1: %%

2: %% Beginning of file 'sample.tex'

3: %%

4: %% Modified 2004 January 9

5: %%

6: %% This is a sample manuscript marked up using the

7: %% AASTeX v5.x LaTeX 2e macros.

8: 

9: %% The first piece of markup in an AASTeX v5.x document

10: %% is the \documentclass command. LaTeX will ignore

11: %% any data that comes before this command.

12: 

13: %% The command below calls the preprint style

14: %% which will produce a one-column, single-spaced document.

15: %% Examples of commands for other substyles follow. Use

16: %% whichever is most appropriate for your purposes.

17: %%

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

19: 

20: %% manuscript produces a one-column, double-spaced document:

21: 

22: %\documentclass[manuscript]{aastex}

23: 

24: %% preprint2 produces a double-column, single-spaced document:

25: 

26: %\documentclass[preprint2]{aastex}

27: 

28: %% Sometimes a paper's abstract is too long to fit on the

29: %% title page in preprint2 mode. When that is the case,

30: %% use the longabstract style option.

31: 

32: %% \documentclass[preprint2,longabstract]{aastex}

33: 

34: %% If you want to create your own macros, you can do so

35: %% using \newcommand. Your macros should appear before

36: %% the \begin{document} command.

37: %%

38: %% If you are submitting to a journal that translates manuscripts

39: %% into SGML, you need to follow certain guidelines when preparing

40: %% your macros. See the AASTeX v5.x Author Guide

41: %% for information.

42: 

43: \newcommand{\lapprox}{{\footnotesize $\buildrel < \over \sim \,$}}

44: \newcommand{\gapprox}{{\footnotesize $\buildrel > \over \sim \,$}}

45: \newcommand{\rb}[1]{\raisebox{0.75ex}[0pt]{#1}}

46: 

47: 

48: %% You can insert a short comment on the title page using the command below.

49: 

50: %\slugcomment{Not to appear in Nonlearned J., 45.}

51: 

52: %% If you wish, you may supply running head information, although

53: %% this information may be modified by the editorial offices.

54: %% The left head contains a list of authors,

55: %% usually a maximum of three (otherwise use et al.).  The right

56: %% head is a modified title of up to roughly 44 characters.

57: %% Running heads will not print in the manuscript style.

58: 

59: \shorttitle{Modeling Extreme Carbon Stars}

60: \shortauthors{Speck et al.}

61: 

62: %% This is the end of the preamble.  Indicate the beginning of the

63: %% paper itself with \begin{document}.

64: 

65: \begin{document}

66: 

67: %% LaTeX will automatically break titles if they run longer than

68: %% one line. However, you may use \\ to force a line break if

69: %% you desire.

70: 

71: %\title{The effect of stellar evolution on dust grain sizes} 

72: %\title{Modeling Extreme Carbon Stars} 

73: %\title{Silicon Carbide Absorption in Extreme Carbon Stars: 

74: %Radiative Transfer Modeling}

75: \title{Silicon carbide absorption features: 

76: dust formation in the outflows of extreme carbon stars}

77: 

78: %% Use \author, \affil, and the \and command to format

79: %% author and affiliation information.

80: %% Note that \email has replaced the old \authoremail command

81: %% from AASTeX v4.0. You can use \email to mark an email address

82: %% anywhere in the paper, not just in the front matter.

83: %% As in the title, use \\ to force line breaks.

84: 

85: \author{ Angela K. Speck, Adrian B. Corman,  

86: Kristina Wakeman \& Caleb H. Wheeler}

87: \affil{Department of Physics \& Astronomy, University of Missouri, Columbia, 

88: MO 65211}

89: \email{speckan@missouri.edu}

90: 

91: \and

92: \author{Grant Thompson}

93: \affil{University of Kentucky, 600 Rose Street, Lexington, Kentucky 40506}

94: 

95: %% Mark off your abstract in the ``abstract'' environment. In the manuscript

96: %% style, abstract will output a Received/Accepted line after the

97: %% title and affiliation information. No date will appear since the author

98: %% does not have this information. The dates will be filled in by the

99: %% editorial office after submission.

100: 

101: \begin{abstract}

102: 

103: Infrared carbon stars without visible counterparts are generally known as 

104: extreme carbon stars. We have selected a subset of these stars 

105: with absorption features in the 10-13\,$\mu$m range, which has been 

106: tentatively attributed to silicon carbide (SiC). We add three new objects 

107: meeting these criterion to the seven previously known, bringing our total 

108: sample to ten sources. We also present the result of radiative transfer 

109: modeling for these stars, comparing these results to those of previous 

110: studies. In order to constrain model parameters, we use published mass-loss 

111: rates, expansion velocities and theoretical dust condensation models 

112: to determine the dust condensation temperature. These show 

113: that the inner dust temperatures of the dust shells for these sources are 

114: significantly higher than previously assumed. This also implies that the 

115: dominant dust species should be graphite instead of amorphous carbon. In 

116: combination with the higher condensation temperature we show that this results 

117: in a much higher acceleration of the dust grains than would be expected from 

118: previous work.  

119: %

120: Our model results suggest that the very optically thick stage of evolution  

121: does not coincide with the timescales for the superwind, but rather, that 

122: this is a very short-lived phase.

123: %

124: Additionally, we compare model and observational 

125: parameters in an attempt to find any correlations. Finally, we show that the 

126: spectrum of one source, IRAS 17534$-$3030, strongly implies that the 

127: 10-13\,$\mu$m feature is due to a solid state rather than a molecular species.

128: 

129: \end{abstract}

130: 

131: %% Keywords should appear after the \end{abstract} command. The uncommented

132: %% example has been keyed in ApJ style. See the instructions to authors

133: %% for the journal to which you are submitting your paper to determine

134: %% what keyword punctuation is appropriate.

135: 

136: %% Authors who wish to have the most important objects in their paper

137: %% linked in the electronic edition to a data center may do so in the

138: %% subject header.  Objects should be in the appropriate "individual"

139: %% headers (e.g. quasars: individual, stars: individual, etc.) with the

140: %% additional provision that the total number of headers, including each

141: %% individual object, not exceed six.  The \objectname{} macro, and its

142: %% alias \object{}, is used to mark each object.  The macro takes the object

143: %% name as its primary argument.  This name will appear in the paper

144: %% and serve as the link's anchor in the electronic edition if the name

145: %% is recognized by the data centers.  The macro also takes an optional

146: %% argument in parentheses in cases where the data center identification

147: %% differs from what is to be printed in the paper.

148: 

149: \keywords{infrared: stars --- stars: carbon --- circumstellar matter --- dust --- stars: AGB and post-AGB}

150: 

151: %% From the front matter, we move on to the body of the paper.

152: %% In the first two sections, notice the use of the natbib \citep

153: %% and \citet commands to identify citations.  The citations are

154: %% tied to the reference list via symbolic KEYs. The KEY corresponds

155: %% to the KEY in the \bibitem in the reference list below. We have

156: %% chosen the first three characters of the first author's name plus

157: %% the last two numeral of the year of publication as our KEY for

158: %% each reference.

159: 

160: \section{Introduction}

161: 

162: \subsection{Physical evolution of intermediate mass stars and their circumstellar shells}

163: 

164: Stars between about 1 and 8\,M$_{\odot}$ will eventually evolve up the 

165: Asymptotic Giant Branch  \citep[AGB;][]{Iben1983}. Because of instabilities in 

166: their interior, AGB stars pulsate and throw off large amounts of mass from 

167: their surface 

168: \citep[e.g.,][]{Vassiliadis1993}. This intensive mass loss produces a circumstellar shell of dust and neutral gas.

169: Once the AGB star has exhausted its outer envelope, the AGB phase ends.

170: At this point, the mass loss virtually stops and the circumstellar shell

171: begins to drift away from the star.

172: At the same time, the central star begins to

173: shrink and heat up from $\sim$3000\,K until it is hot enough to ionize the

174: surrounding gas, at which point the object becomes a planetary nebula (PN).

175: The short-lived post-AGB phase, as the star evolves toward to the PN phase, is

176: also known as the proto- or pre-planetary nebula (PPN) phase.

177:  During the ascent of the AGB, the

178: velocity of the outflowing mass appears to be fairly constant

179: \citep[e.g.,][]{hug88,fong03}. Therefore the dust furthest from the star

180: represents the oldest mass loss, while material closer to the star represents

181: more recent mass loss. Towards the end of the AGB phase

182: the increasing

183:  impact of the thermal pulse cycles leads to an increasing mass-loss

184:  rate 

185: \citep[e.g.,][]{vw93,vil02a}. Such an increase in mass-loss rate (dubbed the

186: {\it superwind}) is necessary to explain the densities seen in typical PNe 

187: \citep{renz81}. Since the invocation of the superwind,

188: many observations of AGB stars and post-AGB stars have supported this 

189: hypothesis \citep[e.g.,][]{km85,wood92}.

190: 

191: \subsection{Chemical evolution of AGB stars and their circumstellar shells}

192: 

193: The chemical composition of the atmospheres of AGB stars is expected 

194: to change as these stars evolve, due to convective dredge up of carbon 

195: produced in the He-burning shell. The amount of carbon relative to oxygen (the 

196: C/O ratio) is critical in determining which types of dust and molecules 

197: are present around 

198: an AGB star. The formation of extremely stable CO molecules will consume 

199: whichever of the two elements is less abundant, leaving only the more abundant 

200: element available for dust formation. Stars start their lives with the cosmic 

201: C/O ratio of $\approx$0.4 and are therefore oxygen-rich. In about a third of 

202: AGB stars, enough carbon will be dredged up to make C/O $>$ 1 and therefore 

203: carbon will dominate the chemistry around these stars, known as carbon stars.

204: %

205: Carbon stars are expected to have circumstellar shells dominated by amorphous 

206: or graphitic carbon grains, which do not have diagnostic infrared features. 

207: Another component of the dust shell around carbon stars, silicon carbide (SiC),

208: does have an infrared spectral feature at $\approx$11$\mu$m and therefore has 

209: been of great interest to researchers seeking to understand the evolution of 

210: the dust shells and infrared features of carbon stars 

211: \citep{Baron1987, Chan1990, Goebel1995, Speck1997, Sloan1998, Speck2005, 

212: Speck2006, Thompson2006}. 

213: 

214: \subsection{Extreme Carbon Stars}

215: \label{ecs_intro}

216: 

217: As carbon stars evolve, mass loss is expected to increase. Consequently, their 

218: circumstellar shells become progressively more optically thick, and

219: eventually the central star is obscured. \citet{Volk1992,Volk2000} 

220: christened such stars ``extreme carbon  stars''. 

221: These stars have also been dubbed ``infrared carbon stars'' 

222: \citep{Groenewegen1994}, and ``very cold carbon stars'' \citep{omont93}.

223: Extreme carbon stars are expected represent that small subset of 

224: carbon-rich AGB stars which are in the superwind phase, just prior to 

225: leaving the AGB. 

226: %

227: Because the superwind phase is short-lived compared to the AGB phase the 

228: number of extreme carbon stars is intrinsically small. Consequently, few 

229: of these objects are known. At present there are $\sim$30 known extreme carbon stars in the Galaxy \citep{Volk1992}

230: compared to $\sim$30,000 known visible carbon stars \citep{skrutskie}. 

231: 

232:  \citet{vh88} attempted to define a way to distinguish between oxygen-rich 

233: and 

234: carbon-rich AGB stars using IRAS color-color space, which was divided into 

235: subsections according to the properties of the dusty shells are these stars 

236: \citep[see Table~1 of][]{vh88}. This was further refined 

237: by \citet{omont93} who identified a population of very cold carbon stars using 

238: HCN and CO observations, and showed that the regions originally designated as 

239: extremely dusty O-rich AGB stars also contain a significant fraction of C-rich 

240: stars. 

241: 

242: 

243: The refinement of the \citet{vh88} color-color diagram by \citet{omont93} 

244: defined subdivisions of the seven zones in color-color space 

245: \citep[see Fig.~1 in][]{omont93}.

246: Cool carbon stars with high mass-loss rates (and little or no SiC emission) 

247: fall into regions III and IV, which had previously been assumed to define 

248: OH-IR stars (i.e. the oxygen-rich counterparts to extreme carbon stars). The 

249: numbered regions have been subdivided into smaller regions denoted by 

250: IIIa1, IIIa2, IIIb1 etc.

251: %

252: %{\tt [ need to say what the subdivisions mean]} 

253: %

254: A subset of the color-color space, covering parts of

255: regions IIIa1c, IIIb2, IIIb2 and VIb is reproduced in Fig.~\ref{irascolor} and 

256: includes our sample stars.

257: 

258: \subsection{SiC absorption features}

259: \label{sicabs}

260: 

261: 

262: SiC has long been predicted to be present in carbon star circumstellar shells, 

263: beginning with condensation theories \citep{Friedman1969,Gilman1969} and 

264: continuing with the prediction of a characteristic SiC $\sim$11$\mu$m  

265: spectral feature \citep{Gilra1971} and then the observational discovery of 

266: an $\sim$11$\mu$m emission feature in many carbon star spectra 

267: \citep{Hackwell1972,Treffers1974}. 

268: %

269: The effect of the evolving dust shell density structure on observed 

270: features, and particularly on the $\sim$11$\mu$m feature, have been discussed 

271: extensively 

272: \citep[see review in][and references therein]{Speck2005}.

273: %

274: As the optical depth of the dust shell increases, self-absorption

275: will diminish the $\sim$11$\mu$m feature and it will 

276: eventually be seen in net absorption. 

277: SiC self-absorption was found to be important in producing accurate 

278: radiative transfer models of extreme carbon stars

279: \citep[e.g.][]{Volk1992}, even though this previous work did not recognize

280: SiC absorption features.

281: %

282: These absorption features are rare and have mostly been ignored in 

283: discussions of 

284: evolutionary sequences in carbon star spectra. In fact the rarity of such 

285: absorption features led to the hypothesis that SiC becomes coated in carbon at 

286: high optical depths \citep[e.g.][]{Baron1987,Chan1990}. 

287: However, meteoritic data and theoretical models do not support this hypothesis 

288: (see \S~\ref{meteor} and \S~\ref{condmod}).

289: 

290: 

291: A few extreme carbon stars have been shown to have an absorption feature at 

292: $\sim11\mu$m which has been tentatively attributed to SiC. This feature was 

293: discovered in the ``prototype'' extreme carbon star AFGL~3068 

294: \citep[hereafter referred to as IRAS\,23166$+$1655;][]{Jones1978}, 

295: and was re-examined by \citet{Speck1997}, which also 

296: identified three additional extreme carbon stars with this feature. 

297: \citet{Clement2003} examined the absorption features of two of these extreme 

298: carbon stars (IRAS\,23166+1655 and IRAS\,02408+5458), 

299: and showed that their 11$\mu$m 

300: absorption features are consistent with $\beta$-SiC\footnote{%

301: %

302: Silicon carbide exists in many ($>$70) different crystal structures, known as 

303: polytypes. See \citet{Speck1997,daulton03,Pitman2007} for a discussion of 

304: the polytypes of SiC.}

305: %

306: nanoparticles. 

307: %

308: The broad absorption features of IRAS\,19548+3035 and IRAS\,21318+5631

309: \citep[also discovered by][]{Speck1997} were attributed to SiC absorption

310: with an interstellar silicate absorption contribution 

311: \citep[see also][]{Groenewegen1996}. This will be discussed further in 

312: \S~\ref{correlsect}.

313: %

314: The absorption features in the spectra of IRAS\,19548+3035 and IRAS\,21318+5631

315: were revisited by \citet{Clement2005} who suggested Si$_3$N$_4$ grains as the 

316: carrier. However, this hypothesis has been shown to be erroneous

317: \citep{Pitman2006}. 

318: 

319: The failure of the Si$_3$N$_4$ hypothesis led \citet{Speck2005} to suggest that

320: amorphous SiC grains may be able to account for the breadth, structure  and 

321: barycentric position of the observed broad 10-13$\mu$m feature in 

322: IRAS\,19548+3035 and IRAS\,21318+5631. 

323: However, the dearth of amorphous presolar SiC grains seems to 

324: preclude this hypothesis (see \S~\ref{meteor}).

325: %

326: %

327: An alternative explanation for this feature is molecular line absorption, 

328: however, currently available line lists are not sufficient to properly 

329: assess this hypothesis \citep[see][and references therein]{Speck2006}.

330: One molecular candidate which has transitions in the correct wavelength 

331: range is C$_3$

332: \citep[e.g.][]{zijl06,jorg}, but the line lists are not readily 

333: available. Furthermore, C$_3$ is expected to be photospheric, rather than circumstellar, 

334: which probably precludes its detection in optically obscured stars.

335: Moreover, the theoretical spectrum of C$_3$ 

336: from \citet{jorg} shows a strong absorption close to the $\sim5\mu$m CO 

337: line, which is stronger than the $\sim11\mu$m feature. As will be seen in

338: \S~\ref{17534}, the spectrum of IRAS\,17534$-$3030 does not show the 5$\mu$m 

339: absorption band and provides evidence that the

340: observed absorption feature is not molecular in origin.

341: 

342: Though previous research has included the effects of SiC self-absorption 

343: \citep[shown to be crucial to produce accurate 

344: models][]{Volk1992,Speck1997,Speck2005}, 

345: no work has been done to directly fit the apparent SiC absorption feature in 

346: radiative transfer models of extreme carbon stars.

347: 

348: 

349: \subsection{Previous Radiative Transfer Models of Extreme Carbon Stars}

350: \label{rtmodels1}

351: 

352: \citet{Volk1992} performed radiative transfer modeling in order to match the

353: Infrared Astronomical Satellite \citep[{\em IRAS};][]{neug84} Low Resolution 

354: Spectrometer (LRS) data for several of extreme carbon stars.

355: They determined that the exact star temperature entered into the 

356: model was not important for the emerging spectra due to the very thick dust 

357: shells around extreme carbon stars 

358: \citep[c.f.][]{DePew2006,Speck2000}.

359: Their models used a fixed composition (a mixture of graphite 

360: and SiC), and a fixed dust condensation temperature. 

361: %

362: \citet{Groenewegen1994} also performed radiative transfer modeling on a larger 

363: set of extreme carbon stars, but these models varied the dust condensation 

364: temperature. Again this was based on {\em IRAS} LRS data.

365: Following up on this, \citet{Groenewegen1995} modeled a large sample of carbon stars

366: using amorphous carbon optical constants \citep{rm91}, 

367: and assumed low dust condensation temperature in a fairly narrow range 

368: (650--900\,K for the extreme carbon stars). Consequently the inner dust radius is 

369: larger than expected. Moreover the resulting models all

370: have relatively low optical depths ($\tau_{11.3\mu \rm m} < 2$). 

371: The optical depth for their IRAS\,23166+1655 model

372: was found to be $\tau_{11.3\mu \rm m} < 1$, 

373: even though this star has an absorption feature at 11$\mu$m.

374: \citet{Groenewegen1998} remodeled these stars, again assuming relatively low 

375: dust condensation temperatures, with similar results.

376: 

377: Finally, \citet{Volk2000} used the improved spectral resolution of the 

378: the Infrared Space Observatory \citep[{\em ISO};][]{kessler} Short Wavelength 

379: Spectrometer \citep[SWS;][]{degraauw96} to examine five extreme carbon stars. 

380: In their modeling study, \citet{Volk2000} allowed the optical depth and radial 

381: dust density distribution to vary; the resulting optical depths were 

382: relatively high (1.4-4.5 at 11.3$\mu$m), and the density of the dust shell was 

383: found to increase rapidly towards the center. This increase was interpreted as 

384: evidence of an increasing mass-loss rate within the last few thousand years, 

385: consistent with the identification of extreme carbon stars as the final stage 

386: of AGB star evolution. While these models did include SiC opacity data, the 

387: 11$\mu$m absorption feature was not recognized and consequently no attempt was 

388: made to fit this feature in these models.

389: 

390: A summary of the parameters of previous models for extreme carbon stars with 

391: 11$\mu$m absorption features in our sample can be 

392: found in Table~\ref{prevmod}. Interestingly, all previous models assume

393: relatively low inner dust temperatures. This will be discussed further in 

394: \S~\ref{condmod}.

395: Furthermore, the modeled dust density distributions suggest a relatively slow 

396: increase in mass-loss rates ($1/r^x$, where $x \approx$ 2.25---3.0). 

397: %This is slower than the predicted increase \citep[e.g.,][]{vw93}.

398: 

399: In reality, dust shells are expected to have heterogeneities and anisotropies 

400: in their density structure 

401: as a result of pulsation-driven dust formation and the 

402: ensuing hydrodynamic turbulent effects \citep[e.g.][]{woitke06}. 

403: These dust formation models suggest that carbon star mass-loss is expected 

404: to be modulated on several timescales, especially that of the pulsation cycle. 

405: % 

406: Furthermore \citet{woitke06} has suggested that 

407: the dynamics in the dust-forming zones around carbon stars lead to 

408: inhomogeneous dust formation, producing fine scale structure in the density of> the dust envelope.

409: In addition, while pulsation shocks are predicted to have a strong effect on 

410: local conditions \citep[e.g.][]{cherchneff06}, 

411: this is not reflected in temporal changes in the IR spectra of carbon stars 

412: \citep{corman08}.

413: As will be seen in \S~\ref{results}, the 

414: spatial scale of the heterogeneities is small and the 

415: timescale for pulsations is short compared to the timescales associated with 

416: even the thinnest dust 

417: shells. Moreover, the inhomogeneities are expected to be wiped out over time 

418: by the hydrodynamic interactions \citep{vil02a,vil02b}. Consequently, we do 

419: not consider these small scale structures in our models.

420: 

421: 

422: \subsection{Meteoritic Evidence}

423: \label{meteor}

424: 

425: The isotopic compositions of certain  grains found in primitive meteorites 

426: indicate that they originated outside the solar system and are thus dubbed 

427: ``presolar''.

428: %

429: Dust grains from AGB stars are found virtually unaltered in

430: these meteorites, demonstrating that these grains become part of the next

431: generation of stars and planets \citep[][and references therein]{cn04}.

432: The precise physical characteristics of these meteoritic dust grains

433: (e.g. sizes, crystal structures, compositions) can be used to help constrain

434: the nature of the dust we see in our astronomical observations.

435: 

436: 

437: \subsubsection{Presolar silicon carbide}

438: \label{presolsic}

439: 

440: Silicon carbide was the first presolar grain to be found in meteorites

441: \citep{Bernatowicz1987} and remains the best studied 

442: \citep[][and references therein]{bern05}. 

443: The most important findings of this work are 

444: (1) that most ($\sim99$\%) of the SiC presolar grains were formed around 

445: carbon stars; 

446: (2) of the AGB SiC grains, \gapprox 95\% appear to originate around low-mass 

447: carbon stars ($<$3\,M$_\odot$), based on nucleosynthesis models of isotopic compositions;

448: (3) that all the SiC grains are crystalline (not amorphous);

449: (4) that nearly all (\gapprox80\%) are of the cubic $\beta$ polytype, with the 

450: remainder comprising the lower temperature 2H polytype;

451: (5) that with one exception, SiC grains have not been 

452: found in the cores of carbon 

453: presolar grains (unlike other carbides: TiC, ZrC, and MoC); and 

454: (6) that the grain size distribution includes both very small and very large 

455: grains (1.5\,nm $\rightarrow$ 26\,$\mu$m), with most grains in the 

456: 0.1--1$\mu$m range.

457: Single-crystal grains can exceed 20$\mu$m in size.

458: Observations of the 11$\mu$m feature have been compared with

459: laboratory spectra of various forms of SiC, and after some false

460: starts it has now been attributed to $\beta$-SiC, matching the information

461: retrieved from meteoritic samples \citep{Speck1999,Clement2003}. 

462: However, there are still some discrepancies between

463: observational and meteoritic evidence (most notably related to

464: grain size).

465: 

466:  \citet{prombo}

467: found a correlation between grain size and the concentration of 

468: {\it s-process} elements in SiC grains taken from the Murchison meteorites. 

469: The Indarch meteorite presolar SiC grains yielded similar results 

470: \citep{jennings}. In both cases, the smaller grains have higher relative 

471: abundances of {\it s-process} elements. This observation may be a result of 

472: different metallicity sources yielding different grain-size distributions

473: \citep{Lagadec2007,Lagadec2008}. Alternatively, it may reflect an evolution 

474: in grain-size with dredge-up \citep{Speck2005}.

475: 

476: 

477: \subsubsection{Presolar ``graphite''}

478: \label{presolc}

479: 

480: In addition to SiC presolar grains, carbon grains are also relatively abundant 

481: and well studied \citep[see][and references therein]{bern05}.

482: Presolar carbon grains are usually referred to as ``graphite'' grains, but 

483: their structures are more complex than this name infers.

484: Presolar graphite is found in two types of spherules classified according to 

485: their external morphologies as ``onion-like'' and ``cauliflower-like''.

486: In general the graphite spherules follow a similar size distribution to the 

487: SiC grains. However, the high-density grains 

488: ($\rho \approx 2.15-2.20$\,g\,$cm^3$) associated with AGB stars have a 

489: mean size of 2$\mu$m. In addition, the AGB presolar graphite spherules span a 

490: larger range of isotopic compositions than the SiC grains, possibly suggesting 

491: that they form at a wider range of times during the AGB phase.

492: 

493: While the presolar SiC grains tend to be single crystals, the graphite grains 

494: regularly contain carbide grains. These carbides are enriched in $s$-process 

495: elements, indicative of formation around late-stage AGB stars.

496: %

497: Many of the ``onion-like'' graphite grains have a core mantle structure in which the 

498: core contains disordered agglomerations of graphene\footnote{%

499: %

500: Graphene is basically a single sheet of graphitic material. If it is 

501: disordered, there are some heptagons and pentagons in place of the regular 

502: hexagonal carbon structure. Graphite is the 3-d structure.}

503: %

504: sheets and PAH\footnote{polycyclic aromatic hydrocarbon}-like 

505: products, while the mantle is composed of well-ordered graphitic concentric 

506: shells. The graphene particles have a typical size of 3-4nm.

507: %

508: The ``cauliflower-like'' graphite grains also have a concentric shell 

509: structure, but it is less well ordered, and is composed primarily of the 

510: disordered graphene.

511: %

512: Whether ``onion'' and ``cauliflower'' graphites are formed in the same 

513: outflows is not known. Both types of grain contain the refractory carbides and 

514: both span the same range of isotopic compositions. 

515: %

516: Whether the ``onion'' or ``cauliflower'' grains are more representative of 

517: grains in the outflows of extreme carbon stars is not known. However, even the 

518: most disordered ``cauliflowers'' or ``onion''-cores are still closer to 

519: graphite 

520: than glassy carbon in structure. The least ordered grains are still considered 

521: to be agglommerations of nano-crystalline grains, rather than truly amorphous

522: (pers.\ comm.\ K. Croat).

523: 

524: \subsubsection{Other presolar carbides}

525: \label{presoltic}

526: 

527: As discussed in \S~\ref{presolsic} and \S~\ref{presolc}, refractory carbides 

528: are found inside ``graphite'' grains but not in SiC grains. Furthermore, SiC 

529: is not one of the carbides found in ``graphite'' grains. The refractory 

530: carbides (TiC, ZrC, MoC and RuC) provide more constraints on the dust 

531: formation processes around carbon stars. In particular, the formation of 

532: ``graphite'' spherules with TiC nuclei limits the range of C/O ratios in which 

533: these grains could form to 1 \lapprox C/O \lapprox 1.2.

534: Meanwhile, the ZrC can form nuclei at higher C/O, but the value still needs to 

535: be less than two.

536: This is consistent with the measured C/O ratios of Galactic carbon stars, 

537: which have an average of 1.15 and a maximum of 1.8 \citep{lamb86}

538: 

539: 

540: \subsection{Investigation}

541: In the present work, we investigate a subset of extreme carbon stars, those 

542: which exhibit the 11$\mu$m absorption feature. Through radiative transfer 

543: modeling, we investigate the nature of these dust shells. We use theoretical 

544: models and meteoritic data to limit the parameter space and thus reduce the 

545: degeneracy within the model results. 

546: In addition, we look for correlations between observed parameters, such as 

547: those that define the 11$\mu$m feature (strength, position, etc) as well as 

548: mass-loss rates and expansion velocities associated with the dust shells.

549: %

550: Finally we determine timescales associated with the dust shells. 

551: 

552: 

553: \section{Observations and Data Processing}

554: \label{obssect}

555: 

556: We investigated 10 extreme carbon stars observed spectroscopically by the ISO 

557: SWS all of which show evidence for an $\sim11\mu$m absorption feature 

558: (see Table~\ref{obstable} and Figs.~\ref{obsfig}, \ref{obsfig2},\ref{obsfig3}

559: and \ref{obsfig4}). 

560: These sources were chosen by searching the ISO archive 

561: for spectra of extreme carbon stars listed in \citet{Volk1992} and selecting 

562: those with an apparent 11$\mu$m absorption feature. In addition, we used the 

563: color-color classification of ``very cold'' carbon stars by \citet{omont93} to 

564: identify further potential sources. Unfortunately most of the potential 

565: sources found in the color-color space (e.g. IRAS\,17583-2291) were not observed

566: by ISO SWS, and the IRAS LRS spectra are too low resolution and/or too noisy to

567:  be used in the present study.

568: %

569: Four of our sources (IRAS\,02408+5458, IRAS\,19548+3035, IRAS\,21318+5631, and 

570: IRAS\,23166 +1655) 

571: were previously studied using ground-based observations and were found to be 

572: consistent with a self-absorbed SiC feature  \citep{Speck1997}. 

573: Two of these sources needed an extra absorption component at $\sim$10$\mu$m, 

574: which were attributed to interstellar absorption. As discussed in 

575: \S~\ref{sicabs}, these have since been the source of some controversy 

576: \citep{Clement2005,Pitman2006}. 

577: %

578: Following the modeling efforts of \citet{Volk2000}, \citet{Speck2005} 

579: identified IRAS\,00210+6221, IRAS\,06582+1507, and IRAS\,17534$-$3030 as exhibiting 

580: an SiC absorption feature. The modeling of \citet{Volk2000} did not include 

581: SiC and did not attempt to fit the 11$\mu$m absorption. Thus, division of the 

582: observed spectra by their respective RT model spectra revealed the 11$\mu$m 

583: absorption feature. 

584: %

585: In this paper we present the discovery of three more potential SiC absorption 

586: features in ISO SWS spectra (IRAS\,01144+6658, IRAS\,03313+6058, and 

587: IRAS\,22303+5950). These were discovered by searching the ISO archive for any 

588: extreme or ``very cold'' carbon stars as determined by their location in the 

589: IRAS color-color space. Those sources that fall within region III without 

590: OH maser emission were examined.

591: %

592: The locations of the stars in our sample in {\em IRAS} color-color space

593: \citep{vh88,omont93} are plotted in Fig~\ref{irascolor}.

594: Interestingly, all sources except IRAS\,19548+3035 

595: plot along a line parallel to and between the 

596: blackbody emission and the $B(T,\lambda,T)\times\lambda^{-1}$ emission lines.

597: 

598: The raw ISO data were extracted from the ISO data archive, and we used the 

599: Off-Line Processing (OLP) pipeline, version 10.1. Individual spectral 

600: sub-bands were cleaned of glitches (caused by cosmic ray particles) and other 

601: bad data sections. Next, they were flat-fielded, sigma-clipped (using the 

602: default values $\sigma$ = 3) and rebinned to the final spectral resolution 

603: ($R=\Delta\lambda/\lambda$), which ranged from 200 to 700, depending on the scanning 

604: speed of the SWS grating during the observation \citep{Leech2003}. The final 

605: spectra are presented in Figure~\ref{obsfig}, which also shows the 

606: best-fitting\footnote{%

607: Best fits are achieved by eye and proceed by examination of the 

608: continuum-divided spectra.}

609: % 

610: blackbody continuum for each spectrum and the resulting 

611: continuum-divided spectra. The blackbody temperatures of the continua at 

612: listed in 

613: Table~\ref{obsparam}.

614: %

615: The continuum-divided spectra clearly show an absorption feature in the 

616: 10-13$\mu$m region, the basic parameters of which (barycentric position, 

617: peak-to-continuum ratio, full width half maximum; FWHM) are listed in 

618: Table~\ref{obsparam}.

619: %

620: The excellent match between the overall spectrum and a single temperature 

621: blackbody suggests that we are seeing an isothermal surface within the dust 

622: shell. This represents the depth at which the shell becomes optically thick. 

623: The lack of extra emission at longer wavelengths suggests that any outlying 

624: dust is low enough in density to have an insignificant contribution to the 

625: overall emission. 

626: %This, in turn, suggests that we are only seeing the {\em superwind}, 

627: %$and the mass-loss rate prior to the onset of the superwind was low.

628: %This will be discussed further in \S~\ref{timescalesect}.

629: 

630: In addition to the {\it ISO\/} SWS spectra, Fig~\ref{obsfig2}, \ref{obsfig3}

631: and \ref{obsfig4} shows the 

632: {\it IRAS\/} LRS spectra and the {\it IRAS\/} 12, 25, 60 and 100$\mu$m 

633: photometry measurements. The difference in the flux levels between the

634: {\it IRAS\/} and  {\it ISO\/} data is not unexpected, since these stars are 

635: variable. However, the shape of the spectrum does not change significantly 

636: between the two observations, suggesting that changes in the stellar 

637: luminosity do not significantly impact the structure and composition 

638: of the dust shells.

639: 

640: In order to determine the cause of the 10-13$\mu$m feature and the factors 

641: that govern its strength, width and position we have tabulated the barycentric 

642: position, feature-to-continuum ratio and equivalent width of the feature 

643: (Table~\ref{obsparam}). In addition, we have also tabulated where the 

644: barycenter of the SiC feature would be if the short wavelength side of the 

645: absorption is due to silicate (as has been postulated, see \S~\ref{sicabs}).

646: This, along with the feature-to-continuum ratio measured at 9.7 and 

647: 11.3$\mu$m, can be found in Table~\ref{obsparam}.

648: 

649: \subsection{The ``30$\mu$m'' feature}

650: 

651: Another prominent spectral feature exhibited by our sample of extreme carbon stars is the  so-called ``30$\mu$m'' feature.

652: This feature is relatively common amongst carbon-rich AGB stars, PPNe 

653: and PNe and  was first discovered in Kuiper Airborne Observatory ({\it KAO}) 

654: spectra of carbon stars and PNe \citep{forrest81}. It has been widely 

655: attributed to magnesium sulfide \citep[MgS; e.g.][]{gm85,nuth85,omont95,begemann94,hony2002}.

656: Modeling this feature is beyond the scope of the present work, but will be 

657: investigated in follow-up modeling. Our models make no attempt to fit the ``30$\mu$m'' feature.

658: %spectrum longwards of $\sim26\mu$m.

659: 

660: 

661: \section{Radiative Transfer Modeling}

662: \label{rtmodeling}

663: 

664: Radiative transfer modeling has been particularly useful in investigating 

665: extreme carbon stars 

666: \citep[see \S~\ref{rtmodels1};][]{Volk1992,Groenewegen1995,Volk2000}. 

667: %

668: We used the 1-D radiative transfer program DUSTY 

669: \citep{Ivezic1995, Nenkova2000}, to determine the effect of dust shell 

670: parameters on the emerging spectra from carbon stars. 

671: %

672: In all cases, the central star was assumed to be at 3000\,K (typical 

673: for an AGB star). \citet{Speck2000} and \citet{DePew2006} showed that 

674: changing this 

675: temperature by $\pm$1000\,K did not significantly change the radiative 

676: transfer model's spectra \citep[c.f.][]{Volk2000}. For simplicity, 

677: dust grains are assumed to be spherical. While DUSTY can include other grain 

678: shapes, this expands parameter space to create more degeneracy between models 

679: and is beyond the scope of the present work. 

680: %

681: 

682: \subsection{Radial dust density distribution}

683: \label{radialdist}

684: 

685: 

686: We assume a radial dust density distribution of 

687: $1/r^2$ which would reflect a constant mass-loss rate.

688: This choice is somewhat controversial and certainly needs justifying.

689: %

690: Current dust formation models suggest that carbon star mass-loss is expected 

691: to be modulated on several timescales, especially that of the pulsation cycle 

692: \citep[][and references therein]{woitke06}. Furthermore \citet{woitke06} has 

693: suggested that 

694: the dynamics in the dust-forming zones around carbon stars lead to 

695: inhomogeneous dust formation, producing fine scale structure in the density of 

696: the dust envelope.

697: %

698: While pulsation shocks are predicted to have a strong effect on local 

699: conditions \citep[e.g.][]{cherchneff06}, 

700: this is not reflected in temporal changes in the IR spectra of carbon stars 

701: \citep{corman08}.

702: %

703: Previous models of extreme carbon stars have included steeper a density 

704: drop-off \citep[see 

705: Table~\ref{prevmod};][]{Volk1992,Volk2000,Groenewegen1995,Groenewegen1998},

706: which is meant to represent the increasing mass-loss rate associated with the

707: onset of the superwind phase. However, \citet{vil02a,vil02b} 

708: showed that 

709: the hydrodynamics in the circumstellar shell wipe out density structure and 

710: leave a shell with a $1/r^2$ density distribution. In addition 

711: \citet{rrh83} showed that carbon star spectra can be well-fitted using such a 

712: density distribution. 

713: %

714: 

715: In our models we assume that modulations in density have 

716: been wiped out or are unimportant in determining the spectrum at these high 

717: optical depths, as we are clearly seeing an outer dust shell surface.

718: %

719: We also assume we are only detecting the dust that has formed since 

720: the onset of the superwind. We assume that there was a 

721: sudden increase in mass loss at some time in the last $\sim$10000\,years 

722: and that this mass-loss rate is now approximately constant, with small scale 

723: fluctuations being unimportant for dust properties.

724: Consequently, the $1/r^2$ density distribution suffices.

725: As will be seen in \S~\ref{results}, the impact of assuming a steeper the 

726: density distribution is to remove dependence on shell size, and thus remove 

727: the ability to out limits on timescales. If we were to adopt a  

728: $1/r^3$ density distribution, all the dust would effectively be 

729: contained close to the star and would reflect the total mass lost over only a 

730: relatively short period.

731: 

732: 

733: 

734: \subsection{Modeling grain-size distributions}

735: \label{gszmodel}

736: 

737: The issue of choosing a grain size distribution is interesting and certainly 

738: requires more discussion. Our initial modeling studies used an MRN distribution

739:  \citep[i.e., $n(a)$ proportional to $a^{-q}$, where $n$ is the number of the 

740: grains in the size interval $(a,a + da)$ and $q=3.5$;$a_{min}=0.005\mu$m; and 

741: $a_{max}=0.25\mu$m;][]{Mathis1977}. 

742: %

743: This was chosen because as will become evident below, 

744: we do not actually know the grain size 

745: distribution and MRN is as plausible as any other. 

746: %

747: However, the MRN distribution was developed for interstellar dust where the 

748: balance of formation and destruction is different from AGB circumstellar 

749: environments. \citet{dsg89} suggested that the grain-size

750: distribution created in the circumstellar environments of AGB stars has a 

751: steeper power law (i.e.\ $a^5$), while \citet[][KMH]{kmh} modified the MRN 

752: distribution to include an exponential fall-off term.

753: 

754: We should also consider the ``observational'' evidence for range and distribution 

755: of grain sizes in carbon star outflows. 

756: %

757: Meteoritic presolar SiC grains from carbon stars have a huge grain-size 

758: distribution, ranging for 1.5nm up to 26$\mu$m, with the majority of grains 

759: ($\sim$70\%) falling in the 0.3--0.7$\mu$m range 

760: \citep[see][]{daulton03,bern05,cn04}. 

761: Half the mass of the presolar SiC found in the Murchison meteorite is in 

762: grains larger than 0.6$\mu$m \citep{virag}. 

763: Carbon presolar grains follow a similar grain size distribution to SiC 

764: \citep{bern05}.

765: However, the sample may be biased 

766: in favor of large grains, which may be more apt to survive the journey through 

767: the ISM and incorporation into a solar system body. 

768: 

769: Conventional theories of 

770: grain growth cannot produce the largest grains. Since AGB stars typically have 

771: mass-loss rates $ < 10^{-5}$ M$_\odot$\,yr$^{-1}$ there should be an upper 

772: limit to the grain sizes of $\sim0.1\mu$m. 

773: However, this assumes an isotropic distribution of material. The existence of

774: meteoritic titanium carbide (TiC) within presolar carbon grains necessitates

775: density inhomogeneities in the gas outflows \citep[c.f., inhomogenities caused 

776: by turbulence in hydrodynamic models of carbon-rich dust formation 

777: regions;][]{woitke06}.

778: %

779: In addition, the grains must be \lapprox 1$\mu$m to produce the 11$\mu$m 

780: feature, implying that 

781: there is a large population of small grains (but not necessarily precluding 

782: large grains). Constraints on grain size were also discussed by \citet{mr87}

783: who found  an upper limit to the grain size of $\sim$0.1$\mu$m based on 

784: polarization measurements of the famous carbon star, IRC+10216. On the other 

785: hand, \citet{Groenewegen1997} used polarization measurements to limit the 

786: grain size to $0.1 \leq a \leq 0.35\mu$m and suggested that very small grains 

787: ($<$80 nm) may not exist around carbon stars (however, smaller SiC grains are found in meteorites). Meanwhile, \citet{jura94} argued 

788: the case for grains larger than 1$\mu$m in the circumstellar shells of 

789: IRC+10216.

790: 

791: Towards the end of the AGB the onset of the superwind may lead to mass-loss 

792: rates as high as a few $\times 10^{-4}$ M$_\odot$\,yr$^{-1}$ 

793: \citep[e.g.][]{hony2003} which could translate into larger grains 

794: \citep[see e.g.][and references therein]{bern05}.

795: However, the relationship between the evolution of carbon stars and the 

796: consequent evolution of grain sizes in their circumstellar shells was discussed

797: by \citet{Speck2005}. They argued that the increased mass-loss rates at the 

798: end of the AGB phase lead to smaller, rather than larger grains as suggested 

799: by meteoritic evidence. Even the 

800: highest observed mass-loss rates cannot account for formation of titanium 

801: carbide (TiC) grains unless the distribution of material is not 

802: spherically symmetric and density enhancements exist \citep{bern05}. Such a 

803: distribution of material makes the concept of grain size distributions even 

804: more complex.

805: 

806: The range of grain sizes used in previous radiative transfer modeling attempts 

807: varies hugely. 

808: %

809: The IRC+10216 models of \citet{griffin1990} required grain sizes limited to 

810: $5 \leq a \leq  50$nm. Similarly \citet{bdg95} required grains smaller than 

811: 50nm.

812: %

813: \citet{Groenewegen1997} also modeled IRC+10216 and found that the spectrum was 

814: best fit using a single grain size of 0.16$\mu$m (rather than a distribution 

815: of grain sizes). The single-grain size approach was also adopted by 

816: \citet{Volk2000} who assume a single grain size of 0.1$\mu$m.

817: This is supported by the success of the early carbon star models of 

818: \citet{rrh83} who also used a single grain size (0.1$\mu$m).

819: However, since extreme carbon stars are expected to be the direct precursors  

820: for post-AGB objects, it may be more appropriate to consider the models of 

821: post-AGB stars. Such modeling efforts have found they need grains up to 

822: millimetre- or even centimetre- in size \citep[e.g.][]{jura2000,meixner2002}.

823: \citet{meixner2002} used a KMH-like distribution with 

824: an effective maximum grain size of 200$\mu$m (hereafter referred to as KMH200).

825: Meanwhile, \citet{szc97} modeled a carbon-rich post-AGB object with a 

826: power-law distribution of grains in the range 0.01--1$\mu$m.

827: 

828: Given the range of possible grain sizes and distributions, it is difficult to 

829: know how best to model the dust.

830: \citet{Groenewegen1995} argued that the specific grain-size is not important 

831: as long as the grains are small enough that the absorption and scattering 

832: properties are independent of grain size. However, given the arguments for a 

833: population of large grains, limiting grains to smaller than 0.1$\mu$m is 

834: unrealistic.

835: 

836: In order to determine the effect of the choice of grain sizes on the model 

837: spectra we have generated models using five additional grains size 

838: distributions:

839: 1) MRN-like with a steeper power law, i.e. $a_{min} = 0.005\mu$m, 

840: $a_{max} = 0.25\mu$m; $q=5$ \citep[as suggested by][]{dsg89};

841: 2) only 0.1$\mu$m-sized grains 

842: \cite[as used or suggested by][]{Volk2000,rrh83,Groenewegen1995,mr87};

843: 3) the dominant meteoritic grain sizes, i.e. 0.1--1$\mu$m only;

844: 4) the standard KMH distribution (i.e. $n(a) \propto a^{-q}e^{a/a_0}$,

845: $a_{min} = 0.005\mu$m, $a_{0} = 0.2\mu$m; $q=3.5$); and finally

846: 5) KMH200: KMH with $a_0 = 200\mu$m.

847: %

848: %

849: The results of this study are shown in Fig.~\ref{gsizefig}, which shows the 

850: effect of changing the grain size distribution while keeping all other 

851: parameters constant.

852: 

853: 

854: If we examine the differences in the spectra generated by changing the 

855: grain-size distributions, this can be understood in terms of the absorption 

856: efficiency of the grains and breadth of the grain size distribution.

857: Changing from our default grain size distribution (MRN) to a modified MRN-like 

858: distribution with a steeper power law drop off ($q=5$) as suggested by 

859: \citet{dsg89} makes very little difference to the model spectral energy distribution (SED).

860: Likewise, switching from MRN to KMH has little effect on the overall shape of 

861: the SED.

862: That the MRN, $q=5$ and the KMH models are so similar is because they are 

863: basically weighted towards the same small grains. While the weighting is 

864: different, the same-sized small grains dominate the SED.

865: %

866: However, changing the size distribution to include larger dust grains as used 

867: in models of post-AGB stars \citep[e.g.][]{meixner2002} has an major effect. 

868: The SED shifts to peak at much shorter wavelengths. This is because of the 

869: reduction in the number of small grains in order to include larger grains. The 

870: proportion of larger 

871: grains is small, but the removal of the smaller grains makes it possible for 

872: the stellar photons to penetrate deeper into the dust shell and provides a 

873: large population of warmer grains, resulting in a warmer SED.

874: In both the KMH200 and 0.1$\mu$m cases there is a lack of very small grains 

875: which absorb a lot of stellar photons and change the temperature distribution 

876: (i.e. 

877: after the first layer of dust the temperature is lower, but if there are no 

878: small grains the stellar photons penetrate further.) Since size distributions 

879: similar to KMH200 are typically associated with long-lived dust disks, this 

880: particular distribution is not considered further.

881: 

882: For the meteoritic grain size distribution, the short wavelength side of the 

883: SED is similar to the default (MRN) model, but now the SED is much narrower. 

884: This can be explained by the very narrow range of grain sizes. There are no 

885: very small grains that can be easily heated and thus the shorter wavelength 

886: emission seen in KMH200 does not occur, but the lack of small grains also 

887: allows deeper penetration of stellar photon leading to a narrower temperature 

888: distribution.

889: 

890: This grain size study suggests that for most adopted grain-size distributions, 

891: the resulting SEDs will be equivalent and we assume the MRN distribution as 

892: ``generic''. However, the meteoritic grain-size distribution can narrow the 

893: overall SED. 

894: %

895: For this reason, our modeling efforts concentrate on the generic (MRN) and ``meteoritic'' grain size distributions, where the ``meteoritic'' is taken to be an extreme amongst the range of reasonable grain-size assumptions.

896: %

897: The impact of choosing different grain size distributions and the implications 

898: of these differences for the potential errors in our models

899: will be discussed in \S~\ref{results}. However, essentially, most grain-size 

900: distributions will yield the same results except for ``meteoritic'' and 

901: KMH200. In both cases changes in optical depth, inner dust temperature and/or 

902: relative geometrical shell thickness can be manipulated to fit the spectrum.

903: There is a degeneracy in model fits due to the relationship between these 

904: three parameters that will be discussed in the next section.

905: As will be seen in the \S~\ref{results}, the grain size effects cannot be 

906: ignored.  In one source the need for ``meteoritic'' grain-size distribution 

907: is clear.

908: 

909: \subsection{Parameter space investigated}

910: 

911: In addition to grain sizes and radial density distribution, the variables investigated with DUSTY are the 

912: inner dust shell temperature (T$_{\rm inner}$), 

913: optical depth (specified at 10$\mu$m; $\tau_{10\mu \rm m}$), 

914: dust composition, and 

915: the geometrical thickness of the dust shell, $\xi = R_{\rm out}$/R$_{\rm in}$, 

916: where

917: R$_{\rm in}$ and R$_{\rm out}$ are the inner and outer radii of the dust 

918: shell, respectively. 

919: %

920: The optical constants for the dust components came from 

921: \citet{Pegourie1988,Hanner1998} and \citet{DL84} for SiC, amorphous carbon

922: and graphite, respectively.

923: %

924: In nearly all cases it was possible to generate more than one model to fit the spectra, consequently we also investigate this degeneracy in parameter space and

925: look for realistic ways to restrict it.

926: 

927: 

928: \subsubsection{Degeneracies in radiative transfer modelling}

929: \label{degen}

930: 

931: 

932: There is a clear degeneracy in the models because of the relationship between

933: certain parameters, e.g. optical depth and geometric shell thickness, or 

934: optical depth and the temperature at the inner edge of the dust shell. 

935: \citet{ie97} discussed these degeneracies and the relationship between the 

936: different input parameters in radiative transfer models, but we 

937: need to understand these relationship  if we are to 

938: understand what our models mean.

939: 

940: The wavelength-dependent optical depth, $\tau_\lambda$ is defined by:

941: 

942: \begin{equation}

943: \label{tau0}

944: %

945: d\tau_\lambda = \rho(r) \kappa_\lambda dr 

946: %

947: \end{equation}

948: 

949: \noindent

950: where $\rho(r)$ is the density of the absorbing/scattering particles 

951: (i.e. the  dust grains) and $\kappa_\lambda$ is the wavelength dependent 

952: opacity of the assemblage of particles along the line of sight.

953: 

954: 

955: \[ \tau_\lambda = \int^{R_{\rm out}}_{R_{\rm in}}{\rho(r) \kappa_\lambda dr} \]

956: 

957: 

958: \noindent

959: where $R_{\rm in}$ is the inner dust shell radius and $R_{\rm out}$ is the 

960: outer dust shell radius.

961: While $\kappa_\lambda$ is dependent on the density distribution of the grains,

962: maintaining the size, shape and composition 

963: (and crystal structure) of the grains means that $\kappa_\lambda$ will not 

964: change significantly. 

965: For simplicity we assume   $\kappa_\lambda$ remains constant.

966: %

967: In our models we assume the density of the dust shell drops off as 1/$r^2$ 

968: from the central star. In addition, our models specify the relative geometrical

969: thickness of the dust shell as $\xi = R_{\rm out}/R_{\rm in}$. Thus we get:

970: 

971: \begin{equation}

972: \label{tau1}

973: %

974: \tau_\lambda = \kappa_\lambda \left[ \frac{\xi - 1}{\xi R_{\rm in}} \right]

975: %

976: \end{equation}

977: 

978: 

979: \noindent

980: The value of $R_{\rm in}$ is set by the values we input for the star's 

981: effective temperature and the inner dust radius (or condensation

982:  temperature, $T_{\rm inner}$).

983: If we know the temperature and luminosity of the star, we can use a 

984: $T(r) \propto r^{-\frac{1}{2}}$ temperature distribution to determine the

985: relationship between  $T_{\rm inner}$ and $R_{\rm in}$ and 

986: substituting into Eq.~\ref{tau1} we get:

987: 

988: \begin{equation}

989: \label{tauTxi}

990: %

991: \tau_\lambda = \sqrt{\frac{16 \pi \sigma}{L_\star}}~\kappa_\lambda T_{\rm inner}^2 \left[ \frac{\xi - 1}{\xi}  \right]

992: %

993: \end{equation}

994: 

995: 

996: 

997: Here we assume that the grains are blackbodies because they are largely 

998: carbon. Including the albedo would allow for a more accurate calculation, but 

999: this would depend on detailed dust parameters (like crystal structure), and the error incurred by our assumption is small 

1000: (i.e.\ significantly less than an order of magnitude).

1001: 

1002: Therefore, according the Eq.~\ref{tauTxi}, for a dust shell with constant 

1003: relative shell thickness ($\xi$) the optical depth 

1004: should increase with the square of the inner dust temperature. Alternatively, 

1005: if the inner dust temperature is fixed, then increasing the relative shell 

1006: thickness should decrease optical depth  a little (as $(\xi - 1) / \xi $). 

1007: This latter effect becomes negligible for large geometric sizes.

1008: 

1009: 

1010: \subsection{Determining the Dust Condensation Temperature}

1011: \label{constraints}

1012: 

1013: \subsubsection{Theoretical dust condensation models}

1014: \label{condmod}

1015: 

1016: As seen in the previous section, the value of the dust condensation 

1017: temperature is a source of degeneracy on radiative transfer modeling.

1018: %

1019: In order to reduce this degeneracy we turn to dust condensation 

1020: theory to determine a theoretical dust condensation temperature 

1021: (i.e. $T_{\rm inner}$) that is appropriate for our stars.

1022: 

1023: As discussed in \S~\ref{meteor}, many presolar grains can be attributed to 

1024: carbon stars and are valuable resources for investigation of dust formation 

1025: regions. For example, in presolar grains titanium carbide (TiC) is found in 

1026: the center of carbon (C) 

1027: grains from AGB stars, but only one SiC grain has been found coated in carbon 

1028: \citep{cn04, bern05}.

1029: %

1030: Consequently there have been many studies of the theoretical condensation 

1031: sequence in Galactic carbon stars in attempts to constrain the physical 

1032: parameters of the dust condensation regions.

1033: These studies showed how the condensation sequence of C, TiC and SiC is 

1034: dependent on various parameters, most notably C/O ratio and gas 

1035: pressure\footnote{%

1036: Gas pressure is a measure of the mass-loss rate ($\dot{M}$) convolved with the 

1037: photospheric temperature (T$_\star$) and outflow velocity ($v_{\rm exp}$).} 

1038: \citep{lf95,sw95}. 

1039: %

1040: %\citet{sw95} found that graphite (carbon) condensation 

1041: %temperature (T$_{cond}$) is strongly dependent on C/O, while the T$_{cond}$ 

1042: %for TiC and SiC are not.

1043: %Conversely, the condensation temperature of carbon is more or less 

1044: %independent of gas pressure, but T$_{cond}$ for TiC and SiC varies strongly 

1045: %with gas pressure. However, \citet{lf95} found that C, TiC and SiC are all 

1046: %strongly dependent on both gas pressure and C/O ratio.

1047: %

1048:  \citet{sw95} argued 

1049: that if carbon forms at a higher condensation temperature, closer to the star 

1050: than SiC, there is a significant decrease in 

1051: the amount of carbon available in the gas, and thus SiC and C do 

1052: not form simultaneously, resulting in naked SiC grains. Therefore, for Galactic

1053: sources, the condensation sequence in the majority of carbon stars should be 

1054: TiC --- C --- SiC, in order to produce the coated TiC grains and uncoated 

1055: (naked) SiC grains seen in the meteoritic presolar grains samples. 

1056: %

1057: Observational evidence for naked SiC grains is discussed by \citet{Speck2005}.

1058: %

1059: \citet{sw95} argued that from kinetic and stellar model considerations, dust 

1060: grains should form in the pressure range 2 $\times 10^{-7} <$ P $< 

1061: 4 \times 10^{-5}$ bars\footnote{%

1062: The expected range of gas pressures in the dust formation zone for 

1063: O-rich AGB stars in the LMC is 10$^{-7}$ bars \lapprox P \lapprox 10$^{-4}$ 

1064: bars \citep{dijk05}}.

1065: %

1066: %Therefore, to form dust in the sequence TiC --- C --- SiC, they need 

1067: %1.04 $<$ C/O $<$ 1.2. For Galactic carbon stars

1068: 

1069: 

1070: \citet{lf95} also modeled the effect of C/O and pressure on the condensation 

1071: sequence in carbon stars, as well as the effect of {\it s}-process and 

1072: nitrogen  abundances. They also briefly discuss the effect of metallicity.

1073: The general trends in condensation temperatures are: 

1074: (1) all condensation temperatures decrease as the gas pressure decreases; and 

1075: (2) At C/O $>$ 1 the condensation temperature of graphite increases with C/O 

1076: (for a given pressure and otherwise constant composition).

1077: 

1078: Figure~\ref{lfptspace} shows how the condensation temperature of carbon and 

1079: SiC vary with the gas pressure in the dust condensation 

1080: zone. For the range plotted carbon always forms first from a cooling gas. 

1081: Very high pressures are required for SiC to form before carbon.

1082: %

1083: For solar metallicity and C/O = 1.05, SiC forms before carbon for 

1084: P $\geq 3.4 \times 10^{-5}$ bars. 

1085: As C/O increases, the minimum pressure required to form SiC first increases.

1086: %

1087: Above C/O $\sim$1.5 carbon always forms before SiC. The exact C/O ratio at 

1088: which 

1089: the carbon forms before SiC depends on pressure. Therefore, in order to 

1090: account for observations of SiC features in the Galaxy, and the presolar 

1091: grain record, we can restrict the P--C/O space such that, for low C/O the gas 

1092: pressure must remain low, but for higher C/O the pressure can be higher. 

1093: This can be used to constrain the dust forming environment around Galactic 

1094: carbon stars. 

1095: 

1096: Studies of carbon star spectra in the Galaxy and the Magellanic Clouds have 

1097: lead to different interpretations with respect to the condensation sequence.

1098: \citet{Lagadec2007} argued for a sequence in which SiC forms before C in the 

1099: Galaxy, whereas SiC and C form together in the Large Magellanic Cloud (LMC) 

1100: and the sequence is reversed (C, then SiC) in the Small Magellanic Cloud (SMC). However, they also suggest that the change in the strength of the features is 

1101: due to the lower number of Si atoms available for SiC formation.

1102: The proposed Galactic condensation sequence is at odds with both the models 

1103: and the meteoritic evidence. \citet{leisenring} support the condensation 

1104: sequence in which C forms before SiC for the 

1105: Galactic carbon stars, while finding that the Magellanic Clouds tended to form 

1106: SiC first, followed by simultaneous condensation of SiC and C.

1107: \citet{Speck2006} used the \citet{lf95} model to explain an unusual LMC carbon 

1108: star spectrum, which suggests that the condensation sequence is sensitive to 

1109: both metallicity and mass-loss rate.

1110: 

1111: 

1112: \subsubsection{P-T space in the condensation zone around extreme carbon stars}

1113: \label{ptcalc}

1114: 

1115: 

1116: In order to constrain the input parameters to our model we need to be able to 

1117: determine the pressure-temperature space in the dust condensation zone.

1118: %

1119: For a mass-losing star with a mass-loss rate $\dot{M}$ and 

1120: an expansion velocity of $v_{\rm exp}$, 

1121: the density $\rho$ of the circumstellar shell 

1122: at a radius $r$ is given by:

1123: 

1124: 

1125: \[ \rho = \frac{\dot{M}}{4 \pi r^{2} v_{\rm exp}} \]

1126: 

1127: 

1128: If we know the temperature and luminosity of the star and the composition of 

1129: the outflowing material we can combine this information with the Ideal Gas Law 

1130: and a $T(r) \propto 1/\sqrt{r}$ temperature distribution to determine 

1131: the

1132: gas pressure at the condensation radius, where the condensation radius is the

1133: distance from the star where the gas has the condensation temperature 

1134: ($T_{\rm cond}$).

1135: 

1136: For simplicity, the solid and gas phases are assume to be at the same 

1137: temperature. While this is clearly a simplification \citep[e.g.][]{chigai}, 

1138: the temperature difference is small compared to the difference needed to 

1139: significantly affect dust formation.

1140: %

1141: We assume that  most of the outflowing material is atomic hydrogen. In fact it will probably be a mixture of atomic and molecular hydrogen (H$_2$) since H$_2$

1142: forms around 2000\,K and the temperature in the outflow is decreasing from 

1143: the stellar surface temperature of $\sim$3000\,K to the dust condensation 

1144: temperature in the 1000--1800\,K range. 

1145: %

1146: %However, without dust grains the gas may not form H$_2$ efficiently, so we assume for simplicity that the all the gas is atomic hydrogen. 

1147: An entirely molecular hydrogen gas would halve 

1148: the gas pressure compared to the the atomic gas. 

1149: %

1150: However, we also assume published CO outflow velocities, which reflect the 

1151: speed of the outflowing material after radiation pressure acceleration. 

1152: Adopting the pre-dust-formation outflow speed (\lapprox 5km/s), would increase 

1153: the pressure.

1154: %

1155: Thus using published values for our sample stars' luminosities, 

1156: mass-loss rates and expansion velocities, we can estimate where their dust 

1157: condensation zones fall in P--T space.

1158: 

1159: 

1160: \subsubsection{Comparison of P--T space for dust condensation models and sample 

1161: stars}

1162: \label{comppt}

1163: 

1164: Figure~\ref{lfptspace} shows how the P--T space for the dust condensation zone 

1165: for our target stars compares to condensation models in P--T space.

1166: It is clear the pressure is never high enough for SiC to form before 

1167: carbon.  Since our sample stars are expected to be the carbon stars with the 

1168: highest mass-loss rates, it implies that this NEVER happens in Galactic carbon 

1169: stars. 

1170: This agrees with the meteoritic evidence which suggests that SiC does

1171: not get coated in carbon and supported the idea that variations in the 

1172: strength of the SiC feature are related to self-absorption.

1173: 

1174: 

1175: The dust condensation temperature is dependent on the pressure in the gas from 

1176: which the dust forms.

1177: In addition, the P--T space occupied by the target sources suggests 

1178: that carbon grains will form at temperatures \gapprox 1600\,K. 

1179: If the C/O is 

1180: very high, then graphite could form as high as $\sim$1800\,K. 

1181: However, Galactic 

1182: carbon stars for which the C/O ratios have been measured show that it is 

1183: in the range 1 to 1.8, with a mean C/O ratio of $\sim$1.15 

1184: \citep{lamb86, olof93a, olof93b}. The precise C/O for our sample is not known,

1185: but even low C/O stars yield $T_{\rm cond}$ \gapprox 1550\,K 

1186: (see Fig.~\ref{lfptspace}).

1187: Previous models of carbon stars have assumed much lower inner dust 

1188: temperatures.

1189: The pressures and temperatures in the gas around these 

1190: stars meet the criteria for forming carbon dust at $T_{\rm cond}$ \gapprox 

1191: 1600\,K, (precise temperature depends on the C/O ratio). Therefore it should 

1192: form at these high temperatures.

1193: 

1194: 

1195: Once dust starts to form, the radial temperature profile will change due to 

1196: absorption of starlight by dust grains.

1197: Therefore, we use the comparison above only to determine the inner dust radius.

1198: Using radial temperature profiles from our models we show that the temperature 

1199: drops significantly more rapidly than $1/\sqrt{r}$ (see Fig.~\ref{lfptspace}), 

1200: which suggests that the 

1201: next condensate (SiC) forms fairly close to the inner dust radius and thus 

1202: mitigates the problem with DUSTY that the grains are assumed to be co-spatial. 

1203: Indeed the models show that the temperature in the dust shell drops to the SiC 

1204: formation temperature at $\sim$1.3R$_{\rm in}$ for all cases, which is small 

1205: compared to the shell thickness, even for the thinnest shells.

1206: 

1207: \subsection{Constraining Dust Composition}

1208: 

1209: Since our stars are carbon-rich, we limit the models to only including 

1210: carbonaceous species such as graphite, amorphous carbon and silicon carbide.

1211: The choice of carbon grains is equivocal. We do not know whether the grains 

1212: composed mostly of carbon are glassy, poly-nanocrystalline or well-ordered.

1213: The meteoritic presolar ``graphite'' grains suggest that circumstellar dust 

1214: can contain either well-ordered graphite or poly-nanocrystalline-graphene 

1215: grains. It should be noted, however, that even the disordered graphene sheets 

1216: are considered to be more graphitic than amorphous. 

1217: 

1218: Amorphous carbon essentially consists of a mixture of $sp^2$ (graphite-like) 

1219: and $sp^3$ (diamond-like) carbon bonds. On heating, $sp^3$ bonds tend to 

1220: convert to $sp^2$ bonds, thus graphitizing the amorphous carbon, but it will 

1221: remain dense like the amorphous phase 

1222: \citep[2.8\,g/cm$^3$][]{comelli,saada,kelires}. 

1223: At $\sim$1300\,K the fraction of 

1224: graphitic bonds is $\sim$90\%. 

1225: %

1226: Comparison of the gas pressure in the circumstellar outflows to graphite 

1227: formation temperature in \S~\ref{comppt} shows that graphite should be able to 

1228: form at temperatures significantly above 1300\,K. Even if solid state carbon 

1229: forms as a chaotic solid, at these temperatures it will quickly anneal to a 

1230: graphitic form.

1231: %

1232: Consequently, we argue that graphite is the better choice of carbonaceous 

1233: material for modeling extreme carbon star dust shells.

1234: 

1235: 

1236: The limitations of the use of the \citet{Pegourie1988} data are discussed in 

1237: \citet{Pitman2007}. Clearly, this data cannot produce the broad absorption 

1238: feature seen in the observations. However, we can use the relative changes in 

1239: composition from star to star as a guide to understanding why these stars 

1240: have the features we see.

1241: 

1242: 

1243: 

1244: \section{Radiative Transfer Modeling Results}

1245: \label{results}

1246: 

1247: The results of the radiative transfer modeling can be seen in 

1248: Fig.~\ref{modelfit1}, \ref{modelfit2} and \ref{modelfit3}. 

1249: The parameters used in each case can be seen in Table~\ref{tabfit1}.

1250: For each source, we have produced models using both MRN (generic) and 

1251: ``meteoritic'' (extreme) grain size distributions.

1252: %

1253: 

1254: \subsection{The effect of grain-size distribution}

1255: \label{gsizemod}

1256: 

1257: Table~\ref{tabfit1} shows the model input parameters for both grainsize distributions. 

1258: In all cases switching from MRN to ``meteoritic'' grainsizes leads to the need for increased SiC component 

1259: (compared to graphite), typically requiring a three- or fourfold increase in the SiC fraction.

1260: 

1261: There are other general trends including the need for decreased optical depths ($\tau_{10\mu \rm m}$) and 

1262: increased geometric shell thickness. However, these trends do not hold for all objects. 

1263: In the cases of IRAS\,00210+6221 and IRAS\,06582+1507, the only difference in parameters between 

1264: the different grain size models is the fraction of SiC. 

1265: IRAS\,01144+6658 is the only source for which optical depth was increased and shell thickness was 

1266: decreased in the ``meteoritic'' model.

1267: 

1268: 

1269: 

1270: \subsection{Dust Shell Thicknesses}

1271: \label{timescalesect}

1272: 

1273: As discussed in \S~\ref{degen}, there is a degeneracy between relative shell 

1274: thickness, inner dust temperature and optical depth. The inner dust 

1275: temperature variability has been limited to 1550--1800\,K by 

1276: theoretical considerations (see \S~\ref{condmod}). 

1277: %

1278: It is possible to tweak parameters such as 

1279: shell thickness, inner dust temperature and optical depth and get almost 

1280: identical models. Fig.~\ref{modelfit23166} shows two almost identical model 

1281: spectra for IRAS\,23166+1655  with different inner dust temperatures, 

1282: optical depths and shell 

1283: thicknesses. However, reducing the inner dust temperature merely requires 

1284: reduction of the optical depth.

1285: %

1286: While increasingly geometrically large dust shells can be accommodated by 

1287: decreasing optical depth (see Eq.~\ref{tauTxi}), at some point, this also 

1288: breaks down, as it leads to a significant population of colder grains which 

1289: emit too much at long wavelengths. In this way we can use the models to place 

1290: to an {\it upper limit to the shell thickness}. 

1291: %

1292: Table~\ref{obsparam2} lists the published expansion velocities for our 

1293: extreme carbon stars.

1294: From the model parameters, the physical value for 

1295: $R_{\rm out}$ can be calculated,

1296: which gives the physical size of the dust shell. 

1297: Since the models give an upper limit to the shell thickness, 

1298: this upper limit to $R_{\rm out}$ together with the expansion velocity ($v_{\rm exp}$) 

1299: was used to calculate the 

1300: time since the outermost edge of the shell was ejected from the star.

1301: The resulting ages of the dust shells are listed in Table~\ref{timescaletab}.

1302: 

1303: As can be seen in Fig.~\ref{gsizefig}, it is possible to accommodate thick 

1304: shells, if we assume a much lower $T_{\rm inner}$ (1000\,K), and assume that 

1305: the dust grain are composed of amorphous rather than graphitic carbon, but 

1306: these parameter values are precluded by the theoretical constraints

1307: on $T_{\rm inner}$ and composition/crystal structure discussed above.

1308: Clearly, since our constraints are theoretical, they may change as hypotheses 

1309: are refined.

1310: 

1311: The timescales for increased mass-loss are model dependent. As can be seen in 

1312: Table~\ref{tabfit1} using a ``meteoritic'' grain size distribution generally 

1313: requires a geometrically thicker shell, as well as lower values for 

1314: $\tau_{10\mu \rm m}$ and higher percentages of SiC. However,

1315: timescales associated with our dust shells are always very short (less than a 

1316: few thousand years) regardless of grain-size distribution. 

1317: Moreover, our timescales are consistent with those derived for the 

1318: ``superwind'' 

1319: seen in post-AGB objects \citep[e.g.][]{skinner97,meixner2004}.

1320: These timescales are too short to be associated with the theoretical

1321: superwind \citep[e.g.][]{vw93}, 

1322: which is expected to last up to $\sim$10\% of the duration of the 

1323: thermally-pulsing AGB phase.

1324: However, if we compare the number of extreme carbon stars to the 

1325: total number of carbon stars in the Galaxy, we find that extreme 

1326: carbon stars constitute only $\sim$0.1\% of the total 

1327: C-star population. The thermally pulsing AGB phase is expected to last

1328: $\sim 10^6$ years, and the time a star spends as a C-stars is even 

1329: shorter \citep[e.g.][]{Lagadec2007}. Consequently, we might expect the 

1330: dust-obscured phase to only last 

1331: $\sim 10^3$ years. This is consistent with the model timescales, 

1332: although the MRN-models still appear to have very short timescales.

1333: %

1334: \citet{LZ08} suggested that the trigger for the superwind is a combination of 

1335: luminosity and carbon abundance. Although the duration of the C-star phase 

1336: maybe be $\sim 3 \times 10^5$\,years, for much of this time, a C-star will be

1337: below the critical carbon abundance required to drive the superwind and 

1338: obscure the star. 

1339: 

1340: How does the extreme carbon star phase fit into the broader C-star evolution?

1341: Many, if not most, carbon-rich post-AGB stars showed marked axisymmetric 

1342: morphologies \citep{meixner1999,waelkens,sahai,soker}. 

1343: %

1344: The extreme carbon stars are expected to be the 

1345: direct precursors of these objects \citep[e.g.][]{skinner98}, but presently 

1346: show little evidence for 

1347: axisymmetry\footnote{one would expect to see more near-IR emission in the 

1348: spectrum of strongly axisymmetric objects due to scattering of starlight 

1349: escaping in the bipolar axis direction.}.

1350: The cause and timing of this axisymmetric structure is not known, but is 

1351: believed to occur at the very end of the AGB phase.

1352: \citet{ds06} showed that significant axisymmetry is not expected to develop 

1353: until the last few tens or hundreds of years of the superwind phase.

1354: The onset of axisymmetry also leads to an optically thicker toroid of dust, as 

1355: the circumstellar shell becomes equatorially enhanced.

1356: We suggest that the very short timescales associated with our model results 

1357: may indicate that the extreme AGB stars are in the process of developing 

1358: axisymmetry, but that this has not developed to the point of allowing large 

1359: amounts of NIR scattered light into the spectrum.

1360: 

1361: 

1362: 

1363: 

1364: \subsection{Dust Shell Density Distribution}

1365: 

1366: In \S~\ref{rtmodeling} we argued the case for maintaining a dust density 

1367: distribution that follows a 1/$r^2$ law. However, it has been argued that the

1368: increasing stellar luminosity and the onset of the superwind phase should 

1369: give rise to a steeper density drop-off.

1370: 

1371: Following the same arguments as shown in \S~\ref{degen}, 

1372: giving rise to Eq.~\ref{tauTxi} we can derive an equation in which the 

1373: exponent of the density power law is a variable.

1374: %

1375: This gives us the relationship between the optical depth ($\tau_\lambda$),

1376: inner dust temperature ($T_{\rm inner}$), relative shell thicks ($\xi$) and 

1377: the exponent of the density power law ($x$):

1378: \begin{equation}

1379: \label{tau_rTxi}

1380: %

1381: \tau_\lambda = \kappa_\lambda \left[ \frac{\xi^{x-1}-1}{(x-1)\xi^{x-1}} \right]\left(\sqrt{\frac{16 \pi \sigma}{L_{\star}}} T_{\rm inner}^2 \right)^{x-1}   

1382: %

1383: \end{equation}

1384: 

1385:  

1386: \noindent

1387: where $\kappa_\lambda$ is the wavelength dependent opacity of the assemblage 

1388: of particles along the line of sight;

1389: $L_\star$ is the luminosity of the star; and

1390: $\sigma$ is the Stefan-Boltzmann constant.

1391: 

1392: As an example of the effect this, we assume the exponent, $x$ = 3 

1393: (which is the highest value in the previous models).

1394: In this case, the relative shell thickness dependence dwindles, and the 

1395: optical depth is essential strongly dependent on the inner dust 

1396: temperature only. 

1397: Since the previous models have severely underestimated the inner 

1398: dust temperature, they have also underestimated the optical depth

1399: \citep[which may explain][]{Groenewegen1995}.

1400: 

1401: The lack of dependence on shell thickness at $x > 2$ leads to a situation 

1402: where most of the dust is essentially confined to a region close to the star 

1403: and the outer dust becomes negligible, and so this confinement to the inner 

1404: region is effectively the same as our assumption that there was a 

1405: sudden increase in mass loss in the recent past.

1406: %

1407: However, we attempted to model our sources with a $1/r^3$ radial 

1408: density distribution and found that we cannot match the shape of the SED 

1409: without restricting the geometrical shell thickness.

1410: Using the $1/r^3$ radial density distribution with the MRN grain-size

1411: distribution and $R_{\rm out}/R_{\rm in}$ \gapprox 20 produces a SED that has 

1412: too much emission longwards of $\sim20\mu$m. 

1413: Therefore, even if the radial density distribution is indicative of a steep 

1414: increase in mass-loss rate the shells still need to be geometrically thin, and 

1415: the shell thickness and percentage of SiC needed are essentially the same as 

1416: for the  $1/r^2$ models

1417: (see Fig.~\ref{modelfit23166} for an example.) 

1418: %

1419: If the ``meteoritic'' grain-size distribution is used in conjunction with a 

1420: $1/r^3$ drop-off, the observed SED cannot be matched. The model SED 

1421: becomes too narrow, and increasing the outer dust-shell radius does not help.

1422: 

1423: 

1424: 

1425: 

1426: \subsection{IRAS~17534$-$3030}

1427: \label{17534}

1428: 

1429: 

1430: IRAS\,17534$-$3030 is exceptional amongst the present sample of extreme carbon stars, 

1431: and certainly requires more discussion. 

1432: As can be seen in the flux-calibrated and continuum-divided spectra in Fig.~\ref{obsfig}, 

1433: IRAS\,17534$-$3030 does not exhibit the usual molecular absorption bands around 13.7$\mu$m 

1434: (due the C$_2$H$_2$) and shortwards of $\sim8\mu$m (due to both C$_2$H$_2$ and HCN), 

1435: which are present in the other sample sources. Its 11--13$\mu$m feature is intermediate 

1436: between the narrow feature exemplified by IRAS\,23166+1655 and the broad feature 

1437: exemplified by IRAS\,19548+3035, indicating that at least some of whatever substance 

1438: causes the broadening is present around IRAS\,17534$-$3030.

1439: One of the suggested carriers for the short wavelength broadening of the 10-13$\mu$m 

1440: absorption feature is C$_3$ \citep{zijl06}. However, C$_3$ is expected to be photospheric, 

1441: rather than circumstellar, which probably precludes its detection in optically obscured stars.

1442: Moreover, the theoretical spectrum of C$_3$ 

1443: from \citet{jorg} shows a strong absorption close to the $\sim5\mu$m CO 

1444: line, which is stronger than the $\sim11\mu$m feature. The spectrum of IRAS\,17534$-$3030

1445: does not show this 5$\mu$m absorption band. This, together with the lack of other 

1446: molecular absorption feature provides evidence that the

1447: broadening of the 10-13$\mu$m feature is not molecular in origin.

1448: 

1449: In addition to the lack of molecular absorption in its spectrum, 

1450: IRAS\,17534$-$3030 is unique is another way: it cannot be modeled with the MRN grain-size distribution.

1451: Modeling of this source requires the ``meteoritic'' grain-size distribution because of its narrow SED.

1452: Whereas our other sources can be fitted with either MRN or ``meteoritic'' grains, IRAS\,17534$-$3030 

1453: cannot.

1454: 

1455: 

1456: \subsection{Impact of the dust condensation temperature}

1457: 

1458: The increase in the inner dust temperature from \lapprox\,1000\,K in previous 

1459: models up to 

1460: \gapprox\,1600\,K decreases the inner dust radius significantly.

1461: Consequently the flux of energy from the star hitting the inner dust radius 

1462: is increased, leading to a greater acceleration and consequently a more 

1463: effectively dust-driven wind. If graphite is formed by annealing of amorphous 

1464: carbon, the grain density should remain high (with $\rho$=2.8g/cm$^3$) and 

1465: the increased acceleration is entirely due to increases flux of stellar 

1466: photons.

1467: However, if graphite forms directly, rather than by 

1468: annealing of amorphous carbon, the radiation pressure effect is further 

1469: enhanced, because graphite grains generally have a 

1470: lower density than amorphous carbon. Assuming 0.1$\mu$m-sized grains,

1471: the acceleration felt by graphite grains (with $\rho$=2.2g/cm$^3$; c.f. 

1472: meteoritic presolar grains; see \S~\ref{presolc}) at 

1473: T$_{\rm inner}$ = 1800\,K is 13 times greater than the 

1474: acceleration felt by amorphous carbon grains (with $\rho$=2.8g/cm$^3$) 

1475: at T$_{\rm inner}$ = 1000\,K.

1476: 

1477: 

1478: 

1479: \subsection{Correlations between observed and model parameters}

1480: \label{correlsect}

1481: 

1482: 

1483: In the course of this investigation we have compiled a large number of 

1484: parameters for these stars. For instance, Table~\ref{obsparam} lists the peak 

1485: position, peak strength, FWHM and equivalent width 

1486: of the $\sim11\mu$m absorption feature.

1487: Table~\ref{obsparam2} lists the published mass-loss rates, luminosities and 

1488: expansion velocities for our sample stars.

1489: 

1490: 

1491: 

1492: In order to understand the physical conditions that give rise to the observed 

1493: 11$\mu$m absorption feature, we have sought correlations between various 

1494: parameters associated with the sample stars (as found in 

1495: Table~\ref{obsparam}, \ref{obsparam2}, \ref{tabfit1}).

1496: We have looked for correlations between each of the following:

1497: mass-loss rate (from CO observations); 

1498: expansion velocity (from CO observations);

1499: stellar luminosity;

1500: strength of the observed absorption;

1501: FWHM of the absorption;

1502: barycentric position of the absorption;

1503: equivalent width of the absorption;

1504: and model parameters.

1505: 

1506: 

1507: %Loup et al 1993 

1508: \citet{Loup1993} showed that, for low mass loss rates there is a 

1509: simple relationship between the [25]-[12] color and mass-loss rates. 

1510: This breaks down at mass-loss rates \gapprox $10^{-5}$\,M$_\odot$yr$^{-1}$. 

1511: It has been suggested that such high mass-loss rate stars at CO-emission 

1512: deficient due to either saturation effects, low kinetic temperatures or 

1513: possibly dramatic recent increases in mass-loss. If we extrapolate the 

1514: trends from low mass-loss rates to determine mass-loss rates from the 

1515: [25]-[12] color we find that these objects should have a mass-loss rate 

1516: in excess of $10^{-4}$\,M$_\odot$yr$^{-1}$, consistent with the high 

1517: modeled optical depths. Because of this relationship, we sought 

1518: correlations between the various observed and modeled parameters and 

1519: the [25]-[12] and [60]-[25] colors. This search yielded only one correlation: 

1520: between the [60]-[25] colors and the dust mass-loss rate from previous models.

1521: It is possible that a better correlation may be found using the 

1522: {\em Manchester Method} i.e. the [6]-[9] color 

1523: \citep[e.g.][]{Sloan2006,zijl06}, however, we suspect that the lack of 

1524: correlation arises because of the intrinsic degeneracy in the modeling.

1525: 

1526: In \S~\ref{sicabs} and \S~\ref{obssect} 

1527: we discussed the possibility that the broadening of the 

1528: 10--13$\mu$m feature might be due to silicate dust.

1529: With this in mind, we re-measured the position  and strength of the 11$\mu$m 

1530: feature assuming that the short wavelength wing is due to silicate. 

1531: This involved

1532: measuring the feature-to-continuum strength at 9.7 and 11.3$\mu$m. These data

1533: are tabulated in Table~\ref{obsparam} and were also included in the 

1534: investigation of correlations between parameters.

1535: 

1536: One pair of parameters that yielded a correlation were

1537: the best-fit blackbody temperature and the model optical depth,

1538: which is demonstrated in Fig~\ref{correlfig}. 

1539: This correlation occurs whether we assume MRN or ``meteoritic'' grainsizes.

1540: %

1541: Since there is a relationship between the optical depth and shell thickness, 

1542: this correlation seems to support models validity.

1543: 

1544: In addition to this relationship between the modeled optical depth and the 

1545: best-fit blackbody temperature, we found two other parameters that the optical 

1546: depth correlates with: modeled percentage of SiC in the dust shell; and the 

1547: calculated timescale of the obscuring dust. In both cases, these correlations only hold for the generic (MRN) grain-sze distribution models.

1548: (shown in Fig.~\ref{correlfigtau}).

1549: The correlation between optical depth and percentage SiC is such that lower 

1550: optical depths 

1551: require more SiC. This in turn suggests that as mass-loss rates increase, the 

1552: SiC component dwindles. This can be interpreted in two ways:

1553: (1) at these high mass-loss rates SiC gets coated by carbon; and

1554: (2) increased mass-loss rates are associated with higher C/O ratios;

1555: carbon is enriched but not silicon, and thus more carbon 

1556: grains can be made, but not more SiC grains.

1557: %

1558: The first option seems unlikely in light of meteoritic evidence (\S~\ref{meteor}) and 

1559: theoretical condensation models (\S~\ref{condmod}).

1560: However, it is possible that there is a metallicity effect in play. 

1561: The presolar grains were formed prior to the formation of the solar system, 

1562: and thus their source stars may have had lower metallicities.

1563: \citet{Speck2006,leisenring} argued that coating of SiC grains is more 

1564: likely in lower metallicity environments, and thus it is difficult to see how 

1565: the higher metallicity objects we are now witnessing could have carbon-coated 

1566: SiC grains.

1567: However, \citet{leisenring} argued that SiC grains form the nucleation seed for MgS, 

1568: which may explain this correlation. 

1569: The second option may be logical if outflows are dust driven. 

1570: If more carbon grains can form, the radiation pressure driving the outflow is more effective.

1571: 

1572: 

1573: The second optical depth correlation is with the calculated timescale (also 

1574: shown in Fig.~\ref{correlfigtau} and again, only for the MRN grain-size 

1575: distribution). In this case, 

1576: the timescale for the dustshell increases with optical depth. 

1577: The oldest shells have the highest optical depth.

1578: %

1579: This may simply be due to the pathlength dependence of optical depth 

1580: (see Eqn.~\ref{tau0}), since $\tau$ depends on the geometric size of the dust 

1581: shell the oldest shells should have the largest optical depth. However, this 

1582: should also lead to a correlation between model dust shell thickness and 

1583: optical depth. No such correlation exists.

1584: %

1585: Another interpretation is that older stars have higher mass-loss rates, and 

1586: thus denser dust shells and higher optical depths. However, this assumes the 

1587: stars all had the same initial mass.

1588: %

1589: This apparent correlation is not strong, and does not holds for ``meteoritic'' 

1590: grains. Consequently, we will not attempt to place too much importance on it.

1591: 

1592: There is no correlation between the strength of the 9.7$\mu$m absorption and 

1593: that at 11.3$\mu$m, implying that the carriers of these features may not be 

1594: related. Furthermore, there is no correlation between the 9.7$\mu$m

1595: strength and distance to the object, which leads us to suggest that the cause 

1596: is not interstellar. Obviously there is a strong correlation between the 

1597: strength of the 9.7$\mu$m absorption and the barycentric position of the 

1598: feature, indicating that this is the cause of the broadening.

1599: 

1600: The most interesting correlations are between the ``isolated'' 

1601: 11.3$\mu$m feature (i.e. the residual feature after taking silicate absorption 

1602: into account) and the fitted blackbody temperature. 

1603: There is a correlation both between the SiC 

1604: barycentric position and the blackbody 

1605: temperature and between the  11.3$\mu$m feature-to-continuum ratio 

1606: (i.e. SiC feature strength) and the blackbody temperature (both shown in 

1607: Fig.~\ref{correlfig}). 

1608: 

1609: The position of the SiC absorption feature moves to shorter wavelengths for 

1610: higher 

1611: blackbody temperatures while also becoming weaker.

1612: This can be understood as being the result of more absorption occuring when 

1613: the surface we see is at a lower temperature, 

1614: which is associated with longer wavelength absorption.

1615: 

1616: We can interpret this result in terms of the self-absorption scheme described 

1617: by \citet{Speck2005}. The shifts in wavelength were attributed to a 

1618: changes in grain size. However, the weaker absorption and warmer blackbody 

1619: temperatures associated with the shortest wavelength peaks is the opposite 

1620: trend to that described by \citet{Speck2005}. Following their scheme, this 

1621: would imply that the weakest absorption is associated with the largest SiC 

1622: grain. The problem with interpreting this observation is that we do not know 

1623: the C/O ratios for these stars. While C/O does not affect the SiC condensation 

1624: temperature, it will affect the graphite condensation temperature. As seen in 

1625: Fig~\ref{modelfit23166}, we can reduce the inner dust temperature and 

1626: compensate by decreasing both the optical depth and the shell thickness.

1627: This would also change the depth into the shell at which the SiC forms.

1628: It is also possible that the shift in position is due to 

1629: incomplete subtraction of the silicate contribution.

1630: Further studies are ongoing, but are beyond the scope of the 

1631: present work.

1632: 

1633: 

1634: 

1635: The assumption that there is a silicate contribution to the spectrum raises 

1636: several questions. Identifying the source of the 9.7$\mu$m contribution to the

1637: absorption feature remains beyond the scope of the present paper, but, 

1638: needless to say, if it is silicate, the specter of dual chemistry would 

1639: require seriously rethinking our current models of dust formation around 

1640: carbon stars. \citet{lf99} suggested that silicates could form around carbon 

1641: stars, and indeed silicate carbon stars exist.

1642: It is possible that the feature may be attributable to some form 

1643: of hydrogenated amorphous carbon, which has been postulated as the source of 

1644: an $\sim9\mu$m emission feature in optical thin carbon star spectra 

1645: \citep[see][and reference therein]{Thompson2006}.

1646: 

1647: No additional correlations were found. The lack of correlation between most of 

1648: the observable or modeled parameters echoes 

1649: the results of \citet{Thompson2006}, who found that there are no trends in the 

1650: parameters associated with 11$\mu$m emission feature in visibly observable 

1651: carbon stars. This suggests that even amongst the extreme carbon stars, 

1652: variations in C/O, $s$-process and nitrogen enhancements make discernment of 

1653: the physical properties associated with the 11$\mu$m feature very difficult.

1654: 

1655: 

1656: \section{Conclusions}

1657: 

1658: 

1659: We have presented three previously unrecognized SiC absorption features in the 

1660: spectra of extreme carbon stars. Together with the seven known SiC absorption 

1661: stars, this brings the total of known extreme carbon stars with SiC absorption 

1662: features to ten.

1663: 

1664: Previous radiative transfer models of extreme carbon stars utilized relatively 

1665: low condensation temperatures. Here, theoretical condensation models have 

1666: been used to justify much higher condensation temperatures. 

1667: In addition, our models use graphite instead of amorphous carbon, because 

1668: of the preferential formation of graphite at higher temperature and the 

1669: meteoritic evidence. Both the higher condensation temperature (through a 

1670: decrease in the inner radius of the dust shell) and to a smaller extent the 

1671: use of graphite will greatly increase the acceleration felt by the dust grains 

1672: in the shell relative to parameters used in previous research. 

1673: %

1674: 

1675: We have shown that grain-size issues cannot be ignored in the production 

1676: of models that accurately fit the observed spectra of extreme carbon stars.  

1677: The size distribution that is needed is not clearly defined because of the 

1678: inherent degeneracy in radiative transfer modeling. Meteoritic grain-size 

1679: distributions are as valid as other size distributions with the advantage of 

1680: being model independent. However, they may underestimate the contribution from 

1681: small grains.

1682: 

1683: With the exception of IRAS 17534$-$3030, all sample stars could be modeled 

1684: with either the generic (MRN) or ``meteoritic grain-size distribution

1685: %

1686: IRAS 17534$-$3030's narrow SED required the use of the ``meteoric'' 

1687: distribution. Furthermore, there is no evidence for any molecular absorption 

1688: in its spectrum. Because the 11$\mu$m 

1689: feature is still present in the absence of the other molecular features, it 

1690: supports the attribution of this feature to a solid state carrier.

1691: 	

1692: The various parameters compiled in the course of this research (both 

1693: through radiative transfer modeling and from observations) have been compared 

1694: in order to identify any correlations, with the result that the cause of 

1695: differences is the spectra cannot be attributed to mass-loss rate or gas

1696: pressure in the dust condensation zone.

1697: In fact the paucity of correlations between parameters echoes 

1698: the results of \citet{Thompson2006} and 

1699: suggests that even amongst the extreme carbon stars, 

1700: variations in C/O, $s$-process and nitrogen enhancements make 11$\mu$m 

1701: a poor probe of the details of dust shell parameters.

1702: 

1703: The timescales associated with the heavy mass-loss experienced by these 

1704: extreme carbon stars are very short (tens to hundred of years) and are not

1705: consistent with timescales for the superwind. This indicates that the heavy 

1706: mass-loss phase of carbon stars is not a direct result of thermal pulse 

1707: (although thermal pulses may be the root cause).

1708: 

1709: 

1710: 

1711: \acknowledgments

1712: 

1713: This work is supported by NSF AST-0607341.

1714: We are very grateful to the referee, Albert Zijlstra, 

1715: for his comments which have significantly improved this paper.

1716: Kevin Volk is also thanked for helpful advice on this paper.

1717: 

1718: %% The reference list follows the main body and any appendices.

1719: %% Use LaTeX's thebibliography environment to mark up your reference list.

1720: %% Note \begin{thebibliography} is followed by an empty set of

1721: %% curly braces.  If you forget this, LaTeX will generate the error

1722: %% "Perhaps a missing \item?".

1723: %%

1724: %% thebibliography produces citations in the text using \bibitem-\cite

1725: %% cross-referencing. Each reference is preceded by a

1726: %% \bibitem command that defines in curly braces the KEY that corresponds

1727: %% to the KEY in the \cite commands (see the first section above).

1728: %% Make sure that you provide a unique KEY for every \bibitem or else the

1729: %% paper will not LaTeX. The square brackets should contain

1730: %% the citation text that LaTeX will insert in

1731: %% place of the \cite commands.

1732: 

1733: %% We have used macros to produce journal name abbreviations.

1734: %% AASTeX provides a number of these for the more frequently-cited journals.

1735: %% See the Author Guide for a list of them.

1736: 

1737: %% Note that the style of the \bibitem labels (in []) is slightly

1738: %% different from previous examples.  The natbib system solves a host

1739: %% of citation expression problems, but it is necessary to clearly

1740: %% delimit the year from the author name used in the citation.

1741: %% See the natbib documentation for more details and options.

1742: 

1743: \begin{thebibliography}{}

1744: 

1745: \bibitem[Amari et al.(2002)]{amari02} Amari S.,

1746: Jennings C., Nguyen A., Stadermann F.J., Zinner E., Lewis R.S.,

1747: 2002, Lunar Planet. Sci. Conf., 33, 120

1748: \bibitem[Bagnulo, Doyle \& Griffin(1995)]{bdg95}Bagnulo, Doyle \& Griffin, 

1749: 1995, \aap, 301, 501.

1750: \bibitem[Baron et al.(1987)]{Baron1987} Baron, Y., de Muizon, M., Papoular, R., \& P\'{e}gouri\'{e}, B. 1987 \aap , 186, 271.

1751: \bibitem[Begemann et al.(1994)]{begemann94} Begemann, B., et al., 1994, \apj, 423, L71.

1752: \bibitem[Bernatowicz et al.(1987)]{Bernatowicz1987} Bernatowicz, T., 

1753: Fraundorf, G., Ming, T., Anders, E., Wopenka, B., Zinner, E., Fraundorf, P.,

1754: 1987, Nature, 330, 728.

1755: \bibitem[Bernatowicz et al.(2006)]{bern05} Bernatowicz et al. 2006, Meteorites 

1756: and the Early Solar System II, D. S. Lauretta and H. Y. McSween Jr. (eds.), 

1757: University of Arizona Press, Tucson, 943 pp., p.109-126.

1758: \bibitem[Chan \& Kwok(1990)]{Chan1990} Chan, S. J., \& Kwok, S. 1990, \aap , 237, 354.

1759: \bibitem[Cherchneff(2006)]{cherchneff06} Cherchneff, I., 2006,\aap, 456, 1001.

1760: \bibitem[Chigai \& Yamamoto(2003)]{chigai} Chigai, T.,  Yamamoto, T. 2003, 

1761: Geochim. Cosmochim. Acta., 67, 64.

1762: 

1763: \bibitem[Clayton \& Nittler(2004)]{cn04}Clayton, D.D., Nittler, L.R., 2004,

1764: ARA\&A, 42, 39.

1765: \bibitem[Cl\'{e}ment et al.(2003)]{Clement2003}Cl\'{e}ment, D., Mutschke, H., Klein, R., \& Henning, T. 2003, \apj, 594, 642

1766: \bibitem[Cl\'{e}ment et al.(2005)]{Clement2005}Cl\'{e}ment, D., Mutschke, H., Klein, R., \& Henning, T. 2005, \apj, 621, 985.

1767: \bibitem[Comelli et al.(1988)]{comelli} Comelli, G. Stohr, J., Robinson, C.J.,  Jark, W. (1988) Phys. Rev. B 38, 7511.

1768: \bibitem[Corman et al.(2008)]{corman08} Corman, A.B., Speck, A.K., Volk, K., Sloan, G.C., Barlow, M.J., 2008, in prep.

1769: \bibitem[Daulton et al.(2003)]{daulton03} Daulton et al. 2003, GeCoA, 67, 4743.

1770: \bibitem[DePew et al.(2006)]{DePew2006} DePew, K., Speck, A.K., Dijkstra, C., 

1771: 2006, \apj, 640, 971.

1772: \bibitem[de Graauw et al.(1996)]{degraauw96}de Graauw, T. et al.  1996, \aap, 

1773: 315, 49.

1774: \bibitem[Dijkstra et al.(2005)]{dijk05} Dijkstra, C., Speck, A. K., Reid, R.B., Abraham, P., 2005, \apjl, 633, L133.

1775: \bibitem[Dijkstra \& Speck(2006)]{ds06} Dijkstra, C., Speck, A. K., 2006, \apj

1776: 651, 288.

1777: \bibitem[Dominik, Sedlmayr, \& Gail(1989)]{dsg89} Dominik, Sedlmayr, \& Gail, 

1778: 1989, \aap, 223, 227.

1779: \bibitem[Draine \& Lee(1984)]{DL84} Draine, B.T., Lee, H.M. 1984, \apj, 285, 89.

1780: \bibitem[Fong, Meixner \& Shah(2003)]{fong03} Fong, D., Meixner, M., Shah, R., 2003, \apj, 582, L39.

1781: \bibitem[Forrest et al.(1981)]{forrest81} Forrest, W. J.; Houck, J. R.; McCarthy, J. F., 1981, \apj, 248, 195.

1782: \bibitem[Friedman(1969)]{Friedman1969} Friedman, C. 1969, Physica 41, 189.

1783: \bibitem[Gilman(1969)]{Gilman1969} Gilman, R. C. 1969, \apj , 155, L185.

1784: \bibitem[Gilra \& Code(1971)]{Gilra1971} Gilra, D. P., \& Code, A. D. 1971, \baas , 3, 379.

1785: \bibitem[Goebel et al.(1995)]{Goebel1995} Goebel, J. H., Cheeseman, P., \& Gerbault, F. 1995, \apj , 449, 246. 

1786: \bibitem[Goebel \& Moseley(1985)]{gm85} Goebel, J. H., Moseley, S.H., 1985, \apj, 290, L35.

1787: \bibitem[Griffin(1990)]{griffin1990} Griffin, I.P., 1990 \mnras, 247, 591.

1788: \bibitem[Groenewegen(1994)]{Groenewegen1994} Groenewegen, M.A.T. 1994, \aap, 

1789: 290, 207.

1790: \bibitem[Groenewegen(1995)]{Groenewegen1995} Groenewegen, M.A.T. 1995, \aap , 293, 463. 

1791: \bibitem[Groenewegen et al.(1996)]{Groenewegen1996} Groenewegen, M.A.T. 1996, \aap, 305, 475.

1792: \bibitem[Groenewegen(1997)]{Groenewegen1997} Groenewegen, M.A.T. 1997, \apss,

1793: 251, 89.

1794: \bibitem[Groenewegen et al.(1998)]{Groenewegen1998} Groenewegen, M.A.T. et al. 1998, \mnras, 293, 18.

1795: \bibitem[Groenewegen et al.(1999)]{Groen1999} Groenewegen, M.A.T. et al. 1999, \aaps, 140, 197.

1796: \bibitem[Groenewegen et al.(2002)]{Groen2002} Groenewegen, M. A. T., Sevenster, M., Spoon, H. W. W., P\'{e}rez, I., 2002, \aap, 390, 511.

1797: \bibitem[Hackwell(1972)]{Hackwell1972} Hackwell, J. A. 1972, \aap , 21, 239.

1798: \bibitem[Hanner(1998)]{Hanner1998} Hanner, M. S. 1988, Infrared Observations of Comets Halley and Wilson and Properties of the Grains, ed. M. S. Hanner (NASA C-3004; Washington: NASA), 22.

1799: \bibitem[Hildebrand(1983)]{hilde83} Hildebrand, R.H., 1983, QJRAS, 24, 267.

1800: \bibitem[Hony et al.(2003)]{hony2003} Hony et al. 2003 \aap, 402, 211.

1801: \bibitem[Hony et al.(2002)]{hony2002} Hony, S., Waters, L.B.F.M., Tielens, A.G.G.M, 2002, \aap, 390, 533

1802: \bibitem[Huggins et al.(1988)]{hug88} Huggins, P.J., Olofsson, H., Johansson, L.E.B., 1988, \apj, 332, 1009.

1803: \bibitem[Iben \& Renzini(1983)]{Iben1983} Iben, I., Jr. \& Renzini, A. 1983, \araa , 21, 271.

1804: \bibitem[Ivezic \& Elitzur(1995)]{Ivezic1995} Ivezic, Z. \& Elitzur, M. 1995, \apj , 445, 415.

1805: \bibitem[Ivezic \& Elitzur(1997)]{ie97} Ivezic, Z. \& Elitzur, M. 1997, \mnras, 287, 799.

1806: \bibitem[Jennings et al.(2002)]{jennings} Jennings, C. L., Savina, M. R., 

1807: Messenger, S., Amari, S., Nichols, R. H., Pellin, M. J., Podosek, F. A. 2002, 

1808: LPSC, 33, A1833

1809: \bibitem[Jones et al.(1978)]{Jones1978} Jones, B., Merrill, K. M., Puetter, R. C., \& Willner, S. P. 1978, \aj, 83, 1437

1810: \bibitem[J{\o}rgensen et al.(2000)]{jorg}J{\o}rgensen, U. G., Hron, J., 

1811: Loidl, R., 2000, \aap, 356, 253.

1812: \bibitem[Jura et al.(2000)]{jura2000} Jura, M., Turner, J. L., Van Dyk, S., Knapp, G. R., 2000, \apjl 528, L105.

1813: \bibitem[Jura(1994)]{jura94} Jura, M. 1994, \apj, 434, 713.

1814: \bibitem[Kelires(1993)]{kelires} Kelires, P.C., 1993, Phys. Rev. B 47, 1829.

1815: \bibitem[Kessler et al.(1996)]{kessler} Kessler, M.F. et al., 1996, \aap, 315, L27 

1816: \bibitem[Kim, Martin \& Hendry(1994)]{kmh} Kim, S.-H., Martin, P. G., 

1817: Hendry, P.D., 1994, \apj, 422, 164.

1818: \bibitem[Knapp \& Morris(1985)]{km85} Knapp, G.R., Morris, M., 1982, \apj, 292, 640.

1819: \bibitem[Lagadec et al.(2007)]{Lagadec2007} Lagadec, E. et al., 2007, \mnras, 376, 1270.

1820: \bibitem[Lagadec et al.(2008)]{Lagadec2008} Lagadec, E. et al., 2008, \mnras, 

1821: 383, 399.

1822: \bibitem[Lagadec \& Zijlstra(2008)]{LZ08} Lagadec, E., Zijlstra, A.A. , 2008, \mnras, in press.

1823: \bibitem[Lambert et al.(1986)]{lamb86} Lambert, D.L., Gustafsson, B., Eriksson, K., Hinkle, K.H., 1986, \apjs, 62, 373.

1824: \bibitem[Leech et al.(2003)]{Leech2003} Leech, K., et al. 2003, The ISO Handbook, Vol. V. SWS - The Short Wavelength Spectrometer Version 2.0.1 (June, 2003). Series edited by T.G. Mueller, J.A.D.L. Blommaert, and P. Garcia-Lario. ESA SP-1262, (Noordwijk: ESA).

1825: \bibitem[Leisenring et al.(2008)]{leisenring} Leisenring, J. M., Kemper, F., Sloan, G. C., 2008, \apj, 681, 1557.

1826: %\bibitem[Lodders \& Amari(2005)]{lod05} Lodders \& Amari (2005)

1827: \bibitem[Lodders \& Fegley(1995)]{lf95} Lodders, K., Fegley, B., 1995, Meteoritics, 30, 661.

1828: \bibitem[Lodders \& Fegley(1999)]{lf99} Lodders, K., Fegley, B., 1999, Asymptotic Giant Branch Stars, IAU Symposium 191, Edited by T. Le Bertre, A. Lebre, and C. Waelkens. p. 279

1829: \bibitem[Loup et al.(1993)]{Loup1993} Loup, C., Forveille, T., Omont, A., 

1830: Paul, J. F., 1993, \aaps, 99, 291.

1831: \bibitem[Mathis et. al(1977)]{Mathis1977} Mathis, J. S., Rumpl, W., \& Nordsieck, K. H. 1977, \apj 217, 425.

1832: \bibitem[Martin \& Rogers(1987)]{mr87} Martin, P.G., Rogers, C., 1987, \apj, 

1833: 322, 374.

1834: \bibitem[Meixner et al.(2004)]{meixner2004} Meixner, M., Zalucha, A., Ueta, T., Fong, D., Justtanont, K.,

1835: 2004, \apj, 614, 371.

1836: \bibitem[Meixner et al.(2002)]{meixner2002} Meixner, M., Ueta, T., 

1837: Bobrowsky, M., Speck, A., 2002, \apj, 571, 936.

1838: \bibitem[Meixner et al.(1999)]{meixner1999} 1999, \apjs, 122, 221.

1839: \bibitem[Nenkova et al.(2000)]{Nenkova2000} Nenkova, M., Ivezic, Z., \& Elitzur, M. 2000 in ASP Conf. Ser. 196, Thermal Emission Spectroscopy and Analysis of Dust, Disks, and Regoliths, ed. M. L. Sitko, A.L. Sprague \& D. K. Lynch (San Fransicso: ASP), 77.

1840: \bibitem[Neugebauer et al.(1984)]{neug84}Neugebauer, G. et al. 1984, Science, 

1841: 224, 1.

1842: \bibitem[Nuth et al.(1985)]{nuth85} Nuth,J.A., et al.\ 1985, \apj, 290, L41.

1843: \bibitem[Olofsson et al.(1993a)]{olof93a} Olofsson, H., Eriksson, K., Gustafsson, B., Carlstroem, U. 1993a, \apjs, 87, 267.

1844: \bibitem[Olofsson et al.(1993b)]{olof93b} Olofsson, H., Eriksson, K., Gustafsson, B., Carlstroem, U. 1993b, \apjs, 87, 305.

1845: \bibitem[Omont et al.(1993)]{omont93} Omont et al. 1993, \aap, 267, 515.

1846: \bibitem[Omont et al.(1995)]{omont95} Omont, A. et al., 1995, \apj, 454, 819

1847: \bibitem[P\'{e}gouri\'{e}(1988)]{Pegourie1988} P\'{e}gouri\'{e}, B. 1988, \aap , 194, 335.

1848: \bibitem[Pitman, Speck \& Hofmeister(2006)]{Pitman2006} Pitman, K.M., Speck, 

1849: A.K., Hofmeister, A.M., 2006, \mnras, 371, 1744.

1850: \bibitem[Pitman et al.(2008)]{Pitman2007} Pitman, K.M., Speck, 

1851: A.K., Hofmeister, A.M., 2008, \aap, 483, 661.

1852: \bibitem[Prombo et al.(1993)]{prombo}Prombo, C.A., Podosek, F.A., Amari, S., 

1853: Lewis, R. S. 1993, ApJ, 410, 393.

1854: \bibitem[Renzini(1981)]{renz81} Renzini, A, 1981, in Physical Processes in Red Giants, eds. I. Iben Jr. \& A. Renzini, Reidel:Dordrecht, pp431.

1855: \bibitem[Rowan-Robinson \& Harris(1983)]{rrh83} Roawn-Robinson, M., Harris, S., 1983, \mnras, 202, 797.

1856: \bibitem[Rouleau \& Martin(1991)]{rm91} Rouleau, F., Martin, P.G., 1991, \apj, 377, 526.

1857: \bibitem[Saada(2000)]{saada} Saada, D., 2000, Ph.D. thesis

1858: \bibitem[Sahai(2004)]{sahai} Sahai, R., 2004, in ASP Conf. Ser. 313, Asymmetrical Planetary Nebulae III: Winds, Structure and the Thunderbird, ed. M. Meixner et al. (San Francisco: ASP) 141.

1859: \bibitem[Sharp \& Wasserburg(1995)]{sw95} Sharp, C.M., Wasserburg, G.J., 1995, GeCoA, 59, 1445.

1860: \bibitem[Skinner et al.(1997)]{skinner97} Skinner C.J., et al., 1997, \aap, 328, 290.

1861: \bibitem[Skinner et al.(1998)]{skinner98} Skinner, C.J., Meixner, M., Bobrowsky, M., 1998, \mnras, 300, L29.

1862: \bibitem[Skrutskie et al.(2001)]{skrutskie} Skrutskie, M. F., Reber, T. J., Murphy, N. W., Weinberg, M. D., 2001, \baas, 33, 1437.

1863: \bibitem[Sloan et al.(2006)]{Sloan2006} Sloan, G.C., et al. 2006, \apj, 645, 1118.

1864: \bibitem[Sloan et al.(1998)]{Sloan1998} Sloan, G. C., Little-Marenin, I. R., \& Price, S. D. 1998, \aj , 115, 809.

1865: \bibitem[Soker \& Subag(2005)]{soker} Soker, N., Subag, E. 2005, \aj, 130, 2717.

1866: \bibitem[Speck et al.(1997)]{Speck1997} Speck, A. K., Barlow, M. J., \& Skinner, C. J. 1997, \mnras , 288, 431.

1867: \bibitem[Speck et al.(1999)]{Speck1999} Speck, A. K., Hofmeister, A.M., \& Barlow, M. J., 1999, \apjl, 513, L87.

1868: \bibitem[Speck et al.(2000)]{Speck2000} Speck, A. K., Barlow, M. J., Sylvester, R.J \& Hofmeister A.M., 2000, \aaps, 146, 437.

1869: \bibitem[Speck et al.(2005)]{Speck2005} Speck, A. K., Thompson, G. D., \& Hofmeister, A. M. 2005, \apj , 634, 426.

1870: \bibitem[Speck et al.(2006)]{Speck2006} Speck, A. K., Cami, J., Markwick-Kemper, C., Leisenring, J., Szczerba, R., Dijkstra, C., Van Dyk, S. \& Meixner, M. 2006, \apj , 650, 892. 

1871: \bibitem[Szczerba et al.(1997)]{szc97} Szczerba, R., Omont, A., Volk, K.,

1872: Cox, P., Kwok, S., 1997, \aap, 317, 859.

1873: \bibitem[Thompson et al.(2006)]{Thompson2006} Thompson, G.D., Corman, A.B., Speck, A.K., Dijkstra, ., 2006, \apj, 652, 1654.

1874: \bibitem[Treffers \& Cohen(1974)]{Treffers1974} Treffers, R., \& Cohen, M. 1974, \apj , 188, 545.

1875: \bibitem[van der Veen \& Habing(1988)]{vh88} van der Veen, W. E. C. J., Habing, H. J., 1988, \aap, 194, 125.

1876: \bibitem[Vassiliadis \& Wood(1993)]{vw93} Vassiliadis, E., Wood, P.R., (VW) 1993, \apj, 413, 641.

1877: \bibitem[Villaver et al.(2002a)]{vil02a}  Villaver, E., Manchado, A., Garcia-Segura, G., 2002a, ApJ, 581, 1204.

1878: \bibitem[Villaver et al.(2002b)]{vil02b}  Villaver, E., Manchado, A., Garcia-Segura, G., 2002b, ApJ, 571, 880.

1879: \bibitem[Virag et al.(1992)]{virag}Virag, A.  et al., 1992 Geochim.\ Cosmochim.\ Acta, 56, 1715.

1880: \bibitem[Volk et al.(1992)]{Volk1992} Volk, K., Kwok, S., \& Langill, P. 1992, \apj , 391, 285.

1881: \bibitem[Volk et al.(2000)]{Volk2000} Volk, K., Xiong, G., \& Kwok, S. 2000, \apj , 530, 408.

1882: \bibitem[Vassiliadis \& Wood(1993)]{Vassiliadis1993} Vassiliadis, E. \& Wood, P. R. 1993, \apj , 413, 641.

1883: \bibitem[Waelkens \& Waters(1999)]{waelkens} Waelkens, C.,  Waters, L. B. F. M. 1999, in IAU Symp. 191, Asymptotic Giant Branch Stars, ed. T. le Bertre, A. Lèbre \& C. Waelkens (San Francisco: ASP), 519.

1884: \bibitem[Woitke(2006)]{woitke06} Woitke, P., 2006, \aap, 452, 537.

1885: \bibitem[Wood et al.(1992)]{wood92} Wood, P.R., et al.\, 1992, \apj, 397, 552.

1886: \bibitem[Zijlstra et al.(2006)]{zijl06} Zijlstra, A., et al. , 2006, \mnras, 370, 1961.

1887: 

1888: 

1889: \end{thebibliography}

1890: 

1891: \clearpage

1892: %Fig1

1893: \begin{figure}[t]

1894: %\resizebox{\hsize}{!}{%

1895: \includegraphics[angle=270,scale=0.75]{f1.eps}%}

1896: \caption{Regions in the IRAS color-color diagram populated by extreme carbon stars. see text for details. 

1897: The black body curve is indicated by the solid grey line.

1898: The modified blackbody [$B(T,\lambda)*\lambda^{-1}$] is indicated by the dashed grey line}

1899: \label{irascolor} 

1900: \end{figure}

1901: 

1902: 

1903: \clearpage

1904: %Fig2

1905: \begin{figure}[t]

1906: %\resizebox{\hsize}{!}{%

1907: \includegraphics[angle=270,scale=0.85]{f2.eps}

1908: \caption{{\it ISO\/} SWS spectra of ten extreme C-stars.

1909: {\it Left Panel\/}: Flux-calibrated spectra and fitted blackbody continua.

1910: Solid line = {\it ISO\/} spectrum;

1911: dashed line = best-fitting blackbody continuum;

1912: The $y$-axis is the flux (F$_\lambda$) in W\,m$^{-2}$\,$\mu$m$^{-1}$;

1913: the $x$-axis is wavelength in $\mu$m.

1914: {\it Right Panel\/}: Continuum-divided spectra.

1915: %Solid line = {\it ISO\/} spectra; dashed line = ground-based spectra from^M

1916: %Speck et al. (1997). 

1917: Table~\ref{obsparam} lists the blackbody temperatures used in each case to 

1918: produce the continuum and the continuum-divided spectra.}

1919: \label{obsfig} 

1920: \end{figure}

1921: 

1922: 

1923: \clearpage

1924: %Fig3

1925: \begin{figure}[t]

1926: %\resizebox{\hsize}{!}{%

1927: \includegraphics[angle=270,scale=0.7]{f3.eps}

1928: \caption{{\it ISO\/} SWS spectra of sample extreme C-stars together with 

1929: {\it IRAS\/} 12, 25, 60 and 100$\mu$m photometry points and {\it IRAS\/} LRS 

1930: spectra.

1931: Solid line = {\it IRAS\/} spectrum;

1932: points =  {\it IRAS\/} photometry points;

1933: dashed line = {\it ISO\/} SWS spectrum.

1934: The $y$-axis is the flux ($\lambda$F$_\lambda$) in W\,m$^{-2}$;

1935: the $x$-axis is wavelength in $\mu$m.}

1936: \label{obsfig2} 

1937: \end{figure}

1938: 

1939: \clearpage

1940: %Fig4

1941: \begin{figure}[t]

1942: %\resizebox{\hsize}{!}{%

1943: %\addtocounter{figure}{-1}

1944: \includegraphics[angle=270,scale=0.7]{f4.eps}

1945: \caption{{\it ISO\/} SWS spectra of sample extreme C-stars together with 

1946: {\it IRAS\/} 12, 25, 60 and 100$\mu$m photometry points and {\it IRAS\/} LRS 

1947: spectra (part 2).

1948: Solid line = {\it IRAS\/} spectrum;

1949: points =  {\it IRAS\/} photometry points;

1950: dashed line = {\it ISO\/} SWS spectrum.

1951: The $y$-axis is the flux ($\lambda$F$_\lambda$) in W\,m$^{-2}$;

1952: the $x$-axis is wavelength in $\mu$m.}

1953: \label{obsfig3} 

1954: \end{figure}

1955: 

1956: \clearpage

1957: %Fig5

1958: \begin{figure}[t]

1959: %\resizebox{\hsize}{!}{%

1960: %\addtocounter{figure}{-1}

1961: \includegraphics[angle=270,scale=0.7]{f5.eps}

1962: \caption{{\it ISO\/} SWS spectra of sample extreme C-stars together with 

1963: {\it IRAS\/} 12, 25, 60 and 100$\mu$m photometry points and {\it IRAS\/} LRS 

1964: spectra (part 3).

1965: Solid line = {\it IRAS\/} spectrum;

1966: points =  {\it IRAS\/} photometry points;

1967: dashed line = {\it ISO\/} SWS spectrum.

1968: The $y$-axis is the flux ($\lambda$F$_\lambda$) in W\,m$^{-2}$;

1969: the $x$-axis is wavelength in $\mu$m.}

1970: \label{obsfig4} 

1971: \end{figure}

1972: 

1973: 

1974: \clearpage

1975: %Fig6

1976: \begin{figure}[t]

1977: %\resizebox{\hsize}{!}{%

1978: \includegraphics[angle=270,scale=0.75]{f6.eps}

1979: \caption{\label{gsizefig} The effect of grain-size distributions on the model spectral energy distribution. In all cases the model parameters are identical except for the grain-size distribution. see \S~\ref{gszmodel} for detailed description of the grain-size distributions. The grey line shows the ISO-SWS spectrum of IRAS\,03313+6058 for comparison.}

1980: \end{figure}

1981: 

1982: \clearpage

1983: %Fig7

1984: \begin{figure}[t]

1985: %\resizebox{\hsize}{!}{%

1986: %

1987: ~~~~~~~\,\includegraphics[angle=270,scale=0.6]{f7a.eps}\\

1988: %\vspace{-25mm}

1989: %\hspace{25mm}

1990: \includegraphics[angle=270,scale=0.6]{f7b.eps}

1991: \caption{\label{lfptspace} Pressure-Temperature space for the dust condensation region around extreme carbon stars (assumes solar metallicity). Grey lines indicate the condensation temperature for a given pressure as calculated by \citet{lf95}. Black lines indicate the P--T paths for the otuflowing gas from our sample stars; ({\it top}) as calculated from the published CO mass-loss rates and expansion velocities (see \S~\ref{ptcalc} for details); and ({\it bottom}) as calculated from the modeled dust temperature radial profile. }

1992: 

1993: %The best model inner dust temperature is indicated by a triangle on each P--T track.}

1994: \end{figure}

1995: 

1996: \clearpage

1997: %Fig8

1998: \begin{figure}[t]

1999: %\resizebox{\hsize}{!}{%

2000: \includegraphics[angle=270,scale=0.7]{f8.eps}

2001: \caption{\label{modelfit1} Best fit models (part 1).

2002: solid line = ISO-SWS spectrum;

2003: dashed line = best fit model using MRN grain-size distribution;

2004: dotted line =  best fit model using ``meteoritic'' grain-size distribution;

2005: $X$-axis is wavelength ($\mu$m);

2006: $y$-axis is flux ($\lambda F_\lambda$) in W\,m$^{-2}$.

2007: In all cases, $T_\star$=3000\,K and $T_{\rm inner}$=1800\,K. 

2008: Model parameters are listed in Table~\ref{tabfit1}}

2009: \end{figure}

2010: 

2011: \clearpage

2012: \begin{figure}[t]

2013: %\addtocounter{figure}{-1}

2014: %\resizebox{\hsize}{!}{%

2015: \includegraphics[angle=270,scale=0.7]{f9.eps}

2016: \caption{\label{modelfit2} Best fit models (part 2) 

2017: solid line = ISO-SWS spectrum;

2018: dashed line = best fit model

2019: $x$-axis is wavelength ($\mu$m);

2020: $y$-axis is flux ($\lambda F_\lambda$) in W\,m$^{-2}$.

2021: In all cases, $T_\star$=3000\,K and $T_{\rm inner}$=1800\,K. 

2022: Model parameters are listed in Table~\ref{tabfit1}}

2023: \end{figure}

2024: 

2025: \clearpage

2026: \begin{figure}[t]

2027: %\addtocounter{figure}{-1}

2028: %\resizebox{\hsize}{!}{%

2029: \includegraphics[angle=270,scale=0.7]{f10.eps}

2030: \caption{\label{modelfit3} Best fit models (part 3) 

2031: solid line = ISO-SWS spectrum;

2032: dashed line = best fit model

2033: $x$-axis is wavelength ($\mu$m);

2034: $y$-axis is flux ($\lambda F_\lambda$) in W\,m$^{-2}$.

2035: In all cases, $T_\star$=3000\,K and $T_{\rm inner}$=1800\,K. 

2036: other parameters are indicated in the legend. 

2037: Parameters of these models are also listed in Table~\ref{tabfit1}}

2038: \end{figure}

2039: 

2040: \clearpage

2041: \begin{figure}[t]

2042: %\resizebox{\hsize}{!}{%

2043: \includegraphics[angle=270,scale=0.7]{f11.eps}

2044: \caption{\label{modelfit23166} Degeneracy in the best fit models for 

2045: IRAS\,23166+1655.  

2046: solid line = ISO-SWS spectrum;

2047: dashed line = best fit model with $T_{\rm inner}$=1800\,K

2048: dotted line = best fit model with $T_{\rm inner}$=1600\,K

2049: $X$-axis is wavelength ($\mu$m);

2050: $y$-axis is flux ($\lambda F_\lambda$) in W\,m$^{-2}$.

2051: In both cases, $T_\star$=3000\,K 

2052: other parameters are indicated in the legend. }

2053: \end{figure}

2054: 

2055: \clearpage

2056: \begin{figure}[t]

2057: %\resizebox{\hsize}{!}{%

2058: \includegraphics[angle=270,scale=1.0]{f12.eps}

2059: \caption{\label{correlfig} Correlations between the best-fit 

2060: blackbody temperature and 

2061: ({\it top}) modeled optical depth;

2062: ({\it middle}) 11.3$\mu$m feature-to-continuum ratio; and

2063: ({\it bottom}) SiC feature position.

2064: Solid lines represent the linear regression fit through the points.

2065: In the top figure, the dotted line represents the linear regression for model fits with MRN grainsize distributions.

2066: Grey points are for meteoritic grainsize models, and grey dashed line is the linear regression fit through the points.}

2067: \end{figure}

2068: 

2069: \clearpage

2070: \begin{figure}[t]

2071: %\resizebox{\hsize}{!}{%

2072: \includegraphics[angle=270,scale=1.0]{f13.eps}

2073: \caption{\label{correlfigtau} Correlations between the modeled optical depth

2074: and 

2075: ({\it top}) the modeled \%age SiC

2076: ({\it bottom}) the calculated superwind timescale from Table~\ref{timescaletab}

2077: Solid lines represent the linear regression fit through the points.}

2078: \end{figure}

2079: 

2080: \clearpage

2081: 

2082: %table1

2083: \begin{table}

2084: \small

2085: \caption{Previous Models of Extreme Carbon stars with 11$\mu$m absorption features \label{prevmod}}

2086: \begin{tabular}{l@{\hspace{1mm}}c@{\hspace{2mm}}c@{\hspace{2mm}}c@{\hspace{1mm}}c@{\hspace{1mm}}c@{\hspace{1mm}}c@{\hspace{1mm}}c@{\hspace{1mm}}c}

2087: \hline

2088: Star& composition$^\dag$& Grain-size &$\tau_{11.3\mu \rm m}$ &	drop-off & R$_{in}^\ast$&  T$_{in}$	& $\dot{M}^\ast$        &	REF\\	

2089: &   (SiC\%)  &  ($\mu$m)             &	                     &	       &($\times 10^{-4}$\,pc)    &	(K)	& M$_\odot$\,yr$^{-1}$	&  \\

2090: \hline

2091: 00210+6221 &	0	& 0.1	     & 	4.5	             & 3.00	 & 0.362/D&	1000	& 12$\times 10^{-5}/D$	&	2	\\	

2092: 01144+6658 &	8	& 0.1	     &	4.84	             &	         &	  &	1000	& 9.50 $\times 10^{-7}$	&	4	\\	

2093: 06582-1507 &	0	& ...        &	5.15	             &	         &	  &	1000	& 1.08$\times 10^{-4}$	&	1	\\	

2094: 	   &	0	& 0.1	     &	2.1	             & 2.25      & 1.23   & 1000	& 2.59$\times 10^{-4}$	&	2	\\	

2095: 17534$-$3030 &	0	& 0.1	     &	4.4	             & 2.50      & 1.39   & 1000	& 8.20$\times 10^{-4}$	&	2	\\	

2096: 19548+3035 &	0	& 0.1	     &	2.5	             & 2.50      & 1.17   & 1000	& 3.89$\times 10^{-4}$	&	2	\\	

2097: 21318+5631 &	0	& ...        &	1.36	             &	         &	  &	700	& 1.10$\times 10^{-4}$	&	3	\\	

2098: 23166+1655 &	0	& 0.1	     &	1.12	             &	         &	  &	650	& 5.50$\times 10^{-7}$	&	4	\\	

2099: 	   &	30	& ...        &	7.85	             &	         &	  &	1000	& 3.30$\times 10^{-5}$	&	1	\\	

2100: 	   &	0	& ...        &	1.19	             &	         &	  &	650	& 1.00$\times 10^{-4}$	&	3	\\	

2101: \hline

2102: \end{tabular}

2103: \begin{tabular}{p{5.8in}}

2104: References 1: \citet{Volk1992}; 2: \citet{Volk2000}; 3: \citet{Groenewegen1995}; 4: \citet{Groenewegen1998}\\

2105: $\ast$ originally quoted as a function of distance. Value quoted here assume 

2106: distances from \citet{Groen2002}, listed in Table~\ref{obsparam2}\\

2107: $\dag$ composition assumes remainder dust is carbon. In all but 

2108: \citet{Volk1992} the carbon is amorphous; 

2109: \citet{Volk1992} uses graphitic carbon.\\

2110: Neither \citet{Volk1992} nor \citet{Groenewegen1995} specify the grains sizes 

2111: used in their models\\

2112: \end{tabular}

2113: \end{table}

2114: 

2115: %table2

2116: \begin{table}

2117: \small

2118: \caption{Target List \label{obstable}}

2119: \begin{tabular}{l@{\hspace{1mm}}l@{\hspace{1mm}}c@{\hspace{1mm}}c@{\hspace{2mm}}c@{\hspace{1mm}}c@{\hspace{1mm}}}

2120: \hline

2121: IRAS &Other & R.A. \footnote{%

2122: Note: Units of right ascension are given in hours, minutes and seconds; 

2123: units of declination are degrees, arcminutes and arcseconds.}& 

2124: Decl.  & TDT  & Date of \\

2125: Number&Names& (J2000)&(J2000) & number & Observation \\

2126: \hline

2127: 00210+6221&CGCS 6006& 00 23 51.2& $+$62 38 16.4 &40401901 & 1996 Dec 24\\

2128: 01144+6658&V829 Cas, AFGL 190, CGCS 6017&01 17 51.6 &+67 13 55.4&68800128&1997 Oct 03 \\

2129: 02408+5458&& 02 44 25.2  & $+$55 11 15& 80002504 & 1998 Jan 24 \\

2130: 03313+6058&CGCS 6061&03 35 30.7 &+61 08 47.2&62301907&1997 Jul 31\\

2131: 06582+1507&CGCS 6193& 07 01 08.44 & $+$15 03 39.8&71002102 & 1997 Oct 26\\

2132: 17534$-$3030&AFGL 5416, CGCS 6690& 17 56 33.1 & $-$30 30 47.1&12102004 & 1996 Mar 17\\

2133: 19548+3035&AFGL 2477, CGCS 6851& 21 50 45.0 & $+$53 15 28.0&56100849 & 1997 May 30 \\

2134: 21318+5631&AFGL 5625S, CGCS 6888& 21 33 22.98 & $+$56 44 35.0&11101103 & 1997 Mar 7\\

2135: 22303+5950&CGCS 6906&22 32 12.8 &+60 06 04.1&77900836&1998 Jan 02\\

2136: 23166+1655&LL Peg, AFGL 3068, CGCS 6913& 23 19 12.39 & $+$17 11 35.4&37900867 & 1996 Nov 29\\

2137: \hline

2138: \end{tabular}

2139: \end{table}

2140: 

2141: 

2142: %table3

2143: \begin{table}

2144: \caption{Observed parameters of the target sources \label{obsparam}}

2145: \small

2146: %\begin{tabular}{l@{\hspace{2.5mm}}c@{\hspace{0.5mm}}c@{\hspace{0.5mm}}c@{\hspace{0.5mm}}c@{\hspace{0.5mm}}c@{\hspace{0.5mm}}c@{\hspace{0.5mm}}c@{\hspace{0.5mm}}c}

2147: \begin{tabular}{lcccccccc}

2148: \hline

2149: IRAS       &$T_{\rm BB}$& Feature & Feature$^1$  &FWHM & Equivalent & 9.7$\mu$m$^2$ & 11.3$\mu$m$^3$ & SiC$^4$\\ 

2150: \rb{Number}& \rb{(K)}  &\rb{Barycenter}& \rb{Strength} & &\rb{Width} & Strength  & Strength & position\\ 

2151: \hline

2152: 00210+6221 & 290K & 10.45 & 0.73 & 2.28&0.60 & 0.76& 0.77 & 11.3\\

2153: 01144+6658 & 280K & 10.51 & 0.82 & 2.32&0.35 & 0.88& 0.87 & 11.3\\

2154: 02408+5458 & 250K & 11.28 & 0.73 & 2.20&0.30 & --- & 0.83 & 11.3\\

2155: 03313+6058 & 350K & 10.12 & 0.87 & 1.72&0.17 & 0.90& 0.97 & 10.8\\

2156: 06582+1507 & 320K & 10.25 & 0.91 & 0.70&0.1  & 0.97& 0.94 & 11.0\\

2157: 17534$-$3030 & 280K & 10.59 & 0.87 & 0.80&0.30 & 0.92& 0.83 & 11.3\\

2158: 19548+3035 & 295K & 10.24 & 0.87 & 2.39&0.25 & 0.88& 0.93 & 11.0\\

2159: 21318+5631 & 300K & 10.20 & 0.84 & 2.90&0.37 & 0.88& 0.93 & 11.0\\

2160: 22303+5950 & 350K & 10.25 & 0.84 & 1.70&0.25 & 0.86& 0.94 & 11.0\\

2161: 23166+1655 & 290K & 11.42 & 0.86 & 2.00&0.21 & --- & 0.88 & 11.4\\

2162: \hline

2163: \end{tabular}

2164: \begin{tabular}{p{6.5in}}

2165: $^1$ Feature strength is the ``peak''-to-continuum ratio and is measured and 

2166: the barycentric position.\\

2167: $^2$ The 9.7$\mu$m strength is the feature-to-continuum ratio measured at 9.7$\mu$m.\\

2168: $^3$ The 11.3$\mu$m strength is the feature-to-continuum ratio measured at 11.3$\mu$m.\\

2169: $^4$ The SiC position, the the approximate barycentric position that the SiC feature 

2170: would have if the short wavelength side of the absorption were due to silicate 9.7$\mu$m 

2171: absorption.\\

2172: \end{tabular}

2173: \end{table}

2174: 

2175: %table4

2176: \begin{table}

2177: \caption{Compilation of CO mass-loss rates and expansion velocities, together with  distances 

2178: and luminosities of sample stars. \label{obsparam2}}

2179: \begin{tabular}{lcccccc}

2180: \hline

2181: Source Name & $v_{\rm exp}$ &	D & $\dot{M}_{\rm gas}$ & $\dot{M}_{\rm dust}$ & L$_{\star}$ \\

2182: 	    &	km/s	    & kpc & M$_\odot$\,yr$^{-1}$& M$_\odot$\,yr$^{-1}$ & L$_\odot$ \\

2183: \hline

2184: 00210+6221* &   16.7        & 3.97& 3.02$\times 10^{-5}$& ---                 &   1$\times 10^{4}$  \\   

2185: 01144+6658  &	18	    & 2.78& 6.38$\times 10^{-5}$&1.51$\times 10^{-7}$ &1.69$\times 10^{4}$\\

2186: 02408+5458* &	11	    & 5.3 & 1.60$\times 10^{-5}$&3.5 $\times 10^{-7}$ &5.70$\times 10^{3}$\\

2187: 03313+6058  &	13.9	    & 5.24& 2.37$\times 10^{-5}$&1.14$\times 10^{-7}$ &1.31$\times 10^{4}$\\

2188: 06582+1507  &	13.7	    & 4.7 & 1.43$\times 10^{-5}$&1.07$\times 10^{-7}$ &1.32$\times 10^{4}$\\

2189: 17534$-$3030  &	19	    & 2	  & 3.76$\times 10^{-5}$&1.06$\times 10^{-7}$ &1.21$\times 10^{4}$\\

2190: 19548+3035  &	22.3	    & 3.38& 1.14$\times 10^{-4}$&2.15$\times 10^{-7}$ &1.32$\times 10^{4}$\\

2191: 21318+5631  &	19.6	    & 1.77& 7.69$\times 10^{-6}$&1.18$\times 10^{-7}$ &1.24$\times 10^{4}$\\

2192: 22303+5950  &	18.3	    & 3.86& 3.19$\times 10^{-4}$&1.09$\times 10^{-7}$ &1.25$\times 10^{4}$\\

2193: 23166+1655  &	15.1	    & 1	  & 1.44$\times 10^{-5}$&8.27$\times 10^{-8}$ &1.10$\times 10^{4}$\\

2194: \hline

2195: \end{tabular}

2196: \begin{tabular}{p{5.5in}}

2197: All data are from \citet{Groen2002} except those marked with $\ast$. 

2198: IRAS\,02408+5458 data come from \citet{Groen1999}.

2199: IRAS\,00210+6221 data are compiled from 

2200: \citet{Volk2000} \citep[$\dot{M}$, which also convolves the distance from ][]{Groen2002}; and

2201: \citet{Volk1992} ($v_{\rm exp}$).

2202: \end{tabular}

2203: \end{table}

2204: 

2205: 

2206: %table 5

2207: \begin{table}

2208: \caption{Model and Derived Parameters \label{tabfit1}}

2209: %\begin{tabular}{lcccp{1.85cm}@{\hspace{7.5mm}}r}

2210: \begin{tabular}{lrrrrcc}

2211: \hline

2212: IRAS  & T$_\star$ &T$_{\rm inner}$&$R_{\rm out}/R_{\rm in}$ & $\tau_{10\mu \rm m}$ &\multicolumn{2}{c}{Composition}\\

2213: Number&           &               &                         &                      &  SiC\% & Graphite \% \\

2214: \hline

2215: \multicolumn{7}{c}{MRN grain size distribution}	\\

2216: 00210+6221	&	3000	&	1800	&	10	&	6	&	10	&	90	\\

2217: 01144+6658	&	3000	&	1800	&	15	&	6.5	&	10	&	90	\\

2218: 02408+5458	&	3000	&	1800	&	20	&	12	&	3	&	97	\\

2219: 03313+6058	&	3000	&	1800	&	15	&	4	&	10	&	90	\\

2220: 06582+1507	&	3000	&	1800	&	10	&	6	&	5	&	95	\\

2221: 17534-3030	&	3000	&	1800	&		&		&		&		\\

2222: 19548+3035	&	3000	&	1800	&	15	&	6.5	&	5	&	95	\\

2223: 21318+5631	&	3000	&	1800	&	10	&	8	&	3	&	97	\\

2224: 22303+5950	&	3000	&	1800	&	15	&	4	&	10	&	90	\\

2225: 23166+1655	&	3000	&	1800	&	15	&	7.5	&	3	&	97	\\

2226: \hline

2227: \multicolumn{7}{c}{Meteoritic grain size distribution}	\\

2228: 00210+6221	&	3000	&	1800	&	10	&	6	&	30	&	70	\\

2229: 01144+6658	&	3000	&	1800	&	10	&	7.5	&	30	&	70	\\

2230: 02408+5458	&	3000	&	1800	&	100	&	8.5	&	12	&	88	\\

2231: 03313+6058	&	3000	&	1800	&	500	&	3.25	&	40	&	60	\\

2232: 06582+1507	&	3000	&	1800	&	10	&	6	&	20	&	80	\\

2233: 17534-3030	&	3000	&	1800	&	20	&	5	&	35	&	65	\\

2234: 19548+3035	&	3000	&	1800	&	100$^\ast$&	4.75	&	25	&	75	\\

2235: 21318+5631	&	3000	&	1800	&	100	&	5	&	20	&	80	\\

2236: 22303+5950	&	3000	&	1800	&	500	&	3.25	&	40	&	60	\\

2237: 23166+1655	&	3000	&	1800	&	50	&	5	&	25	&	75	\\

2238: \hline

2239: \end{tabular}

2240: \begin{tabular}{p{16.8cm}}

2241: $\ast$ Models with $R_{\rm out}/R_{\rm in} > 100$ can be accommodated because the data beyond 26$\mu$m is poor and ignored, but based on 

2242: the similarity of this source to IRAS\,21318+5631, we assume this is a good upper limit.\\

2243: \end{tabular}

2244: \end{table}

2245: 

2246: \begin{table}

2247: \caption{Size and age of the dust shells \label{timescaletab}}

2248: \begin{tabular}{lcccccc}

2249: \hline

2250:         & \multicolumn{3}{c}{MRN size distribution}  & \multicolumn{3}{c}{Meteoritic size distribution}\\

2251: IRAS    & R$_{\rm in}$ &R$_{\rm out}$ & Age    & R$_{\rm in}$ &R$_{\rm out}$ & Age    \\

2252: Number  & ($10^{14}$cm)&($10^{14}$cm) & (yrs)  & ($10^{14}$cm)&($10^{14}$cm) & (yrs)  \\

2253: \hline

2254: 00210	& 2.68	&	26.8	&       50.9  &	2.78	& 	2.78	&	52.8	\\

2255: 01144	& 2.74	&	41.1	&	72.4  &	3.08	& 	30.8	&	54.2	\\

2256: 02408	& 4.11	&	82.2	&	236.8 &	2.99	& 	299	&	861.3	\\

2257: 03313	& 2.17	&	32.6	&	74.2  &	2.10	& 	1050	&	2393.7	\\

2258: 06582	& 2.73	&	27.3	&	63.1  &	2.72	& 	27.2	&	62.9	\\

2259: 17534	& 2.51	&	50.2	&	83.7  &	2.51	& 	50.2	&	83.7	\\

2260: 19548	& 2.78	&	41.7	&	59.3  &	2.38	& 	238 	&	338.2	\\

2261: 21318	& 3.18	&	31.8	&	51.4  &	2.41	& 	241	&	389.6	\\

2262: 22303	& 2.24	&	33.6	&	58.2  &	2.10	& 	1050	&	1818.2	\\

2263: 23166	& 2.88	&	43.2	&	90.7  &	2.45	& 	123	&	257.1	\\

2264: \hline

2265: \end{tabular}

2266: \begin{tabular}{p{6in}}

2267: Calculation of R$_{\rm out}$ is based on the model parameters listed in Table~\ref{tabfit1}. R$_{\rm in}$ comes from the model output. The calculation of the age of the dust shells is then done using these data and the observed expansion velocities listed in Table~\ref{obsparam2}

2268: \end{tabular}

2269: \end{table}

2270: 

2271: \end{document}

2272: