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

2: %\documentclass{emulateapj}

3: 

4: 

5: \newcommand{\ISO}{{\it ISO}}

6: \newcommand{\etal}{ et al. }

7: \newcommand{\simgt}{\ga}

8: \newcommand{\simlt}{\lower.5ex\hbox{$\; \buildrel < \over \sim \;$}}

9: 

10: %\accepted{8}

11: %\journalid{}{}

12: %\articleid{}{}

13: \slugcomment{draft date: \today}

14: 

15: \begin{document}

16: 

17: \title{The Dust cloud around the White Dwarf G 29-38. 2. 

18: Spectrum from 5-40 $\mu$m and 

19: mid-infrared photometric variability}

20: %\subtitle{}

21: 

22: \author{William T. Reach\altaffilmark{1},

23: Carey Lisse\altaffilmark{2},

24: Ted von Hippel\altaffilmark{3,4},

25: %Marc J. Kuchner\altaffilmark{5},

26: %Adam Burrows\altaffilmark{4},

27: Fergal Mullally\altaffilmark{5},

28: %Mukremin Kilic\altaffilmark{3},

29: %D. E. Winget\altaffilmark{3}

30: }

31: 

32: \altaffiltext{1}{Infrared Processing and Analysis Center, MS 220-6, 

33: California Institute of Technology,

34: Pasadena, CA 91125}

35: 

36: \altaffiltext{2}{Planetary Exploration Group, Space Department,

37: Johns Hopkins University, Applied Physics Laboratory, Laurel, MD}

38: 

39: \altaffiltext{3}{Physics Department, Siena College, 515 Loudon Road, Loudonville, NY 12211}

40: 

41: \altaffiltext{4}{Visiting Research Scientist, Department of Physics, Florida International

42: University, 11200 SW 8th St, Miami, FL 33199}

43: 

44: \altaffiltext{5}{Department of Astronomy, University of Texas, 1 University Station C1400, Austin, TX 78712}

45: 

46: 

47: %\altaffiltext{5}{NASA Goddard Space Flight Center, Greenbelt, MD 20771}

48: 

49: 

50: %\altaffiltext{4}{Department of Astronomy and Steward Observatory, University of %Arizona, 933 North Cherry Avenue, Tucson, AZ 85721}

51: 

52: \email{reach@ipac.caltech.edu}

53: 

54: \begin{abstract}

55: We model the mineralogy and distribution of dust around the white dwarf G29-39 using

56: the infrared spectrum from 1-35 $\mu$m.

57: The spectral model for G29-38 dust combines a wide range of materials based on spectral

58: studies of comets and debris disks.

59: In order of their contribution to the mid-infrared emission, the most abundant minerals 

60: around G29-38 are

61: amorphous carbon  ($\lambda<8$ $\mu$m),

62: amorphous and crystalline silicates (5--40 $\mu$m), 

63: water ice (10--15 and 23--35 $\mu$m), 

64: and metal sulfides (18--28 $\mu$m).

65: The amorphous C can be equivalently replaced by other materials (like metallic Fe)

66: with featureless infrared spectra.

67: The best-fitting crystalline silicate is Fe-rich pyroxene.

68: In order to absorb enough starlight to power the observed emission, the disk must either

69: be much thinner than the stellar radius (so that it can be heated from above and below) or

70: it must have an opening angle wider than $2^{\circ}$.

71: A `moderately optically thick' torus model fits well

72: if the dust extends inward to 50 times the white dwarf radius,

73: all grains hotter than 1100 K are vaporized,

74: the optical depth from the star through the disk is $\tau_{\parallel}=5$, and

75: the radial density profile $\propto r^{-2.7}$;

76: the total mass of this model disk is $2\times 10^{19}$ g.

77: A physically thin (less than the white dwarf radius) 

78: and optically thick disk can contribute to the near-infrared continuum only;

79: such a disk cannot explain the longer-wavelength continuum or strong emission features. 

80: The combination of a physically thin,

81: optically-thick inner disk and an outer, physically thick and moderately optically thin

82: cloud or disk produces a reasonably good fit to the spectrum and requires only silicates in the outer

83: cloud.

84: We discuss the mineralogical results in comparison to planetary materials.

85: The silicate composition contains minerals found from cometary spectra and

86: meteorites, but Fe-rich pyroxene is more abundant than enstatite 

87: (Mg-rich pyroxene) or forsterite (Mg-rich olivine) in G29-38 dust, in contrast

88: to what is found in most comet or meteorite mineralogies. 

89: Enstatite meteorites may be the most similar solar system materials to 

90: the G29-38 dust.

91: Finally, we suggest the surviving core of a `hot jupiter' as an alternative (neither cometary nor asteroidal) origin for the 

92: debris, though further theoretical work is needed to determine if this hypothesis is viable.

93: \end{abstract}

94: 

95: \keywords{white dwarfs; stars: individual (G29-38, WD 2326+049); infrared: stars

96: }

97: 

98: \section{Introduction}

99: 

100: The end state of stellar evolution for most types of star is a white dwarf.

101: Planetary material is common around main sequence stars \citep{backman,rhee,trilling}.

102: Part of the planetary system is expected to survive the violent, late stages of stellar

103: evolution \citep{debes02}.

104: Thus it is to be expected that planetary materials are present around white dwarfs.

105: But since white dwarfs are so under-luminous, this material is normally impossible to detect

106: via reflected light or thermal emission, if it is distributed like the Solar System (with

107: its inner portion destroyed out to at least the maximum radius of the red giant/AGB 

108: photosphere). Young, hot white dwarfs, where debris can be detected even at 100 AU from

109: the star \citep{suchu}, are the exception to this rule.

110: Planetary material can become detectable if it moves sufficiently close to a star, for example

111: via gravitational perturbations such as produce comets in the inner Solar System (from their

112: Kuiper Belt or Oort cloud homes) and

113: meteorites on the surface of Earth (from their asteroid belt homes).

114: \citet{jura03} explained how an asteroid can pass sufficiently close to a white dwarf to be tidally

115: disrupted, leaving a disk of dust near the star.

116: Sun-grazing comets in the solar system pass within similar distances.

117: The new and growing class of dusty, metal-rich white dwarfs 

118: (classified as DAZd, where D=white dwarf, A=mostly hydrogen atmosphere, 

119: Z=trace atmospheric metals, d=dusty) allow a unique window into planetary systems 

120: \citep{vonhippel}.

121: 

122: The first-known DAZd star, and the one with the brightest infrared excess, is 

123: Giclas 29-38 (WD 2326+049; G29-38 hereafter).

124: Its effective temperature has been estimated

125: in the range 11,800 to 12,100 K with surface

126: gravity $\log g$ in the range 8.14 to 7.90

127: \citep[e.g.][]{bergeron04,koester05}.

128: In \citet{holberg06}, the model corresponding to the

129: lowest estimated $T$ highest $g$ has radius 8100 km 

130: and mass .69 $M_\odot$, 

131: while the model corresponding to the highest $T$ and lowest $g$ 

132: has radius 9650 km and mass .55 $M_{\odot}$.

133: At wavelengths longer than 2.5 $\mu$m, the spectrum is completely dominated by 

134: non-photospheric thermal emission with color temperatures 300--900 K. 

135: In paper 1 \citep{paperone}, we measured the spectrum of this star from 5.5--14 $\mu$m,

136: revealing a remarkably strong emission feature from 9--11 $\mu$m that is typical

137: of small ($< 2 \mu$m radius) silicate grains. The brightness of the disk permits

138: spectroscopy at sufficient signal-to-noise to study the dust mineralogy. We have now

139: measured the infrared spectrum out to 35 $\mu$m, and in this paper we discuss dust

140: cloud models and compositions in more detail.

141: 

142: After presenting the new observations, analyzing the variability, and compiling the spectral energy distribution in

143: \S\ref{obssec}, we model the disk several different ways.

144: In \S\ref{mineralsec}, we use a wide mix of minerals, and a simple temperature distribution, to

145: determine likely constituents of the disk based on spectral features.

146: In \S\ref{modsec}, we present optically thin shell, moderately optically thick disk, and

147: optically thick thin-disk models.

148: In \S\ref{varsec}, we present constraints based on variability.

149: In \S\ref{discussion}, we tie together the evidence gleaned from comparing the models to

150: the observations.

151: 

152: \section{Observations\label{obssec}}

153: 

154: Table~\ref{obstab} shows a log of observations with the 

155: {\it Spitzer} Space Telescope

156: \citep{werner}. 

157: The Infrared Spectrograph (IRS) \citep{houck}

158: observations were performed in ``staring''  mode,

159: wherein the source was placed on each of two nod positions

160: on each of the slits and orders.

161: IRS data were processed with the S16 pipeline 

162: (with backgrounds subtracted by differencing nods)

163: and extracted

164: using optimal extraction in the {\it Spitzer} Science Center's data

165: analysis tool SPICE. The 4.5, 8, and 24 $\mu$m observations described in Paper 1

166: were supplemented by archival Infrared Array Camera (IRAC) \citep{fazio} 3.6, 4.5, 5.8, and 8 $\mu$m

167: observations.

168: Annular aperture photometry was performed on each IRAC basic calibrated image 

169: with a 4 pixel on-source radius and 8-20 pixel background annulus; array-location-dependent,

170: aperture-loss, and pixel-phase photometric corrections were applied; and the uncertainty

171: of each image's photometry was determined by combining uncertainties due to photon statistics and background removal \citep[see][]{reachcal}.

172: 

173: \subsection{Variability of the mid-infrared flux}

174: 

175: The variability of 

176: mid-infrared flux of G29-38 can be

177: constrained with the IRAC photometry.

178: \footnote{

179: We tested the IRS observations for variability, but there was no clear

180: evidence. The IRS observing strategy provided 6 samples along 60 sec ramps

181: in each subslit.

182: The flux varied by $< 4$\% among the shortest wavelength spectra, 

183: with no wavelength dependence detectable above the noise.}

184: Fluxes are shown in Table~\ref{iractab}.

185: The brightness of the star on each basic calibrated image

186: from each observing sequence was measured, corrected for

187: the array-location-dependent response, then a weighted

188: mean and statistical uncertainty in weighted mean computed.

189: The absolute calibration uncertainties are not included, 

190: because we are comparing fluxes from the same instrument;

191: IRAC photometry has been shown to be stable to 

192: better than 1\% \citep{reachcal}.

193: To obtain the highest signal-to-noise from

194: each image for repeatability,

195: we found an aperture radius of 4 pixels was optimal.

196: The signal-to-noise was 230, 2008, 90, and 130 at

197: 3.6, 4.5, 5.8, and 8 $\mu$m.

198: At 4.5 \& 8 $\mu$m, the star was observed in two

199: independent observations just a few minutes apart,

200: yielding identical mean fluxes; the

201: flux differences are $-0.1\pm 0.2$\% and

202: $-0.1\pm 0.3$\%, respectively.

203: At 3.6 \& 5.8 $\mu$m, the star was observed with

204: different AORs separated by 13 months. 

205: The star was fainter in the second epoch, by

206: $-3.3\pm 0.4$\% at 3.6 $\mu$m

207: (and a statistically insignificant $-1.0\pm 0.8$\%

208: at 5.8 $\mu$m). The brightness difference at 3.6 $\mu$m

209: is likely due to short-term fluctuations in the star,

210: to which we turn now.

211: 

212: The variability of the mid-infrared flux over shorter

213: timescales can be assessed from the individual images

214: taken in 2004 Nov. By the design of IRAC \citep{fazio},

215: the 3.6 and 5.8 $\mu$m channels observe simultaneously,

216: as do the 4.5 and 8 $\mu$m channels. The sequence of

217: events was as follows, with times given relative to the first

218: image at 2004-Nov-26 10:54:11.8 UT: 

219: \begin{itemize}

220: \item[00:00] the star is placed in the 4.5+8 $\mu$m field of view

221: and a short (1.2 sec) frame is taken,

222: \item[00:03] 5 consecutive dithers, consisting of a 30 sec frame

223: and a short telescope slew, are performed,

224: \item[04:34] the telescope is moved to re-center the star in the 

225: 4.5+8 $\mu$m field of view and a short (1.2 sec) frame is taken,

226: \item[04:37] 19 consecutive dithers, consisting of a 30 sec frame

227: and a short telescope slew, are performed,

228: \item[18:22] the telescope is moved to center the star in the 

229: 3.6+5.8 $\mu$m field of view,

230: \item[18:22] 19 consecutive dithers, consisting of a 30 sec frame

231: and a short telescope slew, are performed, then

232: \item[30:51] the sequence is complete.

233: \end{itemize}

234: Figure~\ref{sourcevar} shows the flux versus time.

235: 

236: The scatter in the photometric measurements is significantly

237: larger than their uncertainties, especially at

238: 3.6 $\mu$m. To assess the harmonic content of the time

239: series, we computed the periodogram as described

240: by \citet{scargle}. Figure~\ref{sourceper} shows the

241: periodograms at the 4 IRAC wavelengths.

242: The time series cannot be described by a simple period, 

243: but instead contain variation on a 

244: range of timescales.

245: This is characteristic of ZZ Ceti variable stars

246: \citep{kleinman}.

247: The peak in the 3.6 $\mu$m amplitude at 190 sec is

248: highly significant ($14\sigma$), as is the

249: harmonic content at 300-440 sec. None of the other

250: harmonic content is significant, at the sampling

251: frequency and signal-to-noise level of these

252: observations, except some power ($3.3\sigma$) at $\sim 220$ sec

253: at 4.5 $\mu$m.

254: %The variations between simultaneously-observed

255: %channels appear marginally correlated.

256: %The correlation coefficient between 3.6 and 5.8 $\mu$m

257: %fluxes is 0.30, and the correlation coefficient between

258: %4.5 and 8 $\mu$m fluxes is 0.32.

259: 

260: 

261: The flux variations seen with IRAC are similar to those seen

262: at Palomar by \citet{graham}, who observed

263: simultaneously at wavelengths dominated by the photosphere

264: (B, J) and the infrared excess (K-band).

265: They found fluctuations at periods

266: 181 and 243 sec in K-band, with no corresponding 

267: ones at J-band. Our IRAC 3.6 $\mu$m results confirm 

268: the significant fluctuations at $\sim 190$ sec.

269: The amplitude of the fluctuations $\sim 4$\%

270: at 3.6 $\mu$m, while \citet{graham} found

271: 2.5\% variations at 2.2 $\mu$m.

272: The stellar photosphere must be subtracted before

273: interpreting these results.

274: At 2.2 (3.6) $\mu$m, the photosphere contributes

275: 66\% (18\%) of the total flux. If we assume

276: the infrared photosphere is constant, and the

277: fluctuations are due to the disk, then the

278: amplitude of disk fluctuations 

279: at 2.2 (3.6) $\mu$m is 7\% (5\%),

280: i.e. very similar, with K-band possibly slightly

281: higher in amplitude.

282: Fluctuations at 4.5, 5.8, 8 $\mu$m are not detected

283: in the periodogram,

284: with upper limit $\sim$ 5\%, 5\%, 3\% of the total

285: flux, or 5\%, 5\%, 3\% of the infrared excess

286: above photosphere ($2\sigma$ limits). The upper limit

287: at 8 $\mu$m, and the trend such that fluctuations are

288: most significant at 2.2 and 3.6 $\mu$m, with a rapid 

289: decrease in amplitude at longer wavelengths, are significant

290: and indicate that the fluctuations are due to an

291: emitting region with a relatively high 

292: ($> 1000$ K) color temperature.

293: 

294: \def\oldnote{

295: {\bf NOTE:} discuss variability WITHIN the IRS observations either here or above...if full 8\% 

296: observed variability applies across the wavelength range

297:  then within

298: the spectra there could have been variations; certainly between

299: subslits...

300: original SL1+2 observations were 3 cycles of 60 sec each subslit so they took ~400 sec per subslit to perform; this is long enough that the IRAC-observed variations would have occurred

301: }

302: 

303: \subsection{Compilation of the spectral energy distribution}

304: 

305: In order to model the mineralogy and dust distribution around G29-38, we

306: need to combine the observed data from the near- through mid-infrared,

307: correcting for calibration errors as well as source variability. This is

308: in fact not possible, since the temporal sampling of the photometry is 

309: inadequate, and the observations are not contemporaneous. We therefore

310: must assemble the various portions of the spectrum in such a way as to make

311: them most plausibly `connected'. In principle, there is a scale factor for

312: each wavelength range that depends on the epoch. The amplitude of the scale

313: factor, due to source variability,

314: may reach up to $\sim 10$\% in the near-infrared and should be less

315: than 5\% at longer wavelengths. Calibration uncertainties range from $\sim 3$\%

316: in the optical to 5\% for IRAC and MIPS to 10\% for IRS. 

317: The range of plausible scale factors is then 5-10\% across all wavelengths.

318: 

319: We first ensured that the spectroscopy and photometry were in accord.

320: The IRAC fluxes must be corrected for the spectral shape of the source first.

321: For the IRAC 8 $\mu$m channel, which is fully covered by the IRS spectrum,

322: we integrated the IRS spectrum appropriately 

323: \citep{reachcal} over the passband to calculate

324: the color-correction $K=1.16$ 

325: (where the corrected flux is the `observed' flux from the IRAC calibration

326: divided by $K$)

327: and applied this correction to the photometry.

328: The IRAC 8$\mu$m flux density can then be compared directly to the IRS spectrum:

329: the flux at the nominal wavelength 7.872 $\mu$m is 7.19 mJy from IRAC and 7.30 mJy

330: from IRS, a deviation of only 1.5\%, well within the calibration error budget.

331: The IRAC 3.6, 4.5, and 5.8 $\mu$m flux density required only very small color corrections:

332: the spectrum has a color temperature $\sim 1100$ K at these wavelengths, 

333: for which the corrections are 0.995, 0.996, and 0.996, respectively.

334: The absolute fluxes from IRAC and IRS at 5.831 $\mu$m are 7.76 and 7.54 mJy, showing

335: the instrument cross-calibration is good.

336: The observed MIPS 24 $\mu$m flux was divided by a color correction of 0.97 to

337: account for the wide spectral response of the filter and detector.

338: For IRAC and IRS, we therefore find that there is no need for a relative re-scaling;

339: that is, the measured flux densities averaged over their exposure times at their

340: epoch of observation agree well, whether due to averaging of {\it or} lack of

341: variability.

342: 

343: To remove the white dwarf photosphere from the spectra, we use

344: a model atmosphere for $T=12,000$ K

345: and $\log g=8$ covering 0.35--60 $\mu$m (courtesy D. Koester),

346: normalized using

347: 2MASS photometry ($J=13.132\pm 0.029$, $H=13.075\pm 0.022$, 

348: $K_s=12.689\pm 0.029$) and optical spectrophotometry 

349: from Palomar \citep{greenstein}.

350: In fact, the optical and 2MASS photometry cannot both agree with the model

351: spectrum. If we normalize the white dwarf model at the 2MASS J band, then

352: the model would under-predict the visible flux by 8\%.

353: The near-infrared spectrum obtained by \citet{kilicspec} with the IRTF also disagrees with

354: the 2MASS photometry; if we normalize the white dwarf model by the 2MASS J-band

355: flux, then we under-predict the IRTF spectrum by 18\%.

356: The low flux observed by 2MASS is likely due to variability of the white dwarf.

357: The shape of the IRTF spectrum follows the model closely at wavelengths

358: shorter than 1.6 $\mu$m; the infrared excess is evident at wavelengths beyond 1.7 $\mu$m.

359: To obtain a joined spectral energy distribution we proceed as follows:

360: (1) normalize the white dwarf model to match the optical photometry,

361: (2) scale the 2MASS photometry by a factor 1.087 to match the model at J band,

362: (3) rescale the IRTF spectrum by 0.92 to match the white dwarf model at 0.8--1.3 $\mu$m,

363: (4) rescale the 2MASS photometry and IRTF spectra by a factor of 1.2 so that the red end of

364: the IRTF spectrum (2.4 $\mu$m) is plausibly consistent with the IRAC 3.6 $\mu$m

365: photometry.  

366: (Plausible consistency here was defined as allowing a blackbody to pass through the

367: IRAC 3.6 $\mu$m and 2MASS K$_{s}$ photometry for color temperatures 800--1200 to

368: within the error bars, and keeping the IRTF and 2MASS photometry consistent at 2.17 $\mu$m.)

369: The uncertainties were taken to be a root-sum-square combination of the measurement

370: uncertainties and 10\% of the brightness of the photosphere, to account for

371: scaling uncertainties in the photospheric subtraction.

372: 

373: Figure~\ref{ffit} shows the combined spectrum of the infrared emission after

374: subtracting the stellar photospheric emission.

375: The combined scalings yield near-infrared photometry and spectrum

376: effectively at the same epoch (in terms of variability) 

377: as the mid-infrared photometry.

378: 

379: \section{Mineralogy\label{mineralsec}}

380: 

381: The composition of the dust around G29-38 was determined by comparing the spectra to those of

382: a set of constituent materials, using a fitting method that has

383: been used for comet Hale-Bopp, the Deep Impact ejecta from

384: comet 9P/Tempel 1, and the dust around the stars HD 100546, HD 69380, and HD 113766

385: \citep{lisse06,lisse07a,lisse07b,lisse08}. Details of this model are given in

386: \citet{lisse08}.

387: In this section, we model the emitting region as an optically thin dust torus.

388: We apply this simple model to G29-38 to allow direct comparison to dust in other astronomical

389: systems, but we will revisit the mineralogy below when testing more sophisticated

390: models.

391: 

392: \subsection{Particles sizes and temperatures}

393: A toroidal model is motivated by the single temperature distributions found for many of the stars studied by 

394: \citet{beichman06a} and \citet{chen06} as well as the narrow dust structures found in many of the HST images of debris disks 

395: \citep{kalas05,kalas06}. The best-fit temperature of the smallest dust particles (0.1--1 $\mu$m), which superheat substantially above Local Thermal Equilibrium (LTE), is $T_{max} = 890$ K for the olivines, 850 - 890 K for the pyroxenes, and 930 K for the amorphous carbon. The largest particles in our calculation, with radius 1000 $\mu$m, are set to $T_{LTE}$ = 600 K, and dust of intermediate size is scaled between the two extremes \citep{lisse06}. Using the Tempel 1 ejecta temperatures as a guide (where we found olivines at 340 K and amorphous carbon at 390 K at 1.51 AU from the Sun, where $T_{LTE}$ = 230 K), and allowing for an G29-38 stellar luminosity that is 

396: $2\times 10^{-3} L_{\odot}$, we estimate a location for the hot dust $\sim 150$ stellar radii

397: from the white dwarf. The location of colder dust, capable of supporting a stable water ice component at 200 K, is at $> 10^3$ stellar radii ($>8\times 10^{11}$ cm). 

398: While differing by at least a factor of 9, this range of distances is still small, suggesting material in tight orbit around the WD.

399: The best-fit single continuum temperature to the 

400: 7--35 $\mu$m Spitzer spectrum is 950 K, close to the amorphous carbon temperature. As seen in the Tempel 1 ejecta from the Deep Impact experiment \citep{lisse06}, the amorphous carbon dominates the continuum behavior as it has a featureless emissivity and is the hottest material, thus contributing most to the short-wavelength emission. 

401: %The short wavelength emission is most diagnostic of an SED's continuum temperature, because of the steep falloff of the blackbody on the Wien law side of the emission peak and the relatively shallow, and temperature independent behavior on the long wavelength Rayleigh-Jean side of the peak. 

402: 

403: %\subsection{Size Distribution and Total Mass}

404: The best-fit size distribution

405: $dn/da\propto a^{-3.7\pm 0.2}$ argues for dust surface area dominated by small particles, but

406: dust mass dominated by large particles.

407: %The predominance of small particle surface area is why we see the strong emission features  - the dominant particles are optically thin, and the feature to continuum ratio is high. 

408: A system in collisional equilibrium would demonstrate a PSD  $\propto a^{-3.5}$ \citep{dohnanyi,durda97};

409: for ``real'' systems a size distribution even steeper than $a^{-3.5}$ at small sizes is expected in a collisional cascade 

410: %both because of its truncation at the blow-out limit  \citep{thebault} and 

411: because of the dependence of particle strength on size  \citep{obrien03}.

412: 

413: The total dust mass required to explain the Spitzer IRS spectrum 

414: (i.e., mass in particles of 0.1--10 $\mu$m in size that contribute appreciably to the $\chi^2_\nu$ value of the fit to the infrared spectrum)

415: is $\sim 2 \times 10^{19}$ g.

416: Extrapolating (using the best-fit size distribution)

417: from a maximum size of 10 $\mu$m to 1 cm would

418: increase the mass by a factor of 8.

419: A total cloud mass of order $10^{19}$ g compressed into a solid

420: sphere of average density 2.5 g~cm$^{-3}$ (i.e. rocky silicate material)

421: would have a radius 

422: of 10 km, equivalent to a single, small asteroid or large comet.

423: The surface area of {\it detected} particles is 

424: $5 \times 10^{22}$ cm$^{2}$. 

425: If this dust is in an annulus of inner radius $7\times 10^{10}$ cm 

426: ($90 R_{*}$) extending to twice that radius, 

427: then the areal filling factor of grains viewed from above the disk

428: is of order unity. 

429: The optically-thin model is thus unlikely to apply, although most of the observed emission

430: (in particular the spectral features) must arise from the optically thin parts of the disk.

431: We address the optical depth effects in the modeling sections below, but we proceed first

432: (with

433: caution) to discuss the mineralogy from the optically thin fits.

434: 

435: \subsection{Dust Composition}

436: 

437: Over 80 different species were tested for their presence in the spectra. 

438: The material spectra were selected by their reported presence in meteorites, in situ comet measurements, YSOs, and debris disks \citep{lisse06}. 

439: Consultations with members of the {\it Stardust} team, and examination of the interplanetary dust particle (IDP) literature \citep[cf. review by][]{bradley02}, and the astrominerological literature \citep[cf. review by][]{molsterwaters} pointed to the most likely mineralogical candidates to be found in the dust. Materials with emissions matching the strong features in the Spitzer emissivity spectra were also tested.

440: The list of materials tested against the {\it Spitzer} spectra included multiple silicates in the olivine and pyroxene class (forsterite, fayalite, clino- and orth-enstatite, augite, anorthite, bronzite, diopside, and ferrosilite); phyllosilicates (saponite, serpentine, smectite, montmorillonite, and chlorite); sulfates (gypsum, ferrosulfate, and magnesium sulfate); oxides (various aluminas, spinels, hibonite, magnetite, and hematite); Mg/Fe sulfides (pyrrohtite, troilite, pyrite, and niningerite); carbonate minerals (calcite, aragonite, dolomite, magnesite, and siderite); water ice, clean and with carbon dioxide, carbon monoxide, methane, and ammonia clathrates; carbon dioxide ice; graphitic and amorphous carbon; and 

441: polycyclic aromatic hydrocarbons (PAH).

442: Of these materials, a small, unique subset was found necessary to properly fit the Spitzer data. 

443: Sources for the data included: for {\it silicates},

444:  the Jena Database of Optical Constants for Cosmic Dust\footnote{http://www.astro.uni-jena.de/Laboratory/OCDB} \citep{dorschner,jaeger98,jaeger03}, 

445: the Mars Global Surveyor Thermal Emission Spectrometer database\footnote{http://tes.asu.edu}, and W. Glaccum (private communication), as well as emission spectra 

446: from \citet{koike} and \citet{chihara};

447: for {\it carbonates}, \citet{kemper02};

448: for {\it sulfides}, \citet{keller} and \citet{kimura};

449: for {\it amorphous carbon}, \citet{edoh};

450: for {\it PAH}, \citet{lidraine}. 

451: 

452: We determined the reduced $\chi^{2}$ for the model fit to the data for 

453: thousands of combinations of minerals. All models with reduced chi-squared  values large than the 95\% confidence limit were excluded from consideration. 

454: The range of constituents abundances was determined by varying the amount of a material, and finding where the model exceeded the 95\% confidence limit. 

455: Only constituents with  abundances significantly above zero were included in the final best fit. 

456: Table~\ref{minfittab} shows the composition of the best-fit model. For each entry, 

457: we also show the reduced $\chi^{2}$ for the model if that constituent is deleted.

458: Upper limits for non-detected species are included in Table~\ref{minfittab} based on

459: the 95\% confidence level amplitude of the constituent when it is included in the fit.

460: 

461: 

462: Figure~\ref{specmod} shows the best-fit spectral model for G29-38. 

463: In order of their contribution to

464: the mid-infrared emission, the contributing minerals are

465: amorphous carbon;

466: amorphous olivine;

467: crystalline silicates 

468: ferrosilite (FeSiO$_{3}$),

469: fayalite (Fe$_{2}$SiO$_{4}$), 

470: diopside (MgCaSi$_2$O$_6$), and

471: enstatite (MgSiO$_{3}$);

472: and metal sulfides (Mg$_{10}$Fe$_{90}$S); and

473: water ice.

474: %

475: Figure~\ref{specresid} shows the spectral model after removal

476: of the best-fitting amorphous silicates that dominate the overall

477: spectral shape. The fit to the IRS data (5.2--35 $\mu$m) is excellent,

478: with $\chi^{2}_{\nu}=1.03$.

479: 

480: The composition of the dust surrounding G29-38, as determined by modeling of the Spitzer IRS

481: spectrum, is unusual when compared to circumstellar dust in other environments. 

482: Amorphous olivine is present, as in the other systems that have been modeled with the same technique. 

483: But the relative lack of crystalline olivine, and

484: the strong prevalence of Fe-rich silicates, make the G29-38 material

485: distinct from other systems.

486: The mix of crystalline pyroxenes and amorphous

487: olivines may be indicative of `aged' dusty material. 

488: No PAHs, carbonates, or water gas are

489: seen, reflecting a total lack of primitive nebular

490: material. Some metal sulfides appear to be present, suggesting temperatures as low as $\sim$600

491: K in the observed dust; 600--700 K is the temperature range for vaporization/condensation of

492: ferrosulfides. Given that the best-fit continuum temperature for the spectrum is ~930K

493: (dominated by the short wavelength emission from 0.1 - 1.0 $\mu$m amorphous carbon particles),

494: either the carbon is superheated beyond the temperature of the metal sulfides, which may be

495: present in larger, cooler particles, or there is a distribution of dust locations and effective

496: temperatures, i.e. a disk-like structure. 

497: 

498: \subsection{Water ice} 

499: The detection of water ice emission in the spectrum is curious. 

500: In a vacuum, water ice sublimates at temperatures above 

501: $\sim$200 K. 

502: Thus water ice cannot be in direct physical contact with the rest of the hot dust detected 

503: in the infrared spectrum.

504: Either the dust must be continuously created, or it must reside in a physical location removed from the rest of the dust reservoir. 

505: Since no water vapor is detected at 6 $\mu$m, where there are strong features

506: that should be detected if water were present in significant quantities 

507: \citep{lisse06,woodward07}, 

508: no appreciable ongoing sublimation can be occurring. 

509: Water vapor is ionized within $\sim 10^{3}$ sec,

510: if we scale from the lifetime of water at 1 AU from the Sun

511: %, which is $10^5$ sec 

512: \citep[$\sim 10^6$ s;][]{schleicherahearn84} according to the luminosity of G29-38 ($2\times10^{-3} L_{\odot}$).

513: There would still be a steady state amount of water vapor present

514: if the ice is sublimating.

515: Thus it appears that the water ice is at a large remove from the rest of the warm circumstellar dust, and the water ice is found in the IRS beam (i.e. within 

516: $5\times 10^{14}$ cm [$7\times 10^{5} R_{*}$] of the star).

517: 

518: In paper 1, we showed that a range of temperatures is required to match the

519: photometry out to 24 $\mu$m, with a two-temperature fit having 890 and 290 K color

520: temperatures. Thus we already suspected either a continuous range of dust temperatures or

521: the presence of two dust reservoirs, one hot and one cold.

522: For G29-38, the ice (or ice-coated grains) must be located at 

523: $>10^{12}$ cm ($2000 R_{*}$)from the star, which,

524: while farther than the dust that dominates the mid-infrared spectrum, is still very close to the

525: star. 

526: Cold, icy dust around other stars, at locations comparable to the Solar System's Kuiper Belt,

527: is common, being present around 15--20\% of all stars \citep{bryden06}.

528: The ice detected around G29-38 is  closer to the star than the far-infrared debris disks at 

529: $\sim 100$ AU ($10^{15}$ cm) radius

530: commonly seen around main sequence stars \citep{bryden06}, and the present location

531: of the icy grains would have been within the star

532: when G29-38 was in its red giant or AGB stage.

533: The source of this cold, presumably icy, material is likely to be the same

534: parent bodies that create the dust closer to the star. We will return to the 

535: possible parent body natures in the Discussion section, but we should point out

536: now that presence of H$_{2}$O is well established in asteroids (especially outer belt),

537: meteorites (as water of hydration as well as signatures of aqueously altered mineralogy),

538: comets, and planets. The H$_{2}$O in any of these bodies would freeze if it were

539: liberated by disintegration of these bodies and survive out to $>10^{12}$ cm from the 

540: white dwarf, owing to its small luminosity. 

541: 

542: 

543: \def\extra{

544: THIS MATERIAL LOOKS IRRELEVANT TO G29_-38!!!

545: 

546: Extrapolating the best-fit model spectrum longwards of 35 $\mu$m 

547: using an assumed absorption efficiency 

548: $Q_{abs} = Q_{abs}(35 \mu{\rm m})/\lambda$, 

549: where $\lambda$ is the wavelength, as a lower limit and 

550: $Q_{abs}=$constant as an upper limit, we estimate a 

551: 70 $\mu$m flux of 5 ± 4 mJy, while the Spitzer MIPS measurements of Beichman et al. (2005) yield  a value of 7 ± 3 mJy for the 70 um excess. We thus concur with conclusion that there is no large amount of cold (i.e., Kuiper Belt dust) contributing to the excess emission detected by Spitzer, and that the emission is dominated by the single debris disk at T~ 400 K .

552: }

553: 

554: 

555: \section{Distribution of material around the star\label{modsec}}

556: 

557: Two possible disk models are considered; these are

558: illustrated in Figure~\ref{cartoon} and discussed in

559: turn in the following sections. The models will be fitted

560: to a combination of the {\it Spitzer} spectroscopy and photometry, 

561: the 2MASS photometry, and the near-infrared

562: spectrum, after removal of the photosphere model

563: and scaling to a common epoch (\S2.2).

564: 

565: \subsection{Physically thick disk}

566: 

567: If the optical depth through the dust cloud around G29-38 is optically

568: thin at the wavelengths of the observed emission, then the 

569: spectrum is determined by simply integrating the density distribution,

570: weighted by the Planck function at the local temperature,

571: through the cloud. If the cloud is optically thin at

572: visible wavelengths, where the spectrum of the white dwarf peaks, 

573: then the dust heating is simply determined by the integrated

574: absorption of the star's distance-diluted spectrum by each grain.

575: We can place some constraints on whether such a model can apply. 

576: First, the cloud can only emit as much energy as it absorbs from the star.

577: Let us consider a flattened torus, defined in

578: spherical coordinates $r, \theta$ (with $r$ distance from the 

579: star, $\theta$ the angular separation from the equatorial plane)

580: as having non-zero 

581: density for $R_{1}<r<R_{2}$ and $\theta<\theta_{\frac{1}{2}}$. 

582: Since the dust luminosity is $f=3$\% of the star luminosity, the

583: opening half-angle of the torus must be at least 

584: $\theta_{\frac{1}{2}}>f/2$ radians, i.e. $\theta_{\frac{1}{2}}>0.8^{\circ}$. A thinner torus 

585: simply cannot intercept enough starlight to emit the observed

586: luminosity. (The constraint does not apply to a disk thinner than the

587: radius of the star, nor to a warped disk, as discussed in following sections.)

588: 

589: Of critical importance to the radiative transfer for calculating the emergent

590: spectrum from the disk is its optical depth in the infrared.

591: For a torus with radial mass density profile 

592: $\rho = \rho_{1} (r/R_{1})^{-\alpha}$,

593: the total cloud mass 

594: \begin{equation}

595: M = 4\pi \theta_{\frac{1}{2}} \rho_{1} R_{1}^{3} f_{\alpha},

596: \end{equation}

597: where

598: \begin{eqnarray}

599: f_\alpha &=& \frac{(R_2/R_1)^{3-\alpha}-1}{3-\alpha}  \,\,\,\,\,\,\,\,({\rm for}\,\alpha\neq3)\\

600: &=& \ln(R_{2}/R_{1})  \,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,({\rm for}\,\alpha=3).

601: \end{eqnarray}

602: The vertical optical depth at the inner radius is

603: \begin{equation}

604: \tau_{\perp1} = \frac{3 Q_{IR} M}{16\pi \rho_{d} a R_{1}^{2} f_\alpha}

605: \end{equation}

606: where $a$ is the particle radius and $Q_{IR}$ is the the absorption

607: efficiency averaged over the thermal emission spectrum.

608: Using the cloud mass radius from paper 1 ($R_{1}=1 R_{\odot}$) and

609: scaling the mass in units of $10^{18}$~g, 

610: $\tau_{\perp1} = 0.03 M_{18} Q_{IR} (a/\mu{\rm m})^{-1}$.

611: For particles smaller than the wavelength, the absorption efficiency

612: can be approximated as $Q_{IR}\simeq 2\pi a/\lambda$, 

613: so the optical depth for thermal emission at wavelength $\lambda$

614: is $\tau_{\perp1,\lambda}=0.2 M_{18} (\lambda/\mu{\rm m})^{-1}$,

615: independent of particle size.

616: For thermal emission at 10 $\mu$m, one can neglect radiative transfer out of the disk

617: only if $M<5\times 10^{19}$ g. 

618: %The mass of the optically thin cloud

619: %estimated in paper 1 is well below this limit, as expected.

620: 

621: Dust heating is determined by the propagation of starlight

622: through the disk. The optical depth 

623: from the disk inner boundary to a distance twice as far from the star is

624: \begin{equation}

625: \tau_{\parallel1}= \frac{\tau_{\perp1} Q_{opt}}{\theta_{\frac{1}{2}} Q_{IR}} \frac{1-2^{-\alpha}}{\alpha-1}.

626: \end{equation}

627: For particles larger than 0.1 $\mu$m, $Q_{V}\simeq 1$ for absorption of

628: starlight, and the optical depth from the star to the disk interior is

629: $\tau_{\parallel1}=2.5 M_{18} (\theta_{\frac{1}{2}}/0.8^{\circ})^{-1} a_{\mu{\rm m}}^{-1} (R_1/R_\odot)^{-2}$.

630: %this model disk is optically thin for starlight

631: %propagation if

632: %\begin{equation}

633: %\theta_{\frac{1}{2}} a_{\mu{\rm m}} (R_1/R_\odot)^{-2} > 2^{\circ} M_{18} \,\mu{\rm m}.

634: %\end{equation}

635: Using the mass and radius from paper 1, we find the cloud is optically

636: thin for starlight propagation (and for

637: thermal emission) as long as $\theta_{\frac{1}{2}}a_{\mu{\rm m}}> 2^{\circ} M_{18}$.

638: The presence of the silicate emission feature further requires the particle size $<2 \mu$m, so

639: for a disk that is optically thin to propagation to starlight, we

640: require $M_{18} < \theta_{\frac{1}{2}}$.

641: The cloud must be at least as massive as derived in paper 1 (where light

642: was allowed to propagate through the disk unimpeded), so we finally obtain the constraint

643: $1 < M_{18} < \theta_{\frac{1}{2}}$ for an optically thin disk.

644: If the disk is more massive than the upper limit, it becomes

645: optically thick to propagation of starlight, the temperature 

646: decreases more rapidly with distance from the star, 

647: and the cloud is more dominated by its inner edge. 

648: We will refer to this case as 

649: moderately optically thick, with $0.5<\tau_{\parallel}<10$ so

650: the cloud is optically thin to vertical propagation of 

651: the observed infrared emission but

652: optically thick to radial propagation of starlight (\S\ref{modthicksec}).

653: 

654: \def\extra{

655: The particle size can be constrained somewhat, without resorting to

656: detailed models. In order to produce the 

657: 9--11 $\mu$m silicate

658: emission feature, we require $Q[10\,\mu{\rm m}]<1$, 

659: with the absorption efficiency $Q\simeq 2\pi a/\lambda$ for small particles, so

660: $a_{\mu{\rm m}} < 2$. Larger particles will certainly be present,

661: but the ones producing the mid-infrared emission feature must satisfy 

662: this criterion. 

663: %The mix of sizes is addressed using the full spectral

664: %models as in Paper 1.

665: Particles smaller than 0.1 $\mu$m will both absorb and emit inefficiently. The temperatures of such particles will be boosted over

666: those of greybody grains by a factor 

667: $T/T_{bb}\simeq (\lambda_{IR}/\lambda_{V})^{1/4}\simeq 1.8,$ 

668: where $\lambda_{IR}$ is the (infrared) wavelength of cooling and $\lambda_{V}$

669: is the (visible) wavelength of heating photons. Since the temperature

670: varies with distance from the star as $T\propto r^{-1/2}$ (as long as $\tau_{parallel}<1$), 

671: a cloud of very small particles can be constructed to have color

672: temperature similar to that of a cloud of larger particles,

673: as long as it is larger by a factor 

674: $\sim (\lambda_{IR}/\lambda_{V})^{1/2}\simeq 3$.

675: Such a cloud of submicron particles would show the silicate feature in 

676: emission. It's optical depth for starlight propagation 

677: (constraining the cloud size using the observed color temperature) 

678: would be

679: $\tau_{\parallel1}=3.1 M_{18} (\theta_{\frac{1}{2}}/0.8^{\circ})^{-1}$.

680: Thus a cloud of submicron or few-micron sized particles can account

681: for the emission, with a modest to small optical depth from

682: the star to the edge of the cloud and a small optical depth 

683: emergent from the cloud.

684: }

685: 

686: %The inner radius of an optically thin cloud should be far enough from the star that

687: %grains can survive. For grey grains with a sublimation temperature of 1200 K, the inner radius of %an optically thin cloud around G29-38 is $0.8 R_{\odot} = 74 R_{*}$. 

688: 

689: It appears therefore that a model cloud that is `physically thick' 

690: (i.e. thicker than the white dwarf's diameter)

691: can still be quite flattened, with

692: angular widths as small as $1^{\circ}$, before becoming optically thick. 

693: The distinction between the

694: `physically thick' model \citep{paperone} 

695: and the `physically thin' model \citep{jura07} is

696: that the physically thin model is thinner than the stellar diameter, so that

697: it can be illuminated on its upper and lower surfaces.

698: In order to intercept as much energy as is observed (3\% of the white dwarf's luminosity), 

699: a physically thick model at its inner edge must 

700: have a half-width {\it at least} $10^{9}$ cm, which is larger than the 

701: stellar radius ($\simeq 8\times 10^{8}$ cm).

702: Starlight does not illuminate

703: the surface of the physically thick model, and instead the dust is heated by 

704: stellar photons that must propagate through the disk. 

705: 

706: \subsubsection{Spherical shell model}

707: 

708: The simplest geometry for the cloud around G29-38, and a limiting case for

709: the physically thick model, is a spherical

710: shell. 

711: %An optically thin spherical shell will have the same

712: %emergent spectrum as any optically thin configuration with the same

713: %radial density profile. 

714: A moderately optically thick spherical shell will have the same

715: temperature distribution as a disk with the same radial density variation.

716: We use the spherical shell calculations to measure the

717: radial temperature profile relative to the optically thin case.

718: %For the emergent spectrum, we return to the density distribution,

719: %materials and size-dependent temperatures as in Paper 1.

720: 

721: Spherical shell calculations were performed using DUSTY

722: \footnote{Ivezic, Z., Nenkova, M. \& Elitzur, M., 1999, User Manual for DUSTY, University of Kentucky Internal Report, accessible at http://www.pa.uky.edu/$\sim$moshe/dusty.}

723: for a radial profile $r^{-3}$, 

724: an amorphous carbon or silicate composition,

725: and a range of total optical depths, $\tau$ (at 0.55 $\mu$m),

726: from 0.01 to 10.

727: The inner boundary of the shell is where the dust temperature reaches 1100 K. 

728: (Fits with inner temperature 1000 and 1300 yielded significantly worse fits to the

729: observed spectrum.)

730: Figure~\ref{g29plotshell} shows emergent spectra compared to the observations.

731: The first are visually very good, clearly reproducing the spectral shape, despite 

732: the simplicity of the model;

733: however, the reduced $\chi^2_\nu=2.2$ is statistically poor. The residuals are dominated by

734: structure near the 9--11 $\mu$m silicate feature: the observed minus model residuals have

735: positive peaks at 9.2 $\mu$m and 11.2 $\mu$m and a broad negative trough spanning 8--13 $\mu$m.

736: The deficiencies of this model are thus due to using only one silicate, in contrast

737: to the more detailed mineralogy found in the multi-composition models discussed above.

738: 

739: The luminosity of the shell per unit stellar luminosity

740: is accurately approximated by

741: \begin{equation}

742: \frac{L_{d}}{L_{*}}\equiv f \simeq f_{0}(1-e^{-\tau})

743: \end{equation}

744: where $f_{0}$=0.74 for carbon and 0.57 for silicate grains.

745: The observed luminosity ratio, $f=0.03$, could be explained by the combined

746: carbon and silicate cloud from Figure~\ref{g29plotshell} with optical depth

747: at 5500 \AA\ and 2000 \AA\ of 0.028 and 0.044, respectively.

748: The total extinction toward the star is small but could potentially

749: be measured with precise UV/visible

750: spectrophotometry, in which case presence of carbonaceous grains could be tested by searching

751: for a 2175 \AA\ feature such as seen in the interstellar medium \citep{ref2175}.

752: 

753: The temperature at a given distance from the

754: star is the same for a spherical shell as for a moderately optically thick disk,

755: as long as scattered light and dust thermal emission is a negligible heat source. 

756: Figure~\ref{g29rad} shows the temperature profiles through shells

757: with different optical depths.

758: The temperature versus distance

759: from the star for grey grains in the optically thin limit

760: would follow $T\propto r^{-0.5}$. For real materials,

761: and taking into account radiative transfer, the temperature

762: profiles are significantly different. 

763: Using a power-law approximation $T\propto r^{-\delta}$,

764: the predicted temperature profiles

765: have $0.44 < \delta < 0.48$ as long as $\tau<0.2$. But

766: for higher $\tau$ a single power law is not adequate.

767: %The deviation of the grain temperature from $r^{-0.5}$ is

768: %primarily due to the extinction along the line of sight to the

769: %star; the extinction factor is approximately 

770: %$e^{-R/R_{\tau}}$, where $R_{\tau}$ is the cloud radius beyond which 

771: %$\tau>1$.  

772: We fitted the curves in Figure~\ref{g29rad} 

773: with empirical functions, for use in the moderately optically thick models.

774: \def\extra{

775: Based on this, we found

776: an approximation formula for the deviation of the temperature

777: from the greybody value to be

778: \begin{equation}

779: \frac{T_{car}}{T_{grey}} = (1+2.5\log\tau+0.818\log\tau^{2})

780: e^{-r/62.5 \tau^{0.06}} + (0.99-0.236\log\tau+0.082\log\tau^{2})

781: \end{equation}

782: for carbon grains, and

783: \begin{equation}

784: \frac{T_{sil}}{T_{grey}} = (0.91+2.45\log\tau+1.23\log\tau^{2})

785: e^{-r/R_{\tau}} + (0.90-0.278\log\tau+0.082\log\tau^{2})

786: \end{equation}

787: (where $R_{\tau}=130 R_{*}$ for $\tau<2$ and $R_{\tau}=90 R_{*}$ for

788: $\tau>2$) for silicates.

789: Note that these equations were only fitted to 

790: $0.5 < \tau < 10$ and cannot be extrapolated outside that range. }

791: For reference, the temperature scalings are closely related to the parameter $\Psi$ 

792: defined in the original paper on the self-similarity solution used in DUSTY

793: \citep{ivezic}.

794: 

795: \subsubsection{Moderately optically thick model\label{modthicksec}}

796: 

797: Armed with the temperature profiles, and

798: the wavelength-dependent cross-sections for each mineral, we can now

799: compute the brightness of a flattened cloud

800: that has optical depth $\tau_{\parallel}<10$ and

801: $\tau_{\perp}<1$. 

802: We define this as a `moderately optically thick' disk. 

803: For simplicity we consider a fanned disk, with scale height proportional to 

804: distance from the star, $h=r\tan\theta_{\frac{1}{2}}$.

805: The model does not apply to a thin, flat disk or

806: a disk with a vertical density gradient. 

807: Full treatment of the optically thick disk with

808: vertical density gradient requires a

809: two-dimensional calculation that is beyond our present scope.

810: \def\extra{

811: Linear combinations of a flat disk and an optically thin or moderately

812: optically thick halo can be used to approximate a

813: disk with vertical density gradient; such hybrid models are

814: discussed after the individual models are detailed.

815: }

816: 

817: The moderately optically thick model was calculated for a 

818: subset of the minerals used in \S\ref{mineralsec}:

819: amorphous carbon \citep{zubkocar},

820: amorphous olivine \citep{dorschner},

821: forsterite 

822: \citep[pure-Mg crystalline olivine,][]{jaeger03}, and

823: enstatite \citep[pure-Mg crystalline pyroxene,][]{jaeger98},

824: and bronzite \citep[Fe-rich crystalline pyroxene,][]{henning97}.

825: Based on the results of \S\ref{mineralsec}, we expect the main constituents (in order)

826: to be the amorphous silicates, carbon, and bronzite.

827: Due to the moderately intensive computations and lack of in-hand UV-FIR laboratory data

828: (required to span absorption of the white dwarf spectrum as well as thermal emission), 

829: we did not include as wide a range of minerals as in \S\ref{mineralsec}.

830: For bronzite, we use the calculations for forsterite but scaled 

831: them by the relative small-particle emissivity over the range of wavelengths where the optical constants were available 

832: (6.7--500 $\mu$m).

833: The most

834: important compositions for which we did not calculate the moderately-optically-thick model were magnesium-iron sulfides, and water ice (which in fact

835: cannot exist on grains in this model and requires a separate reservoir, 

836: as discussed in \S\ref{mineralsec}).

837: The temperature of grains of each size (from 0.1 to 1000 $\mu$m)

838: was calculated by balancing radiation from the white dwarf,

839: geometrically diluted by $r^{-2}$ within $20 < r/R_{*} < 70000$,

840: with the grain's thermal emission. 

841: Figure~\ref{tplot} shows the grain temperatures for two materials and three particle sizes in this optically-thin limit.

842: The temperatures were then adjusted

843: using the scale factors derived in the previous section appropriate

844: for the composition, distance from star, and total cloud optical depth. If a grain's temperature exceeds a vaporization temperature, $T_{vap}$, 

845: it's emissivity is set to zero.

846: Values of the vaporization temperature in the 1000-2000 K range are expected for most minerals; evidence for inner edges of YSO disks

847: at distances corresponding to these vaporization temperatures

848: has been found in interferometric observations \citep{monnier}.

849: The emission spectrum of dust at each distance from the star

850: was calculated by integrating over several size distributions: 

851: a power-law $\propto a^{-3.5}$, a Hanner

852: law with slope 3.7 and critical size 7.4 $\mu$m,

853: the size distribution from the coma of comet Halley,

854: and the size distribution of interplanetary meteoroids

855: and lunar microcraters \citep{gruen85}.

856: 

857: The emergent spectrum from the cloud was then calculated by 

858: integrating in spherical coordinates

859: using a radial density

860: distribution $n\propto r^{-\alpha}$ and

861: minimum radius $R_{min}$, with individual models sampled from the ranges

862: $0.3 < \alpha < 6$ and

863: $50 < R_{min}/R_{*} < 1000$.

864: We assume azimuthal symmetry and optically thin infrared emission,

865: so the opening angle of the disk $\theta_{\frac{1}{2}}$ does not

866: affect the spectral shape. 

867: A range of $1000 < T_{vap} < 1600$~K and $0 < \tau_{\parallel} < 10$ were

868: considered.

869: Then for each cloud geometry $(\alpha, R_{min})$, a 

870: linear combination of the models for carbon, olivine, and 

871: forsterite was fitted to the observations, and the $\chi^{2}$ for

872: the mixture computed. 

873: Because these materials vaporize at 

874: $R<50 R_{*}$ for essentially all particle sizes

875: (see Fig.~\ref{tplot}), 

876: all models with $R_{min}<50 R_{*}$ are equivalent.

877: 

878: The best-fitting solution is enumerated in Table~\ref{modthicktab}.

879: %$T_{vap}=1200$ K, $\tau_{\parallel}=5$, $R_{min}\leq 50 R_{*}$, and $\alpha=2.5$. 

880: Given 390 data points in the spectral energy distribution,

881: we expect a `good' fit to have reduced $\chi^{2}_{\nu}\simeq 1$ within dispersion

882: $\sqrt{2/390}=0.07$. The best-fitting model had $\chi^{2}_{\nu}=1.23$.

883: We discuss the residuals, which are

884: localized rather than spanning a wide wavelength range,

885: in the following paragraph;

886: for statistical purposes we assume they are unrelated to the cloud geometry. 

887: The confidence intervals for the parameters, determined by

888: $\Delta\chi^{2}_{\nu}=0.07$, are listed in Table~\ref{modthicktab}.

889: 

890: 

891: The residuals from the radiative transfer model are shown in 

892: Figure~\ref{fitresid}. Coherent features dominate the residuals; 

893: the amount by which they increase $\chi^{2}_{\nu}$ measures their statistical significance.

894: These include the following: 

895: (1) a `W'-shaped pattern between 8 and 10.5 $\mu$m ($\Delta\chi^{2}_{\nu}=0.30$);

896: (2) an `S'-shaped pattern between 15 and 21 $\mu$m ($\Delta\chi^{2}_{\nu}=0.24$); and

897: (3) a `U'-shaped residual between 1 and 3 $\mu$m ($\Delta\chi^{2}_{\nu}=0.20$).

898: Each of these features is statistically significant.

899: Feature (1) is due to mismatching the fundamental silicate Si-O stretch feature;

900: an additional silicate is required which provides a shorter-wavelength silicate feature than

901: Mg-rich olivine.

902: Feature (2) is due to either mismatching the silicate Si-O-Si bending feature or presence

903: of MgO (or Mg$_{x}$Fe$_{1-x}$O with $x>.5$), which has a peak at 18.5 $\mu$m. 

904: Cosmic abundances favor formation of Mg-rich silicates over Mg oxides, so a preferable

905: solution for residual features (1) and (2) is a silicate with bluer fundamental and better-matching 

906: Si-O-Si bending mode features. 

907: 

908: We experimented with the following minerals, inspired by the results of \S\ref{mineralsec}.

909: All models included amorphous carbon and olivine.

910: Additional (third) minerals were included one at a time; the most successful 

911: were bronzite ($\chi^2_\nu=1.22$) and forsterite ($\chi^2_\nu=1.55$), 

912: in order of goodness of fit. These results are

913: consistent with those in \S\ref{mineralsec}, where an Fe-rich pyroxene was

914: shown to be the third-most important mineral.

915: Then for the most successful third minerals, fourth minerals were added; 

916: a good combination was forsterite plus montmorillonite ($\chi^2_\nu=1.23$), 

917: but it was only slightly  better than bronzite model alone. 

918: Other minerals such as amorphous pyroxene yielded negligible improvement to

919: the fit ($\Delta\chi^{2}_{\nu}=0.02$). The materials commonly referred to 

920: as amorphous olivine and amorphous pyroxene are both amorphous silicates and are essentially 

921: indistinguishable.

922: It appears to be significant that Fe-rich pyroxene is more abundant

923: than Mg-rich olivine; this is directly due to

924: Fe-rich pyroxene having a bluer fundamental Si-O stretch (as observed) compared to

925: forsterite.

926: From the laboratory measurements of phyllosilicates by \citet{glotch}, 

927: there are two

928: properties that make them possible explanations for residual features (1) and (2): 

929: phyllosilicates have an

930: Si-O stretch that is bluer than forsterite, and they have a double-peaked Si-O-Si

931: bending mode that is similar to the shape observed in the G29-38 spectrum from 18--20 $\mu$m.

932: 

933: The distinction between the bronzite model and the forsterite+montmorillonite model is

934: mathematically within the errors, but we will refer primarily to the former model

935: since it is simpler and it agrees with the more extensive mineral search 

936: from \S\ref{mineralsec}.

937: The fit could be improved if a more exhaustive mineral search were performed.

938: However, the results of \S\ref{mineralsec} already demonstrate that including

939: a wide range of minerals can explain the silicate feature shape adequately,

940: and we are already at a reduced $\chi^{2}_{\nu}=1.2$ so that the present signal-to-noise

941: does not allow further, unique modeling.

942: \def\extra{

943: Two minerals with significant amplitude in \S\ref{mineralsec} were not included in

944: this section. Sulfides have a broad bump in the 35--40 $\mu$m region 

945: (with that of Mg$_{0.5}$Fe$_{0.5}$S at 37.0 $\mu$m and that of FeS [troilite] at 39.2 $\mu$m);

946: the IRS spectra end at 35 $\mu$m and do not span the red side of that feature, so the presence

947: or state of sulfides in the disk cannot be uniquely determined.

948: We also did not include water ice, which was found to be required in the

949: mineralogical model from \S\ref{mineralsec}. There is still significant H$_2$O in the model

950: in this section, but it is in the form of water of hydration in the phyllosilicate. 

951: }

952: 

953: Feature (3) in the residuals relates to the shape of the inner edge of the disk and the vaporization 

954: temperature. We only explored $T_{vap}$ on a grid with 100 K intervals, and we fixed 

955: it at the same value for all compositions, so the observed, modest deviations from our model 

956: are not particularly surprising in the near-infrared. Further, we did not explore the 

957: composition of the featureless materials, which could affect not only their 

958: $T_{vap}$ but also the slope of their absorption around 3--6 $\mu$m, which determines the

959: shape of the inner edge of the disk spectrum.

960: 

961: 

962: \subsection{Physically thin disk}

963: 

964: A physically thin

965: disk is very {\it optically} thick, so starlight cannot propagate radially through the disk.

966: Such a disk must be so thin that the star can illuminate its surface, or it must be warped such that the surface has clear lines of sight to the star, or a combination of both effects as described by \citet{jura07}.

967: The temperature versus distance from the star scales as $r^{-.75}$ \citep[][e.g.]{chiang}.

968: The flux from an optically thick disk is straightforward to

969: estimate: for G29-38, the disk temperature $T = 8008 (R/R_{*})^{-0.75}$ K,

970: and the stellar radius $R_{*}=8\times 10^{8}$ cm, 

971: %$R_{*}=7.5\times 10^{8}$ cm, 

972: so the model is determined only by the inner and outer radii

973: of the disk. If we furthermore set the inner radius to

974: be that at which dust sublimates, the only free parameter

975: is the outer radius of the disk. 

976: Since emission from

977: the inner radius dominates at the shortest wavelengths,

978: the optically thick model makes robust predictions of 

979: the disk flux at the wavelength where emission from

980: material at the

981: vaporization temperature peaks, i.e. around 3 $\mu$m.

982: The flux at longer wavelengths depends on the outer radius.

983: 

984: As observed from Earth, the disk may of course be inclined with

985: respect to the line of sight \citep{vonhippel}, though previous calculations

986: considered face-on geometry for illustration

987: \citep{jura03,jura07}. The flux will scale as $\cos i$, where

988: $i$ is the angle between the disk axis and the line of sight,

989: until $i \rightarrow \tan^{-1}H/R_{outer}$, where $R_{outer}$ is the

990: outer radius of the disk and $H$ is its scale height. Taking the

991: outer radius $\sim 50 R_{*}$ from the GD 362 

992: model \citep{jura07}, and requiring disk thickness $H<R_{*}$, 

993: the low-inclination limit only applies when $i<1.2^{\circ}$.

994: The nearly edge-on limit is relatively improbable and

995: furthermore would cover the star unless $H<<R_{*}$, in

996: which case the edge-on limit applies to even less probable

997: geometries. 

998: An optically thick disk has a

999: spectral energy distribution determined almost entirely

1000: by the outer radius, and total flux scaling with 

1001: $\cos i$.

1002: 

1003: Using the {\it Spitzer} spectra, it is clear

1004: the simple optically thick model is definitively ruled out

1005: by the presence of a very strong silicate emission feature:

1006: this feature requires an optically thin emitting region.

1007: The feature contributes a significant portion of the disk

1008: luminosity and must have associated continuum. 

1009: Further, the observed spectral energy distribution requires a colder

1010: component with color temperature $\sim 290$ K \citep{paperone} to

1011: explain the flux at 24 $\mu$m.

1012: As a first step toward constraining a possible optically thick disk

1013: around the star, we fitted the thin-disk model to the spectrum excluding

1014: the silicate feature (8--12 $\mu$m), and setting the inner radius as the

1015: location where the grain temperature is 1200 K.  The best fit

1016: has $\chi^{2}_{\nu}=3.8$; this high value is due to the lack of significant 

1017: emergent cold flux from the model.

1018: However we take the constraints on $R_2$ and $i$ as a guide,

1019: with best values $R_{2}/R_{*}=49\pm 5$ and $i=41\pm 3^{\circ}$.

1020: 

1021: Instead of attempting to fit the entire spectrum with the thin-disk model, we

1022: now consider only fitting it to the shorter-wavelength continuum, with an

1023: eye toward adding a cooler, physically-thick silicate-bearing cloud.

1024: Thus we excluded wavelengths longer than  8 $\mu$m and 

1025: fitted the physically thin model to the spectrum of G29-38.

1026: %for a range of $25 < R_{2}/R_{*} < 83$ and $0 < i < 70^{\circ}$.

1027: The near-infrared spectrum, and the decrease from 5--8 $\mu$m, require

1028: the inner edge of the disk is closer to the star than the point that reaches

1029: 1200 K; a better fit is obtained with $T_{vap}=1500$ K so that $R_{1}=9R_{*}$.

1030: Figure~\ref{juramany} shows the constraints on outer radius ($R_{2}$) 

1031: and inclination ($i$). 

1032: The best fit has a low $\chi^{2}_{\nu}=0.8$; it is probably less than 1 due

1033: to overestimation of the uncertainties due to photospheric subtraction in

1034: the near-infrared.

1035: The best fitting outer radius $R_{2}=25R_{*}$ and inclination $23^{\circ}$.

1036: It is notable that the required inclination is roughly in the range

1037: required by \citet{graham} in his model

1038: for the near-infrared timing (which required an inclination

1039: such that the dust temperature pulsations are detectable

1040: while the exciting pulsations on the photosphere are not).

1041: 

1042: An amendment to the physically thin disk model is needed to

1043: improve the fit and explain the silicate emission.

1044: %A thin disk model was earlier applied to G29-38 and found

1045: %to explain the infrared SED with $35<R_{out}<78 R_*$

1046: %\citep{jura03}, similar to the results in the previous paragraph.

1047: %To explain the mid-infrared emission, including silicate emission

1048: %feature, of both G29-38 and GD 362, an optically thin emitting region is

1049: %required. 

1050: One possible solution is to include

1051: an optically thin, effectively-physically-thick region at 

1052: the outer edges of the ring. \citet{jura07} showed that

1053: for GD 362, an extension of the ring that is warped, by only

1054: $\sim 7^{\circ}$, can produce a silicate feature in emission.

1055: Figure~\ref{cartoon} shows their model, compared to the moderately thick model

1056: from \S\ref{modthicksec}.

1057: Indeed this outer region produces both the silicate feature

1058: and a significant fraction of the continuum at 

1059: $\lambda>11$ $\mu$m in their model. The plausibility of

1060: the warp of the outer disk is discussed and justified physically by

1061: \citet{jura07}, who label it as `region III' in their model. 

1062: In many ways, the details 

1063: (mass, temperature, radius, vertical extent)

1064: of this outer

1065: portion of the disk must be similar to the physically

1066: thick model, since they explain the mid-infrared emission

1067: in the same way (an optically thin cloud of silicates). In the

1068: warped disk model, the optically thin region is actually the

1069: upper layer of the disk; i.e. even the warped portion of

1070: the disk may be optically thick, as long as it is has a direct

1071: line of sight to the star.

1072: 

1073: \subsection{Comparing the thick and thin models}

1074: 

1075: The disk spectral energy distribution can be empirically 

1076: decomposed into three major components. One of them is

1077: a  continuum with a relatively hot (890 K) color temperature, peaking

1078: around 4 $\mu$m and dominating the near-infrared emission.

1079: Another component is continuum with a lower color temperature ($\sim 300$ K).

1080: The other major component is the silicate emission feature.

1081: In the \citet{jura07} model, the hot component is the

1082: blackbody disk---hot because it is close to the star---and the

1083: cooler continuum and silicate feature arise in the outer warp region III.

1084: In the physically thick model (in Paper I and \S\ref{modthicksec} above), 

1085: the hot component is

1086: amorphous carbon---hot because it is due to highly absorbing

1087: material---and the cooler component is silicates---cool because the

1088: silicates are more transparent and have strong mid-infrared emission

1089: features that allow them to cool efficiently;

1090: the two materials are colocated.

1091: 

1092: For many plausible configurations of solid material around

1093: the white dwarf, we can consider the cloud as the sum of an

1094: optically thick disk and an optically thin halo or flared disk surface. 

1095:  Indeed, \citet{vinkovic} proved 

1096: mathematically that flared disk models are equivalent

1097: to disk plus halo models. The radial

1098: profile of a spherical halo can be directly related

1099: to the flaring angle of a flared disk. Thus it is not

1100: possible, using the spectral energy distribution alone,

1101: to separate disk and halo (or warped disk) models.

1102: The halo dominates the infrared emission when 

1103: $\tau_{halo}>H/4R$, where $H$ is the flare height and

1104: $R$ the distance from the star \citep{vinkovic}. 

1105: For the wedge-shaped 

1106: `physically thick' model discussed above, this is

1107: equivalent to $\tau_{\parallel}>\tan\theta_{\frac{1}{2}}/4$. 

1108: Equating the emission and absorbed flux (3\% of the stellar

1109: flux), the halo will dominate the infrared emission

1110: when $\tau_{\parallel}>0.004$, which is already required

1111: for both the physically thick and thin models.

1112: 

1113: For the specific case of the G29-38 disk, we made a direct substitution of the

1114: optically-thick disk for the C/Fe component of the moderately-optically-thick model.

1115: The optically-thick disk was taken directly from the model in Figure~\ref{jurabest} (i.e.

1116: the one fitted to wavelengths shorter than 8 $\mu$m), and the silicate components

1117: were taken from the moderately optically thick model in Figure~\ref{fplotbest}. 

1118: The silicate component of the moderately-optically-thick disk was rescaled in 

1119: amplitude in order to match the observed flux

1120: after being added to the optically thick disk. 

1121: The parameters for the best-fitting disk in this combined disk+silicate model are

1122: similar to those derived above using only short-wavelength data; $R_{2}=22R_{*}$ and

1123: $i=29^{\circ}$.

1124: Figure~\ref{thinthick} shows the best fit;

1125: it is very similar to the moderately optically thick model in Figure~\ref{fplotbest}; 

1126: the goodness-of-fit, $\chi^{2}_{\nu}=1.40$, not much worse than 

1127: the moderately optically thick model. Improvements to $\chi^{2}_{\nu}$ as obtained

1128: in \S\ref{mineralsec} could be obtained by including more minerals in the optically thin

1129: region.

1130: 

1131: \section{Timing constraints\label{varsec}}

1132: 

1133: The geometry of the G29-38 disk can been constrained 

1134:  using timing information.

1135: The star is a ZZ Ceti variable with non-radial modes that yield

1136: its optical pulsations. 

1137: \citet{graham} found

1138: pulsations at periods $\sim 200$ sec in the K and L bands that have 

1139: no strong counterpart in J and B-bands.

1140: More detailed optical photometry clearly shows pulsations at

1141: the predicted frequency: see the peak around 5380 $\mu$Hz

1142: from the Whole Earth Telescope observations

1143: \citep{winget}. The weakness of this mode at visible wavelengths,

1144: compared to infrared wavelengths, could be due to the manner in 

1145: which it interacts with the disk.

1146: The B-band light is completely dominated by direct photons from

1147: the star, while the K-band light contains a contribution from

1148: thermal emission by the dust that produces the mid-infrared excess.

1149: The K-band pulsations are attributed to dust temperature variations.

1150: 

1151: As \citet{graham} explained,

1152: to generate observable K and L-band pulsations from a mode that is

1153: not prominent on the photosphere,

1154: the geometry is constrained. 

1155: They invoke a mode of

1156: stellar brightness variations (2$^{nd}$ order spherical harmonic) that

1157: brightens the pole and dims the equator of the star.

1158: If the disk is thin enough that its heating is driven by the equatorial

1159: stellar brightness, then it will track the equatorial photospheric

1160: temperature, rather than the average over the surface.

1161: \citet{graham} show that the spherical harmonic that excites the

1162: disk temperature variations can exist without significant variation

1163: of optical wavelength light if the star is viewed at the

1164: angle where the ratio of the two spherical harmonics is small.

1165: To eliminate J-band variability in their data required a viewing 

1166: angle in the range 45--65$^{\circ}$. 

1167: 

1168: The IRAC observations confirm the significance of the 200 sec period

1169: in the near-infrared. (No modes with periods longer than 500 s are detectable

1170: due to the brevity of the IRAC observations, so

1171: we could not address the 615 sec period that dominates the optical pulsations \citep{winget}.)

1172: The fluctuating portion of the infrared flux, after

1173: subtracting the direct stellar photosphere, is 7\% at 2.2 $\mu$m,

1174: 5\% at 3.6 $\mu$m, and $<3$\% at 8 $\mu$m. The color of the

1175: fluctuations suggests they arise from the inner portion of the

1176: disk, near the dust vaporization temperature.

1177: This explanation works well with the physically thin disk model

1178: with the inner edge of the disk at 1200 K and decreasing temperature

1179: outwards.

1180: The timing constraints rule out a spherical distribution for all 

1181: the dust, though some dust can remain in a spherical distribution

1182: without generating pulsations (because the temperature variations

1183: average out over a sphere).

1184: The observations do not yet rule out a

1185: `physically thick' configuration with a small opening angle ($\theta_{\frac{1}{2}}$).

1186: We consider the \citet{graham} results very

1187: important; they require observational confirmation

1188: and have the promise of revealing the disk structure in

1189: more detail. 

1190: 

1191: 

1192: 

1193: \section{Discussion and Conclusions\label{discussion}}

1194: 

1195: The mid-infrared spectrum of G29-38 is due to a cloud of small particles orbiting the

1196: star within the distance where tidal forces from white dwarf's strong gravity

1197: would destroy a large, weak body. The tidal forces exceed self-gravity and

1198: strength of a non-rotating, rigid spherical body within a distance

1199: \begin{equation}

1200: d/R_{*} = 120 \rho^{-1/3} \left[ 1 + 0.11 \frac{S_{\rm kPa}}{\rho r_{10}^{-2}} \right]

1201: \label{roche}\end{equation}

1202: where $\rho$ is the bodies density (g~cm$^{-3}$), $r_{10}$ is its radius (in units of

1203: 10 km), and $S_{\rm kPa}$ is its strength in kPa. 

1204: If the dust around G29-38 was produced by one body, then the present-day mass 

1205: requires $r_{10}>1$. 

1206: For an asteroid or comet, we expect $0.5<\rho<2$. 

1207: %A negligible strength (0.006 kPa) was inferred from the tidal splitting of comet

1208: %Shoemaker-Levy 9 \citep{asphaug}, but this only applies to the binding between

1209: %major units in that comet.

1210: %Many asteroids and comets are believed to be rubble piles \citep{rubble}, so

1211: %the binding strength can be much higher for smaller-scale units.

1212: Spin rates of near-Earth objects require strengths $\sim 2 r_{10}^{-1/2}$ kPa

1213: to balance centrifugal forces against strength \citep{holsapple}.

1214: The strength term in equation~\ref{roche} is thus negligible for

1215: bodies with $r_{10}\rho>1$, which is likely for the parent bodies of the

1216: G29-38 dust. 

1217: Fragments of parent bodies can survive much closer to the star;

1218: rocks with $S_{kpa}=3000$, $\rho=3$, and could survive against tidal

1219: disruption all the way to the

1220: surface of the star if they are smaller than 0.1 km.

1221: 

1222: In the spectrum of G29-38, the strong emission feature at 9--11 $\mu$m, the color temperature, and the timing information all support the interpretation of the mid-infrared excess as being dominated by a cloud of small silicate particles. The infrared excess at 2--6 $\mu$m is apparently due to a featureless blackbody continuum. Based on our modeling results, this higher-color-temperature emission could be explained either by a highly-absorbing mineral (like solid C or Fe) or by a massive disk of material that is thinner than the white dwarf's diameter. 

1223: The difference between these models, both of which can explain the observed spectral energy distribution reasonably well, is significant: the massive disk model could harbor 

1224: $10^{24}$ g (or more) of dust \citep{jura03}, while the cloud of amorphous C or Fe requires $\sim 10^{19}$ g. There is no {\it a priori} reason that either of these mass estimates should be preferred or rejected. Some $10^{19}$ g of silicates are required in all models. 

1225: 

1226: To understand the origin of the circumstellar material, it is important

1227: to know how much mass is observed. If only $10^{19}$ g of material is present, then the observed infrared excess can be explained by a single small asteroid or a comet. 

1228: If $10^{24}$ g of material is present, then an entire, large asteroid, or numerous

1229: smaller ones, is required. In all cases, the parent body must have been somehow transported from a distance far enough from the star that it could have survived the red giant and AGB phase of the star ($>5\times 10^{13}$ cm) inward to the Roche limit where it would be disrupted by the gravity of the white dwarf ($< 10^{11}$ cm). 

1230: 

1231: \subsection{Nature of the parent bodies}

1232: The mineralogical results can help us relate the dust to possible parent bodies.

1233: We concentrate on the two primary components of the spectrum separately: the silicate material (required in all models in order to explain the 9--11 $\mu$m emission feature) and the highly-absorbing material (required only in the thick-disk models).

1234: 

1235: The composition  of the highly-absorbing material cannot be determined unambiguously from the spectrum. We fitted it in \S\ref{mineralsec} with amorphous C, based on the high cosmic abundance of C. But solid Fe or Si are also plausible, given their high cosmic abundance. 

1236: Indeed, for chondritic (asteroidal and terrestrial) material the abundance of Si and Fe is

1237: much higher than C \citep{juracarbon}.

1238: Mineralogical models for dust around other stars and in comets, using the same methodology as in \S\ref{mineralsec}, do not always show a high abundance of C 

1239: \citep{lisse07a,lisse07b}. For extrasolar systems, a careful subtraction of the photosphere is critical to measuring the high-color-temperature emission (which is characteristic of highly-absorbing material like solid C). 

1240: For G29-38, the high-color-temperature excess (at 3.6 $\mu$m and

1241: longer wavelengths) is so far above photosphere that it must arise from 

1242: circumstellar material, but

1243: at wavelengths shorter than 2 $\mu$m the infrared excess shape depends on the photosphere model. 

1244: %Near-infrared photometry for white dwarfs has been found to deviate significantly, sometimes in the form of a significant {\it deficit} of emission that cannot be explained by a dust cloud 

1245: %\citep{kilicWDphot}.

1246: For G29-38 the mid-infrared data clearly require emission with a color temperature $> 800$ K, whether it is C or Fe.

1247: 

1248: The compositions of the potential parent bodies for the circumstellar material around white dwarfs can be addressed by studies of Solar System bodies.

1249: To date there has been no sample return mission from an asteroid, but meteorites provide direct measures of composition of parts of at least some asteroids. Carbonaceous chondrites 

1250: have some C, but all chondrites are largely silicate mineral (olivine and pyroxene, 

1251: mostly Mg-rich [forsterite and enstatite]), 

1252: with a wide range of other minerals (some Ca and Al-rich)

1253: and metals (often including previously-molten Fe).

1254: Metallic meteorites, commonly found in museums and on the ground, have largely Fe and Ni composition \citep{shearer}.

1255: Comets are likely to have a more primitive composition than carbonaceous chondrites, with 

1256: abundant silicate grains as well as carbonaceous material, based on infrared spectroscopy of cometary dust, laboratory study of cometary interplanetary dust particles, and {\it in situ} mass spectrometry during the 1P/Halley flyby in 1986 \citep{hannerbradley}. 

1257: The most abundant silicate minerals in meteorites are Mg-rich olivines and pyroxenes, 

1258: as well as feldspar and phyllosilicates. Interplanetary dust particles believed

1259: to originate from comets (CP type) are largely composed of phyllosilicates that

1260: require aqueous alteration on their parent body \citep{messenger}.

1261: 

1262: Based on the analogy to Solar System bodies, the dominance of Fe-rich pyroxene mineralogy is

1263: distinct. Comets or outer-main-belt (D-type) asteroids contain Mg-rich olivine and phyllosilicates,

1264: which when combined can reasonably fit the observed spectrum of G29-38.

1265: Both comets and D-type asteroids contain organic material, which would be consistent

1266: with the presence of amorphous C, and the most primitive carbonaceous (CI) 

1267: meteorites are largely composed of phyllosilicates.

1268: But the Mg-rich olivine plus phyllosilicate model is not as good a fit to the data as Fe-rich

1269: pyroxene. There are pyroxene-dominated meteorites, but they are dominated by Mg-rich pyroxene

1270: (hence the name enstatite meteorites for this rare class).

1271: It is worth noting that enstatite chondrites are thought to have formed in reducing conditions and contain

1272: other minerals including niningerite \citep{weisberg}, which was one of the most abundant minerals from

1273: our fit for G29-38 in Tab.~\ref{minfittab}. Thus the physical conditions for formation of the

1274: enstatite meteorite parent bodies may have some relevance to the formation of the G29-38 debris

1275: parent body.

1276: 

1277: In terms of the featureless material that produces the near-infrared continuum (in the physically

1278: thick models), either C or Fe are acceptable to the fits. If the material were C, then a

1279: cometary or D-type asteroid origin would be more likely, whereas with abundant Fe, formation

1280: closer to the star and within a differentiated parent body would be implicated.

1281: 

1282: Some constraints on the composition of the material are obtained from the abundances in the white dwarf atmosphere. 

1283: Solid material at the inner edge of the disk is constantly being vaporized by the stellar radiation. These vapors reach the atmosphere of the star (or are blown out of the system). 

1284: They cannot reside in the stellar atmosphere for long; instead, they diffuse 

1285: rapidly inward, deeper than the photosphere. 

1286: Thus the heavy elements in the stellar atmosphere must be `fresh,'

1287: consistent with an origin from vaporization of circumstellar dust but not with a long-lived stellar atmosphere. 

1288: G29-38 has metals present in its atmosphere.

1289: \citet{juracarbon} discussed the deficiency of C in some externally polluted

1290: white dwarfs, where the abundance of C relative to Fe is more than

1291: 10 times lower than solar. 

1292: CI meteorites have C/Fe 10 times lower than solar; this is commonly

1293: explained by the volatility of C and the high temperature required for chondrite formation

1294: \citep{brearley}.

1295: We note that the abundances of refractory elements in comets

1296: {\it and} asteroids as well as that inferred from exozodiacal dust

1297: is consistently less than solar, in the 7--10 \% range 

1298: \citep{lisse06}. But in a relative abundance, dust collected during the encounter 

1299: with the long-period comet 1P/Halley has

1300: C/Fe abundance ratio similar to Solar \citep{halley}. 

1301: Short-period comets (which are periodically heated to higher temperature) may be more devolatilized and may have abundances more similar to asteroidal material. 

1302: %G29-38 is C-deficient, which favors an asteroidal origin 

1303: %over cometary material that has been maintained at $T<200$ K. 

1304: 

1305: %geometry: dynamical argument on disk flattening

1306: 

1307: \subsection{Disruption of `Hot Jupiter'?}

1308: One intriguing possibility for the origin of the infrared excess is the survival of the core of a giant planet and its subsequent gravitational disruption. 

1309: `Hot Jupiters,' with masses of order

1310: $10^{30}$ g orbiting within $10^{12}$ cm (0.1 AU) of their star,

1311: appear to be fairly common in extrasolar planetary systems: 1.2\% of nearby F, G, and K stars

1312: has such a planet \citep{hotjupiter}.

1313: In this scenario, the planet would become engulfed into a common atmosphere during the red giant phase. Drag from the extended stellar atmosphere would cause the planet to spiral inward toward the stellar core. At the end of the mass-losing phase of the star's evolution, we would be left with a white dwarf, the surviving core of the planet, and the planetary nebula composed of the outer atmosphere from the star. 

1314: 

1315: If a planet began at $<10^{12}$ cm from the star, it is possible

1316: that its remnant could land within the Roche radius. 

1317: The effect of the post-main sequence evolution on the planetary dynamics has not been explored in detail.

1318: The change in mass of the central star could make any borderline-unstable system of

1319: multiple planets unstable \citep{debes02} and could lead to nonlinear orbital perturbations.

1320: A simple estimate of the orbital decay due to gas drag when the planet is within the red

1321: giant atmosphere is made by setting the rate of kinetic energy imparted to the planet,

1322: \begin{equation}

1323: \dot{E} = \frac{1}{2} \rho \pi R_p^2 \left(\frac{GM_*}{a}\right)^{3/2}

1324: \end{equation}

1325: equal to the change in orbital binding energy

1326: \begin{equation}

1327: \dot{U} = \frac{G M_* M_p}{a^2} \frac{{\rm d}a}{{\rm d}t},

1328: \end{equation}

1329: where $M_p$ and $R_p$ are the mass and radius of the planet, $G$ is the gravitational constant,

1330: $a$ is semimajor axis of the presumed-circular orbit, 

1331: $M_*$ is the mass of the star, and $\rho$ is the mass density of

1332: the star at the distance of the planet. Taking for illustration a planet with the mass of Jupiter

1333: and average density 1 g~cm$^{-3}$,

1334: orbiting at a distance of 0.1 AU from a star of mass $2 M_{\odot}$ that has mass density

1335: $10^{-9}$ g~cm$^{-3}$ at 0.1 AU from its center, the timescale for orbital decay 

1336: is $6\times 10^{4}$ yr. This is much smaller than the asymptotic giant branch lifetime 

1337: $\sim 10^{7}$ yr \citep{vassiliadis} for a progenitor star mass of $3.1 M_{\odot}$ \citep{weidemann}.

1338: Thus the inward spiral of hot Jupiter (or other close-in) planetary orbits is likely to be rapid, once

1339: the star has entered its giant phase.

1340: 

1341: 

1342: % Soker & Livio 1984: considered only planets much further from star (5 AU and out), but

1343: % showed they do accrete as well as have mass loss and end up at different distances and sizes

1344: 

1345: Considering only a convective main-sequence primary star, \citet{sasselov} showed

1346: that the orbit of a `hot Jupiter' like OGLE-TR-56b would tidally decay on a timescale of 1--10 Gyr.

1347: Again considering only a main sequence primary star, 

1348: \citet{baraffe} showed that planets with masses less than

1349: $3\times 10^{30}$ will evaporate down to a rocky core in less than 5 Gyr

1350: These effects may combine, with tidal decay enhancing evaporation \citep{erkaev}.

1351: Tidal decay of `hot Jupiter' orbits cannot be too rapid for main sequence stars, or 

1352: else they would be much less common. 

1353: %Thus, even though the closest-orbiting planets the timescales for tidal orbital may be

1354: %comparable to the main sequence lifetime...

1355: The orbital evolution is very sensitive to 

1356: the stellar structure and the planet's orbital eccentricity; the response is

1357: highly nonlinear and may lead to a rapid inward spiral upon perturbation of

1358: the stellar interior structure or eccentricity \citep{ogilvie,jackson08}

1359: Nonetheless, regardless of whether tidal decay can bring planets close to the stellar

1360: surface during main sequence evolution, the gas drag during the red giant phase 

1361: should finish the job.

1362: 

1363: How much of a `hot jupiter' could survive the late stages of its star's evolution is not known.

1364: A brown dwarf was recently discovered around a white dwarf suggesting that some companions can 

1365: survive the red giant phase of the primary star despite being well within the atmosphere

1366: \citep{maxted}. Indeed, the companion is within the Roche distance of the white dwarf

1367: unless it has a mass density greater than 3.6 g~cm$^{-3}$ or is held together by more than

1368: its own gravity.

1369: The composition of the core of an extrasolar giant planet is  difficult to constrain. Mass and radius estimates seem to require a dense core, possibly of high-pressure ice or stony composition 

1370: \citep{burrows07}. This material would be tidally disrupted, as in

1371: the asteroid (or comet) disruption models considered to date. 

1372: The Fe-rich pyroxene mineralogy for the G29-38 debris, which is distinct from that

1373: of most comets and asteroids, does not appear impossible for a rocky planetary core, but

1374: at present there is little more than can be said about whether such a mineralogy is likely.

1375: 

1376: That G29-38 is exceptional

1377: among debris disks (having by far the brightest infrared excess, with 3\% of the star's

1378: luminosity absorbed and radiated by dust)

1379: makes the hypothesis more attractive. Consider the reverse argument: if

1380: 1\% of all stars have `hot jupiters,' what is their expected end state? From this point

1381: of view, it seems inevitable that such remnant planets would generate debris for their

1382: parent star in its white dwarf phase. But this argument assumes that `hot jupiters' 

1383: exist around A-type stars (the progenitor type for G29-38), at least part of the planet can

1384: survive the common envelope phase, the remnant core lands within the Roche radius, and

1385: the debris can survive 500 Myr (the age of G29-38). 

1386: The age problem may not be severe, if 

1387: the remnant planetary core becomes fragmented, with fragments gradually entering the

1388: Roche radius due to collisional disruption and gravitational perturbations.

1389: Further theoretical work is needed to test the viability of this scenario.

1390: 

1391: 

1392: \acknowledgements  

1393: 

1394: WTR gratefully acknowledges discussions on radiative transfer in disks with Moshe Elitzur in May 2007.

1395: WTR gratefully acknowledges discussions on extrasolar giant planets with Adam Burrows in June 2007.

1396: This work is based in part on observations made with the {\it Spitzer Space

1397: Telescope}, which is operated by the Jet Propulsion Laboratory, California

1398: Institute of Technology under NASA contract 1407. 

1399: %Support for this work was

1400: %provided by NASA through award Project NBR: 1269551

1401: %issued by JPL/Caltech to the University of Texas.

1402: 

1403: %\clearpage

1404: 

1405: \begin{thebibliography}{} 

1406: 

1407: %\bibitem[Alcock et al.(1986)]{alcock} Alcock, C., Fristrom, C. C., \& 

1408: %Siegelman, R. 1986, \apj, 302, 462

1409:  

1410: %\bibitem[Alcock \& Illarionov(1980)]{alcock80} 

1411: %Alcock, C., \& Illarionov, A. 1980, ApJ, 235, 534

1412: 

1413: \bibitem[Backman \& Paresce(1993)]{backman}

1414: Backman, D. E., \& Paresce, F. 1993, in Protostars and Planets III, ed. E. H. Levy \& J. I. Lunine (Tucson: Univ. Arizona Press), 1253

1415: 

1416: \bibitem[Baraffe et al.(2004)]{baraffe} Baraffe, I., Selsis, F., Chabrier, G., Barman, T. S., Allard, F., Hauschildt, P. H., Lammer, H. 2004, \aap, 419, L13

1417: 

1418: \bibitem[Beichman et al.(2006)]{beichman06a} 

1419: Beichman C. A., Tanner, A., Bryden, G., Stapelfeldt, K. R., Werner, M. W.,

1420: Rieke, G. H., Trilling, D. E., Lawler, S., Gautier, T. N. 2006, \apj, 639, 1166

1421: 

1422: %\bibitem[Bergeron et al.(1995a)]{postAGBage} Bergeron, P., Wesemael, F., 

1423: %\& Beauchamp, A. 1995, \pasp, 107, 1047

1424: 

1425: \bibitem[Bergeron et al.(2004)]{bergeron04} Bergeron, P., Wesemael, F., 

1426: \& Beauchamp, A. 1995, \pasp, 107, 1047

1427: 

1428: %\bibitem[Bergeron et al.(1995b)]{bergeron95} Bergeron, P.,

1429: %Fontaine, G., Bill\`res, M., Boudreault, S., Green, E. M. 2004, \apj, 600, 404

1430: 

1431: \bibitem[Bradley(2002)]{bradley02} Bradley, J. P. 2002, Highlights Astron., 12, 34

1432: 

1433: \bibitem[Brearley(1998)]{brearley}

1434: Brearley, A. 1998, in {\it Planetary Materials}

1435: (Reviews in Mineralogy Vol. 36), ed. J. J. Papike (Washington: Mineralogical 

1436: Society of America)

1437: 

1438: \bibitem[Bryden et al.(2006)]{bryden06}

1439: Bryden, G. et al. 2006, \apj, 636, 1098

1440: 

1441: \bibitem[Burrows et al.,(2007)]{burrows07}

1442: Burrows, A., Hubeny, I., Budaj, J., Hubbard, W. B. 2007, \apj, 661, 502

1443: 

1444: %\bibitem[Chary et al.(1999)]{chary} Chary, R., Zuckerman, B., \& 

1445: %Becklin, E. E. 1999, in {\it The Universe as Seen

1446: %by ISO}, eds. P. Cox \& M. F. Kessler, ESA-SP 427, p. 289

1447: 

1448: \bibitem[Chen et al.(2006)]{chen06} Chen, C. H., Sargent, B. A., Bohac, C.,

1449: Kim, K. H., Leibensperger, E., Jura, M., Najita, J., Forrest, W. J.,

1450: Watson, D. M., Sloan, G. C., Keller, L. D. 2006, \apjs, 166, 351

1451:  

1452: \bibitem[Chiang \& Goldreich(1997)]{chiang} Chiang, E. I.,

1453: Goldreich, P. 1997, \apj, 490, 368

1454: 

1455: \bibitem[Chihara et al.(2002)]{chihara} Chihara, H., Koike, C., Tsuchiyama, A., Tachibana, S., Sakamoto, D. 2002. \aap, 391, 267

1456: 

1457: %\bibitem[Crovisier et al.(1997)]{crovisier} Crovisier, J.,  Leech, K.,

1458: %Bockelee-Morvan, D., Brooke, T. Y., Hanner, M. S., Altieri, B., 

1459: %Keller, H. U., Lellouch, E. 1997, Science, 275, 1904

1460: 

1461: \bibitem[Debes \& Sigurdsson(2002)]{debes02} Debes, J. H., \& 

1462: Sigurdsson, S. 2002, \apj, 572, 556

1463: 

1464: \bibitem[Dohnanyi(1969)]{dohnanyi} Dohnanyi, J. W. 1969, JGR, 74, 2531

1465: %Collisional models of asteroids and their debris

1466: 

1467: \bibitem[Dorschner et al.(1995)]{dorschner}

1468: Dorschner, J., Begemann, B., Henning, T., Jaeger, C., Mutschke, H.

1469: 1995, \aap, 300, 503

1470: 

1471: %\bibitem[Dupuis et al.(1993)]{dupuis} 

1472: %Dupuis, J., Fontaine, G., \& Wesemael, F.  1993, \apjs, 87, 345

1473: 

1474: \bibitem[Durda and Dermott(1997)]{durda97} Durda, D., Dermott, S. F. 1996, Icarus, 130, 140 

1475: 

1476: \bibitem[Edoh(1983)]{edoh} Edoh, O. 1983, Ph. D. thesis, Univ. Arizona

1477: 

1478: \bibitem[Erkaev et al.(2007)]{erkaev} Erkaev, N. V., Kulikov, Yu. N., Lammer, H.,

1479: Selsis, F., Langmayr, D., Jaritz, G. F., Biernat, H. K. 2007. \aap, 472, 329

1480: 

1481: \bibitem[Fazio et al.(2004)]{fazio}

1482: Fazio, G. G. et al. 2004, \apjs, 154, 10

1483: 

1484: %\bibitem[Fixsen \& Dwek(2002)]{fixsendwek}

1485: %Fixsen, D. J., \& Dwek, E. 2002, ApJ, 578, 1009

1486: 

1487: \bibitem[Glotch et al.(2007)]{glotch}

1488: Glotch, T. D., Rossman, G. R., Aharanson, O. 2007, {\it Icarus}, 192, 605

1489: 

1490: \bibitem[Graham et al.(1990)]{graham}

1491: Graham, J. R., Matthews, K., Neugebauer, G., \& Soifer, B. T. 1990, \apj, 357, 216

1492: 

1493: %\bibitem[Green et al.(1986)]{green}

1494: %Green, R. F., Schmidt, M., and Liebert, J., 1986, \apjs, 61, 305

1495: 

1496: \bibitem[Greenstein \& Liebert(1990)]{greenstein} Greenstein, J. L,

1497: \& Liebert, J. W. 1990, \apj, 360, 662

1498: 

1499: \bibitem[Gr\"un et al.(1985)]{gruen85}

1500: Gr\"un, E., Zook, H. A., Fechtig, H., \& Giese, R. H. 1985. Icarus, 62, 244

1501: 

1502: \bibitem[Hanner \& Bradley(2004)]{hannerbradley}

1503: Hanner, M. S., Bradley, J. P. 2004, in {\it Comets II},

1504: eds. M. C. Festou, H. U. Keller, H. A. Weaver,

1505: Tucson: U. Arizona Press, p. 555

1506: 

1507: %\bibitem[Harrington \& Dahn(1980)]{usno}

1508: %Harrington, R. S., \& Dahn, C. C. \aj, 85, 454

1509: %Riddle,R.K., 1970,Pub.U.S.Naval Obs., Second Series,20,PART 3.

1510: 

1511: \bibitem[Henning and Mutschke(1997)]{henning97} Henning, Th., Mutschke H. 1997. \aap, 327, 743

1512: 

1513: %\bibitem[Holberg et al.(2003)]{holberg} Holberg, J. B., Barstow, M. A.,

1514: %\& Burleigh, M. R. 2003, \apjs, 147, 145

1515: 

1516: \bibitem[Holberg and Bergeron(2006)]{holberg06} 

1517: Holberg, J. B., Bergeron, P. 2006, \aj, 132, 1221

1518: 

1519: \bibitem[Holsapple(2007)]{holsapple}

1520: Holsapple, K. A. 2007.

1521: %Spin limits of Solar System bodies: From the small fast-rotators to 2003 EL6

1522: Icarus, 187, 500

1523: 

1524: \bibitem[Houck et al.(2004)]{houck}

1525: Houck, J. R. et al. 2004, \apjs, 154, 18

1526: 

1527: \bibitem[Ivezi\'c and Elitzur(1997)]{ivezic}

1528: Ivezi\'c, \v{Z}., Elitzur, M. 1997, \mnras, 287, 799

1529: 

1530: \bibitem[Jackson et al.(2007)]{jackson08} Jackson, B., Greenberg, R., Barnes, R. 2008, \apj, 678, 498

1531: 

1532: \bibitem[J\"ager et al.(1998)]{jaeger98}

1533: J\"ager, C., Molster, F. J., Dorschner, J., Henning, Th.,

1534: Mutschke, H., Waters, L. B. F. M. 1998, \aap, 339, 904

1535: 

1536: \bibitem[J\"ager et al.(2003)]{jaeger03}

1537: J\"ager, C., Dorschner, J., Mutschke, H., Posch, Th., \&

1538: Henning, Th. 2003, \aap, 408, 193

1539: 

1540: \bibitem[Jessberger et al.(1988)]{halley}

1541: Jessberger, E. K., Christoforidis, A., \& Kissel, J. 1988, Nature, 332, 691

1542: 

1543: \bibitem[Jura(2003)]{jura03} Jura, M. 2003, \apjl, 584, L91

1544: 

1545: \bibitem[Jura(2006)]{juracarbon} Jura, M. 2003, \apjl, 653, 613

1546: 

1547: \bibitem[Jura(2007)]{jura07} Jura, M., Farihi, J., Zuckerman, B.,

1548: Becklin E. E. 2007, \apj, 133, 1927

1549: 

1550: \bibitem[Kalas et al.(2005)]{kalas05} Kalas, P., Graham, J. R., Clampin, M. 2005, Nature, 435, 1067

1551: 

1552: \bibitem[Kalas et al.(2006)]{kalas06} Kalas, P., Graham, J. R., Clampin, M. C., Fitzgerald, M. P. 2006, \apjl, 637, L57

1553: 

1554: \bibitem[Keller et al.(2002)]{keller} Keller, L., Hony, S., Bradley, J. P., Molster, F. J.,

1555: Waters, L. B. F. M., Bouwman, J., de Koter, A., Brownlee, D. E., Flynn, G. J.,

1556: Henning, T., Mutschke, H. 2002, Nature, 417, 148

1557: 

1558: \bibitem[Kemper et al.(2002)]{kemper02} Kemper, F., J\"ager, C., Waters, L. B. F. M.,

1559: Henning, Th., Molster, F. J., Barlow, M. J., Lim, T., de Koter, A. 2002, Nature, 415, 295

1560: 

1561: \bibitem[Kemper et al.(2004)]{kemper} Kemper, F., Vriend, W. J., \&

1562: Tielens, A. G. G. M. 2004, ApJ, 6009, 826

1563: 

1564: \bibitem[Kilic et al.(2006)]{kilicspec} Kilic, M., von Hippel, T.,

1565: Leggett, S. K., Winget, D. E. 2006, \apj, 646, 474

1566: 

1567: %\bibitem[Kilic et al.(2006a)]{kilicWDphot} Kilic, M., von Hippel, T.,

1568: %Mullally, F., Reach, W. T., Kuchner, M. J., Winget, D. E., Burrows, A. 2006a, \apj, 642, 1051

1569: 

1570: \bibitem[Kimura et al.(2005)]{kimura} Kimura, Y., Kurumada, M., Tamura, K., Koike, C.,

1571: Chihara, H., Kaito, C. 2005, \aap, 442, 507

1572: 

1573: \bibitem[Kleinman et al.(1998)]{kleinman} Kleinman, S. J. et al. 1998, ApJ, 495, 424

1574: 

1575: %\bibitem[Koester et al.(1997)]{koester97}

1576: %	Koester, D., Provencal, J., \& Shipman, H. L. 1997, \aap, 320, L57

1577: 

1578: \bibitem[Koester et al.(2005)]{koester05}

1579: Koester, D., Rollenhagen, K., Napiwotzki, R., Voss, B., Christlieb, N.,

1580: Homeier, D., Reimers, D. 2005, \aap, 432, 1025

1581: 

1582: \bibitem[Koike et al.(2003)]{koike} Koike, C., Chihara, H., Tsuchiyama, A., Suto, H., Sogawa, H., Okuda, H. 2003, \aap, 399, 1101

1583: 

1584: \bibitem[Kuchner et al.(1998)]{kuchner98} Kuchner, M. J., Koresko, C. D., \& Brown, M. E. 1998, \apjl, 508, L81

1585: 

1586: %\bibitem[Kuchner \& Saeger(2005)]{kuchnersaeger} Kuchner, M. J., \& Saeger, S. 2005, submitted to ApJ

1587: 

1588: \bibitem[Li and Draine(2001)]{lidraine} Li, A., Draine, B. T. 2001, \apj, 554, 778

1589: 

1590: \bibitem[Lisse et al.(2006)]{lisse06} Lisse, C. M., et al. 2006, Science, 313, 635

1591: 

1592: \bibitem[Lisse et al.(2007a)]{lisse07a} Lisse, C. M., Kraemer, K. E.,

1593: Nuth III, J. A., Li, A., \& Joswiak, D. 2007a, Icarus, 187, 69

1594: 

1595: \bibitem[Lisse et al.(2007b)]{lisse07b} Lisse, C. M., 

1596: Beichman, C. A., Bryden, G., Wyatt, M. C. 2007b, \apj, 658, 584

1597: 

1598: \bibitem[Lisse et al.(2008)]{lisse08} Lisse, C. M., Chen, C. H., Wyatt, M. C., Morlok, A.

1599: 2008, \apj, 673, 1106

1600: 

1601: \bibitem[Marcy et al.(2005)]{hotjupiter}

1602: Marcy, G., Butler, R. P., Fischer, D., Vogt, S., Wright, J. T.,

1603: Tinney, C. G., Jones, H. R. A. 2005. Prog. Theor. Phys. Suppl., 158, 24

1604: 

1605: \bibitem[Mathis(1994)]{ref2175}

1606: Mathis, J. S. 1994, \apj, 422, 176

1607: 

1608: \bibitem[Maxted et al.(2006)]{maxted} Maxted, P. F. L., Napiwotzki, R., Dobbie, P. D.,

1609: Burleigh, M. R. 2006, Nature, 442, 543

1610: 

1611: %\bibitem[McCook \& Sion(1987)]{mccooksion} McCook, G. P. \& Sion, E. M. 1987, \apjs, 65, 603

1612: 

1613: \bibitem[Messenger, Sandford, \& Brownlee(2006)]{messenger}

1614: Messenger, S., Sandford, S., \& Brownlee, D. 2006, in

1615: {\it Meteorites and the Early Solar System II}, eds. D. S. Lauretta and H. Y. McSween Jr.,

1616: (Tucson:Univ. Arizona Press), p. 187--208

1617: 

1618: \bibitem[Weisberg et al.(2006)]{weisberg}

1619: Weisberg, M. K., McCoy, T. J., Krot, A. N. 2006, in

1620: {\it Meteorites and the Early Solar System II}, eds. D. S. Lauretta and H. Y. McSween Jr.,

1621: (Tucson:Univ. Arizona Press), p. 19--52

1622: 

1623: \bibitem[Molster and Waters(2003)]{molsterwaters} Molster, F. J., Waters, L. B. F. M. 2003, in {it Astromineralogy}, ed. Th. Henning (Lect. Notes. Phys. vol. 609), 121

1624: 

1625: \bibitem[Monnier \& Millan-Gabet(2002)]{monnier}

1626: Monnier, J. D., \& Millan-Gabet, R. 2002, \apj, 579, 694

1627: 

1628: \bibitem[O'Brien and Greenberg(2003)]{obrien03} 

1629: O'Brien, D. P., and Greenberg, R. 2003. Icarus, 164, 334

1630: 

1631: %\bibitem[O'Donovan et al.(2007)]{odonovan} O'Donovan, F. T., et al. 2007, ApJ, 663, L37 

1632: \bibitem[Ogilvie and Lin(2007)]{ogilvie} Ogilvie, G. I., Lin, D. N. C. 2007, \apj, 661, 1180

1633: 

1634: %\bibitem[Reach et al.(2003)]{reachzody} Reach, W. T., Morris, P., 

1635: %	Boulanger, F., \& Okumura, K. 2003, Icarus, 164, 384

1636: 

1637: \bibitem[Reach et al.(2005a)]{paperone} Reach, W. T., 

1638: Kuchner, M. J., von Hippel, T., Burrows, A.,

1639: Mullally, F., Kilic, M., Winget, D. E. 2005a, \apjl, 635, L161

1640: 

1641: \bibitem[Reach et al.(2005b)]{reachcal} Reach, W. T. at 12 coauthors,

1642:  2005b, PASP, 117, 978

1643: 

1644: \bibitem[Rhee et al.(2007)]{rhee}

1645: Rhee, J. H., Song, I., Zuckerman, B., \& McElwain, M. 2007, ApJ, 660, 1556

1646: 

1647: \bibitem[Rieke et al.(2004)]{rieke}

1648: Rieke, G. H. et al. 2004, ApJS, 154, 25

1649: 

1650: \bibitem[Rieke et al.(2005)]{riekeAstar}

1651: Rieke, G. H. et al. 2005, ApJ, 620, 1010

1652: 

1653: \bibitem[Sasselov(2003)]{sasselov} Sasselov, D. D. 2003, ApJ, 596, 1327

1654: 

1655: \bibitem[Scargle(1982)]{scargle} Scargle, J. D. 1982, \apj, 263, 835

1656: 

1657: \bibitem[Schleicher \& A'Hearn(1988)]{schleicherahearn84}

1658: Schleicher, D. G. \& A'Hearn, M. F. 1988, \apj, 331, 1058

1659: 

1660: \bibitem[Shearer, Papike, and Rietmeijer(1998)]{shearer}

1661: Shearer, C. K. Papike, J. J., Rietmeijer, F. J. M. 1998, in {\it Planetary Materials}

1662: (Reviews in Mineralogy Vol. 36), ed. J. J. Papike (Washington: Mineralogical 

1663: Society of America), p. 3.

1664: 

1665: %\bibitem[Sykes \& Greenberg(1986)]{sykes86}

1666: %Sykes, M. V., \& Greenberg, R. 1986, Icarus, 65, 51

1667: 

1668: \bibitem[Su et al.(2007)]{suchu}

1669: Su, K.~Y.~L., et al.\ 2007, \apjl, 657, L41 

1670: 

1671: %\bibitem[Th\'ebault, Augereau, and Beust(2003)]{thebault} 

1672: %Th\'ebault, P., Augereau, J. C., and Beust, H. 2003. \aap, 408, 775

1673: 

1674: %\bibitem[Tokunaga et al.(1990)]{tokunaga}

1675: %Tokunaga, A. T., Becklin, E. E., \& Zukerman, B. 1990, \apjl, 358, L21

1676: 

1677: \bibitem[Trilling et al.(2008)]{trilling}

1678: Trilling, D. E., Bryden, G., Beichman, C. A., Rieke, G. H., Su, K. Y. L., Stansberry, J. A.,

1679: Blaylock, M., Stapelfeldt, K. R., Beeman, J. W., \&  Haller, E. E. 2008. \apj, 674, 1086

1680: 

1681: \bibitem[Vassiliadis and Wood(1993)]{vassiliadis} Vassiliadis, E., Wood, P. R. 1993, \apj, 413, 641

1682: 

1683: \bibitem[Vinkovi\'c et al.(2003)]{vinkovic}

1684: Vinkovi\'c, D., Ivevi\'c, Z., Miroshnichenko, A. S., Elitzer, M.

1685: 2003, \mnras, 346, 1151 

1686: 

1687: \bibitem[von Hippel et al.(2007)]{vonhippel}

1688: von Hippel, T., Kuchner, M. J., Kilic, M., Mullally, F., \& Reach, W. T. 2007, \apj 662, 544

1689: 

1690: \bibitem[Werner et al.(2004)]{werner}

1691: Werner, M. W. et al. 2004, ApJS, 154, 1

1692: 

1693: \bibitem[Weidemann(2000)]{weidemann} Weidemann, V. 2000, \aap, 363, 647

1694: 

1695: \bibitem[Winget et al.(1990)]{winget} Winget, D. E. and 27 coauthors, 1990, \apj, 357, 630

1696: 

1697: \bibitem[Woodward et al.(2007)]{woodward07} Woodward, C. E., Kelley, M. S., Bockel\'ee-Morvan, D., Gehrz, R. D. 2007, \apj, 671, 1065

1698: 

1699: %\bibitem[Zuckerman \& Becklin(1987)]{ZB87}

1700: %Zuckerman, B., \& Becklin, E. E. 1987, Nature, 330, 138

1701: 

1702: %\bibitem[Zuckerman \& Reid(1998)]{zuckreid}

1703: %Zuckerman, B., \& Reid, I. N. 1998, \apjl, 505, L143

1704: %Zuckerman, B., Koester, D., Reid, I. N., \& H\"unsch, M. 2003, \apj, 496, 477

1705: 

1706: %\bibitem[Zuckerman et al.(2003)]{zuckerman}

1707: %Zuckerman, B., Koester, D., Reid, I. N., \& H\"unsch, M. 2003, \apj, 596, 477

1708: 

1709: \bibitem[Zubko et al.(1996)]{zubkocar}

1710: Zubko V.G., Mennella V., Colangeli L., Bussoletti E., 1996,

1711: MNRAS, 282, 1321

1712: 

1713: \end{thebibliography}

1714: 

1715: \clearpage

1716: \begin{table}

1717: \caption[]{{\it Spitzer} Observing log for G29-38}\label{obstab} 

1718: \begin{flushleft} 

1719: \begin{tabular}{llll}

1720: Date & AORID & Instrument & Wavelengths \\ 

1721: \hline

1722: 2004 Nov 26 & 10119424 & IRAC & 4.5, 8 $\mu$m \\

1723: 2005 Nov 26 & 11124224 & IRAC & 3.6, 4.5, 5.8, 8 $\mu$m \\

1724: 2004 Dec 2  & 10149376 & MIPS & 24 $\mu$m \\

1725: 2004 Dec 8  & 10184192 & IRS & 5.2--14.2 $\mu$m\\

1726: 2005 Dec 23 & 13835264 & IRAC & 3.6, 4.5, 5.8, 8 $\mu$m \\

1727: 2006 Jun 30 & 13828096 & IRS & 5.2--36 $\mu$m \\

1728: \end{tabular}

1729: \end{flushleft} 

1730: \end{table}  

1731: 

1732: \clearpage

1733: 

1734: \begin{table}

1735: \caption[]{Mid-infrared Fluxes of G29-39 (mJy)$^a$}\label{iractab} 

1736: \begin{flushleft} 

1737: \begin{tabular}{lccc}

1738: 	&2004 Nov 26 & 2004 Nov 26 & 2005 Dec 23\\

1739:     & 10:54     & 10:58       & 23:23\\

1740: \hline

1741: 3.6 $\mu$m &                & $8.37\pm 0.01$ & $8.10\pm 0.03$ \\

1742: 4.5 $\mu$m & $8.88\pm 0.02$ & $8.87\pm 0.01$ &    \\

1743: 5.8 $\mu$m &                & $8.37\pm 0.02$ & $8.28\pm 0.07$\\

1744: 8.0 $\mu$m & $8.73\pm 0.03$ & $8.72\pm 0.02$ &     \\

1745: \end{tabular}

1746: %\tablenotetext{a}

1747: \noindent\par $^a${Uncertainties are statistical uncertainty

1748: in the weighted mean of the flux measurements from all frames

1749: taken during each observing sequence}

1750: \end{flushleft} 

1751: \end{table}  

1752: 

1753: \clearpage

1754: 

1755: \begin{deluxetable}{lcccccc}

1756: \tablecaption{Composition of the Best-Fit Model\tablenotemark{a} to the {\it Spitzer}/IRS G29-38 Spectrum\label{minfittab}}

1757: %\tablewidth{7truein}

1758: \tabletypesize{\footnotesize}

1759: \tablehead{

1760: \colhead{Species} & \colhead{Weighted\tablenotemark{b}} & \colhead{Density}&  

1761: \colhead{Mol. Wt.} & \colhead{$N_{moles}$\tablenotemark{c}} & 

1762: \colhead{$T_{max}^{d}$}& 

1763: \colhead{$\chi^{2}$ if}\\

1764: & Surface Area & (g~cm$^{-3}$) & (relative) && (K) & excluded 

1765: }

1766: \startdata

1767: \cutinhead{Detections	}

1768: %{\bf Olivines}\\

1769: Amorph Olivine (MgFeSiO$_4$) & 	0.33	 &	3.6	&      172	&	0.69	&   890	&	90.6\\

1770: Fayalite (Fe$_2$SiO$_4$)		&	0.08	&	4.3&	      204&		0.17&	   890&		2.91\\

1771: 

1772: %{\bf Pyroxenes}\\

1773: FerroSilite (Fe$_2$Si$_2$O$_6$)&  0.11&		4.0&	      264&		0.17&	   890&	9.85\\

1774: Diopside (CaMgSi$_2$O$_6$)	 &    	0.05	 &	3.3	&     216&		0.076&	   890&		2.05\\

1775: OrthoEnstatite (Mg$_2$Si$_2$O$_6$)&  	0.04	&	3.2	&      200&	0.064&	   890&		1.98\\

1776: 

1777: %{\bf Metal Sulfides}\\

1778: Niningerite (Mg$_{10}$Fe$_{90}$S)\tablenotemark{e} &0.10&4.5& 84&	0.53&  890&		1.49\\

1779: 

1780: %{\bf Organics}\\

1781: Amorph Carbon (C)&      	0.28&		2.5&       12&		5.83&	   930&	$> 100$\\

1782: 

1783: %{\bf Water}\\

1784: Water-ice (H$_2$O)&     	0.29&		1.0	&     18&	1.61&	   220&		5.82\\

1785: 						

1786: \cutinhead{Upper Limits and Non-Detections}

1787: 

1788: Forsterite[Koike](Mg$_2$SiO$_4$)&  	0.02&3.2&      140&	0.046&   890&	1.15\\

1789: Amorph Pyroxene (MgFeSi$_2$O$_6$)&  	0.00	&3.5&      232&	0.09&   890&		1.04\\

1790: 

1791: %Phyllosilicates

1792: Smectite/Notronite\tablenotemark{f} &     	0.00	&	2.3&	    496&		0.03&	   890&		1.04\\

1793: 

1794: Water Gas (H$_2$O)&    	0.01	&	1.0	&       18&	$\leq 0.00$&   220&		1.04\\

1795: 

1796: %Carbonates

1797: Magnesite (MgCO$_3$)&  	0.00&	3.1&     84&	$\leq 0.00$&	   890&		1.04\\

1798: Siderite (FeCO$_3$)&    	0.00&	3.9&      116&	$\leq 0.00$ & 890&	1.04\\

1799: 

1800: PAH  (C$_{10}$H$_{14}$)& 	0.00&	1.0&    (178)&	$\leq 0.011$&	 N/A	&	1.04\\

1801: \enddata

1802: \tablenotetext{a}{Best-fit model $\chi^2_\nu=1.04$ with power law particle size distribution $dn/da \propto a^{-3.7}$,  5--35 $\mu$m range of fit, 336 degrees of freedom}

1803: \tablenotetext{b}{Weight of the emissivity spectrum of each dust species required to match the G29-38 emissivity spectrum.}

1804: \tablenotetext{c}{$N_{moles}(i)$ is the Density/Molecular Weight $\times$ Normalized Surface Area for mineral $i$. Errors are $\pm 15$\% (1$\sigma$).}

1805: \tablenotetext{d}{All temperatures are $\pm 20$K (1$\sigma$)}

1806: \tablenotetext{e}{We use the name niningerite to refer to Mg$_{x}$Fe$_{1-x}$S.

1807: a niningerite composition of Mg$_{25}$Fe$_{75}$S may fit the data better.}

1808: \tablenotetext{f}{Na$_{0.33}$Fe$_2$(Si,Al)$_4$O$_{10}$(OH)$_2\cdot3$H$_2$O)}

1809: \end{deluxetable}

1810: 

1811: \clearpage

1812: 

1813: 

1814: \begin{table}

1815: \caption[]{Best-fitting moderately-optically-thin model}\label{modthicktab} 

1816: \begin{flushleft} 

1817: \begin{tabular}{lll}

1818: Parameter & Best value & Confidence interval\\

1819: \hline

1820: $T_{vap}$ & 1100 K & 1050 -- 1200 K\\

1821: $\tau_{\parallel}$ & 2 & 1 -- 8\\

1822: $R_{min}$ & 50 & $\le 150$\\

1823: $\alpha$ & 2.7 & 2.4 -- 2.9\\

1824: \end{tabular}

1825: \end{flushleft} 

1826: \end{table}  

1827: 

1828: \clearpage

1829: 

1830: \begin{figure}[th]

1831: \plotone{f1}

1832: \caption[sourcevar]{

1833: Flux versus time for each of the IRAC channels

1834: (from top to bottom: 3.6, 4.5, 5.8, and 8 $\mu$m wavelength).

1835: The time begins with the first frame of the entire

1836: sequence (which combines two AORs) at

1837: 2004 Nov 26 10:54:11.28 UT.

1838: Each point is a flux measurement from a single image,

1839: with the statistical uncertainty of the aperture

1840: photometry performed as described by the

1841: IRAC calibration procedure \citep{reachcal}.

1842: Line segments simply connect the data points.

1843: The high-error points at 4.5 and 8 $\mu$m are the

1844: short frames taken just at the beginning of each

1845: of the two AORs. All four plots are scaled to show the

1846: same dynamic range (from -6\% to +8.4\%, centered

1847: on the median).

1848: \label{sourcevar}}

1849: \end{figure}

1850: 

1851: \begin{figure}[th]

1852: \plotone{f2}

1853: \caption[sourceper]{

1854: Scargle periodograms of the time series from the IRAC

1855: 2004 Nov 26 observations. Each panel has the same scale

1856: both horizontally and vertically; the 3.6 $\mu$m panel

1857: is twice as large because the amplitudes are much higher.

1858: The periods ($2\pi/\omega$, where $\omega$ is the angular

1859: frequency) are displayed on a logarithmic stretch.

1860: \label{sourceper}}

1861: \end{figure}

1862: 

1863: 

1864: \begin{figure}[th]

1865: \plotone{f3}

1866: \caption[ffit]{

1867: Combined spectra and photometry

1868: (using 2MASS, IRTF, IRAC, MIPS, and IRS)

1869: of G29-38 with stellar photosphere removed.

1870: \label{ffit}}

1871: \end{figure}

1872: 

1873: \begin{figure}[th]

1874: \plotone{f4}

1875: \caption[specmod]{

1876: Emissivity spectrum of the dust around G29-38. The observed

1877: spectrum has been divided by a 930 K blackbody and fitted with

1878: a linear combination of 12 minerals. The shape of each mineral's

1879: emissivity, normalized by its fitted amplitude to the

1880: G29-38 emissivity, is shown separately by a colored line

1881: (offset vertically for clarity):

1882: red = amorphous carbon,

1883: bright green = carbonates (zero amplitude),

1884: yellow = PAH (zero amplitude),

1885: light orange = water vapor (zero amplitude),

1886: deep orange = water ice,

1887: olive green = sulfides, represented here by niningerite,

1888: blue = phyllosilicates (zero amplitude),

1889: light blue = crystalline pyroxenes (ferrosilite,

1890: diopside, and orthoenstatite, in order of 20 $\mu$m amplitude),

1891: purple = amorphous olivine, and

1892: dark blue = crystalline olivines (forsterite and fayalite, in order of 20 $\mu$m amplitude).

1893: \label{specmod}}

1894: \end{figure}

1895: 

1896: 

1897: \begin{figure}[th]

1898: \plotone{f5}

1899: \caption[specresid]{

1900: Emissivity spectrum as in Fig.~\ref{specmod}, but after removal of

1901: the best-fitting silicates and carbon.  

1902: The residuals are well fit by a combination of water ice (deep orange, dashed and solid are two 

1903: temperatures) and metal sulfides (olive green). 

1904: The water ice is at temperature $~200$ K, and cannot be in the same location as the 930 K dust.

1905: Whereas the dust is $\sim 10^{11}$ cm from the star, the water ice must be 

1906: further $\sim 10^{13}$ cm, at the outer edge of an extended disk.

1907: \label{specresid}}

1908: \end{figure}

1909: 

1910: 

1911: 

1912: \begin{figure}[th]

1913: \plotone{f6}

1914: \caption[g29v3]{

1915: Cartoons illustrating possible geometries for the G29-38 disk.

1916: The physically thin plus warp model (top) is based on \citet{jura07}'s model for the similar white dwarf GD 362.

1917: The physically thick (bottom) is based on \citet{paperone}

1918: and elaborated in \S\ref{modsec}. 

1919: Distances from the star are labeled, in units of stellar radii.

1920: The distances from the star for

1921: grain vaporization (greybody grains at 1200 K)

1922: and tidal disruption (Roche limit for a

1923: solid body to be tidally disrupted by the white dwarf with

1924: radius $7.5\times 10^{8}$ cm and mass 0.69 $M_{\odot}$)

1925: are indicated.

1926: \label{cartoon}}

1927: \end{figure}

1928: 

1929: \begin{figure}[th]

1930: \epsscale{0.8}

1931: \plotone{f7}

1932: \epsscale{1}

1933: \caption[g29plotshell]{

1934: Spherical shell models for the dust around G29-38. 

1935: Individual models for amorphous olivine and carbon are shown as dashed and dotted

1936: curves, respectively, and a linear combination is shown as the solid curve.

1937: The input white dwarf spectrum, and the photosphere-subtracted infrared observations, are

1938: shown for comparison. 

1939: The DUSTY models have optical depth $\tau(0.55\mu{\rm m})=0.018$ and 0.011 in silicates and

1940: carbon, respectively.

1941: %The DUSTY models have optical depth $\tau(0.55\mu{\rm m})=0.07$, and

1942: %the linear combination

1943: %was 0.41 times the silicate model plus 0.23 times the silicate model.

1944: \label{g29plotshell}}

1945: \end{figure}

1946: 

1947: \begin{figure}[th]

1948: \epsscale{0.8}

1949: \plotone{f8}

1950: \epsscale{1}

1951: \caption[g29rad]{

1952: Temperature profiles in a spherical shell around G29-38.

1953: The top panel is for amorphous olivine, 

1954: and the bottom panel is for amorphous carbon.

1955: All models are computed to an inner temperature of 1200 K.

1956: Each profile is for a different total optical depth, with

1957: values 0.0100, 0.0268, 0.072, 0.193, 0.518, 1.39, 3.73, and 10.

1958: The optically thin models extend closest to the star, are the

1959: warmest at the outer edge of the plot, and are closest to a

1960: straight line ($T\propto r^{-0.5}$).

1961: The optically thick models begin further from the star

1962: and decrease in temperature much more steeply.

1963: \label{g29rad}}

1964: \end{figure}

1965: 

1966: \begin{figure}[th]

1967: \epsscale{0.8}

1968: \plotone{f9}

1969: \epsscale{1}

1970: \caption[tplot]{

1971: Temperatures of olivine (top) and carbon (bottom) grains of three different sizes: 

1972: 0.1 $\mu$m (diamonds),

1973: 1 $\mu$m (triangles), and

1974: 10 $\mu$m (squares). 

1975: The grains are irradiated directly by the white dwarf (i.e.

1976: the cloud is optically thin).

1977: The solid line shows power-law fits of 

1978: the form $T=T_{1} (r/100 R_{*})^{-\delta}$. For olivine (carbon) grains of 

1979: 0.1, 1, and 10 $\mu$m radius, 

1980: $T_1=1905$, 1136, 853 K (1393, 740, 780 K) and

1981: $\delta=0.47$, 0.50, 0.54 (0.37, 0.41, 0.51), respectively.

1982: The horizontal dotted line  indicates the

1983: vaporization temperature (1200 K) in our model.

1984: \label{tplot}}

1985: \end{figure}

1986: 

1987: \begin{figure}[th]

1988: \plotone{f10}

1989: \caption[fplotbest]{

1990: Best-fitting moderately-optically thick disk model (solid line)

1991: combining amorphous carbon (dashed), amorphous olivine (dotted),

1992: crystalline bronzite (dash-dot)

1993: to the observed photosphere-subtracted spectral energy distribution of G29-38. 

1994: \label{fplotbest}}

1995: \end{figure}

1996: 

1997: \begin{figure}[th]

1998: \plotone{f11}

1999: \caption[fitresid]{

2000: Residuals (observed minus model) from the best-fitting 

2001: moderately-optically thick disk model of Fig.~\ref{fplotbest}.

2002: Features discussed in the text are labeled.

2003: \label{fitresid}}

2004: \end{figure}

2005: 

2006: 

2007: \begin{figure}[th]

2008: \plotone{f12}

2009: \caption[juramany]{

2010: Goodness of fit ($\chi^{2}/{\rm d.o.f}$) from the physically-thin, 

2011: optically-thick model for a range of outer radii ($R_{2}$, in units of stellar

2012: radius) and inclination. The best fitting model has inner and outer radii $R_{1}=9$,

2013: $R_{2}=25$, and inclination $23^{\circ}$. 

2014: \label{juramany}}

2015: \end{figure}

2016: 

2017: \begin{figure}[th]

2018: \plotone{f13}

2019: \caption[juramany]{

2020: Best-fitting physically-thin, optically-thick disk model. The model (solid line)

2021: was only fitted to data at wavelengths less than 8 $\mu$m;

2022: it cannot produce the silicate feature and under-predicts

2023: (by far) the longer-wavelength continuum.

2024: \label{jurabest}}

2025: \end{figure}

2026: 

2027: \begin{figure}[th]

2028: \plotone{f14}

2029: \caption[thinthick]{

2030: Best-fitting combined model, with a physically-thin, optically-thick disk 

2031: (cf. Fig~\ref{jurabest}) plus a

2032: silicate-only (no C or Fe) physically-thick disk (cf. Fig.~\ref{fplotbest}).

2033: \label{thinthick}}

2034: \end{figure}

2035: 

2036: \def\extra{

2037: \begin{figure}[th]

2038: \plotone{f15}

2039: \caption[silcompare]{

2040: The relative abundance of olivine versus pyroxene determined from the infrared spectrum of G29-38, circumstellar clouds around 

2041: HD 69830, HD 113766, HD 100546, and HD 163296, and comets 29P/Schwassmann-Wachmann 1 and 73P/Schwassmann-Wachmann 3.

2042: The objects in the lower-left are young stars. 

2043: G29-38 is at the upper-right, together with systems for which dust is inferred to be of asteroidal origin.

2044: \label{silcompare}}

2045: \end{figure}

2046: }

2047: 

2048: 

2049: \end{document}

2050: 

2051: 

2052: