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

2: 

3: \usepackage{natbib}

4: \usepackage{epsfig}

5: 

6: \newcommand{\rhill}{r_{\mathrm{H}}}

7: \newcommand{\msun}{\mathrm{M}_{\sun}}

8: \newcommand{\mearth}{\mathrm{M}_{\oplus}}

9: \newcommand{\AU}{\:\mathrm{AU}}

10: \newcommand{\figref}[1]{Figure \ref{#1}}

11: \newcommand{\eqref}[1]{equation (\ref{#1})}

12: \newcommand{\Eqref}[1]{Equation (\ref{#1})}

13: 

14: \shorttitle{Radiative Transfer in Protoplanetary Disks}

15: \shortauthors{Jang-Condell \& Sasselov}

16: %\slugcomment{DRAFT \today}

17: \slugcomment{submitted to ApJ February 18, 2003 -- revised \today}

18: 

19: \begin{document}

20: \newcommand{\etal}{et al.~}

21: \newcommand{\eg}{e.g.~}

22: \title{Radiative Transfer on Perturbations in Protoplanetary Disks}

23: \author{Hannah Jang-Condell\altaffilmark{1}

24: \and Dimitar D.~Sasselov\altaffilmark{2}}

25: \affil{Harvard-Smithsonian Center for Astrophysics}

26: \affil{60 Garden St., Cambridge, MA 02138}

27: \altaffiltext{1}{hjang@cfa.harvard.edu}

28: \altaffiltext{2}{dsasselov@cfa.harvard.edu}

29: 

30: \newcommand{\tunpert}{133.6\:\mathrm{K}}	

31: % 1:       130.683      135.475      133.596      135.475

32: \newcommand{\tsmmin}{130.7\:\mathrm{K}}

33: \newcommand{\tsmmax}{135.5\:\mathrm{K}}

34: \newcommand{\tsmd}{5\:\mathrm{K}}

35: \newcommand{\losm}{2.2}

36: \newcommand{\hism}{1.4}

37: \newcommand{\dsm}{4}

38: % 10:      123.678      142.308      133.596      157.056

39: \newcommand{\tlgmin}{123.7\:\mathrm{K}}

40: \newcommand{\tlgmax}{142.3\:\mathrm{K}}

41: \newcommand{\tlgd}{19\:\mathrm{K}}

42: \newcommand{\lolg}{7.4}

43: \newcommand{\hilg}{6.5}

44: \newcommand{\dlg}{14}

45: % 20:      122.036      146.574      133.596      153.426

46: \newcommand{\txlmin}{122.0\:\mathrm{K}}

47: \newcommand{\txlmax}{146.6\:\mathrm{K}}

48: \newcommand{\txld}{25\:\mathrm{K}}

49: \newcommand{\hixl}{8.7}

50: \newcommand{\loxl}{9.7}

51: \newcommand{\dxl}{18}

52: 

53: \begin{abstract}

54: We present a method for calculating the radiative 

55: tranfer on a protoplanetary disk perturbed by a protoplanet.  

56: We apply this method to determine the effect on the temperature 

57: structure within the photosphere of a passive circumstellar disk 

58: in the vicinity of a small protoplanet of up to 20 Earth masses.  

59: The gravitational potential of a protoplanet induces a compression 

60: of the disk material near it, resulting in a decrement in the density 

61: at the disk's surface.  Thus, an isodensity contour at the height 

62: of the photosphere takes on the shape of a well.  When such a well is 

63: illuminated by stellar irradiation at grazing incidence, 

64: it results in cooling in a shadowed 

65: region and heating in an exposed region.  For typical stellar and 

66: disk parameters relevant to the epoch of planet formation, 

67: we find that the temperature variation due to a protoplanet at 

68: 1 AU separation from its parent star is about \dsm\% ($\tsmd$) for a planet 

69: of 1 Earth mass, about \dlg\% ($\tlgd$) for planet of 10 Earth masses, 

70: and about \dxl\% ($\txld$) for planet of 20 Earth masses, 

71: We conclude that even such relatively small protoplanets can induce 

72: temperature variations in a passive disk.  Therefore, many of the processes 

73: involved in planet formation should not be modeled with a locally 

74: isothermal equation of state.

75: \end{abstract}

76: 

77: \keywords{planetary systems: protoplanetary disks --- 

78: planetary systems: formation --- radiative transfer}

79: 

80: \section{Introduction}

81: 

82: Planetary systems are formed from rotating protoplanetary disks,

83: which are often modeled

84: in steady-state, in vertical hydrostatic equilibrium, with gas and 

85: dust fully mixed and thermally coupled \citep{kenyon}.  

86: Such disks describe very well the observed properties of

87: T Tauri disks of age $\sim$1~Myr and typical mass accretion rates

88: of $\leq 10^{-8} M_{\odot} {\rm yr}^{-1}$ \citep{hartmann,vertstruct}. 

89: They are passive disks in the sense modeled

90: by \eg \citet{CG}, where the main source of photospheric heating 

91: is irradiation from the parent star, although viscous heating is 

92: still important at small stellar distances and near the midplane.  

93: The temperature in such disks

94: is computed under the assumption that the upper surface of the disk

95: is perfectly concave and smooth at all radii, which is a very good

96: description of an unperturbed disk because thermal and

97: gravitational instabilities are damped very efficiently \citep{thermstab}.

98: 

99: We are interested in planet formation inside such disks, which

100: means that we might be forced to abandon the above assumption of

101: a perfectly smooth surface. The newly formed planet core will

102: distort it, affect the heating and cooling of the disk

103: locally, and could have significant consequences for the further

104: growth and migration of protoplanets. As discussed in \citet{sas_lec}, 

105: the distortion need only be large enough compared to

106: the grazing angle at which the starlight strikes the disk. This

107: small angle has a minimum

108: at $0.4\AU$ and increases significantly only at very large distances.

109: The depth of the depression due to the additional mass of the planet, $m_p$,

110: will be proportional to $(\rhill/h)^3$, where $\rhill$ is the Hill radius,

111: and $h$ is the local scale height, with

112: a shade area dependent on the grazing angle \citep{sas_lec}.

113: 

114: Generally planet formation has been treated numerically in 2-dimensional

115: disks; to date, there is only one recent study of the 3-dimensional

116: effect due to planets embedded in a protoplanetary disk \citep{bate}.

117: Bate \etal model the 3-dimensional response of a gaseous viscous disk

118: with a simple locally isothermal equation of state, without taking 

119: into account irradiation from the central star.  

120: 

121: In this paper we study in detail the development of a compression

122: in a standard passive disk and the radiative transfer in it. 

123: We consider this to be the first step in developing a global

124: 3-dimensional simulation, similar to the one by \citet{bate},

125: but with a realistic equation of state, especially near and inside

126: the protoplanet's Roche lobe.

127: We are explicitly interested in the planetary growth process before a 

128: gap in the disk is formed, hence our approach is limited to 

129: $m_p\lesssim30\:\mearth$, which results in a small perturbation to the 

130: irradiated surfaces of a passive disk.

131: 

132: \section{Heating Sources}

133: 

134: A schematic of our adopted disk model is shown in \figref{flareddisk}.  

135: The disk is flared, so that disk height increases faster than 

136: distance to the star, consistent with 

137: observational evidence and theoretical models

138: \citep{kenyon,CG,vertstruct}.  

139: We assume that the dust is well mixed with the gas 

140: and is the primary source of 

141: opacity at wavelengths characteristic of stellar emission.  

142: The dust absorbs incident stellar radiation at the surface 

143: and re-emits at longer wavelength, where the opacity is much lower.  

144: We are interested in the region of the disk that is optically thick, 

145: at distances of $\lesssim10$ AU \citep{dalessio2}.  

146: This gives rise to three different layers with smooth transitions 

147: in the disk: 

148: the uppermost is optically thin to both stellar and disk radiation, 

149: the interior is optically thick to both stellar and disk radiation, 

150: and the middle layer is a transition region that is optically thick 

151: to stellar radiation, but still optically thin to disk radiation.  

152: We shall refer to this optically thick/thin region as the photosphere.  

153: The reprocessing of stellar radiation in this region 

154: is important for determining the interior temperature.  

155: 

156: \citet{CG} calculate the temperature of only the thin/thin and

157: thick/thick regions using energy conservation considerations, without

158: detailing the temperature structure of the photosphere.

159: \citeauthor{vertstruct} (\citeyear{vertstruct},\citeyear{dalessio2})

160: calculate the detailed temperature structure of the

161: photosphere in disks around young stars in a self-consistent way,

162: including radiative transport and convection. 

163: They find that radiative transfer is the dominant mechanism for energy 

164: transport in the disk photosphere, and the primary heating sources are 

165: stellar irradiation and viscous

166: heating, as shown by comparison to \citet{calvet}, who use the

167: Milne-Strittematter treatment of the superposition of solutions

168: \citep{milne,strittmatter}.  We shall consider stellar radiation and

169: viscous heating to be the primary sources of heating, where the total

170: temperature is arrived at by calculating the sum of the viscous and

171: radiative fluxes, as \(T^4 = T_v^4 + T_r^4\), where $T_v$ is

172: the viscous temperature and $T_r$ is stellar radiation heating

173: temperature.  

174: 

175: \subsection{Viscous Heating}

176: 

177: The temperature due to viscous heating for a constant flux, gray atmosphere is 

178: \begin{equation}

179: T_v^4 = \frac{3F_v}{4\sigma_B}(\tau_d+2/3)

180: \end{equation}

181: where $\sigma_B$ is the Stefan-Boltzmann constant, 

182: $\tau_d$ is the optical depth of the disk's own radiation,

183: and the viscous flux $F_v$ at a distance $a$, for a star of effective 

184: temperature $T_{\star}$, mass $M_{\star}$, and radius $R_{\star}$, 

185: accreting at a rate $\dot{M}_a$ is 

186: \begin{equation}

187: F_v = \frac{3GM_{\star}\dot{M}_a}{4\pi a^3}

188: 	\left[1-\left(\frac{R_{\star}}{a}\right)^{1/2}\right]

189: \end{equation}

190: \citep{pringle}.

191: 

192: 

193: \subsection{Stellar Irradiation}

194: 

195: We shall use a modified form of the Milne-Strittmatter

196: treatment to calculate the effect of heating by stellar irradiation 

197: on a three-dimensional

198: perturbation on a passive accretion disk.

199: We assume that the dust is well-mixed with the gas in the disk and that 

200: they are thermodynamically coupled so that the dust temperature 

201: and gas temperature are the same.  In general, there is no guarantee

202: that these two temperatures are the same, or that the dust is either 

203: well mixed or uniform in composition.

204: 

205: Following the Milne-Strittmatter treatment, the incident radiation 

206: from the star is considered to be at some characteristic wavelength, 

207: typically in the visible, and the radiation emitted by the 

208: disk will be at a longer characteristic wavelength, typically 

209: in the infrared.  

210: The incident, short wavelength radiation will be indicated by the 

211: subscript $s$ (stellar or scattered), and the re-emitted, long wavelength 

212: radiation will be indicated by the subscript $d$ (disk or diffuse).

213: Some fraction of the energy of the incident stellar 

214: radiation, $\sigma$, is scattered 

215: at the same frequency.  The remaining fraction, $\alpha = 1-\sigma$, is 

216: absorbed by dust in the disk and re-emitted at a longer characteristic 

217: wavelength, typically in the infrared.  This gives the 

218: equation of radiative transfer for the 

219: stellar short wavelength radiation,

220: \begin{equation}

221: \frac{1}{\kappa_s \rho} \mathbf{\hat{k}} \cdot \nabla I_s =

222: 	I_s - \frac{\sigma E_0 e^{-\tau_s}}{4\pi}

223: 	- \frac{\sigma \int I_s d\Omega}{4\pi} 

224: \end{equation}

225: where $\rho$ is the density, $\kappa_s$ is the opacity at short wavelengths, 

226: and $E_0$ is the incident energy flux given by 

227: \begin{equation}

228: E_0 = \sigma_B T_{\star}^4 \left(\frac{R_{\star}}{a}\right)^2.

229: \end{equation}

230: The zeroth moment of this equation is 

231: \begin{equation}\label{s0}

232: \frac{1}{\kappa_s \rho} \nabla \cdot \mathbf{F}_s 

233: = \alpha 4\pi J_s - \sigma E_0 e^{-\tau_s},

234: \end{equation}

235: and the first moment, making use of the Eddington approximation, is

236: \begin{equation}\label{s1}

237: \frac{1}{\kappa_s \rho} \nabla J_s = \frac{3\,\mathbf{F}_s}{4\pi}. 

238: \end{equation}

239: 

240: The radiative transfer equation for the radiation absorbed 

241: and re-emitted within the disk is 

242: \begin{equation}

243: \frac{1}{\kappa_d\rho} \mathbf{\hat{k}} \cdot \nabla I_d = I_d - B

244: \end{equation}

245: where $\kappa_d$ is the optical depth at disk radiation wavelengths.  

246: The zeroth and first moments (using the Eddington approximation) are 

247: \begin{equation}\label{d0}

248: \frac{1}{\kappa_d\rho} \nabla \cdot \mathbf{F}_d = 4\pi (J_d - B),

249: \end{equation}

250: \begin{equation}\label{d1}

251: \frac{1}{\kappa_d\rho}\,\nabla J_d = \frac{3\,\mathbf{F}_d}{4\pi}.

252: \end{equation}

253: 

254: 

255: \subsubsection{Plane-parallel Disk}

256: \label{plane-parallel}

257: 

258: In a plane parallel disk atmosphere, this problem is reduced 

259: to one dimension, and we can express everything as a function of 

260: optical depth normal to the surface, $\tau_d$.  We can substitute 

261: \[ \frac{1}{\kappa_d\rho}\nabla \rightarrow \frac{d}{d\tau_d}

262: \quad \mbox{and} \quad

263: \mathbf{F}_{s,d} \rightarrow F_{s,d}. \]

264: If the angle of incidence 

265: of the stellar radiation is $\cos^{-1}\mu_0$, then 

266: \begin{equation}

267: \tau_d = \tau_s\mu_0/q

268: \end{equation}

269: where $q$ is the ratio of opacities

270: \begin{equation}

271: q \equiv \frac{\kappa_s}{\kappa_d}.

272: \end{equation}

273: The condition of zero net flux relates the short and long wavelength 

274: fluxes as 

275: \begin{equation}\label{sd}

276: F_s + F_d = E_0 \mu_0 \exp(-q\tau_d/\mu_0)

277: \end{equation}

278: and the boundary conditions at the surface are that of an isotropic 

279: radiation field, 

280: \begin{equation}

281: 2\pi\,J_s(0) = F_s(0) \quad \mbox{and} \quad

282: 2\pi\,J_d(0) = F_d(0).

283: \end{equation}

284: 

285: Now, equations (\ref{s0}), (\ref{s1}), (\ref{d0}), (\ref{d1}), and 

286: (\ref{sd}) are a closed system of linear differential equations 

287: that we can solve directly for the temperature, 

288: $T_r = (\pi B/\sigma_B)^{1/4}$.  The solution is given in 

289: \citet{calvet} as 

290: \begin{equation}

291: \label{radtemp}

292: B = \frac{\alpha E_0 \mu_0}{4\pi} 

293: [C_1' + C_2' \exp(-q\tau_d/\mu_0) + C_3' \exp(-\beta q\tau_d)]

294: \end{equation}

295: where $\beta\equiv\sqrt{3\alpha}$, 

296: \begin{eqnarray}

297: C_1' &=& (1+C_1)\left(2+\frac{3\mu_0}{q}\right) 

298: 	+ C_2\left(2+\frac{3}{\beta q}\right), \\

299: C_2' &=& \frac{(1+C_1)}{\mu_0} \left(q-\frac{3\mu_0^2}{q}\right), \\

300: C_3' &=& C_2 \beta \left(q - \frac{3}{q\beta^2}\right),

301: \end{eqnarray}

302: and 

303: \begin{eqnarray}

304: C_1 &=& -\frac{3\sigma\mu_0^2}{1-\beta^2\mu_0^2}, \\

305: C_2 &=& \frac{\sigma(2+3\mu_0)}{\beta(1+2\beta/3)(1-\beta^2\mu_0^2)}.

306: \end{eqnarray}

307: 

308: \figref{t_unperturbed} shows how temperature varies with 

309: $\mu_0$ at the surface ($\tau_d=0$) and in the interior 

310: ($\tau_d\rightarrow\infty$).  The temperatures are normalized 

311: to an effective radiant temperature $T_0 = (E_0/\sigma_B)^{1/4}$, 

312: which is the blackbody temperature corresponding to the incident energy flux.  

313: Note that at the surface the temperature remains relatively constant 

314: with $\mu_0$ and is 

315: greater than $T_0$ due to the absorptive properties of the dust, 

316: in accordance with the ``superheated'' surface layer proposed in \citet{CG}.

317: In contrast, the interior temperature is sensitive to the incident angle, 

318: especially at grazing incidence.  Thus, perturbations in the surface 

319: that change the angle of incidence can significantly affect 

320: the interior temperature.

321: 

322: By definition, $\tau_d=2/3$ is where the disk becomes optically 

323: thick to its own radiation.  

324: For $q\gtrsim1$, the temperature at $\tau_d=2/3$ is 

325: can be approximated by the interior temperature, since as 

326: $q\tau_d$ becomes large, the exponentials in \eqref{radtemp}

327: vanish and $B$ becomes independent of $\tau_d$.  

328: This indicates that in the absence of viscous heating, 

329: the disk interior is vertically isothermal for $\tau_d\geq2/3$.

330: 

331: 

332: \subsubsection{Disk with Perturbation}

333: 

334: Now we consider radiative transfer on a perturbed disk.  

335: A point $P(x,y,z)$ within 

336: the disk ``sees'' radiative flux coming from the surface of the disk.

337: Each area element $\delta A$ contributes a 

338: solid angle $\delta\Omega$ of flux to $P$.

339: \figref{angles} shows a schematic of such an area element.  

340: The angle of incidence of radiation on this surface element is 

341: $\cos^{-1}\mu$.  If $\mathbf{k}$ is the vector from $\delta A$ to $P$ and 

342: $\mathbf{\hat{n}}$ is the unit normal to $\delta A$, then 

343: $\nu$ is the cos of the angle between $\mathbf{\hat{n}}$ and $\mathbf{k}$.

344: 

345: We approximate the flux contribution from $\delta A$ as if 

346: it were a surface emitting isotropically with intensity 

347: \[ I = \frac{F_d(\tau_d,\mu)}{\pi}. \]

348: and the corresponding contribution to the total flux is 

349: $ d\mathbf{F} = I \,\mathbf{\hat{k}}\; d\Omega.$

350: Then the total flux at $P$ over all surface 

351: contributions is 

352: \begin{equation}

353: \mathbf{F}_{\mathrm{tot}} = 

354: \frac{1}{\pi} \int F_d \,\mathbf{\hat{k}}\; \delta\Omega.

355: \end{equation}

356: Since $\nabla$ is a linear operator,

357: \begin{equation}

358: \frac{1}{\kappa_d\rho}\nabla\cdot\mathbf{F}_{\mathrm{tot}} =

359: 	\frac{1}{\pi} \int \frac{d F_d}{d\tau_d} \,

360: 		\mathbf{\hat{n}} \cdot \mathbf{\hat{k}}\; \delta\Omega =

361: 	4\pi\left(

362: 	\frac{1}{\pi} \int J_d\, \nu \; \delta\Omega -

363: 	\frac{1}{\pi} \int B\, \nu \; \delta\Omega \right).

364: \end{equation}

365: Defining 

366: \begin{eqnarray}

367: J_{\mathrm{tot}} &=& \frac{1}{\pi} \int J_d(\tau_d,\mu)\,\nu\;\delta\Omega 

368: \quad \mbox{and} \\

369: B_{\mathrm{tot}} &=& \frac{1}{\pi} \int B(\tau_d,\mu)\,\nu\;\delta\Omega 

370: \label{btot}

371: \end{eqnarray}

372: we see that $\mathbf{F}_{\mathrm{tot}}$, $J_{\mathrm{tot}}$, 

373: and $B_{\mathrm{tot}}$ satisfy equations (\ref{d0}) and (\ref{d1}), 

374: and we can calculate the perturbed temperature as 

375: \begin{equation}

376: T_r = \left(\frac{\pi B_{\mathrm{tot}}}{\sigma_B}\right)^{1/4}.

377: \end{equation}

378: 

379: For a plane-parallel surface, $\tau_d$ and $\mu$ are constant, and 

380: $\nu = \cos \theta$,

381: where $\theta$ is the angle with respect to the surface normal.

382: We then recover the solution for the plane-parallel case.  

383: 

384: 

385: This method of calculating radiative transfer can be applied to 

386: any three-dimensional disk configuration, including perturbations 

387: induced by protoplanets.  Sufficiently massive protoplanets will 

388: open a gap in the disk, in which case the effects of shadowing and 

389: illumination across the gap can be calculated for all azimuthal angles.  

390: In the following section, we will calculate radiative transfer on 

391: small perturbations induced by 

392: planets too small to open a gap, but large enough to act on the 

393: disk locally.  

394: 

395: 

396: \section{Hydrostatic Equilibrium}

397: 

398: Now we consider a perturbation induced by the gravitational potential 

399: of a protoplanet on a circumstellar disk.  We shall consider only 

400: the gravitational effects of the protoplanet on hydrostatic 

401: equilibrium and assume that any resonant effects are washed out by 

402: gas drag.  

403: 

404: \subsection{Vertical Density Profile}

405: 

406: In order to find the vertical structure of a non-self-gravitating 

407: gaseous disk orbiting a central star, 

408: we solve the equation of hydrostatic equilibrium:

409: \begin{equation}

410: \nabla P = - \rho \nabla \Phi

411: \end{equation}

412: where $P$ is the gas pressure of the disk and $\Phi$ is the gravitational 

413: potential of the central star.  Since we are interested in the vertical 

414: structure, we consider only the $z$ components and assume that the disk is 

415: isothermal in the $z$ direction.  For a disk without a perturbing 

416: protoplanet, \( \Phi = GM_{\star}/a \) where $M_{\star}$ and $a$ 

417: are the mass and distance of the star, respectively.  

418: If $c_s$ is the isothermal sound speed, then 

419: \begin{equation}

420: c_s^2\,\frac{d\, \ln \rho}{dz} = 

421: - \frac{GM_{\star}}{a^3} z.

422: \label{thediffeq}

423: \end{equation}

424: If we assume that $a\gg z$ and define 

425: \( h \equiv c_s\,a^{3/2}/( GM_{\star})^{1/2}

426: = (c_s/v_{\phi}) a, \)

427: then 

428: \begin{equation}

429: \rho(z) =  \rho_0 \exp\left(-\frac{z^2}{2 h^2}\right),

430: \label{unperturbed}

431: \end{equation}

432: i.e.~the density has a Gaussian distribution with a scale height 

433: of $h$ which is determined by the local temperature. 

434: 

435: The insertion of a planet into a passive disk adds an additional term 

436: to $\Phi$, representing the potential due to the planet:

437: \[ \Phi = GM_{\star}/a + G m_p/r_p \]

438: where $m_p$ and $r_p$ are the mass and distance of the planet.

439: Assuming that the planet is in the midplane ($z=0$),

440: Eq.~(\ref{thediffeq}) becomes

441: \begin{equation}

442: \frac{d\, \ln \rho}{dz} =  -\frac{z}{h^2}

443: 	- \left(\frac{M_p a^3}{M_{\star} h^2}\right) \frac{z}{r_p^3}.

444: \end{equation}

445: The Hill radius is defined as 

446: \( \rhill \equiv (M_p/3M_{\star})^{1/3} a,\)

447: and this equation integrates to 

448: \begin{equation}

449: \rho(z) = \rho_1 \exp\left[-\frac{z^2}{2 h^2}

450: 	+ \frac{3\rhill^3}{h^2 \, r_p}

451: 	\right]

452: \end{equation}

453: where \(r_p = \sqrt{x^2 + y^2 + z^2}\), 

454: with the planet as the coordinate origin.

455: So long as $\rhill \ll h$ and $|z|>\rhill$, 

456: the effect of the planet is a small 

457: perturbation on the density profile.  

458: 

459: We normalize this equation by matching 

460: the density at $z=0$ to the unperturbed density, thus

461: \begin{equation}

462: \rho_p(z) = \rho_0 \exp\left[-\frac{z^2}{2 h^2}

463:         + \frac{3\rhill^2}{h^2}

464: 	  \left(\frac{\rhill}{\sqrt{x^2+y^2+z^2}}

465: 	-\frac{\rhill}{\sqrt{x^2+y^2}}\right)\right]

466: \label{pertdens}

467: \end{equation}

468: 

469: Note that there is a singularity ar $r_p = 0$.   \figref{zprof} shows 

470: the shape of this density profile. 

471: In \figref{zprof}a, we hold $\rhill/h$ fixed at $0.5$ and vary

472: $\sqrt{x^2+y^2}$.  

473: For a disk with $h=0.04\AU$ at $1\AU$ around a central star of

474: $0.5\:\msun$, this corresponds to a protoplanetary mass of

475: $4\:\mearth$.  For

476: $(x^2+y^2)\gtrsim\rhill^2$, the perturbed density profile does not

477: deviate significantly from the unperturbed density profile.  For

478: $(x^2+y^2)\lesssim\rhill^2$, the density profile develops a sharper

479: profile at $z<h$.  

480: In \figref{zprof}b, we hold $\sqrt{x^2+y^2}$ fixed at the Hill radius

481: and vary the mass of the protoplanet, effectively varying $\rhill$.

482: Larger protoplanets have the effect of decreasing the overall density 

483: profile, whereas position controls the shape of the profile, 

484: particularly close to the protoplanet.  

485: Since the flow of gas within the Roche lobe cannot

486: be adequately described hydrostatically, we exclude the region

487: $\sqrt{x^2+y^2} \lesssim \rhill$ from our analysis.

488: 

489: The normalization on \eqref{pertdens} is somewhat arbitrary, but 

490: at the disk heights that are of interest to us, the overall normalization 

491: has small effect.  We demonstrate this as follows: 

492: as $z\gg\rhill$, 

493: \eqref{pertdens} becomes

494: \begin{equation}

495: \rho_p(z\gg\rhill) \approx \rho_0

496: \exp\left(-\frac{z^2}{2 h^2}-\frac{3\rhill^3}{h^2\sqrt{x^2+y^2}}\right),

497: \end{equation}

498: that is, a Gaussian with respect to $z$ with some normalization 

499: that depends on $\rhill$ and $\sqrt{x^2+y^2}$.

500: The height of the photosphere, $H$, is generally several times the 

501: pressure scale height.  If we take $H/h=5$, then the change in 

502: density with vertical distance is 

503: \begin{equation}

504: \label{drho2}

505: \frac{\Delta\rho}{\rho} = 

506: \left|\frac{\Delta z}{\rho}\frac{d\rho}{dz}(z=H)\right|

507: = \frac{25\Delta z}H.

508: \end{equation}

509: So, a density change by a factor of $1/2$ corresponds to a shift in 

510: vertical height of 2\%.  In other words, the density gradient in the 

511: photosphere is so high that changes in the normalization of the 

512: density will not significantly affect our results. 

513:  

514: \subsection{Photosphere Height}

515: \label{photosphere}

516: We calculate the height of the perturbed photosphere from the 

517: perturbed density profile as the isodensity contour at the 

518: density of the unperturbed photosphere.

519: That is, if $H_0$ is the unperturbed photosphere 

520: height and $H=H_0(1-\epsilon)$ is the new photosphere height, then 

521: we set 

522: \[

523: \rho_d(H_0) = \rho_p(H)

524: \]

525: or

526: \begin{equation}

527: -\frac{H_0^2}{2} =

528: -\frac{H^2}{2} + 3\rhill^3 

529: 	  \left(\frac{1}{\sqrt{x^2+y^2+H^2}}-\frac{1}{\sqrt{x^2+y^2}}\right).

530: \end{equation}

531: Setting \( \sqrt{x^2+y^2} = \gamma H_0 \), then

532: to first order in $\epsilon$

533: 

534: \begin{equation}

535: H \approx H_0 \left\{1 - 

536: 	3\left(\frac{\rhill}{H_0}\right)^3

537: 	\left[ \frac{1}{\gamma} - \frac{1}{\sqrt{\gamma^2+1}} \right] 

538: 	\right\} 

539: \end{equation}

540: so that the depth of the perturbation scales as $(\rhill/H_0)^3$.  

541: For typical disk models, the ratio of the photosphere height to the 

542: thermal scale height, $H_0/h$, is between 3 and 5, depending on the 

543: distance from the star and is often taken to be a constant 

544: \citep{kenyon,CG,vertstruct}.  Restating the scaling of the 

545: depth of the perturbation as $\propto(\rhill/h)^3$, we see that this 

546: agrees with the estimate made in \citet{sas_lec}.

547: 

548: The shape of this perturbed photosphere is shown in \figref{hprof} 

549: for varying values of $\rhill$.  The perturbation is small for $r>\rhill$, 

550: and there is a singularity at $r=0$.  As the Hill radius increases, 

551: the depth of the perturbation also increases.  

552: Note that if $H_0/h=5$, then $\rhill/H_0=0.2$ means that the Hill 

553: radius is comparable to the disk pressure scale height and

554: the perturbation is no longer small.  Protoplanets this massive are 

555: likely to open a gap in the disk rather than simply inducing a 

556: perturbation to the photosphere.

557: 

558: 

559: 

560: \section{Disk Shear}

561: 

562: The differential rotation of the disk causes the disk material 

563: near the protoplanet to move through the perturbation at a rate 

564: that depends on the velocity with respect to the protoplanet's 

565: orbit.  Material moving along a given streamline in the disk 

566: may pass through the perturbation too quickly to experience 

567: significant heating and cooling, thereby diminishing the effect 

568: of the perturbation on the temperature structure of the disk.

569: 

570: The velocity of the disk material with respect to 

571: the protoplanet is 

572: $v \simeq a(\Omega-\Omega_p)$ where $\Omega$ and $\Omega_p$ are 

573: the orbital angular velocities of the disk material and the protoplanet,

574: respectively, and $a$ is the orbital radius of the protoplanet.

575: The orbital angular velocity of a gaseous disk is 

576: \begin{equation}\label{streamvel}

577: r^2\Omega^2 = \frac{GM_{\star}}{r} + \frac{1}{\rho}\frac{dP}{dr}.

578: \end{equation}

579: In thin disks, such as we are investigating, the pressure gradient 

580: term is small so the orbital velocity is nearly equal to the 

581: Keplerian velocity.  

582: 

583: The heating/cooling rate of the disk material can be expressed as 

584: \begin{equation}

585: C \frac{\partial T}{\partial t} = F - \sigma T^4

586: \end{equation}

587: where we take $F = \sigma(T_v^4+T_r^4)$ and $C$ is the specific 

588: heat per unit surface area of the disk.

589: We shall adopt a specific heat of $C = k\Sigma/m$ where $k$ is the 

590: Boltzmann constant, $\Sigma$ is the total surface density, and 

591: $m$ is the mean molecular mass.  

592: 

593: We shall assume that the streamlines of the disk material follow 

594: \eqref{streamvel}, and that they are not perturbed by the 

595: protoplanet's graviational potential.  This is an adequate approximation 

596: for disk material outside the protoplanet's Hill radius.

597: Along each streamline, we calculate the total radiative flux, $F$, 

598: at a given position, and using the velocity of the streamline 

599: with respect to the protoplanet along with heating/cooling 

600: rate, we can calculate the steady state temperature at each position 

601: in the disk. 

602: 

603: 

604: \section{Model Parameters}

605: 

606: Dust properties can be parametrized by $\alpha$, $\kappa_s$, and $q$.

607: Then, the temperature profile in a plane-parallel disk is determined 

608: completely by $F_v$, $E_0$, $\mu_0$, and $\tau_d$, where $\tau_d$ is 

609: measured normal to the surface of the photosphere.  

610: Stellar properties determine $F_v$ and $E_0$, 

611: and the disk profile determines $\mu_0$.

612: If $H_0 \propto r^{\xi}$, then the incident angle is determined by 

613: \begin{equation}

614: \mu_0 = \frac{ (\xi-1) H_0/a }{

615: [1+(H_0/a)^2][1+\xi^2(H_0/a)^2]}

616: \end{equation}

617: 

618: A perturbation changes the local angle of incidence and optical depth.  

619: In particular, without plane parallel symmetry, there is no single 

620: optical depth that parametrizes the distance below the surface.  

621: Instead, we sum over different lines of sight, using the optical depth 

622: along each line of sight 

623: to determine the contribution to the disk flux as in \eqref{btot}.  

624: The local angle of incidence depends only on the geometry of the 

625: perturbed surface, while the optical depth also depends on the 

626: density structure along the line of sight.  

627: The shape of the perturbed surface depends on $\rhill$ and $H_0$, 

628: and the density structure depends on the disk pressure scale height, $h$.  

629: 

630: 

631: 

632: \subsection{Fiducial Model}

633: 

634: Motivated by observations of T Tauri stars, we will assume 

635: for our fiducial model that the central star 

636: has mass $M_{\star} = 0.5\;\msun$, radius $R_{\star} = 2 R_{\odot}$, 

637: and effective temperature $T_{\star} = 4000$ K, 

638: and that the protoplanet is at $a = 1\AU$ from the star.  

639: We assume an accretion rate of 

640: $\dot{M}_a = 10^{-8}\;\msun\,\mbox{yr}^{-1}$, so that 

641: $T_v = 72$ K at $\tau_d = 2/3$ at 1 AU.  

642: With these parameters, heating in the photosphere will be dominated 

643: by stellar irradiation rather than viscous heating.  

644: Near the midplane, however, viscous heating will dominate.  

645: Since we only consider the photosphere in our model, 

646: the calculation of the midplane temperature is outside the scope 

647: of this paper.  

648: 

649: We will assume that the fraction of scattered radiation is $\alpha = 0.28$, 

650: with opacity in the optical of 

651: $\kappa_s = 400\;\mbox{cm}^2\mbox{g}^{-1}$, and 

652: ratio of opacities $q = 20$.

653: 

654: We shall assume that $H_0 = 0.16\AU (r/1\AU)^{9/7}$,

655: consistent with \eg \citet{dalessio2}.  

656: In calculating the detailed density structure in the photosphere, 

657: we shall assume that $h=3.4\times10^{-2}\AU$ so that $H_0/h = 5$.  

658: A protoplanet with mass $m_p = 1\;\mearth$ ($10\:\mearth$) will have a 

659: Hill radius of $1.26\times10^{-2}\AU$ ($2.72\times10^{-2}\AU$).  

660: As shown by \citet{bate}, planets with masses $\lesssim 30\:\mearth$ 

661: are not massive enough to open a gap, and our perturbative 

662: treatment is valid.  We have calculated models with 

663: $\:1\mearth$, $10\:\mearth$, and $20\:\mearth$ planets, and present the 

664: results below.

665: 

666: 

667: \subsection{Numerical details} 

668: 

669: We define our coordinate axes so that 

670: $x$ is aligned with the radial direction, 

671: $y$ is aligned with $\phi$ (i.e.~the direction of planet's orbit), 

672: and $z$ is perpendicular to the orbital plane.

673: The coordinate origin is set at the surface of the disk above 

674: the planet position, so that the planet's coordinates are 

675: $(0,0,-H_0)$.  

676: 

677: We calculate the shape of the disk surface in the $x$ and $y$ directions 

678: over a range \( -10\rhill \leq x< 10\rhill \) and 

679: \( -10\rhill \leq y< 10\rhill \), 

680: and use this surface to numerically 

681: integrate temperatures within the photosphere, to a depth of 

682: $z(\tau_d=2/3) = -0.011\AU$.

683: The temperature is calculated at each point within a 

684: $64\times64\times64$ grid.

685: 

686: When summing over surface elements to find $B_\mathrm{tot}$ as in 

687: \eqref{btot}, we can approximate 

688: $\delta\Omega\approx\nu\delta A/\ell^2$ 

689: for $\delta A \ll \ell^2$, where $\ell$ is the distance between 

690: the point $P$ and $\delta A$.  However, 

691: for points close to the surface, this approximation breaks down.  

692: In addition, $\nu$ may have large excursions over $\delta A$.  

693: For this reason, we assume that each $\delta A$ is approximately planar, 

694: so that $B(\tau_d,\mu)$ does not vary, but we calculate 

695: $\int_{\delta\Omega} \nu\: d\Omega$ analytically from the limits 

696: of $\delta A$.  In other words, we calculate

697: \begin{equation}

698: B_\mathrm{tot}

699:  = \frac{1}{\pi} \sum B(\tau_d,\mu) \int\limits_{\delta\Omega} \nu\: d\Omega

700: \end{equation}

701: and solve for the temperature.

702: 

703: 

704: \section{Results}

705: 

706: Figures \ref{planet01}, \ref{planet10},  and \ref{planet20} 

707: show the spatial distribution of temperatures 

708: in the photosphere of the fiducial model for $1\:\mearth$ 

709: $10\:\mearth$, and $20\:\mearth$ planets,

710: respectively.  The colors represent temperatures in Kelvin, 

711: as indicated by the colorbars.

712: The horizontal axis indicates increasing radial distance from the star 

713: so that the photosphere is illuminated from the left.

714: The cross-section is taken at the $\tau_d=2/3$ surface. 

715: The bottommost row and leftmost column show the

716: unperturbed temperature of $\tunpert$ for comparison.  This temperature

717: is also indicated in the colorbar by a dotted white line.  

718: 

719: The direction of the planet's orbit is upward on the plots, so 

720: the interior disk material moves faster, and the 

721: exterior disk material moves slower than the planet.  

722: This causes the heated and cooled regions to shear out.  The 

723: positions of minimum and maximum temperatures are offset from 

724: the planet's position since it takes some time for the 

725: heated/cooled material to come back to the equilibrium 

726: temperature. 

727: 

728: For a $1\:\mearth$ protoplanet, 

729: the shadowing and illumination at the surface results in a slight

730: temperature change, with the area to the right of the

731: planet reaching a maximum temperature of $\tsmmax$, and the area to 

732: the left cooled to a minimun of $\tsmmin$ 

733: -- a change of $-\losm\%$ to $+\hism\%$.  

734: Thus, we see that an Earth-mass planet is too small

735: significantly affect the interior temperature of the disk.

736: 

737: A larger planet will create a larger perturbation, which will lead to

738: a greater temperature change.  

739: \figref{planet10} and \figref{planet20} show the temperature distribution

740: in the photosphere for a $10\:\mearth$ planet and $20\:\mearth$ planet,

741: respectively.  

742: The plots are scaled to the Hill radius, which goes as 

743: $m_p^{1/3}$, so the physical range in Figures \ref{planet10} and 

744: \ref{planet20} is actually larger than that in \figref{planet01}.  

745: The high temperatures close to the Hill radius are due to the low

746: density in this region, which gives a low optical depth to the

747: illuminated surface.  However, since this region is not in hydrostatic

748: equilibrium, we exclude it from our analysis.  Outside the Hill 

749: radius, the temperatures at $\tau_d=2/3$ for a $10\:\mearth$ protoplanet 

750: range from $\tlgmin$ to $\tlgmax$, or by $-\lolg\%$ to $+\hilg\%$.  

751: For a $20\:\mearth$ protoplanet, which is close to but not at the limit 

752: where a gap in the disk can form, the temperature range increase 

753: slightly, with a minimum of $\txlmin$ and maximum of $\txlmax$, 

754: or by $-\loxl\%$ to $+\hixl\%$.  

755: Thus, a protoplanet with mass 10 $\mearth$ 

756: or larger may have a significant effect on the local

757: temperature structure of a protoplanetary disk.

758: 

759: \section{Discussion}

760: 

761: As planet-forming disks are likely to be passive \citep{calvet2002}, 

762: we have explored

763: stellar irradiation effects due to a disk surface perturbation

764: induced by an embedded protoplanet. We find that even a relatively small

765: protoplanet (in the range from 1 to 20 Earth mass) produces enough

766: of a perturbation to cause shadowing. 

767: In particular, we find that in a standard passive disk accreting 

768: at $10^{-8}\:\msun\mathrm{yr}^{-1}$, 

769: the temperature variation induced by a $1\:\mearth$ planet at 1 AU 

770: is $\tsmd$, or only about $\dsm\%$.  However, we find that even though 

771: a protoplanet may not be massive enough to open a gap in a disk, it may 

772: still be able to significantly change the temperature structure in 

773: a the disk material near it.  

774: A $10\:\mearth$ ($20\:\mearth$) planet can induce a 

775: temperature variation of $\tlgd$ ($\txld$), or $\dlg\%$ ($\dxl\%$).

776: 

777: This results in a local temperature

778: perturbation which is large enough to affect the local properties of

779: the disk and the accretion rate, and possibly the migration rate

780: of the protoplanet.  

781: Although the perturbation in temperature may be 

782: too subtle to directly observe, it can have serious consequences for 

783: planet growth and migration.

784: 

785: \citet{wardA} has shown that 

786: the rate of Type I migration is strongly dependent upon the temperature 

787: gradient of the disk.  If the temperature decreases with distance from 

788: the star, then the total net torques result in inward migration 

789: of the secondary.  However, if $k = -d\ln T/d\ln r \lesssim -1$, then 

790: the opposite is true and the secondary will migrate outward.  

791: The temperature perturbation near a protoplanet locally 

792: decreases the value of $k$, and may result in halting or even 

793: reversing Type I migration.  This may resolve the conundrum that 

794: Type I migration timescales are typically much less than observed 

795: disk lifetimes of $\sim10^7$ years \citep{hartmann}.

796: Yet to date, no simulation of planet-disk interactions accounts 

797: for local temperature pertubations as described here.  

798: 

799: We have not yet considered 

800: the behavior of the dust (as a function of size) in

801: the perturbed region.  We expect the dust to remain coupled

802: to the gas and compress its scale height accordingly. In

803: fact crude estimates of dust sedimentation in the vicinity

804: of the planet confirm that. However, a detailed estimate of

805: the sedimentation timescale is necessary in order to evaluate

806: the amount of residual dust, \eg swept by stellar radiation

807: from the inner edge, and its total opacity. These are issues 

808: we plan to address in future work.  

809: 

810: In a forthcoming paper, we will undertake a parameter study 

811: of our model for radiative transfer on a perturbed disk.  

812: By varying attributes of the disk, star, and protoplanet, we 

813: will investigate the possibility that a relatively small protoplanet 

814: which is not large enough to open a gap in a disk might still be 

815: able to affect disk structure.  

816: 

817: 

818: 

819: \acknowledgements

820: We thank Nuria Calvet for numerous helpful comments in the preparation 

821: of this paper.  

822: 

823: \bibliographystyle{apj}

824: \bibliography{apj-jour,/home/hjang/Thesis/Papers/planets}

825: 

826: 

827: \begin{figure}

828: \plotone{f1.eps}

829: \caption{\label{flareddisk}Schematic of a flared disk. 

830: Proportions are not to scale.  The labels thick and thin 

831: refer to optical depth in short and long wavelengths.

832: }

833: \end{figure}

834: 

835: \begin{figure}

836: \plotone{f2.eps}

837: \caption{\label{t_unperturbed}The temperature due solely to stellar 

838: irradiation at the surface of the disk (dashed line) 

839: and in the interior (solid line) 

840: for varying values of the angle of

841: incidence, $\mu_0$, where $T_0 = (E_0/\sigma_B)^{1/4}$.  

842: }

843: \end{figure}

844: 

845: \begin{figure}

846: \epsscale{0.5}

847: \plotone{f3.eps}

848: \caption{\label{angles}Angles involved in calculation of perturbed surface.  

849: See text for details.}

850: \epsscale{1.0}

851: \end{figure}

852: 

853: \begin{figure}

854: \plottwo{f4a.eps}{f4b.eps}

855: \caption{\label{zprof}Perturbed vertical density profile for varying 

856: parameters.  Density is normalized to the midplane density.  Both 

857: plots are to the same scale and show the unperturbed density profile 

858: as a solid line.

859: (a) Density profile for fixed Hill radius and 

860: varying values of $(x^2+y^2)^{1/2}$ as indicated.

861: (b) Density profile for $(x^2+y^2)^{1/2}$ equal to the 

862: Hill radius and varying planetary masses.

863: }

864: \end{figure}

865: 

866: \begin{figure}

867: \plotone{f5.eps}

868: \caption{\label{hprof}Perturbed photosphere height as a function of 

869: distance in the $xy$-plane, $r$, for varying Hill radii.  The solid vertical 

870: line indicates where $r=\rhill$.

871: }

872: \end{figure}

873: 

874: \begin{figure}

875: \plotone{f6.eps}

876: \caption{\label{planet01}Temperature profile of the disk photosphere 

877: near a protoplanet of mass 1 $\mearth$ 

878: at $\tau_d = 2/3$ in the $xy$-plane.  

879: The colors indicate the temperature in Kelvin, scaled as shown in 

880: the colorbar. 

881: The cross in the center represents planet position, 

882: with the arrow indicating the direction of the planet's orbit.

883: The blacked-out circle indicates the extent of the Hill radius.  

884: The dotted white line 

885: in the color bar indicates the unperturbed temperature at 

886: $\tau_d=2/3$.

887: }

888: \end{figure}

889: 

890: \begin{figure}

891: \plotone{f7.eps}

892: \caption{\label{planet10}Same as \figref{planet01}, for a 

893: protoplanet of mass 10 $\mearth$.  Note that the dimensions 

894: on the figures scale with the Hill radius, so the physical sizes of 

895: the plots are $10^{1/3}$ larger than in \figref{planet01}.

896: }

897: \end{figure}

898: 

899: \begin{figure}

900: \plotone{f8.eps}

901: \caption{\label{planet20}Same as \figref{planet01}, for a 

902: protoplanet of mass 20 $\mearth$.}

903: \end{figure}

904: 

905: \end{document}

906: