565: The next panel, $10^5$~yr, shows the further solid surface density

1566: increase between 12 and 18~AU.  Disk annuli that increase in surface

1567: density also heat up, and the methane bump consequently moves outward

1568: for the first $10^5$~yr.

1569: 

1570: \cite{pollack96} presented core accretion models of Jupiter, Saturn, and

1571: Uranus, beginning with solid surface densities of 10, 3, and

1572: 0.75~g~cm$^{-2}$.  While these simulations showed the plausibility of

1573: giant planet formation by core accretion, the planet formation

1574: timescales calculated were too long: Jupiter, Saturn and Uranus took 8,

1575: 9.6, and 16~Myr to form, respectively.  Since runaway solid core growth

1576: rate is directly proportional to surface density (Eq.

1577: \ref{planetgrowth}), we see that having a high surface density speeds up

1578: giant planet formation.

1579: 

1580: If one assumes that nearly 100\% of the available solids were

1581: incorporated into planetesimals, our results give starting solid surface

1582: densities that are far more conducive to giant planet formation than

1583: those from \cite{pollack96}.  (The $\Sigma_{\rm solid}$ of

1584: \cite{pollack96} refers specifically to 100-km planetesimals, not

1585: smaller bodies or grains.)  Table \ref{sdtable} lists solid surface

1586: density as a function of radius and time in the solar nebula.  Note that

1587: these results assume a substantial population of small grains is always

1588: available for ice deposition.

1589: 

1590: We consider $1.5 \times 10^5$~yr the extreme upper limit of the

1591: planetesimal formation timescale, which Hubbard and Blackman [2006]

1592: predict to be $\le 5 \times 10^4$~yr.  However, we continue the disk

1593: simulation out to 2~Myr in order to calculate nebular gas temperature,

1594: density and composition, which are necessary for models of giant planet

1595: formation that build on this work (see, for example, the new Saturn

1596: formation simulation by Dodson-Robinson et al. [2008]).  At Saturn's

1597: heliocentric distance, 9.5~AU, the solid surface density never drops

1598: below 8.7~g~cm$^{-2}$ during the first $1.5 \times 10^5$~yr of disk

1599: evolution.  This large solid surface density should provide more than a

1600: factor of 3 speedup in Saturn's solid growth phase over the Pollack et

1601: al. results.

1602: 

1603: For Uranus, the relative increase in solids is even more pronounced: at

1604: 20~AU, the minimum solid surface density during the first $1.5 \times

1605: 10^5$~yr is 3.3~g~cm$^{-2}$, for a factor of 4 speedup in runaway core

1606: growth.  Moving Uranus' starting position inward to 16~AU, in accordance

1607: with the Nice model of planet migration (Tsiganis et al. 2005), provides

1608: a starting $\Sigma_{\rm solid}$ of not less than 4.9~g~cm$^{-2}$, in

1609: addition to the speedup given by a higher angular-frequency orbit.  The

1610: prospects for Neptune to form rapidly at $\sim 12$~AU, as predicted by

1611: the Nice model, are quite good: $\Sigma_{\rm solid} \ge

1612: 7.1$~g~cm$^{-2}$.

1613: 

1614: \cite{hubickyj05} provided an update to the Pollack et al. simulations

1615: of Jupiter's growth, this time assuming rapid grain settling and

1616: sublimation would lower opacities of the gaseous protoplanetary envelope

1617: and speed up its contraction.  (Even if we assume low-opacity envelopes

1618: for Saturn and Uranus, they still require high solid surface densities

1619: to form quickly, as their solid growth stages both take $\ge 1.5$~Myr in

1620: the Pollack et al. models.)  Hubickyj et al. found that, given a

1621: planetesimal surface density of 10~g~cm$^{-2}$, Jupiter could reach its

1622: present mass in 2.3~Myr.  Our work shows that this starting condition

1623: for Jupiter is not only viable but likely: we predict solid surface

1624: densities not less than 8.2~g~cm$^{-2}$ with refractory material only

1625: and as high as 14.5~g~cm$^{-2}$ after the ice line moves through at $5

1626: \times 10^4$~yr.

1627: 

1628: \section{Conclusions}

1629: \label{conclusions}

1630: 

1631: By combining a viscously evolving protostellar disk model with a kinetic

1632: model of ice formation, we have calculated the time-evolving solid

1633: surface density available for giant planet formation.  We find three

1634: results that are highly favorable for gas- and ice-giant formation:

1635: 

1636: \begin{enumerate}

1637: 

1638: \item The total (gas+solid) surface density distribution evolves toward

1639: uniformity, with an outward-moving wave depositing mass in the giant

1640: planet-forming region between 5 and 20~AU,

1641: 

1642: \item The ammonia and water ice lines coincide, providing a mass

1643: enhancement of 7\% at the snow line over pure water condensation, and

1644: 

1645: \item Methane and CO condensation beyond 12~AU add mass to Uranus and

1646: potentially Neptune's feeding zones.

1647: 

1648: \end{enumerate}

1649: 

1650: The next step is to use a core-accretion model to calculate how fast

1651: Saturn, Uranus and Neptune can form with our enhanced surface densities.

1652: The same methods used by \cite{hubickyj05} should work for Saturn.

1653: However, an updated method of tracking planetesimal scattering may be

1654: required for feeding zones with $R > 15$~AU, where orbital velocities of

1655: planetesimals decrease to $\sim 10$--50 times their escape velocities

1656: and self-stirring begins to affect the dynamics of the planetesimal

1657: disk.  By calculating solid surface density as a function of radius and

1658: time self-consistently with viscous evolution of the protostellar disk,

1659: instead of assuming a single gas/solid ratio covers the entire giant

1660: planet-forming region of the solar nebula, the research presented here

1661: allows core accretion simulations to be fully deterministic.

1662: 

1663: One area that requires more work is the viscous evolution of an

1664: MRI-turbulent disk with a dead zone, where $\alpha$ varies with radius.

1665: Annuli with an inactive midplane would have a lower column-averaged

1666: value of $\alpha$, which could slow or stop the movement of mass from

1667: the inner solar nebula to the giant planet-forming region.  However,

1668: Turner and Sano (2008) find that a toroidal magnetic field component

1669: produced in the midplane by shear drives a laminar accretion flow even

1670: through the nonturbulent dead zone.

1671: 

1672: Since our focus is on increasing the solid mass available for giant

1673: planet cores, we have neglected clathrate hydrates in our chemical

1674: model, preferring to tap into the large mass reservoir provided by ices.

1675: However, noble gas enrichment in the giant planet atmospheres suggests

1676: that clathration is very important, if not for building planet cores,

1677: then at least for determining atmospheric composition during the late

1678: stages of accretion (Hersant et al. 2004).  The presence of PAHs in

1679: carbonaceous meteorites, cometary dust and Iapetus implies that these,

1680: too, might be an important solid mass reservoir.  It is not yet known

1681: whether PAHs survive in the ISM or form primarily in the warm disk.

1682: Woods and Willacy (2007) present a mechanism for benzene formation in

1683: the inner 3~AU of the solar nebula.  We are investigating the

1684: possibility of including PAH chemistry in future work.

1685: 

1686: Finally, we note that the ALMA Design Reference Science Plan contains a

1687: proposal for detecting the snow line in nearby protoplanetary disks.  We

1688: submit that the CO ice line would be an equally valuable detection,

1689: since this work strongly suggests that CO ice should be part of the

1690: solid inventory of the Uranus- and Neptune-building planetesimals.  In a

1691: 1 Myr-old solar nebula analog in Taurus (140 pc), the CO ice line would

1692: lie 6 resolution units from the star when observing the CO ($J = 3

1693: \rightarrow 2$) transition.

1694: 

1695: We thank Julie Moses, Mark Marley and Doug Lin for helpful discussions

1696: about the project design.  S.D.R. was supported by grants from the NSF

1697: Graduate Research Fellowship program and the Achievement Rewards for

1698: College Scientists Foundation.  K.W. and N.T. were supported by the JPL

1699: Research and Technology Development Program.  P.B. received funding from

1700: NSF Grant AST-0507424 and NASA Origins Grant NNX08AH82G.

1701: % G.L. was

1702: %supported by NSF Career Grant AST-0449986 and NASA Planetary Geology and

1703: %Geophysics Program Grant NNG04GK19G.

1704: 

1705: \begin{thebibliography}{}

1706: 

1707: \bibitem[Abe et al.(2000)]{abe00} Abe, Y., Ohtani, E., Okuchi, T.,

1708: Righter, K., Draker, M.\ 2000.\ Water in the Early Earth.\ Origin of the

1709: earth and moon, edited by R.M.~Canup and K.~Righter and 69 collaborating

1710: authors. Tucson: University of Arizona Press, p. 413-433

1711: 

1712: \bibitem[A'Hearn(2008)]{ahearn08} A'Hearn, M.~F.\ 2008.\ Deep Impact and

1713: the Origin and Evolution of Cometary Nuclei.\ Space Science Reviews 56.

1714: 

1715: \bibitem[Aikawa et al.(1996)]{aikawa96} Aikawa, Y., Miyama, S.~M.,

1716: Nakano, T., Umebayashi, T.\ 1996.\ Evolution of Molecular Abundance in

1717: Gaseous Disks around Young Stars: Depletion of CO Molecules.\

1718: Astrophysical Journal 467, 684.

1719: 

1720: \bibitem[Aikawa et al.(2008)]{aikawa08} Aikawa, Y., Wakelam, V., Garrod,

1721: R.~T., Herbst, E.\ 2008.\ Molecular Evolution and Star Formation: From

1722: Prestellar Cores to Protostellar Cores.\ Astrophysical Journal 674, 984.

1723: 

1724: \bibitem[Alexander et al.(2006)]{alexander06} Alexander, R.~D., Clarke,

1725: C.~J., Pringle, J.~E.\ 2006.\ Photoevaporation of protoplanetary

1726: discs - II. Evolutionary models and observable properties.\ Monthly

1727: Notices of the Royal Astronomical Society 369, 229-239.

1728: 

1729: \bibitem[Allamandola and Sandford(1994)]{allamandola94} Allamandola,

1730: L.~J., Sandford, S.~A.\ 1994.\ The First Symposium on the Infrared

1731: Cirrus and Diffuse Interstellar Clouds, 58, 302

1732: 

1733: \bibitem[Allen and Robinson(1997)]{allen77} Allen, M., Robinson, G.~W.

1734: 1977. The molecular composition of dense interstellar clouds.\

1735: Astrophysical Journal 212, 396-415.

1736: 

1737: \bibitem[Anders and Grevesse(1989)]{anders89} Anders, E., Grevesse, N.

1738: 1989.\ Abundances of the elements - Meteoritic and solar.\ Geochimica et

1739: Cosmochimica Acta 53, 197-214

1740: 

1741: \bibitem[Balbus and Hawley(1991)]{balbus91} Balbus, S.~A., Hawley,

1742: J.~F. 1991.\ A powerful local shear instability in weakly magnetized

1743: disks. I - Linear analysis. II - Nonlinear evolution.\ Astrophysical

1744: Journal 376, 214-233.

1745: 

1746: \bibitem[Barucci et al.(1996)]{barucci96} Barucci, M.~A., Fulchignoni,

1747: M., Lazzarin, M.\ 1996.\ Water ice in primitive asteroids? Planetary and

1748: Space Science 44, 1047-1049.

1749: 

1750: \bibitem[Barzel and Biham(2007)]{barzel07} Barzel, B., Biham, O. 2007.\

1751: Efficient stochastic simulations of complex reaction networks on

1752: surfaces.\ Journal of Chemical Physics 127, 4703.

1753: 

1754: \bibitem[Bell et al.(1997)]{bell97} Bell, K. R., Cassen, P. M., Klahr,

1755: H. H., Henning, Th. 1997.\ The Structure and Appearance of Protostellar

1756: Accretion Disks: Limits on Disk Flaring.\ Astrophysical Journal 486,

1757: 372.

1758: 

1759: %\bibitem[Blum et al.(2000)]{blum00} Blum, J., et al. 2006, PhRvL, 85,

1760: %2426

1761: 

1762: \bibitem[Boss(2005)]{boss05} Boss, A.~P. 2005.\ Evolution of the Solar

1763: Nebula. VII. Formation and Survival of Protoplanets Formed by Disk

1764: Instability.\ Astrophysical Journal 629, 535-548.

1765: 

1766: %\bibitem[Botta et al.(2005)]{botta05} Botta, O., Bada, J. L., and

1767: %Ehrenfreund, P. 2002, Asteroids, Comets, and Meteors: ACM 2002, 500, 925

1768: 

1769: \bibitem[Caselli et al.(1998)]{caselli98} Caselli, P., Hasegawa, T.~I.,

1770: Herbst, E. 1998.\ A Proposed Modification of the Rate Equations for

1771: Reactions on Grain Surfaces.\ Astrophysical Journal 495, 309.

1772: 

1773: \bibitem[Cazaux and Tielens(2002)]{cazaux02} Cazaux, S., Tielens,

1774: A.~G.~G.~M. 2002. Molecular Hydrogen Formation in the Interstellar

1775: Medium. Astrophysical Journal 575, L29-L32.

1776: 

1777: \bibitem[Chiang and Goldreich(1997)]{chiang97} Chiang, E.~I., Goldreich,

1778: P.\ 1997.\ Spectral Energy Distributions of T Tauri Stars with Passive

1779: Circumstellar Disks.\ Astrophysical Journal 490, 368.

1780: 

1781: \bibitem[Ciesla(2007)]{ciesla07} Ciesla, F.~J.\ 2007.\ Outward Transport

1782: of High-Temperature Materials Arount the Midplane of the Solar Nebula.\

1783: Science 318, 613.

1784: 

1785: \bibitem[Ciesla and Cuzzi(2006)]{ciesla06} Ciesla, F.~J., Cuzzi, J.~N.\

1786: 2006.\ The evolution of the water distribution in a viscous

1787: protoplanetary disk.\ Icarus 181, 178-204.

1788: 

1789: \bibitem[Charnley and Rodgers(2008)]{charnley08} Charnley, S.~B., 

1790: Rodgers, S.~D.\ 2008.\ Interstellar Reservoirs of Cometary Matter.\

1791: Space Science Reviews 40.

1792: 

1793: \bibitem[Collings et al.(2004)]{collings04} Collings, M.~P., Anderson,

1794: M.~A., Chen, R., Dever, J.~W., Viti, S., Williams, D.~A., McCoustra,

1795: M.~R.~S.\ 2004.\ A laboratory survey of the thermal desorption of

1796: astrophysically relevant molecules.\ Monthly Notices of the Royal

1797: Astronomical Society 354, 1133-1140.

1798: 

1799: \bibitem[D'Alessio et al.(2006)]{dalessio06} D'Alessio, P., Calvet, N.,

1800: Hartmann, L., Franco-Hern\'{a}ndez, R., Serv\'{i}n, H.\ 2006.\ Effects

1801: of Dust Growth and Settling in T Tauri Disks. Astrophysical Journal

1802: 638, 314-335.

1803: 

1804: \bibitem[Desch(2007)]{desch07} Desch, S.~J.\ 2007.\ Mass Distribution

1805: and Planet Formation in the Solar Nebula. Astrophysical Journal 671,

1806: 878-893.

1807: 

1808: \bibitem[Dodson-Robinson et al.(2008)]{dodsonrobinson08}

1809: Dodson-Robinson, S.~E., Bodenheimer, P., Laughlin, G., Willacy, K.,

1810: Turner, N.~J., Beichman, C.~A.\ 2008.\ Saturn Forms by Core Accretion in

1811: 3.4~Myr.\ ArXiv e-prints arXiv:0810.0228 (accepted for publication in

1812: Astrophysical Journal Letters).

1813: 

1814: \bibitem[Dullemond et al.(2006)]{dullemond06} Dullemond, C.~P., Apai,

1815: D., Walch, S.\ 2006.\ Crystalline Silicates as a Probe of Disk Formation

1816: History.\ Astrophysical Journal 640, L67-L70.

1817: 

1818: \bibitem[Ferguson et al.(2005)]{ferguson05} Ferguson, J.~W., Alexander,

1819: D.~R., Allard, F., Barman, T., Bodnarik, J.~G., Hauschildt, P.~T.,

1820: Heffner-Wong, A., Tamanai, A.\ 2005.\ Low-Temperature Opacities.

1821: Astrophysical Journal 623, 585-596.

1822: 

1823: %\bibitem[Fischer and Valenti(2005)]{fischer05} Fischer, D.~A., and

1824: %Valenti, J. 2005, \apj, 622, 1102

1825: 

1826: \bibitem[Flasar et al.(2005)]{flasar05} Flasar, F.~M., and 45

1827: colleagues. 2005.\ Temperatures, Winds, and Composition in the Saturnian

1828: System.\ Science 307, 1247-1251.

1829: 

1830: \bibitem[Fraser et al.(2001)]{fraser01} Fraser, H.~J., Collings, M.~P.,

1831: McCoustra, M.~R.~S., Williams, D.~A.\ 2001.\ Thermal desorption of water

1832: ice in the interstellar medium. Monthly Notices of the Royal

1833: Astronomical Society 327, 1165-1172

1834: 

1835: \bibitem[Freeman et al.(2007)]{freeman07} Freeman, J., Stegman, D.,

1836: May, D.\ 2007.\ The Role of Ammonia in the Evolution of Enceladus.\ AGU

1837: Fall Meeting Abstracts 7.

1838: 

1839: \bibitem[Garrod and Herbst(2006)]{garrod06} Garrod, R. T., Herbst, E.\

1840: 2006.\ Formation of methyl formate and other organic species in the

1841: warm-up phase of hot molecular cores. Astronomy and Astrophysics 457,

1842: 927-936.

1843: 

1844: \bibitem[Grundy et al.(2002)]{grundy02} Grundy, W.~M., Buie, M.~W., 

1845: Spencer, J.~R.\ 2003.\ Spectroscopy of Pluto and Triton at 3-4 Microns:

1846: Possible Evidence for Wide Distribution of Nonvolatile Solids.

1847: Astronomical Journal 124, 2273-2278.

1848: 

1849: \bibitem[Haisch et al.(2001)]{haisch01} Haisch, K.~E., Jr., Lada, E.~A.,

1850: Lada, C.~J.\ 2001.\ Disk Frequencies and Lifetimes in Young Clusters.\

1851: Astrophysical Journal 553, L153-L156.

1852: 

1853: \bibitem[Hallenbeck et al.(2000)]{hallenbeck00} Hallenbeck, S.~L., Nuth,

1854: J.~A. III, Nelson, R.~N.\ 2000.\ Evolving Optical Properties of

1855: Annealing Silicate Grains: From Amorphous Condensate to Crystalline

1856: Mineral.\ Astrophysical Journal 535, 247-255.

1857: 

1858: \bibitem[Harker and Desch(2002)]{harker02} Harker, D.~E., Desch, S.~J.\

1859: 2002.\ Annealing of Silicate Dust by Nebular Shocks at 10~AU.

1860: Astrophysical Journal 565, L109-L112.

1861: 

1862: \bibitem[Hasegawa and Herbst(1993)]{hasegawa93} Hasegawa, T.~I., 

1863: Herbst, E.\ 1993.\ New gas-grain chemical models of quiescent dense

1864: interstellar clouds - The effects of H2 tunnelling reactions and cosmic

1865: ray induced desorption. Monthly Notices of the Royal Astronomical

1866: Society 261, 83-102

1867: 

1868: \bibitem[Hasegawa et al.(1992)]{hasegawa92} Hasegawa, T.~I., Herbst, E.,

1869: Leung, C.~M.\ 1992.\ Models of gas-grain chemistry in dense interstellar

1870: clouds with complex organic molecules.\ Astrophysical Journal Supplement

1871: Series 82, 167-195.

1872: 

1873: \bibitem[Hawley and Stone(1998)]{hawley98} Hawley, J.~F., Stone, J.~M.\

1874: 1998.\ Nonlinear Evolution of the Magnetorotational Instability in

1875: Ion-Neutral Disks.\ Astrophysical Journal 501, 758.

1876: 

1877: \bibitem[Hawley et al.(1999)]{hawley99} Hawley, J.~F., Balbus, S.~A., 

1878: Winters, W.~F.\ 1999.\ Local Hydrodynamic Stability of Accretion Disks.\

1879: Astrophysical Journal 518, 394-404.

1880: 

1881: \bibitem[Hayashi (1981)]{hayashi81} Hayashi, C.\ 1981.\ Structure of the

1882: Solar Nebula, Growth and Decay of Magnetic Fields and Effects of

1883: Magnetic and Turbulent Viscosities on the Nebula.\ Progress of

1884: Theoretical Physics Supplement 70, 35-53.

1885: 

1886: \bibitem[Helling et al.(2000)]{helling00} Helling, Ch., Winters, J.~M.,

1887: Sedlmayr, E.\ 2000.\ Circumstellar dust shells around long-period

1888: variables. VII. The role of molecular opacities.\ Astronomy and

1889: Astrophysics 358, 651-664.

1890: 

1891: \bibitem[Herbst et al.(2005)]{herbst05} Herbst, E., Chang, Q., Cuppen,

1892: H.~M.\ 2005.\ Chemistry on interstellar grains.\ Journal of Physics

1893: Conference Series, 6, 18-35.

1894: 

1895: \bibitem[Hersant et al.(2001)]{hersant01} Hersant, F., Gautier, D., 

1896: Hur\'{e}, J.-M.\ 2001.\ A Two-dimensional Model for the Primordial

1897: Nebula Constrained by D/H Measurements in the Solar System: Implications

1898: for the Formation of Giant Planets.\ Astrophysical Journal 554, 391-407.

1899: 

1900: \bibitem[Hersant et al.(2004)]{hersant04} Hersant, F., Gautier, D., 

1901: Lunine, J.~I.\ 2004.\ Enrichment in volatiles in the giant planets of

1902: the Solar System.\ Planetary and Space Science 52, 623-641.

1903: 

1904: %\bibitem[Hersant et al.(2008)]{hersant08} Hersant, F., Gautier, D.,

1905: %Tobie, G., Lunine, J.\ 2008.\ Interpretation of the carbon abundance in

1906: %Saturn measured by Cassini.\ Planetary and Space Science 56, 1103-1111.

1907: 

1908: \bibitem[Hornekaer et al.(2003)]{hornekaer03} Hornekaer, L., Baurichter,

1909: A., Petrunin, V.~V., Field, D., Luntz, A.~C.\ 2003.\ Importance of

1910: Surface Morphology in Interstellar H$_2$ Formation.\ Science 302,

1911: 1943-1946.

1912: 

1913: \bibitem[Hubbard and Blackman(2006)]{hubbard06} Hubbard, A., Blackman,

1914: E.~G.\ 2006.\ Planetesimal growth in turbulent discs before the onset of

1915: gravitational instability.\ New Astronomy 12, 246-263.

1916: 

1917: \bibitem[Hubickyj et al.(2005)]{hubickyj05} Hubickyj, O., Bodenheimer,

1918: P., Lissauer, J.~J.\ 2005.\ Accretion of the gaseous envelope of Jupiter

1919: around a 5-10 Earth-mass core.\ Icarus 179, 415-431.

1920: 

1921: \bibitem[Hussmann et al.(2006)]{hussmann06} Hussmann, H., Sohl, F., 

1922: Spohn, T.\ 2006.\ Subsurface oceans and deep interiors of medium-sized

1923: outer planet satellites and large trans-neptunian objects. Icarus 185,

1924: 258-273.

1925: 

1926: \bibitem[Ilgner and Nelson(2006)]{ilgner06} Ilgner, M., Nelson, R.~P.\

1927: 2006.\ On the ionisation fraction in protoplanetary disks. II. The

1928: effect of turbulent mixing on gas-phase chemistry.\ Astronomy and

1929: Astrophysics 445, 223-232.

1930: 

1931: \bibitem[Jessberger et al.(1988)]{jessberger88} Jessberger, E. K.,

1932: Christoforidis, A., Kissel, J.\ 1988.\ Aspects of the major element

1933: composition of Halley's dust.\ Nature 332, 691-695.

1934: 

1935: \bibitem[Keller et al.(2006)]{keller06} Keller, L.~P., and 32

1936: colleagues. 2006.\ Infrared Spectroscopy of Comet 81P/Wild 2 Samples

1937: Returned by Stardust.\ Science 314, 1728.

1938: 

1939: \bibitem[Kippenhahn and Weigert(1994)]{kippenhahn94} Kippenhahn, R.,

1940: Weigert, A.\ 1994, Stellar Structure and Evolution (Berlin:

1941: Springer-Verlag)

1942: 

1943: \bibitem[Kretke and Lin(2007)]{kretke07} Kretke, K.~A.,  Lin, D.~N.~C.\

1944: 2007.\ Grain Retention and Formation of Planetesimals near the Snow Line

1945: in MRI-driven Turbulent Protoplanetary Disks.\ Astrophysical Journal

1946: 664, L55-L58.

1947: 

1948: \bibitem[Lodders(2003)]{lodders03} Lodders, K.\ 2003.\ Solar System

1949: Abundances and Condensation Temperatures of the Elements.\ Astrophysical

1950: Journal 591, 1220-1247.

1951: 

1952: \bibitem[Lodders(2004)]{lodders04} Lodders, K.\ 2004.\ Jupiter Formed

1953: with More Tar than Ice.\ Astrophysical Journal 611, 587-597.

1954: 

1955: \bibitem[Lodders and Fegley(1994)]{lodders94} Lodders, K., Fegley, B.\

1956: 1994.\ The origin of carbon monoxide in Neptune's atmosphere.\ Icarus,

1957: 112, 368-375.

1958: 

1959: \bibitem[Lopes et al.(2007)]{lopes07} Lopes, R.~M.~C., and 43

1960: colleagues 2007.\ Cryovolcanic features on Titan's surface as revealed

1961: by the Cassini Titan Radar Mapper.\ Icarus 186, 395-412.

1962: 

1963: \bibitem[Lynden-Bell and Pringle(1974)]{lyndenbell74} Lynden-Bell, D.,

1964: Pringle, J.~E.\ 1974.\ The evolution of viscous discs and the origin of

1965: the nebular variables.\ Monthly Notices of the Royal Astronomical

1966: Society 168, 603-637.

1967: 

1968: %\bibitem[Lyons and Young(2005)]{lyons05} Lyons, J. R., and Young, E. D.

1969: %2005, \nat, 435, 317

1970: 

1971: \bibitem[Lyra et al.(2008)]{lyra08} Lyra, W., Johansen, A., Klahr, H.,

1972: and Piskunov, N.\ 2008.\ Global magnetohydrodynamical models of

1973: turbulence in protoplanetary disks. I. A cylindrical potential on a

1974: Cartesian grid and transport of solids.\ Astronomy and Astrophysics 479,

1975: 883-901.

1976: 

1977: \bibitem[Mihalas (1978)]{mihalas78} Mihalas, D. 1978, Stellar

1978: Atmospheres (San Francisco: Freeman)

1979: 

1980: \bibitem[Millar et al.(1997)]{millar97} Millar, T.~J., Farquhar, P.~R.

1981: A., Willacy, K.\ 1997.\ The UMIST Database for Astrochemistry 1995.\

1982: Astronomy and Astrophysics Supplement Series 121, 139-185.

1983: 

1984: \bibitem[Milsom et al.(1994)]{milsom94} Milsom, J. A., Chen, X., 

1985: Taam, R.\ 1994.\ The vertical structure and stability of accretion disks

1986: surrounding black holes and neutron stars.\ Astrophysical Journal 421,

1987: 668-676.

1988: 

1989: \bibitem[\"{O}berg et al.(2005)]{oberg05} \"{O}berg, K.~I., van

1990: Broekhuizen, F., Fraser, H.~J., Bisschop, S.~E., van Dishoeck, E.~F., 

1991: Schlemmer, S.\ 2005.\ Competition between CO and N$_2$ Desorption from

1992: Interstellar Ices.\ Astrophysical Journal 631, L33-L36.

1993: 

1994: \bibitem[Owen(2000)]{owen00} Owen, T.\ 2000.\ On the origin of Titan's

1995: atmosphere. Planetary and Space Science 48, 747-752.

1996: 

1997: \bibitem[Owen and Encrenaz(2003)]{owen03} Owen, T., Encrenaz, T.\ 2003.\

1998: Element Abundances and Isotope Ratios in the Giant Planets and Titan.\

1999: Space Science Reviews 106, 121-138.

2000: 

2001: \bibitem[Podolak(2003)]{podolak03} Podolak, M.\ 2003.\ The contribution

2002: of small grains to the opacity of protoplanetary atmospheres.\ Icarus

2003: 165, 428-437.

2004: 

2005: \bibitem[Pollack et al.(1996)]{pollack96} Pollack, J.~B., Hubickyj, O.,

2006: Bodenheimer, P., Lissauer, J.~J., Podolak, M., Greenzweig, Y.\ 1996.\

2007: Formation of the Giant Planets by Concurrent Accretion of Solids and

2008: Gas.\ Icarus 124, 62-85.

2009: 

2010: \bibitem[Prentice(2007)]{prentice07} Prentice, A.~J.\ 2007.\ Iapetus: a

2011: Prediction for Bulk Chemical Composition, Internal Physical Structure

2012: and Origin.\ AGU Fall Meeting Abstracts 1432.

2013: 

2014: \bibitem[Pringle(1981)]{pringle81} Pringle, J.~E.\ 1981.\ Accretion

2015: discs in astrophysics.\ Annual Review of Astronomy and Astrophysics 19,

2016: 137-162.

2017: 

2018: \bibitem[Press et al.(1992)]{press92} Press, W. H., Teukolsky, S. A.,

2019: Vetterling, W. T., Flannery, B. P. 1992, Numerical Recipes in Fortran

2020: 77, Second Edition (Cambridge: Cambridge University Press)

2021: 

2022: \bibitem[Reyes-Ruiz et al.(2003)]{reyesruiz03} Reyes-Ruiz, M.,

2023: P\'{e}rez-Tijerina, E., S\'{a}nchez-Salcedo, F.~J.\ 2003.\ The

2024: Magneto-Rotational Instability in Protoplanetary Disks.\ Revista

2025: Mexicana de Astronomia y Astrofisica Conference Series 18, 92-96.

2026: 

2027: %\bibitem[Richter et al.(2006)]{richter06} Richter, F.~M., Mendybaev, R.~

2028: %A., Davis, A.~M.\ 2006.\ Conditions in the protoplanetary disk as seen

2029: %by the type B CAIs.\ Meteoritics and Planetary Science, 41, 83-93.

2030: 

2031: \bibitem[Rivkin et al.(2000)]{rivkin00} Rivkin, A.~S., Howell, E.~S.,

2032: Vilas, F., Lebofsky, L.~A.\ 2002.\ Hydrated Minerals on Asteroids: The

2033: Astronomical Record.\ Asteroids III 235-253.

2034: 

2035: %\bibitem[Robinson et al.(2006)]{robinson06} Robinson, S. E., Laughlin,

2036: %G., Bodenheimer, P., and Fischer, D. 2006, \apj, 643, 484

2037: 

2038: \bibitem[Ruffle and Herbst(2000)]{ruffle00} Ruffle, D.~P., Herbst, E.\

2039: 2000.\ New models of interstellar gas-grain chemistry - I. Surface

2040: diffusion rates.\ Monthly Notices of the Royal Astronomical Society 319,

2041: 837-850.

2042: 

2043: \bibitem[Ruden and Lin(1986)]{ruden86} Ruden, S.~P., Lin, D.~N.~C.\

2044: 1986.\ The global evolution of the primordial solar nebula.\

2045: Astrophysical Journal 308, 883-901.

2046: 

2047: \bibitem[Ruden and Pollack(1991)]{ruden91} Ruden, S.~P., Pollack,

2048: J.~B.\ 1991.\ The dynamical evolution of the protosolar nebula.\

2049: Astrophysical Journal 375, 740-760.

2050: 

2051: \bibitem[Safronov(1969)]{safronov69} Safronov, V.~S. 1969, In Russian,

2052: English translation: NASA-TTF-677, 1972

2053: 

2054: \bibitem[Sandford and Allamandola(1990)]{sandford90} {Sandford, S.~A., 

2055: Allamandola, L.~J.\ 1990.\ The volume- and surface-binding energies of

2056: ice systems containing CO, CO2, and H2O.\ Icarus 87, 188-192.

2057: 

2058: %\bibitem[Scott and Krot(2005)]{scott05} Scott, E.~D.~R., Krot, A.~N.\

2059: %2005.\ Thermal Processing of Silicate Dust in the Solar Nebula: Clues

2060: %from Primitive Chondrite Matrices.\ Astrophysical Journal 623, 571-578.

2061: 

2062: \bibitem[Semenov et al.(2003)]{semenov03} Semenov, D., Henning, Th.,

2063: Helling, Ch., Ilgner, M., Sedlmayr, E.\ 2003.\ Rosseland and Planck mean

2064: opacities for protoplanetary discs.\ Astronomy and Astrophysics 410,

2065: 611-621.

2066: 

2067: \bibitem[Shakura and Syunyaev(1973)]{shakura73} Shakura, N.~I., 

2068: Syunyaev, R.~A.\ 1973.\ Black holes in binary systems. Observational

2069: appearance.\ Astronomy and Astrophysics 24, 337-355

2070: 

2071: \bibitem[Spohn and Schubert(2003)]{spohn03} Spohn, T., Schubert, G.\

2072: 2003.\ Oceans in the icy Galilean satellites of Jupiter? Icarus 161,

2073: 456-467.

2074: 

2075: \bibitem[Stamatellos and Whitworth(2008)]{stamatellos08} Stamatellos,

2076: D., Whitworth, A.~P.\ 2008.\ Can giant planets form by gravitational

2077: fragmentation of discs? Astronomy and Astrophysics 480, 879-887.

2078: 

2079: \bibitem[Stepinski and Valageas(1997)]{stepinski97} Stepinski, T.~F.,

2080: Valageas, P.\ 1997.\ Global evolution of solid matter in turbulent

2081: protoplanetary disks. II. Development of icy planetesimals.\ Astronomy

2082: and Astrophysics 319, 1007-1019.

2083: 

2084: \bibitem[Tielens and Allamandola(1987)]{tielens87} Tielens, A.~G.~G.~M.,

2085: Allamandola, L.~J.\ 1987.\ Composition, structure, and chemistry of

2086: interstellar dust.\ Interstellar Processes 134, 397-469.

2087: 

2088: \bibitem[Tielens and Hagen(1982)]{tielens82} Tielens, A.~G.~G.~M., 

2089: Hagen, W.\ 1982.\ Model calculations of the molecular composition of

2090: interstellar grain mantles.\ Astronomy and Astrophysics 114, 245-260.

2091: 

2092: %\bibitem[Tobie et al.(2005)]{tobie05} Tobie, G., Grasset, O., Lunine, J.

2093: %I., Mocquet, A., Sotin, C.\ 2005.\ Titan's internal structure inferred

2094: %from a coupled thermal-orbital model.\ Icarus 175, 496-502.

2095: 

2096: \bibitem[Tsiganis et al.(2005)]{tsiganis05} Tsiganis, K., Gomes, R.,

2097: Morbidelli, A., Levison, H.~F.\ 2005.\ Origin of the orbital

2098: architecture of the giant planets of the Solar System. Nature 435,

2099: 459-461.

2100: 

2101: \bibitem[Turner and Sano(2008)]{turner08} Turner, N.~J., Sano, T.\

2102: 2008.\ Dead Zone Accretion Flows in Protostellar Disks.\ Astrophysical

2103: Journal 679, L131-L134.

2104: 

2105: %\bibitem[Turner et al.(2007)]{turner07} Turner, N.~J., Sano, T., 

2106: %Dziourkevitch, N.\ 2007.\ Turbulent Mixing and the Dead Zone in

2107: %Protostellar Disks.\ Astrophysical Journal 659, 729-737.

2108: 

2109: \bibitem[Umebayashi and Nakano(1980)]{umebayashi80} Umebayashi, T., 

2110: Nakano, T.\ 1980.\ Recombination of Ions and Electrons on Grains and the

2111: Ionization Degree in Dense Interstellar Clouds.\ Publications of the

2112: Astronomical Society of Japan 32, 405.

2113: 

2114: \bibitem[Umebayashi and Nakano(1981)]{umebayashi81} Umebayashi, T., 

2115: Nakano, T.\ 1981.\ Fluxes of Energetic Particles and the Ionization Rate

2116: in Very Dense Interstellar Clouds.\ Publications of the Astronomical

2117: Society of Japan 33, 617.

2118: 

2119: \bibitem[van Boekel et al.(2004)]{vanboekel04} van Boekel, R., and 22

2120: colleagues. 2004.\ The building blocks of planets within the

2121: `terrestrial' region of protoplanetary disks.\ Nature 432, 479-482.

2122: 

2123: \bibitem[Weidenschilling(1977)]{weidenschilling77} Weidenschilling,

2124: S.~J.\ 1977.\ The distribution of mass in the planetary system and solar

2125: nebula.\ Astrophysics and Space Science 51, 153-158.

2126: 

2127: %\bibitem[Weissman(1996)]{weissman96} Weissman, P.~R.\ 1996.\ The Oort

2128: %Cloud.\ Completing the Inventory of the Solar System, 107, 265-288.

2129: 

2130: \bibitem[Willacy and Langer(2000)]{willacy00} Willacy, K., Langer,

2131: W.~D.\ 2000.\ The Importance of Photoprocessing in Protoplanetary

2132: Disks.\ Astrophysical Journal 544, 903-920.

2133: 

2134: \bibitem[Willacy et al.(2006)]{willacy06} Willacy, K., Langer, W.,

2135: Allen, M., Bryden, G.\ 2006.\ Turbulence-driven Diffusion in

2136: Protoplanetary Disks: Chemical Effects in the Outer Regions.\

2137: Astrophysical Journal 644, 1202-1213.

2138: 

2139: \bibitem[Woods and Willacy(2007)]{woods07} Woods, P.~M., Willacy, K.

2140: 2007.\ Benzene Formation in the Inner Regions of Protostellar Disks.\

2141: Astrophysical Journal 655, L49-L52.

2142: 

2143: \bibitem[Womack et al.(1992)]{womack92} Womack, M., Wyckoff, S., 

2144: Ziurys, L.~M.\ 1992.\ Estimates of N2 abundances in dense molecular

2145: clouds.\ Astrophysical Journal 401, 728-735.

2146: 

2147: \end{thebibliography}

2148: 

2149: %Tables Here

2150: 

2151: \include{freedisk}

2152: \include{solarcomp}

2153: \include{freechem}

2154: \include{binding}

2155: \include{cloudoutput}

2156: \include{iceratios}

2157: \include{sdtable}

2158: 

2159: \clearpage

2160: 

2161: %Figures Here

2162: 

2163: %\begin{figure}

2164: %\centering

2165: %\includegraphics{taucont.jpg}

2166: %\caption{Optical depth as a function of disk height and radius.}

2167: %\label{opdepth}

2168: %\end{figure}

2169: 

2170: \begin{figure}

2171: \centering

2172: \includegraphics[scale=0.9]{smooth.jpg}

2173: \caption{Opacity as a function of temperature and density.  The black

2174: line shows the unsmoothed tables of Ferguson et al. (2005) ($T >

2175: 1000$K)and Semenov et al. (2003) ($T < 700$K).  In the range $700 < T <

2176: 1000$K, we take a weighted average in log-T space of the two tables.

2177: Finally, the green curve shows the smoothed opacity functions we use in

2178: the model.}

2179: \label{opplot}

2180: \end{figure}

2181: 

2182: \begin{figure}

2183: \centering

2184: \includegraphics[scale=0.9]{sdplanet.jpg}

2185: \caption{Surface density as a function of radius and time in the giant

2186: planet-forming region.  Each panel shows a snapshot of the disk surface

2187: density profile in black, with the profiles from previous epochs

2188: retained in gray.  During the first $5 \times 10^5$~yr of evolution, a

2189: wave of local surface density enhancement propagates outward, depositing

2190: mass in the giant planet region of the solar nebula.}

2191: \label{sdplanet}

2192: \end{figure}

2193: 

2194: \begin{figure}

2195: \centering

2196: \includegraphics[scale=0.9]{sdwholedisk.jpg}

2197: \caption{Surface density as a function of radius and time throughout the

2198: whole disk.  Each panel shows a snapshot of the disk surface density

2199: profile in black, with the profiles from previous epochs retained in

2200: gray.  The disk expands from $30$~AU to $> 80$~AU over the course of the

2201: simulation.}

2202: \label{sdwholedisk}

2203: \end{figure}

2204: 

2205: \begin{figure}

2206: \centering

2207: \includegraphics[scale=0.9]{sdann.jpg}

2208: \caption{Surface density as a function of time at 5, 10, 15 and 20~AU.

2209: An outward-propagating wave adds mass to each annulus as it passes

2210: through.  The largest relative surface density increase, $\Delta \Sigma

2211: / Sigma_0$, occurs at 10~AU, near Saturn's feeding zone.}

2212: \label{sdann}

2213: \end{figure}

2214: 

2215: \begin{figure}

2216: \centering

2217: \includegraphics[scale=0.9]{mdot.jpg}

2218: \caption{Time snapshots of accretion rate as a function of radius.

2219: Negative values of $\dot{M}$ indicate ``decretion'' regions of the disk

2220: with net outward mass flow.}

2221: \label{mdot}

2222: \end{figure}

2223: 

2224: \begin{figure}

2225: \centering

2226: \includegraphics[scale=0.9]{midplanetemp.jpg}

2227: \caption{Midplane temperature as a function of radius and time in the

2228: inner 30~AU of the disk.  Outside of 30~AU, the midplane is isothermal

2229: at 20~K, the assumed ambient temperature.  The dashed line at 800~K

2230: shows the minimum temperature for silicate crystallization (Gail 1998);

2231: the dotted line at 160~K shows the sublimation temperature of H$_2$O

2232: from the temperature-programmed desorption experiments of

2233: \cite{collings04}; and the dash-dotted line at 20~K shows the assumed

2234: ambient temperature.}

2235: \label{midplanetemp}

2236: \end{figure}

2237: 

2238: \begin{figure}

2239: \centering

2240: \begin{tabular}{c}

2241: \includegraphics[scale=0.5]{carbon.jpg} \\

2242: \includegraphics[scale=0.5]{nitrogen.jpg} \\

2243: \includegraphics[scale=0.5]{oxygen.jpg} \\

2244: \end{tabular}

2245: \caption{Distribution of C, N and O atoms among dust and most common ice

2246: species.  Proportions are rounded to the nearest 1\%.}

2247: \label{pies}

2248: \end{figure}

2249: 

2250: \begin{figure}

2251: \centering

2252: \includegraphics[scale=0.9]{freezeabun.jpg}

2253: \caption{Time snapshots of ice abundance as a function of radius for

2254: species formed mainly by gas freezeout.  Note that the condensation

2255: fronts are not step functions: the region of partial ice freezeout can

2256: cover up to 3 AU.}

2257: \label{freezeabun}

2258: \end{figure}

2259: 

2260: \begin{figure}

2261: \centering

2262: \includegraphics[scale=0.6]{sulfurnetwork.jpg}

2263: \caption{Reactions linking the most abundant sulfuric ices, H$_2$S,

2264: H$_2$CS and OCS.  The label ``G'' in front of a compound name denotes

2265: that the species is solid, adsorbed on the grains.  The notation $H_2S

2266: \stackrel{H}{\rightarrow} HS$ denotes the reaction $H_2S + H \rightarrow

2267: HS + H_2$; all other reaction pathways can be read analogously to this

2268: example.  Non-sulfuric reaction by-products, such as H$_2$ in the

2269: aforementioned reaction, are not shown.  An unlabeled double arrow, such

2270: as the one connecting GOCS and OCS, denotes a phase change.  Reaction

2271: pairs are not in equilibrium.}

2272: \label{sulfurnetwork}

2273: \end{figure}

2274: 

2275: \begin{figure}

2276: \centering

2277: \includegraphics[scale=0.9]{sulfuric.jpg}

2278: \caption{Abundances of solid NO, N$_2$, and sulfur-carrying species as a

2279: function of radius and time.  We find a banded sulfur ice distribution

2280: in which H$_2$S dominates outside $\sim 7$~AU, OCS is most abundant

2281: between 5 and 7~AU, and H$_2$CS is dominant in the inner nebula.}

2282: \label{sulfuric}

2283: \end{figure}

2284: 

2285: \begin{figure}

2286: \centering

2287: \includegraphics[scale=0.9]{hydrocarbon.jpg}

2288: \caption{Hydrocarbon ice abundances as a function of radius and time.

2289: We group all C$_3$H$_n$ compounds together because our reaction sequence

2290: ends at C$_3$H$_4$, so we cannot differentiate between different levels

2291: of C$_3$ saturation.}

2292: \label{hydrocarbon}

2293: \end{figure}

2294: 

2295: \begin{figure}

2296: \centering

2297: \includegraphics[scale=0.9]{solids.jpg}

2298: \caption{Time snapshots of solid surface density available for planet

2299: formation as a function of radius.  In each panel, the current solid

2300: surface density is plotted in black and previous time snapshots are

2301: retained in gray.}

2302: \label{solids}

2303: \end{figure}

2304: 

2305: 

2306: \end{document}

2307: